Nanopore Technology - Methods and Protocols-Springer US - Humana (2021)

Methods in
Molecular Biology 2186
Monifa A.V. Fahie Editor
Nanopore
Technology
Methods and Protocols
METHODS IN MOLECULAR BIOLOGY
Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, UK
For further volumes:

http://www.springer.com/series/7651
For over 35 years, biological scientists have come to rely on the research protocols and
methodologies in the critically acclaimed Methods in Molecular Biology series. The series was
the first to introduce the step-by-step protocols approach that has become the standard in all
biomedical protocol publishing. Each protocol is provided in readily-reproducible step-by-
step fashion, opening with an introductory overview, a list of the materials and reagents
needed to complete the experiment, and followed by a detailed procedure that is supported
with a helpful notes section offering tips and tricks of the trade as well as troubleshooting
advice. These hallmark features were introduced by series editor Dr. John Walker and
constitute the key ingredient in each and every volume of the Methods in Molecular Biology
series. Tested and trusted, comprehensive and reliable, all protocols from the series are
indexed in PubMed.
Nanopore Technology
Methods and Protocols
Edited by
Monifa A. V. Fahie
Molecular and Cellular Biology Program, University of Massachusetts Amherst, Amherst, MA, USA
Editor
Monifa A. V. Fahie
Molecular and Cellular Biology
Program
University of Massachusetts Amherst
Amherst, MA, USA
ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-0716-0805-0 ISBN 978-1-0716-0806-7 (eBook)
https://doi.org/10.1007/978-1-0716-0806-7
© Springer Science+Business Media, LLC, part of Springer Nature 2021

All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of
translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical
way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This Humana imprint is published by the registered company Springer Science+Business Media, LLC part of Springer
Nature.
The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.
Preface
The field of nanopore technology for single molecule sensing has garnered much attention
with its robust application in DNA sequencing. Much effort over the last two decades in
nanopore design and computational analysis has made nanopore-based DNA sequencing a
real-world (and outer space) technique, not only used in the confines of a research labora-
tory. Nanopore sensing, however, is also used for the analysis of other biological molecules
such as RNA and proteins as well as chemicals such as metabolites, toxins, or drugs.
In recent years, interest in protein analysis and peptide sequencing has gained traction in
and among the members of the nanopore technology field. However, unlike DNA and RNA
sequencing, precise and accurate protein analysis has several obstacles. Therefore, this
volume primarily focuses on a few of the single molecule methods and nanopores developed
for the specific and selective characterization of protein analytes.
Nanopores are nanometer-sized holes. In nature, small holes exist as membrane chan-
nels that act as gatekeepers, allowing or disallowing molecules to enter or exit the cell.
Membrane channels exist in all forms of life. It is here where the field of nanopore
technology gained inspiration from as far back as in the 1980s, where the intrinsic activity
of acetylcholine receptors were being studied or the nonspecific activity of voltage-
dependent anion channels (VDAC) against synthetic molecules were studied in model
lipid bilayers [1, 2]. Not only are nanopores based on protein channels but also solid-state
materials and synthetic DNA origami. Solid-state nanopores, although not represented in
this book, are a class of nanopores that have shown great promise in protein analysis [3–6].
The advantage of synthetic nanopores is the option of size tunability while the advantage of
purified protein nanopores is consistent and precise pore characteristics such as diameter and
asymmetric charge distribution. Hybrid nanopores combine both biological and synthetic
nanopores and are a developing technique that can push the boundaries of protein analysis
by single nanopore sensing technology.
Nanopore sensing is performed in either one of three main strategies. Firstly, early
nanopore research focused heavily on the passage of analytes through the nanopore’s
lumen, a process called translocation. Molecules upon entering the nanopore lumen
would displace water and, therefore, block ion movement through the pore, resulting in a
measurable decrease in ionic current. The size of the current blockages is congruent with the
molecular weight or size of the translocating molecule. Currently, this detection method is
primarily used by solid-state nanopores but also used by protein nanopores, as exampled in
Chapters 3–6, 10, 11, and 13–15.
In recent years, other methods of analyte detection by protein nanopores have become
popular. For example, analyte trapping, which can be considered as incomplete transloca-
tion, has only been successfully performed with protein nanopores with asymmetric lumen
characteristics and constriction sites significantly smaller than the rest of the lumen, i.e.,
goblet-shaped nanopores. One example, the cytolysin A (ClyA) nanopore has been used to
trap protein analytes up to ~40 kDa in size [7, 8]. The protein analytes’ behavior inside the
ClyA nanopore can give information about its compactness or rigidity and can also report on
conformational changes due to ligand–protein interactions.
Finally, protein nanopores can also detect analytes that do not enter its lumen but those
that interact with an external binding site that is either intrinsic to the nanopore or that
v
vi Preface
which has been engineered. Chapters 7–9 give detailed methodologies for this type of
nanopore detection system.
This book, with its focus on nanopore technology and biomolecule characterization,
will hold the interest of the biophysicists, biochemists, bioengineers, and molecular biolo-
gists among us. This book’s contributions come from a collective of 39 scientists from all
over the world, working diligently in this growing field of nanopore sensing application.
This book includes 16 chapters which are grouped into four parts:
Part I consists of four chapters which lay the framework for the foundation of nanopore
technology: nanopore design and nanopore production.
Part II consists of seven chapters discussing various biological nanopores, nanopore
engineering, and their uses in single molecule sensing. In particular, the sensing of
proteins and one example of sugar sensing are outlined.
Part III consists of two chapters outlining computational methods to study intrinsic
nanopore behavior as well as the formulations for characterizing the specific trans-
location activity of a vesicle particle through a nanopore.
Part IV consists of three chapters specifically detailing the use of the technique droplet
interface bilayer (DIB) in nanopore and membrane biophysical studies.
The editor wishes to thank all the contributors for their dedication and patience during
the creation of this book.
Amherst, MA, USA Monifa A. V. Fahie
References
1. Suarez-Isla BA, Wan K, Lindstrom J, Montal M (1983) Single-channel recordings from purified
acetylcholine receptors reconstituted in bilayers formed at the tip of patch pipets. Biochemistry
22:2319–2323
2. Zimmerberg J, Parsegian VA (1986) Polymer inaccessible volume changes during opening and closing
of a voltage-dependent ionic channel. Nature 323:36–39
3. Han A, Schürmann G, Mondin G, Bitterli RA, Hegelbach NG, De Rooij NF, Staufer U (2006) Sensing
protein molecules using nanofabricated pores. Appl Phys Lett 88:1–4
4. Fologea D, Ledden B, McNabb DS, Li J (2007) Electrical characterization of protein molecules by a
solid-state nanopore. Appl Phys Lett 91:53901-1–53901-3
5. Talaga DS, Li J (2009) Single-molecule protein unfolding in solid state nanopores. J Am Chem Soc
131:9287–9297
6. Nir I, Huttner D, Meller A (2015) Direct sensing and discrimination among ubiquitin and ubiquitin
chains using solid-state nanopores. Biophys J 108:2340–2349
7. Meervelt V Van, Soskine M, Maglia G (2014) Detection of two isomeric binding configurations in a
protein aptamer complex with a biological nanopore. ACS Nano 8:12826–12835
8. Willems K, Ruić D, Biesemans A, Galenkamp NS, Van Dorpe P, Maglia G (2019) Engineering and
modeling the electrophoretic trapping of a single protein inside a nanopore. ACS Nano 13:9980–9992
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
PART I DESIGN AND PREPARATION OF BIOLOGICAL BASED NANOPORES

1 Preparation of Fragaceatoxin C (FraC) Nanopores . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Natalie Lisa Mutter, Gang Huang, Nieck Jordy van der Heide,
Florian Leonardus Rudolfus Lucas, Nicole Stéphanie Galenkamp,
Giovanni Maglia, and Carsten Wloka
2 Preparation of Cytolysin A (ClyA) Nanopores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Nicole Stéphanie Galenkamp, Veerle Van Meervelt, Natalie Lisa Mutter,
Nieck Jordy van der Heide, Carsten Wloka, and Giovanni Maglia
3 Building Synthetic Transmembrane Peptide Pores . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Kozhinjampara R. Mahendran
4 Design and Assembly of Membrane-Spanning DNA Nanopores . . . . . . . . . . . . . . 33
Kerstin Göpfrich, Alexander Ohmann, and Ulrich F. Keyser
PART II SINGLE MOLECULE DETECTION AND ANALYSIS OF PROTEIN ANALYTES
5 Determining the Orientation of Porins in Planar Lipid Bilayers . . . . . . . . . . . . . . . 51

Sandra A. Ionescu, Sejeong Lee, and Hagan Bayley
6 Revelation of Function and Inhibition of Wza Through
Single-Channel Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Lingbing Kong
7 Protein Analyte Sensing with an Outer Membrane Protein G
(OmpG) Nanopore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Monifa A. V. Fahie, Bib Yang, Christina M. Chisholm, and Min Chen
8 Nanopore Enzymology to Study Protein Kinases and Their
Inhibition by Small Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Leon Harrington, Leila T. Alexander, Stefan Knapp,
and Hagan Bayley
9 A Selective Activity-Based Approach for Analysis of Enzymes
with an OmpG Nanopore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Monifa A. V. Fahie, Bach Pham, Fanjun Li, and Min Chen
10 Oligonucleotide-Directed Protein Threading Through a
Rigid Nanopore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Garbiñe Celaya and David Rodriguez-Larrea
11 Unfolding and Translocation of Proteins Through an
Alpha-Hemolysin Nanopore by ClpXP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Jeff Nivala, Logan Mulroney, Qing Luan, Robin Abu-Shumays,
and Mark Akeson
vii
viii Contents
PART III COMPUTATIONAL ANALYSIS OF NANOPORE BEHAVIOR
12 Simulation of pH-Dependent, Loop-Based Membrane Protein

Gating Using Pretzel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Alan Perez-Rathke, Monifa A. V. Fahie, Christina M. Chisholm,
Min Chen, and Jie Liang
13 Free Energy Minimization for Vesicle Translocation Through
a Narrow Pore. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Hamid R. Shojaei, Ahad Khaleghi Ardabili, and
Murrugappan Muthukumar
PART IV DROPLET INTERFACE BILAYER SYSTEM

14 Single Ion-Channel Analysis in Droplet Interface Bilayer . . . . . . . . . . . . . . . . . . . . 187
Arash Manafirad
15 Continuous and Rapid Solution Exchange in a Lipid Bilayer
Perfusion System Based on Droplet-Interface Bilayer. . . . . . . . . . . . . . . . . . . . . . . . 197
En-Hsin Lee
16 Protein Transport Studied by a Model Asymmetric Membrane
Army Arranged in a Dimple Chip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Xin Li and Min Chen
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Contributors
ROBIN ABU-SHUMAYS • UC Santa Cruz Genomics Institute, University of California, Santa

Cruz, Santa Cruz, CA, USA
MARK AKESON • UC Santa Cruz Genomics Institute, University of California, Santa Cruz,
Santa Cruz, CA, USA
LEILA T. ALEXANDER • Nuffield Department of Clinical Medicine, Structural Genomics
Consortium and Target Discovery Institute, University of Oxford, Oxford, UK;
Personalized Health Informatics, SIB Swiss Institute of Bioinformatics, Basel, Switzerland
HAGAN BAYLEY • Department of Chemistry, University of Oxford, Oxford, UK
GARBIÑE CELAYA • Department of Biochemistry and Molecular Biology (UPV/EHU),
Biofisika Institute (CSIC, UPV/EHU), Leioa, Spain
MIN CHEN • Department of Chemistry, University of Massachusetts Amherst, Amherst, MA,
USA
CHRISTINA M. CHISHOLM • Molecular and Cellular Biology Program, University of
Massachusetts Amherst, Amherst, MA, USA
MONIFA A. V. FAHIE • Molecular and Cellular Biology Program, University of Massachusetts
Amherst, Amherst, MA, USA
NICOLE STÉPHANIE GALENKAMP • Groningen Biomolecular Sciences and Biotechnology
Institute, University of Groningen, Groningen, The Netherlands
KERSTIN GÖPFRICH • Cavendish Laboratory, University of Cambridge, Cambridge, UK
LEON HARRINGTON • Department of Chemistry, University of Oxford, Oxford, UK; Max
Planck Institute of Biochemistry, Martinsried, Germany
GANG HUANG • Groningen Biomolecular Sciences and Biotechnology Institute, University of
Groningen, Groningen, The Netherlands
SANDRA A. IONESCU • Department of Chemistry, University of Oxford, Oxford, UK
ULRICH F. KEYSER • Cavendish Laboratory, University of Cambridge, Cambridge, UK
AHAD KHALEGHI ARDABILI • School of Engineering and Natural Sciences, Altınbaş
University, Istanbul, Turkey
STEFAN KNAPP • Nuffield Department of Clinical Medicine, Structural Genomics
Consortium and Target Discovery Institute, University of Oxford, Oxford, UK; Institute
for Pharmaceutical Chemistry, Structural Genomics Consortium, and Buchmann Institute
for Molecular Life Sciences, Johann Wolfgang Goethe-University, Frankfurt am Main,
Germany
LINGBING KONG • International Institute of Rare Sugar Research and Education,
Department of Applied Biological Science, Faculty of Agriculture, Kagawa University,
Miki, Kagawa, Japan
EN-HSIN LEE • Department of Chemistry, University of Massachusetts Amherst, Amherst,
MA, USA
SEJEONG LEE • Department of Chemistry, University of Oxford, Oxford, UK
JIE LIANG • Department of Bioengineering, University of Illinois at Chicago, Chicago, IL,
USA
FANJUN LI • Department of Chemistry, University of Massachusetts Amherst, Amherst, MA,
USA
XIN LI • Department of Chemistry, University of Massachusetts Amherst, Amherst, MA, USA
ix
x Contributors
QING LUAN • Department of Chemistry and Biochemistry, University of Notre Dame, Notre
Dame, IN, USA
FLORIAN LEONARDUS RUDOLFUS LUCAS • Groningen Biomolecular Sciences and Biotechnology
Institute, University of Groningen, Groningen, The Netherlands
GIOVANNI MAGLIA • Groningen Biomolecular Sciences and Biotechnology Institute,
University of Groningen, Groningen, The Netherlands
KOZHINJAMPARA R. MAHENDRAN • Membrane Biology Laboratory, Interdisciplinary Research
Program, Rajiv Gandhi Centre for Biotechnology, Thiruvananthapuram, India
ARASH MANAFIRAD • Department of Physics, University of Massachusetts Amherst, Amherst,
MA, USA; Department of Chemistry, University of Massachusetts Amherst, Amherst, MA,
USA
LOGAN MULRONEY • UC Santa Cruz Genomics Institute, University of California, Santa
Cruz, Santa Cruz, CA, USA
MURRUGAPPAN MUTHUKUMAR • Department of Polymer Science and Engineering, University
of Massachusetts Amherst, Amherst, MA, USA
NATALIE LISA MUTTER • Groningen Biomolecular Sciences and Biotechnology Institute,
JEFF NIVALA • Paul G. Allen School of Computer Science and Engineering, University of
Washington, Seattle, WA, USA
ALEXANDER OHMANN • Cavendish Laboratory, University of Cambridge, Cambridge, UK
ALAN PEREZ-RATHKE • Department of Bioengineering, University of Illinois at Chicago,
Chicago, IL, USA
BACH PHAM • Department of Chemistry, University of Massachusetts Amherst, Amherst, MA,
USA
DAVID RODRIGUEZ-LARREA • Department of Biochemistry and Molecular Biology (UPV/
EHU), Biofisika Institute (CSIC, UPV/EHU), Leioa, Spain
HAMID R. SHOJAEI • Department of Polymer Science and Engineering, University of
Massachusetts Amherst, Amherst, MA, USA
NIECK JORDY VAN DER HEIDE • Groningen Biomolecular Sciences and Biotechnology Institute,
VEERLE VAN MEERVELT • Groningen Biomolecular Sciences and Biotechnology Institute,
CARSTEN WLOKA • Groningen Biomolecular Sciences and Biotechnology Institute, University
of Groningen, Groningen, The Netherlands
BIB YANG • Department of Chemistry, University of Massachusetts Amherst, Amherst, MA,
USA
Part I
Design and Preparation of Biological Based Nanopores

Chapter 1
Preparation of Fragaceatoxin C (FraC) Nanopores

Natalie Lisa Mutter, Gang Huang, Nieck Jordy van der Heide,
Florian Leonardus Rudolfus Lucas, Nicole Stéphanie Galenkamp,
Giovanni Maglia, and Carsten Wloka
Abstract
Biological nanopores are an emerging class of biosensors with high-end precision owing to their reproduc-
ible fabrication at the nanometer scale. Most notably, nanopore-based DNA sequencing applications are
currently being commercialized, while nanopore-based proteomics may become a reality in the near future.
Although membrane proteins often prove to be difficult to purify, we describe a straightforward protocol
for the preparation of Fragaceatoxin C (FraC) nanopores, which may have applications for DNA analysis
and nanopore-based proteomics. Recombinantly expressed FraC nanopores are purified via two rounds of
Ni-NTA affinity chromatography before and after oligomerization on sphingomyelin-containing lipo-
somes. Starting from a plasmid vector containing the FraC gene, our method allows the production of
purified nanopores within a week. Afterward, the FraC nanopores can be stored at +4 C for several months,
or frozen.
Key words Nanotechnology, Nanopore, Porin, Actinoporin, Fragaceatoxin C, FraC, Protein purifi-
cation, Protein oligomerization, Electrophysiology, Artificial bilayers
1 Introduction
The understanding of single ion channels was revolutionized over

four decades ago when it was shown that single channel currents
can be recorded [1]. The small ion current flowing through a single
channel in a planar lipid bilayer [2] has been subsequently used for
single-molecule studies. Single channels (or nanopores) with a
β-barrel comprise toxins such as α-hemolysin (αHL) [3, 4], aero-
lysin [5, 6], Mycobacterium smegmatis porin A (MspA) [7], ferric
hydroxamate uptake component A (FhuA) [8], and outer mem-
brane proteins (OmpG [9–11], OmpF [12], OmpC [13]), and
have been used for the detection of small molecules as well as
for DNA, protein, and peptide analysis. Our laboratory introduced
the α-helical Cytolysin A (ClyA) nanopore from Salmonella typhi
for studies of proteins [14] and DNA [15]. More recently the
Monifa A. V. Fahie (ed.), Nanopore Technology: Methods and Protocols, Methods in Molecular Biology, vol. 2186,
https://doi.org/10.1007/978-1-0716-0806-7_1, © Springer Science+Business Media, LLC, part of Springer Nature 2021
3
4 Natalie Lisa Mutter et al.
Fragaceatoxin C (FraC) nanopore [16] has been introduced in our

lab for peptide detection.
The octameric crystal structure of FraC has been disclosed [17]
after being discovered to be an ingredient of the venom of the
strawberry anemone (Actinia fragacea) [18]. This α-helical pore
forms a relatively narrow (V-shaped) trans entry, which proved to
be useful for the analysis of DNA [19] and physiologically impor-
tant proteins and peptides [16, 20].
Based on previous work [17, 21], we describe here an easy-to-
follow protocol to obtain FraC nanopores. In this protocol,
histidine-tagged FraC monomers are recombinantly expressed in
Escherichia coli and purified using Ni-NTA affinity chromatogra-
phy. Monomers are then assembled into oligomers on
sphingomyelin-containing liposomes. In the final step, liposomes
are solubilized and oligomeric FraC nanopores are purified using a
second round of Ni-NTA affinity chromatography. FraC nanopores
can be obtained within a week and stored at 80 C as well as at
4 C for several months.
2 Materials
All solutions are prepared using deionized water (Millipore). Spe-

cial attention should be paid to the final pH of all solutions, in
particular to the high amount of imidazole-containing buffers
(e.g., elution buffers). All reagents can be prepared and stored at
room temperature (~25 C), unless otherwise noted.
2.1 Transformation 1. Sterile Lysogeny broth (LB).

and Expression of FraC 2. Sterile 2 YT medium.
in Escherichia coli
3. 1 M Isopropyl-β-D-thiogalactopyranoside (IPTG). Store at
BL21 (DE3) Strain
20 C.
4. 100 mg/mL Ampicillin. Store at 20 C.
5. LB-Agar plates containing 1% (w/v) glucose (optional) and
100 μg/mL ampicillin. Store at 4 C.
6. 50 μL aliquot of electrocompetent E. cloni® EXPRESS BL21
(DE3) (Lucigen, stored at 80 C).
7. Fragaceatoxin C pT7-SC1 [22] plasmid DNA containing a
C-terminal His6-tag (75–150 ng/μL DNA), stored at 20 C.
8. LB medium containing 1% (w/v) glucose (500 μL).
9. 100 mL Erlenmeyer flask (starter culture).
10. 1 L Erlenmeyer flask (expression culture).
11. Centrifuge capable of maintaining 4 C and 5000 g.
Preparation of FraC Nanopores 5
12. Spectrophotometer.
13. Shaking incubator capable of maintaining 37 and 20 C as well
as 200 rpm.
14. Electroporation device and electroporation cuvettes.
15. Crushed ice.
16. T-shaped spreader, sterile.
2.2 Purification 1. Lysis buffer: 0.015 M Tris–HCl, pH 7.5, 0.150 M NaCl,

of Fragaceatoxin C 0.02 M imidazole, 1 mM MgCl2, and 4 M urea (see Note 1).
Monomers Store at 4 C for up to 1 month. Freshly prepared buffers are
preferred, however.
2. Wash buffer (WB): 0.015 M Tris–HCl, pH 7.5, 0.150 M NaCl,
0.02 M imidazole. Store at 4 C.
3. Elution buffer (EB): 0.015 M Tris–HCl, pH 7.5, 0.150 M
NaCl, 0.300 M imidazole. Store at 4 C (see Note 2).
4. DNase I.
5. Chicken Lysozyme powder.
6. Sonication device.
7. Falcon™ tubes and compatible shaking/turning device.
8. Gravity flow columns with a volume of 1–2 mL.
9. Ni-NTA beads.
2.3 Preparation 1. Dilution buffer (DB): 0.015 M of Tris–HCl, pH 7.5, 0.150 M

of Sphingomyelin: of NaCl. Store at 4 C.
DPhPC Liposomes 2. Sphingomyelin (Brain, Porcine).
3. 1,2-Diphytanoyl-sn-glycero-3-phosphocholine (DPhPC).
4. Pentane (99%).
5. Ethanol, absolute (99.8%).
6. Round-bottom flask (25 mL) and rotary evaporator (or heat
gun).
7. Sonication water bath.
2.4 Oligomerization 1. 5% (w/v) N,N-Dimethyldodecylamine-N-oxide (LDAO).

of Fragaceatoxin C Store at 20 C (room temperature is also acceptable).
Monomers 2. 10% (w/v) Dodecyl-β-D-maltoside (DDM). Store, for exam-
ple, 1 mL aliquots at 20 C.
3. Elution buffer for oligomers (EBO). 0.200 M Na2EDTA,
pH 8.0, 0.02% (w/v) DDM (see Note 3). Store at 4 C.
3 Methods
3.1 Transformation 1. Transform the FraC plasmid DNA into the electrocompetent
and Expression E. cloni® EXPRESS BL21 (DE3) strain for protein expression.
of Fragaceatoxin C Place the electrocompetent cells on ice and add 1 μL plasmid
in an Escherichia coli DNA (75–150 ng/μL DNA) to 50 μL cells. Transfer the cells
BL21 (DE3) Strain to an electroporation cuvette (cooled down on ice) and elec-
troporate (settings: bacteria, 5 ms pulse).
2. Add 500 μL prewarmed LB supplemented with 1% (w/v)
glucose quickly to the cells and incubate for 1 h at 37 C.
3. Plate the transformed cells out on LB plates containing
100 μg/mL ampicillin and 1% (w/v) glucose. Add 300 μL
cells to one plate and 100 μL cells to another plate, to ensure
well-spaced colonies. Incubate the plates at 37 C overnight.
4. The following day (day 2), inoculate 10 mL LB medium sup-
plemented with 100 μg/mL ampicillin with a single colony and
incubate at 37 C and 200 rpm overnight in a 100 mL Erlen-
meyer flask.
5. On day 3, inoculate 200 mL 2 YT medium containing
100 μg/mL ampicillin (expression culture) with 2 mL of the
overnight starter culture. This dilutes the culture 1:100. Incu-
bate the expression culture for 2–3 h at 37 C and 200 rpm
until the optical density at 600 nm wavelength (OD600) is
between 0.6 and 0.8.
6. Transfer the expression culture to 20 C and 200 rpm, add
100 μL IPTG for a final concentration of 0.5 mM IPTG, to
induce protein expression, and continue growth overnight.
3.2 Purification 1. Harvest cells in 50 mL Falcon™ tubes by centrifugation at

of Fragaceatoxin C 5000 g (30 min at 4 C) and discard the supernatant (see
Monomers Note 4).
2. Freeze the bacterial cell pellets for at least 1 h at 80 C (see
Note 5).
3. Solubilize two bacterial cell pellets (100 mL of the expression
culture total) in 20 mL lysis buffer (see Note 6).
4. Add 0.2 units/mL of DNase I and 20 μg/mL of lysozyme to
each resuspended pellet (see Note 7). Incubate the pellets for
30 min at room temperature (~25 C) in a shaking device.
5. Sonicate by putting the tip of the sonication device at least 3 cm
beneath the surface of the liquid (see Note 8). Sonicate on ice at
30% output power for 3 30 s to break the cells open (see
Note 9).
6. Pellet the cell debris by centrifugation for 30 min at 5000 g
and 4 C. Discard the pellet. The protein is in the supernatant.
7. Wash 100 μL of Ni-NTA beads (~200 μL of Ni-NTA slurry)

with wash buffer (WB), add 500 μL WB to the Ni-NTA beads
and centrifuge for 30 s at 10,000 g, discard the supernatant
and repeat three times (see Note 10).
8. Incubate the supernatant from step 6 with the washed Ni-NTA
beads for 1 h at room temperature (~25 C) on a shaking device
or tube rotator to prevent sinking of Ni-NTA beads.
9. Prewash a gravity flow column with about five column volumes
(~ 1 mL, depending on the column used) WB to check the flow
and equilibrate the column.
10. Add supernatant from step 8 containing the incubated
Ni-NTA beads into the column (see Note 10).
11. Wash the column with 10 mL of WB. After washing, remove
the WB by putting the column in a clean microtube and
centrifuge at 1500 g for 30 s (see Note 11).
12. Elute the monomers from the Ni-NTA beads with five elution
steps of 150 μL EB. Incubate for 5–10 min with EB and
centrifuge (1500 g, 30 s) in order to collect elution fractions
in clean microtubes.
13. Check the protein concentration of the monomers with a
spectrophotometer at wavelength 280 nm (see Note 12). The
experimentally determined extinction coefficient is
1 1
38,688 M cm .
14. Store the monomers at 4 C overnight or for up to a week, or
continue with the next steps (see Note 13).
3.3 Preparation 1. Solubilize 25 mg sphingomyelin and 25 mg DPhPC (1:1) in

of Sphingomyelin: approximately 2 mL pentane and about 200 μL of ethanol in a
DPhPC Liposomes round-bottom flask (see Note 14).
2. Make a film of sphingomyelin:DPhPC on the surface of the
round-bottom flask by rotary evaporation. Keep the flask open
for 30 min to let all solvent evaporate (see Note 15).
3. Solubilize the lipid film in 5 mL dilution buffer (DB) to a final
concentration of 10 mg/mL.
4. Briefly sonicate the lipid film solution in a sonication bath and
freeze-thaw at least one time, freezing at 20 C (see Note 16).
3.4 Oligomerization 1. Add sphingomyelin:DPhPC liposomes to FraC monomers in a

of Fragaceatoxin C mass ratio of 10:1 (lipid:protein), mix thoroughly/vortex, and
Monomers incubate for 30 min at 37 C (see Notes 17 and 18).
2. Solubilize proteoliposomes by adding a final concentration of
0.6% (w/v) LDAO. The solution should turn clear
immediately.
3. Dilute the solubilized mixture 20-fold with dilution buffer

supplemented with 0.02% (w/v) DDM.
4. Wash 50 μL of Ni-NTA beads (100 μL of Ni-NTA suspension)
with WB supplemented with 0.02% (w/v) DDM. Add 500 μL
WB containing 0.02% (w/v) DDM to the Ni-NTA beads,
centrifuge for 30 s at 10,000 g, discard the supernatant,
and repeat this wash three times.
5. Incubate the oligomer mixture with the washed Ni-NTA beads
for 1 h at room temperature (~25 C) while shaking (using an
orbital shaker/tube rotator).
6. Prewash a gravity flow column with 5 column volumes (~ 1
mL) WB supplemented with 0.02% (w/v) DDM.
7. Add the incubated Ni-NTA beads to the column (see Note 10).
8. Wash the FraC oligomers incubated with the Ni-NTA beads
with 10 mL of WB supplemented with 0.02% (w/v) DDM.
Remove excess WB containing 0.02% (w/v) DDM by centrifu-
ging at 1500 g for 30 s (see Note 11).
9. Elute the oligomers from the Ni-NTA beads in one elution step
with 100 μL EBO. Incubate the Ni-NTA beads for 5–10 min
with EBO and centrifuge (1500 g, 30 s) to collect the elution
in a clean microtube (see Note 19).
10. Oligomers can be stored for several months at 4 C. Store
oligomers at 80 C for longer periods (see Note 20).
11. Oligomers can be used in electrophysiology as described
before [23].
4 Notes
1. Concentrated HCl can be used first to get close to the required

pH. Then, use less concentrated HCl to avoid the pH getting
below the desired value.
2. When stored at 4 C, the buffer can be used for a longer period
of time.
3. A pH of 8.0 is required to dissolve the Na2EDTA. The NaOH
powder should be used to bring the pH close to the required
pH. Use a lower-concentrated NaOH solution (1 M) to avoid
the pH getting above the required value.
4. Before centrifugation, balance the tubes on a weighing scale to
a difference of less than 0.1 g. Use buffer or transfer cell culture
from one tube to another in order to balance.
5. Pellets frozen at 80 C can be kept for long-term storage,
allowing purification at a later time.
6. Pellets can be combined in this step in a ratio of 1 mL lysis

buffer per 5 mL bacterial cells.
7. The DNase I is added to break down DNA and the lysozyme is
added to break down the bacterial cell wall.
8. Consult manufacturer instructions on how to operate the son-
ication device.
9. It is also possible to sonicate the cells for 2 min continuously on
ice. Do not overheat the sample.
10. Ni-NTA beads settle to the bottom; therefore make sure you
resuspend the Ni-NTA beads before pipetting.
11. After spinning the Ni-NTA beads to dryness, add EB directly.
12. Expect a yield of about 0.5 mg/mL or less from 100 mL
expression culture.
13. Freezing of monomers results in precipitation of the proteins.
14. The solvents should evaporate relatively fast, so make sure the
pentane and ethanol you use contain little water.
15. A rotary evaporator can be used. However, manual turning and
heat application to evaporate solvents, using a heat gun for
example, will achieve the same result. It is crucial that all
solvent is evaporated.
16. Freeze-thawing may be repeated many times. It creates fewer
multilamellar liposomes; thus, they will be more uniform.
17. To yield smaller FraC pores [20], it is also possible to use
sphingomyelin-containing liposomes in a mass ratio of 1:1
instead.
18. Incubation for up to 1 h does not influence the
oligomerization.
19. It is also possible to use a different elution buffer containing
imidazole instead of EDTA. Add 100 μL DB supplemented
with 0.02% (w/v) DDM and 1 M imidazole. Adjust the pH of
the solution to 8.0. Incubate for 5–20 min, and centrifuge
afterward to collect the elution.
20. FraC nanopores can be stored at 4 C for several months. It is
possible to store oligomers at 80 C. After thawing, store at
4 C.
References
1. Neher E, Sakmann B (1976) Single-channel Ca2+ binding reactions. J Gen Physiol

currents recorded from membrane of dener- 82:511–542
vated frog muscle fibres. Nature 260:619–621 3. Division B, Biology C, Cruz S (1996) Charac-
2. Moczydlowski E, Latorre R (1983) Gating terization of individual polynucleotide mole-
kinetics of Ca2+-activated K+ channels from cules using a membrane channel. Proc Natl
rat muscle incorporated into planar lipid Acad Sci 93:13770–13773
bilayers. Evidence for two voltage-dependent
4. Gu L-Q, Braha O, Conlan S, Cheley S, Bayley 14. Wloka C, Van Meervelt V, Van Gelder D,
H (1999) Stochastic sensing of organic ana- Danda N, Jager N, Williams CP, Maglia G
lytes by a pore-forming protein containing (2017) Label-free and real-time detection of
amolecular adapter. Nature 398:686–690 protein ubiquitination with a biological nano-
5. Cao C, Ying YL, Hu ZL, Liao DF, Tian H, pore. ACS Nano 11:4387–4394
Long YT (2016) Discrimination of oligonu- 15. Franceschini L, Brouns T, Willems K,
cleotides of different lengths with a wild-type Carlon E, Maglia G (2016) DNA translocation
aerolysin nanopore. Nat Nanotechnol through nanopores at physiological ionic
11:713–718 strengths requires precise nanoscale engineer-
6. Piguet F, Ouldali H, Pastoriza-Gallego M, ing. ACS Nano 10:8394–8402
Manivet P, Pelta J, Oukhaled A (2018) Identi- 16. Huang G, Willems K, Soskine M, Wloka C,
fication of single amino acid differences in uni- Maglia G (2017) Electro-osmotic capture and
formly charged homopolymeric peptides with ionic discrimination of peptide and protein bio-
aerolysin nanopore. Nat Commun 9:966 markers with FraC nanopores. Nat Commun
7. Manrao EA, Derrington IM, Laszlo AH, Lang- 8:935
ford KW, Hopper MK, Gillgren N, 17. Tanaka K, Caaveiro JMM, Morante K, Gonzá-
Pavlenok M, Niederweis M, Gundlach JH lez-Mañas JM, Tsumoto K (2015) Structural
(2012) Reading DNA at single-nucleotide res- basis for self-assembly of a cytolytic pore lined
olution with a mutant MspA nanopore and by protein and lipid. Nat Commun 6:6337
phi29 DNA polymerase. Nat Biotechnol 18. Bellomio A, Morante K, Barlič A, Gutiérrez-
30:349–353 Aguirre I, Viguera AR, González-Mañas JM
8. Mohammad MM, Iyer R, Howard KR, McPike (2009) Purification, cloning and characteriza-
MP, Borer PN, Movileanu L (2012) Engineer- tion of fragaceatoxin C, a novel actinoporin
ing a rigid protein tunnel for biomolecular from the sea anemone Actinia fragacea. Toxi-
detection. J Am Chem Soc 134:9521–9531 con 54:869–880
9. Fahie MA, Yang B, Mullis M, Holden MA, 19. Wloka C, Mutter NL, Soskine M, Maglia G
Chen M (2015) Selective detection of protein (2016) Alpha-helical fragaceatoxin C nanopore
homologues in serum using an OmpG nano- engineered for double-stranded and single-
pore. Anal Chem 87:11143–11149 stranded nucleic acid analysis. Angew Chem
10. Fahie MA, Yang B, Pham B, Chen M (2016) Int Ed Engl 55:12494–12498
Tuning the selectivity and sensitivity of an 20. Huang G, Voet A, Maglia G (2019) FraC
OmpG nanopore sensor by adjusting ligand nanopores with adjustable diameter identify
tether length. ACS Sensors 1:614–622 the mass of opposite-charge peptides with
11. Kahlstatt J, Reiß P, Halbritter T, Essen LO, 44 dalton resolution. Nat Commun 10:835
Koert U, Heckel A (2018) A light-triggered 21. Tanaka K, Caaveiro JMM, Tsumoto K (2015)
transmembrane porin. Chem Commun Bidirectional transformation of a metamorphic
54:9623–9626 protein between the water-soluble and trans-
12. Ghai I, Bajaj H, Bafna JA, El Damrany Hussein membrane native states. Biochemistry
HA, Winterhalter M, Wagner R (2018) Ampi- 54:6863–6866
cillin permeation across OmpF, the major 22. Miles G, Cheley S, Braha O, Bayley H (2001)
outer-membrane channel in Escherichia coli. J The staphylococcal leukocidin bicomponent
Biol Chem 293:7030–7037 toxin forms large ionic channels. Biochemistry
13. Prajapati JD, Solano CJF, Winterhalter M, 40:8514–8522
Kleinekathöfer U (2018) Enrofloxacin perme- 23. Maglia G, Heron AJ, Stoddart D, Japrung D,
ation pathways across the porin OmpC. J Phys Bayley H (2010) Analysis of single nucleic acid
Chem B 122:1417–1426 molecules with protein nanopores. Methods
Enzymol 475:591–623
Chapter 2
Preparation of Cytolysin A (ClyA) Nanopores

Nicole Stéphanie Galenkamp, Veerle Van Meervelt, Natalie Lisa Mutter,
Nieck Jordy van der Heide, Carsten Wloka, and Giovanni Maglia
Abstract
The ionic currents passing through nanopores can be used to sequence DNA and identify molecules at the
single-molecule level. Recently, researchers have started using nanopores for the detection and analysis of
proteins, providing a new platform for single-molecule enzymology studies and more efficient biomolecular
sensing applications. For this approach, the homo-oligomeric Cytolysin A (ClyA) nanopore has been
demonstrated as a powerful tool. Here, we describe a simple protocol allowing the production of ClyA
nanopores. Monomers of ClyA are expressed in Escherichia coli and oligomerized in the presence of
detergent. Subsequently, different oligomer variants are electrophoretically resolved and stored in a gel
matrix for long-term use.
Key words Nanopores, Cytolysin A, ClyA, Protein purification, Membrane protein, Protein assembly,
Pore-forming toxin, Single-molecule, Electrophysiology, Protein trapping
1 Introduction
In a typical nanopore experiment, a potential is applied across a

nanopore reconstituted into a lipid bilayer. The resulting ionic
current flowing through nanopores provides the signal by which
molecules are recognized, studied, or sequenced. Over the past
20 years, nanopores have been used to sequence DNA, recognize
small molecules, and sample chemical reactions at the single-mole-
cule level [1–9]. Advantages include high resolution on a micro- to
millisecond timescale [10], and the ability to observe native single
molecules with no intrinsic limitation of molecular size or observa-
tion time. Because of their single-molecule nature, nanopore
experiments can unravel processes and transient intermediates nor-
mally hidden in ensemble measurements [11, 12].
Single-molecule analysis with Cytolysin A nanopores has been
introduced 7 years ago by our laboratory [13]. The crystal structure
of the nanopore revealed a 13 nm long nanopore comprising
12 identical protomers [14], with a 5.5 nm cis entry and 3.3 nm
11
12 Nicole Stéphanie Galenkamp et al.
trans entry, which allows the entry of proteins or dsDNA

[15]. Under a negative applied potential, folded proteins between
~20 and ~40 kDa added to the cis side of a dodecameric nanopore
can be accommodated inside its lumen exerted by the electro-
osmotic flow [9, 16–18]. Importantly, the narrower trans entry
of the nanopore (3.3 nm) allows trapping of proteins without the
need of labeling, immobilization, or chemical modification [9, 17–
19]. Furthermore, the ionic current is large enough to permit
working at physiological salt conditions (150 mM), which is impor-
tant when sampling protein–protein, protein–DNA interactions
[20] or dsDNA [15].
Inside the nanopore, proteins can be identified by their specific
ionic current blockades. Notably, the asymmetric environment of
the nanopore lumen also allowed resolving isomeric protein–DNA
interactions [20]. Furthermore, the ionic current across the nano-
pore was exquisitely sensitive to small changes in the structure of
the trapped protein, which in turn allowed sampling protein con-
formational changes, due to binding of substrates [9, 17]. Hence,
ClyA nanopores can be used in single-molecule enzymology studies
[21], while the ability of identifying ligand binding to internalized
protein has been exploited to quantify metabolites in bodily
fluids [18].
In our method, we describe the recombinant expression of
histidine-tagged ClyA monomers in Escherichia coli, purification
of the monomers using Ni-NTA affinity chromatography, and
assembly into oligomers by addition of detergent. Because ClyA
monomers are water-soluble and assemble into nanopores, only
upon addition of a detergent (or liposomes) large amounts of
nanopores can easily be obtained. Finally, we describe how to
resolve the ClyA oligomers by separation on blue-native polyacryl-
amide gel electrophoresis. Within 1 week ClyA nanopores can be
obtained and the nanopores can be stored long-term at 4 or 20 C
when embedded within the gel matrix.
2 Materials
Deionized water is used for all solutions and care is given to the final
pH of the solution. Unless otherwise noted, reagents are prepared
and stored at 25 C.
2.1 Transformation 1. 1 L Lysogeny broth (LB). Weigh 10 g of bacto tryptone, 5 g of

and Expression yeast extract, and 10 g of NaCl and transfer to a 1.5 or 2 L
of ClyA-AS bottle. Add water (about 800–900 mL), dissolve all compo-
in Escherichia coli nents by mixing, and fill up to 1 L with water. Sterilize by
autoclaving. Store at room temperature.
2. 1 L 2 YT-medium. Weigh 16 g of bacto tryptone, 10 g of
bacto yeast extract, and 5 g of NaCl and transfer to a 1.5 or 2 L
Preparation of ClyA Nanopores 13
bottle. Mix components in 900 mL water, then fill up to 1 L

with water. Sterilize by autoclaving. Store at room temperature.
3. 10 mL of 1 M Isopropyl-β-D-thiogalactopyranoside (IPTG).
Add 2.38 g IPTG to 10 mL water and mix well. Aliquots can be
stored at 20 C.
4. 10 mL of 100 mg/mL (1000) ampicillin. Add 1.0 g ampicil-
lin to 10 mL water and mix well. Aliquots can be stored at
20 C.
5. LB-agar plates: LB-agar plates containing 1% (w/v) glucose
and 100 μg/mL ampicillin which should be stored at 4 C. For
1 L of LB agar: add 15 g of bacto agar to 1.0 L of LB medium
and autoclave. Use a temperature-controlled water bath to cool
the LB medium to 60 C and add 1 mL of 100 mg/mL
ampicillin and 10 g of glucose. In a sterile environment, pour
plates (about 30 mL per dish). Dry the plates for 30 min. Store
at 4 C.
6. 50 μL Electrocompetent E. cloni® EXPRESS BL21 (DE3)
(Lucigen, stored at 80 C).
7. pT7-SC1 [22] plasmid containing the ClyA-AS gene with
C-terminal His6-tag (75–150 ng/μL DNA), stored at
20 C. ClyA-AS contains eight mutations relative to the
Salmonella typhi ClyA-WT: C87A, L99Q, E103G, F166Y,
I203V, C285S, K294R, and H307Y (the H307Y mutation is
in the C-terminal His6-tag added for purification).
8. LB medium with 1% (w/v) glucose (400 μL).
9. 100 mL Erlenmeyer flask for the starter culture.
10. 1 L Erlenmeyer flask for the expression culture.
11. Cooled centrifuge (4 C).
12. Spectrophotometer, such as a Nanodrop.
13. Shaking incubator (37 and 20 C).
14. Electroporation device and electroporation cuvettes.
15. Sterile T-shaped spreader.
16. Ice.
2.2 For Purification 1. 1 L Lysis buffer: 15 mM Tris–HCl, pH 7.5, 150 mM NaCl,

of ClyA-AS 10 mM imidazole, pH 8.0, 1 mM MgCl2. Weigh 1.82 g Tris–
HCl, 8.77 g NaCl, 0.68 g imidazole, and 95.2 mg MgCl2, and
transfer to a 1 L glass bottle. Add about 900 mL water to
dissolve all components. Adjust the pH to 7.5 with HCl. Fill
up with water to 1 L and store at room temperature for up to
2 months. Fresh buffers are preferred, however. If the buffer is
stored for a period of time, the pH should be checked regularly.
2. 1 L Wash buffer (WB): 10 mM imidazole, 150 mM NaCl,

15 mM Tris–HCl, pH 7.5. Weigh 1.82 g Tris–HCl, 8.77 g
NaCl, and 0.68 g imidazole, and transfer to a 1 L glass bottle.
Add about 900 mL water and mix until all components are
dissolved. Adjust the pH to 7.5 with HCl. Fill up to 1 L with
water and store at 4 C.
3. 100 mL Elution buffer (EB): 150 mM NaCl, 15 mM Tris–
HCl, pH 7.5 supplemented with 300 mM imidazole, pH 8.
Weigh 182 mg Tris–HCl, 877 mg NaCl, and 2.04 g imidazole.
Transfer to a 100 mL glass bottle. Add about 80 mL water and
mix until all components are dissolved. Adjust the pH to 7.5
with HCl. Fill to 100 mL water and store at 4 C.
4. 1 mL Extraction buffer: 150 mM NaCl, 15 mM Tris–HCl,
pH 7.5 supplemented with 0.2% (w/v) β-dodecylmaltoside
(DDM) and 5 mM EDTA. Weigh 1.82 mg Tris–HCl and
8.77 mg NaCl and transfer to a microtube. Add water to a
volume of 800 μL and mix until all components are dissolved.
Adjust the pH to 7.5 with HCl. Add 20 μL of 10% (w/v) DDM
stock and 1.46 mg of EDTA. Fill to 1 mL and store at 20 C.
5. DNase I (1 U/μL).
6. Lysozyme (Chicken).
7. Vortex.
8. Centrifuge tubes and compatible shaking/rotating device.
9. Sonication device.
10. Ni-NTA resin (e.g., Qiagen).
11. Cooled centrifuge (4 C).
12. Gravity flow columns.
13. Nanodrop or Bradford assay to measure protein concentration.
14. 10% (w/v) DDM stock prepared in deionized water. Weigh 1 g
DDM and transfer to a 10 mL tube. Fill up to 10 mL with
water and mix until all DDM is dissolved. Store aliquots at
20 C.
15. Criterion TGX 4–20% polyacrylamide pre-cast gels (Bio-rad).
16. 20 native running buffer (Invitrogen).
17. 20 cathode buffer additive (Invitrogen).
18. 5 blue native sample buffer (BN SB): mix 2.5 mL of 20
native running buffer with 4 g of glycerol (99%), 2.5 mL of
20 cathode buffer additive and 1 mL of deionized water and
store aliquots at 20 C.
19. Electrophoresis cell and power supply fitting the respective gel
system.
20. Surgical blade.
3 Methods
3.1 Transformation 1. Transform electrocompetent E. cloni® EXPRESS BL21 (DE3)

and Recombinant with pT7-SC1 plasmid DNA containing the ClyA-AS gene. On
Expression of ClyA-AS ice, add 1 μL DNA (75–150 ng/μL plasmid) to 50 μL cells and
in Escherichia coli mix gently. Add the mixture to a precooled electroporation
(Day 1–3) cuvette and electroporate (2 5 ms pulse, setting: bacteria).
2. Immediately add 400 μL pre-warmed (37 C) LB medium
containing 1% (w/v) glucose to the cells and incubate for 1 h
at 37 C.
3. Transfer the cell suspension to a LB agar plate, supplemented
with 1% (w/v) glucose and 100 μg/mL ampicillin. Spread out
the suspension using a T-shaped spreader until the plate is dry.
Incubate the plates at 37 C overnight.
4. On day 2, inoculate a single colony into 10 mL LB-medium
containing 100 μg/mL ampicillin (starter culture) and incubate
in a 50 mL Erlenmeyer flask at 37 C and 200 rpm overnight.
5. On day 3, transfer 1–5 mL of the starter culture to an Erlen-
meyer flask containing 200 mL 2 YT medium supplemented
with 100 μg/mL ampicillin. Incubate the expression culture
for ~2–3 h at 37 C and 200 rpm until OD600 (optical density)
is 0.6–0.8.
6. Add 0.5 mM IPTG to induce protein expression and incubate
the culture overnight in a 25 C shaking incubator at 200 rpm.
3.2 Purification 1. Harvest the bacterial cells by centrifugation in 50 mL centri-

of ClyA-AS (Day 4) fuge tubes at 4000 g for 30 min at 4 C. Discard the
supernatant. Store the pellet at 80 C.
2. Freeze the cell pellet for at least 1 h at 80 C. To allow a more
thorough cell disruption, the pellet should be frozen at 80 C
and thawed two more times. It takes approximately 20 min to
be frozen again.
3. Resuspend the pellets in 20 mL of Lysis buffer supplemented
with 0.2 units/mL DNase I. Use a vortex to homogenize the
solution well.
4. Add 10 μg/mL of lysozyme. Mix well by vortexing and incu-
bate the suspension for 20–30 min at room temperature while
shaking.
5. Place the tube containing the cell suspension on ice and disrupt
the cells by sonication: 3 30 s at 30% output (see Note 1).
6. Spin down debris by centrifugation for 30 min at 5000
g while cooled at 4 C. The protein is in the supernatant.
Transfer the cell lysate to a fresh 50 mL centrifuge tube.
7. During the centrifugation of the previous step, start preparing
the Ni-NTA resin. Take 100 μL of Ni-NTA resin (200 μL of
Ni-NTA suspension) into a microtube and wash the resin by

adding 500 μL of WB. Mix by inverting the tube and centrifuge
the suspension for 30 s at 11,000 g. Carefully remove the
supernatant and add 500 μL of fresh WB. Repeat three times.
8. Add the washed Ni-NTA resin to the cell lysate in the centri-
fuge tube. Incubate for 10–30 min while turning at room
temperature (see Note 2).
9. Prewash the empty gravity flow column (1.2 mL bed volume)
with approximately 5 bed volumes of WB (~5 mL).
10. Transfer the lysate with the incubated Ni-NTA resin stepwise
into the column (see Note 3).
11. Wash the column with at least 50 resin volumes (~5 mL) of WB
and remove excess WB by centrifugation of the column in a
clean microtube for 30 s (max. 2000 g).
12. Place the column into a clean microtube and elute the ClyA
monomers by adding 100 μL of EB. For larger columns, use a
volume of EB that covers the resin volume. Incubate for
5–10 min and elute by a quick (~30–60 s) centrifugation spin
(max. 2000 g). Repeat the elution step ~3–4 times. Store the
monomer solution at 4 C (see Note 4).
3.3 Oligomerization 1. Measure the protein concentration using a spectrophotometer or

and Gel Purification a Bradford assay and dilute the purified protein to 1 mg/mL
of ClyA-AS Nanopores in EB.
(Day 4) 2. Add 0.2% (w/v) DDM and mix well. Incubate for 30 min at
37 C without shaking.
3. Add 5 blue native sample buffer to the purified protein so
that the final concentration in the sample is 1 BN SM (dilute
down five times).
4. Prepare the blue native PAGE. Assemble it into the electropho-
resis cell and fill the container and the top compartment of the
gel with 1 native running buffer. Using a pipette, wash the
wells of the gel with 1 native running buffer into the wells.
Make sure air bubbles are removed.
5. Load 45 μL of sample into each well (depending on the well
size). Then, add 2 mL of 20 cathode buffer additive to the
top compartment and mix gently (final concentration in top
compartment should be 1 cathode buffer additive). Then
allow the gel to run for 10 min at 120 V (see Note 5).
6. After 10 min, take out the gel and discard the cathode buffer
additive solution (blue solution) from the upper compartment.
Wash thoroughly (six times or more) with water in order to
remove the blue solution from the wells.
7. Fill the upper compartment with fresh 1 native running
buffer and insert the gel back in the container.
Fig. 1 Isolation of different ClyA-AS types separated on 4–20% acrylamide BN-PAGE
8. Allow the gel to run for approximately 1 h at 120 V until the

protein bands corresponding to the oligomeric ClyA types are
clearly separated (Fig. 1).
9. Using a surgical blade, cut out the bands corresponding to the
desired oligomeric state (see Note 6). Store the bands at 20 C.
10. Extract the protein from a smaller piece (1/10th) of the gel by
adding 20 μL Extraction buffer. Incubate at room temperature
for 1–2 h. Store the gel-extracted ClyA oligomers for
1–2 months at 4 C.
4 Notes
1. Make sure not to overheat by keeping samples on ice.

2. Incubate for no longer than 45 min. Longer incubation times
will cause the ClyA to aggregate.
3. If the column does not flow through well, a tabletop centrifuge
can be used to facilitate the flow through. Do not use speeds
higher than 2000 g as they might damage the column and
centrifuge.
4. ClyA monomers will aggregate when frozen; therefore do not
freeze the ClyA monomers.
5. The 20 cathode buffer additive should have entered the gel.
6. Oligomeric ClyA proteins show multiple bands in blue-native
polyacrylamide gels (Fig. 1). To extract the 12-mer form of
ClyA (Type I), extract the lowest band from the polyacrylamide
gel. The band above is corresponding to the 13-mer variant of
ClyA (Type II).
References
1. Bezrukov SM, Vodyanoy I, Parsegian VA membrane channel. Proc Natl Acad Sci U S A
(1994) Counting polymers moving through a 93:13770–13773
single ion channel. Nature 370:279–281 3. Wanunu M, Morrison W, Rabin Y, Grosberg
2. Kasianowicz JJ, Brandin E, Branton D, Dea- AY, Meller A (2009) Electrostatic focusing of
mer DW (1996) Characterization of individual unlabelled DNA into nanoscale pores using a
polynucleotide molecules using a salt gradient. Nat Nanotechnol 5:160–165
4. Meller A, Nivon L, Brandin E, Golovchenko J, proteins by selective external association and

Branton D (2000) Rapid nanopore discrimina- pore entry. Nano Lett 12:4895–4900
tion between single polynucleotide molecules. 14. Mueller M, Grauschopf U, Maier T,
Proc Natl Acad Sci U S A 97:1079–1084 Glockshuber R, Ban N (2009) The structure
5. Akeson M, Branton D, Kasianowicz JJ, of a cytolytic alpha-helical toxin pore reveals its
Brandin E, Deamer DW (1999) Microsecond assembly mechanism. Nature 459:726–730
time-scale discrimination among polycytidylic 15. Franceschini L, Brouns T, Willems K,
acid, polyadenylic acid, and polyuridylic acid as Carlon E, Maglia G (2016) DNA translocation
homopolymers or as segments within single through nanopores at physiological ionic
RNA molecules. Biophys J 77:3227–3233 strengths requires precise nanoscale engineer-
6. Clarke J, Wu HC, Jayasinghe L, Patel A, ing. ACS Nano 10:8394–8402
Reid S, Bayley H (2009) Continuous base 16. Soskine M, Biesemans A, De Maeyer M, Maglia
identification for single-molecule nanopore G (2013) Tuning the size and properties of
DNA sequencing. Nat Nanotechnol ClyA nanopores assisted by directed evolution.
4:265–270 J Am Chem Soc 135:13456–13463
7. Luchian T, Shin SH, Bayley H (2003) Single- 17. Van Meervelt V, Soskine M, Singh S,
molecule covalent chemistry with spatially Schuurman-Wolters GK, Wijma HJ,
separated reactants. Angew Chem Int Ed Engl Poolman B, Maglia G (2017) Real-time con-
42:3766–3771 formational changes and controlled orientation
8. Movileanu L, Howorka S, Braha O, Bayley H of native proteins inside a protein nanoreactor.
(2000) Detecting protein analytes that modu- J Am Chem Soc 139:18640–18646
late transmembrane movement of a polymer 18. Galenkamp NS, Soskine M, Hermans J,
chain within a single protein pore. Nat Bio- Wloka C, Maglia G (2018) Direct electrical
technol 18:1091–1095 quantification of glucose and asparagine from
9. Soskine M, Biesemans A, Maglia G (2015) bodily fluids using nanopores. Nat Commun
Single-molecule analyte recognition with 9:4085
ClyA nanopores equipped with internal protein 19. Wloka C, Van Meervelt V, van Gelder D,
adaptors. J Am Chem Soc 137:5793–5797 Danda N, Jager N, Williams CP, Maglia G
10. Maglia G, Heron AJ, Stoddart D, Japrung D, (2017) Label-free and real-time detection of
Bayley H (2010) Analysis of single nucleic acid protein ubiquitination with a biological nano-
molecules with protein nanopores. Methods pore. ACS Nano 11:4387–4394
Enzymol 475:591–623 20. Van Meervelt V, Soskine M, Maglia G (2014)
11. Lin Y, Ying YL, Gao R, Long YT (2018) Detection of two isomeric binding configura-
Single-molecule sensing with nanopore con- tions in a protein-aptamer complex with a
finement: from chemical reactions to biological biological nanopore. ACS Nano
interactions. Chemistry 24:13064–13071 8:12826–12835
12. Ramsay WJ, Bell NAW, Qing Y, Bayley H 21. Willems K, Van Meervelt V, Wloka C, Maglia G
(2018) Single-molecule observation of the (2017) Single-molecule nanopore enzymol-
intermediates in a catalytic cycle. J Am Chem ogy. Philos Trans R Soc Lond B Biol Sci
Soc 140:17538–17546 372:1726
13. Soskine M, Biesemans A, Moeyaert B, 22. Miles G, Cheley S, Braha O, Bayley H (2001)
Cheley S, Bayley H, Maglia G (2012) An engi- The staphylococcal leukocidin bicomponent
neered ClyA nanopore detects folded target toxin forms large ionic channels. Biochemistry
40:8514–8522
Chapter 3
Building Synthetic Transmembrane Peptide Pores

Kozhinjampara R. Mahendran
Abstract
Membrane protein pores have demonstrated applications in nanopore technology. Previous studies have
mostly focused on β-barrel protein pores, whereas α-helix-based transmembrane protein pores are rarely
explored in nanopore applications. Here, we developed a synthetic transmembrane peptide pore built
entirely from short synthetic α-helical peptides. We examined the formation of a stable uniform
ion-selective pore in single-channel electrical recordings. Furthermore, we show that cyclodextrins (CDs)
block the peptide pores and determine the kinetics of CD binding and translocation. We suggest that such
designed synthetic transmembrane pores will be useful for several applications in biotechnology, including
stochastic sensing.
Key words Peptide, Pores, Bilayer, Conductance, Single-channel, Substrates
1 Introduction
Membrane-spanning proteins are found in all major groups of

organisms and involved in a wide variety of biological functions,
including transport of small molecules across the membrane
[1, 2]. They are mostly hydrophobic, usually requiring appropriate
detergents during their extraction and purification from the mem-
brane [1, 3]. Often, they fold into their native structures and can
exert their functions in specific membrane-mimetic conditions [4–
6]. Notably, membrane proteins are engineered extensively for
applications in single-molecule sensing of a wide variety of biomo-
lecules [5, 6]. Previously engineering and redesign of membrane
proteins have been predominantly based on β-barrels [3, 5]. For
example, heptameric α-hemolysin pore (αHL) derived from Staph-
ylococcus aureus has been engineered extensively for the detection of
a wide variety of analytes through stochastic sensing [5, 7]. Most of
the natural ion channels consist of α-helical bundles and charge
selectivity, which has not been produced with engineered αHL
pores [5, 8]. Importantly, alpha-helical pore-forming proteins
include a broad range of antimicrobial peptides (AMPs) and
19
20 Kozhinjampara R. Mahendran
alpha-pore-forming toxins (α-PFTs) with diverse structures and

membrane assembly mechanisms [2, 9–11]. However, engineering
α-helical bundles from existing scaffolds can cause misfolding of
proteins, and therefore α-helical bundles are engineered mainly by
de novo design [12–16]. Also, the interaction of alpha-helical pep-
tides with lipid membrane can produce complex structures hinder-
ing a stable pore formation [12, 13, 16].
Recently, there is a significant interest in developing and engi-
neering synthetic transmembrane pores that can be used for sto-
chastic sensing of small biomolecules [5, 17]. Specifically,
engineering α-helical peptides that can selectively conduct specific
ions can play a major role in nanobiotechnology. Notably, a short
synthetic peptide based on the membrane-spanning domain of
polysaccharide transporter Wza has been previously characterized
[13]. This synthetic pore inserts into the membrane to form an
active stable pore, but the assembly pathway of the pore consists of
several intermediate conductance states [13]. Currently, various
approaches are used for the de novo design of an entire protein
based on the physical principles of protein folding [18]. In particu-
lar, computational methods are used extensively to design possible
structures with molecular-level accuracy for building suitable struc-
tural models for engineering [15, 19, 20]. We propose that a stable,
functional alpha-helical synthetic pore with larger inner diameter
can be used for single-molecule sensing of complex biomolecules.
Here, we focus on an alpha-helical peptide pore, pPorA, based on
short synthetic peptides consisting of 40 amino acids derived from
the porin PorACj produced by a Gram-positive bacteria Corynebac-
terium jeikeium [16, 21].
2 Materials
2.1 Peptide Structure 1. Pore-forming 40 amino acid pPorA peptides (Peptide Protein
Analysis Research Ltd. at >95% purity as lyophilized powder). (MID-
QITEIFGQLGTFLGGFGNIFKGLKDVIETIVKWTAAK).
Solubilize peptides in 10 mM potassium phosphate buffer,
pH 7.4 and store at 4 C for 3 months.
2. MOS-500 spectropolarimeters fitted with Peltier temperature
controllers (Bio Logic Science Instruments) for Circular
Dichroism Analysis.
3. Quartz cuvettes are suitable for circular dichroism
measurements.
4. PBS buffer: 8.2 mM sodium phosphate, 1.8 mM potassium
phosphate, 137 mM sodium chloride, 2.7 mM potassium
chloride at pH 7.4.
Synthetic Peptide Pores 21
5. CD buffer: 8.2 mM sodium phosphate, 1.8 mM potassium

phosphate, 137 mM sodium chloride, 2.7 mM potassium
chloride at pH 7.4 with 1% n-dodecyl β-D-maltoside (DDM).
2.2 Gel Extraction 1. SDS PAGE reagents: 2 Laemmli sample buffer (Bio-Rad),
Any kD™ Mini-PROTEAN® TGX™ precast gel (Bio-Rad),
Precision Plus Protein™ Dual Color Standards (Bio-Rad).
2. Microfilterfuge tubes.
2.3 Single-Channel 1. Pre-painting solution: Use either 10 mg/mL hexadecane in n-

Recording pentane or 10 mg/mL hexadecane in hexane. This stock solu-
tion can be applied directly to Teflon aperture and can be
stored at 20 C for 3 months.
2. Bilayer forming lipid solution: 1,2-diphytanoyl-sn-glycero-3-
phosphocholine (Avanti Polar Lipids) dissolved in pentane
(5 mg/mL). This lipid stock solution can be applied directly
to bilayer chambers and can be stored at 20 C for at least
6 months.
3. Electrolyte solution: Typically contains 1 M KCl, 10 mM
HEPES, pH 7.4. Dithiothreitol (DTT) can be used with elec-
trolyte buffer to facilitate cysteine peptide insertion into the
bilayer.
4. Detergent buffer: 0.1% n-dodecyl β-D-maltoside (DDM) solu-
bilized in 10 mM potassium phosphate buffer pH 7.4 is used
for directly diluting peptides. This solution can be stored at
20 C for 6 months.
5. Analyte pore blocker: octakis-(6-amino-6-deox-
y)-γ-cyclodextrin octahydrochloride (am8γCD, AraChem
Cyclodextrin-Shop) solubilized in 1 M KCl electrolyte buffer
is used for blocking pores. This solution should be prepared
fresh before adding into the bilayer chamber.
6. Bilayer chambers: We use custom-made vertical bilayer cham-
bers made from Teflon or Delrin with a wide variety of shapes
and solution volumes ranging from 500 μL to 2 mL.
7. Polytetrafluoroethylene (Teflon) film (Goodfellow, Cam-
bridge) for the bilayer formation.
8. Teflon film of several diameter apertures (30–100 μm) are
obtained by punching holes into 25-μm thickness polytetra-
fluoroethylene (PTFE) Teflon foils by a high-voltage discharge
BD10A (Electro-Technic Products, US).
9. Ag/AgCl electrodes in 3 M KCl, 3% agarose bridge.
10. Bilayer chamber cleaning solution: KOH, isopropanol, metha-
nol, and ethanol.
11. Equipment: Axopatch 200B patch-clamp amplifier (Molecular
Devices, CA); Digidata 1550B 1440A digitizer with data
acquisition and analysis program (pClamp 10.6); Faraday cage

setup and antivibration table (Holmarc Opto-Mechatronics
Pvt. Ltd.) (see Note 1).
3 Methods
3.1 Peptide 1. Synthesize peptides using solid-phase peptide synthesis and

Synthesis purify to >95% by reversed-phase high-performance liquid
chromatography (HPLC).
2. Confirm their molecular mass by mass spectrometry. Store pep-
tides as lyophilized powders at 20 C. Here, we used 40 amino
acid residue peptides, pPorA: MIDQITEIFGQLGTFLGGFG
NIFKGLKDVIETIVKWTAAK derived from the membrane
porin PorACj of Corynebacterium jeikeium.
3. Design, synthesize, and purify a mutant peptide MIDQ
ITEIFGQLGTFLGGFGNIFCGLKDVIETIVKWTAAK.
3.2 Circular 1. Determine the secondary structures of the peptides in different

Dichroism buffer conditions by circular dichroism (CD) spectroscopy
Spectroscopy using MOS-500 spectropolarimeters fitted with Peltier tem-
perature controllers.
2. Prepare peptide samples as 5 and 50 μM solutions in CD buffer.
Measure CD spectra of peptides only in PBS buffer as a control
to confirm the role of detergent in the proper folding of alpha-
helical peptides. Record spectra using a 1 mm path length
quartz cuvette at 20 C.
3. For each dataset (in deg.), subtract baselines from the same
buffer and cuvette, and then normalize data points for amide
bonds, concentration of peptide and path length of the cuvette
to obtain mean residue ellipticity (MRE; deg. cm2/dmol/res)
(see Note 2). CD spectra should confirm the secondary struc-
tures of wild-type and mutant pPorA, which exist in α-helical
conformation in 1% n-dodecyl β-D-maltoside (DDM) micelles
(Fig. 1a, b).
3.3 SDS PAGE Gel SDS-polyacrylamide gel electrophoresis (SDS-PAGE) is used to

Extraction of pPorA assess the oligomeric state of DDM solubilized mutant pPorA
Oligomers peptides.
1. Solubilize the purified peptides in detergent buffer and prepare
in Laemmli SDS-sample buffer. Load onto an Any kD Mini-
PROTEAN® TGX™ precast gel with a Protein Standard
Marker and perform electrophoresis until dye front is at the
bottom of gel (see Notes 3 and 4).
Fig. 1 CD spectra of pPorA peptide pores. (a) CD spectra at 20 C for wild-type pPorA in PBS with 1% DDM at
5 μM (blue) and 50 μM (red) peptide concentration. (b) CD spectra at 20 C for cysteine mutant pPorA in PBS
with 1% DDM at 5 μM (blue) and 50 μM (red) peptide concentration
2. Visualize the proteins by staining the gel with standard Coo-

massie technique. Determine the electrophoretic mobility and
molecular mass estimate of peptides in SDS-PAGE based on
the protein standards. The pPorA mutant peptides produce
three bands on SDS-PAGE gel, a lower ~4.5 kDa band
corresponding to the monomeric unit, ~9 kDa band
corresponding to the dimeric unit, and a ~35 kDa band
corresponding to a possible octameric pore structure (Fig. 2a).
3. Cut the bands containing corresponding peptide oligomers
from the gel, crush the gel slices, and incubate each separately
in 500 μL phosphate buffer for 45 min at room temperature or
longer.
4. Obtain the cleared protein filtrate by using a microfilterfuge
tube centrifuged at 15,000 g for 5 min. pPorA oligomer is
ready to use in single-channel electrical recording.
3.4 Single-Channel 1. Set up the single-channel electrical recording apparatus with

Electrical Recordings the following parameters: Apply a low-pass filter frequency of
10 kHz and a sampling frequency of 50 kHz. Optional-Filter
the electrical signal at 2 kHz using an 8-pole Bessel digital filter
based on the analysis requirement.
2. Perform single-channel recordings using planar lipid bilayers of
1,2-diphytanoyl-sn-glycero-3-phosphocholine (see Note 5) by
the Montal and Muller technique across a polytetrafluoroethy-
lene (Teflon) film aperture (25-μm thick, ~80 μm in diameter)
[22] (see Note 6).
3. Coat each side of the Teflon film containing the aperture with
1–2 μL pre-painting solution using a simple stereo microscope
(see Note 7).
Fig. 2 Electrical properties of gel extracted cysteine mutant pPorA peptide pores. (a) The mutant pPorA
peptides run on the SDS-PAGE. The red circle indicates the self-assembled mutant pPorA peptide oligomers.
(b) Electrical recording showing single gel extracted mutant pPorA pores at +50 mV and (c) 50 mV. The
corresponding current-amplitude histogram is shown. (d) I–V curve obtained from a single gel extracted
mutant pPorA pores. (e) Gating of gel extracted single mutant pPorA pores at +200 mV. (f) Interaction of gel
extracted single mutant pPorA pores with am8γCD (1 μM, trans) at +50 mV. The current signals were filtered at
10 kHz and sampled at 50 kHz. Electrolyte: 1 M KCl, 10 mM HEPES, pH 7.4
4. Fill cis and trans compartments of the bilayer chamber with

500 μL of electrolyte buffer.
5. Add 5 μL bilayer forming lipid solution into both chambers.
6. Prepare the monolayer in each chamber by lowering the elec-
trolyte level below the aperture and raising it again. Connect
the cis compartment to the grounded electrode while connect
the trans compartment to the live electrode using Ag/AgCl
wires set in 3% agarose containing 3.0 M KCl (see Note 8).
7. Apply a potential difference across the bilayer and test the
stability of the bilayer at different voltages ranging from 50
to 200 mV (see Notes 9 and 10).
8. Mix in purified peptides solubilized in 0.1% DDM to the cis
side of the chamber (see Notes 11 and 12). Apply a voltage
of +200 mV without breaking the bilayer.
9. Monitor pore formation as a stepwise increase in the conduc-
tance of the electrical reading. Analyze the electrical recording
data with the latest version of pClamp software. Reconstitution
of pore-forming peptides into planar lipid bilayer under an

applied potential allows for measuring an ion conductance
that provides an approximate pore size, charge selectivity, or
functional properties such as channel opening and closing or
substrate blocking.
10. A description of the electrical characteristics of the pPorA pore
is as follows:
11. The oligomer ~35 kDa band from the gel rapidly inserts into
the bilayer forming a stable pore of single-channel conductance
of 4.0 0.2 nS at 50 mV (Fig. 2b, c).
12. The current–voltage (I–V) curve shows higher conductance at
negative potential compared to positive potential (Fig. 2d).
13. The gating of the peptide pore is only observed at voltages
higher than 60 mV (Fig. 2e).
14. The oligomeric pore can be blocked with am8γCD (Fig. 2f).
The addition of 1 μM am8γCD to the trans side results in the
ion current blockages only at positive potential establishing a
strong electrostatic interaction between the am8γCD and the
peptide pore.
3.5 Determination 1. Obtain multiple insertions of at least 100 pPorA or pPorA

of Single-Channel mutant oligomers in single-channel recording by adding high
Conductance of pPorA concentrations of peptide pores to the cis chamber provided this
Peptide Pores do not disrupt the bilayer stability (Fig. 3a).
2. Apply a fixed voltage as multiple pores insert into the bilayer.
Re-form the bilayer if it breaks to allow more pores to insert.
3. Determine the current for each pore by the stepwise increase in
current. Generate a histogram and fit with Gaussian to deter-
mine the average unitary conductance that reveals the unifor-
mity of the pores (Fig. 3b–d). A typical pore has a single-
channel unitary conductance (G) of 4.0 0.2 nS at +25
and +50 mV (n ¼ 100). The pore shows slight asymmetry in
the single-channel conductance, i.e., rectification at the posi-
tive applied voltage or higher conductance obtained at negative
voltages than positive voltages (Fig. 3c, d).
4. Acquire the current–voltage (I–V) curve for one single peptide
pore at a time by applying different voltages ranging from
100 to +100 mV (Fig. 4a).
3.6 Charge 1. Acquire ion selectivity measurements using asymmetric salt

Selectivity conditions in planar lipid bilayers. Add 1.0 M KCl in the cis
side and 0.15 M KCl in the trans side of the bilayer
compartment.
Fig. 3 Electrical properties of cysteine mutant pPorA peptide pores. (a) Electrical recording of multiple
insertions of mutant pPorA pores into a planar bilayer at +25 mV. (b) Single mutant pPorA pores insertion
at +50 mV. (c) Electrical recording of single mutant pPorA pores at +50 mV. (d) Electrical recording of single
mutant pPorA pores at 50 mV. The corresponding current amplitude histogram fitted with Gaussian is shown
as an inset. The current signals were filtered at 2 kHz and sampled at 10 kHz. Electrolyte: 1 M KCl, 10 mM
HEPES, pH 7.4
2. Add peptides to the cis side of the bilayer chamber. Here, the
peptide pore formation produces a particular ion current in the
absence of an applied transmembrane voltage.
3. Manually set the ionic current to zero by fine-tuning the
applied voltage. The voltage needed to produce zero current
is termed as the “reverse potential” (Vm), which is used to
estimate the permeability ratio of K+ and Cl ions across the
pore by Goldman-Hodgkin-Katz equation [23].
!
cis
RT P K þ ½K þ þ P Cl ½Cl trans
Vm ¼ ln
F P K þ ½K þ
trans
þ P Cl ½Cl cis
4. In this equation, R is the universal gas constant (8.314 J/K/
mol), T is the temperature in Kelvin (K ¼ C + 273.15), F is the
Faraday’s constant (96,485 C/mol), P K þ is the relative
Fig. 4 Single-channel electrical properties of cysteine mutant pPorA peptide pores. (a) Current–voltage (I–V)
curve obtained from single mutant pPorA pore. (b) Reverse potential derived from the I–V curve of single
mutant pPorA pore. Electrolyte: 1 M KCl, 10 mM MES, pH 6.0, cis/0.15 M KCl, 10 mM MES, pH 6.0, trans. (c)
Electrical recording of single mutant pPorA showing gating at +100 mV. (d) Electrical recording of single
mutant pPorA showing rapid gating at +150 mV. The current signals were filtered at 2 kHz and sampled at
10 kHz. Electrolyte: 1 M KCl, 10 mM HEPES, pH 7.4
membrane permeability for K+, P Cl is the relative membrane

permeability for Cl, [K+]cis is the concentration of K+ ion in
the cis side, [K+]trans is the concentration of K+ in the trans side,
[Cl]cis is the concentration of Cl in the cis side, and [Cl]trans
is the concentration of Cl in the trans side. Ion-selectivity
studies reveal that the mutant pPorA pore is cation-selective,
with a permeability ratio of 10:1 for K+ and Cl ions (Fig. 4b).
3.7 Voltage- 1. Record the ionic behavior of single pPorA pores at various
Dependent Gating voltages for extended periods of time, e.g., 10 min for each
Analysis of pPorA voltages from 10 to 200 mV (see Note 13).
Pores and Substrate- 2. Analyze the ionic current (open and closed states) in pClamp
Blocking Events analysis software. The single-channel recordings should reveal
that the peptide pore is in a non-gating state at voltages lower
3.7.1 Voltage-Dependent
than 50 mV. At higher voltages above 60 mV, the pore
Gating
fluctuated between open and closed conductance state (gating)
with more frequent gating events at the higher voltages
(Fig. 4c, d).
3.7.2 Substrate Blocking 1. Acquire the ionic behavior of a single pPorA peptide at differ-
ent substrate concentrations and applied voltages, recording
for 10 min each substrate concentration/voltage condition (see
Notes 14 and 15).
2. Quantify substrate blocking using a statistical analysis of a
single pore in its blocked and unblocked states. Count the
number of blockage events and the blockage time with a
single-channel search in pClamp analysis software. Obtain the
average blockage time or dwell times by plotting a histogram
that is fit to a standard exponential distribution.
3. Calculate the average residence time of substrate closure (τc)
and the reciprocal of (τc) gives the dissociation rate of the
substrate. The time between successive substrate blockages
will give (τo), and the 1/[(τo)] C, (where C is the concentra-
tion of the substrate) provides the association rate of the
substrate.
3.8 Cleanup After 1. An efficient washing procedure is essential for cleaning the
Each Single-Channel bilayer chambers after each use.
Recording Experiment 2. Wash the bilayer chamber after experiments with large amounts
of distilled water, followed by isopropanol and methanol.
3. Dry the chambers thoroughly under a stream of nitrogen to
evaporate organic solvents.
4. In the case of gel extracted peptides, use 3 M KOH and ethanol
washes to remove any residual peptides from the chip and
Teflon.
4 Notes
1. An acoustic screening (Faraday cage) of the setup, in combina-

tion with an antivibration table, will allow us to measure ion
conductance with better time resolution.
2. The CD spectra of peptides only in PBS buffers show that
peptides exist as insoluble aggregates in the buffer.
Importantly, the peptides are folded in an alpha-helical confor-

mation only in the presence of the detergent.
3. Remove the seal from the precast gel before performing gel
electrophoresis. If the seal is not removed, proteins cannot run
through the gel efficiently.
4. Load the peptides on Any kD Mini-PROTEAN® TGX™ pre-
cast gel (Bio-Rad) with Protein Standard Marker (Bio-Rad) for
resolving the monomer bands precisely. We used Plus Protein
Dual Color Standards for molecular weight estimation due to
increased brightness makes it easier to identify peptides on the
gel in real time. This allows the rapid extraction of the peptides
from the gel without staining the gel.
5. The most widely used lipid for planar lipid membranes is
DPhPC as it has ideal stable bilayer-forming properties.
Other lipids such as α-lecithin result in the formation of unsta-
ble leaky bilayers not suitable for peptide pore reconstitution.
6. The lipid bilayer is a perfect insulator compared to a bilayer in
the presence of a pore that allows the detection of pore con-
ductance and selectivity. The most stable lipid bilayers are
achieved by using a Teflon aperture of small diameter in the
range of ~30 μm, which allows longer recordings for 6–8 h but
requires more patience during protein reconstitution. Also,
better time resolutions can be achieved using smaller
membranes.
7. The Teflon apertures should be viewed using a light micro-
scope for accurate pre-painting of the hole, which is critical for
the formation of the stable bilayers. A small amount of hex-
adecane is used because the Teflon film aperture should be free
of excess hexadecane so that an electrical current can be
recorded between the bilayer compartments before the forma-
tion of the bilayer.
8. The use of a 3 M KCl/agarose bridge between an electrode is
critical for the activity of mutant cysteine peptides and
ion-selectivity experiments. The 3 M KCl agarose bridge
reduces the access resistance and prevents Ag+ ions from the
electrode leaching into the electrolyte solution. Also, double
distilled water should be used to prepare all aqueous buffers
including KCl (electrolyte) which should be filtered through a
0.4-μm membrane.
9. By measuring the capacitance of the membrane we can con-
clude on the stability of the bilayer.
10. In the complete absence of protein pores, the planar lipid
membrane itself can display pore-like events. Therefore, it is
crucial to form DPhPC planar lipid bilayers and record the
current through lipid bilayers at high voltages up to 200 mV

for 5 min to confirm the stability of lipid bilayers before peptide
reconstitution.
11. A low concentration of gel extracted peptides should be added
for the insertion into the lipid bilayer. The addition of a high
concentration of gel extracted peptides can cause unstable lipid
bilayers, specifically at higher voltages. Depending on the con-
centration of the peptide stock, the first insertion will occur
after a few minutes.
12. In the case of cysteine peptides, add ~1 μM DTT to facilitate
rapid pore insertion into the lipid bilayer.
13. Perfuse the bilayer compartment with electrolyte buffer to
avoid numerous insertions of peptides and maintain a single
channel for single-molecule sensing.
14. Usually, cationic substrates, anionic substrates, and neutral
substrates are used to block the pores formed by the peptides.
In the case of charged substrates, blocking with the pore
depends on the external parameters such as the side of the
addition, salt concentration, polarity, and magnitude of the
applied voltage. For example, addition to cationic substrates
to the cis side results in ion current blockages only at negative
voltages as substrates are electrophoretically pulled into the
pore to interact with negatively charged residues in the pore
lumen.
15. Notably, we examined the interaction of cation-selective
mutant pPorA pores with cationic CDs. For example, cationic
cyclic octasaccharide (am8γCD) interacts with the pore in a
voltage-dependent manner. Addition of 1 μM am8γCD to the
cis side of the bilayer compartment resulted in time-resolved
ion current blockages only at negative voltages indicting
electrophoretic pulling of CDs into the pore (Fig. 5a, c).
When the polarity of the voltage reversed to positive potential,
no ion current blockages were observed confirming strong
electrostatic interaction between am8γCD and the pore
(Fig. 5b, d).
Fig. 5 Interaction of cysteine mutant pPorA peptide pores with cationic cyclodextrins. (a) Typical ion current
recordings showing the interaction of single mutant pPorA with am8γCD (1 μM, cis) at 50 mV. (b) No ion
current blockages were observed at +50 mV. (c) Typical ion current recordings showing the interaction of
single mutant pPorA with am8γCD (1 μM, trans) at +50 mV. (d) No ion current blockages were observed at
50 mV. The current signals were filtered at 10 kHz and sampled at 50 kHz. Electrolyte: 1 M KCl, 10 mM
HEPES, pH 7.4
References
1. von Heijne G (2006) Membrane-protein 6. Kasianowicz JJ, Balijepalli AK, Ettedgui J, For-
topology. Nat Rev Mol Cell Biol 7:909–918 stater JH, Wang H, Zhang H, Robertson JW
2. Dal Peraro M, Van der Goot FG (2016) Pore- (2016) Analytical applications for pore-
forming toxins: ancient, but never really out of forming proteins. Biochim Biophys Acta
fashion. Nat Rev Microbiol 14:77–92 1858:593–606
3. Bayley H, Jayasinghe L (2004) Functional 7. Bayley H (2015) Nanopore sequencing: from
engineered channels and pores (review). Mol imagination to reality. Clin Chem 61:25–31
Membr Biol 21:209–220 8. Tajkhorshid E, Nollert P, Jensen MO, Miercke
4. Bayley H (2009) Membrane-protein structure: LJ, O’Connell J, Stroud RM, Schulten K
piercing insights. Nature 459:651–652 (2002) Control of the selectivity of the aqua-
5. Ayub M, Bayley H (2016) Engineered trans- porin water channel family by global orienta-
membrane pores. Curr Opin Chem Biol tional tuning. Science 296:525–530
34:117–126
9. Lear JD, Wasserman ZR, DeGrado WF (1988) 17. Wang S, Zhao Z, Haque F, Guo P (2018)
Synthetic amphiphilic peptide models for pro- Engineering of protein nanopores for sequenc-
tein ion channels. Science 240:1177–1181 ing, chemical or protein sensing and disease
10. Brogden KA (2005) Antimicrobial peptides: diagnosis. Curr Opin Biotechnol 51:80–89
pore formers or metabolic inhibitors in bacte- 18. Huang PS, Feldmeier K, Parmeggiani F,
ria? Nat Rev Microbiol 3:238–250 Velasco DAF, Hocker B, Baker D (2016) De
11. Puthumadathil N, Jayasree P, Santhosh novo design of a four-fold symmetric
Kumar K, Nampoothiri KM, Bajaj H, Mahen- TIM-barrel protein with atomic-level accuracy.
dran KR (2019) Detecting the structural Nat Chem Biol 12:29–34
assembly pathway of human antimicrobial pep- 19. Thomson AR, Wood CW, Burton AJ, Bartlett
tide pores at single-channel level. Biomater Sci GJ, Sessions RB, Brady RL, Woolfson DN
7:3226–3237 (2014) Computational design of water-soluble
12. Woolfson DN (2017) Coiled-coil design: alpha-helical barrels. Science 346:485–488
updated and upgraded. Subcell Biochem 20. Lu P, Min D, DiMaio F, Wei KY, Vahey MD,
82:35–61 Boyken SE, Chen Z, Fallas JA, Ueda G,
13. Mahendran KR, Niitsu A, Kong L, Thomson Sheffler W, Mulligan VK, Xu W, Bowie JU,
AR, Sessions RB, Woolfson DN, Bayley H Baker D (2018) Accurate computational
(2017) A monodisperse transmembrane design of multipass transmembrane proteins.
alpha-helical peptide barrel. Nat Chem Science 359:1042–1046
9:411–419 21. Abdali N, Barth E, Norouzy A, Schulz R, Nau
14. Woolfson DN, Bartlett GJ, Burton AJ, Heal WM, Kleinekathofer U, Tauch A, Benz R
JW, Niitsu A, Thomson AR, Wood CW (2013) Corynebacterium jeikeium jk0268 con-
(2015) De novo protein design: how do we stitutes for the 40 amino acid long PorACj,
expand into the universe of possible protein which forms a homooligomeric and anion-
structures? Curr Opin Struct Biol 33:16–26 selective cell wall channel. PLoS One 8:e75651
15. Joh NH, Wang T, Bhate MP, Acharya R, Wu Y, 22. Gutsmann T, Heimburg T, Keyser U, Mahen-
Grabe M, Hong M, Grigoryan G, DeGrado dran KR, Winterhalter M (2015) Protein
WF (2014) De novo design of a transmem- reconstitution into freestanding planar lipid
brane Zn2+-transporting four-helix bundle. membranes for electrophysiological characteri-
Science 346:1520–1524 zation. Nat Protoc 10:188–198
16. Krishnan RS, Satheesan R, Puthumadathil N, 23. Benz R, Schmid A, Hancock RE (1985) Ion
Kumar KS, Jayasree P, Mahendran KR (2019) selectivity of gram-negative bacterial porins. J
Autonomously assembled synthetic transmem- Bacteriol 162:722–727
brane peptide pore. J Am Chem Soc
141:2949–2959
Chapter 4
Design and Assembly of Membrane-Spanning DNA

Nanopores
Kerstin Göpfrich, Alexander Ohmann, and Ulrich F. Keyser
Abstract
Versatile lipid membrane-inserting nanopores have been made by functionalizing DNA nanostructures
with hydrophobic tags. Here, we outline design and considerations to obtain DNA nanopores with the
desired dimensions and conductance properties. We further provide guidance on their reconstitution into
lipid membranes.
Key words DNA nanotechnology, DNA origami, Nanopores, Ionic conductance, Lipid membrane,
Synthetic ion-channels
1 Introduction
Protein pores traverse the membranes of virtually all living cells,

mediating vital processes like signal transduction or substrate
exchange. Yet these pores featuring nanoscale dimensions have
also found technological applications—from antimicrobial agents
[1] to synthetic cell assembly [2] and, most prominently, label-free
single molecule detection with notable success in DNA sequencing
[3]. Beyond the protein pores found in nature, synthetic
membrane-spanning nanopores have been tailored for specific
applications [4]. While peptide pores are a prominent example,
recent advances in structural DNA nanotechnology led to the
creation of diverse DNA-based membrane-spanning pores
(Fig. 1). Typically, hydrophobic tags are carefully positioned and
covalently liked to the DNA construct. They help overcome the
energy barrier for the insertion of the hydrophilic DNA into the
hydrophobic lipid bilayer. Remarkably, DNA nanopore designs
span three orders of magnitude in conductance and molecular
weight. Voltage [5] and ligand [6] gating has been demonstrated
and DNA pores have been employed for label-free detection of
translocating DNA [7]. With the large toolbox for chemical
33
34 Kerstin Göpfrich et al.
Fig. 1 Membrane-spanning DNA nanopores. (a) Schematic illustration of diverse membrane-spanning DNA
nanopores (literature references from left to right: [11, 12, 14, 15]). Adapted from [8]. (b) TEM micrograph of a
lipid vesicle with multiple DNA origami nanopores as pioneered by Langecker et al. Adapted from [13]. Rep-
rinted with permission
functionalization of DNA, membrane-spanning DNA nanopores

hold great potential for the implementation of versatile functions
[8]. Here, we give step-by-step instruction on the design and
synthesis of a simple DNA nanopore. We further give guidance
and discussion of a more complex DNA structure design.
2 Materials
2.1 DNA Nanopore The following materials are required to assemble a simple DNA
Assembly nanopore as presented and characterized in [9, 10]. It consists of
four interconnected DNA duplexes and is anchored in the lipid
membrane via four cholesterol tags as shown in Fig. 2. Prepare all
solutions using ultrapure water at room temperature.
1. Custom-made unmodified DNA oligonucleotides (standard
desalting purification).
2. Custom-made 30 fluorophore-modified DNA oligonucleotides
(HPLC purification).
3. Custom-made 30 cholesterol-tagged oligonucleotides. An
example list of DNA sequences previously used for the assem-
bly of a four-helix bundle [9] is given in Table 1.
4. TE buffer: 10 mM Tris–HCl, pH 8, 1 mM EDTA.
5. Folding buffer: 10 mM Tris–HCl, 1 mM EDTA, 20 mM
MgCl2, pH 8.0.
6. Microtubes (0.2, 0.6, 1.5 mL).
7. PCR thermocycler.
DNA Nanopore Design 35
Fig. 2 A simple DNA nanopore; (a) side view and (b) top view of a simple DNA nanopore consisting of four
interconnected DNA duplexes. The four cholesterol modifications for membrane anchoring are shielded by six
nucleotide long single-stranded DNA overhangs to avoid aggregation of the structure. Adapted from [9]
Table 1
DNA sequences employed for Cy3-labeled DNA nanostructures assembled with four cholesterol
modifications shielded by six nucleotide long overhangs (orange) Adapted from [9]
DNA
Sequence (5' to 3')
strand
sc1 TTTTTTCCTTTCCACGAACACAGGGTTGTCCGATCCTATATTACGACTCCTTT
sc2 TTTGGGAAGGGGTTCGCAAGTCGCACCCTAAACGA-CholTEG
sc3 TTTTTTTCTTATCCTGCATCGAAAGCTCAATCATGCATCTTT
sc4 TTTATGTTGAAGGCTCAGGATGCA-CholTEG
st1 TTTATCGGACATTCAACATGGAGTCGTGGTGCGACTA-CholTEG
st2 TTTTTTTGCGAACAGGATAAGACGTTTAGAATATAGGTTT-Cy3
st3 TTTTTCGATGCCCCTTCCCGATGCATGAAGGGCATCCTGAGCCACCCA-CholTEG
st4 TTTTTTTGTGTTCGTGGAATTGAGCTTTT-Cy3
2.2 Attachment 1. Unlabeled lipids: 1,2-dioleoyl-sn-glycero-3-phosphocholine

to Giant Unilamellar (DOPC) and 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine
Vesicles (GUVs) (POPC), delivered in chloroform at a stock concentration of
25 mg/mL (Avanti Polar Lipids).
2. GUVs prepared by electroformation with a lipid composition
of 50 mol% DOPC, 50 mol% POPC.
3. 300 mM sucrose solution to prepare GUVs.
4. Fluorescent lipid: Atto488-PE (from Atto-TEC), delivered in
chloroform at a stock concentration of 1 mg/mL. This lipid is
supplemented at 0.5 mol% in GUVs for visualization purposes.
5. Chloroform.
6. Two indium tin oxide (ITO)-coated glass slides (Nanion
Technologies GmbH).
7. Rubber ring as a spacer.

8. Cleaning reagents: acetone, isopropanol, 70% ethanol.
9. DNA nanostructures.
10. GUV buffer: 160 mM glucose, 10 mM Tris–HCl, 1 mM
EDTA, 20 mM MgCl2, pH 8.0.
11. Bovine Serum Albumin (BSA)-coated cover slides for the
observation chamber (to prevent fusion of the GUVs with
the glass surface).
12. Observation chamber (from ibidi) or assembled from a micros-
copy glass slide and a cover slide spaced with, e.g., double-
sided sticky tape.
13. Confocal laser scanning or epifluorescence microscope for
visualization.
14. Electroformation unit (Vesicle Prep Pro, Nanion
Technologies GmbH).
15. Dessicator.
3 Methods
3.1 DNA Nanopore 1. Dissolve lyophilized unmodified DNA oligonucleotides (sc1,

Assembly sc3) and fluorophore-tagged oligonucleotides (st2, st4) to
100 μM in TE buffer according to manufacturer’s instructions.
3.1.1 Oligonucleotide
Aliquot into several tubes. This solution can then be stored at
Preparation
20 C for long-term storage.
2. Dissolve lyophilized cholesterol-tagged DNA oligonucleotides
(sc2, sc4, st1, st3) to 100 μM in ddH2O according to manu-
facturer’s instructions. Aliquot (e.g., 10 μL per aliquot) and
store at 20 C.
3.1.2 DNA Nanopore 3. Thaw one aliquot of each of the eight of the unmodified,
Assembly fluorophore-modified and cholesterol-modified oligonucleo-
tides from Table 1.
4. In particular, heat the four aliquots containing cholesterol-
tagged oligonucleotides to 60 C for 5 min on a hotplate or
in the PCR thermocycler while the other four oligonucleotides
can be thawed at room temperature. Vortex the cholesterol-
modified oligonucleotides directly after heating and pipette
quickly to avoid aggregation.
5. Mix the eight DNA oligonucleotides, no preferential order,
from Table 1 to a final concentration of 1 μM in folding buffer
in a microtube (see Note 1). For example, mix together 1 μL of
each of the eight oligonucleotides (stock at 100 μM) and then
add 92 μL of folding buffer.
6. Place the microtube containing the DNA mixture into a PCR

thermocycler. Heat to 85 C for 5 min, subsequently gradually
cool to 25 C over 18 h. Store assembled DNA nanopores at
4 C or use immediately in a downstream assay.
3.2 Preparation 1. Mix DOPC and POPC lipids (dissolved in chloroform) in a 1:1
of Giant Unilamellar mole ratio and add Atto488 PE. Add chloroform to obtain a
Vesicles (GUVs) by 10 mM lipid mix. Seal all lipid vials with Teflon film to prevent
Electroformation evaporation and oxidation of the lipids. This lipid mix can be
stored at 20 C and reused for future electroformation
preparations.
2. Clean ITO slides by sonicating for 5–10 min in acetone, then in
isopropanol and lastly in ethanol.
3. Pipet 20 μL of the lipid mixture on the conducting side of each
ITO slide (as determined with a voltmeter).
4. Use a clean glass coverslip to spread the lipid mix. Spreading
should look even and reflect light with green colors.
5. Place in desiccator for at least 10 min (better 1–2 h) to evapo-
rate the chloroform and dry the lipid mix.
6. Press a rubber ring onto one of the slides, enclosing the area
where the lipid was spread.
7. Place this cover slide in the electroformation unit and fill
entirely with 300 mM sucrose.
8. Put the second slide on top of the rubber ring so that the two
lipid-coated sides are facing one another with the rubber ring
in between.
9. Apply an AC-current of 3VP-P and a frequency of 5 Hz for 2 h.
This may vary if you are using a different lipid mixture.
10. Remove the top glass slide and pipet up the vesicles (dispersed
in the sucrose solution) with a large-diameter pipette tip imme-
diately after the 2 h incubation is complete and store at 4 C
(see Note 2).
3.3 Attachment 1. Dilute the DNA nanopores to a concentration of 200 nM (1:5

of DNA Nanopore dilution of the initial 1 μM stock). Use the GUV buffer for the
to Giant Unilamellar dilution in order to match the osmolarity of the GUVs. Exam-
Vesicles (GUVs) ple: GUVs were 300 mM sucrose (hence 300 mOsm). The
folding buffer (10 mM Tris–HCl, 1 mM EDTA, 20 mM
MgCl2) has an osmolarity of approx. 70 mOsm. Therefore
the 1:5 dilution should be prepared by mixing the DNA nano-
pores with 358 mM sucrose to match the osmolarity of the
GUVs. This gives an osmolarity of 300 mOsm for the diluted
DNA nanostructures (see Note 3).
Fig. 3 Visualization of DNA nanopore membrane insertion. GUV imaged in bright field (left) and fluorescence
mode, excited at 514 nm (middle and right). At 0 s the DNA nanostructures are added to GUV suspension and
are not yet bound. At 60 s after addition of the DNA nanostructures, a bright ring forms around the vesicle,
indicating rapid folding and membrane attachment of the DNA nanostructures. Adapted from [8]. Reprinted
with permission
2. Add the osmolarity-matched DNA nanopore solution to the

GUV suspension at a ratio of 1:1 (v/v). Membrane attachment
happens within seconds (see Note 4).
3. Transfer the DNA nanopore-GUV mixture to a BSA-coated
observation chamber.
4. Use a confocal laser scanning microscope or an epifluorescence
microscope at an excitation of 514 nm to visualize the attach-
ment of the Cy3-labeled DNA nanopores to the lipid mem-
brane of the GUVs. The fluorescence from the nanopores
should be visible as a fluorescent ring in the confocal plane
(Fig. 3, right panel). The nanopore fluorescence (Cy3 fluores-
cence ¼ Ex max: 550 nm and Em max: 570 nm) should overlay
with the fluorescence of the lipid membrane (LissRhod PE ¼ Ex
max: 545 and Em max: 567) (see Note 5).
5. As a control experiment, prepare DNA nanopores without the
cholesterol modifications. This construct should remain in the
bulk solution around the GUVs but not embedded in the lipid
membrane.
3.4 Guidelines For more advanced applications, design and customize your own
for DNA Nanopore DNA nanopore with the desired features according to the protocol
Design outlined below.
1. Choose the required pore size. The pore size will depend on the
envisioned application of the DNA nanopore: For nanopore
sensing or to create a transmembrane passage for specific mole-
cules, keep in mind that the pore diameter must exceed the size
of the analyte. Note that even if this criterion is met, large
highly charged analytes may still be excluded from the pore
due to the negative charge of the DNA backbone (see Note 6).
Small pores, on the other hand, are preferable to engineer
selectivity and to mimic protein ion channels. The nominal

inner pore diameters of existing DNA nanopores ranges from
sub-nanometer [11] to 6 nm [12], spanning three orders of
magnitude in ionic conductance (from tens of pS to tens of nS).
2. Derive membrane anchoring strategy. Hydrophobic modifica-
tions, covalently linked to the DNA, facilitate the insertion of
DNA nanopores into lipid membranes (see Note 7). Choles-
terol [9, 11–13], tocopherol [7], and ethane [14] are all DNA
modifications commercially available. Alternatives such as in-
house-synthesized porphyrin [15, 16] have been successfully
used as hydrophobic tags for lipid membrane anchoring. Alter-
natively, biotin-functionalized DNA nanopores have been
designed to bind to biotinylated lipids via biotin-streptavidin
linkage [7] (also commercially available). For the design, it is
important to note that the anchoring strategy can also influ-
ence the ion conductance of the pore. Experiments and molec-
ular dynamics simulations have shown that the lipid head
groups tilt toward the membrane-inserted DNA nanostructure
(see Fig. 4a). In this configuration, the hydrophobic lipid tails
are shielded from the hydrophilic DNA. This leads to a toroidal
pore at the DNA-lipid interface. Ions can flow though this
toroidal lipid pore as well as through the central cavity of the
DNA nanopore, increasing the overall conductance (Fig. 4a,
red arrows). The emergence of a continuous lipid bilayer
leaflet allows for lipid molecules to diffuse from one bilayer
sheet to the other (Fig. 4a in blue). For this reason, membrane-
inserted DNA nanostructures have also been shown to act as
scramblases [17]. To confine the ion pathway to the central
Fig. 4 Computer-aided modeling of DNA nanopores. (a) Schematic illustration of a DNA nanopore with
hydrophilic exterior. Ions can pass through the central cavity as well as though the toroidal lipid pore forming
at the DNA-lipid interface [15]. Lipids can be transferred from one bilayer leaflet to the other, leading to
scramblase activity of DNA nanopores [17]. (b) Coarse grained simulation of the free energy of DNA nanopore
insertion as a function of the pore radius and the number of hydrophobic tags (here: cholesterol). Adapted from
[12]. Reprinted with permission. Copyright 2016, American Chemical Society. This direct link is supposed to
be included: https://pubs.acs.org/doi/10.1021/acsnano.6b03759
cavity and to inhibit scrambling, remove the charges of the

DNA backbone in the membrane-spanning section of the
pore by using ethane modifications [14], C12 spacers or
uncharged nucleic acids (e.g., PNA) in the stem section of the
designed DNA pore.
3. Determine the required number of membrane anchors. To trig-
ger self-insertion of the DNA nanopores, the energy cost for
pore formation in the lipid membrane has to be compensated
by the energy gain for the insertion of the hydrophobic tags or
the formation of other molecular bonds. Therefore, it is crucial
to choose the right number of hydrophobic tags. Referencing
the coarse-grained simulations in Fig. 4b, one can extract the
required number of cholesterol tags for a certain pore radius
(the blue region of the plot represents a net energy gain for
pore insertion) (see Note 8).
4. Choose between scaffolded DNA origami or scaffold-free DNA
tiles. This will depend on the application of the DNA nanopore.
Assembly of nanopores from short synthetic single strands of
DNA is useful when the target structure is small. Assembly
from single strands, however, becomes less efficient for large
finite-size assemblies (12–17% yield [18]), which is why scaf-
folded DNA origami is the method of choice for larger nano-
pores. When using a kilobase long DNA scaffold (e.g., the
commonly used DNA from the virophage M13mp18), the
size of the DNA nanostructure falls into the megadalton
regime. However, the achievable concentrations are typically
restricted to nanomolar amounts. Table 2 compares advantages
Table 2
Comparison of scaffolded DNA origami nanopores and scaffold-free DNA nanopores
Scaffolded DNA origami nanopores Scaffold-free DNA nanopores

Molecular weight MDa range kDa range
Number of DNA strands 1 scaffold, typically 200–350 Typically 2–20 synthetic
synthetic DNA oligonucleotides DNA oligonucleotides
Achievable concentrations Typically 1–100 nM Typically 1–10 μM
Modifications Large number of possible attachment Can be placed on all strands,
sites on staples, typically no however limited number
modifications on scaffold
Assembly Thermal annealing for hours to days Thermal annealing or RT
assembly within minutes
Purification Required (e.g., spin filtration) Mostly not required
References [7, 12, 13] [5, 6, 9, 10, 11, 14–16]
and disadvantages of both DNA assembly methods with

corresponding literature references.
5. Conceive target shape. The target shape of the nanopore will
depend on the chosen pore diameter. The pore cavity is typi-
cally lined by a single layer of DNA, although multiple layers
may improve the mechanical stability. The transmembrane sec-
tion of the pore should be at least 5 nm long to fully span the
lipid bilayer. For a scaffold-free design, the transmembrane
section is the only required part. For scaffolded DNA origami,
the remaining scaffold has to be used up. This can be achieved
by extending the pore length far beyond the thickness of the
bilayer up to 40 nm [12, 12], and/or by implementing a base
plate, which lies flat on the membrane and provides space for
the attachment of hydrophobic groups [7]. Such an asymmet-
ric mushroom-shaped pore with a large hydrophilic head will
always insert in the same orientation. To implement the envi-
sioned design, it is useful to employ computer-aided design
tools like caDNAno (see Note 9). While caDNAno has been
developed for scaffolded DNA origami, the scaffold path can
be broken up into short strands if the target structure is
scaffold-free. The same design principles apply as for DNA
nanostructures in general. A useful reference with guidance is
given elsewhere [19]. Once the target shape is drawn, the
design-file can be submitted to CanDo (see Note 10) for
structural and fluctuation analysis. Fluctuations can indicate
areas of weak structural stability, which can be stabilized by
additional crossovers. Following basic DNA origami design
principles [19] and combining caDNAno and CanDo, one
can be relatively certain that a new DNA origami design will
assemble as designed; however, the yield is to be experimentally
determined. Other design tools and software analysis packages
are outlined in Note 11.
6. Optimize the positioning of membrane anchors. For scaffolded
DNA origami, hydrophobic moieties are often placed on a base
plate adjacent to the stem of the pore [12, 13]. By extending
single-stranded DNA overhangs from the DNA origami struc-
ture, it is possible to attach many hydrophobic tags with just
one sequence (to reduce the cost of the DNA synthesis). These
overhangs are typically 15–20 bases long to ensure stable bind-
ing at room temperature. They can be placed in a standard [12]
or a zipper-like arrangement [13]. For scaffold-free designs,
the hydrophobic moieties are often placed on the transmem-
brane section of the pore itself [5, 6, 11, 13, 15]. To avoid twist
and strain, it may be useful to choose positions where the
helices point outward [9]. For both, scaffolded and scaffold-
free designs, it is crucial that the hydrophobic tags cannot all be
inserted without insertion of the DNA nanopore—otherwise
the DNA nanopores will lie flat on the membrane without

additional energy gain for flipping into the ion conducting
orientation. Careful geometrical considerations are required
to design a functional pore. Here, it is important to consider
the linker length and the length of the hydrophobic tag itself.
Porphyrin, for instance, is more compact than cholesterol,
making it better suited for very narrow architectures
[15]. Note that the positioning of the anchors will also affect
diffusion and membrane attachment [20]. To avoid aggrega-
tion of the structures, it may be helpful to shield the hydro-
phobicity of the tags. This can be realized with a dynamic
hinge, hiding the hydrophobic tags prior to membrane inser-
tion [21]. Alternatively, the extension of single-stranded over-
hangs which can wrap around the hydrophobic tags has been
shown to prevent aggregation effectively [9].
7. Design the DNA sequences. Custom DNA sequences can be
purchased commercially. When placing the order, in addition
to the DNA sequence, it is necessary to provide the type of
purification. Standard desalting is sufficient for unmodified
strands, however, HPLC purification is recommended for cho-
lesterol- or fluorophore-tagged DNA sequences. Designing the
correct DNA sequence is crucial. For scaffolded DNA origami
nanopores, the DNA sequences of the staples are predetermined
by the scaffold sequence. However, custom sequences can be
assigned to single-stranded DNA overhangs for the attachment
of hydrophobic groups. Designing the correct DNA sequence is
crucial for both scaffolded DNA and scaffold-free DNA nano-
pores. NUPACK’s thermodynamic analysis can help select
sequences without stable secondary structures and unwanted
complementarities with other strands (see Note 12). A favorable
set of DNA sequences will reduce the assembly times and pro-
duce a higher yield of the target structure.
8. Optional: Implement stimuli-response and/or selectivity. DNA
nanopores provide the exciting possibility to implement
stimuli-response and/or selectivity. The ideal pore for nano-
pore sensing should exhibit a stable low-noise conductance
state. Yet many other applications, including controlled drug
release systems, may require gated pores which are responsive
to external stimuli and/or selective for specific analytes. Due to
the negative charge of the DNA backbone, DNA nanopores
are, to some degree, cation selective. They transport propor-
tionally more cations compared to anions, and charge-based
exclusion of larger analytes is possible [15, 22]. G-quadruplexes
or other ion-sensitive DNA motifs could be attached to the
mouth of the pore in order to engineer more specific selectivity.
In terms of stimuli-response, note that voltage gating has been
observed for all DNA nanopores. A high transmembrane
voltage (typically above 50 mV) favors a lower conductance

state [5, 23]. Ligand gating can be implemented by incorpor-
ating specific binding sites at the pore entry, for instance bind-
ing sites for complementary DNA [6]. In future designs,
versatile stimuli-response could be achieved by exploiting
light-sensitive modifications, pH-responsive DNA motifs,
aptamers or other sequence motifs and functional groups (see
Note 13).
9. DNA nanopore assembly. Following the design and the synthe-
sis of the required DNA sequences, assembly of the DNA
nanopores is carried out according to protocols used for assem-
bly of standard DNA nanostructures [19]. Sequences carrying
hydrophobic tags are often added in excess—either after assem-
bly for DNA origami pores or in the folding mix of DNA tile-
based architectures.
3.5 Guidelines 1. Choose lipid composition. The insertion efficiency does not
for Lipid Membrane depend on the design of the DNA nanopore alone. The lipid
Reconstitution of DNA membrane itself has a strong influence on the insertion char-
Nanopores acteristics. Hydrophobic tags show lipid membrane prefer-
ences. For example, cholesterol and tocopherol insert almost
exclusively into liquid disordered domains composed of unsat-
urated lipids like DOPC at room temperature. Palmitate, on
the other hand, has been shown to insert preferentially into the
liquid ordered phase containing saturated lipids [24]. Charged
lipid membranes may lead to nonspecific orientation absorp-
tion of the DNA nanopores. At the same time, it is important
to select phospholipids which form stable bilayers with minimal
leakage (like DPhPC, DOPC, POPC, EggPC) to avoid mea-
surement artifacts related to ion conduction through transient
lipid pores (see Note 14).
2. Choose experimental system to test for DNA nanopore insertion.
The functionality of DNA nanopores has been tested by elec-
trical ionic conductance measurements [11, 13] as well as
optical observation of transmembrane transport [7]. However,
only electrical measurements can give insights into the ionic
conductance on the single-channel level. For these measure-
ments, a voltage is applied across a lipid membrane while
recording the ionic current. DNA nanopore insertion increases
the permeability of the membrane and hence increases the ionic
current in a stepwise manner. Different setups have been devel-
oped for single-channel ionic current measurements of protein
pores [25]. For DNA nanopore insertion, high membrane
fluidity and high membrane curvature are beneficial. Solvent-
containing membranes have been shown to be suitable for
DNA nanopore insertion [25]. Alternatively, the DNA nano-
pores can be added to GUVs which are then used for patch
clamping. Instead of a stepwise conductance increase, the over-

all conductance of the lipid bilayer patch is measured. Patches
can readily be formed and broken, giving a first indication for
the insertion rate in a high throughput manner [26]. Generally,
it is advisable to use two different setups to confirm the single-
channel conductance of a DNA nanopore, since different prop-
erties such as the lateral membrane tension can influence their
conductance behavior [5, 22].
4 Notes
1. DNA nanopores typically require divalent ions

(10–20 mM Mg2+) for long-term stability and membrane
attachment. Buffers often contain EDTA, a chelator for diva-
lent ions. Without EDTA, the concentration of divalent ions
can be reduced. Alternatively, increased concentrations of
monovalent ions can be used [27]. For some applications,
UV crosslinking may be the method of choice to prevent
degradation [28].
2. If there is a low yield of GUVs, try the following: (a) Prepare
fresh lipid stocks, sucrose solution, use new ITO slides and
then make a fresh batch of GUVs. (b) Remove GUVs immedi-
ately after the electroformation is finished, or else they will fuse
with the glass slides. (c) Try different lipid spreading techni-
ques and spread smaller/larger volumes. Temperature and
humidity can affect the yield of GUVs, so alter the procedure
to suit your lab environment.
3. This ensures that the osmotic pressures inside and outside the
GUV are equilibrated and sufficient MgCl2 is present for
attachment and for the structures to remain intact.
4. Not all DNA nanopores will insert in the desired orientation
perpendicular to the lipid membrane. Generally, only a small
fraction of the pores that are absorbed on the membrane will
create a transmembrane passage for ions.
5. If no insertions can be observed, first ensure that the DNA
nanopores are assembled correctly in the final measurement
buffer (e.g., with a combination of AFM, TEM, gel electro-
phoresis, DLS). Then check if the pores are attached to the
membrane and not aggregated (e.g., by functionalizing them
with a fluorescent dye). Confirm the unilamellarity of the
membrane with a standard widely used protein pore, e.g.,
α-hemolysin. Increase the DNA nanopore concentration and
the observation time. Finally, reconsider your design, in partic-
ular, the number and positioning of the hydrophobic tags. Also
consider using a different measurement setup.
6. Apart from DNA, other types of natural and unnatural nucleic

acids (e.g., RNA, PNA, XNA [29]) may be suitable and inter-
esting for the construction of nanopores as described here.
Both DNA-hybrids and nanopores made entirely of other
forms of nucleic acids are conceivable. Alternatively, DNA
nanotechnology can be used to scaffold the arrangement of
peptides and proteins leading to protein-DNA hybrid
nanopores [30].
7. Addition of surfactants (e.g., Triton, OPOE) can increase the
insertion efficiency of DNA pores, but it also alters the mem-
brane properties and can lead to artifacts related to the forma-
tion of transient lipid pores. Therefore, appropriate control
experiments are particularly crucial if surfactants are employed.
8. Note that aggregation of the DNA nanopores can occur if too
many hydrophobic tags are used [9, 13] or if their positioning
(see Subheading 3.4, step 6) or their sequence design (see
Subheading 3.4, step 7) is not optimal.
9. caDNAno is an open-source DNA origami design software
developed by Douglas et al., available at http://cadnano.org/
[31]. Multiple online tutorials offer guidance for inexperienced
users, and template designs are available for download. caD-
NAno can be installed as a stand-alone program as well as a
plug-in for Autodesk Maya providing an interface for three-
dimensional visualization. Within caDNAno, possible DNA
geometries are currently limited to hexagonal and square
lattices.
10. CanDo is a free online tool to predict the three-dimensional
shape and flexibility of scaffolded and scaffold-free DNA
nanostructures in solution using finite-element analysis
[32]. It is available at http://cando-dna-origami.org/. It can
thus help to make design choices before the more cost- and
time-intensive DNA synthesis. The structural predictions are
regularly improved and updated [33].
11. Other computational tools, including SARSE [34], Tiamat
[35], oxDNA [36], tacoxDNA [37], vHelix [38], DAEDA-
LUS [39], and ENRG MD [40] MrDNA [41], are useful for
design and structure prediction of DNA nanostructures.
12. NUPACK is a free online tool for the analysis and the design of
secondary structures of one or more interacting DNA
sequences available at http://www.nupack.org/ [42]. It is
especially useful to optimize the sequence design for small
scaffold-free DNA nanostructures and sequences that carry
hydrophobic groups.
13. DNA nanopores have also been inserted into solid state nano-
pores [43], where they can serve as adapters to tune the pore
size or to implement molecular recognition [44, 45]. While the
functionalization with hydrophobic tags is normally not neces-

sary in this case, the protocol provided here can also give
guidance on the design of DNA nanopores for solid state
supports.
14. Note that DNA nanopores will influence the lipid bilayer prop-
erties. The absorption of highly charged polymers creates an
asymmetry and influences the spontaneous membrane curva-
ture [46]. Lipid tubulation of GUVs has been observed as a
consequence of the attachment of micromolar concentrations
of DNA nanopores [8].
Acknowledgements
K.G. received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-
Curie grant agreement No. 792270. K.G. further acknowledges
support from the Winton Programme for the Physics of Sustain-
ability, Gates Cambridge and the Oppenheimer PhD studentship.
A.O. was supported by funding from the Engineering and Physical
Sciences Research Council (EPSRC) and through the Vice Chan-
cellor’s Award from the Cambridge Trust. U.F.K. received funding
from an ERC Consolidator Grant (Designerpores 647144) and
Oxford Nanopore Technologies.
References
1. Fernandez-Lopez S, Kim HS, Choi EC, molecular cargo across a biological membrane.
Delgado M, Granja JR, Khasanov A, Nat Nanotechnol 11:152–156
Kraehenbuehl K, Long G, Weinberger DA, 7. Krishnan S, Ziegler D, Arnaut V, Martin TG,
Wilcoxen KM, Ghadiri MR (2001) Antibacter- Kapsner K, Henneberg K, Bausch AR,
ial agents based on the cyclic D,L-α-peptide Dietz H, Simmel FC (2016) Molecular trans-
architecture. Nature 412:452–456 port through large-diameter DNA nanopores.
2. Göpfrich K, Platzman I, Spatz JP (2018) Mas- Nat Commun 7:12787
tering complexity: towards bottom-up con- 8. Göpfrich K (2017) Rational design of
struction of multifunctional eukaryotic DNA-based lipid membrane pores. Disserta-
synthetic cells. Trends Biotechnol 36:938–951 tion, University of Cambridge
3. Deamer D, Akeson M, Branton D (2016) 9. Ohmann A, Göpfrich K, Joshi H, Thompson
Three decades of nanopore sequencing. Nat RF, Sobota D, Ranson NA, Aksimentiev A,
Biotechnol 34:518–524 Keyser UF (2019) Controlling aggregation of
4. Sakai N, Matile S (2013) Synthetic ion chan- cholesterol-modified DNA nanostructures.
nels. Langmuir 29:9031–9040 Nucleic Acids Res 47:11441–11451
5. Seifert A, Göpfrich K, Burns JR, Fertig N, Key- 10. Ohmann A (2020) A synthetic lipid scramblase
ser UF, Howorka S (2015) Bilayer-spanning built from DNA. Dissertation, University of
DNA nanopores with voltage-switching Cambridge, https://www.repository.cam.ac.
between open and closed state. ACS Nano uk/handle/1810/303262
9:1117–1126 11. Göpfrich K, Zettl T, Meijering AE, Hernán-
6. Burns JR, Seifert A, Fertig N, Howorka S dez-Ainsa S, Kocabey S, Liedl T, Keyser UF
(2016) A biomimetic DNA-based channel for (2015) DNA-tile structures induce ionic cur-
the ligand-controlled transport of charged rents through lipid membranes. Nano Lett
15:3134–3138
12. Göpfrich K, Li CY, Ricci M, Bhamidimarri SP, 24. Loew M, Springer R, Scolari S, Altenbrunn F,
Yoo J, Gyenes B, Ohmann A, Winterhalter M, Seitz O, Liebscher J, Huster D, Herrmann A,
Aksimentiev A, Keyser UF (2016) Arbuzova A (2010) Lipid domain specific
Large-conductance transmembrane porin recruitment of lipophilic nucleic acids: a key
made from DNA origami. ACS Nano for switchable functionalization of membranes.
10:8207–8214 J Am Chem Soc 132:16066–16072
13. Langecker M, Arnaut V, Martin TG, List J, 25. Gutsmann T, Heimburg T, Keyser U, Mahen-
Renner S, Mayer M, Dietz H, Simmel FC dran KR, Winterhalter M (2015) Protein
(2012) Synthetic lipid membrane channels reconstitution into freestanding planar lipid
formed by designed DNA nanostructures. Sci- membranes for electrophysiological characteri-
ence 338:932–936 zation. Nat Protoc 10:188–198
14. Burns JR, Stulz E, Howorka S (2013) Self- 26. Göpfrich K, Kulkarni CV, Pambos OJ, Keyser
assembled DNA nanopores that span lipid UF (2013) Lipid nanobilayers to host
bilayers. Nano Lett 13:2351–2356 biological nanopores for DNA translocations.
15. Göpfrich K, Li CY, Mames I, Bhamidimarri SP, Langmuir 29:355–364
Ricci M, Yoo J, Mames A, Ohmann A, 27. Martin TG, Dietz H (2012) Magnesium-free
Winterhalter M, Stulz E, Aksimentiev A, Key- self-assembly of multi-layer DNA objects. Nat
ser UF (2016) Ion channels made from a single Commun 3:1103
membrane-spanning DNA duplex. Nano Lett 28. Gerling T, Kube M, Kick B, Dietz H (2018)
16:4665–4669 Sequence-programmable covalent bonding of
16. Burns JR, Göpfrich K, Wood JW, Thacker VV, designed DNA assemblies. Sci Adv 4:eaau1157
Stulz E, Keyser UF, Howorka S (2013) Lipid- 29. Pinheiro VB, Holliger P (2014) Towards XNA
bilayer-spanning DNA nanopores with a nanotechnology: new materials from synthetic
bifunctional porphyrin anchor. Angew Chem genetic polymers. Trends Biotechnol
Int Ed 52:12069–12072 32:321–328
17. Ohmann A, Li CY, Maffeo C, Al Nahas K, 30. Spruijt E, Tusk SE, Bayley H (2018) DNA
Baumann KN, Göpfrich K, Yoo J, Keyser UF, scaffolds support stable and uniform peptide
Aksimentiev A (2018) A synthetic enzyme built nanopores. Nat Nanotechnol 13:739–745
from DNA flips 107 lipids per second in 31. Douglas SM, Marblestone AH,
biological membranes. Nat Commun 9:2426 Teerapittayanon S, Vazquez A, Church GM,
18. Wei B, Dai M, Yin P (2012) Complex shapes Shih WM (2009) Rapid prototyping of 3D
self-assembled from single-stranded DNA tiles. DNA-origami shapes with caDNAno. Nucleic
Nature 485:623–626 Acids Res 37:5001–5006
19. Wagenbauer KF, Engelhardt FAS, Stahl E, 32. Castro CE, Kilchherr F, Kim DN, Shiao EL,
Hechtl VK, Stömmer P, Seebacher F, Wauer T, Wortmann P, Bathe M, Dietz H
Meregalli L, Ketterer P, Gerling T, Dietz H (2011) A primer to scaffolded DNA origami.
(2017) How we make DNA origami. Nat Methods 8:221–229
Chembiochem 18:1873–1885 33. Pan K, Kim DN, Zhang F, Adendorff MR,
20. Khmelinskaia A, Mücksch J, Petrov EP, Fran- Yan H, Bathe M (2014) Lattice-free prediction
quelim HG, Schwille P (2018) Control of of three-dimensional structure of programmed
membrane binding and diffusion of DNA assemblies. Nat Commun 5:5578
cholesteryl-modified DNA origami nanostruc- 34. Andersen ES, Dong M, Nielsen MM, Jahn K,
tures by DNA spacers. Langmuir Lind-Thomsen A, Mamdouh W, Gothelf KV,
34:14921–14931 Besenbacher F, Kjems J (2008) DNA origami
21. List J, Weber M, Simmel FC (2014) Hydro- design of dolphin-shaped structures with flexi-
phobic actuation of a DNA origami bilayer ble tails. ACS Nano 2:1213–1218
structure. Angew Chem Int Ed 53:4236–4239 35. Williams S, Lund K, Lin C, Wonka P,
22. Maingi V, Burns JR, Uusitalo JJ, Howorka S, Lindsay S, Yan H (2009) Tiamat: a three-
Marrink SJ, Sansom MS (2017) Stability and dimensional editing tool for complex DNA
dynamics of membrane-spanning DNA nanostructures. International workshop on DNA-
pores. Nat Commun 8:1–12 based computers. Springer, Berlin, Heidelberg,
23. Maingi V, Lelimousi M, Howorka S, Sansom 2008
MSP (2015) Gating-like motions and wall 36. Šulc P, Romano F, Ouldridge TE, Rovigatti L,
porosity in a DNA nanopore scaffold revealed Doye JPK, Louis AA (2012) Sequence-
by molecular simulations. ACS Nano dependent thermodynamics of a coarse-
9:11209–11217 grained DNA model. J Chem Phys
137:135101
37. Suma A, Poppleton E, Matthies M, Šulc P, 42. Zadeh JN, Steenberg CD, Bois JS, Wolfe BR,
Romano F, Louis AA, Doye JPK, Pierce MB, Khan AR, Dirks RM and Pierce NA
Micheletti C, Rovigatti L (2019) TacoxDNA: (2011) NUPACK: analysis and design of
a user-friendly web server for simulations of nucleic acid systems. J. Comput. Chem
complex DNA structures, from single strands 32:170–173.
to origami. J Comput Chem 40:2586–2595 43. Bell NAW, Engst CR, Ablay M, Divitini G,
38. Benson E, Mohammed A, Gardell J, Masich S, Ducati C, Liedl T, Keyser UF (2012) DNA
Czeizler E, Orponen P, Högberg B (2015) origami nanopores. Nano Lett 12:512–517
DNA rendering of polyhedral meshes at the 44. Hernández-Ainsa S, Bell NAW, Thacker VV,
nanoscale. Nature 523:441–444 Göpfrich K, Misiunas K, Fuentes-Perez ME,
39. Veneziano R, Ratanalert S, Zhang K, Zhang F, Moreno-Herrero F, Keyser UF (2013) DNA
Yan H, Chiu W, Bathe M (2016) Designer origami nanopores for controlling DNA trans-
nanoscale DNA assemblies programmed from location. ACS Nano 7:6024–6030
the top down. Science 352:1534 45. Wei R, Martin TG, Rant U, Dietz H (2012)
40. Maffeo C, Aksimentiev A (2020) MrDNA: a DNA origami gatekeepers for solid-state nano-
multi-resolution model for predicting the pores. Angew Chem Int Ed 51:4864–4867
structure and dynamics of DNA systems. 46. Nikolov V, Lipowsky R, Dimova R (2007)
Nucleic Acids Res 48(9):5135–5146 Behavior of giant vesicles with anchored DNA
41. Maffeo C, Yoo J, Aksimentiev A (2016) De molecules. Biophys J 92:4356–4368
novo reconstruction of DNA origami struc-
tures through atomistic molecular dynamics
simulation. Nucleic Acids Res 44:3013–3019
Part II
Single Molecule Detection and Analysis of Protein Analytes

Chapter 5
Determining the Orientation of Porins in Planar Lipid

Bilayers
Sandra A. Ionescu, Sejeong Lee, and Hagan Bayley
Abstract
Single-channel planar lipid bilayer (PLB) recording of bacterial porins has revealed molecular details of
transport across the outer membrane of Gram-negative bacteria, including antibiotic permeation and
protein translocation. To explore directional transport processes across cellular membranes, the orientation
of porins or other pore-forming proteins must be established in a lipid bilayer prior to experimentation.
Here, we describe a direct method for determining the orientation of porins in a PLB—with a focus on
E. coli OmpF—by using targeted covalent modification of cysteine mutants. Each of the two possible
orientations can be correlated with the porin conductance asymmetry, such that thereafter an I–V curve
taken at the start of an experiment will suffice to establish orientation.
Key words Porin, Outer membrane protein, OmpF, Planar lipid bilayer, Orientation, Electrophysiol-
ogy, Cysteine labeling, Thiol reagents, Membranes
1 Introduction
Bacterial porins are the most abundant proteins in the outer mem-
brane of Gram-negative bacteria where they mediate the passive
transport of nutrients and toxins, such as antibiotics and bacterio-
cins (strain-specific antibacterial proteins) [1]. Detergent-
solubilized porins can be reconstituted into artificial bilayers,
where they retain properties consistent with their in vivo function.
Single-channel planar lipid bilayer (PLB) recording has become a
powerful tool for measuring the currents that flow through indi-
vidual protein pores. The technique can detect interactions
between pores and other molecules with sub-millisecond resolution
[2], and has been used to investigate porin physiology, including
sugar transport through maltoporin [3], phage binding to FhuA
[4], and antibacterial peptide translocation through OmpF [5]. It
has also been used to develop stochastic sensors [6] based on the
α-hemolysin pore for the detection of kinases and sugars [7, 8]. Bac-
terial porins comprise an attractive class of transmembrane proteins
51
52 Sandra A. Ionescu et al.
for the development of additional sensor elements. Compared to

multi-subunit pore-forming proteins—such as α-hemolysin [9],
which is made up of seven polypeptide chains—each porin barrel
is composed of a single chain. Modifications of the pore at desired
locations can therefore be easily achieved by mutagenesis because
(1) assembly of heteromers is not required and (2) mutations can
be readily introduced over several β-strands. Because porins com-
prise largely symmetrical β-barrel structures, it is feasible for a porin
to insert into a PLB with either the extracellular or periplasmic
loops first. For example, the most abundant porin outer membrane
protein F (OmpF), which adopts only one orientation in vivo [10],
has been shown to insert into a PLB in both orientations
[11, 12]. This is known from the two types of I–V curve asymme-
try, in which the porin conductance is higher at either positive or
negative applied voltages. Asymmetric ion conduction has also
been observed with other porins, including maltoporin [13] and
OmpG [14]. In order to employ the highly informative PLB tech-
nique to investigate directional transport through porins or to
develop new sensors, the orientation of the porin in question
must first be established.
Previously, the orientation of the bacterial porins OmpF [11],
maltoporin [13], OmpG [14], and mitochondrial VDAC [15] were
indirectly inferred from interactions with antibiotics, reducing
reagents and sugars applied from one side or the other of a bilayer,
and by conductance changes in response to applied potentials,
respectively. These approaches are not generally applicable.
In this chapter a general and direct method for establishing
porin orientation is presented based on the authors’ work with the
trimeric E. coli OmpF [12]. The method uses targeted covalent
modification of peripheral cysteine residues—which are not present
in wild-type porins—introduced on the periplasmic and extracellu-
lar loops. Once a single-cysteine mutant has inserted into a PLB, an
I–V curve is measured to establish porin I–V asymmetry. A thiol-
directed PEG reagent (mPEG-maleimide or -orthopyridyl disul-
fide) is then introduced on each side of the bilayer in turn; the
molecular weight of the PEG is chosen such that it cannot diffuse
through the porin. When the reagent is introduced on the side of
the bilayer where the cysteine residue is exposed, a reaction will
occur generating a PEG adduct that causes a concomitant decrease
in the ionic current through the pore [12]. The site of reaction (cis
or trans) can then be correlated to the I–V asymmetry of the
channel to establish the absolute orientation of the porin with
respect to the I–V curve. After the connection between the porin
I–V curve and orientation has been established, an I–V curve
measured at the start of each subsequent experiment is enough to
establish the porin’s orientation. Caveats are that the wild-type
porin must have an asymmetric I–V curve and the chosen cysteine
mutation(s) must not significantly change this asymmetry. This
Orientation of Porins in Planar Lipid Bilayers 53
targeted covalent modification approach might be further applied

to determine the orientation of a wide range of membranes chan-
nels and pores, including toxins, MACPF proteins, and
transporters.
2 Materials
2.1 Solutions Prepare all solutions using ultrapure deionized and double-distilled
and Reagents water. Store buffers and reagents at room temperature unless oth-
erwise indicated. Use Hamilton syringes for the handling of
organic solvents.
1. Electrophysiology recording buffer (used for OmpF): 100 mM
KCl, 20 mM potassium phosphate buffer, pH 7.0. Weigh 7.5 g
KCl (see Note 1) and transfer to a graduated cylinder. Add
24.6 mL of K2HPO4 (0.5 M) and 15.4 mL of KH2PO4
(0.5 M). Add water to 900 mL. If necessary, adjust solution
to pH 7.0 (see Note 2). Make up to 1 L with water. Filter buffer
through a 0.22 μm polyethersulfone (PES) filter before use.
2. TCEP: Prepare a 20 mM tris(2-carboxyethyl)phosphine
(TCEP) solution by dissolving 5 mg in 1 mL water. Make
fresh before use.
3. Hexadecane (1% v/v in pentane): Prepare a 1 mL stock of 10%
anhydrous hexadecane in anhydrous pentane by dissolving
100 μL hexadecane in 900 μL pentane. Dilute the solution
tenfold in pentane and aliquot into glass vials. Use a PTFE-
lined cap to minimize evaporation and store at 20 C.
4. DPhPC lipid (2.5 mg/mL in pentane): For bilayer formation,
1,2-diphytanoyl-sn-glycerol-3-phosphocholine (DPhPC) is
used. Dissolve 25 mg of powdered lipid in 1 mL anhydrous
pentane. Dilute the solution tenfold in pentane and aliquot
into glass vials. Use a PTFE-lined cap to minimize evaporation
and store at 20 C.
5. mPEG-MAL-5K: Prepare a 5 mM solution of methoxy poly-
ethylene glycol-maleimide-5K (mPEG-MAL-5K) by dissolving
0.125 g in 0.5 mL electrophysiology recording buffer. Make
fresh before use.
6. mPEG-OPSS-5K: Prepare a 5 mM solution of methoxy poly-
ethylene glycol-orthopyridyl disulfide-5K (mPEG-OPSS-5K)
by dissolving 0.125 g in 0.5 mL electrophysiology recording
buffer (see Note 3). Make fresh before use.
7. DTT: Prepare a 200 mM dithiothreitol (DTT) solution by
dissolving 31 mg in 1 mL electrophysiology recording buffer.
Make fresh before use.
2.2 Proteins 1. Wild-type (WT) porin: Transform the plasmid containing the
porin gene of interest into E. coli. Express and purify proteins
according to a protocol of choice. For OmpF, express in E. coli
BZB1107 (ompf knockout strain) and purify as previously
described [5]. Store the purified protein in buffer containing
1% (w/v) n-octyl-β-D-glucopyranoside (β-OG) at 80 C.
2. Single-cysteine porin mutants: Introduce single-cysteine muta-
tions on extracellular or periplasmic loops at a position where
the thiol group will be exposed for reaction with thiol-directed
PEG reagents (see Note 4). For OmpF, mutations E29C,
N246C (extracellular), and D221C (periplasmic) have been
successfully labeled after PLB reconstitution [12]. The extra-
cellular and periplasmic designation is based on the orientation
found in vivo [10]. Express according to item 1, supplement-
ing the storage buffer with TCEP or DTT (1 to 2 mM) to
avoid thiol oxidation.
2.3 Electro- 1. Recording compartment: two Delrin (acetal resin) compart-

physiology Setup ments (0.5 mL volume).
2. PTFE film (25 μm thick).
3. RTV silicone coating.
4. PVC-coated copper wire.
5. Faraday cage.
6. Ag/AgCl electrodes: Oxidize silver wire (1.0 mm thick) over-
night in 5% sodium hypochlorite. Solder tip of silver electrodes
to PVC-coated copper wires.
7. Agarose salt bridge: 1.5% low-gelling agarose in 3 M KCl.
Weigh out 1.5 g low-gelling agarose and 22.4 g KCl and add
water to 100 mL. Microwave solution until agarose is fully
dissolved. Transfer 1 mL aliquots into 1.5 mL centrifuge
tubes. One tube is enough to make two electrodes.
8. Recording equipment and software for analysis. The authors
use an Axopatch 200B patch-clamp amplifier connected to a
CV 203BU head stage. Filter data through a 2 kHz low-pass
Bessel filter and digitize with a Digidata 1322A converter at a
sampling frequency of 10 kHz. Perform data analysis with
pClamp 10.3 software.
3 Methods
All experiments can be carried out at room temperature.
3.1 Planar lipid 1. Create a ~50 μm aperture in PTFE film: Suspend the film
bilayer between a 2–5 mm spark-gap and discharge a spark for approx-
recording setup imately 5 s. The aperture can be visualized and the diameter
determined with a light microscope.
Fig. 1 Electrophysiology setup and the orientations of OmpF in planar lipid bilayers. (a) Planar lipid bilayer
(PLB) recording setup. The recording compartment (black) resides in a Faraday cage and has two
compartments: cis at ground and trans to which a voltage is applied. A 3 M KCl agarose salt bridge
(yellow tips) forms an electrical connection between the recording buffer and each Ag/AgCl electrode,
which is soldered to an insulated copper wire (orange). The PLB (usually made from DPhPC) is formed over
a ~50 μm-sized aperture in a PTFE film (white box) sandwiched between the cis and trans compartments.
A porin of interest, e.g., OmpF, can insert into the bilayer from the cis compartment in two possible
orientations. (b) OmpF I–V curves obtained in 0.1 M KCl, 20 mM phosphate buffer, pH 7.0. The two
asymmetric curves represent the two orientations that OmpF can adopt in a PLB. Positive asymmetry
denotes higher conductance at positive applied voltages (blue); negative asymmetry denotes higher
conductance at negative applied voltages (orange)
2. Sandwich the PTFE film between two Delrin compartments

with RTV silicone coating. Allow to set overnight.
3. Make agarose bridges for the Ag/AgCl electrodes: Melt one
aliquot of agarose salt bridge solution. Fill a 100 or 200 μL
pipette tip with molten agarose and insert an electrode contain-
ing a rubber stopper at the top until a seal is formed (Fig. 1a).
Allow agarose to solidify.
4. Connect the electrodes to the recording apparatus. The com-
partment containing the grounded electrode is designated
“cis” and the opposite compartment to which voltage is applied
“trans” (Fig. 1a).
5. Form a bilayer according to the Montal-Mueller solvent-free
method [16, 17]. In brief, paint both sides of the PTFE film
with 1% (v/v) hexadecane using a pipette or capillary and allow
to dry for 10–20 min. Add electrophysiology recording buffer
and TCEP (200 μM final) to both compartments followed by
2–4 μL of lipid solution (2.5 mg/mL) on each side (see Note
5). Wait 5–10 min for the pentane to evaporate, then slowly
lower and raise the buffer level in both of the compartments to
create the bilayer.
3.2 Porin For our setup, cis is at ground and voltage is applied to the trans
Reconstitution side. Protein is introduced into the cis compartment.
and Establishing I–V
1. Reconstitute a single porin (WT or cysteine mutant) into the
Asymmetry bilayer. For OmpF, successful insertion has been realized by the
addition of 50–500 ng of protein in 1% (w/v) β-OG to 500 μL
of electrophysiology recording buffer [6]. Facilitate protein
insertion by stirring, pipette mixing, or applying voltages in
the 150–250 mV range.
2. To prevent further porin insertions, replace 20% of the buffer in
the protein-containing cis compartment with fresh buffer a
minimum of three times. Mix the compartment solution with
a pipette between each exchange.
3. Measure the channel I–V curve by recording the current read-
out at voltages ranging from 100 to +100 mV in 10 mV
increments (see Note 6). Positive asymmetry refers to a channel
that exhibits higher conductance at positive applied potentials,
whereas negative asymmetry refers to a channel that exhibits
higher conductance at negative applied potentials. The two
asymmetries will be assigned to the two possible orientations
of the porin within the bilayer (see Note 7) (Fig. 1b).
3.3 Cysteine Mutant 1. Insert a cysteine mutant porin into a PLB and establish the I–V
Labeling asymmetry according to step 3 of Subheading 3.2.
with mPEG-MAL-5K 2. Record the porin baseline conductance at the applied potential
that will be used for the cysteine-targeted covalent modifica-
tion. For OmpF, use 70 mV (see Note 8).
3. Add mPEG-MAL-5K (1 mM final) to the cis compartment (see
Note 9). A successful reaction between the porin thiol and
mPEG-MAL-5K should produce a stable stepwise drop in
conductance (see Note 10). The number of steps corresponds
to the stoichiometry of the pore. The drop(s) in conductance
should not be reversible. For trimeric OmpF, it is typical to
observe three steps ranging from 1 to 5 pA, depending on the
cysteine mutant [12] (Fig. 2a, TOP). Adduct formation is
typically seen within 5 min of reagent addition, but this may
vary among porins and with the buffer pH.
4. Add mPEG-MAL-5K to the trans compartment according to
step 3. If a stepwise reaction has already been observed upon
reagent addition to the cis side, no further reaction should be
seen upon reagent addition to the trans side (Fig. 2b).
5. If no reaction (current step) is observed on either side of the
bilayer, then the cysteine thiol may be (a) buried in the bilayer
or (b) oxidized or may have reacted with reagent impurities.
Cysteine reactivity can be assessed by carrying out the targeted
covalent modification in bulk solution and analyzing the
Fig. 2 Determining cysteine mutant orientation by targeted covalent modification. (a) OmpF cysteine mutant
inserted in an orientation that exposes the thiol groups to the cis side of the bilayer and exhibits a positive (Pos)
I–V curve asymmetry (left panel). Top: Upon addition of mPEG-MAL-5K to the cis compartment, a PEG adduct
is formed through a thio-ether linkage at each cysteine thiol, causing stepwise drops in the current. In the case
of trimeric OmpF, three steps are observed (associated with levels L0–L3). The addition of DTT does not alter
the conductance, as the thio-ether bond cannot be broken. Bottom: Upon addition of mPEG-OPSS-5K to the cis
compartment, thiol-disulfide exchange initiated by the porin cysteine thiols generates three PEG adducts
(associated with levels L0–L3). The resultant disulfide linkages are cleaved upon cis-side addition of DTT,
returning the conductance to the initial open pore level (L0). (b) OmpF extracellular loop cysteine mutant
inserted in an orientation that exposes the thiol groups to the trans side of the bilayer and exhibits negative
(Neg) asymmetry. Upon addition of a thiol-directed PEG reagent into the cis side of the bilayer, no adduct is
formed (the labeling reaction does not occur) as the bulky PEG molecule cannot pass through the porin.
Therefore, as expected, there is no change in the open pore current (L0)
products by SDS-PAGE. If labeling is efficient in the bulk phase

but is not observed in the PLB recording, then (a) may be the
case: repeat steps 1–4 with a new cysteine mutant. If no or
poor labeling is observed on SDS-PAGE, then (b) may be the
case: the preparation of the cysteine mutant (exposure to oxi-
dizing conditions) and reagent purity should be considered.
6. Note the connection between the I–V curve asymmetry (posi-
tive or negative, as defined in step 3 of Subheading 3.2) and
the side of cysteine labeling (cis or trans).
7. Repeat steps 1–4 several times to ensure that the side of adduct
formation (cis or trans) is dependent on the orientation of the
channel, as determined from the I–V curve. If a cysteine
mutant porin exhibiting positive asymmetry reacts with the
reagent on the cis side of the bilayer, then a porin with the
same cysteine mutation inserted in the opposite orientation
(negative asymmetry) should react with the reagent on the
trans side of the bilayer.
3.4 Cysteine Mutant 1. Repeat steps 1–7 in Subheading 3.3 using mPEG-OPSS-5K
Labeling (1 mM final) (Fig. 2a, BOTTOM). In this case, the PEG
with mPEG-OPSS-5K adduct is attached to the porin via a disulfide bond (see Note
11). The OPSS reaction is slower than the MAL reaction. With
OmpF, the reaction of all three cysteines with the OPSS
reagent is sometimes observed only after 30 min [12].
2. If a stepwise reaction is observed (see Note 12), add DTT
(20 mM final) to the same compartment to which the
mPEG-OPSS-5K successfully reacted with the porin. The
DTT will cleave the disulfide bond to release the PEG adduct,
leading to a stepwise increase in conductance with the same
amplitudes seen during adduct formation.
3.5 Determining 1. Based on the targeted covalent modification results from Sub-
Porin Orientation headings 3.3 and 3.4, establish the orientation of each porin
in a PLB from the I– cysteine mutant relative to the I–V curve, e.g., positive asym-
V Curve metry indicates insertion into the bilayer from the cis compart-
ment with the periplasmic loops first, leaving the extracellular
loops in cis and the periplasmic loops in trans.
2. Ensure that the results are consistent across multiple porin
cysteine mutants. If a particular cysteine mutant exhibits posi-
tive asymmetry upon bilayer insertion and reacts with reagent
added to the cis side, then that same mutant exhibiting negative
asymmetry should react with reagent on the trans side. Like-
wise, if an extracellular cysteine mutant exhibits positive asym-
metry upon bilayer insertion and reacts with reagent added to
the cis side, then a periplasmic cysteine mutant that exhibits
positive asymmetry should react with reagent added to the
trans side only (Fig. 3).
Fig. 3 Orientation of OmpF in planar lipid bilayers summary. The connection between I–V curve asymmetry
and OmpF cysteine mutant bilayer orientation for our setup, in which the cis compartment is at ground and
voltage is applied to the trans compartment. For porins exhibiting positive I–V asymmetry (blue trimers), the
cysteine thiols will be exposed on the cis side for extracellular cysteine mutants and on the trans side for
periplasmic cysteine mutants. For porins exhibiting negative I–V asymmetry (orange trimers), the opposite will
be true. Therefore, for WT OmpF, positive I–V asymmetry means the extracellular loops are exposed in the cis
compartment, leaving the periplasmic loops in trans. The opposite orientation applies for WT OmpF exhibiting
negative I–V asymmetry
3. The absolute orientation of the cysteine mutant porin(s) can

now be determined from a single-channel I–V curve.
4. Identify a porin substrate that shows side-dependent (cis vs
trans addition) interaction patterns. For OmpF, enrofloxacin
can be used [11, 12].
5. Determine if the side-dependent interaction pattern relative to
the channel I–V asymmetry (positive or negative) is consistent
across cysteine mutants and WT to confirm that the relative
orientations are the same (see Note 13).
6. The absolute orientation of the WT porin can now be deter-
mined from a single-channel I–V curve.
4 Notes
1. The authors recommend using ultrapure KCl (99.999% trace

metal basis) to minimize metal impurities that may interact
with protein pores and affect the current read-out. KCl was
chosen because the mobility of K+ and Cl ions are similar. Salt
concentrations higher than 100 mM can be used so long as the
I–V asymmetry of the pore is maintained (see Subheading 3.2).
2. Other buffers with different pH values can also be used. A pH
between 7.0 and 8.0 is recommended to favor thiol deprotona-
tion to the reactive thiolate, while avoiding side reactions
between primary amines and the maleimide reagent.
3. The authors noted that successful reaction between the mPEG-
OPSS reagent and the OmpF cysteine mutant was highly sus-
ceptible to reagent impurities. Ask the supplier for a purity
assessment or repurify the reagent in-house, for example, by
using HPLC, where possible. Impurities remaining after syn-
thesis, such as dipyridyl disulfide, may react with the free porin
cysteine(s) to form a disulfide adduct, thus precluding the
formation of a subsequent PEG adduct via thiol-disulfide
exchange.
4. Multiple periplasmic and extracellular single-cysteine mutants
should be prepared in parallel for testing. Depending on the
position and orientation of a thiol group, adduct formation
may not be favorable (e.g., if the cysteine is buried in or located
near the lipid bilayer) or may not cause a decrease in porin
conductance (e.g., if the adduct is too distant from the ion
conducting pathway). PyMOL was used to choose free and
exposed residues for cysteine mutagenesis based on the
OmpF crystal structure (PDB: 3POX). Avoiding charged resi-
dues is recommended, as the glutamate (E29C) and aspartate
(D221C) mutants showed a bigger deviation from the WT I–V
curve than the asparagine mutant (N246C).
5. Addition of TCEP to the recording compartment prevents
cysteine oxidation during the recording and will react less
readily than DTT with the maleimide group. However, high
concentrations of TCEP (>5 mM) can destabilize the lipid
bilayer.
6. An increase in current asymmetry was noted at higher voltages
(up to 200 mV was tested) for OmpF. This may be a general
trend for the porins and an I–V curve that includes higher
voltages (within the constraints of bilayer stability) can be
measured in cases of weak asymmetry.
7. The two types of I–V curve (positive and negative asymmetry)
should be linked by a 180 rotation, as expected for proteins in
opposite orientations in a symmetric bilayer. The I–V curves of

WT porin and the cysteine mutants should be compared to
ensure that they are similar, i.e., the pore conductance and
asymmetry should not have been significantly affected by the
single mutation. This allows extrapolation of the orientation
obtained with cysteine mutants to the WT porin. For other
methods of confirming the connection between the orientation
of cysteine mutants and WT porin, see Subheading 3.5.
8. Any voltage can be used for the PEG labeling experiments so
long as the bilayer remains stable for the duration of the
experiment (for DPhPC, we recommend 150 mV as the
maximum voltage) and the change in conductance upon PEG
adduct formation can be resolved from the noise in the current
trace. The noise will depend on the diameter of the bilayer and
the acquisition frequency.
9. Because of the susceptibility of the OPSS reaction to impurities
(see Note 3), the authors recommend investigating porin ori-
entation using the maleimide (MAL) reagent first. Orientation
and cysteine specificity can then be confirmed using the OPSS
reagent.
10. Test the mPEG-MAL-5K reaction with WT porin to ensure
there are no off-target interactions between the PEG and the
porin, which may obfuscate results.
11. The labeling reaction is carried out with the reversible mPEG-
OPSS-5K reagent to affirm the results obtained with mPEG-
MAL-5K and to prove cysteine specificity, as primary amines
can also react with maleimides at high pH (>8).
12. The amplitude of the current step(s) observed with the OPSS
reagent should be similar to the current step(s) observed with
the MAL reagent if the same sized PEG is used, e.g., 5 kDa was
used for OmpF [12].
13. Enrofloxacin exhibits side-dependent interactions with OmpF
in the presence of Mg2+. In the case of positive I–V curve
asymmetry and under negative applied potential, enrofloxacin
added from the cis side causes fast blocking events, whereas
trans-side addition does not significantly alter OmpF baseline
conductance [11, 12]. An OmpF exhibiting negative I–V curve
asymmetry should show the opposite interaction pattern, with
fast blocking events observed after enrofloxacin addition from
the trans side.
References
1. Pagès JM, James CE, Winterhalter M (2008) 9. Song L, Hobaugh MR, Shustak C, Cheley S,
The porin and the permeating antibiotic: a Bayley H, Gouaux JE (1996) Structure of
selective diffusion barrier in Gram-negative staphylococcal alpha-hemolysin, a heptameric
bacteria. Nat Rev Microbiol 6:893–903 transmembrane pore. Science 274:1859–1866
2. Qing Y, Pulcu GS, Bell NAW, Bayley H (2018) 10. Hoenger A, Pagès JM, Fourel D, Engel A
Bioorthogonal cycloadditions with (1993) The orientation of porin OmpF in the
sub-millisecond intermediates. Angew Chem outer membrane of Escherichia coli. J Mol Biol
Int Ed 57:1218–1221 233:400–413
3. Kullman L, Winterhalter M, Bezrukov SM 11. Brauser A, Schroeder I, Gutsmann T,
(2002) Transport of maltodextrins through Cosentino C, Moroni A, Hansen UP, Winter-
maltoporin: a single-channel study. Biophys J halter M (2012) Modulation of enrofloxacin
82:803–812 binding in OmpF by Mg2+ as revealed by the
4. Udho E, Jakes KS, Buchanan SK, James KJ, analysis of fast flickering single-porin current. J
Jiang X, Klebba PE, Finkelstein A (2009) Gen Physiol 140:69–82
Reconstitution of bacterial outer membrane 12. Ionescu SA, Lee S, Housden NG, Kaminska R,
TonB-dependent transporters in planar lipid Kleanthous C, Bayley H (2017) Orientation of
bilayer membranes. Proc Natl Acad Sci U S A the OmpF porin in planar lipid bilayers. Chem-
22:21990–21995 biochem 18:554–562
5. Housden NG, Hopper JTS, Lukoyanova N, 13. Danelon C, Brando T, Winterhalter M (2003)
Rodriguez-Larrea D, Wojdyla JA, Klein A, Probing the orientation of reconstituted mal-
Kaminska R, Bayley H, Saibil HR, Robinson toporin channels at the single-protein level. J
CV, Kleanthous C (2013) Intrinsically disor- Biol Chem 278:35542–35551
dered protein threads through the bacterial 14. Chen M, Li QH, Bayley H (2008) Orientation
outer-membrane porin OmpF. Science of the monomeric porin OmpG in planar lipid
340:1570–1574 bilayers. Chembiochem 9:3029–3036
6. Bayley H, Cremer PS (2001) Stochastic sensors 15. Marques EJ, Carneiro CM, Silva AS, Krasilni-
inspired by biology. Nature 413:226–230 kov OV (2004) Does VDAC insert into mem-
7. Harrington L, Cheley S, Alexander LT, branes in random orientation? Biochim
Knapp S, Bayley H (2013) Stochastic detection Biophys Acta 1661:68–77
of Pim protein kinases reveals electrostatically 16. Gutsmann T, Heimburg T, Keyser U, Mahen-
enhanced association of a peptide substrate. dran KR, Winterhalter M (2015) Protein
Proc Natl Acad Sci U S A 110:E4417–E4426 reconstitution into freestanding planar lipid
8. Ramsay WJ, Bayley H (2018) Single-molecule membranes for electrophysiological characteri-
determination of the isomers of D-glucose and zation. Nat Protoc 10:188–198
D-fructose that bind to boronic acids. Angew 17. Zakharian E (2013) Recording of ion channel
Chem Int Ed 57:2841–2845 activity in planar lipid bilayer experiments.
Methods Mol Biol 998:109–118
Chapter 6
Revelation of Function and Inhibition of Wza Through

Single-Channel Studies
Lingbing Kong
Abstract
Antibacterial resistance (AR) is causing more and more bacterial infections that cannot be cured by using
the antibacterial drugs that are currently available. It is predicted that 10 million people will die every year
by 2050 from infections caused by antibacterial resistant strains, surpassing the predicted numbers of deaths
caused by cancer. AR is therefore a global challenge and novel antibacterial strategies are in high demand.
To this end, the work on exploring the pore properties of a bacterial sugar transporter, WzaK30, has led to
the discovery of the first inhibitor against bacterial capsular polysaccharides export.
Recently, single-molecule recapitulation of capsular polysaccharide (CPS) export and pore formation
properties of Wza barrel peptides have also revealed the possibility of a next-generation of Wza strategies.
These strategies are based upon the first examination and understanding of the pore properties of wild-type
(WT) and mutant WzaK30 in single-molecule electrical channel recording. The initially reported experi-
mental procedures have been further developed to enable efficient studies of other Wza homologs that are
more common in bacterial pathogens causing significant bacterial infections. Therefore, this chapter
presents the most recent protocols and logistics behind the research on Wza channel activity, antibacterials,
and strategies. The disciplines covered here include computation, molecular biology, biochemistry, elec-
trophysiology, microbiology, and biophysics.
Key words Single-molecule, Antibacterial, Wza, Capsular polysaccharides, Inhibition
1 Introduction
Bacteria have various defense mechanisms that enable them to resist

environmental challenges, which include immune response agents
inside animal hosts. Among the defensive machineries in Gram-
negative bacteria, the outermost layer called the outer membrane is
often a thick sugar-rich layer which forms a barrier to regulate the
interaction of the organism with environmental stimuli such as
antibodies, chemokines, cytokines, or antibacterials. Capsular poly-
saccharides (CPS) and exopolysaccharides (EPS) are the most com-
mon sugar moieties in the outer membrane of Gram-negative
bacteria. In the biosynthesis of the outer membrane, CPS and
EPS moieties must be exported from the cell [1]. One lipoprotein,
63
64 Lingbing Kong
Wza, is the translocon essential for the CPS export in various

bacteria [2]. Research has shown that blocking channels could be
a strategy to combat pathogens. For example, the anthrax PA63
channel has been studied by single-channel recording and a library
of blockers have been prepared and tested for this channel
[3]. Some of these compounds, especially the positively charged
beta-cyclodextrins, have shown to be effective against the anthrax-
causing bacteria at concentrations in the low nanomolar level.
In a similar fashion, specifically blocking the Wza channel and
therefore the translocation of CPS would inhibit the formation of
the outermost capsule protective layer of the bacteria. In this way,
targeting Wza would serve as a novel antibacterial strategy to
combat the rise of antibacterial resistance [4, 5]. The pore proper-
ties of various mutant WzaK30 pores as well as proteolytically
cleaved WT pores by proteinase K in situ were examined previously
[6]. The WzaK30 mutants were then used in a screen of blockers and
identification of binding sites in the pore. In this chapter, we
outline the techniques used to characterize Wza channel activity
and blockage on a single-molecule level [6, 7].
2 Materials
2.1 Reagents 1. Computational software: PYMOL, Modeller, HOLE,

and Kits for Plasmid AutoDock.
Construction 2. Competent cells for bacterial transformation with WzaK30 and
and Extraction other Wza homologs: XL10-Gold or DH5α E. coli from Agi-
lent or NIPPON GENE CO., LTD.
3. Polymerase chain reaction (PCR) for mutated WzaK30 and
other Wza constructs: Phusion High-Fidelity PCR Kit and
TA Cloning™ Kit with pCR™2.1 Vector.
4. Purification of PCR products: agarose gel electrophoresis with
Gel Extraction Kit.
5. Bacterial culture and plates: The LB-Agar plates and LB
medium are placed in an Incubated/Refrigerated Stackable
Shakers at typically 37 C.
6. Plasmid DNA extraction kits: Miniprep and/or Maxiprep kits.
2.2 Reagents 1. IVTT system: E. coli T7-S30 Extract System for Circular DNA
for Protein Expression kit (Promega).
2. SDS-PAGE: Criterion™ XT Bis-Tris precast gels (4–12% gel).
3. 20 XT MOPS running buffer (Biorad).
4. Gel drier with vacuum (e.g., the Model 583 gel dryer from
Bio Rad).
5. TE buffer: 10 mM Tris–HCl, pH 8, 1 mM EDTA.
Wza Single Channel Studies 65
6. 4 Laemmli Sample Buffer: 200 mM Tris–HCl, pH 6.8,

400 mM DTT, 8% w/v SDS, 0.4% bromophenol blue, 40%
glycerol.
7. Radiolabel: L-(35S)-Methionine.
8. Amino acid mixtures (minus methionine; minus leucine).
9. Centrifuge (25,000 g).
10. Whatman filter paper and X-ray film.
11. Microcentrifuge microfiltration device (0.45 μm pore size).
12. NanoDrop or UV-Vis spectrophotometer.
2.3 Reagents for Wza 1. High salt buffer for initial channel studies: 2 M KCl, 5 mM
Channel Studies HEPES, pH 7.5.
2. Low salt buffer for evaluation of blockers under physiological
conditions: 300 mM KCl, 5 mM HEPES, pH 7.5.
3. Oil: Hexadecane (10% v/v in pentane), Lipid: DphPC
(10 mg/mL in pentane).
4. A planar bilayer setup: Two chamber Delrin chip.
5. Silver/silver chloride electrodes.
6. Methanethiosulfonate (MTSES).
7. Wza blocker: Octakis(6-deoxy-6-amino)cyclomaltooctaose
(am8γCD).
8. Equipment: Axopatch 200B (Axon Instruments), Digidata
1500B converter (Axon Instruments).
2.4 Reagents for In 1. Pro-Q Emerald 300 Glycoprotein Stain Kit (Molecular
Vivo Blocker Inhibition Probes).
(Polysaccharide 2. M9 minimal growth media (Sigma) for growth of E69 strain
Analysis) with or without blocker.
3. Small molecules to be tested for Wza blocking activity.
4. 10% SDS-PAGE gel.
5. Gel imager suitable for detecting fluorescence in 500 nm range.
3 Methods
An outline of the methods referenced here is shown in Scheme 1.

3.1 Computational The availability of a high-resolution full structure of the WT Wza
Analysis protein is crucial for this development. However, the X-ray crystal
for Open-Form Wza structure of WzaK30 is incomplete. It is therefore necessary to create
Mutants and Blocker a homology model. This homology model is used to base the
Design design of Wza mutants and the selection of blockers for screening.
66 Lingbing Kong
Scheme 1 Flowchart to discover initial drug candidates against membrane

protein channels. This flowchart summarizes the efficient development of the
strong inhibitor of the WzaK30 channels and cellular functions in live bacteria. It is
presented in a way that is applicable not only to the Wza homologs but also other
membrane proteins, especially multimeric membrane proteins forming stable
pore structures in cells
1. Download the protein structure from the protein data bank

(PDB): ID 2j58.pdb for WzaK30 from https://www.rcsb.org/
structure/2J58.
2. Generate the homology model for the desired Wza by using
Modeller (see Note 1).
(a) Download the latest version of Modeller from https://
salilab.org/modeller/download_installation.html.
(b) Follow the instructions at: https://salilab.org/modeller/
release.html.
3. Perform local relaxation or energy-minimization of the homol-
ogy model by using Modeller. However, a newer common
practice is to perform molecular dynamics with the protein in
explicit water (see Note 2).
4. Obtain channel or pore diameters by manual measurements in
PyMOL or plotted with software like HOLE. These values,
especially those at the constriction sites (the narrowest sites),
help in the prediction of potential channel current levels, sig-
nal-to-noise ratios, and stability.
5. Design initial channel blockers based on Wza structure using
AutoDock and a template molecule such as γ-cyclodextrin
[6]. The constriction sites of a channel are more stable than
many soluble drug target proteins whose binding sites could be
simply at the interface of a homo-dimeric protein complex.
Most of the pore sizes of a multimeric channel are relatively
unchanged (see Note 3). The criteria to select initial blockers is
based upon size, shape, and charges or polarities.
(a) Size: The relatively rigid constriction site of the Wza pro-
tein is at the alpha-helix barrel. The constriction site of
WT WzaK30, for example, is 17 Å at the barrel. Therefore,
a potential strong blocker should be in the same or similar
size (see Note 4).
(b) Shape: The Wza channels are multimeric and symmetric
protein complexes (eightfold symmetry). The potential
blockers should therefore be ideally cyclic and eightfold
symmetric (see Note 5).
(c) Charges or polarities: The affinities of Wza channels to
potential blockers follow the general binding energies of
bimolecular interactions, i.e., electrostatic interactions >
hydrogen bonds > hydrophobic interactions. In the case
of WzaK30, the residues at the constriction site of the
alpha-helix barrel are composed of amino acids E369
and Y373, whose side chains are negatively charged in
the case of E369 or act as hydrogen donor or acceptor in
the case of the sidechain oxygen on Y373 (see Note 6).
68 Lingbing Kong
3.2 Mutagenesis 1. Using the Wza homology model, identify residues in the Wza
of Wza Channels Using channel to mutate in order to increase the size of the channel
Molecular Biology constriction site. There are two constriction sites at WT
Techniques WzaK30, one at the barrel site at Domain 4 constructed by the
E369 or Y373 residues while the other at a loop site at Domain
1 constructed by the Y110 residues (Fig. 1) (see Notes 7
and 8).
2. Analyze the restriction map of the Wza containing plasmid
especially at the sites of interest to make sure the cloning and
restriction digestions are as predicted.
3. Perform polymerase chain reaction (PCR) to acquire the
desired constructs. The vector used for cloning should ideally
contain an Ampicillin resistance gene. The general procedure
for PCR is a two-tube approach (see Note 9).
(a) In the first PCR reaction, mix the Forward primer that
contains a mutation with the Reverse primer that anneals
to the Ampicillin resistance gene.
Fig. 1 Domains of WzaK30 proteins (PDB: 2j58; modified with Modeller9.14). The
domain in red (Domain 4) is the transmembrane domain facing the extracellular
environment in live bacteria. The domain in blue (Domain 1) is the periplasmic
domain located at the periplasm between the outer membrane and inner
membrane
(b) In the second PCR reaction, mix the Reverse primer that
contains the mutation with the Forward primer that
anneals to the same site in the Ampicillin resistance gene.
4. If desired PCR products obtained, mix the two reaction mix-
tures in a 1:1 mole ratio and then digest with DpnI for 1–3 h.
Transform the PCR products using suitable competent cells
and subsequently miniprep to acquire the desired plasmid
construct.
3.3 Protein 1. Perform IVTT reactions by using the E. coli T7-S30 Extract
Expression System for Circular DNA kit (Promega). Place the reagents on
and Purification ice before use.
2. Adjust the concentration of circular DNA to 100 ng/μL. Mix
together 800 ng DNA with 2.5 μL amino acid mixture minus
Methionine, 2.5 μL amino acid mixture minus Leucine, 2 μL L-
(35S)-Methionine, 20 μL S30 premix and 15 μL T7 S30 extract
(see Note 10).
3. Place the 50 μL reaction mixture at 37 C and incubate for 3 h.
4. Centrifuge the reaction mixture at 25,000 g for 20 min.
Resuspend the pellet, which contains Wza monomer and octa-
mers, in 4 Laemmli Sample Buffer (add 12 μL buffer per
50 μL reaction).
5. Perform SDS-PAGE at 35–50 V at 4 C in 1 MOPS running
buffer, which requires about 10 h or overnight to complete (see
Note 11).
6. Place the gel on Whatman filter paper and dry the gel for about
7 h at room temperature (see Note 12).
7. Identify and extract octameric Wza proteins. Develop an auto-
radiograph of the gel in a darkroom and use the autoradiograph
to locate the Wza octamer on the gel. The Wza octamer runs
larger than 100 kDa [6].
8. Excise the gel and hydrate the dried gel piece attached to filter
paper in 100 μL TE buffer. Remove the filter paper and in an
Eppendorf tube crush the gel into very small pieces with a
homogenization pestle. Incubate the sample overnight at
room temperature (see Note 13).
9. Remove the gel pieces by centrifuging in a microcentrifuge
microfiltration device at around 5000 g for 5 min.
10. Aliquot the filtrate and store the samples at 80 C. The ideal
concentration of the Wza octamer is approximately
100–400 μg/mL measured by using a Nanodrop.
3.4 Single-Molecule 1. Set up the following parameters on the single-channel record-

Electrical Channel ing instrumentation (i.e., the Axopatch 200B patch-clamp
Recording Protocols amplifier, pClamp software)
70 Lingbing Kong
(a) Axopatch 200B patch-clamp amplifier:

CONFIG: WHOLE CELL β ¼ 1.
Output gain: 5.
Lowpass Bessel filter: 1 kHz.
Scaling factor (V/pA): 0.001.
(b) pClamp parameters:
Acquisition mode: Gap-free.
Sampling rate: 5 kHz.
2. Preparation of the bilayer:
(a) Coat the aperture with a thin layer of hexadecane by
adding 2 μL of (1:10) hexadecane pentane solution.
(b) Add 1 mL of high salt buffer or low salt buffer to either
chamber of the planar bilayer setup. Add DTT (200 μM)
to the buffers when a Wza mutant with a cysteine muta-
tion is used (see Note 14).
(c) Add two drops (~10 μL) of DPhPC to each chamber.
Allow 5 min for the pentane to evaporate. Pipet in both
chambers to generate a bilayer.
3. Single-channel recording of Wza channels:
(a) Add Wza octamer stock (1 μL) to the cis chamber (the
control or ground chamber) and mix. Apply a positive
potential to assist insertion of a Wza channel into the
bilayer.
(b) Upon insertion of a single Wza channel, apply positive and
negative electrical potentials (50 mV) from the trans
chamber to obtain the electrical responses of the channel,
the channel properties of the Wza channels.
(c) Confirm the current characteristics of the K375C or
K375C/Y110G mutants: Add 1.5 mM of methanethio-
sulfonate (MTSES) reagent to trans chamber when one or
several Wza proteins insert into the bilayer. The MTSES
reagent typically blocks the pore completely when reacted
with the eight K375C residues of the Wza octamer. To
reverse the MTSES reaction with the Wza cysteines and
reopen the pores, mix in 10 mM DTT to trans chamber
(see Note 15).
(d) Test the blocker efficiency against various octameric Wza
mutants, e.g., WT Wza, Y110G, K375C/Y110G. An
ideal blocker causes the current of the channel to drop
significantly and reversibly (see Notes 16 and 17).
(e) Upon identification of a suitable blocker, test a series of
concentrations of the blocker with the same single chan-
nel to obtain the kinetic details of the binding events.
(f) Perform “Single-channel search” with the single-channel

traces of the Wza channel interacting with the blocker in
pCLAMP 10.0 software. Extract the dwell time of the
“off” events and plot with a suitable bin and fit with an
exponential equation to acquire the τoff of each set of
binding events. koff is then obtained by 1/τoff. Extract
the dwell time of the “on” events in a similar fashion.
Plot the reciprocal of the dwell time of the “on” events,
i.e., 1/τon against different concentrations of the blocker.
The slope of this plot is the kon. Calculate the Kd by
dividing the koff by the kon. At least three repeats (nr) are
required to acquire an error (E) so the dissociation con-
stant is expressed as Kd (averaged) E (n ¼ nr) (see
Note 18).
3.5 Live Bacteria The experimental protocols are optimized for the E. coli K30 E69
Inhibitory Assay strain (see Note 19).
Protocols
1. Bacteria growth and polysaccharide extraction
(a) Grow the target pathogenic bacteria in M9 minimal
medium in appropriate volume at 37 C to reach an
OD600 between 0.5 and 1.0 (see Note 20).
(b) Dilute the culture to give an approximate OD600 ¼ 0.005.
Transfer 495 μL of this bacterial culture to an
Eppendorf tube.
(c) Add a solution of the blocker (5 μL) in a series of final
concentrations from 10 μM up to 1 mM.
(d) Place the samples in an incubator at 37 C with shaking at
300 rpm.
(e) After overnight growth, adjust the bacterial cultures to
OD600 ¼ 0.5.
(f) Transfer 500 μL of the culture from each sample to a new
Eppendorf tube. Centrifuge at 5000 g for 20 min.
(g) Resuspend the pellet in 500 μL PBS. Mix with 500 μL of
PBS-equilibrated phenol.
(h) Heat the suspension to 65 C for 15 min (see Note 21).
Mix the samples every 5 min until a homogeneous solu-
tion is achieved.
(i) Centrifuge the solution at 4 C for 15 min. Transfer the
supernatant (upper layer) to new Eppendorf tubes.
(j) Wash the supernatants with dichloromethane twice to
remove any residual phenol (see Note 22). Store the
supernatant at 80 C.
72 Lingbing Kong
2. SDS-PAGE analysis of CPS content:

(a) Dilute 5 μL polysaccharide extract in 5 μL PBS and then
add 4 Laemmli sample loading buffer (5 μL). Load the
samples onto a 10% SDS-PAGE gel.
(b) Perform SDS-PAGE at constant current 20 mA at room
temperature for about 80 min.
3. Staining and visualization:
(a) Rinse the gel in distilled water and then stain it with the
pro-Q Emerald 300 glycoprotein stain kit following man-
ufacturer’s instructions. For example, oxidize for 30 min,
wash three times, stain for 60 min, wash three times, and
then immediately image.
(b) Image the fluorescence in the 500 nm range with any UV
lamp with excitation around 300 nm.
(c) Analyze the fluorescence intensities of the CPS bands by
using a modern image processing software such as
Photoshop.
(d) Plot the fluorescence intensities against the concentra-
tions of the blocker to obtain the half inhibitory concen-
tration (IC50).
4 Notes
1. The amino acid sequence in the WT WzaK30 used for Modeller

is based on the complete gene sequence of the protein in
E. coli E69.
2. The flexibility of microstructures of a membrane protein
inserted in a relevant lipid bilayer and in an explicit solvent
guides the proper understanding of protein functions in its
natural cellular environment and thus the design and screening
for a potential blocker.
3. It is common to encounter a protein structure from a protein
PDB file or a snapshot out of a molecular-dynamics simulation
with diminished symmetry. It is a routine practice to use per-
fectly circular protein structures and potential blockers for
possible binding evaluations, such as Docking studies. How-
ever, it is important to also evaluate how difficult or easy these
conformations are achieved through conformational changes
from those with minimal energies.
4. Protons need to be added back to the structures so that poten-
tial hydrogen bonds could be evaluated accordingly.
5. Dendrimers with plenty of surface primary amines, such as
DAB-Am-16 and DAB-Am-32, do block irreversibly to the
Wza channels with unaltered barrel structures on an irregular

basis. These molecules do not have much value for use to
inhibit the Wza functions in live bacteria and are therefore
not pursued.
6. The pH of the biological environment of Wza is essential for
the determination of the protonation or the deprotonation
state of the relevant amino acids and the potential blockers.
For example, the E369 residues at WzaK30 should therefore
adopt the deprotonated state (–COO) at a slightly basic solu-
tion or environment (pH 7.5, for example), while the primary
amines at other residues and the C-6 position of blocker
Am8γCD should all be protonated (–NH3+).
7. WT WzaK30 pores do not generate stable ionic currents which
is needed for accurate analysis of a small molecule library screen
[6]. It is necessary therefore to acquire open-form WzaK30
pores through mutagenesis. For general open-form Wza
mutant channels, the guideline for mutagenesis is to acquire
stable channels with minimum changes to the structure of the
pore. This is because the aim of protein engineering is to find
blockers that are more likely to work on the WT target in live
cells. The mutants with minimum changes to the WT Wza
protein will increase the chance of successful blocker screening.
8. The X-ray crystal structure of wild-type WzaK30 (WT-WzaK30)
(PDB ID: 2j58) shows two sites relatively narrower than the
rest, one being the loop region of Domain 4 at the periplasmic
side while the other being around the middle of the alpha-helix
barrel. Previous reports have shown that mutation of residue
Y110 has little effect on Wza protein structure. This point
mutation is selected at the constriction site at the D4 domain,
which has flexible loops. The ring of Y110 residues in the
WzaK30 channel maintains a diameter of 3.7 Å. Using Modeller
to determine the diameter of the Y110G mutant, previous
work reveals that the Y110G mutation yields a new diameter
of 10.5 Å [6]. Other mutants K375C and K375C/Y110G
were also engineered. It is crucial to validate the properties of
new Wza mutants. Therefore, in the case of WzaK30, the intro-
duction of the K375C mutation at the entrance of the pore is to
allow the identification of the current level, signal/noise ratio,
and stability of the WzaK30 channels [6].
9. When the two-tube PCR approach does not afford the desired
constructs, an alternative design of PCR primers for a TA
ligation followed by site-specific digestion and ligation can be
carried out. Generally, the two-tube approach is quick and
efficient for simple constructs.
10. Alternatively, nonradioactive L-Methionine can be used in the
IVTT reaction. After running the gel, stain the gel using
74 Lingbing Kong
standard Coomassie methods. After the gel is stained, excise

the desired band containing Wza proteins and extract the Wza
proteins as outlined in Subheading 3.3, steps 8–10. This
method avoids the use of radioactive materials and there is no
need to establish relevant facilities and regulations.
11. Limiting the applied voltage in SDS-PAGE to 35–50 V
increases the yield of functional octameric WzaK30. A lower
voltage than 35 V tends to trigger errors and lead to failures in
power supply at a later stage. Increasing the voltage may be
required after several hours; otherwise, the power supply may
stop because of too low current.
12. A simplified approach for small gel piece drying is to use
vacuum Büchner funnel connected to an oil pump. In either
way, the gel or gel piece is placed on the filter paper, which is
covered by a small layer of cling film and a silicone rubber sheet
that covers the whole area of the filter paper or any area
necessary to prevent the vacuum from leaking.
13. It is important to leave the crushed gel pieces in a buffer on
bench for at least an hour to allow sufficient diffusion of
protein into the buffer.
14. EDTA has no effect on the channel properties, so there is no
need to include it in the buffers.
15. The reaction of the cysteine residues with MTSES to form
disulfide bonds is usually very rapid, within a few minutes.
The cleavage of the disulfide bonds by addition of DTT takes
much longer. It could take over 10 min for all the disulfide
bonds to be cleaved, as indicated by the current level returning
to the value before the addition of MTSES.
16. Blockers differ in binding constants. Dendrimers that cause
irregular and irreversible blockades are not considered as an
ideal blocker.
17. A charged channel blocker satisfies the Woodhull’s model for
voltage-dependent binding of a charged channel blocker.
18. The repeats refer to individual Wza pores, i.e., three repeats of
the blocker concentration test must be performed by three
independent pores. Therefore, all independent repeats of the
Wza channel should be included equally in the initial evalua-
tion of the channel properties.
19. On-gel carbohydrate staining offers an efficient way to quantify
the capsular polysaccharides (CPS) extracted from live bacteria
grown in the presence or absence of the Wza blockers. The
pro-Q Emerald 300 Glycoprotein Stain Kit (Molecular Probes)
is suitable staining for CPS moieties with at least two vicinal
hydroxyls in each repeating unit. This is because the oxidant in
the pro-Q Emerald 300 Glycoprotein Stain Kit (Molecular
Probes) cleaves the –C(OH)–C(OH)– bond to generate alde-

hyde groups that will react with the primary amine groups of
the fluorescent reagent to generate fluorescent signals at the
CPS band on a SDS-PAGE gel. The K30 CPS repeating unit,
for example, has two pairs of these adjacent hydroxyl groups,
both at the glucuronic acid, i.e., the C2 and C3 hydroxyls and
the C3 and C4 hydroxyls. Other bacteria with different CPS
characteristics should be re-evaluated to determine the best
conditions for carbohydrate staining.
20. M9 minimal medium is much cleaner than LB medium. The
inhibitory effects of the blockers obtained with M9 minimal
medium match well with those in complement-mediated kill-
ing. The results gained in the LB medium are meaningful at the
qualitative level; however, attempts to obtain quantitative data
in LB medium encountered difficulties, probably due to the
complex mixture of the yeast extract. LB medium is therefore
not used for quantitative experiments to determine the inhibi-
tory concentration of a blocker.
21. Use caution when handling phenol. Perform this step in a
chemical hood, as phenol is an irritant.
22. The high-speed centrifuge tubes used here should be made of
polypropylene plastic instead of polycarbonate. The latter
tends to become dissolved in the dichloromethane and leads
to loss of samples.
Acknowledgements
Professor Kong is currently a tenure-track Assistant Professor and

Principal Investigator at Kagawa University and the International
Institute of Rare Sugar Research and Education (IIRSRE). He is
grateful for the start-up funds supported by Kagawa University and
IIRSRE, which enabled the purchase of a patch clamp system and
essential accessories for the establishment of the single-molecule
electrical channel recording studies in the lab. The JSPS grants
(Startup grant 17H06912 and Basic Research (B) grant
18H02109) for Kong are also highly appreciated, which are sup-
porting the establishment of a multidisciplinary lab for the studies
of the function and inhibition of Wza homologs other than WzaK30
that are causing more bacterial infections. Professor Kong also is
grateful to Professor Hagan Bayley at the University of Oxford for
his generous gift of all the Wza plasmids as well as some basic
accessories for the planar bilayer system developed during Professor
Kong’s PhD and postdoctoral research.
76 Lingbing Kong
References
1. Whitfield C (2006) Biosynthesis and assembly of 5. Kong L, Vijayakrishnan B, Kowarik M, Park J,
capsular polysaccharides in Escherichia coli. Zakharova AN, Neiwert L, Faridmoayer A,
Annu Rev Biochem 75:39–68 Davis BG (2016) An antibacterial vaccination
2. Dong C, Beis K, Nesper J, Brunkan- strategy based on a glycoconjugate containing
Lamontagne AL, Clarke BR, Whitfield C, Nai- the core lipopolysacchride tetrasaccharide
smith JH (2006) Wza the translocon for E. coli Hep2Kdo2. Nat Chem 8:242–249
capsular polysaccharides defines a new class of 6. Kong L, Harrington L, Li Q, Cheley S, Davis
membrane protein. Nature 444:226–229 BG, Bayley H (2013) Single-molecule interro-
3. Nestorovich EM, Bezrukov SM (2012) gation of a bacterial sugar transporter allows the
Obstructing toxin pathways by targeted pore discovery of an extracellular inhibitor. Nat Chem
blockage. Chem Rev 112:6388–6430 5:651–659
4. O’Neill J (2016) Tackling drug-resistant infec- 7. Kong L, Almond A, Bayley H, Davis BG (2016)
tions globally: final report and recommenda- Chemical polyglycosylation and nanoliter detec-
tions. Review on Antimicrobial Resistance. tion system enables single-molecule recapitula-
Government of the United Kingdom tion of bacterial sugar export. Nat Chem
8:461–469
Chapter 7
Protein Analyte Sensing with an Outer Membrane Protein G

(OmpG) Nanopore
Monifa A. V. Fahie, Bib Yang, Christina M. Chisholm, and Min Chen
Abstract
Nanopore sensing is a powerful lab-on-a-chip technique that allows for the analysis of biomarkers present in
small sample sizes. In general, nanopore clogging and low detection accuracy arise when the sample
becomes more and more complex such as in blood or lysate. To address this, we developed an OmpG
nanopore that distinguishes among not only different proteins in a mixture but also protein homologs.
Here, we describe this OmpG-based nanopore system that specifically analyzes targets biomarkers in
complex mixtures.
Key words Outer membrane protein G, Sulfhydryl-maleimide chemistry, Protein-protein interac-

tions, Nanopore gating, Extracellular loops, Loop dynamics, Biotin, Antibody, Streptavidin, Avidin,
Electrophysiology
1 Introduction
Nanopore sensing is an attractive analytical tool in the field of

diagnostics. It has been used to identify small molecules [1, 2]
and biomolecules [3–8]. Nanopores have been engineered from
either pore-forming proteins, synthetic materials, or even 3D-DNA
structures [9–23]. One of the most successful applications of nano-
pore sensing has been in nucleic acid sequencing [24]; however
nanopore studies for protein sensing have gained considerable
interest and development recently [24–29]. Proteins can be
detected by (1) translocation or interaction within the walls or
constriction site of the nanopore, (2) binding to external adaptors
where a change in the adaptor is transduced to the nanopore, and
(3) binding directly to the outer surface of the nanopore.
For a translocating protein, the ionic blockade that is generated
in the nanopore is sensitive to a myriad of features on the protein
such as its folding state, monomeric/multimeric state, posttransla-
tional modifications, or distribution of surface charge [8, 30–
33]. Thus, the same protein could generate different ionic
77
78 Monifa A. V. Fahie et al.
blockades as it passes through a nanopore depending on its state.

This can make detection specificity challenging.
Specialized nanopores are those decorated with high affinity
molecules that make them specific to only a few analytes. These
types of nanopores have been mainly used to detect the native/
active state of protein analytes. Once the protein is bound to the
high affinity molecule this can induce a change in the nanopore via
an adaptor [5] or via direct binding [34–37]. The current method
described in this chapter is an example of a specialized nanopore
that discriminates among protein variants through direct interac-
tion with analyte.
Recently, we engineered a nanopore based on outer membrane
protein G (OmpG) from Escherichia coli (E. coli). This nanopore
has a unique feature where it generates a fluctuating current unlike
other common nanopores. We exploited the dynamic nature of the
seven flexible extracellular loops of OmpG to capture specific ana-
lytes (Fig. 1). Analytes are distinguished through changes in loop
dynamics and protein-protein interaction with the OmpG loop
surface [33, 36, 38, 39]. We observed that a single OmpG con-
struct could distinguish among several proteins even structural
homologs (Fig. 2). Because of the sensing power of OmpG, it is
an exciting platform on which to build nanopore sensors for sensi-
tive and specific analysis.
2 Materials
2.1 Cloning 1. pT7-OmpGwt plasmid.

of Cysteine 2. Oligonucleotides for mutagenesis.
Mutant OmpG
3. High fidelity polymerase.
4. Restriction enzymes such as DpnI.
5. Chemically competent DH5α cells.
6. Ampicillin (100 mg/mL).
2.2 Protein 1. Chemically competent BL21 DE3 cells.

Expression 2. LB media (Miller version): Tryptone 10 g/L, yeast extract 5 g/
of OmpG-Cysteine L, sodium chloride 10 g/L.
Mutant
3. Isopropyl β-D-1-thiogalactopyranoside (IPTG).
2.3 Preparation 1. Storage buffer: 50 mM Tris–HCl, pH 8, 150 mM NaCl,

of OmpG-Biotin or 1 mM EDTA.
OmpG-Sulfonamide 2. Lysis buffer: 50 mM Tris–HCl, pH 8, 150 mM NaCl, 200 μg/
Nanopore mL lysozyme, 1 mM EDTA, 3 mM TCEP.
3. Inclusion Body (IB) wash buffer: 50 mM Tris–HCl, pH 8,
1.5 M urea, 1 mM TCEP.
OmpG Nanopore Detection of Protein Analytes 79
(a) Extracellular side Loop 6

Loop 5 D D
Loop 4 Q I Loop 7
Loop 3 S R
T T W E
Loop 2 D E
Loop 1 E P N G N D
R D G
G D G L D N
G E A K Y W E H D
K D M G
Y D S R V D
A
N E
N G D S
G M Y G Y N D S Q
Outer membrane envelope

T F H D
F
E D D T H A
A T G
W
D W K
Y S R
V G V W G F R R F E F
N T D
N K V N F H
L P F Y M E Q L
E A G Y E R A Y
D N Q L E G
I E R M T V Y A
E R R Y I I
E P S E E G
P Q F W Y R R G
Y L T L V
S Y E G K N A T L
M W G F S G
V Y L G I V Y Y
A G L V V
Y A E T R L P Y
G A N Q S N
F L V L L P G
I P F T L Y
N A H Y A Y
G D L V
N R T G S
A I Y F W V T S D
F A Q L F N F
R S D T L H F
H D N E Q
N W F F V G N
W L D
G P D K
D E T
D L
N N
R
E
Periplasmic turns
E
extracellular loops
(b) C224 (c)
open
C224
loop 6
loop 6
10pA 0 pA
1s
periplasmic turns
Fig. 1 Outer membrane protein G structure and characteristic gating pattern. (a) Flat representation of the
OmpG protein amino acid sequence. The peptide sequence of the seven loops are highlighted in color and the
residues used to mutate into cysteine are highlighted in white bold lettering. (b) Top and side view of the OmpG
crystal structure. (c) Characteristic OmpG gating signal in the current recording
4. Binding buffer: 50 mM Tris–HCl, pH 8, 3 mM TCEP,

8 M urea.
5. Probe Sonication system.
6. 0.45 μm filters.
7. Anion exchange columns (Q sepharose).
8. Desalting columns (5 mL, GE).
9. Chromatography gravity/centrifuge columns (15 mL).
10. Falcon tubes (50 mL).
(a) (b)
20s
30
no binding 0nM
I (pA)
20
10
0
30
mAb binding 1nM
I (pA)
20
10
0
30 5nM
I (pA)
20
no binding polyAb.1 mAb polyAb.2 10
0
30 10nM
I (pA)
20
10
// // // 0
Fig. 2 Detection of different antibodies binding with OmpG-biotin construct. (a) Schematic of OmpG and
antibody reversible interactions and corresponding current recording traces of no binding and binding signals.
(b) Concentration-dependent binding of monoclonal antibody (mAb) with OmpG-biotin construct
11. Wash buffer: 50 mM Tris–HCl, pH 8, 3 mM TCEP, 8 M urea,

75 mM NaCl.
12. Elution buffer: 50 mM Tris–HCl, pH 8, 3 mM TCEP, 8 M
urea, 200 mM NaCl.
13. High salt buffer: 50 mM Tris–HCl, pH 8, 3 mM TCEP, 8 M
urea, 500 mM NaCl.
14. Labeling buffer: 50 mM HEPES, pH 7.0, 150 mM NaCl,
8 M urea.
15. Labeling reagents: Maleimide-PEG2-Biotin for preparing
OmpG-biotin, Maleimide-PEG4-sulfonamide for preparing
OmpG-sulfonamide.
16. Refolding buffer: 50 mM Tris–HCl, pH 9.0, 3.25% octyl glu-
coside (OG).
17. Trypsin enzyme. Trypsin activity buffer: 50 mM Tris–HCl,
pH 8, 1 mM CaCl2.
18. Iodoacetamide.
19. Chymotrypsin enzyme. Chymotrypsin activity buffer: 100 mM
Tris–HCl, pH 8, 10 mM CaCl2.
20. Bradford assay reagents.
21. SDS-PAGE materials and electrophoresis equipment.
2.4 Current 1. Phospholipids: 10 mg/mL diphytanoylphosphatidylcholine

Recording (DPhPC) dissolved in pentane.
of OmpG-Biotin or 2. Oil: 10% v/v hexadecane dissolved in pentane.
OmpG-Sulfonamide
3. Polytetrafluoroethylene film (25 μm thickness) for the
Nanopore aperture.
4. Biotin-binding analytes—(a) streptavidin, (b) avidin, (c) mouse

anti-biotin monoclonal antibody clone BTN.4, (d) mouse anti-
biotin monoclonal antibody clone SB58C.
5. Sulfonamide-binding analytes—(a) human carbonic anhydrase
VII, (b) human carbonic anhydrase II.
6. Human blood (10%).
7. Equipment: two chamber chip, Ag/AgCl electrodes, Digidata
1320A/D board (Axon Instruments), Axopatch 200B inte-
grating patch clamp amplifier (Molecular Devices).
8. Software: latest versions of pClamp and Clampfit.
3 Methods
3.1 Substitution of a 1. Introduce single cysteine substitutions on either of OmpG’s

Single Cysteine seven loops by site-directed mutagenesis. The plasmid template
Mutation on OmpG is pT7-OmpG (see Note 1). The mutagenesis primers to
Protein mutate residues on OmpG loops are described in Table 1 (see
Note 2).
2. Use a standard polymerase chain reaction protocol from a
commercially available high-fidelity enzyme such as Phusion
to generate the PCR product with the desired mutation (see
Note 3).
Table 1
List of the oligonucleotides used to generate the OmpG mutants
OmpG mutations Mutagenic oligonucleotides

L1 (E24C) Forward: 50 -TATGGCTGTGATATGGATGGGCTG-30
Reverse: 50 -CATATCACAGCCATAACCCTCGAC-30
L2 (S58C) Forward: 50 -GATTATTGCGCGGGTAAACGTGGAACG-30
Reverse: 50 -ACCCGCGCAATAATCTACCGGCCC-30
L3 (E101C) Forward: 50 -GTTGATTGTCCGGGTAAAGACACG-30
Reverse: 50 -ACCCGGACAATCAACGTAGTGATAAC-30
L4 (T143C) Forward: 50 -CTGAACTGTACCGGTTACGCTGATAC-30
Reverse: 50 -ACCGGTACAGTTCAGATCGTTGGC-30
L5 (S182C) Forward: 50 -GACGACTGCCGCAATAACGGTGAG-30
Reverse: 50 -ATTGCGGCAGTCGTCCATATTGAAG-30
L6 (D224C) Forward: 50 -GGGACTGGCAGTGTGATATTGAACGTGAAG-30
Reverse: 50 -GTTCAATATCACACTGCCAGTCCCAGTTAC-30
L7 (G264C) Forward: 50 -CGACGAATGCGACAGTGATAAATTC-30
Reverse: 50 -ACTGTCGCATTCGTCGTGATCCTG-30
3. Digest the template DNA in the PCR reaction mix with DpnI
enzyme (see Note 4).
4. Transform the DpnI-digested PCR mix into competent E. coli
DH5α cells (or an equivalent cell strain) onto LB-Agar plates
with 100 μg/mL ampicillin.
5. Isolate single clones and identify the desired mutant construct
with restriction enzyme digestion and DNA sequencing.
3.2 Expression 1. Transform the validated pT7-OmpG cysteine mutant plasmid

and Purification into BL21 DE3 E. coli competent cells and plate onto LB-Agar
of Denatured OmpG plates supplemented with 100 μg/mL ampicillin (LB-Agar-
Cysteine Mutant from Amp). Incubate LB-Agar-Amp plate(s) for 14–18 h at 37 C
Inclusion Bodies (see Note 5).
2. Isolate one or more colonies from the BL21 transformation
3.2.1 Expression
and inoculate a starter culture of LB media supplemented with
of OmpG Proteins
100 μg/mL ampicillin. Allow the starter culture to grow 37 C
with agitation at 200 rpm for 12–16 h.
3. Inoculate a larger volume (at least ten times volume of starter
culture) of LB + Amp media with the starter culture. Allow cells
to grow at 37 C, 200 rpm shaking until the OD600 reaches
between 0.5 and 0.8 (see Note 6).
4. Add inducer isopropyl β-D-1-thiogalactopyranoside (IPTG) to
culture at a final concentration of 0.5 mM to induce the protein
expression and continue to shake culture at 37 C (see Note 7).
5. Harvest cells after 3–4 h of IPTG induction by centrifugation
at 3000 g for 20–30 min. Remove LB supernatant and
resuspend cell pellet in either storage buffer or lysis buffer (see
Note 8).
6. If the cell pellet is resuspended in storage buffer, keep pellet in
20 C freezer for approximately 2 months or until ready to
purify (Subheading 3.2.2, step 1). Once cell pellet is resus-
pended in lysis buffer, this lysate cannot be stored or frozen;
the inclusion body must be prepared immediately (Subheading
3.2.2, step 2).
3.2.2 Isolation of OmpG 1. Thaw OmpG cell suspension at room temperature if previously
Inclusion Body frozen in storage buffer (see Note 9). Add lysozyme to a final
concentration of 200 μg/mL and reducing agent (DTT or
TCEP) to a final concentration of 3 mM. Incubate the
OmpG-containing lysate at room temperature for 20 min or
37 C for 10 min to allow the lysozyme to break down the
cell wall.
2. Sonicate the cell mixture on ice to further break the bacterial
membranes and shear the nucleic acid to reduce the lysate’s
viscosity. Pulse 5 s on/5 s off for a total of 7 min (see Note 10).
3. Centrifuge the lysate at 20,000 g for 20 min. Discard the

supernatant.
4. Resuspend the pellet in inclusion body (IB) wash buffer. Incu-
bate at room temperature for 10 min.
5. Centrifuge pellet again at 20,000 g for 15 min. Discard IB
wash supernatant. The resulting inclusion body pellet can be
stored at 20 C for up to 4 months.
3.2.3 Anion Exchange 1. Solubilize the OmpG-containing inclusion body in binding

Purification of Denatured buffer (see Note 11) for at least 30 min at room temperature
OmpG Cysteine Proteins or until pellet is ~90% solubilized (see Note 12).
2. Centrifuge the solubilized inclusion body at 20,000 g for
15 min. Discard the pellet and pass the supernatant through a
0.45 μm filter before purification.
3. Equilibrate anion exchange Q beads with at least 5 column
volume of binding buffer.
4. Apply the filtered inclusion body solution onto the equilibrated
Q beads to allow the OmpG to bind. Collect the flow-through.
5. Wash the beads with 5 column volume of binding buffer.
Collect wash.
6. Wash the beads with 5 column volume of wash buffer. Collect
second wash.
7. Elute protein from beads with 3 column volume of elution
buffer into clean centrifuge tubes.
8. Wash beads with 5 column volume of high salt buffer to
remove any bound proteins or nucleic acids. Collect
third wash.
9. Analyze protein content for each wash or elution collection
with SDS-PAGE. A 12% or 15% gel is suitable for OmpG
analysis.
10. For short-term storage (48 h or less), place eluted protein in
4 C (see Note 13).
11. Clean the Q beads: Wash beads with 10 column volume of
distilled water. Check the manufacturer’s instructions on
appropriate cleaning solvents and protocols for the beads.
Store beads in 20% ethanol at 4 C for future use.
3.3 Ligand Labeling 1. Determine the protein concentration of the eluted OmpG
and Refolding of OmpG fraction(s) with a Bradford assay (see Note 14).
Proteins 2. Use about 0.5–1.0 mg of protein for the labeling reaction.
3.3.1 Conjugation Buffer exchange the OmpG using size exclusion chromatogra-
of Maleimide Ligand phy (SEC) column equilibrated with labeling buffer. Refer to
to OmpG Cysteine Mutant the manufacturer’s recommendations for appropriate volume
of protein to purify with the column. For example, if using a
5 mL HiTrap desalting column (GE), no more than 0.75 mL

protein may be injected onto column in order to efficiently
exchange the buffer (see Note 15).
3. To label the OmpG cysteine mutant with small ligands like
maleimide-PEG2-biotin or maleimide-PEG4-sulfonamide,
incubate the protein with excess ligand in a molar ratio of at
least 1:5 (protein to ligand) at room temperature (~23 C) for
1–2 h with moderate shaking (~100 rpm) (see Note 16).
4. Add DTT to a final concentration of 10 mM to quench the
excess maleimide ligand.
5. Buffer exchange in elution buffer without reducing agent by
SEC to remove excess free ligand.
3.3.2 Refolding 1. Dilute the labeled OmpG mixture with refolding buffer until
of Labeled OmpG Protein the final concentration of urea is approximately 3 M (see Note
17).
2. Incubate samples at 37 C for 48–72 h.
3. Test the refolding efficiency of the labeled OmpG proteins by
heat-denatured gel shift assay. Take one replicate of the
refolded OmpG protein (at least 2 μg) and prepare with
SDS-sample buffer. Take a second replicate of the OmpG
protein, mix with SDS-sample buffer, and boil for 10–15 min
at 95 C.
4. Run both boiled and unboiled proteins on an SDS-PAGE gel
and analyze the shift in apparent molecular weight of the
boiled vs. unboiled sample (Fig. 3).
5. Store refolded OmpG protein in small aliquots (10–20 μL) at
80 C for up to 1 month (see Note 18).
3.4 Testing Labeling 1. After diluting denatured OmpG into refolding buffer, mix
Efficiency of Refolded OmpG and streptavidin in a 1:1 molar ratio and incubate at
or Denatured OmpG 23 C for 2–5 min (see Note 19).
Proteins 2. Add SDS-sample buffer and run samples on a 12% or 15%
3.4.1 Labeling Efficiency
SDS-PAGE gel (Fig. 4). As a negative control, run the labeled
of Biotin-Labeled OmpG
OmpG minus the streptavidin to compare.
(Gel Shift Assay)
Loop 1 Loop 2 Loop 3 Loop 4 Loop 5 Loop 6 Loop 7

- + - + - + - + - + - + - + heat
kDa
34.6-
-denatured
-refolded
Fig. 3 Refolding efficiency gel analysis of OmpG-ligand constructs. Constructs were denatured at 95 C for
10–15 min and compared with refolded construct. Proteins were loaded onto a 12% or 15% SDS-PAGE gel
and protein mobility in gel was analyzed as an indicator for their folding status
Loop 1 Loop 2 Loop 3 Loop 4 Loop 5 Loop 6 Loop 7

- + - + - + - + - + - + - + streptavidin
kDa
-streptavidin-OmpG
55.6- complex
42.7-
34.6- -denatured
OmpG
Fig. 4 Labeling efficiency of OmpG-biotin loop constructs. OmpG-biotin constructs were heat denatured in
SDS-sample buffer at 95 C for 10–15 min. Thereafter, one sample was incubated with 1:1 mole ratio of
streptavidin protein for 5 min at ambient temperature, while no streptavidin was added to the other sample.
The constructs, streptavidin, were run on SDS-PAGE gels and gel mobility shifts in the presence of
streptavidin were analyzed as a marker for labeling efficiency
3. Using an imaging software, such as Image J, perform a densi-

tometric analysis on the OmpG only sample vs. the labeled
OmpG + streptavidin mixture to determine the labeling effi-
ciency estimate.
3.4.2 Proteolytic 1. Digest the labeled denatured OmpG, i.e., OmpG in 8 M urea
Digestion of Labeled with trypsin or chymotrypsin based on the manufacturer’s
OmpG-Sulfonamide protocol for the specific protease. Trypsin digestion is suitable
for Mass Spectrometric for analysis of peptides from loops 2, 3, 4, and 5. To analyze
Analysis loops 1, 6, and 7 digest the OmpG with chymotrypsin instead.
2. Dilute denatured OmpG in trypsin activity buffer so that the
urea concentration is 1 M or lower. Reduce any existing disul-
fide bonds with DTT at a final concentration of 10 mM (see
Note 20). Incubate at 23 C for 30 min, then add iodoaceta-
mide to a final concentration of 50 mM. Incubate at 23 C in
the dark for 1 h with gentle agitation.
3. Use a trypsin:OmpG mole ratio range of 1:20 to 1:75 and
incubate digestion mixture at 23 C for 12–16 h.
4. Alternatively, for cleavage with chymotrypsin, add DTT to a
final concentration of 5 mM to reduce any unlabeled OmpG
dimers in ~8 M urea. Incubate for 5 min.
5. Add 15 mM iodoacetamide and incubate for 15 min at room
temperature in the dark. Dilute the sample with chymotrypsin
activity buffer to obtain <1 M urea and add chymotrypsin to a
final protease:protein mole ratio of 1:20. Allow the digestion
mixture to incubate for 12–16 h at 23 C.
6. Perform mass spectrometric analysis on the digested OmpG
peptides to identify the existence labeled peptides compared to
unlabeled ones (see Note 21).
3.5 Current 1. Generate an artificial bilayer in a two-chamber chip separated

Recording of OmpG by a 25 μm thick polytetrafluoroethylene film with an aperture
Proteins of 100 μm diameter in the center [33, 36].
2. Gently apply 1–2 μL hexadecane/pentane (10% v/v) solution

to the polytetrafluoroethylene film to cover the aperture with a
layer of hexadecane. Incubate at room temperature for 2–5 min
(see Note 22).
3. Add appropriate volume of the buffer of your choice to each
chamber (see Note 23).
4. Carefully drop 10–20 μL of 1,2-diphytanoyl-sn-glycerol-3-
phosphocholine (DPhPC) dissolved in pentane at 10 mg/mL
onto the surface of the buffer to generate a monolayer of
phospholipids. Do not pipet the DPhPC lipids directly into
the buffer as the pentane needs to evaporate; therefore keep
pipet tip above the surface of the buffer. Allow the pentane to
evaporate for 2–5 min.
5. Immerse an Ag/AgCl electrode into the buffer of each cham-
ber, where one chamber is grounded. Slowly pipet up and
down the buffer mixture in each chamber to assist in the
formation of a monolayer in each chamber (see Note 24).
6. Once the bilayer is formed, test the stability at 200–300 mV.
Add OmpG proteins to a final concentration of 0.1–1 nM to
the chamber which is grounded (see Note 25). Apply a voltage
of 200–300 mV to assist pore insertion into the bilayer at the
aperture.
7. Lower the voltage to 20–70 mV for characterization once a
single OmpG protein has inserted into the bilayer. Filter the
current using a Bessel filter at 2 kHz and acquire the current at
a sampling rate of 50 μs.
8. Determine the orientation of the inserted OmpG pore [36, 40]
by analyzing the asymmetrical gating pattern at positive and
negative potentials (Fig. 5 and Table 2).
9. After determining the chamber in which the OmpG loops
reside, add protein analytes to the specific chamber (Table 3).
Mix the solution containing the protein analyte by carefully
pipetting up and down about 1/3rd of the volume of buffer in
the chamber (see Note 26).
10. As a negative control use the unlabeled OmpG to prescreen for
nonspecific interactions of desired protein analytes with the
OmpG nanopore (see Note 27).
11. For the OmpG-sulfonamide sensor, prepare a human blood
lysate sample. First resuspend blood in PBS buffer and centri-
fuge at 400 g for 3 min. Discard the supernatant. Repeat
wash one more time to remove any hemolyzed material. After
washing blood, flash freeze blood cells in liquid nitrogen.
Allow cells to thaw on ice. Centrifuge at 10,000 g to pellet
any cell debris and keep supernatant containing blood lysate.
trans pore
pH 5.0 pH 6.0 pH 7.0
60 80 turns
60
60
cis
I (pA)
I (pA)
I (pA)
+50 mV 40
40 40
20 20 20
0 0 0
0
V
0 0
-20 -20
trans
I (pA)
I (pA)
I (pA)
-50 mV -40 -40
-20
-40
-60 -60 -60 loops
biotin
200 ms
cis pore
pH 5.0 pH 6.0 pH 7.0 biotin
loops
60 80
60
60
I (pA)
I (pA)
I (pA)
+50 mV 40
40 40 cis
20 20 20
0 0 0
0 0 0
V
-20 -20
I (pA)
I (pA)
I (pA)
-20
-50 mV -40 -40 -40
trans
-60 -60 -60 turns
200 ms
Fig. 5 Orientation of OmpG constructs inserted into lipid bilayer. OmpG has asymmetrical gating behavior
termed quiet and noisy. A trans oriented pore is quiet (lower gating frequency) at negative voltage and noisy
(higher gating frequency) at positive voltage
Table 2
Asymmetrical characteristics of a trans OmpG nanopore at various pH in 1 M KCl
pH 5 pH 6 pH 7
Gating
characteristic 50 mV +50 mV 50 mV +50 mV 50 mV +50 mV
Open probability 0.17 0.04 0.50 0.06 0.73 0.10 0.88 0.04 0.94 0.02 0.94 0.02
Gating frequency 50.1 10.6 121.2 19.9 53.2 14.1 92.8 29.9 32.9 7.7 52.7 11.5
(s1)
Ton (ms) 3.4 0.7 4.1 0.9 15.6 3.3 11.3 1.8 29.3 4.7 15.6 3.0
Toff (ms) 15.8 2.6 3.9 0.7 4.5 0.8 1.3 0.2 1.9 0.5 1.1 0.2
12. Add an appropriate volume of blood lysate to the loop facing

chamber and slowly mix with a stir bar or by gently pipetting
up and down. In general, up to 30% maximum of blood lysate
is appropriate to maintain a stable bilayer. Record the OmpG
nanopore before and after addition of blood lysate and take
note of any changes in ionic signal.
Table 3
Characteristic current traces of analyte binding to OmpG nanopore sensor
OmpG construct Loops for sensing Analyte Characteristic signal

OmpG-biotin L6 only Streptavidin
OmpG-biotin All loops except L5 Avidin
OmpG-biotin All loops except L5 SB58C antibody
OmpG-biotin L2, L4, L6, L7 BTN.4 antibody
OmpG-biotin L3 BTN.4 antibody
OmpG-sulfonamide L3 hCAII
OmpG-sulfonamide All loops hCAVII
3.6 Analysis of OmpG 1. To determine the open conductance of OmpG, analyze at least
Current Recording 10 s of OmpG only trace and generate a histogram in Clampfit
Traces Post Recording software analysis program. Apply a Gaussian fit to the histo-
gram to obtain the value for the open conductance.
2. To analyze specific features of OmpG gating activity, use the
event detection tool in Clampfit to retrace the open and closed
behavior of the nanopore. Choose a specific region of the trace
to analyze using a cursor pair to mark the beginning and the
end of the analysis region. Optional: filter the analysis region of
the trace before applying the event detection program. Go to

the filter option in the Analyze drop down and choose low-pass
Gaussian at 1000 Hz 3 dB cutoff. Using single channel
search, assign the current levels for the open (0) state. For
example, if the maximum current of the pore is 60 pA, assign
the open state at ~55 pA. A horizontal line labeled 0 marks the
assignment. Assign the current level for the closed (1) state
relative to the open level. For example, if the open (0) state is
55 pA, choose a desired closed state below this level, for
example, 10 pA below or at 45 pA. A horizontal line labeled
1 marks the assignment. Allow the event detection to start by
selecting Nonstop retracing. Once the event detection is com-
plete, determine by eye whether the retracing satisfactorily
matches the actual current trace. Adjust the current levels
0 and 1 to improve the event retracing if needed. In the Results
window, the open (0) and closed (1) states are listed with the
information about the amplitude of the state, when the state
starts and ends in the trace and therefore the dwell time of the
state. From this list of all open and closed states of the chosen
analysis region, the gating frequency, the closed event duration,
and the open duration can be calculated.
3. To determine the gating frequency, calculate the total number
of closed events and divide it by the total time of the trace
analyzed.
4. To determine the event duration or the inter-event duration
take only the closed (1) or open (0) dwell times from the
Results window, respectively. Using the Extract Data subset
option in the Analyze drop down, copy only the desired values
from the total results into a new sheet. Create a histogram of
either the closed (1) or open (0) dwell times. Fit the histogram
to a standard exponential. Obtain the tau value from the fitting
results which is the average dwell time of either the closed
(1) or open (0) state population.
5. To determine the open probability, take the sum of all the
individual open dwell times and divide it by the total time of
the analyzed trace region, which includes both open and closed
states of the OmpG pore.
4 Notes
1. If OmpG expression in inclusion body is low, ensure that the

signal peptide is removed from the open reading frame of the
gene so that it expresses in the inclusion body and not secreted
to the outer membrane. The signal peptide of OmpG consists
of amino acids 1–21 MKKLLPCTALVMCAGMACAQA;
therefore maintain the start methionine and delete amino
acids 2–21.
2. The mutagenic primers contain at least 15 bp sequence on its 30

end that anneals the template. The 50 end contains the muta-
tion (either deletion, insertion or substitution). Also, at least
15 bp of sequence at the 50 end of each primer anneals to the 50
end of its primer pair. See Table 1 for examples.
3. An alternative PCR method uses primer SC47: 50 -CAGAA
GTGGTCCTGCAACTTTATC (reverse) with mutagenic for-
ward and SC46: 50 -ATAAAGTTGCAGGACCACTTCTG
(forward) with mutagenic reverse to generate two complemen-
tary fragments which when ligated generates the entire plas-
mid. Primers SC46 and SC47 anneal to the AmpR cassette in
the plasmid pT7-OmpG. Mix the two PCR products in a 1:1
molar ratio and subject to DpnI digestion for 3 h to degrade
the parental plasmid. Optional steps are PCR or gel purification
of the PCR products and in vitro ligation of PCR products with
T4 DNA ligase.
4. PCR purify product before DpnI digestion if PCR buffer is not
compatible with enzyme activity.
5. It is recommended to purify OmpG protein from a fresh trans-
formation plate up to 2 weeks old if stored at 4 C, not from
BL21 (DE3) glycerol stocks.
6. To ensure a high yield of OmpG expression, shake culture at
high speed (200–300 rpm) and use only up to 30% the volume
of shaker flask, to keep cells well ventilated and growing
aerobically.
7. OmpG yield lowers when induced at low temperatures, e.g.,
16 C. An OD600 reading up to 1 is acceptable before IPTG
induction in LB at 37 C.
8. Resuspend pellet in 10 mL buffer for every 250 mL culture.
9. When thawing a cell pellet, it is also acceptable to quickly warm
at 37 C until thawed.
10. If a sonicator/microfluidizer is not available after lysis with
lysozyme, add a DNAse I treatment step to break down the
genomic DNA and reduce the lysate’s viscosity. Add 50 kU
DNase I and 3 mM MgCl2 to the mixture and incubate for
30 min at room temperature.
11. OmpG can be purified in denatured form in 6.5–8 M urea.
When preparing high urea concentrations, use caution not to
heat the urea solution above 30 C where the urea starts to
degrade.
12. For efficient solubilization, resuspend inclusion body in bind-
ing buffer and place in beaker with stir bar and stir at constant
speed for at least 30 mins. If the room temperature is higher
than 30 C, solubilize the inclusion body on ice with stirring.
13. After purification, avoid freezing protein at 20 or 80 C

before labeling. The protein refolding efficiency is unchanged
but the labeling efficiency of the cysteine residue on OmpG
may be reduced. Therefore, it is best to purify OmpG-Cys and
then label within 2–3 h of purifying.
14. The Bradford assay is suitable because it is not affected by
buffers containing urea and/or reducing agents.
15. HEPES buffer is used because the –NH2 group on the Tris–
HCl molecule may interfere with maleimide cysteine chemis-
try, competing with the cysteine. Also, the optimum pH range
for maleimide cysteine conjugation is 6.5–7.5.
16. Another option is to incubate OmpG-Cys with maleimide
ligand for 12–18 h at 4 C with gentle agitation. If the ligand
is in powder form, first dissolve in an appropriate solvent such
as DMSO at an appropriate concentration according to manu-
facturer’s instructions.
17. Refolding is most efficient when the protein concentration is
between 0.35 mg/mL (10 μM) and 3.5 mg/mL (100 μM).
Also, a quick refolding procedure is to dilute the denatured
OmpG in a 2:3 v/v ratio, i.e., 2 volumes of denatured OmpG
with 3 volumes of refolding buffer where the final octyl gluco-
side concentration is ~1.95%.
18. Supplement glycerol into the refolded OmpG protein sample
to a final concentration of 20% v/v and mix well for homoge-
neity. Store in small aliquots at 80 C. The glycerol helps to
extend the shelf life of the refolded OmpG protein for
3–6 months as opposed to 1–2 months without glycerol. Aim
for a protein concentration of 0.3–0.8 mg/mL upon storage.
19. The streptavidin/biotin complex is resistant to SDS denatur-
ation. It is recommended to lower the urea concentration in
your denatured OmpG sample to below 6 M prior to mixing
with streptavidin to ensure that the streptavidin-OmpG com-
plex forms and remains intact during electrophoresis. Alterna-
tively, you can test labeling efficiency with your refolded
OmpG sample.
20. This step is to reduce any unlabeled OmpG dimer molecules.
21. The cysteine residues of unlabeled peptides would be modified
with iodoacetamide; thus, the mass is increased by ~57 Da due
to the carbamidomethylation.
22. The incubation period is to allow the pentane to evaporate.
The incubation time will depend on the ambient room tem-
perature where at higher temperature incubate for
shorter time.
23. Using a salt concentration lower than 100 mM is not advisable

as the current produced from OmpG pores will be below 20 pA
at 50 mV and thus decreases the signal-to-noise ratio. Typical
buffers contain either NaCl or KCl ions in a range from 0.15 to
3.0 M.
24. The appropriate working OmpG concentration to allow single
pore insertion should be determined for each batch of protein
purified. Typically diluting OmpG samples to 0.01 mg/mL
and using about 0.5 μL in a 900 μL chamber yields single
pore insertions.
25. If too many pores continue to insert into the bilayer, lower the
OmpG concentration in the chamber by removing some of the
buffer and replacing with fresh buffer. Add more lipids if
bilayer becomes unstable. Repeat the buffer renewal if too
many OmpG pores continue to insert into the bilayer. Another
option is to discard buffer in both chambers, wash with water
and 70–85% ethanol alternately, and prepare a new bilayer.
Dilute the OmpG sample even further before adding to the
new bilayer.
26. Ensure that your pipet tip is as far away from the aperture as
possible so as not to disturb the bilayer or introduce more
OmpG insertions.
27. If there is no signal indicating the detection of protein analyte,
the inserted OmpG may be unlabeled, break and remake
bilayer to induce insertion of another OmpG pore. Other
options, use the appropriate loop(s) to sense proteins, refer to
Table 3; use ligands with short linkers such as PEG2 to PEG4
or no linker if possible.
References
1. Kasianowicz JJ, Bezrukov SM (1995) Proton- chain within a single protein pore. Nat Bio-
ation dynamics of the alpha-toxin ion channel technol 18:1091–1095
from spectral analysis of pH-dependent current 6. Halverson KM, Panchal RG, Nguyen TL,
fluctuations. Biophys J 69:94–105 Gussio R, Little SF, Misakian M, Bavari S,
2. Braha O, Gu LQ, Zhou L, Lu X, Cheley S, Kasianowicz JJ (2005) Anthrax biosensor, pro-
Bayley H (2000) Simultaneous stochastic sens- tective antigen ion channel asymmetric block-
ing of divalent metal ions. Nat Biotechnol ade. J Biol Chem 280:34056–34062
18:1005–1007 7. Hornblower B, Coombs A, Whitaker RD,
3. Bezrukov SM, Vodyanoy I, Parsegian VA Kolomeisky A, Picone SJ, Meller A, Akeson M
(1994) Counting polymers moving through a (2007) Single-molecule analysis of
single ion channel. Nature 370:279–281 DNA-protein complexes using nanopores.
4. Kasianowicz JJ, Henrickson SE, Weetall HH, Nat Methods 4:315–317
Robertson B (2001) Simultaneous multiana- 8. Talaga DS, Li J (2009) Single-molecule protein
lyte detection with a nanometer-scale pore. unfolding in solid state nanopores. J Am Chem
Anal Chem 73:2268–2272 Soc 131:9287–9297
5. Movileanu L, Howorka S, Braha O, Bayley H 9. Braha O, Walker B, Cheley S, Kasianowicz JJ,
(2000) Detecting protein analytes that modu- Song L, Gouaux JE, Bayley H (1997)
late transmembrane movement of a polymer
Designed protein pores as components for bio- 24. Kasianowicz JJ, Balijepalli AK, Ettedgui J, For-
sensors. Chem Biol 4:497–505 stater JH, Wang H, Zhang H, Robertson JWF
10. Wendell D, Jing P, Geng J, Subramaniam V, (2016) Analytical applications for pore-
Lee TJ, Montemagno C, Guo P (2009) Trans- forming proteins. Biochim Biophys Acta Bio-
location of double stranded DNA through membr 1858:593–606
membrane adapted phi29 motor protein nano- 25. Movileanu L (2009) Interrogating single pro-
pore. Nat Nanotechnol 4:765–772 teins through nanopores: challenges and
11. Bell NAW, Keyser UF (2014) Nanopores opportunities. Trends Biotechnol 27:333–341
formed by DNA origami: a review. FEBS Lett 26. Majd S, Yusko EC, Billeh YN, Macrae MX,
588:3564–3570 Yang J, Mayer M (2010) Applications of
12. Dekker C (2007) Solid-state nanopores. Nat biological pores in nanomedicine, sensing,
Nanotechnol 2:209–215 and nanoelectronics. Curr Opin Biotechnol
13. Li J, Stein D, McMullan C, Branton D, Aziz 21:439–476
MJ, Golovchenko JA (2001) Ion-beam sculpt- 27. Oukhaled A, Bacri L, Pastoriza-Gallego M,
ing at nanometre length scales. Nature Betton JM, Pelta J (2012) Sensing proteins
412:166–169 through nanopores: fundamental to applica-
14. Bell NAW, Engst CR, Ablay M, Divitini G, tions. ACS Chem Biol 7:1935–1949
Ducati C, Liedl T, Keyser UF (2012) DNA 28. Shi W, Friedman AK, Baker LA (2017) Nano-
origami nanopores. Nano Lett 12:512–517 pore sensing. Anal Chem 89:157–188
15. Messager L, Burns JR, Kim J, Cecchin D, 29. Willems K, Van Meervelt V, Wloka C, Maglia G
Hindley J, Pyne ALB, Gaitzsch J, Battaglia G, (2017) Single-molecule nanopore enzymol-
Howorka S (2016) Biomimetic hybrid nano- ogy. Philos Trans R Soc B 372:20160230
containers with selective permeability. Angew 30. Nivala J, Mulroney L, Li G, Schreiber J, Akeson
Chem Int Ed 55:11106–11109 M (2014) Discrimination among protein var-
16. Butler TZ, Pavlenok M, Derrington IM, iants using an unfoldase-coupled nanopore.
Niederweis M, Gundlach JH (2008) Single- ACS Nano 8:12365–12375
molecule DNA detection with an engineered 31. Nir I, Huttner D, Meller A (2015) Direct sens-
MspA protein nanopore. Proc Natl Acad Sci ing and discrimination among ubiquitin and
105:20647–20652 ubiquitin chains using solid-state nanopores.
17. Stefureac R, Long Y-T, Kraatz H-B, Howard P, Biophys J 108:2340–2349
Lee JS (2006) Transport of alpha-helical pep- 32. Rosen CB, Rodriguez-Larrea D, Bayley H
tides through alpha-hemolysin and aerolysin (2014) Single-molecule site-specific detection
pores. Biochemistry 45:9172–9179 of protein phosphorylation with a nanopore.
18. Cornell BA, Braach-Maksvytis VL, King LG, Nat Biotechnol 32:179–181
Osman PD, Raguse B, Wieczorek L, Pace RJ 33. Fahie MA, Chen M (2015) Electrostatic inter-
(1997) A biosensor that uses ion-channel actions between OmpG nanopore and analyte
switches. Nature 387:580–583 protein surface can distinguish between glyco-
19. Guo BY, Zeng T, Wu HC (2015) Recent sylated isoforms. J Phys Chem B
advances of DNA sequencing via nanopore- 119:10198–10206
based technologies. Sci Bull 60:287–295 34. Rotem D, Jayasinghe L, Salichou M, Bayley H
20. Han A, Schürmann G, Mondin G, Bitterli RA, (2012) Protein detection by nanopores
Hegelbach NG, De Rooij NF, Staufer U equipped with aptamers. J Am Chem Soc
(2006) Sensing protein molecules using nano- 134:2781–2787
fabricated pores. Appl Phys Lett 88:1–4 35. Harrington L, Alexander LT, Knapp S, Bayley
21. Firnkes M, Pedone D, Knezevic J, H (2015) Pim kinase inhibitors evaluated with
Döblinger M, Rant U, Do M, Rant U (2010) a single-molecule engineered nanopore sensor.
Electrically facilitated translocations of proteins Angew Chem Int Ed 54:8154–8159
through silicon nitride nanopores: conjoint 36. Fahie M, Chisholm C, Chen M (2015)
and competitive action of diffusion. Nano Resolved single-molecule detection of individ-
Lett 10:2162–2167 ual species within a mixture of anti-biotin anti-
22. Chen P, Gu J, Brandin E, Kim YR, Wang Q, bodies using an engineered monomeric
Branton D (2004) Probing single DNA mole- nanopore. ACS Nano 9:1089–1098
cule transport using fabricated nanopores. 37. Pham B, Eron SJ, Hill ME, Li X, Fahie MA,
Nano Lett 4:2293–2298 Hardy JA, Chen M (2019) A nanopore
23. Burns JR, Stulz E, Howorka S (2013) Self- approach for analysis of caspase-7 activity in
assembled DNA nanopores that span lipid cell lysates. Biophys J 117:844–855
bilayers. Nano Lett 13:2351–2356
38. Fahie MA, Yang B, Mullis M, Holden MA, OmpG nanopore sensor by adjusting ligand
Chen M (2015) Selective detection of protein tether length. ACS Sensors 1:614–622
homologues in serum using an OmpG nano- 40. Chen M, Li Q-H, Bayley H (2008) Orienta-
pore. Anal Chem 87:11143–11149 tion of the monomeric porin OmpG in planar
39. Fahie MA, Yang B, Pham B, Chen M (2016) lipid bilayers. Chembiochem 9:3029–3036
Tuning the selectivity and sensitivity of an
Chapter 8
Nanopore Enzymology to Study Protein Kinases

and Their Inhibition by Small Molecules
Leon Harrington, Leila T. Alexander, Stefan Knapp, and Hagan Bayley
Abstract
Nanopore enzymology is a powerful single-molecule technique for the label-free study of enzymes using
engineered protein nanopore sensors. The technique has been applied to protein kinases, where it has
enabled the full repertoire of kinase function to be observed, including: kinetics of substrate binding and
dissociation, product binding and dissociation, nucleotide binding, and reversible phosphorylation. Fur-
ther, minor modifications enable the screening of type I kinase inhibitors and the determination
of inhibition constants in a facile and label-free manner. Here, we describe the design and production of
suitably engineered protein nanopores and their use for the determination of key mechanistic parameters of
kinases. We also provide procedures for the determination of inhibition constants of protein kinase
inhibitors.
Key words Kinase, Kinase inhibitors, Inhibitor screening, Nanopore enzymology, Enzyme kinetics,
Enzyme mechanism, Single molecule
1 Introduction
Engineered protein nanopores have been employed to study a

broad range of analytes including metal ions [1, 2], small organic
molecules [3, 4], polymers such as PEG [5–7], and biopolymers
such as polynucleotides [8–16], peptides [17, 18], and proteins
[19–22]. The staphylococcal pore-forming toxin alpha-hemolysin
has been most extensively used for nanopore sensors due to its
robustness and tolerance to engineering; however there are now a
range of protein pores offering different characteristics such as
larger pore lumens or more favorable pore architectures for analyte
detection. Some examples include ClyA [23], FhuA [24], and
CsgG [25]. Detection can be achieved through several different
strategies, dependent on the nature of the analyte and the particular
protein pore being engineered. In all cases, analyte detection is
achieved by measuring the modulation of ionic current flowing
through single pores reconstituted into artificial lipid bilayers. For
95
96 Leon Harrington et al.
small analytes, an affinity site is typically engineered within the pore

lumen. Polymeric and larger analytes can be detected by their
electrophoretic translocation through the pore. For analytes too
big to be accommodated within the pore lumen, a common strat-
egy is to tether a ligand to a flexible linker (such as PEG or a peptide
linker sequence) attached to the pore mouth [19]. Additionally, it is
possible to trap protein analytes within the mouth of large pores
such as ClyA by the manipulation of electroosmotic flow [26].
Nanopore enzymology is the application of nanopore sensors
to detect enzymes and measure their characteristic properties such
as substrate kinetics and affinities, interactions with cofactors, turn-
over, as well as conformational changes due to ligand binding
[27]. As a single-molecule technique, it offers the ability to measure
properties not normally accessible by conventional methods, such
as the detection and measurement of intermediate steps that are not
rate determining, and the measurement of heterogeneity and dis-
order in reactions. Further, the method does not require the use of
fluorescent or radioactive labels that may be prohibitive to intro-
duce, may perturb the properties of the target enzyme, or cause
artifacts or false positives in screening scenarios. Finally, unlike
many conventional reaction assays there is no reliance on coupled
enzymatic reactions that may themselves be susceptible to inhibi-
tion or other perturbations [28].
The first generation of nanopore protein kinase sensors were
designed to detect cAMP-dependent protein kinase (PKA). As
kinases are too large for the lumen of the alpha-hemolysin pore,
detection was performed proximal to the pore mouth using a
peptide derived from the heat-stable protein kinase inhibitor
(PKIP5–24) attached either through a chemically coupled PEG
linker [29] or through fusion within one of the beta-sheet loops
at the mouth of the αHL pore [30]. Genetic encoding of the sensor
peptide within the αHL pore avoided the difficulties of chemical
attachment, such as incomplete reaction yields; however the sensor
peptide remained conformationally constrained within the beta
loop. We further enhanced this engineering strategy by encoding
a site-selective protease (TEV protease) recognition site, so that the
sensor-containing loop could be cleaved posttranslationally to lib-
erate one end of the sensor peptide, removing these conformational
constraints (see Fig. 1) [31].
We have used these proteolytically cleaved sensors to exten-
sively study protein kinases. As well as PKA, we have used such
sensors to study the Pim family of protein kinases, a small family of
three constitutively active serine/threonine protein kinases. We
have shown that these kinases exhibit very fast association rates
for consensus substrate sequences due to electrostatic enhancement
[31]. It is also possible to discriminate different family members
through the noise spectra of their respective ionic current signals.
When using a substrate sequence as the sensor peptide, we have
Protein Kinases Studied by Nanopore Enzymology 97
αHL
cis
Membrane
Modified
Subunit
trans
S/G linker
Sensor peptide
(e.g. pimtide)
TEV RS
b TEV
αHL S/G linker Pimtide RS S/G linker αHL
G G S S G S G S S K I G G L135
S/G linker αHL

125T GNDSSGSGSSGARKRRRHPSGPPTAENLYFQ
αHL S/G linker Pimtide TEV RS
Fig. 1 Design of TEV protease-cleavable αHL trans loop fusion pores for kinase detection. (a) Illustration of the
protease-cleaved sensor pore indicating the positions of the sensor peptide element (purple), flexible Ser/Gly
linkers (green), and the TEV protease recognition site (“TEV RS,” red). (b) Scheme and sequence of the sensor-
containing pore loop. The Pim kinase consensus substrate “pimtide” is used as an example; however other
sequences can be inserted as desired (see main text). (Adapted from ref. 31)
shown it is possible to observe phosphorylation, dephosphoryla-

tion, and thiophosphorylation [32]. We have also shown that a
pseudosubstrate sequence can be used, which allows the observa-
tion of nucleotide interactions without enzymatic conversion
[33]. An important example is the synergistic binding of ATP,
whereby a significant enhancement in the affinity of the peptide
pseudosubstrate is observed in the presence of ATP. A particular
advantage of our nanopore sensors equipped with pseudosubstrate
peptides is that they allow the observation of individual binding
events that represent formation of both binary and ternary com-

plexes. A range of rate constants and affinities can therefore be
determined for nucleotide interactions [33].
Pseudosubstrate sensors can be applied to determine the inhi-
bition constants of type I (ATP-competitive) protein kinase inhibi-
tors. Such inhibitors bind within the ATP-binding pocket of
protein kinases. However, unlike ATP, they do not synergistically
bind with the pseudosubstrate peptide, and therefore appear as
“binary” complex events. Therefore, it is possible to derive the
inhibition constant (Ki) of the inhibitor by titrating inhibitor
against a fixed concentration of ATP and measuring the modulation
of the populations of binding events corresponding to binary and
ternary complex formation [33].
Here we describe the design and construction of proteolytically
cleaved nanopore sensors. We then describe the procedures to
reconstitute these into lipid bilayers and measure kinase–substrate
interactions. We give here only a light description of the electro-
physiology techniques, as in principle the pores can be reconsti-
tuted in any number of different artificial membranes, and the
techniques are already extensively described elsewhere [34]. Fur-
ther, we describe how experiments can be performed to study
nucleotide interactions. Finally, we describe the application of
nanopore sensors for the evaluation and screening of protein kinase
inhibitors.
2 Materials
Prepare all aqueous solutions with ultrapure water with a resistivity

of 18 MΩ cm at 25 C and reagents of analytical purity.
2.1 Nanopore 1. S30 T7 High-Yield Protein Expression System (Promega).

Preparation This is an in vitro transcription-translation system optimized
for constructs under T7 promoters.
2. Plasmid DNA with pore-encoding genes under control of a T7
promoter. One plasmid should express αHL-D127N and the
other αHL-D127-PLM-TEV (or desired variant using an alter-
native sensor peptide sequence). The plasmid DNA must be
purified by Maxiprep to achieve sufficient purity and quantity
for the in vitro transcription and translation (IVTT) reactions.
35
3. S-Met (1175 Ci/mmol, 10 mCi/mL, PerkinElmer). Local
rules and regulations and appropriate safety measures must be
followed when handling and disposing of radioisotopes.
4. Chloramphenicol (1000): Prepare a 34 mg/mL stock solu-
tion in absolute ethanol. The stock can be stored at 20 C
until required.
5. TEV protease. We recommend using AcTEV protease

(Thermo Fisher Scientific) as it is stable for prolonged incuba-
tion and has activity over a broad temperature (4–30 C) and
pH (6.0–8.5).
6. Laemmli sample buffer (1): 62.5 mM Tris–HCl, pH 6.8,
2.3% (w/v) SDS, 5% (v/v) β-mercaptoethanol, 10% (w/v)
glycerol. A 4 concentrate is usually prepared, from which a
1 solution can be prepared by dilution with water as required
in the protocol.
7. TE buffer: 10 mM Tris–HCl, pH 7.5, 1 mM EDTA, filter
sterilized (0.2 μm membrane).
8. Autoradiography reagents and equipment: Whatman 3M filter
paper, gel drier with vacuum, autoradiography film, film cas-
sette, film developer.
9. Nanopore gel extraction tools: scalpel, pair of forceps, pestle
(e.g., Sigma Pellet Pestle, Merck), 0.2 μm centrifugal filter
(e.g., Ultrafree-MC centrifugal filter units, Merck Millipore).
2.2 Rabbit 1. Rabbit whole blood.

Erythrocyte 2. Hypotonic lysis buffer: 10 mM MOPS, pH 7.4.
Membranes
3. MBSA buffer: 10 mM MOPS, 150 mM NaCl, 0.1% (w/v)
bovine serum albumin, titrated to pH 7.4 with NaOH. After
preparation the buffer should be sterilized by passing through a
0.2 μm membrane filter (e.g., Millipore Stericup vacuum filter
units or similar) and stored at 4 C.
2.3 Single-Channel 1. Single-channel recording buffer: 15 mM MOPS, pH 6.8,

Recording 300 mM KCl, 5 mM DTT (see Note 1).
2. Single-channel special equipment: Ag/AgCl electrodes (with
agar bridges), Delrin compartments with 1 mL compartment
volume, an Axon Axopatch 200B patch-clamp amplifier and
DigiData digitizer.
3. Pim kinase (or desired protein kinase). These enzymes may be
commercially available or can be purified recombinantly such as
Pim-1, -2, and -3, which are described in previous work [31].
4. ATP stock solution (100 mM) buffered to pH 7.0 with Tris
base: Dissolve 551 mg of ATP disodium salt hydrate in 5 mL of
water. Dissolve 1.21 g of Tris base in 10 mL of water to give a
stock solution of 1 M. Add the 1 M Tris base stock solution to
the ATP solution until pH 7.0 is achieved. Add sufficient water
to achieve a final volume of 10 mL. As the stoichiometry of the
water of crystallization for the ATP is not usually known, the
exact final concentration of ATP must be determined spectro-
photometrically at 259 nm using an extinction coefficient of
15,400 M1 cm1. The ATP stock solution should then be
aliquoted and frozen at 80 C. Avoid freeze/thaw and

replace stocks regularly (e.g., every 6–12 months, see Note
2). Other nucleotide solutions may be prepared in a similar
manner.
5. 1 M MgCl2 solution for ATP coordination.
6. Inhibitor stock solutions: Inhibitors should be dissolved in
DMSO where possible. Stock solutions should be stored in
the dark at 20 C. We recommend verification of the identity
and purity of inhibitors obtained commercially or from a third-
party, for example, by LC-MS and/or NMR. Pim kinase inhi-
bitors can be as described in previous work [33].
7. Useful Analysis Software: Clampfit, QuB, OriginPro.
3 Methods
3.1 Design of TEV The following methods are written for the study of Pim kinases (see
Protease-Cleaved Note 3); however the methods should be generally applicable for
Sensor Pores any protein kinase provided a suitable substrate and/or pseudosub-
strate sequence is known.
1. Sensor pores comprise heteroheptamers of αHL-D127N (see
Note 4), where a single subunit is modified to contain the
sensor peptide sequence flanked by flexible Ser/Gly linker
sequences and a tobacco etch virus (TEV) protease recognition
sequence (see Fig. 1). The sensor peptide-modified subunit is
designed to be expressed with an intact sensor-containing loop
that is then posttranslationally cleaved by TEV protease to
liberate the C-terminal end of the sensor peptide.
2. Insert the sequence encoding the desired sensor peptide into
the sensor subunit-encoding gene using standard molecular
cloning methods.
3. Sequence the resulting plasmid to verify successful cloning.
4. Obtain high purity preparations of both the αHL-D127N and
the sensor peptide-modified subunit αHL-D127N-PLM-TEV
(or desired variant using an alternative sensor peptide
sequence) encoding plasmids by using a commercial Maxiprep
system. Check purity and concentration of the plasmid solu-
tions using UV-vis spectroscopy and dilute to 400 ng/μL with
ultrapure water. Store at 20 C until required.
3.2 Preparation To prepare rabbit erythrocyte membranes, buffers should be

of Rabbit Erythrocyte ice-cold throughout.
Membranes
1. Place 10 mL of rabbit blood into a 50 mL centrifuge tube and
add 40 mL MBSA buffer. Centrifuge this mixture at 700 g
for 10 min in a swing bucket rotor. A thin layer of white blood
cells may be visible on top of the pellet of erythrocytes. Remove

the white cells with the supernatant.
2. Gently resuspend the erythrocyte pellet in 50 mL MBSA by
inversion of the tube. Centrifuge again at 600 g for 20 min
and repeat these washing steps five times. Transfer the washed
erythrocyte pellet to a fresh centrifuge tube.
3. Resuspend the pellet in hypotonic lysis buffer (80 mL). Centri-
fuge at 30,000 g for 30 min. The resulting pellet should
consist of three layers. The upper, soft layer contains the mem-
branes. Carefully remove the supernatant, then collect the
upper layer of the pellet without disturbing the lower layers.
4. Resuspend the membrane pellet in MBSA, centrifuge at
30,000 g for 30 min, then remove the supernatant. This
washing with MBSA should be repeated until the pellet
becomes almost colorless. The membranes can then be frozen
as 50 μL aliquots at 80 C.
3.3 Preparation 1. Prepare three in vitro transcription and translation (IVTT)

of TEV reactions. For reaction 1, mix 2.5 μL of 400 ng/μL plasmid
Protease-Cleaved DNA encoding αHL-D127N, 10 μL of S30 premix plus, 1 μL
Sensor Pores of 35S-Met, 9 μL of S30 extract, and 2.5 μL of rabbit erythro-
cyte membranes. This will produce αHL-D127N homohepta-
mers to be used as a marker for SDS-PAGE. For reaction 2, mix
2.5 μL of 400 ng/μL plasmid DNA encoding αHL-D127N-
PLM-TEV (or variant thereof), 10 μL of S30 premix plus, 1 μL
of 35S-Met, 9 μL of S30 extract, and 2.5 μL water. This will
produce αHL-D127N-PLM-TEV monomers. For reaction
3, mix 7.5 μL of 400 ng/μL plasmid DNA encoding αHL-
D127N, 30 μL of S30 premix plus, 3 μL of 35S-Met, 27 μL of
S30 extract, and 7.5 μL water. This will produce αHL-D127N
monomers.
2. All reactions should then be incubated at 37 C for 1 h, for
example, in a water bath or an incubator with constant
agitation.
3. After incubation, halt protein synthesis by adding chloram-
phenicol to reactions 2 and 3, giving a final concentration of
34 μg/mL. Reactions 1 and 3 should be stored on ice until
required.
4. Add 1 μL of TEV protease to reaction 2 and incubate at 4 C
for 4 h.
5. To assemble heteroheptamers, mix reaction 2 with reaction
3 and add 5 μL of rabbit erythrocyte membranes. Incubate
the mixture at room temperature (22–25 C) for 1 h.
6. Centrifuge reaction 1 and the assembled heteroheptamer mix-
ture each at 25,000 g for 10 min at 4 C. Discard the
supernatants and resuspend each membrane pellet in 200 μL

MBSA buffer. Centrifuge again as before, wash each pellet with
a further 200 μL MBSA buffer and centrifuge a final time.
7. Resuspend both pellets in 50 μL Laemmli sample buffer (1).
Both samples should then be electrophoresed in adjacent lanes
on a 5% SDS-PAGE gel (recommended dimensions:
165 220 1 mm). Electrophoresis should be performed
under a constant voltage of 80 V and run overnight (18 h).
8. The gel should then be carefully removed from the glass plates
and dried on Whatman 3M filter paper under vacuum at room
temperature for approximately 4 h in a gel dryer. The gel will be
dried to the Whatman paper. It should be completely dry to
avoid sticking to the autoradiography film; however overdrying
can result in cracking of the gel.
9. Image the dried gel by autoradiography. Under darkroom
conditions, lay an autoradiography film on top of the dried
gel and staple the two together using at least two staples in
different corners and at different angles to each other. These
staples will be used to realign the film and gel together after
development. Seal the cassette and expose for an appropriate
amount of time to achieve a satisfactory exposure (typically
24 h, depending on the remaining activity of the 35S-Met).
10. After exposure, and under dark room conditions, carefully
remove the staples to separate the film from the gel. Develop
the film. Once the film is developed, it can be handled in the
light. The gel should appear as in Fig. 2. The first lane contain-
ing reaction 1 will contain a single band indicating the position
of the homoheptamer. The second lane should contain several
bands, with the uppermost coinciding with the homoheptamer
marker band of the first lane. The next band below this corre-
sponds to the desired heteroheptamer containing a single mod-
ified subunit.
11. Reattach the developed film to the dried gel using the staple
holes for alignment and new staples. Ensure the correct side of
the gel is facing the film.
12. With the film facing toward you, use a scalpel to cut a rectangle
tightly around the bulk of the desired heteroheptamer band,
ensuring you cut through the film, the dried gel, and the filter
paper.
13. Using a pair of forceps, collect the excised rectangle of gel and
filter paper, and place in a 1.5 mL microcentrifuge tube. Do
not include the excised X-ray film. Rehydrate the excised gel
and filter paper with 300 μL of TE buffer for 10 min. Carefully
remove the filter paper from the rehydrated gel slice and dis-
card. Homogenize the gel slice in the buffer with a plastic
pestle. Incubate for a further 50 min at room temperature.
Uncleaved
Cleaved
Marker
(αHL-D127N)7
(αHL-D127N)6(αHL-D127N-PLM-TEV)1
Fig. 2 Example preparative SDS-PAGE for the extraction of heteromeric sensor

pores. The “Marker” lane contains only homoheptamers of αHL-D127N, used as
a reference point. The “Uncleaved” and “Cleaved” lanes contain mixtures of
(αHL-D127N)x(αHL-D127N-PLM-TEV)y where the αHL-D127N-PLM-TEV
subunits in the “Cleaved lane” are posttranslationally cleaved by TEV protease
prior to heptamerization. (Adapted from ref. 31)
14. Remove the supernatant from the gel slice fragments by filtra-
tion through a 0.2 μm centrifugal filter.
15. Prepare 5–10 μL aliquots of the filtrate in microcentrifuge
tubes, flash freeze in liquid nitrogen, and store at 80 C.
3.4 Reconstitution 1. Set up the necessary electrophysiology equipment for the pre-
of Pores ferred lipid bilayer technique. We use the Montal-Mueller
and Single-Channel method for “folded” black lipid membranes [35], using
Recording custom-made Ag/AgCl electrodes (with agar bridges, see
Note 5) and Delrin compartments with 1 mL compartment
volume, with an Axon Axopatch 200B patch-clamp amplifier
and DigiData digitizer, as described in Maglia et al. [34]. While
we have not tested our kinase-sensing nanopores in other
membrane systems, they should be compatible with all com-
mon techniques, provided the addition of analytes and cofac-
tors to the trans side of the membrane during measurement is
feasible.
2. Establish a lipid bilayer and verify its stability and size, for
example, by applying a triangular potential waveform (1 mV/
ms) for a few minutes. This produces a square wave current
signal due to the membrane capacitance, where the current is
proportional to the membrane size.
3. Thaw an aliquot of purified nanopore sensor protein and add
an initial volume of 0.2 μL to the cis compartment. The precise
volume of nanopore required needs to be found empirically, as
this will depend on the particular nanopore mutant, the type of

lipid bilayer being used, and the protein extraction yield during
purification, among other variables. For our Pim kinase sensors
in Montal-Mueller bilayers with 1 mL compartments we typi-
cally add 0.1–0.5 μL, which corresponds to a final concentra-
tion of approximately 1 ng/μL.
4. To promote pore insertion, we apply +150 mV while stirring
the cis compartment. Pore insertion will be marked by a jump
in conductance of the membrane, whereupon stirring should
immediately cease and the membrane potential should be
decreased to 50 mV (see Note 6). If multiple insertions
frequently occur, the amount of nanopore protein added to
the cis compartment should be decreased; the goal is to main-
tain a single pore within the lipid bilayer for the required
duration of the measurement.
5. The inserted pore should be observed for several minutes to
ensure the conductance and current signal (e.g., blockade
kinetics of the sensor peptide) are as expected (see Note 7). If
the pore has not previously been characterized, we recommend
performing an I-V curve measurement from 100 to
+100 mV. The system is then ready for the desired experiment
to be performed. An example current trace of the αHL-
D127N-PLM-TEV sensor is shown in Fig. 3a.
3.5 Substrate 1. With a single nanopore reconstituted in the lipid bilayer, pro-
Binding Kinetics tein kinase may then be added to the trans compartment to
measure binding to the sensor peptide.
2. Ideally, the protein kinase stock solution should be around
1000-fold concentrated versus the final desired concentration,
for example around 100 μM, so that concentrations of
100, 200, 400 nM can be achieved by adding 1, 2, or 4 μL of
100 μM stock to a 1 mL trans compartment (see Note 8).
3. Apply a membrane potential of 50 mV and observe and
record binding of the kinase to the sensor pore. An example
of the current signal resulting from Pim-1 binding to the αHL-
D127N-PLM-TEV sensor is shown in Fig. 3b.
4. Perform a titration of protein kinase, for example 100, 200,
and 400 nM. This helps confirm that the observed signal is due
to the kinase–substrate interaction, as the measured pseudo-
first-order association rate constant will show a linear concen-
tration dependence, whereas the first-order dissociation rate
constant should be independent of kinase concentration. The
bimolecular association rate constant is easily calculated as the
gradient of the plot of the pseudo-first-order association rate
constant vs. kinase concentration. A set of titrations can be
performed sequentially on a single sensor pore, with replicates
performed several times with fresh membranes and pores.
a αHL-D127N-PLM-TEV (cleaved), –50 mV

0 pA
b αHL-D127N-PLM-TEV (cleaved), 91 nM PIM-1 trans

0 pA
15 pA O *
1s
k–1 k'+2 = k+2⋅ [Kinase]
k+1 k–2
Kinase
B1 O1 O2
Fig. 3 Example current traces of the cleaved (αHL-D127N)6(αHL-D127N-PLM-TEV)1 pore (“αHL-D127N-PLM-

TEV” for brevity) under an applied potential of 50 mV (a) before and (b) after the addition of 91 nM PIM-1 to
the trans compartment. The corresponding dwell time histograms for the open pore current level (indicated by
a red “O”) are also shown. (c) Kinetic model. In the absence of kinase, the pore exhibits a two state current
signal, alternating between the open pore current level (state O1) and a blocked state (B1) that occurs due to
transient interaction of the sensor peptide with the pore mouth. In the presence of kinase, an additional state
(O2) in the open pore current level is observed with a longer mean dwell time (as indicated with an asterisk in
the histogram of (b)) that represents the kinase-bound state of the sensor pore. Kinase can only bind the
sensor peptide when it is freely available in solution (i.e., from state O1). Kinase binding to the sensor peptide
holds the pore in the open state until dissociation occurs. As binding occurs some distance from the pore
mouth, no current blockade due to kinase binding is observed. (Adapted from ref. 31)
5. Extract sections of current recording corresponding to specific

kinase concentrations into individual files, for example, in
Clampfit. Use Clampfit to export the extracted files to QuB
LDT format and open in QuB (https://qub.mandelics.com).
Set a dead time appropriate for filtering conditions. We use
0.3 ms for a filter corner frequency of 2 kHz.
6. Idealize the current recording using the SKM method in QuB.
We do not apply the dead time when performing idealization.
7. Fit the kinetic model (Fig. 3c) using the MIL method in QuB,
with the dead time applied. Use the “smooth binning” option
and a square-root y-ordinate for the dwell time histograms.
8. Transfer the fitted rate constants (k0þ2 and k2) to a spreadsheet
or data analysis software (e.g., OriginPro). Plot both rate con-
stants against kinase concentration (in M). The dissociation
rate constant k2 (in s1) should be independent of kinase
concentration. The measured association rate constant is a
pseudo-first-order rate constant (k0þ2 ¼ kþ2 ñ½Kinase ) and

when plotted against concentration should be linearly depen-
dent on concentration. Obtain the bimolecular association rate
constant k+2 (in M1 s1) by linear regression. The dissociation
equilibrium constant can then be calculated as K d ¼ kk2
þ2
.
3.6 Nucleotide 1. When studying the effect of nucleotides such as ATP or its
Effects and Synergistic analog ATP-γ-S, it is important to use a pseudosubstrate sensor
Binding peptide. This allows the repeat observation of binary and ter-
nary complex binding events. If a substrate sensor peptide is
used, phospho-transfer will occur rapidly and irreversibly. We
have recently shown that a protein phosphatase can be used to
dephosphorylate the phospho-pore and potentially establish a
catalytic cycle; however challenges remain to extract quantita-
tive data from such experiments [32]. The following steps are
written for ATP; however the same approach can be used for
other analogs where synergistic binding occurs, such as
ATP-γ-S (N.B. this effect is weaker than for ATP, and metal-
ion dependent) [32].
2. Reconstitute a single pseudosubstrate sensor pore as in Sub-
heading 3.4.
3. Add kinase to the trans compartment with mixing (see Note 8)
and observe binding for several minutes (see Fig. 4a).
4. Add 1 M MgCl2 solution to a final concentration of 10 mM in
the trans compartment, together with the desired concentra-
tion of ATP, with mixing. Apply a potential of 50 mV and
record the current signal for 10–20 min. Synergistic binding
between nucleotide, kinase, and pseudosubstrate will lead to
the ternary complex having a stronger affinity than the binary
complex (absence of nucleotide), resulting in two separate
populations of kinase binding events (see Fig. 4b).
KinaseñMgATP
5. To determine K d , a titration of ATP should be
performed. We recommend a half-log series over at least four
orders of magnitude (e.g., 1–1000 μM). As many kinases pos-
sess significant ATPase activity, it is not recommended to per-
form a sequence of titrations in a single experiment.
6. Once data are collected, extract appropriate sections for analy-
sis in QuB, as described in Subheading 3.5.
7. Using the model shown in Fig. 4c (see Note 9), idealize current
traces using the SKM method of QuB. Then fit the model
using the MIL method. Transfer the fitted rate constants to a
spreadsheet or data analysis software for further analysis.
8. Plot both the pseudo-first-order association rate constants for
binary (k0þ2 ) and ternary (k0þ3 ) complexes against log ATP
concentration. The resulting plots should be sigmoid, with
the pseudo-first-order association rate constant for the binary
a αHL-D127N-PSLM-TEV (cleaved), 162 nM PIM-1 trans

0 pA
b αHL-D127N-PSLM-TEV (cleaved), 162 nM PIM-1 + 10 mM MgCl2 + 10 μM ATP trans
0 pA
* +
15 pA
1s
c Generalised kinetic model
]
se
ina
⋅ [K
+2
k
=
–2
k
+2 Kinase
O2
k'
k–1
k+1
k' +3
=
k +3
B1 O1
⋅ [K
•N
k –3
uc
]
K•Nuc
O3
Fig. 4 Synergistic binding of ATP. Example current traces and open pore current level dwell time histograms
for the cleaved heteroheptameric sensor pore αHL-D127N-PSLM-TEV in the presence of 162 nM PIM-1 (a)
before and (b) after the addition of 10 mM MgCl2 and 10 μM ATP (final concentrations) in the trans
compartment. The introduction of MgATP leads to an additional population of events in the open pore current
level with a mean dwell time in the order of seconds (orange cross) that corresponds to formation of the
ternary complex. The population indicated with the green asterisk corresponds to formation of the binary
complex. (c) Generalized kinetic model for analysis of synergistic nucleotide binding. It is sometimes
necessary to include an additional state in the blocked current level, B2, for kinases that exhibit current
noise when bound (see Note 10). All open states share the same current level, as do the blocked states. We
assume independent binding of apo-kinase and MgATP-bound kinase to the sensor pore. ((a, b) Adapted with
permission from ref. 33. Copyright 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. (c) Adapted with
permission from ref. 32. Copyright 2018 American Chemical Society)
population decreasing with ATP concentration, and the

pseudo-first-order association rate constant for the ternary
population increasing with ATP concentration (see Fig. 5).
9. Use nonlinear regression to fit the following equations:
Binary population:
a b
Fig. 5 Example titration curves for synergistic binding of PIM-1 with MgATP to αHL-D127N-PSLM-TEV. The
ATP concentration-dependence of the pseudo-first-order association rate constants for formation of the (a)
binary (k 0þ2 ) and (b) ternary (k 0þ3 ) complex are plotted together with their fitted curves according to the
equations in Subheading 3.6. The concentration of PIM-1 was 162 nM throughout. Error bars represent
S.D. (n ¼ 3). (Adapted with permission from ref. 33. Copyright 2015 WILEY-VCH Verlag GmbH & Co. KGaA,
Weinheim)
!
k0þ2 ½ATP0
¼ kþ2 ñ 1 KinaseñMgATP
½Pim0 K þ ½ATP
d 0
Ternary population:
!
k0þ3 ½ATP0
¼ kþ3 ñ
½Pim0 KinaseñMgATP
Kd þ ½ATP0
KinaseñMgATP
10. Fitting both curves will provide two values for K d ,
which should be in good agreement, as well as the bimolecu-
lar association rate constants for binary (k+2) and ternary (k+3)
complex formation.
3.7 Determination 1. This method of characterizing kinase inhibitors relies on mea-

of Inhibition Constants suring the modulation of the synergistic binding of ATP due to
of Kinase Inhibitors competitive inhibition at the ATP-binding pocket. It is there-
fore only capable of characterizing and detecting type I protein
kinase inhibitors. We anticipate that type II inhibitors should
be detectable through their modulation of the peptide sub-
strate binding kinetics (as measured in Subheading 3.5); how-
ever this has not yet been demonstrated.
2. Begin by preparing a dilution series of your desired compound.
We recommend the use of a half-log serial dilution, spanning at
least four orders of magnitude (e.g., from 1 nM to 10 μM final
concentration). As these diluted stocks will be further diluted
upon addition to the electrophysiology compartment, this fac-
tor must be taken into account to achieve the correct final
concentrations. We typically add 2 μL of inhibitor to a 1 mL
compartment. Dilutions should be made in the same solvent as

the starting stock solution, which will typically be DMSO. Care
should be taken to prepare the dilution series accurately, avoid-
ing pipetting errors. Dilutions can be stored at 20 C if
required over several days.
3. Obtain a single sensor pore in a planar lipid bilayer, as described
in Subheading 3.4. This must be the pseudosubstrate variant to
avoid phosphorylation of the sensor peptide. With no mem-
brane potential applied, add 200 nM Pim-1 (or an appropriate
concentration of another kinase complementary to the sensor
pore) to the trans compartment and mix the compartment
contents for 30 s (use the same procedure as in Note 8).
Then apply a potential of 50 mV and observe kinase binding
for several minutes to confirm correct functioning of the
sensor pore.
4. With no applied membrane potential, add 1 M MgCl2 to
achieve a final concentration of 10 mM MgCl2, and 100 mM
ATP stock solution to a final concentration of 100 μM ATP in
the trans compartment. Mix the compartment for 30 s, then
apply a potential of 50 mV and confirm visually that synergis-
tic binding is occurring (see Fig. 4b).
5. With no applied potential, add inhibitor to the trans compart-
ment and mix for 30 s (use the procedure in Note 8 to ensure
efficient mixing). Apply a potential of 50 mV and then record
for around 10 min to obtain sufficient binding events for
analysis.
6. Each concentration point of the titration should be performed
in at least triplicate (see Note 10).
7. Once all recordings have been made, analysis can be performed
to determine the inhibition constant (Ki). As discussed in
Subheading 3.5, extract the required sections of recording in
Clampfit, and then export them to QuB file format (LDT).
8. In QuB, idealize recordings using the SKM method and the
model shown in Fig. 4c (see Note 9). Then fit the rate constants
to the idealized data using the MIL method. The resulting rate
constants should then be collated in a spreadsheet or data
analysis software for subsequent fitting.
9. Using data analysis software, plot the pseudo-first-order asso-
ciation rate constant for the binary population (k0þ2 , see Note
11) vs. log inhibitor concentration. Make a similar plot of the
pseudo-first-order association rate constant for the ternary
population (k0þ3) vs. log inhibitor concentration. These should
both be sigmoid curves, with the binary population values
increasing with inhibitor concentration, and those for the ter-
nary population decreasing with inhibitor concentration (see
Fig. 6).
a b
Fig. 6 Example inhibitor titration curves. The dependences of the pseudo-first-order association rate constants
for formation of (a) the combined PIM-1 and PIM-1-inhibitor “binary” population (k 0þ2 , see Note 12), and (b)
the ternary (k 0þ3 ) complex on the concentration of inhibitor “1” are plotted together with their fitted curves
according to the equations in Subheading 3.7. The concentration of PIM-1 was 162 nm throughout. Error bars
represent S.D. (n ¼ 3). (Adapted with permission from ref. 33. Copyright 2015 WILEY-VCH Verlag GmbH &
Co. KGaA, Weinheim)
10. Use nonlinear regression to fit the following equations to each

set of data (see Note 12):
Binary population:

½I 0
0
kþ2 1 þ Ki
¼ kþ2 ñ
½Pim0 ½ATP0 ½I
1 þ KinaseñMgATP þ K 0i
Kd
Ternary population:
k0þ3 k ½ATP0
¼ þ3
½Pim0 PimñMgATP ½ATP0 ½I 0
Kd 1 þ KinaseñMgATP þ K i
Kd
KinaseñMgATP
During fitting K d should be fixed to a previously
determined value (see Subheading 3.6) where possible, to
reduce the degrees of freedom and avoid overfitting.
11. Fitting of both the binary and ternary population datasets
should yield two values for the inhibition constant (Ki) in
good agreement with each other. Fitting also provides the
bimolecular association rate constants for binary (k+2) and
ternary (k+3) complex formation. It is always advisable to use
an alternative method to corroborate determined inhibition
constants.
4 Notes
1. The electrophysiology buffer should be carefully chosen to suit

both the requirements of the kinase, i.e., to maintain normal
activity, but also to meet the requirements of single-channel
electrical recording. High salt concentrations are usually
desired for single-channel electrical recordings as this gives a
higher pore current and hence better signal-to-noise. How-
ever, such high concentrations (e.g., 1–2 M KCl) will likely
inhibit kinase activity. We use 300 mM KCl, which gives ade-
quate signal-to-noise with an applied potential of 50 mV and
filtering at 2 kHz.
2. Due to the single-molecule sensitivity of the experiments to
kinase–nucleotide complexes, the condition of ATP stock solu-
tions is critical. Hydrolysis of ATP stock solutions will give
inaccurate measurements due to the presence of ADP, AMP,
etc. To minimize hydrolysis, the stock solution should be
titrated to neutral pH. We have found titration with the weak
base Tris to be preferable over the strong base NaOH, as this
avoids hydrolysis caused by high local concentrations of
hydroxide ions during titration. To further avoid hydrolysis,
single-use aliquots should be prepared and stored at 80 C for
a maximum of 12 months.
3. For Pim kinase sensor pores, the notation “PLM” stands for
“pimtide loop mutant,” and “PSLM” stands for “pseudosub-
strate loop mutant.” The sequence of pimtide is:
ARKRRRHPSGPPTA, and that of the pseudosubstrate is:
ARKRRRHPAGPPTA.
4. The D127N mutation is necessary to avoid undesirable elec-
trostatic interactions between the αHL pore mouth and the
sensor peptide. WT-αHL possesses three charged residues at
the pore mouth: D127, D128, and K131. This leads to a net
negative charge at the pore mouth under experimental condi-
tions, whereas pimtide has a high net positive charge. When
pimtide-based sensor pores were made in the WT-αHL back-
ground, we did not see the characteristic fast blocking behavior
that we attribute to motions of the sensor peptide, presumably
because electrostatic interactions between the pore mouth and
sensor peptide “latch” the peptide in a fixed conformation. The
D127N mutation neutralizes the net charge of the pore mouth
and allows free motion of the sensor peptide.
5. It is always preferable to use agar bridges, as they reduce the
access resistance, minimize junction potential offsets, and pre-
vent leaching of Ag+ into solution. Agar bridges are essential
when using the reducing agent DTT in the measurement
buffer, which is often required to maintain kinase activity, to
prevent chemical attack and degradation of the electrodes.
6. We found that higher applied membrane potentials increased

the probability of the pore entering a permanently closed state.
This state can only be escaped by reversing the membrane
potential. This may be due to interactions of the sensor peptide
with the pore barrel or a collapse of the barrel itself. We found
50 mV to be the optimal potential for obtaining good signal-
to-noise while avoiding undesired pore closures.
7. Occasionally we see pores insert that exhibit a single conduc-
tance state without the characteristic blockades caused by the
attached sensor peptide. These are likely to be due to contam-
inating traces of the αHL-D127N homoheptamer. They are
easily rejected due to their markedly different current
signatures.
8. As protein stock solutions are comparatively dense due to the
presence of glycerol in their storage buffers, ensure effective
mixing by transferring 200 μL buffer from the trans compart-
ment to a 1.5 mL microcentrifuge tube. Protein kinase stock
solution can then be added and mixed in the tube by gentle
pipetting. Immediately transfer the mixed contents of the tube
back to the trans compartment, which should then be mixed
for 30 s by use of a magnetic stirrer flea operated from under-
neath by a small bar magnet fixed to a variable speed
motor [34].
9. When performing this experiment with PIM-1, we found it
necessary to include an additional blocked state B2 connected
to O3. We have previously identified additional current noise
present during Pim kinase binding to the sensor, which is likely
due to partially resolved current transitions arising from rapid
motions of the unstructured termini of the Pim kinases
[31]. The inclusion of this extra state avoids erroneous ideali-
zation of the long ternary complex events that could give rise to
falsely truncated event durations. This may not be necessary
with other kinases that do not exhibit such fluctuations (e.g.,
PKA).
10. We prefer to perform a new experiment for each inhibitor
concentration, to avoid potential systematic error due to ATP
hydrolysis over the course of the experiment (protein kinases
typically have some innate ATPase activity, the strength of
which will vary depending on the particular kinase). If sequen-
tial titrations are desired in a single experiment, the total dura-
tion of the experiment should be kept as short as possible, and
controls performed to ensure that the effective ATP concen-
tration is not significantly decreased in the latter stages of the
experiment.
11. Although we continue to use the notation k0þ2 for simplicity, it

is important to note that the “binary” population contains
binding events from both apo-kinase and inhibitor-bound
kinase. When type I kinase inhibitors were titrated against
kinase in the absence of MgATP, we saw no change in the
pseudo-first-order association rate constant for kinase binding
or the dissociation rate constant [33]. This allows us to treat
the apo-kinase and inhibitor-bound kinase binding events
together as a single population.
12. These equations make the assumption that Ki [Pim], which
significantly simplifies their form but will not always be true.
However, we have found the more exact forms to give almost
identical results in at least one case where Ki [Pim]; therefore
we suggest the use of the simplified forms where possible [33].
Acknowledgements
This work was supported by grants from the National Institutes of

Health and Oxford Nanopore Technologies. L.H. was supported
in part by a Biotechnology and Biological Sciences Research Coun-
cil doctoral training grant. S.K. and L.T.A. are supported by the
Structural Genomics Consortium, a registered charity (number
1097737) that receives funds from AbbVie, Bayer Pharma AG,
Boehringer Ingelheim, Canada Foundation for Innovation, Eshel-
man Institute for Innovation, Genome Canada, Innovative Medi-
cines Initiative (EU/EFPIA) [ULTRA-DD grant no. 115766],
Janssen, Merck KGaA Darmstadt Germany, MSD, Novartis
Pharma AG, Ontario Ministry of Economic Development and
Innovation, Pfizer, São Paulo Research Foundation-FAPESP,
Takeda, and Wellcome [106169/ZZ14/Z].
References
1. Bayley H, Cremer PS (2001) Stochastic sensors 6. Robertson JWF, Rodrigues CG, Stanford VM,
inspired by biology. Nature 413:226–230 Rubinson KA, Krasilnikov OV, Kasianowicz JJ
2. Hammerstein AF, Shin SH, Bayley H (2010) (2007) Single-molecule mass spectrometry in
Single-molecule kinetics of two-step divalent solution using a solitary nanopore. Proc Natl
cation chelation. Angew Chem Int Ed Engl Acad Sci U S A 104:8207–8211
49:5085–5090 7. Movileanu L, Cheley S, Bayley H (2003) Par-
3. Cheley S, Gu L, Bayley H (2002) Stochastic titioning of individual flexible polymers into a
sensing of nanomolar inositol 1,4,5- nanoscopic protein pore. Biophys J
trisphosphate with an engineered pore. Chem 85:897–910
Biol 9:829–838 8. Cherf GM, Lieberman KR, Rashid H, Lam CE,
4. Guan X, Gu L, Cheley S, Braha O, Bayley H Karplus K, Akeson M (2012) Automated for-
(2005) Stochastic sensing of TNT with a ward and reverse ratcheting of DNA in a nano-
genetically engineered pore. ChemBioChem pore at 5-Å precision. Nat Biotechnol
6:1875–1881 30:344–348
5. Bezrukov SM, Vodyanoy I, Parsegian VA 9. Deamer D, Akeson M, Branton D (2016)
(1994) Counting polymers moving through a Three decades of nanopore sequencing. Nat
single ion channel. Nature 370:279–281 Biotechnol 34:518–524
10. Garalde DR, Snell EA, Jachimowicz D, through an α-hemolysin nanopore. Nat Bio-
Sipos B, Lloyd JH, Bruce M, Pantic N, technol 31:247–250
Admassu T, James P, Warland A, Jordan M, 23. Soskine M, Biesemans A, Moeyaert B,
Ciccone J, Serra S, Keenan J, Martin S, Cheley S, Bayley H, Maglia G (2012) An engi-
McNeill L, Wallace EJ, Jayasinghe L, neered ClyA nanopore detects folded target
Wright C, Blasco J, Young S, Brocklebank D, proteins by selective external association and
Juul S, Clarke J, Heron AJ, Turner DJ (2018) pore entry. Nano Lett 12:4895–4900
Highly parallel direct RNA sequencing on an 24. Mohammad MM, Iyer R, Howard KR, McPike
array of nanopores. Nat Methods 15:201–206 MP, Borer PN, Movileanu L (2012) Engineer-
11. Hayden E (2012) Nanopore genome sequencer ing a rigid protein tunnel for biomolecular
makes its debut. Nature News, https://doi.org/ detection. J Am Chem Soc 134:9521–9531
10.1038/nature.2012.10051 25. Brown CG, Clarke J (2016) Nanopore devel-
12. Jain M, Olsen HE, Paten B, Akeson M (2016) opment at Oxford Nanopore. Nat Biotechnol
The Oxford Nanopore MinION: delivery of 34:810–811
nanopore sequencing to the genomics commu- 26. Soskine M, Biesemans A, Maglia G (2015)
nity. Genome Biol 17:239 Single-molecule analyte recognition with
13. Kasianowicz JJ, Brandin E, Branton D, Deamer ClyA nanopores equipped with internal protein
DW (1996) Characterization of individual adaptors. J Am Chem Soc 137:5793–5797
polynucleotide molecules using a membrane 27. Willems K, Van Meervelt V, Wloka C, Maglia G
channel. Proc Natl Acad Sci U S A (2017) Single-molecule nanopore enzymol-
93:13770–13773 ogy. Philos Trans R Soc Lond Ser B Biol Sci
14. Manrao EA, Derrington IM, Laszlo AH, Lang- 372:1726
ford KW, Hopper MK, Gillgren N, 28. Ma H, Deacon S, Horiuchi K (2008) The chal-
Pavlenok M, Niederweis M, Gundlach JH lenge of selecting protein kinase assays for lead
(2012) Reading DNA at single-nucleotide res- discovery optimization. Expert Opin Drug
olution with a mutant MspA nanopore and Discov 3:607–621
Phi29 DNA polymerase. Nat Biotechnol
30:349–353 29. Xie H, Braha O, Gu LQ, Cheley S, Bayley H
(2005) Single-molecule observation of the cat-
15. Pennisi E (2012) Genome sequencing. Search alytic subunit of cAMP-dependent protein
for pore-fection. Science 336:534–537 kinase binding to an inhibitor peptide. Chem
16. Stoddart D, Heron AJ, Mikhailova E, Biol 12:109–120
Maglia G, Bayley H (2009) Single-nucleotide 30. Cheley S, Xie H, Bayley H (2006) A genetically
discrimination in immobilized DNA oligonu- encoded pore for the stochastic detection of a
cleotides with a biological nanopore. Proc Natl protein kinase. ChemBioChem 7:1923–1927
Acad Sci U S A 106:7702–7707
31. Harrington L, Cheley S, Alexander LT,
17. Movileanu L, Schmittschmitt JP, Scholtz JM, Knapp S, Bayley H (2013) Stochastic detection
Bayley H (2005) Interactions of peptides with a of Pim protein kinases reveals electrostatically
protein pore. Biophys J 89:1030–1045 enhanced association of a peptide substrate.
18. Kukwikila M, Howorka S (2010) Electrically Proc Natl Acad Sci U S A 110:E4417–E4426
sensing protease activity with nanopores. J 32. Harrington L, Alexander LT, Knapp S, Bayley
Phys Condens Matter 22:454103 H (2019) Single-molecule protein phosphory-
19. Movileanu L, Howorka S, Braha O, Bayley H lation and dephosphorylation by nanopore
(2000) Detecting protein analytes that modu- enzymology. ACS Nano 13:633–641
late transmembrane movement of a polymer 33. Harrington L, Alexander LT, Knapp S, Bayley
chain within a single protein pore. Nat Biotech- H (2015) Pim kinase inhibitors evaluated with
nol 18:1091–1095 a single-molecule engineered nanopore sensor.
20. Rodriguez-Larrea D, Bayley H (2013) Multi- Angew Chem Int Ed Engl 54:8154–8159
step protein unfolding during nanopore trans- 34. Maglia G, Heron AJ, Stoddart D, Japrung D,
location. Nat Nanotechnol 2013(8):288–295 Bayley H (2010) Analysis of single nucleic acid
21. Rosen CB, Rodriguez-Larrea D, Bayley H molecules with protein nanopores. Methods
(2014) Single-molecule site-specific detection Enzymol 475:591–623
of protein phosphorylation with a nanopore. 35. Montal M, Mueller P (1972) Formation of
Nat Biotechnol 2014(32):179–181 bimolecular membranes from lipid monolayers
22. Nivala J, Marks DB, Akeson M (2013) and a study of their electrical properties. Proc
Unfoldase-mediated protein translocation Natl Acad Sci U S A 69:3561–3566
Chapter 9
A Selective Activity-Based Approach for Analysis

of Enzymes with an OmpG Nanopore
Monifa A. V. Fahie, Bach Pham, Fanjun Li, and Min Chen
Abstract
Many enzymatic activity assays are based on either (1) identifying and quantifying the enzyme with methods
such as western blot or enzyme-linked substrate assay (ELISA) or (2) quantifying the enzymatic reaction by
monitoring the changing levels of either product or substrate. We have generated an outer membrane
protein G (OmpG)-based nanopore approach to distinguish enzyme identity as well as analyze the enzyme’s
catalytic activity. Here, we engineered an OmpG nanopore with a peptide cut site inserted into one of its
loops to detect proteolytic behavior. In addition, we generated an OmpG nanopore with a single-stranded
DNA attached to a loop for analyzing nucleolytic cleavage. This OmpG nanopore approach may be highly
useful in analyzing specific enzymes in complex biological samples, or in directly determining kinetics of
enzyme-substrate complex association and dissociation.
Key words Nanopore engineering/design, Enzyme activity sensor, OmpG, Nuclease, Caspase, Pro-
tease, DNA ligand, Peptide substrate, Enzyme-substrate complex, Single-channel recording
1 Introduction
Enzymes with catalytic cleaving activities (-ases) are ubiquitous in

biology. They are essential for catabolic and anabolic processes and
therefore can help to (1) maintain the cell’s survival such as with
lysosomal enzymes during autophagic flux [1] or (2) carry out a
sequence of events leading to cell death such as with caspases
during apoptosis [2]. Studies in enzyme activity is popular in bio-
chemistry, where bulk activity assays that either monitor the loss of
substrate or the gain of product are most common. For enzymes
with catalytic cleaving activity a reduction in the substrate’s molec-
ular weight is a feasible readout of enzymatic activity. For substrates
with molecular weights too small to distinguish a reduction in size,
assays that take advantage of chromogenic of fluorogenic substrates
have been utilized.
Nanopores have been previously used to detect enzymes either
directly [3–7] or indirectly through their enzymatic products which
115
can be made more specific with the use of nanopore coupled

enzyme systems [8–16]. Although nanopores have successfully
detected either the enzymes or their products, there are still few
examples that can distinguish enzyme homologs [17]. In addition,
there have been little example of nanopore that can distinguish the
different states of the enzyme/substrate interaction as they prog-
ress through a cleavage reaction. In this chapter, we describe the
analysis of two different cleaving enzymes, an endonuclease and a
protease, by an outer membrane protein G (OmpG) nanopore.
Endonucleases are a class of enzymes which cleave the phos-
phodiester bond within a nucleic acid moiety. They can either
cleave DNA or RNA, whether single stranded or double stranded.
These class of enzymes are involved in processes such as DNA
replication [18], DNA repair [19], and programmed cell death
[20]. Another class of enzymes, proteases, cleave peptide bonds
within a protein molecule and their functions are found in diverse
processes such as in viral infection and autophagy [2, 21].
Here, we describe an OmpG-based nanopore platform for not
only directly detecting enzymes but also capturing specific steps in
their catalytic process which can ultimately be used to selectively
and specifically analyze enzyme behavior in a mixture containing
related enzymes or homologs.
2 Materials
2.1 Cloning 1. pT7-OmpGwt plasmid.

of Mutant OmpG 2. Primers for mutagenesis (see Table 1).
Proteins
3. High fidelity polymerase such as Phusion.
4. Restriction enzymes such as DpnI.
5. Chemically competent E. coli cells such as DH5α.
6. 100 mg/mL Ampicillin dissolved in sterile water stock.
2.2 Protein 1. LB media (Miller high salt version): 10 g/L Tryptone, 5 g/L
Expression of OmpG yeast extract, 10 g/L sodium chloride, 100 μg/mL ampicillin.
Mutants 2. LB Agar plates with 100 μg/mL ampicillin.
3. Chemically competent E. coli cells such as BL21 DE3.
4. Isopropyl β-D-1-thiogalactopyranoside (IPTG).
2.3 Preparation 1. Sonicator, French press, or microfluidizer.

of OmpGcasp or 2. Centrifuge capable of pelleting volumes of 250 mL or more at
OmpGL6neu-DNA 3000 g.
Nanopore
3. Centrifuge capable of pelleting volumes of 50 mL or less at
12,000 g.
4. 0.45 μm filter for filtering lysate supernatant.
Table 1
Oligonucleotides used to generate OmpG nanopores for enzyme sensing
OmpG construct OmpG mutations Sequence

OmpGL6neu-DNA D221S, D224C, D225S, F: CAGTGTAGTATTACCGGCACGGGCCATGATTTTAAC
E227T, R228G, E229T R: GGTAATACTACACTGCCAGCTCCAGTTACTCCAGCG
OmpGcasp H231A F: AGCGATTTTAACCGTGTAGGTTTATTTTAC
R: ACGGTTAAAATCGCTGCCCGTGCCGGTAATAC
H261A F: GATGCCGACGAAGGCGACAGTGATAAATTC
R: ACTGTCGCCTTCGTCGGCATCCTGCCACTC
Deletion of D215 F: GGGCTGCGCTGGAGTAACTGGGACTGGCAG
R: ACTCCAGCGCAGCCCAATGCGCGTATACGG
Insertion of GDEVDG F: GATGAAGTCGACGGTGATATTGAACGTGAAGGCC
R: ATCACCGTCGACTTCATCTCCCTGCCAGTCCCAG
Analyzing proteases or Nucleases with OmpG Nanopores
117
5. Anion exchange columns (GE, 5 mL).

6. HiTrap Desalting columns (GE, 5 mL).
7. Fast protein liquid chromatography (FPLC) such as
AKTA (GE).
8. Lysis buffer: 50 mM Tris–HCl, pH 8, 150 mM NaCl,
1 mM EDTA.
9. Inclusion Body wash buffer: 50 mM Tris–HCl, pH 8,
1.5 M urea.
10. Anion exchange low salt buffer: 50 mM Tris–HCl, pH 8,
50 mM NaCl, 8 M urea.
11. Anion exchange high salt buffer: 50 mM Tris–HCl, pH 8,
500 mM NaCl, 8 M urea.
12. Tris(2-carboxyethyl)phosphine (TCEP) or dithiothreitol
(DTT) for purification of cysteine containing OmpG protein.
13. OmpG conjugation buffer: 50 mM HEPES, pH 7, 150 mM
NaCl, 8 M urea.
14. Thiolated DNA oligonucleotide (23mer).
15. Centricon with 3 kDa molecular weight cutoff.
16. TE buffer: 10 mM Tris–HCl, pH 8, 0.1 mM EDTA.
17. Nonreducing elution buffer: 50 mM Tris–HCl, pH 8, 8 M
urea, 200 mM NaCl.
18. 2-Dipyridyl disulfide (2-PDS) dissolved in acetonitrile.
19. Diethyl ether.
20. Refolding buffer: 50 mM Tris–HCl, pH 9.0, 3.25% octyl glu-
coside (OG).
21. Glycerol (for long-term storage of refolded OmpG proteins).
2.4 Bulk Activity 1. Chromogenic or fluorogenic caspase 3/7 substrate such as

Assays for Caspase or Ac-DEVD-AFC.
Nuclease S1 Enzymes 2. Caspase enzyme.
3. Caspase activity buffer: 50 mM HEPES, 150 mM NaCl,
pH 7.5.
4. Single-stranded complementary DNA oligonucleotide
(23mer).
5. Nuclease S1 enzyme.
6. Nuclease S1 activity buffer: 20 mM sodium acetate, pH 5, 1 M
NaCl, 5 mM CaCl2.
7. 1 M Tris–HCl, pH 8 (to quench the nuclease S1 activity).
8. SDS Polyacrylamide gel electrophoresis equipment and
reagents.
9. UV-Vis spectrophotometer or plate reader.
Analyzing proteases or Nucleases with OmpG Nanopores 119
2.5 Single-Channel 1. 1,2-Diphytanoyl-sn-glycero-3-phosphocholine (DPhPC)

Recording of OmpG lipids.
Nanopores 2. Hexadecane.
3. Pentane to dissolve DPhPC or hexadecane.
4. Teflon film.
5. Silicone Glue.
6. Piezoelectric spark generator.
7. Sandpaper.
8. Air table (to minimize vibrations which lowers signal-to-noise
ratio).
9. A two-chamber chip made from a polyoxymethylene material
such as Delrin.
10. Two Ag/AgCl electrodes.
11. Digidata 1320A/D board (Axon Instruments).
12. Axopatch 200B integrating patch clamp amplifier (Molecular
Devices).
3 Methods
3.1 Cloning of OmpG 1. For the nuclease S1 sensor, mutate six amino acid residues on
Mutants for Enzyme loop 6, one of which includes a cysteine at position
Sensing 224 (OmpGL6neu-C224).
2. For the caspase sensor, make three mutations: H231A,
H261A, and a deletion of D215 on the OmpGwt backbone.
For the fourth mutation, replace the amino acid residue D224
with a caspase cleavage site GDEVDG into loop 6. The muta-
genic oligonucleotides to generate the two different sensors are
found in Table 1.
3. Use the appropriate competent cells such as DH5α for bacterial
transformations of the desired DNA construct.
3.2 Expression 1. Isolate single colonies to inoculate an overnight LB culture

of the Desired containing ampicillin of about 5 mL. Use this overnight culture
Plasmids in BL21 (DE3) to start a larger culture of at least 100 mL on the next day.
Competent Cells 2. Grow culture at 37 C with 200–250 rpm shaking to an OD600
of 0.6–0.8. Depending on how much starter culture used, this
may take 2–4 h shaking incubation.
3. Induce with a final concentration of 0.5 mM IPTG and harvest
the cell pellet after 3–4 h of shaking at 37 C, 200–250 rpm.
Remove the supernatant from the cell pellet. Freeze the cell
pellet at 20 C for up to 6 months.
3.3 Lysis 1. Thaw the harvested cell pellet and resuspend in lysis buffer.
and Inclusion Body N.B. If preparing the nuclease S1 sensor, go directly to Sub-
Preparation heading 3.6 for specific details on the preparation of this
protein.
2. Lyse the cell suspension using either a sonicator, French press,
or microfluidizer. Clear the lysate by centrifugation at
20,000 g, 4 C for 20 min. Discard the supernatant.
3. Resuspend the pellet with inclusion body wash buffer and
centrifuge again at previous speed, temperature, and time.
Discard the supernatant. The inclusion body pellet can be
stored in 20 C freezer for up to 6 months.
3.4 Purification from 1. Dissolve the OmpG-containing inclusion body in anion

Inclusion Body exchange low salt buffer for 30 min using a stir bar and a stir
plate to help break up the inclusion body.
2. Centrifuge at 20,000 g for 20 min to pellet any undissolved
inclusion body. Pass the supernatant through a 0.45 μm filter.
Purify the OmpG proteins with a Q-anionic exchange column.
Wash and elute with anion exchange high salt buffer by apply-
ing a gradient (Fig. 1). OmpG protein elutes around 200 mM
NaCl, thus around 30% gradient.
3. Run the elution fractions on SDS-PAGE to analyze protein
quality. Choose the purest and highly concentrated fractions
to refold (see Subheading 3.5). Denatured proteins may be
stored in 80 C for up to 6 months (if there is no cysteine
mutation).
OmpG elution 500 mM NaCl
2.5-
2.0-
Absorbance 280 (AU)
1.5-
200 mM NaCl
1.0-
0.5- Wash
Flow-Through
75 mM NaCl
0 mM NaCl
0.0-
0 10 20 30 40 50 60 70
Volume (ml)
Fig. 1 Chromatogram of an OmpG anion exchange purification from inclusion bodies. A gradient of sodium
chloride concentration in 8 M urea denaturing buffer is used to sequentially elute the OmpG protein from a Q
sepharose column
3.5 Refolding 1. After collecting the purest OmpG proteins, mix in the refold-
of Denatured OmpG ing buffer dropwise until the final concentration of urea reaches
Proteins 3.0 M or below. For example, with highly concentrated OmpG
samples (50–100 μM) mix 1 mL of purified OmpG with about
2.5 mL of refolding buffer; this typically yields >90% refolded
protein.
2. Incubate the refolding OmpG sample at 25 or 37 C for at least
36 h before using the protein for single-channel recording.
Test refolding efficiency by a gel shift assay [22].
3. For long-term storage, mix refolded proteins with glycerol to a
final concentration of 20% v/v. Dispense the protein in several
small aliquots. Flash freeze in liquid nitrogen and store
refolded proteins at 80 C for up to 6 months.
3.6 Preparation It is recommended that the DNA ligand is activated with 2-dipyr-
of Nuclease Sensor idyl disulfide (2-PDS) (see Note 1) and stored at 20 C before
OmpG Nanopore purifying cysteine containing OmpG. After removing the reducing
agent from purified OmpGL6neu-C224, keep protein on ice while
determining protein concentration and subsequently label OmpG
with the 2-PDS DNA ligand within 2 h to obtain at least 50%
conjugated protein.
3.6.1 Activation 1. Quick centrifuge the dried 23mer DNA at 10,000 g to

of Thiolated DNA Ligand reduce any loss upon opening the vial.
(deprotection) 2. Add 100 μL of TE buffer and vortex to solubilize the DNA.
3. Add 100 μL of 0.5 M TCEP, pH 7 to the DNA and incubate at
room temperature for ~2 h to deprotect the thiol group.
4. Store the reduced DNA mixture in the 20 C freezer until
further use or continue with the 2-dipyridyl disulfide (2-PDS)
conjugation in Subheading 3.6.2 below.
3.6.2 Conjugation 1. Desalt the deprotected ligand to remove the excess reducing
of 2-PDS to Thiolated DNA agent. Apply the DNA mixture to a 5 mL centricon with an
appropriate kDa cutoff according to the DNA molecular
weight.
2. Add 5 mL TE buffer to dilute the DNA. Concentrate the DNA
to ~100 μL on the centricon by centrifuging at an appropriate
speed according to manufacturer’s instructions.
3. Add another 5 mL TE buffer and concentrate the DNA.
Repeat the TE dilution and subsequent concentration two
more times to bring the reducing agent concentration to a
sub-micromolar level.
4. During the final centricon-based desalting step, prepare a
working solution of 50 mM 2-dipyridyl disulfide (2-PDS)
solution in acetonitrile. For example, dissolve 23.3 mg 2-PDS
in 106 μL acetonitrile for a concentration of 1 M. Then take

5 μL of the 1 M solution and dilute in 95 μL acetonitrile to
make a 50 mM solution.
5. Resuspend the resulting ~100 μL desalted DNA in the centri-
con and transfer from the centricon into a 1.5 mL tube. Mix
the desalted DNA with an equal volume of 50 mM 2-PDS.
Incubate for 2 h at room temperature and mix by inverting
every 20–30 min.
6. To remove the excess 2-PDS, mix at least 200 μL of diethyl
ether into the DNA mixture in order to distinguish the two
layers.
7. Centrifuge at top speed for 30 s. Discard ~150–180 μL of the
top layer (the diethyl ether layer) into an organic waste bottle.
8. Repeat the diethyl ether extraction another nine times.
9. Allow the residual diethyl ether in the DNA solution to evapo-
rate for 4–16 h in a chemical hood.
10. Determine the DNA molar concentration and store the acti-
vated 2-PDS DNA in a 20 C freezer until ready to use for
labeling (see Note 2).
3.6.3 Lysis and Inclusion 1. Resuspend the OmpG cell pellet in lysis buffer supplemented
Body Preparation with 3 mM TCEP (or DTT). Lyse cells on ice using any
of OmpGL6neu-C224 mechanical lysis method of your choice.
2. Isolate the inclusion body pellet by centrifuging at 20,000 g
for 20 min. Resuspend the pellet with inclusion body wash
buffer containing 3 mM TCEP and centrifuge again at previous
speed and time. Discard supernatant. The resulting inclusion
body pellet can be stored in 20 C freezer for up to 6 months.
3.6.4 Purification from 1. As in Subheading 3.3, dissolve OmpG-containing inclusion

Inclusion Body body in anion exchange low salt buffer containing 3 mM
TCEP for 30 min with stir bar and stir plate to help break up
the inclusion body for dissolution.
2. Centrifuge to pellet any undissolved inclusion body. Pass super-
natant through a 0.45 μm filter. Purify OmpG proteins with a
Q-anionic exchange column under reducing conditions. Wash
and elute with anion exchange high salt buffer containing
3 mM TCEP by applying a gradient (Fig. 1).
3. Analyze the protein fractions by SDS-PAGE. Determine the
concentration of the purest OmpG containing fractions with an
appropriate protein concentration assay such as the Bradford
assay. Desalt the protein with a 5 mL desalting column into
OmpG conjugation buffer (see Note 3).
3.6.5 Labeling of OmpG 1. Incubate the OmpGL6neu-Cys with activated 2-PDS DNA in
Cysteine Mutant a 1:1 molar ratio (protein:DNA).
with Activated 2-PDS DNA 2. Shake gently at room temperature for 2 h or alternatively for
12–18 h at 4 C.
3. Separate conjugated OmpG-Cys-DNA from unlabeled
OmpG-Cys on a nonreducing SDS-PAGE gel with a percent-
age suitable for significant separation between
conjugated vs. the unlabeled OmpG-Cys (see Note 4).
4. Excise the band of the OmpGL6neu-DNA which runs slower
than unconjugated, at about ~50 kDa (Fig. 2) and extract it
from the gel by crushing gel with a pestle in 3 the gel volume
of nonreducing elution buffer (see Note 5).
5. Incubate with gentle agitation at 25 C for 2–6 h. Continue to
elute protein from crushed gel by storing at 4 C for 14–18 h.
6. Next day, centrifuge the crushed gel mixture at high speed
(at least 10,000 g) for 15 min at room temperature (see
Note 6). Collect the eluted protein.
7. Eluted denatured OmpGL6neu-DNA can be flash frozen in
liquid nitrogen and stored in 80 C for up to 6 months.
Otherwise, go to Subheading 3.5 for refolding of the
DNA-labeled OmpG protein.
3.7 Cleavage Assay 1. Mix caspase and refolded OmpGcasp together in caspase activ-
of OmpG Protease ity buffer in either a 1:20 or up to a 1:1 caspase to OmpG mole
Sensor ratio.
2. Incubate for 0.5–4 h at room temperature (22–25 C) (see
Note 7). Heat denature all samples by boiling at 95 C for
15 min in SDS-PAGE Laemmli sample buffer and then run on
appropriate acrylamide gel (Fig. 3a).
85-
60-
-OmpGL6neu-DNA
-unlabeled OmpGL6neu
40-
25-
15-
kDa
Fig. 2 SDS-PAGE analysis of labeled vs. unlabeled OmpG construct. OmpG

incubated with activated thiolated DNA was run on a 15% gel to separate
labeled vs. unlabeled OmpG proteins
(a) (b)
nS1 - + -
Casp-3 - + - DTT - - +
Casp-7 - - + kDa
-85
-60
OmpG DNA-
OmpG casp - - 40
Cleaved -
Cleaved - - 30 OmpG DNA -40
OmpG casp - 25
-25
kDa
Fig. 3 Cleavage gels of OmpG substrates. (a) Caspases 3 and 7 were incubated with OmpGcasp construct in a
1:20 mole ratio. Proteins were boiled in SDS-sample buffer at 95 C for 10 min and run on gel electrophoresis
to analyze cleavage efficiency. (b) OmpGL6neu-DNA construct was incubated with nuclease S1 (nS1) at pH 5.
Proteins were heat denatured and analyzed by gel shift for cleavage efficiency
3.8 Cleavage Assay 1. Incubate the refolded OmpGL6neu-DNA with nuclease S1

of OmpG Nuclease enzyme in nuclease activity buffer for 60 min at 25 C. Quench
Sensor the reaction by increasing the pH to 8 by adding 1 M Tris–
HCl.
2. Mix the protein reactions with SDS Laemmli sample buffer and
boil at 90–95 C for 10 min to denature all proteins. Analyze
nuclease activity by gel shift in SDS-PAGE.
3. As a positive control, treat OmpGL6neu-DNA in 10 mM
DTT. Of note, when running this positive control on the gel,
ensure to leave empty lanes on both sides of this DTT-treated
sample, as the DTT can bleed into other lanes as the voltage is
applied (Fig. 3b).
3.9 Single-Channel 1. Similar to previous studies, perform single-channel-recording

Recording of OmpG experiments in a two-chamber chip with a 25 μm thick Teflon
Proteins film containing a 100 μm aperture separating the chambers
[23] (Fig. 4).
2. To make an aperture for single-channel recording go to Sub-
heading 3.10. Add 1–3 μL solution of 10% v/v hexadecane
dissolved in pentane to coat the Teflon sheet on both faces (see
Note 8). This will coat the aperture in a thin layer of oil. Allow
the volatile pentane to evaporate in 2–5 min at ambient
temperature.
3. Fill each chamber with buffer to completely submerge the
aperture on both faces. Add approximately 15 μg of phospho-
lipids to each chamber to form a monolayer (see Note 9).
1,2-Diphytanoyl-sn-glycero-3-phosphocholine (DPhPC) dis-
solved in pentane at a concentration of 10 mg/mL is the
recommended phospholipid for these experiments. Allow the
Fig. 4 Schematic of the two-chamber chip used to create DPHPC lipid bilayers. The two chambers are
separated by a Teflon film (yellow) which has a 100 μm hole in the center on which lipid monolayers from each
chamber fuse to create a lipid bilayer. The nanopore of interest nestles in the lipid bilayer creating a small
current when voltage is applied
pentane to air dry for 2–5 min or infuse a gentle stream of

compressed air to accelerate the evaporation.
4. Mix the buffer/phospholipid mixture in each chamber by
slowly pipetting up and down at least 75% of the mixture.
The pipetting aids in the formation of monolayers covering
each face of the aperture to create a bilayer.
5. Connect the chip to the Axopatch 200B integrating patch
clamp amplifier by submerging the silver portion of a silver
chloride electrode into the buffered mixture (not too close to
the aperture) and the “gold” portion to the headstage.
6. When voltage is switched on, it is only applied to one chamber
(trans), while the other chamber remains grounded (cis),
thereby generating opposing positive and negative voltage
across each face of the bilayer. To assess the stability of the
bilayer, see Subheading 3.11 for more details.
Fig. 5 Single-channel recording of a single OmpG pore vs. multiple pore insertions. A raw trace showing the
ionic behavior of a single pore at 50 mV until two other pores insert into the lipid bilayer. The ionic trace
becomes noisy and difficult to analyze as opposed to a single OmpG pore
7. After generating a stable bilayer, to promote insertion of a

single OmpG protein, mix in 2–3 nM final concentration to
either of the chambers (trans or cis) and apply a voltage of
about 200–300 mV across the bilayer until a single pore inserts
into the membrane. Once a single pore inserts, lower the
voltage to 20–60 mV depending on the experiment.
8. Filter the current signal by adjusting the low-pass Bessel filter
(80 dB/decade) on the Axopatch 200B integrating patch
clamp to 2 kHz. For signal acquisition, use a sampling at
100 μs after digitization with the Digidata 1440A/D board.
Record the nanopore activity with pClamp software program at
both positive and negative voltages for at least 5–10 min. Eval-
uate the current trace to determine if it is a single pore insertion
(Fig. 5) and the orientation of the pore [22, 24].
9. After determining pore orientation and recording to character-
ize the apo condition of the nanopore, add the enzyme of
interest into the loop facing chamber and gently mix so that
the bilayer is not disturbed. Record the nanopore and enzyme
mixture at both voltages and take note of any changes in
current signal (Figs. 6 and 7).
3.10 Method for 1. Cut an appropriate square of Teflon film and place it on the
Making an Aperture circumference of the chamber to determine the size of the hole.
for the Two- Draw the outline of the circumference of the opening that the
Chamber Chip Teflon film will sit between the two buffer chambers.
2. Using a sharp pin which can be made from a glass capillary,
make an indent in the middle of the Teflon film. Take the
Teflon film and place on the piezoelectric spark generator so
that the ident is positioned in the center of the electrical spark.
Casp-7
asp-7
OmpGcasp
- - -
+
E+S ES E+P
- Casp-7 + Casp-7 Casp-7 Casp-7

0 Pre-Cleavage OmpG Cleavage
Interaction
-5
I(pA)
-10
-15
-20
100 ms
Fig. 6 A characteristic trace of OmpGcasp nanopore with and without caspase-7. The current signal of a single
OmpGcasp nanopore showing a transient interaction with caspase-7 (in red) followed by the cleavage of the
loop at the recognition site (in blue)
unlabeled pore OmpG-ssDNA OmpG-ssDNA + nuclease S1 post cleavage

0 0 0 0
-20 -20 -20 -20

I (pA)
-40 -40 -40 -40
-60 -60 -60 -60
200 ms
Fig. 7 A discontinuous trace of OmpGL6neu-DNA nanopore with and without nuclease S1. Conjugation of a
single-stranded DNA oligonucleotide onto loop 6 of OmpG blocks the lumen at 50 mV which can act as a
positive marker for labeled pore. Binding of nuclease S1 to OmpGL6neu-DNA construct generates a unique
signal that clearly distinguishes it from the final cleavage of the DNA molecule
(a)
100
I (pA)
0
-100
(b)
100
I (pA) 0
-100
200 ms
Fig. 8 Schematic trace of a stable bilayer. (a) The current with or without a
voltage applied is stable at zero without any fluctuations in current. (b) Seal test
signal of a formed bilayer in the chip aperture. Amplitude is above 100 pA
3. Generate 10–15 sparks with the clicker to create an aperture of

approximately 100 μm in diameter. Fit the Teflon film onto the
chamber and optionally use silicone glue to hold it in place.
Assemble the second chamber.
4. Remove any excess Teflon film with razor blade. Allow the glue
to dry for at least 8 h before using the chip. On the first use of
newly sealed Teflon film, rinse the chip at least 5 times alter-
nating between 70% and 95% ethanol and distilled water, where
your final rinse is 70–95% ethanol. Allow chip to air dry or use a
gentle stream of compressed air to let the ethanol evaporate.
3.11 Method 1. Testing the bilayer formation: After slowly pipetting the buffer
for Testing the Bilayer in each chamber to mix lipids, a consistent current of zero
Formation should be observed (Fig. 8a). Evaluate the size and coverage
and Stability of the bilayer with the bilayer test switch. An oscillating sin like
curve is generated with a corresponding amplitude. An ampli-
tude of 80–300 pA (when the scaled output/output gain (α) is
set to 1) is indicative that a 100 μm bilayer has been formed
(Fig. 8b). The amplitude of the sigmoidal curve informs about
the bilayer size. An amplitude no lower than 80 pA and no
larger than 400 pA will usually generate a stable bilayer for
several hours.
2. Test the bilayer stability: After evaluating bilayer formation
with the bilayer seal test, turn off the seal test switch. Apply a
voltage of 200–300 mV across the bilayer for at least 5 min
and monitor the behavior of the current. A stable bilayer will
remain at zero with little to no fluctuations in current. An

unstable bilayer will either break/overload (i.e., a large current
signal) or there may be several fluctuations in current seen. If
the bilayer breaks, pipet the chamber solution to reform the
bilayer. Apply a high voltage again. Once the bilayer stability is
acceptable, continue with experiment by acquiring and record-
ing a single OmpG nanopore.
3. For any further issues with bilayer generation, refer to Notes
10–14 for troubleshooting tips.
4 Notes
1. The conjugation of 2-dipyridyl disulfide to the thiolated DNA

is to make the DNA more reactive to the reduced cysteines on
the OmpG sample. It therefore lowers the possibility of creat-
ing DNA-DNA disulfide dimers.

2. Activated DNA is stable in 20 C freezer for at least
6 months.
3. Reducing agents must be removed prior to conjugating
OmpGL6neu-C224 with activated 2-PDS DNA.
4. Do not incubate the OmpGL6neu-DNA in SDS-sample buffer
containing any reducing agent or run on a PAGE gel contain-
ing reducing agent. This will break the conjugated bond result-
ing in unlabeled OmpG.
5. An alternative to manually crushing the gel with a small pestle
one can also shred the gel by placing gel in a 0.6 mL tube that
has been punctured at the bottom with a 21 G needle to create
a small hole. Place the 0.6 mL tube into a 1.5 mL tube and
centrifuge at high speed at room temperature.
6. Do not allow the eluted protein to warm up above 30 C, the
temperature at which urea will start to degrade and lower the
protein yield by reducing its solubility.
7. Determine whether the protease or nuclease is active in the
buffer of the refolded OmpG sample. Use a positive control
substrate such as a chromogenic substrate (or a ssDNA for the
nuclease) in the presence of 2% octyl glucoside detergent, 3 M
urea, or Tris–HCl, pH 9. If the protein is not active in such a
buffer, dilute the detergent and urea concentration to a level
that retains the enzyme activity. Use this dilution ratio as a
guide to dilute the refolded OmpG sample before incubating
with enzyme. Diluting the detergent in OmpG sample after it
has successfully refolded does not significantly denature the
OmpG at room temperature.
8. Avoid directly touching the pipet tip to the Teflon film. This
may puncture or bend the film as altering its shape integrity
could make bilayer formation difficult.
9. Never apply dissolved phospholipid solution directly to the
Teflon film. The phospholipids will adsorb onto the Teflon
surface. An oil layer of hexadecane is needed to coat the Teflon
film first. The phospholipids should always be added to the
surface of the buffered solution. The oil coating the Teflon
blocks the flow of aqueous solution from one chamber to
the next and is used to correctly orient the phospholipids in
the aperture, i.e., with the lipid tails facing the Teflon film and
the polar phospho-heads facing the buffered solution.
10. A ~100 μm bilayer has a wave amplitude reading of ~100 pA
when the seal test ext command on the Axopatch 200B is
turned on.
(a) If the amplitude reading is smaller than 100 pA, the
aperture may be smaller than 100 μm. Bilayers smaller
than 100 μm may be too stable and unbreakable. Alterna-
tively, there may be too much oil on the Teflon film. Pipet
some of the buffered solution in the chamber toward the
bilayer to wash some of the oil from the Teflon film. Pipet
the solution (as far as possible from the aperture) to
regenerate the bilayer. Measure the amplitude with the
seal test again. Alternatively, add another 5–10 μL more
phospholipids to each chamber, allow to evaporate, and
mix solution slowly. Retest with the seal test.
(b) If the amplitude is around 10 pA or lower, no bilayer has
formed. There may be too little oil on the Teflon film.
Wash the chip several times alternating between 70% and
95% ethanol and water. Dry the chip and remake the lipid
bilayer.
(c) If the amplitude consistently decreases (shrinking bilayer),
break the bilayer and pipet to reform. There may not be
enough lipids to sufficiently form a monolayer in each
chamber; therefore add another 5–10 μL DPhPC lipids
to each chamber.
(d) Amplitudes consistently larger than 300 pA may be indic-
ative of a deformed aperture possibly due to pipet punc-
turing. Discard the old Teflon film and remake a 100 μm
aperture.
11. A stable bilayer has a consistent current of zero with or without
voltage applied
(a) If the current of the bilayer is nonzero (within 5 pA)
without voltage applied, there may be a calibration error.
Turn the pipet offset switch until the current signal
returns to zero. If the pipet offset does not fix the calibra-
tion, then adjust the headstage offset whole cell β ¼ 1
which is found in the back of the Axopatch 200B device.
(b) If the trace fluctuates around zero causing a wobbly base-
line, the silver electrode may be too long and therefore
causing minute vibrations in the connection. Cut the wire
shorter or wrap some of the wire length around itself so
that it can be taut when connected to the headstage.
(c) If the voltage is applied and a nonzero current is observed,
the aperture may not have been properly cleaned from the
previous experiment, where previous nanopore samples
are contaminating the bilayer. Wash the chip extensively
with 70–95% ethanol and water. For a more intensive
clean, sonicate the chip in either 0.1 M NaOH or 1–5%
detergent for 15–30 min. Rinse extensively with tap
water, rinse again in distilled water, then wash with
70–95% ethanol before drying. Remake bilayer.
12. A stable bilayer does not break when up to 200 mV is applied.
If the bilayer breaks when high voltage (or low voltage) is
applied, it is unstable. Check the expiration date of lipids and
oil. Oxidized lipids or oil can cause unstable bilayers. Water
contamination in the lipids or oil can also cause unstable
bilayers. Alternatively, too much oil or lipids coating the Teflon
film can cause instability in the bilayer. As the lipids or oil is
used multiple times, the pentane evaporates resulting in more
concentrated lipids or oil. It is recommended to seal vials
containing lipids or oil with parafilm to reduce pentane evapo-
ration and/or water contamination when stored in the freezer.
13. A stable bilayer does not break when a single or multiple
nanopores insert. If the bilayer consistently breaks only when
a nanopore inserts, there may be insufficient lipids. Add
another 5–10 μL more phospholipids to each chamber. Retest
with the seal test. Alternatively, there may be too many nano-
pores in the chamber. Dilute the nanopores, by removing an
appropriate volume of buffer from the chamber containing the
nanopores. Replenish with new buffer and lipids proportional
to what was removed. Retest bilayer stability and continue with
recording a single nanopore.
14. A stable bilayer breaks when buffered solution is pipetted
directly toward the aperture or when zapped.
(a) If the bilayer does not break when buffer solution is
pipetted toward the aperture (do not place tip directly
on Teflon film), or when the zap switch is used, there
may be too much oil on the Teflon film. Wash the chip
several times in alternating water and 70–95% ethanol and
remake bilayer with less oil.
(b) If the bilayer is unbreakable, the aperture may be smaller

than 100 μm. With smaller bilayers, it is easy to overload
the aperture with oil. Apply less oil to apertures smaller
than 100 μm or use bilayers at least 100 μm in size.
(c) Another possible cause for the bilayer not breaking is a
bad connection caused by the Ag/AgCl electrode. Over
time the silver electrodes become turn white from multi-
ple use in buffer. Regenerate Ag/AgCl by scraping off the
white deposit with sandpaper to expose the silver under-
neath and then incubate the silver wire in 10% bleach
solution for at least 30 min at room temperature, or
several hours in 4 C.
References
1. Ravanan P, Srikumar IF, Talwar P (2017) activity by nanopore analysis. Biosens Bioelec-
Autophagy: the spotlight for cellular stress tron 62:158–162
responses. Life Sci 188:53–67 11. Kukwikila M, Howorka S (2015) Nanopore-
2. D’Arcy MS (2019) Cell death: a review of the based electrical and label-free sensing of
major forms of apoptosis, necrosis and autop- enzyme activity in blood serum. Anal Chem
hagy. Cell Biol Int 43:582–592 87:9149–9154
3. Xie H, Braha O, Gu LQ, Cheley S, Bayley H 12. Nivala J, Marks DB, Akeson M (2013)
(2005) Single-molecule observation of the cat- Unfoldase-mediated protein translocation
alytic subunit of cAMP-dependent protein through an α-hemolysin nanopore. Nat Bio-
kinase binding to an inhibitor peptide. Chem technol 31:247–250
Biol 12:109–120 13. Nivala J, Mulroney L, Li G, Schreiber J, Akeson
4. Cheley S, Xie H, Bayley H (2006) A genetically M (2014) Discrimination among protein var-
encoded pore for the stochastic detection of a iants using an unfoldase-coupled nanopore.
protein kinase. Chembiochem 7:1923–1927 ACS Nano 8:12365–12375
5. Rotem D, Jayasinghe L, Salichou M, Bayley H 14. Ho C-W, Van Meervelt V, Tsai K-C, De Tem-
(2012) Protein detection by nanopores merman P-J, Mast J, Maglia G (2015) Engi-
equipped with aptamers. J Am Chem Soc neering a nanopore with co-chaperonin
134:2781–2787 function. Sci Adv 1:e1500905
6. Harrington L, Cheley S, Alexander LT, 15. Derrington IM, Craig JM, Stava E, Laszlo AH,
Knapp S, Bayley H (2013) Stochastic detection Ross BC, Brinkerhoff H, Nova IC, Doering K,
of Pim protein kinases reveals electrostatically Tickman BI, Ronaghi M, Mandell JG, Gunder-
enhanced association of a peptide substrate. son KL, Gundlach JH (2015) Subangstrom
Proc Natl Acad Sci U S A 110:E4417–E4426 single-molecule measurements of motor pro-
7. Van Meervelt V, Soskine M, Maglia G (2014) teins using a nanopore. Nat Biotechnol
Detection of two isomeric binding configura- 33:1073–1075
tions in a protein aptamer complex with a 16. Craig JM, Laszlo AH, Brinkerhoff H, Derring-
biological nanopore. ACS Nano ton IM, Noakes MT, Nova IC, Tickman BI,
8:12826–12835 Doering K, de Leeuw NF, Gundlach JH
8. Zhao Q, De Zoysa RSS, Wang D, Jayaward- (2017) Revealing dynamics of helicase translo-
hana DA, Guan X (2009) Real-time monitor- cation on single-stranded DNA using high-
ing of peptide cleavage using a nanopore resolution nanopore tweezers. Proc Natl Acad
probe. J Am Chem Soc 131:6324–6325 Sci 114:11932–11937
9. Mohammad MM, Iyer R, Howard KR, McPike 17. Soskine M, Biesemans A, Moeyaert B,
MP, Borer PN, Movileanu L (2012) Engineer- Cheley S, Bayley H, Maglia G (2012) An engi-
ing a rigid protein tunnel for biomolecular neered ClyA nanopore detects folded target
detection. J Am Chem Soc 134:9521–9531 proteins by selective external association and
10. Wang L, Han Y, Zhou S, Guan X (2014) Real- pore entry. Nano Lett 12:4895–4900
time label-free measurement of HIV-1 protease 18. OKazaki T (2017) Days weaving the lagging
strand synthesis of DNA—a personal
recollection of the discovery of Okazaki frag- proteases: Identification, localization and inhi-
ments and studies on discontinuous replication bitors as potential therapeutics. Eur J Cell Biol
mechanism. Proc Jpn Acad Ser B Phys Biol Sci 94:375–383
93:322–338 22. Fahie M, Chisholm C, Chen M (2015)
19. Wood RD (1996) DNA repair in Eukaryotes. Resolved single-molecule detection of individ-
Annu Rev Biochem 65:135–167 ual species within a mixture of anti-biotin anti-
20. Enari M, Sakahira H, Yokoyama H, Okawa K, bodies using an engineered monomeric
Iwamatsu A, Nagata S (1998) A caspase- nanopore. ACS Nano 9:1089–1098
activated DNase that degrades DNA during 23. Pham B, Eron SJ, Hill ME, Li X, Fahie MA,
apoptosis, and its inhibitor ICAD. Nature Hardy JA, Chen M (2019) A nanopore
391:43–50 approach for analysis of caspase-7 activity in
21. Garten W, Braden C, Arendt A, Peitsch C, cell lysates. Biophys J 117:844–855
Baron J, Lu Y, Pawletko K, Hardes K, 24. Chen M, Li Q-H, Bayley H (2008) Orienta-
Steinmetzer T, Böttcher-Friebertsh€auser E tion of the monomeric porin OmpG in planar
(2015) Influenza virus activating host lipid bilayers. Chembiochem 9:3029–3036
Chapter 10
Oligonucleotide-Directed Protein Threading Through a Rigid

Nanopore
Garbiñe Celaya and David Rodriguez-Larrea
Abstract
Nanopore technology enables the detection and analysis of single protein molecules. The technique
measures the ionic current passing through a single pore inserted in an electrically insulating membrane.
The translocation of the protein molecule through the pore causes a modulation of the ionic current.
Analysis of the ionic current reveals the biophysics of co-translocational unfolding and may be used to infer
the amino acid sequence and posttranslational modifications of the molecule.
Key words Nanopore, Single-molecule, Protein, Sequencing, Co-translocational unfolding
1 Introduction
Nanopore technology measures the ionic current flowing through a

single pore made in an insulating membrane [1]. Typically, the
membrane separates two compartments filled with an electrolyte
solution and an electrical potential is applied across the membrane.
When a molecule interacts with the pore it modulates the flow of
ions through the pore, and this effect can be readily measured
within the pA range. The approach shows both high temporal
resolution (typically <1 ms) and high dynamic range (changes
<0.5% in the open pore current can be measured with high
confidence).
These features, together with the ease of parallelization
(hundreds of pores have been arrayed in a chip the size of a
fingernail), justify an ample research field aiming to develop
nanopore-based applications [2]. Between them, next-generation
DNA sequencing is the best-known example. A single strand of
DNA slowly moving step by step through the pore causes a city-
landscape like ionic current profile that can be translated into the
oligonucleotide sequence [3, 4]. While some improvement in the
base call accuracy is still needed, it offers several promising
135
136 Garbiñe Celaya and David Rodriguez-Larrea
advantages: long read lengths, access to epigenetics, high sequenc-

ing speed, and low sequencing costs, all combined in devices the
size of a chocolate bar [2]. In addition, nanopores either made with
solid materials or with biological components can be coupled with
single-molecule DNA labeling for multiplexed detection of anti-
bodies or DNA-binding proteins [5, 6]. These systems hold enor-
mous potential for future diagnostics and drug-screening
applications. These applications are important, but it is in the
analysis of proteins where nanopores hold more potential to revo-
lutionize basic research. Although state-of-the-art methods to
characterize proteomes have proven powerful [7], they contain
significant limitations: low populated species remain unidentified,
limited quantification of abundance, statistical black-boxes in data
analysis procedures, and problems reconstructing original mole-
cules from the analysis of their fragments (a particularly challenging
scenario when multiple posttranslational modifications are possible
within the same molecule).
Nanopores could potentially sequence full-length single-pro-
tein molecules, including posttranslational modifications and quan-
tifying each of the species [8, 9]. Nonetheless, the pursue of this
challenge can well be harder than DNA sequencing: proteins tend
to be folded (which implies that they typically do not fit through
the narrow pores used for sequencing), weakly charged, and show
20 different amino acids plus a wide variety of posttranslational
modifications. Several approaches have shown that it is possible to
gather information about a protein in its folded state using nano-
pores [10] such as its shape or size but it seems likely that true
sequencing can only be achieved using narrow pores. For that the
protein needs to be targeted toward the pore and unfolded in an
active manner. Nivala and Akeson [11] have efficiently targeted and
translocated multidomain proteins through alpha-hemolysin nano-
pores using a SsrA tag on the C-terminus of the protein substrate
and ClpXP molecular motors placed on the trans side of the mem-
brane. Rodriguez-Larrea and Bayley [12, 13] instead use an oligo-
nucleotide tag on either terminus of a single-domain protein and
drive unfolding and translocation with the electric potential. With
this last approach, it has been proven the feasibility of characterizing
posttranslational modifications on proteins with nanopores [8] and
here, in this chapter, we will cover the experimental details of such
methodology.
2 Materials
2.1 For Purification 1. A pT7 plasmid containing the WT α-Hemolysin (α-HL) from
of α-Hemolysin and Its S. aureus gene.
Heptamers 2. BL21 (DE3) E. coli cells.
Protein Threading with Nanopores 137
3. Luria-Bertani media.
4. Ampicillin (0.1 g/mL) (in pure water).
5. 1 M IPTG (in pure water).
6. 5–15% gradient SDS-PAGE gels.
7. Buffer A: 30 mM Tris–HCl, 50 mM EDTA, pH 8.0.
8. Ammonium sulfate (powder, at least 100 g).
9. Buffer B: 20 mM NaCl, 10 mM sodium acetate, pH 5.0.
10. Anticoagulant: 80 mM citrate, 140 mM dextrose, pH 7.4.
11. PBS: 150 mM sodium chloride, 10 mM sodium phosphate,
pH 7.4.
12. Hypotonic solution (5 mM HEPES, pH 7.4).
13. Buffer C (2 M KCl, 10 mM HEPES, pH 7.4).
14. Buffer D (2 M NaCl, 10 mM HEPES, pH 7.4).
15. 10% SDS solution in pure water.
16. Size exclusion column (we use Superdex 200 16/60 from GE
Healthcare).
2.2 For Generation of 1. Protein of interest with a C-terminal cysteine. One option is to
Protein- choose a protein that naturally has a cysteine group on its
Oligonucleotide C-terminus. Alternatively, express a C-terminal cysteine recom-
Complexes for Single- binant version of the protein of interest.
Molecule Analysis with 2. Oligonucleotide at least 30 nucleotides long, without second-
Nanopores ary structure and with either a 50 or 30 terminal maleimide.
3. Tris(2-carboxyethyl)phosphine (TCEP).
4. Buffer exchange columns (1 mL).
5. Exchange buffer: 10 mM HEPES, pH 7.
6. Elution buffer: 10 mM HEPES, 1 M KCl, pH 7.
7. Anion exchange column (we use monoQ FF from GE
Healthcare).
2.3 For Single- 1. We use an Axotpatch 200B equipped with a Digidata 1440
Molecule (Molecular Devices). We have also tested cheaper and versatile
Measurements new models as is the case of the e4 amplifier from Elements,
which can also do the job.
2. Hexadecane dissolved in pentane (1% v:v).
3. 1,2-Diphytanoyl-sn-glycero-3-phosphocholine (DPhPC) dis-
solved in pentane (5 mg/mL).
4. Pure ethanol (100%).
5. A Teflon film 0.025 mm thick.
6. A needle.
7. A piezoelectric igniter.
8. Glass capillaries (20 μL volume).

9. Glass syringe (20 μL volume).
10. 1 mm thick silver wire.
11. Hypochlorite solution (20% v:v in pure water).
12. Copper wires and gold-plated pins.
13. A chamber made of an electric insulating material (Delrin is
recommended).
14. Recording buffer: This is typically an electrolyte solution con-
taining chloride ions. We frequently use 10 mM HEPES, 2 M
KCl, pH 7.
15. 2% Agarose in 10 mM HEPES, 2 M KCl, pH 7.
16. Silicone glue (optional).
17. Faraday cage.
3 Methods
3.1 Purification of 1. Transform the BL21 cells with the plasmid containing the
Monomeric Hemolysin α-hemolysin gene and plate in LB-agar supplemented with
0.1 mg/mL ampicillin.
2. The day after, grow single colonies in 25 mL of Luria-Bertani
(LB) media supplemented with 0.1 mg/mL ampicillin over-
night under shaking at 25 C.
3. Optional: Take 500 μL of cells and mix with 500 μL of glycerol,
mix thoroughly and store at 80 C. This glycerol stock can be
used as starting point in future purifications.
4. Centrifuge the grown media at 10,000 g for 10 min at 4 C
and resuspend the pellets in 0.5 L of fresh LB supplemented
with 0.1 mg/mL ampicillin. Grow with shaking at 25 C for
2 h.
5. Optional: Take a small aliquot of the culture before induction.
Induce the media with IPTG 1 mM final concentration. Grow
for 4 h at 25 C (see Note 1).
6. Centrifuge the media at 10,000 g for 10 min and resuspend
the pellet of cells in 30 mL of buffer A.
7. Lyse the cells by sonication while immersed in ice. This can be
done with 10 cycles of 30 s sonication followed by 30 s without
sonication.
8. Centrifuge the lysate for 45 min at 10,000 g and 4 C.
9. Keep the supernatant and add enough ammonium sulfate to
reach 75% saturation while stirring at 4 C for 1 h (see Note 2).
10. Centrifuge for 50 min at 40,000 g and 4 C. Keep the pellet
as most of the protein should be found there.
11. Resuspend the pellet in 5 mL of buffer B and dialyze against

the same buffer to remove ammonium sulfate overnight at
4 C.
12. Centrifuge the solution at 40,000 g and 4 C and load the
supernatant in a Superdex 200 16/60 previously equilibrated
in buffer B (see Note 3).
13. Run 5–15% SDS-PAGE from the different fractions to identify
those containing the monomeric hemolysin.
14. Store the fractions containing the monomeric hemolysin at
20 C (see Note 4).
3.2 Preparation of 1. Mix 100 mL of fresh blood in a container with ice-cooled

Rabbit Erythrocyte 50 mL of anticoagulant. Keep the mixture on ice (see Note 5).
Membranes 2. Centrifuge at 500 g for 10 min at 4 C. Discard the plasma
and white blood cell fraction with a pipette to purify the red
blood cell fraction. Add an equal volume of PBS and repeat
three more times.
3. Erythrocyte lysis by hypotonic shock: Add dropwise the pur-
ified red blood cell fraction to 2 L of hypotonic solution at 4 C
with vigorous stirring (see Note 6).
4. Centrifuge at 20,000 g for 10 min at 4 C. Keep the pellet.
5. Resuspend the pellet of membranes in 5 mL of ice-cold hypo-
tonic solution and centrifuge at 20,000 g for 10 min at 4 C,
keep the pellet and repeat three further times.
6. This step is optional but advisable to remove electrostatically
bound proteins. Centrifuge at 20,000 g for 10 min at 4 C
and resuspend the membrane pellet in 2 mL of buffer C. Wait
1 min and centrifuge at 20,000 g for 10 min at 4 C. Repeat
three more times.
7. Resuspend the pellet in 2 mL of buffer C, aliquot, and store at
20 C.
3.3 Preparation of 1. Mix 50 μL of purified rabbit erythrocyte membranes with

Heptameric α- 250 μL of monomeric protein (typically 0.3 mg/mL) at room
Hemolysin temperature. Incubate for 5 min.
2. Centrifuge at 20,000 g for 10 min and discard (or keep for
future use) the supernatant containing monomeric hemolysin.
Resuspend the pellet in 250 μL of buffer D and centrifuge at
20,000 g for 10 min. Discard the supernatant and repeat
three more times.
3. Centrifuge at 20,000 g for 10 min and resuspend the pellet
in 50 μL of buffer D supplemented with SDS to a final concen-
tration of 0.1% (w/v) to extract the heptamers (see Note 7).
4. Centrifuge at 20,000 g for 10 min and keep the supernatant.
Store at 20 C (see Note 8).
3.4 Generation of 1. Mix 2 mg of the protein of interest with 10 mM of a reducing

Protein- agent overnight at 4 C (TCEP is advised because it is less
Oligonucleotide reactive against maleimide).
Complexes for Single- 2. Remove the excess TCEP from the protein solution using
Molecule Analysis with buffer exchange columns.
Nanopores 3. Dissolve 0.1–0.5 mg of the lyophilized maleimide-
oligonucleotide with the solution containing the reduced pro-
tein (see Note 9). Incubate at room temperature for 2 h.
4. Purify the reaction in an anion exchange column with a salt
gradient 0 to 1 M KCl using exchange buffer and elution buffer
(a gradient length of 120 mL is recommended).
5. Identify with SDS-PAGE the fractions containing oligonucleo-
tide-modified protein (Fig. 1b). Since KCl concentrations
higher than 100 mM may precipitate the SDS, dilute your
fractions with water in a 1:5 v/v ratio before running through
SDS-PAGE. Use unmodified protein as a negative control
reference. The oligonucleotide-modified protein will run
slower in SDS-PAGE.
6. Determine the concentration by measuring the absorbance at
260 nm and using the calculated molar extinction coefficient of
the oligonucleotide. Pool the fractions with higher
a b
protein
Ag/AgCl electrodes
ground electrode
trans cis
PTFE film
oligonucleotide
linker
20 seconds
Fig. 1 Setup of single-molecule experiment. (a) Side view of the chambers we use for the single-molecule
experiments. Cis and trans compartments are separated by a PTFE film with a 100–50 μm hole where the
membrane is built. (b) Representation of the protein-oligonucleotide construct used for single-molecule
measurements. Different linkers can be used (maleimide is described in the protocol). (c) 20 s of the ionic
current recording obtained with the protocol described here showing multiple single-molecule events of
unfolding and translocation. The signal was filtered at 5 kHz (low-pass Bessel filter)
concentration. Final concentration of oligonucleotide-

modified protein should be above 1 μM. Store protein for
long-term use at the appropriate storage conditions. This
must be determined empirically for each protein of interest.
3.5 Single-Molecule 1. Immerse 2 pieces, each 2 cm long, of silver wire into the
Measurements hypochlorite solution. Leave them overnight at room temper-
ature (they should look dark gray).
3.5.1 Preparation of the
Silver/Silver Chloride 2. Heat 1 mL of 2% agarose solution and pipette 200 μL. Remove
Electrodes the tip without expelling the solution and immerse 1 cm the
electrodes (the other half should be outside the tip).
3. Connect the electrodes to the female pin of a wire functiona-
lized at both ends with gold-plated pins (male and female
respectively).
3.5.2 Preparation of the 1. Cut a 2 2 cm square-shaped Teflon piece (see Fig. 1a for a side
aperture view representation of our chambers).
2. Touch the center of the Teflon film with a sharp needle. Do not
perforate it, just a gentle touch.
3. Place the film between the electrodes of the piezoelectric
igniter and discharge several times (ten times is usually
enough). Confirm with a microscope that there is a perfect
round-shaped hole. If not, prepare a new piece of Teflon.
4. Place the film between the two compartments of the chamber.
Ensure that there is no water leakage. Sealing with silicone glue
may be beneficial.
5. Rinse with pure ethanol. Remove ethanol with a gentle
gas-source of N2.
3.5.3 Preparation of the 1. Immerse the glass capillary into the hexadecane and allow the
Lipid Bilayer solution to rise through the capillary.
2. Gently touch the Teflon film with the capillary at both sides. Be
sure that some solution is spread through the film. Wait 30 s for
the pentane to evaporate.
3. Add appropriate volume of recording buffer into both com-
partments. Move the chamber to the Faraday cage, immerse
one electrode in each compartment, and connect the wires to
the amplifier (to the headstage if using the Axopatch 200B).
4. Pipette up the solution in each compartment and release it back
rinsing the Teflon gently (see Note 10).
5. Place two drops (10 μL) of DPhPC solution onto the surface of
each compartment. Wait 2 min for the pentane to evaporate.
6. Repeatedly and slowly pipette up and down the solution across
the Teflon aperture in both compartments. When a lipid bilayer
is formed, no ionic current is detected.
3.5.4 Testing the Bilayer 1. Assess the quality of the membrane by applying a voltage
Formation and Stability triangle wave: a 20 mV voltage ramp going up in 20 ms fol-
lowed a 20 mV voltage ramp going down in 20 ms. This cycle is
repeated continuously, and the capacitive current is measured.
The magnitude of the capacitance depends on the membrane
surface but also on the particularities of the chamber. In any
case, a good membrane shows a city-scape like current profile.
It should also break when zapped.
3.5.5 Testing the 1. Set a constant voltage bias of +100 mV across the bilayer and
Nanopore Activity add α-hemolysin heptamers to the cis side (ground). A charac-
teristic single α-hemolysin insertion will cause a stepped
increase in the ionic current of 180–200 pA.
2. Following the first insertion, perfuse the cis chamber with at
least 10 volumes of fresh buffer to dilute the α-hemolysin
concentration in the chamber to prevent multiple pore
insertions.
3. Add 50 μL of the oligonucleotide-protein complex to the cis
chamber and mix thoroughly. Single-molecule unfolding and
translocation events are observable within seconds after the
addition of protein with a voltage bias of 100 mV (Fig. 1c)
(see Notes 11 and 12).
4. If no translocation events are seen even with a stable
α-hemolysin heptamer inserted into a stable bilayer, the protein
of interest may be difficult to unfold through the α-hemolysin
nanopore (see Note 13).
4 Notes
1. We recommend that you run samples before and after IPTG

induction on SDS-PAGE to test protein expression efficiency.
Some colonies produce larger amounts of protein than others,
so it is also useful to do this test on several colonies.
2. Ammonium sulfate is better dissolved if added by a small
amount at a time.
3. When using a size exclusion chromatography, check maximum
loading volumes and never exceed this limit. Instead, load the
sample in several chromatography runs.
4. After size-exclusion, the protein may not be of a high purity
(>90%). This is not a problem as heptamerization in rabbit red
blood cells can still be carried out efficiently. The heptamers
become more and more purified during the wash steps with
rabbit erythrocyte membranes.
5. For high-quality erythrocyte membranes, it is important to

carry out the whole procedure in the hours following blood
collection and to work constantly with ice-cooled solutions.
6. For efficient lysis it is advised to use 1:1000 (v/v) ratio between
the purified red blood cell fraction and the hypotonic solution,
but we have successfully used a 1:200 (v/v) ratio.
7. When extracting the heptamers from the erythrocyte mem-
brane, do not use KCl salt as it may precipitate the SDS.
8. This procedure yields a solution with a high concentration of
heptamers. It is advisable to experimentally determine an
appropriate dilution of the stock such that nanopore insertions
do not occur too frequently (for us a 1:100 dilution is
acceptable).
9. The oligonucleotide has to be more than 23 nucleotides long
to unfold and translocate the protein substrate.
10. You can proceed to the next step when there is no ionic current
between the compartments when the solution is removed but
it does occur when the solution is placed back (this should be
true for both compartments).
11. Voltage biases smaller than 70 mV do not produce unfolding
and translocation events. We typically work in the 70–140 mV
range.
12. Due to the importance of the voltage bias magnitude, be sure
that the electrodes are properly balanced.
13. We have used our approach with unknotted proteins lacking
disulfide bonds as big as 30 kDa. Bigger proteins, knotted
proteins, or proteins containing disulfide bonds may not be
easily unfolded and translocated through the nanopore. These
proteins may need an optimized protocol or engineering by
mutagenesis.
Acknowledgements
D.R.-L. is a recipient of a Ramón y Cajal Fellowship (RYC-2013-

12799). D.R.-L. was funded by MINECO grants BIO2017-
88946-R and BFU2016-81754-ERC (FEDER funds).
References
1. Maglia G, Heron AJ, Stoddart D, Japrung D, 3. Bayley H (2006) Sequencing single molecules
Bayley H (2010) Analysis of single nucleic acid of DNA. Curr Opin Chem Biol 10:628–637
molecules with protein nanopores. Methods 4. Branton D, Deamer DW, Marziali A, Bayley H,
Enzymol 475:591–623 Benner SA, Butler T, Di Ventra M, Garaj S,
2. Check Hayden E (2015) Pint-sized DNA Hibbs A, Huang X, Jovanovich SB, Krstic PS,
sequencer impresses first users. Nature Lindsay S, Ling XS, Mastrangelo CH,
521:15–16 Meller A, Oliver JS, Pershin YV, Ramsey JM,
Riehn R, Soni GV, Tabard-Cossa V, 9. Dong Z, Kennedy E, Hokmabadi M, Timp G

Wanunu M, Wiggin M, Schloss JA (2008) (2017) Discriminating residue substitutions in
The potential and challenges of nanopore a single protein molecule using a
sequencing. Nat Biotechnol 26:1146–1153 sub-nanopore. ACS Nano 11:5440–5452
5. Bell NA, Keyser UF (2016) Digitally encoded 10. Yusko EC, Johnson JM, Majd S, Prangkio P,
DNA nanostructures for multiplexed, single- Rollings RC, Li J, Yang J, Mayer M (2011)
molecule protein sensing with nanopores. Nat Controlling protein translocation through
Nanotechnol 11:645–651 nanopores with bio-inspired fluid walls. Nat
6. Celaya G, Perales-Calvo J, Muga A, Moro F, Nanotechnol 6:253–260
Rodriguez-Larrea D (2017) Label-free, multi- 11. Nivala J, Marks DB, Akeson M (2013)
plexed, single-molecule analysis of protein- Unfoldase-mediated protein translocation
DNA complexes with nanopores. ACS Nano through an α-hemolysin nanopore. Nat Bio-
11:5815–5825 technol 31:247–250
7. Huttlin EL, Jedrychowski MP, Elias JE, 12. Rodriguez-Larrea D, Bayley H (2013) Multi-
Goswami T, Rad R, Beausoleil SA, Villén J, step protein unfoding during nanopore trans-
Haas W, Sowa ME, Gygi SP (2010) A tissue- location. Nat Nanotechnol 8:288–295
specific atlas of mouse protein phosphorylation 13. Rodriguez-Larrea D, Bayley H (2014) Protein
and expression. Cell 143:1174–1189 co-translocational unfolding depends on the
8. Rosen CB, Rodriguez-Larrea D, Bayley H direction of pulling. Nat Commun 5:4841
(2014) Single-molecule site-specific detection
of protein phosphorylation with a nanopore.
Nat Biotechnol 32:179–181
Chapter 11
Unfolding and Translocation of Proteins Through

an Alpha-Hemolysin Nanopore by ClpXP
Jeff Nivala, Logan Mulroney, Qing Luan, Robin Abu-Shumays,
and Mark Akeson
Abstract
Proteins present a significant challenge for nanopore-based sequence analysis. This is partly due to their
stable tertiary structures that must be unfolded for linear translocation, and the absence of regular charge
density. To address these challenges, here we describe how ClpXP, an ATP-dependent protein unfoldase,
can be harnessed to unfold and processively translocate multi-domain protein substrates through an alpha-
hemolysin nanopore sensor. This process results in ionic current patterns that are diagnostic of protein
sequence and structure at the single-molecule level.
Key words Nanopore, Protein folding, Protein sequencing, Single-molecule, Unfoldase,

α-Hemolysin, ClpXP
1 Introduction
Traditional protein analysis methods (e.g., mass spectrometry and

antibody recognition) can achieve high detection specificity and
sensitivity, yet single-molecule resolution is difficult or impossible
to achieve with such technologies [1]. This limitation prevents
complete characterization of complex proteomic samples. Nano-
pore-based analytical methods are a promising approach currently
being explored to characterize protein variation with greater reso-
lution [2, 3]. The potential of nanopores for the analysis of intact,
full-length protein strands is particularly attractive. However,
unfolding and processive translocation of full-length proteins is
difficult to achieve with voltage-mediated force alone. We detail
here a protocol that uses a molecular motor (ClpXP) to unfold and
translocate suitably tagged proteins through an alpha hemolysin
(αHL) nanopore [4–7]. We have shown that this method is sensi-
tive to protein sequence, length, unfolding kinetics, and additional
structure-dependent features that enable discrimination among
145
146 Jeff Nivala et al.
unique protein domains at the single-molecule level using machine

learning tools [7].
An essential requirement of this protocol is a single-channel
nanopore platform. We include here a description of the custom
nanopore setup we have used to successfully conduct these experi-
ments (and has been described previously [8]), but other nanopore
platforms should also be compatible with ClpXP-mediated protein
analysis.
2 Materials
1. Platform, Aperture, and Electrodes: Figure 1 shows the design

of our nanopore platform. The core element is a small, puck-
shaped piece of custom molded Teflon. While the exact dimen-
sions are not critical, the important feature of the puck is the
presence of two wells (cis and trans) that have two small holes
drilled in each of them. The holes permit the insertion of AgCl
electrodes and a U-tube (Cole-Parmer) connecting the two
wells. One end of the U-tube (cis) is where the nanopore
aperture is constructed. The ~20–30 μm aperture is made
with Teflon heat shrink tubing (Cole-Parmer) formed around
a chemically etched needle point. This supports the lipid bilayer
into which an alpha-hemolysin (αHL) protein pore can insert.
1 N NaOH and tungsten needle also needed in the making of
apertures.
2. Stock Lipids: Lipid type is 4ME 16:0 PC (1,2-diphytanoyl-sn-
glycero-3-phosphocholine). Powder comes in a 25 mg/vial.
Fig. 1 The nanopore apparatus. (Figure credit Noah Wilson)

Protein Translocation through Nanopores by ClpXP 147
Add 2.5 mL chloroform to the vial to make a 10 mg/mL stock.

Keep the stock in 20 C freezer (see Note 1).
3. Aperture Coating Reagent: 0.2 mg/mL 4ME 16:0 PC lipid
dissolved in hexane.
4. Aperture preparation reagents: 10% nitric acid in water, glass
petri dish, glass beaker, baking soda solution, syringe,
200 proof ethanol, hexane.
5. 1 Perfusion buffer: 10 mM Hepes pH 7.6, 300 mM KCl,
10 mM MgCl2, 10% glycerol (optional), 1 mM DTT
(optional), 1 mM EDTA (optional).
6. ATP regeneration mix (ARM): 20 mM Hepes pH 7.6,
600 mM KCl, 20 mM MgCl2, 5 mM ATP, 8 mM creatine
phosphate, 0.08 mg/mL creatine phosphokinase, 100 nM
ClpX6, 300 nM ClpP14, 20% glycerol (optional), 2 mM DTT
(optional), 2 mM EDTA (optional).
7. Alpha-hemolysin, ClpX, and ClpP: These proteins can be
expressed and purified from E. coli. This process has been
optimized and described by others [4, 5, 9], so we will not go
into the detailed protocol here. Dissolve ClpP and ClpX prior
in ATP regeneration mix.
8. Substrate proteins: Proteins to be analyzed by this method
need to be tagged with a specific C-terminal sequence that
includes a flexible polyanionic region (including glycine, serine,
glutamic acid, and aspartic acid) and an 11 amino acid ssrA tag
[6, 7]. The former region promotes initial capture and thread-
ing of the protein into the nanopore, while the latter motif
enables recognition of the threaded protein by ClpX. This
tagging can be done genetically [6, 7], or potentially posttran-
slationally with modifications to the design [10].
9. Equipment: A digitizer such as the Digidata 1440 series; a
telegraphed instrument such as the Axopatch 200B;
microscope.
3 Methods
3.1 Aperture 1. Cut a small length (8–10 cm) of heat-resistant tubing for the
and Electrode U-tube and place a small piece of shrink-wrap tubing on
Construction one end.
2. Repeat step 1 until you have 15–20 assembled tubes.
3. Immerse one end of a tungsten needle in a bath of 1 N sodium
hydroxide.
4. Apply a 20 V voltage across the bath which will begin to etch
the tungsten needle. Turn off the voltage once the needle is
molecularly thin (see Note 2).
5. Carefully thread the needle through the heat-resistant tubing

into the shrink-wrap tubing. Do not introduce dust into the
shrink-wrap tubing, as that will alter the aperture. Slowly pull
through a heating element (similar to a blow dryer, though a
flame can also be used) to affix the shrink wrap tubing around
the point of the needle. Keep heating until the needle is sealed
in the shrink-wrap tubing.
6. Pull out the needle in one motion and examine the tubing
under a microscope.
7. Repeat steps 5 and 6 for the remaining preassembled tubes.
8. While viewing the tubing under the microscope, slice it with a
clean razor blade so that the aperture formed by the needle is of
the desired diameter. Use a 15–25 μm thick wire to gauge the
diameter of the aperture (see Notes 3 and 4).
9. If the platform is not level, or the aperture too big, repeat step
8 until you have an aperture that you are satisfied with. Store
the remaining uncut apertures for later use if the one you’re
using begins to fail.
10. Insert the U-tube into a manufactured Teflon puck by thread-
ing the trans side (larger opening, non-sliced side) of the
U-tube through the top of one of the wells of the puck.
Make sure that there is a good seal between the tubing and
the hole of the puck, or the buffer will leak out of the well (see
Note 5).
11. Make a smooth bend in the tubing and thread it through the
bottom of the other well. Cut the trans side of the U-tube so
that the opening is below the top of the well. Again, make sure
there is a good seal (see Notes 5 and 6).
12. To assemble the electrodes, connect a wire between a gold pin
and the Ag or AgCl electrodes (see Note 7).
13. Wrap the wire with shrink-wrap tubing and slowly rotate the
electrodes to create an even seal around the electrode-wire
connection and down the length of the wire.
14. Ensure the electrodes make a tight seal in the Teflon puck, or
buffer will leak out (see Note 8).
15. The aperture and electrodes created here are reusable for many
experiments, and only need to be replaced once they fail. Store
any extra uncut aperture tubes and the tungsten needle for later
use if necessary. Only re-etch the needle if all the apertures are
too large, or the needle broke.
3.2 Aperture 1. To extensively remove residual clogging lipid, boil the aperture
Preparation in 50 mL 10% nitric acid in a 100 mL glass beaker covered with
a glass petri dish on a heating block for 5 min (see Note 9).
2. Allow to cool for 10–15 min, until the beaker is no longer too
hot to hold. To prevent the aperture from falling out of the
beaker, pour the nitric acid out into a container with water and
baking soda and with the glass petri dish cover the top of the
beaker.
3. Wash the aperture with plenty of DDI water. With the glass
petri dish cover the top of the beaker to prevent the aperture
from falling out, pour the wash water into the container with
baking soda.
4. Repeat washing 3 times or until the wash water does not react
(foam) when added to the baking soda water bath.
5. Plug the electrode hole of the cis side (see Fig. 1 for distinguish-
ing the cis and trans side) with a dead-end pipette tip to avoid
leakage.
6. Add DDI water to the cis side. Using a syringe with rubber
tubing attached via a Luer lock, pull the water through the
U-tube from the trans side to get water through the aperture.
7. Repeat washing step 6 at least 3 times.
8. Repeat washing step 6 with 200 proof ethanol.
9. Repeat washing step 6 with hexane.
10. Aspirate the surface of the aperture puck to dry.
11. Connect a syringe to the trans side and pull the plunger to the
maximum volume. Wedge a large pipette tip to keep the
plunger in position to maintain vacuum to completely dry the
aperture and U-tube.
3.3 Preparation 1. Use glass capillary pipette to add 20 μL of the stock 10 mg/mL
of Aperture Coating lipid/chloroform solution in a glass vial. Keep in a desiccator
Reagent under vacuum until the chloroform has completely dried out.
The lipid will appear as a tiny nearly clear round pad at the
bottom of the vial. It takes about an hour to completely dry (see
Note 10).
2. Remove a lipid vial from the desiccator right before starting an
experiment. Use glass capillary pipette to add 100 μL hexane to
the vial to dissolve the lipid. Cap the vial with a sealing cap.
Make sure it seals well. Shake the vial gently until the lipid pad
is completely dissolved into the hexane (see Notes 11 and 12).
3.4 Aperture 1. Put the aperture into a temperature controlling station. cis end
Assembly facing yourself and trans end facing the head stage.
2. Connect the electrodes. Ensure the electrodes make a good seal
with the puck buffer compartments by filling each well with
buffer and checking that no buffer leaks out around the
electrodes.
3. Add 1 μL lipid solution to the cis surface and quickly pull the
coating solution through the aperture by suction with the
syringe (see Note 13).
4. Repeat 3–4 times to finish the lipid coat.
5. Add buffer with ClpXP and ATP-regeneration mix to both
wells. Check and remove trapped air bubbles. Connect a
syringe filled with buffer with ClpXP and ATP-regeneration
mix to the trans end and push buffer through the U-tube into
the cis compartment. Air bubbles will come out first and then
buffer should visibly squirt out of the aperture (see Note 14).
6. Use Kimwipe to carefully clean the excess buffer around the
aperture and add fresh buffer to make a meniscus over the
outer surface of each well. At this point, the Axopatch current
should read overloaded, indicating a completed circuit.
7. Fill the perfusion system with 1 perfusion buffer.
3.5 Lipid Patch 1. Roll 3–5 pieces of tape into a barrel-shape to allow double-
Preparation sided taping. Tape 3–5 microscope glass coverslips into a
petri dish.
2. Add ~0.5 μL the stock 10 mg/mL lipid/chloroform solution
on the taped cover glass. Make 4–5 such patches on the corners
of each glass coverslips.
3. Allow the chloroform to completely air dry, leaving a small
circle of dried lipid. When not using the lipid drops, keep the
petri dish covered to avoid dust contaminating the lipid drops
(see Note 15).
3.6 Lipid Bilayer 1. Add 1 μL hexadecane to a glass coverslip with no lipid drops.
Formation Coat the paintbrush hair in hexadecane by dipping into the
solvent and then rub the solvent into one of the previously
spotted and dried lipid drops. Using the brush, gently scrape
the lipid from the glass coverslip surface and roll the lipid into a
ball. The lipid will not seem to move at the beginning, so be
very patient. Start from the edge and add solvent slowly as
needed. Adjust the stiffness and stickiness of the lipid ball by
adding different amounts of solvent.
2. Pick up the lipid ball with the brush bristle and roll the lipid
around the outer edge of the aperture on the cis end surface. Be
careful not to directly block the aperture with the lipid ball. The
current should remain overloaded while painting the lipid ball
onto the cis surface.
3. When an adequate amount of lipid (see Note 16) has been
applied to the cis surface, use a pipette to blow an air bubble
over the aperture to form a lipid bilayer over the aperture. If a
lipid bilayer forms, the current will read zero or close to zero
(up to ~3 pA).
4. Ensure that a true bilayer and not a lipid clog has formed over
the aperture by zapping the bilayer with rapid high voltage. A
good bilayer will be broken by zapping, while a clog will not.
5. The proper lipid consistency can be determined by repeating
steps 3 and 4 several times. If several consecutive bilayers
rupture when zapped, the lipid is the proper consistency and
a suitable bilayer has been formed with reasonable certainty (see
Notes 17 and 18)
3.7 Single-Pore 1. Add 1 μL of αHL protein (~1 mg/mL stock) to the cis solution
Insertion when a suitable bilayer is formed. Turn on the voltage and wait
for a single pore to incorporate into the bilayer, indicated by a
sudden and discrete increase in current (see Note 19). The
conductance of a single pore is dependent on the temperature,
salt concentration, voltage, and pore protein characteristics. In
standard conditions for these experiments, the single channel
conductance is ~54 pA, with an RMS of <5 pA (see Notes 20–
24).
2. When a good single pore is inserted, turn off the voltage
(turning off the voltage helps to prevent a second channel
inserting) and perfuse the cis solution with 3–5 mL of fresh
buffer (that does not contain ClpXP or ATP) to remove excess
pore protein. After perfusion, turn the voltage back on to
confirm the channel is still present (see Note 25).
3. Substrate proteins can now be added to the cis well at a final
concentration of ~1 μM, and data collection can begin. A
Faraday cage can also now be added to enclose the nanopore
setup to reduce signal noise (see Note 26).
Protocol:
Acquisition mode: Gap-free.
Trial length: Duration: 10 min.
Sampling: sampling rate per signal (Hz) 10,000 ¼ 100 μs.
3.8 Cleanup 1. When the experiment is finished (typically following rupture of

the bilayer), use a Kimwipe to remove all the buffer from both
wells of the puck.
2. Pull out both electrodes and rinse with DDI water to remove
trace buffer.
3. Remove puck from the temperature controller station. If stuck
too tight, turn the temperature to 9.9 C and let the station
cool down. After a while (~10 min), the cooled station will
shrink enough to remove the puck.
4. Remove any remaining buffer from the perfusion system and
rinse the syringes with water twice. Clean the puck and aper-
ture as detailed in Subheading 3.1.
4 Notes
1. Make sure the chloroform is water-free and avoid contact with

plastic wares, including plastic graduated cylinder, falcon tubes,
plastic pipettes, pipette tips, and parafilm (do not use parafilm
to wrap the cap for sealing purpose). Use glass pipettes or glass
graduated cylinder to pipette. If the cap of stock bottle is
plastic, make sure to use ones with foil or Teflon lined. Keep
the lipid stock at 20 C. When the lipid stock is removed from
20 C, wait until it warms to room temperature before open-
ing the cap to avoid condensation building up in the vial.
2. After etching the needle, be careful not to break or snap the
end, as it will be very delicate. Store carefully so the etched end
will not break when handled.
3. When slicing the tubing with a razor blade, make sure you slice
straight up and down to create a level cis platform. Also, you
can always make another slice. It is better to be cautious and
slice farther away from the cone formed by the needle, than to
make too large of an aperture.
4. You can check if you cut an aperture from the tubing by
connecting a syringe to the trans side of the U-tube and try
to push water through the aperture. If you are able to get a
stream of water to come out of the cis end, you have success-
fully formed an aperture. If no water is able to flow through,
you need to remove all the water and slice it again.
5. If the hole in the well is too large, you can use parafilm to wrap
around the tubing to increase the diameter to make a tight seal.
If the hole in the well is too small, you can use a hard metal file
to widen the hole by twisting the file in the hole to sand away
some of the Teflon.
6. Make sure there are no kinks in the tubing as you bend it to fit
between the two wells. If there is a kink, it will make pushing
buffer through the tubing much harder and increases the like-
lihood of introducing air bubbles that can break the circuit.
7. The trans electrode should be as short as possible, but still
make a good connection into the puck and the headstage.
More noise will be introduced to the experiment the longer
that electrode is.
8. If there is not a tight enough seal, you can wrap the ends of the
electrodes in parafilm and wedge it into the well holes.
9. 10% Nitric acid boils at 100 C, but heat the surface to
300–400 C to accelerate boiling.
10. Lipid vials last for 1 week. Typically, six are made at a time and
used throughout the week.
11. Do not use if the lipid has been dissolved for more than 3 h or
the volume decreases significantly, which occurs if the tube
remains uncapped (exposed to open air) for extended periods.
If higher concentration lipid solution is used for coating, the
aperture is more likely to get clogged.
12. The vials of lipid will be good in the desiccator for up to a week.
13. Good coating will not produce visible excessive lipid on the
aperture surface.
14. Make fresh buffer daily. Preferably, do not dilute from stock
solutions. Channel insertions seem less likely with >1 day old
buffers.
15. The dried lipid drops are good for up to 1 week.
16. Air does not stick to Teflon well. When blowing air bubbles
with a pipette over the cis end surface without lipid, the air
bubble will deflect and not stay or adhere. When an adequate
amount of lipid has been applied to the cis end surface, the air
bubble will adhere to the surface and not float away.
17. If clogged, a minor lipid clog can be unclogged by removing
visible excess lipid on top of the aperture with a pipette tip or
paintbrush and zapping often. If the clog persists, connect a
syringe with buffer to the trans end and try to push buffer
through aperture. If the clog is cleared, buffer will flow out of
the aperture and the current will return to being overloaded. If
it is too hard to push buffer through, insert the paintbrush into
the aperture to try to physically remove the excess lipid. If this
does not clear the clog, the aperture needs to be cleaned as
detailed in Subheading 3.8.
18. A suitable bilayer blocks the current completely and stably.
Occasionally, poorly sealed bilayers will form, which are evident
from incomplete current blockage that may increase over time.
These are signs of a leaking bilayer. Reform the bilayer with or
without zapping and see if stable bilayer can be recovered. If
not, it is a sign of insufficient lipid. Apply the lipid brush again
or roll another lipid ball to supplement existing lipid coating
around aperture.
19. If a pore insertion does not happen for a while (~15 min), there
are several strategies that may increase the likelihood of a pore
to insert into the lipid bilayer:
(a) Zap the bilayer and reform it quickly. This runs the risk,
however, of pore protein inserting from the trans side as a
“backward” insertion (the channel conductance will be
the expected but negative value upon reversing the
applied voltage polarity).
(b) Reform the bilayer without zapping. This may bring pore
stuck in the excess lipid on the edges of the aperture into
the aperture. It runs the risk of the reformed bilayer

breaking and then it carries the same issues as zapping
and reforming.
(c) Quickly flip the voltage polarity back-and-forth several
times. Any perturbation to the bilayer runs the risk of
breaking it and allowing pore protein to insert from the
trans compartment.
(d) Increase the voltage above 180 mV to thin the bilayer and
make insertions more likely. This can also break the bilayer
due to the added force being applied to it.
(e) If no insertion events are seen, add additional pore pro-
tein. Adding too much channel protein will make multiple
channel insertions more probable, so be cautious when
adding. If multiple insertions are observed (they occur as
integer multiples of the single channel conductance
value), it is a sign of channel activity. Reform with or
without zapping until single channel is captured.
20. Occasionally, the pore will gate, which appears as a current
dropping close to zero. If that happens, quickly reverse the
voltage back and forth or turn it off for a moment. This can
ungate the protein pore.
21. If the current drifts drastically, turn off the voltage to allow the
bilayer to rest (~5–10 min). If still not fixable, try to reform a
new bilayer and pore.
22. Protein pores are typically asymmetric, meaning the mean
current amplitude under positive and negative voltage will be
different. Symmetric or non-integer multiple currents may
suggest bad bilayer. Backward channels are sometimes cap-
tured, especially if the bilayer has broken after adding protein
pore to the cis well. They are not useful, reform the bilayer to
remove the backward protein pore.
23. Even if the voltage is turned off, the formation of a bilayer can
still be observed (i.e., current level will drop to zero from
overloaded), but protein pore insertion will not be apparent.
Make sure to check that the voltage is on when waiting for an
insertion.
24. The bilayer is thought to be thinner at higher voltages. It is
observed that pore insertions are more likely when there is a
higher applied voltage. However, bilayers are less stable at
higher applied voltages.
25. ClpX activity is optimal at 37 C. But high temperatures make
the bilayer more fragile. Bilayer and channel formation can be
done at 23 C. When a suitable channel is captured, turn off the
voltage and perfuse (to remove ClpXP from the cis solution).
Then the temperature can be raised to 30 C.
26. Fiber optic lighting is a significant noise source; however a

suitable Faraday cage is able to completely block it. Vibration
isolation tables are also helpful in reducing noise; otherwise
walking or other sources of vibration will manifest observable
noise in the pore current signal.
References
1. Restrepo-Pérez L, Joo C, Dekker C (2018) through an α-hemolysin nanopore. Nat Bio-
Paving the way to single-molecule protein technol 31:247–250. https://doi.org/10.
sequencing. Nat Nanotechnol 13:786–796. 1038/nbt.2503
https://doi.org/10.1038/s41565-018-0236- 7. Nivala J, Mulroney L, Li G, Schreiber J, Ake-
6 son M (2014) Discrimination among protein
2. Movileanu L (2009) Interrogating single pro- variants using an unfoldase-coupled nanopore.
teins through nanopores: challenges and ACS Nano 8:12365–12375. https://doi.org/
opportunities. Trends Biotechnol 10.1021/nn5049987
27:333–341. https://doi.org/10.1016/j. 8. Gyarfas B, Olasagasti F, Benner S, Garalde D,
tibtech.2009.02.008 Lieberman KR, Akeson M (2009) Mapping the
3. Oukhaled A, Bacri L, Pastoriza-Gallego M, position of DNA polymerase-bound DNA
Betton JM, Pelta J (2012) Sensing proteins templates in a nanopore at 5 Å resolution.
through nanopores: fundamental to applica- ACS Nano 3:1457–1466. https://doi.org/
tions. ACS Chem Biol 7:1935–1949. https:// 10.1021/nn900303g
doi.org/10.1021/cb300449t 9. Walker B, Krishnasastry M, Zorn L,
4. Aubin-Tam ME, Olivares AO, Sauer RT, Baker Kasianowicz J, Bayley H (1992) Functional
TA, Lang MJ (2011) Single-molecule protein expression of the a-hemolysin of Staphylococ-
unfolding and translocation by an ATP-fueled cus aureus in intact Escherichia coli and in cell
proteolytic machine. Cell 145:257–267. lysates. Deletion of five C-terminal amino acids
https://doi.org/10.1016/j.cell.2011.03.036 selectively impairs hemolytic activity. J Biol
5. Maillard RA, Chistol G, Sen M, Righini M, Chem 267:10902–10909. https://doi.org/
Tan J, Kaiser CM, Hodges C, Martin A, Bus- 10.1002/chem.201405267
tamante C (2011) ClpX(P) generates mechani- 10. Biswas S, Song W, Borges C, Lindsay S, Zhang
cal force to unfold and translocate its protein P (2015) Click addition of a DNA thread to the
substrates. Cell 145:459–469. https://doi. N-termini of peptides for their translocation
org/10.1016/j.cell.2011.04.010 through solid-state nanopores. ACS Nano
6. Nivala J, Marks DB, Akeson M (2013) 9:9652–9664. https://doi.org/10.1021/
Unfoldase-mediated protein translocation acsnano.5b04984
Part III
Computational Analysis of Nanopore Behavior

Chapter 12
Simulation of pH-Dependent, Loop-Based Membrane

Protein Gating Using Pretzel
Alan Perez-Rathke, Monifa A. V. Fahie, Christina M. Chisholm,
Min Chen, and Jie Liang
Abstract
Bacterial porins often exhibit ion conductance and gating behavior which can be modulated by
pH. However, the underlying control mechanism of gating is often complex, and direct inspection of the
protein structure is generally insufficient for full mechanistic understanding. Here we describe Pretzel, a
computational framework that can effectively model loop-based gating events in membrane proteins. Our
method combines Monte Carlo conformational sampling, structure clustering, ensemble energy evalua-
tion, and a topological gating criterion to model the equilibrium gating state under the pH environment of
interest. We discuss details of applying Pretzel to the porin outer membrane protein G (OmpG).
Key words Computer simulation, Protein loop modeling, Sequential Monte Carlo sampling, Clus-
tering, Protein conformation, Ion channel gating, Bacterial outer membrane proteins, OmpG protein
1 Introduction
Bacterial membrane porins, such as outer membrane protein G

(OmpG) [1], show promise for use as nano-sensor devices [2–
4]. Typically, these devices measure the flow of ionic current
through the porin channel to detect and quantify the presence of
a specific molecular species in solution. However, environmental
factors such as pH can alter the gating behavior of these proteins, as
is the case of OmpG, which exhibits pH-dependent gating
behavior [1].
Here we describe how the computational framework, called
Pretzel (Protein Topology of Zoetic Loops) [5], can be applied
to investigate loop-driven gating behavior of membrane proteins
under different pH conditions. Our method allows atomic-level
study of loop-based conformational gating events that may occur
over time scales not amenable to conventional computational
methods such as molecular dynamics and Brownian dynamics simu-
lations. Specifically, we describe detailed application of Pretzel to
159
160 Alan Perez-Rathke et al.
OmpG, a porin in which loop invasion into the lumen is thought to

be a central component of the observed pH-dependent gating
process [2].
2 Materials
Pretzel is a C++ framework with source code and scripts available

from the public git repository: https://github.com/uic-lianglab/
ompg-public. In addition, several data processing steps (e.g., clus-
tering) employ the SCOBY suite of utilities available at: https://
bitbucket.org/aperezrathke/scoby.
2.1 Additional 1. Linux HPC cluster or workstation, equipped with at least 2 GB

Equipment RAM per computational node.
2. G++ compiler version 4.7 (https://gcc.gnu.org/).
3. CMake software version 4.8 (https://cmake.org/).
4. Protein structure as a PDB atom coordinates file, ideally mem-
brane aligned from the OPM database [6]. In particular, mem-
brane aligned OmpG PDBs 2iww and 2iwv [1] may be
downloaded from: https://opm.phar.umich.edu/ (see
Note 1).
5. Optional membrane PDB atom coordinates file from
CHARMM-GUI [7] (http://www.charmm-gui.org/) (see
Note 1).
6. Git software, for obtaining Pretzel and SCOBY source code,
version 1.7.1 (https://git-scm.com).
7. CASTp for detecting cavities and channels within proteins [8]
(http://sts.bioe.uic.edu/castp/).
8. BOOST C++ libraries, required by SCOBY, version 1.58
(https://www.boost.org/).
9. NAMD [9] molecular dynamics software (see Note 12).
2.2 Equipment Setup We assume access is available to a Linux terminal operating a bash
shell. Download and install all listed software. It is recommended to
obtain the Pretzel and SCOBY source codes via the git clone termi-
nal command; unless otherwise stated, all relative file paths are
relative to the Pretzel repository folder retrieved by the clone oper-
ation (./).
To compile the Pretzel and SCOBY source codes:
1. Navigate to Pretzel’s ./scripts folder and run the shell scripts
cmake_init.sh followed by cmake_build.sh.
2. Copy the compiled pretzel binary from ./CMakeBuild to ./ as
all demonstration shell scripts provided assume ./ to be the
location of the compiled binary.
Pretzel Simulation of pH-Dependent Gating 161
3. Follow an analogous procedure to compile the SCOBY source

code. Similarly, copy the scoby binary into Pretzel’s ./ directory.
If additional details are desired for Pretzel and SCOBY compi-
lation, please refer to the readme.md files within the respective git
repositories.
3 Methods
The application of the Pretzel method consists of the following

general stages: fragment (single-loop) building, multi-loop gener-
ation, protonation of ionizable residues, relaxation of structures
and energy evaluation, and gating state assignment (Fig. 1). Briefly,
the fragment building stage consists of producing 3-D conforma-
tion libraries for the backbones of each individual loop region to be
modeled. The next stage, multi-loop generation, utilizes the single-
loop fragment libraries to efficiently generate an ensemble of 3-D
conformations of the protein with multiple loops (see Note 2).
Next, the pKa value is computed for each ionizable residue and
hydrogens are placed according to the pH level for each multi-loop
conformation. Subsequently, the multi-loop conformations have
Fig. 1 Stages of the Pretzel framework

hydrogens and side chains relaxed using an implicit solvent molec-

ular dynamics simulation; the energy of each conformation is cap-
tured after the last relaxation step. Finally, we assess the gating state
at each conformation using a topological criterion.
3.1 Single-Loop 1. Generate the raw (i.e., unclustered), single-loop, backbone

Fragments fragments by running the shell script ./scripts/a_frag_lib/
a_queue_raw.sh (see Notes 3 and 4).
2. Combine the raw fragments into a single PDB file per loop by
running the shell script ./scripts/a_frag_lib/b_cat_raw.sh.
3. Convert the raw PDB fragments into comma-separated values
(CSV) format by running the shell script ./scripts/a_frag_lib/
c_pdb_2_cluster_csv.sh.
4. Convert the raw, plain-text CSV fragments into binary format
by running the shell script ./scripts/a_frag_lib/d_clus-
ter_csv_2_cluster_bin.sh (see Note 5).
5. For clustering efficiency and to limit the effect of noise, per-
form PCA dimensionality reduction on the raw fragments by
running the shell script ./scripts/a_ frag_lib/e_pca_reduce.sh
(see Note 6).
6. To help mitigate conformational sampling bias, cluster the raw
fragments at each loop by running the shell script ./scripts/
a_frag_lib/f_cluster_shwap.sh; then resample from the clusters
by running the shell script ./scripts/a_frag_lib/g_resample_ids.
sh followed by ./scripts/a_frag_lib/h_resample_pdbs.sh (see
Notes 7 and 8).
7. Copy the refined fragment libraries to the location expected by
the multi-loop generation stage by running the shell script ./
scripts/a_frag_lib/i_build_lib_dir.sh.
3.2 Multi-Loop 1. Generate the raw (i.e., unclustered), multi-loop ensemble by

Generation running the shell script ./scripts/b_multi_loop/a_gen_pdbs.sh
(see Note 3).
2. Convert the raw, multi-loop PDBs into CSV format by running
the shell script ./scripts/c_cluster/a_pdb_dirs_2_cluster_csv.sh.
3. Convert the raw, plain-text, multi-loop CSVs into binary for-
mat by running the shell script ./scripts/c_cluster/b_clus-
ter_csv_2_cluster_bin.sh.
4. For multi-loop clustering efficiency, perform PCA dimension-
ality reduction by running the shell script ./scripts/c_cluster/
c_pca_reduce.sh (see Note 6).
5. To help mitigate multi-loop sampling bias, cluster the raw
ensemble by running the shell script ./scripts/c_cluster/d_clus-
ter_shwap.sh; then resample the ensemble by running the shell
script ./scripts/c_cluster/e_resample_ids.sh followed by ./scripts/
c_cluster/f_resample_pdbs.sh (see Note 7).
3.3 Protonate 1. Compute the pKa of each ionizable residue among the multi-
loop ensemble by running the shell script ./scripts/b_multi_-
loop/b_calc_pka.sh (see Note 9).
2. After pKa computation has completed, it is optional but recom-
mended to remove unneeded files generated by PROPKA3
[10] as this will free up disk space that may be used by other
applications. To remove these unneeded files, run the shell
script ./scripts/b_multi_loop/c_clean_propka.sh.
3. Next, for each multi-loop conformation, place hydrogens at all
residues and assign protonation states according to the pH
level by running the shell script ./scripts/c_cluster/h_gen_-
NAMD_topo.sh (see Note 10).
4. Similarly, after hydrogens have been placed, it is optional but
recommended to remove unneeded files generated by NAMD
[9]. To remove these unneeded files, run the shell script ./
scripts/i_psf_clean.sh.
3.4 Relax and Score 1. To relax hydrogens and side chains in implicit solvent, run
the shell script ./scripts/d_relax/a_NAMD_minimize.sh (see
Note 11).
2. To evaluate the potential energy at each relaxed conformation,
run the shell script ./scripts/d_relax/b_NAMD_capture.sh (see
Notes 12 and 13).
3. For easier downstream analysis, concatenate (i.e., combine)
individual conformation energy data files into aggregated
CSV files by running the shell script ./scripts/d_relax/
c_1_cat_NAMD_energy.sh (see Note 14) followed by ./scripts/
d_relax/c_2_munge_NAMD_interactions.sh (see Note 15).
3.5 Assess Gating 1. Convert the NAMD [9] relaxed atomic coordinates at each
conformation back to a hydrogen-free PDB format expected by
CASTp [8] by running the shell script ./scripts/d_relax/
d_1_coor_to_castp_pdb.sh.
2. Compute topological properties at each relaxed conformation
by running the shell script ./scripts/d_relax/e_call_castp.sh (see
Note 16).
3. To assign “open” (conducting) vs. “closed” (nonconducting)
labels to each relaxed conformation, run the shell script ./
scripts/d_relax/f_open_closed_classify.sh (see Note 17).
3.6 Conclusion Application of the Pretzel framework results in energy and gating
state summaries for an ensemble of 3-D protein conformations.
These summaries may be used for further analysis to identify resi-
dues and regions which drive the gating process [5]. Although here
we discuss only details in applying Pretzel to OmpG, one may
modify the provided shell scripts to model other membrane pro-
teins in which loop regions are intrinsic to the gating mechanism.
4 Notes
1. For OmpG modeling, the membrane-aligned PDBs [6] 2iww

and 2iwv are available in the Pretzel git repository folder: ./
ompg/templates/. In addition, corresponding PDBs containing
atom coordinates for DMPC lipid bilayer membranes, gener-
ated using CHARMM-GUI [7], are located in the folder: ./
ompg/membrane/. All shell scripts used for demonstrating
OmpG modeling assume these PDB coordinate files exist in
these folders.
2. The mDiSGro method [11, 12] is used for sampling protein
loop 3-D conformations.
3. The bash shell scripts ./scripts/a_frag_lib/a_queue_raw.sh and
./scripts/b_multi_loop/a_gen_pdbs.sh generate backbone frag-
ments and multi-loop conformations, respectively, by spawning
multiple concurrent (i.e., parallel) pretzel processes. To control
the number of concurrent pretzel processes, edit the
corresponding script to assign the desired value to the
MAX_NPROC variable; it is recommended to not exceed the
number of available logical cores. For detailed pretzel argument
descriptions, please refer to the comments in https://github.
com/uic-lianglab/ompg-public/blob/master/source/
params.h.
4. Other noteworthy variables in ./scripts/a_frag_lib/
a_queue_raw.sh include START (a list of start residue numbers
for each loop region of interest) and END (a list of end residue
numbers for each loop region). To modify the loop regions
simulated, simply change the corresponding entries in the
START and END lists.
5. The purpose of the shell scripts ./scripts/a_frag_lib/
c_pdb_2_cluster_csv.sh and ./scripts/a_frag_lib/d_clus-
ter_csv_2_cluster_bin.sh is to convert the fragment data into a
format usable by the SCOBY clustering utilities.
6. The shell script ./scripts/a_frag_lib/e_pca_reduce.sh has two
noteworthy variables that may be modified through a text editor.
The first variable, MAX_NPROC, controls the number of con-
current scoby instances used for processing each loop’s fragment
library; if RAM or CPU availability is limited, it is recommended
to decrease this value. The second variable, PERC_KEEP, is a
value in range (0.0, 1.0] and controls the percentage of variance
retained by the principal components; the corresponding multi-
loop ensemble script ./scripts/c_cluster/c_pca_reduce.sh also has
an analogous PERC_KEEP variable. For detailed descriptions of
arguments used by the SCOBY PCA reduce utility, please refer
to the script example_pca_reduce.sh at https://bitbucket.org/
aperezrathke/scoby/src.
7. The mDiSGro [12] sampling algorithm used by Pretzel has

inherent sampling bias. Clustering the conformational geome-
tries followed by resampling can help mitigate this bias by
achieving a more uniform coverage of the structural landscape.
SCOBY provides an implementation of a streaming affinity
propagation clustering algorithm [13, 14]; for detailed argu-
ment descriptions please refer to the script example_shwap.sh at
https://bitbucket.org/aperezrathke/scoby/src.
8. A representative clustering of a large set of raw fragments is
shown in Fig. 2. The resampled fragment libraries for OmpG
loops 1 and 6 are shown in Fig. 3.
9. The PROPKA3 method [10] is used for computing acid disso-
ciation constants (pKa). Similar to other scripts, ./scripts/
b_multi_loop/b_calc_pka.sh launches multiple PROPKA
(python) instances in parallel; to adjust RAM and CPU usage,
one may modify the MAX_NPROC variable to increase or
decrease the number of instances used.
Fig. 2 StrAP [13] algorithm clustering of 278,009 single-loop (fragment)

conformations for OmpG loop 6, resulting in 398 clusters (black boxes). Only
the first two principal components (PC1, PC2) for each fragment’s atomic
coordinates are shown
Fig. 3 OmpG single-loop fragment libraries after clustering and resampling: (a)
loop 1 fragment library; (b) loop 6 fragment library. Each library consists of
10,000 loop conformations
10. The shell script ./scripts/c_cluster/h_gen_NAMD_topo.sh uses

the NAMD [9] psfgen utility with a custom Tcl script to append
hydrogen atoms and to protonate ionizable residues according
to the parameter pH and computed pKa. For demonstration
OmpG modeling, the pH levels used are 5 and 7. To simulate
different pH levels, use a text editor to change the PH_LE-
VELS variable within the shell script. Also, similar to other
scripts, the shell script’s MAX_NPROC variable can be mod-
ified to adjust for CPU and RAM usage.
11. Each protonated 3-D conformation undergoes a modicum of
relaxation primarily to resolve hydrogen and side-chain steric
clashes. For each conformation, the “relaxation” consists of a
separate NAMD [9] molecular dynamics simulation in implicit
solvent using the CHARMM36 [15] force field. Experience
suggests that this phase is the most computationally intensive.
To address this, multiple NAMD processes are spawned con-
currently. Similar to other scripts, to control the number of
spawned process (to adjust for CPU and RAM usage), simply
edit the MAX_NPROC variable within ./scripts/d_relax/
a_NAMD_minimize.sh.
12. The potential energy is captured at each relaxed protein con-
formation using a 0-step NAMD [9] simulation. The NAMD
C++ source code has been modified to provide detailed report-
ing of all nonbonded potential energy terms at each residue.
Pre-compiled binaries are provided in the Pretzel git repository
folder: ./tools/namd/bin/. However, if these binaries are not
system compatible, then the modified NAMD source code
(provided in the Pretzel git repository folder: ./tools/namd/
source/) must be compiled and placed in the corresponding ./
tools/namd/bin/ subfolder. See the “Compiling NAMD” sec-
tion within ./tools/namd/source/notes.txt for detailed instruc-
tions on how to compile the NAMD source code; the shell
script ./tools/namd/source/full_build.sh also provides a compi-

lation reference.
13. Evaluation of the potential energy at each relaxed conforma-
tion is computationally intensive. To address this, multiple
NAMD processes are spawned concurrently. Similar to other
scripts, to control the number of spawned process (to adjust for
CPU and RAM usage), simply edit the MAX_NPROC variable
within ./scripts/d_relax/b_NAMD_capture.sh.
14. The shell script ./scripts/d_relax/c_1_cat_NAMD_energy.sh
generates a plain-text, comma-separated values (CSV) energy
summary file, ener.csv, for each simulated pH level. Each CSV
energy summary file may be opened within any spreadsheet
editor or serve as input for downstream analysis in statistical
computing environments such as R, Matlab, or Python. The
energy summary has the following columns: NAME, PH,
POTENTIAL, ELECT, VDW, BOND, IMPRP, BOUND-
ARY, MISC. The first column, NAME, contains the file name
of each relaxed conformation that had its energy evaluated.
The next column, PH, is the pH at which the energy was
captured. The remaining column entries are the corresponding
potential energy terms at each conformation: POTENTIAL is
the total potential energy, ELECT is the electrostatic potential
energy, VDW is van der Waals potential energy, etc. See the
NAMD 2.11 user guide, section 5.1 (available at https://www.
ks.uiuc.edu/Research/namd/2.11/) for descriptions of the
potential energy terms used by the molecular dynamics force
field.
15. The shell script ./scripts/d_relax/c_2_munge_NAMD_interac-
tions.sh generates a plain-text, comma-separated values (CSV)
interactions summary file, ints_electro.10000.csv, of the electro-
static potential energies at each residue-residue pair. The 10000
in the file name indicates that the summary is limited to the
10,000 lowest energy conformations; to change this limit,
modify the MAX_ENERGY_RANK variable within the shell
script. The first two columns of the CSV interactions summary
file are the PDB residue numbers of the two electrostatically
interacting residues; the remaining columns are the observed
electrostatic energy values for the corresponding row’s residue
pair at each of the evaluated conformations.
16. The CASTp method [8] is used to compute topological prop-
erties such as pockets and mouths [16, 17]. These properties
will be used to heuristically identify possible conducting chan-
nels within each relaxed conformation [5]. The shell script ./
scripts/d_relax/e_call_castp.sh assumes the newcast executable
binary is available through the user’s $PATH environment
variable; for queries regarding batch usage of newcast, please
contact uic.lianglab@gmail.com. In addition, the shell script

models conductance by setting the PROBE_RADIUS variable
equal to 2.75, where this value in Angstroms is the van der
Waals radius of a potassium (K+) ion [18]. To modify the
radius of the conducting ion, simply change the PROBE_RA-
DIUS variable in a text editor. Similar to other shell scripts,
multiple newcast instances are spawned to run in parallel; to
adjust CPU and RAM usage, modify the MAX_NPROC vari-
able in a text editor.
17. The shell script ./scripts/d_relax/f_open_closed_classify.sh will
process the CASTp data for each relaxed conformation to
identify possible conducting channels. The output is a
comma-separated values (CSV) summary file, oc.2.75.csv, with
two columns. The first column is the file name of each confor-
mation. The second column is a “1” or “0” entry indicating if
that row’s conformation is in an “open” (conducting) or
“closed” (nonconducting) state, respectively. The 2.75 in the
name of output CSV file is to denote the radius of the con-
ducting K+ ion (see also Note 16); this may be changed by
modifying the PROBE_RADIUS variable within the shell
script.
Acknowledgements
This research was supported by the US National Institutes of

Health grants R01-GM126558, R01-GM079804, and
R35GM127084.
References
1. Yildiz Ö, Vinothkumar KR, Goswami P, Kühl- OmpG pH-dependent gating from loop
brandt W (2006) Structure of the monomeric ensemble and single channel studies. J Am
outer-membrane porin OmpG in the open and Chem Soc 140:1105–1115
closed conformation. EMBO J 25:3702–3713 6. Lomize MA, Pogozheva ID, Joo H, Mosberg
2. Chen M, Khalid S, Sansom MS, Bayley H HI, Lomize AL (2012) OPM database and
(2008) Outer membrane protein G: Engineer- PPM web server: resources for positioning of
ing a quiet pore for biosensing. Proc Natl Acad proteins in membranes. Nucleic Acids Res 40:
Sci USA 105:6272–6277 D370–D376
3. Fahie M, Yang B, Mullis M, Holden MA, Chen 7. Jo S, Kim T, Iyer VG, Im W (2008)
M (2015) Selective detection of protein homo- CHARMM-GUI: a web-based graphical user
logues in serum using an OmpG nanopore. interface for CHARMM. J Comput Chem
Anal Chem 87:11143–11149 29:1859–1865
4. Fahie M, Chen M (2015) Electrostatic interac- 8. Tian W, Chen C, Lei X, Zhao J, Liang J (2018)
tions between OmpG nanopore and analyte CASTp 3.0: computed atlas of surface topog-
protein surface can distinguish between glyco- raphy of proteins. Nucleic Acids Res 46:
sylated isoforms. J Phys Chem B W363–W367
119:10198–10206 9. Phillips JC, Braun R, Wang W, Gumbart J,
5. Perez-Rathke A, Fahie MA, Chisholm C, Tajkhorshid E, Villa E, Chipot C, Skeel RD,
Liang J, Chen M (2018) Mechanism of Kalé L, Schulten K (2005) Scalable molecular
dynamics with NAMD. J Comput Chem knowledge discovery in databases. Springer,

26:1781–1802 Heidelberg, pp 628–643
10. Olsson MHM, Søndergaard CR, 14. Frey BJ, Dueck D (2007) Clustering by passing
Rostkowski M, Jensen JH (2011) PROPKA3: messages between data points. Science
consistent treatment of internal and surface 315:972–976
residues in empirical pKa predictions. J Chem 15. Huang J, MacKerell AD Jr (2013)
Theory Comput 7:525–537 CHARMM36 all-atom additive protein force
11. Tang K, Zhang J, Liang J (2014) Fast protein field: Validation based on comparison to NMR
loop sampling and structure prediction using data. J Comput Chem 34:2135–3145
distance-guided sequential chain-growth 16. Liang J, Edelsbrunner H, Fu P, Sudhakar PV,
Monte Carlo method. PLoS Comput Biol 10: Subramaniam S (1998) Analytical shape com-
e1003539 putation of macromolecules: I. Molecular area
12. Tang K, Wong SW, Liu JS, Zhang J, Liang J and volume through alpha shape. Proteins
(2015) Conformational sampling and structure 33:1–17
prediction of multiple interacting loops in sol- 17. Liang J, Edelsbrunner H, Fu P, Sudhakar PV,
uble and β-barrel membrane proteins using Subramaniam S (1998) Analytical shape com-
multi-loop distance-guided chain-growth putation of macromolecules: II. Inaccessible
Monte Carlo method. Bioinformatics cavities in proteins. Proteins 33:18–29
31:2646–2652 18. Bondi A (1964) van der Waals volumes and
13. Zhang X, Furtlehner C, Sebag M (2008) Data radii. J Phys Chem 68:441–451
streaming with affinity propagation. In: Joint
European conference on machine learning and
Chapter 13
Free Energy Minimization for Vesicle Translocation Through

a Narrow Pore
Hamid R. Shojaei, Ahad Khaleghi Ardabili, and Murrugappan Muthukumar
Abstract
This chapter presents a mathematical formulation for the translocation process of a vesicle through a narrow
pore. The effect of the deformation of the vesicle while passing through the pore causes a penalty in the free
energy, while the existence of an external driving force assists. We formulate the free energy landscape of the
vesicle in terms of bending and stretching energy and use Fokker-Plank formalism to calculate the first-
passage translocation time. We also address various modifications that can be done to this approach to make
it work for different systems.
Key words Translocation, Vesicle surface energy, Fokker-Plank mechanism, Helfrich free energy
1 Introduction
The control of vesicles’ translocation through pores is fundamental

to drug delivery and transdermal applications [1–3]. Although the
exact process varies based on chemical and physical details of the
system, it can be studied using a coarse-grained theoretical model
[4]. The geometric properties of vesicles as manifested by the
curvature and surface tension of the membrane fluctuate due to
the interaction with the pore and other environmental obstacles.
Analyzing the statistical mechanical properties of these macroscopic
features gives us a significant insight into the behavior of these
complex systems. In order to study the passage of a vesicle through
a pore, we adopted the method which was originally introduced by
Deserno and Gelbart [5]. In their model, they minimized the
energy of the vesicle’s adhesion, membrane tension, and curvature,
under fixed volume constraint, to describe the phenomenon of
endocytosis.
171
172 Hamid R. Shojaei et al.
2 Models
For this chapter, we assume the readers already have some knowl-
edge in mathematics and geometry. We consider the vesicle to be
initially a sphere in the form of r0, and assume that it is incompress-
ible and only the area of the vesicle changes during the transloca-
tion thorough the pore, which has the shape of a cylinder with
radius d (d < r0) and length L.
2.1 Translocation We model the vesicle to undergo smooth deformations while pass-
Process Model ing through the pore. The schematics process of the whole passage
Assumption of the translocation of the vesicle from the donor to the receiver
compartment is shown in Fig. 1. When the vesicle reaches the pore
entrance (Fig. 1a), its shape starts to undergo deformation in order
to fill up the pore (Fig. 1b). The vesicle moves further into the pore,
until it completely fills up the pore and enters the receiving com-
partment (Fig. 1c). The process continues until the vesicle has
completely left the donor compartment, but has still filled the
pore partially (Fig. 1d). Finally, the vesicle completely leaves the
pore and reaches to a new relaxed state on the receiving compart-
ment (Fig. 1e).
Fig. 1 Schematics of the translocation of a spherical vesicle of initial radius r0 through a narrow pore of radius
d and length L. (a) Initial state of the vesicle in the donor compartment, (b) partial penetration of the vesicle
into the pore, (c) filling of the pore with the remainder of the vesicle partitioned into both the donor and
receiver compartments, (d) partial filling of the pore in the exit stage, and (e) final state of the vesicle in the
receiver compartment
Vesicle Translocation Through a Pore 173
2.2 Helfrich Free The free energy landscape of the vesicle due to the fluctuation on its
Energy shape from the free vesicle is in the form of Helfrich free energy:
1 1
FH ¼ κ ∮dA ð2H c0 Þ2 þ λ ðA A0 Þ2 : ð1Þ
2 c 2A0
The first term in Eq. 1 comes from bending the surface, while
the second term accounts for the stretching of the surface. κc and λ
are, respectively, bending modulus and stretching modulus. The
bending term in the free energy is the integral
of (2H c0)2 over
the whole surface. Here, H ¼ 12 r11 þ r12 is the mean curvature,
with r1, 2 being the principal radii at any given point on the surface.
The spontaneous curvature c0 comes from the molecular structure
of the surface. In general, there is an energy cost to bend a surface,
which is given by the integral of the quadratic difference between
mean curvature and spontaneous curvature over the whole surface
of the vesicle. Helfrich free energy originally included the Gaussian
curvature, which after integration gives a topological invariant.
However, its contribution to the free energy can be ignored since
in our model we do not take into consideration the topological
changes in the vesicle.
The second term in the free energy comes from the change on
the surface area from when it is relaxed.
2.3 Filling Stage The stage at which the vesicle starts penetrating the pore but has
not entered the receiving compartment is called the filling stage.
During this stage the deformed shape of the vesicle is shown in
Fig. 2a. To avoid edge penalty on the surface, the curvature should
be finite and smooth at all points over the surface, including the
point at which the vesicle meets the pore entrance. This can be
obtained by introducing a hypothetical torus at the contact point
Fig. 2 Various parameters used to define the free energy of the deformed vesicle. (a) Filling stage: r is the
radius of the partial sphere, b is the small radius of the toroidal vesicle. Dotted lines show the cross section of
a hypothetical torus that keeps the curvature of the vesicle smooth. θ is the angle between the center of the
spherical part and the center of the small toroidal ring. (b) Crossing stage: Two toroids are needed. b0 , r0 , and
θ0 correspond to the toroid on the receiver side, and b, r, and θ correspond to the donor side as in part (a)
between the pore and the vesicle. This torus has the inner radius of
b and outer radius of b + d, with b being an unknown parameter.
The dashed-line circle in Fig. 2a shows the cross-section of this
torus. To calculate the volume of the vesicle in the donor side we
add the volume of the partial sphere to the volume of the cone that
fits into the empty space of the partial sphere. This cone has sharp
edges that extend beyond the boundary of the completed sphere;
therefore, you then have to subtract the volume of the partial
toroidal section. The total form of the vesicle can be divided into
multiple parts: a partial sphere with radius r and cutting angle θ, a
partial torus with inner and outer radius b and b + d and angle θ, a
cylinder with radius d and length l (l < L), and a hemisphere with
radius d. The geometrical parameters of the partial shapes are as
follows.
2.3.1 Partial Sphere The volume of a sphere with radius r and a missing conical angle θ is
2
V sphere ¼ π ð1 þ cos θÞr 3 , ð2Þ
3
and the surface of this partial sphere is given by
A sphere ¼ 2π ð1 þ cos θÞr 2 : ð3Þ
The mean radius of curvature on the surface of the sphere is
constant and is equal to r.
2.3.2 Partial Torus The volume of a partial torus with inner and outer radii b and b + d,
between the angles ν ¼ 2π þ θ and π, as it is shown in Fig. 2a, is

2 π 2
V torus ¼ πb ðb þ d Þ θ b cos θ , ð4aÞ
2 3
and the volume of a cone with angle θ and radius of b + d is
1
V cone ¼ π ðb þ d Þ3 cot θ: ð4bÞ
3
The surface area is given by
h i
π
A torus ¼ 2πb ðb þ d Þ θ b cos θ : ð5Þ
2
The mean radius of curvature of the surface of the torus at
angle ν is
b þ d þ 2b cos ν
H ¼ , ð6Þ
2b ðb þ d þ b cos νÞ
where b and θ are related to each other with geometrical
confinement:
bþd
sin θ ¼ , ð7aÞ
r þb
r sin θ d
b¼ : ð7bÞ
1 sin θ
The total volume of the vesicle in the donor compartment will
be equal to Vsphere + Vcone Vtorus.
2.3.3 Cylinder The volume of the cylinder with length l is

V cylinder ¼ πd 2 l: ð8Þ
A cylinder ¼ 2πdl: ð9Þ
The mean radius of curvature on the surface of the sphere is
1
constant and is equal to 2d .
2.3.4 Hemisphere The volume of a hemisphere with radius d is

2 3
V hemisphere ¼ πd : ð10Þ
3
A hemisphere ¼ 2πd 2 : ð11Þ
Now that we have set up different geometrical parameters of
various possible shapes of the vesicle in the filling stage, we can start
reviewing crossing stage.
2.4 Crossing Stage The crossing stage starts when the vesicle enters the receiving
compartment, and it lasts until the vesicle leaves the donor com-
partment. In this stage, the vesicle occupies both the donor and
receiving compartments, as well as the pore completely. The form
of the vesicle for this stage can be defined in a similar manner as the
filling stage. The general form of the vesicle can be divided into a
partial sphere and a partial toroid for either donor or receiving
compartments with different radii and angles on each side and a
cylinder with the form of the pore and the length of L. A schematic
figure of this stage is depicted in Fig. 2b. In this stage, in addition to
0 0 0
b, r, and θ in the donor compartment, we have b , r , and θ for the
0 0 0
receiving compartment. Similar to Eqs. 7a and 7b, b , r , and θ are
related to each other by the following formulas:
b0 þ d
sin θ0 ¼ , ð12aÞ
r 0 þ b0
r 0 sin θ0 d
b0 ¼ : ð12bÞ
1 sin θ0
2.5 Depletion Stage The final stage, at which the vesicle has completely left the donor
compartment but is filling the pore partially is called the depletion
stage. The geometrical structure of the vesicle is the same as the

filling stage, but it happens in the receiving compartment.
2.6 Volume For most biological systems and for bilayer membranes, the vesicle
Constraint is incompressible. However, one can add the compressibility
parameter to the system. The initial volume of a spherical vesicle
with radius r0 is
4 3
V0 ¼ πr : ð13Þ
3 0
In order to parametrize the translocation process, we need to
define the translocation coordinate, α, which is the ratio of the
volume of the vesicle that passed the pore entrance at any given
time (Vpassed) to the initial volume of the vesicle during the translo-
cation process:
V passed
α¼ : ð14Þ
V0
If we define the volume ratio of the pore to the vesicle as β
V pore πd 2 L
β¼ ¼ , ð15Þ
V0 V0
then α ¼ 0 is the starting time and α ¼ 1 + β represents the end of
the translocation process.
2.7 External Force Translocation only happens if there is an external force pushing the
vesicle in the donor compartment toward the receiving compart-
ment through the pore. The most common form of such a force for
vesicles is osmotic pressure. We can consider that an external pres-
sure P1 is applied to the vesicle with volume V0 in the donor
compartment, which is higher than the pressure in the receiver
compartment P2. If we assume that this pressure difference
between donor and receiver compartments is constant, then the
external contribution to the free energy will be
V passed
F ext ¼ ΔPV ¼ ΔPV 0 : ð16Þ
V0
Using Eq. 14 and defining an energy term
f 0 ¼ ΔPV 0 ð1 þ βÞ,
we can now write the external contribution to the free energy in
Eq. 16 as a function of the translocation coordinate
α
F ext ¼ f 0 : ð17Þ
1þβ
In the process of translocation α varies between 0 and 1 + β.
However, when the vesicle is completely out of the donor
compartment, α ¼ 1. And from this time on we

take F ext ¼ f 0 1=1þβ for the rest of the process.
2.8 Fokker-Plank The kinetics of the translocation process can be studied using the
Formalism free energy landscape. Using the translocation coordinate, α, and
the Fokker-Planck formalism, we can calculate the average translo-
cation time
Z Z α1
1 1þβ
τ¼ dα1 dα2 e½F ðα1 ÞF ðα2 Þ=kB T , ð18Þ
κ 0 0
with
F ðαÞ ¼ F H þ F ext , ð19Þ
which are defined in Eqs. 1 and 17. In order to apply Fokker-Planck
formalism we need to assume that the translocation time is longer
than the relaxation time of thevesicle, which is the time required by
the vesicle to get back to its equilibrium form after any distortion in
its shape. This will then satisfy the detailed balance requirement of
Fokker-Planck formalism. As vesicle goes through the pore, there
will be a local friction, and it is incorporated in formula 18 by
parameter κ. This parameter is also set as the unit of time for our
problem.
3 Methods
After knowing all the details, we can calculate the free energy
landscape as a function of the translocation coordinate α. Before
the translocation process starts, the bending and the stretching
terms in the Helfrich free energy are
" 2 #
κc 2 2
F 0,bend ¼ 4πr 0 c0 ¼ κ c ð2π Þð2 r 0 c 0 Þ2 , ð20aÞ
2 r0
F 0,stretch ¼ 0: ð20bÞ
The bending and stretching energy contributions to the free
energy on each stage of translocation process are calculated in the
following sections.
3.1 Filling Stage In the filling stage 0 < α < β. The volume constraint in Eq. 13 can
be written as
V donor ¼ ð1 αÞV 0 ¼ V sphere þ V cone V torus : ð21Þ
The rest of the vesicle can be used to get the length of the
cylinder, which is filled
2 4
πd 2 l þ πd 3 ¼ αV 0 ¼ παr 30 ,
3 3

2 r 30
l¼ d 2α 3 1 : ð22Þ
3 d
Using Eqs. 2, 4a and 4b and substituting b from Eq. 7b, we
obtain r(θ) for any given value of α. Both r and θ should be real.
Also, as it can be seen in Fig. 2a, r and θ are bounded parameters:
d < r < r 0, ð23aÞ
d π
sin 1 <θ< : ð23bÞ
r0 2
Also, since part of the vesicle can practically pass the pore
entrance before touching the edge of the pore, we have a mathe-
matical limit for the minimum value of α:
1 d3
αmin ¼ < α: ð23cÞ
2 r 30
The bending free energy is
F I,bend ¼ F I,b,sphere þ F I,b,torus þ F I,b,cylinder þ F I,b,hemisphere : ð24Þ
The first term is
Z 2
1 2
F I,b,sphere ¼ κc dA c0
2 r
1
¼ κ ð2π Þð2 rc 0 Þ2 ð1 þ cos θÞ: ð25Þ
2 c
The second term is
Z 2
1 b þ d þ 2 b cos ν
F I,b,torus ¼ κc dA c0 ,
2 2b ðb þ d þ b cos νÞ
n
1 π
F I,b,torus ¼ κc 2π ðc 0 b þ 2Þ2 cos θ þ 2πc 0 ðc 0 b þ 2Þ θ ðb þ d Þ
2 2
2 pffiffiffi
2π ðb þ d Þ 1 d θ π
þ pffiffiffipffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi π 2 tan pffiffiffiffiffiffiffiffiffiffiffiffiffiffi tan þ :
b d 2b þ d 2b þ d 2 4
ð26Þ
The last two terms in Eq. 24 are
2
1 1
F I,b,cylinder ¼ κc πd l
2
c0
2 d

1 2 r 30
¼ κc π 2α 3 1 ð1 c 0 d Þ2 , ð27Þ
2 3 d
Z 2
1 2 1
F I,b,hemisphere ¼ κ c dA c 0 ¼ κc ð2π Þð2 c 0 d Þ2 : ð28Þ
2 d 2
The stretching term is
1 1
F I,stretch ¼ λðAI A 0 Þ2 ¼ λðΔA I Þ2 , ð29Þ
2A 0 2A 0
with
ΔAI ¼ A sphere þ A torus þ A cylinder þ A hemisphere 4πr 20 : ð30Þ
Each term in Eq. 29 is already calculated in the materials
section. For any given α, after substituting r(θ) of the volume
constraint, the free energy will be only a function of r (or θ). Now
that we have the free energy as a function of α, we can minimize the
free energy equation with respect to θ and get the free energy
landscape at any α.
3.2 Crossing Stage In the crossing stage, β < α < 1. We introduce the counterpart
parameters, which corresponds to the received compartment with
prime.
V donor ¼ ð1 αÞV 0 ¼ V sphere þ V cone V torus , ð31Þ
V receiver ¼ ð1 α0 ÞV 0 ¼ V 0sphere þ V 0cone V 0torus , ð32Þ
0 0
with (1 α) + (1 α ) + β ¼ 1, or α ¼ 1 + β α. The bending
energy for this stage is
F II,bend ¼ F II,b,sphere þ F II,b,torus þ F II,b,pore þ F 0II,b,sphere
þ F 0II,b,torus : ð33Þ
With assistance of Eqs. 12a and 12b, each term in Eq. 33 is as
follows
1
F II,b,sphere ¼ κ ð2π Þð2 rc 0 Þ2 ð1 þ cos θÞ, ð34aÞ
2 c
1
F 0II,b,sphere ¼ κ ð2π Þð2 r 0 c 0 Þ ð1 þ cos θ0 Þ,
2
ð34bÞ
2 c
n
1 π
F II,b,torus ¼ κc 2π ðc 0 b þ 2Þ2 cos θ þ 2πc 0 ðc 0 b þ 2Þ θ ðb þ d Þ
2 2
pffiffiffi
2π ðb þ d Þ2 1 d θ π
þ pffiffiffipffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi π 2 tan pffiffiffiffiffiffiffiffiffiffiffiffiffiffi tan þ ,
b d 2b þ d 2b þ d 2 4
ð34cÞ
n 2 π
1
F II,b,torus ¼ κ c 2π c 0 b 0 þ 2 cos θ0 þ 2πc 0 c 0 b 0 þ 2 θ0 b 0 þ d
2 2
0 2 pffiffiffi 0 )
2π b þ d 1 d θ π
þ 0 pffiffiffipffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi π 2 tan pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi tan þ ,
b d 2b 0 þ d 2b 0 þ d 2 4
ð34dÞ
2 r 3 2
1 1 1 8
F II,b,pore ¼ κ πd 2 L c0 ¼ κ πβ 0 1 c0d :
2 c d 2 c 3 d3
ð34eÞ
Where L is the length of the pore. The same as filling stage, the
stretching energy ist
1 1
F II,stretch ¼ λ ðA II A 0 Þ2 ¼ λðΔA II Þ2 , ð35aÞ
2A 0 2A 0
with
ΔA II ¼ A sphere þ A torus þ A pore þ A 0sphere þ A 0torus 4πr 20 , ð35bÞ
0 0 0
with r, b, and θ in A, respectively, replaced by r0, b , and θ in A . The
minimization process is the same as the filling stage. Since r should
monotonically and continuously reduce from r0 to d, we can mini-
mize the free energy separately for the donor and receiver para-
0
meters (r and r ) to get the free energy landscape.
3.3 Depletion Stage For 1 <α < 1 + β, we are in the depletion stage. In this stage,
everything is completely the same as the filling stage, by replacing α
0
with α ¼ 1 + β α.
3.4 Free Energy The general form of the free energy, as we calculated in previous
Barrier sections, is shown in Fig. 3. Figure 3a shows the free energy
landscape without an external driving force ( f0 ¼ 0). Figure 3b
shows the free energy landscape with the existence of an external
driving force ( f0 6¼ 0). While we assume the spontaneous curvature
to be 0, there is no restriction on having a nonzero c0.
3.5 Average Having the free energy landscape, we can use Eq. 18 to find the first
Translocation Time passage translocation time. In order to solve the double integral in
Eq. 18, we need to increment α with a value δ. At each step we
calculate the numerical value of the free energy and then sum over
its numerical values. The results for an arbitrary value of κc and λ for
multiple values of r0 is shown in Fig. 4. For example, with the
driving force f0 ¼ 15 and the radius of the vesicle r0 ¼ 1.75 with
bending modulus κc ¼ 3 and stretching modulus λ ¼ 5, the trans-
location time will be equal to 500. Changing any parameter here
has a big impact on the outcomes. In the following paragraph we
give some examples.
3.5.1 Vesicle Radius r0 Let us increase the radius of vesicle to twice its value (i.e., r0 ¼ 2.5).
To achieve the same translocation as in previous example (i.e.,
τ ¼ 500) one needs to increase the external free energy drive f0 by
more than 4.6 times to f0 ¼ 70.
3.5.2 Elastic Moduli κc If all the parameters are the same except elastic modulus κc, then
and λ increasing it by a factor of 2 will increase the translocation time τ up
to six orders of magnitude. The effect of λ is much weaker in
comparison to κ c.
Fig. 3 The free energy barrier with and without an external driving force. (a) The Helfrich free energy landscape
as a function of α. There is no external driving force ( f0 ¼ 0). Three different stages of the vesicle
translocation (filling, crossing, and depletion) is shown. The black line shows the curve when the spontaneous
curvature is zero, and the blue dashed line is for the case when there is a nonzero (c0 6¼ 0). Here (r0 ¼ 2.5d,
β ¼ 0.3, κc ¼ 2 and λ ¼ 5). (b) The Helfrich free energy landscape as a function of α in the presence of an
external driving force ( f0 6¼ 0). While we assume the spontaneous curvature to be 0 (c0 ¼ 0), there is no
restriction on having a nonzero c0. We plotted the curve for different values of stretching and bending moduli.
The black line shows the case when κ c ¼ 2 and λ ¼ 5, the blue dashed line has a different value, λ ¼ 2.5,
and the red dashed line shows the curve when κc ¼ 1 and λ ¼ 5
Fig. 4 The effect of the external driving force f0 on translocation time τ. (a) Shows the case when the vesicle’s
radius is r0 ¼ 1.75 and in (b) r0 ¼ 2.5. The effect of different stretching and bending moduli is depicted in both
(a) and (b)
4 Notes
1. In order to run the program, it is enough to have Python3 in

the system. We use Numpy, Scipy, and matplotlib libraries to
perform the calculations. The program has many comments
within itself. It is designed to be user friendly and people from
different backgrounds will be able to use it at any level they
want. The part which is for the user is at the bottom of the
program; there are possibilities for changing the parameters as
the user wishes. There are a few notes that should be consid-
ered while using the program, which are mentioned in this
section.
VesicleTranslocation(r0, β, Fext, κc, λ) accepts r0, β,
Fext, κc, and λ as input, and as output it prints the energies at
each α and plots the figure which shows the relationship
between free energy and α. The user also needs to enter the
values for the κc and λ, which are bending modulus and a
stretching modulus, respectively.
Users can obtain the program from Github. In order to run
the program, one needs first to go to the section which starts
with “if __name__ ¼¼ ’__main__’:”. In this section the user
can choose parameters in VesicleTranslocation(r0, β,
Fext, κ c, λ) as inputs to start the computation. If the user has
an IDE for python, he/she can simply run the program in IDE.
Otherwise from the terminal the following command will run
the program:
Python vesicletranslocation.py
More information about the program is available at the
Github address (https://github.com/hamid-shojaei/vesicle-
translocation).
2. One needs to pay attention to the range of β. If the size of the
vesicle is smaller than the pore, then β 1 and the free energy
will not change inside the pore. If the length of the pore is zero
(l ¼ 0), we do not have crossing stage, and it will be just filling
and depletion. When there is an external force ( f0 6¼ 0), the
form of the external force will be different if it is in the deple-
tion process.
3. One possible extension to this work is to consider a conical
shape instead of a cylindrical geometry for the pore. The cross-
ing stage formulas should be modified. In this case, Eqs. 4a, 8,
9, 24, and 33 will be modified. It would be interesting to see
the impact of geometry of the pore on the properties of trans-
location since the real biological environment can be far from
perfect.
4. In this work we ignored the electrostatic effects of vesicle-fluid
interface on the translocation process. A more detailed and
comprehensive study which includes the effect of a charged

vesicle is another direction for future works.
5. A different shape shape of the vesicle can be chosen, but then
one needs to change H in Eq. 1 and all the equations in
different stages of translocation, beside those which describes
a cylindrical pore. Choosing a different vesicle shaple can also
potentially impact the translocation process.
Acknowledgements
This research was supported by NSF Grant No. DMR-1713696

and AFOSR Grant No. FA9550-17-1-0160.
References
1. Mezei MM, Gulasekharam V (1980) Lipo- 3. Guo X, Szoka FC (2003) Chemical approaches
somes—a selective drug delivery system for the to triggerable lipid vesicles for drug and gene
topical route of administration I: lotion dosage delivery. Chem Res 36:335–341
form. Life Sci 26:1473–1477 4. Hamid SR, Murugappan M (2016) Transloca-
2. Wachter C, Vierl U, Cevc G (2008) Adaptability tion of an incompressible vesicle through a pore.
and elasticity of the mixed lipid bilayer vesicles J Phys Chem B 120:6102–6109
containing non-ionic surfactant designed for tar- 5. Deserno M, Gelbart WM (2002) Adhesion and
geted drug delivery across the skin. Drug Target wrapping in colloid-vesicle complexes. Phys
16:611–625 Chem B 106:5543–5552
Part IV
Droplet Interface Bilayer System

Chapter 14
Single Ion-Channel Analysis in Droplet Interface Bilayer

Arash Manafirad
Abstract
Droplet interface bilayer (DIB) is a method of fabricating lipid bilayer membrane by contacting two
aqueous droplets coated with a monolayer of lipid molecules in oil media. Lipids coat the droplet surface
either by vesicles fusing to the water-oil interface from the droplet side or diffusing toward the interface
from the oil side, thereby forming a lipid monolayer. With the DIB technique, nanoliter amounts of
aqueous solution is needed and one may obtain two different compositions of monolayers to form
asymmetric bilayer which is difficult to replicate by other in vitro lipid membrane methods. Here, a
DIB-based protocol is reported to fabricate a stable lipid bilayer membrane to perform single-channel
electrophysiology on a pore-forming toxin.
Key words Droplet interface bilayer, Lipid membrane, Electrophysiology, Ion-channel analysis, Lipid
monolayer, DIB
1 Introduction
Screening ion channels for unwanted drug interactions is an expen-

sive and painstaking process. Though methods based on fluores-
cent plate readers have accelerated the screening process,
electrophysiology is still required to record specific channel-drug
interactions. A method called patch clamp electrophysiology
involves direct contact of a glass pipette with an isolated patch of
the cell membrane that is overexpressed with the ion channel of
interest. This method requires live cells and sophisticated instru-
mentation and every time a new cell line must be created for each
channel or mutation to be studied [1, 2].
Different techniques, including supported bilayer membranes
[3–6], the Montal-Mueller lipid membranes [7], and the giant
unilamellar vesicles (GUVs) [8–10], have been developed in recent
decades to fabricate artificial bilayer membranes on which to study
ion channels. Recently, the droplet interface bilayer (DIB)
approach has been developed to incorporate membrane proteins
for in situ analysis [11, 12]. The DIB is a replica of a cell membrane
187
188 Arash Manafirad
that is created by joining two independently formed lipid mono-

layers together upon the contact of aqueous nanoliter sized dro-
plets in oil [11]. Monolayer lipids may originate either from the oil
phase (the lipid-out approach that lipids are dissolved in oil) or as
vesicles in the aqueous phase (the lipid-in approach where lipids are
dispersed in the aqueous phase) [13]. Monolayers created from
vesicles form quickly (~2 min) and allow each droplet to possess a
unique lipid composition, i.e., the possibility of forming asymmet-
ric bilayers which is closer to the asymmetry found in biological
membranes [13]. In some cases, however, the presence of lipids as
vesicles in the aqueous phase may be problematic due to unspecific
interaction with a biological medium. For the lipid-out approach,
the bilayer formation time and success rate highly depend on the
concentration of the lipid(s). Typically, total lipid concentration of
less than 2 mg/mL will be insufficient to form a stable bilayer, and
usually before the formation of the bilayer the droplets immediately
fuse together.
Here in this chapter, the fidelity of the lipid bilayer is tested by
addition and then incorporation of a bacterial toxin called
α-hemolysin in the droplet interface bilayer. We examine toxin
activity through the reversible interaction of γ-cyclodextrin as a
model for ion channel analysis. The α-hemolysin is added to a
droplet (on the working electrode) with a final concentration of
0.1 μg/mL. Similarly, γ-cyclodextrin is added to the second droplet
(on the ground electrode) with a final concentration of 50 μM. We
adopt a combinatorial approach to prepare a rigid lipid bilayer,
which is composed of two different lipids, DPhPC and DPhPE in
a 4:1 mole ratio. This specific lipid composition is shown to possess
an enhanced electrical stability [14].
2 Materials
Before any attempts to prepare the solutions, safety precautions and

waste disposal regulations may be diligently followed. Prepare all
the solutions using 18 MΩ cm ultrapure deionized water. Unless
indicated otherwise, use analytical grade reagents. Pay extra care to
avoid detergent contamination of the glassware and instruments
washed with soap. The presence of detergent will compromise the
stability of the DIB and eventually will disrupt the membrane
bilayer.
1. DIB buffer solution: 120 mM KCl, 60 mM NaCl, 10 mM
MgCl2, and 10 mM HEPES (pH 7.4, titrated with KOH).
The buffer should be stored at 4 C and used no longer than
1 week.
2. 3% (wt/vol) agarose solution: the agarose gel is used to avoid
droplet detachment from the electrodes and provide
Single Ion-Channel Analysis in Droplet Interface Bilayer 189
hydrophilicity. To prepare an agarose solution, dissolve 30 mg

of low melt agarose in 1 mL of DIB buffer solution to obtain a
3% (wt/vol) solution. This solution can be used for several
weeks if stored at 4 C. A metal heat block set at 50–60 C is
needed to melt down the agarose each time before any
experiment.
3. Ag/AgCl electrodes: Cut two 20 mm length of 100 μm diam-
eter silver wire.
4. 15% (wt/vol) Sodium hypochlorite solution.
5. A vacuum chamber or Vacuum desiccator.
6. 10 mg stock solution of DPhPC lipid solution: dissolve 25 mg
of 1,2-diphytanoyl-sn-glycero-3-phosphocholine (DPhPC;
Avanti Polar Lipids) in 2.5 mL of pentane, then store it in a
glass vial wrapped in parafilm in a 20 C freezer.
7. 10 mg stock solution of DPhPE lipid solution: dissolve 25 mg
of 1,2-diphytanoyl-sn-glycero-3-phosphoethanolamine
(DPhPE; Avanti Polar Lipids) in 2.5 mL of chloroform (see
Note 1). Store the solution in a glass vial wrapped in parafilm in
a 20 C freezer.
8. Mini-extruder (Avanti Polar Lipids).
9. Nuclepore track-etched polycarbonate membrane, 0.1-μm
pore size, 19-mm diameter (Whatman).
10. Faraday cage.
11. Micromanipulator: an NMN-21, three-axis micromanipulator
(Narishige, JP).
12. Patch-clamp amplifier (Axopatch 200B; Axon Instruments).
13. Stereomicroscope (AmScope model no. SW-3T24Z-FRL).
14. Extruded sheet poly (methyl methacrylate) (PMMA, 4-mm
thickness).
15. 4 mL-borosilicate glass vials with caps protected by PTFE.
16. Wild-type α-hemolysin pre-heptamer from Staphylococcus
aureus. Upon receiving the toxin, make a stock solution of
20 μg/mL concentration in buffer per manufacturer’s instruc-
tion and split it into 5 μL aliquots and flash-freeze the aliquots
in liquid nitrogen and immediately store them in 80 C
freezer.
17. γ-Cyclodextrin substrate for α-hemolysin translocation testing.
3 Methods
3.1 Lipid-In 1. Take 340 μL of DPhPC stock solution and mix with 80 μL of
Approach Vesicle DPhPE stock in a glass vial.
Preparation
190 Arash Manafirad
2. To this mixture add 100 μL of chloroform and 100 μL of n-

pentane.
3. Swirl the vial gently under a stream of nitrogen to evaporate the
solvents to obtain a lipid film.
4. To remove the trace amount of the organic solvents, vacuum-
dry the lipid film for 2 h in a vacuum chamber or a vacuum
desiccator.
5. Hydrate the lipid film by adding 1 mL of the DIB buffer
solution and vortex the solution with the vial cap firmly closed
and sealed with a parafilm for 10 min (see Note 2).
6. Freeze-thaw (eight cycles) the lipid solution in liquid nitrogen
and a metal heat block warmed to a temperature of 60 C.
7. Extrude the lipid solution 11 passes through the membrane
filter with the Avanti mini-extruder. This protocol would yield
a final lipid concentration of 4 mM.
8. Store the vesicle solution in a plastic vial (see Note 3).
9. This vesicle solution can be stored at 4 C and may be used up
to 1 week.
3.2 Lipid-Out 1. Mix 600 μL of DPhPC with 150 μL of DPhPE in a 25-mL glass
Approach Vesicle vial.
Preparation 2. Evaporate the solvents under a stream of nitrogen gas.
3. Vacuum-dry the formed lipid film for 2 h in a vacuum chamber
or a vacuum desiccator.
4. Resuspend the lipids by adding 1 mL n-hexadecane oil.
5. Stir this solution for a few hours before use. This protocol will
yield a total lipid concentration of 7.5 mg/mL, which is
enough to form a lipid bilayer quite rapidly (<1 h).
3.3 Fabrication 1. Fill the oil bath (e.g., a 1 mL volume PMMA chamber, see
of Droplet Interface Fig. 1) with n-hexadecane or lipid-dissolved hexadecane for
Bilayer the lipid-out approach.
2. Place the chamber in a metal Faraday cage (see Note 4).
3. Prepare the Ag/AgCl electrodes by gently melting one end of
each wire over a flame to obtain a silver ball of ~200 μm in
diameter (see Fig. 2a).
4. Wash the ball end of the electrodes with deionized water.
5. Dip the ball end of each wire into the 15% (wt/vol) sodium
hypochlorite solution for 10 min or until the ball end turns into
a dark gray color. Electrodes are etched to provide a conduct-
ing medium, i.e., activate the electrode surface.
6. Wash the ball end of the electrodes with deionized water several
times.
Fig. 1 Subtractive micromilling (Modela MDX-40A, Roland DG) was used to fabricate a PMMA-based device in
which the droplet interface bilayers were formed
7. Dip the ball end of each electrode into the 3% (wt/vol) agarose
solution to adopt a thin layer of agarose on the surface of the
ball end of the electrodes. Upon cooling, an agarose droplet
forms at the end of each electrode (see Fig. 2a). To avoid an
osmotic shock due to dehydration of the gel, immediately
submerge the agarose-coated electrodes in the hexadecane-
containing oil chamber (see Note 5).
8. Assemble the electrodes to the micromanipulator and attach
them to the head stage of the patch-clamp amplifier with
appropriate insulated connectors (see Note 6).
9. Apply 0.2 μL of either vesicle solution for the lipid in approach
or only DIB buffer solution in the case of the lipid-out
approach to the ball end of agarose-coated electrodes that are
already submerged into the oil. A stereomicroscope can be
helpful to view and locate the droplets.
10. Allow ~2–5 min for the monolayers to form on each electrode.
As the monolayers form the interfacial tension drops and this
results in droplets to sag from the electrode (see Fig. 2b). Thus,
it is important to give the droplets enough time to form defect-
free and tension-free monolayers.
11. Carefully bring the two droplets to the close vicinity of each
other without pushing them hard against each other as this
may cause droplet fusion (see Fig. 2c and Note 7). For the best
192 Arash Manafirad
Fig. 2 DIB formation steps. (a) Electrodes area etched with sodium hypochlorite, coated with agarose gel, and
immediately dipped into the n-hexadecane oil bath. (b) Two 0.2 μL aqueous droplets are applied to both
electrodes. After 3 min both droplets relaxed and sagged. (c) Gently the two droplets with the aid of
micromanipulator are contacted. +50 mV voltage was applied to expedite the bilayer formation process. (d)
After the formation of the lipid bilayer, the droplets are moved away from each other to show the point of
contact. The inset schematically depicts the formation of the bilayer from the monolayers. Scale bar is 0.5 mm
results, you may use a stereomicroscope to view the electrodes

and droplets and follow the procedure.
12. Close the lid of the Faraday cage after contacting the droplets
and wait for 1–2 min for the bilayer to form. The process of
development of the bilayer can be monitored by capacitance
measurement of the lipid bilayer (see Fig. 3a).
3.4 Ion Channel 1. Take out an aliquot of stock α-hemolysin and thaw it on ice (see
Activity Recording Note 8).
2. Dilute the α-hemolysin stock solution with 18 MΩ water to
2 μg/mL.
3. Dilute the 2 μg/mL α-hemolysin solution to 0.1 μg/mL using
the vesicle solution in case of the lipid-in approach or with DIB
buffer solution in the case of the lipid-out approach.
Fig. 3 Ionic activity of an α-hemolysin nanopore in a DIB system. (a) Lipid bilayer membrane capacitance
measurement performed by the electrophysiology recording. The arrow points to the time point that the bilayer
starts to form from the monolayers. (b) Electrophysiology recordings of a single α-hemolysin toxin
incorporated into a DIB and its ionic activity is partially blocked by the entrance of the γ-cyclodextrin substrate
into the pore lumen. Arrow points to the time in which the toxin is inserted into the DIB
4. Dilute the stock solution of the γ-cyclodextrin substrate to a

concentration of 500 μM in deionized water.
5. Dilute this 500 μM γ-cyclodextrin solution further with the
vesicle solution to a final concentration of 50 μM.
6. Pipette 0.2 μL of the working α-hemolysin solution prepared in
step 3 onto the working electrode (the trans electrode).
7. Pipette 0.2 μL of the working 50 μM γ-cyclodextrin solution
onto the ground electrode (the cis electrode).
8. Follow steps 9–11 in Subheading 3.3 Fabrication of Droplet
Interface Bilayer.
194 Arash Manafirad
9. Apply a constant voltage of 50 mV to expedite the bilayer

formation and monitor the progress of bilayer formation by
checking the capacitance (see Fig. 3a).
10. Use the micromanipulators and carefully adjust the inter-
droplet distance so that the capacitance remains stable with
an amplitude of 400 pA (see Note 9).
11. Turn off the capacitance reading (see Note 10).
12. In about 10 min, a single step rise in the signal will appear that
is due to single toxin incorporation in the lipid bilayer. Upon
the toxin incorporation, the γ-cyclodextrin will act to block the
toxin for a short period of time. This blockage, thereby, man-
ifests itself in the signal as steps appearing and disappearing (see
Fig. 3b).
4 Notes
1. We highly recommend using Luer lock glass syringe and a

metal needle when preparing the stock lipid solutions. Using
plastic pipette tips may contaminate the lipid solution.
2. Note that the sonication of the lipid solution in this step is not
recommended.
3. Storing the vesicle solution for a long period of time might
result in loss of lipids as they may stick to the glass container.
4. To minimize the noise and vibrations we recommend placing
the patch-clamp amplifier head stage and the micromanipula-
tors inside the cage as well. Also, all the wiring must be insu-
lated. One tip is to wrap the wiring in aluminum foil.
5. We recommend preparing fresh electrodes before each experi-
ment, even though each electrode can be used for up to
2 weeks.
6. Any disconnection of the head stage from the electrodes or
shorting of the head stage-electrode connection can result in
droplet fusing. Be cautious and secure the electrodes into the
head stage and use a power surge protector for all equipment.
7. The bilayer formation process is a spontaneous process, so
rough positioning of the droplets close together, but not
stacked each other, would suffice to form a stable bilayer for
the experiments.
8. For rapid thawing of the α-hemolysin aliquot one may keep the
aliquot under running cold water. But we recommend gentle
thawing in ice. As this protein is a toxin, it is recommended that
one uses gloves when handling.
9. The capacitance observation to verify the bilayer formation

should be done in less than 20 s if the working droplet contains
ion channels or toxins. These membrane proteins usually
become incorporated seconds after the bilayer has formed.
10. On the Axopatch 200B control panel, in the “commands”
section turn off the external command knob.
References
1. Dunlop J, Bowlby M, Peri R et al (2008) Ion and thermodynamics of lipid biomembranes.
channel screening. Comb Chem High Faraday Discuss 161:591–611
Throughput Screen 11:514–522 9. Wesołowska O, Michalak K, Maniewska J,
2. McManus OB, Garcia ML, Weaver D et al Hendrich AB (2009) Giant unilamellar vesi-
(2012) Ion channel screening. In: Assay guid- cles; a perfect tool to visualize phase separation
ance manual. Eli Lilly & Company and the and lipid rafts in model systems. Acta Biochim
National Center for Advancing Translational Pol 56:33–39
Sciences, Bethseda, MD 10. Fischer A, Franco A, Oberholzer T (2002)
3. Tamm LK, McConnell HM (1985) Supported Giant vesicles as microreactors for enzymatic
phospholipid bilayers. Biophys J 47:105–113 mRNA synthesis. Chembiochem 3:409–417
4. Sackmann E (1996) Supported membranes: 11. Bayley H, Cronin B, Heron A et al (2008)
scientific and practical applications. Science Droplet interface bilayers. Mol BioSyst
271:43–48 4:1191–1208
5. Richter RP, Bérat R, Brisson AR (2006) For- 12. Syeda R, Holden MA, Hwang WL, Bayley H
mation of solid-supported lipid bilayers: an (2008) Screening blockers against a potassium
integrated view. Langmuir 22:3497–3505 channel with a droplet interface bilayer array. J
6. Holden MA, Lein MJ, Manafirad A, Aurian- Am Chem Soc 130:15543–15548
Blajeni DE (2017) Membrane and droplet- 13. Hwang WL, Chen M, Cronin B et al (2008)
interface bilayer systems and methods. US Pat- Asymmetric droplet interface bilayers. J Am
ent, 5 Jan 2017 Chem Soc 130:5878–5879
7. Montal M, Mueller P (1972) Formation of 14. Andersson M, Jackman J, Wilson D et al
bimolecular membranes from lipid monolayers (2011) Vesicle and bilayer formation of diphy-
and a study of their electrical properties. Proc tanoylphosphatidylcholine (DPhPC) and
Natl Acad Sci U S A 69:3561–3566 diphytanoylphosphatidylethanolamine
8. Evans E, Rawicz W, Smith BA (2013) Con- (DPhPE) mixtures and their bilayers’ electrical
cluding remarks Back to the future: mechanics stability. Colloids Surf B Biointerfaces
82:550–561
Chapter 15
Continuous and Rapid Solution Exchange in a Lipid Bilayer

Perfusion System Based on Droplet-Interface Bilayer
En-Hsin Lee
Abstract
Because of the high sensitivity of lipid bilayers to external pressure fluctuations, a major challenge in
functional studies of biological pores or ion channels is the difficulty in exchanging solutions rapidly
while maintaining the stability of the lipid bilayer in a model membrane. Here we describe a droplet-
interface bilayer-based perfusion system that has been routinely used in our research and is currently the
most robust and stable perfusion system that provides prompt solution exchange surrounding a lipid
bilayer. In this model membrane system, solutions can be completely exchanged within 1–2 s to obtain
prompt responses of a lipid bilayer or membrane pores to the membrane environments. Also, our system is
stable enough to sustain continuous perfusions up to at least dozens of minutes. To demonstrate, we show
that acidification-induced protein channel insertion, substrate binding to protein channels, and pH
gradient-driven protein translocation of anthrax toxin can be sequentially initiated by continuous perfu-
sions in our system. Moreover, by rapidly switching the solutions, the protein translocation based on ratchet
mechanisms can be paused and reinitiated iteratively in our system. Overall, this perfusion system provides a
controllable and reliable solution exchange platform for investigations of pores and translocations on lipid
bilayers.
Key words Lipid bilayer, Solution exchange, Perfusion system, Microfluidics, Droplet-interface
bilayer, Ion channels, Proton gradient, Protein transport
1 Introduction
In the past decade, droplet-interface bilayers (DIBs) have been

developed as powerful model membrane techniques for functional
studies of membrane pores and ion channels [1–6]. The lipid
bilayer in a DIB is formed when two small aqueous droplets
(100–300 nL each droplet) with a lipid monolayer at their water-
oil interface are brought into contact. Therefore, this method has a
number of advantages compared with other model membranes,
such as minimized amount of protein or label needed, reduced
time for experiment replication since the droplets can be easily
discarded and remade, easy control of lipid asymmetry [3, 4, 7],
and ability to perform secondary analysis [5] and bilayer networks
197
198 En-Hsin Lee
[8, 9]. However, similar to other existing model membranes, one

of the most challenging limitations in DIB is the incapability of
exchanging aqueous solutions rapidly while maintaining the stabil-
ity of the lipid bilayer. Although it is a common practice to
exchange solutions in planar lipid bilayer system, its milliliter cham-
ber volumes are too large for instantaneous delivery and removal of
a solute for the membrane due to the long diffusive length. The
mechanically fragile nature of the lipid bilayer also makes it impos-
sible to highly increase the flow rate for achieving instant response
of a membrane to its exchanging environment. As a result, attempts
have been made to reconstitute lipid bilayers, including DIBs, into
microfluidic systems for effective perfusions [6, 10–14], which
require high volumetric flow rate and minimum diameter of a
perfusion chamber. However, such systems either suffer from pres-
sure and flow disturbances that easily rupture lipid bilayers or
generally require more than 10 s for solution exchanges, which is
not yet an ideal perfusion rate for instantaneous delivery of a solute.
Here, we present a robust DIB-based perfusion system that has
been routinely used in our research. This perfusion system not only
possesses the advantages of conventional DIB, but also allows the
solution on one side of a lipid bilayer to be completely exchanged
within 1–2 s. As the perfusion flow in our system directs to the
plane of the lipid bilayer, it does not generate stagnant region in the
perfusing “droplet.” Therefore, prompt responses of a membrane
to the perfusions can be obtained. Moreover, the introduction of an
aqueous reservoir makes our perfusion device into an “open loop”
hydraulic system that is less susceptible to undesired pressure fluc-
tuations. This system, hence, shows remarkable stability that can
sustain continuous perfusions up to at least several hours.
One of the applications of our perfusion system is its capability
of switching the proton gradient (ΔpH) across the membrane on
and off iteratively by perfusing solutions with different pH values.
To demonstrate the competence of this perfusion device,
pH-dependent ion channel insertion and ΔpH-driven protein
translocation were performed. Two polypeptides of anthrax toxin
[15–17] were used in our experiments. One is protective antigen
(PA), a pore-forming protein that undergoes conformational rear-
rangement from its pre-pore structure to the transmembrane chan-
nel upon acidification [18, 19]. PA pore utilizes transmembrane
potential or proton gradient as driving force to translocate the
enzymatic components of anthrax toxin, one of which is named
lethal factor [20, 21], into the host cell cytosol via ratchet mechan-
isms [17, 22–26]. Here we used the N-terminal truncation form of
lethal factor (LFN), the minimal translocation-competent leader
domain of the lethal factor [27, 28], for our translocation experi-
ments. We show that by perfusing different buffers and protein
solutions sequentially to the lipid bilayer in our perfusion system,
initiation of PA pore insertion, binding of LFN to PA pores,
Rapid solution Exchange in a DIB 199
initiation of ΔpH-driven translocations, pauses, and reinitiations of

translocations could all be obtained and clearly identified from the
changes in ionic current. Overall, this DIB-based perfusion system
enables us to reconstitute or mimic the changing environment of
cellular membranes under a variety of controls, and thus enhance
the acquisition of delicate responses of ion channels to their sub-
strates or environments.
2 Materials
2.1 Perfusion Device 1. Flow chip: The construction of the flow chip design is shown in
Fig. 1 for fabrication (see Note 1). The flow chip includes a
smaller DIB perfusion port (diameter: 1.5 mm, depth: 4 mm)
and a larger fluid reservoir (diameter: 5 mm, depth: 4.5 mm).
The reservoir connects with the perfusion port through a
lateral channel and serves as a baffle to suppress any effect
from uneven flow rates in the device. The opening for infusions
Fig. 1 Device construction for DIB perfusion system. The aqueous chamber in the flow chip consists of four
openings, including (i) a DIB perfusion port, (ii) a large pressure control reservoir, (iii) an injection pore for
capillaries bundled together, and (iv) a drain. The pressure control reservoir is slightly higher than the
DIB perfusion port and is used to prevent pressure changes under the DIB. The bottom-right panel shows
the sectional diagram of the perfusing device. (Reproduced from ref. 32 as the author of the content.) Two
electrodes are used. A DIB droplet is suspended from the working electrode while the grounded electrode dips
in the reservoir. Solutions are perfused through the bottom dome of the DIB
200 En-Hsin Lee
points up toward the perfusion port, while the reservoir con-

nects to another opening for drain.
2. Capillary inlets: Bundle and glue six fused silica capillaries (e.g.,
outer diameter (OD): 280 μm, inner diameter (ID): 150 μm)
inside a larger tube (e.g., PFA tube with OD: 1/16 in., ID:
0.04 in.), which serves as the tubing sleeve for the connection
with the flow chip (see Note 2). The number of fused silica
capillaries in the bundle can be designated as needed.
3. A tube for outlet drain (see Note 3).
4. Tubing fittings to connect the inlet bundle and the outlet tube
into the flow chip (e.g., one-piece fingertight male nut, 10–32
coned for 1/16 in. OD).
5. Gastight syringes with Luer Locks.
6. Tubing fittings to connect each fused silica capillary to a gas-
tight syringe (see Note 4).
7. Tubing fittings to connect the outlet tube to a gastight syringe
(e.g., a one-piece fingertight male nut and a female to female
Luer adaptor).
8. Syringe pumps. One syringe pump should be able to withdraw
as drain.
2.2 DIB 1. Faraday cage (see Note 5).

Electrophysiology 2. Patch-clamp amplifier.
3. Micromanipulators.
4. Platform for holding the flow chip.
5. Alligator clips for holding the electrodes on the micromanipu-
lators. Solder each to an insulated cable and fit appropriate
connectors for the headstage of the patch-clamp amplifier.
6. Stereomicroscope.
2.3 Buffers 1. Basic buffer: 10 mM MES, 10 mM Tris–HCl, 100 mM KCl,

and Protein Solutions pH 8.5. Degas by vacuum for 3 h or overnight.
2. Acidic buffer: 10 mM MES, 10 mM Tris–HCl, 100 mM KCl,
pH 5.5. Degas by vacuum for 3 h or overnight.
3. Recombinant PA heptameric pre-pore purified as previously
reported [29]. Dilute to a final concentration of ~7 ng/μL in
the basic buffer before use.
4. Recombinant LFN purified as previously reported [28]. Dilute
to a final concentration of ~250 ng/μL in the basic buffer
before use.
2.4 DIB 1. Ag/AgCl electrodes prepared as previously described [2]. In

short, briefly melt one end of a ~2-cm length of 100-μm-
diameter silver wire over a flame to create a small silver ball of
~250 μm in diameter. Immerse the ball end of the wire in 20%
(wt/vol) sodium hypochlorite solution for 30 min. A gray
color indicates that AgCl coating has formed.
2. 2.5 mM 1,2-diphytanoyl-sn-glycero-3-phosphocholine
(DPhPC) vesicle solution in the basic buffer.
3. Approximately 3% (w/v) agarose (low melt) solution prepared
as previously described [2].
4. Hexadecane.
5. Approximately 0.2% (w/v) agarose solution prepared by
directly diluting the 3% agarose in water.
3 Methods
3.1 Perfusion Device 1. Apply a thin layer of ~0.2% (w/v) agarose solution to the inner
Assembly surface of the aqueous chamber (Fig. 1, blue part in the
and Preparation bottom-right panel). Allow the solution to dry out to form a
thin film of agarose on the inner surface of the chamber to
prevent hexadecane or air bubble from being trapped within
the channel (see Note 6).
2. Fix the inlet bundle upright into the perfusion port and the
drain tube into the pressure control reservoir of the flow chip
by corresponding adaptors (Fig. 1). The ends of the fused
capillaries should be only a bit lower than the DIB perfusion
port (Fig. 2a, DIB port).
3. Place and fix the flow chip in a Faraday cage (Fig. 2b). All tubes
extend out through a hole of the Faraday cage to connect with
gastight syringes.
4. Fill the infusion gastight syringes with buffers and protein
solutions, respectively (see Note 7). Avoid air bubbles in the
syringes. Up to six different solutions can be independently
introduced from the infusion syringes. We filled three syringes
with the basic buffer since there were only four solutions
required for LFN translocation experiment.
5. Connect each capillary to an infusion syringe through
corresponding adaptors. Slowly push the air out of each capil-
lary by infusing the solution from the syringe to the perfusion
port. During this process, carefully avoid and remove any air
bubble trapped within the conduits (see Note 8).
6. Fill a gastight syringe with the basic buffer. Connect the syringe
to the outlet tube as a drain. Slowly infuse the basic buffer from
the syringe to fill the outlet tube and the reservoir while
202 En-Hsin Lee
Fig. 2 Actual views of the DIB perfusion device. (a) Top view of the DIB perfusion port and the pressure control
reservoir with two Ag/AgCl electrodes under stereomicroscope. A DIB was formed between the droplet and the
perfusion port. (Reproduced from ref. 32 as the author of the content.) (b) Top view of a Faraday cage
enclosing the flow chip, including two micromanipulators with arms for holding the electrodes. The capillary
inlets and the drain tube thread out through a hole at the bottom of the Faraday cage. A US penny (19 mm) was
placed by the flow chip for scale
carefully avoiding trapping air bubbles within the conduit.

Ensure the syringe has enough drawing space for the
experiment.
7. Pipette the basic buffer to rinse (see Note 9) and fill the
U-shape aqueous chamber. Use a stereomicroscope to assist
with this process.
8. Place each syringe on a syringe pump outside the Faraday cage.
The drain syringe should be placed on the pump that with-
draws. Ground all syringe pumps to avoid electrical noise (see
Note 10).
9. Slowly pipette hexadecane to fill the top of the flow chip as an
oil bath (Fig. 1, yellow part in the bottom-right panel).
3.2 DIB Preparation 1. Prepare an agarose-coated electrode: Briefly dip the ball end of
an Ag/AgCl electrode into 90 C 3% (w/v) agarose. On cool-
ing, the small ball will be encased by a layer of solidified
agarose. Immediately submerge the agarose-coated ball on
the electrode into the oil bath to prevent dehydration.
2. Attach two Ag/AgCl electrodes, one must be agarose-coated,
to the micromanipulators so that the ball end of each electrode
can be suspended within the oil bath. Use an agarose-coated
electrode as the working electrode while the other one is
electrically grounded. Connect both electrodes to the patch-

clamp amplifier via the headstage.
3. Dip the grounded electrode into the large reservoir and the
working electrode into the perfusion port. Correct offset vol-
tages to near zero, which can be performed by applying a
triangular-wave potential to obtain outputs that are visually
vertical lines, and manually adjusting the compensatory ampli-
fier offset until the vertical lines are evenly spaced (see Note
11). Lift the working electrode up from the perfusion port
after compensating the offset.
4. Create a lipid monolayer-encased droplet on the working elec-
trode: Pipette 250 nL vesicle solution onto the agarose-coated
ball of the working electrode in the oil bath. The vesicle solu-
tion will adhere to the hydrogel and suspend as a droplet.
Incubate for ~2 min to allow the self-assembly of lipid mono-
layer (see Note 12). When monolayer forms, the droplet sags
from the electrode with flexible surface due to the reduction of
the surface tension.
5. Create a lipid monolayer-encased perfusion dome: Pipette
1.0 μL of vesicle solution to the top of the perfusion port of
the flow chip. Allow ~5 min for monolayer formation (see Note
13). The aqueous surface becomes slightly bulgy (see Note 14)
and flexible when monolayer forms.
3.3 Bilayer 1. Conventional electrical recordings of ionic current can be made

Formation using Ag/AgCl electrodes connected to patch-clamp amplifier
and Electrical (see Note 15).
Recording 2. Lowering the droplet onto the monolayer dome of the perfu-
sion port via micromanipulator. Under a stereomicroscope,
position the droplet carefully so that it gently “touches” the
dome (see Note 16) (Fig. 2a). Lipid bilayer can now be formed
spontaneously when the two monolayers are brought into
contact (see Note 17).
3. Monitor the bilayer formation via capacitance measurement
[2, 30] (see Note 18). Close the lid of the Faraday cage to
shield the system from external electrical noise (see Note 19).
Lipid bilayer formation can be observed generally within
1–2 min after the contact of the monolayers (see Note 20).
4. Adjust bilayer size by micromanipulation of the droplet posi-
tions while monitoring the capacitance (see Note 21).
3.4 Continuous 1. To exchange buffers, operate an infusing syringe pump and the
and Rapid Solution withdrawing syringe pump simultaneously in the same flow
Exchange rate so that the net volume of the solution in the aqueous
chamber remains constant (see Note 22). In our device, flow
rates up to 1 μL/s (where 1 μL is roughly the volume of the
204 En-Hsin Lee
Fig. 3 Stability test for a DIB underwent buffer exchange in the perfusion system. The blue dashed line and the
red solid line indicate the spans of perfusions of the basic buffer and the acidic buffer through the perfusion
dome of DIB in a flow rate of 0.5 μL/s. The top droplet of the DIB remains constant at the basic condition
throughout the test. Voltages were applied to the DIB membrane as indicated on top. The gray shaded bars
indicate triangular-wave voltages applied to the membrane (see Note 18), which generate square-wave
current outputs that are visually black blocks of currents. As the amplitude of the square-wave current
depends on the diameter of the lipid bilayer [30], the minimal fluctuation of its amplitude over time indicates
the high stability of the DIB during the perfusions. Applications of +50 and 50 mV generated no current,
indicating that there was no leakage on the membrane. (Reproduced from ref. 32 as the author of the content)
perfusion dome) could be achieved without any visible distur-

bance of DIB. The DIB membrane could be stable under
continuous flow up to at least dozens of minutes.
2. Before introducing proteins, test perfusions of buffers should
be routinely performed to ensure the stability of DIB (Fig. 3).
Exchange buffers by turns in a flow rate of 0.5 μL/s. If the
bilayer is stable during the buffer exchange, no current will be
observed when a potential is applied to the membrane (see
Note 23). Also, bilayer size observed will remain stable when
a triangular-wave potential is applied for capacitance measure-
ment (see Note 24).
3. Perfuse the bottom dome with the basic buffer to set the DIB
at a pH-symmetric condition.
4. Apply a small potential (Δψ Δψ droplet Δψ dome; Δψ dome
0 mV) at 20 mV to facilitate ionic current recording (see
Note 25).
Fig. 4 Ionic current changes of acidification-activated PA pore insertion and ΔpH-driven LFN translocations in
the DIB perfusion system. A small voltage ( 20 mV) was applied throughout both recordings. Different bars on
top indicate spans of perfusions through the bottom dome of the DIB, whereas the top droplet remains
constant at the basic condition (pH 8.5). (a) Before point (i), PA pre-pore introduced in basic condition (pH 8.5)
generated no current. Once the acidic buffer (pH 5.5) was introduced, PA pre-pores converted to pores and
inserted on the DIB membrane, which was observed as a rapid increase in current to 1000 pA (which
represented ~800 PA pores) at point (i). Excess proteins were rinsed away by the buffer and the current
became stable as observed at point (ii). From point (ii) to point (iii), the current stably switched between two
different states when two buffers were alternately perfused, indicating a fixed number of PA pores on the DIB
and a stable system sustained repeated perfusions. LFN peptides were then introduced in basic condition and
led to a partially blocked state as observed between points (iii) and (iv). As the acidic buffer was again
introduced to generate a pH gradient across the DIB membrane, translocation was initiated and observed as a
fully blocked state followed by a slowly increased current as shown between points (iv) and (v). The current
restored and could be stably switched again upon buffer exchange after point (v), indicating that the
translocation was complete. (b) Demonstration of rapid solution exchanges during a translocation. With PA
pores inserted on the DIB membrane, LFN peptides were introduced at (i). From points (ii) to (iii), two buffers
were rapidly switched without interruption as described in Note 22. As the pH gradient was switched on and
off accordingly, the translocation was iteratively interrupted until it proceeded to completion at point (iii).
(Reproduced from ref. 32 as the author of the content)
5. Perfuse protein solutions and buffers under a flow rate of

0.5 μL/s (Fig. 4). PA pore insertion, LFN binding, and trans-
location could be observed by monitoring changes in ionic
current.
206 En-Hsin Lee
4 Notes
1. Microchannel architectures of the flow chip can be designed

using CAD softwares (e.g., SolidWorks) and milled on a trans-
parent acrylic plastic with a subtractive computer numerical
control (CNC) milling machine (e.g., Roland MDX-40 3D
milling machine). We used two acrylic segments to assemble
into the microchannel. If such assembly of the acrylic segments
is needed, apply high-vacuum silicone grease on the joints to
prevent leakage before each use. Note that care should be taken
to avoid the contact between silicone grease and the liquid in
the microchannel.
2. Level off the openings of the fused silica capillaries toward the
DIB perfusion port when bundling. Seal at the end of the
tubing sleeve toward the DIB perfusion port with waterproof-
ing glue (e.g., Z-Poxy) so that no solution infuses into the
sleeve during the experiment. Have the capillaries long enough
to extend from the flow chip to the syringe pumps placed
outside the Faraday cage.
3. Have the length of the outlet tube enough to extend from the
flow chip to the syringe pump placed outside the Faraday cage.
Larger inner diameter is recommended for the outlet tube
(e.g., PFA tube with ID: 0.04 in.) so that less backpressure is
generated when draining.
4. Since most capillary tubing connectors are not designed to be
used with fused silica capillaries directly, extra adaptors for the
connections with gastight syringes may be necessary. For exam-
ple, we attached MicroTight fittings to each fused silica capil-
lary with MicroTight sleeve so that it was able to connect to a
short PFA tube, which in turn was connected to a gastight
syringe by a female to female Luer adaptor.
5. The metal box acting as Faraday cage should be capable of
enclosing micromanipulators, headstage for patch-clamp
amplifier, flow chip, and its platform. Bore a hole on the cage
to allow the tubing to extend through (Fig. 2b).
6. Air bubble is one of the most common issues that easily
encountered in microfluidics. Existence of air bubbles leads to
flow instability, undesired pressure, and longer pressure equili-
bration in microfluidics and is therefore detrimental to DIB
stability. Usage of gastight syringes (Subheading 2.1, item 5)
and degassed buffers (Subheading 2.3, items 1 and 2) are both
for purposes of avoiding air bubbles. We find application of
agarose on the surface of the microchannel significantly reduces
entries and trappings of air bubbles in the system.
7. If the concentration of a protein solution does not need to be

extremely precise in the experiment, draw some buffer into the
protein syringe after filling the protein solution and immedi-
ately continue to step 5 of Subheading 3.1. This will reduce the
amount of proteins that contaminate the aqueous chamber
during the device setup.
8. As there might be extra adaptors for capillaries and syringes,
care should be taken to avoid trapping any air bubble within the
connections.
9. The perfusion port must be rinsed thoroughly to remove any
residual proteins pushed out from capillaries. Always add buf-
fers into the perfusion port and remove extra buffers from the
reservoir side. Care should be taken to avoid pushing air into
the conduit when pipetting.
10. Syringe pumps and their power cords may highly increase
electrical noise in electrophysiology recording. See Note 19
with troubleshooting advice for high electrical noise. Lubricate
syringe pumps when needed.
11. When no bilayer acts as a capacitor in the circuit, the triangular-
wave voltage generates a triangular-wave current with the ver-
tices of triangles at infinity, which is visually parallel vertical
lines. If there are potential offsets between the electrodes, the
baseline of the triangular-wave current shifts upward or down-
ward and leads to different spacing between the vertical lines. It
is standard practice to compensate electrode offsets before each
experiment.
12. The droplet can be lifted to the oil-air interface of the oil bath
to facilitate monolayer formation.
13. Formation of lipid monolayer on the perfusion port is generally
slower than the droplet. Fusing the monolayer-encased droplet
on the working electrode into the perfusion port substantially
facilitates the formation of monolayer on the port. Make
another monolayer-encased droplet on working electrode for
DIB. Repeat this fusing and droplet-making processes several
times if needed.
14. If aqueous solution in the perfusion port keeps bulging out
after the monolayer forms, there might be undesired pressure
within the microfluidic system. Troubleshoot with the follow-
ing suggestions: disentangle the capillaries and tubes that are
twisted, ensure the buffers are degased, examine entire system
and remove any air bubble or hexadecane trapped within the
conduits, pipette buffer to flush the aqueous chamber from
perfusion port through lateral channel to the reservoir to
ensure the flow is unobstructed. If a syringe or connection
part needs to be disassembled for air bubble removal, the
syringe end of the tube should be held high to prevent
208 En-Hsin Lee
backflow of hexadecane into the tube. On the other hand, if

buffer is rinsed into the perfusion port, remove extra buffer
from the reservoir side with caution to avoid too-low aqueous
surface, which leads to hexadecane entering the conduits. After
troubleshooting, remake a lipid monolayer on the
perfusion port.
15. Our electrophysiology recordings sampled at 5 kHz with a
low-pass filter at 1 kHz using a digitizer.
16. If the monolayers are well formed, the two surfaces are able to
tolerate certain degree of compression. Nevertheless, the drop-
let and the dome will still fuse together if they are pushed
too hard.
17. If the two surfaces rupture easily, it is most likely that one
(usually the side of the dome) or both lipid monolayers are
insufficiently assembled. Increasing the incubation time for
monolayer formation or applying more vesicle solutions to
the perfusion port as described in Note 13 are common prac-
tices to solve the problem. Contamination may also cause
ruptures of the monolayers. Troubleshoot as previously
described [2]. On the other hand, if the formation of lipid
bilayer is delayed or does not occur at all after the two surfaces
are brought into contact, it may result from too-high concen-
tration of lipid, which leads to multilamellar lipid layers. Dis-
card the top droplet by lifting the electrode up from the oil
bath or pipette off the bulgy dome and remake the monolayers.
18. Apply a triangular-wave potential to the lipid bilayer to obtain a
square wave capacitive current output. The triangular-wave
potential we applied was 100 Hz in frequency. Use a model
cell with a known capacitance of 100 pF to calibrate the
triangular-wave potential so that the amplitude of the square
wave output equals to the capacitance. In this way, the ampli-
tude of the square wave output is positive correlated to the
diameter of the lipid bilayer.
19. As the amount of mechanical parts in the perfusion device is
highly increased comparing to conventional DIB, special care
should be taken to avoid electrical noises. Troubleshoot with
the following advices if electrical noise is high: check whether
the Faraday is grounded, check whether the syringe pumps are
grounded, eliminate vibrations and static electricity within and
surrounding the working stage, change the power cords of the
syringe pumps, or rearrange the electric cords surrounding the
working stage.
20. To accelerate bilayer formation, a potential of +50 mV can be
applied between the electrodes.
21. In our experiments, DPhPC bilayers were adjusted to approxi-

mately 350–500 pF in capacitance, which was monitored as
described in Note 18.
22. While the operation of the infusion pump and the withdrawing
pump should be stopped simultaneously, it is also a common
practice to switch two infusion pumps on and off alternately
while withdrawing solutions continuously to obtain rapid
switches of solutions (Fig. 4b).
23. If currents are observed with bilayer size staying stable during
buffer exchange, potential reasons are osmotic imbalance, the
presence of excess detergent, or contaminations of pore-
forming reagents.
24. When using syringe pumps to drive the microfluidic flow, issues
raised from air bubbles trapped in the conduits are most inten-
sified. Bilayer size shrinks or grows during the perfusions if
there is air bubble in the system. If bilayer size keeps shrinking,
growing, or fluctuating badly upon perfusions, separate the
droplet from the perfusion dome and troubleshoot as
described in Note 14. Also check leakage if bilayer size keeps
shrinking.
25. In our experiment, a potential of 20 mV was far below the
required threshold for potential gradient-driven LFN translo-
cation [29, 31] and did not contribute a significant transloca-
tion force. Instead, the small negative potential facilitated the
entry of positively charged N terminus of LFN into the PA pore
interior. Therefore, translocations observed were only driven
by the proton gradient.
Acknowledgements
This was a part of Ph.D. works advised by Dr. Matthew A. Holden

(University of Massachusetts Amherst). Purified PA pre-pore and
LFN peptides were generous gifts from Dr. R.J. Collier (Harvard
Medical School) and Dr. B.A. Krantz (University of Maryland
Baltimore).
References
1. Bayley H, Cronin B, Heron A et al (2008) 3. Li X, Huang J, Holden MA, Chen M (2017)
Droplet interface bilayers. Mol Biosyst Peptide-mediated membrane transport of mac-
4:1191–1208. https://doi.org/10.1039/ romolecular cargo driven by membrane asym-
b808893d metry. Anal Chem 89:12369–12374. https://
2. Leptihn S, Castell OK, Cronin B et al (2013) doi.org/10.1021/acs.analchem.7b03421
Constructing droplet interface bilayers from 4. Huang J, Lein M, Gunderson C, Holden MA
the contact of aqueous droplets in oil. Nat (2011) Direct quantitation of peptide-
Protoc 8:1048–1057. https://doi.org/10. mediated protein transport across a droplet-
1038/nprot.2013.061 interface bilayer. J Am Chem Soc
210 En-Hsin Lee
133:15818–15821. https://doi.org/10. 16. Collier RJ, Young JAT (2003) Anthrax toxin.
1021/ja2046342 Annu Rev Cell Dev Biol 19:45–70. https://
5. Fischer A, Holden MA, Pentelute BL, Collier doi.org/10.1146/annurev.cellbio.19.111301.
RJ (2011) Ultrasensitive detection of protein 140655
translocated through toxin pores in droplet- 17. Feld GK, Brown MJ, Krantz BA (2012) Ratch-
interface bilayers. Proc Natl Acad Sci U S A eting up protein translocation with anthrax
108:16577–16581. https://doi.org/10. toxin. Protein Sci 21:606–624. https://doi.
1073/pnas.1113074108 org/10.1002/pro.2052
6. Funakoshi K, Suzuki H, Takeuchi S (2006) 18. Collier RJ (2009) Membrane translocation by
Lipid bilayer formation by contacting mono- anthrax toxin. Mol Asp Med 30:413–422.
layers in a microfluidic device for membrane https://doi.org/10.1016/j.mam.2009.06.
protein analysis. Anal Chem 78:8169–8174. 003
https://doi.org/10.1021/ac0613479 19. Naik S, Brock S, Akkaladevi N et al (2013)
7. Hwang WL, Chen M, Cronin B et al (2008) Monitoring the kinetics of the pH-driven tran-
Asymmetric droplet interface bilayers. J Am sition of the anthrax toxin prepore to the pore
Chem Soc 130:5878–5879. https://doi.org/ by biolayer interferometry and surface plasmon
10.1021/ja802089s resonance. Biochemistry 52:6335–6347.
8. Maglia G, Heron AJ, Hwang WL et al (2009) https://doi.org/10.1021/bi400705n
Droplet networks with incorporated protein 20. Pannifer AD, Wong TY, Schwarzenbacher R
diodes show collective properties. Nat Nano- et al (2001) Crystal structure of the anthrax
technol 4:437–440. https://doi.org/10. lethal factor. Nature 414:229–233. https://
1038/nnano.2009.121 doi.org/10.1038/n35101998
9. Holden MA, Needham D, Bayley H (2007) 21. Duesbery NS, Webb CP, Leppla SH et al
Functional bionetworks from nanoliter water (1998) Proteolytic inactivation of MAP-kinase-
droplets. J Am Chem Soc 129:8650–8655. kinase by anthrax lethal factor. Science
https://doi.org/10.1021/ja072292a 280:734–737
10. Shao C, Sun B, Colombini M, DeVoe DL 22. Krantz BA, Finkelstein A, Collier RJ (2006)
(2011) Rapid microfluidic perfusion enabling Protein translocation through the anthrax
kinetic studies of lipid ion channels in a bilayer toxin transmembrane pore is driven by a pro-
lipid membrane chip. Ann Biomed Eng ton gradient. J Mol Biol 355:968–979.
39:2242–2251. https://doi.org/10.1007/ https://doi.org/10.1016/j.jmb.2005.11.030
s10439-011-0323-4 23. Brown MJ, Thoren KL, Krantz BA (2011)
11. Tsuji Y, Kawano R, Osaki T et al (2013) Charge requirements for proton gradient-
Droplet-based lipid bilayer system integrated driven translocation of anthrax toxin. J Biol
with microfluidic channels for solution Chem 286:23189–23199. https://doi.org/
exchange. Lab Chip 13:1476. https://doi. 10.1074/jbc.M111.231167
org/10.1039/c3lc41359d 24. Wynia-Smith SL, Brown MJ, Chirichella G et al
12. Lein M, Huang J, Holden MA (2013) Robust (2012) Electrostatic ratchet in the protective
reagent addition and perfusion strategies for antigen channel promotes anthrax toxin trans-
droplet-interface bilayers. Lab Chip 13:2749. location. J Biol Chem 287:43753–43764.
https://doi.org/10.1039/c3lc41323c https://doi.org/10.1074/jbc.M112.419598
13. Suzuki H, Tabata K, Kato-Yamada Y et al 25. Colby JM, Krantz BA (2015) Peptide probes
(2004) Planar lipid bilayer reconstitution with reveal a hydrophobic steric ratchet in the
a micro-fluidic system. Lab Chip 4:502–505. anthrax toxin protective antigen translocase. J
https://doi.org/10.1039/b405967k Mol Biol 427:3598–3606. https://doi.org/
14. Suzuki H, Tabata KV, Noji H, Takeuchi S 10.1016/j.jmb.2015.09.007
(2007) Electrophysiological recordings of sin- 26. Das D, Krantz BA (2016) Peptide- and proton-
gle ion channels in planar lipid bilayers using a driven allosteric clamps catalyze anthrax toxin
polymethyl methacrylate microfluidic chip. translocation across membranes. Proc Natl
Biosens Bioelectron 22:1111–1115. https:// Acad Sci 113:9611–9616. https://doi.org/
doi.org/10.1016/j.bios.2006.04.013 10.1073/pnas.1600624113
15. Young JAT, Collier RJ (2007) Anthrax toxin: 27. Wesche J, Elliott JL, Falnes PO et al (1998)
receptor binding, internalization, pore forma- Characterization of membrane translocation by
tion, and translocation. Annu Rev Biochem anthrax protective antigen. Biochemistry
76:243–265. https://doi.org/10.1146/ 37:15737–15746. https://doi.org/10.1021/
annurev.biochem.75.103004.142728 bi981436i
28. Cunningham K, Lacy DB, Mogridge J, Collier dynamic control of droplet interface bilayer
RJ (2002) Mapping the lethal factor and edema area. Langmuir 27:14335–14342. https://
factor binding sites on oligomeric anthrax pro- doi.org/10.1021/la203081v
tective antigen. Proc Natl Acad Sci U S A 31. Zhang S, Finkelstein A, Collier RJ (2004) Evi-
99:7049–7053. https://doi.org/10.1073/ dence that translocation of anthrax toxin’s
pnas.062160399 lethal factor is initiated by entry of its N termi-
29. Zhang S, Udho E, Wu Z et al (2004) Protein nus into the protective antigen channel. Proc
translocation through anthrax toxin channels Natl Acad Sci U S A 101:16756–16761.
formed in planar lipid bilayers. Biophys J https://doi.org/10.1073/pnas.0405754101
87:3842–3849. https://doi.org/10.1529/ 32. Lee E-H (2017) Mechanistic studies of proton
biophysj.104.050864 gradient-driven protein translocation by
30. Gross LCM, Heron AJ, Baca SC, Wallace MI droplet-interface bilayer techniques. Disserta-
(2011) Determining membrane capacitance by tion, University of Massachusetts Amherst
Chapter 16
Protein Transport Studied by a Model Asymmetric

Membrane Army Arranged in a Dimple Chip
Xin Li and Min Chen
Abstract
Reconstituted model membrane systems are powerful platforms to tackle interesting problems existing in
membrane biology. One of the barriers to efficient drug delivery, as therapeutics to disease, is the physical
membrane barrier of the cell. Small molecule can typically diffuse through the membrane; however,
biomolecules such as proteins or nucleic acids cannot passively diffuse the bilayer and thus much research
has been geared to engineering protein and/or nucleic acids delivery methods. One delivery method uses
cell penetrating peptides (CPPs). In this chapter, we introduce the model “membrane army” arranged in
dimple chip to study the delivery of β-galactosidase by a CPP known as Pep-1. This method uses droplet
interface bilayer technology (DIB). It accelerates the speed to screen through the working conditions in
CPP-assisted protein translocations because each chip provides dimples that can accommodate 36 pairs of
droplets or 18 model bilayers. We will use one of the successful translocation conditions of β-galactosidase
delivery as the example to illustrate how the model “membrane army” is built and utilized.
Key words Model membrane, Droplet interface bilayers, Screening method, Cell penetrating pep-
tides, Protein delivery, Translocation, Asymmetric bilayer, Protein transport
1 Introduction
Biological membranes are barriers of cells and their internal orga-

nelles and these barriers define the boarder of cells and control the
cellular material exchange with the surrounding environment. The
plasma membrane is usually a liquid crystalline lattice embedded
with transmembrane proteins, glycoproteins, carbohydrates, cho-
lesterol, as well as other molecules [1]. The structure of the plasma
membrane is roughly identified by the fluid mosaic model where
the embedded components have some range of movement within
the planar membrane, directing the cellular essential activities such
as molecular trafficking through the secretory and endocytic path-
ways [2]. However, the complexity of the biological membrane
makes it difficult to use whole cells to study biological membrane
processes and it is quite challenging to isolate meaningful data
213
214 Xin Li and Min Chen
considering the interferences from several pathway-networks.

Model membrane systems are developed with simpler and more
controllable parameters to make the experimental investigation
more tractable.
Model membranes have various forms and are classed by the
lipid content and production. Unilamellar vesicles, however, are
widely used as the model membrane in drug delivery research
[3]. They are spherical and contain a single bilayer of lipids. They
can be made either as small unilamellar vesicles (SUVs,
20–100 nm), large unilamellar vesicles (LUVs, 50–200 nm), or
giant unilamellar vesicles (GUVs, 1–100 μm) [4]. The high sur-
face/volume ratio of SUVs and LUVs make them compatible with
many assays such as binding assays and fluorescence spectrometric
assays [5, 6]. The diameter of GUVs is above the diffraction limit
and so GUVs are usually used in light microscope research. Another
popular membrane system is the supported lipid bilayer (SLB). An
example, the black lipid membrane, is the formation of a bilayer
across an aperture of a supported Teflon film [7]. Nanodiscs, nano-
scale phospholipid bilayers, are another example of a model mem-
brane. These could provide flexibility in lipid bilayer content and
stand out as better model membranes for membrane protein
study [8].
Another model membrane approach is based on immersing
two individual lipid-mono-layer-coated droplets in an oil phase
and then joining the two droplets to create lipid bilayer at the
interface [9]. This droplet interface bilayer (DIB) was proved to
stay stable for weeks. This model membrane system is not only
stable in structure but also compatible for functional membrane
protein insertion. For example, the pore-forming toxin, alpha
hemolysin (αHL), was expressed within the droplet by in vitro
translation and then detected to insert into the DIB by electrophys-
iology technique [10]. The size of the lipid bilayer interface can be
precisely controlled by adjusting the distance between inserted
electrodes [11]. In addition, chemical reagents on both sides of
the bilayer and the bilayer lipid composition can also be adjusted
according to the experimental needs [12]. This DIB system is the
prototype for the model “membrane army” discussed in this
chapter.
The dimple chip-based DIB takes advantage of the high stabil-
ity of DIBs and arranges pairs of droplets in the dimple chip filled
with hexadecane, allowing to screen different conditions at once. In
earlier designs of DIB, both droplets are inserted with electrodes
for tracing the current signal across the bilayer [13]. However, in
this chapter we are more interested in investigating the conditions
that underlie how a small 21-amino acid peptide, Pep-1, is able to
deliver a ~500 kDa protein directly across the membrane rather
than how the membrane is disrupted by translocation [14]. We,
therefore, removed the electrodes and used the fixed distance
Translocation of Protein Across an Asymmetric Membrane Bilayer 215
Fig. 1 Schematic illustrating the dimple chip membrane army methodology. Pep-1-mediated translocation of
proteins is driven by an electropotential difference provided by the asymmetric droplet interface bilayer.
(Reprinted (adapted) with permission from X. Li, J. Huang, M. A. Holden, and M. Chen, “Peptide-Mediated
Membrane Transport of Macromolecular Cargo Driven by Membrane Asymmetry,” Anal. Chem. 2017,
89, 12369–12374. Copyright 2017 American Chemical Society)
between the adjacent wells to consistently confine the bilayer size

from well to well [15].
We can usually screen 36 droplets in 6 columns in one dimple
chip and form 18 DIBs by pushing the adjacent droplets toward
each other. In this chapter, we will describe the “membrane army”
method applied to the Pep-1-assisted β-galactosidase (β-gal) trans-
location (Fig. 1). The components in each pair of droplets are
different, the droplet with Pep-1–β-gal inside is called the source
droplet and the other droplet is the capture droplet. The capture
droplet contains the β-gal substrate, resorufin-β-D-galactopyrano-
side (RG), which could be turned into fluorescent resorufin by
β-gal, so that the captured β-gal can be quantitated by the fluores-
cence intensity of the capture droplet.
2 Materials
Prepare all solutions with ultrapure water and analytical grade

reagents. All solutions are filtered through a sterile PES membrane
with 0.22 μm pore size to remove potential microbial and dust
contaminants.
2.1 Dimple Chip 1. Standard 24-well cell culture plate: it is made of polystyrene
and with high clarity (see Note 1).
2. Toothpaste and cotton swab: The toothpaste should be silica-

based to polish the wells (see Note 2).
3. Computer numerical control (CNC) machine (Roland
MDX-40).
4. 0.5 mm-radius ball-end mill.
5. Microscope glass slides.
2.2 LUVs 1. LUVs buffer: 150 mM NaCl, 10 mM HEPES, pH 7.4. Weigh

2.192 g NaCl and 0.596 g HEPES. Transfer them to 400 mL
beaker. Add about 200 mL water into the beaker. Mix the
reagents with water to dissolve them completely. Weigh
0.400 g NaOH into 10 mL water to make 1 M NaOH solu-
tion. Titrate the 1 M NaOH into the buffer to adjust the buffer
pH to 7.4 (see Note 3). Add water to make the volume of
solution up to 250 mL. Store the buffer at room temperature
(see Note 4).
2. PC solution: Dissolve 25 mg 1,2-diphytanoyl-sn-glycero-3-
phosphocholine (DphPC, Avanti Polar lipids) powder in
2.5 mL pentane to make 10 mg/mL DphPC (see Note 5).
Store in a vial at 20 C.
3. PG solution: Dissolve 25 mg 1,2-diphytanoyl-sn-glycero-3-
phosphocholine (DphPG, Avanti Polar lipids) powder in
2.5 mL pentane to make 10 mg/mL DphPG (see Note 5).
Another negatively charged lipid, PS (1,2-diphytanoyl-sn-gly-
cero-3-phospho-L-serine), can be substituted here for PG.
4. Containers: 4 Dram glass vials with black caps, 1.5 mL Eppen-
dorf tubes.
5. Equipment: Speedvac vacuum concentrator, vortex mixer,
37 C water bath, liquid nitrogen, Avanti 100 nm pore size
polycarbonate extrusion membranes, Avanti Mini extruder set
which includes a mini extruder, heating block, 2 O-Rings,
2 gas-tight syringes, 100 filter supports.
6. Cleaning reagents: Water, 70% ethanol, any commercial
detergent.
2.3 Translocation 1. Pep-1: An amphipathic 21 amino acid peptide (Ac-KETW-

WETWWTEWSQPKKKRKV-Cya) where Ac is an acetyl moi-
ety and Cya is a cysteamide. Add water to dissolve Pep-1
powder to make the final concentration 680 μM according to
manufacturer’s protocol. Aliquot the stock Pep-1 solution into
20 μL in each Eppendorf tube. This process was operated on
ice to conserve the Pep-1 structure/activity. Then flash freeze
the aliquots in liquid nitrogen and store at 20 C.
2. β-gal: Add water to β-gal so that the recommended stock
concentration of β-gal is 20 μM. Aliquot the stock β-gal
solution into 20 μL in each Eppendorf tube. This process was

operated on ice to conserve the protein structure/activity.
Then flash freeze all the aliquots in liquid nitrogen and store
at 20 C.
3. RG: Add water to RG and make 50 mM stock solution. Keep
RG solution in a brown vial and wrap it with aluminum foil to
avoid exposure to light. Store at 20 C.
4. Fluorescence Microscope (e.g., Nikon Eclipse T3i
microscope).
3 Methods
All procedures are operated at room temperature unless otherwise

specified.
3.1 Fabrication 1. Tape the individual well of the plastic plate to the workpiece
of the Dimple Chip table of the computer numerical control (CNC) machine.
2. Install a 0.5 mm-radius ball-end mill into the spindle head.
3. Create the dimple chips by the programmed design. Each
dimple chip contains 12 12 array of dimples. The dimple
arrays are 0.7 mm apart on center and the depth of each
dimples is programmed to be approximately 250 μm.
4. Polish the dimple chip with silica-based toothpaste and cotton
swab. This will minimize the defects created by the drill
machine and will also remove the residual plastic debris. A
smooth dimple surface enhances the stability of the DIB sys-
tem. Figure 2 illustrates a fully fabricated dimple chip.
5. Wash the dimple chip with ample detergent and water until it is
clean.
3.2 Large 1. Prepare DphPC lipids: Aliquot 169 μL of 10 mg/mL DphPC

Unilamellar Vesicles pentane solution as the lipids for the neutral leaflet of the
of 100 nm in Size model membrane. Add the solution to a 4 Dram glass vial.
2. Prepare DphPC/DphPG lipid mixture: Mix 118 μL of 10 mg/
mL DphPC pentane solution and 51 μL of 10 mg/mL DphPG
pentane solution. This lipid mixture is the composition of the
negatively charged leaflet of the model membrane. Add the
mixture in a 4 Dram glass vial.
3. Place the above two lipid solutions in SpeedVac Concentrator
with the vials open with no cap. Desiccate the lipids for 1 h until
no pentane remains (see Note 6).
4. Add 1 mL LUVs buffer into the vials with the dried lipids.
Resuspend the lipids in LUVs buffer thoroughly. The neutral
lipid suspension contains 2 mM DphPC while the negatively
Fig. 2 A CNC-generated dimple chip. A single 24-well with a 6 6 array of

dimples etched onto the bottom. Glass slide attached to bottom of dimple chip
for mounting to microscope. For size comparison a United States of America one
cent coin is used. Each pair of dimples are placed one well away from each other
charged lipid suspension mixture contains 1.4 mM DphPC and

0.6 mM DphPG. The suspensions appear translucent or milky.
5. Mix the suspensions by vortex mixer for 30 min, mixing every
3–5 min to fully hydrate the lipids. As the hydration proceeds,
the liquid becomes more and more transparent. Transfer the
suspensions into 1.5 mL Eppendorf tubes.
6. Prepare a liquid nitrogen bath and a 37 C water bath. Place the
suspensions in liquid nitrogen for 60 s or until frozen. Remove
from liquid nitrogen and place in 37 C water bath until
completely thawed. Vortex suspension to mix. Repeat the alter-
nating liquid nitrogen bath and 37 C bath for a total of five
rounds of freeze-thaw. These freeze-thaw cycles help to hydrate
the lipids into large multilamellar vesicles.
7. Fully assemble the mini-extruder set with a 100 nm pore-sized
membrane (see Note 7).
8. Equilibrate the mini-extruder set: Fill the LUVs buffer into one
of the gas-tight syringes and make sure the syringe plunger on
the other end is set to zero. Push the plunger of the syringe
filled with LUVs buffer inward so that the buffer automatically
fills into the empty syringe through the membrane. Repeat the
passing of LUVs buffer from one syringe to the other for a total
of two passes. Remove the syringe filled with LUVs buffer from
the mini-extruder. Discard the buffer from the syringe (see
Note 8).
9. Repeat step 8 two more times.
10. Fill the hydrated DphPC lipid solution into the same syringe
used to wash the mini extruder with the LUVs buffer.
11. Gently push the plunger of the filled syringe until the suspen-
sion is transferred to the alternate syringe on the other end
completely.
12. Gently push the plunger of the alternate syringe to transfer the
suspension back to the original syringe.
13. Repeat steps 11 and 12 until the suspension is clear. The clarity
of the solution indicates the homogeneity of the vesicles; where
the solution is clear, it contains a high homogeneous mix of
vesicles. In general, 21 passes through the membrane of the
mini-extruder is recommended for 2 mM lipid suspension. A
more concentrated lipid suspension will need more cycles of
back-and-forth extrusion to reach an ideal level of
homogeneity.
14. After 21 passes, the extruded lipid solution is filled in the
alternate syringe. This is to avoid any possible contamination
of vesicle aggregates.
15. Remove the syringe from the extrusion system and transfer the
vesicles into a clean 1.5 mL Eppendorf tube. Store at 4 C.
Vesicles are stable at 4 C for up to 2 weeks.
16. Clean the mini-extrusion set thoroughly in alternating water
and ethanol washes. Disassemble the mini-extruder and the
syringes. Rinse each piece with water and then with 70% etha-
nol twice. After the second 70% ethanol rinse, submerge all the
pieces in a 95% ethanol bath and let sit for 15 min. Rinse the
pieces with water and the 70% ethanol again. Let the final rinse
be with 95% ethanol. Allow the pieces to air-dry for 15–30 min.
17. Once dried, reassemble the mini-extruder to be used in the
preparation of the DphPC/DphPG vesicles. Repeat steps
7–15. After cleaning and drying the mini-extruder pieces and
syringes, store the pieces for future use.
3.3 Preparation 1. Fill the dimple chip with 400 μL hexadecane to create the oil
of the Dimple Chip bath for residing source and capture droplets.
in an Oil Bath Over 2. Take a clean microscope glass slide and position it right over
the Microscope the stage center of the microscope.
3. Drop about 30 μL hexadecane on the center of the glass slide
and then put the oil bath on top of the wet surface. Make sure
the glass slide and the dimple chip are sealed well by the oil and
no air bubbles in between.
3.4 Preparation 1. Thaw an aliquot of the stock Pep-1 and β-gal on ice.
of Source Droplets 2. Dilute Pep-1 to 40 μM by adding 1 μL of 680 μM Pep-1 stock
to 16 μL water.
3. Dilute β-gal to 500 nM by adding 1 μL of 20 μM β-gal stock to
39 μL water.
4. Preincubate a mixture containing 20 μM Pep-1 with 250 nM

β-gal (80:1 mole ratio) for 30 min at room temperature (25 C)
to form the complexes.
5. Dilute 2 μL complexes with 18 μL of PC solution by pipetting
the solution up and down to obtain complexes of 2 μM Pep-1
with 25 nM β-gal.
6. Pipette 0.3 μL of the Pep-1–β-gal complexes in PC solution in
the hexadecane reservoir to create a droplet.
7. Allow the droplet to sit in the oil reservoir for 30 min to allow
for monolayer formation around the droplet. Alternatively, use
a 20 μL pipettor to pipette the droplet back and forth in the oil
reservoir to accelerate the monolayer formed around the drop-
let (see Note 9).
8. Transfer the relaxed source droplet to the oil bath (see Note
10).
9. Repeat steps 4–6 to create more source droplets.
3.5 Preparation 1. Dilute the stock 50 mM RG with water to 500 μM. For
of Capture Droplets example, mix 1 μL 50 mM RG stock with 99 μL water.
2. Resuspend 2 μL diluted RG with 18 μL PC/PG solution.
3. Pipette 0.3 μL of the RG/PC/PG solution back and forth in
the hexadecane reservoir to create a complete PC/PG mono-
layer as described in step 5 in Subheading 3.4.
4. Transfer the relaxed capture droplet to the oil bath with one
well away from the source droplet.
5. Repeat steps 3 and 4 to create more capture droplets.
3.6 Translocation 1. Bring the source and the capture droplets into adjacent dimples
in Dimple Chip on the dimple chip (Fig. 3a). Push the source droplets with a
clean pipette tip to make contact with the adjacent capture
droplets to form the lipid bilayers (Fig. 3b) (see Note 11).
2. Once the bilayer is formed, the meniscus will stop growing
bigger (Fig. 3c). The asymmetric bilayer will drive Pep-1 to
carry β-gal across the bilayer (see Note 12).
3. Incubate the droplet bilayers for at least 2 h. Take a dark field
image which shows the capture droplet is bright, indicating
that β-gal has translocated successfully into the capture droplet
(Fig. 4) (see Note 13).
4. Use ethanol and water to rinse the chip alternatively and then
allow it to air-dry. Place in a drawer to avoid collecting dust.
Fig. 3 Setup of the droplet interface bilayers in a dimple chip. (a) A source droplet containing peptide enzyme
complexes and a capture droplet containing substrate were prepared by mixing each solution with vesicles
separately. (b) The droplets come into contact in a dimple chip under hexadecane and formed a fixed bilayer
based on the parameters of the dimple. (c) Multiple droplet interface bilayer pairs were arranged on a dimple
chip for parallel translocation experiments. Lipid vesicles within each droplet supplied lipids to form mono-
layers that annealed to create a bilayer membrane at the contacting region. The white dashed lines highlight
one row of dimples on the dimple chip. (Adapted with permission from X. Li, J. Huang, M. A. Holden, and
M. Chen, “Peptide-Mediated Membrane Transport of Macromolecular Cargo Driven by Membrane Asymme-
try,” Anal. Chem. 2017, 89, 12369–12374. Copyright 2017 American Chemical Society)
3.7 Calibration 1. As a control, place a capture droplet and a source droplet into
and Turnover Rate the dimple chip but do not make contact between the two
Analysis droplets. Measure the fluorescence increase of the capture
of Pep-1-Assisted droplet over a 2 h period. This baseline fluorescence represents
β-Gal Translocation the spontaneous breakdown of RG.
2. Calibration curve for amount of β-gal translocated: To calibrate
the β-gal turnover rate, mix various concentrations of β-gal
ranging from 0.1 to 2 pM, for example, with a constant
Fig. 4 Images of the DIBs taken by the microscope. (a) Bright field image that
two droplets formed a bilayer in the contacting region. (b) Dark field image of the
source droplet (dark upper) containing Pep-1–β-gal complex and the capture
droplet (white lower) containing the β-gal substrate, resorufin-β-D-
galactopyranoside. After 2 h incubation translocated β-gal hydrolyzes the
substrate in the capture droplet into the fluorescent product, resorufin, which
is seen as a bright white droplet
concentration of 50 μM RG and 1.8 mM DPhPC vesicles in

LUVs buffer (see Note 14).
3. Submerge 0.3-μL of the β-gal/RG/vesicle mixture in hexade-
cane and image the fluorescence intensity of the droplet over a
period of 2 h using a Nikon Eclipse T3i microscope with an
excitation wavelength of 561 nm. The β-gal catalyzed release of
resorufin product will typically remain constant after 2 h has
elapsed as the fluorescence intensity does not change.
4. Plot the fluorescence intensity of the droplets after every
30 min to generate Fig. 5a. Determine the slope from Fig. 5a
for the β-gal turnover rate. Then plot the slope (fluorescence/
min) for each β-gal concentration used to generate Fig. 5b.
5. Use Fig. 5b as a standard curve for your translocation experi-
ments to quantify the translocated β-gal enzyme. Remeasure
the standard curve periodically to verify the activity of the β-gal
enzyme.
4 Notes
1. The fluorescence microscopy-based assay requires the chip to

have high transparency so that when the light transmitted
through the chip, the fluorescence intensity will not be com-
prised by a large degree.
Fig. 5 Calibration curve of β-galactosidase activity. (a) A series of β-gal of different concentrations were mixed
with 50 μM resorufin-β-D-galactopyranoside (RG) substrate and LUVs buffer. The fluorescence signal of the
enzymatic reaction product (resorufin) was recorded over 2 h. (b) The β-gal turnover rates derived from the
slope of fluorescence intensity vs. time in (a) is plotted against the β-gal concentration. The data was fitted
with a linear regression model to derive the β-gal calibration curve. (Adapted with permission from X. Li,
J. Huang, M. A. Holden, and M. Chen, “Peptide-Mediated Membrane Transport of Macromolecular Cargo
Driven by Membrane Asymmetry,” Anal. Chem. 2017, 89, 12369–12374. Copyright 2017 American Chemical
Society)
2. The machine will probably create defects on the dimple chip;

therefore a silica-based toothpaste provides a gentle abrasive to
smoothen any rough patches.
3. If the pH is close to 7.4 yet still a little bit lower than 7.4,
further dilution of 1 M NaOH is recommended to avoid the
pH being over adjusted.
4. No sodium azide is added to this buffer. For long-term usage,
the buffer is recommended to be stored at 4 C.
5. Pentane is volatile with a low boiling point. As the liquids exist
in equilibrium between the gas and liquid states, pre-wetting
the pipette tip could saturate the gas state of pentane inside the
tip and thus give a more accurate volume of pentane when
pipetting.
6. An alternate way to dry the lipids is to coat the bottom of a test
tube and dry the lipids under a gentle stream of pressurized air.
This is an easier way and takes just a few minutes to completely
dry lipids. The challenging part is to control the speed of the air
and the position of the test tube well so that the liquid does not
splash out of the mouth of the test tube. It is highly recom-
mended to wear goggles to avoid any spilled liquid.
7. The mini-extruder set also includes an extruder stand, which
could fix the extruder apparatus inside and equilibrate the
temperature with the heating block. However, the transition
temperatures from gel to liquid for either DphPC or DphPG

are way beyond the working temperature range [16]. We did
not use the extruder stand and the heating block.
8. This step is not only to rinse the extrusion system with the
LUVs buffer but also to check the potential extruder leakage.
9. The droplet might split into small droplets while moving in and
out of the pipette tip; therefore a 20–200 μL pipette is recom-
mended for this step. Gentle pipetting skills are highly
recommended.
10. Before transferring the source droplet into the oil bath, the
droplet needs to be rinsed in the hexadecane to avoid small
buds attached to the source droplet being transferred to the oil
bath. Because these small buds might swim around in the oil
bath and bring contamination to the later translocation
experiments.
11. It is important to maintain the same ionic strength of both
droplets to prevent the creation of a chemical potential across
the bilayer. Using the buffer and reagent conditions described,
the ionic strength is kept the same at 0.077 M.
12. Droplets need to be relaxed first, otherwise there is no way to
form a bilayer. Unrelaxed droplets will immediately fuse into
one once they meet each other.
13. If no translocation is observed, you can alter the bilayer con-
ditions such as lipid concentration, or the lipid ratio mix of the
capture or the source droplet. Also, ensure whether the activity
of the β-gal enzyme is retained.
14. Prepare the different concentrations in parallel, adding each
reagent to the tube one at a time or use a multichannel pipet-
tor. Add the RG last and image the droplets within 60 s of
mixing all the reagents.
Acknowledgements
This work was supported by the US National Science Foundation

under Grant No. 1253565.
References
1. Benga G, Holmes RP (1984) Interactions Canesi S, Mauzeroll J (2016)
between components in biological membranes Ferrocene-modified phospholipid: an innova-
and their implications for membrane function. tive precursor for redox-triggered drug deliv-
Prog Biophys Mol Biol 43:195–257 ery vesicles selective to cancer cells. Langmuir
2. Bloch KE (1983) Sterol, structure and mem- 32:4169–4178
brane function. Crit Rev Biochem 14:47–92 4. Keller H, Worch R, Schwille P (2013) Model
3. Noyhouzer T, L’Homme C, Beaulieu I, membrane systems. Humana Press, Totowa,
Mazurkiewicz S, Kuss S, Kraatz H-B, NJ, pp 417–438
5. Saksena S, Wahlman J, Teis D, Johnson AE, 11. Holden MA, Needham D, Bayley H (2007)
Emr SD (2009) Functional reconstitution of Functional bionetworks from nanoliter water
ESCRT-III assembly and disassembly. Cell droplets. J Am Chem Soc 129:8650–8655
136:97–109 12. Hwang WL, Chen M, Cronin B, Holden MA,
6. Struck DK, Hoekstra D, Pagano RE (1981) Bayley H (2008) Asymmetric droplet interface
Use of resonance energy transfer to monitor bilayers. J Am Chem Soc 130:5878–5879
membrane fusion. Biochemistry 13. Huang J, Lein M, Gunderson C, Holden MA
20:4093–4099 (2011) Direct quantitation of peptide-
7. Mueller P, Rudin DO, Tien HT, Wescott WC mediated protein transport across a droplet–in-
(1962) Reconstitution of excitable cell mem- terface bilayer. J Am Chem Soc
brane structure in vitro. Circulation 133:15818–15821
26:1167–1171 14. Morris MC, Depollier J, Mery J, Heitz F,
8. Yeh V, Lee TY, Chen CW, Kuo PC, Shiue J, Divita G (2001) A peptide carrier for the deliv-
Chu LK, Yu TY (2018) Highly efficient trans- ery of biologically active proteins into mamma-
fer of 7TM membrane protein from native lian cells. Nat Biotechnol 19:1173–1176
membrane to covalently circularized nanodisc. 15. Li X, Huang J, Holden MA, Chen M (2017)
Sci Rep 8:13501 Peptide-mediated membrane transport of mac-
9. Bayley H, Cronin B, Heron A, Holden MA, romolecular cargo driven by membrane asym-
Hwang WL, Syeda R, Thompson J, Wallace M metry. Anal Chem 89:12369–12374
(2008) Droplet interface bilayers. Mol Biosyst 16. Hung WC, Chen FY, Huang HW (2000)
4:1191–1120 Order–disorder transition in bilayers of diphy-
10. Syeda R, Holden MA, Hwang WL, Bayley H tanoyl phosphatidylcholine. Biochim Biophys
(2008) Screening blockers against a potassium Acta Biomembr 1467:198–206
channel with a droplet interface bilayer array. J
Am Chem Soc 130:15543–15548
INDEX
A CanDo .......................................................................41, 45
Capsular polysaccharide (CPS)............................... 63, 64,
Actinia Fragacea............................................................... 4 72, 74, 75
Aerolysin ............................................................................ 3 Carbamidomethylation ................................................... 91
Agarose salt bridge....................................................54, 55
Caspases ......................................115, 118, 119, 123, 124
Air table ......................................................................... 119 CASTp .................................................160, 163, 167, 168
Alpha-helix barrel ......................................................67, 73 Cathode buffer ................................................... 14, 16, 17
Alpha-hemolysin ..................................................... 95, 96,
Cation selective .................................................. 27, 30, 42
136, 145–155, 214 Centricon..................................................... 118, 121, 122
Ampicillin .......................................... 4, 6, 13, 15, 68, 69, Centrifugation .................................. 6, 8, 15, 16, 82, 120
78, 82, 116, 119, 137, 138
Charge selectivity ............................................... 19, 24–27
Anion exchange...............................................79, 83, 118, CHARMM-GUI .................................................. 160, 164
120, 122, 137, 140 Chelator ........................................................................... 44
Antibodies ........................................ 80, 81, 88, 136, 145
Chemical modification .................................................... 12
Antimicrobial peptides (AMPs)...................................... 19 Chloramphenicol.................................................... 98, 101
Apertures ..............................................21, 23, 24, 29, 54, Chloroform ............................................35, 37, 147, 149,
55, 70, 79, 85, 86, 92, 124, 125, 128, 130–132, 150, 152, 189, 190
141, 146–154, 214
Cholesterol ..........................................35, 38, 39, 42, 213
Artificial bilayers ...............................................51, 85, 187 Circular dichroism (CD) .........................................21–25,
Atomic force microscopy (AFM) ................................... 44 28, 30, 31, 65, 73
AutoDock ..................................................................64, 67
Cis .........................................................11, 12, 24, 26, 27,
Autoradiography .................................................... 99, 102 30, 31, 52, 55–59, 61, 70, 102, 104, 125, 126,
Average residence time ................................................... 28 140, 142, 146, 149–154, 193
Avidin.........................................................................81, 88
ClpXP............................................................136, 145–155
Axopatch 200B .................................................21, 54, 65, CMake ........................................................................... 160
69, 70, 81, 99, 102, 119, 125, 126, 130, 131, Coarse-grained theoretical model ................................ 171
141, 147, 189, 195 Computer numerical control (CNC)..........................206,
216, 217
B
Conductance .....................................................20, 24, 25,
Bacterial transformations ....................................... 64, 119 28, 29, 33, 39, 42–44, 52, 55–58, 60, 61, 88, 104,
β-barrel.................................................................. 3, 19, 52 112, 151, 153, 154, 168
Binary and ternary complexes ............................... 98, 105 Confocal laser scanning microscope .............................. 38
Biotin ............................................................................... 91 Conformations .........................22, 29, 72, 111, 161–168
BL21 (DE3) .............................4, 6, 13, 15, 90, 119, 136 Constriction sites ........................................ 67, 68, 73, 77
Black lipid membrane ................................................... 214 Corynebacterium jeikeium ........................................20, 22
Blockages ............................. 25, 28, 30, 31, 64, 153, 194 Coverslips ............................................................... 37, 150
Blue-native polyacrylamide gel electrophoresis Current blockades .................................................. 12, 105
(BN-PAGE) ......................................................... 17 Cy3-labeled ...............................................................35, 38
Blue native sample buffer (BN SB) ................................ 14 Cyclodextrins (CDs) ....................................21–23, 28, 31
BOOST C++ ................................................................. 160 Cysteine label .................................................................. 58
Bradford assay .............................. 14, 16, 80, 83, 91, 122 Cytolysin A (ClyA).................................. 3, 11–17, 95, 96
C D
C++ ....................................................................... 160, 166 Delrin .......................... 21, 54, 55, 65, 99, 102, 119, 138
CaDNAno .................................................................41, 45 Dessicator ........................................................................ 36
https://doi.org/10.1007/978-1-0716-0806-7, © Springer Science+Business Media, LLC, part of Springer Nature 2021
227
NANOPORE TECHNOLOGY: METHODS AND PROTOCOLS
228 Index
Digidata digitizer ................................................... 99, 102 Ferric hydroxamate uptake component A
Dimple chips......................................................... 213–224 (FhuA)....................................................... 3, 51, 95
1,2-Dioleoyl-sn-glycero-3-phosphocholine Fiber optic lighting ....................................................... 155
(DOPC) .................................................. 35, 37, 39 Fokker-Plank ................................................................. 177
1,2-Diphytanoyl-sn-glycero-3-phosphocholine Folded proteins ............................................................... 12
(DPhPC)....................................... 4, 7, 29, 39, 53, Fragaceatoxin C (FraC) ................................................ 3–9
55, 61, 65, 70, 79, 86, 119, 124, 125, 130, 137, Free energy minimization............................................. 171
141, 188–190, 201, 209, 216–219, 222, 224 Freeze-thaw ......................................................7, 190, 218
1,2-Diphytanoyl-sn-glycero-3-phosphocholine
(DPhPG)...........................................216–219, 224 G
1,2-Diphytanoyl-sn-glycero-3-phosphoethanolamine G++ compiler ................................................................ 160
(DPhPE) ................................................... 188–190
β-galactosidase...................................................... 215, 223
DNA .................................................3, 4, 6, 9, 11–13, 15, Gas tight syringe ........................................................... 218
33–46, 64, 69, 82, 90, 98, 101, 116, 118, 119, Gating frequency.......................................................87, 89
121–123, 127, 129, 135, 136
Gel extraction ..................................................... 21–23, 64
DNA motifs ...............................................................42, 43 Giant unilamellar vesicles (GUVs) ..........................35–39,
DNA nanotechnology ..............................................33, 45 44, 46, 187, 214
DNA origami
Git ........................................................160, 161, 164, 166
DNA origami based nanopore ...........................34, 40 Glass capillaries ....................................125, 138, 141, 149
DNase ........................................................ 4, 6, 14, 15, 90 Glass slides .........................................35–37, 44, 216, 218
Dodecyl-β-D-maltoside (DDM) .................................4, 8, Glycoproteins ............................................ 65, 72, 74, 213
9, 14, 16, 21–24
Goldman-Hodgkin-Katz equation................................. 26
Droplet interface bilayer (DIB).......................... 187–193, G-quadruplex .................................................................. 42
197–202, 204–208, 214, 217 Gram negative bacteria ................................................... 51
Dwell times....................................... 28, 71, 89, 105, 107
Gravity flow ................................................. 4, 7, 8, 14, 16
Dynamic light scattering (DLS) ..................................... 44
H
E
Headstage ............................................................ 125, 131,
Elastic moduli................................................................ 180 141, 152, 200, 203, 206
Electroformation ................................................ 35–37, 44 Helfrich free energy .................................... 173, 177, 181
Electroosmotic flow (EOF) ............................................ 96
Heptamers ..................................136, 137, 139, 142, 143
Electrophysiology..........................................8, 53–56, 98, Heteroheptamers ................................................. 100–102
102, 107, 111, 187, 193, 200, 207, 208, 214 Hexadecane ................................................ 21, 29, 53, 55,
Electroporation ................................................5, 6, 13, 15
65, 70, 79, 86, 119, 124, 130, 137, 141, 150,
Enrofloxacin ..............................................................59, 61 190, 191, 201, 202, 207, 208, 214, 218,
Erythrocyte membranes........................................ 99–101, 220–222, 224
139, 142, 143
His6-tag .......................................................................4, 13
Escherichia coli .......................................... 4, 6, 12, 13, 15, HOLEs ....................................................... 21, 29, 64, 67,
52, 54, 64, 69, 70, 72, 78, 82, 116, 136, 147 102, 125, 129, 140, 141, 146, 148, 149, 152,
Ethanol .................................................4, 7, 9, 21, 28, 36,
201, 202, 206
37, 83, 92, 98, 128, 130, 131, 137, 141, 147, Homogenize........................................................... 15, 102
149, 216, 219, 220 Homology model............................................... 65, 67, 68
Event durations ...................................................... 89, 112 Hydrophobic ............................19, 33, 39, 41, 42, 45, 67
Exopolysaccharide (EPS) ................................................ 63
Hydrophobic tags .............................................. 33, 39–46
Extinction coefficient ......................................... 7, 99, 140 Hypotonic lysis....................................................... 99, 101
Extracellular.................................... 52, 54, 57–60, 68, 78
I
F
Inclusion bodies ................................................ 78, 82–83,
Faraday cage ..............................22, 28, 54, 55, 138, 141,
89, 90, 118, 120, 122
151, 155, 189, 190, 192, 200–203, 206 Incubate.....................................................6–9, 15–17, 23,
Faraday’s constant ........................................................... 26 69, 82–86, 90, 91, 101, 102, 121–124, 129, 132,
Fast protein liquid chromatography 139, 140, 203, 220
(FPLC)............................................................... 118
Indium tin oxide (ITO) glass slide ................... 35, 37, 44
Index 229
Inoculate...................................................... 6, 15, 82, 119 Modeller ...................................................... 64, 67, 72, 73
In situ...................................................................... 64, 187 Molecular dynamics ................................. 39, 67, 72, 159,
Inter-event duration........................................................ 89 160, 162, 166, 167
In vitro transcription translation Monomers ....................... 4, 6–9, 12, 16, 17, 29, 69, 101
(IVTT) .....................................64, 69, 73, 98, 101 Montal and Muller technique ........................................ 23
Iodoacetamide .................................................... 80, 85, 91 Mycobacterium smegmatis porin A (MspA) ................... 3
Ion channels ...................... 3, 19, 39, 187–195, 197–199
Ionic current............................................... 11, 12, 26, 28, N
39, 52, 95, 96, 135, 140–143, 159, 199, 203–205
NAMD.................................................160, 163, 166, 167
Isopropyl-β-D-thiogalactopyranoside Nanodiscs ...................................................................... 214
(IPTG) .......................................... 4, 6, 13, 15, 78, Nanopores ............................. 3–9, 11–17, 33–46, 77–92,
82, 90, 116, 119, 137, 138, 142
95–113, 115–132, 135–143, 145–155, 193
I-V curve........................................................................ 104 Nanotechnology........................................................33, 45
Native running buffer ...............................................14, 16
K
Negative potential .................................................. 25, 209
Kd ............................................................................ 71, 106 Next generation DNA sequencing............................... 135
Kinetics ..........................70, 96, 104, 105, 107, 145, 177 Ni-NTA.................................................. 4, 7–9, 12, 14–16
koff .................................................................................... 71 Nuclease S1 .........................................118–120, 124, 127
kon .................................................................................... 71 NUPACK...................................................................42, 45
L O
Labeling efficiency.............................................. 84–85, 91 Octamers....................................................................69, 70
Laemmli ................................................................... 21, 22, Octyl glucopyranoside (OG)................................. 80, 118
65, 69, 72, 99, 102, 123, 124 OD600 ............................................. 6, 15, 70, 82, 90, 119
Large unilamellar vesicles (LUVs) ..............................214, Oligomerization ...........................................4, 7–9, 16, 17
216–218, 222–224 Oligomers ....................................... 4, 8, 9, 12, 17, 22–25
N,N- Dimethyldodecylamine-N-oxide Oligonucleotides ..................................34, 36, 40, 78, 81,
(LDAO) ............................................................. 5, 7 117–119, 127, 135–137, 140, 143
Lethal factor .................................................................. 198 Open probability .......................................................87, 89
Ligand binding..........................................................12, 96 Orientations.......................................................39, 41, 42,
Linux.............................................................................. 160 44, 51–61, 86, 87, 126
Lipid monolayers................................................. 125, 188, Osmolarity ....................................................................... 37
197, 203, 207, 208 Outer membrane proteins (Omps)
Liposomes........................................................... 4, 7, 9, 12 OmpC .......................................................................... 3
LissRhodPE ..................................................................... 38 OmpF.............................................................. 3, 51–61
Loop dynamics ................................................................ 78 OmpG......................................................3, 52, 77–92,
Low pass filter ........................................................ 23, 208 115–132, 159, 160, 163–166
Lysis ............................................... 4, 6, 9, 13, 15, 78, 82,
90, 118, 120, 122, 139, 143 P
Lysogeny broth/Luria broth .....................................4, 12 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine
Lysozyme..................................4, 6, 9, 14, 15, 78, 82, 90 (POPC) ................................................... 35, 37, 43
Partial torus .......................................................... 174, 175
M
Patch clamp ....................................................75, 126, 187
MDiSGro .............................................................. 164, 165 Patch clamp amplifier...........................21, 54, 69, 70, 81,
Mean residue ellipticity (MRE) ...................................... 22 99, 102, 119, 125, 189, 191, 193, 200, 203, 206
Membrane capacitance......................................... 102, 193 Pclamp ............................. 22, 24, 28, 54, 69–71, 81, 126
Membrane permeability............................................26, 27 Pentane .......................................... 4, 7, 9, 21, 53, 55, 65,
Membrane-spanning DNA.......................................33–46 70, 79, 86, 91, 119, 124, 131, 137, 141, 189,
Microfluidics........................................198, 206, 207, 209 216, 217, 223
Microfluidizer.................................................90, 116, 120 Pep-1........................................... 214–216, 218, 220, 222
Micromanipulators..................................... 189, 191–194, Peptide detection .............................................................. 4
200, 202, 203, 206 Peptide nucleic acid (PNA) ......................................40, 45
Mini-extruder .................... 189, 190, 216, 218, 219, 223 Peptide pores ...................................................... 19–31, 33
230 Index
Perfusions ............................................150, 151, 197–209 Refolding ........................... 80, 83–84, 91, 118, 121, 123
Periplasmic...................................... 52, 54, 58–60, 68, 73 Resorufin-β-D-galactopyranoside
Phosphorylation ..................................................... 97, 109 (RG) .........................................215, 217, 220–224
Physiological conditions ................................................. 65 RTV silicone coat ......................................................54, 55
Piezoelectric igniter ............................................. 137, 141
Pim kinases ............................ 97, 99, 100, 104, 111, 112 S
Pimtide ................................................................... 97, 111 Salmonella typhi...........................................................3, 13
Planar lipid bilayer (PLB) ..................... 51, 52, 54–56, 58 Sampling frequency...................................................23, 54
Polyethylene glycol (PEG) ...............................52, 54, 57,
Scaffolded DNA ........................................................40–42
58, 60, 61, 95, 96 SCOBY ................................................160, 161, 164, 165
Polymerase chain reaction (PCR) ....................36, 37, 64, SDS-polyacrylamide gel electrophoresis
68, 69, 73, 81, 82, 90
(SDS-PAGE)................................... 22–24, 58, 64,
Polysaccharide transporter.............................................. 20 65, 69, 72, 74, 75, 80, 83–85, 101–103, 120,
Polytetrafluoroethylene film (Teflon) ........................... 21, 122–124, 137, 139, 140, 142
23, 28, 29, 37, 119, 124, 125, 128, 130, 131,
Shell scripts ................................. 160, 162–164, 166–168
137, 141, 146, 148, 152, 153, 214 Signal to noise ................................................67, 111, 112
PorACj .......................................................................20, 22 Silicone glue ........................................119, 128, 138, 141
Pore forming protein ................................ 19, 52, 77, 198
Silver/silver chloride .............................................. 65, 141
Porins .......................................... 20, 22, 51–62, 159, 160 Single channel recording ................................... 23, 28, 99
Porphyrin...................................................................39, 42 Single channel search ................................................28, 89
Positive potential................................................ 25, 30, 70 Single-loop fragment ........................................... 161, 166
Potential energies ........................................ 163, 166, 167
Single molecule enzymology .......................................... 12
PPorA..................................................... 20, 22–28, 30, 31 Single molecules.....................................3, 11, 19, 20, 30,
Precipitation ...................................................................... 9 33, 69, 70, 75, 96, 111, 136, 137, 140, 145, 146
Pretzel ................................................................... 159–168
Single strand DNA.................................... 35, 41, 42, 127
PROPKA3 ............................................................ 163, 165 Size exclusion chromatography (SEC) ....................83, 84
Proteases ............................................................85, 96, 97, Solid state nanopores ...................................................... 45
99–101, 103, 116, 123, 129
Sonication ........................4, 6, 7, 9, 14, 15, 79, 138, 193
Protective antigen (PA) ......................198, 200, 205, 209 Spectrophotometer .......................... 5, 7, 13, 16, 65, 118
Protein analytes .................................................. 77–92, 96 Sphingomyelin............................................................... 4, 7
Protein assembly .................................................. 149, 201 Square wave ................................................. 102, 204, 208
Protein-DNA interaction................................................ 12
Ssra tag.................................................................. 136, 147
Protein kinase (PKA) ............................................. 96, 112 Staphylococcus aureus ......................................19, 136, 189
Protein-protein interaction............................................. 78 Starter culture................................... 4, 6, 13, 15, 82, 119
Protein purification ...................................... 6, 15, 16, 69,
Stereomicroscope ................................................ 189, 191,
83, 91, 104, 118, 122, 138 192, 200, 202, 203
Protein threading ................................................. 135–143 Streptavidin............................................81, 84, 85, 88, 91
Protein trapping .......................................................... v, 12
Substrates..............................................12, 25, 28, 30, 33,
Proteoliposomes................................................................ 7 59, 96–98, 100, 104, 105, 107, 115, 116, 118,
Proteomics..................................................................... 145 124, 129, 136, 143, 147, 151, 189, 193, 199,
Protomers ........................................................................ 11
215, 221–223
Proton gradient .................................................... 198, 209 Supported lipid bilayer (SLB) ...................................... 214
Pseudosubstrate..................... 97, 98, 100, 105, 109, 111 Surgical blade ............................................................14, 17
PyMOL ............................................................... 60, 64, 67
T
Q
Thiols ............................................ 54, 56, 57, 59, 60, 121
Quantification ............................................................... 136
Tobacco etch virus (TEV) protease ............................. 100
Quartz cuvettes .........................................................20, 22 τoff .................................................................................... 71
QuB ................................................................99, 105, 109 τon .................................................................................... 71
Toothpaste................................................... 216, 217, 223
R
Trans ................................................ 4, 11, 12, 24, 25, 27,
Recombinant ........................................... 12, 15, 137, 200 31, 52, 55–59, 61, 70, 87, 97, 102, 104, 105, 107,
Reconstitution ............................................ 24, 29, 30, 39, 109, 112, 125, 126, 136, 140, 146, 148–150,
54, 56, 102, 104 152–154, 193
Index 231
Transient intermediates .................................................. 11 V
Translocate .................................................. 143, 145, 198
Translocon ....................................................................... 64 Vesicles ..............................................34, 36–38, 171–183,
Transmembrane.......................................... 19–30, 38, 39, 188–191, 193, 201, 203, 208, 214, 217–219,
41, 42, 44, 51, 68, 198, 213 221, 222
Transmission electron microscopy (TEM) ..............34, 44 Vortex .................................. 7, 14, 15, 36, 121, 190, 218
Triangular potential waveform ..................................... 102
W
2-dipyridyl disulfide (2-PDS) .............118, 121–123, 129
Wza ..................................................................... 20, 63–75
U
X
Universal gas constant .................................................... 26
UV crosslinking............................................................... 44 X-ray crystal structure ...............................................65, 73

Nanopore Technology - Methods and Protocols-Springer US - Humana (2021)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nanopore Technology - Methods and Protocols-Springer US - Humana (2021)

Uploaded by

Copyright:

Available Formats

Methods in

Molecular Biology 2186

Monifa A.V. Fahie Editor

For further volumes:

Methods and Protocols

ISSN 1064-3745 ISSN 1940-6029 (electronic)

© Springer Science+Business Media, LLC, part of Springer Nature 2021

Amherst, MA, USA Monifa A. V. Fahie

PART I DESIGN AND PREPARATION OF BIOLOGICAL BASED NANOPORES

PART II SINGLE MOLECULE DETECTION AND ANALYSIS OF PROTEIN ANALYTES

5 Determining the Orientation of Porins in Planar Lipid Bilayers . . . . . . . . . . . . . . . 51

PART III COMPUTATIONAL ANALYSIS OF NANOPORE BEHAVIOR

12 Simulation of pH-Dependent, Loop-Based Membrane Protein

PART IV DROPLET INTERFACE BILAYER SYSTEM

ROBIN ABU-SHUMAYS • UC Santa Cruz Genomics Institute, University of California, Santa

Design and Preparation of Biological Based Nanopores

Preparation of Fragaceatoxin C (FraC) Nanopores

The understanding of single ion channels was revolutionized over

Fragaceatoxin C (FraC) nanopore [16] has been introduced in our

All solutions are prepared using deionized water (Millipore). Spe-

2.1 Transformation 1. Sterile Lysogeny broth (LB).

2.2 Purification 1. Lysis buffer: 0.015 M Tris–HCl, pH 7.5, 0.150 M NaCl,

2.3 Preparation 1. Dilution buffer (DB): 0.015 M of Tris–HCl, pH 7.5, 0.150 M

2.4 Oligomerization 1. 5% (w/v) N,N-Dimethyldodecylamine-N-oxide (LDAO).

3.2 Purification 1. Harvest cells in 50 mL Falcon™ tubes by centrifugation at

7. Wash 100 μL of Ni-NTA beads (~200 μL of Ni-NTA slurry)

3.3 Preparation 1. Solubilize 25 mg sphingomyelin and 25 mg DPhPC (1:1) in

3.4 Oligomerization 1. Add sphingomyelin:DPhPC liposomes to FraC monomers in a

3. Dilute the solubilized mixture 20-fold with dilution buffer

1. Concentrated HCl can be used first to get close to the required

6. Pellets can be combined in this step in a ratio of 1 mL lysis

1. Neher E, Sakmann B (1976) Single-channel Ca2+ binding reactions. J Gen Physiol

Preparation of Cytolysin A (ClyA) Nanopores

In a typical nanopore experiment, a potential is applied across a

trans entry, which allows the entry of proteins or dsDNA

2.1 Transformation 1. 1 L Lysogeny broth (LB). Weigh 10 g of bacto tryptone, 5 g of

bottle. Mix components in 900 mL water, then fill up to 1 L

2.2 For Purification 1. 1 L Lysis buffer: 15 mM Tris–HCl, pH 7.5, 150 mM NaCl,

2. 1 L Wash buffer (WB): 10 mM imidazole, 150 mM NaCl,

3.1 Transformation 1. Transform electrocompetent E. cloni® EXPRESS BL21 (DE3)

3.2 Purification 1. Harvest the bacterial cells by centrifugation in 50 mL centri-

Ni-NTA suspension) into a microtube and wash the resin by

3.3 Oligomerization 1. Measure the protein concentration using a spectrophotometer or

Fig. 1 Isolation of different ClyA-AS types separated on 4–20% acrylamide BN-PAGE

8. Allow the gel to run for approximately 1 h at 120 V until the

1. Make sure not to overheat by keeping samples on ice.

4. Meller A, Nivon L, Brandin E, Golovchenko J, proteins by selective external association and

Building Synthetic Transmembrane Peptide Pores

Key words Peptide, Pores, Bilayer, Conductance, Single-channel, Substrates

Membrane-spanning proteins are found in all major groups of

alpha-pore-forming toxins (α-PFTs) with diverse structures and

5. CD buffer: 8.2 mM sodium phosphate, 1.8 mM potassium

2.3 Single-Channel 1. Pre-painting solution: Use either 10 mg/mL hexadecane in n-

acquisition and analysis program (pClamp 10.6); Faraday cage

3.1 Peptide 1. Synthesize peptides using solid-phase peptide synthesis and

3.2 Circular 1. Determine the secondary structures of the peptides in different

3.3 SDS PAGE Gel SDS-polyacrylamide gel electrophoresis (SDS-PAGE) is used to

2. Visualize the proteins by staining the gel with standard Coo-

3.4 Single-Channel 1. Set up the single-channel electrical recording apparatus with

4. Fill cis and trans compartments of the bilayer chamber with

of pore-forming peptides into planar lipid bilayer under an

3.5 Determination 1. Obtain multiple insertions of at least 100 pPorA or pPorA

3.6 Charge 1. Acquire ion selectivity measurements using asymmetric salt

current through lipid bilayers at high voltages up to 200 mV

1. The authors recommend using ultrapure KCl (99.999% trace