Structure
Article
An Active Site Water Network
in the Plasminogen Activator Pla
from Yersinia pestis
Elif Eren,1 Megan Murphy,2 Jon Goguen,2 and Bert van den Berg1,*
1Program in Molecular Medicine
2Department of Molecular Genetics and Microbiology
University of Massachusetts Medical School, Worcester, MA 01605, USA
*Correspondence: bert.vandenberg@umassmed.edu
DOI 10.1016/j.str.2010.03.013
SUMMARY
The plasminogen activator Pla from Yersinia pestis is
an outer membrane protease (omptin) that is impor-
tant for the virulence of plague. Here, we present
the high-resolution crystal structure of wild-type,
enzymatically active Pla at 1.9 Å. The structure
shows a water molecule located between active
site residues D84 and H208, which likely corresponds
to the nucleophilic water. A number of other water
molecules are present in the active site, linking resi-
dues important for enzymatic activity. The R211 side-
chain in loop L4 is close to the nucleophilic water
and possibly involved in the stabilization of the oxy-
anion intermediate. Subtle conformational changes
of H208 result from the binding of lipopolysaccharide
to the outside of the barrel, explaining the unusual
dependence of omptins on lipopolysaccharide for
activity. The Pla structure suggests a model for the
interaction with plasminogen substrate and provides
a more detailed understanding of the catalytic mech-
anism of omptin proteases.
INTRODUCTION
Yersinia pestis, the causative agent of plague, evolved recently
from a much less virulent species, Yersinia pseudotuberculosis
(Bercouvier and Mollaret, 1984). Y. pestis is transmitted primarily
by the bite of infected fleas. Due to the rapidly ingested but minute
blood meal taken by fleas, and because Y. pestis infection can be
readily suppressed by adaptive immunity (Degen et al., 2007),
efficient transmission requires that the bacteria quickly reach
high densities in the circulation. Y. pestis utilizes several mecha-
nisms, both active and passive, to suppress and evade innate
immune responses. Some of these are shared with Y. pseudotu-
berculosis, while others are not. One important adaptation unique
to Y. pestis is the production of an outer membrane protease
known as Pla. Specific inactivation of this protease results in
a million-fold increase in LD50 for mice infected subcutaneously
(Sodeinde et al., 1992; Sebbane et al. 2006) and also greatly
reduces the virulence of Y. pestis following pneumonic infection
(Lathem et al. 2007; Agar et al. 2009). Pla has been shown to
Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved 809
degrade several potentially important mammalian proteins
in vitro (Yun and Morrissey, 2009; Suomalainen et al., 2007).
In addition, it has been shown to act as both an adhesin and inva-
sin, mediating tight binding to fibronectin and promoting uptake
of the bacteria by nonphagocytic cells in culture (Suomalainen
et al., 2007; Cowan et al. 2000). However, the only activity that
has been strongly associated with virulence is the cleavage (acti-
vation) of plasminogen by Pla to generate plasmin (Degen et al.,
2007; Beesely et al., 1967), an important blood protease that
degrades many blood plasma proteins, most notably fibrin.
Plasmin activity promotes dissemination of Y. pestis from the
primary infection site by inhibiting microabscess formation
(Degen et al., 2007; Sodeinde et al., 1992).
Pla belongs to a unique family of integral outer membrane
proteases known as omptins, which are widely distributed within
the family Enterobacteriaceae. Although omptins share high
sequence identity (50%–75%), their biological functions range
from a possible housekeeping role in Escherichia coli (Haiko
et al., 2009) to the important contribution to virulence in plague
described above. A medium resolution structure of only one
omptin, E. coli OmpT (50% identical to Pla) crystallized from
refolded, inactive protein, has been reported (Vandeputte-
Rutten et al., 2001). Although OmpT was originally classified
as a serine protease (Kramer et al., 2000a), mutagenesis exper-
iments and the crystal structure were inconsistent with this
mechanism (Vandeputte-Rutten et al., 2001; Kramer et al.,
2001). Furthermore, inhibitors of the major classes of prote-
ases are ineffective against omptins (Sugimura and Nishihara,
1988), complicating the development of Pla-based therapeutics.
Molecular dynamics studies suggested a novel catalytic mecha-
nism for omptins in which a water molecule is responsible for
nucleophilic attack on the substrate (Baaden and Sansom,
2004). While consistent with mutagenesis data, no direct evi-
dence for this mechanism exists. Here, we report high-resolution
crystal structures of wild-type Pla protease obtained from enzy-
matically active crystals in order to obtain detailed information
about the catalytic mechanism and to understand the structural
basis of substrate specificity of omptin family members.
RESULTS AND DISCUSSION
Description of the Pla Structure
Crystals of wild-type Pla and the inactive Pla mutant D86A were
obtained using the detergent C8E4. Remarkably, after overnight
 Structure
High-Resolution Structure of Y. pestis Pla

Table 1. Data Collection, Phasing, and Refinement Statistics for Wild-Type Pla and the Inactive D86A Mutant
Wild-Type Pla
OsCl3 (APS ID-23B) Native 1 (NSLS X-25) Native 2 (NSLA X-6A) D86A (APS ID-23D)
Data collection
Space group C2221 P6522 P6522 P21
Cell dimensions (a, b, c) 140.2, 243.3, 129.9 140.3, 140.3, 130.4 140.2, 140.2, 130.5 69.9, 127.8, 107.2,
90, 90.1, 90
Resolution (Å) 50–2.9 (2.95–2.90)a 50–1.9 (1.88–1.9) 50–2.3 (2.34–2.30) 40–2.54 (2.58–2.54)
Rsym (%) 12.9 (61.4) 6.4 (58.5) 10.3 (84.1) 9.8 (60.9)
I/sI 19.8 (3.1) 48.6 (4.7) 31.0 (3.7) 18.8 (1.9)
Redundancy 7.5 (7.5) 12.2 (10.5) 14.2 (12.9) 6.3 (4.2)
Completeness (%) 100.0 (99.9) 100.0 (99.9) 99.4 (99.9) 95.7 (79.2)
Phasing
Number of sites 3
SOLVE Z-score/FOM 12.2/0.25 (3.0 Å)
SHARP Pp/FOMcen, acen 1.11/0.26; 0.10 (2.9 Å)
Refinement
Resolution (Å) 20–1.9 20–2.3 20–2.55
Number of reflections 56,505 31,644 56,736
Rwork/Rfree (%)b 19.5/20.8 19.9/22.6 18.5/24.6
Rmsds
Bonds/angles 0.007/1.23 0.007/1.22 0.008/1.27
Number of atoms
Protein/detergent/water 2177/212/200 2173/194/135 9247/0/125
B-factors
Protein/detergent/water 31/57/47 38/65/49 54/—/49
Ramachandran plot
Favored/allowed/disallowed 93.7/6.3/0.0 91.1/8.5/0.4 83.8/15.6/0.6
Pp, phasing power.
a
Values in parentheses are for the highest resolution shell.
b
Rwork = SjFo–Fcj/SFo. Rfree is the cross-validation of R-factor, with 5% of the total reflections omitted in model refinement.

dialysis against C8E4 the Pla proteins elute as a broad peak close
to the void volume on analytical gel filtration chromatography
columns (see Figure S1 available online; compare with protein
purified in OG). This surprising result shows that crystals of
(membrane) proteins can form from high-molecular weight,
heterogeneous protein aggregates.
An initial model for wild-type Pla was generated from phases
obtained from a single anomalous dispersion (SAD) data set
of an osmium derivative (Table 1; Figure S2). This model was
used to solve two structures of wild-type Pla at high-resolution
(1.9 Å and 2.3 Å) by molecular replacement. Pla is approximately
70 Å long and forms a narrow b-barrel with an elliptical cross-
section. The barrel consists of 10 antiparallel b strands con-
nected by short periplasmic turns and five extracellular loops
(Figure 1). Residues E252–S269 in loop L5 are not visible in the
wild-type Pla structure, possibly due to autocatalytic activity
(Kukkonen et al. 2001) of Pla at residue K261. The entire Pla
molecule is visible in the structure of the active site mutant
D86A, solved at a resolution of 2.55 Å (Table 1). The D86A mutant
has very low activity (Figures 2A and 2B) and therefore does not
undergo autocatalysis. The structures of wild-type Pla and the
D86A mutant are very similar, with rmsd values of 0.45 Å (all
atoms). In addition, and as expected from the high sequence
homology (Figure 1C), the overall structure of Pla is similar to
that of OmpT, with an rmsd of 1.1 Å.
Active Site Water Molecules
The catalytic site residues D84, D86, D206, and H208 are
conserved in all omptins (Hritonenko and Stathopoulos, 2007)
and are located in a cleft on the extracellular surface of the
b-barrel (Figures 1, 3A, and 3B). D84/D86 form a couple and
are located on one side of the barrel, whereas D206/H208 form
the other couple, located on the opposite side of the barrel (Fig-
ures 3A and 3B). The minimum distance between both active site
couples is between D84 and H208 (4.7 Å). A water molecule
(W1) is present between the D84 carboxyl group and the H208
N32 atom (Figures 3A and 3B). This water molecule most likely
corresponds to the catalytic nucleophile that was previously
proposed based on molecular dynamics studies of OmpT
(Baaden and Sansom, 2004). The D206/H208 couple likely forms
a catalytic dyad that activates the nucleophilic water molecule
for attack on the carbonyl carbon of the scissile peptide bond.
While there are other enzymes that utilize a catalytic dyad for
activity (e.g., phospholipase A2) (Scott et al., 1990), the presence
of a catalytic dyad in a protease is unique to omptins. The crucial
importance of D206/H208 for catalysis is supported by

810 Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved
Structure
High-Resolution Structure of Y. pestis Pla

A L3 B L3 density as water molecules is supported by the lower resolution

L2 structure (2.3 Å) of wild-type Pla. Here, density is present only for
L4
L2 the W1 nucleophilic water and one additional water molecule
L4 (Figure S3), separated by 3.6 Å.
L1 Active Site
L1 Pla also has a water molecule (W2) located between D84 and
L5 S7 L5 D86 (Figures 3A and 3B; Figure S3), which is reminiscent of
LPS
Binding Motif aspartic proteases (Brik and Wong, 2003). However, mutagen-
esis studies for Pla show that mutations of D84/D86 result in
proteins that still have some residual activity (Kukkonen et al.,
Extracellular
2001) (Figures 2A and 2B), suggesting a supporting rather than
a central role for these residues in catalysis. This notion is sup-
Outer ported by the pH-activity profiles for Pla (Figure 2C) and OmpT
Membrane
(Kramer et al., 2000b), showing that omptins, unlike aspartic
C proteases, are inactive below pH 5.0. We therefore propose
C Periplasm that omptins should not be classified as aspartic proteases
N N (Suomalainen et al., 2007). D84 and D86 likely perform important
C structural roles, by coordinating the nucleophilic water molecule
S1 L1 either directly (D84) or indirectly via W2 (the distance between
10 20 30 40 50
1
1
A S S Q L I P N I S P D S F T V A A S T G M L S G K S H E M L Y - D A E T G R K I S Q L D W K I K N V Pla
S T E T L - - S F T P D N I N A D I S L G T L S G K T K E R V Y L A E E G G R K V S Q L D W K F N N A OmpT
W1 and W2 is 2.8 Å) (Figure 3B). In addition to the active site resi-
S2 S3 L2
dues, S99 and H101 are two residues located close to the active
60 70 80 90 100 site that have also been shown to be important for omptin activity
51 A I L K G D I S W D P Y S F L T L N A R G W T S L A S G S G N M D D Y D W M N E N Q S E - W T D H S S Pla
50 A I I K G A I N W D L M P Q I S I G A A G W T T L G S R G G N M V D Q D W M D S S N P G T W T D E S R OmpT (Kramer et al., 2000a). The Pla structure provides an explanation
S4 S5 L3 for the importance of S99 and H101, since they coordinate the
110 120 130 140 150
101 H P A T N V N H A N E Y D L N V K G W L L Q D E N Y K A G I T A G Y Q E T R F S W T A T G G S Y S Y N Pla active site residues D84 and D86 via bridging water molecules.
101 H P D T Q L N Y A N E F D L N I K G W L L N E P N Y R L G L M A G Y Q E S R Y S F T A R G G S Y I Y S OmpT

L3 S6 S7
160 170 180 190
152 N G - A Y - - - T G N F P K G V R V I G Y N Q R F S M P Y I G L A G Q Y R I N D F E L N A L F K F S D Pla
152 S E E G F R D D I G S F P N G E R A I G Y K Q R F K M P Y I G L T G S Y R Y E D F E L G G T F K Y S G OmpT
L4 S8
200 210 220 230 240
199 W V R A H D N D E H Y M - - R D L T F R E K T S G S R Y Y G T V I N A G Y Y V T P N A K V F A E F T Y Pla
203 W V E S S D N D E H Y D P G K R I T Y R S K V K D Q N Y Y S V A V N A G Y Y V T P N A K V Y V E G A W OmpT
L5 S10
250 260 270 280 290
248 S K Y D E G K G G T Q T I D K N S G D S V S I G G D A A G I S N K N Y T V T A G L Q Y R F
254 N R V T N K K G N T S L Y D H N N - N T S D Y S K N G A G I E N Y N F I T T A G L K Y T F
Figure 1. Overall Structure of Y. pestis Wild-Type Pla
(A and B) Backbone representation viewed from the side (90 rotated in B), with
b strands colored blue, loop1 salmon, loop 2 olive, loop 3 cyan, loop 4 lime,
and loop 5 purple. Cytoplasmic turns are colored wheat. The active site resi-
dues D84, D86, D206, and H208 (yellow carbons, blue nitrogens, and red
oxygens), as well as R211 (green carbons), are shown as stick models. The
residues of the putative LPS binding site are shown as magenta stick models.
The modeled C8E4 detergent molecules close to the LPS binding site are
colored orange. All figures were made using PYMOL (DeLano, W.L. The
PyMOL Molecular Graphics System, 2002, DeLano Scientific, Palo Alto, CA,
USA).
(C) Sequence alignment of Pla and OmpT with the observed secondary struc-
tures. Identical residues (red), active site residues and catalytic residues (blue)
and R211(Pla)/K217(OmpT) (green) are indicated. Pla secondary structure
elements (loops in specific colors, b strands in blue, and cytoplasmic turns
in wheat) are shown on top. Extracellular loops are labeled L1–L5 and trans-
membrane b strands are labeled S1–S10, with the dark blue regions indicating
the parts that are embedded in the outer membrane.
mutagenesis studies (Kramer et al., 2001; Kukkonen et al., 2001)
(Figure 2B) and by the pH-activity profile, which suggests
involvement of a histidine residue in catalysis (Figure 2C). In the
1.9 Å structure, additional electron density is present close to
W1, which we have modeled and refined as two additional active
site water molecules (Figure 3A). The short distances between
these three water molecules (2.1–2.2 Å) (Figure 3B) suggest the
presence of a mixed water model analogous to that observed
in, e.g., urease (Pearson et al., 2000). The assignment of this
A Model for the Interaction of Pla
with Plasminogen Substrate
S9 Pla cleaves the R561-V562 bond within the sequence PGRVVGG
located in a surface-exposed loop of plasminogen (Sodeinde
et al., 1992; Wang et al., 1998). Placing the scissile peptide
bond approximately at the position of the nucleophilic water
Pla
OmpT molecule, the distribution of residues within the active site cleft
strongly suggests that the arginine side chain of the substrate
will bind in a deep, negatively charged pocket formed by Pla resi-
dues E29, D204, D206, and E217 (Figure 4). The two valine resi-
dues C-terminal to the scissile bond will likely bind in the shallow,
hydrophobic pocket located on the other side of the plane
formed by the active site residues D84, H208, and the nucleo-
philic water molecule (Figure 4).
Another Pla residue that is likely to be important for interaction
with the plasminogen substrate is R211, which is located at the
tip of loop L4 and points inward to the active site. The structure
of OmpT did not give any clues as to the location of the analo-
gous residue K217 since it was disordered, presumably as
a result of the several mutations introduced to aid crystallization
(G216K/K217G). The arginine side chain in Pla is only 6 Å away
from the nucleophilic water molecule and within hydrogen bond-
ing distance (3.2 Å) of one of the additional active site waters
(Figure 3). The location of the arginine side chain makes this
residue very likely to interact with the plasminogen substrate.
We propose that the positive charge on the arginine side chain
could play a role in the stabilization of the oxyanion intermediate
during catalysis, analogous to the role of calcium in phospholi-
pase A2 (Scott et al., 1990). Consistent with this hypothesis,
substitution of R211 for alanine results in low levels of plasmin-
ogen cleavage (Figure 2B). Remarkably, the conservative mutant
R211K is also a poor plasminogen activator (Kukkonen et al.,
2001) (Figure 2B), indicating that the interaction of R211 with
the substrate is highly specific.

Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved 811
 Structure
High-Resolution Structure of Y. pestis Pla

A B

100 100
Plasminogen Activation (%)

Plasminogen Activation (%)

80 80

60 60

40 40

20 20

0 0
Pla Pla D86A D86A Pla D84A D84N D86A D86N H208N R211A R211K
Crystals Crystals

100
Plasminogen Activation (%)

80

60

40

20

0
4 5 6 7 8 9
pH

Figure 2. Plasminogen Activation

(A) Activity of b-OG purified wild-type Pla (Pla), wild-type Pla isolated from crystals (Pla Crystals), b-OG purified D86A mutant (D86A), and D86A mutant isolated
from crystals (D86A Crystals).
(B) Activity of membrane-bound wild-type Pla and active site residue Pla mutants. The activity of D84A and H208N mutants were at background levels.
(C) pH optimum for plasminogen activation. Plasminogen activation values of membrane-bound (,) and b-OG (A) purified wild-type Pla were determined at a pH
range 4.5–8.5. All plasminogen activation values in (A), (B), and (C) correspond to specific activities of proteins and are the mean of three independent measure-
ments carried out with at least two different preparations of purified proteins or membranes. The error bars represent the standard errors (SEM).

Activation of Omptins by Lipopolysaccharide (Ferguson et al., 1998), it has been proposed that omptins have
Omptins require LPS for enzymatic activity (Kukkonen et al., an LPS binding site composed of the residues Y134, E136, R138,
2001, 2004; Kramer et al., 2002) by a mechanism that is unclear. and R171 (numbering for Pla). The previously determined crystal
Based on the structure of the OM protein FhuA with bound LPS structure of E. coli OmpT at 2.6 Å resolution (Vandeputte-Rutten

812 Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved
Structure
High-Resolution Structure of Y. pestis Pla

A B

H101
H101 R211 H101 S99
R211 2.93 R211
2.27 3.05
3.17
2.93 2.65
2.15
S99 H208 S99 H208 2.53 3.29
D84 3.68
2.09 H208
W1 W1 2.50
2.67
D84 D84 2.93
W1 2.56
W2 D206 W2 D206
2.79
2.76
D86 D86 W2
2.47
D206
S41 D86
S41

Figure 3. The Active Site of Wild-Type Pla Contains a Network of Water Molecules
(A) Stereoview of the active site from the extracellular side. The active site residues are shown as stick models in green, whereas the other residues that are impor-
tant for activity are shown in yellow. Water molecules are shown as red crosses. Fo-Fc densities obtained by simulated annealing omit maps are shown as a blue
mesh. The nucleophilic water molecule W1, as well as the W2 water molecule between D84 and D86, is highlighted in boldface.
(B) Schematic diagram showing the hydrogen bonding network of water molecules and amino acids in the active site. Bond lengths of all H-bonds are shown in
blue.

et al., 2001) was obtained from protein refolded from inclusion
bodies in the absence of LPS and therefore corresponds to inac-
tive protein. In contrast, the Pla protein used for crystallization
was purified from the bacterial OM, using LDAO for extraction
and the first gel filtration column (Experimental Procedures).
Remarkably, while the purified wild-type protein used for crystal-
lization was only 5% active, wild-type Pla reisolated from the
crystals had full activity (Figure 2A). The low overall activity of
our purified protein is likely due to the extensive removal of
LPS from the protein by LDAO. Indeed, when LPS or variants
thereof are added to inactive, LDAO-purified Pla, reactivation
of the protein occurs (Figure 5A). The inactivating effect of
LDAO on Pla is exhibited by many detergents, but not by C8E4
or OG (Figure 5B). The fact that the protein in the crystals is fully
active could be explained by assuming that the crystals are
formed by a small population of LPS-bound Pla molecules that
exists in our preparation. This notion is supported by the fact
that there are five striking, parallel tubes of density close to the

-G-R-V-V-G-

A B

Arg211

* *

Figure 4. Putative Model for the Interaction of Pla with Plasminogen

(A) Surface view from the extracellular side, showing the charge distribution across the active site cleft. The putative binding sites for plasminogen residues are
indicated. The position of R211 is indicated.
(B) Stereoview of the active site cleft, with approximately the same orientation as in (A), showing residues E29, D204, D206, and E217 of the negatively charged
binding pocket as magenta stick models, the active site residues D84 and H208 as yellow stick models, and residues S99, H101, R211, Y148, Y150, F159, V175,
I166, and M210 as green stick models (oxygens red, nitrogens blue). The location of the nucleophilic water is indicated with an asterisk. (A) was generated using
the ABSP plug-in within PYMOL.

Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved 813
 Structure
High-Resolution Structure of Y. pestis Pla

A B Figure 5. Mechanism of Pla Activation by

100 100 LPS
(A) Effect of LPS on Pla activity. Plasminogen acti-
Plasminogen Activation (%)

Plasminogen Activation (%)

80
vation by Pla was measured in the presence of
80
KDO2-Lipid A (5:1; KDO2-Lipid A:Pla molar ratio),
E. coli LPS (8:1; LPS:Pla molar ratio) and Y. pestis
60 60 LPS (2.5:1; LPS:Pla molar ratio), after inactivation
of wild-type Pla by LDAO (Pla LDAO). The activity
40 40 of b-OG purified wild-type Pla (Pla) is shown as
a positive control.
(B) Effect of detergents on plasminogen activation
20 20
by Pla. Plasminogen activation by membrane-
bound Pla was measured in the absence of any
0 0 detergent (control) or after incubation with the
Pla Kdo2 Y. pestis E. coli Pla Control OG Tween C9E12 C8E4 DM DDM DHPC FC LDAO
LipA LPS LPS LDAO 20 12 detergents indicated in the figure. Plasminogen
activation values in (A) and (B) correspond to
C D specific activities of proteins. All values are means
of three independent measurements of activity
R171 R138 with at least two different preparations of mem-
branes. Error bars represent the standard error
E136 D84 (SEM).
(C) Closeup side view of the barrel of wild-type Pla,
H208 with the residues in the putative LPS binding site
shown as yellow stick models (nitrogens blue,
oxygens red). Bound acyl chains in the vicinity of
Y134 D206 the LPS binding site that may belong to a partially
D86
ordered bound LPS molecule are shown in
magenta, with Fo-Fc densities obtained by simu-
lated annealing omit maps shown as a blue mesh.
(D) Closeup of the superimposed active sites of
active Pla (yellow) and inactive OmpT (blue),
showing the conformational shift for strand S7
and H208 away from the active site in OmpT.

proposed LPS binding site in the Pla structure (Figure 5C), which tively small, these differences may be large enough to profoundly
could correspond to acyl chains of a partially ordered LPS mole- affect enzymatic activity. We propose that the activating effect of
cule. Alternatively, the tubes of density could correspond to C8E4 LPS consists of pushing strand S7 inward, generating an active
molecules used for crystallization. We consider this latter possi- site with the proper geometry. Importantly, the putative LPS
bility less likely, since the addition of C8E4 to inactive Pla in binding site is located on the same side of the barrel as H208
solution does not result in activation. Moreover, we directly (Figure 1).
confirmed the presence of an equimolar amount of LPS relative
to Pla in the crystals of the active, wild-type protein via the highly Substrate Specificity of Omptin Family Members
sensitive limulus amoebocyte lysate (LAL) assay (Table 2) (Anto- Despite being 50% identical in amino acid sequence and having
nicelli et al., 2004; Taylor et al., 1995), supporting our assignment the same active site residues, OmpT is a very poor activator of
of the tubular densities as acyl chains of a partially ordered LPS human plasminogen both in solution and in membranes (Fig-
molecule bound to Pla. Interestingly, wild-type protein purified in ure 6A) (Haiko et al., 2009; Kukkonen et al., 2001). This observa-
OG contains an excess of bound LPS (Table 2), whereas LDAO- tion raises the question as to what is the structural basis for the
purified protein contains very low amounts of bound LPS. This substrate specificity of omptin family members. A structural
result confirms that enzymatic activity of the wild-type protein comparison of Pla and OmpT (Figure 6B) shows that, although
is directly correlated with the presence of bound LPS, which in both structures are very similar within the membrane-embedded
turn depends on the type of detergent used during protein parts, there are significant differences in the extracellular loops,
extraction and purification. in particular loops L3, L4, and L5. In a previous study (Kukkonen
What are the LPS-induced structural changes that lead to
omptin activation? A superposition of the active site residues
of Pla and inactive OmpT lacking LPS or detergent close to the Table 2. LPS Detection in Wild-Type Pla Preparations and
putative LPS binding site shows very small (<0.3 Å) changes Crystals
for the aspartic acid residues in the active site (D84, D86, and Protein LPS:Protein Molar Ratio SE (%)
D206) (Figure 5D). In contrast, H208 (and neighboring residues Pla-OG 3.00 0.05
in strand S7) is shifted outward by 0.8-0.9 Å in OmpT, away
Pla-LDAO 0.06 0.0012
from the active site (Figure 5D). This shift results in increased
Pla Crystals 1.02 0.07
(0.4–1.0 Å) distances in OmpT between the catalytic dyad resi-
dues D206/H208 and between the two residues (H208/D84) All values are the mean of three independent measurements. Pla-OG,
wild-type Pla purified in OG; Pla-LDAO, wild-type Pla purified in LDAO.
that coordinate the nucleophilic water molecule. Though rela-

814 Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved
Structure
High-Resolution Structure of Y. pestis Pla
A B L3
100
Plasminogen Activation (%)
80
60
40
20 L2
0
Pla OmpT Pla OmpT L1
M M
Figure 6. Structural Differences between OmpT and Pla May Affect Plasminogen Activation
(A) Plasminogen activation of OG-purified Pla (Pla) and OmpT (OmpT), and of membrane-bound Pla (Pla M) and OmpT (OmpT M). Plasminogen activation values
are the mean of three independent measurements carried out with at least two different preparations of purified proteins or membranes. The error bars represent
the standard error (SEM).
(B) Superpositions of D86A Pla (blue) and OmpT (magenta) viewed from the extracellular side (left) and from the side (right), highlighting (with arrows) the longer L3
and L4 loops in OmpT that may be responsible for the poor plasminogen activation by this protease. The active site residues of Pla are shown in yellow (nitrogens
blue, oxygens red).
et al., 2001), OmpT was converted into a reasonably efficient
plasminogen activator (25% Pla activity) by substituting K217
in loop L4 with the corresponding arginine residue present in
Pla (R211; Figure 1C), combined with the removal of D213/
P214 (absent in Pla). The additional introduction of Pla loop L3
in this OmpT mutant results in a hybrid protein with 90% activity
of wild-type Pla (Kukkonen et al., 2001). Thus, loops L3 and L4
are important contributors toward substrate specificity, with
R211 in L4 being the single most important residue. A sequence
comparison between OmpT and Pla shows that the tips of loops
L3 and L4 are indeed among the most divergent areas between
Pla and OmpT, with the Pla loops being two and four residues
shorter, respectively, than the corresponding loops in OmpT
(Figure 1C). The L3 and L4 loops form the entrance to the active
site and are very likely to be involved in plasminogen binding.
Steric hindrance by the longer OmpT loops is therefore likely to
be an important reason for the poor plasminogen activating
capability of OmpT. Pla likely binds with high affinity to plasmin-
ogen (Km for plasminogen activation 120 nM), suggesting that
differences in the interaction surface (such as those existing
between Pla and OmpT) may therefore have a large effect on
the activity of the protease toward plasminogen. Unfortunately,
the very low activity of OmpT toward plasminogen does not allow
determination of Michaelis constants for plasminogen activation
by OmpT.
In summary, the current high-resolution crystal structure of
wild-type Pla provides a detailed view of the active site of an
omptin protease, highlighting the importance of water molecules
for catalytic activity, as well as identifying an unusual network of
water molecules linking the active site residues that explains
previous mutagenesis data. The protein within the crystals is
catalytically active, and a comparison with inactive OmpT sug-
gests that LPS activates omptins by inducing a subtle conforma-
Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved 815
L3
L4
L4
L5
L2
L5
L1
tional change for strand S7 including the active site residue
H208. Due to its surface localization and critical importance
in both bubonic and pneumonic plague (Sodeinde et al., 1992;
Sebbane et al., 2006; Lathem et al., 2007; Agar et al., 2009),
Pla is potential target for development of novel therapeutics,
as are the omptins of other pathogenic enterobacteria. Explora-
tion of this potential and understanding of the catalytic mecha-
nism employed by this unique protease family will be greatly
aided by the availability of high-resolution Pla structures.
EXPERIMENTAL PROCEDURES
Cloning, Expression, and Membrane Preparation of Proteins
The pla gene from Y. pestis and ompT from E. coli were cloned into the E. coli
expression vector pB22 (Guzman et al., 1995; Van den Berg et al., 2004) with
a C-terminal hexa-histidine tag for purification. Active site residue mutations
were made by using the QuickChange Site-Directed Mutagenesis kit (Strata-
gene, La Jolla, CA). DNA sequencing was performed at CFAR DNA
sequencing facility (UMass Medical School, Worcester, MA). C43 cells (Miroux
and Walker, 1996) were transformed with Pla-pB22. The cells were grown to
OD600 1.0 at 37 C and then induced with 0.2% arabinose at 20 C overnight.
Cells were harvested by centrifugation at 4500 rpm for 20 min (Beckman
Coulter, J6-MC). Cell pellets were resuspended in 10 mM Tris-Cl, 50 mM
NaCl (pH 8.0), and cells were lysed by sonication (3 3 40 s intervals) (Branson
Digital Sonifier). Unbroken cells and inclusion cells were removed by centrifu-
gation at 9000 rpm for 20 min (Beckman L8-70M ultracentrifuge). Total
membranes were obtained by centrifugation at 40,000 rpm for 40 min (45 Ti
rotor; Beckman L8-70M ultracentrifuge). Membranes were homogenized in
10 mM Tris, 50 mM NaCl (pH 8.0) and kept at 80 C. Pla and OmpT amounts
in total membranes was determined by comparison of the band intensities in
Western blots with that of purified Pla and OmpT using the spot denso analysis
program (AlphaImager 2200), using serial dilutions to avoid saturation of the
band intensities. In Western blots the C terminus histidine tag was detected
using Penta-His HRP conjugate (QIAGEN, Germantown, MD). Detergent-puri-
fied Pla and OmpT protein concentrations were determined by the BCA
Protein Assay Kit (Thermo Scientific, Rockford, IL).

Purification of Pla for Crystallization
Total membranes of C43 cells overexpressing Pla were diluted to 6 mg/ml in
TSB (20 m Tris, 300 mM NaCl, 10% glycerol [pH 8.0]). Membranes were solu-
bilized by stirring in TSB with 1% LDAO (Anatrace, Maumee, OH) for 1 hour at
4 C. Solubilized membranes were centrifuged at 40,000 rpm for 30 min.
The membrane extract was applied to a 10 ml nickel column (chelating sephar-
ose fast flow; GE Healthcare, Waukesha, WI). The column was washed with
10 column volumes (CV) of TSB containing 0.2% LDAO and 15 mM imidazole.
Pla was eluted with 3 CV TSB containing 0.2% LDAO and 250 mM imidazole.
Pla was further purified by gel filtration chromatography using 10 mM Tris,
50 mM NaCl, 0.05% LDAO (pH 8.0). Purified protein was concentrated to
10 mg/ml and then dialyzed (50 kDa molecular weight cutoff; Spectrapor)
overnight against 10 mM Tris, 100 mM LiCl, 0.4% C8E4 (Anatrace) (pH 8.0).
The Pla mutant D86A was purified as described above and dialyzed against
10 mM Na-acetate, 100 mM LiCl, 0.4% C8E4 (pH 5.0).
Purification of Active Pla and OmpT
Pla and OmpT-containing total membranes were diluted to 6 mg/ml in 10 mM
Tris-Cl, 150 mM NaCl (pH 8.0). Membranes were solubilized in 1% OG
(b-octyl-glucoside; Anatrace) at 55 C for 30 min followed by centrifugation
at 40,000 rpm for 30 min to remove precipitates and unsolubilized membranes.
The membrane extract was applied to a 5 ml nickel column. The column was
washed with 5 CV 10 mM Tris-Cl, 150 mM NaCl (pH 8.0), 1% OG, 15 mM imid-
azole. The protein was eluted with 3 CV 10 mM Tris-Cl, 50 mM NaCl (pH 8.0),
1% OG. The proteins were concentrated to 5 mg/ml and dialyzed against
10 mM Tris-Cl, 50 mM NaCl, 1% OG (pH 8.0). Protein quantification was
carried out using the BCA assay (Thermo Scientific).
Crystallization of Pla and Structure Determination
Wild-type Pla was diluted to 6 mg/ml in 10 mM Tris, 100 mM LiCl, 0.4% C8E4.
Crystals were obtained by hanging drop vapor diffusion at 22 C by adding 1 ml
of protein solution to 1 ml of mother liquor containing 27% PEG 400, 0.1 M
LiSO2, 0.1 M Li-citrate (pH 4.0) (final pH 4.8 after addition of protein).
Zeppelin-shaped crystals appeared between 2 days to 1 week and reached
a maximum size of 300 microns in the longest direction. They belong to
space group P6522 (Table 1) and contain one Pla molecule per asymmetric
unit (Vm 5.3 Å3/Da, corresponding to 77% solvent content). A heavy atom
derivative was prepared by adding 5 mM OsCl3 to the crystal drop for 2 hours,
resulting in a yellow-brown color of the crystals. Interestingly, the soaking of
osmium (and platinum; data not shown) into the crystals resulted in an inability
to scale the data in P6522. Apparently the soaking resulted in a change of
space group, which is now C2221 (three molecules per asymmetric unit; Vm
5.3 Å3/Da, corresponding to 77% solvent content). As shown in Figure S2,
the osmium atoms are bound on the surface of the protein and may be involved
in lattice contacts, providing an explanation for the subtle change in space
group (the packing of the molecules is virtually identical between the two
space groups). Crystals were flash-frozen in liquid nitrogen directly from the
mother liquor. D86A mutant Pla crystals were obtained via the same method,
by adding 0.8 ml of protein (8 mg/ml) to 1 ml of mother liquor containing 16%
PEG 400, 0.1 M Li-citrate (pH 3.5). The cube-shaped crystals appeared after
about 1 week and grew to a maximum size of 150 microns. They belong to
space group P21 and contain four Pla molecules in the asymmetric unit
(Vm 3.4 Å3/Da, corresponding to 64% solvent content). These crystals
were flash-frozen from crystallization solution containing 30% PEG400.
Diffraction data were collected at APS beamlines ID-23B/D and NSLS
beamlines X25 and X6A, and were processed with HKL2000 (Otwinowski
and Minor, 1997) or XDS (Kabsch, 1993). A highly redundant SAD data set at
the osmium peak wavelength (1.13 Å) was collected and used for phasing
by SOLVE (Terwilliger and Berendzen, 1999). Three clear sites (one per Pla
molecule) were found and refined in SHARP (Bricogne et al., 2003). The result-
ing density-modified maps were of excellent quality and allowed automated
model building of 50% of the asymmetric unit by RESOLVE (Terwilliger,
2000). The model (without waters and detergents) was completed by manual
building within COOT (Emsley and Cowtan, 2004) and subjected to a single
round of refinement within PHENIX (Adams et al., 2002). This model was
used as the search model to solve the structures of wild-type Pla and the
D86A mutant by molecular replacement using Phaser (Storoni et al., 2004).
All structures were refined within PHENIX, using a protocol that included
816 Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved
Structure
High-Resolution Structure of Y. pestis Pla
TLS refinement. Simulated annealing Fo-Fc omit maps (Figures 3A and 5C)
were made in CNS1.2 (Brunger et al., 1998), using a start temperature of
1000K and cooling steps of 25K.
Pla and OmpT Activity Assays
Activity measurement of purified proteins was carried out with 2.5 mg Pla and
800 nM plasminogen in 50 mM Tris-Acetate, 50 mM MES (pH 6.5) buffer in
a total volume of 100 ml. Pla and plasminogen were incubated at room temper-
ature for 15 min, after which the reaction pH was changed to pH 8.0 by addition
of 90 ml 1M Tris (pH 8.0). Chromogenic substrate S2250 (Sigma, St. Louis, MO)
was added to the reaction mixture to a final concentration of 5 mM. The absor-
bance change at 405 nm was followed for 30 min (TECAN; Safire, Männedorf,
Switzerland). Background activity of reagents or proteins was subtracted from
all measured activities.
For activity determination of Pla crystals, several large crystals were
collected from the mother liquor using 0.1–0.2 mm loops and washed three
times by serial transfers into 2 ml mother liquor aliquots. The washed crystals
were redissolved in 10 mM Tris, 50 mM NaCl (pH 8.0), 0.4% C8E4. The amount
of protein within the crystals was determined both by BCA assay and spot
densitometry analysis of protein bands run on SDS-PAGE NuPAGE 4%–
12% Bis-Tris Gel; Invitrogen, Carlsbad, CA) gels. The single crystal (approxi-
mately 300 microns in size) that was used for structure determination was
also collected and analyzed in the activity assay. The activity of Pla from crys-
tals was determined as described above except that in these measurements
0.09–0.25 mg of Pla was used.
The determination of Pla activity in membranes was carried out as described
above except that 30 mg total membranes were used in the assay. Activities of
E. coli total membranes obtained from cells transformed with the empty
plasmid were subtracted from all membrane activities. Activities of active
site mutants were normalized according to the expression levels, which dif-
fered by only about 13%. In detergent inhibition assays, total membranes (6
mg/ml protein) were incubated with any of 1% (final concentration) OG,
Tween-20, C12E9, C8E4, DM, DDM, DHPC (di-heptanoyl-phosphatidylcholine),
or LDAO for 10 min at room temperature and 5 ml was transferred into the
activity assay mixture.
For reactivation of pla by Kdo2-LipidA and LPS, Pla was diluted to 0.1 mg/ml
using 10 mM Tris, 50 mM NaCl, 1% OG (pH 8.0). The protein was inactivated
by addition of 0.25% LDAO and incubation at room temperature for 10 min.
The minimum inhibitory concentration of LDAO that would inactive the protein
by more than 90% was determined by titration of protein with different
amounts of LDAO. Subsequently 5 ml of inactivated Pla was added to the
activity assay mixture. Reactivation was performed by titration of inactive
Pla with Kdo2-LipidA, E. coli LPS or Y. pestis LPS.
Determination of the pH-Activity Profile
The pH profile was determined by measuring the activity of OG-purified Pla or
total membranes containing Pla using two different buffer systems ranging
from pH 4.5 to 9.5. The first buffer system used was 50 mM Tris-acetate,
50 mM MES, and the second buffer system was 50 mM Tris-Cl and 50 mM
Bis-Tris. Both systems yielded identical results. In Figure 2C the results
obtained using the Tris-MES buffer system are shown.
LPS Detection Assay
LPS detection in purified Pla proteins extracted with either OG or LDAO and in
Pla crystals was carried out by Limulus Amoebocyte Lysate (LAL) endotoxin
assay kit (GenScript, Piscataway, NJ) according to the manufacturer’s proto-
cols. Pla crystals for assay were obtained and washed as described in the ‘‘Pla
and OmpT activity assays’’ section. For the assays 50 pg of purified protein
was used. LPS to protein molar ratios were calculated considering that 1 ng
endotoxin standard corresponds to 10 enzyme units of activity, and by
assuming an average molecular weight of LPS of 10,000 (Erridge et al., 2002).
ACCESSION NUMBERS
Coordinates and structure factors were deposited in the Protein Data Bank
with the following PDB ID codes: wild-type Pla Native 1, 2X55; wild-type Pla
Native 2, 2X56; and Pla mutant D86A, 2X4M.
Structure
High-Resolution Structure of Y. pestis Pla
SUPPLEMENTAL INFORMATION
Supplemental Information includes three figures and can be found with this
article online at doi:10.1016/j.str.2010.03.013.
ACKNOWLEDGMENTS
We thank Debra Touw for help in crystal data collection and data processing.
We thank the staff of APS beamlines ID-23B/D and NSLS beamlines X25 and
X6A for their assistance during data collection. We also thank the RapiData
2009 instructors (rapid data collection and structure solving) at the NSLS for
their guidance in and help with the collection of preliminary data of Pla crystals.
Received: January 26, 2010
Revised: March 22, 2010
Accepted: March 31, 2010
Published: July 13, 2010
REFERENCES
Adams, P.D., Grosse-Kunstleve, R.W., Hung, L.W., Ioerger, T.R., McCoy, A.J.,
Moriarty, N.W., Read, R.J., Sacchettini, J.C., Sauter, N.K., and Terwilliger, T.C.
(2002). PHENIX: building new software for automated crystallographic struc-
ture determination. Acta Crystallogr. D Biol. Crystallogr. 58, 1948–1954.
Agar, S.L., Sha, J., Foltz, S.M., Erova, T.E., Walberg, K.G., Baze, W.B., Suarez,
G., Peterson, J.W., and Chopra, A.K. (2009). Characterization of the rat pneu-
monic plague model: infection kinetics following aerosolization of Yersinia
pestis CO92. Microbes Infect. 11, 205–214.
Antonicelli, F., Brown, D., Parmentier, M., Drost, E.M., Hirani, N., Rahman, I.,
Donaldson, K., and MacNee, W. (2004). Regulation of LPS-mediated inflam-
mation in vivo and in vitro by the thiol antioxidant Nacystelyn. Am. J. Physiol.
Lung Cell. Mol. Physiol. 286, L1319–L1327.
Baaden, M., and Sansom, M.S. (2004). OmpT: molecular dynamics simula-
tions of an outer membrane enzyme. Biophys. J. 87, 2942–2953.
Beesley, E.D., Brubaker, R.R., Janssen, W.A., and Surgalla, M.J. (1967). Pes-
ticins. 3. Expression of coagulase and mechanism of fibrinolysis. J. Bacteriol.
94, 19–26.
Bercouvier, H., and Mollaret, H.H. (1984). Bergeys manual of systematic
bacteriology. 1, 498-509.
Bricogne, G., Vonrhein, C., Flensburg, C., Schiltz, M., and Paciorek, W. (2003).
Generation, representation and flow of phase information in structure determi-
nation: recent developments in and around SHARP 2.0. Acta Crystallogr. D
Biol. Crystallogr. 59, 2023–2030.
Brik, A., and Wong, C.H. (2003). HIV-1 protease: mechanism and drug
discovery. Org. Biomol. Chem. 1, 5–14.
Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-
Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al.
(1998). Crystallography & NMR system: a new software suite for macromolec-
ular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921.
Cowan, C., Jones, H.A., Kaya, Y.H., Perry, R.D., and Straley, S.C. (2000). Inva-
sion of epithelial cells by Yersinia pestis: evidence for a Y. pestis-specific inva-
sin. Infect. Immun. 68, 4523–4530.
Degen, J.L., Bugge, T.H., and Goguen, J.D. (2007). Fibrin and fibrinolysis in
infection and host defense. J. Thromb. Haemost. 5 (Suppl 1), 24–31.
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
Erridge, C., Bennett-Guerrero, E., and Poxton, I.R. (2002). Structure and
function of lipopolysaccharides. Microbes Infect. 4, 837–851.
Ferguson, A.D., Hofmann, E., Coulton, J.W., Diederichs, K., and Welte, W.
(1998). Siderophore-mediated iron transport: crystal structure of FhuA with
bound lipopolysaccharide. Science 282, 2215–2220.
Guzman, L.M., Belin, D., Carson, M.J., and Beckwith, J. (1995). Tight regula-
tion, modulation, and high-level expression by vectors containing the arabi-
nose PBAD promoter. J. Bacteriol. 177, 4121–4130.
Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved 817
Haiko, J., Suomalainen, M., Ojala, T., Lahteenmaki, K., and Korhonen, T.K.
(2009). Invited review: Breaking barriers–attack on innate immune defences
by omptin surface proteases of enterobacterial pathogens. Innate Immun.
15, 67–80.
Hritonenko, V., and Stathopoulos, C. (2007). Omptin proteins: an expanding
family of outer membrane proteases in Gram-negative Enterobacteriaceae.
Mol. Membr. Biol. 24, 395–406.
Kabsch, W. (1993). Automatic processing of rotation diffraction data from crys-
tals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26,
795–800.
Kramer, R.A., Dekker, N., and Egmond, M.R. (2000a). Identification of active
site serine and histidine residues in Escherichia coli outer membrane protease
OmpT. FEBS Lett. 468, 220–224.
Kramer, R.A., Zandwijken, D., Egmond, M.R., and Dekker, N. (2000b). In vitro
folding, purification and characterization of Escherichia coli outer membrane
protease ompT. Eur. J. Biochem. 267, 885–893.
Kramer, R.A., Vandeputte-Rutten, L., de Roon, G.J., Gros, P., Dekker, N., and
Egmond, M.R. (2001). Identification of essential acidic residues of outer
membrane protease OmpT supports a novel active site. FEBS Lett. 505,
426–430.
Kramer, R.A., Brandenburg, K., Vandeputte-Rutten, L., Werkhoven, M., Gros,
P., Dekker, N., and Egmond, M.R. (2002). Lipopolysaccharide regions involved
in the activation of Escherichia coli outer membrane protease OmpT. Eur. J.
Biochem. 269, 1746–1752.
Kukkonen, M., Lahteenmaki, K., Suomalainen, M., Kalkkinen, N., Emody, L.,
Lang, H., and Korhonen, T.K. (2001). Protein regions important for plasmin-
ogen activation and inactivation of alpha2-antiplasmin in the surface protease
Pla of Yersinia pestis. Mol. Microbiol. 40, 1097–1111.
Kukkonen, M., Suomalainen, M., Kyllonen, P., Lahteenmaki, K., Lang, H.,
Virkola, R., Helander, I.M., Holst, O., and Korhonen, T.K. (2004). Lack of
O-antigen is essential for plasminogen activation by Yersinia pestis and
Salmonella enterica. Mol. Microbiol. 51, 215–225.
Lathem, W.W., Price, P.A., Miller, V.L., and Goldman, W.E. (2007). A plasmin-
ogen-activating protease specifically controls the development of primary
pneumonic plague. Science 315, 509–513.
Miroux, B., and Walker, J.E. (1996). Over-production of proteins in Escherichia
coli: mutant hosts that allow synthesis of some membrane proteins and glob-
ular proteins at high levels. J. Mol. Biol. 260, 289–298.
Otwinowski, Z., and Minor, W. (1997). Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol. 276, 307–326.
Pearson, M.A., Park, I.S., Schaller, R.A., Michel, L.O., Karplus, P.A., and Hau-
singer, R.P. (2000). Kinetic and structural characterization of urease active site
variants. Biochemistry 39, 8575–8584.
Scott, D.L., White, S.P., Otwinowski, Z., Yuan, W., Gelb, M.H., and Sigler, P.B.
(1990). Interfacial catalysis: the mechanism of phospholipase A2. Science 250,
1541–1546.
Sebbane, F., Jarrett, C.O., Gardner, D., Long, D., and Hinnebusch, B.J. (2006).
Role of the Yersinia pestis plasminogen activator in the incidence of distinct
septicemic and bubonic forms of flea-borne plague. Proc. Natl. Acad. Sci.
USA 103, 5526–5530.
Sodeinde, O.A., Subrahmanyam, Y.V., Stark, K., Quan, T., Bao, Y., and
Goguen, J.D. (1992). A surface protease and the invasive character of plague.
Science 258, 1004–1007.
Storoni, L.C., McCoy, A.J., and Read, R.J. (2004). Likelihood-enhanced fast
rotation functions. Acta Crystallogr. D Biol. Crystallogr. 60, 432–438.
Sugimura, K., and Nishihara, T. (1988). Purification, characterization, and
primary structure of Escherichia coli protease VII with specificity for paired
basic residues: identity of protease VII and OmpT. J. Bacteriol. 170,
5625–5632.
Suomalainen, M., Haiko, J., Ramu, P., Lobo, L., Kukkonen, M., Westerlund-
Wikstrom, B., Virkola, R., Lahteenmaki, K., and Korhonen, T.K. (2007). Using
every trick in the book: the Pla surface protease of Yersinia pestis. Adv. Exp.
Med. Biol. 603, 268–278.

Taylor, A.H., Heavner, G., Nedelman, M., Sherris, D., Brunt, E., Knight, D., and
Ghrayeb, J. (1995). Lipopolysaccharide (LPS) neutralizing peptides reveal
a lipid A binding site of LPS binding protein. J. Biol. Chem. 270, 17934–17938.
Terwilliger, T.C. (2000). Maximum-likelihood density modification. Acta Crys-
tallogr. D Biol. Crystallogr. 56, 965–972.
Terwilliger, T.C., and Berendzen, J. (1999). Automated MAD and MIR structure
solution. Acta Crystallogr. D Biol. Crystallogr. 55, 849–861.
Van den Berg, B., Clemons, W.M., Jr., Collinson, I., Modis, Y., Hartmann, E.,
Harrison, S.C., and Rapoport, T.A. (2004). X-ray structure of a protein-
conducting channel. Nature 427, 36–44.
818 Structure 18, 809–818, July 14, 2010 ª2010 Elsevier Ltd All rights reserved
Structure
High-Resolution Structure of Y. pestis Pla
Vandeputte-Rutten, L., Kramer, R.A., Kroon, J., Dekker, N., Egmond, M.R.,
and Gros, P. (2001). Crystal structure of the outer membrane protease
OmpT from Escherichia coli suggests a novel catalytic site. EMBO J. 20,
5033–5039.
Wang, X., Lin, X., Loy, J.A., Tang, J., and Zhang, X.C. (1998). Crystal structure
of the catalytic domain of human plasmin complexed with streptokinase.
Science 281, 1662–1665.
Yun, T.H., and Morrissey, J.H. (2009). Polyphosphate and omptins: novel
bacterial procoagulant agents. J. Cell. Mol. Med. 10.1111/j.1582-4934.2009.
00884.x.