Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Contents lists available at ScienceDirect

BBA - Proteins and Proteomics


journal homepage: www.elsevier.com/locate/bbapap

N-terminus determines activity and specificity of styrene monooxygenase MARK


reductases
Thomas Heinea,b,⁎, Anika Scholtisseka,b, Adrie H. Westphalb, Willem J.H. van Berkelb,
Dirk Tischlera,⁎⁎
a
Environmental Microbiology, Interdisciplinary Ecological Center, TU Bergakadmie Freiberg, Leipziger Straße 29, 09599 Freiberg, Germany
b
Laboratory of Biochemistry, Wageningen University & Research, Stippeneng 4, 6708 WE Wageningen, The Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Styrene monooxygenases (SMOs) are two-enzyme systems that catalyze the enantioselective epoxidation of
Fusion protein styrene to (S)-styrene oxide. The FADH2 co-substrate of the epoxidase component (StyA) is supplied by an
Flavoprotein NADH-dependent flavin reductase (StyB). The genome of Rhodococcus opacus 1CP encodes two SMO systems.
Rhodococcus opacus One system, which we define as E1-type, displays homology to the SMO from Pseudomonas taiwanensis VLB120.
Protein-ligand interaction
The other system, originally reported as a fused system (RoStyA2B), is defined as E2-type. Here we found that
Fluorescence spectroscopy
E1-type RoStyB is inhibited by FMN, while RoStyA2B is known to be active with FMN. To rationalize the ob-
served specificity of RoStyB for FAD, we generated an artificial reductase, designated as RoStyBart, in which the
first 22 amino acid residues of RoStyB were joined to the reductase part of RoStyA2B, while the oxygenase part
(A2) was removed. RoStyBart mainly purified as apo-protein and mimicked RoStyB in being inhibited by FMN.
Pre-incubation with FAD yielded a turnover number at 30 °C of 133.9 ± 3.5 s− 1, one of the highest rates
observed for StyB reductases. RoStyBart holo-enzyme switches to a ping-pong mechanism and fluorescence
analysis indicated for unproductive binding of FMN to the second (co-substrate) binding site. In summary, it is
shown for the first time that optimization of the N-termini of StyB reductases allows the evolution of their
activity and specificity.

1. Introduction and one, designated as E2-type, containing epoxidase-like proteins


homologous to RoStyA1 and a fusion protein similar to RoStyA2B, both
Styrene monooxygenases (SMO; EC 1.14.14.11) form a unique from Rhodococcus opacus 1CP [13,24].
group (group E) of flavoprotein monooxygenases [1–3]. SMOs are two- The SMO reaction starts with StyB, which generates FADH2, which
enzyme systems, consisting of an epoxidase (StyA) and a flavin re- is then in turn used by the monooxygenase [23,25,26]. In StyA, FADH2
ductase (StyB; EC 1.5.1.36). StyA epoxidases are structurally related to reacts with molecular oxygen to form flavin hydroperoxide, which
group A flavoprotein monooxygenases [1,4], while StyB reductases subsequently reacts with styrene to generate the enantiopure epoxide
have structural properties in common with iron reductases [5]. Being [26,27]. After epoxidation, a flavin hydroxide remains, which subse-
involved in the microbial degradation of aromatic compounds in the quently eliminates water and is released as oxidized FAD. Although E1-
soil [6–9], SMOs catalyze the regio- and enantioselective epoxidation of type StyBs are capable of reducing FAD, FMN, and riboflavin, epox-
styrene into (S)-styrene oxide [10]. Others are supposed to be part of idation of styrene by StyAs is only observed with FADH2 as co-sub-
indole detoxification [11]. Interestingly, all SMOs perform sulfoxida- strate.
tion reactions [12–21], which is supposed to be a non-natural activity. So far, ten StyB reductases have been biochemically characterized
In recent years more and more SMOs were identified and described, [21,23–26,28–31]. While being strictly NADH-dependent, they all are
however, no clear nomenclature was introduced. We propose to define homodimers and active with FAD, FMN and riboflavin. The N-termini
two groups as suggested by Montersino et al. 2011: one with the most of the reductases vary in length, which is particularly true for the fusion
common SMOs as E1-type proteins, homologous to PtStyA (epoxidase) proteins (Fig. 1). In the latter, the reductases are linked to the mono-
and PtStyB (reductase) from Pseudomonas taiwanensis VLB120 [22,23] oxygenase components by a shortened N-terminus, which is surprising,


Correspondence to: T. Heine, Environmental Microbiology, Interdisciplinary Ecological Center, TU Bergakadmie Freiberg, Leipziger Straße 29, 09599 Freiberg, Germany.
⁎⁎
Corresponding author.
E-mail addresses: thomas.heine@ioez.tu-freiberg.de (T. Heine), dirk-tischler@email.de (D. Tischler).

http://dx.doi.org/10.1016/j.bbapap.2017.09.004
Received 31 May 2017; Received in revised form 10 August 2017; Accepted 5 September 2017
Available online 06 September 2017
1570-9639/ © 2017 Elsevier B.V. All rights reserved.
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Fig. 1. Multiple sequence alignment of StyB reductases and related flavin:NADH oxidoreductases. The artificial RoStyBart serves as default. RoStyB Rhodococcus opacus 1CP (accession
number and/or PDB ID: AII82582), RoSty(A2)B Rhodococcus opacus 1CP (ACR43974), VpSty(A2)B Variovorax paradoxus EPS (ADU39062), PpStyB Pseudomonas putida S12 (4F07), PtStyB
Pseudomonas taiwanensis VLB120 (AAC23719), PpStyB Pseudomonas putida SN1 (ABB03728), AbStyB Acinetobacter baylyi ADP1 (YP_047251), R5StyB Rhodococcus sp. ST-5 (BAL04133),
R10StyB Rhodococcus sp. ST-10 (BAL04130), SgcE6 Streptomyces globisporus (AAL06698/4R82), PheA2 Geobacillus thermoglucosidasius A7 (AAF66547/1RZ1), NphA2 Rhodococcus sp.
strain PN1 (BAB86379), HpaC Thermus thermophilus HB8 (2ED4), TTHA0420 Thermus thermophilus HB8 (WP_011172509/1Y0A). Structural data are available for marked (*) sequences.
E2-type fusion proteins are truncated for the monooxygenase part. Shadings correspond to the level of conservation (black_100%; dark gray_80%; gray_60%). Scale under the alignment is
numbered according to the sequence of RoStyBart.

because structural studies of other flavin reductases illustrate that this Bacillus sphaericus JS905 [38] and SgcEg from Streptomyces globisporus
region is involved in substrate binding [32,33]. So far, only a single [39], but an inhibition by FMN has never been reported. PheA2 from
crystal structure of a SMO reductase from Pseudomonas putida S12 is Geobacillus thermoglucosidasius A7 is able to reduce FMN but shows no
known (PDB ID: 4F07; PpStyB S12). The first 13 amino acid residues of tight interaction with this co-substrate in fluorescence titration ex-
this protein are not visible in the electron density map, and the func- periments [5,40].
tional role of these residues in catalysis remains to be elucidated. The self-sufficient E2-type RoStyA2B contains fused epoxidase and
Preliminary investigations towards the substrate spectrum of E1- reductase domains, which function together as conventional SMOs
type RoStyB pointed to a preference for FAD and inhibition with FMN (StyA/StyB), but as a single polypeptide. RoStyA2B has a very low
[21]. A substrate preference for FAD was also observed for HpaH C1 flavin reductase activity [21,24,41], which could be due to the fusion of
from Pseudomonas putida [34], pyrrole-2-carboxylate monooxygenase- the reductase to the much larger epoxidase. The N-terminus of the
reductase from Rhodococcus sp. [35], TftC from Burkholderia cepacia RoStyA2B-reductase is about 12 to 22 residues shorter compared to
AC1100 [36], NphA2 from Rhodococcus sp. PN1 [37], PNP-B from StyB reductases found in Pseudomonas, Variovorax and Rhodococcus

1771
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

species [42] (Fig. 1). Most of the studied StyB reductases form homo- 2.5. Gene expression, purification and storage of recombinant proteins
dimers [2,21,23,26,29,30]. RoStyA2B forms oligomers, with a shift to
the dimeric form under reducing conditions. This suggests that both the The expression screening was performed with different media in
epoxidase and reductase in the fusion protein are dimers, constraining 100-mL Erlenmeyer flasks for all constructs, allowing for the potential
the range of positions and conformations of the linkers between the two production of recombinant reductases, in analogy to earlier studies
enzymes [21]. StyB reductases usually follow a sequential mechanism [13,21]. Expression, purification and storage of RoStyA2B and RoStyB
[23,25,30]. However, we could observe a shift to a double-displace- were performed as described earlier [21,24].
ment reaction at higher FAD concentrations in an artificially con- For RoStyBart, expression took place in a 3-liter biofermenter. E. coli
structed mimic of the fusion protein [31]. BL21 (pRoStyBart_P01) was cultivated in LBNB medium [50]
To obtain more insight in the function of the N-terminal part of the (100 μg mL− 1 ampicillin and 50 μg mL− 1 chloramphenicol) at 30 °C
RoStyB proteins, we deleted the monooxygenase part (A2) of RoStyA2B until an OD600 of 0.4 was reached. The high amount of salt and pre-
and added the first 22 residues of RoStyB, resulting in artificial protein sence of betaine can prevent the formation of inclusion bodies. The
RoStyBart. The redesigned reductase has been characterized with re- batch was subsequently cooled to 20 °C. Expression was induced at an
spect to the parental oxidoreductases. OD600 of 0.6 by adding 0.1 mM of isopropyl-β-D-thiogalactopyranoside
(IPTG), and growth was continued for 20 h at 20 °C (120 rpm). Cells
2. Materials and methods were harvested by centrifugation (5000 ×g, 30 min, 4 °C), resuspended
in 10 mM Tris-HCl buffer (pH 7.5) and stored at −80 °C.
2.1. Chemicals and enzymes RoStyBart was obtained as soluble protein. Crude extract was pre-
pared from freshly thawed biomass by disruption in a precooled French
FAD, FMN and riboflavin were purchased from Sigma-Aldrich Pressure cell, followed by centrifugation to remove cell debris
(Steinheim, Germany) and Applichem (Darmstadt, Germany). NADH (50,000 × g, 2 h, 4 °C). The supernatant was applied to a 1-mL HisTrap
was purchased from Gerbu (Heidelberg, Germany). Enzymes for cloning FF column pre-equilibrated in 10 mM Tris-HCl, 0.5 M NaCl, 25 mM
purposes were received from MBI Fermentas (St. Leon-Rot, Germany) imidazole, pH 7.5 (loading buffer), using an ÄKTA fast-performance
and New England Biolabs GmbH (Frankfurt am Main, Germany). liquid chromatography system (GE Healthcare). The column was wa-
shed with 10 column volumes of loading buffer to remove nonspecific
2.2. Bacterial strains, plasmids and culture conditions bound proteins. RoStyBart was eluted with a linear imidazole gradient
up to 500 mM over 30 column volumes. Fractions with NADH:FAD
Escherichia coli strain DH5α and strain BL21 (DE3) pLysS were oxidoreductase activity (see paragraph 2.7.) were pooled and con-
cultivated for cloning and expression purposes as described elsewhere centrated using Sartorius Vivaspin 20 filters (5000 MWCO) at 4 °C. The
[43]. Plasmids are listed in Table 1. concentrate was passed through a 10 mL Econo-Pac 10DG desalting
gravity-flow column (Bio-Rad) to remove remaining imidazole and
2.3. Multiple sequence alignment of StyB reductases and protein structure sodium chloride. Protein obtained was kept in storage buffer (10 mM
modeling of RoStyBart Tris-HCl, 50% [v/v] glycerol, pH 7.5) at − 20 °C.

A multiple sequence alignment of StyB reductases was obtained 2.6. Protein determination, flavin content analysis and analytical gel
using MEGA version 6 [44]. Sequences were obtained from the NCBI filtration
protein database (www.ncbi.nlm.nih.gov). The alignment served as
initiation point for comparative modeling. Here, PheA2 (PDB ID: Recombinant proteins were subjected to discontinuous sodium do-
1RZ1), HpaC (2D36), CobR (4IRA) and a putative oxidoreductase from decyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) [43] in
Rickettsia felis (4L82) served as the structural templates [5,45–47] and order to determine purity and subunit molecular size. Protein con-
an iterative approach was used with Modeller Software 9.14 [48] to centration was determined with the Bradford method [51], using
compute a RoStyBart model-structure. BradfordUltra reagent (Expedeon) and bovine serum albumin (Sigma)
as reference protein.
2.4. Construction of expression clones The presence of non-covalently bound flavin in RoStyBart was
analyzed using RP-HPLC. For that, 50 μL of 70 μM purified protein in
The rostyBart gene (Accession number: MF124795) was purchased 10 mM Tris-HCl (pH 7.5) were heated at 95 °C for 5 min. After cen-
from Eurofins MWG (Ebersberg) in a pEX-A vector system allowing for trifugation (16,000 ×g, 10 min, 4 °C) 5 μL of the supernatant were
ampicillin resistance selection. The DNA sequence was optimized for applied on a Eurospher C18 column (250 mm length by 4 mm i.d., 5 μm
the codon usage and GC content of Acinetobacter baylyi ADP1 with the particle size, 100 Å pore size; Knauer, Germany). RP-HPLC was per-
OPTIMIZER tool [21,49]. 5′-NdeI and 3′-NotI restriction sites were formed with 50 mM sodium acetate (pH 5) using a linear gradient of
added and used for subcloning into pET16bP to obtain the expression 15% to 60% methanol at a flow of 0.7 mL min− 1 to elute the flavins.
construct pRoStyBart_P01, from which recombinant proteins can be 5 μL of 50 μM of each flavin species (net retention volumes: 6.40 mL for
obtained as His10-tagged proteins. FAD, 7.42 mL for FMN and 9.60 mL for riboflavin) were applied as

Table 1
Plasmids used in this study.

Plasmid Relevant characteristic(s) Source or reference

pET16bP pET16b with additional multicloning site; allows production of recombinant proteins with an N-terminal His10-tag U. Wehmeyer
pEX-A-PfStyBart pfstyBart (524-bp NdeI/NotI fragment) cloned into pEX-A MWG Eurofins
pEX-A-RoStyBart rostyBart (557-bp NdeI/NotI fragment) cloned into pEX-A MWG Eurofins
pRoStyA2B_P01a rostyA2B of R. opacus 1CP (1722-bp NdeI/NotI fragment) cloned into pET16bP [24]
pRoStyB_P01a rostyB of R. opacus 1CP (545-bp NdeI/NotI fragment) cloned into pET16bP [21]
pPfStyBart_P01 pfstyBart (524-bp NdeI/NotI fragment) cloned into pET16bP This study
pRoStyBart_P01 rostyBart (557-bp NdeI/NotI fragment) cloned into pET16bP This study

a
The original designations of plasmids and genes were pSRoA2B_P01 and pSRoB_P01, respectively.

1772
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

references. The detection wavelength was 450 nm. fluorescence (excitation wavelength = 280 nm, emission range
Flavin content of recombinant proteins was measured spectro- 290–400 nm) and for FAD fluorescence (excitation wave-
photometrically at a wavelength of 450 nm (assuming a molar ab- length = 450 nm, cutoff wavelength = 455, emission ran-
sorption coefficient for protein-bound flavin, ε450 nm = ge = 476–650 nm). An equimolar amount of FAD was applied to apo-
11.3 mM− 1 cm− 1) [52]. In addition, the flavin fluorescence was protein to generate the holo-form of the proteins. Dissociation constants
measured with excitation at 450 nm and emission detected at 525 nm. of protein-ligand complexes were determined from tryptophan (ex-
Analytical gel filtration of apo- and holo-form of RoStyBart was citation wavelength = 280 nm, cutoff wavelength = 290 nm, emission
performed on an ÄktaFPLC system (GE Healthcare) as described earlier wavelength = 340 nm) and/or flavin (excitation wave-
[21]. The holo-form of RoStyBart was analyzed under oxidized and length = 450 nm, cutoff wavelength = 515 nm, emission wave-
reduced conditions. Identity of the eluted proteins was checked by SDS- length = 525 nm) fluorescence titration experiments. NADH fluores-
PAGE and the standard activity assay. cence was monitored at 460 nm (emission wavelength = 340 nm).
NAD+, NADH, FAD and FMN served as ligands for RoStyBart. Each
2.7. Activity assay and inhibition studies titration step was measured separately in a well at 13 °C covering a
range of 0 to 6 μM of the respective ligand. The enzyme was applied at a
Flavin oxidoreductase activity was determined spectro- concentration of 0.43 μM to 20 mM Tris-HCl, pH 7.5, in a final volume
photometrically (Cary 50, Varian) by quantifying NAD(P)H consump- of 200 μL. To prevent the influence of post flavin-binding events, each
tion at 340 nm (ε340 nm = 6.22 mM− 1 cm− 1) [53]. One unit of en- set up was started by addition of enzyme and immediate data acquisi-
zyme activity is defined as the amount required to oxidize 1 μmol of tion.
NAD(P)H per min. All measurements were carried out in triplicate. The Dissociation constants of enzyme-flavin complexes were obtained
standard assay consisted of 20 mM Tris-HCl, pH 7.5, 60 μM flavin co- by using the nonlinear Generalized Reduced Gradient (GRG) algorithm
substrate (FAD, FMN or riboflavin) and 175 μM NAD(P)H in 1.0 mL. (Solver – Microsoft Excel) by fitting the fluorescence emission data (F)
After incubating the mixture for 10 min at 30 °C, the reaction was to the model described by Eqs. (1) and (2).
started by adding an appropriate amount of enzyme (around 70 nM).
F = fFAD ∗ [FAD]free + fComplex ∗ [Complex] + fEnzyme ∗ [Enzyme]free (1)
For estimating steady-state kinetic parameters, initial reaction rates
were determined using 2.5 to 60 μM FAD and 10 to 175 μM NADH. (K dapp + [FAD]total + [Enzyme]total )
RoStyBart was applied either as apo- or as holo-enzyme where the 1
− ((K dapp + [FAD]total + [Enzyme]total )2 − 4∗ [FAD]total ∗ [Enzyme]total ) 2
purified flavin-free enzyme was considered as apo-form and the enzyme [Complex] =
2
pre-incubated with FAD was designated as holo-form. For the latter, the
(2)
stock enzyme solution (7 μM) was incubated (10 min, room tempera-
ture) in 20 Tris-HCl, pH 7.5 with 60 μM FAD. Kinetic parameters were Here, fFAD, fComplex and fEnzyme are the fluorescence conversion factors
obtained by nonlinear regression analysis applying KaleidaGraph 4.5 for free FAD, formed complex and free enzyme, respectively.
(Synergy Software), assuming Michaelis-Menten kinetics. The time course of flavin binding to RoStyBart was monitored at
The pH preference for holo-RoStyBart was tested in an assay con- 25 °C over 107 min by recording tryptophan fluorescence emission at
taining 100 mM sodium phosphate buffer, 60 μM FAD and 175 μM 340 nm in 20 mM Tris-HCl, pH 7.5. Flavin co-substrates were applied in
NADH, covering a pH-range from 5.4 to 8. In order to elucidate the ten-fold excess over the enzyme. Data fitting and calculation of rate
temperature optimum for the oxidoreductase activity of RoStyBart, the constants was executed with KaleidaGraph 4.5 (Synergy Software) ap-
holo-enzyme was applied to a pre-equilibrated standard assay (20 to plying Eq. (3) for the binding of FAD and Eq. (4) for the binding of
47 °C). For evaluation of the thermal stability, the holo-enzyme was FMN. Here, F is the total fluorescence resulting from S1 and S2, re-
incubated for 10 min in a microcentrifuge tube equilibrated in 20 Tris- presenting the fluorescence of different protein-flavin species that are
HCl, pH 7.5 with 60 μM FAD to the respective temperature (20 to 52 °C; formed, and B, the fluorescence of the remaining assay components.
Thermomixer, Eppendorf), and applied to the standard assay at 30 °C.
F = B + S1 ∗e(−t ∗ kapp1) − S2 ∗e(−t ∗ kapp2) (3)
To evaluate the flavin specificity, RoStyBart (7 μM) was pre-in-
cubated at room temperature in 20 mM Tris-HCl, pH 7.5, containing F = B + S1 ∗e(−t ∗ kapp1) (4)
either 60 μM of FAD, FMN or riboflavin. After a period of 60 min, the
holo-protein obtained was applied to a standard assay containing either
60 μM of FAD, FMN, or riboflavin, or a FAD/FMN (1:1) mixture, re- 3. Results
spectively. RoStyB was treated similarly and assayed for conversion of
FAD or FMN. In addition, oxidoreductase activity was measured, using 3.1. Multiple sequence alignment of StyB reductases and structure modeling
the standard assay, of samples taken periodically up to 60 min from an
incubation at 4 °C or 22 °C of RoStyBart (7 μM) added to 20 mM Tris- A multiple sequence alignment of the selected StyB reductases was
HCl, pH 7.5, containing either 60 μM FAD or 60 μM FMN. computed with MEGA6 [44] (Fig. 1). Sequences were obtained using
RoStyA2B (ACR43974) and RoStyB (AKM21224) as seed sequences for
2.8. Spectrophotometric and fluorescence measurements BLASTP searches [54] in the non-redundant protein database (NCBI).
All StyB primary structures are rather short in length (149 to 182 aa)
Flavin fluorescence titration was applied to determine the dis- and share large portions of sequence similarity. The motifs responsible
sociation constant (Kd) for the complex between RoStyB and FAD (Cary for NAD and flavin binding [32,33,39,55–58] are conserved.
Eclipse fluorescence spectrometer; Varian). RoStyB (3 μM) was titrated Based on the presented sequence alignment, a dimeric model of
with 0 to 10 μM FAD in 20 mM Tris-HCl, pH 7.5, and the FAD fluor- RoStyBart was computed using the crystal structures of PheA2, HpaC,
escence signal was monitored at 525 nm upon excitation at 450 nm CobR and a putative oxidoreductase as templates (Fig. 2). If present in
(about 30 data points per titration step; standard deviation of < 1%). the template structures, the (co-)substrates NADH and FAD were in-
The measurements were done at 13 °C in view of the low stability of cluded in the modeling [5,26]. The single tryptophan (Trp70) of RoS-
RoStyB. tyBart was found to be close to the isoalloxazine ring of FAD (distance
All further spectroscopic measurements were carried out using a about 4 Å) (Fig. 2A).
Black Quartz Microplate (Hellma, clear bottom) and a SpectraMax M2e The multiple sequence alignment depicted in Fig. 1 contains next to
and data were captured with the software SoftMax Pro (both Molecular all characterized StyB reductases also related reductases. In the fol-
Devices, USA). Emission spectra were recorded for tryptophan lowing, all amino acid positions are numbered according to the

1773
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Fig. 2. Surface map (A) and ribbon diagram (B) of the 3D-structural model of RoStyBart with bound FAD and NADH. Flavins are placed according to structural data of PheA2 (PDB ID:
1RZ1; bright orange) and PpStyB S12 (4F07; orange). NADH is positioned according to PheA2 (1RZ1). Protein backbones of the dimers are shaded in green or gray and the N-terminal
residues are shaded in red. In (A), residues that are proposed to be involved in (co-)substrate binding are shaded in orange (FAD), blue (NADH) and a flexible loop (aa 106 to 114) above
the FAD binding site (yellow), respectively. The only tryptophan within the RoStyBart sequence (Trp70) is close to the isoalloxazine ring (about 3.4 to 4 Å; magenta). Arg28 is part of the
N-terminal region and known to form hydrogen bonds towards NADH (indicated by dashed lines). For clarity, only one FAD and one NADH molecule is indicated in the dimer (B). (For
interpretation of the references to color in this figure legend, the reader is referred to the online version of this chapter.)

sequence of RoStyBart. The residues that are required for (co-)substrate studied by Riedel et al. [21]. The RoStyB apo-protein mainly occurs as
binding are highly conserved. The TANx3Sx10S, [STC]xxPP and GDH dimer. The apo- and holo-forms of RoStyA2B exist as higher-order
motifs are localized at the active site or close to it [39,55,56]. Thr54, quarternary complexes in equilibrium with the dimer. However, under
Ala55, Asn56 and Ser/Cys71 are involved in binding the isoalloxazine reduced conditions there is a shift to the dimeric state. In this study, we
ring of the FAD [39,56–58]. Val38, Ala55, Ser105 and Phe167 form a performed analytical gel filtration under similar conditions for RoSty-
hydrophobic pocket to bind the dimethylbenzene moiety of the iso- Bart (Fig. S3). The apo-form of RoStyBart mainly occurs as a monomer.
alloxazine ring [5,32,58]. Whereas the location of the FMN moiety is If incubated with FAD, a clear shift to the dimeric state was detected for
similar, the AMP moiety of FAD can bind differently. In PheA2, the the holo-form of RoStyBart, with minor formation of higher-order
adenine and ribose moieties interact with valine and asparagine re- structures. Under reduced conditions, the holo-enzyme almost solely
sidues (Lys50 and Gln103 in RoStyBart) at a small cleft on the surface of exists as a dimer. If incubated with FMN, there is less tendency to di-
each monomer [5,32]. Interestingly, the lysine residue replacing the merization. Under oxidized conditions, monomer and dimer of FMN-
non-polar valine is only found in StyA2B-like SMOs and AbStyB (Aci- bound RoStyBart are in equilibrium. Under reduced conditions, there is
netobacter baylii ADP1), which might affect the recognition of the AMP a shift to the monomeric-state.
part of FAD [55]. Pro77 and Phe104 may also form hydrogen bonds
with the adenosine moiety of FAD [58]. In PpStyB S12, the adenosine 3.3. Catalytic properties of RoStyB, RoStyA2B and RoStyBart
moiety occupies the active site of an adjacent monomer [26]. This
might be due to a more open conformation of a loop that corresponds to 3.3.1. Enzyme activity with FAD
residues 106–114 in the RoStyBart model (Fig. 2A). This region is The enzymes used here are not active with NADPH, thus, all oxi-
variable in StyB reductases and thus no suggestion can be made how the doreductase activities were assayed using NADH. The steady-state ki-
AMP part of FAD will bind in RoStyBart. netic properties of RoStyA2B and RoStyB with FAD as co-substrate have
There is only one StyB structure available in the database [26]. In been reported before [21,24] and are summarized in Table 2. Worthy of
this structure, the N-terminal residues 1–13 are not visible due to poor note is that pre-incubation of RoStyB with FAD does not change the
electron density [26]. This suggests a high flexibility of the N-terminus. enzyme activity.
As this region is close to the FAD/NADH binding pocket, it could in- RoStyBart is very active with FAD (Fig. 3). However, the Michaelis
fluence the catalytic properties of the enzyme. Furthermore, the flexible constants for FAD and NADH of this redesigned enzyme are quite dif-
N-terminus might also be involved in the formation and transient sta- ferent from that of RoStyB and RoStyA2B (Table 2). The KM, FAD values of
bilization of a StyA-StyB complex [26]. RoStyBart determined with apo-protein (KM,− FAD = 15.5 ± 1.2 μM), or
with apo-protein pre-incubated with FAD (KM,+FAD = 19.1 ± 1.2 μM)
3.2. Purification of RoStyBart and analytical gel filtration are in the same range, as is the case for the KM values for NADH
(KM,NADH − FAD = 110.5 ± 9.9 μM; KM,NADH + FAD = 74.6 ± 4.4 μM).
RoStyBart was purified by metal affinity chromatography. SDS- Pre-incubation with FAD gives a three-fold increase in activity of RoS-
PAGE yielded a single band at a size of approximately 21 kDa (Fig. S1), tyBartholo (kcat = 133.9 ± 3.5 s− 1) compared to RoStyBartapo
in accordance with the theoretical subunit mass of 21.8 kDa (including (kcat = 46.1 ± 2.1 s− 1) (Fig. 3, Table 2).
the N-terminal His10-tag). HPLC analysis of the heat-treated RoStyBart Upon incubation at 20 °C, RoStyBartholo is stable for at least 60 min
preparation showed no peaks at the expected positions of FAD, FMN or (Fig. S4). In a series of experiments where the standard activity of
riboflavin (Fig. S2), thus, the protein is purified mainly in its apo-form. RoStyBartholo was measured at various temperatures, a maximum was
This is supported by the absorption and fluorescence properties of the found at around 40 °C. If RoStyBartholo is pre-incubated at various
purified RoStyBart enzyme. Neither the characteristic flavin absorption temperatures for 10 min, the activity (measured at 30 °C) starts to drop
spectrum nor an emission signal at 525 nm after excitation at 450 nm above 20 °C, an indication that the enzyme is not very thermotolerant
was observed. (Fig. S5). The pH optimum for RoStyBartholo lies in a range of 5.8 to 6.2
The hydrodynamic properties of the SMOs from R. opacus 1CP were (Fig. S5).

1774
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Table 2
Kinetic parameters of RoStyBart compared to its templates and ancestors.

Enzyme Electron donor Electron acceptor Km kcat kcatKm− 1 Reference


μM μM μM s− 1 s− 1 μM− 1

RoStyBartapo NADH (10–175) FAD (60) 110.5 ± 9.9 46.1 ± 2.1 0.417 This study
NADH (175) FAD (2.5–70) 15.5 ± 1.2 34.3 ± 0.9 2.216
RoStyBartholo NADH (10–175) FAD (60) 74.6 ± 4.4 133.9 ± 3.5 1.795 This study
NADH (175) FAD (2.5–70) 19.1 ± 1.2 122.7 ± 2.8 6.419
RoStyA2B NADH (17.5–175) FAD (50) 58.0 ± 9.0 3.9 ± 0.3 0.068 [24]
(R. opacus 1CP) NADH (175) FAD (4–60) 26.0 ± 2.0 5.2 ± 0.2 0.203
NADH (175) FMN (4–60) 67.0 ± 22.0 4.8 ± 1.0 0.071
NADH (175) Riboflavin (4–60) 23.0 ± 2.0 3.0 ± 0.1 0.131
RoStyB NADH (2.5–300) FAD (200) 59.3 ± 4.2 26.9 ± 0.7 0.454 [21]
(R. opacus 1CP) NADH (250) FAD (5–300) 108.3 ± 18.3 32.9 ± 2.4 0.304

(See Supporting Table S1 for kinetic data of other StyB's).

The specific activity of RoStyBartholo depends on the enzyme con- compared to the activity with FAD (Fig. 4A). Furthermore, the activity
centration (Fig. 3C). A significant decrease was observed at con- with FAD and FMN is hardly dependent on the pre-incubation with
centrations below 0.75 μg mL− 1. When, upon dilution in the assay FAD. However, if RoStyBapo is pre-incubated with FMN, little activity is
mixture, dissociation of dimers occurs, the decrease in activity can be found in the reaction with FAD and there is almost no activity with
explained if the formed monomers are not active. A dissociation con- FMN.
stant for the dimer of about 1 nM would describe the observed curve. As With RoStyBart, the activity with FAD and FMN is strongly stimu-
described in paragraph 3.2, dimerization of RoStyBart is stimulated by lated by pre-incubation with FAD (Fig. 4B). This suggests that FAD
flavin binding (cf. Fig. S3). serves as activity enhancing prosthetic group. Both RoStyBartapo and
RoStyBartholo are relatively more active with FMN than the corre-
sponding RoStyB proteins (Fig. 4). However, when RoStyBart is pre-
3.3.2. Enzyme activity with different flavin cofactors incubated with FMN, the enzyme is hardly active with both FAD and
Interesting differences between RoStyB and RoStyBart are seen in FMN, suggesting that FMN functions as an inactive prosthetic group.
activity measurements started with apo-protein and FAD or FMN in the If riboflavin is used as flavin co-substrate for RoStyBartapo no
assay solution (Fig. 4). RoStyB shows only about 10% activity with FMN

Fig. 3. Steady-state kinetic analysis of FAD:NADH oxidor-


eductase activity of RoStyBart. Specific activities are
plotted as function of FAD (A) or NADH (B) concentration
while keeping the amount of the other substrate constant.
Kinetic parameters were calculated using KaleidaGraph 4.5
(Synergy) assuming Michaelis-Menten kinetics. The stan-
dard assays were started with either RoStyBartapo (■; da-
shed line) or RoStyBartholo (●; solid line). (C) Specific ac-
tivity of RoStyBartholo as a function of enzyme
concentration. All measurements were done in triplicate.

1775
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Fig. 4. Relative activities of RoStyB (A) and RoStyBart (B)


with different flavin co-substrates. Enzymes were used as
apo-protein or pre-incubated for 60 min in a 60 μM flavin
containing solution and subsequently applied to the stan-
dard assay containing NADH and the respective flavin. All
measurements were carried out in triplicate. Results are
depicted relative to the respective holo-enzyme.

oxidoreductase activity is observed. Pre-incubation with riboflavin


neither yields activity with FAD, FMN or riboflavin as substrate. This is
again an indication that only FAD serves as activity enhancing pros-
thetic group of RoStyBart.

3.3.3. Kinetic mechanism of RoStyBart


In order to elucidate the kinetic mechanisms of the apo- and holo-
forms of RoStyBart, initial rates were determined over a range of
FAD:NADH ratios and depicted in double reciprocal plots (Fig. S6). In
case of RoStyBartapo, for NADH as well as for FAD as fixed (co-)sub-
strate, the different slopes intersect to the left of the ordinate (Fig. S6A
and Fig. S6B). In the first case, the lines intersect at the X-axis, in-
dicating a constant KM for FAD and a sequential mechanism. In the
latter case, the intersection falls into the second quadrant, which is
typical for a ternary complex comprising mechanism. The intersections
with the X-axis imply a decrease in Km for NADH upon increasing FAD
concentration. Thus, it can be assumed that FAD serves as the leading
substrate. At this stage the observed intersections suggest that
RoStyBartapo obeys a sequential ordered bi-bi mechanism. Further, the
proposed initial binding of FAD fits to a proposed role as a prosthetic
group. This is in accordance with mechanistic studies on RoStyBartholo
(Fig. S6C and Fig. S6D). For NADH as well as for FAD as fixed co-
substrate, parallel slopes are observed, which indicates a switch to a
ping-pong bi-bi mechanism compared to the apo-form.

3.4. Protein-flavin interaction


Fig. 5. (A) Fluorescence emission spectra of FAD (excitation wavelength = 450 nm) for
The binding of FAD to RoStyA2B, RoStyB and RoStyBart was de- RoStyBartholo (orange), RoStyA2Bholo (blue) and RoStyB (green) normalized to the
termined by monitoring flavin and/or tryptophan fluorescence emis- emission maximum of free FAD (black). (B) Fluorescence emission spectra from trypto-
sion. The FAD-fluorescence of RoStyBartholo is almost identical to that phans of RoStyBart (orange), RoStyA2B (blue) and RoStyB (green), whereby apo-protein
of free FAD. In RoStyA2Bholo and RoStyBholo, the FAD-fluorescence is is shown by dashed lines and holo-protein by solid lines (excitation = 280 nm). Curves
about doubled or raised about 50% compared to that of free FAD, re- were measured in duplicates. Assays contained 50 mM Tris-HCl (pH 7.5), 17.5 μM FAD
and 17.25 μM of the respective protein.
spectively (Fig. 5A).
RoStyA2B and RoStyBart share a single tryptophan at position 49
and 70, respectively (Fig. 2). RoStyB, on the other hand, contains two tryptophan to the FAD, thereby diminishing the tryptophan emission.
tryptophans at position 86 and 111 (Fig. 2). The maxima of the tryp- Tryptophan fluorescence studies revealed different kinetics for FAD
tophan emission of RoStyBartapo and RoStyBartholo around 330 nm are and FMN binding to RoStyBartapo. In case of incubation with FMN there
blue-shifted compared to that of the apo- and holo-forms of RoStyB and is a slow decrease in tryptophan fluorescence over time with a rate
RoStyA2B. This suggests that the tryptophan in RoStyBart is located in a constant kapp1 of 0.0233 ± 0.0003 min− 1 (Fig. 6A). For FAD, initially
rather apolar environment which was confirmed by mapping the hy- a relatively fast rise (kapp2 = 0.448 ± 0.02 min− 1) is followed by a
drophobicity onto the model structure (Fig. S8B). Upon FAD binding, slow decrease in tryptophan fluorescence (Fig. 6A). The slow step has
the tryptophan fluorescence of RoStyBartholo decreases by about 50%, an apparent rate constant similar to that of binding of FMN
that of RoStyB by about 30%, and that of RoStyA2Bholo by about 70%, (kapp1 = 0.0206 ± 0.0003 min− 1).
compared to the apo-form (Fig. 5B). The Tryptophan is close to the The dissociation constants of the enzyme-FAD complexes of
isoalloxazine ring (about 3.4 to 4 Å). Likely, in the holo-form of the RoStyBapo and RoStyBartapo were determined by fluorescence titration
proteins, Förster resonance energy transfer takes place (FRET) from the

1776
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Fig. 6. Time dependent change in the tryptophan fluores-


cence intensity of RoStyBart (A) due to the binding of a
flavin co-substrate (excitation at 280 nm; emission at
340 nm). Solid dots refer to FAD and blank dots refer to
FMN. The curves were fitted with Eq. (3) (for FAD; dashed
line) and Eq. (4) (for FMN; solid line). For RoStyBart, one
rate constant kapp1 = 0.0233 ± 0.0003 min− 1 was ob-
tained for the binding of FMN whereas the fitting of the
FAD trace gives a fast component
kapp2 = 0.448 ± 0.02 min− 1 and a slow component
kapp1 = 0.0206 ± 0.0003 min− 1. Every curve was at least
measured twice. (B) Relative fluorescence emission from
tryptophan of RoStyBart (0.43 μM) measured at 340 nm (●
solid line; excitation 290 nm) and FAD measured at 525 nm
(○ dashed line; excitation 450 nm) during titration with
FAD to estimate the dissociation constant of the apo-RoS-
tyBart-FAD complex. Binding of FAD was monitored in case
of RoStyB (3 μM) at 525 nm (□ dashed line; excitation
450 nm).

experiments. To prevent the influence of post ligand-binding events, first SMO reductase, which is strictly FAD-dependent. In contrast to its
each set up was started by addition of enzyme to the FAD solution and family members, FMN is an inhibitor instead of a substrate for RoStyB.
immediate data acquisition (Fig. 6B). Clear binding of FAD was ob- Thus, it can mark an evolutionary event for this enzyme class. All other
served for RoStyBapo and RoStyBartapo. From the tryptophan and flavin StyB reductases are not strictly FAD-dependent and thus may yield
fluorescence titration curves and global fitting, a Kd of 4.3 ± 0.2 μM unproductive side-activities.
was estimated for the RoStyBapo-FAD complex assuming 1:1 binding. In this study, we replaced the N-terminal region of RoStyA2B to get
This is in congruence with previous results that indicate weak binding further insight into the molecular determinants of the flavin specificity
of FAD [21]. For RoStyBart, similar tryptophan and FAD titration ex- of the SMO reductases. The RoStyBart enzyme was predominantly
periments resulted in a Kd of 0.65 ± 0.06 μM for the RoStyBartapo-FAD purified as apo-protein (Fig. S2). This is somewhat surprising in view of
complex. This value is almost one order of magnitude lower than the the proposed function of FAD as prosthetic group (vide infra).
dissociation constant determined for the RoStyA2Bapo-FAD complex Nevertheless, RoStyBart appeared to be one of the most active StyB
(Kd = 5.05 ± 0.52 μM; [13]). enzymes reported today (Fig. 7 and Table 2). Pre-incubation of RoS-
tyBart with FAD led to a strong increase in oxidoreductase activity
4. Discussion (Fig. 3). This activation appeared to be dependent on protein con-
centration and temperature (Fig. S4). In accordance with that, we could
SMOs are playing an important role in the detoxification of styrene observe an initial increase in tryptophan fluorescence followed by a
in aerobic microorganisms. Rhodococcus opacus 1CP uses two SMO slow decrease over time (Fig. 6A). The need for higher concentrations
types, which we here define as E1- and E2-type. E1-type proteins are and temperatures may be due to the need for a dimeric structure with a
homologous to the PtStyA (epoxidase) and PtStyB (reductase) from hydrophobic interface (vide infra). The optimal temperature for the
Pseudomonas taiwanensis VLB120 [22,23] and are usually embedded in activation is located around 20 °C (Fig. S4 and S5A). A temperature
a styrene degradation cluster. E2-type SMOs contain epoxidase-like dependent activation towards the binding of FAD has been described
proteins homologous to RoStyA1 and flavin reductases similar to the before for other dimeric flavoproteins such as D-amino acid oxidase
reductase part of RoStyA2B, both from Rhodococcus opacus 1CP [13,24]. [64] and lipoamide dehydrogenase [65,66].
The E2-type is usually not part of a styrene degradation cluster [21]. The activity of RoStyBart is difficult to compare with other SMO
Recent studies indicate that this second type might be involved in in- reductases as the specific activity depends on the protein concentration.
dole detoxification [11]. This was earlier hypothesized [42] and a PtStyB and PpStyB S12 were shown to reach activities of up to
convergent evolution of these SMO types was concluded [30]. The 700 U mg− 1 upon dilution [23,25,26] (Fig. 7). Interestingly, RoStyBart
phylogenetic analysis (Fig. 1 and Fig. S7) presented here reinforce this shows an opposite behavior (Fig. 3C). The specific activity of this
hypothesis. The motif TANx3Sx10S discriminates E1- or E2-type StyB Rhodococcus enzyme decreases when the protein is diluted. RoSty-
reductases. The alanine residue of this motif varies towards E1-type Bartholo occurs manly in the dimeric state whereas the apo-form tends
reductases that have valine or isoleucine at this position. In most E2- to form monomers and oligomers (Fig. S3). This supports that SMO
type reductases, the second serine of the motif is altered to cysteine. reductases are active as homodimers. The quaternary structure may
This region is supposed to be involved in binding the isoalloxazine ring also be the reason for the remarkable higher activity of the artificial
of the FAD [39,56–58]. The amino acid changes in the TANx3Sx10S construct compared to RoStyA2B. As described earlier [21] RoStyA2B
motif should not have large effects due to their chemical similarity. has two dimer interfaces. Possibly, there is a steric hindrance due to the
Furthermore, only in case of E1-type StyB reductases the GDH motif is large oxygenase part that impedes a correct dimer formation and thus, a
altered to histidine or asparagine at the glycine position. In the RoS- high catalytic activity.
tyBart model, this glycine residue is close to the SxxPP motif of the Intriguingly, the affinity of FAD for RoStyBart (Kd =
adjacent monomer and this change might affect dimer formation. 0.65 ± 0.06 μM) is strongly improved compared to that of RoStyA2B
Several reductases of two-component flavoprotein monooxygenase (Kd = 5.05 ± 0.52 μM) and RoStyB (Kd = 4.31 ± 0.18 μM) (Fig. 6B).
systems display flavin promiscuity while the oxygenase components can The affinity of RoStyBart for FAD is slightly better than found with other
utilize only FAD. Indeed, this behavior was reported for PheA1/PheA2 SMO reductases: PpStyB S12 (1.15 ± 0.14 μM), PtStyB (2.3 ± 0.3 μM)
[40], RebH/RebF [59] and HpaA/HpaC [60]. On the other hand, there and AbStyB (1.82 ± 0.02 μM) [23,26,30]. A dissociation constant in the
are also two-component flavoprotein monooxygenase systems that micromolar range is common for NADH-flavin reductases [60,67–69]
possess FAD specificity within the oxygenase and the reductase: SgcC/ whereas few also bind the FAD with nanomolar affinity [5,70].
SgcE6 [61], NphA1/NphA2 [37,62], TftD/TftC [63], HPAH C2/C1 Several flavin binding modes are feasible as shown for related re-
[34], 4-NP monooxygenase [38] and pyrrole-2-carboxylate mono- ductases. In PheA2, a cofactor FAD binds tightly with the isoalloxazine
oxygenase [35]. In this context, it is important to note that RoStyB is the ring directed to the dimer interface and the adenosine moiety stabilized

1777
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

Fig. 7. Comparison of the highest determined activities (A) for SMO reductases while using FAD and NADH as (co)-substrates. The relative catalytic efficiencies for FAD are shown in (B)
[13,21,23,24,28–30]. AbStyB was assayed at 25 °C and PpStyB SN1 at 37 °C.

by a loop after the third α-helix (which can also be found in the sufficient SMO originating from the same strain, is less active as RoStyB,
RoStyBart model). PheA2 acts according to a ping-pong bi-bi me- but can reduce various flavins [24]. In this study we created an artificial
chanism. After reduction of the FAD cofactor by NADH and subsequent construct designated RoStyBart by joining the N-terminus of RoStyB
release of NAD+, a flavin substrate binds at the NADH-binding site and with the reductase part of RoStyA2B. The resulting RoStyBart combines
takes the electrons from the reduced FAD cofactor [5]. In contrast, the flavin specificity of RoStyB with a higher activity and affinity for
PpStyB S12 does not use a tightly bound FAD cofactor, which might be FAD. This demonstrates for the first time that the NADH:FAD oxidor-
due to a more flexible loop compared to PheA2. Therefore, the ADP eductase activity of an natural E2-type SMO is affected by the N-
moiety of FAD can occupy the NADH binding site of an adjacent terminal fusion to the epoxidase subunit.
monomer (Fig. S8A) [26]. The kinetic studies of RoStyBartapo indicate FAD as leading substrate
A peculiar feature of RoStyB is the inhibition by FMN and riboflavin. in a sequential mechanism. The fluorescence investigations revealed a
This is the first report about such a behavior in NADH:flavin oxidor- fast step upon FAD binding, which does not occur in case of FMN
eductases. Moreover, this property remained in the artificial construct binding. This suggests that the AMP moiety of FAD is needed to trigger
RoStyBart. Fluorescence kinetic data support a difference in binding RoStyBart in order to allow proper NADH binding and oxidation. A
mode between FAD and FMN (Fig. 7). For incubation with FAD, the fast second flavin binding site can be assumed as RoStyBartholo follows a
component (kapp2 of 0.448 ± 0.02 min− 1) might reflect binding as a ping- pong mechanism. Thus, it is likely that a co-substrate FAD binds
prosthetic group to the flavin binding site. This is in accordance with to the NADH-binding site. FMN is supposed to block this site and in-
activation of RoStyBart during incubation with FAD (Fig. S4) which hibits the enzyme. Structural studies and further engineering of the N-
happens in a similar time range. Further, RoStyBartholo shows a ping- terminal region might help understanding the reasons for the change or
pong bi-bi mechanism, implying a second binding site for a flavin co- limitation in the co-substrate specificity.
substrate. Binding of FAD to that site might be reflected by the slow
decrease in fluorescence (kapp1 of 0.0233 ± 0.0003 min− 1). In con- Transparency document
trast, when RoStyBart is incubated with FMN, the fast component is
missing and only the slow component can be determined. Thus, it can The Transparency document associated with this article can be
be suggested that FMN does not bind to the flavin binding site as a found, in online version.
prosthetic group. Further, the rate of this slow decrease in fluorescence
(kapp1 of 0.0206 ± 0.0003 min− 1) is in a similar range when com- Acknowledgements
pared to the binding of FAD to the co-substrate site as well as to the
inactivation of the enzyme (ki, app of 0.046 ± 0.001 min− 1, Fig. S4). It The authors thank for the funding by the European Social Fund and
is likely, that the co-substrate flavin has to bind in a closed conforma- the Saxon Government (GETGEOWEB: 100101363).
tion as with NADH (Fig. 2). As this is not possible for FMN, it will
occupy the second (co-substrate) binding site leading to inhibition of Appendix A. Supplementary data
the enzyme. This state might be stabilized by Arg28 (Fig. S8A). This is
supported by analytical gel-filtration experiments as RoStyBart showed Supplementary data to this article can be found online at http://dx.
less dimer formation if FMN is applied under oxidized conditions and doi.org/10.1016/j.bbapap.2017.09.004.
dissociates to the monomeric state under reduced conditions (Fig. S3B).
This implies that FMN binds preferably to the NADH-binding site References
without stimulating dimer formation. As only the N-terminal part is
different compared to RoStyA2B, this region seems to impact binding of [1] W.J.H. van Berkel, N. Kamerbeek, M. Fraaije, Flavoprotein monooxygenases, a di-
the flavin to the NADH binding site and in case of FMN binding is verse class of oxidative biocatalysts, J. Biotechnol. 124 (2006) 15582–15587.
[2] S. Montersino, D. Tischler, G.T. Gassner, W.J.H. van Berkel, Catalytic and structural
unfavorable for catalysis. features of flavoprotein hydroxylases and epoxidases, Adv. Synth. Catal. 353 (2011)
2301–2319.
5. Conclusion [3] M.M. Huijbers, S. Montersino, A.H. Westphal, D. Tischler, W.J.H. van Berkel, Flavin
dependent monooxygenases, Arch. Biochem. Biophys. 544 (2014) 2–17.
[4] M.L. Mascotti, M.J. Ayub, N. Furnham, J.M. Thornton, R.A. Laskowski, Chopping
RoStyB from Rhodococcus opacus 1CP is the first NAD(P)H-depen- and changing: the evolution of the flavin-dependent monooxygenases, J. Mol. Biol.
dent flavin reductase that is highly specific for FAD. RoStyA2B, a self- 428 (2016) 3131–3146.

1778
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

[5] R.H.H. van den Heuvel, A.H. Westphal, A.J.R. Heck, M.A. Walsh, S. Rovida, W.J.H. van Berkel, A. Mattevi, Reduced Flavin Reductase PheA2 in Complex With
W.J.H. van Berkel, A. Mattevi, Structural studies on flavin reductase PheA2 reveal NAD, 13 (2004), pp. 12860–12867, http://dx.doi.org/10.2210/pdb1rz1/pdb.
binding of NAD in an unusual folded conformation and support novel mechanism of [33] T.R. Russell, S.-C. Tu, Aminobacter aminovorans NADH:flavin oxidoreductase
action, J. Biol. Chem. 279 (2003) 12860–12867. His140: a highly conserved residue critical for NADH binding and utilization,
[6] G. Baggi, M.M. Boga, D. Catelani, E. Galli, V. Treccani, Styrene catabolism by a Biochemist 43 (2004) 12887–12893.
strain of Pseudomonas fluorescens, Syst. Appl. Microbiol. 4 (1983) 141–147. [34] U. Arunachalam, V. Massey, C.S. Vaidyanathan, p-Hydroxyphenylacetate-3-hydro-
[7] F. Beltrametti, A.M. Marconi, G. Bestetti, C. Colombo, E. Galli, M. Ruzzi, E. Zennaro, xylase. A two-protein component enzyme, J. Biol. Chem. 267 (1992) 25848–25855.
Sequencing and functional analysis of styrene catabolism genes from Pseudomonas [35] D. Becker, T. Schräder, J.R. Andreesen, Two-component flavin-dependent pyrrole-
fluorescens ST, Appl. Environ. Microbiol. 63 (1997) 2232–2239. 2-carboxylate monooxygenase from Rhodococcus sp, Eur. J. Biochem. 249 (1997)
[8] P.M. Santos, J.M. Blatny, I. Di Bartolo, S. Valla, E. Zennaro, Physiological analysis of 739–747.
the expression of the styrene degradation gene cluster in Pseudomonas fluorescens [36] L. Xun, Purification and characterization of chlorophenol 4-monooxygenase from
ST, Appl. Environ. Microbiol. 66 (2000) 1305–1310. Burkholderia cepacia AC1100, J. Bacteriol. 178 (1996) 2645–2649.
[9] D. Tischler, Microbial Styrene Degradation, Springer Verlag, 2015. [37] M. Takeo, M. Murakami, S. Niihara, K. Yamamoto, M. Nishimura, D.-i. Kato,
[10] S. Hartmans, M.J. van der Werf, J.A. de Bont, Bacterial degradation of styrene in- S. Negoro, Mechanism of 4-nitrophenol oxidation in Rhodococcus sp. strain PN1:
volving a novel flavin adenine dinucleotide-dependent styrene monooxygenase, characterization of the two-component 4-nitrophenol hydroxylase and regulation of
Appl. Environ. Microbiol. 56 (1990) 1347–1351. its expression, J. Bacteriol. 190 (2008) 7367–7374.
[11] G.-H. Lin, H.-P. Chen, H.-Y. Shu, S.-W. Lee, Detoxification of indole by an indole- [38] V. Kadiyala, J.C. Spain, A two-component monooxygenase catalyzes both the hy-
induced flavoprotein oxygenase from Acinetobacter baumannii, PLoS ONE 10 (2015) droxylation of p-nitrophenol and the oxidative release of nitrite from 4-ni-
e0138798. trocatechol in Bacillus sphaericus JS905, Appl. Environ. Microbiol. 64 (1998)
[12] F. Hollmann, P.-C. Lin, B. Witholt, A. Schmid, Stereospecific biocatalytic epoxida- 2479–2484.
tion: the first example of direct regeneration of a FAD-dependent monooxygenase [39] S.G. van Lanen, S. Lin, G.P. Horsman, B. Shen, Characterization of SgcE6, the flavin
for catalysis, J. Am. Chem. Soc. 125 (2003) 8209–8217. reductase component supporting FAD-dependent halogenation and hydroxylation
[13] D. Tischler, R. Kermer, J.A.D. Groning, S.R. Kaschabek, W.J.H. van Berkel, in the biosynthesis of the enediyne antitumor antibiotic C-1027, FEMS Microbiol.
M. Schlömann, StyA1 and StyA2B from Rhodococcus opacus 1CP: a multifunctional Lett. 300 (2009) 237–241.
styrene monooxygenase system, J. Bacteriol. 192 (2010) 5220–5227. [40] U. Kirchner, A.H. Westphal, R. Müller, W.J.H. van Berkel, Phenol hydroxylase from
[14] L.J. Gursky, J. Nikodinovic-Runic, K.A. Feenstra, K.E. O'Connor, In vitro evolution Bacillus thermoglucosidasius A7, a two-protein component monooxygenase with a
of styrene monooxygenase from Pseudomonas putida CA-3 for improved epoxide dual role for FAD, J. Biol. Chem. 278 (2003) 47545–47553.
synthesis, Appl. Microbiol. Biotechnol. 85 (2010) 995–1004. [41] D. Tischler, M. Schlömann, W.J.H. van Berkel, G.T. Gassner, FAD C(4a)-hydroxide
[15] E.W. van Hellemond, D.B. Janssen, M.W. Fraaije, Discovery of a novel styrene stabilized in a naturally fused styrene monooxygenase, FEBS Lett. 587 (2013)
monooxygenase originating from the metagenome, Appl. Environ. Microbiol. 73 3848–3852.
(2007) 5832–5839. [42] D. Tischler, J.A.D. Gröning, S.R. Kaschabek, M. Schlömann, One-component styrene
[16] H. Toda, R. Imae, N. Itoh, Efficient biocatalysis for the production of enantiopure monooxygenases: an evolutionary view on a rare class of flavoproteins, Appl.
(S)-epoxides using a styrene monooxygenase (SMO) and Leifsonia alcohol dehy- Biochem. Biotechnol. 167 (2012) 931–944.
drogenase (LSADH) system, Tetrahedron Asymmetry 23 (2012) 1542–1549. [43] J. Sambrook, D.W. Russell (Eds.), Molecular Cloning: A Laboratory Manual, Cold
[17] D.R. Boyd, N.D. Sharma, B. McMurray, S.A. Haughey, C.C.R. Allen, J.T.G. Hamilton, Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2001.
W.C. McRoberts, R.A.M. O'Ferrall, J. Nikodinovic-Runic, L.A. Coulombel, [44] K. Tamura, G. Stecher, D. Peterson, A. Filipski, S. Kumar, MEGA6: Molecular evo-
K.E. O'Connor, Bacterial dioxygenase- and monooxygenase-catalysed sulfoxidation lutionary genetics analysis version 6.0, Mol. Biol. Evol. 30 (2013) 2725–2729.
of benzo [b] thiophenes, Org. Biomol. Chem. 10 (2012) 782–790. [45] M. Okai, N. Kudo, W.C. Lee, M. Kamo, K. Nagata, M. Tanokura, Crystal structures of
[18] J. Nikodinovic-Runic, L. Coulombel, D. Francuski, N.D. Sharma, D.R. Boyd, the short-chain flavin reductase HpaC from Sulfolobus tokodaii strain 7 in its three
R.M.O. Ferrall, K.E. O'Connor, The oxidation of alkylaryl sulfides and benzo [b] states: NAD(P)(+)(−)free, NAD(+)(−)bound, and NADP(+)(−)bound,
thiophenes by Escherichia coli cells expressing wild-type and engineered styrene Biochemist 45 (2006) 5103–5110.
monooxygenase from Pseudomonas putida CA-3, Appl. Microbiol. Biotechnol. 97 [46] M.C. Clifton, J.A. Abendroth, Structure of a Putative Oxidoreductase From Rickettsia
(2013) 4849–4858. felis, (2013), http://dx.doi.org/10.2210/pdb4l82/pdb.
[19] R.D. Ceccoli, D.A. Bianchi, D.V. Rial, Flavoprotein monooxygenases for oxidative [47] A.D. Lawrence, A.F. Scott, M.J. Warren, R.W. Pickersgill, CobR in Complex With
biocatalysis: recombinant expression in microbial hosts and applications, Front. FAD, (2014), http://dx.doi.org/10.2210/pdb4ira/pdb.
Microbiol. 5 (2014) 25. [48] A. Fiser, A. Sali, Modeller: Generation and refinement of homology-based protein
[20] C.E. Paul, D. Tischler, A. Riedel, T. Heine, N. Itoh, F. Hollmann, Nonenzymatic structure models, Methods Enzymol. 374 (2003) 461–491.
regeneration of styrene monooxygenase for catalysis, ACS Catal. 5 (2015) [49] P. Puigbo, E. Guzman, A. Romeu, S. Garcia-Vallve, OPTIMIZER: A web server for
2961–2965. optimizing the codon usage of DNA sequences, Nucleic Acids Res. 35 (2007)
[21] A. Riedel, T. Heine, A.H. Westphal, C. Conrad, P. Rathsack, W.J.H. van Berkel, W126–W131.
D. Tischler, Catalytic and hydrodynamic properties of styrene monooxygenases [50] N. Oganesyan, I. Ankoudinova, S.-H. Kim, R. Kim, Effect of osmotic stress and heat
from Rhodococcus opacus 1CP are modulated by cofactor binding, AMB Express 5 shock in recombinant protein overexpression and crystallization, Protein Expr.
(2015) 112. Purif. 52 (2007) 280–285.
[22] S. Panke, B. Witholt, A. Schmid, M.G. Wubbolts, Towards a biocatalyst for (S)- [51] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram
styrene oxide production: characterization of the styrene degradation pathway of quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem.
Pseudomonas sp. strain VLB120, Appl. Environ. Microbiol. 64 (1998) 2032–2043. 72 (1976) 248–254.
[23] K. Otto, K. Hofstetter, M. Röthlisberger, B. Witholt, A. Schmid, Biochemical char- [52] S.K. Chapman, G.A. Reid (Eds.), Flavoprotein Protocols, Humana Press, New Jersey,
acterization of StyAB from Pseudomonas sp. strain VLB120 as a two-component 1999.
flavin-diffusible monooxygenase, J. Bacteriol. 186 (2004) 5292–5302. [53] R.M.C. Dawson, Data for Biochemical Research, Clarendon Press, Oxford, 1986.
[24] D. Tischler, D. Eulberg, S. Lakner, S.R. Kaschabek, W.J.H. van Berkel, [54] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, Basic local alignment
M. Schlömann, Identification of a novel self-sufficient styrene monooxygenase from search tool, J. Mol. Biol. 215 (1990) 403–410.
Rhodococcus opacus 1CP, J. Bacteriol. 191 (2009) 4996–5009. [55] K. Thotsaporn, J. Sucharitakul, J. Wongratana, C. Suadee, P. Chaiyen, Cloning and
[25] A. Kantz, F. Chin, N. Nallamothu, T. Nguyen, G.T. Gassner, Mechanism of flavin expression of p-hydroxyphenylacetate 3-hydroxylase from Acinetobacter baumannii:
transfer and oxygen activation by the two-component flavoenzyme styrene mono- evidence of the divergence of enzymes in the class of two-protein component aro-
oxygenase, Arch. Biochem. Biophys. 442 (2005) 102–116. matic hydroxylases, Biochim. Biophys. Acta Gene Struct. Expr. 1680 (2004) 60–66.
[26] E. Morrison, A. Kantz, G.T. Gassner, M.H. Sazinsky, Structure and mechanism of [56] B. Galán, E. Díaz, M.A. Prieto, J.L. García, Functional analysis of the small com-
styrene monooxygenase reductase: new insight into the FAD-transfer reaction, ponent of the 4-hydroxyphenylacetate 3-monooxygenase of Escherichia coli W: a
Biochemist 52 (2013) 6063–6075. prototype of a new Flavin:NAD(P)H reductase subfamily, J. Bacteriol. 182 (2000)
[27] A. Kantz, G.T. Gassner, Nature of the reaction intermediates in the flavin adenine 627–636.
dinucleotide-dependent epoxidation mechanism of styrene monooxygenase, [57] S.-H. Kim, T. Hisano, W. Iwasaki, A. Ebihara, K. Miki, Crystal structure of the flavin
Biochemist 50 (2011) 523–532. reductase component (HpaC) of 4-hydroxyphenylacetate 3-monooxygenase from
[28] Y.-J. Yeo, Production, purification, and characterization of soluble NADH-flavin Thermus thermophilus HB8: Structural basis for the flavin affinity, Proteins 70 (2008)
oxidoreductase (StyB) from Pseudomonas putida SN1, J. Microbiol. Biotechnol. 19 718–730.
(2009) 362–367. [58] T. Imagawa, T. Tsurumura, Y. Sugimoto, K. Aki, K. Ishidoh, S. Kuramitsu, H. Tsuge,
[29] H. Toda, R. Imae, T. Komio, N. Itoh, Expression and characterization of styrene Structural basis of free reduced flavin generation by flavin reductase from Thermus
monooxygenases of Rhodococcus sp. ST-5 and ST-10 for synthesizing enantiopure thermophilus HB8, J. Biol. Chem. 286 (2011) 44078–44085.
(S)-epoxides, Appl. Microbiol. Biotechnol. 96 (2012) 407–418. [59] E. Yeh, S. Garneau, C.T. Walsh, Robust in vitro activity of RebF and RebH, a two-
[30] J.A.D. Gröning, S.R. Kaschabek, M. Schlömann, D. Tischler, A mechanistic study on component reductase/halogenase, generating 7-chlorotryptophan during re-
SMOB-ADP1: an NADH:flavin oxidoreductase of the two-component styrene beccamycin biosynthesis, Proc. Natl. Acad. Sci. 102 (2005) 3960–3965.
monooxygenase of Acinetobacter baylyi ADP1, Arch. Microbiol. 196 (2014) [60] S. Chakraborty, M. Ortiz-Maldonado, B. Entsch, D.P. Ballou, Studies on the me-
829–845. chanism of p-hydroxyphenylacetate 3-hydroxylase from Pseudomonas aeruginosa: a
[31] T. Heine, K. Tucker, N. Okonkwo, B. Assefa, C. Conrad, A. Scholtissek, system composed of a small flavin reductase and a large flavin-dependent oxyge-
M. Schlömann, G. Gassner, D. Tischler, Engineering styrene monooxygenase for nase, Biochemist 49 (2010) 372–385.
biocatalysis: reductase-epoxidase fusion proteins, Appl. Biochem. Biotechnol. [61] S. Lin, S.G. Van Lanen, B. Shen, Characterization of the two-component, FAD-de-
(2016) 1–21. pendent monooxygenase SgcC that requires carrier protein-tethered substrates for
[32] R.H.H. van den Heuvel, A.H. Westphal, A.J. Heck, M.A. Walsh, S. Rovida, the biosynthesis of the enediyne antitumor antibiotic C-1027, J. Am. Chem. Soc.

1779
T. Heine et al. BBA - Proteins and Proteomics 1865 (2017) 1770–1780

130 (2008) 6616–6623. fluorescens. Kinetics of reconstitution, Eur. J. Biochem. 197 (1991) 769–779.
[62] D. Mitra, C.S. Vaidyanathan, A new 4-nitrophenol 2-hydroxylase from a Nocardia [67] R.J. Parry, W. Li, An NADPH:FAD oxidoreductase from the valanimycin producer,
sp, Biochem. Int. 8 (1984) 609–615. Streptomyces viridifaciens. Cloning, analysis, and overexpression, J. Biol. Chem. 272
[63] B.N. Webb, J.W. Ballinger, E. Kim, S.M. Belchik, K.-S. Lam, B. Youn, M.S. Nissen, (1997) 23303–23311.
L. Xun, C. Kang, Characterization of chlorophenol 4-monooxygenase (TftD) and [68] L. Filisetti, M. Fontecave, V. Niviere, Mechanism and substrate specificity of the
NADH:FAD oxidoreductase (TftC) of Burkholderia cepacia AC1100, J. Biol. Chem. flavin reductase ActVB from Streptomyces coelicolor, J. Biol. Chem. 278 (2003)
285 (2010) 2014–2027. 296–303.
[64] V. Massey, B. Curti, A new method of preparation of D-amino acid oxidase apo- [69] V. Sedláček, T. Klumpler, J. Marek, I. Kučera, Biochemical properties and crystal
protein and a conformational change after its combination with flavin adenine di- structure of the flavin reductase FerA from Paracoccus denitrificans, Microbiol. Res.
nucleotide, J. Biol. Chem. 241 (1966) 3417–3423. 188-189 (2016) 9–22.
[65] J.F. Kalse, C. Veeger, Relation between conformations and activities of lipoamide [70] M.K. Tiwari, R.K. Singh, J.-K. Lee, H. Zhao, Mechanistic studies on the flavin:NADH
dehydrogenase, Biochim. Biophys. Acta, Enzymol. 159 (1968) 244–256. reductase (PrnF) from Pseudomonas fluorescens involved in arylamine oxygenation,
[66] W.J.H. van Berkel, J.A. Benen, M.C. Snoek, On the FAD-induced dimerization of Bioorg. Med. Chem. Lett. 22 (2012) 1344–1347.
apo-lipoamide dehydrogenase from Azotobacter vinelandii and Pseudomonas

1780

You might also like