Structural and Mutagenesis Studies of the Thiamine-Dependent, Ketone-Accepting YerE from Pseudomonas protegens

Article abstract of DOI:10.1002/cbic.201800325

A wide range of thiamine diphosphate (ThDP)-dependent enzymes catalyze the benzoin-type carboligation of pyruvate with aldehydes. A few ThDP-dependent enzymes, such as YerE from Yersinia pseudotuberculosis (YpYerE), are known to accept ketones as acceptor substrates. Catalysis by YpYerE gives access to chiral tertiary alcohols, a group of products difficult to obtain in an enantioenriched form by other means. Hence, knowledge of the three-dimensional structure of the enzyme is crucial to identify structure–activity relationships. However, YpYerE has yet to be crystallized, despite several attempts. Herein, we show that a homologue of YpYerE, namely, PpYerE from Pseudomonas protegens (59 % amino acid identity), displays similar catalytic activity: benzaldehyde and its derivatives as well as ketones are converted into chiral 2-hydroxy ketones by using pyruvate as a donor. To enable comparison of aldehyde- and ketone-accepting enzymes and to guide site-directed mutagenesis studies, PpYerE was crystallized and its structure was determined to a resolution of 1.55 ?.

Full text of DOI:10.1002/cbic.201800325

Accepted Article
Title: Structural and Mutagenesis Studies of the Thiamine-Dependent,
Ketone-Accepting YerE from Pseudomonas protegens
Authors: Jan-Patrick Steitz, Michael Müller, Sabrina Loschonsky, Anna
Baierl, Luzia Wiesli, Michael Richter, Martina Pohl, Gunter
Schneider, Doreen Dobritzsch, and Alexander Fries
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.201800325
Link to VoR: http://dx.doi.org/10.1002/cbic.201800325
A Journal of
www.chembiochem.org

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
Structural and Mutagenesis Studies of the Thiamine-Dependent,
Ketone-Accepting YerE from Pseudomonas protegens
Sabrina Hampel,+[a] Jan-Patrick Steitz,+[a] Anna Baierl,+[b] Patrizia Lehwald,[a] Luzia Wiesli,[c] Michael
Richter,[c,d] Alexander Fries,[a] Martina Pohl,[b] Gunter Schneider,[e] Doreen Dobritzsch,*[f] and Michael
Müller*[a]
Abstract: A wide range of thiamine diphosphate (ThDP) dependent
enzymes catalyze benzoin-type carboligation of pyruvate with
aldehydes. Few ThDP-dependent enzymes, such as YerE from
Yersinia pseudotuberculosis (YpYerE), are known to accept ketones
as acceptor substrates. Catalysis by YpYerE gives access to chiral
tertiary alcohols, a group of products difficult to obtain enantioenriched
by other means. Hence, knowledge of the three-dimensional structure
of the enzyme is crucial to identify structure–activity relationships.
However, YpYerE could not be crystallized despite several attempts.
Herein, we show that a homologue of YpYerE, namely PpYerE from
Pseudomonas protegens (59% amino acid identity), displays similar
catalytic activity: benzaldehyde and derivatives as well as ketones are
converted into chiral 2-hydroxy ketones using pyruvate as donor. To
enable comparison of aldehyde- versus ketone-accepting enzymes
and guide site-directed mutagenesis studies, PpYerE was crystallized
and its structure determined to 1.55 Å resolution.
dependent aldehyde–ketone cross-coupling.[1,2] YpYerE is
involved in the biosynthesis of the branched-chain sugar
yersiniose A, which is part of the O-antigen of the host and other
bacteria.[1,3] The physiologically catalyzed reaction is the transfer
of an activated acetaldehyde, generated by decarboxylation of
pyruvate, to a 3,6-dideoxy-4-ketosugar.[1] In addition to pyruvate,
YpYerE accepts 2-oxobutyrate and acetaldehyde as donors and
a broad range of different acceptor substrates, with moderate to
high enantioselectivity.[2]
In addition to YpYerE, further ThDP-dependent enzymes are
able to accept ketones. Acetohydroxy acid synthases (AHAS) are
a large group of ThDP-dependent enzymes distantly related to
YerE. Physiologically, they catalyze the synthesis of acetolactate
and acetohydroxy acids.[4,5] AHAS accept -keto acids, but are
not known to accept non-activated ketones.
The enzyme acetoin:dichlorophenolindophenol oxido-
reductase (Ao:DCPIP OR) from Bacillus licheniformis (also
named acetylacetoin synthase [AAS][4,6,7]) is able to activate and
transfer 2-hydroxy ketones, e.g. acetoin and methylacetoin, as
well as 1,2-diketones of different chain length (C2 and C3).
Moreover, as with YpYerE, esters of pyruvate can act as acceptor
substrates. Compared to YpYerE, Ao:DCPIP OR affords the
same enantioselectivity for the addition of acetaldehyde to
cyclohexane-1,2-dione, but opposite enantiomers with aromatic
acceptor substrates.[5] Recently, Giovannini et al. reported the
highly enantioselective synthesis of (S)-phenylacetylcarbinol [(S)-
PAC] and derivatives in good yield by the wild-type Ao:DCPIP OR
using methylacetoin as donor substrate.[8] Before that, the
synthesis of (S)-PAC was achieved only by rationally designed
variants of R-selective enzymes.[9]
Introduction
YerE from Yersinia pseudotuberculosis (YpYerE) was the first
characterized biocatalyst for thiamine diphosphate (ThDP)
[a]
Dr. S. Hampel,+ J.-P. Steitz,+ Dr. P. Lehwald, Dr. A. Fries, Prof. Dr.
M. Müller
Institut für Pharmazeutische Wissenschaften
Albert-Ludwigs-Universität Freiburg
Albertstrasse 25, 79104 Freiburg (Germany)
E-mail: michael.mueller@pharmazie.uni-freiburg.de
The related enzyme AcoAB from Bacillus subtilis plays a key
role in the biodegradation of methylacetoin. Metabolic engineer-
ing proved its ability to generate an activated acetaldehyde from
pyruvate or acetoin that is transferred to acetone, generating
methylacetoin.[10] The range of donor and acceptor substrates and
reactivities is rarely comparable with YpYerE.
[b]
[c]
A. Baierl,+ Prof. Dr. M. Pohl
IBG-1: Biotechnology, Forschungszentrum Jülich GmbH,
52425 Jülich (Germany)
L. Wiesli
Swiss Federal Laboratories for Materials Science and Technology
(Empa), Laboratory for Biointerfaces
Lechenfeldstr. 5, 9014 St. Gallen (Switzerland)
ThDP-dependent cyclohexane-1,2-dione hydrolase (CDH)
from Azoarcus sp. strain 22Lin, similar to IolD from different
organisms, might play a role in the catabolism of cyclitols and
inositols. CDH catalyzes the cleavage of C–C bonds and can
catalyze the reverse reaction. The double variant CDH-
H28A/N484A was rationally created to accept ketones as accep-
tor substrates; this variant shares an overlapping substrate range
with YpYerE. Two notable donor substrates accepted by the
double variant are methyl pyruvate and butane-2,3-dione. This
reactivity is similar to the enzymes from acetoin metabolism.[11,12]
Despite its fundamental biocatalytic and biosynthetic
significance, YpYerE has not been structurally characterized as
yet, although numerous attempts to obtain crystals have been
made. Within this paper, we present our efforts to gain deeper
[d]
[e]
[f]
Dr. M. Richter
Fraunhofer Institute for Interfacial Engineering and Biotechnology
IGB, branch BioCat
Schulgasse 11a, 94315 Straubing (Germany)
Prof. Dr. G. Schneider
Department of Medical Biochemistry and Biophysics
Karolinska Institutet
Tomtebodavägen 6, 17177 Stockholm (Sweden)
Prof. Dr. D. Dobritzsch
Department of Chemistry-BMC, Uppsala Universitet
Husargatan 3, 75237 Uppsala (Sweden)
E-mail: doreen.dobritzsch@kemi.uu.se
[+]
These authors contributed equally to this work.
Supporting information for this article is given via a link at the end of the
document.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
insight into the catalysis of ketone-accepting ThDP-dependent
enzymes by identifying an enzyme homologous to YpYerE:
PpYerE from Pseudomonas protegens, for which we were able to
obtain crystals suitable for X-ray structure elucidation, shows
homology to YpYerE concerning the associated biosynthetic gene
cluster and the catalyzed reactions.
In the biosynthesis of CDP-yersiniose A (4), the coupling of CDP-
4-keto-3,6-dideoxy-D-glucose (2) and the ThDP-bound activated
acetaldehyde is followed by a reduction of the ketone 3.[1] Chen
et al. postulated that YpYerF, an NAD(P)H-dependent reductase,
catalyzes this reaction.[1] A gene (pp-yerF) with moderate identity
to yp-yerF (37%) was found. The gene product, annotated as an
epimerase, was named PpYerF. Further in silico investigation has
not given conclusive proof of its function. Nevertheless, the low
sequence homology to ascE (25%), the epimerase from the
ascarylose biosynthetic gene cluster, is noteworthy.[1,13,14] We
found four open reading frames in the biosynthetic gene cluster
harboring the gene pp-yerE, that possibly code for glycosyltrans-
ferases; however, none of them showed any homology to YerG.
Another open reading frame encodes conserved domains that are
common in acyltransferases.
Regrettably, the structure of the physiological donor substrate
cannot be directly deduced from sequence analysis of the
biosynthetic gene cluster. We did not find any genes encoding
enzymes that might play a role in the synthesis of an alternative
donor substrate than pyruvate, e.g., an aldolase that chain-
elongates pyruvate. As pyruvate was a good donor in the in vitro
assays (see below) and is ubiquitous in the cell, we propose that
pyruvate is the physiological donor substrate of PpYerE.
To gain deeper insight into the catalytic mechanism of ketone-
accepting ThDP-dependent enzymes, we investigated whether
PpYerE can substitute YpYerE in several of the reactions known
to be catalyzed by the latter, with pyruvate or acetaldehyde being
used as donor substrate.
Results and Discussion
We identified the putative enzyme PpYerE by BLAST analysis
using the protein sequence of YpYerE. Both proteins are nearly
equal in length with 568 and 565 amino acids, respectively, of
which 336 are identical (59.3%).[13] The biosynthetic gene clusters
for the synthesis of yersiniose A and ascarylose were analyzed
and compared with the putative biosynthetic gene cluster
containing the pp-yerE gene on the genetic level.[1,14] Ascarylose
is a 3,6-dideoxysugar included in the O-antigen of different Y.
pseudotuberculosis strains.
Chen et al. have identified four homologous genes in the bio-
synthetic gene clusters of yersiniose A and ascarylose.[1] Those
genes (ddhA–D) code for four enzymes that catalyze the con-
version of -D-glucose 1-phosphate (1) into CDP-4-keto-3,6-
dideoxy-D-glucose
(2),
namely
-D-glucose-1-phosphate
cytidylyltransferase (Ep), CDP-D-glucose-4,6-dehydratase (Eod),
CDP-6-deoxy-L-threo-D-glycero-4-hexulose-3-dehydratase (E1),
and CDP-6-deoxy-L-threo-D-glycero-4-hexulose-3-dehydratase
reductase (E3) (Figure 1A).[15] Similar four genes are also part of
a homologous biosynthetic gene cluster of P. protegens and,
moreover, are arranged in the same order. Comparison of each
translated protein gives amino acid sequence identities of 40–
80% compared to the proteins from the yersiniose A biosynthetic
gene cluster (Figure 1B). From this observation, we conclude that
the substrate biosynthetically converted by PpYerE is likely CDP-
4-keto-3,6-dideoxy-D-glucose (2) or a related keto sugar.
Formation of phenylacetylcarbinol (PAC, 7)
Syntheses of PAC and derivatives thereof from benzaldehyde(s)
(5) and pyruvate (6) are well-known reactions catalyzed by many
ThDP-dependent enzymes.[16] All of the benzaldehydes 5a–f
tested were accepted and nearly quantitatively converted into the
R-enantiomer of PAC and respective derivatives (7a–f) by
PpYerE as well as YpYerE (Table 1). For the studied conversions,
the enantioselectivity of PpYerE was significantly lower, except
for the reaction of 3-hydroxybenzaldehyde (5f) with 6 (Table 1).
However, we observed a steady decrease of the enantiomeric
excess (ee) in the products over time. We also observed a
decrease of the ee of previously purified (R)-PAC (7a) (without
enzyme) after 1, 3, and 24 hours in buffer, to >99%, 98%, and
94% ee, respectively.
R1, R2
H, H
a
F, H
b
Cl, H Br, H
I, H
e
H, OH
f
c
d
PpYerE
84
76
74
74
68
75
ee (24h) [%] ( conv. (48h) [%])
YpYerE
(96)
97
(97)
94
(98)
94
(99)
94
(>99)
80
(>99)
61
ee (24h) [%] ( conv. (48h) [%])
(82)
(98)
(97)
(97)
(95)
(88)
Table 1. Enzymatic synthesis of PAC and derivatives (7a–f) by PpYerE and
YpYerE. Conditions: benzaldehyde or derivative (5a–f. 20 mM), pyruvate (6, 50
mM), PpYerE [1.0 mg/mL (Bradford)], YpYerE (crude extract), 20–30% DMSO,
KPi buffer (50 mM, pH 8.0, containing 3 mM MgCl2, 0.5 mM ThDP), 30 °C.
Conversion was determined by GC-MS analysis, ee by chiral-phase HPLC.
Figure 1. (A) Biosynthesis of CDP-yersiniose A (4) according to Chen et al. (B)
Schematic representation of the biosynthetic gene cluster of ascarylose and
yersiniose A compared to a section of the cluster harboring PpYerE. a)
Percentage of identity between translated genes of Yersinia pseudotuberculosis
O:VI and Pseudomonas protegens.
Formation of acetolactate and acetoin
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
Scheme 1. Ketone substrates 8–11 and products 12–15 of PpYerE-catalyzed
transformations. Conditions: ketone (8–11, 20 mM), pyruvate (6, 50 mM),
PpYerE (1 mg/mL), KPi buffer (50 mM, pH 8.0, containing 3 mM MgCl2, 0.5 mM
ThDP), 20% DMSO (only added for assays with -tetralone (10) and 1-
phenoxypropan-2-one [11]), total volume 1.5 mL; [a] Cyclohexane-1,2-dione (9,
100 mM), pyruvate (6, 20 mM). Conversions (%) were determined after 24 h by
GC-MS analysis; ee values (%, in brackets) were determined by chiral-phase
HPLC analysis. Results for YpYerE were published by Lehwald et al. and added
for comparison.[2]
Incubation of PpYerE with pyruvate (6) as the sole substrate
yielded nearly racemic acetoin. Experiments with [2-13C]-6
showed the enzymatic synthesis of acetolactate, followed by a
probably non-enzymatic decarboxylation. This is in line with the
(low) enantioselectivity observed for the enzymatic reaction of two
molecules of acetaldehyde [(S)-acetoin, 26% ee] and of pyruvate
(6) and acetaldehyde [(S)-acetoin, 21% ee].
For YpYerE, formation of both products of the homocoupling
of pyruvate (6), acetolactate and acetoin, were shown, too. 1H
NMR data (in D2O) indicated the quantitative synthesis of
acetolactate, and we also identified small amounts of acetoin.
After extraction with ethyl acetate, only the decarboxylation
product acetoin (<5% ee) was identified, which probably is due to
the low solubility of acetolactate in ethyl acetate. Hence, the
In summary, YpYerE and PpYerE exhibit high amino acid
sequence identity, are located in homologous biosynthetic gene
clusters, and behave largely similar with respect to substrate
range and stereoselectivity in biocatalytic C–C bond formations.
Most probably, these enzymes are homologues and are involved
in the biosynthesis of yersiniose or a structurally related derivative.
Hence, we set out to elucidate the structure of both enzymes, but
obtained crystals only of PpYerE.
mechanism of acetolactate formation, followed by
a non-
enzymatic decarboxylation yielding nearly racemic acetoin, is
assumed to be the same for both enzymes.
Structure determination and quality of the models
Cross-coupling of aldehydes and (di)ketones
Crystals of N-terminally His-tagged, full-length PpYerE (2 × 588
amino acids, 129.7 kDa) were grown in the presence of 3 mM
MgCl2 and 1 mM ThDP, using PEG 4000 as precipitant. They
diffract to 1.55 Å resolution, belong to space group P21, and
contain two polypeptide chains (one homodimer) per asymmetric
unit. With unit cell dimension a being equal in length to c, these
crystals fulfill one of the possible sets of requirements that allow
twinning in the P21 space group, and twinning was indeed
observed. Structure determination was achieved by molecular
replacement using a subunit of Arabidopsis thaliana AHAS (PDB
accession code 1ybh, 32% sequence identity) as search model.
The final model was refined to Rcryst/Rfree values of
15.1%/17.4% and has excellent stereochemistry (see Table S1 in
the Supporting Information). It contains one FAD, one ThDP, and
one magnesium ion per chain, and a total of 878 water molecules.
The observed electron density is well-defined for all cofactors and
almost the entire polypeptide chains, with the exception of a ~10-
residue stretch that links the first two domains in both subunits
(A183–192, B182–194), and parts of a loop and helix located at
the entrance to the active site in chain B (B483–491). The N-
terminal His tags, as well as the three C-terminal residues of chain
A, are not visible in the electron density map.
For the synthesis of tertiary alcohols, different cyclic and acyclic
ketones were studied, most of them known substrates for YpYerE
(for all tested ketones, see Table S3 in the Supporting
Information). Pyruvate (6) was used as donor substrate.
Selectivities for cyclic ketones differed between PpYerE and
YpYerE: with PpYerE, the conversion of cyclohexanone (8) was
lower, while cyclopentanone and cycloheptanone were not
accepted at all. Both YpYerE and PpYerE did not accept
cyclooctanone and -tetralone (3,4-dihydro-1(2H)-naphthalen-
one). Cyclohexane-1,2-dione (9) was a substrate of PpYerE,
resulting in low conversion (5%) even after variation of substrate
ratios. -Tetralone (3,4-dihydro-2(1H)-naphthalenone, 10) was
converted by PpYerE to give 37% of the tertiary alcohol 14 with
48% ee. In the analogous transformation by YpYerE, the
conversion was higher (75%) yet the enantioselectivity was lower
(9% ee). Hexane-3,4-dione (18) was converted by PpYerE with a
lower yield (Table 2), compared to YpYerE. In this case, both
products showed the same absolute S-stereochemistry with
inverse specific optical rotation compared to the product obtained
by Ao:DCPIP OR catalysis.[6] Conversion and enantioselectivity
with 1-phenoxypropan-2-one (11) were lower with PpYerE
catalysis than with YpYerE (Scheme 1, Table S3 in the Supporting
Information).
Overall structure
The structures of the two crystallographically independent
monomers per asymmetric unit are essentially identical, with a
root mean square deviation (rmsd) for the corresponding C-
positions of 0.22 Å. The fold of the PpYerE subunit closely
resembles that of other ThDP-dependent enzymes. It is divided
into three domains called Pyr or  domain (1–182), central or 
domain (193–363), and PP or  domain (364–568). All have an
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
/-architecture comprising a central six-stranded, parallel -
sheet with several helices packed against both sides (Figure 2).
Figure 2. Overall structure of PpYerE. (A) The PpYerE dimer. Both subunits
are shown in cartoon representation, with subunit A colored cream, and subunit
B in different shades of green for the three distinct domains (Pyr, forest green;
central, green; PP, yellow-green) and covered by a correspondingly colored
semitransparent surface. The cofactors are shown as sticks, with ThDP in red
and dark blue, and FAD in pink and cyan for subunits A and B, respectively. (B)
The PpYerE subunit in cartoon representation. The three distinct domains are
labelled and colored differently. ThDP and FAD are shown as sticks in brown
and hot pink, respectively. The magnesium ion is shown as a black sphere, and
the metal-coordinating water molecule as a red sphere. Amino acids connecting
the Pyr and central domains are not resolved in electron density.
Figure 3. ThDP modification observed in the crystal structure of PpYerE.
Stereo view of the final 2Fo–Fc map contoured at 1  (blue), and the final Fo–Fc
map contoured at +3  (green) and –3  (red), respectively, around ThDP. Both
maps indicate modification of the cofactor at the C2-position.
For the present structure of PpYerE, additional electron density
was observed extending from the thiazolium ring of ThDP
indicating covalent attachment of an unknown moiety, possibly an
oxygen atom (Figure 3). Thiazolone derivatives of ThDP have
also been observed in other crystal structures of ThDP-dependent
enzymes and are likely a result of the exposure to intense X-ray
radiation during data collection.[17]
PpYerE is a homodimer in the crystalline state as well as in
solution (Figure 2A), a feature it shares with, e.g., transketolase
but not the majority of known ThDP-dependent enzymes, which
instead are homotetramers assembled from two such dimers. The
mode of dimer assembly is shared by all ThDP-dependent
enzymes and is also conserved in PpYerE, with the interface
being formed between the respective Pyr and PP domains of the
two monomers. It buries ~3400 Å2 (15.8%) of the solvent-
accessible surface area of each monomer and involves formation
of 57 hydrogen bonds and two salt bridges.
The closest structural homologue to PpYerE is, according to the
criteria used by PDBeFOLD, the Pseudomonas fluorescens
benzaldehyde lyase (BAL; PDB entry 2ag0, 25.3% sequence
identity), showing a Q-score of 0.60 and an rmsd of 1.8 Å for 501
aligned C-atoms. BAL is a homotetrameric enzyme that does not
bind FAD. Accordingly, the largest structural deviations are
observed in the central FAD-binding domain and regions involved
in subunit interface formation. The DALI server identifies the
homotetrameric A. thaliana AHAS, the search model used in
molecular replacement, as the closest structural homologue, with
a Z-score of 43.4 (PDB accession code 1yi0, rmsd = 2.2 Å for 539
aligned C-positions).
FAD binding
FAD is located in a deep crevice at the interface between all
three domains, with most of the enzyme residues involved in its
binding originating from the central domain, which is therefore
also termed FAD-binding domain. The surface of this crevice is
covered by a ~20 residue segment inserted between strand 4
and helix 5 of the Pyr domain that protrudes from the second
subunit in the dimer. It largely sequesters what would otherwise
be the more exposed face of the flavin cofactor from the
surrounding solvent. FAD is bound in an extended conformation
with the flavin ring placed near the thiazolium ring of ThDP, which
is basically identical to the conformation observed in the crystal
structures
of
other
flavin-containing
ThDP-dependent
enzymes.[18]All residues located within van der Waals distance to
the FAD are shown in Figure S3 in the Supporting information.
As for ScAHAS, no catalytic function for FAD has been
discerned as no electron transfer to or from FAD is required for
the catalyzed carboligation reaction. The hypothesis of FAD being
an evolutionary relict from POX-like ancestors formulated for
AHAS by Chang and Cronan may thus also apply to FAD binding
by PpYerE.[19]
ThDP binding
Binding of the ThDP cofactor is mediated by the Pyr and PP
domains, with residues of the former surrounding the pyrimidine
ring whereas the diphosphate is, primarily via a Mg2+ ion,
anchored in the PP domain of the respective other subunit
(Figure 3), resulting in binding of two ThDP molecules per
homodimer. The two hydrogen bonds between protein and the
ThDP pyrimidine ring, formed by E48 to the N1’ and by G420 to
the N4’ nitrogen atoms, are crucial for catalytic activity. They are
conserved in almost all ThDP-dependent enzymes and are known
to induce the 1’,4’-imino tautomer[15] and orient the 4’-imino group,
respectively, which facilitates the deprotonation of the thiazolium
C2 atom to a carbanion as the first step in the reaction cycle. In
PpYerE, M422 is the hydrophobic residue that stabilizes the
canonical V-conformation of the ThDP.
The substrate-binding site
The active site can be accessed by substrates through a relatively
narrow tunnel, whose walls are formed by residues from the PP
and FAD-binding domains of one subunit and the Pyr domain of
the other subunit. The catalytically crucial C2 and N4’ atoms of
ThDP as well as the C7-methyl group of FAD are thereby most
solvent-exposed. Part of the helix 22 and the following loop that
contribute to formation of the active site entrance are disordered
in the B subunit, which does, however, not affect the regions of
the active site nearest to the cofactors (Figure 4). It is feasible that
the partial disorder of this residue stretch is indicative of a mobility
with functional importance, e.g., in terms of facilitating exchange
of products by substrates after the reaction has occurred and
closing off the active site during catalysis, or of adjusting active
site dimensions to the sizes and shapes of different acceptor
substrates, especially as the corresponding residue stretch in the
crystal structure of ScAHAS is also partially disordered.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
that is blocked by amino acid side chains.[22] The pocket
can be opened by a few amino acid exchanges
enabling antiparallel orientation of substrate side
chains, which is a prerequisite for S-selectivity. This
concept has been successfully applied to the R-
selective pyruvate decarboxylase from Acetobacter
pasteurianus (ApPDC) as well as 2-succinyl-5-
enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate
synthase from both E. coli and B. subtilis (EcMenD,
BsMenD), yielding S-selective variants.[23,24]
In PpYerE the S-pocket is defined by L22, I23, G24,
G25, L476, and V479 (see Figure S2A in the
Supporting Information). V479 is located at the
standard position 477 (according to the standard
numbering scheme for ThDP-dependent decar-
boxylases of the BioCatNet database system), which
was found to be decisive for the inversion of
enantioselectivity of ApPDC, EcMenD, and BsMenD.[25]
In PpYerE the size of the S-pocket is restricted by the
backbone of helix 22 that is shifted about 2 Å into the
area of the potential pocket compared to ApPDC.
Modelling of benzaldehyde (5a) into a pocket that was
opened by a single amino acid exchange of V479 to
alanine and glycine, respectively, indicated that the
available space is not sufficient for antiparallel binding
of 5a, as it clashes with the -helix (see Figure S2B in the
Supporting Information). This was confirmed by experimental
results in which the ee value of (R)-PAC (7a) was not altered for
the variant V479G and only slightly reduced for variant V479A
(Table 2). Consequently, propanal (16) and n-pentanal (17) were
tested as smaller, aliphatic acceptors. The respective products, 3-
hydroxypentan-2-one (19) and 3-hydroxyheptan-2-one (20), are
nature-identical flavor ingredients used for aroma production.[26]
The absolute configuration was determined by CD spectroscopy,
in accordance to the spectrum of (R)-19 recorded by Vinogradov
et al. (see Figure S3 in the Supporting Information).[27]
With respect to the product 3-hydroxypentan-2-one (19), wild-
type PpYerE catalyzed the formation of the S-enantiomer with
40% ee, which could be further increased by the mutation V479A
to 57% ee. Mutation V479G, on the other hand, inverted the
enantioselectivity to an excess of the R-enantiomer (55% ee)
(Table 2). For product 3-hydroxyheptan-2-one (20), wild-type
PpYerE gave a low excess (16%) of the R-enantiomer. Again,
mutation V479A increased the relative amount of the S-
enantiomer (31% ee), while mutation V479G increased the
relative amount of the R-enantiomer (29% ee) (Table 2). Wild-type
PpYerE showed high n-pentanal (17) conversion of 98% within
1 hour. The conversion with variant V479A and V479G was
reduced to 30% and 14%, respectively, indicating a lower
enzymatic activity and/or stability of the variants.
Figure 4. Comparison of PpYerE with AtAHAS (PDB ID: 1yi0).[20] (A) Stereo
view of the superimposed subunits of PpYerE (green) and AtAHAS (gray) in
cartoon representation. ThDP and FAD bound to PpYerE are shown as sticks
in brown and pink, respectively. The sulfonylurea-type inhibitor bound to the
AtAHAS crystal structure selected for the superimposition is shown as sticks in
blue, the ThDP derivative and FAD as black sticks. (B) Stereo view of the
superimposed PpYerE and AtAHAS substrate-binding sites. The substrate- and
cofactor-binding cavity of PpYerE is outlined by a semitransparent surface, and
all PpYerE residues forming its walls near the substrate-binding site are shown
with carbon atoms in yellow when derived from subunit A and green when
belonging to subunit B. ThDP and FAD are depicted with carbon atoms in cyan.
Helix 22 is shown as cartoon, with the residue stretch not resolved in electron
density in one of the subunits indicated in red. The corresponding AtAHAS
secondary structure element is shown in white, and cofactors and amino acid
residues with carbon atoms in white. The labels indicate the PpYerE residue,
with the residue type of the corresponding AtAHAS residue given after a slash
only if it is not conserved.
In the immediate vicinity of the cofactors, the active site of PpYerE
is highly similar to that of AtAHAS (Figure 4). From the residues
forming this part of the cavity, all but L22’ and I23’, are conserved
at (almost) the same position in the two enzymes. The degree of
sequence conservation is significantly lower at the outskirts of the
substrate-binding cavity towards the herbicide-binding site
identified for plant (and also yeast) AHAS. Here, the PpYerE
residues M26, F482, and Y264 are replaced by alanine,
tryptophan, and histidine, respectively, which may indicate
differing substrate specificities or inhibition behavior of both
enzymes.
From the acceptor ketones described above, hexane-3,4-
dione (18) was selected for further investigations with the PpYerE
variants, as it is less sterically demanding than benzaldehyde (5a)
and therefore could theoretically fit into the engineered acceptor-
binding pocket. Due to the nomenclature based on the CIP priority
rules, the parallel orientation of the donor side chain and the
longer acceptor side chain of 18 (-CO-CH2-CH3) before C–C bond
Mutagenesis studies to invert enantioselectivity
PpYerE catalyzes the formation of (R)-PAC (7a) with high
enantioselectivity (Table 1) which can be explained by a parallel
orientation of the donor (pyruvate, 6) and acceptor (benzaldehyde,
5a) side chains prior to C–C bond formation.[21] In many ThDP-
dependent decarboxylases a potential antiparallel acceptor-
binding pocket (also referred to as the “S-pocket”) has been found
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
formation results in the formation of the S-enantiomer. It was thus
expected that opening of the S-pocket would increase the
formation of the R-enantiomer of 3-hydroxy-3-ethyl-hexane-2,4-
dione (21). Wild-type PpYerE showed low enantioselectivity,
which was inverted for variant V479A; ee values for both
conversions were below 30%. Variant V479G gave almost
racemic 21.
Even though high enantioselectivities could not be obtained,
variant V479A showed a shift of the enantioselectivity towards
increased formation of the S-product in all reactions. This
indicates that the pocket concept is generally applicable, although
a high percentage of the acceptor substrate still binds in parallel
orientation. The enantioselectivity of the variants could probably
be increased by destabilization of the parallel orientation, as
described by Westphal et al.[24]
Interestingly, for variant V479G the enantioselectivity was
shifted towards the R-product when propanal (16) and n-pentanal
(17) were used as acceptor substrates, even though more space
should be available within the alternative acceptor-binding pocket.
This might be due to an accumulation of glycine residues, as
glycine is also found in positions 477 and 481 in helix 22 as well
as in positions 24 and 25, lining the potential pocket (see Figure
S2A in the Supporting Information). It can be assumed that
introduction of another glycine destabilizes the secondary
structure and leads to the disintegration of the S-pocket.
Conclusions
Although many ThDP-dependent enzymes catalyze carboligation
reactions of pyruvate as donor with aldehydes as acceptors, only
few ThDP-dependent enzymes, such as YpYerE and PpYerE
(59% amino acid identity), utilize ketones to form tertiary alcohols.
Numerous attempts to crystallize YpYerE have not been
successful; nevertheless, we were able to crystallize PpYerE and
determine its structure to 1.55 Å resolution. PpYerE displays
similar catalytic activity as YpYerE, i.e. benzaldehyde and
derivatives, as well as cyclic and acyclic ketones and 1,2-
diketones, as acceptor and pyruvate as donor were converted into
the respective secondary or tertiary 2-hydroxy ketones. In the
frame of site-directed mutagenesis studies, the S-pocket concept
could also be applied to PpYerE, but was limited for larger
acceptor substrates due to the structurally restricted size of the
pocket.
The detailed structural and catalytic characterization of
PpYerE and comparison with the homologous YpYerE shows that
both enzymes share the same amino acid residues in their active
site, have a very similar substrate range, and are encoded by
genes that are located in highly similar biosynthetic gene clusters.
Hence, our results pave the way for the identification of putatively
related enzymes which should be able to accept ketones as
substrates, too. Moreover, the rational modification of ThDP-
dependent decarboxylases and related enzymes resulting in
ketone-accepting enzymes, as previously shown by Loschonsky
et al. for the variant CDH-H28A/N484A, could enlarge the product
range towards sterically demanding 2-hydroxy ketones.[12]
R1,
R2
C6H5,
H
C2H5,
H
n-C4H9,
H
C3H5O,
C2H5
19*
PpYerE
>99 (R)
(99)
40 (S)
(n.d.)
16 (R)
(98)
Experimental Section
ee [%] (conv. [%])
(n.d.)
Construction of plasmids for gene expression
PpYerE V479A
95 (R)
57 (S)
31 (S)
-27*
The gene ilvb (pp-yerE, UniProtKB accession number Q4K6F7_PSEF5)
was amplified via PCR according to standard protocols from the genomic
DNA isolated from Pseudomonas protegens Pf-5 that was kindly provided
by the Department of Environmental Systems Science ETH (Zürich,
ee [%] (conv. [%])
(88)
(n.d.)
(30)
(n.d.)
PpYerE V479G
>99 (R)
55 (R)
29 (R)
-4*
ee [%] (conv. [%])
(97)
(n.d.)
(14)
(n.d.)
Switzerland).
AATAATCATATGAAAGCCTCGGATGCAGTAGC-3’ introducing an NdeI
restriction site and the reverse primer 5’-
Therefore,
the
forward
primer
5’-
Table 2. Enantioselectivity of PpYerE variants for the carboligation of pyruvate
(6) with benzaldehyde (5a), propanal (16), n-pentanal (17), and 3,4-hexandione
(18). Conditions: acceptor substrate (20–50 mM), pyruvate (6, 50 mM), PpYerE
variant (1.0 mg/mL), 30 °C. Conversions were determined after 1 h by HPLC for
aromatic products and GC-MS for aliphatic products, respectively, ee values
after 1 h by chiral-phase HPLC or chiral-phase GC-MS analysis. *For the
reactions with hexane-3,4-dione (18) the ee value was measured after 24 h.
AATAATGCGGCCGCTCCATCGTGTCGGGGTGATTG-3’ introducing a
NotI restriction site were used (restriction sites underlined). The PCR
product was then inserted as appropriate into the expression vector
pET22b(+) through restriction and ligation steps which resulted in the
plasmid pET22b*Pf-5 that contained the gene coding for the target protein
Ilvb (PpYerE) with a His6 tag attached at the C-terminus. Moreover, to
obtain the N-terminally His6-tag-carrying protein variant, the gene was
further subcloned into the pET28a(+) expression vector. For this purpose,
pET22b*Pf-5 was used as template and the gene was amplified via PCR
As exchange of V479 by alanine and glycine did not result in
an effective opening of the S-pocket in PpYerE, further amino
acids lining the pocket were exchanged to smaller residues which
was also not successful, as tested for products 3-hydroxypentan-
2-one (19) and 3-hydroxyheptan-2-one (20). Replacement of
L476 by alanine and glycine yielded inactive variants. Exchange
of I23 by alanine and glycine did not have a significant impact on
the ee of 19 and increased the formation of the R-enantiomer for
20.
using
the
forward
primer
and
5’-TATATATACATATGAAAG
the reverse primer 5’-
CCTCGGATGCAGTAGC-3’
TATATAGGATCCTTATCCATCGTGTCGGG-3’ harboring an NdeI and
BamHI restriction site (underlined), respectively. This amplified cDNA was
inserted into the vector backbone similarly as described above, yielding
vector pET28b*Pf-5 encoding the target protein Ilvb (PpYerE) with a N-
terminal His6-tag. In frame cloning and correctness of the sequence of the
genes in all vector constructs used for further experiments was confirmed
by DNA sequencing (GATC Biotech, Germany). For plasmid amplification,
Escherichia coli DH5 cells were used routinely.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
Expression of the gene pp-yerE constructs
Biocatalytic transformations
After transformation of the pET22b*Pf-5 or the pET28b*Pf-5 into
BL21(DE3) cells, single clones were used to inoculate 20 mL lysogeny
broth (LB) nutrient medium containing 100 g/mL ampicillin or 50 g/mL
kanamycin, respectively (representative example). These cultures were
grown overnight at 37 °C under agitation (120–180 rpm). After dilution
(1:100) in LB medium, 1L cultures were cultivated under aerobic conditions
in baffled flasks at 37 °C with agitation (120–180 rpm) until an optical
density (OD600) of 0.6 was reached. Then, IPTG was added to a final
concentration of 0.1 mM and expression was carried out overnight at 25 °C
and 120–180 rpm. Cells were harvested via centrifugation at 6000 g at
4 °C for 45–60 min. Lysis of the cells was accomplished by sonication (duty
cycle = 50%, control = 5, Sonifier 250, Branson, Danbury USA) by using 3
times 30 s pulses each with intervals of 30 s on ice in between (in KPi
buffer A: 100 mM, pH 8.0 containing 0.5 mM ThDP, 3 mM MgCl2).
Alternatively, cells were lysed by using CelLytic B Cell Lysis Reagent
(Sigma Aldrich) and incubation for 30–60 min at 200 rpm and 37 °C. Cell
debris was removed by centrifugation (6000 g at 4 °C for 45–60 min). The
supernatant was used for purification of the target protein by immobilized
metal ion affinity chromatography (IMAC) on an Ni-NTA matrix using KPi
buffer B (50 mM, pH 8.0 containing 0.5 mM ThDP and 3 mM MgCl2
supplemented with different concentrations of imidazole: 20 mM and 50
mM for the consecutive washing steps and 300 mM for elution of the target
protein). The target protein eluate was desalted by using PD-10 columns
(SephadexTM G-25 M, GE Healthcare) according to a protocol of the
manufacturer and using KPi buffer A. For crystallization, target proteins
were purified through an additional step via gel permeation
chromatography. Thus, a Superdex S200 10/300 column was equilibrated
overnight with Tris buffer (20 mM, 3 mM MgCl2, 1 mM ThDP, pH 8.0) and a
flow rate of 0.3 mL/min. After loading the protein sample obtained from the
previous IMAC purification and additional desalting, the flow rate was
increased to 0.5 mL/min. The target protein appeared as dimers in solution
and the respective fractions were collected, pooled, and then concentrated
by using AmiconTM Ultra centrifugal filter units (30 K) and Vivaspin 6
centrifugal devices (30 K), each at 3700 g for 25 min at 4 °C. Purification
to homogeneity was confirmed by SDS-PAGE analysis. Full functionality
of the enzymes was demonstrated by measuring activity towards a range
of substrates for both the C-terminally and N-terminally His6-tagged
enzyme.
For all carboligation reactions, purified and lyophilized PpYerE
variants were used. The final protein concentration was adjusted to
1 mg/mL. The preparations were incubated at 30 °C. Samples were
analyzed by chiral-phase HPLC after dilution in acetonitrile or chiral-phase
GC-MS after extraction with ethyl acetate (see the Supporting Information).
Buffer conditions and substrate concentrations depended on the
respective acceptor substrate:
Transformations comparing YpYerE and PpYerE (Table 1 and
Scheme 1)
Pyruvate (6) + benzaldehydes (5a–f): 50 mM potassium phosphate,
0.5 mM ThDP, 3 mM MgCl2, 50 mM pyruvate (6), 20 mM benzaldehyde or
derivative (5a–f), YpYerE (crude extract), 20–30% DMSO, pH 8.0.
Pyruvate (6) + ketones (8–11): 50 mM potassium phosphate, 0.5 mM
ThDP, 3 mM MgCl2, 50 mM pyruvate (6), 20 mM ketone (8–11), 20% DMSO
(only added for assays with -tetralone and 1-phenoxypropan-2-one), pH
8.0.
Transformations with variants (Table 2)
Pyruvate (6) + benzaldehyde (5a): 50 mM potassium phosphate,
0.5 mM ThDP, 3 mM MgCl2, 50 mM pyruvate (6), 20 mM benzaldehyde (5a),
pH 8.
Pyruvate (6) + aliphatic aldehydes (16.17): 100 mM potassium
phosphate, 2.4 mM ThDP, 3 mM MgCl2, 50 M FAD, 50 mM pyruvate (6),
50 mM propanal (16) and n-pentanal (17), respectively, pH 8.
Pyruvate (6) + hexane-3,4-dione (18): 100 mM potassium phosphate,
2.4 mM ThDP, 3 mM MgCl2∙6 H2O, 50 M FAD, 50 mM pyruvate (6), 20 mM
hexane-3,4-dione (18), pH 7.
Crystallization
Initial crystallization screens were performed using 96-well sparse
matrix screens in sitting-drop setups at 20 °C. Several hits obtained with
the Morpheus crystallization screen (Molecular Dimensions) were further
optimized. The crystal used for data collection was obtained by sitting-drop
vapor diffusion against a 1 mL reservoir containing 10% (w/v) PEG 4000,
0.1 M MES/imidazole pH 6.5, 0.3 M MgCl2, 0.3 M CaCl2, and 17% (v/v)
glycerol at 20 °C. The 1 L drop consisted of equal volumes of reservoir
and protein solution (10.3 mg/mL PpYerE in 20 mM Tris pH 8.0, 3 mM
MgCl2, 1 mM ThDP). Crystals appeared within 1 day of equilibration.
Data collection, structure determination, and refinement
Site-directed mutagenesis
PpYerE variants were created with a PCR-based method using synthetic
mutagenic primers and the KOD Hot Start Polymerase Kit (Novagen). PCR
reactions were performed in 50 L and contained 3 ng/L template DNA,
0.5 M of each primer, 200 M dNTP mix, 1.5 mM MgSO4, and 0.02 U/L
KOD polymerase. The PCR temperature profile was: initial denaturation at
95 °C for 120 s, followed by 18 cycles of denaturation (95 °C for 20 s),
annealing (62 °C for 30 s), and elongation (72 °C for 220 s). A final
elongation step at 72 °C for 10 min was added. After the PCR, reaction
samples were purified using the GeneJET PCR Purification Kit (Thermo
Scientifc) according to the manufacturer’s information. Parental DNA was
removed by DpnI digestion on a 20 L scale. Up to 50 ng/L DNA was
incubated with 1 L FastDigest DpnI (Thermo Scientific) at 37 °C for 3 h.
Subsequently, an 8 L sample was used to transform chemically
competent E. coli DH5 cells (Thermo Scientific) according to a standard
protocol. Individual colonies were cultivated in LB medium overnight and
plasmids were isolated with the QIAprep Spin Miniprep Kit (Qiagen). Gene
sequences were confirmed by DNA sequencing (LGC Genomics, Berlin).
Crystals were flash-frozen without additional cryo-protection by rapid
transfer into liquid nitrogen. Crystallographic data were collected at 100 K
at beamline 14.1 of the Berliner Elektronenspeicherring-Gesellschaft für
Synchrotronstrahlung (BESSY) (Berlin, Germany). Details of data
collection and refinement statistics are given in the Supporting Information
(Table S1). The data were originally indexed in space group C2221, but
subsequent analysis revealed them to be twinned (twin operator –l, –k, –
h; twin fraction = 0.44). The crystal thus belongs to space group P21 with
unit cell parameters of a = c = 65.3 Å, b = 139.9 Å,  = 97.9°, and contains
two polypeptide chains (one homodimer) per asymmetric unit. Integration
of the diffraction data was carried out with iMOSFLM.[28] Intensities were
merged and scaled using AIMLESS[29] and structure factor amplitudes
were calculated with CTRUNCATE of the CCP4 suite of programs.[30]
Phases were obtained by molecular replacement using PHASER[31]
with a high-resolution limit of 2.5 Å and a subunit of Arabidopsis thaliana
acetohydroxy acid synthase (PDB accession code 1ybh, 32% sequence
identity)[20] as search model.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
Giovannini, O. Bortolini, A. Cavazzini, R. Greco, G. Fantin, A. Massi,
Green Chem. 2014, 16, 3904.
After initial rigid body and restraint refinement and extension of the
phases to 1.55 Å resolution, iterations of model-building in COOT[32] were
alternated with TLS and restrained refinement with intensity-based twin
refinement in REFMAC5[33] until crystallographic R factor and Rfree
converged. All reflections in the given resolution range (Table S1) were
used with the exception of 5% randomly selected reflections for monitoring
Rfree. Automatically determined local NCS restraints were applied. Water
molecules were added manually or using the search routine implemented
in COOT.
[8]
[9]
P. P. Giovannini, L. A. Lerin, M. Müller, G. Bernacchia, M. de Bastiani,
M. Catani, G. Di Carmine, A. Massi, Adv. Synth. Catal. 2016, 358, 2767.
T. Sehl, S. Bock, L. Marx, Z. Maugeri, L. Walter, R. Westphal, C. Vogel,
U. Menyes, M. Erhardt, M. Müller, M. Pohl, D. Röther, Green Chem.
2017, 19, 380.
[10] X. Jiang, H. Zhang, J. Yang, Y. Zheng, D. Feng, W. Liu, X. Xu, Y. Cao,
H. Zou, R. Zhang, T. Cheng, F. Jiao, M. Xian, Sci. Rep. 2013, 3, 2445.
[11] S. Loschonsky, S. Waltzer, S. Fraas, T. Wacker, S. L. A. Andrade, P. M.
H. Kroneck, M. Müller, ChemBioChem 2014, 15, 389.
[12] S. Loschonsky, T. Wacker, S. Waltzer, P. P. Giovannini, M. J. McLeish,
S. L. A. Andrade, M. Müller, Angew. Chem. Int. Ed. 2014, 53, 14402.
[13] Geneious alignment (alignment with free ends and gaps, cost matrix:
blosum62, gap open penalty: 12, gap extension penalty: 3, Geneious).
[14] J. S. Thorson, S. F. Lo, H. W. Liu, J. Am. Chem. Soc. 1993, 115, 5827.
[15] V. P. Miller, J. S. Thorson, O. Ploux, S. F. Lo, H. W. Liu, Biochemistry
1993, 32, 11934.
[16] a) M. Müller, G. A. Sprenger, M. Pohl, Curr. Opin. Chem. Biol. 2013, 17,
261; b) M. Beigi, E. Gauchenova, L. Walter, S. Waltzer, F. Bonina, T.
Stillger, D. Rother, M. Pohl, M. Müller, Chem. Eur. J. 2016, 22, 13999.
[17] C. L. Berthold, P. Moussatche, N. G. J. Richards, Y. Lindqvist, J. Biol.
Chem. 2005, 280, 41645.
The refined model contains residues 1–182, 193–565, and 1–181,
195–482, 492–568 for chains A and B, respectively, two ThDP, two FAD,
two Mg2+, and 876 water molecules. It has excellent stereochemistry as
determined with RAMPAGE[34] and MOLPROBITY,[35] with 98.8% of the
residues in the favored and none in outlier regions of the Ramachandran
plot, respectively. Further details about model quality are given in Table
S1. Structure comparisons and similarity searches were performed using
the DALI server[36] and PDBeFOLD[37] at the European Bioinformatics
Institute (http://www.ebi.ac.uk/msd-srv/ssm). Molecular surfaces were
analyzed with the Protein Interfaces, Surfaces and Assemblies Service at
the European Bioinformatics Institute.[38] Figures 2–4 were prepared with
PyMol (http://www.pymol.org). The crystallographic data and structure
have been deposited in the Protein Data Bank under accession code 5ahk.
[18] cf. S. S. Pang, R. G. Duggleby, L. W. Guddat, J. Mol. Biol. 2002, 317,
249.
[19] Y. Y. Chang, J. E. Cronan, J. Bacteriol. 1988, 170, 3937.
[20] J. A. McCourt, S. S. Pang, J. King-Scott, L. W. Guddat, R. G. Duggleby,
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 569.
[21] M. Knoll, M. Müller, J. Pleiss, M. Pohl, ChemBioChem 2006, 7, 1928.
[22] P. T. Anastas, Handbook of Green Chemistry, Wiley-VCH, Weinheim,
Germany, 2010.
[23] a) R. Westphal, S. Waltzer, U. Mackfeld, M. Widmann, J. Pleiss, M.
Beigi, M. Müller, D. Rother, M. Pohl, Chem. Commun. 2013, 49, 2061;
b) R. Westphal, S. Jansen, C. Vogel, J. Pleiss, M. Müller, D. Rother, M.
Pohl, ChemCatChem 2014, 6, 1082; c) D. Rother neé Gocke, G. Kolter,
T. Gerhards, C. L. Berthold, E. Gauchenova, M. Knoll, J. Pleiss, M.
Müller, G. Schneider, M. Pohl, ChemCatChem 2011, 3, 1587.
[24] R. Westphal, D. Hahn, U. Mackfeld, S. Waltzer, M. Beigi, M. Widmann,
C. Vogel, J. Pleiss, M. Müller, D. Rother, M. Pohl, ChemCatChem 2013,
5, 3587.
Dedicated to Prof. Wolfgang Steglich on the occasion of
his 85th birthday.
Acknowledgements
[25] a) C. Vogel, M. Widmann, M. Pohl, J. Pleiss, BMC Biochem. 2012, 13,
24; b) P. C. F. Buchholz, C. Vogel, W. Reusch, M. Pohl, D. Rother, A.
C. Spieß, J. Pleiss, ChemBioChem 2016, 17, 2093.
[26] a) A. J. Taylor, D. S. Mottram, Flavour Science. Recent Developments,
Elsevier Reference Monographs, s.l., 1997; b) F. Neuser, H. Zorn, R. G.
Berger, J. Agric. Food Chem. 2000, 48, 6191.
[27] M. Vinogradov, A. Kaplun, M. Vyazmensky, S. Engel, R. Golbik, K.
Tittmann, K. Uhlemann, L. Meshalkina, Z. Barak, G. Hübner et al.,
Analytical biochemistry 2005, 342, 126.
[28] A. G. W. Leslie, H. R. Powell (Ed.) Processing diffraction data with
MOSFLM, in Evolving Methods for Macromolecular Crystallography
(Eds. R. J. Read, J. L. Sussman), Springer, Dordrecht, 2007, pp. 40–52.
[29] P. R. Evans, Acta Crystallogr. Sect. D: Biol. Crystallogr. 2011, 67, 282.
[30] M. D. Winn, C. C. Ballard, K. D. Cowtan, E. J. Dodson, P. Emsley, P. R.
Evans, R. M. Keegan, E. B. Krissinel, A. G. W. Leslie, A. McCoy, S.
McNicholas, G. Murshudov, N. Pannu, E. Potterton, H. Powell, R. Read,
A. Vagin, K. Wilson, Acta Crystallogr. Sect. D: Biol. Crystallogr. 2011,
67, 235.
We thank Sascha Ferlaino, Dr. Maryam Beigi, and Dr. Tobias Wacker,
University of Freiburg, for technical support and helpful discussions, Ömer
Poyraz, Karolinska Institutet, for collection of the crystallographic data, and
Dr. Kay Greenfield for help in improving the manuscript. We acknowledge
access to synchrotron radiation at BESSY, Berlin, Germany and thank the
staff at beamline ID14-1 for support. This work was supported by a grant
from the Swedish Research Council (to G.S.), by grants from the DFG
(FOR 1296), and by a grant from the European Union’s Horizon 2020
Research and Innovation Programme under Grant Agreement No 635595.
Keywords: asymmetric synthesis • biocatalysis • biosynthesis •
C–C coupling • cross-coupling
[31] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C.
Storoni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658.
[32] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Acta Crystallogr. Sect.
D: Biol. Crystallogr. 2010, 66, 486.
[33] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect. D:
Biol. Crystallogr. 1997, 53, 240.
[1]
[2]
H. Chen, Z. Guo, H.-w. Liu, J. Am. Chem. Soc. 1998, 120, 11796.
P. Lehwald, M. Richter, C. Röhr, H.-w. Liu, M. Müller, Angew. Chem.
Int. Ed. 2010, 49, 2389.
[34] S. C. Lovell, I. W. Davis, W. B. Arendall, P. I. W. de Bakker, J. M. Word,
M. G. Prisant, J. S. Richardson, D. C. Richardson, Proteins 2003, 50,
437.
[3]
[4]
H. Grisebach, R. Schmid, Angew. Chem. Int. Ed. Engl. 1972, 11, 159.
P. P. Giovannini, P. Pedrini, V. Venturi, G. Fantin, A. Medici, J. Mol.
Catal. B: Enzym. 2010, 64, 113.
[35] V. B. Chen, W. B. Arendall, J. J. Headd, D. A. Keedy, R. M. Immormino,
G. J. Kapral, L. W. Murray, J. S. Richardson, D. C. Richardson, Acta
Crystallogr. Sect. D: Biol. Crystallogr. 2010, 66, 12.
[5]
G. Bernacchia, O. Bortolini, M. de Bastiani, L. A. Lerin, S. Loschonsky,
A. Massi, M. Müller, P. P. Giovannini, Angew. Chem. Int. Ed. 2015, 54,
7171.
[36] L. Holm, L. M. Laakso, Nucleic Acids Res. 2016, 44, W351-5.
[37] E. Krissinel, K. Henrick, Acta Crystallogr. Sect. D: Biol. Crystallogr.
2004, 60, 2256.
[6]
[7]
P. P. Giovannini, G. Fantin, A. Massi, V. Venturi, P. Pedrini, Org.
Biomol. Chem. 2011, 9, 8038.
a) O. Bortolini, P. P. Giovannini, S. Maietti, A. Massi, P. Pedrini, G.
Sacchetti, V. Venturi, J. Mol. Catal. B: Enzym. 2013, 85-86, 93; b) P. P.
[38] E. Krissinel, K. Henrick, J. Mol. Biol. 2007, 372, 774.
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.

10.1002/cbic.201800325
ChemBioChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
FULL PAPER
We have identified a new ThDP-
dependent enzyme catalyzing
the cross-condensation between
a keto acid and several activated
and non-activated ketones and
have solved its crystal structure.
S. Hampel, J.-P. Steitz, A. Baierl, P.
Lehwald, L. Wiesli, M. Richter, A. Fries,
M. Pohl, G. Schneider, D. Dobritzsch,*
and M. Müller*
Page No. – Page No.
Structural and Mutagenesis Studies
of the Thiamine-Dependent, Ketone-
Accepting YerE from Pseudomonas
protegens
For internal use, please do not delete. Submitted_Manuscript
This article is protected by copyright. All rights reserved.