Chorismatase Mechanisms Reveal Fundamentally Different Types of Reaction in a Single Conserved Protein Fold

Article abstract of DOI:10.1021/jacs.5b05559

Chorismatases are a class of chorismate-converting enzymes involved in the biosynthetic pathways of different natural products, many of them with interesting pharmaceutical characteristics. So far, three subfamilies of chorismatases are described that convert chorismate into different (dihydro-)benzoate derivatives (CH-FkbO, CH-Hyg5, and CH-XanB2). Until now, the detailed enzyme mechanism and the molecular basis for the different reaction products were unknown. Here we show that the CH-FkbO and CH-Hyg5 subfamilies share the same protein fold, but employ fundamentally different reaction mechanisms. While the FkbO reaction is a typical hydrolysis, the Hyg5 reaction proceeds intramolecularly, most likely via an arene oxide intermediate. Two nonconserved active site residues were identified that are responsible for the different reaction mechanisms in CH-FkbO and CH-Hyg5. Further, we propose an additional amino acid residue to be responsible for the discrimination of the CH-XanB2 subfamily, which catalyzes the formation of two different hydroxybenzoate regioisomers, likely in a single active site. A multiple sequence alignment shows that these three crucial amino acid positions are located in conserved motifs and can therefore be used to assign unknown chorismatases to the corresponding subfamily. (Chemical Equation Presented).

Full text of DOI:10.1021/jacs.5b05559

Article
pubs.acs.org/JACS
Chorismatase Mechanisms Reveal Fundamentally Different Types of
Reaction in a Single Conserved Protein Fold
Florian Hubrich,† Puneet Juneja,‡,§ Michael Muller,† Kay Diederichs,‡ Wolfram Welte,‡
̈
and Jennifer N. Andexer*,†
†Institute of Pharmaceutical Sciences, University of Freiburg, Albertstr. 25, 79104 Freiburg, Germany
‡Department of Biology, University of Konstanz, 78457 Konstanz, Germany
S
* Supporting Information
ABSTRACT: Chorismatases are a class of chorismate-converting
enzymes involved in the biosynthetic pathways of different natural
products, many of them with interesting pharmaceutical characteristics.
So far, three subfamilies of chorismatases are described that convert
chorismate into different (dihydro-)benzoate derivatives (CH-FkbO,
CH-Hyg5, and CH-XanB2). Until now, the detailed enzyme
mechanism and the molecular basis for the different reaction products
were unknown. Here we show that the CH-FkbO and CH-Hyg5
subfamilies share the same protein fold, but employ fundamentally
different reaction mechanisms. While the FkbO reaction is a typical
hydrolysis, the Hyg5 reaction proceeds intramolecularly, most likely via
an arene oxide intermediate. Two nonconserved active site residues
were identified that are responsible for the different reaction
mechanisms in CH-FkbO and CH-Hyg5. Further, we propose an
additional amino acid residue to be responsible for the discrimination of the CH-XanB2 subfamily, which catalyzes the formation
of two different hydroxybenzoate regioisomers, likely in a single active site. A multiple sequence alignment shows that these three
crucial amino acid positions are located in conserved motifs and can therefore be used to assign unknown chorismatases to the
corresponding subfamily.
INTRODUCTION
■
Chorismate (1) is a central branching point between primary
and secondary metabolism in most organisms except animals.
Compounds derived from chorismate (1) are important
metabolites or part of complex natural products such as
siderophores, polyketides and terpenoids, some of which are
used as pharmaceuticals.1,2 Additionally, products of choris-
mate-converting enzymes are often chiral and contain multiple
functional groups suitable for follow-up chemistry.3−5
In previous work, we described a new class of bacterial
chorismate-converting enzymes named chorismatases (CH).6
They catalyze the cleavage of chorismate (1) to pyruvate and
(dihydro-)benzoate derivatives (Figure 1); as yet, their
mechanism has not been elucidated in detail.7,8 Depending
on the products produced CH can be grouped in three distinct
subfamilies: CH-FkbO (chorismatases/chorismate hydrolases;
EC 3.3.2.13), such as FkbO from Streptomyces hygroscopicus
subsp. ascomyceticus (accession code AAF86394.1, UniProt
Q9KID9), produce 3,4-trans-dihydroxy-cyclohexa-1,5-dienecar-
boxylate (3,4-trans-CHD, 2). 3,4-trans-CHD (2) serves as
starter unit in the biosynthesis of important polyketidic
immunosuppressants such as rapamycin or tacrolimus.6 In
contrast, CH-Hyg5 (chorismatases/3-hydroxybenzoate syn-
thases; EC 4.1.3.45), such as Hyg5 from S. hygroscopicus
(accession code AAC38060.1, UniProt O30478) produce 3-
Figure 1. Reaction products of the different chorismatase subfamilies.
hydroxybenzoate (3-HBA, 3).6,9 CH-Hyg5 have been identified
in the biosynthetic gene clusters of compounds such as
cuevaene A and the immunosuppressant brasilicardin.6,9,10 3-
HBA (3) has also been described for CH-XanB2 (chorisma-
tases/hydroxybenzoate synthases), such as XanB2 from
Xanthomonas campestris pv campestris, but along with a
chorismate lyase-like (EC 4.1.3.40) activity resulting in 4-
Received: May 29, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
hydroxybenzoate (4-HBA, 4).8,11 CH-XanB2 are found in
different phytopathogenic Xanthomonas strains, where they are
involved in the biosynthesis of xanthomonadins.6,8,12,13
Despite catalyzing reactions with different outcomes, the
different subfamilies of chorismatases show high sequence
identity (26.3−40.4% on the protein level), notably around the
active site that is located at the C-terminus of the enzymes
(Figure S1). Regarding the reaction catalyzed, CH-FkbO are
very similar to isochorismatases, both performing a typical
hydrolysis reaction. However, their sequences and structures
are not related.6,14,15 Consequently, we suggested an
isochorismatase-like hydrolysis mechanism for the CH-FkbO
subfamily. This postulate has been supported by mutagenesis
experiments.7 CH-Hyg5 were assumed to perform a hydrolysis-
elimination mechanism with either 2 or enoyl benzoate as an
intermediate during 3-HBA (3) formation. Due to the trans
configuration of the vicinal diol of chorismate (1) and 3,4-trans-
CHD (2), the elimination reaction would have to proceed in a
syn fashion, which is chemically unfavored (Figure S2).
Nevertheless, syn elimination was described for the choris-
mate-converting MqnA leading to enoyl benzoate.16
Figure 3. Outcome of the chorismatase reactions performed by FkbO
and Hyg5 in 18O-labeled water to test the expected and postulated
mechanisms.19 (A) Hydrolysis reaction of FkbO and (B) intra-
molecular reaction of Hyg5.
Surprisingly, no incorporation of 18O could be detected in
either product [pyruvate and 3-HBA (3)] of the Hyg5 reaction
under these conditions (Figures 3 and S4). This result clearly
contradicts a hydrolysis−elimination mechanism for CH-Hyg5.
Additionally, the two potential intermediates, 3,4-trans-CHD
(2) and enoyl benzoate, were tested as substrates; neither of
them was converted by Hyg5 (Figure S2). The observation that
no external oxygen was incorporated into the Hyg5 products
indicated an intramolecular mechanism for the Hyg5 reaction.
A pericyclic mechanism, as described for chorismate-converting
enzymes producing the 2- and 4-regioisomeric hydroxyben-
zoates,17,18 can also be ruled out due to the position of the
hydroxyl group in 3-HBA (3). This suggested that CH-Hyg5
employ an alternative catalytic strategy to form the products.
In order to elucidate this strategy, we decided to perform
further assays with isotopically labeled chorismate (1).20 We
were specifically interested in gaining information about the
localization and fate of the chorismate C3 and C4 oxygen
atoms in the Hyg5 products 3-HBA (3) and pyruvate.
Therefore, we selectively introduced an 18O-label at the C4-
hydroxyl group of chorismate. The method was based on the
observation that FkbO and Hyg5 are strictly specific for
chorismate (1).21 This allows for a convenient assay of these
enzymes with in situ produced 4-18O-labeled chorismate using
the isochorismate synthase EntC from E. coli. EntC catalyzes
the introduction of a hydroxyl group at the chorismate C4 at
the expense of the isochorismate hydroxyl at C2 (and vice versa
for the reverse reaction), reaching a chorismate/isochorismate
equilibrium (3:2). Starting from pure nonlabeled isochorismate
in H218O, EntC produces a mixture of 2-18O-labeled
isochorismate and 4-18O-labeled chorismate. When added to
this reaction mixture, the chorismatases FkbO and Hyg5
entirely convert the present 4-18O-labeled chorismate derived
from isochorismate into the chorismatase reaction products
pyruvate and 3,4-trans-CHD (2) or 3-HBA (3), respectively
(Figure 4).
RESULTS AND DISCUSSION
■
For the CH-FkbO subfamily an isochorismatase-like hydrolysis
mechanism was suggested where the acidic active site residue
E338FkbO initially protonates the C3′ of chorismate (1,
methylene group at the enol ether moiety). This resulting
carbocation intermediate is nucleophilically attacked by
(activated) water leading to the formation of an unstable
tetrahedral intermediate (hemiketal), which decomposes
spontaneously into the products 3,4-trans-CHD (2) and
pyruvate (Figure 2).7
Figure 2. Hydrolysis mechanism of the CH-FkbO subfamily.
The importance of the postulated catalytic glutamate
(E338FkbO and E334Hyg5) was demonstrated previously for
FkbO using site-directed mutagenesis,7 and the corresponding
Hyg5 variant (E334QHyg5) also shows a complete loss of
activity (Figure S3). This observation supports that an initial
protonation step of the methylene group at C3′ is essential for
the reaction in both subfamilies.
Subsequently, the postulated hydrolysis mechanism was
investigated by performing assays in 18O-labeled water (H218O)
followed by LC-MS analysis. As expected, the FkbO reaction in
the presence of H218O resulted in incorporation of labeled
oxygen into the pyruvate moiety (Figures 3 and S4).
The assay confirmed that FkbO catalyzes the hydrolysis
reaction as proposed: the labeled oxygen from 4-18O-labeled
chorismate is retained in 3,4-trans-CHD (2), as well as the
labeled oxygen from the solvent H218O is incorporated into the
second product pyruvate (Figures 4, S4). However, when FkbO
was replaced by Hyg5 in otherwise identical conditions, the
4-18O-label was found in the 3-hydroxyl of 3-HBA (3), while
pyruvate was unlabeled, as seen before (Figures 4, S4).
B
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 4. In situ synthesis of selectively labeled 4-18O-chorismate (4-18O-1) and formation of 4-18O-3,4-trans-CHD (4-18O-2), as well as 2-18O-
pyruvate (FkbO) or 3-18O-3-HBA (3-18O-3) and pyruvate (Hyg5).
The complete migration of labeled oxygen from C4 in the
substrate chorismate (1) to C3 in the product 3-HBA (3)
supports an intramolecular mechanism, most probably featuring
an arene oxide as intermediate. The rearrangement of the arene
oxide to 3-HBA (3) could proceed via an NIH shift (Figure
5).22 Based on these results, other potential mechanistic
protein in three domains. The location and orientation of the
cocrystallized competitive inhibitor 3-(2-carboxyethyl) ben-
zoate24 in the active site clearly corresponds to the situation in
FkbO (Figure S5). Only two amino acids in close proximity to
the inhibitor differ in the two enzymes: a cysteine (C327Hyg5) in
Hyg5 corresponding to an alanine (A331FkbO) in FkbO, and a
glycine (G240Hyg5) in Hyg5 corresponding to another alanine
(A244FkbO) in FkbO (Figure 6).
How can these differences be responsible for the
fundamentally different molecular mechanisms of the two
chorismatase subfamilies? The active site geometry of Hyg5
suggests that the absence of the methyl group in G240Hyg5
(A244FkbO) causes a small but effective change in the
orientation of the catalytic glutamate side chain by abolishing
sterical hindrance (Figure 6). Due to this orientational change,
E334Hyg5 can activate the 4-hydroxyl moiety of chorismate for a
nucleophilic attack in Hyg5 by deprotonating it. The cysteine in
Hyg5 is likely involved in selective product formation by
directing the putative arene oxide opening to the C3 position
[as opposed to the C4 position leading to 4-HBA (4)] of the
ring via electrostatic and/or steric interaction (Figure 5).
To prove this structure−function relationship, sets of single
and double variants of the two enzymes were constructed to
investigate their influence on the CH-Hyg5 and CH-FkbO
activities (Figure 7). Both Hyg5 single variants, C327AHyg5 and
G240AHyg5, showed decreased activity compared to the
wildtype (6% and 55%, respectively), and resulted in a
complete loss of product selectivity. The FkbO product 3,4-
trans-CHD (2) as well as the XanB2 product 4-HBA (4) were
produced (Figures 7 and S3). The double variant C327A/
G240AHyg5 showed the same behavior, but along with a more
dramatic loss of activity [only 5% of chorismate (1) was
converted after 20 h]. This agrees with our hypothesis that
C327 and G240 are responsible for the selectivity of the CH-
Hyg5 reaction. In the case of the cysteine variant, loss of
electrostatic stabilization by the missing polar moiety and
reduced steric requirement of the side chain leads to the
additional production of 4-HBA (4), through an unselective
opening of the postulated arene oxide intermediate. The
involvement of C327 in directing the selectivity of product
formation was underlined by the crystal structure of the variant
C327SHyg5 with the product 3-HBA (3) (PDB-ID 5A3K, Figure
S5, Table S1). The variant showed comparable characteristics
Figure 5. Postulated arene oxide mechanism of CH-Hyg5: initial
protonation of the enolpyruvyl moiety, intramolecular nucleophilic
attack of the activated 4-hydroxyl group at the C3 position of the
cyclohexadiene ring, formation of an arene oxide intermediate, and a
hydrogen-shift during aromatization leads to the products 3-HBA (3)
and pyruvate.
alternatives such as syn elimination of water or a reaction via an
acetonide five-membered cyclic acetal intermediate could be
clearly ruled out (Figure S2). Examples for the occurrence of
arene oxides in other enzymatic reactions are given in ref 23.
In order to support this theory with structural data, the three-
dimensional structure of Hyg5 from S. hygroscopicus was solved
at 1.9 Å resolution (PDB-ID 5AG3, Figure S5 and Table S1).
Comparison with the structure of FkbO from S. hygroscopicus
subsp. ascomyceticus7 allowed an assessment of the molecular
basis for the different reactions. The overall structure of Hyg5
was almost identical to the one published for FkbO7 [root-
mean-square deviation (r.m.s.d.) of atomic positions = 1.18 Å
with bound inhibitor 3-(2-carboxyethyl) benzoate, r.m.s.d. =
1.40 Å without inhibitor], showing the same partition of the
C
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 6. Overlay of the active sites of FkbO (cyan) and Hyg5 (magenta) with the inhibitor 3-(2-carboxyethyl) benzoate in two different views. Left:
face-on view; right: edge-on view.
extended product range that was more pronounced for the
double variant.
The loss of selectivity experienced in the FkbO and Hyg5
variants suggests that product formation in the CH-XanB2
subfamily is unselective. In this case the opening of the arene
oxide intermediate is not exclusively directed to the C3
position, which results in a mixture of 3-HBA (3) and 4-HBA
(4) as products. Consequently, the formation of 3-HBA (3)
and 4-HBA (4) takes place in one single active site. This is in
contrast to previous postulations: based on sequence align-
ments and data received from experiments with full-length and
truncated XanB2 variants, Zhou et al. proposed that the CH-
XanB2 subfamily possesses two different active sites that
convert chorismate (1) into 3-HBA (3) and 4-HBA (4),
respectively.8
In the next step, we attempted to transfer the knowledge
gained for CH-FkbO and CH-Hyg5 to a general sequential
basis and to get insight in the discrimination between the three
chorismatase subfamilies (CH-FkbO, CH-Hyg5, CH-XanB2).
A Blastp search with an E-value < e−90 25 followed by a multiple
,
Figure 7. Product range of chorismatases and chorismatase variants.
The relative amounts of products [3,4-trans-CHD (2), 3-HBA (3), and
4-HBA (4)] are plotted in relation to all compounds (remaining
substrate and products) detected in the HPLC assay after 20 h
reaction time. All variants show a loss of product selectivity. Residual
activities of the variants in comparison to the corresponding wildtype
sequence alignment, revealed that all amino acid residues
contributing to the formation of the active site are embedded in
conserved sequence motifs (Figure S1). The only two amino
acids that differ between the CH-FkbO and CH-Hyg5
subfamilies are those described above. G240Hyg5 (A244FkbO) is
part of a conserved S-(A/G)-T-A motif, while C327Hyg5
(A331FkbO) is part of a conserved I-(A/C)-R motif (Figure
S1). As shown in this study, these two residues are solely
responsible for the selectivity of product formation in the CH-
FkbO and CH-Hyg5 subfamilies. Surprisingly, the CH-XanB2
subfamily shows the same characteristics as the CH-Hyg5
subfamily.
Nevertheless, based on further analysis of the crystal
structures and the multiple sequence alignment we identified
another subtle difference between CH-FkbO/CH-Hyg5 and
CH-XanB2 in a conserved motif close to the active site.
XanB2A173 (corresponding to G173Hyg5/G181FkbO) deviates
from an otherwise conserved sequence motif: P-A-A-T-(G/A)-
I-G (Figure S1). The higher steric requirement of the alanine
a
b
enzymes are given in the main text. Data from ref 7. FkbO_A331C
was inactive and not properly folded (see also Figure S3).
as observed for C327AHyg5, while the crystal structure revealed
coordination of the 3-HBA (3) hydroxyl group by the
introduced serine side chain (Figure S5). The formation of
the CH-FkbO product 3,4-trans-CHD (2) derived from the
competing hydrolysis reaction can be explained by both the
decreased velocity of the Hyg5 variants and the larger steric
requirements of the Hyg5 variant G240AHyg5 resulting in a
change of orientation of the catalytic glutamate (Figure 6). In
FkbO, the situation is similar: both the single variant
A244GFkbO (38% of wildtype activity) and the double variant
A244G/A331CFkbO (15% of wildtype activity) showed an
D
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
side chain could be responsible for a slightly different
positioning of the substrate in the CH-XanB2 subfamily
leading to an unselective opening of the postulated arene
oxide intermediate. Additionally, sequence comparison and
phylogenetic studies on 134 chorismatases based on these
conserved active site motifs enabled us to assign them as
members of the CH-FkbO, CH-Hyg5 and CH-XanB2
subfamilies (Figure S6).
ACKNOWLEDGMENTS
■
We thank Prof. Dr. Nikolaus Amrhein (ETH Zurich,
̈
Switzerland), Dr. Tobias J. Erb (MPI for Terrestrial Micro-
biology Marburg, Germany), Prof. Dr. Georg Fuchs (University
of Freiburg, Germany), and Dr. Michael Richter (Empa,
Switzerland) for useful input and critical reading of the
manuscript as well as Prof. Dr. V. Prasad Shastri for helpful
comments. We thank Dr. Alexander Fries (University of
Freiburg, Germany) for analytical support. We also thank the
beamline staff at the SLS MX beamlines for excellent support.
CONCLUSION
■
In conclusion, we showed that the chorismatase subfamilies
CH-FkbO and CH-Hyg5 catalyze two fundamentally different
molecular mechanisms despite their conserved protein
structure and almost identical active site architectures. The
mechanism proposed for the CH-Hyg5 subfamily can directly
be used to explain the broader product range of the CH-XanB2
subfamily, which likely converts chorismate via the same arene
oxide intermediate to 3-HBA (3) and 4-HBA (4) in one single
active site. We propose that in total only three amino acid
residues in conserved motifs are responsible for the
discrimination between the CH-FkbO, CH-Hyg5 and CH-
XanB2 subfamilies. Phylogenetic analysis also shows three
distinct clades in the phylogenetic tree (Figure S6); all of the
sequences analyzed contain the expected amino acid residues
that correspond to the three subfamilies. The active site
patterns described in this study complement and confirm the
phylogenetic method by providing a molecular explanation.
With this combined approach it is now possible to clearly assign
a new chorismatase to one of the subfamilies. The existence of
highly conserved motifs will greatly facilitate the detection of
further chorismatases and subsequently novel natural products.
Especially the CH-Hyg5 and CH-XanB2 subfamilies that
provide building blocks with meta-only substitutions will be
promising targets for the identification of new biosynthetic
gene clusters for such compounds.
REFERENCES
■
(1) Dosselaere, F.; Vanderleyden, J. Crit. Rev. Microbiol. 2001, 27, 75.
(2) Floss, H. G. Nat. Prod. Rep. 1997, 14, 433.
(3) Muller, R.; Breuer, M.; Wagener, A.; Schmidt, K.; Leistner, E.
̈
Microbiology 1996, 142, 1005.
(4) Bongaerts, J.; Esser, S.; Lorbach, V.; Al-Momani, L.; Muller, M.
A.; Franke, D.; Grondal, C.; Kurutsch, A.; Bujnicki, R.; Takors, R.;
Raeven, L.; Wubbolts, M.; Bovenberg, R.; Nieger, M.; Schurmann, M.;
Trachtmann, N.; Kozak, S.; Sprenger, G. A.; Muller, M. Angew. Chem.,
̈
̈
̈
Int. Ed. 2011, 50, 7781.
(5) Sprenger, G. A. Appl. Microbiol. Biotechnol. 2007, 75, 739.
(6) Andexer, J. N.; Kendrew, S. G.; Nur-e-Alam, M.; Lazos, O.;
Foster, T. A.; Zimmermann, A.-S.; Warneck, T. D.; Suthar, D.; Coates,
N. J.; Koehn, F. E.; Skotnicki, J. S.; Carter, G. T.; Gregory, M. A.;
Martin, C. J.; Moss, S. J.; Leadlay, P. F.; Wilkinson, B. Proc. Natl. Acad.
Sci. U. S. A. 2011, 108, 4776.
(7) Juneja, P.; Hubrich, F.; Diederichs, K.; Welte, W.; Andexer, J. N.
J. Mol. Biol. 2014, 426, 105.
(8) Zhou, L.; Wang, J.; Wang, J.; Poplawsky, A.; Lin, S.; Zhu, B.;
Chang, C.; Zhou, T.; Zhang, L.; He, Y. Mol. Microbiol. 2013, 87, 80.
(9) Jiang, Y.; Wang, H.; Lu, C.; Ding, Y.; Li, Y.; Shen, Y.
ChemBioChem 2013, 14, 1468.
(10) Hayashi, Y.; Matsuura, N.; Toshima, H.; Itoh, N.; Ishikawa, J.;
Mikami, Y.; Dairi, T. J. Antibiot. 2008, 61, 164.
(11) Schoner, T. A.; Fuchs, S. W.; Reinhold-Hurek, B.; Bode, H. B.
̈
PLoS One 2014, 9, e90922.
(12) He, Y.; Wu, J.; Zhou, L.; Yang, F.; He, Y.; Jiang, B.; Bai, L.; Xu,
Y.; Deng, Z.; Tang, J.; Zhang, L. Mol. Plant-Microbe Interact. 2011, 24,
948.
(13) Zhou, L.; Huang, T.-W.; Wang, J.-Y.; Sun, S.; Chen, G.;
Poplawsky, A.; He, Y.-W. Mol. Plant-Microbe Interact. 2013, 26, 1239.
(14) Rusnak, F.; Liu, J.; Quinn, N.; Berchtold, G. A.; Walsh, C. T.
Biochemistry 1990, 29, 1425.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b05559.
Experimental details, analytical data, and structural data
(PDF)
(15) Parsons, J. F.; Calabrese, K.; Eisenstein, E.; Ladner, J. E.
Biochemistry 2003, 42, 5684.
(16) Mahanta, N.; Fedoseyenko, D.; Dairi, T.; Begley, T. P. J. Am.
Chem. Soc. 2013, 135, 15318.
AUTHOR INFORMATION
■
(17) DeClue, M. S.; Baldridge, K. K.; Kunzler, D. E.; Kast, P.; Hilvert,
̈
D. J. Am. Chem. Soc. 2005, 127, 15002.
Corresponding Author
(18) Wright, S. K.; DeClue, M. S.; Mandal, A.; Lee, L.; Wiest, O.;
Cleland, W. W.; Hilvert, D. J. Am. Chem. Soc. 2005, 127, 12957.
(19) The experimental necessary following conversion of pyruvate
leading to D-lactate by lactate dehydrogenase (LDH) to avoid solvent
exchange is not shown but was done according to Teufel, R.; Gantert,
C.; Voss, M.; Eisenreich, W.; Haehnel, W.; Fuchs, G. J. Biol. Chem.
2011, 286, 11021.
*E-mail: Jennifer.andexer@pharmazie.uni-freiburg.de
Present Address
§Department of Structural Biochemistry, MPI of Molecular
Physiology, 44202 Dortmund, Germany.
Funding
This work was partly funded by the “Juniorprofessorenprog-
(20) Gustin, D. J.; Hilvert, D. J. Org. Chem. 1999, 64 (13), 4935.
ramm Baden-Wurttemberg” (Ministry for Science, Research
̈
(21) Hubrich, F.; Muller, M.; Andexer, J. N. J. Biotechnol. 2014, 191,
93.
(22) Yagi, H.; Jerina, D. M.; Kasperek, G. J.; Bruice, T. C. Proc. Natl.
Acad. Sci. U. S. A. 1972, 69, 1985.
(23) (a) Gunsior, M.; Ravel, J.; Challis, G. L.; Townsend, C. A.
Biochemistry 2004, 43, 663. (b) Boyd, D. R.; Sharma, N. D.; Harrison,
J. S.; Malone, J. F.; McRoberts, W. C.; Hamilton, J. T. G.; Harper, D.
B. Org. Biomol. Chem. 2008, 6, 1251. Teufel, R.; Mascaraque, V.;
̈
and Arts). The research leading to these results has received
funding from the European Community’s Seventh Framework
Programme (FP7/2007-2013) under BioStruct-X (grant agree-
ment N°283570).
Notes
The authors declare no competing financial interest.
E
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Ismail, W.; Voss, M.; Perera, J.; Eisenreich, W.; Haehnel, W.; Fuchs, G.
Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 14390.
(24) Hubrich, F.; Mordhorst, S.; Andexer, J. N. Bioorg. Med. Chem.
Lett. 2013, 23, 1477.
(25) Madden, T. L.; Tatusov, R. L.; Zhang, J. Methods Enzymol. 1996,
266, 131.
F
DOI: 10.1021/jacs.5b05559
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX