α-Hydroxy-β-keto acid rearrangement-decarboxylation: Impact on thiamine diphosphate-dependent enzymatic transformations

Article abstract of DOI:10.1039/c2ob26981c

The thiamine diphosphate (ThDP) dependent MenD catalyzes the reaction of α-ketoglutarate with pyruvate to selectively form 4-hydroxy-5-oxohexanoic acid 2, which seems to be inconsistent with the assumed acyl donor role of the physiological substrate α-KG. In contrast the reaction of α-ketoglutarate with acetaldehyde gives exclusively the expected 5-hydroxy-4-oxo regioisomer 1. These reactions were studied by NMR and CD spectroscopy, which revealed that with pyruvate the observed regioselectivity is due to the rearrangement-decarboxylation of the initially formed α-hydroxy-β-keto acid rather than a donor-acceptor substrate role variation. Further experiments with other ThDP-dependent enzymes, YerE, SucA, and CDH, verified that this degenerate decarboxylation can be linked to the reduced enantioselectivity of acyloins often observed in ThDP-dependent enzymatic transformations.

Full text of DOI:10.1039/c2ob26981c

Organic &
Biomolecular
Chemistry
View Article Online
View Journal | View Issue
COMMUNICATION
α-Hydroxy-β-keto acid rearrangement–decarboxylation:
impact on thiamine diphosphate-dependent enzymatic
transformations†
Cite this: Org. Biomol. Chem., 2013, 11,
252
Received 11th October 2012,
Accepted 8th November 2012
Maryam Beigi,a Sabrina Loschonsky,a Patrizia Lehwald,a Volker Brecht,a
DOI: 10.1039/c2ob26981c
Susana L. A. Andrade,b Finian J. Leeper,c Werner Hummeld and Michael Müller*a
www.rsc.org/obc
The thiamine diphosphate (ThDP) dependent MenD catalyzes the C–C bonds in this way is not easy and enzymes have the poten-
reaction of α-ketoglutarate with pyruvate to selectively form tial of producing chiral products with high enantiomeric
4-hydroxy-5-oxohexanoic acid 2, which seems to be inconsistent excess (ee).
with the assumed acyl donor role of the physiological substrate
MenD (2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-
α-KG. In contrast the reaction of α-ketoglutarate with acet- carboxylate synthase) from E. coli is a ThDP-dependent
aldehyde gives exclusively the expected 5-hydroxy-4-oxo regio- enzyme involved in menaquinone biosynthesis which uses
isomer 1. These reactions were studied by NMR and CD α-ketoglutarate (α-KG) as the natural acyl donor and
spectroscopy, which revealed that with pyruvate the observed
regioselectivity is due to the rearrangement–decarboxylation of
the initially formed α-hydroxy-β-keto acid rather than a donor–
acceptor substrate role variation. Further experiments with other
ThDP-dependent enzymes, YerE, SucA, and CDH, verified that this
degenerate decarboxylation can be linked to the reduced enantio-
selectivity of acyloins often observed in ThDP-dependent enzy-
matic transformations.
Introduction
Thiamine diphosphate (ThDP)-dependent enzymes participate
in numerous biosynthetic pathways and catalyze a wide range
of reactions mainly involved in C–C bond formation or clea-
vage adjacent to a carbonyl group.1,2 One of the basic model
reactions of this enzyme family is the decarboxylation of an
α-keto acid (the acyl donor) to give the active aldehyde inter-
mediate which adds to an aldehyde (the acyl acceptor) to form
the corresponding α-hydroxy ketone (an acyloin), Scheme 1a.
This type of reaction is synthetically attractive as formation of
aInstitute of Pharmaceutical Sciences, Albert-Ludwigs-Universität Freiburg,
Albertstraße 25, 79104 Freiburg, Germany.
E-mail: michael.mueller@pharmazie.uni-freiburg.de; Tel: +49-761-2036335
bInstitute of Organic Chemistry and Biochemistry and BIOSS Center for Biological
Signaling Studies, Albert-Ludwigs-Universität Freiburg, Albertstraße 21, 79104
Freiburg, Germany
cDepartment of Chemistry, University of Cambridge, Lensfield Road, Cambridge,
CB2 1EW, UK
dInstitute of Molecular Enzyme Technology at the Heinrich-Heine-University of
Scheme 1 (a) Mechanism of the ThDP-dependent formation of acyloins. (b)
Two regioisomeric acyloins are formed by MenD from α-KG depending on the
second substrate.
Düsseldorf, Research Centre, 52426 Jülich, Germany
†Electronic supplementary information (ESI) available: Further experimental
details, 13C NMR and CD spectra. See DOI: 10.1039/c2ob26981c
252 | Org. Biomol. Chem., 2013, 11, 252–256
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Communication
isochorismate as the acceptor.3–5 We have previously reported first formed intermediate is 2-hydroxy-2-methyl-3-oxohexane-
that acyloin 1 was formed as the sole product when acet- dioic acid 3 (Scheme 2b). This would be the expected product
aldehyde was incubated with α-KG and MenD, whereas the use if α-KG is the acyl donor and pyruvate the acceptor. On contin-
of pyruvate instead of acetaldehyde led exclusively to the for- ued monitoring of the incubation mixture, the intensity of the
mation of 2, an apparent reversal of the donor–acceptor roles signals for 3 decreased over time while signals for 2 increased
of the two substrates (Scheme 1b).5 However, the donor sub- and became dominant (see ESI†). Extraction after complete
strate spectrum of MenD is known to be narrow and pyruvate consumption of the reactants gave nearly racemic 2 as the
is accepted as a donor only in trace amounts by MenD when main product, in agreement with published data.5
α-KG is absent.5,6 Therefore, we reasoned that the regioselec-
In order to investigate whether the observed decarboxy-
tivity might not be due to a switch in the donor substrate, as it lation occurs through any detectable intermediate, 13C-labeled
appears at first sight, but instead to a selective rearrangement– α-KG was synthesized by oxidation of 13C-labeled L-glutamate
decarboxylation of an initially formed α-hydroxy-β-keto acid using L-glutamate dehydrogenase (L-GluDH) from Clostridium
product. Here we report on experiments using 13C-labeled sub- sp., coupled with NADH oxidase from Lactobacillus brevis7 for
strates to elucidate the mechanism of the selective formation simultaneous regeneration of the cofactor.8 Biotransformation
of 2, and the consequences for similar enzymatic and nonenzy- of the in situ formed [1,2-13C2]α-KG with MenD and pyruvate
matic transformations.
was then followed by 13C NMR spectroscopy. As expected, L-gluta-
mate was fully consumed by the coupled enzyme system;
however, some reduction of pyruvate to lactate was observed as
a side reaction. Hence, pyruvate was added in excess to avoid
substrate limitation.
Results and discussion
Repetition of the NMR experiments using both L-[1,2-13C2]
glutamate and [1-13C], [2-13C] or [1,2-13C2]pyruvate again
demonstrated formation of 3 as the first observed intermedi-
ate, which eventually underwent decarboxylation to give 2 as
the sole final product (Scheme 2c). Sequential 13C NMR
measurements were performed every hour for 48 hours. No
intermediate between 3 and 2 could be detected in these experi-
ments. It is worth noting that the observed regiochemistry of
the decarboxylation product is unlikely to be of thermo-
dynamic origin as incubation of α-KG with acetaldehyde in the
presence of MenD leads to the formation of the isomeric 1 as
the sole product without any detectable traces of 2. Upon
further incubation in D2O for 48 hours, 2 proved to be
unstable. In contrast, 1 was quite stable and the only deuter-
ium incorporation observed was at C-3 (see ESI†).
Incubation of α-KG with [1,2-13C2]-acetaldehyde and MenD led,
as expected, to the formation of 1 as the sole product, with the
13C labels exclusively in the CH3CHOH group (Scheme 2a). No
product 2 was observed.
The use of α-KG with [2-13C] or [1,2-13C2]pyruvate and fol-
lowing the reaction by 13C NMR spectroscopy showed that the
Analysis by in situ CD spectroscopy9 indicated that the
(S)-enantiomer of 3 had been formed, as indicated by the
appearance of a positive CD band at 300 nm. No CD signal was
observed for 1 or 2 indicating that both products were virtually
racemic (see ESI†). This was confirmed by chiral phase LC-MS.
The observation that the decarboxylation of 3 produces
specifically 2 initially suggested that a rearrangement–decarb-
oxylation, as has been proposed for the decarboxylation of
(R)-acetolactate by the acetolactate decarboxylase from Klebsiella
aerogenes,10,11 might also be occurring enzymatically in our
case. Thus, two possible mechanisms are as shown in
Scheme 3, in which decarboxylation, with or without carboxyl
group migration, gives the enediol intermediate, which could
be regioselectively protonated by the enzyme to give only 2.
However, the racemic nature of 2 would be highly unusual for
an enzymatic product. Furthermore, separation of the enzyme
by ultrafiltration (for details, see ESI†) after the formation of
the intermediate 3 still led to formation of only 2 with no trace
of 1. Therefore, the observed decarboxylation is probably non-
enzymatic. However, some impact of the protein on this trans-
formation cannot be excluded as the nonenzymatic reaction
Scheme 2 MenD-catalyzed C–C bond forming reactions with 13C-labeled sub-
strates. In (c) [1,2-13C2]α-KG was made by in situ enzymatic oxidation of L-gluta-
mate. The labeled C atoms are marked with circles and asterisks.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 252–256 | 253

View Article Online
Communication
Organic & Biomolecular Chemistry
Scheme 4 YerE-catalyzed formation of acetohydroxy acids 4 and 5, and sub-
sequent decarboxylation to acyloins 6–8. The labeled C atoms are marked with
an asterisk.
Scheme 3 Possible pathways for the decarboxylation of 3 via (a) direct decarb-
oxylation or (b) tertiary ketol rearrangement, followed by decarboxylation.
never proceeded to completion, whereas it did in the presence
of MenD.
similar results regarding regioselectivity to those observed
with MenD. Additionally, SucA catalyzed the formation of
acetoin using either pyruvate or acetaldehyde as the sole sub-
strate. Incubation of pyruvate with SucA led, via 4, to almost
racemic acetoin 6 with only 8% ee of the (R)-isomer but the ee
increased to 90% when acetoin was prepared using acet-
aldehyde as the substrate.
In contrast, NMR investigations with ThDP-dependent CDH
(cyclohexanedione hydrolase from Azoarcus sp.)15 showed no
formation of acetolactate as an intermediate, starting either
from [1,2-13C]pyruvate as the sole substrate or in combination
with acetaldehyde. Accordingly, highly enantio-enriched
(S)-acetoin [(S)-6] (up to 90% ee) was directly obtained using
pyruvate as the sole substrate. This can be explained by enzy-
matic decarboxylation of pyruvate to give acetaldehyde, which
then acts as the acyl acceptor substrate giving (S)-acetoin.
If the decarboxylation is nonenzymatic, this leaves the ques-
tion of why the protonation of the enediol intermediate is
regiospecific, forming only 2. The probable explanation is
related to the presence of the carboxylic acid functionality in 3,
which could be a general acid catalyst and would be expected
to protonate C-4 much more efficiently than C-5, due to the
more favorable ring size. In support of this, decarboxylation of
3, after the removal of enzyme, at alkaline pH (pH 13, to
ensure the carboxylic acid is fully deprotonated) was found to
produce a mixture of 1 and 2. Unfortunately, further experi-
ments to test these hypotheses using other donor substrates,
such as 2-oxoglutarate 5-monoethyl ester or oxaloacetate (a
nonphysiological donor known to be accepted by MenD in the
presence of 2-fluorobenzaldehyde5), failed since these donors
were not accepted as substrates by MenD with pyruvate as the
acceptor.
α-Keto acids are frequently accepted as acyl acceptor sub-
strates in reactions catalyzed by different ThDP-dependent
enzymes. Therefore, α-hydroxy-β-keto acids are often formed
and similar decarboxylation to that described above can be
expected. In order to test whether this is generally the case,
three further ThDP-dependent enzymes were investigated.
First [2-13C]pyruvate was incubated with 2-oxobutyrate in
the presence of ThDP-dependent YerE.12 This transformation
initially led to the formation of (S)-acetolactate (4) (from two
molecules of pyruvate) and (S)-acetohydroxybutyrate (5) (from
pyruvate as the donor and 2-oxobutyrate as the acceptor)
(Scheme 4).13 The NMR data showed that 2-oxobutyrate is the
acceptor preferred by the enzyme (ratio 5 : 4 = 84 : 16 after
17 h). When this reaction was followed for a longer time, it
was found that besides acetoin (6), acyloin regioisomers 7 and
8 were also formed in a 60 : 40 ratio as the final products
(Scheme 4). Acyloin 8 can only be derived by decarboxylation
of 5, since the possibility of 2-oxobutyrate being the acyl donor
was excluded by further 13C NMR experiments using [1,2-13C2]
pyruvate.
Conclusions
From the results with MenD, YerE and SucA, it can be seen
that the formation and subsequent decarboxylation of
α-hydroxy-β-keto acids is common when α-keto acids are
involved as substrates in ThDP-dependent enzymatic trans-
formations. As shown above, this often leads to the formation
of the mixed regioisomers of the corresponding α-hydroxy
ketones, but also might lead to the selective formation of
a sole product, as in the decarboxylation of 3. We postulate
that the decarboxylation of optically active acetolactate
is the reason for the formation of the nearly racemic
acetoin observed in many ThDP-dependent enzymatic trans-
formations.16
In conclusion, the change in regioselectivity induced by
varying the substrate(s) in ThDP-dependent enzymatic trans-
formations may not necessarily be due to swapping the acyl
donor–acceptor roles of the substrates. The observation of
unexpected products, such as 2, which initially appear to be
the result of acceptor–donor reversal, can lead to misinter-
pretation and incorrect characterization of an enzyme. Under-
standing the selective formation of 2 by MenD paved the way
SucA, the ThDP-dependent E1 subunit of the α-ketoglut-
arate dehydrogenase complex from E. coli K12,14 was studied
next. Incubation of α-KG and labeled pyruvate with SucA led to
254 | Org. Biomol. Chem., 2013, 11, 252–256
This journal is © The Royal Society of Chemistry 2013

View Article Online
Organic & Biomolecular Chemistry
Communication
for the discovery that CDH produces highly enantioenriched 13C NMR kinetic studies with in situ production of 13C-labeled
acetoin from pyruvate as the sole substrate, because it is incap- α-KG
able of producing acetolactate.
MenD (500 μg mL−1), L-GluDH (4 U mL−1), NADH oxidase (25
U mL−1), L-[1,2-13C]glutamic acid (30 mM), NAD+ (5 mM) and
DTT (5 mM) were incubated with 13C-labeled pyruvate (30 mM)
Experimental
General procedures
in the reaction buffer (750 μL) at 30 °C directly in an NMR
tube. The reactions were followed by in situ NMR experiments.
Nuclear magnetic resonance (NMR) spectra were recorded on a
Deuterium labeling experiments
DRX 400 (Bruker) operating at 400 and 100 MHz for 1H and 13
C
The general procedure for formation of 1 or 2 was followed.
Formation of 1 and 2 was confirmed by NMR and the reaction
mixtures were then lyophilized. The solid residues were dis-
solved in D2O (1.5 mL) and were incubated for a further
48 hours.
For 1: δD (CH3OH) 2.8 (br s); δC (CD3OD) 20.0 (CH3), 32.6
(CH2), 35.5 (br, CH2), 74.2 (CH), 181.2 (COOH), 216.1 (br,
CvO).
acquisitions, respectively. Coupling constants (J) are reported
in Hertz (Hz). Circular dichroism (CD) was measured using a
spectral polarimeter J-810 (Jasco International). High-perform-
ance liquid chromatography (HPLC-DAD) was performed on
an HP 1100 chromatography system (Agilent). HPLC-DAD with
MS/MS (LC-MS) was performed using API2000 with a Turbo-
lonSprayTM source, EI, MRM scan (Applied Biosystems). LC on a
chiral phase was performed on an Astec Chirobiotic T 5 μm
column, at 5 °C, eluted at 0.3 mL min−1 with methanol, 0.1%
acetic acid and 0.4% triethylamine. GC on a chiral phase was
performed on GC-2010 (FID) at 70 °C equipped with the injec-
tor AOC-20 (Shimadzu) using a Macherey-Nagel, FS-Lipodex D
column (50.0 m × 0.25 mm).
13C NMR studies of YerE-catalyzed carboligation of pyruvate
and 2-oxobutyrate
YerE (1 mg mL−1) and 2-oxobutyrate (50 mM) were incubated
with [1,2-13C] or [2-13C]pyruvate (50 mM) in the reaction buffer
(750 μL) at 25 °C directly in an NMR tube. 13C NMR spectra
were recorded with a time delay between spectra of 30 min.
Other general procedures are given in the ESI.†
General procedure for formation of 1 or 2
MenD or SucA (500 μg mL−1) and α-KG (30 mM) was incubated
with pyruvate or acetaldehyde (30 mM) in a reaction buffer
(1.5 mL; 50 mM phosphate, 2 mM MgCl2, 0.1 mM ThDP, 10%
(v/v) D2O, pH = 8) at 30 °C and 300 rpm using a thermomixer
(Eppendorf). After 48 h, the sample was used for NMR without
any further purification.
SucA-catalyzed acetoin formation
SucA (700 μg mL−1) and pyruvate or acetaldehyde (50 mM)
were incubated in the reaction buffer (1.5 mL) at 30 °C and
300 rpm using a thermomixer (Eppendorf). After 48 h, the
sample was used for NMR and GC measurements without any
further purification. Chiral GC: tR(R) = 17.6 min, tR(S) =
13.2 min.
For in situ 13C NMR experiments, 15 mM of each substrate
was incubated in 750 μL of the reaction buffer directly in an
NMR tube under the same conditions as above. For selected
kinetic experiments, the time delay between spectra was
60 min.
CDH-catalyzed acetoin formation
CDH (1 mg mL−1) and pyruvate (25 mM) were incubated in the
reaction buffer (1.5 mL; 50 mM MES, 3 mM MgSO4, 0.5 mM
ThDP, 10% (v/v) D2O, pH = 6.5) at 30 °C and 300 rpm using a
thermomixer (Eppendorf). After 24 and 48 h, the sample was
used for GC measurements without any further purification.
For in situ 13C-NMR, [1,2-13C2]pyruvate (25 mM) was incu-
bated in the reaction buffer (750 μL) directly in an NMR tube
under the same conditions as above. The time delay between
spectra was 30 min.
For in situ CD experiments,9 15 mM of each substrate was
incubated in the reaction buffer in a 0.1 cm path length cell at
30 °C. The spectra were first recorded with the buffer to make
sure that there is no background signal. The reactions were
initiated by addition of the enzyme and the spectra were
recorded in the range of 240–350 nm. The time delay between
spectra was 60 min.
5-Hydroxy-4-oxohexanoic acid 1, δH (D2O) 1.26 (3 H, d,
J 7.1), 2.27–2.32 (2 H, m), 2.65–2.70 (2 H, m) and 4.32 (1 H, q,
J 7.1); δC (D2O) 18.4 (CH3), 30.8 (CH2), 34.1 (CH2), 72.8 (CH),
181.3 (COOH) and 215.8 (CvO); Chiral LC-MS: tR(R) = 6.8 min,
tR(S) = 6.2 min; MS/MS: 145 (parent ion, M − H+), 127, 101,
and 83 (fragment ions).
Acknowledgements
This paper is dedicated to Prof. Dr Maria-Regina Kula on the
4-Hydroxy-5-oxohexanoic acid 2, δH (D2O) 1.67–1.76 (2 H, occasion of her 75th birthday. We are grateful to F. Bonina and
m), 1.96–2.05 (2 H, m), 2.14 (3 H, s), 4.22 (1 H, dd, J 8.3, 3.9); S. Waltzer for technical assistance. Carsten Lanzerath is
δC (D2O) 25.3 (CH3), 28.9 (CH2), 29.0 (CH2), 76.4 (CH), 182.0 acknowledged for his kind assistance in providing L-GluDH
(COOH), 214.8 (CvO); Chiral LC-MS: tR(R) = 16.2 min, tR(S) = and NADH-oxidase enzymes. We thank the DFG (FOR 1296)
17.3 min; MS/MS: 145 (parent ion, M − H+), 127, 101, and 83 and the DBU (ChemBioTech: AZ13234-32) for financial
(fragment ions).
support.
This journal is © The Royal Society of Chemistry 2013
Org. Biomol. Chem., 2013, 11, 252–256 | 255

View Article Online
Communication
Organic & Biomolecular Chemistry
11 S. Najmudin, J. T. Andersen, S. A. Patkar, T. V. Borchert,
D. H. G. Crout and V. Fülöp, Acta Crystallogr., Sect. D: Biol.
Crystallogr., 2003, 59, 1073.
12 P. Lehwald, M. Richter, C. Röhr, H. W. Liu and M. Müller,
Angew. Chem., Int. Ed., 2010, 49, 2389.
Notes and references
1 M. Müller, D. Gocke and M. Pohl, FEBS J., 2009, 276, 2894.
2 R. Kluger and K. Tittmann, Chem. Rev., 2008, 108, 1797.
3 J. Popp, J. Bacteriol., 1989, 171, 4349.
4 A. Dawson, P. K. Fyfe and W. N. Hunter, J. Mol. Biol., 2008, 13 D. Chipman, Z. Barak and J. V. Schloss, Biochim.
384, 1353.
Biophys. Acta, Protein Struct. Mol. Enzymol., 1998, 1385,
401.
14 M. G. Darlison, M. E. Spencer and J. R. Guest,
Eur. J. Biochem., 1984, 141, 351.
15 (a) S. Fraas, A. K. Steinbach, A. Tabbert, J. Harder,
U. Ermler, K. Tittmann, A. Meyer and P. M. H. Kroneck,
J. Mol. Catal. B: Enzym., 2009, 61, 47; (b) A. K. Steinbach,
S. Fraas, J. Harder, A. Tabbert, H. Brinkmann, A. Meyer,
U. Ermler and P. M. H. Kroneck, J. Bacteriol., 2011, 193,
6760.
16 S. Bornemann, D. H. G. Crout, H. Dalton,
D. W. Hutchinson, G. Dean, N. Thomson and
M. M. Turner, J. Chem. Soc., Perkin Trans. 1, 1993, 309.
5 A. Kurutsch, M. Richter, V. Brecht, G. A. Sprenger and
M. Müller, J. Mol. Catal. B: Enzym., 2009, 61, 56.
6 M. Fang, B. M. Langman and D. R. J. Palmer, Bioorg. Med.
Chem. Lett., 2010, 20, 5019.
7 B. Geueke, B. Riebel and W. Hummel, Enzyme Microb.
Technol., 2003, 32, 205.
8 P. Ödman, W. B. Wellborn and A. S. Bommarius, Tetra-
hedron: Asymmetry, 2004, 15, 2933.
9 A. Baykal, S. Chakraborty, A. Dodoo and F. Jordan, Bioorg.
Chem., 2006, 34, 380.
10 D. H. G. Crout, S. Davies, R. J. Heath, C. O. Miles,
D. L. Rathbone and B. E. P. Swoboda, Biocatalysis, 1994, 9, 1.
256 | Org. Biomol. Chem., 2013, 11, 252–256
This journal is © The Royal Society of Chemistry 2013