The first mimetic of the transketolase reaction

Article abstract of DOI:10.1002/ejoc.200500986

Although the biocatalytic formation of acyclic α,α′- dihydroxy ketones by transketolase is well documented in the literature, there is currently no one-pot chemical synthesis of these dihydroxy ketones available. Here, we report preliminary results of an

Full text of DOI:10.1002/ejoc.200500986

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200500986
The First Mimetic of the Transketolase Reaction
Mark E. B. Smith,[a] Kirsty Smithies,[a] Tarik Senussi,[b] Paul A. Dalby,[b] and
Helen C. Hailes*[a]
Keywords: Water chemistry / Aldehydes / C–C coupling / Transketolase / Enolates
Although the biocatalytic formation of acyclic α,αЈ-dihydroxy
ketones by transketolase is well documented in the litera-
ture, there is currently no one-pot chemical synthesis of these
dihydroxy ketones available. Here, we report preliminary re-
sults of an atom-efficient one-pot synthesis of α,αЈ-dihydroxy
ketones in water by a mimic of the transketolase reaction.
The formation of a quaternary ammonium enolate is pos-
tulated in this tertiary-amine-mediated carbon–carbon bond-
forming reaction.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Transketolase (TK) (EC 2.2.1.1) is an important and ver- five-step synthetic approach to aromatic and aliphatic dihy-
satile enzyme that has been used in stereospecific carbon– droxy ketones starting from 1,4-dioxene and lithiation with
carbon bond formation.[1] In vivo TK catalyses the transfer tert-butyllithium,[4] a strategy including a ruthenium-cata-
of the C1–C2 hydroxy ketone unit from -xylulose-5-phos- lyzed oxidation of allenes,[5] a double hydroxylation of silyl
phate to -ribose-5-phosphate, in a key step of the pentose enol ethers with m-chloroperbenzoic acid,[6] the utilisation
phosphate pathway, and the enzyme requires magnesium() of 1-chloroalkyl p-tolyl sulfoxides as hydroxycarbonyl anion
ions as well as thiamine pyrophosphate (TPP) as cofac- equivalents,[7] the approach by Enders et al. to enantiomeric
tors.[1] This stereospecific transfer is reversible; however, α,αЈ-dihydroxy ketones using dihydroxyacetone and chiral
Srere et al. reported that β-hydroxypyruvic acid is a sub- auxiliaries,[8] and an organoiron-templated route to cyclic
strate for TK, donating the two-carbon nucleophilic moiety species.[9] The lack of concise, direct strategies, presumably
with the subsequent loss of carbon dioxide, rendering the stems from the difficulties associated with creating a stable
reaction irreversible.[2] The TK-catalysed condensation of hydroxycarbonyl anion equivalent: umpolung approaches
hydroxypyruvate (HPA, 1) with a variety of aldehydes has using dithianes or tert-butyl hydrazones can result in β-eli-
been performed successfully in vitro to generate (S)-2 on a mination of hydroxide or alkoxide upon generation of the
preparatively useful scale (Scheme 1).[1] Dihydroxy ketone hydroxycarbonyl anion.[10] Herein, we report preliminary
functionalities are present in several natural products and results of a remarkable new atom-efficient, one-pot synthe-
are also important synthons for further structural elabora- sis of racemic α,αЈ-dihydroxy ketones by a mimetic of the
tion to a range of compounds including keto sugars.[3]
TK reaction in water.
As part of an ongoing research programme to develop
TK-catalysed carbon–carbon bond-formation methodolo-
gies, a number of high-throughput methods for the screen-
ing of libraries of TK mutants are under development. Dur-
ing investigations when screening for the TK reaction be-
tween propanal (Scheme 1, R = C2H5) and lithium hy-
droxypyruvate (Li-1, LiHPA), at pH = 7.0, a range of dif-
ferent buffers were evaluated. Interestingly, the formation
of 1,3-dihydroxypentan-2-one (2a, R = C2H5) was observed
in some control experiments and was therefore investigated
further. When carrying out experiments using 3-(4-morpho-
lino)propanesulfonic acid (MOPS, 3) buffer with TPP and
Mg2+, but no TK, 2a was generated. However, under com-
parable conditions no reactions were observed when either
tris(hydroxymethyl)aminomethane (TRIS, 4) or glycylgly-
cine (Gly-Gly, 5) were used (Figure 1). Most surprisingly,
the same results were observed using buffers 3–5 when both
TPP and Mg2+ were omitted from the reactions, ruling out
Scheme 1. Transketolase-catalysed reaction using HPA (1) to gener-
ate dihydroxy ketones (S)-2.
Although the biocatalytic formation of dihydroxy
ketones by TK is well documented in the literature, there
is currently no one-pot non-enzymatic synthesis available.
Reported syntheses of α,αЈ-dihydroxy ketones include: a
[a] Department of Chemistry, University College London, Christo-
pher Ingold Laboratories,
20 Gordon Street, London, WC1H 0AJ, UK
Fax: +44-20-7679-7463
E-mail: h.c.hailes@ucl.ac.uk.
[b] Department of Biochemical Engineering, University College
London,
Torrington Place, London WC1E 7JE, UK
Eur. J. Org. Chem. 2006, 1121–1123
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1121

M. E. B. Smith, K. Smithies, T. Senussi, P. A. Dalby, H. C. Hailes
SHORT COMMUNICATION
the mechanistic involvement of TPP and formation of 2a buffer was added. The use of N,NЈ-dimethylpiperazine, N-
via the thiazolium ylide.[11] However, these results indicated methylmorpholine and N-methylpiperidine led to the for-
that MOPS was promoting the carbon–carbon bond-form- mation of 2a (Table 1, Entries 5–7). Morpholine, piperidine,
ing reaction.
triethylamine and sodium 1-decanesulfonate (Table 1, En-
tries 8–11) generated no 2a. This demonstrated that the re-
action required the addition of a cyclic tertiary amine.
When morpholine was used, the secondary amine reacted
with the aldehyde alone generating the enamine. The forma-
tion of side-products resulting from an aldol addition of 2a
to propanal was also observed using cyclic tertiary amines
other than MOPS (Table 1, Entries 5–7). This implied that
buffering the reaction at near neutral conditions by the sul-
fonate group reduces the basicity of the reaction medium
and stops unwanted side reactions. However, upon reducing
the equivalents of aldehyde used from three to one with
N-methylmorpholine, 2a was cleanly formed with no side-
products in 35% yield, and the reaction proceeded more
rapidly than in the MOPS-buffered solution (Table 1, En-
try 12). The lower yield of 2a when using N-methylpiperi-
dine indicated that a second heteroatom at the 4-position
in the ring may be important mechanistically.
From the acceptors used in the transformation, the reac-
tion proceeds with an aliphatic, aromatic and hydroxylated
aliphatic aldehyde. However, changing the donor molecule
had a significant impact on product formation. Although
both lithium and potassium hydroxypyruvate reacted to
give the dihydroxy ketone product in comparable yields, py-
ruvic acid (free acid and sodium salt) failed to give any
reaction under the above conditions. The reaction mimics
the biological transketolase reaction with respect to the
substrates and product, however does not involve the use
of TPP and therefore mechanistically operates by a different
pathway. Based upon all observations, a mechanism for the
reaction has been proposed (Scheme 2) involving tertiary
amine catalysis and quaternary amine enolate formation
which has previously been described, for example, in cyclo-
propanation reactions.[12]
Figure 1. Buffers used in initial screening experiments.
MOPS (3) was then used with alternative aldehydes in
preparative reactions at pH = 7. Using propanal and Li-1
with 3, 2a was isolated in 21–35% yield depending on the
reaction time (Table 1, Entries 1, 2). The use of glycolalde-
hyde and the aromatic acceptor benzaldehyde gave the α,αЈ-
dihydroxy ketones 2b and 2c in 63% and 25% yields, respec-
tively (Table 1, Entries 3, 4). No side products resulting
from the addition of a second aldehyde unit were observed
using any of the aldehydes in the presence of MOPS after
72 h. The highest yields were observed using 1 equiv. of
MOPS.
Table 1. Use of different aldehydes and sulfonate or amine.[a]
Entry
Sulfonate or amine
t [h]
Yield (%)
Side products
2[b]
1
2
3
MOPS (3)
MOPS (3)
MOPS (3)
5
72
72
72
5
5
5
5
5
2a (21)
2a (35)
2b (63)
2c (25)
2a (26)
2a (27)
2a (8)
0
0
0
0
0
8
9
7
0
0
0
0
0
4
MOPS (3)
5
6
7
8
N,NЈ-dimethylpiperazine
N-methylmorpholine
N-methylpiperidine
morpholine
9
piperidine
triethylamine
0
0
0
10
11
12
5
5
5
sodium 1-decanesulfonate
N-methylmorpholine
2a (35)[c]
[a] Aldehyde (3 equiv.), Li-1 (1 equiv.), amine/sufonate (1 equiv.) in
water at 50 m, room temp. [b] Yield of isolated product after puri-
fication by chromatography. [c] Aldehyde (1 equiv.), Li-1 (1 equiv.),
amine (1 equiv.) in water at 50 m, room temp.
The structure of MOPS suggested that a heterocyclic or
sulfonic acid moiety may be required to promote the reac-
tion. Accordingly, a range of secondary and tertiary amines
and sulfonic acids were screened for propanal conversion
with Li-1 to 1,3-dihydroxypentan-2-one (2a), in an attempt
to elucidate the functionality required for dihydroxy ketone
formation. The reactions were carried out in water, and to
Scheme 2. Proposed mechanism of the formation of 2 using MOPS.
In aqueous alkaline media the enolisation and tauto-
avoid complications due to the addition of a buffer, no merisation of 1 has been studied.[13] Thus, the conjugate
1122
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 1121–1123

The First Mimetic of the Transketolase Reaction
SHORT COMMUNICATION
addition of the tertiary amine to the enol 1a of 1, or tauto-
merisation to tartronate semialdehyde (1b) and subsequent
addition of amine to the aldehyde moiety will generate the
enolate 6. From observations to date, the improvement in
yield in the presence of a second heteroatom suggests that
some stabilisation of the nitrogen enolate 6 or other inter-
mediates through hydrogen bonding may be important. Ad-
dition of the aldehyde, proton transfer, then elimination of
the tertiary amine and decarboxylation will, after conver-
sion of the enol to the ketone tautomer, give 2 and should
release the tertiary amine at the end of the reaction. This
postulated mechanism, reminiscent of the Baylis–Hillman
reaction, is further substantiated by the lack of observed
reactivity with pyruvic acid and sodium pyruvate, which
will undergo tautomerisation less readily. Detailed studies
into the mechanistic role of the tertiary amine and catalytic
function in the reaction cycle have yet to be fully estab-
lished.
In summary, a one-pot tertiary-amine-mediated carbon–
carbon bond-forming reaction in water has been observed
that is chemoselective for hydroxypyruvate as a donor
group. Interestingly, the observed donor substrate tolerance
to date is analogous to that observed with transketolase.
We are currently exploring the scope of this reaction since
in principle it should be possible to use the tertiary amine
in catalytic quantities, and it has significant synthetic poten-
tial for the construction of keto sugars and keto sugar ana-
logues. This new methodology to racemic substrates is com-
plementary to current approaches using TK to generate
α,αЈ-dihydroxy ketones enantioselectively, and further
mechanistic studies will be performed.
Acknowledgments
The authors would like to thank the UK Engineering and Physical
Sciences Research Council (EPSRC) for support of the multidisci-
plinary Biocatalysis Integrated with Chemistry and Engineering
(BiCE) programme (GR/S62505/01) (M. E. B. S. and H. K. S.) and
the BBSRC for a studentship (T. S.). Financial support from the 12
industrial partners supporting the BiCE programme is also ac-
knowledged. We also thank Dr. Michael J. Porter for helpful dis-
cussions.
[1] For examples of TK reactions using a range of substrates, see:
a) E. Racker, in: The Enzymes, vol. 5 (Eds.: P. D. Boyer, H.
Lardy, K. Myrzback), Academic Press, New York, 1961, pp.
397–412; b) J. Bolte, C. Demuynck, H. Samaki, Tetrahedron
Lett. 1987, 28, 5525–5528; c) C. Demuynck, J. Bolte, L. Hec-
quet, V. Dalmas, Tetrahedron Lett. 1991, 32, 5085–5088; d) F.
Effenberger, V. Null, T. Ziegler, Tetrahedron Lett. 1992, 33,
5157–5160; e) G. R. Hobbs, M. D. Lilly, N. J. Turner, J. M.
Ward, A. J. Willets, J. M. Woodley, J. Chem. Soc., Perkin Trans.
1 1993, 165–166; f) K. G. Morris, M. E. B. Smith, N. J. Turner,
M. D. Lilly, R. K. Mitra, J. M. Woodley, Tetrahedron: Asym-
metry 1996, 7, 2185–2188; g) R. K. Mitra, J. M. Woodley,
M. D. Lilly, Enzyme Microb. Technol. 1998, 22, 64–70; h) F. T.
Zimmerman, A. Schneider, U. Schörken, G. A. Sprenger, W.-
D. Fessner, Tetrahedron: Asymmetry 1999, 10, 1643–1646.
[2] P. Srere, J. R. Cooper, M. Tabachnick, E. Racker, Arch. Bi-
ochim. Biophys. 1958, 74, 295–305.
[3] a) D. R. Andrews, R. A. Giusto, A. R. Sudhakar, Tetrahedron
Lett. 1996, 37, 3417–3420; b) D. C. Myles, P. J. Andrulis III,
G. M. Whitesides, Tetrahedron Lett. 1991, 32, 4835–4838; c) Y.
Kobori, D. C. Myles, G. M. Whitesides, J. Org. Chem. 1992,
57, 5899–5907.
[4] M. Fetizon, P. Goulaouic, I. Hanna, Tetrahedron Lett. 1985,
26, 4925–4928.
[5] M. Laux, N. Krause, Synlett 1997, 765–766.
[6] Y. Horiguchi, E. Nakamura, I. Kuwajima, Tetrahedron Lett.
1989, 30, 3323–3326.
[7] T. Satoh, K. Onda, K. Yamakawa, J. Org. Chem. 1991, 56,
Experimental Section
4129–4134.
[8] D. Enders, M. Voith, A. Lenzen, Angew. Chem. Int. Ed. 2005,
44, 1304–1325.
[9] A. J. Pearson, K. Chang, J. Org. Chem. 1993, 58, 1228–1237.
[10] M. E. B. Smith, H. C. Hailes, unpublished results.
[11] R. Heck, A. P. Henderson, B. Kohler, J. Retey, B. T. Golding,
Eur. J. Org. Chem. 2001, 2623–2627.
[12] a) C. D. Papageorgiou, S. V. Ley, M. J. Gaunt, Angew. Chem.
Int. Ed. 2003, 42, 828–831; b) C. D. Papageorgiou, M. A. Cub-
illo de Dios, S. V. Ley, M. J. Gaunt, Angew. Chem. Int. Ed.
2004, 43, 4641–4644.
[13] D. Fleury, M. B. Fleury, N. Platzer, Tetrahedron 1981, 37, 493–
501.
Representative Experimental Procedure: Li-1 (3 mmol), propanal
(9 mmol) and the tertiary amine (3 mmol) were stirred in water
(60 mL) at room temp. for the time specified. The water was re-
moved in vacuo and the dry compound loaded onto a flash silica
chromatography column [EtOAc/petroleum ether (boiling range
40–60 °C), 4:1] to give the products indicated (2a–2c) which were
characterised by 1H and 13C spectroscopy, high-resolution mass
spectrometry and IR spectroscopy.[1e,4] The identity of the insepa-
rable mixture of isomeric side products (leading to superimposed
NMR spectra) was confirmed by MS. ESMS: m/z (%) = 194 (88)
[M + NH4]+, 159 (100) [M – H2O]+. HRMS: m/z calcd. for
C8H17O4 [MH]+ 177.11268, found 177.11291.
Received: December 19, 2005
Published Online: January 24, 2006
Eur. J. Org. Chem. 2006, 1121–1123
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1123