Chemoenzymatic synthesis of l-tyrosine derivative for a transketolase assay

Article abstract of DOI:10.1016/j.tetlet.2008.03.099

We have prepared an l-tyrosine derivative bearing a d-threo ketose moiety by a convenient chemoenzymatic route. This compound is of potential interest for developing stereospecific assays for enzymes catalyzing C-C bond cleavage such as transketolase. We showed in vitro by analytical studies (LC/MS and ³¹P NMR) that this compound can release l-tyrosine in the presence of wild type TK extract and bovine serum albumin. This assay is the first step towards a mutant TK selection test that could be developed for yeast cells auxotrophic for l-tyrosine.

Full text of DOI:10.1016/j.tetlet.2008.03.099

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 3229–3233
Chemoenzymatic synthesis of L-tyrosine
derivative for a transketolase assay
a
a
b
a
´
`
`
Franck Charmantray , Virgil Helaine , Angelika Lasikova , Bertrand Legeret ,
Laurence Hecquet a,*
a
`
`
`
´ ˆ
`
Universite´ Blaise Pascal, Laboratoire de Synthese et Etude de Systemes a Interet Biologique, UMR CNRS 6504, 63177 Aubiere Cedex, France
b Slovak University of Technology, Department of Organic Chemistry, Radlinskeo 9, 81237 Bratislava, Slovak Republic
Received 21 December 2007; revised 5 March 2008; accepted 19 March 2008
Available online 23 March 2008
Abstract
We have prepared an L-tyrosine derivative bearing a D-threo ketose moiety by a convenient chemoenzymatic route. This compound is
of potential interest for developing stereospecific assays for enzymes catalyzing C–C bond cleavage such as transketolase. We showed
in vitro by analytical studies (LC/MS and 31P NMR) that this compound can release L-tyrosine in the presence of wild type TK extract
and bovine serum albumin. This assay is the first step towards a mutant TK selection test that could be developed for yeast cells
auxotrophic for L-tyrosine.
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Biocatalysis; Tyrosine; Transketolase; Assay
Transketolase is a useful catalyst for ketose synthesis
through the stereocontrolled formation of the C2–C3 bond
(Scheme 1). TK isolated from spinach leaves,1 baker’s
yeast,2 Escherichia coli3 and Saccharomyces cerevisiae
recombinant TK4 has been used for synthetic purposes.5
TK reaction is reversible except when the donor substrate
is b-hydroxypyruvic acid. This property is very convenient
for synthetic purposes because the decarboxylation of this
donor substrate makes the overall condensation reaction
irreversible. TK catalyses the transfer of a ketol unit from
b-hydroxypyruvic acid to an aldehyde to give a D-threo
(3S,4R) ketose. TK is very specific for ketol as donors
and for hydroxyaldehydes in the (R) configuration as
acceptors. In the course of generating yeast TK with new
or improved substrate specificities, an efficient screening
or selection system is an absolute prerequisite for identify-
ing evolved enzyme variants that display improved proper-
ties. Only screening tests have been developed for this type
of enzyme. Various assays for TK activity determination
have been proposed using detection by spectrophotometry
with NADH-dependent enzymes.6 More recently, a colori-
metric assay was reported in the presence of tetrazolium
Scheme 1.
*
Corresponding author. Tel.: +33 4 73 407871; fax: +33 4 73 407717.
E-mail address: Laurence.HECQUET@univ-bpclermont.fr (L. Hecquet).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.03.099

3230
F. Charmantray et al. / Tetrahedron Letters 49 (2008) 3229–3233
In our case, the appropriate acceptor aldehyde for the
aldol condensation catalyzed by RAMA in the presence
of DHAP was compound 2. To our knowledge, this was
the first time a C-2 amino acid substituted aldehyde was
a substrate of RAMA. The aldehyde was made by ozono-
lysis of the double bond of compound 3. Carboxylic acid
was protected as ethyl ester and amine by usual protecting
groups: Cbz, Boc, acetyl or trifluoroacetyl.
Compounds 4 were either commercially available (4a) or
synthesized from L-tyrosine ethylester (4b, 91%;11 4c,
83%;12 4d, 91%13). Compounds 4a–d were first allylated
in refluxing acetone to give the corresponding intermedi-
ates (3a, 91%;14 3b, 62%; 3c, 91%; 3d, 90%). Subsequent
ozonolysis of the crude products in methylene chloride at
ꢀ78 °C followed by the reduction of the ozonides with
dimethyl sulfide gave the aldehydes (2a, 74%; 2b, 62%;
2c, 32%; 2d, 22%).
Scheme 2.
red.7 To detect modifications of the stereospecificity of
mutant TK, we have developed a fluorogenic screening test
for TK based on the detection of coumarine from stereo-
chemical probes.8 However, these in vitro assays have a
major disadvantage. For screening large libraries of mutant
enzymes, it is necessary to disrupt each clone separately to
determine the catalytic properties of the mutant enzyme,
unlike for selection tests. As a rule, these latter efficiently
combine the enzymatic catalytic activity with a survival
factor, a growth advantage for the micro-organisms. In this
context, we target fused molecules containing a sugar
moiety with varied stereochemistries for mutant TK
recognition and an aminoacid as a possible leaving group
for cell supply.
Aldehydes 2c and 2d in the presence of RAMA did not
react with DHAP, as shown by enzymatic assay.15 This
might be due to the marked hydrophobic character, along
with steric hindrance, of the Cbz and Boc amino protecting
groups.
First, we sought to validate the principle of the assay
in vitro with wild-type TK extract and compound 1 bearing
a D-threo ketose moiety and L-tyrosine as the leaving group
(Scheme 2).
Adding co-solvents such as 10% DMSO, DMF or meth-
anol did not improve the results.
By contrast, aldehydes 2a and 2b underwent efficient
aldol addition with DHAP owing to lower steric hindrance
and better solubility with 10% methanol.
Here we report on the synthesis of compound 1a by a
straightforward chemoenzymatic route, avoiding protec-
tion and deprotection steps, as previously described for
obtaining a D-threo ketose analogue bearing umbelliferone
instead of tyrosine moiety, by a chemical route.9 The assay
was studied in vitro with TK extract from yeast. We used
LC/MS and 31P NMR experiments to monitor both the
first step catalyzed by wild-type TK reaction from 1 and
the second step in which the protected L-tyrosine was
released on BSA catalyzed b-elimination.
We developed a chemoenzymatic strategy in which both
chiral centres (3S,4R) of compound 1 sugar moiety were
introduced together by C–C bond formation in a highly
stereoselective manner, catalyzed by fructose-1,6-bisphos-
phate aldolase from rabbit muscle, RAMA (E.C.
4.1.2.13), in the presence of dihydroxyacetonephosphate
(DHAP) as donor substrate and the suitable aldehyde 2
as acceptor (Scheme 3). This step was followed by dephos-
phorylation of the aldol product catalyzed by acid phos-
phatase (E.C. 3.1.3.2) at pH 4.7. This strategy is most
often used to produce (3S,4R) ketoses such as 13C labelled
sugars, heteroatom-substituted sugars, deoxy sugars,
fluoro sugars, long-chain sugars and cyclitols. Well over
100 aldehydes have been described as acceptor substrates.10
Crude compounds 1a and 1b were obtained after
dephosphorylation catalyzed by acid phosphatase (E.C.
3.1.3.2). NMR analysis of the crude enzymatic mixtures
showed that compound 1a was the main product (89%
according to 1H NMR in the presence of TMSP-d4 as
calibrate). Compound 1b was a minor product owing to
the ester and (or) trifluoroacetyl protecting group hydro-
lysis. So we focused our strategy on compound 1a. A pure
analytical amount of 1a, necessary for in vitro studies, was
obtained after purification by reverse-phase semi-prepara-
tive LC. This is the first time an amino acid branched sugar
has been generated using RAMA as a catalyst, thus extend-
ing its scope for syntheses.
The enzymatic assay was based on TK-catalyzed C2–C3
bond cleavage from donor substrates 1a and subsequent
transfer of the hydroxyacetyl group released to D-ribose-
5-phosphate, the natural acceptor substrate of TK (Scheme
4). This reaction led to the formation of D-sedoheptulose-7-
phosphate along with the a-hydroxyaldehyde 1a0. Accord-
ing to the literature16 and our own results with coumarine
in place of L-tyrosine, the b-elimination reaction from 1a0
Scheme 3. Reagents and conditions: (i) K2CO3, allylbromide; (ii) O3; (iii) DMS; (iv) DHAP, RAMA, pH 7.5; (v) acid phosphatase pH 4.7.

F. Charmantray et al. / Tetrahedron Letters 49 (2008) 3229–3233
3231
Scheme 4.
analogues takes place satisfactorily using bovine serum
albumin (BSA). Under these conditions, the intermediate
1a0 led to the release of protected L-tyrosine 4a. We carried
out 31P NMR and LC/MS experiments to monitor these
two reactions in comparison with the control, that is,
intrinsic chemical stability of compound 1a under the same
conditions. Our aim was to show qualitatively that TK
extract and BSA specifically catalyzed the release of 1a0
and 4a respectively.
31P NMR was suitable for following the conversion of
D-ribose-5-phosphate into D-sedoheptulose-7-phosphate in
the first step of the assay. Reaction progress was followed
by recording 31P NMR spectra decoupled from 1H and 13
C
Fig. 1. Release of compound 1a0 from 1a catalyzed by TK. Controls: 1a
(100 lM), TK extract (1 unit mLꢀ1); M: 1a in Bicine buffer pH 8.2; +: 1a
in Mops buffer pH 7.2. Reactions: 1a (100 lM), D-ribose-5-phosphate
(100 lM), TK extract (1 unit mLꢀ1); }: 1a, N: 1a0 in Bicine buffer 0.1 M
pH 8.2; ꢀ: 1a, j: 1a0 in Mops buffer 0.1M pH 7.2.
and using H3PO4 as internal standard for the calibration.
We showed the stability of both phosphorylated monosac-
charides with TK at pH 7.2 in Mops buffer and without the
donor substrate 1a.
In the presence of 1a, the conversion of D-ribose-5-phos-
phate to D-sedoheptulose-7-phosphate occurred, as shown
by the disappearance of the signal of D-ribose-5-phosphate
at 3.5 ppm and the appearance of the signal of D-sedohep-
tulose-7-phosphate at 4.35 ppm at the same time.
from compound 1a, confirming the results obtained previ-
ously using 31P NMR.
Second, we set out to show that BSA catalyzed the b-
elimination from 1a to protected L-tyrosine 4a (retention
time: 14.5 min; HRMS m/z: [M+Na+]: calculated for
C13H17NO4Na 274.1055, found 274.1049). First of all, we
checked the stability of compound 1a in the presence of
BSA. Figure 2 shows that compound 1a slightly decom-
posed to 4a, the rate being twice as high in Bicine buffer
as in Mops buffer. This background noise could be due
to BSA, as already mentioned on an analogue fluorogenic
compound,8a or to a non-enzymatic hydroxyacetyl group
transfer as recently described by Hailes and co-workers.19
When TK and D-ribose-5-phosphate were added, N-
acetyl-L-tyrosine ethyl ester 4a was released at a rate 10
times higher in Mops and 5 times higher in Bicine than
The 31P NMR shift of D-sedoheptulose-7-phosphate
matched that obtained from an authentic sample prepared
according to a procedure described in the literature.17 This
study demonstrated that TK catalyzed the hydroxyacetyl
group transfer from the donor substrate 1a to D-ribose-5-
phosphate, the natural acceptor substrate of TK.
The experiments were carried out at pH 7.2 (Mops buf-
fer) and pH 8.2 (Bicine buffer) to compare the influence of
the pH on the reaction rates. These pH values were close to
the common values used for the TK reaction and for b-
elimination catalyzed by BSA.18
First, we studied the step catalyzed by TK. We showed
that compound 1a was stable at both pH 7.2 and 8.2, in the
presence of TK extract (Fig. 1). When the acceptor sub-
strate, D-ribose-5-phosphate was added to the reaction
mixture containing compound 1a and TK extract, we
followed the disappearance of 1a (retention time: 11.3 min;
HRMS m/z: [M+Na+]: calculated for C18H25NO8Na
406.1478, found 406.1472) and at the same time the appear-
ance of compound 1a0. This compound was characterized
by its exact mass (retention time: 8.5 min; C18 XTerraÒ
2.1 mm ꢁ 100 mm, Waters. Reaction products were succes-
sively eluted along the elution gradient water/acetonitrile/
formic acid 94.9/5/0.1, v/v/v to water/acetonitrile/formic
acid 5/94.9/0.1 v/v/v. HRMS m/z: [M+Na+]: calculated
for C16H21NO6Na: 346.1267, found: 346.1262). We noted
that the TK-catalyzed reaction was not pH-dependent.
Hence we showed that TK catalyzed C2–C3 bond cleavage
Fig. 2. Release of 4a from 1a catalyzed by TK and BSA. Controls: 1a
(100 lM), BSA (2 mg mLꢀ1); M: 1a in Bicine buffer 0.1 M pH 8.2, ꢀ: 1a in
Mops buffer 0.1 M pH 7.2. Reactions: 1a (100 lM), BSA (2 mg mLꢀ1),
D-ribose-5-phosphate (100 lM), TK extract (1 unit mLꢀ1); N: 4a in Bicine
buffer 0.1 M pH 8.2, j: 4a in Mops buffer 0.1 M pH 7.2.

3232
F. Charmantray et al. / Tetrahedron Letters 49 (2008) 3229–3233
previously. The absolute rates of compound 4a release were
similar in any buffer. Consistent with the difference in the
stability of compound 1a in each buffer, the relative rate
of compound 4a release in Mops buffer was twice as high
as in Bicine buffer. Hence, Mops buffer (pH 7.2) is more
suitable than Bicine buffer (pH 8.2).
LC/MS monitoring enabled us to confirm the efficiency
of the TK-catalyzed first step by identifying a-hydroxyalde-
hyde key intermediate 1a0, and the efficiency of the BSA-
catalyzed second step by identifying protected L-tyrosine
4a released.
tant was evaporated to dryness under vacuum. [a] 5.78 (c
1.2, MeOH) Yield: 89% according the 1H NMR. 1H
NMR (CD3)2CO d (ppm): 1.10 (t, 3H, J = 7 Hz, CH3
ester); 1.82 (s, 3H, CH3 acetyl); 2.78 (dd, 1H, J = 9,
14 Hz, CH part of CH2–Ar); 2.94 (dd, 1H, J = 6, 14 Hz,
CH part of CH2–Ar); 3.89 (dd, 1H, J = 6, 9 Hz, CH part
of CH2–O); 3.99 (dd, 1H, J = 6, 9 Hz, CH part of CH2–
O); 4.02 (q, 2H, J = 7 Hz, CH2 ester); 4.17–4.21 (m, 1H,
CH–OH); 4.28 (d, 1H, J = 2 Hz, CH–OH); 4.39 (d, 1H,
J = 20 Hz, CH part of CH2–OH); 4.43 (d, 1H, J = 20 Hz,
CH part of CH2–OH); 4.50 (dd, 1H, J = 6, 9 Hz, CH);
6.76 (d, 2H, J = 9 Hz, CHAr); 7.01 (d, 2H, J = 9 Hz,
CHAr); 13C NMR (CD3)2CO d (ppm): 14.5 (CH3); 22.3
(CH3); 37.7 (CH2); 55.6 (CH); 62.3 (CH2); 68.0 (CH2);
71.8 (CH2); 72.5 (CH); 77.0 (CH); 115.6 (2CHAr); 130.5
(CAr); 131.3 (2CHAr); 159.1 (CAr); 169.9 (C@O ester);
173.2 (C@O amide); 212.4 (C@O ketone); HRMS m/z:
[M+Na+]: calculated for C18H25NO8Na 406.1478, found
406.1472. A sample of compound 1a was further purified
by HPLC using C18 XTerraÒ column (Waters),
7.6 mm ꢁ 100 mm Isocratic solvent system water/acetoni-
trile; 89/11, v/v, was used to elute 1a at 4.2 mL minꢀ1. In
those conditions 1a was recovered in 4 mg scale as pure
compound according to LC/MS analysis performed on
an analytical XTerra C18 column; 2.1 mm ꢁ 100 mm
(Waters).21
As the ultimate aim could be to develop this TK selec-
tion assay in vivo in yeast cells, the deprotection of 4a by
an enzymatic route can be investigated. Usually, the
enzymes commonly used for hydrolysing an ester and an
acetyl group are proteases and acylases, respectively, as
described for other similarly protected amino acids.11,20
Using commercially available hydrolases, subtilisin and
acylase I, compound 4a was fully converted sequentially
into L-acetyltyrosine and then tyrosine.
In conclusion, we synthesized compound 1a in four steps
from protected L-tyrosine 4a. From compound 1a we per-
formed an enzymatic assay leading, in two steps, to the
release of protected L-tyrosine. We showed by analytical
studies that in the first step, compound 1a was a donor sub-
strate for TK in the presence of D-ribose-5-phosphate as
acceptor substrate. To our knowledge, only compounds
such as the natural substrate D-xylulose-5-phosphate, b-
hydroxypyruvic acid or L-erythrulose, have been consid-
ered as possible donor substrates for TK. Both our previ-
ous8 and current work show that the substrate specificity
of TK for the donor substrate is broader than expected.
We demonstrated the possibility of detecting wild type
TK activity in vitro from compound 1a based on the
release of L-tyrosine. In the future, and to make this test
usable in vivo, a prerequisite will be the limitation of the
background signal for the direct cleavage of 1a into 4a.
For cells both auxotrophic for L-tyrosine and expressing
TK, it should be possible to carry out this assay in vivo.
This strategy could offer the first stereospecific selection
test for TK mutants.
Acknowledgements
We thank MENRT for its financial support (postdoc-
toral fellowship for Dr. A.L.). We thank Professor G.
Schneider (Karolinska Institut, Division of Molecular
Biology, S-17177 Stockholm, Sweden) for the gift of the
yeast strain H402 transformed with pTKL1.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.tetlet.
2008.03.099.
References and notes
Preparation of N-acetyl-O0-(2R,3S,5-trihydroxy-4-oxo-
pentyl)-L-tyrosine ethyl ester (1a): N-Acetyl-O0-(2-oxo-
ethyl)-L-tyrosine ethyl ester 2a (1.1 g, 3.75 mmol) was
dissolved in water/methanol, 1/1, v/v. After dissolution,
7 ml of water were added dropwise. Dihydroxyacetone
phosphate (360 mM, 3.75 mmol, 1 equiv, pH 7) was added
thereafter (to give a 200 mM final substrate concentration)
followed by 385 units of RAMA. The mixture was stirred
for 24 hours at room temperature; the reaction was fol-
lowed by 13C NMR in H2O/D2O, 9/1, v/v; until complete
disappearance of the starting aldehyde. The pH was
adjusted to 4.8 and 370 units of acid phosphatase were
added. The mixture was then stirred overnight. 3 volumes
of methanol were added. The precipitate was discarded
by centrifugation at 8000 rpm and the subsequent superna-
1. Bolte, J.; Demuynck, C.; Hecquet, L.; Samaki, H. Carbohydr. Res.
1990, 206, 79–85.
2. (a) Ziegler, T.; Straub, A.; Effenberger, F. Angew. Chem., Int. Ed.
Engl. 1988, 27, 716–721; (b) Kobori, Y.; Myles, D. C.; Withesides, G.
M. J. Org. Chem. 1991, 22, 5899–5907.
3. (a) Morris, K. G.; Smith, M. E. B.; Turner, N. J. Tetrahedron:
Asymmetry 1996, 7, 2185–2188; (b) Zimmermann, F. T.; Schneider,
A.; Scho¨rken, Y.; Sprenger, G. A.; Fessner, W. D. Tetrahedron:
Asymmetry 1999, 10, 1643–1646.
4. (a) Wikner, C.; Meshalkina, L.; Nilsson, U.; Nikkola, M.; Lindqvist,
Y.; Schneider, G. J. Biol. Chem. 1994, 269, 32144–32150; (b) Nilsson,
´
U.; Hecquet, L.; Gefflaut, T.; Guerard, C.; Schneider, G. FEBS Lett.
1998, 424, 49–52.
´
´
5. (a) Andre, C.; Guerard, C.; Hecquet, L.; Demuynck, C.; Bolte, J. J.
´
Mol. Catal. B: Enzym. 1998, 5, 459–466; (b) Guerard, C.; Alphand,
V.; Archelas, A.; Demuynck, C.; Hecquet, L.; Furstoss, R.; Bolte, J.

F. Charmantray et al. / Tetrahedron Letters 49 (2008) 3229–3233
3233
´
Eur. J. Org. Chem. 1999, 1, 3399–3402; (c) Crestia, D.; Guerard, C.;
13. Delle Monache, G.; Di Giovanni, M. C.; Maggio, F.; Misiti, D.;
Zappia, G. Synthesis 1995, 9, 1155–1158.
14. Tous, G.; Bush, A.; Tous, A.; Jordan, F. J. Med. Chem. 1990, 33,
1620–1634.
15. Bergmeyer, H. U. Methods in Enzymatic Analysis, 34th ed.; Chemie,
1984; pp 146–147.
16. (a) Klein, G.; Reymond, J. L. Helv. Chim. Acta 1999, 82, 400–407; (b)
Badalassi, F.; Wahler, D.; Klein, G.; Crotti, P.; Reymond, J. L.
Angew. Chem., Int. Ed. 2000, 39, 4067–4070; (c) Perez-Carlon, R.;
Jourdain, N.; Reymond, J. L. Chem. Eur. J. 2000, 6, 4154–4162.
Veschambre, H.; Hecquet, L.; Demuynck, C.; Bolte, J. Tetrahedron:
Asymmetry 2001, 12, 869–876.
6. (a) Horecker, B. L.; Smyrniotis, P. Z. Methods Enzymol. 1955, 1, 371–
375; (b) Hecquet, L.; Bolte, J.; Demuynck, C. Biosc. Biotech. Biochem.
1993, 12, 2174–2176; (c) Sevostyanova, I. A.; Solovjeva, O. N.;
Kochetov, G. A. Biochemistry (Mosc). 2006, 71, 560–562.
7. Smith, M. E. B.; Kaulmann, U.; Ward, J. M.; Hailes, H. C. Bioorg.
Med. Chem. 2006, 14, 7062–7065.
´
8. (a) Sevestre, A.; Helaine, V.; Guyot, G.; Martin, C.; Hecquet, L.
Tetrahedron Lett. 2003, 44, 827–830; (b) Sevestre, A.; Charmantray,
17. Lee, S.; Kirschning, A.; Muller, M.; Way, C.; Floss, H. G. J. Mol
¨
´
F.; Helaine, V.; Lasikova, A.; Hecquet, L. Tetrahedron 2006, 62,
Catal. B 1999, 6, 369–377.
3969–3976.
18. Klein, G.; Reymond, J. L. Bioorg. Med. Chem. Lett. 1998, 8, 1113–
1116.
´
9. Gonzalez-Garcia, E.; Helaine, V.; Klein, G.; Schuermann, M.;
Sprenger, G. A.; Fessner, W. D.; Reymond, J. L. Chem. Eur. J.
2003, 9, 893–899.
19. Smith, M. E. B.; Smithies, K.; Senussi, T.; Dalby, P.; Hailes, H. C.
Eur. J. Org. Chem. 2006, 5, 1121–1123.
10. Machajewski, T. D.; Wong, C.-H. Angew. Chem., Int. Ed. 2000, 39,
1352–1374.
11. Bois-Choussy, M.; Zhu, J. J. Org. Chem. 1998, 63, 5662–5665.
12. Hioki, K.; Kinugasa, M.; Kishimoto, M.; Fujiwara, M.; Tani, S.;
Kunishima, M. Synthesis 2006, 12, 1931–1933.
20. Easton, C. J.; Harper, J. B. Tetrahedron Lett. 1998, 39, 5269–5272.
21. Linear gradient elution was used from water/acetonitrile/formic
acid, 94.9/5/0.1, v/v/v to water/acetonitrile/formic acid 5/94.9/0.1,
v/v/v.