Enhancing and reversing the stereoselectivity of Escherichia coli transketolase via single-point mutations

Article abstract of DOI:10.1002/adsc.200800489

Chiral auxiliary methodology and chiral assays have been developed to establish the enantiomeric purities of erythrulose and 1,3-dihydroxypentan-2-one generated using wild-type (WT) Escherichia coli transketolase (TK). L-Erythrulose was formed in 95% ee and (3S)-1,3-dihydroxypentan-2-one in 58% ee. Since the latter compound was formed in moderate ee, TK libraries were screened to identify higher performing mutants. A colorimetric screen and chiral assay were successfully applied to a 96-well format, and new active TK mutants were identified, which gave 1,3-dihydroxypentan-2-one in high stereoselectivities. Remarkably, activesite single-point mutants were identified that were able to both enhance and reverse the stereoselectivity of TK.

Full text of DOI:10.1002/adsc.200800489

FULL PAPERS
DOI: 10.1002/adsc.200800489
Enhancing and Reversing the Stereoselectivity of Escherichia coli
Transketolase via Single-Point Mutations
a
b
a
b
Mark E. B. Smith, Edward G. Hibbert, Alexander B. Jones, Paul A. Dalby,
a,
and Helen C. Hailes *
Department of Chemistry, University College London, Christopher Ingold Laboratories, 20 Gordon Street, London
WC1H 0AJ, U.K.
a
Fax : (+44)-20-7679-7463; e-mail: h.c.hailes@ucl.ac.uk
Department of Biochemical Engineering, University College London, Torrington Place, London WC1E 7JE, U.K.
b
Received: August 5, 2008; Published online: November 5, 2008
Abstract: Chiral auxiliary methodology and chiral to a 96-well format, and new active TK mutants
assays have been developed to establish the enantio- were identified, which gave 1,3-dihydroxypentan-2-
meric purities of erythrulose and 1,3-dihydroxypen- one in high stereoselectivities. Remarkably, active-
tan-2-one generated using wild-type (WT) Escheri- site single-point mutants were identified that were
chia coli transketolase (TK). l-Erythrulose was able to both enhance and reverse the stereoselectiv-
formed in 95% ee and (3S)-1,3-dihydroxypentan-2- ity of TK.
one in 58% ee. Since the latter compound was
formed in moderate ee, TK libraries were screened
to identify higher performing mutants. A colorimet- Keywords: biotransformations; a,a’-dihydroxy ke-
ric screen and chiral assay were successfully applied tones; enzymes; molecular evolution; transketolase
[
2,4]
Introduction
tive enzyme for industrial applications.
Notably,
TK has been used for the continuous production of l-
Transketolase (TK) (EC 2.2.1.1) is an important erythrulose in a membrane reactor and process
enzyme that has been used in stereospecific carbon- screening for the synthesis of d-xylulose 5-phosphate
[
4]
carbon bond formation and requires magnesium(II) by TK has been described. TK has been used with a
ions as well as thiamine diphosphate (ThDP) as cofac- range of a-hydroxyaldehyde acceptors and HPA (free
[
1–6]
tors.
In vivo TK catalyses the transfer of a two- acid or Li salt 1) where it is stereospecific for the a-
[5,6]
carbon ketol unit from d-xylulose 5-phosphate to d- (R)-hydroxyaldehyde.
Furthermore, TK can toler-
Although this transfer is rever- ate a range of non-a-hydroxylated aldehydes, al-
sible, the use of the donor b-hydroxypyruvate (HPA though lower relative rates of reaction (typically 5–
[
1,2]
ribose 5-phosphate.
[
3]
1
) renders the reaction irreversible. Interestingly, 35% of hydroxylated substrates) are normally ob-
E.coli TK more readily accepts 1 than spinach or served. For large-scale processes TK variants are re-
[6]
yeast TK giving up to a 30-fold higher specific activi- quired that have high specific activities and stereose-
[7]
ty. The ability of TK to achieve enantioselective lectivities towards substrates. We are particularly in-
[8]
carbon-carbon bond formation, and tolerance for a terested in using TK in synthetic applications with
wide range of aldehyde acceptors to give (S)-a,a’-di- non-a-hydroxylated aldehydes to expand the use of
hydroxy ketones 2 (Scheme 1), has made it an attrac- this enzyme for the preparation of chiral synthons.
Directed evolution strategies have proven to be ex-
tremely successful in generating enzymes with im-
proved properties, and the use of sequence and crystal
[9]
structure information can be extremely valuable. In
previous work we have used saturation mutagenesis
targeted to the TK active site to identify enhanced
TK activity towards glycolaldehyde (R=CH OH)
2
[10,11]
and the non-a-hydroxylated aldehyde propanal.
Scheme 1. TK-catalysed reaction to generate a,a’-dihydroxy Although mutants with improved activities were de-
ketones.
scribed, stereoselectivities were not established due to
Adv. Synth. Catal. 2008, 350, 2631 – 2638
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2631

Mark E. B. Smith et al.
FULL PAPERS
a lack of chiral assays. The enantioselectivity of the
TK product with chiral a-hydroxy aldehyde acceptors
has previously been monitored by comparison to
known standards and high enantioselectivities gener-
[5]
ally observed. With glycolaldehyde and non-a-hy-
droxylated substrates, although optical rotations have
generally been reported, enantiomeric purities are
[6]
not. For example, when glycolaldehyde was reacted
with 1 and spinach TK to give 3, the authors stated
that the ꢁstereospecificity seems very highꢂ (optical ro-
2
D
5
25
D
[12]
tation data [a] : +7; lit. [a] : +11.3 ), although the
enantiomeric purity of the products was not fully as-
[5a]
sessed. Other reports describing the formation of 4
using TK are communications with brief experimental
data including optical rotation data, however optical
[
6b]
purities were not determined. One exception is the
use of spinach TK and 1 with ethanal, methoxyetha-
nal and methylthioaldehydes: ees were reported in
Scheme 2. Synthesis of 5, (3S)-1,3-dihydroxypentan-2-one
[
6a,13]
the range of 60–76% for the (S)-isomer.
Our ap-
(
4) and (3S)-8. Reagents and conditions: (i) 0.5 mol% Er-
proach was therefore to initially establish suitable
assays to determine the enantiomeric purities of prod-
ucts 3 and 4, when using glycolaldehyde and propanal t-BuLi, EtI; (vi) 6N HCl; (vii) Ac O, pyr, DMAP.
AHCTUNGTREG(NNUN OTf) , Ac O, CH CN, room temperature; (ii) PSTA, DMF,
3
2
3
2
,2-DMP; (iii) NaIO , KHPO ; (iv) SAMP, PhH, reflux; (v)
4 4
2
with WT E.coli TK, and then to screen the TK libra-
ries to identify more stereoselective TK variants under mild reaction conditions has been described. In
where suboptimal optical purities were observed.
particular erbium AHCTUNGTERNNUNG( III) triflate has been highlighted as
an extremely active acylation catalyst. When race-
[
17]
mic 3 was reacted with acetic anhydride in the pres-
Results and Discussion
ence of Er ACHTUNEGRTNNUNG( OTf) (0.5 mol%), the triacetylated prod-
3
uct 5 was cleanly formed after 17 h in 30% yield
(Scheme 2a). Notably, when the reaction was repeat-
ed with a commercially available sample of l-erythru-
Chiral Assays
Despite the many reports describing the synthesis or lose (3), the product 5 had an ee of ꢀ95% by chiral
enzymatic formation of l-erythrulose, and its com- HPLC. In addition, subjecting the sample of l-eryth-
mercial applications, the predominant method for as- rulose to an extended acylation protocol (48 h) gave
[
5a,14,15]
sessing enantiomeric purity is optical rotation.
no deterioration in the observed ee due to racemisa-
(OTf) -cata-
One problem is that when compounds have low opti- tion at the a-position, highlighting that ErAHCUTNGTRENNUNG
3
cal rotation values there is a greater risk of introduc- lysed acylation offers an excellent method for the de-
ing errors due to the limits involved in accurately rivatisation of acid-sensitive chiral alcohols such as er-
measuring the data. For this reason, and for applica- ythrulose (3).
tions in automated screening protocols or when gener-
The synthesis of chemical standards and a chiral
ating small quantities of 3, a method using chiral assay were then established to determine the enantio-
HPLC or GC was sought. For chiral HPLC and GC 3 meric purity and absolute stereochemistry of 4 when
needed to be derivatised and conversion to the triace- using propanal with WT E.coli TK. Using Enderꢂs
tate was investigated. Derivatisation using either chiral auxiliary methodology to generate a,a’-dihy-
acetic
anhydride/pyridine/4-dimethylaminopyridine droxy ketones in high enantioselectivities, aminotriol
[18]
(
DMAP) or acetyl chloride/triethylamine led to the 6 was converted to the SAMP hydrazone (S)-7. Re-
formation of polymerised products. The HCl-cata- action with ethyl iodide and removal of the chiral
lysed esterification of racemic 3, synthesised using a auxiliary and acetal protecting group gave (3S)-4
[
18]
[17]
novel mimetic of the TK reaction recently reported (Scheme 2b). Racemic 4 and (3S)-4 were readily
[
16]
from this laboratory,
with acetic anhydride gave converted to the diacetate 8 and a chiral GC method
racemic triacetate 5 and this material was used to suc- established for screening the enantiomeric purities.
cessfully develop a chiral HPLC assay using a Chiral-
cel AD column. However, using this method to triace-
tylate a commercially available sample of l-erythru- Stereoselectivities of WT E.coli TK
lose (3) led to complete racemisation at the a-posi-
tion. The use of Lewis acid catalysts, including lantha- Compounds 3 and 4 were prepared using WT E.coli
nide triflates, for the acylation of hydroxy groups TK and converted into 5 and 8 as described above
2632
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2631 – 2638

Enhancing and Reversing the Stereoselectivity of Escherichia coli Transketolase
FULL PAPERS
(
Scheme 2). The optical rotation data for 3 was very (E. coli TK) form a pocket and interact with the C-2
similar to that reported by Bolte et al for spinach TK hydroxy group of natural a-hydroxylated aldehyde
2
D
5
25
D
[5a]
[11]
(
[a] : +6.7 measured for 3; Bolte et al. [a] : +7;
substrates. The C-2 hydroxy group of an a-hydroxy-
25
[12]
lit. [a] : +11.3 ). HPLC analysis of 5, and compari- lated aldehyde substrate has been replaced by the
D
son to triacetylated l-erythrulose (l-erythrulose avail- methyl group in propanal, and therefore active mu-
able commercially) indicated that 3 had been formed tants identified from the H26, D469, and H100 struc-
in 95% ee (S major isomer). Similarly, conversion of 4 tural libraries may exert interesting stereochemical ef-
to 8 then analysis of 8 by chiral GC, and comparison fects on the TK reaction and were selected for investi-
to (3S)-8 synthesised as described above using SAMP gation.
chiral auxiliary methodologies, indicated
4
was
Reaction plates of the E. coli TK libraries obtained
formed in 58% ee (S major isomer). This stereoselec- by saturation mutagenesis of H26, D469, and H100
tivity was comparable to previous studies using spi- were incubated with propanal and 1 prior to screening
[
6a,13]
nach TK and similar acceptors.
WT-TK showed for product 4 with a tetrazolium-based colorimetric
[20]
good stereoselectivity towards glycolaldehyde but was assay and for product ee using the GC assay. Both
significantly lower for the non-a-hydroxylated alde- assays were successfully applied to a 96-well format
hyde propanal. To use TK for the preparation of allowing for screening in a high-throughput manner
chiral synthons for applications in synthesis high ste- (representative data for colorimetric assay in Figure 2
reoselectivities are required, and therefore the use of for D469).
engineered TKs were investigated with the acceptor
propanal.
The colorimetric assay and GC-based ee analysis of
the 96-well plates highlighted several interesting fea-
tures (Table 1). The D469 library displayed good ac-
tivities with no inactive mutants generated, and
single-point mutants that enhanced and reversed the
stereoselectivities were identified. It should be noted
that the GC assay was more sensitive (>1–2% bio-
Stereoselectivities with Active-Site Single-Point TK
Mutants
We have previously demonstrated that saturation mu- conversions detected) than the colorimetric assay (>
[
20]
tagenesis targeted to the enzyme active site can en- 8% bioconversions detected ), so mutants with low
hance TK activity towards glycolaldehyde and propa- activities not readily detected due to a faint coloration
[10,11]
nal.
Saturation mutagenesis was targeted to the colorimetric assay, were identified in the GC
active-site residues selected on either structural assay. Mutants showing good ees and conversions to
grounds (residues within 4 ꢃ of postulated binding of 4, from the intense coloration in the colorimetric
substrate) or phylogenetic variability within bacterial screen, were investigated further and sequenced as
and yeast TKs (also within 10 ꢃ of ThDP cofactor). It D469E [90% ee (3S)-4], D469T [64% ee (3S)-4] and
has been reported that D477 in yeast TK, the analo- D469Y [53% ee (3R)-4] (Table 2). Several other mu-
[
11]
gous residue to D469 in E. coli TK, hydrogen bonds tants showed good activities but poor ees were ob-
to the C-2 hydroxy group of a-hydroxylated aldehyde served against propanal, including D469S [12% ee
substrates and is involved in determining the stereose- (3S)-4], and D469A [33% ee (3S)-4].
lectivity of TK (Figure 1, erythrose 4-phosphate ad-
junct). Furthermore, residues D469, H26 and H100 to the D469 library, most mutants lead to the forma-
tion of (3R)-4. Two mutants gave 4 in an ee of >75%
Analysis of the H26 library showed that, in contrast
[19]
in favour of the (3R)-product but none gave the (3S)-
product in high ee. The most stereoselective mutant
also gave good conversions to 4, and was sequenced
as H26Y, which remarkably gave (3R)-4 in 88% ee.
Interestingly, activity of the H100 library was reduced
compared to WT. This may be because H100 (H103
in yeast) has a key binding interaction to HPA 1
which could be lost or modified unfavourably in the
[
21]
library mutants.
Although the (3S)-selectivity was
largely maintained across the library and 4 mutants
demonstrated (3R)-selectivity (10–20% ee), this set of
mutants was not explored further because lower activ-
ities and moderate ees were observed.
Reactions using the high performing mutants were
performed at 50 mM concentration and preparative
Figure 1. Model of donor (1)-substrate (erythrose 4-phos- quantities of 4 isolated to confirm the ee data. The
[11]
phate)-ThDP adjunct in the active site of TK.
specific activities of D469T and D469Y have been re-
Adv. Synth. Catal. 2008, 350, 2631 – 2638
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2633

Mark E. B. Smith et al.
FULL PAPERS
Figure 2. Colorimetric assay for D469 library to give 4.
[a]
Table 1. Stereoselectivity profile for D469 and H26 libraries.
Table 2. Data for high performing D469 and H26 mutants.
Library (3S)-4 se- (3R)-4 se- ee knock-
Inactive
Mutants
Mutant Rate TK Spec. activity Relative ee of
[a]
À1
lective
lective
outs
[mmol/ [mg] [mmol/minmg ] Activity
h]
4
D469
H26
H100
69
29
88
14
42
4
13
12
4
0
13
0
WT
1.49
0.49 0.05
0.26 0.24
0.48 0.22
0.22 0.39
0.25 0.18
1
5
4
8
4
58%
(3S)
64%
[b]
D469T 3.74
D469Y 6.43
[
a]
ee knockouts are for ee of <10%.
(
3S)
53%
3R)
90%
3S)
88%
3R)
(
[11]
ported
and those for the remaining mutants were D469E 5.16
(
determined using clarified lysates of determined TK
concentrations (Table 2). All gave better specific ac-
tivities for the conversion of propanal to 4 than WT:
D469E gave an 8-fold improvement and H26Y, 4-fold
H26Y 2.69
(
[a]
Reactions carried out using clarified lysate of determined
TK concentration with 50 mM HPA, 50 mM propanal,
(
Table 2).
The ability of active-site single-point mutations to
2.4 mM ThDP and 9 mM MgCl in 50 mM Tris at pH 7.0
2
have such a remarkable effect on the stereoselectivity
with a small non-a-hydroxylated acceptor was surpris-
ing. Although combination mutants for other enzymes
have successfully resulted in reversals in stereoselec-
on 300 mL scale.
[b]
[11]
From ref.
tivities, there are few examples reported with single- verse selectivity observed with Asp469Tyr is harder to
point mutations. These include directed evolution rationalise. Similarly, His26Tyr where a positively
[
22]
[23]
studies with hydantoinases,
aldolases,
epoxyge- charged residue is substituted for Tyr, leads to a high
cyclohexa- reversing variant. It is difficult to predict whether the
[24]
[25]
nases,
arylmalonate decarboxylase,
[26]
none monooxygenase,
and hydroxynitrile lyase to unexpectedly high ee is due to steric or electrostatic
form diastereoisomeric products. However, in most effects, or longer range conformational changes.
cases the single-point mutants were not highly stereo-
[
27]
selective and were used in further studies to generate
combination mutants. The enhanced ee with variant Conclusions
Asp469Glu is probably due to the presence of the
extra methylene unit resulting in tighter binding of In summary, assays have been developed to readily
propanal or reaction intermediates. However, the in- screen erythrulose (3) and 1,3-dihydroxypentan-2-one
2634
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2631 – 2638

Enhancing and Reversing the Stereoselectivity of Escherichia coli Transketolase
FULL PAPERS
(
4) for the enantiomeric purities achieved in TK bio- were dissolved in H
2
O (30 mL) and the pH adjusted to 7
with 0.1M NaOH. Following a 20 min enzyme/cofactor pre-
incubation, the 1/glycolaldehyde mixture was added to the
enzyme solution and the mixture was stirred at 258C for
conversions. Using WT-TK, l-erythrulose was formed
in 95% ee and the assay has wide applicability for de-
termining the enantiomeric purity of 3 in biocatalytic
or chemical conversions. The synthesis of (3S)-1,3-di-
hydroxypentan-2-one using chiral auxiliary methodol-
ogy confirmed that this is the major isomer when
using WT E.coli TK, although only a moderate ee was
2
4 h. During this time, the pH was maintained at 7.0 by ad-
dition of 1M HCl using a pH stat (Stat Titrino, Metrohm).
Silica was added and the reaction mixture concentrated to
dryness under vacuum then dry loaded onto a flash silica
column. Following silica column purification (EtOAc:
observed (58%). TK libraries were screened to identi- MeOH, 9:1), l-erythrulose was isolated as an colourless oil;
fy higher performing mutants. A colorimetric screen yield: 288 mg (60%). Analytical data were identical to those
[5a]
25
D
25
D
and chiral assay were successfully applied to a 96-well previously described. [a] : +6.7 (c 1.48, H
format, and new active TK mutants were identified,
O), lit. [a] :
2
[12]
+
11.3 (c 4.0, H O)
2
which gave 4 in high stereoselectivities. Remarkably,
active-site single-point mutants were identified that Erythrulose Triacetate (5)
were able to both enhance and reverse the stereose-
Erythrulose (180 mg, 1.5 mmol) was added to a stirred solu-
tion of Ac O (425 mL, 4.5 mmol) and Er(OTf) (4.6 mg, 0.5
lectivity of TK towards propanal. Modelling and crys-
tallographic studies are now underway to understand
the reasons for these observations.
AHCTUNGTRENNUNG
2
3
mol% vs. erythrulose) in acetonitrile (5 mL). After stirring
for 1 h at 258C, all erythrulose had dissolved. The mixture
was stirred at 258C for a further 16 h, and the mixture then
partitioned between EtOAc (50 mL) and NaHCO3 (satu-
rated) (50 mL). The organic phase was dried, concentrated,
and purified using flash silica chromatography (EtOAc:hex-
Experimental Section
anes, 1:1) to afford 5 as a colourless oil; yield: 109 mg
General Experimental Methods
1
(
30%). H NMR (300 MHz; CDCl ): d=2.06 (3H, s,
3
Unless otherwise noted, solvents and reagents were reagent
grade from commercial suppliers and used without further
purification. Dry THF was obtained using anhydrous alumi-
COCH ), 2.14 (3H, s, COCH ), 2.16 (3H, s, COCH ), 4.36
3
3
3
(1H, dd, J=12.2 and 5.0, CHCH OAc), 4.47 (1H, dd, J=
2
12.2 and 3.4, CHCH OAc), 4.80 (1H, d, J=17.6, CO-
2
[28]
na columns.
All moisture-sensitive reactions were per-
CH OAc), 4.87 (1H, d, J=17.6, COCH OAc), 5.38 (1H, dd,
2
2
1
3
formed under a nitrogen or argon atmosphere using oven-
dried glassware. Reactions were monitored by TLC on Kie-
selgel 60 F254 plates with detection by UV, potassium per-
manganate, and phosphomolybdic acid stains. Flash column
chromatography was carried out using silica gel (particle
J=5.0 and 3.4, CHCH OAc); C NMR (75 MHz; CDCl ):
2
3
d=20.3, 20.5, 20.6 (COCH ), 62.3 (CHCH OAc), 66.6 (CO-
3
2
CH OAc), 74.7 (CHCH OAc), 169.8, 170.0, 170.3 (C=O
2
2
ester), 198.4 (C=O ketone); MS (+HR-FAB): m/z=
+
269.06301 (M+Na ), C H O requires 269.06372. Erythru-
1
0
14
7
1
13
25
size 40–63 mm). H NMR and C NMR spectra were record-
ed in CDCl3 at the field indicated using Bruker
lose triacetate (5) obtained via E.coli WT-TK [a] : À25 (c
D
a
0.5, CHCl ) and was determined to have an ee of (S) 95%
3
AMX300 MHz and AMX400 MHz machine. Coupling con-
stants are measured in Hertz (Hz) and unless otherwise
specified, NMR spectra were recorded at 298 K. Mass spec-
tra were recorded on a Thermo Finnegan MAT 900XP spec-
trometer. Infrared spectra were recorded on a Shamadazu
FTIR-8700 infrared spectrophotometer. Optical rotations
were recorded on an Optical Activity Limited PolAAR2000
by chiral HPLC (see below for method).
HPLC Assay for 5
HPLC analysis was performed on a Varian Prostar instru-
ment equipped with a Chiracel AD chiral column (Daicel,
25 cmꢄ0.46 cm). HPLC conditions: injection volume, 10 mL;
mobile phase, i-PrOH:hexanes, 5:95; flow rate, 0.8 mL/min;
detection, 210 nm; retention times: (R)-erythrulose triace-
tate, 22.0 min, (S)-erythrulose triacetate, 25.4 min. Racemic
erythrulose standards were synthesised as previously de-
2
À1
polarimeter at 589 nm, quoted in deg cm g and concentra-
tion (c) in g/100 mL.
5-Amino-5-hydroxymethyl-2,2-dimethyl-1,3-dioxane and
2
,2-dimethyl-5-oxo-1,3-dioxane were prepared as previously
[29]
[16]
described.
(S)-(+)-1-(2,2-Dimethyl-1,3-dioxan-5-yliden-
scribed and acetylated as described above.
ACHTUNGTRENNUNGa mino)-2-methoxymethylpyrrolidine [(S)-7] was synthesised
[18]
as reported by Enders et al. b-Hydroxypyruvate was pre-
(
S)-(À)-(2,2-Dimethyl-4-ethyl-1,3-dioxan-5-
[30]
[6c]
pared as the lithium salt as previously described. Libraries
were constructed and reaction plates obtained as previously
ylidenamino)-2-methoxymethylpyrrolidine
[10,11]
The reaction was carried out under anhydrous conditions.
To (S)-7 (0.0446 g, 0.180 mmol) and THF (10 mL) in a
three-necked round-bottomed flask was added t-BuLi
reported.
l-Erythrulose using E.coli WT-TK
(
0.14 mL, 0.20 mmol; 15% in pentane) dropwise by syringe
ThDP (88 mg, 192 mmol) and MgCl .6H O (146 mg, 718
at À788C. After stirring for 2 h at this temperature, the mix-
ture was cooled to À1008C and bromoethane (0.020 g,
0.18 mmol) added slowly. After further stirring for 2 h at
À1008C, the mixture was warmed to room temperature over
several hours. The mixture was quenched with pH 7 buffer
2
2
mmol) were dissolved in H O (42 mL) and the pH adjusted
2
to 7 with 0.1M NaOH. To this stirred solution, at 258C, was
[10]
added WT-TK clarified lysate (8 mL, 39 mg of TK), and
the mixture stirred for 20 min. In a separate flask, 1
(
440 mg, 4 mmol) and glycolaldehyde (240 mg, 4 mmol)
solution (2 mL) and diluted with Et O (80 mL). The organic
2
Adv. Synth. Catal. 2008, 350, 2631 – 2638
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2635

Mark E. B. Smith et al.
FULL PAPERS
layer was washed with pH 7 buffer solution (10 mL) and sa-
rated NaHCO3 solution. The product was purified by
column chromatography (EtOAc:hexane, 2:3) to give the
title compound in a quantitative yield. The characterisation
turated NaCl solution (2ꢄ10 mL), dried (MgSO ) and con-
4
centrated under vacuum. The monoalkylated SAMP-hydra-
zone was purified by flash silica chromatography (penta-
2
5
data were identical to those above and [a] : À23.0 (c 1.2,
D
ne:Et O, 5:1) to afford the title compound as a colourless
H O).
2
2
1
oil; yield: 18 mg (35%). H NMR (400 MHz; CDCl ): d=
3
0
2
3
.97 (3H, t, J=7.2, CH CH ), 1.39 [6H, s, C ACHTUNGTRENNG(U CH ) ], 1.50–
2 3 3 2
Screening of Libraries for Activity and Product ee
.08 (6H, m, NCHCH CH , CH ), 2.40 (1H, m, NCHH),
2
2
2
Preparation of cell samples – plate format: The 96-well re-
action plates were thawed from À808C resulting in the
freeze-thaw lysis of the 100 mL cell cultures. A 12ꢄ cofactor
.34 (3H, s, OCH ), 2.98–3.45 (4H, m, CH OCH , NCH,
3
3
2
NCHH), 4.14 (1H, dd, J=15.6 and 2.0, OCHHC=N), 4.25
1H, m, OCHC=N), 4.50 (1H, d, J=15.6, OCHHC=N);
(
13
stock solution (28.8 mM ThDP, 108 mM MgCl , 50 mM Tris,
C NMR (100 MHz; CDCl ): d=9.5 (CH CH ), 22.8, 23.9
2
3
2
3
pH 7.0; 25 mL) was added to each well and the plate was
then incubated at 258C for 20 min. A 4ꢄ HPA stock solu-
tion (200 mM in 50 mM Tris, pH 7.0; 75 mL) and a 3ꢄ
propanal stock solution (150 mM in 50 mM Tris, pH 7.0;
100 mL) were then added to each well and the plate incubat-
ed at 258C for a further 2 d.
and 24.4 [C ACHTUNGTRENNUNG( CH ) ], 25.2, 26.8, 55.6, 59.2, 59.8, 66.7, 72.0,
3 2
7
2
5.7, 100.1 [C ACHTUNGTRENNUNG( CH ) ], 160.4 (C=N); MS (HR-EI): m/z=
3 2
+
71.20097 (M ), C H N O requires 271.20162.
14
27
2
3
[31]
(
3S)-1,3-Dihydroxypentan-2-one (3S)-4
(
S)-(À)-(2,2-Dimethyl-4-ethyl-1,3-dioxan-5-ylidenamino)-2-
Assaying reaction plates for production of 4: 50 mL of re-
action mixture were transferred from each well to a fresh
96-well plate containing 50 mL of 50 mM Tris, pH 7.0 and
20 mg of MP-Carbonate Scavenger Resin (Biotage), which
was added to each well using a resin loader (Radleys). After
3 h, Tris (100 mL, 50 mM, pH 7.0) was added to each well,
with mixing, and 50 mL from each well transferred to a fresh
96-well plate. A plate reader (Fluostar, BMG-labtech) fitted
with an autoinjector was then used to add 2,3,5-triphenylte-
trazolium chloride (20 mL, 0.2% solution in methanol) then
3M NaOH (10 mL). The plate was shaken for 10 s, left for
1 min and an absorbance reading taken of each well at
methoxymethylpyrrolidine (0.100 g, 0.37 mmol) was dis-
solved in pentane (2 mL) and treated with 6N HCl
(
tion followed by TLC analysis. After 20 min the mixture was
extracted with Et O (3ꢄ20 mL), washed with saturated
0.4 mL). The mixture was vigorously stirred and the reac-
2
NaCl solution (20 mL) and dried (MgSO ). The crude prod-
uct was purified by silica gel column chromatography
4
(
Et O:pentane, 1:2) to give (3S)-1,3-dihydroxypentan-2-
2
[6b] 1
one as a colourless oil; yield: 18 mg (40%) . The H, and
13
[6b]
C NMR data were identical to those previously reported
25
and [a] : À44.5 (c 3.1, H O).
D
2
4
85 nm.
Assaying reaction plates for ee determination of 4 – con-
1
,3-Diacetoxypentan-2-one (8): General Procedure
version of 4 into 8: 100 mL of reaction mixture were trans-
ferred from each reaction well to a corresponding set of
vials each containing EtOAc (300 mL). The vials were
shaken and allowed to partition prior to transferring a por-
tion of the organic phase (100 mL) into fresh vials. Pyridine
[containing DMAP (10 mg/mL), 20 mL] was added to each
vial and ee analysis performed. A racemic sample of 4 syn-
for Racemic Sample and Preparative Scale
Biotransformations
1,3-Dihydroxypentan-2-one (100 mg, 0.85 mmol) was dis-
solved in pyridine (3 mL) and cooled to 08C. DMAP (cat.)
was added followed by Ac O (1 mL) and the mixture
warmed to room temperature. After 3 h, the mixture was
evaporated to dryness and the residue re-dissolved in
EtOAc (50 mL). The organic phase was washed with 0.3M
KHSO (50 mL), saturated NaHCO solution (50 mL), dried
2
[16]
thesised as previously described and converted into 8 as
described above.
GC assay for ee determination of 8: GC analysis was per-
formed on a Perkin–Elmer Autosystem XL Gas Chromato-
graph equipped with a b-Dex 225 chiral column (Supelco,
30 mꢄ0.25 mm). GC conditions: injection volume, 1 mL; car-
rier gas, He; carrier gas pressure, 15 psi; injector tempera-
ture, 2508C; oven temperature, 608C then increased at 38C/
min; detector temperature, 3008C; detection, flame ionised
detector (FID). Retention times: (3R)-8, 29.9 min; (3S)-8,
30.3 min.
4
3
(
MgSO ) and the solvent removed under vacuum. The prod-
4
uct was purified by silica column chromatography (EtOA-
c:hexane, 2:3) to yield the titled compound as an oil in
À1
quantitative yield. IR (film): nmax =2977, 2940, 1733 cm ;
1
H NMR (300 MHz; CDCl ): d=0.93 (3H, t, J=7.4,
3
CH CH ), 1.91–1.64 (2H, m, CH CH ), 2.09 (3H, s,
COCH ), 2.10 (3H, s, COCH ), 4.72 (1H, d, J=17.2,
CHHOAc), 4.77 (1H, d, J=17.2, CHHOAc), 5.01 (1H, dd,
J=7.5 and 5.0, CHOAc); C NMR (75 MHz; CDCl ): d=
2
3
2
3
3
3
1
3
3
9
.2 (CH CH ), 20.3, 20.5, 24.0, 66.1 (CH OAc), 77.5
2
3
2
Obtaining Preparatively Useful Quantities of
Selected Mutant Biocatalysts
(
CHOAc), 170.0 (C=O ester), 170.4 (C=O ester), 200.7 (C=
+
O ketone); MS (HR-FAB): m/z=225.07483 (M+Na ),
C H NaO requires 225.07389.
Streak plates onto Luria-Bertani (LB) agar containing
9
14
5
1
50 mg/mL ampicillin were made for potentially interesting
mutants from the relevant library master plate glycerol
stock. Individual colonies were selected and grown in LB
broth (10 mL) containing 150 mg/mL ampicillin for 16 h at
378C, 200 rpm. The 10 mL innoculum was then transferred
into further LB broth (200 mL) containing 150 mg/mL ampi-
cillin and incubated for 16 h at 378C, 200 rpm. An innocu-
lum optical density (600 nm) of 5 is typical after this time.
(
3S)-1,3-Diacetoxypentan-2-one, (3S)-8, via SAMP
(
3S)-1,3-Dihydroxypentan-2-one (18 mg, 0.15 mmol) was
dissolved in pyridine (0.1 mL) and Ac O (0.015 g) and
DMAP (0.020 g) were added. The mixture was stirred for
1
2
8 h at room temperature then the solution was concentrat-
ed under vacuum and partitioned between EtOAc and satu-
2636
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2631 – 2638

Enhancing and Reversing the Stereoselectivity of Escherichia coli Transketolase
FULL PAPERS
The innoculum was centrifuged at 4000 rpm for 20 min at Acknowledgements
108C. The supernatant was discarded and the cell pellet re-
suspended in sodium phosphate (10 mL of buffer per 1 g
wet cell paste, 5 mM, pH 7.0). The suspension was sonicated
The authors would like to thank the UK Engineering and
Physical Sciences Research Council (EPSRC) for support of
the multidisciplinary Bioconversion Integrated with Chemis-
try and Engineering (BiCE) programme (GR/S62505/01)
(M.E.B.S. and E.G.H.) and the EPSRC for a DTA student-
ship for A.B.J. Financial support from the 13 industrial part-
ners supporting the BiCE programme is also acknowledged.
(
Sanyo, Soniprep 150) on ice and the crude lysate centri-
fuged for 20 min at 4000 rpm at 108C. The clarified lysate
was pipetted into Eppendorf flasks and stored at À208C.
Kinetic Characterisation of Selected Mutants
Lysate TK protein concentration: The total protein concen-
tration in lysate samples was measured using a standard
Bradford assay using bovine serum albumin solutions of
known concentrations as standards. The percentage of total
protein present comprising TK was determined by densi-
tometry measurements following the electrophoresis of
lysate samples on a 7.5% SDS-PAGE gel using commercial-
ly available TK (Sigma) as a band standard. The Coomassie-
stained gel was imaged using Imagesoft softmare to calcu-
late the transketolase band density relative to the total band
density.
References
[
1] E. Racker, in: The Enzymes, Vol. 5, (Eds.: P. D. Boyer,
H. Lardy, K. Myrzback), Academic Press, New York,
1
961, pp 397–606.
[
[
2] N. J. Turner, Curr. Opin. Biotechnol. 2000, 11, 527–531.
3] P. Srere, J. R. Cooper, M. Tabachnick, E. Racker, Arch.
Biochim. Biophys. 1958, 74, 295–305.
Initial Rate Data:
A 12ꢄ cofactor stock solution
[
[
4] a) J. Bongs, D. Hahn, U. Schorken, G. A. Sprenger, C.
Wandrey, Biotechnol.Lett. 1997, 19, 213–215; b) J.
Shaeri, R. Wohlgemuth, J. M. Woodley, Org. Proc. Res.
Dev. 2006, 10, 605–610.
(
28.8 mM ThDP, 108 mM MgCl , 50 mM Tris, pH 7.0; 25 mL)
2
was added to lysate (100 mL) and the mixture incubated at
2
5
58C for 20 mins. A 4ꢄ stock solution of 1 (200 mM in
0 mM Tris, pH 7.0; 75 mL) and a 3ꢄ propanal stock solu-
5] a) J. Bolte, C. Demuynck, H. Samaki, Tetrahedron Lett.
tion (150 mM in 50 mM Tris, pH 7.0; 100 mL) were then
added and the reaction sampled and quenched after 1 h and
1
987, 28, 5525–5528; b) D. C. Myles, P. J. Andrulis,
G. M. Whitesides, Tetrahedron Lett. 1991, 32, 4835–
838; c) F. Effenberger, V. Null, T. Ziegler, Tetrahedron
2
0
h by addition of 50 mL of reaction mixture to 50 mL of
.1% TFA in water. The concentration of 4 present in the
4
Lett. 1992, 33, 5157–5160; d) L. Hecquet; J. Bolte, C.
Demuynck, C. Tetrahedron 1994, 50, 8677–8684; e) Y.
Kobori, D. C. Myles, G. M. Whitesides, J. Org. Chem.
samples was determined by HPLC.
1
992, 57, 5899–5907; f) R. K. Mitra, J. M. Woodley,
HPLC Assay for Determining Concentration of 4
M. D. Lilly, Enzyme Microb. Technol. 1998, 22, 64–70.
6] a) C. Demuynck, J. Bolte, L. Hecquet, V. Dalmas, Tet-
rahedron Lett. 1991, 32, 5085–5088; b) 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;
c) 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.
HPLC analysis was performed on a Dionex equipped with
an Aminex HPX-87H ion-exclusion column (250 mmꢄ
[
[
4
0
.6 mm, Bio-Rad Laboratories) maintained at 608C, using
.1% (v/v) trifluoroacetic acid in H O as mobile phase at a
2
flow rate of 0.6 mL/min. Retention time: 4, 15.6 min.
TK-Mediated Synthesis of 4
7] G. A. Sprenger, M. Pohl, J. Mol. Catal. B: Enzyme
ThDP (88 mg, 192 mmol) and MgCl .6H O (146 mg, 718
1999, 6, 145–159.
2
2
mmol) were dissolved in H O (42 mL) and the pH adjusted
[8] C. U. Ingram, M. Bommer, M. E. B. Smith, P. A. Dalby,
J. M. Ward, H. C. Hailes, G. J. Lye, Biotech. Bioeng.
2007, 96, 559–569.
[9] E. G. Hibbert, F. Baganz, H. C. Hailes, J. M. Ward, G. J.
Lye, J. M. Woodley, P. A. Dalby, Biomol. Eng. 2005, 22,
11–19.
2
to 7 with 0.1M NaOH. To this stirred solution, at 258C, was
added TK clarified lysate (8 mL, 0.7–2.3 U) and the mixture
stirred for 20 min. In another flask, 1 (440 mg, 4 mmol) and
propanal (288 mL, 4 mmol) were dissolved in H O (30 mL)
2
and the pH adjusted to 7 with 0.1M NaOH. Following the
2
0 min enzyme/cofactor pre-incubation, the mixture of 1
[10] E. G. Hibbert, T. Senussi, S. J. Costelloe, W. Lei,
M. E. B. Smith, J. M. Ward, H. C. Hailes, P. A. Dalby, J.
Biotech. 2007, 131, 425–432.
[11] E. G. Hibbert, T. O. Senussi, M. E. B. Smith, S. J. Cost-
elloe, J. M. Ward, H. C. Hailes, P. A. Dalby, J. Biotech.
2008, 134, 240–245.
and propanal was added to the enzyme solution and the
mixture stirred at 258C for 24 h. During this time, the pH
was maintained at 7.0 by addition of 1M HCl using a pH
stat (Stat Titrino, Metrohm). Silica was added and the reac-
tion mixture concentrated to dryness before being dry
loaded onto a flash silica gel column. Following column pu-
rification (EtOAc/hexane, 1:1), 4 was isolated as an oil
which became a white solid on standing; yield: 354 mg (70%
for D469E).
[12] R. L. Whistler, L. A. Underkofler, J. Am. Chem. Soc.
1938, 60, 2507–2508.
[13] V. Dalmas, C. Demuynck, Tetrahedron: Asymmetry
1993, 4, 2383–2388.
Wild-type TK afforded (3S)-4 in 58% ee (36% yield).
Mutant D469E afforded (3S)-4 in 90% ee Mutant H26Y af-
forded (3R)-4 in 88% ee (63% yield).
[14] D. Moonmangmee, O. Adachi, E. Shinagawa, H.
Toyama, G. Theeragool, N. Lotong, K. Matsushita,
Biosci. Biotechnol. Biochem. 2002, 66, 307–318.
Adv. Synth. Catal. 2008, 350, 2631 – 2638
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
2637

Mark E. B. Smith et al.
FULL PAPERS
[
15] K. Nishimura, T. Harada, Y. Arita, H. Watanabe, H.
Iwabuki, A. Terada, T. Naganuma, Y. Uzuka, J. Biosci.
Bioeng. 2006, 101, 303–308.
[23] G. J. Williams, S. Domann, A. Nelson, A. Berry, Proc.
Natl. Acad. Sci. USA 2003, 100, 3143–3148.
[24] K. L. Tee, U. Schwaneberg, Angew. Chem. Int. Ed.
2006, 45, 5380–5383.
[
16] M. E. B. Smith, K. Smithies, T. Senussi, P. A. Dalby,
H. C. Hailes, Eur. J. Org. Chem. 2006, 1121–1123.
17] A. Precopio, R. Dalpozzo, A. De Nino, L. Maiuolo, B.
Russo, G. Sindona, Adv. Synth. Catal. 2004, 346, 1465–
[25] Y. Terao, Y. Ijima, K. Miyamoto, H. Ohta, J. Mol. Cat.
[
B: Enz. 2007, 45, 15–20.
[26] M. T. Reetz, B. Brunner, T. Schneider, F. Shultz, C. M.
Clouthier, M. M. Kayser, Angew. Chem. Int. Ed. 2004,
43, 4075–4078.
[27] H. Bꢅhler, B. Miehlich, F. Effenberger, ChemBioChem.
2005, 6, 711–717.
[28] A. B. Pangborn, M. A. Giargello, R. H. Grubbs, R. K.
Rosen, F. J. Timmers, Organometallics 1996, 15, 1518–
1520.
[29] D. C. Forbes, D. G. Ene, M. P. Doyle, Synthesis 1998,
879–882.
1
470.
18] D. Enders, M. Voith, S. J. Ince, Synthesis 2002, 1775–
779.
[
[
[
[
1
19] U. Nilsson, L. Meshalkina, Y. Lindqvist, G. Schneider,
J. Biol. Chem. 1997, 272, 1864–1869.
20] M. E. B. Smith, U. Kaulmann, J. M. Ward, H. C. Hailes,
Bioorg. Med. Chem. 2006, 14, 7062–7065.
21] C. Wikner, L. Meshalkina, U. Nilsson, S. Backstrom, Y.
Lindqvist, G. Schneider, Eur. J. Biochem. 1995, 233,
7
50–755.
22] O. May, P. T. Nguyen, F. H. Arnold, Nat. Biotechnol.
000, 18, 317–320.
[30] D. Enders, B. Bockstiegel, Synthesis 1989, 493–496.
[31] D. Enders, I. Breuer, G. Raabe, Synthesis 2005, 3517–
3530.
[
2
2638
asc.wiley-vch.de
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2008, 350, 2631 – 2638