Asymmetric acyloin condensation catalysed by phenylpyruvate decarboxylase. Part 2: Substrate specificity and purification of the enzyme

Article abstract of DOI:10.1016/S0957-4166(01)00095-7

Phenylpyruvate decarboxylase from Achromobacter eurydice was used to catalyse the asymmetric acyloin condensation of phenylpyruvate 1 with various aldehydes 2 to produce optically active acyloins PhCH₂COCH(OH)R 3. The specific activity of the phenylpyruvate decarboxylase enzyme was increased by a factor of 332 after its purification. The molecular weight of the purified enzyme was shown to be 150 kDa by gel filtration chromatography, while SDS gel electrophoresis showed two sub-units with molecular weights of 90 and 40 kDa. The acyloin condensation yield decreased with increasing chain length for straight chain aliphatic aldehydes from 76% for acetaldehyde to 24% for valeraldehyde. The e.e.s of the acyloin products were 87-98%. Low yields of acyloin products were obtained with chloroacetaldehyde (13%) and glycoaldehyde (16%). Indole-3-pyruvate was a substrate of the enzyme and provided acyloin condensation product 3-hydroxy-1-(3-indolyl)-2-butanone 5 with acetaldehyde in 19% yield, while benzoylformate was not a substrate for the enzyme.

Full text of DOI:10.1016/S0957-4166(01)00095-7

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 571–577
Asymmetric acyloin condensation catalysed by phenylpyruvate
decarboxylase. Part 2: Substrate specificity and purification of
the enzyme
Zhiwei Guo, Animesh Goswami, Venkata B. Nanduri and Ramesh N. Patel*
Process Research & Development, Bristol-Myers Squibb Pharmaceutical Research Institute, New Brunswick, NJ 08903, USA
Received 20 December 2000; accepted 29 January 2001
Abstract—Phenylpyruvate decarboxylase from Achromobacter eurydice was used to catalyse the asymmetric acyloin condensation
of phenylpyruvate 1 with various aldehydes 2 to produce optically active acyloins PhCH2COCH(OH)R 3. The specific activity of
the phenylpyruvate decarboxylase enzyme was increased by a factor of 332 after its purification. The molecular weight of the
purified enzyme was shown to be 150 kDa by gel filtration chromatography, while SDS gel electrophoresis showed two sub-units
with molecular weights of 90 and 40 kDa. The acyloin condensation yield decreased with increasing chain length for straight chain
aliphatic aldehydes from 76% for acetaldehyde to 24% for valeraldehyde. The e.e.s of the acyloin products were 87–98%. Low
yields of acyloin products were obtained with chloroacetaldehyde (13%) and glycoaldehyde (16%). Indole-3-pyruvate was a
substrate of the enzyme and provided acyloin condensation product 3-hydroxy-1-(3-indolyl)-2-butanone 5 with acetaldehyde in
19% yield, while benzoylformate was not a substrate for the enzyme. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
2. Results and discussion
2.1. Enzyme purification
Chiral unymmetrically substituted acyloins (a-hydroxy-
ketones) are important classes of intermediates in
organic synthesis due to their bifunctional nature, and
the fact that they have a stereogenic centre amenable to
further synthetic manipulation. Present chemical meth-
ods for acyloin synthesis generally give racemic prod-
ucts.1,2 Enzyme mediated acyloin formation would
provide an environmentally friendly method to prepare
chiral acyloins.3 Acyloin formations mediated by yeast
pyruvate decarboxylase (EC 4.1.1.1)4,5 and bacterial
benzoylformate decarboxylase (EC 4.1.1.7)6,7 have been
studied by a number of researchers. In our previous
work,8 we reported that phenylpyruvate decarboxylase
(EC 4.1.1.43) catalysed the asymmetric acyloin conden-
sation of phenylpyruvate 1 with acetaldehyde 2a to
produce (R)-(−)-3-hydroxy-1-phenyl-2-butanone 3a
(Scheme 1). Herein, we describe the purification and
immobilisation of the enzyme and the substrate specific-
ity of the enzyme for both acyl donors (a-keto-acids)
and acyl acceptors (aldehydes).
Purification of phenylpyruvate decarboxylase from
Achromobacter eurydice SC16386 was performed. The
specific activity of the enzyme was increased 332-fold
compared to the crude cell extract by six successive
column purification steps (Table 1). The purified
enzyme eluted as a single peak on gel filtration chro-
matography with a molecular weight of 150 kDa. SDS
gel electrophoresis showed two bands, probably corre-
sponding to sub-units with molecular weights of 90 and
40 kDa. By using the most active fraction of
phenylpyruvate decarboxylase from the UNO Q1
column of purification step 4, (R)-(−)-3a was obtained
in 91% yield with an e.e. of 85%. The use of purified
enzyme minimised decarboxylation and other side reac-
tions and afforded higher yields of acyloin products.
2.2. Enzyme immobilisation
The enzyme from the cell extracts of A. eurydice was
immobilised on DE52 ion exchange resin. Immobilised
enzyme was used to run 14 reaction cycles (usually
reaction time of one day per cycle). The activity of the
immobilised enzyme dropped significantly after the first
two cycles (from 0.084 unit in the first to 0.032 unit for
the third cycle). Subsequently, the activity loss per cycle
* Corresponding author. E-mail: ramesh.patel@bms.com
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00095-7

572
Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
OH
R2
CO2H
R2
Phenylpyruvate
decarboxylase
R1
+ R1-CHO
O
O
2
1 or 4
3 or 5
CH2-
1 and 3, R2 =
2 and 3, (R1):
a (Me), b (Et), c (n-Pr), d (n-Bu), e (PhCH2), f (Ph), g (H),
h (Me(CH2)6), i (Me(CH2)8), j (Me(CH2)10), k (PhCH=CH), l (Br3C), m (Me2CH),
n (Me3CCH2), o (BrCH2), p (BrCH2CH2), q (CH2=CH), r (ClCH2), s (HOCH2).
CH2-
R2 =
4 and 5,
R1 = CH3
N
H
Scheme 1.
Table 1. Purification of phenylpyruvate decarboxylase from Achromobacter eurydice SC16386
Purification step
Total protein
(mg)
Total activity
(unit=umol/h)
Specific activity (unit/mg protein)
Purification factor
Crude cell extract
Step 1: DE52
8634
550
28
10.04
3.22
1.98
0.34
47.62
29.00
11.21
0.0055
0.0527
0.4004
1
10
73
Step 2: phenyl sepharose
Step 3: sephacryl S-100/desalting
Step 4: UNO-Q1
Step 5: Superdex 200
Step 6: ISO column/desalting
5.48
0.5458
99
a
a
a
2.21
0.62
1.1162
1.8235
203
332
a Salt interfered with the activity assay and the activity was not determined.
was minimal and appreciable activity (0.013 unit)
remained even after all 14 cycles.
saturation). This partially purified enzyme preparation
enabled the use of higher amounts of protein in concen-
trated form and to obtain higher yields of the acyloin
condensation product. This enzyme preparation was
stable when stored at about 4°C (no change in activity
after 64 days) and was used in the substrate specificity
studies. The rate of product 3a formation by the par-
tially purified enzyme was nearly parallel to substrate 1
consumption. Product 3a was stable under the reaction
condition and was obtained in 76% yield.
2.3. Substrate specificity of different aldehydes as acyl
acceptors
The substrate specificity of the enzyme for different
aldehydes as acyl acceptors was studied by using
phenylpyruvate 1 as the acyl donor (Scheme 1). The
crude cell extract was used in the initial study to
examine the effect of acetaldehyde concentration on the
acyloin condensation. The results are shown in Fig. 1.
An optimum aldehyde concentration was necessary so
as to obtain the highest yield of the acyloin product.
The yield of acyloin product was low at low aldehyde
concentrations, probably due to decarboxylation and
other competing reactions, while at high aldehyde con-
centrations the enzyme was probably inhibited. The
best yields of 34–35% were obtained at acetaldehyde
concentrations of 50–100 mM with 5 mM concentra-
tion of phenylpyruvate using the crude cell extract.
The specificity of the enzyme with different aldehyde
substrates was studied using the partially purified
enzyme. As with acetaldehyde, there was an optimum
concentration for each aldehyde (Fig. 1). The yields of
reactions using the optimum concentration of each
aldehyde were compared (Table 2). With straight chain
aliphatic aldehydes, the yield decreased with increasing
chain length and the yields of acyloin products were in
the following order for different R groups: Me, 76%>
Et, 55%>n-Pr, 38%>n-Bu, 24%. Using longer chain
aldehydes 2h, 2i, 2j, aromatic aldehydes 2e, 2f, 2k,
hindered aldehydes 2l, 2m, 2n, bromoacetaldehyde 2o,
bromopropionaldehyde 2p and acrolin 2q, either none
A partially purified enzyme was prepared from the cell
extract of A. eurydice by DE52 ion exchange chro-
matography and ammonium sulfate precipitation (60%

Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
573
Figure 1. Acyloin condensation of 1 with aldehydes 2a, 2b, 2c, and 2d.
Table 2. Acyloin condensation catalysed by the partially purified enzyme
Product entry
R
Optimum aldehyde concentration (mM)
Yield (%)
Enantiomeric excess (%)
3a
3b
3c
3d
3e–3q
3r
3s
Me
Et
n-Pr
n-Bu
80
200
300
300
76
55
38
24
B3
13
87
98
98
92
a
a
a
ClCH2
HOCH2
50
50
16
a Not determined.
or negligible amounts (0–3%) of the acyloin products
were obtained. The acyloin products were obtained in
13 and 16% yields with chloroacetaldehyde 2r and
glycoaldehyde 2s, respectively. Formaldehyde 2g was
not suitable as a substrate for the enzyme.
2.4. Substrate specificity of different a-keto-acids as
acyl donors
The substrate specificity of the enzyme was also studied
for different a-keto-acids. Enzymatic acyloin condensa-
tion of indole-3-pyruvic acid 4 with acetaldehyde gave
3-hydroxy-1-(3-indolyl)-2-butanone
5 in 19% yield
The substrate specificity observed for phenylpyruvate
decarboxylase is different from benzoylformate decar-
boxylase, though both accept acetaldehyde as a sub-
strate. The later accepts aromatic aldehydes, but does
not accept propionaldehyde, chloroacetaldehyde, and
glycoaldehyde.7,9
(Scheme 1). The acyloin product was isolated and struc-
1
ture 5 was established by 1D and 2D H NMR and
LCMS. Using benzoylformic acid 6 as an acyl donor
and acetaldehyde as an acyl acceptor, no acyloin
product was found. There was no decarboxylation to
form benzaldehyde. Thus bezoylformate is not a suit-
able substrate for phenylpyruvate decarboxylase from
A. eurydice.
HPLC analysis of the enzymatic acyloin condensation
products 3a–3d showed that the first eluting enantiomer
was produced predominantly with each aldehyde,
whilst the condensation was highly enantiospecific
providing 87–98% enantiomeric excesses (e.e.) of the
first eluting enantiomer with different aldehydes. The
(−)-(R)-3a enantiomer has been synthesised previously
and its absolute configuration determined.8 Although
the absolute configurations of the enzymatic acyloin
products 3b–3d were not established conclusively
because the pure (−)-(R)-3a enantiomer was found to
have similar retention times in HPLC to the major, first
eluting products, this indicated that they have (R)-
configuration.
3. Experimental
3.1. Chemicals and microorganisms
Chemicals were purchased from VWR and/or Aldrich.
DE52 cellulose was purchased from Whatman Inc.
Phenylsepharose was obtained from Pharmacia Bio-
tech. The UNO Q1 ion exchange column was from
Bio-Rad. Achromobacter eurydice SC16386 was
obtained and grown in a large scale fermentor as
described previously.8

574
Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
3.2. Analytical methods
3.4. Protein assay
1H NMR spectra were recorded in CDCl3 using a
JEOL Eclipse 400 operating at 400 MHz. When nec-
essary, 2% D2O was added as indicated.
The Bio-Rad protein assay was used to determine
protein concentration.10 Samples containing 1–20 mL
of enzyme fraction were diluted to 0.8 mL with
water. Bio-Rad reagent (0.2 mL) was added and the
solution was mixed thoroughly. The protein concen-
tration (mg/mL) was calculated from the standard
curve with bovine serum albumin as standard.
HPLC method 1 was used for identification and
quantification of different components, while methods
2 and 3 were used for determination of e.e.s
Method 1 was performed on a reversed phase C-18
column (5 mm, 15×0.46 cm, Kromasil) at 40°C with a
flow rate of 1.0 mL/min using gradient elution as
follows: 75% solvent A (0.1% H3PO4 in water) and
25% solvent B (CH3CN) for the first 10 min, then
increasing to 100% solvent B over 20 min. UV detec-
tion was at 210 nm. Retention times for phenylpyru-
vate, phenylacetic acid, and phenylacetaldehyde were
5.0, 6.7, and 8.0 min, respectively. Retention times for
chemically prepared racemic acyloins 3a–3f were 7.5,
13.6, 17.5, 20.0, 20.0, and 18.9 min, respectively.
3.5. Enzyme purification
All purification processes were carried out at about
4°C. A. eurydice cells (100 g) were suspended in
buffer A (500 mL) (100 mM potassium phosphate, 1
mM EDTA, 0.1 mM DTT, 2 mM MgSO4, 0.1 mM
TPP, 10% glycerol, pH 6.8). The cells were disrupted
by three passages through a microfluidiser (Model
110F, Microfluidics Co., Newton, Massachusetts,
USA). The resulting suspension was centrifuged at
27 000 g for 1 h. The active supernatant was collected
as the crude cell extract.
Method 2 for separation of acyloin enantiomers of 3a,
3b, 3c, 3e, and 3f was performed on a Chiralpak
AD column (25×0.46 cm, Daicel) at ambient tempera-
ture with a mixture of hexane–isopropanol–ethanol
(95.4:4.0:0.6) as eluent at a flow rate of 1.0 mL/min
and UV detection at 210 nm. Retention times in min-
utes for the two enantiomers were as follows: 3a
(14.8/18.4), 3b (11.3/16.2), 3c (10.4/12.2), 3e (17.8/
19.1), and 3f (17.5/20.7).
3.5.1. Enzyme purification step 1: DE52 column. The
crude cell extract was applied on to a DE52 cellulose
ion exchange column (14×5.0 cm) previously equili-
brated with buffer A. The column was washed with
buffer A (1 L), followed by gradient elution from
buffer A to buffer B (B is the same as A, except 200
mM w.r.t potassium phosphate—gradient complete in
800 mL), then from buffer B to buffer C (0.8 M
NaCl in buffer B—gradient complete in 800 mL). The
fractions with high specific activity were combined.
Method 3 for separation of enantiomers of 3d was
the same as method 2 except the eluent was a mix-
ture of hexane:iso-propanol (99.2:0.8). Retention
times for the two enantiomers of 3d were 38.0 and
46.7 min.
3.5.2. Enzyme purification step 2: phenylsepharose
column. The active fractions from step 1 were treated
with NaCl (10 g), adjusted to 60% (NH4)2SO4 satura-
tion and centrifuged at 27 000 g for 1 h. The precipi-
tate was dissolved in buffer D (80 mL) (1 M
(NH4)2SO4 in buffer A). The supernatant obtained by
centrifugation at 27 000 g for 1 h was loaded on to a
phenylsepharose column (180 mL) previously equili-
brated with buffer D. The column was washed with
buffer E (450 mL) [0.5 M (NH4)2SO4, in buffer A]
followed by gradient elution from buffer E to buffer
A in 720 mL, then 420 mL of buffer F (the same as
buffer A except 10 mM of potassium phosphate). The
fractions with high specific activity were pooled (40
mL).
LCMS analysis was performed on a reversed phase
C-18 column (5 mm, 5×0.46 cm, Luna) at ambient
temperature with a flow rate of 4.0 mL/min using
gradient elution from solvent A (5% CH3CN–95%
H2O–10 mM NH4OAc) to solvent B (95% CH3CN–
5% H2O–10 mM NH4OAc) in 4 min with positive ion
electrospray (ES+) or negative ion electrospray (ES−)
methods.
3.3. Enzyme activity assay
The enzyme activity was assayed by acyloin conden-
sation of 1 with 2a. A mixture of enzyme fraction
(500 mL), potassium phosphate buffer (100 mL, 500
mM, pH 6.8), MgCl2 (30 mL, 100 mM), thiamine
pyrophosphate (TPP, 20 mL, 50 mM), water (250 mL),
1 (50 mL, 100 mM), and 2a (50 mL, 1 M) was stirred
at 28°C for 20 h. The reaction was terminated by
addition of aqueous HCl (50 mL, 1 M) and MeOH
(950 mL). The resulting mixture was filtrated and sub-
jected to HPLC method 1 to determine the amounts
of product 3a. In some cases, the same procedure was
employed on a smaller scale. One unit of activity was
defined as 1 mmol 3a formed per h.
3.5.3. Enzyme purification step 3: sephacryl S-100
column. The active fractions from step 2 were concen-
trated (50 kDa cut off membrane) and loaded on to a
sephacryl S-100 (100×2.5 cm) gel filtration column
previously equilibrated with buffer G (0.2 M NaCl in
buffer A). The column was eluted with buffer G. The
major activity was found in 70 mL of the eluate, which
was then concentrated (50 kDa cut off membrane) to 3
mL and diluted with buffer A to 15 mL to reduce the
salt concentration.

Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
575
3.5.4. Enzyme purification step 4: UNO-Q1 column. The
3.7. Chemical syntheses of ( )-3a to 3f
enzyme solution from step 3 was concentrated (50 kDa
cut off membrane) to 2 mL and loaded on to an
UNO-Q1 ion exchange column (Bio-Rad, 1 mL). The
column was eluted with buffer A (5 mL) followed by a
gradient elution with increasing buffer H (0.8 M NaCl
in buffer A) from 0 to 60% in 20 mL. The fractions
with high specific activity were pooled (2 mL).
The synthesis of ( )-3a was completed as reported
previously.8 ( )-3b to 3f were synthesised similarly
using the following general procedure. To a stirred
solution of trimethylsilyl cyanide (20 mmol, 1984 mg,
2667 mL) in CH2Cl2 (80 mL) at 0°C, the aldehyde 2 (20
mmol) was added, followed by TEA (2 mmol, 202 mg,
278 mL). The resulting mixture was stirred at 0°C for 2
h and the solvent was removed under reduced pressure.
The residue was dissolved in anhydrous ether (20 mL)
and added to a stirred solution of benzylmagnesium
chloride (20 mL, 1.0 M in ether). The reaction mixture
was stirred under reflux for 2 h and then cooled to
room temperature. Water (20 mL) was added with
caution followed by HCl (4 M, 20 mL) with vigorous
stirring. The progress of the reaction was followed by
TLC. After completion of hydrolysis, the product was
extracted with ethyl acetate (2×40 mL). The extract was
washed successively with water, 5% sodium bicarbon-
ate, water, 1 M HCl, water, and brine, then dried over
MgSO4 and concentrated to dryness.
3.5.5. Enzyme purification step 5: Superdex 200 column.
The pooled fractions from step 4 were concentrated by
ultrafiltration (50 kDa cut off membrane) to 0.2 mL
and loaded onto a Superdex 200 column (Pharmacia,
24 mL) previously equilibrated with buffer A. The
column was eluted with buffer A. The fractions with
high specific activity were combined (2 mL).
3.5.6. Enzyme purification step 6: ISO column. The
pooled fractions from step 5 were adjusted to 1 M
concentration of (NH4)2SO4, loaded onto an ISO
hydrophobic interaction column (Pharmacia, 1 mL)
previously equilibrated with buffer D. The column was
eluted with 5 mL of buffer D followed by gradient
elution from buffer D to buffer A in 10 mL. An activity
assay was performed after (NH4)2SO4 was removed
from each fraction (50 kDa cut off membrane). The
fractions with highest specific activity were pooled (2
mL).
Flash chromatography on silica and elution with
CH2Cl2 gave oily products 3b (1.54 g, 43%); 3c (2.27 g,
59%); 3d (2.14 g, 52%); and 3e (1.62 g, 34%). For 3f,
elution with CH2Cl2–MTBE (19:1) gave a solid, which
was recrystallised from hexane–MTBE (2:1) to afford 3f
(0.43 g, 10%). The results of HPLC analyses are shown
in Section 3.2.
3.5.7. Determination of molecular weight. Gel filtration
column (Superdex 200, Pharmacia, 24 mL) was cali-
brated with standard proteins (Gel filtration chro-
matography standard, Bio-Rad) and eluted with buffer
G (0.2 M NaCl in buffer A). The elution parameter
kav=(Ve−Vo)/(Vt−Vo) was used for the preparation of
calibration curve, where Ve=elution volume for the
protein, Vo=column void volume, and Vt=total bed
volume.11
1
3.7.1. 3-Hydroxy-1-phenyl-2-pentanone 3b. H NMR l
0.92 (t, 3H), 1.64 (m, 1H), 1.95 (m, 1H), 3.34 (m, 1H),
3.76 (d, 1H), 3.79 (d, 1H), 4.25 (m, 1H), 7.1–7.4 (m,
5H); (CDCl3–2% D2O): l 0.92 (t, J=7.3, 3H), 1.64 (m,
1H), 1.93 (m, 1H), 3.76 (d, J=15.6, 1H), 3.79 (d,
J=15.6, 1H), 4.25 (m, 1H), 7.1–7.4 (m, 5H). LCMS
(ES+) m/z 179 (M+H).
1
3.7.2. 3-Hydroxy-1-phenyl-2-hexanone 3c1. H NMR l
0.95 (t, 3H), 1.30–1.75 (m, 3H), 1.90 (m, 1H), 3.32 (d,
1H, D2O exchangeable), 3.75 (d, 1H), 3.84 (d, 1H), 4.32
(m, 1H), 7.1–7.4 (m, 5H). LCMS (ES+) m/z 193 (M+
H).
3.6. Immobilisation of the enzyme and acyloin conden-
sation with the immobilised enzyme
DE52 resin (10 mL bed volume) was equilibrated with
buffer A and added to the crude cell extract of A.
eurydice (20 mL). The mixture was shaken at 20 rpm
for 20 h and then filtered. The adsorbed resins were
washed with buffer A (5×30 mL) and suspended in the
same buffer to give a final volume of 30 mL. One
fourth of the suspension was used to catalyse the
acyloin condensation of 1 with 2a in 14 reaction cycles.
For each cycle, a mixture of the immobilised enzyme,
MgCl2 (120 mL, 100 mM), TPP (80 mL, 50 mM), 1 (200
mL, 100 mM), 2a (200 mL, 1 M), and buffer A (to adjust
the final volume to 10 mL) was shaken at 50 rpm, 28°C
for 20 h and then filtered. The resins were washed with
buffer A (3×10 mL) and then used in the next reaction
cycle. The combined filtrate and washes were acidified
with HCl (1 M, 2 mL) and extracted with MTBE (2×15
mL). The extract was evaporated to dryness. The
residue was dissolved in CH3CN (2 mL) and analysed
by HPLC method 1.
1
3.7.3. 3-Hydroxy-1-phenyl-2-heptanone 3d. H NMR l
0.90 (t, J=7.3, 3H), 1.25–1.48 (m, 4H), 1.58 (m, 1H),
1.88 (m, 1H), 3.38 (brs, 1H), 3.76 (d, J=15.1, 1H), 3.80
(d, J=15.1, 1H), 4.27 (m, 1H), 7.15–7.38 (m, 5H).
LCMS (ES+) m/z 207 (M+H).
1
3.7.4. 3-Hydroxy-1,4-diphenyl-2-butanone 3e. H NMR
l 2.89 (dd, J1=14.2, J2=7.3, 1H), 3.17 (dd, J1=14.2,
J2=4.9, 1H), 3.25 (d, J=5.8, 1H, D2O exchangeable),
3.75 (d, J=15.6, 1H), 3.81 (d, J=15.6, 1H), 4.50 (m,
1H), 7.14–7.38 (m, 10H). LCMS (ES+) m/z 258 (M+
NH4).
1
3.7.5. 1-Hydroxy-1,3-diphenyl acetone 3f. H NMR l
3.62 (d, J=16.1, 1H), 3.66 (d, J=16.1, 1H), 4.24 (d,
J=4.2, 1H, D2O exchangeable), 5.19 (d, J=4.2, 1H; s

576
Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
after D2O exchange), 7.00 (d, 2H), 7.2–7.5 (m, 8H).
LCMS (ES+) m/z 244 (M+NH4).
HCl, and brine. After removal of solvent, the residue
was subjected to HPLC method 1 again to analyse the
non-acidic components, and to HPLC method 2 or 3 to
determine the e.e.
3.8. Chemical synthesis of 1-hydroxy-3-phenyl-2-
propanone 3g
3.12. Effect of acetaldehyde concentration
A solution of phenylacetyl chloride (1.55 g, 10 mmol) in
ether (40 mL) was slowly added to a stirred solution of
diazomethane in ether (220 mL, 70 mmol, alcohol free)
cooled with an ice-salt bath. After 1 h at room temper-
ature, the solvent was removed under reduced pressure.
The residue was dissolved in tert-butyl alcohol (20 mL)
and stirred with boron trifluoride diethyl etherate (1.5
mL) at 70°C for 1 h. The usual work-up and flash
chromatography on silica and elution with hexane–
ethyl acetate (2:1) afforded 3g, retention time 4.8 min
The crude cell extract (500 mL for each sample) was
used in the initial study to examine the effect of acetal-
dehyde concentration. A clear stock solution of 1 M
acetaldehyde in water was used. Duplicate samples were
set up for each aldehyde at concentrations of 0, 5, 10,
20, 50, 100, 200, and 300 mM. The acyloin condensa-
tion was carried out as described in Section 3.11 and
the products were analysed by HPLC method 1.
1
by HPLC method 1. H NMR l 3.00 (brs, 1H, D2O
3.13. Time course of enzymatic acyloin condensation of
1 with 2a
exchangeable), 3.72 (s, 2H), 4.28 (s, 2H), 7.21 (d, 2H),
7.26–7.38 (m, 3H).12 LCMS (ES+) m/z 151 (M+H).
Nine identical reactions were set up with the partially
purified enzyme at an acetaldehyde concentration of 80
mM. The reactions were stopped at 2, 4, 8, 20, 28, 44,
52, 72, and 140 h and analysed by HPLC method 1.
3.9. Chemical syntheses of ( )-2-hydroxy-1-phenyl-
propanone 7
Lactonitrile (20 mmol, 1.42 g, 1.43 mL in 10 mL ether)
was added slowly to a solution of phenylmagnesium
bromide (30 mL, 1.0 M in ether) at 0°C. The resulting
mixture was refluxed for 1 h. Water (10 mL) was added
with caution followed by 10% H2SO4 (20 mL). The
ethereal extract was washed successively with 5%
NaHCO3, water, and brine, and then dried over MgSO4
and concentrated to dryness. Flash chromatography on
silica and elution with CH2Cl2 gave oily product 7 (1.46
g, 49%). Retention time 6.1 min by HPLC method 1.
1H NMR l 1.38 (d, 3H), 3.82 (brd, 1H), 5.08 (m, 1H),
7.41 (t, 2H), 7.52 (t, 1H), 7.83 (d, 2H). LCMS (ES−)
m/z 209 (M+OAc).
3.14. Enzymatic acyloin condensation of 1 with 2b–2g
The partially purified enzyme was used. The 1 M
aldehyde stock solutions were prepared in water for 2b
and 2g and in 50% aqueous ethanol for 2c–2f. The
reactions were set up at aldehyde concentrations of 5,
10, 20, 50, 100, 200, and 300 mM. For reactions at 500,
1000, and 2000 mM aldehyde concentrations neat 2c
and 2d were used. The reactions were analysed by
HPLC method 1 and compared with authentic samples
prepared chemically as described in Sections 3.7 and
3.8.
3.10. Preparation of the partially purified enzyme
3.15. Enzymatic acyloin condensation of 1 with 2h–2q
The active fractions from enzyme purification step 1
were combined, adjusted to 60% (NH4)2SO4 saturation,
and centrifuged at 27 000 g for 1 h. The precipitate was
dialysed against buffer A (2×2 L) to give the partially
purified enzyme preparation, which was used for sub-
strate specificity study.
The partially purified enzyme was used. The aldehydes
2h–2k or 1 M stock solutions of 2l–2q were used to set
up reactions at different aldehyde concentrations. The
reaction products were analysed by HPLC method 1.
HPLC analyses did not show any significant additional
peaks of more than 3 area percent when compared to
the simple decarboxylation reaction and the control
reactions without any enzyme.
3.11. General procedure for enzymatic acyloin
condensation
3.16. Enzymatic acyloin condensation of 1 with 2r and
2s
A mixture of the enzyme (500 mL), potassium phos-
phate buffer (100 mL, 500 mM, pH 6.8), MgCl2 (30 mL,
100 mM), TPP (20 mL, 50 mM), water (to adjust the
final volume to 1 mL), sodium phenylpyruvate (50 mL,
100 mM), and the aldehyde (to the final concentration
as indicated) was stirred at 28°C for 20 h or as indi-
cated. A control reaction without any enzyme was
always carried out for each aldehyde. The reaction was
terminated by addition of aqueous HCl (50 mL, 1 M)
and MeOH (950 mL). The resulting mixture was filtered
and subjected to HPLC method 1 to determine the
amounts of the various components. The mixture was
then extracted with ethyl acetate (2×3 mL). The organic
layer was washed with 5% NaHCO3 (twice), water, 1 M
The partially purified enzyme was used and the reac-
tions were set up at aldehyde 2s concentrations of 5, 10,
20, 50, 100, 200, and 300 mM. The reactions were
analysed by HPLC method 1. Estimated percent yields
of a new product peak at 3.7 min were 8, 12, 15, 16, 16,
15, and 13, respectively. There was 49% remaining of 1
in the sample with an initial aldehyde concentration of
50 mM. The reaction mixture was then extracted with
EtOAc. The extract was washed twice with 5%
NaHCO3, water, 1 M HCl, and brine and then evapo-
rated to dryness to provide product 3s, retention time
1
3.7 min by HPLC method 1. H NMR l 3.6 (brs, OH,

Z. Guo et al. / Tetrahedron: Asymmetry 42 (2001) 571–577
577
D2O exchangeable), 3.8 (d, J=16, 1H), 3.9 (d, J=16,
1H), 4.0 (m, 2H), 4.3 (m, 1H), 7.1 (d, 2H), 7.2–7.4 (m,
3H); LCMS (ES−) m/z 179 (M−H).
mM) as the acyl donor in place of sodium phenylpyru-
vate and acetaldehyde (100 mM) as the acceptor.
HPLC (method 1) analysis and comparison with
authentic samples of 2-hydroxy-1-phenyl-propanone 7
(prepared by conventional chemical synthesis) and
commercial benzaldehyde showed the absence of either
one in the reaction mixture.
When the concentration of 2r was 50 mM, a similar
procedure gave 13% of 3r as the highest yield with a
retention time of 13.8 min, analysed by HPLC method
1; LCMS (ES−) m/z 197/199 with an intensity ratio of
3/1 (M−H).
Acknowledgements
3.17. Enzymatic acyloin condensation catalysed by
active fraction from UNO Q1 column
We are indebted to Mr. Steve Schwarz and Mr. David
Brzozowski for their help in the growth of microorgan-
isms, to Dr. Thomas L. Laporte and Dr. Dana Caz-
zulino for the fermentation and recovery of the cells, to
Mr. K. David Mirfakhrae for his help in the develop-
ment of HPLC methods, and to the BMS Analytical
Department for their help with NMR and LCMS
analyses. We would also like to acknowledge Dr.
Richard Mueller for his comments.
The reaction was conducted under conditions similar to
that for the enzyme activity assay but using 20% of the
materials in each case; the most active fraction from the
UNO Q1 column purification step 4 was used. The
yield of acyloin product 3a was 91% analysed by HPLC
method 1, and the e.e. was 85% determined by HPLC
method 2.
3.18. Enzymatic acyloin condensation of indole-3-pyru-
vic acid as an acyl donor
The reaction was conducted under the same conditions
as stated in Section 3.11, with indole-3-pyruvic acid 4 (5
mM) as the acyl donor in place of sodium phenylpyru-
vate and acetaldehyde (100 mM) as the acceptor. When
compared (HPLC method 1) with the control reaction
without enzyme, a new non-acidic product 5 was found
with a retention time 7.5 min in 19% yield. The reaction
was scaled up by using 50 mL of the enzyme prepara-
tion, which was prepared from the crude cell extract by
ammonium sulfate precipitation (30–60% saturation)
and dialysis against buffer A. The product was isolated
by flash chromatography on silica and elution with
hexane:ethyl acetate (2:1) to give purified acyloin
References
1. Gill, M.; Kiefel, M. J.; Lally, D. A. Tetrahedron Lett.
1986, 27, 1933–1934.
2. Vankar, Y. D.; Chaudhuri, N. C.; Singh, S. P. Synth.
Commun. 1996, 96, 443–473.
3. Ward, O. P.; Baev, M. In Stereoselective Biocatalysis;
Patel, R. N., Ed. Decarboxylases in Sterereoselective
Catalysis; Marcel Dekker: New York, 2000; pp. 267–287.
4. Iding, H.; Siegert, P.; Mesch, K.; Pohl, M. Biochim.
Biophys. Acta 1998, 1385, 307–322.
5. Rogers, P. L.; Shin, H. S.; Wang, B. Adv. Biochem. Eng.
Biotechnol. 1997, 56, 33–59.
6. Prosen, E.; Ward, O. P. J. Ind. Microbiol. 1994, 13,
287–291.
1
product 5. H NMR l 1.44 (d, J=7.0, 3H), 3.94 (s,
2H), 4.40 (q, J=7.0, 1H), 7.10–7.16 (m, 2H), 7.20 (t,
1H), 7.38 (d, 1H), 7.50 (d, 1H), 8.18 (brs, 1H, NH);
1H–1H COSY in agreement with the structure of the
desired product 5; LCMS (ES+) m/z 204 (M+H), (ES−)
m/z 202 (M−H).
7. Iding, H.; Dunnwald, T.; Greiner, L.; Liese, A.; Muller,
M.; Siegert, P.; Grotzinger, J.; Demir, A. S.; Pohl, M.
Chem. Eur. J. 2000, 6, 1483–1495.
8. Guo, Z.; Goswami, A.; Mirfakhrae, K. D.; Patel, R. N.
Tetrahedron: Asymmetry 1999, 10, 4667–4675.
9. Demir, A. S.; Dunnwald, T.; Iding, H.; Pohl, M.; Muller,
M. Tetrahedron: Asymmetry 1999, 10, 4769–4774.
10. Bradford, M. Anal. Biochem. 1976, 72, 248–254.
11. Reiland, J. Methods Enzymol. 1971, 22, 287–321.
12. Wissner, A. J. Org. Chem. 1979, 44, 4617–4622.
3.19. Enzymatic reaction of benzoylformic acid as an
acyl donor
The reaction was conducted under the same conditions
as stated in Section 3.11, with benzoylformic acid 6 (5
.
.