Mutation of cysteine-295 to alanine in secondary alcohol dehydrogenase from thermoanaerobacter ethanolicus affects the enantioselectivity and substrate specificity of ketone reductions

Article abstract of DOI:10.1016/S0968-0896(01)00073-6

The mutation of Cys-295 to alanine in Thermoanaerobacter ethanolicus secondary alcohol dehycrogenase (SADH) was performed to give C295A SADH, on the basis of molecular modeling studies utilizing the X-ray crystal structure coordinates of the highly homologous T. brockii secondary alcohol dehydrogenase (YKF.PDB). This mutant SADH has activity for 2-propanol comparable to wild-type SADH. However, the C295A mutation was found to cause a significant shift of enantioselectivity toward the (S)-configuration in the reduction of some ethynylketones to the corresponding chiral propargyl alcohols. This result confirms our prediction that Cys-295 is part of a small alkyl group binding pocket whose size determines the binding orientation of ketone substrates, and, hence, the stereochemical configuration of the product alcohol. Furthermore, C295A SADH has much higher actifity towards t-butyl and some α-branched ketones than does wild-type SADH. The C295A mutation does not affect the thioester reductase activity of SADH. The broader substrate specificity and altered stereoselectivity for C295A SADH make it a potentially useful tool for asymmetric reductions. Copyright

Full text of DOI:10.1016/S0968-0896(01)00073-6

Bioorganic & Medicinal Chemistry 9 (2001) 1659–1666
Mutation of Cysteine-295 to Alanine in Secondary Alcohol
Dehydrogenase from Thermoanaerobacter ethanolicus Affects the
Enantioselectivity and Substrate Specificity of Ketone Reductions
Christian Heiss,a,y Maris Laivenieks,d J. Gregory Zeikusd and Robert S. Phillipsa,b,c,
*
aDepartment of Chemistry and Molecular Biology, University of Georgia, Athens, GA 30602-2556, USA
bDepartment of Biochemistry and Molecular Biology, University of Georgia, Athens, GA 30602-2556, USA
cCenter for Metalloenzyme Studies, University of Georgia, Athens, GA 30602-2556, USA
dDepartment of Biochemistry, Michigan State University, East Lansing, MI 48824-1319, USA
Received 18 September 2000; accepted 22 January 2001
Abstract—The mutation of Cys-295 to alanine in Thermoanaerobacter ethanolicus secondary alcohol dehydrogenase (SADH) was
performed to give C295A SADH, on the basis of molecular modeling studies utilizing the X-ray crystal structure coordinates of the
highly homologous T. brockii secondary alcohol dehydrogenase (1YKF.PDB). This mutant SADH has activity for 2-propanol
comparable to wild-type SADH. However, the C295A mutation was found to cause a significant shift of enantioselectivity toward
the (S)-configuration in the reduction of some ethynylketones to the corresponding chiral propargyl alcohols. This result confirms
our prediction that Cys-295 is part of a small alkyl group binding pocket whose size determines the binding orientation of ketone
substrates, and, hence, the stereochemical configuration of the product alcohol. Furthermore, C295A SADH has much higher
activity towards t-butyl and some a-branched ketones than does wild-type SADH. The C295A mutation does not affect the thioe-
ster reductase activity of SADH. The broader substrate specificity and altered stereoselectivity for C295A SADH make it a poten-
tially useful tool for asymmetric reductions. # 2001 Elsevier Science Ltd. All rights reserved.
Alcohol dehydrogenases (ADHs) have been used exten-
sively for the asymmetric synthesis of chiral building
blocks because of their high chemoselectivity, enantio-
selectivity, tolerance of a wide spectrum of functional
groups, and their relative ease of use.1 A number of
ADH’s have been isolated from thermophilic organ-
isms, the best studied being that from Thermoanaero
bacter brockii (TBADH)2 Due to their high stability to
heat, organic co-solvents, and immobilisation, these
enzymes have received considerable attention in chiral
synthesis. Although thermostable ADHs have been
shown to reduce many different ketones bearing a vari-
ety of other functionalities enantioselectively to the
corresponding secondary alcohols, there is still a need to
further explore their synthetic scope and to deepen our
knowledge of the physical principles of their stereo-
chemistry.
In an earlier study,3 we investigated the asymmetric
reduction of ethynyl ketones and ketoesters by second-
ary alcohol dehydrogenase (SADH) from T. ethanolicus
[eq (1)], and we found that trifunctional chiral building
blocks are obtained in good yield and excellent
enantiomeric excess. We also showed that although
ethynylketones are substrates for SADH and the ste-
reospecificity of their reduction clearly follows the
established ‘pocket’-model,4,5 these a,b-unsaturated
ketones also irreversibly inactivate the enzyme, which
requires the use of relatively large enzyme concentra-
tions to obtain good conversion. We had postulated
that a nucleophilic residue in the active site might cause
the inactivation by adding to the electron-deficient triple
bond in the ethynylketone substrates. This result stimu-
lated us to perform modeling studies of SADH using the
available coordinates of the highly homologous enzyme,
TBADH. The result of these modeling studies predicted
*Corresponding author at the Department of Chemistry, University of
GA, Athens, GA 30602-2556, USA. Tel.: +1-706-542-1996; fax: +1-
706-542-9454; e-mail: rsphillips@chem.uga.edu
yPresent address: McNeil Specialty Products Company, Athens,
Georgia.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00073-6

1660
C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
that Cys-295 is located in the postulated ‘small alkyl
group’ binding pocket. In the present study, we investi-
gated the effect of the mutation Cys295 ! Ala (C295A)
on the enantioselectivity of ethynylketone reduction and
on substrate specificity of SADH.
residues located in the small pocket, Ala-85, Ile-86, Thr-
153, and Cys-295 are only partially conserved in related
sequences (bold characters, Fig. 2). After our studies
had been completed, the X-ray structure of the apo-
enzyme of TBADH with bound (S)-2-butanol
(1BXZ.PDB) was published,8 and a model for DMSO
bound to TBADH was proposed.9 The substrate bind-
ing site presented in these papers is very similar to our
model shown in Figure 1; however, the potential role of
Cys-295 in substrate binding was not recognized.
Results and Discussion
We examined the X-ray crystal structure6,7 of TBADH
(1YKF.PDB), which is very highly homologous (99%
identity) to SADH, in order to find the location and
orientation in which the substrates bind to the enzyme.
Only one conformation of a good substrate,3 isopropyl
4-oxo-5-hexynoate, was found that was free from severe
steric overlaps upon docking in the active site. In this
structure, presented in Figure 1, the ethynyl group pro-
trudes into a region enclosed by the residues, His-59,
Ala-85, Ile-86, Asp-150, Thr-153, and Cys-295. Thus,
these residues make up the previously proposed ‘small
alkyl binding pocket’.4,5 The alkyl chain bearing the
ester group occupies a space surrounded by Cys-37, Ser-
39, Leu-107, Trp-110, Tyr-267, Gly-293, Leu-294, and
the nicotinamide ring of the cofactor. Cys-37, His-59
and Asp-150, which are conserved in all sequences of
SADH and related enzymes (bold italic characters, Fig. 2),
are ligands to the catalytic zinc ion (Fig. 1). The other
The nicotinamide ring is also very close in space to the
a-hydrogens of the substrate in our model (Fig. 1). This
observation explains why any branching in this position
precludes binding in this orientation, as can be seen not
only in the lack of reactivity of 4-methyl-1-heptyn-3-one
(8 in Table 1), but also in the high enantiomeric excess
obtained in the reduction of 4-methyl-1-pentyn-3-one (2
in Table 1) by wild-type SADH. The model places the
ester carbonyl oxygen of isopropyl 4-oxo-5-hexynoate
within hydrogen bonding distance of Ser-39, and
although no significant difference in enantiomeric excess
or yield between ethynylketoesters and simple ethy-
nylketones of comparable size were found, we have
found that the ketoesters react much faster than the
simple ketones.3 Moreover, this result suggests that
hydrogen bonding may take place between Ser-39 and
Figure 1. Active site model of wild-type SADH with bound substrate, isopropyl 4-oxo-5-hexynoate. Coordinates were taken from 1YKF.PDB.

C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
1661
Figure 2. Alignment of SADH from T. ethanolicus and related sequences. S71131: T. ethanolicus; P14941, T. brockii; P25984, Clostridium beijer-
inckii; P35630, Entamoeba histolytica; P75214, Mycoplasma pneumoniae; P14940, Alcaligenes euthrophus. The zinc ligands are shown in italics, and
the residues which comprise the small alkyl binding pocket are shown in bold.

1662
C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
Table 1. C295A SADH reductions of ethynylketones
The specific activity of C295A SADH (43 U/mg) is only
slightly lower than that of the wild-type enzyme (54 U/
mg)12 indicating that the mutation does not disrupt the
active site structure. We reduced a series of ethynyl
ketones with SADH C295A in order to compare the
outcome with our previous studies with wild-type
SADH3. However, the ethynylketones were found to
inactivate the mutant enzyme in the same manner and
extent as was found with wild-type SADH (data not
shown). Thus, Cys-295 does not appear to be involved
in the mechanism of inactivation. It is possible that the
inactivation is due to reaction of the zinc ligand, Cys-37,
with the ethynylketones. Alternatively, Peretz et al.
reported that TBADH is inactivated by sulhydryl
reagents reacting at Cys-203.13
Entry
Substrate
Yielda,b (%) Abs. conf.a
ee.a (%)
76 (80)
1
39 (32)
88 (50)
39 (0)
S (S)
S (S)
S (S)
S (S)
S (R)
—(R)
S (R)
S (–)
R (R)
2
3
4
5
6
7
8
9
>98 (>98)
85 (85)
76 (51)
56 (50)
—(66)
51 (28)
42 (20)
0 (0)
We observed a significant effect of the C295A mutation
on the enantioselectivity of the reductions, as can be
seen in Table 1. These reactions were performed with
the goal of optimizing chemical yield rather than
enantioselectivity; thus, the reactions cannot be com-
pared quantitatively. Nevertheless, the comparison of
the outcome of the reactions of wild-type and C295A
mutant SADH under otherwise identical conditions is
of practical interest. It is apparent that the stereo-
chemical outcome for small substrates is largely unaf-
fected (entries 1–3), likely because the the reactions are
governed primarily by the degree of steric interaction
with the large pocket (i.e., the nicotinamide ring) rather
than by the size of the small pocket, which can easily
accommodate any of the substituents in this group.
With wild-type SADH, 4-methyl-1-pentyne-3-one (2 in
Table 1) exhibits very high enantioselectivity for reduc-
tion. We propose that this is due both to the ability of
the small alkyl pocket to comfortably accommodate the
isopropyl group as well as to the unfavorable interac-
tions of the a-branching for binding the isopropyl group
in the large pocket. There is no significant difference in
the result of reduction of 2 with wild-type or C295A
mutant SADH (Table 1). However, the yield of product
(but not the ee) from reduction of a t-butyl ketone (3) is
much higher for C295A SADH (Table 1). The steric
interaction with the cofactor should be smallest with the
ethynyl group, due to its linear geometry, hence the
preference for the binding orientation that leads to the
(S)-enantiomers. The substrates of intermediate size (5
and 7 in Table 1), however, experience a reversal of
preferred configuration of the product from (R) for
wild-type to (S) for C295A SADH, indicating that the
small pocket in the mutant SADH is able to accom-
modate significantly larger alkyl groups than wild-type
SADH. Furthermore, substrate 4 shows a marked
increase in stereoselectivity of reduction with C295A
SADH. The most striking evidence for the importance
of the steric interactions of a-substituents with the
dihydronicotinamide ring comes from compound 8,
4-methyl-1-heptyn-3-one, which does not react at a sig-
nificant rate with wild-type SADH. In contrast, 8 is
rapidly reduced by C295A SADH to give a 2:1 mixture
of (3S,4S) and (3S,4R)-diastereomers in good chemical
yield, with an enantiomeric excess for each diastereomer
greater than 98%. Thus, the ‘small alkyl’ pocket is now
large enough to accomodate even the sec-pentyl group,
60 (32)
43 (0)
67 (42)
>98 (–)
60 (82)
23 (35)
aAbsolute configuration. Results of reductions with wild-type SADH
are given in parentheses.
bYields reported are isolated yields.
the amide nitrogen in the proposed physiological sub-
strate of SADH, acetyl CoA.10 Our previous finding
that the mutation Ser-39!Thr shifts the enantiospecifi-
city for 2-butanol and 2-pentanol toward the (R)-enan-
tiomers supports our conclusion that Ser-39 is part of
the large pocket in SADH.11 We hypothesized that Cys-
295 may be responsible for the irreversible inactivation
of the enzyme observed during reaction with some eth-
ynyl ketones.3 If so, replacement of Cys-295 with a non-
nucleophilic residue by site-directed mutagenesis should
therefore abolish the observed inactivation. Cys-295 is
conserved in some, but not all, related ADH sequences.
In the ADH from Entamoeba histolytica, it is replaced
by threonine, and in that from Mycoplasma pneumoniae,
it is replaced with methionine (Fig. 2). We chose the
Cys-295!Ala mutation as the least disruptive mutation
to test this hypothesis. This mutation should further-
more result in an enlargement of the small alkyl binding
pocket, due to the absence of the large sulfur atom,
which has a steric bulk comparable to a CH2, a change
which should be reflected in a change in substrate spe-
cificity and enantioselectivity. Since a larger ‘small
pocket’ would accommodate larger alkyl substituents,
an overall shift of enantioselectivity towards (R)-alco-
hols for simple alkyl ketone reduction and (S)-alcohols
for ethynyl ketone reduction should entail if the model
shown in Figure 1 is correct.

C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
Table 2. Kinetic parameters of thioester reduction by wild-type and
C295A SADH
1663
Experimental
General
Compound
SADH
Km
kcat
kcat/Km
(Â103 MÀ1 sÀ1
(Â10À3 M) (sÀ1
)
)
Enzyme assays and kinetic experiments were performed
on a Varian Cary 1E UV–visible spectrophotometer
equipped with a Peltier-type thermoelectric tempera-
ture-controlled 6Â6 cell cuvette changer. Capillary GC
was performed on a Varian 3300 gas chromatograph
with FI detection (Supelco b-Dex 120 chiral column,
30 mÂ0.250 mm id, 0.25 mm film thickness) programmed
Ethylthioacetate Wild-type
6.0
14
0.6
1.2
37
38
73
50
6.2
2.7
122
42
Ethylthioacetate
Acetyl–CoA
Acetyl–CoA
C295A
Wild-type
C295A
between 40 and 175 ꢀC. H and 13C NMR spectra were
1
whereas binding of this group in the large pocket is still
precluded because of the severe steric interactions
between NADPH and the a-methyl group in this sub-
strate. As the size of the alkyl substituent is increased
further, the influence of the Cys to Ala mutation
decreases substantially, so that an ethynyl ketoester
substrate, 9, reacts similarly with wild-type and mutant
SADH.
taken on a Bruker AC-250 or AC-300 spectrometer,
respectively, using the residual CHCl3 signal (d
7.26 ppm) as internal reference. Optical rotations were
measured on a Rudolph Autopol IV polarimeter.
Enzyme assay
SADH was assayed at 50 ꢀC by following the increase in
absorbance at 340 nm due to the formation of NADPH
(Áe340=6.22 mMÀ1 cmÀ1).5 The assay mixture con-
tained 200 mM 2-propanol and 1.25 mM NADP in
50 mM Tris–HCl buffer, pH 8.0. One unit (U) is the
amount of enzyme required to reduce 1 mmol of NADP
per minute.
Burdette et al.10 found that SADH from T. ethanolicus
acts as a thioester reductase, reducing the acetyl group
of acetyl CoA to give a thiohemiacetal, which releases
CoA and acetaldehyde, which is further reduced to eth-
anol. Hence, it was suggested that acetyl CoA is the
physiological substrate for SADH. Within this context,
we examined whether Cys-295 is required for the
thioesterase mechanism of SADH. If so, the thioester
reductase activity should be greatly reduced or abol-
ished in C295A SADH. However, the kinetic para-
meters for both ethyl thioacetate and acetyl CoA show
only a small difference between wild-type and C295A
SADH (Table 2). These differences are not significant
enough to indicate any involvement of Cys-295 in the
chemical mechanism of thioester reduction. It is inter-
esting to note that acetyl CoA is a significantly better
substrate than ethyl thioacetate, with comparable kcat
values but Km values about 10-fold lower (Table 2). This
result is in accord with our finding that longer chain
ketoesters such as isopropyl 4-oxo-5-hexynoate are bet-
ter substrates for reduction than are alkyl ketones.3
Modeling studies of SADH
Low-energy conformations of isopropyl 4-oxo-5-hexy-
noate, an excellent substrate that gives very high ee of
(R)-alcohol upon reduction,3 were identified using the
Tripos force-field (Sybyl 6.3, Tripos Inc., 1996). All
waters were removed from the protein structure
(1YKF.PDB), and these ketones were then docked into
the active site using the Sybyl 6.3 program in such a way
that the carbonyl oxygen was in close proximity to the
catalytic zinc ion and to the pro-R hydrogen in the 4-posi-
tion of the nicotinamide ring of the NADPH cofactor.
The substrate was oriented such that hydride transfer
would occur on its si-face, since we had observed very
high stereoselectivity for (R)-alcohol in the reaction. In
order to accommodate the substrate, it was necessary to
rotate the nicotinamide ring about the C–N bond
between the pyridine nitrogen and the ribose-C1. A
similar rotation was performed by Kleifeld et al.9 in
modeling the complex of TBADH and DMSO. Com-
parison of the HLADH structures with and without
DMSO bound had shown that a rotation of the nicoti-
namide ring occurs on ligand binding.14,15 After thus
obtaining reasonable starting structures, these were then
energy-minimized using the TRIPOS force-field and
examined for validity based on the absence of steric
overlap.
Conclusions
The results presented herein provide a good indication
of the location and orientation of substrates in the
active site of SADH. The residues that comprise the
large and small alkyl binding pockets have been clearly
identified by modeling studies and confirmed by our
experimental results. These results show that modifica-
tion of the side chains making up the small alkyl bind-
ing pocket by site-directed mutagenesis results in a
significant change in both substrate specificity and in
enantioselectivity of ketone reduction without sig-
nificant loss of activity. Starting with this active-site
model, together with our previous results from the
mutation of Ser39!Thr, further mutations in the small
pocket could be envisioned to result in even higher spe-
cificity for (R)-alcohols. C295A SADH has broader
substrate specificity for a-branched ketones than wild-
type SADH and TBADH, making it a potentially useful
tool in asymmetric reductions.
Sequence alignments of SADH
The amino acid sequence of SADH was used as a query
for a PSI-BLAST search in the nr database at NCBI.
Those sequences with highest similarities were then
aligned using the PILEUP program of GCG.16 For dis-
play, the output from the PILEUP program was run
with BOXSHADE to generate Figure 2.

1664
C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
Preparation of C295A mutant SADH
The residue was taken up in Et2O (40 mL) and the
resulting solution was stirred vigorously overnight with
an aqueous solution of NaF (0.462 g, 11.0 mmol) and
NBu4Cl (0.31 g, 1.1 mmol). The layers were separated
and the aqueous layer extracted with Et2O (3Â25 mL).
The combined organic extracts were dried with Na2SO4,
the solvent was removed in vacuo, and the residue
chromatographed on silica gel.
Point mutations were introduced as described pre-
viously.17 PCR-based T. ethanolicus 39E adhB gene
mutagenesis was performed using plasmid pADHB25-
kan as template. An oligonucleotide primer (KA4N)
was synthesized complementary to the noncoding
strand that included a KpnI restriction site, the native
adhB ribosome binding site, and the adhB translation
initiation codon (50-CGGGGTACCCCGTATTTTA-
Enzymatic reduction of ethynylketones with SADH
C295A
GGAGGTGTTTAATGATGAAAGG-30).
A second
roligonucleotide primer (KA4C) that included the com-
plement to the adhB termination codon and the ApaI
estriction site was synthesized complementary to the cod-
ing strand (50-CAGTCCGGGCCCTTATGCTAATAT-
SADH C295A (600 U) and NADP (1 mg, 1.2 mmol)
were dissolved in 50 mM Tris–HCl buffer, pH 8.0,
(8.5 mL). After preincubation for 10 min at 50 ꢀC, the
appropriate ketone (1.0 mmol) was added in one por-
tion and the mixture was kept at 50 ꢀC. As soon as the
ketone was consumed (by GC analysis), the reaction
mixture was saturated with NaCl and extracted with
Et2O (3Â4 mL). Due to their tendency to form emul-
sions, the extractions had to be centrifuged (10 min at
4000 rpm) prior to separation. The combined extracts
were dried with Na2SO4, the solvent was removed in
vacuo, and the residue chromatographed. The reactions
were stopped immediately upon consumption of the
ketone, so they were not allowed to reach equilbrium,
and the products were at least partially kinetically
controlled.
TACAACAGGTTTG-30).
Complementary
33
oligonucleotide primers that contained the mutated
bases (50-CTATAAAAGGCGGGCTAGCCCCCGGT-
GGACGTC-30) and (50-GACGTCCACCGGGGGC-
TAGCCCGCCTTTTATAG-30) were used together
with the two KA4 end primers to amplify the 50 and 30
ends of the adhB gene. PCR synthesis of mutated partial
and complete adhB gene were performed using TagPlus
PrecisionTM PCR system (Stratagene, LA Jolla, CA).
Obtained constructs were subcloned in the KpnI and
ApaI restricted pBluescriptII KS (ꢁ) and expressed in
Eschericlia coli DH5a. Mutations were verified by whole
gene sequencing using the Thermo SequenaseTM radio-
labeled terminator cycle sequencing kit (USB Corpora-
tion, Cleveland, OH, USA). DNA work was performed
using published protocols.18,19
Determination of optical purity of the alcohols
The alcohol (0.02 mmol) was dissolved in CDCl3
(0.3 mL) in an NMR tube, and a 50 mM solution of
Eu(hfc)3 in CDCl3 (0.5 mL) was added in portions of
0.1 mL. A 1H NMR spectrum was taken after each
addition, and the two signals corresponding to the dia-
stereotopic carbinol protons were integrated. If only one
signal was visible, the enantiomeric excess was assumed
to be greater than 98%. Enantiomeric excess was calcu-
lated as either %(R) À%(S) or %(S) À%(R), depending
on which enantiomer was present in greater amount.
Purification of C295A SADH
For asymmetric reductions. E. coli cells expressing
C295A SADH were grown as described.17 The wet cells
(12 g) were suspended in 50 mM Tris–HCl buffer, pH
8.0, (50 mL) containing 0.1 mM dithiothreitol, soni-
cated, and centrifuged (30 min at 10000Âg, all of the
following centrifugations were done under these condi-
tions). The supernatant was heated in a capped bottle at
70 ꢀC for 1 h, and the thick suspension was then cen-
trifuged. This heat treatment was repeated once more,
whereupon the supernatant had an activity of 92 U/ml
and a specific activity of 27 U/mg protein.
(S)-1-Pentyn-3-ol. Yield 32.7 mg (39%), 76% ee, Rf 0.20
1
(10% ether/pet. ether), H NMR (300 MHz, CDCl3) d
4.32 (dt, J=6.5 Hz, 2.0 Hz, 1H), 2.46 (d, J=1.8 Hz, 1H),
1.74 (m, 3H), 1.02 (t, J=7.4 Hz, 3H).
For determination of kinetic parameters. The enzyme
solution obtained above was further purified by affinity
chromatography on Red-Agarose as previously descri-
bed11 and gave C295A SADH with a specific activity of
43 U/mg protein.
(S)-4-Methyl-1-pentyn-3-ol. Yield 86.3 mg (88%),
>98% ee, Rf 0.22 (10% ether/pet. ether), 1H NMR
(300 MHz, CDCl3) d 4.17 (dd, J=5.7 Hz, 2.0 Hz, 1H),
2.45 (d, J=2.3 Hz, 1H), 1.89 (m, 2H), 1.01 (m, 6H).
(S)-4,4-Dimethyl-1-pentyn-3-ol. Yield 13.2 mg (22%,
based on 53% conversion), 85% ee, Rf 0.22 (10% ether/
pet. ether), [a]D20 À10.0ꢀ (c 1.5, dioxane) (lit.21 [a]2D0
Preparation of ethynylketones.20 A solution of acyl
chloride (10.0 mmol) and bis-TMS-acetylene (1.87 g,
11.0 mmol) in dry CH2Cl2 (25 mL) was cooled to 0 ꢀC
and AlCl3 (1.60 g, 12.0 mmol) was added in portions
during 30 min. The mixture was stirred for 2 h at 0 ꢀC
and for 2 h at room temperature and then poured into a
1:1 mixture of 1 N HCl and ice (50 mL). The layers were
separated and the aqueous phase extracted with CH2Cl2
(3Â30 mL). The combined organic extracts were dried
with Na2SO4 and the solvent was removed in vacuo.
+35.6ꢀ (c 3, dioxane), 90% ee for (R)-enantiomer), H
1
NMR (250 MHz, CDCl3) d 4.02 (d, J=2.1 Hz, 1H), 2.45
(d, J=2.2 Hz, 1H), 1.8 (bs, 1H), 1.00 (s, 9H).
(S)-1-Hexyn-3-ol. Yield 48.1 mg (51%), 76% ee, Rf 0.22
(10% ether/pet. ether), [a]2D0 À9.9ꢀ (c 2.5, CHCl3) (lit.22
[a]2D0 +9.0ꢀ (c 1.0, CHCl3), 75% ee for (R)-enantiomer),
1H NMR (250 MHz, CDCl3) d 4.39 (dt, J=6.5 Hz,

C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
1665
2.2 Hz, 1H), 2.47 (d, J=2.1 Hz, 1H), 1.71 (m, 3H), 1.50
(m, 2H), 0.97 (t, J=7.0 Hz, 3H).
(80 mg, 0.47 mmol) and, at 0 ꢀC, AlCl3 (80 mg). The
mixture was stirred for 2 h at 0 ꢀC and for 2 h at room
temperature and then poured into a 1:1 mixture of 1 M
HCl and ice. The layers were separated, and the aqu-
eous phase extracted with CH2Cl2. The combined
organic extracts were dried with Na2SO4 and the solvent
was removed in vacuo. The residue was taken up in
Et2O and the resulting solution was stirred vigorously
overnight with an aqueous solution of NaF (20 mg,
0.5 mmol) and NBu4Cl (14 mg, 0.05 mmol) in H2O
(2 mL). The layers were separated and the aqueous layer
extracted with Et2O. The combined organic extracts
were dried with Na2SO4, the solvent was removed in
vacuo, and the residue chromatographed on silica gel.
Yield 28.2 mg (48%); ee 78%; 1H NMR (250 MHz,
CDCl3) d 3.22 (s, 1H), 2.59 (m, J=6.8 Hz, 1H), 1.79 (m,
1H), 1.39 (m, 3H), 1.19 (d, J=7.2 Hz, 3H), 0.93 (t,
J=7.2 Hz, 3H); 13C NMR (75 MHz, CDCl3) d 191.1,
80.9, 79.0, 48.1, 34.5, 20.1, 15.5, 13.9.
(S)-5-Methyl-1-hexyn-3-ol. Yield 47.3 mg (42%), 56%
ee, Rf 0.20 (10% ether/pet. ether), [a]2D0 À16.1ꢀ (c 2.3,
dioxane) (lit.21 [a]D25 +28.8ꢀ (c 3, dioxane), 88% ee for
(R)-enantiomer), 1H NMR (300 MHz, CDCl3) d 4.42
(dt, J=7.4 Hz, 2.3 Hz, 1H), 2.46 (d, J=2.1 Hz, 1H), 1.87
(m, 1H), 1.60 (m, 3H), 0.95 (d, J=6.7 Hz, 3H), 0.93 (d,
J=6.8 Hz, 3 H).
(S)-1-Heptyn-3-ol. Yield 67.8 mg (60%), 67% ee, Rf 0.19
(10% ether/pet. ether), [a]2D0 À8.5ꢀ (c 2.6, CHCl3) (lit.22
[a]2D0 À5.5ꢀ (c 0.9, CHCl3), 70% ee for (S)-enantiomer),
1H NMR (250 MHz, CDCl3) d 4.37 (dt, J=7.3 Hz,
2.1 Hz, 1H), 2.47 (d, J=2.1 Hz, 1H), 1.87 (bs, 1H), 1.72
(m, 2H), 1.41 (m, 4H), 0.92 (t, J=7.0 Hz, 3H).
(R)-Methyl 4-Hydroxy-5-hexynoate. Yield 15.8 mg
(23%, based on 48% conversion), 60% ee, Rf 0.20 (20%
ethyl acetate/hexanes), [a]2D0 +7.0ꢀ (c 1.8, CHCl3), H
Preparation of ethyl thioacetate.24 A mixture of NaOH
(4.0 g, 0.1 mol), H2O (7.5 mL), and ethanethiol (8.9 mL,
0.12 mol) was poured onto cracked ice (50 g). Acetic
anhydride (11.8 mL, 0.125 mol) was added with vigor-
ous stirring, which was continued for 5 min, whereby
the thioester separated immediately. The product was
washed with H2O (5Â5 mL), dried with MgSO4, and
distilled, bp 112–114 ꢀC (lit. bp 114–116 ꢀC). Yield, 5.8 g
(56%).
1
NMR (250 MHz, CDCl3) d 4.49 (dt, J=6.0 Hz, 2.0 Hz,
1H), 3.69 (s, 3H), 2.55 (m, 2H), 2.49 (d, J=2.1 Hz, 1H),
2.21–1.98 (m, 3H).
(3S,4S)- and (3S,4R)-4-Methyl-1-heptyn-3-ol. (ꢁ)-4-
Methyl-1-heptyn-3-one (100 mg, 0.81 mmol) was
reduced by C295A SADH. The reaction was monitored
by GC and worked up after 70% conversion. The pro-
duct was found to be a 2:1 mixture of diastereomers.
Starting material [17 mg (17%)] was recovered, and
found by GC to be identical to (R)-4-methyl-1-heptyn-
3-one (see below). Therefore, the major product of the
enzyme reaction was (3S,4S)-4-methyl-1-heptyn-3-ol.
Yield 44 mg (43%), >98% ee, Rf 0.19 (10% ether/pet.
ether), [a]2D0 À10.2ꢀ (c 2.75, dioxane) 1H NMR
(250 MHz, CDCl3) d 4.28 (dt, J=5.5 Hz, 2.1 Hz, 1H),
2.46 (d, J=2.2 Hz, 1H), 1.85–1.2 (m, 6H), 1.02 and 1.00
(2:1) (d, J=6.7 Hz, 3H), 0.92 (t, J=6.9 Hz, 3H).
Acknowledgements
The authors would like to thank Dr. Paul Simon for his
assistance in performing the modeling studies.
References and Notes
Resolution of 2-methylpentanoic acid.23 Lipase from
Candida rugosa (Sigma L1745, 750 units/mg,1.0 g),
1-octanol (130 mg, 1.0 mmol), 2-methylpentanoic acid
(116 mg, 1.0 mmol), and heptane (10 mL) were com-
bined and shaken at room temperature for 14 h. The
reaction progress was followed by TLC (5% ethyl ace-
tate/hexanes containing 0.25% acetic acid, detection by
vanillin) Rf (ester) 0.53, (acid) 0.08, (alcohol) 0.03. The
enzyme was removed by filtration, the filtrate evapo-
1. Wong, C., Whitesides, G. M.; Enzymes in Synthetic Organic
Chemistry; Tetrahedron Organic Chemistry Series, Vol. 12;
Elsevier Science: Oxford, 1994.
2. Lamed, R. J.; Keinan, E.; Zeikus, J. G. Enzyme Microb.
Technol. 1981, 3, 144.
3. Heiss, C.; Phillips, R. S. J. Chem. Soc., Perkin 1 2000, 2821.
4. Keinan, E.; Hafeli, E. K.; Seth, K. K.; Lamed, R. J. Am.
Chem. Soc. 1986, 108, 162.
5. Pham, V. T.; Phillips, R. S. J. Am. Chem. Soc. 1990, 112,
3629.
6. Korkhin, Y.; Frolow, F.; Bogin, O.; Peretz, M.; Kalb, A. J.;
Burstein, Y. Acta Crystallogr. 1996, D52, 882.
7. Khorkin, Y.; Kalb (Gilboa), A. J.; Peretz, M.; Bogin, O.;
Burstein, Y.; Frolow, F. J. Mol. Biol. 1998, 278, 967.
8. Li, C.; Heatwole, J.; Saelaiman, S.; Shoham, M. Proteins
1999, 37, 619.
9. Kleifeld, O.; Frenkel, A.; Bogin, O.; Eisenstein, M.; Brumfeld,
V.; Burstein, Y.; Sagi, I. Biochem. 2000, 39, 7702.
10. Burdette, D. S.; Zeikus, J. G. Biochem. J. 1994, 302, 163.
11. Tripp, A. E.; Burdette, D. S.; Zeikus, J. G.; Phillips, R. S.
J. Am. Chem. Soc. 1998, 120, 5137.
rated,
and
the
components
separated
by
chromatography. Yield of (S)-ester, 73.6 mg (64%),
yield of (R)-acid, 55.0 mg (93%).
(R)-4-Methyl-1-heptyn-3-one.
(R)-2-Methylpentanoic
acid (55.0 mg, 0.47 mmol) and KOH (31 mg) were dis-
solved in MeOH (1.0 mL), allowed to stand for 90 min,
and evaporated to dryness. Benzene (2 mL) was added
and the mixture stirred for 60 min. At 0 ꢀC, DMF (1 mL)
and oxalyl chloride (250 mg, 2 mmol) were added and
stirring was continued at room temperature for 2 h. The
excess of oxalyl chloride was removed by evaporation
with CHCl3. The residue was taken up in CH2Cl2, fil-
tered, and treated with bis-(trimethylsilyl)acetylene
12. Burdette, D. S.; Vieille, C.; Zeikus, J. G. Biochem. J. 1996,
316, 115.
13. Peretz, M.; Weiner, L. M.; Burstein, Y. Protein Sci. 1997,
6, 1074.

1666
C. Heiss et al. / Bioorg. Med. Chem. 9 (2001) 1659–1666
14. Eklund, H.; Branden, C. I. J. Biol. Chem. 1979, 254,
3458.
15. Ecklund, H.; Samma, J. P.; Wallen, L.; Branden, C. I.;
Akeson, A.; Jones, T. A. J. Mol. Biol. 1981, 146, 561.
16. Wisconsin Package Version 10.0; Genetics Computer
Group (GCG): Madison, WI.
19. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D.,
Seidman, J. G., Smith, J. A., Struhl, K., Eds.; Current Proto-
cols in Molecular Biology; Greene Publishing and Wiley-Inter-
science: New York, 1993.
20. Walton, D. R. M.; Waugh, F. J. Organomet. Chem. 1972,
37, 45.
17. Burdette, D. S.; Secundo, F.; Phillips, R. S.; Dong, J.;
Scott, R. A.; Zeikus, J. G. Biochem. J. 1997, 326, 717.
18. Sambrook, J.; Fritch, E. F.; Maniatis, T. In Molecular
Cloning: A Laboratory Manual, 2nd ed.; Nolan, C, Ed.; Cold
Spring Harbor: Cold Spring Harbor, 1989.
21. Vigneron, J.; Bloy, V. Tetrahedron Lett. 1979, 2683.
22. DeMarquez, M. D.; Rowland, P. J.; Scopes, P. M.; Thaller,
V. J. Chem. Res. (M) 1986, 1348.
23. Engel, K.-H. Tetrahedron: Asymmetry 1991, 2, 165.
24. Wenzel, R. J. Am. Chem. Soc. 1937, 59, 1089.