Assessing the Substrate Selectivities and Enantioselectivities of Eight Novel Baeyer-Villiger Monooxygenases toward Alkyl-Substituted Cyclohexanones

Article abstract of DOI:10.1021/jo030253l

Genes encoding eight Baeyer-Villiger monooxygenases have recently been cloned from bacteria inhabiting a wastewater treatment plant. We have carried out a systematic investigation in which each newly cloned enzyme, as well as the cyclohexanone monooxygenase from Acinetobacter sp. NCIB 9871, was used to oxidize 15 different alkyl-substituted cyclohexanones. The panel of substrates included equal numbers of 2-, 3-, and 4-alkyl-substituted compounds to probe each enzyme's stereoselectivity toward a homologous series of synthetically important compounds. For all 4-alkyl-substituted cyclohexanones tested, enzymes were discovered that afforded each of the corresponding (S)-lactones in ≥98% ee. This was also true for the 2-alkyl-substituted cyclohexanones examined. The situation was more complex for 3-akyl-substituted cyclohexanones. In a few cases, single Baeyer-Villiger monooxygenases possessed both high regio- and enantioselectivities toward these compounds. More commonly, however, they showed only one type of selectivity. Nonetheless, enzymes with such properties might be useful as parts of a two-step bioprocess where an initial kinetic resolution is followed by a regioselective oxidation on the isolated, optically pure ketone.

Full text of DOI:10.1021/jo030253l

Assessin g th e Su bstr a te Selectivities a n d En a n tioselectivities of
Eigh t Novel Ba eyer -Villiger Mon ooxygen a ses tow a r d
Alk yl-Su bstitu ted Cycloh exa n on es
Brian G. Kyte,† Pierre Rouvie`re,‡ Qiong Cheng,‡ and J on D. Stewart*,†
127 Chemistry Research Building, University of Florida, Gainesville, Florida 32611, and
DuPont, Inc., Experimental Station, E328, Wilmington, Delaware 19880-0328
jds2@chem.ufl.edu
Received August 7, 2003
Genes encoding eight Baeyer-Villiger monooxygenases have recently been cloned from bacteria
inhabiting a wastewater treatment plant. We have carried out a systematic investigation in which
each newly cloned enzyme, as well as the cyclohexanone monooxygenase from Acinetobacter sp.
NCIB 9871, was used to oxidize 15 different alkyl-substituted cyclohexanones. The panel of
substrates included equal numbers of 2-, 3-, and 4-alkyl-substituted compounds to probe each
enzyme’s stereoselectivity toward a homologous series of synthetically important compounds. For
all 4-alkyl-substituted cyclohexanones tested, enzymes were discovered that afforded each of the
corresponding (S)-lactones in g98% ee. This was also true for the 2-alkyl-substituted cyclohexanones
examined. The situation was more complex for 3-akyl-substituted cyclohexanones. In a few cases,
single Baeyer-Villiger monooxygenases possessed both high regio- and enantioselectivities toward
these compounds. More commonly, however, they showed only one type of selectivity. Nonetheless,
enzymes with such properties might be useful as parts of a two-step bioprocess where an initial
kinetic resolution is followed by a regioselective oxidation on the isolated, optically pure ketone.
In tr od u ction
Villiger oxidants offer a clear strategy for improving
substrate- and stereoselectivities by altering ligand
structures and reaction conditions and, in principle,
provide access to the antipodal product if the metal
ligand(s) can be obtained in mirror image form.
Homochiral ꢀ-caprolactones are widely used building
blocks in organic and polymer synthesis. Among the
variety of possible routes that have been explored,1-5
enantioselective Baeyer-Villiger oxidations of appropri-
ately substituted cyclohexanones constitute a particularly
convenient pathway to these compounds since the ketone
precursors can be prepared by well-established methods.
A number of metal-based Baeyer-Villiger catalysts have
been reported, and some of these show impressive stereo-
selectivities, albeit on a limited range of substrates.6-11
It should be noted, however, that such “chemical” Baeyer-
The stereoselectivities of enzymatic Baeyer-Villiger
oxidations are almost uniformly very high for those
ketones accepted as substrates. Flavoprotein monooxy-
genases utilize atmospheric O2 as the oxidant and yield
water as the waste product, which minimizes the envi-
ronmental burden of these reactions.12-15 Biocatalytic
strategies, however, can also be problematic when it is
necessary to obtain both product enantiomers. This
means that a different enzyme with complementary
stereoselectivity (rather than the enantiomer of an exist-
ing catalyst) must be identified. Consequently, the prac-
tical impact of enzymatic Baeyer-Villiger oxidations in
preparative organic chemistry will be increased signifi-
* To whom correspondence should be addressed: Phone: (352) 846-
0743. Fax: (352) 846-2095.
† University of Florida.
‡ DuPont, Inc.
(1) Pirkle, W. H.; Adams, P. E. J . Org. Chem. 1979, 44, 2169-2175.
(2) Blanco, L.; Guibe´-J ampel, E.; Rousseau, G. Tetrahedron Lett.
1988, 1915-1918.
(3) Fouque, E.; Rousseau, G. Synthesis 1989, 661-666.
(4) Fellous, R.; Lizzani-Cuvelier, L.; Loiseau, M. A.; Sassy, E.
Tetrahedron: Asymmetry 1994, 5, 343-346.
(5) Shioji, K.; Matsuo, A.; Okuma, K.; Nakamura, K.; Ohno, A.
Tetrahedron Lett. 2000, 8799-8802.
(10) Strukul, G.; Varagnolo, A.; Pinna, F. J . Mol. Catal. A: Chem.
1997, 117, 413-423.
(11) Bolm, C.; Beckmann, O.; Cosp, A.; Palazzi, C. Synlett 2001,
1461-1463.
(6) Gusso, A.; Baccin, C.; Pinna, F.; Strukul, G. Organometallics
1994, 13, 3442-3451.
(12) Willetts, A. Trends Biotechnol. 1997, 15, 55-62.
(13) Mihovilovic, M. D.; Mu¨ller, B.; Stanetty, P. Eur. J . Org. Chem.
2002, 3711-3730.
(14) Alphand, V.; Carrea, G.; Wohlgemuth, R.; Furstoss, R.; Woodley,
J . M. Trends Biotechnol. 2003, 21, 318-323.
(15) Kamerbeek, N. M.; J anssen, D. B.; van Berkel, W. J . H.; Fraaije,
M. W. Adv. Synth. Catal. 2003, 345, 667-678.
(7) Bolm, C.; Schlingloff, G.; Weickhardt, K. Ang. Chem., Int. Ed.
Engl. 1994, 33, 1848-1849.
(8) Bolm, C.; Schlingloff, G.; Bienewald, F. J . Mol. Catal. A: Chem.
1997, 117, 347-350.
(9) Bolm, C.; Luong, T. K. K.; Schlingloff, G. Synlett 1997, 1151.
10.1021/jo030253l CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/12/2003
12
J . Org. Chem. 2004, 69, 12-17

Novel Baeyer-Villiger Monooxygenases
cantly by expanding the existing repertoire of known
monooxygenases.
molecular biology strategies to clone a variety of genes
encoding novel Baeyer-Villiger monooxygenases. Their
efforts have dramatically increased the available enzyme
diversity. Unfortunately, with a few exceptions,13,29,34 this
growth in monooxygenase availability has not been
matched by a corresponding increase in knowledge of
their substrate selectivities and enantioselectivities.
Here, we report the systematic characterization of eight
novel Baeyer-Villiger monooxygenases cloned by DuPont
workers from organisms found in a wastewater treatment
plant. While preliminary characterization experiments
had verified that each was a bona fide Baeyer-Villiger
monooxygenase, neither their substrate selectivities nor
enantioselectivities had been established prior to these
studies. We therefore used whole cells of engineered E.
coli strains expressing each of the new enzymes to
determine the outcomes of a series of reactions involving
a homologous series of 2-, 3-, and 4-alkyl-substituted
cyclohexanones. We have previously used this same
series of ketones to profile the properties of the Acineto-
bacter sp. NCIB 9871 cyclohexanone monooxygenase,
which provides a basis for comparison with the new
data.35
Cyclohexanone monooxygenase (E.C. 1.14.13.22) from
Acinetobacter sp. NCIB 9871 has been the most exten-
sively studied Baeyer-Villiger monooxygenase since its
initial isolation and characterization by Trudgill and co-
workers in 1976.16 This enzyme has been shown to carry
out asymmetric Baeyer-Villiger oxidations on a variety
of cyclic ketones with high chemo-, regio-, and enanti-
oselectivities (for recent reviews, see refs 13 and 17). The
enzyme-catalyzed pathway is analogous to the stereo-
electronically governed Criegee mechanism followed in
peracid-mediated Baeyer-Villiger oxidations.18-21 This
makes regioselectivity predictable by familiar rules.
Preparative reactions involving Acinetobacter sp. NCIB
9871 cyclohexanone monooxygenase have been carried
out with the purified or semipurified enzyme (with
provision for regeneration of the essential NADPH co-
factor), by whole cells of wild-type or mutant Acineto-
bacter strains, or with whole cells of recombinant baker’s
yeast or Escherichia coli cells (reviewed in refs 13 and
17). Genetically engineered strains are particularly con-
venient since they minimize the possibility of competing
side-reactions and provide high enzyme levels that
maximize reaction rates. With recent advances in process
development that have significantly increased the volu-
metric productivities of these whole-cell-mediated bio-
conversions,22,23 the major issue in enzymatic Baeyer-
Villiger oxidations is the limited array of products that
can be provided solely by the Acinetobacter sp. NCIB 9871
enzyme.
Resu lts a n d Discu ssion
Most of the substituted cyclohexanones required for
this study were available commercially (Scheme 1). The
remaining compounds were synthesized by standard
methods.36-38 Baeyer-Villiger oxidations of all cyclohex-
anones were carried out on 1 mmol scales with growing
cells of E. coli strains that expressed a single heterologous
monooxygenase enzyme. The appropriate genes were
inserted into standard expression plasmids.30-33 Mo-
nooxygenase production was induced by adding isopro-
pylthio-â-D-galactoside (IPTG) when the cultures reached
the late logarithmic phase of growth. The appropriate
cyclohexanone and a stoichiometric quantity of cyclodex-
trin (if necessary) were added to a final concentration of
10 mM 30 min thereafter. The reactions were allowed to
proceed at room temperature until complete.39 For con-
versions involving kinetic resolutions, 50% completion
was reached within 2-6 h. Oxidations of prochiral
cyclohexanones were completed within 12 h. Traces of
indole were the only byproduct detected in the reaction
mixtures, and chemical yields were uniformly high (ac-
cording to GC analyses that included an internal stan-
dard). Both the extent of conversion and the enantiomeric
Several groups have recently addressed the lack of
diversity in Baeyer-Villiger monooxygenases. Over the
past 30 years, several enzymes have been purified from
bacterial species;24-26 unfortunately, the corresponding
genes were not cloned, with one exception.27 The lack of
a cloned gene makes it impossible to use genetically
engineered strains to overproduce the corresponding
protein. This problem has been addressed by Witholt,28
Lau,29 Rouvie`re,30-32 and Cheng,33 who have utilized
(16) Donoghue, N. A.; Norris, D. B.; Trudgill, P. W. Eur. J . Biochem.
1976, 63, 175-192.
(17) Stewart, J . D. Curr. Org. Chem. 1998, 2, 211-232.
(18) Schwab, J . M. J . Am. Chem. Soc. 1981, 103, 1876-1878.
(19) Schwab, J . M.; Li, W.-b.; Thomas, L. P. J . Am. Chem. Soc. 1983,
105, 4800-4808.
(20) Ryerson, C. C.; Ballou, D. P.; Walsh, C. T. Biochemistry 1982,
21, 2644-2655.
(21) Sheng, D.; Ballou, D. P.; Massey, V. Biochemistry 2001, 40,
11156-11167.
(22) Walton, A. Z.; Stewart, J . D. Biotechnol. Prog. 2002, 18, 262-
(31) Brzostowicz, P. C.; Blasko, M. S.; Rouvie`re, P. E. Appl. Micro-
biol. Biotechnol. 2002, 58, 781-789.
268.
(23) Doig, S. D.; Avenell, P. J .; Bird, P. A.; Gallati, P.; Lander, K.
S.; Lye, G. J .; Wohlgemuth, R.; Woodley, J . M. Biotechnol. Prog. 2002,
18, 1039-1046.
(32) Brzostowicz, P. C.; Walters, D. M.; Thomas, S. M.; Nagarajan,
V.; Rouvie`re, P. E. Appl. Environ. Microbiol. 2003, 69, 334-342.
(33) Kostichka, K.; Thomas, S. M.; Gibson, K. J .; Nagarajan, V.;
Cheng, Q. J . Bacteriol. 2001, 183, 6478-6486.
(24) Griffin, M.; Trudgill, P. W. Biochem. Soc. Trans. 1973, 1, 1255-
1258.
(34) Mihovilovic, M. D.; Rudroff, F.; Mu¨ller, B.; Stanetty, P. Bioorg.
Med. Chem. Lett. 2003, 13, 1479-1482.
(25) Britton, L. N.; Brand, J . M.; Markovetz, A. J . Biochim. Biophys.
Acta 1974, 369, 45-49.
(26) van der Werf, M. J . Biochem. J . 2000, 347, 693-701.
(27) Kamerbeek, N. M.; Moonen, M. J . H.; van der Ven, J . G. M.;
van Berkel, W. J . H.; Fraaije, M. W.; J anssen, D. B. Eur. J . Biochem.
2001, 268, 2547-2557.
(28) van Beilen, J . B.; Mouriane, F.; Seeger, M. A.; Kovac, J .; Li, Z.;
Smits, T. H. M.; Fritsche, U.; Witholt, B. Environ. Microbiol. 2003, 5,
174-182.
(29) Iwaki, H.; Hasegawa, Y.; Wang, S.; Kayser, M. M.; Lau, P. K.
C. Appl. Environ. Microbiol. 2002, 68, 5671-5684.
(30) Brzostowicz, P. C.; Gibson, K. L.; Thomas, S. M.; Blasko, M.
S.; Rouviere, P. E. J . Bacteriol. 2000, 182, 4241-4248.
(35) Stewart, J . D.; Reed, K. W.; Martinez, C. A.; Zhu, J .; Chen, G.;
Kayser, M. M. J . Am. Chem. Soc. 1998, 120, 3541-3548.
(36) Stork, G.; Brizzolara, A.; Landesman, H.; Szmuszkovicz, J .;
Terrell, R. J . Am. Chem. Soc. 1963, 85, 207-222.
(37) Eschenmoser, A.; Schreiber, J .; J ulia, S. A. Helv. Chim. Acta
1953, 36, 482-488.
(38) Stork, G.; Danheiser, R. L. J . Org. Chem. 1973, 38, 1775-1776.
(39) A control reaction in which cyclohexanone and isopropyl-thio-
â-D-galactoside were added to a culture of E. coli TOP10 cells lacking
a plasmid showed no ꢀ-caprolactone formation, demonstrating that the
cloned monooxygenases were responsible for all of the Baeyer-Villiger
oxidations observed.
J . Org. Chem, Vol. 69, No. 1, 2004 13

Kyte et al.
SCHEME 1
purity of the residual starting material and the lactone
product(s) were assessed by GC under conditions that
allowed baseline resolution of the racemic materials.
Parallel GC analysis of authentic standards revealed the
absolute configurations of most species and their order
of elution from a chiral-phase GC column. For reactions
involving kinetic resolutions, at least five substrate
enantiomeric excess values were determined during the
first 50% of the reaction; then, these data were analyzed
by computer fitting to an equation that allowed calcula-
tion of the enantioselectivity value (E).40,41
4-Alk yl-Su bstitu ted Cycloh exa n on es. Depending
on the substitution pattern, Baeyer-Villiger oxidations
of cyclohexanones can either involve desymmetrization
(in the case of mesomeric cyclohexanones 1a -e) or kinetic
resolutions (3a -e and 6a -e) (Scheme 1). The former case
is relatively straightforward since only a single lactone
regioisomer is possible. Acinetobacter sp. NCIB 9871
cyclohexanone monooxygenase generally shows good (S)-
selectivity in conversions of 4-alkyl-substituted cyclohex-
anones; however, two cases are problematic: (S)-2c is
available in only 92% ee, and n-butyl-substituted cyclo-
hexanone 1e is accepted as a substrate only at very
high enzyme concentrations and with low enantio-
selectivity.35,42-44 We hoped that one or more of the new
Baeyer-Villiger monooxygenases would provide solutions
to both of these problems.
The results of enzymatic Baeyer-Villiger oxidations
of 1a -e are shown in Figure 1. Nearly all of the enzyme-
substrate pairs exhibited predominantly (S)-selectivity.
In the few cases where (R)-lactones were formed prefer-
entially, the enantiomeric purities were not synthetically
useful (<60%). On the other hand, the expanded collec-
tion of enzymes did provide the two (S)-ꢀ-caprolactone
building blocks that were not produced effectively by the
original Acinetobacter enzyme (2c and 2e). Cyclohex-
anone monooxygenases from Brevibacterium sp. (ChnB1),
Acidovorax CHX, and Rhodococcus SC1 all oxidized 1c
with very high (S)-enantioselectivity, and the Brevibac-
terium sp. ChnB1 and Rhodococcus SC1 enzymes also
showed complete stereoselectivity in the oxidation of 4-n-
butylcyclohexanone 1e.
2-Ak yl-Su bstitu ted Cycloh exa n on es. As expected,
only the “normal” lactone regioisomer was formed in all
biological oxidations of cyclohexanones 3a -e. Acineto-
bacter sp. NCIB 9871 cyclohexanone monooxygenase
exhibits very high E values favoring the (S)-ketones in
conversions of racemic 3b-e where substituents are
larger than a methyl group.41 It shows only limited
enantioselectivity toward 3a , however (E ) 10). Finding
a solution to this problematic reaction was therefore a
key goal.
We were gratified to discover that the cyclododecanone
monooxygenase from Rhodococcus SC1 oxidized 3a with
almost complete (S)-selectivity (E g 200). This is an
unusual trait since all of the other enzymes showed poor
(40) Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294-7299.
(41) Stewart, J . D.; Reed, K. W.; Zhu, J .; Chen, G.; Kayser, M. M.
J . Org. Chem. 1996, 61, 7652-7653.
(44) The presence of an isopropyl substituent in 1d inverts the (R)-/
(S)- designation because of a change in priority assignments. The
absolute sense of stereoselectivity is the same as in the majority of
cases. This is also true for 3d and 6d . Data for these substrates are
marked by an asterix to highlight this phenomenon.
(42) Taschner, M. J .; Black, D. J . J . Am. Chem. Soc. 1988, 110,
6892-6893.
(43) Taschner, M. J .; Black, D. J .; Chen, Q.-Z. Tetrahedron: Asym-
metry 1993, 4, 1387-1390.
14 J . Org. Chem., Vol. 69, No. 1, 2004

Novel Baeyer-Villiger Monooxygenases
3-Ak lyl-Su bstitu ted Cycloh exa n on es. Baeyer-Vil-
liger oxidations of cyclohexanones 6a -e are more com-
plicated than the examples discussed above since both
bonds flanking the carbonyl have nearly equal electron
density (Scheme 1). In the absence of enzymatic guidance,
oxidations of such ketones afford almost equal amounts
of both regioisomers; when carried out on racemic ke-
tones, four different lactone products result. Acinetobacter
sp. NCIB 9871 cyclohexanone monooxygenase shows very
interesting behavior toward this group of substrates. In
the case of methyl- and ethyl-substituted 6a and 6b,
respectively, the two ketone enantiomers are oxidized
cleanly to divergent lactone regioisomers, although the
process occurs with E values only slightly favoring the
(R)-ketones (2 and 18, respectively).35 By contrast, as the
steric size of the substituent is increased (6c-e), high
regioselectivity is observed (favoring 8 and ent-8), al-
though E values are modest.35 For this series of sub-
strates, it is thus important to specify not only the E
value but also the regioselectivity of the oxidations. The
latter property is reported as percent regioisomeric excess
(% re), which was calculated after all of the ketone had
been consumed. The definition of this quantity is analo-
gous to those for enantiomeric and diastereomeric excess
values.45
F IGURE 1. Baeyer-Villiger oxidations of 4-alkyl-substituted
cyclohexanones 1a -e. Enantiomeric excess values for reactions
allowed to proceed to completion are indicated. For each
substrate, the bar colors indicate the identities of the organ-
isms from which the Baeyer-Villiger monooxygenase was
originally cloned: Brevibacterium sp. ChnB1, black; Brevibac-
terium sp. ChnB2, yellow; Acidovorax CHX, orange; Acineto-
bacter SE19, dark blue; Arthrobacter BP2, green; Rhodococcus
phi1, red; Rhodococcus phi2, pink; Rhodococcus SC1, brown;
Acinetobacter sp. NCIB 9871, light blue. Where a bar is absent,
no reaction occurred and only starting material was observed
by GC. The asterix next to 1d indicates that the apparent
selectivity change is due to a change in substituent priority
numbers.
There are four possible outcomes for enzymatic Baeyer-
Villiger oxidations of 3-alkyl-substituted cyclohexa-
nones: (1) high enantio- and regioselectivity (E g 60, %
re g 98), (2) high enantioselectivity but poor regioselec-
tivity (E g 60, % re < 98), (3) poor enantioselectivity with
high regioselectivity (E < 60, % re g 98), and (4) low
enantio- and regioselectivity (E < 60, % re < 98). While
case one is clearly the most desirable, cases two and three
also have synthetic value. Highly enantioselective oxida-
tions provide access to optically pure ketones by removal
of the undesired antipode (case two). After separation
from the unwanted lactone(s), the remaining ketone can
be converted to a homochiral lactone by subsequent
oxidation using an enzyme with poor enantioselectivity
but high regioselectivity (case three). In our study of
newly cloned Baeyer-Villiger monooxygenases (Figure
3), we found examples of all four classes of behavior, with
type three being most common (15 occurrences). Interest-
ingly, none of the enzymes studied here favored formation
of regioisomer 7. Lactones with the general structure of
7 were either produced in a 1:1 ratio with 8 or they were
the minor component in the final product mixture.
F IGURE 2. Baeyer-Villiger oxidations of 2-alkyl-substituted
cyclohexanones 3a -e. Enantioselectivity values (E) were
determined by computer fitting of GC data obtained from the
complete time course of each reaction to an equation derived
from the theoretical expression relating the enantiomeric
excess of residual substrate (eeS) with the fractional conversion
(c): c ) ([{1 + eeS}E]/[1 - eeS])1/(1-E). The bar colors indicate
the identities of the organisms from which the Baeyer-Villiger
monooxygenases were cloned (same as in Figure 1). Where a
bar is absent, no reaction occurred and only starting material
was observed by GC. The asterix next to 3d indicates that the
apparent selectivity change is due to a change in substituent
priority numbers.
Having access to pairs of enzymes that selectively
deliver either product enantiomer is a key goal in
biocatalysis. The oxidation of 6e by the ChnB1 and
ChnB2 enzymes from Brevibacterium sp. provides an
excellent example of this phenomenon. Both conversions
are highly enantio- and regioselective, affording (R)- or
(S)-8e, respectively. Unfortunately, this behavior was not
general and these two enzymes showed relatively poor
enantioselectivity toward the other 3-substituted cyclo-
hexanones with the exception of the ChnB1-mediated
oxidation of 6a , which favored the (S)-ketone with an E
value of 60. The only other oxidation that combined high
regio- and enantioselectivity involved n-propyl-substi-
enantiomeric discrimination toward this substrate (Fig-
ure 2). In fact, most of the newly cloned Baeyer-Villiger
monooxygenases oxidized 2-alkyl-substituted cyclohex-
anones with lower stereoselectivities than the Acineto-
bacter sp. NCIB 9871 enzyme. This monooxygenase and
that from Rhodococcus SC1 are the most useful for
synthetic purposes since together they allow both the (R)-
ketones and (S)-lactones to be prepared in very high
enantiomeric purities from all of the 2-substituted cyclo-
hexanones examined here.
(45) Percent regioisomeric excess (% re) is defined as the percentage
of the major regioisomer (both enantiomers) minus the percentage of
the minor regioisomer (both enantiomers).
J . Org. Chem, Vol. 69, No. 1, 2004 15

Kyte et al.
F IGURE 3. Baeyer-Villiger oxidations of 3-alkyl-substituted cyclohexanones 6a -e. The bar colors indicate the identities of the
organisms from which the monooxygenases were cloned (same as in Figure 1). Where a bar is absent, no reaction occurred and
only starting material was observed by GC. (A) Enantioselectivity values for reactions of 6a were determined as described previously
(see legend to Figure 2). For other cases, the same equation was applied to data obtained from a single time point during the
reaction time course after derivatization of the residual ketone. (B) The regioselectivities of Baeyer-Villiger oxidations of 6a -e
were determined by allowing reactions to proceed to completion and then subtracting the percentage of the minor regioisomer (7
+ ent-7) from the percentage of the major regioisomer (8 + ent-8). This value, defined as the regioisomeric excess (% re), is depicted
for each enzyme-substrate pair. The asterix next to 6d indicates that the apparent selectivity change is due to a change in
substituent priority numbers.
available lactones in this series, just under half are novel
building blocks made available by one or more of the
newly cloned Baeyer-Villiger monooxygenases.
Baeyer-Villiger oxidations of 4-alkyl-substituted cy-
clohexanones are straightforward reactions that can only
afford single lactone products. We were unable to identify
monooxygenases that provided (R)-lactones from ketones
1a -e with acceptable optical purities, although we were
able to produce (S)-2a -e with uniformly high % ee values
using biocatalysts identified in this study. Discovering
(R)-selective Baeyer-Villiger monooxygenases remains
an important goal that will require a further expanded
set of enzymes.
F IGURE 4. Summary of substituted ꢀ-caprolactones available
The stereoselective oxidation of 2-alkyl-substituted
in homochiral form by enzymatic Baeyer-Villiger oxidations.
cyclohexanones is now an essentially solved problem for
Compounds depicted with a solid green box can be produced
the substrates investigated here. The intrinsic bias
in g98% ee directly by Baeyer-Villiger oxidations using at
toward migration of the more-substituted carbon ensures
high regioselectivity and the use of either the Rhodococ-
cus SC1 (for 3a ) or Acinetobacter sp. NCIB 9871 mo-
nooxygenases (for 3b-e) ensures efficient kinetic reso-
lutions that yield homochiral lactones. The antipodes can
be prepared by subsequent chemical oxidation of the
recovered optically pure ketone, which occurs with reten-
tion of stereochemistry.
least one of the monooxygenases investigated here. Those
shown with a white/green striped pattern are available by an
initial kinetic resolution with one monooxygenase followed by
a peracid-mediated Baeyer-Villiger oxidation or a second
enzymatic oxidation. Plus signs denote ꢀ-caprolactones newly
available in homochiral form using one of the eight novel
monooxygenases explored here. Compounds depicted with red
boxes are not yet available in homochiral form using Baeyer-
Villiger monooxygenases. The asterix next to the entries for
R ) i-Pr/allyl indicates that the apparent selectivity change
is due to a change in substituent priority numbers.
The newly cloned Baeyer-Villiger monooxygenases
contribute most significantly to stereoselective oxidations
of 3-alkyl-substituted cyclohexanones. These are also the
most challenging class of substrates for stereoselective
Baeyer-Villiger reactions. In some cases, e.g., 6c and 6e,
single Baeyer-Villiger monooxygenases display both
high enantio- and regioselectivities, thereby directly
providing homochiral lactones from the corresponding
racemic ketones. In other cases, judicious combinations
of enzymes with differing selectivities could be employed
sequentially to yield homochiral products. For example,
(S)-8b could be prepared by oxidizing racemic 6b with
the Acinetobacter SE19 enzyme until 50% completion (to
remove the undesired antipode), and then subjecting the
isolated (S)-6b to a regioselective Baeyer-Villiger reac-
tion catalyzed by the Brevibacterium sp. ChnB2 enzyme.
tuted 6c, which was converted cleanly by the Arthro-
bacter BP2 monooxygenase to (R)-8c.
Con clu sion
Our long-term goal is to identify one or more Baeyer-
Villiger monooxygenases capable of stereo- and regiose-
lectively oxidizing all of the substituted cyclohexanones
studied here to afford all possible alkyl-substituted
ꢀ-caprolactones in homochiral form. Our results are
summarized in Figure 4, which indicates the spectrum
of substituted ꢀ-caprolactones that can be produced in
g98% ee using either a single enzymatic oxidation or a
chemoenzymatic or two-enzyme combination. Of the 21
16 J . Org. Chem., Vol. 69, No. 1, 2004

Novel Baeyer-Villiger Monooxygenases
(to solubilize the ketone), they were added to a concentration
of 10 mM at this time. The appropriate ketone substrate was
also added to a final concentration of 10 mM. Baeyer-Villiger
oxidations were carried out by shaking the cultures at 150 rpm
at room temperature. Samples for GC analysis were prepared
by taking 100 µL aliquots of the reaction mixture and extract-
ing with 900 µL of EtOAc that also contained 0.01% methyl
benzoate as an internal standard. A 1.0 µL portion of the
organic phase was used for GC analysis.
In reactions of 4-alkyl-substituted cyclohexanones 1a -e,
conversions were allowed to proceed to completion (as deter-
mined by normal-phase GC analysis); then, the optical purities
of the lactones were determined by chiral-phase GC measure-
ments.
Sequences leading to (R)-8a and (S)-8c could be devised
along similar lines.
Our focus in this study was to characterize the sub-
strate selectivities and enantioselectivities of newly
discovered enzymes, rather than on isolating each prod-
uct. Previous efforts using either growing or nongrowing
cells of an E. coli strain overexpressing Acinetobacter sp.
NCIB 9871 cyclohexanone monooxygenase yielded prod-
uct titers of 1-246 and 11 g/L,22,47 respectively. Analogous
applications of the strains investigated here are likely
to be equally straightforward. The use of these novel
Baeyer-Villiger monooxygenases in asymmetric synthe-
sis will be reported in due course.
Baeyer-Villiger oxidations of 2-alkyl-substituted cyclohexa-
nones 3a -e were initially carried out using normal-phase GC
monitoring to determine whether a given enzyme showed
activity toward that substrate and to establish an approximate
time course for the conversion. The bioconversions were then
repeated in order to determine E values accurately. At least
five samples were taken between 0 and 50% fractional conver-
sion (as determined by normal-phase GC analysis), and the
enantiomeric purities of both the remaining ketone and the
lactone product were determined by chiral-phase GC analyses.
These data were fit directly to an analytical expression derived
from Sih’s theoretical analysis of kinetic resolutions.40,41 The
fits were inspected visually to ensure that no systematic
deviations occurred (which would indicate that secondary
processes such as enantioselective degradation had taken place
during the bioconversion). No deviations were observed.
Baeyer-Villiger oxidations of 3-alkyl-substituted cyclohex-
anones 6a -e were first monitored by normal-phase GC to
determine whether a substrate was accepted by a given
monooxygenase. These bioconversions were continued until all
of the ketone substrate had been consumed to provide data
for calculating the regioisomeric excess values.45 All successful
reactions were then repeated to determine the corresponding
E values. The enantiomers of ketone 6a could be separated
cleanly by chiral-phase GC, and E values for this substrate
were determined as described in the previous paragraph. The
enantiomers of 6b-e could not be separated under our
conditions for chiral-phase GC analysis. These reactions were
therefore run to 35-50% completion (as determined by normal-
phase GC analysis), and then remaining substrate and lactone
product(s) were removed by extraction with ethyl acetate (3
× 100 mL). After concentration by rotary evaporation, the
residue was mixed with 1.5 equiv (relative to starting ketone
concentration) of (2R,3R)-butanediol (1.5 mmol, 140 mg), a
catalytic amount of p-toluenesulfonic acid, and 25 mL of
benzene. The reaction mixture was held at reflux for 2 h using
a Dean-Stark trap; then, it was cooled to room temperature,
and an aliquot was analyzed by chiral-phase GC. The enantio-
meric excess value determined for the residual ketone sub-
strate, along with the fractional conversion value (assessed
as described above by normal-phase GC), was used to calculate
the E value.40
Exp er im en ta l Section
Standards of racemic lactones were prepared by oxidation
of ketones with m-CPBA as reported previously.35 Biotrans-
formations were monitored by GC using DB-17 and Chirasil-
Dex CB columns for nonchiral and chiral separations, respec-
tively. The absolute configurations of products and the order
of elution during chiral-phase GC analysis were determined
by comparisons with authentic standards.35 All of the strains
used in these studies can be obtained by contacting Drs. Pierre
Rouvie`re and Qiong Cheng at DuPont, Inc.
Gen er a l P r oced u r e for P r ep a r in g Keton es 1e, 3b, a n d
3e. Pyridinium chlorochromate (67 mmol, 14.4 g) and meth-
ylene chloride (100 mL) were added to a 300 mL round-bottom
flask. Neat alcohol (45 mmol) was added dropwise at room
temperature. After TLC analysis (4:1 hexanes/ethyl ether)
indicated that the reactions were complete (ca. 3 h), the
mixtures were filtered through Celite 545 and washed with
saturated CuSO4 (4 × 100 mL). After drying with Na2SO4, the
solvent was removed by rotary evaporation. No further puri-
fication was required.
Gen er a l P r oced u r e for P r ep a r a tion of 3-Alk yl-Su bsti-
tu ted Cycloh exa n on es 6b-e. Enone intermediates were
prepared from 1,3-cyclohexanedione according to literature
procedures.37,38 The appropriate Grignard reagent (45 mmol)
was added to a solution of 3-isobutoxy-2-cyclohexen-1-one (30
mmol, 5.0 g) in 50 mL of anhydrous diethyl ether at room
temperature over 30 min. After GC analysis indicated that the
reaction had reached completion, the reaction was quenched
by slowly adding 100 mL of 10% H2SO4. The aqueous layer
was extracted with ethyl acetate (3 × 100 mL), and then the
combined organic layers were washed with brine (200 mL),
dried with anhydrous Na2SO4, and concentrated by rotary
evaporation. The enones were purified by vacuum distillation.
Saturated ketones were prepared by catalytic hydrogenation
over palladium on carbon as described previously.47 No further
purification was required.
Gen er a l P r oced u r e for Ba eyer -Villiger Oxid a tion s
Usin g Wh ole Cells of En gin eer ed E. coli Str a in s. Over-
night cultures were grown by adding a single colony of the
appropriate strain to 10 mL of LB medium supplemented with
100 µg/mL ampicillin. After the cultures had shaken overnight
at 37 °C, a 1 mL aliquot was added to a 100 mL portion of LB
medium supplemented with 200 µg/mL of ampicillin (except
for the strain expressing the Rhodococcus SC1 enzyme, which
required 100 µg/mL ampicillin) in a 250 mL Erlenmeyer flask.
The cultures were shaken at 150-200 rpm at 37 °C until they
reached an OD600 between 1.8 and 2.0; then, isopropyl-â-D-
galactoside (IPTG) was added to a final concentration of 0.10
mM. The cultures were shaken at 150 rpm for an additional
30 min at room temperature. If cyclodextrins were required
Ack n ow led gm en t. We thank Cerestar, Inc., for
their generous donation of cyclodextrins. Financial
support by the National Science Foundation (CHE-
0130315) is also gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Detailed experi-
mental procedures for all compounds synthesized in this study,
GC analysis conditions and chromatograms for all racemic
ketones and lactones with elution patterns indicated, proce-
dures for strain storage, and a description of the methods used
to construct an overexpression plasmid for the Rhodococcus
SC1 cyclododecanone monooxygenase. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
(46) Mihovilovic, M. D.; Chen, G.; Wang, S.; Kyte, B.; Rochon, F.;
Kayser, M. D.; Stewart, J . D. J . Org. Chem. 2001, 66, 733-738.
(47) Walton, A. Z.; Stewart, J . D. Biotechnol. Prog. 2003, in press.
J O030253L
J . Org. Chem, Vol. 69, No. 1, 2004 17