On the resolution of chiral substrates by a retro-Claisenase enzyme: Biotransformations of heteroannular bicyclic β-diketones by 6-oxocamphor hydrolase

Article abstract of DOI:10.1002/adsc.200700046

The enzyme 6-oxocamphor hydrolase (OCH) from Rhodococcus sp. NCIMB 9784 catalyses the cleavage of a carbon-carbon bond between two carbonyl groups in both mono- and bicyclic non-enolisable β-diketone substrates. In this mode OCH has been shown to effect the desymmetrisation of both bridged symmetrical bicyclic [2.2.1] and [2.2.2] systems and a series of 1-alkylbicyclo[3.3.0] octane-2,8-diones, yielding chiral substituted cyclopentanone and cyclohexanone products in high optical purity. In the present study, OCH has been challenged with a series of heteroannular substrates including 1-methylbicyclo[4.3.0] nonane-2,9-dione (7a-methylhexa-hydroindene-1,7-dione) in an effort to assess the competence of the enzyme for kinetic resolutions of asymmetric, racemic substrates. OCH was shown to catalyse the resolution of 1-methylbicyclo-[4.3.0] nonane-2,9-dione with an E value of 2.9. The effect of increasing the length of the alkyl chain in the 1-position, or enlarging one of the rings, was to increase the enantioselectivity of the enzyme to 5.7 and 3.1 for the substrates 1-allylbicyclo-[4.3.0]nonane-2,9-dione (7a-allylhexahydroindene-1,7-dione) and 1-methylbicyclo[5.3.0]decane-2,10-dione (8a-methyloctahydroazulene-1,8-dione), respectively. 1-Methylbicyclo[5.4.0]undecane-2,10-dione (9a- methyloctahydrobenzocycloheptene-1,9-dione) was not a substrate for OCH. These experiments constitute the first description of the resolution behaviour of such a retro-Claisenase enzyme, and suggest a maximum steric limit for substrate recognition by OCH.

Full text of DOI:10.1002/adsc.200700046

FULL PAPERS
DOI: 10.1002/adsc.200700046
On the Resolution of Chiral Substrates by a retro-Claisenase
Enzyme: Biotransformations of Heteroannular Bicyclic
b-Diketones by 6-Oxocamphor Hydrolase
Cheryl L. Hill,a Lee Chiang Hung,b Derek J. Smith,b Chandra S. Verma,b
and Gideon Grogana,*
a
York Structural Biology Laboratory, Department ofChemistry, University ofYork, Heslington, York, YO10 5YW, U.K.
Fax : (+44)-1904-328-266; e-mail: grogan@ysbl.york.ac.uk
Bioinformatics Institute, 30 Biopolis Street, # 07-01, Matrix, Singapore 138671
b
Received: January 24, 2007
Abstract: The enzyme 6-oxocamphor hydrolase effect of increasing the length of the alkyl chain in
(OCH) from Rhodococcus sp. NCIMB 9784 catalyses the 1-position, or enlarging one ofthe rings, was to
the cleavage ofa carbon-carbon bond between two
carbonyl groups in both mono- and bicyclic non-eno- and 3.1 for the substrates 1-allylbicyclo-
lisable b-diketone substrates. In this mode OCH has [4.3.0]nonane-2,9-dione (7a-allylhexahydroindene-
increase the enantioselectivity ofthe enzyme to 5.7
G
been shown to effect the desymmetrisation of both 1,7-dione) and 1-methylbicyclo
bridged symmetrical bicyclic [2.2.1] and [2.2.2] sys- dione (8a-methyloctahydroazulene-1,8-dione), re-
tems and a series of1-alkylbicyclo [3.3.0]octane-2,8- spectively. 1-Methylbicyclo[5.4.0]undecane-2,10-
diones, yielding chiral substituted cyclopentanone dione (9a-methyloctahydrobenzocycloheptene-1,9-
A
A
ACHTREUNG
and cyclohexanone products in high optical purity. In dione) was not a substrate for OCH. These experi-
the present study, OCH has been challenged with a ments constitute the first description of the resolu-
series ofheteroannular substrates including 1-
tion behaviour ofsuch a retro-Claisenase enzyme,
methylbicyclo[4.3.0]nonane-2,9-dione (7a-methylhexa- and suggest a maximum steric limit for substrate rec-
ACHTREUNG
hydroindene-1,7-dione) in an effort to assess the ognition by OCH.
competence ofthe enzyme of r kinetic resolutions of
asymmetric, racemic substrates. OCH was shown to Keywords: biotransformations; chemoenzymatic syn-
catalyse
the
resolution
of1-methylbicyclo-
thesis; b-diketones; enzyme catalysis; enzymes;
A
lyases
Introduction
quantitative yields ofoptically enriched keto acid
products (4 and 6) from these biotransformations. We
have also obtained an X-ray crystal structure ofboth
We have recently been evaluating the potential ofthe
C C bond cleaving 6-oxocamphor hydrolase, involved native OCH[3] and a low activity mutant, in which one
À
[4]
in the metabolism of(1 R)-(+)-camphor by Rhodococ- ofthe natural products ofreaction was observed.
cus sp. NCIMB 9874, for the preparative desymmetri- The latter structure has helped to shed light on the
sation ofbicyclic b-diketones.[1] The transformation molecular determinants ofboth mechanism and enan-
catalysed by OCH in nature is the cleavage, by a tiotopic selectivity in reactions catalysed by OCH. At
retro-Claisen reaction, ofthe bicyclic diketone 6-oxo-
this stage, we have proposed a mechanism for the ac-
camphor (1), to a-campholinic acid (2), a reaction tivity ofOCH that is initiated by general base activa-
that proceeds with high diastereoselectivity and enan- tion ofa water molecule by histidine 145. The hydrox-
tioselectivity in favour of the (2R,4S)-enantiomer of ide nucleophile was postulated to attack the carbonyl
the product (Figure 1). This represented an unusual group on the pro-(S) topos ofthe substrate only, as
enzymatic desymmetrisation reaction, and we have the carbonyl group on the other topos is held by an
since applied the enzyme to the desymmetrisation of oxyanion hole formed by the side chains of histidine
a series ofsymmetrical bicyclic b-diketones ofthe
122 and tryptophan 40. A tetrahedral oxyanion is
[2.2.1],[1] [2.2.2][1] (3) and [3.3.0][2] (5) series, and dem- formed, that rearranges to form an enolate intermedi-
onstrated that, in each case, it is possible to obtain ate that is protonated to give the final product. An
Adv. Synth. Catal. 2007, 349, 1353 – 1360
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1353

Cheryl L. Hill et al.
FULL PAPERS
best opportunities for successful substrate recognition.
A general route to this class ofbicyclic b-diketones
was recently published.[5] In this report we detail stud-
ies into the biotransformation of compounds 9, 10
and 12 by OCH, and describe some characteristics of
the kinetic resolution exhibited by the enzyme.
Results and Discussion
As part ofa continuing study into the wider substrate
specificity of OCH, the decalin analogue substrate 7
was synthesised according to the published method.[5]
Disappointingly, it did not prove possible to perform
alkylations ofsubstrate 7 using the standard tech-
niques, and predictably, 7 itselfwas not a substrate for
OCH, nor were the unsubstituted [4.3.0] (8), [5.3.0]
(11) and [5.4.0] (13) bicyclic diketones, the syntheses
[5]
ofwhich were reported in our previous paper,
that
Figure 1. Desymmetrisations ofprochiral bicyclic b-dike-
tones by OCH.
form the basis of compounds 9, 10, 12 and 14. Sub-
strates 9, 10, 12 and 14 were synthesised by alkylation
ofthe parent bicyclic diketones using the procedures
additional measure of prochiral control is afforded by described in the Experimental Section. Milligram
the nascent cyclopentane ring ofthe product of rming
quantities ofcompounds 9, 10, 12 and 14 were tested
for their suitability as substrates, and gratifyingly, re-
a stacking interaction with phenyalanine 82.[4]
With a view to examining the wider substrate spe- actions analysed by TLC against the appropriate con-
cificity of OCH, we decided to synthesise a series of trols revealed disappearance ofsubstrates 9, 10 and
asymmetric, heteroannular bicyclic b-diketones that 12 over a 16–72 h period. Substrate 14 proved not to
would test the ability ofthe enzyme to catalyse the
be a substrate for OCH.
resolution ofracemic substrates. As the substrate spe-
The substrates under consideration had been syn-
cificity of OCH had proven to be quite narrow, ac- thesised in an attempt to explore the substrate specif-
cepting only mono- or bicyclic diketones that were icity ofOCH with respect to the kinetic resolution of
non-enolisable, either through restrictions imposed by racemates. It is useful at this point to consider the
Bredtꢁs rule (in the case ofbicyclo
dione) or by alkylation at the carbon atom between tion ofthese asymmetric substrates by the enzyme. In
the carbonyl groups (as in the series of1-alkylbicyclo- the case ofheteroannular substrates such as 9, 10, 12
[3.3.0]octane-2,8-diones,[2] it was decided to synthesise and 14, it is conceivable that a) the bond between car-
[2.2.2]octane-2,6- possible stereochemical course ofthe biotransofrma-
ACHTREUNG
compounds based around bicyclo [4.3.0], [5.3.0] and bonyls that may be cleaved may be in either the
[5.4.0] systems (Figure 2), as these would present the larger or smaller ring, or both, and that b) two pairs
of enantiomeric product forms may result from cleav-
age ofeither bond. Hence, the maximum possible
number ofproducts rofm the reaction is eight
(Figure 3). The ideal outcome ofany such biotransfor-
mation ofa heteroannular ketone would be one prod-
uct ofeight thereofre, in which case the enzyme
would display first enantio-selectivity, selecting one
enantiomer of the diketone for transformation;
thence regio-selectivity for the bond to be cleaved.
The diastereomeric mix ofthe final product has previ-
ously been found not to be determined by the
enzyme, as, whilst the attack ofbase-activated water
at one topos is enzyme-controlled, the protonation of
the enolate is not. This is evidenced by the same dia-
stereomeric mixture ofenantiomerically enriched
products being recovered from biotransformations of
both the natural substrate 1[1] and the 1-alkylbicyclo-
Figure 2. Compounds described in the present study.
A
1354
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1353 – 1360

On the Resolution ofChiral Substrates by a retro-Claisenase Enzyme
FULL PAPERS
Figure 3. Possible stereochemical outcomes of retro-Claisen cleavage ofheteroannular bicyclic diketone 9 by OCH.
In the case ofsubstrates 9, 10 and 12 therefore, the at m/z=125 was observed for ester 20 was due to the
racemic keto ester products obtained from abiotic presence ofa seven-membered ring with a methyl
1
methanolysis ofthese compounds were to prove val-
uable in assessing the enantioselectivity ofthe
substituent. H NMR ofthe product mixtures result-
ing from these reactions revealed possible diastereo-
enzyme. First, the product mixtures from the metha- meric mixtures, and these were resolved by chiral GC
nolysis of 9, 10 and 12 revealed that in each case, as into four enantiomeric products. The diastereomeric
may be expected, it was the smaller C5 ring ofeach
ratios for the methyl esters 16, 18 and 20 were 2.7:1,
substrate that had been cleaved. This was revealed by 1.7:1 and 1.1:1, respectively. A chiral GC trace illus-
a fragmentation pattern for ester 16 that clearly trating the resolution of the four enantiomeric forms
showed a major peak at m/z=111 corresponding to a of 16 is shown in Figure 4.
six-membered ring with a methyl substituent. A peak
Figure 4. Separation ofdiastereoisomers and enantiomers ofester 16 on Cyclosil-B chiral GC column run isothermally at
1508C.
Adv. Synth. Catal. 2007, 349, 1353 – 1360
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1355

Cheryl L. Hill et al.
FULL PAPERS
Figure 5. Resolution of1-methylbicyclo
G
&
^
dione 9 by 6-oxocamphor hydrolase. = % conversion;
~
% ee diketone substrate* 9; = % ee ester product 16.
* This value determined by abiotic methanolysis ofrecov-
ered residual ketone (see Experimental Section).
Figure 6. Model ofthe (1 R,6S)-enantiomer ofsubstrate 9 in
the active site ofOCH. Distances ofhydrogen bonding in-
teractions (dashed lines) are shown in Šngstroms, including
the accommodation ofthe carbonyl group ofthe six-mem-
bered ring in the oxyanion hole formed by H121 and W40.
A candidate water molecule for attack at the carbonyl
group ofthe ifve-membered ring is shown at a distance of
4.2 Šngstroms (dashed line) from the relevant carbon atom.
The resolution behaviour ofOCH with substrates
9, 10 and 12 was then tested through the periodic
sampling ofenzymatic reactions, using the protocol
described in the Experimental Section. In each case,
GC analysis ofthe methyl ester products 16, 18 and
20 revealed that the diastereomeric mixtures recov-
ered from both abiotic methanolysis and enzyme cata-
lysed retro-Claisen reaction plus methyl esterification, forms of substrate 9 and used them as virtual ligands
exhibited the same diastereomeric excess. This is con- in energy-minimised CHARMM models ofthe OCH
sonant with previously observed results that suggest active site in the complex (see Experimental Section).
that such molecules are present in the most thermo- In the model with the (1R,6S)-enantiomer bound
dynamically stable form as a consequence of protona- (Figure 6), the carbonyl group ofthe six-membered
tion/deprotonation oftheir enol tautomers. A graph
describing the resolution ofsubstrate 9 is shown in histidine 122 and tryptophan 40 as suggested by the
Figure 5. The enantiomeric excess ofthe residual sub-
previously obtained ligand complex.[4] The 6-mem-
strate increases with time, consistent with OCH dis- bered ring stacks against phenylalanine 82, in a simi-
playing a measure ofenantioselectivity with respect lar fashion to the nascent cyclopentane ring of the cy-
ring was constrained in the oxyanion hole formed by
to 9, yet both enantiomers had been consumed after cloalkanoic acid products derived from the [3.3.0]
[2]
48 h. The enantioselectivity as determined by the E series ofsubstrates. In this arrangement, the methyl
[6]
value, using the method ofRakels and co-workers
group ofthe substrate points toward the entrance to
was only 2.9 (E values were calculated using data ob- the active site marked by the gate residue phenylala-
tained at the time point closest to 50% conversion). nine 79, analogous to the proposed binding ofthe
The biotransformation of 10, in which the 1-methyl symmetrical
1-methylbicyclo[3.3.0]octane-2,8-dione
C
ed at a slower rate than that of 9, yet the enantiose- the attack ofa water molecule activated by histidine
lectivity ofthe reaction increased to 5.7 indicative of
a greater level ofdiscrimination ofthe enzyme ofr
145 at the carbonyl carbon ofthe ifve-membered ring
as previously suggested. Transformation of this fav-
each substrate enantiomer. Substrate 12 was process- oured enantiomer would result in a mixture ofthe
ed much more slowly than 9, and not to completion (2S,3S)-trans- and (2R,3S)-cis-keto acid product enan-
under the conditions described, an E value of3.1 was
calculated, and an enantiomeric excess of84% ofthe
residual diketone at a conversion ofapproximately
75% observed.
tiomers of 15. This outcome would be consistent with
the stereochemical outcome ofthe biotransformations
ofthe bicyclo
A
both modelling studies and optical rotation measure-
In order to gain further insight into the enantiose- ments.[2]
lectivity ofOCH with respect to the chiral substrates,
we have modelled the structures ofboth enantiomeric
The poor E values recorded for the resolutions in-
dicate that the disfavoured enantiomer is also recog-
1356
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1353 – 1360

On the Resolution ofChiral Substrates by a retro-Claisenase Enzyme
FULL PAPERS
Conclusions
In this report, we have further examined the biocata-
lytic properties ofthe retro-Claisenase 6-oxocamphor
hydrolase and have established both some steric
limits ofsubstrate acceptance and described ofr the
first time the catalytic behaviour of this unusual
enzyme with respect to the resolution ofchiral sub-
strates. When challenged with a small series ofhetero-
annular substrates, OCH was shown to display a mea-
sure ofenantioselectivity, that we have attributed,
through the use ofmodelling, to comparatively poor
molecular recognition ofthe disafvoured enantiomer
in the active site ofthe enzyme. The results suggest
that, whilst OCH in its native form would not be a
useful catalyst for the resolution of bicyclic diketone
substrates, enough enantioselectivity exists to suggest
that mutation either by rational or random means
may yield an enzyme capable ofimproved perofr-
mance.
Figure 7. Model ofthe (1 S,6R)-enantiomer ofsubstrate 9 in
the active site ofOCH. The carbonyl group ofthe six-mem-
bered ring has once again been modelled to sit in the oxyan-
ion hole formed by H121 and W40. Hydrogen bonding inter-
actions are indicated by dashed lines and labelled in Šng-
stroms. However, in this instance, both candidate water mol-
ecules are somewhat distant from the carbonyl group of the
five-membered ring, suggesting that this enantiomer would
be the disfavoured substrate in the transformation of the
racemate.
Experimental Section
General Remarks
All solvents were distilled before use. All non-aqueous reac-
tions were carried out under oxygen-free nitrogen. Petrol
refers to the fraction of petroleum ether boiling in the range
of40–60 8C. Flash chromatography was carried out using
Davisil Flash Silica 60, 35–60 micron. Thin layer chromatog-
raphy was carried out on commercially available Merck F254
aluminium-backed silica plates. Proton and other NMR
spectra were recorded on a Jeol EX 400 (400 MHz) instru-
ment. Chemical shifts are quoted in parts per million.
Carbon NMR spectra were assigned using DEPT experi-
ments. Chemical ionisation mass spectra were recorded on a
Fisons Analytical (v6) Autospec spectrometer. Chiral GC
analysis ofboth racemic and enzyme-derived esters 16, 18
and 20 was performed on the same instrument, fitted with
an Agilent Cyclosil-B capillary column (30 m”0.25 mm”
0.25 mm); injector temperature 2508C; detector temperature
3208C; column temperature 1508C isothermal for 16; 1208C
isothermal for 18 and 20.
nised by OCH. In order to explain the recognition of
the (1S,6R)-enantiomer in the active site, it would
need to be accommodated in a significantly different
conformation to allow attack at the carbonyl carbon
ofthe five-membered ring. In a model ofthe (1 S,6R)-
enantiomer in the active site (Figure 7), the methyl
group points towards and pushes away the aspartate
154 that is proposed to form part of a putative dyad
that activates histidine 145 for general base catalysis.
When the binding ofthe carbonyl group ofthe six-
membered ring in the oxyanion hole was retained
throughout the simulation, the geometries on the 5-
membered ring are such that the attack by a water ac-
tivated by histidine 145 would not be as readily under-
gone as for the (1R,6S)-enantiomer, as the now dis-
torted geometry ofany putative Asp154-His145 dyad
will make this event less likely. Hence, although the
absolute configuration of the favoured enantiomer in
these resolutions is not known, a knowledge ofthe
structure ofthe active site in complex with the natural
product[4] in conjunction with the modelling studies
are suggestive ofthe (1 R,6S)-enantiomer as being
that transformed most rapidly.
Preparation of Biocatalyst
A crude cell extract from a recombinant strain of E. coli
BL21 (DE3) expressing the camK gene that encodes OCH
was used as the biocatalyst in all experiments. The prepara-
[2]
tion ofthis crude biocatalyst
and the assay used to deter-
mine specific activity of the enzyme have been described
previously.[7] The specific activity of the enzyme preparation
with respect to the assay substrate, bicyclo[2.2.2]octane-2,6-
A
dione, was 9 UmLÀ1, the cell extract routinelÀy1 having a pro-
tein concentration ofapproximately 5 mgmL
.
Adv. Synth. Catal. 2007, 349, 1353 – 1360
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1357

Cheryl L. Hill et al.
FULL PAPERS
(3H, s, CH3); 13C NMR (CDCl3): d=217.0 (C=O), 209.4
(C=O), 65.7 (CCH3), 45.5 (CH), 41.0 (CH2), 37.4 (CH2),
31.3 (CH2), 25.7 (CH2), 25.2 (CH2), 24.7 (CH2), 21.0 (CH3);
EI-MS: m/z=180 [27%, (M)], 152 (77), 123 (100), 110
(100), 55 (88); HR-MS: m/z=180.1155 [C11H16O2 (M+) re-
quires: 180.1150].
1-Methylbicyclo[4.3.0]nonane-2,9-dione (7a-Methyl-
hexahydroindene-1,7-dione) (9)
G
A solution ofdiketone 8 (280 mg, 1.84 mmol), potassium
carbonate (1.53 g, 11.0 mmol) and iodomethane (1.5 mL,
24.8 mmol) in THF (15 mL) was heated to 508C f or 5 h
under nitrogen. The solution was then cooled and water
(15 mL) added. The aqueous phase was extracted with Et2O
(3”10 mL). Combined organic extracts were washed with
brine (10 mL), dried (MgSO4), filtered and solvent evaporat-
ed under reduced pressure to give the crude product. Purifi-
cation by flash column chromatography on silica (9:1)
petrol/EtOAc), gave diketone 9; yield: 200 mg (61%); RF =
1-Methylbicyclo[5.4.0]undecane-2,11-dione
A
(Octahydro-9a-methyl-9aH-benzo[7]annulene-1,9-
dione) (14)
To a suspension ofNaH (53 mg, 60% suspension in nujol,
1.39 mmol) in THF (1 mL) dione 13 (250 mg, 1.39 mmol),
dissolved in THF (5 mL), was added over a period of5 min
under nitrogen. The mixture was left to stir at room temper-
1
0.71 (1:1 petrol/EtOAc); H NMR (400 MHz; CDCl3): d=
2.52–2.26 (4H, m), 2.23–2.16 (1H, m), 2.12–1.99 (3H,m),
1.87–1.78 (2H, m), 1.67–1.56 (2H, m) and (3H, s, CH3);
13C NMR (CDCl3): d=215.3 (C=O), 208.5 (C=O), 64.2
(CCH3), 47.1 (CH), 39.4 (CH2), 34.2 (CH2), 26.8 (CH2), 23.9
(CH2), 18.1 (CH3), 14.2 (CH2); EI-MS: m/z=166 [55%,
(M)], 138 (25), 123 (54), 110 (100), 55 (63); HR-MS: m/z=
166.0994 [C10H14O2 (M)+ requires: 166.0994].
ature for
a
further 30 min. Iodomethane (0.3 mL,
4.17 mmol) was then added and stirring at room tempera-
ture was continued for 2 h. The reaction mixture was then
poured onto 10% (w/v) aqueous KH2PO4 (25 mL), extract-
ed with EtOAc (3”10 mL), dried (MgSO4) and the solvent
removed under reduced pressure to give the crude product.
Purification by flash column chromatography on silica (1:1
petrol/EtOAc) gave diketone derivative 14; yield: 110 mg
1-Allylbicyclo[4.3.0]nonane-2,9-dione (7a-Allyl-
A
1
hexahydo-7aH-indene-1,7-dione) (10)
(41%); RF =0.81 (1:1 petrol/EtOAc); H NMR (400 MHz;
CDCl3): d=2.77–2.70 (1H, m), 2.58–2.38 (3H, m), 2.06–1.98
(1H, m), 1.92–1.67 (10H, m), 1.36 (3H, s, CH3); 13C NMR
(CDCl3): d=213.7 (C=O), 209.4 (C=O), 67.6 (CCH3), 45.4
(CH), 44.3 (CH2), 39.5 (CH2), 28.4 (CH2), 25.1 (CH2), 25.0
(CH2) 23.7 (CH2), 19.6 (CH3).
To a suspension ofNaH (74 mg, 60% suspension in nujol,
1.97 mmol) in THF (1 mL) dione 8 (300 mg, 1.97 mmol), dis-
solved in THF (5 mL), was added over a period of5 min
under nitrogen. The mixture was left to stir at room temper-
ature for
a further 30 min. Allyl bromide (0.5 mL,
5.91 mmol) was added and stirring at room temperature was
continued for 18 h. The reaction mixture was then poured
onto 10% (w/v) aqueous KH2PO4 (50 mL), extracted with
EtOAc (3”15 mL), dried (MgSO4) and the solvent removed
under reduced pressure to give the crude product. Purifica-
tion by flash column chromatography on silica (1:1 petrol/
EtOAc) gave diketone derivative 10; yield: 208 mg (54%),
3-(2-Methyl-3-oxocyclohexyl)propionic Acid Methyl
Ester (16)
Enzyme-catalysed reaction: A solution ofdiketone 9 (30 mg,
0.18 mmol) in ethanol (2 mL) was added to a stirred solu-
tion ofenzyme (225 U) in 50 m M Tris/HCl buffer (50 mL,
pH 7.1). The reaction was left stirring at room temperature
for 24 h. When TLC analysis confirmed the disappearance
ofsubstrate, the mixture was acidified to pH 3–4 and centri-
fuged at 15000 rpm for 20 min to remove precipitated pro-
tein. The supernatant was poured off and extracted with
EtOAc (5”50 mL), dried (MgSO4) and the solvent removed
under reduced pressure to give the crude product, acid 15;
yield: 30 mg (90%); RF =0.15 (1:1 petrol/EtOAc ).
To a stirred solution of 15 (30 mg, 0.16 mmol) in a 1:1
mixture oftoluene/methanol (3 mL) TMS-diazomethane
(0.2 mL, 0.54 mmol) was added dropwise at 08C. The reac-
tion mixture was then allowed to warm to room temperature
and left to stir for 1 h. Volatiles were removed by evapora-
tion under reduced pressure to give the crude product. Pu-
rification by flash column chromatography on silica (9:1
1
RF =0.84 (1:1 petrol/EtOAc); H NMR (400 MHz; CDCl3):
d=5.54 (1H, ddt, J=3.0, 10.0, 17.0 Hz, CH=CH2), 5.08 (2H,
dd, J=10.0, 17.0 Hz, CH=CH2), 2.92 (1H, hept, J=6.0 Hz,
H-5), 2.44–2.11 (10H, m), 1.78–1.70 (2H, m); 13C NMR
(CDCl3): d=214.0 (C=O), 207.2 (C=O), 133.2 (CH=CH2),
118.8 (CH=CH2), 68.0 (C-1), 42.3 (CH2), 39.8 (CH2), 35.7
(CH2), 34.7 (CH2), 26.5 (CH2), 23.9 (CH2), 23.2 (CH2); EI-
MS: m/z=192 [75%, (M)+], 136 (85), 79 (100); HR-MS: m/
z=192.1148 [C12H16O2 (M)+ requires: 192.1150].
1-Methylbicyclo[5.3.0]decane-2,10-dione (8a-Methyl-
G
octahydroazulene-1,8-dione) (12)
A solution ofdiketone 11 (280 mg, 2.67 mmol), potassium
carbonate (1.53 g, 11.0 mmol) and iodomethane (1.6 mL,
36.0 mmol) in THF (15 mL) was heated to 508C f or 5 h petrol/EtOAc) afforded ester 16; yield: 28 mg (87%); RF =
1
under nitrogen. The solution was then cooled and water
(15 mL) added. The aqueous phase was extracted with Et2O
(3”10 mL). Combined organic extracts were washed with
brine (10 mL), dried (MgSO4), filtered and solvent evaporat-
ed under reduced pressure to give the crude product. Purifi-
cation by flash column chromatography on silica (9:1 petrol/
EtOAc) gave diketone 12; yield: 300 mg (62%); RF =0.7
0.63 (1:1 petrol/EtOAc); H NMR (400 MHz; CDCl3): d=
3.68(3H, s, OCH3), 2.45–2.24 (4H, m), 2.15–1.90 (3H, m),
1.72–1.35 (5H, m), and (3H, d, J 6, CH3); 13C NMR
(CDCl3): d=212.9 (C=O), 173.9 (CO2Me), 51.7 (OCH3),
49.8 (CH), 44.2 (CH), 41.4 (CH2), 31.1 (CH2), 29.9 (CH2),
28.9 (CH2), 25.6 (CH2), 11.9 (CH3); EI-MS: m/z=198 [32%,
(M)], 167 (44), 111 (100), 55 (42); HR-MS: m/z=198.1248
[C11H18O3 (M+) requires: 198.1256].
1
(1:1 petrol/EtOAc); H NMR (400 MHz; CDCl3); d=2.66–
2.60 (1H, m) 2.55–2.46 (1H, m), 2.40–2.28 (2H, m), 2.24–
2.17 (1H, m), 2.10–2.02 (1H, m), 1.95–1.57 (7H, m), 1.34
Methanolysis:
A
solution ofdiketone
9
(50 mg,
0.30 mmol) in methanol (2 mL) was added to a stirred solu-
1358
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1353 – 1360

On the Resolution ofChiral Substrates by a retro-Claisenase Enzyme
FULL PAPERS
tion ofsodium methoxide (18 mg, 0.33 mmol) in methanol
(3 mL). The reaction was left to stir at room temperature
for 1 h and then poured onto water (10 mL), extracted with
EtOAc (3”10 mL), dried (MgSO4) and the solvent removed
under reduced pressure to give the crude product. Purifica-
tion by flash column chromatography on silica (9:1 petrol/
EtOAc) gave ester 16; yield: 40 mg (72%).
aliquot ofreaction mixture was removed. This was ifrst ex-
tracted with ethyl acetate (3”10 mL) in order to remove the
residual ketone. The residual ketone was subjected to meth-
anolysis as above, in order to obtain a methyl ester product
that would be easily analysed by chiral GC. The enantiomer-
ic excess ofthe residual ketones were thus determined rfom
analysis oftheir methyl ester derivatives, as it did not prove
possible to resolve the ketone substrates on the chiral GC
columns available. After the first extraction, the aqueous
phase ofthe reaction was acidiifed as above and again ex-
tracted with ethyl acetate in order to obtain the keto acid
product. The crude product was esterified using TMS-diazo-
methane as above, in order to obtain methyl ester products
that could be analysed by chiral GC. Enantiomeric excess
measurements for methyl esters derived from residual
ketone substrate and keto acid product were then plotted in
Figure 5 to give an overall picture ofthe resolution behav-
iour ofthe enzyme.
3-(2-Allyl-3-oxocyclohexyl)propionic Acid Methyl
Ester (18)
Enzyme-catalysed reaction:
A
solution ofdiketone
10
(30 mg, 0.16) in ethanol (2 mL) was added to a stirred solu-
tion ofenzyme (225U) in 50 m M Tris/HCl buffer (50 mL,
pH 7.1). Reaction, product isolation and esterification were
performed as for 16. Purification by flash column chroma-
tography on silica (9:1 petrol/EtOAc) gave ester 18; yield:
28 mg (85%); RF =0.68 (1:1 petrol/EtOAc); 1H NMR
(400 MHz; CDCl3): d=5.75 (1H, ddt, J=3.0, 10.0, 17.0 Hz,
CH=CH2), 5.03 (2H, dd, J=10.0, 17.0 Hz, CH=CH2), 3.68
(3H, s, OCH3), 2.61–1.55 (14H, m); 13C NMR (CDCl3): d=
222.0 (C=O), 173.8 (CO2Me), 136.0 (CH), 116.4 (CH2), 54.7
(CH) 51.7 (OCH3), 41.2 (CH), 41.0 (CH2), 31.5 (CH2), 31.2
(CH2), 28.4 (CH2), 24.6 (CH2), 15.3 (CH2).
Methanolysis: A solution ofdiketone 10 (50 mg, 0.26) in
methanol (2 mL) was added to a stirred solution ofsodium
methoxide (15 mg, 0.27 mmol) in methanol (3 mL) and the
reaction carried out as for 16 above. Purification by flash
column chromatography on silica (9:1 petrol/EtOAc) gave
racemic ester 18; yield: 39 mg (67%).
Computer Modelling
Extensive modelling ofthe separate enantiomeric ofrms of
substrate 9, including detailed molecular dynamics simula-
tions, were carried out and will be presented elsewhere. The
models illustrated in Figure 6 and Figure 7 were built by
using 6-oxocamphor hydrolase pdb coordinate files 1o8u
(native structure) and 1szo (ligand complex). The structure
ofthe ligand was altered using Quanta (Accelrys Inc, San
Diego), removing any overlapping structural water mole-
cules in the process. The two enantiomeric states ofthe
ligand were thus generated and subjected to further optimi-
sations using CHARMM.[8] This was carried out by adding
hydrogen atoms to the protein, followed by a restrained
energy minimisation to remove any remaining steric clashes.
The dynamics that characterise the two enantiomeric states
were probed by carrying out molecular dynamics simula-
tions lasting 4 ns using NAMD software[9] with the
CHARMM27 force field. For this, the model structures
were solvated to a distance of6 Š rfom all protein atoms in
a rectangular box ofwater, and sodium cations were added
to each box to neutralise the overall negative charge ofthe
structures. Molecular dynamics simulations were performed
with a 2 fs time step at a constant temperature of 300 K and
a constant pressure of1 atmosphere under periodic boun-
dary conditions. The Particle Mesh Ewald (PME) method
was used for electrostatics, and a 12 Š cut-off was used for
van der Waals interactions. The TIP3P water model[10] was
used to model the solvent. The ensemble was minimised for
1500 steps using the conjugate gradient method. This was
followed by a heating protocol where the ensemble temper-
ature is increased from 0 K to 300 K over 5 ps. Afterwards,
the backbone atoms ofthe structure were ifxed, while the
side chains and solvent were allowed to move unrestrained
for a further 30 ps, followed by totally unrestrained equili-
bration for 10 ps. The production run was then performed
for 4 ns, with the SHAKE algorithm[11] implemented to re-
strain all hydrogen motions. Coordinates were saved every
1 ps.
3-(2-Methyl-3-oxocycloheptyl)propionic Acid Methyl
Ester (20)
Enzyme-catalysed reaction:
A
solution ofdiketone
12
(30 mg, 0.17 mmol) in ethanol (2 mL) was added to a stirred
solution ofenzyme (20 mL) in 50 m M Tris/HCl buffer
(50 mL, pH 7.1). Reaction, product isolation and esterifica-
tion were performed as for 16. Purification by flash column
chromatography on silica, (9:1 petrol/EtOAc) gave ester 20;
1
yield:29 mg (91%); RF =0.68 (1:1 petrol/EtOAc); H NMR
(400 MHz; CDCl3): d=3.69(3H, s, OCH3), 2.85–2.79 (1H,
m), 2.62–2.55 (1H, m), 2.38–2.18 (4H, m), 1.75–1.32 (8H,
m), 1.07 (3H, d, J=7 Hz, CH3); 13C NMR (CDCl3): d=
215.3 (C=O), 173.8 (CO2Me), 51.6 (OCH3), 49.4 (CH), 43.2
(CH2), 39.8 (CH), 32.4 (CH2), 32.0 (CH2), 25.7 (CH2), 24.8
(CH2), 24.1 (CH2), 13.1 (CH3); EI-MS: m/z=212 [64%,
(M)], 181 (63), 125 (100), 55 (64); HR-MS: m/z=212.1413
[C12H20O3 (M+) requires: 212.1412].
Methanolysis:
A
solution ofdiketone
12 (50 mg,
0.28 mmol) in methanol (2 mL) was added to a stirred solu-
tion ofsodium methoxide (18 mg, 0.31 mmol) in methanol
(3 mL) and the reaction carried out as for 16 above. Purifi-
cation by flash column chromatography on silica (9:1 petrol/
EtOAc) gave racemic ester 20; yield: 38 mg (64%).
Procedure for Detailed Monitoring of Resolutions of
9, 10 and 12
In each case, a solution ofdiketone (100 mg) in ethanol
(2 mL) was added to a stirred solution ofenzyme (450 U) in
50 mM Tris/HCl buffer (100 mL, pH 7.1). At the intervals
shown in Figure 5 for the biotransformation of 9, a 20 mL
Adv. Synth. Catal. 2007, 349, 1353 – 1360
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1359

Cheryl L. Hill et al.
FULL PAPERS
[5] C. L. Hill, M. McGrath, T. Hunt, G. Grogan, Synlett
2006, 309–311.
Acknowledgements
[6] J. L. Rakels, A. J. Straathof, J. J. Heijnen, Enzym.
We gratefully acknowledge the award of a DTA studentship
to C.L.H. from the Engineering and Physical Sciences Re-
search Council (EPSRC).
Microb. Technol. 1993, 15, 1051–1056.
[7] J. P. Bennett, J. L. Whittingham, A. M. Brzozowski,
P. M. Leonard, G. Grogan. Biochemistry 2007, 46,
137–144.
[8] B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J.
States, S. Swaminathan, M. Karplus, J. Comp. Chem.
1983, 4, 1878–2217.
[9] L. Kalꢂ, R. Skeel, M. Bhandarkar, R. Brunner, A.
Gursoy, N. Krawetz, J. Phillips, A. Shinozaki, K. Vara-
darajan, K. Schulten, J. Comp. Phys. 1999, 151, 283–
312.
References
[1] G. Grogan, J. Graf, A. Jones, S. Parsons, N. J. Turner,
S. L. Flitsch, Angew. Chem. Int. Ed. Engl. 2001, 40,
1111–1114.
[2] C. L. Hill, C. S. Verma, G. Grogan, Adv. Synth. Catal.
2007, 349, 916–924.
[3] J. L. Whittingham, J. P. Turkenburg, C. S. Verma, M. A.
Walsh, G. Grogan, J. Biol. Chem. 2003, 278, 1744–
1750.
[4] P. M. Leonard, G. Grogan, J. Biol. Chem. 2004, 279,
31312–31317.
[10] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W.
Impey, M. L. Klein, J. Chem. Phys. 1983, 79, 926–935.
[11] W. F. van Gunsteren, H. J. C. Berendsen, Mol. Phys.
1977, 34, 1311–1332.
1360
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1353 – 1360