Desymmetrisations of 1-alkylbicyclo[3.3.0]octane-2,8-diones by enzymatic retro-Claisen reaction yield optically enriched 2,3-substituted cyclopentanones

Article abstract of DOI:10.1002/adsc.200600468

A series of 1-alkylbicyclo[3.3.0]octane-2,8-diones was transformed by enzymatic retro-Claisen reaction using recombinant 6-oxocamphor hydrolase (OCH) overexpressed in Escherichia coli, to yield optically active 2,3-substituted cyclopentanones with enantiomeric excesses of up to >95%. Whilst the parent substrate, bicyclo[3.3.0]octane-2,8-dione 12, was transformed only very slowly, derivatives 13, 14, 15, 16 and 30 with alkyl chains of varying length in the 1-position were converted rapidly to optically active products with typically 82% de and up to >95% enantiomeric excess. The results confirm the apparent requirement of OCH for non-enolisable diketone substrates, and offer a potential route to decorated cyclopentanone derivatives of multiple chiral centres. Computer modelling of 1-methylbicyclo-[3.3.0]octane-2,8-dione into the active site of OCH suggested that the bicyclic [3.3.0] series substrates were accommodated in the active site in similar orientation with the natural enzyme substrate, 6-oxo-camphor, and would thus yield the (2S,3S)-product series.

Full text of DOI:10.1002/adsc.200600468

FULL PAPERS
DOI: 10.1002/adsc.200600468
Desymmetrisations of 1-Alkylbicyclo[3.3.0]octane-2,8-diones by
G
Enzymatic Retro-Claisen Reaction Yield Optically Enriched 2,3-
Substituted Cyclopentanones
Cheryl L. Hill,a Chandra S. Verma,b and Gideon Grogana,*
a
York Structural Biology Laboratory, Department of Chemistry, University of York, Heslington, York, YO10 5YW, UK
Fax : (+44)-1904-328-266; e-mail: grogan@ysbl.york.ac.uk
BioinformaticsInstitute, 30 BiopolisStreet, #07-01, Matrix, Singapore 138671
b
Received: September 12, 2006
Abstract: A series of 1-alkylbicyclo[3.3.0]octane-2,8- ketone substrates, and offer a potential route to dec-
G
dioneswastranfsormed by enzymatic retro -Claisen orated cyclopentanone derivativesof multiple chiral
reaction using recombinant 6-oxocamphor hydrolase centres. Computer modelling of 1-methylbicyclo-
(OCH) overexpressed in Escherichia coli, to yield
A
optically active 2,3-substituted cyclopentanones with suggested that the bicyclic [3.3.0] series substrates
enantiomeric excesses of up to >95%. Whilst the were accommodated in the active site in similar ori-
parent substrate, bicyclo
A
was transformed only very slowly, derivatives 13, 14, camphor, and would thusyield the (2 S,3S)-product
15, 16 and 30 with alkyl chainsof varying length in
series.
the 1-position were converted rapidly to optically
active productswith typically 82% de and up to Keywords: biotransformations; chemoenzymatic syn-
>95% enantiomeric excess. The results confirm the thesis; b-diketones; enzyme catalysis; enzymes;
apparent requirement of OCH for non-enolisable di- lyases
Introduction
were racemic[4] (Figure 2). When challenged with the
demethylated natural substrate analogues 8 and 10,
The enzymatic desymmetrisation of a prochiral sub- however, the enzyme catalysed the production of pre-
strate can offer an environmentally benign, quantita- dominantly one enantiomer of keto acids 9 and 11. In
tive route to an optically pure product.[1] Whilst these addition to early substrate specificity studies, we also
processes most often employ a carbon-heteroatom purified OCH from the host strain and cloned the
bond hydrolase such as lipase or esterase, to catalyse, gene encoding the enzyme.[5] OCH wasfound to be
for example the selective cleavage of one ester group
of a symmetrical diacetate, the groups of Taschner[2]
and Roberts[3] have reported the enzymatic desym-
metrisations of symmetrical cyclic and bicyclic ke-
tonesuisng enzymatic Baeyer–Villiger reactionsthat
cleave carbon-carbon bonds. In addition, we have re-
À
cently reported an unusual C C bond cleavage reac-
tion in which an enzyme, 6-oxocamphor hydrolase
(OCH) from Rhodococcus sp. NCIMB 9784, catalyses
the desymmetrisation of bicyclic b-diketonesby enzy-
matic retro-Claisen reaction.[4] In nature, OCH cataly-
ses the cleavage of diketone 1 (6-oxocamphor; 2,6-
bornanedione) to yield (2R,4S)-a-campholinic acid 2
and the (2S,4S)-diastereomer 3 in a 6:1 ratio, the
former having an enantiomeric excess of >95%
(Figure 1). Previous studies showed that OCH also ac-
cepted 2,2-alkylated cyclohexane-1,3-diones 4 and 6
Figure 1. Natural biotransformation catalysed by 6-oxo cam-
phor hydrolase as part of the catabolism of (1R)-(+)-cam-
assubstrate,s although the keto acid products
5 and 7 phor by Rhodococcus sp. NCIMB 9784.
916
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 916 – 924

Desymmetrisations of 1-Alkylbicyclo[3.3.0]octane-2,8-diones
A
FULL PAPERS
order to confirm the conservation of prochiral selec-
tivity in the recombinant enzyme, and extended the
substrate specificity to include 1-alkylbicyclo-
A
tuted cyclopentanone derivativeson carbon-carbon
bond cleavage.
Results and Discussion
Overexpression of 6-Oxocamphor Hydrolase
Cloning and sequencing of the gene camK, encoding
OCH, has previously been published, and appears to
form part of a larger operon for camphor metabolism
in Rhodococcus sp. NCIMB 9784.[11] Overexpression
had previously been achieved in Escherichia coli for
the purposes of obtaining protein for crystallisation
and X-ray crystallographic studies.[8] SDS-PAGE anal-
ysis of sequential ammonium sulfate fractions of the
supernatant (Figure 3) obtained after cell disruption
Figure 2. Application of 6-oxocamphor hydrolase to the bio-
transformation of xenobiotic symmetrical b-diketones(ese
ref.[4]).
related to the crotonase superfamily of enzymes, a
group that catalyses a wide range of chemical reac-
tions, each, until the description of OCH, accepting
an acyl-CoA thioester as substrate.[6] The structure of
native OCH[7] and a low kcat pluslow KM mutant
form, His122Ala, which was observed to bind the
minor diasteromeric natural product[8] were deter-
mined and suggested that whilst the overall fold was
typical of some crotonase superfamily members, the
mechanism of action would perhaps be somewhat dif-
ferent. Enzyme kineticsexperiment,s in conjunction
with structural studies, suggested that carbon-carbon
bond cleavage wasachieved by nucleophilic attack of
a water molecule that had been activated by histidine
145 at the carbonyl of the pro-(S) toposof the usb-
strate diketone. General base activation of water had
already been proposed as the initiator for the mecha-
nism of cleavage of a b-diketone in human fumaryl-
acetoacetate hydrolase (FAAH), an enzyme involved
Figure 3. SDS-PAGE of sequential ammonium sulfate cuts
of the cell extract of Escherichia coli BL21(DE3) expressing
A
camK, the gene encoding OCH. Lane 1: Low molecular
weight markers(BioRad); Lane 2: empty; Lane 3: 20%
(NH4)2SO4 cut; Lane 4: 40% (NH4)2SO4 cut; Lane 5: 60%
(NH4)2SO4 cut; Lane 6: 80% (NH4)2SO4 cut.
in mammalian tyrosine metabolism,[9] but in that case, from expression of camK in pET16b(+) in E. coli
a metal ion was also required for catalysis. The mech- BL21(DE3) shows that the gene is very highly ex-
anisms of cleavage of carbon-carbon bonds in b-dike- pressed, and thus provides an excellent source of
AHCTREUNG
[10]
tonesare emerging asintriguingly diveres,
but enzyme for biocatalysis also. The enzyme was typical-
OCH is still the only one whose abilities of prochiral ly used in the form of the crude extract, as further pu-
discrimination, allied with its cofactor independent rification was time consuming and showed to be un-
activity appear to suggest that it may be a useful necessary in the context of interfering activities, of
enzyme for applied biocatalysis. In this report, we which none were detected. The crude extract yielded
have overexpressed the camK gene encoding OCH in a specific enzymatic activity of 9.0 Units per milli-
Escherichia coli and used the enzyme to catalyse the gram, using an assay based on the bicyclic diketone
[5]
desymmetrisation of natural substrate analogue 10, in 10 assubstrate,
where 1 Unit wasequal to the trans-
Adv. Synth. Catal. 2007, 349, 916 – 924
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
917

Cheryl L. Hill et al.
FULL PAPERS
formation of 1 mmol of diketone per minute per milli- Biotransformation of Bicyclo[3.3.0]octane-2,8-diones
E
R
T
G
H
C
A
N
gram of protein.
12 by Recombinant OCH
The transformation of a series of bicyclo[3.3.0] sub-
G
Biotransformation of Bicyclo
10 by Recombinant OCH
A
strates 12–16 was then assessed. A scheme for the bio-
transformation of substrates and preparation of deriv-
atives and racemic standards is shown in Figure 4.
In an effort to establish that the prochiral selectivity 10 mM substrate, dissolved in 2 mL of ethanol was
of OCH had been conserved in the recombinant added to enzyme (225 U) in a total volume of 50 mL
form, a preparative desymmetrisation of 10 wascar-
of Tris/HCl buffer pH 7.1. Reactions were monitored
ried out. After TLC had confirmed that all the sub- by TLC until the substrate spot had disappeared, and
strate had disappeared, the product was esterified the reaction mixture acidified to pH 3.0–4.0 before
using TMS-diazomethane, and a diastereomeric acetal extraction into ethyl acetate. The acid wasfirts con-
formed using (2R,3R)-butanediol, asprevioulsy re-
verted to the methyl ester, then its equivalent acetal
ported.[4] 13C NMR of the acetal of the methyl ester, with (2R,3R)-butanediol as previously described. The
compared with the acetal of the equivalent racemic diastereomeric and enantiomeric excess of derivatised
ester, confirmed that the (S)-acid had been obtained products were then assessed using standard capillary
with an enantiomeric excess of >95%, and that the GC, against standard racemic acetals, that had been
prochiral selectivity of recombinant OCH was the derived from methyl esters themselves derived from
same as the wild-type enzyme.
abiotic methanolysis of the substrate diketones. As an
example, a comparison of GC traces obtained for the
enzyme-derived acetal of 1-ethylbicyclo[3.3.0]octane-
G
Figure 4. Enzymatic desymmetrisations of 1-alkylbicyclo[3.3.0]octane-2,8-diones: Preparation of racemic standards and deriv-
G
atives.
918
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 916 – 924

Desymmetrisations of 1-Alkylbicyclo[3.3.0]octane-2,8-diones
A
FULL PAPERS
2,8-dione 14 and itsequivalent racemate ishsown in
Figure 5.
for the enantiotopic differentiation of equivalent pro-
chiral substrates by abiotic desymmetrisation using
The results obtained for the desymmetrisation of stoichiometric amounts of (À)-ephedrine asnucleo-
compounds 12–16 are shown in Table 1. The parent phile. However, the keto acid productsfrom the
transformation of methyl (13) and prop-2-ynyl (16)
substituted substrates exhibited only 8% ee and 48%
Table 1. Desymmetrisation of 1-alkylbicyclo[3.3.0]octane-
2,8-dionesby 6-oxo camphor hydrolase.
A
ee, respectively.[12]
The biotransformation of 1-(pent-2-ynyl)bicyclo-
[3.3.0]octane-2,8-dione 30 wasperformed asprecursor
Substrate Product Diastereomeric
excess (%)
Enantiomeric
excess (%)
G
to obtaining the natural product (2S,3R)-oxo-2-(2Z-
pentenyl)cyclopent-1-yl-propionic acid 29, or itsdia-
stereomer, a jasmonate analogue previously isolated
12
1317
14
15
16
-
-
-
82
81
86
78
>95
>95
>95
91
from
a
culture of Botryodiploida theobromae[13]
18
19
20
(Figure 6). The biotransformation proceeded smooth-
ly, although a significant if not equal level of hydroly-
sis was also observed in the absence of enzyme. Hy-
drogenation of the alkyne product 31 using Lindlar
compound of thiesrsi,es bicyclo
[3.3.0]octane-2,8- catalyst unfortunately resulted not in the Z-alkene,
dione, is transformed only very slowly, with TLC but in the saturated compound 32, which displayed an
showing only a faint accumulation of a product optical rotation [a]D: +29.48. Chiral GC of the
against buffered controls over a 48-h period. Howev- enzyme-derived methyl ester 33 against the equiva-
er, the methylated derivative 13 wastransformed rap-
lent racemic compound derived by abiotic methanoly-
idly, to yield an acid, the acetal/ester derivative of sis revealed an enantiomeric excess of 75% (de
which exhibited an enantiomeric excess of >95% and 75%), which islower than that obtained for the other
a diastereomeric excess of 82%. It was notable that products, having been compromised by the inherent
the de of the enzyme-derived acetal 25 wasthe same
asthat obtained for the acetal ester acquired by abio-
aqueous instability of the substrate.
In each biotransformation performed in this investi-
tic methanolysis of starting material. The pattern of gation, there istsrong evidence of a highly enantio-
selectivity was repeated for the short series of com- topically selective enzymatic reaction yielding, in each
pounds where the alkyl substituent was ethyl (14), case, products of good to excellent optical purity.
allyl (15) and prop-2-ynyl (16), although in the case of Chiral 2,3-substituted cyclopentanone derivatives
16, the enantiomeric excess of products was sightly form the core of a number of natural productssuch as
lower, perhapsowing to background hydro.liyss
the prostanoids and jasmonate series of plant growth
Düthaler and Maenfisch described in 1985 a method regulators. Homologation of products such as those
Figure 5. GC tracesshowing: top, resolution of racemic acetal derivative 26 on capillary GC; bottom, resolution of optically
active acetal 26 derived from enzymatically acquired product carboxylic acid 18.
Adv. Synth. Catal. 2007, 349, 916 – 924
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
919

Cheryl L. Hill et al.
FULL PAPERS
Figure 7. Model of 1-methylbicyclo[3.3.0]octane-2,8-dione in
G
the active site of OCH, showing possible molecular determi-
nants of prochiral selectivity. His145 is thought to activate a
water molecule for attack at carbonyl group on the pro-(S)-
topos (left as shown) of the substrate. A dashed line marked
3.0 (Šngstroms) indicates the distance from the water to the
carbonyl carbon in the model. The cyclopentanone ring on
the pro-(R)-topos (right as shown) stacks against Phe82. The
carbonyl group on the pro-(R)-toposisbound in an oxyan-
ion hole formed by the side chains of Trp40 and His122. Hy-
drogen bonds are shown with dashed lines, accompanied by
distances (in Šngstroms) in the model.
Figure 6. Biotransformation
of
1-(pent-2-ynyl)bicyclo-
A
carbonyl group on the pro-S-toposof the diketone.
The carbonyl group on the pro-R toposishydrogen
derived from starting material 16 could serve as syn- bonded to both histidine 122 and tryptophan 40,
thonsfor thees or other natural product,s with the which together appear to form a potential oxyanion
crucial elements of stereoselectivity controlled by a hole for the stabilisation of an intermediate enolate in
selective enzymatic process that allows near quantita- the proposed mechanism for OCH. The cyclopenta-
tive yield of product from a symmetrical starting ma- none ring of the product, on the pro-(R)-toposof the
terial.
substrate, is stacked against phenylalanine 82. The
diastereomeric product mixtures that result from
transformation of both the natural substrate and the
Molecular Basis of Prochiral Selectivity – Modelling
series of 1-alkylbicyclo[3.3.0]octane-2,8-dionesstudied
A
1-Methylbicyclo
[3.3.0]octane-2,8-dione into the
here are thought to arise from non-selective protona-
tion of that enolate intermediate and consequent tau-
tomerisation, leading to scrambled stereochemistry at
Active Site of OCH
The structure of the low kcat pluslow KM mutant of the 2-position of the cycloalkanone ring in each case
OCH, His122Ala, has revealed some of the active site – only the stereochemistry at the 3-position appears
mechanistic determinants of prochiral selectivity.[8] to be determined by enzymatic enantiotopic selectivi-
Using this and the native structure of OCH, we have ty. The diastereoisomeric mixture thus appears to be
used the CHARMM program to model the 1- that which is the most thermodynamically stable form
methylbicyclo
active site of native OCH. Although the three-dimen- is known to be most stable for compounds of the
sional structure of 13 isdifferent from that of the nat-
jasmonate series.[14] Whilst the absolute configuration
ural substrate, it appears that many of the interactions of each product hasnot been unambiguoulsy deter-
between substrate and active site may persist in this mined, the series of 1-alkylbicyclo[3.3.0]octane-2,8-
A
AHCTREUNG
putative complex (Figure 7). Histidine 145, mooted to dione substrates is unlikely to invert in the active site
be the residue responsible for general base-catalysed with respect to the known orientation of the natural
activation of the putative nucleophilic water ispois-
substrate and hence the absolute stereochemistry at
tioned near to a candidate water molecule near the the 3-position of the cyclopentanone ring would be
920
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 916 – 924

Desymmetrisations of 1-Alkylbicyclo[3.3.0]octane-2,8-diones
A
FULL PAPERS
(S) asdrawn, with a trans-relationship to the 2-posi- Overexpression and Isolation of Crude Recombinant
tion, which would thusbe ( S), in the thermodynami- 6-Oxocamphor Hydrolase Preparations for
cally favoured diastereomers. Further support is pro- Biotransformations
vided by well-established knowledge of the sign of op-
The gene encoding 6-oxocamphor hydrolase, camK, had
tical rotation of jasmonate derivatives being deter-
mined by the configuration at the 2-position[15]; the
(2S,3R)- or (2S,3S)-configuration in Figure 6 being
dextrorotatory, asobesrved with the natural product
29[13] and carboxylic acid 32 in this study, respectively.
been cloned previously[5] and ligated into the pET16b(+)
plasmid to generate plasmid pGG3.[7] In the present study,
camK wasexpresed in Escherichia coli strain BL21 (DE3).
The plasmid pGG3 was transformed into the bacterium,
which wasgrown on Luria Bertani (LB) agar platescontain-
ing 30 mgmLÀ1 kanamycin, which were grown overnight at
378C. A single colony was used to inoculate 7.5 mL of LB
broth and thissmall culture wasgrown for 18 h at 37 8C with
shaking. This starter culture was then used to inoculate
750 mL of LB broth in a 2-L Erlenmeyer flask. The culture
wasagitated at 150 rpm at 37 8C until the absorbance at
600 nm read 0.5. At thispoint, 1 m M isopropyl thiogalacto-
pyranoside (IPTG) was added, and the culture grown for a
further 3 h at 378C. After thistime, the cellswere harvested
by centrifugation and the resultant cell paste resuspended in
approximately 100 mL 50 mM Tris/HCl buffer pH 7.1 con-
taining 20 mM phenylmethylsulfonyl fluoride (PMSF) and
1 mM dithiothreitol (DTT). The cell suspension was sonicat-
ed for three pulses of 45 s with 1 min intervals at 48C, and
the cell debrisremoved by centrifugation. To the usperna-
tant containing 6-oxocamphor hydrolase was added solid
ammonium sulfate to 60% saturation, and the suspension
stirred for 1 h at 48C. The suspension was centrifuged and
the resultant pellet discarded. To the supernatant was added
solid ammonium sulfate to 80% saturation and the suspen-
sion stirred for 1 h at 48C. The resultant suspension was cen-
trifuged and the pellet obtained was dissolved in a minimum
of cell resuspension buffer. The dissolved protein was then
dialysed against two changes of the same buffer over a
period of 24 h and then snap frozen in liquid nitrogen and
stored at À808C.
Conclusions
In this report, we have demonstrated the use of re-
combinant 6-oxocamphor hydrolase in crude prepara-
tions as a practical biocatalyst for the desymmetrisa-
tion of bicyclic diketones. In addition, we have ex-
panded the substrate specificity of OCH to include 1-
alkylbicyclo[3.3.0]octane-2,8-diones, which, on enzy-
R
matic transformation, yield chiral 2,3-substituted cy-
clopentanoneswith high diatsereomeric and enantio-
meric excess. Whilst the application of this catalyst re-
mains to be optimised in respect of enzyme stability
and other process characteristics, the selectivity of the
enzyme, allied to itsindependence from cofactor,s
render it an unusual and intriguing potential candi-
date for the preparative synthesis of optically active
compounds.
Experimental Section
Bicyclo[2.2.2]octane-2,6-dione 10
A
General Remarks
Bicyclo[2.2.2]octane-2,6-dione 10 wasprepared uisng the
G
procedure described by Almqvist et al.[16] Spectroscopic data
were in accordance with the literature.
All solvents were distilled before use. All non-aqueous reac-
tionswere carried out under oxygen-free nitrogen. Petrol
refersto the fraction of petroleum boiling in the range of
40–608C. Flash chromatography was carried out using Davi-
sil Flash Silica 60, 35–60 micron. Thin layer chromatography
Bicyclo[3.3.0]octane-2,8-dione 12
G
wascarried out on commercially available Merck F
alumi-
Bicyclo[3.3.0]octane-2,8-dione 12 wasprepared for thees
U
254
nium-backed silica plates. Proton and carbon NMR spectra
were recorded on Jeol EX 400 (400 MHz) instrument.
Chemical shifts are quoted in parts per million. Carbon
NMR spectra were assigned using DEPT experiments.
Chemical ionisation mass spectra were recorded on a Fisons
Analytical (v6) Autospec spectrometer. GC analysis of ace-
tals 25, 26, 27 and 28 wasperformed on an Agilent 6890 gas
chromatograph fitted with an HP5 capillary column (30 m”
0.32 mm”0.25 mm): injector temperature 2508C; detector
temperature 3208C; column temperature 1308C for 25, 26
and 28 and 1808C for 27. Chiral GC analysis of both race-
mic and enzyme-derived esters 33 wasperformed 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
1208C for 40 min, then 1508C for 25 min, then 1708C for
60 min.
for thisgeneral classof bicyclic
b-diketones.[18] 1H NMR
(400 MHz, CDCl3): d=3.20 (1H, hept, J=6.0 Hz, H-5), 3.04
(1H, d, J=6.0 Hz, H-1), 2.38–2.20 (6H, m), 1.86–1.76 (2H,
m); 13C NMR (125 MHz; CDCl3): d=207.1 (2C=O), 63.4
(CH-1), 39.4 (CH-5), 37.9 (CH2-3 and CH2-7), 26.2 (CH2-4
and CH2-6); MS (CI; NH3): m/z=138 [100%, (M+NH4)+],
55 (41), 110 (32), 156 (62); found for (M+NH4)+: 156.1023;
C8H10O2 requires: 156.1024.
1-Methylbicyclo[3.3.0]octane-2,8-dione 13
C
1-Methylbicyclo[3.3.0]octane-2,8-dione 13 wasprepared
U
using the method of Düthaler and Maenfisch.[17] 1H NMR
(400 MHz, CDCl3): d=2.71 (1H, hept, J=6.0 Hz, H-5),
2.38–2.20 (6H, m), 1.86–1.76 (2H, m), 1.19 (3H, s, CH3);
Adv. Synth. Catal. 2007, 349, 916 – 924
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
921

Cheryl L. Hill et al.
FULL PAPERS
13C NMR (125 MHz, CDCl3): d=212.8 (2C=O), 46.7 (CH),
36.7 (CH2-3 and CH2-7), 24.6 (CH2-4 and CH2-6), 19.3
(CH3); MS (CI; NH3): m/z=170 [100%, (M+NH4)+], 153
(10); found for (M + NH4)+: 170.1177; C9H12O2 requires:
170.1181.
[92%, (M+NH4)+], 148 (100), 194 (58); found for (M+
NH4)+: 194.1181, C11H12O2 requires: 194.1181].
1-(Pent-2-ynyl)bicyclo[3.3.0]octane-2,8-dione 30
A
Phosphorus tribromide (0.56 mL, 5.94 mmol) was added
dropwise to a stirred, ice-cooled solution of pent-2-yn-1-ol
(1 g, 11.8 mmol) and pyridine (0.1 mL) in ether (22 mL).
After stirring for 22 h at room temperature, the reaction
mixture waspoured on to ice/water (20 mL). The organic
layer wasesparated and the aqueouslayer wasextracted
with ether (3”10 mL). The combined organic extractswere
dried (MgSO4), filtered and evaporated under reduced pres-
sure to afford 1-bromopent-2-yne as a brown oil; yield:
1.52 g (88%). The product wasuesd without further purifi-
cation.
To a suspension of NaH (30 mg, 60% suspension in nujol,
0.725 mmol) in THF (1 mL) dione 12 (100 mg, 0.725 mmol),
dissolved in THF (5 mL), was added in 5 min under nitro-
gen. The mixture wasleft to stir at room temperature for a
further 30 min. 1-Bromopent-2-yne (321 mg, 2.17 mmol) was
added and stirring at room temperature was continued for
18 h. The reaction mixture wasthen poured on to 10%
KH2PO4 (50 mL) and extracted with ethyl acetate (3”
15 mL), dried (MgSO4) and the solvent removed under re-
duced pressure to give the crude product. Purification by
flash column chromatography on silica (1:1 petrol/ethyl ace-
tate) gave diketone 30; yield: 50 mg (34%); Rf (1:1 petrol
ethyl acetate): 0.84; 1H NMR (400 MHz, CDCl3): d=3.13
(1H, hept, J=6.0 Hz, H-5), 2.55 (2H, t, J=2 Hz, CH2-1’),
2.46–2.24 (6H, m) 2.11 (2H, qt, J=2, 8 Hz, CH2-4’), 1.83–
1.74 (2H, m), 1.08 (3H, t, J 8, CH3-5’); 13C NMR (125 MHz,
CDCl3): d=210.9 (2C=O), 84.0 (CH-3’), 74.1 (C-1), 67.9 (C-
2’), 43.7 (CH-5), 37.4 (CH2-3 and CH2-7), 25.1 (CH2-4 and
CH2-6), 23.8 (CH2-1’), 14.0 (CH3), 12.2 (CH2-4); EI-MS:
m/z=205 [20%, (M+H)+], 147 (100) 133 (40); found for
(M+H)+: 205.1228, C13H16O2 requires: 205.1229.
1-Ethylbicyclo[3.3.0]octane-2,8-dione 14
A
To a suspension of NaH (61 mg, 60% suspension in nujol,
1.59 mmol) in THF (1 mL) dione 12 (220 mg, 1.59 mmol),
dissolved in THF (5 mL), was added in 5 min under nitro-
gen. The mixture wasleft to stir at room temperature for a
further 30 min. Iodoethane (0.25 mL, 3.18 mmol) wasadded
and stirring at room temperature was continued for 20 h.
The reaction mixture wasthen poured on to 10% KH PO4
2
(50 mL) and extracted with ethyl acetate (3”15 mL), dried
(MgSO4) and the solvent removed under reduced pressure
to give the crude product. Purification by flash column chro-
matography on silica (1:1 petrol/ethyl acetate) gave diketone
14; yield: 107 mg (40%); Rf (1:1 petrol/ethyl acetate): 0.84;
1H NMR (400 MHz; CDCl3): d=2.91 (1H, hept, J=5.0 Hz,
H-5), 2.43–2.14 (6H, m), 1.76 (2H, q, J=7.5 Hz, CH2CH3),
1.92–1.70 (2H, m), 0.84 (3H, t, J=7.5 Hz, CH2CH3);
13C NMR (125 MHz; CDCl3): d=212.4 (2C=O), 70.4 (C-1),
42.7 (CH-5), 37.6 (CH2-3 and CH2-7), 27.2 (CH2CH3), 25.2
(CH2-4 and CH2-6), 9.6 (CH3); MS (CI; NH3): m/z=167
[8%, (M+H)+], 184 (100); found for (M+NH4)+: 184.1338,
C10H14O2 requires: 184.1338].
1-Allylbicyclo[3.3.0]octane-2,8-dione 15
A
To a suspension of NaH (56 mg, 60% suspension in nujol,
1.45 mmol) in THF (1 mL) dione 12 (200 mg, 1.45 mmol),
dissolved in THF (5 mL), was added in 5 min under nitro-
gen. The mixture wasleft to stir at room temperature for a
further 30 min. Allyl bromide (0.25 mL, 2.90 mmol) was
added and stirring at room temperature was continued for
18 h. The reaction mixture wasthen poured on to 10%
KH2PO4 (50 mL) and extracted with ethyl acetate (3”
15 mL), dried (MgSO4) and the solvent removed under re-
duced pressure to give the crude product. Purification by
flash column chromatography on silica (1:1 petrol/ethyl ace-
tate) gave diketone 15; yield: 140 mg (54%); Rf (1:1 petrol
ethyl acetate): 0.84; 1H NMR (400 MHz, CDCl3): d=5.54
(1H, ddt, J=3.0, 10.0 and 17.0 Hz, CH=CH2), 5.08 (2H, dd,
J 10.0 and 17.0, CH=CH2), 2.92 (1H, hept, J=6.0 Hz, H-5),
2.44–2.11 (8H, m), 1.78–1.70 (2H, m); 13C NMR (125 MHz,
CDCl3): d=211.7 (2C=O), 132.4 (CH=CH2), 119.6 (CH=
CH2), 69.3 (C-1), 42.8 (CH-5), 38.4 (CH2-1’), 37.5 (CH2-3
and CH2-7), 24.8 (CH2CH3), 25.2 (CH2-4 and CH2-6); EI-
MS: m/z=178 [55%, (M)+], 79 (90), 122 (94), 150 (100);
found for (M)+: 178.0993, C11H14O2 requires: 178.0994.
Enzymatic Retro-Claisen Reaction
A solution of diketone (0.05 g) in ethanol (2 mL) was added
to a stirred solution of enzyme (225 U) in buffer (50 mL).
The reaction wastsirred at room temperature for 24 h.
When TLC indicated that all the substrate had been trans-
formed, the reaction mixture wasacidified to pH 3–4 and
centrifuged at 15000 rpm for 20 min in order to remove pre-
cipitated protein. The supernatant was removed and extract-
ed with ethyl acetate (5”50 mL), dried over anhydrous
magnesium sulfate and the solvent removed under reduced
pressure to give the crude product. Crude keto acids 17, 18,
19 and 20 derived in this way were not isolated at this stage,
but were used in this form for methylation.
The acid 31 derived from biotransformation of 30 was
characterised as 3-(2-pent-2-ynyl-3-oxocyclopentyl)-propion-
ic acid (31): Rf (1:1 petrol/ethyl acetate): 0.10; 1H NMR:
(400 MHz, CDCl3): d=2.58–2.34 (5H, m), 2.25–2.05 (6H,
m), 1.86–1.81 (1H, m), 1.79–1.60 (1H, m), 1.45–1.40 (1H,
m), 1.27–1.23 (1H, m) and 1.06 (3H, t, J=7 Hz, CH3);
13C NMR (125 MHz, CDCl3): d=218.5 (C=O), 179.3
(CO2H), 83.6 (C-3’), 75.7 (C-2’), 53.7 (CH), 40.3 (CH), 37.7
(CH2) 31.8 (CH2), 29.2 (CH2), 26.7 (CH2), 17.4 (CH2-1’),
14.0 (CH3), 12.3 (CH2-4); EI-MS: m/z=223 [35%, (M+
1-(Prop-2-ynyl)bicyclo[3.3.0]octane-2,8-dione 16
N
1-(2’-Propynyl)bicyclo[3.3.0]octane-2,8-dione 16 waspre-
G
pared using the method of Düthaler and Maenfisch.[17]
1H NMR (400 MHz, CDCl3): d=3.16 (1H, hept, J=6.0 Hz,
H-5), 2.59 (2H, d, J=2.5 Hz, CH2-1’), 2.47–2.24 (6H, m),
1.98 (1H, t, J=2.5 Hz, CH-3’), 1.87–1.76 (2H, m); 13C NMR
(125 MHz, CDCl3): d=210.5 (2C=O), 79.8 (CH-3’), 70.7 (C-
1), 67.6 (C-2’), 43.7 (CH-5), 37.5 (CH2-3 and CH2-7), 25.2
(CH2-4 and CH2-6), 23.4 (CH2-1’); MS (CI; NH3): m/z=177
922
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 916 – 924

Desymmetrisations of 1-Alkylbicyclo[3.3.0]octane-2,8-diones
A
FULL PAPERS
H)+], 122 (57), 107 (100); found for (M+H)+: 223.1335,
C13H18O3:223.1334].
(1H, m), 1.69–1.59 (1H, m), 1.47–1.37 (1H, m); MS (CI;
NH3): m/z=209 [89%, (M+H)+], 94 (33), 177 (45), 226
(100); found for (M+H)+: 209.1176, C12H16O3 requires:
209.1178].
Acid 31 wasthen hydrogenated and the product 32 char-
acterised, to allow comparison to natural product 29, as 3-
(2-pentyl-3-oxocyclopentyl)-propionic acid (32): Rf (1:1
3-(3-Oxo-2-pentylcyclopentyl)-propionic acid methyl ester
1
1
petrol/ethyl acetate): 0.10; H NMR (400 MHz, CDCl3): d=
(33): Rf (1:1 petrol/ethyl acetate): 0.75; H NMR (400 MHz,
2.54–2.32 (4H, m), 2.24–2.04 (3H, m), 1.92–1.87 (1H, m),
1.77–1.70 (1H, m), 1.59–1.53 (2H, m), 1.43–1.24 (8H, m),
0.87 (3H, t, J=7 Hz, CH3); 13C NMR (125 MHz, CDCl3):
d=220.7 (C=O), 179.3 (CO2H), 54.9 (CH), 40.8 (CH), 37.7
(CH2), 32.1 (CH2), 31.7 (CH2), 29.4 (CH2), 27.9 (CH2), 26.7
(CH2), 26.4 (CH2), 22.4 (CH2), 14.0 (CH3); EI-MS: m/z=
226 [15%, (M)+], 138 (47), 83 (100); [a]D: +29.48.
CDCl3): d=3.67 (3H, s, OMe), 2.50–2.30 (3H, m), 2.17–2.02
(3H, m), 1.90–1.81 (1H, m), 1.75–1.69 (1H, m), 1.58–1.52
(3H, m), 1.42–1.23 (7H, m), 0.87 (3H, t, J=7 Hz, CH3);
13C NMR (125 MHz, CDCl3): d=220.5 (C=O), 173.8
(CO2Me), 54.8 (CH), 51.6 (OMe), 40.8 (CH), 37.8 (CH2),
32.1 (CH2), 31.8 (CH2), 29.7 (CH2), 27.8 (CH2), 26.7 (CH2),
26.4 (CH2), 22.5 (CH2), 14.0 (CH3); EI-MS: m/z=240 [30%,
(M)+], 84 (55), 49 (100); found for (M)+: 240.1725, C14H24O3
requires: 240.1725].
Preparation of Racemic Methyl Ester Product
Standards
Formation of Methyl Esters of Enzyme-Derived Keto
Acids using TMS-Diazomethane
A solution of diketone (0.05 g) in methanol (2 mL) was
added to a stirred solution of sodium methoxide (1.2 equiv)
in methanol (2 mL). The reaction mixture wastsirred at
room temperature for 1 h and then poured on to water
(10 mL), extracted with ethyl acetate (3”10 mL), dried over
anhydrous magnesium sulfate and the solvent removed
under reduced pressure to give the crude methyl ester prod-
uct.
To a stirred solution of the crude carboxylic acid product
(1 equiv.) in a 1:1 mixture of toluene/methanol (3 mL) TMS-
diazomethane (2 equivs.) was added dropwise at 08C. The
reaction mixture wasthen allowed to warm to room temper-
ature and stirred for 1 h. Volatiles were removed by evapo-
ration under reduced pressure to give the crude methyl
ester products 21, 22, 23 and 24 in quantitative yields.
3-(2-Methyl-3-oxocyclopentyl)-propionic acid methyl ester
(21): Yield: 67%; Rf (1:1 petrol/ethyl acetate): 0.55;
1H NMR (400 MHz, CDCl3): d=3.66 (3H, s, OCH3), 2.48–
231 (3H, m), 2.20–2.02 (4H, m), 1.72–1.58 (2H, m), 1.37–
1.32 (1H, m), 1.06 (3H, d, J=7.0 Hz, CH3); 13C NMR
(125 MHz, CDCl3): d=220.5 (C=O), 1.73.9 (CO2Me), 51.8
(OCH3), 50.4 (CH), 44.2 (CH), 37.3 (CH2), 31.9 (CH2) and
29.6 (CH2), 26.9 (CH2), 12.6 (CH3); MS (CI; NH3): m/z=
202 [100%, (M+NH4)+]; found for (M+NH4)+: 202.1439,
C10H16O3 requires: 202.1443.
3-(2-Ethyl-3-oxocyclopentyl)-propionic acid methyl ester
(22): Yield: 69%; Rf (1:1 petrol/ethyl acetate): 0.55;
1H NMR (400 MHz, CDCl3): d=3.70 (3H, s, OCH3), 2.53–
2.30 (3H, m), 2.20–2.00 (3H, m), 1.76–1.25 (6H, m), 0.90
(3H, t, J=7.5 Hz, CH3); 13C NMR (125 MHz, CDCl3): d=
217.8 (C=O), 1.73.9 (CO2Me), 56.0 (CH), 51.8 (OCH3), 40.2
(CH), 38.0 (CH2), 31.9 (CH2) and 29.9 (CH2), 26.8 (CH2),
20.5 (CH2-1’’), 10.9 (CH3); MS (CI; NH3): m/z=199 [4%,
(M+H)+], 181 (11), 216 (100); found for (M+NH4)+:
216.1595, C11H18O3 requires: 216.1560.
Formation of Acetals
A
solution of methyl ester, (2R,3R)-2,3-butanediol
(2 equivs.), tosic acid monohydrate (catalytic) and magnesi-
um sulfate (200 mg) in benzene (7 mL) was refluxed for 5 h.
The reaction was then quenched with saturated sodium bi-
carbonate solution, extracted with ethyl acetate (3”10 mL)
dried over anhydrous magnesium sulfate and the solvent re-
moved under reduced pressure to give the crude product
acetals 25, 26, 27 and 28 in quantitative yields, which were
purified over silica plugs prior to injection in the gas chro-
matograph.
Molecular Modelling
Molecular modelling was used to construct a model for the
docked substrate 13 with the CHARMM program and
force-field.[14] The model for the substrate 13 wasbuilt using
QUANTA (AccelrysInc., San Diego, CA., USA) and the
atom types and charges were assigned using standard
CHARMM parameter sets.[15] The crystal structure PDB ac-
cession code 1o8u was used to set up the initial model of the
3-(2-Allyl-3-oxocyclopentyl)-propionic acid methyl ester
(23): Yield: 73%; Rf (1.1 petrol/acetate): 0.55; 1H NMR
(400 MHz, CDCl3): d=5.69 (1H, ddt, J=3.0, 10.0 and 17.0
Hz, CH=CH2), 5.02 (2H, dd, J=10.0 and 17.0 Hz, CH= protein in its native state. All crystallographic water mole-
CH2) 3.70 (3H, s, OCH3), 2.47–2.30 (4H, m), 2.17–2.00 (4H,
m), 1.91–1.79 (2H, m), 1.59–1.49 (1H, m), 1.43–1.33 (1H,
m); 13C NMR (125 MHz, CDCl3): d=219.6 (C=O), 1.73.9
(CO2Me), 135.3 (CH=), 117.4 (=CH2), 54.6 (CH), 51.8
(OCH3), 40.4 (CH), 37.9 (CH2), 32.2 (CH2-1’’), 31.9 (CH2),
29.6 (CH2), 26.8 (CH2); MS (CI; NH3): m/z=209 [89%,
(M+H)+], 94 (33), 177 (45), 226 (100); found for (M+H)+,
211.1328, C12H18O3 requires: 211.1334.
3-(3-Oxo-2-prop-2-ynylcyclopentyl)-propionic acid methyl
ester (24): Yield: 49%; Rf (1.1 petrol/ethyl acetate): 0.55;
1H NMR (400 MHz, CDCl3): d=3.70 (3H, s, OCH3), 2.62
(1H, ddd, J=2.5, 5.0 and 17.0 Hz, H-1), 2.53–2.36 (4H, m),
2.24–2.06 (4H, m), 1.93 (1H, t, J=2.5 Hz, H-3’’), 1.90–1.84
culeswere retained except for two in the putative active
site. The substrate was initially docked using the stereo-
chemical constraints imposed upon the reaction using the
structure 13. Thisled to a clahs between two water mole-
cules and the substrate and so the water molecules were re-
moved. Of the five water moleculesthat were located in the
active site, two with B values of 32 Š and 35 Š were re-
moved as they clashed with the model substrate. The sub-
strate itself was docked according to information gleaned
from known OCH mechanism and the stereochemical con-
straints of the reaction; the structure of (2S,4S)-a-campho-
linic acid 3 from the crystal structure of the His122Ala
mutant of the enzyme[8] wasuesd asa guide. Hydrogen
Adv. Synth. Catal. 2007, 349, 916 – 924
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
923

Cheryl L. Hill et al.
FULL PAPERS
atomswere built uinsg the HBUILT functionality of
CHARMM.[21] Examination of the energeticsof hydrogen
bond formation by the Hisside chainsled to a model where
two histidines (His45 and His122), both of which form part
of the active site, were protonated. The geometry and ener-
getics of the docked substrates and protein were optimised
by carrying out energy minimisations. These were performed
with atoms within a 10 Š radius around the substrate subject
to decreasing constraints, whereas the atoms beyond this
were restrained with a harmonic force of 10.0 kcalmolÀ1 ŠÀ2.
Minimisations were carried out using the Steepest Descent
and Adopted Basis Newton Raphson algorithms[19] and were
continued until the root mean square magnitude of the
forceswaslower than 10 À3 10.0 kcalmolÀ1 ŠÀ1.
[5] G. Grogan, G. A. Roberts, D. Bougioukou, N. J. Turner,
S. L. Flitsch, J. Biol. Chem. 2001, 275, 12565–12572.
[6] H. M. Holden, M. W. Benning, T. Haller, J. A. Gerlt,
Acc. Chem. Res. 2001, 34, 145–157.
[7] J. L. Whittingham, J. P. Turkenburg, C. S. Verma, M. A.
Walsh, G. Grogan, J. Biol. Chem. 2003, 278, 1744–
1750.
[8] P. M. Leonard, G. Grogan, J. Biol. Chem. 2004, 279,
31312–31317.
[9] D. E. Timm, H. A. Mueller, P. Bhanumoorthy, J. M.
Harp, G. J. Bunick, Structure 1999, 7, 1023–1033.
[10] G. Grogan, Biochem. J. 2005, 388, 721–730.
[11] G. A. Roberts, G. Grogan, N. J. Turner, S. L. Flitsch,
DNA Sequence 2004, 15, 96–103.
[12] R. O. Duthaler, P. Maeinfisch, Helv. Chim. Acta 1984,
67, 845–855.
[13] O. Miersch, J. Schmidt, G. Sembdner, K. Schreiber,
Acknowledgements
Phytochemistry 1989, 28, 1303–1305.
[14] R. A. Creelman, J. E. Mullet, Annu. Rev. Plant. Physi-
ol. Plant Mol. Biol. 1997, 48, 355–381.
[15] L. Holbrook, P. Tung, K. Ward, D. M. Reid, S. Abrams,
N. Lamb, W. Quail, M. M. Moloney, Plant Physiol.
1997, 114, 419–428.
We gratefully acknowledge the award of a DTA studentship
to C. L. H. from the Engineering and Physical Sciences Re-
search Council (EPSRC). G. G. would like to thank the In-
novation and Research Priming Fund, University of York,
for additional funding.
[16] F. Almqvist, L. Eklund, T. Frejd, Synth. Commun.
1993, 23, 1499–1505.
[17] R. O. Duthaler, P. Maienfisch, Helv. Chim. Acta 1984,
67, 856–865.
References
[18] C. L. Hill, M. McGrath, T. Hunt, G. Grogan, Synlett
[1] E. Schoffers, A. Golebiowski, C. R. Johnson, Tetrahe-
dron 1996, 52, 3769–3826.
2006, 309–311.
[19] B. R. Brooks, R. E. Bruccoleri, B. D. Olafson, D. J.
States, S. Swaminathan, M. Karplus, J. Comp. Chem.
1983, 4, 187–217.
[20] F. A. Momany, R. Rone, J. Comp. Chem. 1992, 13,
888–900.
[2] M. J. Taschner, D. J. Black, J. Am. Chem. Soc. 1988,
110, 6892–6893.
[3] R. Gagnon, G. Grogan, E. Groussain, S. Pedragosa-
Moreau, P. F. Richardson, S. M. Roberts, A. J. Willetts,
V. Alphand, J. Lebreton, R. Furstoss, J. Chem. Soc.,
Perkin Trans. 1 1995, 2527–2528.
[4] G. Grogan, J. Graf, A. Jones, S. Parsons, N. J. Turner,
S. L. Flitsch, Angew. Chem. Int. Ed. 2001, 40, 1111–
1114.
[21] A. T. Brunger, M. Karplus, Proteins: Struct. Funct.
Genet. 1988, 4, 148–156.
924
www.asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 916 – 924