Bacterial biotransformation of isoprene and related dienes

Article abstract of DOI:10.1007/s007060070096

The bacterium Pseudomonas putida ML 2 was used in the oxidative biodegradation of the acyclic dienes isoprene, trans-piperylene, cis-piperylene, and 1,3-butadiene. Regioselective dioxygenase-catalyzed dihydroxylation of alkenes yielded vicinal diols in the preferred sequence monosubstituted 〉 cis-disubstituted 〉 gem-disubstituted 〉 trans-disubstituted. The isolated diol metabolites had an excess of the R configuration (9-97% ee), and further diol oxidation was controlled by addition of propylene glycol as an inhibitor. Stereoselectivity using the ML2 strain resulted from both enzymatic asymmetric alkene dihydroxylation and kinetic resolution of diols. Enantioselective oxidation of the allylic secondary alcohol group of R configuration yielded the corresponding unsaturated ketoalcohol; the residual diol was recovered with a large excess (≥ 93% ee) of the S configuration. In addition to the enzymatic diene oxidation steps yielding unsaturated diols and ketoalcohols, evidence was also found of enzymatic alkene hydrogenation to yield saturated ketoalcohols and diols.

Full text of DOI:10.1007/s007060070096

Monatshefte fuÈr Chemie 131, 673±685 (2000)
Bacterial Biotransformation of Isoprene
and Related Dienes
Derek R. Boyd,1;Ã, David Clarke1, Marcel C. Cleij1, John T. G Hamilton2,
and Gary N. Sheldrake1
1
School of Chemistry and QUESTOR Centre, The Queen's University of Belfast, Belfast BT9 5AG,
UK
2
Food Science Division, Department of Agriculture for Northern Ireland, Belfast BT9 5PX, UK
Summary. The bacterium Pseudomonas putida ML 2 was used in the oxidative biodegradation of
the acyclic dienes isoprene, trans-piperylene, cis-piperylene, and 1,3-butadiene. Regioselective
dioxygenase-catalyzed dihydroxylation of alkenes yielded vicinal diols in the preferred sequence
monosubstituted > cis-disubstituted > gem-disubstituted > trans-disubstituted. The isolated diol
metabolites had an excess of the R con®guration (9±97% ee), and further diol oxidation was
controlled by addition of propylene glycol as an inhibitor. Stereoselectivity using the ML2 strain
resulted from both enzymatic asymmetric alkene dihydroxylation and kinetic resolution of diols.
Enantioselective oxidation of the allylic secondary alcohol group of R con®guration yielded the
corresponding unsaturated ketoalcohol; the residual diol was recovered with a large excess (ꢀ 93%
ee) of the S con®guration. In addition to the enzymatic diene oxidation steps yielding unsaturated
diols and ketoalcohols, evidence was also found of enzymatic alkene hydrogenation to yield
saturated ketoalcohols and diols.
Keywords. Asymmetric dihydroxylation; Chiral resolution; Dioxygenases; Enzymes; Isoprene.
Introduction
Methane and isoprene 1 are reported to be present in considerable quantities in the
earth's atmosphere and are the most abundant volatile hydrocarbons in the
environment [1,2]. Biogenic isoprene is produced in large quantities from plants
14
1
(5Â10 g year ), particularly in the tropical rain forests and from the oceans [1±
3]. It has been estimated that isoprene is responsible for ca. 30% of the total non-
methane organic compounds (NMOCs) emitted into the environment in North
America each year [1]. Isoprene can readily be involved in radical reactions under
the in¯uence of UV radiation and will strongly in¯uence atmospheric chemistry.
The isoprene radicals produced in the earth's atmosphere can thus undergo
photochemical reactions with nitrogen oxides (NO, NO2) from combustion sources
resulting in the formation of ozone in the lowest layer of the atmosphere. Whereas
the fate of isoprene in atmospheric chemistry has been studied extensively [4±6],
Ã
Corresponding author

674
D. R. Boyd et al.
its effect on living systems appears to have been mainly con®ned to animals
[7±10]. Isoprene is also produced in large quantities by the petrochemical industry
as a monomer for cis-1,4-polyisoprene elastomer, and in this context the potential
risk of human exposure to higher concentrations of anthropogenic isoprene than
may normally occur in nature through biogenic isoprene has initiated toxicological
studies [7±10]. These animal studies have shown that the ®rst step in the metabolic
pathway of isoprene degradation involves the monooxygenase-catalyzed formation
of mutagenic epoxides. Liver microsomal cytochrome P-450 enzymes and
dioxygen were found to be responsible for both mono- and bis-epoxidation of 1.
Hydrolysis of the isoprene monoepoxide metabolites gave the corresponding diols
2 and 3. Apart from these earlier reports on the metabolism of 1 in eucaryotic cells
[7±9], to our knowledge the metabolic fate of isoprene or related butadienes in
procaryotic (bacterial) cells has not been reported1. Since both plants, a major
biogenic source of isoprene, and bacteria, a major source of hydrocarbon-
degrading enzymes, are found to co-exist within soil, metabolism by soil bacteria
could play an important role as a biological sink controlling the concentration of
biogenic isoprene released into the environment. This has indeed recently been
established for the ®rst time using both in vitro and in vivo studies of the microbial
consumption of atmospheric isoprene by aerobic soil bacteria in a temperate forest
soil [2]. However, this microbial consumption programme did not address the
question of biodegradation pathways or the nature of metabolic intermediates and
is thus complementary to the results obtained in the present study.
Dioxygenase-catalyzed cis-dihydroxylation of mono- and polycyclic aromatic
compounds using Pseudomonas putida UV4 has been carried out in these
laboratories [11±15]. Dioxygenase enzymes are also known to be present in two
wild-type strains of P. putida, i.e. ML 2, a source of benzene dioxygenase (BDO),
and NCIMB 8859, a source of naphthalene dioxygenase (NDO). UV4, a mutant
strain of P. putida contains a dioxygenase (toluene diooxygenase, TDO), but is
de®cient in the cis-diol dehydrogenase enzyme which is responsible for the
conversion of arene cis-dihydrodiols to catechols in the corresponding wild-type
strain. Non-aromatic conjugated cyclic diene and triene substrates of different ring
sizes have been reported to undergo stereoselective alkene cis-dihydroxylation in
the presence of P. putida UV4 [15]. It has also been demonstrated that a cyclic
aliphatic alkene can be dihydroxylated by toluene dioxygenase from P. putida F1
[17]. Dioxygenase-catalyzed asymmetric dihydroxylation of a series of volatile
acyclic conjugated dienes, elucidation of the structure and stereochemistry of the
resultant diols, and the metabolic sequence involved in the formation of other
derived metabolites formed from this class of NMOCs are presented and discussed
here.
Results and Discussion
A major objective of this study was to elucidate the sequence of steps involved in
the metabolism of dienes 1, 7, 15, and 19 by different strains of the soil bacterium
P. putida (Schemes 1±4) and to characterize and, where possible, to assign
1
See note added in proof

Bacterial Biotransformation of Dienes
675
stereochemistry to the metabolites. Due to the volatile nature of dienes 1 (b.p.:
34ꢁC), 7 (b.p.: 42ꢁC), 15 (b.p.: 44ꢁC), and 19 (b.p.: ±5ꢁC) and their limited
solubility in the aqueous culture medium, the isolated yields were expected to be
quite low using our reported standard biotransformation conditions [11, 14, 16].
However, by modi®cation of these biotransformation procedures and utilization of
more sensitive assay methods, it was possible to identify and stereochemically
assign the bioproducts. This was achieved by using a lower incubation temperature
(4ꢁC) and higher cell densities in shake ¯asks (optical density values of ca. 10 at
600 nm), by addition of a competitive inhibitor of the diol dehydrogenase
(propylene glycol) during biotransformations, and by application of chiral
stationary phase GC/FID, GC/MS, and GC/IR methods for analysis of culture
supernatants. Despite these modi®cations the isolated yields of bioproducts were
relatively low (< 10%) and thus currently of limited synthetic value.
Based on reported results using P. putida UV4 and a series of cyclic diene
substrates where the corresponding vic-diols were obtained [15], diols 2, 3, 8, 9,
16, 17, and 20 were expected to be the initial metabolities from the acylic dienes 1,
7, 15, and 19. Authentic samples of these diols were prepared by osmylation
(OsO4) of the corresponding dienes 1, 7, 15, and 19 and were identi®ed by NMR
spectroscopy, GC/MS, and GC/IR analysis and by comparison with literature data.
Enantiomerically enriched (7±79% ee) samples of these diols were also obtained
by catalytic asymmetric dihydroxylation using osmium tetroxide in the presence of
a chiral ligand (AD-mix-ꢀ). This chemical method of asymmetric dihydroxylation
has been studied extensively by Sharpless et al. [18, 19] using a range of alkenes
and dienes, and a model has been developed which allows tentative assignments of
the absolute con®gurations of the product diols to be made.
In view of its importance in the environment, the biotransformation of isoprene
(1) was studied more rigorously than that of butadiene (19) or the diene isomers 7
and 15 (Table 1). Using the ML2 and 8859 strains of P. putida as sources of BDO
and NDO, respectively, and 1 as substrate, diols 2 and 3 were isolated as initial
metabolites. These diols could not be detected using GC/MS analysis and the TDO
enzyme present in the UV 4 strain (Table 1). The regioselectivity of dioxygenase-
catalyzed dihydroxylation was consistently found to favour attack at the
monosubstituted alkene bond relative to the gem-disubstituted alkene bond
Scheme 1

676
D. R. Boyd et al.
Table 1. Concentrations of bioproducts 2, 3, 5, and 6, respectively, and enantiomeric excess values (ee)
obtained from the biodegradation of isoprene (1) by the ML2, UV4, and 8859 strains of P. putida
Propylene [3]/mM ee (3)
glycol
%v/v
[2]/mM ee (2)
[2]/[3]
[5]/mM
[6]/mM
ML2a
ML2a
UV4b
UV4b
8859c
8859c
0
15
0
1.5
1.5
0
40%
44%
6.3
6.4
0
34%
16%
4.2
4.3
2.0
0
0.2
0
0
0
15
0
0.4
0.06
0.2
45%
14%
16%
0.9
0.2
0.6
9%
12%
12%
2.3
3.3
3.0
0
0
0
0
15
0
0
a
b
BDO; TDO; NDO
c
(2/3 ˆ 3.3±4.2). The highest yields of diols 2 (6.3 mM) and 3 (1.5 mM) were
obtained using the ML2 strain. Using the latter strain the yield of diol 2 was further
diminished by oxidative metabolism under the catalytic in¯uence of an alcohol
dehydrogenase enzyme to yield ketoalcohol 4 which was further metabolized to
yield the saturated ketoalcohol 5 and saturated diol 6. Since the tertiary alcohol
group in diol 3 was unaffected by the alcohol dehydrogenase, no further
metabolites were found.
It had earlier been found that propylene glycol can act as a competitive
inhibitor both for a diol dehydrogenase enzyme isolated from the wild-type ML2
strain (ca. 36% activity compared with the parent substrate, benzene cis-glycol)
[20] and for a diol dehydrogenase present in an E. coli recombinant strain cloned
from the ML2 strain (ca. 189% activity compared with the parent substrate,
benzene cis-glycol) [21]. Hence, use of the ML2 strain and addition of propylene
glycol was expected to reduce the rate of formation of ketoalcohol and to give a
more reliable estimate of both regio- and stereoselectivity resulting from
dioxygenase-catalyzed dihydroxylation. The effect of adding propylene glycol
(15% v/v) to biotransformations of 1 with the three P. putida strains is evident from
Table 1. Whereas the yields of diols 2 and 3 increased slightly using the wild-type
ML2 and 8859 strains, ketoalcohol formation was eliminated when the ML2 strain
was used with propylene glycol as additive. A more marked effect was observed
with the UV4 strain where formation of 2 and 3 was increased to levels which were
readily detectable (0.4 and 0.9 mM, respectively). This result suggests that the
mutant UV 4 strain, although lacking the arene cis-dihydrodiol dehydrogenase
enzyme, may contain a further type of dehydrogenase which can oxidize acyclic
substrates like 2 and 3 and which is also effectively inhibited by propylene glycol.
The stereoselectivity of the enzyme-catalyzed dihydroxylation process was
determined by use of a chiral stationary phase (CSP) GC column to resolve the
enantiomers followed by identi®cation using mass and IR spectroscopy. CSPGC/
FID analysis was used to determine ee values of the samples of diols 2, 3, 8, 9, 16,
17, and 20 obtained using both chemical (OsO4 and AD-mix-ꢀ) and biological
(dioxygenase) asymmetric dihydroxylation catalysts. The absolute con®gurations
were tentatively assigned on the basis of (i) the preferred con®gurations of diols
produced using the ꢀ-form of AD-mix and the model developed by Sharpless for

Bacterial Biotransformation of Dienes
677
Table 2. Enantiomeric excess, absolute con®guration, and ratio of diol isomers 2, 3, 8, 9, 16, 17, and 20
obtained by asymmetric dihydroxylation using AD-mix-ꢀ
Diol
ee (%)
Abs. con®g.
Isomer ratio
3
2
21
64
79
65
7
2R
2R
2.0 (2/3)
9
2R, 3R
2R
8
0.36 (8/9)
1.8 (16/17)
17
16
20
2R, 1S
2R
44
45
2R
predicting the enantiofacial selectivity of the dihydroxylation process for different
alkene types [18, 19] (Table 2) and (ii) the elution sequence observed using
cyclodextrin-based CSPGC columns. The late eluting enantiomer of each diol was
assigned the R con®guration at the allylic chiral centre.
The enantiomeric excess values observed for the diols 2, 3, 8, 9, 16, 17, and 20
obtained by asymmetric dihydroxylation using ꢀ-AD-mix (H2O:t-BuOH, 20ꢁC)
were found to be in the range of 7±79% with the lowest value being associated with
dihydroxylation of a cis-disubstituted alkene bond as predicted from earlier studies
[18, 19]. Using the BDO enzyme (from ML2), 1 was asymmetrically
dihydroxylated to yield diols 2 (16% ee in the presence of propylene glycol) and
3 (44% ee in the presence of propylene glycol) with an excess of the R enantiomer.
This enantiomeric excess increased (34% ee) for diol 2 when the diol
dehydrogenase enzyme was present. When the TDO (from UV4) and NDO (from
8859) enzymes were used, the stereoselectivity was again generally low for diols 2
(9±12%) and 3 (14±45%) but also showed a preference for the R con®guration.
When the TDO- and NDO-containing strains of P. putida (UV4 and 8859) and
1 as substrate were studied, the diols 2 and 3 were formed without evidence of
further metabolism to the corresponding ketoalcohol 4 or their derivatives.
Similarly, when the BDO-containing strain (ML2) was used, only a trace of
ketoalcohol 4 was detected by GC/MS and GC/IR analysis; however, in this case
signi®cant quantities of the saturated ketoalcohol 5 (2.0 mM) and diol 6 (0.2 mM)
were identi®ed. The unexpected formation of the latter two metabolities was
con®rmed by addition of an authentic sample of the unsaturated ketoalcohol 4 to
ML2 cultures. The alkene bond in compound 4 was found to be rapidly
hydrogenated to yield the ketol 5 which was in turn slowly reduced to diol 6.
Further evidence for the metabolic sequence alkene dihydroxylation ! allylic
alcohol oxidation ! alkene hydrogenation ! saturated ketone reduction (Scheme
1) was obtained from a time-course study over 24 h which showed the unsaturated
diol 2 to be the most abundant metabolite during the biotransformation, whereas
the saturated diol 6 was only detected after 12 h (Fig. 1).
A time-course study of the biotransformation of the racemic cis-diol
metabolites of isoprene (2 and 3) using the ML2 strain over a 24 h period again
showed that the concentration of 3 remained constant, whereas the proportion of 2
decreased as the amounts of 5 and 6 increased (Fig. 2). CSPGC along with both
FID and MS detection demonstrated that the R enantiomer of 2 was selectively

678
D. R. Boyd et al.
Fig 1. Time-course study of the formation of metabolites 2, 3, 5, and 6 from isoprene (1) using
P. putida ML2
Fig 2. Time-course study of the formation of metabolites 5 and 6 from racemic diols 2 and 3 using
P. putida ML2
oxidized to ketoalcohol 4, and after 24 h the residual diol was mainly of the S
con®guration (95% ee). The enantiopurity of 2 obtained during the biotransforma-
tion of 1 is thus due to the combined effect of asymmetric dihydroxylation and
kinetic resolution each favouring the R con®guration and is an example of
enantiocomplementarity. It has been established from metabolic studies of arene
substrates [20] that the ML2 strain of P. putida contains both benzene dioxygenase
and cis-diol dehydrogenase enzymes which could in the present context account
for the formation of diols 2 and 3 and the ketoalcohol 4. The nature of ML2
enzymes responsible for the ®nal two steps, i.e. hydrogenation to yield ketol 5 and
ketone reduction to yield diol 6, has not been established. Formation of the
saturated ketoalcohol 5 and saturated diol 6 indicates that enzymatic hydrogenation
can occur on the gem-disubstituted alkene.
Based upon the higher yields of diol metabolites obtained and the unexpected
results resulting from both oxidation (2 ! 4) and reduction steps (4 ! 5 ! 6)
using the ML2 strain with 1 as substrate, the related diene substrates trans-
piperylene (7), cis-piperylene (15), and 1,3-butadiene (20) were also examined
using this strain (Schemes 2±4). In accordance with expectations, the trans isomer
7 was found to yield the corresponding diol metabolites 8 (0.6 mM) and 9 (0.3 mM)
with regioselectivity favouring the monosubstitued alkene bond, i.e. the rate of

Bacterial Biotransformation of Dienes
679
Scheme 2
formation of diol 8 was ca. ®ve-fold faster compared with diol 9. In the presence of
propylene glycol, which was added to inhibit further oxidative metabolism, both
diols were found to be formed in relatively low enantiomeric excess (8, 38% ee; 9,
33% ee). Using the methods discussed earlier, the absolute con®gurations were
tentatively assigned as 2R, 3R, and 2R for diols 9 and 8, respectively. In the absence
of propylene glycol, diol 8 was rapidly metabolized in a stereoselective manner to
yield the unsaturated ketoalcohol 10 (0.5 mM) and the saturated ketoalcohol 11
(1.6 mM). The residual diol 8 was found to be of high enantiopurity (93% ee) and
of the 2S con®guration. In a similar manner, diol 9 was further biotransformed but
at a slower rate (presumably via the ketoalcohol 12 which was undetected) to the
saturated ketoalcohol 13 (0.1 mM) and the diol 14 (0.1 mM). The formation of the
saturated ketoalcohols 11 and 13 and of the saturated diol 14 indicates that the
enzyme-catalyzed hydrogenation can occur on both monosubstituted and trans-
disubstituted alkenes. The remaining diol 9 was found to be present with an excess
of the 2S, 3S enantiomer (15% ee). Evidence in support of the metabolic sequences
7 ! 8 ! 10 !11 and 7 ! 9 ! 12 ! 13 ! 14 was obtained by addition of the
Scheme 3

680
D. R. Boyd et al.
racemic diols 8 and 9 individually as substrates and the detection of the
corresponding metabolites 10, 11, 13, and 14. The formation of both diols 8 and 9
with an excess of the R allylic alcohol chiral centre from asymmetric
dihydroxylation (in the presence of an inhibitor) and the preferential oxidation
of this enantiomer to leave mainly the S enantiomer after kinetic resolution again
provides an example of enzymatic enantiocomplementarity, i.e. the use of enzymes
to produce either enantiomer.
The biotransformation of cis-1,3-pentadiene (15) using the ML 2 strain in the
presence of propylene glycol yielded the expected diol metabolites 16 (1.1 mM;
74% ee) and 17 (1.2 mM; 70% ee). As before, the dioxygenase-catalyzed
dihydroxylation appeared to favour diol formation at a monosubstituted alkene
bond over a disubstituted alkene bond (ca. 3-fold rate difference) and of an
enantiomer having an R con®guration at the allylic chiral centre. Preferential
removal of the allylic R enantiomer from diols 16 and 17 was again observed in the
absence of propylene glycol inhibitor, leading to a residual sample of diol 16 (26%
ee, 2R) and diol 17 (97% ee, 2S, 3R). Diol 16 proved to be a particularly good
substrate yielding only the unsaturated ketoalcohol product 18 (4.4 mM), whereas
diol 17 yielded the saturated ketoalcohol 13 (1.0 mM), probably via reduction of
the saturated ketoalcohol 12. It is noteworthy that while enzymatic hydrogenation
was found on the monosubstituted alkene group in the ketoalcohol 12, the cis-
disubstituted alkene group in metabolite 18 was not hydrogenated.
On account of its volatility, butadiene (19) proved to be more dif®cult to study
experimentally using the ML 2 strain of P. putida. However, by bubbling gaseous
19 into heptamethylnonane (0ꢁC) and adding the resulting solution to the culture
medium (4ꢁC), dioxygenase-catalyzed dihydroxylation was found to yield diol 20
(0.2 mM) with a relatively low preference for the 2R enantiomer (25% ee). In the
absence of the diol dehydrogenase inhibitor the 2R enantiomer of diol 20 was
selectively oxidized to yield the saturated ketoalcohol 22 (0.3 mM). The unreacted
diol 20 was found to be mainly of 2S con®guration (95% ee). Thus, the
enantiocomplementarity previously observed during the biotransformations of
dienes 1, 7, and 15 to yield the corresponding chiral diols was again observed for
diene 19. It is noteworthy that the biotransformation of monosubstituted benzene
and naphthalene substrates by P. putida mutant strains has been reported [22] to
yield a single cis-dihydrodiol enantiomer and that the closely coupled cis-diol
dehydrogenase was found to act exclusively on the same enantiomer.
Scheme 4

Bacterial Biotransformation of Dienes
681
Scheme 5
Conclusions
Bacterial biodegradation of isoprene (1) and related diene substrates 7, 15, and 19
using three different strains of the soil bacterium P. putida has generally been
found to proceed via the dioxygenase-catalyzed asymmetric dihydroxylation of an
alkene group to yield all of the possible diols 2, 3, 8, 9, 16, 17, and 20, each with an
excess of the allylic R enantiomer. The rate of dihydroxylation appeared to be
faster on monosubstituted rather than disubstituted alkene bonds. Further diol
dehydrogenase-catalyzed oxidation at the secondary allylic hydroxyl group to yield
unsaturated ketoalcohols 4, 10, 12, 18, and 21 was found to occur preferentially on
the allylic R enantiomer, thus leading to the recovery of diols having an excess of
the allylic S con®guration. The unsaturated ketoalcohol bioproducts were further
metabolized through a reductive process involving enzymatic hydrogenation to
yield saturated ketoalcohols 5, 11, 13, and 22 and ketone reduction to yield diols 6
and 14. A general biodegradation pathway to account for the formation of all
metabolites from dienes 1, 7, 15, and 19 is shown in Scheme 5.
Experimental
NMR analyzis was carried out using both Bruker Avance DPX-300 (300 MHz) and DRX-500
(500 MHz) instruments. All NMR samples were run in CDCl3 using internal TMS as reference.

682
D. R. Boyd et al.
Coupling constants are expressed in Hz. GC/MS analyses were carried out using a Hewlett Packard
6890 gas chromatograph directly linked to a Hewlett Packard 5973 Mass Selective Detector (MSD).
Mass spectroscopic data were obtained by operating the MSD in the electron impact scanning mode
measuring ion currents between m/z ˆ 30 and 400. The instrument was also operated in the chemical
ionization mode using methane as reagent gas (CIMS) and measuring ion currents between m/z ˆ 65
and 400 to determine the molecular ion for each compound. GC/IR analysis was completed using a
Hewlett Packard 5890A gas chromatograph/5965B ± infrared detector/mass selective detector (GC/
IRD/MSD) system with the IRD/MSD in the series con®guration. The IRD scanned the range of
750±4000 cm 1. The gas chromatographs which were ®tted with Supelco fused silica capillary
columns (30 mÂ0.25 mm i.d. with a 0.25mm Beta Dex-120 phase) were equipped with autosamplers.
1
The ovens were programmed at 65ꢁC for 1 min and then ramped at 10ꢁC min to 220ꢁC and held at
this temperature for 5 min. Samples (1 mm3) were injected in the splitless mode into the injector port
which was maintained at 250ꢁC. Helium was employed as the carrier gas at a ¯ow rate of
1 cm3 Á min 1. Quanti®cation of analytes was performed using gas chromatography with ¯ame
ionisation detection (GC/FID). Chromatographic separations for FID were performed using the
following conditions: system A: 10 m Chrompak CP-Chirasil-DEX CB fused silica column (3.3 psi,
isothermal at 100ꢁC); system B: 10 m Chrompak CP-Chirasil-DEX CB fused silica column (3.3 psi,
isothermal at 110ꢁC); system C: 30 m Supelco-ꢀ-DEX 225 column (35 psi, isothermal at 100ꢁC);
system D: 30 m Supelco-ꢀ-DEX 225 column (19 psi; isothermal at 100ꢁC).
The dienes 1, 7, 15, and 19 (> 97% purity) were purchased from Aldrich and used as substrates
without further puri®cation. Diol 20 in racemic form was available from Fluka.
Authentic samples of the known diol metabolites 2, 3, 8, 9, 16, and 17 were obtained by
osmylation. This hydroxylation procedure proved to be rather dif®cult due to the reactivity of the
initially formed diols in the presence of OsO4 and hence resulted in relatively low yields. In a typical
general procedure a solution of diene (50 mM) and N-methylmorpholine-N-oxide (55 mM) in a
mixture of acetone (50 cm3) and water (1.5 cm3) was added to a catalytic amount of OsO4. After
stirring overnight at ambient temperature the solution was concentrated in vacuo. The residual oil
was dissolved in CH2Cl2, dried over MgSO4, concentrated, and puri®ed by ¯ash chromatography
(silica gel) starting with CH2Cl2 (100%) as eluent and gradually increasing the acetone concentration
to 10% v/v. The product was in each case an oil containing a racemic mixture of the two isomeric
diols (2/3 8/9, 16/17). The yield of diols obtained were in the range of 20±30%. The structures of the
individual diol metabolites were con®rmed by comparison of IR, 1H NMR, and MS data comparison
with those of authentic samples. The ML 2 strain of P. putida was obtained from University of
Warwick (Prof. H. Dalton); the NCIMB 8859 strain was available from the Collection of Industrial
and Marine Bacteria Ltd., Aberdeen, UK. Since authentic samples of the cis-diols 2/3, 8/9, 16/17,
and 20 were available and since these were always the initial (and often the major) metabolites, a
summary of the relevant spectroscopic and chromatographic data for each of these known racemic
compounds is provided [4, 8].
(2R/2S)-1,2-Dihydroxy-2-methyl-3-butene (3; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.3 (3H, s, Me), 3.45 (2H, d, J ˆ 11.0, CHHOH), 3.50 (1H, d,
J ˆ 11.0, CHHOH), 3.73 (2H, bs, OH), 5.2 (1H, dd, J ˆ 10.4, 1.0, =CHH), 5.35 (1H, dd, J ˆ 17.6,
1.0, =CHH), 5.9 (1H, dd, J ˆ 10.4, 17.6, ±CH=) ppm; GC/IR: v ˆ 1644, 3095 (H2C=CH), 3626 (OH)
‡
cm 1; GC/CIMS: MH ˆ 103; GC/EIMS: m/z ˆ 71 (100), 43 (82), 41 (27), 31 (18), 87 (3); CSPGC/
FID: system A: 2S-3 (7.8 min), 2R-3 (8.1 min).
(2R/2S)-1,2-Dihydroxy-3-methyl-3-butene (2; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.75 (3H, s, Me), 3.55 (1H, dd, J ˆ 11.0, 7.3, HHCOH), 3.67 (1H,
dd, J ˆ 11.0, 2.8, HHCOH), 4.18 (1H, dd, J ˆ 7.3, 2.8, CHOH), 4.96 (1H, d, J ˆ 0.8, =CHH), 5.06

Bacterial Biotransformation of Dienes
683
(1H, d, J ˆ 0.8, =CHH) ppm; GC/IR: v ˆ 1649, 3085 (C=CH2), 3636 (OH) cm 1; GC/EIMS: m/
z ˆ 102 (M+, 0.5), 71 (100), 43 (92), 41 (59), 39 (43), 31 (25), 87 (1); CSPGC/FID: system A: 2S-2
(15.1 min), 2R-2 (16.3 min).
(2S,3S/2R,3R)-2,3-Dihydroxy-4-pentene (9; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.1 (3H, d, J ˆ 6.3, Me), 3.0 (2H, bs, 2ÂOH), 3.65 (1H, dq, J ˆ 6.3,
6.3, CHOHMe), 3.7 (1H, dd, J ˆ 7.0, 6.3, CHOH), 5.24 (1H, d, J ˆ 10.4, =CHH), 5.36 (1H, d,
J ˆ 17.0, =CHH), 5.85 (1H, ddd, J ˆ 17.0, 10.4, 6.3, H2C=CH) ppm; GC/IR: v ˆ 1644, 3090
(H2C=CH), 3618 (OH) cm 1; GC/EIMS: m/z ˆ 45 (100), 58 (78), 57 (56), 43 (28), 31 (14); CSPGC/
FID: system B: 2S,3S-9 (4.84 min), 2R-3R-9 (5.23 min).
Trans-(2R/2S)-1,2-Dihydroxy-3-pentene (8; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.7 (3H, d, J ˆ 6.5, Me), 2.3 (2H, bs, 2ÂOH), 3.3 (1H, dd, J ˆ 11.0,
7.7, CHH), 3.7 (1H, dd, J ˆ 5.3, 11.0, CHH), 4.2 (1H, ddd, J ˆ 8.4, 7.0, 3.5, CHOH), 5.48 (1H, ddq,
J ˆ 12.6, 6.9, 15.5, HC=CHMe), 5.8 (1H, dq, J ˆ 15.5, 6.5, =CHMe) ppm; GC/IR: v ˆ 1672, (C=C),
‡
3635 (OH) cm 1; GC/EIMS: m/z ˆ 102 (M , 0.5), 71 (100), 43 (42), 41 (40), 39 (26), 53 (26), 31
(21), 87 (0.5); CSPGC/FID: system B: S-8 (7.4 min), R-8 (8.2 min).
(2R3S/2S3R)-2,3-Dihydroxy-4-pentene (17; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.15 (3H, d, J ˆ 6.4, Me), 2.15 (2H, bs, 2ÂOH), 3.88 (1H, dq,
J ˆ 6.4, 3.6, CHMe), 4.09 (1H, dd, J ˆ 6.3, 3.6, CHOH), 5.27 (1H, ddd, J ˆ 10.5, 1.2, 1.2, HHC=),
5.35 (1H, ddd, J ˆ 17.2, 1.2, 1.2, HHC=), 5.91 (1H, ddd, J ˆ 17.2, 10.6, 6.5, =CH) ppm; GC/IR:
v ˆ 1644, 3089 (H2C=CH), 3632 (OH) cm 1; GC/EIMS: m/z ˆ 45 (100), 58 (83), 57 (60), 43 (19);
CSPGC/FID: system C: 2S,3R-17 (9.2 min), 2R,3S-17 (9.5 min).
cis-(2R/2S)-1,2-Dihydroxy-3-pentene (16; C5H10O2)
1H NMR (500 MHz, ꢁ, CDCl3): 1.71 (3H, d, J ˆ 7.0, Me), 2.15 (2H, bs, 2ÂOH), 3.50 (1H, dd,
J ˆ 11.2, 7.9, CHHOH), 3.59 (1H, dd, J ˆ 11.2, 3.7, CHHOH), 4.59 (1H, ddd, J ˆ 8.4, 8.4, 3.6,
1
CHOH), 5.39 (1H, m, =CHMe), 5.65 (1H, m, HC=) ppm; GC/IR: v ˆ 1658 (C=C), 3636 (OH) cm
;
GC/EIMS: m/z ˆ 71 (100), 43 (41), 41 (41), 39 (27), 53 (27), 31 (22), 87 (1); CSPGC/FID: system C:
2S-16 (14.6 min) 2R-16 (15.6 min).
(2R/2S)-1,2-Dihydroxy-3-butene (20; C4H8O2)
1H NMR (500 MHz, ꢁ, CDCl3): 3.50 (1H, dd, J ˆ 11.0, 7.5, CHHOH), 2.95 (2H, bs, 2ÂOH), 3.67
(1H, dd, J ˆ 11.4, 3.4, CHHOH), 4.26 (1H, m, CHOH), 5.23 (1H, ddd, J ˆ 10.6, 1.4, 1.4, HHC=),
5.35 (1H, ddd, J ˆ 17.3, 1.5, 1.5, HHC=), 5.85 (1H, ddd, J ˆ 17.0, 10.4, 5.6), ppm GC/IR: v ˆ 1648
(C=C), 2887, 2938, 3091 (CH), 3641 (OH) cm 1; GC/EIMS: m/z ˆ 57 (100), 58 (65), 31 (62), 39
(44), 70 (37), 42(35); CSPGC/FID: system D: 2S-20 (10.6 min) 2R-20 (11.0 min).
Metabolites 5, 6, 10, 11, 13, 14, 18, and 22 were identi®ed from IR, EIMS and CIMS data
obtained directly after GC separation.
1-Hydroxy-3-methyl-2-butanone (5; C5H10O2)
GC/IR: v ˆ 3530 (OH), 1727 (C=O) cm 1; GC/EIMS: m/z ˆ 102 (M+, 1), 31 (31), 43 (100), 71 (35),
‡
41 (48); GC/CIMS: m/z ˆ 103 (MH ).

684
D. R. Boyd et al.
3-Methyl-1,2-butanediol (6; C5H12O2)
GC/IR: v ˆ 3644 (OH), 2968, 2892 (CH), 1058 cm 1; GC/EIMS: m/z ˆ 73 (100), 55 (70), 43 (60), 61
‡
(40), 31 (39), 41 (25); GC/CIMS: m/z ˆ 87 (MH-H2O ).
(E)-1-Hydroxy-3-penten-2-one (10; C5H8O2)
GC/IR: v ˆ 3525 (OH), 1702 (ꢂ,ꢀ-unsatd. C=O), 1643, 966 (C=C) cm 1; GC/EIMS: m/z ˆ 69 (100),
‡
41 (98), 39 (54); 31 (19); GC/CIMS: m/z ˆ 101 (MH ).
1-Hydroxy-2-pentanone (11; C5H10O2)
‡
GC/IR: v ˆ 3530 (OH), 1728 C=O) cm 1; GC/EIMS: m/z ˆ 102 (M , 6), 43 (100), 71 (51), 31 (34);
‡
GC/CIMS: m/z ˆ 103 (MH ).
2-Hydroxy-3-pentanone (13; C5H10O2)
‡
GC/IR: v ˆ 3535 (OH), 1728 C=O) cm 1; GC/EIMS: m/z ˆ 102 (M , 2), 45 (100), 57 (36), 31 (9);
‡
GC/CIMS: m/z ˆ 103 (MH ).
2,3-Pentanediol (14; C5H12O2)
‡
GC/EIMS: m/z ˆ 59 (100), 31 (68), 45 (60), 41 (31), 75 (14); GC/CIMS: m/z ˆ 87 (MH-H2O ).
(Z)-1-Hydroxy-3-penten-2-one (18; C5H8O2)
‡
GC/IR: v ˆ 3525 (OH), 1701 (ꢂ,ꢀ-unsatd. C=O), 1631 (C=C) cm 1; GC/EIMS: m/z ˆ 100 (M , 2),
‡
69 (100), 41 (79), 39 (52), 31 (21); GC/CIMS: m/z ˆ 101 (MH ).
1-Hydroxy-2-butanone (22; C4H8O2)
‡
GC/IR: v ˆ 3536 (OH), 1733 (C=O) cm 1; GC/EIMS: m/z ˆ 88 (M ,13), 57 (100), 31 (36); GC/
‡
CIMS: m/z ˆ 89 (MH ).
Substrates were metabolized on a small scale using shake ¯ask cultures of the constitutive mutant
strain (UV4) of P. putida and the wild type P. putida NCIMB 8859 according to reported methods
[15, 16, 20]. The wild type P. putida ML2 was grown on benzene as the sole carbon source. To obtain
biomass, a 10 litre Biolab fermenter containing rich media (Luria Broth) was inoculated with a 1%
inoculum of benzene grown cells. The cells were harvested by centrifugation (10000 g for 15 min)
during the late exponential phase of growth and resuspended in potassium phosphate buffer
(100 mM, pH ˆ 7.5) to an optical density of 10 at 600 nm. These cell suspensions were used for the
bioconversions. A typical procedure involved incubation of diene substrates 7, 15, and 19 (60 mm3 in
3 cm3 of culture medium containing the bacterial cells, OD ˆ 10 at 600 nm) in shake ¯asks (10 cm3)
at a temperature of ‡4ꢁC and pH ˆ 7.5 for 24 h. As dienes 7, 15, and 19 had a very low solubility in
water, a two-phase system was formed with the aqueous phase being saturated with the diene. In the
case of 1,3-butadiene (19) the substrate was passed as a gas into heptamethylnonane to yield a
saturated solution at 0ꢁC. Subsequently, a portion (200 mm3) of this solution was added to the
bacterial cell solution. Bioconversions with the chemically synthesized mixtures of racemic diols 2,
3, 8, 9, 16, 17, and 20 as substrates for P. putida ML2 were carried out at a diol concentration of 2.0
g.dm 3. The metabolites were obtained by rigorous extraction of the culture medium with ethyl
acetate after saturation with salt. The extract was concentrated and analyzed by chiral stationary
phase FID, MS, and IR detection. The structures of the cis-diol metabolites 2, 3, 8, 9, 16, 17, and 20
were established by comparison with authentic samples. The remaining ketoalcohols (5, 10, 11, 13,

Bacterial Biotransformation of Dienes
685
18, and 21) and diols (6 and 14) were identi®ed on the basis of comparison with literature data and
from MS and IR spectra.
Note Added in Proof
In a recent report [23] the metabolism of isoprene in a Rhodococcus strain was found to occur via a
monooxygenase-catalyzed epoxidation. No evidence of diene epoxidation was obtained using the
P. putida strains in the present study.
Acknowledgements
We wish to thank Professor H. Dalton for the P. putida ML2 strain, Dr N. I. Bowers for the
preliminary studies on cis-piperylene, and the Technology Development Programme of the European
Regional Development Fund and the IRTU for ®nancial support (MC, DC).
References
[1] Atkinson R, Arey J (1998) Acc Chem Res 31: 574
[2] Cleveland CC, Yavitt JB (1998) Appl Environ Microbiol 64: 172
[3] KonigG,BrundaM,PuxbaumH,HewittCN,DuckhamSC,RudolfJ(1995)JAtmosEnviron29:861
[4] Becker KH, Barnes I, Ruppert L (1994) Physico-Chemical Behaviour of Atmospheric Pollutants
1: 54
[5] Peeters A, Ectors A, Boullart W (1994) Physico-Chemical Behaviour of Atmospheric Pollutants
1: 61
[6] Madronich S, Calvert JG (1990) J Geophys Res 95: 5697
[7] Bleasdale C, Small RD, Watson WP, Wilson J, Golding BT (1996) Toxicology 113: 290
[8] Wistuba D, Weigand K, Peter H (1994) Chem Res Toxicol 7: 336
[9] Longo V, Citti L, Gervasi PG (1985) Toxicology Letters 29: 33
[10] Gervasi PG (1985) Mutat Res 156: 77
[11] Allen CCR, Boyd DR, Dalton H, Sharma ND, Brannigan I, Kerley NA, Sheldrake GN, Taylor ST
(1995) J Chem Soc Chem Commun 117
[12] Allen CCR, Boyd DR, Hempenstall F, Larkin MJ, Sharma ND (1999) Appl and Environ
Microbiol 65: 1335
[13] Boyd DR, Sharma ND, Agarwal RA, Resnick SM, Schocken MJ, Gibson DT, Sayer JM, Yagi H,
Jerina DM (1997) J Chem Soc Perkin Trans 1, 1715
[14] Boyd DR, Sharma ND, Evans TA, Groocock MG, Malone JF, Stevenson PJ, Dalton H (1997)
J Chem Soc Perkin Trans 1, 1879
[15] Bowers NI, Boyd DR, Sharma ND, Kennedy MA, Sheldrake GN, Dalton H (1998) Tetrahedron
Asymm 9: 1831
[16] Boyd DR, Sharma ND, Haughey SA, Kennedy MA, McMurray BT, Sheldrake GN, Allen CCR,
Dalton H, Sproule K (1998) J Chem Soc Perkin Trans 1, 1929
[17] Lange CC, Wackett LP (1997) J Bacteriology 179: 3858
[18] Kolb HC, VanNiewenhze MS, Sharpless KB (1994) Chem Rev 94: 2483
[19] Wang ZM, Kakiuchi K, Sharpless KB (1995) J Org Chem 59: 6897
[20] Allen CCR, Boyd DR, Dalton H, Sharma ND, Walker CE (in preparation)
[21] Fong KPY, Goh CBH, Tan, H-M (1996) J Bacteriol 178: 5592
[22] Allen, CCR, Boyd DR, Dalton H, Sharma ND, Brannigan I, Kerley NA, Sheldrake GN, Taylor
SC (1994) J Chem Soc Chem 117
[23] van Hylckama Vlieg JE, Leemhuis H, Spelberg JH, Janssen DB (2000) J Bacteriol 182: 1956
Received December 20, 1999. Accepted (revised) February 7, 2000