Biocatalytic Asymmetric Hydroxylation of Hydrocarbons with the Topsoil-Microorganism Bacillus megaterium

Article abstract of DOI:10.1021/jo991725s

A Bacillus megaterium strain was isolated from topsoil by a selective screening procedure with allylbenzene as a xenobiotic substrate. This strain performed the hydroxylation chemoselectively (no arene oxidation and overoxidized products) and enantioselectively (up to 99% ee) in the benzylic and nonbenzylic positions of a variety of unfunctionalized arylalkanes. Salycilate and phenobarbital, which are potent inducers of cytochrome P-450 activity, changed the regioselectivity of the microbial CH insertion, without an effect on the enantioselectivity. The biotransformation conditions were optimized in regard to product yield and enantioselectivity by variation of the oxygen-gas supply and the time of the substrate addition. The different product distributions (α- versus β-hydroxylated product) that are obtained on induction of cytochrome P-450 enzyme activity demonstrate the involvement of two or more hydroxylating enzymes with distinct regioselectivities in this biotransformation. An oxygen-rebound mechanism is assumed for the cytochrome P-450-type monooxygenase activity, in which steric interactions between the substrate and the enzyme determine the preferred face of the hydroxy-group transfer to the radical intermediate.

Full text of DOI:10.1021/jo991725s

878
J . Org. Chem. 2000, 65, 878-882
Bioca ta lytic Asym m etr ic Hyd r oxyla tion of Hyd r oca r bon s w ith th e
Top soil-Micr oor ga n ism Ba cillu s m ega ter iu m
Waldemar Adam,† Zoltan Lukacs,*,†,‡ Dag Harmsen,§ Chantu R. Saha-Mo¨ller,† and
Peter Schreier‡
Institutes of Organic Chemistry and Food Chemistry, University of Wu¨rzburg, Am Hubland,
D-97074 Wu¨rzburg, Germany, and Institute of Hygiene and Microbiology, University of Wu¨rzburg,
J osef-Schneider-Str. 2, D-97080 Wu¨rzburg, Germany
Received November 4, 1999
A Bacillus megaterium strain was isolated from topsoil by a selective screening procedure with
allylbenzene as a xenobiotic substrate. This strain performed the hydroxylation chemoselectively
(no arene oxidation and overoxidized products) and enantioselectively (up to 99% ee) in the benzylic
and nonbenzylic positions of a variety of unfunctionalized arylalkanes. Salycilate and phenobarbital,
which are potent inducers of cytochrome P-450 activity, changed the regioselectivity of the microbial
CH insertion, without an effect on the enantioselectivity. The biotransformation conditions were
optimized in regard to product yield and enantioselectivity by variation of the oxygen-gas supply
and the time of the substrate addition. The different product distributions (R- versus â-hydroxylated
product) that are obtained on induction of cytochrome P-450 enzyme activity demonstrate the
involvement of two or more hydroxylating enzymes with distinct regioselectivities in this
biotransformation. An oxygen-rebound mechanism is assumed for the cytochrome P-450-type
monooxygenase activity, in which steric interactions between the substrate and the enzyme
determine the preferred face of the hydroxy-group transfer to the radical intermediate.
In tr od u ction
tivity.9-11 Despite all these efforts, the development of
an effective catalytic asymmetric hydroxylation of un-
functionalized hydrocarbons with broad applicability
remains a challenge.
CH oxidations belong to the most versatile reactions
in organic chemistry, since potentially useful oxyfunc-
tionalized synthons may be obtained from readily acces-
sible hydrocarbons. Only a few conventional synthetic
methods, which require harsh reaction conditions and
yield racemic products, are to date available for this
purpose.1 Therefore, much effort has been previously
expended to develop efficient, catalytic oxidations of
hydrocarbons.2-5 For example, in the past years, metal-
catalyzed asymmetric CH oxidations have been per-
formed by employing chiral auxiliaries6 and optically
active oxidants7,8 to afford enantiomerically enriched
products. Nevertheless, most of the time only a moderate
enantioselectivity has been achieved for unactivated
alkanes. In this context, biomimetic studies have been
carried out to gain insight into the mechanism of the
catalytic CH oxidation and to enhance its stereoselec-
Alternatively, microorganisms have been successfully
applied to the selective oxygenation of unactivated CH
bonds in organic substrates.12 So far, the major emphasis
of such work has focused on the hydroxylation of steroids,
terpenes and other complex natural products.13 For these
biotransformations, mostly fungi like Rhizopus nigricans,
Mortiella isabellina, and Cunninghamella elegans, for
instance, have been employed,14,15 for which high regio-
and enantioselectivities have been reported.16-19 In con-
trast, the hydroxylation of xenobiotics is far more scarce
because, unlike natural products, these compounds are
not readily introduced into the cell for metabolism.
Furthermore, many monooxygenases, particularly the
well-known cytochrome P-450cam from Pseudomonas
putida, are highly substrate-specific, and therefore, their
* To whom correspondence should be addressed. E-mail: adam@
chemie.uni-wuerzburg.de. Fax: +49-931-8884756. http://www-organik.
chemie.uni-wuerzburg.de.
(9) Lim, M. H.; Lee, Y. J .; Goh, Y. M.; Nam, W.; Kim, C. Bull. Chem.
Soc. J pn. 1999, 72, 707-713.
(10) Breslow, R. Acc. Chem. Res. 1995, 28, 146-153.
(11) Moro-oka, Y.; Akita, M. Catal. Today 1998, 41, 327-338.
(12) J ohnson, R. A. In Organic Chemistry-Part C-Oxidation in
Organic Chemistry; Trahanowsky, W. S., Ed.; Academic Press: New
York, 1978.
(13) Holland, H. L. In Biotechnology; Kelly, D. R., Ed.; Wiley-VCH:
Weinheim, 1998; Vol 8a.
(14) Holland, H. L. Organic Synthesis with Oxidative Enzymes; VCH
Publishers: Weinheim, 1992.
(15) Davis, C. R.; J ohnson, R. A.; Cialdella, J . I.; Liggett, W. F.;
Mizsak, S. A.; Marshall, V. P. J . Org. Chem. 1997, 62, 2244-2251.
(16) Holland, H. L. Curr. Opin. Chem. Biol. 1999, 3, 22-27.
(17) Berg, A.; Rafter, J . J . Biochem. J . 1981, 196, 781-786.
(18) Schwab, E.; Bernreuther, A.; Puapoomachareon, P.; Mori, K.;
Schreier, P. Tetrahedron: Asymmetry 1991, 2, 471-479.
(19) Davies, H. G.; Green, R. H.; Kelly, D. R.; Roberts, S. M.
Biotransformations in Preparative Organic Chemistry: The Use of
Isolated Enzymes and Whole Cell Systems in Synthesis; Academic
Press: London, 1989.
† Institute of Organic Chemistry.
‡ Institute of Food Chemistry.
§ Institute of Hygiene and Microbiology.
(1) Haines, A. H. Methods for the Oxidation of Organic Compounds
- Alkanes, Alkenes, Alkynes and Arenes; Academic Press: London,
1985.
(2) Stahl, S. S.; Labinger, J . A.; Bercaw, J . E. Angew. Chem., Int.
Ed. Engl. 1998, 37, 2180-2192.
(3) Barton, D. H. R.; Doller, D. Acc. Chem. Res. 1992, 25, 504-512.
(4) Reiser, O. Angew. Chem., Int. Ed. Engl. 1994, 33, 69-72.
(5) Adam, W.; Curci, R.; D’Accolti, L.; Dinoi, A.; Fusco, C.; Gaspar-
rini, F.; Kluge, R.; Paredes, R.; Schulz, M.; Smerz, A. K.; Veloza, L. A.;
Weinkoetz, S.; Winde, R. Chem.sEur. J . 1997, 3, 105-109.
(6) J acobsen, E. N.; Marko, I.; Mungall, W. S.; Schro¨der, G.;
Sharpless, K. B. J . Am. Chem. Soc. 1988, 110, 1968-1970.
(7) Davis, F. A.; Vishwakarma L. C.; Billmers, J . M.; Finn, J . J . Org.
Chem. 1984, 49, 3241-3243.
(8) Hamada, T.; Irie, R.; Mihara, J .; Hamachi, K.; Katsuki, T.
Tetrahedron 1998, 54, 10017-10028.
10.1021/jo991725s CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/11/2000

Asymmetric Hydroxylation of Hydrocarbons
J . Org. Chem., Vol. 65, No. 3, 2000 879
application is restricted to a limited number of closely
related substrates.12
In the case of simple arylalkanes, the substrates of
interest in the present work, the oxidation of the aromatic
ring or benzylic hydroxylation is usually favored accord-
ing to literature reports.20 For instance, P. putida hy-
droxylates p-ethylphenol to yield the corresponding R-hy-
droxy phenol in an enantiomeric excess (ee) of 39% in
favor of the S enantiomer,14 but Pseudomonas aeruginosa
fails to hydroxylate the sterically more demanding isobu-
tylbenzene.12 Another severe drawback in whole-cell CH
insertions is the overoxidation of the alcohol to either the
ketone or even the carboxylic acid.12 Moreover, most of
the currently known biocatalytic hydroxylations of xe-
nobiotics with whole cells suffer from low conversions and
poor enantioselectivities.
In view of these shortcomings, the purpose of the
present study was to search for microorganisms that
display both high chemo- and enantioselectivity, i.e.,
those that deliver optically pure oxyfunctionalized prod-
ucts without further oxidation. In this contribution, we
present our results for the hydroxylation of arylalkanes
by a Bacillus megaterium strain, which was obtained
from topsoil by a selective screening procedure. We
demonstrate herein for the first time that this microor-
ganism may be used for the oxidative biotransformation
of the unactivated arylalkanes 1a -g to yield selectively,
even on the semipreparative (mg) scale, the correspond-
ing benzylic and nonbenzylic alcohols (no arene oxidation)
in high enantiomeric excess and with low amounts, if any,
of overoxidation products.
F igu r e 1. Growth curves of B. megaterium in the presence
of different amounts of the model substrate 1e.
quence data of the European Molecular Biology Labora-
tory (EMBL) database, which disclosed that an Arthro-
bacter sp., a Pseudomonas sp., and a Bacillus sp. had
been isolated. The Bacillus strain yielded the R-hydroxy-
lation product (benzylic CH insertion, no epoxidation nor
phenyl hydroxylation) of allylbenzene, whereas the other
strains were ineffective for oxidative biotransformations.
Therefore, the effective strain was more rigorously
characterized and identified as a B. megaterium strain.
In an effort to optimize the biotransformation condi-
tions, the growth curve for B. megaterium was recorded
for different substrate amounts, which were added during
the early log-phase growth. For all experiments, isobu-
tylbenzene (1e) was chosen as model substrate (Scheme
1).
Sch em e 1. Hyd r oxyla tion of Mod el Su bstr a te 1e
by B. m ega ter iu m
Resu lts
An adequate screening procedure to select suitable soil
bacteria for the biotransformation of hydrocarbons had
to be developed. Allylbenzene was chosen as the selecting
substrate because it possesses three different oxidation
sites for its detoxification as a biologically harmful
substance: the phenyl ring, the benzylic position, and
the double bond. Applying various amounts (5-25 µL)
of allylbenzene and glucose (0-10 g/L) to soil samples of
different origin, this screening procedure resulted in
three different microbial species. For their identification,
they were subjected to sequencing of the first 300 base
pairs of their small subunit ribosomal RNA gene (16s)
by means of the Taq-cycle-DyeDeoxy-terminator tech-
nique.21 The resulting nucleotide sequences (variable
region V1-V3 of 16s-like-rRNA) were compared to se-
Addition of 67 µmol of the substrate, which is ap-
proximately equivalent to the amount added during
regular biotransformations, did not influence the growth
in comparison to a culture without any substrate added
(Figure 1). However, when the amount of substrate was
increased to 224 µmol, cellular growth was immediately
slowed and the culture rapidly died due to substance
toxicity. Addition of the substrate at different times
during growth led to significantly different yields of the
oxidation products. Thus, substrate addition during the
early log-phase, as well as early stationary phase,
resulted in approximately up to 4-fold less oxidation
product than during the mid-log-phase addition. There-
fore, the latter conditions were chosen as the optimal time
for the addition of the substrate in regard to yield (data
not shown).
(20) Sibbesen, O.; Zhang, Z.; Ortiz de Montellano, P. R. Arch.
Biochem. Biophys. 1998, 353, 285-296.
(21) Harmsen, D.; Rothga¨nger, J .; Singer, C.; Albert, J .; Frosch, M.
Lancet 1999, 353, 291.
The influence of the oxygen-gas supply on the hydroxy-
lation reaction was elucidated for the further optimiza-

880 J . Org. Chem., Vol. 65, No. 3, 2000
Adam et al.
F igu r e 2. Dependence of the extent of hydroxylation for equal amounts of substrate 1e by B. megaterium on the oxygen availability;
a
oxygen-gas influx depends on the rotation velocity (rpm). No product was observed.
Ta ble 1. Regio- a n d En a n tioselectivities of th e
tion of the biotransformation conditions. The availability
Oxygen -Atom In ser tion in to th e C-H Bon d of
of oxygen gas in the growth medium was varried by
Hyd r oca r bon s 1 by B. m ega ter iu m
changing the rotation velocity of the incubation vessel.
Clearly, no products were detected at low (90 rpm)
oxygen-gas saturations (Figure 2). In contrast, relatively
high oxygen-gas influx increased the hydroxylation at the
R position ca. 8-fold, whereas the amount of â product
was merely doubled.
To elucidate the scope of substrate acceptability by B.
megaterium, several hydrocarbons were used for the
biotransformation (Table 1). B. megaterium was incu-
bated with ca. 100 µmol of the respective hydrocarbon
for 17 h, the cells were removed by centrifugation, and
liquid-liquid extraction of the growth medium was
carried out for product isolation. Because the products
and the substrates are quite volatile, the solvent was
removed by distillation through a Vigreux column to
prevent loss due to evaporation. Finally, the products
were submitted to gas-chromatographic analysis and
mass spectrometry. The enantiomeric excess was deter-
mined by multidimensional gas chromatography (MDGC)
on cyclodextrin columns. The absolute configurations
were established by comparison with authentic reference
samples, wherever available, and by employing the
circular-dichroism-exciton-chirality method22 (details will
be published separately).
Ethylbenzene (1a ) and propylbenzene (1b) both yielded
solely the R-hydroxylation product. For 1-phenylethanol
(2a ), an enantiomeric excess of only 19% for the R alcohol
was observed (entry 1), whereas 1-phenylpropanol (2b)
displayed a much higher ee value of 74% for the R
a
Conversion of hydrocarbons determined by GC analysis; error
enantiomer (entry 2). Allylbenzene (1c), the screening
limit ( 2% of the stated values. bRatio of regioisomeric alcohols;
substrate, was hydroxylated to the corresponding alcohol
(R)-2c with an ee value of 70% under preservation of its
in the case of ketone formation, the product distribution is
normalized to 100%; the amount of the corresponding ketone is
given in parentheses. cEnantiomeric excesses (ee values) were
determined by multidimensional gas chromatography on a cyclo-
dextrin column. dSmall amounts (ca. 4%) of cinnamyl alcohol were
observed.
double bond, which is an unexpected feature of this
biotransformation since the structurally related cinnamyl
alcohol was reduced by B. megaterium to the 3-phenyl-
propanol (data not shown). Mechanistically significant
is the fact that ca. 4% cinnamyl alcohol was detected.
The enantiomeric excess of the allyl alcohol 2c is com-
parable to the one with propylbenzene (1b) as substrate
(entry 3). A tertiary butyl group in the side chain
(substrate 1d ) increased the enantiomeric excess to 91%
of the R enantiomer; however, the conversion drops to
18% due to the greater steric hindrance (entry 4).
Branching in the propyl chain results additionally in
â-hydroxylation products. Indeed, for isobutylbenzene
(1e), the 2eâ regioisomer is even preferentially formed
(entry 5). This is highly unusual, because many of the
reported bacterial hydroxylations of arylalkanes in the
aliphatic side chain are restricted to the benzylic posi-
(22) Harada, N.; Nakanishi, K. Circular Dichroic Spectroscopy -
Exciton Coupling in Organic Chemistry; University Science Books: Mill
Valley, 1983.

Asymmetric Hydroxylation of Hydrocarbons
J . Org. Chem., Vol. 65, No. 3, 2000 881
selective aliphatic CH insertions without epoxidation or
phenyl hydroxylation. For the unfunctionalized hydro-
carbons 1a -g, B. megaterium gave the usually largely
favored microbial benzylic hydroxylation,14 but also the
nonbenzylic positions were hydroxylated in large amounts.
The use of known inducers or repressors of cytochrome
P-450 activity revealed that at least two distinct enzymes
are involved, one favors the benzylic hydroxylation, the
other the nonbenzylic one (Figure 3). Thus, the potent
inducers of cytochrome P-450 activity significantly af-
fected the product distribution of this biotransformation.
Salicylate selectively increased the amount of the R-hy-
droxylated product, while the barbiturate enhanced the
â hydroxylation, in comparison to the regioselectivity
found under standard conditions (no additive, Figure 3).
If only one monooxygenase were responsible for all
hydroxylations, then enhancement of its expression
would increase all the regioisomers, but the ratio of the
products would remain unchanged. The involvement of
more than one enzyme is also suggested by a change in
the oxygen supply of the medium (Figure 2). An increase
in oxygen availability (180 rpm) shifts the ratio of R
versus â product in the hydroxylation of the model
substrate 1e toward the benzylic position (proportionally
more regioisomer 2eR), since different oxygen concentra-
tions in the medium affect the expression of the distinct
hydroxylating enzymes on a transcriptional level and
shift the product pattern of regioisomers towards the
favored hydroxylation position of that particular enzyme
which is transcribed preferentially. This means that for
the model substrate 1e, the observed product pattern is
a result of the different cellular availability of hydroxy-
lating enzymes with distinct regioselectivities and not a
result of different selectivities of a single enzyme. In that
context, it is interesting to note that cytochrome P-450
BM-3, a monooxygenase described for B. megaterium, is
one of the few prokaryotic cytochrome P-450 enzymes
inducible by barbiturates.24,26 The known properties of
this enzyme are the ω-1, ω-2, and ω-3 hydroxylation of
fatty acids of chainlengths from C12 to C18, fatty alcohols
and fatty amides, with the requirement of a polar group
in the substrate to hold the methylene group to be
hydroxylated close to the active center.27,28 Since the
enzymatic activity observed in the present case is also
inducible by a barbiturate, but the type of substrate is
completely different, it seems likely that the biotrans-
formations described herein are carried out by cyto-
chrome P-450 mutants, which readily accept purely
hydrophobic substrates.
F igu r e 3. Effect of cytochrome P-450 inducers or repressors
(additives) on the hydroxylation of equal amounts of the
hydrocarbon 1e by B. megaterium.
tion.14 The enantiomeric excess of 2-methyl-1-phenylpro-
pan-1-ol (2eR) was found to be 91% for the R enantiomer.
In contrast, (1-methyl)propylbenzene (1f) yielded the
R-hydroxylation product 2fR in 38% ee for the R enan-
tiomer, as well as the threo (39% ee) and erythro (83%
ee) diastereomers of the â-hydroxylation product 2fâ
(entry 6). The erythro (2R,3R)-1-methyl-1-phenyl-2-pro-
panol (2fâ) eluted significantly later than the other two
alcohols. It is mechanistically important for this C-H
insertion that the chiral starting material 1f remained
racemic throughout this biotransformation.
n-Pentylbenzene (1g) is especially noteworthy, as it
gave all possible methylene-hydroxylation products, but
no terminal oxidation of the side chain. The regioselec-
tivity shows once more no preference for the benzylic
position since the R- and â-regiomeric alcohols were
obtained in approximately equal amounts. CH insertion
at the γ and δ positions is significantly less, with a slight
preference for the δ-methylene group (entry 7). For all
regioisomers, the respective ketone products were formed
in small amounts (3-5%) due to further oxidation of the
alcohols (Table 1, entry 7, values in parentheses). The
enantiomeric excess is very high except for the R-hy-
droxylation product 2gR (42% ee). The R and δ hydroxy-
lations of substrate 1g by B. megaterium proceeded with
a preference for the S enantiomers, whereas the â and γ
insertions afforded the R enantiomers.
To assess whether monooxygenases of the cytochrome
P-450 type participate in these microbial transforma-
tions, known inducers and repressors of such enzyme
activity were tested on substrate 1e (Figure 3). Addition
of salicylate23 to the growth medium effectively induced
R hydroxylation, whereas â hydroxylation was sharply
reduced. Phenobarbital, a barbiturate,24 induced both R
and â hydroxylation, but with a more significant increase
of the â product. Ancymidol,25 a cytochrome P-450
repressor, reduced the yields of both products, but more
effectively that of the â hydroxylation. Therefore, the
involvement of cytochrome P-450 enzymes seems to be
likely.
The enzymes involved in this oxyfunctionalization
display a high enantioselectivity which is strongly de-
pendent on the steric properties of the substrate (Table
1). This is clearly evident for substrates with branching
of the aliphatic side chain (Table 1, entries 4 and 5). A
sterically more demanding alkyl chain poses greater
hindrance to the transfer of the hydroxy group from the
pro-S face of the substrate and, thus, favors the insertion
from the pro-R face. For instance, substrate 1d with a
tert-butyl group and 1e with an isopropyl group are
hydroxylated in an enantiomeric excess of 91%, in favor
of the R enantiomers (Table 1, entries 4 and 5). The
Discu ssion
The topsoil screening process with the unfunctionalized
allylbenzene (1c) delivered a B. megaterium strain for
(26) Liang, Q.; Chen, L.; Fulco, A. J . Biochim. Biophys. Acta 1998,
1380, 183-197.
(27) Cytochrome P-450, 2nd ed.; Omura, T., Ishimura, Y., Fujii-
Kuriyama, Y., Eds.; VCH Publishers: New York, 1993.
(28) Munro, A. W.; Lindsay, J . G. Mol. Microbiol. 1996, 20, 1115-
1125.
(23) Shaw, G.; Kao, H.; Chiang, A. Curr. Microbiol. 1997, 35, 28-
31.
(24) Narhi, L. O.; Fulco, A. J . J . Biol. Chem. 1986, 261, 7160-7169.
(25) Fuchser, J .; Zeeck, A. Liebigs Ann. Chem. 1997, 87-95.

882 J . Org. Chem., Vol. 65, No. 3, 2000
Adam et al.
of the resulting radical. In addition to the crucial steric
effects, the enantioselectivity is attenuated by hydropho-
bic interactions between the enzyme and the substrate
(Figure 4). This is evident from the sharp increase in the
enantioselectivity (Table 1, entries 1 and 2) for the
hydroxylations of ethylbenzene (1a ) compared to propyl-
benzene (1b). The hydrophobic attraction between the
longer alkyl chain and the appropriate amino acid (in
proximity to the prosthetic group) blocks the transfer of
the hydroxy group to the pro-S face of the radical
intermediate and properly aligns the substrate in the
active center of the enzyme to facilitate the hydroxy-
group transfer to the pro-R face (Figure 4). After the
hydroxy group is transferred, the original iron state is
regenerated and reenters the catalytic cycle.
In regard to the sense in the R versus S selectivity,
the biotransformation of pentylbenzene (1g) provides an
unusual dependence for the four possible regioisomeric
hydroxylation products (Table 1, entry 7). Thus, while
the regioisomers 1-phenyl-1-pentanol (2gR) and 5-phenyl-
2-pentanol (2gδ) are S-configured, for the 2gâ and 2gγ
regioisomers the R enantiomers are formed, the usual
enantiomer obtained for all the other substrates em-
ployed (Table 1). This distinguished behavior cannot be
explained in terms of kinetic resolution of the chiral
alcohols through enantioselective oxidation, as the
amounts of ketones found are too low to influence the
enantiomeric excess of the alcohols. Therefore, in the case
of (S)-alcohol formation, either different enzymes are
involved or the substrate with the longer alkyl chain is
positioned with the opposite orientation in the enzyme
cavity to result in the observed reversed enantioselec-
tivity.
In conclusion, we have shown that the B. megaterium
strain employed herein as a nonpathogenic soil microor-
ganism readily hydroxylates a variety of unfunctionalized
arylalkanes selectively at the benzylic as well as non-
benzylic positions, but not the phenyl ring, to yield the
corresponding regiomeric alcohols in high enantiomeric
excesses (up to 99% ee). Additionally, the regioselectivity
of the oxidation may be altered by employing different
inducers, under conservation of the absolute configura-
tion with high enantioselectivities in the hydroxylated
products. Furthermore, it was demonstrated that the
stereochemical course of the hydroxylation is determined
by steric effects between the enzyme and the substrate.
Since for this environmentally benign process the regio-
selectivity may be conveniently adapted to suit different
synthetic goals, further applications in organic synthesis
seem promising and worthwile for the preparation of
optically active oxyfunctionalized building blocks from
readily available simple hydrocarbon substrates.
F igu r e 4. Oxygen-rebound mechanism for the C-H insertion
of unfunctionalized hydrocarbons 1a -g by B. megaterium
monooxygenases, presumably of the cytochrome P-450 type;
the substrate (R ) alkyl) is aligned in the active center of the
enzyme by hydrophobic interaction with amino acids (R1, R2
) alkyl or aryl), and the hydroxy-group transfer to the radical
site is directed through the alignment of the substrate 1 in
the active center by the steric interactions between the alkyl
and aryl groups of the substrate and enzyme.
influence of the three-dimensional structure of the mol-
ecule on the enantiomeric excess is further demonstrated
in the hydroxylation of (1-methyl)propylbenzene (1f). The
oxidation products 2fR and threo-2fâ that display com-
parable chromatographic properties are both obtained
with low ee values, whereas the significantly later eluting
erythro-2fâ diastereomer is formed in much higher enan-
tiomeric excess (Table 1, entry 6). Therefore, the three-
dimensional fit of the substrate in the active center of
the enzyme is decisive in the control of the selectivity
for this biotransformation. If the substrate-enzyme
interaction were relatively loose, then attack from either
prochiral face of the substrate would be equally facile,
since it could position itself randomly in the enzyme
cavity. To achieve high enantioselectivities, besides steric
effects, presumably alignment of the substrate in the
active center of the enzyme by hydrophobic interactions
with appropriate amino acids, e.g., alanine or phenyla-
lanine, appear to be involved (Figure 4).
For the observed monooxygenase activity in the CH
insertion we assume the established catalytic cycle for
cytochrome P-450 enzymes, in which the oxygen-rebound
mechanism applies.28,29 Mechanistically important is the
fact that the alcohol products 2fR and 2fâ of the chiral
substrate 1f (Table 1, entry 6) were found in moderate
to high enantiomeric excesses, whereas the recovered 1f
remained racemic after the biotransformation. The lack
of kinetic resolution in this oxidation signifies that the
abstraction of the H atom by the oxoiron complex (ferryl
intermediate) proceeds without preference of one of the
substrate enantiomers. It may, therefore, be concluded
that in the initial step of the CH insertion a carbon-
centered radical is formed without discrimination be-
tween the different prochiral faces of the starting mate-
rial. The small amount of cinnamyl alcohol, which was
observed as a byproduct of the hydroxylation of allylben-
zene (1c), lends support that radical intermediates
intervene in this oxidation. In the final step, the actual
oxygen insertion, the hydroxy group is transferred from
the iron complex preferentially to the less hindered side
Ack n ow led gm en t. We express our gratitude to Dr.
U. Rdest (Institute of Microbiology, University of
Wu¨rzburg, Germany) for her valuable help and advice.
Furthermore, we thank the Deutsche Forschungsge-
meinschaft (SFB 347 “Selektive Reaktionen Metall-
aktivierter Moleku¨le”) and the Fonds der Chemischen
Industrie for their generous financial support.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures with bacterial screening protocol, biotransformation
conditions, and MDGC columns for the alcohols 2. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(29) Wong, L.-L. Curr. Opin. Chem. Biol. 1998, 2, 263-268.
J O991725S