Engineering cytochrome P450 BM3 for terminal alkane hydroxylation

Article abstract of DOI:10.1002/adsc.200505465

Enzymes that catalyze the terminal hydroxylation of alkanes could be used to produce more valuable chemicals from hydrocarbons. Cytochrome P450 BM3 from Bacillus megaterium hydroxylates medium-chain fatty acids at subterminal positions at high rates. To engineer BM3 for terminal alkane hydroxylation, we performed saturation mutagenesis at selected active-site residues of a BM3 variant that hydroxylates alkanes. Recombination of beneficial mutations generated a library of BM3 mutants that hydroxylate linear alkanes with a wide range of regioselectivities. Mutant 77-9H exhibits 52% selectivity for the terminal position of octane. This regioselectivity is octane-specific and does not transfer to other substrates, including shorter and longer hydrocarbons or fatty acids. These results show that BM3 can be readily molded for regioselective oxidation.

Full text of DOI:10.1002/adsc.200505465

FULL PAPERS
DOI: 10.1002/adsc.200505465
Engineering Cytochrome P450 BM3for Terminal Alkane
Hydroxylation
Peter Meinhold,a Matthew W. Peters,a Adam Hartwick,a Alisha R. Hernandez,a, b
Frances H. Arnolda,*
a
California Institute of Technology, Division of Chemistry and Chemical Engineering, Mail Code 210-41, Pasadena, CA
91125, USA
Fax: (þ1)-626-568-8743, e-mail: frances@cheme.caltech.edu
b
Current address: Colorado College, Department of Chemistry, 1040 N. Nevada Ave, Colorado Springs, CO 80903, USA
Fax: (þ1)-719-389-6182, e-mail: a_hernandez@coloradocollege.edu
Received: December 1, 2005; Accepted: February 7, 2006
Supporting Information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: Enzymes that catalyze the terminal hydrox- lectivities. Mutant 77-9H exhibits 52% selectivity for
ylation of alkanes could be used to produce more val- the terminal position of octane. This regioselectivity
uable chemicals from hydrocarbons. Cytochrome is octane-specific and does not transfer to other sub-
P450 BM3 from Bacillus megaterium hydroxylates strates, including shorter and longer hydrocarbons
medium-chain fatty acids at subterminal positions at or fatty acids. These results show that BM3 can be
high rates. To engineer BM3 for terminal alkane hy- readily molded for regioselective oxidation.
droxylation, we performed saturation mutagenesis at
selected active-site residues of a BM3 variant that hy-
droxylates alkanes. Recombination of beneficial mu- Keywords: alkanes; biocatalysis; cytochrome P450
tations generated a library of BM3 mutants that hy- BM3; directed evolution; high throughput screening;
droxylate linear alkanes with a wide range of regiose- terminal hydroxylation
Introduction
out the hydroxylation reaction[5–7] at the terminal alkane
carbon exclusively, at rates of ~200 minÀ1 [8]
Cytochromes P450 catalyze the insertion of oxygen
.
Efficient routes for regio- and stereoselective hydroxy-
lation of non-activated carbon atoms are sought after into a wide variety of substrates, including n-alkanes,
for production of higher value chemicals from alkanes.[1] via a reactive iron-oxo species chelated by a porphyrin.
Synthetic catalysts are plagued by low selectivity leading Some members of the CYP4[9,10] CYP52,[11–14] CYP86,[15]
to overoxidation[2] and often require harsh reaction con- and CYP153 families[16] hydroxylate terminal methyl
ditions. Enzymes, in contrast, can catalyze the selective groups of fatty acids or n-alkanes. Like AlkB, these en-
oxidation of alkanes to alcohols at room temperature zymes are membrane-bound and contain multiple pro-
and pressure, using dioxygen or peroxide as the oxidant tein components, which complicates their structural
and producing no side products other than water. The and kinetic characterization and makes them difficult
first step of alkane degradation by alkane-assimilating to use in engineered pathways. (Only recently has a solu-
microorganisms involves selective hydroxylation of the ble, terminal alkane hydroxylating cytochrome P450,
alkane at the terminal carbon. Biochemical and genetic classified as CYP153A6, been reported.[17]) The report-
studies have focused on a limited number of enzymes ed rates on alkanes are relatively low. Alkane oxidation
that catalyze this reaction,[3] the best known of which is by CYP52A3 from alkane-assimilating Candida malto-
the Pseudomonas putida GPo1 alkane hydroxylase.[4] sa, for example, proceeds at a rate of 27 minÀ1 for hexa-
Enabling growth on C6 to C12 n-alkanes, this enzyme sys- decane.[14] Similar rates have been reported for mamma-
tem consists of an alkane hydroxylase (AlkB), rubre- lian P450s such as CYP4B1 (33 minÀ1, 26 minÀ1
doxin (AlkG), and rubredoxin reductase (AlkT). 31 minÀ1, 11 minÀ1, 67 minÀ1 for heptane, octane, non-
AlkG and AlkTtransfer electrons to AlkB, a non-
ane, decane and 4-methylheptane, respectively).[10] No
heme diiron integral membrane protein that carries crystal structures are available for any of these P450s.
,
Adv. Synth. Catal. 2006, 348, 763 – 772
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
763

Peter Meinhold et al.
FULL PAPERS
Cytochrome P450 BM3 (CYP102A1) from Bacillus tect by their in situ chemical conversion to dye-sensitive
megaterium hydroxylates medium-chain (C12 to C20) fat- compounds such as aldehydes and ketones. Spectro-
ty acids at extraordinarily high rates of up to scopic monitoring of NADPH consumption rates in
17,100 minÀ1 [18]
Oxygen is inserted at subterminal (w- the presence of the alkane is not useful either, because
.
1, w-2, and w-3) positions in approximately equal it often selects for mutants with NADPH oxidation rates
amounts (the exact regioselectivity varies with the chain that are uncoupled to product formation. We circum-
length of the substrate).[19] BM3 is a soluble enzyme vented these issues in previous work by using surrogate
whose hydroxylase and reductase domains are con- substrates that mimic desired substrates but yield prod-
tained in a single polypeptide. Expressed in functional ucts that are more easily detected. For example, we used
form at high levels in Escherichia coli, this enzyme can dimethyl ether (DME) to screen for mutants more ac-
be incorporated into engineered pathways for whole tive on propane.[21] Here we chose hexyl methyl ether
cell biotransformations[20] orcanbeusedin vitro for selec- (HME) as a surrogate for octane. HME shares most of
tive oxidations.[21,22] Attracted by the possibility of trans- the physical properties of octane, including solubility
À
ferring the high rate of fatty acid hydroxylation to other in buffer, size, shape, and C H bond strength. This
substrates, various groups have engineered BM3 sub- screen is also sensitive to terminal hydroxylation. As
strate specificity[23–28] as well as regio- and enantioselec- shown in Figure 1, hydroxylation at the HME methoxy
tivity.[21,29–31] In previous work, our group converted carbon produces an equivalent of formaldehyde, which
BM3 into a fast and productive alkane hydroxylase with is detected at concentrations as low as 10 mM with the
specificity for different subterminal carbons.[21,32,33] T o ex- dye Purpald.[40] Hydroxylation at the a-carbon, on the
tendthespectrumofBM3-catalyzedreactions we wanted other hand, produces hexanal, which is practically un-
to engineer BM3 for terminal alkane hydroxylation.
reactive towards Purpald at the concentrations encoun-
Protein engineering to obtain high selectivity for ter- tered in the screens (formaldehyde is approximately 42
minal hydroxylation is made challenging by the fact times more sensitive to Purpald than is hexanal). Be-
that BM3 undergoes large conformational changes dur- cause using a surrogate screen always runs the risk of
ing catalysis[34] that are not fully captured by studying the identifying mutants that are more active on the surro-
crystal structures of the heme domain in its substrate- gate but not the desired substrate, it is important to ver-
bound and substrate-free forms.[35–39] Directed evolu- ify that activity on the surrogate (HME) correlates with
tion involving sequential rounds of mutagenesis and activity on the desired substrate (octane) (see below).
high throughput screening has therefore been our meth-
od of choice. Laboratory-evolved BM3 alkane hydroxy-
lase mutant 9-10A[21] contains 14 mutations in the heme Saturation Mutagenesis of Cytochrome P450 BM39-
domain, only one of which, V78A, is located in the active 10A and Recombination of Beneficial Mutations
site. We recently described two other active-site muta-
tions, A82L and A328V, that shifted the regioselectivity Changing substrate specificity or regioselectivity often
of alkane hydroxylation predominantly to a single, sub- requires multiple mutations within the active site,
terminal (w-1) position.[21] Prompted by these results, we some of which may be coupled. Such mutations are
decided to introduce further mutations in the active site not likely to be found by random mutagenesis of the
in order to achieve the desired selectivity for terminal full protein.[41,42] We therefore used the BM3 heme do-
hydroxylation. Starting with BM3 mutant 9-10A, we main crystal structures with the bound substrates palmi-
performed saturation mutagenesis on eleven residues toleic acid[36] and N-palmitoylglycine[38] to identify 11
that should be in close proximity to the substrate. We as- residues with side chains that are within 5 Š of the termi-
sayed these mutants using a screen that is sensitive to the nal eight carbons of the alkyl chains of the bound sub-
regioselectivity of hydroxylation. Beneficial mutations strates: A74, L75, V78, F81, A82, F87, T88, T260, I263,
were recombined to generate a library of ~9,000 func- A264, and A328. Mutations at six of these positions
tional BM3 mutants, from which mutants exhibiting a (74, 78, 82, 87, 264, and 328) were previously known to
range of regioselectivities for alkane hydroxylation influence the activity of BM3. For example, changes to
have been isolated.
F87 in wild-type BM3 affect the distributions of hy-
droxylated fatty acid products[30,43–45] and activity to-
wards non-natural substrates.[29] Residues at positions
74, 87, and 264 have been altered in combination with
other mutations to increase the activity of BM3 towards
a variety of substrates, including octane and polyaro-
matic compounds.[23,25,26,46,47] Mutation of A328 to valine
in the wild-type enzyme has been reported to affect sub-
Results and Discussion
Screening for Regioselective Alkane Hydroxylation
Screening enzyme libraries for alkane hydroxylation is strate binding and activity on fatty acids.[48] We generat-
challenging because the desired (alcohol) products are ed 11 saturation mutagenesis libraries in which all twen-
colorless and are formed at concentrations too low tode- ty amino acids were introduced at each of the 11 identi-
764
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 763 – 772

Engineering Cytochrome P450 BM3 for Terminal Alkane Hydroxylation
FULL PAPERS
Figure 1. a) Screening for terminal alkane hydroxylation using hexyl methyl ether (HME). Terminal (w) oxidation of HME
results in the formation of formaldehyde, which turns dark purple in the presence of Purpald. Hydroxylation of the w-2 carbon
forms hexanal, which turns pink upon reaction with Purpald. Other hydroxylation products do not react with Purpald. b) The
Purpald assay is much more sensitive to formaldehyde than to hexanal.
fied positions in 9-10A. From each library, 174 clones were active towards DME or HME (data not shown). A
were screened for activity towards DME and HME.
total of 16 mutants with improved activity on DME or
Mutations that improved activity towards HME or HME compared to the best BM3 alkane hydroxylase
DME were found at seven positions: L75 (W), A78 to date, mutant 1-12G,[21] were selected for purification
(T, F), A82 (T, S, F, I, C, G), F87 (I, V, L), T88 (C), and further characterization. The active-site amino
T260 (N, L) and A328 (L, F, M) (see Table S1 in Support- acid substitutions and activities on DME and HME for
ing Information). The improved enzymes exhibited dif- four of these mutants are presented in Table 2.
ferent levels of activity and regio- and enantioselectivity
for hydroxylation of octane (Table 1). The primary
products are 2-, 3-, and 4-octanol, but limited amounts Correlation between Activity on HME and Terminal
of 1-octanol are also produced (typically <10% of the Hydroxylation of Octane
total products). Only mutations at position 328 lead to
fractions of 1-octanol higher than 10%: A328L and Rates of product formation, total turnover numbers
A328M produce 13 and 24% 1-octanol, respectively.
(ttn), and regioselectivities for hydroxylation of octane
The beneficial mutations of Table 1 and the wild-type were determined for the mutants in Table 2. The total
residues at these positions were recombined by overlap turnover number describes the number of turnovers
extension PCR with degenerate primers or a mixture of the enzyme catalyzes before it is inactivated. Rates
primers to generate a library containing all possible and mechanism of inactivation were not further exam-
combinations of these mutations. This relatively in- ined for this report.
volved strategy was favored over other DNA recombi-
Mutant 53-5H, which shows increased activity on
nation techniques such as DNA shuffling,[49] because it DME but not on HME (Table 2), exhibits the least ter-
does not bias recombination events towards those sepa- minal hydroxylation (3% of the product is 1-octanol)
rated by longer stretches of DNA. Mutation L75I, which but produces 2-octanol with high selectivity (87%), sim-
affects regioselectivity but not activity, was also included ilar to 1-12G. The mutants that show increased HME ac-
in the recombination, as were previously known benefi- tivity exhibit a range of terminal hydroxylation capabil-
cial mutations A82L and A328V.[21] Mutation A82V was ities, with 1-octanol accounting for as much as 52% of
included as well, due to constraints in the design of the the product for 77-9H. In general, mutants exhibiting
degenerate primer used to construct the library. The re- high total turnover on HME in the screen also produced
sulting recombination library contains approximately more 1-octanol. The total amount of 1-octanol formed
9,000 different mutants of P450 BM3, each of which con- (% 1-octanolÂttn) correlates with the amount of
tains a characteristic active site. More than 70% of these HME converted for the different BM3 mutants (Fig-
Adv. Synth. Catal. 2006, 348, 763 – 772
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
765

Peter Meinhold et al.
FULL PAPERS
Table 1. Regioselectivities, rates of product formation and total turnover numbers of octane hydroxylation reactions catalyzed
by variants of cytochrome P450 BM-3 9-10A containing single active-site mutations.[a]
Mutation[b]
1-ol [%]
2-ol [%]
3-ol [%]
4-ol [%]
Others[c]
Rate[d]
ttn[e]
(9-10A)
L75I
L75W
A78T4
A78F
A78S
A82T5
A82S
A82F
A82I
A82C
A82G
F87I
F87V
F87L
1
51
42
6
21
18
18
45
17
18
44
19
20
21
26
17
6
13
22
21
11
26
23
4
26
17
24
2
31
20
3
26
48
59
42
23
3
1
20
9
500
11
9
540
210
200
4000
380
260
2700
300
140
600
200
360
60
3000
4500
4000
3
3
22
4
5
23
6
4
4
5
7
8
5
7
4
7
6
6
27
37
48
25
40
26
11
26
52
70
52
55
51
67
39
29
88
71
87
3000
2500
340
9
2
5
1
2600
3000
3500
3000
2000
2600
3600
2300
4000
2500
3000
2500
2800
700
1
13
26
4
4
240
60
12
22
10
29
42
0
T88C
2
5
0
0
460
260
160
300
40
T260L
T260S
T260N
A328F
A328M
A328L
6
2
24
13
0
0
0
0
5
0
10
20
1000
[a]
Transformations were carried out using cell lysates. Refer to Experimental section for reaction conditions.
Mutations are in variant 9-10A (with amino acid mutations R47C, V78A, K94I, P142S, T175I, A184V, F205C, S226R,
H236Q, E252G, R255S, A290V, L353V).
Overoxidation products (ketones and aldehydes) with a similar product distribution as the alcohols.
Initial rates of product formation were measured by GC over 60 s and reported as nmol total products/min/nmol protein.
The reactions contained P450 (100 nM), NADPH (500 mM), and octane (4 mM) in ethanol (1%) and potassium phosphate
buffer (0.1 M, pH 8.0). Errors are at most 10%.
Total turnover numbers are reported as nmol product/nmol protein. The reactions contained P450 (25 nM), NADPH
(500 mM), and octane (4 mM) in ethanol (1%) and potassium phosphate buffer (0.1 M, pH 8.0). The reactions were allowed
to proceed for 4 hours, well past the lifetime of the enzymes under these conditions. Errors are at most 10%.
[b]
[c]
[d]
[e]
Table 2. Mutations[a] in selected cytochrome P450 BM3 mutants made by recombining active-site mutations and relative ac-
tivities towards dimethyl ether (DME) and hexyl methyl ether (HME). Product distributions,[b] rates and total turnover num-
bers (ttn) of octane hydroxylation are reported.
Mutant A78 A82 F87 A328 DME[c] HME[c] 1-ol [%] 2-ol [%] 3-ol [%] 4-ol [%] ttn[d] Rate[e] [minÀ1
]
9-10A
1-12G
77-9H
68-8F
13-7C
53-5H
0.6
1.0
0.5
1.0
0
3
51
79
21
11
27
4
3000 540
7500 150
1300 160
L
G
S
V
TG
F
TL
F
L
0.9
1.5
52
45
3
0
L
1.1
F
1.9
17
1.4
1.2
74
0.7
31
3
58
87
8
5
2
1400 180
7400 660
1.4
8
1
2600 310
1
[a]
Mutations were made in parent 9-10A which contains amino acid substitutions R47C, V78A K94I, P142S, T175I, A184V,
F205C, S226R, H236Q, E252G, R255S, A290V, and L353V.
[b]
Product distribution determined as ratio of a specific alcohol product to the total amount of alcohol products (given in %).
Errors are at most 3%. The formation of ketones was also observed but is less than 10% of the total amount of product.
Activities for terminal hydroxylation of dimethyl ether and hexyl methyl ether are reported relative to variant 1-12G.
Measured after 4 hours and reported as nmole octanols/nmole P450 produced (as described in Table 1).
Initial rates, in nmole octanol/nmole P450/min as described in Table 1.
[c]
[d]
[e]
ure 2), showing that activity on HME is a good surrogate terms of overall rate or ttn on octane, is highly active to-
for terminal hydroxylation of octane. Mutant 77-9H, for wards HME in the screen due to its propensity for termi-
example, although not improved relative to 1-12G in nal hydroxlylation.
766
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 763 – 772

Engineering Cytochrome P450 BM3 for Terminal Alkane Hydroxylation
FULL PAPERS
nes of chain length C6 to C10. Product distributions sum-
marized in Table 3 show that mutant 77-9H hydroxy-
lates only octane primarily at the terminal position. Al-
kanes shorter or longer than octane are hydroxylated
mainly at the 2-position.
Wild-type BM3 hydroxylates lauric and palmitic acid
at subterminal positions. Lauric acid is hydroxylated
mainly (48%) at the w-1 position, while palmitic acid is
hydroxylated at the w-2 position (48%).[19] On fatty
acids, mutant 77-9H shows a similar trend, with selectiv-
ity for specific subterminal positions (data not shown).
We altered residues that contact the terminal eight car-
bons of the substrate based on the crystal structure
with bound palmitoylglycine.[38] Amino acid residues
outside this region have not been modified, particularly
those interacting with the carboxyl moiety, and we might
expect changes in regioselectivity towards fatty acids to
be less dramatic than with alkane substrates.
Figure 2. The amount of 1-octanol formed during an octane
hydroxylation reaction correlates with activity on HME
measured during screening (R2 ¼0.64). Mutant 77-9H is de-
^
Total Turnover Numbers, Rates and Coupling
Efficiency
picted with
.
Total turnover numbers and rates of product formation
were measured for hydroxylation of the different al-
kanes by 77-9H (Table 4). Rates of product formation
range from 112 minÀ1 (on heptane) to 742 minÀ1 (on
nonane), similar to that reported for AlkB on octane
(200 minÀ1).[8] The rates of cytochromes P450 that carry
out terminal hydroxylation reactions are an order of
magnitude lower than those of 77-9H.[10,14] Wild-type
BM3 also hydroxylates alkanes at similarly low rates
(e.g., 30 minÀ1 for octane[28]).
To determine ttn, the enzyme was diluted to a concen-
tration at which the P450 is neither oxygen-limited nor
inactivated (at low enzyme concentrations, a decrease
in activity is observed which is likely due to dialysis of
the flavin cofactors, a possible inactivation mechanism
at low P450 concentrations[50]). The ttns and rates of hy-
Table 3. Product distributions[a] for alkane hydroxylation cat-
alyzed by cytochrome P450 BM3 77–9H.
Substrate 1-ol [%] 2-ol [%] 3-ol [%] 4-ol [%] 5-ol [%]
hexane
heptane 21.3
octane
nonane
decane
11.4
83.8
74.3
45.1
88.1
91.8
4.9
4.2
2.8
1.5
2.0
0.2
0.3
0.7
0.7
51.8
9.7
5.5
0.0
0.0
[a]
Product distribution determined as ratio of a specific alco-
hol product to the total amount of alcohol products (given
in %). Errors are at most 3%. The formation of ketones
was also observed but is less than 10% of the total amount
of product.
Variants other than 77-9H also exhibited increased ac- droxylation were found to be substrate-dependent and
tivity on HME (Table 2). None has the selectivity of 77- generally increase with the alkaneꢁs chain length. The
9H for terminal hydroxylation and instead favors con- observed decrease in ttn and rate for decane hydroxyla-
version of octane into 2-octanol. Because the total tion might be an artefact of the limited solubility of dec-
amount of 1-octanol formed correlates with the signal ane in the reaction buffer (~0.4 mM in water with 2%
intensity in the HME screen, an enzyme that catalyzes ethanol).[51]
10Â more total turnovers than 77-9H but produces
What limits the ttn of 77-9H is currently not under-
only one-tenth the fraction of 1-octanol would yield ap- stood. It is well known that unproductive dissociation
proximately the same signal in the screen. In fact, the en- of reduced oxygen species can lead to the formation of
zymes with the highest regioselectivities for 2-octanol superoxide or peroxide, which inactivate the enzyme.
were also the most active in terms of ttn (data not In fact, the enzyme exhibits the most turnovers on non-
shown).
ane, the substrate for which the coupling efficiency is the
highest (66%, Table 4).
Wild-type P450 BM3 tightly regulates electron trans-
fer from the cofactor (NADPH) to the heme. In the ab-
sence of substrate, a weakly-bound water molecule acts
Regioselectivity of Mutant 77-9H on n-Alkanes
Mutant 77-9H with the highest selectivity for terminal as the sixth, axial ligand of the heme iron. Substrate re-
hydroxylation was characterized in reactions with alka- places this water molecule, perturbing the spin-state
Adv. Synth. Catal. 2006, 348, 763 – 772
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
767

Peter Meinhold et al.
FULL PAPERS
Table 4. Total turnovers, product formation rates, NADPH oxidation rates, and coupling efficiencies for alkane hydroxylation
catalyzed by cytochrome P450 BM3 mutant 77-9H.
Substrate
ttn[a]
Rate of product formation
Rate of NADPH oxidation
Coupling efficiency
[%][d]
[b]
[c]
[minÀ1
]
[minÀ1
]
hexane
heptane
octane
nonane
decane
1800ꢀ130
1730ꢀ110
3040ꢀ150
4240ꢀ680
2831ꢀ225
251ꢀ29
112ꢀ9
1560ꢀ60
1630ꢀ20
1630ꢀ30
1150ꢀ100
940ꢀ50
16.1ꢀ2.0
6.9ꢀ0.6
236ꢀ24
742ꢀ30
318ꢀ47
14.5ꢀ1.7
65.6ꢀ9.3
33.4ꢀ7.2
[a]
Total turnover numbers determined as nmol product/nmol protein. The reactions contained P450 (25 nM), an NADPH re-
generation system (166 mM NADPþ , 6.6 U/mL isocitrate dehydrogenase, 41.6 mM isocitrate) and octane (4 mM) in ethanol
(2%) and potassium phosphate buffer (0.1 M, pH 8.0).
[b]
[c]
[d]
Rates of product formation were measured by GC over 20 s as nmol total products/min/nmol protein. The reactions con-
tained P450 (200 nM), NADPH (500 mM) and octane (2 mM) in ethanol (2%) and potassium phosphate buffer (0.1 M,
pH 8.0).
Rates of NADPH oxidation were measured over 20 s at 340 nm as nmol NADPH/min/nmol protein. The reactions con-
tained P450 (200 nM), NADPH (160 mM), and octane (2 mM) in ethanol (2%) and potassium phosphate buffer (0.1 M,
pH 8.0). The background NADPH oxidation rate (without substrate) rate was 220/min.
Coupling efficiency is the ratio of product formation rate to NADPH oxidation rate.
equilibrium of the heme iron in favor of the high-spin port w-hydroxylation of lauric acid,[43] although more re-
form and also increases the heme iron reduction poten- cent work showed that F87A broadens regioselectivity
tial by approximately 130 mV.[52,53] These events trigger and shifts hydroxylation away from the terminal posi-
electron transfer to the substrate-bound heme and start tion.[45,57] The F87A substitution in 9-10A does not
the catalytic cycle. However, with mutant BM3 enzymes lead to terminal hydroxylation activity on alkanes
and non-natural substrates, NADPH consumption is not (data not shown).
necessarily coupled to the formation of product. In-
By directing mutations to 11 residues in the active site,
stead, electrons from NADPH reduce heme-bound di- we generated a set of P450 BM3 mutants that exhibit a
oxygen to water or to peroxide and superoxide.[50] To de- range of regioselectivities for alkane hydroxylation. Sin-
termine the coupling efficiency of mutant 77-9H, rates gle mutations within the active site mainly result in rel-
of NAPDH oxidation upon addition of the alkane sub- atively small changes in the distribution of the alcohol
strates were measured and compared to product forma- products (Table 1). When these mutations are com-
tion rates (Table 4). The rate of product formation is cor- bined, however, more dramatic changes are observed
related to coupling efficiency, i.e., the more the reaction (Table 2). Mutant 77-9H, whose activity on HME is
is coupled to NADPH oxidation, the higher is the rate of even higher than that of 1-12G, has three mutations in
product formation.
the active site, A78T, A82G, and A328L. All three muta-
tions individually increase the fraction of terminal hy-
droxylation (Table 1). Mutation A328L exerts the
most dramatic effect in 9-10A, increasing the terminally
hydroxylated product from 1% to 13%. In contrast,
A78Tproduces predominantly 4-octanol (with 4% 1-oc-
Engineering the Active Site of P450 BM3for Selective
Substrate Oxidation
Linear alkanes are challenging to hydroxylate selective- tanol), while A82G produces 7% 1-octanol. In combina-
ly because they contain no functional groups that help tion, mutations A78Tand A328L (mutant 13-7C in Ta-
fix their orientation with respect to the reactive heme ble 2) increase terminal hydroxylation of octane to
iron-oxo group. Model studies with metalloporphyrins 17%. Addition of A82G further increases the selectivity
show that the regioselectivity of n-alkane hydroxylation for terminal hydroxylation, to 52% in 77-9H.
À
is determined by the relative C H bond dissociation en-
Mutant 77-9H exhibits terminal hydroxylation activi-
ergies if access to the reactive oxygen species is not re- ty only for octane. Other alkane substrates, in compari-
stricted.[54,55] Engineering a P450 for terminal hydroxy- son, are hydroxylated primarily at the subterminal posi-
lase activity must therefore override the inherent specif- tion. For nonane and decane, the 2-alcohols account for
icity of the catalytic species for the methylene groups so ~90% of the products. The active site of 77-9H is there-
that only the terminal methyl carbon is activated.
fore not restricted near the activated oxygen to prevent
Wild-type cytochrome P450 BM3 hydroxylates its subterminal hydroxylation in general. The specificity for
preferred substrates, medium-chain fatty acids, at sub- the terminal methyl group of octane probably reflects
terminal positions. The terminal carbon is not hydroxy- specific interactions between active site residues and oc-
lated.[56] P450 BM3 mutant F87A was reported to sup- tane methylene groups and/or the methyl group at the
768
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 763 – 772

Engineering Cytochrome P450 BM3 for Terminal Alkane Hydroxylation
FULL PAPERS
unreacted end. The hexyl methyl ether substrate used Expression and Purification of P450 BM3
for screening terminal hydroxylation activity has the
The genes encoding P450 BM3 mutants described in this report
same chain length as octane. The mutations that enable
terminal hydroxylation of HME likely act in a similar
fashion to promote terminal hydroxylation of octane.
This result is fully consistent with the first law of directed
evolution: “you get what you screen for.”[58]
were expressed and purified as described previously.[21] En-
zyme concentrations were determined in triplicate from the
CO-difference spectra.[60]
Construction of Saturation Mutagenesis Libraries
Mutations were introduced into cytochrome P450 BM3 mutant
9-10A[21] by PCR overlap extension mutagenesis using Pfu tur-
bo DNA polymerase.[61] The following positions were chosen
for saturation mutagenesis: A74, L75, A78, F81, A82, F87,
T88, T260, L263, I264 and A328. The primers for each satura-
tion library contained all possible combinations of bases, NNN
(N¼A, T, G, or C), at the codon for a particular residue. The
primers in the forward direction for each library were:
Conclusion
The BM3 active-site mutants described here also selec-
tively oxidize substrates other than alkanes and fatty
acids. We recently reported enantioselective terminal al-
kene epoxidation[31] as well as regio- and enantioselec-
tive hydroxylation of 2-cyclopentylbenzoxazole,[22]
yielding potentially valuable intermediates for chemical
synthesis. Mutant 53-5H, which exhibits the highest re-
gioselectivity for octane hydroxylation (Table 2), selec-
tively oxidizes ethane to ethanol, an activity unknown in
naturally-occurring P450s.[28] These findings demon-
strate that the BM3 active site can be efficiently molded
for enantio- and regioselective oxidation of a wide range
of substrates. The effects of mutations, however, are
hard to predict. Because the BM3 active site is highly
amenable to mutation, focusing mutagenesis to the ac-
tive site and identifying improved mutants using high
throughput screening or selection systems will continue
to be an effective strategy for engineering selective BM3
catalysts.
74NNNfor
75NNNfor
(5’-GTCAANNNCTTAAATTTGCACG-3’);
(5’-GTCAAGCGNNNAAATTTGCACG-3’);
78NNNfor (5’-GCTTAAATTTNNNCGTGATTTTGCAG-
G-3’); 81NNNfor (5’-CGTGATNNNGCAGGAGAC-3’);
82NNNfor (5’-CGTGATTTTNNNGGAGAC-3’); 87NNNfor
(5’-GAGACGGGTTANNNACAAGCTGGAC-3’); 88NNN
for
(5’-GGAGACGGGTTATTTNNNAGCTGGACG-3’);
260NNNfor (5’-CAAATTATTNNNTTCTTAATTGCGG-
GAC-3’); 263NNNfor (5’-ACATTCTTANNNGCGGGA-
CACGAAAC-3’); 264NNNfor (5’-ACATTCTTAATTN-
NNGGACACGAAAC-3’); and 328NNNfor (5’-CCAACT-
NNNCCTGCGTTTTCC-3’).
The reverse primers for each of these libraries complement
their corresponding forward primers. For each mutation, two
separate PCRs were performed, each using a perfectly comple-
mentary primer, BamHI-forw (5’-ggaaacaggatccatcgatgc-3’)
and SacI-rev (5’-gtgaaggaataccgccaagc-3’), at the end of the se-
quence and a mutagenic primer. The resulting two overlapping
fragments that contain the mutations were then annealed in a
second PCR to amplify the complete mutated gene. The full
gene was cut with BamHI and SacI restriction enzymes and li-
gated with T4 DNA ligase into pBM3_WT18-6, previously cut
with BamHI and SacI to remove the wild-type BM3 gene. The
ligation mixtures were transformed into E. coli DH5a electro-
competent cells and plated onto Luria-Bertani (LB) agar
plates to form single colonies for picking into 96-well plates.
A library of 174 mutants in two 96-well plates (eight wells con-
taining parent 9-10A, one well not inoculated) was prepared.
Assuming perfect primers, the library size is large enough to
cover all amino acids encoded by at least two codons with a
probability of 99.5%. Only for methionine is the probability
lower (93.5%).
Addendum
While this paper was under review, Urlacher and collab-
orators reported an engineered homologue of P450
BM3 (CYP102A3 from Bacillus subtilis) with terminal
octane hydroxylation activity.[59] None of the mutations
in the CYP102A3 variant are equivalent to those of 77-
9H.
Experimental Section
General Remarks
Construction of Recombination Libraries
All liquid alkanes and product standards and solvents, as well
as isocitrate and isocitrate dehydrogenase were purchased
from Sigma-Aldrich, Inc. (St. Louis, MO). Propane and di-
methyl ether were purchased from Special Gas Services (War-
ren, NJ). Hexyl methyl ether was purchased from TCI America
(Portland, OR). NADPH and NADPþ were obtained from Bi-
ocatalytics, Inc. (Pasadena, CA). Enzymes for molecular biol-
ogy experiments were purchased from Invitrogen (T4 DNA li-
gase), Roche (BamHI, SacI) and Stratagene (Pfu turbo DNA
polymerase)
Beneficial mutations found at seven positions (L75W, A78T/F,
A82T/S/F/I/C/G, F87I/V/L, T88C, T260N/L and A328L/F/M)
and the wild-type amino acids at these positions were recom-
bined sequentially by overlap extension PCR with degenerate
primers (see above) or a mixture of primers (mutation sites in
the primers are in bold letters; the reverse primers for each li-
brary complement corresponding forward primers). The muta-
tion L75I which was not found to increase activity was selected
for its change in regioselectivity and added to the library to in-
Adv. Synth. Catal. 2006, 348, 763 – 772
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
769

Peter Meinhold et al.
FULL PAPERS
crease its diversity. The mutations A82L and A328V were add-
ed to the library because of their previously described benefi-
cial effects.[21] Mutation A82V was added due to constraints
in designing the degenerate primer used to construct the li-
brary.
First, mutations at positions 78 and 82 were introduced into
9–10A using primers 78A/T/S 82A/T/S/F/I/C/G/L/V for* (5’-
GCTTAAATTCRCGCGTGATTTTDBCGGAGACG), and
2% ethanol. After 1 min, an NADPH regenerating system
(300 mL, 1.66 mM NADPþ, 66 U/mL isocitrate dehydrogen-
ase, 416 mM isocitrate) was added to the reaction before the
vial was capped and pressurized with oxygen (1.4 bar). At the
end of the reaction, an aliquot (1400 mL) was removed from
the vial.
To determine product formation rates, the reactions (5 mL)
were performed similarly, except more protein (200 nM) was
78F 82A/T/S/F/I/C/G/L/V for (5’-GCTTAAATTCTTTCGT- used, and NADPH was added (500 mM) instead of a regenera-
GATTTTDBCGGAGACG).
tion system. Additionally, the vials were not pressurized with
oxygen. Aliquots of the reaction (1400 mL) were removed at
20, 40 and 60 s.
Next, mutations at position 260 were introduced using pri-
mers 260Tfor* (5’-CAAATTATTACATTCTTAATTGCGG-
GAC), 260Nfor (5’-CAAATTATTAACTTCTTAATTG-
CGGGAC), and 260Lfor (5’-CAAATTATTCTTTTCT-
TAATTGCGGGAC).
Next, mutations at position 87 and 88 were introduced using
primers 87F/I/V/L 88Tfor* (5’-GAGACGGGTTANTYA-
CAAGCTGGAC), and 87F/I/V/L 88Cfor (5’-GAGAC-
GGGTTANTYTGTAGCTGGAC).
In all cases, the aliquots were quenched with CHCl3 (300 mL)
and HCl (6 M, 100 mL) in a 2 mL microcentrifuge tube. An in-
ternal standard containing 1-hexanol (14 mL, 10 mM) for oc-
tane, nonane and decane reactions or 1-octanol (14 mL,
10 mM) for hexane and heptane reactions was added to the
tube. The tube was vortexed and then centrifuged at 10,000 g
for 2 minutes in a microcentrifuge. The chloroform layer was
removed with a pipette and analyzed by gas chromatography
to determine product concentrations. Control reactions per-
formed by repeating these steps without the addition of sub-
strate revealed no detectable background levels of these specif-
ic products.
Next, mutations at position 328 were introduced using pri-
mers
328Afor*
(5’-CCAACTGCTCCTGCGTTTTCC),
328L/F/Vfor (5’-CCAACTBTTCCTGCGTTTTCC), and
328Mfor (5’-CCAACTATGCCTGCGTTTTCC).
Finally, mutations at position 75 were introduced using pri-
mers 75L/Ifor* (5’-CTTAAGTCAAGCGMTTAAATTC),
and 75Wfor (5’-CTTAAGTCAAGCGTGGAAATTC).
While building the library, we screened and sequenced ten
improved members of a partial library containing mutations
at positions 78, 82 and 260 to verify that it had been constructed
properly. Mutations were identified at all of these sites. We
found all possible amino acids at positions 78 and 260, and
four out of nine amino acids at position 82.
Gas Chromatography
Identification of analytes was performed using purchased
standards. Quantification was based on 5-point calibration
curves using internal standards. All samples were injected at
a volume of 1.0 mL and analyses were performed at least in trip-
licate. Analyses of alkane hydroxylation products were per-
formed on a Hewlett-Packard 5890 Series II Plus gas chroma-
tograph with a flame ionization detector (FID) and fitted with
an HP-7673 autosampler system. Direct analysis of hexane,
heptane, octane, nonane and decane hydroxylation products
was performed on an HP-5 capillary column (cross-linked
5% phenylmethylsiloxane, 30 m length, 0.32 mm ID, 0.25 mm
film thickness) connected to the FID detector. A typical tem-
perature program for separating the alcohol products is
2508C injector, 3008C detector, 508C oven for 3 minutes,
then 108C/minute gradient to 2008C, 258C/minute gradient
to 2508C, then 2508C for 3 minutes.
High Throughput Screening with Alkyl Methyl Ethers
Cell lysates for high throughput screening were prepared as de-
scribed.[21] Screening for activity on dimethyl ether was per-
formed as described.[21] For screening on hexyl methyl ether
(HME), phosphate buffer (120 mL, 0.1 M, pH 8.0) and then
2 mL of 400 mM HME in ethanol was added to 30 mL of E.
coli supernatant. After 1 min of incubation at room tempera-
ture, NADPH (50 mL, 0.8 mM) was added to the diluted lysate
and substrate. Purpald (168 mM in 2 M NaOH) was added
15 min after initiating the reaction to form a purple product
with the formaldehyde that was generated upon demethylation
of the substrate. The purple color was quantified approximate-
ly 15 min later at 550 nm using a Spectramax Plus microtiter
plate reader (Molecular Devices, Sunnyvale, CA). Improved
mutants were selected based on an increased signal compared
to 9-10A (for the saturation mutagenesis libraries) and 1-12G
(for the recombination library).
NADPH Oxidation Rates
Rates of NADPH oxidation were measured at 258C using a Bi-
oSpec-1601 UV-Vis spectrophotometer (Shimadzu, Columbia,
MD) and 1 cm pathlength cuvettes. A typical reaction solution
contained enzyme (800 mL, 200 nM) in potassium phosphate
buffer (0.1 M, pH 8.0) and 2 mM or 4 mM substrate in ethanol
(2%, vol/vol). The reaction was initiated by the addition of
NADPH (200 mL, 800 mM), and the decrease in absorption at
340 nm was monitored. Rates were calculated based on the ex-
tinction coefficient of NADPH (6.22 mMÀ1 cmÀ1). Back-
ground NADPH consumption rates were measured without
substrate.
Alkane Hydroxylation Reactions
For determination of total turnover numbers and regioselectiv-
ities, reactions with liquid alkanes (hexane, heptane, octane,
nonane and decane) were performed for 12 hours at 258C in
20 mL headspace vials, stirred at low speed using magnetic stir-
ring bars. The reactions were carried out in potassium phos-
phate buffer (0.1 M, pH 8.0, 3 mL) containing purified protein
(25 nM). Substrate was added to this solution as 60 mL of
200 mM ethanol solutions to give 4 mM total substrate and
770
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 763 – 772

Engineering Cytochrome P450 BM3 for Terminal Alkane Hydroxylation
FULL PAPERS
[21] M. W. Peters, P. Meinhold, F. H. Arnold, J. Am. Chem.
Soc. 2003, 125, 13442
Acknowledgements
The authors thank Dr. Nathan Dalleska for assistance with gas
chromatography. This work was supported by the National Sci-
ence Foundation (BES-0313567) and by a Caltech Minority Un-
dergraduate Research Fellowship (ARH).
[22] D. F. Munzer, P. Meinhold, M. W. Peters, S. Feichtenhof-
er, H. Griengl, F. H. Arnold, A. Glieder, A. de Raadt,
Chem. Commun. 2005, 20, 2597.
[23] D. Appel, S. Lutz-Wahl, P. Fischer, U. Schwaneberg,
R. D. Schmid, J. Biotechnol. 2001, 88, 167.
[24] A. B. Carmichael, L. L. Wong, Eur. J. Biochem. 2001,
268, 3117.
References
[1] M. Ayala, E. Torres, Appl. Catal. A: General 2004, 272, 1.
[2] H. Arakawa, M. Aresta, J. N. Armor, M. A. Barteau,
E. J. Beckman, A. T. Bell, J. E. Bercaw, C. Creutz, E.
Dinjus, D. A. Dixon, K. Domen, D. L. DuBois, J. Eckert,
E. Fujita, D. H. Gibson, W. A. Goddard, D. W. Good-
man, J. Keller, G. J. Kubas, H. H. Kung, J. E. Lyons,
L. E. Manzer, T. J. Marks, K. Morokuma, K. M. Nicholas,
R. Periana, L. Que, J. Rostrup-Nielson, W. M. H. Sach-
tler, L. D. Schmidt, A. Sen, G. A. Somorjai, P. C. Stair,
B. R. Stults, W. Tumas, Chem. Rev. 2001, 101, 953.
[3] J. B. van Beilen, Z. Li, W. A. Duetz, T. H. M. Smits, B.
Witholt, Oil Gas Sci. Technol. 2003, 58, 427.
[4] J. A. Peterson, D. Basu, M. J. Coon, J. Biol. Chem. 1966,
241, 5162.
[5] E. J. McKenna, M. J. Coon, J. Biol. Chem. 1970, 245,
3882.
[6] J. B. van Beilen, D. Penninga, B. Witholt, J. Biol. Chem.
1992, 267, 9194.
[7] M. Kok, R. Oldenhuis, M. P. G. Vanderlinden, P. Raatjes,
J. Kingma, P. H. Vanlelyveld, B. Witholt, J. Biol. Chem.
1989, 264, 5435.
[8] J. Shanklin, C. Achim, H. Schmidt, B. G. Fox, E. Munck,
Proc. Natl. Acad. Sci. USA 1997, 94, 2981.
[25] Q. S. Li, J. Ogawa, R. D. Schmid, S. Shimizu, Appl. Envi-
ron. Mirobiol. 2001, 67, 5735.
[26] Q. S. Li, U. Schwaneberg, P. Fischer, R. D. Schmid,
Chem. Eur. J. 2000, 6, 1531.
[27] C. F. Oliver, S. Modi, W. U. Primrose, L. Y. Lian, G. C. K.
Roberts, Biochem. J. 1997, 327, 537.
[28] P. Meinhold, M. W. Peters, M. M. Y. Chen, K. Takahashi,
F. H. Arnold, Chembiochem 2005, 6, 1765.
[29] Q. S. Li, J. Ogawa, R. D. Schmid, S. Shimizu, FEBS Lett.
2001, 508.
[30] S. Graham-Lorence, G. L. Truan, J. A. Peterson, J. R.
Falck, S. Wei, C. Helvig, J. H. Capdevila, J. Biol. Chem.
1997, 272, 1127.
[31] T. Kubo, M. W. Peters, P. Meinhold, F. H. Arnold, Chem.
Eur. J. 2006, 12, 1216–1220.
[32] E. T. Farinas, U. Schwaneberg, A. Glieder, F. H. Arnold,
Adv. Synth. Catal. 2001, 343, 601.
[33] A. Glieder, E. T. Farinas, F. H. Arnold, Nat. Biotechnol.
2002, 20, 1135.
[34] S. Modi, M. J. Sutcliffe, W. U. Primrose, L. Y. Lian,
G. C. K. Roberts, Nat. Struct. Biol. 1996, 3, 414.
[35] K. G. Ravichandran, S. S. Bodupalli, C. A. Hasemann,
J. A. Peterson, J. Deisenhofer, Science 1993, 261, 731.
[36] H. Li, T. L. Poulos, Nat. Struct. Biol. 1997, 4, 140.
[37] I. F. Sevrioukova, H. Li, H. Zhang, J. A. Peterson, T. L.
Poulos, Proc. Natl. Acad. Sci. 1999, 96, 1863.
[38] D. C. Haines, D. R. Tomchick, M. Machius, J. A. Peter-
son, Biochemistry 2001, 40, 13456.
[9] A. E. C. M. Simpson, Gen. Pharmacol. 1997, 28, 351.
[10] M. B. Fisher, Y. M. Zheng, A. E. Rettie, Biochem. Bio-
phys. Res. Commun. 1998, 248, 352.
[11] T. Lida, T. Sumita, A. Ohta, M. Takagi,Yeast 2000, 16,
1077.
[12] J. S. Yadav, J. C. Loper, Gene 1999, 226, 139.
[13] M. Ohkuma, T. Zimmer, T. Iida, W.-H. Schunck, A.
Ohta, M. Takagi, J. Biol. Chem. 1998, 273, 3948.
[14] U. Scheller, T. Zimmer, D. Becher, F. Schauer, W. H.
Schunck, J. Biol. Chem. 1998, 273, 32528.
[15] I. Benveniste, N. Tijet, F. Adas, G. Philipps, J. P. Salaun,
F. Durst, Biochem. Biophys. Res. Commun. 1998, 243,
688.
[39] T. Jovanovic, R. Farid, R. A. Friesner, A. E. McDermott,
J. Am. Chem. Soc. 2005, 127, 13548.
[40] H. B. Hopps, Aldrichimica Acta 2000, 33, 28.
[41] J. D. Bloom, J. J. Silberg, C. O. Wilke, D. A. Drummond,
C. Adami, F. H. Arnold, Proc. Natl. Acad. Sci. 2005, 102,
606.
[42] K. L. Morley, R. J. Kazlauskas, Trends Biotechnol. 2005,
23, 231.
[16] T. Maier, H. H. Forster, O. Asperger, U. Hahn, Biochem.
Biophys. Res. Commun. 2001, 286, 652.
[17] J. B. van Beilen, R. Holtackers, D. Luscher, U. Bauer, B.
Witholt, W. A. Duetz, Appl. Environ. Microbiol. 2005,
71, 1737.
[18] A. Noble, C. S. Moles, S. K. Chapman, D. A. Lysek, A. C.
Mackay, G. A. Reid, R. P. Hanzlik, A. W. Munro, Bio-
chem. J. 1999, 339, 371.
[19] S. S. Boddupalli, B. C. Pramanik, C. A. Slaughter, R. W.
Estabrook, J. A. Peterson, Arch. Biochem. Biophys.
1992, 292, 20.
[43] C. F. Oliver, S. Modi, M. J. Sutcliffe, W. U. Primrose,
G. C. K. Roberts, Biochemistry 1997, 36, 1567.
[44] H. Y. Li, T. L. Poulos, Biochim. Biophys. Acta 1999, 1441,
141.
[45] P. C. Cirino, F. H. Arnold, Adv. Synth. Catal. 2002, 344,
932.
[46] Q. S. Li, U. Schwaneberg, M. Fischer, J. Schmitt, J. Pleiss,
S. Lutz-Wahl, R. D. Schmid, Biochim. Biophys. Acta
2001, 1545.
[47] L. A. Cowart, J. R. Falck, J. H. Capdevila, Arch. Bio-
chem. Biophys. 2001, 387, 117.
[20] S. Schneider, M. G. Wubbolts, D. Sanglard, B. Witholt,
Tetrahedron: Asymmetry 1998, 9, 2833.
[48] J. A. Peterson, A. Hedge, S. Graham, D. Mullin, in: Sixth
International Symposium on Cytochrome P450 Biodiver-
sity, University of California, Los Angeles, 2002.
Adv. Synth. Catal. 2006, 348, 763 – 772
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
771

Peter Meinhold et al.
FULL PAPERS
[49] W. P. C. Stemmer, Proc. Natl. Acad. Sci. USA 1994, 91,
10747.
[54] J. F. Bartoli, O. Brigaud, P. Battioni, D. Mansuy, J. Chem.
Soc. Chem. Commun. 1991, 440.
[50] P. R. de Montanello, Cytochrome P450: structure, mecha-
nism, and biochemistry, 2 edn., Plenum Press, New York,
1995.
[51] A. Skrecz, D. Shaw, A. Maczynski, J. Phys. Chem. Refer-
ence Data 1999, 28, 983.
[52] S. N. Daff, S. K. Chapman, K. L. Turner, R. A. Holt, S.
Govindaraj, T. L. Poulos, A. W. Munro, Biochemistry
1997, 36, 13816.
[53] T. W. B. Ost, J. Clark, C. G. Mowat, C. S. Miles, M. D.
Walkinshaw, G. A. Reid, S. K. Chapman, S. Daff, J.
Am. Chem. Soc. 2003, 125, 15010.
[55] B. R. Cook, T. J. Reinert, K. S. Suslick, J. Am. Chem. Soc.
1986, 108, 7281.
[56] Y. Miura, A. J. Fulco, Biochim. Biophys. Acta 1975, 388,
305.
[57] O. Lentz, V. Urlacher, R. D. Schmid, J. Biotechnol. 2004,
108, 41.
[58] L. You, F. H. Arnold, Protein Eng. 1996, 9, 77.
[59] O. Lentz, A. Feenstra, T. Habicher, B. Hauer, R. D.
Schmid, V. B. Urlacher, Chembiochem 2005, 7, 345.
[60] T. Omura, R. Sato, J. Biol. Chem. 1964, 239, 2370.
[61] R. Higuchi, B. Krummel, R. K. Saiki, Nucl. Acids Res.
1988, 16, 7351.
772
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 763 – 772