Baeckvall J.-E. (ed.) Modern Oxidation Methods

Подождите немного. Документ загружается.

Pimaricin is an important antifungal agent used in human therapy for the

treatment of fungal keratitis as well as in the food industry to prevent mold

contamination. The enzymes and the gene cluster responsible for their production

by Streptomyces natalensis was identiﬁed recently. The cluster contains a P450 that is

responsible for ring decoration [99].

Production of epothilones (potential anticancer agents) was carried out with

recombinant Streptomyces strains. Therefore, the entire epothilone biosynthetic gene

cluster from the myxobacterium Sorangium cellulosum was heterologously expressed

in an engineered Streptomyces venezuelae strain. The resulting strains produced

approximately 0.1 mgL

1

epothilone B as a sole product after 4 days cultivation.

Deletion of a gene encoding a cytochrome P450 epoxidase gave rise to a mutant that

selectively produced 0.4 mgL

1

epothilone D [100].

Biosynthesis of the sesquiterpene antibiotic albaﬂavenone in Streptomyces coelicolor

A3(2) was studied in detail. The mechanism and stereochemistry of the enzymatic

formation of epi-isozizaene by multistep cyclization of farnesyl diphosphate was

investigated [101] and a P450 (CYP170A1) was identiﬁed to carry out two sequential

allylic oxidations to convert epi-isozizaene to an epimeric mixture of albaﬂavenols

and thence to albaﬂavenone [102].

While functions of a number of P450s from Streptomyces are known, in the vast

majority of these species the function of P450s so far remains unknown. Efforts are

being made to shed light on this issue. The long-term goal is to identify orphan P450

functions and to establish a strategy for the production of novel secondary metabo-

lites that have new biomedical function [103, 104]. Some crystal structures of P450s

from Streptomyces are already available, for example, that of P450

PikC

from Strepto-

myces venezuelae [105], P450

StaP

(CYP245A1) from Streptomyces sp. TP-A0274 [106],

P450

SU1

(CYP105A1) from Streptomyces griseolus [107], or recently CYP105P1 from

Streptomyces avermitilis [108]. The structural knowledge of these P450s opens up a way

for enzyme optimization and engineering leading to new antibacterial properties.

While Streptomyces produce the majority of known antibiotics, there are also

other bacterial species with interesting activities for pharmaceutical application.

OxyA, OxyB, and OxyC (CYP165A1, CYP165B1, and CYP165C1) from Amycolatopsis

orientalis, for example, are involved in the biosynthesis of the glycopeptide antibiotic

vancomycin [67]. A similar set of P450s with high identity to the oxygenases in

vancomycin biosynthesis (92% to 94%) is involved in the biosynthesis of balhimycin

in Amycolatopsis balhimycina [109].

Another example is P450

eryF

(CYP107A1) from Saccharopolyspora erythraea, which

catalyzes a hydroxylation step leading to the functional molecule in the biosynthesis

of the macrolide antibiotic erythromycin [110]. The crystal structures of OxyB and

OxyC [111, 112] as well as that of P450

eryF

[113] have been published.

Nonrecombinant wild-type strains are often used as biocatalysts for the production

of pharmaceutically relevant metabolites, as in many of the examples described

above. The application of wild-type strains, however, has some drawbacks. These

include product degradation, by-product formation, and catabolite repression [75].

Further, there is little or no possibility of engineering the biocatalyst itself. The

application of recombinant engineered strains, combined with bioprocess and

12.3 Application and Engineering of P450s for the Pharmaceutical Industry

433

biocatalyst engineering, provides an attractive alternative. For industrial biocatalysis

E. coli is in general the organism of choice because of the well-developed molecular

biology techniques and high levels of biocatalyst expression. Besides this, high cell

densities can be achieved during fermentation of E. coli [114]. Recombinant Pseu-

domonas strains are of industrial interest as well, since they could be useful for

biotransformations in the presence of toxic organic solvents. A number of solvent-

tolerant P. putida strains have been isolated and used as hosts for the expression of

P450 genes. An excellent review which includes many examples of the use of

recombinant E. coli and P. putida whole-cell biocatalysts on a scale up to 30 L has

been written by Park [75].

Recently construction and application of a novel P450 library, based on about

250 bacterial P450 genes (about 70% originating from actinomycetes), co-expressed

with putidaredoxin and putidaredoxin reductase in E. coli has been reported [115].

A good example demonstrating the exquisite regioselectivity of P450s, which is

impossible to reach by chemical oxidations, is presented by a screening of this library

with testosterone. Within the screening 24 bacterial P450s were identiﬁed, which

monohydroxylate testosterone regio- and stereoselectively at the 2a-, 2b-, 6b-, 7b-,

11b-, 12b -, 15b-, 16a- and 17-positions [116]. Most of these hydroxylations are

common for both prokaryotic and human P450s. Thus, the identiﬁed bacterial

candidates will be applied for the production of drug metabolites on a preparative

scale.

A screening among 1800 bacterial strains identiﬁed the Mycobacterium sp. strain

HXH-1500, catalyzing regio- and stereoselective hydroxylation of limonene at

carbon-atom C7 to yield the anticancer drug ()-perillyl alcohol. The biocatalytic

production of ()-perillyl alcohol from limonene was performed using the

Mycobacterium sp. P450 alkane hydroxylase (CYP153 family) recombinantly

expressed in P. putida cells [117]. The whole-cell process was performed in a two-

phase system resulting in 6.8 g L

1

of product in the organic phase.

12.3.2

Application of Mammalian P450s for Drug Development

An ongoing ﬁeld of research in drug discovery and drug development is the use of

recombinant human P450s. Human P450s expressed in baculovirus-infected insect

cell lines have been in use for quite some time now to identify human P450s involved

in drug metabolism [118, 119]. They are also used to synthesize drug metabolites in

order to assess the toxicity of potential pharmaceuticals [120–122]. Recent achieve-

ments in the expression of recombinant human P450s in Saccharomyces cerevisiae

[76, 123, 124], Yarrowia lipolytica [125] and Pichia pastoris [126, 127] facilitate their use

for the synthesis of drug metabolites.

Besides the achievements in the expression of mammalian P450s in eukaryotic

organisms, E. coli is an attractive expression system and is of particular interest for

industrial applications (see also Section 12.3.1). Efforts to express human P450s in E.

coli have been made [128–130]; in this case, however, mammalian cDNAs have to be

modiﬁed before they can be expressed [131].

434

12 Biooxidation with Cytochrome P450 Monooxygenases

12.3.2.1 Enhancement of Recombinant Expression in E. coli

Initial attempts to enhance the expression of human P450s in E. coli involved

modiﬁcation of the N-terminal sequence, since this region has been proposed to

be important in determining protein yield [132, 133]. This strategy was applied to

express a number of mammalian P450s including CYP 2E1 [134], 4A4 [135],

6A1 [136], 1A1 [137], as well as recently 20A1 [138] and 4X1 [139]. A different

approach was pursued by truncating the N-terminal hydrophobic region – which is

responsible for the interaction with the membrane – to a greater or lesser extent [140–

142]. Three human P450s (3A4, 2C9, and 1A2) with truncated N-termini were each

co-expressed with human CPR in E. coli using a bicistronic expression system. Intact

E. coli cells and membranes containing P450 and CPR were used in the preparative

synthesis of drug metabolites. The optimized biooxidation conducted on a 1-L scale

yielded 59 mg 6b-hydroxytestosterone (32), 110 mg 4-hydroxydiclofenac (33), and

88 mg acetaminophen (34) [130] (Scheme 12.11).

Further optimization could be achieved by fusions of a bacterial signal peptide (for

example a modiﬁed ompA-leader sequence) to human P450 cDNAs [122]. This leader

is removed during P450-processing, and the native P450 is released. Following this

strategy, 14 recombinant human P450s co-expressed with NADPH P450 reductase in

HOOC

diclofenac 33

CYP2C9

CYP1A2

phenacetin

testosterone

CYP3A4

Scheme 12.11 Oxidation of testosterone, diclofenac, and phenacetin to yield 6b-

hydroxytestosterone (32), 4-hydroxydiclofenac (33) and acetaminophen (34), respectively.

12.3 Application and Engineering of P450s for the Pharmaceutical Industry

435

E. coli have been used by several pharmaceutical companies, both as biocatalysts for

the preparation of metabolites of drug candidates and for high-throughput P450

inhibition screenings. Up to 300 mg of different metabolites could be obtained using

permeabilized recombinant E. coli cells expressing human P450s [143].

12.3.2.2 Enhancement of Activity and Selectivity and Engineering of Novel Activities

Since structure-function relationships for mammalian (and other eukaryotic) P450s

are far less clear than for bacterial ones, most protein engineering efforts conducted

on mammalian P450s so far focus on the identiﬁcation of key residues involved in

substrate binding. Nevertheless, as more crystal structures of mammalian P450s are

becoming available, the structural knowledge can also be used to engineer novel

enzyme properties.

Recent examples include the engineering of CYP2B1 by a combination of random

and site-directed mutagenesis to accept the anti-cancer prodrugs cyclophosphamide

and ifosfamide [144]. Homology modeling and rational design performed on

CYP2B1 allowed the regioselectivity of progesterone 16-a-hydroxylation to switch

to 21-hydroxylation [145].

The ability of human CYP1A2 to catalyze O-demethylation of 7-methoxyresoruﬁn

was improved by three rounds of mutagenesis. The triple-mutant E163K/V193M/

K170Q exhibited turnover rates more than ﬁve times faster than wild-type

CYP1A2 [146].

TheGillamgroupconstructedachimericlibraryofCYP1A1andCYP1A2employing

restriction enzyme-mediated DNA family shufﬂing. They observed different activity

proﬁles toward various luciferin derivatives among active clones, including improved

speciﬁc activity, novel activities, and broadening of substrate range [147, 148].

The substrate spectrum of human CYP2D6 was expanded toward bulky indole

derivatives, like 4- and 5-benzyloxyindoles, by random mutagenesis and site-directed

mutagenesis [149].

12.3.2.3 Construction of Artificial Self-Sufficient Fusion Proteins

A very effective mode to eliminate the reconstitution procedure of redox chains is

to tailor self-sufﬁcient electron transport chains emulating natural fusion enzymes.

A large variety of these artiﬁcial fusion enzymes have been reported. Fusions of P450s

with a diﬂavin (FAD- and FMN-containing) reductase, P450s with ﬂavodoxins, P450s

with ferredoxins and ferredoxin reductases, P450s with dioxygenase reductase-like

enzymes, and even semisynthetic heme-containing ﬂavoproteins that can act as

a monooxygenase have been described [70, 150]. Genes of various mammalian P450s

and their redox partners were used to generate functional multi-domain proteins

with designed properties (molecular Lego) beyond the restrictions imposed by the

naturally occurring protein domains [151].

For example, the N-terminal human CYP2D6 sequence was linked to the

C-terminal human cytochrome reductase module, resulting in a self-sufﬁcient

enzyme exhibiting activity toward a wide range of pharmaceutical compounds such

as bufuralol, metoprolol, or dextromethorphan [152]. Other examples are fusion

proteins of human CYP1A1 with rat CPR [153], human CYP2C9 with BMR [154], and

436

12 Biooxidation with Cytochrome P450 Monooxygenases

human CYP3A4 with rat or human CPR [155]. As cytochrome b

was detected to

remarkably enhance the coupling efﬁciency if co-expressed with the CYP3A4-fusion

systems [156], recombinant three-component systems with CYP3A4/CPR/

cytochrome b

were tailored, maintaining even higher oxidation activity for testos-

terone and nifedipine [157].

12.4

Application of P450s for Synthesis of Fine Chemicals

Another valuable application for P450s, aside from the production of pharmaceu-

tically relevant metabolites, is the fabrication of sought-after ﬁne chemicals and

ﬂavors that are of interest for the food industry [158]. P450s are interesting

biocatalysts for such syntheses, since chirality of the products is often very important.

Celik et al. have described an example of substrate hydroxylation by P450s with

a high degree of regio- and diastereose lectivity [159]. The authors used recombinant

E. coli whole-cell systems, in which cytochromes P450

SU1

, P 450

SU2

, and P450

SOY

were over-expressed with their cognate ferrodoxins for hydroxylation of a-ionone

(35)andb-ionone (38) to their corresponding mono-hydrox ylated derivatives. For

a-ionone the reaction was diastereoselective, yielding only the anti-isomers

(3S,6S)- and (3R,6R)-3-hydrox y-a-ionone (36 and 37). For b-ionone an enantiomeri c

excess of 35% in favor of the (S)-enantiomer (39) was observed (Scheme 12.12).

In another screening of the recombinant P450-library described in Ref. [115] (see

Section 12.3.1), CYP109B1 was identiﬁed to be capable of regioselective hydroxyl-

ation of ( þ )-valencene (48) at allylic C2-position, thereby producing ( þ )-nootkatone

– a sought after fragrance for ﬂavoring that has a strong grapefruit odor and bitter

taste [160, 161] – via nootkatol. By product formation, which accounted for 35% of the

total product during biooxidation with recombinant E. coli in an aqueous milieu, was

signiﬁcantly reduced when aqueous-organic two-liquid-phase systems were set up

O O O

O O

6353

P450

Scheme 12.12 Regio- and diastereoselective hydroxylation of a-ionone (35) and b-ionone (38).

12.4 Application of P450s for Synthesis of Fine Chemicals

437

leading to accumulation of nootkatol and ( þ )-nootkatone of up to 97% of the total

product. Up to 120 mg L

1

of nootkatol and ( þ )-nootkatone were produced within

8 h employing this system [162].

Functional expression of tricistronic constructs consisting of P450

cam

from

P. putida and the auxiliary redox partners putidaredoxin reductase and putidaredoxin

in E. coli has been reported. The transformed whole-cells efﬁciently oxidized (1R)-

( þ )-camphor (10) to 5-exo-hydroxycamphor (11) and limonene to ()-perillyl

alcohol [163]. Since these bioengineered e-cells possess a heterologous self-sufﬁcient

P450 catalytic system, they may have advantages in terms of low cost and high yield

for the production of ﬁne chemicals.

12.5

Engineering of P450s for Biocatalysis

12.5.1

Cofactor Substitution and Regeneration

The application of cytochrome P450 monooxygenases requires functional and

efﬁcient electron suppliers and is dependent on the availability of the costly cofactors

NAD(P)H. Thus, various approaches have been made to substitute or regenerate

NAD(P)H and to increase the coupling efﬁciency for both in vitro and in vivo

applications of P450s. These attempts include direct chemical or electrochemical

reduction of P450s, the use of cheap chemicals to directly replace NAD(P)H, the

development of (enzymatic) cofactor recycling systems [164], and the engineering

of artiﬁcial fusion proteins [150].

12.5.1.1 Cofactor Substitution In Vitro

Direct chemical reduction can be achieved by the use of organometallic complexes

such as Co

sepulchrate trichloride [165, 166] by addition of Ti

III

citrate [167] or by the

strong reducing agent sodium dithionite [168] – an inexpensive agent routinely used

to produce the ferrous-carbonyl form of P450. The limitations of these approaches

are, however, low system efﬁcacy, low sensitivity of mediators to molecular oxygen,

and the production of reactive oxygen species or mediator aggregation.

Peroxides that directly convert the heme iron of P450s to a ferric hydroperoxy

complex by the peroxide shunt (e.g., hydrogen peroxide, cumene peroxide, or tert-

butyl oxide) could be useful for oxidation of various substrates. The essential problem

in utilizing the peroxide shunt for P450 biocatalysis seems to lie in the time-

dependent degradation of the heme and in oxidation of the protein [169, 170].

Methods of directed evolution, like random and site-speciﬁc mutagenesis, were

applied to evolve P450s to enhance the efﬁciency of the peroxide shunt

pathway [171].

Electrochemical reduction of P450s by using an electrode seems to be the most

convenient way of fabricating bioreactors [172] and has been studied in detail for

more than 15 years. However, heme reduction on a cathode is complicated and most

438

12 Biooxidation with Cytochrome P450 Monooxygenases

often insufﬁcient for biocatalysis, which can partly be attributed to the deeply buried

heme iron and partly to the instability of the enzyme upon interaction with the

electrode surface. Improvements to this approach include the modiﬁcation of the

electrode surface [150].

12.5.1.2 Cofactor Regeneration In Vitro

One of the most common approaches to overcome the stoichiometric need for

NAD(P)H involves enzymatic regeneration systems. Strategies for the regeneration

of NADPH-dependent P450s are based on

D/L-isocitrate dehydrogenase [173], an

engineered formate dehydrogenase from Pseudomonas sp. 101 [174], glucose-6-

phosphate dehydrogenase [175], or alcohol dehydrogenase from Thermoanaerobium

brockii [176]. Although these systems work reasonably well, further reduction of costs

for enzymatic oxyfunctionalization can be achieved by engineering of NADPH-

dependent P450s to accept NADH [177, 178]. NADH has the advantage that it is about

ten times less expensive and more stable than NADPH. Furthermore, more NAD

dependent enzymes at lower prices are available for cofactor regeneration. Our group

used homology modeling and site-directed mutagenesis to construct P450

BM3

mutants showing altered cofactor speciﬁcity from NADPH to NADH. The best

mutant, W1046S/S965D, demonstrated high turnover rates (k

cat

¼ 14600 min

1

)

during cytochrome c reduction and low K

for NADH. It had equal afﬁnity to both

NADH and NADPH and displayed a 2.6-times higher activity when NADH was

used [179].

12.5.1.3 Cofactor Regeneration in Whole-Cells

Cofactor concentration within the cell may become a bottleneck for the overall

process if concentrations of the recombinant P450 biocatalyst and/or their activities

reach a very high level [180]. Strategies to overcome this limitation employ the use

of artiﬁcial cofactor recycling systems supporting the endogenous ones and the

employment of native electron transfer systems to increase the coupling efﬁciency of

P450s and reduce metabolic stress for the host organism.

The catalytic efﬁciency of P450

cam

expressed i n E. coli could be increased up to

25-fold by the coupling of its native electron transfe r syst em (putidaredoxin

reductase and putidaredoxin) to enzymatic NADH regeneration catalyzed by

a recombinantly expresse d glycerol dehydrogenase. This whole-cell s ystem was

applied for camphor hydroxylation in aqueous solution, as well as in a biphasic

system with isooctane [181].

A recent approach in this ﬁeld is the construction of a novel E. coli whole-cell

biocatalyst with improved intracellular cofactor regeneration driven by external

glucose [182]. In this system, additional recombinant intracellular NADPH regen-

eration takes place through co-expression of a glucose facilitator from Zymomonas

mobilis for uptake of unphosphorylated glucose and an NADP

-dependent glucose

dehydrogenase from Bacillus megaterium that oxidizes glucose to gluconolactone.

When a quintuple mutant of P450

BM3

that oxyfunctionalizes a-pinene (49) – a cheap

waste product of wood industry – to yield a-pinene oxide, verbenol and myrtenol [183]

was expressed in this system, the engineered strain showed a 9-fold increased initial

12.5 Engineering of P450s for Biocatalysis

439

a-pinene oxide formation rate and a 7-fold increased a-pinene oxide yield in the

presence of glucose compared to glucose-free conditions. Further bioprocess engi-

neering addressed the low water solubility and the toxicity of a-pinene by setting up

an aqueous-organic two-phase bioprocess with diisononyl phthalate as a biocom-

patible organic carrier solvent. With an aqueous-to-organic phase ratio of 3: 2 and

30% (v/v) of a-pinene in the organic phase, a total product concentration of over

1gL

1

was achieved [184]. This process marks a promising step toward a future

application of recombinant microorganisms for the selective oxidation of terpenoids

to value-added products.

12.5.2

Construction of Artificial Fusion Proteins

As in the case of human P450s (described in Section 12.3.2.3), a variety of arti ﬁcial

fusion enzymes have also been generated for bacterial P450s aiming to increase

enzyme activity and coupling efﬁciency.

Concerning P450

cam

, the hemoprotein was fused to its natural electron donors,

putidaredoxin and putidaredoxin reductase in different orders of the individual parts

and variable linkers between them. However, despite increased coupling, the catalytic

activity of the engineered assembly was only 30% of that of the reconstituted wild-type

mixture, reconstituted in a 1 : 1 : 1 ratio with the separate redox partners [185]. The

catalytic activity of the fusion enzyme toward

D-( þ )-camphor (10) could be increased

10-fold by the development of an enzymatic cross-linking of putidaredoxin with a

peptide fragment between putidaredoxin reductase and P450

cam

, accommodating a

reactive glutamyl residue prone to attack by transglutaminase [186].

The putidaredoxin reductase-like unit of the self-sufﬁcient CYP116B2 from Rho-

dococcus sp. (P450RhF) [187] was ligated to the heme domain of P450

cam

, CYP153A, or

CYP203A. The chimeric proteins showed catalytic activity against a variety of

compounds like n-alkanes,

D-( þ )-camphor (10), and 4-hydroxybenzoate [188].

12.5.3

Engineering of New Substrate Specificities

In this chapter focus exclusively on the two best characterized P450s: P450

cam

(CYP101A1) from P. putida [189] and P450

BM3

(CYP102A1) from B. megaterium.

Their structure, catalytic mechanism, and biochemistry have been studied in detail

extensively [10]. Mutagenesis studies on these P450s have led to signiﬁcant improve-

ments in our understanding of general aspects of P450 catalytic function, and

numerous approaches have been undertaken to engineer new substrate speciﬁcities

for these enzymes.

12.5.3.1 P450

cam

from Pseudomonas putida

For P450

cam

it was demonstrated that mutations of the active site residue Y96 to more

hydrophobic ones considerably increased its activity toward the oxidation of hydro-

phobic molecules smaller than camphor (10), like styrene (40) and alkanes [190, 191],

440

12 Biooxidation with Cytochrome P450 Monooxygenases

or larger than camphor, such as naphthalene (41 ) and pyrene (42) [192, 193]. The

Y96Fmutant was capable of regio- and stereoselective hydroxylation of tetralin (43)to

the 1-(R)-alcohol [194]. Using saturation mutagenesis at positions Y96 and F87,

several mutants were constructed with activity toward indole (44) and diphenyl-

methane (45) [195]. Some mutants have been engineered for biodegradation of

environmentally harmful pollutants, such as di-, tri-, tetra-, penta-, and hexachlor-

obenzene (46) [196, 197] (Figure 12.3).

P450

cam

was further engineered to an alkane hydroxylase by step-by-step adap-

tation of the enzyme to smaller n-alkanes beginning with hexane [198], then

proceeding to butane and propane [199], and ﬁnally to ethane (47) [200]. The

best mutant, with eight substitutions, oxidized propane at 500 min

1

with 86%

coupling, which was comparable with that of the wild-type enzyme toward camphor

(10) [200].

Successful attempts to engineer mutants for selective oxidation of cheap terpenes to

expensive oxidized derivatives have also been accomplished. Rational mutants

of P450

cam

exhibited activity toward the sesquiterpene ( þ )-valencene (48) and pro-

duced > 85% of ( þ )-trans-nootkatol and ( þ )-nootkatone [201]. The monoterpene

( þ )-a-pinene (49) was oxidized by P450

cam

mutants to 86% ( þ )-cis-verbenol or to a

mixture of ( þ )-verbenone and ( þ )-cis-verbenol (Figure 12.3) [202]. Verbenol and

verbenone are active pheromones against various beetle species.

P450

cam

ClCl

Y96A

Y96

V247

F87 Y96

L244 V247

Y96F

F87W Y96F T101L

T185M V247L

G248AL294M

L1244M L1358P

F87 Y96

F87W Y96F

V247L

F87 Y96

Y96F

F87 Y96

Figure 12.3 Mutations of P450

cam

leading to altered substrate specificities. WT: wild-type enzyme

without mutations.

12.5 Engineering of P450s for Biocatalysis

441

12.5.3.2 P450

BM3

from Bacillus megaterium

P450

BM3

is an obvious target enzyme for the development of biotechnological

applications, since it is a self-sufﬁcient single-component protein with a high catalytic

activity. The turnover rates of P450

BM3

toward fatty acids are among the highest

activity values reported for P450s [203].

The mutant A74G/F87V/L188Q, designed by saturation mutagenesis, was shown

to oxidize indole, n-octane, highly branched fatty acids and fatty alcohols, poly-

chlorinated dibenzo-p-dioxins, polyaromatic hydrocarbons, styrene, and many other

chemical compounds [204 –208]. The monoterpene geranylacetone was converted by

P450

BM3

(R47L/Y51F/F87V) with high activity (2080 min

1

) and stereoselectivity

(97% ee) to a single product, namely 9,10-epoxygeranylacetone [209]. The mutant

A74E/F87V/P386S exhibited an 80-fold improved activity toward b-ionone (42)

compared to the wild-type enzyme and produced the ﬂavoring (R)-4-hydroxy-b-io-

none as the only product [210].

Using a combination of directed evolution and site-directed mutagenesis Arnold

and coworkers altered the selectivity of P450

BM3

from hydroxylation of dodecane

(C12), ﬁrst to octane (C8) and hexane (C6) and further on to gaseous propane (C3) and

ethane (C2) [211–214]. Some mutants were found with high stereoselectivity, leading

either to (R)- or to (S)-2-octanol [215].

In our group, a systematic analysis of the structures of 29 P450s and 6379 P450

sequences, with the aim of identifying selectivity- and speciﬁcity-determining

residues, led to identiﬁcation of a positively charged heme-interacting residue in

the SRS5, which was present in about 98% of the sequences analyzed. This residue is

located in close vicinity to the heme center and restricts the conformation of the

SRS5. It is preferentially located at position 10 or 11 after the conserved ExxR motif (in

about 95% of the sequences). Replacing this residue by hydrophobic residues

of different size has been shown to change substrate speciﬁcity and regioselectivity

for P450s of different superfamilies [216]. Based on this analysis, a minimal P450

BM3

mutant library of only 24 variants plus wild-type was constructed by combining ﬁve

hydrophobic amino acids (alanine, valine, phenylalanine, leucine, and isoleucine) in

positions 87 and 328. The library was screened with four terpene substrates

geranylacetone, nerylacetone, (4R)-limonene, and ( þ )-valencene (52). Eleven var-

iants demonstrated either a strong shift or improved regio- or stereoselectivity during

oxidation of at least one substrate as compared to P450

BM3

wild-type [217].

Although wild-type P450

BM3

is not able to metabolize any drug-like compound

tested so far, it has been turned by protein design and directed evolution into an

enzyme that oxidizes human drugs [218]. The R47L/F87V/L188Q mutant was shown

to metabolize testosterone, amodiaquine, dextromethorphan, acetaminophen, and

3,4-methylenedioxymethylamphetamine [219]. Several mutants were obtained by

means of directed evolution which are able to convert propranolol, a multi-function

beta-adrenergic blocker [220]. Another mutant was capable of stereo- and regiose-

lective hydroxylation of the peptide group of buspirone to yield (R)-6-hydroxybuspir-

one – an anti-anxiety agent – with > 99.5% ee [221].

Recent approaches in this ﬁeld reported mutants with applicability as biocatalysts

in the production of reactive metabolites from the drugs clozapine, diclofenac, and

442

12 Biooxidation with Cytochrome P450 Monooxygenases