Báo cáo khoa học: The role of cytochrome P450 monooxygenases in microbial fatty acid metabolism pdf

For example, Keywords alkanes; bacteria; biosurfactants; cytochrome P450 monooxygenase; dicarboxylic acids; fatty acids; hydroxylation; oxylipins; polyketides; yeast Correspondence I.. B

The role of cytochrome P450 monooxygenases in

microbial fatty acid metabolism

Inge N A Van Bogaert1, Sara Groeneboer2, Karen Saerens1and Wim Soetaert1

1 Department of Biochemical and Microbial Technology, Laboratory of Industrial Biotechnology and Biocatalysis, Ghent University,

Ghent, Belgium

2 Laboratory for Protein Biochemistry and Biomolecular Engineering, Ghent University, Ghent, Belgium

Introduction

P450 classification and nomenclature

Cytochrome P450s form a vast and divergent family of

enzymes They are heme–thiolate proteins, bearing, in

a hydrophobic pocket, a protoporphyrin IX linked to

the apoprotein by a bond between the heme iron

cen-tre and the sulfur atom of a conserved cysteinyl

resi-due In a typical reaction catalysed by a P450,

molecular oxygen binds to the heme iron for activation

before transfer to the substrate Carbon monoxide can

also bind, leading to a reduced P450 producing a

char-acteristic CO-binding difference spectrum with an

absorbance maximum at 450 nm This P450–CO com-plex is inactive and has given the name to P450 (pig-ment absorbing at 450 nm) [1] Inhibition by CO and reversion of this inhibition by 450 nm light are charac-teristic for most reactions catalysed by P450s

New genes are annotated as a P450 based on the presence of typical conserved domains involved in heme binding and proton transfer [2] P450s have been categorized in families and subfamilies They belong to the same family when they share ‡ 40% amino acid identity and they belong to the same subfamily when they share‡ 55% amino acid identity [3] For example,

Keywords

alkanes; bacteria; biosurfactants;

cytochrome P450 monooxygenase;

dicarboxylic acids; fatty acids; hydroxylation;

oxylipins; polyketides; yeast

Correspondence

I N A Van Bogaert, Department of

Biochemical and Microbial Technology,

Laboratory of Industrial Biotechnology and

Biocatalysis, Faculty of Bioscience

Engineering, Ghent University, Coupure

Links 653, B-9000 Ghent, Belgium

Fax: +32 9 264 62 31

Tel: +32 9 264 60 34

E-mail: inge.vanbogaert@ugent.be

Website: http://www.inbio.be

(Received 22 June 2010, revised 19 August

2010, accepted 16 September 2010)

doi:10.1111/j.1742-4658.2010.07949.x

Cytochrome P450 monooxygenases (P450s) are a diverse collection of enzymes acting on various endogenous and xenobiotic molecules Most of them catalyse hydroxylation reactions and one group of possible substrates are fatty acids and their related structures In this minireview, the signifi-cance of P450s in microbial fatty acid conversion is described Bacteria and yeasts possess various P450 systems involved in alkane and fatty acid deg-radation, and often several enzymes with different activities and specificities are retrieved in one organism Furthermore, P450s take part in the forma-tion of fatty acid-based secondary metabolites Finally, there are a substan-tial number of microbial P450s displaying activity towards fatty acids, but

to which no biological role could be assigned despite the often quite intense research

Abbreviations

GPo1, alkane hydroxylase; P450, cytochrome P450 monooxygenase; psi factor, precocious sexual inducers.

CYP94A1 from the plant Vicia sativa is the first

mem-ber of subfamily A of family 94 (there is more about

the activity of this enzyme in the accompanying

mini-review by Pinot & Beisson [4]) Over 11 000 P450

genes are included in this internationally used

nomen-clature system, and probably an extensive number of

nonclassified genes are present in the emerging pile of

genome sequencing data Besides being vast, the P450

superfamily is also highly divergent For example, in

plant P450s, CYP94 identity may be as low as 16%

with CYP51 and even only 12% with CYP74

Diver-gence is further illustrated by the huge number of

fam-ilies (977) and subfamfam-ilies (2519)

Phylogenetic studies have suggested that the

unicel-lular ancestor of plants had at least three distinct P450

branches [2] The first, sharing a common ancestor

with animal P450s involved in xenobiotic metabolism,

gave rise to Group A, catalysing typical plants

reac-tions (i.e synthesis of lignin, flavonoids, etc.) Group

non-A is composed of two branches The first

origi-nates from a common ancestor with animal CYP4 and

fungal CYP52 fatty acid hydroxylases; plant fatty acid

hydroxylases of the families CYP86 and CYP94

origi-nate from this branch The second branch of Group

non-A shares a common ancestor with animal, fungal

and bacterial sterol oxidases (CYP51) This branch

gave rise to the plant obtusifoliol 14 demethylase

(CYP51) and brassinolide hydroxylases (CYP85 and

CYP90 families)

Microbial P450s and fatty acids

Fatty acids are simple, yet indispensable, molecules to

any living cell Incorporated in phospholipids, they

make up the major part of the plasma membrane and

besides structural roles, they also function as a carbon

or energy source Furthermore, modified fatty acids

are building blocks for other complex molecules or act

as signalling molecules to trigger physiological

changes All these roles and processes require specific

enzymes, and cytochrome P450 monooxygenases

(P450s) make up a significant part of them P450s are

heme–thiolate proteins involved in the hydroxylation

of a wide range of endogenous and xenobiotic

com-pounds They are present in every eukaryotic organism

and in a substantial number of prokaryotes More

than 3800 microbial P450s are known to date [5] and

we estimate that  10–17% of them display activity

towards fatty acids or related structures On the one

hand, these activities are linked to the degradation of

fatty acids and alkanes; metabolization of these latter

compounds is inherently coupled to fatty acid

degrada-tion because the conversion of alkanes to fatty acids is

an essential step in the alkane assimilation process On the other hand, P450s are also involved in the synthe-sis of special fatty acid-based molecules such as sec-ondary metabolites or signal molecules Although the P450s described in this review all act on fatty acid sub-strates, this is not reflected in their overall similarity; according to Nelson’s classification system based on amino acid identity they belong to various families [5] Besides the involvement in different physiological functions, P450s also differ in the position of the hydroxylation; this may occur close to the carboxyl group, giving rise to a- or b-hydroxylated fatty acids (mediated by CYP152), in-chain (e.g CYP1006) or at the terminal or subterminal ending (e.g CYP52) Sev-eral classes of P450s involved in either metabolization

or biosynthesis processes and with different regiospeci-ficities are discussed Whenever possible, the interna-tionally used nomenclature is applied [3]

Fatty acid metabolism Alkane degradation

Microbial populations can break down almost every natural organic compound Even alkanes, which from

a chemical point of view are almost inert molecules, can be degraded and utilized as a carbon source by both bacteria and fungi Traditional aerobic alkane assimilation is initiated by terminal hydroxylation In the subsequent oxidation steps, the corresponding pri-mary alcohol is converted via an aldehyde to a fatty acid, which will enter b-oxidation [6] Depending on the particular microorganism, initial oxidation can be governed by several unrelated alkane hydroxylases Although microbial alkane degradation was first dem-onstrated about a century ago [7], research on this topic was only really boosted in the 1950s and 1960s when the production of single cell protein based on paraffin or alkanes became a hot topic The first alkane hydroxylase (GPo1) was found in the early 1960s in the soil bacteria Pseudomonas oleovorans (later renamed Pseudomonas putida) and was shown to be an integral membrane-bound nonheme di-iron monooxy-genase [8] More recently, related genes have been iso-lated from a broad range of bacteria [9] Whereas various yeasts were also well established as single cell protein producers, it took another few decades before their corresponding alkane hydroxylase enzymes were identified as cytochrome P450 monooxygenases, unre-lated to the bacterial GPo1 system [10] To date, all yeast alkane hydroxylases belong to the CYP52 family (membrane-bound class II) This family contains sev-eral enzymes with demonstrated activity towards

alkanes and⁄ or fatty acids, but also harbours numerous

genes which have not been explored (yet) or are proven

to be pseudogenes Well-studied members of the

CYP52 family are the enzymes of the

alkane-metabo-lizing yeasts Candida tropicalis, Candida maltosa and

Yarrowia lipolytica C tropicalis has at least 18 genes

belonging to this family Craft et al [11] evaluated 10

of them by quantitative competitive reverse

transcrip-tion-PCR; this highly specific technique is required

because of the high identity between the genes Five

genes were clearly induced by octadecane (CYP52A12,

-13, -14, -17 and -18), whereas no transcription was

detected for CYP52A15, -16, -19, -20 and D2 under

those and other conditions Also, C maltosa possesses

several so called P450alk genes: alk1 to alk8 Four

genes are considered to be the primary P450alk genes:

alk1, alk2, alk3 and alk5, corresponding to CYP52A3,

-A5, A4 and -A9, respectively Quadruple mutants

were unable to grow on alkanes as a sole carbon

source, but complementation by one of the four genes

restored growth on hexadecane [12] The corresponding

enzymes differ in their substrate chain-length

specific-ity, giving a possible reason for the multitude of

P450alk genes often present in one organism

Although the induction of P450alk genes in response

to n-alkanes or fatty acids is a common feature among

alkane-assimilating yeasts, the underlying molecular

mechanisms remain largely unknown C maltosa alk2

turned out to be inducible by alkanes, as well as by

the peroxisome proliferator clofibrate The respective

cis-acting elements in the alk2 promoter region were

identified and there are indications of similar motives

in other C maltosa alk promoters [13]

Y lipolytica is another well-studied alkanotrophic

yeast with 12 P450alk genes in its genome Six of the

eight tested genes showed induction by alkanes

Among them, alk1 (CYP52F1) displayed the highest

expression and although single disruptions in the other

genes did not result in yeasts unable to metabolize

alk-anes, Dalk1 mutants are unable to grow on decane In

addition, a Dalk1Dalk2 double mutant cannot grow on

hexadecane either Therefore, it is suggested that the

primary P450alk gene alk1 is required for assimilation

of decane and dodecane, whereas alk2 (CYP52F2) is

involved in the assimilation of molecules longer than

dodecane The other alk genes possibly act on even

longer alkanes or other types of carbon chains [14]

Recently, Ohta’s group shed light on the

transcrip-tional induction of the alk1 gene by alkanes [15]; in

the presence of alkanes, Yas1p and Yas2p, two basic

helix–loop–helix proteins form heterodimers and bind

to the cis-acting alkane-responsive element 1 in the

alk1 promoter The protein complex also binds to

other promoters of genes interfering in alkane degrada-tion such as the acetoacetyl-CoA thiolase gene and the yas1 gene itself This latter binding creates a positive autoregulatory feedback which results in a quick and profound transcriptional response to alkanes Further-more, a third regulatory protein was identified The repressor Yas3p binds specifically to Yas2p when no alkanes are present, but when the yeast is exposed to alkanes, Yas3p is translocated from the nucleus to the endoplasmatic reticulum Conserved motives of the alkane-responsive element 1 sequence were retrieved in the P450alk promoters of C maltosa, C tropicalis and Debaryomyces hansenii, suggesting a common mecha-nism for alkane-responsive induction

Although Cardini and Jurtshuk [16] provided strong indications for the involvement of a bacterial P450 in the hydroxylation of octane in Rhodococcus rhodoch-rous, the role of bacterial P450s in alkane degradation

in addition to the well-established alkane hydroxylase system has long been underestimated The first bacte-rial P450alk was cloned from Acinetobacter calcoaceti-cus in 2001, this class I P450 was assigned to the new family CYP153 [17] Recently, similar enzymes were found in other alkane-utilizing species such as Sphingo-monassp HXN200, Mycobacterium sp HXN1500 and Alcanivorax borkumensis Van Beilen et al [18] were able to demonstrate the functionality of seven of ele-ven genes (CYP153A6, -A7, A11, A13, A14 and -D1)

by functional expression in a DGPo1 P putida strain and its restored ability to grow on alkanes Most alk-anotrophes attack various or mixed substrates with different and often length-specific enzymes, reflected in the variation among yeast CYP52 genes and the bacte-rial alkane hydroxylase system A similar trend can be observed for the bacterial P450alk enzymes For exam-ple, Sphingomonas sp HXN200 possesses five CYP153 genes, three of which show activity towards C5–C10 alkanes, whereas no affinity for these substrates was observed for the other two genes, suggesting that these genes are either pseudogenes or act on different sub-strates such as long-chain alkanes Other organisms possess several types of enzymes; Al borkumensis con-tains two alkane hydroxylases and two CYP153s [19] and whereas the first alkane hydroxylase is essential for the degradation of C6, no clear role could be assigned to the second, but double knockouts resulted

in deficient growth on C8–C16 Because this organism

is, thanks to its efficient and broad-spectrum hydrocar-bon-degrading capacities, a dominant microbe in oil-polluted waters, the CYP153 enzymes are postulated

to cover the rest of the alkane length range

In general, one associates alkane-degrading organ-isms with oil-polluted environments, but in fact

alkane-degrading enzymes are found in many more

organisms than the ones strictly appearing in

oil-related niches Indeed, alkanes are persistent molecules

and this property makes them the perfect components

for natural barriers such as plant and insect cuticles

The major component of plant cuticles is cutin, a

poly-mer consisting of omega hydroxy fatty acids

cross-linked by ester and epoxide bonds, which is

impreg-nated and covered with waxes (more information on

the constitution and biosynthesis of plant envelopes

can be found in the accompanying minireview by Pinot

& Beisson [4]) These cuticular and epicuticular waxes

are a mixture of long-chain alkanes (C16–C30) and

related structures For example, in the rice blast fungus

Magnoporthe oryzae, a putative alkane-degrading

cyto-chrome P450 (MGG_05908.5 or CYP584A2) is

upreg-ulated upon the first stages of infection, together with

other genes regarding the utilization of

nonconvention-al carbon sources [20] These findings suggest that, on

the one hand, alkane degradation is required for

breaking the plant’s defence but, on the other hand,

alkanes serve as nutritional input during the initial

col-onization steps Insect cuticle consists of protein and

chitin, and is covered by a highly resistant lipid layer:

the epicuticle This epiculticle is composed of

hydro-carbons, wax esters, fatty alcohols and fatty acids

Hydrocarbons are the prevalent component and

include alkanes, alkenes and methyl-branched chains

in various ratios, depending on the insect species In

general, alkanes make up the biggest part of the

hydrocarbon moiety and their chain length ranges

between C21 and C35, with a particular preference for

odd-numbered chains [21] The entomopathogenic

fun-gus Metarhizium anisopliae infects a broad range of

insects by direct penetration of the host cuticle and

hence can be exploited as a biological control agent of

pests cDNA microarray analysis of the fungus grown

on cuticular extracts revealed a clear upregulation of

an alkane-inducible cytochrome P450 (AJ273607)

during the first hours of incubation [22] According to amino acid similarity, this gene should be classified in the CYP52 family Several other expressed sequence tag (EST) or microarray cDNA analysis regarding en-tomopathogenic infection concealed involvement of P450s supposed to hydroxylate alkanes and⁄ or fatty acids

Fatty acid degradation – omega oxidation Although cytochrome P450s do not intervene in the degradation of fatty acids in the b-oxidation cycle itself, they do take part in the steps preceding b-oxida-tion (Fig 1) As menb-oxida-tioned in the previous secb-oxida-tion, alkanes can be converted to common fatty acids by initial cytochrome P450 interference However, the same enzymes responsible for terminal alkane oxida-tion often also mediate subterminal oxidaoxida-tion The corresponding secondary alcohols are oxidized to ketones and a Baeyer–Villiger monooxygenase converts them

to esters, which are cleaved to give rise to a primary alcohol and a common fatty acid

Furthermore, common fatty acids can be terminally oxidized by a cytochrome P450 monooxygenase and the resulting hydroxy-fatty acid is further converted to

a dicarboxylic acid, which then enters b-oxidation This so-called x-oxidation occurs both in bacteria and yeasts, and yet is far more documented for this latter group because medium- and long-chain dicarboxylic acids are commercially produced by yeast fermenta-tions to serve as building blocks of, for example, per-fume, polymers, high-quality lubricants or macrolide antibiotics

In this respect, the previously discussed fungal CYP52 enzymes are versatile enzymes; several isoforms exhibit different activities and specificities towards alk-anes, as well as towards fatty acids or related struc-tures An individual CYP52 not only demonstrates a distinct substrate specificity regarding chain lengths,

Fig 1 Assimilation of alkanes and fatty

acids The (putative) involvement of

cyto-chrome P450 monooxygenases is indicated

by grey arrows.

but in addition has preferences for either alkanes or

fatty acids, and similarly, compared with the chain

lengths, alkane⁄ fatty acid specificities of the various

enzymes within species are often overlapping The

C maltosa P450Cm1 (ALK1 or CYP52A3), for

exam-ple, prefers alkanes, whereas P450Cm2 (ALK3 or

CYP52A4) shows highest affinity towards fatty acids

Nevertheless, both enzymes are able to hydroxylate

both substrates However, the C maltosa CYP52A10

and CYP52A11 enzymes are only able to convert fatty

acids [23,24] It is not completely clear which gene(s)

exactly mediate dicarboxylic acid formation, but

Kog-ure et al [25] demonstrated that the alk5 gene

(CYP52A9) was highly induced when comparing a

C maltosa dicarboxylic acid overproducing mutant

with a reference strain Interestingly, P450alk5 is

indeed the isozyme with the strongest tendency to

x-hydroxylate fatty acids By contrast, in vitro

experi-ments after heterologous expression demonstrated that

the highly active CYP52A3, although alkane

prefer-ring, not only converts hexadecane to its primary

product 1-hexadecanol, but also further oxidizes this

component to hexadecanal, hexadecanoic acid,

1,16-hexadecandiol, 16-hydroxyhexadecanoic acid and even

1,16-hexadecanedioic acid, in this way bypassing the

two enzymes normally involved in x-oxidation [26]

The authors did not verify this phenomenon for other

fatty acid-oxidizing C maltosa isozymes, but one can

postulate that CYP52A9 generates a similar oxidation

cascade This assumption is supported by data from

another dicarboxylic acid-producing strain Just like

C maltosa, C tropicalis mutant strains are used in

industrial dicarboxylic acid production, the strains are

among others blocked in b-oxidation by inactivation

of the POX genes, resulting in higher dicarboxylic acid

yields Upon exposure to oleic acid, CYP52A13 and

CYP52A17 are strongly and consistently induced

Again, enzymatic tests after heterologous expression

revealed the enzyme’s abilities to synthesize

dicarbox-ylic acids CYP52A17 shows the greatest overoxidizing

capacities not only regarding substrate chain length,

but also concerning activity; the conversion of oleic

acid to its diacid occurs twice as quickly as the

forma-tion of x-hydroxy oleic acid [27] Despite the in vitro

evidence, there is no clear answer to the question of

which role this P450 oxidation cascade plays in vivo

One can assume that the prevalence of different

enzyme systems contributes to the yeast capacities to

grow efficiently on a broad chain-length range of

alk-anes and fatty acids (e.g C7–C40 for C maltosa)

Another advantage of the P450 bypass is the

circum-vention of H2O2 formation by the fatty alcohol

oxi-dase However, the overoxidation cascade requires

three molecules of NADPH, potentially creating metabolic limitations In classical x-oxidation, only one NADPH molecule is used, whereas the alcohol dehy-drogenase delivers one reducing equivalent (NADH) Members of the CYP52 family are supposed to be all linked to alkane and⁄ or fatty acid hydroxylation, yet these assumptions are made based on the amino acid sequence and some enzymes might be involved in unrelated biological processes CaAlk8, for example, is the only CYP52 member originating from C albicans (CYP52A21) Although it has been shown that the enzyme terminally and subterminally hydroxylates lau-ric, myristic and palmitic acid [28], and is involved in alkane degradation, Panwar et al [29] suggested that CYP52A21 takes part in conferring multidrug resis-tance to the opportunistic pathogen C albicans Dis-ruption of CYP52A21 in the wild-type strain did not lead to a drug-sensitive strain, probably attributed to the presence of several other multidrug resistance mechanisms Nevertheless, the role of CYP52A21 in multidrug resistance was demonstrated by overexpres-sion in Saccharomyces cerevisiae and in a hypersensi-tive C albicans host, rendering the latter resistant to fluconazole, itraconazole and 4-nitroquinoline oxide

In addition, experiments with C albicans microsomes indicate that resistance is caused by CYP52A21-medi-ated drug modification

Besides the yeast species discussed above, Coryne-bacteriumsp is also known as a producer of dicarbox-ylic acids [30] P450s are probably involved, but the exact pathway remains unrevealed By contrast, a spe-cific P450 could be put forward as a candidate for x-oxidation in the cyanobacteria Anabaena variabilis Although CYP110 is induced by alkanes such as hex-adecane, CYP110 does not participate in alkane degra-dation; findings which are supported by the fact that alkanes are toxic for Anabaena variabilis Based on the sequence similarity and binding affinity experiments with fatty acids, it was suggested that CYP110 is related to fatty acid x-oxidation of saturated and (poly)unsaturated fatty acids and subsequent forma-tion of dicarboxylic acids which then undergo b-oxida-tion [31] The alkane inducibility of the cyp110 mRNA was used to design a hexadecane biosensor system [32]

Biosynthesis of a- and b-hydroxylated fatty acids

Not only are hydroxylated fatty acids intermediates in alkane and fatty acid metabolization, they can also be useful components as such a-Hydroxylated long-chain fatty acids, for example, are important constituents of sphingolipids These lipids are essential components of

mammalian cell membranes, but can also be found in

some bacterial and fungal cell membranes

Interest-ingly, bacterial sphingolipid synthesis predominantly

occurs in anaerobic species One such anaerobe is

Sphingomonas paucimobilis Its sphingolipids are rich

in a-hydroxymyristic acid and the enzyme responsible

for the conversion of myristic acid has been identified

as a member of the P450 superfamily: P450SPa or

CYP152B1 [33] However, the enzyme utilizes H2O2

instead of O2 and does not require reducing

equiva-lents, comparable with the peroxide shunt reaction for

common P450s [34] Indeed, P450SPa lacks the critical

residues compulsory for the fulfilment of the typical

monooxygenation reaction; in the distal I helix the

Asp⁄ Glu–Thr proton delivery system is substituted by

Arg–Pro and, albeit the heme-binding cysteine was

retrieved, the preceding consensus motive involved in

electron transfer was modified [35,36] Furthermore, it

sounds quite logical that anaerobic organisms try to

circumvent the use of molecular oxygen The enzyme

is highly specific towards fatty acids; no alkanes, fatty

alcohols or fatty aldehydes are hydroxylated In

addi-tion to myristic acid, the enzyme shows activity to

slightly shorter or longer saturated fatty acids and

ara-chidonic acid [37] Based on a database similarity

search with the P450SPagene, Matsunaga’s group were

able to identify another fatty acid hydroxylase in

Bacil-lus subtillis However, this P450Bsb or CYP152A1

enzyme is less regiospecific; myristic acid is converted

to a mixture of a- and b-hydroxymyristic acid, with

a slightly higher amount of the b-hydroxylated

product [38] A few years ago, P450CLA (CYP152A2)

from the anaerobe Clostridium acetobutylicum was

characterized: like P450Bsb, this enzyme performs both

a- and b-hydroxylations of saturated and unsaturated

fatty acids, but in this particular case the a-position is

preferred [39]

Although there is a link between the occurrence of

fatty acid a-hydroxylation and sphingolipids, some

questions remain: what is the biological role of

P450Bsb in the non-sphingolipid-producing B subtillis

and why is a mixture of a- and b-forms produced? It

was suggested that peroxide-utilizing P450s might be

involved in the oxygen-detoxification system of

Closti-dium acetobutylicum [39] Although anaerobic,

Clos-tridium species can tolerate microoxic conditions

(< 5% O2) Besides the classical oxygen detoxification

systems, heme oxygenases, oxidases and lipid

per-oxidase scavenging enzymes are involved in the

estab-lishment of the anoxic microenvironment [40]

Furthermore, P450SPa turned out to be capable of

hydroxylating phytanic acid (3,7,11,15-tetramethyl

hexadecanoic acid), a degradation product of chlorophyll

[41] This branched fatty acid cannot undergo b-oxidation because of methylation at the b-position

In humans, metabolization occurs by an initial a-oxidation step, resulting in the removal of one car-bon instead of two [42] The subsequent pristanic acid will be entirely degraded by b-oxidation Oxidation of phytanoyl-CoA is mediated by phytanoyl-CoA dioxy-genase, an iron requiring non-heme oxidoreductase Homologue enzymes can be retrieved in a wide variety

of bacteria, questioning the role of P450 in bacterial a-oxidation of phytanic acid

Fatty acid hydroxylating P450s involved in secondary metabolite synthesis

Biosurfactants Biosurfactants are surface-active compounds capable

of reducing interfacial tension between liquids thanks

to their amphiphilic properties Amphiphilic molecules consist of a hydrophilic and a hydrophobic moiety that interacts with the phase boundary in heterogeneous systems, allowing them to, for example, act as a deter-gent, wetting agent or emulsifier of oil⁄ water mixtures

In general, common fatty acids or b-hydroxy fatty acids originating from b-oxidation act as the hydro-phobic part However, in some particular cases, P450 hydroxylated fatty acids make up the hydrophobic tail One such example are the cellobiose lipids produced

by several yeast-like fungi (Fig 2A) Ustilagic acids from the plant-pathogen Ustilago maydis contain 15,16-dihydroxypalmitic acid or 2,15,16-trihydroxypal-mitic acid Teichmann and co-workers [43] elucidated the biosynthetic gene cluster harbouring two P450 genes Upon disruption of the first P450 gene, cyp1, no cellobiose lipids could be detected CYP1 catalyses the conversion of palmitic acid to juniperic acid and this terminal hydroxylation is essential for the covalent binding of the hydroxylated fatty acid to the cellobiose moiety By contrast, Dcyp2 mutant strains retain their capacities to secrete cellobiose lipids Yet, these mole-cules lack the typical hydroxylation at the subterminal position These findings prove that CYP1-dependent x-hydroxylation does not depend on prior subterminal hydroxylation and that both enzymes are highly selec-tive for either the terminal or subterminal position Surprisingly, CYP1 and CYP2 share only 15% amino acid identity and despite their activity towards the ter-minus of fatty acids, they are not classified into the CYP52 family but, based on their amino acid sequence, are assigned CYP5025A1 and CYP5030A1, respectively

No P450 activity is linked to a-hydroxylation; this

reaction is governed by AHD1, a non-heme diiron

oxi-do-reductase cyp1 homologues were retrieved in two

other cellobiose lipids producing organisms,

Pseu-odozyma flocullosa and Pseuodozyma fusiformata [44]

One was not able to provide direct evidence for cyp1

contribution to the biosurfactant synthesis, but a

posi-tive correlation between cyp1 expression and

flocculo-sin synthesis was demonstrated

Sophorolipids are another type of hydroxyfatty acid

containing biosurfactants They consist of a x- or

x-1-hydroxylated fatty acid etherified via its hydroxylgroup

to a sophorose unit (Fig 2B) The fatty acid can be

palmitic, palmitoleic, stearic, oleic or linoleic acid

Typical sophorolipid-producing organisms are the

yeasts Candida bombicola and C apicola P450

involve-ment in sophorolipid biosynthesis was suggested from

a simultaneous increase of cellular P450 content in

C apicola Two cyp52 genes were cloned from C

api-cola (CYP52E1 and CYP52E2) Yet, Southern

hybrid-ization results indicated the existence of additional

cyp52 sequences, making it hard to draw conclusions

on the role of CYP52E1 and -E2 [45] C bombicola, another sophorolipid-producing yeast closely related to

C apicola, harbours at least eight cyp52 genes One of these, CYP52M1, was highly induced upon sophoroli-pid production, suggesting its participation in the sophorolipid biosynthesis pathway [46]

Polyketides Polyketides are a structurally very diverse family of secondary metabolites with different biological activi-ties occurring in bacteria, plants and animals They are synthesized by polymerization of acetyl and propionyl

in a similar process to fatty acid synthesis and can undergo extensive derivatization; in many cases, macr-olidic structures are formed which are further modified

by, for example, several hydroxylation steps (Figs 3 and 4) These hydroxylation steps are very often medi-ated by cytochrome P450 monooxygenases

Discussing biosynthesis of all microbial polyketides would take us too far from the scope of this review, but we focus on two well-described polyketides of which the backbone structure displays fatty acid simi-larity Fumonisin is a mycotoxin produced by several Fusarium species, among others Fusarium verticilloides and Fusarium proliferatum, both widespread plant-pathogens infecting maize and other grains, rendering this mycotoxin a common contaminant of corn Fumonisin is hepatotoxic and nephrotoxic, but its acute toxicity is low The long-term effect of low con-centrations is less clear, but fumonisin is suggested to

be carcinogenic The 17 fumonisin biosynthetic genes are located in a gene cluster and three of them are P450 enzymes The FUM6 protein (CYP505B1) inter-venes in one of the first steps in fumonisin synthesis by hydroxylation of the polyketide-amino acid at C-14 and C-15 (Fig 3) Fum6 deletion mutants of F verticil-loidesare unable to produce fumonisin-like compounds because the hydroxylgroups are required for the esteri-fication of tricarballylic moieties downstream of the biosynthesis pathway FUM6 is a self-sufficient P450 containing a NADPH-dependent reductase domain and belongs to the same family (CYP505) as the first discovered self-sufficient eukaryotic P450: P450foxy from F oxysporum The second cytochrome P450 monooxygenase, FUM2 (CYP65AH1), most likely catalyses fumonisin C-10 hydroxylation, whereas the third, FUM15 (CYP617F1), is suggested to be respon-sible for the synthesis of low levels of a fumonisin iso-form [47]

Well-studied examples of bacterial polyketides are the antifungal components typically produced by soil actinomycetes Nystatins produced by Streptomyces

A

B

Fig 2 Structure of (A) cellobiose lipids produced by Ustilago

may-dis; n = 2 or 4 (B) A common sophorolipid molecule in the acidic

form R=H or COCH 3

noursei are commercialized as an antibiotic to treat

Candidasp and Cryptococcus sp infections (Fig 4A)

The nystatin backbone is composed of a 38-membered

macrolactone ring which can be further modified by

cytochrome P450 enzymes; NysN (CYP105H1)

oxi-dizes the methyl group at C-16 and NysL (CYP161A1)

performs a hydroxylation at the C-10 position DnysL

mutants produce 10-deoxynystatin, but despite the

absence of the hydroxyl group, the product retains its

antifungal activity [48] Streptomyces nodosus

synthe-sizes amphotericin; in addition to its antifungal

proper-ties, this antibiotic is also active against human

immunodeficiency virus, Leishmania parasites and

prion diseases Amphothericin has a 38-membered

macrolacton backbone similar to the one of nystatin

(Fig 4B) and also this biosynthetic cluster contains

two P450 genes, amphL and amphN or CYP161A3 and

CYP105H4 respectively AmphL mediates C-8

hydrox-ylation, whereas AmphN oxidizes the C-16 methyl

group [49] The biosynthetic gene clusters for a number

of polyketides have been elucidated (amphotericin,

nystatin, candicidin, pimaricin and rimocidin) All

these clusters contain a P450 gene homologue to

amp-hN and NysN associated to the C-16 oxidation These

polyketide-specific P450 sequences can be used for

screening purposes This strategy led to the identification

of a nystatin-like gene cluster in Pseudonocardia autrophicacontaining the typical nppL and nppN genes [50] and the isolation of putative polyketide producing actinomycetes [51]

Oxylipins Fatty acids have an established role as building blocks of membranes and triacylglycerols, acting as a structural component or energy reservoir Besides their classical roles, fatty acid derivates act as signal-ling molecules with great physiological significance One such example are oxylipins, molecules originating from oxidized unsaturated fatty acids They are wide-spread in aerobic organisms such as plants, animals and fungi, but also occur in certain bacteria Also the well-described mammal prostaglandins and leuko-trienes belong to the oxylipin family Although oxyli-pin synthesis in mammals and plants is well documented, far less information can be found on microbial oxylipins and reports on the cloning of responsible genes are scarce Oxylipin-forming enzymes are structurally very diverse; in plants they belong to an atypical cytochrome P450 subfamily, whereas most other lipoxygenases are non-heme iron-containing proteins

Fig 3 Part of the biosynthetic pathway of

fumonisin in Fusarium sp R1= H or OH

(hydroxylation is rare and is supposed to be

governed by Fum15p), R 2 = H or OH Steps

with P450 involvement are marked with an

arrow.

Oxylipins of fungal species are believed to be

involved in signalling, but more research is required

to assign specific functions Aspergillus nidulans is a

model organism for the understanding of fungal

development because it has a defined sexual and

asex-ual development cycle Oxylipins regulate the balance

between both cycles Furthermore, they regulate

sec-ondary metabolism and are in this way important for

plant–host colonization and mycotoxin production

[52] The involved oxylipins are called precocious

sex-ual inducers (psi-factors) and are derived from

unsat-urated C18 fatty acids One such psi factor-producing

oxygenases or Ppo enzyme is PpoA, a bifunctional

protein with a fatty acid heme dioxygenase⁄

peroxi-dase domain in its N-terminal region and a P450

heme–thiolate domain in its C-terminal region The

enzyme first oxidizes linoleic acid to

(8R)-hydrox-yperoxyoctadecadienoic acid and transfers this

product in the second reaction step to

5,8-dihydro-xyoctadecadienoic acid by means of the P450 domain

PpoA acts in vitro on unsaturated C16 and C20 fatty

acids as well, and was assigned as CYP6001A1 [53]

Several CYP6001 homologues can be retrieved among

fungi such as other Aspergillus sp., Neurospora sp.,

Fusariumsp and Ustilago maydis, although their

function is not confirmed in these latter species

[54,55]

Another enzyme taking part in oxylipin biosynthesis

is PpoC (CYP1006C1) Like PpoA, two different

heme-containing regions are present, but in contrast to

PpoA the P450 heme–thiloate domain is degenerated;

the conserved cysteine residue known as the fifth heme

iron ligand is replaced by a glycine or phenylalanine in

A nidulans or A fumigatus PpoC, respectively This

is reflected in the enzymes’ activity; whereas PpoA further converts (8R)-hydroxyperoxyoctadecadienoic acid to 5,8-dihydroxyoctadecadienoic acid, PpoC only performs the first reaction step [56]

Oxylipins are also present in bacteria and might take part in stress responses and host–pathogen interactions Most bacterial lipoxygenases are non-heme iron pro-teins, but few plant CYP74-like proteins can be retrieved, for example, in the rhizobacterium Methylo-bacterium nodulans This bifunctional protein possesses

an N-terminal peroxidase region and a C-terminal CYP74-like P450 region Because

Methylobacteri-um nodulansis a root-nodule-forming and nitrogen-fix-ing symbiont of Crotalaria (plants belongnitrogen-fix-ing to the Fabaceae family), it is plausible that this bacterial lipox-ygenase originated from horizontal gene transfer [57]

Fatty acid-acting P450s with unclear biological function

Self-sufficient P450 The majority of the P450 monooxygenases obtain the necessary electrons for oxygen cleavage and substrate hydroxylation via one or two redox partners There are several P450 redox systems which can be classified into different groups according to the components involved (reviewed in [58]) Most eukaryotic micro-somal P450s – among them the previously discussed CYP52 family – use the class II redox system They form a small electron transfer chain together with the NADPH cytochrome P450 reductase Both enzymes

O O

H3C

H3C

H2N

H3C

H3C

H3C

H2N

H3C

CH3

CH3

O

A

B

OH OH

OH OH OH

OH

OH

O

HO

O OH O

H

NysL

NysN

AmphL

O OH

OH

O

OH

O

O

O OH HO

AmphN

Fig 4 Structure of (A) nystatin produced by Streptomyces noursei, (B) amphotericin pro-duced by Streptomyces nodosus P450-mediated hydroxylation or oxidations are marked with an arrow.

are N-terminally anchored to the endoplasmatic

reticu-lum and derived structures, and the body of the

pro-tein is located in the cytoplasmatic space [59]

Cytochrome P450 reductase is a flavoprotein

contain-ing the flavin cofactors FAD and FMN It transfers

the hydride ion of NADPH to the lower redox

poten-tial FAD FAD then transfers single electrons to

FMN, which in turn reduces the cytochrome P450

monooxygenase heme centre as required to activate

molecular oxygen [60] By contrast, most prokaryotic

P450s are cytosolic and communicate with two

sepa-rate redox partners: a NAD(P)H-binding and

FAD-containing reductase and a ferrodoxin or flavodoxin

that transfers the electrons from the reductase to the

P450 heme These redox systems are referred to as

class I Yet, in the 1980s, a catalytically self-sufficient

119 kDa protein was characterized and purified from

Bacillus megaterium: BM-3 or CYP102A1, now

posi-tioned in the class VIII redox family This large

pro-tein is a gene fusion product between a P450 and a

cytochrome P450 reductase, rendering electron transfer

extremely efficient and resulting in a catalytic activity

of 4600 nmol fatty acid per nmol P450 per minute,

whereas most class II systems have catalytic activities

of two or even three orders of magnitude lower [61]

The self-sufficient and soluble properties simplify the

enzyme’s overexpression and purification, and turn it

into an ideal model for spectroscopic and structural

studies The derived models for the heme–substrate

interactions indeed gave substantial input for the

understanding of mammalian P450 systems [62,63]

Despite the available data on the enzyme’s structure

and in vitro substrates, its proper biological role and

natural substrate remain to be revealed BM-3

hydroxylates fatty acids with a chain length between

12 and 18 carbon atoms at the x-1, x-2 or x-3

posi-tion and has highest affinity towards pentadecanoic

and palmitic acid (Km of 2 lm) Unsaturated fatty

acids are even better substrates and besides the typical

x-1, x-2 or x-3 hydroxylation, additional epoxidation

of double bonds can occur Turnover rates of

> 15 000 have been reported for arachidonic acid

(C20:4) hydroxylation Saturated fatty acid amides and

fatty alcohols are hydroxylated as well, yet at lower

efficiencies (reviewed in [64])

BM-3 not only displays structural similarity with

mammalian P450s, it also shows an induction profile

very similar to mammalian CYP4A enzymes These

P450s are fatty acid x-hydroxylases which are induced

by barbirturates and other peroxisomal proliferators

Also in the 5¢-flanking region of the BM-3 gene a

so-called Barbie-box is retrieved; motives occurring in

all barbiturate-inducible genes The regulatory system

includes the positive transcription factor BM3P1, the autoregulated repressor Bm3R1 and several regulatory sites English et al [65] found that the branched fatty acid phytanic acid is not only an inducer of BM-3, but also a substrate that is converted to x-1 hydroxyphy-tanic acid B megaterium is a soil bacterium and the authors state that many plant-derived unsaturated fatty acids are extremely toxic to this bacteria; the induction of BM-3 by phytanic or other fatty acids may contribute to a metabolization or detoxification system However, it must be mentioned that phytanic acid is not toxic to B megaterium, although it is a major vegetative breakdown product occurring in the soil Furthermore, branched chain fatty acids make up 80% of the fatty acid content of the Bacillus sp mem-branes and when the hydroxylation of these substrates was studied in more detail, they were shown to be at least as good substrates as their straight chain ana-logues, having a higher regio- and stereospecific hydroxylation pattern Therefore, it is possible that BM-3 takes part in the oxidative degradation of branched chain fatty acids [66]

Its high catalytic activity, elucidated protein struc-ture and ease of expression and use in in vivo experi-ments have made BM-3 an attractive target for protein engineering with possible biotechnological applications Various mutants are described as being able to act on shorter fatty acids, polycyclic aromatic hydrocarbons and even gaseous alkanes [67–69] or displaying a shift

in the hydroxylation pattern towards the terminal or internal positions [70,71]

Meanwhile, the self-sufficient CYP102A family has been extended with > 10 members, mainly originating from soil bacteria CYP102A2 and -A3 from B subtil-lis, CYP102A5 from B cereus and CYP102A7 from

B licheniformishave been characterized and in general hydroxylate the same substrates as BM-3, sometimes with even higher activities [72] This group of proteins even harbours a sequence of a noncultured soil bacte-rium obtained by screening a metagenome database [73] Again, the biological roles in the different organ-isms remain to be discovered, but it has been demon-strated that CYP102A2 and -A3 are nonessential genes and are not involved in the adaptive response concern-ing fatty acid detoxification [74] The same conclusion can be drawn for CYP102B1 CYP102B1 is a cofactor requiring arachidonic acid hydroxylating and epoxydiz-ing P450 from Streptomyces coelicolor No differences concerning cell development or antibiotic production were observed when comparing Dcyp102b1 strains with wild-type strains, but the lipid profiles of both strains were quite different, suggesting the involve-ment of CYP102B1 in lipid biochemical pathways