Fungal secondary metabolism — from biochemistry to genomics

© 2005 Nature Publishing Group

*University of Wisconsin–Madison, Department of Plant Pathology, 882 Russell Labs, 1630 Linden Drive, Madison, Wisconsin 53706, USA.‡Department of Molecular Biology and Biotechnology, Firth Court, University of Sheffield, S10 2TN, UK.§Department of Cell and Molecular Biology, Tulane University, New Orleans, Louisiana 70118, USA.e-mails: [email protected]; [email protected]; [email protected]:10.1038/nrmicro1286

The fungal kingdom includes many species with unique and unusual biochemical pathways. The products of these pathways include important pharma-ceuticals such as penicillin, cyclosporin and statins; potent poisons, including aflatoxins and tricho-thecenes; and some Janus-faced metabolites that are both toxic and pharmaceutically useful, such as the ERGOT ALKALOIDS. All of these natural products, along with many other low-molecular-weight fun-gal metabolites, are classified together as secondary metabolites and have been reviewed elsewhere1–3.

What are secondary metabolites? Julian Davies once joked that “the simplest definition is that they are not generally included in the standard metabolic charts”4. Secondary metabolites are customarily distinguished from primary metabolites, which are the almost universally distributed compounds of INTERMEDIARY

METABOLISM. Primary metabolism has been the domain of biochemists, whereas until recently, secondary metabolism has largely been the domain of organic

chemists. Secondary metabolites are often bioactive, usually of low molecular weight, and are produced as families of related compounds at restricted parts of the life cycle, with production often correlated with a specific stage of morphological differentiation. Secondary metabolites share the enigmatic properties of cellular dispensability — producer organisms can grow without synthesizing these metabolites — and restricted taxonomic distribution (only a small group of organisms produces each metabolite)5,6.

The systematic study of fungal secondary meta-bolites began in 1922 under the leadership of Harold Raistrick, who eventually characterized more than 200 mould metabolites7. However, it was not until the discovery and development of penicillin BOX 1 that widespread attention was focused on fungal meta bolites. Pharmaceutical companies instigated extensive screening programmes, and by 1950 a treasure trove of microbial products with pharma-ceutical applications had been discovered. This

FUNGAL SECONDARY METABOLISM FROM BIOCHEMISTRY TO GENOMICSNancy P. Keller*, Geoffrey Turner‡ and Joan W. Bennett§

Abstract | Much of natural product chemistry concerns a group of compounds known as secondary metabolites. These low-molecular-weight metabolites often have potent physiological activities. Digitalis, morphine and quinine are plant secondary metabolites, whereas penicillin, cephalosporin, ergotrate and the statins are equally well known fungal secondary metabolites. Although chemically diverse, all secondary metabolites are produced by a few common biosynthetic pathways, often in conjunction with morphological development. Recent advances in molecular biology, bioinformatics and comparative genomics have revealed that the genes encoding specific fungal secondary metabolites are clustered and often located near telomeres. In this review, we address some important questions, including which evolutionary pressures led to gene clustering, why closely related species produce different profiles of secondary metabolites, and whether fungal genomics will accelerate the discovery of new pharmacologically active natural products.

NATURE REVIEWS | MICROBIOLOGY VOLUME 3 | DECEMBER 2005 | 937

R E V I E W S


ERGOT ALKALOID Any of a group of about 30 indole alkaloids obtained from the sclerotial phase of the fungus Claviceps purpurea.

INTERMEDIARY METABOLISMEnzyme-catalysed processes within cells that metabolize macronutrients, carbohydrate, fat and protein.

ALLELOPATHICDescribes secondary metabolites released by plants, bacteria, fungi or viruses that have a direct or indirect, harmful or even beneficial effect on another organism.

search for bioactive secondary metabolites has continued unabated, and thousands of compounds that inhibit the growth of bacteria, fungi, protozoa, parasites, insects, viruses and even human tumour cells have been discovered. Many other molecules with cytotoxic, mutagenic, carcino genic, teratogenic, immunosuppressive, enzyme inhibitory, ALLELOPATHIC and other biological effects also have been found. A recent literature survey of fungal metabolites, which examined 1,500 compounds that were isolated and characterized between 1993 and 2001, showed that

more than half of these molecules had antibacterial, antifungal or antitumour activity8. Given the empha-sis on physiological activity, one common classifica-tion method for secondary metabolites is through defining their impact on human interests, for exam-ple, plant and animal toxins, growth hormones and pharmaceuticals. Another, more chemically rational system of classification reflects the fact that, despite their enormous chemical complexity and diversity, all secondary metabolites arise from a limited number of precursors from primary meta bolism. This review

Box 1 | Penicillin

Penicillin, the first broad-spectrum antibiotic, is the most famous fungal secondary metabolite. Penicillin transformed the practice of medicine, changed the trajectory of pharmaceutical research, influenced the course of World War II and saved countless lives. The penicillin story has been told many times95–98. In 1929, Alexander Fleming discovered the antibacterial action of a ‘mould juice’ from Penicillium notatum and named the biological activity ‘penicillin’. A decade later, Howard Florey, working at Oxford University with Ernst Chain, Norman Heatley and others, purified enough penicillin to demonstrate its clinical efficacy, completing their first experiments during the evacuation of Dunkirk during World War II. Florey and Heatley, fearing a German invasion, travelled clandestinely to the USA with cultures of the mould. Penicillin yields were low and the surface-culture method used for growing the mould was cumbersome, so government scientists in Peoria, Illinois screened fungi from all over the world in search of a higher-yielding strain that could grow in submerged culture. An ardent technician named Mary Hunt, soon dubbed ‘Mouldy Mary’, scoured Peoria markets for mouldy produce, and found the rotting cantaloupe that eventually yielded a strain of Penicillium chrysogenum that was selected for large-scale production99. As scale-up methods were perfected in preparation for D-Day, chemists worked hard to elucidate the chemical structure of the antibiotic. It became apparent that there was more than one penicillin. The main metabolite obtained by surface culture of the Fleming strain of P. notatum was different from the main metabolite obtained by submerged fermentation of the Peoria strain of P. chrysogenum. These metabolites are now known, respectively, as 2-pentenylpenicillin (‘penicillin I’) and benzylpenicillin (‘penicillin II’)100.

It had been expected that a complete chemical synthesis for penicillin would replace fungal biosynthesis in commercial production. Instead, efficient industrial-scale fermentation methods were developed. These fermentation technologies were not only good preparation for industrial production of the streptomycin family of antibiotics by actinomycetes, but have also provided a platform for the biotechnological production of mammalian hormones and other gene-encoded products using genetically engineered microorganisms.

Box 2 | Aflatoxins

Mycotoxins, or mould poisons, are less well known than mushroom poisons but cause a higher incidence of disease. Eating toxic mushrooms is easier to avoid than the inadvertent consumption of mould-contaminated foods. Synthetic contaminants, food additives and pesticide residues get more press attention, but mycotoxins are almost certainly the main non-infectious dietary risk factor in the human food supply101. The most famous mycotoxins are the aflatoxins. These molecules were discovered in the early 1960s when thousands of turkey poults mysteriously died in hatcheries in and around London. All of the dead turkeys had been fed the same Brazilian peanut meal. The meal was heavily contaminated with a common species of mould, so suspicion focused on the fungus, and soon a family of toxic metabolites was isolated. These toxins were named aflatoxins after the producer species, Aspergillus flavus. The four major aflatoxins — aflatoxin B1, B2, G1 and G2 — were identified based on their blue or green fluorescence under ultraviolet light and their relative chromatographic mobility during thin-layer chromatography with silica gel101. Further studies showed that aflatoxin B1 was one of the most toxic and carcinogenic compounds ever discovered102,103. The ease with which A. flavus grew on most major crop plants and the prevalence of aflatoxin contamination of foods and feeds led to major international research efforts that included elucidation of most of the genes in the biosynthetic pathway13,104,105.

There is considerable evidence that the Iraqi government stockpiled aflatoxins as part of their chemical-warfare programme during the 1980s106. Perhaps Saddam Hussein, or one of his henchmen, was influenced by Graham Greene’s 1979 spy novel, The Human Factor107. The plot revolves around the toxicologically improbable murder of a man whose whisky had been laced with aflatoxin. When the victim, a heavy drinker, succumbs to liver cancer, no one suspects foul play. Nonetheless, despite their rightly deserved notoriety, aflatoxins are a poor choice of poison for both novelists and bioterrorists. They do most of their damage in developing countries, where their prevalence in the food supply is an all too common phenomenon108.

938 | DECEMBER 2005 | VOLUME 3 www.nature.com/reviews/micro

R E V I E W S


NH

O

N

S

O

OHO

N

HN

N

N

HN

N

N

N

N

NH

NH

O

O

CH3

O

H3C

CH3O

OOCH3

O

O

CH3

O

H3C O

O

HO

CH3

N

N

OHO

O

CH3

OH

SS

a Peptides

b Alkaloids

c Terpenes

d Polyketides

N

HN

R1

CH3

OHN

N

O

O

R2

N

OH

R3 O N

N

NH

O

O

O

H3C

O

O

O

O

O

O

OH

O

HO

O

CO

OOH

OH

O

OOCH3

O

HNO

HO

HOO

Fusarin C

O

OOHOH

HO

OH

O

OO

O

O

O

O

OCH3

O O

Trichothecene T2 toxinAristolochene

Ergopeptides

CyclosporinGliotoxin

Penicillin G

WAAflatoxin B1

Lovastatin

Gibberellin GA3

Fumitremorgen C

DOMAINIn a polyketide synthase or non-ribosomal peptide synthetase, a stretch of conserved amino acids that defines a specific biochemical function or active site region.

will start by classifying secondary metabolites according to the enzyme classes involved in their bio-synthesis, and ends with a description of the wealth of putative metabolites that have been revealed by sequencing the genomes of three Aspergillus spe-cies: Aspergillus nidulans, Aspergillus fumigatus and Aspergillus oryzae.

Classes of fungal secondary metabolitesThere are several classes of secondary metabolites, discussed below.

Polyketides. Polyketides are the most abundant fungal secondary metabolites. The genetically best-characterized polyketides include the yellow A. nidulans spore-pigment intermediate naphtho-pyrone (WA)9, the carcinogen aflatoxin BOX 2 and the commercially important cholesterol-lowering compound lovastatin10 (FIG. 1). Fungal polyketides are synthesized by type I polyketide synthases (PKSs), which are multi domain proteins that are related to eukaryotic fatty-acid synthases and contain similar DOMAIN structures (FIG. 2). For both enzymes, short-chain carboxylic acids — usually acetyl coenzyme A (acetyl CoA) and malonyl CoA — are condensed to form carbon chains of varying lengths. The main dif-ference between polyketides and fatty acids is the full reduction of the β-carbon in fatty acids, which is an optional event in polyketide synthesis. In the fungal PKSs, the ketoacyl CoA synthase (KS), acyltransferase (AT) and acyl carrier (ACP) domains are essential for polyketide synthesis, whereas the ketoreductase (KR), dehydratase (DH) and enoyl reductase (ER) domains that are required for ketone reduction in fatty acids are not present in all fungal PKS enzymes (FIG. 2). Bacteria also have multidomain, type I PKS enzymes11, for example, the multimodular enzyme DEBS (6-deoxy-erythronolide B synthase) is required for biosynthesis of the antibiotic erythromycin A. It consists of three different subunits, each of which is composed of two MODULES. The six modules function successively in a six-step reaction to incorporate methylmalonyl CoA into a growing polyketide chain, which is then released and cyclized through addition of a thioesterase domain at the end of module six12.

Unlike bacterial PKS type I enzymes, which have separate modules for each methylmalonyl CoA addi-tion, fungal PKSs are limited to one module, with which they can carry out repeated biosynthetic reac-tions, and are therefore called ‘iterative PKSs’. For example, the WA PKS, which has the minimal domain set of KS–AT–ACP, uses repeated cycles of condensa-tion without reduction to generate a heptaketide. The C-terminal region of the enzyme, which has a thio-ester ase domain motif, is responsible for a CLAISENTYPE

CYCLIZATION that results in the formation of the aromatic polyketide naphthopyrone9. How the number of cycles of condensation is controlled to stop at the hexaketide is not understood for this or for any other fungal type I PKS enzyme.

The diversity of fungal polyketide structures results from the number of iteration reactions, the number of reduction reactions, which extender unit is used and, in the case of aromatic polyketides, cyclizations of the nascent polyketide chain. Further variety is achieved by the introduction of many different post-polyketide-synthesis steps. For example, in addition to the PKS and a fatty-acid synthase that is required for a hexanoyl CoA starter unit, in A. nidulans the genes in the cluster

Figure 1 | The main groups of fungal secondary metabolites. Ergopeptides such as ergotamine are tryptophan-derived alkaloids to which peptides are added by a non-ribosomal peptide synthetase (NRPS). T2 toxin is a trichothecene of Fusarium sporotrichioides. WA (a naphthopyrone), is a yellow pigment produced by the polyketide synthase WA, encoded by the wA gene of Aspergillus nidulans. It is converted to the green spore pigment by a laccase. Fusarin C synthesis is dependent on a hybrid PKS (polyketide synthase)–NRPS.


R E V I E W S


KS AT ACP (CYC) (TE)(DH) (MT) (ER) (KR)

β-lactam antibiotics

A A AP C P PC C TE

O S

H2N

SH

L-cysteine

NH2

OHO

O S

L-α-aminoadipate

O S

H2N

L-valine

H2NHN

OOHO N

HO

SH

HO

O

δ-(L-α-aminoadipyl)-L-cysteinyl-D-valine

MODULE In a polyketide synthase or non-ribosomal peptide synthetase, the complete set of domains that is required for one round of chain elongation and modification.

CLAISENTYPE CYCLIZATIONClaisen condensations are a common mechanism in biological systems for synthesis of carbon–carbon bonds. The product is a β-ketoester. A similar reaction is also used to cyclize the heptaketide product of the wA gene to form an aromatic ring.

that is required for sterigmatocystin synthesis encode five monoxygenases, four dehydrogenases, an esterase, an O-methyltransferase, a reductase, an oxidase and a DNA-binding protein13. Similar clusters of genes that code for PKS enzymes and 15 or more genes encoding post-PKS-step enzymes are present in the clusters that are required for aflatoxin synthesis in Aspergillus flavus and Aspergillus parasiticus14,15.

Non-ribosomal peptides. Non-ribosomal peptides are derived from both PROTEINOGENIC AMINO ACIDS and non-proteinogenic amino acids by multidomain, multi modular enzymes named non-ribosomal pep-tide synthetases16 (NRPSs) (FIGS 1,3). Each module in an NRPS contains several domains that allow recognition, activation and then covalent bind-ing of a module-specific amino acid as a thioester to the 4′-phosphopantetheine cofactor, which is attached to each module through a conserved serine. Subsequently, peptide bonds are formed between the

tethered amino acids. The resulting peptide is then released by a thioesterase-like domain that is usually located at the C-terminal end of the final module. The domains that carry out these activities are named A (adenylation), P (pantothenylation/peptidyl carrier), C (condensation/peptide-bond formation) and TE (thioesterase) (FIG. 3). The first fungal NRPS identified, δ-(l-α-aminoadipyl)-l-cysteinyl-d-valine synthetase (ACVS), catalyses the first committed step in the bio-synthesis of the β-lactam antibiotics penicillin and cephalosporin17 BOX 1. ACVS condenses l-α-amino-adipic acid, l-cysteine and l-valine, and epimerizes l-valine to d-valine18 to form a linear peptide (ACV) that is subsequently cyclized to iso penicillin N by the action of isopenicillin N synthase.

Diversity among non-ribosomal peptides arises in the length of peptides produced, whether the peptide is cyclized, and variations in the functions of the domains. Another example of an NRPS that produces a linear peptide is the peptaibol synthetase of Trichoderma virens19. This enzyme produces an 18-residue linear peptide that includes the rare amino acid aminoisobutyric acid. The N terminus is acylated by an enzyme domain that resembles a fatty-acid synthase, and the C-terminal amino acid is hydroxy-lated. In Tolypocladium niveum, an NRPS synthesizes the cyclic undecapeptide cyclosporin20, which is an immuno suppressive drug that is used to treat patients after organ-transplant surgery. This NRPS lacks a TE domain. The peptide is released by cyclization, and some of the modules contain an additional methyl-ation domain between the A and P domains, which catalyses the methylation of amino acids. The enzyme modules and amino-acid sequences are collinear with a repeating pattern (APC)n of domain structure. Other NRPSs in bacteria and fungi have a more diverse domain arrangement. For example, in A. nidulans the NRPS gene that is involved in the biosynthesis of the siderophore ferricrocin, a cyclic hexapeptide with the structure Gly–Ser–Gly–(N5-acetyl–N5-hydroxy-ornithine)3, has the domain structure (APC)3 (PC)2 REF. 21. It is not clear how the modules and domains are used, although some iterative use of modules is hypothesized.

Terpenes. The best-known terpenes are odoriferous plant metabolites such as camphor and turpentine, but fungi also synthesize several important terpenes, including the aristolochenes, carotenoids, gibberellins, indole-diterpenes and trichothecenes. All terpenes are composed of several isoprene units, can be linear or cyclic, saturated or unsaturated, and can be modified in various ways (FIGS 1,4).

Common classes of terpenes include the mono-terpenes, which are generated from geranyl pyro-phosphate (also known as geranyl diphosphate); sesquiterpenes, which are generated from farnesyl pyrophosphate (also known as farnesyl diphosphate); and diterpenes and carotenoids, which are generated from geranylgeranyl pyrophosphate (also known as geranylgeranyl diphosphate). The defining enzyme in

Figure 2 | Fungal polyketide synthase (PKS) domain structure. The minimal structure is KS–AT–ACP. Optional domains are in brackets. Polyketide synthesis is initiated when acetyl and malonyl coenzyme A (CoA) are loaded as thioesters on to the 4′-phosphopantotheine of an acyl carrier (ACP) domain by means of the acyltransferase (AT) domain. Condensation then occurs with another thioester intermediate bound to the ketoacyl CoA synthase (KS) domain, and decarboxylation of the ACP-bound intermediate occurs. The resulting β-ketothioester can then be reduced by the action of the ketoreductase (KR) domain, followed by dehydration by the dehydratase (DH) domain. If an enoyl reductase (ER) domain is present, an unsaturated intermediate is formed. Some PKSs contain a methyltransferase (MT) domain that methylates the α-carbon of the thioester. CYC, cyclase; TE, thioesterase.

Figure 3 | ACV synthetase, a trimodular non-ribosomal peptide synthetase. δ-(L-α-aminoadipyl) L-cysteinyl-D-valine (ACV) synthetase catalyses the first committed step in penicillin and cephalosporin biosynthesis. Each amino acid is recognized and activated by the cognate adenylation domain (A), and attached as a thioester to 4′-phosphopantetheine at the peptidyl carrier domain (P). Peptide bonds are formed with the involvement of the condensation domain (C). The final tripeptide, attached to the peptidyl carrier domain of the C-terminal module, is released by the integrated thioesterase domain (TE), with the L-valine isomerized to D-valine. The tripeptide is subsequently cyclized to isopenicillin N.


R E V I E W S


OP P

P PO

P PO

P PO

P PO

Mevalonate pathway

Secondary metabolites

Indole alkaloids

Monoterpenes

Sesquiterpenes Steroids

Diterpenes Carotenoids

Primary metabolites

Dimethylallyl diphosphate (DMAPP)

OP P

Dimethylallyl diphosphate

Geranyl diphosphate (GPP)

Geranylgeranyl diphosphate (GGPP)

C50

Farnesyl diphosphate (FPP)

P PO

Isopentenyl diphosphate

Coenzyme Q

P PO

P PO

PROTEINOGENIC AMINO ACIDSThose amino acids that are found in proteins and that are coded for in the standard genetic code. Proteinogenic means ‘protein-building’.

PRENYLATION The enzymatic addition of prenyl moieties to secondary metabolic intermediates.

terpene synthesis is terpene cyclase, which is essential for the production of different terpenes from different diphosphates. Although terpene cyclases have struc-tural homology, they have little primary sequence similarity and seem to have diverged relatively rapidly from a common ancestor.

Several fungal terpene cyclases have been char-acterized, including a bifunctional terpene cyclase from Gibberella fujikuroi22, a trichodiene synthase from Fusarium sporotrichioides23 and an aristolo-chene cyclase from Aspergillus terreus and Penicillium roquefortii24. The crystal structures of the trichodiene synthase and aristolochene cyclase have been solved. A conserved DDXXD/E motif (where X is any amino acid) is involved in binding Mg2+ ions, which are required for the enzyme to bind to prenyl phosphates. Trichothecene and aristolochene are the substrates for further modifications, including acetylations, esterifications and oxygenations that generate various sesquiterpenoid mycotoxins.

Carotenoid biosynthesis has been studied in Neurospora crassa25. Only one geranylgeranyl diphos-phate synthase has been found in the genome of N. crassa, in contrast to the two distinct genes encoding geranylgeranyl diphosphate synthase that are found in gibberellin- and indole-diterpene-producing fungi. Of the genes encoding these cyclases, one (named ggs2 in G. fujikuroi and paxG in Penicillium paxilli) is dedicated to secondary metabolic pathways for gib-berellin and paxilline, respectively; this might reflect compartmentalization of primary and secondary terpene metabolism26.

Indole alkaloids. Indole alkaloids are usually derived from tryptophan and dimethylallyl pyrophosphate, although sometimes amino acids other than tryptophan are used as precursors BOX 3. The best-understood pathway is ergotamine synthesis in Claviceps purpurea and related species27. The first committed step is the PRENYLATION of tryptophan by dimethylallyl tryptophan synthetase (DMATS). Following methylation of dimethylallyl tryptophan, a series of oxidation steps proceed through agro clavine to lysergic acid. Lysergic acid is then activated by a single-module NRPS, con-densed with a tripeptide that is produced by a second NRPS, and released as ergotamine.

Other tryptophan-derived alkaloids such as the fumigaclavines and fumitremorgens of A. fumigatus undergo one or more prenylation steps. The details of these pathways are yet to be elucidated, but it is likely that the fumigaclavine biosynthetic pathway proceeds through agroclavine and might therefore have some early steps in common with the ergotamine pathway. Fumitremorgens are derived from proline, tryptophan and dimethylallyl pyrophosphate28. Tryptophan and proline are condensed to form a diketopiperazine called brevianamide F, possibly by means of an unidentified NRPS. It is hypothesized that brevianamides E and A from Penicillium brevicompactum and fumitremorgens A, B and C, and tryprostatins from A. fumigatus are derived from brevianamide F by diverse pathways that require various oxidases, methylases and prenyl transferases.

GeneticsGene clusters. Until recently, research into fungal secondary metabolism was dominated by drug com-panies, which have screening programmes for new and commercially viable bioactive products that are coupled with the structural elucidation of bioactive secondary metabolites. The biochemical pathways that lead to the production of a few specific second-ary metabolites, including aflatoxin, penicillins and ergot alkaloids, have however been analysed using blocked mutants and isotopic tracers29–31. In general, biochemical and genetic studies were rare owing to the formidable technical challenges: the enzymes that synthesize secondary metabolites are only present in minute quantities, and many of the producer fungal species lacked the sexual systems that were required for genetic analysis. The advent of recombinant DNA

Figure 4 | Terpene biosynthetic pathway. Isopentenyl diphosphate and its isomer dimethylallyl diphosphate (DMAPP), products of the mevalonate pathway, are the building blocks (5C isoprene units) for the linear polyprenyl diphosphates, which are precursors of steroids, carotenoids and coenzyme Q in many species. A family of isoprenyl diphosphate synthases is responsible for chain elongation. DMAPP and the isoprenoid intermediates are also the starting points for a wide range of secondary metabolites, including indole alkaloids, monoterpenes, sesquiterpenes and diterpenes. The terpenes are produced by cyclization of the isoprenoids. For example, farnesyl diphosphate is cyclized to produce a large variety of sesquiterpenes, depending on the cyclase structure and function. DMAPP can be added as a side chain to various aromatic compounds. For example, ergotamine is derived from dimethylallyl tryptophan.


R E V I E W S


methodologies in the 1980s enabled dramatic progress in the genetics and biochemistry of fungal secondary metabolism. This rapid progress was facilitated by what is now considered a hallmark characteristic of second-ary metabolic biosynthetic pathways — the grouping of pathway genes in a contiguous cluster.

Metabolic gene clustering was not predicted by the fungal research community. In fact, in the era between the discovery of the bacterial operon and the advent of large-scale eukaryotic gene cloning, it had become dogma that eukaryotic genes that are involved in functionally related pathways are not linked. By 1990 however, this tenet had been abandoned owing to the almost routine discovery of gene clusters in fungi for phenotypes as varied as nutrient use, mating type, pathogenicity and secondary metabolism32,33. In less than a decade, it was shown that the genes for the pro-duction of a broad range of secondary metabolites were located adjacent to one another; in addition, a pathway-specific regulatory gene was often embedded in these gene clusters. Fungal secondary metabolite clusters characterized by gene cloning include aflatoxins14,15, cephalosporin34, compactin35,36, ergot alkaloids27, fumonisin37, gibberellins38,39, HC toxin40, lovastatin10, melanin41,42, paxillin26, penicillin43,44, sterigmatocystin13, sirodesmin45 and trichothecenes46,47.

Transcription factors. The co-regulation of the clusters of genes that code for the synthesis of natural products can, in part, be explained by coordinated transcrip-tional control of biosynthetic genes by ‘narrow’- or

‘broad’-domain transcription factors. The narrow pathway-specific regulators are usually found in the cluster and positively regulate gene expression. These proteins are often Zn(II)2Cys6 zinc binuclear cluster proteins48–50, which are a class of proteins so far only found in fungi. The archetypal protein in this group is AflR (aflatoxin regulator), the Zn(II)2Cys6 protein that is required for aflatoxin and sterigmato cystin bio-synthetic gene activation49–52. Typical for this group of DNA-binding proteins, AflR recognizes and binds to a palindromic sequence found in the promoters of the biosynthetic genes50. Other transcription factors that are encoded in biosynthetic gene clusters include Cys2His2 zinc-finger proteins (Tri6 and MRTRI6 for trichothecene production48) and an ankyrin repeat protein (ToxE for HC-toxin production53). Cluster regulators not found in the cluster itself include a two-peptide forkhead complex (AcFKH1 and CPCR1) for cephalosporin production54 and an HAP-like transcriptional complex (PENR1) for peni-cillin55. Additionally, PENR1 has also been shown to be important in taka-amylase, xylanase and cellobio-hydrolase production56.

Secondary metabolite biosynthesis has long been known to be responsive to environmental cues, including the carbon and nitrogen source, ambient temperature, light and pH57,58. Broad-domain factors are transcription factors that are important in inte-grating cellular responses to these parameters. Several studies59,60 indicate that responses to environmental signals are transmitted through Cys2His2 zinc-finger

Box 3 | Ergot alkaloids

Ergot alkaloids are found in the sclerotia of Claviceps, a genus of plant pathogens that parasitizes rye, wheat and other grasses. When sclerotia are ground and baked into bread, inadvertent human consumption can lead to convulsions, vasoconstriction and possible hallucinations, a disease syndrome variously called St Anthony’s Fire or ergotism. Claviceps sclerotia contain a cocktail of ergot alkaloids from which modern scientists have identified three main groups: the clavine type, the lysergic-acid type and the peptide alkaloids. Members of each group have different physiological properties, some of which have been exploited for human use. For example, ethnomycologists have hypothesized that the Eleusian mysteries, holy rituals carried out in ancient Greece, used an elixir containing water-soluble, hallucinogenic ergot alkaloids. Medieval midwives adopted Claviceps sclerotia to hasten labour and to induce abortion. In the twentieth and twenty-first centuries, ergot alkaloids have been used in obstetrics as well as in the treatment of migraine headaches109.

The famous hallucinogen lysergic acid diethyl amide (LSD) was discovered by Hofmann at the Sandoz Laboratories in 1938, in a research project that involved making novel derivatives of methergine R, a compound then widely used to assuage haemorrhage during childbirth. Hofmann accidentally swallowed a minute quantity of his semi-synthetic derivative and discovered its hallucinogenic properties. For a while, LSD was marketed to psychiatrists, used unsuccessfully to treat schizophrenia, and adopted by the American CIA as a truth serum for investigating suspected communists110. Most famously, during the 1960s, LSD was the favourite recreational drug of the counterculture movement. Popular songs such as the Beatles’ ‘Lucy in the Sky with Diamonds’ and Jefferson Airplane’s ‘White Rabbit’ glamourized its psychedelic effects. By the mid-1970s, however, government efforts to criminalize its use, coupled with increased public awareness about the possibility of ‘bad trips’, succeeded in making LSD less socially acceptable.

One provocative hypothesis developed by historians of the era was that ergotism might have contributed to the notorious seventeenth-century witchcraft trials at Salem, Massachusetts111,112. More recently, Robin Cook, a popular novelist, used the Salem–ergot hypothesis as the plot line for Acceptable Risk113. In this thriller, biotechnologists isolate Claviceps spores from a damp New England cellar, culture the fungus, and extract a new alkaloid that enhances intellect, stamina, libido and ego. Too late, they discover that their ‘billion-dollar drug’ has side effects, turning those who consume it into flesh-eating zombies. In a crude way, this story highlights the paradox of most powerful drugs: often the distinction between a toxin and a remedy is a fraction of a decimal point, the addition of a methyl group or the removal of a double bond.


R E V I E W S


Sterigmatocystin PenicillinConidiation

BrlA AflR LaeA

FlbA FadA PkaA Colonygrowth

TranscriptionalPost-transcriptional

global transcription factors that mediate carbon61

(CreA), nitrogen62 and pH63,64 (PacC) signalling. Broad transcription factors can positively (PacC regulation of penicillin) or negatively (CreA regulation of penicillin) regulate metabolite production and are conserved in all fungi and other eukaryotes. Regulation by both narrow- and broad-domain transcription factors ensures that secondary metabolite pathways can respond to the demands of general cellular metabolism and the presence of specific pathway inducers.

Secondary metabolism and fungal development. An association between natural product formation and morphological development has been observed for decades57,65. For example, one unusual class of mutants in A. parasiticus named ‘sec’ are deficient in both sporulation and aflatoxin production66. Although the genes that encode the enzymes for aflatoxin synthesis are present67, expression of the pathway-specific regu-lator, aflR, is 5–10-fold reduced in the toxigenic strains compared with parental strains68. A similar example involves a mutation in the A. parasiticus fluP gene that results in a fluffy hyphal morphology pheno type, a reduction in the number of asexual spores produced and a reduction in the level of aflatoxin production69.

The correlation of secondary metabolism with fungal development has recently been reviewed70, and the authors made special note of the joint regulation of these processes through a G-protein–protein kinase A signal-transduction pathway. FadA, a G protein in A. nidulans, and PkaA, a protein kinase that is down-stream in the FadA regulatory pathway, are particularly well characterized. Both proteins negatively regulate aflatoxin and sterigmatocystin synthesis71,72 (FIG. 5). PkaA regulates production of sterigmatocystin tran-scriptionally and post-transcriptionally through aflR and its gene product72, and it is postulated that a simi-lar regulatory cascade controls aflatoxin regulation in

A. parasiticus73. AflR is inactivated by PkaA-mediated phosphorylation, and the transcription of aflR is repressed by PkaA activity74; the repression of aflR transcription is mediated by PkaA through a novel methyltransferase, LaeA75 (discussed later).

Increasing research into the roles of G-protein signalling in fungal development indicates that signal-transduction pathways often positively or negatively regulate secondary metabolism. For example, FadA negatively regulates aflatoxin and sterigmatocystin synthesis, but positively regulates penicillin produc-tion in A. nidulans, and a FadA homologue also positively regulates trichothecene production in F. sporotrichioides76. G proteins also regulate secondary metabolism in Botrytis cinerea and Trichoderma atro-viride, among others77,78. In addition to its role in the regulation of secondary metabolism and sporulation, G-protein signalling is crucial for pathogen icity. For instance, deletion of cpg1, which encodes the α-subunit of a heterotrimeric G protein in the chestnut blight fungus Cryphonectria parasitica, results in a marked reduction in the fungal growth rate, reduced levels of spore production, decrease in pigmentation and a loss of virulence79.

Epigenetic controls? The role of narrow pathway-specific regulators and broad global transcription factors such as PacC in penicillin biosynthesis63 and in regulation of secondary metabolite gene clusters — coupled with the G-protein/cAMP/protein-kinase-mediated regulation of sterigmatocystin and aflatoxin biosynthesis — provided the research community with insights into the mechanisms of secondary-metabolite regulation in fungi. None of these findings, however, could explain why metabolite-specific synthesis and regulatory genes were clustered.

The evolution and maintenance of gene clusters has received a great deal of attention in bacteria80–82 and fungi32,33,83,84. Evidence for horizontal transfer of the penicillin cluster from bacteria to fungi was met with great excitement, as it suggested a reason for both the origin and maintenance of metabolic clustering83,84,85. However, other functionally conserved metabolites, such as gibberellin production in the fungus G. fujikuroi and in plants, are unlikely to have arisen from hori-zontal transfer between kingdoms22, and it is likely that several evolutionary mechanisms account for the presence of secondary-metabolite clusters. Although a unifying model to explain the clustering of secondary-metabolite genes in fungi is lacking, the preservation of clusters, now known to be extensive through scru-tiny of fungal genomes, might indicate an underlying advantage to maintaining clustering.

An exciting advance in support of a global require-ment for fungal secondary-metabolite gene clustering has arisen from identification of an Aspergillus methyl-transferase, LaeA75. LaeA was originally identified by complementation of a sterigmatocystin biosynthetic mutant. Deletion and overexpression of laeA in A. nidulans, A. fumigatus and A. terreus strains either silences or increases, respectively, the production

Figure 5 | Integrating signal-transduction controls in spore production (conidiation) and secondary metabolism in Aspergillus nidulans. The model shown uses solid lines to indicate known pathways and dashed lines to indicate hypothesized pathways. FadA, α subunit heterotrimeric G protein; FlbA, a regulatory protein that has an RGS (regulator of G protein signalling) motif; PkaA, catalytic subunit of protein kinase A; BrlA, a conidiation-specific transcription factor; AflR (aflatoxin regulator), a sterigmatocystin/aflatoxin-specific transcription factor.


R E V I E W S


Heterochromatin

Euchromatin

Heterochromatin nucleosome

Euchromatin

EuchromatinEuchromatin

Euchromatin

LaeA

EUCHROMATICDescribes chromosome regions with actively transcribed genes. Generally these regions stain poorly or not at all.

HETEROCHROMATICDescribes chromosomal regions that are generally genetically inert. The chromatin is tightly coiled throughout the cell cycle and stains well.

of several secondary metabolites and the expression of their respective genes 75. Sequence analysis of laeA and analysis of the encoded protein indicated that it is a protein methyltransferase75. The highest sequence similarity is to histone and arginine methyltransferases, which have important roles in the regulation of gene expression, such as defining the boundaries of EUCHRO

MATIC and HETEROCHROMATIC chromosomal domains in the mating locus of yeast and the β-globin locus in mice86–88. Although LaeA functions are not yet fully characterized, we speculate that this protein is involved in chromatin modification, perhaps reminiscent of mating-type locus expression in yeast. Methylation is involved in both gene repression and activation of heterochromatic and euchromatic regions of eukaryo-tic chromosomes89, so one model of LaeA-mediated regulation of clusters invokes a role in heterochromatin repression (FIG. 6).

Bioinformatics and gene predictionsAutomated and manual annotations of the N. crassa, A. fumigatus, A. nidulans and A. oryzae genomes predict gene functions based on homology to char-acterized genes and their products. It is easy to find genes encoding putative NRPSs and PKSs based on protein domain structures. Terpene and alkaloid

biosynthetic genes are detected by sequence similar-ity to fungal terpene cyclases and with the DMATSs of Claviceps species. As terpene cyclase primary sequences have diverged, it is more difficult to detect the genes that encode these proteins.

Oxidoreductases, methylases, acetylases, esterases and transporters are not exclusive to secondary metab-olism, so homology searches are more problematic for these genes. Nevertheless, if good candidate genes encoding these enzymes are found adjacent to signa-ture secondary metabolic genes such as NRPS, PKS and DMATS homologues, there is an improved chance that they are involved in production of a secondary metabolite.

Using these criteria, several important genes have been identified from fungal genomic-DNA-sequence data; a summary is given in TABLE 1. Most of these are present in putative gene clusters. In some cases, we know the function of the cluster from pre-genomic studies, but the function of most of these putative clusters will remain speculative until functional stud-ies have been carried out. More putative secondary metabolite clusters are present than are needed to account for the known products, and some of these clusters might not be expressed under laboratory con-ditions. For example, the aflatoxin cluster of A. oryzae is not expressed under conditions that are favourable to aflatoxin expression in A. flavus and A. parasiticus (reviewed in REF. 33).

Comparative studies allow a confident predic-tion of the likely gliotoxin cluster of A. fumigatus, based on its resemblance to the sirodesmin cluster of Leptosphaeria maculans. Both compounds are piperazines. Sirodesmin is derived from tyrosine and serine, and gliotoxin is derived from phenyl alanine and serine45. Each cluster includes a bimodular NRPS, which is presumably required for condensa-tion of the precursor amino acids. The A. fumigatus genome also has seven significant hits in the public databases (ranging from 22%, expect 5e-15, to 52%, expect e-115 in a BLAST search) with the DMATS of C. purpurea, and the best hit (TIGR assigned number Afu2g18040 accessible at the Central Aspergillus Data Repository (see Online links box)) is present in a clus-ter that contains other genes resembling those found in C. purpurea. No NRPS is present in the putative A. fumigatus cluster, as expected from the absence of reports of ergopeptine synthesis from this species,

Figure 6 | Model of LaeA function. Secondary metabolite gene clusters such as the sterigmatocystin gene cluster are maintained in silenced heterochromatin but are surrounded by expressed euchromatin. The LaeA protein functions to initiate a process that converts heterochromatin to euchromatin, perhaps by interfering with methylases or deacetylases that are associated with heterochromatin. Sterigmatocystin genes are designated with thick arrows, thin arrows indicate genes that are located on either side of the sterigmatocystin cluster and that are not regulated by LaeA.

Table 1 | Summary of secondary-metabolite gene classes in Aspergillus

Protein Aspergillus oryzae Aspergillus fumigatus Aspergillus nidulans

PKS 30 14 27

NRPS 18 14 14

FAS 5 1 6

Sesquiterpene cyclase 1 Not detected 1

DMATS 2 7 2

PKS, polyketide synthase; NRPS, non-ribosomal peptide synthetase; FAS, fatty-acid synthase; DMATS, dimethylallyl tryptophan synthetase.


R E V I E W S


PARALOGUESGenes that are derived from a common ancestor by duplication. They can have related functions.

ORTHOLOGUESGenes that are derived from a common ancestor by a speciation event. They usually have equivalent function in their respective species.

which produces its own group of clavines90,91. Further homologues of DMATS genes are found in this and another cluster, and these might be prenyl transferases that are required for prenylation steps required in fumigaclavine and fumitremorgen biosynthesis.

There are still many PKS and NRPS genes (and associated clusters) for which the putative products cannot be predicted. In bacteria, prediction of NRPS products has been partially successful through the use of a derivation of a module-recognition code that is based on correlation of NRPS adenylation domain structure to a particular amino acid92. However, this algorithm is not sufficiently robust to predict the amino-acid specificity of fungal NRPS modules, partly because of the limited experimental information available. Similarly, it is not yet possible to predict the structure of any polyketide of an iterative, type I fungal PKS through analysis of domain and motif structures alone. For example, automated annotation might label some fungal PKSs as ‘lovastatin biosynthesis’ genes. This classification simply reflects similarity to one of the lovastatin PKS genes, and probably indicates a related PARALOGUE rather than a true ORTHOLOGUE.

Several sequenced fungal genomes contain a hybrid PKS–NRPS gene. This arrangement has been observed previously in bacteria, and illustrates how genome data add to our knowledge of the diversity of PKS and NRPS structures found in nature. A hybrid gene in Fusarium verticilliodes (formerly Fusarium moniliforme) produces a precursor of the toxin fusarin C 93, and a PKS/NRPS is a virulence factor in the rice blast fungus Magnaporthe grisea94. Genes of this hybrid class can be detected in several fungal genomes; however, sequence compari-sons indicate that most are paralogues to the fusarin C gene, and in most cases the products are unknown.

Bioinformatic analysis of the Aspergillus genomes has revealed few terpene biosynthetic genes. A. nidulans

has a homologue of the aristolochene cyclase gene found in P. roquefortii. A. terreus and A. oryzae have a homo-logue of trichothecene cyclase, but no obvious terpene cyclase is detectable in A. fumigatus.

ConclusionsSecondary metabolites are low-molecular-weight natural products with restricted taxonomic distribution, often synthesized after active growth has ceased, which do not have an obvious function in producer species. The β-lactam antibiotic penicillin, which is synthesized by an NRPS, and the polyketides aflatoxin and sterigmato-cystin, which are synthesized through a polyketide pathway, are among the best studied fungal secondary metabolites, and their pathways have become paradigms. The discovery that genes involved in their production are clustered, as are the genes that code for the produc-tion of the vast majority of other secondary metabolites that have been studied, has important implications for gene regulation and evolution. Pragmatically, it means that putative biosynthetic pathway genes for secondary metabolism can easily be detected by in silico analysis of genomic data. Although bioinformatics alone can-not predict pathway products, or even determine if the genes are expressed, these analyses can reveal molecular evidence of many hitherto undiscovered pathways.

The recent characterization of LaeA, which is a global regulator of secondary metabolism, provides a tool for detecting gene clusters, and might reveal novel chromatin-based mechanisms of transcriptional control of these clusters. The laeA gene has been found in all aspergilli examined to date; however, it remains to be seen if LaeA functions in other fungi.

Note added in proofThree recent papers describe the genomes of A. fumi-gatus114, A. nidulans115 and A. oryzae116.

1. Turner, W. B. Fungal Metabolites (Academic Press, London, 1971).

2. Turner, W. B. & Aldridge, D. C. Fungal Metabolites II (Academic Press, London, 1983).

3. Cole, R. & Schweikert, M. Handbook of Secondary Fungal Metabolites Volumes 1–3 (Elsevier, Amsterdam, 2003).

4. Davies, J. Recombinant DNA and the Production of Small Molecules (ASM Press, Washington DC, 1985).

5. Bennett, J. W. & Bentley, R. What’s in a name? Microbial secondary metabolism. Adv. Appl. Microbiol. 34, 1–28 (1989).

6. Ciba Foundation Symposium 171. Secondary Metabolites: Their Function and Evolution (John Wiley & Sons, Chicester, 1992).

7. Raistrick, H. A region of biosynthesis. Proc. R. Soc. Lond. B Biol. Sci. 136, 481–508 (1950).

8. Pelaez, F. Biological activities of fungal metabolites. in Handbook of Industrial Mycology (ed. An, Z.) 49–92 (Marcel Dekker, New York, 2005).

9. Fujii, I., Watanabe, A., Sankawa, U. & Ebizuka, Y. Identification of Claisen cyclase domain in fungal polyketide synthase WA, a naphthopyrone synthase of Aspergillus nidulans. Chem. Biol. 8, 189–197 (2001).

10. Kennedy, J. et al. Modulation of polyketide synthase activity by accessory proteins during lovastatin biosynthesis. Science 284, 1368–1372 (1999). The first biochemical dissection of fungal polyketide synthase and use of the model system A. nidulans to help decipher lovastatin assembly in A. terreus.

11. Bentley, R. & Bennett, J. W. Constructing polyketides: from Collie to combinatorial biosynthesis. Annu. Rev. Microbiol. 53, 411–446 (1999).

12. Donadio S., Staver, M. J., McAlpine, J. B., Swanson, S. J. & Katz, L. Modular organization of gene required for complex polyketide biosynthesis. Science 252, 675–679 (1991).

13. Brown, D. et al. Twenty-five co-regulated transcripts define a sterigmatocystin gene cluster in Aspergillus nidulans. Proc. Natl Acad. Sci. USA 93, 1418–1422 (1996).

14. Yu, J. et al. Clustered pathway genes in aflatoxin bio-synthesis. Appl. Environ. Microbiol. 70, 1253–1262 (2004).

15. Yu, J., D. Bhatnagar, D. & Cleveland, T. D. Completed sequence of aflatoxin pathway gene cluster in Aspergillus parasiticus. FEBS Lett. 564, 126–130 (2004).

16. Finking, R. & Marahiel, M. Biosynthesis of nonribosomal peptides. Annu. Rev. Microbiol. 58, 453–488 (2004).

17. Smith, D. J., Earl, A. J., Turner, G. The multifunctional peptide synthetase performing the first step of penicillin biosynthesis in Penicillium chrysogenum is a 421073 dalton protein similar to Bacillus brevis peptide antibiotic synthetases. EMBO J. 9, 2743–2750 (1990).First indication of the multimodular structure of peptide synthetases: three modules were detected in this tripeptide synthetase.

18. Kallow, W., Kennedy, J., Arezi, B., Turner, G. & von Doehren, H. Thioesterase domain of δ-(L-α-aminoadipyl)-L-cysteinyl-D-valine synthetase: alteration of stereospecificity by site-directed mutagenesis. J. Mol. Biol. 297, 395–408 (2000).

19. Wiest, A. et al. Identification of peptaibols from Trichoderma virens and cloning of a peptaibol synthetase. J. Biol. Chem. 277, 20862–20868 (2002).

20. Weber, G., Schorgendorfer, K., Schneider-Scherzer, E. & Leitner, E. The peptide synthetase catalyzing cyclosporine production in Tolypocladium niveum is

encoded by a giant 45.8-kilobase open reading frame. Curr. Genet. 26, 120–125 (1994).

21. Eisendle, M., Oberegger, H., Zadra, I. & Haas, H. The siderophore system is essential for viability of Aspergillus nidulans: functional analysis of two genes encoding L-ornithine N5-monooxygenase (sidA) and a nonribosomal peptide synthetase (sidC). Mol. Microbiol. 49, 359–375 (2003).

22. Tudzynski, B., Hedden, P. Carrera, E. & Gaskin, P. The 450–4 gene of Gibberella fujikuroi encodes ent-kaurine oxidase in the gibberellin biosynthesis pathway. Appl. Environ. Microbiol. 67, 3514–3522 (2001).

23. Rynkiewicz, M. J., Cane, D. E. & Christianson, D. W. Structure of trichodiene synthase from Fusarium sporotrichioides provides mechanistic inferences on the terpene cyclization cascade. Proc. Natl Acad. Sci. USA 98, 13543–13548 (2001).

24. Carruthers, J., Kang, I., Rynkiewicz, M., Cane, D. & Christianson, D. Crystal structure determination of aristolochene synthase from the cheese mold, Penicillium roquefortii. J. Biol. Chem. 275, 25533–25539 (2000).Structural determination of a fungal terpenecyclase, showing that whereas the primary sequences of terpene cyclases are not well conserved between plants and fungi, the tertiary structure is conserved. Also, terpene cyclases might all be derived from a common ancestor.

25. Schmidhauser, T., Lauter, F., Russo, V. & Yanofsky, C. Cloning, sequence, and photoregulation of al-1, a carotenoid biosynthetic gene of Neurospora crassa. Mol. Cell. Biol. 10, 5064–5070 (1990).


R E V I E W S


26. Young, C., McMillan, L., Telfer, E. & Scott, B. Molecular cloning and genetic analysis of an indole-diterpene gene cluster from Penicillium paxilli. Mol. Microbiol. 39, 754–764 (2001)

27. Tudzynski, P. et al. Evidence for an ergot alkaloid gene cluster in Claviceps purpurea. Mol. Gen. Genet. 261, 133–141 (1999).Identification of the ergot alkaloid gene cluster, which includes an NRPS required for ergotamine biosynthesis.

28. von Nussbaum, F. Stephacidin B — a new stage of complexity within prenylated indole alkaloids from fungi. Angew. Chem. Int. Ed. Engl. 42, 3068–3071 (2003).

29. Bennett, J. W., Chang, P.-K. & Bhatnagar, D. One gene to whole pathway: the role of norsolorinic acid in aflatoxin research. Adv. Appl. Microbiol. 45, 1–15 (1997).

30. Luengo, J. M. & Penalva, M. A. Penicillin Biosynthesis in Aspergillus: 50 Years On (eds Martinelli, S. D & Kinghorn, J. R.) 603–638 (Elsevier, Amsterdam,1994).

31. Rehacek, Z. & Sajdl, P. Ergot Alkaloids: Chemistry, Biological Effects, Biotechnology (Academia, Prague, 1990)

32. Keller, N. & Hohn, T. Metabolic pathway gene clusters in filamentous fungi. Fungal Genet. Biol. 21, 17–29 (1997).

33. Zhang, Y.-Q., Wilkinson, H., Keller, N. P. & Tsitsigiannis, D. Secondary metabolite gene clusters. in Handbook of Industrial Microbiology (ed. An, Z.) 355–386 (Marcel Dekker, New York, 2005).

34. Gutierrez, S., Velasco, J., Fernandez, F. J. & Martin, J. F. The cefG gene of Cephalosporium acremonium is linked to the cefEF gene and encodes a deacetylcephalosporin C acetyltransferase closely related to homoserine O-acetyltransferase. J. Bacteriol. 174, 3056–3064 (1992).

35. Abe, Y. et al. Effect of increased dosage of the ML-236B (compactin) biosynthetic gene cluster on ML-236B production in Penicillium citrinum. Mol. Gen. Genet. 268, 130–137 (2002).

36. Abe, Y., Ono, C., Hosobuchi, M. & Yoshikawa, H. Functional analysis of mlcR, a regulatory gene for ML-236B (compactin) biosynthesis in Penicillium citrinum. Mol. Genet. Genomics 268, 352–361 (2002).

37. Proctor, R. H., Brown, D. W., Plattner, R. D. & Desjardins, A. E. Co-expression of 15 contiguous genes delineates a fumonisin biosynthetic gene cluster in Gibberella moniliformis. Fungal Genet. Biol. 38, 237–249 (2003).

38. Hedden, P., Phillips, A., Rojas, M., Carrera, C. & Tudzynski, B. Gibberellin biosynthesis in plants and fungi: a case of convergent evolution? J. Plant Growth Regul. 20, 319–331 (2002).

39. Tudzynski, B. Biosynthesis of gibberellins in Gibberella fujikuroi: biomolecular aspects. Appl. Environ. Microbiol. 52, 298–310 (1999).

40. Ahn, J., Cheng, Y. & Walton, J. An extended physical map of the TOX2 locus of Cochliobolus carbonum required for biosynthesis of HC-toxin. Fungal Genet. Biol. 35, 31–38 (2002).

41. Kimura, N. & Tsuge, T. Gene cluster involved in melanin biosynthesis of the filamentous fungus Alternaria alternata. J. Bacteriol. 175, 4427–4435 (1993).

42. Tsai, H., Wheeler, M., Chang, Y. & Kwon-Chung, K. A developmentally regulated gene cluster involved in conidial pigment biosynthesis in Aspergillus fumigatus. J. Bacteriol. 181, 6469–6477 (1999).

43. Smith, D. J. et al. β-lactam antibiotic biosynthetic genes have been conserved in clusters in prokaryotes and eukaryotes. EMBO J. 9, 741–747 (1990). Showed that secondary metabolic genes were clustered in filamentous fungi, and revealed the close relationship between the β-lactam biosynthetic genes of bacteria and fungi.

44. Brakhaage, A. A. Molecular regulation of β-lactam biosynthesis in filamentous fungi. Microbiol. Mol. Biol. Rev. 62, 547–585 (1998).

45. Gardiner, D., Cozijnsen, A., Wilson, L., Pedras, M. & Howlett, B. The sirodesmin biosynthetic gene cluster of the plant pathogenic fungus Leptosphaeria maculans. Mol. Microbiol. 53, 1307–1318 (2004).

46. Trapp, S., Hohn T., McCormick, S. & Jarvis, B. Characterization of the gene cluster for biosynthesis of macrocyclic trichothecenes in Myrothecium roridum. Mol. Gen. Genet. 257, 421–432 (1998).

47. Brown, D., McCormick, S., Alexander, N., Proctor, R. & Desjardins, A. A genetic and biochemical approach to study trichothecene diversity in Fusarium sporotrichioides and Fusarium graminearum. Fungal Genet. Biol. 32, 121–133 (2001).

48. Proctor, R, Hohn, T., McCormick, S. & Desjardins, A. Tri6 encodes an unusual zinc finger protein involved in regulation of trichothecene biosynthesis in Fusarium sporotrichioides. Appl. Environ. Microbiol. 61, 1923–1930 (1995).

49. Woloshuk, C. et al. Molecular characterization of aflR, a regulatory locus for aflatoxin biosynthesis. Appl. Environ. Microbiol. 60, 2408–2414 (1994).Identification of the first Zn(II)2Cys6 that regulates a secondary metabolite gene cluster.

50. Fernandes, M., Keller, N. & Adams, T. Sequence-specific binding by Aspergillus nidulans AflR, a C6 zinc cluster protein regulating mycotoxin biosynthesis. Mol. Microbiol. 28, 1355–1365 (1998).

51. Chang, P., Ehrlich, K., Yu, J., Bhatnagar, D. & Cleveland, T. Increased expression of Aspergillus parasiticus aflR, encoding a sequence-specific DNA-binding protein, relieves nitrate inhibition of aflatoxin biosynthesis. Appl. Environ. Microbiol. 61, 2372–2377 (1995).

52. Yu, J. et al. Conservation of structure and function of the aflatoxin regulatory gene aflR from Aspergillus nidulans and A. flavus. Curr. Genet. 29, 549–555 (1996).

53. Pedley, K. & Walton, J. Regulation of cyclic peptide biosynthesis in a plant pathogenic fungus by a novel transcription factor. Proc. Natl Acad. Sci. USA 98, 14174–14179 (2001).

54. Schmitt, E. K., Hoff, B. & Kuck, U. AcFKH1, a novel member of the forkhead family, associates with the RFX transcription factor CPCR1 in the cephalosporin C-producing fungus Acremonium chrysogenum. Gene 342, 269–281 (2004).In contrast to penicillin regulation in A. nidulans, cephalosporin regulation is by a forkhead transcription factor in Acremonium.

55. Litzka, O., Papagiannopolus, P., Davis, M., Hynes, M. & Brakhage, A. The penicillin regulator PENR1 of Aspergillus nidulans is a HAP-like transcriptional complex. Eur. J. Biochem. 251, 758–767 (1998).A HAP-like CCAAT-binding complex regulates penicillin biosynthesis.

56. Brakhage, A. et al. HAP-like CCAAT-binding complexes in filamentous fungi: implications for biotechnology. Fungal Genet. Biol. 27, 243–252 (1999).

57. Bennett, J. & Ciegler, A. (eds) Secondary Metabolism and Differentiation in Fungi. (Marcel Dekker, New York, 1983).

58. Berry, D. R. (ed.) Physiology of Industrial Fungi (Blackwell Scientific Publishing, Oxford, 1988).

59. Ehrlich, K., Montalbano, B. & Cotty, P. Sequence comparison of aflR from different Aspergillus species provides evidence for variability in regulation of aflatoxin production. Fungal Genet. Biol. 38, 63–74 (2003).

60. Tudzynski, B., Homann, V., Feng, B. & Marzluf, G. Isolation, characterization and disruption of the areA nitrogen regulatory gene of Gibberella fujikuroi. Mol. Gen. Genet. 261, 106–114 (1999).

61. Dowzer, C. & Kelly, J. Cloning of the creA gene from Aspergillus nidulans: a gene involved in carbon catabolite repression. Curr. Genet. 15, 457–459 (1989).

62. Kudla, B. et al. The regulatory gene areA mediation nitrogen metabolite repression in Aspergillus nidulans. Mutations affecting specificity of gene activation alter a loop residue of a putative zinc finger. EMBO J. 9, 1355–1364 (1990).

63. Tilburn, J. et al. The Aspergillus PacC zinc finger transcription factor mediates regulation of both acid- and alkaline-expressed genes by ambient pH. EMBO J. 14, 779–790 (1995).Important contribution that showed that pH regulates the penicillin gene cluster through a global transcription factor, PacC.

64. Martin, J. Molecular control of expression of penicillin biosynthesis genes in fungi: regulatory proteins interact with a bidirectional promoter region. J. Bacteriol. 182, 2355–2362 (2000).

65. Luckner, M. Secondary Metabolism in Microorganisms, Plants and Animals (Springer–Verlag, Berlin, 1990).

66. Kale, S., Bhatnagar D. & Bennett, J. Isolation and characterization of morphological variants of Aspergillus parasiticus deficient in secondary metabolite production. Mycol. Res. 98, 645–652 (1994).

67. Kale, S., Cary, J., Bhatnagar, D. & Bennett, J. Characterization of an experimentally induced, nonaflatoxigenic variant strains of Aspergillus parasiticus. Appl. Environ. Microbiol. 62, 3999–3404 (1996).

68. Kale, S. et al. Genetic analysis of morphological variants of Aspergillus parasiticus deficient in secondary metabolite production. Mycol. Res. 107, 831–840 (2003).

69. Zhou, R., Rasooly, R. & Linz, J. Isolation and analysis of fluP, a gene associated with hyphal grown and sporulation in Aspergillus parasiticus. Mol. Gen. Genet. 264, 514–520 (2000).

70. Calvo, A., Wilson, R., Bok, J. & Keller, N. Relationship between secondary metabolism and fungal development. Mol. Microbiol. Rev. 66, 447–459 (2002).

71. Hicks, J., Yu, J., Keller, N. & Adams, T. Aspergillus sporulation and mycotoxin production both require inactivation of the FadA G α protein-dependent signaling pathway. EMBO J. 16, 4916–4923 (1997).This paper reported the genetic connection of sporulation and secondary metabolism through a G-protein signalling pathway.

72. Shimizu, K. & Keller, N. Genetic involvement of a cAMP-dependent protein kinase in a G protein signaling pathway regulating morphological and chemical transitions in Aspergillus nidulans. Genetics 157, 591–600 (2001).

73. Roze, L., Beaudry, R., Keller, N. & Linz, J. Regulation of aflatoxin synthesis by FadA/cAMP/protein kinase A signaling in Aspergillus parasiticus. Mycopathologia 158, 219–232 (2004).

74. Shimizu, K., Hicks, J., Huang T.-P. & Keller, N. P. Pka, Ras and RGS protein interactions regulate activity of AflR, a Zn(II)2Cys6 transcription factor in Aspergillus nidulans. Genetics 165, 1095–1104 (2003).

75. Bok, J. & Keller, N. LaeA, a regulator of secondary metabolism in Aspergillus. Euk. Cell 3, 527–535 (2004).Discovery of novel global regulator of several Aspergillus secondary metabolites.

76. Tag, A. et al. G-protein signalling mediates differential production of toxic secondary metabolites. Mol. Microbiol. 38, 658–665 (2000).

77. Schulze Gronover, C., Schorn, C. & Tydzynski, B. Identification of Botrytis cinerea genes up-regulated during infection and controlled by the G α subunit BCG1 using suppression subtractive hybridization (SSH). Mol. Plant Microbe Interact. 17, 537–546 (2004).

78. Reithner, B. et al. The G protein α subunit Tga1 of Trichoderma atroviride is involved in chitinase formation and differential production of antifungal metabolites. Fungal Genet. Biol. 42, 749–760 (2005).

79. Gao, S. & Nuss, D. Distinct roles for two G protein α subunits in fungal virulence, morphology, and reproduction revealed by targeted gene disruption. Proc. Natl Acad. Sci. USA 93, 14122–14127 (1996).

80. Lawrence, J. G. & Roth, J. R. Selfish operons: horizontal transfer may drive the evolution of gene clusters. Genetics 143, 1843–1860 (1996).

81. Lawrence, J. G. Selfish operons and speciation by gene transfer. Trends Microbiol. 5, 355–359 (1997).

82. Lawrence, J. G. Gene transfer, speciation, and the evolution of bacterial genomes. Curr. Opin. Microbiol. 2, 519–523 (1999).

83. Rosewich, U. & Kistler, H. Role of horizontal gene transfer in the evolution of fungi. Annu. Rev. Phytopathol. 38, 325–363 (2000).

84. Walton, J. J. Horizontal gene transfer and the evolution of secondary metabolite gene clusters in fungi: an hypothesis. Fungal Genet. Biol. 30, 167–171 (2000).

85. Smith, M. W., Feng, D.-F. & Doolittle, R. F. Evolution by acquisition: the case for horizontal gene transfers. Trends Biochem. Sci. 17, 489–493 (1992).

86. Litt, M., Simpson, M., Gaszner, M., Allis, D. & Felsenfeld, G. Correlation between histone lysine methylation and developmental changes at the chicken β-globin locus. Science 293, 2453–2455 (2001).

87. Recillas-Targa, F. et al. Position-effect protection and enhancer blocking by the chicken β-globin insulator are separable activities. Proc. Natl Acad. Sci. USA 99, 6883–6888 (2002).

88. Noma, K., Allis, C. & Grewal, S. Transitions in distinct histone H3 methylation patterns at the heterochromatin domain boundaries. Science 293, 1150–1155 (2001).

89. Lee, D. Y., Teyssier, C., Strahl, B. D. & Stallcup, M. R. Role of protein methylation I regulation of transcription. Endocr. Rev. 26, 147–170 (2005).

90. Spilsbury, J. F. & Wilkinson, S. The isolation of festuclavine and two new clavine alkaloids from Aspergillus fumigatus Fres. J. Chem. Soc. 5, 2085–2091 (1961).

91. Cole, R. J. et al. Mycotoxins produced by Aspergillus fumigatus species isolated from molded silage. J. Agric. Food Chem. 25, 826–830 (1977).

92. Challis, G. L, Ravel, J. & Townsend, C. A. Predictive, structure-based model of amino acid recognition by nonribosomal peptide synthetase adenylation domains. Chem. Biol. 7, 211–224 (2000).

93. Song, Z., Cox, R. J., Lazarus, C. M. & Simpson, T. J. Fusarin C biosynthesis in Fusarium moniliforme and Fusarium venenatum. ChembioChem. 5, 1196–1203 (2004).

94. Bohnert, H. U. et al. A putative polyketide synthase/peptide synthetase from Magnaporthe grisea signals pathogen attack to resistant rice. Plant Cell 16, 2499–2513 (2004).


R E V I E W S


95. Hobby, G. Penicillin: Meeting the Challenge (Yale University Press, New Haven, 1985).

96. Wainwright, M. Miracle Cure: the Story of Penicillin and the Golden Age of Antibiotics (Blackwell Publishing, Oxford, 1990).

97. Bennett, J. & Chung, K. Alexander Fleming and the discovery of penicillin. Adv. Appl. Microbiol. 49, 163–184 (2001).

98. Lax, A. The Mold in Dr Florey’s Coat. The Story of the Penicillin Miracle (Henry Holt & Company, New York, 2004)

99. Scoutaris, M. “Moldy Mary” and the Illinois Fruit and Vegetable Company. Pharm. Hist. 38, 175–177 (1996).

100. Bentley, R. The molecular structure of penicillin. J. Chem. Ed. 81, 1462–1470 (2004).

101. Kuiper-Goodman, T. Food safety: mycotoxins and phycotoxins in perspective. In Mycotoxins and Phycotoxins — Developments in Chemistry, Toxicology and Food Safety. (eds Miraglia, M., van Edmond, H., Brera, C. & Gilbert, J.) 25–48 (Alaken Inc., Fort Collins, 1998)

102. Squire, R. A. Ranking animal carcinogens. A proposed regulatory approach. Science 214, 877–880 (1981).

103. Eaton, D. & Groopman, J. (eds) The Toxicology of Aflatoxins: Human Health, Veterinary, and Agricultural Significance (Academic Press, San Diego, 1998).

104. Payne, G. & Brown, M. Genetics and physiology of aflatoxin biosynthesis. Annu. Rev. Plant Path. 36, 329–362 (1998).

105. Hicks, J., Shimizu, K. & Keller, N. Genetics and biosynthesis of aflatoxins and sterigmatocystin. in The Mycota. Volume XI. Agricultural Applications, (ed. Kempken, F.) 55–69 (Springer–Verlag, Berlin, 2002).

106. Zilinskas, R. A. Iraq’s biological weapons. The past as future? J. Amer. Med. Assoc. 278, 418–424 (1997).

107. Green, G. The Human Factor (Everyman’s Library, London, 1979).

108. Centers for Disease Control and Prevention (CDC). Outbreak of aflatoxin poisoning — eastern and central provinces, Kenya. January–July 2004. Morb. Mortal. Wkly Rep. 53, 790–793 (2004).

109. Bennett, J. W. & Bentley, R. Pride and prejudice: the story of ergot. Persp. Biol. Med. 42, 333–355 (1999).

110. Ulrich, R. F. & Paten, B. M. The rise, decline and fall of LSD. Persp. Biol. Med. 34, 561–578 (1991).

111. Caporeal, L. Ergotism: the Satan loosed in Salem? Science 192, 21–26 (1976).

112. Matossian, M. Ergot and the Salem witchcraft affair. Am. Scientist 70, 355–357 (1982).

113. Cook, R. Acceptable Risk (Barkley, New York, 1996).114. Nierman, W. C. et al. Genomic sequence of the

pathogenic and allergenic filamentous fungus Aspergillus fumigatus. Nature (in the press).

115. Galagan, J. et al. Sequencing and comparative analysis of Aspergillus nidulans. Nature (in the press).

116. Machida, M. et al. Genome sequencing and analysis of Aspergillus oryzae. Nature (in the press).

AcknowledgementsGenomic data for Aspergillus fumigatus were provided by The Institute for Genomic Research and The Wellcome Trust Sanger Institute; genomic data for Aspergillus nidulans were provided by The Broad Institute; and genomic data for Aspergillus oryzae were provided by The National Institute of Advanced Industrial Science and Technology. Coordination of the analyses of these data was enabled by an international collaboration involving more than 50 institutions from 10 countries and coordinated from Manchester, UK.

Competing interests statementThe authors declare no competing financial interests.

Online linksDATABASESThe following terms in this article are linked online to:Entrez: http://www.ncbi.nlm.nih.gov/EntrezAspergillus flavus | Aspergillus parasiticus | Aspergillus terreus | Fusarium sporotrichioides | Magnaporthe grisea | Neurospora crassa

FURTHER INFORMATIONNancy P. Keller’s homepage: http://www.plantpath.wisc.edu/NewFacPage/faculty.htmGeoffrey Turner’s homepage: http://www.shef.ac.uk/mbb/staff/turnerThe Aspergillus website: http://www.aspergillus.man.ac.ukThe Aspergillus nidulans Database: http://www.broad.mit.edu/annotation/fungi/aspergillusCentral Aspergillus Data Repository: http://www.cadre.man.ac.ukDatabase of the Genomes Analysed at the National Institute of Advanced Industrial Science and Technology: http://www.bio.nite.go.jp/dogan/TopThe TIGR Aspergillus fumigatus Genome Project: http://www.tigr.org/tdb/e2k1/afu1The Wellcome Trust Sanger Institute Aspergillus fumigatus Genome Project: http://www.sanger.ac.uk/Projects/A_fumigatusAccess to this interactive links box is free online.


R E V I E W S

Fungal secondary metabolism — from biochemistry to genomics

Documents