Biosynthesis of Fusarubins Accounts for Pigmentation of Fusariumfujikuroi Perithecia

Lena Studt,a,b Philipp Wiemann,a Karin Kleigrewe,b Hans-Ulrich Humpf,b and Bettina Tudzynskia

Institute for Biology and Biotechnology of Plantsa and Institute of Food Chemistry,b Westfälische Wilhelms-University, Münster, Germany

Fusarium fujikuroi produces a variety of secondary metabolites, of which polyketides form the most diverse group. Among theseare the highly pigmented naphthoquinones, which have been shown to possess different functional properties for the fungus. Agroup of naphthoquinones, polyketides related to fusarubin, were identified in Fusarium spp. more than 60 years ago, but nei-ther the genes responsible for their formation nor their biological function has been discovered to date. In addition, although itis known that the sexual fruiting bodies in which the progeny of the fungus develops are darkly colored by a polyketide synthase(PKS)-derived pigment, the structure of this pigment has never been elucidated. Here we present data that link the fusarubin-type polyketides to a defined gene cluster, which we designate fsr, and demonstrate that the fusarubins are the pigments respon-sible for the coloration of the perithecia. We studied their regulation and the function of the single genes within the cluster by acombination of gene replacements and overexpression of the PKS-encoding gene, and we present a model for the biosyntheticpathway of the fusarubins based on these data.

The rice-pathogenic ascomycete Fusarium fujikuroi (teleo-morph, Gibberella fujikuroi MP-C), belonging to the Gibber-

ella fujikuroi species complex, was first isolated from infected riceplants in 1890 (33) and was identified as the causative agent of“bakanae,” or foolish seedling disease, due to its ability to producegibberellic acids (GAs) (34, 44). Typical symptoms of this diseaseare chlorotic, slender, and hyperelongated internodes as well asempty or sterile grains (69). Besides GAs, the fungus produces avariety of other secondary metabolites, such as the red polyketidebikaverin, accompanying the production of terpenoid GAs in liq-uid culture (10). Secondary metabolite genes are often organizedin gene clusters (32), such as those found in F. fujikuroi for theseven GA-biosynthetic genes (7) and the six genes involved inbikaverin biosynthesis (48, 80). It is assumed that spatial proxim-ity facilitates the coregulation of secondary metabolite geneswithin the cluster via, e.g., epigenetic processes, transcriptionalregulation by global regulators, or pathway-specific transcriptionfactors located within a gene cluster (53). Most of the secondarymetabolite gene clusters are silent under laboratory conditions(8). Different approaches have been undertaken for the activationof those gene clusters, e.g., optimization of culture conditions(pH, nutrient availability) (6, 81), overexpression of pathway-spe-cific transcription factors (9) or global transcriptional regulators,such as LaeA (53), and histone modifications affecting epigeneticcontrol (68).

Secondary metabolites can be grouped into four differentclasses depending on their structural properties: polyketides, ter-penes, nonribosomal peptides, and amino acid-derived com-pounds. Among secondary metabolites, polyketides form themost abundant group (37), including most of the green and redfungal pigments, all of which belong to the group of naphthoqui-nones.

These polyketides are synthesized by multidomain nonreduc-ing polyketide synthases (NR-PKSs). NR-PKSs are the minimalfungal PKSs, containing a �-ketoacyl synthase (KS) domain, amalonyl coenzyme A (malonyl-CoA):acyl carrier protein (ACP)transacylase (MAT) domain, and one or two ACP domains, re-sulting in a defined carbon skeleton through several rounds of

elongation (14). Recently, two new domains were identified byCrawford et al.: the starter unit ACP transacylase (SAT) domain,responsible for selecting the starter unit, and the product template(PT) domain. The SAT domain, which has high similarity to theMAT domain, was shown to accept acyl units other than malonyl-CoA as starter units and is widespread among fungal NR-PKSs(15, 16, 17). The PT domain, located between the acyl transferase(AT) and the ACP domain, is responsible for specific aldol cycli-zation and aromatization of poly-�-keto species, thereby deter-mining the chain length and specific cyclization patterns of thepolyketide precursors in fungal NR-PKSs (17, 18). Finally, therespective polyketide is released through various release mecha-nisms (24), of which thioesterase (TE)-mediated product releaseby a canonical TE domain is the most common. This domainoften extends to a domain capable of C-C Claisen cyclization (aTE/CLC domain) in fungal NR-PKSs, e.g., in Aspergillus parasiti-cus PksA (42). In some cases, NR-PKSs contain a reductase-releas-ing (R) domain; in others, e.g., the PKS involved in asperthecinbiosynthesis in Aspergillus nidulans, there is no releasing domainat all (14).

So far, the structures of more than 100 naphthoquinone metabo-lites have been elucidated (52), indicating the structural diversity ofthis group. The ability to produce naphthoquinones is widespreadamong fungal organisms, especially among members of the genusFusarium. To date, only a few of these compounds could be linked toPKS-encoding gene clusters. Among them are the following red pig-ments: bikaverin in F. fujikuroi (48, 79), aurofusarin in Fusariumgraminearum (28, 39, 51), and the perithecial pigments in Fusarium

Received 12 March 2012 Accepted 25 March 2012

Published ahead of print 6 April 2012

Address correspondence to Bettina Tudzynski, tudzynsb@uni-muenster.de, orHans-Ulrich Humpf, humpf@uni-muenster.de.

Supplemental material for this article may be found at http://aem.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AEM.00823-12

4468 aem.asm.org Applied and Environmental Microbiology p. 4468–4480 June 2012 Volume 78 Number 12

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

solani (30), F. graminearum (28), and F. verticillioides (60), whosestructures have not been identified yet.

In F. fujikuroi, the only PKS described is the bikaverin-specificNR-PKS Bik1 (79). In this study, we present the identification andfunctional characterization of a second NR-PKS in F. fujikuroiwith homology to two PKSs from other Fusarium spp., responsi-ble for the pigmentation of perithecia: the Pgl1 proteins from F.graminearum (28) and F. verticillioides (60). Targeted deletion ofthis PKS-encoding gene in two F. fujikuroi strains of opposite mat-ing types, and their subsequent sexual crossing, confirmed its rolein perithecial pigmentation in F. fujikuroi as well. Detailed molec-ular analyses revealed the presence of a gene cluster comprising sixcoregulated genes. Chemical analyses of the products accumu-lated by the single deletion mutants of these genes proved this genecluster to be responsible for the biosynthesis of naphthoquinoneswith structural similarity to fusarubin and allowed us to establishthe biosynthetic pathway. To our knowledge, this is the first reportto pinpoint perithecial pigmentation in Fusarium spp. to the genecluster responsible for the production of fusarubins.

MATERIALS AND METHODSFungal strains and culture conditions. The wild-type (Wt) Fusariumfujikuroi strain IMI58289 (Commonwealth Mycological Institute, Kew,United Kingdom) was used as a parent strain for all knockout experi-ments. As a mating partner, F. fujikuroi C1995 (kindly provided by J. F.Leslie, Kansas State University) was used. For all liquid cultivations, the F.fujikuroi strains were preincubated for 72 h in 300-ml Erlenmeyer flaskswith 100 ml Darken medium (DVK) (22) on a rotary shaker at 28°C and180 rpm. A 500-�l aliquot of this culture was used for inoculation ofsynthetic ICI (Imperial Chemical Industries Ltd., United Kingdom) me-dium (29) with either 6 mM glutamine or 6 mM sodium nitrate, andincubation was carried out for an additional 1 to 12 days. For protoplas-ting, 500 �l of the preincubated culture was taken and transferred to 100ml ICI medium with 10 g/liter fructose instead of glucose and 1 g/liter(NH4)2SO4 as the nitrogen source; this mixture was incubated on a rotaryshaker at 28°C and 180 rpm for no longer than 16 h. For the nitrogen shiftexperiments, the mycelia were grown in ICI medium on a rotary shaker at28°C and 180 rpm. After 4 days, either 33 mM glutamine, 33 mM sodiumnitrate, or no nitrogen was added, and the cultures were incubated for anadditional 30 min. For the identification of the compounds, the funguswas grown for 7 days in ICI medium on a rotary shaker at 28°C and 180rpm. For pH shift experiments, the strains were grown in ICI medium for4 days on a rotary shaker at 28°C and 180 rpm; then they were transferredto ICI medium with no nitrogen source, adjusted to a pH of either 4 or 8,for an additional 2 h (11). For sexual crosses, carrot agar was used accord-ing to the method of Klittich and Leslie (41). For DNA isolation, thefungus was grown on cellophane sheets on solidified complete medium(CM) for 3 days at 28°C. For RNA isolation, the fungus was grown in ICImedium with the desired nitrogen source for 2 to 7 days at 28°C and 180rpm after preincubation in DVK.

HPLC-DAD analyses of the naphthoquinones. For high-perfor-mance liquid chromatography (HPLC)-diode array detector (DAD) anal-ysis, culture fluid from 1- to 12-day-old cultures was filtered over a 0.2-�m-pore-size membrane filter (Millex; Millipore) and was used directlywithout further preparation. The samples were separated on a 150- by3.00-mm, 3-�m Gemini 3u C6-phenyl 110-Å column (Phenomenex,Aschaffenburg, Germany) by using a binary gradient delivered by a Hita-chi L-7100 (Merck) pump with 1% formic acid as solvent A and acetoni-trile (ACN) as solvent B. The binary gradient started with a linear gradientfrom 20% B to 50% B in 15 min, followed by an isocratic step for 5 minand an additional linear gradient step to 75% B in 5 min. After each HPLCrun, the column was washed by increasing the gradient to 100% B and wasequilibrated for 5 min under the starting conditions. The flow rate was setto 0.3 ml min�1. The desired compounds were detected using a Hitachi

L-2455 Elite LaChrom DAD. To compare the chromatograms of differentculture filtrates, the wavelength was set to 450 nm.

HPLC-UV-FTMS analyses of the naphthoquinones. Liquid media of1- to 12-day-old cultures were filtered over a 0.2-�m-pore-size mem-brane filter and were directly analyzed by high-performance liquid chro-matography coupled with Fourier transformation mass spectrometry(FTMS) detection (HPLC-UV-FTMS). To this end, an Accela LC 60057-60010 system (Thermo Fisher Scientific, Bremen, Germany) was linked toa linear trap quadrupole (LTQ) Orbitrap XL mass spectrometer (ThermoFisher Scientific). Data were acquired with Xcalibur 2.07 SP1 (ThermoScientific). The samples were separated on a 150- by 3.00-mm, 3 �mGemini 3u C6-phenyl 110-Å column (Phenomenex, Aschaffenburg, Ger-many) by using a binary gradient as described above. The flow rate was setto 0.3 ml min�1. Ionization was performed with heated electrospray ion-ization. Further mass spectrometer conditions were as follows: capillarytemperature, 225°C; atmospheric pressure chemical ionization (APCI)vaporizer temperature, 250°C; sheath gas flow, 40 units; auxiliary gas flow,20 units; source voltage, 3.5 kV; capillary voltage, 35 V; tube lens, 110 V;multiple 00 offset, �4.00 V; lens 0 voltage, �4.20 V; gate lens offset,�35.00 V; multipole 1 offset, �8.00 V; front lens, �5.25 V. Scan eventswere as follows. (i) A total-ion scan of a mass range from m/z 250 to 800with a resolution of 30,000 in the positive-ion mode (FTMS) was carriedout. (ii) Depending on scan event i, the most intense parent ion wasfragmented in the ion trap mass spectrometry (ITMS). If no parent masswas found, the most intense ion was fragmented: activation type, colli-sion-induced dissociation (CID); normalized collision energy, 35; activa-tion Q, 0.250; isolation width, 2. (iii) Depending on scan event i, the mostintense parent ion of the list was fragmented. If no parent mass was found,the most intense ion was fragmented in the FTMS: activation type, higher-energy collisional dissociation (HCD); normalized collision energy, 50;activation Q, 0.250; isolation width, 1.5; resolution, 15,000. (iv) A total-ion scan of a mass range from m/z 200 to 800 in the negative-ion modewith the ITMS was carried out. The exact masses of the compounds wereidentified if the proton adduct ([M � H]�), and, in addition, the sodiumadduct ([M � Na]�) or the exact mass in the negative-ion mode ([M �H]�) was also detected.

Purification of the naphthoquinoid metabolites. For the purificationof the naphthoquinones, 1 liter of 7-day-old culture fluid of the wild type(Wt) (for final products), the �fsr2 and �fsr3 deletion mutants (for inter-mediates), and the �fsr2-5/gpd::fsr1 strain (for the PKS-derived product)was extracted by solid-phase extraction (SPE) using C18 SPE cartridges(Phenomenex, Aschaffenburg, Germany). Further purification was ac-complished by semipreparative HPLC on a 250- by 10.00-mm, 5 �mGemini 5u C6-phenyl 110-Å column (Phenomenex, Aschaffenburg, Ger-many) by using an isocratic gradient delivered by a Jasco PU-2087 pumpwith 1% formic acid as solvent A and ACN as solvent B. The flow rate wasset to 3 ml min�1. The desired compounds were detected using a JascoUV-2075 UV detector at 450 nm (for final products) or 400 nm (forintermediates and PKS-derived products). Naphthoquinoid metabolites1 to 8 were obtained after semipreparative HPLC in amounts of 10 to 30mg for structure elucidation.

NMR measurements. 1H, 13C, and 2-dimensional (2-D) nuclear mag-netic resonance (NMR) spectra were acquired on a Bruker DPX-400(Bruker BioSpin, Rheinstetten, Germany) NMR spectrometer. Signals arereported in parts per million referenced to dimethyl sulfoxide (DMSO)-d6. For structural elucidation and NMR signal assignment, 2-D NMRexperiments, such as (H,H)-correlated spectroscopy (H,H-COSY), het-eronuclear multiple-quantum correlation (HMQC), and heteronuclearmultiple-bond correlation (HMBC) experiments, were performed. Pulseprograms for these experiments were taken from the software library.

Sequence data and phylogenetic analyses. BLASTP analysis (1) wasperformed using the predicted protein sequences of Fsr1 to Fsr6, as well asFF_03983 and FF_03990, against NCBI nonredundant protein sequences(Table 1). PKS protein sequences were retrieved from the following sources:for F. graminearum, F. oxysporum, and F. verticillioides, the Broad Institute

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4469

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

(www.broadinstitute.org); for F. solani, the Joint Genome Institute (www.jgi-psf.org); and for characterized PKSs of other fungi (14, 32), NCBI(www.ncbi.nlm.nih.gov). KS domains used for phylogenetic analyses wereextracted using the PKS/NRPS Analysis website (http://nrps.igs.umaryland.edu/nrps) (4). Phylogenetic analysis was performed by comparing the KSdomains using the Web-based tool at www.phylogeny.fr (23).

Plasmid constructions. The F. fujikuroi knockout strains were createdusing yeast recombinational cloning (20). The 5= and 3= flanks of thecorresponding genes were amplified with appropriate primer pairs (seeTable S1 in the supplemental material); for the 5= flank, primers 5F and 5Rwere used, and for the 3= flank, primers 3F and 3R were used, based on thegenomic sequence of F. fujikuroi IMI58289 (B. Tudzynski and coworkers,unpublished data). The hygromycin resistance cassette, consisting of thehygromycin B phosphotransferase gene hph (31), driven by the trpC pro-moter, was amplified using the hph-F/hph-R primer pair from the tem-plate pCSN44 (66). Alternatively, the nourseothricin resistance cassette,also driven by the trpC promoter, was used. The nourseothricin resistancecassette was amplified by using plasmid pZPnat1 (GenBank accessionnumber AY631958.1) as a template with the primer pair hphF/hphR-trpC-T2. All primers used are listed in Tables S1 to S4 in the supplementalmaterial. The fragments obtained were cloned into the Saccharomycescerevisiae strain FY834 (82) together with the EcoRI/XhoI-restricted plas-mid pRS426 (19). For restoration of pigment biosynthesis in the �fsr1mutant, the respective gene was amplified with the corresponding termi-nator sequence in four PCRs. All four amplified products were cloned intoSaccharomyces cerevisiae FY834 together with the HindIII-restricted mod-ified plasmid pRS426, which additionally contained a nourseothricin re-sistance cassette driven by an oliC promoter amplified from pNR1 (50)using the primer pair nat-OE-prom/nat-OE-term. The gene of interestitself was driven by the gpd promoter, which was amplified from pveAgfp(80) using the gpd-yeast-for/gpd-yeast-rev primer pair.

Standard molecular methods. For DNA isolation, lyophilized myce-lium was ground to a fine powder in liquid nitrogen, dispersed in extrac-tion buffer according to the method of Cenis (12), and afterwards used forPCR amplification and Southern blot analyses. PCR mixtures contained25 ng genomic DNA, 5 pmol of each primer, 200 nM deoxynucleosidetriphosphates, and 1 U of BioTherm DNA polymerase (GeneCraft

GmbH, Lüdinghausen, Germany) for diagnostic PCR and amplificationof the flanks for yeast recombinational cloning. PCRs were performedwith an initial denaturing step at 94°C for 3 min, followed by 35 cycles of1 min at 94°C, 1 min at 56 to 60°C, and 1 to 2 min at 70°C, and a finalelongation step at 70°C for 10 min. For amplification of the knockoutconstructs, the TaKaRa polymerase kit was used as recommended by themanufacturer.

For restoration of fusarubin biosynthesis in the�fsr1 mutant, as well as forthe identification of the PKS-derived product, fsr1 was amplified in 4 frag-ments by using a proofreading polymerase. Under these conditions, PCRmixtures contained 25 ng genomic DNA, 5 pmol of each primer, and 1 U ofPhusion polymerase (Finnzymes, Thermo Fisher Scientific, Finland).

For Southern blot analysis, genomic DNA was completely digestedusing appropriate enzymes, separated on a 1% (wt/vol) agarose gel, andtransferred to nylon membranes (Nytran SPC; Whatman) by downwardblotting (3). 32P-labeled probes were generated using the randomoligomer-primer method and were hybridized to the membranes over-night at 65°C according to the protocol of Sambrook et al. (64). Afterhybridization, the membrane was washed with 1� SSPE (0.18 M NaCl, 10mM NaH2PO4, and 1 mM EDTA [pH 7])– 0.1% sodium dodecyl sulfate(SDS) at the same temperature.

For RNA isolation, lyophilized mycelium was ground with a mortarand pestle, and the powder was extracted by using the RNAagents total-RNA isolation kit (Promega, Germany) according to the manufacturer’sinstructions. Twenty micrograms per sample was loaded onto a 1% aga-rose gel and was run under denaturing conditions (64). The separatedRNA was transferred to nylon membranes (Nytran SPC; Whatman). 32P-labeled probes were created as described above and were hybridized to themembrane overnight at 65°C. Plasmid DNA from Saccharomyces cerevi-siae was extracted using the yeast plasmid isolation kit (SpeedPrep; Dual-systems Biotech) and was directly used for PCR.

Fungal transformations. Protoplasts were prepared from F. fujikuroiIMI58289 as described previously (74). About 107 protoplasts were trans-formed with 10 �g of the amplified replacement cassettes for the knock-outs of the putative gene cluster (fsr1 to fsr6) and one gene each upstream(FF_03983) and downstream (FF_03990) of the cluster. For the restora-tion of fsr1 and for the identification of the PKS-derived product, the fsr1

TABLE 1 Information about the putative fsr gene cluster

Gene Length (bp) IntronNo. of aminoacids InterPro annotation Domains and motifs Predicted function

FF_03983 1,788 0 595 IPR002937 Amine oxidase Unknown functionfsr1 (FF_03984) 7,069 4 2,286 IPR014031 �-Ketoacyl synthase Nonreducing PKS responsible for the

IPR014043 Acyltransferase domain condensation of seven acetyl-CoAIPR009081 Acyl carrier protein-like unitsIPR013120 Thioesterreductase

fsr2 (FF_03985) 1,209 1 387 IPR001077 O-Methyltransferase, family 2 Methyltransferase responsible for themethylation of hydroxyl groups atC-6 and C-8

fsr3 (FF_03986) 1,660 2 522 IPR002938 Monooxygenase, FAD binding Monooxygenase responsible forreduction/oxidation of the PKS-derived aldehyde as well as forincorporation of hydroxyl groupsat C-5, C-10, and C-13

IPR003042 Aromatic-ring hydroxylase-like

fsr4 (FF_03987) 1,117 2 338 IPR002085 Alcohol dehydrogenasesuperfamily, zinc containing

Unknown function

IPR016040 NAD(P)-binding domainfsr5 (FF_03988) 794 1 256 IPR002198 Short-chain

dehydrogenase/reductaseUnknown function

fsr6 (FF_03989) 1,014 0 338 IPR001138 Zn(II)2C6 fungal-type DNA-binding domain

Zn(II)2Cys6 transcription factor

IPR013700 Aflatoxin-regulatory proteinFF_03990 1,181 1 374 None Unknown function

Studt et al.

4470 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

and �fsr2-5 deletion mutants were transformed with 10 �g of the pRS426vector containing fsr1 driven by the gpd promoter.

For sexual crosses, C1995, the mating partner of F. fujikuroi IMI58289, was also transformed with the knockout cassette for fsr1. Trans-formed protoplasts were regenerated as described by Tudzynski et al. (75).The medium contained the appropriate resistance marker. Single conidialcultures were established from either hygromycin B- or nourseothricin-resistant transformants and were used for DNA isolation and Southernblot analysis.

Nucleotide sequence accession number. The nucleotide and proteinsequences of the fsr gene cluster have been deposited in GenBank underaccession number HE613440.

RESULTSModification of culture condition reveals a second redpolyketide in Fusarium fujikuroi. In F. fujikuroi, the biosynthesisof both GAs and bikaverin is strongly repressed by high levels ofglutamine. In synthetic ICI medium with low levels of glutamine(6 mM), the fungus starts to produce these metabolites after ni-trogen depletion (10, 79). However, when the fungus was grownwith 6 mM sodium nitrate as the sole nitrogen source, only GAswere produced; accumulation of bikaverin was no longer de-tected, due to the alkaline pH (Fig. 1B). Surprisingly, the liquidculture still was deeply red pigmented after approximately 5 daysof cultivation (Fig. 1A). High-performance liquid chromatogra-phy (HPLC) and HPLC coupled with Fourier transformationmass spectrometry detection (HPLC-UV-FTMS) (data notshown) confirmed that the red pigment was not bikaverin (Fig.1B). Measurements of the pH values of the different growth mediaover a time course showed significant differences depending onthe nitrogen source. While the pH was maintained at acidic valuesbetween 5 and 6 with glutamine, the pH increased to values of 6.5to 7.5 in cultures with sodium nitrate (Fig. 1C). The ambient al-

kaline conditions explain the lack of bikaverin biosynthesis, asshown previously (79). The second red pigment, in contrast, wasproduced only if the ambient pH of the liquid culture immediatelyincreased to a neutral or alkaline value. Thus, pH conditions in thefirst 24 h after inoculation seem to determine which of the pig-ments is produced.

Cloning, disruption, and characterization of the responsiblepolyketide synthase-encoding gene (fsr1). Due to the intense redcolor of the cryptic compound, we assumed structural similarityto aromatic pigments synthesized by NR-PKSs, and hence we hy-pothesized the involvement of a nonreducing PKS in the biosyn-thesis of this compound. A BLASTP analysis (1) with the Bik1sequence against the recently sequenced genome of F. fujikuroiIMI58289 (Tudzynski and coworkers, unpublished) revealed asecond NR-PKS with 40% identity (E value, 0) to Bik1. This PKS(Fsr1) consists of 2,286 amino acids (aa), encoded by an openreading frame (ORF) of 7,069 bp, which is interrupted by fourputative introns. ClustalW alignment of Fsr1 with the well char-acterized NR-PKSs responsible for aflatoxin production in A.parasiticus (PksA), perithecial pigment production in F. solani(PksN), and bikaverin production in F. fujikuroi (Bik1) revealedthe presence of the following domains. (i) In the N terminus, therecently identified SAT domain, harboring the active GXCXG site,can be found (Fig. 2A; see also Fig. S1 in the supplemental mate-rial). (ii) The SAT domain is followed by the KS and MAT do-mains (Fig. 2A; see also Fig. S1). (iii) Next, we identified a domainwith high similarity to the newly identified PT domain presentingthe catalytic histidine and aspartate residues (Fig. 2A; see also Fig.S1). (iv) Adjacent to the PT domain, Fsr1 possesses two ACP do-mains (Fig. 2A; see also Fig. S1). (v) At the C terminus, the enzymeexhibits a reductase (R) domain (Fig. 2A) instead of the canonical

FIG 1 Identification of a new group of red pigments by changing standard growth conditions. Wild-type (Wt) F. fujikuroi was grown in ICI medium with either6 mM glutamine (Gln) or 6 mM sodium nitrate (NO3

�) for 5 days. Samples were taken and were used for HPLC-DAD analyses as described in Materials andMethods. (A) Images of the flasks with the Wt after 5 days of inoculation. (B) HPLC chromatogram (detected at 450 nm). Peaks for norbikaverin (compound 1)and bikaverin (compound 2) are labeled in the upper chromatogram. New, unknown compounds in the lower chromatogram are labeled with a question mark.(C) pH values over 5 days. The pH was measured every 12 h.

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4471

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

TE domain. In contrast to TE and TE/CLC domains, R domainsshow sequence similarities to the short-chain dehydrogenase/re-ductase (SDR) superfamily, exhibiting Rossman fold structureand nucleotide binding motifs (36). ClustalW alignment of theFsr1 R domain with characterized R domains from bacterial andfungal PKSs (5, 13, 38, 45) and the R domain of the yeast �-amino-adipate reductase Lys2p (25) shows high amino acid conservation(see Fig. S2 in the supplemental material).

Phylogenetic analysis of the KS domains of characterized fun-gal NR-PKSs revealed a close relatedness of Fsr1 to the Pgl1 pro-

teins of F. verticillioides and F. graminearum, both of which havebeen shown to be responsible for perithecial pigmentation (28,60). However, the chemical structures of these perithecial pig-ments have not been elucidated yet.

In order to investigate whether fsr1 is involved in the produc-tion of the cryptic pigment observed in axenic culture (Fig. 1B),the ORF was deleted by targeted gene replacement. Three inde-pendent knockout mutants were obtained and were designated�fsr1-T4, �fsr1-T6, and �fsr1-T8. Diagnostic PCR (data notshown) and Southern blot analysis (see Fig. S3 in the supplemental

FIG 2 Phylogeny, PKS domains, and the gene cluster. (A) KS domains of fungal PKSs used for the analysis were extracted, and a phylogram was generated asdescribed in Materials and Methods. For characterized PKSs, the leaves of the phylogram are labeled with the organism, protein name, and GenBank accessionnumber (EMBL database). For uncharacterized Fusarium PKSs, the Broad Institute identification is given for F. graminearum (FGSG), F. oxysporum (FOXG),and F. verticillioides (FVEG), and the Joint Genome Institute identification is given for F. solani. The domain architecture of each PKS and the chemical productproduced are given on the right. Branches with branch support values of �0.5 were collapsed. The KS domain of the highly reducing PKS Fum1 was used as anoutgroup. The bar represents 0.3 character change. Organisms: Af, Aspergillus fumigatus; An, A. nidulans; Ap, A. parasiticus; As, Acremonium strictum; At,Aspergillus terreus; Cl, Colletotrichum lagenarium; Cn, Cercospora nicotianae; Ds, Dothistroma septosporum; Ed, Exophiala dermatitidis; Fg, F. graminearum; Ff, F.fujikuroi; Fs, F. solani; Fv, F. verticillioides; Gl, Glarea lozoyensis; Mp, Monascus purpureus; Nsp., Nodulisporium sp. strain ATCC 74245. Domains: SAT, starter unitACP transacylase domain (superscript letters indicate an amino acid other than C at the third position of the active site); KS, �-ketoacyl synthase domain; MAT,malonyl-CoA:ACP transacylase domain; PT, product template domain; ACP, acyl carrier protein domain; TE, thioesterase domain; TE/CLC, thioesterase/Claisen cyclase domain; R, reductase domain; DH, dehydratase domain; ER, enoyl reductase domain; KR, keto reductase domain; CMet, C-methyltransferasedomain. (B) The components of the fsr gene cluster, fsr1 to fsr6, are depicted as filled rectangles. The first genes upstream (FF_03983) and downstream (FF_03990)are shaded. Arrows indicate the direction of translation, and open bars represent introns within the gene.

Studt et al.

4472 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

material) confirmed the absence both of the wild-type gene and ofadditional ectopic integrations of the replacement fragment.When grown in a medium with 6 mM sodium nitrate, none of themutants exhibited red pigmentation, indicating that Fsr1 is re-sponsible for the biosynthesis of the novel red pigment. Since all�fsr1 mutants behaved similarly (see Fig. S4 in the supplementalmaterial), �fsr1-T4 was arbitrarily chosen for subsequent work inthis study. Reintroduction of the complete ORF into the �fsr1mutant resulted in full restoration of pigment biosynthesis, con-firming that Fsr1 is indeed responsible for the formation of thenew red pigments (Fig. 3C).

Identification and characterization of the correspondinggene cluster. Since the biosynthetic genes for many fungal second-ary metabolites are organized in gene clusters (32), we analyzedthe genes in the proximity of fsr1. The five genes downstream offsr1 (Fig. 2B) show homologies to genes known to be typicallyinvolved in the biosynthesis of secondary metabolites (Table 1).To determine whether these genes adjacent to fsr1 are involved inpigment biosynthesis, we studied their expression profiles underbiosynthesis-favoring (6 mM sodium nitrate) and non-biosynthe-sis-favoring (6 mM glutamine) conditions. Northern blot analysesrevealed the coregulation of six genes (fsr1 to fsr6). FF_03983,

upstream of fsr1, and FF_03990, downstream of fsr6, are not ex-pressed and therefore are probably not involved in the biosynthe-sis (Fig. 4).

Regulation of the fsr gene cluster. To gain deeper insight intothe regulation of fsr gene expression, a time course growth exper-iment was performed under pigment-favoring conditions. North-ern blot analyses revealed time-dependent expression of the fsrgenes (shown for fsr1 to fsr3): transcripts are detectable at the thirdday; their intensity peaks at the fourth day and then decreasescontinuously until the signals are almost undetectable at day 6postinoculation (Fig. 5A). To further unravel the pH-dependentexpression, the wild type was grown under biosynthesis-favoringconditions for 5 days before the mycelium was shifted into syn-thetic ICI medium without any nitrogen source and with the pHvalue set either to 4 or to 8. The results clearly show that fsr2 andfsr3 gene expression is repressed under acidic conditions, while thegenes are expressed under alkaline conditions (Fig. 5B).

To study the impact of nitrogen quality and quantity on geneexpression in more detail, the wild type was grown for 5 daysunder pigment production conditions. Then either 33 mM glu-tamine, 33 mM sodium nitrate, or water (no nitrogen) was addedto the cultures, and the effect on fsr gene expression was studied

FIG 3 fsr1 is responsible for the production of fusarubins. Fungal strains weregrown in 6 mM sodium nitrate. Culture filtrates were taken after 6 days ofgrowth, when the wild type (Wt) showed red pigmentation. Samples were useddirectly for HPLC-DAD measurements, as described in Materials and Meth-ods. Shown are the fungal strains in liquid culture with the correspondingHPLC chromatograms (detected at 450 nm). (A) Wild type; (B) �fsr1 deletionmutant; (C) restoration of fsr1 in the deletion mutant (�fsr1/fsr1c). Peaks as-sociated with Fsr1 are outlined in blue.

FIG 4 Coregulation of the putative fsr gene cluster. The wild type (Wt) wasgrown in ICI medium with either 6 mM glutamine (Gln) or 6 mM sodiumnitrate (NO3

�). Lyophilized mycelium was used for Northern blot analyses.The putative genes of the fsr gene cluster (fsr1 to fsr6) and one gene eachupstream (FF_03983) and downstream (FF_03990) of the gene cluster wereused for probing.

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4473

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

after an additional 30 min. Northern blot analyses show that glu-tamine strictly represses pigment biosynthesis, while the genes arestill expressed if sodium nitrate or water is added (Fig. 5C).

Next, we wanted to determine whether single deletions of eachof the fsr genes affect the expression of the other cluster genes, ashas been shown for bikaverin genes (79). For this purpose, dele-tion mutants of all six genes were generated. Surprisingly, deletionof the PKS-encoding gene fsr1 resulted in complete downregula-tion of the other five fsr genes, while single deletions of the fsr2 tofsr5 genes did not affect the expression of any other pathway gene.The deletion of fsr6, encoding a putative Zn(II)2Cys6 transcrip-tion factor, resulted in the expected downregulation of all fsrgenes, including fsr1 (see Fig. S6A in the supplemental material).

Chemical identification of intermediates and final products.Cultivation of the wild type in ICI medium with sodium nitrateas the sole nitrogen source resulted in a complex spectrum ofdifferent pathway-specific compounds (Fig. 6A), all of whichare missing in the �fsr1 deletion mutant (Fig. 6B). The struc-ture of the main compounds (except for compound 2, due toinsufficient accumulation in the liquid culture) was extensivelyelucidated using their exact masses, characteristic UV spectra,and nuclear magnetic resonance (NMR) data. By these meth-ods, we identified the previously described naphthoquinones8-O-methylfusarubin (compound 1), 8-O-methylnectriafu-rone (compound 2), 8-O-methyl-13-hydroxynorjavanicin(compound 3), and 13-hydroxynorjavanicin (compound 5),along with one new compound, which was subsequently desig-nated 8-O-methylanhydrofusarubinlactol (compound 4) (67,70–72). The structures identified were in agreement with pub-lished data. Additionally, we are presenting, for the first time,13C NMR data for compounds 2 to 5 (for detailed spectroscopicdata, see the supplemental material). These compounds andstructurally related metabolites are known to be produced bydifferent Fusarium spp. The first naphthoquinones with a sim-ilar structure were isolated and characterized by Arnstein et al.from Fusarium javanicum and were therefore designated ja-vanicin and oxyjavanicin (2), the latter of which was later re-

named fusarubin (63). O-Demethylanhydrofusarubin was thefirst and, until now, the only naphthoquinone pigment of thefusarubin family to be described in F. fujikuroi (21). To ourknowledge, none of the other compounds presented here havebeen described in F. fujikuroi before.

To confirm that the six coregulated fsr genes are indeed in-volved in the biosynthesis or regulation of pigment production,the product spectra of the �fsr1 to �fsr6 mutants and of deletionmutants of the two border genes, FF_03983 and FF_03990 (Fig.2B), were analyzed. The fact that the FF_03983 and FF_03990deletion strains showed no alteration in the product spectrastrengthened the suggestion made on the basis of our coregulationstudies (Fig. 4) that these genes are not involved in the biosynthe-sis or regulation of the pigments (data not shown).

The pathway-related products synthesized by the �fsr1 to �fsr6deletion mutants were analyzed in detail to demonstrate their in-volvement in the biosynthesis of fusarubins (Fig. 6B). The �fsr1,�fsr2, and �fsr3 mutants are the only mutants showing an altera-tion in the product spectrum, indicating their role in the forma-tion of the final products.

Under pigment-favoring conditions, three major peaks wereaccumulated by the �fsr2 and �fsr3 mutants; they are labeled ascompounds 6 and 7 in the chromatogram of the �fsr2 mutant andcompound 8 in that of the �fsr3 mutant. Compound 6 showed anNMR spectrum similar to that of 6-O-demethyl-5-deoxyfusaru-bin, previously identified from Nectria haematococca (55), butwith slightly different but distinct chemical shifts resulting fromthe position of the hydroxyl group at C-10 instead of C-5. Hence,compound 6 was designated 6-O-demethyl-10-deoxyfusarubin.Compounds 7 and 8 gave similar NMR spectra of the polyketidebackbone, without the hydroxyl groups at positions C-5 and C-10,where the A and B rings are already formed, but not the C ring.The structures show characteristic chemical shifts indicating thepresence of an aldehyde, which was confirmed by 2-D NMR anal-yses. The only difference is the presence of an additional methylgroup at position C-6 in compound 8. The two intermediates weretherefore designated 6-O-demethylfusarubinaldehyde and fus-

FIG 5 The expression of fsr genes depends on culture age, pH, and nitrogen. (A) The wild type (Wt) was grown in ICI medium with 6 mM sodium nitrate. Themycelium was harvested after 2, 3, 4, 5, 6, and 7 days (t/d). Lyophilized mycelium was used for Northern blot analyses. For probing, the fsr1 to fsr3 genes were used.(B) The Wt was grown in ICI medium with 6 mM sodium nitrate for 4 days and was then shifted into a liquid medium set to pH 4 or pH 8, without any nitrogen.The mycelium was harvested after additional inoculation for 2 h and was used for Northern blot analysis. fsr2 and fsr3 were used for probing. (C) The Wt wasgrown in ICI medium with 6 mM sodium nitrate. After 4 days, either 33 mM glutamine [�N (Gln)], 33 mM sodium nitrate [�N (NO3

�)], or water (�N) wasadded. After additional inoculation for 30 min, the mycelium was harvested and was used for Northern blot analyses. The fsr2 to fsr5 genes were used for probing.

Studt et al.

4474 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

arubinaldehyde, respectively (for detailed spectroscopic data, seethe supplemental material). None of these three compounds havebeen described before.

In contrast, deletion of fsr4 or fsr5 did not alter the product spec-trum of the metabolites, indicating that these genes are not involvedin the modification of the PKS-derived precursor under these specificconditions. Interestingly, product formation was enriched in the�fsr4 mutant compared to that in the wild type, suggesting that fsr4might possess a regulatory function in the biosynthesis of fusarubins.Deletion of fsr6, a putative Zn(II)2Cys6 transcription factor, resultedin the total loss of the respective products (Fig. 6B), indicating itsfunction as a pathway-specific transcription factor.

In order to identify the Fsr1-derived product, a mutant inwhich all genes except fsr1 and fsr6 are deleted (�fsr2-5) wasgenerated. Surprisingly, this deletion resulted not only in thedownregulation of the deleted genes, fsr2 to fsr5, but also in thatof the two remaining genes, fsr1 and fsr6 (see Fig. S6B in thesupplemental material). To overcome this effect, an fsr1 over-expression vector (gpdprom:fsr1) was transformed into the�fsr2-5 deletion mutant, resulting in �fsr2-5/OE:fsr1 mutantsoverexpressing fsr1 under the control of the strong A. nidulansgpd (glucose-6-phophatase dehydrogenase) promoter. By useof this approach, one major compound accumulated and wasfound to be identical with 6-O-demethylfusarubinaldehyde(compound 7), found in the �fsr2 mutant (Fig. 6C; for detailedinformation, see the supplemental material). This heptaketide,lacking the methyl group as well as the two hydroxyl groups atC-5 and C-10, is therefore the earliest intermediate in the bio-synthetic pathway of the fusarubins.

The fusarubins are responsible for the pigmentation of theperithecia. Due to the close relatedness of Fsr1 to the Pgl1 proteinsof F. graminearum and F. verticillioides (28, 60) (Fig. 2A), wewanted to investigate whether the fusarubin-type compounds areidentical with the yet unknown perithecial pigments in F. fujiku-roi. Therefore, sexual crosses were performed between the wild-type strain IMI58289 (Mat-1) and F. fujikuroi strain C1995, withthe opposite mating type (Mat-2), on the one hand, and betweenthe IMI58289/�fsr1 and C1995/�fsr1 deletion mutants, on theother hand. For this purpose, the fsr1 knockout fragment was alsotransformed into C1995, and deletion mutants were identified bydiagnostic PCR (data not shown). Microscopic analysis of thefruiting bodies resulting from both sexual crosses clearly showedthat the loss of fsr1 in both mating partners resulted in colorlessperithecia that lack the normal purple pigmentation (Fig. 7).

DISCUSSION

Polyketides represent a highly diverse group of natural productswithin fungal secondary metabolites. They are generated by largemultidomain enzymes, the PKSs (14). Among them are naphtho-quinones, a class of polyketides formed by NR-PKSs. They areoften deeply colored, and many of them show phytotoxic, insec-ticidal, antibacterial, and fungicidal activities (52). Due to theirstructural properties, they are also thought to be involved in pro-tection against extreme environmental conditions, such as UVirradiation and desiccation (30).

Until now, only the red polyketides bikaverin and norbikaverin(40), as well as O-demethylanhydrofusarubin (21), were identifiedas naphthoquinone pigments of F. fujikuroi. In this study, we de-scribe the identification of new secondary metabolites with naph-thoquinoid structures, as well as the corresponding biosyntheticgenes, organized in a gene cluster in the genome of F. fujikuroi.These metabolites are red pigmented like bikaverin, but their pro-duction is differently regulated by external factors such as pH and

FIG 6 Chemical analyses of fungal strains. For analyses of compounds, theindicated fungal strains were grown under conditions favoring fusarubin bio-synthesis (ICI with 6 mM sodium nitrate). Samples were taken after 6 days ofgrowth and were directly used for HPLC-DAD measurement as described inMaterials and Methods. Numbers indicate compounds found in the liquidculture (compound 1, 8-O-methylfusarubin; compound 2, 8-O-methylnec-triafurone; compound 3, 8-O-methyl-13-hydroxynorjavanicin; compound 4,8-O-methylanhydrofusarubinlactol; compound 5, 13-hydroxynorjavanicin;compound 6, 6-O-demethyl-10-deoxyfusarubin; compound 7, 6-O-demeth-ylfusarubinaldehyde; compound 8, fusarubinaldehyde). (A) Chemical analysisof the wild type (Wt). (B) Chemical analyses of the �fsr1 to �fsr6 single dele-tion mutants. (C) Identification of the PKS-derived product by overexpressionof fsr1 in the �fsr2-5 mutant. (D) Structural formulae of compounds identifiedin the wild-type strain (compounds 1 to 5), in the �fsr2 (compounds 6 and 7)and �fsr3 (compound 8) deletion mutants, and in the �fsr2-5/gpd::fsr1 strain(compound 7).

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4475

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

nitrogen sources. The new red pigments were identified as mem-bers of the fusarubin family (Fig. 6D). Although most of thesestructures have been described previously in different Fusariumspp. (52), their biological function and biosynthetic genes haveremained mysterious until now.

Fusarubin-type naphthoquinones are responsible for thecoloration of the fruiting bodies. In this study, we show for thefirst time that the fusarubin-type naphthoquinones (Fig. 6D)are identical with the perithecial pigments (Fig. 7). Perithecia arethe sexual fruiting bodies in which the progeny of the fungus, theascospores, are formed. Here naphthoquinones might function asprotectants, as suggested for their role in F. solani (30), againstharmful environmental conditions, such as reactive oxygen spe-cies (ROS), UV irradiation, desiccation, and fungivorous insects.In A. fumigatus, the conidia of mutants defective in PKS-derivedpigment biosynthesis are more susceptible to killing by macro-phages than are pigmented wild-type conidia (35). Furthermore,studies of A. nidulans have shown that certain insects preferen-tially feed on fungal mutants exhibiting a restricted secondarymetabolite profile, indicating that a fungus capable of the produc-tion of its full secondary metabolite spectrum has an enhancedsurvival rate (61). The colorless perithecia of fsr1 mutants werelarger than the colored perithecia produced by the wild type (Fig.7), suggesting that the red pigments are also important for theintegrity of the perithecial cell wall. A similar correlation wasfound for A. fumigatus, where PKS-derived melanin is requiredfor the correct assembly of the cell wall layers in the conidia (59).Interestingly, in F. fujikuroi, the fusarubin-type pigments are notrestricted to the fruiting bodies but also occur in the culture brothwhen the organism is grown in liquid culture. This raises the ques-tion of whether these secondary metabolites have further func-tions besides protection of the perithecia. One possibility wouldbe a functional redundancy between the red pigments, since fus-arubins and bikaverin are not produced under the same cultureconditions in F. fujikuroi. A similar situation was found for F.graminearum, where deletion of the aurofusarin-specific PKS gene

aur1 affected the expression of pgl1, which is responsible for theformation of pigmented perithecia (28).

Biosynthetic pathway of the fusarubins. We identified themain compounds produced by the wild type as the naphthoqui-nones 8-O-methylfusarubin (compound 1), 8-O-methylnectria-furone (compound 2), 8-O-methyl-13-hydroxynorjavanicin(compound 3), 8-O-methylanhydrofusarubinlactol (compound4), and 13-hydroxynorjavanicin (compound 5) (Fig. 6D) on thebasis of their exact masses, characteristic UV fingerprints, andNMR data (67, 70–72). All of these compounds, except for 8-O-methylanhydrofusarubinlactol (compound 4), which has neverbeen described before, are known metabolites of other Fusariumspp. (52). However, this is the first report of the identification ofthese naphthoquinones in F. fujikuroi. The finding that thesestructurally related but distinctly different naphthoquinones areall products of the same NR-PKS might indicate that most of thesestructures identified in Fusarium spp. are intermediates or finalproducts of one and the same biosynthetic pathway.

To analyze this pathway in detail, it is important to identify thefirst intermediate released by the responsible PKS. Fsr1 shows thetypical domain organization of a nonreducing PKS, consisting of aSAT domain, a KS domain, a MAT domain, and two ACP do-mains, with one difference: the often-found TE or TE/CLC do-main is replaced by an R domain (Fig. 2A). This finding led to thehypothesis that the PKS-derived heptaketide is most likely re-leased as an aldehyde. By overexpressing fsr1 in the �fsr2-5 mutantin an Fsr6-independent manner, we were able to identify the firstintermediate of the pathway as 6-O-demethylfusarubinaldehyde(compound 7) (Fig. 6C and D). This reductive release mechanism,resulting in the formation of an aldehyde intermediate, has beenshown previously for the 3-methylorcinaldehyde synthase Pks1 inAcremonium strictum (5) and AfoE, the NR-PKS of the asperfura-none pathway (13), which also harbor R domains. This mecha-nism has also been shown in bacteria, e.g., for Lgr, the nonribo-somal peptide synthase (NRPS) of the gramicidin pathway inBacillus brevus (38); for SmfC, one NRPS of the SFM-A gene clus-

FIG 7 The fusarubins are responsible for the coloration of the fruiting bodies in F. fujikuroi. Strains were crossed as described in Materials and Methods.Perithecia are indicated by black arrows. Size standards are shown in the lower left corners. (A) Crossing of the wild-type strain IMI58289 with its mating partnerC1995 results in the formation of darkly pigmented fruiting bodies. (B) Crossing of IMI58289/�fsr1 with C1995/�fsr1 results in the formation of colorlessperithecia that are larger than the pigmented perithecia.

Studt et al.

4476 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

ter in Streptomyces lavendulae (45); and for MxcG, the NRPS re-sponsible for the formation of the aldehyde intermediate in themyxochelin biosynthesis of Stigmatella aurantiaca (46). Further-more, the �-aminoadipate reductase Lys2p, which also harbors anR domain, has been shown to be responsible for the formation ofan aldehyde intermediate in the lysine-biosynthetic pathway ofSaccharomyces cerevisiae (25). In the biosynthetic pathway of fus-arubins, the A and B rings of compound 7 are most likely formedby a specific C-4/C-9-type aldol cyclization and aromatizationcatalyzed by the identified PT domain of Fsr1. This is in accor-dance with the mechanism that the PT domain of A. fumigatusPksA catalyzes during aflatoxin formation (15, 18, 47). Since phy-logenetic analysis of the Fsr1 PT domain shows close relatednessto PksA (see Fig. S5 in the supplemental material), this C-4/C-9-type is very likely to occur in F. fujikuroi, but experimentalproof is needed. Parisot et al. described a similar metabolite,6-O-demethylnectriachrysone, isolated from F. solani mutantscreated by random UV mutagenesis (57). This slightly differentstructure most likely results from rearrangement of the releasedaldehyde to a more stable intermediate during the product isola-tion procedure, which was different from our method. This find-ing indicates that 6-O-demethylfusarubinaldehyde (compound7), rather than the proposed fusarubinic acid (54), is also the firstintermediate in F. solani, as was shown for F. fujikuroi in thispaper. This is the first report of an NR-PKS harboring an R do-main showing C-4/C-9 cyclization.

Although a SAT domain was identified in Fsr1, and compound7 is most likely built by a C14 alkyl chain, it remains unclearwhether this alkyl chain consists of malonyl-derived buildingblocks only or whether the SAT domain accepts a starter unit witha different chain length. Investigation of the SAT domain of A.fumigatus PksA provides experimental proof that this SAT do-main accepts a hexanoyl starter unit (15). However, this hexanoylstarter is assembled by a fatty acid synthase (FAS) �-subunit(HexA) and a FAS �-subunit (HexB) that are encoded by genes inclose proximity to pksA in A. fumigatus (78). Other examples ofNR-PKSs harboring a SAT domain that are known to accept starterunits other than malonyl-CoA are found in F. graminearum trichoth-ecene and A. nidulans asperfuranone biosynthesis. In both cases, theNR-PKS accepts the product of a highly reducing PKS (HR-PKS)as a starter unit. Interestingly, the genes encoding the NR-PKS andthe HR-PKS belong to the same gene cluster (13, 39), a situationresembling that of the aflatoxin gene cluster in A. fumigatus. In F.fujikuroi, no HR-PKS- or FAS-encoding genes are found in closeproximity to the fusarubin gene cluster. However, the nature ofthe building blocks assembled by Fsr1 needs experimental proof.This would provide valuable information that might help answerthe question whether the gene involved in starter unit formationneeds to be located in the same gene cluster as the NR-PKS-en-coding gene in case the starter unit is not malonyl-CoA.

The presence of a tandem ACP motif in Fsr1 was also observedin A. nidulans WA, where the two ACP domains were shown to beredundant (26), suggesting a similar situation for Fsr1 in F. fuji-kuroi.

To identify further intermediates of this pathway, single dele-tion mutants of all six pathway genes were generated, and theresulting metabolites were analyzed by HPLC-DAD (Fig. 6B) andHPLC-UV-FTMS (data not shown). Of all the mutants, only the�fsr2 and �fsr3 mutants showed a metabolite spectrum alteredfrom that of the wild type. The structure of the accumulated in-

termediates was additionally elucidated by performing NMR ex-periments. By those experiments, 6-O-demethylfusarubinalde-hyde (compound 7) and 6-O-demethyl-10-deoxyfusarubin(compound 6) were identified in the fsr2 deletion mutant, whilefusarubinaldehyde (compound 8) was the main product of thefsr3 deletion mutant. These results confirmed the hypothesizedfunctions of those two enzymes (Table 1) as an O-methyltrans-ferase (Fsr2) and a monooxygenase (Fsr3), as previously proposedfor mutants created by UV mutagenesis in F. solani (56, 57).

The intermediates found in the �fsr2 and �fsr3 mutants andputative intermediates described in the literature indicate thatFsr2 and Fsr3 function independently of each other. Nevertheless,the fact that the second hydroxyl group is not present at positionC-10 in one of the �fsr2 intermediates (compound 6) indicatesthat the methylation of the hydroxyl group at position C-6 is nec-essary for this step. Furthermore, methylation at C-8 takes placeonly if both hydroxyl groups at C-5 and C-10 are present and doesnot seem to be an essential biosynthetic step, as indicated by theaccumulation of 13-hydroxynorjavanicin (compound 5) in thewild type. A similarly flexible pathway was shown for the biosyn-thesis of bikaverin, where the monomethylated form, nor-bikaverin, coexists with the dimethylated form, bikaverin, in liq-uid culture (79). The accumulation of only one intermediate inthe �fsr3 mutant, instead of the five structurally related but dis-tinct metabolites found in the wild type, suggests that Fsr3 is re-sponsible not only for the hydroxylation steps at positions C-5 andC-10 but also for the reduction of the released aldehyde as well asthe oxidation and subsequent 13-hydroxylation after the cleavageof CO2. This series of possible biosynthetic steps indicates the roleof Fsr3 as a multifunctional monooxygenase. So far, multiple ox-idization reactions by a single enzyme have been described onlyfor cytochrome P450 monooxygenases, including three of the fourP450 monooxygenases in the gibberellin biosynthesis of F. fujiku-roi (62, 76, 77) and Tri4 in the trichothecene biosynthesis of F.graminearum (73). In A. fumigatus, chain shortening of YWA1was shown to be catalyzed by Ayg1, resulting in the formation of1,3,6,8-tetrahydroxynaphthalene (T4HN) (27). Since no proteinwith significant homology to Ayg1 was shown to be encoded in thefsr gene cluster, a similar mechanism is very unlikely. However,whether Fsr3 is capable of multiple oxidation steps, including thecleavage of CO2, awaits experimental proof.

The proteins encoded by fsr4 and fsr5 show homologies to pro-teins harboring an oxidoreductase and a short-chain dehydrogenasedomain, respectively. Surprisingly, no enzymatic function could bedetermined for either of the two, because deletion of their encodinggenes did not alter the product spectrum (Fig. 6B). Remarkably, de-letion of fsr4, but not fsr5, seems to enhance the biosynthesis of theproducts (Fig. 6B), although the fsr1 to fsr3 genes were not upregu-lated in the fsr4 deletion mutant at the time point investigated (seeFig. S6A in the supplemental material). Proteins harboring a short-chain dehydrogenase domain were previously shown to possess reg-ulatory functions in filamentous fungi. Thus, the Nmr1 protein,which also has a predicted short-chain reductase domain, was linkedto the nitrogen-dependent regulation of secondary metabolism in F.fujikuroi. Yeast two-hybrid analyses showed that Nmr1 interacts withthe global nitrogen regulator AreA, thereby inactivating AreA undernitrogen-sufficient conditions, leading to a subsequent downregula-tion of genes responsible for gibberellin biosynthesis (65). Addition-ally, an nmr-like gene, bik4, was identified in the bikaverin gene clus-ter, and the encoded protein, Bik4, was suggested to be a coregulator

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4477

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

of the Zn(II)2Cys6 transcription factor Bik5 (79). However, sincefusarubin derivatives are produced in the perithecia under naturalconditions, the possibility that Fsr4 and Fsr5 catalyze specific enzy-matic steps during perithecium development cannot be ruled out.

Organization of the perithecial pigment gene cluster in F.fujikuroi and related species. The gene cluster for the perithecialpigment in F. solani differs from the gene clusters found in F.verticillioides, F. graminearum (60), and F. fujikuroi (this paper).Remarkably, a core gene cluster consisting of genes encoding thePKS (Pgl1, Fsr1), the methyltransferase (Omt1, Fsr2), and themonooxygenase (Fdm1, Fsr3) is syntenic in all four fusaria,whereas the entire genomic region from the FF_03983 to theFF_03990 gene, including the cluster genes fsr1 to fsr6, is syntenicin F. verticillioides only (60). In F. graminearum and F. solani, thesynteny between fsr3 and fsr4 is abrogated, suggesting that in thesespecies, the respective perithecial pigment gene cluster may con-sist of genes different from those in F. fujikuroi. This constitu-tional difference of the respective gene cluster in F. solani and F.fujikuroi could account for the different production conditionsand the slightly different metabolite spectrum for F. solani (de-scribed by Kurobane et al. [43]) versus F. fujikuroi (this study).

Interestingly, Ma et al., by using a microarray approach, de-scribed seven genes that were coregulated with pgl1 in F. verticil-lioides (49). However, in F. fujikuroi, only five of the homologousgenes, fsr2 to fsr6, adjacent to the PKS-encoding fsr1 gene werefound to be strictly coregulated (Fig. 4), possibly due to the differ-

ent culture conditions used. The restriction of the fusarubin genecluster in F. fujikuroi to the genes fsr1 to fsr6 is confirmed by thefact that deletion of the genes bordering fsr1 and fsr6 had no im-pact on biosynthesis, and these genes were not expressed underany of the conditions tested in this study (Fig. 4).

The two groups of red pigments in F. fujikuroi show con-trasting regulation. Interestingly, the two groups of pigments inF. fujikuroi, bikaverins and fusarubins, are found to be expressedunder contrasting pH conditions. The most remarkable differenceis that bikaverin is produced only under acidic conditions, whilethe fusarubin pigments are formed only when the pH of the cul-ture medium is neutral or alkaline (Fig. 1). Due to the strict re-sponse of fsr gene expression to changing pH conditions, the in-volvement of the pH regulator PacC is currently underinvestigation. In A. nidulans, this C2H2 zinc finger transcriptionfactor is known to preferentially activate target genes at an alkalinepH and to repress genes that are expressed at an acidic pH (58).For the bikaverin genes, noncanonical pH regulation has beendemonstrated: the genes are repressed under alkaline conditionsin a PacC-dependent manner (79).

Another interesting yet unknown mechanism for regulatingthe expression of bikaverin genes is their strict interdependency. Ifany of the biosynthesis genes bik1 to bik3 is deleted, the remaininggenes also are not expressed, due to a yet unknown mechanism(79). This is not the case for the fsr genes, where only the deletionof the PKS-encoding gene, fsr1, or the gene coding for a

FIG 8 Biosynthetic pathway of fusarubins in F. fujikuroi. Shown is the verified route of naphthoquinone formation in F. fujikuroi based on the main compoundsidentified in the liquid culture. The condensation of seven acetyl-CoA units results in the formation of a heptaketide (E) (bound to the ACP domain of theNR-PKS). After rings A and B are formed by aldol-type cyclization, the PKS-derived product is released as 6-O-demethylfusarubinaldehyde (compound 7). Thentwo hydroxyl groups are incorporated by Fsr3, and simultaneously hydroxyl groups are methylated by Fsr2. The aldehyde is, on the one hand, reduced by Fsr3to 8-O-methylfusarubin alcohol, which equilibrates mainly with 8-O-methylfusarubin (compound 1) and only small amounts of 8-O-methylnectriafurone(compound 2) (indicated by the gray reaction arrows). On the other hand, the aldehyde can be oxidized to form 8-O-methylfusarubinic acid, a reaction drivenby Fsr3 equilibrating with 8-O-methylfusarubinlactone, finally resulting in 8-O-methylanhydrofusarubinlactol (compound 4) after a further reduction step andloss of water. 8-O-Methylfusarubinic acid can also undergo decarboxylation, resulting in 8-O-methyl-13-hydroxynorjavanicin (compound 3) after anotherhydroxylation step at C-13. Both steps are most likely also accomplished by Fsr3. Structures highlighted in gray were identified by NMR, mass spectrometry, andUV data in the liquid culture of wild-type F. fujikuroi during the course of this work. The aldehyde identified as the first intermediate in this biosynthetic pathwayis boxed. The solid dark gray arrow indicates the main route of naphthoquinone formation.

Studt et al.

4478 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Zn(II)2Cys6 transcription factor, fsr6, results in a total loss of geneexpression; deletion of any other gene does not affect the expres-sion of the remaining fsr genes (see Fig. S6A in the supplementalmaterial). In addition to these contrasting regulation patterns,there is also a common regulation mechanism: large amounts ofnitrogen repress the expression of bikaverin (79) as well as fusaru-bin (described here) biosynthetic genes. We clearly demonstratedthat this repression is specifically triggered by glutamine (Fig. 5C).How glutamine sensing and subsequent signal transduction workin F. fujikuroi is under investigation.

In conclusion, we describe the identification of five new redpigments in F. fujikuroi that are produced under specific cul-ture conditions (alkaline pH; nitrogen limitation) differentfrom those for bikaverins. By a combination of HPLC, MS, andNMR analyses, we identified 8-O-methylfusarubin (compound1), 8-O-methylnectriafurone (compound 2), 8-O-methyl-13-hy-droxynorjavanicin (compound 3), 8-O-methylanhydrofusaru-binlactol (compound 4), and 13-hydroxynorjavanicin (com-pound 5) as true metabolites of the pigment pathway in F.fujikuroi (67, 70–72). Based on the newly identified structures, wewere able to establish a biosynthetic pathway for fusarubin-likemetabolites: the heptaketide backbone of the polyketide resultsfrom the condensation of seven acetyl-CoA subunits and is sub-sequently released by the PKS Fsr1 as 6-O-demethylfusarubinal-dehyde (compound 7). After several methylation and hydroxyla-tion steps by Fsr2 and Fsr3, respectively, the final productscompounds 1 to 5 accumulate in the liquid culture of wild-type F.fujikuroi, with 8-O-methylfusarubin (compound 1) as the mainproduct (Fig. 8).

In addition, we pinpointed these pigments for the first time to aconcrete gene cluster in the genus Fusarium and demonstrated thatthe fusarubins described here are the formerly unknown perithecialpigments.

ACKNOWLEDGMENTS

This work and the research fellowship of Lena Studt were supported byfunds of the Deutsche Forschungsgesellschaft (DFG), Graduiertenkolleg1409 (GRK1409, Germany).

We thank Barbara Howlett of the School of Botany, University ofMelbourne, for providing the pZPnat1 vector; J. F. Leslie, Kansas StateUniversity, for F. fujikuroi strain C1995; Dominik Wagner for construc-tion of the modified pRS426 vector; Katharina W. von Bargen for helpwith the acquisition of NMR data; and Kathleen Plamper for excellenttechnical support.

REFERENCES1. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local

alignment search tool. J. Mol. Biol. 215:403– 410.2. Arnstein HRV, Cook AH, Lacey MS. 1946. An antibacterial pigment

from Fusarium javanicum. Nature 157:333–334.3. Ausubel FM, et al. 1987. Current protocols in molecular biology. John

Wiley and Sons, New York, NY.4. Bachmann BO, Ravel J. 2009. Methods for in silico prediction of micro-

bial polyketide and nonribosomal peptide biosynthetic pathways fromDNA sequence data. Methods Enzymol. 458:181–217.

5. Bailey AM, et al. 2007. Characterisation of 3-methylorcinaldehyde syn-thase (MOS) in Acremonium strictum: first observation of a reductive re-lease mechanism during polyketide biosynthesis. Chem. Commun.(Camb.) 39:4053– 4055.

6. Bode HB, Bethe B, Höfs R, Zeeck A. 2002. Big effects from small changes:possible ways to explore nature’s chemical diversity. Chembiochem3:619 – 627.

7. Bömke C, Tudzynski B. 2009. Diversity, regulation, and evolution of the

gibberellin biosynthetic pathway in fungi compared to plants and bacteria.Phytochemistry 70:1876 –1893.

8. Brakhage AA, Schroeckh V. 2011. Fungal secondary metabolites—strategies to activate silent gene clusters. Fungal Genet. Biol. 48:15–22.

9. Brakhage AA, et al. 2008. Activation of fungal silent gene clusters: a newavenue to drug discovery. Prog. Drug Res. 66:1–12.

10. Brückner B, Blechschmidt D. 1991. The gibberellin fermentation. Crit.Rev. Biotechnol. 1:163–192.

11. Caracuel Z, Casanova C, Roncero MIG, Di Pietro A, Ramos J. 2003. pHresponse transcription factor PacC controls salt stress tolerance and ex-pression of the P-type Na�-ATPase Ena1 in Fusarium oxysporum. Eu-karyot. Cell 2:1246 –1252.

12. Cenis JL. 1992. Rapid extraction of fungal DNA for PCR amplification.Nucleic Acids Res. 20:2380.

13. Chiang YM, et al. 2009. A gene cluster containing two fungal polyketidesynthases encodes the biosynthetic pathway for a polyketide, asperfura-none, in Aspergillus nidulans. J. Am. Chem. Soc. 131:2965–2970.

14. Chiang Y-M, Oakley BR, Keller NP, Wang CCC. 2010. Unravelingpolyketide synthesis in members of the genus Aspergillus. Appl. Microbiol.Biotechnol. 86:1719 –1736.

15. Crawford JM, Dancy BCR, Hill EA, Udwary DW, Townsend CA. 2006.Identification of a starter unit acyl-carrier protein transacylase domain inan iterative type I polyketide synthase. Proc. Natl. Acad. Sci. U. S. A.103:16728 –16733.

16. Crawford JM, et al. 2008. Deconstruction of iterative multidomainpolyketide synthase function. Science 320:243–246.

17. Crawford JM, Vagstad AL, Ehrlich KC, Townsend CA. 2008. Starter unitspecificity directs genome mining of polyketide synthase pathways infungi. Bioorg. Chem. 36:16 –22.

18. Crawford JM, et al. 2009. Structural basis for biosynthetic programmingof fungal aromatic polyketide cyclization. Nature 461:1139 –1143.

19. Christianson TW, Sikorski RS, Dante M, Shero JH, Hieter P. 1992. Mul-tifunctional yeast high-copy-number shuttle vectors. Gene 110:119–122.

20. Colot HV, et al. 2006. A high-throughput gene knockout procedure forNeurospora reveals functions for multiple transcription factors. Proc. Natl.Acad. Sci. U. S. A. 103:10352–10357.

21. Cross BE, Myers PL, Webster GRB. 1970. New metabolites of Gib-berella fujikuroi. XIV. O-Demethyl-anhydrofusarubin. J. Chem. Soc.7:930.

22. Darken MA, Jensen AAL, Shu P. 1959. Production of gibberellic acid byfermentation. Appl. Microbiol. 7:301–303.

23. Dereeper A, et al. 2008. Phylogeny.fr: robust phylogenetic analysis for thenon-specialist. Nucleic Acids Res. 36:465– 469.

24. Du L, Lou L. 2010. PKS and NRPS release mechanisms. Nat. Prod. Rep.27:255–278.

25. Ehmann DE, Gehring AM, Walsh CT. 1999. Lysine biosynthesis inSaccharomyces cerevisiae: mechanism of �-aminoadipate reductase (Lys2)involves posttranslational phosphopantetheinylation by Lys5. Biochemis-try 38:6171– 6177.

26. Fujii I, Watanabe A, Sankawa U, Ebizuka Y. 2001. Identification ofClaisen cyclase domain in fungal polyketide synthase WA, a naphthopy-rone synthase of Aspergillus nidulans. Chem. Biol. 8:189 –197.

27. Fujii I, et al. 2004. Hydrolytic polyketide shortening by Ayg1p, a novelenzyme involved in fungal melanin biosynthesis. J. Biol. Chem. 279:44613– 44620.

28. Gaffoor I, et al. 2005. Functional analysis of the polyketide synthase genesin the filamentous fungus Gibberella zeae (anamorph Fusariumgraminearum). Eukaryot. Cell 4:1926 –1933.

29. Geissman TA, Verbiscar AJ, Phinney BO, Cragg G. 1966. Studies on thebiosynthesis of gibberellins from (�)-kaurenoic acid in cultures of Gib-berella fujikuroi. Phytochemistry 5:933–947.

30. Graziani S, Vasnier C, Daboussi MJ. 2004. Novel polyketide synthasefrom Nectria haematococca. Appl. Environ. Microbiol. 70:2984 –2988.

31. Gritz L, Davies LJ. 1983. Plasmid-encoded hygromycin B resistance: thesequence of hygromycin B phosphotransferase gene and its expression inEscherichia coli and Saccharomyces cerevisiae. Gene 25:179 –188.

32. Hoffmeister D, Keller NP. 2007. Natural products of filamentous fungi:enzymes, genes, and their regulation. Nat. Prod. Rep. 24:393– 416.

33. Hori S. 1890. Injurious fungi attacking rice plants. Bot. Mag. 4:425– 427.34. Ito S, Kimura J. 1931. Studies on the ‘bakanae’ disease of rice plant. Rep.

Hokkaido Nat. Agric. Exp. Stn. 27:1–99.35. Jahn B, Langfelder K, Schneider U, Schindel C, Brakhage AA. 2002.

PKSP-dependent reduction of phagolysosome fusion and intracellular kill

Fusarubin Biosynthesis and F. fujikuroi Perithecia

June 2012 Volume 78 Number 12 aem.asm.org 4479

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

of Aspergillus fumigatus conidia by human monocyte-derived macro-phages. Cell. Microbiol. 4:793– 803.

36. Kavanagh KL, Jornvall H, Persson B, Oppermann U. 2008. The SDRsuperfamily: functional and structural diversity within a family of meta-bolic and regulatory enzymes. Cell. Mol. Life Sci. 65:3895–3906.

37. Keller NP, Turner G, Bennett JW. 2005. Fungal secondary metabolism—from biochemistry to genomics. Nat. Rev. Microbiol. 3:937–947.

38. Kessler N, Schumann H, Morneweg S, Linne U, Marahiel MA. 2004.The linear pentadecapeptide gramicidin is assembled by four multimodu-lar nonribosomal peptide synthetases that comprise 16 modules with 56catalytic domains. J. Biol. Chem. 279:7413–7419.

39. Kim J-E, et al. 2005. Putative polyketide synthase and laccase genes forbiosynthesis of aurofusarin in Gibberella zeae. Appl. Environ. Microbiol.71:1701–1708.

40. Kjær D, Kjær A, Pedersen C, Bu’lock JD, Smith JR. 1971. Bikaverin andnorbikaverin, benzoxanthentrione pigments of Gibberella fujikuroi. J.Chem. Soc. Perkin 1 16:2792–2797.

41. Klittich CJR, Leslie JF. 1988. Nitrate reduction mutants of Fusariummoniliforme (Gibberella fujikuroi). Genetics 118:417– 423.

42. Korman TP, et al. 2010. Structure and function of an iterative polyketidesynthase thioesterase domain catalyzing Claisen cyclization in aflatoxinbiosynthesis. Proc. Natl. Acad. Sci. U. S. A. 107:6246 – 6251.

43. Kurobane I, Vining LC, McInnes AG, Smith DG. 1978. Diastereoiso-meric 4a,10a-dihydrofusarubins: true metabolites of Fusarium solani.Can. J. Chem. 56:1593–1594.

44. Kurosawa E. 1926. Experimental studies on the nature of the substanceexcreted by the “bakanae” fungus. Trans. Nat. Hist. Soc. Formosa 16:213–227.

45. Li L, et al. 2008. Characterization of the saframycin A gene clusterfrom Streptomyces lavendulae NRRL 11002 revealing a nonribosomalpeptide synthetase system for assembling the unusual tetrapeptidylskeleton in an iterative manner. J. Bacteriol. 190:251–263.

46. Li Y, Weissman KJ, Müller R. 2008. Myxochelin biosynthesis: directevidence for two- and four-electron reduction of a carrier protein-boundthioester. J. Am. Chem. Soc. 130:7554 –7555.

47. Li Y, Xu W, Tang Y. 2010. Classification, prediction, and verification ofthe regioselectivity of fungal polyketide synthase product template do-mains. J. Biol. Chem. 285:22764 –22773.

48. Linnemannstöns P, et al. 2002. The polyketide synthase gene pks4 fromGibberella fujikuroi encodes a key enzyme in the biosynthesis of the redpigment bikaverin. Fungal Genet. Biol. 37:134 –148.

49. Ma L-J, et al. 2010. Comparative genomics reveals mobile pathogenicitychromosomes in Fusarium. Nature 464:367–373.

50. Malonek S, et al. 2004. Functional characterization of two cytochromeP450 monooxygenase genes, P450-1 and P450-4, of the gibberellic acidgene cluster in Fusarium proliferatum (Gibberella fujikuroi MP-D). J. Biol.Chem. 279:25075–25084.

51. Malz S, et al. 2005. Identification of a gene cluster responsible for thebiosynthesis of aurofusarin in the Fusarium graminearum species com-plex. Fungal Genet. Biol. 42:420 – 433.

52. Medentsev AG, Akimenko VK. 1998. Naphthoquinone metabolites ofthe Fungi. Phytochemistry 47:935–959.

53. Palmer JM, Keller NP. 2010. Secondary metabolism in fungi: does chro-mosomal location matter? Curr. Opin. Microbiol. 13:431– 436.

54. Parisot D, Devys M, Barbier M. 1988. Fusarubinic acid, a new naphtho-quinone from the fungus Nectria haematococca. Phytochemistry 27:3002–3004.

55. Parisot D, Devys M, Barbier M. 1991. 6-O-Demethyl-5-deoxyfusarubinand its anhydro derivative produced by a mutant of the fungus Nectriahaematococca blocked in fusarubin biosynthesis. J. Antibiot. 44:103–107.

56. Parisot D, Devys M, Barbier M. 1991. Nectriachrysone, a new metaboliterelated to fusarubin produced by the fungus Nectria haematococca. J.Chem. Soc. Perkin 1 1991:2280 –2281.

57. Parisot D, Devys M, Barbier M. 1992. Heptaketide-derived polyenesfrom the fungus Nectria haematococca. Phytochemistry 31:4357– 4358.

58. Peñalva MA, Tilburn J, Bignell E, Arst HN, Jr. 2008. Ambient pHgene regulation in fungi: making connections. Trends Microbiol. 16:291–300.

59. Pihet M, et al. 2009. Melanin is an essential component to the integrity ofthe cell wall of Aspergillus fumigatus conidia. BMC Microbiol. 9:177. doi:10.1186/1471-2180-9-177.

60. Proctor RH, Butchko RAE, Brown DW, Moretti A. 2007. Functional

characterization, sequence comparisons and distribution of a polyketidesynthase gene required for perithecial pigmentation in some Fusariumspecies. Food Addit. Contam. 24:1076 –1087.

61. Rohlfs M, Albert M, Keller NP, Kempken F. 2007. Secondary chemicalsprotect mould from fungivory. Biol. Lett. 3:523–525.

62. Rojas MC, Hedden P, Gaskin P, Tudzynksi B. 2001. The P450-1 gene ofGibberella fujikuroi encodes a multifunctional enzyme in gibberellin bio-synthesis. Proc. Natl. Acad. Sci. U. S. A. 98:5838 –5843.

63. Ruelius HW, Gauhe A. 1950. Uber Fusarubin, einem Naphthochinon-farbstoff aus Fusarien. Justus Liebigs Ann. Chem. 569:38 –59.

64. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a labora-tory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

65. Schönig B, Brown DW, Oeser B, Tudzynski B. 2008. Cross-specieshybridization with Fusarium verticillioides microarrays reveals new in-sights in Fusarium fujikuroi nitrogen regulation and the role of AreA andNMR. Eukaryot. Cell 7:1831–1846.

66. Staben C, et al. 1989. Use of a bacterial hygromycin B resistance gene as adominant selectable marker in Neurospora crassa transformation. FungalGenet. Newsl. 36:79 – 81.

67. Steyn PS, Wessels PL, Marasas FO. 1979. Pigments from Fusariummoniliforme Sheldon. Structure and 13C nuclear magnetic resonance as-signments of an aza-anthraquinone and three naphthoquinones. Tetra-hedron 35:1551–1555.

68. Strauss J, Reyes-Dominguez Y. 2011. Regulation of secondary metabo-lism by chromatin structure and epigenetic codes. Fungal Genet. Biol.48:62– 69.

69. Syn S, Snyder WC. 1981. The bakanae disease of the rice plant, p104 –113. In Nelson PE, Toussoun TA, Cook RJ (ed), Fusarium: dis-eases, biology and taxonomy. Pennsylvania State University Press,University Park, PA.

70. Tatum JH, Baker RA, Berry RE. 1985. Naphthoquinones produced byFusarium oxysporum isolated from citrus. Phytochemistry 24:457– 459.

71. Tatum JH, Baker RA, Berry RE. 1987. Naphthoquinones and derivativesfrom Fusarium. Phytochemistry 26:795–798.

72. Tatum JH, Baker RA, Berry RE. 1987. Naphthofurans produced byFusarium oxysporum isolated from citrus. Phytochemistry 26:2499 –2500.

73. Tokai T, et al. 2007. Fusarium Tri4 encodes a key multifunctional cyto-chrome P450 monooxygenase for four consecutive oxygenation steps intrichothecene biosynthesis. Biochem. Biophys. Res. Commun. 353:412–417.

74. Tudzynski B, Mende K, WeItring K-M, Kinghorn JR, Unkles SE. 1996.The Gibberella fujikuroi niaD gene encoding nitrate reductase: isolation,sequence, homologous transformation and electrophoretic karyotype lo-cation. Microbiology 142:533–539.

75. Tudzynski B, Homann V, Feng B, Marzluf GA. 1999. Isolation, charac-terization and disruption of the areA nitrogen regulatory gene of Gibber-ella fujikuroi. Mol. Gen. Genet. 261:106 –114.

76. Tudzynski B, Hedden P, Carrera E, Gaskin P. 2001. The P450-4 gene ofGibberella fujikuroi encodes ent-kaurene oxidase in the gibberellin biosyn-thesis pathway. Appl. Environ. Microbiol. 67:3514 –3522.

77. Tudzynski B, Rojas MC, Gaskin P, Hedden P. 2002. The Gibberellafujikuroi gibberellin 20-oxidase is a multifunctional monooxygenase. J.Biol. Chem. 277:21246 –21253.

78. Watanabe CMH, Wilson D, Linz JE, Townsend CA. 1996. Demonstra-tion of the catalytic roles and evidence for the physical association of typeI fatty acid synthases and a polyketide synthase in the biosynthesis ofaflatoxin B1. Chem. Biol. 3:463– 469.

79. Wiemann P, et al. 2009. Biosynthesis of the red pigment bikaverin inFusarium fujikuroi: genes, their function and regulation. Mol. Microbiol.72:931–946.

80. Wiemann P, et al. 2010. FfVel1 and FfLae1, components of a Velvet-likecomplex in Fusarium fujikuroi, affect differentiation, secondary metabo-lism and virulence. Mol. Microbiol. 77:972–994.

81. Williams RB, Henrikson JC, Hoover AR, Lee AE, Cichewicz RH. 2008.Epigenetic remodeling of the fungal secondary metabolome. Org. Biomol.Chem. 6:1895–1897.

82. Winston F, Dollard C, Ricupero-Hovasse SL. 1995. Construction of a set ofconvenient Saccharomyces cerevisiae strains that are isogenic to S288C. Yeast11:53–55.

Studt et al.

4480 aem.asm.org Applied and Environmental Microbiology

on July 6, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from