Genetic bases for variation in structure and biological activity of trichothecene toxins produced by diverse fungi.

Trichothecenes are sesquiterpene toxins produced by diverse but relatively few fungal species in at least three classes of Ascomycetes: Dothideomycetes, Eurotiomycetes, and Sordariomycetes. Approximately 200 structurally distinct trichothecene analogs have been described, but a given fungal species typically produces only a small subset of analogs. All trichothecenes share a core structure consisting of a four-ring nucleus known as 12,13-epoxytrichothec-9-ene. This structure can be substituted at various positions with hydroxyl, acyl, or keto groups to give rise to the diversity of trichothecene structures that has been described. Over the last 30 years, the genetic and biochemical pathways required for trichothecene biosynthesis in several species of the fungi Fusarium and Trichoderma have been elucidated. In addition, phylogenetic and functional analyses of trichothecene biosynthetic (TRI) genes from fungi in multiple genera have provided insights into how acquisition, loss, and changes in functions of TRI genes have given rise to the diversity of trichothecene structures. These analyses also suggest both divergence and convergence of TRI gene function during the evolutionary history of trichothecene biosynthesis. What has driven trichothecene structural diversification remains an unanswered question. However, insight into the role of trichothecenes in plant pathogenesis of Fusarium species and into plant glucosyltransferases that detoxify the toxins by glycosylating them point to a possible driver. Because the glucosyltransferases can have substrate specificity, changes in trichothecene structures produced by a fungus could allow it to evade detoxification by the plant enzymes. Thus, it is possible that advantages conferred by evading detoxification have contributed to trichothecene structural diversification. KEY POINTS : • TRI genes have evolved by diverse processes: loss, acquisition and changes in function. • Some TRI genes have acquired the same function by convergent evolution. • Some other TRI genes have evolved divergently to have different functions. • Some TRI genes were acquired or resulted from diversification in function of other genes. • Substrate specificity of plant glucosyltransferases could drive trichothecene diversity.

Pathogenicity of Fumonisin-producing and Nonproducing Strains of Aspergillus Species in Section Nigri to Maize Ears and Seedlings.

Species of Aspergillus section Nigri are commonly associated with maize kernels, and some strains can produce fumonisin mycotoxins. However, there is little information about the extent to which these fungi contribute to fumonisin contamination in grain, the damage they cause to maize ears, or their effects on maize seed germination and seedling health. We compared fumonisin-producing and nonproducing strains of A. niger, A. welwitschiae, A. phoenicis, A. tubingensis, and A. carbonarius from the United States and Italy in laboratory and field studies to assess their ability to contribute to fumonisin contamination, to cause maize ear rot, and to affect seed germination and seedling growth. In laboratory experiments, some strains of each Aspergillus species reduced germination or seedling growth, but there was high variability among strains within species. There were no consistent differences between fumonisin-producing and nonproducing strains. In field studies in Iowa and Illinois, strains were variable in their ability to cause ear rot symptoms, but this was independent of the ability of the Aspergillus strains to produce fumonisins. Contamination of grain with fumonisins was not consistently increased by inoculation with Aspergillus strains compared with the control, and was much greater in F. verticillioides-inoculated treatments than in Aspergillus-inoculated treatments. However, the ratio of the FB analogs FB2 and FB1 was altered by inoculation with some Aspergillus strains, indicating that FB2 production by Aspergillus strains occurred in the field. These results demonstrate the pathogenic capabilities of strains of Aspergillus in section Nigri, but suggest that their effects on maize ears and seedlings are not related to their ability to produce fumonisins, and that fumonisin contamination of grain caused by Aspergillus spp. is not as significant as that caused by Fusarium spp.

Molecular biodiversity of mycotoxigenic fungi that threaten food safety.

Fungal biodiversity is one of the most important contributors to the occurrence and severity of mycotoxin contamination of crop plants. Phenotypic and metabolic plasticity has enabled mycotoxigenic fungi to colonize a broad range of agriculturally important crops and to adapt to a range of environmental conditions. New mycotoxin-commodity combinations provide evidence for the ability of fungi to adapt to changing conditions and the emergence of genotypes that confer enhanced aggressiveness toward plants and/or altered mycotoxin production profiles. Perhaps the most important contributor to qualitative differences in mycotoxin production among fungi is variation in mycotoxin biosynthetic genes. Molecular genetic and biochemical analyses of toxigenic fungi have elucidated specific differences in biosynthetic genes that are responsible for intra- and inter-specific differences in mycotoxin production. For Aspergillus and Fusarium, the mycotoxigenic genera of greatest concern, variation in biosynthetic genes responsible for production of individual families of mycotoxins appears to be the result of evolutionary adaptation. Examples of such variation have been reported for: a) aflatoxin biosynthetic genes in Aspergillus flavus and Aspergillus parasiticus; b) trichothecene biosynthetic genes within and among Fusarium species; and c) fumonisin biosynthetic genes in Aspergillus and Fusarium species. Understanding the variation in these biosynthetic genes and the basis for variation in mycotoxin production is important for accurate assessment of the risks that fungi pose to food safety and for prevention of mycotoxin contamination of crops in the field and in storage.

LC-MS/MS method for the determination of the fungal pigment bikaverin in maize kernels as an indicator of ear rot.

Bikaverin is a polyketide-derived pigment produced by multiple species of the fungus Fusarium, some of which can cause ear and kernel rot of maize. A method was developed for the analysis of bikaverin by high-performance liquid chromatography (LC) coupled to electrospray ionisation tandem mass spectrometry (MS/MS). The quantitative nature of the LC-MS/MS method was demonstrated over a range of concentrations of bikaverin in maize. For spike-recovery experiments utilising maize spiked with bikaverin to a level 5 µg g⁻¹ of maize, the measured recovery (%) was 70.6 ± 10.4. Based on the utilised method, the limit of detection (based on a signal-to-noise ratio (S/N) > 3) was better than 0.5 µg g⁻¹ from bikaverin spiked into uncontaminated ground maize. Further, the limit of quantitation (LOQ) was 3 µg g⁻¹ (based on S/N > 10) from bikaverin spiked into ground maize. The method was applied to assess contamination of maize with bikaverin following inoculation of developing maize ears with Fusarium verticillioides under agricultural field conditions.

Analysis of fumonisin contamination and the presence of Fusarium in wheat with kernel black point disease in the United States.

The ability of the fungus Fusarium proliferatum to cause kernel black point disease in wheat was previously established, but natural contamination of black point wheat with both F. proliferatum and fumonisin mycotoxins has not been studied in the United States. Low levels of fumonisins were detected in nine of 43 wheat samples with kernel black point disease that were obtained from across the United States. A subset of samples was contaminated with F. proliferatum as well as with F. fujikuroi, F. nygamai, F. thapsinum and F. verticillioides, species closely related to F. proliferatum and morphologically similar to it in that they produce chains of asexual spores, or conidia. Nevertheless, of conidial chain-forming fusaria isolated from symptomatic wheat, F. proliferatum dominated. In greenhouse tests, isolates of F. proliferatum and the other species recovered from wheat samples were able to cause symptoms of kernel black point and, in some cases, low levels of fumonisin contamination of wheat. These data add to the understanding of the risk of fumonisin contamination of wheat and the potential for Fusarium species to cause kernel black point disease and fumonisin contamination of wheat. Further, the results of this study indicate that while US-grown wheat can sporadically be contaminated by fumonisins, the natural contamination levels seem to be low. The observations made provide evidence that fumonisins are not likely to be a factor contributing to the ability of Fusarium to cause kernel black point disease.

Identification of loci and functional characterization of trichothecene biosynthesis genes in filamentous fungi of the genus Trichoderma.

Trichothecenes are mycotoxins produced by Trichoderma, Fusarium, and at least four other genera in the fungal order Hypocreales. Fusarium has a trichothecene biosynthetic gene (TRI) cluster that encodes transport and regulatory proteins as well as most enzymes required for the formation of the mycotoxins. However, little is known about trichothecene biosynthesis in the other genera. Here, we identify and characterize TRI gene orthologues (tri) in Trichoderma arundinaceum and Trichoderma brevicompactum. Our results indicate that both Trichoderma species have a tri cluster that consists of orthologues of seven genes present in the Fusarium TRI cluster. Organization of genes in the cluster is the same in the two Trichoderma species but differs from the organization in Fusarium. Sequence and functional analysis revealed that the gene (tri5) responsible for the first committed step in trichothecene biosynthesis is located outside the cluster in both Trichoderma species rather than inside the cluster as it is in Fusarium. Heterologous expression analysis revealed that two T. arundinaceum cluster genes (tri4 and tri11) differ in function from their Fusarium orthologues. The Tatri4-encoded enzyme catalyzes only three of the four oxygenation reactions catalyzed by the orthologous enzyme in Fusarium. The Tatri11-encoded enzyme catalyzes a completely different reaction (trichothecene C-4 hydroxylation) than the Fusarium orthologue (trichothecene C-15 hydroxylation). The results of this study indicate that although some characteristics of the tri/TRI cluster have been conserved during evolution of Trichoderma and Fusarium, the cluster has undergone marked changes, including gene loss and/or gain, gene rearrangement, and divergence of gene function.

Genomic analysis of Fusarium verticillioides.

Fusarium verticillioides (teleomorph Gibberella moniliformis) can be either an endophyte of maize, causing no visible disease, or a pathogen-causing disease of ears, stalks, roots and seedlings. At any stage, this fungus can synthesize fumonisins, a family of mycotoxins structurally similar to the sphingolipid sphinganine. Ingestion of fumonisin-contaminated maize has been associated with a number of animal diseases, including cancer in rodents, and exposure has been correlated with human oesophageal cancer in some regions of the world, and some evidence suggests that fumonisins are a risk factor for neural tube defects. A primary goal of the authors' laboratory is to eliminate fumonisin contamination of maize and maize products. Understanding how and why these toxins are made and the F. verticillioides-maize disease process will allow one to develop novel strategies to limit tissue destruction (rot) and fumonisin production. To meet this goal, genomic sequence data, expressed sequence tags (ESTs) and microarrays are being used to identify F. verticillioides genes involved in the biosynthesis of toxins and plant pathogenesis. This paper describes the current status of F. verticillioides genomic resources and three approaches being used to mine microarray data from a wild-type strain cultured in liquid fumonisin production medium for 12, 24, 48, 72, 96 and 120h. Taken together, these approaches demonstrate the power of microarray technology to provide information on different biological processes.

Functional characterization, sequence comparisons and distribution of a polyketide synthase gene required for perithecial pigmentation in some Fusarium species.

Polyketides are a structurally diverse class of secondary metabolites produced by bacteria, fungi, plants and animals. The fungal genus Fusarium includes agronomically important plant pathogenic and mycotoxin-producing species and produces numerous polyketides. The study further characterized a polyketide synthase-encoding gene (PKS3 = PGL1) that was previously identified in F. graminearum and F. verticillioides. Disruption of the F. verticillioides PGL1 indicated that it is required for the production of the dark pigment in perithecial walls, as previously shown in F. graminearum. A third PGL1 orthologue was identified in the genomic sequence of N. haematococca (anamorph F. solani f. sp. pisi). Analysis of the carboxy-terminal end of the deduced PGL1 protein indicated that it had a functional domain related to dehydrogenases/reductases that is sometimes present in non-ribosomal peptide synthetases. Comparison of the genomic regions flanking PGL1 in F. graminearum, F. verticillioides and N. haematococca revealed that the extent of gene synteny in this region was greater between F. graminearum and F. verticillioides than between either of these species and N. haematococca. Southern blot analysis indicated that PGL1 occurs widely within the genus Fusarium including species with no known sexual stage.

Wheat kernel black point and fumonisin contamination by Fusarium proliferatum.

Fusarium proliferatum is a major cause of maize ear rot and fumonisin contamination and also can cause wheat kernel black point disease. The primary objective of this study was to determine whether nine F. proliferatum strains from wheat from Nepal can cause black point and fumonisin contamination in wheat kernels. For comparison, the study included three Fusarium strains from US maize. In test 1, all the strains but one produced significant symptoms of kernel black point; two strains decreased kernel yield; and four strains contaminated kernels with fumonisins B(1), B(2) and B(3) as determined by liquid chromatography-mass spectroscopy. Strain Ggm202 from Nepal, which produced the highest levels of fumonisins (mean = 49 microg g(-1)) on five wheat cultivars in test 1, was confirmed to produce fumonisins (mean = 38 microg g(-1)) on two cultivars in test 2. The data indicate a potential for fumonisin contamination of wheat infected with F. proliferatum.

Molecular biology of Fusarium mycotoxins.

As the 20th century ended, Fusarium mycotoxicology entered the age of genomics. With complete genomes of Fusarium graminearum and F. verticillioides and several Fusarium gene expression sequence databases on hand, researchers worldwide are working at a rapid pace to identify mycotoxin biosynthetic and regulatory genes. Seven classes of mycotoxin biosynthetic genes or gene clusters have been identified in Fusarium to date; four are polyketide synthase gene clusters for equisetin, fumonisins, fusarins, and zearalenones. Other Fusarium mycotoxin biosynthetic genes include a terpene cyclase gene cluster for trichothecenes, a cyclic peptide synthetase for enniatins, and a cytochrome P450 for butenolide. From the perspective of the United States Department of Agriculture, the ultimate goal of research on Fusarium molecular biology is to reduce mycotoxins in cereal grains. With this goal in mind, efforts have focused on identifying aspects of mycotoxin biosynthesis and regulation that can be exploited for mycotoxin control. New information on fungal and plant genomes and gene expression will continue to provide information on genes important for fungal-plant interactions and to facilitate the development of targeted approaches for breeding and engineering crops for resistance to Fusarium infection and mycotoxin contamination.

Biosynthesis ofFusarium mycotoxins and genomics ofFusarium verticillioides.

Analyses of mycotoxin biosynthetic genes inFusarium indicate that interspecies variation in trichothecene structure can result from differences in gene function and interspecies variation in fumonisin production/non-production can result from differences in the presence/absence of genes. Such variation is not always correlated with phylogenetic relationships of species as determined by sequencing primary metabolic genes; distantly related species can share the same mycotoxin biosynthetic genotype and resulting phenotype, while more closely related species can differ. These findings provide further evidence that the evolution of mycotoxin biosynthesis inFusarium has not always been congruent with the evolution of species.

Deletion and complementation of the mating type (MAT) locus of the wheat head blight pathogen Gibberella zeae.

Gibberella zeae, a self-fertile, haploid filamentous ascomycete, causes serious epidemics of wheat (Triticum aestivum) head blight worldwide and contaminates grain with trichothecene mycotoxins. Anecdotal evidence dating back to the late 19th century indicates that G. zeae ascospores (sexual spores) are a more important inoculum source than are macroconidia (asexual spores), although the fungus can produce both during wheat head blight epidemics. To develop fungal strains to test this hypothesis, the entire mating type (MAT1) locus was deleted from a self-fertile (MAT1-1/MAT1-2), virulent, trichothecene-producing wild-type strain of G. zeae. The resulting MAT deletion (mat1-1/mat1-2) strains were unable to produce perithecia or ascospores and appeared to be unable to mate with the fertile strain from which they were derived. Complementation of a MAT deletion strain by transformation with a copy of the entire MAT locus resulted in recovery of production of perithecia and ascospores. MAT deletion strains and MAT-complemented strains retained the ability to produce macroconidia that could cause head blight, as assessed by direct injection into wheat heads in greenhouse tests. Availability of MAT-null and MAT-complemented strains provides a means to determine the importance of ascospores in the biology of G. zeae and perhaps to identify novel approaches to control wheat head blight.

FUM1--a gene required for fumonisin biosynthesis but not for maize ear rot and ear infection by Gibberella moniliformis in field tests.

We have analyzed the role of fumonisins in infection of maize (Zea mays) by Gibberella moniliformis (anamorph Fusarium verticillioides) in field tests in Illinois and Iowa, United States. Fumonisin-nonproducing mutants were obtained by disrupting FUM1 (previously FUM5), the gene encoding a polyketide synthase required for fumonisin biosynthesis. Maize ear rot, ear infection, and fumonisin contamination were assessed by silk-channel injection in 1999 and 2000 and also by spray application onto maize silks, injection into maize stalks, and application with maize seeds at planting in 1999. Ear rot was evaluated by visual assessment of whole ears and by calculating percentage of symptomatic kernels by weight. Fumonisin levels in kernels were determined by high-performance liquid chromatography. The presence of applied strains in kernels was determined by analysis of recovered isolates for genetic markers and fumonisin production. Two independent fumonisin-nonproducing (fum1-3 and fum1-4) mutants were similar to their respective fumonisin-producing (FUM1-1) progenitor strains in ability to cause ear rot following silk-channel injection and also were similar in ability to infect maize ears following application by all four methods tested. This evidence confirms that fumonisins are not required for G. moniliformis to cause maize ear rot and ear infection.

Characterization of four clustered and coregulated genes associated with fumonisin biosynthesis in Fusarium verticillioides.

Fumonisins are mycotoxins that cause several fatal animal diseases, including cancer in rats and mice. These toxins are produced by several Fusarium species, including the maize pathogen Fusarium verticillioides, and can accumulate in maize infected with the fungus. We have identified four F. verticillioides genes (FUM6, FUM7, FUM8, and FUM9) adjacent to FUM5, a previously identified polyketide synthase gene that is required for fumonisin biosynthesis. Gene disruption analysis revealed that FUM6 and FUM8 are required for fumonisin production and Northern blot analysis revealed that expression of all four recently identified genes is correlated with fumonisin production. Nucleotide sequence analysis indicated that the predicted FUM6 translation product is most similar to cytochrome P450 monooxygenase-P450 reductase fusion proteins and the predicted products of FUM7, FUM8, and FUM9 are most similar to type III alcohol dehydrogenases, class-II alpha-aminotransferases, and dioxygenases, respectively. Together, these data are consistent with FUM5 through FUM9 being part of a fumonisin biosynthetic gene cluster in F. verticillioides.

A genetic and biochemical approach to study trichothecene diversity in Fusarium sporotrichioides and Fusarium graminearum.

The trichothecenes T-2 toxin and deoxynivalenol (DON) are natural fungal products that are toxic to both animals and plants. Their importance in the pathogenicity of Fusarium spp. on crop plants has inspired efforts to understand the genetic and biochemical mechanisms leading to trichothecene synthesis. In order to better understand T-2 toxin biosynthesis by Fusarium sporotrichioides and DON biosynthesis by F. graminearum, we compared the nucleotide sequence of the 23-kb core trichothecene gene cluster from each organism. This comparative genetic analysis allowed us to predict proteins encoded by two trichothecene genes, TRI9 and TRI10, that had not previously been described from either Fusarium species. Differences in gene structure also were correlated with differences in the types of trichothecenes that the two species produce. Gene disruption experiments showed that F. sporotrichioides TRI7 (FsTRI7) is required for acetylation of the oxygen on C-4 of T-2 toxin. Sequence analysis indicated that F. graminearum TRI7 (FgTRI7) is nonfunctional. This is consistent with the fact that the FgTRI7 product is not required for DON synthesis in F. graminearum because C-4 is not oxygenated.

A polyketide synthase gene required for biosynthesis of fumonisin mycotoxins in Gibberella fujikuroi mating population A.

Fumonisins are toxins associated with several mycotoxicoses and are produced by the maize pathogen Gibberella fujikuroi mating population A (MP-A). Biochemical analyses indicate that fumonisins are a product of either polyketide or fatty acid biosynthesis. To isolate a putative polyketide synthase (PKS) gene involved in fumonisin biosynthesis, we employed PCR with degenerate PKS primers and a cDNA template prepared from a fumonisin-producing culture of G. fujikuroi. Sequence analysis of the single PCR product and its flanking DNA revealed a gene (FUM5) with a 7.8-kb coding region. The predicted FUM5 translation product was highly similar to bacterial and fungal Type I PKSs. Transformation of a cosmid clone carrying FUM5 into G. fujikuroi enhanced production in three strains and restored wild-type production in a fumonisin nonproducing mutant. Disruption of FUM5 reduced fumonisin production by over 99% in G. fujikuroi MP-A. Together, these results indicate that FUM5 is a PKS gene required for fumonisin biosynthesis.

Characterization of a transcriptional activator controlling trichothecene toxin biosynthesis.

Trichothecene biosynthetic pathway genes are localized within a gene cluster in Fusarium sporotrichioides and require the zinc-finger containing protein, TRI6, for expression. We show here that TRI6 is able to bind within the promoter regions of nine different pathway genes and that TRI6 binding is involved in pathway gene activation. TRI6 binding occurs at three distinct sites in the TRI5 promoter, all of which contain the sequence TNAGGCCT. DNA fragments from the promoter regions of six other pathway genes containing this sequence are also substrates for TRI6 binding. Specific nucleotide changes in the TNAGGCCT sequence dramatically reduced TRI6 binding. Analysis of TRI6 binding within the TRI3 and TRI11 promoters and the TRI4-TRI6 intergenic region which do not contain the TNAGGCCT motif suggests that the minimum sequence required for TRI6 binding is YNAGGCC. Two potential TRI6 binding sites, T4A and T4B, were identified within the intergenic region for the divergently transcribed TRI4 and TRI6 genes. Alteration or deletion of the T4A site resulted in the loss of nearly all in vitro TRI6 binding and was correlated with the loss of promoter activity in vivo as measured by the expression of mutant TRI4(p)/GUS fusions. This establishes a physiological role for TRI6 binding and demonstrates that TRI6 is directly involved in the regulation of pathway gene expression. To determine if a predicted Cys2His2 zinc-finger motif at the C-terminus of TRI6 is involved in DNA binding, a C187A mutant was constructed in TRI6 using site-directed mutagenesis. The C187A mutant did not bind promoter DNA fragments, supporting the role of C187 in DNA binding. In addition, a TRI6 homologue in the distantly related macrocyclic trichothecene pathway of Myrothecium roridum (MRTRI6) was also shown to bind to the same TRI5 and TRI4 promoter fragments bound by TRI6. Together, these data confirm our previous proposal that TRI6 is an activator of trichothecene pathway gene expression and that DNA binding employs the C-terminal region of TRI6 containing three predicted Cys2His2 zinc fingers.

Biosynthetic and genetic relationships of B-series fumonisins produced by Gibberella fujikuroi mating population A.

Fumonisins are mycotoxins produced by the maize pathogen Gibberella fujikuroi mating population A and frequently contaminate maize. Wild-type G. fujikuroi produces four B-series fumonisins, FB1, FB2, FR3 and FB4. These toxins are identical in structure except for the number and positions of hydroxyls along their linear carbon backbone. To elucidate the genetic and biosynthetic relationships among these fumonisins, we conducted meiotic and biochemical analyses of G. fujikuroi mutants with altered fumonisin production that resulted from defective alleles at three loci, Fum1, Fum2 and Fum3. These mutants produced either no fumonisins, only FR2 and FB4, or only FR3 and FR4. Genetic analyses revealed the orientation of the Fum loci along linkage group 1 of the fungus. The mutants were grown together in pair-wise combinations to determine if their fumonisin production phenotypes could be complemented. When FR3- and FB2-producing mutants were grown together, complementation occurred. However, when a nonproducing mutant was grown with a FR2- or FB3-producing mutant, complementation did not occur or was incomplete. When purified FR2, FR3, or FB4 was fed to mutant cultures, FR4 was converted primarily to FR2, FR3 was converted to FB1 and FB2 was not converted. The results from these assays suggest a previously unrecognized branch in the fumonisin biosynthetic pathway.

Possible Role of Trichothecene Mycotoxins in Virulence of Fusarium graminearum on Maize.

Trichothecene-producing and -nonproducing Fusarium graminearum strains were tested for their ability to cause Gibberella ear rot in field trials at two locations-Ottawa, Ontario, and Peoria, Illinois-in 1996. Maize ears were inoculated with wild-type or transgenic F. graminearum strains in which the trichothecene biosynthetic pathway had been disabled by the specific disruption of the trichodiene synthase gene and with a derivative revertant strain in which trichothecene production had been restored through recombination. A silk channel inoculation method was employed at both locations. In addition, a kernel puncture inoculation method was used at the Ontario location. Harvested maize ears were analyzed for visual disease severity, grain yield, deoxynivalenol (DON) concentration, and fungal biomass by quantitative polymerase chain reaction (PCR) and/or ergosterol quantitation. There was a significant correlation (r= 0.86) between data obtained from the two different methods of quantifying fungal biomass. The trichothecene-nonproducing strains were still pathogenic but appeared less virulent on maize than the trichothecene-producing progenitor and revertant strains, as assayed by most parameters. This suggests that the trichothecenes may act as virulence factors to enhance the spread of F. graminearum on maize.

Linkage among genes responsible for fumonisin biosynthesis in Gibberella fujikuroi mating population A.

Most naturally occurring strains of the fungus Gibberella fujikuroi mating population A produce high levels of the mycotoxin fumonisin B1 (FB1), which is oxygenated at both carbons C-5 and C-10. Some strains, however, produce only FB2 or FB3, suggesting that they lack the ability to hydroxylate position C-10 or C-5, respectively. Genetic analysis indicates that these different phenotypes are due to single gene defects at closely linked loci designated fum2 and fum3. Further allellism tests indicate that both fum2 and fum3 are closely linked to fum1, a previously identified gene that regulates fumonisin production. The recovery frequency of FB1-producing progency from cross 510 between fum1 and fum2 mutations suggests a map distance of approximately 6.2 cM between these two loci. Amplified fragment length polymorphism analysis of parents and progeny of cross 510 was employed to confirm that the FB1-producing strains are recombinant progeny. We conclude that fum1, fum2, and fum3 constitute a fumonisin biosynthetic gene cluster on chromosome 1 of the restriction fragment length-map of G. fujikuroi.