Essential to the biosynthesis of each mycotoxin is a specific regulatory gene encoding a protein that binds to cis-elements in the promoters of biosynthetic pathway genes. These transcription factors act as positive regulators that assist in recruiting RNA polymerase II to initiate transcription. The aflatoxin gene cluster contains the cis-regulator gene AFLR, which is essential for aflatoxin production (Chang et al., 1993; Payne et al., 1993). AflR belongs to a class of transcription factors referred to as zinc cluster proteins, found only in fungi and primarily in the Ascomycetes (MacPherson et al., 2006). The most extensively studied zinc cluster protein is Gal4 in yeast, which is responsible for activating genes involved in galactose metabolism (Traven et al., 2006). Zinc cluster proteins have six cysteine residues conserved in a motif that binds with two zinc molecules (Macpherson et al., 2006). The amino acid sequence of AflR contains the C-X2-C-X6-C-X6-C-X6-C that aligns perfectly with the Gal4 motif (Woloshuk et al., 1994). AflR binds to the consensus DNA sequence TCG(N5)CGA, which is found in the promoters of many aflatoxin biosynthetic genes (Yu et al., 1996; Fernandes et al., 1998). The fumonisin gene cluster in F. verticillioides also contains a gene (FUM21) encoding a zinc cluster protein that regulates transcription (Brown et al., 2007). Unlike AFLR, which lacks introns, FUM21 remained unrecognized in the cluster until Brown et al. (2007) discovered that the gene contains multiple introns that are alternatively spliced, giving rise to a family of transcripts, many not encoding the complete zinc cluster motif. The authors demonstrated that strains without a functional FUM21 lacked transcripts from two biosynthetic genes (FUM1 and FUM8) and failed to produce fumonisin B1 (FB1), the major form of fumonisin produced in nature. There are no reports describing the putative cis-element in the promoters of the fumonisin pathway genes that is recognized by Fum21. In F. graminearum, the transcription factor gene (TRI6) is located in the gene cluster on chromosome 2 (Proctor et al., 1995; Hohn et al., 1999). Tri6 is a member of the Cys2-His2 zinc finger family of regulators, which are common in all eukaryotes. This class of proteins associates with a single zinc molecule when binding to promoter sequence DNA (MacPherson et al., 2006). Disruption of TRI6 reduces transcription of genes involved in trichothecene biosynthesis (Proctor et al., 1995; Hohn et al., 1999). A consensus cis-promoter element to which Tri6 binds was identified as YNAGGCC (Proctor et al., 1995). Recent microarray and ChIP-Seq experiments indicate that in addition to regulating trichothecene biosynthesis, Tri6 has a more global impact on gene transcription (Seong et al., 2009; Nasmith et al., 2011).
RNA polymerase II complex and transcription
One function of the specific transcription regulators such as AFLR, FUM21, and TRI6 is to recruit the RNA polymerase holoenzyme complex to the gene promoter. This complex is composed of RNA polymerase II, general transcription factors, and Mediator (Holstege et al., 1998; Lewis & Reinberg, 2003). Studies in yeast have provided fundamental understanding of how these various components interact with the gene-specific regulators to initiate transcription (Mitsuzawa & Ishihama, 2004). It is also apparent that many of the signal transduction pathways eventually influence RNA polymerase function through the Mediator complex (Myers & Kornberg, 2000; Borggrefe et al., 2002; Lewis & Reinberg, 2003). With respect to mycotoxin biosynthesis, only a few studies provide insight into the transcription process. During the screening of REMI mutants of F. verticillioides, Shim & Woloshuk (2001) identified a strain with reduced FB1 production on autoclaved maize kernels. The mutant also accumulated a scarlet metabolite in the medium, grew more slowly, and produced very little aerial hyphae and fewer conidia than the wild type (Shim & Woloshuk, 2001). Acidic growth conditions restored conidia production and FB1 biosynthesis. The REMI insertion occurred in the gene FCC1, which encodes a protein similar to the Mediator protein Ssn8 in yeast (also known as Srb11 and Ume3; Myers & Kornberg, 2000; Shim & Woloshuk, 2001). Ssn8p is a C-type cyclin that pairs with a kinase encoded by SSN3 (also known as SRB10 and UME5; Myers & Kornberg, 2000). Bluhm & Woloshuk (2006) cloned the homologue of SSN3 in F. verticillioides (FCK1) and demonstrated by yeast two-hybrid analysis that Fck1 and Fcc1 associate with each other. The phenotype of the FCK1-disruption strain was similar to that of fcc1::REMI strain (Bluhm & Woloshuk, 2006). In yeast, Ssn8 and Ssn3p participate in the regulation of genes in response to environmental stress and nutrient availability. As yeast grows on a fermentable sugar, Gal4 and the C-terminus of polymerase II are phosphorylated by Ssn8, resulting in accelerated transcription of the GAL genes. Under conditions of stress, that is, nonfermentable carbon, Ssn8 and Ssn3 are degraded and Gal4 is not phosphorylated, resulting in increased expression of 173 genes (Cooper et al., 1997; Holstege et al., 1998; Rohde et al., 2000). The impact of FCC1 on gene expression was examined by subtractive hybridization with RNA from the wild type and fcc1::REMI strain, resulting in the identification of 658 unique ESTs (Shim & Woloshuk, 2001). When a microarray constructed with the ESTs was probed with RNA isolated from wild type and fcc1::REMI grown on autoclaved maize kernels, 116 ESTs were found to be differentially expressed (Pirttila et al., 2004). Significant differences were also measured for 166 ESTs during growth at low (3) and high (8) pH conditions.
A mutant of F. graminearum (Δcid1) disrupted in the FCC1 homologue CID1 exhibited many of the same developmental phenotypes as fcc1::REMI strain in F. verticillioides, including severe mycotoxin (DON) reduction (Zhou et al., 2010). The authors demonstrated that a functional CID1 is required for pathogenicity on wheat and maize and resistance to environmental stresses (Ni+, Cd2+, H2O2; Zhou et al., 2010). More recently, the FCK1 homologue was characterized as part of a larger study investigating the 116 protein kinase genes in F. graminearum. Referred to as FgSRB10 (Fg04484), the disruption mutant (Fgsrb10) grew poorly, exhibited reduced conidiation, and was nonpathogenic on wheat (Wang et al., 2011). Also among the kinase genes examined was FgCTK1 (Fg06793), a homologue of yeast CTK1, which is another cyclin-dependent kinase that phosphorylates the C-terminus of RNA polymerase II and is involved in transcript elongation (Ahn et al., 2009). Disruption of FgCTK1 resulted in reduced growth, conidiation, ascospore production, pathogenicity, and DON production (Wang et al., 2011). The cumulative results from these studies suggest that transcription of FB1 and DON biosynthetic genes involves similar Mediator components. No studies have examined the components involved in the RNA polymerase holoenzyme complex in A. flavus. However, studies in A. nidulans (Shimizu & Keller, 2001; Shimizu et al., 2003) demonstrated that phosphorylation of AflR by the cAMP-dependent protein kinase PkaA negatively impacts the function of this regulator. This modification appears to be part of the signaling mechanism that regulates the functional activity of AflR, but it is not known whether this serves as a signal to Mediator or the RNA polymerase II complex.
Epigenetic regulation broadly refers to the changes in gene expression regulated by mechanisms not linked to mutations in the gene sequence (Vaquero et al., 2003). Such changes result in a variety of phenotypes induced in the same genotype often in response to environmental or metabolic cues. The mechanisms most recognized as epigenetic are DNA methylation, histone modifications, RNA interference (RNAi), and chromosome position effects. Current evidence indicates that histone modification is involved in the epigenetic regulation of mycotoxin biosynthesis, but not DNA methylation (Liu et al., 2012). Furthermore, the epigenetic phenomenon affecting mycotoxin biosynthesis is also global across the genome, consequently impacting the expression of many other genes and pathways.
Loss of AFLR expression (LAEA) has provided the clearest evidence for epigenetic regulation of mycotoxin biosynthesis. Butchko et al. (1999) used a creative visual screen to identify mutations that affected sterigmatocystin biosynthesis in A. nidulans. In this screen, a total of 176 isolates failed to produce the orange-colored metabolite norsolorinic acid of the sterigmatocystin pathway, which accumulated in the mutagenized strain. Three mutants that appeared normal with respect to growth and development exhibited no AFLR expression. Bok & Keller (2004) characterized LAEA as the mutation in one of three strains lacking AFLR expression. The gene is located on chromosome 8 in A. nidulans, whereas its homologue in A. flavus is on chromosome 2 (Bok & Keller, 2004). Without a functional LAEA, production of secondary metabolites is repressed, while growth and conidiation are not affected (Bok & Keller, 2004; Georgianna et al., 2010). Recent evidence indicates a clear link between LaeA, chromatin modification, and transcription of the genes involved in sterigmatocystin biosynthesis (Reyes-Dominguez et al., 2010). In addition, over-expression of LAEA revealed gene clusters responsible for the synthesis of other secondary metabolites (Bok et al., 2006). Histone modification, particularly the N-termini of the core histone proteins, is a major mechanism of epigenetic regulation. Modifications, such as methylation, acetylation, phosphorylation, and ubiquitylation, result in the reversible switching between heterochromatin and euchromatin states (Vaquero et al., 2003). Reyes-Dominguez et al. (2010) demonstrated that histone subunit H3 with methylation at lysine 9 (K9) of the N-terminus functions in maintaining a heterochromatin state surrounding the gene cluster responsible for sterigmatocystin biosynthesis, thus rendering the region inaccessible to the RNA polymerase II complex. When conditions are suitable to trigger sterigmatocystin production, the level of methylated H3 subunits decreases, resulting in euchromatin state at the gene cluster locus (Reyes-Dominguez et al., 2010). Although LaeA has a role in this transition, the exact mechanism remains unresolved.
The LAEA homologue (FfLAE1) was recently identified in Fusarium fujikuroi, which causes Bakanae disease of rice (Wiemann et al., 2010). This pathogen is primarily known as the model for studying gibberellin biosynthesis, but it is also capable of producing small amounts of fumonisins (Proctor et al., 2004; Stepien et al., 2011). The ΔFflae1 mutant of F. fujikuroi exhibited reduced gibberellin production, but measurements of fumonisin production were not presented (Wiemann et al., 2010). Significantly, FfLAE1 restored the production of the sterigmatocystin pathway when expressed in the laeA mutant of A. nidulans (Wiemann et al., 2010), demonstrating functional conservation in two fungal species.
Acetylation of the core histone proteins reduces the affinity of histone to DNA due to effects on the charge, whereas deacetylation results in more condensed chromatin structure (Vaquero et al., 2003). The effect of the histone deacetylase gene HDAA on sterigmatocystin biosynthesis in A. nidulans was examined by Shwab et al. (2007). Deletion of HDAA caused the accumulation of STCU and AFLR transcripts 6–12 h sooner than in the wild-type strain, but did not impact fungal growth. Different results were obtained in F. graminearum in which a pair of genes (FLT1 and HDF1) was identified as important components for histone deacetylation (Ding et al., 2009; Li et al., 2011). Deletion of these genes reduced conidia production, pathogenicity, and histone deacetylase activity; however, the mutation of HDF1 was solely responsible for more than 30% reduction in DON production (Li et al., 2011). Two other putative histone deacetylase genes (HDF2 and HDF3) were also studied. The Δhdf3 strain exhibited wild-type growth, conidiation, and pathogenicity, but produced 60% less DON (Li et al., 2011). This difference observed in A. nidulans and F. graminearum is intriguing. Shwab et al. (2007) suggested that HDAA affects the gene cluster near the telomere, but not those at more proximal locations. Deletion of the HDAA homologue in Aspergillus fumigatus up-regulated the production of secondary metabolites whose genes were not located near telomeres (Lee et al., 2009). The deletion also reduced conidia germination and the production of gliotoxin. These results illustrate the complexity of epigenetic regulation.
In F. verticillioides, application of the histone deacetylase inhibitor chostatin A resulted in increased transcription of FUM1 and FUM21 (Visentin et al., 2012). Although not statistically significant, FB1 levels were consistently higher in the chostatin A-treated cultures. Co-precipitation assays with antibodies specific for acetylated histone H4 yielded more promoter DNA from FUM1 and FUM21 under conditions conducive for FB1 production than under normal culture conditions. The results indicated that histone acetylation has an important role in the transcriptional regulation of fumonisin genes (Visentin et al., 2012).
A few recent studies have examined nonhistone proteins that are involved in chromatin modification. The heterochromatin protein HP1 can bind to the K9 methylated H3, and the HP1 ortholog in A. nidulans (HEPA) has a major role in silencing transcription of the sterigmatocystin genes (Reyes-Dominguez et al., 2010). To determine whether HP1 is important for DON production, Reyes-Dominguez et al. (2012) examined a HEP1 mutant of F. graminearum. Growth of the mutant was similar to that of the wild-type strain, but DON production and transcription of TRI5 and TRI6 were significantly reduced (Reyes-Dominguez et al., 2012).
Much attention has been focused on the Velvet (VE) complex and its role in the epigenetic regulation of mycotoxin production. The VE gene family, which was first described in A. nidulans, is major regulator of light response, development, and secondary metabolism (Bayram & Braus, 2012). In A. nidulans, VeA and VelB interact with LaeA to facilitate the epigenetic activity of LaeA (Bayram et al., 2008; Yin & Keller, 2011). The presence of light greatly impacts this interaction resulting in less sterigmatocystin, and the converse is also true (Stinnett et al., 2007; Bayram et al., 2008). In F. verticillioides, mutations in FvVE1, a homologue of VEA, abolish FB1 production and transcription of FUM21 and FUM8 (Myung et al., 2009). Furthermore, FvVE1 appears to be required for disease symptom development in inoculated maize seedlings (Myung et al., 2012). The FvVE1 mutants grew endophytically in maize tissues and failed to produce the necrotic leaf symptoms observed with wild-type strains. Two research groups have recently published the results of their studies on the VeA homologue (FgVEA and FgVE1) in F. graminearum (Jiang et al., 2011; Merhej et al., 2012). As with the F. verticillioides gene, mutation of FgVEA (FgVE1) reduces mycotoxin (DON) production and pathogenicity. Even though FgVEA and FgVE1 designate the same gene (FGSG_11955) in F. graminearum genome, the authors reported contrasting effects on conidia production and germination in their mutant strains. Deletion of the gene (FgVEA) resulted in a significant increase in conidia production, but the time needed for germination was twice that of the wild-type strain PH-1 (Jiang et al., 2011). In contrast, an insertional mutation (FgVE1) resulted in a severe reduction in the number of conidia produced, but spore germination was not affected (Merhej et al., 2012). In F. verticillioides, Fanelli et al. (2012) showed that expression of FvVE1 is affected by light, as observed with A. nidulans. The study demonstrated that fumonisin production and expression of FvVE1, FUM1, and FUM21 are greater under individual color (red to blue) exposure and pulsed light than under continuous white light or dark conditions.
Epigenetic effects also were hypothesized to explain the silencing of aflatoxin production in parasexually derived diploids of A. flavus strain 649 (Woloshuk et al., 1995; Smith et al., 2007). Papa (1979, 1980) described strain 649 as a nonproducer of aflatoxin that exhibits dominant repression of aflatoxin in diploids. He also mapped the mutation (afl1) to the same locus that was later determined as the aflatoxin gene cluster (Foutz et al., 1995). Once the genomic sequence of A. flavus became available, Smith et al. (2007) probed strain 649 by Southern analysis and discovered that the strain has a 317-kb deletion in chromosome 3, which includes the entire aflatoxin gene cluster. The deleted chromosome end was replaced with a duplicated region (939 kb) from chromosome 2. The silencing observed in diploids was hypothesized as a transvection phenomenon (Woloshuk et al., 1995; Smith et al., 2007). First described in Drosophila melanogaster, transvection (or trans-sensing) is the inactivation of alleles due to somatic chromosome pairing and often arises due to genomic rearrangements (Wu & Morris, 1999; Duncan, 2002). Because fungi are haploid, this type of silencing effect is rare. In Neurospora crassa, transvection was described for ASM-1, a gene required for ascospore maturation (Aramayo & Metzenberg, 1996; Aramayo et al., 1996). Deletion of ASM-1 has a recessive effect on growth in the haploid fungus; however, during the diploid phase of meiosis, the mutation exerts a dominant effect resulting in aborted ascospores (Aramayo & Metzenberg, 1996). In strain 649, the repression of gene expression in diploids is confined to the aflatoxin biosynthetic genes with normal expression of genes outside the cluster (Smith et al., 2007). Also, ectopic insertion of AflR restores aflatoxin production in diploids (Smith et al., 2007). In D. melanogaster, several epigenetic mechanisms have been suggested as explanations for transvection (Wu & Morris, 1999). It is reasonable to propose that in diploids formed with strain 649, the heterochromatin state at the aflatoxin cluster is not affected by the expression of LAEA. It remains unclear how the expression of AFLR at another locus restores transcription of the genes in the aflatoxin cluster.