In contrast to most primary metabolism genes, the genes involved in secondary metabolism and certain nutrient utilization pathways are clustered in fungi. Recently a nuclear protein, LaeA, was found to be required for the transcription of several secondary metabolite gene clusters in Aspergillus nidulans. Here we show that LaeA regulation does not extend to nutrient utilization or the spoC1 sporulation clusters. One of the secondary metabolite clusters regulated by LaeA contains the positive regulatory (i.e. aflR) and biosynthetic genes required for biosynthesis of sterigmatocystin (ST), a carcinogenic toxin. Analysis of ST gene cluster expression indicates LaeA regulation of the cluster is location specific as transcription of genes bordering the ST cluster are unaffected in a ΔlaeA mutant and placement of a primary metabolic gene, argB, in the ST cluster resulted in argB silencing in the ΔlaeA background. ST cluster gene expression was remediated when an additional copy of aflR was placed outside of the cluster but not when placed in the cluster. Site-specific mutation of an s-adenosyl methionine (AdoMet) binding site in LaeA generated a ΔlaeA phenotype suggesting the protein to be a methyltransferase.
Filamentous fungi display many unique characteristics that render them of great interest to the research community. Among these characteristics is the production of natural products, or secondary metabolites (Bennett, 1987). These compounds often have obscure or unknown functions in the producing organism but have tremendous importance to humankind. Secondary metabolites display a broad range of useful antibiotic and immunosuppressant activities as well as less desirable phyto- and mycotoxic activities.
These biological properties of natural products have spurred efforts towards identifying genes involved in their synthesis. Accumulating data from these studies dispelled an original premise that fungal metabolic genes would be scattered throughout the genome. In fact, the hallmark of secondary metabolite genes, in contrast to most genes involved in primary metabolism, is that they are clustered in fungal genomes (Keller and Hohn, 1997). Examples of secondary metabolite gene clusters are well exemplified in the genus Aspergillus and include those synthesizing antibiotics (penicillin, PN), pharmaceuticals (lovastatin) and toxins (aflatoxin, AF) and sterigmatocystin (ST) (reviewed in Zhang et al., 2004).
Recently, we identified a nuclear transcriptional regulator, LaeA, of secondary metabolite synthesis through complementation of a ST biosynthesis mutant in Aspergillus nidulans (Bok and Keller, 2004). Sequence analysis of LaeA showed no homology to known transcription factors but rather indicated some similarity to protein methyltransferases. Loss of LaeA silenced ST and PN production in A. nidulans, lovastatin production in Aspergillus terreus and gliotoxin production in Aspergillus fumigatus leading to the hypothesis that LaeA was involved in global regulation of secondary metabolite gene clusters. A role for global regulation of secondary metabolite gene clusters was further substantiated by recent microarray analysis of A. nidulans laeA mutantswhich highlighted the potential of LaeA to identify novel clusters and their metabolites (Bok et al., 2006).
Here we present several lines of evidence demonstrating that LaeA regulation of gene clusters appears specific to secondary metabolite clusters and, furthermore, is chromosome location specific. Placement of a primary metabolic gene, argB, in the ST cluster results in downregulation of argB expression in the ΔlaeA background. Additionally, an extra copy of aflR remediates ST biosynthesis in a ΔlaeA background when placed outside of the ST cluster but not inside the cluster.
Nutrient utilization clusters are not regulated by LaeA
In addition to gene clusters involved in secondary metabolism, fungi also possess catabolic pathways for the utilization of low-molecular-weight nutrients whose genes are often arranged in clusters. To examine the possibility that laeA might also regulate nutrient utilization cluster genes in A. nidulans, we examined transcripts from the proline (Garcia et al., 2004) and nitrate (Narendja et al., 2002) gene clusters in conditions known to induce or repress these genes in the wild-type (WT) strain (Muro-Pastor et al., 1999; Garcia et al., 2004). As shown in Fig. 1, representative transcripts were unaffected in the ΔlaeA background in both non-induced and induced conditions for prnD (Fig. 1A) or niiA (Fig. 1B) gene expression. Additionally, a representative transcript from the conidiation-specific SpoC1 gene cluster (Gwynne et al., 1984) was also examined and found not to be regulated by LaeA at 24 h although there appeared to be a slight increase of expression at 48 h (Fig. 1C).
Genes bordering the ST cluster are not regulated by LaeA
Gene expression data of the A. nidulansΔlaeA strain compared with WT showed that selected genes in the ST cluster (defined by Brown et al., 1996) were downregulated (Bok and Keller, 2004). These findings are extended by recent microarray work illustrating a similar pattern of regulation of secondary metabolite gene clusters where all or nearly all genes within a cluster are down- or upregulated in a ΔlaeA or overexpression laeA strain, respectively, with little effect on genes outside of the cluster (Fig. 2A and Bok et al., 2006). Here we present supportive data of these microarray results by assessing a transcriptional profile of the entire 60 kb ST gene cluster and flanking regions by Northern analysis in a ΔlaeA strain compared with WT (Fig. 2B). Virtually every gene in the cluster is downregulated in the ΔlaeA strain compared with WT. This is contrast to two flanking transcripts that show little if any difference in regulation in the two strains. Thus, LaeA regulation is specific to the cluster region and not border genes.
Remediation of ST gene expression in ΔlaeA is dependent on location of an extra copy of aflR
The gene encoding the ST pathway-specific transcription factor AflR is located in the ST cluster (Brown et al., 1996; Fernandes et al., 1998). An extra copy of aflR, whether in the ST cluster or placed at the trpC locus, increases ST cluster gene expression and subsequent ST production (Fig. 3 and data not shown). We investigated whether an extra copy of aflR could remediate ΔlaeA silencing. As shown in Fig. 3, aflR and stcU (a ST biosynthetic gene formerly called verA and regulated by aflR; Keller et al., 1994) were expressed in a ΔlaeA background when the extra copy of aflR was placed at the trpC locus but not when placed in the ST cluster.
A primary metabolism gene is regulated by LaeA when placed in the ST cluster
To further investigate the possibility that gene regulation by LaeA was location dependent, we identified a gene involved in arginine metabolism, argB encoding ornithine carbamoyltransferase (Berse et al., 1983), that was not regulated by LaeA when argB was located at its native locus (Fig. 4, lanes 3 and 4, and Table 1). Strains were created where argB was removed from its native locus and placed in the ST cluster using its own promoter. Using these strains, a comparison of argB mRNA levels in WT and the ΔlaeA mutant at 10 h, a time period when stc genes would not be expressed but optimal for argB expression, clearly shows argB transcription to be downregulated in the ΔlaeA mutant (Fig. 4, lanes 7 and 8). Quantification of argB showed an approximately 1/3-fold lower expression level in the cluster in the ΔlaeA background (Table 1). argB expression remained depressed in the ΔlaeA strain at later, ST-inducing time points (e.g. 48 and 72 h, data not shown). This downregulation of argB expression was reminiscent, although not as severe, to the silencing of ST gene expression in a ΔlaeA background (Figs 2 and 3 and Bok and Keller, 2004).
Table 1. Quantification of argB expression in Fig. 4.
argB (area volume)
Actin (area volume)
Transcripts were calibrated by ImageQuantTLv 2005 (Amersham Bioscience).
ΔargB::trpC; ΔstcE::argB; ΔlaeA
ΔargB::trpC; ΔstcE::argB; ΔlaeA
LaeA is a putative methyltransferase
The sequence of LaeA shows it contains conserved motifs commonly found in protein methyltransferases (Bok and Keller, 2004). Following procedures used to identify putative methyltransferase function, the AdoMet binding motif – LDLGCGTG – was mutated by replacing the two underscored glycines with alanines based on a modified Hamahata et al. (1996) method. Similarly to the ΔlaeA strain, the strain carrying the mutation in the AdoMet binding site abolished the expression of aflR and stcU required for ST production (Fig. 5). This result supports LaeA as a likely methyltransferase.
The cluster arrangement of fungal genes involved in a unifying metabolic or developmental process has elicited much discussion on origination and maintenance of such clustering (reviewed in Zhang et al., 2004). No one model can explain the finding of such varied gene clusters as nutrient utilization, mating type, pathogenicity islands and secondary metabolism in the fungal genome. However, the identification of LaeA, a nuclear protein shown to transcriptionally activate several secondary metabolite gene clusters in Aspergillus spp. (Bok and Keller, 2004), suggested a possible global mechanism involved in cluster gene regulation. Here we present evidence indicating that LaeA regulation does not extend to nutrient utilization clusters nor the SpoC1 conidiation-specific cluster. Furthermore, the mechanism of LaeA regulation of gene expression appears coupled to chromosome location, as moving genes in or out of the ST cluster (a cluster known to be regulated by LaeA) directly correlated to gain or loss of transcriptional regulation mediated by LaeA.
To address our hypothesis that laeA is specific to secondary metabolite gene cluster regulation, we examined the effects of loss of laeA on expression of a representative gene from two well-characterized primary metabolite gene clusters, the proline and nitrate utilization gene clusters. Our methods were based on well-established conditions that result in inductive or non-inductive conditions for these two clusters as described in Garcia et al. (2004) and Muro-Pastor et al. (1999) respectively. Loss of laeA appeared to have neither an effect on suppressing expression in induced conditions nor activating expression in the non-induced conditions (Fig. 1A and B). We also examined spoC1C expression in ΔlaeA as compared with WT following conditions described for spoC1 cluster expression (Law and Timberlake, 1980; Gwynne et al., 1984). The spoC1 cluster contains genes expressed during early conidiation events, a time frame (24 h) when most secondary metabolites are not expressed. As show in Fig. 1C, spoC1C expression was similar in both strains at 24 h although there did appear to be a slight increase in spoC1C expression in ΔlaeA as compared with WT at 48 h. Together, these results indicate that laeA has little impact on gene expression in these non-secondary metabolite gene clusters. This conclusion is supported by recent microarray analysis of laeA mutants compared with WT (Bok et al., 2006).
To further investigate factors affecting LaeA regulation of secondary metabolite gene clusters, we focused on gene regulation in the ST cluster. This approximately 60 kb cluster is bound on one side by stcA, encoding a polyketide synthase required for generating the ST carbon backbone (Yu and Leonard, 1995), and stcW, encoding a flavin-requiring monooxygenase (Keller et al., 2000), as well as the uncharacterized stcX (Brown et al., 1996). The sixth gene in the cluster is aflR, encoding a positive acting Zn(II)2Cys6 transcription factor required for the transcription of the biosynthetic stc genes (Fernandes et al., 1998). An examination of the expression of most genes in the ST cluster along with two uncharacterized border genes on either side of stcA and stcX, respectively, showed LaeA regulation was limited to the characterized ST cluster (Fig. 2 and Bok et al., 2006). This expression profile was similar to that of an aflR disruption strain (Yu et al., 1996). Thus, theoretically, LaeA regulation of the ST cluster could be entirely mediated by aflR suppression. Our finding that an extra copy of aflR could remediate the ΔlaeA phenotype, but only when placed outside of the ST cluster (Fig. 3), suggests that cluster location of aflR may play the most important role in mediating LaeA regulation for ST cluster expression. This finding also raises questions on LaeA function. If our assumption is that LaeA activity is required for accessibility to ST cluster chromatin, it is not intuitive that a trpC-located aflR should rescue ST gene expression in a ΔlaeA background. However, unlike the loss of aflR, loss of laeA does not result in a complete null ST phenotype. Depending on what media or temperature the fungus is grown, there is some ST production in the laeA mutant (data not shown). We speculate that the ST cluster in the ΔlaeA strain is not entirely closed and, should there be sufficient aflR expression (such as at the trpC locus), the resulting gene product would be able to activate ST gene expression in the ΔlaeA mutant. Our results are reminiscent to findings in Saccharomyces cerevisiae where it was shown that overexpression of the transcription factor PPR1 can suppress the silencing of an ura3 gene placed at the telomere (Aparicio and Gottschling, 1994). However, we currently do not know whether the mechanism underlying these results is related in these two fungi.
In earlier studies we found both the ST and PN clusters were regulated by LaeA (Bok and Keller, 2004; Bok et al., 2006; Fig. 2A). A significant difference in these clusters, one pertinent to our observations with AflR and ST regulation, is that PENR1, the HAP-like transcriptional complex regulating PN biosynthesis, is not located in the PN cluster. It is also involved in expression of numerous non-PN genes (Litzka et al., 1998; Brakhage et al., 1999). Interestingly penR1 was not regulated by LaeA based on our microarray analysis (Bok et al., 2006). Therefore, neither presence of a cluster-located transcription factor nor transcriptional regulation of such a factor is a requirement for LaeA control of secondary metabolite gene cluster expression. We suggest that LaeA exerts another layer of regulation on biosynthetic genes within a cluster in addition to the regulation by the pathway-specific regulator. For example, a study by Liang et al. (1997) showed that the expression of ver-1, an orthologue of A. nidulans stcU and involved in AF biosynthesis in Aspergillus parasiticus, varied depending on location of this gene in the genome. Variation in ver-1 expression as described by Liang et al. (1997) may be in part due to loss of regulation by LaeA.
The importance of chromosome location in LaeA regulation of gene expression was also supported by the argB expression profile. Figure 4 clearly shows argB expression is LaeA mediated when placed in the ST cluster. Although not silenced to the same degree as stc genes, it is nevertheless downregulated in the ΔlaeA background when placed in the ST cluster. We found this downregulation both at an early time point, 10 h, when ST genes are not expressed and at late, ST-compatible time points (48 h; data not shown). This experiment also supports a non-AflR mechanism involved in LaeA control of gene expression in the ST cluster region of the chromosome as argB is not regulated by AflR, nor is aflR expressed at 10 h.
Previous results showed LaeA to be a nuclear-located protein with motifs most similar to those of protein methyltransferases (Hamahata et al., 1996; Bok and Keller, 2004), proteins involved in gene regulation through modification of chromatin structure (Peterson and Laniel, 2004). Here we show that site-specific mutation of the conserved AdoMet-binding motif in LaeA generated a strain with an identical phenotype to the loss of function allele (Fig. 5), thus supporting LaeA as a methyltransferase. Our current efforts are directed towards the possibility that LaeA could be involved in methylation changes of histone proteins and/or their associated activating complexes, and by these means affects transcription of select secondary metabolite gene clusters.
Plasmids were constructed using standard techniques. Turbo (Stratagene) was used for PCR reactions. Primers for PCR and probes are listed in Table 2. Plasmid pDN02 was constructed by ligation of a 2.5 kb ApaI fragment, containing the aflR gene and 400 bp of the native promoter, into pPK1, a pBluescript SK-based plasmid containing a 1.9 kb blunt end ligated SspI fragment containing argB. Plasmid pJW20 was constructed by ligation of the 2.5 kb ApaI fragment containing aflR into pSH96, a pBluescript SK-based plasmid containing a 1.8 kb blunt end-ligated SacI–EcoRI fragment containing the 5′-end of the trpC gene. pJW63.4 was constructed using double-joint PCR reactions (Yu et al., 2004) to introduce mutations in the s-adenosyl methionine (AdoMet) binding site of LaeA (DLGCGTG → DLACATG) based on a previous publication (Hamahata et al., 1996) to inactivate methyltransferase activity. To introduce mutations in the AdoMet binding site, a 1.8 kb 5′ fragment of laeA was amplified by using two primers, LAE1 (Bok and Keller, 2004) and Rmet. Another two primers, Fmet and LAE2 (Bok and Keller, 2004), created a 2 kb 3′ PCR product of laeA. These two purified fragments were mixed and a 3 kb fragment was amplified by two nested primers, Mt1 and OER (Bok and Keller, 2004), containing HindIII restriction enzyme sites at both ends, to yield a final 3 kb fragment containing the entire modified laeA gene. This 3 kb fragment was subcloned into the HindIII site of pSH96 (Wieser and Adams, 1995) to created pJW63.4.
DNA extractions from fungal and bacterial strains, restriction enzyme digestion, gel electrophoresis, blotting, hybridization and probe preparation were performed according to standard methods (Sambrook et al., 1989; Shimizu and Keller, 2001). Total RNA was isolated from lyophilized mycelia using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturers' instructions. RNA blots were hybridized with a 1.3 kb EcoRV–XhoI aflR fragment from pJW20, a 2 kb PCR product by primers stcAF and stcAR, a 0.7 kb SacII–KpnI fragment from pRB7 containing the stcU coding region (Bok and Keller, 2004), a 1.0 kb SacI–SphI argB fragment from pDN02, a 3 kb laeA fragment from pJW45.4 (Bok and Keller, 2004), a 0.9 kb PCR product by primers NiiF and NiiR for niiA, a 0.4 kb PCR product by primers PrnF and PrnR for prnD, a 1 kb PCR product by primers SpoC1CF and SpoC1CR for spoC1 and a 0.4kb PCR product by primers ActF and ActR for the actin gene. Also A. nidulans cosmid pL11C09, which contains most of the ST gene cluster (Brown et al., 1996), was used as a probe for mRNA expression. Flanking transcripts of the ST gene cluster were probed with a 2 kb PCR product amplified by primers 5′flankingF and 5′flankingR for a transcript upstream of stcA (http://www.broad.mit.edu/annotation/fungi/aspergillus/, contig1.132, 243110–245278 bp), and with a 3 kb PCR product amplified by primers 3′flankingF and 3′flankingR for a transcript downstream of stcX (http://www.broad.mit.edu/annotation/fungi/aspergillus/, contig1.132, 180034–182950 bp).
Fungal strains and culture conditions
The fungal strains used in these experiments are shown in Table 3. All strains were maintained as glycerol stocks and were grown at 37°C on solid minimal media plates or in minimal media liquid shake cultures both containing 1% glucose (GMM) as the sole carbon source (Shimizu and Keller, 2001). For the expression of argB, 50 ml of GMM was inoculated with 107 conidia per ml and incubated for 10 h at 37°C, 300 r.p.m. for each fungal strain analysed. For the expression of spoC1C, the method of Law and Timberlake (1980) was followed for induction of conidiophore formation. Briefly, 50 ml of GMM was inoculated with 106 conidia ml−1 of WT or ΔlaeA and incubated for 20 h. Mycelia were harvested by filtration. To induce conidiation, the unwashed mycelia were placed on 9 cm Whatman no. 1 paper on top of solid GMM and incubated at 37°C. Conidia and mycelium were harvested at 24 h and 48 h for mRNA extraction. For niiA expression, the method of Muro-Pastor et al. (1999) was followed. Briefly, 100 ml GMM, substituting 5 mM urea (a non-induced, derepressed condition for niiA expression) for nitrate, was inoculated with 106 conidia per ml of WT or ΔlaeA. To obtain enough mycelia, four flasks of WT and four flasks of ΔlaeA were grown. After incubating for 7 h at 37°C at 300 r.p.m., the mycelia were harvested and washed and incubated in GMM-no nitrogen source medium for 20 min. The mycelia were then transferred to flasks containing either GMM + 5 mM ammonium D-(+) tartarate (non-induced, repressed conditions) or GMM + 10 mM NaNO3 (induced, derepressed conditions). The flasks were incubated for another 2 h in these media after which the mycelia were filtered, washed and used for RNA isolation. A method by Garcia et al. (2004) was followed for prnD experiments. Here 106 conidia per ml of WT and ΔlaeA were inoculated in 100 ml of FMM (minimal medium containing 0.1% fructose instead of glucose) containing 5 mM urea as the nitrogen source. To obtain enough mycelia, four flasks of WT and four flasks of ΔlaeA were grown. After incubating for 8 h at 37°C at 300 r.p.m., the mycelia were harvested, washed and transferred to flasks containing either MMG + 20 mM ammonium D-(+) tartarate (non-induced, repressed conditions) or MMF + 5 mM urea + 20 mM l-proline (induced, derepressed conditions). The flasks were incubated for another 2 h in these media after which the mycelia were filtered, washed and used for RNA isolation. For other expression experiments, we followed our previously published culture method (Shimizu and Keller, 2001). Sexual crosses were performed according to Pontecorvo et al. (1953).
Table 3. Aspergillus nidulans strains used for this study.
Fungal transformations were performed accordingly to standard techniques (Miller et al., 1985), with some minor modifications where protoplasts were embedded in minimal media top agar instead of spread by a glass rod on solid media. To examine location-dependent argB expression, strain TDN02.3 was constructed by transformation of strain RMS011 with plasmid pDN02, and strain TJW57.9 was constructed by transformation of strain RJW33.2 with plasmid pJW20. To identify a putative methyltransferase function of laeA, strain TJW60 was constructed by transformation of strain RJW32.2 with plasmid pJW63.4. To examine location effect of aflR, RDN01.55 and RJW54.8 were created by sexual cross between TJW57.9 and RDIT44.35. RDN04.8 and RDN05.2 were created by sexual crosses between TDN02.3 and RJW33.2, and between TDN02.3 and RDIT30.23 respectively. Genetic backgrounds of created strains were confirmed by Southern blot analyses and/or PCR analyses.
This research was funded by NSF Grant MCB-9874646 and NSF MCB-0236393 to N.P.K. and NIH (MBRS) Grant No. SO6GM08008 to S.P.K.