Introduction

Transcription factors containing bZIP domains (basic region, leucine zipper) are engaged in the regulation of diverse metabolic pathways in eukaryotes. They form homo- and heterodimers through the leucine zipper domains. Different combinations of monomers recognize different sequence elements in the promoters of target genes (Glover and Harrison 1995; Grigoryan and Keating 2008).

Expression of sulfur metabolism genes is controlled by bZIP transcription factors in numerous fungi: CYS3 in Neurospora crassa (Marzluf 1997), Zip1 in Schizosaccharomyces pombe (Harrison et al. 2005), Met4 and Met28 in Saccharomyces cerevisiae (Thomas and Surdin-Kerjan 1997). CYS3 from N. crassa is one of the best characterized bZIP family proteins in filamentous fungi. It is involved in coordinated expression of genes encoding enzymes responsible for acquisition and utilization of sulfur (Marzluf and Metzenberg 1968). CYS3 functions as a homodimer (Kanaan et al. 1992) and recognizes the palindromic sequence 5-ATGRYRYCAT-3 (Li and Marzluf 1996) present in promoters of sulfur assimilation genes (Fu and Marzluf 1990; Fu et al. 1989). In Aspergillus nidulans, an ortholog of CYS3 is MetR activating transcription of sulfur metabolism genes, in particular those involved in sulfate assimilation (Natorff et al. 2003). The Aspergillus fumigatus MetR protein is a bZIP transcription factor important for sulfur assimilation and pathogenicity (Amich et al. 2013).

Transcription factors specific to sulfur metabolism in fungi are controlled by activity of ubiquitin ligases from the SCF family. The role of the SCFMet30 complex in ubiquitination of the Met4 bZIP transcription factor was extensively studied in S. cerevisiae. Depending on growth conditions, ubiquitinated Met4 becomes inactivated (Kaiser et al. 2000) or is directed to degradation (Rouillon et al. 2000). Degradation takes place in response to an excess of sulfur amino acids in minimal media while in rich media the ubiquitinated protein remains stable (Menant et al. 2006). The A. nidulans ubiquitin ligase controlling activity of MetR is encoded by the scon genes (Natorff et al. 1998; Piotrowska et al. 2000). The level of the active MetR protein is apparently dependent on the sulfur status. When cysteine or methionine (which is readily metabolized to cysteine) is abundant, MetR is probably inactivated and/or degraded by the SCFSconB ubiquitin ligase complex, analogously to the S. cerevisiae Met4 protein. In consequence, expression of MetR-regulated genes is repressed to its minimal constitutive level. This regulatory system is known as the sulfur metabolite repression system (SMR) (Paszewski et al. 1994). In contrast to cys-3 in N. crassa, transcription of the metR gene itself is not controlled by SMR. Regulation of bZIP transcription factor stability may be more complex as an ortholog of MetR in S. pombe, Zip1, which is phosphorylated and this phosphorylation leads to an interaction with Pof1 and subsequent degradation (Harrison et al. 2005). Moreover, recent transcriptomic analyses found interaction of sulfur metabolism with metabolism of iron (Amich et al. 2013) and stress responses (Sieńko et al. 2014).

The A. nidulans genome contains 22 proteins, which are annotated in the Broad Institute database (http://www.broadinstitute.org), as containing bZIP domain. Using the MetR amino acid sequence as a query in a BLAST similarity search of the A. nidulans genome reveals that one of these proteins, encoded by the AN5218 open reading frame (ORF), contains a bZIP domain very similar to that of MetR. This finding prompted us to search for a function of the novel gene, which we named metZ. Orthologs of metZ are present in Eurotiales only, indicating specific regulation of sulfur metabolism in this taxon, which distinguishes it from other orders of Ascomycota. The high similarity of the MetR and MetZ bZIP domains suggested that overexpression of metZ could complement some deficiencies of the ΔmetR mutant. To verify this assumption, the metZ promoter was replaced with a MetR-independent one, which led to transcriptional activation of some genes, including those for transporters of sulfate and sulfur amino acids.

Materials and methods

Strains and plasmids

Aspergillus nidulans strains from our collection carrying standard markers (Martinelli 1994), used in the study, are listed in Table 1 along with plasmids and an Escherichia coli strain. The W1 wild-type (WT) strain of A. nidulans (Glasgow) was used as a reference for growth and qPCR experiments. A. nidulans chromosome-specific gene libraries constructed on cosmids pWE15 and pLORIST2 (Brody et al. 1991) were obtained from the Fungal Genetics Stock Center, Kansas, USA.

Table 1 List of strains and plasmids
Full size table

Growth conditions and genetic analysis

For protoplast or DNA isolation, mycelia were grown in complete (C) medium (Cove 1966). For RNA isolation, minimal sulfur free (MM-S) medium (Lukaszkiewicz and Paszewski 1976) supplemented with either 0.1 mM sulfate (low inorganic sulfur), 2 mM sulfate (high inorganic sulfur), 0.3 mM l-methionine (low organic sulfur), 1 or 5 mM l-methionine (repressing conditions) as a sole sulfur source was used. The MM-S medium was also supplemented according to the auxotrophic requirements of the strain employed. Liquid cultures were grown in 100 ml of MM-S medium at 37 °C for 16 h in a rotary shaker (200 rpm). Escherichia coli was grown in standard LB medium supplemented with antibiotics as required (Sambrook et al. 1989).

Nucleic acids manipulations

Standard procedures for plasmid propagation and isolation were according to (Sambrook et al. 1989). Sequences of primers used are listed in Supplementary Table S1. DNA from A. nidulans was isolated by the salting out method by grinding frozen mycelia under liquid nitrogen, followed by immediate suspension in warm STEN buffer (1 % SDS, 100 mM Tris pH 7.5, 50 mM EDTA pH 8, 100 mM NaCl) (Sambrook et al. 1989).

Total RNA was isolated from powdered mycelia using TRI Reagent (Molecular Research Center) according to the manufacturer’s protocol (Chomczyński 1993) and subsequently precipitated with lithium chloride added to a final concentration of 3.42 M as described by (Barlow et al. 1963). RNA for Northern blot analysis (30 μg of total RNA in each lane) was fractionated in 1 % agarose gels containing 1 % formaldehyde, and 1 × NBC buffer (0.5 M boric acid, 10 mM sodium citrate, 50 mM NaOH) and transferred to nylon membranes (BrightStar®-Plus, Ambion) using a Turbo blotter (Schleicher & Schuell, Inc.). 32P-labeling was done with DecaLabel DNA Labeling Kit according to the manufacturer (Thermo Scientific, Fermentas). Hybridization signals were visualized in a Fuji FLA-7000 PhosphorImager. Reverse transcription was performed using the RevertAid™ H Minus First Strand cDNA Synthesis Kit according to the manufacturer’s protocol (Thermo Scientific, Fermentas).

Plasmids bearing the metZ gene under the A. nidulans alcA or trpC promoter were constructed in the pAL3 vector bearing the N. crassa pyr-4 selection cassette (Waring et al. 1989). The entire metZ ORF with 3′UTR (from −12 bp upstream of the ATG start codon through 786 bp after the stop codon) was amplified by PCR, cut and ligated into KpnI–BamHI restriction sites of the vector yielding the PALMZ8 plasmid (Table 1). Next, the SpeI–KpnI fragment of the alcA promoter was replaced with PCR-amplified 377-bp fragment of the trpC promoter (Hamer and Timberlake 1987) yielding the kTRMZPG plasmid (Table 1).

The TOPO5′3′UTRMZPG plasmid (Table 1) for deletion of the entire metZ gene, including its promoter and 3′UTR, comprised the N. crassa pyr-4 selection marker surrounded by the 5′- and 3′-flanking regions of metZ (positions −2,735 through −1,070 relative to the metZ ATG codon and 365 through 2017 relative to the metZ stop codon, respectively). Both flanks of metZ along with the pyr-4 marker were PCR-amplified (sequences of primers are shown in Supplementary Table S1) and cloned initially in the PCR2-TOPOII vector (LifeTechnologies™). Then, the 5′- and 3′-flanks were cut and cloned respectively in the XbaI-NotI and BglII-KpnI restriction sites of the TOPOpyr4 plasmid (Table 1) yielding TOPO5′3′UTRMZPG, which was used for A. nidulans transformation. Deletion of metZ was attempted in the W12 wild-type strain and in strains bearing ΔnkuA (TN02A25) or ΔmetR mutation (RM131), all bearing the pyrG89 mutation.

Transformation of A. nidulans

Mycelia for transformation were collected by filtration, washed with 0.6 M KCl and suspended in 0.6 M KCl buffered with 10 mM potassium phosphate pH 6.5 and containing lytic enzymes: 15–20 mg/ml of Glucanex® 200G (Novozymes), 2 mg/ml of Driselase® (Sigma–Aldrich) and 1 mg/ml of snail acetone powder (Sigma–Aldrich). Protoplasts were prepared and transformed by the PEG method (Kuwano et al. 2008). Transformants were selected for uracil prototrophy on MM-S medium supplemented with 2 mM sulfate and 1.2 M sorbitol.

Microarray analysis

Transcriptomes of two congenic strains, K1 (ΔmetR, pyr-4 +) and TZ12 (ΔmetR, trpC pr::metZ, pyr-4 +), both grown in MM-S supplemented with 0.3 mM methionine, were compared in three biological replicates, each in two technical replicates with dye swap. cRNA probes fluorescently labeled with Cy3 or Cy5 were synthesized using Quick Amp Labeling Kit, two-color (Agilent Technologies) according to manufacturer’s protocol, using 5 μg of total RNA as a template. Labeled probes were hybridized concurrently to A. nidulans custom-designed microarray slides (purchased from Agilent) in an 8 × 15 k format containing oligonucleotides representing all known A. nidulans genes identified in the Aspergillus Genome Database (AspGD) version s06-m01-r07. Following hybridization, the microarrays were scanned with an Axon GenePix 4000B microarray scanner (Molecular Devices, LLC). Feature extraction was done with GenePix Pro 6.1. Raw LogRatio results from all biological and technical replicates were Lowess normalized, the resulting data for each gene were averaged and the statistical significance (p values) was calculated with Acuity 4.0 software. Additional data manipulations were done in Microsoft Excel. A gene was considered to be differentially expressed between the ΔmetR and ΔmetR,trpC pr ::metZ strains if its transcript level differed between the two strains at least twofold (|log2Ratio| > 1) and the probability of such a difference by chance was less than 0.05 (p < 0.05). The resulting lists of differentially expressed genes (Supplementary Table S2) were subjected to further bioinformatics analysis. ORF descriptions were retrieved from the Aspergillus Genome Database (http://www.aspgd.org/). The data have been deposited in NCBI’s Gene Expression Omnibus (Edgar et al. 2002) and are accessible through GEO Series accession number GSE62548 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE62548).

Quantitative real-time RT-PCR

Real-time RT-PCR was performed using the LightCycler® 480 System (Roche Laboratories) with SYBR Green detection, according to the manufacturer’s instructions. The primers used to quantify expression of target genes (Supplementary Table S1) were designed using Primer3 software (Untergasser et al. 2012). Primer specificity was verified by melting curve analysis. qRT-PCRs were performed in triplicate in 96-well plates with each 10-μl reaction mixture containing 5 μl of LightCycler® 480 SYBR Green I Master mix (Roche Laboratories), two primers (3 pmol of each) and 1 μl of diluted template cDNA. cDNA was synthesized from 5 μg of total RNA treated with DNaseI (Roche) using RevertAid™ H Minus M-MuLV Reverse Transcriptase kit (Fermentas) according to supplier’s protocol. The actA actin gene (AN6542) was used as the normalization reference (internal control) for target gene expression ratios. Average cycle thresholds were calculated and the Pfaffl method (Pfaffl 2001) was applied to calculate relative expression with respect to that of actin.

Bioinformatics tools

Homology searches were carried out against the GenBank (release 95.0) database with BLAST algorithm (Altschul and Lipman 1990). DNA and protein sequences were aligned with ClustalX (Thompson et al. 1997). Evolutionary tree based on protein distance matrix was built with the Kitsch program, which is part of the PHYLIP package version 3.6. (Felsenstein 1989). Primers were designed with Clone Manager Suite 9 (Scientific & Educational Software, Cary NC). Predicted dimers formed by bZIP proteins were drawn using the DrawCoil program (http://www.grigoryanlab.org/drawcoil/). Functional categories enriched among genes up- or down-regulated in the ΔmetR strain overexpressing metZ were identified using the FungiFun web server (https://sbi.hki-jena.de/FungiFun/FungiFun.cgi) (Priebe et al. 2011).

Results

The metZ gene was found by a BLAST search of the A. nidulans genome using the MetR protein sequence as a query, which yielded only one additional sequence with e-value amounting 7.0e-20. This gene encoded by the AN5218 ORF was named metZ. The newly identified gene is located on chromosome V and codes for a putative protein of 279 amino acids showing 26 % identity with MetR. The putative MetZ protein contains a bZIP domain showing 77 % identity and 91 % similarity to the bZIP domain of MetR (Fig. 1a). Such a high similarity suggests that MetZ could be a bZIP transcription factor potentially recognizing sequences in DNA similar to the MetR targets.

Fig. 1
figure 1

In silico analysis of Aspergillus nidulans metZ gene and its orthologs. a The evolutionary tree and alignment of the MetZ and MetR bZIP domains from A. nidulans and other fungi: Aspergillus, Gibberella, Magnaporthe, Neurospora, Penicillium, Saccharomyces, Talaromyces. b Exon–intron structure of the metZ gene and its orthologs. Positions of conserved regions in promoters and introns are marked by dark boxes. Arrows indicate primers used in qPCR (see “Materials and methods”). c Alignment of sequences conserved in promoters of metZ and its orthologs in Eurotiales. Conserved nucleotides are shaded and putative CYS3 (MetR)-binding sites are underlined. d Alignment of conserved regions located in introns (at ca. 2/3 of their length) of metZ orthologs in different species of Eurotiales. Conserved sequences and putative CYS3 (MetR)-binding sites are marked as above. e Putative dimers formed by MetR and MetZ proteins. Broken lines indicate ionic interactions stabilizing each dimer

Full size image

The UmetZ and LmetZ primers (Supplementary Table S1) designed to the central part of the metZ gene allowed for identification of the SL10G06 cosmid in the LORIST/pWE library. A 3,470-bp SalI-KpnI fragment bearing the metZ gene was excised and cloned into the pBluescript KS(–) vector yielding the kKSMZ plasmid, then the insert was verified by DNA sequencing. The sequence of the metZ gene was submitted to the EMBL/GenBank database under accession number KJ195521 and it is identical to the sequence of the AN5218 ORF in the A. nidulans FGSC A4 strain. Both metR (Natorff et al. 2003) and metZ genes are interrupted by a single unusually long intron, located in similar position of both genes. The intron in the metZ gene is 570 bp long (Fig. 1b), as determined by comparison of the RT-PCR—generated cDNA obtained on the template of mRNA with the genomic sequence.

Orthologs of MetZ are present in the Eurotiales order only, in contrast to MetR having orthologs in the entire Pezizomycotina subphylum (Fig. 1a). The MetZ protein forms a single branch of evolutionary tree (highlighted by a shaded box on tree in Fig. 1a), which suggests duplication of the ancestral MetR gene during evolution of Eurotiales. Analysis of promoter regions of metZ and its orthologs shows that all of them contain conserved palindromic 5′-ATGRYRYCAT-3′ elements located in the same position (Fig. 1c). Moreover, similar sequence motifs are also found within the long introns of metZ and its orthologs (Fig. 1d). This sequence motif is recognized by the N. crassa CYS3 protein (Li and Marzluf 1996), an ortholog of the MetR protein. Such motifs are also present in the promoters of the A. nidulans genes regulated by MetR. It has been shown earlier that the bZIP domains of MetR and CYS3 are functionally interchangeable (Natorff et al. 2003), thus the two proteins are likely to bind the same sequences.

To identify the function of the metZ gene we tried deleting it (see “Materials and methods”), but failed despite having repeated the procedure several times: all the transformants obtained contained an intact metZ gene. Thus, to identify the physiological function of metZ an alternative approach had to be applied. To study the effect of MetZ on expression of other genes the metZ ORF was fused with strong promoters alcA or trpC and these constructs were introduced into the W12 wild-type strain and the metR deletion mutant (RM131). Each construct partially complemented the defect of the ΔmetR strain because transformants overexpressing metZ, in contrast to the parental ΔmetR strain, did not need methionine for growth on complete medium though they still required it on minimal medium (Fig. 2). Moreover, the ΔmetR mutant overexpressing the metZ gene could be rescued with cysteine when grown on minimal medium, in contrast to the ΔmetR strain requiring supplementation with methionine (Fig. 2). These results suggest that MetZ could induce transcription of genes encoding sulfur amino acids permeases. This assumption is supported by the fact that overexpression of metZ in the wild-type strain resulted in an elevated sensitivity to ethionine (Fig. 2), a toxic analog of methionine. Moreover, overexpression of metZ in the ΔmetR mutant rendered this strain sensitive to chromate and selenate (toxic analogs of sulfate) even under repressive conditions. This result suggests that MetZ can activate transcription of the sB gene encoding sulfate permease.

Fig. 2
figure 2

Growth of A. nidulans strains on solid media supplemented with various sulfur sources. Wild-type strain W1 and the K1 ΔmetR mutant are compared with congenic transformants overexpressing the metZ gene

Full size image

To identify potential targets of the MetZ transcription factor we compared the transcriptome of the ΔmetR strain overexpressing metZ from the trpC promoter with that of the ΔmetR strain transformed with an empty vector, showing very low constitutive level of the metZ transcript. Overexpression of the metZ gene up-regulated 750 genes and down-regulated 552 genes (Supplementary Table S2). To find biological significance of the observed transcriptomic changes we analyzed KEGG pathways, GO terms and the Functional Catalogue categories assigned to all the genes whose transcript levels were changed at least twofold in ΔmetR overexpressing metZ. Those genes belonged to diverse main categories of the Functional Catalogue (Fig. 3). The most frequently represented among the differentially expressed genes was the main category Metabolism, containing 238 up-regulated and 180 down-regulated genes. Some of the up-regulated genes were those involved in sulfur metabolism, including methionine permeases (AN1631, AN8538, and AN12483), arylsulfatase (AN8341), taurine dioxygenases (AN2960, AN4111, and AN4108) and homocysteine synthase (cysD) (Fig. 4, see Supplementary Table S2 for details). However, some genes encoding enzymes of the sulfate assimilation pathway were unaffected by overexpression of metZ (e.g., sA coding for phosphoadenosine phosphosulfate reductase or the AN1752 encoding α-subunit of sulfite reductase) or were even down-regulated (e.g., sC encoding ATP sulfurylase or AN7600 encoding β-subunit of sulfite reductase). Thus, overexpression of metZ leads to elevated transcription of permeases and enzymes incorporating sulfide into sulfur amino acids while transcription of sulfate assimilation genes is decreased (Fig. 4). The simultaneous activation of two genes encoding oxidoreductases transferring electrons from sulfide to quinone (AN1825 and AN8346) could additionally contribute to controlling intracellular sulfide concentration. Interestingly, analysis of KEGG pathways revealed that the metZ overexpressing strain accumulates transcripts encoding three consecutive enzymes of the ubiquinone biosynthesis pathway (AN1743, AN3586 and AN4569). Among other up-regulated genes many are involved in biosynthesis of lipids, which might be components of biological membranes: ergosterol (AN0451, AN1901, AN4094, AN6506, AN8907), fatty acids (AN9407, AN9408) and unsaturated fatty acids (AN1037, AN4135, AN6731, AN7204). GO term analysis of genes up-regulated by overexpression of metZ showed that 58 of them are involved in transmembrane transport, including methionine permeases mentioned above. However, the sB transcript encoding sulfate permease was not detected in the microarray analysis while the results of growth tests suggested that it could be elevated (Fig. 2). Hence, we performed quantitative real-time PCR analysis for the sB transcript and additionally for some other relevant genes to complement the microarray data (Table 2). Levels of the sB and cysD transcripts were almost threefold induced and that of AN1631 (encoding an ortholog of the Saccharomyces cerevisiae Mup1p high-affinity methionine permease) even sevenfold in the strain overexpressing metZ. As expected, overexpression of metZ resulted in at least 70-fold induction of its own transcript. Using two pair of primers we found five-fold stronger up-regulation of the second exon than the first one under starvation conditions (Table 2) suggesting independent expression of the second exon, presumably starting from a putative second promoter located in the intron. This seems likely since the intron does contain conserved sequences that could activate transcription (Fig. 1d).

Fig. 3
figure 3

Categories of Functional Catalogue enriched among genes affected by overexpression of metZ. The number of up- (top) or down-regulated genes (bottom) is also shown

Full size image
Fig. 4
figure 4

Effect of the metZ gene overexpression on sulfur metabolism. Up-regulated genes are indicated by thick arrows, whereas down-regulated ones by empty arrows

Full size image
Table 2 Expression of selected genes in Aspergillus nidulans strains
Full size table

Expression of the metZ gene was studied in more detail using Northern blotting to reveal two bands in the wild-type strain under derepressing conditions (up to 4 h of sulfur starvation) (Fig. 5), supporting the above assumption that there is a second transcription start site located in the metZ intron. Transcription of metZ was strongly regulated by the sulfur metabolite repression system (SMR), being repressed in A. nidulans grown in the presence of methionine or high sulfate and derepressed after a shift to sulfur starvation conditions (Fig. 5). However, under the sulfur starvation conditions expression of only the short metZ transcript was up-regulated as shown by a real-time PCR analysis (Table 2). One should note here that in the overexpression experiment where the full metZ ORF was fused to a strong MetR-independent promoter, only the long (spliced) transcript was overproduced (Table 2; Fig. 5). Transcription of metZ depends on the MetR transcription factor because only a very low level of the metZ transcript was observed in the ΔmetR mutant even under sulfur starvation conditions (Fig. 5). The latter result confirms that the expression of the metZ gene is controlled by SMR and the MetZ protein is a second, beside MetR, transcription factor affecting expression of sulfur metabolism genes.

Fig. 5
figure 5

metZ gene expression in wild-type and ΔmetR strains. Transcript levels under sulfur starvation conditions in wild-type and ΔmetR strains are shown on the three leftmost lanes and the three central lanes, respectively. Overexpression of the metZ gene under the trpC promoter is compared to its endogenous expression in the recipient ΔmetR strain (two rightmost lanes). The actA transcript was used for normalization of RNA loading

Full size image

Discussion

In this study, we describe a novel paralog of the MetR bZIP transcription activator of A. nidulans, encoded by the metZ gene. The deduced amino acid sequences of both MetR and MetZ proteins contain almost identical bZIP domains so they could in principle recognize similar target sequences in DNA and consequently, affect transcription of the same genes. Since bZIP proteins usually form dimers (Pu and Struhl 1993), it seems plausible that MetZ could form the MetZ–MetZ homodimer or heterodimers with other transcription factors. Analysis of putative interactions between leucine zipper domains shows that MetZ and MetR proteins have very similar leucine zippers so they could potentially form homodimers and the MetZ-MetR heterodimer, each of them stabilized by six ionic bonds (Fig. 1e). In S. cerevisiae, two proteins involved in regulation of sulfur metabolism genes (Met4p and Met28p) form a heterodimer through their leucine zipper domains (Lee et al. 2010). However, Met4p and Met28p show low similarity to other fungal bZIP transcription factors involved in regulation of sulfur metabolism (Fig. 1a).

To study the role of MetZ as a putative transcription factor we analyzed transcriptomic effects of the metZ gene overexpression in the ΔmetR background. Microarray analysis revealed elevated expression of 58 genes involved in transmembrane transport including three genes encoding methionine permeases. Elevated expression of these genes could lead to complementation of the methionine requirement of the metR mutant on complete medium. One can speculate that genes encoding methionine permeases are normally activated by the MetR protein. If so, in the ΔmetR strain methionine could only be taken up by a general amino acid permease for which it would compete with other amino acids present in complete medium. Thus, to allow ΔmetR to grow, complete medium must be additionally supplemented with methionine. The induction of high-affinity methionine permeases by the overexpressed MetZ protein could allow methionine to enter the cell from complete medium even in the absence of methionine excess. This assumption is supported by the sensitivity of transformants overexpressing metZ to ethionine, a toxic analog of methionine. Overexpression of metZ induces also expression of the sB gene encoding sulfate permease, which results in elevated sensitivity of metZ transformants to structural analogs of sulfate—chromate and selenate.

On the other hand, decreased expression of the sC gene and the AN7600 ORF encoding β-subunit of sulfite reductase could explain why the metR mutant can not be complemented by overexpression of metZ on minimal medium. Thus, MetZ seems to be involved in the regulation of sulfur metabolism-related genes by fine tuning their expression (Fig. 4).

Microarray analysis of metZ overexpression revealed also up-regulated genes taking part in synthesis of ergosterol, fatty acids and unsaturated fatty acids which might be components of biological membranes. However, it is possible that some changes in the transcript levels could in fact be side effects of the MetZ protein excess resulting in the formation of non-physiological heterodimers with other bZIP transcription factors. Therefore, we have not attempted a more detailed analysis of possibly non-specific effects of gene overexpression.

Of interest are also the unusually long introns and the common exon–intron organization of the metR (Natorff et al. 2003) and metZ genes (Fig. 1b). The large intron of metZ contains several sequence motifs identical with the recognition sequence of the N. crassa CYS3 transcription factor, which could activate transcription from a second start point. Since MetR and MetZ proteins share similar bZIP domains with the N. crassa CYS3 protein (Fig. 1a), it seems likely that they could also recognize similar target sequences in DNA. Indeed, qPCR results (Table 2) have confirmed that expression of the second exon of metZ can be independently driven from an intronic promoter. We can speculate that the N. crassa cys-3 gene has similar expression pattern since two transcripts and two protein bands were detected under derepressing conditions (Fu et al. 1989; Tao and Marzluf 1998). The presence of two types of the metZ transcript, the spliced one and the short one containing only the second exon of the metZ gene suggests complex regulation of sulfur metabolism by MetR and two variants of MetZ. This type of gene organization, with an alternative promoter localized within an intron, has also been observed earlier in other genes, e.g., the human TP53 gene (Marcel et al. 2011).

We found that expression of metZ is MetR-dependent. Since high activity of MetR leads to accumulation of sulfide and activation of stress responses (Sieńko et al. 2014) then simultaneously elevated expression of metZ could decrease stress by lowering production of sulfide. The A. nidulans metZ gene, but not the metR gene, is regulated by the SMR system similarly to the cys-3 gene encoding the single sulfur-specific transcription factor in N. crassa. It appears therefore that the existence of two transcription factors activating sulfur metabolism genes only in Eurotiales discerns their regulation of sulfur metabolism from that of other fungi, and makes it potentially more sophisticated.