To investigate cation adaptation and homoeostasis in Aspergillus nidulans, two transcription-factor-encoding genes have been characterized. The A. nidulans orthologue crzA of the Saccharomyces cerevisiae CRZ1 gene, encoding a transcription factor mediating gene regulation by Ca2+, has been identified and deleted. The crzA deletion phenotype includes extreme sensitivity to alkaline pH, Ca2+ toxicity and aberrant morphology connected with alterations of cell-wall-related phenotypes such as reduced expression of a chitin synthase gene, chsB. A fully functional C-terminally GFP (green fluorescent protein)-tagged form of the CrzA protein is apparently excluded from nuclei in the absence of added Ca2+, but rapidly accumulates in nuclei upon exposure to Ca2+. In addition, the previously identified sltA gene, which has no identifiable homologues in yeasts, was deleted, and the resulting phenotype includes considerably enhanced toxicity by a number of cations other than Ca2+ and also by alkaline pH. Reduced expression of a homologue of the S. cerevisiae P-type ATPase Na+ pump gene ENA1 might partly explain the cation sensitivity of sltA-null strains. Up-regulation of the homologue of the S. cerevisiae vacuolar Ca2+/H+ exchanger gene VCX1 might explain the lack of Ca2+ toxicity to null-sltA mutants, whereas down-regulation of this gene might be responsible for Ca2+ toxicity to crzA-null mutants. Both crzA and sltA encode DNA-binding proteins, and the latter exerts both positive and negative gene regulation.
- Aspergillus nidulans
- calcium signalling
- cation homoeostasis
- cell-wall integrity
- nuclear localization
- transcription factor
Rapid adaptation to different environmental conditions, including appropriate adaptive responses to various cations, is crucial for the survival and proliferation of micro-organisms. The Ca2+/calcineurin/Crz1p signalling pathway is necessary for adaptation to diverse environmental conditions, including high cation levels and alkaline pH, and considerable characterization of it has been performed in Saccharomyces cerevisiae, Schizosaccharomyces pombe, Torulaspora delbrueckii and Candida albicans [1–4]. Sudden exposure of S. cerevisiae to these stress conditions induces a transient increase in cytosolic Ca2+ levels, which triggers activation of the phosphatase calcineurin through binding of the Ca2+-binding protein, calmodulin. S. cerevisiae calcineurin, like that of mammals, is composed of both catalytic and regulatory subunits . Activated calcineurin mediates the dephosphorylation of the cytoplasmically localized transcription factor Crz1p, resulting in its translocation to the nucleus  (reviewed in ).
The nuclear-localized zinc finger Crz1p transcription factor then activates the expression of genes whose promoters contain physiologically relevant CDREs (calcineurin-dependent response elements)  5′-GNGGC(G/T)CA-3′, and up-regulates the expression of various target genes, including some which encode cation transporters that act at the plasma membrane or other membranous organelles [8–10]. We previously found that alkaline-pH-induced up-regulation of the Aspergillus giganteus afp (antifungal protein) gene can be prevented by FK506, an inhibitor of calcineurin . To pursue afp regulation in the context of cation adaptation and homoeostasis, we have chosen to investigate Ca2+-mediated signalling in Aspergillus nidulans, an experimentally more amenable species than A. giganteus.
Ca2+-mediated signalling has been less investigated in filamentous fungi than in yeasts, although some notable work was performed on calcineurin , which has been revisited recently . It was initially reported that the A. nidulans gene encoding the calcineurin catalytic subunit is essential and required for cell-cycle progression [14,15]. Very recently, however, an A. nidulans ΔcnaA mutant strain, which had the calcineurin catalytic subunit deleted, was found to be viable and to display a phenotype similar to that observed for the corresponding Aspergillus fumigatus (ΔcalA) mutant strain, which has severe defects in growth extension, branching and conidiophore development . In addition to displaying abnormal morphogenesis, the A. fumigatus ΔcalA mutant is also impaired in its virulence [16,17]. In Aspergillus oryzae, calcineurin is required for its adaptation to elevated temperature, high Na+ levels and alkalinization . Crz1p orthologues have been recently characterized in the filamentous fungi A. fumigatus and Botrytis cinerea, where they play major roles in tolerance of certain cations [13,19].
Another transcription factor important for cation adaptation and homoeostasis is SltA. An A. nidulans mutation, sltA1, resulting in toxicity by high levels of Na+ and K+, was described previously . Subsequently, cloning of the sltA gene by complementation of the sltA1 mutation showed that it encodes a protein containing three Cys2His2-type zinc fingers and exhibits significant similarity to the Ace1 transcriptional repressor of the cellulase and xylanase genes in Hypocrea jecorinia (Trichoderma reesei) . Although O'Neil et al.  showed that the sltA1 mutation results from a premature stop codon just downstream of the third zinc finger of the predicted protein, they did not investigate the role of the SltA protein in cation tolerance, nor did they determine whether sltA1 has a null phenotype.
In the present paper, we have characterized aspects of the role of the A. nidulans transcription factors CrzA, a homologue of S. cerevisiae Crz1p, and SltA, which has no identifiable S. cerevisiae homologue, but, like CrzA, is involved in cation adaptation and homoeostasis. We show that both CrzA and SltA are capable of binding DNA and that SltA acts positively on transcription of the Ena1p-like Na+ pump gene enaA and negatively on transcription of the putative vacuolar Ca2+/H+ exchanger gene vcxA. The negative regulation of vcxA by SltA is opposed by its transcriptional activation by CrzA, and the positive regulation by CrzA is the over-riding factor, as vcxA mRNA levels in a mutant lacking both of the transcription factors are indistinguishable from those in a mutant lacking only CrzA. Exposure to Ca2+ results in nuclear localization of CrzA.
Strains, growth conditions, molecular techniques and phenotype analysis
A. nidulans strains used are listed in Table 1. Escherichia coli strains DH5α (GibcoBRL) and XL10-Gold® (Stratagene) were used as hosts for the maintenance of the plasmids listed in Supplementary Table S1 (at http://www.BiochemJ.org/bj/414/bj4140419add.htm). General cloning procedures for E. coli were followed. Standard growth medium for A. nidulans, genetic techniques and phenotype testing were as described previously ( and references therein). The gene symbols used were described previously . Reversion alleles of sltA1 were selected in a UV-mutagenized strain of genotype halA24 inoB2 sltA1 by their ability to enable growth on appropriately supplemented complete medium containing 1 M NaH2PO4 (final concentration) and their phenotypes were determined in a selection of outcrossed strains. A. nidulans transformation was performed as described previously ( and references therein). A. nidulans total RNA and genomic DNA preparation, as well as Northern blot and Southern blot analyses, were performed following standard protocols ( and references therein).
Identification of the crzA gene
The CrzA homologue in the A. nidulans genome was identified by Blastp searches using the S. cerevisiae Crz1p sequence. On the basis of sequence conservation and apparently similar function, AN5726 was named crzA. The position of the predicted intron within the crzA gene was verified by RT-PCR (reverse transcription–PCR) and is in agreement with GenBank entry AB259568. Strain J734 was grown in liquid minimal medium and exposed for 2 h to 1 M NaCl, after which total RNA was obtained. Total RNA treated with 7.5 units of DNase I (FPLC pure, Amersham Biosciences) was used as the template for RT–PCR. First-strand synthesis was performed using primer CrzA-9 rev (see Supplementary Table S2 at http://www.BiochemJ.org/bj/414/bj4140419add.htm) and 200 units of ArrayScript ™ Reverse Transcriptase (Ambion). The following PCR using primers CrzA-8 fw and CrzA-9 rev yielded full-length crzA cDNA, which was cloned into pGEM-T-Easy (Promega), resulting in pGEM-T-CrzAcDNA. The insert was subsequently sequenced using primer walking ( and references therein).
Deletion of crzA
The crzA deletion cassette was constructed as a gene-replacement cassette comprising the pyr-4 gene of Neurospora crassa as the selection marker, flanked by 1.5 kb of 5′- and 3′-UTRs (untranslated regions) of the crzA gene. The 5′-UTR of crzA was amplified by PCR using genomic DNA from strain J734 and the primers crznulo-1 and crznulo-2, which introduced NotI and BamHI restriction sites respectively. The resulting PCR fragment was digested with NotI and BamHI and inserted into pBluescriptSK(+) (Stratagene), resulting in pBSKpromCrzA. The pyr-4 selection marker, obtained from plasmid pFB6  by BglII digestion, was inserted into BamHI-linearized pBSKpromCrzA to generate pBSKpromCrzApyr4-fw. The correct orientation of the pyr-4 gene was verified by PCR and restriction analysis. The 3′-UTR of crzA was amplified by PCR using genomic DNA from strain J734 and the primers crznulo-3 and crznulo-4. The resulting PCR fragment was inserted into pGEM-T-Easy and excised using EcoRI digestion. The EcoRI fragment was inserted in the correct orientation into pBSKpromCrzApyr4-fw via the EcoRI restriction site. The final plasmid containing the 6.2 kb crzA deletion cassette was named pBSKdeltaCrzA. This cassette was amplified by PCR using primers crzAnulo1 and crzAnulo-4 and used to transform A. nidulans strain MAD1425. Pyrimidine prototrophic transformants were screened for the crzA deletion by Southern blot and PCR analysis, and strain BER02 was selected for further work.
Deletion of sltA
A cassette comprising the A. fumigatus riboB gene fused to the upstream and downstream UTRs of sltA was used to delete the sltA coding region. A 1.5 kb 5′-UTR of sltA was amplified using the primers sltnul1 and sltnul2, which included NotI and BamHI restriction sites respectively. The PCR fragment generated was digested with NotI and BamHI and cloned into the same restriction sites in pBluescriptSK(+) (Stratagene) to generate pBS-SltA5′. A 1.5 kb 3′-UTR of sltA was amplified using the primers sltnul3 (inserting an EcoRI restriction site) and sltnul4 (containing a XhoI site). The PCR fragment was digested with EcoRI and XhoI and ligated into pBS-SltA5′ to generate pBS-SltA5′3′. The EcoRI 1.9 kb fragment containing the A. fumigatus riboB coding region and 5′- and 3′-UTRs was obtained from p1548, and ligated into pBS-SltA5′3′ digested with EcoRI to create pBSdeltaSltA. pBSdeltaSltA was digested with NotI and XhoI to yield the deletion cassette. The sltA-null strain was generated using MAD1427 as the recipient strain and selecting for transformants which were able to grow in the absence of riboflavin. Riboflavin prototrophs were analysed by Southern blotting, and one transformant carrying the desired insertion was used for further work.
GFP (green fluorescent protein)-tagging of CrzA
Construction of an A. nidulans strain expressing CrzA–GFP was performed following a method described previously , taking advantage of recently available strains which are deficient in heterologous recombination . Amplification from 1.5 kb from the far 3′-end of the gene and the 3′-UTR region was performed using the primer pairs CrzA1/CrzA2 and CrzA3/CrzA4 respectively (see Supplementary Table S2) and fused in a three-way PCR with a fragment amplified with primers CrzA-GFP1 and CrzA-GFP2 which contained the gfp gene together with the selectable pyrG gene of A. fumigatus . The fusion cassette was transformed into the MAD1732 strain following a protocol described previously ( and references therein). Pyrimidine prototrophs were selected and homokaryotic transformants were analysed by Southern blotting for single-copy integration of the gene-replacement cassette.
Steady-state mRNA levels of the A. nidulans ENA1-like genes were detected using specific PCR-amplified genome probes. For AN6642, a probe of 1357 bp using primers An6642 rev and An6642 fw, covering 40% of the ORF (open reading frame), was used. For AN1628, a probe of 948 bp using primers An1628 rev and An1628 fw, covering 29% of the ORF, was used. For AN10982, a probe of 1295 bp using primers An10982 rev and An10982 fw, covering 37% of the ORF, was used. To detect AN0471 (designated vcxA), a 926 bp fragment using primers An0471 rev and An0721 fw, covering 65% of the ORF, was synthesized. For chsB detection (AN2523), a fragment of 719 bp was amplified using primers An2523 rev and An2523 fw, covering 30% of the ORF. Actin transcript detection was used as the probe for the fifth exon, amplifying the genomic DNA with primers Anid-AcnA and Anid-AcnB. For details of oligonucleotide primer sequences see Supplementary Table S2.
Northern-blot analyses used total RNA extracted from wild-type and mutant strains grown in MMA (minimal medium agar) (pH 6.5) containing 1% glucose and 10 mM ammonia as the main carbon and nitrogen sources respectively. To investigate the effects of various cations, the strains were incubated for 18 h at 37 °C, after which the addition of various substances and/or alterations to the pH were performed and incubated for a further time period as stated in the Figure legends.
Quantification of enaA and acnA transcription was performed by exposing the blots to a PhosphorImager screen (Molecular Dynamics) and developing using a FLA-5100 Reader (Fujifilm). Band intensities were measured using Multi-Gauge V3.0 software (Fujifilm).
Construction of GST (glutathione transferase)–CrzA and GST–SltA fusion proteins
The plasmid pGEX-CrzA123(469-603) was constructed to express a GST fusion protein containing the complete zinc-finger region of CrzA (amino acids 469–603). The zinc-finger region of crzA was amplified by PCR using genomic DNA from strain J734 as the template and the primers CrzA-ZF-fw-BamHI and CrzA-ZF-rev-EcoRI. Both oligonucleotide primers introduced the restriction sites for BamHI and EcoRI that allowed in-frame ligation of the PCR fragment into pGEX-2T (Amersham Biosciences) to generate pGEX-CrzA (469-603). The plasmid expressing the GST–SltA (amino acids 381–549) fusion protein was constructed by cloning a PCR amplicon obtained from J734 genomic DNA with the oligonucleotide primers SltA-ZF-fw-BamHI and SltA-ZF-rev-EcoRI (amplifying the entire zinc-finger region of the sltA gene and introducing BamHI and EcoRI restriction sites) into pGEX-2T to generate pGEX-SltA(381-549). Fidelity and the correct orientation of the inserts were verified by DNA sequencing. Fusion proteins were expressed in E. coli DH1 strain and purified by glutathione-affinity chromatography following the manufacturer's instructions (glutathione– Sepharose 4B resin; Amersham Biosciences).
EMSA (electrophoretic mobility-shift assay)
Synthetic double-stranded oligonucleotide primers were generated as described previously . A. giganteus sites afp-CDRE-5 and afp-SLTA-1 were obtained by annealing the oligonucletide primers CDRE-5 with CDRE-5(−) and SltA-1 with SltA-1(−) respectively. T. reesei probes 2C and 2C4 were obtained by annealing the oligonucleotide primers Ace1-2C-1 with Ace1-2C-2 and Ace1-2C4-1 with Ace1-2C4-2 respectively. For the A. nidulans chsB CDRE-1 site, oligonucleotide primers Chs-CDRE-1fw and Chs-CDRE-1rev were annealed (for oligonucleotide primer sequences see Supplementary Table 2).
Synthetic probes were labelled by end-filling using 2 units of Klenow polymerase (Roche) in the presence of a mixture containing dGTP (25 μM final concentration) and [α-32P]dCTP (3000 Ci/mmol, PerkinElmer). Binding reactions were performed as described previously . The binding reaction mixture (final volume of 20 μl) contained binding buffer [25 mM Hepes/KOH (pH 7.9), 50 mM KCl, 5 mM MgCl2, 0.1 mM EDTA, 1 mM DTT (dithiothreitol) and 20% (w/v) glycerol], 1 ng of labelled double-stranded oligonucleotide primers, 3 μg of non-specific competitor [poly(dI–dC) or poly(dA–dT) primers] and varying quantities of protein: GST, GST–CrzA (amino acids 469–603) or E. coli protein extracts for GST–SltA (amino acids 381–549). Various amounts of unlabelled double-stranded oligonucleotides as specific competitors (as indicated in the Figures or the Figure legends) were added to the reaction mixtures before the addition of protein. Reaction mixtures were kept on ice for 30 min, and DNA and DNA–protein complexes were separated by native PAGE (5% gels). Gels were dried and analysed by autoradiography.
AFP susceptibility testing
Determination of the sensitivity of A. nidulans strains to AFP was performed following a method published previously .
Conidia of A. nidulans strains were grown in appropriately supplemented liquid minimal medium placed on to coverslips. Germlings adhering to the coverslips were observed using an Axioskop 50 microscope (Zeiss) equipped with a Camedia Digital Camera C-4000 Zoom (Olympus). Light images were captured with a ×40 objective lens. Images were processed using Adobe Photoshop 6.0 (Adobe). Observations of Aspergillus fluorescent strains were performed in WMM (watch minimal medium) as described previously . The Aspergillus strain expressing CrzA–GFP was cultured on the surface of glass-bottomed Petri dishes immersed in 2.5 ml of appropriately supplemented WMM at 25 °C for 16 h. The medium was replaced by fresh WMM supplemented with or without 10 mM CaCl2, and the green fluorescence was observed using a Leica DMI6000b inverted microscope equipped with a GFP-specific filter (λex, 470 nm and λem, 525 nm), and recorded with an Orca-ER camera (Hamamatsu) driven by the MetaMorph imaging system (Universal Imaging Corporation).
Identification of the crzA gene
The putative A. nidulans homologue of S. cerevisiae CRZ1, AN5726.3, denoted crzA, was identified by Blastp searches using the full-length Crz1p protein sequence. The in silico predicted cDNA sequence of crzA was verified by DNA sequencing of a RT–PCR product obtained from total RNA.
CrzA and Crz1p share 66% identity at the amino acid level within the putative DNA-binding domain, which comprises three zinc fingers, the first two of which conform to the Cys2His2 consensus sequence [(Y/F)XCX2–4CX3(Y/F)X5LX2HX3–4H] (Figure 1A). This degree of sequence conservation suggests that CrzA and Crz1p might bind the same DNA target sequence. The third finger of Crz1p and its homologues, including CrzA, appear to have an atypical Cys2HisCys structure with unusual spacing (Figure 1A) and is currently under study. The 730-amino-acid-residue CrzA protein, in common with its homologues, contains the DNA-binding domain in its C-terminal region. CrzA contains two overlapping putative CDDs (calcineurin-docking domains) (Figure 1B), whereas Crz1p, which is activated by dephosphorylation by calcineurin phosphatase, is thought to contain a single specific recognition sequence PXIXIT [29–31]. Putative Crz1p orthologues are widely distributed amongst the fungi, with sequence conservation notably high among filamentous fungi (Figure 1C and Supplementary Figure S1 at http://www.BiochemJ.org/bj/414/bj414ppppadd.htm). The crzA transcript appears to be expressed at a very low level by Northern blot analysis, but is increased somewhat by exposure to 100 mM CaCl2 for periods of 5–30 min (results not shown). In contrast, no increase in A. fumigatus crzA transcript levels was reported in response to 10 or 60 min exposure to 200 mM CaCl2 . The very modest increase in crzA transcript levels in response to Ca2+ suggests that this regulation might be mainly post-transcriptional, particularly in view of the effect of Ca2+ on CrzA localization (see below).
The phenotype of a crzA-null mutation
To assess the physiological role of CrzA, a strain with crzA deleted was constructed by gene replacement using the N. crassa pyr-4 gene as a selectable marker (see the Experimental section for details). The original transformant and outcrossed ΔcrzA strains are, similarly to the corresponding deletion mutants of A. fumigatus and B. cinerea [13,19], strongly inhibited by elevated Ca2+ concentrations and alkaline pH (Figure 2), thus resembling the behaviour of crz1 mutants [9,32], but to a more extreme extent.
Strains deleted for crzA show delayed germination in minimal medium in which Ca2+ is present only as a trace contaminant (Figure 3A). The presence of 100 mM CaCl2 causes hyphal hyperbranching in ΔcrzA strains (Figure 3B). Similar morphological abnormalities have been noted for the corresponding A. fumigatus and B. cinerea deletion mutants [13,19]. Interestingly, ΔcrzA strains show significant resistance to the toxicity of AFP (Figure 3C), which was recently shown to inhibit chitin biosynthesis, very probably as a result of inhibition of classes III and V chitin synthases . In contrast, ΔcrzA strains show slight hypersensitivities to other cell-wall perturbing compounds such as caspofungin and Congo Red (results not shown), further suggesting a role for CrzA in determining cell-wall structure. Cell-wall abnormalities have also been reported for the orthologous B. cinerea crz1 deletion mutant . The similarities in phenotype between null-crzA and null-chsB A. nidulans strains, which lack a class III chitin synthase [34,35], suggested that chsB might be under the transcriptional control of CrzA. Figure 3(D) shows that whereas transfer into medium containing 100 mM CaCl2 elevates the transcript levels of chsB in a wild-type strain, it fails to do so in a ΔcrzA strain, indicating that up-regulation of chsB in response to Ca2+ is CrzA-dependent. In agreement, the chsB promoter contains two putative CDRE sites (results not shown) conforming to the Crz1p-binding-site consensus sequence . In contrast with the effect of Ca2+, 300 mM LiCl appears to have no effect on chsB transcript levels (Figure 3D).
Another target of CrzA regulation is AN0471.3, denoted vcxAAN0471, which we identified as the closest A. nidulans homologue of S. cerevisiae VCX1 using tBlastn and Blastp searches. In S. cerevisiae, the vacuolar Ca2+/H+ exchanger Vcx1p mediates the import of excess cytosolic Ca2+ into the vacuole . VCX1 transcription is negatively regulated by calcineurin in a Crz1p-independent manner [8,37]. Figure 4(A) shows that vcxAAN0471 transcript levels are increased by 100 mM CaCl2 in a CrzA-dependent manner. As a large number of Ca2+-regulated S. cerevisiae genes show a greater transcriptional response at an early time after alkalinization, which provokes a transient rise in cytosolic Ca2+ levels than at a later time , vcxAAN0471 transcript levels were monitored at both 30 and 60 min following Ca2+ exposure. Indeed, steady-state transcript levels were lower at the later time point (Figure 4A). Expression of VCX1 homologues in A. fumigatus and B. cinerea is also dependent on the respective CRZ1 orthologues [13,19].
Subcellular localization of CrzA
To analyse the subcellular distribution of CrzA, a mutant crzA allele expressing C-terminally GFP-tagged CrzA was constructed by gene replacement using methods described previously [24,25,38]. The protein (CrzA–GFP) encoded by the tagged allele appears to be fully functional, as no Ca2+ sensitivity or growth impairment at an alkaline pH was observable (results not shown). In the absence of Ca2+ addition and any form of stress, CrzA–GFP is exclusively cytosolic, apparently excluded from nuclei (Figure 5). Upon addition of CaCl2 (final concentration of 10 mM), CrzA rapidly accumulates in nuclei, eventually depleting the cytoplasm of fluorescence (Figure 5 and results not shown). Thus, as described previously for the CrzA orthologue Crz1p [6,31,39] (reviewed in ), CrzA appears to be subjected to active nuclear export and import machinery, with an increase in the cytosolic Ca2+ concentration eliciting its nuclear import (Figure 5). Presumably, as in S. cerevisiae , Ca2+ activation of calcineurin results in dephosphorylation of CrzA and a consequent loss of nuclear-export capability. A. fumigatus CrzA also localizes to the nucleus in response to Ca2+ , as does B. cinerea CRZ1 when expressed heterologously in S. cerevisiae .
Further characterization of sltA mutations
In view of the sensitivity of ΔcrzA strains to Ca2+ toxicity, further characterization of mutations in the sltA gene where the sltA1 allele confers sensitivity to the Na+ and K+ cations [20,21,40] seemed warranted. Although SltA homologues are widely distributed in filamentous fungi, we have been unable to detect any in searches of yeast genomes. The sltA gene encodes a 698-amino-acid-residue protein containing three Cys2His2 zinc fingers, and the sltA1 mutation changes Trp502, the second codon following that of the last histidine residue of finger 3, to a stop codon (see Supplementary Figure S2A at http://www.BiochemJ.org/bj/414/bj414ppppadd.htm) .
To determine the phenotype of a null-sltA allele and to see whether sltA1 has a null phenotype, a deletion allele was constructed by gene replacement using the A. fumigatus riboB gene as a selectable maker (see the Experimental section and Supplementary Figure S2B for details). We detected no phenotypic differences between ΔsltA and sltA1, suggesting that at least some portion of the C-terminal 197 amino acid residues is essential to SltA function (results not shown). As shown in Figure 2, the absence of functional SltA results in an elevated sensitivity to Li+, Cs+ and Mg2+, in addition to Na+ and K+, the inability to grow at alkaline pH, and hypersensitivity to neomycin toxicity. In plate tests, ΔcrzAΔsltA double mutants have phenotypes which are the sum of the single mutant phenotypes (Figure 2).
Although overall conservation of the C-terminal sequence beyond the third zinc finger of SltA is low among putative homologues, Trp502 and the residues immediately downstream from it are highly conserved, suggesting that this region might be required for full functionality (Supplementary Figure S2A). To test this possibility, revertants of a sltA1 strain were obtained, after UV mutagenesis, as those which allowed growth in the presence of 1 M NaH2PO4. As well as reversion in codon 502 to the wild-type tryptophan codon, partially functional revertants encoding arginine, leucine, phenylalanine, lysine or cysteine at codon 502 were also obtained (Table 2). Unsurprisingly, reversion to a phenylalanine residue appears to result in the greatest degree of functionality. These reversion results indicate that Trp502 is itself an important residue and that one or more additional important residues is/are located C-terminal to it.
SltA acts negatively in the control of vcxAAN0471 and positively in the control of enaAAN6642
The ΔsltA mutant showed an elevated response in vcxAAN0471 transcript levels in the presence of Ca2+ at both 30 and 60 min, and the response is more pronounced at the latter time point (Figure 4B, and results not shown). However, ΔcrzA is epistatic to ΔsltA in controlling vcxAAN0471, as with the ΔcrzAΔsltA double mutant there is no response in vcxAAN0471 transcript levels (Figure 4B) in the presence of Ca2+, which correlates with the fact that Ca2+ inhibition of the ΔcrzAΔsltA double mutant is very similar to that of the ΔcrzA single mutant (Figure 2).
Using S. cerevisiae Ena1p, Ena2p and Ena5p in Blastp and tBlastn searches of the A. nidulans genome (http://www.broad.mit.edu/annotation/genome/aspergillus_group/Blast.html), three loci whose predicted protein products are closely related to the Ena P-type ATPase Na+ pumps, AN6642.3 (designated enaA ), AN1628.3 and AN10982.3, were identified (results not shown). We have designated these genes enaAAN6642, enaBAN1628 and enaCAN10982. In S. cerevisiae, the Ena transporters are involved in the salt-stress response and are required for growth at alkaline pH [42,43].
Expression of enaAAN6642 requires a combination of elevated concentrations of Na+ and alkaline pH (Figure 6). We have been unable to detect transcripts for enaBAN1628 and enaCAN10982 under these conditions or the genetic backgrounds shown in Figure 6. Neither 300 mM LiCl nor 100 mM CaCl2 is sufficient to induce expression of any of the ena genes (results not shown). Figure 6 shows that SltA plays a significant positive-acting role in enaAAN6642 expression; however, if CrzA plays a role, it is much more marginal.
CrzA and SltA are both DNA-binding proteins
The putative DNA-binding domains, containing the entire zinc-finger regions, amino acid residues 469–603 for CrzA and amino acid residues 381–549 for SltA were expressed in E. coli as GST fusion proteins. Although purification of GST–CrzA (amino acids 469–603) using glutathione-affinity chromatography was feasible, we did not succeed in purifying GST–SltA (amino acids 381–549) despite trying several elution conditions (results not shown). Consequently, bacterial crude protein extracts were used in DNA-binding studies.
To characterize the DNA-binding ability of GST–CrzA (amino acids 469–603), probes conforming to the S. cerevisiae Crz1p consensus-site CDRE  were synthesized (Figure 7A). Two putative CDRE sites were identified in the A. nidulans chsB promoter, and five such sites were identified in the A. giganteus afp promoter, previously suggested to be under the control of the Ca2+–calcineurin signalling pathway . CrzA binds to CDRE-1 from the chsB promoter and to CDRE-5 from the afp promoter (Figures 7B and 7C), confirming that binding sequences for CrzA and Crz1p are very similar or identical.
Because of the sequence similarity of SltA to T. reesei ACE1 in the zinc-finger region , the ability of GST–SltA (amino acids 381–549) to bind to the 2C and mutant 2C4 probes described previously  was tested. Similarly to ACE1 binding, SltA binds to the 2C but not the 2C4 probe (Figure 7D). SltA also binds to the SLTA-1 site of the afp promoter, which conforms to the AGGCA ACE1 consensus site (Figures 7A, 7D and 7E).
As the trinucleotide GGC is common to sequences bound by CrzA and SltA, we tested whether either of the fusion proteins might bind to a site bound by the other. However, we found no evidence for binding of SltA to afp CDRE-5 (Figure 7D) or of CrzA to afp SLTA-1 (Figure 7E).
The two putative transcription factors investigated here are both involved in preventing the toxicity of certain cations. In the absence of functional SltA, A. nidulans is abnormally sensitive to toxicity by Na+, K+, Li+, Cs+ and Mg2+, but not by Ca2+, whereas the absence of CrzA results in the toxicity of Ca2+. Transcript levels of the putative vacuolar Ca2+/H+ exchanger gene vcxA in response to high Ca2+ levels provides a possible explanation for the Ca2+ toxicity to ΔcrzA strains and the lack of Ca2+ toxicity in ΔsltA strains. Whereas ΔsltA strains show an elevated response to Ca2+ in vcxA transcript levels, putatively allowing greater vacuolar storage and thus detoxification of Ca2+, ΔcrzA strains apparently show no Ca2+ response at all, putatively resulting in elevated cytosolic Ca2+ levels. The VcxA homologue in S. cerevisiae, Vcx1p, mediates the vacuolar import of excess cytosolic Ca2+ . One possible factor in cation toxicity in sltA-null strains is a reduction in the induction of the putative cation exporter EnaAAN6642, as the homologous S. cerevisiae plasma-membrane exporter Ena1p enables a detoxification response to Na+, Li+ and alkaline conditions [42,43].
The presence of three putative SltA/ACE1 binding sites 5′ to the enaA coding region and one such site 5′ to the vcxA coding region is consistent with SltA regulation of the transcription levels of these genes. Five putative CDRE sites lie upstream of the vcxA coding region, consistent with CrzA transcriptional regulation of vcxA. One CDRE site lies upstream of the enaA coding region, but, due to a lack of solid evidence that CrzA regulates enaA, its significance, if any, is unclear. Further work will be necessary in order to determine the physiological relevance of these 5′ sites.
This work has highlighted significant differences between A. nidulans and S. cerevisiae. S. cerevisiae lacks an identifiable SltA homologue and, while S. cerevisiae Crz1p is not involved in VCX1 expression, but plays a crucial role in ENA1 expression , A. nidulans CrzA plays a crucial role in vcxA expression (similar to findings in A. fumigatus and B. cinerea), but is hardly involved, if at all, in enaAAN6642 regulation. The role of CrzA in Ca2+-induced chsB expression has drawn attention to a parallel situation in C. albicans, where Ca2+ stimulation of the promoter activity for four different chitinase genes is Crz1p-dependent .
A further point of interest is that SltA appears to play both a positive (in enaAAN6642 regulation) and a negative (in vcxA regulation) role, with the caveat that we have not yet shown that SltA acts directly to regulate the genes in question. The PacC transcription factor, which mediates control of gene expression by ambient pH in A. nidulans, acts as both an activator and a repressor , but its S. cerevisiae homologue Rim101p exerts positive regulation by repressing transcription of genes encoding other transcriptional repressors . Finally, we note that at least three transcription factors are necessary for growth of A. nidulans at alkaline pH: PacC , and, as established here, SltA and CrzA. In the case of SltA, at least one pertinent factor is its role in enaAAN6642 expression.
We thank Barbara Walewska, Susanne Engelhardt, Tatiana Munera Huertas and Elena Reoyo for technical assistance, and Dr Berl Oakley (Department of Molecular Genetics, Ohio State University, Columbus, OH, U.S.A.) for plasmid p1548 containing the A. fumigatus riboB gene. We are grateful to the Biotechnology and Biological Sciences Research Council (BBSRC) for support through grant BB/D52178/1 to H. N. A., and to the Spanish Ministerio de Educación y Ciencia (D. G. I. C. Y. T.) for support through grant BFU2006-04185 to E. A. E. L. A.-B. is a pre-doctoral fellow of the PFPU of the Spanish Ministerio de Educación y Ciencia.
Abbreviations: AFP, antifungal protein; CDRE, calcineurin-dependent response element; GFP, green fluorescent protein; GST, glutathione transferase; ORF, open reading frame; RT–PCR, reverse transcription–PCR; UTR, untranslated region; WMM, watch minimal medium
- © The Authors Journal compilation © 2008 Biochemical Society