Signaling of ambient pH in Aspergillus involves a cysteine protease.

In Aspergillus nidulans, the regulation of gene expression in response to changes in ambient pH is mediated by the PacC zinc finger transcriptional regulator. At alkaline ambient pH, PacC is proteolytically processed to a functional form serving as an activator of alkaline-expressed genes and a repressor of acid-expressed genes. This activation of PacC occurs in response to a signal mediated by the products of the pal genes. Thus, the products of the palA, -B, -C, -F, -H, and -I genes constitute an alkaline ambient pH signal transduction pathway. How the pal signal transduction pathway senses ambient pH and transduces a signal to trigger PacC processing is a fascinating unresolved problem. We have cloned and sequenced the palB gene. The predicted palB gene product has similarity to the catalytic domain of the calpain family of calcium-activated cysteine proteases. We have shown, however, that the PalB protein does not catalyze the final step of proteolytic processing of PacC.

In Aspergillus nidulans, the regulation of gene expression in response to changes in ambient pH is mediated by the PacC zinc finger transcriptional regulator. At alkaline ambient pH, PacC is proteolytically processed to a functional form serving as an activator of alkalineexpressed genes and a repressor of acid-expressed genes. This activation of PacC occurs in response to a signal mediated by the products of the pal genes. Thus, the products of the palA, -B, -C, -F, -H, and -I genes constitute an alkaline ambient pH signal transduction pathway. How the pal signal transduction pathway senses ambient pH and transduces a signal to trigger PacC processing is a fascinating unresolved problem. We have cloned and sequenced the palB gene. The predicted palB gene product has similarity to the catalytic domain of the calpain family of calcium-activated cysteine proteases. We have shown, however, that the PalB protein does not catalyze the final step of proteolytic processing of PacC.
The ascomycete Aspergillus nidulans, like many other microorganisms, grows over a wide pH range (1,2). Crucial to this physiological versatility, A. nidulans has a regulatory system to ensure that gene products that function beyond cell boundaries, such as permeases and extracellular enzymes, are only synthesized at pHs at which they can function effectively so that, for example, acid phosphatase is secreted in acidic environments and alkaline phosphatase in alkaline environments (1,3,4). Mutations in the pacC and palA, -B, -C, -F, -H, and -I genes alter the response to ambient pH. The PacC zinc finger transcriptional regulator directly regulates expression of genes under pH control, and the products of the pal genes constitute a signal transduction pathway which, at alkaline ambient pH, triggers the conversion of PacC to a functional form as a transcriptional activator of alkaline-expressed genes and repressor of acid-expressed genes (1,(3)(4)(5). The absence of the pal signal at acid pH results in derepression of acid-expressed genes and lack of activation of alkaline-expressed genes. We are characterizing the genes of this novel ambient pH signal transduction pathway. We have cloned the palB gene and shown that the predicted PalB protein has sequence similarity to the catalytic domain of the calpain family of calcium-activated cysteine proteases. Thus, a cysteine protease is a component of the alkaline ambient pH signal transduction pathway in Aspergillus.

MATERIALS AND METHODS
A. nidulans Strains and Media-A. nidulans strains carried markers in standard use (6). Standard media were used (1,3,7).
palB Gene Disruption-A 3.2-kb 1 BglII restriction fragment containing the pyr4 gene of Neurospora crassa, which complements the pyrG89 mutation of A. nidulans was cloned into the BamHI site of the plasmid pUC19.18, containing the SphI/KpnI fragment shown in Fig. 1B, to produce pUC19.18⌬BamHI. This plasmid was linearized with SphI prior to transformation of a strain of genotype yA2 pabaA1 pyrG89. Genomic DNAs of pyr ϩ transformants that were unable to grow on pH 8 medium were checked by Southern analysis for the predicted disruption restriction fragment pattern (data not shown). One verified disruption strain was used for meiotic analysis and to construct diploid strains for dominance and complementation analysis.
Sequence Analysis-The palB genomic sequence was determined on both strands by double-stranded sequencing of plasmids constructed in pUC19 using universal and gene-specific primers. The U. S. Biochemical Corp. Sequenase system was used. To identify introns, cDNA clones and PCR fragments generated by reverse transcriptase PCR were sequenced with the U. S. Biochemical Corp. system and with an automated sequencer.
Gel Mobility Shift Experiment-The gel mobility shift experiment was done using growth conditions and mycelial extracts as described (4). RESULTS We cloned the palB gene using linkage group VIII clones from a chromosome-allocated, wild type A. nidulans cosmid library (8). To identify palB ϩ transformants, we used the inability of pal mutants to grow on alkaline pH media (1). The cosmid clone pWE15 03H12 rescued the palB7 strain to give a pal ϩ phenotype. This cosmid clone was shown by hybridization to overlap cosmid clones containing the ivoB gene, which is closely linked (2.1 centimorgans (11)) to palB (data not shown). A 6-kb BamHI restriction fragment of the cosmid was then found to rescue the palB7 strain (Fig. 1A). When this BamHI fragment was inserted into the argB ϩ -containing plasmid pILJ16 and the resulting plasmid (pILJ16.1) used for transformation, all pal ϩ transformants analyzed had BamHI fragmentdirected homologous integration (data not shown). One transformant, showing homologous integration, was crossed to an argB2 strain. All 104 progeny analyzed were palB ϩ , and the argB ϩ and chaA1 markers showed 8% recombination, consistent with the linkage (12) between palB and chaA. This and the result that the larger, overlapping SphI restriction fragment rescued at a higher frequency suggested that the entire palB gene might not be contained within the BamHI fragment and that the BamHI site within the SphI fragment might be within the palB gene (Fig. 1). This BamHI site was therefore used for gene disruption (Fig. 1B). Strains disrupted at the BamHI site are unable to grow on pH 8 medium (Fig. 1B). Also, the reces-sive disruption allele is non-complementing in a diploid with palB7 (Fig. 1B). When a disruptant was crossed to a tubA1 pantoA10 strain, analysis of 104 progeny located the palB disruption mutation between tubA and pantoA, consistent with the map position (6) of palB. These results confirmed that palB had been cloned and that the BamHI site is within the gene.
The sequence for 3.2 kb of the palB-containing genomic region is shown in Fig. 2. This sequence spans the BamHI site within palB. The 3-kb XbaI/KpnI restriction fragment indicated by the striped bar in Fig. 1 and the 700-base pair EcoRI/ BamHI restriction fragment contained within it were used to probe Northern blots of wild type A. nidulans and to screen a cDNA library. Both probes hybridized to a message of approximately 3 kb. The boundaries of the largest cDNA clone isolated are shown by arrows in Fig. 2. We believe this cDNA clone contains all of the palB coding region for the following reasons. The clone contains a poly(A) tail, and all three reading frames upstream of the putative start codon are closed multiple times in the cDNA. Also, the size of this cDNA clone is close to our estimate of the mRNA size. This cDNA clone, a shorter cDNA clone, and reverse transcriptase PCR fragments were sequenced to determine the position of introns (Fig. 2). The resulting open reading frame would encode a protein of 842 amino acids and a molecular mass of 93.9 kDa. Sequence data base searches with this open reading frame using the FASTA program (13) retrieve members of the calpain family of calcium-activated cysteine proteases. Strong sequence similarity between PalB and the calpains is limited to the catalytic region. A multiple sequence alignment of the catalytic region of several calpains and PalB is shown in Fig. 3. In addition to the Cys, His, and Asn residues believed to constitute the calpain catalytic triad (14), starred in Fig. 3, PalB contains many conserved residues within the catalytic domain. In a pairwise comparison with human calpain p94, PalB has 29.1% amino acid identity over 300 amino acids and 32.3% identity over 189 amino acids of the catalytic domain. As indicated in Fig. 3, the BamHI site used for disruption separates the codon for the Cys catalytic residue from those for the His and Asn catalytic residues.
Activation of PacC in response to alkaline pH requires PacC proteolysis (4). A ϳ29-kDa form of PacC, which is functional as a transcriptional activator and repressor, predominates at alkaline pH in wild type strains and irrespective of pH in pacC c mutant strains, which mimic the effects of growth at alkaline pH. This 29-kDa form of PacC migrates as complex I in gel mobility shift experiments using a promoter fragment of the PacC-regulated ipnA (isopenicillin N synthetase) gene and extracts of a wild type strain grown under alkaline conditions (Fig. 4, lane 3) or an alkalinity-mimicking pacC c 202 strain grown under acidic conditions (Fig. 4, lane 4). The full-length (73 kDa) form of PacC, which is inactive as a transcriptional activator and repressor, migrates more slowly as complex II and predominates in gel mobility shift assays with extracts of wild type strains grown under acidic conditions (4) or palB7 strains grown under acidic or neutral conditions (Fig. 4, lanes  5 and 6). Additionally, pal mutants, including palB7 strains, produce greatly reduced amounts of PacC DNA binding activity (Fig. 4, lanes 5 and 6 and Ref. 4). We have shown previously that pacC c mutations are phenotypically epistatic to mutations in the pal signal transduction pathway genes (1,3,5), and the expectation would therefore be that pacC c epistasis would extend to PacC processing. This prediction is testable because proteolytic processing of PacC to yield the 29-kDa form occurs in alkalinity-mimicking pacC c strains in which the primary translation product is already truncated to some extent by pacC c mutations (4). The prediction has been previously con-firmed in the case of a mutation in palA (4). Despite the phenotypic epistasis of pacC c mutations to palB mutations (1, 3), it was necessary to determine experimentally whether processing of PacC to the 29-kDa form is affected by a palB mutation because PalB appears to be a cysteine protease. PalB is clearly not responsible for this processing, as the fully proteolyzed form of PacC predominates at acid pH in gel mobility shifts using extracts of a strain carrying both pacC c 202 and palB7 mutations (Fig. 4, lanes 7 and 8). DISCUSSION We have shown that a protein with sequence similarity to calpains is a component of the alkaline ambient pH signal transduction pathway in Aspergillus. Calpains contain C-terminal EF-hand structures which are responsible for calcium binding and regulation. PalB, like the Drosophila melanogaster Sol (small optic lobes protein) has similarity to the catalytic domain of calpains but not to these calcium-binding domains (15). However, PalB and Sol both have similarity to a putative EF-hand structure located immediately C-terminal to the catalytic domain (overlined in Fig. 3). Some putative EF-hand structures in known calpains deviate from the EF-hand consensus (16,17). Allowing for such deviations, there are also several possible EF-hands in the C-terminal region of PalB.
As hypothesized for the Sol protein (15), PalB might have substrate specificity similar to that of calpains but not be calcium-regulated. In vivo and in vitro substrates for calpains include cytoskeletal proteins, transcription factors, protein kinases, and protein phosphatases (18 -23). Protein kinase C is proteolyzed to a permanently active form by calpain (19). Certain of these substrates are suggestive of signal transduction roles for calpains. Also suggestive of a calpain signal transduction role is the nuclear location of a muscle-specific calpain, which is mutated in one form of muscular dystrophy (24,25).
PalB is required for growth at alkaline but not acidic pH. Although the palB mRNA is in low abundance, we have determined that there is little or no difference in its levels after growth in the pH range 4 -8 (data not shown). The final proteolytic processing step for PacC is not catalyzed by PalB. The pal pathway is, however, responsible for a PacC modification that is required for its conversion to a functional form at alkaline pH (3,4). The alkalinity-mimicking pacC c 202 mutation removes 214 C-terminal residues from PacC (3). This deletion allows proteolytic processing of PacC to the 29-kDa functional form in the absence of the pal signal at acidic pH or in a pal mutant background (Fig. 4 and Ref. 4). The pacC c 202 and other pacC c mutation C-terminal truncations might destabilize intramolecular interactions, allowing proteolysis of PacC to the active form (4). In wild type strains, the pal pathway is thought to introduce a modification of PacC at alkaline pH, disrupting intramolecular interactions to allow activating proteolysis (4). The type of PacC modification mediated by the pal pathway in response to alkaline pH is not known, but an earlier (and more C-terminal) proteolytic cleavage of PacC resulting in susceptibility to further proteolysis might be mediated by PalB. Alternatively, PalB might proteolyze one of the other pal gene products in a signaling cascade. Sequencing of other pal genes and biochemical studies of PalB might help identify the PalB substrate(s).
FIG. 4. PalB is not responsible for the final proteolysis converting PacC to the functional form. Electrophoretic mobility shift assays were performed using a 31-base pair promoter fragment of the PacC-regulated ipnA (isopenicillin A synthetase) gene containing the high affinity ipnA2 PacC-binding site and extracts of various strains as described (4). Lane 1, oligonucleotide alone; lanes 2 and 3, 5 g of protein of wild type (biA1) grown in acidic and alkaline conditions, respectively; lane 4, 5 g of protein of the pabaA1 pacC c 202 strain grown in acidic conditions; lanes 5 and 6, 10 g of protein of the pabaA1 palB7 strain grown in acidic and neutral conditions, respectively (palB7 strains do not grow in alkaline conditions); lanes 7 and 8, 2.5 and 5 g of protein, respectively, of the strain inoB2 pacC c 202 chaA1 palB7 grown in acidic conditions. Solid arrows indicate the high mobility complex I, formed with the functional 29-kDa proteolyzed version of PacC, and the low mobility complex II, formed with the full-length, non-functional version of PacC. Because pacC c 202 results in a substantially shorter full-length version of PacC it increases the mobility of complex II (3,4). The open arrow indicates a complex probably formed by a proteolysis degradation product of the 29-kDa functional version of PacC. As seen from lanes 7 and 8, the level of this putative degradation product is not diminished (and might even be elevated) by the palB7 mutation. The slight amount of complex I formed by neutral grown palB7 extracts (lane 6 and Fig. 7 of Ref. 4) also forms when extracts of a neutral grown palB disruption strain are used (data not shown), indicating that very limited PacC processing can occur in the total absence of PalB. Thus, although the sequence change resulting from the palB7 mutation has yet to be determined, its physiological and biochemical phenotype appears to be that of a null allele.