XprG, a putative p53-like transcriptional activator, regulates production of extracellular proteases in response to nutrient limitation and may also have a role in programmed cell death. To identify genes that may be involved in the XprG regulatory pathway, xprG2 revertants were isolated and shown to carry mutations in genes which we have named sogA-C (suppressors of xprG). The translocation breakpoint in the sogA1 mutant was localized to a homolog of Saccharomyces cerevisiae VPS5 and mapping data indicated that sogB was tightly linked to a VPS17 homolog. Complementation of the sogA1 and sogB1 mutations and identification of nonsense mutations in the sogA2 and sogB1 alleles confirmed the identification. Vps17p and Vps5p are part of a complex involved in sorting of vacuolar proteins in yeast and regulation of cell-surface receptors in mammals. Protease zymograms indicate that mutations in sogA-C permit secretion of intracellular proteases, as in S. cerevisiae vps5 and vps17 mutants. In contrast to S. cerevisiae, the production of intracellular protease was much higher in the mutants. Analysis of serine protease gene expression suggests that an XprG-independent mechanism for regulation of extracellular protease gene expression in response to carbon starvation exists and is activated in the pseudorevertants.
THE extracellular proteases of the filamentous fungus Aspergillus nidulans can be used as a model to study the response to starvation as production of extracellular proteases is repressed in nutrient-sufficient growth medium and is stimulated by the absence of a carbon, nitrogen, or sulfur source, irrespective of whether protein is present (Cohen 1973a). The xprG gene plays a major role in the regulation of extracellular proteases. Loss-of-function mutations in xprG abolish the production of extracellular proteases in response to carbon starvation and the response to nitrogen and sulfur starvation is reduced (Katz et al. 2006). In addition, production of an acid phosphatase in response to phosphate starvation is lost in xprG− mutants indicating that XprG may be involved in a general response to starvation (Katz et al. 2006). A single gain-of-function mutation in xprG has been identified. Strains carrying the xprG1 mutation have increased extracellular protease activity in response to carbon and, to a lesser extent, nitrogen starvation (Katz et al. 1996). The xprG1 mutation results in an amino acid substitution (R186W) in the putative DNA-binding domain of XprG.
XprG belongs to a newly identified group of proteins that contains an Ndt80-like DNA-binding domain and belongs to the family of p53-like transcription factors (pfam.sanger.ac.uk/). Ndt80 activates the transcription of >150 genes during middle meiosis in Saccharomyces cerevisiae (Chu et al. 1998). Nutrient limitation is required for the initiation and completion of meiosis in S. cerevisiae (Esposito and Klapholz 1981). The ndt80 mutants arrest at the point in meiosis past which progression through meiosis cannot be reversed by nutrient supplementation (Xu et al. 1995). Thus Ndt80 may, like XprG, be involved in sensing nutrient status (Katz et al. 2006). Mutations in xprG do not affect meiosis.
Genetic evidence suggests that two noncatalytic hexokinase-like proteins (HxkC and HxkD) are components of the XprG regulatory pathway (Katz et al. 2000; Bernardo et al. 2007). Loss-of-function mutations in hxkC and hxkD produce a similar phenotype to the xprG1 gain-of-function mutation and are suppressed by xprG− mutations. HxkD is a nuclear protein whereas HxkC is associated with mitochondria (Bernardo et al. 2007). The binding of hexokinase to mitochondria blocks apoptosis in human tumor cells (Pastorino et al. 2002) and regulates programmed cell death in plants (Kim et al. 2006). The similarity between XprG and Vib-1, which is required for expression of genes during programmed cell death in Neurospora crassa (Dementhon et al. 2006), suggests that HxkC and XprG may have a role in regulating programmed cell death as well as the response to nutrient depletion in A. nidulans.
AreA is the transcriptional activator that mediates nitrogen metabolite repression in A. nidulans (Kudla et al. 1990). Though XprG is involved in controlling the production of extracellular proteases in response to nitrogen limitation, AreA function is also required (Katz et al. 2006). The genetic data suggests that XprG may modulate AreA-mediated activation of extracellular protease gene expression (Katz et al. 2006). The aim of this study was to investigate whether a similar situation exists with respect to carbon regulation (i.e., Does XprG modulate the activity of a transcription factor that mediates the response to carbon starvation?). The creA gene encodes a transcriptional repressor, which controls carbon catabolite repression in A. nidulans (Dowzer and Kelly 1991). Extracellular protease activity is increased in strains lacking CreA but the effect is XprG dependent, suggesting that XprG acts downstream of CreA (Katz et al. 2008). A genetic approach was then used to try to identify proteins that might be involved in the response to carbon starvation in partnership with XprG. We describe the isolation and characterization of suppressor mutations that restored production of extracellular proteases in response to carbon starvation in an xprG null mutant.
MATERIALS AND METHODS
Aspergillus strains, media, growth conditions, genetic techniques, and transformations:
The Aspergillus nidulans strains used in this study are listed in Table 1. The media used for culture of A. nidulans was as described previously (Katz et al. 1996). The techniques used for genetic manipulation of A. nidulans have been described (Clutterbuck 1974). A. nidulans transformations were performed using the method of Tilburn et al. (1983).
Isolation of mutants:
Conidial suspensions of strain MK198 were irradiated with ultraviolet light (254 nm) as described previously (Katz et al. 1996). Mutagenized conidia were diluted and spread on minimal medium containing 1% skim milk as a carbon source to screen for revertants that produced halos.
Cloning of the sogA and sogB genes:
To clone the sogA gene, nucleotides 154,070 to 156,746 (containing the A. nidulans homolog of VPS5) on contig 1.61 were amplified with Taq DNA polymerase using primers MK253 (5′-GTAGAGATGCATAAGCTTGG-3′) and MK256 (5′-TCAAGAATTCACGACTATGGAGCGAAG-3′). The amplified DNA was digested with EcoRI and NsiI and inserted into pUC18 (Yanisch-Perron et al. 1985). Five different vps5 clones were introduced into the pabaA1 sogA1 xprG2 strain (MK308) by cotransformation with the pabaA plasmid, pPABA-AT6 (R. B. Todd, unpublished results). Paba+ transformants were tested on medium containing milk as a carbon source. Transformants that exhibited a sogA+ xprG2 phenotype were analyzed by Southern blot analysis.
To isolate the sogB gene, nucleotides 5832–9032 (containing the A. nidulans homolog of VPS17) on contig 1.36 were amplified with Taq DNA polymerase using primers MK285 (5′-ACACTGCGCCATCGAGAATG-3′) and MK286 (5′-GATGGTTCAAATGCATACGC-3′). The amplification product was introduced into the pabaA1 sogB1 xprG2 strain (MK310) by cotransformation and analyzed as described for the sogA gene.
Nucleic acid extraction, blotting, and hybridization:
RNA was prepared using the RNeasy Plant Mini kit according to the manufacturer's instructions (QIAGEN). Northern blots were hybridized with 32P-labeled probes prepared using the Prime-a-Gene labeling system (Promega). Genomic DNA was extracted from A. nidulans by the method of Andrianopoulos and Hynes (1988). Southern blot hybridizations were performed using the DIG nonradioactive DNA labeling and detection system (Roche Applied Science).
Protease enzyme assays:
The enzyme assays used to measure protease activity in A. nidulans growth medium were performed as described previously (Katz et al. 1996). The results were expressed in arbitrary units (total absorbance units per gram dry weight of mycelium).
Gel electrophoresis and detection of protease activity:
Nondenaturing polyacrylamide gel electrophoresis and detection of protease activity using a 1% agarose milk overlay was performed as described by Katz et al. (1996). For the analysis of the effect of protease inhibitors, the gels were adjusted to pH 5.3 after electrophoresis and then incubated at 37° in 0.15 m sodium acetate buffer (pH 5.3) containing the inhibitor for 60 min on a shaking platform prior to the application of the agarose milk overlay.
For detection of extracellular proteases, samples of filtered culture medium that contained 1 OD440 from the control strain, MH2, and an equivalent volume (adjusted for differences in mycelial dry weight) of medium from the mutant strains were prepared as described by Katz et al. (2008). For detection of intracellular proteases, powdered lyophilized mycelium was suspended in H2O at a concentration of 10 mg/ml and frozen. After thawing, the mycelial debris was removed by centrifugation. An equal volume of sample buffer (40% glycerol 0.3 m Tris–HCL pH 6.8) and 5 μl 0.1% bromophenol blue were added to duplicate 25-μl aliquots and were loaded in adjacent wells of the gel.
Isolation and genetic characterization of xprG2 pseudorevertants:
The xprG2 allele contains a frameshift mutation in the ninth codon of the predicted open reading frame. To identify suppressors of the xprG2 mutation, conidia from an xprG2 strain (MK198) were irradiated with ultraviolet light and spread on medium containing milk as a carbon source. Four pseudorevertants (MK307–MK310) that produced halos were isolated (Figure 1). Protease enzyme assays showed that the response to carbon starvation was partially restored but there was no difference in extracellular protease levels in response to nitrogen or sulfur starvation in the four pseudorevertants, when compared to the xprG2 strain (MK198) from which they were derived (Figure 2). The final pH of the carbon-free culture medium was the same (pH 6.4) for xprG+, xprG2, and pseudorevertant strains. Thus, the increase in protease activity detected in pseudorevertants was not due to a modification of the pH of the culture medium.
The four pseudorevertants displayed similar altered colony morphology, compact growth, and reduced conidiation. Although the pseudorevertants produced halos on media containing milk as a carbon or nitrogen source, the ability of the pseudorevertants to utilize protein (1% BSA or 1% casein) as a carbon or nitrogen source was not significantly increased. Outcrosses showed that each pseudorevertant carried an extragenic suppressor of the xprG2 mutation. The altered colony morphology, compact growth, and reduced conidiation of the segregants carrying the suppressor mutations in xprG+ and xprG2 genetic backgrounds were similar.
Diploids (MK314–MK317) constructed from the four pseudorevertants and the mapping strain MSF for haploidization analyses exhibited normal colony morphology indicating that the mutations were recessive. The suppressor mutation in xprG2rev3 was located on chromosome II and the mutation in xprG2rev4 was located on chromosome VIII. The mutation in xprG2rev2 was due to a translocation involving chromosomes II and VII and xprG2rev1 carried a rearrangement involving three chromosomes (I, VI, and VIII).
The chromosome II translocation breakpoint in xprG2rev2 and the suppressor mutation in xprG2rev3 were both mapped to within 5 MU of creB. Phenotypic analysis of a diploid (MK387), constructed to test for complementation, showed pseudorevertants 2 and 3 carry allelic suppressor mutations, designated sogA1 (suppressor of xprG) and sogA2, respectively. The xprG2rev4 mutation (designated sogB1) was mapped to 2.1 MU from methE and 10 MU from gatA on chromosome VII. The chromosomal rearrangement in xprG2 pseudorevertant 1 involved chromosomes I, VI, and VIII, whereas sogA was located on chromosome II and sogB on chromosome VII. Thus the suppressor mutation in xprG2rev1 was located in a third gene, sogC.
Identification of the sogA gene:
The genetic mapping data and colony morphology of the xprG2 pseudorevertants suggested that sogA1 and sogA2 might be alleles of acrB, which is also linked to creB (Boase et al. 2003). Three of the four pseudorevertants exhibited increased sensitivity to molybdate similar to acrB− mutants (Figure 1). However, transformation experiments indicated that acrB did not complement the sogA1 mutation. To confirm that sogA1 was not an allele of acrB, p4AcrB was used to probe Southern blots of genomic DNA from sogA1 and sogA+ strains. The results showed that p4AcrB hybridized to fragments that contained the sogA1 translocation breakpoint but that the translocation breakpoint was located in AN3594, rather than in acrB (Figure 3, data not shown). AN3594 is a homolog of the S. cerevisiae VPS5 gene, which encodes a sorting nexin required for the separation of secreted and vacuolar proteins in the late Golgi (Nothwehr and Hindes 1997). To confirm that AN3594 was sogA, clones containing AN3594 were used in cotransformation experiments with xprG2rev2 (MK308) to test whether AN3594 could complement the sogA1 mutation. A plasmid containing the pabaA gene, pPABA-AT6 (R. B. Todd, unpublished results), was used as a selectable marker. Of 23 paba+ transformants, three displayed a phenotype indistinguishable from the sogA+xprG2 strain MK198 (Figure 4). Southern blot analysis confirmed that these three transformants carried the AN3594 plasmids whereas three paba+ transformants that resembled xprG2rev2 did not contain the plasmids (data not shown).
To confirm that xprG2rev3 also carries a mutation in the VPS5 homolog, AN3594, the coding region of the gene was amplified from MK309 and its DNA sequence was determined. A nonsense mutation was identified in codon 383 (supplemental Figure S1).
Identification of the sogB gene:
In S. cerevisiae Vps5p forms a complex with Vps17p (Horazdovsky et al. 1997). An A. nidulans homolog of VPS17 (AN2224) is on chromosome VII in the same position that genetic mapping data indicated as the location of sogB (Figure 3). A similar strategy to that employed for sogA was used to confirm that AN2224 could complement the sogB1 mutation (Figure 4).
The coding region of AN2224 gene was amplified from MK310 to identify the sogB1 mutation. Sequence analysis revealed the presence of a nonsense mutation in codon 262.
Molecular characterization of sogA and sogB:
The intron/exon structure of the AN3594 (sogA) and AN2224 (sogB) coding regions were confirmed by RT–PCR and DNA sequence analysis. The annotation of AN2224 proved to be correct (www.broad.mit.edu/annotation/fungi/aspergillus/index.html) but the second intron in AN3594 was in the wrong position (supplemental Figure S1). The corrected 567-amino-acid SogA sequence shows 27% amino acid identity with S. cerevisiae Vps5p and 49% identity in the 115 amino acid PX (phosphoinositide-binding) domain (http://pfam.sanger.ac.uk/family?acc=PF00787). In contrast, there is no difference in the overall similarity of the 574-amino-acid SogB sequence to Vps17p (38% identity) and the PX domain (35% identity). The nonsense mutations in sogA2 (codon 383) and sogB1 (codon 262) are downstream from the sequences encoding the PX domains of SogA (codons 161–275) and SogB (codons 115–229). Thus, it is possible that sogA2 and sogB1 are not null mutations. The translocation breakpoint in sogA1 is between codons 246 and 294 and, therefore, may also not disrupt the PX domain.
Within the S. cerevisiae VPS5 gene there is a gene (VAM10) on the complementary strand that is required for vacuole morphogenesis (Kato and Wickner 2003). No open reading frame encoding a protein similar to Vam10p was found in sogA.
Expression of the serine protease gene prtA in the revertants:
Loss-of-function mutations in xprG abolish production of extracellular proteases in response to carbon starvation. Northern blot analysis was performed to determine whether extracellular protease gene expression is restored in the four revertants (Figure 5). The blots were probed with the prtA gene, which encodes an extracellular serine protease that is responsible for most of the extracellular protease activity detected in A. nidulans (Katz et al. 1994; vanKuyk et al. 2000). In the xprG+ [wild-type (WT)] strain high levels of prtA transcript were produced during carbon starvation but no transcript was detected in the xprG2 mutant. In the four revertant strains, the prtA transcript was detected but the level was much lower than in the wild-type strain. In the northern blot of RNA produced during nitrogen starvation, high levels of the prtA transcript were observed in the wild-type strain and a faint band was detected in the xprG2 mutant. None of the xprG2 revertants showed higher levels of prtA transcript than the xprG2 mutant indicating that the sog mutations do not restore the response to nitrogen starvation.
Secretion of intracellular proteases:
Mutations in the S. cerevisiae VPS5 gene result in secretion of an intracellular protease that is normally located within the vacuole (Robinson et al. 1988). The proteases found in the culture medium of an A. nidulans xprG+ strain (MH2) during carbon starvation were characterized using protease inhibitors and nondenaturing gel electrophoresis. Six bands of proteolytic activity were observed in culture filtrate from carbon starvation conditions (Figure 6A, lane E). Metal chelating agents (EDTA, 10-o-phenanthroline and DTT) inhibit the proteolytic activity of Band 1 indicating that this band results from the activity of a metalloprotease (Table 2). Bands 2, 3, and 4 have identical susceptibility to serine protease inhibitors (aprotinin, leupeptin, PCMPS, PMSF, TLCK, and trypsin inhibitor), indicating that they are all forms of the same serine protease. These bands were also observed in extracts of mycelia grown in glucose and are likely to be due to the activity of an intracellular serine protease (Figure 6A, lane I). Bands 5 and 6 also show susceptibility to serine protease inhibitors but their susceptibility profiles differ from each other as well as from bands 2, 3, and 4. The identity of band 5 is known. Gene disruption experiments have shown that it is the product of the prtA gene (vanKuyk et al. 2000). None of the bands were affected by the aspartic protease inhibitor pepstatin A.
Analysis of the spectrum of proteases in the filtered culture medium of the xprG2 revertants showed very high levels of the three intracellular protease bands (bands 2, 3, and 4 in Figure 6B, top). These results suggest that proteases that are normally found in the vacuole are secreted in the revertants. A band corresponding to PrtA was detected in the xprG2 sogC1 strain (xprG2 revertant 1) but not in the other three revertants. No bands with protease activity were visible in the xprG2 mutant. High levels of the activity corresponding to intracellular protease bands 2, 3, and 4 were also detected in mycelia (Figure 6B, bottom), indicating that abnormal secretion of this protease has not depleted the intracellular pool.
In eukaryotes, a series of membrane-enclosed compartments are involved in transport of proteins to the cell surface and, in the reverse direction, transport of cell surface proteins marked for degradation. In the late Golgi, newly synthesized proteins destined for the cell surface or secretion (e.g., extracellular proteases) are sorted from vacuole-bound proteins (e.g., intracellular proteases), which move into the endosomal system. Cell-surface proteins targeted by ubiquitination for degradation in the vacuole, also enter the endosomal system from whence they can be retrieved by retrograde transport to the Golgi. In addition to cell-surface proteins, other proteins that are known to be subject to retrograde transport from endosomes to the Golgi in S. cerevisiae include the acid hydrolase receptor Vps10p, the peptidases Kex2p and Ste13p, and the v-Snare Snc1p (reviewed in Bonifacino and Rojas 2006).
Two of the genes identified in this study are putative homologs of the S. cerevisiae vacuolar protein-sorting (VPS) genes, VPS5 and VPS17, which are involved in retrograde transport from endosomes to the Golgi (Horazdovsky et al. 1997). In S. cerevisaie, vps5 and vps17 mutants contain fragmented vacuoles and are defective in the sorting and processing of vacuolar hydrolases including carboxypeptidase Y, proteinase A, and proteinase B (Banta et al. 1988; Robinson et al. 1988). Enzymes that are destined for the vacuole are separated from secreted proteins by binding to a sorting receptor, Vps10p, in a late Golgi compartment (Vida et al. 1993; Marcusson et al. 1994). The vacuolar proteins are then delivered to a prevacuolar endosome. Once its cargo is delivered, Vps10p is returned by the retromer complex back to the Golgi where it is then reused (Seaman et al. 1998). The recycling of Vps10p is essential for the delivery of vacuolar enzymes and this recycling depends on Vps5p and Vps17p, which are components of the retromer complex (Horazdovsky et al. 1997). Vps26p, Vps29p, and Vps35p are also components of the retromer (Bonifacino and Rojas 2006). In vps5 and vps17 mutants, Vps10p is not present in the late Golgi where protein sorting normally occurs and, thus, carboxypeptidase Y is secreted rather than targeted to the vacuole.
The data presented here suggest the A. nidulans vps5 (sogA) and vps17 (sogB) homologs perform a similar role and that mutations in these genes result in the secretion of proteases that are normally intracellular. The sogC gene has not yet been isolated due to the complex nature of the sogC1 mutation. However, a high level of intracellular protease was also detected in the culture medium of the sogC1 mutant, which indicates that sogC is also likely to encode a Vps homolog. The region adjacent to the sogA gene contains a second sorting nexin (AN3584) and appears to contain a cluster of other genes involved in protein trafficking and degradation (supplemental Table S1).
Mammalian genomes encode two homologs of Vps5p (sorting nexins SNX1 and SNX2) but no Vps17p homolog (Kurten et al. 1996; Haft et al. 1998, 2000). The evidence suggests that mammalian sorting nexins, like Vps5p, are components of the mammalian retromer (Arighi et al. 2004; Seaman 2004). Sorting nexins are also involved in the degradation of the epidermal growth factor receptor, though the mechanism by which this occurs is not clear (Carlton et al. 2004; Gullapalli et al. 2004). SNX1 has recently been shown to be involved in retromer-independent transport of a G protein-coupled receptor to the lysosome for degradation (Gullapalli et al. 2006). Both SNX1 and SNX2 associate with a variety of cell surface receptors in vivo and in vitro (Haft et al. 1998; Heydorn et al. 2004). SogA and SogB may also be involved in the degradation of cell surface proteins in A. nidulans. Mutations in both these genes increase sensitivity to molybdate. One mechanism by which this might occur is through an increase in the uptake of molybdate through decreased degradation of a molybdate transporter in the mutants.
Northern blot analysis showed that transcript from the extracellular serine protease gene prtA was detected (albeit at low levels) in the xprG2 revertants. PrtA activity was also detected in one of the four revertants using protease zymograms. This finding suggests that extracellular protease gene expression is activated during carbon starvation by an xprG-independent mechanism in the revertants. The pathways involving creA, creB, and creC are unlikely to be responsible as mutations in these genes do not suppress the xprGΔ1 null mutation (Katz et al. 2008).
The properties of the extracellular and intracellular proteases that were detected in this study are similar to those reported by Cohen (1973b). Cohen identified, in the culture filtrate of A. nidulans, four proteases (α, β, γ, and ε) that were distinguished on the basis of electrophoretic mobility, inhibitor sensitivity, substrate preference, and pH optimum. Extracellular proteases γ and ε were susceptible to serine protease inhibitors and are similar to bands 5 and 6 in this study. Band 5 has been shown to be the prtA gene product (Katz et al. 1994; vanKuyk et al. 2000). Extracellular protease α was not inhibited by any of the protease inhibitors tested by Cohen (1973b) but the mobility of α was altered by EDTA (Cohen 1973b). The protease activity in band 1, which is sensitive to metal-chelating agents and thus is a metallo-protease, probably corresponds to protease α. The fact that Cohen (1973b) treated his samples with protease inhibitors prior to electrophoresis rather than afterwards as in this study and that protease inhibition by EDTA is reversible may account for the differences in the two studies.
Like Cohen (1973b), who found three forms of an intracellular serine protease (β), three bands (2, 3, and 4) of protease activity with identical inhibition profiles were found in mycelial extracts in this study. This protease activity was present at very high levels in the culture medium and mycelia of the xprG2 revertants. It is not known whether there are higher levels of the intracellular protease during carbon starvation in A. nidulans but in A. niger, a potential homolog of the gene encoding bands 2–4 (pepC) is expressed constitutively (Jarai et al. 1994). It remains to be determined whether the secretion of intracellular protease triggers increased production or whether the sog− mutations activate a signaling pathway (e.g., a stress response pathway) that leads to increased expression by some other mechanism.
We gratefully acknowledge Karen-Ann Gray for technical assistance, A. J. Clutterbuck for providing A. nidulans strains, and the Australian Research Council for supporting this project.
- Received August 18, 2008.
- Accepted February 1, 2009.
- Copyright © 2009 by the Genetics Society of America