Characterization of a vacuolar protease in Neurospora crassa and the use of gene RIPing to generate protease-deficient strains.

We have isolated a gene from Neurospora crassa that appears to encode a pepstatin-sensitive protease found both in membranes and in soluble contents of vacuoles. The gene contains two introns and encodes a 396-residue protein with a molecular mass of 42,900 Da. Because of the similarity of the protein to proteinase A in Saccharomyces cerevisiae the gene has been named pep-4. Strains with mutations in the pep-4 gene were generated in vivo by the gene RIPing procedure described by Selker and Garrett (Selker, E. U., and Garrett, P. W. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 6870-6874). The mutant strains were deficient in pepstatin-sensitive protease activity and did not appear to produce a major 42-kDa polypeptide in the vacuole. The mutant strains grew at the same rate as the wild type and had no other observable phenotype. When compared with inactivation of the PEP4 gene of S. cerevisiae, inactivation of the pep-4 gene in N. crassa produced a phenotype that was different in several ways. In N. crassa the mutant strains did not exhibit reduced sporulation or reduced viability after nitrogen starvation, and they had elevated levels of proteinase B and carboxypeptidase activities. The pep-4 gene appears to encode the N. crassa, homolog of proteinase A, but the maturation of vacuolar hydrolases appeared to be less dependent on this protease than has been observed in S. cerevisiae.

The lysosomal compartment of eucaryotic cells contains a large variety of hydrolytic enzymes. Most of these appear to be synthesized in the endoplasmic reticulum as inactive precursors, which are then modified in the Golgi and delivered to the lysosome. Activation occurs in the lysosome by the proteolytic removal of part of the polypeptide chain (1)(2)(3)(4)(5). In Saccharomyces cerevisiae proteinase A, the product of the PEP4 gene, has been shown to play a major role in the activation of vacuolar proteases. For example, cleavage by proteinase A activates proteinase B and carboxypeptidase Y. Proteinase B can, in turn, activate other hydrolases. Thus, proteinase A appears to be at the top of an activation cascade, and cells containing pep4 null mutations are deficient in at least three different proteases (6,7). Proteinase A itself is made as a prepropolypeptide. A 21-amino acid signal sequence is removed in the endoplasmic reticulum, and an additional 55 amino acids are cleaved autocatalytically in the vacuole to activate the enzyme.
We have been investigating the structure and function of vacuoles in Neurospora crassa (8). These organelles are very similar in size and composition to mammalian lysosomes. Because of the high density of vacuoles from N. crassa they can be isolated in good yield and purity. One of the most abundant proteins in the vacuole is a polypeptide of approximately 40 kDa (9). In our efforts to identify subunits of the vacuolar ATPase, we obtained evidence, described in this report, that the abundant 40-kDa protein is proteinase A. (We subsequently identified a 41-kDa subunit of the vacuolar ATPase that co-migrates in polyacrylamide gels with proteinase A (10).) The characterization of proteinase A in N. crassa is important for several reasons. First, to our knowledge, this is the only protein in the vacuolar lumen to be identified in any filamentous fungus. (A gene encoding a serine protease that probably resides in the vacuole has been isolated from Aspergillus niger (11). However, direct information on the location of the gene product has not been reported.) If genes for several vacuolar proteins can be isolated, then the mechanisms of protein targeting to the vacuole can be investigated. Second, investigation of the structure and maturation of proteinase A from N. crassa will permit a test of hypotheses developed in S. cerevisiae. Much of our knowledge of vacuolar biogenesis is based on data obtained with yeast. It is important to see the extent to which these observations can be applied to other organisms. It should be noted that although S. cerevisiae and N. crassa are both fungi, the evolutionary distance between the yeasts and the filamentous fungi is nearly as great as the distance between yeasts and animals (12). Third, the identification of the gene for proteinase A makes it possible to generate protease-deficient strains. As mentioned above some pep4 strains of S. cerevisiae are deficient in multiple proteases. They have normal growth rates and have become widely used laboratory strains. Because of the industrial importance of protein production in filamentous fungi, protease-deficient strains could be very useful.
Preparation and Sequencing of Vacuolar Proteins-Vacuoles were prepared as described (13,14). Vacuoles were lysed by suspension in 1 mM EGTA, adjusted to pH 7.5 with Tris, and centrifuged at 20,000 ϫ g for 5 min to remove high density contaminants, mostly cell wall fragments. The supernatant was centrifuged at 150,000 ϫ g for 10 min. The final supernatant was kept as the "vacuolar contents" fraction, while the pellet contained the vacuolar membranes. Membranes were resuspended in 1 mM EGTA, pH 7.5, at a concentration of approximately 3 mg protein/ml. For some experiments the membrane sector of the vacuolar ATPase was prepared as described (9). Both vacuolar membranes and contents were stored at Ϫ80°C. Vacuolar proteins were further purified by polyacrylamide gel electrophoresis. Tryptic fragments of the proteins were isolated and purified as described (15). To summarize briefly, individual bands were excised from polyacrylamide gels, electroeluted, precipitated with acetone, redissolved in 2 M urea, 0.1 M NH4CO3, pH 7.5, and digested with trypsin (0.1 g, 24 h, 37°C). The tryptic fragments were purified by high pressure liquid chromatography, using a C18 column (Vydac 218TP). Peptides were sequenced in our laboratory using a Porton 2090E Integrated Microsequencing System.
Isolation and Sequencing of the pep-4 Gene-The polymerase chain reaction was used to amplify a portion of the PEP4 gene (see Fig. 1). The 5Ј primer, containing an EcoRI restriction site, was GGAATTCAAG-GT(C/T)(G/A)T(C/T)CT(C/T/G)GA(C/T)AC(C/T/G/A)GG. The 3Ј primer, containing a HindIII restriction site, was the antisense of GG(T/C/ A)AT(C/T)CT(C/T/G)GG(C/T)CT(C/T/G)GG(C/T)TAAGCTTCC. Using N. crassa genomic DNA we amplified a 0.3-kb 1 fragment, which was then used to screen a cDNA library in the -ZAP vector. Two cDNA clones, 1.1 and 1.7 kb, were isolated. The ends of the shorter cDNA, all of the longer cDNA, and the genomic DNA described below, were sequenced (16), using the Cyclone kit (International Biotechnologies, New Haven, CT), to obtain a series of single stranded plasmids with nested deletions. The 3Ј ends of both cDNA clones were polyadenylated. The longer cDNA clone was used to screen a genomic library in the pSV50 vector (17). Two positive cosmid clones were isolated, 6-5-A and 31-5-E. By using the 1.7-kb cDNA to probe blots of digested cosmid DNA, the pep-4 gene was found to lie within two adjacent PstI fragments. Each of these fragments was subcloned into the Bluescript SK ϩ vector (Stratagene, La Jolla, CA). The smaller fragment (0.6 kb) was completely sequenced and was found to contain the 5Ј end of the gene. To identify the rest of the pep-4 gene, half of the larger (3.8-kb) PstI fragment was sequenced.
Inactivation of the pep-4 Gene-A 1.8-kb fragment of the pep-4 gene was isolated by digesting the Bluescript plasmid containing the 3.8-kb genomic fragment with HindIII. (One HindIII site was in the multiple cloning site of the vector, and the other was at the 3Ј end of the gene.) The 1.8-kb HindIII fragment was inserted into the pCSN44 vector (18), yielding pCSN-pep4. pCSN44 contains a gene that confers resistance to hygromycin B. Using a method previously described (19) we transformed N. crassa strain LV10 (his-3, spe-1, aga; Fungal Genetics Stock Center 2760 -3A, kindly provided by R. Davis and J. Pitkin, University of California, Irvine) with pCSN-pep4 and selected hygromycin-resistant colonies. Genomic DNA was prepared from 20 of these transformed strains, using the procedure of Oakley et al. (20). After digestion with EcoRV, the DNA was analyzed by the Southern blotting procedure, using the 1.8-kb HindIII fragment as probe. Three of the transformed strains, designated P3, P6, and P15, appeared to contain a single copy of the pCSN-pep4 plasmid integrated into the genome.
Following the procedure of Selker et al. (21,22) (described more fully in "Results") we generated repeat-induced point mutations in the pep-4 gene. The P3, P6, and P15 strains were mated to wild type strain 74A. After approximately 3 weeks, spores were isolated from each of these matings. As shown below, the pep-4 locus was found to be located within 5 map units of the aga locus, which encodes the enzyme arginase (23). The P3, P6, and P15 strains were aga Ϫ ; therefore, inactivated pep-4 genes were predicted to cosegregate with the aga Ϫ gene. Six aga Ϫ strains, identified as described (23), were isolated from each of the three matings. DNA was isolated, digested with NlaIII, and analyzed by the method of Southern using 32 P-labeled pep-4 cDNA as probe. One of the 18 isolates, designated P6 -45 appeared to have mutations in the pep-4 gene. In this isolate the restriction fragments seen in the wild type control were missing, and a larger, novel fragment appeared. The P6 -45 strain did not contain DNA from the pCSN-pep4, indicating that the integrated plasmid was not closely linked to pep-4 and did not cosegregate with pep-4 in this particular isolate.
To determine if the pep-4 gene was indeed mutated, the P6 -45 strain was mated with wild type 74a. Twelve progeny were analyzed for proteinase A activity (see below). Six had essentially the same activity as the wild-type control, and six had 10 -20% of the activity of wild type.
To obtain additional mutated pep-4 alleles, an additional 15 aga Ϫ progeny were isolated from the original P6 ϫ 74a cross. Two of these with low proteinase A activity, designated pep-4 -47 and pep-4 -49 were used for all experiments described in this report. To analyze the mutations in these strains, genomic DNA was prepared and the pep-4 gene was amplified by the polymerase chain reaction. The positions and sequence of the primers are shown in Fig. 1. The amplified DNA was separated on an agarose gel, purified by the GeneClean procedure (Bio 101, La Jolla CA), and subcloned into the pT7Blue vector (Novagen, Madison, WI). Single-stranded DNA was prepared, essentially by the procedure recommended by the manufacturer. The ends of the mutated pep-4 genes were sequenced, further primers were made, and the process was repeated until the entire amplified region had been sequenced.
Protease Assays-N. crassa cells were grown 3-5 days at 25°C, in 20-ml stationary liquid cultures of Vogel's minimal medium as described (24). Preliminary experiments showed that protease activity was essentially the same after 3, 4, or 5 days of growth. Cells were harvested by filtration through Whatman No. 1 filter paper, rinsed in 0.1 M Tris, pH 7.5, suspended in approximately 10 ml of 0.1 M Tris, pH 7.5, and disrupted in a BeadBeater homogenizer (Biospec Products, Bartlesville, OK) for 1 min in an ice-jacketed chamber. The homogenate was transferred to a 12-ml tube and centrifuged for 10 min at 3000 rpm in a Sorvall SS34 rotor to pellet heavy debris and intact cells. The supernatant was transferred to 1.5-ml microfuge tubes and centrifuged for 20 min at 18,000 ϫ g. The final supernatant was removed with a pipette and placed on ice. For most protease assays the cell extract was used within 1 h of preparation, but the extract also retained significant activity if frozen in liquid nitrogen and stored at Ϫ80°C. Proteinase A and proteinase B were assayed as described (2), except that the hydrolysis of the peptide N-succinyl-L-arginyl-L-prolyl-L-phenylalanyl-L-histidyl-L-leucyl-L-leucyl-L-valyl-L-tyrosine 7-amido-4-methylcoumarin was carried out at 37°C. Protein concentration of the cell extracts was determined by the method of Lowry et al. (25).

RESULTS
Isolation of the pep-4 Gene-N. crassa vacuoles contain an abundant polypeptide of approximately 40 kDa, some of which copurifies with the vacuolar ATPase. Using polyacrylamide gel electrophoresis, we attempted to purify the 41-kDa subunit of the membrane sector of the vacuolar ATPase (9). A tryptic fragment of the protein was sequenced (VVLDTGSSNL) and compared with proteins in the GenBank TM data base. The sequence was nearly identical to residues 106 -115 (VILDTGSSNL) of proteinase A in S. cerevisiae. A second tryptic fragment, prepared from the prominent 40-kDa polypeptide in the soluble contents of vacuoles, had the sequence GIL-GLGYDTV, similar to residues 196 -205 (GILGLGYDTI) of yeast proteinase A. With this information the polymerase chain reaction was used to amplify a fragment of N. crassa genomic DNA. This fragment was, in turn, used to isolate two cDNA clones and genomic clones, as described under "Experimental Procedures." The sequence of the genomic DNA is shown in Fig. 1. By comparing cDNA and genomic DNA we identified two short introns of 63 and 71 base pairs. As is typical for many N. crassa genes, the introns occur early in the protein coding region (26). Also typical was the observation that the cDNAs were polyadenylated at different sites, in this case separated by 67 base pairs. The chromosomal position of the gene was precisely determined by the fortuitous observation that it was present in two cosmid clones that had been previously mapped. Based on the data of Maclung et al. (27) the gene lies on linkage group VII in the middle of a 25-kb region that separates the frq and for genes.
Analysis of the open reading frame showed that the gene encoded a protein with 396 amino acids, 42,900 daltons. When aligned with proteinase A from S. cerevisiae (Fig. 2) 57% of the residues were identical. Proteinase A is an aspartyl protease that has two highly conserved active sites (1, 6, 7). These regions, boxed in Fig. 2, are essentially identical in S. cerevisiae and N. crassa. Also identical are the two putative glycosylation sites and two suggested sites of disulfide bond formation (Cys 116 -Cys 121 and Cys 318 -Cys 351 ). These data strongly suggested that the gene we had cloned encoded the N. crassa homolog of proteinase A; thus, it was named pep-4.
In S. cerevisiae proteinase A is synthesized as a large inactive precursor. The N terminus of the protein is cleaved, first in the endoplasmic reticulum and second in the vacuole to generate the active, mature form of the protein. One of the tryptic peptides that we sequenced corresponded to residues 71-90 in the N. crassa protein (underlined in Fig. 2). This peptide was particularly interesting because it was not preceded in the sequence by an arginine or lysine residue as would be predicted for trypsin cleavage. Furthermore, the beginning of this peptide corresponds precisely to the beginning of the mature form of proteinase A from S. cerevisiae. The N. crassa peptide may have been derived from the N terminus of the mature protein.
Comparison of "presequences" in proteinase A is discussed further below.
Inactivation of the pep-4 Gene-To inactivate the pep-4 gene we took advantage of the RIPing phenomenon (21,22). As described under "Experimental Procedures," we introduced an extra copy of the pep-4 gene into wild type cells by transformation. The transformed cells were then put through a sexual cycle. In N. crassa, haploid cells of different mating type fuse and undergo several rounds of nuclear division before nuclear fusion occurs. In these premeiotic cells, nuclei with two copies of a given gene are subject to a poorly understood process whereby duplicated DNA becomes mutated. GC base pairs are converted to AT base pairs in both the introduced and the endogenous copy of the gene. 5-50% of nuclei with duplicated DNA are affected, and the number of point mutations varies from 1 to more than 50 per kb of DNA (22).
The transformed strains contained an extra copy of a 1.8-kb PstI/HindIII fragment of genomic DNA, which included the entire protein coding region of the pep-4 gene. Haploid progeny of a cross between transformant and wild type were examined for changes in NlaIII restriction sites within the pep-4 gene or for loss of protease activity in cell extracts (see "Experimental Procedures"). Two of the progeny, designated pep-4 -47 and pep-4 -49 were selected for further analysis. Using the polymerase chain reaction, genomic DNA corresponding to the entire protein coding region of the pep-4 gene was amplified. The 5Ј primer was 30 base pairs upstream of the PstI site that marked the end of the duplicated region, thus ensuring that the amplified fragment corresponded to the endogenous copy of the pep-4 gene. In addition, the introduced, extra copy of pep-4 had apparently integrated at a site not closely linked to the pep-4 locus, because the introduced and endogenous pep-4 genes segregated independently (data not shown). The amplified DNA was sequenced as described under "Experimental Procedures" and found to be mutated at many sites. In strain pep-4 -47, 99 nucleotide changes were found, scattered throughout the length of the duplicated region. In strain pep-4 -49, 22 point mutations were detected, all but two in the 5Ј half of the duplicated region. The changes were typical of those observed previously in RIPed genes, GC to AT transitions with cytosine followed by adenine being the most common target.
The effect of these mutations on the protein sequence is shown in Fig. 3. In strain pep-4 -49 all changes were missense mutations. Fourteen nucleotide changes were silent, while the rest were predicted to change seven amino acids, all clustered in the region of residues 145-202. In the more heavily mutated strain, pep-4 -47, a stop codon was introduced after amino acid 31. Furthermore, the highly conserved G in the 5Ј splice site of the first intron was mutated.
Analysis of Protease Activity in Strains with Mutations in the pep-4 Gene-We used whole-cell extracts to assay the hydrolysis of acid-denatured hemoglobin (see "Experimental Procedures"). In S. cerevisiae proteinase A has been reported to be the only protease to catalyze the reaction at acid pH (2). We also tested the effect of pepstatin, a specific inhibitor of aspartyl proteases. As shown in Fig. 4 extract from the wild type cells had significant levels of protease activity (11 g of tyrosine/ min/mg), which was inhibited 85% by pepstatin. The optimum pH for this activity was 3.2, the same as reported for S. cerevisiae (30). Protease activity in the mutant strains, pep-4 -47 and pep-4 -49, was approximately 15% the level observed in the wild-type and was not significantly inhibited by pepstatin. Our interpretation of these data was that the mutant strains had no measurable proteinase A-like activity and that the pepstatinresistant activity came from other enzymes.
Proteinase A has also been reported to hydrolyze the octapeptide N-succinyl-L-arginyl-L-prolyl-L-phenylalanyl-L-histidyl-L-leucyl-L-leucyl-L-valyl-L-tyrosine 7-amido-4-methylcoumarin (2). As shown in Fig. 5, extract from wild type N. crassa hydrolyzed the peptide at significant rates, and the hydrolysis was inhibited 63% by pepstatin. The pep-4 mutant strain had only 36% of the activity of the wild type, and this activity was not significantly inhibited by pepstatin. Thus, the mutant strain appeared to be deficient in all of the proteinase A-like activity as assayed by hydrolysis of this peptide.
We also measured the hydrolysis of hide powder azure, diagnostic of proteinase B activity, and of N-benzoyl-L-tyrosine p-nitroanilide, diagnostic of carboxypeptidase activity (2), as shown in Table I. The mutant strains had significant levels of both types of activity. These experiments were repeated several times; the mutant strains consistently yielded 1.5-2.0-fold higher activity than the wild type.
Analysis of the pep-4 Protein-Vacuoles were prepared from the two mutant strains and from the wild type. The vacuoles were lysed by suspension in 1 mM EGTA. Membranes and soluble contents were separated by centrifugation and analyzed by gel electrophoresis. As shown in Fig. 6 an abundant polypeptide of 40 kDa, presumably proteinase A, was observed in the wild type but not in either of the mutant strains. The pattern of polypeptides was clearly different for the membranes versus the soluble contents, however proteinase A was a major component of both.
The proteins from vacuolar membranes and contents were transferred to a nylon membrane and incubated with a polyclonal antibody, raised against the proteinase A from S. cerevi- siae (6). A band of 40 kDa was bound by the antibody in protein from the wild type strain, but not in the pep-4 mutant strains (data not shown). These results strongly suggested that the 40-kDa polypeptide was proteinase A.
Viability and Fertility of pep-4 Mutant Strains-The growth rate of the pep-4 mutant strains 47 and 49 was not measurably different from the growth rate of the wild type either on solid minimal medium or in vigorously aerated liquid minimal medium (data not shown). Also, in contrast to pep4 mutant stains of S. cerevisiae (28), mutations in the N. crassa pep-4 gene did not significantly affect fertility or viability after nitrogen starvation. A cross between pep-4 -47 and pep-4 -49 produced abundant viable spores. To test the effect of the pep-4 mutation on the survival of cells after nitrogen starvation, 3-day-old cultures of both pep-4 mutant strains and the wild-type strain were transferred to nitrogen-free Vogel's medium. The cultures were incubated at 25°C. After 2, 3, 4, or 5 days the cultures were rinsed and resuspended in normal Vogel's medium containing ammonium nitrate to test for the resumption of growth. All cultures grew well, and mutants were indistinguishable from the wild type. DISCUSSION We have isolated a gene that appears to encode the N. crassa homolog of proteinase A. The high degree of sequence conservation at the putative active site and sensitivity of the pep-4 gene product to inhibition by pepstatin argues strongly that the enzyme is an aspartyl protease. The pep-4 gene product shares approximately 30% sequence identity with mammalian cathepsins, pepsinogens, and renins but is 57% identical to proteinase A from S. cerevisiae. Antibody raised to the yeast proteinase A cross-reacts with the N. crassa polypeptide, the protease was located in the vacuole, and the specificity for substrates, measured in vitro, was the same as reported for yeast proteinase A.
Proteinase A in S. cerevisiae is made as a prepropeptide that undergoes maturation by cleavages in the endoplasmic reticulum and in the vacuole (3, 6, 7). We do not have direct evidence for processing of the N. crassa protein, but one of the tryptic peptides we sequenced was not preceded by the expected lysine or arginine residue. It is likely that this peptide was derived from the N terminus of the mature protein, indicating that the N. crassa and S. cerevisiae precursor proteins are cleaved at the same position. The 20 amino acids preceding the cleavage site are completely dissimilar, but the amino acids at the N terminus of the mature protein are very similar. A recent analysis of the propeptide of proteinase A from S. cerevisiae identified 39 functional mutant forms containing random sequence in the region preceding the cleavage site (29). A single residue, lysine at position 53, was identified as important for processing. This lysine is present in the N. crassa protease, part of a sequence of six amino acids (HLGQKY) identical in both yeast and N. crassa propeptides. Conceivably, the region containing these six amino acids could interact with conserved regions in the mature protein to determine the site of cleavage.
Our data indicated that in N. crassa proteinase A is one of the most abundant proteins in the vacuole. In S. cerevisiae the purified enzyme behaves like a typical soluble protein (30), and in N. crassa most of the enzyme was found in the soluble contents. However, it was surprising to observe that proteinase A was also the most prominent polypeptide in vacuolar membranes. Because of our interest in other proteins in the vacuolar membrane, particularly the proton-pumping ATPase, we wished to generate a strain of N. crassa that lacked proteinase A. We used the RIPing phenomenon, which appears to be unique to N. crassa (22) to inactivate the pep-4 gene. Seven amino acids were changed in one mutant strain, and a stop codon was introduced early in the protein coding region of a second mutant strain. Cell extracts from the pep-4 mutant strains lacked essentially all proteinase-A activity as assessed by hydrolysis of two different substrates. Neither mutant strain had detectable proteinase A polypeptide in the vacuole,   as assayed by gel electrophoresis or by the Western blotting procedure. In the strain with only missense mutations (pep-4 -49) the polypeptide must be quickly degraded or not targeted to the vacuole. Inactivation of the pep-4 gene had no measurable effect on growth rate or viability of the strains. pep-4 mutants of N. crassa differ from pep4 mutants of S. cerevisiae in two significant ways. First, in S. cerevisiae a/␣ diploids that are homozygous for pep4 are sporulation-deficient (28). In N. crassa, crosses between two pep-4 mutant strains produced a normal yield of viable spores. Secondly, pep4 mutants of S. cerevisiae are deficient in several vacuolar proteases, including protease B and carboxypeptidase Y (6,7). In N. crassa these proteases are not deficient in pep-4 mutants. In fact the activities measured were higher than in the wild type strain, as if the loss of proteinase A was compensated by higher levels of other proteases. The multiple protease deficiency in yeast pep4 strains has been explained by the hypothesis that proteinase A is needed for the proteolytic maturation of several other proteases. The requirement does not appear to be absolute, however, because there may be two or more enzymes that can cleave and activate precursor proteins for vacuolar hydrolases. Woolford et al. (31), in a recent study of proteinase A mutants, concluded that "a great deal of redundancy has been built into the system." Our analysis of pep-4 mutants in N. crassa supports this conclusion.
Although the pep-4 mutant strains we have generated retain significant protease activity, they may still be useful for the isolation of proteolytically sensitive proteins. Proteinase A appears to be very abundant in the wild type strains, and a significant amount is tightly associated with the vacuolar membrane. In studies of the vacuolar ATPase we have found that some of the subunits (e.g. the 57-and 98-kDa polypeptides) are particularly sensitive to proteolysis (9). Use of the pep-4 strain for purification of the ATPase and other proteins of the vacuolar membrane may provide a greater yield of undamaged polypeptides.