Identification of proteases with shared functions to the proprotein processing protease Krp1 in the fission yeast Schizosaccharomyces pombe

Authors


Abstract

Many secretory proteins are synthesized as inactive proproteins that undergo proteolytic activation as they travel through the eukaryotic secretory pathway. The best characterized family of processing enzymes are the prohormone convertases or kexins, and these are responsible for the processing of a wide variety of prohormones and other precursors. Recent work has identified other proteases that appear to be involved in proprotein processing, but characterization of these enzymes is at an early stage. Krp1 is the only kexin identified in the fission yeast Schizosaccharomyces pombe, in which it is essential for cell viability. We have used a genetic screen to identify four proteases with specificities that overlap Krp1. Two are serine proteases, one is a zinc metalloprotease (glycoprotease) and one is an aspartyl protease that belongs to the recently described yapsin family of processing enzymes. All four proteases support the growth of a yeast strain lacking Krp1, and each is able to process the P-factor precursor, the only substrate currently known to be processed by Krp1.

Introduction

The Krp1 protease from the fission yeast Schizosaccharomyces pombe is a Ca2+-dependent type I serine endopeptidase that cleaves secretory proproteins on the carboxyl side of Lys–Arg and Arg–Arg motifs. It is a member of the prohormone convertase family of eukaryotic enzymes (for reviews, see Steiner et al., 1992; Zhou et al., 1999), although family members are often called kexins, because the KEX2 protease from the budding yeast Saccharomyces cerevisiae was the first to be characterized. Kexins have been implicated in processing a variety of proproteins as they are transported through the secretory pathway. Krp1 processes the P-factor pheromone precursor, but there must be other substrates, as Krp1 activity is essential for cell viability (Davey et al., 1994). One approach to identifying these substrates might be to isolate multicopy suppressors that allow an S. pombe strain containing a temperature-sensitive Krp1 to grow at the non-permissive temperature. We have previously generated a krp1ts allele that allows growth at 23°C but not at 37°C (Ladds et al., 2000). The enzyme retains some activity at 37°C, as it allows growth when overexpressed, and we would therefore expect overexpressing an essential substrate to allow the cells to overcome the growth defect associated with the reduced activity. This approach would also be expected to identify proteases with activities that can complement Krp1, and here we report the isolation and characterization of four such enzymes.

Results and discussion

A single point mutation generates a temperature-sensitive Krp1

Loss of Krp1 activity is lethal (Davey et al., 1994), and we therefore prepared strains containing a temperature-sensitive version of the protein (Ladds et al., 2000). Briefly, the wild-type krp1 was replaced by a randomly mutated version of the gene generated by performing the polymerase chain reaction (PCR) under conditions of low stringency and then selecting temperature-sensitive transformants that could be rescued by a plasmid-borne krp1. One such isolate contained a single T-to-C mutation at nucleotide 1274 (relative to the ATG of the krp1 gene) that changed the phenylalanine at residue 425 to serine. Recreating this single change in JY383 (krp1+) generated the temperature-sensitive strain JY965 (krp1ts) (Ladds et al., 2000). A three-dimensional structure is not available for any kexin, but predictive methods and computer-assisted modelling place this residue at the junction between the catalytic domain (residues 103–422) and the P domain (residues 426–584). The two domains are thought to fold independently but are then believed to interact through a set of hydrophobic contacts that allow the P domain to stabilize and regulate the catalytic domain (Zhou et al., 1998). We presume that the F425S mutation affects the interaction between these domains and reduces the activity of the enzyme.

There were no obvious differences between the krp1ts strain (JY965) and its isogenic krp1+ parent (JY383) at 23°C (Fig. 1). Both had a doubling time of about 4.3 h in minimal medium and the same barrel-like morphology expected for wild-type S. pombe. Shifting krp1+ cells to 37°C induced a lag in cell division, but the cells soon resumed exponential growth with a doubling time of 2.2 h. Cell shape was unaffected by the shift to 37°C, and cells remained viable throughout the experiment. In contrast, the krp1ts strain stopped dividing soon after the shift to 37°C. Cells underwent one round of division, but there was no further increase in cell number after 6 h. Viability reduced soon after the transfer to 37°C, and no viable cells could be recovered 28 h after the shift.

Figure 1.

Growth, viability and volume of the krp1+ (JY383) and krp1ts (JY965) strains were determined in minimal medium at 23°C and 37°C. Cell concentrations and median cell volumes were determined using a Coulter Channelyser, and viability was assessed by colony formation at 23°C. There was no difference between the strains at 23°C, and only the growth of the krp1ts strain (x) is shown for clarity. Viability of both strains remained high at 23°C and is not shown. The shift from 23°C to 37°C was achieved within 10 min using a water bath. Growth of both the krp1+ (○) and the krp1ts (●) strains at 37°C is indicated, but the krp1+ strain remained viable throughout the course of the experiment, and only the viability of the krp1ts strain (▪) is shown.

Mutant phenotypes often provide clues to the process affected by a mutation, and we have started to investigate the defects in the krp1ts strain. Viability was lost soon after the shift to the non-permissive temperature, and the cells developed aberrant morphologies. Some cells adopted a lemon-like morphology, whereas others exhibited blebbing. Staining with Calcofluor white further highlighted defects in cell wall biosynthesis (Fig. 2). At 23°C, Calcofluor stained the septum in the middle of dividing cells, but the stained material was mislocalized at 37°C and generally found at one end of the non-viable cells. Investigations into the nature and construction of the cell wall in S. pombe are at a relatively early stage (for a review, see Ishiguro, 1998), and it is not yet possible to reconcile the mutant phenotype with any specific part of the process. One possibility is that Krp1 plays a role in activating or targeting the enzymes involved in cell wall construction, but no suitable substrates have yet been identified. It is perhaps significant that cell wall defects also accompany loss of kexin activity in S. cerevisiae and Yarrowia lipolytica. In neither case is the kexin essential for cell growth, but deletion of KEX2 in S. cerevisiae (Komano and Fuller, 1995) or of XPR6 in Y. lipolytica (Enderlin and Ogrydziak, 1994) leads to morphology changes similar to those seen with the krp1ts in S. pombe. Cell wall defects are also seen after inactivation of one of the kexins in the nematode Caenorhabditis elegans (Thacker et al., 1995). It is possible that kexins play a general role in regulating this important activity.

Figure 2.

Morphology of the krp1ts strain (JY965) during exponential growth in minimal medium at 23°C (top) and after 28 h at 37°C (bottom) when few of the cells were viable (see Fig. 1). Unfixed cells were stained with Calcofluor white, embedded in agarose and the fluorescent and interference-contrast images superimposed. Bar = 10 µm.

Multicopy suppression of krp1ts

The role of Krp1 as a proprotein processing enzyme along with the defects observed in the krp1ts strain at the non-permissive temperature suggested that normal yeast growth requires the function of processed forms of one or more unidentified Krp1 substrates. Such substrates might be identified as multicopy suppressors of the krp1ts defect. Although the krp1ts strain is unable to grow at 37°C, the Krp1ts protein is not completely inactive at this temperature, as overexpression of krp1ts allows growth under non-permissive conditions (see Fig. 3). It was therefore expected that overexpressing an essential substrate would also allow the cells to overcome the defect associated with the reduced processing activity. Such an approach was also expected to identify proteases with functions similar to those of Krp1.

Figure 3.

To demonstrate complementation of Krp1, the krp1ts strain JY965 was transformed with various pREP constructs and cultured in minimal medium at 23°C. Approximately 5000 transformants were plated onto medium lacking thiamine (nmt1 promoter is induced) or containing 5 µM thiamine (nmt1 promoter is repressed), and colony formation was recorded after 3 days at 37°C. Expression from the nmt1 promoter is not fully repressed by thiamine. Representative examples are shown for isolates from groups A (can complement growth under both conditions) and B (can complement growth only when the nmt1 promoter is induced). Proteins are from either S. pombe (Sp) or S. cerevisiae (Sc).

Minimal medium lacking leucine was used to select colonies after transformation of JY965 (krp1ts) with a S. pombe cDNA library in the regulated pREP vector. The pREP expression vector allows the expression of particular genes to be under the control of the thiamine-repressible nmt1 promoter (Maundrell, 1993). Transformants were plated onto medium lacking thiamine and incubated at 23°C for 24 h (to induce expression from the nmt1 promoter) before being transferred to 37°C to identify constructs able to rescue the krp1ts defect. Plasmids rescued from these transformants were reintroduced into a fresh krp1ts tester strain and plated at 37°C in the presence or absence of thiamine (Fig. 3). Expression from the nmt1 promoter is not fully repressed by thiamine, and there was sufficient expression from an nmt1–krp1 construct to overcome the krp1ts defect. In contrast, krp1ts could only support growth when the promoter was fully induced in the absence of thiamine. A few of the plasmids isolated in the screen could support growth in the presence of thiamine (group A in Fig. 3), but most were able to complement krp1ts only when the nmt1 promoter was fully induced (group B).

The rescued plasmids were grouped by restriction endonuclease analysis, and partial DNA sequence data were obtained for representative constructs. Most had sequences that were available as part of the S. pombe genome sequencing project, and this facilitated identification of their proposed products. Five of the constructs had the potential to encode proteins with homology to proteases, and these form the basis of this report. The analysis of other constructs will be presented elsewhere (G. Ladds and J. Davey, in preparation).

Serine proteases

We identified three serine proteases: Krp1 (accession number SPAC22E12.09c), Isp6 (SPAC4A8.04) and an enzyme we propose calling Psp3 (the third S.pombeserine protease) (SPAC1006.01). These are related to each other and to serine proteases from S. cerevisiae (Table 1, Fig. 4). Krp1 and KEX2 form a subdivision of this group and belong to the kexin family of proprotein convertases. Given the similarity between the two kexins, it was not surprising that KEX2 supported the growth of the krp1ts strain at 37°C (group A in Fig. 3).

Table 1. Pairwise comparison of the S. pombe serine proteases with each other and with related enzymes from S. cerevisiae.
 Isp6Psp3KEX2PRB1YSP3YCR045C
  1. The full-length sequences were compared using the GCG program bestfit and are shown as percentage identity and percentage similarity.

Krp122 (49)23 (46)40 (57)18 (42)20 (45)25 (48)
Isp6 43 (64)22 (45)47 (65)42 (64)33 (55)
Psp3  22 (47)45 (64)43 (63)35 (57)
Figure 4.

Alignment of serine proteases. Sequences were aligned using the pileup programme (GCG) with a GapWeight of 3.0 and a GapLengthWeight of 0.1. Gaps introduced to maximize the alignment are indicated by –. Residues identical in all the enzymes are indicated by black shading. The alignment also attempts to highlight the differences between the kexins and the other enzymes, and grey shading indicates residues that are either found in both kexins or in at least four of the other five enzymes. The catalytic residues (D H N S) are shown by letters above the sequence (these are at positions D162, H200, N300 and S371 for Krp1). The residues on either side of the predicted presequence cleavage sites are underlined (SC22 in Krp1); these sites have been confirmed for Krp1 (Powner and Davey, 1998) and KEX2 (Wilcox and Fuller, 1991). The dibasic motif used by the kexins to remove the prosequence from the catalytic fragment (KR102 in Krp1) is indicated by a double underline, and the internal dibasic motif that is subsequently cleaved to inactivate the inhibitory prosequence (KR82 in Krp1) is indicated by a dashed underline. The hydrophobic transmembrane regions near the C-terminus of the kexins are indicated by a wavy underline (F672 to I691 in Krp1). Arrowheads below the sequence indicate the mature region of PRB1 (from E281 to a region just upstream of N594) (Van Den Hazel et al., 1996).

The maturation and activity of Krp1 has been well documented (Davey et al., 1994; Powner and Davey, 1998; Ladds et al., 2000). Briefly, Krp1 is synthesized with an N-terminal presequence, a prosequence, a catalytic domain, a P domain and a C-terminal transmembrane domain. The presequence directs the precursor to the secretory pathway and is removed (by cleavage between Ser-21 and Cys-22) during segregation into the endoplasmic reticulum (ER). The prosequence is presumed to play a role in the correct folding of the catalytic domain and is removed in the ER by intramolecular cleavage after the dibasic motif Lys–Arg-102. The cleaved prosequence is not released from the enzyme but remains non-covalently associated with the catalytic domain and acts as an autoinhibitor so that the enzyme is able to become fully active only when the prosequence is subsequently cleaved at an internal dibasic motif (Lys–Arg-82).

Isp6 (induced during sporogenesis in S.pombe) was identified by a screen for cDNAs preferentially expressed during sexual differentiation (Sato et al., 1994). It is not essential for cell viability but is required for conjugation and was proposed to be a protease involved in protein reconstruction during the transition from vegetative growth to sexual differentiation (Sato et al., 1994). Sequence analysis suggests that Isp6 contains an N-terminal presequence to direct the protein into the ER, where cleavage between Ala-24 and Ala-25 is likely to generate the mature protein. The predicted protein contains no other obvious stretch of hydrophobic residues and is likely to be transported through the secretory pathway to either the vacuole or the extracellular medium.

Psp3 is very similar to Isp6 (43% identical, 64% similar). It has an N-terminal presequence, with cleavage predicted to be between Ala-20 and Phe-21, but there is no other stretch of hydrophobic residues, and it is likely to be a soluble protein at some location along the secretory pathway. A short fragment of this construct was identified during a study to sequence non-overlapping cDNA clones from S. pombe (Yoshioka et al., 1997).

Isp6 and Psp3 do not appear to be kexins and are more related to a group of serine proteases that include PRB1, YSP3 and YCR045C from S. cerevisiae. Their similarities extend beyond the catalytic regions, and the S. cerevisiae proteins could provide insights into the maturation of the S. pombe enzymes (for review, see Van Den Hazel et al., 1996). PRB1 is a vacuolar protein with broad substrate specificity and is made as a preproprotein. An N-terminal presequence targets the polypeptide to the ER and is removed by cleavage between Ala-19 and Leu-20. Autocatalytic cleavage between residues Thr-280 and Glu-281 releases a prosequence that remains non-covalently associated with the catalytic fragment and acts as an autoinhibitor so that the enzyme is unable to become active until the prosequence is degraded in the vacuole. Two-step processing of the C-terminus by PEP4 and PRB1 in the vacuole removes about 55 residues to generate a mature enzyme in which the C-terminus is just N-terminal of Asn-594 in the original preproprotein. Much less is known about YSP3, and no cellular function has yet been assigned to this enzyme. It was identified during the genome sequencing project (Sterky et al., 1996) and found to undergo rapid but transient induction upon transfer to sporulation medium (Chu et al., 1998). Sequence analysis of YSP3 predicts a shorter N-terminal prosequence (cleavage predicted to be near Gln-173), and it does not seem to require processing at the C-terminus, as this region in the initial translation product is very similar to the mature PRB1 protein. The two S. pombe enzymes are very similar to YSP3; both have short N-terminal prosequences and do not appear to require processing at the C-terminus.

Aspartyl protease

Our screen also identified an aspartyl protease that could overcome the loss of Krp1. This was previously identified during the S. pombe genome sequencing project (accession number SPCC1795.09). Based on sequence comparisons (Table 2, Fig. 5), we propose calling this Yps1 as the first member of the yapsin family of aspartyl proteases to be described in S. pombe.

Table 2. Pairwise comparison of the S. pombe aspartyl protease Yps1 with yapsins from S. cerevisiae.
 YPS1YPS2YPS3YPS6YPS7
  1. The full-length sequences were compared using the GCG program bestfit and are shown as percentage identity and percentage similarity.

Yps127 (49)27 (46)25 (44)25 (46)23 (45)
Figure 5.

Alignment of yapsins (YPS6 and YPS7 have not formally been shown to be yapsins). Sequences were aligned using the pileup programme (GCG) with a GapWeight of 3.0 and a GapLengthWeight of 0.1. Gaps introduced to maximize the alignment are indicated by –. Black shading indicates residues identical in all the enzymes, and grey shading indicates residues identical in at least four enzymes. The two catalytic aspartate residues are shown by letters above the sequence (these are at positions D85 and D301 for Yps1). The residues on either side of the predicted presequence cleavage sites are underlined (SL18 in Yps1), and basic motifs proposed as cleavage sites for the removal of an N-terminal prosequence are indicated by a double underline (KR45 in Yps1); cleavage at these sites has been confirmed for YPS1 (Cawley et al., 1998) and YPS3 (Olsen et al., 1999). The addition of a glycophosphatidylinositol (GPI) anchor at the C-terminus of the yapsins involves a hydrophobic region (indicated by a wavy underline; A505 to I521 in Yps1) and an attachment residue or ω site (indicated by a double underline; N495 in Yps1) (Nuoffer et al., 1993; Udenfriend and Kodukula, 1995). YPS1, YPS2, YPS3 and YPS6 also contain a dibasic motif (KR547 in YPS1) that is likely to prevent incorporation of the GPI-anchored protein into the cell wall (Hamada et al., 1998; 1999).

Unlike most aspartyl proteases, which cleave at hydrophobic residues, yapsins have a common specificity for basic residues (Egel-Mitani et al., 1990; Komano and Fuller, 1995) and appear to constitute a new family of proprotein processing enzymes. They have been described in mammals (Azaryan et al., 1995) and fish (Mackin et al., 1991), but are best characterized in S. cerevisiae. YPS1 (previously called YAP3) and YPS2 (previously called MKC7) were isolated as multicopy suppressors of KEX2 mutations in S. cerevisiae (Egel-Mitani et al., 1990; Komano and Fuller, 1995), whereas YPS3, YPS6 and YPS7 were identified by homology-based screens of the S. cerevisiae genome sequence (Olsen et al., 1999). A pseudogene encoding 165 amino acids was designated YPS5 and, confusingly, both YPS3 and YPS7 have been referred to as YPS4. Enzymatic analyses of YPS1, YPS2 and YPS3 reveal a common specificity for paired or single basic residue cleavage sites (Egel-Mitani et al., 1990; Bourbonnais et al., 1994; Komano and Fuller, 1995; Ledgerwood et al., 1996; Komano et al., 1999), and the exclusion of basic residues from the P1′ position directs the enzymes to cleave C-terminal to runs of basic residues (Komano et al., 1999). This is particularly relevant to their ability to cleave correctly after dibasic motifs that are normally processed by KEX2. Based on its ability to overcome the loss of Krp1, we suggest that Yps1 is not only structurally similar to the yapsins but also functionally related to these proteins. This suggestion is further strengthened by the fact that the S. cerevisiae YPS1 complements the krp1ts strain when overexpressed (see group B in Fig. 3).

YPS1 is the best-characterized yapsin (Egel-Mitani et al., 1990; Azaryan et al., 1993; Bourbonnais et al., 1994; Cawley et al., 1995; 1998; Ash et al., 1995). It is synthesized with an N-terminal presequence that is removed after translocation into the ER where a prosequence is removed by intramolecular cleavage immediately after Lys–Arg-67 (Cawley et al., 1998). The mature enzyme is also processed, by an as yet unidentified protease, into an α- and β-subunit that remain linked by a disulphide bond between Cys-117 and Cys-186 (Cawley et al., 1998). The N-terminal amino acid of the β-subunit, Asp-145, is situated within a loop domain of YPS1 arising from a large insertion of about 76 residues in comparison with pepsin. Very similar processing is predicted for YPS2; cleavage of the prosequence is likely to occur after Arg-65, and the large loop domain between the two catalytic residues suggests that the mature protein may be cleaved into α- and β-subunits. The prosequence of YPS3 is also removed by cleavage after a dibasic motif (Lys–Arg-47) (Olsen et al., 1999), but it lacks the loop region and probably remains as one polypeptide chain. We have no biochemical data concerning the maturation of the S. pombe Yps1 but predict cleavage of the presequence after Ser-17 and cleavage of the prosequence after Lys–Arg-45. It remains to be investigated whether these cleavages occur and whether removal of the prosequence is an intramolecular event.

Processing at the C-terminus of the S. cerevisiae yapsins leads to the addition of a glycosylphosphatidylinositol (GPI) anchor (Ash et al., 1995; Cawley et al., 1995; Komano and Fuller, 1995). Sequence comparisons (Nuoffer et al., 1993; Udenfriend and Kodukula, 1995) have identified a putative GPI anchor attachment signal at the C-terminus of target proteins. The signal is composed of an attachment site, a spacer domain of approximately 8–12 amino acids and a terminal hydrophobic domain of 10–15 residues. The attachment site (or ω site) is a small amino acid and is usually followed by two other small amino acids. In S. cerevisiae, the most common amino acids for GPI attachment are asparagine and glycine (Caro et al., 1997; Hamada et al., 1998). The addition of GPI anchors occurs in the ER, and the GPI proteins are then transported to the cell surface where some remain attached to the plasma membrane, whereas others are processed further and incorporated into the yeast cell wall. Processing involves removal of the phospholipid moiety and binding of the glycan part of the GPI remnant to β1,6-glucans in the cell wall (Kapteyn et al., 1996; Müller et al., 1996; Van Der Vaart et al., 1997). A dibasic motif just before the ω site appears to act as a negative signal that prevents this additional processing, and proteins containing this motif remain in the plasma membrane (Hamada et al., 1998; 1999). Plasma membrane localization has been confirmed for YPS1 (Ash et al., 1995; Cawley et al., 1995) and YPS2 (Komano and Fuller, 1995) and is predicted for YPS3 and YPS6. In contrast, YPS7 does not contain a dibasic motif upstream of the predicted ω site and would be expected to be incorporated into the cell wall.

GPI biosynthesis is believed to be well conserved throughout evolution, and the mechanism in S. pombe appears to be similar to that in the budding yeast (Colussi and Orlean, 1997; Tiede et al., 1998). We would therefore expect a GPI anchor to be added to Asn-495 in Yps1, and the lack of an adjacent dibasic motif would be consistent with localization to the cell wall. A more detailed molecular analysis of the maturation of Yps1 is under way.

No physiological function has yet been assigned to the yeast yapsins, but their localization to the cell surface and their specificity for basic residues in proproteins would be consistent with a role in the processing of cell wall precursors or precursors of enzymes involved in cell wall synthesis or remodelling. Whether such processing is particularly important under stress conditions remains to be investigated (Komano and Fuller, 1995).

Only one other aspartyl protease has been identified in S. pombe. Sxa1 was identified through a mutant with a phenotype similar to that of a ras1 mutation (Imai and Yamamoto, 1992). It is constitutively expressed in cells of both mating types but is required for efficient mating only in P-cells and was presumed to have some role in adapting to the presence of the M-factor mating pheromone. Biochemical analysis failed to demonstrate any Sxa1-dependent degradation of M-factor, but an sxa1 mutant had much lower general proteolytic activity than a wild-type strain (Ladds et al., 1998). The substrate specificity of Sxa1 is not known, but it does not appear to be a yapsin. Not only is there no great sequence similarity to the S. cerevisiae yapsins, but overexpression of Sxa1 failed to rescue the krp1ts strain.

Glycoprotease

The growth defect of the krp1ts strain was rescued by overexpression of a zinc metalloprotease belonging to the M22 family of peptidases (Rawlings and Barrett, 1995). The relevant sequence was identified during the S. pombe genome sequencing project (accession number SPCC1259.10). Sequence analysis reveals homology to the glycoprotease subclass of the M22 family (Table 3, Fig. 6), and we propose calling the enzyme Pgp1 (S.pombeglycoprotease). A second putative glycoprotease has also been identified during the S. pombe sequencing project, and we propose calling this Pgp2 (accession number SPBC16D10.03). Overexpression of Pgp2 failed to rescue the krp1ts strain.

Table 3. Pairwise comparison of the S. pombe glycoproteases with each other and with related enzymes from S. cerevisiae.
 Pgp2QRI7YKR038c
  1. The full-length sequences were compared using the GCG program bestfit and are shown as percentage identity and percentage similarity.

Pgp126 (48)41 (61)24 (49)
Pgp2 26 (51)67 (81)
Figure 6.

Alignment of glycoproteases. Sequences were aligned using the pileup programme (GCG) with a GapWeight of 3.0 and a GapLengthWeight of 0.1. Gaps introduced to maximize the alignment are indicated by –. Residues identical in all the enzymes are indicated by black shading, and those conserved between each pair of enzymes are indicated by grey shading. The metal-binding residues are highlighted above the sequence (H157 and H161 in Pgp1).

Glycoproteases have been reported in bacteria, plants, nematodes, yeast and mammals. The prototypic member of the family is O-sialoglycoprotein endopeptidase from Pasteurella haemolytica (Abdullah et al., 1991; 1992). This neutral metallopeptidase is highly specific for O-sialoglycoproteins and does not cleave unglycosylated proteins, desialylated glycoproteins or glycoproteins that are only N-glycosylated. Two glycoproteases have been identified by sequencing projects in S. cerevisiae, but there is almost no functional information available about these enzymes. QRI7, for example, is known only not to be essential for cell viability (Simon et al., 1994; Saiz et al., 1996), whereas YKR038C is known to be transcribed at low levels in both haploid and diploid cells grown under standard laboratory conditions (Richard et al., 1997). There is no information available about the substrate specificity of the yeast enzymes, but it is perhaps significant that cleavage of glycophorin A by the P. haemolytica glycoprotease is primarily after Arg-31 (Sutherland et al., 1992). A similar specificity for Pgp1 could explain its ability to recognize substrates processed by Krp1.

Complementing the loss of Krp1

Our initial characterization of the proteases was performed in the krp1ts strain and might be complicated by the low level of Krp1 activity that remains in these cells. The plasmids were therefore introduced into a diploid S. pombe strain in which one copy of the krp1 gene is replaced by the ura4+ cassette (JY324, krp1+/krp1::ura4+) (Davey et al., 1994). Sporulating this diploid does not normally generate viable progeny that are krp1 (and hence ura4+), but these can be obtained by first transforming the diploid with a plasmid-borne copy of krp1. JY324 was transformed with each of the plasmids identified during the screen, and spores were plated on medium lacking both uracil (to select for krp1::ura4+ progeny) and thiamine (to induce expression from nmt1). Each of the proteases identified during the screen was able to support the growth of the krp1::ura4+ progeny, suggesting that they could complement the complete loss of Krp1 activity. To confirm this suggestion and to prepare more stable and better characterized strains lacking the chromosomal krp1, we used the ura4+ cassette to disrupt krp1 in isolates of JY383 containing pREP vectors expressing the different proteases. Southern blot analyses confirmed disruption of the chromosomal krp1 in each strain, and their ability to grow demonstrated that each protease could complement Krp1 (Table 4).

Table 4. Δkrp1 strains containing pREP3X-protease constructs.
 Krp1Isp6Psp3Yps1Pgp1
  1. Growth of strains lacking krp1 (Δkrp1) but containing pREP vectors expressing the different proteases was monitored at 29°C in minimal medium lacking thiamine (nmt1 promoter is induced). For comparison, an isogenic wild-type strain (krp1+) has a generation time of 3.6 h under these conditions.

Generation time (h)3.545.34.84.9

Substrate specificity

Protease specificity was monitored using fluorogenic peptides that contain the putative cleavage site coupled to methyl coumarin so that processing liberates a fluorescent reporter. Total cell extracts were prepared from cultures of the Δkrp1 strains containing pREP vectors expressing the different proteases, and these were assayed against a range of substrates containing various basic motifs (Table 5). Total extracts (rather than, for example, microsomal fractions) avoid complications that may be caused by the different proteases being localized to different cellular compartments. We are confident that the use of such extracts did not alter the specificity of the enzymes because our results for Krp1 are almost identical to those reported previously for Krp1 in yeast microsomes (Davey et al., 1994), in membranes prepared from injected frog oocytes (Davey et al., 1994), and in a cell-free system (Powner and Davey, 1998). The results confirmed that Krp1 was specific for dibasic motifs involving arginine residues and preferred Lys–Arg to Arg–Arg. There was limited cleavage of Lys–Lys motifs and no detectable processing after single basic residues. Isp6 was also specific for dibasic motifs, although it appeared to prefer cleavage after lysine rather than arginine. Psp3, the third serine protease, cleaved all the substrates used in the study. The ability of Yps1 to cleave after basic residues confirms that it is a member of the yapsin family (Egel-Mitani et al., 1990; Bourbonnais et al., 1994; Komano and Fuller, 1995; Ledgerwood et al., 1996; Komano et al., 1999).

Table 5. Total cell extracts prepared from strains lacking krp1 (Δkrp1) but containing pREP vectors expressing the different proteases were assayed for their ability to process fluorogenic substrates.
 Krp1Isp6Psp3Yps1Pgp1
  1. Reactions were performed at 23°C, and samples were preincubated at this temperature for 5 min before mixing. All assays contained equivalent amounts of the different extracts (as judged by protein content) and were performed under initial rate conditions. Activities are expressed as percentages of the cleavage rate of 25 pmol min−1 obtained for Krp1 against Boc.Arg–Thr–Lys–Arg–MCA (100%). Extracts were prepared and assayed on at least three occasions, and each result was within the range indicated.

Boc.Arg–Thr–Lys–Arg–MCA10020–2570–7560–65< 1
Boc.Gln–Arg–Arg–MCA50–555–1075–8060–65< 1
Boc.Gly–Lys–Lys–MCA1–550–5545–50< 1< 1
Boc.Gln–Gly–Arg–MCA< 1< 140–4550–55< 1

Many explanations can be advanced for the apparent inactivity of Pgp1. It is possible, for example, that the protease inhibitors included in the extract and assay buffers inhibit Pgp1. Unfortunately, none of the extracts prepared in the absence of inhibitors was active, and we were unable to investigate this possibility further. Another explanation could be that Pgp1, like other glycoproteases, does not process unglycosylated proteins and would not therefore cleave the short peptides available for monitoring enzyme specificity. Glycosylated substrates such as glycophorin A could be used to investigate the activity of Pgp1, but this is beyond the scope of this current study.

Processing of the P-factor precursor

P-factor is an unmodified peptide of 23 amino acids, which is encoded by the map2 gene (Imai and Yamamoto, 1994). Sequence analysis predicts a primary translation product with an N-terminal presequence and four repeats of the mature P-factor sequence that are separated by spacer regions containing pairs of basic residues (Lys–Arg). Processing of the Map2 product by Krp1 releases the individual subunits (Davey et al., 1994) that are then trimmed to generate the mature pheromone. To investigate the ability of the various proteases to process the pheromone precursor, Δkrp1 strains containing pREP vectors expressing the different proteases were assayed for the release of P-factor.

Pheromone stimulation is required for S. pombe to undergo meiosis and sporulation (Willer et al., 1995), and a normal heterozygous diploid cell sporulates in nitrogen-free medium because it produces both pheromones and both receptors and will autostimulate its response pathway. In contrast, mutant strains defective in the early subfunctions of the mating-type locus will only sporulate on exposure to added pheromone. The diploid strain JY403 (mat1-Pm/mat1-Pc) is defective in the meiosis-specific subfunction of the mat1-P mating locus and is unable to produce either P-factor or the M-factor receptor. It will sporulate only if supplied with P-factor (Egel et al., 1994). Haploid strains being assayed for P-factor secretion were mixed with an excess of the diploid tester strain and spotted onto nitrogen-free medium. After 48 h, the plates were exposed to iodine vapour to stain the halos of azygotic asci surrounding the cells secreting P-factor (Fig. 7). All the strains produced halos, suggesting that each protease could process the Lys–Arg motif within the P-factor precursor. Processing by Pgp1 appears to contradict our earlier enzymatic analysis (Table 5), but Map2 is glycosylated (Davey et al., 1994) and may be a suitable substrate for this enzyme. The width of the sporulating halos can provide a semi-quantitative estimate of pheromone production (Willer et al., 1995), and the smaller halos observed with strains containing Isp6 would be consistent with the lower activity recorded for this enzyme against substrates containing Lys–Arg motifs.

Figure 7.

P-factor was assayed by its ability to induce meiosis and sporulation in a diploid mat1-M/mat1-Pc strain (JY403). The haploid strains being assayed for P-factor secretion and the responsive diploid strain were grown to stationary phase in MSL and mixed in a 50-fold excess of the diploid strain. Dilutions of this mixture were spotted on MSA plates (MSL with agar) and incubated for 48 h. Plates were exposed to iodine vapour, which stains the halos of azygotic asci that surround the cells secreting P-factor, and photographed. The width of the halos provides a semi-quantitative estimate of P-factor production. The haploid strains used were JY1103 (mat1-P, Δmat2,3::LEU2, leu1-32, ura4-D18, krp1::ura4+ with pREP-Krp1), JY1104 (with pREP-Isp6), JY1105 (with pREP-Psp6), JY1106 (with pREP-Yps1) and JY1107 (with pREP-Pgp1).

Experimental procedures

Strains, reagents and general methods

The S. pombe strains used in this study are listed in Table 6. The diploid strain (JY403) (Egel et al., 1994) has a stable mating type because the silent mating-type cassettes (mat2 and mat3) have been deleted, and the mat1 cassettes are unable to receive the double-strand break usually associated with mating-type switching and mitotic recombination. The strain is homozygous for the sxa2 mutation (Imai and Yamamoto, 1992) and is unable to produce the Sxa2 carboxypeptidase that degrades extracellular P-factor (Ladds and Davey, 2000). Construction and characterization of strains containing a temperature-sensitive Krp1 are described elsewhere (Ladds et al., 2000). General yeast procedures were performed as described previously (Davey et al., 1995), using lithium acetate for the transformation of yeast. Culture media used were YE (yeast extract; for routine cell growth), PM (a defined minimal medium) and MSL (a minimal medium that promotes sexual differentiation; Egel et al., 1994). DNA manipulations were performed by standard methods. All constructs were sequenced by the dideoxynucleotide method using double-stranded DNA as template and a series of oligonucleotide primers designed to generate overlapping sequence data. Oligonucleotides were synthesized by Alta Bioscience. Unless stated otherwise, amplification by PCR used Pwo DNA polymerase (from Pyrococcus woesei) according to the supplier's instructions (Boehringer Mannheim Biochemicals). This polymerase has a 3′−5′ exonuclease (proofreading) activity that reduces the introduction of errors during amplification. Most sequence analysis was performed by the GCG package. Prediction of presequence cleavage sites was performed on-line at http://www.cbs. dtu.dk/services/SignalP/ (Nielsen et al., 1997), and prediction of transmembrane domains at http://www.biokemi.su.se/~server/toppred2/ (Von Heijne, 1992).

Table 6. S. pombe strains used in this study.
StrainGenotype
JY383 mat1-P, Δmat2,3::LEU2 , leu1-32, ura4-D18
JY4032n mat1-M int-H1::ura4+, Δmat2,3::LEU2, leu1-32, ade6-M216, sxa2-563
mat1-Pc::int-1, Δmat2,3::LEU2, leu1-32, ura4-D18, sxa2-563
JY654 mat1-P, Δmat2,3::LEU2 , leu1-32, ura4-D18, krp1 ts and ura4+
JY965 mat1-P, Δmat2,3::LEU2 , leu1-32, ura4-D18, krp1 ts
JY1103 mat1-P, Δmat2,3::LEU2 , leu1-32, ura4-D18, krp1::ura4 + with pREP-Krp1
JY1104As JY1103 but with pREP-Isp6
JY1105As JY1103 but with pREP-Psp6
JY1106As JY1103 but with pREP-Yps1
JY1107As JY1103 but with pREP-Pgp1

Monitoring growth, viability and morphology

Minimal medium was inoculated (at ≈ 104 cells ml−1) with exponentially growing wild-type (JY383) or krp1ts (JY965) strains and incubated with aeration at 23°C until ≈ 105 cells ml−1. Cultures were divided in two and one-half transferred to 37°C. Incubations were in water baths to minimize the time taken to shift temperatures – control experiments found that a 100 ml culture at 23°C took less than 10 min to reach 37°C. Cell numbers were monitored using a Coulter Channelyser (Beckman Coulter), and viability was determined by colony formation on YE at 23°C. Morphology was recorded 28 h after the shift to 37°C. Unfixed cells were stained with Calcofluor white (0.1 mg ml−1), embedded in agarose and viewed using a Zeiss Axioskop.

Screen for multicopy suppressors of krp1ts

A krp1ts strain (JY965) was transformed with an S. pombe cDNA library in pREP3 (a gift from Masayuki Yamamoto, University of Tokyo). The library was prepared from a homothallic (h90) strain 4 h after transfer to nitrogen-free medium, has an average insert of about 1800 bp and contains ≈ 150 000 isolates. Transformants were plated onto medium lacking thiamine (to induce expression from the nmt1 promoter) and incubated at 23°C for 24 h to allow expression of the relevant cDNAs. Plates were then transferred to 37°C to identify constructs able to rescue the krp1ts defect.

Constructs from S. pombe

Construction of pREP vectors containing wild-type Krp1 (JD561), the temperature-sensitive Krp1ts (JD1016) (containing the F425S mutation) and the inactive Krp1[S371A] (JD1075) (the catalytically important serine residue at position 371 is converted to alanine; Powner and Davey, 1998) has been described previously (Ladds et al., 2000). The sxa1 open reading frame (ORF) was amplified from genomic DNA using the sense primer JO373 (ggggatccaccATGAAGGCTTCTTTCTTTGTATTTGC; lower case letters are not complementary to sxa1, the primer includes an emboldened BamHI site and changes the sequence immediately upstream of the initiator ATG from ATAAT to one that is expected to be more favourable for translation; Yun et al., 1996) and the antisense primer JO374 (ggggatccACTCAAGCGAAAAGTAAAGAGATCAGAG; introduces a BamHI site immediately downstream of the TCA stop anticodon) and cloned into the BamHI site of the pREP vector to give JD686. The pgp2 ORF was amplified from genomic DNA using the sense primer JO1164 (ggggatccaccATGGGAAAACCTTTAATTG; introduces a BamHI site and a sequence predicted to give improved translation immediately upstream of the ATG) and the antisense primer JO1065 (ggggatccAGTATTAATCCCTCCATG; introduces a BamHI site just downstream of the TTA stop anticodon) and cloned into the BamHI site of the pREP vector to give JD1569.

Constructs from S. cerevisiae

The KEX2 ORF was amplified from genomic DNA using the sense primer JO901 (ggcatATGAAAGTGAGGAAATATATT AC; introduces an NdeI site immediately upstream of the ATG) and the antisense primer JO902 (ggcatatgTCACGATCGTCCGGAAGATG; introduces an NdeI site immediately downstream of the TCA stop anticodon) and cloned into the NdeI site of the pREP vector to give JD1568. The YPS1 ORF was amplified from genomic DNA using the sense primer JO907 (ggcatATGAAACTGAAAACTGTAAGATC; introduces an NdeI site immediately upstream of the ATG) and the antisense primer JO908 (ggcatatgTCAGATGAATGCAAAAAGAAGAG; introduces an NdeI site immediately downstream of the TCA stop anticodon) and cloned into the NdeI site of the pREP vector to give JD1168.

Disruption of krp1

JY383 (mat1-P, Δmat2,3::LEU2, leu1-32, ura4-D18) was transformed first with pREP vectors expressing the different proteases and then with a linearized krp1::ura4+ construct (from JD638). Disruption of the chromosomal krp1 locus was confirmed by Southern blotting to generate JY1103 (mat1-P, Δmat2,3::LEU2, leu1-32, ura4-D18, krp1::ura4+ with pREP-Krp1), JY1104 (with pREP-Isp6), JY1105 (with pREP-Psp6), JY1106 (with pREP-Yps1) and JY1107 (with pREP-Pgp1).

Protease assays

Yeast strains (Δkrp1) containing pREP vectors expressing the different proteases were grown in minimal medium lacking thiamine to ≈ 1 × 107 cells ml−1. Cells were washed once in extract buffer (EB) (100 mM Tris-HCl, pH 6.8, 10 mM EDTA) and then resuspended as a thick paste in EBplus (EB with protease inhibitors: 1 mM phenylmethylsulphonyl fluoride, 1 mM N-α-tosyl-l-lysine chloromethyl ketone, 1 mM tosyl-l-phenylalanine chloromethyl ketone, 1 mM leupeptin and 0.1 mM pepstatin). Acid-washed glass beads (425–600 µm in diameter) were added, and cells were homogenized by vortexing. Reaction mixtures were assembled in two halves, one containing cell extract and the other containing 100 µM fluorogenic peptide in assay buffer (100 mM HEPES, pH 6.5, 2 mM CaCl2, 0.5% Triton X-100 and the same protease inhibitors as in EBplus). Fluorogenic peptides were purchased from Peninsula Laboratories (Boc.Gln–Arg–Arg–MCA, Boc.Gly–Lys–Lys–MCA and Boc.Arg–Thr–Lys–Arg–MCA) or Sigma Chemical (Boc.Gln–Gly–Arg-MCA). Reactions were performed at 23°C, and samples were preincubated at this temperature for 5 min before mixing. The increase in fluorescence was monitored with a luminescence spectrometer (model LS-5; Perkin-Elmer) at an excitation wavelength of 365 nm and an emission wavelength of 460 nm.

Halo assays for P-factor

The responsive diploid strain (JY403) and the haploid strains being assayed for P-factor secretion (Δkrp1 containing pREP constructs expressing the different proteases) were grown to stationary phase in MSL and mixed in a 50-fold excess of the diploid strain. Dilutions were spotted on MSA plates (MSL with agar), incubated for 48 h and exposed to iodine vapour.

Acknowledgements

We thank the Wellcome Trust (049472), the Cancer Research Campaign (SP1972) and the Lister Institute of Preventive Medicine (J.D.) for funding. Dale Powner and Kevin Davis were involved in the early stages of this project. and we thank Masayuki Yamamoto (University of Tokyo), Olaf Nielsen and Richard Egel (University of Copenhagen) for plasmids and strains.

Ancillary