Purification and Characterization of a Novel Metalloendopeptidase from Saccharomyces cerevisiaet

We previously identified an activity in the soluble fraction of the yeast Saccharomyces cerevisiae that is a candidate for catalyzing the proteolytic maturation of farnesylated -CXXX precursor polypeptides. We describe here a 1259-fold purification of this activity by chromatography on DEAE-cellulose, hydroxylapatite, phenyl-Sepharose, and Sephacryl S-200. Sodium dodecyl sulfate gel electrophoresis of this preparation demonstrated a single 68-kDa polypeptide chain. The experimentally determined N-terminal amino acid sequence was identical at all 20 positions with residues 28-47 of the deduced sequence of the S. cerevisiae YCL57w gene product. This analysis suggests that the YCL57w gene encodes this enzyme and that the initial translation product may contain a leader peptide. Its complete deduced amino acid sequence shows significant homology to a number of zinc metallopeptidases and is most closely related to rat metalloendopeptidase 24.15 (E.C. 3.4.24.15), an enzyme that preferentially cleaves after hydrophobic residues. Using the purified yeast enzyme, we show a unique cleavage site in the peptides bradykinin and P-neoendorphin four residues from the C-terminus on the carboxyl side of a hydrophobic amino acid. The cleavage pattern for neurotensin revealed a major site three residues from the C-terminus also on the carboxyl side of a hydrophobic residue and a minor site four residues from the C-terminus of the peptide. This specificity is similar to that of rat endopeptidase 24.15 and may explain why the farnesylated peptide employed in our studies is a good substrate for the yeast enzyme. However, subcellular localization experiments revealed that this activity does not appear to be cytosolic; its distribution follows that of Golgi and vacuolar marker enzymes. These results suggest that this enzyme may not be involved in the physiological processing of cytosolic -CXXX containing proteins but may be important in other types of protein or peptide metabolism. In eukaryotic cells, a number of proteins and polypeptides whose initial translation product contain a -Cys-Xaa-XaaXaa sequence at the C-terminus, where -Xaa represents a variety of amino acid residues, can be targeted for a series of posttranslational modifications that include sequential lipidation with either a CIS farnesyl or C ~ O geranylgeranyl moiety on the cysteine residue, proteolysis of the three terminal amino acids, and a-carboxyl methyl esterification of the newly exposed cysteine residue (Clarke, 1992; Cox & Der, 1992; Schafer & Rine, 1992). The initial lipidation step has been widely studied and is well characterized in several systems [for reviews, see Maltese (1990), Casey (1992), Glomset et al. (1992), and Sinensky and Lutz (1992)l. The final methylation step has also been characterized in yeast (Hrycyna & Clarke, 1990; Hrycyna et al., 199 1) and mammalian cells (Stephenson & Clarke, 1990, 1992; Perez-Sala et al., 1991; Volker et ai., 1991). On the other hand, less is known about the intermediate processing step involving the proteolytic removal of the three terminal amino acids from the isoprenylated precursor polypeptide. Indirect evidence for this proteolytic event was inferred from the isolation of a-carboxyl methylated isoprenylcysteine derivatives from proteins whose initial translation products contain an additional three C-terminal residues (Ong et al., 1989; Ota & Clarke, 1989; Stimmel et al., 1990; Yamaneet al., 1990, 1991). Guiterrez et al. (1989) directly showed the removal of the three terminal amino acids from This work was supported in part by Grant GM-26020 from the National Institutes of Health to S.C. C.A.H. was supported in part by United States Public Health Service Training Grant GM-07185. * To whom correspondence should be addressed. @ Abstract published in Advance ACS Abstracts, October 1 , 1993. 0006-2960/93/043211 293%04.00/0 the mammalian p21ras protein. By site-directed mutagenesis, a unique tryptophan residue was introduced into the protein precursor at each of the three C-terminal positions, and the loss of the residue was monitored during biosynthetic processing in vivo. Similarly, Fujiyama and Tamanoi (1 990) demonstrated loss of the three C-terminal residues from the Saccharomyces cerevisiae RAS2 protein during its posttranslational biosynthetic processing. Finally, a membraneassociated proteolytic activity in canine microsomes has been identified that reportedly increases the membrane binding of farnesylated p21H-" 2-fold over the nonproteolyzed species (Hancock et al., 1991). Enzyme activities responsible for the cleavage of the three C-terminal amino acids in mammalian systems have recently been biochemically characterized. Ashby et al. (1992) identified a membrane-associated endoproteolytic activity in rat liver that releases the terminal three amino acids as a tripeptide and is specific for a farnesylated peptide substrate. A microsomal enzymatic activity from calf liver, specific for lipidated substrates, has also been identified that can specifically cleave a farnesylated tetrapeptide between the isoprenylated cysteine residue and the adjacent residue as well as the related farnesylated triand dipeptides (Ma et al., 1992; Ma & Rando, 1992). More recently, a distinct activity from brain microsomal membranes capable of sequentially removing the three terminal amino acids from a chemically synthesized C-terminal heptapeptide of mouseN-ras protein was identified (Akopyan et al., 1992). This activity has been characterized as a novel thiol-dependent, serine type carboxypeptidase that displays a higher affinity for the farnesylated substrate than its non-farnesylated analog (Akopyan et al., 1992). Inter@ 1993 American Chemical Society 11294 Biochemistry, Vol. 32, No. 42, 1993 estingly, no soluble proteolytic activities capable of cleaving the three terminal amino acid residues from farnesylated precursors have been identified in mammalian systems. Proteolytic activities recognizing synthetic isoprenylated peptide substrates have also been characterized in the yeast S. cerevisiae. We have identified at least three distinct activities in yeast that can catalyze the cleavage of the three COOH-terminal amino acids from the synthetic peptide substrate N-acetyl-KSKTK[S-farnesyl-Cys] VIM in vitro, a membrane-associated enzyme and two soluble activities, one of which has been identified as vacuolar carboxypeptidase Y (Hrycyna & Clarke, 1992). The membrane-associated enzyme is similar in its inhibitor specificity to an activity identified by Ashby et al. (1992) that catalyzes the removal of the three terminal amino acids as a tripeptide from a farnesylated peptide substrate in vitro. We characterized the partially purified soluble activity, also identified by Ashby et al. (1992), as a 110-kDa enzyme that initially appeared to be a metallocarboxypeptidase cleaving both farnesylated and nonfarnesylated -CXXX containing peptides but not unrelated peptides (Hrycyna & Clarke, 1992). In this work, we describe the purification of this soluble activity, its biochemical characterization, and the identification of the gene encoding this protein. Our results suggest that the activity is not, in fact, a carboxypeptidase but instead represents a novel noncytosolic zinc metalloendopeptidase encoded by the YCL57w gene on chromosome I11 in S. cerevisiae. The localization of this enzyme suggests that it may not play a role in the posttranslational maturation of isoprenylated protein precursors but instead may be important in peptide metabolism in yeast. EXPERIMENTALPROCEDURES Yeast Strains, Media, and Growth Conditions. S. cerevisiae strains used in this study are ABYSl (MATapral prbl prcl cpsl ade) (Achstetter et al., 1984) and SM1188 (MATa leu2 ura3 hislcanl s te l4k:TRPI) (Hrycynaet al., 1991). Unless otherwise indicated, strains were propagated on YEPD medium containing 1% (w/v) yeast extract (Difco), 2% (w/ v) Bacto-Peptone (Difco), and 2% (w/v) D-glucose. Preparation of Crude Soluble Fraction from S. cerevisiae. Cells were grown to an ODm of 0.9-1.2. The soluble fraction was prepared essentially as described previously (Hrycyna & Clarke, 1990) except 0.6 M mannitol in 10 mM Tris-HC1, pH 7.45, was used as the lysis buffer (Jazwinski, 1990). Synthetic Peptide Substrates. The synthetic peptide N-[ *4C]Acetyl-~-Ly~-~-Ser-~-Lys-~-Thr-~-Lys-[S-trans-transfarnesyl-~-Cys]-~-Val-~-Ile-~-Met (N-Ac-KSKTK[S-farnesyl-CysIVIM was synthesized as previously described (Hrycyna & Clarke, 1992). Bradykinin, 8-neoendorphin, and neurotensin were purchased from Sigma. In Vitro Coupled Peptidase Assay and Inhibition Assays. Peptidase enzyme activity assays and inhibition assays were performed essentially as previously described (Hrycyna & Clarke, 1992) except that the potential inhibitor was added directly to the reaction mixture without prior preincubation with the peptidase activity. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) Analysis. SDS-PAGE was performed using the buffer system described by Laemmli (1 970). Hrycyna and Clarke Electrophoresis sample buffer was added to aliquots of the soluble fraction or active column chromatography fractions [final concentrations of 60 mM Tris-HC1, pH 6.8,2% (w/v) SDS, 12% (v/v) glycerol, 0.001% Bromophenol Blue and 0.7 M 8-mercaptoethanol] and heated for 5 min at 100 OC. Electrophoresis was carried out on 8% (w/v) acrylamide, 0.28% N,N-methylenebisacrylamide resolving gels using high molecular weight markers (Bio-Rad) or prestained high molecular weight markers (Bio-Rad) as standards. Gels were stained in 50% methanol (v/v), 10% acetic acid (v/v), and 0.1% Coomassie brilliant blue (w/v) for 1 h and destained overnight in 10% acetic acid (v/v), and 5% methanol (v/v). Subsequently, the protein bands were further visualized by silver staining (Ausubel et al., 1991). Analysis of Cleavage of Synthetic Peptide Substrates. Bradykinin, neurotensin, and j3-neoendorphin were dissolved in water to a final concentration of 1 mg/mL and purified on a C4 Vydac reverse-phase column (300-A pore diameter, 4.6 mm inner diameter, 250-mm length) monitored by UV absorption at 214 nm. The column was equilibrated in 100% solvent A at room temperature (solvent A is 0.1% trifluoroacetic acid in water, and solvent B is 0.1% trifluoroacetic acid, 99% acetonitrile, and 0.9% water) and eluted with a linear 2.4% solvent B/min gradient at a flow rate of 1 mL/min. The amino acid composition of each peptide was determined by a modification of the method of Jones and Gilligan (1983). Solvent C is 0.1 M sodium acetate, pH 7,22/tetrahydrofuran/ methanol (895:10:95), and solvent D is 100% methanol. The o-phthalaldehyde derivitizing solution is 0.4% (w/v) o-phthalaldehyde (Fluka), 10% (v/v) methanol, 0.4% j3-mercaptoethanol, and 0.8 mg/mL Brij-35 (Sigma) in 0.4 M potassium borate, pH 10.4. The o-phthalaldehyde-derivatized amino acids were separated on a Waters Resolve CIS reversephase column (3.9" inner diameter X 150" length, 5-pm spherical silica) equilibrated in 100% solvent C. The column was eluted using the following gradient at a flow rate of 1.7 mL/min: 0-1 min, &20% solvent D; 1-1 2 min, isocratic with 20% solvent D; 12-17 min, 2040% solvent D; 17-20 min, isocratic with 40% solvent D; 20-25 min, 4060% solvent D; 25-31 min, isocratic with 60% solvent D; 31-32 min, 60100% solvent D. The column was then washed for 5 min in 100% solvent D and re-equilibrated for 15 min in 100% solvent C prior to the next injection. Proline content was determined for each of the peptides by the method of Cooper et al. (1984) on an analytical Econosphere C1g reverse phase column (Alltech/Applied Scientific, 3.9-mm inner diameter X 150mm length) except that the total volume of reagents used in the reaction was scaled down to 400 pL from 1.6 mL. In both analyses, fluorescence was monitored by a Gilson model 121 fluorometer. Quantification of the derivatized amino acids was based on the fluorescence of amino acid standards (Pierce Chemical Co., Standard H; 50-75 pmol). Each of these peptides, bradykinin (9.4 nmol), 8-neoendorphin (9.1 nmol), and neurotensin (6.0 nmol), was incubated in the presence of purified soluble peptidase (0.03,0.09, and 0.09 pg, respectively) at 31 OC for 35 min, 1.75 h, and 3 h, respectively. The purified soluble peptidase used is the pool of Sephacryl S-200 fractions 33 and 34 in 100 mM Tris-HC1, pH 7.53, from a fractionation of pooled phenyl-Sepharose fractions 180 and 18 1. The total volume of the bradykinincontaining reaction was 30 pL whereas the total volumes of the reactions with 8-neoendorphin and neurotensin were 40 pL. After the incubation, the entire reaction mixture was separated on a C4 Vydac reverse-phase column and monitored by UV absorption at 214 nm. Each of the peptide products Abbreviations: HPLC, high-performance liquid chromatography; PVDF, polyvinylidene difluoride; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; EDTA, (ethylenedinitrilo) tetraacetic acid; pHMB, phydroxymercuribenzoate; PMSF, phenylmethanesulfonyl fluoride.