key: cord-0749876-imkozbf1
authors: nan
title: Structure, biosynthesis, and localization of dipeptidyl aminopeptidase B, an integral membrane glycoprotein of the yeast vacuole
date: 1989-04-01
journal: J Cell Biol
DOI: nan
sha: a7a5f2d3cff5287f3c65c24cb77de1caf5872687
doc_id: 749876
cord_uid: imkozbf1

We have characterized the structure, biogenesis, and localization of dipeptidyl aminopeptidase B (DPAP B), a membrane protein of the yeast vacuole. An antibody specific for DPAP B recognizes a 120-kD glycoprotein in yeast that behaves like an integral membrane protein in that it is not removed from membranes by high pH Na2CO3 treatment. Inspection of the deduced amino acid sequence of DPAP B reveals a hydrophobic domain near the NH2 terminus that could potentially span a lipid bilayer. The in vitro enzymatic activity and apparent molecular weight of DPAP B are unaffected by the allelic state of PEP4, a gene essential for the proteolytic activation of a number of soluble vacuolar hydrolases. DPAP B is synthesized as a glycosylated precursor that is converted to the mature 120-kD species by carbohydrate addition. The precursor form of DPAP B accumulates in sec mutants (Novick, P., C. Field, and R. Schekman. 1980. Cell. 21:205-215) that are blocked at the ER (sec18) or Golgi apparatus (sec7), but not at secretory vesicles (sec1). Immunolocalization of DPAP B in wild-type or sec1 mutant cells shows that the protein resides in the vacuolar membrane. However, it is present in non-vacuolar compartments in sec18 and sec7 cells, confirming that the delivery of DPAP B is blocked in these mutants. Interestingly, DPAP B appears to stain the nuclear envelope in a sec18 mutant, which is consistent with the accumulation of DPAP B in the ER membrane at the restrictive temperature. These results suggest that soluble and membrane-bound vacuolar proteins use the same stages of the secretory pathway for their transport.

sor form of DPAP B accumulates in sec mutants (Novick, P., C. Field, and R. Schekman. 1980. Cell. 21:205-215) that are blocked at the ER (sec18) or Golgi apparatus (sec7), but not at secretory vesicles (secl). Immunolocalization of DPAP B in wild-type or secl mutant cells shows that the protein resides in the vacuolar membrane. However, it is present in nonvacuolar compartments in sec18 and sec7 cells, confirming that the delivery of DPAP B is blocked in these mutants. Interestingly, DPAP B appears to stain the nuclear envelope in a sec18 mutant, which is consistent with the accumulation of DPAP B in the ER membrane at the restrictive temperature. These results suggest that soluble and membrane-bound vacuolar proteins use the same stages of the secretory pathway for their transport.

T HE vacuole of the yeast Saccharomyces cerevisiae is considered to be the equivalent of the lysosome of mammalian cells because it contains a large number of hydrolytic enzymes and has an acidic pH (reviewed in Rothman and Stevens, 1988) . The biosynthesis and sorting of soluble vacuolar hydrolases, such as carboxypeptidase Y (CPY) t and proteinase A, have been examined. The localization of soluble proteins to the vacuole requires the early stages of the secretory pathway; i.e., ER and Golgi apparatus (Stevens et al., 1982) . In addition, localization determinants on CPY and proteinase A, as well as genes necessary for the efficient sorting of these proteins, have been identified (Vails et al., 1987; Johnson et al., 1987; Klionsky et al., 1988; Bankaitis et al., 1986; Robinson et al., 1988; Rothman et al., 1989) .

Much less is known about the sorting of membrane proteins to the vacuole. Studies on mammalian cells indicate that soluble and membrane-bound lysosomal proteins are sorted by distinct mechanisms (von Figura and Hasilik, 1986) . Several lysosomal membrane proteins have been identified (Chen et al., 1985; Lewis et al., 1985; Barriocanal et al., 1986; Lippincott-Schwartz and Fambrough, 1986; Waheed et al., 1988) , and of those examined, none possess the mannose-6-phosphate moiety required for sorting soluble lysosomal proteins (yon Figura and Hasilik, 1986) . Furthermore, several membrane proteins are found at normal levels in the lysosomes of I-cell fibroblasts, which missort soluble lysosomal hydrolases to the cell surface.

Several activities associated with the vacuolar membrane of yeast have been characterized (Rothman and Stevens, 1988) , including transport systems for Ca 2÷ (Ohsumi and Anraku, 1981) and basic amino acids (Ohsumi and Anraku, 1983) , a protein-translocating ATPase (Uehida et al., 1985; Kane et al., 1989) , an ct-mannosidase (Opheim, 1978) , and dipeptidyl aminopeptidase (DPAP) B (Suarez Rendueles et al., 1981; Julius et al., 1983; Bordallo et al., 1984) . Nothing is known about the transport or posttranslational modifications for any of the vacuolar membrane proteins. Analysis of the localization of ct-mannosidase and DPAP B in yeast mutants that missort soluble vacuolar proteins suggested that most of these mutants do not missort membrane-bound vacuolar proteins (Bankaitis et al., 1986; Robinson et al., 1988; Rothman and Stevens, unpublished results) . Thus, soluble and membrane-bound proteins appear to be directed to the vacuole by distinct mechanisms.

As a first step in the study of vacuolar membrane protein sorting and transport, we have characterized the structure, biogenesis, and localization of DPAP B. We show that DPAP B is an integral membrane protein that, like soluble vacuolar proteins, uses the early stages of the secretory pathway for its transport to the vacuole.

The genotypes of the Escherichia coli and S. cerevisiae strains used in these studies are listed in Table I . Yeast cultures were grown in minimal medium supplemented with the appropriate nutrients as previously described .

Restriction endonucleases and other enzymes used for cloning and sequencing were from New England Biolabs ( 

Restriction endonuclease and Bal 31 exonuclease digestions and ligations were performed as recommended by the suppliers. Plasmid purification, agarose gel electmplforesis, and DNA-mediated transformation of E. coil were performed according to standard procedures (Maniatis et al., 1982) . Plasmids were introduced into yeast cells by the lithium acetate transforma-tion method (Ito et al., 1983) . E. coli strain MC1061 was used for all plasmid manipulations and JM103 was used for M13 phage work. DNA sequencing was performed by the chain termination method of Sanger et al. (1977) using Ml3mpl8 and mpl9 vectors (Yanisch-Perron et al., 1985) . In one case, sequencing was primed using a synthetic oligonucleotide corresponding to nucleotides 441-459 of DAP2 (see Fig. 2 ), prepared at the University of Oregon Biotechnology Laboratory on a DNA synthesizer (model 380B; Applied Biosystems, Inc., Foster City, CA) as described (Ito et al., 1982) . DNA and amino acid sequence analyses were performed using the sequence analysis software package (version 4) of the University of Wisconsin Genetics Computer Group.

Plasmid pRG1 was constmc~xl by inserting the 4.7-kb Bam HI-Pst I (blunt-ended by treating with T4 DNA polymerase) DAP2 fragment (Fig.  1) , isolated from the original DAP2 plasmid YEpl3-GS13-4 (Julius et al., 1983) , into the Barn HI-Sma I sites of pUC12. Plasmid pGP3, a multicopy 2-#m plasmid containing the DAP2 gene, was constructed by inserting the 4.7-kb Bam HI-Pst I (blunt-ended) fragment from pRG1 into the Bam HI-Pvu II sites of YEp24.

A chromosomal DAP2 deletion was constructed by transplacing the genomic copy with the cloned gene into which the LEU2 gene was inserted in place of 1.3 kb of the DAP2 coding region (Rothstein, 1983;  Fig. 1 ). Plasmid pRGI was digested with BstE II and Kpn I, treated with T4 DNA polymerase to make blunt ends, and ligated with a 2.1-kb Hpa I fragment containing the LEU2 gene. The resulting plasmid (pGP2) was cut with Barn HI and Pst I and used to transform the yeast strain JHRY20-2C (leu2-3, 1eu2-112). Stable LEU+ transformants were purified and screened for the lack of thermolabile dipeptidyl aminopeptidase activity, assayed as described previously (Julius et al., 1983) .

Plasmid pC JR6, a multicopy 2-#m yeast vector containing the DPAP B coding region fused to the GAL/promoter, was constructed by inserting a 400-bp Eco RI-Hind III GAL/ promoter fragment (Johnston and Davis, 1984) into the Eco RI and Hind III sites of pSEYS, and subsequently subcloning into the Hind III site of this plasmid (pCJR5) the 2.9 kbp Hind IIl fragment of DAP2 (Fig. 1) , resulting in the fusion of the GAL/ promoter at position -90 of DAP2.

DPAP B antigen for production of antiserum was made in E. coli by fusing a portion of the DAP2 coding region into the E. coli expression vector priSe5, which contains the T4 phage lysozyme gene under the control of the TAC promoter (Muchmore et al., 1988 ; kindly provided by R. Dahlquist, University of Oregon). Plasmid pRGI was digested with Stu I, treated briefly with Bal 31 exonuclease, and further digested with Hind III. Fragments of "~1,650 bp were cloned into the Sna BI-Hind III sites of priSe5, resulting in the production of a hybrid protein consisting of the NH2-terminal seven residues of T4 phage lysozyme fused to the COOH-terminal half of DPAP B. Transformants were screened for an isopropyl ~-D-thiogalactopyranoside-(IPTG) inducible protein species by SDS-PAGE (Laemmli, 1970) of total E. coil cell extracts. A plasmid was isolated (pRG2) which produced a large amount (*5 % lof total cell protein) of a 55-kD species upon induction with isopropyl B-D-thiogalactopyranoside.

E. coil cells containing plasmid pRG2 were grown in LB broth plus ampicillin to saturation, then diluted 1:100 in 250 ml of fresh medium, and grown for 1 h at 37°C. Isopropyl/~-D-thiogalactopyranoside was added to 1 mM, and incubation was continued for 3 h before collecting the cells by cenlrifugation. The cells were lysed and the fusion protein was isolated from the insoluble fraction essentially as previously described (Kleid et al., 1981 ). An SDS-PAGE gel slice containing the fusion protein was resuspended in SDS elution buffer (1% SDS, 50 mM Tris-HCI, pH 8.0, 1 mM EDTA, 1 mM DTT, 100 mM NaCI) in a homogenizer (Dounce; Wheaton Instruments Div., Millville, NJ), incubated for 2 h at 37°C, and centrifuged at 12,000 g for 10 min. The supernatant was saved, and the pellet was resuspended in 5 ml of elution buffer and treated as above. The supernatants were combined, and the protein was precipitated by adding 4 vol of acetone, incubating for 30 min at -80°C, and centrifuging at 15,000 g for 10 rain.

A DPAP B-specific antibody was prepared by injecting DPAP B antigen into New Zealand White Rabbits (250 ftg DPAP B antigen per rabbit per injection) essentially as described (Vaitukaitis, 1981) . These antibodies were affinity purified as previously described (Stevens et aL, 1982) . Antiserum to phosphoglycerate kinase was prepared as described .

Yeast cells containing plasmid pGP3 were grown in minimal medium at 30°C to mid log phase. 12 OD600 U of cells (1.2 × 10 s cells) were spheroplasted as described previously , lysed in 50 mM sodium phosphate, pH Z5, 2 mM EDTA, and 0.5 mM PMSF, and diluted 100-fold in ice-cold 100 mM sodium carbonate, pH 11.5. An aliquot representing the whole extract fraction was removed, and after 30 rain on ice the lysate was centrifuged for 3 h at 100,000 g. The supernatant and whole extract fractions were neutralized with 1 N acetic acid, then precipitated in 5% TCA on ice. Precipitated protein was pelleted, washed twice in diethylether, dried, resuspended in sample buffer, and heated at 70°C until dissolved. The pellet fraction was homogenized in 50 mM Tris-HCl, pH 6.8, and solubilized by addition of SDS to a final concentration of 4%. The sample was incubated at 60°C until complete solubilization had occurred; i.e., the suspension became completely transparent. Sample buffer (without SDS) was added, and the samples were subjected to SDS-PAGE on an 8% polyacrylamide gel. Electroblotting of the gel onto nitrocellulose was carried out according to Burnette (1981) using a Trans-blot Apparatus from Bio-Pad Laboratories (Richmond, CA). The transferred proteins were detected by probing the nitrocellulose with antibodies to the appropriate antigens and subsequent labeling with 0.3 #Ci of IzSI-protein A. The dried nitrocellulose was exposed to film (XAR-5; Eastman Kodak Co., Rochester, NY) at -80°C.

Yeast cells were grown in minima[ media plus 50 tzM Na2SO4 to midlog phase, spun down in a clinical centrifuge, washed once and resuspended in sulfate-free media, and incubated at the appropriate labeling temperature for 15-30 min at 8 OD6o0 U/ml. 35S-H~SO4 was added (0.5-1.0 mCi) and cells were labeled and then chased with 10 mM Na2SO4 (except for the experiments in Fig. 5 B, which were chased with 10 mM Na~SO4, 1 mM cysteine, 1 mM methionine, and 0.1 mg/ml cycloheximide). 20D60o U of cells were then sedimented, converted to spheroplasts, and lysed in 0.1 ml of 1% SDS by heating for 3 min at 100°C. 0.9 ml of immunoprecipitation buffer was added, followed by 0.1 ml of IgG Sorb (prepared as suggested by the manufacturer). The mixture was centrifuged 15 min in a microfuge (model 235B; Fisher Scientific Co., Pittsburgh, PA) and the supernatant was treated with 2/~1 of affinity purified DPAP B antibody and incubated 60 min on ice. The immune complexes were sedimented with IgG Sorb and washed as before (Stevens et al., 1982) , and the final IgG Sorb cell pellet was resuspended in 0.1 ml of 1% SDS and heated 3 min at 100°C. 0.9 ml of immunoprecipitation buffer and 0.1 ml of IgG Sorb was added, and the immunoprecipitation protocol was repeated. Sample buffer was added to the final IgG Sorb pellet and the samples were heated for 3 min at 100°C and centrifuged. Supernatant fractions were transferred to new tubes and aiiquots (normalized for the amount of 35S-H2SO4 incorporated) were analyzed by SDS-PAGE and fluorography as described previously .

Endo F treatment (Elder and Alexander, 1982) was carried out on immunoprecipitated samples of DPAP B essentially as described . A time course of endo F treatment was performed by adding 300 mU ofendo F to DPAP B immunoprecipitated from SEY5186 cells (containing plasmid pGP3) and withdrawing aliquots at the times indicated in Fig. 6 .

Yeast cells containing a disruption of the chromosomal DAP2 locus and transformed with the plasmid pC JR6 were grown in minimal media plus 2% raffinose to an OD600 of 1.0, and galactose was added directly to the cultures to a final concentration of 3%. The cultures were then incubated at either 34°C (SEC +, secl8, and secl) or 38°C (SEC + and secT) for 2 h and the cultures were put on ice. The cells were fixed with formaldehyde, spheroplasted, and prepared for immunofluorescence as described (Kilmartin and Adams, 1984; Adams and Pringle, 1984; Wittenberg et al., 1987) . In some cases, fixed spheroplasts were treated with SDS by resuspending in 1 ml of 1.2 M sorbitol/PBS (prepared as described by Adams and PringIe, 1984) plus 5 % SDS, mixed gently, centrifuged at 8,000 g, and washed twice in 1.2 M sorbitol. FITC-conjugated sheep anti-rabbit IgG second antibody was used for staining DPAP B antibody. Nuclei were stained with 4'6diamidino-2-phenyl-indole (DAPI) as described (Kilmartin and Adams, 1984) .

Cells were viewed and photographed using a Zeiss Axioplan Photomicroscope (equipped for epifluorescence at excitation wavelengths appropriate for DAPI and FITC fluorescence) and film and developer (TMAX-400; Eastman Kodak Co.). Photomicrographs are shown at a magnification of 1,000.

Yeast cells contain two membrane-associated DPAP activities that differ in their subcellular locations and thermal stabilities (Suarez Rendueles et al., 1981; Julius et al., 1983). DPAP B is a thermolabile enzyme associated with the vacuolar membrane (Bordallo et al., 1984) . DPAP A, the STE/3 gene product, is a nonvacuolar, heat-stable enzyme that participates in the proteolytic maturation of the mating pheromone a-factor (Fuller et al., 1988 stel3 mutant (Sprague, G., and Ira Herskowitz, unpublished results; Julius et al., 1983) . Another gene, called STE/Y (also called DPP2), was cloned by its ability to complement the or-factor maturation defect of a ste13 mutant when expressed from a high copy number plasmid (Sprague, G., and I. Herskowitz, unpublished results; Julius et al., 1983) . The gene dosage of DPP2 had a direct effect on the specific activity of DPAP B in extracts of yeast cells (Julius et al., 1983; Sprague, G., unpublished results) . Strains transformed with a multicopy plasmid containing DPP2 on a 4.7-kb Bam HI-Pst I fragment (Fig. 1) showed a 10-20-fold increase in DPAP B activity, whereas strains carrying a disrupted genomic copy of DPP2 (dpp2-za2::LEU2, see Materials and Methods; see Fig. 4 ) showed no measurable DPAP B activity in assays of whole cell extracts (data not shown). Also, a dpp2-A2::LEU2/DPP2 diploid strain had •50% of the DPAP B activity of a DPP2/DPP2 diploid (data not shown).

Mutants deficient in DPAP B activity have been previously described (Suarez Rendueles and Wolf, 1987) . These mutants, including one with an exceptionally thermolabile DPAP B activity, harbored mutations that fell into a single complementation group, called DAP2, and failed to complement the dpp2-za2 mutation. Thus, DPP2 and DAF2 define the same locus, which encodes DPAP B. Henceforth, DAP2 will denote the gene encoding DPAP B. No phenotype other than the lack of DPAP B was observed in strains disrupted at the DAP2 locus (Sprague, G., unpublished results; data not shown), consistent with the results of Suarez Rendueles and Wolf (1987). 

We deduced the primary structure of DPAP B by sequencing DAF2. Fig. 1 shows the restriction map of DAP2 and the sequencing strategy for the 3,135-bp Barn HI-Acc I fragment.

The complete nucleotide and deduced amino acid sequences are presented in Fig. 2 . 459 bp of 5' and 153 bp of 3' flanking sequences are included in addition to the 2,523 bp of coding sequence. Translation of DPAP B is predicted to initiate at the methionine codon numbered 1 (Fig. 2) . The sequenced region contains a long open reading frame coding for a protein of 841 amino acids, with eight canonical sites for asparaginelinked (N-linked) glycosylation. A 24-bp segment located from -210 to -187 relative to the putative translation initiating codon may function as a reiterated TATA-Iike box for initiation of transcription. An identical sequence is found in a similar location upstream of the yeast STE3 gene (Hagen et al., 1986) . The predicted molecular mass from the deduced amino acid sequence is 96,429 D, which is similar to the apparent molecular mass of DPAP B after removal of N-linked carbohydrate (see below).

Hydropathy analysis of the amino acid sequence (Kyte and Doolittle, 1982;  Fig. 3 A) reveals a single hydrophobic domain, beginning 30 residues from the NHz terminus, of sufficient length and hydrophobicity to span a lipid bilayer (Adams and Rose, 1985; Davis and Model, 1985) . Thus, the predicted topology of DPAP B is that of a type II (or group B; Fig. 3 B) integral membrane protein (Garoff, 1985; Singer et al., 1987) , with an NH2-terminal cytoplasmic do-main and a lumenal COOH terminus. The hydrophobic domain presumably functions as both an ER-targeting signal and a membrane anchor, as is true of other type II integral membrane proteins, such as the neuraminidase of influenza virus (Bos et al., 1984) and the transferrin receptor (Zerial et al., 1986) . Further support for this disposition of DPAP B in the membrane is the observation that DPAP B is a glycoprotein (see below), and that all sites for N-linked glycosylation are included in the COOH-terminal domain, which therefore must be the lumenal domain.

No significant sequence similarity has been found between DPAP B and any other vacuolar or tysosomal protein for which sequence information is available. However, strong homology has been found between DPAP B and DPAP A, the product of the STE/3 gene (Flanagan, C., and J. Thorner, personal communication; see below).

To facilitate the biochemical characterization of DPAP B, a DPAP B-specific antibody was prepared (see Materials and Methods). Lanes 1 and 2 of Fig. 4 show that the antibody is specific for the DAP2 gene product in yeast. Isogenic DAP2 and dap2-A2 cells were labeled with 355-H2804 for 30 min and chased for 30 min in the presence of 10 mM NazSO4, and immunoprecipitations were performed from cell extracts. Two species were precipitated from the DAP2 strain (Fig. 4, lane 2) : a major species of ,x,120 kD and a lesser amount of a ll0-kD form. Both of these bands were absent from the dap2-za2 strain (lane/), indicating that they represent the products of the DA/~ locus. 2) JHRY20-2CA3 (A) and JHRY20-2C (+) cells were 35S labeled for 30 min and chased 30 min in the presence of Na2SO4. (Lanes 3-10) JHRY20-2C cells were pulse labeled for 5 min and chased in unlabeled Na2SO4, cysteine, methionine, and cycloheximide for the times indicated. Labeled cells were collected, converted to spheroplasts, lysed in 1% SDS, and immunoprecipitated with DPAP B antibody. Immunoprecipitates were analyzed on 6% polyacrylamide SDS gels without (-Endo F) or with (+Endo F) earlier treatment with endo E

The deduced amino acid sequence predicts that DPAP B is an integral membrane protein. To test this prediction biochemically, yeast spheroplasts were treated with 100 mM sodium carbonate, pH 11.5, and centrifuged at 100,000 g. Under these conditions, only intrinsic membrane proteins pellet with membranes, whereas soluble and peripheral membrane proteins are recovered in the supernatant (Steck and Yu, 1973; Fujiki et al., 1982) . Fig. 5 shows western blot analysis of the total, pellet, and supernatant fractions using DPAP B-specific and phosphoglycerate kinase-(PGK) specific antibodies. The DPAP B antigen was found in the membrane fraction, whereas the soluble protein PGK did not sediment with membranes. These data confirm that DPAP B is an integral membrane protein.

The soluble vacuolar proteins analyzed to date, such as CPY, are synthesized as glycoproteins (Rothman and Stevens, 1988) . CPY first appears as a glycosylated precursor which receives further glycosyl residues in the Golgi apparatus before being proteolytically cleaved to the mature form upon delivery to the vacuole (Stevens et al., 1982) . To follow the biosynthesis of DPAP B, a culture of DAP2 cells was pulse labeled for 5 min in 35S-H2SO4 and chased for 0, 5, 10, and 30 min (Fig. 4, lanes 3-6) . DPAP B initially appeared as two distinct bands of 110 and 113 kD (lane 3) which chased to the broad 120-kD species seen in lane 2. A small amount of the ll0-kD form remained after a 30-min chase (lanes 2 and 6). Longer times of up to 60 min were required to convert this species to the fully mature 120-kD form (data not shown).

Treatment of the immunoprecipitated samples with endo F (Fig. 4, lanes 7-10) , which removes N-linked glycosyl residues, showed that the only detectable difference in the species seen in lanes 3-6 is in the amount of N-linked carbohydrate present on DPAP B. At each chase time point, labeled DPAP B migrated as a 96-kD species after endo F treatment, which is very close to the molecular mass predicted from the deduced amino acid sequence. A small amount of a 99-kD species was also detected in lanes 8-10;

this appears to correspond to one residual carbohydrate chain remaining on DPAP B since it disappears upon longer incubation with endo F (see Fig. 6 ). Thus, DPAP B is modified in a manner similar to soluble vacuolar proteins, such as CPY, in that its carbohydrate moieties undergo only modest extension in the Golgi apparatus, and do not receive the extensive outer chain mannose residues found on secreted proteins such as invertase (Esmon et al., 1981; Runge, 1988) .

The appearance of two distinct DPAP B species at the earliest times in its biogenesis is similar to that seen with invertase, which is initially synthesized as several discrete species that differ in the number of N-linked glycosyl chains added per protein molecule (Esmon et al., 1981) . To determine how many of the eight potential sites for N-linked glycosylation are modified on DPAP B, limited endo F digestion was per- cells containing the DAP2 gene on a multicopy plasmid were converted to spheroplasts, lysed, and diluted into 100 mM sodium carbonate, pH 11.5. A portion of the total extract (Tot) was saved, and the rest was separated into membrane (Mb) and soluble (Sol) fractions by centrifugation at 100,000 g for 3 h. The three fractions were subjected to SDS-PAGE, and the gel was analyzed by immunoblotting with DPAP B and phosphoglycerate kinase antibodies and subsequent incubation with 12~I-protein A. An autoradiogram of the immunoblots is shown, with the positions of the protein molecular mass standards indicated. The different glycosylated species are marked, with the corresponding number of glycosyl residues shown.

The lournal of Cell Biology, Volume 108, 1989 •368 formed on the precursor, which was immunoprecipitated from a secl8 mutant at 34°C (see below). Fig. 6 suggests that the precursor forms seen in lane 3 of Fig. 4 correspond to DPAP B with six and seven glycosyl residues added, although forms with five and eight oligosaccharide chains can also be seen (Fig. 6 , 0-rain time point). Therefore, DPAP B can be modified at all eight or as few as five sites, but the majority of molecules receives six or seven glycosyl chains.

The PEP4 gene function is necessary for the activation of a number of vacuolar hydrolases (Jones et al., 1982) . Mutations at this locus result in the failure to proteolytically process the large molecular mass, inactive forms of soluble vacuolar proteins such as CPY (Rothman and Stevens, 1988) . Suarez Rendueles et al. (1981) showed that total DPAP (DPAP A and DPAP B) activity was similar in a wild-type strain and a pep4-3 mutant. To examine further the effect of a pep4 mutation on DPAP B, enzyme assays and immunoprecipitations of DPAP B were performed using isogenic PEP4 andpep4-A2 strains. No significant effect on the activity of DPAP B was detected in the pep4-A2 strain as compared to the PEP4 strain (data not shown), confirming that DPAP B is not synthesized as an inactive precursor. This result is consistent with the observation that DPAP B which had accumulated in the ER ofsecl8 cells (see below) was enzymatically active (data not shown). Fig. 7 shows that DPAP B immunoprecipitated from isogenic PEP4 and pep4-A2 strains comigrates before and after digestion with endo E although a minor breakdown product, not seen in the pep4-A2 strain, was detected in the PEP4 strain after endo F treatment. These data argue that the biogenesis of DPAP B is independent of the allelic state of the PEP4 gene.

The temperature-sensitive sec mutations define sequential steps in the secretory pathway in yeast, since cells carrying these mutations are blocked for protein secretion at distinct stages (Novick et al., 1980; Novick et al., 1981) . The localization of CPY to the vacuole is prevented in sec mutants that are blocked at either the ER or Golgi apparatus (secI8 and secT, respectively; Stevens et al., 1982) , and the protein accumulates as an inactive precursor that lacks Golgi-mediated carbohydrate modifications, secl cells, which at 37°C are blocked at a late secretory vesicle stage of the pathway, transport CPY to the vacuole at all temperatures. Thus, secretory and soluble vacuolar proteins diverge after the sec7-defined block and before the secl-blocked stage. To determine if the maturation of DPAP B was affected in these mutants, DPAP Figure 7 . Effect of a pep4 mutation on DPAP B. Strains JHRY20-2C and JHRY20-2CA2 were 35S labeled for 15 min and chased for 30 min at 30°C. Labeled cells were collected, and DPAP B was immunoprecipitated and endo F treated as before.

Effect of sec mutations on DPAP B. Strains were 35S labeled at their nonpermissive temperatures (X2180, HMSF176, and HMSF1, 34°C; SF294-2B, 38°C) for 15 min, chased for 30 min in the presence of Na2SO4 at the same temperature, and DPAP B was immunoprecipitated as before.

B was immunoprecipitated after labeling the cells at the nonpermissive temperature of 34°C (Salminen and Novick, 1987) , except for the sec7mutant, whose restrictive temperature is 38°C (data not shown). Wild-type cells contained the 120-kD mature form of DPAP B at either 34°C (Fig. 8 ) or 38°C (data not shown), secl cells also contained mature DPAP B, whereas secl8 cells accumulated the DPAP B precursor. Thus, the precursor form seen after a short labeling period (Fig. 4, lane 3) is the ER-form of DPAP B. The sec7 strain accumulated a mixture of core-glycosylated and mature forms of DPAP B. This suggests that the block in transport is at an early stage within the Golgi apparatus such that DPAP B is only partially processed by the mannosyl transferases (Runge, 1988) . This phenotype is similar to that seen with invertase, which accumulates as a heterogeneous population of core-and hyper-glycosylated forms in sec7 cells at the restrictive temperature (Esmon et al., 1981) . The accumulation of the DPAP B precursor in secl8 and secTcells was thermoreversible: after labeling cells at the restrictive temperature, shifting the cultures to 23°C, and chasing for 3 h in the absence of new protein synthesis, DPAP B was completely converted to the mature 120-kD form (data not shown). This result, whichwas also observed with CPY accumulated in secl8 and sec7 cells (Stevens et al., 1982) , suggests that the accumulated DPAP B precursor resides early in the pathway and is able to transit the Golgi apparatus upon shifting to the permissive temperature.

To determine whether the accumulation of the DPAP B precursor in sec18 and see7 mutants was correlated with a failure to deliver the protein to the vacuole, DPAP B was localized by indirect immunofluorescence microscopy. For these studies, DPAP B synthesis was placed under the control of the inducible GAL/promoter (Johnston and Davis, 1984) i allowing analysis of nascent DPAP B synthesis after shifting to the nonpcrmissive temperature. A multicopy plasmid conraining the GALI-DAP2 gene fusion was transformed into wild-type and sec mutant strains which had been disrupted at the chromosomal DAP2 locus. Transformed strains were then shifted to the restrictive temperatures in galactosecontaining media and incubated for 2 h. This procedure typically resulted in a 4-10-fold induction of DPAP B activity relative to a strain containing a single genomic copy of DAF2 (data not shown). No DPAP B activity or antigen was detected when these strains were grown in the absence of galactose (data not shown). The cells were fixed in formaldehyde, spheroplasted, and incubated with affinity-purified DPAP B Figure 9 . DPAP B localization by immunofluorescence microscopy. Strains JHRY20-1A (SEC*), CJRY21-3BA1 (secl8), CJRY23-2AA1 (sec7), and CJRY22-6BA I (secl), each containing the GAL1-DAP2 gene fusion on a multicopy plasmid, were shifted to their respective nonpermissive temperatures and treated for induction of the GAL/promoter for 2 h as described in Materials and Methods. The cells were fixed, converted to spheroplasts, and stained with DAPI and afffinity-purified DPAP B antibody. Cells were viewed by Nomarski differential interference optics (A, D, G, and J), and under DAPI (B, E, H, and K) and FITC (C, E I, and L) excitation wavelengths. antibody, followed by FITC-conjugated sheep anti-goat antibody (Fig. 9 ). Nomarski optics revealed the position of the vacuole, and the nucleus was localized by DAPI staining. Wild-type cells grown at either 34°C (Fig. 9 , A-C) or 38°C (data not shown) exhibited fluorescent labeling of DPAP B around the perimeter of the vacuole, consistent with the fact that DPAP B is a vacuolar membrane protein. A similar labeling pattern was seen in secl cells (Fig. 9 , J-L), indicat-ing that the secl block is later in the secretory pathway than the point at which sorting of vacuolar membrane proteins occurs. The staining pattern seen in sec18 (Fig. 9 , D-F) and sect (Fig. 9 , G-I) cells was distinctly nonvacuolar. Interestingly, the sec18 strain showed DPAP B staining around the nucleus, apparently in the nuclear envelope, which probably reflects the fact that the nuclear envelope and the ER are continuous membranes. These data indicate that transport of DPAP B to the vacuole is prevented in sec mutants that are blocked in protein transport early in the secretory pathway.

DPAP B is the first protein of the yeast vacuolar membrane whose structure and biosynthesis has been characterized. Of the vacuolar membrane-associated activities that have been described in yeast, only the proton-translocating ATPase (Uchida et al., 1985; Kane et al., 1989 ) and u-mannosidase (Yoshihisa et al., 1988) have been extensively characterized biochemically. The proton-translocating ATPase is a multimeric enzyme consisting of both integral and peripheral membrane proteins (Kane et al., 1989) . ~Mannosidase, which is frequently used as a marker for the vacuolar membrane, is efficiently removed from membranes by treatment with high-pH sodium carbonate, suggesting that it is peripherally associated with the vacuolar membrane (Yoshihisa et al., 1988) . Thus, DPAP B appears to be a more appropriate marker for intrinsic vacuolar membrane proteins.

In a previous biochemical study, DPAP B was reported to be a 40-kD protein (Garcia Alvarez et al., 1985) . In the present work we have found that DPAP B is a 120-kD glycoprotein, that this species is absent from a strain with a disruption at DAP2, and that the DAP2 sequence encodes a protein predicted to have a molecular mass of 96,429 D, which corresponds closely to that actually observed for deglycosylated DPAP B. The 40-kD species seen previously may have been a contaminating protein or a proteolytic fragment of DPAP B.

As is the case for the soluble vacuolar protein CPY (Stevens et al., 1982) , the delivery of DPAP B to the vacuole is blocked in sec mutants that affect the early stages of the secretory pathway, sec18 (ER-blocked) and sec7 (Golgiblocked) cells accumulate DPAP B in nonvacuolar compartments at the restrictive temperature. However, secl ceils, which are blocked at a late secretory vesicle stage, transport DPAP B to the vacuole at all temperatures, indicating that DPAP B is transported directly from the Golgi apparatus to the vacuole. This rules out models in which vacuolar membrane proteins are first delivered to the plasma membrane, and suggests that, in yeast, soluble and membrane-bound proteins traverse the same compartments of the secretory pathway en route to the vacuole (Stevens et al., 1982) , as is the case for the delivery of proteins to the lysosome in mammalian cells (Griffiths et al., 1988) .

The activity and apparent molecular mass of DPAP B are unaffected by a null mutation at the PEP4 locus. DPAP B is active at all stages of its transport and does not appear to undergo proteolytic cleavage upon delivery to the vacuole. In this respect, DPAP B differs from the soluble vacuolar hydrolases studied thus far, which are synthesized as zymogens and are activated in a PEP4-dependent fashion (Rothman and Stevens, 1988) . Interestingly, the other vacuolar membrane activities that have been examined, i.e., ot-mannosidase (Jones et al., 1982) and the proton-translocating ATPase (Yamashiro, C., and T. Stevens, unpublished results), are also PEP4 independent.

Overproduction of the soluble vacuolar proteins CPY and proteinase A results in their secretion, presumably due to the saturation of a component of the sorting apparatus . This observation provided a means of isolating mutations that disrupt the sorting process, both linked and unlinked to the structural genes of the soluble vacuolar proteins (Vails et al., 1987; Bankaitis et al., 1986; Robinson et al., 1988; Rothman et al., 1989) . 20-fold overproduction of DPAP B does not result in increased cell surface activity of this enzyme (our unpublished results). However, at this level of expression DPAP B does suppress the defect in u-factor maturation of a MATtx stel3 mutant (Sprague, G., and I. Herskowitz, unpublished observation; Julius et al., 1983) . A reasonable interpretation of this result is that DPAP B is mislocalized into the late secretory pathway when overproduced, allowing it to encounter and process the c~-factor precursor to the mature form. Based on this assumption, mutations have been isolated that result in secretion of mature ct-factor in a stel3 background. These mutants do not cause DPAP B to be overproduced, and may define genes that encode components of the sorting apparatus for vacuolar membrane proteins (Pohlig, G., and T. Stevens, unpublished results) .

The sorting signal(s) that directs membrane proteins to the vacuole has yet to be defined. The sorting information could reside in any of three distinct domains; i.e., the cytoplasmic, membrane, and/or lumenal domains. Moreover, each of these domains has been shown to be important for proper intracellular targeting of particular membrane proteins. The lumenal domains of the influenza virus hemagglutinin and vesicular stomatitis virus G protein have been implicated as important for the polarized secretion of these proteins to the apical and basolateral membranes, respectively (Roth et al., 1987; McQueen et al., 1987) . Retention of the coronavirus E1 glycoprotein in the Golgi apparatus requires a specific transmembrane region (Machamer and Rose, 1987) , and an intact cytoplasmic tail is necessary for endocytosis of the LDL and transferrin receptors (Davis et al., 1986; Rothenberger et al., 1987) . Preliminary experiments have shown that the cytoplasmic tail and transmembrane region of DPAP B contain vacuolar sorting information; i.e., fusion of the NH:-terminal 47 residues of DPAP B to the NH2 terminus of the cytoplasmic form of invertase results in the delivery of this hybrid protein to the vacuolar membrane (Roberts, C., and T. Stevens, unpublished observation) . This approach has been previously used to map the sorting domains of CPY (Johnson et al., 1987) Figure 10 . Comparison of DPAP B and DPAP A. Models for the disposition of the two proteins in their respective membranes are shown, with their subcellular locations listed at right. The NH2 and COOH termini of the proteins are indicated, as are the regions of extensive amino acid identity, with the percentage of identical amino acids given.

. Further experiments are aimed at the identification of the specific amino acid residues in this region that are necessary for targeting DPAP B to the vacuole. The DNA sequence of the STE/3 gene, which encodes DPAP A (Julius et al., 1983) , has recently been determined (Flanagan, C., and J. Thorner, personal communication). The deduced amino acid sequence predicts a protein of 931 amino acids with striking similarities to DPAP B (Fig. 10) . Both proteins are predicted to have type II membrane topologies, with short cytoplasmic tails and ,0800 amino acid lumenal domains. The lumenal domains of these proteins are very similar; in particular, the COOH-terminal 240 amino acids are 48 % identical, presumably reflecting the similar enzymatic functions of these proteins in vitro and in vivo (Julius et al., 1983) . However, the cytoplasmic and transmembrane domains are unrelated, possibly reflecting the different subcellular addresses of DPAPA and DPAP B. This sorting signal hypothesis can be tested by exchanging these segments of the two proteins, an approach that has been useful in studying the sorting of membrane proteins in polarized cells (Roth et al., 1987; McQueen et al., 1987) .

Relationship of actin and tubulin distribution to bud growth in wild-type and morphogenetic-mutant Saccharomyces cerevisiae

Structural requirements of a membranespanning domain for protein anchoring and cell surface transport

Isolation of yeast mutants defective in protein targeting to the vacuole

Biosynthesis, glycosylation, movement through the Golgi system, and transport to lysosomes by an N-linked carbohydrate-independent mechanism of three lysosomal integral membrane proteins

Localization of the thermosensitive X-prolyl dipeptidyl aminopeptidase in the vacuolar membrane of Saccharomyces cerevisiae. FEBS (Fed. Eur

NH2-terminal hydrophobic region of influenza virus neuraminidase provides the signal function in translocation

Western blotting": electrophoretic transfer of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A

Analysis of gene control signals by DNA fusion and cloning in Escherichia coll

Identification of two lysosomal membrane glycoproteins

An artificial anchor domain: hydrophobicity sutfices to stop transfer

The J. D. mutation in familial hypercholesterolemia: amino acid substitution in cytoplasmic domain impedes internalization of LDL receptors

Endo-B-N-acetylglucosaminidase F: endoglycosidase from Flavobacterium meningosepticum that cleaves both high-mannose and complex glycoproteins

Invertase B-galactosidase hybrid proteins fail to be transported from the endoplasmic reticulum in Saccharomyces cerevisiae

Compartmentalized assembly of oligosaccharides on exported glycoprotein in yeast

Isolation of intracellular membranes by means of sodium carbonate treatment: application to endoplasmic reticulum

Enzymes required for yeast prohormone processing

Purification and characterization of a thermosensitive X-prolyl dipeptidyl aminopeptidase (dipeptidyl aminopeptidase yscV) from Saccharomyces cerevisiae

Using recombinant DNA techniques to study protein targeting in the eucaryotic cell

The mannose 6-phosphate receptor and the biogenesis of lysosomes

Evidence the yeast STE3 gene encodes a receptor for the peptide pheromone a factor: gene sequence and implications for the structure of the presumed receptor

Transformation of intact yeast cells treated with alkali cations

Solid phase synthesis of polynucleotides VI. Further studies on polystyrene copolymers for the solid support

Distinct sequence determinants direct intracellular sorting and modification of a yeast vacuolar proteinase

Sequences that regulate the divergent GAL1-GALIO promoter in Saccharomyces cerevisiae

PEP4 gene function is required for expression of several vacuolar hydrolases in Saccharomyces cerevisiae

Yeast ot factor is processed from a larger precursor polypeptide: the essential role of a membrane-bound dipeptidyl aminopeptidase

Protein sorting in yeast: the role of the vacuolar proton-translocating ATPase

Structural rearrangements of tubulin and actin during the cell cycle of the yeast Saccharomyces

Cloned viral protein vaccine for foot-and-mouth disease: responses in cattle and swine

Intracellular sorting and processing of a yeast vacuolar hydrolase: proteinase A propeptide contains vacuolar targeting information

A simple method for displaying the hydropathic character of a protein

Cleavage of structural proteins during the assembly of the head of bacteriophage T4

Glycoproteins of the lysosomal membrane

Lysosomal membrane dynamics: structure and interorganelle movement of a major lysosomal membrane protein

A specific transmembrane domain of a coronavirus El glycoprotein is required for its retention in the Golgi region

Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory

Basolateral expression of a chimeric protein in which the transmembrane and cytoplasmic domains of vesicular stomatitis virus G protein have been replaced by those of the influenza virus hemagglutinin

New M 13 vectors for cloning

Expression and ~SN labeling of proteins for proton and nitrogen-15 NMR

Order of events in the yeast secretory pathway

Identification of 23 complemenration groups required for post-translational events in the yeast secretory pathway

Export of major cell surface proteins is blocked in yeast secretory mutants

Active transport of basic amino acids driven by a proton motive force in vacuolar membrane vesicles of Saccharomyces cerevisiae

Calcium transport driven by a proton motive force in vacuolar membrane vesicles of Saccharomyces cerevisiae

ct-D-mannosidase ofSaccharomyces cerevisiae: characterization and modulation of activity

Protein sorting in yeast: isolation of mutants defective in the delivery and processing of multiple vacuolar hydrolases

The large external domain is suIfieient for the correct sorting of secreted or chimeric influenza virus hemagglutinins in polarized monkey kidney cells

Endocytosis of the transferrin receptor requires the cytoplasmic domain but not its phosphorylation site

Protein sorting in yeast

Protein sorting in yeast: mutants defective in vacuole biogenesis mislocalize vacuolar proteins into the late secretory pathway

Protein sorting and biogenesis of the lysosome-like vacuole in yeast

Overproduction-induced mislocalization of a yeast vacuolar protein allows isolation of its structural gene

Characterization of genes required for protein sorting, vacuolar function, and endocytosis in the yeast Saccharomyces cerevisiae

One step gene disruption in yeast

Post-translational modification during protein secretion

A ras-like protein is required for a post-Gotgi event in yeast secretion

DNA sequencing with chain terminating inhibitors

On the transfer of integral membrane proteins into membranes

Selective solubilization of proteins from red blood cell membranes by protein perturbants

Early stages in the yeast secretory pathway are required for transport of carboxypeptidase Y to the vacuole

Gene dosage-dependent secretion of yeast vacuolar carboxypeptidase Y

A new X-prolyl-dipeptidyl aminopeptidase from yeast associated with a particulate fraction. FEBS (Fed. Eur

Identification of the structural gene for dipeptidyl aminopeptidase yscV (DAP2) of Saccharomyees cerevisiae

Purification and properties of H÷-translocating, Mg÷-adenosine triphosphatase from vacuolar membranes of Saccharomyces cerevisiae

Production of antisera with small doses of immunogen: multiple intradermal injections

Protein sorting in yeast: the localization determinant of yeast vacuolar carboxypeptidase Y resides in the propeptide

Lysosomal enzymes and their receptors

Human lysosomal acid phosphatase is transported as a transmembrane protein to lysosomes in transfected baby hamster kidney cells

Subcellular localization of a protein kinase required for cell cycle initiation in Saccharomyces cerevisiae: evidence for an association between the CDC28 gene product and the insoluble cytoplasmic matrix

Improved M13 phage cloning vectors and host strains: nucleotide sequences of the MI3 mpl9 and put19 vectors

Solubilization and purification of ot-mannosidase, a marker enzyme of vacuolar membranes in Saccharomyces cerevisiae

The transmembrahe segment of the human transferrin receptor functions as a signal peptide

We would like to thank George Sprague for the gift of the plasmid containing the DAP2 clone and for many useful discussions; Jeremy Thorner for communicating results before publication; Rick Dahlquist for providing the plasmid prise5; Ruskin Gould for excellent technical assistance; Harry Howard for expert assistance with the fluorescence microscopy; and Christopher Raymond, Nicholas Davis, Karen Moore, and Patricia Kane for comments on the manuscript.

Received for publication 20 December 1988.