Compartment Acidification Is Required for Efficient Sorting of Proteins ...

THEJOURNAL

OF

BIOLOGICAL CHEMISTRY

VOl. 267, No. 5, Issue of February 15,pp. 3416-3422.1992 Printed in U.S.A.

0 1992 by The American Society for Biochemistry and Molecular Biology, Inc

Compartment Acidification Is Required for Efficient Sorting of Proteins to the Vacuolein Saccharomyces cereuisiae” (Received for publication, June 28,1991)

Daniel J. KlionskySO, Hannah Nelsonll, and Nathan Nelsonll From the $.Department of Microbiology, University of California, Davis, California 95616 and the llRoche Institute of Molecular Biology, Roche Research Center, Nutley, New Jersey07110

The vacuole of the yeast Saccharomyces cerevisiae man et al., 1989) and may serve as a trigger for the autoacticontains a proton-translocating ATPase that acidifies vation of proteinase A (Woolford et al., 1986), a model based the vacuolar lumen and generates pH agradient across in part on the pH-dependent autocatalysis of the homologous the vacuole membrane. We have investigated the role mammalian enzymes pepsinogen (James and Sielecki, 1968) of compartmentacidification of the vacuolar system in and lysosomal cathepsin D (Hasilik et al., 1982). Similarly, the sorting of vacuolar proteins. Strains with chro- compartment acidification has been proposed to play a role mosomal disruptions of the genes encoding theA, B, or in protein sorting analogous to the acidification-dependent c subunit of the vacuolar ATPase are unable to acidify sorting and receptor recycling of certain lysosomal proteins their vacuoles. These vat mutant strains accumulate (Mellman et al., 1986). However, recent determinations of the and secrete precursor forms of the soluble vacuolar hydrolases carboxypeptidaseY and proteinase A. The vacuolar pH (Prestonet al., 1989; Yamashiro et al., 1990) and kinetics of secretionsuggests that missorting occurs inanalyses of protein processing in mutants that are impaired the Golgi complex orin post-Golgi vesicles. The pres- in their ability to acidify their vacuoles (Yamashiro et al., ence of mature forms of the vacuolar proteins within 1990) suggest that pH may play a more limited role in zymothe cell indicates that vat mutations do not cause de- gen activation. In addition, in vitro studies indicate that fects in zymogen processing. Precursor forms of the aspects of the tonoplasm other than the pHmay be required membrane-associated vacuolar hydrolase alkaline to initiate proteolytic processing (Mechler et al., 1987). There are also some discrepancies concerning the imporphosphatase are also accumulated in vat mutant cells but to a lesserextent, suggesting that sortingof vacu- tance of compartment acidification inprotein sorting. It olar membrane proteins is less sensitive to changes in should be noted that the Golgi complex (and perhaps other the lumenalpH. A similar typeof missorting defect cancompartments involved in vacuolar protein transport)is likely be induced in wild-type cells at pH 7.5. These results to contain a vacuolar type ATPase (Moriyama and Nelson, indicate that acidification of the vacuolar system is 1989b). When we refer to compartment acidification by vacimportant for efficient sorting of proteins to the vac- uolar type ATPases, we mean acidification of this vacuolar uole. system, and not just the vacuole itself. This is of particular importance considering the proposed role of the trans-Golgi network in the sorting of secretory pathway proteins (Griffiths and Simons, 1986). Neutralization of the pH of the The yeast vacuole is integrally involved in awide variety of vacuolar system by a protonophore such as CCCP’ or by cellular functions (Klionsky et al., 1990).The vacuole is partly acidotropic agents like the weak base ammonium acetate analogous tothe mammalian lysosome; bothare low pH results in the missorting of vacuolar proteins (Banta et al., compartments that contain numerous hydrolytic enzymes 1988).These chemicals act in a nonspecific manner to disrupt (Kornfeld and Mellman, 1989). The reduced vacuolar pH is membrane potentials and/or eliminate pH gradients in all maintained by a proton-translocating ATPase. The vacuolar intracellularcompartments,Inhibition of vacuolar type ATPase is a multisubunit complexcomposed of approxi- ATPases with the drug bafilomycin AI (Bowman et al., 1988) mately eight different polypeptides (Kane et al., 1989; Mori- provides a more specific means of raising the vacuolar pH. yama and Nelson, 1989a; Uchida et al., 1985). Genes encoding Treatment with bafilomycin Al causes an increase in the three of these subunits (A, B, and c) have been cloned, and vacuolar pH and results in themissorting of soluble vacuolar their nucleotide sequences have been determined (Hirata et hydrolases, but apparently not those of the vacuole membrane al., 1990; Nelson and Nelson, 1989; Nelson et al., 1989; Shih (Banta et al., 1988; Klionsky and Emr, 1989). The role of et al., 1988; Yamashiro et al., 1990). The pH gradient gener- vacuole acidification has also been analyzed genetically. Cerated by the vacuolar ATPase is important in a number of tain mutants that were isolated on the basis of missorting processes including amino acid and ion uptake (Klionsky et phenotypes (ups) or decreased peptidase activity (pep) are al., 1990). It has also been proposed that the low pH of the compromised in their ability to acidify the vacuole (Banta et vacuole is important for processing precursor proteins (Roth- al., 1988, Preston et al., 1989; Rothman et al., 1989).Recently, specific mutants have been generated that are defective in * This work was supported by the University of California Cancer Research Coordinating Committee and United States Public Health vacuole acidification (uph) or that contain disruptions in Service Grant DK43684 from the National Institutes of Health (to genes encoding subunits of the vacuolar ATPase (Nelson and D. J. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 5 To whom correspondence should be addressed. Tel.: 916-7520277;Fax: 916-752-9014.

’ The abbreviations used are: CCCP, carbonyl cyanide m-chlorophenylhydrazone; PCR, polymerase chain reaction; YNB, yeast nitrogen base; MES, 2-(N-morpholino)ethanesulfonicacid; MOPS, 3(N-morpho1ino)propanesulfonicacid; SDS, sodium dodecyl sulfate; ER, endoplasmic reticulum.

3416

Compartment Acidification and Nelson, 1990; Preston et al., 1989; Yamashiro et al., 1990). Studies of a uat2 (uatE) mutant strain led to the conclusion that acidification of the vacuolar system is not important for the sorting of vacuolar membrane proteins and is only required for highly efficient targeting of the soluble vacuolar hydrolases (Yamashiro et al., 1990). A preliminary analysisof protein targeting in stains with mutations in the VATA or VATc genes suggested a more substantial defect in soluble vacuolar protein sorting (Nelson and Nelson, 1990). To address these differences and to develop a more complete understanding of the role of compartment acidification in protein sorting, we have carried out a detailed analysis of the effect of disruptions of VAT genes on the processing and localization of both soluble and membrane-associated vacuolar proteins. We have found that mutants that are defective in compartment acidification are able to process precursor vacuolar proteins, but exhibit significant missorting defects. A missorting defect can also be induced in wild-type cells by growing them under conditions that mimic the effects of the uat mutations. EXPERIMENTAL PROCEDURES

Sorting Protein

3417

MOPS. To prepare media at higher pH values, NaOH was added to adjust the pH. Media at pH 3.5 and 4.5 were buffered with 100 mM MES, and the pH was adjusted with NaOH. Modified Wickerham's minimal medium was prepared as described previously (Johnson et al., 1987). Materials-Lyticase was obtained from Enzogenetics (Corvallis, OR); Tran3%-label was from I C N a,-macroglobulin was from Boehringer Mannheim; Autofluor was from National Diagnostics; and all other chemicals were from Sigma. Antisera to proteinase A, carboxypeptidase Y, alkaline phosphate, and invertase were prepared as described previously (Klionsky et al., 1988; Klionsky and Emr, 1989). Antisera to the F, ,&subunit and glyceraldehyde-3-phosphate dehydrogenase were generously supplied by Dr. Michael Douglas (University of North Carolina, Chapel Hill, NC) and Dr. Michael Yaffe (University of California, San Diego). Bafilomycin A, was generously provided by Dr. Karlheinz Altendorf (Universitat Osnabruck, Osnabruck, Federal Republic of Germany). Labeling and Immunoprecipitation-The procedures used for the preparation, labeling, and fractionation of yeast cells and spheroplasts were modifications of procedures described previously (Klionsky et al., 1988; Robinson et al., 1988). To label spheroplasts, cells were grown in YNB, 2% glucose buffered at pH 5.5 to mid-log phase. Approximately 4 units of cells a t Amnm were collected by centrifugation and suspended in 0.1 M Tris sulfate (pH 9.4), 10 mM dithiothreitol and incubated at 30 "C for 5 min. Cellswere centrifuged again and suspended in Wickerham's minimal medium that was adjusted to pH 6.5 and that contained 1.3 M sorbitol. Lyticase was added to a finalactivity of 30 units/A unit at 600 nm, and the cultures were incubated at 30 "C for 20 min to remove the cell wall. Spheroplasts were pelleted and resuspended in YNB, 2% glucose that was buffered at pH 5.5 and that contained 1.3 M sorbitol. Spheroplasts were labeled with Tran3%label (0.2 pCi/ml) and chased with a mixture of methionine (8 mM) and cysteine (4 mM). The labeled culture was separated into intracellular (spheroplast) and extracellular (periplasm plus medium) fractions following centrifugation a t 3000 rpm for 2 min. The separated fractions were trichloroacetic acid-precipitated and suspended in MES/urea resuspension buffer (50 mM sodium phosphate, 25 mM MES (pH 7.0), 1%SDS, 3M urea, 0.5% 0-mercaptoethanol, 1mM sodium azide). Double immunoprecipitations were carried out as described previously (Bedwell et al., 1987; Klionsky et al., 1988). Radiolabeled immunoprecipitates were suspended in SDS sample buffer, boiled for 4 min, and electrophoresed on SDS-polyacrylamide gels. After electrophoresis, gels were fixed, and treated with Autofluor. Densitometry was performed on a Shimadzu CS9000 dual-wavelength flying-spot scanner. To label intact cells, cultures were grown and labeled in YNB, 2% glucose buffered at pH5.5. After labeling and chase, the entire culture was precipitated with trichloroacetic acid or separated intocell pellet and media supernatant fractions by centrifugation at 7000 rpm for 5 min prior to trichloroacetic acid precipitation. Following trichloroacetic acid precipitation, pellets were suspended in MES/urea resuspension buffer and immunoprecipitated. To examine the effect of altering the pH on the sorting of vacuolar proteins, yeast cells were grown in YNB (pH 5.5). Cells were collected by centrifugation, washed, and resuspended in YNB buffered at the appropriate pH. Cells were allowedto continue growing at 30 "C for designated lengths of time after the pHshift prior to theaddition of radioactive label.

Strains and Plasmids-The Saccharomycescerevisiae yeast strains used were W303-1B (MATa ku2 hk3 d e 2 trpl urd') and W303-1B with LEU2 chromosomal disruptions in the VATA, VATB, or VATc gene (uatA-Al,uatB-Al, or uatc-Al) (Nelson and Nelson, 1990; Noumi et al., 1991). The VATA gene encoding subunit A of the yeast vacuolar ATPase was cloned by amplification from genomic DNA of yeast strain W303 by PCR (Fig. 1). The information for the construction of the primers came from the published sequence of the TFPl gene (Shih et al., 1988). The TFPl gene encodes both the 69-kDa subunit of the vacuolar ATPase and a 454-amino acid protein of 50 kDa(Kane et al., 1990). The gene encoding the 50-kDa protein interrupts the sequence that codes for the vacuolar ATPase subunit A. A two-step PCR approach was used to clone the VATA gene, encoding the 69-kDa subunit A, without encoding the 50-kDa protein. First, PCR was performed with yeast DNA, a primer (200 pmol) containing a BamHI site and corresponding to the sense strand at positions 736-760 of the TFPl gene, and an antisense primer (200 pmol) corresponding to positions 1777-1797. This generated a fragment encoding the amino-terminal portionof the VATA gene (encompassing amino acids 1-284), whichis located 5' to the50-kDa protein coding sequence. Second, a separate PCR reaction was performed with yeast DNA, a 41-nucleotide primer (200 pmol) corresponding to the sense strand a t positions 1778-1797 3160-3180, and an antisense primer (200 pmol) corresponding to positions 4430-4452 and containing an NdeI site. This produced a fragment encoding the carboxyl-terminal half of the VATA gene (coding for amino acids 285-617), located 3' to the gene for the 50-kDa protein. The resulting two DNA fragments of 1.1 and 1.3 kilobases contain sequence information only for the VATA gene and do not contain coding information for the 50-kDa protein. These fragments were diluted 1,000-fold, and 1r l of each was used for amplification with the two primers (200 pmol) containing the BamHI and NdeI restriction sites. The resulting 2.4 kilobase DNA fragment was cloned into theBamHI-NdeI sites of RESULTS the plasmid YPNl (Noumi et al., 1991). This construct ofVATA in YPNl was used for expressing subunit A of the vacuolar ATPase in Vacuolar Protein Sorting in uat Mutants-Most vacuolar yeast. Disruption of VATA in the yeast chromosome was performed by cutting the resulting plasmid with EcoRI to remove two DNA proteins transit through a portion of the secretory pathway. fragments of0.4 and 1.1 kilobases. The EcoRI sites were filled in All of the characterized soluble vacuolar hydrolases such as using the Klenow fragment of DNA polymerase; and theyeast LEU2 proteinase A and carboxypeptidase Y undergo various types gene, excised with HpaI, was cloned into theblunt-ended sites. A 4.3- of glycosyl and proteolytic processing (Klionsky e t al.,1990). kilobase DNA fragment containing the LEU2 gene and the flanking These proteins receive core glycosylation in the endoplasmic regions of VATA was excised by BamHI and NdeI restriction enzymes and was used for disrupting VATA, in addition to the gene encoding reticulum (ER), generating the p l form, followed by subsethe 50-kDa protein, as previously described (Noumi et al., 1991). The quent carbohydrate trimming and elongation in the Golgi absence of the 50-kDa protein does not affect the vacuolar ATPase complex that result in the higher molecular mass p2 form. since deletion of only that portion of the TFPl gene that encodes the The proteolytic modifications typically include the removal 50-kDa protein does not affect the function of subunit A (Kane et al., of a propeptide segment upon or just before delivery of the 1990). The disruption was verified by Southern blots, and theresulting mutant had a phenotype identical to the other disruptants in precursor proteinsto thevacuole. These modifications provide a convenient way to assess the location of vacuolar proteins genes encoding the vacuolar ATPase (Nelson and Nelson, 1990). Media-Cells were grown in yeast nitrogen base (YNB) containing within the secretory pathway and to determine the kinetics 2% glucose and buffered at pH 5.5 with 50 mM MES and 50 mM of vacuolar delivery. In addition, a vacuolar membrane pro-

+

3418

Compartment Acidification

tein, alkaline phosphatase, undergoes similar types of processing events. To determine the effects of disruptions of genes encoding different vacuolar ATPase subunits, we examined the processing and localization of several vacuolar hydrolases. The wild-type yeast strain W303-1B or the isogenic strains carrying chromosomal disruptionsin the VATA (Fig. l), VATE, or VATc gene (Nelson and Nelson, 1990) were grown in medium buffered at pH 5.5. The cells were converted to spheroplastsprior to labeling and chase and were subsequently separated into intracellular (spheroplast) and extracellular (periplasm plus medium) fractions (Fig. 2). All three of the vat mutant strainsaccumulated high levels of precursor E

E

EHH

H

I kb

FIG. 1. Cloning and disruption of VATA gene encoding 69kDa subunit A of yeast vacuolar ATPase. The VATA gene was cloned by PCR as described under “Experimental Procedures.” Primers used for amplification of the TFPl gene were designed to amplify only sequences corresponding to VATA, and not the50-kDa protein also encoded by TFPl (Kane etal.,1990).A plasmid in which the LEU2 gene replaced the indicated portion of VATA was used to generate the vatA null mutant as described under “Experimental Procedures.” Stipled boxes represent the coding regions for VATA; the shingled box indicates the location of the coding region for the 50-kDa protein; and the hatched box shows the inserted LEU2 gene. B, BamHI; E, EcoRI; H, HindIII; N , NdeI. kb, kilobase. vatB vatA VAT Fraction:

I

E

1,

I

E

I

I

E

I

vatc

I

E

proALPmALP‘ % Precursor: p2 CPY\

plcPYmCPY % Recursor:

4

67

56

65

7

71

60

65

proPrAm h A -I % Recursor:

FIG. 2. Alkaline phosphatase, carboxypeptidase Y, and proteinase A are accumulated as precursor forms in vat mutant strains. Strain W303-1B (VAT) and theisogenic strains with chromosomal disruptions in the VATA, VATB, or VATc gene were grown in YNB buffered at pH 5.5 and converted to spheroplasts as described under “Experimental Procedures.” After labeling a t 30 “C for 30 min in the presence of Tran3’S-label, chase was initiated by the addition of methionine/cysteine ( 8 4 mM final concentration) and allowed to continue for 1 h. After the chase, the intracellular ( I ; spheroplast pellet) and extracelluar ( E ; supernatant) fractions were separated and precipitated with trichloroacetic acid. Double immunoprecipitations were performed with antisera to alkaline phosphatase (ALP), carboxypeptidase Y ( C P Y ) , or proteinase A (PrA). Densitometric analysis wasused to determine the approximate percent of total precursor forms present in both the intracellular and extracellular fractions, and thecorresponding values are shown below each pair of lanes. The values presented are averages from four independent experiments. The positions of the precursor and mature ( m )forms of the vacuolar hydrolases are indicated.

and Protein Sorting forms of carboxypeptidase Y and proteinase A(Fig. 2) as well as anadditional soluble vacuolar hydrolase, proteinase B(data not shown). This is in contrast to a previous analysis of the effect of a vat2 (vatE) disruption on vacuolar protein sorting (Yamashiro et al., 1990). Substantialamounts of the soluble vacuolar hydrolases were also found in the extracellular fraction from all three mutants (Table I). In general, less proteinase A was found in the extracellular fraction than carboxypeptidase Y. This is similar to previous results with vacuolar protein sorting ups mutant strains that indicated that carboxypeptidase Y is secreted to a greater extent thanis proteinase Aor B (Robinson et al., 1988). The extracellular fraction contained very little ( 6 % ) of the mature forms of any of the vacuolar proteins. This indicates that theprecursor protein (p2 form) found inthe extracellular fraction hadbeen secreted from the cell and was not the result of cell lysis. We also monitored the location of a cytosolic marker enzyme, glyceraldehyde-3phosphate dehydrogenase. Only minor amounts of this enzyme were found in the extracellular fraction (Table I), indicating that the level of lysis was insufficient to account for the amountof precursor vacuolar protein in this fraction. All three of the vat mutations also resulted in the accumulation of the precursor form of the vacuolar membrane protein alkaline phosphatase (Fig. 2). Even though precursor alkaline phosphatase was accumulated in vat mutant strains, thelevel of precursor was substantially less than was seen for carboxypeptidase Y and proteinaseA (Fig. 2). It was previously shown that alkaline phosphatase is delivered to the vacuole and processed with normal kinetics in the presence of the drug bafilomycin A, (Klionsky and Emr, 1989). Bafilomycin is an inhibitor of vacuolar type ATPases and results in precursor accumulation and missorting of the soluble vacuolar hydrolases (Banta et al., 1988; Bowman et al., 1988). In a wildtype strain,bafilomycin caused essentially the same degree of missorting of carboxypeptidase Y and proteinase A as was caused by the vat mutations. A similar differential effect on the sortingof membrane versus soluble proteins was seen with CCCP. CCCP is a protonophore that disrupts the membrane potential and abolishes the pH gradient in all intracellular compartments. Treatment with CCCP resulted in a greater accumulation of precursor forms of proteinase Aand carboxypeptidase Y than did the Avatc mutation (data notshown). CCCP had much less of an effect on alkaline phosphatase, however, in agreement with the relative insensitivity of this protein to changes in pH and/ormembrane potential. TABLEI Percent of vacuolar and cytosolic proteins in the extracellular fraction of spheroplasts Yeast spheroplasts were labeled for 30 min, followed by a 60-min chase. After separation into intracellular and extracellular fractions, the samples were immunoprecipitated as described under “Experimental Procedures.” The amount of the indicated vacuolar precursor proteins present in the extracellular fraction is expressed as the percentage of the total (precursor plus mature) protein found in both the intracellular and extracellular fractions. Values for carboxypeptidase Y (CPY) and proteinase A (PrA) represent averages from three independent experiments. The glyceraldehyde-3-phosphate dehydrogenase (G3PDH) values are based on an immunoprecipitation of the samples shown in Fig. 2. Percentage of protein in extracellular fraction Enzyme VAT

ProCPY ProPrA