Synthetic Genetic Interactions With Temperature-Sensitive Clathrin in Saccharomyces cerevisiae: Roles for Synaptojanin-Like Inp53p and Dynamin-Related Vps1p in Clathrin-Dependent Protein Sorting at the trans-Golgi Network

Corresponding author: Gregory S. Payne, UCLA School of Medicine, Department of Biological Chemistry, 33-247 CHS, P.O. Box 951737, Los Angeles, CA 90095-1737., gpayne{at}mednet.ucla.edu (E-mail)

Communicating editor: D. BOTSTEIN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Clathrin is involved in selective protein transport at the Golgi apparatus and the plasma membrane. To further understand the molecular mechanisms underlying clathrin-mediated protein transport pathways, we initiated a genetic screen for mutations that display synthetic growth defects when combined with a temperature-sensitive allele of the clathrin heavy chain gene (chc1-521) in Saccharomyces cerevisiae. Mutations, when present in cells with wild-type clathrin, were analyzed for effects on mating pheromone -factor precursor maturation and sorting of the vacuolar protein carboxypeptidase Y as measures of protein sorting at the yeast trans-Golgi network (TGN) compartment. By these criteria, two classes of mutants were obtained, those with and those without defects in protein sorting at the TGN. One mutant with unaltered protein sorting at the TGN contains a mutation in PTC1, a type 2c serine/threonine phosphatase with widespread influences. The collection of mutants displaying TGN sorting defects includes members with mutations in previously identified vacuolar protein sorting genes (VPS), including the dynamin family member VPS1. Striking genetic interactions were observed by combining temperature-sensitive alleles of CHC1 and VPS1, supporting the model that Vps1p is involved in clathrin-mediated vesicle formation at the TGN. Also in the spectrum of mutants with TGN sorting defects are isolates with mutations in the following: RIC1, encoding a product originally proposed to participate in ribosome biogenesis; LUV1, encoding a product potentially involved in vacuole and microtubule organization; and INP53, encoding a synaptojanin-like inositol polyphosphate 5-phosphatase. Disruption of INP53, but not the related INP51 and INP52 genes, resulted in -factor maturation defects and exacerbated -factor maturation defects when combined with chc1-521. Our findings implicate a wide variety of proteins in clathrin-dependent processes and provide evidence for the selective involvement of Inp53p in clathrin-mediated protein sorting at the TGN.

EUKARYOTIC cells contain multiple membrane-bounded compartments, each with a distinct composition and cellular function. To maintain the integrity of each organelle, the cell selectively transports proteins to the appropriate resident compartment. Examples of such selective transport occur in the secretory pathway. Proteins that reside in secretory pathway organelles can be actively retained and/or retrieved from subsequent compartments, while secreted and plasma membrane proteins traverse the pathway (ROTHMAN and WIELAND 1996 ; KAISER et al. 1997 ). Further sorting occurs at organelles such as the trans-Golgi network (TGN) where pathways branch to multiple destinations (TRAUB and KORNFELD 1997 ). The complex molecular mechanisms necessary for different steps in selective protein transport remain incompletely defined.

Coat proteins constitute a class of molecules associated with the cytoplasmic face of organelles and transport vesicles. Coat protein complexes are believed to drive membrane deformation during vesicle formation and impart specificity to protein transport reactions by selecting appropriate proteins for incorporation as cargo into nascent transport vesicles (SCHEKMAN and ORCI 1996 ; ROBINSON 1997 ). Distinct coat complexes have been defined, including COPI, COPII, and clathrin coats. COPI and COPII coats have well-characterized roles in selective transport at the early stages of the secretory pathway and probably participate in other secretory and endocytic steps (BARLOWE 1998 ; LOWE and KREIS 1998 ). In contrast, clathrin coats are involved in receptor-mediated endocytosis, localization of resident membrane proteins to the TGN, and transport of proteins to the lysosome/vacuole (WILSBACH and PAYNE 1993A ; SCHMID 1997 ; MOLLOY et al. 1999 ).

Clathrin and clathrin adaptors (APs) are the major structural components of clathrin coats (SCHMID 1997 ). Clathrin is a three-legged molecule, termed a triskelion, made up of heavy (Chc) and light chains (Clc). Triskelions assemble together to form the outer polyhedral shell of the coat. Clathin APs are heterotetramic complexes that bridge the clathrin shell to the membrane. APs act in recruitment of clathrin to appropriate membranes and in cargo selection during vesicle formation. At least two distinct AP complexes interact with clathrin, AP-1 at the TGN and AP-2 at the plasma membrane (HIRST and ROBINSON 1998 ). Whether the newly described AP-3 complex participates in clathrin-coated vesicle formation is currently unclear (SIMPSON et al. 1996 ; DELL'ANGELICA et al. 1998 ). In addition to APs, a growing list of accessory proteins has been identified that associate with clathrin coats and/or function in clathrin-coated vesicle formation (PISHVAEE and PAYNE 1998 ). Recent studies in yeast demonstrate clathrin function in the absence of AP complexes (HUANG et al. 1999 ; YEUNG et al. 1999 ). These findings further implicate the participation of additional factors in clathrin-coated vesicle formation.

The yeast Saccharomyces cerevisiae contains a single clathrin heavy chain gene (CHC1) and a single clathrin light chain gene (CLC1; PAYNE and SCHEKMAN 1985 ; SILVEIRA et al. 1990 ; LEMMON et al. 1991 ). Strains lacking clathrin heavy chain (chc1) grow slowly or, in certain genetic backgrounds, are inviable (PAYNE and SCHEKMAN 1985 ; LEMMON and JONES 1987 ). In addition to slow growth, phenotypes of viable chc1 strains include mislocalization of TGN membrane proteins and retarded receptor-mediated endocytosis (PAYNE et al. 1988 ; PAYNE and SCHEKMAN 1989 ). To characterize the immediate effects of inactivating clathrin heavy chain, a recessive, temperature-sensitive allele of CHC1, chc1-521, was generated previously. At the nonpermissive temperature, chc1-521 cells rapidly exhibit protein trafficking defects commensurate with those in chc1 cells, and over time also grow slowly (SEEGER and PAYNE 1992A ; TAN et al. 1993 ). These results support roles for clathrin in Golgi membrane protein localization and receptor-mediated endocytosis. In contrast to chc1 cells, which sort newly synthesized soluble vacuolar proteins from the TGN at near wild-type levels, chc1-521 cells display an immediate sorting defect leading to secretion of vacuolar protein precursors (SEEGER and PAYNE 1992B ). However, by an undefined mechanism, vacuolar protein sorting gradually recovers in chc1-521 cells incubated at the nonpermissive temperature, reaching efficiencies observed in chc1 and wild-type cells. Taken together, the combined phenotypes of chc1 mutants argue that clathrin is required for protein sorting at the TGN and receptor-mediated endocytosis at the plasma membrane.

Screens for mutations that enhance specific phenotypes in a sensitized mutant background can be an effective strategy to identify genes whose products act in the same pathway (GUARENTE 1993 ). Previously, this type of approach was carried out in a screen for mutations that confer lethality in a chc1 strain, yielding lesions in numerous loci (MUNN et al. 1991 ). However, the specificity of the mutations was uncertain because of the possibility that subtle defects in pathways unrelated to clathrin could cause inviability when present in already debilitated chc1 cells. To circumvent this issue, we have taken advantage of the properties of cells expressing the temperature-sensitive chc1-521 allele to characterize clathrin-dependent pathways. In these cells, the severity of growth and protein transport defects depends on the incubation temperature (SEEGER and PAYNE 1992A ). At the permissive temperature of 24°, chc1-521 cells grow at wild-type rates and display little or no abnormality in clathrin-mediated protein traffic. At 30°, mutant cells continue to grow at near wild-type rates, but compromised clathrin function is evidenced by the partial mislocalization of TGN membrane proteins. Incubation at 37° is required to produce growth and trafficking defects approximating those of chc1 cells. This graded response to temperature suggested that the chc1-521 allele would sensitize cells to mutations in other components of clathrin-dependent pathways. In support of this idea, disruptions of genes encoding subunits of the AP-1 complex, which have no observed detrimental effects alone, accentuate growth and TGN membrane protein localization defects in chc1-521 cells (PHAN et al. 1994 ; RAD et al. 1995 ; STEPP et al. 1995 ; YEUNG et al. 1999 ). Consistent with a specific, if subtle, role for AP-1 at the Golgi apparatus, AP-1 subunit gene disruptions do not influence endocytosis in chc1-521 cells.

Here we describe isolation and characterization of mutations that cause severe growth defects when combined with chc1-521 at the semipermissive temperature of 30°. This type of genetic interaction in yeast, where two mutations in combination result in significantly greater defects than either alone, is referred to as synthetic (GUARENTE 1993 ). Accordingly we have termed the mutations tcs (temperature-sensitive clathrin synthetic mutation). The collection of tcs mutations defines a set of gene products that may participate in clathrin-dependent transport pathways. Identification of tcs mutations in the genes encoding the dynamin-related GTPase Vps1p and the synaptojanin-like inositol-5-phosphatase family member Inp53p prompted a more ex-tensive examination of their genetic interactions with clathrin. The results offer functional evidence supporting roles for Vps1p and Inp53p in clathrin-mediated protein trafficking at the TGN.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plasmids and nucleic acid techniques:
Plasmid constructions were performed using standard molecular biology techniques (SAMBROOK et al. 1989 ). pULE-CHC1, carrying the LYS2, URA3, and CHC1 genes, was constructed as follows. An XbaI-HindIII fragment carrying the LYS2 gene was inserted into pRS316 (SIKORSKI and HIETER 1989 ) to create pRS316-LYS2A. A ClaI-SalI fragment with the 5' end of CHC1 and a ClaI fragment with the 3' end of CHC1 from pCHCc102 (PAYNE et al. 1987 ) were sequentially inserted into pRS316-LYS2A to produce pULE-CHC1. A 6.9-kb SmaI-SalI fragment containing CHC1 was removed from pCHC-XS (MUNN et al. 1991 ) and ligated into pRS423 (CHRISTIANSON et al. 1992 ) to create p423-CHC1. YIpCHC521Cla contains the ClaI-SalI fragment of the chc1-521 allele (SEEGER and PAYNE 1992B ) in the integrating plasmid YIp5 (URA3). To generate p6-2(15a)BP, a 4.5-kb PvuII-BamHI fragment containing the RIC1 open reading frame was isolated from a genomic library clone (the BamHI fragment was generated by ligation of the Sau3A-cleaved genomic fragment to the BamHI-cleaved vector) and subcloned into the BamHI and SacI sites in pRS315 (SIKORSKI and HIETER 1989 ) after converting the SacI end to a blunt form with T4 DNA polymerase (blunt-end). p426-RIC1 resulted by ligation of a BamHI-XbaI fragment and an XbaI-HindIII fragment from p6-2(15a)BP into pRS426 (CHRISTIANSON et al. 1992 ) cleaved with BamHI and HindIII. A SalI-NotI fragment containing RIC1 was subcloned from p426-RIC1 into pRS315 to form p315-RIC1. pric11 is pBluescript II KS⁺ (Stratagene, La Jolla, CA) containing a 4.0-kb HindIII-BamHI RIC1 fragment from p6-2(15a)BP where a 2.2-kb EcoRI-SpeI fragment containing most of the RIC1 coding sequence is replaced by the TRP1 gene. A 3.6-kb BamHI-SalI genomic fragment containing YDR027c/LUV1 was subcloned into pRS315 to produce p45-5(SB). ptcs3-1 is pBluescript II KS⁺ containing the 3.6-kb YDR027c/LUV1 fragment where a 2.5-kb BglII-ClaI fragment containing most of the LUV1 coding sequence is replaced by the TRP1 gene. pPTC1-2 was constructed by subcloning a 1.26-kb EcoRI-BamHI genomic fragment containing PTC1 into pRS314 (SIKORSKI and HIETER 1989 ). pptc1-1 is pBluescript II KS⁺ containing the 1.26-kb PTC1 fragment where a 0.81-kb NdeI-BamHI fragment containing most of the PTC1 gene was replaced with TRP1. p313-INP53 contains a 4.6-kb HincII-XbaI genomic fragment which includes the INP53 open reading frame subcloned into pRS313 (SIKORSKI and HIETER 1989 ) cleaved with EcoRV and XbaI. The 4.6-kb fragment from p313-INP53 was liberated by cleavage with XhoI and SacI and ligated into pRS315 to generate p315-INP53.

Strains, media, and genetic techniques:
Strains used in this study are shown in Table 1. GPY982 was constructed by transforming SEY6210 with YIpCHC521Cla linearized with XbaI and selecting for integrants on SD -ura medium. Loss of URA3 along with the wild-type 3' end of CHC1 was selected on 5-fluoroorotic acid (5-FOA) medium. GPY1010-5B is a meiotic progeny from a cross of SEY6211 and GPY982.

To generate a disruption in the INP53 gene, the primer pairs 5'-TGGGGCGAAGAATATCTAGTTATCCACTCCTTCATAGAATGATTGTACTGAGAGTGCACC-3', 5'-GGCGCAAATCCTGATCCAAAC-3' and 5'-CATTTTGGGGTCAATGGCTGCCATGAGTCTAAAGTCATATCATCTGTGCGGTATTTCACACCG-3', 5'-CGGCTGGTCGCTAATCGTTG-3' were used to generate two overlapping PCR products using pRS303 (SIKORSKI and HIETER 1989 ) as the template. These PCR products were cotransformed into SEY6210, GPY982, and/or GPY1056. Disruption of the INP53 locus by homologous recombination was confirmed by Southern blot analysis and/or PCR. Strains disrupted at the INP51 locus were generated in a similar fashion using the primer pairs 5'-GTTGTGTTAATCGTATGAATTCGAAGCACATTTCACTACAATGATTGTACTGAGAGTGCACC-3', 5'-GGCGCAAATCCTGATCCAAAC-3' and 5'-GTGGCTCATCTTCGTTCTCAACGAATGGATTGGGATCTCCATCTGTGCGGTATTTCACACCG-3', 5'-CGGCTGGTCGCTAATCGTTG-3' in PCR reactions using pRS303 as the template. INP52 was disrupted with the HIS3 gene similarly using primer pairs 5'-ACGCAAAGGCAGCAGAATCAAAAACAAATACTCAGTAGCTATGATTGTACTGAGAGTGCACC-3', 5'-GGCGCAAATCCTGATCCAAAC-3' and 5'-GTAACACAATTTAATTGGGGTCGCAAGGCTTCAATGGATGAACATCTGTGCGGTATTTCACACCG-3', 5'-CGGCTGGTCGCTAATCGTTG-3' using pRS303 as template. INP52 was disrupted with the TRP1 gene using the primer pairs 5'-ACGCAAAGGCAGCAGAATCAAAAACAAATACTCAGTAGCTATGATTGTACTGAGAGTGCACC-3', 5'-GGTATTCTTGCCACGACTCATC-3' and 5'-GTAACACAATTTAATTGGGGTCGCAAGGCTTCAATGGATGAACATCTGTGTGGTATTTCACACCG-3', 5'-CAGAATGTGCTCTAGATTCGg-3' using pRS304 (SIKORSKI and HIETER 1989 ) as template.

RIC1 was disrupted to generate GPY1480 by transforming SEY6210 with pric1-1 digested with XhoI. Similarly a SalI-BamHI fragment from ptcs3-1 was used to disrupt LUV1, and an EcoRI fragment from pptc1-1 was used to disrupt PTC1. To create a ptc1 hog1 double mutant (GPY1371), a ClaI-BamHI fragment from pDHG16 (provided by H. Saito, Harvard Medical School, Boston, MA) was transformed into a strain carrying the ptc1 allele found in our screen which had been backcrossed to a wild-type strain. All disruptions were verified by Southern blot analysis or by PCR. Strains bearing a chromosomal copy of the vps1-ts allele were constructed with pCAV40, an integrating vector containing the vps1-ts allele (provided by T. Stevens, University of Oregon, Eugene, OR). pCAV40 was digested with EcoRI and transformed into CHC1 cells (GPY1100) or chc1-521 cells (GPY418). Ura⁺ transformants were plated onto 5-FOA-containing media, and resulting colonies were assayed for secretion of carboxypeptidase Y (CPY) at 37°. A collection of VPS mutants was kindly provided by B. Horazdovsky (The University of Texas Southwestern Medical Center, Dallas, TX).

YPD medium is 1% Bacto-yeast extract, 2% Bactopeptone, and 2% dextrose. SD is 0.67% yeast nitrogen base without amino acids and 2% dextrose. Supplemented SD is SD with 40 µg/ml adenine, 30 µg/ml leucine, 30 µg/ml lysine, 20 µg/ml histidine, 20 µg/ml uracil, and 20 µg/ml tryptophan. SD -ura, SD -his, and SD -trp are supplemented SD without uracil, histidine, or tryptophan, respectively. SD CAA medium is supplemented SD with 5 mg/ml vitamin assay casamino acid mix. SD CAA -ura is SD CAA without uracil. SDYE is supplemented SD with 0.2% yeast extract. 5-FOA and -aminoadipate (-AA) media were prepared as described previously (ROSE et al. 1990 ). Cell densities in liquid culture were measured in a 1-cm plastic cuvette using a Beckman Instruments DU-62 spectrophotometer (Beckman Instruments, Fullerton, CA). One A₅₀₀ unit is equivalent to 2.3 x 10⁷ cells/ml.

Standard techniques for yeast mating, sporulation, and tetrad analysis were used (GUTHRIE and FINK 1991 ). DNA transformations were performed as previously described (GIETZ and SCHIESTL 1995 ).

Mutagenesis and genetic screen for tcs mutants:
The plasmid loss strategy relied on the toxic effects of 5-FOA and -AA in cells that express the URA3 and LYS2 products, respectively (BASSON et al. 1987 ). GPY1056 was grown at 30° in SD CAA -ura medium to stationary phase and plated at a density of 1000 cells/plate. Plates were exposed to UV irradiation, and irradiated cells were allowed to recover at 30°. This treatment resulted in 16% cell viability. Approximately 16,000 UV-irradiated colonies were replica plated onto -AA medium and incubated at 30°. Colonies that did not grow on -AA-containing medium were patched onto YPD agar, allowed to grow at 30°, and replica plated onto 5-FOA-containing medium. Out of 77 potential tcs mutants, 25 remained unable to grow on 5-FOA when transformed with a HIS3 CHC1 plasmid, indicating that these mutants are sensitive to 5-FOA even when expressing wild-type CHC1. The remaining 52 mutants were successively backcrossed to GPY1056 or GPY1057 three times. Segregants from the third backcross were used for all assays except for tcs2, TCS9-1, and tcs10 for which the original mutagenized strains were used.

Filter overlay blot:
Secretion of p2CPY was assessed using a filter overlay blot assay performed essentially as described previously (WILSBACH and PAYNE 1993B ) except antibody against p2CPY was used (provided by T. Stevens).

Metabolic labeling and immunoprecipitation:
For metabolic labeling of -factor, cells were grown to midlogarithmic phase in SD CAA -ura at 24° or 30°. Labeling and immunoprecipitation was performed as described by SEEGER and PAYNE 1992B except that labeling time was as indicated in figure legends. ₂-Macroglobulin (10 µg/ml) was added to experiments shown in Fig 7 to stabilize secreted pheromone.

View larger version (64K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Maturation of -factor is incomplete in a subset of tcs mutants. Wild-type (WT, GPY1056) and tcs1-15 mutant strains (GPY2172-GPY2186) carrying the chc1-521 allele and a plasmid-borne copy of CHC1 and URA3 were grown in SD CAA -ura overnight at 30°. tcs13 was grown in SDYE media overnight at 24° and shifted to 30° before metabolic labeling due to poor growth in SD CAA -ura at 30°. Cells were metabolically labeled with [35S]methionine/cysteine for 45 min at 30°. -Factor was immunoprecipitated from the culture supernatant and subjected to SDS-PAGE and autoradiography.

View larger version (66K): [in this window] [in a new window] [Download PPT slide]   Figure 2. p2CPY is secreted by a subset of tcs mutants. Wild-type (WT, GPY1056) and tcs mutant strains (GPY2172-GPY2186) carrying the chc1-521 allele and a plasmid with CHC1 and URA3 were grown in SD CAA -ura media to saturation at 24°. Cells were diluted to 1 x 106 cells/ml, and 3 µl was spotted onto YPD agar. After 2 days of growth at 24°, cells were replica plated onto YPD agar, overlaid with a nitrocellulose filter, and incubated overnight at 30°. Filters were probed with a monoclonal antibody to p2CPY.

View larger version (66K): [in this window] [in a new window] [Download PPT slide]   Figure 3. A subset of tcs mutants displays abnormal vacuolar morphology. Wild-type (WT, GPY1056), tcs1 (GPY2172), tcs3 (GPY2174), TCS9-1 (GPY2180), and tcs15 (GPY2186) strains were grown overnight at 30° in SD CAA -ura to mid-logarithmic phase. Cells were incubated with FM4-64 for 15 min and resuspended in fresh medium devoid of dye and allowed to internalize the dye for 45 min at 30° before viewing. FM4-64 fluorescence (left) and differential interference contrast (DIC) optics (right) are shown.

View larger version (49K): [in this window] [in a new window] [Download PPT slide]   Figure 4. A temperature-sensitive allele of VPS1 accentuates the growth defect of chc1-521 cells. Wild-type (WT, GPY1100), chc1-521 (GPY418), vps1-ts (GPY832), and chc1-521 vps1-ts (GPY775) strains were grown overnight to saturation at 24° in YPD. Serial dilutions of each culture were spotted on YPD agar and incubated at 24° or 37°.

View larger version (34K): [in this window] [in a new window] [Download PPT slide]   Figure 5. The -factor maturation defect of chc1-521 cells is enhanced by the vps1-ts allele. Wild-type (WT, GPY1100), chc1-521 (GPY418), vps1-ts (GPY832), and chc1-521 vps1-ts (GPY775) strains were incubated at 24° to midlogarithmic growth. After a 2-hr shift to 30°, cells were metabolically labeled at 30° for 45 min, and -factor was immunoprecipitated as described in the legend to Fig 1.

View larger version (73K): [in this window] [in a new window] [Download PPT slide]   Figure 6. Complementation of tcs mutants with genomic DNA fragments. Strains carrying tcs mutant alleles were transformed with plasmids containing the indicated gene isolated from a genomic library or the parental vector (+ vector). Transformants were grown to saturation at 24° in medium selecting for the appropriate plasmid, and serial dilutions were spotted onto supplemented SD (-5-FOA) and 5-FOA-containing media (+5-FOA) and allowed to grow at 30° for 3 days.

View larger version (45K): [in this window] [in a new window] [Download PPT slide]   Figure 7. Low temperature and the chc1-521 allele accentuate the -factor maturation defect of inp53 cells. (A) Wild-type (WT, SEY6210), inp51 (GPY2078), inp52 (GPY2062), and inp53 (GPY1876) strains were grown at 24° to midlogarithmic phase, shifted to 20° or 30° for 15 min, and labeled with [35S]methionine/cysteine for 45 min at 20° or 30 min at 30°. -Factor was immunoprecipitated from the culture supernatant and subjected to SDS-PAGE and autoradiography. (B) Wild-type (WT, SEY6210), chc1-521 (GPY982), inp51 (GPY2078), chc1-521 inp51 (GPY2141), inp52 (GPY2062), chc1-521 inp52 (GPY2142), inp53 (GPY1876), and chc1-521 inp53 (GPY2143) strains were grown to midlogarithmic phase in SDYE media and metabolically labeled with [35S]methionine/cysteine for 45 min at 24°. -Factor was immunoprecipitated from the culture supernatant and subjected to SDS-PAGE and autoradiography.

FM4-64 labeling:
Yeast cells were labeled with FM4-64 essentially as described by VIDA and EMR 1995 . Cells growing in midlogarithmic phase were incubated in the presence of 40 µM FM4-64 at a cell concentration of 2 x 10⁸ to 4 x 10⁸ cells/ml for 15 min at 30°. Cells were harvested by centrifugation at 700 x g for 3 min, resuspended at 1 x 10⁸ to 2 x 10⁸ cells/ml in fresh medium, and allowed to internalize dye for 45 min at 30°. Cells were then collected by centrifugation, resuspended at 5 x 10⁷ to 1 x 10⁸ cells/ml, and viewed on concanavalin-A-coated microscope slides using a Nikon FX-A Microphot microscope.

Cloning tcs mutants:
tcs mutants were transformed with a single copy genomic library (no. 77162; American Type Culture Collection, Manassas, VA), and transformants were screened for growth on 5-FOA-containing medium. DNA from 5-FOA-resistant transformants was prepared as described previously (ROSE et al. 1990 ), and plasmids were introduced into Escherichia coli by electroporation. Genomic inserts were identified by DNA sequencing. The minimal complementing region of each genomic clone was determined by subcloning individual open reading frames followed by complementation analysis on 5-FOA-containing medium.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of mutations that display synthetic lethality with chc1-521:
A plasmid loss strategy was adopted to identify tcs mutations in chc1-521 cells at 30°. For this approach, a centromere-containing plasmid carrying wild-type CHC1, URA3, and LYS2 genes was introduced into cells with chc1-521, ura3, and lys2 mutant alleles at the chromosomal loci. Previously, we observed that viable chc1 cells transformed with a URA3 CHC1 plasmid were inviable on 5-FOA, presumably because the 5-FOA selection procedure imposes sufficient stress to inhibit the growth of debilitated clathrin-deficient cells that lose the plasmid (G. PAYNE, unpublished results). On the basis of this finding we anticipated that if a tcs mutation together with chc1-521 at 30° causes a severe loss of clathrin function, then the mutant cells should be inviable on 5-FOA or -AA.

To isolate tcs mutants, plasmid-containing cells were mutagenized with ultraviolet irradiation and allowed to grow into colonies, and the colonies were screened for growth on -AA- and 5-FOA-containing medium. Of 16,000 colonies formed from mutagenized cells, 52 tcs candidate strains were backcrossed three times to the parental strain, and those that yielded 2:2 segregation of the 5-FOA-sensitive phenotype were analyzed further. Complementation tests defined one dominant mutation (TCS9-1) and 14 recessive tcs complementation groups (Table 2). The majority of complementation groups contained a single isolate, indicating that the screen was subsaturating (Table 2).

A subset of tcs mutants affects clathrin-mediated protein trafficking events:
The effect of tcs mutations in a CHC1 background on clathrin-dependent protein transport processes was assessed. First, maturation of the mating pheromone -factor precursor was examined as a measure of proper localization of the TGN membrane protein Kex2p. Kex2p is responsible for the cleavages that initiate proteolytic maturation of the -factor precursor in the TGN (FULLER et al. 1988 ). In cells with defective clathrin heavy chain, Kex2p is mislocalized to the cell surface (PAYNE and SCHEKMAN 1989 ). The resulting depletion of TGN-localized Kex2p leads to inefficient -factor precursor maturation and secretion of the highly glycosylated form of pheromone. Other mutations, such as vacuolar protein sorting (vps) mutations, that affect trafficking between the TGN and endosomes can also lead to Kex2p mislocalization and attendant defects in -factor precursor maturation (WILSBACH and PAYNE 1993B ). Therefore, secretion of highly glycosylated -factor precursor serves as a convenient and reliable indicator of Kex2p mislocalization. Secreted forms of -factor were detected by radiolabeling tcs cells with [³⁵S]methionine, immunoprecipitating -factor from the culture supernatant, and subjecting the immunoprecipitates to SDS-PAGE. The mature 13-amino-acid -factor peptide migrates to the bottom of the gel, while the highly glycosylated form remains near the top (Fig 1). Incompletely processed forms of -factor can also be observed migrating immediately above the mature form. Of the 15 mutant strains, 9 secreted detectable levels of highly glycosylated -factor precursor, ranging from a severe defect in tcs1-3 mutants to a minor defect in the TCS9-1 mutant (Fig 1; Table 2). These data suggest that tcs1 through TCS9-1 mutations affect localization of the TGN membrane protein Kex2p.

The integrity of vacuolar protein transport was determined by monitoring biosynthesis of the vacuolar hydrolase, CPY. Newly synthesized CPY is translocated into the endoplasmic reticulum and core glycosylated to form a 67-kD species (p1CPY). Upon transit through the Golgi apparatus, p1CPY is further glycosylated to form a 69-kD species (p2CPY; STEVENS et al. 1982 ). In the TGN, p2CPY is recognized by a receptor, Pep1p/Vps10p, and diverted from the secretory pathway into vesicles targeted to endosomes. In prevacuolar endosomes, Pep1p/Vps10p is thought to dissociate from p2CPY allowing vesicle-mediated retrieval of Pep1p/Vps10p to the TGN (MARCUSSON et al. 1994 ; COOPER and STEVENS 1996 ). Dissociated p2CPY continues to the vacuole where proteolytical maturation generates a 61-kD species (STEVENS et al. 1982 ). Inefficient localization of Pep1p/Vps10p to the TGN and/or defects in the TGN to endosome to vacuole pathway result in secretion of p2CPY (BRYANT and STEVENS 1998 ). The tcs mutants were tested for secretion of p2CPY using a filter overlay and immunoblotting with a p2CPY-specific monoclonal antibody. Missorting of p2CPY was detected in tcs1 through TCS9-1 and in tcs12 strains (Fig 2). Qualitatively strong defects were apparent in tcs2, tcs5, tcs6, tcs7, tcs8, and TCS9-1 strains (Fig 2; Table 2). tcs1 and tcs3 strains exhibited milder phenotypes, while tcs4 and tcs12 strains secreted slight but reproducible levels of p2CPY (Fig 2; Table 2). Except for tcs12 (but see below), the same set of mutants were defective in -factor precursor maturation and p2CPY sorting, though the relative extent of the defects in each process differed between strains (Table 2). We define this set as class 1 mutants. The synthetic growth defects of tcs mutations with the chc1-521 allele and effects of the class 1 mutations by themselves on -factor maturation and p2CPY sorting suggest that products of the class 1 TCS genes influence protein transport pathways between the TGN and endosomes.

tcs mutations in VPS genes:
Studies of vps mutants that missort p2CPY have resulted in identification of >50 genes involved in vacuolar protein transport from the TGN to vacuoles (BRYANT and STEVENS 1998 ). To determine whether tcs mutations occurred in previously known VPS genes, p2CPY secretion was used as an assay in complementation tests between tcs mutants and a gallery of vps mutants. Diploids from noncomplementing crosses were induced to undergo meiosis and subjected to linkage analysis. By these criteria vps1(tcs2), vps5(tcs5), pep12/vps6(tcs7), vps17(tcs8), and vps21(tcs6) were present in the class 1 tcs collection (Table 2).

Trafficking defects in vps strains can result in morphological changes to the vacuole. On the basis of vacuole morphology, the vps mutants have been classified into six groups, A–F (RAYMOND et al. 1992 ). The vps mutants identified in the tcs screen represent three classes, B, D, and F. Class B (vps5, vps17) is characterized by fragmented vacuoles, class D (pep12, vps21) by large single vacuoles, and class F (vps1) by large vacuoles encircled by smaller vacuoles. Vacuolar morphology in the tcs collection was visualized with the lipophilic vital dye FM4-64, which stains the vacuole membrane (VIDA and EMR 1995 ). In general, the vacuole morphology of strains with tcs mutations in known VPS genes corresponded to earlier classifications of vps mutants. Highly fragmented vacuoles, similar to class B vps mutants, were also observed in tcs1, tcs3, and tcs15 cells (Fig 3). In addition, TCS9-1 cells displayed an abnormally large percentage of cells containing single, large vacuoles (Fig 3). An increase in cell size was also observed in the TCS9-1 strain (Fig 3, compare DIC panel with other panels). The vacuole morphology defects observed in tcs strains are consistent with a role of these gene products in membrane traffic to the vacuole.

Of the three tcs mutant strains exhibiting both strong -factor maturation and CPY sorting defects (tcs1, tcs2, tcs3), only tcs2 represented a known vps locus, vps1. VPS1 encodes a member of the dynamin family of GTPases (VATER et al. 1992 ). In mammalian cells, dynamin is involved in scission of invaginated clathrin-coated pits at the plasma membrane to form free clathrin-coated vesicles (SCHMID 1997 ; SCHMID et al. 1998 ). By analogy, Vps1p and clathrin could function together at the TGN in yeast. Consistent with this idea, similar defects in TGN membrane protein localization have been described in vps1 and chc1-521 mutants (WILSBACH and PAYNE 1993B ; NOTHWEHR et al. 1995 ). To probe the relationship of Vps1p and clathrin in more detail, we constructed a strain containing chc1-521 and a temperature-sensitive allele of VPS1 (vps1-ts; VATER et al. 1992 ). Growth and -factor maturation were evaluated in the double mutant, congenic single mutants, and the wild-type strain. At 24°, growth of the single mutants was commensurate with the wild-type strain, but the double mutant grew at a slightly slower rate. At 37°, the chc1-ts cells grew somewhat more slowly than the vps1-ts and wild-type strains. Compared to the single mutants, growth of the double mutant was severely affected at 37° (Fig 4). A striking synthetic effect of the double mutant combination on -factor precursor maturation was also apparent (Fig 5). At 30°, only mature -factor was secreted by the wild-type and vps1-ts strains. The congenic chc1-521 strain used in this comparison secreted a minor amount of highly glycosylated precursor and intermediate cleavage products (16%). In contrast, the double mutant secreted a substantial level of precursor forms (47%). The synergistic effects of CHC1 and VPS1 conditional alleles indicate a sensitive functional interdependence of Vps1p and clathrin in cell growth and Kex2p localization in the TGN. These results, together with the identification of tcs mutations in other VPS genes whose products are known to function in transport between the TGN and endosomes, demonstrate that the tcs screen can be an effective approach to define proteins which act in clathrin-dependent transport pathways.

RIC1 and LUV1/RKI1 are class 1 TCS genes:
Four class 1 tcs mutants (tcs1, tcs3, tcs4, and TCS9-1) did not correspond to a previously identified vps complementation group. Two of these mutants, tcs1 and tcs3, displayed growth defects at elevated temperatures (37°) as well as -factor maturation defects and CPY sorting defects at lower temperatures (24° and 30°). On the basis of the relatively severe -factor maturation defects in these strains, we focused on identifying the mutant loci. To isolate wild-type versions of tcs1 and tcs3, mutant strains were transformed with a genomic library carried by a centromere-containing plasmid. The resulting transformants were screened for the ability to grow on medium containing 5-FOA. Genomic DNA fragments complementing the 5-FOA growth defect were subsequently dissected to identify the complementing gene (Fig 6).

A single open reading frame containing RIC1 complemented all mutant phenotypes of tcs1 mutants (Fig 6; data not shown). A mutant allele of RIC1 was identified by MIZUTA et al. 1997 in a screen for genes involved in ribosome synthesis. This ric1 mutant was reported to be temperature sensitive for growth and to exhibit reduced levels of transcripts encoding both ri-bosomal protein genes and ribosomal RNA after shift to the nonpermissive temperature (37°; MIZUTA et al. 1997 ). However, many secretory pathway mutants display similar ribosome synthesis defects, raising the possibility that reduced synthesis of ribosome components in ric1 cells is a secondary consequence of a defect in protein trafficking (see DISCUSSION; MIZUTA and WARNER 1994 ). The primary sequence of the 1056-amino-acid Ric1 protein reveals no significant homology to known proteins or motifs. Disruption of RIC1 (ric1) produced phenotypes identical to those of the ric1 mutant isolated in the tcs screen (data not shown). A cross between the ric1 strain and a strain carrying the ric1/tcs1 allele resulted in a diploid which displayed ric1 mutant phenotypes, providing further evidence that the mutant locus in the tcs1 strain is RIC1.

Complementation of the tcs3 mutant resulted in isolation of LUV1/RKI1. Introduction of LUV1 on a centromeric plasmid was sufficient for complete complementation of the tcs3 mutant phenotypes (Fig 6; data not shown). LUV1 also has been isolated through a screen for mutations that cause synthetic growth defects with a calcineurin mutant (M. CONBOY and M. CYERT, personal communication) and (as RKI1) in a screen for mutations affecting microtubules (SMITH et al. 1998 ). LUV1 encodes an open reading frame that has the potential to encode a 101.5-kD protein with a predicted coiled-coil domain. A BLAST database search with the Luv1p amino acid sequence identified two related open reading frames of unknown function in other organisms: Schizosaccharomyces pombe (E value = 4 x 10^-12, accession no. CAB16266) and Arabidopsis thaliana (E value = 3 x 10^-9, accession no. CAA16926; ALTSCHUL et al. 1997 ). We constructed a strain carrying a disruption of TCS3. The phenotypes of this strain mirrored those of the original tcs3 mutant, including a growth defect at elevated temperatures, secretion of highly glycosylated -factor, and secretion of p2CPY (data not shown). Mutant phenotypes were not complemented in a diploid from a cross between a tcs3 strain and a strain carrying the original tcs3 allele. We conclude that mutation of the evolutionarily conserved product of LUV1/RKI1 is responsible for the phenotypes caused by tcs3.

PTC1 is a class 2 TCS gene:
A genomic fragment was also isolated that rescued the chc1-521-dependent 5-FOA growth defect of the class 2 tcs mutant tcs11. PTC1 was sufficient for rescue of the 5-FOA growth defect (Fig 6). A cross between a ptc1 chc1-521 strain carrying a CHC1 URA3 plasmid and a tcs11 chc1-521 strain also carrying this plasmid resulted in a diploid unable to grow on 5-FOA-containing medium. The lack of complementation argues that tcs11 is a mutant allele of PTC1. PTC1 encodes a type 2C serine/threonine phosphatase implicated in multiple cellular functions, including osmotic stress response, tRNA biosynthesis, and mitochondrial inheritance (VAN ZYL et al. 1989 ; MAEDA et al. 1993 , MAEDA et al. 1994 ; ROEDER et al. 1998 ). A strain with a disrupted copy of PTC1 displayed normal -factor maturation, CPY sorting, endocytosis as assayed by -factor internalization, and vacuolar morphology (data not shown). Therefore, at present the only connection between PTC1 and protein trafficking is the synthetic growth defect with chc1-521.

We considered the possibility that synthetic growth defects of ptc1 and chc1-521 derive from the role of Ptc1p in osmotic stress response. Cells challenged by high external osmolarity increase internal osmolarity by activating a MAP kinase signal transduction pathway (the HOG pathway) that includes the MAP kinase Hog1p and the MAP kinase kinase Pbs2p (BREWSTER et al. 1993 ). Genetic experiments suggest that Ptc1p can downregulate the HOG pathway through activities on Hog1p and/or Pbs2p (MAEDA et al. 1994 ). Thus, the absence of Ptc1p could allow increased basal activity of the HOG pathway MAP kinase, leading to an increase in internal osmolarity. If plasma membrane composition was altered by the defect in clathrin, an increase in internal osmolarity might lead to cell lysis. As a test of this possibility, Hog1p was eliminated by gene disruption in the ptc1 chc1-ts strain carrying the CHC1 URA3 plasmid. As a control, HOG1 was also disrupted in a chc1-521 carrying the CHC1 URA3 plasmid. Disruption of HOG1 did not suppress the 5-FOA growth defect in the ptc1 chc1-ts strain and did not itself cause synthetic growth defects with chc1-521. These findings argue against the idea that synthetic lethality caused by ptc1 is due to upregulation of the HOG pathway.

INP53 is a TCS gene:
The library fragment that complemented tcs12 carried INP53/SJL3/SOP2, one of three genes (INP51, INP52, or INP53) encoding synaptojanin-like inositol polyphosphate 5-phosphatases (SRINIVASAN et al. 1997 ; SINGER-KRUGER et al. 1998 ; STOLZ et al. 1998 ; GUO et al. 1999 ). INP53 was identified previously in a screen for mutations that suppress vacuolar targeting of a mutant form of the plasma membrane ATPase (LUO and CHANG 1997 ). The mutant allele of INP53 identified in the screen, designated sop2, caused a subtle -factor maturation defect at 24°. The tcs12 mutant allele isolated in our screen was originally assigned to class 2 based on the absence of an -factor maturation defect at 30°. Recognition of tcs12 as inp53 prompted us to reevaluate the -factor maturation phenotype of inp53 cells and compare the effects of inp53 to inp51 and inp52. For this purpose, isogenic strains were generated carrying disruptions of either INP51, INP52, or INP53. We examined -factor maturation after shifting cells from 24° to either 30° or 20°, choosing 20° by reasoning that a defect in a process involving the lipid bilayer might be exaggerated at lower temperatures. Wild-type, inp51, and inp52 strains secreted exclusively mature -factor at 30° (Fig 7A). At this temperature, a minor maturation defect was apparent in the inp53 strain (Fig 7A). At 20°, the defect in inp53 cells was more pronounced, whereas the other two inp mutants were essentially unaffected by this temperature shift (Fig 7A). These results prompted reassignment of tcs12 as a class 1 mutation and suggest a specific role for Inp53p in Kex2p localization.

To explore the specificity of genetic interactions between inp53 and chc1-521, the INP genes were individually disrupted in a chc1-521 strain carrying the CHC1 URA3 plasmid. As shown in Fig 8A, only the combination of inp53 with chc1-521 resulted in inviability on 5-FOA medium at 30°. Combining inp51 with chc1-ts did not prevent growth on 5-FOA, but the colonies were smaller than those from the chc1-521 control strain. The inp52 allele had no effect when combined with chc1-521. As an alternative approach to examining synthetic growth defects, a diploid strain heterozygous for inp53 and homozygous for chc1-521 was induced to sporulate, and the meiotic progeny were subjected to tetrad analysis. Each tetrad yielded four viable segregants, indicating that the inp53 chc1-521 combination is not lethal when double mutants are obtained by this method. This finding was not totally unexpected given earlier findings with chc1 strains, indicating that loss of a complementing plasmid on 5-FOA is a more stringent growth condition than direct incubation of a mutant on standard medium (see above). The viability of inp53 chc1-521 meiotic progeny encouraged us to disrupt the individual INP genes directly in chc1-521 haploids, generating a set of congenic double mutants. Growth of each double mutant at 24° was equivalent to the wild-type growth of the parental chc1-ts strain (Fig 8B). A striking growth defect was observed when the inp53 chc1-521 strain was incubated at 30° and 37° (Fig 8B). Growth of the inp51 chc1-521 strain was also compromised at the elevated temperatures but less than growth of the inp53 chc1-521 strain. The inp52 chc1-521 strain mimicked the chc1-521 parental strain at 30° and 37°. The specificity of the inp chc1-521 interactions was probed further by analyzing -factor maturation in each double mutant at the permissive growth temperature, 24° (Fig 7B). Little or no defect was apparent in any of the single mutants (Fig 7B). In the double mutants, pairing chc1-521 with either inp51 or inp52 had only a marginal effect on -factor maturation, but combining chc1-521 and inp53 produced a strong pheromone maturation defect (42% precursor forms). These results are consistent with effects of the individual inp mutations on -factor maturation and point to a specific role for Inp53p in clathrin-mediated TGN localization of Kex2p. It remains to be determined whether, like inp53, viable combinations of other class 2 tcs mutations with chc1-521 can be isolated and whether such combinations display TGN sorting defects.

View larger version (49K): [in this window] [in a new window] [Download PPT slide]   Figure 8. inp53 accentuates the growth defect of chc1-521 cells. (A) chc1-521 (GPY1056), chc1-521 inp51 (GPY2064), chc1-521 inp52 (GPY2162), and chc1-521 inp53 (GPY1877) strains, all of which carry the pULE-CHC1 plasmid, were grown to saturation in YPD media at 24°. Serial dilutions of each culture were spotted onto 5-FOA media and allowed to grow at 30°. (B) chc1-521 (GPY982), chc1-521 inp51 (GPY2141), chc1-521 inp52 (GPY2142), and chc1-521 inp53 (GPY2143) strains were grown to saturation at 24° in YPD media. Serial dilutions of each culture were spotted onto YPD agar and allowed to grow at 24°, 30°, or 37°.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

As a genetic strategy to identify proteins involved in clathrin-dependent protein transport pathways, we carried out a screen for mutations that cause synthetic growth defects in a strain expressing a partially functional clathrin heavy chain. The tcs mutations recovered in this screen divide into two classes based on their effects on protein trafficking in the TGN/endosome system. Class 1 mutations cause defects in -factor maturation, a reliable signature of Kex2p mislocalization, and defects in biosynthetic sorting of CPY to the vacuole. Measurements of Kex2p stability in selected class 1 mutants indicate higher than normal rates of Kex2p vacuolar degradation (data not shown), supporting conclusions based on the -factor maturation assay. Class 2 mutations do not affect Kex2p localization or CPY sorting.

One of the genes represented in the tcs collection was INP53, encoding a type II inositol polyphosphate 5-phosphatase. Classification of Inps is based primarily on substrate specificity of animal cell enzymes, and type II Inps characteristically are able to hydrolyze the C5 phosphate from the inositol moiety of both phosphatidylinositol 4,5-bisphosphate (PI[4,5]P₂) and phosphatidylinositol 3,4,5-trisphosphate (PI[3,4,5]P₂) (ERNEUX et al. 1998 ). A subset of type II INPs is further distinguished by the presence of a domain related to the yeast protein Sac1p, a protein implicated in phospholipid metabolism and vesicular trafficking (CLEVES et al. 1989 ; WHITTERS et al. 1993 ). Recently, the Sac1 domain was found to encode a novel polyphosphoinositide phosphatase activity, indicating that this subgroup of type II Inps has the potential to express two distinct inositol phosphatase activities from one polypeptide (GUO et al. 1999 ). Synaptojanin, the founding member of the Sac1 domain containing INP subfamily, has been proposed to act in clathrin-dependent endocytosis in nerve terminals (MCPHERSON et al. 1996 ; HAFFNER et al. 1997 ). Of the four yeast proteins with homology to INPs, three, Inp51–53, also carry an N-terminal Sac1p domain and are alternatively referred to as synaptojanin-like (Sjl; SRINIVASAN et al. 1997 ; SINGER-KRUGER et al. 1998 ; STOLZ et al. 1998 ). However, the Sac1-like domain in Inp51p is catalytically inactive (GUO et al. 1999 ). Systematic analyses of endocytic and vacuolar sorting pathways in strains with individual disruptions of the INP genes have been carried out. These studies revealed only subtle vacuolar morphology anomalies in inp52 or inp53 cells, but no defects in CPY sorting or in endocytic uptake of the lipophilic endocytic tracer FM4-64 (SRINIVASAN et al. 1997 ; SINGER-KRUGER et al. 1998 ; STOLZ et al. 1998 ). Double mutant combinations have no effect on CPY sorting but result in varying levels of endocytic defects; the inp51 inp52 combination is more detrimental than inp52 inp53, while the inp51 inp53 pair is innocuous (SRINIVASAN et al. 1997 ; SINGER-KRUGER et al. 1998 ; STOLZ et al. 1998 ). Although results from double mutants offer some distinction between the Inps, particularly with respect to endocytosis, the absence of informative phenotypes in single mutants has hindered functional definition of specific roles for any of the individual Inps in vesicular transport events. Our results indicate that, of the three synaptojanin-like INP genes, only disruption of INP53 results in an -factor maturation defect at 20° and 30°, accentuation of -factor maturation defects when combined with chc1-521 at 24°, and an inability to grow on 5-FOA when introduced into chc1-521 cells carrying a CHC1 URA3 plasmid. Additionally, on standard media, the combination of inp53 with chc1-521 showed stronger synthetic growth defects than either inp51 or inp52. These findings argue that Inp53p specifically affects the clathrin-dependent TGN to endosome traffic pathway. Our studies confirm and extend the work of LUO and CHANG 1997 , who detected minor levels of secreted -factor precursor and marginally decreased steady-state levels of Kex2p in inp53 cells at 24°, but did not analyze mutations in other INP genes. By analogy to the proposed function of synaptojanin in endocytic clathrin-coated vesicle traffic, we suggest that Inp53p participates in clathrin-coated vesicle traffic from the TGN.

The synthetic growth defect of chc1-521 inp51 cells, although less severe than that of chc1-521 inp53 cells, suggests that Inp51p could also participate in a clathrin-dependent transport process. Since inp51 alone, or in combination with chc1-521, did not alter -factor maturation, Inp51p is more likely to be involved in endocytosis. In support of this idea, inp51 causes synthetic lethality when combined with a temperature-sensitive allele of PAN1 (WENDLAND and EMR 1998 ). Mutations in PAN1 affect endocytosis, and the Pan1 protein is homologous to the mammalian clathrin accessory protein, Eps15 (WENDLAND et al. 1996 ; TANG et al. 1997 ). Furthermore, inp51 combined with inp52 causes severe endocytosis defects (SINGER-KRUGER et al. 1998 ). Unlike inp51 and inp53, inp52 did not display synthetic effects with chc1-521. Thus, if Inp52p is involved in clathrin-dependent transport, its contribution is not substantial enough to be detected by our assays.

A number of class 1 tcs mutations occurred in known VPS genes, including VPS1. The genetic interaction between VPS1 and CHC1 extends to temperature-sensitive alleles, which produce synthetic defects in growth and -factor maturation, suggesting a strong connection between the functions of Vps1p and clathrin. Vps1p is a member of the dynamin family of GTPases (VATER et al. 1992 ). In mammalian cells, dynamin assembles into a ring around the necks of invaginated clathrin-coated pits at the plasma membrane (TAKEI et al. 1995 ). It is generally agreed that GTP hydrolysis is necessary to sever the connection between the nascent vesicle and plasma membrane, though the model that dynamin itself is the severing agent has been recently challenged (SWEITZER and HINSHAW 1998 ; SEVER et al. 1999 ). Given the interaction of dynamin and clathrin in mammalian cell endocytosis, it is reasonable to posit a similar relationship between Vps1p and clathrin. Consistent with this view, vps1 and chc1 cells share a similar, and distinctive, defect in TGN membrane protein localization. In both types of mutant cells, TGN membrane proteins are mislocalized to the cell surface. In chc1 cells, the additional endocytic defect causes accumulation at the cell surface, whereas in vps1 cells the TGN proteins are internalized and delivered to the vacuole (SEEGER and PAYNE 1992B ; NOTHWEHR et al. 1995 ). Since TGN membrane protein localization involves cycling between the TGN and endosomes, plasma membrane mislocalization in vps1 and chc1 cells suggests that these mutations block the endosome-targeted pathway at the TGN. The observation of synthetic effects on -factor maturation by temperature-sensitive alleles of VPS1 and CHC1 now establishes a functional connection between the gene products in TGN membrane protein localization. Additional experiments, particularly subcellular localization of Vps1p and clathrin, will be needed to test for physical links.

Mutant alleles of four other well-characterized VPS genes were identified in the tcs screen (vps21, pep12/vps6, vps5, and vps17). Vps21p, a small GTPase of the Ypt/rab family, and Pep12p, an endosomal t-SNARE, are proposed to act in targeting and fusion of TGN-derived vesicles with endosomes (HORAZDOVSKY et al. 1994 ; BECHERER et al. 1996 ). Interestingly, Vps5p and Vps17p form a subcomplex within a multimeric assembly of Vps proteins thought to constitute a vesicle coat essential for endosome to TGN traffic (HORAZDOVSKY et al. 1997 ; SEAMAN et al. 1998 ). Considering the current model that clathrin functions in vesicular transport between the TGN and endosomes, it is not surprising to find genetic interactions between chc1-521 and mutations that affect other stages of TGN-endosome membrane trafficking. Since the tcs screen was not carried out to saturation, a more systematic examination will be needed to determine whether the vps/tcs mutations identify a subset of VPS genes particularly sensitive to clathrin deficiencies or whether genetic interactions with chc1-521 will prove to be a ubiquitous VPS feature.

Two class 1 genes, RIC1 and LUV1/RKI1, have not been implicated previously in vesicular transport. RIC1 was originally identified in a screen for temperature-sensitive mutants with reduced synthesis of ribosome components. At the elevated temperature, transcript levels of both ribosomal proteins and RNA were lowered (MIZUTA et al. 1997 ). However, a similar reduction in ribosomal protein and rRNA synthesis has been documented for ts mutations affecting different stages of the secretory pathway, revealing a regulatory pathway connecting secretory pathway defects to ribosome biosynthesis (MIZUTA and WARNER 1994 ). Several observations suggest that the defects in synthesis of ribosomal components in ric1 cells are similarly indirect, stemming from a primary defect in vesicle trafficking. First, although the ric1/tcs1 strain is temperature sensitive for growth, we observed -factor maturation defects, CPY missorting, and vacuole fragmentation at permissive growth temperatures where no obvious defects in protein synthesis were apparent. Second, in the genetic background used in the tcs screen, disruption of RIC1, or introduction of the ric1 allele identified in the ribosome synthesis screen, yielded the same spectrum of trafficking defects at permissive growth temperatures that were detected in the ric1/tcs1 cells (E. BENSEN and G. PAYNE, unpublished results). The primary sequence of Ric1p does not offer clues about molecular function. Further experiments analyzing ric1 mutants and the Ric1 protein are in progress.

TCS3/LUV1/RKI1 (which we will refer to as LUV1) has been identified in multiple screens. In our studies, the original tcs3 allele and a LUV1 disruption cause a substantial -factor maturation defect, CPY missorting, and vacuole fragmentation at permissive growth temperatures, suggesting a role for Luv1p in vesicle traffic between the TGN and endosomes. A luv1 allele was also isolated in a screen for mutations that cause synthetic growth effects with a disruption of the regulatory subunit of calcineurin encoded by CNB1 (M. CONBOY and M. CYERT, personal communication). This analysis demonstrated vacuole fragmentation in luv1 cells, partial secretion of the Golgi form of CPY, and sensitivity of cell growth to a variety of ions (M. CONBOY and M. CYERT, personal communication). In addition, Luv1p has been implicated recently in microtubule function. Cells harboring luv1/rki1 are hypersensitive to microtubule-depolymerizing drugs, and at the nonpermissive growth temperature they display a loss of microtubule structures. Furthermore, Luv1p can bind to Rbl2p, a protein associated with free ß-tubulin subunits in cells (ARCHER et al. 1995 ; SMITH et al. 1998 ). The pleiotropic defects in cells lacking Luv1p make it difficult to ascribe a direct role for the protein in any one of the affected processes. It may be that Luv1p provides a single cellular function that impacts multiple aspects of cell physiology. Alternatively, the protein may directly function in distinct pathways. For example, the view that Rbl2p acts in managing assembly of tubulin heterodimers could indicate a general function for Luv1p in regulating assembly of multimeric protein assemblies. Isolation of a conditional allele of LUV1 and further examination of Luv1p-interacting partners may allow these possibilities to be distinguished.

The class 2 tcs mutation in the gene encoding the type 2C serine/threonine protein phosphatase Ptc1p caused synthetic growth defects with chc1-521 but did not affect TGN-endosome protein transport. Synthetic effects can result from mutations in genes whose products act in the same process or pathway, but they can also occur when mutant proteins function in distinct processes (GUARENTE 1993 ). Particularly when considering effects on cell growth, the possibility of indirect interactions cannot be minimized. For class 1 mutations, autonomous effects on clathrin-mediated trafficking steps link the products of these TCS genes to protein sorting at the TGN. For class 2 mutations, connections to clathrin are more tenuous. One possibility is that these mutations affect endocytosis, but preliminary experiments do not indicate strong effects on the trafficking of the mating pheromone a-factor receptor (E. BEN-SEN, unpublished results). Published studies of Ptc1p function do not offer significant clues to the basis of its synthetic effects with chc1-521. Ptc1p has been proposed to participate in a multitude of cellular functions, including osmoregulation, tRNA synthesis, DNA recombination, cell wall ß-glucan assembly, and mitochondrial inheritance (VAN ZYL et al. 1989 ; MAEDA et al. 1994 ; HUANG and SYMINGTON 1995 ; JIANG et al. 1995 ; ROEDER et al. 1998 ). It was possible that the synthetic growth defects in ptc1 chc1-521 cells are due to upregulation of basal osmotic stress response (HOG) pathway activity, thereby eliciting an increase in internal osmolarity that could indirectly intensify defects caused by the chc1-521 allele. However, disruption of HOG1 did not restore growth to chc1-521 ptc1 cells, making this scenario unlikely. This experiment also suggests that the effects of ptc1 are not mediated through the recently discovered role of Hog1p in Golgi membrane protein localization (REYNOLDS et al. 1998 ). Thus, given the pleiotropic activities of Ptc1p, a direct role for this phosphatase in clathrin-dependent trafficking is possible but remains to be established.

In summary, characterization of tcs mutants demonstrates that the screen for mutations that cause synthetic growth interactions with chc1-521 constitutes a robust method for identifying proteins that influence clathrin-dependent transport pathways. Our findings strengthen connections between Vps and clathrin function, offer new insights into the specificity of inositol polyphosphate 5-phosphatases, and raise the possibility of membrane-trafficking roles for proteins thought to act in other processes. These results provide a genetic foundation to guide molecular analysis of the proteins identified through the tcs screen, and suggest that further application of the screen will allow novel insights into clathrin-mediated traffic routes.

	ACKNOWLEDGMENTS

We thank Bruce Horazdovsky, Haruo Saito, and Tom Stevens for providing strains, plasmids, and antibodies. We acknowledge Cara Capuano and Kevin Roberg for assistance with cloning the tcs mutations. We are also grateful to Mike Juvet, Ken Oyadomari, and Audrey Nakamura for technical assistance. We thank Dan Rube for critically reading the manuscript. This work was supported in part by the Predoctoral Training Program in Genetic Mechanisms at UCLA (T32-GM07104) and a UCLA Dissertation Year fellowship to E.S.B. and National Institutes of Health grant GM-39040 to G.S.P.

Manuscript received July 12, 1999; Accepted for publication September 15, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALTSCHUL, S. F., T. L.