An arf1D Synthetic Lethal Screen Identifies a New Clathrin Heavy Chain Conditional Allele That Perturbs Vacuolar Protein Transport in Saccharomyces cerevisiae

ADP-ribosylation factor (ARF) is a small GTP-binding protein that is thought to regulate the assembly of coat proteins on transport vesicles. To identify factors that functionally interact with ARF, we have performed a genetic screen in Saccharomyces cerevisiae for mutations that exhibit synthetic lethality with an arf1D allele and defined seven genes by complementation tests (SWA1-7 for synthetically lethal with arf1D). Most of the swa mutants exhibit phenotypes comparable to arf1D mutants such as temperature-conditional growth, hypersensitivity to fluoride ions, and partial protein transport and glycosylation defects. Here, we report that swa5-1 is a new temperature-sensitive allele of the clathrin heavy chain gene (chc1-5), which carries a frameshift mutation near the 39 end of the CHC1 open reading frame. This genetic interaction between arf1 and chc1 provides in vivo evidence for a role for ARF in clathrin coat assembly. Surprisingly, strains harboring chc1-5 exhibited a significant defect in transport of carboxypeptidase Y or carboxypeptidase S to the vacuole that was not observed in other chc1 ts mutants. The kinetics of invertase secretion or transport of alkaline phosphatase to the vacuole were not significantly affected in the chc1-5 mutant, further implicating clathrin specifically in the Golgi to vacuole transport pathway for carboxypeptidase Y. PROTEIN transport between distinct organelles in clathrin (Simpson et al. 1996; Panek et al. 1997) aleukaryotic cells is carried out by membrane-bound though a recent study reported a direct interaction bevesicles. Formation of transport vesicles involves assemtween AP-3 and the clathrin heavy chain in vitro (Dell’ bly of cytosolic coat proteins onto the donor membrane, Angelica et al. 1998). selective packaging of the cargo proteins and finally Small GTP-binding proteins are required to initiate budding of the coated vesicles. Several types of vesicle (or prime) assembly of the coats. As Sar1p functions to coats that mediate different protein transport steps have recruit COPII to the ER membrane, ADP-ribosylation been studied in detail (reviewed in Schekman and Orci factor (ARF) appears to be required on the Golgi mem1996). The Sec23/24p and Sec13/31p complexes form brane to recruit COPI or AP-1/clathrin coats (SchekCOPII, which coats vesicles that bud from the endoman and Orci 1996). The involvement of ARF in plasmic reticulum (ER). Coatomer, a heptameric (a,b,clathrin/AP-1 recruitment was first suggested by the b9,g,d,ε,z-COP) protein complex, coats COPI vesicles observation that AP-1 dissociated from Golgi memthat mediate transport from the Golgi to the ER and branes of cells treated with brefeldin A (Robinson and between Golgi compartments. Two types of coat proKreis 1992; Wong and Brodsky 1992), a drug that is teins contain clathrin, which is composed of heavy thought to specifically inhibit an ARF guanine nucleochains and light chains, associated with different adaptide exchange factor (Donaldson et al. 1992; Helms tor protein (AP) complexes. In mammalian cells, clathand Rothman 1992; Morinaga et al. 1996; Peyroche rin/AP-1 coated vesicles bud from the trans-Golgi netet al. 1996). In vitro, the binding of AP-1 to Golgi memwork and deliver cargo to endosomal compartments, branes and subsequent recruitment of clathrin is while clathrin/AP-2 drives endocytosis from the plasma dependent upon the GTPgS-activated form of ARF membrane. Recently, a third adaptor complex (AP-3) (Stamnes and Rothman 1993; Traub et al. 1993). Howhas been identified (Simpson et al. 1996), which in yeast ever, GTPgS-activated ARF can also mediate recruitappears to mediate the delivery of proteins from the ment of AP-2 and COPI to endosomal membranes (Seatrans-Golgi network directly to the vacuole (Cowles et al. man et al. 1993; Whitney et al. 1995), and COPI to 1997; Piper et al. 1997; Stepp et al. 1997). This AP-3 comER membranes (Bednarek et al. 1995) and can inhibit plex has been suggested to function independently of endosome-endosome fusion (Lenhard et al. 1992) and nuclear envelope reassembly (Boman et al. 1992). The in vivo significance of these observations is still not clear. Corresponding author: Todd R. Graham, Department of Molecular In addition, treatment of cells with brefeldin A causes Biology, Box 1820 Station B, Vanderbilt University, Nashville, TN 37235. E-mail: [email protected] the dissociation of many proteins peripherally associGenetics 150: 577–589 (October 1998) 578 C.-Y. Chen and T. R. Graham (Kranz and Holm 1990) with 6211 arf1D (Gaynor et al. 1998), ated with the Golgi (Kooy et al. 1992; Podos et al. 1994; respectively. The swa5-1 mutant (CCY 2017) was backcrossed Misumi et al. 1997) and it is not known if this represents three times with a wild-type yeast strain (SEY6210) to produce a direct requirement for ARF in Golgi binding or is an CCY620-5. Yeast cells were grown on yeast extract, peptone, indirect consequence of perturbing Golgi structure and and dextrose (YPD), synthetic minimal (SD) media supplemented as necessary (Sherman 1991). FOA (59-fluoroorotic function. acid; Sigma, St. Louis) media counterselective against growth In yeast Saccharomyces cerevisiae, ARF is encoded by two of Ura1 cells were prepared as described (Sikorski and Boeke genes, ARF1 and ARF2, which encode proteins with 96% 1991). Cells were grown overnight in liquid SD media conidentity that are probably redundant in function taining 0.2% yeast extract and required supplements for meta(Stearns et al. 1990a). Double arf1D arf2D mutants are bolic labeling experiments. Plasmids pCC8218 and pCC616 were generated as follows: inviable, indicating that ARF is an essential protein in a 1.8-kb EcoRI-PstI fragment carryingARF1 was subcloned from yeast. The Arf2 protein is only expressed at 10% of the pRB1297 (Stearns et al. 1990a) into the polylinker of pPolyIII level of Arf1 protein, and strains carrying a deletion of (Lathe et al. 1987) to produce pPolyIII-ARF1. An z1.8-kb the ARF2 gene show a wild-type phenotype (Stearns et NotI fragment from pPolyIII-ARF1 was inserted into pCH1153 al. 1990a,b). Strains harboring a deletion of ARF1 grow (Kranz and Holm 1990) to produce pCC8218 and an z1.8kb PstI-BamHI fragment from pPolyIII-ARF1 was subcloned well, yet exhibit modest defects in protein secretion and into pRS315 (Sikorski and Hieter 1989) to produce pCC616. modification (Stearns et al. 1990a), and, perhaps more To produce pCC416, a 7.7-kb SalI-NruI fragment carrying significantly, arf1D mutants exhibit a substantial alterCHC1 was subcloned from pCC1-2 (isolated from a genomic ation in the structure of Golgi and endosomal compartlibrary) into the SalI-SmaI sites of pRS416 (Sikorski and ments (Gaynor et al. 1998). This suggests that ARF plays Hieter 1989). The isogenic chc1-ts strains were prepared by targeted intea role not only in protein transport but also in organelle gration of 59 truncated chc1 alleles into the CHC1 locus (demembrane dynamics (Gaynor et al. 1998). Strains harpicted in Figure 7B). YIpchc521DCla (Tan et al. 1993) was boring the arf1D null mutation combined with mutaused to prepare the 6210 chc1-521 strain. To prepare pCC306 tions in RET1, SEC21, and SEC27, which respectively chc1-5, a 7.4-kb SalI-NotI fragment carrying chc1-5 was subencode a-, g-, and b9-COP subunits of the coatomer cloned from pCC416-1 into pRS306 (Sikorski and Hieter complex, were found to display a synthetic growth de1989). A 7.6-kb AatII-SalI fragment carrying chc1-D57 was subcloned from pD57 (Lemmon et al. 1991) into the SmaI-SalI fect, whereas double mutants harboring arf1D comsites of pRS306 to produce pCC306 chc1-D57. The 59 ends of bined with several other sec mutations do not exhibit the chc1 genes were deleted by digesting the pCC306 plasmids synthetic growth defects (Stearns et al. 1990b; Gaynor with BglII and ClaI, blunt-ending and ligating to generate the et al. 1998). This provides in vivo evidence that ARF plays 39 chc1 series of plasmids. To prepare p 39chc1-D43, two PCR a specific role in coatomer function; however, there is primers were designed to delete the C-terminal 43 amino acids of Chc1p: one is upstream of the last BamHI site in the CHC1 no such genetic evidence yet for ARF in the assembly open reading frame (59-ACAAATTTGACCAATTGGGATTGof AP-1/clathrin coats. 39); the other introduces two stop codons and a SalI site after To identify other proteins that functionally interact codon 1610 (59-CGTCGAGTCGACTCATTATTTTTTTATGG with Arf1p, we have employed a genetic screen to search AGATTTCAAATGGC-39). After PCR amplification, the prodfor the mutations that display synthetic lethality in comucts were digested with SalI and BamHI and subcloned into the SalI-BamHI sites of p 39chc1-5. To generate the 6210 chc1 bination with the arf1D null mutation and have identimutants, the p 39chc1 plasmids were linearized with AftII and fied seven complementation groups (SWA1-7 for syntransformed into SEY6210 to target integration into the chrothetically lethal with arf1D). In this study, we focused mosomal CHC1 locus. Ura1 transformants were tested for ts on characterization of SWA5, which was found to be growth. Total DNA was then isolated from ts transformants allelic to the clathrin heavy chain gene (CHC1). This and the correct integration event was confirmed by PCR and genetic interaction between arf1D and chc1 provides in DNA sequencing. Isolation and characterization of swa mutants: To start the vivo support for a role for ARF in clathrin coat assembly. screen, CCY2011 and CCY2804 were grown at 308 in liquid Surprisingly, chc1-5 (swa5-1) mutants exhibited markSD medium lacking uracil to stationary phase and treated with edly slow transport kinetics for carboxypeptidase Y 3% ethyl methanesulfonate (Fluka, St. Louis) for 30 min at (CPY) that was notobserved in three other chc1 tempera308 (Lawrence 1991), which resulted in z60–80% survival. ture-sensitive (ts) mutants. The effect of chc1-5 on CPY Cells were plated on YPD to yield z500 colonies per plate and incubated at 308 for 5–6 days. Uniformly red colonies were transport was much more substantial than the effect streaked twice onto YPD plates to confirm the nonsectoring on invertase or alkaline phosphatase transport, further phenotype and finally onto FOA medium to confirm plasmidimplicating clathrin in transport of CPY from the Golgi dependent growth. Mutant strains were crossed with the pato the vacuole. rental strain of opposite mating type and an ARF1 ARF2 ade2 ade3 strain, and the resulting diploids were replica-plated onto FOA medium to determine if mutations were recessive and if the nonsectoring phenotype resulted from plasmid integraMATERIALS AND METHODS tion, respectively. All of the diploids were FOA indicating that the mutations were recessive and that none of the mutants Media, strains, and plasmid construction: Yeast strains and plasmids used in this study are listed in Table 1. CCY2011C recombined the URA3 and ADE3 genes into a chromosome. For complementation analysis, mutants of opposite mating and CCY2804C are meiotic progeny from crosses of CH1305 (Kranz and Holm 1990) with 6210 arf1D and of CH1304 type were intercrossed and the resulting diploids were replica579 arf1 Synthetic Lethality With chc1