Mammalian Sly1 regulates syntaxin 5 function in endoplasmic reticulum to Golgi transport.

Members of the syntaxin gene family are components of protein complexes which regulate vesicle docking and/or fusion during transport of cargo through the secretory pathway of eukaryotic cells. We have previously demonstrated that syntaxin 5 is specifically required for endoplasmic reticulum to Golgi transport (Dascher, C., Matteson, J., and Balch, W. E. (1994) J. Biol. Chem. 269, 29363-29366). To extend these observations we have now cloned a protein from rat liver membranes which forms a native complex with syntaxin 5. We demonstrate that this protein is the mammalian homologue to yeast Sly1p, previously identified as a protein which genetically and biochemically interacts with the small GTPase Ypt1p and Sed5p, proteins involved in docking/fusion in the early secretory pathway of yeast. Using transient expression we find that overexpression of rat liver Sly1 (rSly1) can neutralize the dominant negative effects of excess syntaxin 5 on endoplasmic reticulum to Golgi transport. These results suggest that rSly1 functions to positively regulate syntaxin 5 function.

Members of the syntaxin gene family are components of protein complexes which regulate vesicle docking and/or fusion during transport of cargo through the secretory pathway of eukaryotic cells. We have previously demonstrated that syntaxin 5 is specifically required for endoplasmic reticulum to Golgi transport (Dascher, C., Matteson, J., and Balch, W. E. (1994) J. Biol. Chem. 269, 29363-29366). To extend these observations we have now cloned a protein from rat liver membranes which forms a native complex with syntaxin 5. We demonstrate that this protein is the mammalian homologue to yeast Sly1p, previously identified as a protein which genetically and biochemically interacts with the small GTPase Ypt1p and Sed5p, proteins involved in docking/fusion in the early secretory pathway of yeast. Using transient expression we find that overexpression of rat liver Sly1 (rSly1) can neutralize the dominant negative effects of excess syntaxin 5 on endoplasmic reticulum to Golgi transport. These results suggest that rSly1 functions to positively regulate syntaxin 5 function.
Transport through the exocytic pathway is mediated by vesicular carriers (1). While a number of coat proteins, including clathrin, COPII, and COPI, have been shown to mediate vesicle budding from different cellular compartments (2)(3)(4), the biochemical machinery essential to promote vesicle docking and fusion is less well understood. One of the proteins participating in these events is the yeast protein Sly1p which was identified as a single copy suppressor of the loss of YPT1 function in yeast when carrying a point mutation (5). Sly1p is a member of a family of related proteins which includes the neuronal specific forms of Munc-18/n-Sec1/rb-Sec1 (6 -9) as well as ubiquitously expressed Munc-18 isoforms in mammalian cells (10,11), Sec1p involved in vesicle delivery to the bud site in yeast (12,13) and homologues in Drosophila (Rop) (14) and Caenorhabditis elegans (Unc18) (15) (for a recent review see Ref. 16). Yeast Sly1p and mammalian Munc-18/n-Sec1/rb-Sec1 have been shown to form complexes with different members of the syntaxin gene family including yeast Sed5 (17,18) and the mammalian syntaxins 1-4 homologues (8 -10), respectively, and in the case of Sly1p, the small GTPase Ypt1 (5,(17)(18)(19)(20)(21)(22)(23)(24)(25).
We have previously demonstrated that the Sed5 mammalian homologue, syntaxin 5, is essential for ER 1 to Golgi transport (21). To extend our analysis of syntaxin 5 function in transport in mammalian cells, we have immunoisolated a protein from rat liver membranes which forms a complex with syntaxin 5. Cloning of the protein based on peptide sequences led to the identification of a mammalian homologue to yeast Sly1p. We provide functional evidence that this protein is required for ER to Golgi transport in vivo.
Peptide Sequencing-1.2 g of rat liver crude microsomes were used for immunoprecipitations. SDS-PAGE resolved a 42-kDa band (p42), corresponding to syntaxin 5 antigen, and a 66-kDa co-immunoprecipitating polypeptide band (p66) were excised and digested with trypsin. Tryptic peptides were separated, and selected peptides were subjected to chemical microsequencing. One peptide of the 42-kDa band (PVSAL-PLAPNNLGG) perfectly matched the sequence of syntaxin 5 (amino acids 161-174) confirming the identity of the syntaxin 5 antigen. Four tryptic peptide sequences were obtained from the 66-kDa band: (I) MLNFNVPHVKNS, (II) VPAVYFVPTEE, (III) GTAAEMVAVK, and (IV) SNPETDDY. A search using the Blast WWW Server of the National Center for Biotechnology Information (NCBI) of GenBank TM and EMBL data bases (25) with peptide (II) identified p66 as the potential rat homologue of the Sly1p of Saccharomyces cerevisiae (5) and its homologue in C. elegans (26); GenBank TM accession number Z35640.
cDNA Cloning and Sequencing-Data bases of expressed sequence tags (EST) and nonredundant PDB, GenBank TM and EMBL sequences were searched using the Blast program provided by NCBI. Two partially sequenced human cDNAs, a human fetal heart EST cDNA (Gen-* This work was supported in part by Grant GM 42336 from the National Institutes of Health (to W. E. B.), by Shared Instrumentation Grants RRO7273 and RR08176, and the Lucille P. Markey Charitable Trust. This is TSRI Manuscript Number 10002-CB. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) U57687. Bank TM accession number R57959) and a human HepG2 3Ј region cDNA (GenBank TM accession number D16890), were found that encoded predicted amino acid sequences 90% identical with peptides I and II of p66. The deduced polypeptides showed 38 to 60% identity with the yeast and C. elegans Sly1 proteins over 73 to 88 amino acid overlaps. For the amplification of rat Sly1 (rSly1) sequences by the polymerase chain reaction (PCR), two oligonucleotide primers were synthesized according to the human cDNA sequences. Oligonucleotide P1 was generated corresponding to nucleotides 62-93 (5Ј-CGTATGTTGAATT-TCAATGTGCCTCATATTAA-3Ј; sense primer) of the human fetal heart EST cDNA showing 90% identity with peptide I (see Fig. 2) of p66. Oligonucleotide P2 was synthesized according to nucleotides 76 -114 (5Ј-AAGATTCTGATATTCAATGTAGTTGCCTCCTCCCACCAC-3Ј; antisense primer) of the human HepG2 3Ј region cDNA. Rat first strand cDNA was generated from total RNA of NRK cells using the Superscript Pre-amplification System for first strand cDNA synthesis (Life Technologies, Inc.). A 1.8-kilobase fragment was amplified using the NRK cDNA as a template and the Expand Long Template PCR system (Boehringer Mannheim) with the set of primers described above. The PCR product was subcloned into the TA cloning vector pCRII (Invitrogen) and sequenced by the chain termination method (27). Based on the partial rSly1 cDNA sequence information, a second set of oligonucleotide primers was designed to be used in the amplification of the missing 5Ј-and 3Ј-cDNA ends (5Ј-and 3Ј-RACE): 3Ј-RACE primer P3, 5Ј-CGATCTCCATACAAATGTCGCCACTGCTG-3Ј; 5Ј RACE antisense primer P4, 5Ј-AGGGTCAGATATGACGTCGAGAAGGGAC-3Ј. RACE reactions were performed using rat liver poly(A) ϩ RNA (Clontech) and the Marathon cDNA Amplification kit (Clontech). PCR products were subcloned into pCRII, and sequenced. In total, 6 clones were fully sequenced. The first methionine of the clone with the longest open reading frame is immediately preceded by a stop codon and a Kozak (28) consensus sequence making it likely to be the initiation codon. Sequence alignments were performed with the multiple alignment program Clustal W 1.5 (29) using the default parameters.
Generation of Expression Constructs and Transient Expression-Nterminal Myc-epitope tagged expression constructs of syntaxin 5, syntaxin 5-11 and rSly1, which allowed expression from the T7 promoter in transient expression experiments (30), were prepared as described (30).

RESULTS AND DISCUSSION
To identify proteins which interact with syntaxin 5, we prepared a crude microsomal fraction containing smooth and rough microsomes as well as Golgi membranes as a source for potential syntaxin 5 complexes. Detergent extracts of crude membranes partially purified using gel filtration (see "Experimental Procedures") were incubated with protein A beads covalently coupled to affinity-purified antibodies to syntaxin 5. The beads were washed extensively with increasing concentrations of salt prior to elution at low pH. Fig. 1 presents a silver-stained gel of eluates from beads attached to total IgG prepared from preimmune serum (Fig. 1b) and beads coupled to two different affinity-purified antibodies specific for syntaxin 5 (Fig. 1, c and d). Western blot analysis of extracts with a syntaxin 5 antibody showed that the beads bound more than 90% of the detergent-extractable syntaxin 5 (data not shown) which migrated as a 42-kDa protein (Fig. 1, c and d). In addition to syntaxin 5, a second prominent polypeptide of 66 kDa was routinely observed to co-immunoprecipitate with syntaxin 5 (Fig. 1, c and d). p66 was not precipitated with either preimmune serum IgGs coupled to beads (Fig. 1a) or when protein A-Sepharose beads alone were incubated with the detergent extracts (data not shown).
To characterize the two proteins recovered from the syntaxin 5-specific antibody-coated beads, bands were excised from Coomassie-stained gels and digested with trypsin, separated by high pressure liquid chromatography, and microsequenced. The sequence of one peptide derived from the 42-kDa band confirmed the identity of this protein as syntaxin 5. Sequences from 3 major tryptic peptides (I, II, and III in Fig. 2) derived from p66 showed between 50 and 90% identities with the Sly1p of S. cerevisiae (5) and C. elegans (26). These results suggest that p66 is the potential rat homologue of yeast Sly1p.
The tryptic peptides were used to search data bases of EST cDNAs and nonredundant sequences (see "Experimental Procedures"). Two potential human homologues were found. The deduced amino acid sequences were between 35 and 60% identical with the yeast and C. elegans Sly1 protein. Based on the sequence of the two human cDNAs, oligonucleotide primers were designed for the amplification of the rSly1 sequence using NRK cDNA as a template (see "Experimental Procedures"). The 5Ј and 3Ј cDNA ends missing in the original clone were amplified using rat liver cDNA as a template to yield the complete clone. The deduced sequence for rSly1 is shown in Fig.  2. The rSly1 open reading frame predicts a ϳ72-kDa hydrophilic protein, which is in agreement with the size of the precipitated polypeptide of 66 kDa on SDS-PAGE. Rsly1 displays 30% identity with S. cerevisiae Sly1p and 42% identity with the Sly1 homologue of C. elegans when optimally aligned (Fig. 3). Identities were distributed over the entire length of the protein with divergent sequences evident principally in the amino-and carboxyl-terminal portions of the three proteins (Fig. 3). The sequence shows only weak (ϳ18%) identity with Sec1-related proteins functioning at the plasma membrane, emphasizing a potential role for rSly1 in ER to Golgi transport. This is consistent with previous studies in yeast which have established a strong interaction between the syntaxin 5 homologue Sed5p and Sly1p in ER to Golgi transport (17).
To explore the potential role of rSly1 in syntaxin 5 function, we took advantage of a transient expression system which we have used previously to examine the function of syntaxin 5 and the small GTPases Rab1, ARF1, and Sar1 in vesicle-mediated transport from the ER to the Golgi apparatus in mammalian cells in vivo (21, 30 -32). For this purpose, we used a recombinant T7 vaccinia virus system to transiently co-express the specific cDNA of interest with the vesicular stomatitis virus glycoprotein (VSV-G). VSV-G is a type I transmembrane protein containing two N-linked carbohydrate chains. Vectorial transport of VSV-G from the ER to and through sequential cis-, medial-, and trans-Golgi compartments can be measured by the processing of its two oligosaccharide chains from the high mannose (Man 9 ) endoglycosidase H (endo H)-sensitive form found in the ER and pre-Golgi intermediates (R S in Fig. 4) to endo H-resistant forms found in the Golgi stack (R 1 and R T in Fig. 4). The transient endo H-resistant (R 1 ) form (Fig. 4) corresponds to the transport of VSV-G to the early cis/medial-Golgi compartment where one or both of the oligosaccharide chains becomes processed by the action of resident ␣-1,2-mannosidases and glycosyltransferases. Subsequent transport of VSV-G to trans-Golgi compartments results in the appearance of the fully processed R T form containing two complex, sialic acid-containing endo H-resistant oligosaccharides (Fig. 4). The sequential FIG. 1. Identification of a p66-rSly1 complex with syntaxin 5. Sepharose beads coupled to total preimmune IgG (lane b) or two different affinity-purified syntaxin antibodies (0612 and 0613) (lanes c and d,  respectively) were incubated overnight with a detergent extract of rat liver microsomes as described under "Experimental Procedures." The bound protein was released as described (17), the eluate was precipitated by trichloroacetic acid, separated on SDS-PAGE, and silverstained. p42/syntaxin 5 (Syn5) and p66/rSly1 are indicated by arrows. Molecular mass marker proteins are shown in lane a. Asterisks mark heavy and light chains of IgGs.

Sly1 Function in ER to Golgi Transport
appearance of each of these processing intermediates allows us to assess the differential requirement for factors regulating ER to Golgi and/or intra-Golgi transport.
BHK-21 cells infected with the recombinant vaccinia virus were co-transfected with VSV-G and various combinations of expression vectors carrying the Myc-tagged rSly1, syntaxin 5 or syntaxin 5-11, a truncated, cytosolic form of syntaxin 5 lacking the transmembrane domain. After incubation for 3-6 h to allow time for protein expression, transfected cells were incubated with [ S 35]methionine for 10 min to label VSV-G in the ER, followed by a chase in the presence of unlabeled methionine for 60 min to promote the migration of VSV-G to the Golgi where it becomes processed to endo H-resistant R 1 and R T forms. Expression of recombinant Myc-rSly1 or Myc-syntaxin 5 was followed by Western blot analysis using the anti-Myc monoclonal antibody or a specific syntaxin 5 antibody, respectively. Overexpression of Myc-rSly1 had no effect on the processing of VSV-G to the R T form (Fig. 4A), suggesting that elevated levels of the protein do not have a dominant negative effect on carrier vesicle function throughout the early secretory pathway. In contrast, expression of the Myc-tagged full-length syntaxin 5 strongly inhibited ER to Golgi transport, but not intra-Golgi transport as indicated by the efficient processing of VSV-G to the R T form (Fig. 4B). This result is in agreement with previous studies where hemagglutinin (HA)-tagged syntaxin 5 was found to potently inhibit ER to Golgi transport (21).
The fact that yeast Sly1p shows biochemical interactions with Sed5p (17) and our ability to detect a prominent syntaxin 5-rSly1 complex in rat liver membranes (Fig. 1) led us to examine if rSly1 will interact with syntaxin 5 to suppress the dominant negative effects of the protein during overexpression. Co-expression revealed that Myc-rSly1 and syntaxin 5 (or the soluble, truncated syntaxin 5-11) could be co-immunoprecipitated from lysates using either anti-Myc or anti-syntaxin 5 antibodies (data not shown), demonstrating that the identified clone expressed a protein capable of forming the specific complex observed in vivo. Furthermore, as shown in Fig. 4, coexpression of Myc-rSly1 with a level of syntaxin 5 sufficient to block transport (Fig. 4B), significantly suppressed inhibition (Fig. 4C). In this case, the level of VSV-G processing to the R T form was at least 2-fold higher than that observed in the absence of rSly1 (compare the value of R T in B at 5 h (16%) to the value of R T in C at 6 h (35%)). Even more potent suppression was observed when the truncated form of syntaxin 5, syntaxin 5-11, was co-expressed with rSly1. As shown in Fig.  4D, and in agreement with previous results (21), this cytosolic form of syntaxin interferes with transport between the ER and the Golgi as well as between compartments of the Golgi stack. This more general effect is characterized by the accumulation of VSV-G in both the R 1 and R T forms (Fig. 4D). In the presence of rSly1, the partial inhibition of transport by syntaxin 5-11 between all compartments was completely suppressed, with VSV-G accumulating exclusively in the R T form (compare R 1 and R T in D at 5 h to R 1 and R T at 6 h in E). These results suggest that rSly1 is likely to interact with syntaxin 5 through its cytosolic domain in vivo.
In this study, we have purified to homogeneity, cloned, and sequenced the rat homologue to yeast Sly1p. This protein, originally isolated as a suppressor to the loss of Ypt1p function (5), is recognized to play an essential, but as yet undefined role in ER to Golgi transport in yeast (5,20). Our biochemical data argue strongly for a direct role of rSly1 in syntaxin 5 function in mammalian cells. Not only was the protein purified as a native complex with syntaxin 5 in detergent extracts of rat liver membranes enriched in ER and Golgi membranes, but overexpression of rSly1 significantly neutralized the inhibitory effects of excess wild-type or truncated syntaxin 5, a protein previously demonstrated to be specific for ER to Golgi transport (21). This is consistent with the biochemical interactions observed between Sly1p and Sed5p in yeast (17). We also noted that overexpression of rSly1 by itself had no dominant negative effect on ER to Golgi transport. This result is in contrast to the ability of the Drosophila Rop protein, a Munc-18/n-Sec1/rb-Sec1 homologue, to cause a decrease in neurotransmitter release (14). Our results suggest that either the two members of the gene family regulate syntaxin function differently, or that the mechanisms involved in the assembly and disassembly of the SNARE complex in constitutive versus regulated secretory pathways are distinct.
The mechanism by which rSly1 functions in vesicle docking and/or fusion remains to be elucidated. One possible model consistent with our results is that rSly1 positively regulates syntaxin 5 function. rSly1 binding to syntaxin 5 would be required for establishing the correct interactions promoting the assembly of a docking-fusion complex. In this model, overexpression of syntaxin 5 would be expected to deplete the endogenous pool of rSly1, a condition which would then allow uncomplexed syntaxin 5 to impede the normal assembly of protein complexes involved in vesicle docking and/or fusion. Overexpression of rSly1 would be predicted to correct this deficiency. Given that function of the rSly1 in complex with syntaxin 5 may likely to be regulated by Rab1 (5,20), we now are poised to explore this requirement with purified proteins.