Portal de Periódicos da CAPES

Two syntaxin homologues in the TGN/endosomal system of yeast

1998; Springer Nature; Volume: 17; Issue: 1 Linguagem: Inglês

10.1093/emboj/17.1.113

1460-2075

Joost C. M. Holthuis,

Plant Reproductive Biology

Article1 January 1998free access Two syntaxin homologues in the TGN/endosomal system of yeast Joost C.M. Holthuis Joost C.M. Holthuis MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Benjamin J. Nichols Benjamin J. Nichols MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Sadhana Dhruvakumar Sadhana Dhruvakumar MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Hugh R.B. Pelham Corresponding Author Hugh R.B. Pelham MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Joost C.M. Holthuis Joost C.M. Holthuis MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Benjamin J. Nichols Benjamin J. Nichols MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Sadhana Dhruvakumar Sadhana Dhruvakumar MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Hugh R.B. Pelham Corresponding Author Hugh R.B. Pelham MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB Search for more papers by this author Author Information Joost C.M. Holthuis1, Benjamin J. Nichols1, Sadhana Dhruvakumar1 and Hugh R.B. Pelham 1 1MRC Laboratory of Molecular Biology, Hills Road, Cambridge, CB2 2QH GB ‡Joost Holthuis and Ben Nichols contributed equally to this work *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:113-126https://doi.org/10.1093/emboj/17.1.113 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Intracellular membrane traffic is thought to be regulated in part by SNAREs, integral membrane proteins on transport vesicles (v-SNAREs) and target organelles (t-SNAREs) that bind to each other and mediate bilayer fusion. All known SNARE-mediated fusion events involve a member of the syntaxin family of t-SNAREs. Sequence comparisons identify eight such proteins encoded in the yeast genome, of which six have been characterized. We describe here the remaining two, Tlg1p and Tlg2p. These have the expected biochemical properties of t-SNAREs, and are located in separable compartments which correspond to a putative early endosome and the yeast equivalent of the TGN, respectively. They co-precipitate with the v-SNARE Vti1p, which is implicated in Golgi–endosome traffic and, remarkably, binds to five different syntaxins. Tlg1p also binds the plasma membrane v-SNARE Snc1p. Both Tlg1p and Tlg2p are required for efficient endocytosis and to maintain normal levels of TGN proteins. However, neither is required for intra-Golgi traffic. Since no further syntaxins have been identified in yeast, this implies that the Golgi apparatus can function with a single syntaxin, Sed5p. Introduction The endocytic and secretory pathways of eucaryotic cells consist of multiple membrane-bound compartments which communicate by means of transport vesicles or by direct fusion. There has been much interest in determining how these compartments are formed and maintained, and how proteins pass between them. Analysis has been complicated by the morphological differences between cell types and species, and by the variety of traffic routes that have been detected or proposed. Central to the secretory pathway is the Golgi apparatus, which consists of a variable number of cisternae through which secreted proteins pass in an ordered fashion (Farquhar, 1985). At the exit point from this organelle is a structure known in animal cells as the trans-Golgi network (TGN), which is thought to be the first point at which the biosynthetic pathway is connected to the endocytic pathway. Progress in unravelling the membrane fusion events in cells has been greatly stimulated by the identification of SNARE proteins. These are (mostly) integral membrane proteins found on vesicles and organelles which interact by the formation of helix bundles; these interactions are a necessary prelude to most known fusion events, and are thought to contribute significantly to the specificity of membrane recognition (Söllner et al., 1993a,b; reviewed by Rothman, 1994; Bennett, 1995; Rothman and Wieland, 1996). The interactions are controlled in part by soluble proteins, NSF and SNAPs, which act at multiple steps in the secretory pathway. SNAREs have been divided conceptually into those associated primarily with vesicles (v-SNAREs) and those on target membranes (t-SNAREs). This is a convenient distinction but not an absolute one—the neuronal t-SNARE syntaxin 1 is found not only on the plasma membrane but also on synaptic vesicles (Walch-Solimena et al., 1995), and in the case of homotypic vacuole fusion in yeast, t- and v-SNAREs operate in equivalent membranes (Nichols et al., 1997). SNAREs also fall into families related in structure and sequence. Of these, the most widespread and well-defined is the syntaxin family of t-SNAREs (Weimbs et al., 1997); all known SNARE-dependent fusion steps so far examined involve a member of this family. A full understanding of the number, distribution and interactions of SNAREs should provide considerable insight into the organization of the intracellular membrane system, and the traffic routes between organelles. An issue of particular interest is the mode of transport through the Golgi apparatus. Two main types of model have been proposed (reviewed by Farquhar, 1985). In the vesicular transport model, cisternae with a stable existence communicate by means of vesicular traffic (Rothman and Wieland, 1996). This is likely to be bidirectional, to allow recycling of proteins such as v-SNAREs. In the cisternal maturation model, based largely on electron microscopic studies, cisternae form de novo by fusion of ER-derived vesicles, mature while passing through the Golgi stack, and disintegrate into vesicles again at the level of the TGN (see Farquhar, 1985). This model requires resident Golgi enzymes to be continually recycled to earlier compartments by retrograde vesicular traffic (Bannykh and Balch, 1997; Mironov et al., 1997; Pelham, 1998). Both models predict a requirement for SNAREs, and indeed it is well established that NSF and SNAP (Sec18p and Sec17p in yeast) are required for intra-Golgi transport as measured in mammalian cell-free systems and in vivo in yeast (Graham and Emr, 1991; Rothman, 1994; Rothman and Wieland, 1996). However, the two models predict somewhat different roles for the SNAREs. A full knowledge of Golgi SNAREs should give clues to the number of different compartments and transport steps, provide constraints for the models, and perhaps distinguish between them (Schekman and Mellman, 1997). The completion of the yeast genome sequence offers a chance to define all the SNAREs in a eucaryotic cell and to attempt to fit them into such models. We have focused on the syntaxin family of t-SNAREs, because these proteins are the easiest to identify by sequence homology. We and others, using a variety of algorithms, have been able to detect just eight putative syntaxins in yeast (Weimbs et al., 1997). Six of these have already been described: Ufe1p in the ER, Sso1p and Sso2p on the plasma membrane, Vam3p on the vacuole, Pep12p in an endosomal compartment, and Sed5p in the earliest part of the Golgi apparatus (Hardwick and Pelham, 1992; Aalto et al., 1993; Becherer et al., 1996; Lewis and Pelham, 1996; Nichols et al., 1997; Wada et al., 1997). In this paper we characterize the remaining two syntaxins, Tlg1p and Tlg2p (t-SNAREs affecting a late Golgi compartment). We show that both are required for efficient endocytosis and for the maintenance of resident proteins in the TGN. The TGN does not have a morphological definition in yeast, but is defined as the compartment where proteins destined for the cell surface are sorted from those destined for delivery to the vacuole, and contains processing proteases such as Kex2p and DPAP A (Ste13p). All known TGN proteins in yeast are integral membrane proteins which are thought to cycle continuously through endosomes (Stack et al., 1995; Cooper and Stevens, 1996; Bryant and Stevens, 1997). Tlg2p co-localizes with TGN markers, while Tlg1p appears to be in an early endosomal compartment through which TGN proteins cycle. Both the Tlg proteins interact with v-SNAREs implicated in Golgi–endosome and Golgi–plasma membrane traffic. However, they do not bind the putative intra-Golgi v-SNAREs Sft1p and Gos1p, and neither protein is essential for transport through the Golgi apparatus. These results indicate that Tlg1p and Tlg2p mediate interactions between the exocytic and endocytic pathways, and also imply that traffic through the Golgi apparatus can be sustained with a single Golgi syntaxin, Sed5p, a result which seems more compatible with a cisternal maturation model than with the standard vesicular model for anterograde intra-Golgi transport. Results Identification of TLG1 and TLG2 The most conserved domain of the syntaxins is the ∼60-residue coiled-coil motif adjacent to the C-terminal membrane anchor. BLAST searches with this portion of known yeast syntaxin family members identified two new yeast genes, corresponding to ORFs YDR468c and YOL018c, which we have named TLG1 and TLG2 (Figure 1). Tlg2p shows strongest homology to the endosomal and vacuolar syntaxins Pep12p and Vam3p, although it is unusual in having a substantial extracytoplasmic domain of 62 residues. Tlg1p is one of the most divergent members of the family, but retains the overall structure characteristic of the syntaxins. A more sophisticated profile search for syntaxin homologues found the same two proteins (Weimbs et al., 1997). The only other related proteins found in the yeast genome were the putative SNAREs Bet1p, Sft1p, Vam7p, Sec9p, and a protein related to Sec9p. Though showing some sequence similarity to the syntaxins these are clearly more divergent, and none of them shares the domain structure of the syntaxins, which in addition to the conserved region includes two other coiled-coil motifs and a C-terminal transmembrane domain (Weimbs et al., 1997). Thus, Tlg1p and Tlg2p are the only syntaxins in yeast that have not previously been characterized. Figure 1.Sequence similarities between yeast syntaxin family members. Only the C-terminal coiled-coil region is shown; the heptad repeats of predominantly hydrophobic residues are indicated by the black dots. Residues that are identical or similar in at least half the sequences are shown on a black or grey background respectively. Download figure Download PowerPoint Location of Tlg1p and Tlg2p As a preliminary guide to their intracellular location, we expressed epitope-tagged versions of Tlg1p and Tlg2p from centromere vectors and visualized them by immunofluorescence. Both proteins revealed a punctate staining pattern, which is characteristic of the yeast Golgi apparatus, and examples of stained cells are shown in Figure 2A and B. Since no syntaxin had previously been assigned to the yeast equivalent of the TGN, we co-stained cells for Kex2p, a TGN marker (Stack et al., 1995). Endogenous levels of Kex2p are difficult to detect (Redding et al., 1991), so we transformed the cells with a multi-copy vector containing the KEX2 gene. Moderate overexpression has previously been shown not to alter the qualitative pattern of Kex2p immunofluorescence, nor to prevent its co-localization with the Golgi marker Sec7p (Franzusoff et al., 1991; Redding et al., 1991). As illustrated in Figure 2A, some Tlg1p-positive structures seemed to contain Kex2p, but in general the staining patterns were not very similar. There was a much more striking similarity between the Tlg2p and Kex2p patterns, although close examination showed that there were also some differences; an example is shown in Figure 2B. These results suggested that Tlg2p might be associated with the TGN, with Tlg1p having a somewhat different distribution. Figure 2.Immunofluorescent localization of Tlg1p and Tlg2p. In (A) and (B), cells expressing myc-tagged Tlg1p or Tlg2p from a CEN vector, and Kex2p from a multicopy vector, were double-labelled with a monoclonal antibody to the myc epitope (9E10) and a polyclonal to Kex2p. (A) shows two cells, (B) a single cell. In (C), HA-tagged Tlg2p expressed from the endogenous gene was labelled with a monoclonal antibody, and endogenous Sed5p was detected with a specific antiserum. Five cells and their buds are visible, at slightly lower magnification than (A) and (B). Cells were imaged in one plane by confocal microscopy. The colour image was produced by combining high-contrast images of the fluorescent signals, to allow their positions to be compared. The precision of this comparison is limited by differences in relative intensity, and slight differences in depth of focus between the channels. Download figure Download PowerPoint We also compared Tlg2p with the early Golgi t-SNARE, Sed5p. In this case, we were able to detect the endogenous Sed5p, and used a strain in which sequences encoding a C-terminal influenza epitope tag had been introduced into the chromosomal TLG2 gene, thus ensuring normal levels of expression of Tlg2p. Both proteins gave the expected punctate staining pattern, but they were clearly present in different structures (Figure 2C). To assess by an independent method the locations of endogenous Tlg1p and Tlg2p, we prepared specific antisera and used them to follow the distribution of the proteins during subcellular fractionation. As shown in Figure 3A, each antiserum detected a single major band on immunoblots, and this band was absent in strains from which the corresponding genes had been deleted (see below). Figure 3B shows an example of the fractionation of wild-type cells on an equilibrium sucrose density gradient, which was adjusted to give optimal resolution of the Tlg-containing membranes. All the data shown are from a single set of gradient fractions, but for clarity the results are presented in four panels, each containing the profile of Tlg2p. Figure 3.Subcellular fractionation of Tlg1p and Tlg2p. (A) Antibody detection of Tlg1p, Tlg2p and Pep12p. Equal amounts of total cell extracts prepared from parental (SEY6210) and appropriate deletion strains were resolved on an SDS–polyacrylamide gel, blotted and probed with polyclonal antibodies raised against recombinant Tlg1p, Tlg2p and Pep12p. The positions of size markers (kDa) are indicated. (B) Sucrose gradient fractionation of membranes. A high-speed membrane pellet (100 000 g) from SEY6210 cells was fractionated on a sucrose density gradient as described in Materials and methods. Fractions were assayed for Kex2p enzyme activity and, by immunoblotting, for Tlg1p, Tlg2p, Pep12p, Sed5p, Gos1p and Vam3p. Quantitation was by densitometric scanning of fluorograms produced from the immunoblots. The same fractions were analysed in each panel, the profile of Tlg2p being shown in each to aid comparison. Download figure Download PowerPoint Figure 3B (top panel) shows that Tlg1p and Tlg2p were at least partially separated and peaked at different densities. However, the distributions overlapped and we reproducibly observed a shoulder on the Tlg2p profile in the position of the Tlg1p peak. Endogenous Kex2p activity was present in a major peak coincident with Tlg2p (fractions 9 and 10) together with a minor peak in the position of Tlg1p (fraction 13). Immunoblotting revealed that a second TGN marker, DPAP A (Nothwehr et al., 1993), coincided closely with the Tlg2p profile (Figure 3B, bottom panel). As expected from the immunofluorescence studies, Tlg2p did not co-fractionate with Sed5p (Figure 3B). We also examined the distribution of the Golgi v-SNARE Gos1p, which binds Sed5p. Gos1p has not previously been localized, but its mammalian homologue GOS28 is found in the medial/trans region of the Golgi apparatus (Nagahama et al., 1996), and in agreement with this the yeast protein peaked at a density midway between that of Sed5p and Tlg2p. This profile is consistent with Gos1p being in both the Sed5p and Tlg2p compartments, or in membranes of intermediate density. Tlg2p was separated well from the vacuolar t-SNARE Vam3p (Nichols et al., 1997; Wada et al., 1997), and at least partially from Pep12p, the t-SNARE for a late endosomal/prevacuolar compartment (Becherer et al., 1996) (Figure 3B). Since TGN proteins are thought to recycle through late endosomes, we sought to address the possibility that some of the Tlg2p was present in the Pep12p compartment. Pep12p is not essential for growth, but in its absence transport of hydrolases to the vacuole is severely impaired (Becherer et al., 1996), due to a failure of vesicles to fuse with the prevacuolar/late endosome compartment. One would therefore expect the removal of Pep12p to strongly affect the fractionation behaviour of this compartment. However, analysis of a pep12 deletion strain revealed a distribution of Kex2p, Tlg1p and Tlg2p indistinguishable from that in wild-type cells (Figure 4). Thus, it seems unlikely that Tlg2p is present in the Pep12p compartment under normal conditions, or that passage through this compartment is required for Kex2p or Tlg1p to achieve their usual steady-state distribution. Figure 4.Subcellular fractionation of Tlg1p and Tlg2p in Δpep12 cells. A high-speed membrane pellet (100 000 g) from Δpep12 cells was fractionated on a sucrose density gradient identical to that shown in Figure 3. Fractions were assayed for Kex2p enzyme activity and, by immunoblotting, for Tlg1p and Tlg2p. Quantitation was by densitometric scanning of fluorograms produced from the immunoblots. Download figure Download PowerPoint The combination of subcellular fractionation and immunofluorescence strongly suggests that Tlg2p is predominantly associated with the TGN. In contrast, Tlg1p did not co-fractionate precisely with any of the markers tested. It was clearly separable from the bulk of the TGN, earlier Golgi markers, late endosomes and the vacuole, as well as the plasma membrane (which is found in a sharp peak at fraction 15 in these gradients). It did coincide with the minor Kex2p peak but, as suggested by the immunofluorescence data, much of the Kex2p was elsewhere. Hence, Tlg1p seems to be concentrated in a previously undefined compartment, which may be accessible to Kex2p. Phenotypes of tlg deletion mutants To investigate the functions of Tlg1p and Tlg2p, we prepared strains in which the corresponding open reading frames were completely removed. Neither gene proved essential. The tlg2 disruptants were superficially normal and grew only slightly slower than the parental strain, with a doubling time at 25°C of 2.0 versus 1.9 h, and at 37°C of 2.0 versus 1.6 h. The tlg1 disruptants grew well at 25°C (2.0 h doubling time), but could barely grow at 37°C. At semi-permissive temperatures they tended to produce large misshapen cells, and showed inefficient separation of daughters from mothers. Strains lacking both genes had a phenotype only slightly more exaggerated than that of the Δtlg1 mutants (2.2 h doubling time at 25°C). The viability of these mutant strains contrasts strongly with the lethal effects of removing the early Golgi t-SNARE, Sed5p, or the Golgi v-SNARE Sft1p (Hardwick and Pelham, 1992; Banfield et al., 1995). Electron microscopy of Δtlg1 and Δtlg1 Δtlg2 cells grown at 30°C confirmed that they were much larger than normal and frequently misshapen, with abnormal bud scars (Figure 5C and D). The vacuoles were also rather more fragmented than usual, a common phenotype of mutations affecting the secretory or endocytic pathways. In addition, the cells accumulated some darkly staining circular structures (indicated by arrowheads in Figure 5). These differed from the characteristic cup-shaped Golgi membranes seen in an sft1 mutant (Banfield et al., 1995) and their identity is unknown. Incubation of the cells at 37°C for 2 h before fixation did not change their EM phenotype substantially (data not shown). By comparison with Δtlg1 cells, the Δtlg2 mutant showed few abnormalities other than some vacuolar fragmentation and the presence of a few of the darkly staining structures (Figure 5B). Figure 5.Electron micrographs of parental SEY6210 (A), Δtlg2 (B), Δtlg1 (C) and Δtlg1 Δtlg2 (D) cells. Note the 100 nm exocytic vesicles in the bud tip in (A), abnormal membranous structures [rare in (B), more common in (C) and (D); arrowheads] and fragmented vacuoles (V) in the mutants. The abnormal membranes are heterogeneous, with some circular structures displaying a double membrane, reminiscent of Berkeley bodies, while other structures are smaller and uniformly densely stained, often with irregular outlines. Bar, 1 μm. Download figure Download PowerPoint To search for possible genetic interactions, we also prepared all possible double deletions of TLG1, TLG2 and genes encoding the other non-essential syntaxins in the endocytic pathway, namely PEP12 and VAM3. All pairwise combinations were viable except for the pep12 tlg1 double mutant. This strain could be sustained with a plasmid bearing one of the deleted genes and the URA3 marker, but fluoro-orotic acid (FOA) selection for loss of the plasmid yielded only extremely rare colonies, which presumably contain suppressor mutations (data not shown). It thus appears that the functions of Pep12p and Tlg1p are in some way related. Loss of TGN proteins Since Tlg2p, and perhaps to some extent Tlg1p, appear to reside in compartments containing Kex2p, we investigated the fate of Kex2p when they were removed. Strikingly, the level of Kex2p activity was reduced by 75% in Δtlg1 cells (Figure 6, left). Substantial reductions were also observed in Δtlg2 cells and in the double mutant. This loss was evidently due to degradation by vacuolar proteases, because additional deletion of the PEP4 gene restored the activity in the double mutant cells to normal (Figure 6, centre). Immunoblotting showed that DPAP A was also depleted in the tlg mutants, though less dramatically than Kex2p (Figure 6, right). Since degradation is the normal fate of Kex2p and DPAP A mutants that lack a specific retrieval signal (Wilcox et al., 1992; Nothwehr et al., 1993), we infer that the tlg mutants have a defect in the retrieval pathway to the TGN, and that TGN proteins are consequently lost to the vacuole. Figure 6.Loss of TGN proteins in tlg deletion strains. Kex2p activity and DPAP A protein levels in total cell lysates of the deletion strains are expressed as a percentage of those measured in the appropriate parental strains (SEY6210 or Δpep4). Deletion of PEP4 had a negligible effect on Kex2p activity. The data shown are means (± standard deviation) of at least four independent experiments Download figure Download PowerPoint Endocytosis defects Loss of Kex2p is a phenotype common to a subset of mutants that have endocytosis defects (Luo and Chang, 1997). We therefore examined endocytosis in the tlg mutants. A ready test is provided by the internalization and subsequent degradation in the vacuole of the alpha-factor receptor, a process that is triggered by the addition of the alpha-factor pheromone (Schandel and Jenness, 1994; Hicke and Riezman, 1996). As shown in Figure 7, the receptor is normally degraded rapidly. Cells lacking either Tlg1p or Tlg2p showed significantly reduced kinetics of receptor degradation, the effect being greatest in Δtlg1 cells. Hence, efficient delivery of the receptor to the vacuole requires the action of both Tlg2p and Tlg1p. Figure 7.Impairment of ligand-induced degradation of α-factor receptor in tlg deletion strains. (A) Immunoblot of Ste2p (α-factor receptor) from extracts of parental (SEY6211) and Δtlg1 Δtlg2 cells prepared before (0 min) or after exposure to α-factor. Ste2p migrates as a doublet of ∼47 kDa. (B) Ste2p levels in parental cells (SEY6211) and tlg mutants before and after exposure to α-factor. Ste2p levels in cell extracts were quantified by immunoblotting and densitometric scanning of fluorographs, and expressed as % of receptor present before α-factor exposure (0 min). The results shown are the average of three independent experiments. Download figure Download PowerPoint Secretory pathway function The TGN receives traffic both from the earlier part of the Golgi apparatus and from the endocytic pathway, and a t-SNARE present in the TGN could be involved in both processes. However, the viability of the tlg mutants, and the lack of any clear accumulation of Golgi membranes in EM images suggests that neither Tlg1p nor Tlg2p is essential for secretion. As a more sensitive test, we examined the maturation of the vacuolar protease carboxypeptidase Y (CPY). CPY appears as a p1 precursor in the ER, is modified to a slightly larger p2 form in the Golgi apparatus, and then passes through the TGN to late endosomes and eventually the vacuole, where it is proteolytically processed to its mature form (reviewed by Stack et al., 1995). Pulse–chase analysis showed that CPY maturation in the Δtlg1 Δtlg2 double mutant was essentially normal, though slightly slower than in the parental cells (Figure 8A). Each single mutant showed at most a slight delay (Figure 8B), and incubation of Δtlg1 cells at 37°C did not increase the severity of the phenotype. Hence, forward traffic of CPY through the Golgi complex was largely unaffected by removal of Tlg1p and Tlg2p. Figure 8.CPY sorting and processing in tlg deletion strains. (A) Pulse–chase analysis of parental (SEY6210) and Δtlg1Δtlg2 strains. Cells were labelled for 4 min at 30°C and chased for the times indicated (min), before immunoprecipitation of CPY. The ER (p1), Golgi (p2) and mature, vacuolar (m) forms of CPY are indicated. (B) Rate of CPY processing in single tlg deletion strains. Pulse–chase analysis was carried out as in (A), but one batch of Δtlg1 cells was preincubated for 2 h at 37°C, and pulse–chased at 37°C. Comparison of the amounts of p1 and p2 relative to mature CPY after 10 and 20 min of chase reveals only a slight decrease in the processing rate in each mutant compared with the parental strain, even at 37°C. (C) CPY secretion from parental (SEY6210), Δtlg1, Δtlg2, Δtlg1Δtlg2 and Δpep12 strains. Cells were grown overnight at 30°C in contact with nitrocellulose and secreted CPY detected by immunostaining. While Δtlg1 and Δtlg1Δtlg2 cells secreted slightly more CPY than the parental strain, far more was secreted by cells lacking the pre-vacuolar t-SNARE Pep12p. Download figure Download PowerPoint The efficient processing of CPY in the mutants indicated that there was little mis-sorting to the cell surface, and indeed we failed to detect any labelled CPY released from the cells. This was confirmed by a colony blotting assay (Figure 8C). This assay detected slightly more CPY released by Δtlg1 and the double mutant than by wild-type cells, but the effect was insignificant compared with the mis-sorting in Δpep12 cells. It seems that diversion of CPY from the secretory pathway, a process that depends on the shuttling of a receptor between late endosomes and the Golgi complex (Cooper and Stevens, 1996), does not require the Tlg proteins. We next assayed the secretion of the periplasmic enzyme invertase. Since our standard strains lack the chromosomal invertase gene, we transformed the Δtlg1 and Δtlg2 mutants with a plasmid expressing high levels of the secreted form of invertase. Assay of internal and total invertase activity showed that more than 80% of the invertase was secreted in each case. Examination of the low levels of internal invertase by immunoblotting showed that there was no significant accumulation of glycosylated precursors in the mutants, even when Δtlg1 cells were incubated at 37°C (Figure 9A). This indicates that there is no block in transport at any point in the pathway from the ER to the cell surface in the mutants. However, ∼10% of the total invertase synthesized accumulated as an unglycosylated precursor which had failed to enter the ER. We have observed a similar phenomenon with a variety of strains overexpressing invertase, apparently because the translocation capacity of the ER is easily saturated. The immunoblot also revealed that the Δtlg1 cells made invertase with slightly shorter outer polysaccharide chains than normal (Figure 9A). This suggests some reduction in the levels of mannosyl transferases in the Golgi apparatus, reminiscent of the dramatic loss of Kex2p from this strain. Figure 9.Biosynthesis and secretion of glycoproteins by tlg deletion strains. (A) Intracellular invertase in tlg deletion strains. An immunoblot of myc-tagged intracellular invertase in parental strain (SEY6210), Δtlg1 and Δtlg2 strains were grown at 25°C and, where shown, preincubated at 37°C for 2 h before lysis. The positions of size markers (kDa) are shown. Golgi-modified invertase is indicated with a vertical bar. The invertase in the Δtlg1 strain migrates more rapidly than the normal mature form, but is still considerably larger than the ER precursor (∼80 kDa). The asterisk marks a background band that is seen also in cells not expressing the myc-tagged invertase, and the arrow indicates the non-glycosylated cytoplasmic form of invertase. The Δtlg2 sample contained slightly more protein than the others in this experiment. (B) Radiolabelled glycoprotein secreted by tlg deletion strains. Cells were labelled with Tran35S-label for 5 and 15 m