INTRODUCTION

The doubling of a cell requires the balanced doubling of all components of the cell. Most attention is generally placed on the doubling of the genetic apparatus, with its complex set of controls, based largely on the cyclins and associated kinases, as well as specific checkpoints to ensure accurate transmission of the genomes to the daughter cells.

Less understood but no less important is the doubling of the other components of the cell, among them the variety of different membranes and the complement of ribosomes. Balance must be maintained. In a complex system of parallel reactions, balance can only be ensured if there is cross-talk between the parallel reactions. An example of such cross-talk is apparent in the recent finding that the synthesis of ribosomes is dependent on the continued functioning of the secretory pathway (1, 2). Inactivation of any tested component of the secretory pathway, from the ER,¹ e.g., Sec63p, to the plasma membrane, e.g., Sec1p, leads to rapid repression of the transcription of both ribosomal RNA and ribosomal protein genes (1). Indeed, the coupling is sufficiently tight that one can distinguish a tight from a leaky sec mutant by their relative effects on the level of mRNA encoding ribosomal proteins. A similar repression occurs when the secretory pathway is inhibited by the drugs tunicamycin or brefeldin A (1).

Although the secretory pathway carries out a variety of functions, a major one is the production of the cell's membranes, in particular the plasma membrane. We surmise that the observations described above reflect the cell's attempt to maintain balance between the synthesis of cell membranes and the synthesis of ribosomes. This repression prevents the runaway production of ribosomes and therefore of protein synthesis capacity when the cell membrane system is under stress.

Several of the many proteins that participate in the secretory pathway are small GTPases that function as molecular switches in protein trafficking. They can be divided into two groups, the ARF/Sar family and the Sec4/Ypt/Rab family, which in mammals has at least 30 members (reviewed in Refs. 3 and 4). They are localized at distinct places along the secretory and endocytosis pathways, suggesting a quite general role in protein trafficking. In Saccharomyces cerevisiae, the most studied examples are Sec4p and Ypt1p (3). Ypt1p is required for ER to Golgi transport, apparently to facilitate the attachment and/or fusion of ER-derived vesicles to Golgi membranes (4, 5), and perhaps also in cis-medial Golgi transport (6). Sec4p is involved in transport from the Golgi to the plasma membrane (4, 5). One of the mammalian members of this family is Rab6, which by immunolocalization and functional studies has been implicated both in late intra-Golgi transport (7) and in budding from the TGN (8) (reviewed in Ref. 9).

Our continuing study of the regulation of ribosome synthesis has led us to clone the gene encoding the S. cerevisiae homologue of Rab6, which complements a temperature-sensitive mutant in which ribosome synthesis is repressed at the nonpermissive temperature. (While this paper was in preparation, a new release to GenBank^TM revealed that this gene is identical to YPT6 (), which had been described (8, 10) but whose sequence had not been made public.) Ypt6p is essential only at elevated temperatures. Assays of invertase secretion and carboxypeptidase Y (CPY) maturation suggested that Ypt6p is involved either in ER to Golgi or in cis- to medial-Golgi transport, in distinct contrast to its apparent role in mammalian cells. Isolation of second site suppressors implicated two additional proteins, Ssd1p and Imh1p, in the secretory pathway. Four additional mutants identified by the repression of ribosome synthesis were also found to be defective in CPY maturation, suggesting that they, too, are in genes encoding components of the secretory pathway.

MATERIALS AND METHODS

All the strains used are derivatives of W303 (11). J1003.6A(MATa, leu2-3, 112, ura3-1, can1-100, ade2-1, ssd1-1) is a haploid derivative of W303. 169ts is a temperature-sensitive strain derived by EMS mutagenesis (MAT, leu2-3, 112, ura3-1, can1-100, ade2-1, his3-11, 15, ssd1-1) (12).

J1006 cells were treated with ethyl-methane sulfonate to an approximate killing ratio of 45%. 517 temperature-sensitive strains were generated that can grow at 23 °C but not at 37 °C. Northern analysis was applied to screen for the mutants that show repressed transcription of ribosomal protein genes (1, 12). In this report, we focus on 169ts, one of several with this phenotype.

169ts cells were transformed with a CEN library carrying a URA3 marker (13) and selected on drop-out Ura plates at 37 °C. Screening for cells in which growth at 37 °C was dependent on the plasmid was accomplished using plates containing 5-flouroorotic acid at 1 mg/ml (14). Plasmid DNA was isolated from the transformed cells and was amplified in Escherichia coli. DNA from different colonies was re-transformed into 169ts. Those plasmids that can support the growth of 169ts at 37 °C were analyzed by restriction mapping and Southern blotting.

Two oligonucleotides, 5-AGCTGTTGATTCTGAACAGTA-3 and 5-GACGCACACAAAGAGTTC-3, were designed to clone the ypt6 gene of 169ts. Genomic DNA was isolated from 169ts and was used as template for polymerase chain reaction, using DNA polymerase Tli (Promega), which has proofreading ability. The polymerase chain reaction product was cloned into Bluescript KS+ and was sequenced using T7 and T3 primers.

An integrative plasmid YipYPT6 containing a EcoRI/ClaI fragment of wild type YPT6 was digested with AflIII, which cuts YPT6 uniquely. The linearized DNA was used to transform competent 169ts cells, and the cells were selected on drop-out Ura plates at 23 °C. The Ura⁺ cells were tested for the growth at 37 °C, and Southern analysis was applied to ensure that the integration had occurred at the right locus. The positive cells were crossed to wild type cells. The diploid cells were sporulated, the asci were dissected, and spores were tested for growth at 37 °C and on drop-out Ura plates and for integration of the wild type YPT6 gene by Southern analysis.

The EcoRI/ClaI (see Fig. 3) fragment was cloned into plasmid pRS316. A HindIII fragment containing the promoter and most of the coding region of YPT6 was replaced with a 1.1-kb HindIII fragment containing URA3. The EcoRI/ClaI fragment from the resulting construct was excised, purified, and used to transform wild type diploid cells. Ura⁺ cells were selected at 23 °C, and proper integration into the YPT6 locus was demonstrated by Southern analysis. Following sporulation, asci were dissected, and the spores grown at 23 °C on YPD plates.

RNA samples were isolated from log-phase yeast (15) and fractionated on 1.5% agarose gel (16). ³²P-Labeled antisense RNAs were used to detect mRNA derived from ACT1 and RPL32. ³²P-Labeled oligonucleotides were used to detect RNA derived from TCM1, RPL4A, RPS10A, PYK1, ENO1, and SNR17, encoding U3 RNA.

Wild type and mutant cells were grown overnight to log-phase at 23 °C in synthetic medium lacking methionine and cysteine. Half of the culture was shifted to 37 °C, whereas the rest remained at 23 °C. After 90 min cells from five A₆₀₀ units of culture were harvested and resuspended in 1.0 ml of prewarmed fresh medium containing 150 µCi of Expre³⁵S³⁵S (Dupont NEN). After 30 min of labeling at either 23 or 37 °C, the radioactive amino acids were chased for 15 min at the same temperature by adding 10 µl of a solution containing 0.1% cysteine, 0.4% methionine, and 100 mM (NH₄)₂SO₄. The chase was stopped by the addition of cycloheximide to 100 µg/ml and NaN₃ to 10 mM, and the culture was chilled on ice. Cell extracts were made according to Rose et al. (17) and immunoprecipitated using 2.0 µl of anti-CPY antibodies (generous gift of T. Stevens) at 4 °C overnight followed by adding protein A-Sepharose. Immunoprecipitated CPY proteins were fractionated on a 9% SDS-polyacrylamide gel.

Yeast cells grown to log-phase overnight in YPD medium at 23 °C were washed with and resuspended in YP containing 0.1% dextrose. Half of the culture was put at 37 °C; half remained at 23 °C. After 2 h of growth, 2 units A₆₀₀ of cells were collected and resuspend in 1 ml of 10 mM NaN₃, 25 mM Tris, pH 7.4. Half of the cells were stored on ice for external invertase assay, while the other half was assayed for internal invertase (18). Briefly, 0.5 ml of solution containing 2.8 M sorbitol, 50 mM potassium P_i, pH 7.5, 2 µl of -mercaptoethanol, and 5 µl of zymolyase (1 mg/ml in distilled H₂O) was added to the 0.5 ml of cells, and the mixture was incubated at 30 °C for 1 h. The cells were collected by centrifugation and resuspended in 0.5 ml of solution containing 1.4 M sorbitol, 25 mM potassium P_i, pH 7.5. Then the cells were lysed by adding 0.5 ml of 1% Triton X-100. For the invertase assay, 40 µl of either the intact cells or the lysate were combined with 0.1 M sodium acetate, pH 5.0 to 75 µl. The reaction was started by adding 25 µl of 0.5 M sucrose. After incubation at 37 °C for 20 min, the reaction was terminated by adding 150 µl of 0.2 M K₂HPO₄ and chilled on ice. The reaction mix was heated at 100 °C for 3 min, and 1 ml of a solution containing 0.3 mg/ml o-dianisidine, 0.05 mg/ml glucose oxidase, 0.01 mg/ml peroxidase, and 2 mM N-ethyl maleimide was added to the reaction. After incubation at 37 °C for 30 min, 1 ml of 6 M HCl was added to the reaction, and the A₅₄₀ was determined.

A construct that contains myc-tagged invertase under TPI1 promoter in an integration vector was kindly provided by Dr N. Dean. The plasmid was digested with XhoI whose unique site is within the URA3 gene. The linearized DNA was used to transform 169ts and sec18 cells, and transformants were selected on Ura plates at 23 °C. Proper integration was confirmed by Southern analysis. WT and mnn10 that carry the same integration were kindly provided by Dr. N. Dean. For Western analysis, a log-phase culture was split in half. Half remained at 23 °C, while the other half was shifted to 37 °C for 2 h. Cells were harvested, and cell extract was made according to Doseff and Arndt (19). Briefly, cells were washed with cold buffer A (100 mM Tris-HCl, pH 7.5, 200 mM NaCl, 1 mM EDTA, 5% glycerol, 0.5 mM dithiothreitol) and were resuspended in buffer A containing 1 mM phenylmethylsulfonyl fluoride (Sigma), 1.2 µg/ml leupeptin (Sigma), and 2 µl/ml aprotinin (Sigma). Cells were lysed by vortexing five times with glass beads for 30 s. Then the same volume of buffer B (50 mM Tris-HCl, pH 7.5, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 200 mM NaCl) was added. The cell extract was fractionated on a 7.5% SDS-polyacrylamide gel and then transferred to a nylon membrane (Millipore). The first antibody is the 9E10 anti-myc monoclonal, and the second is anti-mouse IgG (Amersham Corp.). An ECL kit was used to detect the invertase.

RESULTS

169ts is one of several mutants isolated from the screen of a bank of 517 ts mutants for those in which the levels of mRNA encoding ribosomal proteins appeared to be preferentially repressed at the nonpermissive temperature (see Ref. 1 and ``Materials and Methods''). The mutant has been back-crossed twice with wild type to confirm the co-segregation of the recessive ts and ribosomal protein mRNA phenotypes. The basis for choosing ts169 is shown in Fig. 1A, in which it is clear that the transcription of ribosomal protein genes is repressed at nonpermissive temperature. Although wild type cells show a transient repression of transcription of ribosomal protein genes that has been well documented (20, 21), mutant cells show a rapid and total loss of ribosomal protein mRNAs. Yet mRNAs derived from nonribosomal protein genes such as PYK1, ENO1, and ACT1 are hardly affected. These results suggest that the repression of transcription is relatively specific to ribosomal protein genes.

To ask whether rRNA transcription was similarly affected, we carried out pulse labeling studies (Fig. 1B). In wild type cells, one can see the 35 S initial transcript, as well as the 27 and 20 S intermediates and the 25 and 18 S final rRNA products. In 169ts cells at the permissive temperature, transcription and processing occur, albeit at a reduced rate. At 37 °C, however, the transcription of rRNA is severely repressed, and the few transcripts made are not processed. Based upon the phenotypes described, 169ts appeared likely to be carrying a mutant allele of a gene controlling the formation of ribosomes in S. cerevisiae.

The gene whose mutation is responsible for temperature sensitivity and repression of ribosomal protein synthesis was cloned using a library based on the CEN plasmid YCp50 (13) to complement the temperature sensitivity of 169ts. From about 40,000 Ura⁺ transformants, seven were able to grow at 37 °C. Using restriction map and Southern analysis (data not shown), they can be divided into groups A, B, and C. After re-transforming these plasmids back into 169ts, the growth of the transformed cells and wild type cells were compared at 37 °C (Fig. 2). Group A permits 169ts to grow nearly as well as wild type; Group B and especially group C support the growth of 169ts less well. We assumed that A carried the wild type allele of 169ts mutant gene and that B and C were single-copy suppressors. A 2.5-kb EcoRV fragment of group A was found to rescue the growth of 169ts at 37 °C (Fig. 3A). This DNA fragment, found on cosmid 8479 from Chromosome XII () carries only one complete gene. This gene encodes a protein 62.5% identical to the human Rab6 (Fig. 3B) and 68.6% identical to the Schizosaccharomyces pombe Ryh1 (22), both implicated in the secretory pathway. It has subsequently become clear that it is identical to the S. cerevisiae gene designated YPT6 (8, 10); to avoid confusion, we have adopted that nomenclature. This gene encodes a protein of 215 amino acids and has the consensus C-terminal sequence, CXC, used for post-translational addition of geranylgeranyl groups (8, 23). The sequence similarity suggests that Ypt6p is involved in the secretory pathway in S. cerevisiae.

To confirm that ypt6 is the mutant gene responsible for the phenotype shown in 169ts, wild type YPT6 was cloned into the URA3 integration vector YIp5. The plasmid was linearized within the YPT6 gene and used to transform competent 169ts cells. Ura⁺ cells, shown by Southern analysis to have undergone homologous recombination at the YPT6 locus, can grow at 37 °C. They were crossed to W303, the diploid cells were dissected, and the spores were grown on YPD plates at 23 °C. Among the spores derived from the 12 tetrads checked, none were ts, suggesting that YPT6 is or is very tightly linked to the mutant gene.

To study further the function of Ypt6p in S. cerevisiae, a null YPT6 strain was constructed (see ``Methods and Materials''). In a plasmid containing the 4-kb EcoRI/ClaI fragment, a 1-kb HindIII fragment containing the promoter and part of the coding region of YPT6 was replaced by a 1.1-kb fragment carrying the URA3 gene. The DNA fragment containing URA3 was used to transform wild type diploid cells. Homologous recombinants were identified by Southern blotting. The Ura<sup>+</sup> diploid cells were sporulated, the asci were dissected, and the spores were grown on YPD plates at 23 °C. Although all four spores were viable, the two Ura<sup>+</sup> colonies are distinctly smaller. Southern analysis confirmed that the two smaller colonies derive from spores containing the disrupted gene. Like 169ts, the null strain grows well at temperatures up to 34 °C, but at 37 °C it doubles only once. Based upon these results, we expected 169ts to carry a nonsense mutation in YPT6. The mutant gene was sequenced from a polymerase chain reaction fragment derived from amplification of the ypt6 gene of 169ts using a pair of oligonucleotides that flank the gene (see ``Materials and Methods''). A single mutation was found: a transition at codon 64 of TGG (Trp) to TGA (amber), resulting in a truncated Ypt6p (see Fig. 3B). This result further proves that Ypt6p is essential only at elevated temperatures.

Because the YPT6 null strain behaves like 169ts in terms of growth, we tested the transcription of ribosomal protein genes at 23 °C and 37 °C. The results are shown in Fig. 1A. As expected, the null strain shows repressed transcription of ribosomal protein genes at 37 °C, while transcription of nonribosomal protein genes is relatively unaffected. Surprisingly, the repression of transcription at 37 °C is not as severe as that in 169ts, suggesting that the truncated Ypt6p has some negative activity.

Based upon the sequence homology with Rab6, as well as on our previous observation that continued secretion is essential for ribosome synthesis (1), Ypt6p seemed likely to be involved in protein transport in S. cerevisiae. We therefore asked if 169ts was impaired in secretion. SUC2, encoding invertase, has two transcripts. The minor one is transcribed constitutively and its protein is soluble and stays in the cytoplasm; the major one is repressed by glucose, and its protein is transported to the cell surface along the secretory pathway. In derepressed wild type cells, invertase is mainly outside the cell. In mutants of the secretion pathway, the invertase remains largely within the cell (18, 24).

Mutant and wild type cultures were grown at 23 °C to log-phase and then transferred to 37 °C in derepression medium for 2 h. The external and internal invertase activity were measured (18, 24) (Fig. 4). At the nonpermissive temperature, invertase secretion is reduced to 15% of normal in the mutant strain, suggesting that Ypt6p is indeed involved in the secretory pathway. Invertase secretion is considerably less impaired in the null strain than in 169ts, suggesting again that truncated Ypt6p can have a negative effect on secretion.

To confirm the secretion defect and to determine the site at which Ypt6p functions, the glycosylation of CPY was analyzed in 169ts and the null strain. CPY is a vacuolar protein, synthesized as a proenzyme that is core-glycosylated in the ER (form p1), further glycosylated in the Golgi (form p2), and cleaved to mature CPY (form m) by Pep4p in the vacuole (Ref. 25 and Fig. 5). Analysis of the accumulation of isoforms of CPY can suggest the functional location of a mutant protein within the pathway (26). Thus in a pulse-chase experiment (Fig. 5), wild type cells show mostly mature CPY, whereas pep4 cells accumulate the p2 form within the vacuole.

In both 169ts and ypt6-null strains subjected at nonpermissive temperature to a pulse-chase with radioactive amino acids, there was accumulation of two forms of CPY (Fig. 5), one migrating as p1 and the other as m. Although unprocessed proCPY comigrates with the m form, treatment with endoglycosidase H demonstrated the band to represent authentic mature CPY (data not shown). We conclude that CPY maturation is partially defective at 37 °C and that Ypt6p is probably involved either in ER to Golgi transport or in a cis- to medial-Golgi step. In the case of CPY, 169ts and the null strain appear equivalent. YPT6 in a CEN plasmid complements the secretion defect in both 169ts (Fig. 5) and the null strain (not shown).

To determine more precisely the step of the secretory pathway at which Ypt6p is required, the glycosylation of invertase was analyzed. A myc-tagged invertase gene under the TPI1 promoter (kindly provided by N. Dean) was integrated by homologous recombination into the URA3 locus of each of the indicated strains. Extracts were made from cultures growing at 23 °C and after 120 min at 37 °C were subjected to Western analysis using the myc-specific monoclonal antibody 9E10 (Fig. 6). Because the TPI1 promoter is constitutively on, the invertase visualized in each lane represents molecules synthesized at both the permissive and nonpermissive temperatures. Wild type cells accumulate a broad band of polyglycosylated invertase secreted into the periplasmic space. In the mutant sec18 at the nonpermissive temperature, a more rapidly migrating form accumulates that has been core-glycosylated in the ER but has not been transported to the Golgi (27, 28). In the mutant mnn10 at both temperatures, invertase accumulates in the medial Golgi due to a deficiency of an glycosyl transferase (29) whose precise activity has not been determined (30). (Note that MNN10 has recently been identified in a screen for cell cycle mutants and rechristened BED1 (31).) Because some Golgi modifications have occurred, the molecules migrate more slowly than the invertase from the sec18 strain but still much faster than that from wild type.

Fig. 6 shows that at the nonpermissive temperature the 169ts cells, carrying the ypt6 truncation, accumulated invertase that has a mobility similar to that of the ER form (sec18ts) and greater than that of the medial Golgi form seen in the mnn10 cells, although there is some intermediate material, presumably due to the leakiness of the phenotype as seen in the CPY experiment. We conclude that Ypt6p acts primarily either at a step between the ER and the Golgi or at a very early step within the Golgi, certainly before the MNN10 step.

During the cloning of YPT6, we isolated two second site suppressors. Plasmid B supports relatively good growth of 169ts at 37 °C. Determination of sequences at the ends of the insert in this plasmid revealed that it was derived from chromosome IV (). The insert contains a single open reading frame, SSD1 (Fig. 7A) (32). This gene has been isolated in numerous suppressor screens and is also known as MCS1 (33), SRK1 (34), and SSL1 (35). SSD1 on a CEN plasmid can complement the secretion defect of 169ts or of the null as tested by CPY accumulation (Fig. 5). Presumably as a result, the repression of ribosomal protein genes is partially suppressed (Fig. 8).

It has recently been shown that W303, our wild type strain, carries the defective ssd1-d allele (32). The suppression of the growth defect of 169ts demonstrates that a mutant or null ypt6 shows synthetic lethality with ssd1-d at temperatures above 34 °C. The wild type, SSD1-v, was first isolated as a suppressor of the deletion of a putative protein phosphatase implicated in cell cycle progression (32) and later was found to suppress mutant alleles of a number of genes including PDE2, BCY1, INS1, and BEM1 (32, 34, 36). Although the mechanism by which Ssd1p functions is not clear, evidence suggests that it is probably involved in cell growth and morphogenesis (19). Fig. 5 is, however, the first evidence that Ssd1p can influence the secretory pathway and in turn ribosomal protein synthesis (Fig. 8).

A second, weaker suppressor of the ypt6 mutation in 169ts, plasmid C, was found to reside on chromosome XII (cosmid L2142, ). Within this insert are six complete genes including CDC25 (Fig. 7B). By multiple-step subcloning, we identified the gene that can suppress 169ts was identified as ORF L2142.5, not previously observed biologically. Since it shares with integrins and myosins significant homology, we term it IMH1. Although IMH1 on a CEN plasmid has little apparent effect on the leaky inhibition of CPY maturation in a ypt6 mutant or null strain (Fig. 5), it nevertheless restores growth and substantially rescues the transcription of ribosomal protein genes from ypt6-induced repression (Fig. 8). Because it was conceivable that Imh1p exerted its effect by suppressing the mutant phenotype of ssd1-d, we transformed the plasmid containing IMH1 into strains containing mutant alleles of SIT4 and LAS1, which grow normally in SSD1-v strains but not in ssd1-d strains. The presence of IMH1 on a CEN plasmid rescued neither the sit4 nor las1 mutants. Therefore Imh1p is unlikely to act through Ssd1p.

The original screen for ts mutants that appeared to be specifically repressed for their expression of mRNA for ribosomal proteins uncovered only a few (1). Because two of these were in fact defective in genes of the secretory pathway, SLY1 (1) and YPT6, we asked if others might also be involved in secretion. Fig. 9 shows that by the criteria of CPY glycosylation and maturation, 257ts, 271ts, and 367ts are clearly defective in secretion, whereas 394ts is moderately defective. Furthermore each of them induces KAR2 (data not shown), a characteristic of mutants in the early secretory pathway (37). No clearer evidence is needed for the close interlocking of the secretory pathway with ribosome synthesis than the observation that four more mutants, selected for insufficient ribosome biosynthesis, have serious defects in the secretory pathway.

DISCUSSION

This report describes the cloning of YPT6, encoding the Rab6 homologue of S. cerevisiae, by complementing a temperature-sensitive mutant identified by the repression of ribosome synthesis at the restrictive temperature. Our studies provide three lines of evidence that Ypt6p is involved in the secretory pathway of S. cerevisiae at elevated temperatures. The maturation and transport to the vacuole of carboxypeptidase Y is impaired in the absence of functional Ypt6p (Fig. 5), although this phenotype is rather leaky (compare Fig. 5 with Fig. 9). The secretion of invertase is inhibited in the absence of Ypt6p (Fig. 4), the molecules accumulating apparently in the ER or the cis-Golgi (Fig. 6). Finally, electron microscopic examination of 169ts cells maintained at the restrictive temperature for 90 min revealed the development of many vesicular structures and an apparent loss of vacuolar contents (data not shown). Surprisingly, Ypt6p is essential only at elevated temperatures, as is its homologue in S. pombe (22). Is this a characteristic of the secretory pathway, or is there a protein with redundant function that itself is temperature-sensitive?

Functional studies of Rab6p in mammalian systems suggest that it participates in late intra-Golgi transport, i.e., between the cis-Golgi  medial/trans-Golgi compartments (7), and perhaps also in budding of exocytic vesicles from the TGN (38). Our data summarized above show that in S. cerevisiae deficiency of Ypt6p causes a block at an earlier step in the secretory pathway. The CPY and invertase molecules that escape this block, however, are matured normally, suggesting that Ypt6p does not participate in transport of secretory proteins beyond the medial Golgi. Yet mammalian Rab6 can support the growth of a YPT6 null strain (8). It remains to be seen if this molecule has different roles in mammalian compared with yeast secretion, or, more likely, if the assays used in the two systems are not fully congruent.

The ypt6 gene of 169ts carries a nonsense mutation at codon 64, generating a truncated Ypt6p. Genetically, this mutation is recessive; wild type YPT6 on a CEN plasmid can rescue the growth, the secretion defects, and the repression of ribosomal protein synthesis at nonpermissive temperature. Nevertheless, the truncated form of Ypt6p has detectable negative effects in comparison with a null strain, both in secretion of invertase (Fig. 4) and in the repression of ribosomal protein gene transcription (Fig. 1A). It is unlikely that 169ts carries a mutation in a second gene because it has been repeatedly back-crossed to the wild type parental strain. Ypt6p, like other members of the small G-proteins, has five highly conserved regions, termed G-1 to G-5, that are critical for GTP-induced conformational change, GTP hydrolysis, and GTP/GDP exchange (39). The truncation in the ypt6 allele of 169ts occurs in the middle of domain G-3. Perhaps the truncated Ypt6p sequesters proteins that interact with its N-terminal domain.

Ypt6p is essential for growth at high temperatures only in the absence of functional Ssd1p. W303, the parental strain from which 169ts was derived, was shown to contain the ssd1-d allele (32). In 169ts SSD1-v on a CEN plasmid can suppress the growth defect (Fig. 2), the secretion defects (Fig. 5), and the repression of ribosomal protein transcription (Fig. 1A). SSD1-v was isolated originally as a suppressor of the deletion of SIT4, required for the start of S phase (32). Later, SSD1-v was found to suppress mutations in many genes including BCY1, (32), INS1 (34), SLK1 (35), BEM2 (36), and LAS1 (19). Although the products of these genes are involved in cell growth or morphogenesis, our evidence implicates Ssd1p in the secretory pathway as well. The biochemical activity of Ssd1p remains totally unclear.

IMH1 on a CEN plasmid can suppress partially the growth defect of 169ts, as well as the repression of transcription of ribosomal protein genes. At first it appears puzzling that IMH1 does not suppress the defect in CPY maturation (Fig. 5). Yet ts169 is only marginally temperature-sensitive, growing well at 34 °C but not at 37 °C. Because IMH1 is only a partial suppressor, perhaps it is not surprising that it would have different effects on different assays. It is likely that Imh1p plays a structural role, because it shares some homology with integrins, myosins, and Uso1p, a cytoskeletal protein implicated in the secretory pathway (40, 41).

Although the work described above has contributed to our understanding of the secretory pathway, it is important to remember that the mutants were identified originally as being defective in ribosome synthesis at the nonpermissive temperature. Indeed, our screen of such mutants has now identified six that are implicated in some aspect of the secretory pathway. Thus, there is a coordinate repression of transcription encompassing 100 ribosomal RNA genes and >100 ribosomal protein genes (Fig. 1; Ref. 1)² in all mutants we have examined that are defective in CPY maturation, i.e., in 169ts (ypt6), in 312ts (sly1), as well as in 257ts, 271ts, 367ts, and 394ts, whose mutant genes are as yet uncloned (Fig. 9). It is striking that this profound cellular response went undetected through more than a decade of study of the secretory pathway. Yet, in retrospect, it is hardly surprising that two macromolecular systems that are responsible for such a large proportion of the cell's mass would have mechanisms for communicating.

What is the basis for the dependence of ribosome synthesis on the secretory pathway? In experiments to be reported elsewhere, we show that the coupling is neither through the IRE1-mediated pathway by which KAR2 is induced by accumulation of imperfect proteins in the ER (37), nor through known pathways regulating ribosome synthesis, such as the stringent response to amino acid deprivation (42), nor through the protein kinase A response to carbon source (43).³ Thus we surmise that an as yet unidentified regulatory pathway connects ribosome synthesis to the secretory pathway. It remains to be seen whether the repression of ribosome synthesis represents a stress response to insufficient plasma membrane synthesis or a monitoring of the health of the secretory pathway in some more direct way.