A Saccharomyces cerevisiae Genome-Wide Mutant Screen for Altered Sensitivity to K1 Killer Toxin

Corresponding author: Howard Bussey, McGill University, 1205 Ave. Docteur Penfield, Montreal, Quebec H3A 1B1, Canada., howard.bussey{at}mcgill.ca (E-mail)

Communicating editor: M. JOHNSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Using the set of Saccharomyces cerevisiae mutants individually deleted for 5718 yeast genes, we screened for altered sensitivity to the antifungal protein, K1 killer toxin, that binds to a cell wall ß-glucan receptor and subsequently forms lethal pores in the plasma membrane. Mutations in 268 genes, including 42 in genes of unknown function, had a phenotype, often mild, with 186 showing resistance and 82 hypersensitivity compared to wild type. Only 15 of these genes were previously known to cause a toxin phenotype when mutated. Mutants for 144 genes were analyzed for alkali-soluble ß-glucan levels; 63 showed alterations. Further, mutants for 118 genes with altered toxin sensitivity were screened for SDS, hygromycin B, and calcofluor white sensitivity as indicators of cell surface defects; 88 showed some additional defect. There is a markedly nonrandom functional distribution of the mutants. Many genes affect specific areas of cellular activity, including cell wall glucan and mannoprotein synthesis, secretory pathway trafficking, lipid and sterol biosynthesis, and cell surface signal transduction, and offer new insights into these processes and their integration.

THE sequenced and analyzed Saccharomyces cerevisiae genome has enabled a program of precise targeted gene disruption, resulting in a collection of mutant strains deficient in each gene (WINZELER et al. 1999 ; GIAVER et al. 2002 ; see also: http://sequence-www.stanford.edu/group/yeast/yeast_deletion_project/deletions3.html). Such a collection promotes the discovery of cellular roles for genes by facilitating the characterization of mutant phenotypes and allows a comprehensive examination of the genetic complexity of a phenotype. We have used the S. cerevisiae gene disruption set to screen for K1 killer toxin phenotypes. Toxin resistance has been extensively studied by classical genetics, and many genes have been identified. This toxin is encoded on the M1 satellite virion of the L dsRNA virus of S. cerevisiae (WICKNER 1996 ). Toxin sensitivity results from binding of the protein to the cell surface and its subsequent action at the plasma membrane promoting a lethal loss of cellular ions (reviewed in BUSSEY 1991 ; BREINIG et al. 2002 ). Defects in the genes involved in these processes may change cellular sensitivity to this toxin, and known resistant mutants define genes whose products are involved in cell wall synthesis and regulation (SHAHINIAN and BUSSEY 2000 ). Here we describe the results of global screens of haploids and homozygous and heterozygous diploid mutants for altered K1 toxin sensitivity.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and media:
Wild-type strains were BY4742 (MAT) and BY4743 (MATa/MAT; BRACHMANN et al. 1998 ), except where noted in Fig 1B, which also presents some results from strain SEY6210 (ROBINSON et al. 1988 ). Deletant strains were from the Saccharomyces Genome Deletion Consortium (GIAVER et al. 2002 ) and are available at Research Genetics (http://www.resgen.com/products/YEASTD.php3; see Table 1 for complete genotype descriptions). Haploid big1 and pkc1 mutants were obtained by dissection of the heterozygous diploid strains on media supplemented with 0.6 M and 1.0 M sorbitol, respectively. To improve spore viability of pkc1 tetrads, 1.0 M sorbitol was added during the zymolyase treatment of asci. Yeast were grown in standard YPD medium (SHERMAN 1991 ), unless otherwise stated. YPD/G418 medium, used to pregrow the mutants for 18 hr on 2% agar plates, is made of YPD supplemented with 200 mg/liter geneticin (GIBCO-BRL, Grand Island, NY). To test for drug sensitivity, YPD plates contained 25 or 50 µg/ml of calcofluor white, 30 or 80 µg/ml of hygromycin B, or 0.05% SDS.

View larger version (27K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Killer toxin sensitivity and quantification of major cell wall polymers of different strains. (A) A total of 5 µl of toxin was spotted onto agar seeded with a fresh culture of each strain (see MATERIALS AND METHODS). The mutant "killing zone" diameter was compared to the corresponding wild type and expressed as a percentage (see MATERIALS AND METHODS). (B) Measurement of cell wall ß-1,6- and ß-1,3-glucan levels was performed by extraction and fractionation of these polymers from cell wall preparations, followed by quantification of the alkali-insoluble fractions. The haploid mutants were from the Saccharomyces Genome Deletion Consortium (sla1 and big1) or from strains HAB880 and HAB900, respectively, for mnn9 and fks1 mutants (see Table 4). To facilitate comparison, the values of alkali-insoluble glucans were expressed as percentages of the corresponding wild-type level. The data represent averages of at least three independent experiments with standard deviations not exceeding 10%.

K1 killer toxin assay:
K1 toxin sensitivity was measured as follows (for details see BROWN et al. 1994 ). Yeast mutant strains (haploid MAT as well as the homozygous and heterozygous diploids) were pregrown for 18 hr at 30° on YPD/G418 in parallel with corresponding wild types pregrown on YPD. To control for variation in toxin activity between experiments, three wild-type controls were incorporated into every batch of mutants tested (100–600 mutants/batch). Approximately 1 x 10⁶ cells were resuspended in 100 µl of sterile water, of which 5 µl was used to inoculate 5 ml of molten YPD agar medium (1% agar, 0.001% methylene blue, and 1x Halvorson buffered at pH 4.7) held at 45°. Sorbitol was supplemented for big1 and pkc1 mutants, and for a wild-type control, as described above. This medium was quickly poured into 60- x 15-mm petri dishes and allowed to cool for 1 hr at room temperature. Then 5 µl K1 killer toxin (1000x stock diluted 1:10; BROWN et al. 1994 ) was spotted on the surface of the solidified medium. The plates were incubated overnight at 18° followed by 24 hr at 30° (48 hr for slow growth mutants). For each mutant showing a "killing" or "death" zone different from wild type, a picture comparing the mutant and appropriate control was taken with the IS-500 Digital Imaging System, version 2.02 (Alpha-Innotech). Two measurements of the killing zone were made with PhotoShop 4.0 and the average was saved in a database (FileMaker Pro 5.0) together with the picture. Mutants with a killing zone <90% or >110% were retested up to four times to confirm the observed phenotype. These percentages were determined as [(mutant killing zone diameter)/(wild-type killing zone diameter) x 100]. A subset of mutants showing killing zones <75% or >115% was selected for further characterization.

K1 toxin survival assay:
To determine cell survival after toxin treatment, 200 µl of a cell culture grown to log phase in YPD pH 4.7 and adjusted to OD₆₀₀ 0.5 was incubated with 50 µl toxin (1000x stock diluted up to 1:25) for 3 hr at 18° on a labquake. Percentage of surviving cells was calculated following plating onto YPD agar after incubation with toxin and counting colonies after 2 days at 30°.

Drug phenotype assay:
Drug sensitivity was determined by spotting diluted cultures on plates containing various drugs as described (RAM et al. 1994 ; LUSSIER et al. 1997 ). Briefly, 5 ml of liquid YPD medium, inoculated with freshly grown cells on YPD/G418, was incubated overnight at 30°. The cell density of these exponentially growing cultures was standardized with water at an OD₆₀₀ of between 0.485 and 0.515, and 2 µl of a set of 10-fold serial dilutions were spotted on YPD supplemented with calcofluor white, hygromycin B, or SDS (see Strains and media for drug concentrations). Hypersensitivity or resistance was monitored for each drug after 48 and 72 hr growth at 30°. The cells were also spotted on a control plate (YPD without drug), which allowed a comparison with the growth rate of the mutants after 24 hr growth at 30°. Pictures of all conditions tested were downloaded into a FileMaker 5.0 database (see above for details).

Cell wall composition analysis:
Total ß-glucan analysis: Haploid strains used for alkali-insoluble ß-glucan determinations were MAT sla1 and big1, respectively, obtained or derived from the Saccharomyces Genome Deletion Consortium (see Strains and media above) and compared to wild-type strain BY4742, while mnn9 (HAB880) and fks1 (HAB900) were compared to parental strain SEY6210. Crude cell walls were isolated and the levels of alkali-insoluble ß-1,3-glucan and ß-1,6-glucan quantified as previously described (DIJKGRAAF et al. 2002 ). The big1 mutant and the corresponding wild type were grown in medium containing 0.6 M sorbitol to provide osmotic support.

Alkali-soluble ß-1,6- and ß-1,3-glucan analysis: Alkali-soluble ß-1,6- and ß-1,3-glucan immunodetection was performed as described by LUSSIER et al. 1998 and summarized here. Yeast were pregrown on YPD/G418 for 18 hr at 30°, grown for 24 hr at 30° in 10 ml YPD liquid, and harvested by a 10-min centrifugation at 1860 x g. Cell pellets were washed with 5 ml of water and resuspended in 100 µl of water plus 100 µl of glass beads. The cells were then subjected to five cycles of vortexing for 30 sec, interspersed with 30-sec incubations on ice. Total cellular protein of the lysate was determined with the Bradford assay (BRADFORD 1976 ; Bio-Rad, Mississauga, ON, Canada) prior to alkali extraction (1.5 N NaOH, 1 hr, 75°). A set of 1:2 serial dilutions of the alkali-soluble fractions were then spotted on Hybond-C nitrocellulose membrane (Amersham, Oakville, ON, Canada). The immunoblotting was performed in Tris-buffered saline Tween containing 5% nonfat dried milk powder using either a 2000-fold dilution of affinity-purified rabbit anti-ß-1,6-glucan primary antibody (LUSSIER et al. 1998 ) or a 1000-fold dilution of anti-ß-1,3-glucan primary antibody (Biosupplies Australia Pty, Victoria, Australia), both with a 2000-fold dilution of horseradish peroxidase goat anti-rabbit secondary antibody (Amersham). The membranes were developed with a chemiluminescence detection kit (Amersham). Dot blots were scanned with a UMAX Astra 1220s scanner and signals were quantitated with Adobe Photoshop software, using the histogram function. The level of ß-1,6- and ß-1,3-glucan for each mutant was estimated by a comparison with a wild-type dilution series, with mutants classified by ranges of 20–35% (see footnote in Table 5).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

K1 toxin sensitivity of deletion mutants:
The toxin sensitivities of deletion mutants for 5718 genes were compared to those of the parental strain. The screen was performed on haploid (MAT) and homozygous diploid mutants, with toxin sensitivity being almost identical in both backgrounds. The heterozygous diploid collection was also tested, with the finding that 42 genes have a haploinsufficient toxin phenotype. The individual deletion of most genes has no effect on toxin sensitivity. These aphenotypic mutations include genes with a wide range of other phenotypes, such as slow growth and respiratory deficiency, and provide an important control for trivial cellular alterations that might affect the killing zone phenotype. For mutants in almost all genes, despite some affliction, the killer phenotype is wild type. Mutants in 268 genes (4.7%) have a phenotype distinct from wild-type toxin sensitivity, with 15 of these genes previously known to have such a phenotype (SHAHINIAN and BUSSEY 2000 ; DE GROOT et al. 2001 ). Table 2 Table 3 Table 4 list these mutants in functional groupings. A given gene is listed just once although some could be included in more than one category. Although the phenotypes are significant and reproducible, most null mutants have partial phenotypes. For example, among 155 haploid-resistant mutants, only 30 are fully resistant at the toxin concentration used (Table 2). Toxin sensitivity can be suppressed or enhanced in mutants, leading to resistance or hypersensitivity. Toxin resistance, which was always found as a recessive phenotype, is likely caused by a loss of function of some component needed for toxin action. In hypersensitive mutants the mutation synthetically enhances toxin lethality and can be functionally informative. Among the mutations resulting in a toxin phenotype, 42 were in uncharacterized open reading frames (ORFs) of unknown function. Of these, 3 were given a KRE (killer toxin resistant) number, and 8 genes with hypersensitive mutants were called FYV (function required for yeast viability upon toxin exposure) and given a number (Table 2 Table 3 Table 4).

Mutants for 118 genes with toxin phenotypes were examined for altered sensitivity to SDS, calcofluor white, and hygromycin B as hypersensitivity or resistance to these compounds is indicative of cell surface defects (LUSSIER et al. 1997 ; ROSS-MACDONALD et al. 1999 ). Mutants in 88 of these genes showed some additional phenotype (Table 2 Table 3 Table 4), independently suggesting that they have some cell surface perturbation. As ß-1,6-glucan is the primary component of the cell wall receptor for the toxin, mutants in 144 genes with toxin phenotypes were examined for alkali-soluble ß-1,6- and ß-1,3-glucan levels (Table 5) with 63 showing an altered level of one or both polymers. Genes previously identified as killer resistant provide positive controls for this global screen (Table 2 and Table 4). A number of characterized genes not known to have altered toxin sensitivity were found, suggesting that they have roles in cell wall or surface organization. Most mutants fall into a limited set of functional classes and define specific areas of cellular biology, some of which are described below (see also Table 2 and Table 3).

Glucan synthesis:
The yeast cell wall is made principally of four components: mannoproteins, chitin, ß-1,3-glucan, and ß-1,6-glucan (ORLEAN 1997 ; LIPKE and OVALLE 1998 ). Protein mannosylation and ß-1,6-glucan synthesis defects are known to lead to toxin resistance by altering the cell wall receptor for the toxin (see SHAHINIAN and BUSSEY 2000 for a review; BREINIG et al. 2002 ). Many mutations resulting in resistance to the K1 toxin have a reduced amount of ß-1,6-glucan in the cell wall and show slow growth or inviability depending on the severity of the defect, and we anticipated finding new genes affecting these processes. A complex pattern of glucan phenotypes was found among the mutants examined for alkali-extractable ß-1,6- and ß-1,3-glucan levels, with reduced or elevated amounts of one or both polymers found (Table 5). Of mutants in 63 genes with glucan phenotypes, 55 had effects on ß-1,6-glucan levels, with the remaining 8 having ß-1,3-glucan-specific alterations. Of the 55 with ß-1,6-glucan phenotypes, 40 also had some ß-1,3-glucan phenotype, with 15 showing a ß-1,6-glucan-specific phenotype. Principal findings are outlined below.

ß-1,6-Glucan reduced:
Mutants for five genes showing partial toxin resistance had specific but partial alkali-soluble ß-1,6-glucan reductions. Among these was the ß-1,3-glucan synthesis-associated gene FKS1, and this mutant also had reduced levels of alkali-insoluble ß-1,3-glucan (Fig 1B; Table 5). The involvement of Fks1p in both ß-1,3- and ß-1,6-glucan biogenesis has been studied further (DIJKGRAAF et al. 2002 ). Mutants in CNE1 encoding yeast calnexin have less ß-1,6-glucan (SHAHINIAN et al. 1998 ), and this is also a mutant phenotype of the uncharacterized gene YKL037W encoding a small integral membrane protein.

ß-1,6-Glucan reduced with altered ß-1,3-glucan:
big1 mutants had greatly reduced levels of ß-1,6-glucan and an increase in ß-1,3-glucan. BIG1 is a conditional essential gene retaining partial viability on medium with osmotic support (BICKLE et al. 1998 ). Heterozygous big1/BIG1 diploids showed haploinsufficient toxin resistance (Table 4), and haploid mutant cells grew very slowly on medium containing 0.6 M sorbitol and were toxin resistant (Fig 1A). Determination of the amount of alkali-insoluble glucan in the cell wall of a big1 mutant showed that the ß-1,6-glucan was 5% of wild-type levels (Fig 1B). The amount of ß-1,3-glucan in big1 mutants increased, possibly through some wall compensatory mechanism. Elsewhere, we have extended work on the role of Big1p in ß-1,6-glucan biogenesis (AZUMA et al. 2002 ).

Mutants for 13 genes had reductions in both alkali-soluble ß-1,6- and ß-1,3-glucan (Table 5), and three are described briefly below. smi1/knr4 mutants are resistant both to the K1 toxin and to the K9 toxin from Hansenula mrakii and have wall glucan defects and a reduced in vitro glucan synthase activity (HONG et al. 1994A , HONG et al. 1994B ). Smi1p localized to cytoplasmic patches near the presumptive bud site in unbudded cells and at the site of bud emergence (MARTIN et al. 1999 ) and may act in the polarization of glucan synthetic components. CSF1 (YLR087C) encodes an integral membrane protein that may be a plasma membrane carrier. The null mutant is hypersensitive to K1 toxin, calcofluor white, SDS, and hygromycin; TOKAI et al. 2000 showed the mutant to be salt and hydrogen peroxide sensitive with low temperature defects in growth and the uptake of glucose and leucine. LAS21 (YJL062W) participates in glycosylphosphatidylinositol (GPI) synthesis, adding an ethanolamine phosphate to the -1,6-linked mannose of the GPI mannose core (BENACHOUR et al. 1999 ). As this mannose core is the site of attachment of the ß-1,6-glucan moiety to GPI-linked cell wall proteins, altered levels of ß-1,6-glucan might be expected, although the basis of neither the ß-1,3-glucan defect nor the mutant hypersensitivity to K1 toxin is evident, indicating a need for further work.

ß-1,6-Glucan elevated:
Killer mutants in 33 genes had elevated levels of ß-1,6-glucan (Table 5). A group of ß-1,6-glucan overproducers are mutant in genes involved in assembly of the outer fungal-specific -1,6-glucan chain of N-glycosyl chains (mnn9, mnn10, and anp1; see Table 5 and Fig 2). Mutants in these genes are hypersensitive to killer toxin and are described further in N-glycosylation below. A contrasting group of resistant mutants overproducing ß-1,6-glucan (and to a lesser extent, ß-1,3-glucan) are in a subgroup of genes involved in cortical actin assembly and endocytosis (Table 2 and sla1 mutant in Fig 1). Our results are consistent with work reporting thickened cell walls in some of these mutants (for a review see PRUYNE and BRETSCHER 2000 ). Cell wall synthesis is normally restricted to the growing bud, but in these mutants new material is added inappropriately to the mother cell, resulting in a thickened wall (LI et al. 2002 ). It is surprising that cells with thickened cell walls and more ß-1,6-glucan can be killer toxin resistant, since resistance typically arises through loss of cell wall ß-1,6-glucan and less binding of the toxin. One explanation is that more toxin is bound to the walls, reducing its effective concentration, a resistance mechanism proposed for the SMKT toxin of Pichia farinosa (SUZUKI and SHIMMA 1999 ). A second explanation is that the thickened cell wall blocks toxin access to the plasma membrane.

View larger version (40K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Schematic summary of N-glycan biosynthesis in yeast. N-glycosyl precursor assembly is initiated in the endoplasmic reticulum. At the stage of GlcNAc2Man9, three glucose residues are serially transferred from the Dol-P-Glc donor to the N-glycan by the glucosyltransferases Alg6p, Alg8p, and Alg10p. Glucosylation is required for efficient transfer of the N-glycan to target proteins by a complex that includes Ost3p. The glucose residues are subsequently trimmed by the sequential action of glucosidases I and II, Cwh41p and Rot2p, respectively. N-linked oligosaccharides undergo further maturation in the Golgi, where addition of the fungal-specific "outer-chain" is initiated by Och1p and elaborated by various enzymes, including the mannan polymerase complex (adapted from ORLEAN 1997 ; SHAHINIAN and BUSSEY 2000 ). Arrows indicate activation and bars indicate negative effects. (*) indicates essential genes; i.e., only heterozygous mutants were tested. Genes whose deletion causes toxin hypersensitivity, red; resistance, blue; no phenotype, yellow; not tested, white. ß-Glucans are shown as follows: ß-1,6-glucan and ß-1,3-glucan both reduced; ß-1,6-glucan reduced and ß-1,3-glucan wild type; ß-1,6-glucan and ß-1,3-glucan both wild type; ß-1,6-glucan elevated and ß-1,3-glucan wild type; ß-1,6-glucan elevated and ß-1,3-glucan reduced. Mnn2p, Alg10p, and Hoc1p are not listed in Table 2 or Table 3; they are resistant or hypersensitive to K1 toxin, but fall outside of the chosen ranges.

Mutants for other genes that specifically overproduce alkali-soluble ß-1,6-glucan have broadly acting gene products, with mutants expected to be pleiotropic and their effects indirect. These include MAP1 encoding one of an essential pair of methionine aminopeptidases; this mutant is killer toxin, calcofluor white, hygromycin, and SDS hypersensitive (Table 3) and has a random budding pattern (NI and SNYDER 2001 ). ERG4 encodes an oxidoreductase required for ergosterol synthesis. This mutant is partially toxin resistant, hypersensitive to calcofluor white, hygromycin, and SDS (Table 2), and has a random budding pattern (NI and SNYDER 2001 ). ERV14 (YGL054C) and ERV41 (YNL067C) encode COPII vesicle coat proteins involved in endoplasmic reticulum (ER)-to-Golgi trafficking (OTTE et al. 2001 ), and both show toxin resistance. Mutants in four genes of unknown function also overproduce alkali-soluble ß-1,6-glucan (Table 5). Two of these genes, BUD27 (YFL023W) and BUD30 (YDL151C), have random budding patterns when mutated (NI and SNYDER 2001 ), and both are hypersensitive to killer toxin. FYV5 (YCL058C) encodes a predicted small integral membrane protein, with the mutant sensitive to sorbitol and low temperature (BIANCHI et al. 1999 ) and K1 toxin hypersensitive. Finally, the null mutant of YGL007C has partial killer toxin resistance (Table 2).

N-glycosylation:
Defects in N-glucosylation and its processing can lead to partial toxin resistance and reduced levels of ß-1,6-glucan (ROMERO et al. 1997 ; SHAHINIAN et al. 1998 ). Our results extend this finding to many other genes whose products are involved in the biosynthesis and elaboration of the Glc₃Man₉GlcNAc₂ oligosaccharide precursor of N-glycoproteins (Table 2 and Table 3; Fig 2). If Golgi synthesis of the fungal-specific -1,6-mannose outer arm of the N-chain is blocked by mutation in OCH1 or in MNN9, MNN10, or ANP1 of the mannan polymerase complex, toxin hypersensitivity results, concomitant with higher levels of ß-1,6-glucan in the cell wall (Table 3 and Table 5; Fig 1 and Fig 2; and see MAGNELLI et al. 2002 for mnn9). The glucan levels observed in an och1 mutant were similar to those obtained in a mnn9 mutant (not shown). To explore this further we determined the alkali-soluble glucan levels for other mutants in the mannan polymerase complex and the outer chain -mannosyltransferases (Fig 2), irrespective of toxin phenotype. A mutant in mnn11, part of the -1,6-mannose-synthesizing mannan polymerase complex, also showed elevated glucan levels, as did mnn2 encoding the major -1,2-mannosyltransferase that initiates mannose branching from the -1,6-glucan backbone. However, a mutant in mnn5, whose gene product extends the -1,2-mannose branches from the -1,6-glucan backbone, had reduced levels of both ß-glucans. Previous work showed that a small amount of glucan is attached to the N-chain structure (TKACZ 1984 ; VAN RINSUM et al. 1991 ; KOLLAR et al. 1997 ), and a genetic study by SHAHINIAN et al. 1998 also suggested this possibility. Our results show that core N-chain processing is required for wild-type ß-1,6-glucan levels, while absence of the outer -1,6-linked mannose side chain or its first -1,2-mannose branch can result in an increase in cell wall ß-1,6-glucan. However, mutants in later mannosylation steps in elaborating branches from the outer -1,6-linked mannose side chain have no effect or lead to reduced ß-glucan levels.

Lipid and sterol synthesis and ion homeostasis:
Mutants for 10 genes involved in the biosynthesis or regulation of lipids or sterols show partial toxin resistance (Table 2). These mutants have defects in membrane structure, possibly affecting the efficiency of insertion of the toxin into the plasma membrane or altering the cellular membrane potential leading to reduced toxin-induced ion permeability. Pertinently, defects in the ATP-dependent Drs2p and Atp2p membrane channels involved in cation and proton pumping confer toxin resistance. The altered membrane composition in lipid or sterol mutants could also affect secretory pathway function, possibly linking their partial toxin resistance phenotypes to those found in protein trafficking and secretion (Table 2). For example, KES1 is implicated in ergosterol biology and can partially suppress the toxin resistance of a kre11-1 mutant, with Kre11p being involved in Golgi vesicular transport as a subunit of the TRAPP II complex (JIANG et al. 1994 ; SACHER et al. 2001 ).

High-osmolarity and stress response pathways:
To survive hyperosmotic conditions, S. cerevisiae increases cellular glycerol levels by activation of the high-osmolarity glycerol (HOG) mitogen-activated protein kinase (MAPK) pathway. Such activation leads to elevated transcription of genes required to cope with stress conditions, including the synthesis of glycerol with a resultant increase in internal osmolarity (POSAS et al. 1998 ; REP et al. 2000 ). Mutants with an inactive HOG pathway are toxin hypersensitive, while deletion of protein phosphatases, such as PTP3, PTC1, or PTC3, which act negatively on the pathway, lead to resistance (Table 3 and Table 2, respectively; Fig 3B). Deletion of HOG1 resulted in a killing zone diameter almost twice that of the wild type. For such large killing zones, the diameter is limited by the diffusion rate of the protein toxin and greatly underestimates increased mutant sensitivity. To quantify sensitivity in a hog1 mutant, toxin-induced cell mortality was measured using a cell survival assay (see MATERIALS AND METHODS). A 10,000-fold reduction in cell viability was found when compared to the wild type. Previous estimates indicate that 3 x 10⁴ molecules of toxin are required to kill a wild-type cell (BUSSEY et al. 1979 ). We compared the sensitivity of the hog1 parental wild type from the deletion collection (strain BY4742) with strain S14a, on which the original lethal dose estimate was made, and found the strains to be of similar sensitivity (data not shown). Thus, just a few toxin molecules per cell are required to kill a hog1 mutant, indicating that a functional HOG pathway provides cells with a powerful way to ameliorate the effects of this toxin.

View larger version (36K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Schematic summary of signal transduction pathways involved in osmoadaptive responses and cell wall synthesis in yeast. (A) Exposure to high extracellular osmolarity triggers an adaptive response mediated by two pathways that converge at Pbs2p. One arm of the pathway involves the binding of Pbs2p to plasma membrane protein Sho1p. Pbs2p is phosphorylated by the Ste11p MAPKKK, through a process requiring Cdc42p, Ste50p, and Ste20p (DESMOND et al. 2000 ). A second pathway involves the two-component osmosensor module Sln1p-Ypd1p-Ssk1p, which activates Pbs2p via a pair of related MAPKKK proteins, Ssk2p and Ssk22p. Activation of this MAPK cascade culminates at Hog1p with Hog1p-dependent activation of the Rck2p protein kinase and activation and inactivation of transcription factors. The model also outlines the action of some negative regulators of the pathway (POSAS et al. 1998 ; REP et al. 1999 , REP et al. 2000 ; BILSLAND-MARCHESAN et al. 2000 ; and references therein). (B) Environmental stresses cause changes in cell wall state, which are detected by the Wsc proteins and Mid2p and Mtl1p. The information is transmitted to Rho1p by the guanine nucleotide exchange factors Rom1p and Rom2p. Tor2p is also an activator of Rho1p, whereas Sac7p and Bem2p are GTPase-activating proteins for Rho1p. Activated, GTP-bound Rho1p interacts with a transcription factor (Skn7p) and regulates the activity of proteins involved in cytoskeleton assembly (Bni1p), cell wall synthesis (Fks1p), and signal transduction (Pkc1p). Pkc1p in turn activates the cell integrity MAP kinase pathway and independently on "another arm" effects Rap1p-dependent transcriptional repression of ribosomal protein genes (LI et al. 2000 ; PHILIP and LEVIN 2001 ; and references therein). For the color-coding scheme, see Fig 2.

The sequence of action of the K1 toxin begins with its binding to ß-1,6-glucan cell wall receptors (SHAHINIAN and BUSSEY 2000 ). In a second step, the toxin inserts into the plasma membrane in a receptor-dependent process (BREINIG et al. 2002 ) and forms pores causing the leakage of ions and cellular metabolites, leading to cell death (MARTINAC et al. 1990 ; AHMED et al. 1999 ). To explore the defect in a hog1 mutant we asked where it occurred in the path of action of the toxin, by examining its epistasis in double-mutant combinations of hog1 with the toxin-resistant cell wall mutants kre1 and kre2, both of which block synthesis of the cell wall receptor. A kre1 hog1 mutant was as fully resistant as a kre1 single mutant, and a kre2 hog1 mutant was nearly so. Thus, defects in the cell wall receptor preventing binding of the toxin are dominant over the hypersensitivity of the hog1 mutant. This result is consistent with hypersensitivity occurring through some downstream effect such as ion homeostasis and/or lethal pore formation. One consequence of the activation of the HOG pathway is the induced expression of the glycerol-3-phosphate dehydrogenase Gpd1p, required in glycerol biosynthesis (ALBERTYN et al. 1994 ). To test whether impaired glycerol production was the basis of the hog1 mutant hypersensitivity, a gpd1 gpd2 double deletion mutant was made to reduce glycerol synthesis (GARCIA-RODRIGUEZ et al. 2000 ). This mutant had wild-type toxin sensitivity (data not shown). In further efforts to identify the downstream effectors of Hog1p responsible for the basal toxin resistance, we examined deletion mutants in the known transcription factors of the pathway, namely Msn1p, Msn2p, Msn4p, Hot1p, Sko1p, and Rck2p (PROFT and SERRANO 1999 ; REP et al. 1999 , REP et al. 2000 ; BILSLAND-MARCHESAN et al. 2000 ). All were wild type in sensitivity, as was the msn2 msn4 double mutant.

Cell integrity signaling:
In response to cell wall alterations, S. cerevisiae stimulates the Mpk1/Slt2p MAP kinase by activation of a cell integrity signaling pathway under the control of PKC1 (Fig 3B). Loss of function of this pathway results in deficiencies in cell wall construction and cell lysis phenotypes, which can be partially suppressed by osmotic stabilizers (LEVIN and BARTLETT-HEUBUSCH 1992 ; PARAVICINI et al. 1992 ; ROEMER et al. 1994 ). Consistent with playing a key role in cell surface integrity, a pkc1 haploid mutant kept alive by osmotic support is extremely sensitive to the toxin. However, most of the upstream activators of Pkc1p and all known downstream MAPK signaling components of the cell integrity pathway show no toxin phenotype (see Fig 3B). The absence of phenotype for the upstream integral plasma membrane activators of the pathway may be explained by the functional redundancy of the components (VERNA et al. 1997 ; KETELA et al. 1999 ; PHILIP and LEVIN 2001 ). Rho1p, the GTP-binding protein involved in relaying the signal from the plasma membrane to Pkc1p, is essential and the heterozygote has a wild-type phenotype. However, in the MAP kinase cascade downstream of Pkc1p, the kinase Bck1p and the MAP kinase Mpk1p are unique and nonessential (LEVIN and ERREDE 1995 ). The absence of a toxin phenotype upon mutation of these components indicates that hypersensitivity of a pkc1 mutant is not caused by the absence of activation of the MPK1 MAP kinase pathway, but in some other way (Fig 3B).

Ribosomal subunit proteins:
Defects in many ribosomal subunit proteins lead to toxin hypersensitivity. Of the 32 small ribosomal subunit genes, 8 are found as single copy and 24 are duplicated, for a total of 56 ORFs (PLANTA and MAGER 1998 ). Toxin hypersensitivity is observed for mutants for 21 of the duplicated genes (Table 3 and Table 4). A single deletion of either copy often shows hypersensitivity. In some cases only one of the duplicated gene mutants shows the phenotype (RPS0B, 4B, 10A, 17A, 19B, 23B), suggesting that they have distinct functions. Since some phenotypes were relatively weak (killing zone diameters <115% of the wild type), not all mutants are listed in Table 3. Of the 8 single-copy genes of the small ribosomal subunit, heterozygous deletions in just 2 essential genes, RPS13 and RPS15, gave toxin hypersensitivity (Table 4). The toxin hypersensitivity phenotype was more prevalent among mutants in the small subunit (43%) than among those in the large (16%). A total of 46 genes encode the large ribosomal subunit proteins, among which 35 are duplicated (PLANTA and MAGER 1998 ); 12 of the duplicated genes show toxin hypersensitivity when mutated (Table 3 and Table 4).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The ability to directly establish a phenotype-to-gene relationship is a great enabling strength of the mutant collection. Moreover, since each gene can be examined simply by testing a mutant, partial or weak phenotypes can be readily analyzed (BENNETT et al. 2001 ; NI and SNYDER 2001 ). The collection allows comprehensive screening and a knowledge of which genes have been examined, overcoming many of the limitations of a classical random mutant screen. Despite the extensive use of random screens for toxin resistance these failed to saturate the genome, as we have found mutants in many new genes. In addition, the mutant collection allows one to know which genes remain to be tested and, importantly, which genes do not have phenotypes. Such comprehensive testing can turn up the unexpected, as illustrated by a few examples. The extent of the relationships between cell wall polymers was unanticipated. Wall glucan work normally focuses on one or the other glucan synthetic pathway, and these are implicitly seen to be specific. Yet fks1 mutants, defective for a component of the ß-1,3-glucan synthase, are affected for both ß-1,3- and ß-1,6-glucan (Fig 1A), as are a large number of other mutants (Table 5). These interactions likely indicate synthetic or regulatory links between these polymers. The mnn9 mutation, which blocks synthesis of the outer -1,6-mannose arm of N-glycans, was assumed specific and has been used to simplify structural analyses of glucomannoproteins in the cell wall (VAN RINSUM et al. 1991 ; MONTIJN et al. 1994 ). The fact that a mnn9 mutation has other secondary effects that increase the amount of glucan in the wall is an unexpected complication, with the possibility that previous work analyzed structures absent from wild-type cells. Electrophysiological work links the Tok1p potassium channel with toxin action (AHMED et al. 1999 ). In the deletion mutant collection used here neither the haploid MATa or MAT nor the diploid heterozygous or homozygous deletion of this gene had a phenotype. Thus, in this strain background Tok1p has no detectable role in toxin action, indicating that despite the ability of the toxin to activate conductance of Tok1p, this channel protein cannot be the only target for the K1 toxin and is not a significant in vivo target in this sensitive strain. Having mutants in all cellular pathways allows the pursuit of phenotype through functional modules and has value in making such connections. Some specific examples are discussed below.

Functional clustering:
The screen identified several examples of interactions that connect biological functions into larger cellular processes, sometimes already known in detail. For example, toxin phenotypes trace the relationship between almost every biosynthetic step of the N-glycosyl moiety of glycoproteins. The cytoskeletal mutants provide an example of a less well-characterized connectivity. Here a set of mutants in cytoskeletal processes has a common toxin resistance phenotype that correlates with mother cells showing abnormal wall proliferation. This wall phenotype, which is not a general one for all cytoskeletal defects, has been reported for individual genes (see PRUYNE and BRETSCHER 2000 ). This functional cluster of genes, which may function in limiting wall growth to daughter cells, offers insight into a new facet of morphogenesis.

The HOG pathway buffers toxin action:
Mutants in hog1 are close to being maximally sensitive to the toxin, dying at 1 molecule/cell, while in a HOG1 strain, four orders of magnitude more toxin is needed to kill a cell. How is this HOG1-dependent resistance achieved? One possibility is that the HOG pathway is stress induced as the toxin causes ion loss. Activation of this signaling pathway may result in changes in membrane conductance, intracellular osmotic pressure, or some other stress response, which can act to reduce the efficiency of the toxin in promoting loss of cellular ions. Although the toxin sensitivity of a gpd1 gpd2 double mutant is similar to wild-type cells, the possible involvement of Hog1p-dependent osmoadaptation cannot be excluded. Consistent with this scenario, GARCÍA-RODRIGUEZ et al. (2000) observed increased intracellular glycerol levels after treatment with the cell-wall-perturbing agent calcofluor white, independent of the action of GPD1 and GPD2. An alternative explanation that there is some constitutive HOG1-dependent effect on cell wall synthesis seems less likely on the basis of the following observations. Epistatic tests using kre1 hog1 and kre2 hog1 mutants are consistent with the HOG pathway acting at the membrane or intracellularly, as cell wall mutants are epistatic to the hog1 defect and remain toxin resistant in double mutants. Deficiencies in the HOG pathway result in extreme toxin sensitivity, and we reasoned that mutations in genes regulated by this pathway might also cause hypersensitivity. In looking for candidates, it is striking that some components specific to the RNA polymerase II complex (e.g., Gal11p, Med2p, Rpb4p, Rpb3p, Rpb7p, Srb5p, and Srb2p) or components shared between RNA polymerases I, II, and III (e.g., Rpb8p, Rpc10p, and Rpo26p) all display a strong toxin hypersensitivity, similar to that of HOG pathway mutants (Table 3 and Table 4). Is this response specific to the HOG pathway? Among the MAPK pathways in yeast (HUNTER and PLOWMAN 1997 ; GUSTIN et al. 1998 ), only the HOG pathway exhibits toxin hypersensitivity. Mutants in SMK1, MPK1, and YKL161c, which encode, respectively, the MAP kinase of the sporulation pathway, the cell integrity pathway, and a putative uncharacterized pathway, are not toxin hypersensitive. Similarly, a null mutation in the MAP kinase kinase encoding gene STE7, which is involved in both the haploid mating and invasive pathways, has no effect on toxin sensitivity. These observations suggest a possible connection between the signaling elements of the HOG pathway and the activity of the RNA polymerase II complex. To investigate which potential target genes of Hog1p are responsible for the hypersensitivity, we looked for toxin phenotypes resulting from mutations in genes known to be induced by osmotic shock (REP et al. 2000 ). None of these genes have an effect comparable to a hog1 mutant. Similar results were obtained for genes whose mRNA level is affected by a mutation of HOG1. However, among the genes whose mRNA level is diminished after a shift to high osmolarity (REP et al. 2000 ), ASC1 had a significant hypersensitivity (Table 3). ASC1 encodes a 40S small subunit ribosomal protein, one of many small ribosomal protein encoding genes that, when mutated, show toxin hypersensitivity (see Table 3 and below). Together, these observations suggest that, if the phenotype observed in a hog1 mutant results from a defect in expression, it is not through a single gene but may originate from a combined deficiency in more than one gene.

Signaling components involved in toxin sensitivity:
Although the HOG pathway is the only MAP kinase cascade showing a toxin phenotype, two upstream activators of MAPK pathways were identified in the screen: SSK1 and PKC1. The toxin hypersensitivity of an ssk1 mutant is consistent with its place upstream of the HOG signal transduction cascade. However, no toxin phenotype is found for the components of the cell integrity MAPK pathway signaling downstream of PKC1, namely, the sequentially acting kinases Bck1p, the redundant pair Mkk1p and Mkk2p, and the Mpk1p MAP kinase (Fig 3B). This raises the question of how Pkc1p signals in producing a normal response to the toxin. Previous genetic analysis suggested a bifurcation of the signaling downstream of PKC1 (ERREDE and LEVIN 1993 ; HELLIWELL et al. 1998 ). Our data are consistent with such a model since some "other arm" of the PKC pathway, distinct from the Bck1p-dependent arm, is responsible for the toxin phenotype. Additional evidence for an alternative pathway comes from studies on the coordination of cell growth and ribosome synthesis, where a block in protein secretion reduces ribosomal protein gene transcription (MIZUTA and WARNER 1994 ; NIERRAS and WARNER 1999 ). This mechanism is: (i) dependent on Pkc1p activity; (ii) not mediated by the cell integrity pathway MAPK cascade (BCK1 or MPK1); and (iii) blocked by rap1-17, a silencing-defective allele of RAP1 (LI et al. 2000 ). We found that a heterozygous rap1 mutant exhibits haploinsufficient toxin hypersensitivity (Fig 3B), providing additional support for Rap1p being an effector of Pkc1p.

Ribosomal subunit mutants show toxin sensitivity:
The coupling of protein secretion to ribosome synthesis through the PKC pathway (NIERRAS and WARNER 1999 ; LI et al. 2000 ) raises the possibility of regulation operating in the reverse direction: that is, defects in protein synthesis mediated predominantly through 40S ribosomal subunit proteins might affect protein secretion and cell wall synthesis. The binding of the rough ER ribosomes to Sec61p of the signal recognition particle is through the 60S ribosomal subunit (BECKMANN et al. 1997 ), and fewer mutants in 60S ribosomal proteins have toxin phenotypes, arguing that the coupling step in itself is unlikely to be the primary site of any such effect. A more mundane alternative explanation is that nonessential defects in protein synthesis through loss of redundant ribosomal proteins have nonspecific knock-on effects on protein secretion/cell wall synthesis through failure to make enough of a component required for protein secretion.

Strength and limitations of comprehensive phenotyping with the collection:
In addition to phenotypic clustering of genes, the simple discovery of biological roles for genes through phenotype remains an important part of this screen. For example, a number of mutants in poorly characterized genes have ß-glucan phenotypes that warrant investigation. The yeast disruption mutant collection has limitations. Duplicated genes and gene families having synthetic phenotypes but no phenotype when individually deleted will be overlooked. Also, the 1105 essential genes representing 18.7% of the yeast genome (GIAVER et al. 2002 ) cannot be screened directly. Haploinsufficiency phenotypes in heterozygotes disrupted in one copy of an essential gene provide a partial solution, as in the case of BIG1. In our screen such haploinsufficiency was found in the heterozygous mutants of 42 genes, but we still do not know the full extent of the involvement of essential genes in cell surface biology. A set of conditional lethal mutants in all essential genes would improve the value of the collection for screening these genes.

	FOOTNOTES

¹ Present address: Institute of Biochemistry, Swiss Federal Institute of Technology, Zurich CH-8093, Switzerland.
² Present address: Department of Bioapplied Chemistry, Osaka City University, 3-3-138 Sugimoto, Sumiyoshi-ku Osaka, 558-8585, Japan.

	ACKNOWLEDGMENTS

We thank Angela Chu and Ron Davis (Stanford University) and Sally Dow (Rosetta Inpharmatics) for providing strains and assistance. We also thank Ashley Coughlin, Steeve Veronneau, Marc Lussier, and Terry Roemer for discussions and contributions and Robin Green and Federico Angioni for comments on the manuscript. Supported by operating and CRD grants from the Natural Sciences and Engineering Research Council of Canada. N.P. was a predoctoral fellow of the Fonds pour la Formation de Chercheurs et l'Aide à la Recherche (Quebec).

Manuscript received November 5, 2002; Accepted for publication December 12, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AHMED, A., F. SESTI, N. ILAN, T. M. SHIH, and S. L. STURLEY et al., 1999  A molecular target for viral killer toxin: TOK1 potassium channels. Cell 99:283-291.[Medline]

ALBERTYN, J., S. HOHMANN, J. M. THEVELEIN, and B. A. PRIOR, 1994  GPD1, which encodes glycerol-3-phosphate dehydrogenase, is essential for growth under osmotic stress in Saccharomyces cerevisiae, and its expression is regulated by the high-osmolarity glycerol response pathway. Mol. Cell. Biol. 14:4135-4144.[Abstract/Free Full Text]

AZUMA, M., N. PAGE, J. LEVINSON, and H. BUSSEY, 2002  Saccharomyces cerevisiae Big1p, a putative endoplasmic reticulum membrane protein required for normal levels of cell wall ß-1,6-glucan. Yeast 19:783-793.[Medline]

BECKMANN, R., D. BUBECK, R. GRASSUCCI, P. PENCZEK, and A. VERSCHOOR et al., 1997  Alignment of conduits for the nascent polypeptide chain in the ribosome-Sec61 complex. Science 278:2123-2126.[Abstract/Free Full Text]

BENACHOUR, A., G. SPIROS, I. FLURY, F. REGGIORI, and E. CANIVENE-GANSEL et al., 1999  Deletion of GPI7, a yeast gene required for addition of a side chain to the glycosylphosphatidylinositol (GPI) core structure, affects GPI protein transport, remodelling, and cell wall integrity. J. Biol. Chem. 274:15251-15261.[Abstract/Free Full Text]

BENNETT, C. B., L. K. LEWIS, G. KARTHIKEYAN, K. S. LOBACHEV, and Y. H. JIN et al., 2001  Genes required for ionizing radiation resistance in yeast. Nat. Genet. 29:426-434.[Medline]

BIANCHI, M. M., G. SARTORI, M. VANDENBOL, A. KANIAK, and D. UCCELLETTI et al., 1999  How to bring orphan genes into functional families. Yeast 15:513-526.[Medline]

BICKLE, M., P. A. DELLEY, A. SCHMIDT, and M. N. HALL, 1998  Cell wall integrity modulates RHO1 activity via the exchange factor ROM2.. EMBO J. 17:2235-2245.[Medline]

BILSLAND-MARCHESAN, E., J. ARINO, H. SAITO, P. SUNNERHAGEN, and F. POSAS, 2000  Rck2 kinase is a substrate for the osmotic stress-activated mitogen-activated protein kinase Hog1. Mol. Cell. Biol. 20:3887-3895.[Abstract/Free Full Text]

BRACHMANN, C. B., A. DAVIES, G. J. COST, E. CAPUTO, and J. LI et al., 1998  Designer deletion strains derived from Saccharomyces cerevisiae S288C: a useful set of strains and plasmids for PCR-mediated gene disruption and other applications. Yeast 14:115-132.[Medline]

BRADFORD, M. M., 1976  A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254.[Medline]

BREINIG, F., D. J. TIPPER, and M. J. SCHMITT, 2002  Kre1p, the plasma membrane receptor for the yeast K1 viral toxin. Cell 108:395-405.[Medline]

BROWN, J. L., T. ROEMER, M. LUSSIER, A. M. SDICU and H. BUSSEY, 1994 The K1 killer toxin: molecular and genetic applications to secretion and cell surface assembly, pp. 217–231 in Molecular Genetics of Yeast: A Practical Approach, edited by J. R. JOHNSTON. IRL Press/Oxford University Press, Oxford.

BUSSEY, H., 1991  K1 killer toxin, a pore-forming protein from yeast. Mol. Microbiol. 5:2339-2343.[Medline]

BUSSEY, H., D. SAVILLE, K. HUTCHINS, and R. G. PALFREE, 1979  Binding of yeast killer toxin to a cell wall receptor on sensitive Saccharomyces cerevisiae.. J. Bacteriol. 140:888-892.[Abstract/Free Full Text]

DE GROOT, P. W. J., C. RUIZ, C. R. VÁZQUEZ DE ALDANA, E. DUEAS, and V. J. CID et al., 2001  A genomic approach for the identification and classification of genes involved in cell wall formation and its regulation in Saccharomyces cerevisiae.. Comp. Funct. Genom. 2:124-142.

DESMOND, C., D. C. RAITT, F. POSAS, and H. SAITO, 2000  Yeast Cdc42 GTPase and Ste20 PAK-like kinase regulate Sho1-dependent activation of the Hog1 MAPK pathway. EMBO J. 19:4623-4631.[Medline]

DIJKGRAAF, G. J. P., M. ABE, Y. OHYA, and H. BUSSEY, 2002  Mutations in Fks1p affect the cell wall content of ß-1,3- and ß-1,6-glucan in Saccharomyces cerevisiae. Yeast 19:671-690.[Medline]

ERREDE, B. and D. E. LEVIN, 1993  A conserved kinase cascade for MAP kinase activation in yeast. Curr. Opin. Cell Biol. 5:254-260.[Medline]

GARCÍA-RODRIGUEZ, L. J., A. DURÁN, and C. RONCERO, 2000  Calcofluor antifungal action depends on chitin and a functional high-osmolarity glycerol response (HOG) pathway: evidence for a physiological role of the Saccharomyces cerevisiae HOG pathway under noninducing conditions. J. Bacteriol. 182:2428-2437.[Abstract/Free Full Text]

GIAVER, G., A. M. CHU, L. NI, C. CONNELLY, and L. RILES et al., 2002  Functional profiling of the Saccharomyces cerevisiae genome. Nature 418:387-391.[Medline]

GUSTIN, M. C., J. ALBERTYN, M. ALEXANDER, and K. DAVENPORT, 1998  MAP kinase pathways in the yeast Saccharomyces cerevisiae.. Microbiol. Mol. Biol. Rev. 62:1264-1300.[Abstract/Free Full Text]

HELLIWELL, S. B., A. SCHMIDT, Y. OHYA, and M. N. HALL, 1998  The Rho1 effector of Pkc1, but not Bni1, mediates signaling from Tor2 to the actin cytoskeleton. Curr. Biol. 8:1211-1214.[Medline]

HONG, Z., P. MANN, N. H. BROWN, L. E. TRAN, and K. J. SHAW et al., 1994a  Cloning and characterization of KNR4, a yeast gene involved in (1,3)-ß-glucan synthesis. Mol. Cell. Biol. 14:1017-1025.[Abstract/Free Full Text]

HONG, Z., P. MANN, K. J. SHAW, and B. DIDOMENICO, 1994b  Analysis of ß-glucans and chitin in a Saccharomyces cerevisiae cell wall mutant using high-performance liquid chromatography. Yeast 10:1083-1092.[Medline]

HUNTER, T. and G. D. PLOWMAN, 1997  The protein kinases of budding yeast: six score and more. Trends Biochem. Sci. 22:18-22.[Medline]

JIANG, B., J. L. BROWN, J. SHERATON, N. FORTIN, and H. BUSSEY, 1994  A new family of yeast genes implicated in ergosterol synthesis is related to the human oxysterol binding protein. Yeast 10:341-353.[Medline]

KETELA, T., R. GREEN, and H. BUSSEY, 1999  Saccharomyces cerevisiae Mid2p is a potential cell wall stress sensor and upstream activator of the PKC1-MPK1 cell integrity pathway. J. Bacteriol. 181:3330-3340.[Abstract/Free Full Text]

KOLLAR, R., B. B. REINHOLD, E. PETRAKOVA, H. J. YEH, and G. ASHWELL et al., 1997  Architecture of the yeast cell wall. Beta (16)-glucan interconnects mannoprotein, beta(13)-glucan, and chitin. J. Biol. Chem. 272:17762-17775.[Abstract/Free Full Text]

LEVIN, D. E. and E. BARTLETT-HEUBUSCH, 1992  Mutants in the S. cerevisiae PKC1 gene display a cell cycle-specific osmotic stability defect. J. Cell Biol. 116:1221-1229.[Abstract/Free Full Text]

LEVIN, D. E. and B. ERREDE, 1995  The proliferation of MAP kinase signaling pathways in yeast. Curr. Opin. Cell Biol. 7:197-202.[Medline]

LI, H., N. PAGE, and H. BUSSEY, 2002  Actin patch assembly proteins Las17p and Sla1p restrict cell wall growth to daughter cells, and interact with the cis Golgi protein Kre6p. Yeast 19:1097-1112.[Medline]

LI, Y., R. D. MOIR, I. K. SETHY-CORACI, J. R. WARNER, and I. M. WILLIS, 2000  Repression of ribosome and tRNA synthesis in secretion-defective cells is signaled by a novel branch of the cell integrity pathway. Mol. Cell. Biol. 20:3843-3851.[Abstract/Free Full Text]

LIPKE, P. N. and R. OVALLE, 1998  Cell wall architecture in yeast: new structure and new challenges. J. Bacteriol. 180:3735-3740.[Free Full Text]

LUSSIER, M., A. M. WHITE, J. SHERATON, T. DI PAOLO, and J. TREADWELL et al., 1997  Large scale identification of genes involved in cell surface biosynthesis and architecture in Saccharomyces cerevisiae.. Genetics 147:435-450.[Abstract]

LUSSIER, M., A. M. SDICU, S. SHAHINIAN, and H. BUSSEY, 1998  The Candida albicans KRE9 gene is required for cell wall ß-1, 6-glucan synthesis and is essential for growth on glucose. Proc. Natl. Acad. Sci. USA 95:9825-9830.[Abstract/Free Full Text]

MAGNELLI, P., J. F. CIPOLLO, and C. ABEIJON, 2002  A refined method for the determination of Saccharomyces cerevisiae cell wall composition and ß-1,6-glucan fine structure. Anal. Biochem. 301:136-150.[Medline]

MARTIN, H., A. DAGKESSAMANSKIA, G. SATCHANSKA, N. DALLIES, and J. FRANCOIS, 1999  KNR4, a suppressor of Saccharomyces cerevisiae cwh mutants, is involved in the transcriptional control of chitin synthase genes. Microbiology 145:249-258.[Abstract]

MARTINAC, B., H. ZHU, A. KUBALSKI, X. ZHOU, and M. CULBERTSON et al., 1990  Yeast K1 killer toxin forms ion channels in sensitive yeast spheroplasts and in artificial liposomes. Proc. Natl. Acad. Sci. USA 87:6228-6232.[Abstract/Free Full Text]

MIZUTA, K. and J. R. WARNER, 1994  Continued functioning of the secretory pathway is essential for ribosome synthesis. Mol. Cell. Biol. 14:2493-2502.[Abstract/Free Full Text]

MONTIJN, R. C., J. VAN RINSUM, F. A. VAN SCHAGEN, and F. M. KLIS, 1994  Glucomannoproteins in the cell wall of Saccharomyces cerevisiae contain a novel type of carbohydrate side chain. J. Biol. Chem. 269:19338-19342.[Abstract/Free Full Text]

NI, L. and M. SNYDER, 2001  A genomic study of the bipolar bud site selection pattern in Saccharomyces cerevisiae.. Mol. Biol. Cell 12:2147-2170.[Abstract/Free Full Text]

NIERRAS, C. R. and J. R. WARNER, 1999  Protein kinase C enables the regulatory circuit that connects membrane synthesis to ribosome synthesis in Saccharomyces cerevisiae.. J. Biol. Chem. 274:13235-13241.[Abstract/Free Full Text]

ORLEAN, P., 1997 Biogenesis of yeast wall and surface components, pp. 229–362 in Molecular and Cellular Biology of the Yeast Saccharomyces, Vol. 3: Cell Cycle and Cell Biology, edited by J. R. PRINGLE, J. R. BROACH and E. W. JONES. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

OTTE, S., W. J. BELDEN, M. HEIDTMAN, J. LIU, and O. N. JENSEN et al., 2001  Erv41p and Erv46p: new components of COP II vesicles involved in transport between the ER and Golgi complex. J. Cell Biol. 152:503-518.[Abstract/Free Full Text]

PARAVICINI, G., M. COOPER, L. FRIEDLI, D. J. SMITH, and J. L. CARPENTIER et al., 1992  The osmotic integrity of the yeast cell requires a functional PKC1 gene product. Mol. Cell. Biol. 12:4896-4905.[Abstract/Free Full Text]

PHILIP, B. and D. E. LEVIN, 2001  Wsc1 and Mid2 are cell surface sensors for cell wall integrity signaling that act through Rom2, a guanine nucleotide exchange factor for Rho1. Mol. Cell. Biol. 21:271-280.[Abstract/Free Full Text]

PLANTA, R. J. and W. H. MAGER, 1998  The list of cytoplasmic ribosomal proteins of Saccharomyces cerevisiae.. Yeast 14:471-477.[Medline]

POSAS, F., M. TAKEKAWA, and H. SAITO, 1998  Signal transduction by MAP kinase cascades in budding yeast. Curr. Opin. Microbiol. 1:175-182.[Medline]

PROFT, M. and R. SERRANO, 1999  Repressors and upstream repressing sequences of the stress-regulated ENA1 gene in S. cerevisiae: bZIP protein Sko1p confers HOG-dependent osmotic regulation. Mol. Cell. Biol. 19:537-546.[Abstract/Free Full Text]

PRUYNE, D. and A. BRETSCHER, 2000  Polarization of cell growth in yeast. J. Cell Sci. 113:571-585.[Abstract]

RAM, A. F., A. WOLTERS, R. TEN HOOPEN, and F. M. KLIS, 1994  A new approach for isolating cell wall mutants in Saccharomyces cerevisiae by screening for hypersensitivity to calcofluor white. Yeast 10:1019-1030.[Medline]

R