Chitin synthase III: Synthetic lethal mutants and “stress related” chitin synthesis that bypasses the CSD3/CHS6 localization pathway

Abstract

We screened Saccharomyces strains for mutants that are synthetically lethal with deletion of the major chitin synthase gene CHS3. In addition to finding, not surprisingly, that mutations in major cell wall-related genes such as FKS1 (glucan synthase) and mutations in any of the Golgi glycosylation complex genes (MNN9 family) are lethal in combination with chs3Δ, we found that a mutation in Srv2p, a bifunctional regulatory gene, is notably lethal in the chs3 deletion. In extending studies of fks1-chitin synthase 3 interactions, we made the surprising discovery that deletion of CSD3/CHS6, a gene normally required for Chs3p delivery and activity in vivo, was not lethal with fks1 and, in fact, that lack of Csd3p/Chs6p did not decrease the high level of stress-related chitin made in the fks1 mutant. This finding suggests that “stress response” chitin synthesis proceeds through an alternate Chs3p targeting pathway.

The cell wall of the budding yeast, Saccharomyces cerevisiae, is composed primarily of mannoproteins, β-1,3 glucan, β-1,6 glucan, and chitin. In vegetative cells chitin, a fibrous polymer of N-acetylglucosamine, is found in a ring at the base of the emerging bud, at the septum, and in the lateral wall. Although chitin makes up only 1–2% of the dry weight of the cell wall it is important for cell wall integrity (1). Mutations that affect chitin synthesis cause osmotic sensitivity (2, 3), abnormal morphology, aggregation, and growth arrest with elongated buds (4–6).

There are three chitin synthase genes in S. cerevisiae: CHS1, (7), CHS2 (8), and CHS3 (4). Chs1p is thought to be a repair enzyme, synthesizing chitin in response to an acid-induced increase in chitinase activity after separation of mother and daughter cells (9). Chs2p is localized to the mother-bud junction and functions to synthesize chitin in the primary septum (5, 6, 8). The CHS3 gene encodes the catalytic subunit of the major chitin synthase CSIII, which is responsible for the synthesis of more than 90% of the chitin in the cell wall (2, 5, 6).

Additional chitin synthesis-related genes have been identified that are required for chitin synthase III activity. The product of the CHS4 (CSD4, CAL2, SKT5) gene is required for chitin synthase III activity both in vivo and in vitro (2, 10). The Chs4p may have a dual role, as a limiting subunit of a Chs3p-containing complex and as an activator of Chs3p (11). In addition, Chs4p links Chs3p to the septins, specifically those encoded by CDC10, through Bni4p (12). Chs5p (Cal3p) also is required for chitin synthase III activity both in vivo and in vitro (13, 14). Chs5p may play a similar role to Chs4p or may function in the localization of Chs3p. Santos et al. (13, 14) found that in the absence of Chs5p, Chs3p fails to localize to the neck or the site of the incipient bud. Csd3/Chs6p is required for chitin synthesis in vivo, but not for chitin synthase III activity in vitro (2, 11) and may be involved in transport of newly synthesized Chs3p from chitosomes to the plasma membrane (15, 16). The recently discovered Chs7p (17) is responsible for export of Chs3p from the endoplasmic reticulum.

The cell wall is a structure that is constantly undergoing change throughout the cell cycle and in response to external stimuli. Chitin synthase III is responsible for the synthesis of chitin during bud emergence and growth, mating, and spore formation (2). Chitin is deposited in a ring, at the site of bud emergence, late in the G₁ phase of the cell cycle. It has been shown by labeling with wheat-germ agglutinin conjugated to FITC as well as staining with Calcofluor that the lateral walls of mother cells contain much more chitin than the lateral walls of buds (5, 6, 18). The chitin in the lateral walls also is made by Chs3p, is cell cycle regulated, and occurs, for the most part, in the later stages of the cell cycle. In addition, Pammer et al. (19) found that transcript levels of CHS3 do not change significantly during the cell cycle, indicating that Chs3p is regulated posttranscriptionally. These findings indicate that the synthesis of chitin is regulated both spatially with respect to morphogenesis and temporally with respect to the cell cycle.

A synthetic lethal mutant screen is useful in identifying functionally related proteins, interacting proteins, regulatory proteins, and proteins that belong to the same complex (20, 21). Bender and Pringle (22) used a synthetic lethal screen to identify two new genes involved in bud emergence, BEM1 and BEM2, by screening for mutants that required plasmid-borne MSB1 for viability. Costigan et al. (23) used a similar screen to identify SLK1, a gene that when mutated requires SPA2 for vegetative growth. Both SPA2 and SKL1 play a role in polarized cell growth. Because deletions of CHS3 are viable, we screened for genes that when mutated are lethal only in combination with a chs3 deletion.

MATERIALS AND METHODS

Yeast Strains, Media, and Plasmids.

Yeast were grown in either rich medium (yeast extract/peptone/dextrose, YPD) or in synthetic minimal (SD) that have been described (24). Other media were SD⁺ (SD plus adenine, histidine, uracil, leucine, lysine, and tryptophan), SD⁺ −ura (as SD⁺ but lacking uracil), SD⁺ −trp (as SD⁺ but lacking tryptophan), and SD⁺ −leu −trp (as SD but lacking both leucine and tryptophan). Calcofluor white-containing medium was made as described (2). SD⁺ medium was used for all drug tests with the exception of Nikkomycin Z. For Nikkomycin Z drug tests, SDA⁺ medium, which contains allantoin (1 mg/ml final concentration) as the nitrogen source, was used (25). Solid media contained 2% agar (Difco), unless otherwise noted. All strains were grown at 26°C.

Strain Construction.

Standard procedures of yeast genetics were used (24). Yeast transformations were done by using the lithium acetate method (26). Standard methods were used for the construction of plasmids. Escherichia coli strain DH5α was used for transformation and plasmid construction. E. coli were transformed by using the procedure of Inoue et al. (27).

Plasmids.

pBK101. The 3.7-kb BamHI/NheI fragment of pDK255 (28) containing ADE3 was cloned into the BamHI/SpeI sites in the multiple cloning site of the TRP1-marked CEN6 vector pRS314 (29) to make pBK101.

pBK102.

pCSD2–15, a 5.5-kb ClaI-BamHI fragment of pCSD2–3 cloned into pSK (Stratagene), was cut with BamHI and SalI. The 5.5-kb band containing CHS3 (CSD2) was gel purified and ligated to BamHI/SalI-digested pBK101 to make pBK102.

pCSD2–3.

Bulawa (2) subcloned the 5.4-kb ClaI–BamHI fragment of pCSD2 into the same sites of the CEN6/ARSH4 vector pRS316 (29).

pRS316.

URA3-marked CEN6/ARSH4 vector was constructed as described (29).

p12a-1.

This plasmid (derived from a CEN4 URA3 genomic library) complements SRV2 mutants.

p13d-3.

This plasmid (derived from a CEN4 URA3 genomic library) complements FKS1 mutants.

p13a-1.

This plasmid (derived from a CEN4 URA3 genomic library) complements ANP1 mutants.

Yeast Genomic Library.

The CEN4 URA3-marked yeast genomic library was a generous gift from the Young Laboratory (Whitehead Institute, Cambridge, MA) and is described elsewhere (30).

Synthetic Lethal Screen.

Strain PRY487 was grown on SD⁺−leu−trp. Individual colonies were suspended in 2 ml of SD⁺−leu−trp and grown for 90 min at 26°C. A 1-ml aliquot of each suspension was sonicated briefly to disperse clumps. Cell counts were done by using a hemocytometer. Suspensions were diluted in H₂O and plated at 2 × 10³ cells/plate on YPD. These cells (total of 1 × 10 ⁵ cells) were mutagenized with UV irradiation to a viability of 6% (60-sec exposure, 40 cm from source, lamp output 10 erg/mm²), and lids were replaced and incubated at 26°C in the dark.

Characterization of Putative Synthetic Lethal Mutants.

Standard genetic procedures were used to determine the recessiveness or dominance of the mutants. Nonsectoring Sect⁻ mutants were mated to PRY 398, and the resulting diploids were tested for their ability to sector by streaking on YPD. Tetrad analysis was performed on diploids whose sectoring phenotype was recessive to confirm that both Sect⁻ and Sect⁺ were recoverable. This analysis also served to backcross the mutagenized strains.

Putative synthetic lethals were transformed with pRS316 and the CEN plasmid pCSD2–3, which carries a wild-type copy of CHS3 to verify the requirement for the plasmid-borne CHS3 for viability.

Agar Diffusion Assay for Drug Sensitivity.

Agar diffusion assays were used to test synthetic lethal mutants for sensitivity to the following drugs: Hygromycin B, Amphotericin B, Nikkomycin Z, tunicamycin, the Echinocandin L-733,560, sodium orthovanadate, and FK-506. The procedure used was a slightly modified version of that described by Island et al. (25).

Cloning of the Gene Complementing the Sect⁻ Phenotype.

We used the drug sensitivity of the synthetic lethal strains to facilitate the cloning of the genes, which, when mutated, are synthetically lethal with a CHS3 deletion. Synthetic lethal strains hypersensitive to Nikkomycin Z were transformed with the CEN4 URA3 library and plated on SD⁺−ura−trp at a density of 1 × 10³ cells/plate. After 3–4 days, transformants were replica plated to SDA⁺ −ura−trp, plus 15 mg/ml Nikkomycin Z. After 1–2 days incubation at 26°C, any transformants that had grown on the Nikkomycin Z plate were streaked for singles on YPD and YPD plus 500 mg/ml Calcofluor white. Single colonies were retested for the presence of the plasmid(s) and the ability to grow on plates with 15 mg/ml Nikkomycin Z. Transformants that were able to sector on YPD and resistant to Calcofluor and Nikkomycin Z underwent plasmid rescue (31).

Synthetic lethal strains that were sensitive to Calcofluor white were transformed with the CEN4 URA3 library and plated on SD⁺ −ura plates as described above. The SD⁺−ura plates contain tryptophan so there is no auxotrophic requirement for maintaining the TRP1-marked plasmid carrying the CHS3 gene. After 3–5 days, colonies were replica-plated onto SD⁺ −ura plus 700 mg/ml Calcofluor white. Calcofluor-resistant colonies from library-transformed strains were restreaked on YPD, YPD plus 700 mg/ml Calcofluor, SD⁺ −ura, SD⁺ −trp, and SD⁺ plus 1 mg/ml 5-fluoroorotic acid (32). Complementing library plasmid DNA was prepared from the yeast strain by the method of Hoffman and Winston (31).

Cell Wall Assays.

Chitin was measured by the method described in Bulawa (2). Other cell wall assays were done by using a slightly modified version of the procedure described by Castro et al. (33).

β-Glucan Synthase Assays.

Enzyme preparation and assays were done as described with minor modifications (33). Briefly, early logarithmic cells were resuspended in 1 mM EDTA (pH 8) and lysed in a BeadBeater. The crude lysate was centrifuged at low speed to remove unbroken cells and cell wall debris. After a high-speed spin, washed pellets were resuspended in buffer plus glycerol and stored at −20°C. β-1,3 Glucan synthase reactions were done as described (33). The amount of [¹⁴C]-glucose incorporated into acid-insoluble glucan was determined by using a Millipore filter method. The amount of [¹⁴C]-glucose incorporated into insoluble glucan trapped on the filter was determined by liquid scintillation counting.

RESULTS

Characterization of Putative Synthetic Lethal Strains.

To show that the requirement for the plasmid is caused by the presence of CHS3 and not by other sequences on the plasmid, putative synthetic lethal mutant strains were transformed with pRS316, a URA3-marked plasmid having a backbone identical to that of pBK102. Transformation with pRS316 should not effect the Sect⁻ phenotype of true synthetic lethals because it does not carry CHS3.

Putative synthetic lethal strains also were transformed with the CEN plasmid pCSD2–3, which carries a wild-type copy of CHS3 but not ADE3. True synthetic lethals should sector when they acquire pCSD2–3 because either plasmid can satisfy the requirement for CHS3. Of the 23 putative synthetic lethal mutant strains tested, 10 behaved as defined for true synthetic lethals when transformed with the tester plasmids. Of these 10, four mutants grew very slowly (1b-1, 1d-1, 5c-2, 7e-1) and one mutant, (9c-5) failed to grow in top agar. For the present study, the remaining five mutant strains were further characterized.

Agar Diffusion Assay for Drug Sensitivity.

The five synthetic lethal mutant strains that were recessive and sectored only when transformed with a plasmid carrying CHS3 were tested for growth in the presence of cell wall-specific drugs. The drugs included Nikkomycin Z, an inhibitor of chitin synthase III; Amphotericin B, a polyene that damages cell membranes; tunicamycin, a specific inhibitor of N-glycosylation; sodium orthovanadate (vanadate), resistance to which often indicates defects in glycosylation; Hygromycin B, an aminoglycoside; Echinocandin, an inhibitor of β-1,3-glucan synthesis; and FK 506, a calcineurin inhibitor. Mutations in FKS1 are known to be hypersensitive to FK506.

Relative to the parental strain (PRY487) three mutant strains (3d-2, 13a-1, and 13d-3) showed increased sensitivity to Nikkomycin Z. Two mutant strains (12a-1 and 12e-1) showed increased sensitivity to Amphotericin B. Only 12e-1 showed increased sensitivity to tunicamycin. None of the mutant strains showed any significant variation from the parent strain when tested for sensitivity to the β-1,3-glucan synthase inhibitor Echinocandin. All five mutant strains (3d-2, 12a-1, 12e-1, 13a-1, and 13d-3) showed increased sensitivity to the aminoglycoside Hygromycin B. Dean (34) demonstrated that abnormal glycosylation results in sensitivity to aminoglycosides, which suggests that all five mutant strains are defective in glycosylation. Four of the synthetic lethal strains showed resistance to vanadate, (3d-2, 12a-1, 13a-1, and 13d-3). Only 13d-3 showed an increased sensitivity to FK 506.

Four mutant strains (3d-2, 12a-1, 13a-1, and 13d-3) exhibit simultaneous growth resistance to vanadate and sensitivity to Hygromycin B, which is characteristic of mutants defective in Golgi-specific glycosylation (35). According to Dean (34) even mutants with defects in the early steps of glycosylation are sensitive to Hygromycin B. Only the 12e-1 mutant strain showed hypersensitivity to tunicamycin. This is also the only mutant that is not resistant to vanadate.

Isolation of Complementing Clone and Database Search.

The Nikkomycin Z hypersensitivity of 13a-1 and 13d-3 was used to facilitate the cloning of the complementing insert from the URA3-marked CEN4 library. Initial sequence data obtained from the complementing insert was used to search existing databases. The gene mutated in 13d-3 was identified as FKS1, which is identical to ETG1 (36), CWH53 (37), PBR1 (33), and CND1 (38). The FKS1 gene encodes a subunit of the S. cerevisiae β-1,3-glucan synthase.

For 13a-1, a blast query identified a sequence present on chromosome V. Because the library inserts average 9 kb in length, most contain several genes. The entire DNA sequence of this clone was downloaded and analyzed for ORFs. Using subcloned portions of the insert followed by complementation analysis, the mutated gene in 13a-1 was determined to be ANP1.

Unlike 13a-1 and 13d-3, 12a-1 showed no sensitivity to Nikkomycin Z. The sensitivity of 12a-1 to Calcofluor white, in addition to the acquisition of the ability to grow on SD without uracil was used to select transformants with a complementing insert. Subcloning and complementation analysis resulted in the identification of the complementing gene in 12a-1 as SRV2 (CAP1 and END14), which has been implicated in the transmission of cAMP-mediated signals via the RAS/adenylyl cyclase pathway (39–41) and more recently shown to bind to the Src homology 3 domain of the actin binding protein, Abp1p. The SRV2 gene may provide a link between growth signals and the cytoskeleton (42).

Complementation Analysis.

Plasmids carrying the isolated complementing gene were used to verify the mutant phenotypes as well as additional members of the complementation group. The plasmid p13d-3, containing the wild-type clone of FKS1, was used to transform all five synthetic lethal mutant strains, including 13d-3. The plasmid p13d-3 complemented the sectoring phenotype and drug profile of the original mutant, 13d-3, as well as two additional mutants, 3d-2 and 12e-1. When p12a-1 or p13a-1 were used to transform the putative synthetic lethal strains, only the original mutant strain was complemented for both the sectoring phenotype and drug profile. Thus, we have identified at least three different complementation groups as being essential in combination with a CHS3 deletion: FKS1 (three members), ANP1 (one member), and SRV2 (one member).

Cell Wall Composition.

The cell wall composition of the mutant strains might suggest the cause of the synthetic lethal interaction with chs3. Strains were grown in YPD supplemented with radioactive glucose. Cells were harvested and fractionated (see Materials and Methods), and the radioactivity incorporated into the alkali-insoluble fraction and the mannan cell wall fraction was determined (Table 1). The incorporation of label into the alkali-insoluble fraction, containing alkali-insoluble glucan and chitin, was 20–30% lower for the fks1 mutant strains than for the control strains PRY485 and PRY487. The fks1 strains showed a level of incorporation of label into mannan comparable to the control strains. Incorporation of label into the alkali-insoluble fraction for the anp1 strain was approximately equal to the control strains, but incorporation of label into mannan was reduced by 75%. The cell wall composition data for the srv2 strain was comparable to the control strains. Chitin measurements and their significance are discussed below.

Table 1.

Cell wall composition

Strain	Mutation	[14C]glucose incorporated	Mannan, %
		Alkali-insoluble fraction, %
PRY487	Wild type	30	18
PRY485	chs3	27	21
13d-3	fks1	20	21
12e-1	fks1	23	17
3d-2	fks1	20	21
12a-1	srv2	32	19
13a-1	anp1	32	5

Results are expressed as percent incorporation of radioacivity from [14C]glucose into cell wall polysaccharides (cpm incorporated per fraction/total counts per min incorporated) × 100. 

β-1,3-Glucan Synthase Activity.

Glucan is a major component of the yeast cell wall. To better understand the synthetic relationship between our mutants and chs3, we measured their β-1,3-glucan synthase activity. Results are summarized in Table 2. Compared with the control strain PRY485, all three members of our fks1 complementation group, (3d-2, 12e-1, and 13d-3), show a significant reduction in glucan synthase activity, the activity being 15–38% of the control. This finding is consistent with previously published data (33). The anp1 mutant strain (13a-1) shows a less marked decrease in enzyme activity (66% of the control). Unlike the fks1 and anp1 strains, the srv2 mutant strain (12a-1) shows an increase in enzyme activity, 126% of the control strain.

Table 2.

β-1,3-glucan synthase activity

Strain	Mutation	Activity (% of control)
PRY485	chs3	≡100
3d-2	fks1	15 (±1)
12e-1	fks1	38 (±6)
13d-3	fks1	21 (±3)
12a-1	srv2	126 (±10)
13a-1	anp1	66 (±3)

β-1,3-glucan synthase activity was determined as described, and units of activity were normalized to protein. Activity is expressed as percent of activity of the control strain PRY485. Results are the mean, where n = 4, and error is the SEM. 

Other Synthetic Lethal Interactions.

Our mutant screen has shown that chs3∷LEU2 is synthetically lethal with anp1. The ANP1 gene is homologous to VAN1 and MNN9. Together, the three constitute a family of genes required for proper Golgi function in S. cerevisiae (43). To determine whether or not the other members of this gene family have the same synthetic lethal interaction with chs3, heterozygous diploids were constructed with mnn9 and van1. After tetrad dissection, we examined the segregation of markers and the viability of the spores. The results show that chs3 is indeed lethal in combination with both mnn9 and van1.

The synthetic lethality of the fks1 complementation group along with the Nikkomycin Z hypersensitivity for two of the three mutants suggests that a simultaneous decrease in glucan and chitin is lethal. If true, one would expect fks1 strains to be lethal in combination with other mutations that result in a loss of chitin synthesized by Chs3p. To test this, strain PRY581 carrying the original mutant fks1 from 13d-3 (but not the chs3∷LEU2 disruption) was mated to the chs4–4∷LEU2 strain PRY582, the chs3∷LEU2 strain PRY502, and the csd3/chs6∷TRP1 strain PRY404. Diploids were dissected and the tetrads were analyzed for segregation of markers and viability. For PRY581/PRY582, 12 double mutant (chs4∷LEU2 fksl) spores were expected. Based on the lack of Leu⁺ Nikkomycin Z hypersensitive spores, no double mutants were recovered. Therefore, fks1 is synthetically lethal in combination with a deletion of CHS4. The cross of PRY581 (fksl) to PRY502 (chs3∷LEU2) is a reconstruction of the original synthetic lethal strain. For PRY581/PRY502, 12 double mutant spores (chs3∷LEU2 fks1) were expected but none were recovered. This result confirms the synthetic lethality of the double mutant, chs3 fks1. The synthetic lethality of fks1 with both chs3 and chs4 supports the idea that a simultaneous reduction in β-(1,3)-glucan and chitin (specifically chitin synthesized by CSIII) is lethal.

For PRY581/PRY404, of the eight double mutant (csd3/chs6∷TRP1 fks1) spores expected, five were viable. Unlike chs3 and chs4 strains, csd3/chs6 deletion strains do have chitin synthase activity in vitro. Three tetrads were tested for their ability to grow on YPD plus Calcofluor. Of the 12 spores tested, 10 were Calcofluor sensitive. The excess of Calcofluor-sensitive spores is consistent with the idea that Chs3p is functional in these strains. When analyzed for chitin content the double mutants were found to have the high level of chitin found in fks1 mutants (Table 3), in striking contrast to csd3/chs6 mutants that have little, if any, chitin made by the CHS3 complex.

Table 3.

Chitin levels in fks1-csd3/chs6 double mutants

Strain	Chitin, nmoles on GlcNAc/mg
PRY224  (WT)	6.7
PRY223  (WT)	8.4
PRY512  (chs3)	0.9
PRY514  (chs4)	1.1
PRY513  (csd3/chs6)	1.9
PRY404  (csd3/chs6)	1.9
PRY581  (fks1)	23.6
CBY108.1  (fks1)	32.4
112.6A  (fks1, csd3/chs6)	47.5
112.10C  (fks1, csd3/chs6)	27.1
112.9B  (fks1, csd3/chs6)	28.1

Chitin levels for cells grown in YPD were measured as described by Bulawa (2) and are expressed as nmoles of acetylglucosamine per mg of cell wet weight. WT, wild type. 

DISCUSSION

MNN9 Family.

The type II membrane proteins Anp1p, Van1p, and Mnn9p, are required for proper Golgi glycoprotein processing in S. cerevisiae (44). We found that chs3 is lethal in combination with deletions of any of the three. Jungmann and Munro (44) have shown that Anp1p, Van1p, and Mnn9p colocalize within the cis Golgi and form two distinct complexes, each complex containing Mnn9p and either Anp1p or Van1p. Both complexes have α-1,6-mannosyltransferase activity. As expected, our anp1 mutant 13a-1 is resistant to vanadate and sensitive to Hygromycin B. In addition, strain 13a-1 is hypersensitive to Nikkomycin Z, which suggests the need for chitin in these mutants. Determination of β-1,3-glucan synthase activity in the 13a-1 strain showed a reduced level of enzyme activity relative to the control. For this strain, there is a concomitant glycosylation defect along with a decrease in enzyme activity.

Recently, Mondesert et al. (45) reported the isolation of morphogenesis checkpoint-dependent (mcd) mutants that are defective in growth but have normal actin organization. One of the mutated genes they cloned and identified as ANP1. Mondesert et al. suggest that N-linked glycosylation is needed not only for the mannoproteins required for cell wall construction during bud emergence but also to direct secretion to the presumptive site of bud emergence and to the emerging bud. Consequently the lethality of chs3 and anp1 may be caused by a weakened cell wall, which results from the general decrease in mannan or might be caused by a need for N-glycosylation of protein/proteins as part of a signaling pathway involved in polarization of secretion during the cell cycle.

SRV2.

SRV2 encodes a 526-aa protein that has at least three functional domains. The N-terminal domain (amino acids 1–192) binds to adenylyl cyclase and is necessary and sufficient for the phenotypes associated with activated regulatory gene RAS. SRV2 is required for RAS-activated adenylyl cyclase activity (39) but mutations that make cell viability independent of the production of cAMP do not suppress the lethality of null alleles. Therefore SRV2 must provide an essential function to the cell that is independent of the production of cAMP. The C-terminal domain is required for normal cellular morphology and response to nutrient extremes (40). The C-terminal domain binds to monomeric actin and has a cytoskeletal regulatory function in vivo (46). The middle third of SRV2 contains a proline-rich region (amino acids 273–286), which has been shown to bind to the Src homology 3 domain of the actin binding protein, Abp1p (42). SRV2 may play a role in maintaining the integrity of cellular membranes because it is localized to cell membranes (47) and cortical actin patches (42) that are thought to be regions of the cell that are actively growing. Because current speculation favors the concept that cell wall synthesis is controlled by the PKC1 and the PBS2-HOG1 regulatory pathways, it will be important to determine where the Srv2p interacts with these two pathways. That it does indeed interact is suggested by the abnormally swollen cell morphology of srv2 null mutants (40). The cell wall composition analysis shows that cell walls of the 12a-1 strain carrying srv2 are similar to those of the control strain, suggesting that the synthetic lethality is not caused by a simple decrease in wall polymers.

We have shown previously (4) that under certain growth conditions, specifically growth on YPD, the double mutant chs2 chs3 is lethal. If the Srv2p is required for proper localization and/or functioning of the Chs2p, strains carrying a mutated srv2 gene would require a functional CHS3 gene for viability. Another possible explanation for the data would be that Srv2p is required for proper localization and functioning of the 10-nm filaments encoded by the cell division cycle genes CDC3, CDC10, CDC11, and CDC12. The results of our unpublished tetrad analyses, done for crosses of a chs3 deletion strain with cdc3, cdc10, cdc11, or cdc12, suggest that chs3 is lethal in combination with cdc3, cdc11, and cdc12 but not with cdc10. This result is interesting and suggests that the principle function of Cdc10p may be binding of the Chs3p complex to the septin ring, and that it otherwise may be dispensable.

FKS1.

We were not surprised to find three fks1 mutants among our synthetically lethal isolates, because lack of both chitin and part of the cell wall β-1,3-glucan should be fatal. FKS1 encodes a subunit of one of the two enzymes that synthesize 1,3-β-glucan, a major structural component of the cell wall. Mutations in this gene have been isolated in other screens designed to identify cell wall alterations, e.g., hypersensitivity to FK506 (48), resistance to Echinocandins (49, 50), and sensitivity to Calcofluor white (37). As expected, all three of our fks1 mutants have reduced β-1,3-glucan synthase activity relative to the control strain PRY485. The residual enzyme activity is thought to be catalyzed by FKS2, a homolog of FKS1 (Fks2p is 88% identical to Fks1p). In continuing this investigation we found, as expected, that deletion of CHS4, a probable second subunit of the Chs3p enzyme complex, is also lethal in fks1 mutants.

At least seven known Saccharomyces mutations that weaken the cell wall lead to substantial increases in the level of chitin deposited in the lateral wall, probably by Chs3p (51). Fks1 and gas1 (52) mutants are among the most prominent strains showing increased chitin synthesis, with chitin levels as much as an order of magnitude higher than those found in wild-type cells. Deposition of high tensile strength chitin fibers in the wall may be a physiologically important “stress response” programmed into the cell, a stereotyped response that probably should be differentiated from “delocalized” chitin synthesis. The latter frequently is associated with disorganization of the actin cytoskeleton (53).

Although nothing is known about the mechanism of stress response chitin synthesis, one would assume that it proceeds by the normal pathways of Chs3p targeting. This pathway involves exit of Chs3p from the endoplasmic reticulum under the control of Chs7p (17), movement to chitosomes, and the plasma membrane mediated by Csd3/Chs6p (16) and probably Chs5p, and then recycling from the plasma membrane to targeted areas such as the site of incipient bud formation and the lateral wall through an endocytic route in vesicles that contain the syntaxin Tlg1p (54).

In spite of strong evidence supporting the operation of this pathway under normal conditions, the elevated chitin level in the csd3/chs6 fks1 double mutant demonstrates that the normal role of CSD3/CHS6 is being bypassed where increased chitin synthesis is required. Clearly an alternate means of transport and activation is being used. Definition of this pathway and its regulation represents a new and exciting area of investigation.

ABBREVIATIONS

YPD

yeast extract/peptone/dextrose

SD

synthetic minimal medium