Identification of SSF1, CNS1, and HCH1 as multicopy suppressors of a Saccharomyces cerevisiae Hsp90 loss-of-function mutation

Abstract

Hsp90 functions in a multicomponent chaperone system to promote the maturation and maintenance of a diverse, but specific, set of target proteins that play key roles in the regulation of cell growth and development. To identify additional components of the Hsp90 chaperone system and its targets, we searched for multicopy suppressors of various temperature-sensitive mutations in the yeast Hsp90 gene, HSP82. Three suppressors were isolated for one Hsp90 mutant (glutamate → lysine at amino acid 381). Each exhibited a unique, allele-specific pattern of suppression with other Hsp90 mutants and had unique structural and biological properties. SSF1 is a member of an essential gene family and functions in the response to mating pheromones. CNS1 is an essential gene that encodes a component of the Hsp90 chaperone machinery. The role of HCH1 is unknown; its sequence has no strong homology to any protein of known function. SSF1 and CNS1 were weak suppressors, whereas HCH1 restored wild-type growth rates at all temperatures tested to cells expressing the E381K mutant. Overexpression of CNS1 or HCH1, but not SSF1, enhanced the maturation of a heterologous Hsp90 target protein, p60^v-src. These results suggest that like Cns1p, Hch1p is a general modulator of Hsp90 chaperone functions, whereas Ssf1p likely is either an Hsp90 target protein or functions in the same pathway as an Hsp90 target protein.

The structure and function of the Hsp90 family of molecular chaperones has been conserved for over one billion years of evolution. Hsp90 proteins from organisms as diverse as the budding yeast, Saccharomyces cerevisiae and humans share 60% amino acid identity. Moreover, the crystal structures of the N-terminal domains of the S. cerevisiae and bovine Hsp90 proteins are superimposable (1–3). Hsp90 is essential for viability in all three of the eukaryotic organisms in which it has been tested, S. cerevisiae (4), Drosophila melanogaster (5), and Schizosaccharomyces pombe (6). Both the Drosophila and human family members can substitute for the endogenous Hsp90 of S. cerevisiae (7–8), and the yeast Hsp90 proteins can effectively chaperone heterologously expressed target proteins from other organisms (8–12). Furthermore, Hsp90 functions as part of a dynamic chaperone complex, the other known components of which are also highly conserved (13–16).

In vitro, Hsp90 exhibits an intrinsic, general chaperone activity, preventing the aggregation of inherently unstable or chemically denatured proteins (17–23). In such assays, Hsp90 is required in large excess and cannot refold denatured proteins. Rather, it maintains these proteins in nonaggregated, folding-competent conformations that can be acted on by other chaperones (18, 20, 21, 23, 24).

In vivo, at normal temperatures, Hsp90 is a highly selective chaperone, with even closely related proteins such as p60^c-src and p60^v-src varying greatly in their dependence on its function (12, 25). Most known Hsp90 target proteins are components of signal transduction pathways, and include steroid hormone receptors, serine/threonine kinases, tyrosine kinases, and transcription factors (13–16). The functions of signaling proteins require them to exist in inactive conformations that are poised for immediate activation. Biochemical and genetic data from many laboratories suggest that Hsp90 helps its target proteins achieve and maintain the inherently unstable conformations that are required for their function (8, 10, 12, 14–16).

Perhaps the most extensively studied chaperone function of Hsp90 is its role in the maturation of the progesterone (PR) and glucocorticoid (GR) receptors, with which Hsp90 forms stoichiometric complexes. Association with Hsp90 and several other cochaperones serves to keep these receptors in inactive conformations that can bind ligand with high affinity (8, 10, 14–16). Studies in reticulocyte lysates indicate that PR maturation involves at least nine proteins and occurs through a cycle of sequential interactions with distinct chaperone complexes, several of which contain Hsp90 (13, 26).

In the yeast S. cerevisiae, two nearly identical genes, HSP82 and HSC82, encode Hsp90 proteins. Although Hsp90 is extremely abundant in yeast, its levels can be reduced at least one order of magnitude without affecting cell viability at 25°C. Much higher concentrations are required for growth at higher temperatures (4). Previously, we isolated eight S. cerevisiae Hsp90 temperature-sensitive (ts) mutants (10), each of which contains a single amino acid substitution and has a distinct phenotype with respect to cell morphology and growth. As is the case with most ts mutations in heat-shock proteins, seven of the eight are partial loss-of-function mutants: they have reduced activity at all temperatures, and are ts because they cannot fulfill the increased demand for Hsp90 function at high temperatures. One, however, behaves as a classic ts mutant: it has nearly wild-type activity at 25°C, but rapidly loses function on shift to high temperature. This mutant, G170D, contains a substitution of aspartic acid for glycine at amino acid position 170 (10). The instability of G170D at high temperatures is consistent with the location of residue 170 in the hydrophobic core of the N-terminal domain of Hsp90 (1–3).

The unique G170D mutant allowed us to examine the breadth of Hsp90’s chaperone functions in vivo. In temperature-shift experiments, Hsp90 does not appear to protect proteins from thermal inactivation. Hsp90 does, however, facilitate the refolding of a heterologous heat-damaged test substrate, firefly luciferase (11). Pharmacological inhibition of vertebrate Hsp90 produced similar results (18). These data suggest that the function of Hsp90 during stress is to interact with proteins as they unfold and to maintain them in conformations that are amenable to rapid refolding after the stress has been removed. This function may be related to the above-mentioned general chaperone activity of Hsp90 with unfolded test substrates in vitro.

Experiments with G170D also demonstrated that Hsp90 is not required for the de novo folding of most cellular proteins in yeast. Rather, Hsp90 functions in the maturation and maintenance of certain target proteins that have difficulty achieving stable structures (10, 11). Presumably, yeast cells require Hsp90 function because one or more Hsp90 target proteins are essential for viability. Given the diversity of mammalian Hsp90 target proteins (13–16) and the distinct phenotypes of different S. cerevisiae Hsp90 point mutants (9, 10, 27), it seems likely that the yeast Hsp90 target proteins will include a wide variety of signal transducers that function in many pathways. Indeed, one such target has recently been identified; Ste11p is a MEK kinase involved in several MAP kinase signaling pathways in yeast (ref. 28; B. Errede, personal communication). The diversity of Hsp90 target proteins and their central importance as regulators of growth and development makes Hsp90 a particularly interesting chaperone to study. However, the number and range of its targets and the multicomponent, dynamic nature of the chaperone complexes in which it functions present formidable problems in elucidating its molecular mechanism.

We and others have employed a variety of biochemical methods to isolate and identify proteins that interact with Hsp90. Although these methods have been very successful, they are only applicable to proteins that are fairly abundant and form stable complexes with Hsp90 or its known substrates (13, 15, 16, 29). Genetic screens for enhancers or suppressors of Hsp90 mutations provide another tool for identifying proteins that function in the same pathway. Indeed, ydj1, a member of the dnaJ family of chaperones, was identified in a screen for mutations that exhibit synthetic lethality with G170D. This finding in turn led to the discovery that Ydj1p plays an important role in the maturation of Hsp90 target proteins (30, 31).

Here we continued the search for Hsp90 target proteins and cochaperones, searching for suppressors of yeast Hsp90 ts point mutants. Because reduction of Hsp90 function would affect the activity of several Hsp90 target proteins, it seemed likely that overexpression of a single target would have only subtle effects on cell growth. Furthermore, overexpression of the Hsp90 cochaperone STI1 suppresses the ts growth phenotype of some Hsp90 mutants, but it exacerbates the growth defects of others (32). Therefore, we searched for suppressors of four phenotypically distinct mutants each containing a single amino acid substitution in a different region of Hsp90 and performed our screens at the lowest possible nonpermissive temperature. This approach allowed us to identify three additional members of the Hsp90 chaperone network, each with unique properties.

MATERIALS AND METHODS

Plasmid and Strain Construction.

Integrating plasmids containing wild-type or mutant HSP82 sequences regulated by the strong constitutive promoter GPD were constructed by subcloning the ClaI-NotI fragments from the corresponding single-copy expression plasmids (10) into pRS303 (34). These plasmids are designated piHGpd/P82 (wild-type HSP82), piHGpd/T22I (threonine at amino acid residue 22 replaced by isoleucine), piHGpd/A41V, piHGpd/G81S, piHGpd/T101I, piHGpd/G170D, piHGpd/G313S, piHGpd/E381K and piHGpd/A587T.

Strains iP82a, iT22Ia (T22I), iA41Va (A41V), iG81Sa (G81S), iT101Ia (T101I), iG170Da (G170D), iG313Sa (G313S), iE381Ka (E381K), and iA587Ta (A587T) are derivatives of W303 (MATa ade2–1 can1–100 his3–12, 16 leu2–3, 112 trp1–1 ura3–1; R. Rothstein, Columbia University). In these strains, the endogenous Hsp90 genes, HSC82 and HSP82 are deleted and wild-type or mutant HSP82 sequences, regulated by the strong constitutive promoter GPD, are integrated into the genome. To create these strains, the integrating plasmids described above were linearized with NheI and transformed into ΔPCLDa (10). Transformants were grown on 5-fluoroorotic acid (5-FOA) medium to select for those that had lost the wild-type HSC82 expression plasmid from ΔPCLDa. The ilep1a strain was constructed in a similar manner except that in this case wild-type HSP82 sequences under the control of a low-expression promoter were integrated into the genome (29).

Yeast and bacterial transformations were obtained by using lithium acetate (35) and calcium chloride (36) methods, respectively, or by using electroporation with a Gene Pulser (Bio-Rad).

Yeast Media.

Cells were cultured in SD, SR, or Sgal (20 g/liter glucose, raffinose, or galactose, respectively, with 5 g/liter ammonium sulfate and 1.7 g/liter yeast nitrogen base without amino acids, supplemented with essential amino acids and nucleotides) or 5-FOA (1.7 g/l yeast nitrogen base without amino acids and ammonium sulfate, 1 g/l proline, 20 g/liter glucose, 10 mg/liter uracil, and 600 mg/liter 5-FOA, supplemented with essential amino acids and nucleotides (37). Solid media contained 20 g/liter Difco Bacto agar.

Screen for Multicopy Suppressors of Hsp90 ts Mutants.

Hsp90 ts mutant strains T101I, G170D, G313S, and E381K were transformed with 1 μg of a genomic library (gift of Y. Kimura, Tokyo Metropolitan Institute of Medical Science; library was constructed from strain 5CG2 (27) in pTV3 (38), a high copy number plasmid marked with TRP1). Transformants that grew on SD minus tryptophan at the lowest temperature that was nonpermissive for growth (36°C for G170D and E381K, 37°C for T101I and G313S) were selected and retested. To ensure that the suppression was caused by the plasmid DNA, plasmids were recovered, transformed into the appropriate strain, and tested for growth at the nonpermissive temperature.

The inserts in the three suppressor plasmids, E20, E26 and E28 were sequenced by the University of Chicago Cancer Research Center DNA Sequencing Facility by using primers specific to vector sequences on either side of the insert. Sequences were compared with those in the yeast genome by using the Saccharomyces Genome Database World Wide Web site (http://genome-www.stanford.edu/Saccharomyces/). All three suppressor plasmids contained multiple ORFs. To identify the specific genes responsible for the suppression, individual ORFs from each suppressor were subcloned into pTV3 and tested for the ability to suppress the growth defect of E381K cells. The original suppressor plasmids E20, E26, and E28 were used in all of the experiments described in this paper.

Assessment of p60^v-src Activity and Accumulation.

The galactose-inducible p60^v-src expression plasmid Y316v-src (39) was transformed into E381K cells containing either a plasmid encoding wild-type Hsp82 (pTGpd/P82; ref. 10), an empty vector (pTV3), or one of the three suppressor plasmids E20, E26, or E28. Activity and accumulation of p60^v-src was assessed as described (10).

RESULTS

Isolation of Multicopy Suppressors of Hsp90 ts Mutants.

Hsp90 interacts with its target proteins in the context of a dynamic, multicomponent chaperone complex (13–16). Structurally, Hsp90 contains multiple domains that presumably interact with different proteins in different ways. Indeed, Bohen and Yamamoto (9) isolated a mutation in the S. cerevisae HSP82 gene, E431K, that specifically affects interactions with a single target protein, GR. Therefore, rather than searching for suppressors of general Hsp90 loss-of-function mutations (deletions and promoter rearrangements), we instead searched for suppressors of four phenotypically distinct ts mutants, each containing a single amino acid substitution in a different region of Hsp90 (T101I, G170D, G313S, and E381K).

Because all of these mutations are recessive (10), a large number of the plasmids recovered in a suppressor screen would likely encode one of the two yeast Hsp90 proteins, Hsp82p and Hsc82p. To prevent this, we used a genomic library constructed from a strain in which HSC82 had been disrupted and HSP82 had been replaced with a construct containing HSP82 coding sequences under the control of the galactose-inducible promoter GAL1 (27). Because HSP82 and HSC82 are an essential gene family in yeast, this strain is only viable when grown in galactose media. More importantly, the only wild-type Hsp90 genes present in the genomic library constructed from this strain are under the control of the GAL1 promoter. By performing our screens in glucose media, which represses the GAL1 promoter, we avoided recovery of the wild-type Hsp90 genes present in this library.

To ensure that weak suppressors would not be missed, we looked for transformants that could grow at the lowest nonpermissive temperature for each of the hsp82 point mutants. G170D and E381K were screened at 36°C, whereas T101I and G313S were screened at 37°C. Approximately 5,000 transformants were tested for their ability to suppress each of the four mutants. No suppressors of T101I, G170D, or G313S were isolated. We did, however, identify three plasmids (E20, E26, and E28) that suppressed E381K.

Identification of Suppressor Proteins.

Sequence analysis revealed that E20 contained 3.4 kb of chromosome 8 (228060–231468) (Fig. 1). This region encodes only one complete ORF, SSF1, but also contains the first 1,104 bp of the 1,631-bp ORF RRP3 and the first 497 bp of the 842-bp ORF YHR067w. E26 contained 4.8 kb of chromosome 2 (546677–551480). This region encodes three complete ORFs, RIB7, RPB5, and CNS1. It also contains the last 528 bp of the 875-bp ORF SPP381 and the first 422 bp of the 2,096-bp ORF YBR156c. E28 contained 6.3 kb of chromosome 14 (105358–111625). This region encodes three complete ORFs, POP3, YNL281w, and ERG24. It also contains the last 1,335 bp of the 1,511-bp ORF WSC2 and the first 710 bp of the 1,275-bp ORF YNL279w.

Figure 1.

Maps of individual suppressor plasmids and the fragments tested for suppressor activity. Fragments of (Top) E20, (Middle) E26, and (Bottom) E28 that provided suppressor activity are shown.

The specific genes responsible for the suppression were identified by subcloning individual ORFs from each suppressor (Fig. 1). A ScaI-KpnI fragment containing the entire ORF of SSF1, 275 bases of upstream sequences (including the first 103 bases of YHRO65c) and 16 bases of downstream sequences contained the suppressing function of the E20 plasmid. An EcoRI fragment containing the entire CNS1 ORF, 696 bp of 5′-untranslated sequences and 175 bp of 3′-untranslated sequences possessed the suppressing function of the E26 plasmid. An NcoI-MscI fragment containing the entire YNL281w (HCH1–High Copy Hsp90 suppressor) ORF and 398 bp of upstream sequences (including the last 207 bp of YNL282w) and 333 bp of downstream sequences (including the last 166 bp of ERG24) contained the suppressing activity of plasmid E28. This analysis revealed that the three multicopy suppressors of E381K were SSF1, CNS1, and HCH1.

SSF1, CNS1, and HCH1 Suppress the ts Growth Phenotype of E381K Cells to Different Extents.

E381K cells have a severe growth defect. They grow at half the rate of wild-type yeast at 25°C, even more slowly at intermediate temperatures (34–35°C), and not at all above 35°C (10). The E381K mutation is recessive; introduction of a plasmid encoding wild-type HSP82 into E381K cells restored normal growth rates at all temperatures tested (25–37°C) (Fig. 2, and data not shown). A plasmid encoding HCH1 also restored a wild-type growth rate to E381K cells up to 37°C, the highest temperature tested. Plasmids encoding either SSF1 or CNS1 increased the maximal growth temperature of E381K cells to 36°C and resulted in faster, but still not wild-type, growth rates at temperatures between 25°C and 35°C. Haploid E381K cells of both mating types responded equally to the suppressors. Diploid E381K cells, however, did not respond as well to the CNS1 and HCH1 plasmids, and their growth was completely unaffected by the SSF1 plasmid (Table 1).

Figure 2.

Effects of SSF1, CNS1, and HCH1 on the growth of E381K cells. E381K cells containing an empty vector or a vector expressing HSP82, SSF1, CNS1, or HCH1 were grown at either 25°C (Right) or 35°C (Left) for 3 days on SD medium lacking tryptophan.

Table 1.

Suppression of haploid and diploid E381K cells by SSF1, CNS1, and HCH1

Plasmid	Mata	Matα	Mata/α
Vector	−	−	−
HSP82 (Hsp90)	++++	++++	++++
SSF1	+	+	−
CNS1	+	+	+/−
HCH1	++++	++++	++

SSF1, CNS1, and HCH1 are Allele-Specific Suppressors.

Next, we asked whether overexpression of SSF1, CNS1, or HCH1 could suppress the growth defects of the other seven ts point mutants in our collection. We also examined the effect of the suppressors on the growth of cells that express high levels of wild-type HSP82 and are not ts (strain iP82a) and of cells that express very low levels of wild-type HSP82 and are therefore ts (strain ilep1a). Growth was assessed at permissive (25°C), semipermissive (34–36°C, depending on the mutant), and nonpermissive (36–37°C, depending on the mutant) temperatures.

The growth of iP82a cells was not affected by any of the suppressors at any temperature tested (Table 2, and data not shown). Introduction of the plasmid encoding SSF1 enhanced the growth of another mutant, A41V, at semipermissive temperatures, but had no effect on its maximal growth temperature. SSF1 had no effect on the growth of any of the other six point mutants in our collection, nor did it affect the growth of ilep1a cells. Overexpression of CNS1 also affected only one other mutant, T22I. The growth rate of T22I cells was increased at 25°C, but remained unchanged at higher temperatures. In marked contrast, overexpression of HCH1 had profound effects on the growth of most of the mutants in our collection. It reduced the growth rates of four mutants, A41V, G170D, G313S, A587T, and of ilep1a cells. It had no effect on two mutants, G81S and T101I, and suppressed the growth defect of one other mutant, T22I. As with E381K, overexpression of HCH1 completely suppressed the growth defect of T22I cells at all temperatures tested (Table 2, and data not shown). Thus, each suppressor exhibits a distinct pattern of allele specificity.

Table 2.

Allele-specific effects of SSF1, CNS1, and HCH1 on the growth of Hsp82 mutants

Mutant	SSF1	CNS1	HCH1	STI1
E381K	↑	↑	⇈	−
iP82a	−	−	−
T22I	−	↑	⇈	↑
A41V	↑	−	↓	↓
G81S	−	−	−	−
T101I	−	−	−	↓
G170D	−	−	↓	↑
G313S	−	−	↓	↑
A587T	−	−	↓	↑
ilep1a	−	−	↓

⇈, Greatly enhanced growth; ↑, somewhat enhanced growth; −, no effect; ↓, inhibited growth. 

Overexpression of CNS1 or HCH1 Increases the Activity of a Heterologous Target Protein in E381K Cells.

The primary function of Hsp90 during normal cell growth appears to be to promote and maintain the activity of a highly selective group of target proteins (11). This function is aided by at least nine other proteins that form the Hsp90 chaperone machine (13–16). Therefore, it seemed likely that the suppressors we isolated were either (i) Hsp90 target proteins or proteins that function in the same biochemical pathway as an Hsp90 target protein or (ii) proteins that function in or modulate the activity of the dynamic Hsp90 chaperone complex. The biological activities of the former should be affected by Hsp90 mutations; the latter should affect the activity of other Hsp90 target proteins. Because the functions of Cns1p, Ssf1p, and Hch1p are not completely understood, we were unable to directly determine whether their activities are affected by Hsp90 mutations. We could, however, determine whether Cns1p, Ssf1p, or Hch1p affected the capacity of Hsp90 to function in the proper maturation of a known target protein.

To do so we asked whether the overexpression of these proteins enhanced the maturation the oncogenic tyrosine kinase p60^v-src. This target protein was chosen for three reasons. First, its maturation is highly dependent on Hsp90 (12). Second, E381K cells have a severe defect in p60^v-src maturation, even at 25°C (10). Third, yeast cells have very low levels of endogenous tyrosine kinase activity, making p60^v-src activity easy to detect by Western blot analysis with an anti-phosphotyrosine antibody.

E381K cells containing either an empty vector, one of the three suppressors, or wild-type HSP82 were transformed with a GAL1-p60^v-src expression plasmid. Tyrosine phosphorylation of total cellular proteins was assessed 6 hours after raffinose-grown cells were shifted to galactose-containing media. As previously described (10, 12), in cells expressing wild-type HSP82, p60^v-src promiscuously phosphorylated a broad array of yeast proteins (Fig. 3). As expected, the activity of p60^v-src was severely reduced in E381K cells containing the empty vector. This reduction in activity was accompanied by a decrease in p60^v-src accumulation (Fig. 3), consistent with previous observations that Hsp90 plays a role in both the stability and activation of p60^v-src (10, 12). Introduction of the SSF1 plasmid into E381K cells had no effect on either p60^v-src activity or accumulation. In marked contrast, overexpression of either CNS1 or HCH1 increased p60^v-src activity. In both cases, the increased activity was accompanied by increased accumulation of p60^v-src protein. The ability of CNS1 and HCH1 to both promote the maturation of a heterologous Hsp90 target protein in E381K cells and enhance the growth of E381K cells in the absence of p60^v-src indicates they increase general Hsp90 function. This, in turn, strongly suggests that these suppressor proteins are either components of the Hsp90 chaperone machine or modulators of its activity.

Figure 3.

Effects of SSF1, CNS1, and HCH1 on p60^v-src activity in E381K cells. E381K cells expressing an empty vector or a vector expressing HSP82, SSF1, CNS1, or HCH1 were transformed with a galactose-inducible p60^v-src expression vector. Cells grown to mid-logarithmic phase at 25°C in raffinose-containing medium were shifted to galactose-containing medium for 6 hours to induce expression of p60^v-src. Total cellular proteins were isolated, separated on 7.5% SDS/PAGE gels and transferred to Immobilon-P membranes. (A) Equal sample loading was confirmed by Coomassie blue staining. (B) p60^v-src activity was assessed by reacting the blot with an antibody specific for phosphotyrosine. (C) p60^v-src accumulation was assessed by reacting the blot with an antibody specific for p60^v-src.

DISCUSSION

The chaperone activities of Hsp90 are believed to be essential for the maturation of a specific set of critical target proteins. The identities of these target proteins, however, are just beginning to be elucidated. Moreover, it is by no means clear whether we have yet identified all, or even most, of the cochaperones and modulators of Hsp90 function. In an attempt to identify Hsp90 target proteins and modulators of Hsp90 function in yeast, we searched for genes that, when overexpressed, would suppress the ts growth defect of several distinct Hsp90 mutants. We identified three multicopy suppressors, SSF1, HCH1, and CNS1. Each suppressor produced a distinct pattern of allele specificity and each had a different character. Our data suggest that SSF1 is likely either to be an Hsp90 target protein or to function in the same pathway as an Hsp90 target protein, whereas CNS1 and HCH1 likely are involved in regulating or promoting general Hsp90 chaperone function.

SSF1: A Potential Hsp90 Target Protein.

SSF1 was originally isolated in a screen for proteins that, when overexpressed, increase the mating efficiency of a ste4 ts mutant (40). A closely related gene, SSF2, (the two proteins share 94% amino acid identity), also was identified. Null mutations in either SSF1 or SSF2 have no detectable phenotype, but mutations in both genes are lethal (40). Depletion of Ssf gene products from growing cultures significantly impaired the ability of the cells to mate (40). Recently, it was determined that this mating defect is a consequence of a failure to carry out the morphological response to pheromone (41). Thus, Ssf1p has essential functions during both mating and vegetative growth.

Overexpression of SSF1 did not suppress the defect in p60^v-src maturation that is observed in E381K cells, suggesting that Ssf1p does not affect the general chaperone functions of Hsp90. It is possible, however, that Ssf1p is a cochaperone that affects only a subset of Hsp90 functions. Alternatively, Ssf1p, or another protein that functions in the same biochemical pathway as Ssf1p, may be an Hsp90 target protein. This is consistent with the observations that (i) Ssf1p functions in the response to mating pheromones, (ii) the suppressor effects of Ssf1p, but not Cns1p and Hch1p, are specific to haploid cells, (iii) Hsp90 mutants are defective in pheromone signaling (28), and (iv) Ste11p has recently been identified as an Hsp90 target protein (ref. 28; B. Errede, personal communiation). Ste11p is a MEK kinase involved in several MAP kinase-signaling pathways in yeast, including the pheromone response pathway. Thus, overexpression of Ssf1p may suppress the effects of the E381K mutation through effects on one or more of the MAP kinase pathways.

CNS1: An Hsp90 Hetercomplex Member.

CNS1 was identified as an essential ORF on chromonsome 2 (42). Spores containing null mutations in CNS1 germinate but only undergo 2–3 divisions. Interestingly, the cells in these spores have elongated buds (42), a morphology that is remarkably similar to that of certain Hsp90 ts mutants (27). CNS1 codes for a 386-aa protein that contains three tetratricopeptide repeat (TPR) domains. Several members of the Hsp90 heterocomplex, including Sti1p and the immunophilins, contain TPR domains, and these domains are required for their interactions with Hsp90 (16, 43–47).

Overexpression of CNS1 partially suppressed the defect in p60^v-src maturation in E381K cells, suggesting that Cns1p influences the chaperone functions of Hsp90. While this manuscript was in preparation, CNS1 was reported to be a multicopy suppressor of the slow-growth phenotype of cells lacking the Hsp90 cochaperone CPR7 (48, 49). In addition to restoring rapid growth to cpr7 mutant cells, overexpression of CNS1 rescued their defect in Hsp90 target protein maturation (49). Moreover, Cns1p was found in protein complexes with both Hsp90 and Cpr7 (48, 49). Thus, it seems very likely that CNS1 is a member of the Hsp90 chaperone machinery.

HCH1: A Modulator of Hsp90 Function.

HCH1 was identified as a 153-aa ORF (YNL281w) of unknown function. It shares 37% sequence identity and 55% sequence homology to the N-terminal half of a 350-aa protein encoded by an uncharacterized yeast ORF, YDR214w. Otherwise, HCH1 has no striking homology to any known proteins. Disruption of HCH1 produced no obvious growth defect, nor did it exhibit an obvious defect in mating or sporulation (B. Scherens, personal communication).

HCH1 was the strongest suppressor identified in our screen. It completely suppressed the growth defect of both E381K and T22I cells. Interestingly, overexpression of HCH1 exacerbated the growth defect of A41V, G170D, G313S, and A587T cells, and had no effect on G81S and T101I cells. Such allele specificity is reminiscent of, though not identical to, that observed for the Hsp90 cochaperone STI1 (32). Overexpression of STI1 enhances the growth of T22I, G170D, G313S, and A587T, reduces the growth of A41V and T101I, and has no effect on G81S and E381K. Mutations in STI1 inhibit the maturation of diverse Hsp90 target proteins in yeast (32), and Sti1p is found associated with Hsp90 complexes in yeast and mammalian cells (13–16, 32). Thus, for STI1, this type of allele-specificity is consistent with Hsp90 functioning at several steps in target protein maturation and with the point mutations affecting different steps—before, during, or after interaction with Sti1p. The same likely is true for HCH1. Indeed, overexpression of HCH1 enhanced the maturation of the heterologous Hsp90 target protein p60^v-src in E381K cells. Although further investigation of HCH1 and its relative YDR214w is required, our data suggest that Hch1p either is a modulator of Hsp90 function or is itself a chaperone that can affect target protein maturation.

Our approach, employing point mutations in different domains of Hsp90 and taking extreme care to ensure that even subtle suppressors would be isolated, has yielded promising candidates for proteins involved in Hsp90 function. Moreover, the patterns of allele-specific suppression uncovered in this study and in earlier work (32) on STI1 suggest that the various Hsp90 point mutations we have isolated will provide valuable tools for dissecting specific steps in the complex, multicomponent Hsp90 chaperone machinery. Notably, our screens were not saturating: four point mutants were used in this study, but suppressors of only one of them, E381K, were isolated; each suppressor was only obtained once; and finally, we did not isolate STI1, a known multicopy suppressor of the G170D mutation (32). It seems likely, therefore, that further screening using these and other Hsp90 point mutants, particularly mutants distributed on different surfaces of each domain, will provide a powerful mechanism for identifying additional Hsp90 target proteins and cofactors. Furthermore, recent work in Drosophila demonstrates that different genetic backgrounds contain preexisting genetic variants that sensitize different growth and differentiation pathways to Hsp90 defects (50). Searching for suppressors of Hsp90 point mutations in different genetic backgrounds of S. cerevisiae may provide a richer source of target proteins (and proteins that function in target protein pathways) than a search in any individual strain.

ABBREVIATION

ts

temperature sensitive