INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Hsp90, a highly conserved, abundant, and essential cytosolic molecular chaperone, functions in the maturation of a limited set of substrate proteins. Hsp90 was originally found to be involved in the activation of proteins involved in signaling pathways such as steroid hormone receptors and the oncogenic tyrosine kinase v-Src. Recent studies have revealed interaction with the hepatitis B retrovirus, nitric oxide synthase, telomerase, and the mutant form of p53 (7, 23), suggesting that Hsp90 may be involved in the maturation of many diverse substrates.

Hsp90 functions with a number of cochaperones in an apparently dynamic but ordered pathway leading to substrate maturation (48). Much of what is known about the interaction of Hsp90 and cochaperones with substrate proteins comes from studies of steroid hormone receptors (45). Besides promoting receptor activity, Hsp90 and cochaperones prevent steroid receptor activity in the absence of hormone, as receptors bound to Hsp90 cannot bind DNA and are therefore inactive as transcription factors. In vitro studies in reticulocyte lysates have led to the identification of proteins required for steroid receptor activation. Purified receptor is unable to bind hormone, but incubation in reticulocyte lysate results in a form of the receptor bound to Hsp90, p23, and immunophilins of the cyclophilin and FK506-binding protein classes (45). This mature form is capable of binding hormone, at which point chaperones are released, forming transcriptionally active receptor. Other components of the reticulocyte lysate that interact with the receptor in this pathway include Hsp70, Hop, and a member of the Hsp40 family (45). While the specific function of each of these proteins in complex assembly is unclear, Hsp70 and Hsp40 appear to function very early in the pathway.

Steroid receptors have three domains: an amino-terminal activation domain (AF-1), a DNA binding domain, and a carboxy-terminal hormone binding domain (HBD). The HBD contains the hormone binding site, the Hsp90 binding site, and a second activation domain, AF-2. Hormone binding triggers a series of events, including release of Hsp90 and cochaperones, DNA binding, and activation of AF-1 and AF-2. Mutational analysis of the estrogen receptor (ER) and the glucocorticoid receptor (GR) revealed that the two activation domains AF-1 and AF-2 are regulated separately (21, 37). AF-2 is strictly hormone dependent, as hormone binding is necessary to cause the conformation change that exposes the AF-2 site. However, truncation mutants of the ER and the GR that contain the amino-terminal AF-1 but lack the HBD are constitutively active (20, 29, 51). This observation led to the concept that the HBD suppresses AF-1 activity in the absence of hormone. In support of this role, fusion of the HBD of either the ER or the GR to heterologous DNA binding and activation domains conferred both hormonal control and Hsp90 binding (44, 46). The mechanism by which the HBD suppresses AF-1 activity is unknown but is likely a consequence of Hsp90 binding to the HBD.

In vivo studies of Saccharomyces cerevisiae have recapitulated the requirement for Hsp90 and cochaperones in steroid receptor maturation (42). While S. cerevisiae does not contain endogenous steroid receptors, steroid receptors expressed in yeast bind homologs of the proteins mentioned aboveHsp82 (Hsp90), Ssa (Hsp70), Sti1 (Hop), Sba1 (p23), Ydj1 (Hsp40), and Cpr6/7 (Cyp-40)and activate transcription in a hormone-dependent manner (17, 27). Mutations in the yeast genes encoding Hsc82/Hsp82, Sti1, Ydj1, and Sba1 affect the ability of steroid hormone receptors expressed in yeast to activate reporter genes in response to hormone (4, 17, 42). The maturation of v-Src in S. cerevisiae is also affected by mutations in the Hsp90 pathway. v-Src is toxic to wild-type yeast, due to the aberrant tyrosine phosphorylation of yeast proteins. Expression of v-Src in yeast strains containing mutations in hsp82, ydj1, sti1, or sba1 results in decreased lethality and/or decreased phosphotyrosine activity relative to wild-type yeast (17, 27). Therefore, S. cerevisiae is a useful model for studying the interactions of Hsp90 and cochaperones with substrate proteins.

The focus of this report is the role of Ydj1 in Hsp90 pathways. Ydj1 is a member of the Hsp40 class of molecular chaperones, an important class of chaperones found in virtually every organism and cellular compartment (12, 28). Hsp40s regulate the function of partner Hsp70s by stimulating the ATPase activity of Hsp70. In some cases, Hsp40s also direct partner Hsp70s to interact with substrate polypeptides. Ydj1 is known to stimulate the ATP hydrolysis of the cytosolic Hsp70 Ssa and is capable of binding some denatured substrates (14, 35). Ydj1 is not essential in S. cerevisiae, but disruption of YDJ1 results in slow-growing cells that grow poorly at 30°C and very poorly in liquid media at all temperatures (1, 9). Ydj1 is a class 1 DnaJ homolog (12), sharing the general domain structure of Escherichia coli DnaJ, in which the amino-terminal J domain is followed by a G/F-rich region (constituting the J + G/F region), a zinc finger region, and a less conserved carboxy-terminal domain. The amino-terminal J + G/F region is sufficient to stimulate the ATPase activity of Ssa, while the substrate binding region has been localized to the zinc finger region and/or the carboxy-terminal region (35, 49). Ydj1 also possesses a CAAX box, which specifies the posttranslational addition of a farnesyl group to the carboxy terminus of Ydj1, a modification that facilitates attachment of Ydj1 to endoplasmic reticulum membranes (11).

Previous reports have shown differing roles for Ydj1 in Hsp90 pathways. Supporting a role in activating steroid receptors, Caplan et al. (10) found that the ydj1-151 mutation decreased the ability of the androgen receptor to activate a reporter gene in the presence of hormone. However, Kimura et al. (29) found that the ydj1-G315D mutation greatly increased the activity of the GR and ER in the absence of hormone, suggesting Ydj1 has a role in repressing receptor activity in the absence of hormone. As different Ydj1 mutants and different receptors were used in these studies, the generality of these results and the role of Ydj1 is unresolved. To clarify the role of Ydj1 in Hsp90 pathways in vivo, we have used multiple Ydj1 mutants to examine the role of Ydj1 in the maturation of three different Hsp90 substrates, v-Src, the GR, and the ER.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Generation of Ydj1 mutants. With the exception of C406S, which was a gift of Avrom Caplan, Ydj1 mutants were selected from a random PCR-mutagenized library of YDJ1 in one of two genetic screens. The initial screen, which yielded mutants N104, N134, N172, N206, and G315E, was designed to isolate Ydj1 mutants that were capable of growth in the presence of v-Src at 30°C. The majority of mutants isolated in this screen were truncation mutants. Ydj1 mutants F47L, G153R, N274, and N363 were isolated in a screen for temperature-sensitive Ydj1 mutants that produced proteins of 30 kDa or larger.

Construction of ydj1 library. The 3.6-kb SacI-MunI fragment of YDJ1 was cloned into the SacI-EcoRI sites of the centromeric plasmid pRS316 to create pRS316-YDJ1. A library of PCR-generated ydj1 mutants was generated with Taq polymerase under standard PCR conditions using the primers Ydj1-forward (5'-CTTTTGATAGAACATAA-3') and Ydj1-reverse (5'-TTGGTACCATTATTACTTTAC-3'). The Ydj1-reverse primer introduces a KpnI site downstream of the stop codon. The PCR products were digested with EcoRI-KpnI and cloned into the same sites of pRS316-YDJ1. The mutagenized 1.4-kb EcoRI-KpnI fragment encompasses 40 bp upstream of the ATG through 100 bp downstream of the stop codon.

(i) Screen 1. The mutagenized URA3-marked YDJ1 library was transformed into strain JJ160 (ydj1::HIS3) expressing a LEU2-marked multicopy plasmid expressing v-Src under the control of a GAL1 promoter (pLv-src; a gift from F. Boschelli [5]). Transformants were plated onto leucine-uracil-selective medium plates with 2% galactose as the carbon source and grown at 30°C for 5 days. Colonies that appeared were restreaked on selective galactose medium to verify the phenotype. Cell lysates from these strains were then checked for expression of Ydj1, using an antibody raised against full-length Ydj1. Only candidate mutant strains expressing near-wild-type levels of Ydj1 were chosen for further study.

(ii) Screen 2. To obtain mutants F47L, G153R, N274, and N363, the ydj1 library was subcloned into the LEU2-marked pRS315 vector. Library DNA was transformed into strain JJ160 expressing pRS316-YDJ1. After 2 days, transformants were streaked onto plates containing 5'-fluoro-orotic acid to counterselect for the presence of wild-type YDJ1. Cell lysates from strains exhibiting temperature-sensitive growth were tested for the presence of Ydj1 by immunoblot analysis. Mutant strains producing proteins that migrated between 30 and 44 kDa were selected for further study.

v-Src expression. ydj1 disruption strain JJ160 expressing various Ydj1 mutants was transformed with a 2µm plasmid expressing GAL1-v-src (pBv-src) or the control plasmid (pB656) (5). Yeast cultures were grown overnight at 30°C in raffinose-uracil dropout medium until mid-log phase; 20% galactose was added to a final concentration of 2%. After 6 h, cultures were harvested for growth assays or preparation of cell lysates for immunoblot analysis. Experiments shown were performed using v-Src overexpressed on a multicopy 2µm plasmid, but the same results were obtained if v-Src was expressed from a low-copy-number centromeric plasmid (data not shown).

Plasmids. These experiments were greatly aided by generous gifts of the following plasmids from Didier Picard, University of Geneva: the wild-type GR (pTCA/GZ), mutant GR (pG/N768) and wild-type ER [pG/ER(G)] expression plasmids and their corresponding reporter plasmids pUC55-26X and pUCSS-ERE; the two-hybrid fusion constructs expressing the Gal4 DNA binding domain fused to the ER HBD [pTCA/GAL4(1-93).ER] and the Gal4 activation domain fused to SRC-1 (pGAD424-SRC1); and the chimeric plasmids pHCA/GAL4(1-93).ER.VP16 and pTCA/GAL4(1-93).GR.VP16. The pHCA-TRP/GAL4(1-93).ER.VP16 plasmid was constructed from pHCA/GAL4(1-93).ER.VP16 by moving a 3.0-kb EcoRI/NotI fragment into the pRS314 vector. A plasmid containing the AF-1 mutation 108-317 (25) was a gift from the laboratory of Keith Yamamoto, and the YEpCUP1-HSE-M-lacZ plasmid was a gift from Dennis Thiele (50). Plasmid ZF3 was constructed by placing lacZ coding sequences under control of the SSB2 promoter.

-Galactosidase assays. -Galactosidase assays were performed as described elsewhere (39). Briefly, cultures were grown in selective medium overnight, diluted to an optical density at 600 nm (OD₆₀₀) of 0.05 in fresh selective medium, and grown overnight at 25°C. As described, hormone (final concentration, 10 µM deoxycholate [DOC] or 0.1 µM -estradiol; Sigma) diluted 1:1,000 from stocks in ethanol or ethanol alone was added prior to the second night of incubation. Cells were harvested at an OD₆₀₀ of 0.2 to 1.0. -Galactosidase units were calculated as 10³ × OD₄₂₀ divided by the OD₆₀₀ × elapsed time (in minutes). Wild-type and ydj1 mutant cells expressing the GR showed similar dose-response curves at DOC concentrations between 0.1 and 10 µM. The results shown are for cells grown overnight at 25°C in the presence of 10 µM DOC or 0.1 µM -estradiol, which is the minimum saturation level for wild-type cells. Additional experiments with either a 30°C incubation or 2-h hormone induction in the ydj1 mutant strains showed similar results in overall GR or ER activation of a reporter gene relative to wild type.

Yeast strains. The following yeast strains were used in this study: PJ51-3a (a trp1-1 ura3-1 leu2-3,112 his3-11,15 ade2-1 can1-100 GAL2⁺ met2-1 lys2-2); JJ160 (a trp-1 ura3-1 leu2-3,112 his3-11,15 ade2-1 can1-100 GAL2⁺ met2-1 lys2-2 ydj1::HIS3); PJ69-4a (a trp1-901 leu2-3,112 ura3-52 his3-200 gal4 gal80 GAL2-ADE2 LYS2::GAL-HIS3 met2::GAL7-lacZ); JJ257 ( trp1 leu2-3,112 ura3-52 his3-11,15 GAL2 lys1 lys2 ydj1::HIS3); and JJ290 ( trp1 leu2-3,112 ura3-52 his3 gal4 gal80 GAL2-ADE2 lys2 met2::GAL7-lacZ ydj1::HIS3).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotting. The v-Src antibody (15) was a generous gift of Frank Boschelli. Polyclonal antiserum was prepared against full-length purified yeast Ydj1 by standard procedures. The antibody against Mge1 has been described previously (38). The GR antibody was obtained from Affinity Bioreagents (catalog no. MA1-510). The anti-phosphotyrosine residue antibody was obtained from Upstate Biologicals (catalog no. 4G10).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The J + G/F region of Ydj1 is sufficient for wild-type growth at 30°C. To better understand the in vivo requirements for Ydj1 function, we analyzed an array of Ydj1 mutants in a number of in vivo assays. Figure 1 lists the six carboxy-terminal truncations and three point mutants that we obtained using two separate genetic screens (see Materials and Methods). Also included in this list is the mutation C406S, which inactivates the CAAX box, thus preventing the farnesylation modification of Ydj1 (11). All mutants produce Ydj1 protein levels similar to wild-type cells (data not shown). Plasmids carrying each of these mutants were transformed into a ydj1::HIS3 strain, and the resulting transformants were assayed for growth. Yeast cells containing the J-domain mutation F47L were the only mutant cells that failed to exhibit wild-type growth at 23°C (data not shown) and 30°C (Fig. 1). Yeast expressing even the shortest truncation mutant, N104, which contains only the J + G/F region, grew as well as the wild type at 30°C, the optimal growth temperature for S. cerevisiae. However, none of the mutants were able to grow at 37°C (Fig. 1). The biggest growth difference between the mutants is evident at the intermediate temperature of 34°C. Only truncations containing the entire zinc finger region were able to grow at 34°C, suggesting that this domain has an essential in vivo role at elevated temperatures.

View larger version (53K): [in this window] [in a new window]   FIG. 1.   Schematic of the domain structure of Ydj1 and the Ydj1 mutants used in the study. Ydj1 mutants were obtained as described in Materials and Methods. The ydj1 disruption strain JJ160 expressing various Ydj1 mutants from low-copy-number plasmids was grown overnight at 25°C and then serially diluted 1:10 prior to plating onto YPD. YPD plates were grown for 2 days at 30, 34, or 37°C.

The carboxy terminus of Ydj1 is required for v-Src activity. v-Src binds Hsp90 and cochaperones shortly after synthesis and remains bound during transport to the plasma membrane, where it is released from chaperones prior to becoming active as a phosphotyrosine kinase (6). The expression of v-Src is toxic to wild-type yeast but not to yeast containing deletions or mutations of genes involved in the Hsp90 pathway (27). Many of these ydj1 mutants were isolated in a screen for ydj1 mutants that were defective in v-Src processing (see Materials and Methods). We further tested the effect of the ydj1 mutants N134, N274, N363, C406S, and G315E on v-Src maturation. Wild-type and ydj1 mutant yeast strains containing a multicopy plasmid expressing v-Src under the control of a GAL1 promoter or a control vector were assayed for growth on galactose-based media (Fig. 2A). In addition, cell lysates from wild-type and mutant yeast strains were prepared to determine the activity of v-Src using an anti-phosphotyrosine residue antibody (Fig. 2B) and the level of v-Src protein using an antibody against v-Src (Fig. 2C).

View larger version (36K): [in this window] [in a new window]   FIG. 2.   v-Src phenotype of ydj1 mutant strains. Wild-type (WT) or ydj1 mutant yeast cultures expressing either the GAL1-v-src (pBv-src) multicopy plasmid or the control plasmid (pB656) were grown overnight in selective medium containing raffinose as the carbon source. v-Src expression was induced by addition of 20% galactose to a final concentration of 2%. Cells were harvested 6 h after induction. (A) Yeast cultures were serially diluted 1:10, plated on selective medium containing galactose as the carbon source, and incubated for 2 days at 30°C. (B) Immunoblot of yeast lysates using anti-phosphotyrosine residue antibody 4G10 (Upstate Biologicals). (C) Immunoblot of yeast lysates using anti-v-Src antibody.

GR activity in ydj1 mutant strains. To further investigate whether the carboxy terminus of Ydj1 is required for function in Hsp90 pathways, we assayed the activity of another Hsp90 substrate, the GR, in ydj1 mutant strains. Although not native to S. cerevisiae, steroid receptors expressed in yeast activate transcription of reporter genes in a hormone-dependent manner. The function of the GR was assayed in the ydj1 null strain expressing the rat GR on one plasmid along with a plasmid expressing the lacZ reporter gene under the control of the glucocorticoid response element (GRE). In wild-type strains, very low levels of -galactosidase activity were produced in the absence of hormone, while hormone addition resulted in an approximate 26-fold induction. In contrast, in a ydj1 null strain, the basal activity of the GR was elevated 28-fold, reaching the activity of wild-type cells in the presence of hormone (Fig. 3). GR activity in the null strain increased further in the presence of hormone, surpassing the level of wild-type cells. Restoring wild-type YDJ1 on a plasmid dramatically reduced the basal levels, resulting in only a twofold increase over cells with a wild-type chromosomal copy of YDJ1.

View larger version (10K): [in this window] [in a new window]   FIG. 3.   ydj1 null strain exhibits elevated GR activity in the absence of hormone. Yeast cultures were grown in selective medium overnight at 25°C and then diluted into fresh medium with the addition of ethanol or 10 µM DOC. -Galactosidase assays were performed after overnight (16-h) incubation. Experiments were repeated at least twice. A representative experiment with triplicate samples is shown. Wild-type (WT; PJ51-3A) and ydj1 disruption (JJ160) strains were transformed with plasmids expressing the GR (pTCA/GZ) and corresponding GRE-lacZ reporter plasmid (pUC55-26X). Activity in the wild type, the ydj1 null strain, and a ydj1 null strain with wild-type YDJ1 (+ YDJ1) supplied on a low-copy-number plasmid were determined.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   GR activity in ydj1 mutant strains. (A) ydj1 disruption strain JJ160 containing GR (pTCA/GZ) and corresponding GRE-lacZ reporter plasmid (pUC55-26X) was transformed with indicated ydj1 mutants expressed on low-copy-number plasmids. ydj1 indicates the null strain, and YDJ1 indicates the presence of wild-type YDJ1 on a plasmid. Wild-type (WT) strain (PJ51-3A) is included for comparison. Assays were performed as described in the legend to Fig. 3 and Materials and Methods. (B) Immunoblot showing level of GR expressed in indicated ydj1 mutant strains. As a control for protein loading, the immunoblot was reprobed with antibodies against the mitochondrial GrpE homolog, Mge1.

ER activity in ydj1 mutant strains. We next examined whether the basal activity of another steroid hormone receptor, the ER, is elevated in ydj1 mutant strains. Like the GR, the ER was expressed from one plasmid while the estrogen response element (ERE) upstream of the lacZ gene was present on another plasmid. The ydj1 null strain and some of the ydj1 mutant strains were examined for ER activation of the lacZ reporter gene. The effects of ydj1 mutations on ER basal activity, shown in Fig. 5A, were similar to those obtained for the GR. The basal activity of the ER in the absence of hormone increased, ranging from 2-fold in the case of N172 to 18-fold for N363. For many mutants, the ER activity in the presence of hormone surpassed that of wild-type cells. However, in the absence of any Ydj1 or in the presence of N172 or N206, the overall activity in the presence of hormone was only 40, 37, or 54% of wild-type levels, respectively, suggesting some ydj1 mutants have decreased hormone-dependent activation of the ER. Thus, elevated basal activity of steroid receptors in the presence of ydj1 mutants is shared between the ER and the GR. These results indicate that a major role of Ydj1 is to facilitate repression of receptor activity in the absence of hormone. However, the ydj1 null and some ydj1 mutants displayed a defect in hormonal activation of the ER but not the GR, suggesting the requirement for Ydj1 in hormonal activation may be receptor dependent (compare Fig. 4A and 5A).

View larger version (14K): [in this window] [in a new window]   FIG. 5.   ER activity in ydj1 mutant strains. (A) ydj1 disruption strain JJ160 containing plasmids expressing the ER [pG/ER(G)] and the corresponding ERE-lacZ reporter construct (pUCSS-ERE) was transformed with indicated ydj1 mutants expressed on low-copy-number plasmids. ydj1 indicates the null strain, and WT indicates the presence of wild-type YDJ1 on a plasmid. Assays were performed as described in the legend to Fig. 3 and Materials and Methods except that 0.1 µM -estradiol was used instead of 10 µM DOC. (B and C) ydj1 disruption strain containing indicated ydj1 mutants was transformed with a plasmid expressing lacZ under the constitutive promoter SSB2 (B) or the inducible CUP1 promoter containing a mutation in the heat shock element (C). (B) -Galactosidase assays were performed after overnight incubation at 25°C. (C) -Galactosidase assays were performed after an overnight incubation in the presence or absence of 0.1 mM CuSO4.

ER AF-2 activity is unaffected by mutations in YDJ1. We examined whether the source of receptor activity in the absence of hormone arose from the amino-terminal AF-1 domain or the AF-2 domain, located in the HBD. First we determined whether the AF-2 domain in ydj1 mutant cells has properties similar to those of hormone-bound receptor. Upon ligand binding and Hsp90 release, the HBD undergoes a large conformational change, exposing AF-2, which is the site of interaction for a number of transcriptional coactivators (19, 24). One such coactivator, SRC-1, was originally isolated in a yeast two-hybrid screen designed to find proteins that interact with the HBD in a hormone-dependent manner (41). We confirmed that we could observe a hormone-dependent two-hybrid interaction between a SRC-1-Gal4 activation domain fusion and a Gal4 DNA binding domain-ER-HBD fusion in a wild-type strain PJ69-4a (data not shown), commonly used in two-hybrid assays (26). A hormone-dependent two-hybrid interaction between a Gal4 DNA binding domain-GR-HBD fusion and the SRC-1 fusion was also observed (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 6.   Two-hybrid interaction between SRC-1 and the ER HBD in ydj1 mutant strains. The Gal4 DNA binding domain (DBD)-ER-HBD fusion plasmid (pTCA-Gal4.ER) and the Gal4 activation domain (AD)-SRC-1 plasmid (424-SRC-1) were transformed into the ydj1 disruption strain JJ290 and strain JJ290 expressing wild-type (WT) YDJ1 or mutant ydj1 from low-copy-number plasmids. -Galactosidase assays were performed after overnight incubation in the presence or absence of 0.1 µM -estradiol.

GR containing AF-2 mutation but not AF-1 mutation displays elevated basal activity. The above results suggest that basal activity of steroid receptors in ydj1 mutant strains does not arise from hormone-independent activation of the AF-2 domain. In addition to AF-2, steroid receptors contain an amino-terminal activation domain, AF-1 (Fig. 7A) (2, 21). To localize the source of receptor activity, we used mutations in the GR that abolish activity of either AF-2 (GR N768) or AF-1 (108-317). In wild-type cells, each of these mutations dramatically reduced GR activity (Fig. 7A), which agrees with earlier findings that the AF-1 and AF-2 domains activate transcription synergistically in yeast (21, 51). First, we tested whether a GR truncation mutant that preserves the Hsp90 binding site but disrupts the AF-2 activation domain had elevated basal activity in ydj1 mutant strains. For this test, we used the previously described GR truncation mutant GR N768 (20), which removes 27 amino acids from the carboxy terminus of the HBD, disrupting helix 11, the core of the AF-2 activating domain (19, 24). In yeast and mammalian cells, GR N768 was unable to activate transcription from a GRE either in the presence or in the absence of hormone (18). GR N768 was transformed into wild-type and ydj1 mutant strains and assayed for GR activity (Fig. 7B). As expected, the GR N768 mutation ablated the hormone response in wild-type cells. In the ydj1 mutant strains, however, the basal activity of GR N768 was as high as in wild-type GR, although no additional increase in response to hormone was observed. These results, combined with those presented in Fig. 6, indicate that increased basal activity is not due to hormone-independent activation of the AF-2 domain.

View larger version (25K): [in this window] [in a new window]   FIG. 7.   Activity of mutant GR in ydj1 mutant strains. (A) Locations of AF-2 (GR N768) and AF-1 (108-317) mutations in the GR and corresponding activity in wild-type yeast. DBD, DNA binding domain. (B and C) Similar to Fig. 4A except that a plasmid expressing mutant GR N768 (B) or GR 108-317 (C) was transformed into the indicated wild-type (WT) or ydj1 mutant strain containing the GRE reporter plasmid.

Heterologous activation domains fused to the HBD of the ER or GR exhibit elevated basal activity in ydj1 mutant strains. The AF-1 site is normally repressed in steroid receptors in the absence of hormone. However, deletion of the HBD of either the ER (51) or the GR (20) results in constitutive AF-1 activity, suggesting that the HBD represses the AF-1 site. The HBD of the ER or GR is able to confer hormonal regulation and Hsp90 binding on proteins containing heterologous activation and DNA binding domains (44, 46), indicating that HBDs and Hsp90 can repress a variety of nearby activation domains.

View larger version (21K): [in this window] [in a new window]   FIG. 8.   Chimeric proteins containing the HBD of the GR or the ER display elevated basal activity in ydj1 mutant strains. ydj1 disruption strain JJ290 expressing either ydj1-N363 or wild-type (WT) YDJ1 was transformed with plasmids expressing chimeric proteins containing the VP16 activation domain and Gal4 DNA binding domain (DBD) fused to the HBD or either the GR [pTCA/GAL4(1-93).GR.VP16] or ER [pHCA-TRP/GAL4(1-93).ER.VP16].

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

To better understand the role of Ydj1 in Hsp90 pathways, we isolated a number of Ydj1 mutants that confer a temperature-sensitive phenotype in the absence of wild-type YDJ1 and analyzed their specific defects in the maturation of diverse Hsp90 substrates.

The J + G/F region confers wild-type growth at 30°C. Both genetic and biochemical evidence indicates that the J + G/F region of Hsp40s interacts with Hsp70s (12, 28). Our results provide further evidence for the in vivo significance of this portion of Ydj1. The only mutation used in our study that fails to exhibit wild-type growth at 23°C (data not shown) and 30°C is F47L, which contains a mutation in a highly conserved residue of helix 3 in the J domain (28). This result suggests that the interaction between Ydj1 and Hsp70 is critical for both growth and the maturation of Hsp90 substrates.

v-Src maturation requires the carboxy terminus of Ydj1. Even though yeast expressing N104 grows well at 30°C, the activity of the Hsp90 substrates v-Src, ER, and GR is severely affected at both 25 and 30°C. These results suggest that the carboxy terminus of Ydj1 has specific functions in Hsp90 pathways, but these functions are not essential at 30°C. Within the carboxy terminus of Ydj1 lies the zinc finger region, the neighboring less conserved region, and the farnesylation signal (Fig. 1). The zinc finger region must be present for wild-type growth at the intermediate temperature of 34°C. The farnesylation signal has previously been shown to be required for growth at 37°C (11), and so it is not surprising that all of the truncation mutants fail to grow at 37°C. We constructed pairs of truncation mutations containing wild-type (CASQ) or mutant (SASQ) farnesylation signals after amino acids 209 and 310. The addition of the farnesylation signal did not restore 37°C growth (data not shown). Additionally, the truncation pairs displayed similar v-Src and GR activities (data not shown), indicating that sequences in the carboxy terminus in addition to the farnesylation signal are likely required for growth at elevated temperatures and maturation of substrates.

Steroid receptors have increased basal activity in ydj1 mutant strains. The common feature of all ydj1 mutant strains was increased steroid receptor activity in the absence of hormone. This effect on steroid receptors appears to be unique to ydj1 mutations, as no previously described mutations in genes encoding other components of the Hsp90 pathway have affected the basal activity of steroid receptors (4, 16, 17, 42). In addition, it seems unlikely that ydj1 mutations indirectly affect steroid receptor activity, as activation of other signaling pathways does not lead to hormone-independent activation of the GR in yeast (43). Basal receptor activity does not appear to come from unmasked AF-2 domains, as there is no detectable two-hybrid interaction between the transcriptional coactivator SRC-1 and the HBD of the ER in the absence of hormone. Additionally, mutant GR N768 containing an AF-2 mutation retained elevated basal activity in ydj1 mutant strains, whereas GR 108-317 containing an AF-1 mutation did not, indicating that the activity arises from the AF-1 domain. Consistent with this interpretation, chimeric proteins containing heterologous activation and DNA binding domains fused to the ER or GR HBD also exhibited constitutive activity in ydj1 mutant strains. These results suggest that increased basal activity in ydj1 mutant strains arises from a defect in the ability of the HBD to repress the amino-terminal AF-1 domain.

Model of Ydj1 function in Hsp90 pathways. The data presented in this report can be integrated into a working model of Ydj1 function in the Hsp90 pathway. According to our model, interaction of Ydj1 with Hsp90 substrates is a critical step resulting in entry of substrates into the Hsp90 folding pathway. In the absence of functional Ydj1, transfer of the receptor to the Hsp90 pathway is not complete, and some of the receptor escapes the regulation imposed by the Hsp90 association, becoming capable of binding DNA and activating transcription. Our results suggest that the interaction of Ydj1 with the HBD of steroid receptors is required for strict hormonal control. When Ydj1 is not functional, the HBD is unable to suppress the activity of AF-1, resulting in active receptor. We have not shown that this interaction is direct, but Ydj1 is capable of binding nonnative polypeptides (14, 35). The ability to bind substrate proteins for presentation to other chaperones would be consistent with the function of DnaJ, which binds ³² or other substrates before presentation to Hsp70 for subsequent folding (22, 30, 32).