The Hsp90 Cochaperones Cpr6, Cpr7, and Cns1 Interact with the Intact Ribosome

Abstract

The abundant molecular chaperone Hsp90 is essential for the folding and stabilization of hundreds of distinct client proteins. Hsp90 is assisted by multiple cochaperones that modulate Hsp90's ATPase activity and/or promote client interaction, but the in vivo functions of many of these cochaperones are largely unknown. We found that Cpr6, Cpr7, and Cns1 interact with the intact ribosome and that Saccharomyces cerevisiae lacking CPR7 or containing mutations in CNS1 exhibited sensitivity to the translation inhibitor hygromycin. Cpr6 contains a peptidyl-prolyl isomerase (PPIase) domain and a tetratricopeptide repeat (TPR) domain flanked by charged regions. Truncation or alteration of basic residues near the carboxy terminus of Cpr6 disrupted ribosome interaction. Cns1 contains an amino-terminal TPR domain and a poorly characterized carboxy-terminal domain. The isolated carboxy-terminal domain was able to interact with the ribosome. Although loss of CPR6 does not cause noticeable growth defects, overexpression of CPR6 results in enhanced growth defects in cells expressing the temperature-sensitive cns1-G90D mutation (the G-to-D change at position 90 encoded by cns1). Cpr6 mutants that exhibit reduced ribosome interaction failed to cause growth defects, indicating that ribosome interaction is required for in vivo functions of Cpr6. Together, these results represent a novel link between the Hsp90 molecular-chaperone machine and protein synthesis.

INTRODUCTION

Molecular chaperones assist in the folding of other proteins during synthesis, as well as upon denaturation or misfolding. Prior results suggest that one type of molecular chaperone assists in folding of newly translated proteins, while a different set of molecular chaperones help proteins fold after denaturation (1). The abundant essential chaperone Hsp90 does not associate with nascent chains and does not appear to have a general role in protein folding. Instead, Hsp90 is required for the folding and activation of a specific subset of cellular proteins (2–4). Although it is unlikely that it plays a general role in ribosomal function, Hsp90 has been shown to be involved in polysome stability and to have physical or genetic interactions with select proteins with ribosomal or preribosomal functions (5–8).

In Saccharomyces cerevisiae, Hsp90 is assisted by over 10 cochaperones that target clients to Hsp90; directly or indirectly modulate Hsp90 ATPase activity; or have other, less defined functions (9). Many of the cochaperones contain tetratricopeptide repeat (TPR) domains and compete for binding to the conserved carboxy-terminal MEEVD sequence of Hsp90. Three of the TPR-containing cochaperones are Cpr6, Cpr7, and Cns1. Loss of CPR6 does not cause noticeable growth defects; loss of CPR7 results in slow, temperature-sensitive growth; and CNS1 is essential for viability (10–12). Cpr6 and Cpr7 share 38% amino acid identity yet exhibit different functions in vivo and in vitro (13–16). In particular, overexpression of CPR7 rescues the temperature-sensitive defect of cns1-G90D (the G-to-D change at position 90 encoded by cns1) cells (17), while overexpression of CPR6 exacerbates the growth defects (18). Further, overexpression of CNS1, but not CPR6, was able to rescue growth defects of cpr7 cells (11–13). Yeast chaperones have been characterized as chaperones linked to protein synthesis (CLIPS) and heat shock proteins (HSPs). Transcription of HSPs is induced by multiple environmental stressors, while transcription of CLIPS is repressed under the same conditions. The pattern of transcriptional activation of CPR6 was very similar to that of HSP82, one of two Hsp90 genes in yeast, as well as genes encoding the Sti1 and Sba1 cochaperones (1). However, Cpr6 and/or Cpr7 was previously shown to associate with translating ribosomes, and loss of CPR7 conferred hypersensitivity to the translation inhibitor hygromycin (1), suggesting a potential link between Hsp90 cochaperones and the ribosome.

We previously demonstrated that isolation of Cpr6 results in copurification of Hsp90, Hsp70, and Ura2, an Hsp90 client protein required for pyrimidine biosynthesis (14). In this study, we found that isolation of Cpr6 also results in copurification of components of both the large and small ribosomal subunits. Further, select mutations within the TPR domain that disrupt Cpr6 interaction with Hsp90 stabilized ribosome interaction. Cns1 and Cpr7 also interacted with the ribosome, although detection of Cpr7 interaction required the presence of the stabilizing mutation in the TPR domain. Although the functional significance of the interaction remains unclear, our results indicate a novel link between Hsp90 cochaperones and the intact ribosome.

MATERIALS AND METHODS

Media, chemicals, and antibodies.

Yeast cells were grown in either YPD (yeast extract-peptone-dextrose) or defined synthetic complete medium supplemented with 2% dextrose. Growth was examined by spotting 10-fold serial dilutions of yeast cultures onto the appropriate media, followed by incubation for 2 days at the indicated temperature. 5-Fluoroorotic acid (5-FOA) was obtained from Toronto Research Chemicals. Anti-Zuo1 antiserum was a gift from Elizabeth Craig (University of Wisconsin—Madison) (19). Hygromycin B was obtained from Sigma. Anti-TAP antibody was obtained from Pierce.

Yeast strains.

Cpr6-ribosome interaction was observed in three different strain backgrounds: S288C, W303, and BY4741. S288C strains expressing RPS0A-TAP or RPL8A-TAP were obtained from Open Biosystems (20). Wild-type (WT), ssz1, zuo1, egd1, egd2, and btt1 strains in the BY4741 background were also from Open Biosystems. Strains JJ762 (WT), JJ1138 (cpr6), JJ110 (cpr6 hsc82 hsp82/YEp24HSP82), JJ1093 (ura2), and JJ21 (cns1::TRP1/YCp50-CNS1), JJ1115 (cpr7), and JJ816 (hsc82 hsp82/YEp24HSP82), which are isogenic to W303, have been described previously (13, 14, 18, 21).

Plasmids.

pRS416GPDHis-Cpr6, pRS414GPDHis-Cpr6, pRS415GPDHis-Sti1, pRS416GPDHis-Cpr7 (WT and Cpr7 193–393), pRS416GPDHis-Cns1, YEP24-CNS1, and pRS317-CNS1 (WT and mutant) were previously described (13, 14, 18, 22). Plasmids expressing untagged WT or mutant untagged Hsp90 (the Hsp82 isoform) (pRS314 HSP82 WT or ΔMEEVD) were transformed into the indicated cpr6 hsc82 hsp82 strain harboring the YEp24-HSP82 plasmid. Transformants were grown in the presence of 5-FOA to lose the YEp24-HSP82 plasmid. To assess hygromycin sensitivity, His-tagged WT or mutant Hsc82 (pRS313GPDHis-HSC82) was transformed into JJ816, and the YEp24-HSP82 plasmid was cured by plating in the presence of 5-FOA. SBA1 was cloned into pRS416GPDHis using engineered BamHI and EcoRI sites. pRS416GPDHis-Zuo1 was a gift from Elizabeth Craig (University of Wisconsin). The plasmid expressing His-Hsp82 WT was a gift from Len Neckers (National Cancer Institute). Amino acid mutations were constructed using site-directed mutagenesis. All mutations were confirmed using DNA sequencing. The mutagenic oligonucleotide sequences are available upon request.

Isolation of His-Cpr6 or other His-tagged complexes.

Briefly, cells were grown overnight and harvested at an optical density at 600 nm (OD₆₀₀) of 1.2 to 2.0. The cell pellets were resuspended in lysis buffer (20 mM Tris, pH 7.5, 100 mM KCl [or other concentrations of KCl as noted], 5 mM MgCl₂, 5 mM imidazole containing a protease inhibitor tablet [Roche Applied Science]) and were disrupted in the presence of glass beads with 8 30-s pulses. The cell lysates were incubated with nickel resin (1 h with rocking; 4°C), followed by washes with lysis buffer plus 35 mM imidazole and 0.1% Tween 20. Proteins were eluted from the nickel resin by boiling in SDS-PAGE sample buffer, and protein complexes were separated by gel electrophoresis (generally 12.5% acrylamide), followed by Coomassie blue staining. Alternatively, proteins were transferred to nitrocellulose and chemiluminescence immunoblots were performed according to the manufacture's suggestions (Pierce, Rockford, IL).

Mass spectrometry fingerprint analysis.

Ribosomal proteins were identified at the Environmental Biotechnology Institute at the University of Idaho. The analysis of peptides was done using reverse-phase liquid chromatography on a Waters nanoAcquity Ultra Performance Liquid Chromatograph (UPLC). This was followed by tandem mass spectrometry (MS-MS) using a Waters Micromass Q-Tof Premier quadrupole-time of flight mass spectrometer using a nanospray electrospray ionization inlet. The smear of proteins in the 15- to 40-kDa range was excised from the Coomassie-stained gel and dried using acetonitrile. Sequence grade trypsin (Sigma-Aldrich) was used to digest the bands at 37°C for 20 h. Peptides were extracted three times using 100 mM ammonium bicarbonate for one extraction and 50% acetonitrile containing 5% formic acid for the other two extractions. The extracts were evaporated to complete dryness and concentrated in 20 μl 5% acetonitrile and 0.1% formic acid. Peptides were identified using the MASCOT program with a peptide and MS-MS tolerance of 0.2 Da. The complete list of proteins with scores over 25 is in Table S1 in the supplemental material.

RESULTS

The Hsp90 cochaperone Cpr6 binds the intact ribosome.

We previously showed that purification of the Hsp90 cochaperone Cpr6 from yeast results in the copurification of Hsp90, Hsp70 (Ssa family), and the Hsp90 client Ura2 (14). Basic residues in the TPR domain of Cpr6 interact with carboxy-terminal EEVD residues of Hsp70 and/or Hsp90 in what has been termed a carboxylate clamp (23). The crystal structure of bovine Cyp40, the homolog of Cpr6, is available, and prior analysis identified amino acids in the predicted EEVD-binding groove of Cyp40 that were important for Hsp90 and Hsp70 binding (24–26). Mutation of a Cpr6 residue predicted to contact the EEVD sequences (K309) disrupted Hsp70 (Ssa family) and Hsp90 (Hsc82 and Hsp82) interaction without disrupting Cpr6-Ura2 interaction (14).

Because the TPR domain plus flanking regions of Cpr6 was required for interaction with Ura2, Hsp90, and Hsp70 (14), we examined the impact of alanine substitution for additional residues within the TPR domain. In agreement with a prior study (25), the K223A, F230A, L281A, and K309A alterations disrupted Hsp90 and Hsp70 interaction, and K231A resulted in reduced interaction (Fig. 1A). In contrast, alteration of residue E300, which is outside the predicted EEVD-binding groove, had little or no effect. None of the mutations disrupted the Cpr6-Ura2 interaction. Surprisingly, isolation of His–Cpr6-F230A also resulted in copurification of multiple proteins in the 15- to 40-kDa range. Reduced levels of the same proteins were observed in cells expressing K309A. Mass spectrometry analysis identified those proteins as components of both the large and small ribosomal subunits (Fig. 1B; see Table S1 in the supplemental material). Zuo1 is a chaperone that binds the large subunit near the exit tunnel (19, 27). An immunoblot confirmed a low level of Zuo1 interaction with WT Cpr6 and enhanced interaction with Cpr6-F230A and -K309A.

FIG 1.

Cpr6 alteration affects Hsp90, Hsp70, and ribosome interaction. (A) WT or mutant His-Cpr6 complexes were isolated from Δcpr6 strain JJ1138 in buffer containing 100 mM KCl. Proteins eluted from the nickel resin were analyzed using Coomassie blue staining (top) or immunoblot analysis (bottom). The bands marked with a bracket are unidentified proteins that bind nonspecifically to nickel resin used as internal loading controls. Lane −, cells contained empty vector. (B) Mass spectrometry analysis identified multiple components of the large and small ribosomal subunits.

We next isolated His-Cpr6 from a commercially available strain expressing TAP-tagged RPS0A, a component of the small subunit, from the normal chromosomal location (20). As shown in Fig. 2A, low levels of RPS0A were present in WT His-Cpr6 complexes, whereas greatly enhanced levels copurified with Cpr6-F230A. Similar results were obtained in a strain expressing RPL8A-TAP, a component of the large subunit (data not shown; see Fig. 7B). Reducing the salt concentration in the lysis buffer resulted in similar levels of ribosome interaction with Cpr6 WT and -F230A (Fig. 2B). Together, these results indicate that Cpr6 WT and -F230A interact with the assembled 80S ribosome in both W303 and S288C backgrounds.

FIG 2.

Effect of altered KCl concentration on Cpr6-ribosome interaction. (A) His-Cpr6 complexes were isolated from a strain expressing RPS0A-TAP from the endogenous chromosomal location. RPS0A-TAP bound to resin was recognized by an anti-TAP antibody. (B) As in panel A, except that 10 mM KCl was used in the lysis and wash buffers instead of 100 mM KCl. (C) His–Cpr6-F230A complexes were isolated from the RPS0A-TAP strain lysed in the presence of the indicated concentrations of KCl. Lanes −, cells contained empty vector.

FIG 7.

Cns1 also interacts with the ribosome. (A) WT or mutant His-Cns1 complexes were isolated from the RPS0A-TAP strain in the presence of 10 mM KCl. (B) The indicated His-tagged protein complexes were isolated from the RPL8A-TAP strain in the presence of 10 mM KCl. Lanes −, cells contained empty vector.

The results shown in Fig. 2A and B suggest that WT Cpr6 is loosely bound to intact ribosomes, dissociating above 10 mM KCl. We examined the effect of increasing salt on Cpr6-F230A–ribosome interaction. As shown in Fig. 2C, Cpr6-F230A dissociated from the ribosome at concentrations over 100 mM KCl. Similar results were obtained in strain JJ1138 (not shown). The reason that the F230A mutation stabilizes ribosome interaction against increasing salt concentrations remains unknown. For comparison, interaction of the Hsp70 Ssb1 with the ribosome is very stable, persisting in the presence of 1 M KCl (19). Surprisingly, the Cpr6-Ura2 interaction was not disrupted by increasing salt concentrations.

We conducted additional experiments to determine whether Cpr6-ribosome interaction was regulated by Hsp90, other ribosome-associated chaperones, or Ura2. The terminal MEEVD residues of Hsp90 are critical for stable interaction with Sti1, Ppt1, Cpr6, and Cpr7 (9). Hsp82ΔMEEVD and WT Hsp82 exhibit similar steady-state levels (14, 28), and cells expressing Hsp82ΔMEEVD as the only Hsp90 in the cell do not exhibit any obvious growth defects. Since ribosome interaction was stabilized upon alteration of residues in the TPR-binding groove, we examined the impact of deletion of the MEEVD sequence. Although Cpr6 failed to stably interact with Hsp82ΔMEEVD, ribosome interaction was not affected (Fig. 3A).

FIG 3.

Cpr6-ribosome interaction was unaffected by mutation of Hsp90 or deletion of genes encoding ribosome-associated chaperones or Ura2. (A) His-Cpr6 complexes were isolated from a cpr6 hsc82 hsp82 strain expressing WT Hsp82 or Hsp82ΔMEEVD. (B) His-Cpr6 WT or -F230A complexes were isolated from strain JJ1138 (Δcpr6), BY4741, or BY4741 containing deletions in SSZ1, ZUO1, EDG1, EDG2, or BTT1. (C) His-Cpr6 complexes were isolated from strain JJ762 (WT) or JJ1093 (ura2::HIS3). Lanes −, cells contained empty vector.

Yeast contains two ribosome-tethered chaperone complexes, RAC and NAC, which interact with nascent chains (29). We tested the possibility that Cpr6 loosely binds to RAC or NAC components in order to facilitate interaction with newly synthetized proteins. As shown in Fig. 3B, deletion of individual components of RAC (SSZ1 and ZUO1) or NAC (EDG1, EDG2, and BTT1) had little or no effect on Cpr6-F230A interaction with the ribosome. Similar effects were observed with WT Cpr6 (not shown). Since Cpr6 appears to directly interact with Ura2, yet another possibility was that Cpr6 bound nascent Ura2, resulting in coisolation of the ribosome. Although the interaction of WT Cpr6 with the ribosome was slightly reduced, ribosome association of Cpr6-F230A was unaltered in cells that lack URA2 (Fig. 3C). Together, these results indicate that Cpr6-ribosome interaction is not dependent on Hsp90 and Ura2 interaction and that it is unlikely that RAC or NAC mediates ribosome interaction.

Basic amino acids in the carboxy terminus of Cpr6 are required for ribosome interaction.

We took advantage of available truncated forms of His-Cpr6 (Fig. 4A) (14) to determine which part of Cpr6 interacts with the ribosome. As shown in Fig. 4B, Cpr6 171–371, but not Cpr6 1–212, bound Hsp90, Hsp70, and Ura2, as well as WT Cpr6. Cpr6 1–358 bound Hsp90 at a reduced level. None of the truncated constructs bound the ribosome as well as intact Cpr6. We introduced the F230A alteration to facilitate narrowing down the interacting region. As shown in Fig. 4C, Cpr6 171–371:F230A was able to interact with the ribosome, but Cpr6 1–358:F230A was not. This suggests that the ribosome-binding site is located near the carboxy terminus of Cpr6. Of note, Cpr6 1–358 bound Hsp90 better under low-salt conditions (Fig. 4B) than under higher-salt conditions (Fig. 4C).

FIG 4.

The carboxy terminus of Cpr6 is required for ribosome interaction. (A) Schematic of WT and truncated Cpr6 constructs. (B) His-Cpr6 complexes were isolated from the RPS0A-TAP strain in the presence of 10 mM KCl. (C) Full-length or truncated His-Cpr6 complexes were isolated from the cpr6 strain in the presence of 100 mM KCl. (D) The indicated His-tagged Cpr6 or Cpr7 constructs were isolated from the RPS0A-TAP strain in the presence of 10 mM KCl. Lanes −, cells contained empty vector.

An earlier study showed that Cpr6 and/or Cpr7 specifically comigrated with polysomes, but the antibody used in those experiments did not distinguish between Cpr6 and Cpr7 (1). As previously shown, isolation of His-Cpr7 does not result in isolation of Hsp90 or Ura2 (14). Cpr7 also failed to interact with the ribosome unless it contained the F251A alteration that is homologous to Cpr6-F230A (Fig. 4D). This also indicates that Cpr6 and Cpr7 are similarly affected by alteration of a homologous residue in the TPR domain.

Cpr6 1–358, which contains a deletion of the terminal 13 amino acid residues (aa) of Cpr6, exhibited reduced ribosome interaction. Eight of the last 20 amino acids of Cpr6 are lysine or arginine residues. As shown in Fig. 5A, only some of them are conserved in Cpr7. We tested the impact of mutation of basic residues on ribosome interaction in the context of WT Cpr6 (Fig. 5B) or Cpr6-F230A (Fig. 5C). Either changing the three lysines in the KAKK sequence to alanines (AAAA) or reversing the charges of the two adjacent lysine residues (KAEE) reduced ribosome interaction without affecting Ura2, Hsp90, or Hsp70 interaction. The DMFS mutant, which reverses the charge of the conserved lysine closest to the carboxy terminus, had little effect in the context of WT Cpr6. However, all three mutations disrupted ribosome interaction under the more stringent conditions (Fig. 5C).

FIG 5.

Identification of basic residues required for ribosome interaction. (A) Sequences of the carboxy termini of Cpr7 and Cpr6 with targeted residues of Cpr6 underlined. (B) His-Cpr6 containing the indicated mutations in basic residues isolated from the RPS0A-TAP strain in the presence of 10 mM KCl. (C) His–Cpr6-F230A cells containing additional mutations in basic residues were isolated from the the RPS0A-TAP strain in the presence of 100 mM KCl. Lanes −, cells contained empty vector.

We recently showed that overexpression of CPR6 results in negative growth in cells expressing a temperature-sensitive mutation in CNS1, cns1-G90D (18). We examined whether there is a correlation between ribosome interaction and ability to cause negative growth. As shown in Fig. 6A, cpr6-F230A and cpr6 1-358 failed to cause negative growth in cns1-G90D cells. Similar effects were observed with the cpr6-KAEE and -DMFS mutants. This indicates that disruption of either Hsp90 or ribosome interaction is able to relieve the negative effects of CPR6 overexpression. We also examined whether the peptidyl-prolyl isomerase (PPIase) domain was required for Cpr6 to cause negative growth. However, Cpr6 171–371 and 212–371, both of which lack the PPIase domain, caused a strong negative effect (Fig. 6B).

FIG 6.

Effect of CPR6 mutation on growth of cns1-G90D cells. (A and B) Strain JJ21 (cns1::TRP1) expressing cns1-G90D under its own promoter was transformed with plasmids expressing WT or mutant His-Cpr6 or empty vector (−). After overnight growth in selective media, 10-fold serial dilutions were plated and grown for 2 days at the indicated temperatures.

Cns1 also interacts with the ribosome.

Since the ability of Cpr6 to cause a growth defect in cns1-G90D cells is linked to ribosome interaction, we determined whether Cns1 also interacts with the ribosome. As shown in Fig. 7A, isolation of His-Cns1 results in copurification of Hsp90. His-Cns1 also interacts with the ribosome, as judged by gel staining and copurification of RPS0A. We next examined the abilities of altered forms of Cns1 to interact with the ribosome. Cns1-G90D, which contains an alteration in a conserved residue of the TPR domain (30), exhibited reduced Hsp90 interaction but maintained ribosome interaction. In contrast, Cns1 1–212, which contains the TPR domain plus the flanking region, interacted with Hsp90 but not the ribosome. Cns1 213–385, which lacks the TPR domain, interacted with the ribosome. We further examined whether Hsp90, Sti1, or Sba1 interacted with the ribosome. For comparison, we included the known ribosome-associated chaperone Zuo1. As expected, His-Sti1 bound Hsp90 and Hsp70. His-Sba1 interaction with Hsp90 was not observed due to the lack of ATP (9). Similar levels of Cpr6, Sti1, Sba1, Hsp82, and Zuo1 bound nickel resin (Fig. 7B). However, only Cpr6 and Zuo1 showed robust ribosome interaction.

Cells containing deletions of ribosome-associated chaperones are sensitive to hygromycin, a translation inhibitor (31). A prior study showed that, unlike cells lacking HSP82, STI1, or CPR6, cells lacking CPR7 were sensitive to hygromycin (1). As shown in Fig. 8A, cpr7 cells exhibited slow growth at all three temperatures and were unable to grow in the presence of hygromycin. Transformation with WT CPR7, cpr7 193-393 (analogous to Cpr6 171–371) (13), or cpr7-F251A, rescued all the growth defects. This is consistent with a prior report that the PPIase domain of Cpr7 is dispensable for most known functions (32). The cns1 1-121 and cns1-G90D mutations were previously known to cause temperature-sensitive growth (17). Both mutations also caused strong growth defects in the presence of hygromycin, indicating that the ability of Cns1 to interact with both Hsp90 and the ribosome is required for growth in the presence of the drug. Cells lacking CPR6 were not sensitive to hygromycin under these conditions (not shown).

FIG 8.

Growth defects of cochaperone and Hsp90 mutant cells. (A) cpr7 (JJ1115) cells expressing WT or mutant His-tagged Cpr7 or JJ21 (cns1::TRP1) cells expressing WT or mutant CNS1 under its own promoter were grown in rich media for 2 days at the indicated temperatures. YPD plates containing 60 mM hygromycin B were grown for 2 days at 30°C. (B) As in panel A, except a plasmid expressing WT or mutant His-tagged HSC82 was expressed in hsc82 hsp82 cells (JJ816).

To assess whether Hsp90 is involved in the ribosomal functions of cochaperones, we examined previously characterized hsc82 mutations (6, 21). When expressed as the only Hsp90 in the cell, hsc82-G309S, hsc82-A583T, and hsc82-I588AM589A cause similar growth defects at 37°C. Of the three mutants, only hsc82-I588AM589A also showed a strong growth defect in the presence of hygromycin. Of note, hsc82-I588AM589A, cpr7, and cns1 mutant cells also exhibit slow growth at 25°C (Fig. 8A). Yeast lacking individual components of RAC also exhibits cold sensitivity (33). It is tempting to speculate that the cold sensitivity of cpr7, cns1, and hsc82 mutant strains is related to defects in ribosomal functions, but more work is needed to establish that connection. Regardless, the hygromycin defect of hsc82-I588AM589A cells suggests that Hsp90 and cochaperones cooperate to promote ribosome function.

DISCUSSION

Hsp90 mediates the folding and activation of proteins as a complex molecular machine, requiring ATP binding and hydrolysis, posttranslational modifications, and multiple cochaperones (9, 34). Cochaperones bind distinct domains and/or specific conformations of Hsp90. In addition to interacting with Hsp90, some cochaperones interact directly with clients. Cdc37 and Sgt1 have important roles targeting protein kinases or leucine-rich-repeat-containing proteins, respectively, to Hsp90 (35, 36). Cpr6 or homologs have been shown to be important for RNA-induced silencing, lipid trafficking, and viral function (37–39). Here, we show that Cpr6, Cns1, and Cpr7, but not other tested cochaperones, interact with intact ribosomes, providing new evidence for specialized cochaperone functions.

Cpr6 contains a PPIase domain and a TPR domain flanked by charged regions. The TPR domain plus flanking regions (aa 171 to 371) contains separate binding sites for Hsp90, Ura2, and the ribosome. Hsp90 binds residues in the established EEVD-binding cleft (23, 25). Basic residues in the last 15 amino acids of Cpr6 were required for ribosome interaction. A chimeric protein that contains the first 271 amino acids of Cpr6 and the last ∼100 amino acids of Cpr7 (His-Cpr6-6A/7B) failed to interact with Ura2 or to rescue growth defects caused by loss of CPR6 but was able to rescue growth defects caused by loss of CPR7 (14). This suggests that the last 100 aa of Cpr6 dictate Ura2 interaction and Cpr6-specific functions. Preliminary results also suggest that, similar to Cpr7, Cpr6-6A/7B failed to interact with the ribosome unless it contained the F230A alteration (not shown). Further analysis is required to determine whether basic amino acid residues in the carboxy terminus of Cpr7 also mediate ribosome interaction. The TPR domains and/or flanking regions of three additional Hsp90 cochaperones, Sgt1, Tah1, and AIP, are also critical for interactions with proteins other than Hsp90 (40–42). Surprisingly, the PPIase domains of both Cpr6 and Cpr7 appear to be dispensable for known functions. However, it is possible that the PPIase domain stabilizes the interaction of WT Cpr6 with the ribosome, since Cpr6 171–373 exhibited a slight reduction in ribosome interaction compared to WT Cpr6. The PPIase domain of Cpr6 was not required to cause negative growth in the cns1-G90D strain. Mutation of a residue (R64) predicted to be required for catalytic activity (32) did not disrupt Cpr6-ribosome interaction, and overexpression of cpr6-R64A, like WT CPR6, caused growth defects in a cns1-G90D strain (not shown). The ribosome-associated chaperone trigger factor of Escherichia coli also contains a PPIase domain. Although not essential for folding of cytosolic proteins, the PPIase domain of the trigger factor assists in folding some substrates (43, 44). More recently, part of the PPIase domain that overlaps the site of enzymatic activity was shown to interact with an unfolded substrate (45).

Mutation or overexpression of CNS1 is known to affect Hsp90 function in yeast (30, 46, 47). Human and Drosophila homologs (TTC4 and Dpit47, respectively), interact with Hsp90, as well as DNA polymerase alpha, the replication protein Cdc6, the histone acetyltransferase MYST/MOF, and the transcriptional initiation factor TFIIIB (48–52). No specific clients of yeast Cns1 have been previously identified. Sequences outside the TPR domain are required for growth at elevated temperature, ribosome interaction, and growth in the presence of hygromycin. There is limited homology between the carboxy terminus of Cpr6 and sequences adjacent to the TPR domain of Cns1. However, most of those residues are missing in Cns1 212–385, which was sufficient for ribosome interaction. Thus, Cpr6 and Cns1 appear to use different sequences to interact with the ribosome. An altered version of TTC4 found in cancer cells contained a mutation located outside the TPR domain (51), and the temperature-sensitive mutation cns1-3C contains three amino acid alterations in the carboxy terminus, D260G, E324G, and L330S (17). Additional work is needed to narrow down the site of ribosomal interaction in Cns1 and to determine whether any of those mutations disrupt ribosome interaction.

Interaction of Cpr6, Cns1, and Cpr7 with the intact ribosome suggests Hsp90 involvement. Although we did not detect Hsp90 interaction with intact ribosomes, cells expressing the hsc82-I588AM589A mutations exhibited growth defects in the presence of hygromycin, suggesting that mutation may affect ribosomal functions. Hsc82 residues I588 and M589 are homologous to Hsp82 I592 and M593, which are located in a putative client-binding site (53). Interestingly, similar to hsc82-I588AM589A, some mutations in hsp82 residues in the client-binding site also resulted in cold-sensitive growth. Further work is needed to determine whether those mutations result in sensitivity to hygromycin and/or ribosomal defects. Alternatively, ribosomal functions of Cns1 and Cpr6 may be related to the ability of both of the proteins to bind Hsp70 of the Ssa family (14, 26, 54). The Hsp70 nucleotide exchange factor Snl1 was recently found to interact with intact ribosomes, with a presumed binding site on the large subunit (55). The function of that interaction is unclear, but endoplasmic reticulum (ER) membrane localization of Snl1 suggests a potential role in protein transport. Alternatively, both Cpr7 and Cns1 interact with Hsp104 during respiratory growth (56), and we have detected Hsp104 in complex with His-Cpr6 (not shown), so it is possible that Hsp104 mediates ribosome interaction.

The gene expression tool SPELL (Serial Pattern of Expression Levels Locator) (57) indicates that CNS1 and CPR7 share transcriptional regulation patterns similar to those of genes encoding proteins required for ribosome function (see Table S2 in the supplemental material). Both the cns1-G90D and cns1 1-212 mutants showed hypersensitivity to hygromycin, indicating that loss of Hsp90 interaction or ribosome interaction affects Cns1 function in that assay. Deletion of CPR7 also confers sensitivity to hygromycin. Cpr6 is functionally linked to the ribosome, since Cpr6 mutants that fail to interact with the ribosome are unable to cause growth defects in cns1-G90D cells. Independent evidence comes from a prior study that found that Cpr6 and/or Cpr7 bound to polysomes (1). Our model is that Cns1 and Cpr7 have ribosomal functions in unstressed cells. Overexpression of the stress-inducible CPR6 may disrupt Cpr7 or Cns1 functions by competing for either Hsp90 or ribosomes. Enhanced ribosomal interaction of Cpr6-F230A and Cpr7-F251A suggests that Cpr6 and Cpr7 interaction may also be regulated by posttranslational modification of nearby residues. Preliminary results suggest that a mild heat shock (10 min at 42°C) had no effect on Cpr6-ribosome interaction, but we have not tested more severe conditions. Two recent papers indicate that ribosome pausing occurs during heat shock (58, 59). Although Hsp90 does not appear to play a significant role, reduced levels of Hsp70-ribosome interaction play a role in ribosomal pausing. It is possible that Cpr6 and/or Cns1 modulates Hsp70 function in that process by targeting Hsp70 to misfolded proteins during times of cellular stress.

Further analysis is needed to determine whether Cns1, Cpr6, and Cpr7 are involved in previously described ribosomal functions linked to Hsp90 or other cochaperones (5–8, 60). Possible cochaperone functions are interactions with nascent chains that are also Hsp90 clients, regulating polysome stability, assisting in the folding of a ribosomal protein, or differentially modulating ribosomal function in response to stress. Our results do not support a direct interaction with components of RAC and NAC, but further experiments are required to determine whether the cochaperones associate with ribosome-bound nascent chains. It is intriguing that the interaction of WT Cpr6 with the ribosome appeared to be stabilized by Ura2, but further studies are needed to confirm those results. Two-hybrid analysis and other genome-wide analyses uncovered an interaction of Cns1 with elongation factor 2, which catalyzes ribosomal translocation during protein synthesis (30, 61). We have not examined whether Cpr6, Cpr7, and/or Cns1 directly interacts with elongation factor 2. We also do not yet know whether ribosome interaction is linked to the chaperoning ability of Cpr6 and Cpr7 (15).

TPR-containing cochaperones competitively interact with the carboxy-terminal EEVD sequences of Hsp90. Each of the cochaperones has varied abilities to modulate the ATPase activity of Hsp90 (or Hsp70) and may have additional functionalities, such as PPIase domains (Cpr6 and Cpr7) or phosphatase domains (Ppt1). Despite having sequence similarity, Cpr6 and Cpr7 have different in vivo functions. A dimer of Hsp90 is able to bind two different TPR cochaperones at the same time, resulting in a progression of different Hsp90-cochaperone complexes during client folding (9, 62). However, only some ternary complexes appear possible. For instance, Cpr6 is able to form ternary complexes with Hsp90 and Aha1, but Cpr7 is not (16). Cns1 and Cpr7 are found in the same complexes in the cell and have overlapping cellular functions. There is also evidence that Cns1 and Cpr7 interact in the absence of Hsp90 (11, 12, 17).

The ability of CPR6 to cause growth defects in cns1 mutant strains requires ribosome interaction. This suggests that cochaperones may compete with one another, either by binding Hsp90 or by binding to the same ribosomal client. We do not have evidence that Cpr6 and Cns1 directly compete for interaction with the same ribosomal protein, and experiments to determine whether overexpression of Cns1 disrupted Cpr6 interaction, or vice versa, did not provide consistent results. Alternatively, Cns1 interaction may cause a conformational change or steric hindrance that prevents Cpr6 interaction, or vice versa. However, differential expression patterns of these cochaperones raises the intriguing possibility that cochaperone-client complexes vary depending on growth conditions. This finding is similar to a prior report that some Hsp90 cochaperones interact with Hsp104 during respiratory growth (56). Either of these cases likely results in altered activity of subsets of client proteins. In summary, these studies indicate that there is much to be learned about the specificity and in vivo functions of TPR-containing Hsp90 cochaperones.