Genetic and biochemical analysis of p23 and ansamycin antibiotics in the function of Hsp90-dependent signaling proteins

Abstract

The ubiquitous molecular chaperone Hsp90 acts in concert with a cohort of associated proteins to facilitate the functional maturation of a number of cellular signaling proteins, such as steroid hormone receptors and oncogene tyrosine kinases. The Hsp90-associated protein p23 is required for the assembly of functional steroid aporeceptor complexes in cell lysates, and Hsp90-binding ansamycin antibiotics disrupt the activity of Hsp90-dependent signaling proteins in cultured mammalian cells and prevent the association of p23 with Hsp90-receptor heterocomplexes; these observations have led to the hypotheses that p23 is required for the maturation of Hsp90 target proteins and that ansamycin antibiotics abrogate the activity of such proteins by disrupting the interaction of p23 with Hsp90. In this study, I demonstrate that ansamycin antibiotics disrupt the function of Hsp90 target proteins expressed in yeast cells; prevent the assembly of Sba1, a yeast p23-like protein, into steroid receptor-Hsp90 complexes; and result in the assembly of receptor-Hsp90 complexes that are defective for ligand binding. To assess the role of p23 in Hsp90 target protein function, I show that the activity of Hsp90 target proteins is unaffected by deletion of SBA1. Interestingly, steroid receptor activity in cells lacking Sba1 displays increased sensitivity to ansamycin antibiotics, and this phenotype is rescued by the expression of human p23 in yeast cells. These findings indicate that Hsp90-dependent signaling proteins can achieve a functional conformation in vivo in the absence of p23. Furthermore, while the presence of p23 decreases the sensitivity of Hsp90-dependent processes to ansamycin treatment, ansamycin antibiotics disrupt signaling through some mechanism other than altering the Hsp90-p23 interaction.

FIG. 1

MI inhibits ligand response of steroid and nuclear receptors in a dose-dependent manner. (A) Yeast cells expressing GR, PR, or RAR and containing the appropriate β-Gal reporter plasmid were treated with 10 μM Dex, 100 nM progesterone, or 10 μM retinoic acid, respectively, in the presence of increasing concentrations of MI. (B) MI (100 μM) does not affect the activity of a constitutive form of GR (N556) or induction of the GAL1 promoter in response to galactose. Data are the means and ranges for two independent transformants and are representative of three or more experiments. β-Gal activity was determined as described in Materials and Methods. Data for F620S mutant rat GR are represented here and in subsequent experiments (see Materials and Methods); MI treatment has a similar effect on the ligand response of wild-type rat GR (data not shown). A 100% activity for GR, PR, or RAR was approximately 2,100, 1,400, or 45 U, respectively.

FIG. 2

Tyrosine kinase activity and protein level of pp60^v-src are decreased in the presence of MI. Yeast cells containing a galactose-inducible expression vector were grown in glucose or galactose with or without MI (50 μM) as indicated. Samples of whole-cell extract were separated by SDS-PAGE; pp60^v-src protein and phosphotyrosine were detected by immunoblotting with the appropriate antibodies. The amount of sample loaded for each blot was normalized to total protein. The approximate positions of the protein molecular weight markers are indicated. The band in lane 1 represents nonspecific cross-reactivity of the anti-pp60^v-src antibody with a yeast protein and is seen in extracts of yeast cells lacking the pp60^v-src expression plasmid (data not shown).

FIG. 3

Yeast Hsp-Hsc82 specifically binds immobilized geldanamycin. Whole-yeast-cell extracts of strains expressing human Hsp90 (lanes 1 and 2) or yeast Hsp82 and Hsc82 (lanes 3 and 4) were incubated with GA-coupled beads with (lanes 2 and 4) or without (lanes 1 and 3) the prior addition of MI (10 μM). GA-coupled beads were subsequently washed, and bound proteins were eluted, separated by SDS-PAGE, and detected by silver staining. Numbers at the left of the figure indicate the approximate positions of protein molecular weight markers. The identities of the major specifically bound proteins as human Hsp90 (lane 1) and yeast Hsp-Hsc82 (lane 3) were confirmed by immunoblotting (data not shown).

FIG. 4

Comparison of yeast Sba1 and human p23 protein sequences. Protein sequences for yeast Sba1 and human p23 are shown. Sba1 and p23 sequences are 28% identical and 54% similar over the entire 160-amino-acid length of human p23. Sequences were aligned by the Gap program to demonstrate maximum similarity (22). Symbols: |, an exact amino acid match; :, a conserved substitution; …, gaps introduced into the amino acid sequence to optimize the pairing of regions of similarity.

FIG. 5

MI alters glucocorticoid aporeceptor complexes in yeast cells. GR was immunoprecipitated under nondenaturing conditions from whole-cell extracts of yeast cells expressing GR and epitope-tagged yeast Sba1 (Sba1-Flag) in the absence or presence of 10 μM Dex and/or 50 μM MI (lanes 1 to 4). GR and associated proteins in immunoprecipitates were resolved by SDS-PAGE and detected by immunoblotting with antibodies to specific proteins (as described in Materials and Methods). In parallel, extracts of yeast cells expressing human PR and epitope-tagged Sba1 in the absence or presence of 50 μM MI were exposed to anti-GR antibody resin, the resulting precipitates were resolved by SDS-PAGE, and specific proteins in these precipitates were detected to provide an estimate of nonspecific binding (lanes 5 and 6). The total amounts of proteins analyzed were similar in each type of extract (lanes 7 to 12). Whole-cell extract of each type was normalized for total protein concentration and separated by SDS-PAGE, and specific proteins were then detected by immunoblotting. Data for a given protein are from a single experiment and are representative of at least three independent experiments.

FIG. 6

MI reduces Dex binding by yeast cells expressing GR in a dose-dependent manner. Yeast cells expressing GR were preincubated with various concentrations of MI or vehicle only for 2 h at 30°C, followed by a 2-h incubation with [³H]Dex with or without [¹H]Dex competitor. Cells were washed, and the specific counts bound were determined (see Materials and Methods). Each datum point represents the mean ± range for two independent samples from a given experiment normalized to maximum specific counts bound and is representative of three experiments. A 100% binding is approximately 3,100 sites per cell. Yeast cells not expressing GR display no specific binding.

FIG. 7

GR signaling in response to Dex is unaffected by SBA1 deletion. Yeast strains that are wild type, those with SBA1 deleted (Δsba1), or those with SBA1 deleted and with a plasmid expressing epitope-tagged Sba1 (SBA1-Flag) were incubated overnight with various concentrations of Dex. All cell types express GR and contain a GR-inducible lacZ reporter. Cells were harvested and β-Gal activities were measured (see Materials and Methods). Data, expressed as arbitrary β-Gal units, represent the means ± ranges of two independent samples from a given experiment and are representative of several experiments.

FIG. 8

GR signaling in SBA1 deletion strains displays increased sensitivity to ansamycins; expression of human p23 rescues this increased sensitivity. Yeast strains with the chromosomal SBA1 gene deleted and containing plasmids expressing no Sba1 (Δsba1), epitope-tagged Sba1 (Sba1-Flag), or human p23 were incubated overnight at 30°C with Dex (10 μM) with or without MII, HA, or MI. All strains express GR and contain a GR-inducible lacZ reporter. Cells were harvested, and β-Gal activities were measured (see Materials and Methods). Data, expressed as arbitrary β-Gal units, represent the means ± ranges of two independent samples from a given experiment and are representative of several experiments. The response of the parent yeast strain, with the chromosomal SBA1 gene intact, is comparable to those of the SBA1-Flag and human p23 strains (data not shown).

FIG. 9

Dex binding in cells lacking Sba1 displays increased sensitivity to MII. Yeast strains with the chromosomal SBA1 gene deleted and containing plasmids expressing no Sba1 (Δsba1) or epitope-tagged Sba1 (SBA1-Flag) were preincubated with various concentrations of MII or vehicle only for 2 h at 30°C, followed by a 2-h incubation with [³H]Dex with or without [¹H]Dex competitor. Cells were washed, and specific counts bound were determined (see Materials and Methods). Each datum point represents the mean ± range for two independent samples from a given experiment normalized to maximum specific counts bound, and results are representative of three experiments. Total binding is approximately 3,100 sites per cell for SBA1-Flag and 1,900 sites per cell for Δsba1.