Expression of a unique drug-resistant Hsp90 ortholog by the nematode Caenorhabditis elegans

Abstract

In all species studied to date, the function of heat shock protein 90 (Hsp90), a ubiquitous and evolutionarily conserved molecular chaperone, is inhibited selectively by the natural product drugs geldanamycin (GA) and radicicol. Crystal structures of the N-terminal region of yeast and human Hsp90 have revealed that these compounds interact with the chaperone in a Bergerat-type adenine nucleotide-binding fold shared throughout the gyrase, Hsp90, histidine kinase mutL (GHKL) superfamily of adenosine triphosphatases. To better understand the consequences of disrupting Hsp90 function in a genetically tractable multicellular organism, we exposed the soil-dwelling nematode Caenorhabditis elegans to GA under a variety of conditions designed to optimize drug uptake. Mutations in the gene encoding C elegans Hsp90 affect larval viability, dauer development, fertility, and life span. However, exposure of worms to GA produced no discernable phenotypes, although the amino acid sequence of worm Hsp90 is 85% homologous to that of human Hsp90. Consistent with this observation, we found that solid phase-immobilized GA failed to bind worm Hsp90 from worm protein extracts or when translated in a rabbit reticulocyte lysate system. Further, affinity precipitation studies using chimeric worm-vertebrate fusion proteins or worm C-terminal truncations expressed in reticulocyte lysate revealed that the conserved nucleotide-binding fold of worm Hsp90 exhibits the novel ability to bind adenosine triphosphate but not GA. Despite its unusual GA resistance, worm Hsp90 appeared fully functional when expressed in a vertebrate background. It heterodimerized with its vertebrate counterpart and showed no evidence of compromising its essential cellular functions. Heterologous expression of worm Hsp90 in tumor cells, however, did not render them GA resistant. These findings provide new insights into the nature of unusual N-terminal nucleotide-binding fold of Hsp90 and suggest that target-related drug resistance is unlikely to emerge in patients receiving GA-like chemotherapeutic agents.

Fig 1.

Immobilized geldanamycin (GA) does not bind the heat shock protein 90 (Hsp90) ortholog expressed by C elegans. GA was derivatized and immobilized on agarose beads as described in Materials and Methods. (A) Tissue extracts. Beads were incubated in nonionic detergent lysates prepared from either mixed-stage worm cultures or the human cancer cell line PC-3M. Lysates were supplemented with soluble GA (18 μM) or an equal volume of dimethyl sulfoxide (DMSO) vehicle as indicated before addition of beads. Bound proteins were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblotted for Hsp90. As a control, total lysates (20 μg protein/lane) was also analyzed (lysate lanes) to verify the migration position and immunoreactivity of Hsp90 derived from the 2 sources. The migration positions of molecular size markers are indicated to the right of the panel. (B) Transcription-translation (TNT®) reactions. Constructs encoding either full-length chick or worm Hsp90 were translated in rabbit reticulocyte lysate supplemented with [³⁵S]methionine. GA beads were then incubated in these translation reactions after addition of either soluble GA (18 μM) or DMSO vehicle as indicated. Bound proteins were fractionated by SDS-PAGE and the radioactive translation products observed by autoradiography. As a control, aliquots (1 μL) of each 50 μL reaction were also analyzed to verify the presence of approximately equivalent amounts of the relevant translation product in the lysate. Molecular sizes are indicated to the right of the panel. Data presented are representative of experiments performed on 3 independent occasions

Fig 2.

Amino acid sequence alignment of heat shock protein 90 (Hsp90) orthologs. Identical and conservatively substituted residues in human Hsp90 (hHsp90, NCBI accession #NM 005348), C elegans DAF-21 (wHsp90, accession #074225), and S cerevisiae Hsp82 (yHsp90, accession #K01387) are highlighted in black. The highly conserved N-terminal nucleotide-binding domains are indicated by the dashed box

Fig 3.

Binding of heat shock protein 90 (Hsp90) chimeric fusion proteins to immobilized geldanamycin (GA). Transcription-translation reactions in reticulocyte lysate were programmed as indicated with plasmid constructs encoding either unmodified chick and worm proteins or chimeric proteins consisting of chick N-terminus fused to worm C-terminus (c/w) or worm N-terminus fused to chick C-terminus (w/c). The junction point for both fusions was located within the amino acid sequence QLFFRALL, which is identical in vertebrates and worms (see: Fig 2, wHsp90 residues 305–312). Left panel: After translation in the presence of [³⁵S]methionine, reactions were supplemented with soluble GA (18 μM) or an equal volume of dimethyl sulfoxide. GA beads were then added and incubated at 4°C with gentle agitation for 60 minutes. After extensive washing, bound proteins were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and radioactive translation products observed by autoradiography. Right panel: Aliquots of each translation reaction were analyzed by SDS-PAGE and autoradiography to monitor the relative amounts of each translation product generated in the various reactions. The “luc” lane contained material from a reaction that was programmed with plasmid encoding firefly luciferase as a positive translation control. The same experimental design was repeated on 2 other occasions, with similar results.

Fig 4.

The N-terminus of worm heat shock protein 90 (Hsp90) binds immobilized adenosine triphosphate (ATP) but not geldanamycin (GA). Transcription-translation reactions were programmed with plasmid constructs encoding the indicated chick or worm proteins in the presence of [³⁵S]methionine. (A) Binding to immobilized GA. Right panel: After translation, lysates were supplemented with soluble GA (18 μM) or dimethyl sulfoxide as indicated and the ability of GA beads to affinity precipitate translation products evaluated as in Figure 3. The migration positions of molecular size markers are indicated to the right of the panel. Left panel: Aliquots of each translation reaction were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography to verify the presence of the expected product in the lysate. (B) Binding to immobilized ATP. Reactions were programmed with plasmids encoding the N-terminus of either chick (amino acids 1–222) or worm (amino acids 1–210) Hsp90. Right panel: Reactions were supplemented with soluble ATP (10 mM) or an equal volume of water. Beads to which ATP was covalently bound by its γ-phosphate were then added and incubated at 4°C for 60 minutes with gentle agitation. Beads were washed and bound translation products observed by SDS-PAGE and autoradiography. Left panel: Aliquots of each translation reaction were analyzed on the same gel to verify the presence of the expected product in the lysate. Similar results were obtained when the same precipitations were performed on 2 other occasions.

Fig 5.

Evidence for dimerization of worm and vertebrate heat shock protein 90 (Hsp90). Full-length worm protein (left panel) or w/c chimera (right panel) was translated in the presence of [³⁵S]methionine. After supplementation of the reaction mixtures with either geldanamycin (18 μM) or dimethyl sulfoxide, immunoprecipitation of endogenous rabbit Hsp90 was performed using antibody raised to an amino-terminal epitope in vertebrate Hsp90α that does not exist in the worm ortholog (SPA-771, Stressgen). In some precipitations, primary antibody was omitted as indicated to demonstrate specificity of the precipitation conditions used. Coprecipitation of radioactive worm translation products was observed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and autoradiography. This experiment was repeated once, with similar results.

Fig 6.

Expression of worm heat shock protein 90 (Hsp90) in human cells. Replicate wells of human embryonic kidney cells (HEK 293 cells) were transfected with plasmids encoding either worm or chick Hsp90 under the transcriptional control of strong, constitutive cytomegalovirus promoter elements. As a control, wells were also transfected with the empty vector backbone, pcDNA 3.1(+). Forty-eight hours after transfection, cells were harvested by incubation in trypsin-ethylenediaminetetraacetic acid solution. Panel A: Cells from duplicate transfections with each plasmid construct were lysed in TNES buffer, and equal amounts of total cellular protein were analyzed by Western blotting with anti-Hsp90 antibody (clone AC88, Stressgen) followed by peroxidase-conjugated goat anti-mouse antibody and chemiluminescent detection. Panel B: Cells from duplicate transfections that had been refed 18 hours previously with medium containing geldanamycin (GA) (2 μM) or dimethyl sulfoxide were stained for flow cytometry to determine their relative levels of cell surface insulin-like growth factor receptor type 1. The data presented are the median channel fluorescence in arbitrary units of 10 000 events from the unimodal population of cells analyzed for each treatment condition. Only 1 value is reported for the “Vector control/(+) GA treatment” group in this experiment because the other sample for this condition was lost during processing. This experiment with minor modifications was repeated on 2 other occasions, with similar results.