Inhibition of Hsp90 function delays and impairs recovery from heat shock Roger F Duncan 1,2 1 Department of Molecular Pharmacology and Toxicology, University of Southern California School of Pharmacy, Los Angeles, CA, USA 2 Department of Molecular Microbiology and Immunology, University of Southern California School of Medicine Los Angeles, CA, USA Exposure of cells to stress inactivates proteins and can result in cell death, unless protective responses occur [1,2]. Virtually all living organisms from bacteria to humans counteract the detrimental effects of cell stress by inducing a highly conserved response, termed the heat shock response or stress response [3]. The response was initially described and detailed in heat shocked Drosophila cells [4,5], among others. Subse- quent studies have shown that a wide array of stressful circumstances, including oxidative stress, abnormal protein accumulation, and endoplasmic reticulum per- turbations, induce a coordinated molecular response with two distinct facets: first, many metabolic proces- ses, including mRNA processing and translation, are severely inhibited, presumably to minimize the accu- mulation of deleterious protein forms. Second, there is a massive, rapid synthesis of a small group of proteins termed the heat shock proteins (Hsps), or stress pro- teins, that requires unique transcriptional induction and mRNA processing, and preferential translation [3]. A key feature of the heat shock response is its sup- pression following restoration of normal environmental conditions [3,6]. This entails the restoration of normal metabolic functions, such as protein synthesis, termin- ation of Hsp synthesis, and reduction in Hsp abun- dance. The latter features appear to be required Keywords geldanamycin; heat shock; herbimycin; Hsp90; Hsp function Correspondence R. F. Duncan, School of Medicine, Department of Molecular Microbiology and Immunology, University of Southern California School of Pharmacy Los Angeles, CA, USA Fax: +1 323 442 1681 Tel: + 323 442 1449 E-mail: rduncan@usc.edu (Received 6 July 2005, accepted 17 August 2005) doi:10.1111/j.1742-4658.2005.04921.x The induction of the heat shock response as well as its termination is auto- regulated by heat shock protein activities. In this study we have investi- gated whether Hsp90 functional protein levels influence the characteristics and duration of the heat shock response. Treatment of cells with several benzoquinone ansamycin inhibitors of Hsp90 (geldanamycin, herbimycin A) activated a heat shock response in the absence of heat shock, as reported previously. Pretreatment of cells with the Hsp90 inhibitors significantly delayed the rate of restoration of normal protein synthesis following a brief heat shock. Concurrently, the rate of Hsp synthesis and accumulation was substantially increased and prolonged. The cessation of heat shock protein synthesis did not occur until the levels of Hsp70 were substantially elevated relative to its standard threshold for autoregulation. The elevated levels of Hsps 22–28 (the small Hsps) and Hsp70 are not able to promote thermo- tolerance when Hsp90 activity is repressed by ansamycins; rather a suppres- sion of thermotolerance is observed. These results suggest that a multicomponent protein chaperone complex involving both Hsp90 and Hsp70 signals the cessation of heat shock protein synthesis, the restoration of normal translation, and likely the establishment of thermotolerance. Impaired function of either component is sufficient to alter the heat shock response. Abbreviations 2DGE, two-dimensional IEF ⁄ SDS ⁄ PAGE; eIF, eukaryotic initiation factor; EMSA, electrophoretic mobility shift assay; HSF, heat shock transcription factor; Hsp, heat shock protein; PI3K, phosphatidylinositol 3-kinase; SM, Schneider’s Drosophila growth medium. 5244 FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS because some Hsps are detrimental to normal growth and hence are sequestered or degraded during recovery [3,7]. Molecular programs that regulate the termination of the heat shock response thus significantly influence cellular and organismal health. The accumulation of a threshold level of Hsp70 sig- nals the cessation of Hsp mRNA transcription and the restoration of the normal rate of protein synthesis. This has been termed ‘autoregulation’ of the stress response, and is proposed to reflect the ability of Hsps to inhibit the heat shock transcription factor (HSF) [6,8,9]. Hsp70-dependent autoregulation of the stress response has been amply documented in cells from yeast to Drosophila to humans [8,10,11]. In this study we have investigated whether and how Hsp90 influences the duration or characteristics of the stress response, about which little is known to date. Hsp90 has been directly implicated in the regulation of HSF, partnering with Hsp70 to control its activity [12]. Through application of benzoquinone ansamycin anti- biotics that compete with ATP for the Hsp90 nucleotide binding site and inhibit Hsp90 activity [13,14], the role(s) of Hsp90 in the heat shock response can be straightforwardly investigated. We find that in Droso- phila cells latent HSF is activated by Hsp90 inhibition leading to a ‘heat shock response’ in the absence of stress, as reported previously for other cell types. When ansamycin-treated cells are heat shocked, the restoration of a normal metabolic function, protein synthesis, is sig- nificantly repressed, even though supra-physiological levels of Hsp70 accumulate due to the failure of these cells to terminate the heat shock response. In the absence of Hsp90 function cells cannot establish a thermo- tolerant state, even with the accentuated abundance of the other classes of Hsps. It has been well-established that Hsp70 collaborates with Hsp90 to carry out many of its chaperone functions [15,16], and these results indi- cate that such collaboration is required for efficient cessation of the heat shock response. Results Benzoquinone ansamycins induce a heat shock response in the absence of heat shock We hypothesize that Hsp90 function regulates facets of the heat shock response, as has been documented for Hsp70 [3,6,8]. Ansamycin Hsp90 inhibitors were used to determine whether two key elements of the heat shock response – (a) induction and suppression of heat shock protein synthesis, and (b) normal protein synthe- sis recovery – were altered when Hsp90 function was abrogated. Initially, the effects of ansamycins on non- heat protein synthesis rate and type were examined to exclude the possibility that effects observed during recovery from heat shock resulted from a general inhi- bition of the translational machinery. First, there was no overall inhibition or activation of translation rate as assessed by trichloroacetic acid (TCA) precipitation of pulse-labeled protein over 6 h of treatment (data not shown) and IEF ⁄ SDS ⁄ PAGE analysis (Fig. 1), thus excluding the possibility that a deficit in protein synthesis recovery from heat shock could be due to a direct effect of the inhibitors on the translational machinery. Second, there is a striking synthetic induc- tion of a limited number of proteins (Fig. 1). Benzo- quinone antibiotics induce the expression of heat shock proteins at the transcriptional level in mamma- lian cells, presumably because Hsp90 sequesters HSF in an inactive state [17–19]. Our results show the same induction of Hsps occurs in Drosophila, based on the precise comigration of the ansamycin-induced proteins with heat-induced proteins. Similar results were obtained with both ansamycins: geldanamycin Fig. 1. Ansamycins induce Hsp synthesis in the absence of heat shock in Drosophila cells. Drosophila S2 cells were incubated with 1 lgÆmL )1 geldanamycin or 1 lM herbimycin A, or left untreated. Cells were pulse-labeled for 15 min at intervals with 20 lCiÆmL )1 [ 35 S]methionine. Protein samples were prepared as described for 2DGE (see Experimental procedures) and analyzed by IEF ⁄ SDS ⁄ PAGE. Equal amounts of protein (equal cell numbers) were loaded onto each first dimension gel (based on Bradford assays, and confirmed by Co- omassie Brilliant Blue staining of the gels after electrophoresis). Labeled proteins were detected by autoradiography. The basic region of the gel is to the right, the acidic end to the left. Cells labeled after 6 h incubation without (A) or with (B) geldanamycin. (C, D) Cells labeled after 2 h incubation with or without herbimycin A. Only sectors containing Hsp70 ⁄ 68 (left sectors in C, D) and Hsp22 ⁄ 23 (right sectors in C, D) are shown. (E) EMSA analysis of ansmycin-treated and heat shocked cells. Whole cell lysates were prepared from cultures treated without or with herbimycin A, as well as a positive control heat-shocked culture. Each lysate was incubated with a 32 P-labeled DNA probe containing the HS element, analyzed by PAGE and autoradiography [41]; only the region corresponding to the migration of protein-bound, retarded mobility probe is shown. (F) Northern analysis. RNA was extracted from cultures treated without or with herbimycin A for 3 h, displayed by formaldehyde ⁄ agarose gel electrophoresis, probed with a 32 P-labeled DNA probe for Hsp70, and visualized by autoradiography. Only the region corresponding to the migration of Hsp70 mRNA is shown. (G) Western EMSA analysis of HSF phosphorylation. Extracts were pre- pared from cells treated with geldanamycin (upper) or herbimycin A (lower) were analyzed by SDS ⁄ PAGE and immunoblotting with anti-HSF sera; retarded bands have been correlated with HSF phosphorylation and activation [22]. An extract from cells heat shocked for 30 min at 36 °C is shown at the lower right, for reference. (P)HSF, phosphorylated HSF isoforms. nsb, nonspecific band. R. F. Duncan Ansamycins inhibit heat shock recovery FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS 5245 (Fig. 1A,B) and herbimycin A (Fig. 1C,D). The activa- tion of Hsp transcription by ansamycins at normal temperature is not subject to autoregulation as little to no difference in Hsp synthetic rate was observed between 2 and 6 h of treatment; e.g. panel B shows abundant Hsp synthesis after 6 h of treatment based on pulse-label analysis. Notably, as will be detailed subsequently, suppression of Hsp synthesis is observed when ansamycin-treated cells are heat shocked, then recovered. The activation of HSF was confirmed by DNA-binding gel shift assay (Fig. 1E) and northern blot analysis using Hsp70 probes (Fig. 1F). Curiously, HSF phosphorylation, which is frequently used as a surrogate marker for HSF activation, was not detected following ansamycin treatment and robust Hsp tran- scription (Fig. 1G), though phosphorylated isoforms can be readily detected following heat shock (Fig. 1G, lower right lane). A B F G E C D Ansamycins inhibit heat shock recovery R. F. Duncan 5246 FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS We were also interested to determine if evidence for Hsp90-mediated regulation of expression of other genes, either at a transcriptional or translational level, could be discerned that might implicate other path- ways that could influence the heat shock response. However, careful comparative examination of the 2D IEF ⁄ SDS ⁄ PAGE spot patterns of several untreated and treated samples failed to uncover any additional reproducibly altered spots (e.g. Fig. 1A,B). Benzoquinone ansamycins delay heat shock recovery and prolong Hsp synthesis To determine whether Hsp90 function affects characteristics of the heat shock response, geldanamycin-treated Drosophila S2 cells were heat shocked for 30 min at 36 °C, then restored to their nor- mal growth temperature. Samples were pulse-labeled for 15 min with [ 35 S]methionine ⁄ cysteine at intervals during heat shock and recovery. In keeping with the observation that geldanamycin did not inhibit protein synthesis per se, no difference in protein synthesis rate during heat shock or early in the recovery phase was observed (Fig. 2A). However, geldanamycin treatment causes a modest inhibition of protein synthesis rate during later recov- ery that is more evident at 3–4 h (Fig. 2A; at recovery times of 2 h and thereafter differences between untreated and ansamycin-treated cells were signifi- cantly negative at the P ¼ 0.05 level based on a paired t-test). As will be described, this overall measurement significantly underestimates the delay in recovery, because the synthesis of heat shock proteins dominates the overall measurement. Recovery rates for specific nonheat shock mRNA translation requires more pre- cise analysis, as will be presented below. Herbimycin treatment during heat shock and recovery produced results indistinguishable from those obtained with gel- danamycin (data not shown). To begin to address how the translation of specific mRNAs was affected, pulse- labeled samples were analyzed by one-dimensional SDS ⁄ PAGE. This provides an informative overview of translational changes, but is insufficient to accurately quantify individual protein synthetic changes. Several features stand out. First, the duration of heat shock protein synthesis is extended (Fig. 3); this is most evi- dent in the Hsp70 band, though detectable for all clas- ses of Hsp synthesis. Second, there is a delay in the restoration of normal translation. This is most evident in the 1D analysis in lanes wherein significant restor- ation is just beginning, e.g. in the 2 h recovery samples in Fig. 3B. These two events are likely coupled and both result from inadequate Hsp90 activity. To precisely characterize and quantify the influence of ansamycin treatments on normal translation recov- ery and Hsp synthesis, cells were pulse-labeled as des- cribed and proteins analyzed by two-dimensional IEF ⁄ SDS ⁄ PAGE, autoradiography, and densitometry (Fig. 4). From a qualitative perspective a significant impairment in normal translation restoration is evident in the reduced intensity of virtually all nonheat protein spots. To provide quantitative data a random sample of A B Fig. 2. Ansamycin treatment suppresses the recovery of normal pro- tein synthesis following heat shock. Drosophila S2 cells were pre- treated with ansamycin for 10 min, then heat shocked in a water bath equilibrated to 36 °C for 30 min, then returned to normal growth temperature for the duration of the experiment. Aliquots were pulse- labeled for 15 min with 20 lCiÆmL )1 [ 35 S]methionine at intervals. (A) Protein samples prepared as described (see Experimental proce- dures) were quantified for incorporation by TCA precipitation on GF ⁄ C filters and liquid scintillation counting. Results represent the average and standard deviation of 3–5 replicates with values from geldanamycin- and herbimycin A-treated samples combined. (B) Spe- cific protein synthesis was quantified by 2D IEF ⁄ SDS ⁄ PAGE, auto- radiography, and densitometry of an array of 14 nonheat shock protein spots (black bars; see Fig. 4 for identification) from samples pulse-labeled 3 h into recovery. R. F. Duncan Ansamycins inhibit heat shock recovery FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS 5247 spots (identified in Fig. 4A) was measured by densito- metry; the extent to which recovery of these spots was retarded is reported in Fig. 2B. For spots 1–9, the range of inhibition is  70–95%, with an average of 81% (black bars). Recovery of a minor class of spots was less affected by ansamycin treatment. Selected members of this group, labeled a-e, were inhibited only  40–60% (right five gradient-filled bars). In summary, ansamycin treatment significantly delays restoration of cell func- tion following heat shock, implicating Hsp90 activity in this process. Heat shock proteins 90, 70, 22 and 23 were similarly quantified. The rate of synthesis of these proteins achieves a higher maximum and remains activated longer in the ansamycin-treated cells (Fig. 4C–I; gel sectors are shown in panels C–E, and densitometric quantitation in panels F–I). For example, Hsp70 synthe- sis rapidly achieves its maximum synthetic rate in untreated cells at 1–2 h, and then rapidly declines, whereas in ansamycin-treated cells its synthetic rate at 2 h is approximately threefold elevated relative to untreated cells, and its synthesis at 3 h remains high, AB Fig. 3. Ansamycin effects on Hsp synthesis and normal protein synthesis recovery analyzed by SDS ⁄ PAGE. Cells were treated and labeled as described in Fig. 2, legend. Extracts were prepared using hSDS buffer (see Experimental procedures) and analyzed by SDS ⁄ PAGE and autoradiography. Equal amounts of protein (equal cell numbers) were loaded into each lane (based on Bradford assays, and confirmed by Coomassie Brilliant Blue staining of the gel after electrophoresis). (A) Cells pretreated with (right) or without (left) 1 lgÆmL )1 geldanamycin. (B) Cells pretreated with (right) or without (left) 1 l M herbimycin A. The location of bands corresponding to heat shock proteins Hsp83, Hsp70, Hsp26 ⁄ 28, and Hsp22 ⁄ 23, as well as actin, are indicted to the right. The extent of induction of Hsp90 during recovery varies; the herbimycin panels (without, with treatment) represent an example of extensive induction, which is not due to the herbimycin treatment itself (as it is equally observed in the untreated portion of the panel). This analysis has been repeated in part or completely three times; a repre- sentative result is shown. Fig. 4. Ansamycin effects on Hsp synthesis and normal protein synthesis recovery analyzed by 2-dimensional IEF ⁄ SDS ⁄ PAGE. Cells were treated and labeled as described in Fig. 2, legend. Protein samples were prepared as described for 2DGE (see Experimental procedures) and analyzed by IEF ⁄ SDS ⁄ PAGE. Equal amounts of protein (equal cell numbers) were loaded onto each first dimension gel (based on Bradford assays, and confirmed by Coomassie Brilliant Blue staining of the gels after electrophoresis), exposed to film and labeled proteins detected by autoradiography. For quantitation, films were scanned with a densitometer, and the intensity of spots determined using the Labworks software (UV Products). The basic region of the gel is to the right, the acidic end to the left. (A, B) Cells treated without or with herbimycin A beginning 10 min prior to heat shock ⁄ recovery; the 3 h recovery samples are shown. The location of spots corresponding to heat shock proteins Hsp90, Hsp70, Hsp26 ⁄ 28, and Hsp22 ⁄ 23, and nonheat shock protein actin, are indicated with arrows. These presentations are overexposed relative to the ones used for densitometric quantification (panels F–I, and Fig. 2B) to emphasize nonheat shock protein synthe- sis. Correspondingly, Hsp synthetic differences are significantly underrepresented due to film saturation. 14 nonheat shock protein spots used for monitoring nonheat shock protein recovery are indicated in panel A with numbers, letters above the spots (with the exception of ‘9’, which is written to the right of the spot); see text for details. (C–E) Sectors from 2D gels analyzing protein labeled at the indicated times. (F–I) Densitometric quantitation of Hsp synthesis. Black bars, without treatment; gradient-filled bars, ansamycin-treated. Ansamycins inhibit heat shock recovery R. F. Duncan 5248 FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS C AB DE GF HI R. F. Duncan Ansamycins inhibit heat shock recovery FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS 5249 similar to the rate in untreated cells 1–2 h earlier (and  fourfold higher than the rate in untreated cells at 3 h). For the other Hsps the onset of repression during the recovery interval is delayed relative to untreated cells, such that the maximum rate is synthesis occurs 1–2 h later. These elevated synthetic rates result in increased Hsp accumulation, as described in the next section. The prolonged rate of Hsp protein synthesis is paral- leled by a prolonged activation of the heat shock tran- scription factor HSF, as assessed by HSF phosphorylation (Fig. 5). Slower migrating HSF iso- forms are induced by heat shock and increase to  1h of recovery, then rapidly decline such that they are no longer detectable by 3 h recovery (Fig. 5A). On the other hand, phosphorylated HSF isoforms peak later (at  2 h into recovery) in ansamycin-treated cells, and remain in their activated form after they have disappeared from the untreated cells. These results confirm that ansamycins cause a prolongation of the heat shock response, which can be considered a deficit in autoregulation as Hsp70 levels are high. Notably, the transcription and translation of Hsp70 mRNA is largely suppressed by 5 h recovery in ansa- mycin-treated cells exposed to heat shock ⁄ recovery, whereas in cells exposed to ansamycin treatment at normal temperature its gene expression continues un- abated (Fig. 1). This suggests that high levels of Hsp70 are sufficient to suppress HSF activity even in the absence of Hsp90, albeit less efficiently than in its pres- ence. Similarly, restoration of normal metabolic func- tion (e.g. protein synthesis) occurs in the absence of Hsp90 function, but with delayed kinetics. The abundance of Hsps ultimately determines their functional significance. The influence of ansamycins on Hsp abundance during heat shock recovery was deter- mined by Coomassie staining 2DGE-resolved proteins, and densitometry. Throughout the first hours of recov- ery no significant differences in abundance of any Hsp was observed (data not shown), consistent with the sim- ilar rates of synthesis. Protein abundance at times cor- responding to the maximum Hsp accumulation (4–5 of recovery) showed modest to significantly elevated levels of Hsps (Fig. 6). For Hsp70, the major Drosophila heat shock protein and the one that primarily determines heat shock sensitivity and resistance, ansamycin treat- ment increased its levels by approximately threefold (see also Fig. 7, inset, for confirmation by western blot analysis). Hsp22 was approximately twofold increased, whereas Hsp23 and Hsp90 were 20–40% increased. These results suggest that ansamycins might enhance thermotolerance, with the caveat that Hsp90 is non- functional and might mitigate benefits provided by higher levels of the other Hsp classes. This issue is addressed in the next section. Benzoquinone ansamycins decrease thermotoler- ance To determine whether the increased Hsp abundance caused by ansamycin treatment had functional conse- quences on cells’ ability to tolerate stress, restoration of cell proliferation following a moderately severe chal- lenge heat shock (30 min at 38 °C) was determined, with or without ansamycin pretreatment. As noted above the contrasting influences of various Hsp classes could AB Fig. 5. Ansamycin prolongs the activity of HSF as assessed by phosphorylation. Cells were treated as described in Fig. 2, legend. Extracts were prepared using hSDS buffer (see Experimental procedures) and analyzed by SDS ⁄ PAGE and immunoblotting with antisera to Drosophila HSF. Equal amounts of protein (equal cell numbers) were loaded into each lane (based on Bradford assays, and confirmed by Amido schwartz staining of the gel after electrophoretic transfer). Cells without (A) or with (B) pretreatment with 1 l M herbimycin A. Locations of the major HSF band as well as the more slowly migrating phosphorylated isoforms [(P)HSF] are indicated to the right of panel B. Three to four differentiated bands are detected on the original film; nsb, nonspecific band. Ansamycins inhibit heat shock recovery R. F. Duncan 5250 FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS produce accentuated thermotolerance, reduced thermo- tolerance if Hsp90 function is crucial, or no distinguishable effect if the influences of distinct Hsp classes counterbalance each other. Without ansamycin treatment or a tolerance-indu- cing heat shock (control conditions) cell numbers progressively increase over several days (Fig. 7, open triangles; top curve). Addition of ansamycin to the cul- ture medium at reseeding had no effect, by itself, on cell growth (open diamond, day 4). Hence its suppres- sion of cell proliferation, described below, cannot be ascribed to a basal toxic effect. The heat shock ⁄ recovery protocol described in previous experiments produces no distinguishable effect on subsequent cell growth at normal growth temperature (solid diamonds). The more severe challenge heat shock modestly delays cell growth (open squares). In cells that have undergone a tolerance-inducing heat shock ⁄ recovery cycle, as detailed in the previous figures, the severe heat shock appears less detrimental as under these conditions cells increase in number indistinguishably from wholly untreated control cells (open circles). In summary, proliferation under the four cited regimens was indistinguishable. If cells are pretreated with ansamycin before the severe challenge heat shock then cell proliferation is significantly retarded (solid squares). Under these conditions modestly elevated levels of Hsps are syn- thesized, but the activity of Hsp90, and perhaps Hsp70 indirectly, is repressed. The elevated levels of Hsps provide no discernable stress-resistance benefits. If cells are both ansamycin-treated and provided a priming heat shock-recovery cycle, leading to sub- stantially higher levels of all Hsps (e.g. Fig. 7, inset), then the growth retardation is less than with ansa- mycin treatment alone (compare solid circles (both A C B Fig. 6. Accumulation of Hsps in ansamycin-treated and untreated cells following heat shock. Cells were treated and labeled as described in Fig. 2, legend. Protein samples were prepared as described for 2DGE (see Experimental procedures) and analyzed by IEF ⁄ SDS ⁄ PAGE. Equal amounts of protein (equal cell numbers) were loaded onto each first dimension gel (based on Bradford assays, and confirmed by Coomassie Brilliant Blue staining of the gels after electrophoresis). Only a sector of the entire gel is shown in the panels. (A) Sectors containing Hsp90 and Hsp70 from samples 4 h and 5 h of recovery from heat shock. The locations of Hsp68 and Grp78 (also referred to as Hsc72 in Droso- phila) are indicated as well. (B) Sectors containing Hsp22 and Hsp23, as in A. (C) Densitometric quantitation of Hsp abundance. The maxi- mum amount for each Hsp protein was set to 100%. Black bars, without treatment; gradient-filled bars, ansamycin-treated. R. F. Duncan Ansamycins inhibit heat shock recovery FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS 5251 treatments) to solid squares), but remains distinguish- ably reduced relative to cells heat shocked ⁄ recovered in the absence of ansamycin (compare open to solid circles). Similar results were obtained when the ansa- mycins were removed after  12 h exposure. Thus, even though heat shock protein production is accen- tuated by ansamycin treatment the abundant Hsps rather correlate with reduced thermotolerance when Hsp90 is rendered nonfunctional, implicating Hsp90 in restoration of cell proliferation following heat shock. Because ansamyin did not suppress cell prolif- eration in nonheat shocked cells, the results suggest that Hsp90 function is specifically needed to restore function to critical, heat-impaired targets. Results by others have suggested that ansamycin-induced Hsp expression can provide thermotolerance to mamma- lian cells (e.g. [18]). The discrepant results may reflect differential reliance on Hsp90 function in establishing thermotolerance, paralleling observations that the primary thermotolerance-inducing Hsp dif- fers between species ⁄ cell type [20]. Discussion Molecular chaperones have distinct, specific roles in cel- lular physiology as well as coordinated functions in pro- tein folding and refolding. In Drosophila, Hsp70 is the principal stress-induced chaperone and its activity has been shown to determine many features of the heat shock response [3]. In particular experimental manipula- tions that restrict Hsp70 accumulation or activity pro- long the heat shock response [3,6,8]. In this study we have addressed the requirement for Hsp90 function dur- ing the early inductive events of the heat shock response, in its cessation via autoregulation by Hsps, and for the activity of Hsps in providing thermotolerance. The results suggest that Hsp90 activity contributes to all aspects of the heat shock response, perhaps as part of a chaperone protein refolding system in coordination with Hsp70 [15,16,21]. Inhibition of Hsp90 by ansamycins does not influ- ence normal protein synthesis, either on a global or protein-specific scale. As reported by others studying mammalian cells [17–19], ansamycin treatment acti- vates HSF leading to the production of Hsps in Drosophila. Intriguingly, this appears to occur in the absence of detectable HSF phosphorylation, contrary to other chemically induced activations of HSF such as are caused by proteasome inhibitors [22]. Suppres- sion of Hsp90 function prolongs the activity of HSF and the synthesis of Hsp mRNAs following heat shock, paralleling the effect of blocking Hsp70 func- tion. Ultimately, HSF is inactivated suggesting that Hsp70 is sufficient to autoregulate HSF transcription even in the absence of Hsp90 function. As a conse- quence of ansamycin pretreatment, the rate of Hsp synthesis is maintained for a longer duration and excess amounts of each of the major classes of Hsps accumulate. However, the enhanced Hsp complement does not enhance restoration of cell proliferation fol- lowing a moderately severe heat shock, but rather sub- stantially reduces it. This suggests that Hsp90 function is critical for thermotolerance and heat shock recovery. This impairment is not a general consequence of Hsp90 inhibition, as retarded cell growth was only observed following heat shock, and not in cells main- tained at their normal growth temperature (the repor- ted cytotoxic effects of ansamycin treatment on tumor cells (e.g [23,24]). were not observed in Drosophila S2 cells). Thus, the effects are specific to the refolding of Fig. 7. Ansamycin-treated cells overexpressing Hsps are sensitized to cell growth inhibition by heat shock. Drosophila S2 cells were untreated or treated with ansamycin (Ansa Rx) and ⁄ or heat shocked in a water bath equilibrated to 36 °C for 30 min followed by recovery at normal growth temperature for 5 h (PreHS) and ⁄ or heat shocked in a water bath equilibrated to 38 °C for 30 min (chal- lenge, or Ch). Following the challenge heat shock (or in its absence for nonshocked cells) cells were placed in a 23 °C incubator for up to 4 days. Cell samples were removed at intervals for determin- ation of cell number by counting in a hemocytometer. The symbol designations are shown to right of the 4 d points. Experiments were performed using either geldanamycin (1 lgÆmL )1 ) or herbimy- cin A (1 l M); cell counts are combined. The day-by-day analysis for each condition was repeated three to eight times, and the average is shown. Deviation for all points overlaps, and so is not presented for clarity. For the 4 d points the deviations were: No challenge (n), ± 161; no challenge, PreHS (m), ± 284; PreHs (s), ± 100; chal- lenge only (h), ± 152; PreHs, Herb Rx (d), ± 191; no PreHs, Herb Rx (n), ± 152. Cells treated with ansamycin only [e, shown only for 4 d (no PreHS, no challenge)] increased cell numbers indistinguish- ably from the control cells (‘no challenge: n’). Inset: protein sam- ples removed after 5 h recovery from the PreHS, or after an equivalent incubation, without or with ansamycin treatment. (– ⁄ –), no treatment; (A ⁄ –), ansamycin, no PreHS; (– ⁄ H), PreHS, no ansa- mycin; (A ⁄ H), ansamycin present prior to PreHS and for the 5 h of recovery. Ansamycins inhibit heat shock recovery R. F. Duncan 5252 FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS heat-denatured proteins, or some other detrimental consequence of heat shock. Inhibition of Hsp90 function by ansamycins also sig- nificantly delays the recovery of normal, nonheat shock mRNA translation following heat shock. Global translation rates are modestly reduced during recovery (Fig. 2), but nonheat shock mRNA translation rates as determined by 2DGE analysis of specific protein spots reveals that the translation of most mRNAs is inhib- ited by 80% (Fig. 2) during intermediate stages of recovery ( 1–3 h). The quantitative discrepancy arises because heat shock mRNAs are super-induced by ansamycin treatment and constitute the majority of overall translation, masking the extensive inhibition of most nonheat shock mRNAs’ translation. A subset of nonheat shock mRNAs recovers more rapidly. These translation recovery patterns likely are linked to molecular events that regulate heat shock-mediated translation repression and subsequent restoration of the affected factors to their active state. As the molecu- lar mechanism for translation inhibition in Drosophila is not currently known, formulating a more precise hypothesis becomes speculative. For example, if inhibi- tion involves diminished eIF4F activity, as substantial evidence suggests [25–28], then as eIF4F activity is progressively restored during recovery mRNAs with the least dependence on eIF4F (e.g. those with rela- tively unstructured 5¢ untranslated regions) would be predicted to recover activity earlier. In the event where eIF4F activity restoration is delayed, as may occur when chaperone activity is inhibited by ansa- mycin treatment, mRNAs with very little requirement for eIF4F, such as the heat shock mRNAs and the subset of nonheat shock mRNAs with diminished eIF4F requirements, will be preferentially translated. Hsp90 functions as a homodimer to fold several key classes of proteins, including protein kinases and tran- scription factors [15,16,21]. In these roles it collaborates with Hsp70-family proteins, which bind to the newly synthesized proteins and transfer them to the Hsp90 folding machine via the intermediates Hop (for tran- scription factors) or cdc37 (for protein kinases) for final folding and stabilization of the activation-primed con- formation. Numerous specific protein substrates of this coordinated folding system have been identified, but with specific regards to protein synthetic events during the heat shock response the eIF2a protein kinase and Akt are noteworthy. eIF2 a phosphorylation has been implicated by many in the repression of protein synthe- sis during heat shock [25,29–31], and thermotolerance has been shown to correlate with accelerated restor- ation of eIF2a to its active state [31]. Hsp70 and Hsp90 association with the heat shock-activated eIF2a kinase regulates its activity [32,33]. Hence, impairment via Hsp90 and the ansamycins provides an avenue through which the heat shock response could be influenced. In addition, the PI3K ⁄ Akt kinase axis regulates protein synthesis rate and type in several circumstances, inclu- ding growth factor stimulation and recovery from stress [34,35]. Suppressing its activity via ansamycins [36] may delay recovery from heat shock paralleling the reduc- tion in recovery from other forms of stress observed by directly inhibiting targets of Akt such as Tsc2 and mTOR with U0126 and rapamycin, respectively [35]. A final potential target of Hsp90 is eIF4F itself, that has been observed to enter into heavy-sedimenting particu- late granules in heat-shocked mammalian cells [28,37]. If Hsp90 is involved in resolubilizing eIF4F from par- ticulate storage granules, and if this is critical for recov- ery of normal translation, then Hsp90 inhibition by ansamycin would be predicted to delay restoration fav- oring the translation of Hsp mRNAs and nonheat shock mRNAs with relatively low 5¢ untranslated region secondary structure. The influence of Hsp90 on heat shock ⁄ recovery is only moderate, suggesting that the other heat shock pro- teins that accumulate in supra-physiological amounts in ansamycin-treated cells can compensate for Hsp90 though less efficiently. This mirrors results in other cir- cumstances documenting partial redundancy in protein folding systems, based in part on the ability of an over- expressed chaperone class to substitute for a deficit in a wholly different class [21]. The contribution of Hsp90 function to the heat shock response is clearly documen- ted in this study, though in keeping with previous char- acterizations of the Drosophila heat shock response Hsp70 predominates. Considering that both heat and ansamycin treatments are used as cancer treatments [38], these results suggest that their coapplication could provide cooperative, more efficacious cell killing. Experimental procedures Chemicals Chemicals were purchased from Sigma Chemical Company (St Louis, MO, USA) unless otherwise indicated. Drug treatment, heat shock, [ 35 S]methionine labeling, and protein extraction and analysis Schneider S2 cells were cultured at 22–24 °C in Schneider’s Drosophila Medium (SM; Invitrogen, Inc., Carlsbad, CA, USA) containing 10% (v ⁄ v) fetal bovine serum, 20 mm l-glutamine, 100 unitsÆmL )1 penicillin, 0.1 mgÆmL )1 strepto- mycin, 0.25 mgÆmL )1 amphotericin B (Invitrogen). At the R. F. Duncan Ansamycins inhibit heat shock recovery FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS 5253 [...]... thermotolerance and heat -shock regulation in Drosophila New Biol 3, 1106–1120 7 Li D & Duncan RF (1995) Rapid turnover of heat shock protein 70 in Drosophila: stabilization during FEBS Journal 272 (2005) 5244–5256 ª 2005 FEBS R F Duncan 8 9 10 11 12 13 14 15 16 17 18 19 20 21 heat shock ,and consequent effects on accumulation Eur J Biochem 231, 454–465 DiDomenico BJ, Bugaisky GE & Lindquist S (1982) The heat shock. .. is self-regulated at both the transcriptional and post-transcriptional levels Cell 31, 593–603 Morimoto RI (1998) Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators Genes Dev 12, 3788–3796 Stone DE & Craig EA (1990) Self-regulation of 70-kilodalton heat shock proteins in Saccharomyces cerevisiae Mol... and structural characterization of the ATP ⁄ ADP-binding site in the Hsp90 molecular chaperone Cell 90, 65–75 Grenert JP, Sullivan WP, Fadden P, Haystead TA, Clark J, Mimnaugh E et al (1997) The amino-terminal domain of heat shock protein 90 (hsp90) that binds geldanamycin is an ATP ⁄ ADP switch domain that regulates hsp90 conformation J Biol Chem 272, 23843–23850 Pratt WB (1993) The role of heat shock. .. inhibit heat shock recovery 22 Kim D, Kim SH & Li GC (1999) Proteasome inhibitors MG132 and lactacystin hyperphosphorylate HSF1 and induce hsp70 and hsp27 expression Biochem Biophys Res Comm 254, 264–268 23 Bagatell R, Paine-Murrieta GD, Taylor CW, Pulcini EJ, Akinaga S, Benjamin IJ & Whitesell L (2000) Induction of a heat shock factor 1-dependent stress response alters the cytotoxic activity of hsp90- binding... of the heme-regulated eIF-2a kinase with heat shock proteins in rabbit reticulocyte lysates J Biol Chem 267, 18160–18167 33 Rose DW, Welch WJ, Kramer G & Hardesty B (1989) Possible involvement of the 90-kDa heat shock protein in the regulation of protein synthesis J Biol Chem 264, 6239–6244 5255 Ansamycins inhibit heat shock recovery 34 Mendez R, Myers MGJ, White MF & Rhoads RT (1996) Stimulation of. .. in regulating the function, folding, and trafficking of the glucorticoid receptor J Biol Chem 268, 21455–21458 Pratt WB & Toft DO (2003) Regulation of signaling protein function and trafficking by the hsp90 ⁄ hsp70-based chaperone machinery Exp Biol Med 228, 111–133 Murakami Y, Uehara Y, Yamamoto C, Fukazawa H & Mizuno S (1991) Induction of hsp 72 ⁄ 73 by herbimycin A, an inhibitor of transformation... Biol 10, 1622–1632 Mizzen LA & Welch WJ (1988) Characterization of the thermotolerant cell I Effects on protein synthesis activity and the regulation of heat- shock protein 70 expression J Cell Biol 106, 1105–1116 Zou J, Guo Y, Guettouche T, Smith DF & Voellmy R (1998) Repression of heat shock transcription factor HSF1 activation by HSP90 (HSP90 complex) that forms a stress-sensitive complex with HSF1... CW, Akinaga S & Whitesell L (2001) Destabilization of steroid receptors by heat shock protein 90-binding drugs: a ligand-independent approach to hormonal therapy of breast cancer Clin Canc Res 7, 2076– 2084 25 Duncan R & Hershey JWB (1984) Heat- shock induced translational alterations in HeLa cells: initiation factor modifications and the inhibition of translation J Biol Chem 259, 11882–11889 26 Panniers... Merrick WC & Henshaw EC (1985) Mechanism of inhibition of polypeptide chain initiation in heat- shocked Ehrlich cells involves reduction of eukaryotic initiation factor 4F activity J Biol Chem 260, 9648–9653 27 Hess MA & Duncan RF (1996) Sequence and structure determinants of Drosophila Hsp70 mRNA translation: 5¢-UTR secondary structure specifically inhibits heat shock protein mRNA translation Nucleic... (1993) Expression of a phosphorylation-resistant eukaryotic initiation factor 2 alpha-subunit mitigates heat shock inhibition of protein synthesis J Biol Chem 268, 12946–12951 31 Chang GC, Liu R, Panniers R & Li GC (1994) Rat fibroblasts transfected with the human 70-kDa heat shock gene exhibit altered translation and eukaryotic initiation factor 2 alpha phosphorylation following heat shock Int J Hyperthermia . Inhibition of Hsp90 function delays and impairs recovery from heat shock Roger F Duncan 1,2 1 Department of Molecular Pharmacology and Toxicology, University of Southern California School of. detrimental consequence of heat shock. Inhibition of Hsp90 function by ansamycins also sig- nificantly delays the recovery of normal, nonheat shock mRNA translation following heat shock. Global translation. two key elements of the heat shock response – (a) induction and suppression of heat shock protein synthesis, and (b) normal protein synthe- sis recovery – were altered when Hsp90 function was abrogated.