Multiple components of the HSP90 chaperone complex function in regulation of heat shock factor 1 In vivo

Abstract

Rapid and transient activation of heat shock genes in response to stress is mediated in eukaryotes by the heat shock transcription factor HSF1. It is well established that cells maintain a dynamic equilibrium between inactive HSF1 monomers and transcriptionally active trimers, but little is known about the mechanism linking HSF1 to reception of various stress stimuli or the factors controlling oligomerization. Recent reports have revealed that HSP90 regulates key steps in the HSF1 activation-deactivation process. Here, we tested the hypothesis that components of the HSP90 chaperone machine, known to function in the folding and maturation of steroid receptors, might also participate in HSF1 regulation. Mobility supershift assays using antibodies against chaperone components demonstrate that active HSF1 trimers exist in a heterocomplex with HSP90, p23, and FKBP52. Functional in vivo experiments in Xenopus oocytes indicate that components of the HSF1 heterocomplex, as well as other components of the HSP90 cochaperone machine, are involved in regulating oligomeric transitions. Elevation of the cellular levels of cochaperones affected the time of HSF1 deactivation during recovery: attenuation was delayed by immunophilins, and accelerated by HSP90, Hsp/c70, Hip, or Hop. In immunotargeting experiments with microinjected antibodies, disruption of HSP90, Hip, Hop, p23, FKBP51, and FKBP52 delayed attenuation. In addition, HSF1 was activated under nonstress conditions after immunotargeting of HSP90 and p23, evidence that these proteins remain associated with HSF1 monomers and function in their repression in vivo. The remarkable similarity of HSF1 complex chaperones identified here (HSP90, p23, and FKBP52) and components in mature steroid receptor complexes suggests that HSF1 oligomerization is regulated by a foldosome-type mechanism similar to steroid receptor pathways. The current evidence leads us to propose a model in which HSF1, HSP90 and p23 comprise a core heterocomplex required for rapid conformational switching through interaction with a dynamic series of HSP90 subcomplexes.

FIG. 1

Recognition of the heat-shock-activated DNA-binding form of HSF1 in gel mobility supershift assays with antibodies against components of the HSP90 chaperone complex. (A) Antibodies (Ab) against HSP90 and the array of cochaperones (indicated above the lanes) were incubated with identical aliquots of a heat-shocked oocyte extract in DNA-binding reactions with ³²P-labeled HSE probe, and the electrophoretic migration of HSF1-HSE complexes was analyzed. (B) Same as above except that oocytes were stressed with 70 mM salicylate for 1 h. The position of the nonsupershifted heat-induced HSF1 complex is indicated at the left.

FIG. 2

Components of the HSP90 chaperone complex are present in the oocyte nucleus. Oocytes were incubated under nonshock (NS) or heat shock (HS) conditions, and proteins from single intact oocytes or from the nuclear and cytoplasmic fractions were separated by SDS-PAGE. Chaperone components were detected with corresponding antibodies (see Materials and Methods) by immunoblotting. In these blots, nuclear lanes contained extract approximately equivalent to one nucleus, except for the IκB blot, in which nucleus lanes contained the equivalent of 10 nuclei.

FIG. 3

Effects of elevated levels of HSP90 and Hsp/c70 on the HSE-binding activity of HSF1 under nonshock conditions, during heat shock, and during recovery. The levels of Hsp/c70 (top) and HSP90 (middle) were elevated by direct microinjection of purified proteins into nuclei. Similar amounts of BSA were injected into nuclei (bottom), and then oocytes were incubated at nonshock (NS) temperatures, heat shocked (time zero), or heat shocked and allowed to recover at the nonshock temperature (for times indicated at the top). Left, gel mobility shift assays comparing HSF1 activity in oocytes with elevated Hsps and identically treated uninjected controls; right, immunoblots showing relative levels of HSP90 or Hsp/c70 in uninjected (U) and injected (Inj.) oocyte nuclei.

FIG. 4

Effects of elevated levels of components of the HSP90 chaperone machine on HSF1. The levels of Hip and Hop (A) and the immunophilins Cyp-40, FKBP51, and FKBP52 or p23 (B) were increased by microinjection of corresponding mRNAs, and HSF1 was compared with uninjected controls by gel shift assay (left). Immunoblots showing relative levels of cochaperones in mRNA-injected oocytes (Inj.) and uninjected (U) controls are shown on the right.

FIG. 5

Effect on HSF1 after immunotargeting of HSP90 chaperone components with microinjected antibodies. (A) Left, gel mobility shift assay of uninjected or antibody (Ab)-injected oocytes that were incubated at nonshock temperatures (NS), heat shocked for 1 h (time zero), or heat shocked and then allowed to recover at the nonshock temperature (for times indicated at the top). Right, gel mobility shift assay showing HSF1 induction upon HSP90 or p23 antibody injection. (B) Gel mobility shift assay showing the effect of microinjected antibodies against NF-κB, IκB, PTP-1B, and CREB on HSF1. (C) Immunoblots showing presence of antibodies in injected oocytes. Antibodies were detected with the appropriate secondary antibodies.

FIG. 6

Effect of microinjected chaperone antibodies on HSF1-mediated transcription in unstressed cells. Oocytes were injected with hsp70-CAT and antibodies (Ab.) against each component of the HSP90 complex as indicated. CAT expression in antibody-injected oocytes was compared to that in similarly treated nonshocked (NS) or heat-shocked (33°C, 1 h; HS) samples without injected antibody (−). CAT assays were performed as described in Materials and Methods.