Manipulation of protein kinases reveals different mechanisms for upregulation of heat shock proteins in motor neurons and non-neuronal cells

Abstract

Motor neurons have a high threshold for induction of heat shock proteins (Hsps) in response to stress, a property associated with impaired ability to activate heat shock transcription factor 1 (Hsf1). Hyperphosphorylation of Hsf1 has been established as a requirement for transactivation of heat shock genes. This study demonstrated that the impaired heat shock response in motor neurons is not due to altered phosphorylation of Hsf1 by kinases previously shown to affect activation of Hsf1 in other cells (PKC, GSK3β, ERK1, CaMKIIα). However, a constitutively active form of CaMKIV induced robust expression of Hsp70, as well as transcription of a GFP reporter gene driven by the human inducible Hsp70 promoter in unstressed motor neurons, but not in mouse embryonic fibroblasts. The results point to novel mechanisms of activation of heat shock genes in motor neurons that have relevance to exploitation of endogenous stress responses therapeutically.

Introduction

A common cellular response to physiological and environmental stress is increased production of heat shock proteins (Hsps) (Bukau and Horwich, 1998, Hartl, 1996, Morimoto and Santoro, 1998). Evidence points to a chaperone-based mechanism controlling transactivation of genes encoding stress-inducible heat shock genes, whereby activation of the major, stress-responsive transcription factor, heat shock transcription factor 1 (Hsf1) is tightly controlled at a series of checkpoints (Voellmy, 2004). The first checkpoint is sequestration of monomeric Hsf1 in the cytoplasm by Hsp90/multichaperone complexes that repress the activity of two nuclear localization signals and several domains that promote trimerization (Baler et al., 1993, Rabindran et al., 1993, Zou et al., 1998). Damaged proteins (e.g., thermally denatured, oxidized, etc.) compete for constitutive, complex-forming Hsps, relieving this inhibition and allowing nuclear accumulation of Hsf1 homotrimers, which bind to heat shock elements (HSE) upstream of genes encoding stress response proteins including Hsps (Pelham, 1982, Wu, 1984, Zuo et al., 1995). The second checkpoint is that DNA binding is not sufficient for initiation of transcription. A subsequent reaction that involves hyperphosphorylation of Hsf1 is required to achieve the transcriptionally active state (Sarge et al., 1993, Sorger et al., 1987, Sorger and Pelham, 1988, Xia and Voellmy, 1997). There is also evidence of another Hsp90/multichaperone complex in the nucleus that binds and inhibits a subset of DNA-bound Hsf1 trimers (Guo et al., 2001). The displacement of these inhibitory complexes from Hsf1 by the nuclear protein Daxx greatly enhanced transactivation of heat shock genes in HeLa and HEK293 cells (Boellmann et al., 2004).

Stress-induced hyperphosphorylation of Hsf1 occurs concomitant with transactivation of heat shock genes and is detected by retardation of the electrophoretic mobility of Hsf1 in SDS-PAGE and Western blotting (Hensold et al., 1990, Xia and Voellmy, 1997). The importance of hyperphosphorylation is also demonstrated by studies showing impairment of Hsp70 induction (Erdos and Lee, 1994, Yamamoto et al., 1994) and Hsf1 activity (Baek et al., 2001, Ohnishi et al., 1999, Xia and Voellmy, 1997) by protein kinase inhibitors and activation by protein phosphatase inhibitors (Chang et al., 1993, Xia and Voellmy, 1997). Several studies have addressed the role of phosphorylation at specific amino acid residues and the kinases involved (Table 1).

Most studies of Hsf1 regulation have been carried out using homogenous, rapidly dividing cell lines. However, many post-mitotic neuronal subtypes fail to induce Hsps in response to stress (Brown and Rush, 1999, Kaarniranta et al., 2002, Manzerra and Brown, 1996, Marcuccilli et al., 1996, Maroni et al., 2003, Tonkiss and Calderwood, 2005). Motor neurons, the primary target cells in diseases such as amyotrophic lateral sclerosis (ALS), have a high threshold for inducing a heat shock response both in vivo (Manzerra and Brown, 1992, Manzerra and Brown, 1996) and in primary culture (Batulan et al., 2003). Non-steroidal anti-inflammatory drugs (NSAIDs), which in non-neuronal cells increase transactivation of heat shock genes in response to stress by enhancing DNA binding of Hsf1 trimers (Cotto et al., 1996, Jurivich et al., 1992, Jurivich et al., 1995), failed to overcome the high threshold for the heat shock response in motor neurons (Batulan et al., 2005). Overexpression of wild-type Hsf1 (Hsf1^wt) was also unable to complement the deficient stress response despite its accumulation in the nucleus of motor neurons (Batulan et al., 2003). However, overexpression of a constitutively active form of Hsf1 (Hsf1^act), with the regulatory domain deleted (amino acids 203–315), induced robust expression of Hsp70 and Hsp40 in cultured motor neurons (Batulan et al., 2006), suggesting that impaired phosphorylation of key residues in this region may account for the high threshold of activation.

This study examined whether manipulating phosphorylation at residues in the regulatory domain of DNA-bound Hsf1 trimers could trigger heat shock response in motor neurons by testing the previously described activation effect of phosphorylation at Ser²³⁰ by calcium/calmodulin-dependent kinase IIα (CaMKIIα) and inhibitory action of Ser³⁰³ phosphorylation by glycogen synthase kinase 3β (GSK3β). Protein kinase C (PKC), which had a variable effect on Hsf1 activation in previous studies, was also tested. Activation of PKC by phorbol ester, inhibition of GSK3β by lithium chloride (LiCl), and overexpression of two CaMKIIα splice variants all failed to induce Hsp70 in the presence or absence of heat shock. However, overexpression of a constitutively active form of CaMKIV (CaMKIV^act) did induce expression of Hsp70 in motor neurons, but likely through an Hsf1-independent mechanism.