13 February 2010. Histone deacetylase (HDAC) enzymes, best known as organizers of chromatin and regulators of gene expression, have been getting a lot of attention as targets for boosting neuronal survival and function. While the neuroprotective roles of HDACs have been mostly attributed to their effects on transcriptional regulation or DNA repair (see ARF related news story), that may not be the whole story. A paper out yesterday in Molecular Cell shows an unexpected link between HDAC4 and the toxicity of polyglutamate (poly-Q)-containing proteins, one that goes via chaperone proteins. In the work, Harm Kampinga and colleagues at the University of Groningen, The Netherlands, identify two heat shock proteins belonging to the DNAJ family of chaperones as potent protectors against the toxicity of poly-Q proteins. They then show that HDAC4 deacetylates and upregulates the chaperones’ activity. The results reveal a new activity for HDACs in protein homeostasis, and could open up new angles of attack on poly-Q diseases.
To identify chaperones that help cells withstand the insult of glutamine repeats, first authors Jurre Hageman and Maria Rujano systematically compared 22 different heat shock proteins (HSPs) for their ability to block aggregation of a 119-glutamine repeat-containing fragment of the huntingtin protein in cell overexpression experiments. They found that most of the chaperones of the Hsp70 family, which are good at protein refolding, were not very good at blocking aggregation. Instead, their survey revealed that two members of the Hsp40 family (DNAJB6b and DNAJB8) caused a dramatic reduction in aggregation based on protein solubility in gels, filter trap assay, or by confocal microscopy imaging of cytosolic inclusions. Along with that, the proteins were able to protect against poly-Q toxicity in cells in culture and in vivo in Xenopus tadpoles. Further, downregulation of endogenous DNAJB6b enhanced poly-Q aggregation and toxicity in cultured cells. The DNAJB6b and DNAJB8 proteins seemed to be general suppressors of poly-Q aggregations, since they worked against two other disease-causing poly-Q-expanded proteins, ataxin-3 and the androgen receptor. Their activity may also be highly conserved—some time ago the Drosophila ortholog of DNAJB6b was found to modify poly-Q toxicity in a fly model of Huntington disease (Kazemi-Esfarjani and Benzer, 2000).
How do the proteins work? The chaperones were unable to break up pre-formed aggregates, but seemed to prevent inclusions of newly expressed protein by forming large, disperse structures that included the poly-Q peptides. “We find they are very effective in the initial phase. As soon as aggregation starts, DNAJ proteins are there,” Kampinga told ARF.
In the canonical chaperone pathway, Hsp70 proteins serve as the refolding engine, and DNAJ proteins function to direct substrates to that engine. In this case, however, the activities of DNAJB6b and DNAJB8 largely bypass the classical folding pathway. Their activity did not require the induction of the heat shock response, and deletion of the N-terminal J domain, required for interaction with HSP70 and related proteins, only partly reduced their ability to prevent aggregation. However, a C-terminal region that functions in substrate binding was required for activity, and, in fact, was fully active on its own. That region, called the SSF-SST domain, was necessary for formation of the large chaperone complexes that associated with poly-Q peptides.
An Assist From Deacetylases
The researchers knew from previous work that the SSF-SST domain of DNAJB6b interacted with the histone deacetylase HDAC4 (Dai et al., 2005). In coexpression experiments, they confirmed the interaction and showed that the DNAJB6b and DNAJB8 proteins could bind two other deacetylases as well, HDAC6 and Sirt2. The association was functional, they showed, as the general HDAC inhibitor Trichostatin A blocked the anti-aggregation activity of DNAJB6b and DNAJB8. In addition, downregulation of HDAC4 was toxic to cells expressing the poly-Q protein, and DNAJB8 no longer could protect those cells.
Two conserved lysine residues in C-terminal region of DNAJB8 proved to be acetylated, and replacement of those sites with nonpolar alanine reduced deaggregation activity. The mutants could still bind substrate and form a poly-disperse network, suggesting that the lysine residues, and by extension HDAC4, were important not for binding, but for further processing of the peptide-chaperone complex.
Kampinga believes that the DNAJB6b and DNAJB8 serve a special function among chaperones, in that, rather than mediating protein refolding of full-length poly-Q-containing proteins, they act on poly-Q-containing peptides. “We think there is not a big problem in folding the full-length protein, but when it goes to get degraded, the proteasome spits out these poly-Q peptides because it cannot cleave them. This is controversial, but the idea is that those [peptides] may initiate the aggregation,” he said. The idea is that the DNAJB proteins would sequester those dangerous peptides, and Kampinga says he has some evidence to support that idea. “In collaboration with Eric Reits [Academic Medical Center, Amsterdam, The Netherlands], we found that exclusively DNAJB6b and DNAJB8 can prevent aggregation of synthetic poly-Q peptides,” he told ARF. “We really believe these are peptide scavengers.”
If that is the case, might the chaperones work for other toxic peptides, for example Aβ? Kampinga says they are working on this question, although it is not clear if the cytosolic chaperones would ever get at Aβ, which is mostly extracellular. “Although there are data now to suggest that extracellular Aβ needs to be internalized to become neurotoxic, we do not know whether, if that is true, the peptides need to be released from endosomes in order to become toxic,” Kampinga said. “Only in the latter case, one may envision that they then can be chaperoned and detoxified by the cytosolic DNAJB proteins we described.”
The results could explain the reported neuroprotective action of HDAC4 (Majdzadeh et al., 2008; Chen and Cepko, 2009), the authors say. However, with HDACs, things get complicated very quickly. HDAC inhibitors have been shown to be effective in fly and mouse models of Huntington disease (see ARF related news story on Steffan et al., 2001, and Thomas et al., 2008). HDAC inhibitors also increase memory function, possibly via HDAC2 (see ARF related news story on Guan et al., 2009). In contrast, other results suggest that inhibition of HDAC1 leads to neurodegeneration, at least in mice (see ARF related news story). Even as clinical work moves ahead on inhibitors (see ARF related news story), Kampinga’s work reminds us there is much more to learn about the substrate specificity and functional roles of the various HDACs that are becoming therapeutic targets.
A better alternative might be to target the chaperones directly, Kampinga says. HDAC effects are global, so he likes the idea of focusing on DNAJB6b or DNAJB8, rather than their modifiers. Their current work includes a study of the effects of boosting DNAJB6b expression in a Huntington’s mouse model, and screening for compounds that can increase the activity or level of chaperone proteins.—Pat McCaffrey.
Hageman J, Rujano MA, van Waarde MAWH, Kakkar V, Dirks RP, Govorukhina N, Oosterveld-Hut HMJ, Lubsen NH, Kampinga HH. A DNAJB chaperone subfamily with HDAC-dependent activities suppresses toxic protein aggregation. Molecular Cell. 2010 Feb 12; 37:355-369. Abstract