A number of studies have shown that class IIa HDACs are capable of participating in homotypic interactions. MITR associates with HDAC1, -3, -4, and -5 (6). Similarly, HDAC7 binds to HDAC1, -2, -4, and -5, as well as itself (15). A recent study presented a computational analysis, using the COILS program, of the predicted secondary structure of HDAC4 (70). The analysis assigned a high probability of a coiled-coil dimerization domain to the region encompassing amino acids 90 to 179 of HDAC4, a region that is fairly well conserved between HDACs 4, 5, and 9 but is absent in HDAC7. Deletion of this region disrupted the previously observed localization of HDAC4 to discrete nuclear bodies, the ability of HDAC4 to self-associate in vitro, and its sumoylation (37,70). However, this region also contains the MEF2-interacting domain of HDAC4, so whether or not the change in localization really has to do with self-aggregation rather than binding to MEF2 is an open question. Also, HDAC7, which diverges from the other family members through most of the putative dimerization region, is still found in the discrete nuclear bodies. It is nonetheless pleasing to speculate that this dimeric region of HDAC4, encompassing the MEF2 binding site, might interface in a bivalent manner with the similarly dimeric MADS/MEF2 domain. Resolution of the question must await either a solution of the MEF2-HDAC4 structures or an analysis of self-association in a more physiologic context.
Was this article helpful?