As discussed above, ATP-dependent remodeling factors are crucial components of the machineries that deposit histones and generate patterns of nucleosomes with diverse composition (). Apart from histones and their variants, however, there are other abundant non-histone architectural proteins, such as HMG proteins or heterochromatin protein 1 (HP1), that associate with the chromatin and shape its structure and dynamics. Do these factors also require motor proteins to bind correctly to the nucleosome fiber? Although this intriguing question has not been investigated in great detail so far, some recent studies provide evidence that suggests this may indeed be the case.
The HMG proteins are among those non-histone architectural proteins that have been studied most extensively [9
]. HMG proteins generally act to decrease the compactness of the chromatin fiber and therefore render chromatin more accessible to regulatory factors [147
]. They bind to chromatin in a highly dynamic and reversible way either by directly contacting the nucleosome and/or via co-factors. There are three subfamilies of HMG proteins, termed HMGA, HMGB and HMGN [9
]. Members of the HMGN group, in particular, have been shown to bind to nucleosomes at the entry/exit sites of the DNA and therefore compete with the linker histone H1 for nucleosome binding sites [150
]. They also exhibit exchange dynamics that are similar to those of H1 [152
]. As detailed above, H1 incorporation into chromatin is strongly dependent on the ISWI chromatin assembly factor [61
]. By analogy, HMGN proteins might also require an ATP-dependent factor for efficient chromatin association. In a recent study addressing the effects of HMGN1 and HMGN2 on chromatin remodeling by the ATP-dependent factors ACF and the SWI/SNF-family protein BRG1, it was shown that ACF can assemble extended periodic nucleosome arrays containing HMGN proteins in vitro
]. Although in vivo
studies have not yet been carried out, these experiments provide an intriguing hint for a possible function of ACF and potentially other chromatin remodeling factors in the assembly of not only histones but also of non-histone architectural proteins into chromatin.
Another candidate remodeling factor for the incorporation of non-histone chromosomal proteins may be ATRX. As discussed above, ATRX has recently been characterized to be required for the incorporation of histone H3.3 into pericentric and telomeric chromatin [100
]. Yet, ATRX has also been shown to physically interact with HP1, which is an abundant protein localized in heterochromatin [154
]. Two recent studies provided evidence for a function of ATRX in the loading of HP1 to chromatin. In Drosophila
, deletion of ATRX resulted in the loss of HP1α from pericentricβ-heterochromatin [157
]. Along the same lines, depletion of ATRX in mouse ES cells led to a strong decrease of HP1α localization at telomeric chromatin [110
]. Although these findings point to a role of ATRX in the association of HP1α with heterochromatin, biochemical studies will be necessary to determine, if indeed ATRX uses its catalytic activity to incorporate HP1 or if the observed phenotypes are the result of recruitment defects. Previous in vitro
experiments have demonstrated that the ACF remodeling factor greatly stimulates the association of HP1 with reconstituted nucleosome arrays. This stimulation, however, was found to be dependent on the Acf1 subunit of the complex and did not involve ATP-hydrolysis [158
]. Nevertheless, although the evidence is somewhat circumstantial at the moment, it will be interesting to elucidate whether non-histone architectural proteins are actual targets of ATP-dependent chromatin remodeling machines.