Abstract
Chromatin conformation is tightly regulated within eukaryotic cells for the control of gene expression. Post-translational modifications, such as the addition of ubiquitin to a histone H2A C-terminal lysine residue (H2AK119Ub), play an important role in chromatin remodelling. Deposition of the H2AK119Ub modification is associated with chromatin condensation. The removal of this modification is catalysed by the Polycomb Repressive Deubiquitinase (PR-DUB) complex. The core of the mammalian PR-DUB complex is made up of the BRCA1-associated protein 1 (BAP1) and one of three Additional Sex Combs-like 1–3 (ASXL1–3) proteins, which interact to convey the deubiquitinase activity of the complex. Truncated variants of the ASXL proteins, lacking a C-terminal Plant homeodomain (PHD), have been identified across a range of cancers and developmental disorders. Although the BAP1 and ASXL interaction is well- characterised, this study aimed to investigate the regulatory roles of the ASXL PHD and the HB1, ASXL, and Restriction Endonuclease Helix-Turn-Helix (HARE-HTH) domains.
The PHD domain is hypothesised to regulate PR-DUB localisation via protein interactions. In previous work, structural analysis of the PHD domain has proven challenging. Therefore, for further study of the PHD domain to be carried out, possible interaction partners that could stabilise the PHD domain in complex were investigated. Within this work, the previously reported interaction between the PHD domain and the Methyl-CpG binding Domain (MBD) of MBD5 and MBD6 was successfully reproduced with recombinant proteins. The PHD-MBD complex and additional putative interactors were studied in silico and in vitro to assess the possible binding partners. Of the proposed interactors, only the PHD-MBD interaction was validated in vitro. Through AlphaFold 3 modelling, the PHD-MBD complex was predicted to bind two zinc ions, with an atypical hybrid zinc-binding site predicted at the interface of the domains. In vitro validation experiments supported this prediction, suggesting non- canonical roles for both the ASXL PHD and MBD domains of MBD5 and MBD6.
To investigate the putative DNA-binding activity of the HARE-HTH domain, AlphaFold 3 modelling of interaction with a previously proposed GC-rich DNA target sequence was performed. Interaction of the HARE-HTH domain with unmodified and methylated variants of a previously identified GC-rich sequence was explored in silico and in vitro, via an analytical ultracentrifugation experiment. From AlphaFold 3 modelling, interaction with methylated variants of the GC-rich sequence appeared possible. However, the results of the analytical ultracentrifugation experiment were inconclusive for binding.
Overall, these results contribute to the growing understanding of how ASXL1–3 proteins mediate the localisation of the PR-DUB complex. The PHD domain is suggested to contribute to the regulation of PR-DUB complex localisation via its stable interaction with the MBD domains of MBD5 and MBD6, forming a non-canonical hybrid zinc-binding site at the complex interface. From predictions of the HARE-HTH domain DNA-binding activity, interaction with methylated GC-rich sequences appeared possible. This interaction would provide a potential mechanism of PR-DUB complex localisation to the DNA, however, further investigation is required to validate this prediction.