The multi-functionality of UHRF1: epigenome maintenance and preservation of genome integrity

Abstract

During S phase, the cooperation between the macromolecular complexes regulating DNA synthesis, epigenetic information maintenance and DNA repair is advantageous for cells, as they can rapidly detect DNA damage and initiate the DNA damage response (DDR). UHRF1 is a fundamental epigenetic regulator; its ability to coordinate DNA methylation and histone code is unique across proteomes of different species. Recently, UHRF1's role in DNA damage repair has been explored and recognized to be as important as its role in maintaining the epigenome. UHRF1 is a sensor for interstrand crosslinks and a determinant for the switch towards homologous recombination in the repair of double-strand breaks; its loss results in enhanced sensitivity to DNA damage. These functions are finely regulated by specific post-translational modifications and are mediated by the SRA domain, which binds to damaged DNA, and the RING domain. Here, we review recent studies on the role of UHRF1 in DDR focusing on how it recognizes DNA damage and cooperates with other proteins in its repair. We then discuss how UHRF1's epigenetic abilities in reading and writing histone modifications, or its interactions with ncRNAs, could interlace with its role in DDR.

Figure 1.

Domains, interactors and post-translational modifications of UHRF1. (A) UHRF1 is a large multidomain protein consisting of five domains: ubiquitin-like domain (UBL), tandem tudor domain (TTD), plant homeodomain (PHD), SET and RING-associated domain (SRA) and really interesting new gene domain (RING), connected by linker regions that undergo post-translational modifications, conferring different conformational states to UHRF1 that regulate its stability and functions. Numbers below the structure show the amino acid position in human UHRF1 Isoform 1. (B) Through these domains UHRF1 can interact with different factors and recognise a pattern of histone modifications. Via the SRA domain, UHRF1 binds DNMT1, HDAC1 and PARP1, and also hemi-methylated DNA, while via the SRA and RING domains it binds to TIP60. Via PHD it recognizes unmodified H3R2 and H3K4 on chromatin, while via PHD and TTD it recognizes trimethylated H3K9. Dashed lines represent domain-specific interactions (C) UHRF1 is subject to different PTMs; in particular, it is phosphorylated, methylated and ubiquitinated in specific sites. These modifications are involved in its stabilization (S108, S639, K385, K500) and/or in the functionality (S661, K385) of the protein. S108 is phosphorylated by CK1δ, mainly following DNA damage; S639 is phosphorylated by the CDK1-cyclin B complex in M phase; K385 is methylated by SET8, required in the G2-M transition. All these modifications determine UHRF1 proteasome-dependent degradation via ubiquitination of K500 or of other unknown residues by UHRF1 or SCF^β-TrCP. K385 is methylated also by SET7 in S phase, in response to DSB; S661 is phosphorylated by CDK2/cyclin A during S phase, again in response to DBS. Both are required for UHRF1 recruitment at the site of DNA damage. Above are reported the specific factors responsible for the highlighted PTMs in the different cellular context; the connections are shown as arrows. The question marks on the arrow between UHRF1/SCF^β-TrCP and K500ub are determined by the absence of direct evidence. The SCF^β-TrCP complex ubiquitinate UHRF1 following phosphorylation of S108 by CK1δ, but the exact residue was not determined. Following methylation of K385 by SET8, ubiquitination of UHRF1 on K500 was observed, but the responsible E3 enzyme was not identified. Since UHRF1 can auto-ubiquitinate via RING domain, the exact E3 ligase responsible for K500ub remains to be elucidated.

Figure 2.

Mechanisms of DNA repair during S phase. The pathway chosen to recognize and repair DNA damage is strictly dependent on the specific type of damage, the cellular environment and the phase of the cell cycle. During S phase, factors of Fanconi anaemia pathway and homologous recombination have their highest expression. (A) In DNA ICLs the two complementary DNA strands are covalently bound to each other and cannot be separated. FANCM is phosphorylated by ATR and recruits the FA core complex on ICL site, that in turn ubiquitinates the FANCD2/FANCI heterodimer; once ubiquitinated they perform the nucleolytic incision at replication forks necessary to release the ICL from one of the two parental strands by binding SLX4. SLX4 recruits and activates endo­nucleases XPF–ERCC1, MUS81–EME1 and SLX1. In the complementary strand the lesion is bypassed while the corrected strand is ligated via MMR mechanism. This intact duplex will be the template for the repair of the DSB created in the other strand via homologous recombination. (B) DSB repair pathway choice is determined by the processing of DNA ends; 5′-to-3′ nucleolytic resection, leaving long 3′ DNA tails addresses the repair to HR. BRCA1 promotes the removal of 53BP1 from the damaged DNA, allowing resection by recruiting phosphorylated CtIP and the nucleolytic MRN complex. Phosphorylation of MRN activates it, starting the resection of 5′ ends together with EXO1 (and DNA2/BLM). The RPA complex coats the 3′ tails generated from resection, protecting them from further processing. RAD51, with the assistance of BRCA2, replaces RPA and catalyses homologous pairing and DNA strand exchange. Following strand invasion, the DNA synthesis proceeds, and the damage is repaired on the basis of the homologous sequence, mainly via synthesis-dependent strand annealing.

Figure 3.

UHRF1 roles in DNA damage response pathways.(A) During S phase UHRF1 can act as a sensor for interstrand crosslinks. It binds the damaged region through its SRA domain, together with UHRF2 (step 1), and recruits FANCD2 through RING domain (step 2). Once ubiquitinated by the FA core complex, FANCD2/FANCI activate the FA pathway (step 3). UHRF1 could also cooperate in the final step of ICR resolution by recruiting nucleases such as MUS81/EME1, via RING domain (step 4). The double-strand lesion produced by such nucleases is repaired in S phase via homologous recombination. (B) UHRF1 is involved in the recognition of double-strand breaks during S phase. Following DNA damage, UHRF1 is phosphorylated by CDK2/cyclin A on S661 and is subsequently methylated by SET7 on K385 (step 1). These two modifications are necessary for the recruitment of UHRF1 on damaged sites; phosphorylation of S661 is essential for UHRF1 interaction with BARD domain of BRCA1, methylation of K385 for the interaction with PARP1, which is also methylated by SET7 (step 2). Phosphorylated UHRF1 poly-ubiquitinates RIF1, dissociating it from 53BP1 and removing it from the damage; the removal of 53BP1 activates the 5′-to-3′ processing of DNA ends leading to the formation of 3′ single-strand tails recognized by the RPA complex and directing the repair towards HR (step 3). Methylated UHRF1 is also responsible for poly-ubiquitination of PCNA at K164. While mono-ubiquitination is commonly linked to processes of DNA damage tolerance pathways (DDT), poly-ubiquitination could be determined by persistence of PCNA on damaged end, representing a signal of the switch towards HR pathway (step 4).

Figure 4.

Hypothetical model representing the complexity of UHRF1 known or potential roles/interactions in DNA damage repair. Numerous aspects remain to be investigated about UHRF1 interactions and roles during the DNA damage response, as some of its known epigenetic abilities could be also involved in the detection of DNA damage and choice of repair pathway. The model attempts to integrate these different functions to show the complexity that could hide behind every player in the DDR, specifically UHRF1. (A) Recognition and binding of H3K9me3 and H3R2me0 by the TTD and PHD domains could be critical to UHRF1 positioning on damaged DNA; the presence of H3K9me3 could reinforce UHRF1 binding to TIP60, blocking TIP60 interaction with P53 and avoiding the activation of the apoptosis process. TIP60 would be free to acetylate ATM, starting the DDR. (B) The RING domain-dependent ubiquitination of H3K18 and H3K23 could facilitate HR repair by blocking their acetylation, detected in DSB repair via NHEJ, or functioning as docking sites for HR proteins. Trimethylation of H4K20, performed in the presence of UHRF1 and PARP1, could prevent 53BP1 recognition of H4K20me1/2, further impairing the NHEJ repair pathway. (C) UHRF1 is a known RNA-binding protein via SRA domain. Following DNA damage, it could also bind the RNA:DNA hybrids formed in S/G2 phase together with BRCA1, possibly via TTD, facilitating HR; the hypothesised mechanism is similar to the binding of RNA:RNA duplex by the Tudor domain of 53BP1 that directs the repair towards NHEJ.