5-hydroxymethylcytosine in DNA repair: A new player or a red herring?

ABSTRACT

Active DNA demethylation performed by ten-eleven translocation (TET) enzymes produces 5-hydroxymethylcytosines, 5-formylcytosines, and 5-carboxylcytosines. Recent observations suggest that 5-hydroxymethylcytosine is a stable epigenetic mark rather than merely an intermediate of DNA demethylation. However, the clear functional role of this new epigenetic player is elusive. The contribution of 5-hydroxymethylation to DNA repair is being discussed currently. Recently, Jiang and colleagues have demonstrated that DNA damage response-activated ATR kinase phosphorylates TET3 in mammalian cells and promotes DNA demethylation and 5-hydroxymethylcytosine accumulation. Moreover, TET3 catalytic activity is important for proper DNA repair and cell survival. Here, we discuss recent studies on the potential role of 5-hydroxymethylation in DNA repair and genome integrity maintenance.

DNA cytosine (C) methylation (leading to 5-methylcytosine (5mC) formation) in vertebrates is a well-studied regulatory mechanism, participating in the control of gene expression, silencing of mobile genetic elements, X chromosome inactivation, etc.^1,2 Embryonic lethality of mutant maintenance or de novo DNA methyltransferases (DNMTs) supports the importance of DNA methylation in development.³ Moreover, lack of maintenance DNA methyltransferase (DNMT1) activity results in the death of differentiated mouse and human cells even in culture.⁴ A role of DNA methylation in maintaining genome stability is contradictory. DNA methylation-dependent silencing of DNA repair genes is associated with the development of a variety of tumors.⁵ On the other hand, some experimental results suggest that DNMT1 is directly involved in DNA repair and cell cycle checkpoint pathways.^6-8 It has been generally accepted for a long time that DNA demethylation could only occur passively through dilution of 5mC epigenetic marks in the course of DNA replication.⁴ However, active DNA demethylation pathways have been discovered recently. The role of activation induced deaminase (AID) and its paralogs in active DNA demethylation is a subject of intensive debates.^9,10 Although genetic studies showed that AID could function in DNA demethylation in particular cell types,^9,11 biochemical data question the relevance of these observations as AID deaminates C much more efficiently than 5mC.^12,13 The most appealing hypothesis suggests an indirect role of AID in active DNA demethylation. Specifically, AID may deaminate C and induce local DNA resynthesis as part of excision or mismatch repair.^14-16 In the course of this resynthesis adjacent 5mC may be converted to unmodified C. Another active DNA demethylation pathway is orchestrated by ten-eleven translocation (TET) methylcytosine dioxygenases, which sequentially convert 5mC into 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC) and 5-carboxylcytosine.^17-20 Final substitution of the modified bases (thymine and TET enzymes products) to C is usually performed by base excision repair machinery.^9,21 Since AID is not constitutively expressed in most tissues (except lymphoid and malignant cells), TET-dependent 5mC oxidation appears to be the main known pathway of active DNA demethylation. Recently it has been reported that 5hmC and 5fC could be considered also as stable epigenetic marks.^22,23 The discovery of specific 5hmC readers (other than the base excision repair factors) fueled the discussion on the functionality of 5hmC.²⁴ However, subsequent biochemical studies supported the specificity of only one 5hmC reader – UHFR2 E3 ubiquitin-protein ligase.²⁵ Thus, determination the functional role of 5hmC by studying its readers seems to be preliminary. Many of the known 5-hydroxymethylation-related effects wrap around the phenomenon of chromatin “opening” – the complex process necessary and/or accompanying activation of gene expression and its physiologically relevant consequences such as differentiation and dedifferentiation, induction of pluripotency, etc.^26-31 However, such effects might not be directly linked to the 5hmC formation, but simply be a consequence of active DNA demethylation. Therefore, investigation of the functional role of a possible new epigenetic mark, 5hmC, is an actual challenge.

The contribution of 5-hydroxymethylation to DNA repair is being discussed for a year. Phosphatidylinositol 3-kinase-related kinases (PIKKs)-dependent increase of 5hmC was reported by Jiali Li group at Kunming Institute of Zoology (Kunming, China).^32,33 According to the report published in the recent issue of EMBO reports,³³ DNA damage response (DDR)-activated ATR kinase phosphorylates TET3 in mammalian cells and promotes DNA demethylation and 5hmC accumulation (Fig. 1A).³³ Specifically, it was shown that DDR, initiated by ultraviolet (UV) or DNA topoisomerase I inhibitor camptothecin, led to a moderate (∼1.5-fold) increase in global 5hmC 3 hours post-treatment. Based on the results obtained, the authors stated that TET3 catalytic activity is important for proper DNA repair and cell survival.³³ Earlier the same group reported ATM-dependent TET1-mediated 5hmC production in mouse Purkinje cells (Fig. 1B).³² Together, these results clearly show that PIKKs, such as ATM and ATR, can phosphorylate TET enzymes (TET1 and 3, accordingly) and stimulate 5hmC production. While it is not clear how ATM-dependent phosphorylation affects TET1,³² the ATR-dependent TET3 phosphorylation prevents its degradation.³³ It appears that TET enzymes and/or the 5hmC epigenetic mark participate in a variety of DNA repair processes. This is apparent from the list of DNA damage triggers, used in these 2 studies: the ATM-dependent DDR was induced by the topoisomerase II poison etoposide, which is known to make double-stranded DNA breaks (DSBs); ATR-dependent DDR was provoked by either topoisomerase I inhibitor camptothecin (induces single-stranded DNA breaks (SSBs) and replication stress), or by UV (induces different types of lesions including pyrimidine dimers, SSBs, DSBs, oxidized bases, clustered lesions). However, the functional role of TET-mediated DNA oxidation in DNA repair pathways remained elusive. Though in TET-deficient cells ATR- and ATM-induced DDR signaling stayed unaffected,^32,33 the repair of the DNA lesions was somehow compromised.³³ Unfortunately, the authors did not analyze the temporal dynamics of DNA repair in cells lacking TET3 or its catalytic activity, to show whether such an inhibitory effect is DNA repair inhibition or delay. It is also unclear whether PIKK-induced DNA demethylation occurs specifically at sites of DNA damage. Only partial, if any, colocalization of ATR-induced 5hmC with histone γH2AX, well-known DDR marker, has been demonstrated.³³ Thus, a pan-nuclear 5hmC production might have no functional significance by itself, but simply reflect global DNA demethylation accompanying, for example, DNA repair gene expression activation.

Figure 1.

DNA damage induces TET-mediated DNA demethylation and stable 5hmC production. Jiang and colleagues demonstrated that different DNA damaging agents induce either ATR-dependent TET3-mediated or ATM-dependent TET1-mediated DNA 5-hydroxymethylation (A-B).^32,33 This 5-hydroxymethylation is pan-nuclear and is sparsely colocalized with DNA repair foci. By contrast, Kafer and colleagues reported that TET2 produces 5hmC perfectly colocalized with DNA damage sites (C).³⁴

Trying to puzzle this out, we should mention another paper focused on the potential role of 5hmC in DNA repair. The paper by Kafer et al. published a year ago in the Cell Reports,³⁴ presents a set of mainly immunocytological data that is not in perfect agreement with the abovementioned studies. Using laser microirradiation and prolonged aphidicolin treatment, Kafer et al. showed that active DNA demethylation and 5hmC production occurs specifically at sites of DNA lesions (Fig. 1C).³⁴ Perfect colocalization of 5hmC with several repair factors (53BP1, Rad51, γH2AX) was shown in human HeLa, HCC827 and A594 cells, exposed to microirradiation or aphidicolin. Moreover, 5hmC was enriched in so-called 53BP1 nuclear bodies even in non-treated cells. It was demonstrated that TET2 was responsible for this DNA damage-induced 5hmC production. Overall, this study showed the importance of TET enzymes in maintaining genome integrity under replication stress. The fact that active DNA demethylation may occur specifically at sites of DNA damage suggests that TET enzymes and/or 5hmC may have a direct role in DNA repair. Nevertheless, the mechanistic understanding of this phenomenon is far from being complete.

In spite of some limitations, the findings of Li and Carlton groups are still inspiring and fueling the fields of epigenetics and DNA repair. The 5hmC mark is the first reported DNA covalent modification arising at sites of DNA damage, and it is quite attractive to ascribe the mechanistic role in DNA repair to it. Several questions should be answered to advance this newly emerging field. It is important to find out whether TET activation/5hmC production is dependent on the types of DNA lesions. The possibility that 5hmC production is a by-product of TETs activation at sites of DNA damage necessary to remove 5mC should be addressed. Finally, it is necessary to consider a scenario that the contribution to the DNA repair is made mostly by TET enzymes but not 5hmC. Identification of the DNA damage-induced 5hmC readers and interacting partners of TET proteins is of crucial importance for the understanding of both 5hmC and TET proteins potential role in DNA repair.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

The work of OLK and SVR has been supported by the Russian Science Foundation (14–24–00022).