|Home | About | Journals | Submit | Contact Us | Français|
Heterochromatin protein 1 (HP1) family members are versatile proteins involved in transcription, chromatin organization, and replication. Recent findings now have implicated HP1 proteins in the DNA damage response as well. Cell-biological approaches showed that reducing the levels of all three HP1 isoforms enhances DNA repair, possibly due to heterochromatin relaxation. Additionally, HP1 is phosphorylated in response to DNA damage, which was suggested to initiate the DNA damage response. These findings have led to the conclusion that heterochromatic proteins are inhibitory to repair and that their dissociation from heterochromatin may facilitate repair. In contrast with an inhibitory role, a more active role for HP1 in DNA repair also was proposed based on the finding that all HP1 isoforms are recruited to UV-induced lesions, oxidative lesions, and DNA breaks. The loss of HP1 renders nematodes highly sensitive to DNA damage, and mice lacking HP1β suffer from genomic instability, suggesting that the loss of HP1 is not necessarily beneficial for repair. These findings raise the possibility that HP1 facilitates DNA repair by reorganizing chromatin, which may involve interactions between phosphorylated HP1 and other DNA damage response proteins. Taken together, these studies illustrate an emerging role of HP1 proteins in the response to genotoxic stress.
Mammalian cells contain three closely related heterochromatin protein 1 (HP1) isoforms: HP1α, HP1β, and HP1γ. The HP1 proteins localize to constitutive heterochromatin, such as centromeres and telomeres, and rapidly exchange between freely diffusing and chromatin-bound states within several seconds (6). All three HP1 proteins dimerize through their C-terminal chromoshadow domain (CSD), leading to homo- and heterodimeric HP1 molecules (20). Upon dimerization, two CSDs create a hydrophobic interaction surface that binds proteins containing a PXVXL peptide motif (9, 26). The chromodomain (CD) located at the N terminus of HP1 binds with moderate affinity (KD [equilibrium dissociation constant] ~ 1 μM) to histone H3 methylated at lysine 9 (H3K9me) in vitro. The binding of HP1 to constitutive heterochromatin depends on the enzymatic activity of methyltransferase Suv3-9H1/2 (KMT1A/B), which trimethylates H3 at lysine 9 (6), while binding to euchromatin depends on H3K9me2 mediated by methyltransferase G9a (KMT1C) (24, 25). Additional histone methylation-independent interactions between HP1 and chromatin depend on the CSD, possibly through interactions with PxVxL-containing chromatin proteins, and involve interactions with an RNA component, possibly through the hinge motif of HP1 (16). Mammalian HP1 proteins modulate gene transcription in both euchromatin and heterochromatin, play roles in DNA replication, and are involved in the assembly and maintenance of constitutive heterochromatin (15, 21, 23).
A number of recent studies have linked HP1 proteins to the DNA damage response (3, 10, 14, 29). Some studies concluded that HP1 is inhibitory to repair associated with constitutive heterochromatin, suggesting that DNA repair benefits from its dissociation from heterochromatin (3, 10). In contrast, other studies revealed that HP1 proteins may be involved in facilitating DNA repair pathways activated by oxidative lesions (29), UV-induced DNA lesions, and chromosomal breaks (14).
The irradiation of cells with UV light results in lesions in the DNA (6-4 photoproduct [6-4 PP] and cyclobutane dimer [CPD]). In mammals, these lesions are removed exclusively by the nucleotide excision repair (NER) system. NER is initiated either by the stalling of RNA Pol II at a lesion (transcription-coupled repair) or by the recognition of DNA lesions by the XPC complex (global genome repair). Following lesion recognition, TFIIH, presumably in cooperation with RPA, unwinds the DNA near the lesion (7, 8). XPA, ERCC1-XPF, and XPG are recruited to the NER complex and catalyze incisions on both sides of the lesion, releasing an ~30-nucleotide fragment containing the lesion (5, 28). The generated single-stranded DNA gap is subsequently filled in by the DNA replication machinery (18).
It was shown recently that all three HP1 homologues (HP1α, HP1β, and HP1γ) are recruited to UV-induced DNA lesions in both human and mouse cells (Fig. (Fig.1A)1A) (14). Surprisingly, the recruitment of HP1 proteins does not depend on the recognition and/or processing of lesions through either global genome repair or transcription-coupled repair. In addition, the CD-mediated interaction with H3K9me3 is dispensable for recruitment to UV lesions, while the CSD is essential to target HP1 to lesions (14). This suggests that a PXVXL-containing protein binds lesions independently of lesion recognition proteins in NER (XPC or DDB2) and subsequently recruit HP1 (see Fig. Fig.3).3). Alternatively, HP1 proteins may recognize structural chromatin alterations caused by UV lesions (see Fig. Fig.3);3); this is reminiscent of the recruitment of HP1 to perturbed heterochromatin (30).
The importance of HP1 in the response to UV damage is underscored by the severe UV sensitivity of nematodes lacking functional HP1 proteins (14). Nematodes encode two HP1 (HPL-1/2) isoforms, which have redundant roles in UV lesion repair. These findings suggest a general requirement of HP1 proteins in the response to UV lesions, which may involve the unfolding of higher-order chromatin to facilitate repair.
Oxidative events, ionizing radiation (IR), and alkylating agents cause base alterations that give rise to oxidative DNA lesions. A frequently occurring highly mutagenic lesion is 8-oxoguanine (8-oxyG), although other oxidative lesions are poorly mutagenic but strongly inhibit replicative DNA polymerases, making these lesion highly cytotoxic (1). Base excision repair (BER) removes oxidative lesions by employing a variety of DNA glycosylases that each act on a different subset of base alterations (13). Recent results from the Dobrucki laboratory showed that HP1α, HPβ, and HPγ are recruited to sites of oxidative damage together with OGG1 and XRCC1, which are known to be involved in BER (29). The HP1 proteins slowly accumulate for ~30 min following damage induction, after which the bound levels gradually decrease within 10 h. Additional studies are needed to address whether the recruitment of HP1 depends on functional BER and whether the loss of HP1 proteins leads to an increased sensitivity to oxidative DNA damage.
Double-strand DNA breaks (DSBs) are induced by different sources, including IR and antitumor agents (e.g., etoposide). The presence of DSBs in the genome triggers the activation of cell cycle checkpoints and DNA repair pathways (4). The repair of DSBs is mediated by homologous recombination repair (HR), which requires a homologous template and therefore is active mainly during S and G2 phases of the cell cycle, or by nonhomologous end joining (NHEJ), which repairs DSBs mainly in G1 (12).
Several studies have established that chromatin (in particular, heterochromatin) is inhibitory to repair (10, 19). Recently, an ATM-dependent signaling cascade was uncovered that involves the phosphorylation of KAP-1, resulting in chromatin relaxation (10, 31). This signaling event is no longer required for efficient repair in Suv3-9h1/h2-deficient cells or upon the knockdown of KAP-1 or HP1 proteins (10). It was suggested that reducing the levels of HP1 proteins would result in the relaxation of heterochromatin, thereby bypassing the requirement for ATM signaling.
Based on these results, it would be expected that the loss of HP1 is beneficial to repair in heterochromatin (10, 11). Indeed, the increased mobility of HP1β at damaged sites in heterochromatin was observed recently based on the transient dispersal of green fluorescent protein-HP1β following the microirradiation of cells that were sensitized to laser light with Hoechst (3). The dispersal of HP1β at damaged sites was suggested to be the result of the phosphorylation of the T51 residue. The phosphorylation of this conserved residue, which is located in the CD of HP1, is triggered by IR, etoposide treatment, and laser-assisted methods (3), and it inhibits the interaction between the CD and H3K9me2 peptide in vitro. Based on these findings, the authors concluded that the DNA damage-induced T51 phosphorylation of HP1β triggers its dissociation from chromatin to facilitate the DNA damage response. Following the transient dissociation of HP1β from H3K9me (~5 min), HP1β was found to reaccumulate at damaged sites, leading to the restoration of the bound levels of HP1 at heterochromatic sites (3). Although not experimentally addressed by Ayoub and coworkers, their model would predict that the reaccumulation of HP1β to damaged sites depends on the binding to H3K9me and therefore on the CD of HP1β.
Another recent study that addressed the role of HP1 proteins in the DNA damage response suggested an alternative explanation for the observed reaccumulation of HP1β at damaged sites. In this study, several methods to inflict DNA damage, including α-particles, soft X rays, and laser-assisted methods, were employed to study the behavior of all three HP1 proteins in response to DSBs (14). HP1 proteins were shown to accumulate de novo at sites of DNA damage. Additionally, the recruitment of HP1 proteins was mediated by the CSD, while HP1 recruitment was independent of the CD and independent of binding to H3K9me (Fig. (Fig.1B).1B). Although it cannot be excluded that there is some HP1 dissociation from H3K9me at damaged sites, these results indicate that there is a net increase of HP1 at damaged chromatin (Fig. (Fig.22).
These results are not inconsistent with the data from the Venkitaraman laboratory, because they also found the accumulation of HP1β at sites of DNA damage (3). However, given that HP1 was found to be targeted to damaged sites in a CD-independent manner, the available data strongly suggest that the accumulation of HP1 reflects H3K9me-independent binding to damaged sites rather than the rebinding of HP1 to H3K9me sites (3, 14). It should be noted that the transient dissociation of HP1 proteins from damaged sites, as reported by Ayoub and coworkers, was not observed in the study from the van Driel laboratory (14), even when the same methodology was applied. Differences in the numbers of iterations of a high-intensity laser used to inflict local damage may underlie this discrepancy.
What is the function of the T51 phosphorylation of HP1 following DNA damage? Ayoub and coworkers suggested that T51 phosphorylation results in the transient dissociation of HP1β from damaged sites (3). In this model, the function of T51P would be to relax heterochromatin transiently to facilitate the DNA damage response. However, phosphorylated HP1β appears to remain localized to the site of DNA damage (3), which is unexpected if T51P lowers the affinity of HP1 to chromatin. Additionally, the authors of this study were unaware of the recruitment of HP1 to damaged areas, which does not depend on the CD (containing the T51 residue) or on H3K9 methylation (14). An alternative function of T51P could be to create an epitope on damage-bound HP1, which allows phosphorylation-dependent interactions with proteins that are involved in the DNA damage response (Fig. (Fig.3).3). Interestingly, a number of proteins that contain BRCT and/or FHA domains, which mediate phosphorylated protein interactions, are known to associate with sites of DNA damage, including MDC1, NBS1, RNF8, and TopBP1 (17). Of particular interest is the FHA domain, because it specifically interacts with phosphorylated threonine residues. Thus, proteins harboring an FHA domain may interact with damage-bound HP1 through phosphorylation-dependent interactions to facilitate DNA repair (Fig. (Fig.3).3). Additionally, as the targeting of HP1 to chromatin causes heterochromatinization (27), the phosphorylation of T51 may ensure that the recruited HP1 molecules do not induce the formation of heterochromatin at sites of DNA damage. Studying the time-dependent distribution of phosphorylated HP1β in the nucleus after DNA damage may shed more light on the function of this modification.
The function of HP1 in the DNA damage response was addressed recently by monitoring the sensitivity of single and double HP1 mutant nematodes to DNA damage (14). Remarkably, the sensitivity of double mutants to IR is comparable to the sensitivity of wild-type animals. However, animals lacking only one of the HP1 isoforms (HPL-1) are less sensitive to IR than wild-type animals, while animals lacking the other isoform (HPL-2) are extremely sensitive to IR. This suggests opposing functions of the HP1 isoforms in response to IR. One possibility is that the decreased sensitivity of HPL-1 animals to IR is due to an altered, possibly less compact chromatin structure in these mutants, similar to what presumably occurs upon reducing the levels of mammalian HP1 (10). Similarly to this phenomenon, mammalian cells with the reduced expression of linker H1, leading to a more accessible chromatin structure, were shown to display decreased sensitivity to IR and the enhanced signaling of ATR kinase (19). On the other hand, the extreme sensitivity of HPL-2 mutant worms suggests that this HP1 isoform has an active role in the repair of DNA breaks. Although it is not clear if these results can be extrapolated to mammalian cells, mammalian HP1 isoforms also may have divergent roles in genome maintenance. Indeed, neuronal cells that were derived from HP1β-deficient mice but not from HP1α-deficient animals display genomic instability (2). How do these results compare to the lack of repair phenotype in cells with the reduced levels of HP1 observed by Goodarzi and coworkers (10)? The latter study used G0 cells, meaning that DSBs are repaired mainly by NHEJ. One possibility is that loss of HP1 impacts HR more than it does NHEJ. Another possibility is that the knockdown of HP1 proteins is not complete, resulting in the residual activity of HP1 proteins in repair. Finally, it could be that differences between mammalian cells and nematodes underlie the seemingly different phenotypes on IR sensitivity caused by the loss of HP1 proteins. Future studies using the recently developed HP1α, HP1β, and double mutant mouse cells will help elucidate the different roles of HP1 isoforms in the DNA damage response (2).
The novel role of the HP1 proteins in the DNA damage response expands our understanding of the versatility of HP1 proteins in chromatin biology. Some important questions remain unanswered. What function do the different HP1 isoforms have in the DNA damage response? How are HP1 proteins recruited to areas of DNA damage? What is the role of HP1 phosphorylation? Is the function of HP1 in DNA repair somehow related to the interaction between KAP-1 and the CSD of HP1? The findings discussed in this minireview will open up new venues of research, which ultimately will give insight into the mechanisms that render chromatin permissive to DNA repair.
We thank Remus T. Dame, Nico P. Dantuma, Maartje C. Brink, Roel van Driel, and Jiri Lukas for the critical reading of the manuscript and Patrick Lane for artwork.
This work was supported by the Danish National Research Foundation (C.D.) and The Netherlands Organization for Scientific Research, NWO (grant 2007/09198/ALW/825.07.042 to M.S.L.).
Christoffel Dinant studied biology at the University of Amsterdam and received his Ph.D. from Erasmus University Rotterdam, where he defended his thesis, A Microscopic Study of the DNA Damage Response, under the supervision of J. H. J. Hoeijmakers, W. Vermeulen, and A. B. Houtsmuller. Currently, he is a postdoctoral fellow in the group of Jiri Lukas in the Centre for Genotoxic Stress Research, Copenhagen, Denmark. His research centers on the chromatin response to DNA damage and he has a particular interest in protein dynamics in the living cell nucleus.
Martijn Luijsterburg is a molecular biologist with a strong interest in the functioning of multiprotein complexes on the chromatin fiber in living cells. In particular, he is interested in the assembly and functioning of DNA repair complexes and mechanisms that render chromatin permissive to DNA repair. He received his M.S. in molecular biology from the Vrije Universiteit in Amsterdam and his Ph.D. in cell biology from the University of Amsterdam, Amsterdam, The Netherlands. At the Karolinska Institutet in Stockholm, Sweden, he currently studies mechanisms of DNA damage-induced signaling through ubiquitylation of histones.
Published ahead of print on 5 October 2009.