NEDD8 is an important regulatory factor in many biological processes. However, the substrates for neddylation, and the relationship between the ubiquitin and NEDD8 pathways remain largely unknown. Here, we show that NEDD8 is covalently conjugated to histone 2A (H2A), and that neddylation of H2A antagonizes its ubiquitylation. NEDD8 suppresses ubiquitylation of H2A, and a decreased level of free NEDD8 promotes H2A ubiquitylation. Furthermore, we found that the E3 ligase RNF168 promotes both H2A ubiquitylation and neddylation. Interestingly, RNF168 is itself a substrate for NEDD8, and neddylation of RNF168 is necessary for its E3 ubiquitin activity. Inhibition of RNF168 neddylation impairs the interaction between RNF168 and its E2 enzyme Ubc13 (also known as UBE2N). Moreover, in response to DNA damage, the level of H2A neddylation decreased with an increase in the ubiquitylation of H2A, which facilitates DNA damage repair. During the later stages of damage repair, H2A neddylation increased gradually, whereas ubiquitylation decreased to basal levels. Mechanistically, NEDD8 negatively regulates the DNA damage repair process through suppression of the ubiquitylation of H2A and γH2AX, which further blocks the recruitment of the damage response protein BRCA1. Our findings elucidate the relationship of H2A ubiquitylation and neddylation, and suggest a novel modulatory approach to DNA damage repair through the neddylation pathway.
Neural precursor cell expressed developmentally downregulated 8 (NEDD8) was first discovered as a gene that was downregulated during mouse brain development (Kamitani et al., 1997). As a member of the ubiquitin like protein (UBL) family, NEDD8 has the greatest amino acid sequence identity to ubiquitin but a distinct conjugation pathway (Kerscher et al., 2006). Genetic studies have demonstrated the essential roles of NEDD8 in cell viability and developmental processes in organisms ranging from yeast to mouse (Hochstrasser, 1998; Jones and Candido, 2000; Lammer et al., 1998; Liakopoulos et al., 1998; Osaka et al., 2000; Ou et al., 2002; Tateishi et al., 2001). Investigations into molecular mechanisms indicate a widespread function of NEDD8 in regulating ubiquitin E3 ligase activity, transcription factor activity and protein stability. Neddylation of cullins, scaffold proteins of the Skp–Cullin–F-box (SCF) ubiquitin ligase complex, promotes recruitment of an E2 conjugating enzyme and results in increased ubiquitylation activity and the degradation of substrates (Kawakami et al., 2001; Pan et al., 2004; Wimuttisuk and Singer, 2007). MDM2-mediated neddylation of p53 suppresses its transcriptional activity, and NF-κB activity is controlled by neddylation at several stages of the pathway (Dohmesen et al., 2008; Noguchi et al., 2011; Singh et al., 2007; Xirodimas et al., 2004). Neddylation is also reported to protect ribosomal proteins from destabilization (Xirodimas et al., 2008; Zhang et al., 2012).
The relationship between the ubiquitin and NEDD8 pathways remains an intricate question. Most substrates can be both ubiquitylated and neddylated, and these two modifications usually occur at the same time. Some E3 ligases and deconjugating enzymes have dual specificity for both ubiquitylation and neddylation, suggesting that these two modifications function in a similar manner. For example, MDM2 is an E3 ligase that mediates conjugation of both ubiquitin and NEDD8 to p53, and both modifications suppress the transcriptional activity of p53 (Xirodimas et al., 2004). Neddylation antagonizes ubiquitylation and blocks substrates from ubiquitin-directed degradation. c-Cbl promotes neddylation of the TGF-β type II receptor and suppresses its K48-linked ubiquitylation and degradation (Zuo et al., 2013).
Ubiquitylation is an early and important signal in the DNA damage repair pathway. Histone 2A (H2A), and its variant H2AX, are modified by K63-linked polyubiquitin chains via multiple E3 ligases, including RING2 (also known as RNF2), RNF8 and RNF168 (Fang et al., 2004; Gatti et al., 2012; Huen et al., 2007; Mailand et al., 2007; Mattiroli et al., 2012; Oestergaard et al., 2012; Pinato et al., 2009; Plans et al., 2008), which facilitates the recruitment of DNA damage repair proteins, such as RAP80 and BRCA1. SUMO and NEDD8 are also involved in the DNA damage response (Guzzo et al., 2012; Ma et al., 2013). H4 is strongly neddylated at the N-terminal tail by the E3 ligase RNF111 (Ma et al., 2013). Here, we report that neddylation of H2A is mediated by RNF168, which competes with ubiquitylation of H2A. NEDD8 suppresses ubiquitylation of H2A and H2AX, and blocks the recruitment of BRCA1 at DNA damage repair sites. Moreover, RNF168 is also a target of NEDD8, and downregulation of RNF168 neddylation by the deneddylating enzyme NEDP1, or mutants of the NEDD8-conjugating enzyme UBC12, also inhibits H2A and H2AX ubiquitylation.
The histones H2A, H2B and H4 are substrates for neddylation
Several groups have tried to identify new substrates for neddylation using proteomics approaches (Jones et al., 2008; Li et al., 2006; Norman and Shiekhattar, 2006; Xirodimas et al., 2008). We analyzed potential neddylation substrates that had been identified by different laboratories and found that histones appeared with a high abundance and frequency. Immunostaining analysis indicated that, similar to histones, NEDD8 mainly existed in the nucleus (supplementary material Fig. S1A). We, therefore, investigated whether histones can be neddylated. Both co-immunoprecipitation and His-tag pulldown assays showed that H2A was clearly neddylated (Fig. 1A). It is reported that conjugation of overexpressed NEDD8 relies on the ubiquitin E1 enzyme Ube1 (also known as UBA1) but not the NEDD8-activating E1 enzyme (Hjerpe et al., 2012). We next used MLN4924, a specific neddylation E1 inhibitor, to examine whether neddylation of H2A could be inhibited. Indeed, His-tag pulldown analysis revealed that MLN4924 suppressed H2A neddylation (Fig. 1B, left). To further confirm the neddylation of H2A, we detected endogenous H2A and its variant γH2AX by His–NEDD8 pulldown assay, the conjugation of ubiquitin to H2A and γH2AX was also detected as a positive control. We found that both endogenous H2A and γH2AX were neddylated and that treatment with MLN4924 blocked the conjugation of NEDD8 (Fig. 1B, right). Interestingly, ubiquitylation of γH2AX was also reduced after treatment with MLN4924 (Fig. 1B). These data indicate that H2A is neddylated.
Next, we examined the endogenous modification of H2B, H3 and H4. H2B and H4 were both ubiquitylated and neddylated, whereas H3 was not, and the deneddylase NEDP1 (also known as SENP8) deconjugated NEDD8 from the histone substrates (Fig. 1C). These data indicate that H2A, H3 and H4 are targets of NEDD8 on the chromatin. Recently, the neddylation of H4 in response to DNA damage has been reported (Ma et al., 2013), and owing to the important function of H2A modification in DNA damage response, we thus focused our work on H2A to elucidate the mechanism of H2A neddylation and its function in VP16-induced DNA damage, and to also understand the correlation with H2A ubiquitylation.
RNF168 is a neddylation E3 ligase for H2A, and NEDP1 is a deneddylase
To identify the E3 ligase required for the neddylation of H2A, we examined the reported H2A ubiquitin E3 ligases – including RNF8, RNF168 and RING2 – as many E3 ligases are reported to have dual activity. The results showed that RNF168 strongly promoted the conjugation of NEDD8 to H2A. RING2 exhibited a weak activity, whereas RNF8 suppressed the conjugation (Fig. 2A). A NEDD8-ΔGG mutant that lacks the last two amino acids and, subsequently, cannot be conjugated to lysines was used as a negative control. To further confirm the role of RNF168 in H2A neddylation, we next assessed the variation in H2A neddylation when RNF168 was knocked down. As expected, ablation of RNF168 by using RNA interference (RNAi) clearly weakened the endogenous neddylation of H2A (Fig. 2B). These data suggest that RNF168 functions as an E3 ligase to mediate the neddylation of H2A.
NEDP1 is a specific deneddylating enzyme, we thus tried to address whether NEDP1 is a deneddylase for H2A. Indeed, NEDP1 deconjugated NEDD8 from H2A (Fig. 2C). When the catalytic cysteine at position 163 was mutated to serine in NEDP1, the deneddylation activity was impaired and the level of H2A neddylation was increased (Fig. 2D). Because ubiquitin specific protease 7 (USP7) has been reported to suppress ubiquitylation of H2A (de Bie et al., 2010; Maertens et al., 2010), we also examined the effect of USP7 on H2A neddylation. Interestingly, exogenous expression of USP7 weakened NEDD8 conjugation, whereas NEDP1 only weakly affected ubiquitin conjugations (Fig. 2C). These results indicate that NEDP1 is a deneddylase for H2A and that USP7 might also function as a deneddylase for H2A.
H2A is modified competitively by NEDD8 and ubiquitin
To examine the sites on H2A to which NEDD8 is conjugated, we constructed a series of mutants in which Lys119 and/or Lys120, two reported ubiquitin conjugation sites, were mutated to arginine, and the levels of neddylation and ubiquitylation were investigated. The results showed that H2A ubiquitylation was mostly abolished by the K119R/K120R double mutant (supplementary material Fig. S2A), whereas H2A neddylation increased in both the K119R and K120R single mutants (supplementary material Fig. S2B). It is probable that the K119R/K120R mutant could not completely abolish H2A ubiquitylation because of the conjugation of ubiquitin at the K13 and K15 residues (Mattiroli et al., 2012). The mutation constructs, in which either all the five lysines at the C-terminal tail were mutated or the C-terminal tail was truncated, could not completely abolish H2A neddylation (supplementary material Fig. S2B). These results indicate that H2A is neddylated at multiple sites. The increased neddylation of H2A when K119 or K120 was mutated suggests that an antagonistic relationship exists between ubiquitylation and neddylation.
We next investigated whether NEDD8 does, indeed, influence H2A ubiquitylation. As shown in Fig. 3A, in resting cells, H2A was ubiquitylated by endogenous ubiquitin, which is in accordance with previous reports that have noted that 10–15% of H2A is ubiquitylated (Goldknopf et al., 1975; Hunt and Dayhoff, 1977). When ubiquitin or NEDD8 was co-expressed with H2A, an extra band above endogenously ubiquitylated H2A was detected, which represents H2A that has been modified with the tagged ubiquitin or NEDD8, and exogenously expressed ubiquitin or NEDD8 suppressed endogenous H2A ubiquitylation (Fig. 3A). This was similar to the observation that endogenous H2A ubiquitylation was inhibited in cells that overexpressed NEDD8 (Fig. 2A). Consistently, as an E3 ligase for both H2A ubiquitylation and neddylation, RNF168 promoted H2A neddylation and suppressed H2A ubiquitylation in cells that overexpressed NEDD8 (Fig. 2A). This is because of the altered ratio of free NEDD8 to ubiquitin, resulting from the ectopic expression of NEDD8. To further confirm this finding, we co-transfected H2A together with His- or HA-tagged ubiquitin and NEDD8 and analyzed the effects of these two small molecules on H2A modification. As expected, overexpression of NEDD8 suppressed H2A ubiquitylation, and, consistently, the ectopic expression of ubiquitin also suppressed H2A neddylation (Fig. 3B).
To explore this finding under physiological conditions, we constructed two UBC12 (also known as UBE2M) mutants, C111S and C111A. Both mutants are known to lack neddylation E2 enzyme activity; however, UBC12-C111S can bind to NEDD8 and reduce the ratio of free NEDD8 to ubiquitin, whereas UBC12-C111A does not (Wada et al., 2000). These two mutants dominantly suppressed, but could not completely abolish, the NEDD8 conjugation pathway owing to the existence of endogenous UBC12 proteins. As shown in Fig. 3C, the inhibition of the NEDD8 pathway – by treatment with MLN4924 or the expression of NEDP1 or the UBC12 mutants – suppressed H2A ubiquitylation. Of the two UBC12 mutants, UBC12-C111A had a greater inhibitory effect, whereas UBC12-C111S showed a weaker effect. Expression of UBC12-C111S reduced the amount of endogenous free NEDD8 and, thus, resulted in increased H2A ubiquitylation (Fig. 3C). These data demonstrate the competitive relationship between NEDD8 and ubiquitin.
Neddylation modulates DNA damage repair
Ubiquitylation of H2A and γH2AX is an important event in DNA damage repair. We were interested in whether neddylation plays a role in this process. We treated the ubiquitin- or NEDD8-transfected cells with VP16, a drug that can induce DNA double- and single-strand breaks, and examined the ubiquitylation of H2A and γH2AX, and the neddylation of H2A under different conditions. The results showed that treatment with VP16 led to increased H2A ubiquitylation but decreased H2A neddylation (Fig. 4A). γH2AX ubiquitylation was greatly increased in response to treatment with VP16, whereas the overexpression of NEDD8 suppressed this increase. We further examined the effect of treatment with VP16 on the ubiquitylation and neddylation of endogenous H2A and obtained a result that was consistent with our previous data (Fig. 4B).
The aforementioned findings indicate that NEDD8 is conjugated to H2A in resting cells and is removed during the early stages of DNA damage. By contrast, the level of H2A ubiquitylation is increased, which is believed to facilitate the recruitment of DNA damage repair proteins. At later stages of DNA damage, induced by treatment with VP16, the neddylation of H2A increased gradually, whereas ubiquitylation of H2A returned to the basal level (Fig. 4C). Therefore, it is reasonable to conclude that the DNA damage repair process is negatively regulated by H2A neddylation to terminate the repair process.
In response to DNA damage, H2A and γH2AX are polyubiquitylated, which recruits downstream DNA damage response factors. Next, we investigated the effect of ectopically expressed NEDD8 or the UBC12 mutants C111S and C111A on VP16-induced γH2AX ubiquitylation (Fig. 4D). The ubiquitin pulldown assay showed that overexpression of NEDD8 inhibited γH2AX ubiquitylation (Fig. 4D, upper panel), which resulted from the suppressive effect of NEDD8 on ubiquitin conjugation, as described in Fig. 3C. This was consistent with the level of endogenous ubiquitylated γH2AX that we detected by using western blot analysis (Fig. 4D, lower panel). UBC12-C111A exhibited a stronger inhibitory effect than UBC12-C111S, which also correlates with the competitive relationship between ubiquitin and NEDD8. Similar results were obtained in HCT116 cells and HeLa cells (supplementary material Figs S1C, S2C). We also tested the effect of neddylation on the subcellular localization of γH2AX and BRCA1 in cells that had been treated with VP16. As expected, γH2AX and BRCA1 were barely detected in control cells (supplementary material Fig. S1B). The accumulation of γH2AX and BRCA1 at DNA damage repair sites decreased in cells that overexpressed NEDD8 or the UBC12 mutants C111S and C111A (Fig. 4E), which is consistent with the results shown in Fig. 3C. Our experiments show that neddylation antagonizes the ubiquitylation of H2A and γH2AX and that inhibition of neddylation, by using UBC12 mutants, also decreases ubiquitylation of these histones, strongly suggesting that neddylation regulates DNA damage repair.
RNF168 is both neddylated and ubiquitylated and neddylation of RNF168 regulates its E3 ligase activity
The observation that inhibition of the neddylation pathway (through the treatment of cells with MLN4924, or the overexpression of NEDP1 or the UBC12 mutants) prevents H2A/γH2AX ubiquitylation seems to contradict the competitive relationship between ubiquitin and NEDD8. This is, possibly, due to the inhibition of the neddylation of other proteins that are involved in the DNA damage response, such as H2A E3 ligases. To investigate this hypothesis, we examined whether the E3 ligases of H2A – including RING2, RNF8 and RNF168 – are neddylated. The results showed that all of these E3 ligases were both ubiquitylated and neddylated, and that, of these proteins, there was a greater amount of the modified RNF168 protein (Fig. 5A).
It has been reported that neddylation regulates E3 ligase activity (Kawakami et al., 2001; Osaka et al., 2000; Pan et al., 2004); therefore, we examined the influence of the neddylation of RNF168 on the ubiquitylation of H2A. We co-transfected HEK293T cells with RNF168 and NEDP1 and found that NEDP1 substantially abolished the conjugation of NEDD8 to RNF168 and impaired its ubiquitin E3 ligase activity (Fig. 5B). Additionally, we assessed the RNF168 ubiquitin E3 ligase activity in response to DNA damage. RNF168 strongly promoted H2A and γH2AX ubiquitylation upon the treatment of cells with VP16, whereas NEDP1 inhibited RNF168-mediated ubiquitylation of H2A and γH2AX (Fig. 5C). The expression of the UBC12-C111A and UBC12-C111S mutants also suppressed RNF168 E3 ligase activity (Fig. 5D). Taken together, these data confirm that neddylation of RNF168 regulates its ubiquitin E3 ligase activity.
DNA damage repair modulates RNF168 neddylation
To further elucidate the function of RNF168 ubiquitin and/or NEDD8 conjugation in DNA damage repair, we investigated the changes in the ubiquitylation and neddylation of endogenous RNF168 in response to DNA damage. His-tag pulldown analysis showed that the treatment of cells with VP16 resulted in an upregulation of the ubiquitylation, but a reduction in the neddylation, of endogenous RNF168 (Fig. 6A).
Next, we examined the effect of neddylation on the interaction between RNF168 and E2 Ubc13 (Fig. 6B). The interaction of RNF168 with Ubc13 increased within 5 min but decreased in 30 min upon treatment with VP16. We also found that RNF168 interacted with NEDP1, and NEDP1 suppressed the interaction between RNF168 and Ubc13 in response to DNA damage, which strongly suggested that neddylation affects the interaction between RNF168 and the E2 ubiquitin-conjugating enzyme.
Furthermore, we generated a series of RNF168 mutation constructs (supplementary material Fig. S3B). Lysines at positions 112, 126, 140, 227, 232, 441 and 445 were mutated to arginine, which were selected based on the prediction of sites for ubiquitin conjugation (UbPred) and analysis of the conservation of the amino acid sequence (ESPript). We found that all mutations had a similar effect on the conjugation of ubiquitin and NEDD8 to RNF168 in comparison to that of the wild-type protein (supplementary material Fig. S3B). These results suggest that RNF168 is probably modified on multiple lysines by NEDD8, which is consistent with other substrates that are neddylated. Although these lysine residues can now be excluded, the ubiquitylation site of RNF168 still remains to be determined. Combined with the results presented in Fig. 5, we demonstrate that neddylation plays an important role in the regulation of RNF168 activity.
NEDD8 plays important regulatory functions in cell growth, viability and development. Here, we studied the function and mechanism of H2A neddylation. By investigating the competitive relationship between the ubiquitylation and neddylation of H2A, we propose a negative regulatory mechanism of DNA damage repair that is mediated by the neddylation of H2A. As the model shown in Fig. 7, H2A is both ubiquitylated and neddylated in resting cells. Upon DNA damage, H2A is predominantly ubiquitylated, and neddylation of the protein is reduced in order to facilitate the subsequent repair process. At the later stages of DNA damage repair, neddylation of H2A increases and ubiquitylation decreases to the basal level. Additionally, we identified RNF168 as an E3 ligase that mediates neddylation of H2A, that RNF168 was also strongly neddylated and that this NEDD8 conjugation was essential for the interaction of RNF168 with the E2 enzyme Ubc13.
H2A ubiquitylation is a key event in the DNA damage repair process as many DNA repair proteins, through their ubiquitin interacting motif (UIM), are recruited to sites of DNA damage by binding to the ubiquitin chains of H2A and γH2AX (Gatti et al., 2012; Pinato et al., 2009). Here, we demonstrated that H2A neddylation antagonizes ubiquitylation and suppresses the recruitment of the DNA damage repair protein BRCA1. We found that NEDD8 was conjugated to H2A in resting cells and was removed when damage occurred, which is likely to facilitate DNA damage repair. Along with DNA damage repair, the level of H2A neddylation increased, whereas that of H2A ubiquitylation decreased (Fig. 4C); therefore, it is reasonable to conclude that the DNA damage repair process is negatively regulated by H2A neddylation.
Under conditions of both ectopic and endogenous protein expression, we demonstrated the competitive relationship between ubiquitin and NEDD8. Investigation of the site on H2A that is required for neddylation revealed that mutation, or truncation, of lysine residues in the C-terminus of H2A did not completely abolish H2A neddylation, suggesting that NEDD8 can be conjugated to multiple sites on H2A. These data indicate that NEDD8 is conjugated to H2A on arbitrary lysines and that NEDD8 conjugation inhibits ubiquitylation of H2A. Because the C-terminal tail of H2A contains only ∼10 amino acids, whereas mature ubiquitin and NEDD8 have 76 amino acids, it is possible that the conjugation of NEDD8 has a steric effect and spatially blocks the conjugation of ubiquitin to H2A.
Interestingly, in addition to functioning as an E3 ligase for H2A ubiquitylation, we demonstrated that RNF168 is an E3 ligase that mediates H2A neddylation. Similar to our observation, MDM2 acts as an E3 ligase to mediate both p53 ubiquitylation and neddylation, and these two modifications regulate p53 activity coordinately (Abida et al., 2007; Carter and Vousden, 2008; Harper, 2004). We examined the substrate selection of RNF168 upon treatment with VP16 and found that RNF168 strongly promotes H2A ubiquitylation (Fig. 5C) but not neddylation (supplementary material Fig. S3A). This observation indicates that, at early stages of the response to DNA damage, RNF168 preferentially conjugates ubiquitin to H2A, which is in agreement with the fact that the ubiquitylation of H2A is essential for the recruitment of other DNA damage repair proteins at lesion sites (Huen et al., 2007; Mailand et al., 2007; Mattiroli et al., 2012; Pinato et al., 2009). Furthermore, we found that RNF168 is itself a substrate of neddylation, which is consistent with the finding that the UIM of RNF168 directly binds to NEDD8 (Ma et al., 2013). More importantly, our data here indicate that neddylation of RNF168 is essential for its ubiquitin E3 ligase activity, and that RNF168 neddylation regulates H2A ubiquitylation and further affects DNA damage repair. The studies on the neddylation of other E3 ligases, such as cullin and MDM2 (Kawakami et al., 2001; Pan et al., 2004; Xirodimas et al., 2004), suggest that neddylation promotes the interaction between the E3 and E2 enzymes, and stabilizes the E3 ligase. Here, we demonstrated that RNF168 interacts with NEDP1, and inhibition by NEDP1 of the neddylation of RNF168 suppresses the interaction between RNF168 and the E2 enzyme Ubc13, which further reduces the E3 ligase activity of RNF168 (Fig. 6). RNF168-catalyzed ubiquitylation of H2A and γH2AX is a very early signal, which occurs within 10 min, in the DNA damage response (Doil et al., 2009; Mailand et al., 2007). This is supported by our observation that the interaction between RNF168 and Ubc13 increased within 5 min following treatment with VP16 (Fig. 6). Upon termination of the ubiquitylation of H2A and γH2AX, the E3 ligase activity of RNF168 gradually decreased, reflected by the result that the RNF168–Ubc13 interaction decreased 30 min after treatment with VP16 (Fig. 6B), and this process was regulated by neddylation of RNF168. At the early stages of the response to DNA damage, a certain amount of RNF168 neddylation assures the E3 ligase activity of RNF168, whereas, following H2A and γH2AX ubiquitylation, neddylation of RNF168 decreased 30 min after treatment with VP16 (Fig. 6A). The reduced neddylation of RNF168 further impaired the interaction between RNF168 and its E2 enzyme Ubc13. We also noted that the decreased neddylation of RNF168 was accompanied by the increased ubiquitylation of RNF168, 30 min after VP16 treatment (Fig. 6A). Further studies are necessary in order to elucidate whether the neddylation of RNF168 also antagonizes the ubiquitylation of the protein.
The role of neddylation in the DNA damage repair network is complicated. The E3 ligase RNF111 has been reported to catalyze neddylation at DNA damage sites, and deprivation of RNF111 impairs the DNA repair response. Neddylation also occurred on lesion sites that could be recognized by the UIMs of RNF168 (Ma et al., 2013). Proteomic analyses indicate that a large number of DNA damage response proteins are involved in neddylation (Liao et al., 2011). Cullins are neddylation substrates that have been well studied (Chiba and Tanaka, 2004; Duda et al., 2008; Huang et al., 2009). Cul4A and Cul4B participate in UV-induced DNA damage repair (Guerrero-Santoro et al., 2008; Kapetanaki et al., 2006; Wang et al., 2006) but are not reported in VP16-induced DNA damage repair. We thus examined the effect of Cul4A and Cul4B on H2A ubiquitylation in response to treatment with VP16. Surprisingly, Cul4B had little effect, whereas Cul4A strongly suppressed H2A ubiquitylation (supplementary material Fig. S4). This unexpected result indicates a role for Cul4A in VP16-induced DNA damage repair which will require further investigation.
Here, we confirm the role of protein neddylation in the DNA damage response. On the one hand, neddylation of H2A antagonizes its ubiquitylation and suppresses the DNA damage response; on the other hand, the neddylation of RNF168 promotes its E3 ubiquitin ligase activity. Downregulation of the neddylation of RNF168 impairs its E3 ligase activity and inhibits ubiquitylation of H2A and γH2AX. Taken together, our findings suggest a newly identified regulatory mechanism for DNA damage repair through the neddylation pathway.
MATERIALS AND METHODS
Human H2A; the H2A mutants H2A-K119R, H2A-K120R, H2A-K119R/K120 (H2A-2KR), K119R/K120R/K126R (H2A-3KR), K119R/K120R/K126R/K128R (H2A-4KR), K119R/K120R/K126R/K128R/K130R (H2A-5KR); and H2A-1-118 mutants were cloned into the pCDNA vector with three Flag-tag repeats (3Flag). The cDNA encoding human ubiquitin and NEDD8 were cloned into pEF1-C-6His and pRK7-HA vectors. NEDP1, UBC12-C111S and UBC12-C111A were cloned into the pCDNA-3Flag vector.
Antibodies and reagents
Antibodies were purchased from the following companies: mouse monoclonal antibodies against Flag (F3165) or HA (H9658) were from Sigma-Aldrich, mouse monoclonal antibodies against the His tag (D291-3) and a rabbit monoclonal against actin (PM053) were from Medical and Biological Laboratories (Nagoya, Japan), a rabbit monoclonal antibody against H2A (39209) was from Active Motif, rabbit monoclonal antibodies against γH2AX (BS4760), H2B (BS1657), H3 (BS1174), H4 (BS1662) and BRCA1 (BS1036) were from Bioworlde, a rabbit monoclonal antibody against RNF168 (TA306771) was from Origene. MLN4924 was a gift from Millennium Pharmaceuticals. VP16 (E1383) was from Sigma-Aldrich.
Cell culture and transfection
HEK293T and HeLa cells were cultured in Iscove's Modified Dulbecco's Medium (that had been supplemented with 10% fetal bovine serum) for 24 h at 37°C under 5% CO2 and then transfected with the indicated plasmids using MegaTran 1.0 transfection reagent (Origene) according to the manufacturer's instructions.
His-tag pulldown assay
The His-tag pulldown assay was performed to examine the modification of histones as follows. Cells were cultured in 10-cm dishes and transfected with the indicated plasmids. At 48 h after transfection, cells were harvested and lysed in 6 ml of lysis buffer (6 M guanidinium-HCl, 0.1 M Na2HPO4 NaH2PO4, 0.01 M Tris-HCl, pH 8.0, 5 mM imidazole and 10 mM β-mercaptoethanol) for 30 min, 2% of the cells, lysed by using NP-40 lysis buffer, were used as inputs. The lysate was incubated with 70 µl of Ni2+ beads for 4 h and then washed four times with 1 ml buffer I (6 M guanidinium-HCl, 0.1 M Na2HPO4 NaH2PO4, 0.01 M Tris-HCl, pH 8.0 and 10 mM β-mercaptoethanol), buffer II (8 M urea, 0.1 M Na2HPO4 NaH2PO4, 0.01 M Tris-HCl, pH 8.0 and 10 mM β-mercaptoethanol), buffer III (8 M urea, 0.1 M Na2HPO4 NaH2PO4, 0.01 M Tris-HCl, pH 6.3 and 10 mM β-mercaptoethanol plus 0.2% Triton X-100) and buffer IV (8 M urea, 0.1 M Na2HPO4 NaH2PO4, 0.01 M Tris-HCl, pH 6.3, 10 mM β-mercaptoethanol plus 0.1% Triton X-100). 30 µl of elution buffer was used to elute modified proteins. The samples, with an equal volume of 2×SDS loading buffer, were analyzed by SDS-PAGE and western blot analysis.
Cells were cultured in 10-cm dishes and transfected with the indicated plasmids. At 48 h later, cells were washed with 10 ml of pre-chilled PBS and lysed for 60 min in 1 ml NP-40 lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1% NP-40, pH 7.4) with 5 µl of protease inhibitor cocktail. The lysates were then pre-cleared with 40 µl of protein G. Usually, 5 µg of antibody or IgG was used to bind to the bait proteins for 5 h, or overnight at 4°C, and then incubated with 80 µl protein G for a further 3 h. Finally, the protein G was washed with 1 ml of NP-40 lysis buffer three times and then heated for 10 min at 96°C with 30 µl 2×SDS loading buffer. Samples were analyzed by SDS-PAGE and western blotting.
HeLa cells were cultured on glass coverslips in 6-well plates. 24 h later, cells were washed three times with pre-chilled PBS and then fixed with 1 ml of pre-chilled methanol for 10 min at −20°C. After washing with pre-chilled PBS three times, the cells were blocked with 1% bovine serum albumin (BSA) for 1 h, followed by incubation with primary antibodies that had been diluted in 1% BSA for 1 h at 37°C. After washing with pre-chilled PBS three times, the cells were incubated for 1 h at 37°C with secondary antibodies that had been diluted in 1% BSA and then washed with pre-chilled PBS three times. Finally, 20 µl of mounting solution was used to mount cells. All of the images were captured by using a LSM 710 NLO and DuoScan microscope (CarlZeiss, Germany).
The quantified data were analyzed by using one-way analysis of variance (ANOVA). All statistical procedures were performed using SPSS software (version 20). Data are represented as means±s.e.m. The P value was determined by Student's t test and a value of P<0.05 was considered statistically significant.
We thank Jiri Lukas and Claudia Lukas (Institute of Cancer Biology and Centre for Genotoxic Stress Research, Denmark) for the gift of plasmids of RNF8 and RNF168. We sincerely thank Yongqun Zhu (Zhejiang University, China) for providing the Ubc13 plasmid and Qianzheng Zhu (The Ohio State University, OH) for the Cul4A construct. We thank Millennium Pharmaceuticals for providing MLN4924.
T.T.L., J.H.G. and X.F.Z. designed the experiments, interpreted data and prepared the manuscript. T.T.L. and J.H.G. performed most of the experiments. Z.J.H. and X.H. performed some experiments.
This work was supported by grants from the National Science Foundation of China [grant numbers 30930020 and 31170709]; the National High Technology and Development Program of China 973 programs [grant number 2010CB911800]; Doctoral Fund of Ministry of Education of China [grant number 20130001130003]; and the International Centre for Genetic Engineering and Biotechnology [grant number CRP/CHN09-01].
The authors declare no competing interests.