|Home | About | Journals | Submit | Contact Us | Français|
The DNA damage response (DDR) and DNA repair are critical for maintaining genomic stability and evading many human diseases [1, 2]. Recent findings indicate accumulation of SUN1, a nuclear envelope (NE) protein, is a significant pathogenic event in Emery-Dreifuss muscular dystrophy and Hutchinson-Gilford progeria syndrome, both caused by mutations in LMNA [3, 4]. However, roles of mammalian SUN proteins in mitotic cell division and genomic stability are unknown. Here we report that the inner NE proteins SUN1 and SUN2 may play a redundant role in DDR. Mouse embryonic fibroblasts from Sun1−/−Sun2−/− mice displayed premature proliferation arrest in S phase of cell cycle, increased apoptosis and DNA damage, and decreased perinuclear heterochromatin, indicating genome instability. Furthermore, activation of ATM and H2A.X, early events in DDR, were impaired in Sun1−/−Sun2−/− fibroblasts. A biochemical screen identified interactions between SUN1/2 and DNA-dependent protein kinase (DNAPK) complex that functions in DNA nonhomologous end joining repair and possibly in DDR [2, 5, 6]. Knockdown of DNAPK reduced ATM activation in NIH3T3 cells, consistent with a potential role of SUN1/2-DNAPK interaction during DDR. SUN1/2 could affect DDR by localizing certain nuclear factors to the NE or by mediating the communication between nuclear and cytoplasmic events.
SUN proteins are inner nuclear membrane proteins with their N-terminal region localized in the nucleoplasm and their C-terminal SUN domain in the lumen of the nuclear envelope (NE) [7–9]. We have previously used mouse genetics to analyze the physiological functions of SUN1 and SUN2 and found that Sun1−/−Sun2−/− mice died shortly after birth [10, 11]. Although the neonatal death phenotype was partly rescued by expressing SUN1 in the nervous system, the surviving mice still displayed multiple defects including growth retardation [10, 12], prompting us to examine the function of SUN1 and SUN2 in mitotic cell division and genomic stability in MEFs.
The MEFs were isolated from embryos at embryonic day 14.5 (E14.5). MEFs from the Sun1−/−Sun2−/−, but not Sun1−/− or Sun2−/− mice, proliferated significantly slower than wild-type MEFs after passage 5 (Figure S1 and and1A).1A). Cell-cycle analysis on unsynchronized cells from passage 6 via flow cytometry showed that the G0/G1-phase fraction was only slightly increased and the S-phase fraction was slightly reduced in Sun1−/−Sun2−/− MEFs (Figure 1B). In contrast, using bromodeoxyuridine (BrdU) to label the replicative DNA in S-phase cells, we observed that the percentage of proliferative S-phase cells in Sun1−/−Sun2−/− MEFs was less than half that of wild-type MEFs (Figure 1C), suggesting an S-phase arrest in Sun1−/−Sun2−/− MEFs. Furthermore, there were an increased number of annexin V-positive cells in Sun1−/−Sun2−/− MEFs at passage 6 (Figure 1D), indicating an increase in apoptosis. These results raised the possibility that DNA damage accumulated more rapidly in Sun1−/−Sun2−/− MEFs.
To detect the potential genomic instability in Sun1−/−Sun2−/− MEFs, we carried out single cell electrophoresis to observe the level of DNA damage. In the absence of methyl methanesulfonate (MMS), which induces DNA damage , there was no significant difference in the tail moment between wild type and Sun1−/−Sun2−/− MEFs. After treatment with MMS, we observed a significant increase in the number of Sun1−/−Sun2−/− MEFs with prominent comet tails, indicative of DNA fragmentation (Figure 1E). In addition, using transmission electronic microscopy (TEM), we found that the perinuclear heterochromatin was decreased in Sun1−/−Sun2−/− MEFs (Figure 1F). These results suggested that SUN1 and SUN2 have roles in maintaining genomic stability, possibly by affecting DDR and/or DNA repair.
Phosphorylation of ataxia telangiectasia mutated protein (ATM) and H2A.X, a histone H2A variant, at Ser139 (i.e., γ-H2A.X) are among the earliest events to occur in response to DNA damage [2, 14, 15]. These early DDR events lead to activation of DNA repair factors and cell-cycle checkpoints, ensuring the proper repair of sites of DNA damage [2, 15, 16]. We obtained three pieces of data to indicate that the early events in DDR are affected in Sun1−/−Sun2−/− MEFs. First, the expression level of γ-H2A.X was significantly reduced in Sun1−/−Sun2−/− MEFs (Figure 2A). In addition, the level of phosphorylated Chk1, a cell-cycle checkpoint factor downstream of the DDR pathway, was also reduced (Figure 2A). Second, while ATM was seen to be activated by 0.1 μM of hydroxyurea (HU) in wild-type MEFs, it was not activated by HU in Sun1−/−Sun2−/− MEFs (Figure 2B). Third, we found that the cell-division cycle of Sun1−/−Sun2−/− MEFs was not blocked at the G2/M phase following treatment with 200 ng/μl of mitomycin C (MMC) (Figure 2C), indicating that the mutant cells failed to properly respond to DNA damage. However, due to the lack of a suitable antibody for mouse non-phosphorylated ATM, we could not exclude the possibility that the observed decrease of phosphorylated ATM was partly due to apoptosis induced ATM degradation . However, such an effect of apoptosis is unlikely to be significant because our analysis using Annexin V indicated that apoptosis was not dramatically increased in SUN1/2 DKO MEFs (only 2.2% compared to 1.3% in wild type; Figure 1D). To confirm the defect of Sun1−/−Sun2−/− MEFs in DDR, we examined the sensitivity of Sun1−/−Sun2−/− MEFs to various DNA damaging agents. Although Sun1−/−Sun2−/− MEFs exhibited no significant abnormality in their response to γ-irradiation, they exhibited increased sensitivity to MMS and MMC (Figure S2). These results suggest that SUN1 and SUN2 (SUN1/2) have a prominent role in DDR to specific types of DNA damage.
To search for the mechanism of SUN1/2 function in DDR, we screened for SUN1 interacting proteins by applying tandem affinity purification and MALDI-MS/MS proteomic analysis . The effectiveness of this approach was indicated by the identification of, among only 27 candidate proteins (Table S1), three KASH domain proteins (Syne-1/Nesprin-1, Syne-2/Nesprin-2, and Nesprin-3) that have all been well characterized as biochemical and functional partners of SUN1/2 [10, 11, 19–21]. In addition, several cytoskeleton proteins and emerin were also identified (Table S1), and this was likely due to their interactions with the Syne/Nesprin proteins .
DNAPKcs, the catalytic subunit of the DNAPK holoenzyme, which also includes Ku70 and Ku80 as the regulatory subunits , was one of candidate SUN1-associated proteins. The DNAPK holoenzyme has been studied extensively for its role in the nonhomologous end-joining repair pathway [2, 5, 24]. Recently, DNAPKcs was found to interact with the Hutchinson-Gilford progeria syndrome (HGPS) mutant version of Lamin A/C, linking its function to HGPS-related DNA instability and cell aging . Although DNAPKcs has been reported to have a role in the phosphorylation of H2A.X in experiments using DNAPKcs−/− MEFs , its function in the early steps of DDR and the potential mechanism of such a role are not clear.
Based on co-immunoprecipitation (co-IP) and Western blot analysis, we confirmed that DNAPKcs was associated with both SUN1 and SUN2 (Figure 3A and 3B). Similar experiments showed that both Ku70 and Ku80 also interacted with SUN1 and SUN2 (Figure 3C–3F). We further examined the localization of these proteins by immunofluorescence staining of tagged proteins expressed from transformed plasmids. Consistent with previous studies, DNAPKcs, Ku70, and Ku80 were localized uniformly in the nucleus (Figure 4, S4A–G [25, 26]). Dual-staining analysis with SUN1/2 and these components of the DNAPK complex indicated that these proteins colocalized at a low level along the inner side of the NE (Figure 4A and S4A–G). However, we did not observe an increase in this colocalization after HU treatment (Figure 4B). We further compared the localization of endogenous Ku70 in wild-type and Sun1−/−Sun2−/− MEFs, but also did not observe a significant difference under standard culturing conditions with or without the HU treatment (Figure S4H–K). Since we cannot make a conclusion about the function of the interaction between SUN1/2 and DNAPKc, and their co-localization in DDR, the mechanism by which the DNAPK complex interacts with SUN1/2 remains to be understood.
Given the well-known function of the DNAPK complex in DNA repair, the interaction between SUN1/2 and the DNAPK complex may suggest that SUN1/2 has a function downstream of DDR, especially in DNA repair, which is consistent with the suggestion that Lamin A/C has a role in DNA repair. However, the data presented above indicated a role for SUN1/2 in an early step of DDR. Using shRNA to knockdown the DNAPKcs mRNA level in NIH3T3 cells (Figure 3G), we observed a reduction of ATM and H2A.X phosphorylation when the cells were treated with HU (Figure 3H), suggesting that the interaction between SUN1/2 and the DNAPK complex is potentially involved in mediating the role of SUN1/2 in DDR. However, we cannot exclude the possibility that the reduction of ATM and H2AX phosphorylation in this experiment is solely caused by knocking down DNAPKcs. Due to the lack of an appropriate antibody against the mouse phosphorylated DNAPKcs, we could not examine whether SUN1/2 play roles in activating the DNAPKcs in DDR.
The mammalian SUN1 and SUN2 proteins have been studied for their roles in nuclear migration and anchorage as well as in anchoring meiotic telomeres to the NE during animal development [10–12, 27, 28]. In this study, we showed that these two inner NE proteins also have a significant function in DDR. Like their roles in anchoring myonuclei and neuronal migration, SUN1 and SUN2 functions in DDR are likely redundant; only MEFs from double knockout mice display obvious defects. We can speculate on a potential model for their function based on our limited observations and the available information.
The identification of the interaction between SUN1/2 and the DNAPK complex provides an important mechanistic clue. Because DNAPK is better known for its function in DNA repair, we can consider two different hypotheses regarding the function of this interaction. One hypothesis is that SUN1/2 may interact with DNAPK for their function in DNA repair, and a defect in this function was masked by the defect in the earlier DDR events in Sun1−/−Sun2−/− MEFs. SUN1/2 function in DDR would thus be mediated by factors that are yet to be determined. An alternative hypothesis is that the DNAPK complex also has a significant role in DDR and its interaction with SUN1/2 is critical for such a function. This hypothesis is consistent with a previous report that DNAPKcs has a role in H2A.X phosphorylation  and our result that shRNA knockdown of DNAPKcs compromised ATM and H2A.X activation in NIH3T3 cells (Figure 3H). It is conceivable that SUN1/2 functions in DDR by localizing DNA damage sites or certain DDR factors to the NE. In a yeast study, Ku70, a regulatory subunit of DNAPK, was shown to recognize and recruit the site of DNA damage to the NE in an Mps3-dependent manner , but it is not clear whether this NE localization is for DDR or DNA repair. In this study, we observed the colocalization between DNAPK components and SUN1/2 in mammalian cells (Figure 4 and S4A–G), indicating a similar function to their yeast counterparts. However, we did not observe an increase of this colocalization after HU treatment (Figure 4 and S4H–K)). In addition, the localization of endogenous Ku70 was not changed in Sun1−/−Sun2−/− MEFs (Figure S4H–K). Therefore, it is possible that the SUN1/2 interaction with the DNAPK complex is a constitutive cellular event required for proper DDR and the interaction is not required for just the NE localization of DNAPK.
SUN proteins are known to form the NE complex with outer NE KASH- domain proteins that interact with cytoplasmic factors [10–12, 19, 20, 22]. Therefore, an alternative model for the role of SUN1/2 roles in DDR could be that they mediate the communication between nuclear and cytoplasmic events. In our search for SUN1/2 interacting factors, we also identified the Ca2+-binding protein reticulocalcin-2 (Rcn2), which has been suggested to be localized in the lumen of the endoplasmic reticulum (ER)  and has been shown to have a role in activating ERK1/2 in a recent report . Interestingly, Sun1−/−Sun2−/− MEFs displayed impaired ERK activation after HU and MMC treatment (Figure S3E–F). Our analysis using co-IP and immunostaining confirmed the interaction between SUN1/2 and Rcn2 and indicated that they co-localized on to the NE (Figure S3G–K). However, when NIH3T3 cells were treated with shRNA against Rcn2, we failed to identify any effect on ATM activation or subsequent after the treatment to induce DDR induction (Figure S3L–N). Though this negative result is not sufficient to exclude a role of for Rcn2 in DDR due to potential redundant functions, the physiological role of the interaction between SUN1/2 and Rcn2 is currently unclear.
Lamin A/C are part of the nuclear lamina located inside the nuclear inner membrane, and their functions have been linked to many important cellular events [32–34]. Both SUN1 and SUN2 have been shown to interact with Lamin A/C [20, 35, 36], and the HGPS-associated Lamin A/C mutations have been shown to impair the interaction between Lamin A/C and SUN1/2 . These data raise a possibility that SUN1/2 may function in DDR through this interaction with Lamin A/C. However, several studies on Lamin A/C contradict such a model. For example, the HGPS mutant version of Lamin A/C (termed progerin), but not wild type, was found to interact with DNAPKcs in a recent study , even though the mutant Lamin A/C cannot bind to SUN1/2 . Furthermore, unlike Sun1−/−Sun2−/−, Lamin A/C mutations were found to cause an increase in γ–H2A.X levels, which were attributed to defective DNA repair [38, 39]. Therefore, the role of the SUN1/2 interaction with Lamin A/C in DDR is still unclear. Chen et al. recently reported that accumulation of SUN1 is a pathogenic event in Emerry-Dreifuss musclular dystrophy and Hutchinson-Gilford progeria, which are caused by mutations in LMNA . Eliminating or reducing SUN1 was found to significantly relieve some pathological phenotypes characterized in mouse models of these diseases. Our results may provide valuable insight into the potential mechanism underlying these observations. We show SUN1 and SUN2 act redundantly to promote DDR, while LMNA mutations were shown to cause potential increases in DDR . Therefore, it is logical to propose that some of the disease phenotypes are caused by hyperactivity in DDR as the result of abnormally high level of SUN1. Mutating SUN1 is expected to only reduce the level of DDR, but the reduction may be sufficient to neutralize the effect of the LMNA mutations. Further studies are needed to uncover the molecular mechanism by which SUN1/2 affect DDR.
We prepared MEFs from E14.5 embryos and cultured them in DMEM (Invitrogen) supplemented with 10% FBS, 1% L-glutamine, and 1% penicillin-streptomycin (Invitrogen). For the continuous passage assay, we plated MEFs at a density of 3×105 cells in a 6-cm plate. We then counted and replated the cell number every 3 days. The BrdU incorporation assay was carried out according to a standard protocol . Briefly, 5×104 MEFs were plated in each well of a 6-well plate. After incubation for 24 hours, they were treated with 10 mg/ml BrdU (Sigma) for 4 hours. The cells were then harvested and stained with a FITC-conjugated anti-BrdU antibody (Caltag) and propidium iodide (PI) (Sigma) or 7-amino-actinomycin D (7AAD). The cell-cycle distribution was analyzed using a FACSCalibur flow cytometer (BD Biosciences) and CellQuest (BD Biosciences) and FlowJo (Tree Star) software.
Data were calculated using an unpaired two-tailed Student t-test and presented as means ± S.E.M.
We thank C. Xu, J. Yao, B. Tan, X. Huang, and the EM facility at Fudan Medical School for assistance and contributions to this study; Y. Jin, Q. Lei, and D. Chen for providing materials; Y. Xiong, K. Guan, T. Su, A. K. Sewell, B. Yin, X. Wu, W. Tao, K. Deng, L. Sun, and members of IDM for valuable comments and discussions. This work was supported by an outstanding graduate student researcher award from the Ministry of Education of China to KL, grants from National Natural Science Foundation of China (No.30871233), National Basic Research Program of China (973-2006CB806700), and the National Hi-Tech Research and Development Program of China (863-2007AA022101). TX and MH are HHMI investigators.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.