Regulation of ionizing radiation-induced Rad52 nuclear foci formation by c-Abl-mediated phosphorylation

We show tyrosine associates with Rad52 on tyrosine the very same site of Rad52 is phosphorylated upon exposure of cells to ionizing radiation (IR). Functional significance of c-Abl-dependent phosphorylation of Rad52 is underscored our findings that cells expressing the phosphorylation-resistant Rad52 mutant, where the was replaced by in response to IR. IR-induced Rad52 nuclear foci formation is markedly suppressed by the expression of dominant negative c-Abl. Together, our data support a mode of post-translational regulation of Rad52 mediated by the c-Abl tyrosine kinase.

Structural homologues of the RAD52 group of genes have been cloned from vertebrates (2). Since these proteins share similar biochemical properties with their yeast homologs (3), and since vertebrate cells deficient in the Rad54 (4,5) or Rad51 paralog genes (6,7) are sensitive to IR and show mitotic recombination defect phenotypes, basic molecular machinery of HR is believed to be conserved through evolution. However, mutant phenotypes of RAD52 epistasis group genes in vertebrate are not always identical to those in yeast.
The most prominent examples are Rad51 and Rad52. Rad51 gene knockout mice are embryonic lethal (8,9) and depletion of Rad51 proteins from cells is accompanied by an accumulation of cytologically detectable chromosome aberrations and subsequent cell death, indicating the essential role of Rad51 in the maintenance of chromosomal DNA during normal cell cycle (10). On the contrary, Rad52 gene knockout mice are normal and fertile and Rad52 deficient cells only show slight decrease in gene targeting frequency and don't exhibit any IR sensitive phenotype (11,12). However, it was recently shown that simultaneous depletion of Rad52 and Xrcc3 proteins renders chicken DT40 cells non-viable and provokes extensive chromosomal breaks (13) Plasmids. Vectors expressing kinase active or inactive c-Abl have been reported previously (21). cDNAs of human Rad51, DFF45 and wild type and mutant Rad52 were subcloned into pcDNA3-Flag and/or pEGFP-C1 vectors. The series of tyrosine to phenylalanine point mutants of human Rad52 were generated by two-step PCR using a set of 30-nucleotide primers carrying the mutated nucleotide. Proteins were visualized with an enhanced chemiluminescence detection system (Perkin Elmer Life Sciences).
In vitro kinase assay. Recombinant kinase active c-Abl was prepared from baculovirus-infected insect cells as described previously (23). Immuno-purified  Given the findings that Rad52 binds to Rad51 (18,24), with which c-Abl also associates (17,18), we were interested in determining whether there was any interaction between c-Abl and Rad52 using IP-Western analysis. A plasmid encoding Flag-c-Abl was coexpressed with a GFP-Rad52 expressing vector in 293T cells for testing of this possibility. Cell lysates were prepared 24 h posttransfection for immunoprecipitation with an anti-Flag antibody. Anti-GFP immunoblotting analysis of the immunocomplexes revealed an apparent association of Rad52 with the c-Abl protein (Fig. 1A, lane 3). Because of its Cterminal nuclear export sequence (NES), the c-Abl protein actively shuttles between the nucleus and the cytoplasm and consequently distributes to both compartments (25). The Rad52 protein is, however, almost exclusively nuclear localized (26). It was therefore of interest to examine where these two proteins co-localize in the cell. c-Abl C-terminal deletion mutant that lacks the NES and is mainly nuclear-localized (21) was tested for its ability to bind to Rad52. The result indicated that this c-Abl lacking the NES exhibited a markedly increased association with the Rad52 protein when compared with that of wild type c-Abl

c-Abl phosphorylates Rad52
The in vivo association of c-Abl with Rad52 prompted us to ask whether the Rad52 protein was phosphorylated as a result of binding to the c-Abl tyrosine kinase. To accomplish this, we coexpressed kinase active (KA) or kinase dead (KD) c-Abl (21) with Flag-tagged Rad52 in 293T cells. Flag-Rad51 and Flag-DFF45 were again included as a positive and negative controls, respectively.
We then substituted each of the 8-tyrosine residues with phenylalanine in order to identify the c-Abl phosphorylation site. When coexpressed with the kinaseactive c-Abl, only Flag-Rad52(Y104F) exhibited nearly completely diminished reactivity towards anti-P-Tyr, whereas the other mutants remained phosphorylated to the extent comparable to that of wild type Rad52 (Fig. 2C, lane supporting our hypothesis that Rad52 phosphorylation induced by IR is a c-Ablmediated event.

Binding of Rad52 to c-Abl is augmented by phosphorylation
Given the fact that the SH2 domain displays a high affinity towards to the phosphorylated tyrosine residue, we were interested in knowing whether Rad52 tyrosine phosphorylation affected its association with c-Abl, which contains a

The IR-induced GFP-Rad52 nuclear foci formation is regulated by c-Abl
A well-characterized feature of cellular homologous recombinational repair of DSBs is the formation of nuclear foci that contain the DSBs repair proteins (26- 28). It has been shown that stably expressed GFP-Rad52 can mimic the endogenous Rad52 protein in forming nuclear foci in response to IR (26,29).
Having demonstrated that IR-induced Rad52 phosphorylation was mediated by c-Abl, we were interested in determining whether tyrosine phosphorylation of Rad52 could affect IR-induced nuclear foci formation. To address this issue, we once again utilized stable Rad52 expressing CHO cell lines that we had previously established. Consistent with the published results, exposure of these cells to IR was associated with the induction of the Rad52 nuclear foci formation. Rad52 proteins form homo-heptameric ring structure in vitro (32). Even though tyrosine 104 is located within the homo-oligomerization domain of Rad52 (33), no apparent effect of phosphorylation on Rad52 homo-complex formation was detected (not shown). Rad52 tyrosine phosphorylation, however, markedly augments its association with the c-Abl protein, which has been shown to interact with ATM. Further studies are necessary to investigate whether this c-Ablmediated phosphorylation can influence the interaction between Rad52 and other proteins that contribute to the IR-induced nuclear foci formation.
It has been shown that Rad52-mediated HR repair is tightly associated with the S phase during which an identical chromosome, which is needed as a template for the repair, becomes available (34). Interestingly, DNA damageinduced c-Abl activation also occurs during S phase (35), suggesting a temporal correlation between these two events. Additionally, ATM, which is partially responsible for IR-induced c-Abl activation (19,20), contributes to the regulation of HR repair (15). It will be of interest to examine whether the ATM-dependent regulation of HR is mediated through c-Abl. Recently, constitutive association between c-Abl and Brca1 was reported (36). This association was disrupted following exposure to IR in an ATM-dependent manner. c-Abl kinase activity seems to be repressed by this association, since BRCA1-mutated cells exhibit constitutive high c-Abl kinase activity that cannot be further increased upon exposure to IR (36). Brca1 has also been shown to promote HR-mediated DNA repair (37). The defect in HR repair in BRCA1-mutated cells may be partially attributed to the deregulated c-Abl kinase and tyrosine hyper-phosphorylation of Rad52. Further study will be required to address this possibility.
In summary, our data provide compelling evidence to support a functional role for c-Abl in the post-translational regulation of IR-induced Rad52 nuclear foci formation.