Response to RAG-Mediated V(D)J Cleavage by NBS1 and γ-H2AX

See allHide authors and affiliations

Science  08 Dec 2000:
Vol. 290, Issue 5498, pp. 1962-1964
DOI: 10.1126/science.290.5498.1962


Genetic disorders affecting cellular responses to DNA damage are characterized by high rates of translocations involving antigen receptor loci and increased susceptibility to lymphoid malignancies. We report that the Nijmegen breakage syndrome protein (NBS1) and histone γ-H2AX, which associate with irradiation-induced DNA double-strand breaks (DSBs), are also found at sites of V(D)J (variable, diversity, joining) recombination–induced DSBs. In developing thymocytes, NBS1 and γ-H2AX form nuclear foci that colocalize with the T cell receptor α locus in response to recombination activating gene (RAG) protein–mediated V(D)J cleavage. Our results suggest that surveillance of T cell receptor recombination intermediates by NBS1 and γ-H2AX may be important for preventing oncogenic translocations.

V(D)J recombination is initiated by lymphoid-specific recombination activating gene 1 (RAG1) and RAG2 proteins, which introduce DSBs precisely between immunoglobulin and T cell receptor (TCR) coding gene segments and flanking recombination signal sequences. RAG-mediated cleavage generates four broken-end intermediates: two blunt signal ends and two covalently closed coding hairpin ends (1). The subsequent resolution of V(D)J ends into coding and signal joints requires ubiquitously expressed factors that function in general DSB repair (2,3). Although V(D)J recombination generates DNA damage, it has been presumed that broken DNA intermediates, which associate with RAG proteins within a postcleavage synaptic complex (4, 5), are sequestered from the DNA damage surveillance machinery. Primary DNA damage sensors include histone H2AX, which becomes rapidly phosphorylated (γ-H2AX) in response to external damage (6, 7), and the MRE11/RAD50/NBS1 complex, which forms ionizing irradiation-induced foci at DSBs (8,9). Although γ-H2AX and MRE11/RAD50/NBS1 appear to play an important role in monitoring chromosome integrity, the physiological conditions that activate these DNA damage surveillance/signaling factors have not been described.

To determine whether NBS1 and γ-H2AX are present at antigen receptor gene–specific breaks introduced during V(D)J recombination, we examined wild-type thymocytes by immunofluorescence analysis (10). We found that approximately 20% of freshly isolated thymocytes showed intense NBS1 and γ-H2AX foci (Fig. 1, A and B). Dual immunostaining revealed that H2AX was generally phosphorylated in the same nuclear domains where NBS1 foci were found (Fig. 1, I through K). The majority of thymocytes with intense NBS1/γ-H2AX staining contained one distinct spot, although cells were occasionally found to contain two, and less frequently, three or more foci. In contrast, multiple foci were distributed throughout the nucleus of irradiated thymocytes (Fig. 1, G and H). NBS1/γ-H2AX foci were present in immature CD4+CD8+ double-positive (DP) thymocytes (Fig. 1, C and D), whereas mature CD4+ single-positive (SP) populations exhibited a diffuse nuclear staining pattern for both proteins (Fig. 1, E and F) (11). Because NBS1/γ-H2AX foci are specific to DSBs (7, 9), these results indicate that mature T cells contain intact genomes, whereas approximately 20% of immature DP thymocytes have at least one DSB at any given time that is recognizable by NBS1/γ-H2AX.

Figure 1

NBS1 and γ-H2AX form foci in immature thymocytes. (A through F) Freshly isolated wild-type thymocytes were stained for NBS1 [(A), (C), (E); green] or γ-H2AX [(B), (D), (F); green], and counterstained with Topro-3 (red). Whole, double positive (DP), and single positive (SP) thymocytes were optically sectioned at 0.5-μm intervals and recombined into a maximum projection. Cells were imaged under identical conditions. (G and H) Bcl-2 transgenic thymocytes (which are resistant to irradiation-induced apoptosis) were exposed to 5 Gy, incubated at 37°C for 24 hours before fixation, stained for NBS1 [(G), green] or γ-H2AX [(H), green], and counterstained with Topro-3 (red). (I through K) Double staining with NBS1 [(I), green], γ-H2AX [(J), red], and colocalization [(K), yellow] in freshly isolated wild-type thymocytes. Scale bar: 10 μm [(A) through (F)]; 3 μm [(G), (H), and (K)].

In contrast to mature lymphocyte populations, which have completed V(D)J recombination, a fraction of immature DP thymocytes is actively undergoing rearrangement of TCRα genes. Recombination at the TCRα locus ceases during positive selection when successful TCR–major histocompatibility complex (MHC) ligand interactions down-regulate RAG1 and RAG2 gene expression (12, 13). To determine whether NBS1 foci formation requires V(D)J recombination in DP cells, we analyzed thymocytes from mice expressing a productively rearranged TCR transgene, termed AND, which is positively selected by MHC class II I-Ab(14).AND+MHC −/− thymocytes, which do not encounter their intrathymic selecting ligands, cannot be positively selected and therefore continue to rearrange their endogenous TCRα genes. NBS1 foci were detected in 20% ofAND+MHC −/− thymocytes, similar to the percentage observed in wild-type thymocytes (Fig. 2A). InAND+MHC+ thymocytes, where intrathymic TCR ligation terminates RAG1/RAG2 expression, NBS1 foci were detectable, but at a lower frequency (approximately 5%) than in the nonselecting MHC −/−background (Fig. 2B). By contrast, inAND+RAG2 −/− thymocytes, NBS1 foci were undetectable (Fig. 2C), consistent with the absence of V(D)J rearrangement in these mice. Thus, foci formation in immature thymocytes requires RAG cleavage and is suppressed by intrathymic TCR-ligand interactions that terminate RAG expression.

Figure 2

Foci formation in thymocytes requires V(D)J recombination. Distribution of NBS1 is compared in TCR transgenic thymocytes from AND+MHC−/− (A), AND+MHC+ (B), andAND+RAG2 −/− (C) mice. Data were obtained as in Fig. 1. Scale bar, 10 μm.

To determine whether NBS1/γ-H2AX foci were present at sites of active V(D)J recombination, we used an immunofluorescence in situ hybridization approach that allowed simultaneous visualization of DNA and protein within intact thymocytes (15, 16). The TCRα locus was identified using a combination of Vα and Cα probes, and NBS1 and γ-H2AX were detected by immunofluorescence. Coincidence of signals with one or both TCRα alleles was detected in 76% of the cells that contained NBS1 foci (N > 100), and in 77% of the cells that contained γ-H2AX foci (N > 100) (Fig. 3). In the remaining 23 to 24% of cells, approximately half exhibited only one TCRα hybridizing signal, leaving the possibility that NBS1 or γ-H2AX were localized to the undetected TCRα locus. In the remaining half (approximately 2.5% of thymocytes), NBS1 (or γ-H2AX) did not show colocalization with either detectable TCRα locus. Although these foci might be associated with other recombinationally active loci (e.g., TCRβ or TCRγ), the vast majority of NBS1 and γ-H2AX foci were coincident with the TCRα locus.

Figure 3

Colocalization of NBS1 and γ-H2AX foci with the TCRα gene in freshly isolated thymocytes. Freshly isolated wild-type thymocytes were stained with anti–γ-H2AX or anti-NBS1 antibodies [(IF) red] followed by FISH (green) detection of the TCRα locus with probes specific for the variable and constant regions. The cells were visualized by differential interference contrast (DIC) microscopy and the three images merged to determine colocalization (yellow). Fluorescence images represent a single optical section.

The observation that NBS1 and γ-H2AX are highly enriched in overlapping nuclear foci in immature thymocytes near RAG-induced DSBs was unexpected, because specific breaks introduced during V(D)J recombination do not appear to activate ATM- or p53-dependent DNA damage checkpoints, nor are these DNA damage sensors required for normal V(D)J recombination (17–19). Furthermore, V(D)J recombination coding-end intermediates have an extremely short half-life and are rare in wild-type mice (20–23). By contrast, 20% of thymocytes have broken signal ends at the TCRα locus (24). Thus, there are two possible, though not mutually exclusive, explanations for foci formation. First, NBS1/γ-H2AX foci may form within the postcleavage synaptic complex (4, 5) and then remain associated with unresolved signal ends, which are stable for several hours before being joined (5, 20). Alternatively, NBS1/γ-H2AX foci may persist on resolved coding junctions as residual footprints of already completed V(D)J recombination. In either case, RAG-induced NBS1/γ-H2AX foci provide a direct visualization of the ongoing process of antigen receptor rearrangement, and constitute the first evidence that DNA damage–sensing pathways are activated during normal V(D)J recombination. Further understanding of how NBS1 and γ-H2AX coordinate repair, signaling, and surveillance functions in response to intrinsic DNA damage during V(D)J recombination will provide insight into the mechanisms underlying chromosomal translocations and malignant transformation in chromosomal breakage disorders.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: andre_nussenzweig{at}


View Abstract

Navigate This Article