RAG2 unleashed: lymphocytes beware

Abstract

The programmed degradation of the RAG2 enzyme upon entry to S phase restricts V(D)J recombination to the G0–G1 phase of the cell cycle. In this issue of Immunity Zhang et al. (2011) show that this is critical to prevent lymphoma formation.

Adaptive immunity relies on an enormously diverse repertoire of antigen receptors. This diversity of recognition molecules allows immune responses with exquisitely high specificity against a plethora of invading pathogen species. This repertoire is generated by DNA recombination, and hence relies on processes that each and every cell is programmed to avoid. This protection is central to survival as DNA recombination when gone awry can result in translocations that activate oncogenes or inactivate tumor suppressors. As lymphocyte development involves bursts of cellular proliferation, such chromosomal aberrations in individual progenitors are readily passed on to large numbers of daughter cells. Compared to post-mitotic cells, this feature dramatically increases the chances of individual lymphocytes to acquire additional deleterious mutations towards becoming a tumor cell. Hence adaptive immunity needs to maintain the balance between the benefits and dangers of DNA recombination.

The primary immunoglobulin (Ig) and T cell receptor (TCR) repertoires are generated by V(D)J recombination, the assembly of functional Ig and TCR genes from individual variable (V), diversity (D), and joining (J) gene segments. This process occurs in two steps: cleavage and joining. During the cleavage phase, the RAG complex formed by the products of recombination activating genes 1 and 2 (Rag1 and Rag2) cuts the DNA synchronously at the ends of two gene segments. In the subsequent joining phase the RAG complex collaborates with the non-homologous end-joining (NHEJ) machinery to facilitate the ligation of the DNA ends in a productive manner (Fig. 1). Mishaps that ultimately lead to lymphomas can occur at both these steps by illegitimate cleavage or illegitimate joining. A good example for the first category is the erroneous RAG-mediated cutting of a unique DNA sequence in the BCL2 major break point region that bears no resemblance to the cognate substrate of the RAG complex (Raghavan et al., 2004), and subsequent joining of the BCL2 gene to the IGH locus resulting in t(14;18) translocations. A clear example for the second category is the illegitimate joining of the DNA ends in the absence of NHEJ. XRCC4 is an essential cofactor for the ligation of DNA ends in this repair pathway, and Xrcc4^−/−Trpp53^−/− mice develop pro-B cell lymphomas that are characterized by translocations linking the Igh to the amplified Myc gene locus (Gao et al., 2000).

Figure 1.

Schematic representation of V(D)J recombination in the context of the cell cycle. Gene segments are shown as filled boxes, and recombination signals sequences (RSS) as triangles. For the joining phase, only gene segments are shown, as the excised RSS-flanked DNA fragments are omitted for clarity. RAG1 protein expression is constant throughout the cell cycle, whereas RAG2 is degraded at the onset of S phase. The T490A mutation (red line) keeps RAG2 stable even in S and G2-M. RAG cleavage products are largely repaired by non-homologous end joining (NHEJ) during G01–G1 phase, with alternative end-joining (A-EJ) serving as a back-up pathway. At the entry to S phase, degradation of RAG2 facilitates the completion of the joining phase of V(D)J recombination, and subsequently no new DNA cleavage is initiated. In contrast, RAG2 T490A does not affect V(D)J recombination in G0–G1, but it now allows for DNA cleavage in S phase. Although some end-joining might still occur exclusively by canonical NHEJ activities, DNA repair at this point largely involves homologous recombination (HR) repair factors. The latter scenario can lead to translocations, but the survival of respective cells is suppressed by p53.

The report by Zhang et al. now adds a worst case scenario to this theme, the combination of both categories: illegitimate cleavage of the correct genes segments but at the wrong time, combined with illegitimate joining by the wrong DNA repair pathway, homologous recombination (HR). Threonine 490 in RAG2 is a target of the cyclin A-Cdk2 kinase and is phosphorylated at the G1 to S transition. This marks RAG2 for poly-ubiquitination by the Skp2-SCF enzyme complex, and ultimately leads to its destruction by the proteasome (Desiderio, 2010) (Fig. 1). In the RAG2 T490A knock-in mouse strain reported in this issue, RAG2 is not subjected to cell-cycle dependent degradation (Zhang et al., 2011). As expected, both RAG1 and RAG2 remain present throughout the cell cycle—not only in G0–G1 but also in S and G2-M phase. Somewhat surprisingly, lymphocyte development in these Rag2^T490A/T490A mice appeared largely normal, but in depth analysis revealed an increase in the number of aberrant V(D)J recombination joints. Importantly, Skp2^−/− mice that are also unable to degrade RAG2 at the G1-S transition showed a very similar phenotype, suggesting that it is caused by the incorrect timing and not by the T490A mutation per se. Strikingly, the aberrant joints lacked microhomologies that are a hallmark of a back-up DNA repair mechanism frequently referred to as “alternative end-joining” (A-EJ). This alternative repair pathway kicks in when one or more of the canonical NHEJ factors are inactive, and becomes activated when RAG1 and RAG2 fail to collaborate with the NHEJ pathway (Corneo et al., 2007). In contrast, the T490A mutation probably presents the cell with DNA breaks at a point in the cell cycle at which HR takes the lead over NHEJ (Branzei and Foiani, 2008). During S-G2 phase, sister chromatids are readily available to serve as templates for the faithful repair of DNA breaks by HR. The random selection of gene segments during V(D)J recombination precludes the existence of templates to guide HR, and this readily explains the aberrant joints in Rag2^T490A/T490A mice. One question that remains open is whether the lack of RAG2 degradation at the G1-S transition in the T490A mutant simply results in spilling over of unrepaired DNA breaks into S phase, or whether bona-fide RAG cleavage continues throughout the cell cycle. But in both scenarios the involvement of HR in S phase is the key event towards the aberrant joint formation.

Although the V(D)J recombination defects in Rag2^T490A/T490A mice seemed rather modest, the increase in apoptosis in their thymuses provided a clear hint towards something more significant. Apoptosis can be triggered by an accumulation of unresolved DNA lesions in a p53-dependent manner, and this might be just what was happening in these mice. Thus Zheng et al. crossed the Rag2^T490A/T490A mice onto the p53 (Trp53)-deficient background to unmask potentially unrepaired DNA damage. Trp53^−/− mice developed T lymphomas by 20 weeks of age, and unexpectedly, the Rag2^T490A/^T490A Trp53^−/− mice showed tumors with almost identical latency, as if the RAG2 mutation had no effect at all. The nature of these tumors, however, was quite distinct: they originated from both the B and the T cell lineage, and almost all of them exhibited clear evidence for clonal translocations involving the respective lineage-specific antigen receptor gene loci. This indicates that erroneous V(D)J recombination involving an illegitimate choice of DNA repair can promote lymphomagenesis. In summary, the RAG2 T490A mutation is the very first mutant RAG protein described thus far that allows for almost normal V(D)J recombination and hence lymphocyte development, but at the same time acts as an oncogene.

These observations immediately raise the question of whether such RAG2 mutations might also be present in human lymphomas. Similarly, as RAG2 destruction is mediated by Skp2, it is conceivable that mutations in Skp2 are associated with lymphomas, but no such link has been reported thus far. As Zhang et al. pointed out, the inhibitory effects of Skp2-deficiency on cell cycle progression might suppress formation of respective lymphoma. It remains worthwhile, however, to screen for both Skp2 and RAG2 mutations in human lymphomas as this might be a defining feature of a subgroup of these tumors.

The concept that the DNA break initiating enzyme is pivotal for directing subsequent DNA repair to avoid chromosomal aberration is probably also important during class switch recombination (CSR). This programmed DNA recombination process, like V(D)J recombination, relies largely on NHEJ but also on A-EJ (Stavnezer et al., 2008). Examples of CSR-associated translocations resulting from mis-directed formation of DNA breaks, and impaired DNA repair are abundant. As canonical CSR is likely completed prior to entry into S phase, it is conceivable that the uncontrolled generation of DNA breaks throughout S phase could invoke the undesired use of HR in this scenario as well. The ultimate outcome would again be translocations leading to lymphomas. Whether these intriguing possibilities are indeed correct and how CSR might be restricted to G0–G1 will hopefully be resolved by future studies.

Lastly, the importance of cell cycle regulated RAG activity has a direct implication on the current view of how the V(D)J recombinase and adaptive immunity evolved. A widely accepted model proposes that RAG1 and RAG2 originated from an ancient transposon (Flajnik and Kasahara, 2010). Such mobile DNA elements are sometimes referred to as “selfish DNA” as their only purposes is to maintain their own existence. The lifecycle of a transposon starts with its transcription and the synthesis of the encoded transposase. This enzyme then excises the very gene that harbors its genetic information from its current genomic location, and integrates it into a new chromosomal site. Importantly, transposases do not care about the DNA break they leave behind, as long as the cell is able to repair it to ensure its survival. Thus it is tempting to speculate that the original RAG transposon lacked the cell cycle regulation that is so essential for V(D)J recombination. This in turn suggests that the taming of the RAG transposase involved acquiring the ability to break DNA only when the joining of the resulting ends is channeled into the appropriate repair pathway, NHEJ. Hence the restriction of RAG activity to the G0–G1 phase of the cell cycle was in all likelihood a key step towards adaptive immunity in jawed vertebrates.