Mediator of DNA Damage Checkpoint Protein 1 Facilitates V(D)J Recombination in Cells Lacking DNA Repair Factor XLF

DNA double-strand breaks (DSBs) trigger the Ataxia telangiectasia mutated (ATM)-dependent DNA damage response (DDR), which consists of histone H2AX, MDC1, RNF168, 53BP1, PTIP, RIF1, Rev7, and Shieldin. Early stages of B and T lymphocyte development are dependent on recombination activating gene (RAG)-induced DSBs that form the basis for further V(D)J recombination. Non-homologous end joining (NHEJ) pathway factors recognize, process, and ligate DSBs. Based on numerous loss-of-function studies, DDR factors were thought to be dispensable for the V(D)J recombination. In particular, mice lacking Mediator of DNA Damage Checkpoint Protein 1 (MDC1) possessed nearly wild-type levels of mature B and T lymphocytes in the spleen, thymus, and bone marrow. NHEJ factor XRCC4-like factor (XLF)/Cernunnos is functionally redundant with ATM, histone H2AX, and p53-binding protein 1 (53BP1) during the lymphocyte development in mice. Here, we genetically inactivated MDC1, XLF, or both MDC1 and XLF in murine vAbl pro-B cell lines and, using chromosomally integrated substrates, demonstrated that MDC1 stimulates the V(D)J recombination in cells lacking XLF. Moreover, combined inactivation of MDC1 and XLF in mice resulted in synthetic lethality. Together, these findings suggest that MDC1 and XLF are functionally redundant during the mouse development, in general, and the V(D)J recombination, in particular.


Variable (V), Diversity (D) and Joining (J) Gene Segments Recombination (V(D)J Recombination) Assays
Based on Chromosomally Integrated pMX Cassettes V(D)J recombination assays were performed using chromosomally-integrated pMX inversion (pMX-INV) and pMX deletion (pMX-DEL) substrates, as previously described [34,42,43,49,50]. In the pMX-INV cassette, the green fluorescent protein (GFP) gene is placed in the reversed orientation and the GFP protein is not expressed. Upon the RAG-induced V(D)J recombination, the GFP gene is placed in the sense orientation leading to the GFP protein expression. The GFP protein is then detected by flow cytometry to estimate the V(D)J recombination efficiency in indicated vAbl cells [42,49,50]. For the Southern blot-based experiments, we used chromosomally-integrated pMX-DEL cassettes. During the V(D)J recombination, the pMX-DEL CJ cassette results in an intermediate product with hairpin-sealed coding ends that require Artemis nuclease activity to open the hairpins prior DNA ligase 4-dependent DNA ligation, leading to coding joints (CJ). On the contrary, the pMX-DEL SJ cassette results in the RAG-dependent generation of blunt signal ends (SE) that can be directly ligated by DNA ligase 4 and do not require Artemis nuclease activity, leading to signal joints (SJ) [34,42,43,49,50].
Both vAbl and HAP1 cells were counted every 24 h using a Countess™ Automated Cell Counter (Invitrogen, Carlsbad, CA, USA) with Trypan blue staining (Invitrogen, Carlsbad, CA, USA) and bright-field detection. Statistical analyses were performed using GraphPad Prism 8 (La Jolla, CA, USA), one-way analysis of variance (ANOVA), and t-test.

Robust V(D)J Recombination in Progenitor-B Cells Lacking Mediator of DNA Damage Checkpoint Protein 1 (MDC1)
Mice lacking MDC1 possess nearly wild-type levels of B and T lymphocytes [40]. Combined inactivation of MDC1 and Artemis suggests that MDC1 protects or stabilizes RAG-induced DSBs before ligation. In particular, the vAbl cells lacking MDC1 and Artemis possess ATM-dependent degradation of free DNA ends during the attempted V(D)J recombination [55]. To further determine the impact of MDC1 on the V(D)J recombination, we inter-crossed MDC1 +/− Eµ-Bcl2 + mice and isolated the cells from the bone marrow of three-week-old MDC1 −/− Eµ-Bcl2 + animals. We then established Abelson murine leukemia virus kinase-transformed pro-B cells (vAbl) and chromosomally-integrated either pMX-INV or pMX-DEL V(D)J recombination cassettes, as described previously [18,34,42,43,49]. Similar to wild type (WT) controls, two independently generated MDC1-deficient vAbl cell lines possessed robust coding-end (CE) and signal end (SE) joining (Supplementary Figure S1). We concluded that MDC1 is dispensable for the V(D)J recombination in WT vAbl progenitor B cells.  (Table 1). Strikingly, we detected no MDC1 −/− XLF −/− double knockout pups, and the final genotype distribution was 34:70:0 (1:2:0) ( Table 1). We concluded that combined inactivation of MDC1 and XLF results in embryonic lethality.   Figure 1C). Inactivation of the MDC1 gene in human HAP1 cells resulted in proliferation rates similar to WT cells at 24-72 h, and reduced proliferation rates at 96 and 120 h ( Figure 1D).

Reduced V(D)J Recombination Efficiency in vAbl Pro-B Cells Lacking both MDC1 and XLF
To determine the impact of MDC1 on V(D)J recombination, we chromosomally-integrated the cassette-carrying GFP gene in reverse orientation and flanked by DNA sequences recognized by RAG (pMX-INV) [49,50] (Figure 2A). To induce the RAG expression, we exposed the cells to the vAbl kinase inhibitor STI571 (Gleevec). Upon a successful V(D)J recombination event, the cells expressing GFP were detectable by flow cytometry [42,49,50]. The cells lacking MDC1 possessed relatively high levels of V(D)J recombination reflected by GFP expression (29%), which was in the range of WT and XLF −/− cell lines (34% and 37%, respectively) ( Figure 2B-D). Strikingly, combined inactivation of MDC1 and XLF resulted in a significantly reduced proportion of GFP-expressing vAbl cells when compared to WT and single knockout controls (average levels of 20%; ****, p < 0.0001). Double knockout DNA-PKcs −/− XLF −/− vAbl cells were used as a negative control to establish background levels of the experiments (0% of GFP-positive cells) [34]. We concluded that MDC1 is stimulating the V(D)J recombination in XLF-deficient cells, due to functional complementarity between MDC1 and XLF in this process.

Discussion
Inactivation of RAG and most of the known NHEJ factor genes in mice leads to immunodeficiency [12,56]. Recently, we and others found that single inactivation of XLF, PAXX, or Mri genes results in mice with the nearly normal immune system, due to the overlapping functions between XLF and PAXX [26][27][28][29]33], as well as XLF and Mri [12,13] (Table 2). The ATM-dependent DDR pathway was initially thought to be dispensable for the V(D)J recombination, although more recent studies using combined genetic inactivation of XLF and ATM [42], as well as DNA-PKcs and ATM [21,57], revealed that ATM is indeed involved in the early stages of B and T lymphocyte development and its function is partially compensated by XLF and DNA-PKcs. Later, we and others found that ATM substrates, H2AX and 53BP1, are also required for B and T lymphocyte development due to their functions in V(D)J recombination [42][43][44] (Table 2). Here, we show that another ATM substrate, MDC1, is involved in the V(D)J recombination and its function is compensated in WT cells by XLF. Combined inactivation of ATM and XLF, or 53BP1 and XLF, resulted in immunodeficient mice of smaller sizes than single knockouts or wild-type controls, with abrogated NHEJ, resembling Ku70 −/− or Ku80 −/− knockouts [1,[42][43][44]. Differently, combined inactivation of DNA-PKcs and XLF [34,35], H2AX and XLF [42], or MDC1 and XLF ( [33]; and this study) resulted in embryonic lethality in mice, challenging genetic interaction studies in vivo (Table 2). One option to overcome this obstacle is to develop conditional knockouts allowing inactivation of DNA-PKcs, XLF, or MDC1 in developing B and T lymphocytes in adult mice. An alternative option is to develop more complex mouse models using. for example, p53 −/− or p53 +/− backgrounds, allowing for the rescue of embryonic lethality (e.g., References [33,35]).
Knocking out genes of interest in cell lines may complement and sometimes substitute in vivo experiments using transgenic mice. In particular, vAbl cell lines can be modified using the CRISPR/Cas9 gene-editing approach and serve as a model system to elucidate the specific roles of a particular gene (e.g., References [30][31][32]50]). Moreover, human, nearly haploid HAP1 cells derived from the KMB-7 cell lines have been recently used to develop genetically-modified cells (e.g., References [13,33,51,54]).
It becomes more accepted that the DDR pathway contributes to the V(D)J recombination in developing B and T lymphocytes [1,34,[42][43][44]. However, the mechanistic aspects underlying the specific roles of the DDR factors in this process remain unclear. One can speculate that DDR factors share the functions with XLF, e.g., by stabilizing the DNA repair complex or supporting timely recruitment and dissociation of the NHEJ factors. The DDR pathway may also contribute to distinct but complementary XLF aspects of the DNA repair, e.g., by recruiting the downstream enzymes, supporting the DNA damage-induced post-translational modifications of DNA repair factors and histones, or protecting the free DNA ends from the nuclease-dependent processing before the DNA ligation step [1,34,[42][43][44]55]. In particular, the role of MDC1 during the V(D)J recombination might be to stabilize the DNA repair complex, to protect the free DNA ends, to ensure efficient recruitment of downstream DDR factors, such as 53BP1, PTIP, RIF1, Shieldin, etc. [1,[42][43][44]47,55,58], or to exit from the G1 phase of the cell cycle following the RAG-induced DSB [59]. Further research is required to identify specific roles of MDC1 and XLF in DNA repair. Table 2. Impact of NHEJ-deficiency on V(D)J recombination in mice.
The proliferation rate of vAbl cells lacking both XLF and MDC1 was reduced when compared to single-deficient and WT controls ( Figure 1) at 72 h. Moreover, proliferation rates of MDC1-deficient cells were also reduced when compared to WT, although not significant. Furthermore, the lack of MDC1 alone resulted in significantly reduced proliferation rates of human HAP1 cells at 96 and 120 h ( Figure 1). These observations may suggest that, first, the lack of MDC1 is compensated by the presence of XLF in murine cells, and second, that the MDC1 is required for efficient DNA repair and proliferation of human cells, likely by supporting the cell cycle progression and DNA damage tolerance [47,59].

Conclusions
Multiple DDR factors are involved in the V(D)J recombination. Due to the functional redundancy between the DDR and NHEJ pathways, complex genetic in vivo and in vitro models will be appropriate to uncover specific functions of DDR factors in B and T lymphocyte development and further elucidate mechanisms underlying their roles.

Conflicts of Interest:
The authors declare no conflict of interest.