The histone methyltransferase Setd2 is indispensable for V(D)J recombination

The diverse repertoire of T cell receptors (TCR) and immunoglobulins is generated through the somatic rearrangement of respective V, D and J gene segments, termed V(D)J recombination, during early T or B cell development. However, epigenetic regulation of V(D)J recombination is still not fully understood. Here we show that the deficiency of Setd2, a histone methyltransferase that catalyzes lysine 36 trimethylation on histone 3 (H3K36me3) in mice, causes a severe developmental block of thymocytes at the CD4−CD8− DN3 stage. While H3K36me3 is normally enriched at the TCRβ locus, Setd2 deficiency reduces TCRβ H3K36me3 and suppresses TCRβ V(D)J rearrangement by impairing RAG1 binding to TCRβ loci and the DNA double-strand break repair. Similarly, Setd2 ablation also impairs immunoglobulin V(D)J rearrangement to induce B cell development block at the pro-B stage. Lastly, SETD2 is frequently mutated in patients with primary immunodeficiency. Our study thus demonstrates that Setd2 is required for optimal V(D)J recombination and normal lymphocyte development.

blocked at the DN3 3 and the pro-B cell stages respectively. The authors propose that these developmental defects indicate a role for Setd2 in V(D)J recombination involving the TCRβ locus and the Igh locus. A potential mechanism is proposed in which H3K36me3 mediates the recruitment of Rag1. Other alterations of Setd2 deletion in mice are suggested, impaired DNA double-strand breaks repair system, that could play a role in this phenotype. They finally identified potential consequences of Setd2 polymorphisms in the development of human primary immunodeficiency. This is an interesting paper and of significant interest. A number of issues need to be addressed to further improve the manuscript.
Major concerns. 1. The use of a B-cell specific promoter (i.e.CD19-cre or Mb1-cre) driving CRE expression crossed with the Sedt2 fl/fl mice is needed to specifically address the role of Setd2 in B-cell development.

Reply:
We thank the reviewer very much for the precise summary of our manuscript and positive comments of our work. Following the reviewer's advice, we crossed the Cd19-Cre mice with the Setd2 f/f mice to specifically knockout Setd2 in B cell. As shown in supplementary Fig.6 of the revised manuscript, we found that specific ablation of Setd2 in the B cell lineage resulted in a developmental block at the pro-B stage, phenotypically consistent with the B cell defects in the Mx1-Cre + ;Setd2 f/f mouse model as shown by us previously. We also performed semi-quantitative PCR assays to examine the V(D)J rearrangement from sorted pro-B cells of Cd19-Cre + ;Setd2 f/f and control mice. As shown in supplementary Fig.6c, the V(D)J recombination was impaired in Setd2 null B cells.

The data supporting the specificity for effects of V(D)J recombination in B cells is incomplete (H3K36me3 and RAG1 ChIP in the Igh loci).
Reply: To address this concern, we have performed new ChIP assays of H3K36me3 and Rag1 at Igh loci in sorted B cells from control and Cd19-Cre + ;Setd2 f/f mice. We found that the deposition of H3K36me3 and Rag1 showed an overlapping pattern, similar to what we observed at TCRβ loci. ChIP experiments results have been incorporated to the revised manuscript as supplementary Fig.   7.
3. Fig 5. ChIP on H3K36me3 needs to be done for Setd2-/-cells to compare the background levels of this modification.
Reply: As suggested, we performed additional ChIP experiments on sorted Setd2 null B and T cells to determine the background levels of H3K36me3. We can barely detect enrichment of H3K36me3 either at IgH or TCRβ loci. The new data has been included to supplementary Fig.2a of the revise manuscript. 4. Is the recruitment of Rag1 by H3K36me3 specific? From the data in fig 6C, it seems that Rag1 recruitment is affected to various degrees or not at all.

Reply:
We greatly appreciated the reviewer's insightful comment. The V(D)J rearrangement is accurately regulated by multiple factors. In our manuscript, we found that Setd2 ablation results in removal of the H3K36me3 modification and decreased (but not complete ablation of) Rag1 recruitment to RSSs during V(D)J rearrangement. Therefore, we do not exclude other factors besides H3K36me3 in contributing to the recruitment of Rag1 to the RSSs. We have added a short discussion on this point in line 382 to 390 of the revised manuscript text.
5. Statistical analysis for correlation between H3K36me3 and Rag1 deposition needs to be done.

Reply:
Following the reviewer's advice, we have used the Pearson's Correlation Coefficient method to measure the correlation between H3K36me3 and Rag1 deposition at RSSs. We found that the correlation r value is 0.7898, and the p value is 0.010 between H3K36me3 and Rag1 deposition on TCRβ locus, which indicates a strong positive correlation.
6. Recruitment of Rag2 by H3K4me3 is well established. Control ChIP experiments are required to address if Rag1 binding changes in Setd2-/-mice are caused by a reduction in H3K4me3 levels.

Reply:
We thank the reviewer's expert view. According to the reviewer's advice, we performed new immunoblotting experiments to examine the H3K4me3 level in Setd2 +/+ and Setd2 -/lymphocytes. As shown in supplementary Figure. 8b, c of the revised manuscript, Setd2 deletion did not cause alterations in H3K4me3 and Rag2 (Fig5a,b and supplementary Fig. 8a). In addition, we performed ChIP assays of H3K4me3 at TCRβ and Igh loci, and did not detect significant reduction of H3K4me3 depositions at TCRβ and Igh genes by Setd2 knockout. This data is incorporated to the revised manuscript as supplementary Figure. 8. Collectively, these data suggest that the aberrant V(D)J rearrangement in Setd2 deficient lymphocytes is caused by ablation of H3K36me3 without affecting H3K4me3.

Figure legends need better description instead of stating the results.
Reply: We thank the reviewer for the careful reading of our manuscript, we have expanded the figure legends and included more detailed description in the revised manuscript.
Minor concerns. 1. The text needs work. There are several typos, multiple grammatical, and some sentences are confusing. The introduction has some references that are not up to date.