The H2B deubiquitinase Usp22 promotes antibody class switch recombination by facilitating non-homologous end joining

Abstract

Class switch recombination (CSR) has a fundamental function during humoral immune response and involves the induction and subsequent repair of DNA breaks in the immunoglobulin (Ig) switch regions. Here we show the role of Usp22, the SAGA complex deubiquitinase that removes ubiquitin from H2B-K120, in the repair of programmed DNA breaks in vivo. Ablation of Usp22 in primary B cells results in defects in γH2AX and impairs the classical non-homologous end joining (c-NHEJ), affecting both V(D)J recombination and CSR. Surprisingly, Usp22 depletion causes defects in CSR to various Ig isotypes, but not IgA. We further demonstrate that IgG CSR primarily relies on c-NHEJ, whereas CSR to IgA is more reliant on the alternative end joining pathway, indicating that CSR to different isotypes involves distinct DNA repair pathways. Hence, Usp22 is the first deubiquitinase reported to regulate both V(D)J recombination and CSR in vivo by facilitating c-NHEJ.

Fig. 1

Spleen B-cell profiles and total Ig isotype analysis of Usp22 KO or WT mice. a Schematic showing the dynamic regulation of histone H2B ubiquitination (Ub). Addition of ubiquitination to H2B is mediated by RNF20/RNF40, while the removal of ubiquitination from H2B is mediated by SAGA complex that is composed of Usp22, Eny2, Atxn7, and Atxn7L3. b The structure of targeted usp22 locus. The protein-encoding exon 2 (133 bp) of the usp22 gene was flanked by loxP sites (triangles). Usp22^flox/flox mice are bred to CD19-cre or Mb1-cre mice to delete exon 2, leading to disruption of usp22 gene expression specifically in B cells. c Spleen B cells were purified from CD19-cre-Usp22 KO mice or WT littermates. qPCR analysis of Usp22 mRNA was performed with resting spleen B cells or B cells that were stimulated with LPS for 3 days ex vivo (n = 3 mice per group). Data represent two independent experiments. d Spleen B cells from CD19-cre-Usp22 KO and WT littermate mice were stimulated with LPS for 2.5 days, exposed to 8 Grays of γ-radiation, and then collected at various time points for western blot analysis of H2B monoubiquitination (H2BK120Ub). Data represent four independent experiments. e Absolute numbers analysis of spleen B cells from CD19-cre-Usp22 KO and WT littermates. FO, follicular B cells; MZ, marginal zone B cells. Data represent two independent experiments each with three to four mice per group. f Serum analysis of various Ig isotypes in unimmunized CD19-cre-Usp22 KO and WT littermates, as well as fecal IgA from the same mice (n = 10 mice per group). Data in c, e, and f were presented as mean ± SEM and were analyzed using two-tailed unpaired Student’s t-test. *p < 0.05 and **p < 0.01; NS, not significant

Fig. 2

Usp22 KO mice exhibited defective IgG/IgE CSR but not IgA CSR. a NP₂₅-CGG plus alum was intraperitoneally injected into CD19-cre Usp22 KO or WT littermate mice. At d22 post immunization, mice were subjected to NP-specific IgG⁺ ELISPOT assay to measure the amount of NP-specific IgG⁺ antibody secreting cells (ASC) in the spleen (n = 7 mice per group). Data were combined from two independent experiments. b Rotavirus (RV) was orally administered to CD19-cre-Usp22 KO or WT littermates. Fecal samples were collected at the indicated time-points post-RV infection, followed by fecal supernatant preparation. The 1:2 dilution of fecal supernatant was used for fecal anti-RV IgA ELISA assay. Data was analyzed by two-way ANOVA. c ELISPOT assays were performed at d23 post-RV infection, to measure the numbers of IgA⁺ RV-specific antibody secreting cells in lamina propria (LP) and bone marrow (BM), respectively. Data in b and c represent two independent experiments each with three mice per group. d Splenic B cells from CD19-cre-Usp22 KO or WT littermates were ex vivo induced to switch to IgG1, IgG2b, IgG3, IgE, and IgA (n = 3 mice per group). e Same as d, except that Mb1-cre Usp22 KO mice were used (n = 3–4 mice per group). Data in d and e represent three independent experiments; data in a and c–e were analyzed using two-tailed unpaired Student’s t-test. Data were presented as mean ± SEM. *p < 0.05 and ***p < 0.001; NS, not significant

Fig. 3

Impaired γH2AX levels in spleen B cells from CD19-cre-Usp22 KO mice. a AID mRNA measured by qPCR of d3-LPS stimulated spleen B cells from CD19-cre-Usp22 KO or WT littermates (left panel; n = 5 mice per group). AID blot analysis of d3-LPS stimulated spleen B cells from Mb1-cre or CD19-cre Usp22 KO or their WT littermates (right panel; n = 2–3 mice per group); AID-KO cells were used as control and data represent two or three independent experiments. b Spleen B cells from CD19-cre-Usp22 KO and WT littermates were stimulated for IgG3 CSR (by LPS) or for IgA CSR (by LPS+ IL-4+ TGF-β+ IL-5+ anti-IgD dextran) for 2.5 days, exposed to γ-radiation, and then collected for γH2AX immunofluorescent staining. Line within micrograph represents 10 μm. The number of γH2AX foci per cell was analyzed 1 or 4 h post irradiation for Usp22-WT (black circles) and CD19-cre-Usp22-KO (white circles) (bottom panel), and representative images for IgG3-CSR group (γH2AX in green and DAPI in blue) are shown in the top panel. Approximately 30–50 cells were analyzed per sample and data represent two or three independent experiments. c Representative flow cytometry plots (left panel) and compiled cell cycle analysis (right panel) of Usp22-WT and CD19-cre-Usp22-KO spleen B cells. Cells were stimulated with LPS for 2.5 days and then incubated with EdU for 4 h, before FACS staining (n = 3 samples per group). d Purified splenic B cells from CD19-cre Usp22-KO or WT littermates was stimulated with LPS and growth of these cells was monitored over time (n = 3 samples per group). Data were tested by two-way ANOVA. e CFSE-pulsed spleen B cells from CD19-cre-Usp22 KO or WT littermates were stimulated by LPS plus IL-4 for CSR to IgG1 for 4 days. The gating (G2 to G11) is based on the intensity of CFSE and the percentage of IgG1 switch per gate is summarized (n = 3 samples per group). Data in c–e represent two independent experiments; data in a–c and e were analyzed using two-tailed unpaired Student’s t-test. Data were presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001

Fig. 4

Microhomology analysis of switch junctions of IgG1- or IgA-switched spleen B cells. Spleen B cells isolated from CD19-cre-Usp22 KO or WT littermates were induced for IgG1 (a and c) or IgA (b and d) CSR ex vivo for 4 days. IgG1 and IgA switch junctions were cloned from these cells, and subjected to sequencing. Approximately 50–60 unique switch junction sequences were analyzed per group. a Sμ/Sγ1 or b Sμ/Sα junctions were categorized based on their microhomology usage: near-blunt joins (0–2 bp), joins with microhomology (3–5 bp or 6 + bp), or insertions. The microhomology length (MH) used by each unique switch junction sequence are summarized (right panel). c Distribution of breakpoints across μ or γ1 switch regions for IgG1-switched Usp22-WT or -KO cells. d Distribution of breakpoints across μ or α switch regions for IgA-switched Usp22-WT or –KO cells. Data in a–d were combined from three mice per group; data in a and b were presented as mean ± SEM and were analyzed using two-tailed unpaired Student’s t-test. NS, not significant

Fig. 5

IgG1 CSR primarily relies on c-NHEJ, whereas IgA CSR is more dependent on A-EJ. a Spleen B cells isolated from 53BP1 KO or WT littermates were induced for ex vivo IgG1, IgG2b, IgG3, or IgA CSR for 4 days. 53BP1^−/− B cells (deficient in c-NHEJ) exhibited a less severe defect in IgA CSR (~ 60% reduction) compared with IgG CSR (~ 80–90% reduction). Data represent two independent experiments each with three to four mice per group. b A-196 (10 µM) was used to inhibit c-NHEJ during the ex vivo CSR of WT spleen B cells and A-197 is the control compound. Inhibition of c-NHEJ markedly reduced IgG1 CSR, but not IgA CSR (n = 4 samples per group). Data represent three independent experiments. c, d The switch region sequence of Sμ on the IgH locus of C57BL/6 background mice were compared with that of c Sα or d Sγ1. Dot Matrix analysis of the mouse switch regions. The dots represent homologies between two switch regions with a search length of 30 bp and defined percentage of identity. Data in a and b were presented as mean ± SEM and were analyzed using two-tailed unpaired Student’s t-test. ***p < 0.001; NS, not significant