CFP1 promotes B cell class switch recombination through the regulation of S region germline transcription and proliferation

Summary

Antibody class switch recombination (CSR) is a transcription-coupled process involving targeted DNA deamination and coordinated DNA repair in the immunoglobulin heavy-chain (Igh) constant gene locus. The epigenetic regulation mechanism of CSR is incompletely understood. Here, we show that the histone methyltransferase component CFP1 is critically required for CSR in activated B cells. Although CFP1 ablation does not negatively affect Aicda expression or DSB repair, it impairs switch (S) region germline transcription and disrupts AID activity-mediated hypermutation. CFP1 deficiency reshapes the global H3K4me3 landscape at promoters of cell cycle-related genes, including Pcna, leading to its mis-regulation, impaired cell proliferation, and genome instabilities. Therefore, CFP1 coordinates chromatin states and transcription to ensure effective CSR via epigenetic regulation.

Graphical abstract

Highlights

• •CFP1 promotes germline transcription at Sμ and Sγ1
• •CFP1 deficiency impairs AID targeting and SHM
• •CFP1 and H3K4me3 correlate with B cell proliferation
• •CFP1 safeguards genome stability in activated B cells

Immunology; Transcriptomics

Introduction

Class switch recombination (CSR) is an intrachromosomal recombination event that occurs within G-rich tandem repetitive switch (S) regions positioned 5′ of each heavy-chain constant (CH) gene (except Cδ) in the IgH locus.¹ The initiation of CSR depends on the coordinated action of B cell activation signals, cytokine microenvironments and DNA repair mechanisms.² Specifically, stimulation through CD40 and/or TLR signaling provides activation signals that drive CSR, enabling the switch from IgM to antibody isotypes such as IgG1, IgA, IgE, and IgG3 through precisely regulated intrachromosomal deletion.³

A critical prerequisite for CSR is germline transcription (GLT) that is initiated from the I promoters, proceeds through the I exons and CH exons/introns.^1,4 GLT activation recruits activation-induced cytidine deaminase (AID) to single-stranded DNA in the Sμ and downstream S regions to catalyze cytosine deamination, converting deoxycytidines (dC) to deoxyuridines (dU).² The resulting G-U mismatches are subsequently processed by base-excision repair (BER) and mismatch repair (MMR) machineries to generate staggered double-strand breaks (DSBs) at participating S regions.^5,6 AID-initiated DSBs engage the canonical DNA damage response (DDR), recruiting critical mediators including ATM kinase, γH2AX, RIF1, and 53BP1 that protect broken DNA ends from excessive resection⁷ and guide their repair predominantly in a deletional fashion⁸ via classical non-homologous end joining (C-NHEJ), with alternative end joining (A-EJ) as a backup pathway.³

As transcription occurs on chromatin templates, it is conceivable that epigenetic regulation plays important roles in CSR. Emerging evidence has revealed the crucial roles of chromatin dynamics in B cell activation and CSR,^9,10 but the underlying mechanisms are only partly understood. The trimethylation of histone H3 at lysine 4 (H3K4me3), catalyzed by the Set1/COMPASS-family of histone lysine methyltransferases (HMTs), is a hallmark of transcriptionally active promoters.¹¹ The CXXC-finger protein 1 (CFP1) is a unique subunit of Setd1A/B that directs the complexes to CpG-rich promoters.¹² Highly overlapping with Setd1A in binding near transcription start sites, CFP1 is essential for recruiting and defining occupancy of the Setd1 complex on chromatin to place H3K4me3 modification.¹³ Loss of CFP1 significantly reduces SET1A chromatin binding and impairs global H3K4me3 patterns across diverse cell types, including embryonic stem cells,¹² T lymphocytes,¹⁴ and oocytes,¹⁵ underscoring the dependency of CFP1 on the methyltransferase complex. That being said, the functional relationship between H3K4me3 and transcription activation remains debated.^11,16 Our recent study indicated that CFP1 controls early B cell development¹⁷ and germinal center (GC) dynamics through the H3K4me3-mediated transcriptional regulation of lineage-specific genes.¹⁸ The role of CFP1 in CSR is less clear. A prior study indicated that while CFP1 is required for H3K4me3 installation in S regions, its knockdown did not affect GLT but instead reduced S region DSB formation.¹⁹ Conversely, the MLL3/4 (KMT2C/2D) complex component PTIP has been implicated in CSR in part by establishing H3K4me3 marks at S regions for GLT activation.²⁰ Therefore, the precise roles for H3K4me3 modification in CSR-activated B cells remain to be further elucidated.

In this study, we establish CFP1 as a critical regulator of CSR. Through CRISPR/Cas9-mediated deletion in murine cell lines and conditional knockout in primary B cells, we demonstrate that CFP1 deficiency disrupts GLT while leaving Aicda expression, enzymatic activity, and DSB repair intact. Furthermore, we reveal that CFP1 orchestrates the transcriptional program of activated B cells by mediating H3K4me3 modifications at genes governing cell cycle progression, proliferation, and B cell activation, thereby linking epigenetic regulation to CSR efficiency.

Results

CFP1 is critical for class-switch recombination in activated B cells

To determine whether CFP1 plays any role in CSR, we first generated Cfp1-deficient CH12F3 cells, a well-established line to model IgM-to-IgA switching upon stimulation with anti-CD40, IL-4, and TGF-β (CIT),²¹ by using CRISPR-Cas9 and obtained two independent knockout clones (Figures S1A and S1B). CFP1 deficiency led to a marked reduction in IgA switching efficiency, from ∼40% in wild-type controls to ∼16% in knockout clones, consistent with a previous report¹⁹ (Figure 1A). To further validate this phenotype in primary B cells, we generated conditional Cfp1 knockout mice by crossing Cfp1^flox/flox mice¹⁴ with Cd21-cre mice to achieve specific and efficient deletion of Cfp1 in mature B cells (Figures S1C and S1D). Upon ex vivo stimulation of splenic naive B cells with LPS, IL-4, TGF-β, IL-5 and δ-dextran, the frequency of IgA⁺ cells was reduced by 8.7-fold in CFP1-deficient cells compared with Cfp1^f/f controls, accompanied with an over 6-fold reduction in switching to IgG1 (Figure 1B). Similarly, stimulation with LPS/IL-4 resulted in markedly reduced switching to IgG1 (Figure 1C), and with LPS alone led to impaired IgG3 switching (Figure 1D). These findings demonstrate that CFP1 is broadly required for efficient CSR in primary B cells. Furthermore, digestion-circularization quantitative PCR (DC-qPCR)²² showed nearly complete loss of Sμ-Sγ1 circular DNA in CFP1-deficient B cells following LPS/IL-4 stimulation (Figures 1E and 1F), further confirming a severe defect in CSR efficiency upon Cfp1 deletion.

Figure 1.

CFP1 is critical for class-switch recombination in activated B cells

(A) Representative flow cytometry plots and ratio of IgA CSR efficiency in CFP1-deficient CH12F3 cells under CIT stimulation for 3 days (n = 3).

(B and C) Representative flow cytometry plots and ratio of IgG1 and IgA switching in Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells under LPS/IL-4/TGF-β/IL-5/δ-glucoside (n = 4) (B) or LPS/IL-4 conditions (n = 6–7) (C).

(D) Representative flow cytometry plots and ratio of the IgG3 switching in Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells under LPS conditions (n = 7–8).

(E) Schematic of DC-PCR with genomic DNA from stimulated B cells.

(F) Detection of deletional Sμ-Sγ1 joining by DC-PCR in stimulated splenic B cells from Cfp1^fl/fl and Cfp1^fl/flCd21^Cre (n = 3), Chrnb1 served as the internal control. For all panels showing quantitative data, data are shown as mean ± SEM. p values by unpaired t test (two-tailed).

CFP1 is required for optimal S region GLT and SHM at S regions

Isotype specificity during CSR is directed by GLT across switch regions, which facilitates the recruitment of AID to sites undergoing DSBs and recombination.^2,23 To assess the role of CFP1 in GLT, we measured initiating (I) transcripts from switch region promoters and spliced GLTs by qPCR (Figure 2A). CFP1-deficient activated primary B cells exhibited roughly 30–50% reductions in both initiating and spliced transcripts at the μ and γ1 loci (Figure 2B), indicating that CFP1 is required for optimal GLT activation at switch regions. However, given the relatively modest GLT reduction and drastic CSR defect of CFP1-deficient cells, additional regulatory mechanisms likely contribute to the pronounced CSR deficiency phenotype.

Figure 2.

CFP1 is required for optimal S region GLT and AID targeting

(A and B) Illustration shows the location of RT-qPCR primers for spliced and initiating Igh switch transcripts (A) and quantification of these transcripts (B) in Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells (n = 3).

(C) RT-qPCR analysis of Aicda gene in Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells stimulated under LPS/IL-4 for 2 days (n = 6).

(D) Schematic of the 5′ Sμ SHM-seq workflow in splenic B cells from Aicda^−/−, Cfp1^fl/fl and Cfp1^fl/flCd21^Cre mice (Upper). Mutation profiles in the above B cells stimulated with LPS/IL4 for 4 days (Lower). Green bars indicate SEM. Positions of AGCT and WRC (W: A/T; R: A/G) motifs are marked with orange and yellow bars, respectively.

(E) Normalized mutation frequency in Cfp1^fl/fl and Cfp1^fl/flCd21^Cre mice (n = 3).

(F) Gene expression in splenic B cells stimulated with LPS/IL-4 for 2 days by mRNA-seq (n = 3). For all panels showing quantitative data, data are shown as mean ± SEM. p values by unpaired t test (two-tailed).

AID-induced DSBs within transcribed S regions are essential for CSR.² Thus, the observed CSR defect in Cfp1-deficient cells could potentially result from impaired AID expression, enzymatic activity, or downstream DSB repair. CFP1 loss does not reduce but rather elevates Aicda transcript levels in activated primary B cells (Figure 2C). As AID lesions in activated B cells can also be converted to somatic mutations, to further evaluate the effect of CFP1 on AID functionality, we measured the mutation frequency of the 5′-Sμ region in primary B cells under LPS + IL-4 stimulation using a high-throughput sequencing-based SHM-seq assay²⁴ (Figure 2D). While control B cells accumulated mutation rates up to 10⁻³/base/generation enriched at the preferred WRC motifs, Cfp1-deficient B cells exhibited an approximately 3-fold reduction in mutation rates (Figures 2D and 2E), indicating defective AID-mediated mutagenesis. To further determine whether the defect resulted from impaired generation or repair of AID-lesions, we next examined the expression of BER and MMR genes. mRNA-seq of LPS/IL-4–stimulated B cells revealed unchanged Ung expression and no significant reduction, or in some cases mild upregulation of core MMR genes including Msh2, Msh6, Exo1, Mlh2, and Pms2 following CFP1 loss (Figure 2F), indicating that neither the BER nor MMR pathway is transcriptionally compromised in the absence of CFP1. Although we cannot rule out the possibility of reduced targeting and enzymatic activity of the NER/BER proteins at S regions in the absence of CFP1, the diminished S region GLT and reduced SHM collectively suggest that CFP1 is required for optimal GLT at S regions, which could in turn lower AID targeting.

CFP1 deficiency does not affect the patterns of DSB end joining

We next investigated if CFP1 plays any role in the joining of DSBs. Although NHEJ is the primary repair pathway for CSR-associated DSBs, A-EJ pathways also contribute to repair.⁷ NHEJ pathway generally operates with minimal DNA end resection and relies on the 53BP1–RIF1 complex to protect DSB termini, whereas A-EJ is characterized by extensive end resection and frequent use of short microhomology (MH) at junctions.^7,8,25,26,27 To assess whether CFP1 deficiency affects these repair pathways at the transcriptional level, we first quantified mRNA levels of 53bp1 and Rif1, along with core NHEJ factors Xrcc5, Nhej1, Xrcc4, and Lig4 (Figure 3A). The expression of these genes was unchanged or slightly elevated in CFP1-deficient cells. Likewise, A-EJ components Ercc1 and Lig1 were also unaffected or modestly upregulated (Figure 3A), indicating that CFP1 does not regulate the expression of key end-joining factors. To directly examine whether CFP1 plays a role in DSB end joining, we performed a Cas9-mediated IgG1 switching assay in CH12F3 cells electroporated with Sμ/Sγ1-targeting sgRNAs²⁸ (Figure 3B). CFP1-deficient cells exhibited comparable end-joining efficiency to WT cells at 48 h post-electroporation, excluding a role in the DSB end joining (Figures 3C and 3D). To further investigate the impact of CFP1 deficiency on AID-initiated physiological S region DSB joining, we evaluated Sμ-Sγ1 and Sμ-Sε junction profiles in LPS/IL-4-activated primary B cells by using high-throughput genome-wide translocation sequencing (HTGTS).^8,29 Much fewer S-S junctions were recovered from CFP1-deficient B cells, implicating less efficient generation or repair of S region breaks (Figure 3E). The overall distribution of deletional versus inversional Sμ-Sγ1 junctions was comparable between CFP1-deficient and controls, where deletions predominate (Figures 3E and 3F). Unlike the DDR and core NHEJ-deficient mutants,²⁶ CFP1-deficient B cells exhibited no significant difference in the frequency of junctions involving end resection of core Sγ1 DSBs (Figure 3F) and microhomology (MH ≥ 4 bp) (Figure 3G). Collectively, these data demonstrate that CFP1 loss does not affect the end-joining pattern of S region breaks.

Figure 3.

CFP1 knockout does not affect DSB end joining

(A) DSB repair-related gene expressions in splenic B cells stimulated with LPS/IL-4 for 2 days assessed by mRNA-seq (n = 3).

(B) Schematic of Cas9/sgRNA-mediated IgG1 switching analysis.

(C and D) Representative flow cytometry plots (C) and quantification (D) of the IgG1 class switching in CH12F3 cell nucleofection of sgRNAs targeting Sμ and Sγ1 for 48 h (n = 7).

(E, F, and G) Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells stimulated with LPS/IL-4 for 4 days (n = 3). (E) Distribution of Sμ-Sγ1 and Sμ-Sε junctions recovered from stimulated B cells. Numbers in parentheses denote total unique junctions in the indicated regions. The area between the dotted lines indicates the core S region. (F) The ratio of inversional versus deletional junctions in Sγ1 and Sε regions (left) and percentage of end resection in Sγ1 and Sε regions (right) (G). The percentage of MH utilization within Sγ1 and Sε regions. For all panels showing quantitative data, data are shown as mean ± SEM. p values by unpaired t test (two-tailed) in (A) and (F). p values by Dunnett’s multiple comparisons test in (D).

Cfp1 ablation impairs cell proliferation

Since cell proliferation is essential for efficient class switching,³⁰ we next examined whether CFP1 regulates B cell proliferation by CFSE labeling followed by 96-h culture with LPS/IL-4 or anti-CD40/IL-4. CFP1-deficient B cells exhibited severe proliferation defect under either condition (Figures 4A and 4C). B cells underwent up to eight divisions within 96 h of stimulation. The frequency of switched cells at each division was significantly reduced upon CFP1 loss (Figures 4B and 4D). Viable cell counts were also markedly lower in Cfp1-deficient B cells under either LPS/IL-4 or anti-CD40/IL-4 stimulation (Figures 4E and 4G). Consistent with the observed proliferation defect following Cfp1 knockout, cell cycle analysis via propidium-iodide (PI) staining revealed that LPS/IL-4 or CD40/IL4 stimulated Cfp1^fl/fl Cd21^Cre B cells exhibited a higher fraction of cells in the G1 phase and a concomitant reduction in S/G2-M phase compared with controls (Figures 4F and 4H). This cell cycle arrest aligns with the observed reduction in CSR efficiency upon CFP1 loss. Collectively, these data support that CFP1 deficiency impairs class switching by restricting B cell proliferation.

Figure 4.

Cfp1 ablation impairs cell proliferation

(A) Representative flow cytometry plots and proliferation index analysis of CFSE-labeled Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells cultured with LPS/IL-4 for 96 h (n = 5–6).

(B) Representative flow cytometry plots and the percentage of IgG1 class switching per cell division, tracked by CFSE dilution after 96 h of LPS/IL-4 stimulation (n = 5–6).

(C) Representative flow cytometry plots and proliferation index analysis of CFSE-labeled Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells cultured with CD40/IL-4 for 96 h (n = 5–6).

(D) Representative flow cytometry plots and the percentage of IgG1 per cell division, tracked by CFSE dilution after 96 h of CD40/IL-4 stimulation (n = 5–6).

(E and G) Live cell counts of Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells at 0 h, 48 h, 72 h, and 96 h after LPS/IL-4 (E, n = 3) or CD40/IL4 stimulation (G, n = 3).

(F and H) Cell-cycle analysis of splenic B cells stimulated with LPS/IL-4 (F, n = 4–5) or CD40/IL4 (H, n = 3) for 48 h. Bars represent the mean values in (B), (D), (F), and (H). For all panels showing quantitative data, data are shown as mean ± SEM. p values by unpaired t test (two-tailed).

Cfp1 regulates cell proliferation-related gene expression through H3K4me3 modification

To investigate whether H3K4me3 modification-related function of CFP1 regulates gene expression in activated B cells during class switching, we performed the CUT&Tag assay and identified 13,815 H3K4me3 peaks genome-wide (Figure S2A). The vast majority of H3K4me3 peaks were retained upon CFP1 loss, and the signal intensity remained largely unchanged (Figure S2B). Differential binding analysis by DiffBind (FDR < 0.05) revealed 2,871 increased and 1,817 decreased H3K4me3 peaks upon CFP1 depletion (Figure S2C). Notably, over 90% of decreased peaks and over 70% of increased peaks localized to gene promoter regions (Figure S2D). Aligning with these H3K4me3 changes, mRNA-seq profiling identified 1,164 upregulated and 656 downregulated genes in CFP1-deficient activated B cells relative to controls (p value < 0.05; absolute log2 fold change >0.5) (Figure 5A). Gene set enrichment analysis (GSEA) showed that the top three significantly downregulated pathways upon CFP1 loss were E2F_targets, G2M_checkpoint, and MYC_targets_V1 (Figure 5B), consistent with observed proliferation defects. Differentially expressed genes positively correlated with changes in H3K4me3 levels (Figure 5C), and genes corresponding to significantly altered H3K4me3 peaks exhibited concordant transcriptional changes (Figure S2E). Moreover, genes losing H3K4me3 upon CFP1 deletion showed significantly higher expression in control cells than those gaining this modification (Figure S2E). Genes with concurrent reductions in mRNA and H3K4me3 (“both down”) were enriched in pathways including antigen receptor-mediated signaling, cell proliferation, immune response, and lymphocyte activation (Figures 5D; S3A and S3B), whereas genes with increased mRNA and H3K4me3 (“both up”) were enriched in pathways such as Wnt signaling, neuron development and cell adhesion (Figures 5D, S3C and S3D).

Figure 5.

Cfp1 regulates cell proliferation-related gene expression through H3K4me3 modification

Cfp1^fl/fl and Cfp1^fl/flCd21^Cre splenic B cells are stimulated with LPS/IL-4 for 48 h for bulk mRNA-seq and CUT&Tag; n = 3.

(A) Volcano plot of differentially expressed genes (p value < 0.05, absolute log₂ FC > 0.5).

(B) GSEA enrichment plot for pathways ranked by P.adjust.

(C) Scatterplot showing the correlation between differential gene expression (DEGs in Figure 5A and altered promoter H3K4me3 levels in Cfp1-KO versus WT stimulated B cells. Pearson’s correlation coefficient is indicated.

(D) GO enrichment in “both down” and “both up” genes shows the top five pathways ranked by q-values for each category.

(E) Heatmap for mRNA and H3K4me3 signals of DEGs from Figure 5B.

(F) Genome browser tracks of mRNA and H3K4me3, for example, positive and negative cell proliferation-related genes.

(G) Live cell counts of CH12F3 cells transduced with lentivirus expressing the Pcna or empty vector (EV) shRNA at 0 h, 24 h, 36 h, and 48 h cultured in vitro (n = 3).

(H) Representative flow cytometry plots and ratio of IgA CSR efficiency in EGFP⁺ CH12F3 cells under CIT stimulation (n = 3). Bars represent the mean ± SD in (G). For all panels showing quantitative data, data are shown as mean ± SEM. p values by Dunnett’s multiple comparisons test. p values based on 48 h data in (G). See also Figure S4.

Notably, CFP1-deficient B cells under LPS/IL-4 stimulation exhibited significantly impaired H3K4me3 modification at Sγ1 and at Sμ and Eμ, accompanied by a subtle but significant reduction in Sγ1 and Sμ transcript levels (Figures S3E–S3G), consistent with the qRT-PCR results (Figure 2B). This data suggests that CFP1 supports CSR, at least in part, through H3K4me3-mediated transcriptional regulation of S region GLT.

CFP1-mediated H3K4me3 affects RNA Pol II dynamics in activated B cells

Recent studies have shown that H3K4me3 regulates the release of paused RNA polymerase II (Pol II),¹⁶ which facilitates the recruitment of AID to S region DNA targets.³¹ To assess whether Cfp1-deficiency in activated B cells affected Pol II association, we performed CUT&Tag analysis to examine Pol II binding at genomic loci with altered H3K4me3 levels upon CFP1 loss. Genomic regions with reduced H3K4me3 exhibited a decrease in Pol II occupancy, particularly at promoter-proximity (Figures S2F–S2G). Conversely, regions with increased H3K4me3 marks showed elevated Pol II association (Figures S2F–S2G). The Pol II pausing index was calculated as the ratio of its signal density (RPKM) at the proximal promoter region (−200 bp to +200 bp relative to the TSS) to that in the gene body (+400 bp from the TSS to the TES) (Figure S2H). Genes were separated into two groups: those with upregulated H3K4me3 and those with downregulated H3K4me3 following CFP1 loss. CFP1 deletion differentially affected the Pol II pausing index based on H3K4me3 changes. In genes with downregulated H3K4me3, the Pol II pausing index significantly reduced, whereas it increased in genes with upregulated H3K4me3 (Figure S2I). Analysis of Pol II distribution revealed that these changes in the pausing index were specifically driven by alterations in promoter-bound Pol II, but not in the gene body (Figure S2J). We concluded that CFP1-mediated H3K4me3 affects Pol II dynamics that are correlated with transcription levels.

Cell cycle defect by PCNA downregulation causes CSR impairment

To further validate downstream candidate genes of CFP1 and investigate the causal relationship between cell proliferation and CSR, we attempted to examine the impact of interfering with Pcna and Cks2, two “both down” genes identified from the RNA-seq and H3K4me3 CUT&Tag analyses in Cfp1-deficient primary B cells (Figures 5E and 5F; S3H and S3I), on CSR. To this end, we performed shRNA-mediated knockdown of these genes in CH12F3 cells and then assessed the effect on proliferation and IgA switching. Although both Pcna and Cks2 were efficiently knocked down, only PCNA deficiency recapitulated the proliferation and IgA CSR defects observed in CFP1-deficient CH12F3 cells (Figures 5G and 5H; S4A), whereas Cks2 knockdown had no significant effect (Figure S4B–S4D). These results confirm that robust cell proliferation is essential for class switching and demonstrate that Pcna expression is required for CFP1-dependent proliferation and CSR, whereas Cks2 appears to play a less critical role in CH12F3 cells.

CFP1 suppresses AID-independent genome instability in activated B cells

Our data suggested a role for CFP1 in promoting cell cycle progression in CSR-activated B cells. To address whether the proliferation defect by CFP1 loss is cytokine-dependent or pre-existing G1/S replication disorder, we analyzed the proliferation and cell cycle status of control and Cfp1-deficient B cells in an AID-deficient background (to preclude AID-induced DNA lesions). Unstimulated Aicda^−/− Cfp1^−/− CH12F3 cells exhibited a mild but significantly slower proliferation rate than the control Aicda^−/− cells (Figure S5A). Cell cycle progression analysis indicated a modest but consistent delay in G1/S transition, characterized by increased G1 population, a normal S phase, and a slightly reduced G2/M fraction (Figure S5B). Stimulation of Cfp1-deficient cells with CIT cytokines exacerbated this defect, causing a more severe proliferation impairment (Figure S5C) and a significant reduction in the proportions of cells in S and G2/M phases (Figure S5D). These results indicate that while Cfp1 deficiency causes a mild defect in baseline cell cycle entry, pronounced progression failure requires cytokine stimulation. To further address whether cytokine stimulation induces general replication abnormalities independent of AID, we took a parallel approach to stimulate Cfp1^f/f control and Cfp1^f/fCD21^cre splenic B cells with RP105, which promotes cell survival/proliferation without inducing AID.³² Under this condition, Cfp1-deficient cells failed to proliferate while the control cells expanded steadily (Figure S5E). Cell cycle analysis confirmed a significant accumulation in G1 and concomitant reduction in S and G2/M phases in Cfp1-deficient cells (Figure S5F). This cell cycle arrest implicated the accumulation of stimulation-induced, AID-independent genomic instabilities that preclude successful replication.

Finally, we analyzed γH2AX foci formation to determine whether CFP1-deficiency promotes general genome instabilities in proliferating B cells. Consistent with relatively normal cell cycle progression, very few γH2AX foci were observed in unstimulated control and Cfp1-deficient CH12F3 cells (Figures S6A and S6B), as well as primary Cfp1^f/f and Cfp1^f/fCD21^cre splenic B cells (Figure S6C). In sharp contrast, RP105 stimulation significantly increased γH2AX foci, an effect that was markedly exacerbated upon Cfp1 loss (Figures S6C and S6D). These findings collectively demonstrated that Cfp1 deficiency enhances AID-independent genome instabilities following B cell stimulation, implicating crucial roles for CFP1 in the suppression of cell activation-initiated, replication-associated DNA damages.

Discussion

Our study establishes CFP1 as a critical epigenetic regulator that orchestrates class-switch recombination to IgG1, IgA, and IgG3 in B cells. CFP1 plays dual roles in CSR by activating optimal GLT and supporting robust B-cell proliferation by modulating H3K4me3 to maintain the expression of cell-cycle-related genes.

A previous study showed that CFP1 is required for CSR by regulating S region DSB formation in CH12F3 cells.¹⁹ Our study indicated that CFP1 deficiency does not impact the efficiency of Cas9-mediated IgG1 switching by NHEJ, or the S-S junction profiles in regard to DSB joining pathway choice, suggesting that CFP1 does not regulate the end joining phase of DSB repair. However, given the apparent difference between Cas9-generated breaks and AID-initiated S region DSBs, and that our HTGTS assay only captures successfully joined sequences, the indistinguishable Cas9-CSR and S-S junction profile in CFP1-deficient B cells does not exclude a role for CFP1 in the generation of S region breaks. Indeed, reduced GLT at Sμ/Sγ1, diminished mutation frequency of 5′-Sμ region, as well as impaired recovery of S-S junctions in CFP1-deficient activated primary B cells, is consistent with the notion that CFP1 is required for AID recruitment to S regions for initiating DSBs via GLT regulation.

Prior studies suggest GLT-associated H3K4me3 contributes to S region chromatin configuration.^10,20 B cells deficient in PTIP similarly exhibited markedly reduced H3K4me3 at the recipient Sγ1 and Sγ3 regions. A noticeable difference, however, is that the CFP1-deficient cells showed significantly reduced H3K4me3 at Sμ, Sε, and 3′RR regions across the entire IgH, whereas PTIP-deficient cells displayed no such changes compared to controls.^10,20 This is in line with the much broader impact of CFP1 depletion on global H3K4me3 modifications relative to PTIP-deficiency, suggesting non-redundant roles for the CFP1-containing Setd1A/B and the PTIP-containing MLL3/4 complex in depositing H3K4me3 marks both globally and at the IgH. Our study, consistent with a prior report,¹⁹ demonstrated that the CFP1-interacting Setd1A/B complexes are the primary enzymes to deposit H3K4me3 in activated B cells. The mechanism underlying the selective effect of the recipient Sγ regions for PTIP is unclear but may involve its specific recruitment by certain transcription factors to these sites.^10,20 Given that MLL3/4 predominantly catalyzes H3K4 mono- and di-methylation,^33,34 an indirect effect of PTIP on the H3K4me3 modification at recipient Sγ regions cannot be ruled out. Furthermore, the much more severe defect in Sγ GLT in PTIP-deficient cells relative to CFP1-ablation suggests that PTIP may possess unique functions beyond H3K4me3 in regulating transcription at the specific S region for CSR.

Recent studies revealed that H3K4me3 is critical for lineage-associated gene transcription. In mouse embryonic stem cells, the depletion of CFP1 leads to a significant reduction in H3K4me3 at approximately half of its target loci, yet only a small subset of these genes altered transcription.³⁵ In oocytes, CFP1 deficiency results in a global loss of H3K4me3 modifications with limited impact on transcription.¹⁵ In contrast, T cells require CFP1 for maintaining both global H3K4me3 and RNA levels with the selective preservation of H3K4me3 at gene promoters essential for survival.¹⁴ Our recent study in GC B cells indicated that CFP1 is required for the transcriptional control of B-lineage and cell cycle genes, highlighting the cell-lineage dependent role of CFP1 in the epigenetic regulation of transcription.¹⁸ This study extended the cell type-specific paradigm to CSR-activated B cells, and move beyond reaffirming the correlation between proliferation and CSR by establishing a CFP1-regulated epigenetic program to attain the proliferative competence for successful CSR in activated B cells.³⁰

Moreover, our study uncovered a distinct regulation mode of H3K4me3 on Pol II dynamics. Pol II pausing was significantly reduced in genes with downregulated H3K4me3, which is driven by alterations in promoter-bound Pol II. Despite reduced pausing, genes exhibiting reduced H3K4me3 upon CFP1 loss showed significantly reduced mRNA levels, suggesting that the primary effect of CFP1-mediated H3K4me3 in activated B cells is on transcription initiation, but not the extension step. This conclusion contrasts with models proposed by Wang et al.¹⁶ and Hu et al.,³⁶ in which promoter-proximal H3K4me3 facilitates the release of paused Pol II to enable transcriptional elongation. The reason for this discrepancy could possibly be attributed to several key differences in the assay system. First, while we deleted the unique component of the Setd1A/B complex, the cited studies depleted subunits shared by all members of the COMPASS family of HMTs. Functional redundancy from other HMTs may maintain H3K4me3 modification in our system. Second, we employed a permanent knockout of CFP1 via Cre-mediated recombination, creating a chronic and adapted state, in contrast to the acute protein degradation of DPY30/RBBP5 used by cited studies. Accordingly, the cited studies observed more severe loss of H3K4me3. As highlighted by Hu et al.,³⁶ the maintenance of Pol II at promoters exhibits dose dependency on H3K4me2/3. The residual H3K4me3 in our system may remain above this critical threshold and is sufficient for releasing paused Pol II into gene bodies. In addition, a key limitation of current studies, including ours and those cited, is the measurement of total RNA Pol II rather than distinguishing its functionally distinct states--paused, active, and poised. This likely contributes to the disparate Pol II distribution profiles at the promoters and gene bodies across different studies. To precisely define the impact of H3K4me3 on the pause-release of Pol II and transcriptional dynamics, antibodies specific to different phosphorylation sites of the Pol II CTD to resolve these distinct functional states should be used in future studies.

Finally, our study implicated roles of CFP1 in suppressing general, AID-independent genome instabilities induced by B cell activation. Cytokine stimulation triggers rapid cell proliferation and robust transcription activation. The resultant increase in γH2AX foci may therefore stem from accelerated cell cycling and/or more frequent collision between the replication and transcription machineries. Critically, CFP1 ablation exacerbates these stimulation-induced instabilities, suggesting a role for CFP1 in suppressing replication-coupled DSBs that warrants further investigation. It will be important to determine whether CFP1’s function related to histone H3K4 trimethylation is involved in this process. Furthermore, although our study ruled out a role for CFP1 in NHEJ of S region breaks, we cannot yet exclude the possibility that CFP1 directly or indirectly regulates HR, the predominant pathway for repairing replication-coupled DSBs.

In conclusion, this study defines CFP1 as a key regulator for CSR, likely through H3K4me3-mediated transcriptional regulation of S region GLT and proliferation of activated B cells.

Limitations of the study

Several limitations in our study should be acknowledged. First, our CSR-HTGTS junction analysis relied on the recovery of successful S-S joining. Due to the severely impaired CSR, only a small number of junctions were obtained from CFP1-deficient B cells, which may not fully capture the global repair landscape of AID-induced breaks and limit statistical power. Second, Cas9-induced DSBs, while useful for probing NHEJ repair capacity, do not fully recapitulate the chromatin context and the nature of staggering S region breaks. Third, PCNA knockdown in CH12 cells phenocopies CFP1 loss; however, this observation was made in a cell line, and its generalizability to primary B cells, as well as the possible involvement of other CFP1 targets, is still open. Finally, although CFP1 loss results in reduced H3K4me3 at switch regions and proliferation-associated genes, the precise molecular mechanism by which CFP1 is recruited to these loci and couples H3K4me3 to transcription has not been established; thus, the data primarily support correlative rather than causal relationships.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Junchao Dong (dongjch@mail.sysu.edu.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Data from RNA-seq and CUT&Tag have been deposited in the NCBI Sequence Read Archive (SRA) under the accession code SRA: PRJNA1305694 (https://www.ncbi.nlm.nih.gov/Traces/study/?acc=PRJNA1305694).
• •Data from the mutation analysis of HTGTS and 5′Sμ regions have been deposited in NCBI Sequence Read Archive (SRA) under the accession code SRA: PRJNA1362705 (https://www.ncbi.nlm.nih.gov/Traces/study/?acc=PRJNA1362705).
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We thank Dr. Fei-long Meng at Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, for kindly providing Cd21^Cre mice. We thank Dr. Leng-Siew Yeap of Shanghai Jiaotong University Medical School for assistance with the SHM-seq data analysis. We thank the Core Facility of Medical Science (Guangzhou Campus), Sun Yat-sen University, for its help in equipment use and data analysis. This work was supported by the National Natural Science Foundation of China (32070892, 32270934 to J.D.), the General Program of Guangdong Basic and Applied Basic Research Foundation (2023A1515012425 to J.D.), and the Sanming Project of Medicine in Shenzhen (SZSM202011004 to C.C.). W.B. is supported by the China Postdoctoral Science Foundation (2023M743995), the Postdoctoral Fellowship Program of CPSF (GZB20240883), and Guangdong Basic and Applied Basic Research Foundation (2023A1515111048).

Author contributions

J.D. conceived the project, J.D. and W.B. designed the experiments, S.G., Y.Z., W.B., Y.Y., J.B., M.L., Y.C., B.Z., R.Z., and C.C. performed the experiments, Y.Z., S.G., W.B., and J.D. analyzed the data, S.G., Y.Z., and J.D. wrote the manuscript with help from W.B. All authors participated in the discussion and helped revise the manuscripts.

Declaration of interests

The authors declare no competing interest.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit Recombinant Monoclonal CGBP antibody	Abcam	Cat# ab198977; RRID: AB_3101764
Mouse anti-β-actin monoclonal antibody	Proteintech	Cat# 66009-1-lg; RRID: AB_2687938
Anti-trimethyl-Histone H3 (Lys4) Rabbit Antibody	Merck	Cat# 07-473; RRID: AB_1977252
Rpb1 CTD (4H8) Mouse Monoclonal Antibody	Cell Signaling	Cat# 2629S; RRID: AB_2167468
HRP-conjugated Goat Anti-Rabbit IgG(H + L)	Proteintech	Cat# SA00001-2; RRID: AB_2722564
Anti-mouse IgG, HRP-linked Antibody	Cell Signaling	Cat# 7076S; RRID: AB_330924
Rabbit Anti-Mouse lgG H&L Antibody	Abcam	Cat# Ab46540; RRID: AB_2614925
Guinea Pig Anti-Rabbit IgG H&L Antibody	Antibodies Online	Cat# ABIN101961; RRID: AB_10775589
Purified NA/LE Rat Anti-Mouse CD180	BD Pharmingen	Cat# 562191; RRID: AB_11153851
Phospho-Histone H2AX-S139 Rabbit mAb	Abclonal	Cat# AP1555; RRID: AB_3698593
Cy3-conjugated Goat anti-Rabbit IgG (H + L)	Abclonal	Cat# AS007; RRID: AB_2769089
Rat monoclonal anti-CD40	eBioscience	Cat# 16-0401-86; RRID: AB_468943
Anti-IgD dextran(mouse)	Fina Biosolutions	Cat# 0001
APC Rat Anti-Mouse IgM	BD Pharmingen	Cat# 550676; RRID: AB_398464
IgA Monoclonal Antibody (mA-6E1), PE	eBioscience	Cat# 12-4204-83; RRID: AB_465918
BB700 Rat Anti-Mouse IgG3	BD OptiBuild	Cat# 745825; RRID: AB_2743274
Biotin anti-mouse IgA Antibody	BioLegend	Cat# 407004; RRID: AB_315079
Biotin anti-mouse IgG1 Antibody	BioLegend	Cat# 406604; RRID: AB_315063
Biotin anti-mouse IgG3 Antibody	BioLegend	Cat# 406803; RRID: AB_315070
PE-Streptavidin	BioLegend	Cat# 405204
APC Rat Anti-Mouse CD45R/B220	BD Pharmingen	Cat# 561880; RRID: AB_398531
Chemicals, peptides, and recombinant proteins
Propidium Iodide Solution	Sigma-Aldrich	Cat# P4864
Hoechst 33342	Invitrogen	Cat# H3570
ProLongTM Gold Antifade Mountant	Invitrogen	Cat# P10144
Recombinant murine IL4	Peprotech	Cat# 96-214-14
Human TGF-β1	Peprotech	Cat# 96-100-21
Lipopolysaccharide Escherichia Coli O111:B4	Sigma-Aldrich	Cat# LPS25
T4 DNA Ligase	Thermo Scientific	Cat# EL0014
EcoRI-HF	New England Biolabs	Cat# R3101
Recombinant Murine IL-5	Peprotech	Cat# 215-15-25
Fixable Viability Dye eFluor 506	eBioscience	Cat# 65-0866-14
NIR Fixable Viability Dye	Biolegend	Cat# 423106
RPMI 1640 medium	Corning	Cat# 10040-CV
Fetal bovine serum	Vazyme	Cat# F103-01
β-mercaptoethanol	Amresco	Cat# 0482
L-Glutamine	Corning	Cat# 25-005-CI
Sodium pyruvate	Corning	Cat# 25-000-CIR
HEPES	Corning	Cat# 25-060-CI
MEM non-essential amino acid	Corning	Cat# 25-025-CI
Penicillin streptomycin	Corning	Cat# 30-002-CI
DMEM medium	Corning	Cat# 10-013-CV
TaKaRa Taq	Takara Bio	Cat# R001B
Phusion High-Fidelity DNA Plymerase	Thermo Scientific	Cat# F530XL
Dynabeads MyOne streptavidin C1 beads	Invitrogen	Cat# 65002
Complete, Mini Protease Inhibitor Cocktail	Roche	Cat# 11836153001
Critical commercial assays
Single Cytofunnel™ with White Filter Cards	Thermo Scientific	Cat# FIS59-910-40
ChamQ Universal SYBR qPCR Master Mix	Vazyme	Cat# Q711-02
BD Cytofix/Cytoperm Kit	BD Biosciences	Cat# 554714
DNA transfection reagent	Neofect	Cat# TF20121201
Mouse Pan B Cell Isolation Kit	Vazyme	Cat# CS202-01
HiScript III RT SuperMix	Vazyme	Cat# R323
CellTrace Cell Proliferation Kits	Invitrogen	Cat# C34571
VAHTS Universal V8 RNA-seq Library Prep Kit for Illumina kit	Vazyme	Cat# NR605-02
Deposited data
RNA-seq and CUT&Tag data	This paper	PRJNA1305694
Mutation analysis of 5′Sμ regions and HTGTS sequencing	This paper	PRJNA1362705
Experimental models: Cell lines
CH12F3	Nakamura et al.	RRID:CVCL E067
HEK293T	ATCC	Cat# CRL-3216
Experimental models: Organisms/strains
Cd21cre	The Jackson Laboratory	Cat# 006368
Cfp1fl/fl	Cao et al14	N/A
Aicda−/−	Muramatsu et al37	N/A
Oligonucleotides
gRNA sequences (see Table S1)	This paper	N/A
shRNA sequences (see Table S1)	This paper	N/A
PCR primers (see Table S1)	This paper	N/A
Recombinant DNA
PSpCas9(BB)-2A-GFP (pX458)	Addgene	48138
EGFP-tagged pLKO.1 vector	This paper	N/A
Software and algorithms
FlowJo v10	Tree Star, Inc.	https://www.flowjo.com/
Prism v9	GraphPad Software Inc	http://www.graphpad.com/
ImageLab v3.0	Bio-Rad Laboratories, Inc.	https://www.bio-rad.com/
IGV v2.16.0	Broad Institute	https://igv.org/
R (v4.0.0 and v4.4.2)	R Foundation for Statistical Computing	https://www.r-project.org/
Trim Galore v0.6.10	Babraham Bioinformatics	https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/
Hisat2 v2.1.0	Daehwan Kim et al.	http://daehwankimlab.github.io/hisat2/
Samtools v1.6	Wellcome Sanger Institute	http://www.htslib.org/
FeatureCounts v2.0.0	Liao Y et al.	https://subread.sourceforge.net/
DESeq2 v1.46.0	Love, M.I. et al.	https://bioconductor.org/packages/release/bioc/html/DESeq2.html
ClusterProfiler v4.14.4	Yu et al.	https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html
Bowtie2 v2.2.5	Langmead and Salzberg	https://bowtie-bio.sourceforge.net/bowtie2/index.shtml
MACS2 v2.2.6	Zhang et al.	https://pypi.org/project/MACS2/
IDR v2.0.4.2	ENCODE Project Consortium	https://github.com/nboley/idr
DiffBind v3.16.0	Ross-Innes et al.	https://bioconductor.org/packages/release/bioc/html/DiffBind.html
Homer v4.11.1	Heinz, S. et al.	http://homer.ucsd.edu/homer/
Seurat v5.1.0	Hao et al.	https://satijalab.org/seurat/
SingleCellExperiment v1.28.1	Bioconductor Core Team	https://bioconductor.org/packages/release/bioc/html/SingleCellExperiment.html
Other
MoFlo Astrios EQs cell sorter	Beckman Coulter, Inc.	N/A
CytoFLEX flow cytometer	Beckman Coulter, Inc.	N/A
Qubit 4 Fluorometer	Invitrogen	N/A
LightCycler 480 II Real-Time PCR System	Roche Diagnostics GmbH	N/A

Experimental model and study participant details

Mice

All mice were crossed onto C57BL/6J (000664, Jax) background and housed under specific-pathogen-free conditions at the animal facility of the Laboratory Animal Center of Sun Yat-sen University. All animal procedures were approved by the Animal Care and Ethics Committee of Sun Yat-sen University (approval no. SYSU-IACUC-2025-001066). All mouse experiments were conducted using mice 8 to 12 weeks old, data were pooled from both sexes for analysis. The following mouse strains on a C57BL/6J background were used: Cfp1^fl/fl,14 Cd21^Cre38 and Aicda^−/−.³⁷

Cell lines

CH12F3-derived cell lines were generated by CRISPR/Cas9-mediated gene targeting. The CH12F3 and its derived cells were maintained in RPMI 1640 media (Corning, 10040-CV) supplemented with 15% FBS (F103, Vazyme), 50 mΜ β-mercaptoethanol (0482-250 mL, Amresco), 2 mM L-Glutamine (25-005-CI, Corning), 1 mM sodium pyruvate (25-000-CI, Corning), 20 μM HEPES (25-060-CI, Corning), 1 × MEM non-essential amino acid (25-025-CI, Corning), 1 × penicillin streptomycin (30-002-CI, Corning) at 37°C with 5% CO₂.

Method details

Generation of CFP1 knockout cell lines via CRISPR/Cas9 system

In order to obtain CFP1 knockout, a total of 3 × 10⁶ CH12F3 cells were co-transfected with 2 μg of paired pSpCas9(BB)-2A-GFP (pX458, Addgene #48138) plasmids encoding sgRNAs that target the 5′ and 3′ intronic regions flanking the gene of interest. Electroporation was performed using the 4D-nucleofector X (Lonza, solution SF, protocol CA-137). At 48 h post-nucleofection, the GFP-positive cells were sorted into 96-well plates. Single-cell clones were screened by PCR and immunoblotting and were confirmed by Sanger sequencing.

Class switch recombination and proliferation assays

CH12F3 cell and its mutants stimulation were performed with anti-CD40 (1 mg/mL, eBioscience), IL-4 (20 ng/mL, PeproTech) plus TGF-β (1 ng/mL, PeproTech) for 72 h. Mature splenic B cells isolated using Mouse Pan B Cell Isolation Kit (CS202-01, Vazyme) were cultured in RPMI 1640 media (Corning, 10040-CV) supplemented with 15% FBS (F103, Vazyme), 50 mΜ β-mercaptoethanol (0482-250 mL, Amresco), 2 mM L-Glutamine (25-005-CI, Corning), 1 mM sodium pyruvate (25-000-CI, Corning), 20 μM HEPES (25-060-CI, Corning), 1 × MEM non-essential amino acid (25-025-CI, Corning), 1 × penicillin streptomycin (30-002-CI, Corning). The B cells were cultured at 5 × 10⁵ cells/mL in the lymphocyte medium described above in cell culture plate (JET Biofil) and were treated with LPS (5 μg/mL, Sigma), IL-4 (20 ng/mL, PeproTech), IL-5 (1.5 ng/mL, PeproTech), anti–δ-dextran (0.9 ng/mL, FinaBio) and TGF-β (6 ng/mL, PeproTech) so that CSR would be induced to IgA, with LPS (5 μg/mL, Sigma) plus IL-4 (20 ng/mL, PeproTech) or anti-CD40 (0.5 μg/mL, eBioscience) plus IL-4 (20 ng/mL, PeproTech) to IgG1, with LPS (5 μg/mL, Sigma) to IgG3. After stimulation for 96 h, cells were collected and stained for the surface makers with the indicated antibodies and analyzed by flow cytometry.

For the proliferation assay, splenic B cells were stained with CellTrace CFSE (C34554, Invitrogen) according to the manufacturer’s instructions then were stimulated with LPS + IL-4 or anti-CD40 + IL-4 for 96 h, CellTrace Violet dilution was assessed by flow cytometry.

For cell cycle analysis, splenic naive were stimulated with LPS + IL-4 or anti-CD40 + IL-4 for 48 h or RP105 (2.5 μg/mL, BD) for 48 h and 72 h at 37°C with 5% CO2, cells were collected and fixed with ice-cold 70% methanol at −20 °C for 3 h, treated with 10 μg/mL RNase A for 30 min at 37 °C, and then resuspended in PBS supplemented with 50 μg/mL PI (P4868, Sigma-Aldrich) to indicate DNA content. Data were collected on the Cytoflex Flow Cytometer (Beckman Coulter) and analyzed with the CytExpert (Version 2.3.1) or with the FlowJo software (Version 10.8.1).

shRNA knockdowns

HEK293T cells were transfected with EGFP-tagged pLKO.1 vector containing shRNA sequence targeting Pcna or Cks2, pMD2.G, and psPAX2 with DNA transfection reagent (Neofect, TF20121201). The lentivirus supernatant was collected 48 h after transfection and removed cell debris by centrifugation. Viral supernatants were mixed with CH12F3 and infected by centrifugation at 32°C 1,000×g 60 min. All the shRNA sequences are listed in Table S1.

Cas9-mediated CSR assay

For Cas9-mediated CSR in CH12F3 cells, sgRNAs targeting Sμ or Sγ1²⁸ was respectively cloned into pX458 and electroporated into WT, Cfp1−/− CH12F3 cells using the 4D nucleofector X unit (Lonza, solution SF, protocol CA-137). The percentage of IgG1 CSR in GFP+ subsets at 48 h post-nucleofection was analyzed.

RT-qPCR

Cells were collected and resuspended in TRIzol reagent (15596026, Invitrogen), total RNA was extracted. Purified RNA was analyzed on Nanodrop and Qubit Fluorometric Quantification for quality check and concentration measurement, respectively. HiScript III RT SuperMix for qPCR (+gDNA wiper) (R323, Vazyme) was then used to synthesize cDNA from the purified total RNA. RT-qPCR was conducted with ChamQ Universal SYBR qPCR Master Mix (Q711-02, Vazyme) on a LightCycler480 Real-Time PCR System. Hprt mRNA was used as an internal control. Primers for quantitative RT-qPCR are listed in Table S1.

DC-PCR

The DC–PCR assay was performed as described previously.^8,22 In brief, Cells stimulated with LPS plus IL-4 for 96 h were collected, DNA was extracted using chloroform extraction and precipitated with ethanol. 4 μg of genomic DNA was digested overnight with 20U of EcoRI (R3101, New England Biolabs). Ligations using T4 DNA ligase (EL0014, Thermo Scientific) were performed under diluted conditions to promote circularization. Digested DNA was ligated overnight at 16°C with a concentration of 1.8 ng/μL in a total volume of 100 μL per reaction. All ligation reactions were pooled, column purified, concentrated on Qubit Fluorometric Quantification. PCR was then performed using ChamQ Universal SYBR qPCR Master Mix (Q711-02, Vazyme) with 30 ng DNA on a LightCycler480 Real-Time PCR System. Primers were designed to amplify the Sμ-Sγ1 rearrangements that occur during CSR to IgG1 in direct chromosomal joining of Sμ-Sγ1 with excision of circular DNA. As a control for EcoRI digestion and circularization of input DNA, amplification of an EcoRI fragment of nicotinic acetylcholine receptor B subunit gene (Chrnb1) was performed. Primers for DC-PCR are listed in Table S1.

Mutation analysis of 5′Sμ region

To evaluate AID targeting in 5′Sμ region, splenic B cells were stimulated with LPS plus IL-4 for 4 days. The 5′-region of Sμ was PCR-amplified from gDNA with indicated primers in Table S1. The PCR product was further tagged with P5 and P7 index primers and sequenced on DNBSEQ-T7 platform with 150 bp paired-end reads. Data were analyzed based on the existing SHM pipeline.³⁹ The raw reads were mapped to the mouse genome mm10/GRCm38 using Bowtie2 (Version 2.3.5.1), and substitutions/deletions/insertions were called with the SAM file. The SHM pipeline reports the mutation frequency of each nucleotide. Inside the sequenced region, sites with a mutation frequency higher than 0.1% were picked for mutation spectrum analysis.

Western blot

To extract whole-cell lysates, at least 1 × 10⁵ cells were collected and washed once with ice-cold PBS and resuspended in 1% SDS buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA, 1%SDS) supplemented with 1 mM protease inhibitors PMSF and incubated on ice for 10 min. Lysates were separated by 12% Bis-Tris precast gels. After electrophoretic transfer to Immobilon PVDF membrane (IPVH00010, Millipore), membranes were blocked the 5% non-fat milk in PBS containing 0.1% Tween 20. Membranes were incubated overnight at 4°C with the following primary antibodies: Rabbit CFP1 (Anti-CGBP, ab198977, Abcam), Mouse β-actin (66009-1-Ig, Proteintech). Membranes were then incubated with HRP-conjugated anti-rabbit secondary antibody (SA00001-2, proteintech) or anti-mouse secondary antibody (7076, Cell Signaling Technology) and detected by ChemiDoc Touch Imaging System (Bio-Rad). Uncropped western blot images as Source Data in the Supplementary Information.

γH2AX IF

Relevant single-cell suspensions were processed by cytospin onto polylysine slides and fixed with 4% paraformaldehyde solution at 4°C for 15 min, followed by a 5-min permeabilization step using 0.2% Triton X-100. Slides were blocked in 5% bovine serum albumin in 0.05% PBST for 1 h at room temperature and stained with anti-γH2AX-Ser139 (1:1000) (AP1555; Abclone) overnight at 4°C followed by Cy3-conjugated goat anti-rabbit antibody (1:500) (AS007; Abclone) for 60 min at room temperature. Cell nuclei were stained with Hoechst 33342 (1:1000) (H3570, Invitrogen), and slides were mounted using cells were mounted onto slides using ProLong^TM Gold Antifade Mountant (P10144, Invitrogen) and examined with a ZEISS LSM 880 with Airyscan confocal microscope fit with 63× objectives.

HTGTS

HTGTS analysis was performed as previously described.^8,29 Briefly, 6 μg genomic DNA from 96 h LPS/IL-4 stimulated mouse splenic B cells was sonicated into fragments ranging from 300 to 1000 bp on an M220 Focused-ultrasonicator (Covaris, M220), and amplified by LAM-PCR with biotin-conjugated 5′ Sμ primer. Biotinylated PCR products were captured by Dynabeads MyOne streptavidin C1 beads (Invitrogen), ligated with bridge adapters on-bead. The ligated products were amplified by second-PCR to add adaptor. The ligated products were amplified by nested PCR using barcode primers, followed by a third round of PCR using P5-I5 and P7-I7 primers. 300–1,000 bp PCR products were enriched by gel separation and sequenced by paired-end 150 bp sequencing on DNBSEQ-T7 platform. The HTGTS data were analyzed and plotted as described.⁸ Primers used for library construction are listed in Table S1.

RNA-sequencing and analysis

Total RNA was isolated using Trizol-chloroform extraction. RNA integrity was assessed using the Qsep1 Bio-Fragment Analyzer system (Bioptic). RNA-seq libraries were prepared from 100 ng of purified RNA using the VAHTS Universal V8 RNA-seq Library Prep Kit for Illumina (NR605-02, Vazyme) following the manufacturer’s protocol with standard dNTPs. Pooled barcoded libraries were sequenced on an Illumina NovaSeq 6000 platform (150 bp paired-end). Library size distribution was verified using the Qsep1 system. RNA-seq was performed using biological replicates for all samples. Trim Galore (version 20.04.6) was used for adapter trimming of raw data, while FastQC (Version 0.11.9) and MultiQC (Version 1.12) were employed to assess sequencing quality. The clean reads were mapped to the mouse genome mm10/GRCm38 using Hisat2 (Version 2.1.0). Differential gene expression analysis of raw featurecounts was performed using R package DESeq2. Sorted BAM files were generated using samtools (Version 1.10.2) and then density bigWig files were generated by using the deepTools (Version 3.5.1) with a normalizing scale factor to calculate RPKM. IGV (version 2.18.2) was used to display specialized features among genomic regions with BigWig files as the input. The heatmap and volcano map were drawn using the R package ggplot2.

CUT&Tag and analysis

CUT&Tag was conducted as previously described.⁴⁰ The initial input of 1 × 10⁵ fresh cells was bound to activated Concanavalin A beads (BP531, Bang Laboratories), and incubated with 0.5 μL anti-H3K4me3 antibody (07–473, Millipore) or 0.5 μL Pol II antibody (Cell Signaling Technology, 2629 T) for 2 h at room temperature with rotation, 0.5 μL anti-rabbit IgG antibody (ABIN101961, Antibodies Online) were performed sequentially for 1 h at room temperature with rotation, then cells were washed three times with 500 μL Dig wash buffer, next 0.5 μL Hyperactive pA-Tn5 Transposase (S603, Vazyme) were added and incubated for 1 h. Afterward, the sample was washed three times with 500 μL of Dig-300 wash buffer, and 300 μL of tagmentation buffer was used for pA-Tn5 activation, the reaction products were subsequently subjected to DNA extraction. After PCR amplification was performed using the primers listed in Table S1, libraries were purified with 1.3 × DNA clean beads (N411, Vazyme) to get fragments larger than 250 bp. Barcoded libraries were pooled and sequenced on the DNBSEQ-T7 platform with 150 bp paired-end reads.

Trim Galore was used for adapter trimming of raw data. Alignments against the mm10 reference genome using BWA (Version 0.7.17) and sorted BAM files were generated using samtools. H3K4me3 peaks were called using MACS3 with p-values < 1e−5. We used the broad-peak function for a definition of broad peaks. The multiBigwigSummary feature of deepTools was used to calculate the RPKM values of H3K4me3 peaks for individual genes. The density bigWig files were generated by using the deeptTools with a normalizing scale factor to calculate RPKM values. IGV was used to display specialized features among genomic regions with BigWig files as the input. R package DiffBind was used for calculating differential peaks, and R package ChipSeeker was employed for annotating the peaks.

Pol II pausing index

The Pol II pausing index was calculated as previously described.⁴¹ In brief, Pol II density (RPKM) were mapped within the proximal promoter (−200bp; +200bp with respect to TSS) and the gene body (+400bp from TSS; TES). The Pol II pausing index was calculated using the ratio of Pol II density in the region (TSS± 200 bp) to that in the gene body. Genes shorter than 400 bp were excluded from the pausing index analysis.

Quantification and statistical analysis

Except when noted otherwise, p-values were determined using unpaired Student’s t-tests, Dunnett’s multiple comparisons test, unpaired Wilcoxon rank sum test or Kolmogorov-Smirnov tests. The specific statistical test used in each case is indicated in corresponding figure legends. p-values < 0.05 were considered significant. For all panels showing quantitative data, data are shown as mean ± SEM.