B cell TET2 recruits OGT for nuclear TET2 and H2B O-GlcNAcylation to drive AID and BLIMP-1 expression and maturation of the antibody response Authors: Shili Li, Yibo Peng, Zhimin Yang, Xin Li, Zhenming Xu, Carlos E. Rivera, Paolo Casali Published: 2026 ABSTRACT SUMMARYMaturation of antibody responses entails B cell Aicda/AID and Prdm1/BLIMP-1 expression for SHM/CSR, plasma cell differentiation, and production of class-switched high-affinity antibodies. We determined that TET1, TET2, and TET3 are not expressed in resting naïve B lymphocytes, and only TET2 is induced in differentiating B cells for AID and BLIMP-1 expression. B cell TET2 recruits OGT, a metabolic sensor and sole protein O-GlcNAcylator, for O-GlcNAcylation of itself and chromatin H2B-S112. TET2 O-GlcNAcylation supports TET2-mediated active DNA demethylation (5mC oxidation to 5hmC) of Aicda and Prdm1 loci. This, together with these loci H2B-S112 O-GlcNAcylation, promotes Aicda/AID and Prdm1/BLIMP-1 expression for maturation of T-dependent and T-independent antibody responses. TET2 recruits OGT through its C-terminal-end, as evidenced by Tet2OGT-Δmut C-terminal-end deletion mutant and AlphaFold 3.0-modeled TET2-DNA-OGT complex. Finally, B cell TET2 takes metabolic cues for Aicda/AID and Prdm1/BLIMP-1 expression, as shown by fumarate inhibition and vitamin C activation of TET2 in humanized THX mice and Tcrβ−/−Tcrδ−/− mice. INTRODUCTION Maturation of the antibody response entails B cell somatic hypermutation (SHM)/class-switch DNA recombination (CSR), plasma cell differentiation, and generation of specific memory B cells (MBCs). SHM and CSR are initiated by activation-induced cytidine deaminase (AID), encoded by Aicda.1 AID introduces nicks and double-stranded breaks in immunoglobulin loci DNA, the repair of which leads to the generation of clonal sub-mutants making antibodies with higher antigen affinity and new and diverse effector functions.2,3 Plasma cells, whose differentiation and maintenance require expression of BLIMP-1, which is encoded by Prdm1, are terminally differentiated elements producing large amounts of antigen-specific antibodies. MBCs express a high-affinity and mostly class-switched B cell receptor (BCR) and are reactivated by cognate antigen to proliferate and differentiate, thereby playing a critical role in anamnestic and specific antibody responses.4,5 In B cells, AID and BLIMP-1 expression is under stringent transcriptional regulation. Aicda transcription is initiated by binding of NF-κB and HoxC46–9 to the Aicda promoter,3,10,11 while Prdm1 transcription is initiated by binding of STAT3 and IRF4 to the Prdm1 promoter.12 Transcription of Aicda and Prdm1 is also regulated by epigenetic factors, such as histone deacetylases (HDACs), particularly HDAC Class III SIRT1, noncoding miRNAs, including miR-155, miR-181b, miR-93, and miR-9,13–19 and, as more recently shown, ten-eleven translocation (TET) methylcytosine dioxygenases.20,21 These enzymes are Fe++- and α-KG (α-ketoglutarate)-dependent and can catalyze genome-wide active DNA demethylation21–23 by sequentially oxidizing 5mC to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC). DNA 5hmC is used as an indicator of TET active DNA demethylation as it is the most stable TET-catalyzed intermediate and up to 1,000-fold more abundant than 5fC and 5caC.21 TETs display tumor-suppressor activity, and their mutations can lead to neoplastic transformation, particularly in the lymphoid lineage.20,21 TET1 is predominant in embryonic stem cells (ESCs).21,23–25 In mouse and zebrafish ESCs, it can interact with O-linked N-acetylglucosamine (GlcNAc) transferase (OGT), which enables this active DNA demethylator’s catalytic function.26 OGT, encoded by the X chromosome Ogt locus, is the sole enzyme attaching a GlcNAc sugar to Ser and Thr residues of nuclear and cytoplasmic proteins.27,28 As a critical post-translational modification (PTM), O-GlcNAcylation can affect protein structure and stability, and, therefore, function.27,28 OGT plays a critical role in regulating self-renewal and differentiation of ESCs, hematopoietic stem cells (HSCs), and immune cells,28–31 TET2 and TET3 are expressed in myeloid or lymphoid immune cells. In myeloid cells, Tet2 deletion, as in LysMCreTet2fl/fl mice, upregulated proinflammatory gene expression in macrophages.21 In T lymphoid lineage, combined Tet2/Tet3 deletion, as in CD4CreTet2fl/flTet3fl/fl mice, resulted in unregulated invariant natural killer T (iNKT) cell expansion.21,32 In contrast, deletion of Tet2 or Tet3 alone resulted in no or a mild phenotype, one entailing only a moderate increase in iNKT cells in thymus, spleen, and lymph nodes.21 Also, combined Tet2/Tet3 deletion in regulatory T cells, as in Foxp3CreTet2fl/flTet3fl/fl mice, impaired cell lineage specification to a greater extent than individual Tet2 or Tet3 deletion,21,23 suggesting that TET2 and TET3 are functionally redundant, at least in T lymphocytes. Combined acute Tet2/Tet3 deletion in all cell types, as in CreERT2Tet2fl/flTet3fl/flRosa26YFPLSL mice, decreased DNA 5hmC genome-wide, including the Aicda locus, leading to decreased Aicda expression and reduced NP-specific IgG1+B cell response to NP-OVA.33 Combined Tet2/Tet3 deletion in Pro-B cells, as in Mb1CreTet2fl/flTet3fl/fl mice, or in all B cells, as in CD19CreTet2fl/flTet3fl/fl mice, also led to a block of Pro-to Pre-B cell development or perturbation in mature B cell homeostasis.34,35 Finally, pan-B cell Tet2 deletion, as in VavCreTet2fl/fl mice was associated with decreased 5hmC deposition in the Prdm1 but not the Aicda locus.36,37 We have determined here that TET2, not TET3, is expressed to induce both Aicda and Prdm1 in mouse and human B cells, mediating a specific antibody response. The critical role of B cell TET2 and exclusion of TET3 in maturation of T-dependent and T-independent antibody responses was addressed using AicdaCreTet2fl/fl and AicdaCreTet3fl/fl mice. An integrative map of TET2-mediated DNA 5mC to 5hmC oxidation of Aicda and Prdm1 loci was constructed by combining our MeDIP/hMeDIP-Seq, TET2 ChIP-qPCR, and MeDIP/hMeDIP-qPCR findings from Tet2−/− B cells and AicdaCreTet2fl/flB cells with existing data.33,38,39 Prompted by the suggested TET1 interaction with OGT and subsequent protein O-GlcNAcylation in ESCs,26 we hypothesized that B cell TET2 recruits OGT through a non-catalytic adaptor function. TET2-recruited OGT would O-GlcNAcylate TET2 itself for TET2-catalyzed active DNA demethylation of Aicda and Prdm1 loci. In these loci, TET2-recruited OGT would also O-GlcNAcylate chromatin H2B-S112, leading to AID and BLIMP-1 expression, SHM/CSR, and plasma cell differentiation. This was addressed by deleting TET2 or OGT in activated and differentiating AicdaCreTet2fl/fl or AicdaCreOgtfl/fl B cells in vivo and in vitro. TET and OGT interacting factors in activated B cells were identified by immunoprecipitation (IP) of nuclear proteins with anti-TET2 and anti-OGT antibodies. TET2 interaction with OGT and DNA at Aicda and Prdm1 loci was visualized in a 3D-modeled TET2-OGT-DNA complex by leveraging the power of AlphaFold 3.0. To analyze OGT contribution to overall TET2 function, the ability of TET2 C-terminal-end (OGT-binding) deletion Tet2OGT-Δmut mutant to rescue CSR and plasma cell differentiation in Tet2−/− B cells was compared to that of the conserved signature catalytic HxD motif Tet2HxD-mut mutant, dual Tet2HxD-mut/OGT-Δmut mutant, or Tet2WT retroviral construct. TET2 catalytic activity (active DNA demethylation) and OGT catalytic activity (O-GlcNAcylation of TET2 and H2B-S112) in mouse and human B cells were also dissected using TET2 inhibitor TET-IN-C3540 and OGT inhibitor ST 045849.41 Finally, B cell TET2 ability to transduce metabolic cues that modulate maturation of specific antibody responses was analyzed in humanized THX mice and Tcrβ−/−Tcrδ−/− mice treated with TET2 inhibitor fumarate, a tricarboxylic acid (TCA) cycle intermediate, or TET2 activator vitamin C. RESULTS TET1, TET2, and TET3 in mouse and human B cells TET1 is predominant in ESCs, while TET2 and TET3 are in HSCs and immune cells.20,21,23 TET2 and TET3, however, are different in structure and have been suggested to play different roles in HSCs and immune cells.20,21,23 To analyze Tet1/TET1, Tet2/TET2 and Tet3/TET3 expression in B lymphocytes, we stimulated mouse naïve IgM+IgD+B cells with LPS plus IL-4 and cultured them in vitro. In resting naïve B cells, Tet transcripts were expressed at low levels, and TET proteins were not detectable (Figures 1A and 1B). After B cell activation, TET1 and TET3 remained undetectable. In contrast, TET2 increased in expression, together with Aicda/AID and Prdm1/BLIMP-1, and was phosphorylated and acetylated throughout 96 h. Similarly, naïve human (hu)IgM+IgD+B cells increased intracellular TET2 expression as they differentiated to germinal center (GC)-like huCD19+CD27−CD38+B cells and huCD19loCD27+CD38+plasma cells upon stimulation with huCD154, huIL-2, huIL-4, and huIL-21 (Figure 1C). Similar information on B cell TET2 expression was gathered from spleen and mesenteric lymph nodes (MLNs) of C57BL/6 mice immunized with NP-CGG conjugated hapten as well as human tonsils (Figure 1D). The lack of TET1 expression in B cells was consistent with a lack of TET1 role in immune cells.21,23–25,42 Similar to TET1, TET3 was not detected in differentiating B cells. However, because combined Tet2/Tet3 deletion in all cell types, as in CreERT2Tet2fl/flTet3fl/flRosa26YFPLSL mice, was shown to impair Aicda expression,33 we immunized AicdaCreTet3fl/fl mice, which delete Tet3 in activated and differentiating B cells, with NP-CGG to exclude any B cell TET3 contribution to maturation of the antibody response. Such mice showed levels of total and high-affinity NP4-binding IgM, IgG1, and IgA comparable to AicdaCreTet3+/+mice, a reflection of comparable class-switched IgG1+B cells, including NP5-specific IgG1+B cells in spleen, and plasmablasts and plasma cells in spleen and bone marrow (BM). AicdaCreTet3fl/fl mice also showed no apoptotic B cells as well as proportions of class-switched IgG3+, IgG1+, IgG2b+, and IgA+B cells, IgG3+, IgG1+, IgG2b+, and IgA+ class-switched (sw)MBCs and plasmablasts in MLNs and Peyer’s patches, comparable to AicdaCreTet3+/+ mice (Figure S1A). AicdaCreTet3fl/fl mice also expressed B cell AID and BLIMP-1 at levels comparable to AicdaCreTet3+/+ mice. They also showed normal IgA+ or IgD+B cells in intestinal lamina propria and normal IgM+B cells in Peyer’s patches (Figure S1B). The normal CSR and plasma cell differentiation of naïve IgM+IgD+AicdaCreTet3fl/flB cells stimulated in vitro with LPS or LPS plus IL-4 further confirmed that Tet3 deletion does not impact B cell differentiation (Figure S1C). Thus, naïve B cells do not express TET proteins; B cells undergoing GC and plasma cell differentiation express TET2, not TET3 or TET1. B cell TET3 deletion, as in AicdaCreTet3fl/fl mice, did not impair AID and BLIMP-1 expression, CSR, plasma cell differentiation, generation of swMBCs, or overall maturation of the antibody response. B cell Tet2 deletion impairs maturation of the antibody response Prompted by selective TET2 expression in mouse and human B cells, we analyzed baseline B cell differentiation and antibody levels in AicdaCreTet2fl/fl mice, which delete Tet2 in activated and differentiating B cells. Non-intentionally immunized AicdaCreTet2fl/fl mice showed lower basal levels of IgG, IgA, and IgE but not IgM than AicdaCreTet2+/+ mice and fewer IgG3+, IgG1+, IgG2b+, and IgA+B cells, swMBCs, and plasmablasts in MLNs and Peyer’s patches, with no altered cell survival (Figures S2A–S2C). They also showed fewer IgM, IgG1, and IgA antibody-secreting cells (ASCs) in MLNs, fewer IgM+ and IgA+B cells in intestinal lamina propria than AicdaCreTet2+/+ mice, and reduced free and bacteria-bound IgD43 and IgA in feces (Figures S2D and S2E). To analyze the B cell TET2 role in the maturation of T-dependent antibody response, we immunized AicdaCreTet2fl/fl mice with NP-CGG. These mice made lower levels of total and NP4-specific IgG1 and IgA and fewer IgG1+ and IgA+ASCs than AicdaCreTet2+/+mice, but comparable levels of total and NP4-binding IgM and numbers of IgM+ASCs in spleen and BM (Figures 2A, S3A, and S3B). NP-CGG-immunized AicdaCreTet2fl/fl mice, however, displayed seemingly normal spleen GCs, GC B cells, proliferating B cells, and lack of apoptotic B cells as AicdaCreTet2+/+ mice, showing that Tet2 deletion did not impact GC formation, cell proliferation, or survival (Figures S3C–S3F). Nevertheless, AicdaCreTet2fl/fl mice showed no IgA+ or IgD+B cells in intestinal lamina propria and normal IgM+B cells in Peyer’s patches (Figure 2B). They showed reduced IgG1+GC B cells, spleen and BM plasma cells, spleen IgG1+B cells, including NP5-specific IgG1+B cells and NP5-specific IgG1+swMBCs, compared to AicdaCreTet2+/+ mice (Figures 2C and S3D–S3H). AicdaCreTet2fl/fl B cells expressed lower levels of Aicda and Prdm1 transcripts than AicdaCreTet2+/+B cells but comparable Irf4 and Bach2 transcripts (Figure 2D); enforced Aicda or Prdm1 expression in Tet2−/− B cells rescued CSR and plasma cell differentiation, while enforced Tet2 expression in Tet2+/+B cells boosted CSR and plasma cell differentiation (Figure 2E). AicdaCreTet2fl/fl mice IgG1+B cells accumulated somatic point-mutations (V1-72DJH-Cγ1 transcripts44), with high replacement (R) to silent (S) mutation ratios, at 5.9-fold lower frequency than AicdaCreTet2+/+ mice (Figure 2F). To analyze the B cell TET2 role in the maturation of the T-independent antibody response, we immunized AicdaCreTet2fl/fl mice with NP-LPS.45 These mice made lower levels of total and NP4-specific IgG3 and IgG2b, reflecting fewer class-switched IgG3+B cells, IgG2b+B cells, and NP5-specific IgG+B cells in MLNs than AicdaCreTet2+/+ mice but comparable levels of total and NP4-binding IgM (Figures S4A and S4B). NP-LPS-immunized AicdaCreTet2fl/fl mice also showed reduced NP5-specific IgG+swMBCs and plasmablasts in MLNs. Such AicdaCreTet2fl/fl mice accumulated somatic point-mutations, at high R:S ratios, in B cell V1-72DJH-Cγ3 and V1-72DJH-Cγ2b transcripts, at a 4.3- and 5.5-fold lower frequency than their AicdaCreTet2+/+ mouse counterparts (Figure S4C). To verify that the reduced B cell AID and BLIMP-1 expression in AicdaCreTet2fl/fl mouse B cells was due to impaired active DNA demethylation of Aicda and Prdm1 loci regulatory regions, we stimulated AicdaCreTet2fl/flB cells in a T-dependent (CD154 plus IL-4) or T-independent (LPS, LPS plus IL-4, LPS plus TGF-β, IL-4, IL-5, and anti-δ mAb/dex) fashion in vitro. These B cells reduced active DNA demethylation of Aicda and Prdm1 promoters and enhancers, and decreased Aicda/AID and Prdm1/BLIMP-1 expression (Figures 2G, 2H, and S5A). They also reduced CSR (lower levels of post-recombination Iμ–Cγ1, Iμ–Cγ3, and Iμ-Cε transcripts) in spite of normal proliferation, and reduced proportions of IgG3+, IgG1+, or IgA+B cells and plasma cells, and made lower IgG1, IgG3, or IgE levels (Figures S5B–S5D). Thus, differentiation stage-specific B cell-intrinsic Tet2 deletion decreases B cell Aicda and Prdm1 loci active DNA demethylation and transcription, AID and BLIMP-1 expression, SHM/CSR, and plasma cell differentiation. It reduces basal blood and gut-associated class-switched antibody levels and impairs the maturation of T-dependent and T-independent antibody responses, with no obvious impact on GC B cell proliferation or GC formation. B cell Ogt deletion impairs maturation of the antibody response The suggested binding of OGT to TET1 for TET1 catalytic function in mouse and zebrafish ESCs26 prompted us to hypothesize that TET2 recruits OGT, the sole known enzyme that mediates protein O-GlcNAcylation, for B cell TET2-mediated AID and BLIMP-1 expression, peripheral B cell differentiation, and maturation of antibody response. We immunized AicdaCreOgtfl/fl mice, which delete Ogt in activated and differentiating B cells, with NP-CGG. These mice developed spleen GCs normal in size and morphology (Figure 3A). They also showed GC B cells in spleen, Peyer’s patches, and MLNs comparable to AicdaCreOgt+/+ mice, albeit reducing IgG1+ GC B cells, class-switched IgG1+, IgA+B cells, NP5-specific IgG1+B cells, and plasma cells (Figures 3A and 3B). They also showed lower levels of NP4-specific IgG1 and IgA, as well as fewer IgD and IgA in intestinal lamina propria, but normal NP4-binding IgM (Figures 3C and 3D). Further, AicdaCreOgtfl/fl mice accumulated somatic point-mutations, with high R:S mutation ratios, in B cell V1-72DJH-Cγ1 transcripts, at 1.7-fold lower frequency than AicdaCreOgt+/+ mice (Figure 3E). The relevance of OGT to overall TET2 function was emphasized by the negative impact of B cell Ogt deletion on AID and BLIMP-1 expression, CSR, and plasma cell differentiation. Likewise, AicdaCreOgtfl/flB cells stimulated in vitro with LPS, LPS plus IL-4, or LPS plus TGF-β, IL-4, IL-5, and anti-δ mAb/dex decreased intracellular AID and BLIMP-1 expression, resulting in reduced CSR to IgG3, IgG1, and IgA as well as plasma cell differentiation (Figure 4A). Thus, differentiation stage-specific B cell-intrinsic Ogt deletion impairs AID and Blimp-1 expression, SHM/CSR, and plasma cell differentiation, thereby dampening the maturation of antigen-specific antibody response. B cell TET2 recruits OGT for its O-GlcNAcylation and catalytic DNA 5mC to 5hmC oxidation Next, we set out to show that reduced AID and BLIMP-1 expression in activated and differentiating AicdaCreOgtfl/flB cells is due to loss of TET2 O-GlcNAcylation and impaired TET2 catalytic DNA 5mC to 5hmC oxidation of Aicda and Prdm1 loci. We stimulated AicdaCreOgtfl/fl and AicdaCreTet2fl/flB cells with LPS plus IL-4 and cultured them in vitro. OGT deletion in AicdaCreOgtfl/flB cells showed impaired 5mC to 5hmC oxidations of Aicda and Prdm1 promoter and enhancer regions to an extent comparable to that of TET2 deletion in AicdaCreTet2fl/flB cells (Figure 4B). In LPS plus IL-4-stimulated AicdaCreOgtfl/flB cells, OGT deletion denied TET2 O-GlcNAcylation and reduced overall TET2 levels (Figure 4C). TET2 was O-GlcNAcylated through 96 h post-activation (Figures 4D and S6A). TET2 O-GlcNAcylation was further supported by the identification of potential O-GlcNAcylated sites throughout mouse and human TET2, as predicted using YinOYang and O-GlcNAcPRED-DL,47 an algorithm and a deep learning tool for protein O-GlcNAcylation (Figures 4E and S6B; Table S1). In mouse TET2, we identified 38 Ser and Thr residues (out of 320 total Ser/Thr) as potential O-GlcNAcylation sites using a 0.8 threshold, higher than the 0.5 and 0.6 thresholds generally used as thresholds for predicted protein PTM, including O-GlcNA-cylation.46–48 In human TET2, we identified 46 Ser and Thr residues (out of 323 total Ser/Thr) as O-GlcNAcylation sites, at > 0.8 threshold. Further, a TET2:PROSER1:OGT49 interaction was identified by native co-IP in LPS plus IL-4-stimulated mouse naïve IgM+IgD+B cells. In these B cell nuclei, TET2 interacted with OGT, CREBBP, and PROSER1, while it did not, as expected, with HDAC1.21 Confirmation of nuclear TET2:PROSER1:OGT49 interaction came from anti-OGT IP, which brought down TET2 and PROSER1 but not CREBBP or HDAC1 (Figure 4F). Thus, B cell TET2 recruits OGT for its O-GlcNAcylation, which supports TET2’s catalytic function. B cell TET2 recruits OGT for O-GlcNAcylation of Aicda and Prdm1 loci chromatin H2B-S112 Having determined that TET2 recruits OGT in B cell nuclei for its O-GlcNAcylation and catalytic function, we further addressed the potential role of TET2 in recruiting OGT to Aicda and Prdm1 loci for these loci’s chromatin H2B-S112 O-GlcNAcylation.50 We stimulated AicdaCreTet2fl/flB and AicdaCreTet2+/+B cells with LPS plus IL-4 and culture them in vitro. In AicdaCreTet2+/+B cells, OGT was detected at Aicda and Prdm1 promoters and enhancers (Figure 4G), at which DNA regions TET2 was also detected (Figure S7A). In such B cells, OGT catalyzed O-GlcNAcylation of Aicda and Prdm1 loci regulatory regions chromatin H2B-S112, a previously unrecognized histone modification in the B cell epigenome. The instrumental role of TET2 in OGT recruitment and subsequent H2B-S112 O-GlcNAcylation was further supported by failure to detect OGT and chromatin H2B-S112 O-GlcNAcylation in Aicda and Prdm1 regulatory regions in AicdaCreTet2fl/flB cells. Thus, B cell TET2 recruits OGT for its O-GlcNAcylation and active DNA demethylation (catalytic function) as well as for O-GlcNAcylation of chromatin H2B-S112 at Aicda and Prdm1 regulatory regions (non-catalytic, adaptor function). 3D modeling of TET2-OGT-DNA complex at Aicda Area IV or Prdm1 Area II Having shown TET2-OGT catalytic and non-catalytic functions in B cell Aicda and Prdm1 loci expression, we set up to gain insight into the TET2-OGT-DNA complex. We first determined the Tet2 deletion impact on dynamic 5mC to 5hmC oxidation at the same C nucleotide residues of Aicda and Prdm1 loci. As compared to Tet2+/+B cells, Tet2−/− B cells stimulated with LPS or CD154 plus IL-4 decreased genome-wide DNA 5hmC and increased 5mC, as revealed by anti-5mC and anti-5hmC Ab immune dot-blots and MeDIP-Seq and hMeDIP-Seq (Figures S7B and S7C). By detailing 5mC and 5hmC distribution and levels, we constructed a map of Aicda and Prdm1 loci and annotated Areas of active DNA demethylation, denoted as Areas I–V (Figure S7D). Aicda Area I corresponded to previously reported region 1 (Aicda promoter), while Areas II, III and IV included regions 2, 3, and 4.38 Area IV also included TetE1 and Area V TetE2.33 Prdm1 Areas I and IV included promoter regions 1A and 1B.39 Prdm1 Areas II, III, and V, not previously recognized as regulatory regions, comprised low frequency of CpG islands,51 thereby identifying potential enhancer regions. Because no crystallography structure of TET2-OGT or TET2-OGT-DNA complex is currently available, we used AlphaFold 3.0, an advanced deep learning platform with updated diffusion architecture and expanded capacity of predicting biomolecular ensembles involving proteins and nucleic acids,52 to 3D model the TET2-OGT-DNA complex at Aicda Area IV and Prdm1 Area II. Mouse TET2 isoform 2, including the critical Cys-N, Cys-C, DSBH (double-stranded β-helix), low complexity insert, and DSBH 1042–1920 aa sequence,53 bound to Aicda locus Area IV and Prdm1 locus Area II. In both Aicda and Prdm1 loci, TET2 accommodated double-stranded DNA through its conserved signature H1303xD1305 motif in a catalytic cavity (Figure 5A). The angle of such TET2 catalytic cavity varied depending on different TET2 isoforms and DNA structures (Aicda Area IV or Prdm1 Area II) (Figures S8 and S9), leading to different TET2-OGT-DNA complexes. In both Aicda Area IV and Prdm1 Area II DNA, TET2 extensively interacts with OGT through its C-terminal-end, thereby validating our structural inference extrapolated from the suggested TET1-OGT interaction in ESCs.26 Deletion of TET2 OGT-binding C-terminal end impairs CSR and plasma cell differentiation to a degree comparable to that of mutation of TET2 catalytic H1303YxD1305A motif As suggested by our TET2-OGT-DNA 3D model involving Aicda and Prdm1 loci regulatory regions, TET2 interacts with OGT through its C-terminal-end. To analyze the role of TET2 C-terminal-end in TET2 OGT recruitment and TET2-modulated B cell CSR and plasma cell differentiation, we generated a mouse Tet2WT isoform 2 construct encoding conserved signature catalytic HxD motif and the whole C-terminal-end (E990-V1920),54,55 a TET2 isoform 2 catalytic Tet2HxD-mut mutant (H1303YxD1305A) construct defective in DNA demethylation activity (tested in neoplastic kidney 293T cells54), a TET2 isoform 2 C-terminal-end deletion Tet2OGT-Δmut mutant (G1876-V1920) construct encoding a TET2 protein putatively incapable of recruiting OGT, and a dual Tet2 HxD-mut/OGT-Δmut mutant construct encoding a TET2 devoid of both catalytic HxD site activity and OGT-binding adaptor function (Figure 5B). TET2 C-terminal-end deletion was based on findings in ESCs that TET1 binds OGT through its C-terminal end, which is conserved in mouse and human TET1, TET2, and TET326; pMIG-Empty, Tet2WT, Tet2HxD-mut, Tet2OGT-Δmut, and Tet2HxD-mut/OGT-Δmut constructs were used to transduce Tet2−/−B cells, which were cultured after induction by LPS plus IL-4. Tet2−/− B cells transduced by Tet2WT, Tet2HxD-mut, Tet2OGT-Δmut, and Tet2HxD-mut/OGT-Δmut constructs exhibit a comparable level of TET2 expression, while Tet2−/− B cells transduced by the pMIG-Empty construct did not express TET2 (Figure 5C). In induced Tet2−/− B cells, enforced Tet2HxD-mut expression rescued CSR to IgG1 and plasma cell differentiation by 69.7% and 52.3%, respectively, of enforced Tet2WT expression (Figure 5C), to levels comparable to those of Tet2HxD-mut/HxD-mutB cells (Figures S6C and S6D). Similarly, enforced expression of Tet2OGT-Δmut, which included an unmutated conserved HxD motif, rescued CSR to IgG1 and plasma cell differentiation by approximately 53.9% and 41.0%, respectively, as inefficiently as Tet2HxD-mut catalytic HxD site mutant. Finally, enforced Tet2HxD-mut/OGT-Δmut expression in Tet2−/− B cells resulted in more limited IgG1 CSR rescue but plasma cell differentiation comparable to individual Tet2HxD-mut or Tet2OGT-Δmut mutants. Small molecule TET2 catalytic inhibitor TET-IN-C35 and/or OGT inhibitor ST 045849 impair B cell AICDA and PRDM1 5hmC oxidation, reduce AID and BLIMP-1 expression, CSR, and plasma cell differentiation To extend the analysis of TET2-OGT interaction to human B cells, we stimulated naïve huIgM+IgD+B cells with CpG plus huIL-2, huIL-4, and huIL-21 and cultured them in the presence of TET2 catalytic inhibitor TET-IN-C35,40 OGT inhibitor ST 04584941 or both small molecule inhibitors. Individual TET2 and OGT inhibitors impaired huIgG CSR and plasma cell differentiation to comparable degrees, reflecting comparable reductions of TET2-catalyzed 5mC to 5hmC oxidation of AICDA and PRDM1 loci regulatory regions, resulting in similar reductions in AICDA/AID and PRDM1/BLIMP-1 expression (Figures S10A–S10C). In addition to TET2-catalyzed active DNA demethylation, the OGT inhibitor impaired AICDA and PRDM1 loci H2B-S112 O-GlcNAcylation. Combined TET2 and OGT inhibitors reduced 5mC to 5hmC oxidation and H2B-S112 O-GlcNAcylation, expression of AICDA/AID and PRDM1/BLIMP-1, CSR, and plasma cell differentiation to degrees comparable to that of either TET2 or OGT individual inhibitor (Figures S10A–S10D). By contrast, in similar human B cells, the TET2 inhibitor resulted in deficient active DNA demethylation but effective chromatin H2B-S112 O-GlcNAcylation of AICDA and PRDM1 promoter and enhancer regions. TET-IN-C35 and ST 045849 inhibition of TET2-catalyzed 5mC to 5hmC oxidation of Aicda and Prdm1 loci, Aicda/AID and Prdm1/BLIMP-1 expression, CSR, and plasma cell differentiation was reproduced in mouse B cells stimulated by LPS plus IL-4 (Figures S11A–S11C). Thus, specific small molecule inhibition of TET2 catalytic activity and/or OGT function impairs 5mC to 5hmC oxidation of AICDA/Aicda and PRDM1/Prdm1 loci regulatory regions to dampen human and mouse B cell AID and BLIMP-1 expression, CSR, and plasma cell differentiation. Fumarate inhibits B cell TET2 to impair AID and BLIMP-1 expression, SHM/CSR, plasma cell and swMBC differentiation, and maturation of human and mouse T-dependent and T-independent antibody responses TET2 catalytic activity requires Fe++ and α-KG co-factors.23,56 We addressed the ability of fumarate, a TCA cycle metabolite that hampers α-KG binding to the TET2 catalytic domain,57 to impair the maturation of human antibody response. Humanized THX mice58 were immunized with NP-CGG and fed dimethyl fumarate (DMF) in water. Upon oral administration, DMF is hydrolyzed by esterase to monomethyl fumarate, which is further converted to fumarate.59 In these THX mice, DMF impaired intracellular human B cell TET2, AID and BLIMP-1 expression and reduced total and NP4-specific huIgG1 and huIgA but not NP4-binding huIgM as compared to THX mice fed nil-water (Figures 6A and 6B), a reflection of reduced total and NP5-specific huIgG1+ and huIgA+B cells, total and class-switched human plasmablasts as well as total and swMBCs60 in spleen and BM (Figures 6C and 6D). DMF-fed THX mice huIgG1+B cells accumulated somatic point-mutations at a lower frequency and lower R:S ratios than nil-water fed THX mice (Figure 6E). DMF inhibition of B cell TET2-mediated AID and BLIMP-1 expression, SHM/CSR, plasma cell differentiation, and generation of swMBCs also extended to maturation of T-independent antibody response in NP-LPS-immunized Tcrβ−/− Tcrδ−/− mice,45 in which any contribution of TET2-expressing T cells was excluded (Figures S12A–S12D). Thus, metabolic inhibition of TET2 by fumarate impairs maturation of human and mouse T-dependent and T-independent antibody responses by dampening B cell AID and BLIMP-1 expression, SHM/CSR, and differentiation of plasma cells and swMBCs. Vitamin C boosts B cell TET2 catalytic activity, Aicda/AICDA and Prdm1/PRDM1 expression, SHM/CSR, plasma cell differentiation, and T cell-independent anti-NP-LPS antibody response in Tcrβ−/− Tcrδ−/− mice Vitamin C would activate TETs by replenishing the cellular Fe++ pool through Fe+++ reduction, as TET proteins utilize a Fe4+-oxo intermediate for catalytic activity and stepwise 5mC oxidation.23 Combined Tet2/Tet3 deletion in CD19CreTet2fl/flTet3fl/fl or Cr2CreTet2fl/flTet3fl/flB cells has suggested a TET2/TET3-dependent increase in BLIMP-1 and plasma cell differentiation by vitamin C in vitro.61,62 We furthered the analysis of vitamin C-induced B cell TET2 activity and BLIMP-1 expression in vivo and extended it to include AID expression. Further, to rule out any potential role of T cells that can also express TET2,21 we immunized Tcrβ−/− Tcrδ−/− mice with (T-independent) NP-LPS45 and treated them with vitamin C. These Tcrβ−/− Tcrδ−/− mice responded with elevated NP4- and LPS-specific IgG3 and IgG2b, but not NP4- or LPS-binding IgM, as compared to non-vitamin C-treated mice. In vitamin C-treated Tcrβ−/− Tcrδ−/− mice, increased levels of NP4-specific IgG3 and IgG2b antibodies reflected increased numbers of NP5-specific IgG3+ and IgG2b+ASCs in spleen and BM, as well as increased total and NP5-specific IgG3+ and IgG2b+B cells, and increased plasma cells in spleen and MLNs, in spite of comparable proportions of GC B cells (Figures 7A and 7B). NP-LPS-immunized Tcrβ−/− Tcrδ−/− mice IgG3+ and IgG2b+B cells increased AID and BLIMP-1, but not, as expected, TET2 expression and accumulated somatic point-mutations (2.19 × 10−2 and 1.35 × 10−2 change/base with high R:S mutation ratios) at a 3.4- and 1.9 fold greater frequency than IgG3+ and IgG2b+B cells of non-vitamin C-treated mice (Figures 7C, 7D, and S12E). B cell-intrinsic TET2 metabolic boost by vitamin C was reproduced in purified mouse naÿve IgM+IgD+B cells stimulated by LPS plus IL-4. These B cells increased Aicda and Prdm1 expression, CSR to IgG1, and plasma cell differentiation (Figure 7E), reflecting increased TET2 recruitment to Aicda and Prdm1 loci, enhanced DNA 5mC to 5hmC oxidation, and increased chromatin H2B-S112 O-GlcNAcylation of these gene loci. A similar vitamin C activity extended to human B cells. Naÿve huIg-M+IgD+B cells stimulated by CpG plus huIL-2, huIL-4, and huIL-21 and cultured in the presence of vitamin C increased 5mC to 5hmC oxidation, expression of AICDA and PRDM1, boosted CSR and plasma cell differentiation, and increased proportion of MBCs (Figure 6F). Thus, metabolic activation of TET2 by vitamin C enhances Aicda/AICDA and Prdm1/PRDM1 loci OGT-dependent O-GlcNAcylation and TET2-catalyzed active DNA demethylation for these loci transcription, AID and BLIMP-1 expression, SHM/CSR, and plasma cell differentiation in a B cell-intrinsic, T cell-independent fashion. DISCUSSION Our findings show that TETs are not detectable in resting naÿve B cells and only TET2 is induced in mouse and human B cells, in a differentiation-stage-specific fashion. B cell differentiation stage-specific TET2 expression was reflected in vivo in GC B cells and plasmablasts in spleen and MLNs of NP-CGG-immunized C57BL/6 mice, NP-LPS-immunized Tcrβ−/− Tcrδ−/− mice, NP-CGG-immunized humanized THX mice, and human tonsils. In B cells undergoing GC-like differentiation, TET2 was expressed, phosphorylated, acetylated, and O-GlcNAcylated for Aicda/AID and Prdm1/BLIMP-1 expression. O-GlcNAcylation of B cell TET2, as mediated by TET2-recruited OGT, would be critical for this epigenetic factor’s catalytic activity at Aicda/AICDA and Prdm1/PRDM1 loci regulatory regions. OGT, as recruited by TET2, also serves as a histone modifier (H2B-S112 O-GlcNAcylation) to facilitate chormatin accessibility and transcription of such loci. TET2 recruits OGT in the B cell nucleus to promote AID and BLIMP-1 expression, SHM/CSR, and plasma cell differentiation for maturation of mouse and human T-dependent and T-independent antibody responses. TET proteins, particularly TET1, play an important role in DNA demethylation during ESC differentiation.25 TET2 and TET3 exert functions in HSCs and immune cells, particularly T cells, affecting active DNA demethylation in physiological and pathological processes.20,21,63–66 The different impacts of combined Tet2/Tet3 deletion versus individual Tet2 or Tet3 deletion on iNKT cell differentiation led to the postulation of a reciprocal redundant activity of these two epigenetic modifiers.21,32 In B cells, only the impact of combined Tet2/Tet3 deletion, not individual Tet2 or Tet3 deletion, on NP-binding B cells, as in NP-OVA-immunized CreERT2Tet2fl/flTet3fl/flRosa26YFPLSL mice, was explored.33 Our findings that B cell TET2 (AicdaCreTet2fl/fl), not TET3 (AicdaCreTet3fl/fl), deletion impairs maturation of T-dependent and T-independent antibody responses support the contention that either TET3 plays no role in peripheral B cell differentiation or TET3 does, but its absence (Tet3 ablation) is fully compensated by TET2. In contrast, TET3 cannot compensate for TET2 deletion in B cells. TET2 role in B cell differentiation is further supported by full rescue of CSR to IgG1 by enforced TET2 expression in combined Tet2/Tet3 deletion CreERT2Tet2fl/flTet3fl/flRosa26YFPLSLB cells.33 Mouse and human TET2 differ from TET1 and TET3 in that it lost its CXXC zinc finger domain, which is retained in TET1 and TET3 N-end, likely due to a chromosomal inversion event that gave rise to the discrete Tet2/TET2 CXXC4 or IDAX gene23; IDAX CXXC domain binds to unmethylated CpG dinucleotides in DNA, to which it recruits TET2.55 In hematopoietic cells’ malignant transformation, including acute myeloid lymphoma (AML), diffuse large B cell lymphoma (DLBCL, a GC B cell malignancy), peripheral T cell lymphoma (PTCL), and angioimmunoblastic T cell leukemia (AITL), mutations leading to loss-of-function are more frequent in coding regions of TET2 than TET1 or TET3.21,36,67 Interestingly, TET2 loss-of-function has also been observed in the absence of TET2 mutations, suggesting a critical yet unclear mechanism, possibly involving protein PTM, in modulating this epigenetic modifier’s function in lymphoid cells malignant transformation.20 OGT is the only known enzyme that can attach a GlcNAc sugar to Ser and Thr in cytoplasmic, mitochondrial, and nuclear proteins.27,28 As a metabolic sensor, OGT relies on the concentration of cellular UDP-GlcNAc, the end-product of the hexosamine biosynthetic pathway, to catalyze protein O-GlcNAcylation.30 Nucleocytoplasmic OGT consists of an N-terminal protein-protein contact domain and a C-terminal catalytic glycotransferase domain.41 O-GlcNAcylation occurs reciprocally or sequentially with phosphorylation in the same or neighboring residues, potentiates protein acetylation, and blocks ubiquitin-mediated protein degradation.28,68–74 By O-GlcNAcylating TET2, OGT would enable TET2-catalyzed active DNA demethylation in mouse and human B cell Aicda and Prdm1 loci regulatory regions. Together with phosphorylation and acetylation, O-GlcNAcylation would contribute to stabilizing TET2 for catalytic activity in differentiating peripheral B cells. In mouse TET2, we identified 38 O-GlcNAcylation sites, which include all the 21 O-GlcNAcylation sites previously detected by mass spectrometry in mouse TET2,75 supporting the validity of our analysis. In human TET2, we identified 46 high-confidence O-GlcNAcylation sites. Notably, our predicted O-GlcNAcylation sites in both mouse and human TET2 were enriched in the N terminus and low-complexity insertion domain, regions known to preferentially harbor PTMs, as assessed by spectrometry.75 Although in vitro DNA demethylation assay for O-GlcNAcylated TET2 was not included in our study, loss of TET2 OGT-recruitment adaptor function, as shown in Tet2OGT-Δmut (OGT-binding) deletion mutant, impaired TET2’s ability to restore CSR and plasma cell differentiation to an extent comparable to that of Tet2HxD-mut (catalytic) mutant. Further, in AicdaCreOgtfl/flB cells, loss of TET2 O-GlcNAcylation and reduced TET2 overall protein impaired TET2-catalyzed 5mC to 5hmC oxidation in Aicda and Prdm1 loci promoters and enhancers to levels comparable to those of AicdaCreTet2fl/flB cells. Accordingly, Ogt deletion in AicdaCreOgtfl/fl mice impaired maturation of the antibody response, including B cell AID and BLIMP-1 expression, SHM/CSR, and plasma cell differentiation, by an order of magnitude comparable to that of AicdaCreTet2fl/fl mice. O-GlcNAcylation contribution to TET2-catalyzed active DNA demethylation in mouse and human B cells, as shown here, mirrors O-GlcNAcylation contribution to TET1-catalyzed active DNA demethylation, as evidenced by mouse and zebrafish ESCs by Ogt deletion or TET1 T535 (O-GlcNAcylation site) mutation, which decreases TET1 level and function despite normal Tet1 expression.76 Interestingly, while TET2 mediates Aicda/AICDA and Prdm1/PRDM1 expression in mouse and human B cells, TET1 can repress genes in mouse ESCs by interacting with repression-associated chromatin factors Sin3A and CHD4.76 In such ESCs, OGT may prevent TET1-dependent expression of transposable elements and protect heterochromatin integrity,77 as occurring in cellular neoplastic transformation and aging, suggesting that OGT-mediated O-GlcNAcylation of TETs may lead to TET-mediated gene activation or repression depending on cell type or context (heterochromatin versus euchromatin).77 TET2 recruits OGT through direct and/or indirect interaction through PROSER1 bridging, as evidenced by native protein co-IPs of nuclear extracts from B cells undergoing CSR and plasma cell differentiation. OGT recruitment, as dependent on TET2, resulted in O-GlcNAcylation of not only TET2 but also chromatin H2B-S11250 of Aicda and Prdm1 loci promoter and enhancer regions. Accordingly, Tet2 deletion, as in AicdaCreTet2fl/flB cells, not only aborted OGT recruitment to Aicda and Prdm1 loci regulatory regions but also reduced chromatin H2B-S112 O-GlcNAcylation of such loci, further outlining a critical and previously unrecognized TET2 non-catalytic OGT recruitment adaptor function in physiologically differentiating B cells. In human promyelocytic leukemia cells, H2B-S112 O-GlcNAcylation has been shown to promote H2B monoubiquitination at Lys120, leading to transcriptional activation.68 Importantly, OGT-enabled H2B-S112 O-GlcNAcylation would contribute, together with TET2-catalyzed active DNA demethylation, to Aicda and Prdm1 loci epigenomic changes and, therefore, expression, as also evidenced in mouse and human B cells treated with TET2 and/or OGT catalytic inhibitor. The relative contribution of OGT-mediated chromatin H2B-S112 O-GlcNAcylation and TET2 active DNA demethylation to Aicda and Prdm1 loci epigenome and, therefore, transcription in GC-differentiating B cells would deserve further investigation. Interestingly, OGT has been suggested to O-GlcNAcylate BCR signaling transducer Lyn for B cell homeostasis, and Ogt deletion by CD19Cre promoted Pre-B cell apoptosis, thereby reducing follicular B cell numbers.78 By contrast, NP-CGG-immunized AicdaCreOgtfl/fl mice, which delete Ogt in activated and differentiating B cells, maintained normal proportions of GC B cells in spleen, MLNs, and Peyer’s patches, further indicating that Ogt deletion reduces SHM/CSR and plasma cell differentiation by negating B cell nuclear O-GlcNAcylation of TET2 and chromatin H2B-S112, without impacting GC B cell survival. While excluding the effect of OGT deletion on GC B cell survival and proliferation, our experiments focused on TET2 and did not address the impact of OGT loss on global protein modification in GC B cells. Further supporting the contribution of O-GlcNAcylation to TET2-catalyzed active DNA demethylation, OGT ST 045849 small molecule inhibitor impaired 5mC to 5hmC oxidation in mouse and human B cells Aicda/AICDA and Prdm1/PRDM1 loci regulatory regions as efficiently as TET catalytic TET-IN-C35 small molecule inhibitor. Combined acute Tet2/Tet3 deletion in all cell types, as in CreERT2Tet2fl/flTet3fl/flRosa26YFPLSL mice, decreased Aicda locus super enhancer DNA 5hmC deposition,33 and pan-B cell Tet2 deletion, as in VavCreTet2fl/fl mice, was associated with decreased 5hmC deposition in Prdm1 but not in Aicda locus.36,37 As shown here, differentiating B cell-specific Tet2 deletion, as in AicdaCreTet2fl/fl mice and AicdaCreTet2fl/fl B cell in vitro, impaired 5mC to 5hmC oxidation in both Aicda and Prdm1 loci promoters and enhancers. AlphaFold 3.0 modeled TET2 isoform 2-OGT-DNA complex involving Aicda Area IV and Prdm1 Area II, as newly identified by our integrative map of our current and others’ reported data,33,38,39 furthers the visualization of crystallized a TET2-DNA structure53 and expands it to include OGT. It revealed a shallow TET2 cavity53 that centers on the signature conserved HxD motif to accommodate Aicda Area IV and Prdm1 Area II. Such TET2 cavity hints at the catalytic mechanism by which this mononuclear non-heme iron oxygenase actively demethylates Aicda and Prdm1 DNA after having locked onto it. 3D modeling of TET2-OGT-DNA complex, as further involving TET2 isoforms 1 and X2, also revealed a conserved spatial structure and identical amino acid residues mediating C-terminal-end interaction with OGT, further supporting the functional relevance of this TET2-OGT-binding interface. TET2 isoforms 1 and X2 adopt a conformation allowing for a catalytic cavity similar to that of TET2 isoform 2 for Aicda and Prdm1 active demethylation. Finally, TET2-OGT-DNA complex AlphaFold 3.0 modeling shows that TET2 C-terminal-end residues (in all three TET2 isoforms) exhibit extensive contact with OGT, similar to TET1-OGT interaction in zebrafish ESCs,26 further emphasizing the role of TET C-terminal end in recruiting OGT. As we show, the catalytic activity of TET2 at AICDA and PRDM1 loci regulatory regions requires O-GlcNAcylation of this active DNA demethylator, as mediated by metabolic sensor OGT. As we further show, TET2 directly takes cues from the metabolic environment to modulate B cell AID and BLIMP-1 expression, as it relies on co-factors Fe++ and α-KG21 for its catalytic activity. Replenished cellular Fe++ pool, as mediated by vitamin C reduction of Fe+++, leads to enhanced TETs catalytic activity.57,61,79 Our findings further the scope of plasma cell differentiation studies in vitro61,62 and extend them to address the vitamin C impact on SHM/CSR and plasma cell differentiation in vivo. In NP-LPS-immunized Tcrβ−/− Tcrδ−/− mice fed a regular diet, they revealed a T cell-independent and B cell-intrinsic vitamin C boost of class-switched, high-affinity antibodies to NP and to LPS—Gulo−/− mice, deficient in L-gulono-γ-lactone oxidase, were not chosen for this study to avoid potential confounding effects related to vitamin C deficiency, as such mice develop human scurvy-like disease.80–82 Our findings also show that vitamin C increases B cell nuclear 5mC to 5hmC oxidation for active DNA demethylation, as well as OGT-mediated AICDA and PRDM1 loci H2B-S112 O-GlcNAcylation to increase mouse and human AID and BLIMP-1 expression. Finally, they determined that vitamin C also increases TET2 deposition at AICDA and PRDM1 loci, likely by increasing Fe++ and Fe++-mediated TET2 stabilization at these loci regulatory regions. As both components of the TCA cycle, TET2 co-factor α-KG and its antagonist fumarate, a potent TETs catalytic inhibitor,57,79 also underline the metabolic regulation of this DNA active demethylator. DMF dampened maturation of T-dependent and T-independent antibody responses by impairing B cell SHM/CSR, plasma cell differentiation, and generation of swMBCs in humanized THX mice and Tcrβ−/− Tcrδ−/− mice, in which B cell TET2 expression was reduced together with AID and BLIMP-1. The reduced numbers of swMBCs could merely reflect reduced CSR and/or impaired overall generation of MBCs. Our findings that B cell TET2 recruits OGT to play a central role in maturation of the antibody response in mouse and human B cells provide important insight into peripheral physiological B cell differentiation and may open new avenues of investigation of dysfunctional TET2 conditions, as in neoplastic B cell transformation. In vivo B cell TET2 modulation by vitamin C and fumarate further highlights the importance of the metabolic context in the physiological regulation of TET2 epigenetic modifier. As a TET2 activator, vitamin C could find a therapeutic application in conditions of B cell TET2 functional deficiency, such as occurring in certain patients with DLBCL or other B cell lymphomas.21,23,83 DMF, a therapeutic for relapsing multiple sclerosis and psoriasis,59 could be used to dampen B cell hyperreactivity, as occurring in autoantibody-mediated systemic autoimmunity of systemic lupus erythematosus,84,85 a condition predominantly affecting females over males (10:1). In this light, because Ogt/OGT is an X-linked gene, OGT-enabled TET2 catalytic activity for B cell differentiation would be intricately linked to X chromosome dosage effect and sex-biased susceptibility. Importantly, as shown here, in addition to OGT, TET2 catalytic activity is controlled by the physiologic TCA metabolite fumarate. Limitations of the study Our experiments provide a string of evidence that deletion of OGT or inhibition of OGT by a small molecule leads to defective active DNA demethylation of B cell Aicda and Prdm1 loci regulatory regions, as well as reduced AID and BLIMP-1 expression, CSR, and plasma cell differentiation to a comparable level to what resulted from TET2 deletion. They, however, do not quantify the relative contribution of H2B-S112 O-GlcNAcylation to overall B cell AID and BLIMP-1 expression. This could be addressed by further quantitative, sequential biochemical experiments. RESOURCE AVAILABILITY Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Paolo Casali (pcasali@uthscsa.edu). Materials availability This study did not generate unique reagents. STAR★METHODS EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS Study approval Buffy coat was obtained upon informed consent from donors, as per protocol of the South Texas Blood and Tissue Center, San Antonio, TX. Human umbilical cord blood was collected from full-term, normally developed male and female newborns from healthy puerperae at the Department of Obstetrics and Gynecology, The University of Texas Long School of Medicine, The University of Texas at San Antonio Health Science Center, and obtained upon informed consent, as per IRB Protocol STUDY00000707. All experiments involving mice were performed in compliance with the animal protocol approved by The University of Texas at San Antonio Health Science Center IACUC Protocol 20200019AR. Mice C57BL/6 (RRID: IMSR_JAX:000664), Tcrβ−/− Tcrδ−/− (B6.129P– Tcrβ−/− Tcrδ−/−; RRID: IMSR_JAX:002121), Tet2fl/fl (B6;129S-Tet2t-m1.1Iaai/J, RRID: IMSR_JAX:017573), Tet3fl/fl (B6(Cg)-Tet3tm1.1Rao/J, RRID: IMSR_JAX:031015), Ogtfl/fl (B6.129-Ogttm1Gwh/J, RRID: IMSR_JAX:004860), BAC transgenic AicdaCre (B6;FVB-Tg(Aicda-Cre)1Rcas/J, RRID: IMSR_JAX:018422) and Tet2−/− (B6(Cg)-Tet2tm1.2Rao/J, RRID: IMSR_JAX:023359) mice were purchased from The Jackson Laboratory. Tet2fl/fl, Tet3fl/fl and Ogtfl/fl mice were crossed with AicdaCre mice to generate activated B cell-specific knockout AicdaCreTet2fl/fl, AicdaCreTet3fl/fl and AicdaCreOgtfl/fl mice, respectively. AicdaCreTet2+/+, AicdaCreTet3+/+ and AicdaCreOgt+/+ were used as controls. BAC transgenic AicdaCre mice, as constructed by introducing a bacterial Cre recombinase gene in lieu of Aicda exon 1 in a supplementary Aicda locus, under control of Aicda promoter/enhancers within the BAC transgene, show full expression of Aicda.16,86 These mice carry two functional endogenous Aicda loci. Tet2HxD-mut/HxD-mut mice were provided by Dr. Minjiang Xu from The University of Texas at San Antonio Health Science Center. Humanized THX mice were constructed by grafting non-γ-irradiated, genetically myeloablated NBSGW (NOD.Cg-KitW−41JTyr+PrkdcscidIl2rgtm1Wjl/ThomJ, RRID: IMSR_JAX:026622) mouse neonates (within 72 h of birth) intracardially (left ventricle) with human cord blood huCD34+ cells and subsequent hormonal conditioning, as we described.58 As we described in our original paper,58 our THX mice reconstitute a functional human lymphoid and myeloid immune system, including marginal zone B cells, GC B cells, follicular helper T cells and neutrophils, and develop well-formed lymph nodes and intestinal lymphoid tissues, including Peyer’s patches, as well as human thymic epithelial cells. These mice harbor diverse human B cell and T cell antigen receptor repertoires and are capable of mounting mature T cell-dependent and T cell-independent antibody responses, including SHM, CSR, plasma cell and memory B cell differentiation. In all experiments, male and female C57BL/6, Tcrβ−/− Tcrδ−/−, Tet2−/−, AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+, AicdaCreOgtfl/fl and Tet2HxD-mut/HxD-mut mice were 8–12 weeks of age and used in virtually equal proportions. Male and female humanized THX mice were 20–22 weeks of age. All mice were housed and bred in our pathogen-free vivarium at The University of Texas at San Antonio Health Science Center. Housing rooms were maintained at a 14-h light/10-h dark cycle and controlled temperature of approximately 22°C–23°C with 40–60% humidity. Food (Teklad LM-485 Sterilizable Mouse/Rat diet, 7912, Inotiv) and water were sterilized. METHOD DETAILS Mouse and human specific antibody responses The maturation of T-dependent and/or T-independent antibody responses was analyzed in AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+, AicdaCreOgtfl/fl and Tcrβ−/− Tcrδ−/− mice (8–12 weeks of age), as well as humanized THX mice (20–22 weeks of age). AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+ and AicdaCreOgtfl/fl mice were injected intraperitoneally (i.p.) with 4-hydroxy-3-nitrophenylacetyl (NP) conjugated to chicken gamma globulin (NP16-CGG, 16 NP molecules conjugated per one CGG molecule; N-5055C-5, Biosearch Technologies) on day 0 (100 μg in 100 μL alum, Inject Alum Adjuvant, 77161, Thermo Fisher Scientific), boosted on day 14 (100 μg in 100 μL PBS) and euthanized on day 28. AicdaCreTet2+/+ and AicdaCreTet2fl/fl mice were injected i.p. with NP0.4-LPS (NP conjugated to E. coli O55:B5 LPS, 2 NP molecules conjugated per 5 LPS molecules, Biosearch Technologies) on day 0 (20 μg in 100 μL PBS), boosted on day 21 (20 μg in 100 μL PBS) and euthanized on day 28. Ten humanized THX mice with a fully developed human immune system (human CD45+ cells consistently ≥90%), fulfilling the criteria and experimental timeline established in our THX studies,58 were used in the Nil/DMF treatment experiments. All 10 THX mice were comparable in terms of human immune system development at baseline. Specifically, the 5 THX mice to be treated by Nil showed 91.1% ± 0.9% human CD45+ cells, 76.8% ± 4.0% human CD19+ B cells and 15.3% ± 3.1% human CD3+ T cells (% of human CD45+ cells); the 5 THX mice to be treated by DMF showed 90.5% ± 0.7% human CD45+ cells, 74.8% ± 3.2% human CD19+ B cells and 16.1% ± 2.4% human CD3+ T cells (% of human CD45+ cells). THX mice in the Nil and DMF groups were age-matched, randomly assigned to treatment groups, and included both female and male animals. THX mice were fed DMF (2 mM = 0.5 mg/mL), a dimethyl ester of fumarate and potent TETs catalytic site inhibitor, and E2 (0.75 μM) or E2 (0.75 μM) alone in drinking water ad libitum on day 0 and continued thereafter. DMF is an orally bioavailable dimethyl ester of the TCA cycle intermediate fumaric acid. Upon oral administration, the compound undergoes initial metabolism by esterase, yielding its primary active metabolite MMF, and, thereby, the physiological TCA intermediate, fumarate.59 These THX mice were injected i.p. with NP16-CGG on day 7 (25 μg in 25 μL alum), boosted (25 μg in 25 μL PBS) on days 14, 21, and 28 and euthanized on day 35. Tcrβ−/− Tcrδ−/− mice, in which any contribution of TET2-expressing T cells21 is excluded, were fed DMF (2 mM), or nil in drinking water ad libitum on day 0 and continued thereafter to analyze the TET2 inhibitory effect by DMF in a T-independent antibody response. Such mice were injected i.p. with NP0.4-LPS on day 7 (25 μg in 25 μL alum), boosted (25 μg in 25 μL PBS) on day 21 and euthanized on day 35. For boosting TET2 activity in a T-independent antibody response, Tcrβ−/− Tcrδ−/− mice were injected i.p. with vitamin C (80 mg in 400 μL PBS) or PBS on days −1, 1, 3 and 5. Such mice were immunized i.p. with NP0.4-LPS on day 0 (50 μg in 50 μL alum), boosted (25 μg in 25 μL PBS) on days 14 and 23 and euthanized on day 44. Serum was collected at indicated time points for titration of total and NP-specific as well as LPS-specific mouse or human antibodies using enzyme-linked immunosorbent assays (ELISAs).6,7,10,15,16,18,19,87–89 VHDJH-CH transcripts from mouse and human B cells were analyzed for SHM/CSR as described in “mouse and human SHM” and “mouse and human B cell transcripts.” Mouse B cells and in vitro B cell differentiation including CSR Cells from mouse bone marrow, spleen, MLNs and/or Peyer’s patches were suspended in ACK Lysis Buffer (BP10-548E, Lonza) for erythrocyte lysis. The lysis reaction was quenched with FBS-RPMI (RPMI-1640 medium, 10-040-CV, Corning, supplemented with 10% v/v Hyclone fetal bovine serum (42Q7980K, Gibco) and 1% v/v antibiotic-antimycotic solution (penicillin-streptomycin and amphotericin B, SV30079.01, Cytiva Life Sciences)), followed by centrifugation. Erythrocyte-free cells were resuspended in FBS-RPMI for further preparation or analysis.15,16,18,19,43,45 To analyze B cell differentiation (naÿve B cells, class-switched B cells, GC B cells, plasmablasts, plasma cells, memory B cells) including CSR, naÿve IgM+IgD+B cells were isolated from spleen of C57BL/6, Tet2−/−, Tet2HxD-mut/HxD-mut, AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+ and AicdaCreOgtfl/fl mice by negative selection using the EasySep Mouse B Cell Isolation Kit (19854, StemCell Technologies) following the manufacturer’s protocol, yielding at least 98% IgM+IgD+B cells. Mouse naÿve IgM+IgD+B cells were cultured in FBS-RPMI (5.0 x 105 cells ml−1) supplemented with 50 mM β-mercaptoethanol for 72 h (for RNA transcript analysis) or for up to 96 h (for fluorescence flow cytometry analysis) upon stimulation with E. coli O55:B5 LPS (4 μg mL−1, L2637, Sigma-Aldrich) for CSR to IgG3, and LPS or CD154 (1 U ml−1; obtained from membrane fragments of baculovirus-infected Sf21 insect cells) plus IL-4 (5.0 ng mL−1, R&D Systems) for CSR to IgG1. For CSR to IgA, naÿve IgM+IgD+B cells were stimulated with E. coli LPS plus IL-4, IL-5 (3 ng mL−1, R&D Systems), TGF-β (2 ng mL−1, R&D Systems), and anti-mouse IgD monoclonal antibody (mAb)/dextran (30 ng mL−1; Fina Biosolutions).10,15,16,18,19,43,45,87 Following a 72 h or 96 h culture, cells were analyzed by flow cytometry to detect class-switched B cells, GC B cells, plasmablasts, plasma cells or memory B cells, as described in “Mouse and human cells flow cytometry”. Human B cells and in vitro B cell differentiation including CSR Human peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats obtained from healthy male and female human donors (18- to 65-year-olds) of different races and ethnic backgrounds (South Texas Blood and Tissue Center, San Antonio, TX), as previously described.13,43,45,58,60 Buffy coat was diluted 1:3 in sterile PBS (pH 7.4, BP3991, Thermo Fisher Scientific) and applied to a Histopaque-1077 density gradient (10771, Sigma Aldrich) in 50 mL SepMate tubes (85450, STEMCELL Technologies), which were spun at 800×g for 15 min. Recovered PBMCs were washed in PBS and resuspended in FBS-RPMI for further preparation or analysis. To analyze human (hu)B cell differentiation, naÿve huCD19+IgM+IgD+B cells were isolated from PBMCs by negative selection using EasySep Human Naÿve B Cell Isolation Kit (17254, STEMCELL Technologies), according to the manufacturer’s protocol, yielding at least 98% huCD19+IgM+IgD+B cells. After pelleting, B cells were resuspended in FBS-RPMI before further analysis or stimulation. Naÿve huIgM+IgD+B cells were cultured in FBS-RPMI (5.0 x 105 cells ml−1) for 72 h (for RNA transcript analysis) or up to 96 h (for fluorescence flow cytometry analysis) upon stimulation with membrane-CD154 (3.0 U ml−1) or CpG ODN 2395 (Eurofins Genomics, 2.5 μg mL−1) plus huIL-2 (589102, BioLegend, 100 ng mL−1), huIL-4 (574002, BioLegend, 20 ng mL−1) and huIL-21 (571202, BioLegend, 50 ng mL−1).15,16,19,43,45,58 Following 96 h culture, human cells were analyzed by flow cytometry to detect class-switched GC B cells, plasmablasts, plasma cells or memory B cells, as described in “mouse and human cells flow cytometry”. Mouse and human cells flow cytometry Cell surface FACS analysis. Mononuclear cells from mouse bone marrow, spleen, MLNs and/or Peyer’s patches were stained with fluorochrome-conjugated anti-mouse mAbs in Hank’s Buffered Salt Solution (HBSS, MT21022CM, Thermo Fisher Scientific) plus 0.1% bovine serum albumin (BSA, BP1600-100, Thermo Fisher Scientific; BSA-HBSS) for 20 min. In vitro-stimulated and/or ex vivo mouse mononuclear cells were stained with PE rat IgG2a anti-mouse (mo)B220 mAb (clone RA3-6B2, 50-112-9644, eBioscience, 1:1,000 dilution), BV650 rat IgG2a anti-moB220 mAb (clone RA3-6B2, 103241, BioLegend, 1:1,000 dilution), BV421 rat IgG2a anti-moCD19 mAb (clone 6D5, 115537, BioLegend, 1:1,000 dilution), PerCP/Cyanine5.5 rat IgM anti-moGL-7 mAb (clone GL7, 144609, BioLegend, 1:1,000 dilution), BV510 hamster IgG2 anti-moCD95 mAb (clone Jo2, 563646, BD Biosciences, 1:1,000), Pacific Blue rat IgG2b anti-moCD38 mAb (clone S21016F, 165615, BioLegend, 1:1,000 dilution), PE/Cyanine7 rat IgG2a anti-moCD138 mAb (clone 281-2; 142513, BioLegend, 1:1,000 dilution), PE mouse IgG1 anti-moIgM mAb (clone AF6-78, 553521, BD Biosciences, 1:1,000 dilution), APC goat IgG anti-moIgG Ab (clone Poly4053, 405336, BioLegend, 1:1,000 dilution), APC rat IgG1 anti-moIgG1 mAb (clone X56, 550874, BioLegend, 1:1,000 dilution), FITC rat IgG2a anti-moIgG3 mAb (clone R40-82, 553403, BioLegend, 1:1,000 dilution), biotin rat IgG anti-moIgG3 mAb (clone RMG3-1, 406803, BioLegend, 1:1,000 dilution) reacted with APC-streptavidin (17-4317-82, eBioscience, 1:1,000 dilution), FITC rat IgG1 anti-moIgA mAb (clone C10-3, 559354, BD Biosciences, 1:1,000 dilution), FITC rat IgG1 anti-moIgG2b mAb (clone RMG2b-1, 406705, BioLegend, 1:1,000 dilution) or 7-aminoactinomycin D (7-AAD, 559925, BD Biosciences). 7-AAD− cells were gated as live cells in all flow cytometry analyses. Among live cells, mouse naÿve B cells, class-switched B cells, GC B cells, plasmablasts, plasma cells and class-switched memory B cells were identified as CD19+IgM+IgD+, CD19+IgG+ (IgG3+, IgG1+ or IgG2b+) or CD19+IgA+, CD19+(B220+)CD95+GL7+, CD19+(B220+)CD138+, CD19lo/−(B220+/−) CD138+ and CD19+B220−IgG+, CD19+B220−IgA+ or CD19+CD38+IgG1+, respectively (Figures S13 and S14). After washing, cells were resuspended in BSA-HBSS for flow cytometry.6,7,10,15,16,18,19,43,45,87–89 Human mononuclear cells from humanized THX mice or peripheral blood of healthy humans were stained with fluorochrome-conjugated anti-human mAbs in BSA-HBSS for 20 min. In vitro-stimulated and/or ex vivo human mononuclear cells were stained with Alexa Fluor 700 mouse IgG1 anti-huCD45 mAb (clone HI30, 304023, BioLegend, 1:1,000 dilution), FITC mouse IgG1 anti-huCD45 mAb (clone 30-F11, 368507, BioLegend, 1:100 dilution), PE/Cyanine7 mouse IgG1 anti-huCD19 mAb (clone HIB19, 302216, BioLegend, 1:100 dilution), BV510 mouse IgG1 anti-huCD138 mAb (clone MI15, 356517, BioLegend, 1:100 dilution), BV510 mouse IgG1 anti-huIgM mAb (clone MHM-88, 314521, BioLegend, 1:100 dilution), BV650 mouse IgG1 anti-huIgM mAb (clone MHM-88, 314525, BioLegend, 1:100 dilution), BV785 mouse IgG2a anti-huIgD mAb (clone IA6-2, 348241, BioLegend, 1:100 dilution), BV421 rat IgG2a anti-huIgG mAb (clone M1310G05, 410703, BioLegend, 1:100 dilution), FITC mouse IgG1 anti-huIgG mAb (clone G18-145, 555786, BD Pharmingen, 1:100 dilution), APC mouse IgG1 anti-huIgA mAb (clone IS11-8E10, 130-113-427, Miltenyi Biotec, 1:100 dilution), FITC goat IgG anti-huIgA Ab (31577, Invitrogen, 1:100 dilution), PE mouse IgG1 anti-huCD27 mAb (clone M-T271, 356405, BioLegend, 1:100 dilution), BV650 mouse IgG1 anti-huCD38 mAb (clone HB-7, 356619, BioLegend, 1:100 dilution) or 7-AAD (559925, BD Biosciences). 7-AAD− cells were gated as live cells in all flow cytometry analyses. Among live cells, human naÿve B cells, class-switched B cells, GC B cells, plasmablasts, plasma cells and memory B cells were identified as huCD45+CD19+IgM+IgD+, huCD19+IgG+ or huCD19+IgA+, huCD19+CD27−CD38+, huCD19+CD27+CD38+, huCD19−CD27+CD38+ and huCD19+CD27+CD38−, respectively (Figures S15 and S16). After washing, cells were resuspended in BSA-HBSS for flow cytometry. B cell division FACS analysis. Naïve IgM+IgD+B cells (1 x 107 cells ml−1) from spleen of AicdaCreTet2+/+ and AicdaCreTet2fl/fl mice were suspended in 3 mL PBS with 2.5 μM CellTrace Yellow (C34567, Thermo Fisher Scientific) followed by incubation at 37°C for 20 min. After washing with FBS-RPMI, B cells were cultured in FBS-RPMI (0.3 x 106 cells ml−1) for 72 h upon stimulation with E. coli LPS or CD154 plus IL-4. Cultured cells were then stained with PE rat IgG2a anti-moB220 mAb (1:1,000 dilution), APC rat IgG1 anti-moIgG1 mAb (1:1,000 dilution) and 7-AAD to quantify proportions of B220+IgG1+B cells after each cell division. Intracellular FACS analysis. Mouse B cells and plasmablasts (2.0 x 106 cells) were surface stained with anti-moCD19 and anti-moCD138 mAbs (1:1,000 dilution), as well as Fixable Viability Dye eFluor 780 (65-0865-14, Thermo Fisher Scientific). Human B cells and plasmablasts (2.0 x 106 cells) from humanized THX mice or healthy humans were surface stained with anti-huCD45, anti-huCD19, anti-huCD27, anti-huCD38 and anti-huCD138 mAbs (1:100 dilution), as well as Fixable Viability Dye eFluor 780. After washing, mouse and human B cells/plasmablasts were fixed by resuspension in Cytofix/Cytoperm buffer (554655, BD Biosciences, 250 μL) and incubated at 4°C for 45 min. After washing twice in BD Perm/Wash buffer (554723, BD Biosciences) for permeabilization, cells were stained with Alexa Fluor 647 rabbit IgG anti-mouse/human AID Ab (bs-7855R-A647, Bioss), PE rat IgG2a anti-moBLIMP-1 mAb (clone 5E7, 150005, BioLegend), Andy Fluor 647-anti-mouse/human TET2 (AF647-TET2, synthesized using iLink Andy Fluor 647 Antibody Labeling Kit, L038, ABP Biosciences), or Alexa Fluor 488 mouse IgG1 anti-huBLIMP-1 mAb (clone 646702, IC36081G, R&D Systems) in BD Perm/Wash buffer for 30 min at 4°C. After washing again twice in BD Perm/Wash buffer, cells were resuspended in BSA-HBSS for flow cytometry (Figures S14 and S16). Ex vivo mouse and human B cell, plasmablast and plasma cell surface markers t-SNE analysis (t-distributed Stochastic Neighbor Embedding is a non-linear dimensionality reduction technique primarily used for visualizing high-dimensional data in 2D or 3D). Surface CD19, IgM, IgD, IgG, IgA, GL-7, CD27, CD38, CD138 and intracellular TET2 protein levels in mouse and human cells, as analyzed by FACS, are indicated by heatmaps using a red-white-blue continuous color scale (red: increased expression, white: baseline expression, blue: decreased expression). Cell surface markers and intracellular TET2 expression profiles of spleen and mesenteric lymph nodes (MLNs) mononuclear cells in NP-CGG-immunized C57BL/6 mice (n = 3) were integrated to create t-SNE maps including mouse naïve B cells (CD19+IgDhiIgMhiGL-7−IgG/IgA−CD38+CD138−), GC B cells (CD19+IgDhiIgMloGL-7+IgG/IgA−CD38−CD138−), late GC B cells (CD19+IgD−IgM+GL-7+IgG/IgAloCD38−CD138lo), plasmablasts (PBs, CD19+IgD−IgM+GL-7−IgG/IgA+CD38−CD138hi), plasma cells (PCs, CD19−IgD−IgM−GL-7−IgG/IgA−CD38−CD138hi) and swMBCs (CD19+IgD−IgMloGL-7−IgG/IgA+CD38+CD138−). Similarly, cell surface markers and intracellular TET2 expression profiles inhuman tonsils (n = 3 subjects) mononuclear cells were integrated to generate t-SNE maps including human transitional B cells (huCD19+IgD+IgM+CD27−CD38+), naïve B cells (huCD19+ IgD+IgM−CD27−CD38−), early GC B cells (huCD19+IgD+IgMloCD27−CD38+), GC B cells (huCD19+IgD−IgM−CD27loCD38+), double-negative B cells (DN, huCD19+IgD−IgM−CD27−CD38−), swMBCs (huCD19loIgD−IgM−CD27+CD38−) and plasma cells (huCD19loIgD−IgM−CD27+CD38+). In t-SNE maps, different pseudocolors were assigned to identify different cell subsets. For in vivo mouse B cell and GC B cell proliferation analyses, NP16-CGG-immunized AicdaCreTet2 +/+ and AicdaCreTet2fl/fl mice were injected i.p. with bromodeoxyuridine (BrdU, 1 mg in 200 μL PBS) on days 10 and 11 (within a 16 h interval) and euthanized 4 h post-BrdU injection. Spleen B cells (CD19+) and GC B cells (B220+CD95+GL-7+IgG+ or B220+CD95+GL-7+IgA+) were surface stained with PE rat IgG2a anti-moB220 mAb (1:1,000 dilution), BV510 hamster IgG2 anti-moCD95 mAb (1:1,000 dilution) and PerCP/Cyanine5.5 rat IgM anti-moGL-7 mAb (1:1,000 dilution), followed by Fixable Viability Dye eFluor 780. After washing, cells were fixed and permeabilized as described above, then stained with APC-anti-mouse/human BrdU Ab (552598, APC BrdU Flow Kit, BD Biosciences) in BD Perm/Wash buffer for 30 min at 4°C. Cells were washed and resuspended in BSA-HBSS for flow cytometry. Flow cytometry analyses were performed using single cell suspensions and using a FACSCelesta flow cytometer (BD Biosciences) with FACSDiva v9.4 software (BD Biosciences). Cells were appropriately gated by forward and side scattering to exclude dead cells and debris. Data were acquired and analyzed using FlowJo v10.9 software (Tree Star). Dimension reduction was performed using t-distributed stochastic neighbor embedding (t-SNE) and overlayed with FlowSOM annotated eight meta-clusters. Surface B220, CD19, IgM, IgD, IgG, IgA, GL-7, CD27, CD38, CD95, CD138 expression and intracellular TET2, AID and BLIMP-1 expression were analyzed. Phenotypic exploration of meta-clusters identified by t-SNE and FlowSOM was conducted and validated by conventional biaxial manual gating from original flow data.90–93 Flow cytometry was used to detect bacteria-bound IgD and IgA in AicdaCreTet2+/+ and AicdaCreTet2fl/fl mice, as we described.15,43,58 Briefly, fecal bacterial pellets were resuspended in 1 mL of PBS containing 1.0% (w/v) BSA, fixed with 7.2% formaldehyde for 10 min at room temperature, washed with PBS, stained with BV421 rat IgG2a anti-moIgD mAb (clone 11-26c.2a; 405725, BioLegend, 1:1,000 dilution) or FITC rat IgG1 anti-moIgA mAb (C10-3, BD Biosciences, 1:1,000 dilution) on ice for 30 min, washed again and then resuspended in PBS containing 0.2 μg mL−1 DAPI for flow cytometry analysis. All events revealed by DAPI were considered as bacteria. Mouse and human antibody ELISAs To measure total or specific mouse IgM, IgG3, IgG1, IgG2b, IgE and IgA in NP16-CGG- and NP0.4-LPS-immunized AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+, AicdaCreOgtfl/fl and Tcrβ−/− Tcrδ−/− mice, sera were diluted 10,000-fold (total IgM, IgG1, IgG3 or IgG2b), 1,000-fold (Total IgA, NP-specific IgM, IgG1 or IgG3) or 100-fold (Total IgE, NP-specific IgG2b and IgA) in PBS containing 0.05% v/v Tween 20 (BP337-500, Thermo Fisher Scientific; PBS-Tween 20), followed by serial 2-fold dilution. Serially diluted sera were incubated at room temperature in 96-well plates coated with pre-adsorbed goat IgG anti-moIgM Ab (1020-01, SouthernBiotech, 1.0 μg mL−1), goat IgG anti-moIgG Ab (IgG3, IgG1, IgG2b, 1030-01, SouthernBiotech, 1.0 μg mL−1), goat IgG anti-moIgA Ab (1040-01, SouthernBiotech, 1.0 μg mL−1), goat IgG anti-moIgE Ab (A90-115A, Bethyl Laboratories, 1.0 μg mL−1), NP5-BSA (average of 5 NP molecules per one BSA molecule, Biosearch Technologies, 1.0 μg mL−1), BSA (Biosearch Technologies, 1.0 μg mL−1) or E. coli O55:B5 LPS (L2637, Sigma-Aldrich, 1 μg mL−1) in 0.1 M sodium carbonate/bicarbonate buffer at pH 9.6. After washing plates with PBS-Tween 20, bound antibodies were detected with biotinylated goat IgG anti-moIgM Ab (1020-08, SouthernBiotech, 1:5,000 dilution), goat IgG anti-moIgG3 Ab (1100-08, SouthernBiotech, 1:5,000 dilution), goat IgG anti-moIgG1 Ab (1070-08, SouthernBiotech, 1:5,000 dilution), goat IgG anti-moIgG2b (1090-08, SouthernBiotech, 1:5,000 dilution), goat IgG anti-moIgA Ab (1040-08, SouthernBiotech, 1:5,000 dilution) or goat IgG anti-moIgE Ab (1110-08, SouthernBiotech, 1:5,000 dilution), followed by reaction with horseradish peroxidase (HRP)-labeled streptavidin (405210, BioLegend), development with O-phenylenediamine (OPD, P8806-50TAB, Sigma-Aldrich) or 3,3′,5,5′-Tetramethylbenzidine (TMB, 421101, BioLegend) substrate and measurement of converted substrate absorbance at 492 nm or 450 nm, respectively. Total mouse immunoglobulin concentrations or specific mouse antibody titers were calculated from OD readings (using a reference curve constructed with known antibody concentrations, BioTek Gen5 Software v3.16) and expressed as μg equivalents per mL (μg eq ml−1), ng eq ml−1 or relative units (RUs, defined as the dilution factor needed to reach 50% saturation binding) using GraphPad Prism v10.4.2 software or Excel v16.89 (Microsoft) software. To measure free fecal IgA in AicdaCreTet2+/+ and AicdaCreTet2fl/fl mice, feces (10 mg) were suspended in 100 μL PBS (filtered through 0.2 μm filter), homogenized and centrifuged at 400×g for 5 min to remove large particles. The supernatant was centrifuged at 8,000×g for 10 min and diluted 200-fold in PBS-Tween 20, followed by serial 2-fold dilution. Serially diluted fecal supernatant was incubated at room temperature in 96-well plates coated with pre-adsorbed goat IgG anti-moIgA Ab (1.0 μg mL−1) in 0.1 M sodium carbonate/bicarbonate buffer at pH 9.6. After washing plates with PBS-Tween 20, bound antibodies were detected with biotinylated goat IgG anti-moIgA Ab (1:5,000 dilution), followed by reaction with HRP-labeled streptavidin, development with OPD substrate and measurement of converted substrate absorbance at 492 nm. Concentration of free IgA was calculated from OD readings and expressed as ng eq ml−1 using GraphPad Prism v10.4.2 software or Excel v16.89 (Microsoft) software. To measure total or specific huIgM, huIgG1 and huIgA in NP16-CGG-immunized humanized THX mice, sera were diluted 400-fold or 20-fold in PBS-Tween 20, followed by serial 2-fold dilution. Serially diluted samples were incubated at room temperature in 96-well plates coated with pre-adsorbed goat IgG anti-huIgM Ab (2020-01, SouthernBiotech, 1.0 μg mL−1), goat IgG anti-huIgG Ab (huIgG1, huIgG2, huIgG3 and huIgG4, 2015-01, SouthernBiotech, 1.0 μg mL−1), goat IgG anti-huIgA Ab (2050-01, SouthernBiotech, 1.0 μg mL−1), NP5-BSA (1.0 μg mL−1) or BSA (1.0 μg mL−1) in 0.1 M sodium carbonate/bicarbonate buffer at pH 9.6. After washing plates with PBS-Tween 20, bound antibodies were detected with biotinylated goat IgG anti-huIgM Ab (2020-08, SouthernBiotech, 1:500 dilution), mouse IgG2b anti-huIgG1 mAb (clone G17-1, 555869, BD Biosciences, 1:500 dilution) or goat IgG anti-huIgA Ab (2050-08, SouthernBiotech, 1:500 dilution), followed by reaction with HRP-labeled streptavidin, development with OPD or TMB substrate and measurement of converted substrate absorbance at 492 nm or 450 nm, respectively. Total human immunoglobulin concentrations or specific human antibody titers were calculated from OD readings and expressed as μg eq ml−1 or RUs using GraphPad Prism v10.4.2 software or Excel v16.89 (Microsoft) software. ASCs ASCs (plasmablasts/plasma cells) were detected by ELISPOT. Spleen or bone marrow cells from AicdaCreTet2+/+, AicdaCreTet2fl/fl and Tcrβ−/− Tcrδ−/− mice were suspended in FBS-RPMI then cultured at 37°C overnight in 96 well PVDF Multi-Screen filter plates (activated with 35% ethanol, MAIPS4510, Millipore) coated with goat IgG anti-moIgM Ab, goat IgG anti-moIgG Ab, goat IgG anti-moIgA Ab or NP5-BSA (all 5 μg mL−1). Spleen and bone marrow cells were plated at 1.25 x 105 and 2.5 x 105 cells per well to analyze total and specific ASCs, respectively. After removing supernatants, plates were incubated with biotinylated goat IgG anti-moIgM Ab (1:1,000 dilution), goat IgG anti-moIgG1 Ab (1:1,000 dilution) or goat IgG anti-moIgA Ab (1:1,000 dilution) for 2 h then, after washing, incubated with HRP-conjugated streptavidin, followed by Vectastain AEC peroxidase substrate (SK-4200, Vector Laboratories). In each experiment, ELISPOT images were generated using data from two independent mice. Individual ASC spots were enumerated using a CTL Immunospot Analyzer and software (CTL ImmunoCapture Software v7.0.21.9, Cellular Technology).15,16,43,45,58,94 Mouse and human B cell antibody SHM Mouse B cell SHM. To analyze SHM in the NP16-CGG- or NP-LPS-induced mouse antibody responses, RNA (1–2 μg) was extracted from AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreOgt+/+, AicdaCreOgtfl/fl and Tcrβ−/− Tcrδ−/− mouse B cells using the RNeasy Mini Kit (74104, Qiagen). cDNA was synthesized from total RNA using SuperScript III First-Strand Synthesis System (18080051, Invitrogen) with oligo-dT primer. Rearranged V1-72DJH-Cγ3, V1-72DJH-Cγ1 or V1-72DJH-Cγ2b cDNA encoding the anti-NP VDJ-IgH chain was amplified using a V1-72 leader-specific forward primer together with a nested Cγ3-, Cγ1-, or Cγ2b-specific reverse primer tagged with Illumina clustering adapters and Phusion high-fidelity DNA polymerase (M0530S, New England BioLabs).15 PCR amplification conditions were 98°C for 10 s, 60°C for 45 s, and 72°C for 1 min for 30 cycles. The cDNA amplicons were further amplified using index PCR involving Illumina clustering adapters and beads cleanup, quantified and then loaded onto Illumina MiSeq system using the 300-bp pair-end sequencing module. Somatic point-mutations in recombined V1-72DJH-Cγ3, V1-72DJH-Cγ1 and V1-72DJH-Cγ2b transcripts were analyzed using IMGT/HighV-QUEST v1.9.2 (The International ImMunoGeneTics Information System, http://www.imgt.org/HighV-QUEST/home.action) and corrected for polymerase and sequencing error rates (0.008) to calculate the frequency of somatic point-mutations (change/base).15,43,45,58,95 Pie charts depicted proportions of V1-72DJH-Cγ3, V1-72DJH-Cγ1 and V1-72DJH-Cγ2b transcripts carrying different numbers of point-mutations; slice gray gradients depict increasing numbers of point-mutations (from 0 to more than 12 per transcript). Histograms depicted means of total, S and R V1-72 mutation frequencies (change/base) in FR1, CDR1, FR2, CDR2 and FR3 of V1-72DJH-Cγ3, V1-72DJH-Cγ1 and V1-72DJH-Cγ2b transcripts. Boxplots depicted overall V1-72 point-mutation frequency in V1-72DJH-Cγ3, V1-72DJH-Cγ1 and V1-72DJH-Cγ2b transcripts. Human B cell SHM. To analyze SHM in human NP16-CGG-induced antibody response, RNA (1–2 μg) was extracted from THX mice human B cells, and cDNA was synthesized as described above. Rearranged huV3DJH-Cγ1 cDNA was amplified using a huV3 leader-specific forward primer together with a nested huCγ1-specific reverse primer tagged with Illumina overhang adapters and Phusion high-fidelity DNA polymerase (M0530S, New England BioLabs) – amplification of huIgH V3 genes was chosen as this family includes gene members of high sequence similarity to mouse V1-72 (V186.2/V3 gene), the gene encoding the most efficient “NP-binding” mouse IgH V segment (https://www.imgt.org/ligmdb/view?id=J00239).44,95 PCR amplification conditions were 98°C for 10 s, 60°C for 45 s and 72°C for 1 min for 30 cycles. The cDNA amplicons were further amplified, sequenced and analyzed for somatic point-mutations as described above. Somatic point-mutations in recombined huV3DJH-Cγ1 were analyzed using IMGT/HighV-QUEST v1.9.2 (The International ImMunoGeneTics Information System, http://www.imgt.org/HighV-QUEST/home.action) and corrected for polymerase and sequencing error rates (0.008) to calculate the frequency of somatic point-mutations (change/base)15,43,45,58,95. Pie charts depicted proportions of huV3DJH-Cγ1 transcripts carrying different numbers of point-mutations; slice gray gradients depict increasing numbers of point-mutations (from 0 to more than 12 per transcript). Histograms depicted means of total, S and R V1-72 mutation frequencies (change/base) in FR1, CDR1, FR2, CDR2 and FR3 of huV3DJH-Cγ1 transcripts. Boxplots depicted overall point-mutation frequency in huV3DJH-Cγ1 transcripts. To analyze human B cell clonality, rearranged huV3DJH-Cγ1 transcripts (up to 250,000 sequences) were analyzed by Illumina MiSeq amplicon sequencing and segregated based on identical human VH gene segment, the same and unique IgH CDR3 together with the same JH sequence. Each discrete clone was depicted as an individual rectangle or square (unique color), whose Area reflects B cell clone size, as inferred from the sum of identical huV3DJH-Cγ1 transcripts (TreeMap, Microsoft Excel v16.89 and IMGT/HighV-QUEST statistic data). To analyze intraclonal diversification, shared and unique point-mutations in huV3DJH-Cγ1 transcripts within each B cell clone were used to construct genealogical trees (phylogenetic maps), revealing sequential multi-step accumulation of point-mutations from unmutated progenitors, and allowing for detailed intraclonal diversification analysis. Genealogical trees were constructed by uploading FASTA files of all segregated huV3DJH-Cγ1 transcripts onto PHYLOViZ Online v2.0 (http://www.phyloviz.net), which uses a JAVA implementation of the Feil’s goe-BURST algorithm rules for visualization of multiple phylogenetic inference trees. Mouse and human B cell transcripts To analyze mouse Tet1, Tet2, Tet3, Aicda, Prdm1, Irf4, Bach2, germline Iμ-Cμ, Iγ3-Cγ3, Iγ1-Cγ1, Iε-Cε, and post-recombination Iμ-Cγ3, Iμ-Cγ1, Iμ-Cε transcripts, RNA was extracted from in vitro-stimulated AicdaCreTet2fl/fl and AicdaCreTet2fl/fl mouse B cells (2.0-5.0 x 106 cells) using the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. Residual DNA was removed using gDNA eliminator columns (Qiagen). cDNA was synthesized from 1 to 2 μg total RNA using SuperScript III First-Strand Synthesis System (18080051, Invitrogen) and oligo-dT primer. Transcript expression was analyzed by SYBR Green dye (IQ SYBR Green Super-mix, 115010139, Bio-Rad) incorporation in PCR reactions involving specific forward and reverse primers (Table S2).16,43 Reactions were performed in an iCycler (Bio-Rad) real-time qPCR system using the following amplification cycles: 95°C for 15 s, 40 cycles at 94°C for 10 s, 60°C for 30 s, and 72°C for 30 s – data acquisition was performed during this 72°C extension step (Bio-Rad CFX Manager Software v3.1). Melting curve analysis was performed from 72°C to 95°C. The 2−ΔCt method (2−ΔCt = 2–[Ct(Gapdh)-Ct(target gene)]) was used to analyze levels of transcripts and data were normalized to levels of Gapdh. To analyze human AICDA and PRDM1 transcripts, RNA was extracted from in vitro-stimulated human B cells (CpG plus huIL-2, huIL-4 and huIL-21) using the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. cDNA synthesis was performed as described above, followed by PCR reactions involving specific human forward and reverse primers (Table S2). Reactions were performed as described above. The 2−ΔCt method (2−ΔCt = 2−[Ct(HPRT1)-Ct(target gene)]) was used to determine transcript levels and data were normalized to levels of HPRT1.6,7,10,15,16,18,19,43,45,87–89 Aicda and Prdm1 constructs and expression Aicda and Prdm151 coding region cDNAs were amplified from E. coli LPS-stimulated mouse B cells using specific forward and reverse primers (Table S2). Aicda and Prdm1 amplicons were digested with XhoI and EcoRI restriction enzymes and cloned into similarly digested pMIG retroviral expression vectors (9044, Addgene) containing enhanced green fluorescent protein (GFP). The resulting constructs containing Aicda or Prdm1 and Gfp genes separated by an internal ribosome entry site (IRES), which allows bicistronic expression of AID or BLIMP-1 and GFP. “Empty” pMIG vectors or pMIG vectors encoding GFP and Aicda or Prdm1 were used with the pCL-Eco retrovirus-packaging vector (NBP2-29540, Novus Biologicals) to transfect human embryonic kidney 293T cells (HEK293T) by calcium phosphate precipitation (ProFection Mammalian Transfection System, E1200, Promega). Virus-containing supernatant was used to transduce mouse Tet2−/− B cells after a 12 h LPS activation, as we previously described.6,16 Transduced B cells were then cultured for 96 h upon stimulation with LPS plus IL-4, followed by flow cytometry analysis of IgG1+B cells and CD19+CD138+ plasmablasts – cells pre-gated for GFP+ and 7-AAD− live cells. Tet2 mutant constructs and expression Wild-type TET2 construct (Tet2WT, aa 990–1920), HxD catalytic site mutant TET2 construct (Tet2HxD-mut, aa 990–1920 with mutated H1303Y and D1305A), C-terminal-end deletion mutant TET2 construct (Tet2OGT-Δmut, aa 990–1875)and double HxD catalytic site and C-terminal-end deletion mutant TET2 construct (Tet2HxD-mut/OGT-Δmut, aa 990–1875 with mutated H1303Y and D1305A) were generated by PCR using specific forward and reverse primers, including a target-specific (HxD or C-terminal-end) mutagenizing primer (Table S2) and the whole mouse TET2 isoform 2,51,55 from E. coli LPS-stimulated mouse B cells, or partial DNA sequence as template (inserted into pMIG vector). Tet2WT construct was synthetized by amplifying Tet2 nucleotide sequence encoding aa 990–1920, ending with “C-terminal-end” and including the “conserved signature TET2 HxD motif”, the putative catalytic site conserved in mouse and human TET1, TET2 and TET3, as identified by transduction of neoplastic kidney 293T cells.54 Tet2WT construct also includes coding region for Cys-rich regions and double-stranded β-helix (DSBH) regions. TET2 isoform 2 H1303xD1305 catalytic site corresponds to TET2 isoform 1 H1295xD129726 and TET2 isoform X2 H1302xD130454 catalytic sites. The catalytic Tet2HxD-mut mutant construct, which contained two point-mutations in the conserved signature HxD motif (H1303YxD1305A) was constructed by PCR using site-directed mutagenizing primer (Table S2); OGT-binding deletion mutant Tet2OGT-Δmut construct was generated by PCR using specific primers amplifying TET2 990 through 1875 sequence (TET2 isoform 2 G1876-V1920 C-terminal-end residues are identical to TET2 isoform 1 G1868-V191226 and TET2 isoform X2 G1875-V191954 residues); finally, the dual mutant Tet2HxD-mut/OGT-Δmut construct was generated by combined and sequential PCR amplifications using the OGT-binding deletion mutant Tet2OGT-Δmut construct as a template and specific primers including the site-directed mutagenizing primer. Due to the lack of TET2 constructs N-terminal aa 1–990, TET2 protein expressed by Tet2WT, Tet2HxD-mut, Tet2OGT-Δmut and Tet2HxD-mut/OGT-Δmut constructs were approximately 121 kDa. Primers incorporated restriction site sequences (XhoI and EcoRI) for insertion in pMIG retroviral expression vector containing a GFP coding sequence, as we described.89 pMIG vectors encoding GFP and Tet2WT, Tet2HxD-mut, Tet2OGT-Δmut or Tet2HxD-mut/OGT-Δmut were used with pCL-Eco retrovirus-packaging vector (NBP2-29540, Novus Biologicals) to transfect human embryonic kidney 293T cells (HEK293T) by calcium phosphate precipitation (ProFection Mammalian Transfection System, E1200, Promega). Virus-containing supernatant was used to transduce mouse Tet2−/− B cells after a 12 h LPS activation, as we previously described.6,16 Upon stimulation with LPS plus IL-4, transduced B cells were cultured for 96 h, followed by flow cytometry analysis (cells pre-gated for GFP+ and 7-AAD−) of intracellular TET2 expression, cell surface IgG1+ and CD19+CD138+ (plasmablasts). ChIP-qRT-PCR ChIP assays were performed as previously described.15,16,43,94 Briefly, B cells (1.0 × 107) were treated with formaldehyde (1% v/v) for 10 min at 25°C to crosslink chromatin. After quenching with 100 mM glycine (pH 8.0) and washing with cold PBS containing protease inhibitors (Roche), B cells were resuspended in lysis buffer (20 mM Tris-HCl, 200 mM NaCl, 2 mM EDTA, 0.1% w/v SDS and protease inhibitors, pH 8.0). Chromatin was fragmented by sonication (DNA fragments of about 200–1,000 bp in length), pre-cleared with Pierce Protein A agarose beads (20333, Thermo Fisher Scientific) and incubated with rabbit IgG anti-mouse/human TET2 Ab (A20365, ABclonal, 5 μg mL−1), rabbit IgG anti-mouse/human OGT Ab (A1990, ABclonal, 1:100 dilution), rabbit IgG anti-human H2B (GLCNAC S112) Ab (GS0022, GlycoScientific, 1:100 dilution), or control rabbit or mouse IgG mAb (1:100 dilution) with irrelevant specificities at 4°C overnight. An aliquot of chromatin (10% of the volume used for ChIP per antibody) was used as 10% input for assay normalization. Immune complexes were precipitated by Protein A agarose beads, washed and eluted (50 mM Tris-HCl, 0.5% SDS, 200 mM NaCl, 100 μg mL−1 proteinase K, pH 8.0), followed by de-crosslinking at 65°C for 4 h. DNA was purified using a QIAquick PCR Purification Kit (28104, Qiagen). Aicda and Prdm1 promoter or enhancer region DNA was amplified from immunoprecipitated chromatin by qPCR using specific forward and reverse primers (custom synthesized by Eurofins Genomics, Table S2). Reactions were performed in an iCycler (Bio-Rad) real-time qPCR system using the following amplification cycles: 95°C for 15 s, 40 cycles at 94°C for 10 s, 60°C for 30 s, and 72°C for 30 s – data acquisition was performed during this 72°C extension step (Bio-Rad CFX Manager Software v3.1). Melting curve analysis was performed from 72°C to 95°C. The percentage of input recovered after immunoprecipitation was calculated using the 2−ΔCt method (2−ΔCt = 2−[Ct(ChIP)-Ct(input)] x 100). Data were normalized to input chromatin DNA and depicted as relative abundance of each amplicon. MeDIP-qRT-PCR and hMeDIP-qRT-PCR DNA methylation and hydroxymethylation were analyzed by methylated DNA immunoprecipitation (MeDIP) and hydroxymethylated DNA immunoprecipitation (hMeDIP). Genomic DNA was extracted from B cells using DNeasy Blood & Tissue Kit (69504, Qiagen) according to the manufacturer’s instructions. DNA was fragmented by restriction digestion with MseI (R0525S, New England Biolabs) at 37°C overnight, followed by heat inactivation of MseI at 65°C for 20 min. An aliquot of fragmented DNA was set aside as 10% input DNA for normalization of data. Fragmented DNA (0.1–1.0 μg) was incubated with mouse IgG1 anti-mouse/human 5-methylcytosine (5mC) mAb (clone A1, 61479, Active Motif) or rabbit IgG anti-mouse/human 5-hydroxymethylcytosine (5hmC) Ab (39092, Active Motif) in the presence of “bridging antibody” for mouse IgG and Protein A (53017, Active Motif), and protease inhibitor cocktail (55010, Active Motif) at 4°C overnight with rotation. Immune complexes were precipitated by Protein G magnetic beads, then washed and eluted (50 mM Tris-HCl, 0.5% SDS, 200 mM NaCl, 20 μg mL−1 proteinase K, pH 8.0), followed by incubation at 55°C for 2 h. DNA was purified using a QIAquick PCR purification kit (Qiagen). For MeDIP- and hMeDIP-qRT-PCR, Aicda and Prdm1 promoter or enhancer region DNA was amplified from the immunoprecipitated chromatin using specific forward and reverse primers (custom synthesized by Eurofins Genomics, Table S2). Reactions and quantification of 5mC- or 5hmC-enriched DNA were performed as described in “ChIP-qRT-PCR”. MeDIP-seq and hMeDIP-Seq For genome-wide mapping of 5mC- and 5hmC-enriched regions, including Aicda and Prdm1 loci, 0.1 ng–20 ng DNA was fragmented to 200 bp fragments using Covaris (520145, shearing parameter 2 min in LV tubes using 15 μL microTUBE-15 AFA Beads Screwcap), and subjected to MeDIP/hMeDIP-Seq library preparation using the NEBNext Ultra II DNA Library Preparation Kit for Illumina (E7645L, New England Biolabs). DNA (MeDIP/hMeDIP-seq) libraries were assessed for quality, size and quantification by Bioanalyzer (Agilent) and Qubit (Invitrogen), pooled, and then sequenced using the Illumina NovaSeq 6000 100PE platform. MeDNA and hMeDNA immunoblotting Genomic DNA methylation and hydroxymethylation were analyzed by blotting followed by detection with anti-5mC and anti-5hmC Abs.43 Total genomic DNA was extracted from LPS- or CD154-stimulated AicdaCreTet2+/+ and AicdaCreTet2fl/fl B cells using DNeasy Blood and Tissue Kit. DNA was digested by MseI enzyme and denatured at 98°C for 3 min, followed by cold shock in ice. Denatured DNA was blotted onto Amersham Hybond-N+ membranes (Cytiva) at 2-fold serial dilutions (300, 150, 75, 37.5 and 19 ng) using BioDot Microfiltration Apparatus (1706545, Bio-Rad) and fixed by UV cross-linking. The blotted membrane was then incubated overnight at 4°C with mouse IgG1 anti-mouse/human 5mC mAb (clone A1, 61479, Active Motif, 1:5,000 dilution) or rabbit IgG anti-mouse/human 5hmC Ab (39092, Active Motif, 1:5,000 dilution). After washing thrice for 10 min each with TBS-Tween 20 (0.05%), the blotted membrane was incubated at room temperature for 2 h with HRP-conjugated goat IgG anti-rabbit IgG secondary antibody (7074P2, Cell Signaling Technology, 1:5000). After washing again thrice with TBS-Tween 20, bound HRP-conjugated Abs were detected using Amersham ECL Prime Western Blotting Detecting Reagent (RPN2232, Cytiva) and proteins visualized using a PXi Touch Imaging System (Syngene).15,16,43,94 Western immunoblotting, TET2 phosphorylation, acetylation and O-GlcNAcylation TET1, TET2, TET3, AID and BLIMP-1 immunoblotting. In vitro-activated C57BL/6, AicdaCreTet2+/+ and AicdaCreTet2fl/fl mouse B cells were lysed in Laemmli buffer. Cell extracts containing equal amounts of protein (50–100 μg) were fractionated by 10% SDS-poly-acrylamide gel electrophoresis (SDS-PAGE), followed by protein transfer onto polyvinylidene difluoride (PVDF) membranes (88518, Thermo Fisher Scientific) for 90 min (250 mA) at 4°C. Membranes were blocked with 3% BSA in TBS-Tween 20 (0.05%) at room temperature for 30 min then incubated overnight at 4°C with rabbit IgG anti-mouse/human TET1 Ab (A1506, ABclonal, 1:2,000 dilution), rabbit IgG anti-mouse/human TET2 Ab (A1526, ABclonal, 1:2000), rabbit IgG anti-mouse/human TET3 Ab (A21930, ABclonal, 1:2,000 dilution), mouse IgG1 anti-mouse/human AID mAb (39–2500, Thermo Fisher Scientific, 1:1,000 dilution) or rat IgG2a anti-mouse/human BLIMP-1 mAb (clone, 6D3, 14–5963-82, Thermo Fisher Scientific, 1:1,000 dilution). After washing thrice for 10 min each with TBS-Tween 20 (0.05%), membranes were incubated at room temperature for 2 h with HRP-conjugated goat IgG anti-rabbit IgG (7074P2, Cell Signaling Technology, 1:5000) secondary Ab, HRP-conjugated horse IgG anti-mouse IgG (7076P2, Cell Signaling Technology, 1:2000) secondary Ab or HRP-conjugated goat IgG anti-rat IgG (7077S, Cell Signaling Technology, 1:2000) secondary Ab. After washing again thrice with TBS-Tween 20 (0.05%), bound HRP-conjugated Abs were detected using Western Lightning Plus-Enhanced chemiluminescence (ECL) reagents (NEL103E001EA, Revvity) and X-ray film. Densitometry was performed using Fiji v1.0 software. TET2 phosphorylation and acetylation and TET2 O-GlcNAcylation. Mouse and human TET2 protein were immunoprecipitated by rabbit IgG anti-mouse/human TET2 Ab (A1526, ABclonal, 1:2000). Immunoprecipitated TET2 were fractionated by 10% SDS-PAGE in equal amounts (50–100 μg). The fractionated proteins were transferred onto PVDF membranes (88518, Thermo Fisher Scientific) for 90 min (250 mA) at 4°C. Membranes were blocked with 3% BSA in TBS-Tween 20 (0.05%) at room temperature for 30 min then incubated overnight at 4°C with rabbit IgG anti-mouse/human TET2 Ab (A1526, ABclonal, 1:2000), mouse IgG2b anti-mouse/human phosphorylated Tyr mAb (clone PY20, 309301, BioLegend, 1:1,000 dilution), mouse IgG2b anti-mouse/human acetylated Lys mAb (clone 15G10, 623401, BioLegend, 1:1,000 dilution), mouse IgG1 anti-mouse/human O-GlcNAc Ser and Thr mAb (clone RL2, 677902, BioLegend, 1:2,000 dilution) or mouse IgG2b anti-mouse/human β-ACTIN mAb (clone 2F1-1, 643802, BioLegend, 1:1,000 dilution). After washing thrice for 10 min each with TBS-Tween 20 (0.05%), membranes were incubated at room temperature with HRP-conjugated horse IgG anti-mouse IgG secondary Ab (7076P2,Cell Signaling Technology, 1:2000) or HRP-conjugated goat IgG anti-rabbit IgG secondary Ab (7074P2,Cell Signaling Technology, 1:5000). After washing again thrice with TBS-Tween 20 (0.05%), bound HRP-conjugated Abs were detected using Amersham ECL Prime Western Blotting Detecting Reagent (RPN2232, Cytiva) and proteins visualized using a PXi Touch Imaging System (Syngene).15,16,43,94 Native TET2 and OGT Co-IP TET2 and OGT native Co-IP. In vitro-activated C57BL/6, AicdaCreOgt+/+ and AicdaCreOgtfl/fl mouse B cells (LPS plus IL-4) as well as in vitro-activated human B cells (CpG plus huIL-2, huIL-4 and huIL-21) were washed twice with cold PBS and pelleted (5,000 rpm at 4°C for 5 min). To extract nuclear proteins, fresh cell pellets were then processed with NE-PER Nuclear and Cytoplasmic Extraction Reagents (78833, Thermo Fisher Scientific). Nuclear extracts were incubated with Pierce Protein G magnetic beads (88848, Thermo Fisher Scientific) at 4°C for 2 h to remove non-specific binding, followed by incubation with rabbit IgG anti-mouse/human TET2 Ab (A1526, Abclonal, 5 μg mL−1) or rabbit IgG anti-mouse/human OGT Ab (A1990, Abclonal, 5 μg mL−1) at 4°C overnight. Protein-Ab complexes were washed thrice with wash buffer (10 mM Tris-HCl, 10 mM NaCl, 1 mM EDTA, 1% Triton X-100 and protease inhibitors, pH 8.0), then eluted from Protein G magnetic beads by incubation with Laemmli buffer (1610747, Bio-Rad) at 55°C for 10 min while mixing, followed by magnetic separation. Immunoprecipitated proteins were transferred onto PVDF membranes (88518, Thermo Fisher Scientific) for 90 min (250 mA) at 4°C. Membranes were blocked with 3% BSA in TBS-Tween 20 (0.05%) at room temperature for 30 min then incubated overnight at 4°C with rabbit IgG anti-mouse/human TET2 Ab (A1526, ABclonal, 1:2000) rabbit IgG anti-mouse/human OGT Ab (A1990, ABclonal, 1:2000), mouse IgG1 anti-mouse/human HDAC1 Ab (A18304, ABclonal, 1:2000), rabbit IgG anti-mouse/human PROSER1 Ab (PA5-72025, Thermo Fisher Scientific, 1:2000) or rabbit IgG anti-mouse/human CREBBP Ab (A26780, ABclonal, 1:2000). After washing thrice for 10 min each with TBS-Tween 20 (0.05%), membranes were incubated at room temperature with HRP-conjugated goat IgG anti-rabbit IgG secondary Ab (7074P2,Cell Signaling Technology, 1:5000). After washing again thrice with TBS-Tween 20 (0.05%), bound HRP-conjugated Abs were detected using Amersham ECL Prime Western Blotting Detecting Reagent (RPN2232, Cytiva) and proteins visualized using a PXi Touch Imaging System (Syngene).15,16,43,94 TET2 and OGT small molecule inhibitors and activator (vitamin C) TET-IN-35 (AOB11121, AOBIOUS, 4 mM), ST 045849 (SML2702, Sigma-Aldrich, 55 mM) or vitamin C (ascorbic acid, A15613-36, Thermo Fisher Scientific, 1.4 mM) was added to in vitro mouse and human B cell microcultures for TET2 catalytic inhibition, OGT function inhibition or TET2 activation studies, respectively. The IC50 of ST 045849 for inhibition of nuclear OGT is 53 ± 7 μM, as determined using a high-throughput fluorescent UDP-GlcNAc displacement biochemical assay designed to identify potential small molecule OGT inhibitors. As a biochemical screening assay does not necessarily provide appropriate information for a cell-culture system, we performed a titration to determine the optimal ST 045849 OGT inhibitory concentration applicable to mouse and human B cell cultures (experiments of Figures S10 and S11). All ST 045849 concentrations (25, 50, and 55 mM) explored were compatible with comparable levels of cell viability. A 55 mM concentration was chosen to achieve consistent and measurable suppression of O-GlcNAcylation in activated mouse and human B cells in conjunction with full B cell viability (Figure S17). TET-IN-C35 inhibits TET proteins’ catalytic activity without disrupting their adaptor function.40 Specifically, TET-IN-C35 binds to TET2 at residue T1259 and more efficiently inhibits (IC50 = 1.2 μM) TET2 catalytic activity in somatic cell reprograming than it does TET1 (IC50 = 3.5 μM) or TET3 (IC50 = 2.3 μM).40 ST 045849 inhibits OGT catalytic activity, i.e., protein O-GlcNAcylation,41 although its OGT binding site has not been precisely identified, and exerts potent inhibitory effects on prostate cancer cells and on hepatic cells.96 Fluorescence microscopy To detect IgM+ and IgA+ cells in spleen and Peyer’s patches of NP16-CGG-immunized AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreOgt+/+ and AicdaCreOgtfl/fl mice, tissues, including intestines “Swiss-rolls”, were fixed with 4% paraformaldehyde and embedded in paraffin, as we described.6,7,10,15,16,18,19,43,45,87–89 10 mm sections were cut and heated at 80°C to adhere to glass slides, washed 4 times in xylene (214736-1L, Millipore Sigma) for 2 min, dehydrated twice with 100% ethanol for 1 min, dehydrated twice with 95% ethanol for 1 min and washed twice in water for 1 min. Antigens were unmasked using 2 mM EDTA (15-575-020, Thermo Fisher Scientific) at 90°C for 30 min, followed by a cooling step at 25°C, thrice washing with TBS (Tris-buffered saline, 15-567-027, Thermo Fisher Scientific) and blocking with 10% BSA for 15 min. Slides were washed again thrice with TBS and then stained with PE rat IgG2a anti-moIgM mAb (clone II/41, 12-5790-82, Thermo Fisher Scientific, 1:100 dilution) PerCP rat IgG2a anti-moIgD mAb (clone 11-26c.21, 405735, Biolegend, 1:100) and FITC rat IgG1 anti-moIgA mAb (clone mA-6E1, 11–4204, Thermo Fisher Scientific, 1:200 dilution) for 2 h in a dark moist chamber. After washing thrice with 0.1% Triton X-100 (T9284, Sigma Aldrich) in TBS, slides were air dried, and coverslips were mounted using ProLong Gold Antifade Reagent with DAPI (P36935, Thermo Fisher Scientific). Fluorescence images were captured using a Zeiss Axio Imager-V.1 fluorescence microscope (ZEN Microscopy Software v3.9×, 20× objective).6,7,10,15,16,18,19,43,45,87–89 To analyze GC formation and GC B cells, spleens from NP16-CGG-immunized AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreOgt+/+ and AicdaCreOgtfl/fl mice were harvested on day 10 and snap frozen in Tissue-Tek O.C.T. Compound (4583, Sakura). Spleen sections (10 μm) were prepared by cryostat and loaded onto positively charged slides, fixed in cold acetone and stained with PerCP/Cyanine5.5 rat IgM anti-moGL7 mAb (1:100 dilution), FITC rat IgG2a anti-moCD19 mAb (clone 6D5, 115505, BioLegend, 1:100 dilution) and APC rat IgG2b anti-moCD3 mAb (clone 17A2, 100235, BioLegend, 1:100 dilution) for 1 h at 25°C in a moist chamber. Cover slips were mounted on slides using ProLong Gold Antifade Reagent with DAPI. Images were captured using a Zeiss Axio Imager-V.1 fluorescence microscope ZEN Microscopy Software v3.9×, 20× objective) and presented using Fiji (ImageJ2 v2.16.0/1.54p). In immunofluorescence experiments to identify class-switched B cells in mouse Peyer’s patches, surface IgM,IgD and IgA expression were denoted by magenta, orange and green pseudocolors. Cell nuclei, as identified by DAPI, was denoted by blue pseudocolor. To identify GC B cells and T cells in mouse spleen, surface CD19, CD3 and GL-7 expression was denoted by green, magenta and yellow pseudocolors. Cell nuclei, as stained by DAPI, were identified by blue pseudocolor. TET2-DNA-OGT AlphaFold 3.0 modeling Predicted structural models of TET2-DNA-OGT complex were generated using AlphaFold 3.0, a deep learning-based structure prediction framework capable of modeling multi-component biomolecular assemblies, including protein-protein and protein-nucleic acid complexes. AlphaFold 3.0 infers three-dimensional structures directly from primary sequence information by leveraging evolutionary relationships and learned physical constraints and provides confidence metrics to evaluate the reliability of predicted interactions.52 Protein sequences corresponding to murine TET2 (isoform 2: NP_001333665.1; isoform 1: NP_001035490.2; isoform X2: XP_006501349.1) and murine OGT (UQT69321.1) were retrieved from NCBI database. TET2 isoforms were modeled to account for potential isoform-specific structural differences and to ensure that predicted interactions with OGT were not dependent on a single protein variant. To incorporate DNA into the complex, a 35-base-pair double-stranded DNA sequence derived from regulatory regions of Aicda and Prdm1 loci was used. Structure predictions were performed using AlphaFold 3.0 in multimer mode, which allows simultaneous modeling of multiple macromolecules within a single prediction. Default parameters, which included the usage of structural templates with release dates before September 20, 2021, were used for all runs, without the use of experimentally derived structural templates or externally imposed restraints. This ensured that predicted interactions were derived solely from AlphaFold’s learned representations and sequence-based inference rather than from prior structural knowledge. For each modeling run, AlphaFold 3.0 generated multiple candidate structures ranked by internal confidence metrics. The highest-ranked models were selected for visualization. Because no crystallography or other determined structure of TET2–OGT or TET2-OGT-DNA complex is currently available, model reliability was evaluated using AlphaFold 3.0 confidence metrics rather than direct structural comparison. Molecular interface-predicted template modeling (ipTM) score was used to assess confidence in the predicted interfaces between interacting components. The ipTM metric provides a quantitative estimate of how reliably AlphaFold predicts the relative positioning of interacting molecules within a complex, with higher values indicating greater confidence in interfacial architecture. All analyzed models exhibited ipTM values greater than 0.6, consistent with a stable and biologically plausible interaction interface. Predicted aligned error (PAE) matrices were also examined to assess confidence in the relative spatial positioning of individual residues and domains. PAE values represent AlphaFold’s estimated error in the predicted distance between pairs of residues, allowing evaluation of domain-level organization and inter-domain flexibility. PAE analysis revealed low predicted error within the C-terminal region of TET2, beginning at approximately amino acid 1208 and extending through the C terminus, encompassing the catalytic domain and adjacent structural elements. In contrast, higher predicted error was observed in the N-terminal region of TET2, consistent with increased conformational flexibility and reduced structural constraint in these regions. Predicted structures were further examined for interfacial features, including hydrogen bonding, hydrophobic contacts, and estimated contact surface Area, using UCSF ChimeraX (https://www.cgl.ucsf.edu/chimerax/). Structural visualizations were generated from representative high-confidence models for figure preparation. AlphaFold model demonstrates substantially improved accuracy over many previous specialized tools: far greater accuracy for protein–ligand interactions compared with state-of-the-art docking tools, much higher accuracy for protein–nucleic acid interactions compared with nucleic-acid-specific predictors and substantially higher antibody–antigen prediction accuracy compared with AlphaFold-Multimer v.2.3,52 showing high-accuracy modeling across biomolecular space is possible within a single unified deep-learning framework. Mouse and human TET2 O-GlcNAcylation sites Putative O-GlcNAcylation sites within mouse and human TET2 sequences were predicted using two computational tools: YinOYang v1.2 (services.healthtech.dtu.dk/services/YinOYang-1.2/) and O-GlcNAcPRED-DL (https://www.djangoproject.com).47 YinOYang employs a neural network trained on 40 experimentally validated O-GlcNAcylation acceptor sites and assesses sequence context and surface accessibility to estimate the likelihood of O-GlcNAcylation modification. Prediction confidence was categorized based on threshold scores defined by the program. For enhanced prediction sensitivity and accuracy, O-GlcNAcPRED-DL, a deep learning–based approach, was also applied. This tool integrates four protein sequence encoding schemes and leverages an ensemble of convolutional neural networks (CNN) and bidirectional long short-term memory (BiLSTM) networks to generate weighted output probabilities, enabling robust identification of potential O-GlcNAcylation sites. Predicted O-GlcNAcylation Ser and Thr sites in mouse and human TET2 were visualized as lollipops. O-GlcNAcylation scores are indicated by lollipops stick height and head size. Red lollipop heads denote predicted O-GlcNAcylation sites at ≥ 0.8 threshold. This is significantly higher than the 0.5 and 0.6 that have been generally used as thresholds for predicted protein PTM, including O-GlcNAcylation.46–48 Sample size and randomization Sample size in all experiments is reported in main and supplementary figure legends. In each experiment, at least 5 mice per group were used to ensure proper biological replicates. Sample size calculations were performed using power analysis, which accounts for effect size, standard deviation, type 1 error and 80% power in a two-sample t test with a 5% significance level (two-sided test). G power software v3.1.9.7 was used for these calculations. Replication: Biological replicates were used in all experiments. Randomization: After matching for sex and age, AicdaCreTet2+/+, AicdaCreTet2fl/fl, AicdaCreTet3+/+, AicdaCreTet3fl/fl, AicdaCreOgt+/+, AicdaCreOgtfl/fl, Tcrβ−/− Tcrδ−/− and humanized THX mice were randomly assigned to appropriate groups. Blinding Experimenters were blinded to group allocation for both data collection and analysis whenever possible. They were not blinded to group allocation during experimental sample collection. Data exclusion No data were excluded from analysis in vivo and in vitro. Bioinformatics FASTQ files for each sample were generated by demultiplexing with Bcl2fastq2 v2.20. Sequencing reads from FASTQ files underwent quality control by trimming adaptors with Trim Galore v0.6.10.97 Bioinformatic analysis was initiated with the alignment of paired-end sequenced FASTQ files to the UC Santa Cruz mouse reference genome (mm10) using Bowtie2 v2.4.4 (https://bowtie-bio.sourceforge.net/bowtie2/index.shtml). Subsequently, duplicate reads were identified and removed using PICARD MarkDuplicates v3.2. Filtering of unmapped sequences and low-quality alignments was carried out using Samtools v1.12 (https://www.ncbi.nlm.nih.gov/pubmed/33590861) with parameter –F 1804. Peak detection was performed using MACS2 v2.2.9.1 with parameter –bedpe. For visualization, bigWig files were generated using BamCoverage command from deepTools v3.5.5 (https://deeptools.readthedocs.io/en/3.5.5/). Specific gene locus tracks were visualized through IGV browser v2.16.2 (https://igv.org/app/). Finally, heatmaps were plotted by deepTools v3.5.3 using computeMatrix with skipzeros parameters followed by plotHeatmap with customized color list command – colorList . QUANTIFICARION AND STATISTICAL ANALYSIS All statistical analyses and calculation of p values were performed using Microsoft Excel v16.89 or GraphPad Prism v10.4.2: p < 0.05 (*), p < 0.01 (**), and p < 0.001 (***). p > 0.05 was considered not significant (ns). Unless otherwise indicated, data were analyzed using two-tailed, unpaired Student’s t-tests, assuming independent samples. Error bars represent standard error of the mean (s.e.m.). Numbers of biological replicates are indicated for each experiment in the figure legends. Western blot band intensities (Figure 1B) were quantified using ImageJ software. Intracellular TET2, AID, and BLIMP-1 protein expression levels (measured as MFI) (Figures 1C, 2E, 2G, S10C, and S12E) were compared using two-tailed, unpaired Student’s t-tests unless otherwise noted. Experiments were performed with three independent biological replicates. Data are presented as truncated violin plots or histograms generated using GraphPad Prism v10.4.2. In truncated violin plots, each dot represents one mouse and the upper and lower edges indicate the maximum and minimum data values, respectively; middle line indicates the median. In histograms, each dot represents an individual mouse or one in vitro B cell microculture and bars depict mean ± s.e.m. Mouse and human immunoglobulin concentrations and antibody titers (Figures 2A, 3C, 7A, S8A, S1A, S2A, S3A, S4A, S5D, and S12A) were analyzed using two-tailed, unpaired Student’s t-tests unless otherwise noted. Experiments were performed using at least five independent biological replicates unless otherwise specified. Data are presented as floating bar plots generated using GraphPad Prism v10.4.2. In floating bar plots, each dot represents one mouse or one in vitro B cell microculture. The top and bottom edges indicate the full range of the data. Differences in B cell mRNA transcript expression (Figure 2D) were analyzed using two-tailed, unpaired Student’s t-tests and visualized as heatmap using Microsoft Excel v16.89. In the heatmap, levels of transcripts were depicted using a red-white-blue continuous color scale (red: increased expression, white: baseline expression, blue: decreased expression). Experiments were performed using three biological replicates. Additional analyses of B cell transcript expression (Figures 6F and S5D, S10A, and S11B), including Aicda/AICDA and Prdm1/PRDM1 transcripts, were analyzed using two-tailed, unpaired Student’s t-tests unless otherwise noted and performed in three biological replicates. Data are presented as histograms generated using GraphPad Prism v10.4.2. Each dot represents an individual mouse or one in vitro B cell microculture and bars depict mean ± s.e.m. DNA methylation and hydroxymethylation levels of Aicda/AICDA and Prdm1/PRDM1 regulatory regions (Figures 4B, 6F, 7E, S5A, S10B, S10C, and S11C), as well as O-GlcNAcylated H2B-S112 levels of the same loci (Figures 4G and 6D), were analyzed using two-tailed, unpaired Student’s t-tests unless otherwise specified. All experiments were performed using three independent biological replicates. Quantification of TET2 and OGT recruitment, DNA methylation, hydroxymethylation and histone H2B-S112 O-GlcNAcylation at Aicda/AICDA and Prdm1/PRDM1 regulatory regions are depicted as truncated violin plots generated using GraphPad Prism v10.4.2. In truncated violin plots, data are enriched DNA expressed as relative to input DNA. Each dot represents relative DNA 5mC and 5hmC levels, or the levels of recruited OGT or O-GlcNAc H2B-S112 at Aicda/AICDA and Prdm1/PRDM1 regulatory regions from one in vitro B cell microculture. Top and bottom edges of the plot indicate the maximum and minimum values, respectively, and line indicates median. Flow cytometric analyses of B cell populations, including total GC B cells, class-switched GC B cells, class-switched memory B cells, IgG1+ and IgA+ B cells, plasmablasts and plasma cells (Figures 2C, 2E, 2G, 6C, 6D, and 7E; Figures S1C, S2E, S3D, S3E, S3G, S4A, S5D, S10A, S11A, and S12A), were analyzed using two-tailed, unpaired Student’s t-tests unless otherwise specified, using three independent biological replicates. Intracellular TET2 expression measured as mean fluorescence intensity by flow cytometry (Figures 1A and 1C) was analyzed using the same statistical approach. Quantification of naïve, class-switched, GC, memory B cells, plasmablasts and plasma cells is depicted as violin plots generated using GraphPad Prism v10.4.2. Each dot represents an individual mouse. The top and bottom edges indicate the full range of data; line indicates median. Point-mutation frequencies in mouse V1-72DJH-Cγ1, V1-72DJH-Cγ3 and V1-72DJH-Cγ2b transcripts, as well as human huV3DJH-Cγ1 transcripts (Figures 2F, 3E, 6C, S4C, and S12D), were analyzed using two-tailed, unpaired Student’s t-tests. Data are presented as pie charts, histograms and boxplots generated using Microsoft Excel v16.89 and GraphPad Prism v10.4.2. Pie charts: slices depicted proportions of specific (mouse: V1-72DJH-Cγ1, V1–72DJH-Cγ2b and V1-72DJH-Cγ3; human: huV3DJH-Cγ1) transcripts carrying different numbers of point-mutations; slice gray gradients depict increasing numbers of point-mutations (from 0 to more than 12 per transcript); the overall mutation frequency is listed below each pie Histograms depict means of total, S and R mutation frequencies (change/base) in FR1, CDR1, FR2, CDR2 and FR3 of specific transcripts. Boxplots depicts overall point-mutation frequency in specific transcripts. In boxplots, data points are displayed as individual dots, each dot representing a single sequence. Box includes the 2nd and 3rd quartiles (top 75% to bottom 25%) of data, line indicates median and cross indicates geometrical mean of data, whiskers show the spread of data (95%–5%) and dots above and below whiskers are outliers. Supplementary Material Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2026.117370.