PDGFA-associated protein 1 protects mature B lymphocytes from stress-induced cell death and promotes antibody gene diversification

The diversity of our Ig gene repertoire is the result of antibody diversification reactions occurring at different stages of B lymphocyte development (Dudley et al., 2005; Methot and Di Noia, 2017). Developing B cells in the bone marrow randomly assemble different gene segments (known as variable, V; diversity, D; and joining, J, genes) at the Ig heavy (Igh) and light (Igl) chain loci via V(D)J recombination (Roth, 2014; Dudley et al., 2005). This process generates unique antibody gene receptors with the potential to collectively recognize a formidable number of antigens. Mature B cells further diversify their Ig genes in the periphery via somatic hypermutation (SHM) and class switch recombination (CSR; Pavri and Nussenzweig, 2011; Methot and Di Noia, 2017). SHM introduces point mutations into the variable V(D)J region of the Ig genes to generate higher-affinity variants. CSR recombines the Igh constant (C) regions to replace the C portion of the IgM heavy chain with one of the alternative isotypes (IgG, IgA, and IgE), thus diversifying the Ig effector function. SHM and CSR are crucial to mount protective humoral responses, as evidenced by primary human immunodeficiency syndromes that are caused by defects in these reactions (Durandy et al., 2013).

SHM and CSR are both dependent on the B cell–specific enzyme activation-induced cytidine deaminase (AID; Revy et al., 2000; Muramatsu et al., 2000). AID deaminates cytosine residues to uracil in single-stranded DNA stretches at the variable regions of both Igh and Igl loci during SHM, and within special recombining elements (switch [S] regions) of the Igh during CSR (Bransteitter et al., 2003; Chaudhuri et al., 2003; Petersen-Mahrt et al., 2002; Pham et al., 2003; Sohail et al., 2003; Ramiro et al., 2003; Dickerson et al., 2003; Matthews et al., 2014; Methot and Di Noia, 2017). The resulting U:G mismatches are differentially processed to generate either mutations in the variable regions or DNA breaks in the S regions (Peled et al., 2008; Methot and Di Noia, 2017; Matthews et al., 2014). AID expression is induced when resting mature (naïve) B cells are activated by the antigen and T cell interactions (Zhou et al., 2003; Sayegh et al., 2003; Gonda et al., 2003; Cunningham et al., 2004; Dedeoglu et al., 2004; Muramatsu et al., 1999). Antigen stimulation reprograms naïve B cells to exit the quiescent state, expand their cellular biomass, and undergo a proliferative burst within transient and anatomically distinct structures in secondary lymphoid organs called germinal centers (GCs; Cyster and Allen, 2019; Victora and Nussenzweig, 2012). AID expression peaks in GC B cells (Cattoretti et al., 2006; Crouch et al., 2007; Roco et al., 2019). The GC reaction represents the end stage of B cell development, as GC B cells differentiate into memory B cells or long-lived plasma cells that secrete high-affinity antibodies.

The integrated stress response (ISR) is a homeostatic program activated by a variety of physiological and pathological stresses to promote cellular recovery (Ron and Walter, 2007; Pakos-Zebrucka et al., 2016). These stresses include both cell-intrinsic and -extrinsic stimuli, such as ER stress, mitochondrial dysfunction, hypoxia, and amino acid deprivation (Quirós et al., 2017; Harding et al., 1999, 2003; Dever et al., 1992; Rzymski et al., 2010; Ye et al., 2010). All forms of stress converge into the phosphorylation of the α subunit of the eukaryotic translation initiation factor 2 (eIF2α) on serine 51 (Donnelly et al., 2013). This event causes a reduction in global protein synthesis while allowing the preferential translation of few selected genes including the ISR effector activating transcription factor 4 (Atf4; Hinnebusch, 2000; Harding et al., 2000; Scheuner et al., 2001; Lu et al., 2004). Atf4 induces the transcriptional up-regulation of stress-responsive genes and rewires cell metabolism toward the recovery of cellular homeostasis (Harding et al., 2003). The inhibition of general protein translation in the early stage of the ISR is followed by a later phase of translational recovery, which restores protein synthesis once the stress is resolved to support cell survival (Brostrom and Brostrom, 1998; Novoa et al., 2003; Kojima et al., 2003; Marciniak et al., 2004; Brostrom et al., 1989; Ma and Hendershot, 2003).

Although the ISR is an adaptive program meant to restore cellular homeostasis and promote cell survival, under conditions of severe or prolonged stress, it induces cell death by activating proapoptotic pathways (Zou et al., 2008; Puthalakath et al., 2007; Teske et al., 2013; Wang et al., 2009; Ohoka et al., 2005; Gupta et al., 2012; Hiramatsu et al., 2014; Marciniak et al., 2004). Furthermore, the sustained protein synthesis during chronic stress causes proteotoxicity and leads to cell death (Han et al., 2013; Krokowski et al., 2013). Here, we identified Pdap1 as an essential regulator of mature B cell physiology that functions by countering chronic activation of the ISR. Pdap1 ablation in mature B cells causes sustained expression of Atf4, long-term induction of the ISR transcriptional program, and cell death. Furthermore, Pdap1 is essential for efficient induction of AID expression and physiological levels of CSR and SHM.

To identify novel modulators of mature B cell physiology, we performed loss-of-CSR screens via somatic gene targeting by CRISPR-Cas9 in the B cell lymphoma line CH12. CH12 cells share several properties of resting B cells and have been used to study key aspects of late B cell differentiation, including B cell activation, antibody diversification by CSR, and immunoglobulin secretion (Arnold et al., 1983; Stockdale et al., 1987; Wiest et al., 1990; Corley et al., 1985; Bishop and Haughton, 1986; Nakamura et al., 1996; Ovnic and Corley, 1987; Fagone et al., 2007; Gass et al., 2002; Dufort et al., 2014; LoCascio et al., 1984a, 1984b). Upon cytokine stimulation, CH12 cells express AID and undergo CSR to IgA with high efficiency (Nakamura et al., 1996). CRISPR-Cas9–mediated deletion of factors required for CSR reduces the efficiency of isotype switching in this cell line (Delgado-Benito et al., 2018). Among all tested candidates, targeting of Pdap1 resulted in a considerable reduction of CSR (Fig. 1, A and B). To confirm the results of the loss-of-CSR screen, we generated Pdap1-deficient CH12 clonal derivative cell lines, which included both indel knockout (Pdap1^−/−) and in-frame deletion mutant (Pdap1^mut) clones (Fig. 1, C and D). In agreement with the CSR defect observed in bulk CH12 cultures, CSR was impaired in Pdap1-deficient clonal derivatives (Fig. 1 E). We concluded that Pdap1 supports efficient CSR in CH12 cells.

Pdap1 is a 28-kD phosphoprotein highly conserved in vertebrates. It was originally identified as a casein kinase II substrate and a weak interactor of platelet-derived growth factor A (PDGF-A; Shen et al., 1996; Fischer and Schubert, 1996). More recently, Pdap1 was described as an RNA-binding protein in several RNA–protein interactome studies (Castello et al., 2012, 2016; Trendel et al., 2019; Baltz et al., 2012; Iadevaia et al., 2020). However, the precise cellular functions of Pdap1 and its involvement in adaptive immunity are unknown.

To verify Pdap1 involvement in CSR in vivo and elucidate the underlying mechanism, we generated a mouse model bearing a conditional Pdap1^F allele and bred it to Cd19^Cre/+ mice to specifically ablate Pdap1 expression at the early stages of B cell differentiation (Rickert et al., 1997; Fig. S1, A and B). B cell development was largely unaffected in Pdap1^F/FCd19^Cre/+ mice (Fig. S1, C and D). Although incomplete Cd19^Cre-mediated recombination of Pdap1^F in B cell progenitors could still mask a potential phenotype, these results suggest that Pdap1 is dispensable for V(D)J recombination and early B cell differentiation. However, we observed a pronounced reduction in the number of splenic resting mature B cells (Fig. S1 E). To confirm the intrinsic CSR defect caused by Pdap1 deficiency, we isolated resting splenocytes from Pdap1^F/FCd19^Cre/+ mice and monitored their capability to undergo CSR upon in vitro stimulation. We assessed CSR under conditions that induce switching to IgG1, IgG3, IgG2b, and IgA and found that Pdap1-deficient B cells displayed reduced levels of CSR for all tested isotypes (Fig. 2, A–D). Western blot (WB) analysis confirmed the near-complete abrogation of Pdap1 expression in these cells (Fig. 2 E). Altogether, these data indicate that Pdap1 is required for efficient CSR in primary B cells irrespective of the stimulation condition.

CSR is dependent on cell proliferation (Hasbold et al., 1998, 1999; Hodgkin et al., 1996; Deenick et al., 1999). To determine if the reduced CSR efficiency of Pdap1-deficient B cells is caused by an underlying defect in cell proliferation, we monitored the proliferation capabilities of splenocyte cultures by cell tracking dye dilution. The CellTrace Violet dilution profiles of Pdap1^F/FCd19^Cre/+ B cells were indistinguishable from the control counterparts under all stimulation conditions (Fig. S2, A–D). Furthermore, class switching was reduced in Pdap1-deficient B cells independently of the number of cell divisions (Fig. S2 E). We concluded that Pdap1 is dispensable for B cell proliferation, and that the CSR defect of Pdap1^F/FCd19^Cre/+ B cells is not due to reduced proliferation capabilities.

The efficiency of CSR is directly linked to the levels of AID expression (Dorsett et al., 2008; Takizawa et al., 2008; Teng et al., 2008). Furthermore, AID targeting is dependent on noncoding transcription across the S regions (germline transcription; GLT), which exposes single-stranded DNA stretches that are the substrate of AID-mediated deamination (Chaudhuri et al., 2003; Dickerson et al., 2003; Ramiro et al., 2003). Therefore, we monitored Aicda mRNA and GLT levels in activated B cells by quantitative RT-PCR (qPCR) analysis. We found that Aicda transcript levels were consistently reduced in Pdap1^F/FCd19^Cre/+ B cells compared with controls across all stimulation conditions (Fig. 3 A). Transcription of donor Sμ region was not affected by Pdap1 deletion (Fig. 3 B), whereas acceptor S region transcription exhibited a varied phenotype, with reduced levels of GLTγ1 and GLTα, and minimally affected or unaltered expression for GLTγ3 and GLTγ2b, respectively (Fig. 3 C). Analogously, analysis of Aicda and germline transcripts in Pdap1-deficient CH12 cell lines showed reduced Aicda mRNA levels but unaffected Sμ and Sα region transcription (Fig. S3, A–C). In agreement with the reduction of Aicda mRNA, Pdap1-deficient B cells expressed lower levels of AID protein upon activation compared with control cells (Fig. 3 D).

AID-mediated deamination in the donor Sμ and recombining acceptor S region leads to the formation of DNA double-strand breaks (DSBs), which are ultimately repaired by end joining to ensure productive CSR events (Matthews et al., 2014; Boboila et al., 2012). The reduced AID expression and isotype-specific GLT phenotypes of Pdap1-deficient B cells converge into a pre-DSB defect, since DSB formation at S regions relies on both AID levels and efficient targeting. These results, however, did not exclude the possibility that Pdap1 might play an additional role during CSR downstream break formation, for example in repair of S region DSBs. To test this possibility, we overexpressed AID in Pdap1-deficient B cells and monitored the levels of CSR (Fig. 3 E). To this end, we transduced control and Pdap1^F/FCd19^Cre/+ splenocytes with a construct expressing AID fused to the hormone-binding domain of the modified estrogen receptor (AID-ER). Provision of 4-hydroxytamoxifen (4-HT) induces the translocation of the cytoplasm-accumulated AID-ER fusion protein into the nucleus and initiates CSR. We found that 4-HT addition rescued the CSR defect of Pdap1^F/FCd19^Cre/+ cells to levels equivalent to AID-ER–reconstituted Aicda^−/− splenocytes in cultures stimulated to undergo class switching to IgG2b (Fig. 3 E), which exhibited unaffected GLTγ2b (Fig. 3 C). In contrast, the rescue of CSR was not significant in cells activated to switch to IgG1, with severely reduced GLTγ1 in the absence of Pdap1 (Fig. 3, C and E), and therefore impaired in AID targeting even under conditions of increased AID levels. Altogether, these data indicate that the CSR defect of Pdap1-deficient B cells is caused by impaired formation of AID-induced DSBs, which is caused primarily by reduced AID expression, with the additional contribution of impaired GLT in an isotype-specific manner.

Pdap1 has been recently identified in RNA-protein capture experiments as an RNA binding protein (Castello et al., 2012, 2016; Trendel et al., 2019; Baltz et al., 2012; Iadevaia et al., 2020). Therefore, we considered the possibility that Pdap1 might contribute to the posttranscriptional regulation of AID expression. Because Pdap1 is localized primarily in the cytoplasm (Fig. S4), we first tested whether Pdap1 controls the degradation of Aicda mRNA. To do so, we treated Cd19^Cre/+ and Pdap1^F/FCd19^Cre/+ splenocyte cultures with Actinomycin D to inhibit the de novo transcription of AID and measured Aicda mRNA decay by qPCR. We found that depletion of Pdap1 did not affect the kinetics of Aicda mRNA degradation (Fig. 4, A and B). This result also indirectly shows that Pdap1 deficiency does not alter decapping or deadenylation of Aicda transcripts since a defect in any of these processes would result in accelerated decay. We concluded that Pdap1 is dispensable for Aicda transcript stability. Next, we monitored Aicda transcript splicing. Several splice variants for AID have been identified, with the full-length transcript being the only splice isoform encoding a functional AID protein (van Maldegem et al., 2009, 2010; Sala et al., 2015). To assess whether Pdap1 modulates Aicda splicing, we performed RNA next-generation sequencing (RNA-seq) of activated B cells induced to undergo CSR to either IgG1 or IgG2b and analyzed the pattern of Aicda exon usage. We found no difference in the profile of Aicda exon usage between Cd19^Cre/+ and Pdap1^F/FCd19^Cre/+ B cells (Fig. 4 C). Furthermore, primary Aicda transcript levels in Pdap1^F/FCd19^Cre/+ splenocytes were reduced compared with Cd19^Cre/+ cells to an extent similar to that observed for the full-length spliced product (Figs. 4 D and 3 A). We concluded that ablation of Pdap1 does not affect Aicda mRNA splicing. Altogether, these findings indicate that Pdap1 supports physiological levels of AID expression, and that the CSR defect of Pdap1-deficient B cells is caused, at least in part, by impaired induction of AID expression following activation rather than a defect in Aicda posttranscriptional regulation.

To define the mechanism responsible for the defective induction of AID expression in Pdap1-deficient cells, we compared the transcriptome profiles of Cd19^Cre/+ and Pdap1^F/FCd19^Cre/+ splenocytes stimulated with either LPS and IL-4 or LPS, B cell–activating factor (BAFF), and TGFβ (Table S1). The use of two stimulation conditions allowed us to zoom in common differentially regulated pathways and added a temporal dimension to the experiment, since LPS-BAFF-TGFβ–stimulated splenocytes proliferate faster than LPS-IL-4–activated ones (Fig. S2 F). To identify differentially expressed genes, we set the significance level of false discovery rate (FDR) to <0.05 (Fig. 5 A). The number of differentially regulated genes was higher in LPS-IL-4–stimulated cultures than in LPS-BAFF-TGFβ–activated cells (1,227 versus 173). Furthermore, the number of genes up-regulated in Pdap1^F/FCd19^Cre/+ was considerably higher than the down-regulated ones in LPS-BAFF-TGFβ–stimulated cultures (124 up- versus 49 down-regulated) but was evenly distributed between the two categories in the LPS-IL-4 stimulation condition (561 up- versus 666 down-regulated).

Pathway enrichment analysis of the down-regulated genes in Pdap1^F/FCd19^Cre/+ B cells in the LPS-IL-4 dataset identified the protein translation–related categories “ribosome,” “ribosome biogenesis in eukaryotes,” and “RNA transport” as the most significant categories (Fig. 5 B). Furthermore, analysis of individual genes showed that the most significantly down-regulated ones under both stimulations were factors essential for CSR and/or genes highly expressed or induced in activated B cells (e.g., Aicda, Batf, Ccl22, Ighg3, and Tmem176b; Fig. 5 C and Table S1). In contrast, the top up-regulated categories in both datasets were related to metabolic pathways of amino acid and aminoacyl-tRNA biosynthesis (Fig. 5 B). Among the most significantly up-regulated genes, we found key enzymes for the biosynthesis of asparagine (Asns), cysteine (Cth), and serine (Psat1, Psph), glutathione metabolism (Chac1), amino acid transporters (Slc1a4, Slc6a9), and several other metabolic and cytoprotective genes (Fig. 5 C, Fig. S5 A, and Table S1).

These data indicate that, in addition to the inhibition of key factors of the CSR program, Pdap1 deficiency in activated B cells is accompanied by transcriptional changes that reflect initial repression of global protein synthesis, up-regulation of cellular stress response genes, and metabolic rewiring to aid translational recovery. These changes represent key aspects of the transcriptional program of the ISR (Ron and Walter, 2007; Pakos-Zebrucka et al., 2016). To confirm the activation of the ISR in Pdap1-deficient B cells, we monitored the kinetics of eIF2α phosphorylation and Atf4 expression in resting and in vitro–activated splenocytes. The eIF2α-Atf4 pathway is normally suppressed in splenic B cells and is induced only at later time points after activation, when splenocyte cultures exhibit a general fitness decline (Fig. 5 D and Fig. S5 B, control cells; Zhu et al., 2019). In contrast to control cells, resting B cells from Pdap1^F/FCd19^Cre/+ spleens displayed a marked phosphorylation of eIF2α and detectable levels of Atf4 protein, thus indicating a baseline activation of the ISR (Fig. 5 D and Fig. S5 B). The ER stress-activated eIF2α kinase PKR-like ER kinase (Perk) also showed autophosphorylation-dependent activation in resting splenocytes (Fig. 5 D and Fig. S5 B). eIF2α phosphorylation was suppressed following activation and was noticeable only in late-stage cultures when eIF2α levels increased considerably in both control and Pdap1-deficient cells (Fig. 5 D and Fig. S5 B). In agreement with a previous report (Zhu et al., 2019), Atf4 was markedly expressed only at late time points after activation in control splenocytes (Fig. 5 D). In contrast, Atf4 protein levels in Pdap1^F/FCd19^Cre/+ cells increased considerably during the first 24 h after activation from the baseline expression observed in resting B cells (Fig. 5 D). Finally, up-regulation of Atf4 expression in the absence of Pdap1 occurred not only at the translation level but also via transcriptional induction (Fig. 5 C and Fig. S5 A). We concluded that Pdap1 deficiency results in up-regulation of the ISR core component Atf4. Altogether, these data indicate that Pdap1 ablation in mature B cells induces Atf4-dependent expression of stress response genes and down-regulation of the CSR program.

The ISR is an adaptive pathway meant to restore cellular homeostasis following cell intrinsic and extrinsic stresses (Harding et al., 1999, 2000, 2003; Brostrom et al., 1996; Dever et al., 1992; Ron and Walter, 2007; Pakos-Zebrucka et al., 2016). However, when the stress is severe in either intensity or duration and overwhelms the response adaptive capacity, the ISR activates the apoptotic cell death program (Pakos-Zebrucka et al., 2016). Furthermore, down-regulation studies have implicated Pdap1 in cell survival and apoptosis resistance of cancer cell lines (Sharma et al., 2016; Weston et al., 2018). Therefore, we tested whether activation of the stress response in the absence of Pdap1 would induce apoptosis in B cells. To do so, we monitored caspase activation in splenocytes by CaspGLOW staining. Caspase activity increased after stimulation and was higher in Pdap1-deficient cells compared with controls at 24 and 48 h after activation (Fig. 6 A). Accordingly, despite the unaffected proliferation capability (Fig. S2), the number of live cells in Pdap1^F/FCd19^Cre/+ splenocyte cultures was reduced 48 and 72 h after activation (Fig. 6 B). Finally, no differences were detected in mitochondrial mass or respiration capacity compared with controls (Fig. 6 C and Fig. S5 C), thus indicating that activation of the ISR is not caused by mitochondrial disfunction. However, Pdap1^F/FCd19^Cre/+ cultures displayed an increased proportion of cells with low mitochondrial membrane potential at early time points after activation (24 and 48 h; Fig. 6 D). Hence, a fraction of Pdap1-deficient B cells may undergo mitochondrial depolarization, an event usually associated with apoptosis induction (Gottlieb et al., 2003).

Next, we tested whether Pdap1 deletion would affect the survival of activated B cells in vivo. To this end, we measured the percentage of GC B cells in Peyer’s patches of Cd19^Cre/+ and Pdap1^F/FCd19^Cre/+ mice. Peyer’s patches are specialized secondary lymphoid tissues that line the wall of the small intestine. Because of their chronic exposure to an enormous variety of food- and microbiome-derived antigens, Peyer’s patches display continual GC activity and are key to the induction of mucosal IgA antibody responses (Reboldi and Cyster, 2016). We found that the percentage of GC B cells was reduced in Peyer’s patches of Pdap1^F/FCd19^Cre/+ mice compared with the control group (Fig. 6 E). In agreement with the CSR defect exhibited by Pdap1^F/FCd19^Cre/+ splenic B cells following activation in vitro (Fig. 2), the percentage of switched IgA GC B cells was considerably decreased in Pdap1^F/FCd19^Cre/+ mice (Fig. 6 E). Altogether, these data suggest that Pdap1 is required to support survival of activated B cells.

Plasma cell differentiation represents the terminal phase of B cell development and is regulated by a transcriptional program that represses B cell identity while promoting the expression of plasma cell signature genes (Nutt et al., 2015; Shi et al., 2015; Minnich et al., 2016). This process is accompanied by massive expansion of the ER and the up-regulation of molecular chaperone and folding enzyme expression. The expansion of the secretory network is essential to accommodate the demands of increased immunoglobulin synthesis and secretion and is dependent on the activation of the unfolded protein response (UPR; Zhang et al., 2005; Gass et al., 2002, 2008; Iwakoshi et al., 2003; van Anken et al., 2003; Shaffer et al., 2004; Taubenheim et al., 2012; Todd et al., 2009). The UPR is a ubiquitous signaling network that senses and responds to the accumulation of misfolded proteins in the ER (Ron and Walter, 2007; Walter and Ron, 2011; Schröder and Kaufman, 2005). This response is mediated by three ER-resident proteins which, in addition to Perk, comprise the inositol-requiring protein kinase/endoribonuclease-1 (Ire1) and Atf6 (Tirasophon et al., 1998; Wang et al., 1998; Haze et al., 1999, 2001). The UPR ultimately reestablishes protein homeostasis by integrating Perk-eIF2α–dependent attenuation of global protein synthesis with Ire1- and Atf6-mediated transcription of factors and enzymes that increase the protein folding and degradation capabilities of the ER.

The Ire1 and Atf6 branches of the UPR are activated during normal differentiation of naïve B cells into plasma cells, and the Ire1 arm is essential for the expansion of their secretory network (Zhang et al., 2005; Gass et al., 2002, 2008; Iwakoshi et al., 2003; van Anken et al., 2003; Aragon et al., 2012). In contrast, the Perk arm of the UPR is dispensable for plasma cell development (Zhang et al., 2005; Gass et al., 2008). Furthermore, this pathway is suppressed during differentiation of naïve B cells into plasma cells (Ma et al., 2010; Gass et al., 2008; Zhang et al., 2005), and unrestricted Perk-eIF2α-Atf4 signaling in activated B cells blocks the formation of plasma cells (Zhu et al., 2019). Given the up-regulated expression of Atf4 in Pdap1-deficient B cells, we considered the possibility that plasma cell development might be impaired in the absence of Pdap1. To this end, we analyzed the plasma cell compartment of Pdap1^F/FCd19^Cre/+ mice. We found that the percentage and number of plasma cells (CD138⁺TACI⁺) in bone marrow and spleen of unimmunized mice was similar between control and Pdap1^F/FCd19^Cre/+ mice (Fig. 7 A).

Activation of naïve B cells with specific combinations of stimuli in vitro induces differentiation into plasmablasts. Plasmablasts are proliferating antibody-secreting cells (ASCs) whose differentiation is driven by the same transcriptional program responsible for plasma cell development (Nutt et al., 2015). Therefore, we tested the formation of plasmablasts (B220⁺CD138⁺) in vitro under the same stimulation conditions used for our previous analyses. We observed no significant difference in the percentage of plasmablasts between controls and Pdap1^F/FCd19^Cre/+ cultures (Fig. 7 B). Accordingly, we did not detect any apparent change in the activation of the plasmablast and plasma cell gene signature between the two genotypes in the RNA-seq analysis (Table S1). We concluded that Atf4 up-regulation in activated Pdap1^F/FCd19^Cre/+ B cells does not inhibit the generation of plasmablasts and plasma cells, and that Pdap1 is not required for the B cell developmental program controlling plasmablast and plasma cell formation and identity.

The ER expansion and increase in protein processing and folding capabilities of plasma cells are regulated by the transcription factor Xbp1-s (spliced Xbp1; Shaffer et al., 2004). Xbp1 deficiency in B cells does not interfere with the development of plasma cells but impairs immunoglobulin secretion because of defective expansion of the secretory pathway (Hu et al., 2009; Todd et al., 2009; Taubenheim et al., 2012; Shaffer et al., 2004). Xbp1-s is expressed as a result of an unconventional splicing event mediated by Ire1 that removes a 26-nt segment from Xbp1 mRNA (unspliced Xbp1 transcript, Xbp1-u), thus changing the transcript reading frame in the resulting spliced Xbp1 transcript (Xbp1-s; Fig. 7 C; Yoshida et al., 2001). Xbp1-s translocates into the nucleus and activates the transcription of genes responsible for the expansion of the secretory pathway (Shaffer et al., 2004; Tellier et al., 2016). We found no difference in Xbp1 mRNA splicing and Xbp1-s expression levels between control and Pdap1^F/FCd19^Cre/+ splenocyte cultures (Fig. 7, C and D). These results suggest that the transcriptional program controlling ER network expansion in plasma cells is not affected by Pdap1 deletion. Finally, ELISpot analysis did not detect any significant difference in the number or spot size of ASCs between the two groups (Fig. 7 E), thus indicating that antibody production is not affected by Pdap1 deficiency. We concluded that Pdap1 is dispensable for the differentiation programs that establish plasma cell identity and function under steady-state conditions.

Aicda is a haploinsufficient gene, and Aicda^+/− mice exhibit defects in both antibody diversification reactions initiated by AID, CSR, and SHM (McBride et al., 2008; Sernández et al., 2008; Takizawa et al., 2008). Given the reduced levels of AID expressed by Pdap1^F/FCd19^Cre/+ B cells following activation, we considered the possibility that Pdap1 depletion might also affect SHM. To test this hypothesis, we sequenced the intronic regions downstream of the J_H4 and J_K5 elements in GC B cells sorted from Peyer’s patches of aged Pdap1^F/FCd19^Cre/+ mice (Jolly et al., 1997). Because of their chronic exposure to microbial antigens, GC B cells in Peyer’s patches accumulate a high number of mutations over time (González-Fernández et al., 1994). Accordingly, we found that GC B cells from control mice displayed a high mutation frequency at both J_H4 and J_K5 intronic regions (Fig. 8, A–D). The vast majority of sequences contained mutations (86%, 79 of 92 for J_H4, and 83%, 104 of 125 for J_K5), and a considerable portion of B cells were heavily mutated (Fig. 8, E and F). In contrast, Pdap1^F/FCd19^Cre/+ GC B cells exhibited a significantly lower mutation frequency at both J_H4 and J_K5 introns (Fig. 8, A–D). The distribution of mutations per sequence was skewed toward an increase in the proportion of sequences bearing zero or fewer than five mutations (50%, 54 of 107 for J_H4, and 65%, 83 of 127 for J_K5), whereas the number of highly mutated clones was concomitantly reduced (Fig. 8, E and F). In agreement with a role of Pdap1 in promoting AID expression rather than the processing of AID-induced lesions, the profile of mutations was not affected by Pdap1 deletion (Fig. 8, G and H).

Altogether, these findings indicate that ablation of Pdap1 impairs the efficiency of SHM of heavy and light chain Ig loci. Therefore, Pdap1 is required to support both antibody diversification reactions occurring in mature B cells, namely CSR and SHM.

In this study, we report the identification of a novel factor required for mature B cell homeostasis and function. Pdap1-deficient B cells develop normally but display reduced cell viability at the mature stage both in vitro and in vivo. Furthermore, they fail to efficiently induce AID expression and cannot support physiological levels of CSR and SHM.

Pdap1 deficiency in mature B cells phenocopies key aspects of the cellular response to chronic stress. Stress-induced eIF2α phosphorylation drives the inhibition of global protein synthesis, while allowing the preferential translation of stress-responsive genes like Atf4 (Harding et al., 2000; Hinnebusch, 2000; Scheuner et al., 2001; Lu et al., 2004). The transcriptional program induced by Atf4 comprises cytoprotective genes as well as the up-regulation of biosynthetic pathways to reprogram cellular metabolism and aid homeostasis recovery. This anabolic program encompasses metabolic pathways of amino acid biosynthesis and the expression of amino acid transporter and aminoacyl-tRNA synthetase genes, and supports recovery of translation under conditions of acute stress (Harding et al., 2003; Rendleman et al., 2018; Quirós et al., 2017; Krokowski et al., 2013; Han et al., 2013). However, during chronic stress, Atf4-dependent increase in protein synthesis leads to proteotoxicity, oxidative stress, and cell death (Han et al., 2013; Krokowski et al., 2013). Pdap1-deficient B cells express Atf4 and exhibit the same transcriptional signature of Atf4-mediated translational recovery (Fig. 5, Fig. S5, and Table S1). Furthermore, the expression of this anabolic program correlates with increased apoptosis of Pdap1^F/FCd19^Cre/+ B cells. Therefore, our data are consistent with a model in which ablation of Pdap1 in activated B cells leads to cell death via induction of an Atf4-dependent chronic stress response. In addition, resting B cells from Pdap1-deficient mice displayed phosphorylation of Perk and its downstream target eIF2α as well as Atf4 expression, thus indicating constitutive activation of the Perk-mediated ISR pathway. This result is in agreement with the reduced number of resting B cells isolated from spleens of Pdap1^F/FCd19^Cre/+ (Fig. S1 E) and indicates that Pdap1 deficiency is also detrimental in naïve B cells. Therefore, Pdap1 is required to protect B lymphocytes from stress-induced cell death under both resting and activated conditions.

Interestingly, while Atf4 expression increased steadily over time after activation in Pdap1-deficient B cells, phosphorylation of Perk and eIF2α was no longer detectable. This observation has two important implications. First, the up-regulation of Atf4 expression in Pdap1-ablated cells is uncoupled from Perk-eIF2α phosphorylation following activation. This result is in agreement with the identification of a phospho-eIF2α–independent regulation of Atf4 expression observed during conditions of prolonged cellular stress (Guan et al., 2014). This regulation is likely under the control of mTORC1, which is activated in B cells after stimulation (Powell et al., 2012; Limon and Fruman, 2012; Iwata et al., 2017), and has been shown to promote Atf4 translation in a phospho-eIF2α–independent manner in other cellular contexts (Park et al., 2017; Ben-Sahra et al., 2016). Second, the dephosphorylation of Perk and eIF2α occurs at early time points after activation. This observation implies the existence of an active mechanism operating immediately upon B cell activation that physiologically inhibits the Perk-eIF2α signaling cascade. In line with this point, although mature B cells retain the potential to activate all three arms of the UPR in response to pharmacological triggers of the pathway, the Perk branch is suppressed during normal differentiation of activated B cells into plasma cells (Ma et al., 2010; Gass et al., 2008; Zhang et al., 2005). Moreover, nonphysiological Perk-eIF2α-Atf4 signaling after B cell activation even blocks the formation of plasma cells (Zhu et al., 2019). In Pdap1^F/FCd19^Cre/+ splenocytes, the Perk-eIF2α-Atf4 pathway is already active under resting conditions, and the increased loss of mitochondrial membrane potential and apoptosis were observed at early, but not late, stages after activation. The earlier kinetics of Atf4 induction in our experimental conditions would explain why Pdap1 deficiency leads to cell death of naïve and early activated B cells but does not affect the plasma cell compartment. It is important to note that the analysis of plasma cell differentiation and function was performed under steady-state conditions; therefore, we cannot completely exclude the possibility that Pdap1 deficiency might affect plasma cell differentiation following immunization.

The activation of the Atf4-dependent stress response in Pdap1^F/FCd19^Cre/+ B cells was accompanied by down-regulation of key factors driving antibody diversification, including AID (Muramatsu et al., 2000; Revy et al., 2000) and the transcription factor Batf, which has been shown to directly control AID expression and other aspects of the CSR program (Ise et al., 2011). Defective induction of Aicda expression in activated Pdap1-deficient B cells was common to all stimulation conditions (Figs. 3 A and 5 C) and likely represents the dominant cause for the reduced CSR and SHM efficiency. Aicda is an haploinsufficient gene, and the efficiency of CSR is directly proportional to AID levels. Mice heterozygous for a null Aicda allele display defects in both CSR and SHM (McBride et al., 2008; Sernández et al., 2008; Takizawa et al., 2008). Analogously, Pdap1 deficiency reduced Aicda expression to approximately half of WT levels in activated B cells and is associated with a similar reduction in the efficiency of both antibody diversification reactions. Furthermore, overexpression of AID in LPS-BAFF-TGFβ–stimulated splenocytes, which did not exhibit any additional defect in GLT, rescued the IgG2b CSR defect of Pdap1^F/FCd19^Cre/+ B cells to a considerable extent.

Pdap1 might control activation of the ISR and the antibody diversification program via independent mechanisms. We found that Pdap1 is localized primarily in the cytoplasm in different cell types. Considering this observation and that Pdap1 was recently described as a RNA-binding protein (Castello et al., 2012, 2016; Iadevaia et al., 2020; Trendel et al., 2019; Baltz et al., 2012), it is tempting to hypothesize that it might play a role in mRNA turnover. In regard to AID, however, we showed that Pdap1 promotes its efficient expression but is dispensable for Aicda transcript stability. Therefore, a potential involvement of Pdap1 in the regulation of AID levels via control of mRNA turnover would occur indirectly, e.g., via modulation of the expression of intermediate factors. Although we cannot exclude the possibility that Pdap1 might regulate the expression of AID independently from its protective role against chronic ISR activation, we favor the hypothesis that the two processes are tightly linked. It is intriguing to speculate that the sustained induction of the ISR in Pdap1^F/FCd19^Cre/+ B cells might actively interfere with the CSR and SHM programs. In this respect, limiting AID expression under conditions of cellular stress might represent a mechanism to alleviate the additional burden imposed by AID-induced genotoxic stress on activated B cells. Analysis of humoral responses upon genetic deletion of ISR components on a Pdap1-deficient background would provide the basis for a new exciting line of investigation into the relationship between response to cellular stresses and antibody gene diversification in mature B cells.

Cd19^Cre (Cd19^tm1(cre)Cgn; Rickert et al., 1997) and Aicda^−/− (Aicda^{tm1Hon/tm1Hon}; Muramatsu et al., 2000) mice were previously described and were maintained on a C57BL/6 background. The conditional Pdap1^F allele bears LoxP sites flanking Exon 2 (ENSMUST00000031627.8) and was generated by CRISPR-Cas9–mediated knock-in microinjection of single-cell embryos on a C57BL/6N strain background (Max Delbrück Center [MDC] Transgenics platform). Germline transmission was confirmed, and positive pups were bred with Cd19^Cre mice to generate Pdap1^F/FCd19^Cre/+ mice. Mice were maintained in a specific pathogen-free barrier facility under standardized conditions (20°C ± 2°C; 55 ± 15% humidity) on a 12-h light/dark cycle. 7–27-wk-old mice of both genders were used in age-matched groups for the experiments. Controls included littermates and mice from inbred strains. All experiments were performed in compliance with European Union directive 2010/63/EU and in agreement with Landesamt für Gesundheit und Soziales directives (Berlin, Germany). Primers used for genotyping the Pdap1^F allele are listed in Table S2.

B lymphocytes were isolated from mouse spleens using anti-CD43 MicroBeads (Miltenyi Biotec) and grown in RPMI 1640 (Life Technologies) supplemented with 10% FBS, 10 mM Hepes (Life Technologies), 1 mM sodium pyruvate (Life Technologies), 1× Antibiotic Antimycotic (Life Technologies), 2 mM l-glutamine (Life Technologies), and 1× 2-mercaptoethanol (Life Technologies) at 37°C and 5% CO₂.

The cell lines used for this study were CH12 (CH12F3, mouse; Nakamura et al., 1996); WT (R1 and R2) and Pdap1-deficient (P1, P2, and P3) CH12 clonal derivatives (mouse; this paper); BOSC23 (human; Pear et al., 1993); and immortalized mouse embryonic fibroblasts (iMEFs; Bothmer et al., 2011). CH12 cells were grown in RPMI 1640 supplemented with 10% FBS, 10 mM Hepes, 1 mM sodium pyruvate, 1× Antibiotic Antimycotic, 2 mM l-glutamine, and 1× 2-mercaptoethanol at 37°C and 5% CO₂. BOSC23 cells and iMEFs were cultured in DMEM (Life Technologies) supplemented with 10% FBS, 2 mM l-glutamine, and penicillin-streptomycin (Life Technologies) at 37°C and 5% CO₂.

CH12 cells were stimulated to undergo CSR to IgA by treatment with 1–5 µg/ml αCD40 (BioLegend), 5 ng/ml TGFβ (R&D Systems), and 5 ng/ml mouse recombinant IL-4 for 48 h. B lymphocytes were stimulated to undergo class switching with 5 µg/ml LPS (Sigma-Aldrich) and 5 ng/ml mouse recombinant IL-4 (Sigma-Aldrich) for CSR to IgG1; 5 µg/ml LPS only for CSR to IgG3; 5 µg/ml LPS, 10 ng/ml BAFF (PeproTech), and 2 ng/ml TGFβ for CSR to IgG2b; or 5 µg/ml LPS, 10 ng/ml BAFF, 2 ng/ml TGFβ, and 1.5 ng/ml recombinant murine IL-5 (PeproTech) for CSR to IgA. For class switching analysis, cell suspensions were stained with fluorochrome-conjugated anti-IgG1, anti-IgG3 (BD Biosciences), anti-IgG2b (BioLegend), or anti-IgA (Southern Biotech).

The pMX-AID-ER-IRES-GFP retroviral vector was a kind gift from Qiao Wang (The Rockefeller University, New York, NY). Splenocyte infections were performed as follows. The HEK293T derivative cell line BOSC23 was transfected with pCL-Eco and pMX-IRES-GFP or pMX-AID-ER-IRES-GFP retroviral vectors using FuGENE HD Transfection Reagent (Promega) to generate viral particle–containing supernatants. B cells were activated with 5 µg/ml LPS and 5 ng/ml IL-4 (IgG1) or 5 µg/ml LPS, 10 ng/ml BAFF, and 2 ng/ml TGFβ (IgG2b) and transduced twice with the viral supernatant, 1 and 2 d after isolation. 4-HT was added 72 h after activation at a final concentration of 0.05 µM, and CSR was assessed 24 h later on the GFP⁺-gated populations. The pMX-Pdap1-3xFlag construct was generated by cloning the cDNA for human Pdap1 into pMX retroviral vector with a C-terminal 3xFlag tag. Primers used to clone the pMX-Pdap1-3xFlag construct are listed in Table S2. iMEFs infections were performed as indicated above for splenocytes, except that cells were transduced with pCL-Eco and pMX-EV/pMX-Pdap1-3xFlag-viral supernatants four times over 2 d. 48 h after the first transduction, iMEFs were selected using 1.5 µg/ml puromycin (Sigma-Aldrich).

For analysis of B cell development and differentiation, spleen and bone marrow cell suspensions were incubated with ACK lysis buffer (Thermo Fisher Scientific) for erythrocyte depletion. Subsequently, cells were blocked with TruStain fcX (BioLegend) for 10 min at 4°C and labeled with fluorochrome-conjugated antibodies to determine the surface expression of CD43, IgM, CD21/CD35, and IgD (BD Biosciences), and CD23, CD3, and CD19 (BioLegend) in PBS supplemented with 3% FBS (PBS/FBS) for 20 min at 4°C. Cells were then washed, resuspended in PBS/FBS, and analyzed.

For analysis of plasma cell differentiation in vitro, splenic B cells were isolated by immunomagnetic depletion of CD43⁺ cells and cultured at a density of 10⁶ cells/ml in the presence of either 20 µg/ml LPS (Sigma-Aldrich; L2880-10MG) and 25 ng/ml IL-4, or 20 µg/ml LPS, 10 ng/ml BAFF, and 2 ng/ml TGFβ. To determine the percentage of plasmablasts, cells were harvested after 96 h, blocked with TruStain fcX for 10 min at 4°C, and stained for surface expression of B220 (BioLegend; 103245) and CD138 (BioLegend; 142506) in MACS buffer (PBS supplemented with 0.5% BSA and 2 mM EDTA) for 20 min at 4°C. Cells were then resuspended in FACS buffer (PBS supplemented with 3% FCS and 1 mM EDTA) containing 1 µg/ml propidium iodide and analyzed.

For analysis of the plasma cell compartment in vivo, 9-, 14-, and 25-wk-old mice were sacrificed to isolate the spleen and hind legs. Total splenocytes and bone marrow cells were isolated by smashing the respective organs, and erythrocytes were depleted by incubation with Gey’s ABC solution for 3 min on ice. For surface staining, 5 × 10⁶ cells were first blocked with TruStain fcX for 10 min at 4°C and then stained for the following markers for 20 min at 4°C in MACS buffer: CD138, TACI (BD PharMingen; 558410), CD93 (BioLegend; 136506), CD19 (BioLegend), and B220. Cells were resuspended in FACS buffer containing 1 µg/ml propidium iodide and analyzed.

For ELISpot analysis, MultiScreen_HTS-IP plates (Merck) were activated with 35% ethanol in PBS for 1 min, washed three times with PBS, and finally coated with 1 µg/ml anti-Igκ (Southern Biotech) and anti-Igλ (Southern Biotech) in PBS at 4°C overnight. The following day, 2 × 10⁵ total splenocytes or BM cells (as well as serial fivefold dilutions) were cultured overnight in RPMI 1640 supplemented with 10% FCS (Sigma-Aldrich) and penicillin-streptomycin. The next day, plates were washed thoroughly six times with PBS containing 0.1% Tween20 and incubated with 1 µg/ml biotinylated antibodies against IgM (Southern Biotech) for 2 h at 37°C. Plates were washed three times under running distilled water and incubated with 0.3 U/ml streptavidin-AP (Roche) for 30 min at room temperature. For detection of spots, plates were washed three times under running distilled water, equilibrated in AP buffer (100 mM Tris-HCl, pH 9.0, 150 mM NaCl, and 1 mM MgCl₂), and developed using NBT/BCIP substrate mix (Promega) diluted in AP buffer. Spots were scanned and counted using the ImmunoSpot Series 6 Alfa Analyzer and ImmunoCapture Image Acquisition as well as ImmunoSpot Analysis software (C.T.L.) with the following parameters: sensitivity 176, spot separation 1.00, diffuseness largest, background balance 40, minimum spot size 0.0015 mm², and maximum spot size 0.6976 mm².

For assessment of GC B cells and CSR in vivo, cell suspensions derived from Peyer’s patches were blocked with TruStain fcX for 10 min at 4°C. Subsequently, cells were stained with fluorochrome-conjugated anti-CD19, anti-B220/CD45R, anti-CD38 (BioLegend), anti-Fas/CD95, and anti-IgA (BD Biosciences) for 20 min at 4°C in PBS/FBS. Cells were washed, resuspended in PBS/FBS, and analyzed. All samples were acquired on a LSRFortessa cell analyzer (BD Biosciences).

Resting B cells and CH12 were fixed with 4% paraformaldehyde (Sigma-Aldrich) and permeabilized with PBST (0.1% Tween 20 in PBS 1×) containing 0.2% IGEPAL CA-630 (Sigma-Aldrich). Cells were stained with antibodies against Pdap1 (Sigma-Aldrich), rat anti-mouse IgM-PE (BD Biosciences), and goat anti-rabbit Alexa Fluor 488 (Abcam). Cells were incubated with permeable nuclear dye (Hoechst dye 33258, Thermo Fisher Scientific) and transferred onto slides for imaging.

iMEFs infected with either empty pMX vector (EV) or pMX-Pdap1-3xFlag construct were grown on coverslips overnight. Upon fixation with 4% paraformaldehyde and permeabilization with 0.5% Triton X-100, cells were stained with mouse anti-Flag M2 (Sigma-Aldrich), goat anti-mouse Alexa Fluor 546 (Invitrogen), and Hoechst dye 33258. Images were acquired using inverted LSM700 AxioObserver laser scanning confocal microscope (Zeiss), with Plan-Apochromat 63×/1.40 Oil Ph3 objective for primary B cells and CH12, and EC Plan-Neofluar 40×/1.30 Oil Ph3 objective for iMEFs.

For cell proliferation analysis by cell tracking dye dilution, primary B cells were pulsed with 5 µM CellTrace Violet (Thermo Fisher Scientific) for 10 min at 37°C. Apoptosis analysis was performed by using CaspGLOW Fluorescein Active Caspase Staining Kit (BioVision) according to the manufacturer’s instructions. Samples were acquired on a LSRFortessa cell analyzer (BD Biosciences). Mitochondrial mass and membrane potential were measured via staining with MitoTracker Green and DeepRed (Thermo Fisher Scientific), respectively, according to the manufacturer’s instructions. Samples were acquired on a LSRFortessa cell analyzer (BD Biosciences).

For the loss-of-CSR assay in bulk CH12 cultures, three gRNAs against Pdap1 gene were cloned into the U6 cassette of a variant of the original pX330 plasmid (pX330-U6-Chimeric_BB-CBh-hSpCas9, Addgene; 42230) modified to express Cas9^WT-T2A-GFP (kind gift from Van Trung Chu, MDC; gPdap1-1/3), and three gRNA pairs were cloned into tandem U6 cassettes in a mutated version of pX330 expressing Cas9^D10A (Nickase-a/c). CH12 cells were transfected with the Cas9-gRNAs expressing constructs (either individually or in a pooled format) via electroporation with Neon Transfection System (Thermo Fisher Scientific), sorted for GFP-positive cells after 40 h, and left to recover for 72 h before activation for CSR analysis. For the generation of Pdap1-deficient CH12 clonal derivatives, the Nickases-a/c constructs were individually electroporated into CH12. Single GFP-positive cells were sorted in 96-well plates after 40 h, and clones were allowed to grow for 12 d before CSR analysis in 96-well format and expansion of selected clones. Controls for gRNA-nucleofected CH12 were cells nucleofected with either empty vector or gRNAs against random sequences not present in the mouse genome (single random gRNA-Cas9^WT as controls for gPdap1-1/3, and random gRNAs pairs-Cas9^D10A as controls for Nickase-a/c). Selected clones R1/R2 and P1/3 were validated at the level of genomic scar and protein expression. The sequences of the gRNAs used in these studies (this paper; Delgado-Benito et al., 2018) are listed in Table S2.

WB analysis of protein levels was performed on whole-cell lysates prepared by lysis in radioimmunoprecipitation assay buffer (Sigma-Aldrich) supplemented with Complete EDTA free proteinase inhibitors (Roche). Pierce Phosphatase Inhibitor Mini Tablets (Thermo Fisher Scientific) were added to the lysis buffer for the analysis of phosphorylated Perk and eIF2α. The antibodies used for WB analysis are anti-Pdap1 (Sigma-Aldrich), anti-tubulin (Abcam), anti-β-actin (Sigma), anti-AID (mAID-2; Thermo Fisher Scientific), anti-phospho-Perk (pThr980; Thermo Fisher Scientific), anti-Perk (Cell Signaling Technology), anti-phospho-eIF2α (pSer51; Cell Signaling Technology), anti-eIF2α (Santa Cruz), anti-Atf4 (Cell Signaling Technology), and anti-Xbp1-s (Cell Signaling Technology). WB signals were quantified using ImageJ (National Institutes of Health), and normalization was performed against loading controls.

mRNA levels for AID and postspliced germline transcripts were measured as follows. Total RNA was extracted from splenocyte cultures 48 h after activation using TRIzol (Invitrogen) according to manufacturer’s instructions and retro-transcribed with SuperScript VILO cDNA Synthesis Kit (Invitrogen). Genomic DNA was removed by RapidOut DNA Removal Kit (Thermo Fisher Scientific). Transcripts were amplified using StepOnePlus Real-Time PCR System (Applied Biosystems) with Luna Universal qPCR Mastermix (NEB). For analysis of Aicda mRNA decay, Actinomycin D was added to the culture medium 48 h after activation at a final concentration of 10 µg/ml for 0.5, 1, 1.5, 2, 3, and 4 h. qPCR analyses were normalized to Ubc or Gapdh (Aicda mRNA and GLTs) or Gapdh (Aicda mRNA decay). Primers used for qPCR (this paper; Xu et al., 2015; Muramatsu et al., 2000; Albershardt et al., 2012) are listed in Table S2.

For electrophoretic analysis of Xbp1 splicing, Xbp1 transcripts were amplified using the primers listed in Table S2 (Madaro et al., 2013) and either mock-digested or digested with PstI enzyme (NEB) for 30 min at 37°C. The intact spliced (Xbp1-s) transcript RT-PCR product and the unspliced (Xbp1-u) transcript digestion products were visualized on 2% agarose gel.

RNA-seq was performed on three mice per genotype. Splenocytes were cultured in LPS and IL-4 (IgG1) or LPS, BAFF, and TGFβ (IgG2b) for 48 h. Cells were collected by centrifugation, and RNA was extracted with TRIzol (Invitrogen) according to manufacturer’s instructions. Ribosomal RNA was depleted using Ribo-Zero Gold rRNA Removal Kit (Illumina). Libraries were prepared with TruSeq Stranded Total RNA Library Prep Kit Gold (Illumina) and run in one lane of a flow cell of HiSeq 4000 (Illumina).

The RNA-seq data were analyzed using the pigx-rnaseq pipeline (Wurmus et al., 2018). STAR (Dobin et al., 2013) mapped the reads to the GRCm38 assembly for mouse by Ensembl, and HTSeq (Anders et al., 2015) was used to count the transcript abundance. Differential expression analysis was done using the DESeq2 (Love et al., 2014) package for R, which uses the Wald test for significance.

For splicing analysis of Aicda and Pdap1 genes, we used the edgeR (Robinson et al., 2010) package for R to determine differentially expressed exons. The annotation was provided by a filtered version of the GRCm38 gene annotation, which contained unique, Havana-annotated exons.

Single-cell suspensions from Peyer’s patches of 24–27-wk-old mice were first incubated with TruStain fcX, and then labeled with antibodies conjugated with B220-FITC (BioLegend), CD19-APC/Pacific Blue (BioLegend), CD38-Alexa Fluor 700 (Thermo Fisher Scientific), and CD95/Fas-PE (BD Biosciences). Non-GC (CD38⁺ Fas⁻) and GC B cells (CD38⁻ Fas⁺) were sorted on an Aria BD sorter. Genomic DNA was extracted, and the 5′ portions of J_H4 (Igh) and J_K5 (Igk) introns were amplified by PCR using Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific). The 800-bp J_H4 and 700-bp J_K5 PCR products were gel extracted, cloned into a pCR2.1 vector using the TOPO TA Cloning Kit (Invitrogen), and sequenced. Mutations were quantified over 510-bp downstream J_H4 and 536-bp downstream J_K5 gene segments. Primers used for SHM analysis (Rouaud et al., 2013; Sander et al., 2015) are listed in Table S2.

Oxygen consumption rate and extracellular acidification rate were measured with the Seahorse XF96 metabolite analyzer using the Seahorse XF Cell Mito Stress Test kit (Agilent). Primary B cells were isolated, counted, and plated at the indicated cell densities directly on Seahorse cell culture plates coated with poly-l-lysine (Sigma-Aldrich). Before measurement, cells were incubated for a maximum of 1 h in Seahorse XF DMEM medium pH 7.4 (Agilent) supplemented with 10 mM glucose (Agilent), 1 mM sodium pyruvate (Agilent), and 2 mM l-glutamine (Agilent).

The statistical significance of differences between groups/datasets was determined by the Mann–Whitney U test for all data presented in this study, with the following exceptions. For the RNA-seq analysis, the Wald test of the DESeq2 package for R was used, and we considered genes with an FDR < 0.05 to be significantly differentially expressed. Statistical details of experiments can be found in the figure legends.

The deep-sequencing data reported in this paper (RNA-seq) has been deposited in the GEO repository under accession no. GSE141876.

Fig. S1 shows that Pdap1 is dispensable for B cell development. Fig. S2 shows that Pdap1 is dispensable for B cell proliferation. Fig. S3 shows that Aicda mRNA levels are reduced in Pdap1-deficient CH12 cell lines. Fig. S4 shows that Pdap1 is localized primarily in the cytoplasm in B cells and MEFs. Fig. S5 shows the stress response signature of Pdap1-deficient B cells. Table S1 contains the pairwise comparison of differentially regulated genes in control versus Pdap1^F/FCd19^Cre/+ splenocytes. Table S2 lists the oligonucleotides used in this study.

We thank all members of the Di Virgilio laboratory for their feedback and discussion; V. Coralluzzo (Di Virgilio Lab, MDC, Berlin) for support with genotyping; Y. Dramaretska (Gargiulo Lab, MDC) for helpful tips on the RNA-seq protocol; the MDC Transgenics platform and Dr. R. Kühn for generation of the Pdap1^F mouse allele; the MDC FACS Core Facility and Dr. H.P. Rahn for assistance with cell sorting; K. Rajewsky (MDC) for feedback and discussion; and N. Zampieri (MDC) for critical reading of the manuscript and advice.

This work was supported by the European Research Council (grant 638897 to M. Di Virgilio), the Helmholtz-Gemeinschaft Zukunftsthema “Immunology and Inflammation” (ZT-0027 to M. Di Virgilio), and the Deutsche Krebshilfe (grant 70112800 to M. Janz). M. Di Virgilio is a Helmholtz Young Investigators Group leader (Helmholtz Association).

Author contributions: M. Di Virgilio and V. Delgado-Benito conceived the project idea and designed the experiments; M. Di Virgilio, V. Delgado-Benito, and M. Berruezo-Llacuna analyzed and interpreted the data; V. Delgado-Benito, M. Berruezo-Llacuna, W. Winkler, D. Sundaravinayagam, S. Balasubramanian, M. Caganova, A. Rahjouei, M.-T. Henke, M. Driesner, and L. Keller performed the experiments; R. Altwasser performed all RNA-seq data analyses; R. Graf contributed to design of the list of gRNAs used for the initial loss-of-CSR screen; A. Prigione, M. Janz, A. Akalin, and M. Di Virgilio supervised the experiments or analyses performed in their respective groups; M. Di Virgilio obtained the funding for the project and wrote the manuscript.