Toll-Like Receptor 9 Antagonizes Antibody Affinity Maturation

Articles https://doi.org/10.1038/s41590-018-0052-z

Toll-like receptor 9 antagonizes antibody affinity maturation

Munir Akkaya 1, Billur Akkaya 2, Ann S. Kim 1, Pietro Miozzo 1, Haewon Sohn1, Mirna Pena1, Alexander S. Roesler1, Brandon P. Theall 1, Travis Henke1, Juraj Kabat3, Jinghua Lu1, David W. Dorward4, Eric Dahlstrom4, Jeff Skinner 1, Louis H. Miller5 and Susan K. Pierce 1*

Key events in T cell–dependent antibody responses, including affinity maturation, are dependent on the B cell’s presentation of antigen to helper T cells at critical checkpoints in germinal-center formation in secondary lymphoid organs. Here we found that signaling via Toll-like receptor 9 (TLR9) blocked the ability of antigen-specific B cells to capture, process and present antigen and to activate antigen-specific helper T cells in vitro. In a mouse model in vivo and in a human clinical trial, the TLR9 agonist CpG enhanced the magnitude of the antibody response to a protein vaccine but failed to promote affinity maturation. Thus, TLR9 signaling might enhance antibody titers at the expense of the ability of B cells to engage in germinal-center events that are highly dependent on B cells’ capture and presentation of antigen.

ey checkpoints in T cell–dependent antibody responses in seem to be required for adjuvant-enhanced T cell–dependent anti- vivo are dependent on antigen-specific B cell–T cell interac- gen-specific antibody responses in vivo11. Although adjuvants can tions. The initiation of T cell–dependent antibody responses influence several qualities of antibody responses, including kinet- K 12 occurs in secondary lymphoid organs and is dependent on the ics, magnitude and breadth , the ability of adjuvants to augment stable interaction of antigen-primed helper T cells with activated B cell GC responses, including somatic hypermutation events, has antigen-specific B cells through complexes of peptide and major remained unexplored until recently. BCR responses specific to the histocompatibility complex (MHC) class II presented on the B cell human immunodeficiency virus envelope protein in nonhuman surface1–4. Depending in part on the quality of the B cell–helper T primates responding to immunization with envelope protein for- cell interaction, B cells either enter germinal centers (GCs) or dif- mulated with eight different adjuvants, including ones that target ferentiate into short-lived plasma cells (PCs) and GC-independent TLR pathways, have now been compared13. The adjuvants increased memory B cells2. Within GCs, the competitive process of affinity the titer of antibodies specific to envelope protein but, remark- selection occurs on the basis of the ability of B cell receptors (BCRs) ably, did not increase the frequency of somatic hypermutation to capture, process and present antigen to follicular helper T cells essential for the development of broadly neutralizing antibodies

(TFH cells). The B cells’ successful presentation of antigen to TFH cells in this system. ultimately results in the differentiation of GC B cells into long-lived Here we provide evidence that although TLR9 signaling memory B cells and PCs. enhanced BCR-induced B cell proliferation and differentiation into B cells also express germline-encoded Toll-like receptors (TLRs) antibody-secreting cells, TLR9 signaling reduced the ability of B that respond to microbial products that have pathogen-associated cells to capture, process and present antigen and to interact with molecular patterns5–7. The dual expression of the BCR and TLRs and activate antigen-specific helper T cells in vitro, which was det- allows B cells to modulate the outcome of antigen encounter in the rimental to the establishment of high-affinity, long-lived antibody presence of pathogens5,6. Indeed, TLR9 signaling has been shown responses in vivo. to enhance the response of B cells to antigens coupled to the TLR9 agonist CpG, in terms of proliferation and differentiation into anti- Results body-secreting cells both in vitro and in vivo. In addition to sig- The effect of TLR9 signaling on BCR-dependent activation of naling for B cell activation, the BCR functions to transport bound B cells. We assessed the effect of CpG on several antigen-driven antigen to specialized intracellular compartments in which the anti- BCR signaling functions in mouse splenic B cells purified by ʹ gen is degraded and the resulting peptides are bound to MHC class negative selection and treated with soluble polyclonal F(ab)2 II molecules that are presented on the cell surface for recognition by antibody specific for immunoglobulin M (anti-IgM), CpG or antigen-specific helper T cells8. At present, whether TLR signaling anti-IgM plus CpG. The effects assessed included the following: influences the processing and presentation of antigen by the BCR phosphorylation of kinases in the BCR signaling pathway (Fig. remains unknown. 1a–d and Supplementary Fig. 1a–c); Ca2+ responses (Fig. 1e and Although TLR signaling has the potential to amplify humoral Supplementary Fig. 1d); the expression of cytokine-encoding responses to antigen9,10, B cell–intrinsic TLR signaling does not mRNA at 3 h after stimulation (Fig. 1f) and of cytokines (as protein)

1Laboratory of Immunogenetics, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, MD, USA. 2Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, USA. 3Research Technologies Branch, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD, USA. 4Research Technologies Branch, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, MT, USA. 5Laboratory of Malaria and Vector Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Rockville, MD, USA. *e-mail: [email protected]

CpG Anti-IgM CpG + anti-IgM ab2.5 2.5 cd4.0 2.5 3.5 2.0 2.0 3.0 2.0 2.5 1.5 1.5 2.0 1.5

p-Btk MFI (fold) 1.5 p-Akt MFI (fold) p-Syk MFI (fold) p-p38 MFI (fold) 1.0 1.0 1.0 1.0 20 40 60 80 100 20 40 60 80 100 20 40 60 80 100 0 20 40 60 80 100 Time (min) Time (min) Time (min) Time (min) e 5 CpG 4 Anti-IgM 3 CpG + anti-IgM 2 Fluo-3 / Fura Red 1 200 400 600 Time (s) f 400 g 2,500 CpG US 300 Anti-IgM 2,000 **** CpG 200 CpG +anti-IgM Anti-IgM 100 1,500 CpG +anti-IgM 60 **** 1,000 *** *** 40 500

Cytokine (pg/ml) **** **** **** mRNA expression (fold) 20 80 40 ** ** 0 0 Ifnb1 Il2 Il6 Il12a Ifng Il4 Tgfb1 Il10 Tnf Il1b WT KO WT KO WT KO WT KO IL-2 IL-6 TNF IL-10 US CpG Anti-IgM CpG + anti-IgM hij 8 100 25 WT 21 80 6 17 60 13 9 4 40 5 IgG (ng/ml) IgM ( μ g/ml) 2 Proliferation (%) 1 20 0.3 KO 0.1 0 0 01234567 0 12345 1.5 3.0 0.05 0.15 0.25 0.35 Time (d) Time (d) CpG (μM) CpG CpG + anti-IgM klm 1.5 6 60

1.0 4 40

0.5 2 20

0.0 0 0 Bcl6 mRNA expression (fold) Aicda mRNA expression (fold) 01234 Prdm1 mRNA expression (fold) 01234 01234 Time (d) Time (d) Time (d)

Fig. 1 | The effect of TLR9 signaling on the outcome of B cell responses to antigen. Analysis of purified mouse splenic B cells stimulated in vitro with anti-IgM (2 µg/ml (a–d) or 5 µg/ml (e–g,i–m)) or CpG (1 μM) alone or in combination. a–d, Abundance of phosphorylated kinases in individual wild- type B cell samples fixed and barcoded with combinations of B220-specific antibodies19, pooled, permeabilized and stained with mAbs specific for the phosphorylated (p-) kinases Syk (a), Btk (b), p38 (c) and Akt (d); results are presented as mean fluorescent intensity (MFI) in stimulated B cells relative to that in unstimulated B cells. e, Calcium flux (measured by flow cytometry) in wild-type B cells loaded with the Ca2+ sensor dyes Furo-red and Fluo-4 and stimulated (key). f, Expression of mRNA encoding various cytokines (horizontal axis) in wild-type B cells stimulated for 4 h (key); results are presented relative to those of unstimulated B cells. g, ELISA of cytokines (below plot) in the culture supernatants of wild-type (WT) or TLR9-deficient (KO) B cells (horizontal axis) left unstimulated (US) or stimulated in vitro for 18 h (for analysis of IL-6) or for 24 h (for analysis of TNF, IL-2 and IL-10) with anti-IgM and/or CpG (key). h, Proliferation of wild-type or TLR9-deficient B cells (bracketing at right end of lines) stimulated in vitro with a sub-optimal concentration (1 µg/ml) of anti-IgM (or no anti-IgM) and increasing concentrations (0–3 µM; horizontal axis) of CpG (key), assessed after 46 h of culture. i,j, ELISA of IgM (i) and IgG (j) in supernatants of wild-type B cells left unstimulated or stimulated for 7 d in vitro (key), assessed for samples depleted of IgG+ B cells (Supplementary Fig. 1g) in j. k–m, Expression of Bcl6 mRNA (k), Prdm1 mRNA (l) and Aicda mRNA (m) in wild-type B cells stimulated for 0–4 d (horizontal axis) with anti-IgM alone or with anti-IgM plus CpG (key); results are presented relative to those of unstimulated B cells at time 0. Each symbol (f,g) represents an individual biological replicate; dashed horizontal lines (f) indicate the mean. **, 0.001 < P ≤ 0.01; ***, 0.0001 < P ≤ 0.001; ****P ≤ 0.0001 (two-sided unpaired t-test). Data are representative of three independent experiments with biological duplicates (a–d; mean and s.d.) or biological triplicates (f–m; mean and s.d. in h–m).

abCpG with anti-IgM or CpG resulted in different patterns of kinase 4 90 min CpG 2.5 No CpG phosphorylation, the result of treatment with both anti-IgM and No CpG 60 min 3 2.0 CpG was additive and was dependent on the concentration of anti- 1.5 10 min IgM and, for CpG responses, on TLR9 expression (Fig. 1a–c and 2 Supplementary Fig. 1a–c). In contrast, B cells underwent Ca2+ flux 1.0

pHrodo MFI (fold) 1 only in response to anti-IgM (Fig. 1e and Supplementary Fig. 1d), pHrodo MFI (fold)

0 10 30 60 90 and this response was unaffected by CpG. CpG and anti-IgM acted 120 160 WT KO WT KO WT KO Time (min) in synergy to induce high expression of the cytokines IL-2 and TNF at the level of both mRNA and protein, whereas the CpG-induced c 10 min HEL CD71 Colocalized Merge+DAPI 0.8 expression of IL-6 and IL-10 was substantially antagonized by anti- P = 0.0878 IgM (Fig. 1f,g). CpG enhanced, in a TLR9-dependent fashion, the No CpG 0.6 proliferation of B cells treated sub-optimally with anti-IgM (Fig. 1h and Supplementary Fig. 1e,f). CpG induced the secretion of IgM by 0.4 B cells that was weakly antagonized by anti-IgM (Fig. 1i), whereas

HEL-CD71 anti-IgM enhanced the CpG-induced secretion of IgG by B cell 0.2 CpG populations depleted of IgG+ B cells (Fig. 1j and Supplementary

(3D colocalization index) 0.0 Fig. 1g). Together these results demonstrated the ability of CpG to alter the B cell response to BCR crosslinking and the ability of BCR CpG No CpG crosslinking to alter the B cell response to CpG. d 60 min We also measured by qPCR the transcription of genes encoding three factors that influence the PC fate of B cells14. CpG alone and HEL LAMP-1 Colocalized Merge+DAPI P = 0.0015 0.8 CpG plus anti-IgM decreased the expression of Bcl6 (which encodes a key transcriptional repressor of PC differentiation whose expres- No CpG 0.6 sion is critical for the maintenance of B cell GC reactions) (Fig. 1k),

0.4 but such treatment increased the expression of Prdm1 (which encodes BLIMP-1, a transcription factor that promotes PC differ-

CpG HEL–LAMP-1 0.2 entiation) (Fig. 1l) and Aicda (which encodes the deaminase AID and is upregulated when B cells differentiate toward PCs) (Fig. 1m). (3D colocalization index) 0.0 Together these results provided evidence that TLR9 signaling had

CpG the potential to drive B cells toward PC differentiation and away No CpG from GC responses. e 60 min

HEL H2M Colocalized Merge+DAPI 0.8 P = 0.0009 Internalization and trafficking of soluble antigen by the BCR. We assessed the ability of the BCRs to internalize soluble antigen, either No CpG 0.6 anti-IgM (internalized by C57BL/6 B cells) or hen egg lysozyme (HEL) (internalized by HEL-specific B cells from MD4 mice, which 0.4 have transgenic expression of a HEL-specific BCR), in the presence HEL-H2M CpG 0.2 or absence of CpG. CpG did not affect the rate or magnitude of the internalization of antigen by BCRs in either case (Supplementary

(3D colocalization index) 0.0 Fig. 2a,b). We characterized the intracellular vesicles into which anti-IgM was internalized by incubating B cells with anti-IgM- CpG No CpG coated electron-dense metal particles and imaging the cells by transmission electron microscopy and tomography. Quantification Fig. 2 | TLR9 signaling antagonizes the trafficking of BCR-bound antigen of three-dimensional reconstructions of the images showed into late endosomal compartments. a, MFI of the pH-sensitive fluorescent anti-IgM+ particles in morphologically similar compartments in dye pHrodo (measured by flow cytometry) in HEL-specific MD4 B cells CpG-treated cells and untreated cells (Supplementary Fig. 2c–g). incubated for 0–160 min (horizontal axis) at 37 °C with pHrodo-conjugated Once internalized into early endosomes, the BCR traffics antigen HEL in the presence or absence of CpG (key); results are presented to increasingly acidic late endosomes. We measured the fluorescence relative to those obtained before incubation. b, MFI of pHrodo in TLR9- intensity (FI) of HEL-specific MD4 B cells internalizing soluble sufficient (‘wild-type’) MD4 and TLR9-deficient MD4 B cells (horizontal HEL conjugated to a pH-sensitive dye, the FI of which increases axis) incubated for 10, 60 or 90 min (above plot) as in a (key); results with decreases in pH. For both CpG-treated B cells and untreated presented as in a. c–e, Confocal microscopy (left) of MD4 B cells that B cells, the FI increased at the same rate for the first 20 min at 37 °C were immobilized on coverslips and incubated with rhodamine-conjugated (Fig. 2a,b and Supplementary Fig. 3a). In the absence of CpG, the FI HEL in the presence or absence of CpG (1 μM) (left margin), then fixed, of the B cells continued to increase with time. In contrast, the FI of permeabilized and stained with mAbs specific for CD71 (c), LAMP1 (d) or the CpG-treated B cells increased little after 30 min, and the block in H2M (e). Right, three-dimensional (3D) colocalization indices of HEL with trafficking was TLR9 dependent (Fig. 2a,b). CD71 at 10 min (c), of HEL with LAMP-1 at 60 min (d) or of HEL with H2M We further characterized, by confocal microscopy, the endo- at 60 min (e) in cells as at left (n = 25 cells per group). Scale bars (left), somal compartments into which the BCR internalized rhoda- 1 µm. Each symbol (b–e) represents an individual biological replicate; mine-conjugated HEL in the presence or absence of CpG (Fig. dashed horizontal lines (c–e) indicate the mean. P values (above plots, 2c–e). Cells were fixed and then were stained with the DNA- c–e), two-sided Mann-Whitney U-test. Data are representative of three binding dye DAPI (to label nuclei) and with antibodies specific independent experiments (mean ± s.d. in a). for the following: the transferrin receptor CD71, a marker of early endosomes; LAMP1, a marker of late endosomes; or H2M, at 18–24 h (Fig. 1g); the proliferative responses of B cells (Fig. 1h); an MHC class II chaperone present in antigen-processing com- and concentration of IgM (Fig. 1i) and IgG (Fig. 1j) in culture partments. Three-dimensional colocalization of HEL with CD71, supernatants over 7 d. We determined that although treatment LAMP1 or H2M was quantified by software for the analysis of

b a ) 60 CpG (20 min) Time (s) 0918 36 63 117 180 306 2

40 No CpG 20 No CpG CpG

Contact area ( μ m 0 0 CpG 90 (20 min) 180 270 360 450 540 630 Time (s)

) 60 2 CpG (60 min)

40 CpG (60min) 20

Contact area ( μ m 0 0 90 180 270 360 450 540 630 No CpG CpG Time (s) c d CpG (20 min) CpG (60 min) CpG (20 min) CpG (60 min) 3.0 1.6 1.6 P = 0.0048 P < 0.0001 P < 0.0001 P = 0.6609 3 2.5 1.4 1.4 2.0 2 1.2 1.2 1.5

1 1.0 1.0 HEL MFI (normalized) 1.0 HEL MFI (normalized) BCR MFI (normalized) BCR MFI (normalized) 0 0 0 0 90 90 90 90 180 270 360 450 540 630 180 270 360 450 540 630 180 270 360 450 540 630 180 270 360 450 540 630 Time (s) Time (s) Time (s) Time (s) e f

P = 0.0003 P < 0.0001 120 HEL 90 20 60 10 30 )

20 7 2.5 15 2.0 1.5 10 (TFI × 10 1.0

5 Internalized HE L

HEL + CpG Internalized HEL (% ) 0.5 0 0

CpG CpG No CpG No CpG BCR HEL Membrane

ghi P = 0.117 0.8 1.0 P = 0.0008

0.6 0.8 0.6 0.4 HEL HEL + CpG 0.4 0.2

HEL-BC R 0.2

0.0 HEL-membrane 0.0 (3D colocalization index) –0.2 (3D colocalization index) –0.2

CpG CpG No CpG No CpG

Fig. 3 | The effect of TLR9 signaling on B cell responses to membrane bound antigen. a, Total internal reflection fluorescence microscopy of HEL-specific MD4 B cells left untreated (No CpG) or pretreated for 20 or 60 min with 1 µM CpG (left margin), labeled with DyLight 649–Fab anti-IgM and placed for 0–306 s (above images) on planar lipid bilayers containing Alexa Fluor 488–HEL (Supplementary Video 1a–c). Scale bar (below images), 12.45 µm. b, Quantification of the contact area of B cells pretreated as in a (above plots) or left untreated (key) and placed for 0–630 s (horizontal axis) on a HEL- containing planar lipid bilayer as in a. c,d, Accumulation of the BCR (c) and HEL (d) in the area of contact of the B cell with the planar lipid bilayer as in b. e, Confocal z-stack images (reconstituted as an x and z side view by maximal projection along the y axis) of HEL-specific MD4 B cells labeled with DyLight 649–Fab anti-IgM and left untreated (HEL) or treated with 1 µM CpG (HEL + CpG), then placed for 30 min at 37 °C on a CM-DiI-labeled PMS (green) containing Alexa Fluor 488–HEL (magenta; internalized HEL, cyan). Scale bars, 2 µm. f, Proportion of HEL in the bilayer as in e that was internalized (left), and total internalized HEL, presented as total fluorescence intensity (TFI) (right). g, Three-dimensional (3D)-reconstructed image of confocal z-stacks showing the BCR (cyan), HEL (magenta) and the PMS (yellow) (Supplementary Video 2a,b) in cells as in e. Scale bars, 2 µm. h,i, Quantification of the colocalization of the BCR and HEL (h) and HEL and the PMS (i) in cells as in e. Each symbol (f,h,i) represents an individual cell; dashed horizontal lines indicate the mean. P values (above plots, c,d,f,h,i), two-sided quadratic regression (c,d) or two-sided Mann-Whitney U-test (f,h,i). Data represent three independent experiments.

Gene expression (fold) Protein expression (fold) a cd –2 0 2 4 –4 –2 0 2 4 CpG Anti-lgM Mitotic roles of polo like kinase CD223 Cell cycle regulation by BTG family CD69 Death receptor signaling Apoptosis signaling CD71 Cell cycle g2/m dna check point regulation CD25 541 410 394 Cyclins and cell cycle regulation GARP Cell cycle G1/s checkpoint regulation CD279 Telomerase signaling Role of CHK proteins in cell cycle check Notch 2 19,443 p53 signaling CD155 ERK signaling LAP 3,866 1,092 NFKb signaling CD262 Jak Stat signaling p38 signaling CD63 Protein kinase A signaling CD98 609 EIF2 signaling CD44 PI3K/AKT signaling CD267 STAT3 pathway CD274 PI3K pathway in B cells Phospholipase C signaling Mac-3 mTOR signaling DcTRAIL-R2 CpG + anti-lgM SAPK/JNK signaling 4-1BB Ligand Acute phase response signaling CD81 B Cell activating factor signaling IL-6 signaling CD107a Tgfb signaling CD49e Interferon signaling CD205 CpG iNOS signaling CD127 IGF-1 signaling b 15 US April mediated signaling CD317 CXCR4 signaling CD49b Anti-lgM TNFR2 signaling CD132 10 TNFR1 signaling CD268 CpG + anti-lgM IL-8 signaling

8%) IL-2 signaling CD86 IL-1 signaling CD83 5 FcgRIIB signaling in B cells MHC-II Integrin signaling

iance: CD27 B cell receptor signaling ar 0 Pattern recognition receptors GITR Toll like receptor signaling CD196 CD27 signaling in lymphocytes CD54 –5 CD40 signaling CD49f PC2 (v Rho A signaling cdc42 signaling CD5 Rho family GTPases CD120b –10 Rac signaling CD147 Regulation of actin motility by Rho CD36 –40 –20 0 20 40 Actin cytoskeleton signaling Actin nucleation by ARP-WASP complex Allergin-1 PCI (variance: 88%) RhoGDI signaling Tim-1 CD31 CD272 CpG CD30 CpG + Anti-lgM CD1d anti-lgM CD66a LPAM-1 e f US CpG CD39 CD49d Anti-lgM CpG + anti-lgM CD21,CD35 HEL HEL + CpG CD256 100 P = 0.0008 CD275 4 CD90.2 Galectin-9 80 lgD 3 CD51

60 MFI CD61 k PIR-A/B 2 Integrin B7 40 CD178 ed to mode CD23 HEL–I-A

W3.18 / isotype) 1 20 CD138 (A maliz CD62L

Nor 0 0 3 3 4 5 3 3 4 5 –10 0––10 –10 –10 10 0 10 10 10 CpG CpG + PE Anti-lgM anti-lgM Isotype AW3. 18

g 4,000 US CpG Anti-lgM h 2,500

2,000 3,000 ** * * ** * ** ** *** ** 1,500 *** **** **** 2,000 *** **** 1,000 **** CD86 (MFI) MHCII (MFT) 1,000 500

0 0

ycin ycin Akt IV osine Akt IV osine tmannin tmannin PD98059 r PD98059 r SB202190SP600125 MHY1485Perif SB202190SP600125 MHY1485Perif No inhibitor RapamWo No inhibitor RapamWo

Fig. 4 | The effect of TLR9 and BCR signaling on B cell transcription and the expression of B cell surface proteins. a, Quantification of genes with expression (by RNA-seq analysis) differentially affected in purified B cells left unstimulated or stimulated for 4 h in vitro with CpG (1 µM) or anti-IgM (5 µg/ml) or a combination of those (along perimeter), showing the overlap of those groups. b, Principle-component analysis of gene expression as in a. c, Ingenuity Pathway Analysis of gene expression (key) in cells as in a (below plot), relative to that of unstimulated B cells, presented as log2-scale enhanced and diminished transcriptional regulation of various pathways (left margin) in response to such stimulation. d, Surface expression of various proteins (assessed with fluorophore-conjugated antibodies) in purified B cells cultured for 24 h as in a (below plot) and barcoded, showing results for proteins with an increase in expression of threefold or more or a decrease in expression of 30% or more, relative to that of untreated B cells (log2 scale). e, Flow cytometry of HEL–I-Ak in HEL-specific MD4 B cells stimulated for 24 h with HEL (1 µg/ml) or HEL plus CpG (1 µM) (above plots), assessed with phycoerythrin (PE)-labeled mAb AW3.18 (specific for HEL–I-Ak) or isotype-matched control mAb (Isotype) (key). f, MFI of HEL–I-Ak in HEL-specific MD4 B cells left unstimulated or stimulated for 24 h with HEL (1 µg/ml) or CpG (1 µM) or both (key), presented as the ratio of the MFI of cells stained with mAb AW3.18 to that of cells stained with the isotype-matched control mAb (AW3.18/isotype). g,h, Surface expression of MHC class II (MHCII) (g) and CD86 (h) on B cells left unstimulated or stimulated with CpG (1 µM) or anti-IgM (5 µg/ml) (key) in the presence or absence of the following kinase inhibitors or activator (horizontal axis): 50 µM PD98059 (inhibits MEK1 and MEK2), 10 µM SB202198 (inhibits p38), 30 µM SP600125 (inhibits JNK), 5 µM Akt IV (inhibits Akt), 10 µM MHY1485 (activates mTOR; inhibits autophagy), 20 µM perifosine (inhibits Akt), 100 nM rapamycine (inhibits mTOR) or 50 nM wortmannin (inhibits PI3K). Each symbol (f,g,h) represents an individual biological replicate. *, 0.01 < P ≤ 0.05; **, 0.001 < P ≤ 0.01; ***, 0.0001 < P ≤ 0.001; ****P ≤ 0.0001, compared with the condition of no inhibitor or activator (two-sided unpaired t-test). Data are representative of one experiment in triplicate (a–c), two independent experiments in triplicate (d,g,h; mean in g,h) or two experiments (e,f; mean in f).

ab c 2 4 × 104 2.0 × 10–2 P = 0.288 P = 0.0041 6 × 10 P = 0.9083 P = 0.0023 P = 0.0638 P = 0.0296 3 × 104 1.5 × 10–2 US 4 × 102 CpG 2 × 104 –2 (s) 1.0 × 10 T cell T cell HEL 2 k length ( m m) 2 × 10 4 k speed ( μ m/s) 1 × 10 5.0 × 10–3 HEL + CpG trac trac

Colocalization duration 0 0 0 de 3A9 MD4 HEL CpG 3A9 MD4 HEL CpG T cells B cells (μg/ml) (μM) T cells B cells (μg/ml) (μM) ++–– ++––

++– 2.0 ++–2.0

+–– 2.0 +––2.0

+––– +–––

+–2.0 – +–2.0 –

++2.0 – ++2.0 –

++2.0 2.0 ++2.0 2.0

++2.0 0.2 ++2.0 0.2

++0.5 – ++0.5 –

++0.5 2.0 ++0.5 2.0

++0.5 0.2 ++0.5 0.2 –103 0 103 104 105 –103 0 103 104 105 CD44 (APC) e450 f WT MD4 B cells KO MD4 B cells WT MD4 B cells KO MD4 B cells h 100 WT 3A9 T cells WT 3A9 T cells KO 3A9 T cells KO 3A9 T cells Anti-CD3 80 US Anti-CD28 60 HEL Anti-CD3 ed to mode

ed to mode 40 HEL + CpG aliz aliz 20 US No rm

No rm 3 0 3 4 5 0 10 10 10 3 3 4 5 3 3 4 5 3 3 4 5 3 3 4 5 –10 –10 0 10 10 10 –10 0 10 10 10 –10 0 10 10 10 –10 0 10 10 10 CD44 (APC) e450 g i 100 Anti-CD3 80 Anti-CD28

60 Anti-CD3 ed to mode

ed to mode 40 aliz aliz 20 US No rm 3 3 4 5 No rm 0 No CpG 0 10 10 10 –10 –103 0 103 104 105 –103 0 103 104 105 –103 0 103 104 105 –103 0 103 104 105 e450 CpG CD44 (AF700)

Fig. 5 | TLR9 signaling decreases the ability of antigen-specific B cells to interact with and activate antigen-specific helper T cells in response to soluble antigen. a–c, Duration of B cell–T cell colocalization (a), T cell track speed (b) and T cell track length (c) for CMTPX-stained 3A9 T cells cultured (at a ratio of 1:1) with CMFDA-stained HEL-specific MD4 B cells in the presence or absence (US) of CpG (1µM) and/or HEL (1 µg/ml) (key), imaged for 24 h at rate of 1 frame per 10–20 min. d,e, CD44 expression (d) and proliferation (assessed with the cell-proliferation dye e450) (e) of HEL-specific 3A9 T cells stimulated for 72 h with various combinations (grid at left) of HEL (0.5 or 2.0 µg/ml), CpG (0.2 or 2.0 µM) and HEL-specific MD4 B cells. APC, allophycocyanin. f,g, Proliferation (f) and CD44 expression (g) of wild-type or TLR9-deficient HEL-specific 3A9 CD4+ T cells cultured for 72 h with wild-type or TLR9-deficient HEL-specific MD4 B cells (above plots) in the presence or absence of HEL (1 µg/ml) with or without CpG (1 µM) (key). AF700, Alexa Fluor 700. h,i, CD44 expression (h) and proliferation (i) of CD4+ T cells left unstimulated or stimulated for 72 h with anti-CD3 with or without anti-CD28 (right margin) in the presence or absence of 1 µM CpG (key). Each symbol (a–c) represents an individual event; dashed horizontal lines indicate the mean. P values (above plots, a–c), two-sided Welch’s t-test. Data represent two independent experiments in triplicate (f,g) or are representative of three experiments (a–e) or three independent experiments in triplicate (h,i). three-dimensional microscopy (Fig. 2c–e and Supplementary Fig. 3b,c) to 60 min, and treatment with CpG significantly dimin- Fig. 3b,c). HEL colocalized with CD71 in early endosomes at ished the colocalization at 60 min (Fig. 2d,e). Together these data 10 min, and treatment with CpG had no significant effect on indicated that TLR9 signaling compromised the BCR’s deliv- this colocalization (Fig. 2c). The colocalization of HEL with ery of antigen into late endosomes and the antigen-processing either LAMP1 or H2M increased from 10 min (Supplementary compartments.

TLR9 signaling affects B cell responses to membrane-bound anti- programs. An Ingenuity Pathway Analysis of the genes expressed gen. We assessed the effect of CpG treatment on B cell responses to differentially in B cells treated with anti-IgM or CpG (alone or antigen incorporated into fluid planar lipid bilayers. HEL-specific together) relative to their expression in unstimulated cells showed MD4 B cells were labeled with Fab anti-IgM conjugated to the fluo- that a variety of pathways were affected, including those related to rophore Dylight 649 and were placed on bilayers containing HEL cell proliferation and apoptosis, signal transduction, responses to labeled with the fluorescent dye Alexa Fluor 488, and live-cell imag- cytokines and cytoskeleton organization (Fig. 4c). ing was carried out by total internal reflection fluorescence micros- We also quantified by flow cytometry the B cell surface expres- copy15–17 (Fig. 3a and Supplementary Video 1a–c). When pretreated sion of approximately 250 mouse cell-surface proteins using a bar- for 20 min with CpG, the B cells spread more robustly over the HEL- coding strategy19 and a kit with fluorophore-conjugated antibodies containing bilayer, covering larger areas, relative to the spreading of for the screening of cell-surface molecules (Fig. 4d). The expression untreated B cells (Fig. 3a,b), but they did not form a well-organized of 63 surface proteins changed significantly after treatment relative central synapse and accumulated less BCR (Fig. 3c) and HEL (Fig. 3d) to their expression on unstimulated cells. We observed synergy (up in the contact area. B cells pretreated for 60 min with CpG acted to 55-fold) between CpG and anti-IgM that induced an increase similarly, but they contracted after 90 s and accumulated less BCR in 48% of the surface proteins; antagonism of anti-IgM responses than the untreated B cells did but accumulated amounts of HEL by CpG for 9.5% of markers; and antagonism of CpG responses by comparable to those accumulated by the untreated B cells (Fig. 3d), anti-IgM for 8% of surface proteins. The expression of 30% of the suggestive of an antigen-independent component of BCR accumu- surface proteins decreased in response to CpG and/or anti-IgM. lation, as described earlier18. Together these results supported the conclusion that signaling We also assessed the BCR-dependent internalization of HEL through the BCR alone, TLR9 alone or both together resulted in incorporated into plasma membrane sheets (PMSs) by HEL-specific distinct outcomes. MD4 B cells. B cells were labeled with Dylight 649–Fab anti-IgM We further analyzed the effect of CpG and anti-IgM on expres- and were placed for 30 min at 37 °C on PMSs labeled with fluores- sion of the early activation marker CD69 and three cell-surface cent dye CM-DiI (used for monitoring cell movement), into which molecules that are critical to B cell antigen presentation and acti- Alexa Fluor 488–HEL was incorporated. Confocal z-stack images vation of helper T cells: MHC class II and the costimulatory mol- were obtained and reconstituted as an x view and a z view (Fig. 3e) ecules CD86 and CD80 (Supplementary Fig. 4a–g). The surface with software program codes that quantify only the internalized expression of MHC class II increased approximately eight- to ten- antigen, excluding antigen tethered on the PMS surface (Fig. 3f). fold at 24–48 h after B cells were treated with anti-IgM alone; this In the absence of CpG, B cells internalized HEL from the PMS, but was antagonized by the addition of CpG, via a TLR9-dependent in the presence of CpG, the internalization of HEL was much less, process (Supplementary Fig. 4c,f). However, treatment with CpG as measured by either the proportion of the HEL in the bilayer that alone induced a four- to fivefold increase in the surface expression was internalized or the total HEL internalized (Fig. 3f). We also of MHC class II (Supplementary Fig. 4c,f), which indicated that determined the degree of colocalization of the BCR and HEL, and CpG initiated at least two signaling pathways through TLR9: one of HEL and fragments of the PMS, in three-dimensional recon- that antagonized BCR-induced expression of MHC class II, and one structed images of the confocal z-stacks of B cells placed on HEL- that independently enhanced the expression of MHC class II. We containing PMSs for 30 min (Fig. 3g and Supplementary Video 2a,b). also assessed the effect of CpG on the ability of HEL-specific B cells HEL, BCR and fragments of the PMS above the contact area with from MD4 mice to process HEL and present complexes of HEL pep- the PMS surface indicated internalization (Fig. 3g). In the absence tide and the MHC class II molecule I-Ak (HEL–I-Ak) on the B cell of CpG, B cells were frequently observed in contact areas from surface; for this, we used a monoclonal antibody (mAb) specific for which the PMS had been cleared, in contrast to CpG-treated B cells, HEL–I-Ak. After 24 h of incubation with HEL, HEL-specific B cells which were observed in areas covered by the PMS. Videos of three- significantly increased their cell-surface abundance of HEL–I-Ak dimensional surface images showed that for both CpG-treated by approximately 2.5-fold, and the addition of CpG blocked this cells and untreated cells, little antigen was retained on the B cell increase (Fig. 4e,f). surface. In the xy and xz cross sections of B cells not treated with Neither treatment with CpG nor treatment with anti-IgM had CpG, the BCR and HEL and PMS fragments to which HEL was a large effect on CD80 expression at 24 h (Supplementary Fig. 4b); bound were evidently inside the cell (Supplementary Video 2a,b) however, after 48 h, the cell-surface expression of CD80 increased where the BCR and HEL colocalized (Fig. 3h), as did HEL and PMS approximately twofold for cells treated with CpG alone, anti-IgM fragments (Fig. 3i). However, for CpG-treated B cells, although the alone or both together, and this showed neither antagonism nor BCR and HEL were present intracellularly, very few PMS fragments synergy (Supplementary Fig. 4b). In contrast, treatment with anti- were evident (Fig. 3g). The colocalization of HEL and the BCR in IgM induced a large increase in the expression of CD86 that was CpG-treated B cells was similar to that in untreated cells (Fig. 3h); blocked by CpG (Supplementary Fig. 4a,e). The regulation of the however, the colocalization of HEL and fragments of the PMS was surface expression of CD69 showed a pattern consistent with a sim- significantly lower in CpG-treated B cells than in untreated cells ple additive effect of anti-IgM and CpG (Supplementary Fig. 4d,g). (Fig. 3i). The lack of PMS fragments in CpG-treated B cells sug- We assessed the effects of several inhibitors of kinases in the BCR gested that antigen taken up by BCRs in the presence of CpG did and TLR9 signaling pathways on the induced expression of MHC not involve pulling the antigen and attached PMS into the cell but class II and CD86. The most consistent effects of the inhibitors were was probably the result of the internalization of antigen that was not on BCR-induced changes in the expression of both MHC class II tightly attached to the PMS. and CD86, although these effects were only partial (Fig. 4g,h). In contrast, the TLR9-induced changes in the expression of MHC class Gene transcription and the expression of B cell surface markers. II and CD86 were unaffected by the inhibitors, which suggested, in RNA-based next-generation sequencing (RNA-seq) analysis was both cases, the possibility of redundancy in the signaling pathways. carried out on untreated B cells and B cells treated for 4 h with anti- Together these results provided evidence that signaling through IgM or CpG alone or together. Venn diagramming (Fig. 4a) and the BCR, TLR9 or both triggered distinct transcriptional programs principle-component analysis (Fig. 4b) of the results showed notable and had differential effects on the expression of cell-surface mark- separation of the transcription profiles of B cells under the different ers. Of greatest interest was the antagonistic effect of TLR9 signaling stimulation conditions. Thus, treatment of B cells with anti-IgM on BCR-induced increases in the abundance of peptide–MHC class and/or CpG resulted in the activation of distinct transcriptional II complexes and the expression of CD86.

a bc 100 100

NIH3T3 cells B cells 80 80 HEL MD4

60 60 20 min HEL WT

40 40 Mock MD4 ed to mode ed to mode 20 20 Mock WT maliz maliz HEL MD4 Nor Nor 0 0 –103 0 103 104 105 –103 0 103 104 105 AF647 APC HEL WT 40 min NIH3T3-mock + isotype (AF647) NIH3T3-mock + isotype (APC) Mock MD4 NIH3T3-mock + anti-CD4 (AF647) NIH3T3-mock + anti-HEL (APC)

NIH3T3-HEL + isotype Cont. (AF647) NIH3T3-HEL + isotype (APC) Mock WT

NIH3T3-HEL + anti-CD4 (AF647) NIH3T3-HEL + anti-HEL (APC) –103 0 103 104 105 HEL acquired by B cells (APC) NIH3T3-mock deP < 0.0001 fg NIH3T3-HEL 30 4 100 100 P = 0.0021 20 P < 0.0001 )

80 80 4 3 10 P < 0.0001 60 60 old) 6 2 40 MFI (f 40 4 ed to mode ed to mode 1 CD44 MFI (× 10 20 20

2 maliz maliz

Nor 0 0 Nor 0 0 3 3 4 5 3 3 4 5 k CpG ––++ ––++ ––++ –10 0 10 10 10 –10 0 10 10 10 e450 CD44 (APC) CD86 MHCII CD69 NIH3T3-moc NIH3T3-HEL NIH3T3-mock NIH3T3-HEL NIH3T3-HEL + CpG NIH3T3-HEL + CpG

Fig. 6 | TLR9 signaling decreases the ability of antigen-specific B cells to activate antigen-specific helper T cells in response to membrane bound antigen. a,b, Surface expression of the HEL–CD4 chimeric membrane protein on NIH3T3-HEL or NIH3T3-mock cells (key), quantified with an Alexa Fluor 647 (AF647)-labeled mAb specific for CD4 (a) or an allophycocyanin (APC)-labeled mAb specific for HEL (b) or the corresponding isotype-matched control mAb (key). c, Flow cytometry of the acquisition of HEL by HEL-specific MD4 B cells or nonspecific (WT) B cells (right margin) incubated for 20 or 40 min (left margin) with NIH3T3-mock or NIH3T3-HEL cells (right margin), then fixed, permeabilized and stained with a HEL-specific mAb. d, Expression of CD86, MHC class II and CD69 (below plots) by HEL-specific MD4 B cells incubated for 24 h with NIH3T3-mock or NIH3T3-HEL cells (key) in the presence (+) or absence (–) of 1 µM CpG (horizontal axis); results are presented relative to those of B cells cocultured with NIH3T3-mock in the absence of CpG. e–g, Proliferation (e) and CD44 expression (f,g) of HEL-specific 3A9 CD4+ T cells cultured for 72 h with HEL-specific MD4 B cells and either NIH3T3-mock cells or NIH3T3-HEL cells in the presence or absence of CpG (1 µM). Each symbol (d,g) represents an individual biological replicate. P values (above plots, d,g), two-sided unpaired t-test. Data are representative of three independent experiments in triplicate (mean values).

CpG blocks B cell–helper T cell interactions in vitro. We assessed of either 0.5 µg/ml or 2.0 µg/ml), HEL-specific 3A9 T cells were the effect of CpG on the length of time antigen-specific B cells and T induced to increase their expression of the activation and memory cells interacted in vitro in the presence of antigen. We labeled HEL- marker CD44 (Fig. 5d and Supplementary Fig. 4h) and proliferate specific T cells from 3A9 mice (which have transgenic expression of (Fig. 5e). The addition of CpG to the cultures (at a concentration of a HEL-specific TCR) with the cell-tracking fluorescent dye CMTPX either 2.0 µM or 0.2 µM) significantly reduced both CD44 expres- and labeled HEL-specific B cells from MD4 mice with cell-tracking sion and T cell proliferation. In the absence of B cells, T cells did fluorescent dye CMFDA, then incubated those cells in culture with not respond to either HEL or CpG. The ability of CpG to antago- HEL alone, CpG alone or HEL plus CpG, imaging the cells continu- nize the activation of HEL-specific T cells by HEL-specific B cells ously over a period of 24 h at a rate of 1 frame per 10–20 min. HEL was dependent on the expression of TLR9 by B cells but not on the induced a significant increase in the duration of colocalization of expression of TLR9 by T cells (Fig. 5f,g). TLR9-sufficent (‘wild- T cells and B cells, and this was antagonized by CpG (Fig. 5a). The type’) or TLR9-deficient 3A9 T cells cultured with wild-type MD4 presence of HEL also resulted in a decrease in both the speed and B cells and HEL proliferated and increased their expression of the length of the tracks of T cells, an effect that was mitigated by CD44; this was blocked by the presence of CpG. In contrast, CpG CpG (Fig. 5b,c). had no effect on the proliferation or CD44 expression of either We assessed the effect of TLR9 signaling on the ability of anti- wild-type 3A9 T cells or TLR9-deficient 3A9 T cells induced by gen-specific B cells to activate antigen-specific T cells in vitro. In the TLR9-deficient MD4 B cells. CpG had no effect on the ability presence of HEL-specific MD4 B cells and HEL (at a concentration of wild-type CD4+ T cells to increase their expression of CD44

μMT–WT μMT–MyD88-KO a μMT–MyD88-KO μMT–WT 5 bcP = 0.0012 d 10 2.0 40 P < 0.0001 3 × 105 P = 0.0096 4 10 2.53 3.39 ) 103 6 1.5 30 2 × 105 102 0 1.0 20 –102 5 CD95 (Percp-e710) 3 4 5 3 4 5 1 × 10 0 10 10 10 0 10 10 10 0.5 10 GL-7 (e660) GC B cells (× 10 NP-specific GC B cells

250 K NP-specific GC B cells (% ) 27.5 8.58 0.0 0 0 200 K μMT–WT μMT–MyD88-KO 150 K fg h 5 4 P < 0.0001 4 × 10 P = 0.0185 100 K ) P < 0.0001 6 40 – 50 K 5 B cells IgD

– 3 × 10

3 + FS C 0 30 3 3 4 5 3 3 4 5 IgD

–10 0 10 10 10 –10 0 10 10 10 + 2 × 105 2 20

NP (PE) B cells (× 10 – B cells (%) 5 e MT–MyD88-KO MT–WT Ig D μ μ + 1 10 1 × 10

5 NP-specific IgG

10 2.72 5.36 IgG

0 0 NP-specific IgG 0 104 μMT–WT μMT–MyD88-KO μMT–MyD88-KO 103 jk l 1.5 200 0 P = 0.0007 )

6 P = 0.0333 3

IgG (AF488 ) –10 3 3 4 5 3 3 4 5 150 –10 0 10 10 10 –10 0 10 10 10 1.0 IgD (BV786) 250 K 100

38.5 6.60 splenocytes) 200 K 6 0.5 MT–WT 50 μ 150 K (per 10 100 K PC-lineage cells (× 10

0.0 NP-specific IgG-secreting cells 0 50 K

FSC 0 –103 0 103 104 105 –103 0 103 104 105 μMT–WT μMT–MyD88-KO NP (PE) n 10 NS NS i μMT–MyD88-KO μMT–WT 5 8 NS

10 0.81 2.08 ) 3 * ** 4 6 *** 10 * 4 103 *** (AU × 10 *** 2 NS

0 NP-specific total IgG 0 CD138 (PE) 10 20 30 40 50 60 70 80 90 100 00103 104 102 103 104 105 Time (d) B220 (AF700)

μMT–WT μMT–MyD88-KO m o 10 NS * ** ** ** ** *** ***** 12 * * 8

) *** ) 3

* 3 ** 6 * 8 ** * *** 4 (AU × 10 4 (AU × 10 NP-specific IgM 2 NS

0 0 10 20 30 40 50 60 70 80 90 NP-specific high affinity IgG 10 20 30 40 50 60 70 80 90 100 Time (d) Time (d)

Fig. 7 | B cell–intrinsic expression of MyD88 affects the outcome of T cell dependent antibody response in vivo. a, Flow cytometry of splenocytes from chimeric mice with MyD88-deficient B cells (µMT–MyD88-KO) or wild-type B cells (µMT–WT) (above plots) (Supplementary Fig. 5a) on day 14 after intraperitoneal immunization with NP-CGG (100 µg per mouse) adsorbed onto alum (100 µl per mouse) and CpG (65 µg per mouse) (gating strategy, Supplementary Fig. 5c). Numbers adjacent to outlined areas indicate percent GC B cells (CD95+GL-7+) among B cells (top row) or NP-specific cells among GC B cells (bottom row). b–d, Total GC B cells (b), frequency of NP-specific GC B cells among all GC B cells (c) and total NP-specific GC B cells (d) from chimeras as in a (key). e, Flow cytometry of cells from chimeras as in a (above plots). Numbers adjacent to outlined areas indicate percent IgG+IgD– B cells among B cells (top row) or NP-specific cells among those cells (bottom row). f–h, Total IgG+IgD– cells (f), frequency of NP-specific B cells among all IgG+IgD– B cells (g) and total NP-specific IgG+IgD– B cells (h) from chimeras as in e (key). i, Flow cytometry of cells from chimeras as in a (above plots). Numbers above outlined areas indicate percent cells of the plasma cell lineage (CD138+B220neg–lo). j,k, Total PC-lineage cells (j) and NP-specific IgG– secreting cells (k) among cells from chimeras as in i (key). (l) ELISPOT analysis of cells from chimeras as in k (left margin). m–o, NP-specific IgM (m), NP-specific IgG (n) and high-affinity NP-specific IgG (o) in serum from chimeric mice as in a (key) on days 7–91 after immunization as in a (horizontal axis), presented as arbitrary units (AU) calculated from serial dilutions of pooled serum. Each symbol (b–d,f–h,j,k) represents an individual mouse; dashed horizontal lines indicate the mean. P values (above plots in b–d,f–h,j,k); NS, not significant (P > 0.05); *, 0.01 < P ≤ 0.05; **, 0.001 < P ≤ 0.01; and ***, 0.0001 < P ≤ 0.001 (two-sided Welch’s t-test). Data are pooled from two independent experiments (a–l) or represent two independent experiments (m–o; mean and s.d. of n = 8 mice per group).

a No CpG c 8 × 10–8 2,500 100 CpG Glycine pH elution gradient (%) P = 0.4726

2,000 80 6 × 10–8 1,500 60 (M) D (mAU) 1,000 40 –8 280 4 × 10 A

500 20 Apparent K 2 × 10–8 0 0 0306090 120 Elution volume (ml) 0 No CpG CpG

b d 1.5 × 10–2 300 No CpG CpG P = 0.3058

200 1.0 × 10–2 (1/s) of f k 100 –3

Response units 5.0 × 10

0 0.0 0 200 400 600 No CpGCpG Time (s)

Fig. 8 | CpG does not induce high-affinity antibodies in humans. a, Chromatography profiles of serum obtained from human subjects on day 70 after immunization with the blood-stage malaria vaccine candidate AMA1-C1, formulated on aluminum-hydroxide gel with or without 564 µg CpG 7909 (key), assessing IgG antibodies purified with protein G columns (left vertical axis); IgG results are presented (in milli-absorbance units (mAU)) as absorbance at

280 nm (A280). b, Surface plasmon resonance analysis of the binding of PfAMA1 and IgG (70 nM) in serum from donors as in a (key), presented as binding sensorgrams. c,d, Surface plasmon resonance analysis of serially diluted purified antibodies from donors as in b (horizontal axis), showing the apparent binding affinity (KD) (c) and off-rate (koff) values (d). Each symbol (c,d) represents an individual donor. P values (above plots, c,d), two-sided Welch’s t-test. Data are representative of one experiment (a), three sample runs (b) or the average of three sample runs (c,d).

(Fig. 5h) or proliferate (Fig. 5i) in response to stimulation with mice in which the B cells were either wild-type or deficient in the plate-bound anti-CD3 and anti-CD28. TLR9 adaptor MyD88 and all other hematopoietic cells were wild- We also assessed the effect of CpG on the ability of B cells to type (Supplementary Fig. 5a). Reconstitution in the two chimeras respond to and capture HEL from cell surfaces and to process and was similar (Supplementary Fig. 5b). MyD88 deficiency in B cells present it to HEL-specific 3A9 T cells using NIH3T3 mouse fibro- is not predicted to affect the intrinsic ability of B cells to participate blasts stably expressing a chimeric membrane protein of HEL and in T cell–dependent antibody responses, as robust T cell–depen- the monomorphic co-receptor CD4 via a retroviral vector (NIH3T3- dent antibody responses are elicited by immunization of MyD88- HEL cells) or transduced with a mock vector (NIH3T3-mock cells) deficient mice in vivo11. We immunized the chimeric mice with (Fig. 6a,b). HEL-specific MD4 mouse B cells (but not nonspecific alum-adsorbed nitrophenyl–chicken γ-globulin (NP-CGG) plus wild-type B cells) captured HEL, as detected by flow cytometry with CpG, harvested the spleens on day 14 after immunization and ana- a HEL-specific mAb (Fig. 6c), and increased their expression of lyzed the cells by flow cytometry according to the appropriate gating CD86, MHC class II and CD69 (Fig. 6d). For both CD86 and MHC strategy (Supplementary Fig. 5c). Although the spleen of chimeras class II, the presence of CpG blocked this increase, but it had a syner- with wild-type cells had a larger total number of GC B cells than gistic effect on the increase in CD69 expression (Fig. 6d). Co-culture that in the spleen of chimeras with MyD88-deficient cells (Fig. 7a,b), of HEL-specific MD4 B cells and 3A9 T cells in the presence of the frequency and total number of GC B cells that were NP spe- NIH3T3-HEL cells resulted in the activation of T cells to proliferate cific was significantly higher in chimeras with MyD88-deficient (Fig. 6e) and increase their expression of CD44 (Fig. 6f,g), but such cells than in chimeras with wild-type cells (Fig. 7c,d). Similarly, the co-culture in the presence of NIH3T3-mock cells did not, and both total number of IgG-class-switched B cells was higher in chimeras of those effects were diminished in the presence of CpG. with wild-type cells than in chimeras with MyD88-deficient cells Together these results demonstrated that in the presence of CpG, (Fig. 7e,f), but the opposite was true for the frequency and total antigen-specific B cells were unable to present soluble antigen or number of NP-specific B cells (Fig. 7g,h). The total number of cells antigen captured from cell surfaces and activate antigen-specific of the PC lineage was greater in chimeras with wild-type cells than helper T cells. in chimeras with MyD88-deficient cells (Fig. 7i,j), but the number of NP-specific antibody–secreting cells was greater in the spleen of The effect of CpG on B cell responses in vivo. To investigate the chimeras with MyD88-deficient cells than in that of chimeras with effect of CpG on B cell responses in vivo, we generated chimeric wild-type cells (Fig. 7k,l). We also determined the total number

+ + + of TFH cells (CD4 CXCR5 PD1 ) in the spleen of the chimeric by TFH cells is dependent on the density of peptide–MHC complexes mice at 14 d after immunization and found no significant differ- on the B cell surface, reflective of the B cell’s affinity-dependent abil- ence between the chimeras in the number of TFH cells, although ity to compete for the capture of antigen from follicular dendritic 3 the chimeras with wild-type cells tended to have more TFH cells cells . Thus, the processing and presentation of antigen by B cells (Supplementary Fig. 5d,e). However, without tools to determine the controls, to a large extent, the clonal selection of B cells. frequency of TFH cells that were antigen specific, that observation Given the central role of the gathering, processing and presenta- was difficult to interpret, given the differences observed for total tion of antigen by B cells in the generation of high-affinity, long-lived GC B cells versus antigen-specific GC B cells, isotype-switched B antibody responses, it is likely that this B cell function is regulated to cells and PCs (Fig. 7). ensure appropriate and timely responses to antigenic challenge, par- Chimeras with wild-type B cells showed higher NP-specific IgM ticularly during infection. However, little is known about the role serum responses than those of chimeras with MyD88-deficient B of receptors of the innate immune system that respond to micro- cells (Fig. 7m). For total NP-specific IgG, although the peak lev- bial products containing pathogen-associated molecular patterns in els at day 30 after immunization were not significantly different regulating these processes. We found that TLR9 signaling promoted in immunized chimeras (Fig. 7n), the levels of NP-specific IgG B cell proliferation and the differentiation of B cells into antibody dropped faster in chimeras with wild-type B cells than in those with secreting cells, as previously shown7,22,23, and, in addition, that it MyD88-deficient B cells. The most notable differences were in the triggered changes in gene expression, including the downregulation levels and persistence of high-affinity NP-specific IgG (Fig. 7o), of the gene encoding BCL6, a transcriptional repressor of PC differ- which reached higher peak levels and persisted for weeks longer in entiation, and upregulation of the gene encoding BLIMP1, a tran- chimeras with MyD88-deficient B cells than in those with wild-type scription factor that promotes PC differentiation. However, TLR9 B cells (Fig. 7o). Thus, CpG drove the production of IgM and short- signaling greatly diminished the ability of antigen-specific B cells lived antigen-specific IgG responses that did not undergo affinity to gather, process and present antigen and to interact with and acti- maturation in vivo. vate antigen-specific CD4+ T cells in vitro. In the presence of CpG, antigen-specific B cells also failed to upregulate their expression of The effect of CpG on antibody affinity in humans. A longitudinal important costimulatory ligands for T cells, including CD86. The study has been done of the antibody response to a candidate malaria ability of CpG to diminish the ability of B cells to capture antigen vaccine administered with or without CpG to healthy malaria-naive from membranes and cell surfaces, as shown here, is particularly people in the US (trial NCT00320658 at https://www.clinicaltrials. relevant to the regulation of B cell responses in vivo, where the B gov/)20. The vaccine was composed of recombinant Plasmodium fal- cells in lymphoid organs access most antigens on the surface of fol- ciparum apical membrane protein 1 (PfAMA1) formulated on the licular dendritic cells24. adjuvant aluminum-hydroxide gel and mixed with 564 µg of CpG Such findings have implications for the outcome of T cell–depen- 7909. The inclusion of CpG in the vaccine resulted in significantly dent antibody responses in vivo. Our data suggest that if B cells higher titers of PfAMA1-specific IgG but did not have an effect on encounter TLR9 ligands during an infection, the B cells would have the longevity of the responses20. diminished ability to gather, process and present antigen to antigen- Using serum samples from that trial20, we determined the effect activated helper T cells at the T cell–B cell border of lymphoid organs. of the inclusion of CpG in the vaccine on the affinity of the IgG As a consequence, B cells would be directed to GC-independent path- response. Serum was collected at 70 d after vaccination from ten ways and would yield ‘affinity-unselected’ PCs and GC-independent people who received the vaccine alone and ten who received the vac- memory B cells. It might be advantageous to rapidly produce anti- cine containing CpG, and IgG was purified on protein G columns bodies with a wide range of affinities in response to activation of (Fig. 8). The chromatographic profiles of IgG from all people were TLR9 during an acute infection and to leave an unselected broad similar, and the IgG antibodies seemed to be very pure (Fig. 8a). B cell repertoire for future antigen challenge2. Such a mechanism The purified IgG antibodies were analyzed by BIACore with might underlie the observation that, in a nonhuman primate model, PfAMA1 recombinant protein adsorbed onto CM5 chips. There although adjuvants that stimulate receptors of the innate immune was no difference between people who received the vaccine alone system, including TLRs, were effective in enhancing the magnitude and those who received the vaccine containing CpG, in terms of of antibody responses to protein immunogens, remarkably, these apparent affinity (Fig. 8b,c) or off rate (Fig. 8b,d). Thus, although were not able to increase somatic hypermutation13. the inclusion of CpG resulted in an increase in the levels of antigen- We have presented evidence both in mice and in humans con- specific antibody20, CpG caused no detectable increase on the affin- sistent with that model. We compared the responses of chimeric ity of these antibodies. mice containing wild-type B cells with the responses of chimeric mice containing MyD88-deficient B cells after immunization with Discussion an alum-adsorbed protein antigen (NP-CGG) given together with It was established over 30 years ago that the processing and presen- CpG. The ability of B cells to respond to CpG resulted in short-lived tation of antigen to CD4+ helper T cells by antigen-specific B cells low-affinity IgG responses and reduced numbers of antigen-specific is highly efficient21. However, although it was clear that antigen- GC B cells. We provided evidence that in humans, the inclusion of specific B cells are able to efficiently process and present antigen, CpG in a recombinant protein vaccine formulated on alum resulted the role of antigen processing and presentation by B cells in T cell– in increased levels of antigen-specific antibodies but had no detect- dependent antibody responses was not elucidated until recently. able effect on the apparent affinity of the antibodies, relative to the Several studies have now provided evidence that the ability of B response of people who received the vaccine alone. To our knowl- cells to process and present antigen to antigen-activated helper T edge, the studies presented here and a published study13 are the only cells and to TFH cells provides critical checkpoints for the entry of ones to address the effect of TLR ligands as adjuvants on antibody B cells into the GC and for the affinity selection of B cells express- affinity maturation. It will be of interest to determine the generality ing somatically hypermutated immunoglobulin-encoding genes of these findings for other vaccines and vaccine regimens. in the GC1–4. The quality of the B cell’s presentation of antigen to antigen-activated T cells at the T cell–B cell border in lymphoid Methods organs determines, in part, if a B cell will enter the GC or differenti- Methods, including statements of data availability and any asso- ate along a GC-independent pathway to become short-lived PCs or ciated accession codes and references, are available at https://doi. GC-independent memory B cells2. In the GC light zone, selection org/10.1038/s41590-018-0052-z.

Received: 9 March 2017; Accepted: 10 January 2018; 18. Liu, W., Meckel, T., Tolar, P., Sohn, H. W. & Pierce, S. K. Intrinsic properties Published online: 23 February 2018 of immunoglobulin IgG1 isotype-switched B cell receptors promote microclustering and the initiation of signaling. Immunity 32, 778–789 (2010). 19. Akkaya, B. et al. A simple, versatile antibody-based barcoding method for fow cytometry. J. Immunol. 197, 2027–2038 (2016). References 20. Crompton, P. D. et al. Te TLR9 ligand CpG promotes the acquisition of 1. De Silva, N. S. & Klein, U. Dynamics of B cells in germinal centres. Nat. Rev. Plasmodium falciparum-specifc memory B cells in malaria-naive individuals. Immunol. 15, 137–148 (2015). J. Immunol. 182, 3318–3326 (2009). 2. Kurosaki, T., Kometani, K. & Ise, W. Memory B cells. Nat. Rev. Immunol. 15, 21. Lanzavecchia, A. Antigen presentation by B lymphocytes: a critical step in 149–159 (2015). T-B collaboration. Curr. Top. Microbiol. Immunol. 130, 65–78 (1986). 3. Mesin, L., Ersching, J. & Victora, G. D. Germinal center B cell dynamics. 22. Eckl-Dorna, J. & Batista, F. D. BCR-mediated uptake of antigen linked to Immunity 45, 471–482 (2016). TLR9 ligand stimulates B-cell proliferation and antigen-specifc plasma cell 4. Bannard, O. & Cyster, J. G. Germinal centers: programmed for afnity formation. Blood 113, 3969–3977 (2009). maturation and antibody diversifcation. Curr. Opin. Immunol. 45, 23. Genestier, L. et al. TLR agonists selectively promote terminal plasma cell 21–30 (2017). diferentiation of B cell subsets specialized in thymus-independent responses. 5. DeFranco, A. L., Rookhuizen, D. C. & Hou, B. Contribution of Toll-like J. Immunol. 178, 7779–7786 (2007). receptor signaling to germinal center antibody responses. Immunol. Rev. 247, 24. Hoogeboom, R. & Tolar, P. Molecular mechanisms of B cell antigen gathering 64–72 (2012). and endocytosis. Curr. Top. Microbiol. Immunol. 393, 45–63 (2016). 6. Rawlings, D. J., Schwartz, M. A., Jackson, S. W. & Meyer-Bahlburg, A. Integration of B cell responses through Toll-like receptors and antigen Acknowledgements receptors. Nat. Rev. Immunol. 12, 282–294 (2012). We thank S. Akira (Osaka University) for TLR9-deficient mice; I. Gery (US National 7. Akkaya, M. et al. B cells produce type 1 IFNs in response to the TLR9 agonist Institutes of Health) for 3A9 mice; N. Barclay (University of Oxford) for mouse mAb CpG-A conjugated to cationic lipids. J. Immunol. 199, OX68; P. Roche (US National Institutes of Health) for the hybridoma secreting mouse 931–940 (2017). mAb AW3.18; L. Lantz (US National Institutes of Health) for producing purified mouse 8. Roche, P. A. & Furuta, K. Te ins and outs of MHC class II-mediated antigen mAb AW3.18; J. Charles-Guery and D. Hudrisier (University of Toulouse) for mouse processing and presentation. Nat. Rev. Immunol. 15, 203–216 (2015). mAb F10.6; O. Voss (US National Institutes of Health) for the NIH3T3 mouse fibroblast 9. Rookhuizen, D. C. & DeFranco, A. L. Toll-like receptor 9 signaling acts on cell line; P. Tolar (The Francis Crick Institute) for Matlab code; and J. Brzostowski and multiple elements of the germinal center to enhance antibody responses. Proc. S. Bolland for suggestions. Supported by the Intramural Research Program of the US Natl Acad. Sci. USA 111, E3224–E3233 (2014). National Institutes of Health, National Institute of Allergy and Infectious Diseases. 10. Akkaya, M. et al. T cell-dependent antigen adjuvanted with DOTAP-CpG-B but not DOTAP-CpG-A induces robust germinal center responses and high afnity antibodies in mice. Eur. J. Immunol. 47, 1890–1899 (2017). Author contributions 11. Gavin, A. L. et al. Adjuvant-enhanced antibody responses in the absence of M.A. and S.K.P. conceived of the project, designed the experiments and wrote the toll-like receptor signaling. Science 314, 1936–1938 (2006). manuscript; M.A., B.A., A.S.K., P.M., H.S., M.P., A.S.R., B.P.T., T.H. and J.L. carried out 12. Cofman, R. L., Sher, A. & Seder, R. A. Vaccine adjuvants: putting innate the experiments; M.A., B.A., A.S.K., P.M., H.S., J.K., J.L., D.W.D., E.D., J.S., L.H.M. and immunity to work. Immunity 33, 492–503 (2010). S.K.P. analyzed the data; and S.K.P. secured funding. 13. Francica, J. R. et al. Analysis of immunoglobulin transcripts and hypermutation following SHIV(AD8) infection and protein-plus-adjuvant Competing interests immunization. Nat. Commun. 6, 6565 (2015). The authors declare no competing financial interests. 14. Basso, K. & Dalla-Favera, R. Roles of BCL6 in normal and transformed germinal center B cells. Immunol. Rev. 247, 172–183 (2012). 15. Fleire, S. J. et al. B cell ligand discrimination through a spreading and Additional information contraction response. Science 312, 738–741 (2006). Supplementary information is available for this paper at https://doi.org/10.1038/ 16. Weber, M. et al. Phospholipase C-γ2 and Vav cooperate within signaling s41590-018-0052-z. microclusters to propagate B cell spreading in response to membrane-bound Reprints and permissions information is available at www.nature.com/reprints. antigen. J. Exp. Med. 205, 853–868 (2008). 17. Tolar, P., Sohn, H. W., Liu, W. & Pierce, S. K. Te molecular assembly and Correspondence and requests for materials should be addressed to S.K.P. organization of signaling active B-cell receptor oligomers. Immunol. Rev. 232, Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in 34–41 (2009). published maps and institutional affiliations.

Methods lymph nodes of 3A9 mice. Both cells were labeled with Cell Proliferation Dye 4 Mice. C57BL/6, B10.BR-H2K2H2-T18a/SgSnJArc, B cell–defcient (B6.129S2- eFluor 450 (ThermoFisher), mixed at 1:1 ratio (7.5 × 10 each) in complete medium Ighmtm1Cgn/J) (µMT) mice, B6.129S2-Ighmtm1Cgn/J, MyD88-defcient mice and and incubated in 96-well plates at 37 °C in the presence of HEL (1 µg/ml) and/or B6 CD45.1+ congenic wild-type mice were purchased from Jackson Laboratory. CpG (1 µM). Stimulation by membrane-bound antigens was carried out by plating HEL- specifc BCR transgenic mice (MD4)25, were purchased from Taconic Farms. NIH3T3 cells, transfected to stably express HEL or mock transfected, in 96-well 4 TLR9-defcient mice were provided by S. Akira (Osaka University, Osaka, Japan). plates, 5 × 10 cells per well, 1 d before adding the B cell–T cell mixture. Mice with transgenic expression of a HEL-specifc, MHC class II I-Ak–restricted Visualization of real time interactions between B cells and T cells was carried TCR (3A9)26, were provided by I. Gery (NEI/NIH, Bethesda, MD, USA). Both out by labeling B cells and T cells with CellTracker Green CMFDA and CellTracker MD4 mice and 3A9 mice were kept hemizygous for the transgenic alleles in an Red CMTPX (ThermoFisher), respectively. Labeled cells were plated into 48-well 5 MHC class II I-Ak or I-Ab haplotype by crossing 3A9 or MD4 mice with wild-type plates at 1:1 ratio (1 × 10 each) and stimulated as indicated above. Interactions mice of the relevant haplotype. TLR9-defcient MD4 and 3A9 mice were generated between cells were recorded as 10- to 20-min time-lapse photos by IncuCyte by crossing the MD4 and 3A9 mice with TLR9-defcient mice for at least three ZOOM system (Essen Bioscience) for 24 h. Image analysis was performed generations. TLR9-defcient mice on the MHC class II I-Ak haplotype were using Imaris Software. generated as the F2 generation of a cross between TLR9-defcient mice (I-Ab) and B10.BR-H2K2H2-T18a/SgSnJArc (I-Ak) mice. Tese TLR9-defcient I-Ak mice Human study population and vaccination procedures. Serum was obtained were further crossed (at least four generations) with TLR9-defcient MD4 or 3A9 70 d after immunization from malaria-naive adults enrolled in a phase 1 clinical mice to obtain TLR9-defcient MD4 or 3A9 mice in the I-Ak haplotype. Mice were trial of the blood-stage malaria vaccine candidate AMA1-C1 (80 µg), formulated 32 bred and maintained in NIH animal facility according to Animal Care and Use on Alhydrogel and mixed with CpG 7909 (564 µg) (trial NCT00320658 at 20 Committee Standards. https://www.clinicaltrials.gov) as detailed elsewhere . The trial was conducted under Investigational New Drug Applications reviewed by the US Food and Antibodies and proteins. Antibodies specific for the following mouse molecules Drug Administration and was reviewed and approved by the National Institute of were purchased from BioLegend: CD19 (Clone:6D5) (conjugated to Alexa Fluor Allergy and Infectious Diseases Review Board and by the University of Rochester 488, Alexa Fluor 700, BV650, PeCy7 or BV421); CD45R/B220 (Clone:RA3-6B2) Institutional Review Board and funding agency. Written informed consent was (conjugated to PE, Alexa Fluor 488, Alexa Fluor 647, Alexa Fluor 700, BV421, obtained from all participants. Serum samples from 20 subjects were randomly BV785, BV605 or BV510); CD80 (Clone:16-10A1) (conjugated to Alexa Fluor 647); selected for analysis; half of the subjects had been vaccinated with CpG7909- CD86 (Clone: GL1) (conjugated to PERCP, PERCP-Cy5.5 or Alexa Fluor 488); containing vaccine and half had been vaccinated with the vaccine alone. IgM (Clone:RMM-1) (conjugated to BV605); CD69 (Clone: H1.2F3) (conjugated to PE or BV605); CD4 (Clone:RM4-5) (conjugated to Alexa Fluor 488 or BV605); Generation of stable cell lines expressing HEL chimeric protein. NIH3T3 cells CD44 (Clone:IM7) (conjugated to Alexa Fluor 700 or Alexa Fluor 647); MHC stably expressing HEL on their surface were generated by transfection with a class II I-Ab (Clone:KH74) (conjugated to Alexa Fluor 488); MHC class II I-Ak construct encoding a protein chimera containing on its amino terminus the entire (Clone: 10-3.6)(conjugated to PE); IgG1 (Clone: RMG1-1)(conjugated to biotin or HEL protein (NCBI protein database accession code AAA48943.1), followed by Alexa Fluor 647) and IgG2a (Clone:RMG2a-62) (conjugated to Alexa Fluor 647), a serine-tyrosine linker sequence and then by the third and fourth transmembrane CD138 (Clone:281-2) (conjugated to PE), CD45.2 (Clone:104) (conjugated to and intracellular domains of rat CD4 (residues 206–457) (NCBI protein database BV650 or APC), CD45.1(Clone:A20) (conjugated to Alexa Fluor 700, Alexa Fluor accession code P05540.1). A codon optimized final construct was generated by 488 or APC), CD3e (Clone: 145-2C11), CD28 (Clone:37.51) and streptavidin (PE Geneart Gene Synthesis (ThermoFisher) and was then subcloned into a pFB-Hygro 33 conjugated). Isotype-matched control antibodies to mouse IgG1 (Clone:MG1-45) retroviral vector . Transduction of NIH3T3 cells using the Phoenix-Eco system 34 and IgG2a (Clone:MG2a-53) were purchased from BioLegend. Antibodies and selection of the stably transfected cells were done as previously described . specific for the following were purchased from BD Biosciences: p-Btk (p-Y223) (Clone:N35-86) (conjugated to PE); pMAPK/p38 (pT180/pY182) (Clone:36/ Bone marrow chimeras and immunizatios. To generate mice with B cell p38) (conjugated to Alexa Fluor 647); p-Syk (pY352) (Clone: 17 A/P-ZAP70) lineage–specific TLR9-deficiency, lethally irradiated 7- to 8-week-old C57BL/6 7 (conjugated to Alexa Fluor 488); pAkt (pS473) (Clone:M89-61) (conjugated to mice (CD45.1+ haplotype) were reconstituted with 1 × 10 bone marrow cells, Alexa Fluor 647); H2-M (Clone: 2E5A) (unconjugated) and CD71 (unconjugated) of which 90% were from µMT mice and 10% were from either MyD88-deficient (Clone C2F2) and, anti-mouse IgD (Clone:11-26 c.2a) conjugated to BV786. mice or wild-type mice. Mice were allowed to reconstitute for at least 8 weeks before intraperitoneal immunization with both 100 g NP -CGG (Biosearch Polyclonal F(ab′)2 anti-mouse IgG (conjugated to Alexa Fluor 647 or Alexa Fluor µ 15 Technologies) adsorbed into 100 l Alum (ThermoFisher) in PBS and 65 g 488) and polyclonal F(ab′)2 anti-mouse IgM (unconjugated and conjugated to µ µ Alexa Fluor 647) were purchased from Jackson ImmunoResearch. Antibodies CpG B (ODN 1826) (Invivogen) in PBS. specific for mouse CD95 (Clone:15A7) (conjugated to Percp e710) and mouse IgM Clone:eb121-15F9) (conjugated to e450) were purchased from ThermoFisher. Signaling pathway analysis. B cells isolated from mouse spleen were cultured Antibodies specific for LAMP-1 (Clone:1D4B) were purchased from Santa Cruz in 96-well plates at 37 °C with the appropriate stimulus. At each time point, cells Biotechnologies. PE-conjugated NP was purchased from Biosearch Technologies. were fixed with 4% PFA containing HBSS buffer for 10 min at 37 °C, followed The mouse mAb (OX68) specific for the third and fourth domains of rat CD4, by permeabilization on ice using pre-chilled Perm Buffer III (BD Biosciences) as described27, was provided by N. Barclay (University of Oxford, Oxford UK). for 30 min. Cells cultured with different stimuli were barcoded by staining with The hybridoma secreting mouse mAb (AW3.18) specific for the HEL peptide antibodies to B220 (identified above) conjugated to different fluorochromes. (residues 4862)–MHC class II I-Ak complex28, was provided by P. Roche Following barcoding as described19, samples from the same time point, each having (NCI/NIH, Bethesda, MD, USA) and purified antibody was produced by unique B220 staining patterns, were pooled into one well for intracellular staining L. Lantz (NIAID/NIH, Rockville, MD,USA). The mouse mAb (F10.6) specific for using phosphorylated kinase–specific antibodies (identified above). For flow HEL29, was provided by J. Charles-Guery and D. Hudrisier (University of Toulouse, cytometry, each sample was gated based on the barcodes. For each phosphorylated Toulouse, France). Fluorescent labeling of the unconjugated antibodies was done kinase–specific antibody, the net mean fluorescence intensity (MFI) was calculated using a LYNX kit purchased from Bio-Rad. by subtraction of the MFI of the fluorescence-minus-one (FMO) control from the MFI of the actual samples. The change in the phosphorylation content (fold value) Flow cytometry. Flow cytometry was performed on a BD LSRII flow cytometer was determined as the ratio of the net MFI of a stimulated condition to the net MFI and data were analyzed in Flowjo software (Tree Star). During analysis, dead of the unstimulated sample at the same time point. Changes in calcium influx into cells were excluded by staining with LIVE/DEAD fixable dead cell stain kits the cell as a response to different stimulation conditions were analyzed using 20 (ThermoFisher). Fluo-4 and Fura red dyes (ThermoFisher) as explained .

Cells. B cells and naive CD4+ T cells were isolated from mouse spleen using mouse Characterization of BCR-mediated antigen internalization. To quantify B cell and naive T cell isolation kits (Miltenyi Biotech) according to previously internalization of antigen from solution, B cells isolated from mice with transgenic published optimized protocols30,31. NIH3T3 mouse fibroblast cell line (ATCC expression of a HEL-specific BCR (MD4) were incubated with saturating amounts CRL-1658) was provided by O. Voss. Phoenix Eco retroviral packaging cell line (4 µg/ml) of soluble HEL (Sigma-Aldrich) on ice for 30 min. Unbound antigen was (ATCC CRL-3214) was purchased from American Type Culture Collection removed by repeated washes at 4 °C. Cells were cultured in complete medium with (ATCC). All cells were maintained in RPMI 1640 media supplemented with or without CpG at 37 °C, and aliquots were taken with time and transferred to ice 50 µM 2-mercaptoethanol, 50 U/ml penicillin, 50 µM streptomycin, 2 mM l- cold FACS buffer (HBSS supplemented with 5% FCS, 10 mM HEPES and 10 mM glutamine, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate and 10 mM sodium azide). Cells were stained with Alexa Fluor 647-conjugated antibody HEPES unless specified otherwise. The pathway inhibitors SP600125, PD98059, specific for mouse IgM and analyzed by flow cytometry. Downmodulation Wortmannin, Rapamycin, SB202190 were purchased from Invivogen. MHY1485 of the surface BCR was measured using the following formula: % receptor and Akt Inhibitor IV (Akt IV) were purchased from Sigma-Aldrich. downmodulation = 100 × [net MFI (tx)/net MFI (t0)]. Alternatively, B cells isolated C57BL/6 mouse spleens were incubated on ice B cell–T cell co-culture. B cells were isolated from the spleens of MD4 mice on with Alexa Fluor 647-conjugated anti-IgM (identified above) for 30 min, then the MHC class II I-Ak haplotype. Naive T cells were isolated from the spleen and were washed and cultured at 37 °C for increasing lengths of time. At each time

Nature Immunology | www.nature.com/natureimmunology © 2018 Nature America Inc., part of Springer Nature. All rights reserved. Articles NaTuRe ImmunoLoGy point, aliquots were transferred into ice cold RPMI1640 supplemented with BSA colocalization analyses, Imaris software (Bitplane, Belfast, UK) was used as (0.5%) either at pH 7.4 or at pH 2.5 to strip surface-bound anti-IgM. Following described39. To obtain the colocalization coefficient from internal compartments fixation with 4% PFA, cells were analyzed by flow cytometry and the portion of of the cells, only z-slices above the PMSs surface determined by Matlab code BCR internalized was determined using the following formula: % internalized described above were used for the three-dimensional reconstitution and calculation BCR = 100 × [net stripped MFI (tx)—net stripped MFI of 0 min time point]/[net of three-dimensional Pearson’s colocalization coefficient between internalized non-stripped MFI (tx)—net stripped MFI (t0)]. antigen and membrane taken from the PMS or between the internalized antigen To characterize the pH of the intracellular compartment into which the and BCR. BCR-bound HEL was internalized, the pH-sensitive fluorescent dye pHrodo Red Avidin (ThermoFisher) was conjugated to biotinylated HEL (SigmaAldrich) Confocal microscopy. Splenic B cells, purified from MD4 transgenic mice were according to manufacturer’s guidelines. B cells isolated from MD4 mice were attached to poly-l-lysine–coated cover glass. Cells were incubated at 37 °C in incubated on ice with pHrodo-HEL for 30 min, and then were washed and cultured complete medium containing Rhodamine-conjugated HEL (Nanocs) with or at 37 °C for increasing lengths of time in phenol red–free RPMI complete medium without CpG for increasing times. Cells were fixed with fixation buffer (4% PFA (Gibco-ThermoFisher) with or without CpG. At each time point, aliquots were in PBS at 37 °C) for 10 min, permeabilized with PBS containing 0.5% saponin and analyzed with flow cytometry to determine both the increase in total internalized 5% BSA at RT for 20 min, washed and incubated with PBS containing 10% normal pHrodo-HEL and the decrease in pH of the endosomal compartment. rat serum and 0.5% saponin at RT for 1 h. Cells were stained with fluorescent mAb specific for CD71, LAMP1 and H2M (identified above) and were stained Preparation of antigen-containing lipid bilayers and membrane sheets. Planar for early (CD71+) endosomes and late (LAMP1+) endosomes, as well as for the lipid bilayers (PLBs) were prepared as described35. In brief, PLBs containing biotin antigen-processing compartment (H2M+), using fluorescence-labeled monoclonal lipid were formed on the acid-cleaned glass bottom of an eight-well Lab-Tek antibodies diluted in blocking buffer mounted on microscope slides using ProLong chamber from 10 µM small unilamellar vesicles consisting of DOPC and DOPE- Gold Antifade Mountant with DAPI (Thermofisher). A Zeiss LSM780 microscope cap-biotin (Avanti Polar Lipids). HEL was incorporated into the PLB by adding equipped with a 63 × A Zeiss Plan-Apochromat 633 oil-immersion objective was 10 nM Alexa Fluor 488–labeled biotinylated HEL into PLB-containing chambers, used for image acquisition. Three-dimensional colocalization of rhodamine-HEL incubating for 30 min at RT, and then washing to remove unbound HEL. with the fluorescent antibodies was calculated using Imaris software. Plasma membrane sheets (PMSs) were prepared as described36. In brief, 5 1.5 × 10 293 A cells were seeded into a chamber of an eight-chamber Lab-Tek Electron microscopy. For transmission electron microscopy (TEM) and electron II chambered #1.5 German coverglass system (Nunc) and were cultured overnight. tomography (ET), B cells were fixed by suspension in modified Karnovsky’s The chambers were filled with blocking solution (HBSS containing 2% BSA and fixative prepared with 10 ml PFA (4%), 20 ml sodium phosphate buffer (0.2 M), 2+ 2.5 mM Ca ) and the cells were sonicated with a 130 W, 20 kHz probe-typed 2 ml glutaraldehyde (50%) and 8 ml distilled water (Electron Microscopy Sciences) ultrasonic processor at 22% amp for 5 s at a time by placing the probe right beneath and were processed using microwave irradiation as previously described40, with the top of buffer. To tether antigen on the PMSs, PMSs were incubated sequentially the following exceptions: between each step, cells were recovered by centrifugation for 30 min with 0.5 g/ml biotinylated annexin V (Biolegend), 5 g/ml streptavidin µ µ for 5 min at 800 × g. Fixed and dehydrated cells were infiltrated with Araldite (Invitrogen) and 50 nM Alexa Fluor 488–conjugated biotinylated HEL with resin (SPI). Sections for tomography were cut at a thickness of 200 nm and were washing between additions with 15 ml buffer (1 × HBSS containing 0.2% BSA imaged without post-section staining. Tomographic tilt series were collected on an 2+ and 2.5 mM Ca ). Ultrascan 4000 camera (Gatan), using SerialEM acquisition and control software (University of Colorado, Boulder, CO)41. Volumetric reconstructions Imaging and image analysis of B cells responding to membrane-bound antigen. were performed using Batchruntomo script (University of Colorado). Particle To image BCR clustering and antigen recruitment to the contact area between counts and volumetric measurements were conducted using IMOD software cells and PLBs and PMSs, time-lapse live-cell total internal reflection fluorescence (University of Colorado)42. Volumetric models were rendered using Amira for imaging was performed using an Olympus IX-81 (Olympus) equipped with a TIR Biosciences (FEI). controlling system, EMCCD camera and 100 × objective lens (Olympus) controlled by Metamorph software (Molecular Devices) as detailed elsewhere18,37, 38. Time- ELISAs and ELISPOT. Concentrations of cytokines (IL-2, IL-6, IL-10 and TNF) lapse live-cell imaging was done at 37 °C on a heated stage with an objective lens and immunoglobulins (IgG and IgM) secreted by purified splenic B cells in culture heater and a heating chamber with image acquisition at 2 s intervals for 20 min. were measured by analysis of the supernatants with relevant Ready-Set-Go! ELISA All acquired images were background-subtracted before further analysis. For the kits (ThermoFisher) according to the manufacturer’s guidelines. analysis of BCR accumulation in the contact area with time after the cells contacted The relative affinities of NP-specific antibodies were determined by the the PLB or PMS, the change in intensity (fold value) was calculated by the mean application of 100 l diluted mouse serum (1:1,000 to 1:100,000) onto Nunc fluorescence intensity (MFI) of the BCR in the contact area at each time point µ MaxiSorp 96-well flat-bottom plates (ThermoFisher) previously coated with 10 divided by the MFI of the first BCR cluster appeared on the field after setting µ g per well of either NP -BSA (to capture high affinity antibodies) or NP -BSA the threshold above background signal using Image Pro Plus software 4 30 (to capture diverse affinity antibodies) (Biosearch Technologies). Plates were (Media Cybernetics). For analysis of antigen recruitment into the contact washed, and bound IgG was detected using HRP-conjugated anti-IgG (identified area with time after the cells contacted the PLB or PMS, the change in intensity above; Jackson ImmunoResearch) developed with 3,3 ,5,5 -tetramethylbenzidine (fold value) was calculated by the MFI of antigen in the cell contact area at a given ′ ′ (TMB) substrate (ThermoFisher). The reaction was quenched using a stop reagent time point divided by the MFI of antigen before the B cell contacted the PLB- or purchased from Sigma-Aldrich. PMS-containing fluorescence-labeled antigen. The OD at 450 nm was measured on a Spectramax plate reader (Molecular For the acquisition of three color confocal z-stack images, a ZEISS 780 LSM Devices). ELISPOT analysis of NP-specific antibody–secreting cells were carried confocal microscope (ZEISS) with 63 1.45 NA oil objective lens was used × out according to the protocol explained elsewhere10, and plates were evaluated and with three laser lines, 488, 56, and 633 nm, at the optimal 240 nm interval to the imaged by Zellnet Consulting. z-direction as described39. In brief, MD4 mouse splenic B cells were labeled with DyLight649-conjugated goat Fab fragments specific for IgM and were put into a chamber containing a CM-DiI-labeled PMS bearing Alexa Fluor 488–labeled Quantitative RT-PCR (qPCR) and RNA-seq analys-s. For qPCR analyses, RNA antigen at 37 °C for 30 min in the presence or absence of CpG and fixed with 4% was isolated from mouse splenic B cells stimulated in vitro using RNeasy mini or PFA. z-stack three color channel images were acquired of the whole cells on the micro Kits (Qiagen) according to manufacturer’s guidelines. cDNA was generated PMSs to obtain the spatial information of internalized antigen, membrane and the using iScript Reverse Transcription Supermix (Bio-Rad). qPCR assays containing BCR. To quantify the percentage of antigen internalization or total fluorescence primers and Taqman probes were purchased from Integrated DNA Technologies, intensity (TFI) of internalized antigen, Matlab software (Mathworks) and Matlab qPCR assays containing primers only were purchased from Eurofins MWG code obtained from P. Tolar (The Francis Crick Institute, London, UK) were used Operon. Sequences of these primers and probes are in Supplementary Table 1. as described in detail36 after splitting three channel images into single channel iQ SYBR Green Supermix (Biorad) or Platinum Quantitative PCR SuperMix-UDG image using Image J. Internalized antigen was quantified after defining the surface (ThermoFisher) were used for amplification of target genes. The change in gene of PMSs according to the code. In brief, two-channel confocal z-stack images for expression (fold value; target vs control) was calculated as previously shown43. antigen and membrane were separated using Image J program and saved as tif files For RNA-seq analysis, 250 µl of purified mouse B cells cultured in full medium for the use of this Matlab code. Using the code, the PMS that had incorporated for 4 h were mixed with 750 µl of Trizol LS (Thermofisher) and RNA was isolated antigen and was outside the cell contact area was considered a background region according to the manufacturer’s recommendations. Total RNA was prepared of interest, and the MFI of the background area was obtained to exclude scattered for next-generation sequencing using a TruSeq Stranded mRNA-Seq library fluorescence from the surface antigen of PMSs. To distinguish the internalized preparation kit (Illumina). The quality and size of the final purified libraries were antigen from the surface antigen, surface plane of the PMS from z-stack series tested on BioAnalzyer DNA 1000 chips (Agilent Technologies.). Libraries were was determined. Regions of interest including whole cell area were drawn on quantified by qPCR using a Kapa Quantification Kit for Illumina sequencing the basis of DIC images and were used to determine the proportion of antigen (Kapa Biosystems) and then were normalized to 2.0 nM stocks. These samples internalization or total fluorescence intensity (TFI) of internalized antigen. For were pooled equally and were clustered on the HiSeq 2500 (Illumina) using the three-dimensional image reconstitution and three-dimensional Pearson’s 12 pM of template for a 2 × 100 bp rapid sequencing run. Raw next-generation

Nature Immunology | www.nature.com/natureimmunology © 2018 Nature America Inc., part of Springer Nature. All rights reserved. NaTuRe ImmunoLoGy Articles sequencing reads were first prepped by removal any illumina adaptor sequences 29. Cauerhf, A., Goldbaum, F. A. & Braden, B. C. Structural mechanism for using Cutadapt v1.1244 and then were trimmed and filtered for quality, q = 18, and afnity maturation of an anti-lysozyme antibody. Proc. Natl. Acad. Sci. USA length, min length 35 bp using FastX Tool Kit v0.0.13 (Hannon Lab, Cold Spring 101, 3539–3544 (2004). Harbor Laboratory). Trimmed reads for each replicates were then mapped to Mus 30. Traba, J., Miozzo, P., Akkaya, B., Pierce, S. K. & Akkaya, M. An optimized musculus mm10 genome using Tophat2 v2.1.145 set to report only matched pairs. protocol to analyze glycolysis and mitochondrial respiration in lymphocytes. Final transcripts based on the combined replicates, and then differentials for each J. Vis. Exp. 117, e54918 (2016). comparison were generated with the Cufflinks suite45 using pooled normalization 31. Akkaya, B. et al. Ex-vivo iTreg diferentiation revisited: Convenient for each experimental condition. Changes in the transcriptional regulation of alternatives to existing strategies. J. Immunol. Meth. 441, 67–71 (2017). various pathways were analyzed using Ingenuity Pathway Analysis software 32. Mullen, G. E. et al. Phase 1 trial of AMA1-C1/Alhydrogel plus CPG 7909: an (Qiagen). asexual blood-stage vaccine for Plasmodium falciparum malaria. PLoS ONE 3, e2940 (2008). Purification and surface plasmon resonance (BIAcore) binding analysis of 33. Tomlinson, D. C., L’Hôte, C. G., Kennedy, W., Pitt, E. & Knowles, M. A. antibodies to PfAMA1. Total serum IgG antibodies from PfAMA1-immunized Alternative splicing of fbroblast growth factor receptor 3 produces a secreted subjects were purified by protein G affinity chromatography. In brief, 0.5 ml serum isoform that inhibits fbroblast growth factor-induced proliferation and is from each individual subject was loaded onto a 5 ml HiTrap Protein G column (GE repressed in urothelial carcinoma cell lines. Cancer. Res. 65, healthcare) and was subjected to a standard elution with a glycine pH gradient. The 10441–10449 (2005). eluted IgG proteins were pooled together and dialyzed against 10 mM HEPES (pH 34. Akkaya, M., Aknin, M. L., Akkaya, B. & Barclay, A. N. Dissection of agonistic 7.5) and 0.15 M NaCl for binding studies. and blocking efects of CD200 receptor antibodies. PLoS ONE 8, Surface plasmon resonance measurements were performed using a BIAcore e63325 (2013). 3000 instrument and were analyzed with BIAevaluation 4.1 software (Biacore AB). 35. Sohn, H. W., Tolar, P., Brzostowski, J. & Pierce, S. K. A method for analyzing To measure the affinity, PfAMA1 protein was immobilized on caboxylated dextran protein-protein interactions in the plasma membrane of live B cells by CM5 chips (Biacore AB) to 500–1,500 response units using a primary amine fuorescence resonance energy transfer imaging as acquired by total internal coupling and a flow rate of 10 µl/min in 10 mM sodium acetate (pH 4.5). The refection fuorescence microscopy. Methods. Mol. Biol. 591, analyte consisted of serial dilutions of IgG antibodies between 16.0 nM and 75 nM 159–183 (2010). in a buffer containing 10 mM HEPES (pH 7.4) and 0.15 M NaCl. The dissociation 36. Natkanski, E. et al. B cells use mechanical energy to discriminate antigen constants were obtained by kinetic curve fitting using BIAevaluation 4.1 (BIAcore). afnities. Science 340, 1587–1590 (2013). 37. Liu, W., Meckel, T., Tolar, P., Sohn, H. W. & Pierce, S. K. Antigen afnity Statistical analysis. The statistical significance of differences between experimental discrimination is an intrinsic function of the B cell receptor. J. Exp. Med. 207, groups was measured using GraphPad Prism software. The methods used for each 1095–1111 (2010). experiment and the P value ranges were indicated in the relevant figure legends. 38. Davey, A. M. & Pierce, S. K. Intrinsic diferences in the initiation of B cell receptor signaling favor responses of human IgG+ memory B cells over IgM+ Life Sciences Reporting Summary. Further information on experimental design naive B cells. J. Immunol. 188, 3332–3341 (2012). and reagents is available in the Life Sciences Reporting Summary. 39. Wang, J., Sohn, H., Sun, G., Milner, J. D. & Pierce, S. K. Te autoinhibitory Data availability statement. The data that support the findings of the study C-terminal SH2 domain of phospholipase C-γ2 stabilizes B cell receptor are available from the corresponding author upon request. RNA-seq data were signalosome assembly. Sci. Signal. 7, ra89 (2014). uploaded to the SRA database under project number PRJNA422889. 40. Oferdahl, D. K., Dorward, D. W., Hansen, B. T. & Bloom, M. E. A three-dimensional comparison of tick-borne favivirus infection in mammalian and tick cell lines. PLoS ONE 7, e47912 (2012). References 41. Mastronarde, D. N. Automated electron microscope tomography using robust 25. Mason, D. Y., Jones, M. & Goodnow, C. C. Development and follicular prediction of specimen movements. J. Struct. Biol. 152, 36–51 (2005). localization of tolerant B lymphocytes in lysozyme/anti-lysozyme IgM/IgD 42. Kremer, J. R., Mastronarde, D. N. & McIntosh, J. R. Computer visualization transgenic mice. Int. Immunol. 4, 163–175 (1992). of three-dimensional image data using IMOD. J. Struct. Biol. 116, 71–76 26. Ho, W. Y., Cooke, M. P., Goodnow, C. C. & Davis, M. M. Resting and anergic (1996). B cells are defective in CD28-dependent costimulation of naive CD4+ T cells. 43. Gordon, E. B. et al. Targeting glutamine metabolism rescues mice from J. Exp. Med. 179, 1539–1549 (1994). late-stage cerebral malaria. Proc. Natl Acad. Sci. USA 112, 13075–13080 27. Akkaya, M., Kwong, L. S., Akkaya, E., Hatherley, D. & Barclay, A. N. Rabbit (2015). CD200R binds host CD200 but not CD200-like proteins from poxviruses. 44. Martin, M. Cutadapt removes adapter sequences from high-throughput Virology 488, 1–8 (2016). sequencing reads. EMBnet J. 17, 10–12 (2011). 28. Dadaglio, G., Nelson, C. A., Deck, M. B., Petzold, S. J. & Unanue, E. R. 45. Trapnell, C. et al. Diferential gene and transcript expression analysis of Characterization and quantitation of peptide-MHC complexes produced from RNA-seq experiments with TopHat and Cufinks. Nat. Protoc. 7, hen egg lysozyme using a monoclonal antibody. Immunity 6, 727–738 (1997). 562–578 (2012).

Corresponding author(s): Susan K. Pierce Initial submission Revised version Final submission Life Sciences Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form is intended for publication with all accepted life science papers and provides structure for consistency and transparency in reporting. Every life science submission will use this form; some list items might not apply to an individual manuscript, but all fields must be completed for clarity. For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

` Experimental design 1. Sample size Describe how sample size was determined. Based on earlier studies 2. Data exclusions Describe any data exclusions. No data was excluded 3. Replication Describe whether the experimental findings were Experiments were reliably reproduced reliably reproduced. 4. Randomization Describe how samples/organisms/participants were For chimeric mouse experiments mice were randomly allocated to groups allocated into experimental groups. 5. Blinding Describe whether the investigators were blinded to Our animal technician, who did the injections did not know the details of the group allocation during data collection and/or analysis. experiment. Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

6. Statistical parameters For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or in the Methods section if additional space is needed). n/a Confirmed

The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.) A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly A statement indicating how many times each experiment was replicated The statistical test(s) used and whether they are one- or two-sided (note: only common tests should be described solely by name; more complex techniques should be described in the Methods section) A description of any assumptions or corrections, such as an adjustment for multiple comparisons The test results (e.g. P values) given as exact values whenever possible and with confidence intervals noted A clear description of statistics including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range) Clearly defined error bars June 2017 See the web collection on statistics for biologists for further resources and guidance.

` Software Policy information about availability of computer code 7. Software Describe the software used to analyze the data in this Statistical significance was analyzed in GraphPad Prism Versions 6 and 7,

1 study. Flow cytometry data was analyzed in Flowjo ver.10. Imaging data were analyzed in

Imaris and ImageJ softwares nature research | life sciences reporting summary

For manuscripts utilizing custom algorithms or software that are central to the paper but not yet described in the published literature, software must be made available to editors and reviewers upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). Nature Methods guidance for providing algorithms and software for publication provides further information on this topic.

` Materials and reagents Policy information about availability of materials 8. Materials availability Indicate whether there are restrictions on availability of There are no restrictions unique materials or if these materials are only available for distribution by a for-profit company. 9. Antibodies Describe the antibodies used and how they were validated Antibody specificity, commercial source, clone name and fluorescent dye for use in the system under study (i.e. assay and species). conjugates of antibodies were clearly identified in the methods section. Antibodies were selected from clones that are regularly used for similar studies.

10. Eukaryotic cell lines a. State the source of each eukaryotic cell line used. NIH3T3 (ATCC-obtained from a colleague), Phoenix Eco (ATCC)

b. Describe the method of cell line authentication used. No authentication was made

c. Report whether the cell lines were tested for Yes. They were free of mycoplasma mycoplasma contamination.

d. If any of the cell lines used are listed in the database No misidentified cell lines were used. of commonly misidentified cell lines maintained by ICLAC, provide a scientific rationale for their use.

` Animals and human research participants Policy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines 11. Description of research animals Provide details on animals and/or animal-derived Female mice 8-12 weeks were used for experiments except for bone marrow materials used in the study. chimeric mice. C57BL/6 (JAX0664), MyD88 KO (JAX 09088), B10.BR (JAX 0465), MuMT (JAX002288), CD45.1+ WT mice ( JAX 2014) were purchased from Jackson Laboratories. TLR9 KO mouse in BL/6 background was obtained from Dr. Shizuo Akira, 3A9 mouse was obtained from Dr. Igal Gery, MD4 mice were purchased from Taconic farms. Related breedings were carried out at NIAID animal facilities and genotypings were done according to previously published protocols.

Policy information about studies involving human research participants 12. Description of human research participants Describe the covariate-relevant population Human participants were as detailed in the original manuscript as well as online in characteristics of the human research participants. the following location: www.clinicaltrials.gov no. NCT00320658 June 2017

2 June 2017 nature research | flow cytometry reporting summary 1 s of Final submission Revised version Initial submission Corresponding author(s):Corresponding Susan K. Pierce experiment) FSC-H/ FSC-A was used to gate the singlets. FSC-H/ Live Dead experiment) FSC-H/ FSC-A was used to on the experimental set up. was used to gate the alive cells. Rest depends 7 and also additional information Gating strategy is exemplified in Figure results and figure legends sections. was provided in the relevant methods, used either as is or after B cell purification and culturing. Details in the used either as is or after B cell purification methods section. Flow JO version 10 B cell puritiy was checked by For experiments involving purified B cells, B220. Then gating for CD19+ staining the cells with live/dead CD19 and B220+ cells among the singlet live population.

identical markers).

populations within post-sort fractions. the flow cytometry data.

Methodological details Data presentation

4. A numerical value for number of cells or percentage (with statistics) is provided. for number of cells or percentage (with 4. A numerical value 3. All plots are contour plots with outliers or pseudocolor plots. plots with outliers or pseudocolor 3. All plots are contour 2. The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysi axes only for bottom left plot of group clearly visible. Include numbers along 2. The axis scales are 1. The axis labels state the marker and fluorochrome used (e.g. CD4-FITC). the marker and fluorochrome used 1. The axis labels state Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information. Tick this box to confirm that a figure exemplifying the gating strategy is provided

9. Describe the gating strategy used.the splenocytes/B cells (depending on FSC-A SSC-A was used to gate 8. Describe the abundance of the relevant cell 8. Describe the abundance of the relevant 7. Describe the software used to collect and analyze 7. Describe the software used to collect 6. Identify the instrument used for data collection.6. Identify the instrument used for data BD LSR-II 5. Describe the sample preparation. and RBCs were lysed. Splenocytes were Spleens were harvested, mashed For all flow cytometry data, confirm that: For all flow cytometry `

Form fields will expand as needed. Please do not leave fields blank. Please do not will expand as needed. Form fields Flow Cytometry Reporting Summary Reporting Cytometry Flow