Fig 2: Acidification destabilizes binding of CD80 to CD28, but not to CTLA4.a, b Superimposition of acidification-induced changes in electrostatic potential (ΔE) on monomeric human CD28 or CTLA4 or respective dimers interacting with CD80. Ligand-interaction and homodimer regions of CD28 and CTLA4 are marked in pink and cyan, respectively. Histidines are marked in red. c, d Molecular dynamic simulations of monomeric CD28 (C) or CTLA4 (D) interacting with CD80 receptor-binding domain. Ligand interaction sites of CD28 and CTLA4 are highlighted in pink. Note that acidification induces a reorientation of the ligand-interaction site of CD28, but not of CTLA4. e Cartesian coordinates of normal vector tips of CD28 and CTLA4 interacting with CD80. f Umbrella sampling simulations of monomeric receptor-ligand interactions. Shaded areas depict error estimation. g Molecular dynamic simulations of CD28 homodimers at pH 7 vs. 5. Ligand-interaction and homodimer regions are marked in pink and cyan, respectively. Histidines are marked in red. h Inhibition of acidification stabilizes endosomal conformation of CD28. CTLA4-TagGFP or CD28/CTLA4(IC)-TagGFP expressing 58αβ A2 T cells were cultured with or without Bafilomycin. CD80-Fc staining was performed on fixed, permeabilized cells. Plot depicts increase of αIgG-AF647-signal normalized to TagGFP gMFI upon Bafilomycin treatment. i, j Analysis of CD28-CD80- (i) and CTLA4-CD80- (j) stability upon acidification. 58αβ A2 T cells expressing CD28 or CTLA4 were incubated on ice with titrated amounts of CD80-Fc and subsequently treated with PBS (pH 7.2) or citric buffer (pH 5.5). After washing bound CD80-Fc was detected using αIgG-AF647 antibody by flow cytometry. See also Figures S5–8. Statistics: 6 h: Pooled data from 4 independent experiments. Two-sided paired t-test (n = 4 biologically independent samples per group). i: Representative data from 1 of 2 independent experiments. Extra sum-of-squares F test. p-value: 3 × 10−15 (n = 3 biologically independent samples per group). j: Representative data from 1 of 2 independent experiments. Extra sum-of-squares F test (n = 3 biologically independent samples per group). Source data are provided as a Source data file.

Fig 3: Extra- and intracellular domains of CD28 and CTLA4 modulate fate of trogocytosed CD80.a Composition of CD28/CTLA4(IC) or CTLA4/CD28(IC) hybrid receptors compared to native CD28 and CTLA4. b–g 58αβ A2 T cells expressing native or hybrid receptors co-cultured with CHO cells expressing CD80-TagRFP-pHluorin were analyzed by flow-cytometry (b–e) or confocal microscopy (f, g). Proportion of T cells acquiring CD80 (b) and amount of trogocytosed CD80 detectable per acquiring cell (c). Note that while only few T cells expressing the CD28/CTLA4(IC) hybrid receptor trogocytose CD80 (shown in b), the amount per acquiring cell is comparable to T cells expressing native CTLA4 (shown in c). d Increase in TagRFP signal intensity (gMFI) upon addition of Bafilomycin during trogocytosis of CD80-TagRFP. e pHluorin/TagRFP ratio of CD80-TagRFP-pHluorin trogocytosed via different receptors. Note that while receptors containing the CTLA4 intracellular domain target trogocytosed CD80 to more acidic compartments (shown in e), blockade of lysosomal function only affects CD80 trogocytosed by receptors containing the CTLA4 extracellular domain (shown in d). Quantification of co-localization of trogocytosed CD80 with wheat germ agglutinin lectin (WGA-L) (f) or lysobisphosphatidic acid (LBPA) (g). Statistics: b, c: Pooled data from 6 independent experiments. Repeated measures one-way ANOVA with Tukey’s correction (n = 6 biologically independent samples per group). d: Pooled data from 4 independent experiments. Two-way repeated-measure ANOVA with Šídák’s, treated vs untreated (n = 4 biologically independent samples per group). e: Pooled data from 7 independent experiments. Repeated measures one-way ANOVA with Tukey’s correction (n = 7 biologically independent samples per group). f: Kruskal-Wallis test with Dunn’s correction: p-values: CD28 vs. CTLA4: 7.4 × 10−8; CD28 vs. CD28/CTLA4(IC): 3 × 10−9; CD28 vs. CTLA4/CD28(IC): 0.99. (CD28: n = 28 cells, CTLA4: n = 25 cells, CD28/CTLA4(IC) and CTLA4/CD28(IC): n = 35 cells each, all examined over 2 independent experiments). g: Pooled data from two independent experiments. Brown-Forsythe/Welch’s ANOVA with Dunnett’s T3 correction. p-values: CD28 vs. CTLA4: 1 × 10−9; CD28 vs. CD28/CTLA4(IC): 0.0048; CD28 vs. CTLA4/CD28(IC): 0.11; CTLA4 vs. CD28/CTLA4(IC): 2.22 × 10−5 (CD28 and CTLA4: n = 16 cells each, CD28/CTLA4(IC) and CTLA4/CD28(IC): n = 29 cells each, all examined over 2 independent experiments). Source data are provided as a Source data file.

Fig 4: Dissecting mechanisms underlying trogocytosis.a Cartoons depicting putative mechanisms of trogocytosis via intraluminal vesicle (ILV) generation or membrane fusion between donor and recipient cells. b CD80 experiences different pH-levels upon trogocytosis via CD28 versus CTLA4. CD28- or CTLA4-expressing 58αβ A2 T cells were co-cultured with CD80-TagRFP-pHluorin expressing CHO cells. Quantification of pHluorin/TagRFP ratio and extrapolation of pH-levels experienced by trogocytosed CD80 based on standard curve shown in Figure S3. c, d Analysis of membrane fusion by bi-molecular fluorescence complementation (BIFC) assay. c Cartoon depicting putative outcomes of BIFC trogocytosis assay. CD8+ T cells (d) or CD28- or CTLA4-expressing 58αβ A2 T cells (e) expressing cytosolic GFP1–10 were co-cultured with CD80-mScarlet-GFP11-expressing CHO cells. Positive control: GFP1–10 transgenic T cells transduced with CD80-mScarlet-GFP11 (d). Negative control: GFP1–10 transgenic T cells cultured with CD80-mScarlet-transgenic CHO cells. Plot shows increase of GFP gMFI over negative control after 24 h (e). f Fluorescent tag attached to CD80 cytoplasmic domain is cleavable by extracellular protease treatment. CD28-expressing 58αβ A2 T cells having trogocytosed CD80-IgG-mScarlet were treated with or without trypsin. Negative control: T cells cultured without CHO cells. Representative flow cytometry plots depicting loss of trogocytosed mScarlet and cell surface CD80 signal upon trypsinization-treatment. Note that loss of both signals marks the extracellular location of the cytoplasmic domain of trogocytosed CD80. See also Figures S3 and S4. Statistics: b: pHluorin/TagRFP: pooled data from 13 independent experiments. Two-sided Wilcoxon matched-pairs signed rank test (n = 13 biologically independent samples per group) pH: two-sided paired t test, pooled data from 3 independent experiments (n = 3 biologically independent samples per group). 4e: Pooled data from 3 independent experiments. Two-way repeated-measure ANOVA with Šídák’s correction, sample vs control (n = 6 biologically independent samples per group). Source data are provided as a Source data file.

Fig 5: CD28 and CTLA4 trogocytose CD80, but direct it towards different fates.a–d Trogocytosis of CD80 (a, c) or CD86 (b, d) by naive WT, Cd28−/− or Ctla4−/− CD8+ T cells (a, b) or WT CD8+ T cells upon CD28- or CTLA4-blockade (c, d). e Quantification of CD80 trogocytosed by GFP-, Cd28- or Ctla4-transduced Cd80−/−Cd86−/− CD8+ T cells. f, g Expression of CTLA4 (f) and its correlation with trogocytosed CD80 on T cell surface (g) of titrated numbers of Cd80−/−Cd86−/− P14 CD8+ T cells activated in flat-, u-, or v-bottom plates for 28 h. To correct for different trogocytosis rates, surface CD80 levels were normalized to total acquired CD80-TagRFP (αCD80-PE-Cy7/TagRFP). h Quantification of trogocytosed CD80 on surface of GFP-, Cd28- or Ctla4-transduced Cd80−/−Cd86−/− CD8+ T cells. Note that only CD28-, but not CTLA4-overexpression in T cells increases CD80-surface expression. i Images and quantification of co-localization of wheat germ agglutinin lectin (WGA-L) surface staining with CD80 trogocytosed by CD28- or CTLA4-expressing 58αβ T cells. Scale bar 2 µm. j Trogocytosis of CD80 via CD28 but not via CTLA4 enables co-stimulation of neighboring cells. Cd28- or Ctla4-transduced Cd80−/−Cd86−/− B cells were co-cultured with CD80-mScarlet expressing MEFs, FACS-sorted for mScarlet-expression, gp33-antigen-loaded and co-cultured with Cd80−/−Cd86−/− P14 CD8+ T cells with or without addition of CTLA4-Fc (“specificity control”) or agonistic αCD28 antibodies (“positive control”). Plots depict day 3 T cell counts (left) and CD25 expression (right). See also Figure S2. Statistics: a: Pooled data from 3 independent experiments. Brown-Forsythe/Welch’s ANOVA with Dunnett’s T3 correction. WT and Cd28−/− (n = 6 biologically independent samples), Ctla4−/− (n = 4 biologically independent samples). b: Pooled data from 3 independent experiments. Brown-Forsythe/Welch’s ANOVA with Dunnett’s T3 correction. p-values: WT vs. Cd28−/−: 0.00001; WT vs. Ctla4−/−: 0.27 WT (n = 9 biologically independent samples), CD28−/− (n = 4 biologically independent samples), Ctla4−/− (n = 7 biologically independent samples). c: Pooled data from 3 independent experiments. Brown-Forsythe/Welch’s ANOVA with Dunnett’s T3 correction. p-values: medium vs. α-CD28-Fab: 1 × 10−15; medium vs. α-CTLA4-Fab: 0.84 (medium and α-CD28-Fab: n = 10 biologically independent samples, α-CTLA4-Fab: n = 8 biologically independent samples). d: Pooled data from 3 independent experiments. Brown-Forsythe/Welch’s ANOVA with Dunnett’s T3 correction. p-values: medium vs. α-CD28-Fab: 4 × 10−8; medium vs. α-CTLA4-Fab: 0.61. (medium and α-CD28-Fab: n = 11 biologically independent samples, α-CTLA4-Fab: n = 8 biologically independent samples). e: Pooled data from 3 independent experiments. Repeated-measures one-way ANOVA with Dunnett’s T3 correction. p-values: GFP vs. Cd28: 4 × 10−7; GFP vs. Ctla4: 0.000017 (n = 10 biologically independent samples per group). f: Representative data from 1 of 2 independent experiments. (n = 3 biologically independent samples per group). g: Representative data from 1 of 2 independent experiments (n = 3 biologically independent samples per group). h: Pooled data from 3 independent experiments. Friedmann test with Dunn’s correction. p-values: GFP vs. Cd28: 0.015; GFP vs. Ctla4: p value: 0.36 (n = 10 biologically independent samples per group). i: Two-sided Mann–Whitney test; p-value: 2 × 10−10 (CD28: n = 28 cells and CTLA4: n = 25 cells, examined over 2 independent experiments). j: Representative data from 1 of 3 independent experiments. Repeated-measures one-way ANOVA with Holm-Šídák’s correction (n = 3 biologically independent samples per group). Source data are provided as a Source data file.