Tripartite phase separation of two signal effectors with vesicles priming B cell responsiveness

Abstract

Antibody-mediated immune responses rely on antigen recognition by the B cell antigen receptor (BCR) and the proper engagement of its intracellular signal effector proteins. Src homology (SH) 2 domain-containing leukocyte protein of 65 kDa (SLP65) is the key scaffold protein mediating BCR signaling. In resting B cells, SLP65 colocalizes with Cbl-interacting protein of 85 kDa (CIN85) in cytoplasmic granules whose formation is not fully understood. Here we show that effective B cell activation requires tripartite phase separation of SLP65, CIN85, and lipid vesicles into droplets via vesicle binding of SLP65 and promiscuous interactions between nine SH3 domains of the trimeric CIN85 and the proline-rich motifs (PRMs) of SLP65. Vesicles are clustered and the dynamical structure of SLP65 persists in the droplet phase in vitro. Our results demonstrate that phase separation driven by concerted transient interactions between scaffold proteins and vesicles is a cellular mechanism to concentrate and organize signal transducers.

Fig. 1. Tripartite phase separation of SLP65 with CIN85 and SUVs.

a Confocal fluorescence microscopy of SLP65-deficient DT40 B cells that had been reconstituted with a citrine fusion protein of either wild-type SLP65 (left), a variant lacking N-terminal 45 amino acids (ΔN, middle), or a CIN85 binding-deficient variant (ΔPRM, right). b Domain architecture of the different constructs of SLP65 and CIN85 studied here (PRR: proline-rich region, CC: coiled coil). c–e Confocal fluorescence microscopy of a mixture of different concentrations of Atto 430LS-tagged SLP65_1-330 with the corresponding equal concentrations of either (c) CIN85_1-333, (d) CIN85_Δ57, or (e) CIN85_Δ57 and SUVs made from 1 mM of phospholipids (see Method). The SUVs had a lipid composition of DOPC:DOPE:DOPS = 65:25:10 mol% and a mean hydrodynamic radius of 23 nm. Concentrations of CIN85_Δ57 refer to the monomeric concentration. Index of dispersion of the fluorescence intensity were shown as box-and-whisker plots (12 ROIs and two independent samples, n = 24). The line inside the box and the edges of the box correspond to the median, first and third quartiles, respectively. The outliers with value ≥1.5 times the interquartile range away from the top or bottom of the box are denoted by circles. Source data are provided as a Source Data file. f Different concentrations of Atto 430LS-tagged SLP65_40-330 were mixed with the corresponding equal concentrations of CIN85_Δ57 together with 1 mM of SUVs. g Mixture of 5 μM of Atto 430LS-tagged SLP65_1-330 and 1 mM of SUVs without CIN85 (left) or with 5 μM of CIN85_Δ57 (right). Scale bars equal to 10 μm (a, c–g).

Fig. 2. Phase separation modulated by vesicle curvature and lipid composition.

a A mixture of recombinant SLP65 with either SUVs (Rh ≈ 20 nm) or LUVs (Rh ≈ 60 nm) was fractionated by a sucrose gradient flotation assay, and the different layers were subjected to anti-SLP65 immunoblot analysis (apparent MW indicated in kDa). Vesicles will float to the top and are found in layers 1 and 2. Bar graph shows the percentage of SLP65 associated with vesicles based on the immunoblot signal intensity. b Confocal fluorescence microscopy of a mixture of different concentrations of Atto 430LS-tagged SLP65_1-330 and CIN85_Δ57 together with SUVs containing DOPC:DOPE of 75:25 mol% (b, top) and DOPC:DOPE:DOPS:cholesterol of 35:25:10:30 mol% (b, bottom), respectively. (Scale bar = 10 μm) Source data are provided as a Source Data file. c SLP65-deficient DT40 B cells were transduced with constructs encoding citrine-tagged SLP65_ΔN fusion proteins with the N-BAR domain of Amphiphysin. Wild-type SLP65 and SLP65_ΔN-expressing cells served as positive and negative controls, respectively. The domain structures of the expressed SLP65 variants are depicted in the upper part of (c). Cells were loaded with the Ca²⁺-sensitive fluorophore Indo-1 and Ca²⁺ flux was monitored by flow cytometry. Cells were stimulated with anti-BCR antibody (M4) at the indicated time point. d Confocal fluorescence microscopy of cells transduced with N-BAR-SLP65_ΔN were either left untreated (upper panel) or stimulated via their BCR (lower panel).

Fig. 3. Cryo-electron tomogram of the tripartite droplets.

a A mixture of 1 μM of SLP65_1-330 and 1 μM of CIN85_Δ57 together with 1 mM of SUVs was placed on a carbon grid and plunged frozen after incubation. The droplet shows a clear increase in concentration of vesicles toward its center with a steep fall-off at the edges. Golden spheres represent vesicles that were manually segmented for 3D-visualization in AMIRA 6.2 and processed using an in-house python script and Matlab. b Histogram of distances between vesicles with an average vesicle to vesicle distance of 4.92 ± 0.1 nm (mean ± SEM; n = 2349). c Histogram of the vesicle’s radii with a mean radius of 9.10 ± 0.03 nm (mean ± SEM; n = 1718). Source data are provided as a Source Data file.

Fig. 4. Structural properties of phase-separated SLP65 and its interactions with CIN85.

a ¹⁵N-HSQC spectra of 60 µM or 15 µM of ¹⁵N-labeled SLP65 only (black) were overlaid by the spectra of SLP65 after addition of unlabeled CIN85_1-333 (pink), CIN85_Δ57 (orange), SUVs (green), or SUVs and CIN85_Δ57 (cyan) at the indicated concentrations. b Bright field images of the respective NMR samples denoted by colored boxes. (Scale bar = 10 µm) c Residue-specific peak intensity ratio of the ¹⁵N-HSQC spectra shown in a of SLP65 with CIN85_1-333 (pink), CIN85_Δ57 (orange), SUVs (green), or SUVs and CIN85_Δ57 (cyan) over SLP65 only at the corresponding concentrations. Error bar indicates uncertainty propagated from the peak’s signal-to-noise. Proline residues were marked in red on top of the chart, while PRMs that interact with CIN85 were highlighted with their sequences written. The five tyrosine residues that are phosphorylated by Syk (Y72, Y84, Y96, Y178, and Y189) were pointed out by inverted triangle. Source data are provided as a Source Data file. d Zoomed-in regions of ¹⁵N-HSQC spectra of 60 µM SLP65 titrated with increasing amounts of CIN85_1-333 were overlaid to show either shifting or disappearance of the respective peaks associated with different PRMs.