Arp2/3, activated by N-WASP but not WASP, is required for B-cell contraction.

Splenic B-cells were incubated with planar lipid bilayers coated with monobiotinylated Fab’ fragment of goat anti-mouse IgG+M (Fab’-PLB) in the absence and presence of various inhibitors and imaged live at 37°C by interference reflection microscopy (IRM). The B-cell plasma membrane area contacting Fab’-PLB (B-cell contact zone) was measured using IRM images and custom MATLAB scripts. (A) Representative IRM images of splenic B-cells from C57BL/6 mice treated with CK-689 or CK-666 (50 µM) before (0 min) and after maximal spreading (2 min). (B) Representative plots of the B-cell contact area versus time from one contracting cell and one non-contracting cell. (C) Percentages (±SEM) of B-cells that underwent contraction after treatment with CK-666 or CK-689. (D) Representative IRM images of splenic B-cells from C57BL/6 mice treated with DMSO or Wiskostatin (Wisko, 10 µM) 10 min before and during incubation with Fab’-PLB. (E) Percentages (±SEM) of B-cells that underwent contraction after treatment with Wisko or DMSO. (F) Representative IRM images of splenic B-cells from flox control and B-cell-specific N-WASP knockout (cNKO) mice. (G) Percentages (±SEM) of cNKO or flox control B-cells that underwent contraction. (H) Representative IRM images of splenic B-cells from WT or WASP knockout mice (WKO). (I) Percentages (±SEM) of WKO or WT B-cells that underwent contraction. Data were generated from three independent experiments, with ∼25 cells per condition, per experiment. Scale bar, 2 µm. *p <0.05, ***p<0.001, by paired student’s t-test.

Figure supplement 1. B-cells spread and contract on Fab’-coated-planar lipid bilayers.

Figure supplement 2. CK-666 significantly decreases Arp2/3 recruitment to the B-cell contact zone.

Figure supplement 3. CK-666 treatment before but not after maximal B-cell spreading decreases the spreading kinetics.

Figure supplement 4. Wiskostatin treatment inhibits N-WASP activation while enhancing WASP activation in B-cells.

Figure 1–Video 1. Effects of CK-666, Wiskostatin, conditional N-WASP knockout, and WASP knockout on B-cell contraction.

Arp2/3, downstream of N-WASP, generates inner F-actin foci, driving B-cell contraction.

(A-C) WT splenic B-cells were treated with CK-689 or CK-666 (50 µM) during incubation with Fab’-PLB (A and B), and flox control and cNKO B-cells were incubated with Fab’-PLB at 37°C (C). Cells were fixed at 2 and 4 min, permeabilized, stained for F-actin with phalloidin, and analyzed using TIRF. Shown are representative TIRF images of phalloidin staining in the contact zone of CK-689 and CK-666-treated B-cells (A, left panels) and fluorescence intensity (FI) profiles of phalloidin staining along a line crossing cells (A, right panels). Green arrows indicate lamellipodial F-actin, and purple arrows indicate inner F-actin foci forming a ring-like structure. Percentages of cells (per image) (±SEM) with inner F-actin foci forming ring-like distribution among CK-689-versus CK-666-treated cells (B) and flox control versus cNKO B-cells (C) before (2 min) and after (4 min) contraction were determined by visual inspection of phalloidin FI line-profiles across the B-cell contact zones. Data were generated from 3 independent experiments with 5 images per condition per experiment and ∼15 cells per image. (D-G) Inner F-actin foci were identified by their diameter (≥250 nm), peak FI (≥2 fold of no foci area, and location (1 µm away from the outer edge) and quantified as the number per cell using TIRF images. (D) Shown are representative images of splenic B-cells from LifeAct-GFP-expressing mice treated with CK-689 or CK-666 from 0 min during incubation with Fab’-PLB at 37°C (top) and the average number (±SEM) of inner LifeAct-GFP foci per cell (bottom) at 2 min. (E) LifeAct-GFP B-cells were treated with CK-666 at 2 min. Shown are representative TIRF images of LifeAct-GFP in the contact zone of B-cells (top) and the average number (±SEM) of inner LifeAct-GFP foci (bottom) in the same cell 10s before and 20s after CK-666 treatment. (F) Shown are representative TIRF images of phalloidin-staining in the contact zone of flox control and cNKO B-cells after incubating with Fab’-PLB for 2 min (top) and the average number (±SEM) of inner F-actin foci per cell (bottom). (H) Shown are representative TIRF images of WT and WKO B-cells expressing LifeAct-GFP incubated with Fab’-PLB for 2 min and the average number (±SEM) of inner F-actin foci per cell (bottom). Data points represent individual cells from three independent experiments with 10 (D, F, and G) or 6 (E) cells per condition per experiment. Scale bar, 2 µm. ** p <0.01, *** p<0.001, by non-parametric student’s t-test.

Figure 2-Video 1. Effects of CK-666 and WKO on F-actin foci formation.

Inner F-actin foci are originated from lamellipodia behind the spreading membrane.

Mouse splenic B-cells from LifeAct-GFP transgenic mice were treated with DMSO or Wisko (10 µM), imaged live using TIRF during incubation with Fab’-PLB at 37°C, and analyzed using kymographs generated by NIH ImageJ. (A) One frame of TIRF time-lapse images of LifeAct-GFP in the contact zone of B-cells treated with DMSO or Wisko. Lines indicate eight kymographs that were randomly generated from each cell. (B) Representative kymographs were generated from TIRF time-lapse images of LifeAct-GFP at the red line in (A). Top panel, a contracting cell. Arrows indicate the start of contraction with inner F-actin foci originating from lamellipodia. Bottom panel, a non-contracting cell. Lamellipodia-derived inner F-actin foci were identified by their LifeAct-GFP FI ≥2 fold of their nearby region, migrating out of the lamellipodial F-actin toward the center of the contact zone, and trackable for >8 sec. (C) Percentages (±SEM) of kymographs showing inner F-actin foci originating from lamellipodia per cell that did and did not undergo contraction. Data were generated from 3 independent experiments with ∼10 cells per condition per experiment. (D) A histogram of inner F-actin foci emerging (expressed as percentages of the total events, blue line) over time relative to the time of B-cell contraction (defined as 0 sec, indicated by a purple dash line and arrow). Data were generated from 5 independent experiments with ∼9 cells per condition per experiment. (E) Percentage (±SEM) of inner F-actin foci originated from lamellipodia observed in 8 randomly positioned kymographs of each DMSO- or Wisko-treated WT or untreated WKO B-cell. Data were generated from 3 independent experiments with ∼10 cells per condition per experiment. Scale bars, 2 µm. *** p<0.001, by non-parametric student’s t-test.

Figure supplement 1. Emerging of Inner F-actin foci from lamellipodia.

N-WASP-activated Arp2/3 sustains the lifetime and the centripetal movement of inner F-actin foci.

LifeAct-GFP-expressing B-cells were incubated with or without various inhibitors and imaged live by TIRF during interaction with Fab’-PLB at 37°C. Three randomly positioned kymographs were generated for each cell using time-lapse images. (A) A representative kymograph from TIRF time-lapse images of a DMSO-treated B-cell (Left panels). Inner F-actin foci were identified as described in Figure 3, and those that emerged during the 60 sec window after maximal spreading (white rectangles in B) and can be tracked for ≥ 4 sec (dashed lines) in individual kymographs were analyzed. Relative lifetimes of inner F-actin foci were measured using the duration each focus could be detected in a kymograph. The relative distances traveled by the foci were measured using the displacement of each focus in a kymograph. Relative speed was calculated for each inner F-actin focus by dividing its relative distance by its relative lifetime (right panels). (B-D) B-cells were treated with CK-689 or CK-666 (50 µM) from the beginning of the incubation with Fab’-PLB (0 min). Shown are representative kymographs (B), relative lifetimes (C), and relative speed (D) of inner F-actin foci in CK-689-versus CK-666-treated B-cells. (E-G) B-cells were treated with CK-666 at maximal spreading (2 min). Shown is a representative kymograph of a CK-666-treated cell (E), relative lifetimes (F), and relative speeds (G) of inner F-actin foci in 30-sec windows before the inhibition and 10 sec after the inhibition in the same cells (linked by blue lines). (H-J) B-cells were treated with DMSO or Wisko (10 µM) 10 min before and during interaction with Fab’-PLB. Shown is a representative kymograph of a Wisko-treated B-cell (H), relative lifetime (I), and relative speed (J) of inner F-actin foci in DMSO versus Wisko-treated B-cells. (K-M) LifeAct-GFP-expressing WT and WKO B-cells were incubated with Fab’-PLB. Shown are a representative kymograph of a WKO B-cell (K), relative lifetime (L), and relative speed (M) of WT versus WKO B-cells. Data points represent the averaged values from inner F-actin foci in individual cells, with 3 kymographs per cell and ∼12 cells per condition per experiment from 3 independent experiments. Scale bar, 2 µm. * p <0.05, ** p <0.01, *** p<0.001, by non-parametric and paired student’s t-test.

Inner F-actin foci, generated by N-WASP-activated Arp2/3, facilitate NMII recruitment and ring-like structure formation.

(A-C) B-cells from mice expressing the GFP fusion of non-muscle myosin IIA (GFP-NMIIA) transgene were treated with DMSO or Wisko (10 µM) 10 min before and during incubation with Fab’-PLB. The B-cell contact zones were imaged live using TIRF. Shown are representative TIRF images of DMSO- and Wisko-treated B-cells at 30 sec (during spreading) and 2 min 30 sec (after maximal spreading) post landing (A, Scale bars, 2 µm), the averaged GFP-NMIIA MFI (±SEM) (B), and the initial rates of increasing (±SEM) of GFP-NMIIA in the contact zone (the slope of the initial GFP-NMIIA MFI versus time curves of individual cells) (C). Data points represent individual cells from 3 independent experiments with ∼6 cells per condition per experiment. * p <0.05, ***p<0.001, by Kolmogorov-Smirnov test (B) or non-parametric student’s t-test (C). (D) Primary B-cells from mice expressing both GFP-NMIIA and LifeAct-RFP transgenes were incubated with Fab’-PLB at 37°C and imaged live by TIRF. Shown are a representative TIRF image of a cell and a kymograph generated from time-lapse TIRF images at the yellow line. The purple arrow indicates the starting point of contraction, the white arrow GFP-NMIIA recruitment proximal to the spreading membrane, and the yellow arrow an F-actin (LifeAct-RFP) focus originating at the lamellipodia and moving away from the spreading membrane. (E, F) Primary B-cells from flox control, WKO, and cNKO mice were incubated with Fab’-PLB for indicated times. Cells were fixed, permeabilized, stained for NMII light chain, and imaged by IRM and TIRF. Shown are representative IRM and TIRF images (E) and percentages (±SD) of B-cells with the NMII ring-like structure in individual images (F), identified by visual inspection. The data were generated from 3 independent experiments with 5 images per condition per experiment. Scale bars, 2 µm. * p <0.05, ** p <0.01, *** p<0.001, by non-parametric student’s t-test.

Figure 5-Video 1. Wiskostatin treatment inhibits NMII ring-like structure formation.

B-cell contraction increases the molecular density within BCR clusters.

Primary B-cells from WT mice were treated with CK-689 or CK-666 from the beginning of the incubation with AF546-Fab’-PLB (0 min) or at the maximal B-cell spreading (2 min) (A, D, E, H, I, N, O). WT B-cells were treated with DMSO or Wisko (10 µM) 10 min before and during the incubation with AF546-Fab’-PLB (B, F, J, P). B-cells from flox control and cNKO mice were incubated with AF546-Fab’-PLB (C, G, K, Q). The B-cell contact zones were imaged live by TIRF. (A-C) Representative time-lapse images at 30 sec (during B-cell spreading) and 2 min 30 sec (after maximal spreading) after cell landed. Scale bars, 2 µm. (D-G) The MFI of AF546-Fab’ in the contact zone was plotted over time. Purple rectangles indicate the contraction phase. (H-K) Rates of AF546-Fab’ MFI increases in the B-cell contact zone before and during contraction were determined by the slope of AF546-Fab’ MFI versus time plots. Data points represent individual cells and were generated from 3 independent experiments with 6-12 cells per experiment. * p <0.05, ** p <0.01, *** p<0.001, by Kolmogorov-Smirnov test (D-G) or paired student’s t-test (H-K). (L) A representative frame from a time-lapse of a CK-689-treated B-cell (left) shows AF546-Fab’ clusters 30 sec after contraction began, and enlarged time-lapse images (right) show a single AF546-Fab’ cluster over a 40-sec time window after contraction began. AF546 clusters were identified using the criteria: ≥250 nm in diameter, ≥1.1 fold of FI outside the B-cell contact zone, and trackable for ≥ 20 sec. Scale bars, 2 µm. (M) The peak FI of an AF546-Fab’ cluster was measured over time, and the increasing rate of AF546-Fab’ peak FI of this cluster was determined by the slope of the plot. (N-Q) The rates (±SEM) of increase in AF546-Fab’ peak FI in individual clusters were compared between during and after contraction (N), between B-cells treated with CK-689 and CK-666 from 0 min (O), between DMSO- and Wisko-treated B-cells (P), and between flox control and cNKO B-cells (Q) after B-cells reached maximal spreading. Data points represent individual cells, the averaged slopes of clusters detected in one B-cell, from 3 independent experiments with 6∼12 cells per condition per experiment. *p <0.05, ** p <0.01, ***p<0.001, by non-parametric student’s t-test.

Figure supplement 1. Fab’-PLB, but not Tf-PLB, induces BCR clustering and phosphorylation.

Figure supplement 2. Tracking and analyzing AF546-Fab’ clusters in the B-cell contact zone.

Figure 6-Video 1. Inhibition of B-cell contraction reduces the molecular density within BCR clusters.

Increased molecular density in BCR clusters leads to reductions in BCR phosphorylation.

(A-G) Flox control and cNKO B-cells incubated with AF546-Fab’-PLB were fixed at 1, 3, 5, and 7 min, permeabilized, stained for phosphorylated CD79a (pCD79a, Tyr182), and imaged using TIRF and IRM. (A) Representative IRM and TIRF images of a flox control versus a cNKO B-cell at 7 min. Scale bars, 2 µm. (B-D) Ratios of pCD79a MFI relative to AF546-Fab’ MFI were plotted against AF546-Fab’ peak FI in individual AF546-Fab’ clusters in the contact zone of flox control (B), cNKO B-cells (C), or flox control and cNKO B-cells overlay (D). AF546-Fab’ clusters were identified as described in Figure 6 and Figure 6-figure supplement 2, from an equal number of cells at the 4 time points. Data points represent individual AF546-Fab’ clusters with an equal number of clusters from each time point. The black diamond symbols represent the average ratios of pCD79a MFI to Fab’ MFI in individual BCR-Fab’ clusters within the indicated Fab’ peak FI range. The brown square symbols represent the fraction of the AF546-Fab’ clusters out of the total, within the indicated Fab’ peak FI range. Data were generated from 3 independent experiments with ∼20 cells and ≥125 clusters per condition per experiment. * p <0.05, *** p<0.001, by non-parametric student’s t-test. (E-G) The MFI (±SEM) of AF546-Fab’ (E) and pCD79a (F) and the MFI ratio (±SEM) of pCD79a relative to AF546-Fab’ (G) in individual AF546-Fab’ clusters at indicated times were compared between flox control and cNKO B-cells and between different times. (H-K) WT B-cells treated with CK-689 or CK-666 after 2 min-incubation with AF546-Fab’-PLB. (H) Representative IRM and TIRF images of a CK-689-versus a CK-666-treated B-cell at 7 min. Scale bars, 2 µm. (I-K) The MFI (±SEM) of AF546-Fab’ (I) and pCD79a (J) and the MFI ratio (±SEM) of pCD79a relative to AF546-Fab’ (K) in individual AF546-Fab’ clusters were compared between CK-689- and CK-666-treated B-cells after 7 min stimulation. Data points represent individual clusters. Horizontal solid lines in the violin plots represent the mean, while the dotted lines represent the quartiles of the distribution. Data were generated from 3 independent experiments with ∼20 cells per condition per experiment. * p <0.05, *** p<0.001, by non-parametric student’s t-test. The p-values in D were corrected using the Benjamini-Hochberg/Yekutieli method for false discovery rate control.

Effects of BCR-Fab’ density on the association of Syk with BCR-Fab’ clusters and its phosphorylation.

Primary B-cells from flox control and cNKO mice were incubated with AF546-Fab’-PLB at 37°C, fixed at 3 or 7 min, stained for total Syk and phosphorylated Syk (pSyk Y519/520), and imaged by TIRF. (A-C) Ratios of Syk MFI relative to AF546-Fab’ MFI were plotted against AF546-Fab’ peak IF in individual AF546-Fab’ clusters in the contact zone of Flox control (A) and cNKO B-cells (B) and their overlay (C). AF546-Fab’ clusters were identified as described in Figure 6 and Figure 6-figure supplement 2 from an equal number of cells after 3- and 7-min stimulation. Data points represent individual AF546-Fab’ clusters with an equal number of clusters from each time point. The black diamond symbols represent the average ratios of Syk MFI to Fab’ MFI in individual AF546-Fab’ clusters with indicated Fab’ peak FI ranges. The brown square symbols represent the fraction of the AF546-Fab’ clusters out of the total at indicated Fab’ peak PI ranges. (D-F) Ratios of pSyk MFI relative to AF546-Fab’ MFI were plotted against AF546-Fab’ peak IF in individual AF546-Fab’ clusters in the contact zone of flox control (D) and cNKO B-cells (E) and their overlay (F). (G-I) MFI ratios of pSyk relative to AF546-Fab’ were plotted against AF546-Fab’ peak IF in individual pSyk puncta in the contact zone of Flox control (G) and cNKO B-cells (H) and their overlay (I). pSyk puncta were identified using the criteria: FI ≥1.3 fold of the background outside the B-cell contact zone and diameter ≥250 nm. The black diamond symbols represent the average ratios of pSyk MFI to Fab’ MFI in individual AF546-Fab’ clusters (D-F) or individual pSyk puncta (G-I) at indicated Fab’ peak FI ranges. The brown square symbols represent the fraction of the AF546-Fab’ clusters out of the total at indicated Fab’ peak PI ranges. Data points represent individual AF546-Fab’ clusters (D-F) or pSyk puncta (G and H), collected from an equal number of cells at 3 and 7 min. Data were generated from 3 independent experiments with ∼23 cells per condition per experiment. * p <0.05, ** p <0.01, *** p<0.001, by non-parametric student’s t-test, between AF546-Fab’ cluster group with different Fab’ peak FI ranges. The p-values in C, F and I were corrected using the Benjamini-Hochberg/Yekutieli method for false discovery rate control.

The effects of BCR-Fab’ density on the association of SHIP-1 with BCR-Fab’ clusters and its phosphorylation.

Primary B-cells from flox control and cNKO mice were incubated with AF546-Fab’-PLB at 37°C, fixed at 3 or 7 min, stained for total SHIP-1 and phosphorylated SHIP-1 (pSHIP-1 Tyr1020), and imaged by TIRF. (A-C) Ratios of SHIP-1 MFI relative to AF546-Fab’ MFI were plotted against AF546-Fab’ peak IF in individual AF546-Fab’ clusters in the contact zone of Flox control (A) and cNKO B-cells (B). An overlay of flox control and cNKO plots is also shown (C). AF546-Fab’ clusters were identified as described in Figure 6 and Figure 6-figure supplement 2 from an equal number of cells after 3 and 7 min stimulation. Data points represent individual AF546-Fab’ clusters with an equal number of clusters from each time point. The black diamond symbols represent the average ratios of SHIP-1 MFI to Fab’ MFI in individual AF546-Fab’ clusters within indicated Fab’ peak FI ranges. The brown square symbols represent the fraction of the AF546-Fab’ clusters out of the total at indicated Fab’ peak FI ranges. (D-F) Ratios of pSHIP-1 MFI relative to AF546-Fab’ MFI were plotted against AF546-Fab’ peak IF in individual AF546-Fab’ clusters in the contact zone of flox control (D) and cNKO B-cells (E) and their overlay (F). (G-I) MFI ratios of pSHIP-1 relative to AF546-Fab’ were plotted against AF546-Fab’ peak IF in individual pSHIP-1 puncta in the contact zone of Flox control (G) and cNKO B-cells (H) and their overlay (I). The black diamond symbols represent the average ratios of pSHIP-1 MFI to Fab’ MFI in individual AF546-Fab’ clusters (D-F) or individual pSHIP-1 puncta (G-I) at indicated Fab’ peak FI ranges. The brown square symbols represent the fraction of the total clusters at indicated Fab’ peak FI ranges. pSHIP-1 puncta were identified using the criteria: FI ≥1.5 fold of the background outside the B-cell contact zone and diameter ≥250 nm. Data points represent individual AF546-Fab’ clusters (D-F) and pSHIP-1 puncta (G and H), collected from an equal number of cells at 3- and 7-min. Data were generated from 3 independent experiments with ∼23 cells per condition per experiment. * p <0.05, ** p <0.01, *** p<0.001, by non-parametric student’s t-test, between AF546-Fab’ cluster group with different Fab’ peak FI ranges. The p-values in C, F, and I were corrected using the Benjamini-Hochberg/Yekutieli method for false discovery rate control.