Coro1B and Coro1C regulate lamellipodia dynamics and cell motility by tuning branched actin turnover

Abstract

Actin filament dynamics must be precisely controlled in cells to execute behaviors such as vesicular trafficking, cytokinesis, and migration. Coronins are conserved actin-binding proteins that regulate several actin-dependent subcellular processes. Here, we describe a new conditional knockout cell line for two ubiquitous coronins, Coro1B and Coro1C. These coronins, which strongly co-localize with Arp2/3-branched actin, require Arp2/3 activity for proper subcellular localization. Coronin null cells have altered lamellipodial protrusion dynamics due to increased branched actin density and reduced actin turnover within lamellipodia, leading to defective haptotaxis. Surprisingly, excessive cofilin accumulates in coronin null lamellipodia, a result that is inconsistent with the current models of coronin-cofilin functional interaction. However, consistent with coronins playing a pro-cofilin role, coronin null cells have increased F-actin levels. Lastly, we demonstrate that the loss of coronins increases accompanied by an increase in cellular contractility. Together, our observations reveal that coronins are critical for proper turnover of branched actin networks and that decreased actin turnover leads to increased cellular contractility.

Figure 1.

Generation and characterization of Coro1B and Coro1C matched pair cell lines. (A) Schematic representation of 4-hydroxytamoxifen-induced deletion of Coro1B-GFP and Coro1C. (B) Western blot analysis of matched-pair MTFs with and without Coro1B-GFP and Coro1C. (C) Confocal micrographs showing mixed populations of Parental and Null cells immunostained for Coro1C and phalloidin (F-actin visualization). Only parental cells express Coro1B–GFP and Coro1C. White asterisks denote null cells. Scale bar, 25 μm. (D and E) Representative images for Coro1B-GFP, Coro1C, Arpc2, and F-actin in parental cells. (D) Insets show Coro1B-GFP and Coro1C in magnified regions of the lamellipodia. Scale bar, 5 μm. (E) Parental cells after 30-min treatment with 150 μM CK666 or DMSO. Yellow arrowheads denote the lack of Coro1B or Coro1C along linear actin cables. Scale bar, 30 μm. Source data are available for this figure: SourceData F1.

Figure S1.

F-actin regulators Coro1B and Coro1C are regulators of cell motility in fibroblasts. (A and B) Relative expression of (A) human Coro1B-GFP and (B) endogenous Coro1C in parental and null populations normalized to GAPDH loading control. Plotted as mean ± SEM. (C) Cell growth represented as number of cells (×10⁵) in parental and null populations over time. Three independent experiments were performed in duplicates. (D) Cell spread area (in microns squared) of parental and null cells over 3 h after plating on 10 μg/ml FN. Two independent experiments were performed; n > 300 cells per time point. C and D are plotted as mean with 95% CI. (E) Total distance traveled by parental (n = 223) and null (n = 300) cells in microns. (F) Schematic of mixed parental and null cell populations in the microfluidic haptotaxis chamber. (G) Velocity in microns per hour (left) and total distance traveled in microns (right) of parental (n = 102) and null (n = 139) cells from haptotaxis assays. (H) Protrusion distance of parental (n = 17) and null (n = 16) cells plated on 10 μg/ml uniform FN. For beeswarm superplots, means of experimental replicates are color-coded and overlayed on violin plots respresenting cumulative cell level data. Error bars denote SEM. Student’s t tests were performed for graphs A and B and E–H. *P value = 0.028; **P value = 0.0013; ***P value = 0.0007; ****P value <0.0001.

Figure 2.

Loss of Coro1B/1C affects random and directed migration and lamellipodia dynamics. (A) Sample plot depicting the tracks of parentals (cyan) and null (magenta) cells migrating on 10 μg/ml FN. (B) Beeswarm superplot depicting random migration velocity of parental (n = 223) and null (n = 300) fibroblasts. The mean of each biological replicate is color-coded and overlayed on violin plots. Error bars represent the standard error of mean. (C) Rose plots for parental (n = 102) and null (n = 139) cells on a FN gradient. (D) FMI graph (mean ± 95% confidence interval) for haptotaxis of cells on FN gradient. *P value = 0.04. (E) Example kymographs for parental and null cells. White and black arrows represent protrusion and retraction, respectively. Scale bar, 5 μm. (F–H) Beeswarm superplots of lamellipodial dynamics in parental (n = 17) and null (n = 16) cells showing (F) protrusion rate in microns per minute, (G) retraction rate in micron per minute and, (H) protrusion duration in minutes. For all graphs, Student’s t tests were performed. Error bars represent the SEM. ****P < 0.0001.

Figure 3.

Deletion of Coro1B/1C impact F-actin dynamics at the lamellipodia. (A) Immunofluorescent staining of F-actin in parental and null cells; scale bar, 10 μm. Lower panels are 2× magnifications of boxed regions above; scale bar, 5 μm. (B) Beeswarm superplot depicting the width of actin at the edge of parental and null cells. Widths are calculated from maximum pixel intensities within 5 μm (lines) of the edge. (C) Representative images of Arpc2 localization in parental and null cells. Scale bar, 30 μm. Top right insets are magnifications of yellow boxed region. Scale bar, 5 μm. (D) Percentage of cell edge positive for Arpc2 signal (length of Arpc2-positive lamellipodia/cell perimeter × 100) for parental (n = 47) and null (n = 54) cells. (E) Fluorescence intensity of Arpc2 within 5 μm of the leading edge of parental (n = 45) and null (n = 34) cells. (F) Representative live-cell confocal micrographs of mScarlet-β-actin and Coro1B-GFP in parental and null cells (left). Scale bar, 5 μm. Corresponding kymographs on the right show mScarlet-β-actin returning to the bleached region indicated with a yellow dashed line in the panels on the left. Images were acquired roughly every 1 s and bleached at t = 3 s and t = 33 s as indicated by red arrowheads. (G) Quantification of polymerization rates in microns per minute of parental (n = 75) and null (n = 71) cells. For all beeswarm superplot graphs, Student’s t tests were performed and error bars denote SEM. *P value <0.041, **P value = 0.0018.

Figure 4.

F-actin levels and cofilin activity are impacted by loss of Coro1B and Coro1C. (A) Immunofluorescent staining of parental and null populations for F-actin; Scale bar, 100 μm. (B) Integrated pixel density of phalloidin staining in fixed parental (n = 1,271) and null (n = 1,154) cells from A (top). Blots of total actin from matched whole cell lysates below. Ratios of actin levels relative to the HSC70 loading control are indicated below the blot. (C) Representative still frames from live cell imaging of parental and null cells expressing cofilin-mScarlet. Scale bar,10 μm. (D) Immunofluorescent staining for endogenous cofilin in parental and null cells. Scale bar, 10 μm. (E) Width of cofilin at the edge of parental (n = 46) and null (n = 48) cells. (F) Fold change in polymerization rate of parental and null cells after treatment with 100 nM jasplakinolide. Data is displayed using a logarithmic transformation to the base 2 and color-coded by experimental replicate. For bar graph, an unpaired Student’s t test with Welch’s correction was performed and error bars represent standard deviation. For beeswarm superplots, the mean of experimental replicates are color-coded and overlayed on violin plots representing cumulative cell level data. Error bars represent SEM. For all graphs, Student’s t tests were performed, unless otherwise stated. *P value = 0.04, **P value = 0.0032, ****P value <0.0001. Source data are available for this figure: SourceData F4.

Figure S2.

Deletion of Coro1B and Coro1C do not significantly impact Arpc2 and cofilin levels in fibroblasts. (A) Western blot analysis of Arpc2 in parental and null cells. (B) Width of Arpc2 at the edge in microns of parental (n = 22) and null (n = 24) cells. (C) Blot analysis of total cofilin and phospho-cofilin (p-cofilin) in parental and null cells. Average ratio of p-cofilin to total cofilin levels are depicted below blot. (D) Quantification of p-cofilin/cofilin ratios from parental and null populations. Each dot represents an independent experiment and error bars denote SD. (E) Quantification of the ratio between cofilin and F-actin fluorescence at the leading edge of parental (n = 46) and null (n = 46) cells. (F) Fluorescence intensity profiles of cofilin and actin in parental (n = 7) and null (n = 8) cells within 3 μm from the cell edge. Black dotted lines denote the average peak in F-actin intensity and black arrows depict a similar peak in cofilin fluorescence in parental and null cells. (G) Quantification of polymerization rates in microns per minute of parental and null cells in the presence and absence of 100 nM jasplakinolide; n > 25. Individual cell data from three experimental replicates are overlayed on bar graph, and error bars denote SD. Two-way ANOVA was performed. For all beeswarm superplots, the mean of experimental replicates are color-coded and overlayed on violin plots representing cumulative cell level data. For all graphs, Student’s t tests were performed, and error bars denote SEM, unless otherwise stated. *P value <0.04, ****P value <0.0001. Source data are available for this figure: SourceData FS2.

Figure 5.

Loss of Coro1B/1C increases contractility. (A) Representative collagen contractility gels. (B) Quantification of collagen gel area, 24 h post-plating, from control (n = 10), parental (n = 24), and null (n = 20) population, acquired from three independent experiments. One-way ANOVA with Dunn’s multiple comparison test performed. Error bars show 95% C.I. (C) Representative traction force maps of parental and null cells plated on 8 kPa polyacrylamide hydrogels. Scale bar, 50 μm. Scale (right) shows traction force magnitude in pascals. (D) Average strain energy density (in femtojoules per micron squared) of parental (n = 40) and null (n = 24) cells extracted from traction force map in C. Data represents two independent experiments with two technical replicates in each experiment. (E) Immunofluorescent staining of vinculin and F-actin in parental and null populations. Scale bar, 100 μm. Lower panels are 2× magnifications of vinculin and F-actin from boxed regions. (F) Graph depicts number of vinculin-positive adhesions per cell; parental (n = 19) and null (n = 18) cells. (G–I) Quantification of (G) vinculin fluorescence intensity; (H) area (in micron squared); and (I) aspect ratio of parental (n = 49) and null (n = 50) cells. For all beeswarm superplots, the mean of experimental replicates are color-coded and overlayed on violin plots representing cumulative cell-level data. Error bars denote SEM. For all graphs, Student’s t tests were performed. **P value <0.005, ***P value <0.0005, ****P value <0.0001.

Figure S3.

Deletion of Coro1B and Coro1C do not directly affect NMIIA localization and activity. (A) Still frames of parental and null cells (top) plated on 8 kPa polyacrylamide gels. Corresponding still images of fluorescent beads below. (B) Western blot analysis of myosin light chain (MLC) and phosphorylated MLC (pMLC) in parental and null populations. pMLC/MLC ratios are provided below. (C) Bar graph representing quantification of pMLC to total MLC ratios across five independent experiments. (D and E) Integrated fluorescence intensity of (D) MLC and (E) pMLC in parental (n = 181) and null (n = 193) cells. (F) Representative immunofluorescent staining of non-muscle myosin IIA, vinculin and F-actin in parental and null cells; Scale bar, 10 μm. Yellow arrowheads mark accumulation of NMIIA along actin bundles associated with focal adhesions. For all beeswarm superplots, the mean of experimental replicates are color-coded and overlayed on violin plots representing cumulative cell level data. For all graphs, Student’s t tests were performed, and error bars denote SEM, unless otherwise stated. Source data are available for this figure: SourceData FS3.