Rho GTPase activity crosstalk mediated by Arhgef11 and Arhgef12 coordinates cell protrusion-retraction cycles

Abstract

Rho GTPases play a key role in the spatio-temporal coordination of cytoskeletal dynamics during cell migration. Here, we directly investigate crosstalk between the major Rho GTPases Rho, Rac and Cdc42 by combining rapid activity perturbation with activity measurements in mammalian cells. These studies reveal that Rac stimulates Rho activity. Direct measurement of spatio-temporal activity patterns show that Rac activity is tightly and precisely coupled to local cell protrusions, followed by Rho activation during retraction. Furthermore, we find that the Rho-activating Lbc-type GEFs Arhgef11 and Arhgef12 are enriched at transient cell protrusions and retractions and recruited to the plasma membrane by active Rac. In addition, their depletion reduces activity crosstalk, cell protrusion-retraction dynamics and migration distance and increases migration directionality. Thus, our study shows that Arhgef11 and Arhgef12 facilitate exploratory cell migration by coordinating cell protrusion and retraction by coupling the activity of the associated regulators Rac and Rho.

Fig. 1. A general method for rapid perturbation of Rho GTPase activity in living cells.

a Schematic of rapid, reversible Rho GTPase activity perturbation via chemically-induced dimerization and readout of cell morphology and cytoskeletal organization by an actin-reporter. b Representative frames from TIRF microscopy time series of mCherry-Actin obtained 30 s before, 24 min during and 24 min after Rho GTPase activation in Neuro-2a neuroblastoma cells (see also Supplementary Movie 1). Yellow arrows point to cell areas that reversibly generate protrusions during Rac1 or Cdc42 activation, and yellow arrowheads point to areas that undergo reversible retraction during RhoA activation. Observations are representative for 3 independent repetitions with a total of at least 40 cells per condition (exact numbers of cells are indicated in individual panels). Numbers in middle panels indicate percentage ± standard error of the mean of cells that initiate protrusion (Rac1/Cdc42) or retraction (RhoA) after addition of dimerizer. Numbers in panels indicate percentage of reacting cells that showed a phenotypic reversal. Scale bar: 10 μm; 0.26 μm/pixel; CFP cyan fluorescent protein, BFP blue fluorescent protein, RFP red fluorescent protein, FKBP’ FK506-binding protein with F36V mutation, eDHFR E. coli dihydrofolate reductase, SLF’ synthetic ligand of FKBP’, TMP eDHFR interacting small molecule trimethoprim.

Fig. 2. Analysis of Rho GTPase crosstalk in living cells.

a Schematic of rapid activity perturbation and combined activity measurement strategy. b–e Analysis of perturbation-response relationships of Rho GTPase activity in Neuro-2a cells. b, c Representative TIRF images before dimerizer addition (b, top) and Rac1 perturbation and uncorrected, raw Rho sensor and raw control sensor signal kinetics (c, bottom) corresponding to orange boxes (see also Supplementary Movie 2). All constructs are predominantly cytosolic and homogenously distributed in the cell bodies and neurite-like protrusions. d Average perturbation and control-corrected activity sensor signal kinetics for selected crosstalk combinations (see Supplementary Fig. 2 for all combinations). e Quantification of average sensor signal changes during Rho GTPase activity perturbation. Responses at time points before (pre), and 5 min after dimerizer addition (on) are shown. f Influence diagram that summarizes significant activity response measurements at 5 min after dimerizer addition. All observations and measurements are based on at least 3 independent repetitions with a total of at least 28 cells per condition (exact numbers of cells are indicated in individual panels). Scale bars: 5 µm; 0.26 μm/pixel; ****P < 0.0001; Student’s t-Test. Error bars represent standard error of the mean. YFP yellow fluorescent protein, GBD GTPase-binding domain, All statistical tests were two-sided. Source data are provided as a Source Data file.

Fig. 3. Rac/Rho crosstalk is commonly observed in adherent mammalian cell lines and can trigger a dynamic Rho activity response.

a Scheme of optogenetic approach to measure Rac/Rho activity crosstalk. b–f Measurement of Rho activity dynamics during rapid optogenetic activation of Rac. b, f Rho activity response to Rac1 perturbation in the commonly used cell lines NIH3T3, HeLa, N2a and U2OS. The difference between 5 measurements before and 5 measurements after the onset of illumination is shown, each corresponding to a time frame of 25 s. c Typical Rho activity response dynamics observed in A431 cells (n = 49 cells from 3 independent experiments). Left: representative TIRF images from video microscopy time series. Right: Rho activity dynamics corresponding to black boxes in left panels. 67% of cells showed a discernible, positive, reversible Rho activity response. In particular, 29% showed a continuous Rho activation during Rac1 photoactivation (top panels), 38% of the cells showed a single, transient activity pulse response and 25% showed no response (middle panels). 8% of cells showed a negative response (bottom panels). d Heatmap representation of all Rho activity responses measured in A431 cells. Color represents % change of the Rho activity sensor. e Measurement of average Rho activity sensor kinetics before, during and after Rac1 activation in A431 cells, corresponding to data shown in (f). g Measurement of average Rho activity sensor kinetics corresponding to data shown in (e) in peripheral or central cell attachment areas. Scale bars: 10 µm; 0.26 μm/pixel; ****P < 0.0001; ***P < 0.001; Student’s t-Test. Error bars represent standard error of the mean. The time range of optogenetic activation is indicated in plots with the label “445 nm ON”. All statistical tests were two-sided. Source data are provided as a Source Data file.

Fig. 4. Sequential Rac and Rho activation is tightly coupled to cell protrusion-retraction cycles in space and time.

a Representative A431 cell expressing a cytosolic cell volume marker (mCitrine) and an improved Rac activity sensor (mCherry-3xp67^phoxGBD; top; see also Supplementary Movie 3; n = 23 cells from 3 independent experiments). The white arrow marks the direction of a local cell protrusion. b Automated tracing of the cell border using a modified version of the ADAPT plugin. c Maps generated by the ADAPT plugin that represent the spatio-temporal dynamics of the Rac sensor signal between the green lines in (b) (left) and of the cell edge velocity (right). The yellow arrows in (c) point to the local protrusion that occurs at the position of the cell area marked by the yellow arrow in (b). Red areas in the velocity map correspond to local cell protrusions, blue areas to local cell retractions. d Plot of Rac sensor signals and cell edge velocity corresponding to the yellow dotted line in (c). e Representative TIRF images (left) of A431 cells that generate spontaneous protrusion-retraction cycles and express the Rac or Rho GTPase activity sensors and the cell volume marker (see also Supplementary Movies 4 and 5). White arrows represent the protrusion direction. Kymographs (right) correspond to white arrows in TIRF images. f Crosscorrelation between Rac sensor signal and cell edge velocity plotted against the time shift between these measurements. g Enrichment of Rac and Rho sensor signals in protrusions (>0.075 μm/min) and retractions (<−0.075 μm/min). Values are normalized to average control sensor enrichment measurements. n = 3 independent experiments with >21 cells per condition. Error bars represent standard error of the mean. Measurements corresponding to individual cells are shown in Supplementary Fig. 5c, d. Scale bars: 10 µm; 0.26 μm/pixel. PH Pleckstrin homology domain, DH Dbl homology domain, Source data are provided as a Source Data file.

Fig. 5. Identification of Arhgef11 and Arhgef12 as Rac effectors in local cell protrusion-retraction cycles.

a Schematic representation of a hypothetical mechanism, by which Lbc-type GEFs could mediate Rac1/Rho activity crosstalk. b TIRF microscopy images (top panels) and protrusion-retraction enrichment functions (bottom panels) for representative cells that express Lbc-type GEFs (CMV-GEF, green) and a cytosolic cell volume marker that acts as a control construct (delCMV-mCitrine, magenta; n > 10 cells from 3 independent experiments). White and yellow arrows point to local cell retractions and protrusions, respectively. c Representative TIRF images (left) of A431 cells that generate spontaneous protrusion-retraction cycles and express Arhgef11 and Arhgef12 fused to mCherry and the cytosolic cell volume marker (mCitrine, see also Supplementary Movies 6 and 7). White arrows represent the protrusion direction. Kymographs (right) correspond to white arrows in TIRF images. d Crosscorrelation (left) between Arhgef11/Arhgef12 signals and cell edge velocity plotted against the time shift between these measurements, and enrichment (right) of Arhgef11/Arhgef12 signals in protrusions and retractions. Arhgef11/Arhgef12 enrichment values are normalized to average control construct enrichment measurements. n = 3 independent experiments with >22 cells per condition e Direct comparison of signal enrichment of active Rac, Arhgef11, Arhgef12 and active Rho relative to the time period of cell protrusion. Black arrows indicate the time point of maximal sensor or GEF enrichment. The measurements shown in this panel are identical to measurements shown in Fig. 4g and Fig. 5d. Error bars represent standard error of the mean. Measurements corresponding to individual cells for panels (d) and (e) are shown in Supplementary Fig. 5c, d. PDZ PSD-95 Dlg ZO-1 domain, RGS Regulator of G protein signaling domain, FAB F-actin binding domain, Source data are provided as a Source Data file.

Fig. 6. Characterization of Rac-dependent Arhgef11 and Arhgef12 plasma membrane recruitment.

a Schematic representation of the Arhgef11 constructs that were used investigate the mechanism of Rac-stimulated plasma membrane recruitment. b–f Measurement of full length or mutant Lbc-type GEF plasma membrane recruitment during Rac activation in A431 cells that co-express PA-Rac1 (n = 3 independent experiments). b, d, e Measurement of recruitment kinetics. c, f Quantification of recruitment in the 25 s time frame during photoactivation (early response) or 1 min after photoactivation (late response). d Measurement of Arhgef11/12 plasma membrane recruitment in peripheral vs central cell attachment areas. g, h Quantification of average Rho activity sensor kinetics before, during and after Rac1 activation in A431 cells that co-express the Rho activity sensor, PA-Rac1 and control or Arhgef11/12 targeting siRNA oligonucleotides. g Measurement of average Rho activity sensor kinetics, corresponding to data shown in (h). h Quantification of the Rho activity response in the 25 s time frame during photoactivation. ****P < 0.0001; **P < 0.01; *P < 0.05; One-way ANOVA with Tukey’s (c, f) or Holm-Sidak’s (h) post test. Error bars represent standard error of the mean. Scale bars: 10 µm; 0.26 μm/pixel. All statistical tests were two-sided. Source data are provided as a Source Data file.

Fig. 7. Arhgef11 and Arhgef12 mediate Rac-dependent Rho activation and the spatio-temporal coordination of local cell protrusion-retraction cycles.

a–d Quantification of protrusion and retraction dynamics in A431 cells with increased (a, b) or decreased (c, d) Arhgef11 and Arhgef12 expression levels. a, c TIRF images (top panels) and cell edge velocity maps (bottom panels) of representative cells that express CMV-mCherry-Arhgef11/Arhgef12 and delCMV-mCitrine (a), or Arhgef11/Ahrgef12 targeting siRNA and delCMV-mCherry (c). b, d Quantification of protrusion-retraction (P-R) cycle duration based on cell edge velocity measurements corresponding to panels (a) and (c), respectively (a, b n = 3 independent experiments with >26 cells per condition, c, d n = 3 independent experiments with >105 cells per condition). Differences in the average values obtained for the two control conditions are presumably due to the experimental protocols, which differ significantly between (b) and (d) (single vs. dual transfection; see “Methods” for details). e Schematic representation of distance (magenta) and displacement (green) for typical spontaneous exploratory cell migration. The distance corresponds to the length of the cell migration trajectory which leads to the indicated displacement between the start and end locations. The directionality is defined as the ratio between these length measurements. f–h Quantification of distance (f), displacement (g) and directionality (h) of A431 cell trajectories over a 4 h time period in control and Arhgef11/Arhgef12 depleted cells (n = 3 independent experiments with >491 cells per condition). (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; One-way ANOVA with Holm-Sidak’s post test). Images were recorded at a frame rate of 1.5/min (a–d) or 1/min (f–h). Error bars represent standard error of the mean. Scale bars: 10 µm; 0.26 μm/pixel. All statistical tests were two-sided. Source data are provided as a Source Data file.

Fig. 8. Proposed mechanism for the generation of protrusion-retraction cycles.

a Schematic for spatio-temporal events that couple cell protrusion and retraction and the signal molecules that mediate this coupling. b Effect of increasing or decreasing Arhgef11/12 levels on local Rac/Rho crosstalk and cell morphodynamics.