Ras suppression potentiates rear actomyosin contractility-driven cell polarization and migration

Abstract

Ras has been extensively studied as a promoter of cell proliferation, while few studies have explored its role in migration. To investigate the direct and immediate effects of Ras activity on cell motility or polarity, we focused on RasGAPs, C2GAPB in Dictyostelium amoebae and RASAL3 in HL-60 neutrophils and macrophages. In both cellular systems, optically recruiting the respective RasGAP to the cell front extinguished pre-existing protrusions and changed migration direction. Surprisingly, however, when these respective RasGAPs were recruited uniformly to the membrane, cells polarized and moved more rapidly while targeting to the back exaggerated these effects. These unexpected outcomes of attenuating Ras activity naturally had strong, context-dependent consequences for chemotaxis. The RasGAP-mediated polarization depended critically on myosin II activity and commenced with contraction at the cell rear, followed by sustained TORC2-dependent actin polymerization at the front. These experimental results were captured by computational simulations in which Ras levels control front- and back-promoting feedback loops. The discovery that inhibiting Ras activity can produce counterintuitive effects on cell migration has important implications for future drug-design strategies targeting oncogenic Ras.

Introduction

Ras GTPases play a vital role in transmitting signals within cells, influencing growth and survival, and activating mutations in Ras genes are found in ~30% of all cancers^1–3. Therefore, targeting mutant Ras has become a major focus in cancer drug development. While extensive research has been dedicated to understanding the role of Ras mutations in promoting cancer growth, the impact of these mutations on metastasis has received less attention. Nevertheless, studies conducted in model systems, Dictyostelium amoebae and human neutrophils, suggest that local Ras activity plays a direct and immediate role in tuning polarity and motility^4–9.

Ras is activated both by stimulation of G-protein coupled receptors and spontaneously at cellular protrusions^6,10–13. Ras function is spatiotemporally controlled by its activators, RasGEFs, and inhibitors, RasGAPs, which might be expected to create ‘front’ and ‘back’, respectively, during migration^11,14–19. However, Ras activity must be carefully balanced, as constitutively active Ras expression leads to hyper-activation of PI3K/TORC2-PKB pathways, causing significant cell spreading and defective migration^5,20,21. Additionally, knockout/knockdown of various RasGAPs have demonstrated their significant impact on Ras and protrusive activities^{8,11,17,18,22,23}. For example, genetic deletion of NF1 in Dictyostelium or humans causes increased macropinocytosis or severe neurofibromatosis type 1, respectively^24–26.

Understanding the physiological relevance of manipulating Ras activity is of paramount importance, but knockout studies have proven relatively ineffective, possibly due to redundancy. For example, Dictyostelium lacking Ras isoforms continue to grow and directed migration remains little affected^27–29. Furthermore, while new treatments with small molecule inhibitors targeting constitutively active KRasG12C show promise, they also pose significant challenges^30–32. A potentially powerful alternative route of suppressing Ras would be to activate RasGAPs. Indeed, in immune cells, locally activating a RasGAP could inhibit Ras and halt chemotaxis⁴.

We designed a series of studies to inhibit Ras activity using RasGAP proteins and see its immediate effects on cell behavior and shape. Surprisingly, different RasGAPs in amoebae and leukocytes can polarize cells and improve migration. Our evidence suggests that polarization was due to an increase in suppression of Ras activity leading to increased contraction at the back. We show that a direct linear relationship does not exist between Ras activation and motility, and an optimal level of activated Ras potentiates migration and polarity.

Results

Reduction of Ras and PI3K activities at cell fronts shuts off protrusions

We examined the effects of C2GAPB in modulating Ras and downstream PI3K activities in growing Dictyostelium cells. Since we encountered challenges in expressing C2GAPB using standard methods, we developed a doxycycline-inducible system to improve expression. We first compared Ras/PI3K activities by expressing fluorescence-labeled biosensors, RBD or PHcrac, respectively, with or without C2GAPB expression. Confocal imaging of midline of cells revealed broad patches of RBD/PIP3 at the protrusions in absence of C2GAPB. However, upon overnight induction of C2GAPB expression, these patches reduced in size (Fig. 1a–c). Ras/PIP3 patches underlie protrusions which mediate cell movement and macropinocytosis^11,33,34. Uptake measurements showed that macropinocytosis was significantly reduced upon C2GAPB expression (Extended Data fig. 1a,b). Alternatively, RBD patches at cell fronts were strongly diminished upon inducing expression of dominant negative RasG (RasG S17N) (Extended Data fig. 1c).

Figure 1. C2GAPB inhibits Ras and PI3K activities in single and electrofused Dictyostelium.

Confocal images of vegetative Dictyostelium single cells expressing (a) GFP-RBD (biosensor for activated Ras; green) or (b) PHcrac-YFP (biosensor for PIP3; green) before and after doxycycline-induced mRFPmars-C2GAPB (red) expression. Pink arrows highlight fronts of these cells. Scale bars represent 5 µm. (c) Cartoon summarizes our observations in (a) and (b) that both Ras activation and PIP3 level on the cell membrane significantly reduce with C2GAPB expression. Time-lapse confocal images of vegetative Dictyostelium electrofused or ‘giant’ cells expressing (d) GFP-RBD (green) or (g) PHcrac-YFP (green) before and after doxycycline-induced mRFPmars-C2GAPB (red) expression. Pink arrows point at reduced RBD or PHcrac waves in presence of C2GAPB. Time in min:sec format. Scale bars represent 5 µm. RBD or PHcrac wave area (e or h) and duration (f or i) before (‘-Dox’; black) and after (‘+Dox’; red) doxycycline-induced mRFPmars-C2GAPB expression. n=10 (-Dox, f), n=27 (+Dox, f), n=10 (-Dox, i), or n=31 (+Dox, i) waves pooled from 10 cells, examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (Two-sided Mann-Whitney test; no adjustments were made for multiple comparisons). The boxes (f, i) extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). (j) Cartoon depicting reduction in size and duration of RBD and PHcrac waves on the basal surface of giant Dictyostelium cells, after overnight doxycycline induction of C2GAPB. Source numerical data are provided.

Ras and PI3K activities form cortical waves on the basal surface of electrofused, giant Dictyostelium cells^{12,13,35–40}. We have used waves in electrofused cells or migration in single cells interchangeably throughout this study. Confocal images of migrating cells offer a cross-sectional view, but in 3D, protrusions extend outwardly while expanding laterally, as a traveling wave. Wave characteristics strongly correlate with protrusion types and migration mode. Upon inducing C2GAPB expression, Ras/PIP3 waves became smaller and more transient (Fig. 1d,g,j and Supplementary Videos 1–4). In absence of C2GAPB, RBD/PHcrac wave area at steady-state varied between 35–80% of total cell area, but with C2GAPB, wave area decreased to less than 20% (Fig. 1e,h). Mean duration of a wave also decreased from 5 minutes to less than a minute (Fig. 1f,i). Conversely, in C2GAPB-null cells, broad, propagating RBD waves were observed, validating that C2GAPB suppresses Ras activity (Extended Data fig. 1d)¹¹.

Next, C2GAPB was optically recruited to cellular protrusions at the leading edge. Consequently, mature protrusions quickly vanished causing the cell to contract, while new protrusions emerged at the former back, and the cell started to move in the opposite direction (Fig. 2a,c). Angular histogram analyses revealed that the probability of nascent protrusion formation was highest at ~120–150 degrees away from C2GAPB recruitment area (Fig. 2d). To confirm that polarity reversal was due specifically to Ras suppression, we locally recruited CAAX-deleted RasG S17N to pre-existing cell fronts which diminished them immediately (Extended Data fig. 1e). In contrast, empty vector recruitment did not block production of new protrusions, and the cell continued to move in its original direction (Fig. 2e,f).

Figure 2. C2GAPB or RASAL3 recruitment shuts off protrusions by inhibiting Ras/PI3K activities in Dictyostelium or neutrophils.

Cartoon illustrating (a) opto-C2GAPB or (b) opto-RASAL3 recruitment to the cell front after locally applying blue light. Time-lapse confocal images of vegetative Dictyostelium single cells expressing (c) mRFPmars-SspB R73Q-C2GAPB or (e) tgRFPt-SspB R73Q-Ctrl (control without C2GAPB). C2GAPB or Ctrl is recruited to migrating cell front by applying 488 nm laser near it, as shown by dashed white box. White arrows denote existing older protrusions whereas pink arrows highlight emerging newer protrusions. Time in min:sec format. Scale bars represent 5 µm. Polar histogram demonstrates higher probability of fresh protrusion formation away from C2GAPB recruitment area (d) whereas for Ctrl, new protrusions form near the recruitment area (f). n=45 protrusions pooled from 11 cells (d) and n=76 protrusions from 13 cells for (f), examined over 3 independent experiments. (g, h) Time-lapse confocal images of vegetative Dictyostelium single cells expressing mRFPmars-SspB R73Q-C2GAPB (upper panel) and (g) RBD-YFP or (h) PHcrac-YFP (both lower panels) after 488 nm laser was switched on globally. RBD or PHcrac status at ‘00:00’ is considered as control timepoint since C2GAPB was not recruited yet. Pink arrows denote RBD or PHcrac patches in cells. Time in min:sec format. Scale bars represent 5 µm. (i, j) Time-lapse confocal images of Dictyostelium electrofused cells expressing mRFPmars-SspB R73Q-C2GAPB (upper panel; red) and (i) RBD-YFP or (j) PHcrac-YFP (both lower panels; yellow) before and after 488 nm laser was switched on globally. White arrows highlight C2GAPB recruitment in red channel whereas pink arrows denote RBD or PHcrac waves near bottom cell surface. Time in min:sec format. Scale bars represent 10 µm. (k) Time-lapse confocal images of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel). RASAL3 was recruited to the cell front by applying 488 nm laser near it, as shown by dashed white box. White arrows denote existing older protrusions whereas pink arrows highlight emerging newer protrusions. Time in min:sec format. Scale bars represent 5 µm. (l) Polar histogram demonstrates higher probability of fresh protrusion formation away from RASAL3 recruitment area; n=30 protrusions pooled from 12 cells over 3 independent replicates. Source numerical data are provided.

Since C2GAPB had a strong effect on cell behavior, we checked whether it was having this effect by reducing Ras/PI3K activities. While unrecruited cells displayed multiple, small RBD/PHcrac patches (‘00:00’ in Fig. 2g,h), once C2GAPB was recruited it caused a simultaneous reduction in activated Ras/PIP3 levels at the protrusions (‘00:15’-’2:30’ in Fig. 2g,h). Similarly, activated Ras/PIP3 propagating waves substantially reduced upon C2GAPB recruitment (Fig. 2i,j and Supplementary Videos 5 and 6). Once we switched off blue light, RBD waves recovered within a minute (Supplementary Video 5). In latrunculin-treated cells, dynamic RBD membrane patches disappeared ~30 seconds of C2GAPB recruitment (Extended Data fig. 1f).

We extended our investigation to differentiated neutrophils to assess the conservation of RasGAP function on cytoskeletal remodeling. We previously developed a membrane recruitable RASAL3 in neutrophils (Fig. 2b)⁴. Optically recruiting RASAL3 to F-actin-rich front, marked by LifeAct, caused mature protrusions to immediately disappear. Simultaneously, a new broad front emerged at the opposite end, causing cell to migrate away (Fig. 2k and Supplementary Video 7). New protrusions appeared at ~130–170 degrees from the recruitment site (Fig. 2l). Alternatively, recruiting KRas4B S17N ΔCAAX to mature fronts caused them to fold and form fresh ones instantaneously at the erstwhile back (Extended Data fig. 1g,h and Supplementary Video 8). Altogether, these observations demonstrate that suppressing Ras activity at the cell front caused immediate shutdown of local signaling, and consequently protrusions and migration.

Global Ras suppression induces polarity and enhances cell migration

Given the results from Figs. 1 and 2, we expected that ectopic RasGAP expression would strongly inhibit cellular activity and migration. Surprisingly, inducing C2GAPB expression resulted in a more polarized phenotype (Fig. 3a, also see Fig. 1a,b). Moreover, polarized C2GAPB-expressing cells exhibited accelerated movement compared to control cells expressing empty vector (Fig. 3b,c and Supplementary Video 9). Quantification revealed 2-fold increase in average cell speed and aspect ratio, a proxy for cell polarity, but not basal cell area (Fig. 3d–f). This unexpected C2GAPB-induced polarity was a result of Ras suppression since inducible RasG S17N expression similarly improved polarity and migration (Extended Data fig. 1c, i and j, and Supplementary Video 10).

Figure 3. C2GAPB expression polarizes Dictyostelium and improves random cell migration.

(a) Confocal images demonstrating doxycycline-induced mRFPmars-C2GAPB expression (red) polarize vegetative Dictyostelium cell (shown in DIC). Scale bars represent 5 µm. (b) Time-lapse confocal images of vegetative Dictyostelium cells expressing tgRFPt-Ctrl (empty vector control, red; top panel) or mRFPmars-C2GAPB (red; bottom panel) after overnight doxycycline treatment. Time in ‘min’ format. Scale bars represent 5 µm. (c) Color-coded (1-min interval) outlines of the Ctrl- or C2GAPB-expressing cell shown in (b). Box-and-whisker plots of (d) average cell speed, (e) aspect ratio, and (f) cell area before (black; -DOX) or after (red; +DOX) overnight doxycycline induction of C2GAPB. n=35 (-Dox) or n=34 (+Dox) cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (d,e); ns denotes non-significant difference, P=0.4167 (f) (Two-sided Mann-Whitney test). The boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Time in min:sec format. Scale bars represent 5 µm. Time-lapse confocal images of vegetative Dictyostelium cells expressing (g) GFP-LimE_∆coil (LimE, F-actin biosensor; green) or (k) myosin II-GFP (green), before (-DOX) and after (+DOX) overnight doxycycline-induction of C2GAPB (red; not shown here). White or pink arrows denote cells of interest. Time in ‘min’ format. Scale bars represent 5 µm. Representative linescan of (h) LimE or (l) myosin II intensity of cells in (g) or (k) respectively. Representative kymograph of cortical (i, j) LimE or (m, n) myosin II intensity in cells in (g) or (k) respectively. A linear color map shows that blue is the lowest LimE or myosin II intensity whereas yellow is the highest. (o) Cartoon summarizing the polarizing effects of expressing C2GAPB in Dictyostelium. Unpolarized cells with small, transient protrusions become polarized with a distinct F-actin front and a myosin II-labelled back after C2GAPB is expressed. Source numerical data are provided.

Wild-type Dictyostelium cells typically displayed multiple, transient, small F-actin fronts, whereas C2GAPB-null cells demonstrated an unpolarized morphology with wide fronts (Fig. 3g and Extended Data fig. 2a). However, in C2GAPB-expressing cells, there was a single, broad, polarized LimE-rich protrusion persistently localized at the front (Fig. 3g,o). Similarly, C2GAPB expression caused myosin II to localize predominantly at the cell back, while a more diffused distribution was generally observed in uninduced cells (Fig. 3k,o). These observations were supported by linescan and kymograph analyses (Fig. 3h–j and l–n). This counterintuitive result could be recapitulated with RasG S17N expression (Extended Data fig. 2b). Although polarity acquisition is a defining characteristic of Dictyostelium differentiation, C2GAPB-induced polarization neither required development nor GPCR signaling. Also, C2GAPB improved polarity and migration in Gβ-null cells which are typically unpolarized with low motility (Extended Data fig. 2c and d)^41,42.

We next checked whether recruiting C2GAPB to the cell membrane, in closer association with Ras, would exaggerate its polarization effects. We noticed that expressing the recruitable C2GAPB protein, fused to SspB, only moderately induced polarity (Fig. 4b), as compared to the native C2GAPB (Figs. 1 and 3). The SspB tag may be responsible for a weaker C2GAPB function, but was not explored here. However, global recruitment of C2GAPB-SspB triggered a robust ‘instant polarization’ response within a minute resulting in a 2-fold increase in average migration speed and aspect ratio, but not basal cell area (Fig. 4a–e and Supplementary Videos 11 and 12). Once blue laser was switched off, cells reverted to their original morphology within minutes (Fig. 4f–h and Supplementary Video 12). There were several F-actin rich fronts in unrecruited cells (‘00:00’ in Fig. 4i). C2GAPB recruitment first reduced LimE signals everywhere on the membrane, then as the cell polarized, a single and persistent actin polymerization site grew at the leading front (Fig. 4i and j). We also successfully polarized Gβ-null cells with C2GAPB recruitment (Extended Data fig. 2e and Supplementary Video 13).

Figure 4. Global C2GAPB recruitment polarizes Dictyostelium and enhances cell migration.

(a) Cartoon illustrating mechanism of opto-C2GAPB global recruitment on Dictyostelium cell membrane with SspB-iLID optogenetic system. (b) Time-lapse confocal images of vegetative Dictyostelium cell expressing tgRFPt-SspB R73Q-Ctrl (control without C2GAPB; left panels) or mRFPmars-SspB R73Q-C2GAPB (right panels), before or after 488 nm laser was switched on globally. Bottom panels demonstrate color-coded (at 1-min interval) outlines of the Ctrl- or C2GAPB-recruited cell. Time in min:sec format. (f) Color-coded (1-min interval) outlines of a representative cell in presence of intermittent 488 nm light. Scale bars represent 10 µm. Box-and-whisker plots of (c, g) average cell speed, (d, h) aspect ratio, and (e) cell area, before (black) or after (red) C2GAPB global recruitment. For plots in g and h, laser was switched on or off multiple times. (c-e, g, h) The boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention. Connecting lines are provided between paired data points obtained from the same cell, before or after C2GAPB recruitment (GraphPad Prism 8). (c-e,g,h) n=24 opto-C2GAPB-expressing cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (c,d,g,h); ns denotes non-significant difference, P=0.1011 (e), P=0.7257 (g), P=0.0557 (h) (Two-sided Wilcoxon signed-rank test). (i) Time-lapse confocal images of vegetative Dictyostelium single cells co-expressing mRFPmars-SspB R73Q-C2GAPB (upper panel) and LimE-YFP (middle panel) after 488 nm laser was switched on globally. LimE patches are highlighted with pink arrows. ‘00:00’ is considered as control timepoint since C2GAPB has not been recruited yet. DIC channel (bottom panel) shows change in cell polarity with C2GAPB recruitment. Time in min:sec format. Scale bars represent 5 µm. (j) Cartoon illustrates phenomenon in cell in (i). (k) Time-lapse confocal images of electrofused Dictyostelium giant cells co-expressing mRFPmars-SspB R73Q-C2GAPB (upper panel; red) and LimE-YFP (lower panel; yellow) before or after laser was switched on globally. White arrows highlight C2GAPB recruitment in red channel whereas pink arrows denote LimE propagating waves near the bottom cell surface. Time in min:sec format. Scale bars represent 10 µm. (l) Cartoon illustrates phenomenon in giant cell in (k). Source numerical data are provided.

Alternatively, globally recruiting RasG S17N ΔCAAX caused Dictyostelium to polarize and migrate rapidly whereas recruiting empty vector did not induce any change (Extended Data fig. 3a–e, Fig. 4b and Supplementary Video 14). Although, these effects were relatively milder compared to C2GAPB recruitment, both showed a similar trend. In neutrophils, KRas4B S17N ΔCAAX recruitment formed a long uropod, localized protrusions at the front, and improved migration. Non-recruiting cells had multiple, transient smaller protrusions and did not move much (Extended Data fig. 3f–m and Supplementary Video 15).

Next, we validated these C2GAPB-induced effects on LimE wave patterns on the basal surface of electrofused cells. Like single cells, expressing recruitable C2GAPB modestly affected wave patterns (Fig. 4k). However, upon recruitment, F-actin waves largely disappeared throughout the cell except for one region on the basal membrane where LimE signal remained strong and did not propagate (Fig. 4k and l, and Supplementary Video 16). Thus, our data shows that both signaling and cytoskeletal waves are largely extinguished but cytoskeletal activity remains persistently confined to one location on the membrane, presumably indicating sustained polarity.

It is well-established that response of growth-stage amoebae to folic acid is slow and less efficient, presumably due to their unpolarized morphology and transient protrusions^43,44. We therefore assessed whether C2GAPB-induced polarization could improve folic acid chemotaxis. Within 5 minutes of setting up a folic acid gradient, C2GAPB-expressing cells polarized and moved persistently towards the source, whereas non-expressing cells in the same population hardly moved. (Extended Data fig. 4 and Supplementary Video 17).

Ras attenuation at the cell rear leads to even stronger polarization

Previously, we showed that recruiting RASAL3 to the front in neutrophils extinguished mature protrusions and made the cells move away (Fig. 2k and l). When we continued to extinguish any new protrusions, the cell rounded up completely and did not move (Fig. 5a). However, once we stopped, the cell continued to move away persistently from the last recruitment area (Fig. 5a and b and Supplementary Video 18). Strikingly, RASAL3 simultaneously rearranged itself to the back region, presumably dragging the CIBN-CAAX membrane anchor with it. The previously non-polarized, migration-incompetent cell now had a stable back and a sustained front which allowed it to migrate (Fig. 5a and b).

Figure 5. Localization of recruited RASAL3 to the back of the neutrophil led to even stronger polarization.

(a) Time-lapse confocal images of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel). Unpolarized, non-migratory neutrophil, post-RASAL3 recruitment over entire periphery, caused cell to shrink and protrusions disappeared. Once laser was off, RASAL3 self-arranged to the back causing cell to polarize and migrate. Pink arrows highlight protrusions. Region of illumination is shown by dashed white box. Time in min:sec format. Scale bars represent 5 µm. (b) Color-coded (1-min intervals) outlines of cell in (a). (c) Time-lapse confocal images of neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after laser was switched on globally. Time in min:sec format. Scale bars represent 5 µm. (d) Color-coded (1-min intervals) outlines of cell in (c). (e) Representative kymograph of cortical LifeAct intensity before or after RASAL3 recruitment. Linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of kymograph is 24 mins. Cartoon depicts recruitment, F-actin polymerization or cell shape corresponding to kymograph. Box-and-whisker plots of (f) cell area, (i) cell speed, and (j) aspect ratio, before (black) or after (red) recruitment. n=16 cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (Two-sided Wilcoxon signed-rank test; f,i,j). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention. Connecting lines are provided between paired data points obtained from the same cell, before or after RASAL3 recruitment (GraphPad Prism 8). Centroid tracks (n=15 cells examined over 3 independent experiments) showing random motility before (g) or after (h) recruitment. Each track lasts 5 mins and was reset to same origin. (k) Time-lapse images of neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel) placed in a fMLP gradient before or after laser was switched on. Pink arrows denote uropod in recruited cell. Time in min:sec format. Scale bars represent 5 µm. (l) Centroid tracks of chemotaxing neutrophils (n=11 cells examined over 3 independent experiments) before or after RASAL3 recruitment. Each track lasts 2 mins and was reset to same origin. Normalized speed of cell shown in (k). Source numerical data are provided.

This spontaneous ‘back’ rearrangement of RASAL3 made it necessary to repeatedly extinguish new fronts to prevent cell movement, since when RASAL3 was globally recruited over the entire cell periphery, it rapidly localized to the rear and polarized the cell. This self-arranged back localization was dependent on cytoskeletal dynamics (Extended Data fig. 5a and b) and was mediated through C-terminal tail of RASAL3 (Extended Data fig. 5c–e). Due to this strong back localization, neutrophils developed long adhesive uropods at RASAL3-enriched regions, became highly polarized and migrated (Fig. 5c and d and Supplementary Video 19). Although F-actin levels increased at the front with RASAL3 recruitment, it did not appear to be as high when activated Ras/RasGEF was recruited (Fig. 5c and e)⁴. Across the population, RASAL3-recruited cells moved more extensively with a 1.4-fold increase in average speed. Recruitment also induced 1.44- and 1.26-fold improvement in cell area and polarity, respectively (Fig. 5f–j). The GAP domain fused to its C-terminal tail was responsible for these effects (Extended Data fig. 5f–h). No such change was observed when only the CRY2PHR component was recruited (Extended Data fig. 5i–k and Supplementary Video 20; also Fig. S3H–N in⁴). Similarly, when we recruited C2GAPB or RasG S17N ΔCAAX to the Dictyostelium rear, cells moved away (Extended Data fig. 6a–c). Adding to the generality of these findings, RASAL3 back recruitment polarized differentiated macrophages to migrate rapidly (Extended Data fig. 6d and Supplementary Video 21).

Next, we explored the role of RASAL3-mediated polarization in neutrophil chemotaxis. Cells moving along a fMLP gradient, upon RASAL3 recruitment, contracted and shrank immediately. After a few seconds, cells formed distinct tails, repolarized, and migrated in random directions, typically away from the attractant gradient (Fig. 5k and l and Supplementary Video 22). Eventually, these RASAL3-polarized cells turned and resumed migration towards the attractant source and moved similarly to unrecruited cells (Extended Data fig. 6e and f and Fig. 5l). In summary, Ras suppression improves chemotaxis to folic acid in Dictyostelium, while in neutrophils, it initially blocks chemotaxis but then has little additional effect. These differential effects may be explained by the fact that amoebae are not strongly polarized by folic acid so an increase in polarity is beneficial, whereas since neutrophils are strongly polarized by fMLP, a further increase in polarity has minimal effect.

RasGAP-induced polarization is accompanied by rear actomyosin contractility

We next assessed whether the RasGAP-induced protrusions required Arp2/3-mediated actin. We used Arp2/3 inhibitor, CK-666, which completely removed protrusions in neutrophils (Extended Data fig. 6g)⁴. When recruited RASAL3 localized to the back, interestingly, the CK-666-treated cells made long, thin protrusions which did not display LifeAct biosensor at the tips. Rather, the biosensor appeared to be along the lateral edges of these narrow protrusions (Extended Data fig. 6g and h). Similarly, long blebs were induced upon C2GAPB expression in CK-666-treated Dictyostelium cells, but not in control cells without C2GAPB (Extended Data fig. 6i and j). These data suggest that RasGAP-induced protrusions at the front are mediated, at least in part, by increased contraction at the rear.

To test whether contraction was mediated by actomyosin, we induced C2GAPB expression in myosin II heavy chain-null (mhcA^-) mutant of Dictyostelium. The mhcA^- cells are generally flattened, unpolarized, and multinucleated cells^45,46. C2GAPB expression had little effect on improving this existing phenotype (Fig. 6a). Furthermore, C2GAPB recruitment also failed to polarize mhcA^- cells (Fig. 6b). Similarly, in C2GAPB-recruited polarized cells, blebbistatin (myosin II inhibitor) treatment diminished polarity although it was not as effective as complete loss of myosin in mhcA^- cells (Extended Data fig. 7a)⁴⁷.

Figure 6. RasGAP-mediated polarization is accompanied by actomyosin contractility at the back.

(a) Time-lapse images of vegetative mhcA^- Dictyostelium expressing mRFPmars-C2GAPB post-doxycycline treatment, with color-coded (1-min interval) outlines. Time in min:sec format. Scale bars: 5 µm. (b) Time-lapse images of mhcA^- Dictyostelium expressing opto-C2GAPB, before or after recruitment. White and pink arrows denote cells of interest. Time in min:sec format. Scale bars: 5 µm. Immunofluorescence images of neutrophil expressing CRY2PHR-mCherry-RASAL3 (red) and LifeAct-miRFP703 (cyan), and stained with anti-phospho-MRLC2 Ser19 (c), anti-myosin IIA heavy chain (h), or anti-myosin IIB heavy chain (i), before or after recruitment. Pink arrows denote MRLC2 phosphorylation or myosin IIA/B enrichment. Scale bars represent 5 µm. (d) Bar graph showing back-front intensity ratio of F-actin and myosin phosphorylation in (c). Data are presented as mean values ± SD. n=11 cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (Two-sided Mann-Whitney test). Immunoblot comparing phospho-MRLC2 Ser19 (18 kDa; e) or myosin IIA/B (230 kDa, n=2 independent experiments; g) in opto-RASAL3-expressing population, before or after recruitment. GAPDH (36 kDa) was loading control. (f) Densitometric analysis of fold change of phospho-MRLC2 Ser19 level in population in (e). n=4 independent experiments; asterisks indicate significant difference, **P ≤ 0.0024 (Two-sided unpaired t-test). Data are presented as mean values ± SD. (j) Strategy for testing effects of blebbistatin on RASAL3-mediated actomyosin contraction and migration. (k) Time-lapse images of blebbistatin-treated neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after recruitment. Time in min:sec format. Scale bars represent 5 µm. Box-and-whisker plots of (l) cell speed, (o) cell area, and (p) aspect ratio, before (black) or after (red) recruitment. n=12 cells examined over 3 independent experiments; ‘ns’ indicates non-significant difference, P=0.3013 (l), P=0.1763 (o), P=0.2334 (p) (Two-sided Wilcoxon signed-rank test). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention. Connecting lines are provided between paired data points obtained from the same blebbistatin-treated cell, before or after RASAL3 recruitment (GraphPad Prism 8). Centroid tracks of blebbistatin-treated cells (n=11 cells examined over 3 independent experiments) showing motility before (m) or after (n) recruitment. Each track lasts 5 mins and was reset to same origin. Source numerical data and unprocessed Western blots are provided.

Is Ras suppression-mediated actomyosin contractility necessary for neutrophil polarization? Upon RASAL3 recruitment, phosphorylated myosin regulatory light chain 2 (MRLC2) was highly enriched at the uropod, across the cell population (Fig. 6c–f). Although myosin II expression remained unaltered, RASAL3 recruitment strongly localized myosin IIA and IIB isoforms to the cell rear (Fig. 6g–i). Additionally, treatment with blebbistatin resulted in complete loss of cell movement and polarity which could not be recovered with RASAL3 recruitment (Fig. 6j–p and Extended Data fig. 7b and Supplementary Video 23)⁴⁸. Interestingly, in presence of a sub-optimal dose of blebbistatin, once RASAL3 was recruited, it moved to a particular region which became the new back. This local Ras suppression could overcome low blebbistatin inhibition, increased contraction, and polarized cells to move (Extended Data fig. 7c–k and Supplementary Video 24).

Treating neutrophils with low doses of ROCK inhibitor, Y27632⁴⁹, reduced myosin II phosphorylation but did not completely abolish it. Although basal motility was stalled, RASAL3 recruitment to the back mostly overcame the effects. RASAL3-recruited cells polarized and migrated due to increased myosin phosphorylation and rear contraction (Extended Data fig. 8a–f, Supplementary Video 25). However, once treated with a higher Y27632 dose⁵⁰ which abolished myosin phosphorylation completely, RASAL3 back recruitment was unable to polarize cells (Extended Data fig. 8l–u, Supplementary Video 26). Altogether, ROCK-mediated myosin IIA and IIB activity generates contraction at the neutrophil rear in response to Ras attenuation. Additionally, the effects of Ras suppression on actomyosin in Dictyostelium and neutrophils are similar.

RASAL3 induces mTORC2-mediated actin polymerization at the front

Back localization of recruited RASAL3 induced F-actin to localize to a stable front which caused neutrophils to polarize and migrate persistently (Fig. 5c–j). To confirm that RASAL3 at the back is directly affecting protrusive activity at the front, we pharmacologically targeted front signaling pathways before recruiting RASAL3. First, we inhibited PI3K/PIP3 signaling using pan-PI3K inhibitor, LY294002, which stalled neutrophil polarity and basal motility (Fig. 7a and b)⁴. Within 5 minutes of applying blue light, RASAL3 moved to the back of LY294002-treated cells and polarized them to move by generating broad F-actin-rich lamellipodium but without an appreciable uropod (Fig. 7b and c, f and g; Supplementary Video 27). Kymograph showed that LifeAct signal increased considerably along with increase in cell area (Fig. 7d). RASAL3 recruitment caused 1.75-, 1.49-, or 1.43-fold improvement in cell speed, basal area or aspect ratio (Fig. 7e, h and i). Similarly, RASAL3 recruitment polarized PI3Kγ inhibitor-treated cells for migration (Extended Data fig. 9 and Supplementary Video 28)^4,51.

Figure 7. RASAL3 polarizes cells by localizing actin polymerization at the front through mTORC2.

Strategy for testing effects of (a) pan-PI3K inhibitor, LY294002, or (j) mTOR inhibitor, PP242 on RASAL3-directed actin polymerization and motility. Time-lapse confocal images of (b) LY294002- or (k) PP242-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after 488 nm laser was turned on globally. Time in min:sec format. Scale bars represent 5 µm. (c, l) Color-coded (1-min intervals) outlines of cells in (b) and (k), respectively. Representative kymographs of cortical LifeAct intensity in (d) LY294002- or (m) PP242-treated RASAL3-expressing neutrophil before or after laser was turned on. A linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of the kymographs are 12 and 20 mins, respectively. Cartoons depict recruitment, actin polymerization or cell shape corresponding to the kymographs. Box-and-whisker plots of (e or n) cell speed, (h or q) cell area, and (i or r) aspect ratio, before (black) and after (red) RASAL3 recruitment in LY294002- or PP242-treated cells. (e,h,i,n,q,r) n=13 LY294002- or PP242-treated cells examined over 3 independent experiments; asterisks indicate significant difference, ***P=0.0002 (e,h,i), **P=0.0081 (n); ns denotes not significant, P=0.0574 (q), P=0.5417 (r) (Two-sided Wilcoxon signed-rank test). The boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention. Connecting lines are provided between paired data points obtained from the same LY294002- or PP242-treated cell, before or after RASAL3 recruitment (GraphPad Prism 8). Centroid tracks of LY294002- or PP242-treated neutrophils (n=12 cells examined over 3 independent experiments) showing random motility before (f or o) or after (g or p) recruitment. Each track lasts 5 mins and was reset to same origin. Source numerical data are provided.

Next, we inhibited mTORC2 signaling with PP242 (Fig. 7j)⁵². Inhibitor treatment caused cells to round up (Fig. 7k)^4,53. Once RASAL3 was recruited to the back, PP242-treated cells displayed weak, intermittent polarity causing them to move slightly, but not persistently (Fig. 7k, l, o and p and Supplementary Video 29). Although we noticed a 1.3-fold improvement in average speed, we did not see significant change in cell area or polarity (Fig. 7m and n, q and r).

Model can explain counterintuitive effects of reducing Ras activity on cell polarization

To understand the role of counterintuitive effects of the RasGAPs, we turned to a legacy model of the excitable network that has successfully recreated cell migratory and ventral wave behavior^37,54–56 and incorporated the new information from this experimental study (Fig. 8a). In the model, complementary inhibition between Ras and PIP2 acts as a positive feedback loop^11,57 and a competing negative feedback loop involves Ras and PKB^37,57 (Fig. 8a). These three elements form the core of the Signal Transduction Excitable Network (STEN). We incorporated two further cytoskeleton-dependent feedbacks to capture the polarization that is observed over time^54,58. There is a front-promoting loop from PKB through branched actin to Ras. There is a back-promoting loop that involves PIP2 and myosin which increases RasGAP activity and, in turn, inhibits Ras. Model simulations displayed characteristic excitable behavior, including propagating waves that annihilate when they meet (Fig. 8b and Supplementary Video 30). We next modelled the effect of increasing/decreasing RasGAP activity as changes in inhibition rate of Ras activity, thus altering the activation threshold of the excitable system (Fig. 8c and Extended Data fig. 10a–c). Whereas simulations in which RasGAP was reduced showed increased Ras activity and resulted in large, fast waves compared to WT levels of RasGAP, those with increased RasGAP showed small, slow waves (Fig. 8b and Extended Data fig. 10d; Supplementary Video 30). The 2D simulations described above model the basal surface of electrofused cells. Experimentally, we only increased RasGAP which reduced Ras activity and wave size or speed, as did the model (Fig 8b). These observations were quantitated by summing all Ras activity throughout the simulations (Fig. 8d). In simulations of a single cell, in which activity is measured along the perimeter (Fig. 8e), we saw high spontaneous activity all around the perimeter before changes in RasGAP level. Halfway through the simulation, RasGAP activity was increased leading to an overall decrease in Ras activity. Thereafter, single streaks representing one or two stationary waves appeared in the kymograph (Fig. 8e). A heatmap of Ras and PKB activity showed much lower activity after GAP addition (Fig. 8f and Extended Data fig. 10e). Simulations also considered the effect of GAP addition on cell movement (see Methods). Without RasGAP, cells showed high activity but moved minimally from their initial location. After GAP addition, the kymograph showed less activity, but greater polarization and the trajectories displayed greater dispersion (Fig. 8g, Extended Data fig. 10f and g). In the kymograph of actomyosin response, we saw multiple streaks of myosin activity along the perimeter prior to GAP addition. However, post-GAP addition, myosin was uniform except for the single cell front (Fig. 8h and Extended Data fig. 8h).

Figure 8. STEN simulations show that increasing RasGAP reduces Ras activity, but cells still polarize.

(a) Schematic of signal transduction excitable network and feedback loops regulating cell polarity. GAP hydrolyses active RasGTP to inactive RasGDP. GAP is part of the back-mediated loop because this inhibition increases Ras inactivation rate. Based on many experimental observations, the interactions shown summarize the invariable direction of movement of the activities and concentrations of components and are not meant to convey direct interactions, and are representative of a series of back effectors which behave similarly. (b) With WT RasGAP levels, EN firings were seen and generated waves of high Ras activity (red) moving at ~10 µm/min and absence of PIP2 (green). Wave trailing edge was marked by high PKB (blue). Whereas lowering RasGAP by 20% increased wave size (~50%) and speed (~100%), a 10% increase had the opposite effect: firings were still seen, but waves broke up frequently and were smaller (~75%) and slower (~40%). The simulation denotes a square where each side is 40 µm; time is in seconds. (c) Cartoon of the phase-plane diagram of the Ras-PKB EN showing curves on which Ras and PKB stay constant, the equilibrium (black circle) and activation threshold of two EN systems, representing WT and lowered GAP conditions. The latter has a larger threshold. (d) Total Ras level over the simulations (n=5 for each condition) for varying RasGAP from 80–140% of WT levels. After ~30% increase in RasGAP, no firings were observed and hence the Ras level plateaus. (e) Kymograph of one-dimensional simulations representing perimeter of smaller cells represented by an environment in which RasGAP strength was increased 50% halfway during the simulation. (f) 2D histograms showing total activity across cell perimeter in the Ras-PKB phase plane for 70 seconds each for WT conditions and after increasing GAP by 50% (n=10). (g) Trajectory of the center-of-mass of a single cell across a 100-second time interval before and after GAP recruitment. (h) Kymographs of normalized myosin species corresponding to the cell in (e). Myosin appears in high PIP2 regions showing a positive correlation between the two species.

Discussion

Our optogenetic study on RasGAPs, C2GAPB and RASAL3, expressed in Dictyostelium and leukocytes, respectively, yielded significant insights. Notably, locally dampening Ras activity could extinguish cellular protrusions, reverse pre-existing polarity, and impede migration. However, contrary to expectations, a global reduction in Ras activity heightened both polarity and motility. Furthermore, in Dictyostelium, targeting C2GAPB to the cell back amplified migration and polarity, whereas in neutrophils and macrophages, spontaneous movement of RASAL3 to the back generated uropods and localized F-actin protrusions at the front. Thus, polarity could be regulated by manipulating Ras activity along the membrane and, therefore, a cell must maintain an optimum Ras level to achieve directed migration. Although Ras and its downstream pathways are typically associated with longer-term growth control, our studies establish an immediate and evolutionarily conserved role in cell polarity and motility.

The counterintuitive effect of Ras suppression in promoting polarity can be understood by including a ‘back-promoting’ loop to our pre-existing model. Ras plays a central role in regulating two strong opposing feedback loops consisting of molecules, such as myosin and PI(4,5)P2 that typically form the back, and PIP3 and F-actin that generate the front (Fig. 8)^{8,11,18,22,23}. Our observation on the effects on RBD/PHcrac wave activity with RasGAP strongly suggests that reducing Ras activity raises the activation threshold of the system. The increased threshold causes the broad, propagating cortical waves to break up into smaller, short-lived waves in electrofused cells. Analogously, in single cells, RasGAPs suppress multiple, co-existing protrusions all around the periphery and confine them to a single protrusion. Obviously, if all protrusions were shut down that would block migration, which we were able to achieve in the model. However, this was not accomplished experimentally, presumably because our RasGAPs were not strong enough to raise the activation threshold enough.

Previous studies have suggested that increases in actomyosin contraction at the back, occurring spontaneously or locally triggered by optogenetic RhoA or RGS4, can initiate polarity and cause cells to move away^50,59–62. Contraction-driven migration also occurs through Ras/ERK-mediated myosin light chain kinase activation and subsequent MRLC phosphorylation⁶³. We demonstrate that spatiotemporal regulation of Ras signaling pathways not only controls F-actin protrusive activity at the front, but also directly coordinates non-muscle myosin II-dependent contraction at the back. In neutrophils, contraction possibly works through ROCK-mediated MRLC2 phosphorylation. Even when Arp2/3 was inhibited, the ability of back suppression to polarize a cell was indicated. Without F-actin, instead of making normal protrusions, suppressing Ras activity at the back caused formation of long bleb-like structures.

In addition to increasing contractility at the back, Ras inhibition at the back triggered increased F-actin polymerization at the front. We noted that while inhibitors of signaling downstream of Ras suppressed protrusions and stopped cells from moving, this could be partially overcome by recruiting RasGAP to the rear. These RasGAP-mediated long-range effects on the cell front were mediated primarily through mTORC2, rather than PI3K/PIP3. Traditionally, these possibilities have been difficult to distinguish since conventional chemotactic gradient studies lack sufficient spatiotemporal resolution to determine chronology of front and back formation during symmetry breaking^60,62,64. However, we showed that back suppression triggers polarization which simultaneously activates the front.

Our findings have important implications for cancer treatment, where targeting Ras may not always be beneficial. Caution must be exercised to avoid, while attempting to abrogate cell proliferation, forcing cells into a more polarized migratory state. The increased migration could induce cells to exit epithelium and metastasize. Consistently, it is well-known that metastasizing cells do not readily divide^65–67. A deeper understanding of different roles that Ras play in cell migration versus growth is essential for developing therapeutic strategies.

Methods

Reagents and inhibitors

200 µg/mL fibronectin stock (Sigma-Aldrich; F4759-2MG) was prepared in sterile water, followed by dilution in PBS. Folic acid (Sigma-Aldrich; #329823065) was dissolved in sterile water, with help of 2M NaOH, to prepare a 1.25 mM stock solution. N-Formyl-Met-Leu-Phe (fMLP, Sigma-Aldrich; #47729) was dissolved in DMSO (Sigma Aldrich; #D2650) to make a stock solution of 10 mM. Phorbol 12-myristate 13-acetate (PMA, Sigma-Aldrich; #P8139) was dissolved in DMSO to make a 1 mM stock solution. 20 mM AS605240 (Sigma-Aldrich; #A0233), 20 mM PP242 (EMD Millipore; #475988), 50 mM LY294002 (Thermo Fisher; #PHZ1144), 50 mM CK-666 (EMD Millipore; #182515), 50 mM blebbistatin (Peprotech; #8567182), 5 mM latrunculin A (Enzo Life Sciences; BML-T119-0100), 10 mM latrunculin B (Sigma-Aldrich; #428020), or 5 mM Y27632 (Sigma-Aldrich; #688001) stock solution was made in DMSO. Jasplakinolide (Sigma-Aldrich; #420127) was available as a ready-made 1 mM stock. Hygromycin B (Thermo Fisher Scientific; #10687010) or G418 sulphate (Thermo Fisher Scientific; #10131035) was purchased as 50 mg/mL stock solution whereas blasticidine S (Sigma-Aldrich; #15205) or puromycin (Sigma-Aldrich; #P8833) was dissolved in sterile water to make stock solutions of 10 mg/mL or 2.5 mg/mL, respectively. Doxycycline hyclate (Sigma; #D9891-1G) was dissolved in sterile water to make a stock of 5 mg/mL. 50 mg/mL TRITC-dextran (Sigma-Aldrich; #T1162) was made in sterile water. All stock solutions were aliquoted and stored at −20℃. According to experimental requirements, further dilutions were made in development buffer (DB), PBS, or growth medium before adding to cells.

Plasmid construction

All DNA oligonucleotides were purchased from Sigma-Aldrich and are listed in Supplementary Table 1. Dictyostelium C2GAPB (RasGAP2) gene was cloned in KF2 expression plasmid¹¹. RasG S17N-expressing plasmid was obtained from dictyBase (#254). Using these constructs, we subcloned C2GAPB or RasG S17N into doxycycline-inducible pDM335 plasmid (dictyBase #523) using BglII/SpeI restriction digestion to generate mRFPmars-C2GAPB/pDM335, GFP-C2GAPB/pDM335, and mRFPmars-RasG S17N/pDM335 constructs. SspB R73Q ORF was amplified from tgRFPt-SspB R73Q plasmid (Addgene #60416, RRID: Addgene_60416) and then subcloned into C2GAPB/pDM335 at the BglII site to generate the opto-C2GAPB construct, mRFPmars-SspB R73Q-C2GAPB⁶⁸. tgRFPt-SspB R73Q ORF was introduced into pCV5 vector to generate tgRFPt-SspB R73Q-Ctrl/pCV5 construct (Addgene #201761)⁴. CAAX-deleted RasG S17N ORF was introduced to opto-Ctrl construct by restriction digestion to generate opto-RasG S17N ΔCAAX. Similarly, tgRFPt-SspB R73Q ORF was introduced into pDM335 to generate tgRFPt-SspB R73Q-Ctrl construct. N150-Venus-iLID/pDM358 construct was made previously⁴. This was used to subclone PHcrac-YFP, LimE_∆coil-YFP, or RBD-YFP ORF to generate dual expressing N150-Venus-iLID/PHcrac-YFP, N150-Venus-iLID/LimE-YFP, or N150-Venus-iLID/RBD-YFP construct, respectively. The shuttle vector, pDM344 (dictyBase #551), was used for this purpose. Constructs for GFP-RBD and GFP- LimE_∆coil were procured from Firtel lab (UCSD) and Marriott lab (University of Wisconsin-Madison), respectively, whereas myosin-GFP/pDRH was obtained from Robinson lab (School of Medicine, JHU)^5,69,70.

CRY2PHR-mCherry-RASAL3/pPB (Addgene #201755), CIBN-CAAX/pLJM1 (Addgene #201749) or LifeAct-miRFP703/pLJM1 (Addgene #201750) construct was made previously⁴. DNA sequences (2064–3036 or 1276–3036 bases) encoding last 322 or 585 amino acids of RASAL3, and CAAX-deleted KRas4B S17N (KRas4B S17N ΔCAAX) were PCR-amplified (from Addgene #201755 or #83156, RRID:Addgene_83156) and cloned into BspEI/NotI sites of PiggyBac™ transposon plasmid to generate CRY2PHR-mCherry-RASAL3_689–1011/pPB, CRY2PHR-mCherry-RASAL3_426–1011/pPB or CRY2PHR-mCherry-KRas4B S17N ΔCAAX/pPB construct, respectively (PiggyBac™ transposon system was gifted by Collins lab, UC Davis)^71,72. Constructs were verified by diagnostic restriction digestion and sequenced at the JHMI Synthesis and Sequencing Facility.

Cell culture

Wild-type Dictyostelium discoideum cells of the AX2 strain (dictyBase #DBS0235521) was obtained from the Kay lab (MRC Laboratory of Molecular Biology, UK). Gβ-null (Gβ^-) cells were created in our lab previously⁴¹. Myosin heavy chain-null strain (mhcA^-) was obtained from Robinson lab (School of Medicine, JHU)⁷³. C2GAPB-null strain (C2GAPB^-) was generated in our lab¹¹. All lines were cultured axenically in HL5 medium (lab stock) at 22°C. Growth-stage cells were used for imaging⁷⁴.

Human HL-60 cell line (ATCC #CCL-240; RRID:CVCL_0002) was obtained from Weiner lab (UCSF) and cultured in RPMI 1640 medium (Gibco #22400–089) supplemented with 15% heat-inactivated fetal bovine serum (FBS; Thermo Fisher #16140071)^4,75. To obtain migration-competent neutrophils, wildtype or stable lines were differentiated in presence of 1.3% DMSO over 5–7 days^4,75. Differentiated cells are an effective model to study human neutrophils⁷⁶. To differentiate HL-60 cells into macrophages, cells were incubated with 32 nM PMA for 48–72 hours^77,78. Cells were grown in humidified conditions at 5% CO₂ and 37⁰C.

Electroporation

Dictyostelium stable lines were generated by electroporating 2 µg DNA in 10⁷ cells using chilled 0.1-cm cuvette (BioRad; #1652089) at 0.85 kV/25 µF twice with 5-sec interval, followed by antibiotic selection over 3–4 weeks^4,79. HL-60 cell line stably co-expressing CIBN-CAAX, opto-RASAL3, and LifeAct-miRFP703 was generated previously^4,79. Here, we introduced 5 µg opto-RASAL3_689–1011, -RASAL3_426–1011, or -KRas4B S17N ΔCAAX construct with 5 µg piggyBac transposase plasmid in 2×10⁶ CIBN-CAAX- and LifeAct-miRFP703-expressing HL-60 stable cells using Neon™ transfection system 100 µL kit (Thermo Fisher #MPK10025). Dictyostelium cells, at a density of 1.5×10⁷ cells/mL, were rolled for 30 mins and electroporated in 4-mm cuvette (BioRad; #1652088) at 1,000 V/3 µF once followed by 1,000 V/1 µF thrice with 2-sec interval to generate electrofused cells^35,36.

Confocal microscopy

Vegetative Dictyostelium or differentiated macrophages were adhered on 8-well coverslip chamber for 40 min. Differentiated neutrophils, pre-treated with heat-killed Klebsiella aerogenes, were adhered to fibronectin-coated chambers for 40 mins^4,79. Next, fresh DB or RPMI 1640 medium was added to attached cells and used for imaging. To induce C2GAPB expression in Dictyostelium, doxycycline (50 µg/mL) was added 8 hr prior to imaging. For macropinocytosis assay, cells were incubated with TRITC-dextran for 4 mins, washed, and imaged subsequently^33,80. All live- or fixed-cell imaging was acquired with 0.3–0.5% (Dictyostelium) or 0.2–10% (HL-60) laser intensity using following microscopes: (1) Zeiss LSM780-FCS single-point, laser scanning confocal microscope with 780-Quasar; 34-channel spectral, high-sensitivity gallium arsenide phosphide detectors supported with ZEN Black software, and (2) Zeiss LSM800 GaAsP single-point laser scanning confocal microscope with wide-field camera supported with ZEN Blue software. All images were acquired with either 63X/1.40 PlanApo oil or 40X/1.30 PlanNeofluar oil DIC objective, along with digital zoom. In single cells, confocal imaging was performed at a middle plane of the cells, whereas in electrofused cells, laser was focused near the bottom surface to visualize cortical waves^4,35,79,81. For inhibitor experiments, neutrophils were treated with 20 µM AS605240, 20 µM PP242, 50 µM LY294002, 50 µM CK-666, 20 or 75 µM blebbistatin, and 10 or 50 µM Y27632 for atleast 10 mins before imaging. For pre-treatment with JLY cocktail, neutrophils were incubated with 10 μM Y27632 for 10 min. Cells were then treated with 8 μM jasplakinolide and 5 μM latrunculin B without changing the final concentration of Y27632^82,83. In Dictyostelium, 50 µM blebbistatin or CK-666 was added during imaging.

Optogenetics

Optogenetic experiments with vegetative Dictyostelium or differentiated HL-60 cells were done in absence of chemoattractant, except for chemotaxis assays. Throughout image acquisition, solid state laser (561 nm excitation and 579–632 nm emission) was used for visualizing proteins or recruitable effectors fused to mCherry, mRFPmars, or tgRFPt tag whereas a diode laser (633 nm excitation and 659–709 nm emission) was used to capture miRFP703 expression. Images were acquired for 5–10 mins, after which 450/488 nm excitation laser was switched on globally to activate recruitment. Image acquisition and photoactivation was done at ~7 sec intervals. Using the T-PMT associated with the red channel, we acquired DIC images. The interactive photobleaching module on Zeiss LSM800 was used to perform local recruitment experiments. A small ROI was placed in front or back of migrating cells, and bleached with 488 nm laser (laser power of ~0.5% or 7% for Dictyostelium or HL-60 cells, respectively) in multiple iteration. Time interval of photoactivation and image acquisition was ~10 secs^{4,35,79,81,84}.

Indirect immunofluorescence

Differentiated opto-RASAL3 expressing HL-60 neutrophils were attached on 8-well coverslip chamber and illuminated with 6×10 second pulses of 420–480 nm LED using a hand-held lamp (#LY-A180). Immediately, non-illuminated and illuminated samples were incubated with a fixative [3.7% buffered paraformaldehyde (Fisher Scientific; #T353-500), 0.25% glutaraldehyde (Sigma-Aldrich; #340855), 0.1% Triton X-100 (Sigma-Aldrich; #T8787) in serum-supplemented RPMI 1640] for 10 mins at RT. Following quenching with sodium borohydride (Fisher Scientific; #S678-10), samples were blocked with 3% BSA (Sigma Aldrich; #A-9647) for 1 hour at RT. Endogenous phospho-myosin regulatory light chain 2 (pMRLC2 Ser19), myosin heavy chain IIA (myosin IIA) or myosin heavy chain IIB (myosin IIB) localization was detected by incubating them for 1 hour at RT with respective antibodies: mouse anti-pMRLC2 Ser19 (1:200, Cell Signaling; #3675, RRID:AB_2250969), rabbit anti-myosin IIA (1:100, Sigma-Aldrich; #M8064, RRID:AB_260673) or rabbit anti-myosin IIB antibody (1:50, Sigma-Aldrich; #M7939, RRID:AB_260669). Finally, samples were incubated with Alexa Fluor 488 goat anti-mouse (Thermo Fisher; #A11001, RRID:AB_2534069) or anti-rabbit (Thermo Fisher; #A11008, RRID:AB_143165) antibody (1:1000) for 40 mins at RT, before imaging.

2D chemotaxis assay

Chemotaxis assays with doxycycline pre-treated, vegetative mRFPmars-C2GAPB-expressing Dictyostelium or differentiated opto-RASAL3 expressing neutrophils were performed on µ-Slide Chemotaxis Collagen IV chamber (Ibidi; #80322) using freshly-prepared 10 nM fMLP or 100 nM folic acid as chemoattractant⁴.

SDS-PAGE and Western blotting

Wildtype or mRFPmars-C2GAPB-expressing Dictyostelium was developed and 6×10⁶ cells were collected every hour, upto 8 hours, for western blot analysis^79,85. Samples were resuspended and lysed in pre-chilled RIPA buffer (supplemented with protease inhibitor cocktail; Thermo Scientific, # 89900) for 20 mins, and then incubated with 3× Laemmili sample buffer (lab stock) at RT for 5 mins. For neutrophils, 10⁷ differentiated opto-RASAL3 expressing cells were resuspended in 1 mL growth medium. For inhibitor experiments, cell suspension was incubated with 10 or 50 µM Y27632 for 10 mins. Next, samples were illuminated with a 420–480 nm LED lamp. Non-illuminated and untreated or Y27632-treated samples were maintained throughout as control. Samples were lysed by boiling in sample buffer.

Sample equivalent to 1.5×10⁶ cells were loaded into pre-cast 4–15% polyacrylamide gel and immunoblotting was performed as per lab protocol^4,79. Endogenous cAR1 (~44 kDa), phospho-MRLC2 Ser19 (18 kDa), MRLC2 (18 kDa), myosin IIA (230 kDa), myosin IIB (230 kDa), or GAPDH (36 kDa) expression was detected by incubating PVDF membrane (Bio-Rad; 162–0262) with rabbit anti-cAR1 (1:1000; generated in our lab⁸⁶), mouse anti-pMRLC2 Ser19 (1:1000), rabbit anti-MRLC2 (1:1000, Cell Signaling; #3672, RRID:AB_10692513), rabbit anti-myosin IIA (1:1000), rabbit anti-myosin IIB (1:200), or rabbit anti-GAPDH antibody (1:3000, Thermo Fisher; #PA1-987, RRID:AB_2107311) overnight at 4⁰C, followed by goat anti-rabbit IRDye 680RD-conjugated (1:10,000; Li-Cor; #925–68071, RRID:AB_2721181) or goat anti-mouse IRDye 800CW-conjugated (1:10,000; Li-Cor; #925–32210, RRID:AB_2687825) secondary antibody for 1 hour in dark. Odyssey CLx imaging system (Li-Cor) detected near-infrared signal from blots.

Image analysis

All images were analyzed with Fiji/ImageJ 1.52i (NIH), Python 3.10, and MATLAB 2019b (MathWorks, Natick, MA) software⁸⁷. We utilized GraphPad Prism 8 (GraphPad software, CA, USA), OriginPro 9.0 (Originlab Corporation, MA, USA) and Microsoft Excel (Microsoft, WA, USA) for plotting our results^4,35,79.

To get the ratio of wave area to cell area in Fig. 1e and h, images were first binarized using ImageJ, by adjusting threshold to cover all pixels of the wave or cell. The range was not reset and the ‘Calculate threshold for each image’ option was unchecked. Subsequently, using the “Analyze Particle” function, a size-based thresholding was applied (to exclude non-cell particles) and cell masks were generated. Next, ‘Fill holes’, ‘Erode’ and ‘Dilate’ options were applied, sequentially and judiciously, to obtain the proper binarized mask for waves or cells^4,35. Ratio of wave area to cell area for each frame was obtained and plotted with time. Duration of waves for Fig. 1f and i was obtained by counting number of frames from when a wave starts to when it ends, and then multiplying it with the time interval.

Cell outline overlays (Figs. 3c, 4b and f, 5b and d, 6a, 7c and i, and Extended Data figs. 1i and j, 2d and e,3g, 5j, 7e, 8d and n, 9c) were obtained from segmented cells^4,35. For membrane kymographs (Figs. 3i and j, m and n, 5e, 7d and m, and Extended Data figs. 3h, 5k, 7f, 8e and o, 9d), cells were segmented against the background following standard image-processing steps with custom code written in MATLAB^4,35,79. Next, kymographs were created from segmented cells using a custom-written MATLAB function⁵⁸. For linescan analysis (Fig. 3h and l), a straight-line segment (5-pixel width) was drawn using ‘Straight’ tool in ImageJ^35,79.

Local protrusion (Fig. 2d,f,l and Extended Data fig. 1h) and cell migration (Figs. 5f–j, l, 6l–p, 7e–i, n–r, Extended Data figs. 3i–m, 4b and c, 6f, 7g–k, 8g–k, q–u, 9e–i) analyses were performed as detailed previously^4,35,79. For cell speed, area and aspect ratio quantifications (Figs. 3d–f, 4c and e, g and h, 3b–e), cells were segmented against the background following standard image-processing steps with custom code written in Python based on package scikit-image, Trackpy v0.6.2 and PyImageJ 1.4.1. Human supervision was involved during the process with an integration to Fiji, and we obtained the X-position, Y-position, cell area, and aspect ratio for each frame of each cell. Distance between adjacent frames was calculated from X-position and Y-position, and was divided by time interval to obtain an instantaneous velocity. The average cell speeds were calculated by taking an average of all instantaneous velocities. For Extended Data fig 1b, macropinocytosis uptake was analyzed by outlining each cell and quantifying the total TRITC fluorescence signal within each outline divided by the cell area⁸⁰. For Extended Data fig 6h, numbers of blebs for each cell, before or after recruitment, were quantified by summing them over a min⁸⁸.

Back/front intensity ratios for F-actin and pMRLC2 Ser19 in fixed samples (Fig. 6d) were quantified in Image J. Using ‘freehand’ tool, a ROI was outlined at the cell back (488 nm channel) and signal intensity for Alexa Fluor 488 (pMRLC2) was measured. Next, this ROI was added to ‘ROI Manager’ and the same region in the 633 nm channel was outlined to measure the corresponding LifeAct-miRFP703 (F-actin) signal. Similarly, ROIs were drawn at the front of a cell in both channels and signal intensities for F-actin and pMRLC2 were quantified. Using ‘back’ and ‘front’ intensity values, graphs were generated. Protein bands between different immunoblot samples were quantified by densitometric analysis using ImageJ (Fig. 6e and f).

Simulations

The simulations are based on a model, previously described, in which three interacting species, RasGTP, PIP2, and PKB, form an excitable network³⁷. In this model, Ras and PIP2 are complementary – wherever Ras activity is high, PIP2 is low and vice versa. This is achieved through mutually inhibitory connections which give rise to the traditional positive feedback loop seen in the canonical excitable networks. In our model, PKB acts as the traditional refractory species: it is activated by Ras but provides slow, negative feedback to the latter. The network connections are shown in Extended Data fig. 10a. The concentrations of each of these molecules is described by stochastic, reaction-diffusion partial differential equations:

$$\frac{\partial \left[\text{Ras}\right]}{\partial t}=-\left(a_1+a_2\left[\text{PKB}\right]\right)\left[\text{Ras}\right]+\frac{a_3}{1+a_4^2{\left[\text{PIP}2\right]}^2}+a_5+w_{\text{Ras}}+D_{\text{Ras}}{\nabla }^2\left[\text{Ras}\right]$$

$$\frac{\partial \left[\text{PIP}2\right]}{\partial t}=-\left(b_1+b_2\left[\text{Ras}\right]\right)\left[\text{PIP}2\right]+b_3+w_{\text{PIP}2}+D_{\text{PIP}2}{\nabla }^2\left[\text{PIP}2\right]$$

$$\frac{\partial \left[\text{PKB}\right]}{\partial t}=-c_1\left[\text{PKB}\right]+c_2\left[\text{Ras}\right]+w_{\text{PKB}}+D_{\text{PKB}}{\nabla }^2\left[\text{PKB}\right]$$

In each of these equations, the final term represents the diffusion of the species, where $D_{*}$ is the respective diffusion coefficient and ${\nabla }^2$ is the spatial Laplacian (in one or two dimensions). The second-to-last terms represent molecular noise. Our model assumes a Langevin approximation in which the size of the noise is based on the reaction terms⁸⁹. For example, in the case of PKB, the noise is given by

$$w_{\text{PKB}}\left(t\right)=\alpha \sqrt{c_1\left[\text{PKB}\right]+c_2\left[\text{Ras}\right]}w\left(t\right)$$

where $w\left(t\right)$ is a zero-mean, unit-variance Gaussian, Brown noise process. In the simulations, the size of this noise was adjusted with the empirical parameter $\alpha$.

In addition to the EN dynamics described above, we incorporated two other terms related to cell polarization^37,54. These feedback loops come from $\text{PIP2}$ and $\text{PKB}$ and denote actions at the front $\left(P_F\right)$ and rear $\left(P_R\right)$ respectively

$$\frac{\partial \left[P_F\right]}{\partial t}=-p_1\left[P_F\right]+p_2\left[\text{PKB}\right]+w_{P_F}+D_{P_F}{\nabla }^2\left[P_F\right]$$

$$\frac{\partial [P_B]}{\partial t}=-p_3\left[P_B\right]+p_4\left[\text{PIP}2\right]+w_{P_B}+D_{P_B}{\nabla }^2\left[P_B\right]$$

Both these terms modify the terms in the RasGTP equation related to hydrolysis:

$$\left(a_1+a_2\left[\text{PKB}\right]\right)\left[\text{Ras}\right]\mapsto \frac{\left(a_1+a_2\left[\text{PKB}\right]+a_{P_B}\left[P_B\right]\right)\left[\text{Ras}\right]}{1+a_{P_F}\left[P_F\right]}$$

having the effect of increasing front and back contributions. Lastly, we note that changes in RasGAP contributions are modeled as changes in the parameters $a_1$ and $a_2$.

Parameter values, listed in Supplementary Table 2, were obtained to recreate experimental observations on the level of excitable activities (i.e., number of firings per unit time, wave propagation, etc.).

The system is largely robust to parameter changes, in that variations from the nominal values below continue to give rise to excitable behavior. To test this, we note that excitable behavior arises from the stochastic crossing of a threshold that is determined by the nullclines of the differential equations (Extended Data fig. 10b). To measure the level of robustness, we varied each of the parameters $a_1,\dots ,\hspace{0.33em}{a}_5,b_1,\dots b_3,c_1$ and $c_2$ from their nominal values and computed the size of the threshold (Extended Data fig. 10c). Some, like $a_2$, $b_2$ and $c_1$ are quite robust, allowing large changes in either direction. Others allow large changes only in one direction.

To simulate the movement of cells, we followed the procedure outlined previously⁹⁰. Briefly, Ras activity along the perimeter of the cell (in a one-dimensional simulation) was thresholded to generate a force normal to the cell surface. The vector sum of these forces around the perimeter were used to generate a net force scaled so that it was in the range of experimentally observed protrusive stresses 0.5–5 nN/μm, and then fed into a viscoelastic model⁹¹ of Dictyostelium mechanics:

$$\ddot{x}+\left(\frac{k_c}{{\gamma }_c}\right)\dot{x}=\left(\frac{1}{{\gamma }_c}+\frac{1}{{\gamma }_a}\right){\dot{\sigma }}_x+\left(\frac{k_c}{{\gamma }_c}\right){\sigma }_x$$

Here, ${\sigma }_x$ is the $x$-component of the stress and the viscoelastic parameters are ${\gamma }_a=6.09\frac{\text{nNs}}{{\text{μm}}^3}$, ${\gamma }_c=0.064\hspace{0.33em}\text{nNs}/{\text{μm}}^3$, and $k_c=0.098\hspace{0.33em}\text{nN}/{\text{μm}}^3$. A similar formula is used to denote the displacement in the $y$-direction.

Simulations were run on MATLAB 2023a (Mathworks, Natick, MA) on custom-code based on the Itô solution in the Stochastic Differential Equation toolbox (http://sdetoolbox.sourceforge.net). Two-dimensional simulations were used to recreate the observed wave patterns of larger electrofused cells, and so assume a grid 40 µm × 40 µm with a spacing of 0.4 µm × 0.4 µm per grid point (i.e., 100 × 100 points) and zero flux boundary conditions. The one-dimensional simulations aim to recreate the membrane fluorescence observed in single-cell confocal images. The dimension is therefore smaller, assuming a cell radius of 5 µm and a spacing of 0.25 µm, resulting in $2\pi \times 5/0.25\approx 126$ points along the perimeter, and periodic boundary conditions.

Statistics & Reproducibility

Statistical analyses were executed using unpaired or paired 2-tailed non-parametric tests on GraphPad Prism 8. Results are expressed as mean ± SD from at least 3 independent experiments. ns denotes P>0.05, * denotes P ≤ 0.05, ** denotes P ≤ 0.01, *** denotes P ≤ 0.001, *** denotes P ≤ 0.0001. Tukey’s convention was used to plot box and whisker plots. Statistical test details are indicated in the figure legends. No statistical methods were used to pre-determine sample sizes, but our sample sizes are similar to those reported in previous publications^4,35,79. Each micrograph, including images presented in Figs. 1a, b; 3a; 4f; 6c, h, i; 7c, l and Extended Data Fig. 1a, shows a representative image or image series from n>3 independent experiments. The experiments were not randomized; randomization is not relevant to our study since all experiments were conducted on cultured cells. Any variations observed between treatment groups are not attributed to sampling bias. Individual data points in each plot are available in source datasheet and data distribution was assumed to be normal but this was not formally tested. No data were excluded from the analyses. Data collection and analysis were not performed blind to the conditions of the experiments.

Manuscript writing

ChatGPT v3.5 software (https://chat.openai.com/chat) was employed in the writing of the Results section. Each figure was manually examined and a series of bullet points explaining each panel of the figure was written by the authors. Then ChatGPT was asked to convert the bullet points to better sentences with correct grammar. Next, the text was carefully reviewed by each author where incorrect sentences were removed. Finally, the manuscript was run on several plagiarism software platforms which returned no instances.

Extended Data

Extended Data figure 1. C2GAPB or RasG S17N regulates Ras and protrusive activities.

(a) Representative images of vegetative Dictyostelium before (top panel, ‘-DOX’) and after (bottom panel, ‘+DOX’) doxycycline-induced GFP-C2GAPB (green) expression. Cells (DIC) were treated with TRITC-dextran (red) before imaging. Scale bars: 10 µm. (b) Macropinocytosis uptake measurements, before (black) and after (red) C2GAPB expression. n=103 cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (Two-sided Mann-Whitney test). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). (c) Time-lapse images of vegetative Dictyostelium expressing mRFPmars-RasG S17N (red, bottom panel, ‘+DOX’) and RBD-GFP (green) after doxycycline treatment. Pink arrows denote RBD patches without RasG S17N. Time in ‘min’ format. Scale bars represent 5 µm. (d) Time-lapse images of C2GAPB-null Dictyostelium expressing RBD-GFP. Time in ‘min’ format. Scale bars: 5 µm. Time-lapse images of Dictyostelium or neutrophil expressing mRFPmars-SspB R73Q-RasG S17N ΔCAAX (e) or CRY2PHR-mCherry-KRAS4B S17N ΔCAAX (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel) (g). RasG or KRAS4B S17N ΔCAAX was recruited to cell front by applying laser near it, as shown by dashed white box. White arrows denote existing protrusions whereas pink arrows highlight newer protrusions. Time in min:sec format. Scale bars: 5 µm. (h) Polar histogram demonstrates higher probability of fresh protrusion formation away from recruitment area; n=14 cells and n=32 protrusions. (f) Time-lapse images of latrunculin A-treated Dictyostelium expressing mRFPmars-SspB R73Q-C2GAPB (red; bottom panel) and RBD-YFP (yellow; top panel) before or after global recruitment. Pink arrows indicate RBD patches whereas white arrows denote recruitment. Time in min:sec format. Scale bars: 5 µm. (i, j) Color-coded (2-mins intervals) outlines of cells in (c). Source numerical data are provided.

Extended Data figure 2. C2GAPB or RasG S17N-induced polarization localizes actin polymerization to a single front, and does not require development or GPCR signaling.

(a) Time-lapse confocal images of vegetative C2GAPB-null Dictyostelium cells expressing LimE-GFP. Time in ‘min’ format. Scale bars represent 5 µm. (b) Time-lapse confocal images of vegetative Dictyostelium cells expressing mRFPmars-RasG S17N (red) and LimE-GFP (green) after overnight doxycycline treatment. Pink arrows denote LimE-rich F-actin patches at the cellular protrusions. Time in ‘min’ format. Scale bars represent 5 µm. (c) Representative western blot (n=2 independent experiments) showing endogenous expression of cAR1 (~44 kDa) in developing wildtype (WT) or C2GAPB-expressing Dictyostelium cells during starvation (0–8 hrs). cAR1 appears as a doublet denoting its two forms, unmodified (lower band) and phosphorylated (upper band). (d) Time-lapse confocal images of vegetative Gβ-null Dictyostelium cells expressing mRFPmars-C2GAPB (bottom panel) or tgRFPt-Ctrl (control without C2GAPB; top panel) after overnight doxycycline treatment. Time in ‘sec’ format. Scale bars represent 5 µm. Color-coded (at 1-min interval) outlines of the C2GAPB- or Ctrl-expressing cell. (e) Time-lapse confocal images of Gβ-null Dictyostelium cell expressing mRFPmars-SspB R73Q-C2GAPB, before or after 488 nm laser was switched on globally. Time in min:sec format. Scale bars represent 5 µm. Color-coded (at 1-min interval) outlines of this cell is provided.

Extended Data figure 3. Global recruitment of dominant negative Ras isoforms improves polarity and migration in Dictyostelium and neutrophils.

(a) Time-lapse confocal images of vegetative Dictyostelium expressing mRFPmars-SspB R73Q-RasG S17N ΔCAAX before or after laser was switched on globally. Pink arrows denote recruitment causing increased movement. Time in min:sec format. Scale bars: 5 µm. Box-and-whisker plots of (b, d) average cell speed, and (c, e) cell area, before (black) or after (red) opto-Ctrl (control) or opto-RasG S17N ΔCAAX recruitment. n=10 (Ctrl) or n=7 cells (RasG S17N ΔCAAX) examined over 3 independent experiments; asterisks indicate significant difference, *P=0.0313 (d), *P=0.0156 (e); ns denotes non-significant difference, P=0.2730 (b), P=0.7394 (c) (Two-sided Wilcoxon signed-rank test). (f) Time-lapse images of neutrophil expressing CRY2PHR-mCherry-KRAS4B S17N ΔCAAX (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after recruitment. Pink arrows denote recruited cell; white arrows indicate control cell where there was no recruitment. Time in min:sec format. Scale bars: 5 µm. (g) Color-coded (1-min intervals) outlines of recruited cell in (f). (h) Representative kymograph of cortical LifeAct intensity before or after recruitment. Linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of kymograph is 22 mins. Cartoon depicts recruitment, F-actin polymerization or cell shape corresponding to kymograph. Box-and-whisker plots of (i) cell area, (l) average speed, and (m) aspect ratio, before (black) or after (red) recruitment. n=14 cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.001 (Two-sided Wilcoxon signed-rank test). Centroid tracks (n=13 cells) showing motility before (j) or after (k) recruitment. Each track lasts 5 mins and was reset to same origin. (b-e,i,l,m) Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Source numerical data are provided.

Extended Data figure 4. C2GAPB-induced polarity promotes vegetative Dictyostelium migration in folic acid gradients.

(a) Time-lapse confocal images of vegetative Dictyostelium cells, with (pink arrow) or without (white arrow; right panel) mRFPmars-SspB R73Q-C2GAPB inducible expression, chemotaxing to 100 nM folic acid. ‘+/- folic acid’ denotes that chemoattractant source was on the left-hand side of this field. Time in min:sec format. Scale bars represent 10 µm. (b) Centroid tracks of cells with (blue) or without (orange) C2GAPB expression (n=20 cells) in the same population migrating to a folic acid gradient. Each track lasts at least 20 mins and was reset to same origin. Source numerical data are provided.

Extended Data figure 5. RASAL3 back localization is dependent on cytoskeletal dynamics and its C-terminal tail whereas it induces polarity through its GAP domain.

(a) Time-lapse confocal images of JLY cocktail-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), after 488 nm laser was turned on globally. Time in min:sec format. Scale bars represent 5 µm. (b) Cartoon shows that JLY treatment caused RASAL3 to recruit uniformly instead of localizing to the back. (c, f) Schematics showing RASAL3 protein sequence (1–1011 amino acid long). Two truncation mutants were generated in this study: (c) RASAL3_689–1011 which consists of only the C-terminal tail (689–1011 amino acids) and (f) RASAL3_426–1011 consisting of the GAP domain along with the C-terminal tail (426–1011 amino acids). (d, g) Time-lapse confocal images of differentiated HL-60 neutrophil expressing (d) CRY2PHR-mCherry-RASAL3 (689–1011 aa) or (g) CRY2PHR-mCherry-RASAL3 (426–1011 aa) (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after 488 nm laser was turned on globally. Time in min:sec format. Scale bars represent 5 µm. (e, h) Cartoons demonstrate phenomenon observed with recruiting CRY2PHR-mCherry-RASAL3 (689–1011 aa) or CRY2PHR-mCherry-RASAL3 (426–1011 aa) in differentiated neutrophils. (i) Time-lapse confocal images of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-CTRL (control without RASAL3, red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after 488 nm laser was globally applied. Time in min:sec format. Scale bars represent 5 µm. (j) Color-coded (at 1-min intervals) outlines of the cell shown in (i). (k) Representative kymograph of cortical LifeAct intensity in CTRL-expressing neutrophil before or after 488 nm laser was switched on. A linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of the kymograph is 17 mins.

Extended Data figure 6. Ras suppression at the back improves polarity and contraction-driven migration.

Representative images of Dictyostelium expressing (a) mRFPmars-SspB R73Q-C2GAPB or (c) mRFPmars-SspB R73Q-RasG S17N ΔCAAX recruited to the back by applying laser near it (dashed white box). Time in min:sec format. Scale bars: 5 µm. (b) Cartoon demonstrates phenomenon in (a) and (c). (d) Time-lapse images of macrophage expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after global recruitment. Time in min:sec format. Scale bars represent 5 µm. (e) Time-lapse images of neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), placed in fMLP gradient, post-global recruitment. Chemoattractant was located on right-hand side of the field. Time in min:sec format. Scale bars: 5 µm. (f) Centroid tracks of chemotaxing opto-RASAL3-recruited neutrophils (n=10). Tracks last 2 mins and was reset to same origin. (g) Time-lapse images of CK666-treated neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after recruitment. Pink arrows denote long blebs which appear after recruited RASAL3 localized to back. Time in min:sec format. Scale bars: 5 µm. (h) Quantification of number of blebs per cell within a minute, before (black) or after (red) recruitment. n=37 cells examined over 3 independent experiments; asterisks indicate significant difference, ****P ≤ 0.0001 (Two-sided Wilcoxon signed-rank test). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Time-lapse images of Dictyostelium expressing mRFPmars-C2GAPB (red; top panel) and myosin II-GFP (green; middle panel) before (j) or after (i) doxycycline treatment. CK666 was added during imaging. Since there is no C2GAPB expression without doxycycline, red panel is not shown in (j). Pink arrows denote long blebs (DIC) in C2GAPB-expressing cells after CK666 was added. Time in min:sec format. Scale bars: 5 µm. Source numerical data are provided.

Extended Data figure 7. RasGAP recruitment cannot overcome blebbistatin inhibition but partially overcomes a sub-optimal dose.

(a) Time-lapse images of Dictyostelium expressing mRFPmars-SspB R73Q-C2GAPB (red; top panel), before or after global recruitment. 50 µM blebbistatin was added during imaging. Pink arrows denote cell of interest (DIC). Time in min:sec format. Scale bars: 5 µm. (b) Time-lapse images of neutrophil expressing LifeAct-miRFP703 (cyan; lower panel), before or after low-dose blebbistatin (20 µM) treatment. Time in min:sec format. Scale bars: 5 µm. (c) Strategy for testing low-dose blebbistatin on RASAL3-directed myosin II contraction and migration. (d) Time-lapse images of 20 µM blebbistatin-treated neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after recruitment. Time in min:sec format. Scale bars: 5 µm. (e) Color-coded (1-min intervals) outlines of cell in (d). (f) Representative kymograph of cortical LifeAct intensity before or after recruitment. Linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of kymograph is 16 mins. Cartoon summarizes recruitment, actin polymerization or cell shape corresponding to kymograph. Box-and-whisker plots of (g) average speed, (j) basal area, and (k) aspect ratio, before (black) and after (red) recruitment in blebbistatin-treated cells. n=11 cells examined over 3 independent experiments; asterisks indicate significant difference, ***P=0.001 (Two-sided Wilcoxon signed-rank test). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Centroid tracks of 20 µM blebbistatin-treated neutrophils (n=10 cells) showing motility before (h) or after (i) RASAL3 recruitment. Each track lasts 5 mins and was reset to same origin. Source numerical data are provided.

Extended Data figure 8. RasGAP-induced polarization works through ROCK-mediated myosin II phosphorylation.

(a) Time-lapse images of neutrophil expressing LifeAct-miRFP703 (cyan; lower panel), before or after low-dose Y27632 (10 µM) treatment. Time in min:sec format. Scale bars: 5 µm. Strategy for testing (b) low (10 µM) or (l) high (50 µM) Y27632 dose on RASAL3-directed myosin II contraction and migration. Time-lapse images of (c) 10 µM or (m) 50 µM Y27632-treated neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after global recruitment. Time in min:sec format. Scale bars: 5 µm. (d, n) Color-coded (1-min intervals) outlines of cells in (c) or (m). (e, o) Representative kymograph of cortical LifeAct intensity in (c) or (m). Linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of kymograph is 16 (e) or 26 (o) mins. Immunoblot comparing phospho-MRLC2 Ser19 (18 kDa) in cell lysates of 10 µM (f) or 50 µM (p) Y27632-treated, opto-RASAL3-expressing population, before or after recruitment (n=3 independent experiments). Quantifications of (g, q) cell speed, (j, t) area, and (k, u) aspect ratio, before (black) and after (red) recruitment in 10 or 50 µM Y27632-treated cells. n=13 (g,j,k) or n=11 cells (q,t,u) examined over 3 independent experiments; asterisks indicate significant difference, ***P=0.0002 (g,j,k); ns denotes non-significant difference P=0.4648 (q), P=0.2402 (t), P=0.3652 (u) (Two-sided Wilcoxon signed-rank test). Boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Centroid tracks of 10 µM (n=12 cells) or 50 µM (n=10 cells) Y27632-treated neutrophils showing random motility before (h, r) or after (i, s) RASAL3 recruitment. Each track lasts 5 mins and was reset to same origin. Source numerical data and unprocessed blots are provided.

Extended Data figure 9. Global RASAL3 recruitment overcomes PI3Kγ inhibition.

(a) Strategy for testing effect of PI3Kγ inhibitor, AS605240, on RASAL3-directed actin polymerization and motility. (b) Time-lapse confocal images of AS605240-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; upper panel) and LifeAct-miRFP703 (cyan; lower panel), before or after 488 nm laser was turned on globally. Time in min:sec format. Scale bars represent 5 µm. (c) Color-coded (1-min intervals) outlines of the cells shown in (b). (d) Representative kymograph of cortical LifeAct intensity in AS605240-treated RASAL3-expressing neutrophil before or after 488 nm laser was turned on. A linear color map shows that blue is the lowest LifeAct intensity whereas yellow is the highest. Duration of the kymograph is 19 mins. Cartoon summarizes membrane recruitment, actin polymerization or cell shape status corresponding to the kymograph. Box-and-whisker plots of (e) cell speed, (h) cell area, and (i) aspect ratio, before (black) and after (red) RASAL3 recruitment in AS605240-treated cells. n=13 cells examined over 3 independent experiments; asterisks indicate significant difference, ***P=0.0002 (Two-sided Wilcoxon signed-rank test). The boxes extend from 25^th to 75^th percentiles, median is at the center, and whiskers and outliers are graphed according to Tukey’s convention (GraphPad Prism 8). Centroid tracks of AS605240-treated neutrophils (n=12) showing random motility before (f) or after (g) RASAL3 recruitment. Each track lasts 5 mins and was reset to same origin. Source numerical data are provided.

Extended Data figure 10. Effect of GAP on EN system threshold and feedback loops that affect cell polarization.

(a) Model schematic of signal transduction pathway showing relevant parameters of Ras-PIP2-PKB excitable network. (b) Phase-plane diagram of Ras-PKB excitable network showing curves on which Ras and PKB stay constant, the equilibrium and the threshold of the EN system. The inset depicts two sets of parameter values in which the system is no longer excitable. (c) Sensitivity analysis of the system. The heatmap shows the change in threshold values normalized with respect to WT threshold for each of the ten parameters (a) of the EN. (d) Top: Thresholds in presence and absence (WT) of GAP on the Ras-PKB phase-plane. Bottom: Variation of value of the threshold between 0–1.5 times the WT value for two parameters $a_1$ and $a_2$ and their combined effect to captures the GAP effect. In our simulations the GAP value of the parameters is 1.5 times the WT value. (e) Histogram of cumulative Ras (left) and PKB (right) across 10 simulations before and after GAP recruitment. (f) Kymograph showing Ras activity around cell perimeter as a function of time (x-axis; 0–240 seconds). The bottom superimposes binarized locations where Ras> 0.25. Points before 140 seconds correspond to WT, and after 140 seconds correspond to global GAP recruitment. (g) The trajectory of CM of a single cell before and after GAP recruitment. The two plots show the path taken across different time segments of the simulation (denoted by the color) before and after recruitment of GAP. Absent colors from the graph shown in the legend, denote no movement in that interval. The thresholded Ras kymographs of panel f were used to generate trajectories in g. (h) Kymographs of the myosin and front polarity terms. The dotted line shows the time at which the GAP value was increased.

Supplementary Material

Suppl. Video 1

Time-lapse confocal microscopy of GFP-RBD (green; left panel) cortical waves at the substrate-attached surface in electrofused, giant Dictyostelium cell, before overnight doxycycline treatment (‘-DOX’ at top of the video). Without doxycycline, mRFPmars-C2GAPB (red; right panel) was not expressed in the cell. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 2

Time-lapse confocal microscopy of GFP-RBD (green; left panel) cortical waves at the substrate-attached surface in electrofused, giant Dictyostelium cells, after overnight doxycycline treatment (‘+DOX’ at top of the video). Doxycycline induced mRFPmars-C2GAPB (red; right panel) expression in the cell. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 3

Time-lapse confocal microscopy of PHcrac-YFP (green; left panel) cortical waves at the substrate-attached surface in electrofused, giant Dictyostelium cell, before overnight doxycycline treatment (‘-DOX’ at top of the video). Without doxycycline, mRFPmars-C2GAPB (red; right panel) was not expressed in the cell. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 4

Time-lapse confocal microscopy of PHcrac-YFP (green; left panel) cortical waves at the substrate-attached surface in electrofused, giant Dictyostelium cells, after overnight doxycycline treatment (‘+DOX’ at top of the video). Doxycycline induced mRFPmars-C2GAPB (red; right panel) expression in the cell. Top left corner shows the time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 5

Time-lapse confocal microscopy of RBD-YFP (yellow; left panel) cortical waves at the substrate-attached surface of electrofused, giant Dictyostelium cells, before or after global membrane recruitment of mRFPmars-SspB R73Q-C2GAPB (‘Opto-C2GAPB’; red, right panel). Global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘05:50’ and ‘30:00’. Propagating RBD waves extinguished with C2GAPB recruitment, and recovered when 488nm laser was switched off again. Top left corner shows the time in min:sec format. Scale bars represent 10 µm.

Suppl. Video 6

Time-lapse confocal microscopy of PHcrac-YFP (yellow; left panel) cortical waves at the substrate-attached surface of electrofused, giant Dictyostelium cells, before or after global membrane recruitment of mRFPmars-SspB R73Q-C2GAPB (‘Opto-C2GAPB’; red, right panel). Global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘16:20’ or ‘02:00’ in movie 1 or 2, respectively. Propagating PHcrac waves extinguished with C2GAPB recruitment. Top left corner shows the time in min:sec format. Scale bars represent 10 µm.

Suppl. Video 7

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel). RASAL3 was recruited to the membrane anchor, CIBN-CAAX (untagged), at the cell front by intermittently applying blue (488 nm) laser near it, as shown with the white box in the red channel. The disappearance of cellular protrusions at the RASAL3 recruitment site is denoted with solid white box in the blue channel. The top right corner shows time in min:sec format. Neutrophils were not exposed to chemoattractants during this experiment. Scale bar: 5 µm.

Suppl. Video 8

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-KRas4B S17N ΔCAAX (red; left panel) and LifeAct-miRFP703 (cyan; right panel). KRas4B S17N was recruited to the membrane anchor, CIBN-CAAX (untagged), at the cell front by intermittently applying blue (488 nm) laser near it, as shown with the white box in the red channel. The disappearance of cellular protrusions at the KRas4B S17N recruitment site is denoted with solid white box in the blue channel. The top right corner shows time in min:sec format. Neutrophils were not exposed to chemoattractants during this experiment. Scale bar: 5 µm.

Supp. Video 9

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cells before (-Dox; left panel) and after overnight doxycycline treatment (+Dox; right panel). Doxycycline induced mRFPmars-C2GAPB (red; right panel) expression in cells. The top right corner of each panel shows time in min:sec format. Cells were not exposed to chemoattractants during this experiment. Scale bar represents 20 µm.

Suppl. Video 10

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cell co-expressed with GFP-RBD (green; left panel) and mRFPmars-RasG S17N ΔCAAX (red; middle panel). With mRFPmars-RasGS17N expression (first movie, ‘+Dox’), cell became polarized with no apparent RBD patches, and migrated rapidly. Without mRFPmars-RasG S17N expression (second movie, ‘-Dox’), cell made several transient protrusions with RBD patches but no productive migration. Scale bar represents 5 µm.

Suppl. Video 11

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cell before or after global membrane recruitment of tgRFPt-SspB R73Q-Ctrl (Opto-Ctrl). Global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘04:00’. Top left corner shows time in min:sec format. Cell was not exposed to any chemoattractant during this experiment. Scale bar represents 5 µm.

Suppl. Video 12

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cell before or after global membrane recruitment of mRFPmars-SspB R73Q-C2GAPB (Opto-C2GAPB). In movie 1, global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘02:00’. In movie 2, blue laser was switched on or off multiple times during the course of the experiment, as denoted with “488nm on” or “488nm off” at the top of the movie. Top left corner shows time in min:sec format. Cells were not exposed to any chemoattractant during this experiment. Scale bar represents 5 µm.

Suppl. Video 13

Time-lapse confocal microscopy of vegetative Dictyostelium Gβ null (Gβ^-) cell before or after global membrane recruitment of mRFPmars-SspB R73Q-C2GAPB (Opto-C2GAPB). Global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘08:20’. Top left corner shows time in min:sec format. Cell was not exposed to any chemoattractant during this experiment. Scale bar represents 5 µm.

Suppl. Video 14

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cell before or after global membrane recruitment of mRFPmars-SspB R73Q-RasG S17N ΔCAAX (Opto-RasG S17N). Global recruitment was initiated when blue laser was applied globally (“488nm on” at bottom of the video) at ‘04:20’. Top left corner shows time in min:sec format. Cell was not exposed to any chemoattractant during this experiment. Scale bar represents 5 µm.

Suppl. Video 15

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-KRas4B S17N ΔCAAX (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was applied globally. The membrane anchor, untagged CIBN-CAAX, was expressed. Top left corner shows time in min:sec format. To start recruitment (red; left panel), the laser was switched on at ‘07:28’ once ‘488 nm ON’ appears at the top of the video. A non-recruitable cell, which presumably did not express CIBN-CAAX, was used as control in this video. Cells were not exposed to chemoattractant during the experiment. Scale bar represents 5 µm.

Suppl. Video 16

Time-lapse confocal microscopy of LimE_∆coil-YFP (yellow; left panel) cortical waves at the substrate-attached surface of electrofused, giant Dictyostelium cells, before or after global membrane recruitment of mRFPmars-SspB R73Q-C2GAPB (‘Opto-C2GAPB’; red, right panel). Global recruitment was initiated when blue laser was applied globally (“488nm on” at top of the video) at ‘15:00’. Propagating LimE waves extinguished, with C2GAPB recruitment, except for one standing wave in the middle. Top left corner shows the time in min:sec format. Scale bar represents 10 µm.

Suppl. Video 17

Time-lapse confocal microscopy of vegetative Dictyostelium AX2 cells chemotaxing to folic acid gradient. ‘+/- folic acid’ denotes a left to right gradient. mRFPmars-C2GAPB (red) expressing or non-expressing cell (DIC) in doxycycline-treated population exhibited different chemotactic behavior to 100 nM folic acid gradient. The top right corner shows time in min:sec format. Scale bar represents 10 µm.

Suppl. Video 18

Time-lapse confocal microscopy of differentiated, unpolarized HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel). RASAL3 was recruited to the membrane anchor, CIBN-CAAX (untagged), at the transient protrusions by intermittently applying blue (488 nm) laser near it, as shown with the white box in the red channel. The disappearance of cellular protrusions at the RASAL3 recruitment site is denoted with solid white box in the blue channel. 488 nm laser was switched off ‘03:09’ onwards to visualize the effects of RASAL3 self-rearrangement on the membrane. The top right corner shows time in min:sec format. Neutrophils were not exposed to chemoattractants during this experiment. Scale bar: 5 µm.

Suppl. Video 19

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was applied globally. The membrane anchor, untagged CIBN-CAAX, was expressed. Top left corner shows time in min:sec format. To start recruitment (red; left panel), the laser was switched on at ‘07:07’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to chemoattractant during the experiment. Scale bar represents 5 µm.

Suppl. Video 20

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-CTRL (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was also expressed. Top left corner shows time in min:sec format. To start recruitment (red; left panel), the laser was switched on at ‘05:43’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to chemoattractant during the experiment. Scale bar represents 5 µm.

Suppl. Video 21

Time-lapse confocal microscopy of differentiated HL-60 macrophage expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was applied globally. The membrane anchor, untagged CIBN-CAAX, was expressed. Top left corner shows time in min:sec format. To start recruitment (red; left panel), the laser was switched on at ‘04:33’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to chemoattractant during the experiment. Scale bar represents 5 µm.

Suppl. Video 22

Time-lapse confocal microscopy of differentiated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), chemotaxing to 5 nM fMLP gradient. ‘-/+ fMLP’ denotes a right to left gradient. Membrane anchor, untagged CIBN-CAAX, was also expressed. Top left corner shows time in min:sec format. To start recruitment (red; left panel), the laser was switched on at ‘01:59’ once ‘488 nm ON’ appears at the top of the video. Scale bar represents 5 µm.

Suppl. Video 23

Time-lapse confocal microscopy of differentiated, blebbistatin-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was also expressed. Cell was treated with 75 µM blebbistatin atleast 10 mins before imaging. Pink arrow denotes the cell of interest. To start recruitment (red; left panel), the laser was switched on at ‘05:15’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to chemoattractant during the experiment. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 24

Time-lapse confocal microscopy of differentiated, blebbistatin-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was also expressed. Cell was treated with 20 µM blebbistatin atleast 10 mins before imaging. Pink arrow denotes the cell of interest. To start recruitment (red; left panel), the laser was switched on at ‘03:23’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to chemoattractant during the experiment. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 25

Time-lapse confocal microscopy of differentiated, Y27632-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was also expressed. Cell was treated with 10 µM Y27632 atleast 10 mins before imaging. To start recruitment (red; left panel), the laser was switched on at ‘04:12’ once ‘488 nm ON’ appears at the top of the video. Neutrophil was not exposed to any chemoattractant during imaging. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 26

Time-lapse confocal microscopy of differentiated, Y27632-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was also expressed. Cell was treated with 50 µM Y27632 atleast 10 mins before imaging. To start recruitment (red; left panel), the laser was switched on at ‘07:00’ once ‘488 nm ON’ appears at the top of the video. Neutrophil was not exposed to any chemoattractant during imaging. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 27

Time-lapse confocal microscopy of differentiated, LY294002-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was applied globally. The membrane anchor, untagged CIBN-CAAX, was expressed as well. Cell was treated with 50 µM LY294002 atleast 10 mins before imaging. To start recruitment (red; left panel), the laser was switched on at ‘04:12’ once ‘488 nm ON’ appears at the top of the video. Neutrophil was not exposed to any chemoattractant during imaging. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 28

Time-lapse confocal microscopy of differentiated, AS605240-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was expressed here. Cell was treated with 20 µM AS605240 atleast 10 mins before imaging. To start recruitment (red; left panel), the laser was turned on at ‘03:58’ once ‘488 nm ON’ appears at the top of the video. Neutrophil was not exposed to any chemoattractant during imaging. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 29

Time-lapse confocal microscopy of differentiated, PP242-treated HL-60 neutrophil expressing CRY2PHR-mCherry-RASAL3 (red; left panel) and LifeAct-miRFP703 (cyan; right panel), before or after 488 nm laser was turned on globally. Membrane anchor, untagged CIBN-CAAX, was expressed. Cell was treated with 20 µM PP242, 10 mins before imaging. To start recruitment (red; left panel), the laser was turned on at ‘06:04’ once ‘488 nm ON’ appears at the top of the video. Cell was not exposed to any chemoattractant during imaging. Top left corner shows time in min:sec format. Scale bar represents 5 µm.

Suppl. Video 30

Two-dimensional simulation results of the excitable network for varying RasGAP levels. The video shows three different simulations, with 80% WT, WT and 110%-RasGAP levels, sequentially. The area is a square with sides 40 μm long and the time-stamp denotes seconds. The three colors correspond to Ras (red), PIP2 (green) and PKB (blue).

Data availability

All data needed to evaluate the conclusions are provided in the main text and figures, extended data figures, or supplementary tables and videos. All unprocessed immunoblots or raw data and associated statistical calculations are provided along with this study. Data in Fig. 2k and l were reanalyzed here from our previous study⁴. Source data are provided with this study. All other data supporting the findings of this study are available from the corresponding authors on reasonable request.