Molecular clones of the subgroup A feline leukemia virus FeLV-A/Glasgow-1 have been obtained. Nucleotide sequence analysis of the 3' end of the proviral genome and comparison with the published sequence of FeLV-B/Gardner-Arnstein showed that the most extensive differences are located within the 5' domain of the env gene. Within this domain, several divergent regions of env are separated by more conserved segments. The 3' end of env is highly conserved, with only a single amino acid coding difference in p15env. The proviral long terminal repeats are also highly conserved, differing by only eight base substitutions and one base insertion. Specific probes constructed from the FeLV-A or FeLV-B env genes were used to compare the env genes of various exogenous FeLV isolates and the endogenous FeLV-related proviruses of normal cat DNA. An FeLV-A-derived env probe showed no hybridization to normal cat DNA but detected all FeLV-A and FeLV-C isolates tested. In contrast, an FeLV-B env probe detected independent FeLV-B isolates and a family of endogenous FeLV-related proviruses. Our observations provide strong evidence to support the hypothesis that FeLV-B viruses have arisen by recombination between FeLV-A and endogenous proviral elements in cat DNA.
We demonstrate a spectroscopic imaging based super-resolution approach by separating the overlapping diffraction spots into several detectors during a single scanning period and taking advantage of the size-dependent emission wavelength in nanoparticles.This approach has been tested using off-the-shelf quantum dots (Invitrogen Qdot) and inhouse novel ultra-small (~3 nm) Ge QDs.Furthermore, we developed a method-specific Gaussian fitting and maximum likelihood estimation based on a Matlab algorithm for fast QD localisation.This methodology results in a three-fold improvement in the number of localised QDs compared to non-spectroscopic images.With the addition of advanced ultrasmall Ge probes, the number can be improved even further, giving at least 1.5 times improvement when compared to Qdots.Using a standard scanning confocal microscope we achieved a data acquisition rate of 200 ms per image frame.This is an improvement on single molecule localisation super-resolution microscopy where repeated image capture limits the imaging speed, and the size of fluorescence probes limits the possible theoretical localisation resolution.We show that our spectral deconvolution approach has a potential to deliver data acquisition rates on the ms scale thus providing super-resolution in live systems.
The sequential and regulated recruitment of leukocytes into tissues by chemoattractants is essential for effective clearance of pathogens and healing. The Rho GTPases Cdc42, Rac, and Rho are important for establishing and maintaining migratory polarity. Most chemoattractants for phagocytes signal either through seven transmembrane G-protein-coupled receptors (GPCRs) or tyrosine kinase receptors. Y721 is the most important for chemotaxis because it recruits phospholipase-C-γ (PL C-γ) and the p85 subunit of class 1A PI3Ks, both of which are implicated in the initiation of chemotaxis. Several intracellular signaling complexes contribute to the polarization of phagocytes in response to chemoattractants, and they probably act together to allow optimal chemotaxis. Cdc42 is implicated in multiple types of cell polarity, including axon specification, yeast mating, and epithelial polarity. There are several PLC isoforms, of which PLCβ2 and PLCβ3 are activated by GPCR signaling in neutrophils, whereas PLCβ isoforms are activated by tyrosine kinase receptors. Polarity signals act to initiate polarization of cells, but subsequent maintenance of polarity could be achieved by Rac and Rho without the requirement for additional signals. Rho and Rac refine each other's activity during cell polarization and migration, balancing actin polymerization, cell contraction, and adhesion essential for chemotaxis. Our current understanding of chemotaxis indicates that several signaling pathways act in concert to induce cell polarization, including Cdc42, Par proteins, PAK/PIX, and PI3Ks. The design and testing of inhibitors of signal transduction molecules involved in migration and chemotaxis will be an important goal for the future.
The cystic fibrosis transmembrane conductance regulator ( CFTR ) gene shows a complex pattern of expression, with temporal and spatial regulation that is not accounted for by elements in the promoter. One approach to identifying the regulatory elements for CFTR is the mapping of DNase I hypersensitive sites (DHS) within the locus. We previously identified at least 12 clusters of DHS across the CFTR gene and here further evaluate DHS in introns 2, 3, 10, 16, 17a, 18, 20 and 21 to assess their functional importance in regulation of CFTR gene expression. Transient transfections of enhan‐ cer/reporter constructs containing the DHS regions showed that those in introns 20 and 21 augmented the activity of the CFTR promoter. Structural analysis of the DNA sequence at the DHS suggested that only the one intron 21 might be caused by inherent DNA structures. Cell specificity of the DHS suggested a role for the DHS in introns 2 and 18 in CFTR expression in some pancreatic duct cells. Finally, regulatory elements at the DHS in introns 10 and 18 may contribute to upregulation of CFTR gene transcription by forskolin and mitomycin C, respectively. These data support a model of regulation of expression of the CFTR gene in which multiple elements contribute to tightly co‐ordinated expression in vivo .
Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Migrating cells need to coordinate distinct leading and trailing edge dynamics but the underlying mechanisms are unclear. Here, we combine experiments and mathematical modeling to elaborate the minimal autonomous biochemical machinery necessary and sufficient for this dynamic coordination and cell movement. RhoA activates Rac1 via DIA and inhibits Rac1 via ROCK, while Rac1 inhibits RhoA through PAK. Our data suggest that in motile, polarized cells, RhoA–ROCK interactions prevail at the rear, whereas RhoA-DIA interactions dominate at the front where Rac1/Rho oscillations drive protrusions and retractions. At the rear, high RhoA and low Rac1 activities are maintained until a wave of oscillatory GTPase activities from the cell front reaches the rear, inducing transient GTPase oscillations and RhoA activity spikes. After the rear retracts, the initial GTPase pattern resumes. Our findings show how periodic, propagating GTPase waves coordinate distinct GTPase patterns at the leading and trailing edge dynamics in moving cells. Introduction Cell migration relies on the coordination of actin dynamics at the leading and the trailing edges (Ridley et al., 2003). During the mesenchymal type of migration, protrusive filamentous actin (F-actin) is cyclically polymerized/depolymerized at the cell’s leading edge, whereas the contractile, actomyosin-enriched trailing edge forms the rear. The leading edge protrudes and retracts multiple times, until the protrusions, known as lamellipodia, are stabilized by adhering to the extracellular matrix (Ridley, 2001). Subsequently, the cell rear detaches and contracts allowing the cell body to be pulled toward the front. Core biochemical mechanisms of this dynamic cycle are governed by the Rho family of small GTPases (Jaffe and Hall, 2005). Two members of this family, Ras homolog family member A (RhoA) and Ras-related C3 botulinum toxin substrate 1 (Rac1), control protrusions and retractions at the leading edge as well as the contractility at the rear (Felmet et al., 2005; Heasman and Ridley, 2008; Machacek et al., 2009). RhoGTPases cycle between an active, GTP-loaded ‘on’ state and an inactive, GDP-loaded ‘off’ state. Switches between on and off states are tightly regulated by (i) guanine nucleotide exchange factors (GEFs) that facilitate GDP/GTP exchange thereby activating GTPases and (ii) GTPase activating proteins (GAPs) that stimulate GTP hydrolysis and transition to a GDP-bound state. A canonic description of mesenchymal cell migration portrays mutually separated zones of Rac1-GTP and RhoA-GTP in polarized cells where Rac1-GTP dominates at the leading edge and RhoA-GTP dominates at the contracted cell rear (Holmes and Edelstein-Keshet, 2016; Holmes et al., 2017; Kunida et al., 2012; Kurokawa and Matsuda, 2005; Pertz et al., 2006; Wang et al., 2013; Zmurchok and Holmes, 2020). This distinct distribution of RhoA and Rac1 activities along polarized cells is explained by a mutual antagonism of RhoA and Rac1 (Edelstein-Keshet et al., 2013; Mori et al., 2008) mediated by downstream effectors of these GTPases (Byrne et al., 2016; Guilluy et al., 2011a; Pertz, 2010). The Rac1 effector, p21 associated kinase (PAK), phosphorylates and inhibits multiple RhoA-specific GEFs, including p115-RhoGEF, GEF-H1 and Net1 (Alberts et al., 2005; Guilluy et al., 2011a; Rosenfeldt et al., 2006). In addition, active Rac1 binds and activates p190RhoGAP, which decreases RhoA activity (Guilluy et al., 2011a). In turn, RhoA-GTP recruits the Rho-associated kinase (ROCK), which phosphorylates and activates Rac-specific GAPs, such as FilGAP and ArhGAP22, thereby inhibiting Rac1 (Guilluy et al., 2011a; Ohta et al., 2006; Sanz-Moreno et al., 2008). This mutual inhibition of RhoA and Rac1 may lead to a bistable behavior where a system can switch between two stable steady states, in which GTPase activities alternate between high and low values (Kholodenko, 2006; Mori et al., 2008). The existence of bistable switches is supported by experiments, where inhibition of the Rac1 effector PAK maintains both high RhoA and low Rac1 activities and associated morphological changes even after the inhibition is released (Byrne et al., 2016). At the same time, RhoA and Rac1 do not behave antagonistically at the leading edge of migrating cells. Here, RhoA activation is rapidly followed by Rac1 activation, tracking a protrusion-retraction cycle (Machacek et al., 2009). This Rac1 activation at the leading edge is mediated by the downstream RhoA effector, Diaphanous related formin-1 (DIA), that was shown to localize to the membrane ruffles of motile cells (Tkachenko et al., 2011; Watanabe et al., 1997). Thus, in contrast to the RhoA effector ROCK, which inhibits Rac1 in the other cell segments, the RhoA effector DIA can stimulate Rac1 activity at the leading edge. If at the leading edge RhoA activates Rac1 but Rac1 inhibits RhoA, this intertwined network circuitry of positive and negative loops will force the network to periodically change RhoA and Rac1 activities, giving rise to self-perpetuating oscillations with a constant amplitude and frequency (Kholodenko, 2006; Tsyganov et al., 2012). By contrast, at the trailing edge and cell body, the mutual RhoA and Rac1 inhibition results in the maintenance of a (quasi)steady state with high RhoA activity and low Rac1 activity. But, how can these different dynamics coexist? More importantly, how are these dynamics coordinated within the cell? Despite decades of research that have painstakingly characterized dynamic Rho and Rac behaviors in cell motility (Holmes and Edelstein-Keshet, 2012), we do not know what dynamic features are necessary and sufficient to achieve the biological effect of cell motility, and how different dynamics at the front and rear are coordinated. Here, we first elucidated the spatial profiles of RhoA-Rac1 interactions in motile MDA-MB-231 breast cancer cells. Using proximity ligation assays (PLA), we show that the concentration of complexes formed by RhoA and its downstream effectors DIA and ROCK depends on the spatial location along the longitudinal axis of polarized cells. RhoA primarily interacts with DIA at the cell leading edge, whereas RhoA - ROCK interactions are the strongest at the cell rear. Based on these findings, we built a mathematical model to analyze RhoA-Rac1 signaling in space and time. The model predicts and the experiments corroborate that at the cell front the GTPase network exhibits oscillatory behavior with high average Rac1-GTP, whereas at the cell rear there is a (quasi)steady state with high RhoA-GTP and low Rac. The front and rear are connected by periodic, propagating GTPase waves. When the wave reaches the rear, RhoA-GTP transiently oscillates and then, following the rear retraction, the GTPase network dynamic pattern returns to the original state. Our model and experimental results show how different GTPase dynamics at the leading edge and the trailing edge can govern distinct cytoskeleton processes and how moving cells reconcile these different dynamics. The RhoA-Rac1 interaction network model defines minimal, autonomous biochemical machinery that is necessary and sufficient for biologically observed modes of cell movement. Results Spatially variable topology of the RhoA-Rac1 interaction network The Rac1 effector PAK inhibits RhoA, and the RhoA effector ROCK inhibits Rac1 (Guilluy et al., 2011a). Here, we tested how the other RhoA effector, DIA, influences the Rac1 and RhoA activities. We first downregulated DIA using small interfering RNA (siRNA) and measured the resulting changes in the Rac1-GTP and RhoA-GTP levels. Downregulation of DIA increased the RhoA abundance and decreased Rac1 abundance, while decreasing relative activities of both RhoA and Rac1 (Figure 1—figure supplement 1, panels A and B). The decrease of relative Rac1 and RhoA activities induced by DIA knockdown shows that DIA activates Rac1 and also supports the existence of a positive feedback loop between DIA and RhoA described earlier (Kitzing et al., 2007). In addition, the GTPase network features another positive feedback from PAK to Rac1 through several molecular mechanisms (Baird et al., 2005; DerMardirossian et al., 2004; Feng et al., 2002; Obermeier et al., 1998). Summing up the interactions between RhoA and Rac1 mediated by their effectors ROCK and PAK (Byrne et al., 2016) and RhoA - Rac1 interactions through DIA, we arrive at the intertwined negative and positive feedback circuitry of the RhoA-Rac1 network shown in Figure 1—figure supplement 1, panel C. To explain the distinct GTPase activities at the leading and trailing edges, we hypothesized that these diverse feedforward and feedback mechanisms may be spatially controlled. Therefore, we explored how the interactions of active RhoA with its effectors vary spatially in polarized MDA-MB-231 cells. Using a proximity ligation assay (PLA), which visualizes protein interactions in situ (Gustafsdottir et al., 2005; Söderberg et al., 2006), we measured RhoA-DIA and RhoA-ROCK complexes (Figure 1A and B). Based on the commonly considered morphology of the long, narrow cell rear and the wide leading edge (Caswell and Zech, 2018), we segmented each polarized cell into three parts: the rear (about 20% of the cell length), intermediate region (next 70% of the cell length), and front (the rest 10% of the length). The density of the RhoA-effector complexes was quantified by dividing the number of PLA reactions by the area of the corresponding compartment. Figure 1 with 1 supplement see all Download asset Open asset Differential localization of the RhoA-DIA and RhoA-ROCK1 protein complexes determine spatially resolved signaling topology. (A, B) Representative PLA images. Each red spot within a cell represents a fluorescent signal from a single RhoA-DIA1 (A) or RhoA-ROCK1 (B) complex. Yellow lines indicate bounds for the leading edge, intermediate region and rear. Bar graphs at the right show the average density of these complexes in different cell regions (the rear, middle and leading edge)± S.E.M. of four independent experiments with 25 cells analyzed per experiment. The asterisk * indicates that p<0.05 calculated using unpaired t-test. (C, D) Representative images of DIA1 and ROCK1 immunostaining. Bar graphs at the right show quantified immunostaining density signals for different cellular compartments ± S.E.M. of four independent experiments with one cell analyzed per experiment. The asterisk * indicates that p<0.05 calculated using unpaired t-test. (E) A schematic wiring diagram of the RhoA-Rac1 network, showing positive (blue) and negative (magenta) feedback loops. Spatially varying RhoA interactions with its effectors DIA and ROCK are shown by dashed lines. The results show that the RhoA-DIA complexes are predominantly localized at the cell front, whereas their density is markedly decreased at the rear (Figure 1A). In contrast, the density of the RhoA-ROCK complexes increases toward the cell rear and decreases at the leading edge (Figure 1B). These results are in line with protein staining data in polarized cells, which suggest that DIA is mainly localized at the leading edge (Figure 1C), whereas ROCK is abundant at the rear and cell body (Figure 1D; Brandt et al., 2007; Goulimari et al., 2005; Newell-Litwa et al., 2015; Watanabe et al., 1997; Wheeler and Ridley, 2004). For MDA-MB-231 cells, our quantitative proteomics data showed that the RhoA abundance is at least 10-fold larger than the abundance of DIA and ROCK isoforms combined (Byrne et al., 2016). Thus, as shown in the Modeling section of Materials and methods, the RhoA-effector concentrations depend approximately linearly on the DIA and ROCK abundances. Taken together, these results suggest a protein interaction circuitry of the GTPase network, where competing effector interactions are spatially controlled (Figure 1E). In order to analyze how this differential spatial arrangement of GTPase-effector interactions can accomplish the dynamic coordination between the leading and trailing edges, we constructed a mechanistic mathematical model and populated it by quantitative mass spectrometry data on protein abundances (Supplementary file 1). Analyzing the dynamics of the RhoA-Rac1 interaction network The changes in ROCK and DIA abundances along the longitudinal axis of polarized cells (Figure 1C and D) could plausibly encode the distinct RhoA-Rac1 temporal behaviors in different cellular segments. Therefore, we explored these possible dynamics of the GTPase network for different DIA and ROCK abundances prevailing at different spatial positions along the cell length. We first used a spatially localized, compartmentalized model where different DIA and ROCK abundances corresponded to distinct spatial locations (see Modeling section of Materials and methods for a detailed description of this model). Using the model, we partitioned a plane of the ROCK and DIA abundances into the areas of different temporal dynamics of RhoA and Rac1 activities (Figure 2A). This partitioning is a two-parameter bifurcation diagram where the regions of distinct GTPase dynamics are separated by bifurcation boundaries at which abrupt, dramatic changes in the dynamic behavior occur (Holmes and Edelstein-Keshet, 2016). The blue region 1 in Figure 2A corresponds to the self-perpetuating oscillations of the RhoA and Rac1 activities at the leading edge. The ROCK abundance is markedly lower and the DIA abundance is higher at the leading edge than in the cell body (Figure 1C and D). Thus, a combination of Rac1 activation by RhoA via DIA and RhoA inhibition by Rac1 via PAK (Figure 2B) results in sustained oscillations of RhoA and Rac1 activities at the leading edge (Figure 2D). This periodic Rac1 activation drives actin polymerization at the leading edge pushing protrusion-retraction cycles (Machacek et al., 2009; Martin et al., 2016; Pertz, 2010; Tkachenko et al., 2011). Figure 2 with 3 supplements see all Download asset Open asset A mathematical model of the RhoA-Rac1 network predicts dramatically distinct dynamic regimes for different DIA and ROCK abundances. (A) Distinct dynamic regimes of the RhoA-Rac1 network dynamics for different DIA and ROCK abundances. Oscillations of RhoA and Rac1 activity exist within area 1 (regime 1). In area 3, sustained GTPase oscillations and a stable steady state with high RhoA and low Rac1 activities coexist. Regimes 0, 2, 5 and 6 have only one stable steady state. Notably, regime 2 is excitable. Steady state solutions with high RhoA activity exist in areas 2–4, and 6–8. Stable steady state solutions with high Rac1 activity exist in areas 0 and 5–8. Regimes 4, 7 and 8 are bistable with two stable steady states. (B, C) Wiring diagrams of the RhoA-Rac1 network for the cell leading edge (B) and the cell body and rear (C). Dashed blue lines indicate weak activating connections. (D–F) Typical time courses of RhoA and Rac1 activity in regimes 1 (D), and 2 (E). (F) In area 3, depending on the initial state, the GTPase network evolves either to a stable steady state (right) or a stable oscillatory regime (left). The green region 2 in Figure 2A is an area of stable high RhoA and low Rac1 activities at the rear and intermediate cell regions. Within this region, RhoA inhibits Rac1 via ROCK, and Rac1 inhibits RhoA via PAK (Figure 2C). After perturbations, the GTPase network converges to steady-state levels of high RhoA-GTP and particularly low Rac1-GTP (Figure 2E). Unlike other dynamical regimes with only a single stable steady state, region 2 corresponds to an excitable an medium, which cannot generate pulses itself, but supports the propagation of excitable activity pulses (see Materials and methods section). The red region 3 corresponds to the coexistence of GTPase oscillations and a stable steady state with high RhoA and low Rac1 activities. Depending on the initial state, the GTPase network evolves to different dynamic regimes. If the initial state has high RhoA-GTP and low Rac1-GTP, the GTPase network progresses to a stable steady state, but if the initial state has low RhoA-GTP and high Rac1-GTP, the network will develop sustained oscillations (Figure 2F). This region 3 is termed a BiDR (Bi-Dynamic-Regimes) by analogy with a bi-stable region where two stable steady states coexist and the system can evolve to any of these states depending on the initial state (Kholodenko, 2006). However, in contrast with bistable regimes only one of two stable regimes is a stable steady state in the BiDR region, whereas the other dynamic regime is a limit cycle that generates stable oscillations. In addition to these dynamic regimes, the spatially localized model predicts other emergent non-linear dynamic behaviors (Figure 2A, Figure 2—figure supplement 1, panels A-D, and Figure 2—figure supplement 2), which the GTPase network may execute under large perturbations of the RhoA and Rac1 effector abundances to coordinate GTPase signaling at the leading and trailing edges (see Modeling section of Materials and methods for a detail description of these regimes). Therefore, we next analyzed how the leading and trailing edge GTPase dynamics are coupled. Spatiotemporal dynamics of the RhoA-Rac1 network reconciles the distinct temporal behaviors at the cell front and rear Different active GTPase concentrations in the cell rear and the leading edge induce diffusion fluxes (Das et al., 2015), which in turn influence the emerging behavior of these GTPases and coordinate their dynamics in distinct cellular segments. As a multitude of dynamic behaviors is possible, we systematically explored the behavior of the RhoA-Rac1 network in space and time using a spatiotemporal model of the GTPase network interactions (referred to as a reaction-diffusion model, see Materials and methods). Starting from experimental observations to rationalize which behaviors are likely with physiological boundaries, we digitized 2D images of polarized cells and incorporated the DIA and ROCK abundances as functions of the spatial coordinate along the cell length, based on the quantitative imaging data (Figure 3A–C). Figure 3 with 1 supplement see all Download asset Open asset Spatial propagation of RhoA and Rac1 activities during cell motility. (A) A 2-D calculation domain obtained by digitizing cell images. Different cellular compartments are indicated. The x-axis represents the direction of cell polarization, the y-axis represents the perpendicular direction. (B, C) The abundance profiles of DIA and ROCK used in simulations (red lines) are superimposed on the experimental spatial profiles (bar graphs in Figure 1C and D). (D–G) Model-predicted spatial patterns of the RhoA and Rac1 activities for different phases of the cell movement cycle. (D, F) Rac1 and RhoA activity snapshots during a protrusion-retraction cycle at the leading edge (t = 175 s from the start of the moving cycle). (E, G) represent snapshots when the Rac1 and RhoA activity wave have spread over the entire cell, reaching the rear (t = 1518 s). (H) The RhoA activity at the leading edge and cell body during a protrusion-retraction phase measured by RhoA FRET probe in space and time. The arrows compare model-predicted and experimentally measured patterns, indicating zones of RhoA oscillatory and high constant activities and a ‘dark zone’ of low RhoA activity. (I) Spatiotemporal pattern of the RhoA activity during further RhoA wave propagation into the cell. (J) The number of RhoA activity bursts at the cell body and rear during 10 min measured using the RhoA FRET probe. Error bars represent 1st and 3rd quartiles, *** indicate p<0.001 calculated using unpaired t-test. (K–M) Fluorescent microscopy images of Rac1 activity (red), combined with staining for F-actin (phalloidin, white) and the nucleus (DAPI, blue) in fixed cells for different phases of the cell movement cycle; (K) a protrusion-retraction cycle at the leading edge, and (L, M) present Rac1 activity wave propagation into the cell body. The images (L, M) were obtained by super-resolution microscopy. The model predicts autonomous, repeating cycles of the spatiotemporal GTPase dynamics (Figure 3D–G and Video 1). For a substantial part of a dynamic cycle, high RhoA-GTP and low Rac1-GTP persist at the cell rear and maintain the rear contraction, whereas active RhoA and Rac1 oscillate at the leading edge, resulting in actin (de)polarization cycles and protrusion-retraction cycles (Figure 3D and F; Wang et al., 2013). At the same time, a wave of oscillating Rac1 and RhoA activities slowly propagates from the leading edge toward the cell rear (Figure 3E and G). Between the oscillatory RhoA-GTP zone and the areas of high RhoA activity, a zone of low RhoA activity emerges (Figure 3F). As time progresses, the wave of oscillating GTPase activities and the area of low RhoA activity spread to the rear (Figure 3—figure supplement 1, panels A and B), leading to re-arrangement of the cytoskeleton (Warner et al., 2019). Because of the oscillations, zones of low Rac1 activities emerge, which give rise to high RhoA-GTP that interacts with ROCK and leads to the rear retraction (Video 1). Subsequently, RhoA returns to its initial high stable activity, and the dynamic pattern of RhoA-GTP and Rac1-GTP over the entire cell returns to its initial state. These model simulations could plausibly explain how the different GTPase dynamics at the cell front and rear are coordinated to enable successful cell migration. Video 1 Download asset This video cannot be played in place because your browser does support HTML5 video. You may still download the video for offline viewing. Download as MPEG-4 Download as WebM Download as Ogg Model-predicted spatiotemporal activity patterns of RhoA and Rac1. Therefore, it was important to test the prediction arising from the model simulations in biological experiments. For this, we used cells stably expressing the mTFP-YFP RhoA-GTP FRET-probe (Kim et al., 2015) allowing us to determine the RhoA-GTP dynamics using ratiometric, live-cell spinning disk microscopy. We imaged the cells with a frequency of one image every 5 s and constrained the measurement time to 10 min to limit phototoxic effects. Due to this time limitation, a full cycle of cellular movement (around 45 min on average, Video 2) could not be followed in an individual cell, and the full spatiotemporal RhoA activity cycle during a cell movement was compiled from several cells observed in different phases of cellular movement. In the initial phase of the cell movement cycle, the spatiotemporal RhoA activity showed three different zones: (i) oscillations at the leading edge, (ii) dark zone of low activity and (iii) light zone of high activity (Figure 3H and Figure 3—figure supplement 1, panel C) in the cell body and rear, matching the model prediction (Figsure 3F and Figure 3—figure supplement 1, panel C). As time progressed, the GTPase activity wave propagated further into the cell (Figure 3I), forming zones of high and low RhoA activities. In the space-time coordinates, the slope of the boundaries of these zones suggests that they travel from the leading edge to the cell rear, confirming the model predictions (Video 1 and Figure 3I). When the wave of oscillatory GTPase activities finally reaches the cell rear, it induces several RhoA-GTP spikes (Figure 3G and I), periods of low RhoA activity (Figure 3—figure supplement 1, panels A-B and D), and subsequent return to the original, high RhoA-GTP at the rear and part of the cell body (Figure 3F and H). Figure 3—figure supplement 1, panel D experimentally captures this transition from a low RhoA activity to the original high activity as the final step of the cell movement cycle predicted by the model. Video 2 Download asset This video cannot be played in place because your browser does support HTML5 video. You may still download the video for offline viewing. Download as MPEG-4 Download as WebM Download as Ogg Live-cell imaging of cel movement cycles. Red color represents staining of the nuclei. Frame increment is 10 min. The model predicts that during a single cellular movement cycle, multiple bursts of RhoA activity appear at the leading edge, whereas at the cell rear, RhoA activity bursts occur only after the RhoA-Rac1 wave has spread through the cell (Video 1). Measuring the number of RhoA bursts at the leading edge and cell rear during observation time (10 min) corroborated model predictions, showing a ca. fivefold larger number of bursts at the leading edge than at the cell rear (Figure 3J). On average, at the leading edge a burst of RhoA activity happens every minute, while at the cell rear only 1 or 2 bursts happen during 10 min (Figure 3J). Although spatially resolved Rac1 activity can be determined using exogenous probes, they dramatically change the cell shape when expressed (Pertz, 2010). However, endogenous Rac1-GTP can be reliably detected by immunostaining with a conformation-specific Rac1-GTP antibody. Rac1 was mainly active at the leading edge with lower activity in the space between the nucleus and cell rear (Figure 3K), similar to the patterns observed in the model for protrusion-retraction cycles (Figure 3D). The GTPase waves can be detected using super-resolution imaging. These images corroborated the Rac1-GTP presence towards the cell nucleus and rear (see super-resolution images in Figure 3L–M and Figure 3—figure supplement 1, panel E). The series of images shown in Figure 3K–M and Figure 3—figure supplement 1, panel E is consistent with the concept of traveling Rac1-GTP waves predicted by the model. The spatiotemporal activation dynamics of Rac1 and RhoA underlie the morphological events during cell migration, that is protrusion-retraction cycles at the front and the retraction cycle at the rear (Ridley et al., 2003; Video 2). These mechanical processes, involving cytoskeleton proteins, can be coordinated by periodic propagating waves of RhoGTPase activities described by our model. Hysteresis of Rac1 and RhoA activities and cell shape features We previously showed that PAK inhibition could change the cell shape of MDA-MB-231 cells from mesenchymal to amoeboid (Byrne et al., 2016). The mesenchymal mode of migration features an elongated cell morphology and high Rac1 activity, whereas the amoeboid mode is hallmarked by a rounded morphology and high RhoA activity (Sanz-Moreno et al., 2008). These morphologies and migration types are mutually exclusive but can transition into each other. Our previous study showed that this transition correlated with the hysteresis of active RhoA and Rac1 upon PAK inhibition (Byrne et al., 2016). Hysteresis is the hallmark of bistability: if a parameter, such as the PAK abundance, reaches a threshold value, then the system flips from one stable state to another stable state, at which it remains for a prolonged period of time even when this parameter has returned to its initial value (Markevich et al., 2004; Sha et al., 2003). Our model now allows us to examine the exact spatiotemporal kinetics of the GTPase network in response to changes in PAK abundance or activity. Varying PAK causes Rac1 and RhoA activities to move through different dynamic regimes (shown by the line connecting points I – II – III in Figure 4A). In unperturbed cells, GTPase activities oscillate at the leading edge. This initial network state corresponds to point I in region 1n and unperturbed ROCK, PAK and DIA abundances and activities (the point I coordinates are (1, 1) in Figure 4A). Because Rac1 and RhoA are difficult to target for therapeutic interventions, we used a small molecule PAK inhibitor (IPA-3) in our previous study (Byrne et al., 2016). As PAK abundance gradually decreases (or PAK inhibition increases), the system moves from the oscillatory region 1 to the BiDR region 3, before reaching a bistable regime (regions 7 and 8), as shown by point II. In the BiDR region, (i) a stable high RhoA-GTP, low Rac1-GTP state and (ii) a stable oscillatory state with a high average Rac1-GTP coexist at the leading edge (Figure 2F and Figure 2—figure supplement 2, panel D). While moving from point I into area 3, the system continues to display the stable oscillatory state with high average Rac1-GTP. In the bistable regions 7 and 8, two stable states co-exist (i) high RhoA-GTP, low Rac1-GTP and (ii) low RhoA-GTP, high Rac1-GTP (Figure 2—figure supplement 2, panels H and I). Entering area 7 from the BiDR area 3, the system relaxes to the steady state with the higher Rac1-GTP level. Only with the further PAK decrease, a saddle-node bifurcation (see Materials and methods) shifts the system to the alternative steady state with the much lower Rac1-GTP level. Figure 4 with 1 supplement see all Download asset Open asset Hysteresis of the RhoA and Rac1 activities are manifested upon PAK inhibition and recapitulated by a spatiotemporal model. (A) Distinct dynamic regimes of the RhoA-Rac1 network for different DIA and ROCK abundances. Colors and numbers of dynamic regimes are the same as in Figure 2A. (B, C) Model-predicted dependencies of the RhoA and Rac1 activities on the PAK abundance for gradually decreasing (blue) and increasing (red) PAK abundances. The network evolution occurs through two different routes (blue and red curves in B and C). It is calculated by averaging the GTPase activities over the time and cell volume based on western blot data reported in our previous study (Byrne et al., 2016). Points I, II and III shown in black (A) are also indicated on the network trajectories (B, C). (D–F) Snapshots of si
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)