|Home | About | Journals | Submit | Contact Us | Français|
Directional cell migration requires the coordination of actin assembly and membrane remodeling. The exocyst is an octameric protein complex essential for exocytosis and plasma membrane remodeling [1,2]. A component of the exocyst, Exo70, directly interacts with the Arp2/3 complex, a core nucleating factor for the generation of branched actin networks for cell morphogenesis and migration [3-9]. Using in vitro actin polymerization assay and time-lapse TIRF microscopy, we found Exo70 functions as a kinetic activator of the Arp2/3 complex that promotes actin filament nucleation and branching. We further found that the effect of Exo70 on actin is mediated by promoting the interaction of Arp2/3 complex with WAVE2, a member of the N-WASP/WAVE family of nucleation promoting factors (NPFs). At the cellular level, the stimulatory effect of Exo70 on Arp2/3 is required for lamellipodia formation and maintaining directional persistence of cell migration. Our findings provide a novel mechanism for regulating actin polymerization and branching for effective membrane protrusion during cell morphogenesis and migration.
We have previously shown that Exo70 directly interacts with the Arp2/3 complex . However, the molecular consequence and functional implication of this interaction is unknown. Here, we purified Exo70 and the Arp2/3 complex (Figure S1A, B) and performed polymerization assays using pyrene-labeled actin. Since the WAVE complexes are usually inactive , we used the recombinant WAVE2 as the nucleation-promoting factor in our assay. Exo70 alone or together with the Arp2/3 complex did not affect F-actin assembly. WAVE2 stimulated the Arp2/3 complex; further addition of Exo70 significantly enhanced actin polymerization (Figure 1A). The concentration of barbed ends in our experiment setting (2μM G-actin, 25 nM WAVE2, and 15 nM Arp2/3) was 0.22 nM in the absence of Exo70, and was 0.41 nM in the presence of Exo70. The stimulatory effect of Exo70 is dose-dependent (Figure 1B). Furthermore, in the presence of saturating concentration of Exo70 (1 μM), significantly less WAVE2 is required (reduced from 25 nM to 6.25 nM) for Arp2/3 activation (Figure S1C).
We have previously found that Exo70 lacking residues 628-630 had a weaker interaction with the Arp2/3 complex . Indeed, Exo70(Δ628–630) had a much lower stimulatory effect on actin polymerization than the wild type Exo70 (Figure 1A). Addition of Exo70 reduced the time to reach half maximal polymerization by 2.4-fold, whereas Exo70(Δ628–630) had a much weaker effect (Figure 1C). The maximal rate of actin polymerization was increased by 100% with Exo70, but only 34% with Exo70(Δ628–630) (Figure 1D).
Upon activation by WAVE2, the Arp2/3 complex generates new actin branches on the side of pre-existing actin filaments, which can be monitored by total internal reflection fluorescence microscopy (TIRFM) . Here we examined the effect of Exo70 on actin branching in real-time using TIRFM. To better visualize branching and elongation of the new filaments, we performed dual-color imaging to better differentiate the pre-existing and newly formed actin filaments. Rhodamine-labeled G-actin was allowed to polymerize for 3 mins and captured onto the surface of a coverslip coated with NEM-myosin. Then Cy5-labeled G-actin together with Arp2/3, WAVE2, wild type or mutant Exo70 was flown into the chamber to replace Rhodamine-labeled G-actin. The newly generated actin filaments were monitored over time. In the presence of Arp2/3 and WAVE2, branches were generated from the sides of actin filaments (Figure 2A, upper panel; see Movie S1 for both audio and visual presentation of the generation of new branches; see Figure S2 for enlarged images). Addition of Exo70 significantly stimulated branch formation (Figure 2A, middle panel; Movie S2 and Figure S2). The Exo70(Δ628–630) mutant barely had any stimulatory effect (Figure 2A, bottom panel; Movie S3). Addition of Exo70 in the absence of WAVE2 had no effect on actin (Movie S4). After 240 seconds, the branching ratio (number of branches/number of total actin filaments) was 6-fold higher in the presence of Exo70 (Figure 2B). The lengths of the newly generated actin filaments were also measured. In reactions with Exo70, the actin filaments were slightly shorter than those without Exo70 (Figure 2C), suggesting that, when the total amount of G-actin is limited (0.8 μM of 8% Cy5-labeled G-actin, in contrast to 2 μM G-actin in the pyrene actin assay), an increase in actin branching results in a reduction of the final length of the filaments.
Since the effect of Exo70 is shown only in the presence of WAVE2, Exo70 itself is unlikely to be an actin nucleator or a nucleation-promoting factor (NPF). Then what is the molecular mechanism by which Exo70 stimulates the Arp2/3 complex? Exo70 binds to the Arp2/3 complex through its ARPC1 subunit (a.k.a. p40 or Arc40), which physically contacts N-WASP/WAVE [12-14]. It is thus possible that Exo70 affects the WAVE2-Arp2/3 interaction, which in turn promotes the Arp2/3 complex activation. To test this possibility, we first performed in vitro binding assay using GST-WAVE2 and purified Arp2/3 complex (Figure 3A). In the presence of Exo70, the amount of Arp2/3 bound to GST-WAVE2 increased by 3.5-fold (Figure 3B). As a negative control, GST-Rabin8, a fusion protein with a similar molecular weight to Exo70, did not bind to the Arp2/3 complex. The affinity of the Arp2/3 complex to WAVE2 is much higher in the presence of Exo70 (kd=210 nM) than without Exo70 (kd=470 nM) (Figure S3A). To examine whether Exo70 affects the WAVE2-Arp2/3 interaction in cells, immunoprecipitation experiments were carried out using cells expressing FLAG-WAVE2 (Figure 3C). When Exo70 was knocked down by siRNA, only 44% of Arp3 was co-precipitated with WAVE2 comparing with control siRNA-treated cells (Figure 3D). Reciprocally, when ARPC1 was knocked down by siRNA, endogenous Exo70 was no longer co-precipitated with WAVE2 (Figure 3E). Consistent with the finding, the Exo70(Δ628–630) mutant interacts much weakly with WAVE2 in comparison to the wild type Exo70 as tested by co-immunoprecipitation experiments (Figure 3F) and pull down experiments (Figure S3B). Taken together, our data strongly suggested that Exo70 positively regulates the interaction between the Arp2/3 complex and WAVE2, and provide a molecular basis for our observed stimulatory effect of Exo70 on Arp2/3 activation.
To examine the physiological function of Exo70 in the stimulation of the Arp2/3 complex, we generated human MDA-MB-231 stable cell lines expressing GFP-tagged rat wild type Exo70 and a Exo70(Δ628-630) mutant that is defective in stimulating the Arp2/3 complex (see above). The endogenous Exo70 was then knocked down by siRNA in these cells (Figure S4A). The levels of other exocyst subunits, WAVE2 and Arp3 were not affected in the knockdown cells (Figure S4B), and Exo70(Δ628-630) binds to Sec8 to the same extent as the wild type (Figure S4C). As shown in Figure 4A, the Exo70-knockdown cells expressing Exo70(Δ628-630) failed to form lamellipodia as indicated by Arp3 and F-actin staining. The “lamellipodia ratio”, defined as the lamellipodia length divided by the total cell perimeter, was significantly lower in the knockdown cells expressing Exo70(Δ628-630) (Figure 4B). The migratory properties of these cells were also examined. In transwell assays, the number of mutant Exo70 cells migrated to the lower chamber was much smaller than the control cells (Figure 4C). Also, using a wound-healing assay, we found that Exo70 siRNA knockdown cells expressing Exo70(Δ628-630) took more time for wound closure (Figure 4D and Figure S4D). By tracking the trajectories of individual cells over time, we found that cells expressing Exo70(Δ628-630) were not only slower in migration, but also less capable of maintaining the direction of their movement (Figure 4E, F). The directional persistence, quantified as the ratio of net displacement to the total migration distance (“D:T ratio”), was much smaller in cells expressing Exo70(Δ628-630) (Figure 4E). These results suggest that the function of Exo70 in stimulating the Arp2/3 complex plays an important role in membrane protrusion formation and directional cell migration.
Our experiments, for the first time, revealed that Exo70 is able to kinetically stimulate the Arp2/3 complex in the presence of WAVE2 for actin polymerization and branching. Exo70 binds to the Arp2/3 complex through ARPC1, the subunit that physically contacts N-WASP/WAVE for activation [12-14]. Our data show that Exo70 stimulates the interaction between the Arp2/3 complex and WAVE2, which may provide a molecular mechanism for the observed stimulatory effect of Exo70 on the Arp2/3 complex activation. During membrane protrusion formation, associated with actin assembly is dynamic membrane reorganization [15,16]. Exo70, as a member of the exocyst complex, is involved in post-Golgi exocytosis and membrane remodeling (; Guo lab unpublished data). The effect of Exo70 on Arp2/3-mediated actin polymerization and branching may couple actin dynamics and membrane reorganization for cell morphogenesis and effective migration.
We have previously shown that the Exo70-Arp2/3 complex interaction is stronger in cells treated with EGF , or expressing Y527F c-Src , which are known to stimulate membrane protrusion and cell migration. These observations suggest that, in cells, the effect of Exo70 on the Arp2/3 complex is up-regulated by signaling events to levels above what we observed here using recombinant proteins. It is possible that the activation of Exo70 and/or the Arp2/3 complex by small GTP-binding proteins or kinases in response to growth factors stimulates their interaction. Future investigation of this regulation will lead to better understanding of many processes such as chemotaxis and tumor cell invasion.
Detailed information on plasmids, antibodies, cell culture, protein purification, immunoprecipitation and in vitro binding assay, fluorescence microscopy, wound-healing assay, transwell assay, individual cell movement tracking is described in the Supplemental Experimental Procedures.
Pyrene actin polymerization assay was performed as previously described [19, 20]. The kinetics of actin polymerization was monitored by pyrene fluorescence with excitation at 370 nm and emission at 410 nm. Proteins to be tested for each reaction (15 nM Arp2/3, 25 nM WAVE2, 1 μM of GST, wild type or mutant Exo70 were mixed and diluted in 90 μl polymerization buffer (60 mM KC1, 2.5 mM NaCl, 0.6 mM MgCl2, 5 mM Tris-HCl, pH 7.5, 2.5 mM HEPES, pH 7.1, 0.5 mM EGTA, 30 μM CaCl2, 0.2 mM ATP, and 0.3 mM NaN3), then pre-incubated for 30 mins at 4°C prior to the addition to the actin assembly mix. 20% pyrenyl-labeled actin was prepared using pyrene-labeled muscle actin and unlabeled muscle actin (Cytoskeleton Inc.). 11.5 μl G-actin (20.9 μM, 20% pyrenyl-labeled) in G-actin buffer (Triethylamine, 0.3mM CaCl2, 0.1mM EDTA, 0.7mM ATP, 6.25mM NaN3) was converted to Mg-G-actin by mixing with 18.5 μl Mg2+ converting buffer (1.6 mM Tris-HCl, pH8, 0.18 mM EGTA, and 0.51 mM MgCl2) for 5 min. The converted actin mixture was then diluted into 90μl protein mixture with a final G-actin concentration of 2 μM, and immediately transferred into a cuvette and monitored for the increase in fluorescence at 410 nm in a fluorescence spectrophotometer. Polymerization kinetics was anlyzed using Excel (Microsoft). The polymerization rate was represented as the maximal slope of the elongation phase of each curve. Statistical analyses were performed using the paired Student's t-test (n=3). Measurement of barbed ends is performed as previously described .
TIRF-based actin branching assay was performed as previously described  with modifications. Total internal reflection excitation was generated on a Nikon TE2000U inverted microscope with an in-house TIRF illuminator using objective-type (Nikon 100×, Plan-achromat 1.49 NA oil immersion objective) TIRF excitation. Images were collected using a back-illuminated electron multiplying-CCD camera (Cascade-512B, Photometrics). Two-color ALEX (Alternating-laser excitation)  was used for fluorescence excitation. A solid-state 532-nm laser (CrystaLaser) and a Helium-Neon red laser (NEC, 10mW max) were used for excitation of Rhodamine-actin and Cy5-actin, respectively. Frame rate was set at 0.5 frame per second. To minimize photo-bleaching, laser intensities were reduced either by a neutral density filter or an intensity attenuator comprising a half-wave plate and a polarizer. Exposure time was set at ~ 400 ms and laser framing were synchronized with the camera shutter through a pulse sent from the camera.
Experiments were carried out in flow chambers assembled by mounting a pre-cleaned coverslip to a standard glass slide using Dow Corning 732 multi-purpose silicone sealant as spacers. Before use, the flow chambers were coated with NEM-myosin by flowing in 50 nM NEM-myosin in high salt buffer (500 mM KC1) and incubating for 1.5 min. The flow chambers were then treated with 1% BSA for 4 min to prevent non-specific binding to the surface. 15 nM Arp2/3, 25 nM WAVE2, 250 nM wild type or mutant Exo70 were mixed in the actin polymerization buffer and incubated at 4°C for 30 mins. 2 μM 6% Rhodamine-labeled actin was first converted to Mg-G-actin, and added to the flow chamber for polymerization. Rhodamine-labeled actin polymerization was allowed to proceed for 3 mins. Then a polymerization buffer containing 0.8 μM 8% Cy5-labeled Mg-G-actin together with 15 nM Arp2/3 complex, 25 nM WAVE2, 250 nM wild type or mutant Exo70 were added into the chamber to replace Rhodamine-labeled actin in the solution. Images were collected immediately after the addition of Cy5-labeled actin. For each ALEX cycle, Cy5 actin was imaged for 9 frames and Rhodamine actin 2 frames. Images were recorded for 10 min after the polymerization started. Average lengths of newly formed Cy5-actin filaments were measured over time from 25 individual actin filaments for samples with or without Exo70 using ImageJ 1.73v software (http://rsb.info.nih.gov/ij/). Statistical analyses were performed using Student's t-test.
We thank Drs. Sally Zigmond, Bruce Goode, Shu-Chan Hsu, Shanshan Feng, Peng Yue, Michael Rosen, Henry Higgs, Giorgio Scita, Jeffery Peterson, and Roberto Dominguez for reagents and helpful discussions. The research is supported by National Institutes of Health grants (GM085146) to W.G. and T.S. Y.Z. is supported by an American Heart Association pre-doctoral fellowship. The JEOL 1011 transmission electron microscope was supported by the NIH Shared Instrumentation Grant to T.S.
Author Contributions: J.L., Y.Z. and W.G. conceived the project and designed experiments; J.L., Y.Z., C.Y., B.H. and Y.S. designed and/or carried out some of the experiments; J.L., Y.Z. and W.G. analyzed the data and wrote the manuscript. J.L. and Y.Z. contributed equally to this work.
Competing financial interests: The authors declare no competing financial interests.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.