Staurosporine (STS) has been known as a classic protein kinase C inhibitor and is a broad-spectrum inhibitor targeting over 250 protein kinases. In this study, we observed that STS treatment induced drastic morphologic changes, such as elongation of a very large number of nonbranched, actin-based long cell protrusions that reached up to 30 µm in an hour without caspase activation or PARP cleavage in fibroblasts and epithelial cells. These cell protrusions were elongated not only from the free cell edge but also from the cell-cell junctions. The elongation of STS-dependent protrusions was required for ATP hydrolysis and was dependent on myosin-X and fascin but independent of Cdc42 and VASP. Interestingly, in the presence of an actin polymerization inhibitor, namely, cytochalasin D, latrunculin A, or jasplakinolide, STS treatment induced excess tubulin polymerization, which resulted in the formation of many extra-long microtubule (MT)–based protrusions toward the outside of the cell. The unique MT-based protrusions were thick and linear compared with the STS-induced filaments or stationary filopodia. These protrusions, which were composed of microtubules, have been scarcely observed in cultured non-neuronal cells. Taken together, our findings revealed that STS-sensitive kinases are essential for the maintenance of normal cell morphology, and a common unidentified molecular mechanism is involved in the formation of the following two different types of protrusions: actin-based filaments and MT-based shafts.
Spatiotemporal regulation of cell protrusion is involved in morphologic dynamics. Cellular protrusions are formed mainly through elongation of cytoskeletal filaments in the extracellular direction and are universally observed during differentiation, proliferation, migration, apoptosis, and expression of physiologic function in living cells. Several types of cell protrusions have been observed, including filopodia, microvilli, stereocilia, bristle, invadopodia, podosome, tunneling nanotube (TNT), microtentacle, and nerve axon (Revenu et al., 2004; Murphy and Courtneidge, 2011; Charpentier et al., 2014). Representative components of the cytoskeleton include actin, tubulin, and vimentin, which form distinct types of cytoskeletal filament, actin filament (AF), microtubule (MT), and intermediate filament, respectively (Blanchoin et al., 2014). Coordinate rearrangement of cytoskeletal filaments are required for the active morphologic changes of growing cells.
The length of AFs depends on the balance between stabilization and destabilization of actin polymerization, which are regulated by adaptor proteins, such as actin-modulating proteins (Rotty et al., 2013). To date, actin-modulating proteins that have been reported include profilin and thymosin β4 as G-actin–binding proteins, cofilin/ADF as a depolymerizing protein, CAPZ as a capping protein, Arp2/3 complex as a branching protein, fascin as a bundling protein, and VASP and formin as polymerization-promoting proteins (dos Remedios et al., 2003). These adaptor proteins are modulated by Rho family small GTPases, including Rho, Rac, and Cdc42, as well as kinases and phosphatases regulated by these GTPases, and contribute to the spatiotemporal dynamics of the cytoskeleton (Heasman and Ridley, 2008). The AF length control mechanism mostly remains to be elucidated (Mogilner and Rubinstein, 2005). Interestingly, the presence of long filopodia reaching a length of up to 150 µm has also been demonstrated in the development process of the limb bud (Sanders et al., 2013). These long filopodia are considered to be small GTPase-independent protrusions (Gousset et al., 2013).
MT dynamics are also regulated through the stabilization/destabilization balance. MTs assume two forms, depending on the location of the initiation point, centrosomal MT, and noncentrosomal MT, and both forms require stabilization of the MT minus end for formation of the initiation point. The γ-tubulin ring complex contributes to stabilization of the minus end in the centrosomal MT form (Kollman et al., 2011), while CAMSAP3 has been identified as a stabilization factor for the noncentrosomal MT form (Meng et al., 2008). In MTs with the stabilized minus end, their length is controlled by dynamic instability between polymerization and depolymerization of the plus end (van der Vaart et al., 2009).
Since actin and tubulin tend to polymerize nonenzymatically and spontaneously, cells are equipped with complicated molecular systems for efficient regulation of polymerization and depolymerization. Many of these regulatory factors are modulated by protein phosphorylation (Govek et al., 2005). Therefore, universal inhibition of these kinases appears to allow for observation of cell morphology in the absence of physiologic regulation. To prove this hypothesis, we observed morphologic changes of cells using staurosporine (STS), which has been known as a classic protein kinase C (PKC) inhibitor and is a broad-spectrum inhibitor targeting over 250 protein kinases (Karaman et al., 2008). The STS-treated cells showed elongation of a very large number of nonbranched, actin-based protrusions reaching a length as long as 30 µm in an hour. These protrusions extended from the contact area between cells, and no retraction was observed. Interestingly, we found that combination of STS and an actin polymerization inhibitor promoted tubulin polymerization, resulting in formation of cellular protrusions with a new morphology of extremely long, MT-based protrusions. These protrusions, which were composed of microtubules, have seldom been observed in cultured non-neuronal cells. We propose that there is a previously unidentified mechanism to regulate the length and composition of protrusions involved in the STS-sensitive factors.
Materials and Methods
Reagents and Plasmids.
All reagents and plasmids used in this study are listed in Supplemental Tables 1–3. All primers were obtained from Sigma-Aldrich (St. Louis, MO). All polymerase chain reaction products were completely sequenced.
Cell Culture and Transfections.
HeLa, 3Y1, COS7, HepG2, and H1299 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin at 37°C, 5% CO2. Primary human nasal epithelial cells were prepared as previously described (Someya et al., 2013). Lipofectamine plus reagent (Invitrogen, Carlsbad, CA) and RNAiMAX reagent (Invitrogen) were used for plasmid and small interfering RNA transfection, respectively, according to the manufacturer’s instructions. For RNA interference experiments, HeLa cells were plated at 1 × 105 cells per well of a 24-well plate. The cells were used 48–72 hours after siRNA transfection. For STS treatment, cells grown on a glass bottom dish (Iwaki, Tokyo, Japan) or a glass cover slip (Matsunami, Osaka, Japan) were gently washed once with a glucose-free Krebs’ solution (KH buffer; 120 mM NaCl, 5 mM KCl, 0.62 mM MgSO4, 1.8 mM CaCl2, 10 mM HEPES, pH 7.2) at room temperature, and then treated with the indicated concentrations of STS in the absence or presence of the indicated compounds at room temperature for the indicated times. The cells were then used for the experiments as described below.
Live-Cell Imaging and Immunofluorescence Microscopy.
For live-cell imaging, the cells that were prepared as described above were placed under atmospheric conditions at room temperature on the microscope stage. Images were acquired with the LSM-510 META microscope system (Carl Zeiss, Thornwood, NJ), employing the channel or lambda mode using a Plan-Apochromat 63×/1.40 Oil DIC M27 oil immersion objective, and ZEN software (Carl Zeiss). For fixed-cell imaging, the cells that were prepared as described above were fixed in an equal volume of 10% trichloroacetic acid (Sigma-Aldrich) for 20 minutes at room temperature, and then washed four times with phosphate-buffered saline (PBS), followed by permeabilization for 2 minutes in PBS containing 0.5% saponin (ICN, Costa Mesa, CA) and 1% fatty acid–free bovine serum albumin (Sigma-Aldrich). Primary and secondary antibodies were diluted in a Can Get Signal solution (Toyobo, Osaka, Japan). The cell staining was performed at room temperature in the dark for 60 minutes. H33342 [2′-(4-ethoxyphenyl)-5-(4-methylpiperazin-1-yl)-1H,1′H-2,5′-bibenzo[d]imidazole] was added to the secondary antibody staining solution. The coverslips were mounted in Mowiol 4-88 (Calbiochem, San Diego, CA) with 1,4-diazabicyclo[2.2.2]octane (Sigma-Aldrich). Fixed samples were imaged as described above. Live-cell images, fixed-cell images, and time-lapse movies were processed using the ImageJ program (http://imagej.nih.gov/ij/) with self-written macros. Figures were generated with Adobe Photoshop CS6 and Illustrator CS6 (San Jose, CA). The length and number of cell protrusions were measured using ImageJ and represented as a dot-plot (modified scatter plot) in Microsoft Excel (Seattle, WA) using self-written macros. The Student’s t test was done based on 30 randomly chosen data sets from all raw data using Stat Plus software (AnalystSoft; http://www.analystsoft.com/en/) with self-written macros.
Transmission Electron Microscopy.
Transmission electron microscopy (TEM) samples were prepared as described previously (Someya et al., 2013). In brief, cells grown on eight-well chamber slides were treated with STS and then fixed in an equal volume of 5% glutaraldehyde/0.1 M PBS (pH 7.3) overnight at 4°C, postfixed in 2% osmium tetroxide in the same buffer, dehydrated in a graded ethanol series, and embedded in Epon 812. Ultrathin sections were then cut on a Sorvall Ultramicrotome MT-5000 (Dupont-Sorvall, Newtown, CT). The sections were stained with uranyl acetate followed by lead citrate and examined at 80 kV with a transmission electron microscope (H7500; Hitachi, Tokyo, Japan).
Cells were washed twice with cold PBS and lysed in an extraction buffer [50 mM Tric-HCl, pH 7.5, 150 mM NaCl, 2 mM vanadate, 1 mM EDTA, complete protease inhibitor cocktail (Roche, Basel, Switzerland), 0.1% NP40], followed by centrifugation at 20,000g for 15 minutes at 4°C. The resulting supernatant was analyzed by SDS-PAGE, transferred to a polyvinylidene fluoride membrane, and then washed in 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, and 0.1% Tween20 and blocked with 1% nonfat milk in 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, and 0.1% Tween20 for 5 minutes at 65°C. The blotted membrane was incubated with a primary antibody for 1 hour at room temperature or overnight at 4°C, followed by incubation with a peroxidase-conjugated secondary antibody for 1 hour at room temperature. Signals were detected with an enhanced chemiluminescence–plus reagent (GE Healthcare, Piscataway, NJ) in LAS1000 (Fujifilm, Tokyo, Japan).
STS Induces Elongation of Many Long Cell Protrusions.
We observed STS-induced morphologic changes in multiple cells in their living state and found common features thereof. For example, a 30-minute treatment of HeLa cells with STS at room temperature induced marked elongation of many nonbranched, long cell protrusions (Fig. 1A). Such long cell protrusions were not formed under experimental conditions that did not contain STS (Fig. 1B). Time-lapse images of STS-treated cells indicated that the STS-induced protrusions extended while continuously moving to and fro without any retraction (Fig. 1C; Supplemental Movies 1 and 2). In the absence of the cell membrane dye FM1-43 [N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino) styryl) pyridinium dibromide], the elongation rate of these protrusions was 0.82 µm/min 30 minutes after STS addition and decreased over time (Fig. 1D), but spontaneous vibration was observed even more than 4 hours after the addition of STS. The length of the elongated protrusions was dependent on the STS concentration, which ranged from 25 nM to 1 µM (Fig. 1E). Many of the known cell protrusions are formed in membrane regions not in contact with other cells (Vasioukhin et al., 2000). Interestingly, the STS-induced protrusions were also formed from regions involved in cell-to-cell contact (Fig. 1F). These STS-dependent changes in cell morphology have also been observed in the rat fibroblast cell 3Y1 (Fig. 1G), African Green Monkey SV40-transfected kidney fibroblast cell COS7 (Fig. 1H), human liver cancer cell HepG2 (Fig. 1I), human non–small-cell lung cancer cell H1299 (Fig. 1J), and primary human nasal epithelial cells (Supplemental Fig. 1). From these results, STS was demonstrated to induce morphologic changes of marked elongation of many long cell protrusions in a short period of time in multiple cell types.
Cell Protrusions That Elongate in an STS-Dependent Manner Are Composed of Actin Filaments.
Despite the fact that STS shows inhibitory activity against many kinases, including PKC, no long protrusion was observed in the presence of another protein kinase inhibitor, as indicated in Fig. 2 (left panels); moreover, cotreatment with STS and the indicated inhibitor did not repress the elongation of STS-dependent protrusions (right graphs). STS has also been known to induce caspase-3/7–dependent apoptosis (Jiang et al., 2013), but no increase in caspase-3/7 activity was observed within the time range of 0–4 hours, which is required for the elongation of STS-dependent protrusion (Fig. 3A). To clarify the constituents of the STS-dependent cell protrusion, we observed the cells using TEM. As shown in Fig. 3B, these protrusions that were formed on HeLa cells were suggested to contain many long actin filaments (left panel, arrows). These filaments were oriented parallel to the direction of the elongation, and little branching was observed. No long protrusion was observed in the control (right panel). To identify that the protrusions are composed of actin filaments, the cells were transfected with a green fluorescent protein (GFP)–actin plasmid, and its intracellular localization in living cells was observed with a confocal microscope. In steady-state HeLa cells, GFP-actin was shown to localize at focal contacts and stress fibers or distribute diffusely in the cytoplasm (Fig. 3C, lower panels). On the other hand, in the presence of STS, GFP-actin was not only strongly localized in the vicinity of cell margins, but also distributed over the entire region of long protrusions (Fig. 3C, upper panels). In addition, the actin stress fibers observed before the addition of STS disappeared upon addition of STS. STS did not induce an appreciable change in the shape of the nucleus within the observation period. Next, we observed endogenous actin localization. Since STS-induced protrusions were susceptible to aldehyde fixation and alcohol fixation (unpublished observations), the cells were fixed with trichloroacetic acid (TCA) in this experiment (Hayashi et al., 1999). While antibodies to endogenous actin recognize both F-actin and G-actin in TCA-fixed cells, actin was strongly localized to the cell margins under the conditions of this experiment (Fig. 3D, right panel). In the presence of STS, endogenous actin, like the GFP-actin described above, was distributed in the vicinity of cell margins and over the entire elongating region of protrusions (Fig. 3D, left panel). From these results, cell protrusions that elongate extremely in an STS-dependent manner were demonstrated to be filopodia-like protrusions composed of actin filaments.
ATP Supply Is Required for the Elongation of STS-Dependent Cell Protrusions.
A supply of actin monomers bound by ATP is essential for polymerization of actin (Pollard, 1986). ATP is produced through both oxidative phosphorylation and glucose metabolism, and the increased ATP level is linked with accelerated cell functions (Vander Heiden et al., 2009). Since STS inhibits kinase activity nonselectively and broadly, STS treatment was predicted to cause an increase in the intracellular ATP level. Before investigating whether STS changes the intracellular ATP level, we examined whether the cellular ATP content was maintained by either oxidative phosphorylation or aerobic glycosylation. As shown in Fig. 4A, the intracellular ATP level decreased to 15% of the initial level in the presence of the glucose metabolism inhibitor 2-deoxyglucose (2-DG) for 2 hours, as indicated in magenta, while control was maintained at 90% or higher (in green). Therefore, ATP production in these cells was found to be almost entirely dependent on glucose metabolism. Then, we observed morphologic changes occurring in ATP-depleted cells upon STS administration. As shown in Fig. 4B, STS-dependent, markedly long protrusions were not formed in the presence of 2-DG compared with the absence of 2-DG (Fig. 4D). 2-DG caused little morphologic change in steady-state cells (Fig. 4, C and E). Next, we investigated whether STS induced an increase in the intracellular ATP level. As shown in Fig. 4F, STS treatment resulted in an immediate consumption of approximately 10% of ATP followed by a modest decline, but at least STS treatment did not increase the total intracellular ATP level under the above-mentioned conditions (open and closed triangles). We tried to search for additional intracellular ATP-depleting agents in our preliminary trial and successfully found that the NADPH oxidase inhibitor diphenylene iodide (DPI) depleted intracellular ATP in HeLa cells. As shown in Fig. 4F, the ATP level that was reduced through the administration of DPI (open circles) was restored by the addition of glucose (closed circles). Therefore, we used DPI for a recoverable ATP-depleting reagent. We then observed morphologic changes occurring in these cells upon administration of STS. DPI caused almost no morphologic changes in steady-state HeLa cells (Fig. 4G). STS administration in the presence of DPI resulted in complete inhibition of the elongation of STS-dependent protrusions, and the cells showed healthy cell morphology similar to that of control (Fig. 4H). This inhibitory effect was eliminated by the addition of glucose (Fig. 4I). With respect to the characteristics of the protrusions observed here, no significant difference was found in the length or number of protrusions (Fig. 4J, lane 6) compared with the absence of DPI (Fig. 4J, lane 2). In addition, the loading of glucose without DPI did not affect the length or number of STS-induced protrusions (Fig. 4J, lane 3). From these results, ATP was found to be essential for the elongation of STS-dependent protrusions. Our data also revealed that the broad inhibition of kinases by STS does not necessarily immediately trigger accumulation of intracellular ATP.
Myosin-X and Fascin Are Involved in the Elongation of STS-Dependent Cell Protrusions.
Many actin-modulating proteins related to the formation of filopodia have been reported to date (dos Remedios et al., 2003). To investigate whether these proteins are involved in the elongation of STS-dependent protrusions, we knocked down some actin-modulating proteins using siRNA in HeLa cells (Fig. 5A). As shown in Fig. 5B, elongation of protrusions was significantly inhibited in Myosin-X (Myo10) (lane 3) and fascin (lane 4) knockdown cells compared with control (lane 2), while elongation was not inhibited by the knockdown of Cdc42 (lane 5), VASP (lane 6), profilin (lane 7), and cofilin (lane 8). An additive effect was observed for the inhibition of the extension of protrusions by double knockdown of Myo10 and fascin (lane 9). This effect was not significant but obviously showed a trend (P > 0.05). Interestingly, the inhibition was eliminated by double knockdown of Myo10 and Cdc42 (lane 10). No additive effect was observed for double knockdown of VASP (lane 11), profilin (lane 12), cofilin (lane 13), and Myo10. Fascin and Myo10 have been reported to localize to the entire protrusion and tip of the filopodia, respectively (Nagy et al., 2008). These results indicate that STS induces the elongation of long protrusions by promoting actin polymerization involving Myo10 and accelerating convergence of actin filaments involving fascin, but the elongation was independent of Cdc42, VASP, cofilin, and profilin.
Actin Polymerization Inhibitor Induces Elongation of Thick Cell Protrusions in the Copresence of STS.
With the expectation that the actin polymerization inhibitor cytochalasin D (CD) prevents the formation of STS-dependent protrusions, we administered STS and CD simultaneously to steady-state HeLa cells and observed the cell morphology in the living state 2 hours after administration in the presence of FM1-43 using a confocal microscope. The result showed that the cell protrusions did not disappear upon coadministration of STS and CD (Fig. 6A). No appreciable change in cell morphology was observed with CD alone (Fig. 6B). Under these experimental conditions, no significant change in the shape of the nucleus was observed. It was noteworthy that the protrusions that were induced with the combination of two compounds (Fig. 6C) were thicker and straighter than the protrusions induced by STS alone (Fig. 6D) and the steady-state cell protrusions (Fig. 6E). The apparent diameters of these protrusions were measured in confocal microscopy images (Fig. 6F). The diameter of the protrusions induced with the combination of STS and CD was significantly greater than the diameter of the protrusions induced by STS alone or in the steady state (t test; P < 0.001). The length of the protrusions induced by the combination of the two compounds was markedly longer (P < 0.001) than the steady-state cell protrusions, but significantly shorter (P < 0.05) than the protrusions induced by STS alone (Fig. 6G). Next, we observed the elongation process of these protrusions continuously in living cells. Because cell membrane dyes FM1-43 and FM4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl) pyridinium dibromide] negatively affected the efficient elongation of protrusions (unpublished observations), we observed the process in HeLa cells expressing a GFP-fusion protein of the PLC-Δ1 PH domain, which interacts specifically with cell membrane lipids (Yeung et al., 2006). As shown in Fig. 6H and Supplemental Movie 3, the protrusions elongated continuously without retraction in the copresence of STS and CD in cells expressing a GFP-fusion protein of the PLC-Δ1 PH domain, but were reduced in flexibility and the number of protrusions compared with the STS-induced ones (Fig. 1A) and terminated in a bulbous tip. These morphologic changes were also observed in the presence of the alternative actin polymerization inhibitors latrunculin A or jasplakinolide (Fig. 6, I–N). From these results, cotreatment with STS and the actin polymerization inhibitor was demonstrated to induce elongation of thick and long protrusions.
Cotreatment with STS and CD Causes Extracellular Protrusion of Markedly Long Microtubules.
We observed the cell protrusions formed in the presence of both STS and CD using TEM to clarify their constituents. As shown in Fig. 7A, these protrusions that formed on HeLa cells were suggested to contain many microtubules that were oriented parallel to the direction of the elongation (left panel, arrowheads). Interestingly, a bulbous expansion, which contains endoplasmic reticulum, ribosomes, and glycogen granules, was observed at the tip of almost all of the elongated protrusions (middle and right panels). Furthermore, these protrusions were sensitive to the microtubule polymerization inhibitor nocodazole (Fig. 7B). To identify that these protrusions are composed of microtubules, the cells were transfected to express GFP tubulin and its intracellular localization was observed in living cells with confocal microscopy. Upon cotreatment with STS and CD, GFP tubulin excessively polymerized to form robust microtubules, and one end of these microtubules protruded outside the cell (Fig. 7C). With stimulation with STS alone, GFP tubulin polymerized to form microtubules, but these did not protrude outside the cell and did not colocalize with thin actin-based protrusions (Fig. 7D). In steady-state HeLa cells, GFP tubulin showed a similar localization to endogenous tubulin (Fig. 7E). STS and CD or nocodazole did not induce appreciable changes in the shape of the nucleus within the observation period. Since the protrusions induced by cotreatment with STS and CD were also susceptible to aldehyde and alcohol fixation, as described above (Fig. 3D), we attempted to observe localization of endogenous tubulin using the TCA fixation method. However, antitubulin antibodies usable in TCA-fixed cells were not commercially available, and thus we were unable to show localization of endogenous tubulin in these protrusions by immunofluorescent staining. Next, we investigated the polarity of the extra-long MT-based protrusions. End binding 1 (EB1) is known as a microtubule plus end tracking protein (Mimori-Kiyosue et al., 2000). GFP-tagged EB1 shows an apparent comet moving on microtubules toward their plus end. Thus, we observed the dynamics of EB1-GFP to investigate the polarity of the growing MT-based protrusions. As shown in Fig. 7F and Supplemental Movie 4, a large number of comet-like signals were observed in the long elongated protrusions extending from the inside of the cell toward the outside of the cell. As we did not observe EB1-GFP migrating on protrusions toward the inside of the cell, these MT-based protrusions were shown to continuously elongate toward the outside of the cell. In contrast, since the treatment with STS alone did not extend MT-based protrusions toward the outside of the cells, migration of EB1-GFP was limited to the inside of the cell (Fig. 7G). Lastly, we investigated whether there was a correlation between the two types of protrusions elongated in an STS-dependent manner, i.e., the protrusions composed of actin filament and the protrusions composed of microtubule shafts. HeLa cells expressing EB1-GFP were prepared and treated with STS for 60 minutes, followed by additional administration of CD. These processes were monitored continuously in living cells. As shown in Fig. 7H, elongation of MT-based protrusions was initiated along with migration of EB1-GFP approximately 3 minutes after the addition of CD (arrowheads). These MT-based protrusions elongated independently of the STS-dependent actin-based protrusions (arrows). In addition, CD did not induce contraction of the actin-based protrusions that already formed. From these results, the long MT-based extracellular protrusions were formed by STS treatment under a limited G-actin supply condition.
Although STS is widely used as an apoptosis inducer, the induction of apoptosis appears to have little association with the formation of protrusions since induction of the elongation of protrusions required a sufficiently lower concentration of STS than activation of caspase3/7 in the HeLa cells used in this experiment and the treatment time was also shorter (Fig. 3A). Our result is not contradictory to a number of previous reports in terms of the time course of caspase3/7 activation by STS (Jiang et al., 2013). STS has also been reported to transiently increase the intracellular ATP level (Zamaraeva et al., 2005), but it was not reproduced in our conditions (Fig. 4F). This may be related to the Warburg effect, whereby ATP derived from the glycolytic pathway accounts for around 90% in the cells used in our experiment (Fig. 4A), which is significantly greater than ATP derived from oxidative phosphorylation. STS inhibits the function of proteins dependent on the coenzyme ATP by invading the active center as an ATP mimetic (Duda et al., 2010). In contrast to a report stating that STS facilitates actin polymerization by directly binding to the ATP-binding cassette of synapsin I (Defranchi et al., 2010), actin-based protrusions were not extended by STS in the absence of ATP (Fig. 4J). The result suggests that this molecular mechanism involves an ATP-consuming, STS-insensitive factor.
Filopodial protrusions extend against retrograde actin flow in the steady state (Yamashiro et al., 2014). Many of the cellular protrusions involved in cell migration are formed transiently and disappear subsequently. These protrusions rarely exceed 10 µm (Mogilner and Rubinstein, 2005). On the other hand, filopodial protrusions, which formed as intercellular communication pathways, are known to include long ones (Sanders et al., 2013). TNT is a known type of long-chain, actin-based protrusion (Rustom et al., 2004). TNT is a canalicular bridge formed in many cells, including neurons and immune cells, and has been reported to be involved in intercellular exchange of organelles and signaling molecules between the bridged cells (Hurtig et al., 2010). Although the mechanism of TNT formation has not been elucidated fully, Cdc42-independent and Myo10- and fascin-dependent actin-based protrusion has been suggested to be involved therein (Gousset et al., 2013). However, little has been known about the mechanism controlling the length of protrusions (Mogilner and Rubinstein, 2005). While the actin-based protrusions formed upon STS treatment were extremely long, their elongation velocity approximated the velocity of the long protrusions mentioned above (Fig. 1). Moreover, the protrusions found by Sanders et al. are susceptible to aldehyde fixation. Because the actin-derived protrusions we obtained are also susceptible to paraformaldehyde fixation, they may have morphologic similarities. Only glutaraldehyde (Figs. 3B and 7A) and TCA (Fig. 3D) allowed for fixation of cells while maintaining the protrusion morphologies in this study, but we could hardly conduct any protein localization analysis by immunofluorescence staining because these fixation methods reduced the antibody recognition ability. Indeed, there was no detectable immunofluorescence staining by using commercially available antibodies against tubulin, fascin, Myo10, myosin II, vimentin, gelsolin, cofilin, and profilin. This may be also a factor preventing analysis of the molecular mechanism.
Many of the signals related to the formation of filopodial protrusions are transmitted through activation of Rho family small GTPases (Mattila and Lappalainen, 2008). In this study, Cdc42-knockdown cells showed no influence on the elongation of STS-dependent protrusions (Fig. 5). Thus, STS is suggested to target a protein located downstream of Cdc42. N-WASP, an effector of Cdc42, recruits the Arp2/3 complex and G-actin through the VCA domain (Miki et al., 1998). The report shows that the expression of the VCA domain enhances the formation of actin polymerization nuclei without regulation by Cdc42. In the cells transfected to express the VCA domain, elongation of MT-based protrusions, but of actin-based protrusions, was induced (Supplemental Fig. 2A). Tubulin stabilization was presumably facilitated by an unknown mechanism because the VCA domain excluded G-actin, which is required for elongation of actin-based protrusions. This phenomenon was reproduced by using the G-actin–depleting agent latrunculin A or jasplakinolide (Fig. 6, I–N). The presence of an unknown factor recognizing the G-actin level and regulating MT polymerization is predicted, and identification of the factor merits further research.
The signal from small GTPases phosphorylates LIMK through phosphorylation of ROCK and PAK, leading to phosphorylation of cofilin. Dephosphorylated cofilin depolymerizes a part of the existing cytoskeleton to secure G-actin, a material for cellular protrusions. Elongation of actin-based protrusions by STS was not suppressed in HeLa cells expressing the S3E mutant of cofilin (Supplemental Fig. 2B) (Suurna et al., 2006). Because a similar result was obtained in cofilin-knockdown cells (Fig. 5), STS is strongly suggested to induce formation of protrusions in a cofilin-independent manner. Although cellular protrusions can be formed independent of cofilin in vitro (Lee et al., 2010), details of the mechanism underlying cofilin-independent formation of protrusions in live cells have not been elucidated sufficiently.
Ena/VASP is involved in the enhancement of actin polymerization and bundling (Bachmann et al., 1999). We prepared two dominant negative mutants of VASP and EVH2- and EVH1-deleted mutants as well as the constitutive active mutant S157/S239/S278/DDD (Zhuang et al., 2004), and treated the cells expressing each of these mutants with STS. However, no influence on the formation of actin-based protrusions was noted (Supplemental Fig. 2, C–E). Because a similar result was obtained in VASP-knockdown cells (Fig. 5), STS is suggested to act in a VASP-independent manner.
Formin family proteins are involved in the enhancement of linearization of AFs through a supply of G-actin by profilin (Watanabe et al., 1999). We prepared Y59A, R88L, and H133S mutants of profilin (Wittenmayer et al., 2004), expressed each of them in cells, and stimulated the cells with STS. However, STS induced elongation of actin-based protrusions in all cases (Supplemental Fig. 2, F–H). Because a similar result was also obtained in profilin-knockdown cells (Fig. 5), STS-induced elongation of actin-based protrusions is suggested to occur independently of the formin-profilin pathway. Formin family proteins have also been reported to contribute to MT stabilization (Chesarone et al., 2010). In this study, MT-based cellular protrusions were not formed even in the presence of constitutive active DIAPH1 (Supplemental Fig. 2I) (Palazzo et al., 2001). A further detailed investigation is necessary to determine whether DIAPH1 is involved in the elongation of MTs in the extracellular direction.
Bundling of AFs by fascin contributes to formation of stable filopodial protrusions, and the activity is reduced by PKC-catalyzed phosphorylation of the Ser39 residue (Vignjevic et al., 2006). In HeLa cells expressing wild-type fascin, the protrusions extended independently of STS, and fascin was localized in the protrusions (Supplemental Fig. 2J). In addition, even longer protrusions were formed upon STS stimulation. Even though the S39E mutant does not localize in the protrusions, STS stimulation caused elongation of the protrusions (Supplemental Fig. 2K). However, because knockdown of endogenous fascin led to suppression of the STS-induced elongation of protrusions (Fig. 5), STS is suggested to be partly involved in the mechanism for the suppressive activity of fascin on the elongation of the protrusion in the steady state.
Myo10 is an unconventional motor protein localized in the tips of filopodial protrusions. Although it is considered to be involved in the formation of the protrusion, the mechanism is not fully understood (Kerber and Cheney, 2011). Myo10 has also been reported to enhance formation of cellular protrusions when overexpressed (Bohil et al., 2006). The STS-induced elongation of protrusions was partially inhibited by knockdown of endogenous Myo10 (Fig. 5). This suggests that the Myo10 function is also involved in part in the elongation of STS-dependent protrusions. Interestingly, Myo10 has been reported to migrate efficiently on the actin bundle cross-linked by fascin (Nagy et al., 2008). It is suggested that an unknown mechanism is available to suppress excessive elongation of protrusions induced by these proteins, and STS acts to deactivate its function. However, identification of the factor requires further investigation.
While STS has been reported to induce formation of branched protrusions in platelets (Nakamura et al., 1995) and cultured neurons (Kitamura et al., 2003), it is unknown whether these protrusions are identical to the linear protrusions observed in the present study. Branched protrusions were not observed in HeLa, COS7, HepG2, or H1299 cells (Fig. 1, G–J). CD-dependent extension of MT-based protrusions in Drosophila S2 cells has been reported (Ling et al., 2004), but few findings regarding MTs as cellular protrusions are available. Clostridium difficile toxin induces depolymerization of actin through ADP-ribosylation and results in elongation of MT-based protrusions (Schwan et al., 2014). The morphology of protrusions, particularly a globular form at the tip of the protrusion, is similar to that observed in the present study (Figs. 6H and 7A). Although the mechanism of MT-based protrusion formation has not been elucidated fully, the interaction between the cortical actin and CLIP170/CLASP2 at the plus end of MT has been suggested to be involved in the elongation of microtubules (Plestant et al., 2014). It remains to be examined further how proteins acting on both actin-based protrusions and MTs (Dent and Gertler, 2003) and MT-associated proteins behave in the presence of STS.
The present work provided us findings suggesting that determination of the length of the protrusion is under physiologic regulation. Moreover, it is significant to have shown that even MTs are capable of extending as cellular protrusions. The fact that the regulatory factors of these protrusions were modulated by STS-susceptible kinases provides a significant advantage in further elucidation of the mechanism controlling formation and retraction of protrusions.
The authors thank Ryo Miyata, Tsuyoshi Kono, and Takuya Kakuki for helpful discussions and preliminary experiments. We also thank Mikie Kodama from the Kojima laboratory for technical assistance.
Participated in research design: Kohno, Ninomiya, Kojima.
Conducted experiments: Kohno, Ninomiya, Kikuchi, Konno.
Performed data analysis: Kohno Ninomiya, Kikuchi, Konno.
Wrote or contributed to the writing of the manuscript: Kohno, Kojima
- Received November 19, 2014.
- Accepted February 13, 2015.
This work was supported in part by grants from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. The authors declare no competing financial interests.
- actin filament
- cytochalasin D
- diphenylene iodide
- N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino) styryl) pyridinium dibromide
- N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl) pyridinium dibromide
- 2-[1-(3-dimethylaminopropyl)-1H-indol-3-yl]-3-(1H-indol-3-yl) maleimide
- green fluorescent protein
- 1-(5-isoquinolinylsulfonyl)-2-methylpiperazine dihydrochloride
- phosphate-buffered saline
- small interfering RNA
- trichloroacetic acid
- transmission electron microscopy
- tunneling nanotube
- trans-4-[(1R)-1-aminoethyl]-N-4-pyridinylcyclohexanecarboxamide dihydrochloride
- Copyright © 2015 by The American Society for Pharmacology and Experimental Therapeutics