Rab7a Regulates Cell Migration Through Rac1 and Vimentin

Rab7a regulates cell migration through Rac1 and vimentin

Azzurra Margiotta1,2, Cinzia Progida2*, Oddmund Bakke2* and Cecilia Bucci1* 1Department of Biological and Environmental Sciences and Technologies, (DiSTeBA) University of Salento, Via Provinciale Monteroni 165, 73100 Lecce, Italy 2 Department of Biosciences, Centre for Immune Regulation, University of Oslo, Blindernveien 31, 0371 Oslo, Norway

* Corresponding authors. Cecilia Bucci, Dipartimento di Scienze e Tecnologie Biologiche ed

Ambientali, Università del Salento, Via Provinciale Monteroni 165, 73100 Lecce, Italy. Tel.: +39 0832 298900; Fax: +39 0832 298626; E-mail: [email protected] Cinzia Progida and Oddmund Bakke, Department of Biosciences, Centre for Immune Regulation, University of Oslo, Blindernveien 31, 0371 Oslo, Norway, tel +47922854441. E-mail: [email protected]; [email protected]

Abstract

Rab7a, a small GTPase of the Rab family, is localized to late endosomes and controls late endocytic trafficking. The discovery of several Rab7a interacting proteins revealed that Rab7a function is closely connected to cytoskeletal elements. Indeed, Rab7a recruits on vesicles RILP and FYCO that are responsible for the movement of Rab7a-positive vesicles and/or organelles on microtubule tracks, but also directly interacts with Rac1, a fundamental regulator of actin cytoskeleton, and with peripherin and vimentin, two intermediate filament proteins. Considering all these interactions and, in particular, the fact that Rac1 and vimentin are key factors for cellular motility, we investigated a possible role of Rab7a in cell migration. We show here that Rab7a is needed for cell migration as Rab7a depletion causes slower migration of NCI H1299 cells affecting cell velocity and directness. Rab7a depletion negatively affects adhesion and spreading onto fibronectin substrates, altering β1- integrin activation, localization and intracellular trafficking, and myosin X localization. In fact, Rab7a-depleted cells show 40% less filopodia and active integrin accumulates at the leading edge of migrating cells. Furthermore, Rab7a depletion decreases the amount of active Rac1 but not its abundance and reduces the number of cells with vimentin filaments facing the wound, indicating that Rab7a has a role in the orientation of vimentin filaments during migration. In conclusion, our results demonstrate a key role of Rab7a in the regulation of different aspects of cell migration.

Keywords: Rab7a; cell migration; vimentin; Rac1; myosin X; β1-integrin.

Abbreviations: HA, Hemagglutinin; EGFP, Enhanced Green Fluorescent Protein; GST, Glutathione S-transferase; TIRF, Total Internal Reflection Fluorescence.

1. Introduction Rab7a is a small GTPase of the Rab family, ubiquitously expressed, localized mainly to late endosomes and regulating transport to endocytic degradative compartments [1, 2]. Rab7 has been called Rab7a to distinguish it from Rab7b, a Rab protein that only partially colocalizes with Rab7a and controls transport from late endosomes to the TGN [3, 4]. Rab7a is required for lysosome, phagolysosome and autolysosome biogenesis and it has been shown to be important also for channel trafficking, growth factor-independent survival and apoptosis [5-12]. A number of Rab7a effectors have being identified and characterized [13]. Among them, RILP (Rab interacting lysosomal protein) regulates microtubule minus-end directed transport with the help of ORP1L (OSBP (oxysterol-binding protein) related protein) by recruiting the dynein/dynactin complex, while FYCO1 (FYVE and coiled-coil domain containing 1) controls plus-end directed transport [14-16]. Thus Rab7a, through the interaction with different effectors, is able to recruit microtubule motors on vesicles determining vesicle movement along microtubule tracks. Recently, we discovered that Rab7a interacts also with two members of the intermediate filament proteins, vimentin and peripherin [17, 18]. We demonstrated that the interaction is direct and that Rab7a regulates vimentin and peripherin filament assembly [17, 18]. Overexpression of Rab7a wt or expression of the constitutively active Rab7a Q67L mutant protein increased vimentin phosphorylation and caused redistribution of vimentin in the soluble fraction, while, coherently, Rab7a-depletion determined a decreased phosphorylation of vimentin and increased the amount of filamentous vimentin [17, 18]. A functional direct interaction between Rab7a and Rac1, a small GTPase regulating actin cytoskeleton, has also been described [19]. The Rab7a-Rac1 interaction mediates late endosomal transport between microtubules and microfilaments in order to regulate ruffled border formation in osteoclasts but also E-cadherin turnover and stability of cell-cell contacts [19, 20]. Coordination between Rab7a and Rac1 seems also crucial in order to integrate autophagy with intracellular trafficking and signaling [19-21]. Altogether these interactions indicate a strong link between Rab7a and cytoskeletal elements suggesting that integration between cytoskeleton and late endocytic trafficking is important and is mediated by Rab7a. Moreover, both Rac1 and vimentin, two Rab7a interacting proteins, are key regulators of cell motility [22, 23]. Furthermore, recent evidence suggests that Rab7a could be involved in cell migration. In fact, it has been demonstrated that the Rab7a-dependent lysosomal degradation affects the final phase of migration of immature neurons during the development of

3 cerebral cortex in vivo [24, 25]. However, how and to which extent Rab7 is involved in the regulation of cell migration it is not known. Therefore, in this study, we investigated the role of Rab7a in cell migration. We established that Rab7a depletion strongly alters cell motility affecting cell velocity and cell directness. In particular, we demonstrated that Rab7a interferes with cell adhesion, by altering β1-integrin activation, distribution and trafficking and decreasing Rac1 activation. In addition, Rab7a depletion affects filopodia formation, by mislocalizing myosin X, and alters vimentin organization during cell migration. In conclusion, Rab7a works as regulator of migration acting on the fundamental players of this process.

2. Materials and methods

2.1. Cells and reagents NCI H1299 cells (ATCC CRL-5803; human lung carcinoma) were grown in DMEM supplemented with 10% FBS, 2 mM L-glutamine, 100 U/ml penicillin and 10 mg/ml streptomycin in an incubator at 37°C, under 5% CO2. Cells were confirmed to be contamination-free. Chemicals were from Sigma-Aldrich (St Louis, MO, USA) and tissue culture reagents were from Sigma-Aldrich, Gibco (Waltham, MA, USA), Lonza (Basel, Switzerland) and Biological Industries (Beit-Haemek, Israel).

2.2. Constructs and antibodies pEGFP-C1, pEGFP-Rab7a T22N, pEGFP-Rab7a, pCDNA3_2XHA-Rab7a wt, pCDNA3_2XHA- Rab7a T22N, pCDNA3_2XHA-Rab7a Q67L and mCherry-Rab7a have been described previously [4, 10, 26]. pHcEGFP-β1-integrin [27], pEGFP-Myosin X [28], and pEGFP-Rac1 [29] were a kind gift from Martin J. Humphries, Richard E. Cheney, and Giorgio Scita, respectively. mApple- Integrin-Beta1-N-18 was a gift from Michael Davidson (Addgene plasmid # 54914). Primary antibodies used in this study were rabbit polyclonal anti-HA (1:500, ab9110) and rabbit polyclonal anti-giantin (1:1000, ab24586) from Abcam (Cambridge, UK); mouse monoclonal anti- Rac1 (1:500, ARC03) from Cytoskeleton (Denver, CO, USA); mouse monoclonal anti-β1-integrin (1:100, sc18887), mouse monoclonal anti-active β1-integrin (1:50 for immunofluorescence analyses or 1:3000 for immunoblot analyses, sc59827), goat polyclonal anti-EEA1 (1:50 for immunofluorescence analyses or 1:100 for immunoblot analyses, sc6415), rabbit polyclonal anti- Rab4 (1:200, sc312), rabbit polyclonal anti-Rab5 (1:200, sc309), goat polyclonal anti-Rab11 (1:200, sc6565), mouse monoclonal anti-rab7a (1:500, sc376362) and mouse monoclonal anti- 4 vimentin (1:100, sc6260) from Santa Cruz Biotechnology (Santa Cruz, CA, USA); rat monoclonal anti-inactive β1-integrin (1:1000, 552828) from BD Biosciences (East Rutherford, NJ, USA); rabbit polyclonal anti-myosin X (1:100, 22430002) from Novus biologicals; mouse monoclonal anti- Rab7b (1:300, H00338382-M01) from Abnova (Taipei City, Taiwan); mouse monoclonal anti- tubulin (1:10000, T5168) was from Sigma-Aldrich. Secondary antibodies conjugated to fluorochromes (1:200 for immunofluorescence analyses) or horseradish peroxidase (HRP, 1:5000 for immunoblot analyses) were from Invitrogen (Carlsbad, CA, USA) or Santa Cruz Biotechnology.

2.3. Transfection and RNA interference Transfection was performed using Metafectene Pro from Biontex (Martinsried, Germany) or Lipofectamine 2000 from Invitrogen according to the protocol provided by the manifacturer. The cells were then analysed after 24 hours of transfection. Transfection of cells with siRNA was performed using RNAiMAX from Invitrogen or Metafectene SI from Biontex following the manufacturer's instructions. siRNAs were purchased from MWG-Biotech (Ebersberg, Germany). Rab7a siRNA efficiency in silencing was reported previously [26]: sense sequence 5'- GGAUGACCUCUAGGAAGAATT-3' and antisense sequence 5'- UUCUUCCUAGAGGUCAUCCTT-3'. Control RNA was used as a negative control: sense sequence 5'-ACUUCGAGCGUGCAUGGCUTT-3' and antisense sequence 5'- AGCCAUGCACGCUCGAAGUTT-3'.

2.4. Wound-healing assay Confluent monolayers of NCI H1299 cells were scratched by a pipette tip. Cells were imaged for a period of 9 hours with a 20X objective on an Olympus Fluoview 1000 IX-81 inverted confocal laser scanning microscope. During the observation cells were maintened in DMEM without phenol red at

37°C under 5% CO2. Images were acquired every 30 minutes. Cell nuclei were tracked by using the Manual Tracking plugin of ImageJ software (National Institutes of Health) and cell velocity and directness were calculated by using the Chemotaxis and Migration Tool software (Ibidi).

2.5. Cell adhesion assay Control and transfected cells were seeded onto fibronectin-coated coverslips (BD Biosciences), fixed 30 minutes later with 3% paraformaldehyde, permeabilized with 0.25% saponin in PBS and stained with Rhodamine-conjugated phalloidin, Hoeschst 33258 and anti-HA antibody. Coverslips were then mounted and examined using an Olympus FluoView FV1000 microscope. Cell areas were determined by using ImageJ software (National Institutes of Health).

2.6. Golgi reorientation measurements Wound-healing assays were performed by scratching confluent monolayers of cells seeded on round coverslips. After 3 hours of incubation at 37°C under 5% CO2 cells were fixed as previously described [30] and stained with Rhodamine-conjugated phalloidin, anti-giantin antibody and Hoeschst 33258. Cells were viewed with a 63X PlanApo NA 1.42 objective on an Olympus FluoView FV1000 microscope with the FV1000 software. Cells with the Golgi in the 1/3 of the cell facing the wound were considered to be polarized.

2.7. Vimentin reorientation measurements NCI H1299 cells were grown on coverslips. Scratches at different time points were performed in each sample. After three hours from the first scratch cells were permeabilized and fixed as described previously [30]. Cells were incubated with primary antibodies for 20 minutes at room temperature. After washing with 0.25% saponin, cells were then incubated with secondary antibodies for 20 minutes in the dark at room temperature. Coverslips were mounted in mowiol. Cells were viewed with a 63X PlanApo NA 1.42 objective on an Olympus FluoView FV1000 microscope with the FV1000 software. Vimentin filaments in the area between the nucleus and the leading edge of migrating cells were considered to be oriented. The percentage of cells which showed oriented vimentin filaments at the final and the initial time points was plotted in the graph.

2.8. TIRF microscopy For total internal reflection fluorescence microscopy NCI H1299 cells were grown on MatTek glass-bottomed dishes (Ashland, USA), fixed with 3% paraformaldehyde and permeabilized with 0.2% Triton-X-100 in PBS for 15 minutes as previously described [31]. Cells were then incubated with antibodies and Rhodamine-conjugated phalloidin for 20 minutes at room temperature. After washing, cells were then incubated with secondary antibodies for 20 minutes in the dark at room temperature. Cells were observed with a Leica DMI600B system with a 63x objective (NA 1.47). The TIRF evanescent field depth was set to 110nm for all the excitation wavelengths.

2.9. EEA1 and active β1-integrin colocalization assay NCI H1299 were transfected with scrambled and Rab7a siRNA. The cells were then fixed and permeabilized as described previously [32]. Briefly cells were fixed with 3% paraformaldehyde, permeabilized with 0.1% saponin and stained for active β1-integrin and EEA1. Appropriate secondary antibodies were used. Coverslips were then mounted and observed with a Zeiss LSM 700

6 confocal microscope. Weighted colocalization coefficient of active β1-integrin and EEA1 was calculated.

2.10. Live cell microscopy NCI H1299 cells were grown on MatTek glass-bottomed dishes, washed with PBS and then incubated in DMEM without phenol red. Live cells were maintened at 37°C under 5% CO2 and imaged with a 63X PlanApo NA 1.42 objective on an Olympus IX-71 microscope equipped with a CSU22 spinning-disk confocal unit (Yokogawa), an Ixon EMCCD camera (Andor) and the Andor iQ1.8 software.

2.11. Western blotting NCI H1299 cells were lysed with RIPA buffer (50mM Tris-HCl, pH 8.0, with 150mM sodium chloride, 1.0% Igepal CA-630 (NP-40), 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate) plus protease inhibitor cocktail (Roche, Mannheim Germany). Proteins were separated using SDS-PAGE, transferred onto polyvinylidene fluoride (PVDF) membrane from Millipore (Billerica, MA) and probed with the specific primary antibodies diluted in 2% milk in PBS followed by HRP-conjugated secondary antibodies. In the case of β1-integrin abundance, western blot analysis was conducted under non-reducing conditions. Bands were visualized by using western blot Luminol Reagent (Santa Cruz) or WesternBright ECL kit (Advansta, Menlo Park, CA, USA).

2.12. Rac1 activity assay Rac1 activity was tested following the manifacturer's protocol for Rac1 Pull-down Activation Assay Biochem Kit (BK035, Cytoskeleton). Briefly, GTP-bound Rac1 was immunoprecipitated from cell lysates with GST-Pak-PBD using glutathione affinity beads. After washing, the beads were subjected to western blot analysis using anti-Rac1 antibody to detect GTP-bound Rac1. Total amount of Rac1 was detected by immunoblotting of the whole cell lysates.

2.13. Cell proliferation and viability assays NCI H1299 cells were treated with either control or Rab7a siRNA, plated out at equal confluence and grown for 24 hours. Cells were then collected and stained with trypan blue dye at different time points over a 9 hour time period. The number of cells and the percentage of living cells were calculated. Quantification of cell viability was carried out using a tetrazolium salt (MTT (3-(4,5-

7 dimethylthiazolyl-2)-2,5-diphenyltetrazolium bromide)) based colorimetric assay, which is based on the mitochondrial-dependent conversion of the yellow tetrazolium salt to purple formazan crystals by metabolically active cells. Briefly, 100 µl of MTT solution (0,5 mg/mL) was added to the cells after 24 hours of treatment with control RNA or Rab7a siRNA. Cell viability has been evaluated at different time points. Cells were then re-incubated for 3 hours. Finally, 100 µl of

DMSO was added to each well and absorbance was measured at 570 nm.

2.14. Statistical analysis Data were statistically analysed using Student's t-test (GraphPad Prism4 software) (*p<0.05, **p<0.01 and ***p<0.001). Error bars represent s.e.m. Experiments were performed at least in triplicate.

3. Results 3.1. Depletion of Rab7a delays wound closure In order to investigate if and how Rab7a could affect cell migration we performed a wound- healing assay on confluent monolayers of NCI H1299 cells transfected with control RNA or siRNA against Rab7a in order to deplete cells from Rab7a. Transfected monolayers at confluence were scratched and the cells migrating toward the wound were imaged at time intervals of 30 min over a 9 hours time period. Quantitative analysis of velocity, directness and Euclidean distance have been calculated by tracking cell nuclei by using the Manual Tracking plugin of Image J (National Institutes of Health) and by analysing the migratory tracks with the Chemiotaxis and Migration tool software (Ibidi). Datasets from the Manual Tracking plugin of Image J were imported into the Chemiotaxis and Migration tool software and settings relative to the number of slices composing the movie, the imaging time interval and the x/y pixel size were applied. Each cell trajectory was analysed for accumulated distance (which represents the cell path), Euclidean distance (length of the straight line between cell start and end point), directness (Euclidean distance/Accumulated distance ratio), and cell velocity (Accumulated distance/time). These parameters have been calculated through the indicated software as described before [33-35]. Interestingly, Rab7a-depleted cells migrated slower than control cells (Supplementary Movie S1; Fig. 1A). In support of this, other migration parameters such as accumulated distance, Euclidean distance and cell velocity were strongly decreased compared to control cells (Fig. 1B-D). Velocity and accumulated distance were decreased of ~40% in Rab7a-depleted cells compared to control cells, whereas Euclidean distance 8 was ~50% lower upon silencing of Rab7a (Fig. 1B-D). Importantly, also directness, which measures the ability of a cell to move persistently along its migratory path, was affected in cells lacking Rab7a with a reduction of about 20% (Fig. 1E). Silencing of Rab7a was checked by western blot analysis and a strong reduction of Rab7a abundance of ~70% was observed upon Rab7a depletion (Fig. 1F-G). To exclude possible off-target effect, Western blot analysis of control and Rab7a-depleted cells, using antibodies against Rab proteins that share high sequence identity with Rab7, was performed, demonstrating that silencing of Rab7a did not affect the abundance of Rab4, Rab5, Rab7b, Rab9 and Rab11 (Supplementary Fig. S1). In order to exclude a possible effect of Rab7a depletion on cell viability and proliferation that could affect migratory parameters, we analysed these processes in control and Rab7a-depleted cells over the nine hour time period of the wound-healing experiment. We observed that cell proliferation was not altered upon Rab7a silencing (Supplementary Fig. S1). Moreover, the percentage of living Rab7a-depleted cells was similar to that of control cells during the monitored time interval (Supplementary Fig. S1). Viability of Rab7a-depleted cells, relative to that of control cells, was also tested by MTT-based assay and was similar over the nine hour time period (Supplementary Fig. S1). Our data of Rab7a involvement in cell migration was confirmed by experiments on the migration of individual cells (Supplementary Fig. S2; Supplementary Movie S2). Sparse cultures of control and Rab7a-depleted cells have been imaged every 15 minutes over a 12 hour time period (Supplementary Fig. S2; Supplementary Movie S2). In line with the results of the wound-healing experiments, velocity, Euclidean distance and accumulated distance were strongly reduced upon Rab7a silencing (Supplementary Fig. S2). Conversely to wound-healing assay where cells are stimulated to migrate in a specific direction in order to close the wound, individual cells in sparse cultures migrate in random direction. In line with this, directness of control and Rab7a-depleted cells was similar in sparse cultures (Supplementary Fig. S2). We then evaluated a possible role of Rab7a on cell polarization, an event happening during migration, by measuring the orientation of the Golgi apparatus into the direction of the wound during migration. In the majority of Rab7a-depleted cells the Golgi apparatus was correctly oriented similarly to control cells, thus depletion of Rab7a did not interfere with Golgi orientation (Supplementary Fig. S3). These data demonstrate that Rab7a is required for proper cell motility but not for Golgi orientation.

3.2. Loss of Rab7a reduces cell spreading 9

Having observed that Rab7a affects cell motility, we further investigated its role in this process. We first analysed whether Rab7a is important for cell spreading, a key step for cell migration. In particular, we evaluated the ability of control and Rab7a-depleted cells to spread out on fibronectin-coated coverslips in 30 min looking at the actin cytoskeleton stained by rhodamine- phalloidin. Interestingly, as shown in Fig. 2, Rab7a-depleted cells showed lower ability to spread compared to control cells. In particular, the average cell area of Rab7a-depleted cells was of 90 µm2, about 60% lower than the average cell area of control cells, which showed an average cell area of ~220 µm2 (Fig. 2B). To confirm that the inability of Rab7a-depleted cells to spread onto fibronectin-coated dishes was specifically due to Rab7a depletion, we also performed rescue experiments. Transient expression of HA-tagged Rab7a wt in Rab7a-depleted cells was sufficient to re-induce, although partially, cell spreading onto fibronectin while it did not change spreading of control cells (Fig. 2B). Indeed, the average cell area was reduced of only 35% in cells silenced for Rab7 and expressing HA-tagged Rab7 wt resulting in a 50% rescue. Analysis of untransfected and transfected cells within the same sample confirmed the results. Furthermore, 50% recovery of spreading was detected also in Rab7a-silenced cells expressing HA-tagged Rab7a Q67L mutant (Fig. 2C), confirming the role of Rab7a in spreading (Fig. 2). In contrast, as expected, expression of the dominant negative Rab7a T22N mutant in Rab7a-depleted cells was not able to recover spreading onto fibronectin (Fig. 2C). These data demonstrate that Rab7a is important for cell spreading to fibronectin.

3.3 Rab7a alters filopodia formation and β1-integrin localization and activation Integrins are heterodimeric receptors with a main role in cell adhesion and spreading [36]. In order to exert this function, integrin localization at protrusions at the leading edge is essential [37-39]. As we have demonstrated that Rab7a depletion impairs cell spreading to fibronectin, we next investigated whether there is a link between Rab7a and integrin. In live cells co-transfected with GFP-β1-integrin and mCherry-Rab7a we could observe Rab7a vesicles positive for β1- integrin, suggesting that intracellular transport of β1-integrin is mediated by Rab7a (Fig. 3A; Supplementary Movie S3). β1-integrin localizes both intracellularly and at the plasma membrane [31]. However, its activity and function are dependent on the conformation of the receptor and the localization of active and inactive β1-integrin is different [31]. In fact, inactive β1-integrin is mainly localized at cell surface, where it is activated [31]. Active β1-integrin binds ligands of the extracellular matrix, it is subsequently internalized, inactivated and recycled contributing to cell migration [31, 40]. As a consequence, most of active β1-integrin is usually intracellular [31, 40].

Therefore, we investigated the localization of the active and the inactive forms of β1- integrin in control and Rab7a-depleted cells during migration using specific antibodies able to recognize and discriminate the active and inactive forms of β1-integrin, by total internal reflection fluorescence (TIRF) microscopy. TIRF analysis is particularly suitable for studying the localization of molecules near the plasma membrane. Indeed, it allowed us to focus on β1-integrin localization at the plasma membrane in migrating cells. The ability of monoclonal anti-β1-integrin antibodies to detect conformation-dependent epitopes has been previously defined [31, 41]. In addition, we also stained cells with rhodamine-conjugated phalloidin in order to visualize filopodial protrusions, which are important during cell migration to explore the surrounding environment by making new transient cell-extracellular matrix adhesions and where integrins can localize. Surprisingly, TIRF analysis of cells stained for actin showed that Rab7a depletion strongly affected filopodia formation (Fig. 3B). We observed 64% of control cells having filopodia, whereas only 37% of cells silenced for Rab7a had filopodia. Thus, Rab7a depletion caused a ~40% reduction of cells with filopodia at the leading edge compared to control cells (Fig. 3C). Furthermore, we detected an increase of about 80% of the number of Rab7-depleted cells with the active form of β1-integrin at the level of the plasma membrane facing the wound, compared to control cells (Fig. 3D). Higher integrin activation at the leading edge could prevent migration due to the inability to form new adhesions through integrin recycling. Phase contrast images of migrating control and Rab7a-depleted cells stained for actin confirmed that Rab7a silencing impairs filopodia formation (Supplementary Fig. S4). We verified that the actin structures visible in control cells were filopodia by looking at the localization of myosin X, a motor protein typically localized at filopodial tips [42], and at the ability to extend and retract of such structures (Supplementary Fig. S5). Given that only active β1-integrin associates with Rab7a-positive vesicles [31] and that we saw an altered distribution of this form of integrin, we further investigated if Rab7a affected β1- integrin activation using specific monoclonal antibodies against active β1-integrin in western blot experiments under non-reducing conditions. While the total amount of integrin was similar in control and Rab7a-depleted cells, Rab7a-depletion caused a statistically significant decrease of active β1-integrin (Fig. 3E-F). In fact, the amount of active β1-integrin was 25% lower in cells silenced for Rab7a compared to control cells (Fig. 3F). Usually inactive β1-integrin is mostly localized at the plasma membrane while active β1- integrin seems to be more cytoplasmic [43]. Therefore, we evaluated the abundance of active β1- integrin in the cytoplasm of NCI H1299 cells transfected with control RNA or siRNA against Rab7a. We found that active β1-integrin was less abundant in the cytoplasm after Rab7a depletion but the staining was more vesicular (Fig. 4A-B) and colocalization with the early endosomal marker

EEA1, whose abundance is unchanged after Rab7a depletion, increased (Fig. 4C-F). In fact, there was a two times higher colocalization of active β1-integrin with EEA1 in Rab7a-depleted cells compared to control cells (Fig. 4D). We confirmed this result by the expression of the dominant negative mutant of Rab7a, Rab7a T22N, which also increased the colocalization between EEA1 and active β1-integrin of about twice (Supplementary Fig. S6). All together these data show that Rab7a is essential for filopodia formation, proper β1- integrin localization, trafficking and activation, and consequently cell spreading.

3.4. Rab7a alters myosin X localization to filopodia tips As TIRF analysis showed that Rab7a depletion inhibits filopodia formation, we next investigated how Rab7a affects formation of protrusions. Myosin X is an actin motor known to promote filopodia formation [44, 45]. In order to understand whether the decrease in the number of filopodia observed after Rab7a depletion is related to myosin X function, we performed live imaging experiments in NCI H1299 cells transfected with mCherry-Rab7a and GFP-myosin X. GFP-myosin X is mostly localized to filopodia tips. However, a small fraction of myosin X localizes to late endosomes/lysosomes and this fraction increases after Rab7a overexpression [46]. In agreement with that, we could detect myosin X also intracellularly on Rab7a positive endosomes (Fig. 5A). Interestingly, these endosomes transported myosin X to the cell periphery, where Rab7a was released (Fig. 5A, Supplementary Movie S4). This observation prompted us to investigate whether the decreased number of filopodia in Rab7a-depleted cells was a consequence of altered myosin X transport. We therefore looked at GFP-myosin X localization in cells silenced for Rab7a. Strikingly, while in about 80% of the control cells myosin X localized mostly at the filopodia tip, in 50% of the cells treated with siRNA against Rab7a, myosin X was predominantly intracellular (Fig. 5B-C). We excluded that the increase of intracellular myosin X after Rab7a depletion was caused by an increase in the total amount of myosin X by western blot of either overexpressed (Fig. 5D-E) or endogenous myosin X (Fig. 5F-G) which showed that Rab7a silencing did not affect myosin X abundance. As β1-integrin is important for cell adhesion and spreading and it is transported to filopodia tips by myosin X [45], two processes affected by Rab7a depletion, we next investigated the intracellular localization of GFP-myosin X relative to β1-integrin in cells knocked down for Rab7a. Intriguingly, we found that the myosin accumulated intracellularly in endosome positive for β1- integrin (Fig. 5H, Supplementary Movie S5).

These results indicate that Rab7a is important for proper filopodia formation possibly by regulating myosin X-dependent transport towards filopodia tips.

3.5. Loss of Rab7a influences Rac1 activation state Our data show that Rab7a interferes with β1-integrin transport and activation. Interestingly, Rab7a interacts with Rac1, a regulator of actin cytoskeleton and cell migration, and Rac1 activation is regulated by integrin β1 [19, 47, 48]. Therefore, we investigated if, in our system, alterations of Rab7a expression affected Rac1 abundance or activation. We first analysed whether modulation of Rab7a expression levels affects Rac1 abundance but we couldn’t detect any differences in Rac1 protein amount upon overexpression or silencing of Rab7a (Fig. 6A). Then, using live microscopy we established that Rac1 and Rab7a co-localize on the same vesicles and move together in the cell (Fig. 6B-C; Supplementary Movie S6). Furthermore, Rab7a co-localized with Rac1 in proximity of the plasma membrane and Rac1 localized also at filopodia (Fig. 6B-C; Supplementary Movie S6). In β1-integrin-null nerves and Schwann cells Rac1 activation is impaired [49, 50] and the abundance of Rac1-GTP in Schwann cells seems to be dependent on the binding of β1-integrin with an extracellular matrix protein [50]. Therefore, we evaluated whether Rab7a could affect Rac1 activation through the regulation of β1-integrin trafficking. In Rab7a-depleted cells the total amount of Rac1 was unchanged compared to control cells, but we detected lower levels of Rac1-GTP (Fig. 6D). Indeed, the amount of active Rac1 decreased of about 65% upon Rab7a-depletion (Fig. 6E). These data demonstrate that Rab7a colocalizes with Rac1 on vesicles near the plasma membrane and in the perinuclear region and depletion of Rab7a strongly inhibits Rac1 activation. All together our results indicate that Rab7a regulates cell migration through its roles in cell adhesion and filopodia formation by modulating integrins and Rac1 activity and transport.

3.6. Rab7a regulates vimentin filaments orientation during cell migration Several reports indicate a close link between Rac1 and vimentin. In fact, vimentin regulates Rac1 activation while activation of Rac1 induces vimentin phosphorylation and disassembly [51- 54]. Furthermore, we have previously shown that Rab7a interacts with vimentin and that depletion of Rab7a determines a lower phosphorylated state of the head domain of vimentin and a higher abundance of vimentin in the insoluble, filamentous fraction [17]. As both Rab7a and Rac1 regulate vimentin assembly and since it has previously been shown that vimentin filaments are oriented towards the leading edge during migration [55], we tested whether vimentin filament orientation was impaired in Rab7a-depleted cells in a wound healing assay. In control migrating cells there was about 50% of cells showing vimentin facing the wound, whereas after 3 hours about 85% of cells

13 had vimentin oriented towards the leading edge. Therefore, there was an increment of about 60% of the number of cells with vimentin facing the wound compared to the cells before migration. Interestingly, this value was much lower in Rab7a-depleted cells, where only 66% of migrating cells (after 3 hours) showed oriented vimentin compared to the 55% of Rab7a-depleted cells at the initial time point (0 hours). The increment of the number of cells with vimentin facing the wound was ~ 25% in cells silenced for Rab7a (Fig. 7A-B). In addition, not only vimentin but also about 80% of Rab7a-positive vesicles in migrating cells expressing GFP-Rab7a were localized between the nucleus and the leading edge, facing the wound (Fig. 7C). These data demonstrate that Rab7a is important for vimentin orientation during migration.

4. Discussion It has been demonstrated that Rab7a directly interacts with important regulators of cell motility such as Rac1, a small GTPase controlling actin cytoskeleton and signal transduction, and vimentin, an intermediate filament protein [17, 19]. Furthermore other data indicate that, during development of cerebral cortex, lysosomal degradation is controlled by Rab7a influencing the final phase of migration of immature neurons [24, 25]. In particular, the distances from the nuclei to the pial edge of the cortical plate were increased when Rab7a was lacking or Rab7a T22N was expressed [24, 25]. Based on this, we investigated further whether Rab7a indeed could regulate cell migration. Our findings revealed that Rab7a is fundamental for cell movement as Rab7a-depleted cells migrated slower, and migration parameters, such as accumulated distance, euclidean distance, cell velocity and directness, were altered (Supplementary Movie S1; Fig. 1A-E). Furthermore, our data indicate that Rab7a is essential for cell migration as Rab7a is required for trafficking and activity of key regulatory molecules of different steps of cell migration. First of all we showed that Rab7a is required for cell spreading as its depletion affected spreading on fibronectin (Fig. 2). This finding is not surprising considering that Rab proteins are responsible for trafficking of molecules and receptors important for cell adhesion and spreading and, indeed, a number of other Rabs are involved in these functions. For instance, overexpression of Rab2 caused increased adhesion of neurons that, in turn, increased strongly neurite outgrowth, while Rab11-regulated recycling of E-cadherin impacts on cadherin based adhesion [56, 57]. Also Rab7b, regulating actin remodeling, and Rab10, controlling endocytosis of a hyaluronan synthase, influence strongly cell adhesion and spreading on different extracellular matrix [58, 59]. Thus Rab7a, regulating trafficking of adhesion molecules and receptors, similarly to other Rabs, influences cell spreading. However, the mechanism by which Rab7a regulates adhesion molecules 14 should be different as Rab7a is involved in transport to degradative compartments such as late endosomes and lysosomes. Therefore, in order to understand the molecular mechanism underlying this regulation, we looked at key players in spreading and filopodia formation. The production of filopodia at the leading edge to explore the surrounding area is a fundamental step in migrating cells. Formation of filopodia is mediated by active Rac1 that controls actin cytoskeleton rearrangements and integrins, cell surface receptors involved in adhesion, spreading and migration, are transported to filopodia tips by myosin X, an unconventional myosin [45, 60]. Interestingly, we established that Rab7a depletion strongly reduces formation of filopodia (Fig. 3C), affecting myosin X localization (Fig. 5B-C). Our findings corroborate and are consistent with previous data demonstrating that filopodia’s number and length are directly regulated by myosin X [46]. Indeed, it was suggested that myosin X could be transported to the cell membrane via Rab7a-positive vesicles, where the machinery responsible for filopodia formation might be assembled, and that these vesicles could be responsible for targeting myosin X to the correct sites of filopodia during their formation [46]. In fact, we observed colocalization of myosin X on Rab7a- positive vesicles that loose their Rab7a coat once approaching cell periphery and established that, in Rab7a-depleted cells, myosin X is not anymore localized to the filopodia tips (Fig. 5A-C), demonstrating a key role of Rab7a in myosin X trafficking and targeting to filopodia and, thus, in the formation of filopodia. Filopodia used to probe the environment in migrating cells are usually rich of integrins. Integrins are fundamental molecules in cell migration and we demonstrated that Rab7a depletion affects β1-integrin activation and trafficking. In fact, our data confirmed that Rab7a colocalizes with β1-integrin (Fig. 3A), as previously shown [31], and indicated that Rab7a depletion decreases the total amount of active β1-integrin (Fig. 3E-F), causes loss of β1-integrin rich filopodia (Fig. 3B- C) and induces localization of active β1-integrin at the leading edge of migrating cells (Fig. 3D). Activation of integrins is regulated by a conformational switch and this is an important mechanism in order to control spatially and temporally ligand affinity states of integrins, hence regulating their function [61]. Indeed, inactive integrins exist in a bent form having low affinity for ligands while active integrins exist in an extended form displaying intermediate and high affinity for ligands [61]. Important sites of activation are represented by protrusions at the leading edge, filopodia and lamellipodia, where inactive integrins are targeted before being activated in order to contribute to adhesion, spreading and thus to cell motility [37-39]. Rab7a depletion does not alter the total amount of β1-integrin but decreases the amount of active β1-integrin (Fig. 3E-F), possibly because formation of filopodia is inhibited by blockage of delivery of myosin X (Fig. 5A-C) and, therefore, integrins cannot reach filopodia, an important site of activation, and this could result in a

15 general decrease of β1-integrin activation. In line with this, we observed an intracellular accumulation of myosin X in vesicle positive for β1-integrin in cells depleted for Rab7a (Fig. 5B- H). Our data show that Rab7a is essential for proper β1-integrin trafficking and, as a consequence, active Rab7a is needed for formation of integrin-rich filopodia. Previously, it has been demonstrated that β1-integrin trafficking and functions in migration was dependent on the activity of several other Rabs such as Rab1, Rab4, Rab5, Rab11, Rab21 and Rab25 [62-67]. Rab1 regulates trafficking from ER to Golgi and between Golgi compartments and it controls β1-integrin localization in lipid raft affecting β1-integrin recycling [66]. Rab4, Rab5, Rab11, Rab21 and Rab25 work in the early steps of endocytosis, regulating transport to early sorting endosomes or recycling to the plasma membrane thus regulating the continuous internalization and recycling of β1-integrins necessary for cell migration [62-65, 68]. Similarly to Rab11, Rab25 is present on recycling endosomes and on the trans Golgi network controlling integrin recycling [68], but it has also been demonstrated that Rab25 collaborates to promote α5β1 recycling of active integrins from late endosomes and lysosomes [69]. Rab7a, instead, controls late endocytic trafficking being important for transport from early endosomes to late endosomes and lysosomes and for endosome maturation [10, 70, 71]. Indeed, we detected an accumulation of integrins at the level of early endosomes after Rab7a depletion (Fig. 4C-F). This could, in turn, lead to increase recycling of active integrins to the plasma membrane via a short loop, mediated by Rab4, and/or via perinuclear recycling compartment mediated by Rab11 or Rab25, explaining the increased amount of active integrins present at the leading edge upon Rab7a depletion (Fig. 3D). The increase of active β1-integrin at leading edge (measured by TIRF microscopy, Fig. 3B) and on early endosomes (measured by co-localization of EEA1, Fig. 4C) in Rab7a-depleted cells is accompanied by a decrease of the total amount of active integrin (measured by Western blot, Fig. 3E) and also by a decreased staining of active integrin in the cytoplasm (measured by immunofluorescence, Fig. 4A) further corroborating these data and suggesting that Rab7a is fundamental for active β1-integrin localization and trafficking. It has been demonstrated that active integrins, after initially activating Rac1 probably through a GEF (Guanine nucleotide Exchange Factor), subsequently recruit a complex involving a GAP (GTPase-activating protein) which down-regulates Rac1 activity in order to limit protrusive activity during cell migration [60, 72]. Accordingly to these and also more recent findings [73], our data demonstrate that Rab7a depletion determines the accumulation of active β1-integrins at the leading edge of migrating cells (Fig. 3D) and causes a strong decreases in the amount of active Rac1 (Fig. 6D-E). 16

The creation of a polarized network of intermediate filaments coordinated with microtubules is fundamental for cell migration [55]. Intermediate filament assembly is dependent on their phosphorylation state. In fact, unphosphorylated vimentin is filamentous and phosphorylation is important to induce depolymerization, which is fundamental for filament reorganization. Importantly, the small GTPase Rab7a regulates the assembly of vimentin affecting also its phosphorylation state [17, 18, 74]. Interestingly, in migrating cells, Rab7a-positive vesicles localize in the perinuclear region facing the wound while vimentin filaments are reorganized and reoriented in the direction of the wound (Fig. 7A-C), in line with what has been previously demonstrated [55]. In contrast, Rab7a depletion in migrating cells causes a decrease in the number of migrating cells with vimentin oriented in the direction of the wound (Fig. 7A-B) possibly due to a reduced ability of the cell to reorganize vimentin cytoskeleton because of the lack of Rab7a. These data further reinforce the role of Rab7a in migration and explain previous data indicating that Rab7a depletion causes a decrease, but not a complete inhibition, of vimentin phosphorylation at various sites accompanied by an increase of insoluble (filamentous) vimentin [17]. In support of this, it has been previously demonstrated that alterations of cytoskeleton components dynamics affect cell migration [75]. Notably, Rac1 and vimentin functions are connected [51-54]. Indeed, Rac1 regulates reorganization of vimentin filaments, as Rac1 activation causes phosphorylation of vimentin at Ser38 inducing vimentin filament disassembly [52-54]. On the other hand, vimentin controls Rac1 activation acting on a Rac1 exchange factor [51]. Thus, in Rab7a-depleted cells, Rac1 activation is impaired (Fig. 6D-E) and this, in turn, affect vimentin phosphorylation and assembly, altering vimentin filament reorganization and reorientation at the leading edge, thus impairing cell migration. Therefore, our data suggest that Rab7a could influence vimentin phosphorylation and assembly acting on the Rac1 GTPase. In addition, the localization of Rab7a in the perinuclear region facing the wound during migration is in agreement with the findings that Rab7a regulates the trafficking and localization of β1-integrin and myosin X at the leading edge (Fig. 3A-D; Fig.5). The effects of Rab7a on trafficking of myosin X and integrins, on activation of integrins and Rac1 and on reorganization of vimentin (Fig. 8) revealed a fundamental role of this small GTPase in the regulation of cell migration.

Competing interests The authors declare that they have no conflicting financial interest.

Transparency document The transparency document associated with this article can be found in online version. 17

Acknowledgements We thank Giorgio Scita (IFOM, Milan, Italy), Richard E. Cheney (University of North Carolina, Chapel Hill, USA) and Martin J. Humphries (University of Manchester, UK) for the kind gift of the pEGFP-Rac1, pEGFP-Myosin X and pHcEGFP-β1-integrin plasmids, respectively.

Funding The financial support of AIRC (Associazione Italiana per la Ricerca sul Cancro, Investigator Grant 2013 N. 14709 to C.B.), MIUR (Ministero dell'Istruzione, dell'Università e della Ricerca, PRIN2010-2011 to C.B.), Telethon Italy (Grant GGP16037 to C.B.), Regione Puglia (SISTEMA Project T7WGSJ3), the Norwegian Cancer Society (grants 5760850 to C.P. and 4604944 to O.B.), and the Research Council of Norway (grants 239903 to C.P., 230779 to O.B., and through its Centre of Excellence funding scheme, project number 179573) is gratefully acknowledged. We also thank the NorMIC Oslo imaging platform, Department of Biosciences, University of Oslo and the 2HE-PONa3_00334 project for the Zeiss LSM700 confocal microscope.

References

[1] S.R. Pfeffer, Rab GTPase regulation of membrane identity., Curr Opin Cell Biol, 25 (2013) 414-419. [2] M. Zhang, L. Chen, S. Wang, T. Wang, Rab7: roles in membrane trafficking and disease, Biosci Rep, 29 (2009) 193-209. [3] C. Progida, L. Cogli, F. Piro, A. De Luca, O. Bakke, C. Bucci, Rab7b controls trafficking from endosomes to the TGN, J Cell Sci, 123 (2010) 1480-1491. [4] C. Progida, M.S. Nielsen, G. Koster, C. Bucci, O. Bakke, Dynamics of Rab7b-dependent transport of sorting receptors, Traffic, 13 (2012) 1273-1285. [5] G. Seebohm, N. Strutz-Seebohm, R. Birkin, G. Dell, C. Bucci, M.R. Spinosa, R. Baltaev, A.F. Mack, G. Korniychuk, A. Choudhury, D. Marks, R.E. Pagano, B. Attali, A. Pfeufer, R.S. Kass, M.C. Sanguinetti, J.M. Tavare, F. Lang, Regulation of endocytic recycling of KCNQ1/KCNE1 potassium channels, Circ Res, 100 (2007) 686-692. [6] G. Seebohm, N. Strutz-Seebohm, O.N. Ureche, U. Henrion, R. Baltaev, A.F. Mack, G. Korniychuk, K. Steinke, D. Tapken, A. Pfeufer, S. Kääb, C. Bucci, B. Attali, J. Merot, J.M. Tavare, U.C. Hoppe, M.C. Sanguinetti, F. Lang, Long QT syndrome-associated mutations in KCNQ1 and KCNE1 subunits disrupt normal endosomal recycling of IKs channels, Circ Res, 103 (2008) 1451-1457. [7] A.L. Edinger, R.M. Cinalli, C.B. Thompson, Rab7 prevents growth factor-independent survival by inhibiting cell-autonomous nutrient transporter expression, Dev Cell, 5 (2003) 571-582. [8] K. Romero Rosales, E.R. Peralta, G.G. Guenther, S.Y. Wong, A.L. Edinger, Rab7 activation by growth factor withdrawal contributes to the induction of apoptosis, Mol Biol Cell, 20 (2009) 2831-2840. [9] C. Bucci, R. Frunzio, L. Chiariotti, A.L. Brown, M.M. Rechler, C.B. Bruni, A new member of the ras gene superfamily identified in a rat liver cell line, Nucleic Acids Res, 16 (1988) 9979- 9993. [10] C. Bucci, P. Thomsen, P. Nicoziani, J. McCarthy, B. van Deurs, Rab7: a key to lysosome biogenesis, Mol Biol Cell, 11 (2000) 467-480. [11] R. Harrison, C. Bucci, O. Vieira, T. Schroer, S. Grinstein, Phagosomes fuse with late endosomes and/or lysosomes by extension of membrane protrusions along microtubules: role of Rab7 and RILP, Mol Cell Biol, 23 (2003) 6494-6506. [12] S. Jager, C. Bucci, I. Tanida, T. Ueno, E. Kominami, P. Saftig, E.L. Eskelinen, Role for Rab7 in maturation of late autophagic vacuoles, J Cell Sci, 117 (2004) 4837-4848. [13] T. Wang, Z. Ming, W. Xiaochun, W. Hong, Rab7: role of its protein interaction cascades in endo-lysosomal traffic, Cell Signal, 23 (2011) 516-521. [14] G. Cantalupo, P. Alifano, V. Roberti, C.B. Bruni, C. Bucci, Rab-interacting lysosomal protein (RILP): the Rab7 effector required for transport to lysosomes, EMBO J, 20 (2001) 683-693. [15] M. Johansson, N. Rocha, W. Zwart, I. Jordens, L. Janssen, C. Kuijl, V.M. Olkkonen, J. Neefjes, Activation of endosomal dynein motors by stepwise assembly of Rab7-RILP-p150Glued, ORP1L, and the receptor betalll spectrin, J Cell Biol, 176 (2007) 459-471. [16] S. Pankiv, E.A. Alemu, A. Brech, J.A. Bruun, T. Lamark, A. Overvatn, G. Bjørkøy, T. Johansen, FYCO1 is a Rab7 effector that binds to LC3 and PI3P to mediate microtubule plus end-directed vesicle transport, J Cell Biol, 188 (2010) 253-269. [17] L. Cogli, C. Progida, R. Bramato, C. Bucci, Vimentin phosphorylation and assembly are regulated by the small GTPase Rab7a, Biochim Biophys Acta, 1833 (2013) 1283-1293.

[18] L. Cogli, C. Progida, C.L. Thomas, B. Spencer-Dene, C. Donno, G. Schiavo, C. Bucci, Charcot- Marie-Tooth type 2B disease-causing RAB7A mutant proteins show altered interaction with the neuronal intermediate filament peripherin, Acta Neuropathol, 25 (2013) 257-272. [19] Y. Sun, K.G. Buki, O. Ettala, J.P. Vaaraniemi, H.K. Vaananen, Possible role of direct Rac1- Rab7 interaction in ruffled border formation of osteoclasts, J Biol Chem, 280 (2005) 32356- 32361. [20] M.A. Frasa, F.C. Maximiano, K. Smolarczyk, R.E. Francis, M.E. Betson, E. Lozano, J. Goldenring, M.C. Seabra, A. Rak, M.R. Ahmadian, V.M. Braga, Armus is a Rac1 effector that inactivates Rab7 and regulates E-cadherin degradation, Curr Biol, 20 (2010) 198-208. [21] B. Carroll, N. Mohd-Naim, F. Maximiano, M.A. Frasa, J. McCormack, M. Finelli, S.B. Thoresen, L. Perdios, R. Daigaku, R.E. Francis, C. Futter, I. Dikic, V.M. Braga, The TBC/RabGAP Armus coordinates Rac1 and Rab7 functions during autophagy, Dev Cell, 25 (2013) 15-28. [22] C.D. Lawson, K. Burridge, The on-off relationship of Rho and Rac during integrin- mediated adhesion and cell migration, Small GTPases, 5 (2014) e27958. [23] B.M. Chung, J.D. Rotty, P.A. Coulombe, Networking galore: intermediate filaments and cell migration, Curr Opin Cell Biol, 25 (2013) 600-612. [24] T. Kawauchi, K. Sekine, M. Shikanai, K. Chihama, K. Tomita, K. Kubo, K. Nakajima, Y. Nabeshima, M. Hoshino, Rab GTPases-dependent endocytic pathways regulate neuronal migration and maturation through N-cadherin trafficking, Neuron, 67 (2010) 588-602. [25] T. Kawauchi, Regulation of cell adhesion and migration in cortical neurons: Not only Rho but also Rab family small GTPases, Small GTPases, 2 (2011) 36-40. [26] M.R. Spinosa, C. Progida, A. De Luca, A.M.R. Colucci, P. Alifano, C. Bucci, Functional characterization of Rab7 mutant proteins associated with Charcot-Marie-Tooth type 2B disease, J Neurosci, 28 (2008) 1640-1648. [27] M. Parsons, A.J. Messent, J.D. Humphries, N.O. Deakin, M.J. Humphries, Quantification of integrin receptor agonism by fluorescence lifetime imaging, J Cell Sci, 121 (2008) 265-271. [28] J.S. Berg, R.E. Cheney, Myosin-X is an unconventional myosin that undergoes intrafilopodial motility, Nat Cell Biol, 4 (2002) 246-250. [29] A. Palamidessi, E. Frittoli, M. Garré, M. Faretta, M. Mione, I. Testa, A. Diaspro, L. Lanzetti, G. Scita, P.P. Di Fiore, Endocytic trafficking of Rac is required for the spatial restriction of signaling in cell migration, Cell, 134 (2008) 135-147. [30] C. Bucci, R.G. Parton, I.H. Mather, H. Stunnenberg, K. Simons, B. Hoflack, M. Zerial, The small GTPase rab5 functions as a regulatory factor in the early endocytic pathway, Cell, 70 (1992) 715-728. [31] A. Arjonen, J. Alanko, S. Veltel, J. Ivaska, Distinct recycling of active and inactive 1 integrins., Traffic, 13 (2012) 610-625. [32] F. Steinberg, K.J. Heesom, M.D. Bass, P.J. Cullen, SNX17 protects integrins fromβ degradation by sorting between lysosomal and recycling pathways, J Cell Biol, 197 (2012) 219-230. [33] K. Poliakova, A. Adebola, C.L. Leung, B. Favre, R.K. Liem, I. Schepens, L. Borradori, BPAG1a and b associate with EB1 and EB3 and modulate vesicular transport, Golgi apparatus structure, and cell migration in C2.7 myoblasts, PLoS One, 9 (2014) e107535. [34] C. Angelucci, G. Maulucci, G. Lama, G. Proietti, A. Colabianchi, M. Papi, A. Maiorana, M. De Spirito, A. Micera, O.B. Balzamino, A. Di Leone, R. Masetti, G. Sica, Epithelial-stromal interactions in human breast cancer: effects on adhesion, plasma membrane fluidity and migration speed and directness, PLoS One, 7 (2012) e50804. [35] P. Outeda, D.L. Huso, S.A. Fisher, M.K. Halushka, H. Kim, F. Qian, G.G. Germino, T. Watnick, Polycystin signaling is required for directed endothelial cell migration and lymphatic development, Cell Rep, 7 (2014). [36] J. Ivaska, Unanchoring integrins in focal adhesions, Nat Cell Biol, 14 (2012) 981-983. 20

[37] F. Lagarrigue, P. Vikas Anekal, H.S. Lee, A.I. Bachir, J.N. Ablack, A.F. Horwitz, M.H. Ginsberg, A RIAM/lamellipodin-talin-integrin complex forms the tip of sticky fingers that guide cell migration, Nat Commun, 6 (2015) 8492. [38] W.B. Kiosses, S.J. Shattil, N. Pampori, M.A. Schwartz, Rac recruits high-affinity integrin v 3 to lamellipodia in endothelial cell migration, Nat Cell Biol, 3 (2001) 316-320. [39] C.G. Galbraith, K.M. Yamada, J.A. Galbraith, Polymerizing actin fibers position integrins αprimedβ to probe for adhesion sites, Science, 315 (2007) 992-995. [40] S.P. Palecek, J.C. Loftus, M.H. Ginsberg, D.A. Lauffenburger, A.F. Horwitz, Integrin-ligand binding properties govern cell migration speed through cell-substratum adhesiveness, Nature, 385 (1997) 537-540. [41] A.P. Mould, S.K. Akiyama, M.J. Humphries, The inhibitory anti-beta1 integrin monoclonal antibody 13 recognizes an epitope that is attenuated by ligand occupancy. Evidence for allosteric inhibition of integrin function, J Biol Chem, 271 (1996) 20365-20374. [42] P.K. Mattila, P. Lappalainen, Filopodia: molecular architecture and cellular functions, Nat Rev Mol Cell Biol, 9 (2008) 446-454. [43] N. De Franceschi, H. Hamidi, J. Alanko, P. Sahgal, J. Ivaska, Integrin traffic - the update, J Cell Sci, 128 (2015) 839-852. [44] A.B. Bohil, B.W. Robertson, R.E. Cheney, Myosin-X is a molecular motor that functions in filopodia formation, Proc Natl Acad Sci U S A, 103 (2006) 12411-12416. [45] H. Zhang, J.S. Berg, Z. Li, Y. Wang, P. Lång, A.D. Sousa, A. Bhaskar, R.E. Cheney, S. Strömblad, Myosin-X provides a motor-based link between integrins and the cytoskeleton, Nat Cell Biol, 6 (2004) 523-531. [46] L. Plantard, A. Arjonen, J.G. Lock, G. Nurani, J. Ivaska, S. Stromblad, PtdIns(3,4,5)P3 is a regulator of myosin-X localization and filopodia formation, J Cell Sci, 123 (2010) 3525-3534. [47] A.L. Berrier, R. Martinez, G.M. Bokoch, S.E. LaFlamme, The integrin beta tail is required and sufficient to regulate adhesion signaling to Rac1, J Cell Sci, 115 (2002) 4285-4291. [48] E.E. Bosco, J.C. Mulloy, Y. Zheng, Rac1 GTPase: a "Rac" of all trades., Cell Mol Life Sci, 66 (2009) 370-374. [49] A. Nodari, D. Zambroni, A. Quattrini, F.A. Court, A. D'Urso, A. Recchia, V.L. Tybulewicz, L. Wrabetz, M.L. Feltri, Beta1 integrin activates Rac1 in Schwann cells to generate radial lamellae during axonal sorting and myelination, J Cell Biol, 177 (2007) 1063-1075. [50] Z. Zhang, B. Yu, Y. Gu, S. Zhou, T. Qian, Y. Wang, G. Ding, F. Ding, X. Gu, Fibroblast-derived tenascin-C promotes Schwann cell migration through 1-integrin dependent pathway during peripheral nerve regeneration, Glia, 64 (2016) 374-385. [51] L.S. Havel, E.R. Kline, A.M. Salgueiro, A.I. Marcus,β Vimentin regulates lung cancer cell adhesion through a VAV2-Rac1 pathway to control focal adhesion kinase activity, Oncogene, 34 (2015) 1979-1990. [52] M. Meriane, S. Mary, F. Comunale, E. Vignal, P. Fort, C. Gauthier-Rouviére, Cdc42Hs and Rac1 GTPases induce the collapse of the vimentin intermediate filament network, J Biol Chem, 275 (2000) 33046-33052. [53] S.Y. Lee, E.J. Song, H.J. Kim, H.J. Kang, J.H. Kim, K.J. Lee, Rac1 regulates heat shock responses by reorganization of vimentin filaments: identification using MALDI-TOF MS, Cell Death Differ, 8 (2001) 1093-1102. [54] B.T. Helfand, M.G. Mendez, S.N. Murthy, D.K. Shumaker, B. Grin, S. Mahammad, U. Aebi, T. Wedig, Y.I. Wu, K.M. Hahn, M. Inagaki, H. Herrmann, R.D. Goldman, Vimentin organization modulates the formation of lamellipodia, Mol Biol Cell, 22 (2011) 1274-1289. [55] Y. Sakamoto, B. Boëda, S. Etienne-Manneville, APC binds intermediate filaments and is required for their reorganization during cell migration, J Cell Biol, 200 (2013) 249-258.

[56] J. Ayala, N. Touchot, A. Zahraoui, A. Tavitian, A. Prochiantz, The product of rab2, a small GTP binding protein, increases neuronal adhesion, and neurite growth in vitro, Neuron, 4 (1990) 797-805. [57] J.G. Lock, J.L. Stow, Rab11 in recycling endosomes regulates the sorting and basolateral transport of E-cadherin, Mol Biol Cell, 16 (2005) 1744-1755. [58] M. Borg, O. Bakke, C. Progida, A novel interaction between Rab7b and actomyosin reveals a dual role in intracellular transport and cell migration, J Cell Sci, 127 (2014) 4927-4939. [59] A.J. Deen, K. Rilla, S. Oikari, R. Kärnä, G. Bart, J. Häyrinen, A.R. Bathina, A. Ropponen, K. Makkonen, R.H. Tammi, M.I. Tammi, Rab10-mediated endocytosis of the hyaluronan synthase HAS3 regulates hyaluronan synthesis and cell adhesion to collagen, J Biol Chem, 289 (2014) 8375-8389. [60] G. Jacquemet, M.R. Morgan, A. Byron, J.D. Humphries, C.K. Choi, C.S. Chen, P.T. Caswell, M.J. Humphries, Rac1 is deactivated at integrin activation sites through an IQGAP1–filamin-A– RacGAP1 pathway, J Cell Sci, 126 (2013) 4121-4135. [61] C. Margadant, H.N. Monsuur, J.C. Norman, A. Sonnenberg, Mechanisms of integrin activation and trafficking, Curr Opin Cell Biol, 23 (2011) 607-614. [62] T. Pellinen, A. Arjonen, K. Vuoriluoto, K. Kallio, J.A. Fransen, J. Ivaska, Small GTPase Rab21 regulates cell adhesion and controls endosomal traffic of beta1-integrins, J Cell Biol, 173 (2006) 767-780. [63] T. Pellinen, S. Tuomi, A. Arjonen, M. Wolf, H. Edgren, H. Meyer, R. Grosse, T. Kitzing, J.K. Rantala, O. Kallioniemi, R. Fässler, M. Kallio, J. Ivaska, Integrin trafficking regulated by Rab21 is necessary for cytokinesis, Dev Cell, 15 (2008) 371-385. [64] C. Sandri, F. Caccavari, D. Valdembri, C. Camillo, S. Veltel, M. Santambrogio, L. Lanzetti, F. Bussolino, J. Ivaska, S. G., The R-Ras/RIN2/Rab5 complex controls endothelial cell adhesion and morphogenesis via active integrin endocytosis and Rac signaling, Cell Res, 22 (2012) 1479-1501. [65] A.M. Powelka, J. Sun, J. Li, M. Gao, L.M. Shaw, A. Sonnenberg, V.W. Hsu, Stimulation- dependent recycling of integrin beta1 regulated by ARF6 and Rab11, Traffic, 5 (2004) 20-36. [66] C. Wang, Y. Yoo, H. Fan, E. Kim, K.L. Guan, J.L. Guan, Regulation of Integrin 1 recycling to lipid rafts by Rab1a to promote cell migration, J Biol Chem, 285 (2010) 29398-29405. [67] M. Roberts, S. Barry, A. Woods, P. van der Sluijs, J. Norman, PDGF-regulatedβ rab4- dependent recycling of alphavbeta3 integrin from early endosomes is necessary for cell adhesion and spreading, Curr Biol, 11 (2001) 1392-1402. [68] R. Agarwal, I. Jurisica, G.B. Mills, K.W. Cheng, The emerging role of the RAB25 small GTPase in cancer, Traffic, 10 (2009) 1561-1568. [69] M.A. Dozynkiewicz, N.B. Jamieson, I. Macpherson, J. Grindlay, P.V. van den Berghe, A. von Thun, J.P. Morton, C. Gourley, P. Timpson, C. Nixon, C.J. McKay, R. Carter, D. Strachan, K. Anderson, O.J. Sansom, P.T. Caswell, J.C. Norman, Rab25 and CLIC3 collaborate to promote integrin recycling from late endosomes/lysosomes and drive cancer progression., Dev Cell, 22 (2012) 131-145. [70] E. Girard, D. Chmiest, N. Fournier, L. Johannes, J.L. Paul, B. Vedie, C. Lamaze, Rab7 is functionally required for selective cargo sorting at the early endosome, Traffic, 15 (2014) 309-326. [71] Q. Sun, W. Westphal, K.N. Wong, I. Tan, Q. Zhong, Rubicon controls endosome maturation as a Rab7 effector, Proc Natl Acad Sci U S A, 107 (2010) 19338-19343. [72] M.A. Del Pozo, M.A. Schwartz, Rac, membrane heterogeneity, caveolin and regulation of growth by integrins, Trends Cell Biol, 17 (2007) 246-250. [73] A. Mascia, F. Gentile, A. Izzo, N. Mollo, M. De Luca, C. Bucci, L. Nitsch, G. Calì, Rab7 Regulates CDH1 Endocytosis, Circular Dorsal Ruffles Genesis and Thyroglobulin Internalization in a Thyroid Cell Line, J Cell Physiol, 231 (2016) 1695-1708. 22

[74] A. Margiotta, C. Bucci, Role of intermediate filaments in vesicular traffic, Cells, 5 (2016) E20. [75] A. Ganguly, H. Yang, R. Sharma, K.D. Patel, F. Cabral, The role of microtubules and their dynamics in cell migration, J Biol Chem, 287 (2012) 43359-43369.

Figure legends

Fig. 1. Rab7a regulates cell motility. NCI H1299 cells transfected with control RNA or with Rab7a siRNA were imaged during a wound-healing assay. Cell migration was tracked every 30 min over a 9 hours period. (A) Images referring to the initial time point (0 hours) or after the closure of the wound (9 hours) are shown. The migratory tracks of some representative cells are shown for control cells and Rab7a-depleted cells. Individual tracks are shown so that each starts at the origin (0,0). Migration velocity (B) accumulated distance (C), euclidean distance (D) and directness (E) have been quantified. Accumulated distance is the total distance that cells travelled in a certain amount of time. Directness measures the tendency of the cells to travel following a direction (in contrast to random movements) and it is calculated by comparing the euclidean distance and the accumulated distance. Data represents the mean ± s.e.m. of four independent experiments (n=60). (F) Lysates of NCI H1299 cells transfected with control RNA or siRNA against Rab7a were subjected to western blot analysis using anti-Rab7a and anti-tubulin antibodies. (G) Quantification of Rab7a abundance. Data represent the mean ± s.e.m. of three independent experiments. Scale bar = 50 µm.

Fig. 2. Rab7a is required for cell spreading. (A) NCI H1299 cells transfected with control RNA, siRNA against Rab7a, or silenced and transfected afterward with HA-Rab7a wt, HA-Rab7a T22N or HA-Rab7a Q67L were plated on fibronectin-coated coverslips and left to adhere for 30 minutes. Samples were subsequently fixed and stained with rhodamine-conjugated phalloidin and anti-HA antibody. Confocal images and the relative transmission images are shown. Scale bar = 10 µm. (B) Quantification of the average area (in µm2) of NCI H1299 cells transfected as indicatedafter 30 minutes of adhesion on fibronectin. Data represent the mean ± s.e.m. of at least three different experiments (n=50). (C) Quantification of the average area (in µm2) of NCI H1299 cells silenced for Rab7a and transfected afterward with HA-Rab7a wt, HA-Rab7a T22N or HA-Rab7a Q67L as indicated. Data represent the mean ± s.e.m. of at least three different experiments (n=50).

Fig. 3. Rab7a depletion alters filopodia formation, localization of active β-1 integrin during wound healing and the activation state of β1-integrin. (A) NCI H1299 cells co-transfected with GFP β1- integrin and mcherry-Rab7a were imaged with a Spinning Disk confocal microscope. Right panel shows merged double fluorescence. Enlargements of the area surrounding the vesicle indicated by the arrow are shown. (B) Total internal reflection fluorescence microscopy (TIRF) images of migrating NCI H1299 cells transfected with control RNA or siRNA against Rab7a and stained using rhodamine-conjugated phalloidin and antibodies against inactive and active β-1 integrin. 24

Merge of actin (red) and inactive β-1 integrin (green) is shown (Merge1). Merge of inactive β-1 integrin (green) and active β-1 integrin (magenta) is shown (Merge2). Images were taken at 110 nm depth. (C) Quantification of the number of cells showing well-structured filopodia in migrating cells. Data represent the mean ± s.e.m. of four different experiments (n=50). (D) Quantification of the number of cells showing active β-1 integrin accumulation at the leading edge of migrating cells. Data represent the mean ± s.e.m. of three different experiments (n=50). (E) Lysates of NCI H1299 cells transfected with control RNA or siRNA against Rab7a were subjected to western blot analysis under non-reducing conditions using anti-active β1-integrin, anti-β1-integrin, anti-Rab7a and anti- tubulin antibodies. (F) Quantification of active β1-integrin abundance. Data represent the mean ± s.e.m. of four independent experiments. Scale bar = 10 µm.

Fig. 4. Rab7a silencing impairs active β1-integrin cytoplasmic localization. (A) Confocal images of NCI H1299 cells transfected with control RNA or siRNA against Rab7a stained with anti-active β1- integrin antibody (red) and DAPI (blue). Enlargements of the indicated areas are shown. (B) Quantification of active β1-integrin intensity. Data represent the mean ± s.e.m. of four independent experiments. (C) NCI H1299 cells transfected with control RNA or siRNA against Rab7a were stained with anti-EEA1 antibody (green) and anti-active β1-integrin antibody (red). Enlargements of the indicated areas are shown. (D) Quantification of colocalization between EEA1 and active β1- integrin. Data represent the mean ± s.e.m. of three independent experiments (n=50). (E) Lysates of NCI H1299 cells transfected with control RNA or siRNA against Rab7a were subjected to western blot analysis using anti-EEA1, anti-Rab7a and anti-tubulin antibodies. (F) Quantification of EEA1 abundance. Data represent the mean ± s.e.m. of three independent experiments. Scale bar = 10 µm.

Fig. 5. Rab7a regulates myosin X localization to filopodia tip. (A) NCI H1299 cells were co- transfected with mCherry-Rab7a and GFP-myosin X and imaged using a spinning-disk confocal microscope at the indicated time points. The red and white arrows indicate a vesicle positive for both mCherry-Rab7a and GFP-myosin X moving toward the cell periphery. Scale bar: 10 µm. (B-C) NCI H1299 cells were treated with control RNA or siRNA against Rab7a and transfected with GFP-myosin X and (B) samples were fixed and stained with Rhodamine-conjugated phalloidin. xy (large image), xz and yz maximum intensity projections of confocal images are shown. Scale bar: 10 µm. (C) Quantification of the percentage of cells with intracellular GFP- myosin X. Data represents the mean ± s.e.m. of three independent experiments (n = 50). (D-E) NCI H1299 cells were treated with either control RNA or with siRNA against Rab7a and transfected with GFP-myosin X and (D) cell lysates were analysed by western blotting against

25 myosin X, Rab7a and tubulin (serving as a loading control). (E) Quantification of myosin X expression in cells transfected with GFP-myosin X. The intensities of the bands were quantified by densitometry, normalized against the amount of tubulin, and plotted relative to the intensities obtained in cells transfected with control RNA. The values represent the mean ± s.d. of three independent experiments. (F-G) NCI H1299 were treated with either control RNA or with siRNA against Rab7a and (F) cell lysates were analysed by western blotting against myosin X, Rab7a and tubulin (serving as a loading control). (G) Quantification of the endogenous myosin X expression. The intensities of the bands were quantified by densitometry, normalized against the amount of tubulin, and plotted relative to the intensities obtained in cells transfected with control RNA. The values represent the mean ± s.d. of three independent experiments. (H) NCI H1299 treated with siRNA against Rab7a and co-transfected with GFP-myosin X and mApple-β1-integrin were imaged using a spinning-disk confocal microscope at the indicated time points. The red and white arrows indicate vesicles positive for both GFP-myosin X and mApple-β1- integrin. Scale bar: 10 µm

Fig. 6. Rab7a modulates Rac1 activation state. (A) Lysates of control NCI H1299 cells or of cells either overexpressing or silenced for Rab7a were subjected to western blot analysis using anti-Rac1, anti-Rab7a and anti-tubulin antibodies. Quantification of Rac1 abundance is shown. Data represent the mean ± s.e.m. of at least three independent experiments. (B-C) NCI H1299 cells co-transfected with GFP-Rac1 and mcherry-Rab7a were imaged with a Spinning Disk confocal microscope and enlargements of the indicated area are shown. Subsequent frames for the two channels and the merge are shown in C. (D) Lysates of NCI H1299 control or Rab7a-depleted cells were subjected to Rac1 activation assay and then subjected to western blot analysis using anti-Rac1, anti-Rab7a and anti-tubulin antibodies. (E) Quantification of Rac1-GTP in control or Rab7a-depleted cells. Data represent the mean ± s.e.m. of three independent experiments. Scale bar = 10 µm.

Fig. 7. Rab7a alters vimentin filaments reorganization and re-orientation during migration. (A) NCI H1299 cells transfected with control RNA or siRNA against Rab7a were stained after wound healing assay with vimentin (red) and Hoechst (blue) at 0 hours and 3 hours from the start of the experiment. * indicates a polarized cell with vimentin filaments oriented into the direction of the wound. (B) Quantification of the percentage of cells with oriented vimentin in control and Rab7a- depleted cells at the initial (0 hours) and final (3 hours) time points (n=50). (C) NCI H1299 cells

26 transfected with GFP-Rab7a wt were stained after wound healing assay with vimentin (red) and Hoechst (blue). Scale bar = 10 µm.

Fig. 8. Proposed model for the role of Rab7a in cell migration. Processes affected by Rab7a- depletion in migrating cells. Rab7a controls cell migration by regulating vimentin filament assembly and organization, Rac1 activity, myosin X localization and β1-integrin localization, activation and trafficking.

SUPPLEMENTARY FIGURE AND MOVIE LEGENDS FOR

Rab7a regulates cell migration through Rac1 and vimentin Azzurra Margiotta1,2, Cinzia Progida2*, Oddmund Bakke2 and Cecilia Bucci1* 1Department of Biological and Environmental Sciences and Technologies, (DiSTeBA) University of Salento, Via Provinciale Monteroni 165, 73100 Lecce, Italy 2 Department of Biosciences, Centre for Immune Regulation, University of Oslo, Blindernveien 31, 0371 Oslo, Norway

* Corresponding authors. Cecilia Bucci, Dipartimento di Scienze e Tecnologie Biologiche ed Ambientali, Università del

Salento, Via Provinciale Monteroni 165, 73100 Lecce, Italy. Tel.: +39 0832 298900; Fax: +39 0832 298626; E-mail: [email protected] and Cinzia Progida, Department of Biosciences, Centre for Immune Regulation, University of Oslo, Blindernveien 31, 0371 Oslo, Norway, tel +47922854441. E-mail: [email protected]

Fig. S1. Rab7a depletion does not affect neither other Rabs abundance nor cell proliferation. (A) Lysates of control or Rab7a-depleted cells were subjected to western blot analysis using anti Rab7a, anti-tubulin, anti-Rab4, anti-Rab5, anti-Rab7b, anti-Rab9 and anti-Rab11 antibodies. Quantification of Rab4 (B), Rab5 (C), Rab9 (D), Rab11 (E) and Rab7b (F) abundance relative to tubulin are shown. Data represent the mean ± s.e.m. of at least three independent experiments. (G) Proliferation assay of control and Rab7a-depleted cells. Cells were grown for 24 hours and then number of cells at different time points was calculated. Quantification of three independent experiments is shown. (H) Control and Rab7a-depleted cells were grown for 24 hours. Subsequently cells were stained at different time points with Trypan Blue Solution and percentage of living cells was calculated. Quantification of three independent experiments is shown. (I) MTT-based colorimetric assay of cell viability of NCI H1299 cells treated for 24 hours with either control RNA or Rab7a siRNA and then analysed at different time points.

Fig. S2. Rab7a depletion affects random cell migration. NCI H1299 cells were transfected with control RNA or with siRNA against Rab7a. Random cell migration was quantified by tracking single cells every 15 min over a 12 hours period. (A) Individual migratory trajectories from 30 representative cells are shown for control cells and Rab7a-depleted cells. Quantification of (B) cell migration velocity, (C) directness, (D) accumulated distance, and (E) euclidean distance are represented as mean ± s.e.m. of three independent experiments (n=100).

Fig. S3. Golgi apparatus orientation during wound healing. (A) NCI H1299 cells transfected with control RNA or siRNA against Rab7a were stained after wound healing assay with anti-giantin antibody (green), rhodamine-conjugated phalloidin (red) and Hoechst (blue). * indicates a polarized cell with Golgi oriented into the direction of the wound. (B) Scheme representing a polarized cell. The cell is considered to have Golgi oriented into the direction of the wound only if this is in the 1/3 of the cell facing the wound. (C) Quantification of NCI H1299 cells transfected with control RNA or Rab7a siRNA showing oriented Golgi into the direction of the wound after three hours of migration. Data represent the mean ± s.e.m. of three independent experiments (n=50). Bars = 10 µm.

Fig. S4. Silencing of Rab7a impairs filopodia formation. Confocal images of NCI H1299 cells transfected with control RNA or siRNA against Rab7a and stained for actin. Relative transmission images are shown. Scale bar = 10 µm.

Fig. S5. Filopodia dynamics in control cells. NCI H1299 cells transfected with LifeAct-RFP and GFP-myosin X were imaged using a spinning-disk confocal microscope at the indicated time points Enlargements of the indicated area show filopodia dynamics. Scale bar = 10 µm

Fig. S6. Rab7a T22N dominant negative mutant impairs localization of active β1-integrin. (A) NCI H1299 cells transfected with GFP or GFP-Rab7a T22N were stained with anti-EEA1 (blue) and anti-active β1-integrin (red) antibodies. Scale bar = 10 µm. (B) Quantification of colocalization between EEA1 and active β1-integrin. Data represent the mean ± s.e.m. of three independent experiments (n=50).

Movie S1. Silencing of Rab7a affects wound closure. Monolayers of NCI H1299 cells transfected either with control RNA (top) or Rab7a siRNA (bottom) were subjected to wound-healing assay. Cell migration was imaged with a Olympus confocal microscope every 30 minutes.

Movie S2. NCI H1299 cells transfected either with control RNA (left) or siRNA against Rab7a (right) were imaged with an Olympus Fluoview 1000 IX81 inverted confocal laser scanning microscope every 15 minutes. Individual cell tracks are shown as overlay.

Movie S3. Rab7a and β1-integrin overlap and move together. NCI H1299 cells co-transfected with GFP-β1-integrin (green) and mCherry-Rab7a wt (red) were imaged with a Spinning Disk confocal microscope at 2 second intervals.

Movie S4. Myosin X is transported to the cell periphery in Rab7a-positive endosomes. NCI H1299 cells co-transfected with GFP-myosin X (green) and mCherry-Rab7a wt (red) were imaged with a Spinning Disk confocal microscope at 5 second intervals. Movie S5. Myosin X localizes intracellularly together with β1-integrin in cell depleted for Rab7a. NCI H1299 cells silenced for Rab7a were co-transfected with GFP-myosin X (green) and mApple- β1-integrin (red) and imaged with a Spinning Disk confocal microscope at 2 second intervals.

Movie S6. Rab7a and Rac1 overlap and move together. NCI H1299 cells co-transfected with GFP- Rac1 (green) and mCherry-Rab7a wt (red) were imaged with a Spinning Disk confocal microscope at 2 second intervals.