- Email: [email protected]
The atypical Rho GTPase RhoD is a regulator of actin cytoskeleton dynamics and directed cell migration
Experimental Cell Research (xxxx) xxxx–xxxx
Contents lists available at ScienceDirect
Experimental Cell Research journal homepage: www.elsevier.com/locate/yexcr
The atypical Rho GTPase RhoD is a regulator of actin cytoskeleton dynamics and directed cell migration Magdalena Bloma, Katarina Reisa, Johan Heldinb, Johan Kreugerc, Pontus Aspenströma, a b c
⁎
Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, Sweden Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala SE-751 22 Uppsala, Sweden Department of Medical Cell Biology, Science for Life Laboratory, Uppsala University, SE-751 23 Uppsala, Sweden
A R T I C L E I N F O
A BS T RAC T
Keywords: RhoD Rho GTPase Actin Stress fiber Cell migration
RhoD belongs to the Rho GTPases, a protein family responsible for the regulation and organization of the actin cytoskeleton, and, consequently, many cellular processes like cell migration, cell division and vesicle trafficking. Here, we demonstrate that the actin cytoskeleton is dynamically regulated by increased or decreased protein levels of RhoD. Ectopic expression of RhoD has previously been shown to give an intertwined weave of actin filaments. We show that this RhoD-dependent effect is detected in several cell types and results in a less dynamic actin filament system. In contrast, RhoD depletion leads to increased actin filament-containing structures, such as cortical actin, stress fibers and edge ruffles. Moreover, vital cellular functions such as cell migration and proliferation are defective when RhoD is silenced. Taken together, we present data suggesting that RhoD is an important component in the control of actin dynamics and directed cell migration.
1. Introduction The actin cytoskeleton machinery is involved in many cellular processes, such as cell migration, morphogenesis, vesicle transport and cell division. A diverse range of proteins have been found to regulate actin filament reorganization, which is required for its highly dynamic properties. The Rho GTPases is a branch of the Ras protein superfamily and are known regulators of the actin cytoskeleton. Out of its 20 members, a limited number have been well studied, in particular RhoA, Cdc42 and Rac1, while the functions of the remaining members are poorly understood. One of the least studied proteins in this family, RhoD, was discovered more than 20 years ago as a protein regulating endosome dynamics and trafficking [1,2]. Many Rho GTPases are well conserved through evolution. RhoD, however, is a recent family member which most likely appeared by a gene duplication event in therians and is the only Rho GTPase to be expressed exclusively in mammals [3]. The Rho GTPases are small GTPases, which act as molecular switches, and bind to their effector proteins when in an active conformation, i.e. in a GTP-bound state. When GTP is hydrolyzed, a conformational change prevents interaction with effector proteins. The ability to bind and hydrolyze nucleotides is in turn regulated by three groups of proteins; the GEFs (Guanine nucleotide Exchange Factors), which exchange GDP for GTP, the GAPs (GTPase Activating Protein) which catalyze the GTP hydrolysis and GDIs
⁎
(Guanosine nucleotide Dissociation Inhibitors), which sequester and inhibit the protein in an inactive GDP-bound state. However, such regulators have not been found for all Rho GTPases, including RhoD. Instead, regulation of those Rho GTPases has been proposed to occur at the transcriptional level and by protein degradation [4,5]. The Rho GTPases can be categorized into two groups, classical and atypical proteins [6]. An atypical Rho GTPase has a slower GTP hydrolysis than the classical ones (such as RhoA, Cdc42 and Rac1), or no hydrolysis at all, rendering the protein constitutively active. One example of an atypical Rho GTPase is RhoBTB2 (also called DBC2), which has an aberrant domain structure and might not bind GTP [7,8]. The Rnd1, 2 and 3 are examples of Rho GTPases with low hydrolysis activity, leading to a constitutive active conformation of the proteins [9,10]. Wrch1 is another example of an atypical Rho GTPase, which is constitutively active due to a high intrinsic nucleotide exchange activity. Since GTP exceeds the level of GDP by the factor of ten in the cell, the protein will quickly bind new GTP after hydrolysis, leaving the protein in an active state [11]. RhoD was first categorized as a classical Rho GTPase, but was later found to possess a high intrinsic nucleotide exchange activity and is therefore now considered an atypical Rho GTPase, which is also true for the RhoD-like GTPase Rif [12]. Most of the 20 Rho GTPase members have been found to control various aspects of actin cytoskeletal dynamics, which can be clearly
Correspondence to: Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Nobels väg 16, Box 280, SE-171 77 Stockholm, Sweden. E-mail address: [email protected] (P. Aspenström).
http://dx.doi.org/10.1016/j.yexcr.2017.02.013 Received 22 November 2016; Received in revised form 7 February 2017; Accepted 10 February 2017 0014-4827/ © 2017 Elsevier Inc. All rights reserved.
Please cite this article as: Blom, M., Experimental Cell Research (2017), http://dx.doi.org/10.1016/j.yexcr.2017.02.013
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
migrating distances of the cells were analyzed using the Gradientech Tracking Tool software (Gradientech AB, Uppsala, Sweden).
seen when ectopically expressed in cells [13]. In addition to the original finding of RhoD as a regulator of endosome dynamics, the protein has been discovered to play roles in filopodia formation, ER-to-Golgi transport, Golgi homeostasis and receptor trafficking [13–16]. Processes like cell cycle regulation, centrosome duplication and cell migration have also been proven to be influenced by RhoD [17–19]. In this study, we widen the perspective of RhoD and show that increase or decrease of RhoD protein levels induces different actin cytoskeletal phenotypes in different cell types, and that fundamental cellular functions such as cell migration, chemotaxis and proliferation are to various degrees regulated by RhoD.
2.3. MCC – Microfluid Chemotaxis Chamber The procedure essentially followed the published protocol [20]. BJ hTERT SV40T were transfected with control or RhoD siRNA 48 h prior experimental onset as described above. Cells were seeded in serum-free media overnight on dishes that had been pre-coated with 0,2% gelatin for 15 min. A hill-shaped flow-dependent gradient of 0–20 ng/ml PDGF-BB in DMEM was applied to the starved cells. Phase-contrast pictures were acquired with a 10 x objective every 5 min for 50 cycles (approximately 4 h) using a Zeiss Axiovert 200 microscope. The migrating distances of the cells were analyzed using the Gradientech Tracking Tool software (Gradientech AB, Uppsala, Sweden). The acquired pictures were divided in three equal sections, where only cells in the first and third sections were quantified, as these two regions hold approximately linear gradients (in contrast to the middle section where the level of PDGF-BB is relatively even) (see Fig. 6A). Hence, the migration of cells towards increasing concentration of PDGF-BB occurs in two directions. In order to avoid obtaining a misleading total migrating distance after the four hours of acquisition, cells migrating out or into the field of view were excluded. In addition, cells dying or dividing during the acquired time were excluded. To quantify the fraction of cells migrating towards the gradient or perpendicular to the gradient, cells migrating in and out of the field of view were included. Data was analyzed using a two-way unpaired Student's t-test with equal variance.
2. Materials and methods 2.1. Cell culture, transfection and immunofluorescence BJ hTERT SV40T, HeLa, U2OS and U251MG were cultured in Dulbecco's modified Eagle's medium supplemented with 10% Fetal Bovine Serum (FBS) at 37 °C and 5% CO2. When starved, cells were cultured overnight in DMEM media without serum. For plasmid transfections, JetPEI transfection reagent (PolyPlus Transfection) was used according to manufacturer's protocol. 24 h post transfection, cells were fixed in 3% paraformaldehyde for 20 min at 37 °C, followed by permeabilization in 0.2% Triton-X100/PBS and blocking in 5% FBS/ PBS for 1 h. Samples were incubated 1 h each with primary and secondary antibodies, which were diluted in blocking solution. Myctagged murine RhoD wild-type, RhoD G26V, RhoD T31N and RhoDSAAX (with cysteines at positions 206 and 207 replaced with serines) have been described before [16]. The murine RhoD T31N-SAAX mutant was subcloned into the pRK5-Myc vector and the murine RhoD G26V into the pRK5-mCherry vector. GFP-LifeAct was a generous gift from Roger Karlsson, Stockholm University, Stockholm, Sweden. For siRNA transfections, SilentFect reagent (BioRad) was used according to manufacturer's protocol. 20 nM of RhoD siRNA; GAAGUGAAUCAUUUCUGCAtt or control siRNA (Ambion-Applied Biosystems) was used per transfection. siRNA-silenced cells were analyzed 48 h after transfection if not stated otherwise. The following antibodies were used: mouse anti-phospho-tyrosine (PY99) 1:200 and rabbit anti-Myc 1:50 (Santa Cruz Biotechnology), mouse anti-Myc 1:500 (Nordic BioSite), aminomethylcoumarin (AMCA)-conjugated goat anti-mouse 1:50 (Jackson ImmunoResearch Laboratories), Alexa Fluor 488-conjugated donkey anti-rabbit 1:1000 (Invitrogen Molecular Probes). F-actin was stained either with tetramethylrhodamine B isothiocyanate (TRITC)-conjugated phalloidin 1:250 (Sigma) or Alexa Fluor 488-conjugated phalloidin 1:500 (Invitrogen Molecular Probes). Pictures were acquired with a Zeiss AxioVert 40 CFL microscope equipped with a Zeiss AxioCAM MRm digital camera and the AxioVision software. Quantifications were performed in blind when possible.
2.4. Scratch-wound assay The IncuCyte Zoom Scratch Wound assay (Essen Bioscience) was used to follow the migration of BJ hTERT SV40T cells, during wound closure. The scratches were introduced using a wound maker, which creates wounds of equal width. This was done 48 h post siRNA transfection and pictures were acquired every two hours with a 10 x phase-contrast objective. The IncuCyte scratch-wound analysis software allowed quantification of the increasing cell confluence inside the wound. The percentage of the closing wound at each time point was analyzed using paired two-way Student's t-test with equal variance. 2.5. Cell proliferation assay The IncuCyte Zoom Cell Proliferation assay was used to monitor the cell growth rate of control and RhoD siRNA-treated BJ hTERT SV40T cells. Since the assay continued for longer than 48 h, the cells had been pre-transfected with siRNA for only 24 h at the experimental onset to make sure the knock down effect would last throughout the experiment. The cells, seeded in a multi-well plate, were scanned every second hour with a 10 x phase contrast objective to follow the growth rate. The IncuCyte proliferation software application was used to quantify the time required for cells to grow from 20% to 80% of confluence. The initial values were 15.3–25.8% for control-silenced samples and 14.3–24.6% for RhoD-silenced samples. Three samples of the control were excluded since their starting confluence values were below 10%, leading to a total of 9 and 12 quantified fields of view for control and RhoD siRNA treated cells respectively.
2.2. Live cell imaging 2.2.1. Actin dynamics BJ hTERT SV40T and U251MG cells were transfected with GFPLifeAct together with an mCherry-vector, or mCherry-tagged RhoD G26V. GFP-LifeAct was also transfected into cells treated with control or RhoD siRNA. Cells were kept at 37 °C and 5% CO2 during acquisition.
2.6. qPCR Total cellular RNA of HeLa, U2OS and U251MG cells was isolated with Quick-RNA MiniPrep Kit (Zymo Research) and reverse transcribed with SuperScript VILO cDNA synthesis kit (Invitrogen). The following primers were used: human RhoD (forward 5′TGGTCAACCTGCAAGTGAA-3′ and reverse 5′- GGAGGCGGTCATAGT CATC-3′), and human GAPDH (forward 5′-GGA AGG TGA AGG TCG
2.2.2. Mitosis and cell migration To follow random cell migration and to see if mitosis was hampered in RhoD silenced BJ hTERT SV40T, cells were seeded at low confluence. 48 h after siRNA transfection, the cells were filmed during 24 h. Cells were kept at 37 °C and 5% CO2 during acquisition. The 2
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
GAG TCA-3′ and reverse 5′-ATG GGT GGA ATC ATA TTG GAA CA-3′). The relative levels of RhoD and GAPDH transcripts were determined with the LC FastStart DNA Master SYBR Green I kit (Roche) in a
LightCycler 2.0 instrument (Roche) using the standard curve method. Total cellular RNA of BJ hTERT SV40T cells was isolated with E.Z.N.A total RNA kit I (VWR) and reverse transcribed with the iScript assay
Fig. 1. – Ectopic expression of RhoD leads to dissolution of stress fibers and filopodia formation. (A) Myc-empty vector and (B-C) Myc-RhoD transfection in HeLa, U2OS and U251MG cells. Filamentous actin was visualized with TRITC-conjugated phalloidin and Myc-RhoD with rabbit anti-Myc antibody followed by an Alexa Fluor 488-conjugated antirabbit antibody. Scale bar, 20 µm. (D) Quantification of actin stress fibers in Hela, U2OS and U251MG cells expressing Myc-empty vector (EV) or Myc-RhoD. n =100 cells / condition. Distinction was made between abundant (++SF), moderate (+SF) and no stress fibers (see Fig. S2). (E) Box-and-whisker plot displaying the number of actin protrusions per cell in HeLa, U2OS and U251MG cells expressing Myc-empty vector (EV) or Myc-RhoD. n = 50 cells / condition.
3
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 2. – RhoD silencing leads to increased stress fibers and edge ruffles in combination with altered focal adhesion distribution. (A-C) HeLa, (D-F) U2OS, and (G-I) U251MG cells were treated with control or RhoD siRNA for 48 h or 72 h before fixation. Arrows in (D) show examples of restricted areas with short actin fibers. (J-K) siRNA treated U251MG cells were starved overnight. The samples were stained with Alexa Fluor 488-conjugated phalloidin to visualize filamentous actin. (L) The percentages of U2OS cells, with clear cortical actin staining. The quantification using a two-way unpaired Student's t-test included 616 cells for control and 623 cells for RhoD siRNA treatment from three independent experiments. Error bars represent standard deviation. ***= p < 0.001 (M) Graph shows number of U251MG cells from three independent experiments with at least one ruffle. n(control siRNA, + serum) = 1087 cells, n(control siRNA, - serum) = 1090 cells, n(RhoD siRNA, + serum) =958 cells, n(RhoD siRNA, - serum) = 1042 cells, n(empty vector, + serum) =539 cells, n(Myc-RhoD, + serum) =477 cells. Error bars represent standard deviation. *= p < 0.05, **= p < 0.01, ***= p < 0.001, n.s. = non-significant. (N-O) Focal adhesion distribution alterations after RhoD silence in U2OS and (P-Q) U251MG cells. Focal adhesions were stained with mouse anti phospho-tyrosine (PY99) antibody followed by an AMCA-conjugated goat anti-mouse antibody. Scale bar, 20 µm.
4
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 3. – Ectopic expression of RhoD leads to less dynamic actin. U251MG cells were transfected with GFP-LifeAct together with (A) mCherry-empty vector and (B) mCherryRhoD G26V. Representative cells were filmed during 120 s to follow actin dynamics. Arrows show examples of dynamic actin region (in A) or non-dynamic region (in B). Scale bar, 10 µm.
siRNA (Fig. S1A,B), we could see a similar effect on actin in HeLa cells (Fig. 2A-B), but even more pronounced in U2OS cells, where an increased number of stress fibers as well as cortical actin filaments were detected (Fig. 2 D-E). Only 9% of control cells exhibited clear cortical actin staining, while RhoD silencing increased this proportion to 74% (Fig. 2L). In addition, many U2OS cells showed areas with short actin fibers in control siRNA-treated cells, which were lost when RhoD was silenced. U251MG cells, on the other hand, showed a different actin structure and demonstrated a dramatic increase in the formation of edge ruffles when RhoD was silenced (Fig. 2G-H, M). Loss of RhoD led to an increase from 39% to 71% of cells with at least one ruffle. Interestingly, a significantly decreased number of cells (19%) showed ruffles when control cells were starved overnight, while starved RhoD silenced cells still showed ruffles to the same extent (66%). This suggests that the induction of edge ruffles in control cells is more dependent upon external stimuli as opposed to edge ruffles induced by loss of RhoD (Fig. 2J-K, M). Conversely, expression of RhoD eliminated ruffles in most cells (from 40% down to 7%) (Fig. 2M). Silencing of RhoD for a prolonged time period (72 h) markedly altered the actin phenotype in HeLa cells, which showed a reduced cell size, in addition to the induction of edge ruffles, resembling the phenotype in U251MG cells. (Fig. 2C,F,I). 72 h of control siRNA treatment did not alter the actin phenotype (data not shown). Altogether, these results illustrate that increased or decreased RhoD protein levels have drastic, but variable effects on the actin architecture in different cell types. We next stained for phospho-tyrosine, which was used as a marker for focal adhesions. The distribution of focal adhesions mirrored the actin phenotype when RhoD was depleted. In U2OS cells, where the cortical actin increased, the focal adhesions were distributed along the cell border (Fig. 2N-O). In U251MG cells, where loss of RhoD led to an increased amount of edge ruffles, the focal adhesions were accumulated into bigger assemblies (Fig. 2P-Q).
(BioRad). qPCR was performed using the SsoFast EvaGreen supermix (Bio-Rad) on a MiniOpticon system (Bio-Rad) with the primer sets mentioned above. The resulting data was analyzed using the Bio-Rad CFX manager software. 3. Results 3.1. Increased or decreased RhoD protein levels lead to diverse cytoskeletal effects in different cell types Some Rho GTPases have a cell-type specific expression pattern, while RhoD appears to be widely expressed. High expression has been found for instance in kidney, liver, stomach, intestines, uterus, bladder and lung [2,3]. It has previously been reported that ectopic expression of RhoD in human foreskin fibroblasts, BJ hTERT SV40T, and porcine aortic endothelial cells leads to reorganization of the actin filament system causing stress fiber dissolution and induction of filopodia [21]. In order to investigate if this was a general effect, we overexpressed Myc-RhoD in different cell lines. In agreement with previous findings, RhoD expression in HeLa (human cervical carcinoma), U2OS (human osteosarcoma) and U251MG (human glioblastoma) cells led to thin and intertwined actin filaments with less cortical actin or stress fibers as compared to control cells transfected with empty vector (Fig. 1A-C). This was further confirmed by quantifying the RhoD-induced effects on the stress fiber organization (Fig. 1D, S2). Another characteristic phenotype of increased RhoD activity is the formation of actin-containing filopodia. This phenotype could be seen in HeLa and U251MG cells, but to a lesser extent in U2OS cells, suggesting different mechanisms for the two phenotypes (Fig. 1B-C, E). We have previously reported that silencing of RhoD results in actin reorganization leading to an increase in cortical actin filaments in BJ hTERT SV40T [21]. After silencing RhoD by treating cells 48 h with
5
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
dynamic actin. (Movies 5–8). Supplementary material related to this article can be found online at: doi:10.1016/j.yexcr.2017.02.013.
3.2. Increased or decreased RhoD protein levels affect actin dynamics To analyze the effect of RhoD on actin dynamics in live cells, we cotransfected LifeAct-GFP (F-actin marker) and RhoD transiently in U251MG cells. The cells were filmed for 120 s to follow rearrangements of actin-containing structures. We could see that cells transfected with an empty mCherry-vector had dynamic cell edges, especially at the lamellipodia (Fig. 3A, Movie 1). In contrast, cells transfected with mCherry-RhoD G26V (constitutively active mutant) showed dynamic filopodia, but other actin structures were rigid and few edge ruffles could be observed (Fig. 3B, Movie 2). We also transfected cells with mCherry-RhoD wild-type (wt), but the cells were more prone to round up and die. However, we have previously seen that expression of RhoD wt and the G26V mutant give similar effects on the actin cytoskeleton (Fig. S3). Furthermore, we treated U251MG cells with control and RhoD siRNA, but could not detect a prominent difference in the actin dynamics, although an increased number of edge ruffles were observed in line with previous results (Movies 3–4). We repeated the same procedure in BJ fibroblasts with similar results; cells expressing an empty vector or treated with control or RhoD siRNA showed a dynamic actin cytoskeleton, whereas RhoD G26V expressing cells had a less
3.3. Defective localization of RhoD leads to induction of actin filament bundles Most Rho GTPases have a CAAX motif (where C stands for cysteine, A for aliphatic amino acid and X for any amino acid) in the C-terminal. It is believed that the RhoD CAAX-motif is farnesylated, enabling the protein to be inserted into a lipid bilayer [22]. RhoD is known to localize to the plasma membrane, to endosomes and to the Golgi apparatus, all of which are lipid bilayer structures [14,16]. However, when mutating to serines the wild-type cysteines in, and adjacent to, the CAAX motif (we call this mutant SAAX), RhoD can no longer be farnesylated. We have previously shown, that this modification of the CAAX-motif leads to a loss of RhoD localization to endosomes in BJ fibroblasts [16]. In the current study, we compared the effects on the actin cytoskeleton when expressing RhoD CAAX and RhoD SAAX in HeLa cells (Fig. 4A-B). RhoD SAAX is still able to induce filopodia, indicating that it still localizes to the plasma membrane. Interestingly,
Fig. 4. – Defective localization of RhoD leads to increased stress fiber formation. HeLa cells were transfected with Myc-constructs expressing (A) RhoD wt CAAX, (B) RhoD wt SAAX, (C) RhoD T31N CAAX or (D) RhoD T31N SAAX. RhoD was visualized with a mouse anti-Myc antibody followed by an Alexa Fluor 488-conjugated anti-mouse antibody. Filamentous actin was visualized with TRITC-conjugated phalloidin. Scale bar, 20 µm.
6
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
RhoD SAAX induces bundles of actin filaments spanning the entire cell body. These fibers appear less contractile compared to stress fibers induced by RhoA, which often render the cell small and contracted [13]. This shows the intrinsic potential of RhoD to regulate the actin cytoskeleton, and that the effect on actin reorganization depends on its subcellular localization. When expressing a dominant negative mutant of RhoD (RhoD T31N CAAX) or RhoD T31N SAAX, no effect on actin filament organization could be detected, suggesting that the GTPase activity is crucial for RhoD’ s effect on actin reorganization (Fig. 4C-D).
SV40T cells, in which RhoD was silenced (Fig. S1C), did not close the wound as fast as control cells (Fig. 5A-C). We next sought to analyze if this decrease in the migratory property was caused by a defective migratory potential or by a derailed ability for directed cell migration. To study this, we analyzed the chemotactic capability in a Microfluid Chemotaxis Chamber as described in Materials and Methods. Control and RhoD-silenced BJ fibroblasts were monitored during exposure to a PDGF-BB chemoattractant gradient. A hill-shaped PDGF-BB gradient was created from the edges toward the middle of the monitored area. After four hours, a substantial fraction of the control siRNA-treated cells migrated to the middle section, whereas the RhoD-silenced cells to a greater extent did not respond to the PDGF-BB gradient (Fig. 6A-B). The migratory route for each individual cell was quantified as described in Material and Methods. We could see that the total migrated distance for RhoD-silenced cells did not differ from control cells, in fact, on average, they migrated a longer distance, although the difference was not statistically significant (Fig. 6D). However, the Euclidean distance did not differ (Fig. S4) When analyzing the directed migration distance, i.e. the distance in the direction of the gradient, RhoD-silenced cells did not migrate as far as control cells (Fig. 6E). Since the cells had different total migrating distances between experiments, we normalized the directed migration distance to the total migration distance, giving us comparable data between experiments. We could see that this migration quota (normalized directed migration distance) was close to 0.4 for control-silenced cells, but only near 0.2 for RhoD-silenced cells
3.4. RhoD silencing leads to deficient directed migration as well as decreased cell proliferation Since cellular movement is fully dependent upon the actin dynamics and RhoD clearly regulates several aspects of it, we next investigated cell migration. We have previously found a clear effect on the actin filament system when silencing RhoD in BJ fibroblasts [21]. Fibroblast cells are useful when studying migration, as the cells are highly motile and show a distinct actin cytoskeleton. One informative way to measure the ability of cells to migrate is the so called scratch-wound assay, where a scratch is made in a confluent monolayer of cells. Thereafter, the capacity of the cells to migrate and close the wound is analyzed. Here, we used the IncuCyte Zoom system, as this application provides formation of equally sized wounds, and a standardized way of analyzing. We could see that after 8 h, BJ hTERT
Fig. 5. – RhoD silencing leads to reduced wound closure. A scratch-wound assay was performed using the IncuCyte Zoom system. BJ hTERT SV40T cells were transfected with control or RhoD siRNA 48 h prior the scratching. The wound closure in (A) control siRNA-treated and (B) RhoD siRNA-treated cells was monitored for 8 h. Scale bar, 300 µm. (C) The percentage of the wound closure at each time point was analyzed using paired two-way Student's t-test with equal variance. Histogram shows mean values and error bars represent standard deviation of 24 wounds for each condition. **= p < 0.01***= p < 0.001, n.s. = non-significant.
7
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 6. – RhoD silencing leads to less efficient directed cell migration and decreased cell proliferation. The ability of BJ hTERT SV40T cells to migrate towards a PDGF gradient was measured with the MCC method as described in Materials and Methods. (A-B) RhoD or control silenced cells were exposed to a 20 ng/ml PDGF gradient (as indicated) during 4 h. Pictures were acquired every 5 min. (C) A polar plot, from one out of three experiments, shows direction and relative migration distances for control and RhoD-silenced cells exposed to a PDGF-BB gradient. Positive values mark the distance towards the gradient and negative values away from gradient. (D-F) Total migration distance, directed migration distance and normalized directed migration distance from three independent experiments were analyzed. n =385 cells for control siRNA-treated cells and n =291 cells for RhoD siRNAtreated cells. (G) The percentage of cells migrating towards the gradient or (H) perpendicular to the gradient was quantified. In G and H, cells moving in and out of the frame were included. n = 449 for control siRNA-treated cells and n =362 for RhoD siRNA-treated cells. Data was analyzed using two-way unpaired Student's t-test with equal variance. Histograms show mean values and error bars standard deviations of three independent experiments. **= p < 0.01, n.s. = non-significant. (I) To measure the cell proliferation rate, an IncuCyte Zoom proliferation assay was performed. Cells were transfected with RhoD or control siRNA 24 h prior to acquisition onset. Graph shows the required time for cells to grow from 20% to 80% of cell confluence. Error bars represent standard deviations.
compared to control cells. We also quantified the fraction of cells moving towards or away from the gradient (regardless of migration distances). Under these conditions, where 50% represents random cell migration, 77% of the control cells moved towards the gradient,
(Fig. 6F). Fig. 6C shows a polar plot displaying the direction and migration distances of cells from one out of three experiments. The distribution of the dots displays the incapacity of RhoD-silenced cells to perform polarized migration as they are more scattered in the plot 8
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
and cortical actin to increased SF and cortical actin to increased membrane ruffling and a strong ruffling associated with loss of SF. During cell migration, the actin cytoskeleton undergoes dynamic rearrangements to facilitate movement of the cell. Since RhoD clearly regulates the cytoskeleton, we wanted to find out how RhoD depletion affected cell migration. Our results from the chemotaxis experiments demonstrate that cells lacking RhoD did not have a decreased capability to migrate compared to control cells, however, they were unable to migrate as efficiently towards a PDGF-BB gradient. We have previously reported that RhoD silenced fibroblasts demonstrate a weaker activation of downstream PDGF-BB signaling pathway [16]. This is likely one of the causes that contributes to a decreased ability of RhoD-silenced cells to migrate in a directed manner. Without considering distance, we also found that a decreased fraction of RhoD knock down cells (63%) were migrating towards the stimulus as compared to control cells (77%). The difference might appear small, but one has to bear in mind that 50% corresponds to random migration, and an even lower number represents migration away from the gradient. Hence, we can conclude that lack of RhoD does not inhibit the cell migration machinery in BJ fibroblasts, but it impedes the ability of the cell to respond to an external migratory stimulus (Fig. 6). We also analyzed migration in a scratch-wound assay. When scratches are made, the migration of the outer layer of cells facing the wound is stimulated to fill the wound. Cells lacking RhoD also migrated to fill the wound, although less efficiently, compared to control cells (Fig. 5A-B) When performing a scratch-wound assay, one has to take into account that a difference in cell proliferation rate can give misleading results. We found that RhoD-silenced cells had a slower proliferation rate that could in theory influence the outcome (Figs. 5C, 6I). Importantly, we could detect a significant difference in wound closure between control and RhoD siRNA-treated cells already after 6 h and this is most likely too a short time period for differences in proliferation rate to interfere with the results. As we saw in the chemotaxis experiments, cells lacking RhoD did not show any signs of inhibited migration capability, but their ability to sense orientation was clearly affected. In order to migrate in a directed way, the polarization machinery of a cell must reorganize its Golgi towards the wound [27]. Since RhoD localizes to Golgi, and influences Golgi homeostasis [14], we postulate that cells lacking RhoD can have a defective Golgi reorientation, which in turn will affect the capability of the cells to polarize and migrate in a directed way.
whereas only 63% of the RhoD silenced cells did so. (Fig. 6G). For comparison, there was no difference between the two conditions when we analyzed the fractions of the cells migrating perpendicularly to the gradient (6 H). This is reflected in Fig. 6C, where there is an equal distribution of cells on left and right side of the origin. We also analyzed random migration of single BJ hTERT SV40 T cells seeded at low density in a multi-well plate. The RhoD-silenced cells migrated a longer distance compared to the control-silenced cells and, under these experimental settings, the difference was significant. However, there was no difference in the Euclidean distance, suggesting that the control cells migrated in a more persistent manner (Fig. S5A-B). Taken together, these results suggest that cells lacking RhoD do not have a decreased capacity to migrate compared to control cells; however, their capability to migrate in a polarized way is hampered. In addition to migration, actin dynamics plays a crucial part in cell division. To analyze if RhoD knock down leads to alterations in cell growth, we used IncuCyte Zoom Proliferation assay and measured the cell confluence over time. We concluded that RhoD silenced cells required an additional 12 h compared to control cells to grow from 20% to 80% of cell confluence (58 h and 46 h respectively) (Fig. 6I). However, we could not detect a slower cell division measured by live cell imaging from when the cells rounded up until they spread out postmitosis (Fig. S6), suggesting that the prolonged cell cycle is not due to defective mitosis. 4. Discussion A correctly regulated actin cytoskeleton is crucial for proper cell function and survival. Thus, altered regulation of Rho GTPases can lead to innumerous pathological conditions. Many studies have shown that an altered activation of Rho GTPases or their regulators are found in various tumor tissues [23], neurodegenerative diseases [24] and inflammatory disorders [25]. It is therefore of great importance to understand, in detail, how the different Rho GTPases regulate the actin cytoskeleton. When RhoD was first discovered, it soon became evident that also this Rho GTPase plays a role in actin cytoskeleton regulation, however, in a manner that differs from the three most studied Rho family members, RhoA, Cdc42 and Rac1. For example, ectopic expression of RhoA is well known to induce stress fibers, whereas ectopic expression of RhoD leads to dissolution of stress fibers and the formation of thin fibers [21]. In fact, in this study, we could see that also other actin-containing structures, such as cortical actin, lamellipodia and edge ruffles were suppressed when RhoD was overexpressed. Instead, the actin filament organization resembled a weave of intertwined threads often in combination with filopodia formation (Fig. 1). Also Rif (RhoF), the closest relative to RhoD, strongly induce filopodia but interestingly, Rif has an opposite effect on the actin cytoskeleton within the cell body. While Rif induces stress fiber formation, RhoD induces their dissolution [26]. In addition, depleting cells of RhoD altered the organization of the actin cytoskeleton. We have previously shown that RhoD silencing in human fibroblasts leads to increased cortical actin and lamellipodia [21]. In this study, we show that depletion of RhoD in HeLa and U2OS cells resulted in increased cortical actin and stress fiber formation, while the effect in U251MG cells rather gave rise to an increase of edge ruffles. When treating U2OS cells with RhoD siRNA, areas with short actin fibers were lost, suggesting an incorporation of these short fibers into stress fibers. Interestingly, in HeLa cells, loss of RhoD showed a spectrum of actin filament structures. After 48 h of RhoD silence, thin actin filaments had evolved into thick stress fibers and cortical actin, whereas an additional 24 h of siRNA-treatment resulted in formation of edge ruffles (Fig. 2). The actin organization is clearly different in different cell-types, hence a different response to RhoD silencing in the cell lines tested. The cell-types examined in the current study respond along a gradual scheme as depicted in Supplementary Fig. 7. There seems to be a gradual shift of actin phenotypes from few stress fibers
5. Conclusions In this study, we present data contributing to broaden the knowledge of RhoD. We show that increased as well as decreased RhoD protein levels have drastic effects on a variety of actin-containing structures. A decreased level of RhoD shifts the actin filament organization into the formation of cortical bundles, stress fibers and edge ruffles, whereas increased levels lead to thin actin fibers and filopodia. Furthermore, an altered RhoD protein level affects actin-driven cell processes, such as cell migration. We believe that findings from this study can enlighten aspects of cell migration and actin dynamics that are central both in physiological and pathological conditions. Acknowledgement We thank Noemi Nagy for verifying the efficiency of siRNA treatment by qPCR. The study was supported from Karolinska Institutet (PA, MB, KR) and grants from the Swedish Cancer Society (JK-2014/820, PA2014/0644). Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.yexcr.2017.02.013. 9
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
References [1] P. Chavrier, K. Simons, M. Zerial, The complexity of the Rab and Rho GTP-binding protein subfamilies revealed by a PCR cloning approach, Gene 112 (1992) 261–264. [2] C. Murphy, R. Saffrich, M. Grummt, H. Gournier, V. Rybin, M. Rubino, P. Auvinen, A. Lutcke, R.G. Parton, M. Zerial, Endosome dynamics regulated by a Rho protein, Nature 384 (1996) 427–432. [3] A. Boureux, E. Vignal, S. Faure, P. Fort, Evolution of the Rho family of ras-like GTPases in eukaryotes, Mol. Biol. Evol. 24 (2007) 203–216. [4] A. Wilkins, Q. Ping, C.L. Carpenter, RhoBTB2 is a substrate of the mammalian Cul3 ubiquitin ligase complex, Genes Dev. 18 (2004) 856–861. [5] R. Azzarelli, F. Guillemot, E. Pacary, Function and regulation of Rnd proteins in cortical projection neuron migration, Front. Neurosci. 9 (2015) 19. [6] P. Aspenstrom, A. Ruusala, D. Pacholsky, Taking Rho GTPases to the next level: the cellular functions of atypical Rho GTPases, Exp. Cell Res. 313 (2007) 3673–3679. [7] F.K. Chang, N. Sato, N. Kobayashi-Simorowski, T. Yoshihara, J.L. Meth, M. Hamaguchi, DBC2 is essential for transporting vesicular stomatitis virus glycoprotein, J. Mol. Biol. 364 (2006) 302–308. [8] J. Berthold, K. Schenkova, S. Ramos, Y. Miura, M. Furukawa, P. Aspenstrom, F. Rivero, Characterization of RhoBTB-dependent Cul3 ubiquitin ligase complexes–evidence for an autoregulatory mechanism, Exp. Cell Res. 314 (2008) 3453–3465. [9] R. Foster, K.Q. Hu, Y. Lu, K.M. Nolan, J. Thissen, J. Settleman, Identification of a novel human Rho protein with unusual properties: gtpase deficiency and in vivo farnesylation, Mol. Cell Biol. 16 (1996) 2689–2699. [10] P. Chardin, Function and regulation of Rnd proteins, Nat. Rev. Mol. Cell Biol. 7 (2006) 54–62. [11] J. Saras, P. Wollberg, P. Aspenstrom, Wrch1 is a GTPase-deficient Cdc42-like protein with unusual binding characteristics and cellular effects, Exp. Cell Res 299 (2004) 356–369. [12] M. Jaiswal, E.K. Fansa, R. Dvorsky, M.R. Ahmadian, New insight into the molecular switch mechanism of human Rho family proteins: shifting a paradigm, Biol. Chem. 394 (2013) 89–95. [13] P. Aspenstrom, A. Fransson, J. Saras, Rho GTPases have diverse effects on the organization of the actin filament system, Biochem. J. 377 (2004) 327–337. [14] M. Blom, K. Reis, V. Nehru, H. Blom, A.K. Gad, P. Aspenstrom, RhoD is a Golgi component with a role in anterograde protein transport from the ER to the plasma membrane, Exp. Cell Res. 333 (2015) 208–219. [15] K. Koizumi, K. Takano, A. Kaneyasu, H. Watanabe-Takano, E. Tokuda, T. Abe,
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] [24]
[25] [26]
[27]
10
N. Watanabe, T. Takenawa, T. Endo, RhoD activated by fibroblast growth factor induces cytoneme-like cellular protrusions through mDia3C, Mol. Biol. Cell 23 (2012) 4647–4661. V. Nehru, O. Voytyuk, J. Lennartsson, P. Aspenstrom, RhoD binds the Rab5 effector Rabankyrin-5 and has a role in trafficking of the platelet-derived growth factor receptor, Traffic 14 (2013) 1242–1254. K. Tsubakimoto, K. Matsumoto, H. Abe, J. Ishii, M. Amano, K. Kaibuchi, T. Endo, Small GTPase RhoD suppresses cell migration and cytokinesis, Oncogene 18 (1999) 2431–2440. A. Kyrkou, M. Soufi, R. Bahtz, C. Ferguson, M. Bai, R.G. Parton, I. Hoffmann, M. Zerial, T. Fotsis, C. Murphy, RhoD participates in the regulation of cell-cycle progression and centrosome duplication, Oncogene 32 (2013) 1831–1842. A. Kyrkou, M. Soufi, R. Bahtz, C. Ferguson, M. Bai, R.G. Parton, I. Hoffmann, M. Zerial, T. Fotsis, C. Murphy, The RhoD to centrosomal duplication, Small GTPases 4 (2013) 116–122. I. Barkefors, S. Le Jan, L. Jakobsson, E. Hejll, G. Carlson, H. Johansson, J. Jarvius, J.W. Park, N. Li Jeon, J. Kreuger, Endothelial cell migration in stable gradients of vascular endothelial growth factor A and fibroblast growth factor 2: effects on chemotaxis and chemokinesis, J. Biol. Chem. 283 (2008) 13905–13912. A.K. Gad, V. Nehru, A. Ruusala, P. Aspenstrom, RhoD regulates cytoskeletal dynamics via the actin nucleation-promoting factor WASp homologue associated with actin Golgi membranes and microtubules, Mol. Biol. Cell 23 (2012) 4807–4819. P.J. Roberts, N. Mitin, P.J. Keller, E.J. Chenette, J.P. Madigan, R.O. Currin, A.D. Cox, O. Wilson, P. Kirschmeier, C.J. Der, Rho Family GTPase modification and dependence on CAAX motif-signaled posttranslational modification, J. Biol. Chem. 283 (2008) 25150–25163. S.I. Ellenbroek, J.G. Collard, Rho GTPases: functions and association with cancer, Clin. Exp. Metastas-. 24 (2007) 657–672. T.R. Stankiewicz, D.A. Linseman, Rho family GTPases: key players in neuronal development, neuronal survival, and neurodegeneration, Front. Cell. Neurosci. 8 (2014) 314. M. Biro, M.A. Munoz, W. Weninger, Targeting Rho-GTPases in immune cell migration and inflammation, Br. J. Pharmacol. 171 (2014) 5491–5506. L. Fan, S. Pellegrin, A. Scott, H. Mellor, The small GTPase rif is an alternative trigger for the formation of actin stress fibers in epithelial cells, J. Cell Sci. 123 (2010) 1247–1252. A. Kupfer, D. Louvard, S.J. Singer, Polarization of the Golgi apparatus and the microtubule-organizing center in cultured fibroblasts at the edge of an experimental wound, Proc. Natl. Acad. Sci. USA 79 (1982) 2603–2607.
Contents lists available at ScienceDirect
Experimental Cell Research journal homepage: www.elsevier.com/locate/yexcr
The atypical Rho GTPase RhoD is a regulator of actin cytoskeleton dynamics and directed cell migration Magdalena Bloma, Katarina Reisa, Johan Heldinb, Johan Kreugerc, Pontus Aspenströma, a b c
⁎
Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, Sweden Department of Immunology, Genetics and Pathology, Science for Life Laboratory, Uppsala SE-751 22 Uppsala, Sweden Department of Medical Cell Biology, Science for Life Laboratory, Uppsala University, SE-751 23 Uppsala, Sweden
A R T I C L E I N F O
A BS T RAC T
Keywords: RhoD Rho GTPase Actin Stress fiber Cell migration
RhoD belongs to the Rho GTPases, a protein family responsible for the regulation and organization of the actin cytoskeleton, and, consequently, many cellular processes like cell migration, cell division and vesicle trafficking. Here, we demonstrate that the actin cytoskeleton is dynamically regulated by increased or decreased protein levels of RhoD. Ectopic expression of RhoD has previously been shown to give an intertwined weave of actin filaments. We show that this RhoD-dependent effect is detected in several cell types and results in a less dynamic actin filament system. In contrast, RhoD depletion leads to increased actin filament-containing structures, such as cortical actin, stress fibers and edge ruffles. Moreover, vital cellular functions such as cell migration and proliferation are defective when RhoD is silenced. Taken together, we present data suggesting that RhoD is an important component in the control of actin dynamics and directed cell migration.
1. Introduction The actin cytoskeleton machinery is involved in many cellular processes, such as cell migration, morphogenesis, vesicle transport and cell division. A diverse range of proteins have been found to regulate actin filament reorganization, which is required for its highly dynamic properties. The Rho GTPases is a branch of the Ras protein superfamily and are known regulators of the actin cytoskeleton. Out of its 20 members, a limited number have been well studied, in particular RhoA, Cdc42 and Rac1, while the functions of the remaining members are poorly understood. One of the least studied proteins in this family, RhoD, was discovered more than 20 years ago as a protein regulating endosome dynamics and trafficking [1,2]. Many Rho GTPases are well conserved through evolution. RhoD, however, is a recent family member which most likely appeared by a gene duplication event in therians and is the only Rho GTPase to be expressed exclusively in mammals [3]. The Rho GTPases are small GTPases, which act as molecular switches, and bind to their effector proteins when in an active conformation, i.e. in a GTP-bound state. When GTP is hydrolyzed, a conformational change prevents interaction with effector proteins. The ability to bind and hydrolyze nucleotides is in turn regulated by three groups of proteins; the GEFs (Guanine nucleotide Exchange Factors), which exchange GDP for GTP, the GAPs (GTPase Activating Protein) which catalyze the GTP hydrolysis and GDIs
⁎
(Guanosine nucleotide Dissociation Inhibitors), which sequester and inhibit the protein in an inactive GDP-bound state. However, such regulators have not been found for all Rho GTPases, including RhoD. Instead, regulation of those Rho GTPases has been proposed to occur at the transcriptional level and by protein degradation [4,5]. The Rho GTPases can be categorized into two groups, classical and atypical proteins [6]. An atypical Rho GTPase has a slower GTP hydrolysis than the classical ones (such as RhoA, Cdc42 and Rac1), or no hydrolysis at all, rendering the protein constitutively active. One example of an atypical Rho GTPase is RhoBTB2 (also called DBC2), which has an aberrant domain structure and might not bind GTP [7,8]. The Rnd1, 2 and 3 are examples of Rho GTPases with low hydrolysis activity, leading to a constitutive active conformation of the proteins [9,10]. Wrch1 is another example of an atypical Rho GTPase, which is constitutively active due to a high intrinsic nucleotide exchange activity. Since GTP exceeds the level of GDP by the factor of ten in the cell, the protein will quickly bind new GTP after hydrolysis, leaving the protein in an active state [11]. RhoD was first categorized as a classical Rho GTPase, but was later found to possess a high intrinsic nucleotide exchange activity and is therefore now considered an atypical Rho GTPase, which is also true for the RhoD-like GTPase Rif [12]. Most of the 20 Rho GTPase members have been found to control various aspects of actin cytoskeletal dynamics, which can be clearly
Correspondence to: Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Nobels väg 16, Box 280, SE-171 77 Stockholm, Sweden. E-mail address: [email protected] (P. Aspenström).
http://dx.doi.org/10.1016/j.yexcr.2017.02.013 Received 22 November 2016; Received in revised form 7 February 2017; Accepted 10 February 2017 0014-4827/ © 2017 Elsevier Inc. All rights reserved.
Please cite this article as: Blom, M., Experimental Cell Research (2017), http://dx.doi.org/10.1016/j.yexcr.2017.02.013
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
migrating distances of the cells were analyzed using the Gradientech Tracking Tool software (Gradientech AB, Uppsala, Sweden).
seen when ectopically expressed in cells [13]. In addition to the original finding of RhoD as a regulator of endosome dynamics, the protein has been discovered to play roles in filopodia formation, ER-to-Golgi transport, Golgi homeostasis and receptor trafficking [13–16]. Processes like cell cycle regulation, centrosome duplication and cell migration have also been proven to be influenced by RhoD [17–19]. In this study, we widen the perspective of RhoD and show that increase or decrease of RhoD protein levels induces different actin cytoskeletal phenotypes in different cell types, and that fundamental cellular functions such as cell migration, chemotaxis and proliferation are to various degrees regulated by RhoD.
2.3. MCC – Microfluid Chemotaxis Chamber The procedure essentially followed the published protocol [20]. BJ hTERT SV40T were transfected with control or RhoD siRNA 48 h prior experimental onset as described above. Cells were seeded in serum-free media overnight on dishes that had been pre-coated with 0,2% gelatin for 15 min. A hill-shaped flow-dependent gradient of 0–20 ng/ml PDGF-BB in DMEM was applied to the starved cells. Phase-contrast pictures were acquired with a 10 x objective every 5 min for 50 cycles (approximately 4 h) using a Zeiss Axiovert 200 microscope. The migrating distances of the cells were analyzed using the Gradientech Tracking Tool software (Gradientech AB, Uppsala, Sweden). The acquired pictures were divided in three equal sections, where only cells in the first and third sections were quantified, as these two regions hold approximately linear gradients (in contrast to the middle section where the level of PDGF-BB is relatively even) (see Fig. 6A). Hence, the migration of cells towards increasing concentration of PDGF-BB occurs in two directions. In order to avoid obtaining a misleading total migrating distance after the four hours of acquisition, cells migrating out or into the field of view were excluded. In addition, cells dying or dividing during the acquired time were excluded. To quantify the fraction of cells migrating towards the gradient or perpendicular to the gradient, cells migrating in and out of the field of view were included. Data was analyzed using a two-way unpaired Student's t-test with equal variance.
2. Materials and methods 2.1. Cell culture, transfection and immunofluorescence BJ hTERT SV40T, HeLa, U2OS and U251MG were cultured in Dulbecco's modified Eagle's medium supplemented with 10% Fetal Bovine Serum (FBS) at 37 °C and 5% CO2. When starved, cells were cultured overnight in DMEM media without serum. For plasmid transfections, JetPEI transfection reagent (PolyPlus Transfection) was used according to manufacturer's protocol. 24 h post transfection, cells were fixed in 3% paraformaldehyde for 20 min at 37 °C, followed by permeabilization in 0.2% Triton-X100/PBS and blocking in 5% FBS/ PBS for 1 h. Samples were incubated 1 h each with primary and secondary antibodies, which were diluted in blocking solution. Myctagged murine RhoD wild-type, RhoD G26V, RhoD T31N and RhoDSAAX (with cysteines at positions 206 and 207 replaced with serines) have been described before [16]. The murine RhoD T31N-SAAX mutant was subcloned into the pRK5-Myc vector and the murine RhoD G26V into the pRK5-mCherry vector. GFP-LifeAct was a generous gift from Roger Karlsson, Stockholm University, Stockholm, Sweden. For siRNA transfections, SilentFect reagent (BioRad) was used according to manufacturer's protocol. 20 nM of RhoD siRNA; GAAGUGAAUCAUUUCUGCAtt or control siRNA (Ambion-Applied Biosystems) was used per transfection. siRNA-silenced cells were analyzed 48 h after transfection if not stated otherwise. The following antibodies were used: mouse anti-phospho-tyrosine (PY99) 1:200 and rabbit anti-Myc 1:50 (Santa Cruz Biotechnology), mouse anti-Myc 1:500 (Nordic BioSite), aminomethylcoumarin (AMCA)-conjugated goat anti-mouse 1:50 (Jackson ImmunoResearch Laboratories), Alexa Fluor 488-conjugated donkey anti-rabbit 1:1000 (Invitrogen Molecular Probes). F-actin was stained either with tetramethylrhodamine B isothiocyanate (TRITC)-conjugated phalloidin 1:250 (Sigma) or Alexa Fluor 488-conjugated phalloidin 1:500 (Invitrogen Molecular Probes). Pictures were acquired with a Zeiss AxioVert 40 CFL microscope equipped with a Zeiss AxioCAM MRm digital camera and the AxioVision software. Quantifications were performed in blind when possible.
2.4. Scratch-wound assay The IncuCyte Zoom Scratch Wound assay (Essen Bioscience) was used to follow the migration of BJ hTERT SV40T cells, during wound closure. The scratches were introduced using a wound maker, which creates wounds of equal width. This was done 48 h post siRNA transfection and pictures were acquired every two hours with a 10 x phase-contrast objective. The IncuCyte scratch-wound analysis software allowed quantification of the increasing cell confluence inside the wound. The percentage of the closing wound at each time point was analyzed using paired two-way Student's t-test with equal variance. 2.5. Cell proliferation assay The IncuCyte Zoom Cell Proliferation assay was used to monitor the cell growth rate of control and RhoD siRNA-treated BJ hTERT SV40T cells. Since the assay continued for longer than 48 h, the cells had been pre-transfected with siRNA for only 24 h at the experimental onset to make sure the knock down effect would last throughout the experiment. The cells, seeded in a multi-well plate, were scanned every second hour with a 10 x phase contrast objective to follow the growth rate. The IncuCyte proliferation software application was used to quantify the time required for cells to grow from 20% to 80% of confluence. The initial values were 15.3–25.8% for control-silenced samples and 14.3–24.6% for RhoD-silenced samples. Three samples of the control were excluded since their starting confluence values were below 10%, leading to a total of 9 and 12 quantified fields of view for control and RhoD siRNA treated cells respectively.
2.2. Live cell imaging 2.2.1. Actin dynamics BJ hTERT SV40T and U251MG cells were transfected with GFPLifeAct together with an mCherry-vector, or mCherry-tagged RhoD G26V. GFP-LifeAct was also transfected into cells treated with control or RhoD siRNA. Cells were kept at 37 °C and 5% CO2 during acquisition.
2.6. qPCR Total cellular RNA of HeLa, U2OS and U251MG cells was isolated with Quick-RNA MiniPrep Kit (Zymo Research) and reverse transcribed with SuperScript VILO cDNA synthesis kit (Invitrogen). The following primers were used: human RhoD (forward 5′TGGTCAACCTGCAAGTGAA-3′ and reverse 5′- GGAGGCGGTCATAGT CATC-3′), and human GAPDH (forward 5′-GGA AGG TGA AGG TCG
2.2.2. Mitosis and cell migration To follow random cell migration and to see if mitosis was hampered in RhoD silenced BJ hTERT SV40T, cells were seeded at low confluence. 48 h after siRNA transfection, the cells were filmed during 24 h. Cells were kept at 37 °C and 5% CO2 during acquisition. The 2
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
GAG TCA-3′ and reverse 5′-ATG GGT GGA ATC ATA TTG GAA CA-3′). The relative levels of RhoD and GAPDH transcripts were determined with the LC FastStart DNA Master SYBR Green I kit (Roche) in a
LightCycler 2.0 instrument (Roche) using the standard curve method. Total cellular RNA of BJ hTERT SV40T cells was isolated with E.Z.N.A total RNA kit I (VWR) and reverse transcribed with the iScript assay
Fig. 1. – Ectopic expression of RhoD leads to dissolution of stress fibers and filopodia formation. (A) Myc-empty vector and (B-C) Myc-RhoD transfection in HeLa, U2OS and U251MG cells. Filamentous actin was visualized with TRITC-conjugated phalloidin and Myc-RhoD with rabbit anti-Myc antibody followed by an Alexa Fluor 488-conjugated antirabbit antibody. Scale bar, 20 µm. (D) Quantification of actin stress fibers in Hela, U2OS and U251MG cells expressing Myc-empty vector (EV) or Myc-RhoD. n =100 cells / condition. Distinction was made between abundant (++SF), moderate (+SF) and no stress fibers (see Fig. S2). (E) Box-and-whisker plot displaying the number of actin protrusions per cell in HeLa, U2OS and U251MG cells expressing Myc-empty vector (EV) or Myc-RhoD. n = 50 cells / condition.
3
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 2. – RhoD silencing leads to increased stress fibers and edge ruffles in combination with altered focal adhesion distribution. (A-C) HeLa, (D-F) U2OS, and (G-I) U251MG cells were treated with control or RhoD siRNA for 48 h or 72 h before fixation. Arrows in (D) show examples of restricted areas with short actin fibers. (J-K) siRNA treated U251MG cells were starved overnight. The samples were stained with Alexa Fluor 488-conjugated phalloidin to visualize filamentous actin. (L) The percentages of U2OS cells, with clear cortical actin staining. The quantification using a two-way unpaired Student's t-test included 616 cells for control and 623 cells for RhoD siRNA treatment from three independent experiments. Error bars represent standard deviation. ***= p < 0.001 (M) Graph shows number of U251MG cells from three independent experiments with at least one ruffle. n(control siRNA, + serum) = 1087 cells, n(control siRNA, - serum) = 1090 cells, n(RhoD siRNA, + serum) =958 cells, n(RhoD siRNA, - serum) = 1042 cells, n(empty vector, + serum) =539 cells, n(Myc-RhoD, + serum) =477 cells. Error bars represent standard deviation. *= p < 0.05, **= p < 0.01, ***= p < 0.001, n.s. = non-significant. (N-O) Focal adhesion distribution alterations after RhoD silence in U2OS and (P-Q) U251MG cells. Focal adhesions were stained with mouse anti phospho-tyrosine (PY99) antibody followed by an AMCA-conjugated goat anti-mouse antibody. Scale bar, 20 µm.
4
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 3. – Ectopic expression of RhoD leads to less dynamic actin. U251MG cells were transfected with GFP-LifeAct together with (A) mCherry-empty vector and (B) mCherryRhoD G26V. Representative cells were filmed during 120 s to follow actin dynamics. Arrows show examples of dynamic actin region (in A) or non-dynamic region (in B). Scale bar, 10 µm.
siRNA (Fig. S1A,B), we could see a similar effect on actin in HeLa cells (Fig. 2A-B), but even more pronounced in U2OS cells, where an increased number of stress fibers as well as cortical actin filaments were detected (Fig. 2 D-E). Only 9% of control cells exhibited clear cortical actin staining, while RhoD silencing increased this proportion to 74% (Fig. 2L). In addition, many U2OS cells showed areas with short actin fibers in control siRNA-treated cells, which were lost when RhoD was silenced. U251MG cells, on the other hand, showed a different actin structure and demonstrated a dramatic increase in the formation of edge ruffles when RhoD was silenced (Fig. 2G-H, M). Loss of RhoD led to an increase from 39% to 71% of cells with at least one ruffle. Interestingly, a significantly decreased number of cells (19%) showed ruffles when control cells were starved overnight, while starved RhoD silenced cells still showed ruffles to the same extent (66%). This suggests that the induction of edge ruffles in control cells is more dependent upon external stimuli as opposed to edge ruffles induced by loss of RhoD (Fig. 2J-K, M). Conversely, expression of RhoD eliminated ruffles in most cells (from 40% down to 7%) (Fig. 2M). Silencing of RhoD for a prolonged time period (72 h) markedly altered the actin phenotype in HeLa cells, which showed a reduced cell size, in addition to the induction of edge ruffles, resembling the phenotype in U251MG cells. (Fig. 2C,F,I). 72 h of control siRNA treatment did not alter the actin phenotype (data not shown). Altogether, these results illustrate that increased or decreased RhoD protein levels have drastic, but variable effects on the actin architecture in different cell types. We next stained for phospho-tyrosine, which was used as a marker for focal adhesions. The distribution of focal adhesions mirrored the actin phenotype when RhoD was depleted. In U2OS cells, where the cortical actin increased, the focal adhesions were distributed along the cell border (Fig. 2N-O). In U251MG cells, where loss of RhoD led to an increased amount of edge ruffles, the focal adhesions were accumulated into bigger assemblies (Fig. 2P-Q).
(BioRad). qPCR was performed using the SsoFast EvaGreen supermix (Bio-Rad) on a MiniOpticon system (Bio-Rad) with the primer sets mentioned above. The resulting data was analyzed using the Bio-Rad CFX manager software. 3. Results 3.1. Increased or decreased RhoD protein levels lead to diverse cytoskeletal effects in different cell types Some Rho GTPases have a cell-type specific expression pattern, while RhoD appears to be widely expressed. High expression has been found for instance in kidney, liver, stomach, intestines, uterus, bladder and lung [2,3]. It has previously been reported that ectopic expression of RhoD in human foreskin fibroblasts, BJ hTERT SV40T, and porcine aortic endothelial cells leads to reorganization of the actin filament system causing stress fiber dissolution and induction of filopodia [21]. In order to investigate if this was a general effect, we overexpressed Myc-RhoD in different cell lines. In agreement with previous findings, RhoD expression in HeLa (human cervical carcinoma), U2OS (human osteosarcoma) and U251MG (human glioblastoma) cells led to thin and intertwined actin filaments with less cortical actin or stress fibers as compared to control cells transfected with empty vector (Fig. 1A-C). This was further confirmed by quantifying the RhoD-induced effects on the stress fiber organization (Fig. 1D, S2). Another characteristic phenotype of increased RhoD activity is the formation of actin-containing filopodia. This phenotype could be seen in HeLa and U251MG cells, but to a lesser extent in U2OS cells, suggesting different mechanisms for the two phenotypes (Fig. 1B-C, E). We have previously reported that silencing of RhoD results in actin reorganization leading to an increase in cortical actin filaments in BJ hTERT SV40T [21]. After silencing RhoD by treating cells 48 h with
5
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
dynamic actin. (Movies 5–8). Supplementary material related to this article can be found online at: doi:10.1016/j.yexcr.2017.02.013.
3.2. Increased or decreased RhoD protein levels affect actin dynamics To analyze the effect of RhoD on actin dynamics in live cells, we cotransfected LifeAct-GFP (F-actin marker) and RhoD transiently in U251MG cells. The cells were filmed for 120 s to follow rearrangements of actin-containing structures. We could see that cells transfected with an empty mCherry-vector had dynamic cell edges, especially at the lamellipodia (Fig. 3A, Movie 1). In contrast, cells transfected with mCherry-RhoD G26V (constitutively active mutant) showed dynamic filopodia, but other actin structures were rigid and few edge ruffles could be observed (Fig. 3B, Movie 2). We also transfected cells with mCherry-RhoD wild-type (wt), but the cells were more prone to round up and die. However, we have previously seen that expression of RhoD wt and the G26V mutant give similar effects on the actin cytoskeleton (Fig. S3). Furthermore, we treated U251MG cells with control and RhoD siRNA, but could not detect a prominent difference in the actin dynamics, although an increased number of edge ruffles were observed in line with previous results (Movies 3–4). We repeated the same procedure in BJ fibroblasts with similar results; cells expressing an empty vector or treated with control or RhoD siRNA showed a dynamic actin cytoskeleton, whereas RhoD G26V expressing cells had a less
3.3. Defective localization of RhoD leads to induction of actin filament bundles Most Rho GTPases have a CAAX motif (where C stands for cysteine, A for aliphatic amino acid and X for any amino acid) in the C-terminal. It is believed that the RhoD CAAX-motif is farnesylated, enabling the protein to be inserted into a lipid bilayer [22]. RhoD is known to localize to the plasma membrane, to endosomes and to the Golgi apparatus, all of which are lipid bilayer structures [14,16]. However, when mutating to serines the wild-type cysteines in, and adjacent to, the CAAX motif (we call this mutant SAAX), RhoD can no longer be farnesylated. We have previously shown, that this modification of the CAAX-motif leads to a loss of RhoD localization to endosomes in BJ fibroblasts [16]. In the current study, we compared the effects on the actin cytoskeleton when expressing RhoD CAAX and RhoD SAAX in HeLa cells (Fig. 4A-B). RhoD SAAX is still able to induce filopodia, indicating that it still localizes to the plasma membrane. Interestingly,
Fig. 4. – Defective localization of RhoD leads to increased stress fiber formation. HeLa cells were transfected with Myc-constructs expressing (A) RhoD wt CAAX, (B) RhoD wt SAAX, (C) RhoD T31N CAAX or (D) RhoD T31N SAAX. RhoD was visualized with a mouse anti-Myc antibody followed by an Alexa Fluor 488-conjugated anti-mouse antibody. Filamentous actin was visualized with TRITC-conjugated phalloidin. Scale bar, 20 µm.
6
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
RhoD SAAX induces bundles of actin filaments spanning the entire cell body. These fibers appear less contractile compared to stress fibers induced by RhoA, which often render the cell small and contracted [13]. This shows the intrinsic potential of RhoD to regulate the actin cytoskeleton, and that the effect on actin reorganization depends on its subcellular localization. When expressing a dominant negative mutant of RhoD (RhoD T31N CAAX) or RhoD T31N SAAX, no effect on actin filament organization could be detected, suggesting that the GTPase activity is crucial for RhoD’ s effect on actin reorganization (Fig. 4C-D).
SV40T cells, in which RhoD was silenced (Fig. S1C), did not close the wound as fast as control cells (Fig. 5A-C). We next sought to analyze if this decrease in the migratory property was caused by a defective migratory potential or by a derailed ability for directed cell migration. To study this, we analyzed the chemotactic capability in a Microfluid Chemotaxis Chamber as described in Materials and Methods. Control and RhoD-silenced BJ fibroblasts were monitored during exposure to a PDGF-BB chemoattractant gradient. A hill-shaped PDGF-BB gradient was created from the edges toward the middle of the monitored area. After four hours, a substantial fraction of the control siRNA-treated cells migrated to the middle section, whereas the RhoD-silenced cells to a greater extent did not respond to the PDGF-BB gradient (Fig. 6A-B). The migratory route for each individual cell was quantified as described in Material and Methods. We could see that the total migrated distance for RhoD-silenced cells did not differ from control cells, in fact, on average, they migrated a longer distance, although the difference was not statistically significant (Fig. 6D). However, the Euclidean distance did not differ (Fig. S4) When analyzing the directed migration distance, i.e. the distance in the direction of the gradient, RhoD-silenced cells did not migrate as far as control cells (Fig. 6E). Since the cells had different total migrating distances between experiments, we normalized the directed migration distance to the total migration distance, giving us comparable data between experiments. We could see that this migration quota (normalized directed migration distance) was close to 0.4 for control-silenced cells, but only near 0.2 for RhoD-silenced cells
3.4. RhoD silencing leads to deficient directed migration as well as decreased cell proliferation Since cellular movement is fully dependent upon the actin dynamics and RhoD clearly regulates several aspects of it, we next investigated cell migration. We have previously found a clear effect on the actin filament system when silencing RhoD in BJ fibroblasts [21]. Fibroblast cells are useful when studying migration, as the cells are highly motile and show a distinct actin cytoskeleton. One informative way to measure the ability of cells to migrate is the so called scratch-wound assay, where a scratch is made in a confluent monolayer of cells. Thereafter, the capacity of the cells to migrate and close the wound is analyzed. Here, we used the IncuCyte Zoom system, as this application provides formation of equally sized wounds, and a standardized way of analyzing. We could see that after 8 h, BJ hTERT
Fig. 5. – RhoD silencing leads to reduced wound closure. A scratch-wound assay was performed using the IncuCyte Zoom system. BJ hTERT SV40T cells were transfected with control or RhoD siRNA 48 h prior the scratching. The wound closure in (A) control siRNA-treated and (B) RhoD siRNA-treated cells was monitored for 8 h. Scale bar, 300 µm. (C) The percentage of the wound closure at each time point was analyzed using paired two-way Student's t-test with equal variance. Histogram shows mean values and error bars represent standard deviation of 24 wounds for each condition. **= p < 0.01***= p < 0.001, n.s. = non-significant.
7
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
Fig. 6. – RhoD silencing leads to less efficient directed cell migration and decreased cell proliferation. The ability of BJ hTERT SV40T cells to migrate towards a PDGF gradient was measured with the MCC method as described in Materials and Methods. (A-B) RhoD or control silenced cells were exposed to a 20 ng/ml PDGF gradient (as indicated) during 4 h. Pictures were acquired every 5 min. (C) A polar plot, from one out of three experiments, shows direction and relative migration distances for control and RhoD-silenced cells exposed to a PDGF-BB gradient. Positive values mark the distance towards the gradient and negative values away from gradient. (D-F) Total migration distance, directed migration distance and normalized directed migration distance from three independent experiments were analyzed. n =385 cells for control siRNA-treated cells and n =291 cells for RhoD siRNAtreated cells. (G) The percentage of cells migrating towards the gradient or (H) perpendicular to the gradient was quantified. In G and H, cells moving in and out of the frame were included. n = 449 for control siRNA-treated cells and n =362 for RhoD siRNA-treated cells. Data was analyzed using two-way unpaired Student's t-test with equal variance. Histograms show mean values and error bars standard deviations of three independent experiments. **= p < 0.01, n.s. = non-significant. (I) To measure the cell proliferation rate, an IncuCyte Zoom proliferation assay was performed. Cells were transfected with RhoD or control siRNA 24 h prior to acquisition onset. Graph shows the required time for cells to grow from 20% to 80% of cell confluence. Error bars represent standard deviations.
compared to control cells. We also quantified the fraction of cells moving towards or away from the gradient (regardless of migration distances). Under these conditions, where 50% represents random cell migration, 77% of the control cells moved towards the gradient,
(Fig. 6F). Fig. 6C shows a polar plot displaying the direction and migration distances of cells from one out of three experiments. The distribution of the dots displays the incapacity of RhoD-silenced cells to perform polarized migration as they are more scattered in the plot 8
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
and cortical actin to increased SF and cortical actin to increased membrane ruffling and a strong ruffling associated with loss of SF. During cell migration, the actin cytoskeleton undergoes dynamic rearrangements to facilitate movement of the cell. Since RhoD clearly regulates the cytoskeleton, we wanted to find out how RhoD depletion affected cell migration. Our results from the chemotaxis experiments demonstrate that cells lacking RhoD did not have a decreased capability to migrate compared to control cells, however, they were unable to migrate as efficiently towards a PDGF-BB gradient. We have previously reported that RhoD silenced fibroblasts demonstrate a weaker activation of downstream PDGF-BB signaling pathway [16]. This is likely one of the causes that contributes to a decreased ability of RhoD-silenced cells to migrate in a directed manner. Without considering distance, we also found that a decreased fraction of RhoD knock down cells (63%) were migrating towards the stimulus as compared to control cells (77%). The difference might appear small, but one has to bear in mind that 50% corresponds to random migration, and an even lower number represents migration away from the gradient. Hence, we can conclude that lack of RhoD does not inhibit the cell migration machinery in BJ fibroblasts, but it impedes the ability of the cell to respond to an external migratory stimulus (Fig. 6). We also analyzed migration in a scratch-wound assay. When scratches are made, the migration of the outer layer of cells facing the wound is stimulated to fill the wound. Cells lacking RhoD also migrated to fill the wound, although less efficiently, compared to control cells (Fig. 5A-B) When performing a scratch-wound assay, one has to take into account that a difference in cell proliferation rate can give misleading results. We found that RhoD-silenced cells had a slower proliferation rate that could in theory influence the outcome (Figs. 5C, 6I). Importantly, we could detect a significant difference in wound closure between control and RhoD siRNA-treated cells already after 6 h and this is most likely too a short time period for differences in proliferation rate to interfere with the results. As we saw in the chemotaxis experiments, cells lacking RhoD did not show any signs of inhibited migration capability, but their ability to sense orientation was clearly affected. In order to migrate in a directed way, the polarization machinery of a cell must reorganize its Golgi towards the wound [27]. Since RhoD localizes to Golgi, and influences Golgi homeostasis [14], we postulate that cells lacking RhoD can have a defective Golgi reorientation, which in turn will affect the capability of the cells to polarize and migrate in a directed way.
whereas only 63% of the RhoD silenced cells did so. (Fig. 6G). For comparison, there was no difference between the two conditions when we analyzed the fractions of the cells migrating perpendicularly to the gradient (6 H). This is reflected in Fig. 6C, where there is an equal distribution of cells on left and right side of the origin. We also analyzed random migration of single BJ hTERT SV40 T cells seeded at low density in a multi-well plate. The RhoD-silenced cells migrated a longer distance compared to the control-silenced cells and, under these experimental settings, the difference was significant. However, there was no difference in the Euclidean distance, suggesting that the control cells migrated in a more persistent manner (Fig. S5A-B). Taken together, these results suggest that cells lacking RhoD do not have a decreased capacity to migrate compared to control cells; however, their capability to migrate in a polarized way is hampered. In addition to migration, actin dynamics plays a crucial part in cell division. To analyze if RhoD knock down leads to alterations in cell growth, we used IncuCyte Zoom Proliferation assay and measured the cell confluence over time. We concluded that RhoD silenced cells required an additional 12 h compared to control cells to grow from 20% to 80% of cell confluence (58 h and 46 h respectively) (Fig. 6I). However, we could not detect a slower cell division measured by live cell imaging from when the cells rounded up until they spread out postmitosis (Fig. S6), suggesting that the prolonged cell cycle is not due to defective mitosis. 4. Discussion A correctly regulated actin cytoskeleton is crucial for proper cell function and survival. Thus, altered regulation of Rho GTPases can lead to innumerous pathological conditions. Many studies have shown that an altered activation of Rho GTPases or their regulators are found in various tumor tissues [23], neurodegenerative diseases [24] and inflammatory disorders [25]. It is therefore of great importance to understand, in detail, how the different Rho GTPases regulate the actin cytoskeleton. When RhoD was first discovered, it soon became evident that also this Rho GTPase plays a role in actin cytoskeleton regulation, however, in a manner that differs from the three most studied Rho family members, RhoA, Cdc42 and Rac1. For example, ectopic expression of RhoA is well known to induce stress fibers, whereas ectopic expression of RhoD leads to dissolution of stress fibers and the formation of thin fibers [21]. In fact, in this study, we could see that also other actin-containing structures, such as cortical actin, lamellipodia and edge ruffles were suppressed when RhoD was overexpressed. Instead, the actin filament organization resembled a weave of intertwined threads often in combination with filopodia formation (Fig. 1). Also Rif (RhoF), the closest relative to RhoD, strongly induce filopodia but interestingly, Rif has an opposite effect on the actin cytoskeleton within the cell body. While Rif induces stress fiber formation, RhoD induces their dissolution [26]. In addition, depleting cells of RhoD altered the organization of the actin cytoskeleton. We have previously shown that RhoD silencing in human fibroblasts leads to increased cortical actin and lamellipodia [21]. In this study, we show that depletion of RhoD in HeLa and U2OS cells resulted in increased cortical actin and stress fiber formation, while the effect in U251MG cells rather gave rise to an increase of edge ruffles. When treating U2OS cells with RhoD siRNA, areas with short actin fibers were lost, suggesting an incorporation of these short fibers into stress fibers. Interestingly, in HeLa cells, loss of RhoD showed a spectrum of actin filament structures. After 48 h of RhoD silence, thin actin filaments had evolved into thick stress fibers and cortical actin, whereas an additional 24 h of siRNA-treatment resulted in formation of edge ruffles (Fig. 2). The actin organization is clearly different in different cell-types, hence a different response to RhoD silencing in the cell lines tested. The cell-types examined in the current study respond along a gradual scheme as depicted in Supplementary Fig. 7. There seems to be a gradual shift of actin phenotypes from few stress fibers
5. Conclusions In this study, we present data contributing to broaden the knowledge of RhoD. We show that increased as well as decreased RhoD protein levels have drastic effects on a variety of actin-containing structures. A decreased level of RhoD shifts the actin filament organization into the formation of cortical bundles, stress fibers and edge ruffles, whereas increased levels lead to thin actin fibers and filopodia. Furthermore, an altered RhoD protein level affects actin-driven cell processes, such as cell migration. We believe that findings from this study can enlighten aspects of cell migration and actin dynamics that are central both in physiological and pathological conditions. Acknowledgement We thank Noemi Nagy for verifying the efficiency of siRNA treatment by qPCR. The study was supported from Karolinska Institutet (PA, MB, KR) and grants from the Swedish Cancer Society (JK-2014/820, PA2014/0644). Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.yexcr.2017.02.013. 9
Experimental Cell Research (xxxx) xxxx–xxxx
M. Blom et al.
References [1] P. Chavrier, K. Simons, M. Zerial, The complexity of the Rab and Rho GTP-binding protein subfamilies revealed by a PCR cloning approach, Gene 112 (1992) 261–264. [2] C. Murphy, R. Saffrich, M. Grummt, H. Gournier, V. Rybin, M. Rubino, P. Auvinen, A. Lutcke, R.G. Parton, M. Zerial, Endosome dynamics regulated by a Rho protein, Nature 384 (1996) 427–432. [3] A. Boureux, E. Vignal, S. Faure, P. Fort, Evolution of the Rho family of ras-like GTPases in eukaryotes, Mol. Biol. Evol. 24 (2007) 203–216. [4] A. Wilkins, Q. Ping, C.L. Carpenter, RhoBTB2 is a substrate of the mammalian Cul3 ubiquitin ligase complex, Genes Dev. 18 (2004) 856–861. [5] R. Azzarelli, F. Guillemot, E. Pacary, Function and regulation of Rnd proteins in cortical projection neuron migration, Front. Neurosci. 9 (2015) 19. [6] P. Aspenstrom, A. Ruusala, D. Pacholsky, Taking Rho GTPases to the next level: the cellular functions of atypical Rho GTPases, Exp. Cell Res. 313 (2007) 3673–3679. [7] F.K. Chang, N. Sato, N. Kobayashi-Simorowski, T. Yoshihara, J.L. Meth, M. Hamaguchi, DBC2 is essential for transporting vesicular stomatitis virus glycoprotein, J. Mol. Biol. 364 (2006) 302–308. [8] J. Berthold, K. Schenkova, S. Ramos, Y. Miura, M. Furukawa, P. Aspenstrom, F. Rivero, Characterization of RhoBTB-dependent Cul3 ubiquitin ligase complexes–evidence for an autoregulatory mechanism, Exp. Cell Res. 314 (2008) 3453–3465. [9] R. Foster, K.Q. Hu, Y. Lu, K.M. Nolan, J. Thissen, J. Settleman, Identification of a novel human Rho protein with unusual properties: gtpase deficiency and in vivo farnesylation, Mol. Cell Biol. 16 (1996) 2689–2699. [10] P. Chardin, Function and regulation of Rnd proteins, Nat. Rev. Mol. Cell Biol. 7 (2006) 54–62. [11] J. Saras, P. Wollberg, P. Aspenstrom, Wrch1 is a GTPase-deficient Cdc42-like protein with unusual binding characteristics and cellular effects, Exp. Cell Res 299 (2004) 356–369. [12] M. Jaiswal, E.K. Fansa, R. Dvorsky, M.R. Ahmadian, New insight into the molecular switch mechanism of human Rho family proteins: shifting a paradigm, Biol. Chem. 394 (2013) 89–95. [13] P. Aspenstrom, A. Fransson, J. Saras, Rho GTPases have diverse effects on the organization of the actin filament system, Biochem. J. 377 (2004) 327–337. [14] M. Blom, K. Reis, V. Nehru, H. Blom, A.K. Gad, P. Aspenstrom, RhoD is a Golgi component with a role in anterograde protein transport from the ER to the plasma membrane, Exp. Cell Res. 333 (2015) 208–219. [15] K. Koizumi, K. Takano, A. Kaneyasu, H. Watanabe-Takano, E. Tokuda, T. Abe,
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] [24]
[25] [26]
[27]
10
N. Watanabe, T. Takenawa, T. Endo, RhoD activated by fibroblast growth factor induces cytoneme-like cellular protrusions through mDia3C, Mol. Biol. Cell 23 (2012) 4647–4661. V. Nehru, O. Voytyuk, J. Lennartsson, P. Aspenstrom, RhoD binds the Rab5 effector Rabankyrin-5 and has a role in trafficking of the platelet-derived growth factor receptor, Traffic 14 (2013) 1242–1254. K. Tsubakimoto, K. Matsumoto, H. Abe, J. Ishii, M. Amano, K. Kaibuchi, T. Endo, Small GTPase RhoD suppresses cell migration and cytokinesis, Oncogene 18 (1999) 2431–2440. A. Kyrkou, M. Soufi, R. Bahtz, C. Ferguson, M. Bai, R.G. Parton, I. Hoffmann, M. Zerial, T. Fotsis, C. Murphy, RhoD participates in the regulation of cell-cycle progression and centrosome duplication, Oncogene 32 (2013) 1831–1842. A. Kyrkou, M. Soufi, R. Bahtz, C. Ferguson, M. Bai, R.G. Parton, I. Hoffmann, M. Zerial, T. Fotsis, C. Murphy, The RhoD to centrosomal duplication, Small GTPases 4 (2013) 116–122. I. Barkefors, S. Le Jan, L. Jakobsson, E. Hejll, G. Carlson, H. Johansson, J. Jarvius, J.W. Park, N. Li Jeon, J. Kreuger, Endothelial cell migration in stable gradients of vascular endothelial growth factor A and fibroblast growth factor 2: effects on chemotaxis and chemokinesis, J. Biol. Chem. 283 (2008) 13905–13912. A.K. Gad, V. Nehru, A. Ruusala, P. Aspenstrom, RhoD regulates cytoskeletal dynamics via the actin nucleation-promoting factor WASp homologue associated with actin Golgi membranes and microtubules, Mol. Biol. Cell 23 (2012) 4807–4819. P.J. Roberts, N. Mitin, P.J. Keller, E.J. Chenette, J.P. Madigan, R.O. Currin, A.D. Cox, O. Wilson, P. Kirschmeier, C.J. Der, Rho Family GTPase modification and dependence on CAAX motif-signaled posttranslational modification, J. Biol. Chem. 283 (2008) 25150–25163. S.I. Ellenbroek, J.G. Collard, Rho GTPases: functions and association with cancer, Clin. Exp. Metastas-. 24 (2007) 657–672. T.R. Stankiewicz, D.A. Linseman, Rho family GTPases: key players in neuronal development, neuronal survival, and neurodegeneration, Front. Cell. Neurosci. 8 (2014) 314. M. Biro, M.A. Munoz, W. Weninger, Targeting Rho-GTPases in immune cell migration and inflammation, Br. J. Pharmacol. 171 (2014) 5491–5506. L. Fan, S. Pellegrin, A. Scott, H. Mellor, The small GTPase rif is an alternative trigger for the formation of actin stress fibers in epithelial cells, J. Cell Sci. 123 (2010) 1247–1252. A. Kupfer, D. Louvard, S.J. Singer, Polarization of the Golgi apparatus and the microtubule-organizing center in cultured fibroblasts at the edge of an experimental wound, Proc. Natl. Acad. Sci. USA 79 (1982) 2603–2607.