E XP ER I ME NTAL C E LL RE S E ARCH

333 (2 015) 2 08 – 219

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

Research Article

RhoD is a Golgi component with a role in anterograde protein transport from the ER to the plasma membrane$ Magdalena Bloma, Katarina Reisa, Vishal Nehrua, Hans Blomb, Annica K.B. Gada, Pontus Aspenströma,n a

Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, SE-171 77 Stockholm, Sweden Applied Physics, Royal Institute of Technology, Science for Life Laboratory, SE-171 21 Stockholm, Sweden

b

article information

abstract

Article Chronology:

RhoD is a member of the Rho GTPase family and it coordinates actin dynamics and membrane

Received 21 November 2014

trafficking. Activation of RhoD results in formation of filopodia, dissolution of stress fibers, and

Received in revised form

the subsequent formation of short actin bundles. In addition, RhoD localizes to early endosomes

24 February 2015

and recycling endosomes, and has a regulatory role in endosome trafficking. In this study, we

Accepted 26 February 2015

report on a function of RhoD in the regulation of Golgi homeostasis. We show that manipulation

Available online 5 March 2015

of protein and activation levels of RhoD, as well as of its binding partner WHAMM, result in

Keywords:

derailed localization of Golgi stacks. Moreover, vesicle trafficking from the endoplasmic reticulum

Rho GTPases

to the plasma membrane via the Golgi apparatus measured by the VSV-G protein is severely

RhoD

hampered by manipulation of RhoD or WHAMM. In summary, our studies demonstrate a novel

WHAMM

role for this member of the Rho GTPases in the regulation of Golgi function.

Golgi

& 2015 Elsevier Inc. All rights reserved.

VSV-G

Introduction RhoD is a member of the Rho family of small GTPases, and it has been implicated in integrating actin reorganization and endosome motility [1,2]. RhoD and the closely related RhoF (also known as Rho in filopodia; Rif) have been shown to have atypical features. Although these proteins can cycle between active GTP-bound and inactive GDPbound conformations, RhoD and RhoF have a greater intrinsic exchange activity compared to the other Rho GTPases [3,4]. In this respect, they resemble the tumor-associated Rac1 splice variant Rac1b [5]. These observations suggest that they are not controlled by the

☆

same regime as the classical Rho GTPases, such as RhoA, Rac1 and Cdc42. These latter GTPases are regulated by guanine nucleotide exchange factors (GEFs) that catalyze the exchange of GDP for GTP, and GTPase activating proteins (GAPs) that stimulate the intrinsic GTPase activity of these proteins [6]. To date, no GEFs or GAPs have been identified for RhoD or Rif. It is therefore likely that the activity of RhoD is regulated by other mechanisms, possibly at the level of transcription or by post-translational modifications. Several studies have demonstrated that RhoD activation triggers the formation of filopodia and thin bundles of actin filaments [2,7]. In addition, RhoD has a role in the regulation of cytokinesis, and

Running title: RhoD in anterograde transport. Correspondence to: Pontus Aspenström, Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Nobels väg 16, Box 280, SE-171 77 Stockholm, Sweden. Fax: þ46 8 330498. E-mail address: [email protected] (P. Aspenström). n

http://dx.doi.org/10.1016/j.yexcr.2015.02.023 0014-4827/& 2015 Elsevier Inc. All rights reserved.

E XP E RI ME N TAL C E L L RE S E ARCH

the disruption of RhoD activity leads to the formation of multinucleated cells [8]. Moreover, RhoD is involved in the regulation of G1/S-phase progression, and expression of the constitutively active RhoD mutant RhoD/G26V results in derailed centriole duplication [9]. In the present study, we show that endogenous RhoD localizes to the Golgi apparatus and a function of RhoD in Golgi homeostasis was suggested by the finding that ectopic expression of RhoD results in dissolution of the Golgi apparatus and redistribution of Golgiderived vesicles throughout the cytoplasm. In addition, we show that interference with the RhoD activity results in defective transport from the ER to the plasma membrane via the Golgi apparatus. We have previously shown that RhoD binds to the actin nucleation-promoting factor WASp homolog associated with Actin, Membranes and Microtubules (WHAMM) [7]. Interestingly, WHAMM was originally identified as a Golgi-localized protein with a role in the regulation of endoplasmic reticulum (ER)-to-Golgi transport [10]. Our data implicate a previously unknown RhoD-dependent pathway, with a role in the regulation of Golgi homeostasis and function, and WHAMM is a possible candidate to function in such a Rho-regulated pathway.

Materials and methods Antibodies, reagents, and constructs A guinea pig anti-WHAMM antibody was a generous gift from Dr Kenneth Campellone (University of Connecticut, CT, USA). The mouse anti-GalT antibody was a generous gift from Dr Ulla Mandel (University of Copenhagen, Denmark). The following commercial antibodies were used: mouse monoclonal anti-Flag (M2), rabbit anti-FLAG, mouse monoclonal anti-α-tubulin and rabbit anti-RhoD (Sigma-Aldrich); rabbit anti-Myc (Santa Cruz Biotechnology); mouse anti-c-Myc (Covance); mouse monoclonal anti-GM130 (BD Biosciences); rabbit anti-mannosidase II (Millipore); mouse and rabbit IgG anti-horse-radish-peroxidase (GE Healthcare); rabbit anti-GFP (Life Technologies); tetramethyl rhodamine isothiocyanate (TRITC)-conjugated anti-mouse, aminomethylcoumarin acetate (AMCA)-conjugated anti-rabbit and TRITC-conjugated anti-guinea pig (Jackson ImmunoResearch Laboratories); TRITC-conjugated anti-rabbit (DAKO); AlexaFluor 488-conjugated anti-mouse, AlexaFluor568-conjugated anti-mouse and anti-rabbit, AlexaFluor594-conjugated anti-mouse, and Atto647Nconjugated anti-rabbit (Invitrogen-Molecular Probes); STAR635-conjugated anti-rabbit (Abberior). Alexa Fluor 488-conjugated phalloidin was from Invitrogen-Molecular Probes, and 40 ,6-diamidino-2-phenylindole dihydrochloride (DAPI) was from Sigma Aldrich. Endoglycosidase H and Complete EDTA-free protease inhibitor cocktail were from Roche. EGFP-VSV-G was a generous gift from Dr. Alberto Luini (TIGEM, Naples, Italy) and FLAG-tagged ArfGAP was a generous gift from Dr Helge Gad (Stockholm University, Sweden).

333 (2 015) 2 08 – 219

209

Knockdown of RhoD and WHAMM expression was induced by transfection of the BJ/SV40T cells with siRNAs against RhoD or WHAMM (CGGUGUUUGAGCGGUACAUtt: RhoD siRNA#2; GAAGUGAAUCAUUUCUGCAtt: RhoD siRNA#3; GAAAAUCAUCGGUUCAGAUtt: WHAMM siRNA), or with a control siRNA (Ambion-Applied Biosystems), using the SilentFect transfection reagent (BioRad). The cells were left for 48 h prior to use in the various assays. To interfere with Golgi homeostasis, BJ/SV40T cells were treated with 1 μg/ml Brefeldin A (Sigma-Aldrich) for 20 min, or 5 μM nocodazole (Sigma-Aldrich) for 1 h prior to fixation. For the immunofluorescence analysis, the cells were fixed in 3% paraformaldehyde in phosphate buffer saline (PBS) for 25 min at 37 1C, and washed with PBS. Cells stained for endogenous WHAMM were instead fixed in ice-cold methanol for 3.5 min. The fixed cells were permeabilized in 0.2% Triton X-100 in PBS for 5 min, washed in PBS, and incubated in 5% FBS in PBS for 30 min at room temperature. Primary and secondary antibodies were diluted in PBS containing 5% FBS. The cells were incubated with the primary and secondary antibodies for intervals of 1 h, followed by washing in PBS. The coverslips were mounted on object slides by the use of FluoromountG (Southern Biotechnology Associates) or, for STED analysis, with Mowiol 4-88 (Sigma-Aldrich). The cells stained with the antibody against Mannosidase II were permeabilized in 0.5% saponin in PBS for 5 min and in the consecutive steps, 0.1% saponin was included in the blocking solution, the washing solution and in the antibody incubation solutions. The cells were photographed under a Zeiss AxioVert 40 CFL microscope equipped with a Zeiss AxioCAM MRm digital camera and the AxioVision software. The cellular effects induced by ectopic expression were determined by the microscopy analysis. The Golgi was considered to be dispersed when the Golgi ribbon became fragmented into vesicles. We avoided quantifying cells that were undergoing mitosis since the Golgi becomes fragmented into vesicles during the process. At least 100 cells were scored for each transfection condition. The entire procedure was repeated at least three times, to allow the statistical analysis of the observed differences in cell morphology. The statistical analysis was performed using Students' t-tests.

qPCR Total cellular RNA was isolated from cell cultures by the Trizol phenol/chloroform extraction method. The RNA was reverse transcribed using Omniscript reverse transcriptase (Qiagene) and oligo (dT) primers (Life Technologies). The following primers were used: human RhoD (forward 50 -TGGTCAACCTGCAAGTGAA-30 and reverse 50 -GGAGGCGGTCATAGTCATC-30 ), and human GAPDH (forward 50 GGA AGG TGA AGG TCG GAG TCA-30 and reverse 50 -ATG GGT GGA ATC ATA TTG GAA CA-30 ). 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.

Cell culture, transfection and immunofluorescence Protein transport assay BJ human foreskin fibroblast stably transfected with hTERT and SV40 large T antigen (BJ/SV40T cells) and green monkey COS1 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. The cells were cultured at 37 1C in 5% CO2. The cells were transfected using Lipofectamine (Invitrogen) or JetPEI reagents (PolyPlus Transfection), according to the protocol provided by the manufacturer.

In essence, the original protocol from Presley et al. was used [11]. In brief, the cells were transfected with EGFP-VSV-G alone or together with the RhoD mutants or WHAMM. In the experiments using siRNAs, the cells were transfected with EGFP-VSV-G 24 h after transfection of the siRNAs. The cells were then kept at 40 1C overnight, and the transport of EGFP-VSV-G from the ER to the plasma

210

EX PE R IM E NTAL C ELL RE S EA RC H

membrane via the Golgi apparatus was initiated by transfer of the cells to 32 1C for the periods described in Figs. 4 and 6. At least 100 cells were analyzed for each condition and the experiments were repeated at least three times. The cells were scored as VSV-G in ER when no traces of other localizations were noticed. The Golgi localization was somewhat difficult to assay since manipulation of RhoD and WHAMM resulted in dispersion of the Golgi and, as a result, the VSV-G staining appeared vesicular. The vesicles were however positive for Golgi markers and therefore we named them Golgi-derived vesicles. The cells with VSV-G localized to the plasma membrane also had staining in Golgi-derived vesicles, but we reasoned that the plasma membrane localization demonstrated that the transport was in fact intact. The ERþGVþplasma membrane (PM) localization makes up 100%.

Endo H assay and Western blotting The BJ/SV40T cells were transfected in duplicate with the constructs described in Fig. 4, and incubated at 40 1C for 24 h. One set of cells was transferred to 32 1C for 30 min, and another set remained at 40 1C during this time. Both sets of cells were then placed on ice, washed in ice-cold PBS, and scraped off using a rubber policeman. The cells were then divided in two pre-chilled micro centrifuge tubes (“Tube 1” and “Tube 2”). The cells were pelleted by centrifugation at 4 1C and the PBS was removed. “Tube 1” samples were resuspended in lysis buffer (20 mM Hepes, pH 7.5, 1% Triton X-100, 10% glycerol, 100 mM NaCl, 5 mM EDTA, and protease inhibitor cocktail) and incubated on ice for 20 min. These samples were thereafter centrifuged at RCF 16,000  g for 5 min at 4 1C. The supernatants were transferred into fresh tubes and SDSPAGE loading dye was added. The “Tube 2” cells were resuspended in 500 ml ice-cold PBS, and saponin was added to a final concentration of 0.1%. The samples were incubated for 1 min by slowly turning the tubes, followed by a brief centrifugation at full speed to pellet the cells. The saponin was removed and the cells were re-suspended in 20 ml 10 mM Na2HPO4, pH 5.5 and 1 ml EndoH (1U/200 ml) was added, followed by incubation at 37 1C for 1 h. The reaction was terminated by adding SDS-PAGE loading dye to the samples. The samples were subjected to SDS-PAGE and the proteins were transferred to nitrocellulose membranes (Hybond-C Extra, GE Healthcare). Immunoblotting was performed with GFP-, Flag- and Myc-specific antibodies, followed by horseradishperoxidase-conjugated anti-mouse or anti-rabbit secondary antibodies. The immunoblots were developed using Luminol Western blotting substrate (Santa Cruz).

Stimulated emission–depletion super-resolution microscopy Super-resolution STED images were acquired using a Leica SP5 TCS STED system equipped with two pulsed excitation lasers (exc1: 531 nm; exc2: 640 nm) and a high-power tunable near-infrared laser (selected STED wavelength 765 nm). We used a Leica STED 100  /1.4 oil objective, which was chromatically red-shifted to provide a better focal overlap of all of the excitation and STED wavelengths. The photons emitted from the AlexaFluor594 and Atto647N or STAR635 fluorophores were collected on two singlephoton sensitive avalanche photodiodes, using a pinhole size of 0.8 Airy units to partly remove background and stray light. Separation of the emitted wavelengths was achieved using a dichroic mirror and

333 (2 015 ) 20 8 –219

two band-pass filters (em1: 582/75 nm; em2: 680/40 nm). The STED images were collected in sequential mode, with a scanning speed of 400 Hz and applying line averaging four times. The size of the images was set to 25.83 mm  25.83 mm (zoom 6). Calibration images of 40 nm fluorescent beads (FluoSpheres, Invitrogen) showed a maximum averaged resolution of 46 nm full-width-halfmaximum (n¼ 158) for the Leica STED system.

Results RhoD regulates Golgi homeostasis To study the localization of endogenous RhoD, the cells of a human foreskin fibroblast cell line (BJ/SV40T cells) [12] were stained with a RhoD-specific antibody. The immunostaining data demonstrated an accumulation of RhoD in the perinuclear region in vicinity of the microtubule organizing center, as revealed by counterstaining with an antibody against α-tubulin (Fig. 1A). This staining pattern resembled the Golgi apparatus and co-staining with the cis-Golgi protein GM130 demonstrated that RhoD localizes to the Golgi area (Fig. 1B). Low-level ectopic expression of wild-type Myc-tagged RhoD was also accumulated in the Golgi (Fig. 1C), however, higher expression levels resulted in Golgi dispersion (Fig. 3F). We next transfected the cells with a plasmid that expresses ArfGAP, a protein previously shown to localize to the Golgi apparatus [13], and observed that endogenous RhoD, as well as ectopically expressed RhoD, localized together with ArfGAP in the Golgi area (Fig. 1D and E). However RhoD did not overlap completely with GM130 or with ArfGAP, which indicated that the proteins to an extent occupy separate compartments within the Golgi apparatus. In this regard, the RhoD localization resembled that of the Golgi membrane protein mannosidase II, [14], which exhibited only partial overlap with GM130 (Fig. 1F). To get a higher resolution view of the RhoD Golgi localization, we analyzed the cells by stimulated emission–depletion (STED) super resolution microscopy. This demonstrated that RhoD localizes to the Golgi area but it co-localized to lesser extent with GM130 (Fig. 1G). A higher degree of co-localization was found with the trans-Golgi marker β-galactosyltransferase (GalT), suggesting that RhoD is occupying a part of the Golgi apparatus that partially coincide with the trans-Golgi (Fig. 1H). To obtain further insight into the Golgi localization of RhoD, we treated the BJ/SV40T fibroblasts with the Golgi disrupting drugs brefeldin A (BFA) and nocodazole (NZ). BFA is an inhibitor of ER-toGolgi transport, and NZ treatment results in depolymerizing of microtubules, which in turn perturbs the organization of the Golgi apparatus [15,16]. Both of these treatments resulted in dispersion of the RhoD-positive vesicles (Fig. 2A). Interestingly, the RhoD bindingpartner WHAMM has previously been shown to localize to the Golgi apparatus [10], and we found that the localization of the endogenous protein overlapped with GM130 (Fig. 2B), however, even low-level ectopic expression of WHAMM efficiently resulted in Golgi dispersion (Fig. 2C). Similar to RhoD, the WHAMM localization to Golgi was disrupted by BFA and NZ treatments, but WHAMM appears to coreside with GM130 also after Golgi disruption (Fig. 2D). Unfortunately, RhoD and WHAMM could not be analyzed in the same cells as the anti-RhoD antibody only works with paraformaldehyde fixation, and the anti-WHAMM antibody only with methanol fixation. It is clear

E XP E RI ME N TAL C E L L RE S E ARCH

333 (2 015) 2 08 – 219

211

Fig. 1 – RhoD localizes to the Golgi apparatus. (A) Endogenous RhoD was detected using a rabbit anti-RhoD antibody followed by a TRITCconjugated anti-rabbit antibody. Microtubules were detected using a mouse anti-α-tubulin antibody followed by an AlexaFluor488conjugated anti-mouse antibody. Scale bar, 10 μm. (B) Endogenous RhoD was detected using a rabbit anti-RhoD antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by a TRITC-conjugated anti-mouse antibody. Nuclei were visualized using DAPI. Scale bar, 20 lm. (C) Ectopically expressed wild-type RhoD was detected using a rabbit anti-Myc antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Scale bar, 20 μm. (D) Endogenous RhoD was detected using a rabbit anti-RhoD antibody followed by a TRITC-conjugated anti-rabbit antibody. FLAG-ArfGAP was detected using a mouse anti-FLAG antibody followed by an AlexaFluor488-conjugated anti-mouse antibody. Scale bar, 20 μm. (E) Myctagged wild-type RhoD was detected using a rabbit anti-Myc antibody followed by an AlexaFluor568-conjugated anti-rabbit antibody. FLAG-ArfGAP was detected using a mouse anti-FLAG antibody followed by an AlexaFluor488-conjugated anti-mouse antibody. Scale bar, 20 μm. (F) Mannosidase II (ManII) was detected using a rabbit anti-ManII antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. GM130 was detected using a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Scale bar, 20 μm. (G and H) STED analysis of RhoD. (G) RhoD was detected using a rabbit anti-RhoD antibody followed by an Atto647Nconjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor594conjugated anti-mouse antibody. Scale bar, 2 μm. (H) RhoD was detected using a rabbit anti-RhoD antibody followed by a STAR635conjugated anti rabbit antibody. GalT was detected using a mouse anti-GalT antibody followed by an AlexaFluor594-conjugated antimouse antibody. Scale bar, 2 μm.

that, although RhoD and WHAMM both localized to the Golgi apparatus, they occupy domains of the Golgi that do not fully overlap. Ectopically expressed RhoD has been shown to localize to early endosomes and to the plasma membrane [2,17]. To study further the localization of exogenously expressed RhoD, the BJ/SV40T cells (data not shown) and COS1 cells were transiently transfected with wild-type RhoD or the constitutively active RhoD/G26V and the

dominant negative RhoD/T31N mutants (Fig. 3A–D). The G26V mutation locks RhoD in a GTP-bound conformation, whereas the T31N represents a GDP-bound conformation [6]. The cells were analyzed for co-localization of the RhoD variants and GM130. Wild-type RhoD localized to the Golgi area but also to vesicles and to the plasma membrane (Fig. 3B). In contrast, RhoD/G26V localized predominantly to vesicles in agreement with earlier

212

EX PE R IM E NTAL C ELL RE S EA RC H

333 (2 015 ) 20 8 –219

Fig. 2 – RhoD Golgi localization is disrupted by brefeldin A and nocodazole. (A) BJ/SV40T cells were treated with 1 μg/ml brefeldin A (BFA) for 20 min or 5 μM nocodazole (NZ) for 1 h. Endogenous RhoD was detected using a rabbit anti-RhoD antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Scale bar, 20 μm. (B) Endogenous WHAMM was detected using a guinea pig anti-WHAMM antibody followed by a TRITC-conjugated anti-guinea pig antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor488-conjugated anti-mouse antibody. Scale bar, 20 μm. (C) FLAG-WHAMM was detected using a rabbit anti-FLAG antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Scale bar, 20 μm. (D) BFA and NZ treatments were performed as in panel (A). Endogenous WHAMM was detected using a guinea pig anti-WHAMM antibody followed by a TRITC-conjugated anti-guinea pig antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor488-conjugated anti-mouse antibody. Scale bar, 20 μm.

observations, where RhoD was found on endosomes [2,17]. One reason for this apparent dichotomy became clear when cells expressing these RhoD mutants were stained with antibodies against the Golgi marker GM130. A marked dispersion of the Golgi apparatus was noted in over 44% of the cells expressing RhoD/G26V and 50% of the cells expressing RhoD/T31N (Fig. 3C, D and F). Expression of RhoDwt also induced Golgi dispersion when expressed at higher levels compared to Fig. 1C and E. A similar dispersion was noted when these cells where stained for the trans-Golgi marker TGN46 (data not shown). We also analyzed whether overexpression of WHAMM had a similar effect on Golgi homeostasis in COS1 cells. In agreement with published observations by Campellone et al. [10], a dispersion of the Golgi was found in 95% of the WHAMM-expressing cells (Fig. 3E and F).

RhoD controls ER-to-plasma membrane transport We argued that the effects on Golgi homeostasis caused by ectopic expression of RhoD would interfere with the protein transport from the ER to the plasma membrane. To investigate this hypothesis, an EGFP-tagged temperature-sensitive mutant of the vesicular stomatitis virus protein G (VSV-G) was used to transfect COS1 cells. At 40 1C, this mutant EGFP-VSV-G is trapped in the ER. However, when the

temperature is shifted to the permissive temperature of 32 1C, EGFPVSV-G is transported via the Golgi apparatus to the plasma membrane [11]. In agreement with this model, an initial accumulation of VSV-G in the ER was observed in cells transfected with the EGFP-VSV-G expressing plasmid alone. After 15 min at the permissive temperature (32 1C), VSV-G was translocated to the Golgi apparatus. The ER localization essentially disappeared after 60 min, and the majority of the VSV-G was at the plasma membrane, with trace amounts remaining in the Golgi apparatus (Fig. 4A and E). In contrast, when the cells were co-transfected with EGFP-VSV-G and the constitutively active RhoD/G26V or the dominant-negative RhoD/T31N, virtually no VSV-G was accumulated in a compartment resembling the Golgi after 15 min, instead VSV-G appeared in vesicular structures (Fig. 4B and C). Since ectopic expression of RhoD mutants triggered Golgi dispersion, we analyzed if these vesicles were positive for a Golgi marker. A majority of the VSV-G vesicles were also positive for GM130 and therefore we referred to them in our analysis as Golgi-derived vesicles (GV). We found that the VSV-G was essentially trapped in these vesicles after 60 min in the cells co-transfected with RhoD/G26V or RhoD/T31N and that significantly less VSV-G reached the plasma membrane (Fig. 4F and G). To further analyze the effects of RhoD on ER-to-Golgi transport of VSV-G, we performed the so called endoglycosidase H (Endo H) protection assay [18]. This assay is based on the

E XP E RI ME N TAL C E L L RE S E ARCH

333 (2 015) 2 08 – 219

213

Fig. 3 – RhoD is involved in the regulation of Golgi homeostasis. (A) Mock-transfected control cells, i.e. transfected with empty Myc-tag vector. The Golgi apparatus was visualized using a mouse anti-GM130 antibody followed by AlexaFluor488-conjugated anti-mouse antibody. Scale bar, 20 lm. (B–E) Myc-tagged wild-type RhoD (B), Myc-tagged RhoD/G26V (C), Myc-tagged RhoD/T31N or FLAG-tagged WHAMM (E) was transfected into COS1 cells. Myc-tagged RhoD/G26V and Myc-tagged RhoD/T31N, or FLAG-tagged WHAMM were detected using rabbit anti-Myc or anti-FLAG antibodies followed by a TRITC-conjugated anti-rabbit antibody. The Golgi apparatus was visualized using a mouse anti-GM130 antibody followed by AlexaFluor488-conjugated anti-mouse antibodies. The white asterisk marks transfected cells. Scale bar, 20 μm. (F) Quantification of dispersion of the Golgi apparatus caused by ectopic expression of RhoD and WHAMM, as performed by the microscopy analysis. A hundred cells for each condition were scored for Golgi dispersion. The columns represent analysis from three independent experiments. Data are means 7standard deviation. ***, po0.001.

finding that the oligoglycosylated VSV-G is no longer accessible for Endo H activity when it has reached the Golgi compartments [19]. At 40 1C, VSV-G localizes to the ER, and as a result, EndoH effectively cleaves the oligosaccharide modification, which is seen as a downshift in the mobility of VSV-G when analyzed by SDS-PAGE and Western blotting (Fig. 4I). When cells expressing VSV-G alone are incubated at 32 1C for 30 min, the VSV-G is predominantly localized to the Golgi apparatus, and as a result, a significant portion of the protein becomes resistant to the Endo H cleavage. This can be seen as the appearance of a higher molecular weight band on Western blotting (Fig. 4I). In contrast, when VSV-G was co-expressed with RhoD/G26V or RhoD/ T31N, the VSV-G was no longer protected against EndoH, as a downshift is seen (Fig. 4J and K). This suggests that in cells expressing RhoD mutants, the VSV-G localized to Golgi-derived vesicles is not protected against EndoH digestion and that the Golgi-derived vesicles did not fully retain the function of the intact Golgi apparatus. Ectopic expression of WHAMM has previously been shown to inhibit ER-to-Golgi transport of VSV-G [10]. To investigate if WHAMM

has the same effect in our cell system, we co-transfected COS1 cells with the FLAG-WHAMM and EGFP-VSV-G expression vectors. Indeed, similar to the mutant RhoD, WHAMM expression resulted in the formation of Golgi-derived vesicles, the ER-to-Golgi transport was significantly delayed and no VSV-G reached the plasma membrane (Fig. 4D). Also, the VSV-G that localized to the Golgi-derived vesicles was not protected against EndoH digestion (Fig. 4H and L).

Knockdown of RhoD affects Golgi homeostasis We next analyzed whether and how silencing of RhoD affects homeostasis of the Golgi apparatus. To this end, the BJ/SV40T cells were transfected with two different small-interfering (si) RNAs targeting RhoD. As the anti-RhoD antibody does not work for Western blotting, we validated the function of the siRNAs using immunofluorescence staining. The RhoD-specific siRNAs significantly quenched RhoD staining (Fig. 5A). Knockdown was also confirmed by qPCR and showed that the RhoD siRNA#2 was more

214

EX PE R IM E NTAL C ELL RE S EA RC H

333 (2 015 ) 20 8 –219

Fig. 4 – RhoD and WHAMM have roles in anterograde protein transport. The effects of RhoD and its binding partner WHAMM on protein transport from the ER to the plasma membrane via the Golgi apparatus was determined by transfecting EGFP-tagged VSVG alone (A), or together with Myc-tagged RhoD/G26V (B) Myc-tagged RhoD/T31N (C), or FLAG-tagged WHAMM (D). The transfected cells were kept at 40 1C overnight and then moved to 32 1C for 0, 15 or 60 min prior to fixing. VSV-G was visualized through the presence of the EGFP tag. GM130 was visualized using a mouse antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Myc-tagged RhoD/T31N was visualized using a rabbit anti-Myc antibody followed by an AMCA-conjugated anti-rabbit antibody. FLAG-tagged WHAMM was visualized using a rabbit anti-FLAG antibody followed by an AMCA-conjugated anti-rabbit antibody. Scale bar, 20 μm. Quantification of the effects on the transport of EGFP-VSV-G from the ER to the Golgi apparatus and to the plasma membrane was determined by microscopy analysis, as described in Materials and methods. Data are means 7standard deviation. The cellular effects on VSV-G transport caused by expression of RhoD or WHAMM were compared to cells transfected with VSV-G alone using Students’ t-test. ***, po0.001; **, po0.01; *, po0.05; n.s., non-significant. (E–H) The Endo H protection assay was performed as described in Materials and methods. Cells were collected after 0 or 30 min and the lysates were subjected to Endo H treatment for 1 h followed by SDS-PAGE and Western blotting with antibodies against GFP to detect EFGP-VSV-G (I), Myc to detect Myc-RhoD/G26V (J), Myc-RhoD/T31N (K), or FLAG to detect FLAG-WHAMM (L).

E XP E RI ME N TAL C E L L RE S E ARCH

333 (2 015) 2 08 – 219

215

Fig. 5 – Silencing of RhoD results in derailed Golgi organization. (A) The effect by RhoD silencing on the organization of the Golgi apparatus in BJ/SV40T fibroblasts was analyzed using a mouse antibody against GM130 followed by a TRITC-conjugated anti-mouse antibody. The efficiency of RhoD silencing was analyzed in cells treated with control or RhoD-specific siRNAs. The cells were stained for endogenous RhoD using a rabbit anti-RhoD antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. Scale bar, 20 μm. (B) Expression levels of RhoD mRNA 48 h after transfection with either siRNA targeting RhoD or control siRNA was analyzed by qPCR. The values were normalized against the expression level of GAPDH. (C) Quantification of the effect on the Golgi apparatus caused by RhoD or WHAMM silencing was determined using microscopy as described in Materials and methods. Data are means 7standard deviation. ***, po0.001. (D) The efficiency of the WHAMM-specific siRNA was revealed by transfecting siRNAs into BJ/SV40T fibroblasts. WHAMM was detected using a rabbit anti-WHAMM antibody followed by an HRP-conjugated antirabbit antibody. (E) The effect on the organization of the Golgi apparatus and the localization of RhoD when silencing WHAMM in BJ/SV40T fibroblasts was analyzed using a rabbit anti-RhoD antibody followed by an AlexaFluor488-conjugated anti-rabbit antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated antimouse antibody. Scale bar, 20 μm. (F) The effect on the organization of the Golgi apparatus and the localization of WHAMM when silencing RhoD in BJ/SV40T fibroblasts was analyzed using a guinea pig anti-WHAMM antibody followed by a TRITC-conjugated anti-guinea pig antibody. The Golgi apparatus was detected using a mouse anti-GM130 antibody followed by an AlexaFluor488conjugated anti-mouse antibody. Scale bar, 20 μm. (G) BJ/SV40T cells were transfected with control siRNA or siRNA targeting RhoD or WHAMM. Thereafter, the cells were transfected with empty Myc-vector or vectors expressing wild-type Myc-RhoD, Myc-RhoD/ T31N or Myc-WHAMM. Quantification of the effect on the Golgi apparatus was determined using microscopy as described in Materials and methods. Data are means 7standard deviation. **, po0.01; *, po0.05; n.s., non-significant.

efficient than siRNA#3 (Fig. 5B). BJ/SV40T cells were transfected with the RhoD-specific siRNAs and the organization of the Golgi was analyzed with antibodies against GM130. The knockdown of RhoD resulted in a significant dispersion of the Golgi apparatus in

37% of the cells (Fig. 5A, quantified in C). A similar result was seen when the siRNA-treated cells were stained for TGN46 (data not shown). We also analyzed whether WHAMM is required for Golgi homeostasis, using siRNA. Western blotting demonstrated that

216

EX PE R IM E NTAL C ELL RE S EA RC H

the siRNAs targeting WHAMM effectively knocked down the expression of WHAMM (Fig. 5D). Silencing of WHAMM resulted in dispersion of the Golgi apparatus in 26% of the cells i.e. with less efficiency compared to the silencing of RhoD (Fig. 5E, quantified in C). The simultaneous knockdown of RhoD and WHAMM aggravated the effect and caused Golgi dispersion in 65% of the cells (Fig. 5C). Notably, when RhoD was knocked down, WHAMM was dispersed in the cytoplasm (Fig. 5F). Conversely, knockdown of WHAMM resulted in dispersion of RhoD (Fig. 5E). To study the relationship between RhoD and WHAMM in the regulation of Golgi homeostasis, we transfected RhoD or WHAMM in cells treated with RhoD- or WHAMM-specific siRNAs. WHAMM expression in WHAMM silenced cells did not rescue the Golgi dispersion phenotype, indicating that WHAMM expression might need to be precisely balanced in cells. In contrast, wild-type RhoD expression in RhoD silenced cells resulted in a partial, but significant, reversion (from 44% in RhoD siRNAþempty vectorexpressing cells to 32% in RhoD siRNAþwild-type Rho-expressing cells) (Fig. 5G). Expression of RhoD/T31N in RhoD siRNA cells did not rescue the phenotype. A similar result was seen with RhoD/ G26V (data not shown), indicating that RhoD needs to be able to cycle between GTP- and GDP-bound conformations in order to rescue the phenotype caused by RhoD depletion. We also tested if the WHAMM-induced Golgi dispersion was dependent on RhoD. However, WHAMM expression induced Golgi dispersion to the same extent in control cells as in RhoD-depleted cells. In contrast, the Golgi dispersion caused by wild-type RhoD was significantly decreased in cells depleted for WHAMM (from 31% to 14%) (Fig. 5G). This might suggest that WHAMM is functioning downstream of RhoD in a pathway controlling Golgi homeostasis. Finally, we tested whether silencing of RhoD or WHAMM results in derailed ER-to-plasma membrane transport. For these experiments, the BJ/SV40T cells were treated with control, RhoD or WHAMMspecific siRNAs, and transfected with EGFP-VSV-G 24 h later (Fig. 6A– C). siRNA treatment resulted in a Golgi dispersion but the VSV-G could still be transported to vesicles positive for GM130, indicating that they are derived from the Golgi. The efficiency of the anterograde transport was however affected and, after 60 min, a larger portion of the VSV-G resided in the Golgi-derived vesicles and significantly less VSV-G reached the plasma membrane compared to in control cells (Fig. 6A–C). Notably, cells transfected with both RhoD and WHAMM siRNA were less efficient in transporting VSV-G from the Golgi to the plasma membrane than either of the single knockdowns (Fig. 6D).

Discussion RhoD has previously been shown to localize to early endosomes and to recycling endosomes, and expression of the constitutively active RhoD/G26V mutant has been shown to negatively regulate endosome trafficking [2,17,20]. In addition to these effects, we have occasionally observed an accumulation of ectopically expressed RhoD in the perinuclear area, although these overexpression experiments did not indicate that RhoD localizes to the Golgi apparatus. The explanation for this apparent contradiction appears to reside in the finding that ectopic expression of RhoD causes a rather dramatic dispersion of the Golgi apparatus, which makes Golgi localization difficult to study under the overexpression conditions. When we expressed wild-type RhoD at low level, we clearly noticed Golgi localization. The RhoD

333 (2 015 ) 20 8 –219

binding partner WHAMM was originally identified as a Golgi-enriched protein [10]. WHAMM is a multi-domain protein with multiple functions and the N-terminal part of WHAMM binds phosphorylated inositides and appears to mediate the localization to the Golgi apparatus. Intriguingly, this part of the protein also binds RhoD [7]. Moreover, a central coiled-coil domain binds microtubules, and the Cterminal domain consists of a WCA domain, which mediates binding to the Arp2/3 complex and can trigger actin polymerization [10]. With these characteristics, WHAMM can integrate cytoskeletal dynamics with Golgi homeostasis. We have previously identified WHAMM as a RhoD binding protein [7], which makes WHAMM an ideal candidate in a RhoD-regulated pathway in the control of Golgi homeostasis. Here, we demonstrate that endogenous RhoD and WHAMM localize to the Golgi apparatus and have obvious roles in anterograde transport of vesicles from the ER via the Golgi to the plasma membrane. Ectopic expression of WHAMM, and of constitutive active and dominant negative mutants of RhoD resulted in deregulated Golgi homeostasis and in Golgi dispersion. However, it is clear that the dispersed Golgi can to some extent receive VSV-G from the ER, indicating that the Golgi is still, at least partially, functional. The EndoH protection assay, however, indicated that this dispersed Golgi is not fully functional, since it cannot protect VSV-G from EndoH digestion. Campellone et al. reported that WHAMM expression resulted in a significantly reduced ER-to-Golgi transport [10]. Our analysis demonstrated that the dispersed Golgi, caused by the ectopic expression of WHAMM and also RhoD mutants, still can receive VSVG from the ER, but the process is much slower. After 60 min at permissive temperature, only about 20% of the RhoD mutant expressing cells, and only 1% of the WHAMM expressing cells show VSV-G staining in the plasma membrane, as compared to 71% for the control (Fig. 4). Regarding RhoD/G26V, this is in line with the previous finding that this mutant can interfere with vesicle transport [2]. Both RhoD and WHAMM silencing gave rise to an altered Golgi homeostasis. However, this phenotype is not necessarily associated with an impaired VSV-G transport from ER to the plasma membrane. Suga et al. found that silencing of Syntaxin 5, a protein residing in the ER-to-Golgi compartment, gives a fragmented Golgi, but does not affect the VSV-G transport [21]. We could see that silencing of RhoD or WHAMM did not affect the transport between ER and Golgi. However, the subsequent transport from Golgi to the plasma membrane was significantly delayed, in particular in cells with both reduced RhoD and WHAMM expression. In absence of RhoD or WHAMM, the VSV-G reached the plasma membrane in 45– 55% of the cells compared to 85% in control cells after 60 min at 32 1C. For the double knockdown, the corresponding number is down to 33%, revealing the importance of RhoD and WHAMM in this process (Fig. 6). Since the trans-Golgi is an exit site for proteins transported to the plasma membrane, it is interesting that RhoD was found to co-localize with the trans-Golgi marker GalT to a greater extent than the cis-Golgi marker GM130, when analyzed with STED microscopy (Fig. 1G and H). Knockdown of both RhoD and WHAMM was not enough to fully abrogate the plasma membrane localization of VSV-G. This suggests that there are other factors involved in this process. One such factor could be JMY, which is a WHAMM-related protein. A recent study by Schlüter et al. showed that JMY is involved in vesicle transport from the trans-Golgi network [22]. Among the other Rho GTPases, both Cdc42 and RhoA have been found to localize to the Golgi area. Erickson et al. reported that Cdc42 is a component of what they referred to as the BFA-sensitive compartment of the Golgi apparatus [23]. Treatment of BFA resulted

E XP E RI ME N TAL C E L L RE S E ARCH

333 (2 015) 2 08 – 219

217

Fig. 6 – Silencing of RhoD results in derailed ER-to-Golgi transport BJ/SV40T cells treated with control siRNA. (A) or siRNA against RhoD (B), WHAMM (C) or both RhoD and WHAMM (D) were after 24 h transfected with EGFP-VSVG. The transfected cells were kept at 40 1C overnight and moved to 32 1C for 0, 15 or 60 min prior to fixing. The EGFP-VSV-G was visualized by the presence of the fluorescent protein and the organization of the Golgi apparatus was visualized by a mouse anti-GM130 antibody followed by an AlexaFluor568-conjugated anti-mouse antibody. Scale bar, 20 μm. Quantification of the effect on the transport of EGFP-VSV-G from the ER to the Golgi apparatus and finally the plasma membrane was carried out using microscopy analysis, as described in Materials and methods. Data are means 7standard deviation. The cellular effects on VSV-G transport caused by silencing of RhoD and WHAMM were compared to cells transfected with VSV-G alone using Students’ t-test. ***, po0.001; **, po0.01.

218

EX PE R IM E NTAL C ELL RE S EA RC H

in cytoplasmic redistribution of Cdc42, similar to what we have seen with RhoD after BFA treatment [23]. In addition, Cdc42 has a role in transport from the trans-Golgi network, to maintain transport to the apical side of epithelial cells [24]. One link between Cdc42 and the Golgi is via the Cdc42-specific GEF Tuba. GM130 binds Tuba and this interaction controls the level of Cdc42 activation, which in turn controls the organization and function of the centrosome [25]. Another link between Cdc42 and the Golgi apparatus is mediated by the Cdc42-GEF intersectin. Interruption of the Cdc42:intersectin interaction with a small molecule inhibitor results in disorganized Golgi [26]. In similarity to the atypical Rho GTPase Wrch1 and the tumor-related splice variant Rac1b, RhoD has the capability to cycle between its active and inactive conformations in the absence of GEFs and GAPs [4,27]. Although the mechanisms underlying the RhoDdependent effects on Golgi homeostasis are not entirely clear, it is possible that RhoD collaborate with WHAMM in the control of Golgi homeostasis. Alternatively, RhoD might signal to Cdc42 via a so far unidentified signaling pathway. There are additional observations suggesting that Rho GTPases have key functions in Golgi homeostasis. The RhoA effector protein Citron-N has also been shown to have a role in the regulation of Golgi dynamics [28]. Citron-N is a neuronal-specific splice variant of the RhoA effector Citron-K, which is a Ser/Thr protein kinase. Citron-N associates with the Golgi apparatus in hippocampal neurons, and the expression of a Citron-N mutant that cannot bind RhoA resulted in disruption of the Golgi ribbon [28]. In addition, a minor portion of RhoA was found to reside in the Golgi area, and inhibition of RhoA with the botulinum toxin C3 resulted in Golgi fragmentation. It has not been tested whether RhoD could bind Citron-N, and whether such an interaction can provide a link to Golgi dynamics. Taken together, our data highlight the involvement of the newly discovered signaling pathway involving RhoD and its binding partner WHAMM in the regulation of Golgi dynamics. Several observations indicate key roles for the cytoskeleton in the regulation of Golgi function. For instance, there is a close correlation between the organization of microtubules and Golgi dynamics [29]. In addition, actin polymerization is an important determinant for the maintenance of Golgi homeostasis, as actin disrupting drugs have been shown to interfere with Golgi function [30]. RhoD and WHAMM can potentially regulate Golgi dynamics independent of each other. However, the fact that WHAMM is a RhoD-binding protein, together with the current observation that WHAMM depletion interferes with the RhoDinduced Golgi dispersion, whereas WHAMM causes Golgi dispersion also in RhoD-depleted cells, suggest that they, at least partially, work in a pathway where WHAMM is acting downstream of RhoD.

Acknowledgments We thank Daniel Salamon (Karolinska Institutet) for expert help on the qPCR analysis. PA was supported by grants from the Swedish Cancer Society (Grant number 110549), the Swedish Research Council (Grant number K201066P20582044) and Karolinska Institutet. AG was supported by grants from Karolinska Institutet, Alex and Eva Wallström Foundation and OE and Edla Johansson Foundation. KR was supported by grants from Karolinska Institutet and OE and Edla Johansson Foundation.

333 (2 015 ) 20 8 –219

references [1] P. Chavrier, K. Simins, 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. Lütcke, R.G. Parton, M. Zerial, Endosome dynamics regulated by a Rho protein, Nature 384 (1996) 427– 432. [3] S. Ellis, H. Mellor, The novel Rho-family GTPase rif regulates coordinated actin-based membrane rearrangements, Curr. Biol. 10 (2000) 1387–1390. [4] 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. [5] D. Fiegen, L.C. Haeusler, L. Blumenstein, U. Herbrand, R. Dvorsky, I.R. Vetterand, M.R. Ahmadian, Alternative splicing of Rac1 generates Rac1b, a self-activating GTPase, J. Biol. Chem. 279 (2004) 4743–4749. [6] A.B. Jaffe, A. Hall, Rho GTPases: biochemistry and biology, Annu. Rev. Cell Dev. Biol. 21 (2005) 247–269. [7] A.K.B Gad, V. Nehru, A. Ruusala, P. Aspenström, RhoD regulates cytoskeletal dynamics via the actin-nucleation-promoting factor WHAMM, Mol. Biol. Cell 23 (2012) 4807–4819. [8] K. Tsubikamoto, 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. [9] 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. [10] K.G. Campellone, N.J. Webb, E.A. Znameroski, M.D. Welch, WHAMM is an Arp2/3 complex activator that binds microtubules and functions in ER to Golgi transport, Cell 134 (2008) 148–161. [11] J.F. Presley, N.B. Cole, T.A. Schroer, K. Hirschberg, K.J. Zaal, J. Lippincott-Schwartz, ER to-Golgi transport visualised in living cells, Nature 389 (1997) 81–85. [12] W.C. Hahn, C.M. Counter, A.S. Lundberg, R.L. Beijersbergen M.W. Brooks, R.A. Weinberg, Creation of human tumour cells with defined genetic elements, Nature 400 (1999) 464–468. [13] E. Cukierman, I. Huber, M. Rotman, D. Cassel, The ARF1 GTPaseactivating protein zinc finger motif and Golgi complex localization, Science 270 (1995) 1999–2002. [14] P.M. Novikoff, D.R. Tulsiani, O. Touster, A. Yam, A.B Novikoff, Immunocytochemical localization of alpha-D-mannosidase II in the Golgi apparatus of rat liver, Proc. Natl. Acad. Sci. USA 80 (1983) 4364–4368. [15] A.A. Rogalski, S.J. Singer, Associations of elements of the Golgi apparatus with microtubules, J. Cell Biol. 99 (1984) 1092–1100. [16] T. Fujiwara, K. Oda, S. Yokota, A. Takatsuki, Y. Ikehara, Brefeldin A causes disassembly of the Golgi complex and accumulation of secretory proteins in the endoplasmic reticulum, J. Biol. Chem. 263 (1988) 18545–18552. [17] V. Nehru, O. Voytyuk, J. Lennartsson, P. Aspenström, 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. [18] R. Schwaninger, H. Plutner, H.W. Davidson, S. Pind, W.E. Balch, Transport of proteins between endoplasmic reticulum and Golgi compartments in semi-intact cells, Methods Enzymol. 219 (1992) 110–124. [19] A.L. Tarentino, T.H. Plummer Jr, F. Maley, The release of intact oligosaccharides from specific glycoproteins by endo-beta-Nacetylglucosaminidase H, J. Biol. Chem. 249 (1974) 818–824. [20] C. Murphy, R. Saffrich, J.C. Olivo-Marin, A. Giner, W. Ansorge T. Fotsis, M. Zerial, Dual function of RhoD in vesicular movement and cell motility, Eur. J. Cell Biol. 80 (2001) 391–398.

E XP E RI ME N TAL C E L L RE S E ARCH

[21] K. Suga, H. Hattori, A. Saito, K. Akagawa., RNA interferencemediated silencing of the syntaxin 5 gene induces Golgi fragmentation but capable of transporting vesicles, FEBS Lett. 579 (2005) 4226–4234. [22] K. Schlüter, D. Waschbüsch, M. Anft, D. Hügging, S. Kind, J. Hänisch, G. Lakisic, A. Gautreau, A. Barnekow, T.E. Stradal., JMY is involved in anterograde vesicle trafficking from the trans-Golgi network, Eur. J. Cell Biol. 93 (2014) 194–204. [23] J.W. Erickson, C.-j. Zhang, R.A. Kahn, T. Evans, R.A. Cerione, Mammalian Cdc42 is a brefeldin A-sensitive component of the Golgi apparatus, J. Biol. Chem. 271 (1996) 26850–26854. [24] A. Müsch, D. Cohen, G. Kreitzer, E. Rodriguez-Boulan, cdc42 regulates the exit of apical and basolateral proteins from the trans-Golgi network, EMBO J. 20 (2001) 2171–2179. [25] A. Kodani, I. Kristensen, L. Huang, C. Sütterlin, GM130-dependent control of Cdc42 activity at the Golgi regulates centrosome organization, Mol. Biol. Cell 20 (2009) 1192–1200.

333 (2 015) 2 08 – 219

219

[26] A. Friesland, Y. Zhao, Y.H. Chen, L. Wang, H. Zhou, Q. Lu, Small molecule targeting Cdc42-intersectin interaction disrupts Golgi organization and suppresses cell motility, Proc. Natl. Acad. Sci. USA 110 (2013) 1261-126. [27] J. Saras, P. Wollberg, P. Aspenström, Wrch1 is a GTPase-deficient Cdc42-like protein with unusual binding characteristics and cellular effects, Exp. Cell Res. 299 (2004) 356–369. [28] P. Camera, J.S. da Silva, G. Griffiths, M.G. Giuffrida, L. Ferrara, V. Schubert, S. Imarisio, L. Silengo, C.G. Dotti, F. Di Cunto, Citron-N is a neuronal Rhoassociated protein involved in Golgi organization through actin cytoskeleton regulation, Nat. Cell Biol. 5 (2003) 1071–1078. [29] J. Thyberg, S. Moskalewski, Role of microtubules in the organization of the Golgi complex, Exp. Cell Res. 246 (1999) 263–279. [30] F. Valderrama, T. Babià, I. Ayala, J.W. Kok, J. Renau-Piqueras G. Egea, Actin microfilaments are essential for the cytological positioning and morphology of the Golgi complex, Eur. J. Cell Biol. 76 (1998) 9–17.