FEBS Letters 587 (2013) 328–338

journal homepage: www.FEBSLetters .org

Localization of Rab proteins to peroxisomes: A proteomics andimmunofluorescence study

0014-5793/$36.00 � 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.febslet.2012.12.025

⇑ Corresponding author. Address: Biochemistry Center Heidelberg (BZH), Uni-versity of Heidelberg, Im Neuenheimer Feld 328, 69120 Heidelberg, Germany.

E-mail address: wilhelm.just@bzh.uni-heidelberg.de (W.W. Just).1 Present address: Eijkman Institute of Molecular Biology, Jakarta, Indonesia.

Thomas Gronemeyer a, Sebastian Wiese b, Sören Grinhagens a, Lukas Schollenberger c, Ari Satyagraha c,1,Lukas A. Huber d, Helmut E. Meyer e, Bettina Warscheid b, Wilhelm W. Just c,⇑a Department of Molecular Genetics and Cell Biology, Ulm University, Germanyb Faculty of Biology, Functional Proteomics, and BIOSS Centre for Biological Signalling Studies, University of Freiburg, Germanyc Biochemistry Center Heidelberg (BZH), University of Heidelberg, Germanyd Innsbruck Biocenter, Innsbruck Medical University, Austriae Medizinisches Proteom-Center, Ruhr-Universität Bochum, Germany

a r t i c l e i n f o

Article history:Received 28 November 2012Revised 19 December 2012Accepted 19 December 2012Available online 16 January 2013

Edited by Gianni Cesareni

Keywords:PeroxisomeRab6Rab10Rab14Rab18ProteomicsMass spectrometryImmunofluorescencePeroxisome proliferationSubcellular localization

a b s t r a c t

A proteomics screen was initiated to identify Rab proteins regulating transport to and away fromperoxisomes. Mass spectrometry-based protein correlation profiling of rat liver organelles andimmunofluorescence analysis of the peroxisome candidate Rab proteins revealed Rab6, Rab10,Rab14 and Rab18 to associate with the peroxisomal membrane. While Rab14 localized to peroxi-somes predominantly in its dominant-active form, other Rab proteins associated with peroxisomesin both their GTP- and GDP-bound state. In summary, our data suggest that Rab6, Rab10, Rab14 andRab18 associate with the peroxisomal compartment and similar as previously shown for Rab8,Rab18 in its GDP-bound state favors peroxisome proliferation.

Structured summary of protein interactions:bifunctional enzyme, PEX11alpha, Rab-18, Rab-14, Rab-6A, Rab-10 and Rab-2A colocalize by cosedimen-tation through density gradient (View interaction)Catalase and Rab18 colocalize by fluorescence microscopy (View interaction)Rab14 and Catalase colocalize by fluorescence microscopy (View interaction)Rab6 and Catalase colocalize by fluorescence microscopy (View interaction)Rab10 and Catalase colocalize by fluorescence microscopy (View interaction)

� 2013 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction Synthesis of ether phospholipids including plasmalogens starts in

Functions of mammalian peroxisomes relate to oxidative degra-dation of metabolites, such as very long chain fatty acids includingpristanic acid, trihydroxycholestanoic acid and long-chain dicar-boxylic acids, D-amino acids, pipecolate, glycolate, hydroxy acids,spermidine and epoxides [1]. Besides the oxidative catabolism thatgenerates highly reactive oxygen species like hydrogen peroxide,superoxide anion and peroxinitrite, peroxisomes are implicatedin the synthesis of ether phospholipids and possibly isoprenoids[2–5]. Many of these functions for completion require activities lo-cated in other subcellular compartments. For example, very longchain fatty acid-CoAs may be chain-shortened to medium chain-CoAs, converted to carnitine derivatives and after export from per-oxisomes taken up by mitochondria for further degradation [1].

peroxisomes but is completed by enzymes located at the endoplas-mic reticulum (ER). Thus, peroxisomes metabolically communicatewith other cellular compartments. Do they also physically interactwith them? Recent studies indicate both in mammalian cells andin Saccharomyces cerevisiae vesicular transport of peroxisomalmembrane proteins (PMPs), such as Pex3p and Pex19p from theER to peroxisomes [6–8]. Further support for vesicular cycling ofcomponents between peroxisomes and the ER may be derivedfrom studies demonstrating ARF1 and COP I coat-binding to per-oxisomes [9–11] that may serve the retrograde transport of ER res-ident proteins and/or the vesicular exchange of materials withother cellular compartments, e.g. lipid droplets LDs) [12–15].

To be motile within cells, peroxisomes are transiently trans-ported both to the cell periphery and back to the cell center [16–19]. Failure of peroxisome–microtubule interactions results inclustering of peroxisomes [20]. Microtubule-based peroxisomalmotility is regulated both inside cells by RhoA and from the outsideby a signaling cascade including ATP and lysophosphatidic acidreceptor co-stimulation and heterotrimeric Gi/Go proteins

T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338 329

[16,17]. Recent studies further demonstrate peroxisomal associa-tion with actin, non-muscle myosin IIA, Rho kinase II and Rab8 pro-viding evidence that in addition to microtubules peroxisomes alsointeract with the acto–myosin complex dependent of the state ofactivity of RhoA [21].

Different to previous views that envisaged peroxisomes as rep-resenting a more or less autonomous cell compartment, recent ad-vances indicate multiple intracellular interactions and functionalexchange between peroxisomes and other organelles. As many ofthese activities are known to require regulatory switches, we initi-ated a proteomics study aiming at detecting peroxisomal associa-tions with small GTPases other than RhoA and Rab8. Byimmunofluorescence (IF) analysis we now confirm localization ofRab6, Rab10, Rab14 and Rab18 to the peroxisomal compartment.

2. Materials and methods

2.1. Isolation of peroxisomes

Highly purified peroxisomes from rat liver were isolated as pre-viously described [22]. Aliquots were stored in small aliquots at�80 �C prior to use.

2.2. Proteomics analyses of peroxisomes and peroxisomal membranes

For protein profiling Nycodenz gradient fractions were individ-ually treated with 100 mM NaCO3 and the resulting membranepellets subjected to SDS–PAGE. For further analysis the gel regionbetween 15–30 kDa was cut into three slices and after trypsintreatment the proteins were subjected to proteomics analysis asdescribed previously [21].

For peptide and protein identification, MS/MS datasets weregenerally correlated with the rat International Protein Index(IPI; www.ebi.ac.uk) database using the MASCOT algorithm. Pro-teins were identified based on at least one unique peptide witha false positive rate below 5%. Proteins of interest were semi-quantitatively followed across the density gradient fractions bycalculating the respective spectral counts and plotting theseagainst the respective gradient fractions to generate proteinprofiles [23].

2.3. Constructs and cloning

Templates encoding Rab2A (human), Rab18 (mouse; 99% se-quence identity with human Rab18) and Rab6 (human) werekindly provided by Kirill Alexandrov (MPI for Molecular Physiol-ogy, Dortmund, GER) [24]. Rab genes were amplified by PCR andsubsequently cloned into pDsRed-C1-monomer (BD Biosciences/Clontech, Heidelberg, GER) using the Sac1/Xma1 or Xho1/Xma1restriction sites. HA-tags (amino acids 96–108 from human influ-enza hemagglutinin) were inserted by PCR primers between theDsRed and the Rab ORF.

HA-Rab6 wildtype (wt) and HA-Rab18wt were constructed byremoving the DsRed by restriction with Age1 and Bgl2, filling upthe sticky ends with Klenow fragment treatment and subsequentblunt end ligation of the plasmid. HA-Rab6Q65L and HA-Rab6-N126I expression constructs [25] were a kind gift of GuangyuWu (Louisiana State University, New Orleans, USA). GFP-Rab10,GFP-Rab10Q68L and GFP-Rab10T23N were as previously published[26] and kindly provided by Kai Simons (MPI of Molecular Cell Biol-ogy, Dresden, GER). GFP-Rab18, GFP-Rab18Q67L and GFP-Rab18S22N were from Sally Martin (University of Queensland,Brisbane, AUS) [27] and Cherry-Rab14wt, Cherry-Rab14Q70L andCherry-Rab14S25N from Mary McCaffrey (Dept. of Biochemistry,University College Cork, IE) [28]. The peroxisomal marker plasmidpEGFP-SKL was from BD Biosciences/Clontech (Heidelberg, GER).

2.4. Cell culture

Human hepatocellular carcinoma (Huh7) wt cells were grownin DMEM containing 10% FCS (Gibco/Invitrogen, Karsruhe, GER).One day before transfection, cells were seeded on glass cover slipsat a density of 2.5 � 105 cells/ml. The medium was changed toDMEM containing 5% FCS and transfection performed using cal-cium phosphate. After 4 h the cells were incubated with 10% glyc-erol in DPBS (Gibco) for 1 min and grown for further 48 h in DMEMcontaining 10% FCS. Transfected cells were fixed for 10 min with 3%paraformaldehyde in DPBS, washed with DPBS and finally mountedon glass slides using VectaShield hard set mounting medium (Vec-tor Laboratories, Burlingame, CA, USA).

For stably expressing tagged Rab constructs, cells were seeded in6 well culture plates (TPP, Trasadingen, CH) at a density of2.5 � 105 cells/ml and transfected as described above. Twenty-fourhours after transfection cells were selected and maintained in med-ium containing 400 and 250 mg/l Geneticin (Gibco), respectively.

2.5. Immunofluorescence and microscopy

Cells grown on glass cover slips were fixed in 3% PFA and perme-abilized with 1% TX-100 for 5 min at room temperature. After block-ing with 10% BSA, cells were incubated for 1 h at 37 �C with guineapig anti-catalase primary antibody followed by anti-guinea pigAlexa546- or Alexa488-labeled secondary antibody (Invitrogen).Cover slips were mounted on glass slides using VectaShield. Digito-nin permeabilization was carried out prior to fixation using a con-centration of 80 lg/ml (Applichem, Darmstadt, DE) in DPBS [29].

HA-tagged Rab constructs were visualized by staining with amouse anti-HA antibody (Sigma, Taufkirchen, DE) and an anti-mouse Atto488- (Sigma) or Alexa546-labeled (Invitrogen) second-ary antibody. For IF demonstration of GFP the mouse anti-GFP anti-body (Roche Diagnostics, Mannheim, GER) was decorated with ananti-mouse Alexa546-labeled secondary antibody (Invitrogen).

Microscopy was done on an Observer SD confocal microscope(Zeiss, Göttingen, GER) equipped with 488 and 547 nm diode lasersand a 63-fold Plan-Apochromat objective with lens aperture 1.4.For image acquisition and processing the AxioVision 4.8.1 software(Zeiss) was used.

2.6. Image analysis

Prior to quantification of co-localizing structures, usually 6–15z-layers per stack were filtered using the Gauss filter implementedin the image acquisition software in order to remove noise. Subse-quently, stacks of the 488 and 547 nm channels were separatelyexported as 8-bit grey-scale images in TIF format and analyzedusing ImageJ (http://rsb.info.nih.gov/ij/) and the OBCOL plugin[30]. The total number of peroxisomes per cell was determinedwith the ‘‘3D-objects-counter’’ plugin for ImageJ.

The Pearson coefficient comparing the intensity distribution ofshapes between two different channels was used to determine co-localization. Voxels with a coefficient >0.5 were considered as co-local-izing. A value equal to 0.5 implies that 50% of the voxels in one imageoverlap with the corresponding voxels in the other image [31].

3. Results

3.1. Analytical studies

In pilot experiments analyzing the association of small GTP-bind-ing proteins to peroxisomes, isolated peroxisomes were first incu-bated in the presence of cytosol, an ATP-regenerating system andthe non-hydrolyzable GTP analogue GMP-PNP before re-isolatingand subjecting to 1D- or 2D-electrophoresis [32]. After transfer onto

Table 1Annotation of Rab proteins and Rab functional groups.

Group II Group VI Group VIII

Rab 2 Rab 6 Rab 8Rab 14 Rab10

330 T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338

nitrocellulose sheets, proteins were visualized by autoradiographyin presence of 32P-GTP [28]. While 1D-gels exhibited radio-labeledGTP-binding proteins in discrete regions of the gel at 30, 27, 24and 20 kDa, 2D-electrophoresis revealed multiple radioactive spotsout of which RhoA and Arf1 were identified by mass spectrometry(data not shown). To further characterize the association of Rab pro-teins to peroxisomal membranes, the post-nuclear supernatant of arat liver homogenate was separated by equilibrium density gradientcentrifugation and Rab protein abundance analyzed by a proteomicsscreen [33]. Using bifunctional protein and Pex11a as markers forthe peroxisomal matrix and membrane, respectively, protein abun-dance profiling [34] identified Rab2, Rab6, Rab10, Rab14, and Rab18as potential peroxisomal constituents (Fig. 1). Except Rab8 [21],these Rab proteins were also detected in the low-density micro-somal region of the gradient, however their maximal concentrationwas clearly correlated with the peroxisomal fraction.

Fig. 1. Protein abundance profiling for Rab2A, Rab6A, Rab10, Rab14, and Rab18. The pgradient centrifugation and the relative abundance of the Rab proteins (A) and the peropeptides using spectral counts. Bifunctional enzyme and Pex11a were used as markers f

Based on tree topology, eight functional groups of co-segre-gating Rab proteins were recently described that might reflectsequence similarities, cellular localization and/or functional cor-relation [35]. The Rab proteins identified by the present studyto supposedly target to peroxisomes were assigned to Rab func-tional group II (Rab14), VI (Rab6) and VIII (Rab8 and Rab10)(Table 1). So far Rab18 has not been assigned to any of thesegroups.

ost-nuclear fraction of a rat liver homogenate was separated by Nycodenz densityxisomal marker proteins (B) were determined by LC/MS analyses of specific trypticor the peroxisomal matrix and membrane (gradient fractions 3 and 4), respectively.

Fig. 2. IF co-localization in Huh7 cells stably expressing EGFP-Rab18wt (A), the Q67L (B) and S22N (C) mutants with catalase that was used as a peroxisomal marker. Forvisualization of catalase, guinea pig anti-catalase antibody was decorated with an Alexa546-labeled secondary antibody. The number of co-localizing structures (D) and theaverage number of total peroxisomes per cell (E) were determined by quantitative image analysis. The images were assembled from z-projections, the scale bar represents10 lm.

T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338 331

3.2. Peroxisomal localization of Rab18 and Rab10

Proper functioning of Rab proteins strictly depends on theirpost-translational modification by C-terminal prenylation. Differ-ent to other small GTPases that are prenylated by farnesyltransfer-ase (FT) and geranylgeranyl transferase type I (GGT-I), Rab family

members are geranylgeranylated by RGGT/GGT-II [36]. Whereasboth FT and GGT-I transfer the isoprenoid residue to the cysteineresidue of a C-terminal CaaX motif, RGGT/GGT-II recognizes Rabproteins in complex with Rab escort protein (REP) facilitating addi-tion to the Rab protein of two geranylgeranyl moieties. However,there is a small number of Rab family members, such as Rab8

Fig. 3. Differential permeabilization of Huh7 cells with digitonin. While Triton X-100 at a concentration of 1% permeabilized all cellular membranes (A), digitonin at aconcentration of 80 mg/ml permeabilized the cholesterol-rich PM but not the peroxisomal membrane (B). Thus, EGFP-SKL transiently expressed in Huh7 cells and importedinto the peroxisomal lumen was detected after Triton X-100 (mid, A) but not digitonin permeabilization (B). For staining EGFP-SKL the primary anti-EGFP antibody wasdecorated with an Alexa-546-labeled secondary antibody. Images were assembled from z-projections, the scale bar represents 10 lm.

332 T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338

and Rab18 that naturally possess a C-terminal CaaX motif [37].Interestingly, the aaX tripeptides of mammalian Rab8 and Rab18are consensus with a peroxisomal targeting signal 1 (PTS1), SLLand SVL, respectively. Both signals have been shown to interactin the two-hybrid system with the PTS1 receptor Pex5p [38]. More-over, mono-cysteine Rabs were subject to CaaX proteolysis by Rasand a-factor converting enzyme 1 (Rce1) and modified by isopre-nylcysteine carboxymethyltransferase (Icmt). Both enzymes werelocated at the ER. Importantly, in the absence of Rce1 and Icmtthe topology of EGFP-tagged and over-expressed Rab-CaaX pro-teins was unaffected [36] suggesting that SLL and SVL do not mis-target Rab8 and Rab18 to peroxisomes.

We first investigated the intracellular distribution of Rab18,particularly its EGFP-tagged wt, dominant-active (Q67L) and dom-inant-inactive (S22N) forms in Huh7 cells (Fig. 2). IF images ob-tained by confocal microscopy suggested co-localization of theexpression products with catalase-positive structures. A high de-gree of co-localization was found with both the dominant-activeand -inactive mutants (Fig. 2B–D). In cells stably transfected withthe dominant-negative mutant, the total number of peroxisomeswas increased by about 50% (Fig. 2E). A similar result was obtainedwith Rab8 [21]. Thus, both Rab8 and Rab18 localize to peroxisomesand overexpression of their dominant-negative mutants favor per-oxisome proliferation.

These observations raised the question as to the implication ofthe PTS1 in this recruitment [39], although normally the aaX C-ter-minal peptide is proteolytically cleaved off during post-translationalprocessing by Rce1. On the other hand, should the PTS1 be involvedin the peroxisomal localization, both Rab8 and Rab18 are expectedto be imported into the peroxisomal matrix by following the Pex5-PTS1 pathway. We therefore used Rab18 (with terminal aaX pep-tide) and Rab6 (without terminal aaX peptide) to study the peroxi-somal topology of the newly synthesized Rab proteins.

By differential permeabilization of Huh7 cells with Triton X-100and digitonin [29], we first investigated peroxisomal import ofEGFP-SKL (Fig. 3). In Triton X-100-permeabilized cells (Fig. 3A),EGFP-SKL was localized in punctate structures typical for peroxi-somes. The peroxisomal structures were visible by native EGFP

fluorescence (left) and by using an anti-GFP antibody (mid). Bothstructures completely co-localized (right). In digitonin-permeab-lized cells (Fig. 3B), the native EGFP fluorescence also produced aperoxisomal pattern (left), however as expected, the anti-EGFPantibody did not, as it was incapable to penetrate the non-perme-abilized peroxisomal membrane.

Subsequently, we investigated the subcellular distribution ofHA-Rab6wt and HA-Rab18wt in Huh7 cells after permeabilizationwith digitonin (Fig. 4). Both constructs were well expressed andusing an antibody directed towards the HA-tag both Rab6(Fig. 4A) and Rab18 (Fig. 4B) co-localized with the peroxisomalmarker EGFP-SKL. While the peroxisomal compartment was recog-nized by the endogenous EGFP fluorescence, the anti-catalase anti-body (Fig. 4C) could not stain peroxisomes due to the non-permeabilized peroxisomal membrane. A quite similar result wasobtained for EGFP-Rab8 (data not shown). These observations sug-gest that the newly expressed HA-Rab6, HA-Rab18 and EGFP-Rab8were not imported by a Pex5p-dependent pathway into the perox-isomal matrix but instead were bound to the cytosolic face of theperoxisomal membrane.

Unlike Rab8, Rab10 that also belongs to the Rab functionalgroup VIII [35] has a typical Rab C-terminal dicysteine prenylationmotif XXCC that not at all mediates Pex5p-supported PTS1 import.Expression of EGFP-Rab10 in Huh7 cells resulted in peroxisomaltargeting of the expression product as revealed by co-localizationof the EGFP signal with that of catalase. Both wt, dominant-activeand -inactive forms of EGFP-Rab10 localized to peroxisomes(Fig. 5A–C). Compared to wt and the dominant-active protein(Q68L), the dominant-inactive mutant (T23N) in addition to per-oxisomes also localized to the Golgi as revealed by the increasein perinuclear staining (Fig. 5C and D). Expressing the three con-structs, however, did not affect the total number of cellular peroxi-somes (Fig. 5E).

3.3. Subcellular distribution of Rab2 and Rab14

As mentioned before, both Rab2 and Rab14 belong to Rabfunctional group II that also includes Rab4, Rab11 and Rab25

Fig. 4. Localization of Rab6 and Rab18 in Huh7 cells after differential permeabilization with digitonin. The HA-tagged wt constructs were transiently co-transfected withEGFP-SKL that was used as a peroxisomal marker. Both HA-Rab6 (A) and HA-Rab18 (B) were localized to the cytoplasmic face of the peroxisomal membrane. Catalase residingin the peroxisomal matrix could not be detected (C). The images were assembled from z-projections, the scale bar represents 10 lm.

T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338 333

[35]. Compared with the other Rabs found in the peroxisomal frac-tion, Rab2 revealed the lowest spectral counts (Fig. 1). Actually,transfecting Huh7 cells with DsRed-Rab2Awt and comparing theflorescence signal with that of the peroxisomal marker EGFP-SKL,we could not get any indication for its peroxisomal localization(Fig. 6). The obtained pattern rather resembled an ER/Golgi topol-ogy [40].

Different to Rab2A, stable expression of Cherry-Rab14 in Huh7cells clearly resulted in localization of the protein to the peroxi-somal compartment (Fig. 7). Comparing the expression patters ofwt (Fig. 7A), dominant-active (Q70L, Fig. 7B) and -inactive (S25N,Fig. 7C) Rab14 showed peroxisomal localization predominantly inits dominant-active form. In this case, a nearly complete peroxi-somal topology was observed. In contrast, the dominant-inactiveRab14 (S25N) accumulated at intracellular structures that local-ized closely around the cell nucleus rather than to peroxisomes(Fig. 7C).

3.4. Peroxisomal localization of EGFP-Rab6

Rab6, a member of the Rab functional group VI, like Rab10 andRab14 carries a C-terminal XCXC prenylation sequence allowingdouble geranylgeranylation for faithful intracellular targeting. Sofar Rab6 was reported to localize to the Golgi and to cytoplasmicvesicles and to control fusion and transport of secretory vesicles

[41]. To study its intracellular distribution, HA-tagged wt, domi-nant-active (Q72L) and -inactive (N126I) Rab6 constructs weretransfected into Huh7 cells and cell lines stably expressing the con-structs established. All corresponding proteins localized to peroxi-somes (Fig. 8). As revealed by differential permeabilization withdigitonin, Rab6wt was bound to the cytosolic face of the mem-brane (Fig. 4A; see above). However, the GTP- and GDP-boundRab6 also recruited to vesicular structures other than peroxisomes(Fig. 8B–D). Most likely these structures include Golgi-derived ves-icles [41], as they, at least partially, co-localize with the Golgi mar-ker GM-130 (data not shown).

4. Discussion

Rab proteins are known to function as essential regulators ofvesicle trafficking and as such are indispensable in the organizationof intracellular compartmentalization. By accomplishing transportof components between cellular compartments, Rab proteins areinvolved in diverse processes, such as protein secretion, receptorrecycling, vesicle fission and fusion and lipid distribution. It istherefore not surprising that usually more than one Rab familymember associates with a distinct compartment. For example,recruitment to the ER and the Golgi has been reported for Rab1,Rab6, Rab18, Rab30, Rab34, Rab43 and Rab2, Rab6, Rab10, Rab11,Rab12, Rab14, Rab34, Rab36, respectively [42,43]. As Rab proteins

Fig. 5. IF co-localization in Huh7 cells stably expressing EGFP-Rab10wt (A), the Q68L (B) and T23A (C) mutants with catalase used as a peroxisomal marker. The guinea pigprimary anti-catalase antibody and was stained with an Alexa546-labeled secondary antibody. The number of co-localizing structures (D) and the average number of totalperoxisomes per cell (E) was determined by quantitative image analysis. Images were assembled from z-projections, the scale bar represents 10 lm.

334 T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338

frequently accompany their target membranes during transport,distinct Rab proteins might be found at more than one intracellularlocation. In regulating maturation of early endosomes, for example,Rab5 accompanies the endocytic membrane through variousstages and by recruiting Rab7 facilitates the conversion of earlyto late endosomes [44]. Hence Rab5 has been found associatedwith both early and late endosomes. A rather broad distributionis also reported for Rab6A and Rab6A’ that were shown to coordi-

nate retrograde endosome–Golgi–ER transport and thus localize toall three compartments [45,46].

As mentioned before, Rab6 regulates trafficking from early recy-cling endosomes up to the ER but in addition catalyzes fission ofGolgi-exiting transport carriers by a myosin II/F-actin-dependentmechanism [47]. Inhibition of either myosin II or Rab6 function im-paired fission of Rab6 carriers and anterograde and retrograde traf-ficking of cargo. Interestingly, myosin IIA and F-actin were recently

Fig. 6. Intracellular localization in Huh7 cells of transiently expressed DsRed-Rab2Awt. To investigate the Rab2 localization to peroxisomes, DsRed-Rab2A was co-transfectedwith the peroxisomal marker EGFP-SKL. No indication for a peroxisomal localization of Rab2 was obtained. Images were assembled from z-projections, the scale barrepresents 10 lm.

Fig. 7. IF co-localization of Huh7 cells stably expressing Cherry-Rab14wt (A), the Q70L (B) and S25N (C) mutants with catalase used as a peroxisomal marker. The primaryanti-catalase antibodies were stained with Alexa488-labeled secondary antibodies. While the dominant-active Q70L mutant assembled with peroxisomes (B), the dominant-inactive one did not but instead strongly associated to a peri-nuclear membrane network (C). The images were assembled from z-projections, the scale bar represents 10 lm.

T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338 335

localized to peroxisomes in an ATP- and GTP-dependent mannerin vivo and in vitro [21]. The GTP-dependence might well reflecttriggering of this process by a GTPase, such as Rab6 that in concertwith the acto–myosin complex might favor peroxisomal vesicula-tion. A retrograde vesicular peroxisome-ER transport is stronglysupported by recent studies following the import pathway of earlycomponents of the peroxisomal protein import machinery, such as

Pex3p and Pex19p connecting the peroxisomal compartment tothe endomembrane system [6–8]. It might also fit to previousobservations demonstrating peroxisomal recruitment of Arf1 andthe COPI coat as mentioned at the beginning [9–11].

In a previous work, we reported that a 50-GFP-Rab8A fusion pro-tein associated to both peroxisomes and the Golgi and in case ofthe dominant-inactive Rab8A (T22N) led to a remarkable increase

Fig. 8. IF co-localization in Huh7 cells stably expressing HA-Rab6wt (A), the Q72L (B) and N126I (C) mutants with catalase used as a peroxisomal marker. While catalase wasvisualized by guinea pig primary anti-catalase and secondary Alexa546-labeled antibodies, Rab6 was identified via the HA-tag and an Atto488-labeled secondary antibody.The number of co-localizing structures (D) and the average number of total peroxisomes per cell (E) were determined by quantitative image analysis. The images wereassembled from z-projections, the scale bar represents 10 lm.

336 T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338

in the number of peroxisomes [21]. Functionally Rab8 has beenlinked to the constitutive and regulated transport of melanosomes,the regulated secretion of ACTH by interacting with TRIP8b, a pro-tein homologous to Pex5p the peroxisomal targeting signal 1(PTS1) receptor and a recycling pathway localizing on organelleswith Arf6, MyoVb and MyoVc [48–51]. MyoV was also identified

as an effector for Rab10 and Rab11 and all three Rab proteins wereimplicated in regulating different pathways for recycling proteinsto the plasma membrane (PM) [51,52]. Whereas Rab11 was re-quired for transferrin recycling, both Rab8 and Rab11 were neces-sary for apical membrane sorting and de novo lumen formation[52]. Thus, active (Rab8, Rab11) and inactive (Rab10) Rab proteins

T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338 337

cooperate in the functioning of specific membrane trafficking path-ways. In Caenorhabditis elegans, Rab10 co-localized on recyclingendosomes with Arf6 and CNT-1, an Arf6 GAP, regulating transportby activating type I phosphatidylinositol-4-phosphate (PtdIns4P)5-kinase. Interestingly, we recently reported that various phos-phatidylinositides including PtdIns4,5P2 were synthesized inmammalian peroxisomal membranes [53]. Moreover, in S. cerevisi-ae Arf1 and Arf3, the latter representing the yeast homolog ofmammalian Arf6, regulate peroxisome abundance in a positiveand negative manner, respectively [9]. These observations suggestthat Rab10 might be involved in trafficking of both recycling endo-somes and vesicles derived from peroxisomal membranes.

There is convincing evidence that Rab18 localizes to LDs there-by triggering the amount of adipocyte differentiation-related pro-tein (ADRP) and inducing close apposition of the droplets to theER most likely facilitating lipid exchange [27,54]. The peroxisomallocalization of Rab18 might thus indicate a similar role for Rab18on peroxisomes that were recently shown to form extensive phys-ical contact with LDs promoting the coupling of lipolysis in LDswith peroxisomal fatty acid oxidation [13].

Rab14 as well as Rab2, Rab8 and Rab10 were recently found tobe targets of the Rab GAP Akt substrate of 160-kDa (AS160) and assuch participate in GLUT4 translocation to the PM [55]. Rab8 andRab14 dominated in muscle, whereas Rab10 was active in adiposetissue possibly by facilitating recruitment of a myosin isoform[56,57]. As silencing of AS160 increased the level of PM GLUT4and glucose uptake in adipocytes, the Rab proteins in their activeGTP-bound state might contribute to GLUT4 retention and thusprevention of its exocytosis [55]. Applied to the peroxisomal sys-tem, these data suggest that dependent of their state of activityRab10/Rab14 might be implicated in impeding or favoring peroxi-somal vesiculation.

The present IF analysis demonstrating Rab6, Rab10, Rab14 andRab18 localization to peroxisomes might represent a first step to-ward a more comprehensive understanding of peroxisomal traf-ficking. However, future studies particularly on thecharacterization of the Rab-GAPs and Rab-GEFs and the peroxi-somal effectors involved are needed in order to allow for a moredetailed insight into the role these Rab proteins play in peroxisomebiogenesis and the communication of peroxisomes with other cel-lular compartments.

Acknowledgements

We are grateful to Nils Johnsson (University of Ulm) for scien-tific assistance and to Guangyu Wu (Louisiana State University),Kai Simons (MPI Dresden), Sally Martin (University of Queensland)and Mary Mc Caffrey (University College Cork) for kindly providingvarious Rab constructs. The work was supported by the DeutscheForschungsgemeinschaft, the Excellence Initiative of the GermanFederal & State Governments (EXC 294 BIOSS) and the FP6 Euro-pean Union Project ‘‘Peroxisome’’ (Grant LSHG-CT-2004-512018).

References

[1] Wanders, R.J. and Waterham, H.R. (2006) Biochemistry of mammalianperoxisomes revisited. Annu. Rev. Biochem. 75, 295–332.

[2] Gorgas, K., Teigler, A., Komljenovic, D. and Just, W.W. (2006) The ether lipid-deficient mouse: tracking down plasmalogen functions. Biochim. Biophys.Acta 1763, 1511–1526.

[3] Hajra, A.K. (1995) Glycerolipid biosynthesis in peroxisomes (microbodies).Prog. Lipid Res. 34, 343–364.

[4] Kovacs, W.J. and Krisans, S. (2003) Cholesterol biosynthesis and regulation:role of peroxisomes. Adv. Exp. Med. Biol. 544, 315–327.

[5] Nagan, N. and Zoeller, R.A. (2001) Plasmalogens: biosynthesis and functions.Prog. Lipid Res. 40, 199–229.

[6] Hoepfner, D., Schildknegt, D., Braakman, I., Philippsen, P. and Tabak, H.F.(2005) Contribution of the endoplasmic reticulum to peroxisome formation.Cell 122, 85–95.

[7] Lam, S.K., Yoda, N. and Schekman, R. (2011) A vesicle carrier that mediatesperoxisome protein traffic from the endoplasmic reticulum. Proc. Natl. Acad.Sci. USA 107, 21523–21528.

[8] van der Zand, A., Gent, J., Braakman, I. and Tabak, H.F. (2012) Biochemicallydistinct vesicles from the endoplasmic reticulum fuse to form peroxisomes.Cell 149, 397–409.

[9] Lay, D., Grosshans, B.L., Heid, H., Gorgas, K. and Just, W.W. (2005) Binding andfunctions of ADP-ribosylation factor on mammalian and yeast peroxisomes. J.Biol. Chem. 280, 34489–34499.

[10] Lay, D., Gorgas, K. and Just, W.W. (2006) Peroxisome biogenesis: where Arfand coatomer might be involved. Biochim. Biophys. Acta 1763, 1678–1687.

[11] Passreiter, M., Anton, M., Lay, D., Frank, R., Harter, C., Wieland, F.T., Gorgas, K.and Just, W.W. (1998) Peroxisome biogenesis: involvement of ARF andcoatomer. J. Cell Biol. 141, 373–383.

[12] Beller, M., Sztalryd, C., Southall, N., Bell, M., Jackle, H., Auld, D.S. and Oliver, B.(2008) COPI complex is a regulator of lipid homeostasis. PLoS Biol. 6, e292.

[13] Binns, D. et al. (2006) An intimate collaboration between peroxisomes andlipid bodies. J. Cell Biol. 173, 719–731.

[14] Blanchette-Mackie, E.J. et al. (1995) Perilipin is located on the surface layer ofintracellular lipid droplets in adipocytes. J. Lipid Res. 36, 1211–1226.

[15] Soni, K.G., Mardones, G.A., Sougrat, R., Smirnova, E., Jackson, C.L. andBonifacino, J.S. (2009) Coatomer-dependent protein delivery to lipiddroplets. J. Cell Sci. 122, 1834–1841.

[16] Huber, C., Saffrich, R., Anton, M., Passreiter, M., Ansorge, W., Gorgas, K. andJust, W. (1997) A heterotrimeric G protein-phospholipase A2 signaling cascadeis involved in the regulation of peroxisomal motility in CHO cells. J. Cell Sci.110 (Pt 23), 2955–2968.

[17] Huber, C.M., Saffrich, R., Ansorge, W. and Just, W.W. (1999) Receptor-mediatedregulation of peroxisomal motility in CHO and endothelial cells. EMBO J. 18,5476–5485.

[18] Rapp, S., Saffrich, R., Anton, M., Jakle, U., Ansorge, W., Gorgas, K. and Just, W.W.(1996) Microtubule-based peroxisome movement. J. Cell Sci. 109, 837–849.

[19] Wiemer, E.A., Wenzel, T., Deerinck, T.J., Ellisman, M.H. and Subramani, S.(1997) Visualization of the peroxisomal compartment in living mammaliancells: dynamic behavior and association with microtubules. J. Cell Biol. 136,71–80.

[20] Nguyen, T., Bjorkman, J., Paton, B.C. and Crane, D.I. (2006) Failure ofmicrotubule-mediated peroxisome division and trafficking in disorders withreduced peroxisome abundance. J. Cell Sci. 119, 636–645.

[21] Schollenberger, L. et al. (2010) RhoA regulates peroxisome association tomicrotubules and the actin cytoskeleton. PLoS One 5, e13886.

[22] Hartl, F.U., Just, W.W., Koster, A. and Schimassek, H. (1985) Improved isolationand purification of rat liver peroxisomes by combined rate zonal andequilibrium density centrifugation. Arch. Biochem. Biophys. 237, 124–134.

[23] Liu, H., Sadygov, R.G. and Yates 3rd., J.R. (2004) A model for random samplingand estimation of relative protein abundance in shotgun proteomics. Anal.Chem. 76, 4193–4201.

[24] Bergbrede, T. et al. (2009) Biophysical analysis of the interaction of Rab6aGTPase with its effector domains. J. Biol. Chem. 284, 2628–2635.

[25] Dong, C. and Wu, G. (2007) Regulation of anterograde transport of adrenergicand angiotensin II receptors by Rab2 and Rab6 GTPases. Cell Signal. 19, 2388–2399.

[26] Schuck, S., Gerl, M.J., Ang, A., Manninen, A., Keller, P., Mellman, I. and Simons,K. (2007) Rab10 is involved in basolateral transport in polarized Madin-Darbycanine kidney cells. Traffic 8, 47–60.

[27] Martin, S., Driessen, K., Nixon, S.J., Zerial, M. and Parton, R.G. (2005) Regulatedlocalization of Rab18 to lipid droplets: effects of lipolytic stimulation andinhibition of lipid droplet catabolism. J. Biol. Chem. 280, 42325–42335.

[28] Kelly, E.E., Horgan, C.P., Adams, C., Patzer, T.M., Ni Shuilleabhain, D.M.,Norman, J.C. and McCaffrey, M.W. (2010) Class I Rab11-family interactingproteins are binding targets for the Rab14 GTPase. Biol. Cell 102, 51–62.

[29] Wolvetang, E.J., Tager, J.M. and Wanders, R.J. (1990) Latency of theperoxisomal enzyme acyl-CoA:dihydroxyacetonephosphate acyltransferasein digitonin-permeabilized fibroblasts: the effect of ATP and ATPaseinhibitors. Biochem. Biophys. Res. Commun. 170, 1135–1143.

[30] Woodcroft, B.J., Hammond, L., Stow, J.L. and Hamilton, N.A. (2009) Automatedorganelle-based colocalization in whole-cell imaging. Cytometry A 75, 941–950.

[31] Manders, E.M., Hoebe, R., Strackee, J., Vossepoel, A.M. and Aten, J.A. (1996)Largest contour segmentation: a tool for the localization of spots in confocalimages. Cytometry 23, 15–21.

[32] Huber, L.A. et al. (1994) Mapping of Ras-related GTP-binding proteins by GTPoverlay following two-dimensional gel electrophoresis. Proc. Natl. Acad. Sci.USA 91, 7874–7878.

[33] Wiese, S. et al. (2007) Proteomics characterization of mouse kidneyperoxisomes by tandem mass spectrometry and protein correlationprofiling. Mol. Cell. Proteomics 6, 2045–2057.

[34] Foster, L.J., de Hoog, C.L., Zhang, Y., Xie, X., Mootha, V.K. and Mann, M. (2006) Amammalian organelle map by protein correlation profiling. Cell 125, 187–199.

[35] Pereira-Leal, J.B. and Seabra, M.C. (2001) Evolution of the Rab family of smallGTP-binding proteins. J. Mol. Biol. 313, 889–901.

[36] Lane, K.T. and Beese, L.S. (2006) Thematic review series: lipid posttranslationalmodifications. Structural biology of protein farnesyltransferase andgeranylgeranyltransferase type I. J. Lipid Res. 47, 681–699.

[37] Joberty, G., Tavitian, A. and Zahraoui, A. (1993) Isoprenylation of Rab proteinspossessing a C-terminal CaaX motif. FEBS Lett. 330, 323–328.

338 T. Gronemeyer et al. / FEBS Letters 587 (2013) 328–338

[38] Lametschwandtner, G., Brocard, C., Fransen, M., Van Veldhoven, P., Berger, J.and Hartig, A. (1998) The difference in recognition of terminal tripeptides asperoxisomal targeting signal 1 between yeast and human is due to differentaffinities of their receptor Pex5p to the cognate signal and to residues adjacentto it. J. Biol. Chem. 273, 33635–33643.

[39] Fransen, M., Amery, L., Hartig, A., Brees, C., Rabijns, A., Mannaerts, G.P. and VanVeldhoven, P.P. (2008) Comparison of the PTS1- and Rab8b-binding propertiesof Pex5p and Pex5Rp/TRIP8b. Biochim. Biophys. Acta 1783, 864–873.

[40] Tisdale, E.J., Bourne, J.R., Khosravi-Far, R., Der, C.J. and Balch, W.E. (1992) GTP-binding mutants of rab1 and rab2 are potent inhibitors of vesicular transportfrom the endoplasmic reticulum to the Golgi complex. J. Cell Biol. 119, 749–761.

[41] Grigoriev, I. et al. (2011) Rab6, Rab8, and MICAL3 cooperate in controllingdocking and fusion of exocytotic carriers. Curr. Biol. 21, 967–974.

[42] Liu, S. and Storrie, B. (2012) Are Rab proteins the link between Golgiorganization and membrane trafficking? Cell. Mol. Life Sci. 69, 4093–4106.

[43] Zerial, M. and McBride, H. (2001) Rab proteins as membrane organizers. Nat.Rev. Mol. Cell Biol. 2, 107–117.

[44] Huotari, J. and Helenius, A. (2011) Endosome maturation. EMBO J. 30, 3481–3500.

[45] Kano, F., Yamauchi, S., Yoshida, Y., Watanabe-Takahashi, M., Nishikawa, K.,Nakamura, N. and Murata, M. (2009) Yip1A regulates the COPI-independentretrograde transport from the Golgi complex to the ER. J. Cell Sci. 122, 2218–2227.

[46] Martinez, O., Antony, C., Pehau-Arnaudet, G., Berger, E.G., Salamero, J. andGoud, B. (1997) GTP-bound forms of rab6 induce the redistribution of Golgiproteins into the endoplasmic reticulum. Proc. Natl. Acad. Sci. USA 94, 1828–1833.

[47] Miserey-Lenkei, S., Chalancon, G., Bardin, S., Formstecher, E., Goud, B. andEchard, A. (2010) Rab and actomyosin-dependent fission of transport vesiclesat the Golgi complex. Nat. Cell Biol. 12, 645–654.

[48] Chakraborty, A.K., Funasaka, Y., Araki, K., Horikawa, T. and Ichihashi, M. (2003)Evidence that the small GTPase Rab8 is involved in melanosome traffic anddendrite extension in B16 melanoma cells. Cell Tissue Res. 314, 381–388.

[49] Chen, S., Liang, M.C., Chia, J.N., Ngsee, J.K. and Ting, A.E. (2001) Rab8b and itsinteracting partner TRIP8b are involved in regulated secretion in AtT20 cells. J.Biol. Chem. 276, 13209–13216.

[50] Jacobs, D.T., Weigert, R., Grode, K.D., Donaldson, J.G. and Cheney, R.E. (2009)Myosin Vc is a molecular motor that functions in secretory granule trafficking.Mol. Biol. Cell 20, 4471–4488.

[51] Peranen, J. (2011) Rab8 GTPase as a regulator of cell shape. Cytoskeleton 68,527–539.

[52] Roland, J.T., Bryant, D.M., Datta, A., Itzen, A., Mostov, K.E. and Goldenring, J.R.(2011) Rab GTPase–Myo5B complexes control membrane recycling andepithelial polarization. Proc. Natl. Acad. Sci. USA 108, 2789–2794.

[53] Jeynov, B., Lay, D., Schmidt, F., Tahirovic, S. and Just, W.W. (2006)Phosphoinositide synthesis and degradation in isolated rat liverperoxisomes. FEBS Lett. 580, 5917–5924.

[54] Ozeki, S., Cheng, J., Tauchi-Sato, K., Hatano, N., Taniguchi, H. and Fujimoto, T.(2005) Rab18 localizes to lipid droplets and induces their close apposition tothe endoplasmic reticulum-derived membrane. J. Cell Sci. 118, 2601–2611.

[55] Dugani, C.B. and Klip, A. (2005) Glucose transporter 4: cycling, compartmentsand controversies. EMBO Rep. 6, 1137–1142.

[56] Miinea, C.P., Sano, H., Kane, S., Sano, E., Fukuda, M., Peranen, J., Lane, W.S. andLienhard, G.E. (2005) AS160, the Akt substrate regulating GLUT4 translocation,has a functional Rab GTPase-activating protein domain. Biochem. J. 391, 87–93.

[57] Roland, J.T., Lapierre, L.A. and Goldenring, J.R. (2009) Alternative splicing inclass V myosins determines association with Rab10. J. Biol. Chem. 284, 1213–1223.