Rab11b resides in a vesicular compartment distinct from Rab11a in parietal cells and other epithelial cells

Available online at www.sciencedirect.com R

Experimental Cell Research 290 (2003) 322–331

www.elsevier.com/locate/yexcr

Rab11b resides in a vesicular compartment distinct from Rab11a in parietal cells and other epithelial cells Lynne A. Lapierre,* Matthew C. Dorn, C. Faith Zimmerman, Jennifer Navarre, Jason O. Burnette, and James R. Goldenring Departments of Medicine, Surgery, and Cellular Biology and Anatomy, Institute of Molecular Medicine and Genetics, Medical College of Georgia, and the Augusta VA Medical Center, Augusta, GA 30912, USA Received 30 January 2003, revised version received 24 April 2003

Abstract The Rab11 family of small GTPases is composed of three members, Rab11a, Rab11b, and Rab25. While recent work on Rab11a and Rab25 has yielded some insights into their function, Rab11b has received little attention. Therefore, we sought to examine the distribution of endogenous Rab11b in epithelial cells. In rabbit gastric parietal cells, unlike Rab11a, Rab11b did not colocalize or coisolate with H⫹/K⫹-ATPase. In MDCK cells, endogenous Rab11b localized to an apical pericentrisomal region distinct from Rab11a. The microtubule agents nocodazole and taxol dramatically alter Rab11a’s localization in the cell, while effects on Rab11b’s distribution were less apparent. These results indicate that in contrast to Rab11a, the Rab11b compartment in the apical region is not as dependent upon microtubules. While Rab11a is known to regulate transferrin trafficking in nonpolarized cells and IgA trafficking in polarized cells, Rab11b exhibited little colocalization with either of these cargoes. Thus, while Rab11a and Rab11b share high sequence homology, they appear to reside within distinct vesicle compartments. © 2003 Elsevier Inc. All rights reserved. Keywords: Rab11b; Rab11a; Parietal cell; Apical recycling

Introduction The Rab proteins are a subfamily of the Ras superfamily of small GTPases and are known to regulate specific trafficking pathways within cells [1]. The trafficking of vesicles within eukaryotic cells is an extremely complex endeavor as illustrated by the identification of over 50 members of the Rab family. A large number of these Rabs are present along the endocytic pathway, suggesting a high degree of complexity within this pathway. In polarized epithelial cells, the process of apical recycling can be separated into three stages. First, following initial internalization into the early endosome the membrane with accompanying proteins enters into the sorting endosome, where fluid-phase cargo and ligands are dissociated

* Corresponding author. Department of Surgery/Section of Surgical Sciences, 4150A MRBIII, Vanderbilt University Medical Center, 465 21st Ave. S., Nashville, TN 37232-8720, USA. Fax: ⫹1-615-343-1591. E-mail address: [email protected] (L.A. Lapierre). 0014-4827/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0014-4827(03)00340-9

from their receptors and routed into the lysosome pathway. Other membrane proteins (such as transferrin receptor and polymeric IgA receptor) are not trafficked into the lysosome pathway for degradation, but are sent instead to the second stage of recycling, the recycling endosome. Proteins designated for recycling back to the apical membrane are then transported through the third stage, the pericentriolar apical recycling endosome, a tubulovesicular compartment where the membrane proteins are repackaged for transport back to the apical plasma membrane [2]. In many polarized epithelial cells there is an added layer of complexity in that recycling can occur both from the apical and the basolateral membranes as well as through a transcytotic pathway between the two membranes. Antibodies against Rab11a label the recycling endosome in nonpolarized cells [3], a subapical population of vesicles in epithelial tissue [4], and the pericentriolar apical recycling compartment in polarized cells [5]. These results indicate that Rab11a is involved in membrane recycling in both polarized and nonpolarized cells.

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

There are currently three identified members of the mammalian Rab11 family, Rab11a, Rab11b, and Rab25 [6], that mainly differ within a 20-amino acid stretch near the carboxyl terminus. While Rab11a [4] and Rab11b [7] are ubiquitously expressed, Rab25 is expressed only in certain epithelial cells [8], indicating that individual members of the family may be involved in different pathways. Rab11a was the first of the family cloned and so most of the work done on this family has focused on Rab11a. Studies in nonpolarized cells have established that Rab11a is a marker of the plasma membrane recycling system involved in trafficking of the transferrin receptor [3,9,10]. A functional Rab11a is necessary for entrance of transferrin into the recycling compartment and for recycling of the vesicles back to the plasma membrane [10]. A Rab11 has also been associated with the trans-Golgi network and vesicle movement out of the Golgi [11,12]. In polarized MDCK cells both Rab11a and Rab25 are associated with the apical recycling endosome and the trafficking of polymeric IgA, but not transferrin [5, 13, 14]. One of the best models of apical recycling is the gastric parietal cell. This cell is highly adapted for the extensive and efficient regulated recycling of the proton pump to the apical surface. In these cells, Rab11a cosegregated with the H⫹/K⫹-ATPase in resting cells and redistributed to the secretory canaliculus with the proton pump upon stimulation with histamine or forskolin [15,16]. While a number of investigations have focused on the function of Rab11a within the apical recycling system, little is known about Rab11b. Recently we have produced a polyclonal antiserum that is specific for Rab11b. With this specific antiserum we have sought to elucidate the distribution of endogenous Rab11b and its potential overlap with Rab11a. Unlike Rab11a, Rab11b did not colocalize or coisolate with the proton pump in parietal cells. While Rab11b does localize to an apical pericentriolar region near Rab11a in MDCK cells, the overlap is incomplete. Rab11b demonstrates a dependence on the microtubule cytoskeleton that is different than that of Rab11a, and the majority of Rab11b does not colocalize with either transferrin receptor or the polymeric IgA receptor. Thus, while Rab11a and Rab11b share high sequence homology and localize to adjacent pericentriolar apical regions, they appear to mark distinct vesicle populations.

Materials and methods Production of anti-Rab11b antisera and antibodies used A rabbit polyclonal antisera was produced against a peptide spanning amino acids 181–213 at the carboxyl terminus of the murine Rab11b sequence. The Rab11b and Rab11a sequences differ by 16 of 34 amino acids in this stretch. The peptide was attached to keyhole limpet hemacyanin for the immunization protocol. The peptide conjugation and antisera production were performed by the Berkeley An-

323

tibody Co. (BabCO). The Rab11b antiserum was purified by affinity chromatography using the same peptide as for the immunization (Pierce). The antibody was eluted with 100 mM triethylamine, pH 11.7, 150 mM NaCl and then neutralized with 3 M Tris, pH 6.5, to reach a final pH of 8.0. Both a mouse monoclonal (8H10; [4]) and a rabbit polyclonal antisera (Zymed) specific for Rab11a were used. The sheep anti-rabbit IgA receptor was a kind gift from Dr. Curtis Okamato (Univ. Southern CA, Los Angeles, CA). A rat monoclonal anti-ZO-1 and a mouse monoclonal antitransferrin receptor (CD71) were from Chemicon. The mouse anti-H⫹/K⫹-ATPase (12.18) was a kind gift from Dr. Adam Smolka (Medical University of South Carolina, Charleston, SC). The secondary antibodies were speciesspecific Alexa-488-goat anti-rabbit IgG (Molecular Probes), Cy3-donkey anti-rat IgG or anti-mouse IgG, and Cy5-donkey anti-mouse IgG or anti-sheep IgG (Jackson Immunological). Actin was visualized using Alexa-568 –Phalloidin (Molecular Probes). Antibody blocking Antibodies were preincubated with or without 5 ␮g/ml of either the Rab11b or the Rab11a peptide in primary dilution buffer for 1 h prior to incubation with Western blots and immunofluorescence. The Rab11b peptide was the same one used for the production and affinity purification of the anti-Rab11b antisera, while the Rab11a peptide was the paralogous human sequence (amino acids 181–213). Western blotting Protein samples were resolved on 12% SDS–polyacrylamide gels following a standard Laemmli protocol [17]. All incubations were performed at room temperature. Proteins were transferred onto Immobilon-P membranes (Millipore) and blocked for 1 h with 5% DMP/TBS-T (5% dry milk powder, Tris-buffered saline, 0.05% Tween-20). The blots were incubated with primary antibody diluted in 2.5% DMP/TBS-T either for 1 h (anti-Rab11b) or overnight (anti-Rab11a), washed four times for 10 min in 2.5% DMP/ TBS-T, and incubated for 1 h with horseradish perioxidaselabeled donkey anti-rabbit secondary (Jackson Immunological) diluted in 2.5% DMP/TBS-T. The blots were then washed three times with TBS-T followed by one time with TBS and specific labeling was detected by enhanced chemiluminescence (Supersignal, Pierce) with autoradiography using Kodak BioMax ML film. Rabbit gastric gland isolation and treatments Gastric glands were isolated from the fundic mucosa of New Zealand White rabbits as previously described [18]. The glands were then treated with either 100 ␮M ranitidine (resting; Sigma) or 100 ␮M histamine (stimulated; Sigma) for 30 min at 37°C with rocking. After stimulation, cells

324

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

were rapidly washed in PBS and then fixed in 4% paraformaldehyde for 15 min at room temperature with rocking. GFP-Rab11 construction, transfection, and production of stable lines The full-length cDNAs of murine Rab11b or rabbit Rab11a were cloned into the pEGFP-C2 vector (Clontech). The resulting chimeras contained the GFP on the amino terminus of the Rab proteins. The GFP-Rab chimeras were transfected into Madin–Darby canine kidney (MDCK) cells with Effectene (Qiagen) following the manufacturer’s protocol. The DNA (1 ␮g/60-mm plate) complex was incubated with the cells overnight. The cells were then washed, refed, and allowed to recover 48 h before selection. Cells were then trypsinized and replated to limiting dilution under selection with 0.5 mg/ml G-418 (Cellgro). Individual colonies were identified and clonally expanded based on their GFP fluorescence, as previously described for GFP-Rab11a [19]. Cell culture and treatments The MDCK cells and the stable GFP-Rab11a and GFPRab11b cell lines were grown in D-MEM supplemented with 10% heat inactivated fetal bovine sera, L-glutamine (2 mM), and penicillin/streptomycin solution. Both GFPRab11a and GFP-Rab11a cell lines were also supplemented with G-418 (0.5 mg/ml). Cells were split 1:20 onto 12-mm, 0.4-␮m pore Transwell clear filters (Costar, clear polyester) and allowed to polarize for 4 days before treatments and staining. The HeLa cells were grown in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine sera, Lglutamine, glucose (2 g/liter), and penicillin/streptomycin solution. Cells were split 1:25 onto 25-mm glass coverslips and were used 1–2 days postplating. All cell culture reagents were from Cellgro with the exception of the FBS (GIBCO/BRL). For the nocodazole treatment, cells were preincubated at 4°C for 30 min, and nocodazole (Calbiochem) was then added to a final concentration of 33 ␮M (from a 33 mM stock in DMSO) and incubated for a further 30 min at 4°C. The cells were then shifted to 37°C, 5% CO2 and incubated a further 60 min. For the Taxol treatment, cells were treated with Taxol (Alexis Corp.) at a final concentration of 5 ␮M (from a 5 mM stock in DMSO) for 4 h at 37°C, 5% CO2. Fixation and immunocytochemistry Cells were washed with PBS and then fixed with 4% paraformaldehyde for 30 min at 4°C. The cells were permealized and blocked with 5% normal donkey sera (Jackson Immunological), 0.3% Triton X-100 in PBS for 20 min and then incubated with the primary antibodies diluted in block for 2 h at room temperature or overnight at 4°C. The cells were washed with PBS and then incubated with the appropriate secondary antibodies for 1 h at room temperature. Cells were

mounted in Prolong Antifade solution (Molecular Probes). The HeLa cells were imaged using a Zeiss Axiophot microscope outfitted with a Spot camera. The MDCK cells were imaged with a scanning confocal fluorescence microscope (Molecular Dynamics, Sunnyvale, CA). Section series (thirty 0.29 ␮m optical sections) were performed twice, first using dual imaging with the 488- and 647-nm excitation laser lines, 530 DF30 and 660 EFLP emission filters, and a 650 beam splitter to visualize the Alexa-488 or GFP with the Cy5 fluorochromes. The cells were then reimaged with a 568-nm excitation laser line and either a 570 EFLP or a 600 DF40 emission filter and a 565 beam splitter to visualize the Cy3 or Alexa-568 fluorochromes, respectively. The gastric glands were imaged as single 0.29-␮m optical sections using the scanning confocal microscope. Transferrin loading MDCK cells grown on transwells were incubated for 1 h in serum-free medium contain 50 ␮g/ml of Cy3-labeled human transferrin (a kind gift from Dr. James E. Casanova, Univ. of Virginia, Charlottesville, VA) and then fixed and stained as above. Tubulovesicle preparation Gastric tublovesicles were prepared from resting rabbit gastric mucosa, as previously described [16,20]. Male New Zealand White rabbits were anesthetized by intravenous administration of ketamine, and their stomachs were perfused under high pressure with oxygenated PBS and removed. The gastric mucosa was scraped off the serosa with a metal spatuala, minced with scissors, and homogenized in 5 Vol of homogenization buffer [113 mM mannitol, 37 mM sucrose, 0.4 mM EDTA, 5 mM MES, pH 6.7, 5 mM benzamidine, 0.1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF)] plus protease inhibitor cocktail (Sigma). Immunoadsorption of H⫹/K⫹-ATPase tubulovesicles As previously described [16], for a single immunoadsorption experiment, 750 ␮g of Dynabeads was washed three times in PBS–1% BSA, blocked for 30 min at 4°C in PBS–1% BSA, and washed twice in PBS– 0.1% BSA. The beads were incubated with either the H⫹/K⫹-ATPase antibody (12.18) or an isotype-matched mouse IgG overnight at 4°C. The loaded beads were washed four times for 30 min at room temperature with PBS– 0.1% BSA and incubated with the p20 fraction of the tubulovesicle preparation (50 ␮g of protein) for 2 h at room temperature in PBS– 0.1% BSA plus protease inhibitor cocktail (Sigma). After 2 h the unbound material was removed and centrifuged at 100,000g for 1 h at 4°C, and the pellet resuspended in SDS sample buffer. The beads were washed four times with PBS– 0.1% BSA and the bound material was eluted from the beads in SDS sample buffer. All samples were heated at 65°C for 15

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

Fig. 1. The anti-Rab11b antiserum did not cross-react with Rab11a. A rabbit polyclonal anti-serum was produced against a peptide spanning amino acids 181–213 at the C terminus of the human sequence. A Western blot of rabbit parietal cell lysate was probed with either the polyclonal anti-Rab11b (b) or a monoclonal anti-Rab11a (a) preincubated with or without the indicated peptides. These results are representative of two separate experiments.

min. The proteins were separated on a 10% SDS–PAGE gel, electrophoretically transferred to Immobilon-P (Millipore), and analyzed by Western blotting. The blot was cut in half at the 60-kDa range. The top part of the blot was probed for H⫹/K⫹-ATPase. The bottom part of the blot was probed first with anti-Rab11a (8H10) and then stripped and reprobed with polyclonal rabbit anti-Rab11b. Primary antibodies were detected with Fc fragment-specific secondary antibodies conjugated to horseradish peroxidase (Jackson ImmunoResearch Labs). Bands were visualized using ECL substrate (SuperSignal, Pierce) and Kodak BioMax film.

Results and discussion The anti-Rab11b antiserum did not cross-react with Rab11a A rabbit polyclonal antiserum was produced against a peptide spanning amino acids 181–213 at the carboxyl terminus of the murine Rab11b sequence, where the Rab11b and Rab11a sequences differ by 16 of 34 amino acids. Strips of blots containing parietal cell lysate were probed either with the monoclonal anti-Rab11a (8H10; [4]) or with the affinity-purified polyclonal anti-Rab11b and competed with peptide (Fig. 1). The immunoreactivities were only blocked by their corresponding peptide, indicating the specificity of the antibody preparations. Immunofluorescence staining of endogenous Rab11a or Rab11b, in MDCK or HeLa cells, was also blocked only when the corresponding peptide was preincubated with the matching antibody (results not shown). The Rab11a and Rab11b antibodies also exhibited specificity against the GFP-Rab chimeras. As shown in Fig. 5 below, the anti-Rab11b antibody reacted with the GFP-Rab11b chimera on Western blot and not with the GFP-Rab11a chimera on Western blot and by immunofluorescence. Similar antigen specificity was observed with the Rab11a antibody. Rab11b does not colocalize or coisolate with H⫹/K⫹ATPase The gastric parietal cell represents the most prominent example of an amplified apical recycling system. This cell is

325

highly adapted for the extensive and efficient movement of the proton pump (H⫹/K⫹-ATPase) to the apical plasma membrane upon stimulation with either histamine or forskolin. Upon stimulation, a massive fusion event occurs between the tubulovesicles containing H⫹/K⫹-ATPase and the apical canaliculus eliciting a fivefold increase in the apical plasma membrane surface area [21, 22]. This massive event is reversible upon cessation of the simulation, and the H⫹/K⫹-ATPase is rapidly reincorporated into tubulovesicles that are competent for another round of fusion [23–25]. H⫹/K⫹-ATPase-containing tubulovesicles therefore represent perhaps the most prominent source of apical recycling membranes. Rab11a’s role in this event has been well documented, first by its colocalization with H⫹/K⫹ATPase and, second, by its movement with H⫹/K⫹-ATPase to the apical membrane upon stimulation [16, 26]. Overexpression of the GTP-binding deficient Rab11aN124I mutant inhibits parietal cell secretion [27]. In contrast with Rab11a, Rab11b did not colocalize with H⫹/K⫹-ATPase in resting parietal cells (Fig. 2A (R)). Upon stimulation of the parietal cells, there is a fusion of tubulovesicles containing H⫹/K⫹ATPase with the secretory canaliculus membrane and a remodeling of the underlying actin [21, 25, 28]. As shown in Fig. 2A (S), in contrast to Rab11a, Rab11b does not move to the apical membrane with H⫹/K⫹-ATPase [16, 26], but rather localizes in areas adjacent to the apical membrane. To explore biochemically the question of Rab11b localization in parietal cells, we fractionated gastric mucosa and probed the fractions for Rab11a, Rab11b, and H⫹/K⫹-ATPase. Fig. 2B illustrates that while Rab11b has a pattern similar to that of Rab11a and H⫹/K⫹ATPase there were some differences. The most striking difference was the presence of detectable Rab11b in the 100,000g supernatant, a fraction in which we routinely observe very little Rab11a or H⫹/K⫹-ATPase. Since there was significant overlap of Rab11b and H⫹/K⫹ATPase in the tubulovesicle fractionation, we studied the coisolation of H⫹/K⫹-ATPase with both Rab11 species. H⫹/K⫹-ATPase-containing tubulovesicles were immunoisolated from the P20 fraction using a monoclonal antibody against H⫹/K⫹-ATPase (12.18) and anti-mouse immunoglobulin magnetic beads [16]. H⫹/K⫹-ATPase and both Rab11 species remained in the supernatant in samples incubated with beads containing an isotypematched nonspecific, mouse IgG1 (Fig. 2C (CB and CU)). As previously reported [16] the proton pump was isolated with the anti-H⫹/K⫹-ATPase loaded beads, along with a significant portion of the Rab11a. In contrast, none of the Rab11b was isolated with the H⫹/K⫹ATPase-containing vesicles (Fig. 2C (HB and HU)). This mucosal preparation contains other gastric gland cells beside parietal cells, but only the parietal cells have H⫹/K⫹-ATPase, so the Rab11a that did not coisolate is likely associated with other vesicle populations from nonparietal cells.

326

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

Fig. 2. Rab11b does not colocalize or coisolate with H⫹/K⫹-ATPase. (A) Isolated rabbit gastric glands (R, resting or S, stimulated) were triple-labeled for Rab11b, H⫹/K⫹-ATPase, and actin. While Rab11a colocalized and moved upon stimulation with H⫹/K⫹-ATPase and actin [16, 26], Rab11b did not. Bar ⫽ 10 ␮M. These results are representative of five separate experiments. (B) Rab11b and Rab11a exhibited slightly different membrane subfractionation patterns within a tubulovescular preparation. Membrane fractions of rabbit gastric mucosa were resolved on gels (H⫹/K⫹-ATPase, 10%, 5 ␮g or Rabs, 12%, 25 ␮g). While the majority of Rab11a, Rab11b, and H⫹/K⫹-ATPase was in the 100,000g pellet and the P20 and P27 sucrose fractions, more Rab11b was observed in the 100,000g supernatant than either Rab11a or H⫹/K⫹-ATPase. H, homogenate; SO, 50g supernatant; P0, 50g pellet; P1, 100g pellet; P2, 10,000g pellet; P3, 100,000g pellet; S3, 100,000g supernatant; P20, 20% sucrose gradient interface; P27, 27% sucrose gradient interface; P33, 33% sucrose gradient interface; CM, sucrose gradient pellet. These results are representative of four separate fractionations. (C) Rab11b did not coisolate with H⫹/K⫹-ATPase. Vesicles containing H⫹/K⫹-ATPase were immunoisolated using the p20 fraction of a rabbit gastric mucosa tubulovesicle fractionation. Both bound and unbound samples were resolved on a 10% SDS–PAGE gel. A significant amount of Rab11a was found with H⫹/K⫹-ATPase in the bound fraction, while all of the Rab11b was detected in the unbound fraction. CB, control bound; CU, control unbound; HB, H⫹/K⫹-ATPase bound; HU, H⫹/K⫹-ATPase unbound. Results are representative of four separate isolations.

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

327

ners” of cells, Rab11b exhibited little movement away from the central region (Fig. 3 (T, arrowhead)). Previous studies have suggested an association of a Rab11 family member with the Golgi apparatus [3,31]. However, neither Rab11a nor Rab11b exhibited any close colocalization with the three Golgi markers AP-1, p58, or TGN38 (data not shown). In addition, Brefeldin A treatment did not affect Rab11b (data not shown), again indicating that Rab11b is not involved in Golgi trafficking. These results indicated that while both Rab11a and Rab11b localized to the same apical region in the cell, they were actually in distinct vesicular compartments, as indicated by the differential effects of microtubule stability on the distribution of the two Rabs. Rab11b did not colocalize with two known Rab11aassociated cargoes, the TfR in nonpolarized cells and the pIgAR in polarized cells Fig. 3. The effects of microtubule stability upon the localization of endogenous Rab11b. MDCK cell lines were plated on Transwell filters and used 4 days postconfluence. Cells were either nontreated (C), nocodazoletreated (N), or taxol-treated (T), fixed, and stained with rabbit anti-Rab11b, mouse anti-Rab11a (8H10) and rat anti-ZO-1. The white arrowheads in the Taxol (T) panels indicate endogenous Rab11a in the corners of cells in the Rab11a panel and the equivalent area in the Rab11b and ZO-1 panels. The black arrowheads on the right indicate the position of the X/Z section. Bar ⫽ 4 ␮M. These results are representative of five separate experiments.

Rab11b localizes to an apical, pericentrosomal compartment that is distinct from the Rab11a compartment Rab11a’s involvement with both the plasma membrane recycling system in nonpolarized cells and the apical recycling system in polarized cells is well established [3,5]. In polarized cells, the structure of the apical recycling system is dependent upon the state of the microtubules [29,30]. Previous work has shown that disruption of microtubules with nocodazole or stabilization of the microtubules with taxol greatly influences the localization of Rab11a and Rab25 in MDCK cells [5]. We therefore studied the localization of Rab11b in untreated and microtubule-manipulated, polarized MDCK cells. In untreated MDCK cells, both Rab11a and Rab11b localized to an apical, pericentrosomal region (Fig. 3). ␥-Tubulin staining also localized to this area (data not shown). Although Rab11a and Rab11b did stain a similar region of the cell, the patterns were generally nonoverlapping (Fig. 3 (C)). Whether Rab11aand Rab11b-containing vesicles are adjacent to each other in this region or in different regions of the same tubulovesicle cannot be determined at this resolution. While nocodazole dispersed the majority of the Rab11a throughout the cell, the effect on Rab11b distribution was less pronounced (Fig. 3 (N)). Similarly, while treatment with Taxol moved the majority of the Rab11a to the subapical “cor-

In addition to different morphological dynamics, we also studied the association of Rab11a and Rab11b with two important recycled cargoes: transferrin receptor and polymeric IgA receptor. Previous work has shown that Rab11a is involved in the trafficking of transferrin receptor (TfR) in nonpolarized cells [3,9,10], but not in polarized MDCK cells [13]. A previous report suggested that overexpression of GFP-Rab11b altered trafficking of the transferrin receptor in transfected 293 cells [32]. We investigated whether the endogenous Rab11b is involved in TfR trafficking in polarized (MDCK) or nonpolarized (HeLa) cells by first determining whether Rab11b colocalized with the TfR. In HeLa cells (Fig. 4A) dual-stained for Rab11b and the TfR, staining for both Rab11b and the TfR in the perinuclear region was observed, but the patterns were nonoverlapping. This finding is more evident in the enlargements of an individual cell (Fig. 4A): The Rab11b pattern has a more broadly tubular appearance, while the TfR appears more concentrated. There is some overlap of the two patterns in the very central region and could indicate a small population of colocalization that cannot be resolved at this resolution. Previous investigations have demonstrated that, in polarized MDCK cells, transferrin receptor does not recycle to the basolateral membrane through Rab11a-containing recycling vesicles [2,19]. To assess the association of Rab11b with transferrin trafficking in polarized cells, polarized MDCK cells were loaded with Cy3-labeled human transferrin and then costained for Rab11b and ZO-1 (Fig. 4B). The majority of the transferrin localized to a basolateral compartment distinct from the Rab11b compartment, indicating that, similar to Rab11a, Rab11b does not participate in the trafficking of transferrin in polarized cells. Since previous work [5,13] has shown that Rab11a is involved in regulating polymeric IgA receptor (pIgAR) trafficking, we investigated whether Rab11b might also be involved. A subline of MDCK cells that stably express the

328

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

Fig. 4. Rab11b did not colocalize with the transferrin receptor (TfR) in nonpolarized cells or internalized transferrin or the IgAR in polarized cells. (A) HeLa cells were stained for Rab11b (double green) and TfR (CD71; double red). The second row of panels are enlargements of an individual cell from the panel above. (B) MDCK cells were grown on Transwell filters and loaded from the basolateral medium with Cy3-transferrin (triple red), fixed, and stained for Rab11b (triple green) and ZO-1 (triple blue). (C) While a majority of the Rab11a colocalized with the polymeric IgA receptor (pIgAR), only a small subset of Rab11b did. pWE cells, a subline of MDCK cells that express the polymeric pIgAR, were grown on Transwell filters and triple-labeled for IgAR, Rab11a, and Rab11b. The bottom panels are pairwise composites of the three staining patterns. The first letter indicates the green pattern and second letter indicates the red pattern; a, Rab11a; b, Rab11b; R, pIgAR. Bar ⫽ 4 ␮M. Each of these results are representative of five separate experiments.

polymeric IgA receptor (pWE cells) [29] was triple-stained for pIgAR, Rab11a, and Rab11b (Fig. 4C). While the majority of the Rab11a staining colocalized with the pIgAR (a/R bottom row) staining, only a small subset of the Rab11b staining overlapped with pIgAR (b/R bottom row). When taken together with the differences observed with microtubule dynamics, there is little evidence for an association of Rab11b with the Rab11a-containing apical recycling system in polarized MDCK cells. Despite the major differences in distribution, there

was a subtle overlap of Rab11b, Rab11a, and IgAR that may indicate a potential intersection of the Rab11a and Rab11b pathways or just a regional convergence in the pericentrosomal area. A potential point of convergence was also seen at the spreading edges of HeLa cells. We observed tight staining for Rab11b, Rab11a, and the transferrin receptor at the tips of spreading lamellapodia in HeLa cells, especially in subconfluent cells. We observed colocalization of Rab11b and TfR at the tips of cell extensions where Rab11a was also present (data not

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

329

Fig. 5. Overexpression of GFP-Rab11b displaced endogenous Rab11a, while overexpression of GFP-Rab11a did not displace Rab11b. (A) MDCK stable lines expressing the indicated GFP-Rab chimeras were grown on Transwell filters and stained for ZO-1 and either endogenous Rab11a or Rab11b. In the GFP-Rab11b row the arrowhead indicates Rab11a staining in a low GFP-Rab11b-expressing cell and the equivalent area in the GFP and ZO-1 panels, indicating the perinuclear endogenous Rab11a staining. The arrow indicates GFP-Rab11b staining and the equivalents are in the Rab11a and ZO-1 panels indicating the lack of endogenous, perinuclear Rab11a staining. Bar ⫽ 4 ␮M. These results are representative of six separate experiments utilizing three different cloned cell lines. (B) Overexpression of either GFP-Rab11a or GFP-Rab11b decreased the amount of endogenous Rab11a found on vesicles, but did not decrease the endogenous Rab11b. Parent MDCK cells and two stable lines, overexpressing either GFP-Rab11a or GFP-Rab11b, were mechanically sheared and the resulting homogenates were fractionated by differential centrifugation at the indicated speeds; 20 ␮g of protein from each fraction was resolved on a 12% SDS–PAGE gel. The blot was sequentially probed for Rab11a (8H10), Rab11b, and finally for GFP with stripping between probes. The fractions are as follows; H, starting homogenate; PI, 1000g pellet; P5, 5000g pellet; P15, 15,000g pellet; P100, 100,000g pellet; S100, 100,000g supernatant. These results are representative of two separate experiments.

shown). This apparent colocalization could indicate that Rab11b interacts with the TfR trafficking pathway in cell extensions. Alternatively, since these spreading cell extensions are areas of growth where membrane is being added, the colocalization may reflect a general highly active point for vesicle insertion into the membrane. The

general property of vesicles trafficked by members of the Rab11 family may be analogous to the use of Rab11 family members in lower eukaryotes (i.e., YPT31p/32p in yeast) in the process of budding and division [33,34]. Takai and colleagues [35] have noted the importance of Rab11-BP/ Raphillin-11 in the migration of nonpolarized cells.

330

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

Overexpression of GFP-Rab11b displaced endogenous Rab11a, while overexpression of GFP-Rab11a did not displace Rab11b Since our studies with Rab11b and the transferrin receptor appeared to contradict the findings of Schlierf et al. [32], we sought to study the behavior of GFP-Rab11b in transfected MDCK cells. Immunostaining of one of three stable MDCK cell lines expressing wild-type Rab11b tagged with GFP at the amino terminus is shown in Fig. 5A (top row). This region of the monolayer was chosen because it contained both cells overexpressing the GFP-Rab11b (one is marked with an arrow) and cells with nondetectable levels of GFP-Rab11b (one is marked with an arrowhead). Interestingly, in the cells overexpressing GFP-Rab11b we observed a decrease and/or dispersal of the endogenous Rab11a immunoreactivity. Three independent, stable GFPRab11b cell lines demonstrated the same pattern of dispersed endogenous Rab11a staining. These results suggest that the overexpression of Rab11b may displace Rab11a from vesicles within the Rab11a trafficking pathway. In contrast the staining pattern of endogenous Rab11b in a GFP-Rab11a line was similar to the pattern seen in MDCK cells (compare Fig. 5A with Fig. 3C). As in Fig. 3C, there are areas where GFP-Rab11a and endogenous Rab11b appear to be colocalized, although at this resolution we cannot determine whether they are contained on the same tubulovesicle or adjacent ones. To investigate biochemically the observation that overexpression of GFP-Rab11b dispersed the endogenous Rab11a, we fractionated the GFP-Rab cell lines and probed for the endogenous Rabs (Fig. 5B). Within the parent MDCK line, the majority of both Rab11a and Rab11b were found in the 15,000g pellet. In the GFP-Rab11a line, there was a slight decrease in the levels of Rab11b, but there was a significant decrease in the amount of endogenous Rab11a. A more marked decrease in Rab11a was observed in the GFP-Rab11b-expressing cells, confirming the results of the immunofluorescence staining. Interestingly, the endogenous Rab11b was not decreased. These results would indicate that endogenous Rab11a is more sensitive to displacement than endogenous Rab11b. It is important to note that displaced Rab11a did not appear in the 100,000g supernatant fraction of either GFP cell line. These results suggest that either Rab11a is not enriched enough in the cytosol to be detected or the cytosolic Rab11a is quickly degraded. Schlierf et al. reported that the TfR could be found in the same area of Vero cells as overexpressed GFP-Rab11b [32]. We observed no colocalization of endogenous Rab11b and transferrin receptor in the pericentriolar region of nonpolarized HeLa cells or with transferrin in polarized MDCK cells. Based on this lack of colocalization, Rab11b is likely not a major regulator of transferrin trafficking in either polarized or nonpolarized cells. However, we have demonstrated here that overexpression of GFP-Rab11b in MDCK cells led to displacement of Rab11a from its normal asso-

ciation with recycling vesicles. Based on the displacement of Rab11a in the MDCK cells and the building evidence of multiple shared effectors [19,36,37], overexpression of GFP-Rab11b may not reflect the endogenous function of Rab11b. Furthermore, overexpression of Rab11b could have an indirect effect on the Rab11a pathway by efficiently competing for shared effectors. These results demonstrate that interpretation of Rab function using overexpression studies must be compared with the endogenous pattern of Rab protein expression. In the case of Rab11b, trafficking studies using overexpression of either the wild type or mutations cannot be interpreted reliably. The expanding list of shared effectors also supports the existence of convergence points. Our laboratory and others have currently identified eight proteins that interact with all three members of the Rab11 family. Six belong to the same family, the Rab11 family of interacting proteins (Rab11FIP), which consist of Rab11-FIP1, Rab11-FIP2, Rab11FIP3 [37], Rap11-FIP4 [38] and the closely related pp75/ Rip11 [36, 39] and RCP [40]. The Rab11 effector Rab11BP/ Rabphilin-11 [35, 41] and the motor protein Myosin Vb [19] also interact with all members of the Rab11 family. Since the distribution of particular Rab11-FIP proteins also is variable among tissues and cells, pairing of Rab11-interacting proteins with individual Rab11 family members is likely responsible for cell-specific trafficking phenotypes. In summary, we have compared the endogenous distribution of Rab11b and Rab11a in gastric parietal cells as well as in HeLa cells and polarized MDCK cells. The results indicate that Rab11b is not associated with the Rab11acontaining portion of the apical recycling system. In light of the observations of areas of some overlap between Rab11a, IgAR, and Rab11b, and the growing list of shared effectors, some points of convergence may exist between Rab11b and Rab11a trafficking vesicles. Further investigation will be required to identify the cargo and pathway that Rab11b does regulate. Nevertheless, the evidence presented here indicates that endogenous Rab11b resides on membrane elements distinct from those regulated by Rab11a.

Acknowledgments This work was supported by grants to J.R.G. from NIH NIDDK (DK48370 and DK43405) and a Veterans Administration Merit Award. All microscopy was performed in the Imaging Core of the Institute for Molecular Medicine and Genetics at the Medical College of Georgia.

References [1] J. Armstrong, How do Rab proteins function in membrane traffic?, IJBCB 32 (2000) 303–307. [2] P.S. Brown, E. Wang, B. Aroeti, S.J. Chapin, K.E. Mostov, K.W. Dunn, Definition of distinct compartments in polarized Madin–Darby

L.A. Lapierre et al. / Experimental Cell Research 290 (2003) 322–331

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17] [18]

[19]

[20]

[21]

canine kidney (MDCK)cells for membrane-volume sorting, polarized sorting and apical recycling, Traffic 1 (2000) 124 –140. O. Ullrich, S. Reinsch, S. Urbe, M. Zerial, R.G. Parton, Rab11 regulates recycling through the pericentriolar recycling endosome, J. Cell Biol. 135 (1996) 913–924. J.R. Goldenring, J. Smith, H.D. Vaughan, P. Cameron, W. Hawkins, J. Navarre, Rab11 is an apically located small GTP-binding protein in epithelial tissues, Am. J. Physiol. 270 (33) (1996) G515–G525. J.E. Casanova, X. Wang, R. Kumar, S.G. Bhartur, J. Navarre, J.E. Woodrum, Y. Altschuler, G.S. Ray, J.R. Goldenring, Association of Rab25 and Rab11a with the apical recycling system of polarized Madin–Darby canine kidney cells, Mol. Biol. Cell 10 (1999) 47– 61. S.G. Bhartur, B.C. Calhoun, J. Woodrum, J. Kurkjian, S. Iyer, F. Lai, J.R. Goldenring, Genomic structure of murine Rab11 family members, Biochem. Biophys. Res. Commun. 269 (2000) 611– 617. F. Lai, L. Stubbs, K. Artzt, Molecular analysis of mouse Rab11b: a new type of mammalian YPT/Rab protein, Genomics 22 (1994) 610 – 616. J.R. Goldenring, K.R. Shen, H.D. Vaughan, I.M. Modling, Identification of a small GTP-binding protein, Rab25, expressed in the gastrointestinal mucosa. kidney and lung, J. Biol. Chem. 268 (1993) 18419 –18422. E.G. Green, E. Ramm, N.M. Riley, D.J. Spiro, J.R. Goldenring, M. Wessling-Resnick, Rab11 is associated with transferrin-containing recycling compartments in K562 cells, Biochem. Biophys. Res. Commun. 239 (1997) 612– 616. M. Ren, G. Xu, J. Zeng, C. De Lemos-Charandini, M. Adesnik, D.D. Sabatini, Hydrolysis of GTP on Rab11 is required for the direct delivery of transferrin from the pericentriolar recycling compartment to the cell surface but not from sorting endosomes, Proc. Natl. Acad. Sci. USA 95 (1998) 6187– 6192. S. Urbe, L.A. Huber, M. Zerial, S.A. Tooze, R.G. Parton, Rab11, a small GTPase associated with both constitutive and regulated secretory pathways in PC12 cells, FEBS Lett. 334 (1993) 175–182. W., Chen, A. Wandinger-Ness, Expression and functional analyses of Rab8 and Rab11a in exocytic transport from trans-Golgi network, in: W.E. Balch, C.J. Der, A. Hall (Eds.), Methods in Enzymology, Academic Press, San Diego. X. Wang, R. Kumar, J. Navarre, J.E. Casanova, J.R. Goldenring, Regulation of vesicle trafficking in Madin–Darby canine kidney cells by Rab11a and Rab25, J. Biol. Chem. 275 (2000) 29138 –29146. S.-M. Leung, W.G. Ruiz, G. Apodaca, Sorting of membrane and fluid at the apical pole of polarized Madin–Darby canine kidney cells, Mol. Biol. Cell 11 (2000) 2131–2150. J.R. Goldenring, C.J. Soroka, K.R. Shen, L.H. Tang, W. Rodriguez, H.D. Vaughan, S.A. Stoch, I.M. Modlin, Enrichment of rab11, a small GTP-binding protein, in gastric parietal cells, Am. J. Physiol. 267 (1994) G187–G194. B.C. Calhoun, L.A. Lapierre, C.S. Chew, J.R. Goldenring, Rab11a redistribution to apical secretory canaliculus during stimulation of gastric parietal cells, Am. J. Physiol. 275 (1998) C163–C170. U.K. Laemmli, Cleavage of structual proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680 – 685. T. Berglindh, K.J. Obrink, A method for preparing islotaed glands from the rabbit gastric mucosa, Acta Physiol. Scand. 96 (1976) 150 –159. L.A. Lapierre, R. Kumar, C.M. Hales, J. Navarre, S.G. Bhartur, J.O. Burnette, J.D.W. Provance, J.A. Mercer, M. Bahler, J.R. Goldenring, Myosin Vb is associated with and regulates plasma membrane recycling systems, Mol. Biol. Cell 12 (2001) 1843–1857. J.M.J. Crothers, D.C. Chow, J.G. Forte, Omeprazole decreases H⫹K⫹-ATPase protein and increase permeablity of oxyntic secretory membrane in rabbits, Am. J. Physiol. 265 (1993) G231–G241. T. Urushidani, J.G. Forte, Stimulation-associated redistribution of H⫹-K⫹-ATPase activity in isolated gastric glands, Am. J. Physiol. 252 (1987) G458 –G465.

331

[22] C.J. Soroka, C.S. Chew, I.M. Modlin, D. Hanzel, A. Smolka, J.R. Goldenring, Characterization of membrane and cytoskeletal compartment in cultured parietal cells using immunofluorescence and confocal microscopy, Eur. J. Cell Biol. 60 (1993) 76 – 87. [23] J.G. Forte, D.X. Yao, The membrane recruitment-and-recycling hypothesis of gastric HCl secretion, Trends Cell Biol. 6 (1996) 45– 48. [24] T.M. Forte, T.E. Machen, J.G. Forte, Ultrastructural changes in oxyntic cells associated with secretory function: a membrane recycling hypothesis, Gastroenterology 73 (1977) 941–955. [25] C. Okamoto, J.G. Forte, Vesicular trafficking machinery, the actin cytoskeleton, and H⫹-K⫹-ATPase recycling in the gastric parietal cell, J. Physiol. 532 (2001) 287–296. [26] B.C. Calhoun, J.R. Goldenring, Rab proteins in gastric parietal cells: evidence of the membrane recycling hypothesis, Yale J. Biol. Med. 69 (1996) 1– 8. [27] J.G. Duman, K. Tyagarajan, M.S. Kolsi, H.-P.H. Moore, J.G. Forte, Expression of rab11a N124I in gastric parietal cells inhibits stimulatory recruitment of the H⫹-K⫹-ATPase, Am. J. Physiol. 277 (1999) C361–C372. [28] J.G. Forte, B. Ly, Q. Rong, S. Ogihara, M. Ramilo, B. Agnew, X. Yao, State of actin in gastric parietal cells, Am. J. Physiol. 274 (1998) C97–C104. [29] K. Mostov, G. Apodaca, B. Aroeti, C. Okamoto, Plasma membrane protein sorting in polarized epithelial cells, J. Cell Biol. 116 (1992) 577–583. [30] G. Apodaca, Endocytic traffic in polarized epithelial cells: role of the actin and microtubule cytoskeleton, Traffic 2 (2001) 149 –159. [31] C. Bucci, A. Wandinger-Ness, A. Lutcke, M. Chiariello, C.B.Z.M. Bruni, Rab5a is a common component of the basolateral and apical endocytic machinery in polarized epithelial cells, Proc. Natl. Acad. Sci. USA 91 (1994) 5061–5065. [32] B. Schlierf, G.H. Fey, J. Hauber, G.M. Hocke, O. Rosorius, Rab11b is essential for recycling of transferrin to the plasma membrane, Exp. Cell Res. 259 (2000) 257–265. [33] M. Benli, F. Doring, D.G. Robinson, X. Yang, D. Gallwitz, Two GTPase isoforms, Ypt31p and Ypt32p, are essential for Golgi function in yeast, EMBO J. 15 (1996) 6460 – 6475. [34] G. Jedd, J. Mulholland, N. Segev, Two new Ypt GTPases are required for exit from the yeast trans-Golgi compartment, J. Cell Biol. 137 (1997) 563–580. [35] A. Mammoto, T. Ohtsuka, I. Hotta, T. Sasaki, Y. Takai, Rab11BP/ Rabphilin-11, a downstream target of Rab11 small G protein implicated in vesicle recycling, J. Biol. Chem. 274 (1999) 25517–25524. [36] R. Prekeris, J. Klumperman, R.H. Scheller, A Rab11/Rip11 protein complex regulates apical membrane trafficking via recycling endosomes, Mol. Cell 6 (2000) 1437–1448. [37] C.M. Hales, R. Griner, M.C. Dorn, D. Hardy, R. Kumar, J. Navarre, E.K.C. Chan, L.A. Lapierre, J.R. Goldenring, Identification and characterization of a family of Rab11 interacting proteins, J. Biol. Chem. 276 (2001) 39067–39075. [38] D.M. Wallace, A.J. Lindsay, A.G. Hendrick, M.W. McCaffrey, The novel Rab11-FIP/RIP/RCP family of proteins displays extensive homo- and hetero-interacting abilities, Biochem. Biophys. Res. Commun. 292 (2002) 909 –915. [39] D. Wang, J.P. Buyon, W. Zhu, E.K. Chan, Defining a novel 75-kDa phosphoprotein associated with SS-A/Ro and identification of distinct human autoantibodies, J. Clin. Invest. 104 (1999) 1265–1275. [40] A.J. Lindsay, A.G. Hendrick, G. Cantalupo, F. Senic-Matuglia, B. Goud, C. Bucci, M.W. McCaffrey, Rab coupling protein (RCP), a novel Rab4 and Rab11 effector protein, J. Biol. Chem. 277 (2002) 12190 –12199. [41] J. Zeng, M. Ren, D. Gravotta, C. De Lemos-Chiarandini, M. Lui, H. Erdjument-Bromage, P. Tempst, G. Xu, T.H. Shen, T. Morimoto, M. Adesnik, D.D. Sabatini, Identification of a putative effector protein for Rab11 that participates in transferrin recycling, Proc. Natl. Acad. Sci. USA 96 (1999) 2840 –2845.