Coordination of Rab8 and Rab11 in primary ciliogenesis

Abstract

Primary cilia are microtubule-based membrane projections located at the surface of many cells. Defects in primary cilia formation have been implicated in a number of genetic disorders, such as Bardet-Biedl Syndrome and Polycystic Kidney Disease. Recent studies have demonstrated that polarized vesicular transport involving Rab8 and its guanine nucleotide-exchange factor Rabin8 is essential for primary ciliogenesis. Here we report that Rabin8 is a direct downstream effector of Rab11, which functions in membrane trafficking from the trans-Golgi network and recycling endosomes. Rab11, in its GTP-bound form, interacts with Rabin8 and kinetically stimulates the guanine nucleotide-exchange activity of Rabin8 toward Rab8. Rab11 is enriched at the base of the primary cilia and inhibition of Rab11 function by a dominant-negative mutant or RNA interference blocks primary ciliogenesis. Our results suggest that Rab GTPases coordinate with each other in the regulation of vesicular trafficking during primary ciliogenesis.

Primary cilia are centriole-derived, microtubule-based membrane projections at the surface of many cells (1). Primary cilia serve important sensory and signaling functions through the membrane proteins localized on their surface. Defects in primary cilia formation have been implicated in a number of genetic disorders, such as Bardet-Biedl Syndrome and polycystic kidney disease (2, 3). The formation of primary cilia, known as primary ciliogenesis, involves the assembly of microtubule substructures (“axonemes”) and directional transport of proteins to the cilia surface by membrane trafficking. Recently, a number of studies using different model systems have demonstrated the role of vesicular trafficking in primary ciliogenesis (4, 5). Among the proteins identified, Rab8, a member of the Rab family of small GTPases and a regulator of membrane traffic from endosomal compartments to the cell surface (6), plays an essential role in primary ciliogenesis (7–9). The guanine nucleotide exchange factor (GEF) that activates Rab8 is Rabin8 (10). Rabin8 has been shown to interact with the BBSome, a protein complex implicated in Bardet-Biedl Syndrome, and regulate cilia formation (8).

Here, we report that the GTP-bound form of Rab11, which regulates vesicular trafficking from the trans-Golgi network (TGN) and recycling endosomes to the plasma membrane (11–14), directly interacts with Rabin8 and stimulates the guanine nucleotide-exchange activity of Rabin8 toward Rab8. A population of Rab11 is localized at the base of the primary cilia, and inhibition of Rab11 blocks ciliogenesis. Our results suggest that Rab GTPases coordinate their actions during primary ciliogenesis.

Results

Rabin8 Directly Interacts with the GTP-Bound Form of Rab11.

A previous study conducted in HeLa cells showed that Rab8 partially colocalizes with Rab11, but not with Rab4 or Rab5 (15). Rab11 is localized to the TGN and recycling endosomes and has been shown to play an important role in controlling vesicular transport from the TGN and the recycling endosomes to the plasma membrane (11–14). In the budding yeast Saccharomyces cerevisiae, the homologs of Rab11 and Rabin8 are Ypt32p and Sec2p, respectively (10, 16). The GTP-bound form of Ypt32p has been shown to interact with Sec2p, the GEF for the Rab protein Sec4p (Rab8 homolog), which is involved in exocytosis (16, 17). We investigated the interaction between Rab11 and Rabin8 using recombinant fusion proteins. GST-tagged Rabin8 and Hisx6-tagged Rab11a were purified from bacteria and used in in vitro binding experiments. As shown in Fig. 1A, Rabin8 bound to Rab11a more strongly in the presence of GTPγS than GDPβS. As a control, Rabin8 did not bind to Rab5a, which is known to regulate early endosomal trafficking (18). To further test the preferential binding of GTP-Rab11a to Rabin8, we took advantage of Rab11a mutant variants. The Rab11a[Q70L] mutant mimics the GTP-bound activated form of Rab11a; the Rab11a[S25N] mutant mimics the nucleotide-free or GDP-bound form of Rab11a. As shown in Fig. 1B, Rab11a[Q70L] strongly interacted with Rabin8, whereas Rab11a[S25N] showed almost no binding to Rabin8. The stronger binding of Rabin8 to Rab11a[Q70L], compared with the wild-type Rab11a, probably resulted from more stable binding of GTP to Rab11a[Q70L], which is deficient in GTP hydrolysis. This result further indicates that Rabin8 interacts with the GTP-bound form of Rab11.

Fig. 1.

Rabin8 interacts with Rab11. (A) Binding of Hisx6-tagged wild-type and mutant Rab11a to GST-Rabin8. Rab11a preferentially bound to Rabin8 in the presence of GTPγS versus GTPβS. As a control, Rabin8 did not bind to Rab5a. Rab11a was detected by the anti-Hisx6 antibody. (B) Rabin8 bound to Rab11a[Q70L] stronger than to wild-type Rab11a. Rab11a[S25N] barely bound to Rabin8. (C) Diagram of GST fusion constructs of Rabin8 with serial C-terminal deletions. Numbers indicate the amino acid sequences. The GEF domain (amino acids 144–245) on Rabin8 is marked in black. (D) Coomassie Blue-stained gel showing the different Rabin8 fusion proteins coupled to Glutathione Sepharose. GST alone was used as a control. Rab11a was expressed as a Hisx6-tagged fusion protein. Binding of Rab11 to the individual Rabin8 constructs was detected with a monoclonal antibody against the Hisx6 epitope.

To identify the region of Rabin8 that binds to Rab11a, we constructed GST-Rabin8 fusion proteins with serial truncations at its C terminus (Fig. 1C). As shown in Fig. 1D, Rab11a interacted with full-length Rabin8, and to a lesser extent with Rabin8 (amino acids 1–400) and Rabin8 (amino acids 1–331). Further deletion of the C terminus abolished this interaction. Thus, the region containing amino acids 262 to 331 of Rabin8 is critical for its interaction with Rab11a. This region is adjacent to the coiled-coil domain (amino acids 144–245) of Rabin8, which has been shown to mediate guanine nucleotide exchange of Rab8 (19, 20).

Rab11 Stimulates the Guanine Nucleotide-Exchange Activity of Rabin8.

Rabin8 activates Rab8 by promoting the dissociation of GDP from, and subsequent reloading of GTP onto, Rab8 (10). The interaction of the GTP-bound form of Rab11 with Rabin8 led us to test whether it stimulates the GEF activity of Rabin8 toward Rab8. A rate-limiting step in guanine nucleotide exchange is the dissociation of GDP from Rab8 for subsequent GTP-loading (10). We thus tested whether GDP dissociation from Rab8 is promoted in the presence of the activated form of Rab11. Because Rab11a[Q70L] interacts more strongly with Rabin8 (Fig. 1), we used Hisx6-tagged Rab11a[Q70L] in our GEF assay. Recombinant NusA-Hisx6-Rab8a fusion protein was purified from bacteria and Rab8a was cleaved from the NusA-Hisx6 tag using thrombin. Rabin8 was purified as a GST-Rabin8 fusion protein and cleaved from the GST tag using thrombin. Trx-Hisx6-S-tagged Rab11a[Q70L] was purified as described above (see Materials and Methods for details). The purification profile of Rab8 and Rabin8 is similar to what was reported previously (10) (Fig. 2A). The additional bands shown in the gel [also observed in a previous study by Hattula et al. (10)] were most likely the degradation products of Rabin8 or Rab8 based on their recognition by the Rabin8 or Rab8 antibodies (see Fig. S1 for details). Purified Rab8a was preloaded with [³H]GDP and then the nucleotide dissociation reaction was monitored over time (Materials and Methods). To examine the effect of Rab11a[Q70L], Rabin8 was first incubated with an excess amount of Rab11a[Q70L] fusion protein at room temperature for 45 min to allow binding of Rabin8 to Rab11a[Q70L], and then used in the [³H]GDP dissociation reaction. As shown in Fig. 2B, Rabin8 promoted the dissociation of [³H]GDP from Rab8a, consistent with previous reports (10, 21). Addition of Rab11a[Q70L] markedly stimulated the dissociation of [³H]GDP from Rab8a. Rab11a[Q70L] alone cannot stimulate the release from Rab8a (Fig. S2A), suggesting that its stimulatory effect was through Rabin8. Rabin8 does not stimulate the release of GDP from Rab11a (Fig. S2B). To test whether the stimulatory effect on Rabin8 is specific for Rab11, we examined the effect of the activated forms of Rab5 (Rab5a[Q79L]) and Rab3 (Rab3a[Q81L]) on the GEF activity of Rabin8. Rab5a regulates early endosomal trafficking (18); Rab3a is involved in regulated secretion at the plasma membrane (22). As shown in Fig. 2C, neither Rab5a[Q79L] nor Rab3a[Q81L] showed any stimulatory effect on Rabin8-mediated dissociation of [³H]GDP from Rab8a. In addition, neither Rab5a[Q79L] nor Rab3a[Q81L] alone can stimulate the release from Rab8a (Fig. S2A). The slight increase of [³H]GDP retention on the filter in the presence of Rab11a[Q70L], Rab5a[Q79L], and Rab3a[Q81L] probably resulted from the binding of [³H]GDP by these proteins, although an excess amount of unlabeled GDP present in the buffer helped to lessen this problem (Fig. S2) (Materials and Methods).

Fig. 2.

Rab11 stimulates the GEF activity of Rabin8. (A) Coomassie Blue-stained SDS/PAGE showing purified Rabin8, Rab8, and Trx-Hisx6-S-tagged Rab11a[Q70L], Rab5a[Q79L], Rab3a[Q81L]. Molecular weight (MW) is indicated to the left. The asterisks indicate the positions of Rabin8 cleaved from the GST tag and Rab8 cleaved from the NusA-Hisx6 tag, respectively. The additional bands present in the purification are most likely the degradation products of Rabin8 or Rab8, based on their recognition by the anti-Rabin8 or Rab8 antibodies (see Fig. S1 for details). Trx-Hisx6-S-tagged Rab11a[Q70L], Rab5a[Q79L], and Rab3a[Q81L] have the predicted molecular weights, as they are fusions of the Rab proteins with the 18 kDa Trx-Hisx6-S tag. (B) The release of [³H]GDP from Rab8 catalyzed by Rabin8 in the presence and absence of Rab11a[Q70L] was analyzed as described in Materials and Methods. Rab11a[Q70L] stimulated the release of [³H]GDP from Rab8 in the presence of Rabin8 (red). Rab11a[Q70L] significantly enhanced the GEF activity of Rabin8 toward Rab8 (P < 0.01, n = 3). (C) Rabin8-mediated [³H]GDP release from Rab8 was tested in the absence (black bars) and presence (white bars) of Rab5a[Q79L] or Rab3a[Q81L] at 5 min and 20 min (n = 3). Rab5a[Q79L] or Rab3a[Q81L] did not stimulate Rabin8-mediated release of [³H]GDP from Rab8. The data were analyzed using Student's t test and presented as SEM (n = 3).

Localization of Rab11a to Primary Cilia.

Rab8 was previously reported to be localized to the primary cilia in hTERT-RPE1 (Infinity telomerase-immortalized human Retinal Pigment Epithelial) cells, which have been extensively used to study primary ciliogenesis (7, 8). We examined the localization of Rab11 with respect to primary cilia. hTERT-RPE1 cells were costained with antibodies against endogenous Rab11 and acetylated tubulin, a marker for primary cilia. As shown in Fig. 3A, a population of Rab11 is concentrated around the primary cilia. This finding is further confirmed by the colocalization of transfected GFP-Rab11a and pericentrin, a marker for the basal body (Fig. 3B). Rab8 has been previously shown to be distributed along the cilia membrane (7, 8). Rab11, on the other hand, does not localize on the surface of the primary cilia with Rab8 (7). Here we found that GFP-Rab11a was localized at the base rather than the surface of the Rab8-marked cilia projection (Fig. 3C). We also detected Rabin8 at the basal body (Fig. 3D), consistent with previous observations (8). The anti-Rabin8 antibody also stained the nucleus, similar to the previous report (10).

Fig. 3.

Localization of Rab11 at the base of the primary cilia. The localization of endogenous Rab11a (A) and expressed GFP-Rab11a (B–D) was examined in hTERT-RPE1 cells together with the cilia marker acetylated tubulin (A, green); the basal body marker pericentrin (B, red), the cilia membrane marker Rab8a (C, red). Rabin8 was stained with the affinity-purified anti-Rabin8 antibody as previously described (10) (D, red). Merged images are shown to the right. (Scale bars, 10 μm.) Cell boundaries are outlined.

Rab11 Regulates Primary Ciliogenesis.

The interaction of Rab11 with Rabin8, together with the localization of Rab11 to the basal body area, led us to investigate whether inhibition of Rab11 in cells blocks ciliogenesis. hTERT-RPE1 cells were transfected with GFP-tagged Rab11a[S25N], a dominant-negative Rab11 mutant (7, 8, 11, 12, 23, 24). Cilia length was determined by immunostaining acetylated α-tubulin. We found that GFP-Rab11a[S25N]-expressing cells had much shorter cilia than the untransfected cells [1.4 ± 0.2 μm vs. 3.3 ± 0.1 μm (mean ± SEM), n = 50; P < 0.01] (Fig. 4 A and B). As a control, cells transfected with GFP-Rab5a[S34N] had relatively normal cilia length [3.0 ± 0.2 μm (mean ± SEM), n = 50] (Fig. 4 A and B), consistent with the previous report (8).

Fig. 4.

Inhibition of Rab11 function in cells blocks primary ciliogenesis. (A) hTERT-RPE1 cells were transfected with the dominant-negative Rab11a[S25N] and GFP-Rab5a[S34N] mutants. The length of primary cilia in these cells was determined by immunostaining of acetylated α-tubulin (red). GFP-Rab11a[S25N]-expressing cells (green) had much shorter cilia than the untransfected cells in the same field. As a control, cells transfected with GFP-Rab5a[S34N] had normal cilia. Nuclei were stained with DAPI (blue) and the merged images are shown to the right. (Scale bar, 10 μm.) (B) Quantification of cilia length in untransfected, GFP-Rab11a[S25N]- and GFP-Rab5a[S34N]-expressing cells. The data were analyzed using Student's t test and presented as SEM (n = 50, P < 0.01) when comparing GFP-Rab11a[S25N]-expressing cells with the other groups. (C) hTERT-RPE1 cells were treated with siRNA oligos targeting human Rab11a and Rab11b. Cells treated with siRNA oligos against Luciferase (LUC) were used as controls. For the rescue experiment, the Rab11 knockdown cells were transfected with a GFP-Rab11a variant (RAB11’) with nucleotide mutations that mismatch the Rab11 oligos (SI Materials and Methods). Rab11 levels were analyzed by Western blot using the anti-Rab11 polyclonal antibody. (D) The primary cilia in Rab11 knockdown cells were significantly shorter than those in LUC siRNA control cells. Cells transfected with GFP-Rab11a' variant (RAB1' Rescue, green) have near normal length cilia. Primary cilia were immunostained with a mouse monoclonal antibody against acetylated α-tubulin (red). The nuclei were stained with DAPI (blue). The merged images are shown to the right. (Scale bar, 10 μm.) (E) Quantification of cilia length in different cells. The data were analyzed using Student's t test and presented as SEM, n = 60, P < 0.01.

In addition to inhibiting Rab11 using the dominant-negative mutant, we also inhibited Rab11 by RNA interference. There are two major Rab11 isoforms, Rab11a and Rab11b, in human cells (13, 25). siRNA oligos targeting these two human Rab11 isoforms were synthesized using sequences as previously described (25). hTERT-RPE1 cells were treated with siRNA oligos targeting Rab11 and Luciferase (as a control). As shown in Fig. 4C, the Rab11 protein level was knocked down by ∼60% in hTERT-RPE1 cells as analyzed by Western blot using anti-Rab11 polyclonal antibodies, similar to the previous report (25). The primary cilia in Rab11 knockdown cells were shorter than those in the control cells, as revealed by immunofluorescence microscopy of acetylated α-tubulin [2.3 ± 0.1 μm vs. 3.7 ± 0.1 μm (mean ± SEM), n = 60; P < 0.01] (Fig. 4 D and E). Rab11 knockdown cells expressing a Rab11a variant (RAB11') with nucleotide mutations that mismatch the Rab11 oligos (SI Materials and Methods) had near normal-length cilia (3.3 ± 0.1 μm, n = 60; P < 0.01). In addition to the shortening of cilia length, as measured by acetylated tubulin staining, we also found that the average length of Rab8 fluorescence along the cilia is shorter in Rab11 knockdown cells (2.5 ± 0.1 μm, n = 60) than in the control cells (3.8 ± 0.1 μm, n = 60; P < 0.01) (Fig. S3). We did not see a complete loss of Rab8 staining along the cilium. This result could be because of insufficient knockdown of Rab11 in these cells. It is also possible that factors in addition to Rab11 regulate the ciliary localization of Rab8.

The BBSome component BBS1 has previously been shown to also interact with Rabin8 (8). We thus examined the relationship between Rab11 and the BBS1 in respect to Rabin8 binding. Using a binding assay similar to the previous study (8), we found that amino acid 262–331 of Rabin8 is important for its interaction with both BBS1 and Rab11 (Figs. 1 and 5A). In cells expressing the activated form of Rab11 (Rab11a[Q71L]), more BBS1 coimmunoprecipitated with Rabin8 (Fig. 5B). On the other hand, in cells expressing the dominant-negative form of Rab11 (Rab11a[S25N]), less BBS1 coimmunoprecipitated with Rabin8. This result suggests that Rab11 plays a positive role in the association of BBS1 with Rabin8. Future work is needed to investigate the molecular details of the Rab11-Rabin8-BBSome interaction during ciliogenesis.

Fig. 5.

Effect of Rab11 on Rabin8-BBS1 interaction. (A) Mapping the domain of Rabin8 that binds to BBS1. Cell lysates from hTERT-RPE1 cells expressing myc-BBS1 were incubated with GST or GST-Rabin8 fusion proteins with serial truncations at the C terminus (Upper, Ponceau S. staining). The bound myc-BBS1 was analyzed by Western blot using the anti-myc monoclonal antibody. Myc-BBS1 interacts with full-length Rabin8, Rabin8 (amino acid 1–400) and Rabin8 (amino acid 1–331). Further deletion of the C terminus abolished this interaction. Molecular weights (MW, in kilodaltons) are indicated to the right. (B) Expression of activated Rab11 promotes the association of BBS1 with Rabin8. hTERT-RPE1 cells were cotransfected with pEGFP-C1, pEGFPC1-myc-Rab11a, pEGFPC1-myc-Rab11a[Q70L], or pEGFPC1-myc-Rab11a[S25N] together with myc-BBS1. Two days after transfection, cells were harvested and immunoprecipitation assay was performed with anti-Rabin8 antibody. The amount of myc-BBS1 coimmunoprecipitated with Rabin8 was detected by Western blot using the anti-Myc monoclonal antibody. (Upper) The levels of myc-BBS1 and the GFP-myc-tagged Rab11 variants in the cell. (Lower) Myc-BBS1 coimmunoprecipitated by the Rabin8 antibody from the cell lysates. Myc-BBS1 binds to Rabin8 more strongly in cells expressing the Rab11a[Q70L] mutant.

Discussion

The generation of primary cilia involves microtubule organization and polarized membrane trafficking (4, 5). Here we show that two Rab GTPases coordinate in their actions during primary ciliogenesis. The GEF for Rab8, Rabin8, is a direct downstream effector of Rab11. The GTP-bound form of Rab11 physically interacts with Rabin8 through a region adjacent to its GEF domain. Furthermore, Rab11, in its activated form, stimulates the GEF activity of Rabin8 toward Rab8. The stimulatory effect on Rabin8 was specific to Rab11 because neither Rab5 (involved in early endosomal trafficking) nor Rab3a (involved in regulated exocytosis) had any effect on Rabin8-mediated guanine nucleotide exchange toward Rab8. A previous study in yeast demonstrated that the GTP-bound form of Ypt32p interacts with Sec2p, and this interaction mediates the recruitment of Sec2 to the post-Golgi secretory vesicles for the activation of Sec4p (17). It was speculated that this Rab cascade coordinates vesicular trafficking between individual transport steps to ensure proper transition, and thus coordination of the secretory pathway (17, 26). Here, we show the coordination of Rab proteins in mammalian cells that involves stimulation of the GEF activity of a downstream Rab protein. This coordination provides a molecular mechanism for activation of downstream trafficking, thus coupling cargo transport from the TGN/recycling endosomes (involving Rab11) to later vesicle docking and fusion at the plasma membrane (involving Rab8). Although we focus on the activation of Rabin8 by Rab11 here, we do not exclude the possibility that Rab11 is involved in the recruitment of Rabin8 to vesicles in a manner similar to the recruitment of Sec2p to vesicles by Ypt32p in yeast (17).

Because Rab8 and Rabin8 are regulators of membrane trafficking for primary ciliogenesis, we speculate that Rab11 also regulates ciliogenesis. Our study revealed that a population of Rab11 is localized around the base of the primary cilia; inhibition of Rab11 blocked ciliogenesis. Furthermore, we found that Rab11 and BBS1 bind to the same region of Rabin8, and activated Rab11 seems to promote the association of BBS1 to Rabin8. It was previously shown that Rab11 mediates the trafficking of rhodopsin to the rhabdomere, a cilium-derived domain within the apical membrane surface of photoreceptor cells in Drosophila (23). Recently, TGN and recycling endosomes have been implicated in the trafficking of rhodopsin transport carriers to cilia-derived photoreceptor-rod outer segments in retinal cells (27, 28). Rab11, together with FIP3, Arf4, and the Arf GTPase-activating protein ASAP1, participate in the formation of a complex involved in the selection and packaging of cargos designated to the cilia (28). These observations, together with the work described here, indicate that membrane trafficking from the TGN and recycling endosomes mediated by Rab11 plays an important role in ciliogenesis.

The functions of Rab proteins are carried out by their downstream effectors. A known downstream effector of the Rab proteins is the exocyst, an octameric protein complex that mediates vesicle tethering at the plasma membrane (29). The exocyst interacts with Rab11 and is implicated in the trafficking from the TGN (30) or recycling endosomes (24, 31, 32) to the plasma membrane. Recently, the exocyst has also been implicated in ciliogenesis (28, 33–36). Future studies are needed to elucidate the molecular mechanisms by which the Rab proteins function together with the exocyst complex during primary ciliogenesis.

In summary, we have demonstrated that Rab11 and Rab8 communicate with each other through Rabin8, and play an important role in primary ciliogenesis. Future studies using different model systems will provide further insights into the functions of the Rab proteins in primary ciliogenesis, and shed light on the etiologies of ciliopathies.

Materials and Methods

Binding Assays.

The generation of Rabin8 and Rab11 fusion proteins is described in SI Materials and Methods. Purified Rab11a was incubated with GST-Rabin8 immobilized on Glutathione Sepharose in binding buffer (20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 10 mM MgCl₂, 1% Triton X-100, and 1 mg/mL BSA) for 2.5 h at 4 °C. After four washes with binding buffer, bound Rab11 was analyzed using SDS/PAGE and Western blot using a monoclonal antibody against the Hisx6 epitope. Pulldown of myc-BBS1 from cell lysates using GST-Rabin8 fragments was carried out as previously described (8). Coimmunoprecipitation of myc-BBS1 and Rabin8 was carried out using anti-Rabin8 polyclonal antibody (10). The amount of myc-BBS1 was detected by anti-myc monoclonal antibody 9E10.

GEF Activity Assays.

The NusA-Hisx6-Rab8a fusion protein was expressed in the pET43 vector and purified from bacteria as previously described (10). Rab8 was then cleaved from the NusA-Hisx6 tag using thrombin overnight at 4 °C. Recombinant GST-Rabin8 was expressed and purified as described above and cleaved from the GST tag using thrombin (see Fig. S1 for Western blot verification). Rab11a[Q70L], Rab5a[Q79L], and Rab3a[Q81L] were expressed in the pET32a(+) vector and the plasmids were all sequence confirmed. These Rab proteins were purified from bacteria as Trx-Hisx6-S-tagged (18 kDa added to the molecular weight) fusion proteins.

GEF activity of Rabin8 toward Rab8 was assayed based on the previously described protocol with modifications (10). Fifteen picomoles of purified Rab8 was first labeled with 100 pmol [³H]GDP (14.2 Ci/mmol, Perkin-Elmer) in a preloading buffer (20 mM Hepes, pH 7.2, 1 mM EDTA, 1mM DTT, 10 mM MgCl₂) for 30 min at 30 °C. Rabin8 was incubated with or without 0.5 nmol of Rab11a[Q70L], Rab5a[Q79L], or Rab3a[Q81L] in a buffer containing 20 mM Hepes, pH 7.2, 1 mM DTT, 10 mM MgCl₂ for 40 min to allow binding. The protein mixtures were then transferred to the “GDP-saturation” buffer (20 mM Hepes, pH 7.2, 2 mM GDP, 1 mM DTT, 10 mM MgCl₂) before added to [³H]GDP-labeled Rab8 in the guanine-exchange assay. The use of excess unlabeled GDP in this step minimizes the binding of Rab11a[Q70L], Rab5a[Q79L], or Rab3a[Q81L] to [³H]GDP in later steps. The release of [³H]GDP from Rab8 was started by adding purified Rabin8 (5 pmol) (in the absence or presence of Rab11a[Q70L], Rab5a[Q79L], or Rab3a[Q81L]) to a final reaction buffer containing 20 mM Hepes, pH 7.2, 1 mM DTT, 2 mM GDP, 1 mM EDTA, 10 mM MgCl₂ at 25 °C in a total volume of 50 μL. Ten microliter samples were taken at various time points and immediately diluted into 1.5 mL ice-cold washing buffer (20 mM Tris-HCl, pH 8.0, 20 mM NaCl, 1 mM DTT, 10 mM MgCl₂). The samples were then applied to wet nitrocellulose filters mounted on a vacuum manifold (Millipore) and washed four times with 3 mL ice-cold washing buffer. After drying, 4 mL scintillation fluid (EcoLite+, MP Biomedicals) was added and the amounts of [³H]GDP remaining on the filter were measured in a scintillation counter (Beckman). The data were statistically analyzed using Student's t test (n = 3).

Cell Culture and RNA Interference.

hTERT-RPE1 cells were grown in DMEM/F12 with 10% FBS. The induction of cilia was carried out as described previously (8, 26, 37). Briefly, cells were grown in media containing 10% FBS overnight, and then starved for 48 h in Opti-MEM to induce cilia formation. Cilia length was measured and statistically analyzed using Student's t test. Plasmids were transfected into cells using FuGene 6 (Roche). Rab11a, Rab5a, or Rabin8 were cloned in-frame in pEGFP-C1, pJ3EGFP or pJ3-Myc vector for expression in cells. siRNA sequences targeting Rab11a and Rab11b and Luciferase (as a control) are included in the SI Materials and Methods. For siRNA treatment, cells were grown in DMEM/F12 plus 10% FBS. siRNA transfection was carried out in Opti-MEM with Lipofectamine 2000 (Invitrogen). Cells were fixed 48 to 72 h after transfection. The remaining cells on plates were lysed for Western blot analysis using a rabbit anti-Rab11 antibody (US Biological).

Immunofluorescence.

For cilia staining, cells were fixed with 4% paraformaldehyde in PBS for 10 min followed by fixation with ice-cold methanol for 3 min. The cells were then permeabilized with 0.5% Triton X-100 in PBS for 10 min, blocked in PBS with 3% BSA, and incubated sequentially with primary and secondary antibodies. Mouse anti-acetylated α-tubulin monoclonal antibody (6-11B-1) and rabbit anti-γ-tubulin polyclonal antibody were purchased from Sigma-Aldrich. Rabbit anti-Rabin8 antibody was affinity purified as previously described (10). Rabbit anti-Rab11 polyclonal antibody was purchased from US Biological. Mouse anti-pericentrin antibody was purchased from Abcam. Fluorescence-labeled secondary antibodies were purchased from Invitrogen. Cells were imaged with a Leica DM IRB microscope (63× objective) with a high-resolution CCD camera (model ORCA-ER, Hamamatsu Photonics). Images were processed with Adobe Photoshop (Adobe Systems).