Activation of Rab8 guanine nucleotide exchange factor Rabin8 by ERK1/2 in response to EGF signaling

Significance

Rab8 is a key regulator of exocytosis. To understand how Rab8 itself is regulated in cells, it is pivotal to study how its guanine nucleotide exchange factor (GEF) Rabin8 is activated. Here we demonstrate that Rabin8 has an autoinhibitory conformation. Phosphorylation of Rabin8 by ERK1/2 in response to EGF relieves the autoinhibition of Rabin8, thus promoting its GEF activity on Rab8. In cells, blocking Rabin8 phosphorylation inhibits vesicular trafficking to the plasma membrane. Our work identifies a molecular mechanism for Rabin8 activation and reveals a regulatory pathway that controls Rab8 activation and vesicular trafficking in response to extracellular signaling.

Abstract

Exocytosis is tightly regulated in many cellular processes, from neurite expansion to tumor proliferation. Rab8, a member of the Rab family of small GTPases, plays an important role in membrane trafficking from the trans-Golgi network and recycling endosomes to the plasma membrane. Rabin8 is a guanine nucleotide exchange factor (GEF) and major activator of Rab8. Investigating how Rabin8 is activated in cells is thus pivotal to the understanding of the regulation of exocytosis. Here we show that phosphorylation serves as an important mechanism for Rabin8 activation. We identified Rabin8 as a direct phospho-substrate of ERK1/2 in response to EGF signaling. At the molecular level, ERK phosphorylation relieves the autoinhibition of Rabin8, thus promoting its GEF activity. We further demonstrate that blocking ERK1/2-mediated phosphorylation of Rabin8 inhibits transferrin recycling to the plasma membrane. Together, our results suggest that ERK1/2 activate Rabin8 to regulate vesicular trafficking to the plasma membrane in response to extracellular signaling.

Rab GTPases constitute the largest family of small GTPases and are important regulators of membrane trafficking in eukaryotic cells (1). Rab proteins cycle between the GDP- and GTP-bound states. Guanine nucleotide exchange factors (GEFs) stimulate the dissociation of GDP from Rab proteins to allow subsequent GTP loading, thus switching the Rab proteins to their active conformation (2–6). Although a number of GEFs have been identified as critical regulators of specific Rab GTPases, how these GEFs are regulated in cells remains unclear.

Rab8a and its paralogue, Rab8b (henceforth referred together as “Rab8”) are members of the Rab family of proteins that mediate membrane trafficking from the trans-Golgi network (TGN) and recycling endosomes to the plasma membrane (7–12). Rabin8 is a GEF and major activator of Rab8 in cells (8, 13). To understand the molecular mechanism of Rab8 activation, it is pivotal to understand how Rabin8 itself is regulated in cells. Pioneering studies in yeast demonstrated that Sec2p, the yeast homolog of Rabin8, is a direct downstream effector of Ypt32p (the yeast homolog of Rab11), which mediates the budding of vesicles from TGN (14). Ypt32p, together with phosphatidylinositol 4-phosphate [PI(4)P], mediates the membrane recruitment of Sec2p (15). Similar to its yeast counterparts, Rabin8 is a direct downstream effector of Rab11, which mediates the generation of secretory vesicles from TGN and recycling endosomes (16, 17). Our previous studies have shown that Rab11 kinetically stimulates the GEF activity of Rabin8 toward Rab8 (16, 18). Although studies in yeast and mammalian cells have revealed different types of regulation of Sec2p/Rabin8 by Ypt32p/Rab11 (i.e., membrane recruitment vs. kinetic activation), the Rab cascade serves to coordinate the generation of vesicles from the donor compartments to its subsequent transport and fusion (3, 4, 19, 20). The Rab11-Rabin8-Rab8 signaling cascade is implicated in a number of cell biological processes, such as primary ciliogenesis and cystogenesis (16, 17, 20, 21). Despite the recent progress in this field, the molecular nature of Rabin8 activation remains unknown.

In this study, using biochemical and biophysical approaches, we demonstrate that Rabin8 has an autoinhibitory conformation. We identify Rabin8 as a direct phospho-substrate of ERK1/2, the principle kinases in the Ras-MEK-ERK cascade in response to EGF signaling. ERK1/2 phosphorylation relieves the autoinhibition and kinetically stimulates the GEF activity of Rabin8 toward Rab8, thus regulating vesicular trafficking to the plasma membrane.

Results

Rabin8 Is Phosphorylated by ERK1/2.

In an effort to understand how exocytosis is regulated in cells, we have searched for potential phosphorylation of proteins functioning in the exocytic pathway. Sequence analyses and previous mass spectrometry studies led to the identification of two phospho-peptides “¹²VNLAS*PTS*PDLLGVYESG²⁹” and “²⁴³VLSSS*PTS*PTQEPLPGGK²⁶⁰” (S* indicates the phosphorylated serine residues) in Rabin8 (22) (Fig. 1A). These sequences match the consensus ERK1/2 phosphorylation motif (“S/T-P” or “P-x-S/T-P,” where “x” indicates any amino acid). Because ERK1/2 was recently shown to regulate exocytosis (23), we decided to further investigate the phosphorylation of Rabin8 by ERK1/2 both in vitro and in cells. First, recombinant GST, Rabin8, GST-Rab11a, and GST-Rab8a were purified from bacteria and incubated with constitutively activated ERK2 (“ERK2-CA”; Materials and Methods) in the presence of [³²P]γ-ATP. As shown in Fig. 1B, Rabin8, but not GST, GST-Rab8a, or GST-Rab11a, was phosphorylated by ERK2-CA. To test the involvement of these four serine residues (S16, S19, S247, and S250) in ERK-mediated phosphorylation, we performed site-directed mutagenesis to substitute these serine residues with alanine (“Rabin8-4A”). The mutant was then expressed as a GST fusion protein and purified from bacteria. As shown in Fig. 1 C and D, ERK2-CA phosphorylated Rabin8. However, the phosphorylation of the Rabin8-4A mutant was significantly reduced. As a negative control, the ERK2 kinase-dead mutant (“ERK2-KD”) (24) failed to phosphorylate either wild-type Rabin8 or Rabin8-4A. Further analyses of the phosphorylation of the four serine residues using Rabin8 fragments are included in Fig. S1. Our results suggest that the four serine residues, though they may not be the only sites phosphorylated by ERK1/2, are critical for ERK1/2 phosphorylation.

Fig. 1.

ERK1/2 phosphorylate Rabin8 in vitro and in cells. (A) Schematic representation of Rabin8 domains. Asterisks indicate the potential phosphorylation sites at Serine 16, 19, 247, and 250. (B) ERK2 phosphorylates Rabin8, but not GST, GST-Rab11a, or GST-Rab8a in vitro. Purified GST, GST-Rab11a, Rabin8, or GST-Rab8a were incubated with ERK2-CA in the presence of [³²P]γ-ATP in an in vitro kinase assay. The samples were subjected to SDS/PAGE and autoradiography. Molecular weights (“MW”) in kDa are indicated to the left. (C) Rabin8 serine 16, 19, 247, and 250 are involved in ERK2 phosphorylation in vitro. Recombinant Rabin8 and Rabin8-4A mutant (S16/19/247/250A) were incubated with ERK2-CA and ERK2-KD. Samples were analyzed by SDS/PAGE and autoradiography. The upper panel is a Coomassie Blue-stained gel showing the purified wild-type and mutant Rabin8 used in the kinase assay. The lower panel shows the phosphorylation of Rabin8 by autoradiogram. (D) The levels of Rabin8 in vitro phosphorylation were quantified and normalized to the levels of Rabin8 proteins. The numbers in the y axis indicate the relative signal intensities on the autoradiogram. Error bar, SD. **P < 0.01; ***P < 0.001; n = 3. (E) EGF stimulates the phosphorylation of Rabin8 in cells. HEK293T cells expressing Flag-Rabin8 were treated with EGF for 5 min or treated with U0126 for 30 min before EGF treatment. Flag-Rabin8 proteins were immunoprecipitated from cell lysates and detected by Western blotting using anti-Flag antibody or anti-ERK1/2 phospho-substrate antibody. The levels of total ERK1/2 and phospho-ERK1/2 were also determined by Western blotting. (F) Quantification of the levels of Rabin8 phosphorylation in the above experiments. A.U., arbitrary units. Error bar, SD. *P < 0.05; n = 3. (G) The four serine residues on Rabin8 (S16, S19, S247, and S250) are required for phosphorylation by ERK1/2 in HEK293T cells. HEK293T cells were transfected with Flag vector control, Flag-tagged Rabin8, or Rabin8-4A. Immunoprecipitation was performed using the anti-Flag antibody. The immunoprecipitated proteins were analyzed with the anti-ERK1/2 phospho-substrate antibody and the anti-Flag antibody. Flag-Rabin8-4A had a significantly reduced level of phosphorylation compared with wild-type Rabin8. (H) Quantification of the levels of Rabin8 phosphorylation. Error bar, SD. **P < 0.01; n = 3.

In cells, ERK1/2 are activated in response to EGF (25). We therefore examined the phosphorylation of Rabin8 in cells upon EGF treatment. HEK293T cells expressing Flag-tagged Rabin8 were treated with EGF for 5 min. Immunoprecipitation experiments were then carried out with anti-Flag antibodies. Phosphorylation of Flag-Rabin8 was detected with an anti-ERK1/2 phospho-substrate antibody. A strong phospho-Rabin8 signal was detected in cells treated with EGF (Fig. 1 E and F). As a positive control, ERK1/2 were also phosphorylated in response to EGF as previously described (25). Because ERK1 and ERK2 are the only known substrates of MEK1 (26), U0126, a specific inhibitor of MEK1, has been commonly used to block the Ras-MEK-ERK pathway, especially ERK1/2 activation. Reduced levels of phosphorylation were detected in cells treated with U0126 (Fig. 1 E and F). We further found that phosphorylation was significantly decreased in the Rabin8-4A mutant (Fig. 1 G and H). Together, the data suggest that Rabin8 is phosphorylated by ERK in response to EGF.

ERK1/2 Phosphorylation Stimulates the GEF Activity of Rabin8.

We next examined whether phosphorylation of Rabin8 by ERK1/2 affects Rabin8 GEF activity by monitoring [³H]GDP release from Rab8. NusA-Hisx6-tagged Rab8a and GST-tagged Rabin8, Rabin8-4A, and Rabin8-4D (the four serine residues were changed to aspartic acid) were purified from bacteria, and the NusA-Hisx6 tag or GST tag was removed by thrombin or PreScission protease, respectively (Fig. 2A). For the GDP release assay in vitro, we first incubated Rabin8 with ERK2-CA in the presence of ATP for 30 min for phosphorylation. Rab8a was preloaded with [³H]GDP, and the amount of bound [³H]GDP was monitored at various time points as previously described (16, 18). As shown in Fig. 2B, Rabin8 promoted the dissociation of [³H]GDP from Rab8a. Rabin8 pretreated with ERK2-CA further accelerated the dissociation of [³H]GDP from Rab8a, suggesting a stronger GEF activity of Rabin8.

Fig. 2.

Phosphorylation stimulates Rabin8 GEF activity. (A) Coomassie Blue-stained gel showing purified Rab8, Rabin8, Rabin8-4A, and Rabin8-4D used in the GEF assay. MW is indicated to the left. (B) The percentage of [³H]GDP bound to Rab8 was measured over time. Phosphorylation of Rabin8 by ERK2-CA significantly enhanced the GEF activity of Rabin8 toward Rab8. (C) Analysis of the release of [³H]GDP from Rab8 catalyzed by Rabin8, Rabin8-4A, or Rabin8-4D. Rabin8-4D was more potent than Rabin8 (squares) or Rabin8-4A (triangles) in promoting GDP release from Rab8. Results are representative of three independent experiments.

To further confirm the stimulatory effect of ERK phosphorylation on Rabin8, we performed a GDP release assay using Rabin8, Rabin8-4A, and Rabin8-4D. As shown in Fig. 2C, Rabin8-4D caused faster dissociation of [³H]GDP from Rab8a compared with wild-type Rabin8 and Rabin8-4A. Phosphomimetic mutation on S247 and 250 alone, which are located adjacent to the Rab8 GEF domain, clearly increased the GEF activity (Fig. S2). Collectively, these data suggest that phosphorylation of Rabin8 by ERK2 stimulates its GEF activity toward Rab8.

EGF Stimulates Rab8 Activation in Cells.

Because ERK1/2 phosphorylation stimulates the GEF activity of Rabin8 in vitro, we examined whether phosphorylation of Rabin8 activates Rab8 in cells. JFC1 (synaptotagmin-like protein 1) is a Rab8 effector that specifically interacts with the GTP-bound form of Rab8 (27, 28). We thus used GST-JFC1 purified from bacteria to pull down the GTP-bound Rab8 from HeLa cell lysates. As shown in Fig. 3 A and B, a higher level of Rab8 was pulled down by GST-JFC1 from HeLa cells treated with EGF. Pretreatment of cells with U0126 decreased the Rab8-GTP level. Our data suggest that Rab8 activity is up-regulated in cells in response to EGF.

Fig. 3.

Rab8 is activated in cells in response to EGF. (A) HeLa cells were serum-starved overnight and then treated with EGF for 5 min or incubated with U0126 before EGF treatment. Cell lysates were incubated with purified GST-JFC1 fusion protein. The amounts of GTP-Rab8 bound to GST-JFC1 were analyzed by Western blotting with anti-Rab8 antibody. The cell lysates were also analyzed for total Rab8, ERK1/2, and phospho-ERK1/2. (B) Quantification of GTP-Rab8 levels in the above experiments. The amounts of GTP-Rab8 were normalized to the control level. The intensity of the bands was quantified by ImageJ and analyzed using the Student t test. Values are presented as mean ± SD. *P < 0.05; **P < 0.01; n = 3. (C) GTP-Rab8 was pulled down by GST-JFC1 from HeLa cells stably expressing GFP-tagged wild-type Rabin8, Rabin8-4A, or Rabin8-4D. (D) Quantification of GTP-Rab8. The levels of GTP-Rab8 are normalized to the level of GTP-Rab8 in cells expressing GFP-Rabin8 (n = 3).

To further test whether Rabin8 phosphorylation plays a role in regulating Rab8 activity, we established HeLa cell lines stably expressing wild-type Rabin8, phosphorylation-deficient mutant (“Rabin8-4A”), or phosphorylation-mimic mutant (“Rabin8-4D”). HeLa cells were starved overnight to minimize the influence of growth factor in the medium. The cells were then lysed and incubated with GST-JFC1 for the pull-down assay to detect the activated GTP-bound form of Rab8. As shown in Fig. 3 C and D, a higher level of Rab8 were in their active form in cells expressing Rabin8-4D mutant compared with those expressing Rabin8 or Rabin8-4A. Taken together, our results suggest that Rab8 activity is regulated by ERK phosphorylation of Rabin8.

Conformational Changes of Rabin8 Induced by ERK Phosphorylation.

We have previously shown that a Rabin8 mutant (“Rabin8Δ(300-305)”) lacking a highly conserved region (amino acid 300–305 “SLYNEF”) after the GEF domain displayed higher binding affinity for Rab8 and showed stronger GEF activity than wild-type Rabin8 (18). It is possible that Rabin8 adopts an autoinhibitory conformation, with its C terminus masking the GEF catalytic domain. To test this hypothesis, we used bioluminescence resonance energy transfer (BRET) technology (29–31). Rabin8 was tagged with NanoLuc and HaloTag (Promega) at its N and C terminus, respectively (Fig. 4A). Upon application of substrate furimazine, NanoLuc produces luminescence that serves as resonance energy donor (λem = 460 nm). If Rabin8 is in a closed conformation, the energy will be transferred to the NanoBRET ligand on HaloTag and produce luminescence with maximal emission at 620 nm (Fig. 4A). To test this system, we first assayed the t-SRARE protein, syntaxin-4 (STX4), which is presumably in a closed conformation according to structure studies of syntaxins (32, 33). As shown in Fig. 4B, tagged STX4 and Rabin8 emitted BRET signals upon addition of furimazine (Fig. 4 B and C). As a negative control, cleavage of the HaloTag by TEV protease results in loss of BRET signal on STX4 or Rabin8 fusion protein. It was previously shown that Rabin8Δ(300-305) has stronger GEF activity than wild-type Rabin8 (18). In our BRET assay, this “gain-of-function” mutant had significantly decreased BRET signal compared with wild-type Rabin8 (Fig. 4D) [shown as “BRET ratio,” the ratio of acceptor and donor bioluminescence emission (34); details in Materials and Methods]. The result is consistent with the observation that deletion of the “SLYNEF” motif relieves the GEF domain from autoinhibition (18).

Fig. 4.

Phosphorylation by ERK relieves the autoinhibition of Rabin8. (A) Schematic diagram showing the use of BRET in analyzing Rabin8 autoinhibition. NanoLuc (BRET donor) and HaloTag (BRET acceptor) were fused to the N and C terminus of Rabin8, respectively. If Rabin8 is in a closed conformation, BRET will occur owing to the close proximity between the donor and acceptor. However, the BRET signal will be significantly decreased when Rabin8 switches to an “open” conformation. (B and C) Intramolecular BRET signals detected for syntaxin-4 (“STX4”) and Rabin8 in the range of 590–650 nm. STX4 and Rabin8 were expressed using the pNLHT vector and purified from E. coli. STX4 was used as a positive control. Both Rabin8 and STX4 show positive BRET signals in the assay (red curves). As negative controls, STX4 and Rabin8 fail to be excited (blue curves) upon HaloTag removal by TEV protease. (D) BRET analysis of Rabin8 and Rabin8Δ(300-305). The luminescence intensities at 460 nm and 620 nm were measured for the donor and acceptor, respectively. The excitation was shown as a normalized ratio of RLU_620nm/RLU_460nm of substrate/ligand group minus the RLU_620nm/RLU_460nm of substrate-only group (Materials and Methods). The BRET ratios were normalized. Error bars, SD. *P < 0.05; n = 3. (E) Purified Rabin8-NLHT proteins were first incubated with ERK2-KD or ERK2-CA and then used in the BRET assay. BRET ratios were normalized and compared. **P < 0.01; n = 5. (F) Protein samples in E were subjected to SDS/PAGE, followed by Western blotting with the anti-Rabin8 antibody and the anti-ERK1/2 phospho-substrate antibody. (G) The luminescence intensities of Rabin8 and Rabin8-4D were measured. The BRET ratios were calculated and compared. **P < 0.01; n = 3.

To examine the effect of ERK phosphorylation on Rabin8, we incubated the Rabin8 fusion protein with ERK2-CA or ERK2-KD and measured the BRET signal. Rabin8 treated with ERK2-CA demonstrated significantly reduced BRET signal compared with that treated with ERK2-KD (Fig. 4 E and F). Similarly, the phosphomimetic Rabin8-4D mutant showed lower BRET signal compared with wild-type Rabin8 (Fig. 4G). Taken together, these data suggest that ERK phosphorylation relieves the autoinhibition of Rabin8, which is consistent with the enhanced GEF activity of phosphorylated Rabin8 shown in Fig. 2.

Phosphorylation of Rabin8 Promotes Its Binding to Rab8.

To examine the effect of phosphorylation of Rabin8 on its binding to Rab8, we incubated GST-Rabin8 with ERK2-CA or ERK-KD for 30 min and then used them to pull down GFP-Rab8a[T22N] from cell lysates. The Rab8a[T22N] mutant was used as it preferentially interacts with Rabin8 (8, 16). As shown in Fig. 5 A and B, significantly more Rab8a[T22N] bound to GST-Rabin8 pretreated with ERK2-CA. The binding data further suggest that ERK phosphorylation relieves the autoinhibition of Rabin8, allowing its subsequent interaction with Rab8.

Fig. 5.

Phosphorylation of Rabin8 promotes its binding to Rab8 but inhibits its interaction with Rab11. (A) Purified GST-Rabin8 was treated with ERK2-KD or ERK2-CA and then conjugated to glutathione Sepharose to test the binding to GFP-Rab8a[T22N] expressed in HEK293T cells. Ponceau S staining shows the GST-Rabin8 or GST on beads. MWs are shown to the left. The proteins bound to the beads were detected using anti-ERK1/2 phospho-substrate antibody or anti-GFP monoclonal antibody. (B) Quantification of GFP-Rab8a[T22N] bound to GST-Rabin8 with different treatments. Error bars, SD. *P < 0.05; n = 3. (C) Purified Rab11a or Rab11a[Q70L] was incubated with GST-Rabin8 that was pretreated with either ERK2-CA or ERK2-KD. The phosphorylation of Rabin8 and the bound Rab11a were analyzed by Western blotting. Ponceau S staining shows the amounts of GST-Rabin8 and GST used in the binding assay. MWs are shown to the right. (D) Quantification of Rab11a bound to GST-Rabin8 with different treatments. *P < 0.05; **P < 0.01; n = 3. (E) The binding of recombinant Rab11a[Q70L] to Rabin8, Rabin8-4A, and Rabin8-4D was examined. Coomassie Blue staining shows the GST-tagged Rabin8 variants, and GST was used as negative control. The bound Rab11a[Q70L] was analyzed by Western blotting using anti-Rab11a antibody.

Phosphorylation of Rabin8 Inhibits Its Interaction with Rab11a.

We have previously shown that the GTP-bound form of Rab11a interacts with Rabin8 and promotes its GEF activity (16). In addition, Rabin8 activation seems to inversely correlate with Rab11a binding (18). Recently it was shown that phosphorylation of Sec2p (the yeast homolog of Rabin8) by an unknown kinase(s) inhibits its binding to Ypt32p (the yeast homolog of Rab11) (35). Here we tested whether phosphorylation of Rabin8 by ERK affects its interaction with Rab11. GST-tagged Rabin8 was preincubated with ERK2-CA or ERK2-KD and used in the in vitro binding assay with wild-type Rab11a or constitutively active Rab11a[Q70L]. As shown in Fig. 5 C and D, phosphorylated Rabin8 significantly reduced binding to Rab11a and Rab11a[Q70L]. Furthermore, the phosphomimetic Rabin8-4D mutant showed a remarkably decreased interaction with Rab11a[Q70L] compared with Rabin8 or Rabin8-4A (Fig. 5E). These results suggest that phosphorylation of Rabin8 by ERK reduces its interaction with Rab11a.

Rabin8 Phosphorylation Regulates Transferrin Recycling to the Plasma Membrane.

Because ERK phosphorylation of Rabin8 stimulates its GEF activity, we examined whether Rabin8 phosphorylation regulates membrane trafficking from recycling endosomes to the plasma membrane. We used the standard transferrin (Tf) recycling assay, which has been commonly used to study protein trafficking to the plasma membrane. First, we examined whether Rabin8 is indeed involved in Tf recycling from the recycling endosomes to the plasma membrane. Endogenous Rabin8 in hTERT-RPE1 cells was knocked down using siRNA. The cells were pulsed with Alexa Fluor 488-Tf on ice for 5 min and chased with complete media at 37 °C to track Tf trafficking over time. Most of the Tf was exocytosed after 45–60 min chase in control cells. However, there was significant accumulation of Tf at perinuclear regions in the cells with Rabin8 knockdown (Fig. 6 A–C). Immunostaining showed that the accumulated Tf mostly colocalized with Rab11, a marker for the recycling endosomes (Fig. 6 D and E), suggesting that Rabin8 controls Tf exocytosis from the recycling endosome to plasma membrane.

Fig. 6.

Rabin8 phosphorylation regulates transferrin recycling. (A) Rabin8 regulates Tf recycling. hTERT-RPE1 cells were treated with either Luciferase siRNA or Rabin8 siRNA and then used in Tf recycling assay. Cells were fixed at indicated time points for microscopy. (Scale bar, 10 µm.) (B) Western blotting showing the knockdown of Rabin8 in cells. Rab8 and actin were used as controls. (C) The amounts of Tf retained in cells were quantified using ImageJ. Error bars, SD. One hundred cells were analyzed in each experiment (n = 3). *P < 0.05. (D) Tf was retained in recycling endosomes (colocalized with Rab11) in Rabin8 knockdown cells but exocytosed in control cells after chase for 60 min. Higher-magnification views of the boxed areas are shown under each image. (Scale bar, 10 µm.) (E) Fluorescence intensity of Rab11 and Tf signals along the line were analyzed using ImageJ. Tf partially colocalized with Rab11. (Scale bar: 10 μm.) (F) hTERT-RPE1 cells stably expressing siRNA-resistant, GFP-tagged Rabin8, Rabin8-4A, or Rabin8-4D were treated with siRNA against endogenous Rabin8 and subjected to transferrin recycling assay. Exocytosis of Tf was delayed in cells expressing Rabin8-4A. (Scale bar, 10 µm.) (G) The amounts of cell-associated Tf in F were quantified using ImageJ and plotted in the bar graph. Three independent experiments were performed, and 100 cells were analyzed for each experiment. *P < 0.05.

To examine whether Rabin8 phosphorylation regulates Tf recycling, hTERT-RPE1 cells stably expressing siRNA-resistant GFP-Rabin8, Rabin8-4A, or Rabin8-4D were transfected with siRNA to knock down the endogenous Rabin8. The cells were incubated with Alexa Fluor 594-Tf 48 h after transfection and chased for indicated periods of time. As shown in Fig. 6 F and G, most of the Tf were recycled back to the plasma membrane and exocytosed in cells expressing Rabin8 or Rabin8-4D. In contrast, a significant amount of the internalized Tf was retained in the perinuclear region in Rabin8-4A–expressing cells. Tf recycling in Rabin8-4D cells seems to be similar to those expressing wild-type Rabin8. This may be due to the existing basal level of wild-type Rabin8 phosphorylation in cells in the culture medium. Together, our result suggests that Rabin8 phosphorylation is important for Tf recycling.

Discussion

Rab8 is a master regulator of exocytic trafficking. To understand how Rab8 is activated in cells, it is pivotal to understand the molecular mechanisms of Rabin8 activation and how Rabin8 is regulated in cells. In this study, our BRET analyses and GEF assays suggest that Rabin8 is in a self-inhibitory conformation; phosphorylation by ERK1/2 relieves such inhibition and leads to Rabin8 activation.

In cells the GEFs are controlled by a multitude of signaling pathways, which coordinate membrane trafficking with other cellular activities during many physiological processes such as cell proliferation, morphogenesis, and migration. Our study shows that Rabin8 is phosphorylated by ERK1/2 in response to EGF. Four serine residues (S16, S19, S247, and S250) mediate the phosphorylation. It is possible that phosphorylation of these residues leads to conformational changes on Rabin8, which relieves its autoinhibition. Our BRET assay and GEF activity analyses are indeed consistent with this speculation. Using transferrin recycling assay, we further demonstrate that ERK1/2 phosphorylation of Rabin8 regulates vesicular trafficking to the plasma membrane.

Although our study focuses on the effect of ERK1/2 on the activation of Rabin8, previous studies have shown the phosphorylation of Rabin8 by NDR2 at a different site, S272 (36). NDR2 phosphorylation was implicated in primary ciliogenesis in mammalian cells (36). In flies, this phosphorylation event was shown to be important for dendrite arborization (37). Phosphorylation of Rabin8 at S272 promotes binding to the downstream exocyst subunit Sec15 and inhibits binding to phosphatidylserine. In the budding yeast Saccharomyces cerevisiae, the Rabin8 homolog Sec2p is phosphorylated at multiple sites (35). Although the responsible kinases remain unknown, phosphorylation of Sec2p switches its binding from Ypt32p and PI(4)P at the Golgi to Sec15p. Therefore, the phosphorylation is thought to facilitate maturation of the secretory vesicle and promote their directional transport (35). In Candida albicans, Sec2p was shown to be phosphorylated by the cell cycle kinase Cdc28, and this phosphorylation is important for hyphal development (38). Here, to our knowledge, for the first time, we show that phosphorylation of Rabin8 by ERK1/2 serves as an important activation mechanism in cells. We also detected the dissociation of phosphorylated Rabin8 from Rab11, which functions upstream in the recruitment and/or activation of Rabin8 (16, 18). It is likely that Rabin8 and Sec2p are direct targets of many kinases that function in different signaling contexts, and regulation of exocytosis through phosphorylation of Rabin8/Sec2p by these kinases is crucial for a range of cellular processes.

The Ras-MEK-ERK signal pathway is a basic signaling system that orchestrates various cellular activities to accomplish a wide range of physiological functions (25). A number of studies have implicated MEK-ERK in the regulation of vesicle trafficking (reviewed in ref. 39). For exocytosis, it was recently shown that ERK1/2 phosphorylate the exocyst subunit Exo70 and promote exocyst complex assembly (23). Together, ERK1/2 respond to extracellular signals to phosphorylate key components of the secretory pathway, thus optimally regulating membrane trafficking.

Materials and Methods

Cell Culture, Reagents, and Plasmids.

Information for cell culture, reagents, plasmids, and RNAi experiments is included SI Materials and Methods.

In Vitro Kinase Assays.

Rab8a, Rab11a, Rabin8, and Rabin8-4A mutant were expressed as GST fusion proteins in the pGEX-6P-1 vector and purified from bacteria using glutathione resin (GE Healthcare). Hisx6-ERK2-CA and Hisx6-ERK2-KD were expressed using an ERK2-MEK1 coexpression system whereby MEK1 phosphorylates and activates ERK2 (24). Purified Rabin8 proteins were incubated with Hisx6-ERK2-CA in the kinase buffer [20 mM Hepes (pH 7.5), 100 mM KCl, 5 mM MgCl₂, 1 mM DTT, 1 mM NaF, and 1 mM PMSF] in the presence of 10 µCi [³²P]γ-ATP for 30 min at 30 °C. Samples were analyzed by SDS/PAGE and autoradiography.

Rab8 [³H]GDP Release Assay.

The NusA-Hisx6-Rab8a and GST-Rabin8 fusion proteins were purified from bacteria as previously described (16). GEF activity of Rabin8 was assayed according to the previously described protocol (16). The data were statistically analyzed using the Student t test (n = 3).

Detection of GTP-Bound Rab8 in Cell Lysates.

GTP-bound Rab8 in cell lysates was detected by JFC1 pull-down assay as reported previously (27), with some modifications. HeLa cells treated with EGF or U0126 were lysed in lysis buffer [20 mM Tris⋅HCl (pH 7.4), 100 mM KCl, 5 mM MgCl₂, 0.5% Triton X-100, 1 mM DTT, 1 mM PMSF, serine/threonine phosphatase inhibitor mixture, and protease inhibitor mixture]. The cell lysates were incubated with GST or GST-JFC1 at 4 °C for 4 h, followed by washing four times with binding buffer. Bound GTP-Rab8 was analyzed using SDS/PAGE and immunoblotting using an anti-Rab8 antibody, and cell lysates were subjected to Western blotting for the detection of ERK1/2, phos-ERK1/2, and Rab8. Quantification of bindings was performed using ImageJ software. The data were statistically analyzed using the Student t test (n = 3).

Fluorescence Microscopy.

Cells were grown on glass coverslips and fixed in 4% paraformaldehyde in PBS for 15 min and then permeabilized in 0.1% Triton X-100 in PBS for 5 min. After washing, the cells were blocked in PBS with 5% (vol/vol) FBS for 30 min. Then cells were incubated with primary antibodies for 1 h, followed by washing and incubation for another hour with fluorescence labeled secondary antibodies (Life Technologies). Fluorescence observation was done using a Leica DMI 6000B inverted microscope equipped with a DFC350 FX camera and a 63× objective.

Transferrin Recycling Assays.

hTERT-RPE1 cells grown on coverslips were washed with MEM and incubated with serum-free MEM for 30 min. Alexa Fluor 488 or Alexa Fluor 594-labeled human transferrin (Life Technologies) was incubated with cells for 5 min at 50 µg/mL at 4 °C. Cells were then washed three times in ice-cold media and incubated with fresh complete media at 37 °C for different time points and immediately fixed for microscopy. Fluorescence intensity of transferrin retained in cells was measured using ImageJ software.

BRET Assays.

Syntaxin-4, Rabin8, Rabin8Δ(300-305), and Rabin8-4D were constructed into the pNLHT vector and purified from Escherichia coli using TALON metal affinity resin (Clontech). The fusion proteins were dialyzed in PBS overnight at 4 °C. NanoBRET Ligand (Promega) and Furimazine (the substrate of NanoLuc) were diluted to 200 nM and 20 µM in PBS, respectively. The purified proteins were used at 10 nM as the final concentration in the reactions. The indicated amounts of proteins were incubated with 100 µL of diluted ligand or PBS in a 96-well plate for 1 min at room temperature. Diluted (100 µL) substrate was then added and mixed thoroughly, and the samples were immediately used for luminescence reading at 620 nm (acceptor) and 460 nm (donor) on a Varioskan Luminometer or Gemini EM Fluorescence Microplate Reader. The calculation of BRET ratio (see below) was performed as previously described (34) (RLU, relative luminescence units):

$$\begin{array}{l}\mathrm{BRET}\hspace{0.17em}\mathrm{RATIO}={\mathrm{RLU}}_{620\hspace{0.17em}\mathrm{nm}}\left(\mathrm{substrate}+\mathrm{ligand}\right)/{\mathrm{RLU}}_{460\hspace{0.17em}\mathrm{nm}}\left(\mathrm{substrate}+\mathrm{ligand}\right)\\ -{\mathrm{RLU}}_{620\hspace{0.17em}\mathrm{nm}}\left(\mathrm{substrate}\right)/{\mathrm{RLU}}_{460\hspace{0.17em}\mathrm{nm}}\left(\mathrm{substrate}\right).\end{array}$$

Binding Assays.

GST-Rabin8 was incubated with purified Hisx6-ERK2-CA or Hisx6-ERK2-KD in the kinase buffer for phosphorylation as described and then incubated with glutathione beads for 2 h. The beads were cleaned by washing in binding buffer [20 mM Tris⋅HCl (pH 7.4), 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 10 mM MgCl₂, and 1% Triton X-100] and used directly for binding. HEK293T cells transfected with GFP-Rab8a[T22N] were lysed for binding. Mammalian cell lysates expressing GFP-Rab8a[T22N] or bacterially purified Trx-Hisx6-S-Rab11a or Trx-Hisx6-S-Rab11a[Q70L] were incubated with GST-Rabin8 in the binding buffer for 2 h at 4 °C. Bound GFP-Rab8a[T22N] or Trx-Hisx6-S-Rab11a was analyzed by Western blotting.