RalB uncoupling from exocyst is required for endothelial Weibel-Palade body exocytosis

Abstract

Ras-like (Ral) GTPases play essential regulatory roles in many cellular processes, including exocytosis. Cycling between GDP- and GTP-bound states, Ral GTPases function as molecular switches and regulate effectors, specifically the multisubunit tethering complex exocyst. Here, we show that Ral isoform RalB, but not RalA, is necessary for regulated exocytosis of Weibel-Palade bodies (WPBs), the specialized endothelial secretory granules that store hemostatic protein von Willebrand factor. Remarkably, unlike typical small GTPase–effector interactions, RalB binds exocyst in GDP-bound state. Upon endothelial cell stimulation, exocyst is uncoupled from RalB-GTP resulting in WPB tethering and exocytosis. Furthermore, we report that protein kinase C (PKC)-dependent phosphorylation of the C-terminal hypervariable region (HVR) of RalB modulates its interaction with exocyst. Exocyst preferentially interacts with phosphorylated RalB in resting endothelium. Dephosphorylation of RalB either by endothelial cell stimulation, or PKC inhibition, or expression of nonphosphorylatable mutant at a specific serine residue of RalB HVR, disengages exocyst and augments WPB exocytosis, resembling a RalB exocyst-binding site mutant. In summary, uncoupling of exocyst from RalB promotes endothelial WPB exocytosis. Our data show that RalB may be more dynamically regulated by phosphorylation and may confer distinct functionality given the high degree of homology and the shared set of effector protein between the two Ral isoforms.

• Functioning as molecular switches, Ral GTPases control various cellular functions via specific effectors such as multisubunit tethering complex exocyst; however, it remains unclear how exocyst binding to Ral is regulated.

• Studying endothelial cells, the authors find that RalB, but not RalA, controls regulated exocytosis of specialized granules Weibel-Palade bodies (WPBs). Unusually, unlike typical small GTPase–effector interactions, RalB-GDP binds exocyst, while exocyst is uncoupled from RalB-GTP upon endothelial stimulation, facilitating WPB exocytosis. Furthermore, PKC-dependent phosphorylation of RalB hypervariable region (HVR) promotes exocyst binding in resting cells; RalB dephosphorylation upon cell stimulation reduces exocyst binding facilitating WPB exocytosis.

• RalB uncoupling from exocyst promotes endothelial WPB exocytosis and RalB is dynamically regulated by post-translational phosphorylation.

INTRODUCTION

Regulated exocytosis is fundamental to normal functioning of specialized eukaryotic cells (Gerber and Sudhof, 2002). Examples include blood cells as well as endocrine cells that contain specialized secretory granules, which, in response to specific stimuli, release cargoes that play essential roles in processes such as inflammation and metabolism (Sharda and Flaumenhaft, 2018; Cassatella et al., 2019; Campbell and Newgard, 2021). Endothelial cells that line the vasculature are critical to maintaining vascular integrity, but also release numerous bioactive products in response to vascular injury by the process of regulated exocytosis. von Willebrand factor (vWF), a large multimeric multifunctional plasma glycoprotein, is one such factor that is indispensable for hemostasis and thrombosis (Moake et al., 1986; Federici, 2003; Springer, 2014). Deficiency of functional vWF is the most common bleeding disorder encountered in humans, but elevated levels of vWF also increase cardiovascular morbidity (Spiel et al., 2008; Sadler et al., 2022). Additionally, vWF plays important roles in angiogenesis, smooth muscle cell proliferation, and pathophysiology of vascular inflammation and cancer metastasis (Kawecki et al., 2017; Randi et al., 2018; Patmore et al., 2020; Drakeford et al., 2022). Therefore, regulated localized release of vWF is important for maintenance of vascular integrity but also for numerous other pathophysiologic processes.

vWF in endothelial cells is stored in specialized secretory granules, Weibel-Palade bodies (WPBs) (McCormack et al., 2017; Nass et al., 2021). vWF forms C-terminal dimers immediately after synthesis in the endoplasmic reticulum followed by N-terminal multimerization in the trans-Golgi network (TGN) (McCormack et al., 2017; Nass et al., 2021). Nascent immature WPBs arise from budding at the TGN, facilitated by tubulating vWF multimers, and then relocate along microtubules to the periphery of the cell, where F-actin anchored WPBs await stimulatory signals to exocytose and release their cargoes extracellularly (Zenner et al., 2007). The process of maturation is marked by increasing WPB electron density caused by further multimerization, compaction, and organization of the vWF multimers into elongated tubules that impart WPBs their characteristic “cigar” shape, 0.5–5 µm in length (McCormack et al., 2017). The process from WPB biogenesis to maturation and ultimately its evoked exocytosis requires coordination of a large and complex trafficking machinery that remains to be completely defined (Schillemans et al., 2019; Hordijk et al., 2024). Specifically, the mechanisms underlying regulated WPB fusion with the plasma membrane in response to endothelial cell stimulation is not fully understood (Schillemans et al., 2019; Hordijk et al., 2024).

Small GTPases of the Ras superfamily are a large group of monomeric proteins that bind and hydrolyze GTP to GDP and act as master regulators of various essential cellular processes including vesicle trafficking (Wennerberg et al., 2005). GDP to GTP exchange and subsequent hydrolysis are accelerated by specific guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs), respectively. GTP-bound small GTPases bind and activate downstream effectors, thereby acting as molecular switches turning on and off corresponding downstream functions. Ral belongs to the Ras branch of Ras superfamily of GTPases, and exists in two highly related isoforms, RalA and RalB (Chardin and Tavitian, 1986, 1989). Despite a high overall amino acid sequence identity of 83% and a shared set of effectors, RalA and RalB possess isoform-specific functions (Hazelett et al., 2011). This includes differential effects of RalA and RalB on actin organization, induced gene expression, cell survival, autophagy, endocytosis, and exocytosis (Apken and Oeckinghaus, 2021). Consequently, RalA and B control many physiologic functions including glucose and lipid metabolism, immune responses, neuronal plasticity, and oncogenesis among others (Nakashima et al., 1999; Martin and Der, 2012; Gentry et al., 2014; Shirakawa and Horiuchi, 2015; Apken and Oeckinghaus, 2021). Ral has previously been shown to regulate WPB exocytosis (de Leeuw et al., 1999; de Leeuw et al., 2001; Rondaij et al., 2008) Specifically, inhibition of Ral by a small inhibitory peptide resulted in impaired vWF release. However, the specific Ral isoform involved and how Ral regulates WPB exocytosis remains unknown.

One of the major effectors of RalA and B in their diverse functions is the exocyst (Moskalenko et al., 2002; Moskalenko et al., 2003). The exocyst belongs to a family of multisubunit tethering complexes that promote granule fusion by physically tethering vesicles to the target membrane (Lepore et al., 2018). Exocyst comprises eight subunits, EXOC1-8 (sec3, sec5, sec6, sec8, sec10, sec15, Exo70, and Exo84 in yeast), and is implicated in vesicular trafficking, exocytosis, cytokinesis, ciliogenesis, and cell division (Wu and Guo, 2015). Ral is not required for exocyst subunit synthesis but promotes exocyst assembly on secretory granules and exocyst function (Spiczka and Yeaman, 2008; Hazelett et al., 2011). Ral specifically interacts with exocyst subunits EXOC2 (sec5) and EXOC8 (Exo84) competitively and controls their downstream cellular functions (Hazelett et al., 2011). We have previously shown that exocyst is present on WPBs and its depletion augments vWF release, but the precise role of exocyst in WPB exocytosis and regulation of its function remain unclear (Sharda et al., 2020).

This study focused on the control of exocyst-dependent WPB exocytosis by Ral GTPases. We find that RalB but not RalA is necessary for WPB exocytosis. GTP-loaded RalB triggers WPB tethering to the plasma membrane and vWF release. Notably, we find that RalB interacts in its GDP-bound state with the exocyst in resting endothelium and that protein kinase C (PKC)-dependent RalB phosphorylation enhances its interaction with exocyst. Subsequent uncoupling of exocyst from GTP-loaded RalB in activated endothelium triggers WPB tethering to the plasma membrane, facilitating WPB exocytosis and vWF release. Our results highlight the mechanism by which RalB controls the tethering and ultimate secretion of WPBs and illuminate points of control that may be manipulated for therapeutic gain.

RESULTS

RalB is required for regulated WPB exocytosis

We have previously shown that exocyst regulates WPB exocytosis (Sharda et al., 2020). Exocyst is a major effector of Ral GTPases (Moskalenko et al., 2002; Sugihara et al., 2002; Moskalenko et al., 2003), and several data suggest that Ral GTPases regulate WPB release, but which Ral isoform functions in this process has not been established (de Leeuw et al., 1999; de Leeuw et al., 2001; Rondaij et al., 2008). To evaluate this further, first we used specific siRNAs to deplete RalA and RalB in human umbilical vein endothelial cells (HUVECs) and studied vWF release. We found that depletion of RalB, but not of RalA, significantly impaired WPB exocytosis by both Ca²⁺- (thrombin and histamine) and cAMP-dependent (forskolin) secretagogues (Figure 1, A and B; Supplemental Figure S1A). Simultaneous knockdown (KD) of RalA and RalB had similar effect to that of RalB KD. Neither RalA nor RalB KD affected basal vWF release (Figure 1B). Furthermore, impairment of vWF release by RalB KD was rescued by expression of siRNA-resistant RalB (Figure 1, C–E).

FIGURE 1:

RalB is switched “on” upon thrombin stimulation and regulates WPD exocytosis. (A) HUVECs depleted of RalA or RalB or both, RalA and RalB, using specific siRNAs and compared with control siRNA. Western blots of total cell lysates showing RalA and/or RalB and vWF expression 48 h after transfection, with GAPDH as loading control. (B) Bar graphs showing total vWF antigen in media estimated by ELISA either at 6 h after media change (basal) or 30 min after 1 U/ml thrombin, 100 mM histamine or 20 mM forskolin treatment, 48 h after siRNA transfection (n = 3;**P< 0.01; ****P< 0.0001; P values were determined by two-way ANOVA with Tukey post hoc analysis). (C) Native, siRNA-resistant, and amino acid sequences of the target site of RalB siRNA used in A and D. Mutated codons are indicated in red. (D) HUVECs expressing wildtype or siRNA-resistant FLAG-tagged RalB were transfected with RalB siRNA. Western blots of total cell lysates showing RalB expression 48 h after transfection, with β-actin as loading control. (E) Bar graphs showing total vWF antigen in media as estimated by ELISA 30 min after 1 U/ml thrombin stimulation 48 h after RalB depletion (n = 2; **P< 0.01; P values were determined by two-way ANOVA with Tukey post hoc analysis). (F) Western blots of total cell lysates showing RalA, RalB, and vWF expression in dihydroartemisinin (DHA)- or DMSO-treated cells (indicated concentrations for 2 h), with β-actin as loading control. (G) Bar graphs showing total vWF antigen in media as estimated by ELISA 30 min after 1 U/ml thrombin stimulation in DHA-treated cells as compared with controls (n = 3; **P< 0.01, ***P< 0.001; ****P< 0.0001; P values were determined by one-way ANOVA with Dunnett post hoc analysis). (H) Total cell lysates of resting and thrombin-treated (1 U/ml for 2 min) HUVECs were incubated with RalBP1-RBD conjugated agarose slurry and eluates immunoblotted with anti-RalB antibody to determine the total GTP-loaded RalB in each condition. RalB expression in lysates (5% of input) and stain free image of RalBP1-RBD shown as loading control. (I) Bar graphs showing densitometric analysis of RalB-RBD pulldown assays (n = 3; **P< 0.01; P value was determined by unpaired t test).

We further confirmed the regulatory role of RalB in evoked WPB exocytosis by using a small-molecule inhibitor, the antimalarial drug dihydroartemisinin, which has been previously shown to specifically degrade cellular RalB but not RalA (Shi et al., 2020). Treatment of HUVECs with dihydroartemisinin for 2 h had no effect on RalA levels but selectively degraded RalB in a dose-dependent manner, without affecting vWF expression (Figure 1F; Supplemental Figure S1B) (Dong et al., 2021). This was associated with a significant impairment in thrombin-stimulated WPB exocytosis (Figure 1G). These data suggest that RalB, but not RalA, is necessary for WPB exocytosis and vWF release.

Next, we assessed the activation status of RalB (fraction of RalB bound to GTP) in resting versus stimulated HUVECs. Secretory cell stimulation results in release of RalGDS, a Ral GEF, from beta-arrestin and consequently RalGDS catalyzes GTP exchange in Ral, thereby switching Ral “on” (Apken and Oeckinghaus, 2021). Using a functional assay that depends on binding of RalB-GTP to the Ral-binding domain (RBD) of its specific effector Ral-binding protein 1 (RalBP1) (Cantor et al., 1995; Jullien-Flores et al., 1995; Park and Weinberg, 1995; Wolthuis et al., 1998), we estimated the fraction of RalB bound to GTP in resting HUVECs compared with HUVECS stimulated with thrombin for 2 min, a time previously shown to maximize Ral activation (Hofer et al., 1998). Our data show that that GTP loading of RalB is significantly increased by thrombin stimulation of endothelial cells (Figure 1, H and I). This result implies that endothelial stimulation switches “on” RalB that subsequently facilitates WPB exocytosis.

RalB associates with mature WPBs

Ral GTPases have been previously reported to localize to WPBs (de Leeuw et al., 1999). To determine whether RalB specifically localizes to WPBs, we performed confocal immunofluorescence microscopy (IFM) with anti-RalB and anti-vWF antibodies (to detect WPBs) on fixed permeabilized HUVECs. WPBs appeared as cigar-shaped vWF-positive structures with broad cytoplasmic distribution; RalB colocalized with vWF on these WPBs (index of correlation 0.73, SD 0.045) but was also present on the plasma membrane and cytoplasm (Figure 2A). Labeling for RalB also overlapped with other markers of mature WPBs, including P-selectin and Angiopoietin-2 (Supplemental Figure S2A). Furthermore, RalA depletion did not affect RalB expression or localization (Figure 2C). These data indicate that a substantial cohort of RalB is present on WPBs in resting HUVECs. Anti-RalA antibody did not perform well enough to convincingly assess RalA labeling by IFM in HUVECs.

FIGURE 2:

RalB is present on mature WPBs. (A–C) IF analysis of HUVECs with anti-vWF and anti-RalB antibodies 48 h after transfection with RalB or RalA siRNA and compared with control siRNA. (A) Micrograph showing presence of RalB (red) in WPBs labeled with vWF (green) in control HUVECs. DAPI (blue) showing nuclei. Inset depicting magnified image of a single WPB. Scale bar, 15 µm. (B) Micrograph showing unaffected distribution of WPBs (vWF; green) in RalB-depleted (red) HUVECs. Scale bar 15 µm. (C) Micrograph showing presence of RalB (red) in WPBs labeled with vWF (green) in RalA-depleted HUVECs. DAPI (blue) showing nuclei. Inset depicting magnified image of WPBs. Scale bar, 15 µm. (D) Bar graphs showing number of WPBs per cell (left), distribution of WPBs per cell (middle), and WPB size (right) 48 h after transfection with RalB siRNA compared with controls (ns = not significant; P value was determined by unpaired t test). (E) IF performed with anti-vWF and anti-Rab27a antibodies 48 h after transfection with RalB siRNA. Micrograph showing presence of Rab27A (red) in WPBs labeled with vWF (green). DAPI (blue) showing nuclei. Inset showing magnified image of WPBs. Scale bar, 15 µm.

Depletion of many components of WPB trafficking machinery can impair WPB maturation (Sharda et al., 2020). To test whether RalB also play a role in WPB maturation, we analyzed RalB siRNA-treated HUVECs by IFM for vWF. RalB depletion did not affect the general distribution or number of WPBs as compared with control siRNA (Figure 2, B and D). Mature WPBs in RalB depleted cells showed a typical cytoplasmic distribution and were positive for Rab27a, a marker of WPB maturity (Figure 2E). This implies that RalB does not participate in WPB biogenesis. Although it remains unclear at what point during maturation do WPB acquire RalB, given its functional role only in WPB exocytosis but not WPB biogenesis, RalB is likely recruited to mature WPBs, similar to Rab27a (Hannah et al., 2003).

Expression of constitutively active RalB targets WPBs to the plasma membrane and augments vWF release

GTP-loaded Ras-like GTPases are inactivated through the activity of specific GAPs that enhance the intrinsic rate of GTP hydrolysis. RalGAP is the GAP specific to Ral GTPases (Emkey et al., 1991; Gridley et al., 2006; Shirakawa et al., 2009; Chen et al., 2011b). The glycine to valine mutation at the 23^rd residue of both Ral isoforms (Ral-G23V) eliminates RalGAP binding (Emkey et al., 1991). Consequently, Ral-G23V exists primarily in the constitutively activated GTP-bound state. To determine how activated RalB affects WPB exocytosis, we expressed RalB-G23V in HUVECs under the control of a Cytomegalovirus (CMV) promoter using lentiviral transduction. HUVECs expressing RalB-G23V demonstrated significant augmentation of vWF release as compared with HUVECs expressing RalB-WT even in the absence of thrombin stimulation (Figure 3A). Results of a RalBP1-RBD pulldown assay confirmed that RalB-G23V was largely GTP-bound (Supplemental Figure S3A). These data suggest that RalB activation stimulates WPB exocytosis and vWF release. On the other hand, expression of RalA-WT or constitutively active RalA-G23V did not affect vWF release (Supplemental Figure S3B).

FIGURE 3:

RalB activation tethers WPBs to the plasma membrane. (A) Expression of autoactivated RalB (RalB-G23V) in HUVECs augments vWF release as compared with RalB-WT. vWF release from nontransduced HUVECs shown as controls. Bar graphs showing total vWF antigen in media estimated by ELISA from resting (90 min after media change) and thrombin stimulated HUVECs (1 U/ml thrombin for 30 min) (n = 3; **P< 0.01 or ****P< 0.0001; P values were determined by two-way ANOVA with Sidak-Holm post hoc analysis). (B) CD63-EGFP expressing HUVECs transduced with lentiviral particles containing RalB-WT or RalB-G23V. WPBs appear as cigar-shaped structures (white arrowheads) distributed among round CD63-positive endolysosomes. WPBs appear tethered to the plasma membrane in RalB-G23V cells as compared with cytoplasmic distribution in RalB-WT cells. Scale bar, 10 µm. (C–D) IF analysis of RalB-G23V and RalB-WT expressing HUVECs labeled with anti-vWF antibody. (C) Bar graph showing cells expressing >15 WPBs per cell as a percentage of total in RalB-G23V expressing HUVECs compared with RalB-WT–expressing HUVECs. A total of 100 random cells counted from each condition from three separate experiments (**P< 0.01; P value was determined by unpaired t test). (D) Bar graph showing WPB distribution in RalB-G23V expressing HUVECs as compared with RalB-WT. The percentage of WPBs present outside 30-µm circle was estimated from 30 randomly selected cells from both conditions from two separate experiments (****P< 0.0001; P value was determined by unpaired t test).

To validate and extend this result, we expressed constitutively active RalB-G23V in HUVECs expressing CD63-EGFP, which labels mature WPBs as well as late endosomes and lysosomes, and performed live cell microscopy to assess WPB trafficking within cells (Sharda et al., 2020). Relative to RalB-WT, RalB-G23V–expressing cells had fewer number of WPBs per cell (Figure 3, B and C).

Notably, WPBs in RalB-G23V cells were found predominantly at the cell periphery, as compared with the broad cytoplasmic distribution in RalB-WT HUVECs (Figure 3D). Many WPBs in RalB-G23V–expressing HUVECs congregated close to the plasma membrane in groups of two to three, reminiscent of compound WPB fusion (Schillemans et al., 2019). Circular CD63-positive endolysosomes formed bulk of the signal and many appeared to be larger in RalB-G23V cells. These enlarged CD63-positive structures were identified as autophagosomes by electron microscopy (EM) (data not shown), which RalB activation is also known to promote (Bodemann et al., 2011). These data suggest that the active form of RalB directly facilitates WPB tethering and fusion.

EXOC2 associates with RalB-GDP in resting endothelial cells via a defined RBD

Given that exocyst is a major Ral effector, we next determined whether RalB-dependent WPB tethering is mediated by exocyst. First, we compared the localization of RalB and exocyst, labeled with an antibody to EXOC7, on WPBs by confocal IFM. The data confirmed that both RalB and exocyst can be detected on the majority of vWF-labeled WPBs in paraformaldehyde fixed and permeabilized HUVECs (index of correlation 0.70, SD 0.07) (Figure 4A). Additionally, we performed immunoelectron microscopy to study the localization of RalB and exocyst ultrastructurally. Because our anti-RalB antibody demonstrated poor labeling of glutaraldehyde-fixed sections used for immunogold labeling, we expressed FLAG-tagged RalB-WT in HUVECs. Transmission electron microscopy (TEM) analysis of sections that were double immunogold labeled for FLAG and EXOC4 confirmed the presence of RalB and exocyst on mature WPB cigars (Figure 4B). Interestingly, RalB and EXOC4 labeling was not uniform but seemed to be localized on a few spots on WPB surface. These could potentially represent the points of anchoring to target membranes.

FIGURE 4:

Exocyst interacts with GDP-loaded RalB. (A) Confocal IF micrographs showing presence of RalB (purple) and EXOC7 (red) with vWF (green) in mature WPBs. Insets depict magnified image of the selected area showing mature WPBs. DAPI (blue) showing nucleus. Scale bar, 10 µm. (B) TEM of HUVECs with double immunogold labeling with anti-EXOC4 (15 nm) and FLAG (10 nm) antibodies. Transmission electron micrograph showing both EXOC4 and RalB decorating mature electron-dense WPBs (white asterisks). Inset showing magnified image of the selected area. Scale bar, 100 nm. (C) Western blots showing IP of resting and thrombin stimulated (1 U/ml for 2 min) total HUVEC lysates with anti-RalB antibody, resolved by SDS–PAGE and immunoblotted with anti-RalB, anti-EXOC2, anti-EXOC6, or anti-EXOC8 antibodies. Total HUVEC lysate (2.5% of IP) used for input. (D) Western blots showing IP of resting and thrombin stimulated (1 U/ml for 2 min) total HUVEC lysates with anti-EXOC4 antibody, resolved by SDS–PAGE and immunoblotted with anti-RalB or anti-EXOC2 antibody. Total HUVEC lysate (5% of IP) used for input. (E) Total cell lysates of RalB-WT, RalB-G23V and RalB-S28N expressing HUVECs incubated with magnetic beads conjugated with anti-FLAG antibody and eluates immunoblotted with anti-FLAG or anti-EXOC2 antibodies. Total cell lysates (5% of IP) used for input. (F) Total cell lysates of resting and thrombin-treated (1 U/ml for 2 min) HUVECs incubated with EXOC2-RBD conjugated agarose slurry and eluates immunoblotted with anti-RalB antibody. Ponceau S staining showing EXOC2-RBD as loading control.

To confirm that RalB physically interacts with exocyst, we first performed immunoprecipitation (IP) experiments on total cell lysates of native HUVECs that were either resting or thrombin-stimulated, to enhance the levels of activated GTP-bound RalB. RalA and B specifically interact with exocyst subunits EXOC2 (sec5) and EXOC8 (Exo84), and interaction of EXOC2 and EXOC8 with RalB can be competitive (Moskalenko et al., 2002; Sugihara et al., 2002; Moskalenko et al., 2003; Jin et al., 2005). RalB was immunoprecipitated using a specific antibody bound to dynabeads, and then eluates resolved by SDS–PAGE and immunoblotted for RalB and the exocyst subunits EXOC2 and EXOC8. RalB, as well as exocyst subunits EXOC2 and EXOC8 were coimmunoprecipitated under both conditions, but, surprisingly, significantly more from resting lysates as compared with thrombin-stimulated HUVEC lysates (Figure 4C; Supplemental Figure S4A). We also immunoblotted for EXOC6 (sec15), another component of exocyst subcomplex II, which also coimmunoprecipitated predominantly with RalB in resting HUVEC lysates (Wu and Guo, 2015). This suggests that RalB interacts with and controls the function of the exocyst holo-complex in HUVECs. We also carried out reverse IP experiments with exocyst. Here we were limited by the performance of EXOC2 and EXOC8 antibodies that we used for immunoblotting above, but EXOC4 antibody worked well. Similar to our results presented above, significantly more RalB coimmunoprecipitated with EXOC4 from resting total HUVEC lysates (Figure 4D; Supplemental Figure S4B).

To corroborate these results, we expressed FLAG-tagged wild-type (RalB-WT), constitutively active (RalB-G23V) and dominant negative (RalB-S28N) RalB in HUVECs, immunoprecipitated RalB with anti-FLAG antibodies, and immunoblotted the eluates for RalB and EXOC2. Despite similar expression levels of the FLAG-RalB variants in the lysates, EXOC2 was detected at much higher levels in the anti-FLAG immunoprecipitates from cells expressing RalB-WT and RalB S28N than from cells expressing RalB-G23V (Figure 4E; Supplemental Figure S4C). We also sent eluates from FLAG pulldown experiments for mass spectrometry analysis. EXOC2 was identified as a RalB interacting protein, but the peptide counts were significantly higher in RalB-WT as compared with RalB-G23V eluates (Supplemental File S1). These data support our conclusion that exocyst binds more avidly to RalB-GDP than to RalB-GTP in HUVECs. Moreover, RalB binds to the exocyst holo-complex in these cells.

Previously, exocyst was shown to bind preferentially to GTP-loaded RalA in solution via the N-terminal 99 residues of the EXOC2 subunit (EXOC2-RBD) (Fukai et al., 2003; Mott et al., 2003). Notably, RalA was truncated to remove its C-terminal hypervariable region (HVR) in these previous reports, but it has since become clear that Ral HVR, which spans residues 177-202, imparts critical functional differences between RalA and RalB and modulates Ral activity, localization and effector interactions (Wu et al., 2005; Sablina et al., 2007; Wang et al., 2010; Kashatus et al., 2011; Martin et al., 2012). To further determine exocyst–RalB interaction in HUVECs, we studied binding of native RalB in HUVEC lysates with EXOC2-RBD. For this, we purified GST-tagged EXOC2-RBD from E. coli (Supplemental Figure S5A) and tested whether it could pull down RalB from lysates of resting and thrombin-stimulated HUVEC lysates. As observed for intact native exocyst complex in total cell lysates, EXOC2-RBD precipitated significantly more RalB from lysates of resting HUVECs than from thrombin-stimulated cells (Figure 4F; Supplemental Figure S4D). This confirms that EXOC2 preferentially binds GDP-loaded RalB in HUVEC lysates. Treatment of lysates with GDP or the nonhydrolysable GTP analogue, GTPgS, did not reduce interaction of EXOC2 with RalB in resting endothelial lysates (Supplemental Figure S5B). This suggests that either other regulatory factors, such as RalB posttranslational modifications, likely modulate the RalB–exocyst interaction or that localized GEF and GAP activity is necessary to modulate RalB guanine nucleotide loading.

PKC-dependent RalB phosphorylation modulates exocyst binding

Posttranslational modifications of Ras and Ral, such as phosphorylation and ubiquitination, regulate GDP/GTP-loading, effector binding and localization (Bivona et al., 2006; Sablina et al., 2007; Wang et al., 2010; Martin et al., 2012; Neyraud et al., 2012; Simicek et al., 2013). Phosphorylation of RalA at S194 was initially shown to modulate effector interaction and localization but also RalA activity (Wu et al., 2005; Sablina et al., 2007; Kashatus et al., 2011). Subsequently, HVR of RalB was also shown to be phosphorylated. In particular, PKC-dependent phosphorylation of serine residues at positions 192, and more so 198, in the RalB HVR were shown to increase RalB-GTP loading as well as binding to RalBP1 and exocyst in cancer cells (Wang et al., 2010; Martin et al., 2012). Thrombin activates multiple PKC isoforms and phosphatases in endothelial cells (Tan et al., 2003; Li et al., 2004; Xie et al., 2016). Therefore, we tested whether thrombin- and PKC-dependent RalB phosphorylation might also be modulating the unexpected binding of EXOC2 with RalB-GDP in HUVECs. We first tested the effect of pharmacologic PKC modulation on RalB phosphorylation and exocyst binding. HUVECs were treated with either bryostatin-1, a PKC activator, or sotrastaurin, a pan-PKC inhibitor, as described previously (Martin et al., 2012; Hung et al., 2015). RalB immunoprecipitated from these total cell lysates were resolved by SDS–PAGE and blotted with anti-phosphoserine, anti-RalB, or anti-EXOC2 antibodies. Our data show that in resting HUVECs, a band that comigrates with RalB is serine-phosphorylated, but that the phosphorylated band is dramatically reduced by thrombin stimulation (Figure 5, A and B). EXOC2 coprecipitation correlated with the intensity of the phosphorylated band. Treatment with the PKC inhibitor sotrastaurin reduced the phosphorylated band and EXOC2 coprecipitation in resting cells, whereas treatment with the PKC activator bryostatin-1 increased the phosphorylated band and enhanced exocyst binding, even in thrombin-stimulated cells (Figure 5, A and B). We interpret these data to conclude that PKC-mediated phosphorylation of RalB in resting cells promotes exocyst binding, whereas thrombin-stimulated dephosphorylation antagonizes exocyst binding and can be reversed by constitutive PKC activity.

FIGURE 5:

PKC-dependent RalB phosphorylation modulates exocyst binding and WPB exocytosis. (A) Thrombin stimulation dephosphorylates RalB and reduces EXOC2 binding. IP of total cell lysates of resting and thrombin stimulated HUVECs (1 U/ml for 2 min) with and without PKC activator bryostatin-1 (100 nM for 10 min) or PKC inhibitor sotrastaurin (500 nM for 1 h) performed with anti-RalB antibody, resolved by SDS–PAGE and immunoblotted with anti-phosphoserine, anti-RalB and anti-EXOC2 antibodies. Total HUVEC lysate (5% of IP) used as input. (B) Densitometric quantification of immunoprecipitated EXOC2 and phospho-RalB in A (normalized to resting untreated lysates) (n = 3 [except two experiments each for thrombin plus bryostatin-1 or sotrastaurin treated cells]; *P< 0.05; **P< 0.01; ***P< 0.001; P values were determined by one-way ANOVA with Dunnett post hoc analysis). (C) Total cell lysates of RalB-WT, RalB-S198D, and RalB-S198A expressing HUVECs incubated with magnetic beads conjugated with anti-FLAG antibody and eluates immunoblotted with anti-Flag and anti-EXOC2 antibodies. Total HUVEC lysate (5% of IP) used as input. (D) Densitometric quantification of EXOC2 coimmunoprecipitated with RalB in C, normalized to EXOC2 levels in RalB-WT (n = 3; ****P< 0.0001; P values were determined by one-way ANOVA with Dunnett post hoc analysis). (E) Bar graphs showing total vWF antigen in media estimated by ELISA collected from resting HUVECs (after 90 min of media change) expressing RalB-WT, RalB-S198D, and RalB-S198A (n = 3; ****P< 0.0001; P values were determined by one-way ANOVA with Bonferroni post hoc analysis).

RalB phosphorylation at S198, and to a lesser extent S192, in the C-terminal HVR have been shown to modulate RalB function (Wang et al., 2010; Martin et al., 2012). To test whether serine phosphorylation at S198 modulates exocyst binding and WPB exocytosis, we used site-directed mutagenesis to convert S198 to non-phosphorylatable alanine (S198A) or the phosphomimetic aspartic acid (S198D) in FLAG-tagged RalB-WT and expressed them in HUVECs. By FLAG IP from cell lysates, RalB-S198D enhanced exocyst binding relative to RalB-WT, while RalB-S198A decreased exocyst binding (Figure 5, C and D), consistent with resting and thrombin-stimulated HUVECs as well as with the effects of activating or inhibiting PKC (Figure 5A). Analyses of these cells for vWF release showed that the effect of these mutations at S198 on WPB exocytosis reflected the status of RalB-exocyst binding. Relative to cells expressing RalB-WT, cells expressing RalB-S198D had impaired WPB exocytosis, whereas cells expressing RalB-S198A showed significantly augmented WPB exocytosis and vWF release (Figure 5E). Combined, these data suggest PKC-dependent posttranslational phosphorylation of RalB enhances its interaction with EXOC2, impeding downstream WPB release.

Uncoupling exocyst from RalB triggers WPB tethering and vWF release

We interpret these data to suggest that the uncoupling of exocyst from GTP-loaded RalB in stimulated endothelial cells triggers WPB tethering and exocytosis. To evaluate this model further, we exploited a mutation, D49E, previously shown to reduce the Ral–exocyst interaction (Moskalenko et al., 2002). We generated the D49E mutation in FLAG-tagged forms of both RalB-WT and RalB-G23V, and expressed them in HUVECs by lentiviral transduction. By coimmunoprecipitation of cell lysates with anti-FLAG antibodies, the RalB-D49E mutation impaired EXOC2 binding for both WT and G23V variants, as expected (Figure 6A). In particular, low levels of EXOC2 were coprecipitated with RalB-G23V, and the addition of the D49E mutation essentially eliminated this binding. To test how RalB-exocyst binding affected WPB exocytosis, we assessed vWF release in these cells by estimating vWF antigen levels in culture media. Consistent with our hypothesis, cells expressing the D49E mutation showed significantly augmented basal vWF release as compared with cells expressing RalB-WT (Figure 6B). By contrast, cells expressing RalB with both the D49E and G23V mutations did not further augment vWF release beyond that of cells expressing the G23V mutation alone. These data support the conclusion that poor binding of RalB to exocyst supports enhanced WPB secretion.

FIGURE 6:

RalB uncoupled exocyst triggers WPB exocytosis. (A) Total cell lysates of RalB-WT and RalB-G23V and their respective D49E mutants incubated with magnetic beads conjugated with anti-FLAG antibody and eluates immunoblotted with anti-FLAG and EXOC2 antibodies. Total HUVEC lysate (2.5% of IP) used as input. (B) Bar graphs showing total vWF antigen in media from resting HUVECs (after 90 min of media change) expressing RalB-WT, RalB-WT D49E, RalB-G23V or RalB-G23V D49E (n = 3; *P< 0.01; P values were determined by one-way ANOVA with Tukey post hoc analysis). (C) Total cell lysates of HUVECs expressing RalB-WT and RalB-WT D49E, RalB-WT E38R or RalB-WT A48W incubated with magnetic beads conjugated with anti-FLAG antibody and eluates immunoblotted with anti-FLAG, EXOC2, and EXOC8 antibodies. Total HUVEC lysate (5% of IP) used as input. (D) Densitometric quantification of EXOC2 and EXOC8 coimmunoprecipitated with FLAG-tagged RalB in C, normalized to EXOC2 or EXOC8 levels in RalB-WT (n = 3; ****P< 0.0001; P values were determined by one-way ANOVA with Dunnett post hoc analysis). (E) Bar graphs showing total vWF antigen in media from resting HUVECs (after 90 min of media change) expressing RalB-WT, RalB-WT D49E, RalB-WT E38R or RalB-WT A48W (n = 3; ***P< 0.001; ****P< 0.0001; P values were determined by one-way ANOVA with Dunnett post hoc analysis). (F) IF analysis of RalB-WT and RalB-WT D49E expressing HUVECs labeled with anti-vWF antibody and anti-CD31 antibody. Micrographs showing WPBs labeled with vWF (red) found predominantly at the plasma membrane in RalB-D49E mutants as compared with cytoplasmic distribution in RalB-WT. DAPI (blue) depicting nuclei. Scale bar, 10 µm. (G) Bar graph comparing % cells with >15 WPB in RalB-WT compared with RalB-D49E expressing HUVECs (**** P< 0.0001; P value was determined by unpaired t test). (H) Bar graph showing WPB distribution in RalB-WT D49E expressing HUVECs as compared with RalB-WT. The percentage of WPBs present outside 30-µm circle was estimated from 30 randomly selected cells from both conditions from two separate experiments (****P< 0.0001; P value was determined by unpaired t test).

To further characterize the role of RalB–EXOC2 and RalB–EXOC8 interactions separately in WPB exocytosis, we generated partial loss-of-function variants RalB-E38R and RalB-A48W that have been shown to selectively reduce EXOC2 and EXOC8 interactions, respectively (Jin et al., 2005; Cascone et al., 2008). FLAG-tagged RalB-WT E38R and RalB-WT A48W mutants were expressed in HUVECs by lentiviral transduction and IP studies carried out on cell lysates. By coimmunoprecipitation of cell lysates with anti-FLAG antibodies, the RalB-E38R mutation not only impaired EXOC2 binding, but also EXOC8 binding (Figure 6, C and D). Similarly, RalB-A48W mutation impaired both EXOC8 and EXOC2 binding, again suggesting that RalB interacts with the exocyst holo-complex required for WPB exocytosis. Furthermore, these mutations significantly augmented vWF release, as seen with RalB-D49E mutation (Figure 6E).

Last, we assessed WPB localization in cells expressing RalB-WT or RalB-WT-D49E cells by IFM with antibodies to the WPB component vWF. As compared with the typical broad cytoplasmic distribution of mature WPBs in RalB-WT–expressing cells, WPBs in the cells expressing RalB-WT-D49E had tendency to be in the periphery, at the plasma membrane like RalB-G23V–expressing HUVECs (Figure 6, F–H). These data confirm that RalB uncoupling from exocyst triggers WPB tethering and facilitates exocytosis.

DISCUSSION

Acting as molecular switches, Ral small GTPases integrate extracellular stimuli with downstream cellular processes, including vesicular trafficking, through specific effectors. In this work, we have made two interdependent but equally important observations. First, we provide evidence that Ral isoform RalB positively regulates WPB exocytosis directly through its binding partner exocyst complex. But more notably, we demonstrate that exocyst can interact with GDP-bound RalB, as in resting endothelium. Once switched “on,” exocyst disengages from GTP-bound RalB, and RalB uncoupling from exocyst promotes WPB exocytosis and vWF release. Moreover, this dynamic association of RalB with exocyst is modulated by posttranslational phosphorylation of RalB (Figure 7).

FIGURE 7:

Proposed model for RalB-exocyst dependent WPB tethering. (A) In resting HUVECs, RalB is GDP-loaded as well as phosphorylated by protein kinase C in its C-terminal HVR. Upon endothelial cell stimulation by thrombin, RalB becomes GTP-loaded catalyzed by GEF, as well as dephosphorylated by activation of phosphatases. Consequently, exocyst is uncoupled from dephosphorylated and GTP-bound RalB, and RalB uncoupling from exocyst promotes WPB tethering to the plasma membrane and vWF release. (B) RalB mutations which render RalB autoactivated (G23V; top), phosphodeficient (S198A; middle) or uncoupled from exocyst (D49E; bottom), trigger WPB tethering and vWF release in the absence of endothelial cells stimulation. (GEF, guanine nucleotide exchange factor; GAP, GTPase-activating protein; P, phosphate; PKC, protein kinase C; PAR, protease activated receptor).

The paradigm for small GTPases as molecular switches is their dual-mode conformation—the GTP-loaded “on” state, where small GTPases bind their effectors, triggering downstream functions, and GDP-loaded “off” state. Our data support a role for activated GTP-bound RalB triggering endothelial WPB exocytosis, but contrary to the accepted paradigm, by disengaging with its binding partner exocyst. There are other sporadic examples of GDP-loaded GTPase interacting with its binding partner. For example, Slp4a preferentially interacts with GDP-bound Rab27a to inhibit dense-core vesicle exocytosis in PC12 cells (Fukuda, 2003). The effector binding occurs at the switch domains that is conserved in all small GTPases including Ral (Wennerberg et al., 2005). The conformational change that occurs with GTP/GDP cycling acts as a sensor and allows for specific interaction with effectors, typically with GTP-bound forms of GTPases (Vetter and Wittinghofer, 2001). The binding site of EXOC2 is located in the switch I region where as that of EXOC8 is located on both switch I and II regions of Ral, and previously, RalA was reported to bind and control exocyst function in insulin secretion as well as GLUT4 trafficking when GTP-loaded (Chen et al., 2007; Lopez et al., 2008). In contrast to these studies, our data clearly show that the exocyst interacts with GDP-bound RalB and functions as a holo-complex in resting endothelial cells and disengage from GTP-bound RalB following endothelial cell stimulation and intracellular calcium mobilization. We confirmed this by showing interaction of native RalB with purified EXOC2 Ral binding domain preferentially from resting lysates as compared with stimulated lysates. Furthermore, treatment with either GDP or nonhydrolysable GTP analogue GTPgS did not alter this dynamic interaction.

There is increasing understanding of how posttranslational modifications, particularly phosphorylation, regulate effector selection and affinity for many known Rabs but also Ras and Ral (Bivona et al., 2006; Sablina et al., 2007; Wang et al., 2010; Martin et al., 2012; Xu et al., 2021; Pfeffer, 2023). Phosphorylation has been shown to regulate both RalA and RalB localization, as well as activity in a cell- and function-specific manner, in addition to ubiquitination in the case of RalB (Wu et al., 2005; Wang et al., 2010; Martin et al., 2012; Simicek et al., 2013). Our data also show that PKC-dependent phosphorylation of C-terminal HVR of RalB promotes exocyst binding in resting HUVECs, and RalB dephosphorylation that occurs in thrombin-stimulated HUVECs reduces exocyst binding, increasing vWF release. Correspondingly, modifications of S198 in RalB HVR affects exocyst binding and WPB exocytosis; the phosphodeficient S198A mutation reduces exocyst binding augmenting WPB exocytosis, while the phosphomimetic S198D mutation enhances exocyst binding and impairs WPB exocytosis. Our studies in native HUVEC lysates confirm these findings. RalB is phosphorylated in resting state, when it binds exocyst; thrombin stimulation dephosphorylates RalB, reducing exocyst binding. While RalB phosphorylation can affect GTP loading, posttranslational phosphorylation has also been shown to regulate RalB localization. Therefore, the reduced binding of EXOC2 to dephosphorylated RalB that we notice could be secondary to the relative abundance of exocyst in the subcellular compartment where RalB relocalizes (Wang et al., 2010; Martin et al., 2012). Structural determinants of Ral–EXOC2 and Ral–EXOC8 interaction have been described previously (Fukai et al., 2003; Mott et al., 2003; Jin et al., 2005). These studies showed that a truncated version of RalA binds these exocyst subunits when GTP-loaded, but notably, these studies used a RalA with its C-terminal HVR removed. This would eliminate the phosphorylation sites in both RalA and RalB that are critical in regulating Ral function allosterically.

Classic PKC isoform PKC-α was previously shown to phosphorylate S192 and S198 residues in RalB HVR in a colorectal cancer cell line (Martin et al., 2012). PKC-α is expressed in HUVECs and therefore may be involved in RalB phosphorylation in endothelial cells (Lorenzi et al., 2008). The effect of PKC inhibition on evoked vWF release has also been previously characterized but the data remain inconclusive. For example, the effect of staurosporine, a pan-PKC inhibitor, on histamine-induced vWF release has been variable, although specific PKC-α inhibition or depletion has been shown to augment histamine-induced vWF release (Carew et al., 1992; Lorenzi et al., 2008). The effect on thrombin-induced vWF release has been dependent on the inhibitor employed (Carew et al., 1992; Lorenzi et al., 2008). These differences could be due to other protein targets of PKC, proteins that also participate in WPB exocytosis and vWF release. Although, protein phosphatase 2Ab has been shown to dephosphorylate RalA after PKA-mediated RalA phosphorylation, the identity of phosphatase/s that may be involved in RalB dephosphorylation remain unknown (Sablina et al., 2007). Protein phosphatase 2B was previously shown to regulate munc18c phosphorylation in endothelial cells, but whether RalB could also be a substrate of this phosphatase is unknown (Nolasco et al., 2009).

In addition to Ral phosphorylation modulating exocyst interaction, posttranslational modifications of exocyst have also been shown to affect its interaction with RalA, which we have not evaluated in this study (Chen et al., 2011a). Ral has been previously shown to be required for assembly of exocyst on secretory granules (Moskalenko et al., 2003). Given that RalB depletion impairs endothelial WPB exocytosis, it is likely that RalB is required for exocyst assembly on WPBs, but the precise mechanism remains unclear. Also uncertain is the interplay of other small GTPases that are present on mature WPBs, particularly Rab27a that is also known to bind and regulate exocyst (Fukuda, 2003; Bierings et al., 2012; Fukuda, 2013; Holthenrich et al., 2019; El-Mansi et al., 2023). Last, the role of RalBP1, a Ral effector that acts as a GAP for Rho GTPases which regulate actin dynamics, remains to be determined in WPB exocytosis (Apken and Oeckinghaus, 2021).

Pharmacologic modulation of endogenous vWF release would be beneficial in treatment of von Willebrand disease but also in other pathophysiologic conditions where vWF plays essential roles, but the molecular regulation of WPB secretion remains elusive. We propose that exocyst uncoupling from RalB is required for endothelial WPB exocytosis and vWF release, and that RalB phosphorylation is a major determinant of RalB–exocyst interaction. Given high degree of homology and the shared set of effectors between RalA and RalB, posttranslational modifications may confer distinct functionality to these GTPases.

MATERIALS AND METHODS

Request a protocol through Bio-protocol

Reagents

Fatty acid–free BSA, Tween-20, 2-mercaptoethanol, triton X-100, methanol, hydrochloric acid, 0.1% gelatin solution, and bryostatin-1 were obtained from Millipore Sigma. DynaBeads Protein G, MaxiSorp immuno plate, 1-Step Ultra TMB-ELISA, 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI), Restore Western Blot Stripping Buffer, puromycin dihydrochloride, bacitracin, and lipofectamine RNAiMax as well as 3000 were obtained from Invitrogen.

Laemmli sample buffer, 4–15% TGX SDS-polyacrylamide gels were obtained from Bio-Rad. EBM-2 basal medium, HUVECs, and HUVECs bullet kits were obtained from Lonza. Human α-thrombin was obtained from Haematologic Technologies. Blotto nonfat dry milk was obtained from ChemCruz. Calcium- and magnesium-free PBS, pH 7.4, Opti-MEM reduced serum media, DMEM, FBS, and penicillin/streptomycin were obtained from Life Technologies. Streptavidin-horseradish peroxidase (HRP) was obtained from Abcam. Sotrastaurin was obtained from Selleckchem.

Antibodies

The primary antibodies are listed in the table below.

Antigen	Host	Vendor
RalA	Rabbit monoclonal	Abcam (ab12667)
RalB	Rabbit monoclonal	Abcam (ab129077)
EXOC2	Rabbit polyclonal	Proteintech (12751-1-AP)
EXOC4	Mouse monoclonal	MilliporeSigma (MABC570)
EXOC6	Rabbit polyclonal	Proteintech (12712-1-AP)
EXOC7	Mouse monoclonal	Kerafast (ED2001)
EXOC8	Rabbit polyclonal	Novus (NBP1-91444)
vWF	Rabbit polyclonal	Dako
vWF	Mouse monoclonal	Thermo Fisher Scientific (MA5-14029)
FLAG	Mouse monoclonal	SigmaMillipore (F1804)
P-selectin	Mouse monoclonal	Thermo Fisher Scientific (14-0628-82)
Phosphoserine	Rabbit polyclonal	SigmaMillipore (AB1603)
CD31	Rabbit recombinant multiclonal	Abcam (ab281583)
Angiopoietin	Rabbit polyclonal	Proteintech (24613-1-AP)
β-Actin	Mouse monoclonal	Thermo Fisher Scientific (MA1-140)

Species-specific HRP-conjugated secondary antibodies as well as Verablot secondary for IP experiments were obtained from Abcam. Species-specific Alexa Fluor 488, 568, and 647 antibodies were obtained from Invitrogen.

Cell culture

HUVECs were cultured in EBM-2 basal media supplemented with bullet kit containing endothelial growth supplements (complete media) in gelatin-coated tissue culture–treated flasks or plates in 37°C incubator with 5% CO₂. Only passage 1–5 were used for experiments.

Plasmids, RNAi, transfection, and transduction

RalA (s11758) and RalB (s11761 and s11762) siRNA were purchased from Life Technologies. For siRNA-transfection, HUVECs were grown to confluency on 6- or 12-well plates in complete media. siRNAs were prepared by mixing 10 µM of negative control, RalB, or RalA siRNA with lipofectamine RNAiMax transfection reagent in OptiMEM low-serum media for 10 min to a final concentration of 5–10 pM, prior to the addition to the cells. The transfection media was replaced with complete media after 6–8 h of transfection. Experiments were performed on RalB and RalA KD cells 48 h posttransfection.

pLA-CMV-N-Flag-RalBWT (plasmid#50990), pDONR223-RalA-WT-V5, pLX301 (plasmid#25895), pEGFP-C2-CD63 (plasmid#62964) were obtained from Addgene.

To make lentiviruses, first, 12 µg of plasmid was incubated with 25 µl of lipofectamine 3000 together with packaging plasmids (9 µg psPAX2 [Addgene#12260], 3 µg PMD2 [Addgene#12259], 3 µg pRSV-Rev [Addgene#12253]) in 1 ml of Opti-MEM low serum medium for 45 min. Lenti-X-293T cells cultured to 70–80% confluency in 10 cm dishes in 9 ml of growth media were then transfected with this mix. Growth media was replaced after overnight transfection. Lentivirus was then harvested at 48 and 72 h posttransfection and either added directly to HUVECs in culture or concentrated with Lenti-X concentrator (Takara) and stored in −80°C for future use.

For viral transduction, P1 HUVECs were grown to 80–90% confluency and either the growth media was replaced with freshly harvested virus-containing media or with frozen concentrated virus in complete media. After overnight incubation, media was replaced with fresh complete media. Antibiotic selection was performed after 48 h of infection.

Site-directed mutagenesis and TOPO and gateway cloning

Site-directed mutagenesis was carried out either using Q5 site-directed mutagenesis kit (New England Bio) or In-fusion cloning (Takara) according to the manufacturers’ protocols. Primers for site-directed mutagenesis were designed based on the protocol employed.

pEGFP-C2-CD63 and pDONR224-RalA-WT-V5 were cloned into pLX301 vector using pENTR/D-TOPO cloning kit (Life Technologies) followed by gateway cloning per manufacturer's protocol. RBD of EXOC2 was also TOPO cloned into pET16 vector using pENTR/D-TOPO cloning kit.

vWF ELISA

HUVECs were grown on 6- or 12-well tissue culture plates until confluent. The cells were serum starved for 90 min at 37°C prior to stimulation with 1 U/ml human alpha thrombin for 30 min. The media containing releasates were immediately frozen or further processed for vWF antigen detection. A 96-well flat bottom MaxiSorp immuno plate was coated with 1 µg/ml anti-vWF antibody (Dako rabbit polyclonal) in a sodium carbonate/bicarbonate buffer pH 9.5 for 1 h at room temperature prior to blocking with 3% wt/vol BSA in phosphate-buffered saline with tween (PBS-T) for 30 min. Pooled plasma from healthy human donors containing vWF antigen were used as a standard by serial dilution in PBS. vWF antigen in the diluted releasates was then captured onto the wells for 1 h at room temperature. The antibody–vWF complexes were washed with PBS-T (0.1%) three times followed by incubation with an anti-vWF antibody (Dako rabbit polyclonal) conjugated with biotin for 60 min at room temperature. A total of 100 ng/ml streptavidin-HRP was added in 1% wt/vol BSA in PBS-T for 45 min at room temperature followed by five washes with PBS-T. TMB solution was then added until noticeable development of color. The reaction was stopped with 1N hydrochloric acid and spectrophotometrically read at 450 nm.

Confocal microscopy

Confocal IFM was performed on HUVECs cultured on chambered coverslips, fixed with 4% paraformaldehyde (PFA) for 10 min, permeabilized with 0.25% triton X-100 for 10 min, blocked with 3% wt/vol BSA, incubated with primary antibodies overnight at 4°C, followed by fluorophore-conjugated secondary antibodies for 45 min and counterstained with DAPI for 2 min. Coverslips were mounted using Prolong Gold mounting media and imaged with a LEICA SP8 upright microscope with 63×1.35 normal aperture oil objective. LAS-X software was used to analyze imaging data. ImageJ colocalization colormap plugin was used to quantify correlation.

Live-cell imaging was carried out on cells plated on glass chambered coverslips with a Yokagawa spinning disk confocal on an inverted Nikon Ti fluorescence microscope fitted with OkoLab 37°C, 5%CO₂ cage microscope incubator with 100×1.25 normal aperture oil objective. Images were further processed in ImageJ.

Quantification of WPB number, distribution, and size

For quantification of WPB size and distribution, immunofluorescence (IF) analysis was performed and image acquired at the same plane in each cell. For WPB number, proportion of cells with >15 vWF-positive structures were estimated from 20 random cells each from 8 to 12 individual wells of stained chamber slides. For WPB distribution, a 30-µm circle was placed at the cell nucleus and the percentage of WPBs present inside and outside the circle was determined. A total of 30 random cells from 2 to 3 individual experiments for each condition were selected. WPB size was determined for 150 nonoverlapping, peripherally distributed WPBs, from 15 random cells, obtained from each condition in ImageJ.

Electron microscopy

Immunogold electron microscopy was performed based on the Tokayasu method. Briefly, for preparation of cryosections, the cells were removed from the dish with Trypsin, fixed with 4% paraformaldehyde + 0.1% glutaraldehyde (in 0.1M sodium phosphate buffer, pH 7.4), washed with PBS and then placed in PBS containing 0.2M glycine to quench free aldehyde groups (15 min). Samples were then frozen in liquid nitrogen and sectioned at −120°C and the sections were transferred to formvar/carbon-coated copper grids. The double immunogold labeling was carried out at room temperature on a piece of parafilm. All antibodies and protein A gold (University Medical Center, Utrecht, the Netherlands) were diluted in 1% BSA in PBS. Grids were blocked with 1% BSA for 10 min and incubated with primary antibodies for 30 min. Double immunogold labeling was completed by sequential incubation with the first species-specific secondary antibody followed by 15 nm Protein-A gold, a fixation step in between, and finally the other secondary antibody followed by10 nm Protein-A gold. The grids were examined in a JEOL 1200EX electron microscope and images were recorded with an AMT 2k CCD camera.

IP

Lysates for IP were prepared using IP lysis buffer (Invitrogen) or magnesium lysis buffer (Millipore Sigma). Lysates were incubated with 2.5 µg anti-RalB or EXOC4 antibody-bound protein G Dynabeads or Anti-Flag M2 magnetic beads overnight, washed with PBS-T three times and boiled with reduced Laemmli sample buffer. Eluted immunoprecipitates were either sent for mass spectrometry or resolved with SDS–PAGE followed by immunoblotting.

Abbreviations used:

EM

electron microscopy

IFM

immunofluorescence microscopy

GAP

GTPase activating proteins

GDP

guanosine diphosphate

GEF

guanosine nucleotide exchange factors

GTP

guanosine triphosphate

GTPase

guanosine triphosphatase

HUVEC

human umbilical vein endothelial cells

HVR

Hypervariable region

PKC

protein kinase C

RalA

Ras-like or Ras-related protein-A

RalB

Ras-like or Ras-related protein-B

RalBP1

RalA-binding protein 1

vWF

von Willebrand factor

WPB

Weibel-Palade body.