RalA promotes a direct exocyst–Par6 interaction to regulate polarity in neuronal development

Cell polarization is a key event in neuronal development, controlling neurogenesis (Farkas and Huttner, 2008), neuronal migration and synaptogenesis (Solecki et al., 2006; Valiente and Marín, 2010). Proper localization of polarity determinants, directed vesicle traffic, and regulated cytoskeletal dynamics are required for polarization in many cell types, including neurons and neural progenitors. Small GTPases coordinate these events to establish cell polarity in response to intra- and extracellular cues (Mellman and Nelson, 2008; McCaffrey and Macara, 2009; Hall and Lalli, 2010). Among these, the Ras-like GTPase Ral, of which there are two isoforms, RalA and RalB, has emerged as an important polarity regulator in a variety of contexts, including basolateral membrane trafficking and tight junction formation in epithelial cells (Shipitsin and Feig, 2004; Hazelett et al., 2011), polarized migration of fibroblasts and cancer cells (Rossé et al., 2006; Spiczka and Yeaman, 2008) and tumorigenesis (Camonis and White, 2005; Lim et al., 2005; Oxford et al., 2005; Bodemann and White, 2008; Hazelett and Yeaman, 2012). During brain development, Ral is involved in asymmetric division of neuroblasts (Carmena et al., 2011), neuronal migration in the neocortex (Jossin and Cooper, 2011), neurite branching (Lalli and Hall, 2005) and activity-dependent spine growth (Teodoro et al., 2013). RalA also regulates axon initiation in cortical neurons by promoting an interaction between one of its effectors, the exocyst, and the partitioning-defective (Par) Par3–Par6–aPKC (atypical protein kinase C) complex (Lalli, 2009), an evolutionarily conserved master regulator of polarity (Goldstein and Macara, 2007). Despite their importance, the underlying molecular mechanisms by which RalA exerts its regulatory function are still unclear.

The exocyst is an evolutionarily conserved octameric protein complex consisting of Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70 and Exo84 (TerBush and Novick, 1995; Hsu et al., 1996; TerBush et al., 1996; Guo et al., 1999). Assembly of the exocyst complex mediates vesicle targeting and tethering at the plasma membrane for exocytosis (Hsu et al., 2004; Munson and Novick, 2006; He and Guo, 2009). Inhibiting exocyst function disrupts polarization in different contexts, including localized exocytosis in Saccharomyces cerevisiae (He and Guo, 2009), basolateral protein delivery in epithelial cells (Grindstaff et al., 1998), ciliogenesis (Zuo et al., 2009; Feng et al., 2012), and neurite outgrowth and branching (Vega and Hsu, 2001; Murthy et al., 2003; Lalli and Hall, 2005). Notably, the exocyst can co-immunoprecipitate with Par complex components and together they can regulate cell motility (Rosse et al., 2009), axon specification (Lalli, 2009) and ciliogenesis (Zuo et al., 2009). However, the molecular mechanisms of this interaction remain unknown.

Active Ral regulates exocyst assembly by directly binding to Sec5 and Exo84 (Moskalenko et al., 2002; Sugihara et al., 2002; Moskalenko et al., 2003) to control migration, membrane traffic, and cytoskeletal dynamics (Hertzog and Chavrier, 2011; Liu and Guo, 2012). These processes also play a fundamental role in neurogenesis, which persists in the postnatal brain in two niches, the dentate gyrus of the hippocampus and the subventricular zone (SVZ) (Zhao et al., 2008). SVZ stem cells generate neuroblasts that move rostrally, forming the ‘rostral migratory stream’ (RMS), towards the olfactory bulb (OB), where they differentiate into inhibitory interneurons (Luskin, 1993; Lois and Alvarez-Buylla, 1994; Doetsch and Alvarez-Buylla, 1996; Lledo et al., 2006). This polarized migration is crucial for postnatal neurogenesis, controlling the fate and integration of new neurons into a pre-existing synaptic network (Belvindrah et al., 2011).

Here, we describe a role for RalA and the exocyst in SVZ-derived neuroblast polarity and migration by using RMS explants, in vivo postnatal electroporation of RalA mutants and genetic deletion of Rala. We uncover a mechanism in which active RalA promotes the direct binding between Exo84 and Par6. Disrupting this interaction impairs polarity in postnatal migratory neuroblasts and embryonic neurons.

RalA regulates neurite branching and neuronal polarization in embryonic neurons (Lalli and Hall, 2005; Lalli, 2009). We found a punctate distribution of RalA in postnatal SVZ-derived rat migratory neuroblasts (Fig. 1A; supplementary material Fig. S1A,B). We inhibited RalA expression in these cells using a validated siRNA approach (Lalli and Hall, 2005; Lalli, 2009). Neuroblasts display a 50–70% decrease in RalA levels 48–72 hours after nucleofection with a siRNA oligonucleotide compared to control siRNA-nucleofected cells (supplementary material Fig. S1A–D). RalA depletion did not cause an increase in apoptosis, as shown by similarly low levels of cleaved caspase 3 positivity in control and RalA siRNA-nucleofected cells (supplementary material Fig. S1E). Given the involvement of Ral in cell motility (Oxford et al., 2005; Rossé et al., 2006; Spiczka and Yeaman, 2008), we examined whether its depletion affects neuroblast migration using 3D culture in Matrigel, a system recapitulating the neuroblast mode of migration observed in vivo (Wichterle et al., 1997). Given that the most effective RalA depletion was observed between 48 and 72 hours after siRNA nucleofection, we re-aggregated neuroblasts co-nucleofected with GFP and control or RalA siRNA oligonucleotides in hanging drops over 5 hours, cultured them in suspension for 48 hours, embedded them in Matrigel and let them migrate for 24 hours before fixation (Falenta et al., 2013). RalA knockdown caused a ∼60% reduction in migration (Fig. 1B,C). Control cells in the same aggregates (GFP-negative, stained in red for F-actin) migrated to a similar extent in all conditions (Fig. 1B, middle column, and Fig. 1C). Co-nucleofection of a plasmid encoding a siRNA-resistant wild-type (WT) RalA and GFP restored migration to control levels, showing the specificity of the RalA siRNA effect (Fig. 1B, bottom row, and Fig. 1C).

During neuroblast migration, nucleokinesis follows the extension of a leading process (Schaar and McConnell, 2005). Quantitative tracking analysis from time-lapse imaging over a 4-hour period (supplementary material Fig. S2A) showed a substantial reduction in migrated distance and speed for RalA-depleted neuroblasts (supplementary material Fig. S2B,C). RalA-depleted cells extended dynamic leading processes that frequently retracted, impairing nucleokinesis. Control nuclei displayed large saltatory movements (jumps >5 µm) (supplementary material Fig. S2D), consistent with previous observations (Schaar and McConnell, 2005; Oudin et al., 2011). RalA depletion caused a ∼6-fold reduction in the percentage of large nucleokinetic jumps, which were largely replaced by small oscillations (supplementary material Fig. S2D,E). Thus, RalA might be involved in regulating neuroblast migration.

To assess the role of RalA in RMS neuroblasts in the intact brain, we altered RalA activity by in vivo postnatal electroporation of Myc-tagged RalA versions – wild type (WT), dominant-negative (DN, RalAS28N), constitutively active (CA, RalAQ72L) and ‘fast cycling’ (between the GDP- and GTP-bound state; FC, RalAF39L) (Lalli, 2009) cloned into pCAG-IRES-EGFP (Niwa et al., 1991; Causeret et al., 2009). The empty vector (EV), or pCX-EGFP (Oudin et al., 2011) served as controls. Electroporation in the lateral ventricle of P2 mouse pups allows transfection of a neuroblast subpopulation in the intact SVZ (Boutin et al., 2008; Sonego et al., 2013a). Brains were immunostained for GFP 5 days later to examine SVZ-derived neuroblasts in the RMS. Co-expression of Myc-tagged RalA and GFP was confirmed by immunostaining (data not shown). Over 95% of neuroblasts expressing WT, CA or FC RalA extended their leading process towards the OB, similar to control cells expressing only GFP (Fig. 2A). Instead, neuroblasts expressing DN RalA displayed a ∼25% shorter leading process, with over a 3-fold increase in the percentage of cells oriented away from the OB (Fig. 2B,C). Both CA and FC RalA significantly increased process length (Fig. 2B), consistent with the idea that active RalA elicits protrusion extension.

To analyze how the different RalA variants affected migration, we counted the number of labelled cells in three different regions along the rostrocaudal axis of the RMS 5 days after electroporation: the descending arm (area 1), the horizontal arm (area 2) and the core of the OB (area 3) (Fig. 2D). Expression of DN RalA caused a ∼50% increase in the percentage of cells found in area 2, suggesting disrupted migration dynamics (Fig. 2E). On average, we observed similar amounts of GFP, CA or DN RalA-expressing cells along the entire RMS (data not shown), consistent with a lack of effect on cell viability. Taken together, these results suggest that RalA activity regulates the morphology and polarity of RMS neuroblasts in vivo.

We further investigated the role of RalA in RMS neuroblasts using a RalA conditional knockout model based on the Cre-lox system (Rala^lox/lox), because Rala null mice are embryonic lethal (Peschard et al., 2012). We electroporated Rala^lox/loxmice with pCAG-Cre-IRES2-EGFP, allowing simultaneous expression of Cre recombinase and GFP (Woodhead et al., 2006), to genetically delete Rala in a subset of RMS neuroblasts in an otherwise wild-type (wt) environment. Brain slices were stained for GFP 5 days after electroporation (Fig. 3A,B). Cre expression alone did not affect neuroblast migration in wt mice (supplementary material Fig. S3A). As expected, expressing only GFP (by electroporating pCX-EGFP) in Rala^lox/lox animals did not affect neuroblast morphology, polarity or migration (supplementary material Fig. S3B,C), ruling out any abnormalities due to genetic background. To monitor RalA expression, we immunostained Rala^lox/lox neuroblasts transfected with pCAG-Cre-IRES2-EGFP for RalA. Visible reduction of RalA expression [46.70%±4.0 (±s.e.m.) decrease in RalA fluorescence intensity levels in GFP-positive cells compared to untransfected neuroblasts; P<0.001, n = 25 GFP-positive and 25 GFP-negative cells] was only detected 5 days after Cre transfection (Fig. S3D), when neuroblasts have already started to reach the OB. At this time point, deletion of Rala altered neuroblast morphology, causing a ∼25% decrease in leading process length and a significant increase in cells lacking a process (Fig. 3C,D). In addition, we observed a >6-fold increase in the amount of misoriented neuroblasts compared to wt animals (Fig. 3E). There was no significant difference in the number of cells reaching the OB, possibly owing to a long half-life of RalA in vivo and/or to the time necessary to achieve full Cre expression and function (unpublished data). To test the function of the related RalB isoform on RMS neuroblast morphology and migration, we electroporated pCX-EGFP in Rala^lox/lox; Ralb^−/− mice (Peschard et al., 2012). Lack of RalB did not cause detectable changes in neuroblast morphology, orientation or migration (Fig. 3). These results support the conclusion that RalA, but not RalB, is required for proper morphology and orientation of neuroblasts in vivo.

Expression of either CA or FC RalA increased elongation of the leading process (Fig. 2), similar to active RalA-enhanced axon extension involving the exocyst (Lalli and Hall, 2005). To determine whether RalA regulates neuroblast morphology and migration through the exocyst, we electroporated CA RalA or an Exo84-uncoupled CA RalA version (RalAQ72LA48W) (Fukai et al., 2003; Hazelett et al., 2011) and examined RMS neuroblasts 5 days later. To assess the specificity of the effects caused by Exo84-uncoupled RalA, we also electroporated CA RalA uncoupled from RalBP1 (RalAQ72LD49N), another Ral effector (Jullien-Flores et al., 2000; Lalli and Hall, 2005) (Fig. 4A). Although all RalA mutants increased process length (Fig. 4B), only Exo84-uncoupled RalA (A48W) caused a substantial increase in the proportion of cells lacking a leading process (Fig. 4C), and significantly increased the number of misoriented neuroblasts (Fig. 4D). Electroporation of another RalA mutant uncoupled from both Sec5 and Exo84 (RalAQ72LD49E) (Moskalenko et al., 2003) similarly increased the percentage of misoriented neuroblasts (10.18%±2.62; P<0.01; n = 5 brains), reinforcing the role of the exocyst in regulating neuroblast polarity downstream of RalA. Uncoupling RalA from Exo84 also impaired migration, causing a ∼30% decrease in the proportion of cells found in the OB compared to both CA and RalBP1-uncoupled RalA (Fig. 4E). To assess the longer-term consequences of this effect, we analysed the distribution of GFP-positive cells found in the OB 14 days after in vivo electroporation. At this later time point, most of the control cells have radially migrated in the OB (Fig. 5A) and started to differentiate into interneurons. About 75% of control GFP-labelled cells were found in the outer OB (Fig. 5A–C, area B) whereas ∼25% were in the OB core (Fig. 5A–C, area A). Expression of CA RalA led to a ∼15% decrease in the proportion of cells found in the outer OB, whereas expression of Exo84-uncoupled RalA caused a more pronounced effect (∼35% decrease) (Fig. 5C). Uncoupling RalA from Exo84 therefore affects neuroblast polarity, morphology and migration, ultimately disrupting the normal distribution of newborn neurons in the OB.

Given that the interaction between the exocyst and the Par complex contributes to different polarization events (Lalli, 2009; Zuo et al., 2009), we asked whether a similar interaction could also be detected in RMS lysates, which are positive for the migrating neuroblast marker doublecortin (Dcx) (Koizumi et al., 2006) (Fig. 5D). An anti-Sec6 antibody immunoprecipitated not only the other exocyst subunits Sec8 and Exo84 but also Par3, supporting the presence of an endogenous exocyst–Par interaction in the RMS (Fig. 5D).

We investigated the molecular mechanism underlying the exocyst–Par interaction by testing whether exocyst subunits can directly bind a Par complex component. We first examined Exo84, a direct binding partner of active RalA (Moskalenko et al., 2003), and potentially of Par3 and Par6 because of a putative canonical PDZ-binding motif present at the C-terminus of Exo84, which could bind the PDZ domains of Par3 and Par6 (Joberty et al., 2000). Rat Exo84 was translated in vitro in the presence of [³⁵S]methionine and incubated with different GST-tagged PDZ domains of Par3 and Par6 [Par3-PDZ1, Par3-PDZ2, Par3-PDZ3, Par3-PDZ(2+3) and Par6-PDZ] (Lin et al., 2000). Exo84 specifically bound to GST–Par6-PDZ (Fig. 6A). This interaction was confirmed with a GST pull-down assay using hTERT-RPE1 cell lysates, where endogenous Exo84 bound to GST-Par6-PDZ (Fig. 6B).

To examine whether Exo84 mediates the interaction between Par6 and the exocyst, a GST pull-down assay was carried out in hTERT-RPE1 cells transfected with Exo84 siRNA. We found that Exo84 levels decreased by almost 90%, whereas the levels of another exocyst subunit, Sec8, remained unchanged (Fig. 6C). In control cells, GST–Par6-PDZ precipitated both Exo84 and Sec8. Exo84 depletion abolished the ability of Par6-PDZ to bind Sec8. In contrast, depleting Sec8 did not affect binding of Exo84 to Par6-PDZ. Importantly, the Par6-PDZ–Exo84 interaction was substantially reduced in RalA-depleted cells (Fig. 6D). Given that RalA activity modulates the exocyst–Par association (Lalli, 2009), we tested whether RalA regulates the direct interaction between Exo84 and Par6. The presence of constitutively active RalA (RalA23V) enhanced the binding between Exo84 and Par6 by ∼50% compared to control samples, whereas dominant-negative RalA (RalA28N) did not affect the Exo84–Par6 interaction (Fig. 6E,F).

Besides a classical C-terminal PDZ-binding domain, Exo84 also contains an internal PDZ-binding motif (E³⁴⁵GAVDLLD³⁵²) located close to the Ral-binding domain (Fig. 7A). This non-canonical PDZ-binding sequence is also found in Pals1, another polarity regulator interacting with Par6-PDZ (Penkert et al., 2004). We therefore translated full-length (FL) Exo84 in vitro and Exo84 mutants lacking either the C-terminal PDZ-binding motif (CΔ3) or harboring mutations in the internal motif at two key residues crucial for PDZ binding (INT-mt; V³⁴⁸D³⁴⁹→A³⁴⁸A³⁴⁹). These Exo84 variants were subjected to a GST pull-down assay with recombinant GST–Par6-PDZ. Both Exo84FL and CΔ3 bound to Par6, whereas Exo84INT-mt failed to bind to GST–Par6-PDZ, suggesting that the interaction between the Par6-PDZ and Exo84 is mediated by the internal PDZ-binding motif of Exo84 (Fig. 7B). Similar results were obtained by a GST pull-down assay using recombinant GST–Par6-PDZ and cell lysates from hTERT-RPE1 cells expressing different GFP-tagged Exo84 proteins (Fig. 7C). As a control, we confirmed that mutating the internal PDZ-binding domain of Exo84 did not affect its ability to bind one of its known interacting partners, Exo70 (Fig. 7D). We also examined the ability of Exo84FL, Exo84INT-mt and Exo84CΔ3 to bind Par6-PDZ in the presence of active RalA. Whereas Exo84FL and Exo84CΔ3 interacted with GST–Par6-PDZ, Exo84INT-mt remained unable to bind Par6-PDZ (supplementary material Fig. S4A).

To test whether the interaction between the non-canonical PDZ-binding motif of Exo84 and Par6-PDZ is direct, we expressed a His₆-tagged Exo84 fragment (INT-frag, aa 251–450) containing the internal PDZ-binding motif from bacteria. Recombinant GST or GST–Par6-PDZ was incubated with a His₆-tagged Exo84 INT-frag or Exo84 INT-frag harboring the V³⁴⁸D³⁴⁹→A³⁴⁸A³⁴⁹ mutations (INT-frag-mt, aa 251–450). Exo84 INT-frag was sufficient to interact with GST–Par6-PDZ, whereas Exo84 INT-frag-mt lost its ability to bind GST–Par6-PDZ (Fig. 7E). Overexpression of a GFP-tagged Exo84 INT-frag decreased the endogenous exocyst–Par interaction in primary cortical neurons (Fig. 7F). Taken together, these results show that Exo84 directly interacts with Par6-PDZ through its non-canonical PDZ-binding motif, and that residues 348 and 349 are crucial for this interaction.

We asked whether the Exo84–Par6 interaction regulates neuroblast morphology, polarity and migration. RMS neuroblasts were nucleofected with GFP, GFP-tagged Exo84 INT-frag or INT-frag-mt. Cells were re-aggregated in clusters, embedded in Matrigel 24 hours after nucleofection and left to migrate for 24 hours. Expression of Exo84 INT-frag caused a ∼40% decrease in migration distance, whereas Exo84 INT-frag-mt caused a more modest inhibitory effect (Fig. 8A,B). Overexpression of Exo84 INT-frag, but not INT-frag-mt also significantly impaired axon specification in embryonic cortical neurons, a process relying on the exocyst–Par interaction (Lalli, 2009) (supplementary material Fig. S4B,C).

To test the functional role of the exocyst–Par interaction in vivo, a pCAG-IRES-EGFP vector expressing Exo84 INT-frag or INT-frag-mt was electroporated in mouse pups. Brains were collected 5 days later and stained for GFP. Expression of Exo84 INT-frag visibly altered the morphology of migrating neuroblasts, many of which displayed ectopic branching (Fig. 8C). Whereas Exo84 INT-frag and INT-frag-mt slightly increased process length (Fig. 8D), only Exo84 INT-frag caused a ∼4-fold increase in the amount of misoriented cells (Fig. 8E) and significantly increased the percentage of branched cells compared to GFP-expressing control neuroblasts (Fig. 8F). Similar relative percentages of cells were detected in areas 1–3 of the RMS 5 days after electroporation of all three constructs (data not shown). Taken together, these results point to a crucial functional role for the Exo84–Par6 interaction in the regulation of neuroblast morphology and polarity.

We have uncovered a role for the small GTPase RalA and the exocyst complex in the regulation of neural progenitor polarity using primary cultures, in vivo electroporation of RalA mutants and, for the first time, inducible genetic deletion of Rala in the brain. Our study highlights a novel molecular mechanism controlling cell polarity. By binding to its effector Exo84, RalA promotes the direct interaction between Exo84 and the PDZ domain of Par6 through an internal non-canonical PDZ-binding motif. A functional role for this interaction was further demonstrated in the polarization of postnatal migrating neuroblasts and embryonic neurons.

Modulation of RalA activity is important for proper neuroblast morphology and polarity in vivo, as shown by the disruption of these features in neuroblasts expressing inactive RalA. We could not properly assess the effect of Rala genetic deletion on migration because a substantial decrease of RalA protein levels was only observed at 5 days after Cre transfection, when most of the electroporated neuroblasts had already migrated along the RMS. This delay might be due to the time required for Cre-mediated excision but also suggests that RalA has a prolonged half-life in neuroblasts in vivo and that reduced levels are sufficient for normal motility. Moreover, compensatory mechanisms might act in vivo following Rala loss. Still, downregulation of RalA in Rala^lox/lox mice significantly disrupted neuroblast morphology and polarity, whereas genetic deletion of Ralb did not have any significant effect (Fig. 3), indicating some different roles between the two Ral isoforms, which can act redundantly in other systems (Lalli and Hall, 2005; Peschard et al., 2012). This evidence, together with the defects caused by siRNA-mediated RalA depletion in vitro and by expression of inactive or exocyst-uncoupled RalA mutants in vivo, strongly supports a role for this small GTPase in regulating neuroblast morphology and polarity. This might have important implications for the maturation of neuroblasts into neurons (Belvindrah et al., 2011; Sanai et al., 2011) and for their ability to re-route to injured brain areas (Arvidsson et al., 2002; Sundholm-Peters et al., 2005). Potential modulators of RalA activity in migrating neuroblasts could be cell adhesion molecules, such as N-cadherin or integrins, which are enriched in the RMS (Yagita et al., 2009; Kazanis et al., 2010) and have been linked to RalA activation in other systems (Lalli and Hall, 2005; Jossin and Cooper, 2011). Another interesting candidate is cannabinoid signaling, which we have recently identified as a major regulator of neural progenitor migration in vivo (Oudin et al., 2011) and has been associated with RalA in pathways controlling neurite outgrowth (He et al., 2005).

The exocyst is a crucial RalA effector in neuroblast migration, as shown by the delayed migration of neuroblasts expressing Exo84-uncoupled RalA, not only along the RMS, but also at later stages in the OB, where neuroblasts migrate radially and differentiate. Indeed, perturbing the RalA–Exo84 interaction significantly disrupted the final distribution of newborn neurons in the OB (Fig. 5), which is likely to affect their fate and synaptic integration, as in the case of another migration regulator, doublecortin (Belvindrah et al., 2011). RMS neuroblast migration is saltatory, alternating process extension with nucleokinesis, with a very dynamic leading protrusion (Nam et al., 2007; Sonego et al., 2013b). Interestingly, perturbing the RalA–Exo84 interaction impaired neuroblast migration, and caused both process loss and misorientation. These observations suggest that the RalA–exocyst interaction might be involved in multiple functions during neuroblast migration, such as process initiation and orientation in response to extracellular stimuli along the RMS (Lalli, 2013).

RalA and the exocyst are involved in neuronal migration in the developing neocortex (Letinic et al., 2009; Jossin and Cooper, 2011). At the molecular level, however, the consequence of RalA–exocyst interaction is still unclear. Our work has now shown that active RalA promotes the interaction between the exocyst subunit Exo84 and the Par complex component Par6. Like the polarity protein Pals1, Exo84 binds the PDZ domain of Par6 through an internal, non-canonical PDZ-binding motif. Importantly, expression of this Exo84 internal fragment disrupts the exocyst–Par interaction and affects neuroblast morphology and orientation to a similar extent to Rala genetic deletion or Exo84-uncoupling of RalA. Expression of the same Exo84 fragment also impairs axon specification in polarizing cortical neurons, a process relying on a RalA-modulated interaction between the exocyst and the Par complex (Lalli, 2009). Furthermore, mutating two crucial residues in the internal PDZ-binding motif of Exo84 abolishes the disrupting effect, supporting the functional importance of this region in cell polarization. Given that the Exo84 Ral-binding domain is located close to the internal PDZ-binding motif, binding of active RalA could cause a conformational change in Exo84 to enhance its interaction with Par6-PDZ.

In epithelial cysts and neuronal progenitors, Par6 is recruited to the cortex by active Cdc42, which binds to its CRIB domain (Joberty et al., 2000; Chen et al., 2006). Interestingly, binding of Pals1 with Par6-PDZ occurs independently of Cdc42 binding to Par6 (Penkert et al., 2004). Localized recruitment of Par6 can occur independently of Cdc42, as reported in dendritic spines (Zhang and Macara, 2008). Future studies are needed to address a potential link between RalA and Cdc42 in migrating neuroblasts and in polarizing neurons, given the similar localization of both GTPases at the nascent axon tip (Schwamborn and Püschel, 2004; Lalli, 2009).

Expressing the Exo84 PDZ-binding motif did not disrupt migration as did Exo84-uncoupled RalA. This might seem surprising; however, based on our data the Exo84 internal PDZ-binding motif is not required for exocyst complex association. In the RMS, RalA-promoted exocyst assembly might serve multiple functions, such as the polarized delivery of integrins (Spiczka and Yeaman, 2008) and matrix metalloproteinases (Liu et al., 2009), which can regulate migration along the RMS (Murase and Horwitz, 2002; Belvindrah et al., 2007; Bovetti et al., 2007), and the recruitment of IQGAP1 (Sakurai-Yageta et al., 2008), a signaling scaffold protein involved in neuroblast migration (Balenci et al., 2007). In addition, our single-end-point assay in fixed brain slices, which can detect strong migration phenotypes, might lack the sensitivity needed to uncover subtler defects in cell dynamics. Future studies using live imaging of brain slice cultures will help address this point.

In summary, we have uncovered a novel signaling mechanism involving RalA, Exo84 and Par6 for the control of cell polarity. Although our study focuses on the RalA-promoted exocyst–Par interaction in a neuronal context, it is likely that such a regulatory mechanism operates in other polarization events, such as the establishment of epithelial asymmetry, directed cell migration or immune synapse formation (Grindstaff et al., 1998; Ludford-Menting et al., 2005; Hertzog and Chavrier, 2011).

Antibodies used were: mouse anti-Sec6 and anti-tau-1 (Calbiochem), mouse anti-RalA and anti-Sec8 (BD Biosciences), rabbit anti-Par3 (Upstate), rabbit anti-cleaved caspase 3 (Cell Signaling), rabbit anti-GFP (Invitrogen), mouse anti-GFP (Proteintech), mouse anti-Myc (Abcam). Mouse anti-Sec8 and anti-Exo84 were gifts from Shu-Chan Hsu (Rutgers University, Piscataway, NJ). Alexa-Fluor-488- and Texas-Red-conjugated secondary antibodies and Texas-Red–phalloidin were from Invitrogen. HRP-conjugated antibodies were from Dako. Cell culture reagents were from Invitrogen and all other reagents were from Sigma unless otherwise specified.

All procedures were performed in accordance with UK Home Office regulations (Animals Scientific Procedures Act, 1986). P2–P3 CD1 mouse pups were obtained from Charles River. P5–P7 Sprague Dawley rat pups were obtained from Harlan. The generation of Rala^lox/lox and Rala^lox/lox; Ralb^−/− mice in FVB background was as previously described (Peschard et al., 2012).

Electroporation was performed in P2 mice as previously described (Oudin et al., 2011). Animals were subjected to five electrical pulses of 100 V for 50 milliseconds with 850-millisecond intervals using a CUY21SC electroporator (Nepagene) (Sonego et al., 2013a).

E18 cortical neurons were nucleofected and cultured as described previously (Lalli, 2009). Neurons extending single processes that were at least twice as long as other neurites and tau-1-positive were scored as polarized. At least 110 cells were analyzed for every nucleofected construct in each of three independent experiments.

The RMS was dissected from P5–P7 mice or rats (Oudin et al., 2011). Following brief trituration, cells were re-suspended in pre-equilibrated Neurobasal complete medium (containing B27 supplement, 2 mM L-glutamine and 0.6% glucose) or nucleofected (see below). Cells were plated onto polyornithine- and laminin-coated 6-well plates (1×10⁶ cells/well) and maintained in Neurobasal complete medium at 37°C under 5% CO₂. Reaggregated nucleofected neuroblasts were embedded in Matrigel as previously described (Oudin et al., 2011).

Dissociated RMS neuroblasts were resuspended in mouse/rat neuron nucleofection solution (Lonza) according to the manufacturer's instructions. Each sample was mixed with 3 µg DNA or 9 µg siRNA, and nucleofected using program G-013. For western blot analysis, cells were plated onto polyornithine- and laminin-coated six-well plates (1×10⁶ cells/well). For migration assays, cells were re-suspended in 30 µl of DMEM + 10% FCS, pipetted as a drop inside a p35 dish lid, and inverted over a dish containing Neurobasal complete medium. Hanging drops were transferred into the medium 5 hours later and cultured in suspension for 48 hours before embedding in growth-factor-reduced Phenol-Red-free Matrigel (BD) on glass coverslips (Falenta et al., 2013). Neuroblasts were allowed to migrate for 24 hours in Neurobasal complete medium at 37°C under 5% CO₂ before fixation.

Matrigel-embedded neuroblasts were fixed in 4% PFA for 40 minutes, washed with PBS, blocked and permeabilized with 15% goat serum, 0.3% Triton X-100, 0.1% BSA in phosphate-buffered saline (PBS) for 1 hour at room temperature. Primary antibodies were diluted in the same solution and incubated overnight at 4°C. Coverslips were washed, incubated with secondary antibodies plus Hoechst 33342 for 2 hours at room temperature and mounted with mounting medium (Dako).

Electroporated brains were hemisected, fixed in 4% PFA overnight at 4°C, embedded in 15% gelatin, and left for 1 hour at room temperature followed by 2 hours at 4°C. After fixing in 4% PFA overnight at 4°C, they were cut on a Vibratome (Leica) into 75-µm-thick sagittal sections. Slices were blocked and permeabilized for 1 hour at room temperature in 1% BSA, 0.1% Triton X-100, 0.1% sodium azide in PBS and incubated overnight at 4°C with primary antibodies diluted in the same solution. After washing in PBS, slices were incubated with secondary antibodies for 3 hours, washed in PBS and mounted. z-stacks (images taken every 0.5–1.5 µm) of the RMS were captured on a Zeiss LSM 710 confocal microscope using EC Plan Neofluar 10× 0.3 NA and 40× 1.3 NA objectives and Zeiss Zen software. All images were digitally processed in Photoshop, and where needed the brightness and contrast features were modified for the entire field of view to better visualize cell morphology.

The morphology of GFP-expressing neuroblasts was examined from confocal z-stack projections of 75-µm-thick sagittal brain slices immunostained with an anti-GFP antibody. Process length was measured as the distance from the base of the cell body to the tip of the leading process using ImageJ. Cells were considered lacking a process if they did not display any visible protrusion extending from the cell soma, or branched if they displayed one or more primary and/or secondary branches from the leading process. To analyze orientation, we traced a line in the middle of the cell body running perpendicular to the direction of migration along the RMS towards the OB. Neuroblasts with their leading process extending in the sector opposite to the direction of migration were considered misoriented. At least 250 cells were analyzed from each brain, and between three and six brains were analyzed for each electroporated construct or genotype.

To analyze migration 5 days after electroporation, the RMS was classified into three anatomically distinct areas (area 1, descending arm of the RMS; area 2, RMS horizontal arm; area 3, core of the OB) as previously described (Belvindrah et al., 2011). Projections of confocal z-stacks were used to count the number of GFP-labeled neuroblasts in areas 1–3 in all slices containing the RMS (typically four slices per brain). The percentage of cells in each area (relative to the total number of labeled cells along the entire RMS) was calculated for statistical comparison among samples. Between three and eight brains were analyzed per genotype or electroporated construct.

To analyze the distribution of GFP-labeled cells in the OB, a mask outlining the inner (‘A’) and outer (‘B’) OB areas (minor axis, 1.25 and 3 mm; major axis, 0.73 and 1.7 mm, respectively) was aligned with the edge of the OB on confocal projections of sagittal brain slices (Fig. 5B). For each brain, the percentage of cells in the two OB areas was calculated after counting GFP-positive cells in all the slices containing the OB (typically five slices per brain).

Images of immunostained aggregates were captured on a Zeiss Axioplan 2 microscope equipped with an ApoTome module and a Zeiss MRm Axiocam CCD camera using A-Plan 5× 0.12 NA, A-Plan 10× 0.25 NA and Plan Apo 20× 0.75 NA objectives and Axiovision software. The distance from the edge of an aggregate to the furthest migrated neuroblast perpendicular to the edge was measured at six points around the entire periphery of the aggregate using ImageJ. For each experiment 10 to 20 isolated aggregates were analyzed.

For time-lapse imaging, nucleofected and re-aggregated neuroblasts were embedded in Matrigel in a 4-chamber 35-mm Hi-Q4 culture dish (Nikon) and imaged in Neurobasal complete medium. Images were captured by the Nikon Biostation IM software using a Nikon Biostation (an automatic multipoint time-lapse imaging system with controlled environment maintained at 37°C under 5% CO₂) with a 20× 0.8 NA objective every 3 minutes for 24 hours. Tracking analysis was performed using Volocity software (Perkin Elmer) as previously described (Oudin et al., 2011), and limited to the first 7 hours of filming. Neuroblasts emerged from aggregates at different times during the filming period, and cells were tracked for their first 4 hours of migration upon exiting aggregates. At least 40 cells were tracked for each condition in each experiment.

Rat Exo84 was subcloned in pcDNA3.1 and in vitro translated with the TNT quick-coupled system (Promega) in the presence of [³⁵S]methionine. A total of 5 µl of the TNT reaction mixture was incubated with 10 µg bacterially purified PDZ domains of Par3 and Par6 in 0.5 ml of binding buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 2 mM MgCl₂, 1% Triton X-100) for 2 hours at 4°C. After four washes with binding buffer, bound Exo84 was analyzed by SDS-PAGE and autoradiography. In some cases in vitro translated Exo84 was incubated with 1 µg of either purified RalA23V or RalA28N in 0.1 ml of buffer (20 mM Tris-HCl pH 7.4, 100 mM NaCl, 5 mM EDTA, 1 mM dithiothreitol (DTT) and 10 mM MgCl₂) for 1 hour at 25°C before adding purified Par3 or Par6 PDZ domains. For GST–PDZ pull-down assays, hTERT-RPE1 cells were lysed in a buffer containing 20 mM Tris-HCl pH 7.5, 25 mM KCl, 1 mM MgCl₂, 0.5 mM EGTA, 1 mM DTT, 0.5% Triton X-100 and protease inhibitors. The cell lysate (1 mg of total protein) was mixed with 20 µl (50% v/v) of glutathione–Sepharose conjugated with 10 µg of GST–Par3 or –Par6 PDZ domains for 2 hours at 4°C. Beads were then washed five times with lysis buffer. Bound Exo84, Sec8 and RalA were monitored by western blotting. For binding experiments with recombinant fusion proteins, bacterially purified GST-tagged PDZ domains and His₆-tagged Exo84 proteins were incubated in binding buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 2 mM MgCl₂, 1% Triton X-100) for 2 hours at 4°C. After four washes with binding buffer, bound Exo84 was analyzed by SDS-PAGE and western blotting.

Statistical analysis was performed using Student's t-test for dual comparison and one-way ANOVA for multiple comparison with SigmaPlot 12.0 (Systat Software Inc.). Differences were considered statistically significant if P<0.05.

pCX-EGFP was a kind gift from Masaru Okabe and Jun-ichi Miyazaki (Osaka University, Osaka, Japan). pCAG-Cre-IRES2-GFP (Woodhead et al., 2006) (Addgene plasmid 26646) was provided by Anjen Chenn (Northwestern University, Evanston, IL). For in vivo electroporation, the DNA sequences encoding Myc-tagged human WT RalA, constitutively active (CA) RalA (RalAQ72L), fast cycling (FC) RalA (RalAF39L), and dominant-negative (DN) RalA (RalAS28N) were amplified by PCR from the original constructs in pRK5-myc (Lalli, 2009) using the primers 5′-GAATTGGCTAGCATGGAGCAGAAGCTGATCTCCGAGGAGG-3′ and 5′-TCTGCAGATCGATTTATAAAATGCAGCATCTTTCTCTG-3′, and subcloned into pCAG-IRES-EGFP (a kind gift from M. Hoshino, National Institute of Neuroscience, Tokyo, Japan) (Causeret et al., 2009). The different effector-uncoupled CA RalA versions were obtained from site-directed mutagenesis of pCAG-CARalA-IRES-EGFP using the following primers for Exo84-uncoupled RalAQ72LA48W: 5′-AGGACTATGAGCCTACCAAATGGGACAGCT/et al/et al/tgciqACCGGAAGAAGG-3′ and 5′-CCTTCTTCGGAGCTGTCCCATTTGGTAGGCTCATAGTCCT-3′; and for RalBP1-uncoupled RalAQ72LD49N, 5′-GCCTACCAAAGCAAACAGCTACCGGAAGAAGG-3′ and 5′-CCTTCTTCCGGTAGCTGTTTGCTTTGGTAGGC-3′.

The Exo84, Par3 and Par6 constructs used for in vitro translation, immunoprecipitation and direct binding assays were: pcDNA3.1-Exo84 (1–716); pcDNA3.1-Exo84-CΔ3 (1–713); pcDNA3.1-Exo84-INT-mt (1–716, V³⁴⁸D→AA³⁴⁹); pEGFPC2-Exo84 (1–716); pEGFPC2-Exo84-CΔ3 (1–713); pEGFPC2-Exo84-INT-mt (1–716, V³⁴⁸D→AA³⁴⁹).

Par3-PDZ1, Par3-PDZ2, Par3-PDZ3, Par3-PDZ(2+3) and Par6C-PDZ in pGEX4T-1 (Lin et al., 2000) were kindly provided by Tony Pawson (Samuel Lunenfeld Research Institute, Toronto, Canada).

To generate His₆-tagged Exo84 fusion proteins, the internal Exo84 fragment (INT-frag, aa 251–450) and the Exo84 internal fragment mutant (INT-frag-mt, 251–450 aa; V³⁴⁸D→AA³⁴⁹) were amplified by PCR using primers incorporating appropriate restriction sites, and cloned in frame into pET32a(+) to generate His₆–Exo84-INT-frag and His₆–Exo84-INT-frag-mt. His₆-tagged RalA23V and RalA28N were generated by cloning into pET32a(+). The sequences encoding Exo84 INT-frag and Exo84 INT-frag-mt were inserted into pCAG-IRES-EGFP for in vivo electroporation. All constructs were confirmed by DNA sequencing.

For siRNA experiments in rat neuroblasts, we used a validated siRNA oligonucleotide targeting RalA [AGACTACGCTGCAATTAGA (Dharmacon) (Lalli, 2009)]. A non-targeting siRNA oligonucleotide was used as a control (AGGUAGUGUAAUCGCCUUGTT). For rescue experiments we co-nucleofected rat neuroblasts with siRNA oligonucleotides together with either pCAG-IRES-EGFP or with pCAG-wtRalA-IRES-EGFP, which contains an siRNA-resistant version of wild-type simian RalA with two mismatches in the region targeted by the RalA siRNA oligonucleotide.

Immortalized human retinal pigmented epithelial cells (hTERT-RPE1) were grown to 50% confluence and transfected using Lipofectamine 2000 with siRNA duplexes targeting human Exo84 (5′-GGTGCCACTTTACTCTATA-3′), human RalA (5′-CCAAATGAAGGGATACCAT-3′) and human Sec8 (5′-AGAACCTGCTTTCATGCAA-3′). We used the Luciferase siRNA duplex GL2 (5′-AACGTACGCGGAATACTTCGA-3′) targeting luciferase as a control. The efficiency of protein knockdown was determined by western blotting.

The RMS from P7 rat pups was carefully dissected and homogenized in 100 µl of immunoprecipitation (IP) buffer containing 150 mM NaCl, 10 mM Tris-HCl pH 7.4, 1 mM EDTA, 1 mM EGTA pH 8.0, 1 mM MgCl₂, 1 mM Na₃VO₄, 10 mM NaF, 1 mM PMSF and one Complete protease-inhibitor tablet/10 ml (Roche). After incubating for 30 minutes on ice, homogenates were cleared by centrifugation at 19,280 g for 15 minutes at 4°C and pre-cleared for 1 hours with protein-G–Sepharose on a rotating wheel at 4°C. Protein concentration was measured in supernatants using a Bradford assay (Bio-Rad Laboratories). Extracts were then incubated with 2 µg of anti-Sec6 antibody or control IgG and protein-G–Sepharose overnight at 4°C. After centrifugation for 5 minutes at 560 g, immunocomplexes were washed four times with IP buffer containing 250 mM NaCl and 0.5% Triton X-100, eluted in boiling Laemlli sample buffer, and examined by SDS-PAGE and western blot analysis with the indicated antibodies as previously described (Lalli, 2009). A similar protocol was followed to immunoprecipitate Sec6 from embryonic rat cortical neurons nucleofected with either GFP or GFP-Exo84 INT-frag (∼23×10⁶ cells per IP).

We thank Anjen Chenn, Mikio Hoshino, Shu-chan Hsu, Jun-ichi Miyazaki, Masaru Okabe, Tony Pawson and Charles Yeaman for providing reagents and Matt Grubb, Alan Hall and Giampietro Schiavo for valuable comments on the manuscript.