The ins and outs of the Arf4-based ciliary membrane-targeting complex

ABSTRACT

The small GTPase Arf4-based ciliary membrane-targeting complex recognizes specific targeting signals within sensory receptors and regulates their directed movement to primary cilia. Activated Arf4 directly binds the VxPx ciliary-targeting signal (CTS) of the light-sensing receptor rhodopsin. Recent findings revealed that at the trans-Golgi, marked by the small GTPase Rab6, activated Arf4 forms a functional complex with rhodopsin and the Arf guanine nucleotide exchange factor (GEF) GBF1, providing positive feedback that drives further Arf4 activation in ciliary trafficking. Arf4 function is conserved across diverse cell types; however, it appears that not all its aspects are conserved across species, as mouse Arf4 is a natural mutant in the conserved α3 helix, which is essential for its interaction with rhodopsin. Generally, activated Arf4 regulates the assembly of the targeting nexus containing the Arf GAP ASAP1 and the Rab11a–FIP3–Rabin8 dual effector complex, which controls the assembly of the highly conserved Rab11a–Rabin8–Rab8 ciliary-targeting module. It was recently found that this module interacts with the R-SNARE VAMP7, likely in its activated, c-Src-phosphorylated form. Rab11 and Rab8 bind VAMP7 regulatory longin domain (LD), whereas Rabin8 interacts with the SNARE domain, capturing VAMP7 for delivery to the ciliary base and subsequent pairing with the cognate SNAREs syntaxin 3 and SNAP-25. This review will focus on the implications of these novel findings that further illuminate the role of well-ordered Arf and Rab interaction networks in targeting of sensory receptors to primary cilia.

Abbreviations: CTS: Ciliary-Targeting Signal; GAP: GTPase Activating Protein; GEF: Guanine Nucleotide Exchange Factor; RTC(s), Rhodopsin Transport Carrier(s); SNARE: Soluble N-ethylmaleimide-sensitive Factor Attachment Protein Receptor; TGN: Trans-Golgi Network.

Introduction

Primary cilia are ubiquitous organelles that originate from the basal bodies and arise in different shapes to fit their precise function in sensory receptor signalling. The exceptionally elaborate primary cilia of the retinal rod photoreceptor cells, the rod outer segments (ROS), concentrate the G-protein-coupled receptor rhodopsin and the accompanying phototransduction apparatus into thousands of light-sensitive membranous discs, creating a distinctive microenvironment that supports the optimal detection and propagation of the visual signal [1,2]. The light-sensitive ROS membranes are continuously subjected to photo-oxidative damage, necessitating their constant renewal to maintain photoreceptor homeostasis [3,4]. A major challenge during this process is the precise linkage of the ciliary cargo uptake into the Golgi-to-cilia-directed membrane carriers, with the assembly of the membrane trafficking machinery responsible for their ultimate fusion with the periciliary plasma membrane [3,5]. Through continuous remodelling, the conserved Arf4-based ciliary-targeting complex – containing the Arf GAP ASAP1 and the Rab11a–FIP3–Rabin8 dual effector complex, which activates a regulator of carrier fusion Rab8 – provides an ever-changing platform that guides the ciliary cargo to its final destination [6–21]. Two recent studies in retinal rod photoreceptors further delineated the assembly of the targeting complex, and revealed essential roles of the Arf GEF GBF1 and the R-SNARE VAMP7, in the initial and final stages of ciliary pathway, respectively [22,23].

Arf GTPases and their regulators

Arf4 belongs to the family of small GTPases that includes Arf, Arf-like (Arl) and Sar proteins [24]. Arfs regulate membrane trafficking, lipid metabolism, organelle morphology and cytoskeleton dynamics [25]. The mammalian Arfs consist of six isoforms (Arf1–Arf6); Arfs1–5 are associated with the Golgi and Arf6 functions at the plasma membrane [26]. Arf4, Arl3, Arl6 and Arl13b are also specifically implicated in membrane targeting to primary cilia, dysfunction of which is a known cause of human genetic diseases and syndromic disorders known as ciliopathies [6,26–30]. Arf GTPases function through the cycles of GTP binding and hydrolysis that are regulated by Arf guanine nucleotide exchange factors (GEFs) and GTPase activating proteins (GAPs) [26,31–35]. GEFs and GAPs control membrane association and signalling pathways of Arf GTPases through activation cascades and positive-feedback loops [36,37]. The prototypical Arf, Arf1, in its GDP-bound form associates with membranes through the N-terminal myristoyl group; however, GEF activation and GTP binding cause a conformational transition, termed the ‘myristoyl switch’ that tightly couples Arf1 activation with stable membrane association [38–43]. Unlike Arf1, GDP-bound Arf4 and Arf5 can stably associate with membranes without Arf GEF, using the N-terminal amphipathic helix and a specific residue in the conserved domain [44,45], and this difference may partially underlie their specific functions that differ from those of Arf1.

In their active, membrane-bound form, Arfs interact with multiple effectors such as coat proteins and lipid modifying enzymes [46]. Because Arfs have intrinsically low rate of GTP hydrolysis, inactivation and termination of Arf signalling is executed by Arf GAPs, [33,35], which in this way couple the proof-reading of cargo incorporation to budding of transport carriers [47]. Particularly, the Arf GAPs ASAP1 belongs to a family of multifunctional proteins containing pleckstrin–homology (PH), ankyrin repeats, proline-rich and SH3 domains, which serve as large scaffolds for the assembly of signalling complexes [33,35,48]. ASAP1 functions both as an Arf GAP and an Arf effector, because its N-terminal BAR (Bin/amphiphysin/Rvs) domain, together with the GAP domain, acts as a coincidence detector that senses the membrane phospholipid composition and the presence of GTP-Arf and responds by increasing membrane curvature, which may facilitate carrier budding [49,50].

Assembly of the Arf4-mediated ciliary membrane-targeting complex at the Golgi/TGN

The fundamental step that initiates the assembly of the rhodopsin ciliary-targeting complex involves activated Arf4, which interacts directly with the rhodopsin C-terminal VxPx ciliary-targeting signal (CTS) at the Golgi/TGN [51] (Figure 1). The specific GEF that activates Arf4 in transport to the cilia was recently identified as the Arf GEF GBF1 [22], which will be described in detail below. The activated Arf4 and rhodopsin next form a complex with the Arf GAP ASAP1 [6]. ASAP1 functions both as an Arf4 GAP and its effector, as evidenced by the Arf4 mutant deficient in ASAP1 hydrolysis that causes retinal degeneration, even though other Arf GAPs can mediate GTP hydrolysis [6]. ASAP1 also controls selective binding of Rab11a and the Rab11-Arf effector FIP3, which assists with GTP hydrolysis on Arf4 [6–8,52]. The Arf4-dependent Stage I of the complex assembly is completed by proofreading of the rhodopsin CTS VxPx coupled to Arf4 inactivation, permitting subsequent rhodopsin uptake into ciliary-targeted rhodopsin transport carriers (RTCs) (Figure 1).

Figure 1.

Model depicting the assembly and conversion of the Arf4-based rhodopsin ciliary-targeting complex at the Golgi/TGN

The assembly of the trafficking complex can be divided into two stages: the Arf4-dependent Stage I and the post-Arf4 Stage II. In Stage I, activated Arf4 recognizes the CTS VxPx of rhodopsin. Arf4, rhodopsin, and Rab6 cooperate to localize and activate the Arf GEF GBF1 at the trans-Golgi. Through positive feedback, Arf4 is further activated by GBF1. Coincidence detection of activated Arf4, acidic phospholipids and PIP₂ recruits the Arf GAP ASAP1 to the Golgi/TGN, where it forms a complex with rhodopsin and Arf4, and selectively binds Rab11 and FIP3. Proofreading of CTS VxPx that regulates rhodopsin uptake into RTCs is coupled with GTP hydrolysis on Arf4 by ASAP1, assisted by FIP3, and followed by the departure of inactivated Arf4. In Stage II, ASAP1 serves as a platform for the assembly of the Rab11-FIP3-Rabin8 dual effector complex. Following GTP hydrolysis on Arf4, FIP3 departs the complex and the R-SNARE VAMP7 likely replaces it. Concurrently, Rabin8 activates GDP-bound Rab8. Rabin8 is phosphorylated by NDR2 kinase (asterisk), which increases its affinity for the Sec15 component of the exocyst membrane-tethering complex. VAMP7 is phosphorylated by the c-Src within the LD (asterisk) and activated for pairing with Q SNAREs syntaxin3 and SNAP-25, which ultimately drives RTC fusion with the periciliary plasma membrane.

In Stage II of the complex assembly, formation of nascent RTCs is coupled to the assembly of the membrane trafficking machinery responsible for their fusion, directing RTCs to the cilium (Figure 1). Following Arf4 inactivation and dissociation, ASAP1, Rab11a and FIP3 remain associated at the TGN where they recruit the Rab8 GEF Rabin8 [7,8]. This leads to the formation of a distinctive Rab11–FIP3–Rabin8 dual effector complex unique to Rab11, the ramifications of which in membrane trafficking have been reviewed previously [10]. Rab11–FIP3–Rabin8 complex assembles through multiple weak interactions that create a high-avidity complex, thus promoting cooperation between the two effectors in the execution of Rab11-directed functions [8–10]. Whereas Rab11 effector FIP3 stimulates the GAP activity of ASAP1 [52], the Rab11-mediated recruitment of Rabin8, the Rab8-specific GEF, initiates the Rab11-Rabin8–Rab8 ciliogenesis cascade [17–19]. Finally, the Rab11-Rabin8-Rab8 module directs Golgi-to-cilia trafficking, by capturing the RTC R-SNARE VAMP7 through direct binding of Rab8 and Rab11 to the VAMP7 regulatory longin domain (LD), and Rabin8 to its SNARE motif. With the R-SNARE aboard, RTCs are equipped for fusion via VAMP7 pairing with plasma membrane Q SNAREs syntaxin 3 and SNAP-25 [22].

The role of Arf4 in the ciliary membrane-targeting complex

Arf4 was identified as a regulatory protein that specifically binds the rhodopsin CTS VxPx [51]. In those experiments, the effects of blocking the action of Arf4 were functionally equivalent to blocking the rhodopsin CTS VxPx, mutations in which cause severe forms of autosomal dominant retinitis pigmentosa (adRP) [51,53,54]. Subsequent studies in transgenic frogs demonstrated that the expression of the Arf4I46D mutant, deficient in GTP hydrolysis by ASAP1, causes dysfunctional rhodopsin trafficking and rapid retinal degeneration [6], verifying a crucial role of the VxPx motif and the Arf4-based targeting complex in ciliary trafficking. Supporting this paradigm for sorting into transport vesicles, a specific CTS and Arf4 provide the directionality and increase the efficiency of ciliary transport of several sensory receptors [16,20,21,55,56]. However, the role of Arf4 in rhodopsin trafficking has been challenged using conditional knockout mouse retinas [57]. In our recently published study [23], we extensively addressed the issues of significance when comparing the data from the two published Arf4 in vivo models, the frog and the mouse, which are briefly summarized here: (i) Absence of a gene vs. a dominant negative action often have different effects, as exemplified by differences between rhodopsin mutant and knockout mouse models [58–62]. (ii) The volume of membrane trafficking in the frog eye exceeds by an order of magnitude that of the rodent rods [63]. (iii) Mouse models do not always recapture retinal membrane trafficking disease phenotype e.g. despite a relatively faithful manifestation of the hearing and balance disorders found in Usher syndrome, none of the Usher 1 mouse models undergo retinal degeneration [64]. Neither the frog nor mouse models are authentic representations of the human eye, but both aid in dissecting disease-related processes. The most likely explanation for the apparent discrepancy between the two models is that in the conditional mouse Arf4 knockout a compensatory mechanism allows rhodopsin trafficking to proceed, perhaps at a suboptimal level, because the mouse model system has low demands on membrane trafficking volumes.

Conversely, by magnifying the role of trafficking through its high volumes of membrane synthesis, the well-established frog model system revealed the stages and molecular machineries involved in the vectorial transport of ciliary cargo. Our research using the frog model has shown that the Arf4 α3 helix, encompassing the IQEAAEELQKML peptide, directly crosslinks to rhodopsin during membrane trafficking [51]. Notably, the human and frog Arf4 are 95% identical and anti-frog Arf4 antibody readily recognizes human Arf4 [8]. While reflecting on a potential compensatory mechanism in use in mouse Arf4 KO, we discovered that mouse is a natural mutant in the crucial Arf4 α3 helix (Figure 2). This is significant because the VxPx CTS of rhodopsin is surrounded by the conserved positively charged residues on the cytoplasmic surface of rhodopsin, which likely interact with the negatively charged residues in Arf4 α3 helix, resulting in an increase in affinity of the Arf4-Rhodopsin interaction probably without providing specificity (Figure 2(a,b). The specificity is likely provided by the essential hydrophobic C-terminal VxPx CTS that could interact with the neighbouring hydrophobic patch of Arf4 (Figure 2(a,b). As shown in Figure 2(c), the alignment of Arf4 sequences from several species reveals high conservation of the amino acid residues comprising Arf4 α3 helix. Compared to the frog, the human Arf4 has only conservative substitutions, whereas the bovine sequence is identical to that of the frog. In mouse and rat two acidic amino acid residues are changed to hydrophobic amino acids, creating a hydrophobic stretch instead of a negatively charged one (Figure 2(b)), which potentially conveys dissimilar functional advantage to rodent Arf4. This type of substitution, typically used in mutagenesis experiments to abolish protein-protein interactions, likely disrupts rhodopsin-Arf4 interactions. In fact, there is high likelihood that the mouse Arf4 either does not interact with rhodopsin, or does so with diminished affinity.

Figure 2.

Arf4 α3 helix, which participates in rhodopsin interactions, is evolutionary conserved, with the exception of mouse and rat

(a) The crystal structure of bovine rhodopsin (PDB code 1F88) is shown in orange cartoon representation. Residues forming a basic patch near the C-terminal tail are shown in stick representation and labelled according to the bovine sequence. The position of the C-terminal QVAPA tail is indicated (note that the linker connecting the C-terminal tail to the rhodopsin core domain was not resolved in the crystal structure). (b) Crystal structure of human Arf4 (PDB code 1Z6X) is shown as a green cartoon representation and that of mouse Arf4 (modelled based on the human Arf4 structure) as a cyan cartoon representation. GDP (sticks) and a magnesium ion (pink ball) mark the GTPase active site of Arf4. Residues forming the acidic patch on human Arf4 and neighbouring hydrophobic residues are shown as sticks and labelled according to the sequence. (c) Sequence alignment of the region corresponding to helix 3 of Arf4 and Arf5 and the surrounding residues, from different species. The acidic residues substituted with hydrophobic residues in mouse and rat Arf4 are boxed.

Interestingly, unlike Arf4, Arf5 is 100% conserved within the species compared in Figure 2(c). Mouse Arf5 is a possible orthologue of human and frog Arf4, as it possesses the acidic residues potentially important for interactions with rhodopsin (Figure 2(c)). Arf4 and Arf5 could be interchangeable in mouse photoreceptors, but not in cells with high volumes of cargo transiting through the secretory pathway [57], where directionality and efficiency are of essence. Notably, GBF1 has high exchange activity on Arf5 [65] and its activation of both Arf4 and Arf5 initiates an activation cascade at the TGN [66]. If Arf5 indeed functions in mouse rod photoreceptors as the human and frog Arf4 orthologue, the distinct downstream functions of Arf4 vs. Arf5 may influence the subsequent steps in rhodopsin trafficking. They may consequently diverge from the highly conserved Arf4-mediated pathway that includes Rab11 and Rab8, which were also reported to be dispensable in the mouse KOs [67]. In contrast to that study, our study in transgenic frogs showed that the Rab8 T22N dominant-negative mutant, deficient in GTP binding, caused dysfunctional rhodopsin trafficking and rapid retinal degeneration, demonstrating an essential role for Rab8 in rhodopsin trafficking [12], consistent with the role of Rab8 as a regulator of carrier fusion with the periciliary plasma membrane, which is dysfunctional in human genetic diseases and ciliopathies [13–15,30,68–70]. These examples indicate that mouse knockout models addressing intracellular processes regulated by Arf and Rab GTPases are subject to further verification that can be accomplished by selecting for such studies the animal models in which these processes are conserved.

Activation of Arf4 by the Arf GEF GBF1

The spatiotemporally restricted activation of Arf4 through GTP binding is crucial to its function in complex assembly; however, available information about the signals that Arf GEFs recognize in order to activate Arfs is still scarce. These signals were examined in a recent study, which established that sensory receptor cargo, such as rhodopsin, directly promotes its own intracellular progression to the cilia by providing input to the specific large Arf GEF, GBF1, to activate a cognate Arf, Arf4, in ciliary trafficking [23]. In the likely scenario, this process starts with multiple random activation events that initially generate a small number of active Arf4 clusters at the Golgi/TGN. Next, a functional ternary complex – sensitive to the GBF1 inhibitor Golgicide A – is formed between GBF1, activated Arf4, and the ciliary cargo, rhodopsin [23], further activating Arf4 through the operation of an autocatalytic amplification mechanism (positive feedback) [71], building up levels of active Arf4, which leads to the assembly of the ciliary-targeting complex. The emergence of newly synthesized rhodopsin in the Golgi is essential for this process, as blocking of rhodopsin transit through the Golgi by cycloheximide abolishes all interactions between GBF1, Arf4 and rhodopsin, indicating the requirement of cargo influx for the complex formation in vivo [23].

GBF1 belongs to the BIG/GBF family of Golgi-localized large Arf GEFs, which comprises multifunctional scaffold proteins that contain the highly conserved Sec7 domain involved in the nucleotide-exchange activity. Arf GEFs are auto-inhibited in the cytosol [72]. Their membrane association and catalytic activity are controlled by cooperative allosteric regulation via coincidence detection by multiple regulatory domains: DCB (dimerization and cyclophilin binding) domain, a HUS domain (homology upstream of Sec7 domain) and HDS (homology downstream of Sec7) domains, which integrate direct inputs from membranes and multiple activated Arfs and Rabs [34,36]. For example, the prototypic Arf GEF, the yeast Sec7, is differentially regulated by two Arfs (Arl1 and Arf1) and two Rabs (YPT1/Rab1 and YPT31/32/Rab11), interacting with different regulatory domains and collectively mediating its TGN localization and allosteric activation [73–76].

It has been reported that both GBF1 and Arf4 function within the early Golgi, and at the TGN [6,7,44,77–83]. The Golgi localization of GBF1 is mediated by the HDS1 and HDS2 domains that appear to act as PIP-binding domains [84–87]. In photoreceptors, the N-terminal DCB-HUS domain of GBF1 directly interacts with Arf4 and with newly synthesised rhodopsin [23]. Notably, the DCB domain of GBF1 also directs Golgi localization through interactions with Golgi-associated Rabs, like Rab1b [88,89]. In support of this mode of action, our recent study revealed that in photoreceptors GBF1 nearly completely colocalizes with Rab6 at the trans-Golgi [23]. We thus hypothesized that Rab6 may control the trans-Golgi membrane association of GBF1.

Interaction of the Arf GEF GBF1 with the small GTPase Rab6

Rab6 is one of the most conserved Rab GTPases throughout evolution and the most abundant Rab protein associated with the Golgi complex. The two ubiquitous isoforms, Rab6A and Rab6A’, regulate transport in and out of the Golgi, including anterograde transport between the Golgi and the plasma membrane [90–92]. Rab6 was the first and the most abundant Rab GTPase identified as a potential regulator of rhodopsin trafficking [93]. Surprisingly, Rab6 function in rhodopsin trafficking is conserved in Drosophila photoreceptors as well, despite many differences between invertebrate and vertebrate photoreceptor cells [94–96]. The downstream effectors of Rab6 in photoreceptors are not known, thus we tested the notion that activated Rab6 controls trans-Golgi membrane association of GBF1. We discovered that GBF1 overlaps with Rab6 at the trans-Golgi significantly better than with GM130 at the cis-Golgi, as revealed by pixel colocalization analysis performed within the Golgi [23]. We thus examined their direct interaction and now demonstrate that GTPγS-activated Rab6, and the Rab6Q72L mutant directly and specifically interact with the DCB-HUS domain of GBF1 (Figure 3(a–c)). We additionally show, using an established retinal cell-free system [97,98], that the peptide mimicking the Rab6 switch 1 domain responsible for recruiting specific effectors (AA 37–52) [99] completely inhibits rhodopsin trafficking in vitro (Figure 3(d)). Peptides corresponding either to the highly homologous domain of Rab7 (AA 33–47), or to the Rab6 SF3 domain that is recognized by the Rab escort protein (REP) involved in prenylation [100] (AA 107–120)(Figure 3(e)) have no effect. Thus, a relevant binding partner and effector of Rab6 at the Golgi/TGN is another important regulator involved in RTC budding. We suggest that it is GBF1, which was previously reported as one of the top interactors of active Rab6 [101]. Several other known effectors link Rab6 to the molecular motors [102,103], including Myosin II, which, along with KIF20A, regulates fission of Rab6-positive carriers and exit from the Golgi [104,105]. Rab6 interacts specifically with the transport protein particle II (TRAPPII) complex, which is essential for ciliogenesis through its interaction with Rabin8 and activation of Rab11 [19,101,106]. One of the Rab6 effectors, Rab6IP1, directly links Rab6 and Rab11, whereas other effectors enable Rab6-postive carriers to acquire Rab8, which regulates their fusion [107–110]. In this context, photoreceptor Rab6 plays a crucial role in the Golgi exit of membrane cargo directed to the primary cilium, in part by concentrating GBF1 at Golgi exit sites where it could both sense the emerging cargo, and recruit and activate Arf4 for the ciliary cargo delivery via the Rab11-Rabin8-Rab8 pathway.

Figure 3.

Assembly of the Arf4-mediated ciliary membrane-targeting complex at the trans-Golgi: the Arf GEF GBF1 is an effector of Rab6

(a) Schematic of GBF1. DCB-HUS (AA 1–710) and Sec7-HDS1 (AA 695–1066) are indicated. (b) GST-DCB-HUS, or GST (Ctrl), were incubated with recombinant human Rab6 bound to GTPγS, or with Rab6Q72L or T22N mutants. Bound Rab6 was detected by immunoblotting. The GST fusion proteins were detected with anti-GST antibody. Arrowheads point to the GST-fusion proteins used in pulldowns. (c) GST-DCB-HUS, GST-Sec7-HDS1, or GST (Ctrl), were incubated with Rab6 bound to GTPγS, or with Rab6Q72L mutant and bound Rab6 was detected as above. (d) Frog retinas were pulse-labelled for 60 min and retinal PNS was incubated for 30 min with 50 µM peptides, as indicated in the panel, prior to a 2 h cell-free chase; photoreceptor membranes were fractionated into Golgi, TGN and RTCs as described [6,98], and radiolabelled proteins analysed by SDS-PAGE and autoradiography ([35S]-Rh). The Rab6 effector peptide mimicking switch I arrested rhodopsin in the Golgi, where Rab6 and GBF1 are localized [23,93] and potently and specifically inhibited its uptake into RTCs. The Rab7 switch I and the Rab6 SF3 peptide had no effect. (e) The ribbon model of Rab6. Switch I is coloured blue, and switch II is red. Modified from ref [110]. SF3 domain is indicated.

RTC fusion: Rabs and their effectors provide framework for SNARE assembly

Activation of Rab8 by Rabin8 renders nascent RTCs competent for fusion with the periciliary plasma membrane, because Rab8 regulates the final stages of polarized membrane traffic, fusion of ciliary-targeted carriers and ciliogenesis [7,11–15,30,68]. Phosphorylation of T72 of the switch II region of Rab8 by the Parkinson’s disease kinase LRRK2 results in a failure of Rabin8-mediated activation of Rab8 [111] and consequently inhibits cilium formation [112,113]. Rabin8 is phosphorylated at S272 by NDR2 kinase (also known as STK38L), which regulates the switch in binding specificity of Rabin8 from PS to the Sec15 component of the exocyst complex that mediates carrier tethering at the periciliary plasma membrane [114]. Mutations in NDR2 affecting Rabin8 function cause canine retinal degeneration corresponding to human ciliopathy Leber congenital amaurosis (LCA) characterized by early-onset blindness [115,116], highlighting the central role of the Rab11-Rabin8-Rab8 pathway.

The fusion event through which RTCs deliver rhodopsin to the ciliary base is driven by the pairing of plasma membrane SNAREs syntaxin 3 and SNAP-25, regulated by omega-3 docosahexaenoic acid (DHA) [117]. The long sought after RTC R-SNARE that pairs with syntaxin 3 and SNAP-25 is now identified as VAMP7 or TI-VAMP [22]. Overall, VAMP7 is known to regulate fusion of transport carriers with the plasma membrane; plasma membrane expansion and neurite outgrowth; Golgi homeostasis and Golgi-to-plasma membrane transport; selective apical exocytosis and polarity, as well as ciliogenesis [118–128]. VAMP7 possesses a N-terminal regulatory LD, which is generally involved in binding small GTPases [129–131]. The closed auto-inhibited conformation of VAMP7, in which LD folds back onto the SNARE motif, is stabilized in part by tyrosine 45 in the LD [119,132]. Vps9-ankyrin repeat protein/Ankrd27 (VARP) stabilizes VAMP7 LD in the closed conformation substantially diminishing SNARE complex formation [133]. Phosphorylation at tyrosine 45 by c-Src kinase activates VAMP7 Q-SNARE binding and complex assembly [134]. In photoreceptors, c-Src phosphorylation occurs at the Golgi, indicating that VAMP7 is trafficking to the cilium in its activated form [22].

During RTC formation and trafficking, VAMP7 interacts with the Rab11-Rabin8-Rab8 trafficking module. Rab11 and Rab8 bind VAMP7 LD, whereas Rabin8 interacts with the SNARE domain [22]. FIP3 regulates VAMP7 access to Rab11, as the interactions between FIP3 and VAMP7 with Rab11 are mutually exclusive. Rabin8 and VARP directly interact in ciliary trafficking of VAMP7, and their respective affinities for VAMP7 may be regulated by its c-Src phosphorylation and activation. To better understand interactions of VAMP7 with its numerous partners, we wanted to visualize the topography of the putative Rab11-Rabin8-VAMP7 complex. We modelled it by superposition of the structure of the SRP receptor [130] as a model for small GTPase-LD interaction, onto the structure of Rab11-Rabin8-FIP3 complex [9] and VAMP7-VARP complex [133] (Figure 4). This model predicts that VAMP7 LD binds to the canonical effector-binding site of Rab11a and overlaps with FIP3. Strikingly, if VAMP7 bound to Rab11-Rabin8 complex is in the inactive closed conformation, in which the VAMP7 SNARE domain folds over the LD [133], the arginine (R150) at the zero layer would point straight towards and clash with the most C-terminal α5 helix of Rabin8 [22]. We postulate that phosphorylation by c-Src kinase induces conformational change in VAMP7 to remove the potential steric hindrance and allow for binding to Rabin8. This, in turn would enable the Rab11-Rabin8-Rab8 module to capture VAMP7 for delivery to the ciliary base, and subsequent pairing with the cognate SNAREs syntaxin 3 and SNAP-25, to regulate the fusion event through which RTCs deliver rhodopsin in the final stages of ciliary-directed trafficking.

Figure 4.

A model of the putative Rab11-Rabin8-VAMP7 complex

Model was generated with PyMOL, using the structure of the SRP receptor (PDB code: 2FH5) [130], Rab11-Rabin8-FIP3 complex (PDB code: 4UJ3) [9], and VAMP7–VARP complex (PDB code: 4B93) [133]. Positions of FIP3 and VARP in the crystalized complexes are indicated in light gray. VAMP7 SNARE domain, LD, and residues Y45 and R150 are indicated. Modified from reference [22]. Conformational change of VAMP7 associated with c-Src phosphorylation at Y45 is predicted to rearrange the SNARE domain and reposition R150 at the zero layer, thus optimizing binding of activated VAMP7 to Rabin8.

Concluding remarks

Communication between Arf and Rab GTPases via regulatory protein and effector networks ultimately leads to the directional delivery of ciliary membrane cargo. Ciliary sensory receptors, such as rhodopsin, are active participants in these processes through direct interactions with select components of the ciliary trafficking complexes. While these interactions are highly evolutionary conserved, the species conservation is unclear in mouse models of ciliary trafficking in photoreceptor cells. Future research will reveal the level of conservation of the Arf4-based ciliary-targeting complex in the assembly of the membrane trafficking machinery fundamental to the biogenesis of cilia and cilia-derived sensory organelles, and its involvement in ciliopathies and other degenerative diseases.

Funding Statement

This work was supported by the NIH [grant number EY-12421 to DD] and by the Novo Nordisk Foundation [grant number NNF15OC0014164 to EL].

Disclosure statement

No potential conflict of interest was reported by the authors.