CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

The Conserved Bardet-Biedl Syndrome Assemble a Coat that Traffics Membrane Proteins to Cilia

Hua Jin,1 Susan Roehl White,1 Toshinobu Shida,1 Stefan Schulz,2 Mike Aguiar,3 Steven P. Gygi,3 J. Fernando Bazan,4 and Maxence V. Nachury1,* 1Department of Molecular and Cellular Physiology, Stanford University School of Medicine, Stanford, CA 94305-5345, USA 2Institute of Pharmacology and Toxicology, Friedrich-Schiller-University, D-07743 Jena, Germany 3Department of Cell Biology, Harvard Medical School, Boston, MA 02115, USA 4Department of Engineering, Genentech, CA 94080, USA *Correspondence: [email protected] DOI 10.1016/j.cell.2010.05.015

SUMMARY et al., 2001; Nachury et al., 2007). The sorting step most likely relies on the recognition of a ciliary targeting signal (CTS) by a sort- The BBSome is a complex of Bardet-Biedl Syndrome ing complex and CTSs have been identified in several ciliary (BBS) proteins that shares common structural membrane proteins (Tam et al., 2000; Berbari et al., 2008b; Follit elements with COPI, COPII, and clathrin coats. Here, et al., 2010). While the GTPases Arf4 and Rab8 have been shown we show that the BBSome constitutes a coat complex to recognize the CTSs of rhodopsin and fibrocystin, respectively that sorts membrane proteins to primary cilia. The (Mazelova et al., 2009; Follit et al., 2010), it is expected that coat BBSome is the major effector of the Arf-like GTPase complexes resembling COPI, COPII, and clathrin carry out the sorting of membrane proteins to cilia. Canonical coat complexes Arl6/BBS3, and the BBSome and GTP-bound Arl6 are recruited to membranes by phosphoinositides (PIPs) and—in colocalize at ciliary punctae in an interdependent GTP most cases—by a GTP-bound Arf family GTPase and the direct manner. Strikingly, Arl6 -mediated recruitment of recognition of sorting signals by coat complexes ensures that the BBSome to synthetic liposomes produces distinct coat polymerization packages transmembrane cargoes into patches of polymerized coat apposed onto the lipid a carrier vesicle (McMahon and Mills, 2004). Thus far, no coat bilayer. Finally, the ciliary targeting signal of somato- complex has been identified for trafficking to cilia. statin receptor 3 needs to be directly recognized Recently, we discovered the BBSome, an octameric complex by the BBSome in order to mediate targeting of consisting of the seven highly conserved Bardet-Biedl syndrome membrane proteins to cilia. Thus, we propose that (BBS) proteins BBS1, BBS2, BBS4, BBS5, BBS7, BBS8, and trafficking of BBSome cargoes to cilia entails the BBS9 and of the novel protein BBIP10 (Nachury et al., 2007; coupling of BBSome coat polymerization to the Loktev et al., 2008). Bardet-Biedl Syndrome is an autosomal recessive disorder characterized by retinal degeneration, poly- recognition of sorting signals by the BBSome. dactyly, kidney cysts, and obesity that can be caused by muta- tions in any of 14 known and whose etiology is associated INTRODUCTION with dysfunction (Fliegauf et al., 2007). Since the BBSome binds Rabin8 and associates with the ciliary membrane and Primary cilia are microtubule-based projections found on nearly since BBS5 binds PIPs on protein-lipid overlays, we have pro- every cell in the human body. Since primary cilia are required for posed that the BBSome functions in vesicular trafficking to the phototransduction, olfaction, planar cell polarity, and Hedgehog cilium (Nachury et al., 2007). In support of this hypothesis, the signaling and since the receptor for each of these signaling path- G protein-coupled receptor (GPCR) Somatostatin receptor 3 ways has been localized to the primary cilium (Fliegauf et al., (SSTR3) fails to reach the cilium of hippocampal neurons in 2007), the targeting of signaling receptors to cilia is thought to and bbs4 knockout mice (Berbari et al., 2008a). However, be crucial for signal sensing and transduction. Yet, our under- a role for the BBSome in vesicular transport remains controver- standing of membrane protein targeting to cilia remains frag- sial, with alternative roles in microtubule anchoring (Kim et al., mentary (Nachury et al., 2010). 2004), intraflagellar transport (Ou et al., 2005; Lechtreck et al., The delivery of membrane proteins to cilia sequentially entails 2009), and ubiquitination (Gerdes et al., 2007) having been sorting and packaging into carrier vesicles, docking, and fusion proposed. Regardless of the model considered, the definite of vesicles with the base of the cilium and intraflagellar transport molecular activity of the BBSome remains unknown. (IFT) from cilia base to cilia tip (Rosenbaum and Witman, 2002). Interestingly, BBS3 encodes the small Arf-like GTPase Arl6, The step of docking and fusion requires the GTPase Rab8 and which is not part of the BBSome and whose function remains un- its guanine nucleotide exchange factor (GEF) Rabin8 (Moritz characterized (Fan et al., 2004; Chiang et al., 2004). We now show

1208 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. A Eluate Figure 1. The BBSome Is the Major Effector C FT Eluate of Arl6 in Retinal Extracts GTP GTP GDP (A) Arl6 specifically captures the BBSome from

GTP GDP GTP GDP retinal extracts. Bovine retinal extracts were Load Arl6 Arl6 GST GTP GTP MW loaded onto GST-Arl6DN16[Q73L] (‘‘Arl6 ’’), Load Arl6 Arl6 GST Arl6 Arl6 GST (kDa) GST-Arl6DN16GDP (‘‘Arl6GDP’’), or GST columns. * Eluates were resolved by SDS-PAGE and silver 260 BBS4 stained. Eight bands unique to the Arl6GTP eluate 160 1 eq. 2 eq. (red dots) were excised, and protein identification carried out by mass spectrometry revealed these 110 • • BBS9 eight proteins to be the eight subunits of the 80 BBS7 160 BBSome (listed on the right side of the gel). • • BBS2 D 110 (B) The BBSome is one of only two protein entities 60 BBS1 80 GTP • • BBS8 recovered by Arl6 chromatography. Direct • • BBS4 60 50 ‘‘in-solution’’ mass spectrometry analysis of the GTP 50 Input Arl6 eluate identified 186 proteins each repre- 40 • • BBS5 α sented by 1 to 136 peptides. One hundred 40 ( -Myc blot) seventy-seven of these proteins were also identi- 30 30 fied by direct analysis of the Arl6GDP eluate and/ or of the GST eluate and therefore represent 20 20 contaminants. The nine proteins identified only in GTP 15 MW the Arl6 eluate are listed in the table together

10 • • BBIP10 (kDa) β prop ct with the total number of peptides and the number

Silver stain BBS1 BBS2 BBS4 BBS5 BBS7 BBS8 BBS9

BBIP10 of unique peptides identified by tandem mass 160 spectrometry for each protein. BBS1 BBS1 Protein Total Unique 110 (C) One volume equivalent of retinal extract (Load), B BBS1 136 41 80 one volume equivalent of each flow-through (FT), BBS9 136 39 60 GTP BBS7 87 31 Arl6 and two volume equivalents of each eluates from BBS4 81 31 50 capture the Arl6 affinity chromatographies were immuno- BBS2 87 26 40 (α-Myc blot) blotted for the BBSome subunit BBS4. Over 75% BBS8 40 13 of BBS4 is depleted by Arl6GTP column, and close BBS5 37 13 30 GTP TACT1 4 4 to 50% of BBS4 is recovered in the Arl6 eluate. BBIP10 2 1 20 The asterisk denotes a nonspecific band. (D) Arl6GTP specifically interacts with the b propeller domain of BBS1. Extracts of HEK293 cells trans- fected with Myc-tagged BBSome subunits or the b propeller (aa 1-430) or C-terminal domains (aa 431–593) of BBS1 were applied to beads decorated with GST-Arl6DN16[Q73L]GTP. Total cell extracts (top) and captured materials (bottom) were immunoblotted for Myc. The BBSome subunit most efficiently recovered on Arl6GTP beads is BBS1, and within BBS1, only the N-terminal domain binds Arl6GTP. that Arl6GTP recruits the BBSome onto membranes to assemble Arl6GTP column and recovered in the Arl6GTP eluate, while no an electron-dense coat and that the BBSome sorts SSTR3 to cilia BBSome binding to Arl6GDP and GST was detected (Figure 1C). by directly recognizing SSTR3’s CTS. Thus, the BBSome consti- Thus, the BBSome is the major Arl6 effector in retinal extracts. tutes a coat complex that sorts membrane proteins to cilia. We further confirmed the BBSome-Arl6GTP interaction by show- ing that BBS1 was the BBSome subunit most efficiently captured GTP RESULTS by Arl6 and by mapping the interaction domain to the N terminus of BBS1 (Figure 1D). Since Arl6 is the only BBS The BBSome Is the Major Effector of Arl6 in Retinal besides the BBSome subunits to be universally conserved in Extracts ciliated organisms, these results tie all of the conserved BBS proteins into two connected biochemical units and suggest a We first sought to identify effectors of Arl6 by affinity chromatog- conserved function for the BBSome/Arl6GTP interaction. raphy. Mutations were introduced into Arl6 to preclude GTP hydrolysis (Q73L) and to limit aggregation of the GTP-bound form (DN16). We chose retinal extract as a starting material Fold Recognition Analyses Reveal Coat-like Structural because of the tremendous rates of membrane protein traf- Elements in the BBSome ficking to cilia in photoreceptors. Remarkably, eight protein The finding that the BBSome is the major effector of an Arf-like bands were recovered specifically in the eluate of the Arl6GTP GTPase suggested that the molecular activity of the BBSome column and were identified as the eight subunits of the BBSome may be related to that of coat complexes. We therefore set out (Figure 1A). Further, direct ‘‘in-solution’’ mass spectrometry to validate the BBSome coat hypothesis at the biochemical, analysis of the eluates failed to identify any Arl6 effector besides structural, and functional levels. First, we explored the structural the BBSome subunits. TACT1, the only other protein specifically anatomy of the BBSome using sensitive structure-prediction recovered in the Arl6GTP eluate, binds directly to the BBSome algorithms. We extended previous findings (Kim et al., 2004) and likely binds to Arl6GTP indirectly (Figure 1B; T.S. and and discovered that BBS4 and BBS8 are almost entirely M.V.N., unpublished data). Furthermore, immunoblotting comprised of TPR repeats (Jı´nek et al., 2004) and are therefore showed that over 75% of the BBSome was depleted by the predicted to fold into extended rod-shaped a solenoids

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1209 A Figure 2. The BBSome and Canonical Coat Complexes Share a Related Structural Organization (A) Schematics of the predicted domain organiza- tion of each BBSome subunit. BBS4 and BBS8 have 13 and 12.5 tetratricopeptide repeats (TPR), respectively. BBS1, BBS2, BBS7, and BBS9 each consist of a b propeller followed by an amphi- pathic helical linker and a g-adaptin ear domain B (GAE). In BBS2, BBS7, and BBS9, the GAE is fol- lowed by an a/b platform domain and a helix domain. In BBS1, a four-helix bundle is inserted between the second and third blades. BBS5 contains two pleckstrin homology (PH) domains and a three-helix bundle, while BBIP10 is pre- dicted to fold into two a helices. (B) Structure-based alignment of the proposed GAE modules of BBS1, BBS2, BBS7, and BBS9. C The structures used to generate the model of the GAE domain of BBS1 were those of GGA1, AP2g1, AP2b2, AP2a, and gCOP. Only the N- and C-terminal six b strands of the Ig-like folds (respectively labeled A–C and E–G) are shown, in a color gradient that matches the chain topology of the human BBS1 GAE domain model. While D the primary sequence identity between solved or predicted structures remains quite low, hydro- phobic positions (highlighted) are well conserved. See the Extended Experimental Procedures for details on fold recognitions and secondary struc- ture predictions. (C) The platform-like modules of BBS2, BBS7, and BBS9 are topological variants of the appendage domains. The C-terminal platform domains of gCOP, bCOP, AP2b2, and AP2a were structurally aligned to the proposed platform-like modules of BBS2, BBS7, and BBS9 as above. b strands are shown as arrows (labeled I–M) and helices as cylin- ders (labeled 1–3), color matched to the AP2a structure and the modeled fold of the human BBS7 platform-like module. Intriguingly, b strand H and helix 1 are missing from BBS2/7/9, but in turn these gain an additional, conserved C-terminal b strand (labeled N) that is predicted to form an edge strand (gray) in the platform b sheet. (D) Recurring membrane recruitment machinery and structural elements of the canonical coats and of the BBSome. The PIPs that participate in coat complexes binding to membranes in vitro are listed in orange. The Arf-like GTPases that recruit the coat complexes to membranes are listed in black. The b propeller, a sole- noid, and appendage domains are listed in blue, red, and green, respectively (McMahon and Mills, 2004).

(Figure 2A) . Meanwhile, BBS1, BBS2, BBS7, and BBS9 share Arl6 Is Found in Punctae at the Ciliary Membrane a related b propeller fold (Chaudhuri et al., 2008) in their N termini Together with the BBSome (Figure 2A) and a domain distantly related to the immunoglobulin We next sought to identify the compartment(s) where the (Ig)-like b sandwich fold of the g-adaptin ear (GAE) motif in their C BBSome/Arl6GTP interaction takes place. We raised a polyclonal termini (Figure 2B). In BBS2, BBS7, and BBS9, the GAE domain is antibody against Arl6 and validated its specificity by immuno- further accompanied by an a/b platform domain (Figure 2C). In blotting lysates of RPE1-hTERT (RPE) cells transfected with several clathrin adaptors and in COPI, the GAE motif—either by Arl6 small interfering RNA (siRNA) (Figure S1A available online). itself of fused to the a/b platform—constitutes the so-called RPE cells grow a primary cilium when switched into quiescence, appendage domain that recruits either regulators of coat and we previously showed that the BBSome subunits BBS4 and assembly or factors that program the coated vesicle for subse- BBIP10 localize to cilia in RPE cells. We extended these findings quent targeting events (McMahon and Mills, 2004). In the by showing that endogenous BBS1 (Figure 3C), BBS2 BBSome, Rabin8 binds to the C terminus of BBS1, and it is (Figure S2A) and BBS5 (Figure S2B), all localized to cilia. Thus, conceivable that Rabin8 serves as a BBSome accessory factor. we conclude that the BBSome is present in mammalian cilia as Since rigid a solenoids and b propellers form the architectural a complex as was recently shown in Chlamydomonas (Lechtreck scaffolds of COPII and clathrin cages (Stagg et al., 2007), the et al., 2009). Remarkably, our anti-Arl6 antibody stained cilia abundance of b propellers, a solenoids, and appendage domains (Figure 3A), and cilia staining was lost after siRNA-mediated inside the BBSome suggests an ancient evolutionary relationship depletion of Arl6 (Figure 4A, left panels). To accurately determine between the BBSome and canonical coat complexes (Figure 2D). the distribution of Arl6 and the BBSome within cilia, a structure

1210 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. we found that the BBSome subunits BBS1 and BBIP10 failed to A acTub Arl6 Arl6 acTub DNA localize to cilia when we depleted Arl6 by siRNA (Figures 4A– 4C). While treatment with two different siRNAs targeting Arl6 dramatically decreased the abundance of Arl6 protein, the abun- dance of the BBSome subunit BBS4 remained unaffected (Figure 4B), and BBS4 still migrated as part of a 500 kDa complex B on size-exclusion chromatography in the absence of Arl6 Arl6 acTub Arl6 (Figure S3A). Thus, Arl6 is specifically required for BBSome local- ization to cilia but not for BBSome assembly. Next, we generated clonal RPE cell lines expressing moderate levels of Arl6 variants to determine whether GTP binding and hydrolysis by Arl6 were acTub required for targeting the BBSome to cilia (Figure 4Eand Arl6 DNA Figure S3B). While Arl6-GFP and Arl6[Q73L]-GFP were both found C inside cilia, Arl6[T31R]-GFP, a variant deficient in GTP but not GDP binding (Kobayashi et al., 2009), was absent from primary cilia BBS1 BBS1 Arl6 acTub BBS1 Arl6 (Figure 4D). Toassess the contribution ofArl6-GFPto BBSome tar- Arl6 acTub geting to cilia, we selectively depleted endogenous Arl6 using an DNA siRNA targeting the 30 untranslated region (UTR) of Arl6 messenger RNA (mRNA) (Figure 4E). While the localization pattern of the Arl6- GFP variants remained unaffected, only Arl6-GFP and Arl6[Q73L]- GFP supported BBSome targeting to cilia (Figures 4D and 4F). BBIP10 Arl6 acTub BBIP10 Arl6 BBIP10 Furthermore, measurements of BBS1 immunoreactivity inside cilia Arl6 showed that Arl6[Q73L]-GFP recruited a greater amount of acTub DNA BBSome to the cilium than Arl6-GFP (Figure S3C). We conclude that Arl6 and BBSome targeting to cilia both require Arl6 binding to GTP but not Arl6 GTPase activity. Conversely, depletion of the Figure 3. Arl6 Is Found in Punctae Together with the BBSome Inside BBSome subunits BBS2, BBS4, and BBS5 resulted in a dramatic Primary Cilia decrease of Arl6 staining within cilia (Figure S4A). (A) Arl6 localizes to the primary cilium in RPE cells. Serum-starved RPE cells Interestingly, we noted that the fraction of cells with BBSome- were immunostained for Arl6 and acetylated a-tubulin (acTub). Scale bars or Arl6-positive cilia varied from 15% to 60% depending on the represent 5 mm. experiment (compare Figures 4C and 4F and Figure S4A). While (B) Arl6 is found in punctae flanking the inside primary cilia. RPE cells the source of the variability remains unknown, we hypothesized immunostained as in (A) were imaged by structured illumination microscopy. that the levels of Arl6 and the BBSome in most cilia fall below The scale bar represents 5 mm. A 3D view of the same cilium is shown in Movie S1. the detection threshold of our traditional immunofluorescence (C) Arl6 colocalizes with the BBSome inside cilia. Serum-starved RPE cells protocol. We therefore developed a method that decreases stably expressing GFP-tagged Arl6 were immunostained for BBS1 (top), background staining while preserving the signals in cilia (see BBIP10 (bottom), and acetylated a-tubulin (acTub). GFP-Arl6 was visualized the Extended Experimental Procedures and Figures S4C–S4G), in the green channel without antibody staining. Scale bars represent 5 mm. and we now find that nearly every RPE cilia stains positive for A line scan through these cilia is shown in Figures S1B and S1C. See also Arl6 and BBS1 (Figure S4G). This increase in the proportion of Figure S2. Arl6- and BBS1-positive cilia is not simply an artifact of the new staining procedure, since Arl6 and BBS1 staining are still lost whose 300 nm diameter cannot be resolved by conventional light from cilia when Arl6 is depleted by siRNA. Together, these results microscopy, we resorted to structured illumination microscopy, demonstrate the interdependence of Arl6GTP and BBSome tar- a ‘‘super-resolution’’ technique that lowers the optical resolution geting to cilia and suggest that Arl6GTP and the BBSome syner- to less than 50 nm (Schermelleh et al., 2008). Arl6 staining gize in binding to the membrane of the cilium. appeared in a pattern of punctae flanking the microtubule axoneme that likely correspond to small membrane-associated Arl6GTP Is Necessary and Sufficient to Efficiently Recruit patches (Figure 3B and Movie S1). Further, deconvolution the BBSome to Liposomes microscopy could resolve a discrete pattern of Arl6 staining We first tested whether Arl6 behaves like Arf1 and Sar1, i.e., binds within cilia that precisely mirrored the distribution of the BBSome to membranes upon GTP binding by exposing an amphipathic subunits BBS1 and BBIP10 (Figure 3C and Figures S1B and helix that inserts itself in the lipid shell and terminates in a basic S1C). Thus, the interaction between Arl6 and the BBSome may collar that interacts with phospholipid headgroups (Gillingham take place within these intraciliary patches. and Munro, 2007). Helical representation of the N terminus of Arl6 demonstrates the amphipathic nature of the N terminal helix Arl6GTP and the BBSome Are Dependent on One Another of Arl6 and its termination with several positively charged residues for Targeting to Cilia (Figures 5A and 5B). Since Arl6 is predicted not to be myristoy- Together with our biochemical data, the colocalization studies lated (Gillingham and Munro, 2007), we expressed Arl6 in bacteria suggested that Arl6GTP may recruit the BBSome to cilia. Here, and purified it to homogeneity (Figure S5A). As mammalian cilia

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1211 Figure 4. Arl6GTP Targets the BBSome to A Arl6 acTub DNA BBS1 acTub DNA BBIP10 acTub DNA B ORF UTR Primary Cilia siControlsiArl6siControlsiArl6 (A) Arl6 is required for BBSome localization to cilia. Arl6 RPE cells treated with control siRNA or siRNA tar- geting the coding sequence (ORF) of Arl6 were BBS4 serum starved and immunostained for Arl6, siControl Actin BBS1, or BBIP10. Cilia were visualized by staining C with acetylated a-tubulin (acTub) antibody. Insets 20 show an enlargement of the Arl6, BBS1, or siControl 15 BBIP10 channel in the region around the cilium. siArl6ORF 10 (B) Protein extracts from cells treated with siRNAs

ORF targeting the Arl6 ORF, the Arl6 30 UTR or control 5 siRNA were immunoblotted for Arl6, BBS4, and siArl6 % positive cilia 0 actin (loading control). Arl6 protein is depleted by Arl6 BBS1 BBIP10 both siRNAs targeting Arl6, while BBS4 protein cilium marker levels remain unaffected. D RPE-hTERT Arl6-GFP Arl6[Q73L]-GFP Arl6[T31R]-GFP (C) RPE cells were prepared and stained as in (A). At least 150 cilia per experiment were counted, and the percentage of Arl6-, BBS1-, or BBIP10- positive cilia was plotted. Error bars represents standard deviations (SDs) between three indepen- dent experiments. siControl AcTub (D) GTP-binding but not hydrolysis by Arl6 is required for BBSome targeting to cilia. RPE, BBS1 RPE-[Arl6-GFP], RPE-[Arl6(Q73L)-GFP], RPE- [Arl6(T31R)-GFP] clonal cell lines treated with Arl6-GFP control siRNA or siRNA against Arl6 30 UTR were

UTR serum starved and immunostained for BBS1 and acetylated a-tubulin (acTub). Arl6-GFP was visual-

siArl6 ized in the green channel without antibody staining. (E) Protein extracts from cells prepared as in (D) were immunoblotted for Arl6. Endogenous Arl6, Arl6- GFP, and a nonspecific band of 70 kDa (double E F 40 35 asterisk) that serves as a loading control are shown. 30 siControl All three panels are extracted from the same expo- RPE-hTERT RPE-hTERT 25 siArl6UTR sure scanshown in Figure S3B.The asterisk denotes -[Arl6-GFP] -[Arl6-GFP] 20 a weak nonspecific band that runs slightly below RPE-hTERTWT Q73LT31RRPE-hTERTWT Q73LT31R 15 Arl6-GFP. Quantitation of signal intensities shows Arl6-GFP * 10 that Arl6-GFP, Arl6[Q73L]-GFP, Arl6[T31R]-GFP Arl6 5 are respectively 7.4-, 5.3- and 2.3-fold more abun- % BBS1-positive cilia 0 dant than endogenous Arl6. When cells are treated ** WT with Arl6UTR siRNA, endogenous Arl6 is depleted Q73L T31R UTR siControl siArl6 RPE-hTERT and Arl6-GFP proteins are slightly upregulated. RPE-hTERT -[Arl6-GFP] (F) BBS1-positive cilia in the experiment shown in (D) were counted. At least 150 cilia were counted for each condition. Error bars represent the SD between microscope fields. See also Figures S3 and S4. cannot be isolated in sufficient quantities to generate a pure lipid as a monodisperse complex devoid of aggregates by rate zonal fraction, we conducted sedimentation assays with liposomes sedimentation (Figure S5B). We then performed sedimentation made from brain lipids (Folch fraction I). Here, we found that assays with purified BBSome, Arl6, guanine nucleotides, and recombinant Arl6 efficiently bound to liposomes in the presence liposomes made from brain lipids and found that efficient binding of the slowly hydrolyzable analog GMP-PNP but not in the pres- of the BBSome to liposomes required Arl6 and GMP-PNP ence of GDP or when the N-terminal amphipathic helix was (Figure 5E). We have thus reconstituted the recruitment of the removed (Figure 5C). We therefore conclude that Arl6 conforms BBSome to membranes from purified components in vitro and to the Arf1/Sar1 paradigm and interacts with membranes through no protein factor other than GTP-bound Arl6 is required for this its N-terminal amphipathic helix when GTP bound. binding. To determine the minimal requirements for BBSome binding to liposomes, we next needed a highly purified BBSome fraction. Phospholipid Requirement for BBSome Binding Such a fraction was obtained by fractionating retinal extract to Liposomes over the Arl6GTP column (Figure 1A) followed by cation exchange Given the robust interaction between Arl6GTP and the BBSome on chromatography (Figure 5D). The purified BBSome was nearly one hand and between Arl6GTP and liposomes on the other hand, free of contaminants as assessed by silver staining and behaved it was conceivable that the BBSome recruitment to liposomes

1212 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 5 GTP 12 9 Figure 5. Arl6 Recruits the BBSome to G D A 1 V 2 + liposome + liposome N M G C Liposomes Made from Pure Lipids 13 8 S L C GMP-PNP GDP GMP-PNP GDP (A) Helical-wheel representation of the N-terminal R 6 SPSPS PSPSPS P 4 L 13 amino acids of Arl6. Hydrophobic, nonpolar resi- L L 11 L 10 dues are clustered on one side of the helix, while L 3 7 charged or polar residues are on the opposite B Arl6 Arl6ΔN side. Nonpolar residues are yellow, polar residues Sar1A 1 MSFIFEWIYNGFSSVLQFLGLYK 23 Arf1 1 MGNIFANLFKGLFGKK 16 are purple, acidic residues are red, basic residues Arl1 1 MGGFFSSIFSSLFGTR 16 liposome pellets are blue, and glycine is gray. The diameter of each Arl6 1 MGLLDRLSVLLGLKK 15 E ++++ Arl6 circle is proportional to the bulk of each residue. ++ ++BBSome MW +++ GDP (B) Sequence alignment of the N terminus from D (kDa) elution BBSome + + + GMP-PNP select Arf/Arl family members. Hydrophobic amino acids are highlighted in dark gray, and basic resi- MW Load Flowthrough14 15 16 17 18 19 20 21 22 116 (kDa) 97 • dues are light green. • 116 • (C) Arl6GTP binds to liposomes through its N ter- 97 • • 66 •BBSome minus. Liposomes (20 mg) made from brain lipids • 66 m D • were incubated with 2 M Arl6 or Arl6 N in the pres- • 45 ence of 100 mM GMP-PNP or GDP in a 100 ml reac- 45 • • tion at 30 C for 1 hr. The reactions were centrifuged 31 at 385,000 3 g for 30 min at 24C and equal 31 ave portions of the resulting supernatants (S) and pellets • Arl6 21 21 (P) were resolved by SDS-PAGE and stained with 14 14 Coomassie. As control for protein precipitation 6 • during the course of the experiment, Arl6 or Arl6DN PC/PE/ were incubated without liposome and processed F PC/PE/PS/PA/PI/ G PS/PA/PI/ 3 PC/PE PI(3,4)P as above. 2 2 2 2 +G+ DP (D) Purification of retinal BBSome. Eluates from the MW GTP MW PI + + GMP-PNP (kDa) PI3P PI4P PI5P PI(3,4)PPI(3,5)PPI(4,5)PPI(3,4,5)P (kDa) Arl6 affinity column were loaded onto a cation 116 97 • exchange column (MonoS), and the BBSome (red 97 • • dots mark subunits) was eluted with a salt gradient. • Fractions were analyzed by SDS-PAGE and silver 66 • BBSome 66 • • staining. BBSome • • (E) The BBSome binds to liposomes in an Arl6- 45 45 • and GTP-dependent manner. Various combina- • tions of Arl6 (0.5 mM), GMP-PNP, or GDP (100 mM) 31 31 were incubated with 4 mg brain lipid liposomes in 21 • Arl6 a50ml reaction at 30C for 30 min. Reactions were 21 • Arl6 14 diluted to 100 ml, supplemented with BBSome 14 Arl6 + BBSome + GMP-PNP Arl6 + BBSome (50 nM final), and returned to 30 C for a further 15 min. Liposomes and bound proteins were sedi- mented at 140,000 3 gave for 30 min at 24 C, and pellets were resolved by SDS-PAGE and stained with silver. Red dots denote BBSome subunits (except for BBIP10 which was not resolved), while the blue dot denotes Arl6. (F) Phosphoinositide specificity of BBSome binding to liposomes. Liposomes (167 mM final lipid concentration) containing 3 mol% of various PIPs were incu- bated with Arl6 (0.25 mM), BBSome (50 nM), and GMP-PNP (100 mM) in a 60 ml reaction before flotation on iodixanol gradients. Liposome-bound proteins were analyzed by SDS-PAGE and silver staining. Although Arl6 binding was similar for all eight liposomes, BBSome binding was maximal when liposomes contained

PI(3,4)P2. Quantification of Arl6 and BBSome binding is shown in Figure S5C. The composition of PC/PE/PS/PA/PI liposomes is as follows: 53% DOPC, 22 mol% DOPE, 1 mol% Texas-Red DHPE, 8 mol% DOPS, 5 mol% DOPA, and 11 mol% DPPI. PIPs were substituted for 3 mol% PI in PIP liposomes. (G) The BBSome requires acidic phospholipids to efficiently bind to liposomes. Liposomes (167 mM final lipid concentration) were incubated with Arl6, BBSome (50 nM), and GMP-PNP or GDP (100 mM) in a 60 ml reaction before flotation on iodixanol gradients. In order to achieve similar recoveries of Arl6 with different liposomes, 1.25 mM Arl6 were used for PC/PE liposomes, while 0.25 mM Arl6 were used for PI(3,4)P2 liposomes. Liposome compositions are as follows. PC/ PE:

88 mol% DOPC, 11 mol% DOPE, 1 mol% Texas-Red DHPE. PC/ PE/ PS/ PA/ PI/ PI(3,4)P2: see (F). See also Figure S5. was strictly indirect and did not involve any contact between the that allows for the efficient capture of COPI, COPII, and exomer BBSome and lipid headgroups. However, COPI and COPII coats coat complexes and that functions in COPI and COPII budding and clathrin adaptors have all been shown to directly contact reactions (Matsuoka et al., 1998; Spang et al., 1998; Wang acidic phospholipids or specific PIPs (Matsuoka et al., 1998; et al., 2006). The phospholipid part of the major mix contains Spang et al., 1998; Bremser et al., 1999; McMahon and Mills, 76 mol% of neutral phospholipids (53 mol% phosphatidylcholine 2004), and those contacts are probably important in the sculpting [PC] and 23 mol % phosphatidylethanolamine [PE]) and 24 mol% of buds and vesicles by the polymerizing coat. To determine of acidic phospholipids (8 mol% phosphatidylserine [PS], 5 mol% whether specific lipids are required for the binding of the BBSome phosphatidic acid [PA], and 11 mol% phosphatidylinositol [PI]). to membranes, we made liposomes from synthetic phospho- The cholesterol/phospholipid molar ratio is 23/77, and all acyl lipids. Given that the lipid composition of mammalian cilia is not chains are oleoyl (18:1) to keep fluidity high through the 4C– known, we started with a base mixture (dubbed ‘‘major mix’’) 30C range of temperatures. To preclude any background signal

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1213 stemming from protein precipitation during the course of the Arl6+BBSome+GMP-PNP incubation with liposomes, we isolated the protein complexes A B bound to liposomes by buoyant density flotation on miniature iodixanol step gradients. Initial experiments showed moderate binding of the BBSome to major mix liposomes in the presence of Arl6GMP-PNP (Figure 5F, lane 1). Since we have previously shown that the BBSome subunit BBS5 binds to PIPs on protein lipid overlays, we replaced a portion of the PI in the major mix with one of seven individual PIPs. While the specificity of BBSome binding for a specific PIP was somewhat variable (Figure S5C), multiphos- Arl6+BBSome+GDP phorylated PIPs [in particular PI(3,4)P2] enhanced BBSome binding to liposomes by as much as 3-fold. This enhanced C D BBSome binding did not result from a general stickiness of PIP liposomes as BBSome binding to PI(3,4)P2 liposomes was strongly Arl6 and GMP-PNP dependent (Figure 5G, lanes 3 and 4, and Figure S5D). We note that recombinant BBS5 exhibits a different PIP specificity on lipid blot overlays (Nachury et al., 2007) than the BBSome does on liposomes and conclude that BBS5 and individual lipids taken out of their physiological envi- ronments may not faithfully recapitulate the specificity of the BBSome complex for lipids in hydrated bilayers. Arl6+GMP-PNP The preference for a lipid bearing a strong negative charge by E F the BBSome suggested that other acidic phospholipids in the major mix might participate in BBSome recruitment to membranes. We therefore tested Arl6- and GMP-PNP-depen- dent binding of the BBSome to liposomes made from neutral lipids. Since PC/PE liposomes were less effective than GMP-PNP PI(3,4)P2 liposomes at recruiting Arl6 , the molarity of Arl6 was adjusted to recover similar amounts of Arl6GMP-PNP regardless of the liposome composition. Despite the 5-fold increase in Arl6 molarity, the binding of Arl6 to liposomes was still strongly dependent upon the addition of GMP-PNP (Figure 5G, Figure 6. Morphological Analysis of Liposomes Incubated with lanes 1 and 2). While the amount of Arl6GMP-PNP recovered re- BBSome, Arl6, and GMP-PNP mained relatively unchanged, BBSome binding was drastically (A) Major mix liposomes (108 mM final lipid concentration) containing 3 mol% reduced in the absence of charges on the bilayer surface PI(3,4)P2 were incubated with Arl6 (0.25 mM), BBSome (50 nM), and GMP- m m (Figure 5G, lanes 1 and 3). Thus, Arl6GMP-PNP binds membranes PNP (100 M) in a 120 l reaction at 30 C for 30 min and processed for thin GMP-PNP section electron microscopy. Distinct coated patches (arrow heads) decorate through neutral and acidic phospholipids and Arl6 and the outer layer of several liposomes. acidic lipids (in particular multiphosphorylated PIPs) synergize (B) Magnified view of a coated patch from (A). Repeat units are highlighted. to recruit the BBSome to membranes. (C) The same liposomes as in (A) were incubated with Arl6 (0.25 mM), BBSome (50 nM), and GDP (100 mM). No coated patches were seen in this preparation. Ultrastructure of Liposomes Incubated with BBSome (D) Magnified view of a liposome from (C). and Arl6GTP (E) The same liposomes as in (A) were incubated with Arl6 (0.25 mM) and GMP- PNP (100 mM). No coated patches were seen in this preparation. We next investigated the morphological consequences of (F) Magnified view of a liposome from (E). BBSome binding to liposomes. Incubation of COPI, COPII, cla- Scale bars in represent 50 nm. See also Figure S6. thrin, and exomer coat components with liposomes leads to the formation of coated profiles and, in the case of COPI and CO- to the exomer coat, no membrane deformation was seen with the PII, the budding of 50 nm diameter coated vesicles (Matsuoka BBSome coat proteins, and coated profiles retained the normal et al., 1998; Spang et al., 1998; Bremser et al., 1999; Wang liposome curvature. When GMP-PNP was replaced with GDP et al., 2006). Liposomes consisted of uni- and multilamellar (Figures 6C and 6D) or when BBSome was omitted (Figures 6E structures with smooth bilayer surfaces when visualized by and 6F), no coated profiles were visible. Thus, upon recruitment thin-section electron microscopy. After incubation with to membranes by Arl6, the BBSome appears to polymerize into BBSome, Arl6, and GMP-PNP, close to 10% of all liposomes an electron-dense coat associated with the liposome surface. showed coated surfaces (Figure 6A). The coated profiles appeared as continuous and well-delineated patches clearly separated from noncoated surfaces (Figure 6A and Figure S6), SSTR3i3 Is an Arl6- and BBSome-Dependent CTS and repeat units could be distinguished at high magnification The polymerization of the BBSome into a membrane coat (Figure 6B), suggesting formation of an ordered polymer. Similar strongly implied that the BBSome sorts specific membrane

1214 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. proteins (i.e., cargoes) inside the cell. A strong candidate for by capture on avidin beads. Remarkably, we found that the levels BBSome cargo is SSTR3, which is lost from cilia in bbs2/ of surface-exposed CD8a-SSTR3i3 remained unchanged in the and bbs4/ hippocampal neurons (Berbari et al., 2008a). We absence of Arl6 or BBS4 (Figure 7H, ‘‘Surface,’’ and Figures extended these results by showing that the number of SSTR3- S7D and S7E). Thus, we conclude that a prototypical BBSome positive cilia decreases dramatically when Arl6 is depleted cargo normally localized in the ciliary membrane accumulates at from cultured hippocampal neurons by lentivirus-mediated short the plasma membrane in the absence of BBSome coat function. hairpin RNA (shRNA) (Figures 7A and 7B). Together, these results establish SSTR3i3 as an Arl6- and The hypothesis that SSTR3 is a bona fide BBSome cargo BBSome-dependent CTS and suggest that the BBSome carries predicts that the BBSome directly recognizes the CTS of out the trafficking of SSTR3i3-containing cargoes from the SSTR3, which was previously mapped to the third intracellular plasma membrane to the ciliary membrane. loop (i3) (Berbari et al., 2008b). We tested this prediction by expressing SSTR3i3 fused to GST in bacteria and conducting DISCUSSION a GST capture assay with purified retinal BBSome. While GST- SSTR3i3 efficiently captured the purified BBSome, GST alone The BBSome as a Planar Coat for Lateral Transport or GST fused to the third intracellular loop of the closely related between Plasma and Ciliary Membranes GPCR SSTR5 failed to recover detectable amounts of BBSome Since the BBSome forms a coat and reads the sorting signals of (Figure 7C). Further, each BBSome subunit expressed in HEK its cargoes to direct them to the cilium, the decision to sort cell bound to SSTR3i3 but not to SSTR5i3 (Figure S7A), indicating membrane proteins such as SSTR3 toward the cilium is almost that SSTR3i3 is not recognized by a single BBSome subunit but certainly made on the compartment where the BBSome coat rather by the holo-BBSome, which assembles around BBSome assembles. Since a synthetic BBSome cargo is detected at subunits expressed in HEK cells (Seo et al., 2010). the plasma membrane in the absence of BBSome function, Sequence analysis of multiple GPCRs targeted to cilia has sug- a plausible model would have BBSome cargoes diffuse laterally gested that the CTS of ciliary GPCRs centers around the in the plasma membrane until their sorting signals become conserved motif AX[S/A]XQ, and mutation of the first and fifth recognized by the BBSome. The ensuing assembly of a planar amino acids of this motif to phenylalanine within SSTR3i3 may BBSome coat would cluster these cargoes into a patch that prevent targeting of a SSTR5-SSTR3i3 chimera to cilia (Berbari can be dragged through the periciliary diffusion barrier sepa- et al., 2008b). Surprisingly, SSTR3i3[AQ-FF] bound to the BBSome rating plasma and ciliary membranes (Nachury et al., 2010). more efficiently than its wild-type counterpart (Figure 7D). We Such a scenario may explain how transmembrane proteins therefore targeted alternative amino acids within the AP[S/A]CQ such as Smoothened enter the cilium by lateral diffusion from motifs of SSTR3i3 (Figure 7E and Figure S7B) and found that the the plasma membrane (Milenkovic et al., 2009). cysteine at the fourth position was the only amino acid required Once inside the cilium, the patch of BBSome coat is predicted for BBSome binding (Figure 7D). To then assess the CTS activity to become transported by the IFT machinery. In nematodes, of the various SSTR3/5i3 variants, we spliced them into the cyto- Chlamydomonas, and human cells, the BBSome has been plasmic tail of the plasma membrane protein CD8a and tran- shown to undergo intraflagellar motility at the same rates as siently expressed the chimeras in IMCD3 cells. While CD8a and known IFT polypeptides, and it has been proposed that the CD8a-SSTR5i3 failed to efficiently target to cilia, CD8a-SSTR3i3 BBSome functions as an adaptor between the IFT complexes was transported to cilia in more than 90% of transfected cells. and IFT cargoes (Blacque et al., 2004; Nachury et al., 2007; Most importantly, the targeting of CD8a-SSTR3i3[C-A] to cilia Lechtreck et al., 2009). The observation that the BBSome was severely impaired compared to CD8a-SSTR3i3 (Figure 7F), assembles a coat in the absence of IFT polypeptides suggests while targeting of CD8a-SSTR3i3[AQ-FF] to cilia remained unaf- that polymerization of a BBSome layer could drive polymeriza- fected (Figure S7B). Thus, the molecular recognition of SSTR3i3 tion of an IFT layer. These BBSome/IFT patches would possess by the BBSome is essential for the full CTS activity of SSTR3i3. one layer of BBSome coat connected to the membrane-bound Finally, we wished to pinpoint the compartment where CD8a- cargoes and one layer of IFT-A and -B complexes bound to SSTR3i3 accumulates in the absence of BBSome or Arl6 func- the IFT motors kinesin II or 1b. The existence of a bilay- tion. To this end, we generated a stable cell line expressing ered IFT/BBSome coat would explain why IFT-A and -B low levels of CD8a-SSTR3i3. While CD8a-SSTR3i3 was targeted complexes fail to maintain cohesion in the absence of BBSome to cilia in >95% of cells treated with a control siRNA, depletion of function (Ou et al., 2005). Arl6 or BBS4 led to a pronounced decrease in ciliary targeting of While Arl6 and the BBSome are generally not required for cilio- CD8a-SSTR3i3 (Figure 7G and Figure S7C). Importantly, the total genesis (see the Extended Discussion), IFT function is universally levels of CD8a-SSTR3i3 remained unchanged by depletion of Arl6 required for cilium formation. A possible explanation resides in or BBS4 (Figure 7H, ‘‘Total’’), thus supporting the interpretation the fact that BBSome only transports a specific set of transmem- that the BBSome coat sorts the synthetic cargo to cilia rather brane proteins to cilia while the IFT complexes are likely required than stabilizes it. Interestingly, examination of CD8a-SSTR3i3 for all transport processes inside cilia. localization in cells depleted of Arl6 or BBS4 revealed significant plasma membrane staining not observed in cells treated with The BBSome as a Canonical Coat control siRNA (Figure S7F). To rigorously test whether CD8a- Unlike COPI and COPII coat formation, BBSome coat assembly SSTR3i3 accumulated at the plasma membrane in the absence in vitro did not sculpt membranes into buds and 50 nm vesicles. of BBSome or Arl6, we conducted surface biotinylation followed While the BBSome may strictly resemble the clathrin plaques

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1215 A AC III SSTR3 DNA ACIII SSTR3 B 70 60 50 40 Vehicle 30 20 10 0 SSTR3-positive cilia (%) Vehicle shArl6

shArl6 Arl6

Actin

i3 D SSTR3 i3 i3 i3 C Input GST SSTR3SSTR5 IInputWT [AQ-FF][Q-A][PQ-AA][CQ-AA][C-A]SSTR5GST BBS4 Myc-BBS2 BBS9 10 BBS8 Ponceau S

10 MW (kDa) 3.5 Ponceau S MW

(kDa) 3.5 E SSTR3i3 VVKVRSTTRRVRAPSCQWVQAPACQRRRRSERRVTR SSTR3i3[C-A] VVKVRSTTRRVRAPSAQWVQAPAAQRRRRSERRVTR SSTR5i3 VVKVKAAGMRVGS------SRRRRSERKVTR

F Glu-tubulin CD8α chimera 100 90 80 70 60

50 40 30 20 targeting to cilia (%) 10 0 CD8α CD8α-SSTR3i3 CD8α-SSTR3i3[C-A] CD8α-SSTR5i3

G Glu-tubulin CD8α-SSTR3i3 100 90 H 80 70 60 siControlsiArl6 siBBS4 50 Surface 40 30 20 Total 10 targeting to cilia (%) 0 siControl siArl6 siBBS4

Figure 7. Direct Recognition of SSTR3i3 by the BBSome Imbues SSTR3i3 with Ciliary Targeting Activity (A) Arl6 is required for SSTR3 trafficking to cilia. Hippocampal neuron cultures were infected with shArl6-lentivirus or control lentivirus at DIV2 (2 days in vitro) and immunostained for adenylate cyclase III (ACIII), a marker of neuronal cilia, and SSTR3 at DIV8. Stacks of ten Z sections were acquired at 0.5 mm interval with a633/1.4 NA objective, deconvolved by constrained iterative, and projected using sum over z axis. Scale bars represent 5 mm. (B) Top: at least 156 cilia were counted for each condition in the experiment shown in (A). Error bars represent the SD between microscope fields. Bottom: protein extracts from hippocampal neuron cultures prepared as in (A) were immunoblotted for Arl6 and actin (loading control). At least 75% of Arl6 protein was depleted upon infection with shArl6-lentivirus compared to control virus. (C) The BBSome directly recognizes the CTS of SSTR3. A MonoS-purified BBSome fraction was mixed with beads coated with GST, GST-SSTR3i3, or GST-SSTR5i3, and the material cleaved off from GST was resolved on SDS-PAGE and probed with antibodies against BBS4, BBS8, and BBS9. The bottom part of the membrane was stained with Ponceau S to show equivalent amounts of fusion moiety cleaved off from GST. Four and a half input equivalents of each eluate were loaded. (D) Generation of a point mutant in SSTR3i3 with decreased BBSome binding. Extracts of HEK cells transfected with Myc-BBS2 were applied to beads decorated with GST-SSTR3i3 variants, GST-SSTR5i3, or GST for capture assays. Total cell extracts (Input) and captured materials were immunoblotted for Myc. The bottom part of the membrane was stained with Ponceau S to show similar amounts of fusion moiety cleaved off from GST. Fifty input equivalents of each eluate were loaded. Low amounts of SSTR3i3[AQ-FF] are recovered after cleavage because GST-SSTR3i3[AQ-FF] is partially degraded inside E. coli. Nonetheless, the large amounts of Myc- BBS2 recovered with SSTR3i3[AQ-FF] demonstrate that SSTR3i3[AQ-FF] binds the BBSome more tightly than SSTR3i3. (E) Sequence of select SSTR3/5i3 variants used in (D) and (F). The segments that differ between SSTR3i3 and SSTR5i3 are highlighted, the conserved AX[A/S]XQ motifs are shown in bold and the mutated amino acids are red. The complete set of sequences is shown in Figure S7B. (F) The CTS activity of SSTR3i3 depends upon intact BBSome binding. CD8a, CD8a-SSTR3i3, CD8a-SSTR3i3[C-A], or CD8a-SSTR5i3 were transiently transfected into IMCD3 cells, and cells were serum starved to induce ciliation. Surface exposed CD8a chimeras were visualized by incubating cells with the OKT8 antibody directed against the extracellular domain of CD8a before fixation. Cilia were visualized by Glu-Tubulin staining. Insets show an enlargement of the CD8a channel in the region around the cilium. Images were acquired with a 633/1.4 NA objective and planes containing cilia are shown. At least 85 ciliated and transfected cells per experiment were counted and the percentages of CD8a chimeras targeting to cilia were plotted. Error bars represents the SD between microscopic fields. Scale bars represent 5 mm.

1216 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. that cluster cargoes at the plasma membrane without deforming somes. Most strikingly, the biochemical output of the BBSome membranes (Saffarian et al., 2009), technical issues may have binding to membranes is the assembly of a thin electron-dense prevented vesicle budding upon BBSome coat formation in our coat on the bilayer surface whose morphology is clearly distinct in vitro system. First, there may be specific soluble factors that from COPI, COPII, clathrin, and exomer coats. While the process assist the BBSome coat in deforming membranes. In the we have recapitulated in the test tube informs the minimal example of clathrin, the large GTPase dynamin and local actin requirements for BBSome coat formation, BBSome coat polymerization are required for the invagination and scission of assembly in vivo must occur with a high degree of spatial and clathrin-coated pits and patches (Koenig and Ikeda, 1996; temporal specificity. First, the loading of GTP onto Arl6 is likely Saffarian et al., 2009). Second, the presentation of sorting to be rate-limiting for BBSome coat assembly in vivo, and an Arl6- signals on the surface of the lipid bilayer and their capture by GEF may locally activate Arl6 to enable BBSome coat assembly the BBSome may be necessary for budding coupled to BBSome in vivo. Second, PCM-1, a major BBSome-associated protein, coat assembly, as is the case for COPI when using liposomes does not copurify with the BBSome on Arl6GTP chromatography. mimicking Golgi membranes (Bremser et al., 1999). Finally, the PCM-1 may therefore prevent BBSome binding to Arl6GTP, and concentrations of BBSome used in the present study (50 nM) removal of PCM-1 from the BBSome may be a prerequisite for may not be sufficiently high to permit the assembly of a BBSome BBSome coat assembly. Third, the requirement for acidic coat competent to deform membranes. COPI-mediated budding lipids—possibly specific PIPs—to efficiently recruit the BBSome was conducted using concentrations approaching 1.5 mMof to membranes in vitro suggests that local lipid composition may coatomer (Bremser et al., 1999). dictate where the BBSome coat assembles. While no PIPs have If the BBSome assembles a canonical coat that sculpts been detected in cilia thus far, the 5-phosphatase INPP5E that membranes into buds and vesicles, where would this budding converts PI(3,4,5)P3 to PI(3,4)P2 is found in cilia, and loss of reaction take place within the cell? Since the steady-state local- INPP5E leads to several closely related to BBS ization of all known coat complexes is the organelle where they (Jacoby et al., 2009; Bielas et al., 2009). Furthermore, loss of sort cargo and deform membranes, one would predict that the the PI(3,4)P2-binding protein tubby leads to obesity and retinal BBSome buds endocytic vesicles off of the ciliary membrane. degeneration in mice (Santagata et al., 2001). Since tubby genet- As the passage of vesicles between the ciliary lumen and the ically interacts with bbs-1 in worms (Mak et al., 2006), it is is topologically unfeasible (Nachury et al., 2010), tempting to speculate that tubby and the BBSome may function it appears more plausible that the BBSome buds vesicles off of on the same membrane compartment. the base of the cilium to remove membrane proteins from cilia, The BBSome coat model suggests that the variety of symp- as suggested by Lechtreck et al. (2009). However, to account toms found in BBS patients is likely to result from the failure to for BBSome-mediated sorting of SSTR3i3 to cilia, one needs to transport signaling receptors to the cilium. While the relevance invoke the budding of vesicles by the BBSome from a donor of SSTR3 to the etiology of BBS is currently unknown (Einstein compartment that is distinct from cilia. Since we have never et al., 2010), the observation that the leptin receptor interacts observed any BBSome or Arl6 staining on endomembrane with BBS1 (Seo et al., 2009) provides a tantalizing hypothesis compartments or at the plasma membrane in a number of for the molecular basis of obesity in BBS, namely that leptin different cell lines using well-validated antibodies, we would signaling may take place within primary cilia in a BBSome-depen- need to postulate that, unlike all known coat complexes, the dent manner. The discovery of signaling receptors transported by steady-state localization of the BBSome does not correspond the BBSome promises to uncover the signaling defects that to the donor compartment for BBSome-mediated budding. underlie BBS and to provide mechanistic insights into the inter- Although unlikely, it is formally possible that the population of play between ciliary trafficking and signaling pathways. BBSome and Arl6 we observe within cilia corresponds to a slowly recycling pool that becomes injected into cilia after the fusion of EXPERIMENTAL PROCEDURES BBSome-coated vesicles with the base of the cilium. Antibodies and Reagents Restricting BBSome Coat Assembly within the Cell Arl6 antibodies were raised in rabbits and purified according to standard protocols. All other antibodies are described in the Extended Experimental Regardless of the model considered, only one protein, the small Procedures. All chemicals were purchased from Sigma unless otherwise indi- GTPase Arl6, is necessary and sufficient for recruiting the cated. 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC), 1,2-dioleoyl-sn- BBSome to membranes. No other soluble protein and no glycero-3-phosphoethanolamine (DOPE), 1,2-dioleoyl-sn-glycero-3-phos- membrane proteins are required for BBSome binding to lipo- pho-L-serine (DOPS), and 1,2-dioleoyl-sn-glycero-3-phosphate (DOPA) were

(G) Ciliary targeting of CD8a-SSTR3i3 is BBSome dependent. IMCD3-[CD8a-SSTR3i3] cells were treated with control siRNA or siRNA against Arl6 or BBS4 and serum starved to induce ciliation. Surface exposed CD8a-SSTR3i3 was visualized with the OKT8 antibody and cilia were stained with Glu-Tubulin as in (F). Insets show an enlargement of the CD8a channel in the region around the cilium. Images were acquired with a 1003/1.4 NA objective and planes containing cilia are shown. At least 110 cilia per treatment were counted and the percentages of CD8a-SSTR3i3-positive cilia were plotted. Error bars represent the SD between microscopic fields. Scale bars represent 5 mm. (H) A missorted BBSome cargo accumulates at the plasma membrane. Surface biotinylation was performed on IMCD3-[CD8a-SSTR3i3] cells treated as in (G). Equal portions of cell lysates were either immunoprecipitated with the OKT8 antibody (Total) or captured with Neutravidin (Surface). The strips were excised from the same membrane. Arl6 and BBS4 were efficiently depleted by siRNA treatment as shown in Figure S7C. Note that total or surface exposed CD8a-SSTR3i3 protein level is not changed in siArl6- or siBBS4-treated cells. See also Figure S7.

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1217 purchased from Avanti polar lipids. PI and PIPs were from Echelon or AG rino and Jon Mulholland for assistance with EM; members of Dick Tsien’s labo- Scientific. Texas Red 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanol- ratory for help with hippocampal neuron culture; Brian Kobilka and Juan Jose amine (DHPE) was from Invitrogen. Chemicals for electron microscopy were Fung for help with liposome preparation; Val Sheffield for the BBS1 antibody; from EM Sciences. Melanie Tallent for the SSTR3 knockout; Joachim Seemann for the CD8a cDNA; and Applied Precision for OMX imaging. This work was supported in Immunofluorescence and Microscopy part by grants to M.V.N. from the American Heart Association (AHA- Cells were fixed in PBS containing 4% paraformaldehyde for 5 min at 37C, fol- 0930365N), the March of Dimes (5-FY09-112), National Institutes of Health/ lowed by extraction with cold methanol at 20C for 5 min and processing for National Institute of General Medical Sciences (R01GM089933), a Sloan immunofluorescence as described (Nachury et al., 2007). Nuclei were stained Research Fellowship (BR-5014) and a Klingenstein Fellowship and by an with Hoechst 33258. Unless otherwise indicated in the figure legend, stacks of NIH/NIGMS fellowship to S.R.W (F32GM089218). J.F.B. is an employee of 24 Z sections were acquired at 0.25 mm interval with a 1003/1.45 NA objective Genentech, Inc. and deconvolved by constrained iterative, and the section containing the cilium was selected for each figure panel. Enhanced immunofluorescence Received: December 3, 2009 and CD8 staining are detailed in the Extended Experimental Procedures. Revised: March 19, 2010 Accepted: April 16, 2010 Affinity Chromatography and BBSome Purification Published: June 24, 2010 Affinity chromatography onto immobilized GST-Arl6DN16 was performed as in Christoforidis and Zerial (2000) with modifications. Bovine retinas (50 g) were REFERENCES thawed in 150 ml NS500 (25 mM Tris [pH 8.0], 500 mM KCl, 5 mM MgCl2, 1 mM DTT) supplemented with 250 mM sucrose and protease inhibitors Berbari, N.F., Lewis, J.S., Bishop, G.A., Askwith, C.C., and Mykytyn, K. (1 mM AEBSF, 10 mg/ml each of Leupeptin, Pepstatin A, Bestatin), homoge- (2008a). Bardet-Biedl syndrome proteins are required for the localization of nized by douncing and centrifuged for 2 hr at 184,000 3 gave in a Ti70 rotor. G protein-coupled receptors to primary cilia. Proc. Natl. Acad. Sci. USA 105, The retinal extract was loaded onto 1 ml GSTrap HP columns (GE) previously 4242–4246. saturated with GST fusion proteins, and columns were washed with 20 ml Berbari, N.F., Johnson, A.D., Lewis, J.S., Askwith, C.C., and Mykytyn, K. NS500 containing 50 mM nucleotide and eluted at 22C with four column (2008b). Identification of ciliary localization sequences within the third intracel- volumes of EB (25 mM Tris [pH 8.0], 2.5 M NaCl, 10 mM EDTA, 5 mM lular loop of G protein-coupled receptors. Mol. Biol. Cell 19, 1540–1547. MgCl2, 1 mM DTT, protease inhibitors). Eluates were run on SDS-PAGE or concentrated by methanol/chloroform precipitation for nanoscale microcapil- Bielas, S.L., Silhavy, J.L., Brancati, F., Kisseleva, M.V., Al-Gazali, L., Sztriha, lary reverse-phase liquid chromatography with electrospray ionization tandem L., Bayoumi, R.A., Zaki, M.S., Abdel-Aleem, A., Rosti, R.O., et al. (2009). Muta- mass spectrometry (LC-MS/MS) as previously described (Haas et al., 2006). tions in INPP5E, encoding inositol polyphosphate-5-phosphatase E, link phos- For BBSome purification, the eluate of the Arl6GTP column was then dialyzed phatidyl inositol signaling to the ciliopathies. Nat. Genet. 41, 1032–1036. against four successive buffers of decreasing ionic strength for 45 min each Blacque, O.E., Reardon, M.J., Li, C., McCarthy, J., Mahjoub, M.R., Ansley,

(final buffer: 25 mM HEPES [pH 7.0], 100 mM NaCl, 5 mM MgCl2,1mM S.J., Badano, J.L., Mah, A.K., Beales, P.L., Davidson, W.S., et al. (2004). DTT) and cleared by ultracentrifugation. The dialyzed Arl6 eluate was then frac- Loss of C. elegans BBS-7 and BBS-8 protein function results in cilia defects tionated on a Mono S PC 1.6/5 column (GE) equilibrated in buffer H5 (25 mM and compromised intraflagellar transport. Genes Dev. 18, 1630–1642. HEPES [pH 7.0], 50 mM NaCl) and developed with a gradient of 50 mM to Bremser, M., Nickel, W., Schweikert, M., Ravazzola, M., Amherdt, M., Hughes, 500 mM NaCl spanning nine column volumes. C.A., So¨ llner, T.H., Rothman, J.E., and Wieland, F.T. (1999). Coupling of coat GST pulldowns are described in the Extended Experimental Procedures. assembly and vesicle budding to packaging of putative cargo receptors. Cell 96, 495–506. Preparation of Liposomes, Binding Assays, and Ultrastructural Chaudhuri, I., So¨ ding, J., and Lupas, A.N. (2008). Evolution of the b-propeller Analysis fold. Proteins 71, 795–803. Liposomes were prepared according to Matsuoka et al. (1998) with minor modifications as detailed in the Extended Experimental Procedures. Lipo- Chiang, A.P., Nishimura, D., Searby, C., Elbedour, K., Carmi, R., Ferguson, somes were extruded through polycarbonate filters (100 nm pore size for brain A.L., Secrist, J., Braun, T., Casavant, T., Stone, E.M., and Sheffield, V.C. lipids and 400 nm pore size for synthetic lipids) and kept at 4C. All liposomes (2004). Comparative genomic analysis identifies an ADP-ribosylation factor- made from synthetic lipids contained 23 mol% of cholesterol and 77 mol% like gene as the cause of Bardet-Biedl syndrome (BBS3). Am. J. Hum. Genet. phospholipids, and Texas Red DHPE was included to normalize lipid concen- 75, 475–484. trations across all liposome stocks and to normalize the rate of liposome Christoforidis, S., and Zerial, M. (2000). Purification and identification of novel recovery in flotation assays. Liposome pelleting assays were conducted in Rab effectors using affinity chromatography. Methods 20, 403–410. HKSM buffer (20 mM HEPES [pH 7.0], 150 mM KOAc, 250 mM Sorbitol, Einstein, E.B., Patterson, C.A., Hon, B.J., Regan, K.A., Reddi, J., Melnikoff, 3.5 mM MgCl2) supplemented with nucleotides. Liposome flotation assays D.E., Mateer, M.J., Schulz, S., Johnson, B.N., and Tallent, M.K. (2010). were conducted as detailed in the Extended Experimental Procedures. Ultra- Somatostatin signaling in neuronal cilia is critical for object recognition structural analysis of protein/liposome mixtures by thin-section electron memory. J. Neurosci. 30, 4306–4314. microscopy was preformed as described Matsuoka et al. (1998). Fan, Y., Esmail, M.A., Ansley, S.J., Blacque, O.E., Boroevich, K., Ross, A.J., Moore, S.J., Badano, J.L., May-Simera, H., Compton, D.S., et al. (2004). Muta- SUPPLEMENTAL INFORMATION tions in a member of the Ras superfamily of small GTP-binding proteins causes Bardet-Biedl syndrome. Nat. Genet. 36, 989–993. Supplemental Information includes Extended Discussion, Extended Experi- Fliegauf, M., Benzing, T., and Omran, H. (2007). When cilia go bad: cilia defects mental Procedures, seven figures, and one movie and can be found with and ciliopathies. Nat. Rev. Mol. Cell Biol. 8, 880–893. this article online at doi:10.1016/j.cell.2010.05.015. Follit, J.A., Li, L., Vucica, Y., and Pazour, G.J. (2010). The cytoplasmic tail of ACKNOWLEDGMENTS fibrocystin contains a ciliary targeting sequence. J. Cell Biol. 188, 21–28. Gerdes, J.M., Liu, Y., Zaghloul, N.A., Leitch, C.C., Lawson, S.S., Kato, M., We thank Suzanne Pfeffer and Scott Seeley for comments on the manuscript; Beachy, P.A., Beales, P.L., DeMartino, G.N., Fisher, S., et al. (2007). Disruption Randy Schekman, Sean Studer, Susan Hamamoto, Ken Matsuoka and Chris of the compromises proteasomal function and perturbs intracel- Fromme for advice on liposome preparation and processing for EM; John Per- lular Wnt response. Nat. Genet. 39, 1350–1360.

1218 Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. Gillingham, A.K., and Munro, S. (2007). The small G proteins of the Arf family Moritz, O.L., Tam, B.M., Hurd, L.L., Pera¨ nen, J., Deretic, D., and Papermaster, and their regulators. Annu. Rev. Cell Dev. Biol. 23, 579–611. D.S. (2001). Mutant rab8 Impairs docking and fusion of rhodopsin-bearing Haas, W., Faherty, B.K., Gerber, S.A., Elias, J.E., Beausoleil, S.A., Bakalarski, post-Golgi membranes and causes cell death of transgenic Xenopus rods. C.E., Li, X., Ville´ n, J., and Gygi, S.P. (2006). Optimization and use of peptide Mol. Biol. Cell 12, 2341–2351. mass measurement accuracy in shotgun proteomics. Mol. Cell. Proteomics Nachury, M.V., Loktev, A.V., Zhang, Q., Westlake, C.J., Pera¨ nen, J., Merdes, 5, 1326–1337. A., Slusarski, D.C., Scheller, R.H., Bazan, J.F., Sheffield, V.C., and Jackson, Jacoby, M., Cox, J.J., Gayral, S., Hampshire, D.J., Ayub, M., Blockmans, M., P.K. (2007). A core complex of BBS proteins cooperates with the GTPase Pernot, E., Kisseleva, M.V., Compe` re, P., Schiffmann, S.N., et al. (2009). Rab8 to promote ciliary membrane biogenesis. Cell 129, 1201–1213. INPP5E mutations cause primary cilium signaling defects, ciliary instability Nachury, M.V., Seeley, E.S., and Jin, H. (2010). Trafficking to the ciliary and ciliopathies in human and mouse. Nat. Genet. 41, 1027–1031. membrane: How to get across the periciliary diffusion barrier? Annu. Rev. Jı´nek, M., Rehwinkel, J., Lazarus, B.D., Izaurralde, E., Hanover, J.A., and Cell Dev. Biol. 26, in press. 10.1146/annurev.cellbio.042308.113337. Conti, E. (2004). The superhelical TPR-repeat domain of O-linked GlcNAc transferase exhibits structural similarities to importin alpha. Nat. Struct. Mol. Ou, G., Blacque, O.E., Snow, J.J., Leroux, M.R., and Scholey, J.M. (2005). Biol. 11, 1001–1007. Functional coordination of intraflagellar transport motors. Nature 436, Kim, J.C., Badano, J.L., Sibold, S., Esmail, M.A., Hill, J., Hoskins, B.E., Leitch, 583–587. C.C., Venner, K., Ansley, S.J., Ross, A.J., et al. (2004). The Bardet-Biedl Rosenbaum, J.L., and Witman, G.B. (2002). Intraflagellar transport. Nat. Rev. protein BBS4 targets cargo to the pericentriolar region and is required for Mol. Cell Biol. 3, 813–825. microtubule anchoring and cell cycle progression. Nat. Genet. 36, 462–470. Saffarian, S., Cocucci, E., and Kirchhausen, T. (2009). Distinct dynamics of Kobayashi, T., Hori, Y., Ueda, N., Kajiho, H., Muraoka, S., Shima, F., Kataoka, endocytic clathrin-coated pits and coated plaques. PLoS Biol. 7, e1000191. T., Kontani, K., and Katada, T. (2009). Biochemical characterization of missense mutations in the Arf/Arl-family small GTPase Arl6 causing Bardet- Santagata, S., Boggon, T.J., Baird, C.L., Gomez, C.A., Zhao, J., Shan, W.S., Biedl syndrome. Biochem. Biophys. Res. Commun. 381, 439–442. Myszka, D.G., and Shapiro, L. (2001). G-protein signaling through tubby Koenig, J.H., and Ikeda, K. (1996). Synaptic vesicles have two distinct recy- proteins. Science 292, 2041–2050. cling pathways. J. Cell Biol. 135, 797–808. Schermelleh, L., Carlton, P.M., Haase, S., Shao, L., Winoto, L., Kner, P., Burke, Lechtreck, K.F., Johnson, E.C., Sakai, T., Cochran, D., Ballif, B.A., Rush, J., B., Cardoso, M.C., Agard, D.A., Gustafsson, M.G., et al. (2008). Subdiffraction Pazour, G.J., Ikebe, M., and Witman, G.B. (2009). The Chlamydomonas rein- multicolor imaging of the nuclear periphery with 3D structured illumination hardtii BBSome is an IFT cargo required for export of specific signaling microscopy. Science 320, 1332–1336. proteins from flagella. J. Cell Biol. 187, 1117–1132. Seo, S., Guo, D.-F., Bugge, K., Morgan, D.A., Rahmouni, K., and Sheffield, Loktev, A.V., Zhang, Q., Beck, J.S., Searby, C.C., Scheetz, T.E., Bazan, J.F., V.C. (2009). Requirement of Bardet-Biedl syndrome proteins for leptin Slusarski, D.C., Sheffield, V.C., Jackson, P.K., and Nachury, M.V. (2008). receptor signaling. Hum. Mol. Genet. 18, 1323–1331. A BBSome subunit links ciliogenesis, microtubule stability, and acetylation. Dev. Cell 15, 854–865. Seo, S., Baye, L.M., Schulz, N.P., Beck, J.S., Zhang, Q., Slusarski, D.C., and Mak, H.Y., Nelson, L.S., Basson, M., Johnson, C.D., and Ruvkun, G. (2006). Sheffield, V.C. (2010). BBS6, BBS10, and BBS12 form a complex with CCT/ Polygenic control of Caenorhabditis elegans fat storage. Nat. Genet. 38, TRiC family chaperonins and mediate BBSome assembly. Proc. Natl. Acad. 363–368. Sci. USA 107, 1488–1493. Matsuoka, K., Orci, L., Amherdt, M., Bednarek, S.Y., Hamamoto, S., Schek- Spang, A., Matsuoka, K., Hamamoto, S., Schekman, R., and Orci, L. (1998). man, R., and Yeung, T. (1998). COPII-coated vesicle formation reconstituted Coatomer, Arf1p, and nucleotide are required to bud coat protein complex with purified coat proteins and chemically defined liposomes. Cell 93, I-coated vesicles from large synthetic liposomes. Proc. Natl. Acad. Sci. USA 263–275. 95, 11199–11204. Mazelova, J., Astuto-Gribble, L., Inoue, H., Tam, B.M., Schonteich, E., Preke- Stagg, S.M., LaPointe, P., and Balch, W.E. (2007). Structural design of cage ris, R., Moritz, O.L., Randazzo, P.A., and Deretic, D. (2009). Ciliary targeting and coat scaffolds that direct membrane traffic. Curr. Opin. Struct. Biol. 17, motif VxPx directs assembly of a trafficking module through Arf4. EMBO J. 221–228. 28, 183–192. McMahon, H.T., and Mills, I.G. (2004). COP and clathrin-coated vesicle Tam, B.M., Moritz, O.L., Hurd, L.B., and Papermaster, D.S. (2000). Identifica- budding: different pathways, common approaches. Curr. Opin. Cell Biol. 16, tion of an outer segment targeting signal in the COOH terminus of rhodopsin 379–391. using transgenic Xenopus laevis. J. Cell Biol. 151, 1369–1380. Milenkovic, L., Scott, M.P., and Rohatgi, R. (2009). Lateral transport of Wang, C.-W., Hamamoto, S., Orci, L., and Schekman, R. (2006). Exomer: Smoothened from the plasma membrane to the membrane of the cilium. A coat complex for transport of select membrane proteins from the trans-Golgi J. Cell Biol. 187, 365–374. network to the plasma membrane in yeast. J. Cell Biol. 174, 973–983.

Cell 141, 1208–1219, June 25, 2010 ª2010 Elsevier Inc. 1219