BBS1 is involved in retrograde trafficking of ciliary GPCRs in the context of the BBSome complex

Protein trafficking within cilia is mediated by the intraflagellar transport (IFT) machinery composed of large protein complexes. The BBSome consists of eight BBS proteins encoded by causative genes of Bardet-Biedl syndrome (BBS), and has been implicated in the trafficking of ciliary membrane proteins, including G protein-coupled receptors (GPCRs), by connecting the IFT machinery to cargo GPCRs. Membrane recruitment of the BBSome to promote cargo trafficking has been proposed to be regulated by the Arf-like small GTPase ARL6/BBS3, through its interaction with the BBS1 subunit of the BBSome. We here investigated how the BBSome core subcomplex composed of BBS1, BBS2, BBS7, and BBS9 assembles and interacts with ARL6, and found that the ARL6–BBS1 interaction is reinforced by BBS9. BBS1-knockout (KO) cells showed defects in the ciliary entry of other BBSome subunits and ARL6, and in ciliary retrograde trafficking and the export of the GPCRs, Smoothened and GPR161. The trafficking defect of these GPCRs was rescued by the exogenous expression of wild-type BBS1, but not by its mutant lacking BBS9-binding ability. Our data thus indicate that the intact BBSome is required for retrograde trafficking of GPCRs out of cilia.


Introduction
Cilia are specialized cell surface projections that function as cellular antennae by perceiving extracellular stimuli, such as fluid flow, and by receiving and transducing developmental signals, such as the Hedgehog (Hh) signal [1,2]. A cilium is composed of a microtubule-based scaffold called the axoneme, which is surrounded by a ciliary membrane that is continuous with the plasma membrane. Defects in ciliary assembly and functions therefore cause a variety of congenital disorders, such as Bardet-Biedl syndrome (BBS), Joubert syndrome, nephronophthisis, and Meckel syndrome, which are collectively referred to as ciliopathies [3,4]. These are pleiotropic disorders characterized by a broad spectrum of symptoms, including polycystic kidney, retinal degeneration, polydactyly, morbid obesity, and mental retardation.
The composition of proteins and lipids in cilia are greatly different from those of the cell body, because the transition zone at the base of cilia serve as a permeability/diffusion barrier PLOS

Establishment of KO cell lines using the CRISPR/Cas9 system
The strategy for disruption of genes in hTERT-RPE1 cells (ATCC, CRL-4000) by the CRISPR/ Cas9 system using homology-independent DNA repair (a version 2 method) was previously described in detail [25]; also see [16,24,26,27]. Briefly, single-guide RNA (sgRNA) sequences targeting the human BBS1 gene (see S3 Table) were designed using CRISPR design [28]. Double-stranded oligonucleotides for these sequences were inserted separately into the all-in-one sgRNA expression vector peSpCAS9(1.1)-2×sgRNA (Addgene ID 80768). hTERT-RPE1 cells were grown on a 12-well plate to approximately 3.0 × 10 5 cells, transfected with 1 μg of the sgRNA vector and 0.25 μg of the donor knock-in vector pDonor-tBFP-NLS-Neo(universal) (Addgene ID 80767) using X-tremeGENE9 DNA transfection reagent (Roche Applied Science), and cultured in the presence of G418 (600 μg/mL). Colonies of the cells carrying nuclear tBFP signals were isolated. Genomic DNA from these isolated cells was subjected to PCR using KOD FX Neo DNA polymerase (TOYOBO). Three sets of primers (S3 Table) were used to distinguish the three integration modes of the donor knock-in vector: forward integration (S2A and S2C Fig, b and b'), reverse integration (c and c'), and no integration with a small indel (a and a') (see [25]

Preparation of cells stably expressing mChe-tagged BBS1 constructs
Lentiviral vectors were prepared as described previously [29]. Briefly, pRRLsinPPT-mChe-BBS1(WT) or its mutant was transfected into HEK293T cells using Polyethylenimine Max along with the packaging plasmids (pRSV-REV, pMD2.g, and pMDL/pRRE; kind gifts from Peter McPherson, McGill University [30]). Culture medium was replaced 8 h after transfection, and collected at 24, 36, and 48 h after transfection. The culture medium containing viral particles was passed through a 0.45-μm filter and centrifuged at 32,000 × g at 4˚C for 4 h. Precipitated lentiviral particles were resuspended in Opti-MEM (Invitrogen) and stored at −80˚C until use. BBS1-KO cells that express mChe-BBS1(WT) or its mutant were prepared by adding a lentiviral suspension to the culture medium.

Immunofluorescence analysis
Induction of ciliogenesis and subsequent immunofluorescence analysis of hTERT-RPE1 cells were carried out as described previously [16,27]. The stained cells were observed using an Axiovert 200M microscope (Carl Zeiss). Statistical analyses were performed using JMP Pro 12 software (SAS Institute).

Modes of interactions involving BBSome core subunits and ARL6
In our previous study [17], we analyzed the 64 possible combinations of eight BBSome subunits tagged with EGFP and tRFP/mChe, by taking advantage of the VIP assay, and found that BBS1, BBS2, BBS7, and BBS9 constitute the core subcomplex (S1 Fig). These four core subunits share common domain organizations reminiscent of those of coat protein complexes [19]. All four subunits are predicted to have a BP fold in their N-terminal half ( Fig 1A); for BBS1 and BBS9, their seven-bladed BP structures were confirmed by X-ray crystallography [31,32]. In their C-terminal half, these four subunits are predicted to have a GAE domain, which is followed by an α/β-platform (PF) domain (except for BBS1). Because BP folds are found in subunits of the COPI coat protein complex, which are subunits which participate in cargo recognition [33,34], and because GAE domains are found in clathrin adaptor proteins, the AP-1 γ-subunit, and GGA proteins [35], Nachury and colleagues proposed that the BBSome functions like a coat protein complex [19].
To delineate the interaction modes among the core subunits, we divided the WT BBS1, BBS2, BBS7, and BBS9 proteins into their N-terminal BP domain and their C-terminal (CT) region containing the GAE-like and α/β-PF domains, as schematically shown in Fig 1A, coexpressed them as EGFP-and mChe/tRFP-fusions in HEK293T cells, and subjected them to the VIP assay. In the VIP assay, protein-protein interactions can be visually detected by analyzing whether the mChe/tRFP-fused protein is coimmunoprecipitated with the EGFP-fused protein under a microscope [17]. As shown in Fig 1B, the BBS9 CT region fused to EGFP interacted with the tRFP-fused CT regions of BBS1 and BBS2 (row 3 and 6, respectively). On the other hand, the BP domain of BBS9 interacted with BBS5 and BBS8 (Fig 1B, row 7 and 8, respectively). The BBS2-BBS7 interaction was mediated by their CT regions (Fig 1C, bottom row).
The VIP assay suggested that the BP domain of BBS1 interacted mainly with the BBS7 CT region (Fig 1D, bottom row), although the BBS1 BP domain also exhibited an interaction with the BBS7 BP region (row 3). We also performed conventional immunoblotting analysis to confirm the VIP data, because a very recent study using the yeast two-hybrid system did not detect the BBS1-BBS7 interaction [36]. As shown in Fig 1E, the CT region of BBS7 made a major contribution to its interaction with the BBS1 BP domain (lane 11), although the BBS7 BP region also exhibited an interaction with the BBS1 BP domain (lane 8).
In agreement with a previous crystallographic study [32], ARL6ΔN15(Q73L) was coprecipitated with the BP domain, but not the CT region, of BBS1 ( Fig 1F); we here used an ARL6ΔN15 (residues 1-15 deleted) construct, like in previous studies [19,32], as the presence of an N-terminal amphipathic helix generally hampers interactions of the ARF/ARL family GTPases with their effectors, at least in vitro.
Based on the data shown in Fig 1B-1F, we predicted the interaction model shown in Fig  1G. The BBS7-BBS2, BBS2-BBS9, and BBS9-BBS1 interactions are mediated by their CT regions, whereas the interaction between BBS1 and BBS7 is mediated mainly by their BP domain and CT region, respectively. The BBS1 BP domain also participates in its interaction with ARL6.
The CT region of BBS1 contains an α-helix region followed by a GAE-like domain. A BBS1 construct lacking the α-helix region (Δα; Fig 2A) retained the ability to interact with BBS9 ( Fig  2B, column 2). By contrast, another BBS1 construct lacking the C-terminal 18-amino acids of the GAE-like domain (BBS1(1-575); Fig 2A) did not interact with BBS9 ( Fig 2B, column 3). Thus, at least a part of the GAE-like domain of BBS1 participates in its interaction with BBS9.
Lorentzen and colleagues previously reported the crystal structure of Chlamydomonas ARL6ΔN15 in complex with Chlamydomonas BBS1(1-425) [32]. They constructed some Lysates prepared from the transfected cells were subjected to the VIP assay, as described in Materials and Methods. (C) Lysates were prepared from HEK293T cells coexpressing an EGFP-fused BBS2 construct and mChe-fused BBS7 construct, as indicated, and subjected to the VIP assay. (D, E) HEK293T cells were cotransfected with expression vectors for an EGFP-fused BBS1 construct and mChe or an mChe-fused BBS7 construct, as indicated. Lysates were prepared from the transfected cells and processed for the VIP assay (D) or SDS-PAGE followed by immunoblotting analysis with an anti-RFP antibody, which reacts with mChe, or an anti-GFP antibody (E). (F) Lysates prepared from HEK293T cells coexpressing an EGFP- mutants of the human BBS1 BP domain on the basis of the crystal structure and found that an I399E or R404A mutation in human BBS1 abolished its interaction with ARL6; Ile 399 and Arg 404 are located within the bipartite blade 1 of the BBS1 BP domain [32]. By the VIP assay and conventional immunoblotting analysis, we confirmed their data; the I399E or R404A mutant of BBS1 could not interact with ARL6ΔN15(Q73L) (Fig 2C and 2D, lanes 3 and 4). Somewhat unexpectedly, BBS1(1-575) demonstrated an attenuated interaction with ARL6ΔN15(Q73L) compared with BBS1(WT) (Fig 2C and 2D, compare lane 5 with lane 2); this will be discussed later (see below). By contrast, both BBS1 point mutants retained their ability to interact with BBS7 ( Fig 2E and 2F, lanes 2 and 3). Thus, it is likely that the BBS1 BP domain interacts with ARL6 and BBS7, at least in part via distinct interfaces. fused BBS1 construct, as indicated, and tRFP-fused ARL6ΔN15(Q73L), were processed for the VIP assay. (G) Schematic representation of the predicted model of interactions involving BBSome subunits and ARL6.
https://doi.org/10.1371/journal.pone.0195005.g001 Lysates prepared from HEK293T cells coexpressing the EGFP-BBS9 construct and mChe-fused BBS1 construct, as indicated, were subjected to the VIP assay. (C, D) HEK293T cells were cotransfected with expression vectors for EGFP-fused ARL6ΔN15(Q73L) and a tRFP-fused BBS1 construct,as indicated. Lysates were prepared from the transfected cells and processed for the VIP assay (C) or SDS-PAGE followed by immunoblotting analysis with an anti-tRFP or anti-GFP antibody (D). (E, F) Lysates prepared from HEK293T cells cotransfected with expression vectors for EGFP-BBS7 and a tRFP-fused BBS1 construct as indicated were processed for the VIP assay (E) or SDS-PAGE followed by immunoblotting analysis with an anti-tRFP or anti-GFP antibody (F).

Impaired retrograde trafficking of ciliary GPCRs in BBS1-KO cells
Smoothened (SMO) and GPR161 are seven-pass transmembrane GPCRs involved in Hh signaling [1,2]. Under basal conditions, SMO is absent from cilia, whereas GPR161 on the ciliary membrane negatively regulates Hh signaling. When the Hh pathway is stimulated, for example, by treating ciliated cells with a small molecule activator, Smoothened Agonist (SAG), SMO enters cilia and GPR161 exit cilia; consequently, the negative regulation of the Hh signaling is canceled.
Sheffield and colleagues previously reported that, in cells derived from Arl6-KO mice, SMO is significantly accumulated within cilia even under basal conditions [20]. On the other hand, Nachury and colleagues reported that GPR161 was retained in the cilia of Arl6-KO IMCD3 cells even when the cells were treated with SAG [21]. These observations taken together suggested that retrograde trafficking and/or export of these ciliary GPCRs are impaired in the absence of ARL6.
In this study, we established BBS1-KO hTERT-RPE1 cell lines using a CRISPR/Cas9 system with our original modifications (the version 2 method; see [25]) and compared their phenotypes with those of control RPE1 cells, as ARL6 directly interacts with BBS1 and regulates BBSome function. Two ARL6 (Fig 3D), BBS5 (Fig 3G), and BBS9 ( Fig 3J) were uniformly distributed within cilia in 10%-20% of control RPE1 cells (also see Fig 3S-3U). In marked contrast, localization of these BBS proteins was substantially altered in the BBS1-KO cell lines: ARL6 was no longer found inside cilia, but its localization at the ciliary base was, at least partially, retained (Fig 3E and 3F; also see Fig 3S); BBS5 localization within cilia and at the ciliary base was abolished (Fig 3H and  3I, also see Fig 3T); ciliary localization of BBS9 was also abolished, and BBS9-positive aggregates were often observed around the base (Fig 3K and 3L, also see Fig 3U). Although the identity of these aggregates is unclear, Sheffield and colleagues also reported the presence of BBS9-positive and BBS8-positive aggregates around the ciliary base in RPE1 cells treated with BBS1 siRNA [22]. Overall, it is thus likely that the BBSome and ARL6 cannot enter cilia in the absence of BBS1.
As the BBSome is believed to move within cilia in association with IFT particles containing the IFT-A and IFT-B complexes [12], we also analyzed the localization of IFT-A and IFT-B subunits in the BBS1-KO cell lines. However, the localization of IFT88 (an IFT-B subunit) or IFT140 (an IFT-A subunit) was not apparently altered in the absence of BBS1 compared with control RPE1 cells; IFT88 was mainly localized at the ciliary base with a minor proportion found along cilia (Fig 3M-3O), whereas the majority of IFT140 is found at the base (Fig 3P-3R). It is likely that intraciliary movement of the IFT-A or IFT-B complex is not dependent on the BBSome, although the BBSome moves in association with IFT particles.
We then compared the localization of SMO and GPR161 under basal (-SAG) and SAGtreated (+SAG) conditions. Under basal conditions, the ciliary localization of SMO was not detected in control RPE1 cells (Fig 4A), but was substantially increased in the BBS1-KO cell lines (Fig 4B and 4C, also see Fig 4M). Upon stimulation with SAG, SMO entered cilia in control RPE1 cells (Fig 4D, also see Fig 4M), and its ciliary localization was further enhanced in the BBS1-KO cells (Fig 4E and 4F, also see Fig 4M). These observations suggest two possibilities: one is that the BBSome suppresses ciliary entry of SMO; and the other is that, even under basal conditions, SMO undergoes constitutive cycling between the ciliary and plasma membranes to keep its ciliary level low, and that a block in its retrograde trafficking and/or exit from cilia in the absence of BBS1 might result in significant retention of SMO on the ciliary membrane. In view of the GPR161 data (see below), we favor the latter possibility.
On the other hand, in control cells, GPR161 was found within most cilia under basal conditions (Fig 4G), and its localization to cilia was significantly decreased under SAG-stimulated conditions (Fig 4J). Under basal conditions, localization of GPR161 within cilia in the BBS1-KO cell lines (Fig 4H and 4I; also see Fig 4N) was significantly higher than that in control cells (Fig 4G). Furthermore, unlike in control cells, GPR161 was retained within cilia even when the BBS1-KO cells were treated with SAG (Fig 4K and 4L; also see Fig 4N). The altered localization of SMO and GPR161 in the BBS1-KO cells under basal and SAG-stimulated conditions is reminiscent of our previous study on cells lacking IFT139, which is a subunit of the IFT-A complex [16]; in IFT139-KO cells, retrograde trafficking and/or export of ciliary GPCRs, including SMO and GPR161, was severely impaired. These observations together indicate that, in the absence of BBS1, retrograde trafficking and/or ciliary exit of the GPCRs involved in Hh signaling is impaired.

BBS1 mediates retrograde trafficking of GPCRs in the context of the BBSome via its interaction with BBS9
To exclude the potential off-target effects of the CRISPR/Cas9 system, we then performed rescue experiments. BBS1-KO cells were infected with a lentiviral vector for the stable expression of mChe-tagged BBS1(WT) or its mutant. In contrast to the BBS1-KO (#B1-1-23) cell line without exogenous BBS1 expression (Fig 4B, 4E, 4H and 4K), SMO was excluded from and GPR161 was localized within cilia (Fig 5A and 5I; also see Fig 5Q and 5R) in the #B1-1-23 cell line expressing mChe-BBS1(WT) under basal conditions, as observed in control cells (Fig 4A  and 4G). Upon stimulation with SAG, SMO entered and GPR161 exited cilia (Fig 5E and 5M; also see Fig 5Q and 5R) in the mChe-BBS1(WT)-expressing BBS1-KO cell line, similarly to in control cells (Fig 4D and 4J). Thus, the impaired localization of SMO and GPR161 in the BBS1-KO cell line under both basal and SAG-stimulated conditions was rescued by the exogenous expression of BBS1(WT).
In striking contrast, the exogenous expression of mChe-tagged BBS1(1-575), which cannot interact with BBS9, did not restore the normal localization of SMO or GPR161 in the BBS1-KO cell line; a substantial, although low, level of SMO was found within cilia under basal conditions (Fig 5D; also see Fig 5Q), and an increased level of GPR161 within cilia was found under basal conditions, and the level was maintained even when the KO cells were stimulated with SAG (Fig 5P; also see Fig 5R). These data indicate that the interaction of BBS1 with BBS9, in other words, the integrity of the BBSome core subcomplex (see S1 Fig), is essential for BBSome function to mediate retrograde trafficking of ciliary GPCRs and/or their exit from cilia.

ARL6 interacts with the BBSome via BBS1 with the aid of BBS9
In contrast to the failed recovery of GPCR retrograde trafficking by the expression of BBS1(1-575), in BBS1-KO cells exogenously expressing a BBS1 mutant defective in ARL6 binding substantially rescued the abnormal localization of SMO and GPR161. Namely, in the BBS1-KO cell line expressing mChe-BBS1(I399E) or mChe-BBS1(R404A), SMO was excluded from cilia under basal conditions (Fig 5B and 5C; also see Fig 5Q). Ciliary exit of GPR161 upon SAG treatment was promoted in BBS1-KO cells expressing mChe-BBS1(I399E) or mChe-BBS1 (R404A) (Fig 5N and 5O), compared with those expressing mChe-BBS1(1-575) (Fig 5P; also see Fig 5R). Thus, these BBS1 point mutants appeared to be functional, at least partly, with regard to GPCR trafficking, even though they are defective in ARL6 binding (Fig 2C and 2D). Given that ARL6 demonstrates a binary interaction only with BBS1 among the BBSome subunits (S4 Fig), the results of these rescue experiments were intriguing.
In an attempt to address the apparent contradiction of the data obtained by the rescue experiments of the BBS1-KO cells (Fig 5) with the biochemical interaction data (Fig 2C and  2D), we took advantage of the VIP assay to analyze whether the ARL6-BBS1 interaction can be enhanced in the presence of other BBSome subunit(s). For this purpose, tRFP-fused ARL6ΔN15(Q73L) and either EGFP-fused BBS1(WT), BBS1(I399E), or BBS1(R404A) were co-expressed with tBFP fusions of all the other BBSome subunits, other core subunits (BBS2, BBS7, and BBS9), or linker subunits (BBS4, BBS8, and BBS18) in HEK293T cells. As shown in Fig 6A, the interaction of EGFP-BBS1(WT) with tRFP-ARL6 appeared to be promoted in the presence of tBFP-fused all the other BBSome subunits (BBS2/BBS4/BBS5/BBS7/BBS8/BBS9/ BBS18; Fig 6A, column 1) or other core subunits (BBS2/BBS7/BBS9; column 4) compared with that in the presence of tBFP-fused linker subunits (BBS4/BBS8/BBS18) (column 7), the latter which does not demonstrate a direct interaction with BBS1. Importantly, substantial interaction of BBS1(I399E) or BBS1(R404A) with ARL6 was detectable in the presence of all other BBSome subunits or all other core subunits, but not in the presence of linker subunits (compare columns 1-6 with 7-9).
The data shown in Fig 1B indicate that the BBS9 CT region interacts with the BBS1 CT region. We then examined whether the BBS9 CT region is sufficient for promoting the ARL6 interaction with BBS1. As shown in Fig 6D and 6E, not only the WT construct (lane 2) but also the CT region (lane 4) of BBS9 considerably enhanced the interaction of BBS1(WT) with ARL6. The enhancement of the ARL6-BBS1 interaction by BBS9 was confirmed to be mediated by the interaction of BBS9 with BBS1, since the BBS9 CT region did not promote the interaction of ARL6 with BBS1(1-575) (compare lane 6 with lane 4), which lacks the BBS9-binding ability (Fig 2B).
These VIP and immunoblotting data suggest that BBS9 can reinforce the interaction of ARL6 with BBS1, although it does not directly interact with ARL6 (see Discussion). Furthermore, these construct, as indicated, and tRFP-ARL6ΔN15(Q73L) together with tBFP-fusion vectors for all BBSome subunits (excepting BBS1), all core subunits (excepting BBS1), or all linker subunits. Lysates prepared from the transfected cells were processed for the VIP assay. Beads bearing fluorescent fusion proteins were observed using an A1R-MP confocal laser-scanning microscope (Nikon). (B and C) HEK293T cells were transfected with expression vectors for an EGFP-fused BBS1 construct as indicated and tRFP-ARL6ΔN15(Q73L) together with tBFP-fused BBS2, BBS7, or BBS9. Lysates prepared from the transfected cells were processed for the VIP assay (B) or immunoblotting analysis (C) with anti-tRFP antibody, which reacts with tRFP and tBFP, and very weakly cross-reacts with EGFP (indicated by an asterisk) and with an anti-GFP antibody. (D and E) Lysates prepared from HEK293T cells cotransfected with expression vectors for an EGFP-fused BBS1 construct as indicated, and tRFP-ARL6ΔN15(Q73L) together with tRFP-fused BBS2 or a BBS9 construct as indicated were processed for the VIP assay (D) or immunoblotting analysis (E) with anti-tRFP antibody and anti-GFP antibody. The anti-tRFP antibody very weakly cross-reacts with EGFP (indicated by an asterisk).

Discussion
Protein trafficking within cilia is mediated by the IFT machinery composed of large protein complexes. The BBSome consists of eight BBS proteins encoded by causative genes of BBS, and has been implicated in the trafficking of ciliary membrane proteins, including GPCRs, by connecting the IFT machinery and cargo GPCRs. The membrane recruitment and coat-like assembly of the BBSome to promote cargo trafficking has been proposed to be regulated by the Arflike small GTPase ARL6/BBS3, through its interaction with the BBS1 subunit. Using the VIPbased method, we here systematically investigated how the BBSome core subcomplex composed of BBS1, BBS2, BBS7, and BBS9 assembles and interacts with ARL6 (Fig 1). The data presented here showed that the CT regions containing the GAE and PF domains, but not the BP domains (except for that of BBS1) of these core subunits mainly participate in the assembly of the core subcomplex ( Fig 1G). In other words, the BP domains of the core subunits are free from core subcomplex assembly. Taking into account the fact that the BP domains of the α-COP and δ-COP subunits of the COPI complex are responsible for recognition of cargo molecules [33,34], it is tempting to speculate that the BP domains of the BBSome subunits participate in cargo recognition, although our attempts to find interactions between the BBSome and candidate cargo molecules have so far been unsuccessful. While this manuscript was in preparation, Klink et al. reported that a recombinant BBSome semi-complex, which contained BBS1 and BBS9 but lacked BBS2 and BBS7, bound in vitro to synthetic peptides derived from ciliary GPCRs, SMO and SSTR3 [37]. Given that, in the genomes of Drosophila species, the BBS2 and BBS7 genes are absent [38], BBS1 and/or BBS9 might play pivotal roles in cargo recognition, although it remains possible that other subunits also play some role.
Unexpectedly, our VIP-based analysis, supported by conventional immunoblotting analysis, also demonstrated that although BBS1 directly interacts with ARL6 via its BP domain as shown by a previous crystallographic study [32], the ARL6-BBS1 interaction can be indirectly strengthened by BBS9 (Fig 6B and 6C). In view of the facts that BBS1 interacts with BBS9 via its CT region ( Fig 1B) and that BBS9 did not show a direct interaction with ARL6 (S4 Fig), how BBS9 supports the ARL6-BBS1 interaction is an interesting issue to address. One possible explanation is that the BBS1 protein on its own adopts a closed conformation, but upon binding of BBS9 to its CT region, the BBS1 protein undergoes a change in conformation so that ARL6 is now accessible to its BP domain. In support of this speculation, the interaction of ARL6 with the BP domain construct of BBS1 appears to be stronger than that with the BBS1 (WT) construct (Fig 1F). Another possibility is that BBS9 can somehow stabilize the ARL6-BBS1 dimer. If so, formation of the ARL6-BBS1 dimer, and subsequent BBS9 binding, can trigger the assembly of the whole BBSome complex. In any case, our data suggest that the ARL6-BBS1 interaction is maximally functional in the context of the BBSome complex.
In this study, we also established BBS1-KO RPE1 cells and showed that the absence of BBS1 impairs retrograde trafficking and/or export of GPR161, and possibly SMO (Fig 4). This phenotype is in line with that reported for ARL6-KO cells [20,21]. The impaired trafficking of these ciliary GPCRs was rescued by the exogenous expression of BBS1(WT), but not by its mutant, BBS1(1-575), defective in BBS9 binding due to the lack of only 18-amino acids from the C-terminus ( Fig 5). As BBS1(1-575) retains the ability to interact with BBS7 and ARL6, the data of rescue experiments indicate that the integrity of the whole BBSome complex is crucial for its role in ciliary protein trafficking.
On the other hand, two BBS1 mutants, BBS1(I399E) and BBS1(R404A), which are defective in the binary interaction with ARL6 (Fig 2B and 2C) [32], were unexpectedly found to rescue the impaired GPCR trafficking in BBS1-KO cells (Fig 5). Given that ARL6-KO cells [20,21] show apparently the same phenotype as that of BBS1-KO cells (Fig 5), it was intriguing that the BBS1 mutants defective in ARL6 binding were able to restore the impaired GPCR trafficking in BBS1-KO cells. However, we finally found that the BBS1 mutants demonstrate a substantial, although limited, interaction with ARL6 in the presence of BBS9, as described above. Therefore, how ARL6 is implicated in BBSome function; namely, whether it regulates the assembly of the BBSome or is a stoichiometric component of the BBSome will be an interesting issue to address in the future, although these roles are not mutually exclusive.  manuscript. This work was supported in part by Grants-in-Aid for Scientific Research on Innovative Areas "Cilia and Centrosome" from the Ministry of Education, Culture, Sports, Science and Technology, Japan (grant number 15H01211 to K.N.); grants from the Japan Society for the Promotion of Science (JSPS) (grant numbers 15H04370 to K.N., 15K07929 to Y.K., and 16J03865 to S.N.); and grants from the Astellas Foundation for Research on Metabolic Disorders to K.N., and from the Takeda Science Foundation and the Uehara Memorial Foundation to Y.K. S.N. was supported by a JSPS Research Fellowship.