Interaction of Transducin with Uncoordinated 119 Protein (UNC119)

The key visual G protein, transducin undergoes bi-directional translocations between the outer segment (OS) and inner compartments of rod photoreceptors in a light-dependent manner thereby contributing to adaptation and neuroprotection of rods. A mammalian uncoordinated 119 protein (UNC119), also known as Retina Gene 4 protein (RG4), has been recently implicated in transducin transport to the OS in the dark through its interaction with the N-acylated GTP-bound transducin-α subunit (Gαt1). Here, we demonstrate that the interaction of human UNC119 (HRG4) with transducin is dependent on the N-acylation, but does not require the GTP-bound form of Gαt1. The lipid specificity of UNC119 is unique: UNC119 bound the myristoylated N terminus of Gαt1 with much higher affinity than a prenylated substrate, whereas the homologous prenyl-binding protein PrBP/δ did not interact with the myristoylated peptide. UNC119 was capable of interacting with Gαt1GDP as well as with heterotrimeric transducin (Gt). This interaction of UNC119 with Gt led to displacement of Gβ1γ1 from the heterotrimer. Furthermore, UNC119 facilitated solubilization of Gt from dark-adapted rod OS membranes. Consistent with these observations, UNC119 inhibited rhodopsin-dependent activation of Gt, but had no effect on the GTP-hydrolysis by Gαt1. A model for the role of UNC119 in the IS→OS translocation of Gt is proposed based on the UNC119 ability to dissociate Gt subunits from each other and the membrane. We also found that UNC119 inhibited activation of Go by D2 dopamine receptor in cultured cells. Thus, UNC119 may play conserved inhibitory role in regulation of GPCR-G protein signaling in non-visual tissues.

In rod photoreceptors, exposure to bright light causes translocation of the visual G protein, transducin from the photosensitive outer segments (OS) 2 to the inner compartments of the cells (reviewed in Refs. [1][2][3]. The light-dependent translocation of transducin is thought to play an important role in light-adaptation and neuroprotection (4,5). Significant advances have been made in understanding the mechanism of this phenomenon. The current evidence supports a simple diffusion model, whereby the activation of transducin by photoexcited rhodopsin (R*) causes dissociation of transducin-␣ (G␣ t1 ) and G␤ 1 ␥ 1 subunits allowing them to diffuse into the inner segment (1)(2)(3)(4)(5)(6)(7)(8)(9). However, translocated transducin must return to the OS during dark adaptation to restore rod sensitivity. This retrograde translocation occurs on a relatively slow time scale with a halflife of 2.5 h (4). The precise mechanism of transducin return to the OS in the dark is not known. Formation of heterotrimeric G t in the inner segment (IS) appears to be a prerequisite for correct transport of transducin to the OS. Heterotrimeric G t forms in the IS in the absence of R* following hydrolysis of G␣ t1 -bound GTP. GTP and GTP␥S both caused light-dependent transducin redistribution from the OS in permeabilized retinas, but only GTP-translocated G t returned to the OS in the dark (8). Likewise, the time course of the IS3 OS transport in the dark was exceedingly slow for the GTPase deficient G␣ t1 Q200L mutant (6). The kinetics of G␣ t1 and G␤ 1 ␥ 1 return to the OS in the dark is identical, supporting the transport of G t as the heterotrimer (4). In dark-adapted G␣ t1 knock-out mice G␤ 1 ␥ 1 is mislocalized and spread throughout the photoreceptor cells (10). Furthermore, G␣ t1 is severely down-regulated and mislocalized in G␥ 1 knock-out mice (11). Thus, G␣ t1 and G␤ 1 ␥ 1 depend on each other for proper targeting.
Transducin may return to the OS in dark by a motor-driven mechanism, diffusion or a combination of the two transport modes. The apparent energy-independence of the G t return transport suggests diffusion as a primary mode (12). G t is modified with two lipid anchors, fatty acyl at the N-terminal Gly of G␣ t1 and thioether-linked farnesyl attached to the C terminus of G␥ 1 (13)(14)(15). Two lipid anchors are generally thought to be sufficient for stable membrane attachment of a protein that precludes protein dissociation and diffusion in the cytosol (16 -19). Intriguingly, G t is capable of significant interdisc transfer in the OS (20,21). We hypothesized that the longitudinal diffusion of G␣ t1 ␤ 1 ␥ 1 , which may allow to "refill" apical discs with G t during dark adaptation, is facilitated by sequestration of one the lipid anchor(s) via interactions with a binding partner (21). The same mechanism may underlie the entire IS3 OS route of G t in the dark. Potential partners include phosducin and prenylbinding protein PrBP/␦ also known as PDE␦ (22)(23)(24)(25). PrBP/␦, in particular, was shown to be critical for the IS3 OS transport of a number of prenylated proteins (26).
UNC119, a mammalian ortholog of Caenorhabditis elegans unc-119 (27), also known as Retina Gene 4 protein (RG4), has recently emerged as the protein essential for transducin transport in darkness (28). Human UNC119 (HRG4) was originally identified in a screen for candidate retinal degeneration genes (29). The C-terminal domain of UNC119 shares significant sequence and structural homology with PrBP/␦ (28,30). UNC119 is relatively abundant in the photoreceptor synapses and the IS (31), but also is expressed in a number of other tissues (32,33). Truncation mutation in UNC119 that deletes the PrBP/␦ homology domain has been linked to cone-rod dystrophy in human patients (34). A transgenic mouse model of the human mutation revealed severe synaptic degeneration (34). In contrast, a knock-out mouse model of UNC119 (MRG4) revealed a different dysfunction at distal IS/OS regions that caused a slowly progressing retinal degeneration (35). Recent proteomic study has reported reduction in UNC119 expression as one of the possible culprits contributing to retinal degeneration in a mouse transgenic model (36). UNC119 was reported to interact with the N terminus of the GTP-bound G␣ t1 in an acylation-dependent manner (28). Importantly, the return of G t to the OS in the dark was impaired in UNC119 knock-out mice (28). UNC119 was found to inhibit the GTPase activity of G␣ t1 . Hence, the proposed model for transducin return to the OS in darkness is based on diffusion of the stable UNC119-G␣ t1 GTP complex (28). Here, we examined the interaction of human UNC119 with transducin and the lipid specificity of UNC119 in comparison to that of PrBP/␦ in order to gain insights into the mechanism of IS3 OS transport of G t in the dark.
Preparation of ROS Membranes, G␣ t1 GDP, G␤ 1␥1 , and Chimeric G␣ t1 -Bovine ROS membranes were prepared as previously described (37). Urea-washed ROS membranes (uROS) were prepared according to protocol in Yamanaka et al. (38). G␣ t1 GDP was prepared and purified according to published protocol (39). Recombinant G␤ 1 ␥ 1 were expressed using the baculovirus/sf9 cell system. The G␤ 1 baculoviral stock was obtained from Dr. S. Chen (University of Iowa). To generate G␥ 1 baculoviruses, the G␥ 1 cDNA was PCR amplified from bovine retinal library with the introduction of the coding sequence for the N-terminal His 6 tag, and cloned into the pFast HTb vector using RsrII/NheI sites. Generation of the recombinant bacmids, transfection of Sf9 cells, and viral amplifications were carried out according to the manufacturer's recommendations (Invitrogen). For expression of the G␤ 1 ␥ 1 heterodimer, Sf9 cell cultures (2 ϫ 10 6 cells/ml) were co-infected with G␤ 1 and G␥ 1 baculoviruses at MOI of 4 -6. The G␤ 1 ␥ 1 heterodimer was purified using affinity chromatography on Ni-NTA resin (Novagen) as previously described (40).
The G␣ ti chimera (G␣ t1 *) with the His 6 sequence inserted between Met 115 and Pro 116 of the helical domain of Ghi8 (41) was generated to allow the N-terminal myristoylation of the protein. The G␣ t1 1-115 sequence was PCR-amplified from the Chi8 template using a 5Ј-primer with an NcoI site and a 3Ј-primer coding the His 6 sequence added to G␣ t1 specific sequence. The Chi8 116 -350 sequence was PCR-amplified using a 5Ј-primer with the His 6 -sequence added to the G␣ t1specific sequence, and a 3Ј-primer containing an XhoI site. The two resulting PCR products were used in the PCR reaction with the flanking primers, and the PCR product was subcloned into the pET15b vector using the NcoI/XhoI sites. G␣ t1 * was expressed and purified as previously described (41). To obtain myristoylated G␣ t1 * (myrG␣ t1 *), BL21-codon plus Escherichia coli cells were co-transformed with kanamycin-resistant plasmid pbb131 expressing yeast N-myristoyl transferase (42).
Cloning, Expression, and Purification of Human UNC119 and PrBP/␦-For expression of the His 6 -tagged UNC119, UNC119  , and PrBP/␦, corresponding cDNAs were PCR amplified from a human retina cDNA library and subcloned into the pET15b vector using NcoI/XhoI sites (UNC119) or NdeI/BamHI sites (PrBP/␦). The N-terminally StrepII-tagged UNC119  was amplified from a human retina cDNA library with a 5Ј primer coding an NcoI site and the tag sequence WSHPQFEK with the SGG linker and a 3Ј primer containing a XhoI site. The PCR product was subcloned sites into the pET15b vector using NcoI/XhoI sites. Protein expression in BL21-codon plus E. coli cells was induced with the addition of 30 M IPTG. UNC119 and StrepII-UNC119  were expressed overnight at 16°C, whereas UNC119  and PrBP/␦ were expressed for 5 h at 30°C. The His 6 -tagged proteins were purified on Ni-NTA resin (Novagen). StrepII-UNC119 55-240 was purified using StrepTactin Superflow Agarose (Novagen). UNC119  and PrBP/␦ were additionally purified by gel-filtration on a Superdex 75 Sepharose column, and UNC119 was purified by an ion-exchange chromatography on a Uno-Q1 column (Bio-Rad).
Preparation of Hypotonic, GTP, and GTP␥S Extracts of ROS Membranes-To extract native transducin without exposure to photobleached R* and dissociation-reassociation of G␣ t1 and G␤ 1 ␥ 1 (43), ROS membranes (200 l, 160 M R) were washed two times with 2 ml of isotonic buffer A (20 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 1 mM DTT, 0.1 mM PMSF and 120 mM KCl) under dim red light. The pellet after centrifugation was resuspended in 1 ml of hypotonic buffer B (5 mM Tris-HCl, pH 7.5, 0.1 mM EDTA, 1 mM DTT, and 0.1 mM PMSF) and homogenized in a 2 ml tube using a disposable pestle under dim red light. To obtain GTP or GTP␥S extracts of G t , ROS membranes (200 l, 160 M R) were photobleached under fluorescent light for 30 min on ice and resuspended in 1 ml of buffer A containing either 100 M GTP or GTP␥S and homogenized in a 2-ml microcentrifuge tube using a disposable pestle. The resulting mixtures were centrifuged at 125,000 ϫ g for 30 min at 4°C. Supernatants of hypotonic, GTP, and GTP␥S extracts were collected and dialyzed against buffer A at 4°C for 4 h. Dialyzed extracts were centrifuged at 125,000 ϫ g for 30 min at 4°C, aliquoted, and stored at Ϫ20°C.
ROS Extraction with UNC119-ROS membranes washed twice with buffer A were resuspended in 1 ml buffer A containing 5 M purified UNC119 and incubated on ice for 30 min in the dark. The mixture was homogenized using a disposable pestle and centrifuged at 125,000 ϫ g for 30 min at 4°C. Cleared supernatant was dialyzed against buffer A at 4°C for 4 h. Dialyzed UNC119 extract of ROS was cleared by centrifugation, aliquoted, and stored at Ϫ20°C.
Pull-down Assays-StrepII-UNC119 55-240 was expressed as described above. Cell pellet was resuspended in Strep-Tactin binding buffer C (100 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA) and then disrupted by ultrasonication. The cleared supernatant after centrifugation (100,000 ϫ g, 60 min) was incubated with 200 l of StrepTactin Superflow Agarose resin for 1 h at 4°C with gentle shaking. StrepII-UNC119 55-240bound resin was washed thoroughly with buffer C to remove unbound proteins. For pull-down assay, 15 l of StrepII-UNC119 55-240 -bound resin was incubated with 5 g each of purified G␣ t1 GDP, G␣ t1 *, and myrG␣ t1 * or with hypotonic, GTP, and GTP␥S ROS extracts (10 g protein each) for 30 min at room temperature in 1.5-ml microcentrifuge tubes. Resin was washed three times with buffer A to remove unbound proteins and 30 l of SDS-PAGE sample buffer was added to the resin. The samples were separated on 4 -12% gradient Bis-tris NuPage gels (Invitrogen).
Native PAGE-Samples were prepared by incubating hypotonic, GTP, and GTP␥S ROS extracts (typically 10 g proteins each) with 5 M His 6 -tagged UNC119 for 30 min at room temperature. Samples were run in 7.5% Mini-PROTEAN TGX precast acrylamide gels (Bio-Rad) at 25°C with constant voltage of 150 V for 45 min.
Single Turnover GTPase Assays-GTPase assays were carried out in suspensions of uROS (10 M R*) reconstituted with hypotonic ROS extract containing 1 M G t in 20 mM HEPES buffer (pH 7.4), containing 100 mM NaCl and 8 mM MgSO 4 similar as described (44). The GTPase reactions were initiated by addition of 100 nM [␥-32 P]GTP. After a 5-s interval, the reaction was stopped or allowed to proceed with or without the addition of UNC119 (3 or 12 M) for the indicated time intervals. The reactions were quenched by addition of 100 l of 7% perchloric acid. Nucleotides were then precipitated using charcoal, and free 32 P i formed was measured by liquid scintillation counting. After subtraction of fraction of GTP hydrolyzed at 5 s, the data were fit with Equation 1, where k is the rate constant of GTP hydrolysis.
Fast Kinetic BRET Assay-Agonist-dependent cellular measurements of bioluminescence resonance energy transfer (BRET) between masGRKct-Rluc8 and G␤ 1 ␥ 2 -Venus were performed to visualize the action of G protein signaling in living cells as previously described with slight modification (45,46). 293T/17 were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), MEM non-essential amino acids, 1 mM sodium pyruvate, and antibiotics (100 units/ml penicillin and 100 g/ml streptomycin) at 37°C in a humidified incubator containing 5% CO 2 . For transfection, cells were seeded into 6-cm dishes at a density of 4 ϫ 10 6 cells/dish. After 4 h, expression constructs (total 5 g/dish) were transfected into the cells using Lipofectamine LTX (8 l/dish) and PLUS (5 l/dish) reagents. D2 receptors, G␣o, Venus155-239-G␤, Venus1-155-G␥ 2 , masGRKct-Rluc8, myc-tagged UNC119 constructs were transfected at a 1:2:1:1: 1:6 ratio. BRET sensor constructs were gifts from Dr. Nevin A. Lambert (Department of Pharmacology and Toxicology, Medical College of Georgia). Empty vector was used to normalize the amount of transfected DNA. The cells were used for experiments at 16 -24 h after transfection. BRET measurements were made using a microplate reader (POLARstar Omega; BMG Labtech) equipped with two emission photomultiplier tubes, allowing us to detect two emissions simultaneously with highest possible resolution of 50 ms for every data point. All measurements were performed at room temperature.
Fitting of the experimental data were performed with nonlinear least squares criteria using GraphPad Prizm Software. The K d values are expressed as mean Ϯ S.E. for three independent measurements.

RESULTS
Interaction of UNC119 (HRG4) with Transducin Is Dependent on the N-terminal Lipid Modification and Independent of the Activation State of G␣ t1 -For the initial characterization of UNC119 interactions with transducin we utilized the N-terminally truncated Strep-tagged UNC119, UNC119  , which includes the sequence homologous to PrBP/␦. Pull-down experiments using UNC119  demonstrate that it binds to G␣ t1 GDP purified from bovine ROS (Fig. 1). To determine if myristoylation of G␣ t1 is required for the interaction with UNC119 55-240 , we constructed chimeric G␣ t1 (G␣ t1 *) corresponding to the N-terminally His-tagged Chi8 (41), but with the His 6 -tag placed within the helical domain of G␣ t1 between Met 115 and Pro 116 . Myristoylated G␣ t1 * (myrG␣ t1 *) was obtained using co-expression with N-myristoyl transferase (NMT) (42). In the pull-down assay, myrG␣ t1 * bound UNC119 55-240 similar to bovine G␣ t1 GDP. Only trace amounts of nonmyristoylated G␣ t1 * precipitated with the UNC119 55-240 resin (Fig. 1). Thus, the interaction of UNC119 55-240 with G␣ t1 is largely dependent on the N-acylation of the protein.
Interactions of UNC119 with holo-G t and activated G␣ t1 GTP␥S were investigated using extracts of ROS membranes with GTP and GTP␥S. Following extraction with GTP, G␣ t1 hydrolyzes the nucleotide leading to the formation of heterotrimeric G␣ t1 GDP⅐G␤ 1 ␥ 1 or G t . The pull-down experiments showed that UNC119 55-240 interacted with G␣ t1 GDP from heterotrimeric G t and G␣ t1 GTP␥S ( Fig. 2A). The ratio of G␤ 1 ␥ 1 to G␣ t1 in the pull-down complex was markedly reduced compared with that in the ROS extracts ( Fig. 2A). The trace amounts of G␤ 1 ␥ 1 were precipitated not as a complex with G␣ t1, but rather due to nonspecific interaction with the Strep-Tactin agarose resin (Fig. 2B). Therefore, interaction with UNC119 with G t apparently leads to dissociation of G␣ t1 and G␤ 1 ␥ 1 . Recoverin, a prominent myristoylated protein in photoreceptor cells, was present in the ROS extracts, but negligible in the pull-down preparations (Fig. 2C).
Interaction of UNC119 with Native Transducin-Extraction of ROS with GTP yields "reconstituted" transducin whose properties differ from properties of "native" transducin (43). "Native" transducin is obtained without exposure to photobleached R* and dissociation-reassociation of G␣ t1 and G␤ 1 ␥ 1 (43). To probe the interaction of UNC119 with "native" G t , the UNC119 55-240 pull-downs were carried out using the hypotonic extract of "dark" ROS membranes. Similarly to the pulldowns from the GTP and GTP␥S ROS extracts, G␣ t1 and minor quantities of G␤ 1 ␥ 1 were pelleted from the hypotonic ROS extract with the StrepII-UNC119 55-240 resin (Fig. 3A). Thus, UNC119 55-240 was able to compete with G␤ 1 ␥ 1 for the interaction with G␣ t1 in "native" transducin as well. Next, we examined the effect of UNC119 on "native" G t on "dark" ROS membranes. Addition of UNC119 or UNC119  to ROS membranes in isotonic buffer caused significant release of G␣ t1 and G␤ 1 ␥ 1 into soluble fraction (Fig. 3B). The levels of recoverin in the soluble fraction remained unchanged in the presence of UNC119 (Fig. 3B). These data indicate that UNC119 interacts with membrane-bound G t and detaches G␣ t1 and G␤ 1 ␥ 1 from the membrane.
Further analysis of UNC119 interactions with "reconstituted" G t , G␣ t1 GTP␥S, and "native" transducin was performed by native gel electrophoresis. G␣ t1 from the GTP-and hypotonic ROS extracts (Fig. 4, A and C) migrated similarly and slightly faster than G␣ t1 in the GTP␥S extract (Fig. 4B). In the all three extract preparations, G␤ 1 ␥ 1 did not co-migrate with G␣ t1 , as it was previously shown for the "reconstituted" G t (43). Addition of UNC119 to each of the ROS extracts strikingly shifted positions of G␣ t1 on the gel toward the position of UNC119. The G␣ t1 band shifted to similar slow migrating bands in the GTP-or hypotonic ROS extract and isotonic UNC119 extract of dark ROS membranes (Fig. 4, A, C, and D). A somewhat different G␣ t1 pattern was seen in the GTP␥S extract with addition of UNC119 (Fig. 4B). Migration of G␤ 1 ␥ 1 on native gel was not significantly altered in the presence of UNC119 (Fig. 4).
Effects of UNC119 on G-protein Activation and Inactivation-The apparent ability of UNC119 to interact with heterotrimeric G t and dissociate G␣ t1 and G␤ 1 ␥ 1 suggests that it might interfere with transducin activation by R*. Indeed, the GTP␥S-binding assays indicated that UNC119 inhibits transducin activation by R*-containing uROS (Fig. 5A). UNC119 is also capable of interaction with the active GTP␥S-bound conformation of G␣ t1 and may affect the GTP hydrolysis by G␣ t1 . Single-turnover GTPase assay revealed no effect of UNC119 on the G␣ t1 GTPase catalytic rate (Fig. 5B).
To investigate the function of UNC119 in living cells, we employed a fast kinetic BRET assay of heterotrimeric G o acti-vation/inactivation by the D2 dopamine receptor D2R in transfected HEK293 cells (45,46). In this study we tested the effect of UNC119 proteins on D2R-G o -mediated signaling by measuring amplitude as well as kinetics of agonist-induced interaction between G␤ 1 ␥ 2 -Venus and its effector construct, masGRKct-Rluc8, as a readout of G protein activation and deactivation state. We first observed that the dopamine-induced maximum amplitude was decreased in UNC119-expressing cells (Fig. 6B). Also, UNC119 significantly reduced the rate of G o activation, but had very little effect on the G-protein inactivation (Fig. 6, A and C). In our time-resolved measurements, UNC119 exerts an effect on the D2-G o signaling that is supporting the observations of negative regulation of GEF activity by UNC119 described in biochemical experiments (Fig. 5A).
Comparison of the Lipid Binding Properties of UNC119 and PrBP/␦-To quantitatively assess the interactions of UNC119 with G␣ t1 , myristoylated N-terminal G␣ t1 peptide, [MYR]-GCG-ASAEEK with Ala3 substituted for Cys was synthesized to allow the labeling with a fluorescence probe, BC. Addition of both UNC119 and UNC119 55-240 to [MYR]-GC[BC]GASAEEK led to large, dose-dependent increases of the probe fluorescence (Fig. 7). From the fluorescence binding curves, the K d values for UNC119 and UNC119  were 185 Ϯ 15 nM and 270 Ϯ 10 nM, respectively (Fig. 7B). The binding of BC-labeled peptide to UNC119 was specific since no fluorescence change occurred on addition of UNC119 preincubated with excess unlabeled [MYR]-GCGASAEEK (not shown). Comparable affinities of UNC119 and UNC119 55-240 for the G␣ t1 N terminus suggest that it binds primarily to the UNC119 region homologous to PrBB/␦. Next   Homology between UNC119 and PrBP/␦ suggests that UNC119 may potentially bind prenylated proteins. This possibility was examined with the binding assay employing fluorescently labeled farnesylated Cys probe, farnesyl-Cys-AMCA. Binding of UNC119 and UNC119  to farnesyl-Cys-AMCA increased the probe fluorescence in a dose dependent manner (Fig. 8). No fluorescence increase was observed when UNC119 was preincubated with unlabeled farnesylated probe (not shown). The K d of UNC119 binding to farnesyl-Cys-AMCA was estimated at 3.8 M, which is an order of magnitude higher than the K d for the UNC119 interaction with the myristoylated N terminus of G␣ t1 . The same assay yielded a K d of 0.55 M for farnesyl-Cys-AMCA binding to PrBP/␦ (Fig. 8). A comparable binding affinity (K d 0.36 M) was obtained for farnesyl-Cys-AMCA binding to PrBP/␦ using the Trp-fluorescence energy transfer assay (supplemental Fig. S1). This affinity is similar to the affinity of dansyl farnesyl binding to PrBP/␦ reported previously (47). Thus, farnesyl-Cys-AMCA is an appropriate reporter of the interactions of UNC119 and PrBP/␦ with farnesylated probes.

DISCUSSION
The energy independence and the massive amounts of transducin returning to the rod outer segment during dark adaptation following exposure to bright light favor some sort of a diffusion mechanism as a transducin transport mode (4,12). Appreciable inter-disc diffusion of heterotrimeric G t modified with two lipid anchors might be possible when one or both of the lipids are sequestered by lipid-binding proteins (21). Indeed, Zhang et al. have recently demonstrated that UNC119 interacts with the N-acylated G␣ t1 GTP, and that the G t trans-   port in the dark is impaired in the UNC119 (MRG4) knock-out mice (28). A "restricted" diffusion model has been proposed in which the rate-limiting step is defined by very slow spontaneous GTP/GDP exchange on G t in the IS (28). UNC119 then interacts with and stabilizes G␣ t1 GTP by inhibiting intrinsic GTPase activity and allowing the UNC119-G␣ t1 GTP complex to diffuse to the OS. Diffusion of G␤ 1 ␥ 1 released by spontaneous activation is facilitated by PrBP/␦ (28). Although a slow spontaneous nucleotide exchange in the absence of R* (k Ͻ 10 Ϫ4 s Ϫ1 ) was observed in isolated preparations of G t (48), it is unclear that it occurs in G t docked to IS membranes under in vivo like conditions. Our results suggest an alternative model for transducin transport to the OS in the dark (supplemental Fig. S2). UNC119 is capable of interacting with heterotrimeric G t (Figs. 2-4). It facilitates transducin release from the membrane accompanied by dissociation of UNC119-G␣ t1 and G␤ 1 ␥ 1 . The co-crystal structure of UNC119 with the lauroylated N-terminal G␣ t1 peptide shows UNC119 interactions with the lipid as well as the N-terminal 6 -10 residues of G␣ t1 (28). Thus, UNC119 apparently disrupts the interaction of transducin subunits by sterically occluding the G␤ 1 ␥ 1 -binding site within the N-terminal ␣N-helix of G␣ t1 (49). We found that UNC119 interacts weakly with farnesylated probe, farnesyl-Cys-AMCA (Fig. 8). Still, in agreement with the previous study, UNC119 did not appreciably bind farnesylated G␤ 1 ␥ 1 (28). In contrast, PrBP/␦ bound farnesyl-Cys-AMCA much more potently then UNC119 (Fig. 8). G␤ 1 ␥ 1 is partially mislocalized in PrBP/␦ knock-out mice (26). Therefore, PrBP/␦ is a probable facilitator of diffusion of G␤ 1 ␥ 1 released from G t by UNC119. Phosducin, another photoreceptor G␤ 1 ␥ 1 -binding protein, seems to play a different role. Phosducin assists light-dependent translocation of transducin from the OS to the IS by sequestering G␤ 1 ␥ 1 (22). However, in the IS, phosducin is phosphorylated in the dark and releases G␤ 1 ␥ 1 (23,50). Thus, phosphorylation of phosducin serves as a trigger for the formation of heterotrimeric G t in the IS (2,23). Formation and membrane association of G t act as a transducin IS "sink" and as a starting point for the G t to the OS return (2,28). Supporting this hypothesis, transducin return to the OS in the dark is delayed in mice expressing phosphorylation-deficient phosducin (51).
Our model of transducin return to the OS in the dark does not assume the requirement for G t spontaneous activation in the IS. The kinetics of transducin IS3 OS translocation are much slower then the light-induced OS3 IS translocation (4). What is the rate-limiting step in this process? One possibility is that the molar ratio of UNC119 to light-translocated G t in the IS is low. As a result, only a small fraction of G t is solubilized by UNC119 from IS membranes and is diffusing at any given moment during dark adaptation. The levels of UNC119 in the OS are clearly much lower than they are in the IS (35). Accordingly, in the OS the equilibrium between UNC119-G␣ t1 and G␣ t1 ␤ 1 ␥ 1 shifts toward the heterotrimer which associates with disc membranes. It remains to be investigated whether UNC119 assists the axial diffusion of G t in the OS. In addition to transducin trafficking, UNC119 may play important roles in regulation of vesicle and ciliary trafficking processes in photo-  receptor cells. Protein myristoylation may mediate UNC119 binding to known partners such as Arf-like proteins 2 and 3 (Arl/2/3) (52,53). Arl2/3 are present in the IS and the connecting cilium, respectively, and are shown to regulate microtubule dependent processes (52,54,55).
Besides transducin, other G proteins from the G i family (G␣ i , G␣ o , G␣ gust ) are myristoylated (16,17). UNC119 inhibited R*-dependent activation of G t and D2R-dependent activation of G o , indicating that UNC119 may inhibit GPCR-dependent activation of the G i family G proteins (Figs. 5 and 6). The mechanism presumably involves G-protein membrane extraction and subunit dissociation. We found no effect of UNC119 on the G␣ t1 GTPase activity under single turnover conditions that allow measurements of GTP hydrolysis independent of the GTP binding reaction. Furthermore, UNC119 did not affect inactivation of G␣ o in cultured cells (Fig. 6). These results contrast the reported inhibitory effect of UNC119 on transducin GTPase activity observed by multi-turnover GTPase assay (28). The ability of UNC119 to inhibit R*-dependent GTP␥S binding to G t possibly explains this discrepancy (Fig. 5). Low-level expression of UNC119 was reported in adrenal gland, cerebellum, cultured fibroblasts, T-cells, lung, and kidney (32,33). Our findings suggest that apart from its role as a cofactor of G-protein trafficking in ciliary sensory cells (28), UNC119 may modulate signal transduction from GPCRs to G proteins in different cell types and tissues.