Portal de Periódicos da CAPES

Conformational Changes in the Phosphorylated C-terminal Domain of Rhodopsin during Rhodopsin Arrestin Interactions

2004; Elsevier BV; Volume: 279; Issue: 49 Linguagem: Inglês

10.1074/jbc.m407341200

1083-351X

Oleg G. Kisselev, Maureen A. Downs, J. Hugh McDowell, Paul A. Hargrave,

Photoreceptor and optogenetics research

Phosphorylation of activated G-protein-coupled receptors and the subsequent binding of arrestin mark major molecular events of homologous desensitization. In the visual system, interactions between arrestin and the phosphorylated rhodopsin are pivotal for proper termination of visual signals. By using high resolution proton nuclear magnetic resonance spectroscopy of the phosphorylated C terminus of rhodopsin, represented by a synthetic 7-phosphopolypeptide, we show that the arrestin-bound conformation is a well ordered helix-loop structure connected to rhodopsin via a flexible linker. In a model of the rhodopsin-arrestin complex, the phosphates point in the direction of arrestin and form a continuous negatively charged surface, which is stabilized by a number of positively charged lysine and arginine residues of arrestin. Opposite to the mostly extended structure of the unphosphorylated C-terminal domain of rhodopsin, the arrestin-bound C-terminal helix is a compact domain that occupies a central position between the cytoplasmic loops and occludes the key binding sites of transducin. In conjunction with other binding sites, the helix-loop structure provides a mechanism of shielding phosphates in the center of the rhodopsin-arrestin complex and appears critical in guiding arrestin for high affinity binding with rhodopsin. Phosphorylation of activated G-protein-coupled receptors and the subsequent binding of arrestin mark major molecular events of homologous desensitization. In the visual system, interactions between arrestin and the phosphorylated rhodopsin are pivotal for proper termination of visual signals. By using high resolution proton nuclear magnetic resonance spectroscopy of the phosphorylated C terminus of rhodopsin, represented by a synthetic 7-phosphopolypeptide, we show that the arrestin-bound conformation is a well ordered helix-loop structure connected to rhodopsin via a flexible linker. In a model of the rhodopsin-arrestin complex, the phosphates point in the direction of arrestin and form a continuous negatively charged surface, which is stabilized by a number of positively charged lysine and arginine residues of arrestin. Opposite to the mostly extended structure of the unphosphorylated C-terminal domain of rhodopsin, the arrestin-bound C-terminal helix is a compact domain that occupies a central position between the cytoplasmic loops and occludes the key binding sites of transducin. In conjunction with other binding sites, the helix-loop structure provides a mechanism of shielding phosphates in the center of the rhodopsin-arrestin complex and appears critical in guiding arrestin for high affinity binding with rhodopsin. Following activation by a variety of sensory stimuli, such as hormones, neurotransmitters, or light, G-protein-coupled receptors (GPCRs) 1The abbreviations used are: GPCRs, G-protein-coupled receptors; GTPγS, guanosine 5′-3-O-(thio)triphosphate; GRKs, GPCR kinases; 7PP, 7-phosphopolypeptide; ROS, rod outer segment; PDB, Protein Data Bank. are deactivated by multiple phosphorylations and subsequent binding of a regulatory protein arrestin (1Luttrell L.M. Lefkowitz R.J. J. Cell Sci. 2002; 115: 455-465Crossref PubMed Google Scholar, 2Shenoy S.K. Lefkowitz R.J. Biochem. J. 2003; 375: 503-515Crossref PubMed Scopus (339) Google Scholar). Deactivation of the active receptor is obligatory and ensures the quantum character of the response to an extracellular signal. GPCR kinases (GRKs) and arrestin proteins are receptor-specific but share universal mechanisms of action, with prototypical rhodopsin kinase and visual arrestin involved in the termination of visual signal transduction (3Arshavsky V.Y. Trends Neurosci. 2002; 25: 124-126Abstract Full Text Full Text PDF PubMed Scopus (52) Google Scholar, 4Maeda T. Imanishi Y. Palczewski K. Prog. Retin. Eye Res. 2003; 22: 417-434Crossref PubMed Scopus (125) Google Scholar). Quenching of the photoresponse in the retinal photoreceptor cells proceeds through phosphorylation of multiple serine and threonine residues at the C terminus of light-activated rhodopsin by GRK1 and high affinity binding of visual arrestin. In vitro studies have implicated all seven serine and threonine residues within the Rh-(334–343) C-terminal stretch as possible substrates of the phosphorylation reaction. The question of the exact number and position of residues phosphorylated in vivo, however, remains uncertain possibly due to the inherent difficulties of controlling dephosphorylation, different kinetics of phosphorylation/dephosphorylation at individual residues, possible contribution of kinases other than GRK1, and other secondary factors. Either Ser-334 (5Ohguro H. Van Hooser J.P. Milam A.H. Palczewski K. J. Biol. Chem. 1995; 270: 14259-14262Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar) or Ser-343 (6Kennedy M.J. Lee K.A. Niemi G.A. Craven K.B. Garwin G.G. Saari J.C. Hurley J.B. Neuron. 2001; 31: 87-101Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar) is phosphorylated initially. The majority of studies provides strong evidence, however, that multiple phosphorylation is required for reproducible deactivation (5Ohguro H. Van Hooser J.P. Milam A.H. Palczewski K. J. Biol. Chem. 1995; 270: 14259-14262Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar, 6Kennedy M.J. Lee K.A. Niemi G.A. Craven K.B. Garwin G.G. Saari J.C. Hurley J.B. Neuron. 2001; 31: 87-101Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar, 7Ohguro H. Palczewski K. Ericsson L.H. Walsh K.A. Johnson R.S. Biochemistry. 1993; 32: 5718-5724Crossref PubMed Scopus (152) Google Scholar, 8Mendez A. Burns M.E. Roca A. Lem J. Wu L.W. Simon M.I. Baylor D.A. Chen J. Neuron. 2000; 28: 153-164Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar). In order to circumvent the current uncertainty about the exact sequence of phosphorylation events, we have used a model synthetic peptide phosphorylated at all seven positions, which mimics the major biological properties of phosphorylated rhodopsin. Conceptually, it is thought that arrestin directly competes for the binding site of transducin (9Pfister C. Chabre M. Plouet J. Tuyen V.V. De Kozak Y. Faure J.P. Kühn H. Science. 1985; 228: 891-893Crossref PubMed Scopus (194) Google Scholar, 10Kühn H. Wilden U. J. Recept. Res. 1987; 7: 283-298Crossref PubMed Scopus (82) Google Scholar, 11Krupnick J.G. Gurevich V.V. Benovic J.L. J. Biol. Chem. 1997; 272: 18125-18131Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar), a heterotrimeric GTP-binding protein responsible for relaying the signal to the cyclic GMP phosphodiesterase, an enzyme of the intracellular second messenger system. The mechanism of arrestin binding and signal shut-off at the molecular level, however, is not understood. The C-terminal region of rhodopsin is resolved poorly in available crystal structures. Phosphorylation of multiple residues can potentially have significant impact on the conformation of the C terminus and prearrange the Rh-(330–348) domain for effective interactions with arrestin. NMR studies using model peptides with various numbers of phosphates (12Dorey M. Hargrave P.A. McDowell J.H. Arendt A. Vogt T. Bhawsar N. Albert A.D. Yeagle P.L. Biochim. Biophys. Acta. 1999; 1416: 217-224Crossref PubMed Scopus (17) Google Scholar) and full protein (13Getmanova E. Patel A.B. Klein-Seetharaman J. Loewen M.C. Reeves P.J. Friedman N. Sheves M. Smith S.O. Khorana H.G. Biochemistry. 2004; 43: 1126-1133Crossref PubMed Scopus (37) Google Scholar) showed little ordering of this region of rhodopsin upon phosphorylation. Whether the phospho-Rh-(330–348) assumes a defined conformation in the rhodopsin-arrestin complex remains unclear. We posed the question whether the phosphorylated C terminus of rhodopsin is simply passive bait for arrestin or if it is directly involved in shaping the cytoplasmic surface of the receptor into a conformation associated with termination of a signaling state. We used high resolution proton NMR to study the dynamics of the model synthetic peptide, representing the fully phosphorylated region of rhodopsin Rh-(330–348), 7PP, in solution and in an arrestin-bound state. Rigid body docking of rhodopsin, containing the C-terminal region in a conformation determined from this experimental study, to arrestin allowed us to propose a model of the rhodopsin-arrestin complex. Arrestin and Phosphorylated Peptide—Arrestin was prepared by the method of Buczylko and Palczewski with modifications as described earlier (14Puig J. Arendt A. Tomson F.L. Abdulaeva G. Miller R. Hargrave P.A. McDowell J.H. FEBS Lett. 1995; 362: 185-188Crossref PubMed Scopus (72) Google Scholar). Peptide Rh-(330–348) from the sequence of bovine rhodopsin was synthesized by standard Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry on an Applied Biosystem model 431A peptide synthesizer. Multiply phosphorylated peptide 7-phospho-Rh-(330–348) (7PP) was synthesized on phenylacetamidomethyl polystyrene resin using t-butoxycarbonyl-O-(diphenylphosphono)-serine and -threonine as described (15Arendt A. McDowell J.H. Abdulaeva G. Hargrave P.A. Protein Pept. Lett. 1996; 3: 361-368Google Scholar). The peptides were homogeneous by high pressure liquid chromatography and showed the expected mass when examined by matrix-assisted laser desorption ionization time-of-flight-mass spectrometry (15Arendt A. McDowell J.H. Abdulaeva G. Hargrave P.A. Protein Pept. Lett. 1996; 3: 361-368Google Scholar). Rhodopsin Binding—Various amounts of the Gtβγ were reconstituted with 3 μg of Gtα and 30 μg of urea-washed ROS membranes in 100 μl of buffer ROS-ISO (10 mm Tris, pH 7.4, 100 mm NaCl, 5 mm MgCl2, 1 mm dithiothreitol, 0.1 mm phenylmethylsulfonyl fluoride) on ice in the absence or presence of 5 μg of bovine arrestin and 100 μm 7PP. The reaction was initiated by exposure to light. Urea-washed ROS membranes were centrifuged at 109,000 × g, 4 °C for 10 min in a TLA-100.3 rotor on a Beckman TL-100 ultracentrifuge. The pellet was washed twice with buffer ROS-ISO. Urea-washed ROS membranes with Gt bound was resuspended in buffer ROS-Hypo (10 mm Tris-HCl, pH 7.4, 0.5 mm MgCl2, 1 mm dithiothreitol, 0.1 mm phenylmethylsulfonyl fluoride) and 250 μm GTPγS, incubated on ice for 30 min, and centrifuged. The supernatant was analyzed for the presence of G-protein subunits by immunoblotting. NMR and Structure Calculations—NMR sample preparation and data processing were essentially as we described previously (16Kisselev O.G. McDowell J.H. Hargrave P.A. FEBS Lett. 2004; 564: 307-311Crossref PubMed Scopus (20) Google Scholar). NMR samples contained 0.16 mm purified arrestin, 1.77 mm 7PP, 7-phospho-Rh-(330–348), or unphosphorylated Rh-(330–348) in sodium phosphate buffer, 0.1 m, pH 6.5, and 10% D2O in a total volume of 0.6 ml. Two-dimensional high resolution proton Transferred Nuclear Overhauser Effect Spectra were acquired at 4 °C on the Varian Unity-600 spectrophotometer as described earlier (17Kisselev O.G. Kao J. Ponder J.W. Fann Y.C. Gautam N. Marshall G.R. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 4270-4275Crossref PubMed Scopus (163) Google Scholar). Data were processed off-line using VNMR 5.2. Total correlation spectroscopy (18Braunschweiler L. Ernst R.R. J. Magn. Res. 1983; 53: 521-528Crossref Scopus (3112) Google Scholar) (two-dimensional TOCSY, MLEV-17 mixing sequence of 120 ms, flanked by two 2-ms trim pulses, 0.5-s pre-acquisition delay, and 1.0-s presaturation) and two-dimensional NOESY (19Kumar A. Ernst R.R. Wüthrich K. Biochem. Biophys. Res. Commun. 1980; 95: 1-6Crossref PubMed Scopus (2032) Google Scholar) (Tmix = 250 ms, 2 × 280 × 2049 data matrix with 16 scans per t1, using WATERGATE water suppression protocol) were used for sequence-specific and stereo-specific assignments. NOEs were classified into weak, medium, and strong based on the peak volume and translated to corresponding interproton distances of 1.9–5.0, 1.9–3.5, and 1.9–2.7 Å, respectively. One hundred ninety eight constraints were used for structure calculations, which involved distance geometry (DISTGEOM of TINKER 3.9 (20Hodsdon M.E. Ponder J.W. Cistola D.P. J. Mol. Biol. 1996; 264: 585-602Crossref PubMed Scopus (140) Google Scholar)), 1000 K restrained molecular dynamics, simulated annealing (ANNEAL, 15 ps total time), and structure refinement (NEWTON, 0.001 root mean square gradient). Calculations utilized CHARMM force field. In order to identify biologically relevant conformations of 7PP in the arrestin-bound state, the structure calculations based on the NMR-derived constraints were performed on the full set of x-ray coordinates of rhodopsin (21Palczewski K. Kumasaka T. Hori T. Behnke C.A. Motoshima H. Fox B.A. Le Trong I. Teller D.C. Okada T. Stenkamp R.E. Yamamoto M. Miyano M. Science. 2000; 289: 739-745Crossref PubMed Scopus (5056) Google Scholar) with unresolved parts of loop C3 and the C-terminal region rebuilt in Insight II. The starting conformation of a rebuilt C terminus was mostly extended. Only region Rh-(324–348) was allowed to move during calculations. For the final energy refinements, restrained molecular dynamics, simulated annealing, and energy minimization protocols were applied to Rh-(330–348) separate from rhodopsin coordinates. One hundred structures were generated independently. Fifteen models were chosen based on the lowest energy and best local geometry and superimposed using main chain atoms. Coordinates of the ensemble and the NMR constraints file were deposited as PDB code 1TQK. Phosphates were added to a representative model in Insight II. Molecular rigid body docking was done manually in Insight II by using surface complementation and electrostatic and hydrophobic interactions as guidance. The final model was chosen based on the strongest charge-charge interactions that avoided steric clashes. Calculation of electrostatic potential, molecular surfaces, and graphics were in MOLMOL (22Koradi R. Billeter M. Wüthrich K. J. Mol. Graphics. 1996; 14: 51-55Crossref PubMed Scopus (6498) Google Scholar). 7-Phospho-Rh-(330–348) Peptide Mimics Phosphorylated C Terminus of Rhodopsin—On a functional level, the C terminus of bovine rhodopsin, the region spanning amino acids 330–348, which contains three serine and four threonine residues, acts as an independent domain with characteristics closely linked to the signal shut-off. Ablation of this region by genetic methods in mice leads to a phenotype in which photo-signal termination is dramatically compromised (23Chen J. Makino C.L. Peachey N.S. Baylor D.A. Simon M.I. Science. 1995; 267: 374-377Crossref PubMed Scopus (275) Google Scholar). The soluble polypeptide Rh-(330–348) has been synthesized chemically in fully phosphorylated form (7PP (15Arendt A. McDowell J.H. Abdulaeva G. Hargrave P.A. Protein Pept. Lett. 1996; 3: 361-368Google Scholar)) and shown to duplicate major functions of the native phosphorylated rhodopsin. The fully phosphorylated form was chosen because multiple phosphorylation was observed in vivo (6Kennedy M.J. Lee K.A. Niemi G.A. Craven K.B. Garwin G.G. Saari J.C. Hurley J.B. Neuron. 2001; 31: 87-101Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar, 8Mendez A. Burns M.E. Roca A. Lem J. Wu L.W. Simon M.I. Baylor D.A. Chen J. Neuron. 2000; 28: 153-164Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar), and both serine and threonine residues have been shown to be required for activation of the action of arrestin (5Ohguro H. Van Hooser J.P. Milam A.H. Palczewski K. J. Biol. Chem. 1995; 270: 14259-14262Abstract Full Text Full Text PDF PubMed Scopus (173) Google Scholar, 24Brannock M.T. Weng K. Robinson P.R. Biochemistry. 1999; 38: 3770-3777Crossref PubMed Scopus (26) Google Scholar, 25McDowell J.H. Nawrocki J.P. Hargrave P.A. Biochemistry. 1993; 32: 4968-4974Crossref PubMed Scopus (86) Google Scholar, 26McDowell J.H. Robinson P.R. Miller R.L. Brannock M.T. Arendt A. Smith W.C. Hargrave P.A. Investig. Ophthalmol. Vis. Sci. 2001; 42: 1439-1443PubMed Google Scholar). 7PP inhibited phototransduction from photoactivated unphosphorylated rhodopsin in the presence of arrestin, based on the phosphodiesterase assay (27McDowell J.H. Smith W.C. Miller R.L. Popp M.P. Arendt A. Abdulaeva G. Hargrave P.A. Biochemistry. 1999; 38: 6119-6125Crossref PubMed Scopus (34) Google Scholar). It also induced conformational changes in arrestin typically seen upon interaction of arrestin with native phosphorylated rhodopsin (14Puig J. Arendt A. Tomson F.L. Abdulaeva G. Miller R. Hargrave P.A. McDowell J.H. FEBS Lett. 1995; 362: 185-188Crossref PubMed Scopus (72) Google Scholar). As predicted from the previous data, we show here that in direct competition experiments, binding of transducin to light-activated unphosphorylated rhodopsin in membranes is inhibited by addition of arrestin and 7PP together but not arrestin or 7PP alone (Fig. 1). These experiments demonstrate close functional approximation between the effects of rhodopsin phosphorylation in vivo and the effects of the model phosphopeptide in the presence of unphosphorylated rhodopsin in vitro. They also indicate that major structural features of the phosphorylated C terminus of rhodopsin are preserved in 7PP and that it can be used to study the structural dynamics of the rhodopsin-arrestin interface and the mechanism of signal shut-off. NMR Structures of 7PP in the Arrestin-bound State—To identify the structural basis of an inhibitory effect of arrestin in the arrestin-rhodopsin complex, high resolution proton NMR spectroscopy was used to determine the solution structure of 7PP and the changes in 7PP conformation upon binding to arrestin. As we reported, no secondary structure elements can be recognized in 7PP in solution, indicating high flexibility and a disordered state of the molecule (16Kisselev O.G. McDowell J.H. Hargrave P.A. FEBS Lett. 2004; 564: 307-311Crossref PubMed Scopus (20) Google Scholar). The result is not surprising, because of the mostly extended conformation of the C-terminal region in available x-ray structures of rhodopsin with the average B-factor values for the Rh-(330–348) region of 70.4 ± 8.8 (PDB code 1HZX) (21Palczewski K. Kumasaka T. Hori T. Behnke C.A. Motoshima H. Fox B.A. Le Trong I. Teller D.C. Okada T. Stenkamp R.E. Yamamoto M. Miyano M. Science. 2000; 289: 739-745Crossref PubMed Scopus (5056) Google Scholar). The disorder of 7PP in solution is also consistent with previous solution NMR experiments of model peptides (12Dorey M. Hargrave P.A. McDowell J.H. Arendt A. Vogt T. Bhawsar N. Albert A.D. Yeagle P.L. Biochim. Biophys. Acta. 1999; 1416: 217-224Crossref PubMed Scopus (17) Google Scholar) and a recent study (13Getmanova E. Patel A.B. Klein-Seetharaman J. Loewen M.C. Reeves P.J. Friedman N. Sheves M. Smith S.O. Khorana H.G. Biochemistry. 2004; 43: 1126-1133Crossref PubMed Scopus (37) Google Scholar) combining solution and solid state NMR of native phosphorylated rhodopsin. Addition of arrestin under the same experimental conditions resulted in striking changes in 7PP conformation, as evident from the NOESY spectra before and after addition of arrestin (Fig. 2a) (16Kisselev O.G. McDowell J.H. Hargrave P.A. FEBS Lett. 2004; 564: 307-311Crossref PubMed Scopus (20) Google Scholar). One hundred ninety eight constraints were identified and used in generation of a first subset of structures consistent with the NMR data by distance geometry and high temperature simulated annealing. The summary of the experimental constraints is shown in Fig. 2b. The analysis of a family of 7PP conformations in an arrestin-bound state showed a well ordered C-terminal helix for residues Glu-341 to Ala-346 and a fairly disordered loop for residues Asp-330 to Thr-340 (Fig. 2). Energy refinements lead to significant improvements of the local geometry and the overall quality of structures, compared with the initial set of models reported. The latest models have more than 83% of residues in the most favored regions of the Ramachandran plot and about 17% of residues in additionally allowed regions (Fig. 3), compared with 71 and 29%, respectively, for the first generation of structures (16Kisselev O.G. McDowell J.H. Hargrave P.A. FEBS Lett. 2004; 564: 307-311Crossref PubMed Scopus (20) Google Scholar).Fig. 3Ramachandran plot and structure statistics for the refined ensemble of arrestin-bound 7PP.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Restricted mobility and tighter structural organization of the C-terminal region, evident from small root mean square deviation values for residues 340–348 (Fig. 3), point to a more prominent role of the C-terminal helix in arrestin binding. Because any structuring is absent in an unphosphorylated peptide or 7PP without arrestin, it appears the C-terminal helix containing phosphates at Thr-342 and Ser-343 is ultimately involved in binding and, thus, is more constrained by the geometry of the binding site on arrestin. This observation correlates well with the proposed sequence of phosphorylation events progressing from Ser-343 to Ser-334 (6Kennedy M.J. Lee K.A. Niemi G.A. Craven K.B. Garwin G.G. Saari J.C. Hurley J.B. Neuron. 2001; 31: 87-101Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar). The first phosphates added are known to have the most effect on the affinity of binding. We predict that the structural elements of the helix loop structure of 7PP would be detectable with only Ser-343 phosphorylated. In support of this prediction, Ser-343 was shown to be phosphorylated most rapidly in vivo to produce the dominant monophosphorylated form of rhodopsin within the first seconds after light activation (6Kennedy M.J. Lee K.A. Niemi G.A. Craven K.B. Garwin G.G. Saari J.C. Hurley J.B. Neuron. 2001; 31: 87-101Abstract Full Text Full Text PDF PubMed Scopus (129) Google Scholar). A prominent feature of the arrestin-bound structure is the position of seven phosphates that are clustered on the loop connecting the disordered N-terminal region and helical C terminus. Remarkably, in the models calculated in the context of the whole rhodopsin, the phosphate groups are facing in one direction, away from rhodopsin, forming a unified negatively charged surface (Fig. 4). Carboxyl groups of Asp-330, Glu-332, and Glu-341 additionally contribute to this surface. Such orientation of phosphates is not observed for the peptide in the absence of arrestin and would be highly unlikely in the aqueous environment for the 7PP polypeptide alone, because of the strong negative charge repulsion. It appears possible because of the interactions of the phosphate groups with the complementary positively charged binding site of arrestin. A Model of Rhodopsin-Arrestin Interactions—Modeling of the cytoplasmic surface of rhodopsin based on the NMR-derived constraints shows that, in the arrestin-bound state, the C-terminal helix occupies a central position between three cytoplasmic loops of rhodopsin. This crevice is thought to open up after rhodopsin photoactivation to form a site for transducin docking (28Farrens D.L. Altenbach C. Yang K. Hubbell W.L. Khorana H.G. Science. 1996; 274: 768-770Crossref PubMed Scopus (1117) Google Scholar, 29Sheikh S.P. Zvyaga T.A. Lichtarge O. Sakmar T.P. Bourne H.R. Nature. 1996; 383: 347-350Crossref PubMed Scopus (399) Google Scholar). The arrestin-bound conformation also completely occludes the N terminus of helix eight, one of the major sites of transducin interaction (30Marin E.P. Krishna A.G. Zvyaga T.A. Isele J. Siebert F. Sakmar T.P. J. Biol. Chem. 2000; 275: 1930-1936Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar). Formation of a compact helical structure at the C terminus of rhodopsin upon arrestin binding may facilitate interactions between arrestin and additional sites of arrestin interactions on loops two and three. As mentioned above, the concerted position of phosphates in an arrestin-bound conformation of 7PP shows that the phosphates must be completely shielded from solvent in the arrestin-rhodopsin complex. This conclusion is consistent with data demonstrating that arrestin prevents de-phosphorylation of rhodopsin by phosphatase (31Palczewski K. McDowell J.H. Jakes S. Ingebritsen T.S. Hargrave P.A. J. Biol. Chem. 1989; 264: 15770-15773Abstract Full Text PDF PubMed Google Scholar). To visualize the position of the phosphates and to estimate the overall geometric plausibility of the rhodopsin-arrestin complex, we performed molecular rigid body docking of available crystal structures of arrestin (PDB 1CF1) with phosphorylated rhodopsin models produced by this study. The model of the complex does not take into consideration global conformational changes in arrestin, which are expected based on the biochemical studies (32Gurevich V.V. Gurevich E.V. Trends Pharmacol. Sci. 2004; 25: 105-111Abstract Full Text Full Text PDF PubMed Scopus (290) Google Scholar), but rather reflects initial stages of the contact between the two proteins. A prominent saddle on arrestin with the highest concentration of positively charged lysine and arginine residues was chosen as a complementary site for the phosphates on the rhodopsin C-terminal domain (Fig. 5). The polar core of this site contains Lys-14, Lys-15, and Arg-175, which are proposed to interact with phosphates on rhodopsin (33Vishnivetskiy S.A. Schubert C. Climaco G.C. Gurevich Y.V. Velez M.G. Gurevich V.V. J. Biol. Chem. 2000; 275: 41049-41057Abstract Full Text Full Text PDF PubMed Scopus (154) Google Scholar). Previous data have shown that interaction of the phosphate groups with Arg-175 in the region 166–179 triggers the electrostatic switch in arrestin (33Vishnivetskiy S.A. Schubert C. Climaco G.C. Gurevich Y.V. Velez M.G. Gurevich V.V. J. Biol. Chem. 2000; 275: 41049-41057Abstract Full Text Full Text PDF PubMed Scopus (154) Google Scholar). The sequence of molecular events that follows can be described most closely as the induced fit mechanism of interactions, with both the rhodopsin C terminus and arrestin changing conformations for the most precise fit. In fact, conformational changes and repositioning of the rhodopsin C terminus may help to direct arrestin for the most specific docking with rhodopsin (34Raman D. Osawa S. Gurevich V.V. Weiss E.R. J. Neurochem. 2003; 84: 1040-1050Crossref PubMed Scopus (30) Google Scholar), making possible interactions with other regions of arrestin, such as the 109–130 stretch (35Smith W.C. McDowell J.H. Dugger D.R. Miller R. Arendt A. Popp M.P. Hargrave P.A. Biochemistry. 1999; 38: 2752-2761Crossref PubMed Scopus (21) Google Scholar) or other regions, that become available due to local and global rearrangements in arrestin. Analogies with the Ball-and-Chain Mechanism of Inactivation of Potassium Channels—Involvement of a built-in domain of a transmembrane signaling protein in the mechanism of inactivation struck us as remarkably similar in concept to the ball-and-chain mechanism of inactivation of voltage-gated potassium channels, described originally for the Drosophila Shaker B channel (36Hoshi T. Zagotta W.N. Aldrich R.W. Science. 1990; 250: 533-538Crossref PubMed Scopus (1281) Google Scholar), and then for a variety of other channels (37Solaro C.R. Lingle C.J. Science. 1992; 257: 1694-1698Crossref PubMed Scopus (96) Google Scholar). The inactivation domain, the ball, blocks the channel pore reversibly and is connected to the channel via a flexible linker, the chain. In both cases, inactivation occurs by a cytoplasmic terminal domain of a transmembrane protein, the N-terminal inactivation gate in the case of the channels, and the C-terminal domain in the case of GPCRs. In both instances, the inactivation properties of the domain can be regulated by phosphorylation (38Antz C. Bauer T. Kalbacher H. Frank R. Covarrubias M. Kalbitzer H.R. Ruppersberg J.P. Baukrowitz T. Fakler B. Nat. Struct. Biol. 1999; 6: 146-150Crossref PubMed Scopus (66) Google Scholar), and the domains function fairly independently of the rest of the protein. Their removal proteolytically or by genetic methods leads to a complete loss of inactivation. The inactivation is restored when the domains are added as independently synthesized polypeptides (14Puig J. Arendt A. Tomson F.L. Abdulaeva G. Miller R. Hargrave P.A. McDowell J.H. FEBS Lett. 1995; 362: 185-188Crossref PubMed Scopus (72) Google Scholar, 39Zagotta W.N. Hoshi T. Aldrich R.W. Science. 1990; 250: 568-571Crossref PubMed Scopus (611) Google Scholar). What makes the mechanism of GPCR inactivation distinct is the obligatory requirement for an auxiliary protein, arrestin, which induces the active "ball" conformation of the phosphorylated C terminus, as we show in this study. In fact, arrestin binding appears to enhance the original nucleation effect of the rhodopsin C terminus in a snowball manner, which completely excludes interactions with transducin. Inactivation gates of several potassium channels have been studied by NMR. Although fairly diverse in their three-dimensional structure, the ball-and-chain domains of the voltage-dependent potassium channel RCK4 (Kv1.4) (40Antz C. Geyer M. Fakler B. Schott M.K. Guy H.R. Frank R. Ruppersberg J.P. Kalbitzer H.R. Nature. 1997; 385: 272-275Crossref PubMed Scopus (100) Google Scholar) and the β2-subunit of a large conductance Ca2+ and voltage-dependent potassium channel (41Bentrop D. Beyermann M. Wissmann R. Fakler B. J. Biol. Chem. 2001; 276: 42116-42121Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar) show striking resemblance to the arrestin-bound structures of the rhodopsin C terminus. Despite little primary sequence similarities, and obvious structural variations, the helix-loop hairpin motif can be readily recognized (Fig. 6). There is no reason to suspect an evolutionary link between channels and GPCRs. More likely, the remarkable similarity between the two systems in concept and structural detail represents a fascinating example of evolutionary convergence of the molecular domains evolved to perform similar regulatory functions. Overall principles of rhodopsin inactivation, including multiple phosphorylation of the C-terminal serine and threonine residues by rhodopsin kinase and the binding of arrestin, are understood fairly well conceptually. Mechanistic details of rhodopsin-arrestin interactions at the atomic level, however, are lacking. Previous studies have focused on the mechanisms of arrestin activation and conformational changes in arrestin accompanying its binding to phosphorylated rhodopsin. We show in this study that binding of the fully phosphorylated polypeptide, representing the last 19 residues of bovine rhodopsin, to arrestin, leads to the peptide folding into a helix-loop structure. When modeled in the context of the whole rhodopsin, this compact structure occupies a central position between cytoplasmic loops of rhodopsin. This unique orientation of the rhodopsin C terminus and the bound arrestin shields phosphates in the center of the complex, occludes major sites of transducin binding, and is associated with termination of a signaling state. Because both GPCRs and arrestins are families of proteins with conserved functions, we predict that diverse C-terminal regions of other GPCRs should exhibit structural properties similar to those identified in this study.