Structural Requirements for the Stabilization of Metarhodopsin II by the C Terminus of the a subunit of Transducin*

The retinal receptor rhodopsin undergoes a conformational change upon light excitation to form metarhodopsin II (Meta II), which allows interaction and activation of its cognate G protein, transducin (G t ). A C-terminal 11-amino acid peptide from transducin, G t a -(340–350), has been shown to both bind and stabilize the Meta II conformation, mimicking heterotrimeric G t . Us- ing a combinatorial library we identified analogs of G t a -(340–350) that bound light-activated rhodopsin with high affinity (Martin, E. L., Rens-Domiano, S., Schatz, P. J., and Hamm, H. E. (1996) J. Biol. Chem. 271, 361–366). We have made peptides with key substitutions either on the background of the native G t a -(340–350) sequence or on the high affinity sequences and used the stabilization of Meta II as a tool to determine which amino acids are critical in G protein-rhodopsin interaction. Removal of the positive charge at the N termini by acylation or delocalization of the charge by K to R substitution en-hances the affinity of the G t a -(340–350) peptides for Meta II, whereas a decrease was observed following C-terminal

Heterotrimeric guanine nucleotide-binding proteins (G proteins) 1 are critical regulatory proteins in a variety of cell signaling pathways. Stimulation of a G protein-coupled receptor by an appropriate agonist results in conformational changes leading to its interaction with a heterotrimeric G protein, catalysis of GDP release, and subsequent G protein activation (for reviews see Refs. [1][2][3]. In the visual system, the retinal light receptor rhodopsin is activated by absorption of a single photon of light followed by interaction with and activation of the G protein transducin (G t ), leading to stimulation of 3Ј,5Ј cyclic GMP phosphodiesterase (for review see Ref. 4). In the absence of additional guanine nucleotides, the G protein enhances agonist binding to a G protein-coupled receptor. Binding of either GDP or GTP to the G protein disrupts the high affinity complex or active state of the G protein-coupled receptor (5,6).
The receptor-G protein interface has been defined in some detail and involves portions of the intracellular loops and juxtamembrane regions of the G protein-coupled receptor with several regions on heterotrimeric G proteins (for review see Ref. 7). Distinct regions on G ␣ that are involved in receptor recognition, GTP binding and hydrolysis, guanine nucleotideinduced conformational changes, and effector interaction have been elucidated using diverse studies including disruption by ADP ribosylation (8), binding of antibodies (9 -11), proteolytic mapping (12), alanine scanning (13,14), peptide or minigene studies (15)(16)(17), and studies of chimeric G ␣ proteins (18 -21). Researchers have determined that the N terminus, C terminus, and parts of the ␣5 helix of G␣ are important sites for receptor recognition.
For rhodopsin-transducin interactions, all three intracellular loops of rhodopsin have been implicated (22)(23)(24)(25)(26)(27)(28) as well as the C-terminal tail (29). The receptor appears to interact with both the ␣ (30, 31) and ␤␥ subunit (32)(33)(34) of heterotrimeric G t . On the ␣ subunit of G t , multiple sites of contact have been identified (reviewed in Ref. 35). Three regions, the N-terminal 23 residues, an internal sequence (G t -(311-323)), and the Cterminal 11 amino acids, were identified by peptide competition (31). Using chimeric proteins, the C terminus and residues 299 -314 of G t were shown to contribute to rhodopsin binding (36 -38). Of these sites of interaction, the C terminus of G protein ␣ subunits has been the most extensively investigated. Rhodopsin-G t interaction can be disrupted by a number of treatments that block the C terminus of G t␣ , including pertussis toxin-catalyzed ADP ribosylation (8) and binding by an antibody (39). The C-terminal peptide G t␣ -(340 -350) has been shown to directly bind to and stabilize Meta II (31), mimicking the entire holo-G protein. Furthermore, selective mutagenesis of this C-terminal G t␣ region leads to alterations in G protein function (14,40,41).
Important insights into the mechanism of G protein-mediated signal transduction have been provided by the crystallization and high resolution structure determination of G protein ␣ (42)(43)(44)(45)(46)(47) and ␤␥ (48) subunits as well as heterotrimeric G ␣␤␥ complexes (49,50). However, in many of the crystal structures the final residues of the G ␣ C terminus are disordered and not visible. NMR studies indicate that the C-terminal 11-amino acid peptide of G t␣ has no structure in solution, but it takes on significant structure when it is bound to either excited (light  1 The abbreviations used are: G protein, guanine nucleotide-binding protein; G t , transducin; Meta II, metarhodopsin II; R*, activated receptor; Fmoc, N-(9-fluorenyl)methoxycarbonyl; HPLC, high pressure liquid chromatography; Ac, acetylated; GTP␥S, guanosine 5Ј-O-(thiotriphosphate). activated) or unexcited rhodopsin (51)(52)(53), suggesting a direct physical interaction between the C-terminal residues of G t␣ and rhodopsin. However, detailed structures of the G t␣ peptidereceptor or G t␣ peptide-R* complexes are still uncertain because of methodological limitations. 2 In an alternative approach to defining important determinants in the C terminus, Martin et al. (54) used a combinatorial peptide library to identify high affinity analogs of G t␣ -(340 -350) that bound to light-activated rhodopsin. Sequences derived from panning the biased library demonstrated the presence of certain positions in which amino acids were absolutely conserved (Cys-347 and Gly-348). Based on work by Dratz et al. (51), as well as peptide substitutions, these residues are predicted to be part of a type IIЈ ␤ turn, which is thought to be required for establishing Meta II stabilization. Also highlighted by the combinatorial library screening procedure was that the Lys-341 in the native sequence was selected against. In nearly 70% of the sequences obtained from the panning, and in all of the highest affinity sequences, Lys-341 was changed to a noncharged group (54). It was hypothesized that a positive charge at position 341 in G t might decrease its affinity for rhodopsin, leading to a faster "off" rate and a higher rate of G protein activation. Given that other G proteins do not have a positive charge at this position, the presence of a charged residue in G t has implications for the evolved properties of the signaling mechanisms.
In this study we have examined the structural basis for the invariance of Gly and Cys at the Ϫ3 and Ϫ4 positions, respectively, and the loss of a positively charged residue at the Ϫ10 position in the high affinity analogs from the combinatorial library. The structural features predicted by NMR were also further explored with analog peptide studies of the functional interaction of these peptides with activated rhodopsin leading to Meta II stabilization.

MATERIALS AND METHODS
Peptide Synthesis-G t␣ -(340 -350) analogs were synthesized by the solid-phase Merrifield method using Fmoc chemistry on Milligen 9050 or Applied Biosystems peptide synthesizers. Peptides were purified by reverse-phase HPLC using C4 or C18 preparative columns and an acetonitrile, 6 mM HCl, water gradient. The purified peptides were subjected to fast-atom bombardment or electrospray mass spectrometry to determine authenticity and analytical reverse-phase HPLC to determine purity. Fmoc-2-amino butyric acid was purchased from Advanced ChemTech (Louisville, KY).
Protein Preparation-Washed bovine rod outer segments-containing rhodopsin were prepared from fresh bovine retinas using sucrose gradient centrifugation and washing of the membranes with EDTA to remove peripherally bound proteins (55). Heterotrimeric G t was prepared as described previously by Stryer et al. (56).
"Extra" Meta II Assay-The absorbance spectra of washed rod outer segments (5 M) were measured in the presence of varying concentrations of G t␣ -(340 -350) analogs or heterotrimeric G t using an SLM Aminco DW2000 spectrophotometer as described previously (51). Samples were mixed in Meta II buffer (50 mM HEPES, pH 8.2, 100 mM NaC1, 1 mM dithiothreitol, 1 mM MgC1 2 ). The samples were maintained at 5.4°C using a water-jacketed cuvette holder and refrigerated circulator. After a dark-adapted spectrum was measured, a flash of light bleaching 10% of the rhodopsin was presented, and after a 1-min incubation a second spectrum was measured. The difference between the two spectra was then calculated. Extra Meta II was calculated as the difference between the absorbance at 390 nm and that at 440 nm. Measurements were done in duplicate for each individual experiment and calculated as a percentage of the extra Meta II produced by 2 M heterotrimeric G protein measured on the same day. The average Ϯ the standard error of the mean were calculated using GraphPad PRISM (Version 3.0). Dose-response curves for the G t␣ -(340 -350) peptide analogs or heterotrimeric G t were generated by nonlinear regression using the following sigmoidal dose-response (variable slope) equation, also known as a four-parameter logistic equation, to obtain the EC 50 values and slope (Hillslope).
Bottom is the Y value at the bottom plateau, which was set to 0, and Top is the Y value at the top plateau. LogEC 50 is the logarithm of the EC 50 , the concentration that gives a response halfway between Bottom and Top. The variable Hillslope controls the slope of the curve such that (i) when Hillslope equals 1, the equation generates a standard dose-response curve, (ii) when Hillslope is less than 1.0, the curve is more shallow, and (iii) when Hillslope is more than 1.0, the curve is steeper. Decay Assay-The absorbance spectra of washed rod outer segments (10 M) were measured in an SLM Aminco DW2000 spectrophotometer in decay buffer (10 mM K 2 PO 4 , pH 6.5, 0.1 M KCl, 0.1 mM EDTA, 1 mM dithiothreitol) in the presence of either 1 or 2 mM G t␣ -(340 -350) analog peptides. The spectra were measured in the dark, then completely bleached in room light. Spectra for the bleached samples were measured at 30-min intervals over a 6-h time period. Finally, 260 mM HCl was added to protonate the retinal Schiff base in rhodopsin and leave free retinal unprotonated.
Molecular Dynamics-Calculations were carried out on Silicon Graphics workstations using the program DISCOVER (MSI, Inc.) and the CVFF forcefield. After amino acid substitutions the energy was minimized using steepest descents, and then 250 ps of molecular dynamics were run at 300 K, pH 7 with a distance-dependent dielectric constant.

RESULTS AND DISCUSSION
The three-dimensional structures of heterotrimeric G proteins have provided detailed information about their subunit and receptor interactions and have suggested mechanisms for GTP hydrolysis. In most of the crystal structures the final residues of the G ␣ C terminus are disordered and not visible. However, in one of the crystal structures of G t␣ ⅐GTP␥S (45) the ordered C-terminal residues 343-349 form an ␣-helix and make van der Waals contacts with residues of the ␣2/␤4 loop. It is not clear whether this is a crystal-packing artifact or an indication of a preferred conformation of the C terminus. Additionally, for both G s␣ (47) and the G i␣ ⅐RGS4 complex (57), the C terminus of G ␣ is an extension of the ␣5 helix. In the G i␣ ⅐RGS4 complex (57) the extended ␣5 helix was stabilized by the N-terminal helix as well as by crystal contacts, whereas for G s␣ (47) the extended ␣5 helix is in close proximity with the ␣4-␤6 loop. Sunahara et al. (47) suggest that the divergence of the G ␣ C-terminal structures may contribute to receptor selectivity.
NMR studies indicate that the C-terminal 11-amino acid peptide of G t␣ has little or no structure in solution, but it takes on significant structure when it is bound to rhodopsin (51-53), indicating a direct physical interaction between the C terminus of G t␣ and its receptor. However, detailed structures of the G t␣ peptide-receptor or G t␣ peptide-R* complexes are still uncertain because of methodological limitations. 2 It has not been possible to refine the published bound peptide structures (51,52) by comparing experimental NMR data to that calculated from the bound structures, indicating that there are significant errors in the NMR distance constraints. In addition, many of the predictions for amino acid substitutions that would be tolerated or favorable for the Kisselev et al. (53) proposed R*-bound structure were not born out by experiment (58). When more accurate NMR data are obtained, it may be possible to refine the protein-bound peptide structures to within the uncertainty in 1.6-Å resolution x-ray structures. 2 For an alternative approach, we identified potent high affinity sequences related to the C terminus of G t␣ using a "peptides-on-plasmids" combinatorial technique (59) in which a library with greater than 10 9 different peptide sequences was tested for binding to light-activated rhodopsin (54). We have now expanded on observations from the combinatorial screening by making specific point mutations within the C-terminal peptide sequence to clarify how the C terminus of G t␣ interacts with and stabilizes the activated rhodopsin species Meta II.
Rhodopsin can be measured spectrophotometrically in many of its light-induced conformational states (for review see Ref. 4). The binding of G t to light-activated rhodopsin stabilizes an active signaling state of the receptor (R*) that can be measured spectrophotometrically (60,61). The active state can be stabilized by the R*-catalyzed loss of GDP from G t , leading to an empty guanine nucleotide binding pocket (62). Addition of either GTP or GDP promotes the loss of the active R* state, as measured by the loss of Meta II stabilization (62)(63)(64). The biologically active Meta II state can be differentiated from its precursor, Meta I, by the extra Meta II assay. This assay makes use of the observation that under conditions of slightly alkaline pH and low temperatures Meta I is strongly favored in the absence of G t . Meta II is stabilized in the presence of G t and can be measured spectrophotometrically. We have exploited the ability of G t␣ -(340 -350) C-terminal peptide analogs to stabilize Meta II in the same manner as G t to investigate the interface between G proteins and their agonist-activated receptors.
The Role of Free Amino Groups in the Peptide Stabilization of Meta II-The native G t␣ -(340 -350) peptide is of relatively low potency in its ability to interact with and stabilize Meta II, with an EC 50 of 1209 M (Fig. 1, Table I). Two similar peptide analogs were found to have increased potencies at stabilizing Meta II; G t␣ -(340 -350)-K341R, with one fewer amino group, displayed an EC 50 of 180 M, and Ac-G t␣ -(340 -350)-K341R, with two fewer amino groups, had an EC 50 of 163 M. The potencies of G t␣ -(340 -350)-K341R and Ac-G t␣ -(340 -350)-K341R are 6.7-and 7.4-fold higher, respectively, than that of native G t␣ -340 -350 peptide ( Table I) Table I).
We have shown previously that the increase in potency of G t␣ -(340 -350)-K341R as compared with the native G t␣ -(340 -350) peptide is associated with a slower rate of Meta II decay in the presence of a Lys to Arg change at position 341 (51). It was hypothesized that the N terminus of the native peptide can attack the retinal Schiff's base linkage of Meta II nonspecifically (51). Thus, we compared the rate of Meta II decay in the presence of various peptide analogs to ascertain whether the increase in potency observed might be associated with changes in the rate of Meta II decay. The half-life for Meta II in the presence of 2 mM Ac-G t␣ -(340 -350)-K341R was 654 min versus a half-life of 702 min in the presence of 2 mM Ac-G t␣ -(340 -350)-K341R-K345R, indicating that there was little difference in their rate of Meta II decay over a 6-h period (data not shown). This is different from the results obtained for unacetylated G t␣ -(340 -350)-K341R and G t␣ -(340 -350) peptides, which attack the retinal Schiff's base and result in a much shorter half-life for Meta II when compared with Ac-G t␣ -(340 -350)-K341R (51). Thus, the increased potency for the peptide in which both Lys residues have been changed to Arg (Ac-G t␣ -(340 -350)-K341R-K345R) appears to be due to an increase in affinity for Meta II rather than an increased stability of Meta II.
We have also investigated the role of the free N terminus on activity using two of the high affinity analogs, peptide 23 (VLEDLKSCGLF) and peptide 24 (MLKNLKDCGMF), identified by combinatorial screening (54). Peptides 23 and 24 contain free N-terminal amino groups. Acylation of the N terminus of peptide 23 to neutralize the free N-terminal amino group decreases the EC 50 only slightly, whereas acylation of peptide 24 decreases its EC 50 16-fold (Fig. 2, Table I Table I). Taken together, the data in Table I suggest that removal of a localized positive charge by acylation of the N terminus or delocalization of the positive charge that occurs when Lys is replaced with Arg can enhance the affinity of certain G t␣ -(340 -350) peptide analogs for their interaction with Meta II.
The Role of the Negatively Charged C Terminus in the Peptide Stabilization of Meta II-NMR studies of the G t␣ peptide analogs bound to rhodopsin indicate substantial structural changes upon light excitation, suggesting that the C-terminal carboxyl group shifts its orientation upon interaction with the activated receptor (51)(52)(53). We tested the importance of the free carboxyl group of G t␣ in maintaining Meta II stabilization by amidating the Ac-G t␣ -(340 -350)-K341R peptide to neutralize the C-terminal negative charge. This resulted in a peptide (Ac-G t␣ -(340 -350)-K341R-Amide) that stabilized Meta II but was 2-fold less potent than the same peptide with a free C terminus (Fig. 1, Table I). Screening of the peptide library showed that the C-terminal final seven amino acids were the most invariant, with the fourth round clones having identities at these positions ranging from 72 to 100% (54). These are the same residues that are disordered in most of the crystal structures. It is reasonable to speculate that the C terminus of G t␣ directly contacts the activated receptor, and our data supports the idea that the negative charge at the C terminus may participate in this interaction.
Roles of Cys-347 and Gly-348 in the Peptide Stabilization of Meta II-The Cys residue at position 347 of G t␣ has long been known to play a critical role in the interaction between G t␣ and rhodopsin. It is the site of pertussis toxin-catalyzed ADP ribosylation, which leads to an uncoupling of G ␣i family G proteins from their cognate receptors (8). That Cys-347 and Gly-348 are critical residues in rhodopsin binding was apparent from the combinatorial library study (54), in which all of the Meta IIbinding peptide analogs sequenced from the fourth round were conserved at these positions. This phenomenon suggests that other amino acids could not substitute effectively for the Cys- To identify the role of the Cys side chain in its interaction with Meta II, we initially synthesized a Cys to Ala analog, Ac-G t␣ -(340 -350)-K341R-C347A. This analog showed a substantial decrease in its ability to stabilize Meta II. Meta II could be stabilized maximally to 77% of control by 1.8 mM Ac-G t␣ -(340 -350)-K341R-C347A peptide with an EC 50 of 579 M (Fig. 3, Table II). This loss of Meta II stabilization is similar to results obtained by Osawa and Weiss (41), who showed that the same mutation in the whole G ␣ resulted in a loss of rhodopsin binding to recombinant mutant G t . However, Garcia et al. (40) have reported that in their recombinant assay system, this same mutation had no deleterious effect on rhodopsin binding and rhodopsin-stimulated G protein activation.
To further test the role of the side chain of Cys-347, we synthesized a Cys to Ser analog, Ac-G t␣ -(340 -350)-K341R-C347S, because Ser has a size similar to the sulfhydryl moiety. To ascertain whether the anionic form of the sulfhydryl moiety is important in the binding pocket environment, a Cys to Asp analog, Ac-Gt␣-(340 -350)-K341R-C347D, was synthesized. At 1.8 mM the Ac-G t␣ -(340 -350)-K341R-C347S peptide stabilized only 40% of Meta II, and the Ac-G t␣ -(340 -350)-K341R-C347D peptide stabilized only 25% of Meta II, as compared with the control Ac-G t␣ -(340 -350)-K341R peptide (Fig. 3). The EC 50 values obtained for both peptides (Ac-G t␣ -(340 -350)-K341R-C347S and Ac-G t␣ -(340 -350)-K341R-C347D) were greater than 1 mM (Table II). To determine whether it is the hydrophobicity of Cys that is important, we substituted Cys-347 with 2-aminobutyric acid (Abu), a compound that replaces the sulfhydryl group of the cysteine with a methyl group and approximately mimics cysteine in both size and hydrophobicity. This -SH to -CH 3 peptide (Ac-G t␣ -(340 -350)-K341R-C347Abu) was able to stabilize Meta II with essentially the same potency as the Ac-G t␣ -(340 -350)-K341R parent peptide, with an EC 50 of 127 M versus 163 M (Fig. 3, Table II). Therefore, the data suggest that the Cys-347 side chain of G t␣ interacts in a hydrophobic manner with Meta II.
To test the requirement for a Gly residue at position 348, we substituted this position with an Ala. A structural feature of the NMR-based model of the rhodopsin-bound peptide is the presence of a ␤-turn between Cys-347 and Phe-350 with Gly-  -(340 -350) analogs The dose response for the ability of each of the listed G t␣ analogs to stabilize Meta II was measured as described under "Materials and Methods." Bold letters indicate amino acid residues that differ relative to the native G t␣ -(340 -350) sequence. Dose-response curves were analyzed by non-linear regression using a sigmoidal dose-response variable slope equation (GraphPad Prism) to obtain the EC 50 values, S.E., and slope. The number of independent experiments, done in duplicate, is listed as n. The EC 50 obtained for heterotrimeric G protein was 0.28 Ϯ 1.12 M, with a slope of 1.19 (n ϭ 8).

Peptide
Sequence EC 50  348 in the n ϩ 1 position in the dark-bound conformation (51) and the light-bound conformation (52,53). Consistent with this observation, peptides in which Gly was substituted with L-Leu, which would be predicted to break the ␤-turn, lost functional activity (51). Meanwhile, D-Ala, predicted to maintain the peptide structure, was almost as potent as the parent peptide (51). Using molecular dynamics calculations to estimate the relative energy to form the ␤Ϫturn, with Gly being taken as 0 kcal mol Ϫ1 , we found that the energy required would be 0 kcal mol Ϫ1 for a Gly at position 348, 3.2 kcal mol Ϫ1 for D-Ala, 12 kcal mol Ϫ1 for L-Ala, and 88.2 kcal mol Ϫ1 for L-Leu. We therefore made a synthetic peptide with a Gly to Ala substitution (Ac-G t␣ -(340 -350)-K341R-G348A). The L-Ala substitution at this position is much milder than that of L-Leu, and it was of interest to see whether this analog could stabilize Meta II. As predicted by the energy calculation, the peptide with the Gly to Ala substitution at position 348 was only minimally capable of stabilizing Meta II, with an EC 50 value of nearly 1 mM and maximal stabilization of 15% at 1.8 mM (Fig. 4, Table II). Thus, the Gly at position 348 appears to be essential for the ability of the C-terminal peptide to effectively stabilize Meta II. This result is consistent with the presence of a type IIЈ ␤-turn at the C terminus of the peptide. Alternatively, the binding pocket on rhodopsin may not be able to tolerate bulky side chains at position 348. This alternative interpretation is not as likely, because substitution of the Gly at this position with D-Ala is equipotent at stabilizing Meta II.
The Role of Lys-341 in the Peptide Stabilization of Meta II-The six highest affinity peptide analogs from the combinatorial library had all replaced the positive charge at position 341 with a hydrophobic residue (54). To further test the idea that a hydrophobic residue at position 341 increases the affinity for Meta II, we substituted a Leu for Lys at position 341 in the Ac-G t␣ -(340 -350)-K341R-C347Abu peptide and tested its ability to stabilize Meta II (Fig. 5). We found that these changes substantially enhanced the EC 50 such that the EC 50 of Ac-G t␣ -(340 -350)-K341R-C347Abu is 127 M, whereas that of Ac-G t␣ -(340 -350)-K341L-C347Abu is 44 M. When peptides were tested with the K341L substitution that had also been substituted with a hydrophic residue at position 347 (Met or Val), we found EC 50 values of 19.6 M for Ac-G t␣ -(340 -350)-K341L-C347M and 43.6 M for Ac-G t␣ -(340 -350)-K341L-C347V (Table  III). These values were 4 -8-fold more potent than the native acetylated peptide, indicating the importance of a hydrophobic residue at position 341.
Similarly, to characterize the hydrophobic site at Cys-347 in the context of the high affinity Ac-peptide 23 peptide sequence, which also has a Leu at position 341, we introduced substitutions (Ac-peptide 23-C347Abu, Ac-peptide 23-C347M, Ac-peptide 23-C347S). Table III -(340 -350) analogs The dose response for the ability of each of the listed G t␣ analogs to stabilize Meta II was measured as described under "Materials and Methods." Bold letters indicate amino acid residues that differ relative to the native G t␣ -(340 -350) sequence. The ⌬ symbol represents 2-aminobutyric acid (Abu). Dose-response curves were analyzed by non-linear regression using a sigmoidal dose-response variable slope equation (GraphPad Prism) to obtain the EC 50 values, S.E., and slope. The number of independent experiments, done in duplicate, is listed as n.
can be substituted with any of those residues, confirming that it interacts in a hydrophobic manner with the receptor. However, the changes did cause a decrease in the potency (11.0 M EC 50 for Ac-peptide 23-C347M, 2.0 M EC 50 for Ac-peptide 23-C347V, and 10.6 M EC 50 for Ac-peptide 23-C347Abu compared with 0.5 M EC 50 for Ac-peptide 23). Therefore, even though these peptides were all more potent than the native peptide at binding Meta II, changes at the critical 347 position diminished the affinity of the Ac-peptide 23 analogs for the receptor. Thus, the detailed fit of many residues within the peptide binding site on rhodopsin appear to be critical for stabilization of Meta II.
Our results corroborate that the Lys at position 341 decreases the affinity of native G t␣ for rhodopsin. One might wonder why such a disruptive residue at the site of receptor interaction would be present. Perhaps it is important for the G protein to have a lower affinity for the receptor so that the receptor can be a more efficient catalyst and thus more rapidly catalyze activation of more G proteins following ligand activation. We are currently testing this idea in full-length G ␣ by making a K341L point mutation. We predict that a G protein with this single mutation will have a higher affinity for rhodopsin.
The Role of Random Linker in the Peptide Stabilization of Meta II-To examine the effect of the random 4-amino acid linker present on the high affinity peptide analogs identified in the combinatorial screening, we synthesized and tested the full-length 15-mer peptides for clones 23 and 24 (peptides 23 and 24). The 15-mer of peptide 24 long was found to be 15-fold more potent than the corresponding 11-mer peptide (Fig. 2, Table IV), whereas the difference is only 1.5-fold for the 15-mer and 11-mer of peptide 23 (Table IV). Thus, although no obvious consensus sequences were present in the 4-amino acid linker region of the sequenced clones (54), the region may contribute to the affinity for light-activated rhodopsin in some cases.
The results presented here confirm that the C terminus of G ␣ is critical in stabilizing the active conformation of the receptor. Understanding the changes that can be tolerated in this region is essential to furthering our knowledge of how G proteincoupled receptors interact with their cognate G proteins. Previous work from our laboratory indicates that C-terminal G ␣ peptides can compete with G protein for binding receptor and therefore potently block signal transduction (16). By screening a receptor of interest, it should be possible to select for peptide sequences with high affinity for the receptor that will competitively inhibit receptor-G protein interaction and potently block signal transduction through that receptor. Peptide sequences so obtained can then be expressed in cells as minigenes (17,65), providing a facile approach for studying G protein signaling.  -(340 -350) analogs The dose response for the ability of each of the listed peptide analogs to stabilize Meta II was measured as described under "Materials and Methods." Bold letters indicate amino acid residues that differ relative to the native G t␣ -(340 -350) sequence. The ⌬ symbol represents 2-aminobutyric acid (Abu). Dose-response curves were analyzed by non-linear regression using a sigmoidal dose-response variable slope equation (GraphPad Prism) to obtain the EC 50 values, S.E., and slope. The number of independent experiments, done in duplicate, is listed as n.

Peptide
Sequence EC 50 -(340 -350) analogs The dose response for the ability of each of the listed peptide analogs to stabilize Meta II was measured as described under "Materials and Methods." Bold letters indicate amino acid residues that differ relative to the native G t␣ -(340 -350) sequence. Dose-response curves were analyzed by non-linear regression using a sigmoidal dose-response variable slope equation (GraphPad Prism) to obtain the EC 50 values, S.E., and slope. The number of independent experiments, done in duplicate, is listed as n.