Three cytoplasmic loops of rhodopsin interact with ... - Europe PMC

6 downloads 0 Views 1MB Size Report
Jun 8, 1989 - [MIIG] = [Ro][Go]/{[Go]- Kd( + KlK2[Po]2)}. [2]. In the presence of two peptides (P1 and P2), the weak interaction with the other binding sites can ...
Proc. Nati. Acad. Sci. USA Vol. 86, pp. 6878-6882, September 1989 Biochemistry

Three cytoplasmic loops of rhodopsin interact with transducin (vision/receptor/guanine nucleotide-binding protein/signal transduction/competing peptide)

B. KONIG*, A. ARENDTt, J. H. MCDOWELLt, M. KAHLERT*, P. A.

HARGRAVEtt, AND K. P. HOFMANN*

*Institut fur Biophysik und Strahlenbiologie der Universitat Freiburg, Albertstrasse 23, D-7800 Freiburg, Federal Republic of Germany; and tDepartment of Ophthalmology and tDepartment of Biochemistry and Molecular Biology, University of Florida, Gainesville, FL 32610

Communicated by L. M. Beidler, June 8, 1989 (received for review February 8, 1989)

binding of Gt to photolyzed rhodopsin therefore abolish the extra MII. The extent of reduction of the extra MII by a competitor measures the effectiveness of the competitor in disrupting the MII-Gt interaction. This assay represents a method of high sensitivity and high specificity to probe the nature of the interaction between photolyzed rhodopsin and Gt. We previously showed (9) that selected peptides from the sequence of the Gt a subunit (Gta) can interfere with binding of Gt to photolyzed rhodopsin, thus allowing assignment of these peptides to the region of the Gta sequence that binds to rhodopsin. In the work described here, we tested peptides from the rhodopsin sequence in order to see which ones reduce the level of extra MII. Such peptides presumably would do so by simulating a region of rhodopsin's surface that interacts with Gt, thus interfering with Gt binding to MII.

Rhodopsin is a member of an ancient class of ABSTRACT receptors that transduce signals through their interaction with guanine nucleotide-binding proteins (G proteins). We have mapped the sites of interaction of rhodopsin with its G protein, which by analogy suggests how other members of this class of receptors may interact with their G proteins. Three regions of rhodopsin's cytoplasmic surface interact with the rod cell G protein transducin (Ga. These are (i) the second cytoplasmic loop, which connects rhodopsin helices m and IV, (U) the third cytoplasmic loop, which connects rhodopsin helices V and VI, and (iu) a putative fourth cytoplasmic loop formed by amino acids 310-321, as the carboxyl-terminal sequence emerges from helix VII and anchors to the lipid bilayer via palmitoylcysteines 322 and 323. Evidence for these regions of interaction of rhodopsin and Gt comes from the ability of synthetic peptides comprising these regions to compete with metarhodopsin II for binding to Gt. A spectroscopic assay that measures the "extra MIU" caused by Gt binding was used to measure the extent of binding of Gt in the presence of competing peptides. The three peptides corresponding to the second, third, and fourth cytoplasmic loops competed effectively with metarhodopsin H, exhibiting Kd values in the 2 jtM range; 11 additional peptides comprising all remaining surface regions of rhodopsin failed to compete even at 200 ,M. Any two peptides that were effective competitors showed a synergistic effect, having 15 times higher effectiveness when mixed than when assayed separately. A mathematical model was developed to describe this behavior.

MATERIALS AND METHODS Spectrophotometric Assay. Binding of Gt to MII was measured as in refs. 7-9. The assay was performed at pH 8 and 4°C, conditions under which only a small, control amount of MII is formed in the absence of Gt. The full extra MII signal in the presence of Gt corresponds to a 60% MII fraction of the total photoexcited rhodopsin. The final levels of MII formation minus the control level (no Gt present) are a direct measure of the rhodopsin-Gt complexes formed. When normalized to the undisturbed full extra MII signal, they yield the relative amount of Gt that is able to interact with rhodopsin. Washed disk membranes were prepared from bovine rod photoreceptors as described (8) and were suspended in buffer A (100 mM NaCl/2 mM MgCl2/2 mM CaC12/0.2 mM EDTA/ 1 mM dithiothreitol/40 mM Hepes, pH 8.0) in a cuvette at 4°C; all measurements were made at a final rhodopsin concentration of 1.5 ,M. Purified (10) Gt (0.75 AM final concentration) and peptides (see below), also in buffer A, were added to the disk membranes. Most measurements were performed with purified Gt; use of low-ionic-strength extract gave identical results. In one case (peptide CIV), dithiothreitol was replaced by 2 mM ascorbate. Control experiments with other peptides using ascorbate instead of dithiothreitol gave the same results. Peptide LII was dissolved in a minimum quantity of ethanol because of its limited solubility in buffer. Peptide Synthesis. Peptides were synthesized by the solidphase Merrifield method with an Applied Protein Technologies synthesizer (Cambridge, MA) and were purified by high-performance liquid chromatography (11). Peptides from rhodopsin's cytoplasmic (C) surface or lumenal (L) surface in the disk membrane (see Fig. 4) had the following amino acid sequences: CI, residues 61-75; CII, 141-153; CIII, 230-252; CIV, 310-321; CV, 323-334; CVI, 327-338; CVII, 337-348; LI, 13-23; LII, 96-115; LIII, 188-203; LIV, 276-286. An analog of peptide CII was made in which the amino acids

Rhodopsin is the best-studied receptor protein of that class of signal-transducing receptors that act via guanine nucleotidebinding proteins, or G proteins. Other members of this class include the adrenergic receptors (1), the muscarinic acetylcholine receptors (2), the substance K receptor (3), and dozens of other receptors including those for neurotransmitters, peptide hormones, and other regulatory factors (4). In the retinal rod cell rhodopsin is excited by light and undergoes a change in conformation that allows it to activate the G protein transducin (Gt), which in turn activates a cGMP phosphodiesterase (reviewed by Stryer, ref. 5). In this report we identify those sites on the surface of photoexcited rhodopsin that are involved in the interaction with Gt. Similar surface sites on homologous receptors may serve to bind and to activate their G proteins. When rhodopsin is stimulated by light it relaxes within milliseconds to an equilibrium between two tautomeric forms, metarhodopsin I and II (MI and MII) (reviewed in ref. 6). MII interacts strongly with Gt, shifting the equilibrium to form "extra MII" (6-8). The extra amount of MII is a stoichiometric measure of the MII-Gt complexes formed and can be easily measured spectroscopically following a flash of light due to the large difference between the absorption maxima of MI and MII (8). Substances that compete for The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Abbreviations: G protein, guanine nucleotide-binding protein; Gt, transducin; MI and MII, metarhodopsins I and II. 6878

Biochemistry: K6nig et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

141-152 were "scrambled," or assembled randomly (sequence Phe-Ser-His-Gly-Arg-Met-Asn-Pro-Glu-Lys-AsnPhe).

RESULTS Measurements of Extra MII Show How Much Gt Is Bound to MIR. When rhodopsin is photolyzed, an amount of MII is formed that is characteristic of the particular reaction conditions (Fig. 1A, trace i). Additional MII is formed when Gt is present, due to its binding to MIT and shifting the MI/MII equilibrium (Fig. lA, trace ii). We found that certain peptides from the rhodopsin sequence were able to inhibit this effect of Gt and reduce the amount of extra MIT. Rhodopsin peptide CIV, which comprises rhodopsin's cytoplasmic sequence 310-321, reduced the amount of extra MII when present at 15 and 75 ILM (Fig. 1B, traces i and ii). When present at 300 ILM, peptide CIV completely abolished the extra MII so that the amount of MII formed was the same as in the absence of Gt (Fig. 1B, trace iii). This effect is specific for only certain

A

Mll

(ii)

Extra MlI

B (i) (ii) (iii)

~ ~ ~ ~ .Im C

0

x

2 sec

FIG. 1. Flash-induced formation of MIT in suspensions of bovine rod photoreceptor disk membranes in the presence of various concentrations of synthetic peptides from the rhodopsin sequence. Isolated washed disk membranes were combined with purified Gt (0.75 ;LM) and the respective amount of synthetic peptide. After a 4-min incubation, a flash of light (532 nm) bleaching 5% of rhodopsin was presented. Signals are the absorbance difference at 380 nm (A.,, of MIT) and 417 nm (isosbestic point of MI/MII). Total rhodopsin concentration was 1.5 ,uM; photolyzed rhodopsin per flash, 0.075 .M; concentration of Gt, 0.75 AuM; effective path length, 7 mm; temperature, 40C; pH 8.0. All signals are the average of two recordings. (A) Trace i, rhodopsin photolyzed in the absence of Gt, producing MIT. Trace ii, when Gt is present it binds to MII and shifts the equilibrium to produce an amount of extra MIT that is the final level minus the control level. (B) Effect of peptide CIV concentration on MIT formation. Trace i, 15 EM; trace ii, 75 AuM; trace iii, 300 JIM. (C) Photolysis of rhodopsin in the presence of Gt and 200 AM peptide CI shows that MIT is formed with the same kinetics and in the same amount as in the absence of peptide CI.

6879

rhodopsin peptides; e.g., when peptide CI, which connects rhodopsin helices I and II, was added to the rhodopsin/Gt mixture, the amount of extra MII formed on photolysis was not affected. Note that the signals obtained for the high concentration of effective peptide are kinetically identical to those of the control. This indicates that in the presence of peptide, the photochemical reaction proceeds normally to the MI/MII equilibrium. We conclude that the peptide prevents interaction of Gt with MIT but does not affect formation of MIT itself. Peptides from Three Regions of Rhodopsin's Cytoplasmic Surface Are Effective Competitors. Fig. 2 shows the dependence of extra MII formation on peptide concentration for 11 peptides that comprise the aqueous-exposed surface of rhodopsin. Three of the peptides (CII, CIII, and CIV) were effective in preventing G, binding to rhodopsin (Fig. 2). These peptides are from the cytoplasmic surface ofrhodopsin. They form loops connecting helices III and IV (CII), helices V and VI (CIII), and helix VII and its site of anchoring in the disk membrane (CIV, palmitic acids esterified to Cys-322 and Cys-323; ref. 13) (see Fig. 4). By contrast, peptides representing other regions of the cytoplasmic surface were ineffective in this assay: e.g., the loop connecting helices I and II (CT) and the carboxyl-terminal region 323-348 (CV, CVI, and CVII; Fig. 2). As expected, none of the peptides representing regions of rhodopsin's surface exposed to the disk lumen were able to perturb the binding of Gt to MII (Fig. 2). Peptide CII with a "scrambled" sequence was completely ineffective; the same was found for a peptide comprising only the amino-terminal part (residues 231-241) of the CIII sequence (data not shown). Modification of Cys-316 in peptide CIV had no effect on the competitive behavior of that peptide, in agreement with the lack of effect of thorough SH modification on rhodopsin-G, interaction (14). Data points for the cases in which peptides competed for Gt binding to photoactivated rhodopsin (competition curves) are best fit by a hyperbolic relation with a Hill coefficient of =2 (values are 1.83, 2.11, and 2.40 for CII, CIII, and CIV, respectively). A linear hyperbola (exponent 1) fails to fit the data for any of the peptides. The Competition Is Synergistic. We examined the effect of two and three different peptides when applied simultaneously. If two different peptides were acting independently, the competition curve (Fig. 3) would be shifted by a factor of =2 to the left, as a result of the 2 times higher total peptide concentration. The experiment, however, resulted in a much larger shift. For CII and CIII applied in equimolar concentration, the measured curve was shifted by a factor of 30, or 15 times more than on a linear superposition basis (Fig. 3). Addition of a third peptide (CIV) caused only the small effect expected from the higher peptide concentration. The pronounced shift was found for any combination of two of the effective peptides (data not shown).

DISCUSSION Identification of the Gt Interactive Sites on Rhodopsin. Our

results clearly demonstrate which of the sites on rhodopsin's surface are involved in G, interaction as well as which ones are not involved. Fig. 4 shows a modified model of the rhodopsin structure that displays the loops that are involved in rhodopsin-Gt interaction. All of the surface regions of rhodopsin or related receptors have been previously suggested as sites of interaction with G proteins on various grounds. A preponderance of evidence has supported involvement of the third cytoplasmic loop, CIII, based on results of limited proteolysis (18) and sitedirected mutagenesis (19) of rhodopsin and of the /3adrenergic receptor (20). However, considerable variability in size and sequence of this loop among different receptor

6880

Biochemistry: K6nig et al.

Proc. Natl. Acad. Sci. USA 86 (1989)

Anf

4U I

I

i

61-75

20-

s

* i

a

o = 323-334 (CV) o = 327-338 (CVI) A = 337-348 (CVII)

230-252 (CIII)

(CI)

10 -

p

Co

A n 19111

C

0-

0 .0

U

I

I

30

40

co

111

1.

1

extra MIl

30-

141-153 (CII)

2010-

1,

1

1

.AA I

0

11 I1

.

~

.

.

0. -;

x

x

0

x

o

= =

A

=

13-23 (LI) 96-115 (LII) 188-203 (Lil)

x

=

276-286 (LIV)

o

310-321 (CIV) X

h

m

2

A

A

X6

A

Mil i

10-2

1111-

lo-1

10-2

10-1

10-2

10-1

Pepude, mM

FIG. 2. Competition between synthetic rhodopsin peptides and photoactivated rhodopsin for binding to Gt. As described in Fig. 1, a control amount of MII is formed in the absence of Gt (MII, lower solid line in each panel). In the presence of Gt, stable binding of Gt to photoactivated rhodopsin causes stabilization of the MIT form that is observed as an enhanced level of 380-nm-absorbing species (extra MII, upper solid line). Data points [mean SEM, n = 4 or 8 (for CII-CIV)] represent the measured A380 as a function of peptide concentration in the presence (filled symbols) or absence (open symbols) of Gt. The measured level of A30 minus the control level is a linear measure of the amount of rhodopsin-Gt complexes (12). Solid lines are hyperbolic fits to the data with Hill coefficients of 1.83, 2.11, and 2.46 for CII, CIII, and CIV, respectively. ±

types and experiments with anti-CIII antibodies (21) have

raised questions about this conclusion. The first cytoplasmic loop, CI, is well conserved in sequence among rhodopsins from Drosophila to man and has been suggested for a well conserved function such as Gt binding for that reason (22). The second cytoplasmic loop, CII, has only recently been implicated in G-protein binding, by competition of antibody and Gt for rhodopsin binding (21) and by site-specific mutagenesis of the ,B-adrenergic receptor (23). As for the carboxyl terminus, removal of 12-24 amino acids by proteolysis did not diminish the ability of rhodopsin to bind and activate G, (18, 24). This argued against a role for this region in G. interaction; however, such a role has been suggested based on antibody binding (21) and peptide competition (25) experiments.

The rhodopsin sequence 310-323 has only recently been recognized to constitute a fourth cytoplasmic loop due to its proposed anchoring in the membrane by palmitic acid esterified to Cys-322 and Cys-323 (13). It is accessible to antibody binding in membrane-bound rhodopsin and is structurally related among different receptors; antibody/peptide competition experiments have shown relatedness of this region of rhodopsin, the M1 muscarinic receptor, and the opiate receptor (D. Newton, G. Adamus, P.A.H. and W. Klee, unpublished data). Its role in Gt binding has been suggested by peptide competition experiments (25) and site-specific mutagenesis of the 8-adrenergic receptor (23) and is firmly supported by our study.

40-

c

301

._

:._

20 -CII+CIII

CH

E-

10-CII+CIII+CIV 0-

0.3

111111 1

1 111111 10 Peptide, pgM

111111

100

FIG. 3. Synergistic competition of two or three different synthetic rhodopsin peptides. The peptides were applied in equimolar concentration (given on the abscissa); the control levels of MII formation (upper and lower solid lines) and of the data points (mean + SEM, n 4) are as in Fig. 2. Solid lines are hyperbolic fits as in Fig. 2. The resulting Hill coefficients are 2.0 for peptide CII alone (compare Fig. 2), 2.7 for CII plus CIII, and 3.0 for CII plus CIII plus CIV. =

FIG. 4. Topographic model of rhodopsin indicating regions CII, CIII, and CIV that interact with Gt. Filled circles represent sequences of peptides that interact with Gt; hatched circles represent sequences that do not. Rhodopsin is a single-chain integral membrane protein that spans the lipid bilayer seven times (15-17). Hydrophilic segments LI-LIV face the lumen of the disk vesicle and peptide sequences CI-CVII face the rod cell cytoplasm. CIV is shown as a fourth cytoplasmic loop that arises from the proposed anchoring of the carboxyl-terminal tail to the membrane by palmitate attached to residues Cys-322 and Cys-323 (13).

Biochemistry: K6nig et al. Mechanism of MU Stabilization. Evaluation of the competition curves for one, two, or three peptides (Figs. 3 and 4) can give some insight into the properties of the rhodopsin-GU interaction system that goes beyond the number of binding sites. The most obvious finding is that two different peptides compete with a 15 times higher efficiency than one single peptide at the same total concentration. We interpret this as evidence that two different peptides act synergistically when binding at their respective sites on Gt. Several explanations may be found for this behavior, and the synergism may be based on a concerted action of peptides (for example, mutual stabilization) or an allosteric effect on Gt or rhodopsin. However, there is in addition the shape of the competition curve that reflects a Hill coefficient of 2. The only interpretation that we have found to explain both the synergistic shift and the shape of the competition curve is the instability of the active rhodopsin conformation itself. We start from the stabilization of MI1 by interaction with Gt that leads to the measured extra MI. Occupation by peptides of two different binding sites on Gt is so effective because it leaves only one site to interact with MIL. We assume that interaction of one loop with Gt is insufficient to stabilize MIT. In the Appendix, we present a model that is quantitatively consistent with the data. It explains the IC50 values of one, two, and three different peptides and the Hill coefficients of the competition curves by interaction of the peptides with their principal site and with two other binding sites of much lower affinity. This fits with recent evidence from the activation of various G proteins by receptor fragments (D. Newton, G. Adamus, P.A.H. and W. Klee, unpublished data). By using mass action formulae, the equilibrium dissociation constants (Kd values) for binding of the peptides can be calculated. One obtains values of 2 AM for binding of synthetic peptides to their binding sites. Conclusions on the Signaling State of Rhodopsin. Independent of a special reaction model, the data show that two or three peptides bind to their sites on Gt almost as strongly as activated rhodopsin. This is evident from the fact that they effectively compete with it for Gt at a concentration of 2 pM (Fig. 3), which is only 1 order of magnitude higher than the Kd for binding of MIT (0.15 pM). It follows that a conformation of the peptide fragment that is able to bind can also be adopted by the isolated synthetic peptide. Thus the local conformation in the loop parts of the sequence does not appear to be any special "active" conformation and therefore cannot provide an essential element of signal transduction. What, then, distinguishes between the binding and nonbinding (or fitting and nonfitting) states of rhodopsin? We assume that the three sites that we have identified must be in the proper conformation and relative position (which they have on the surface of MII, but not on MI or rhodopsin). This relative position of the relevant loops in the overall conformation of MIT constitutes the "signaling state" of the receptor in the signal-transduction process. Our reaction model above argues for this notion. It explains the observed synergism of peptide competition by a strong correlation between the binding sites in MII: binding to one site does not influence its equilibrium with MI, whereas forcing two sites into the right position, by binding to Gt, is sufficient to stabilize MII, presumably including the tertiary structure around the third site. Recent evidence for independent binding of Gt ,B subunit to the rhodopsin surface (26) raises the question which of the counterparts of the interacting rhodopsin loops identified in this study may be located at the ,B subunit. It is interesting that activation of the holoprotein and release of the a subunit relieves extra MIT formation (6), which implies that the remaining ,1 subunit alone cannot stabilize MIT. In the frame of our model above this would suggest that only one of the

Proc. Natl. Acad. Sci. USA 86 (1989)

6881

binding sites is at the G-protein /3 subunit and that two sites are at the a subunit.

Use of the well studied rhodopsin-Gt system makes possible a variety of investigations that are inaccessible with other members of this class of G-protein-linked receptors. Materials are available in quantity and the extra MII measurement is a sensitive and specific assay for G-proteinreceptor interaction. These results with rhodopsin should serve to aid our understanding of the mechanism of action of this class of signal-transducing proteins.

APPENDIX Quantitative Evaluation of Peptide Competition. In evaluating the data one must note that even a single peptide is able to block MIT formation completely, although with a 15 times higher Km. It does not behave as expected for weak competition and effective binding with the Kd of the pair. This fact and the Hill coefficient of 2 of the competition curve (Fig. 3) indicate that peptides bind not only to their principal site but also to the other sites, albeit with much lower affinity. To calculate mass action, we start from the measured level of extra MIT. It is proportional to the concentration of MII-Gt complexes ([MIIG]) that is given by

[MIIGU

=

[Ro]([GO]

-

[GP])/([Go]

-

[GP]

+

Kd).

[1]

[Go] is the total concentration of G protein; [RO] is the total concentration of flash-activated rhodopsin; Kd is the [R0]related dissociation constant. Kd = Kd(MII)(1 + 1/K), where K = [MII]/[MI]. With the MII-related dissociation constant Kd(MII) = 0.15 uM and K = 0.1 at 4TC, Kd = 1.5 ILM. [GP] is the concentration of G protein that is unavailable for MIT binding due to the competing peptide. In the case [Go] >> [RO], [GP] can be treated as independent of [MIIG]. (Experimental conditions are [Go] = 0.75 ,uM and [RO] = 0.075 ,uM.) The competition curve for a single peptide fits a hyperbolic equation with a Hill coefficient of 2. This is not consistent with a simple model where the binding of one peptide prevents the binding of MIT. However, both this finding and the observed synergistic action of two peptides can be explained by the following assumptions: (i) at least two peptides must be bound to block the binding of MII; (ii) each peptide binds not only to its principal site but also, with a much lower affinity, to two other binding sites. In the presence of one peptide (P) that binds strongly to site 1 on the G protein and weakly to sites 2 and 3, the mass action law yields _ [G1P] [G12P] + [G13P] [G][P] ([G2P] [G123P] [G23P][P] 2= K3

[G2P]

_

+

[G3P])[P]

[G3P]

[G][P] [G][P] _ [G12P] + [G13P] + [G23P] ([G1P] + [G2P] + [G3P])[P]

_

[G123P] ([G12P] + [G13P])[P]

where K1 and K2 (= K3) are the constants for binding of the peptide to its principal binding site and to the other two binding sites, respectively; the indices in G1P, G12P, and G123P denote

Biochemistry: K6nig et al.

6882

Proc. Natl. Acad. Sci. USA 86 (1989)

the binding sites occupied by peptide (for example, G1P means site 1 is occupied and sites 2 and 3 are free). The concentration of free G protein is given by

[GI = [Go] - ([G1P] + [G2P] + [G3P]) - ([G12P] + [G13P] + [G23P])

-

[G123P].

Under the condition [P0] >> [Go] this can be expressed as

[GI = [Go] - (K1 + 2K2)[GM[Po] -

(K1K2 +

2K22)[G][Pof

-

2KK22[G][Po]3.

With Eq. 1, [Go] - [GP] = [GI + [G1P] + [G2P] + [G3P] (one peptide bound to the G protein is competitively ineffective), and K2[PO] > [Go]. With Eq. 1 and [Go] - [GP] = [G] + [GP1] + [GP2] this yields [MIIG]

=

[Ro[GoG]/{[Go]

+

Kd(l +

2K12[Po]2)}.

[3]

An analogous consideration for three peptides (P1, P2, and P3) leads to

With IC50 = 1.8 juM for two peptides and Eq. 6, one obtains a dissociation constant 1/K1 = 2.1 ,uM. The same value (1/K1 = 2.2 gM) is obtained from Eq. 7 and IC50 = 1.05 juM for three peptides. This confirms the validity of the model. Using this value and IC50 = 60 ,uM for one peptide, Eq. 5 yields 1/K2 = 1.1mM. The technical assistance of I. Baumle is gratefully acknowledged. This work was supported in part by the Deutsche Forschungsgemeinschaft (SFB 60 and SFB 325), by Grants EY06225 and EY06226 from the National Eye Institute of the National Institutes of Health, and by an unrestricted grant from Research to Prevent Blindness, Inc., to the University of Florida Department of Ophthalmology. P.A.H. was supported by a Jules and Doris Stein Professorship from Research to Prevent Blindness, Inc. 1. Dohlman, H. G., Caron, M. G. & Lefkowitz, R. J. (1987) Biochemistry 26, 2657-2664. 2. Peralta, E. G., Ashkenazi, A., Einslow, J. W., Smith, D. H., Ramachandran, J. & Capon, D. J. (1987) EMBO J. 6, 3923-

3929. 3. Masu, Y., Nakayama, K., Tamaki, H., Harada, Y., Kuno, M. & Nakanishi, S. (1987) Nature (London) 329, 836-838. 4. Bimbaumer, L., Codina, J., Mattera, R., Yatani, A., Scherer, N., Toro, M.-J. & Brown, A. M. (1987) Kidney Int. 32, Suppl. 23, S14-S37. 5. Stryer, L. (1986) Annu. Rev. Neurosci. 9, 87-119. 6. Hofmann, K. P. (1986) Photobiochem. Photobiophys. 13, 309327. 7. Emeis, D. & Hofmann, K. P. (1981) FEBS Lett. 136, 201-207. 8. Emeis, D., Kuhn, H., Reichert, J. & Hofmann, K. P. (1982) FEBS Lett. 143, 29-34. 9. Hamm, H. E., Deretic, D., Arendt, A., Hargrave, P. A., Konig, B. & Hofmann, K. P. (1988) Science 241, 832-835. 10. Fung, B. K.-K., Hurley, J. B. & Stryer, L. (1981) Proc. Nati. Acad. Sci. USA 78, 152-156. 11. Adamus, G., Arendt, A., Zam, Z. S., McDowell, J. H. & Hargrave, P. A. (1988) Peptide Res. 1, 42-47. 12. Schleicher, A., Kuhn, H. & Hofmann, K. P. (1989) Biochemistry 28, 1770-1775. 13. Ovchinnikov, Y. A., Abdulaev, N. G. & Bogachuk, A. S. (1988) FEBS Lett. 230, 1-5. 14. Hofmann, K. P. & Reichert, J. (1985) J. Biol. Chem. 260,

7990-7995.

[MIIG]

=

[R0][G0]/{[G0]

+

Kd(1 + 6K12[Po]2 +

6K13[Po]3)}. [4]

One has to be aware that in the case of two or three peptides the condition [P0] >> [Go] on which Eqs. 2-4 are based is no longer strictly valid under the experimental conditions; this causes a deviation in the direction of steeper functions and slightly higher IC50 values. Consistently, the hyperbolic fit to the data from two and three peptides yields somewhat higher exponents than predicted by Eqs. 3 and 4 (Fig. 3). In the frame of this model, the affinities of the peptides for their principal and other binding sites can be determined from the IC50 values; at [P0] = IC50, [MIIG] = 1[R0][G0]/([G0] + Kd). Together with Eqs. 2-4 this yields

K1K21C502

=

K12IC502

=

[Go]/Kd

+

1,

for one peptide;

[5]

([Go]/Kd + 1), for two peptides; [6]

K12C502+ K13IC503 = 6 ([GO]/Kd

+

1), for three peptides. [7]

15. Ovchinnikov, Y. A. (1982) FEBS Lett. 148, 179-191. 16. Dratz, E. A. & Hargrave, P. A. (1983) Trends Biochem. Sci. 8, 128-131. 17. Findlay, J. B. C. (1986) Photobiochem. Photobiophys. 13, 213228. 18. Kuhn, H. & Hargrave, P. A. (1981) Biochemistry 20, 24102417. 19. Franke, R. R., Sakmar, T. P., Oprian, D. D. & Khorana, H. G. (1988) J. Biol. Chem. 263, 2119-2122. 20. Strader, C. D., Dixon, R. A. F., Cheung, A. H., Candelore, M. R., Blake, A. D. & Sigal, I. S. (1987) J. Biol. Chem. 262, 16439-16443. 21. Weiss, E. R., Kelleher, D. J. & Johnson, G. L. (1988) J. Biol. Chem. 263, 6150-6154. 22. Applebury, M. L. & Hargrave, P. A. (1986) Vision Res. 26, 1881-1895. 23. O'Dowd, D. F., Hnatovich, M., Regan, J. W., Leader, W. M., Caron, M. G. & Lefkowitz, R. J. (1988) J. Biol. Chem. 263, 15985-15992. 24. Wehner, M. & Kuhn, H. (1987) Adv. Biosci. 62, 345-351. 25. Takemoto, D. J., Morrision, D., Davis, L. C. & Takemoto, L. J. (1986) Biochem. J. 235, 309-312. 26. Kelleher, D. J. & Johnson, G. L. (1989) Mol. Pharmacol. 34, 452-460.