Receptor and βγ Binding Sites in the α Subunit of the Retinal G Protein Transducin

See allHide authors and affiliations

Science  17 Jan 1997:
Vol. 275, Issue 5298, pp. 381-384
DOI: 10.1126/science.275.5298.381


Transmembrane receptors for hormones, neurotransmitters, light, and odorants mediate their cellular effects by activating heterotrimeric guanine nucleotide-binding proteins (G proteins). Crystal structures have revealed contact surfaces between G protein subunits, but not the surfaces or molecular mechanism through which Gαβγ responds to activation by transmembrane receptors. Such a surface was identified from the results of testing 100 mutant α subunits of the retinal G protein transducin for their ability to interact with rhodopsin. Sites at which alanine substitutions impaired this interaction mapped to two distinct Gα surfaces: a βγ-binding surface and a putative receptor-interacting surface. On the basis of these results a mechanism for receptor-catalyzed exchange of guanosine diphosphate for guanosine triphosphate is proposed.

Heterotrimeric G proteins relay signals from seven-transmembrane-spanning receptors (7TMRs) to cellular enzymes and ion channels. Activated by photons, odors, and many hormones and neurotransmitters, 7TMRs catalyze replacement by guanosine triphosphate (GTP) of guanosine diphosphate (GDP) bound to Gα subunits, causing dissociation of the Gα·GTP and βγ subunits, which in turn relay the signal to downstream effectors (1). Activation is robust; in the case of transducin (Gt), rhodopsin induces an ∼107-fold increase in the basal (unstimulated) rate of GTP-GDP exchange (2). Crystal structures (3, 4, 5) have defined conserved folds of G proteins, as well as the trimeric structure, GTP-induced conformational change, and a plausible catalytic mechanism for GTP hydrolysis, which turns off the signal. It is not known, however, how the G protein interacts with its 7TMR and responds by exchanging GDP for GTP.

Biochemical and molecular genetic approaches have defined three key features of this interaction: (i) The 7TMR interacts directly with both βγ and Gα (6, 7). (ii) The COOH-terminal tail of Gα (∼10 residues) interacts directly with the receptor (7, 8, 9, 10). (iii) In addition to the COOH-terminal tail, the interaction involves other residues in Gα not yet identified (7, 9). Here we report results of a comprehensive molecular genetic strategy that confirm these findings and identify a putative 7TMR-interacting surface of Gα, and we suggest a mechanism for 7TMR-triggered conformational change.

We used two assays of G protein activation to test 100 Gα mutants in which alanine replaced individual amino acid residues in the α subunit of Gtt) [92 single and 8 double replacements (11)]. In the crystal structure, almost all of these residues are oriented toward the solvent (12). The first assay measured the ability of rhodopsin, in the presence of all-trans-retinal, to promote guanosine 5′-O-(3′-triotriphosphate) (GTP-γ-S) binding to αt in microsomal extracts from COS-7 cells transiently expressing αt, β1, γ1, and opsin. Binding of GTP-γ-S induces a conformational change that is detected by the accumulation of a trypsin-resistant αt fragment (Fig. 1) (10). The second assay (Fig. 2) assessed the ability of αt, translated in vitro and labeled with [35S]methionine, to bind photoactivated rhodopsin in membranes prepared from retinal rod outer segments (13). Activation by AlF4 served as a functional control in both assays.

Fig. 1.

Activation of normal and representative mutant αt molecules by rhodopsin and AlF4 in microsomes. (A) Microsomal fractions from COS-7 cells expressing αt alanine mutants, bovine opsin, β1, and γ1 were incubated with all-trans-retinal under ambient light. Samples were removed at the indicated times before and after the addition of GTP-γ-S (10 μM) and treated with trypsin, as described (10). Lanes labeled “−Retinal” show the slow rate of GTP-γ-S-for-GDP exchange in the absence of photoactivated rhodopsin. Lanes labeled “AlF4” represent incubations that contained AlF4, but not GTP-γ-S (10). Arrows indicate the 31-kD trypsin-resistant αt fragments. Standard lanes represent percentages of the original microsomal extract, not treated with trypsin. Abbreviations: WT, normal αt; mRR, moderately impaired receptor response; sRR, severely impaired receptor response; WT*, normal interaction when assay results are corrected for a moderate decrease in protection by AlF4; I, indeterminable activation by rhodopsin. For the mutant designations, A is Ala, I is Ile, K is Lys, and Y is Tyr, and the number indicates the position of the mutated residue. (B) Quantitative analysis (by densitometry of immunoblots) of αt protection by activated rhodopsin (R*) or AlF4. For each mutant, values represent the mean ± 2 SEM of at least four separate experiments, performed on microsomal membranes derived from at least two separate transfections. The shaded area depicts the range (mean ± 2 SEM) of results with normal αt (WT) in the presence of all-trans-retinal. (C) Values of activation time courses, corrected for the amount of protection by AlF4 [calculated as the quotient of the uncorrected values divided by the mean values for AlF4 protection (11)].

Fig. 2.

Binding to photoactivated rhodopsin of normal and mutant αt translated in vitro. (A) Samples of the diluted translation mix, containing [35S]methionine-labeled αt, were incubated with rod outer segment membranes in ambient light, in the presence or absence of GTP-γ-S, as described (13). Standard lanes represent percentages of total labeled αt in the translation extract. Samples were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography (13). For the mutant designations, D is Asp and L is Leu. (B) Quantitative analysis (by PhosphorImager) of binding results. R* binding depicts the percentage of total radiolabeled αt bound to rhodopsin in the absence (−) or presence (+) of 100 μM GTP-γ-S. For each mutant, values represent the mean ± 2 SEM of at least four separate experiments, performed on in vitro-translated αt derived from at least two separate preparations. (C) Autoradiograms and quantitative analysis of the percent of αt protected from trypsin by AlF4; arrows indicate the 31-kD trypsin-resistant αt fragments. Std., standard (100% equivalent). (D) Corrected values of binding experiments for representative alanine mutants. Bars represent the difference between binding in the absence and the presence of GTP-γ-S, normalized with respect to protection by AlF4 (11).

The microsomal assay revealed unequivocally impaired responses to rhodopsin in 24 of the 88 mutants (11) in which the effect of rhodopsin could be determined. We subclassified these (11) as severely or moderately impaired in receptor response (sRR or mRR); Fig. 1 depicts a representative sRR and an mRR phenotype. The remaining 64 mutants showed normal (WT) responses. Of the 54 in vitro-translated αt mutants in which we could assess binding to photoactivated rhodopsin, 22 had impaired receptor binding (RB), whereas 32 phenotypes were WT. Autoradiograms and quantitation (by PhosphorImager) of αt bound to rhodopsin, in the absence or presence of GTP-γ-S, for four representative mutants is shown in Fig. 2. In the presence of GTP-γ-S, considerably smaller amounts of [35S]methionine-labeled normal (unmutated) αt bound to rhodopsin, presumably because rhodopsin efficiently induces αt to bind GTP-γ-S, putting it into a stable conformation that binds poorly to both βγ and to rhodopsin. For all mutants, classification into WT or RB categories was the same whether we assessed total αt binding or the difference between binding in the absence or presence of GTP-γ-S (14).

As a control in both assays, αt mutants were tested for susceptibility to protection from tryptic cleavage by a receptor-independent stimulus, AlF4 [which mimics the γ-phosphate of GTP, thereby switching GDP-bound Gα into a conformation that resists trypsin cleavage (3, 4, 10, 15)]. For some mutants in both assays, AlF4 protection was reduced but reproducible. In these cases we assessed the rhodopsin interaction phenotype by normalizing it relative to the degree of protection effected by AlF4 (Figs. 1 and 2, mutants I339A and D311A, respectively). With the normalization procedure we classified 15 and 7 mutant phenotypes as WT in the activation and binding assays, respectively. By the criterion of resistance to trypsin, small numbers of mutants failed to respond to AlF4 under each assay condition (12 of 100 and 12 of 66 in the microsomal and in vitro translation assays, respectively) (16). Rhodopsin activation or binding could not be assessed for these mutants, which are designated (11) indeterminable (I) (Y316A in Fig. 1, L349A in Fig. 2). The phenotypes of all but seven mutants could be determined in at least one of the assays (16). The informative 93 mutants were thus designated either R or WT, to indicate whether or not the mutation impaired interaction with rhodopsin (11).

To construct a three-dimensional (3D) model of G protein activation by 7TMRs (Fig. 3), we assigned to 34 R residues a specific role as components of binding surfaces that may interact directly with βγ or rhodopsin. Although any alanine substitution could indirectly hinder the interaction by blocking rhodopsin-induced conformational change, rather than by preventing binding of αt to rhodopsin, assignments of several residues into the R category are in keeping with previous evidence (5, 7, 9, 17).

Fig. 3.

Alanine mutations and 3D structure of αt. (A to C) Views of αt rotated successively by 90° about the vertical axis, which is parallel to the postulated (5, 18) plane of the plasma membrane; the membrane presumably is located just to the right of the model in (A) and just to the left of the model in (C) (vertical yellow stripes). Panel (A) represents the face of αt that binds to βγ and (B) the face of αt that would be “seen” by the receptor in the membrane, whereas (C) depicts the face of αt opposite to that seen in (A). Residues at sites where alanine substitution impaired interaction with rhodopsin are colored cyan (βγ-interacting surface, outlined in magenta), red (putative rhodopsin-interacting surface, outlined in dark blue), and green [other residues, as indicated in (24)]. Beige or black indicate, respectively, sites at which mutations produced WT or indeterminate phenotypes. In (A) and (C), GDP (yellow) can be glimpsed between the two Gα domains (the α-helical domain is colored gray). Ct, COOH-terminal residues 344 to 350 of αt; not visible in crystals of the heterotrimer (5), these are taken from a separate crystal structure (3). Nt, NH2-terminus. (D and E) Ribbon representations of αt in a complex (5) with subunit β (dark blue; γ not shown). In (D) and (E) αt is tilted ∼90° around the horizontal axis with respect to (A), so that the NH2-terminal α-helix points toward the viewer. Panel (D) shows locations of sites with an R phenotype, colored as in (A) to (C). Panel (E) highlights structural elements postulated to transmit conformational change from the receptor to the guanine nucleotide-binding pocket. These include α2 and β1 through β3 (cyan), which communicate with βγ, and β6 and α5 (red), proposed as key conformation transmitters between the β6-α5 loop (green, contacting GDP) and parts of the protein that interact directly with membrane-bound receptor, R (α4-β6 loop, mid-α5, and the COOH-terminus).

Mapped onto the crystal structure (5) of αt, these 34 R residues form two sets of clusters. Our model assigns these to putative surfaces for interaction with βγ or with the 7TMR (Fig. 3, A through D). Both surfaces of αt are probably similar in location and function to corresponding surfaces of other Gα molecules (1, 4, 5, 7). Indeed, the two surfaces functionally defined in our experiments closely match surfaces predicted (18) by inference from mapping residues conserved among Gα subunits onto the αt 3D structure.

Rhodopsin and other 7TMRs appear to “act at a distance” (18) in catalyzing GTP-for-GDP exchange, in that intracellular loops of most 7TMRs are too short to allow direct interaction with the guanine nucleotide-binding site, which is probably located ∼30 Å from the plasma membrane (Fig. 3, A and C). On the basis of the proposed βγ- and 7TMR-interacting surfaces, we devised an explanation for propagation of the conformational signal from the receptor to the nucleotide-binding pocket.

Most of the αt residues that directly contact βγ in crystals of the heterotrimer (5) showed an R phenotype (17), confirming that the 17 R residues (cyan-colored in Fig. 3, A through C) coincide with the βγ-binding surface of αt. βγ-interacting residues are distributed among four elements of secondary structure (11), three of which (Fig. 3, D and E) stretch from the postulated plane of the plasma membrane (5, 18) to the GDP-binding pocket. It is likely that βγ cooperates with the 7TMR to open the nucleotide-binding site, either actively, by inducing conformational change in Gα, or indirectly, by enhancing affinity of the Gαβγ complex for the 7TMR; in the latter role, βγ could also serve as a stabilizer or fulcrum to make Gα more susceptible to a separate action of the 7TMR.

A putative 7TMR-interacting surface (outlined in dark blue in Fig. 3) is located near the membrane, at the ends of α5 and β6 (that is, the COOH-terminal tail and the α4-β6 loop, respectively; Fig. 3, A through C). Additional mutation-sensitive sites located in both β6 and α5 extend part way toward the guanine nucleotide-binding pocket (11). We propose (Fig. 3, D and E) that 7TMR-induced changes in the conformations of the β6 strand and the α5 helix are communicated to the β6-α5 loop (green in Fig. 3E), in which side chains of a short sequence of amino acids contact the guanine ring of GDP (3, 4, 18, 19).

Although not yet explicitly tested, this idea is in keeping with our present observations and with results of studies that probed functions of β6, α5, and the β6-α5 loop (20, 21, 22, 23). Specifically, several mutations in α5, at locations distant from the nucleotide-binding pocket, increase rates of GDP dissociation and its replacement by GTP (20). Amino acid substitutions in the β6-α5 loop increase the GDP dissociation rate in several guanosine triphosphatases (GTPases), promoting spontaneous GTP-for-GDP exchange (21, 22, 23). Such mutations dramatically activate both p21ras and αs—producing, respectively, neoplastic transformation (22) and a human endocrine syndrome, testotoxicosis-pseudohypoparathyroidism (23). Thus 7TMR-induced conformational change may be propagated from the membrane through β6 and α5 to the loop that connects them, right in the GDP-binding pocket (Fig. 3, D and E).

Web Supplement

The accompanying table lists phenotypes produced by 100 alanine substitution mutations in αt. Phenotypes are summarized for assays of αt activation in microsomes of cells expressing rhodopsin and Gt (10, 12) (Fig. 1 legend) and for binding of in vitro-translated αt to photoactivated rhodopsin in rod outer segment membranes (13) (Fig. 2 legend). The overall phenotype was judged to be impaired in rhodopsin interaction (R) if such a defect was observed in either assay and indeterminate (I) only if the ability of AlF4- to protect αt from tryptic degradation was severely reduced in both assays (below 15 and 10% in the microsomal and in vitro translation assays, respectively); all other phenotypes were termed wild type (WT). The next column (secondary structure) indicates the structural element to which each mutated residue belongs (5), including α helix (α), β strands (β), or loops (L) between these (for example, β6/ α5 L). A plus sign in this column indicates that the residue contacts βγ in crystals of the heterotrimer (5, 18). The far right-hand column indicates the putative interaction—with βγ or rhodopsin (Rhod)—we attribute to the normal residue located at a site where mutation produced an R phenotype. Two additional residues are required for myristoylation of αt (Myris), and two more interact with GDP (19).

Specific criteria in the activation assay are as follows: Alanine mutants were considered impaired in activation if mutant values did not fall within 2 SEM of values for unmutated αt; sRR and mRR mutants showed <20% or 20 to 40% of normal activation, respectively, after a 4-min exposure to GTP-γ-S in membranes containing active rhodopsin. In the binding assay, impaired binding (RB) indicates that binding values do not overlap with the WT value, ±2 SEM.

For some mutants in both assays, AlF4- protection was reduced but reproducible. In these cases we assessed the rhodopsin interaction phenotype by normalizing it relative to the degree of protection effected by AlF4-. An asterisk (WT*) indicates that the receptor interaction in an assay is considered normal when results are corrected for moderate decreases in protection by AlF4-. Average levels of AlF4- protection for the alanine mutants tested in the rhodopsin activation assay for the five phenotypes were as follows: WT, 55%; WT*, 32%; mRR, 55%; sRR, 41%; and I, 4%; in this assay we placed mutants in the WT* category only if a corrected activation curve similar to that seen with unmutated αt resulted from dividing the fraction of mutant αt protected (by exposure to all-trans-retinal plus GTP-γ-S) by the corresponding AlF4—protected value. Average levels of AlF4- protection for the alanine mutants tested in the receptor binding assay were WT, 25%; WT*, 19%; RB, 26%; and I, 3%.

Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr. The abbreviation C-ter refers to COOH-terminus.

View this table:


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
    In this theoretical study, residues in cluster 2 and cluster 1 correspond, respectively, to the βγ- and putative 7TMR-interacting surfaces identified in the present work.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
View Abstract

Stay Connected to Science

Navigate This Article