INTRODUCTION

Certain extracellular signals, including hormones, neurotransmitters, neuromodulators, chemokines, odorants, and light, activate a class of receptors that initiate cellular effects via activation of heterotrimeric G proteins. Agonist binding to the G protein-coupled receptors leads to conformational changes that promote a tighter interaction with specific G proteins, catalysis of GDP release, and subsequent G protein activation. In the absence of guanine nucleotides, agonist binding to the receptor is stabilized by the bound G protein. The structural basis of the ternary complex among agonist, receptor, and heterotrimeric G protein is an active area of study. Extensive mutagenesis experiments, as well as peptide competition investigations for a variety of G protein-coupled receptors, have led to an understanding that the second and third cytoplasmic loops and, in some circumstances, the putative fourth loop, as well as portions of  helices VI and VII, are important in recognition of cognate G proteins (1, 2, 3). It has been shown that the heterotrimeric G protein, rather than just the  or subunits, is required for the interaction, but studies pointing out the importance of specific regions are thus far limited to the  subunits (4, 5, 6, 7).

The crystal structures of the active (GTPS¹ and GDP AlF₄-bound) (8, 9, 10) and inactive (GDP-bound) (11, 12) forms of the  subunits of transducin (G_t) and G_i1 have been reported. Analysis of the two crystal forms has established the nature of the conformational change induced by the exchange of GTP for GDP and the switch mechanism by which the presence or absence of the -phosphate defines the active or inactive state of the _t subunit (9, 11). The high resolution crystal structures of the heterotrimeric G proteins, G_t and G_i1, provide a fundamental context for understanding how a heterotrimer interacts with the membrane and with activated receptors (13, 14). The molecular mechanisms involved in the conformational changes of the  subunit and the nucleotide exchange induced by the heterotrimeric G protein interacting with activated receptors are of great interest. Denker et al. (15) suggested that agonist binding to the receptor initiates the exchange of nucleotide from G_o by movement of the _o subunit carboxyl-terminal region. The guanine ring of GDP interacts with the residues TCAT, 30 residues from the carboxyl terminus, in a loop at the amino terminus of the final -helix (5). The receptor-stimulated movement of -helix 5 may allow for the release of GDP. Consistent with this suggestion, disturbing the interactions between the conserved TCAT region of the  subunit and the guanine ring of GDP has been shown to decrease GDP affinity (16, 17).

It has been known for many years that the presence of G proteins increases agonist affinity to receptors, while binding of either GDP or GTP to the G protein complex disrupts the high affinity agonist binding state (18, 19). Thus, one of the primary ways of measuring the ternary complex formed by the agonist, receptor, and G protein, is by analyzing the agonist affinity for the receptor. G protein binding to the light-activated rhodopsin (Rh) receptor induces a high affinity state that can be measured either as decreased ability to remove the G protein from the membrane (20) or as stabilization of the active signaling state of rhodopsin, metarhodopsin II (Rh*) (21, 22). The high affinity state is induced by Rh*-catalyzed loss of GDP, leading to an empty guanine nucleotide binding pocket (23). The addition of either GTP or GDP can promote the loss of the high affinity state, measured by centrifugation (20, 24, 25) or decay of Rh* (23, 26, 27). The mechanisms involved in initiating the allosteric modulation of the agonist binding site on the receptor upon guanine nucleotide binding to the G protein are still unknown.

The studies presented here investigate the following areas: 1) determining the regions on G_t, besides those already implicated, involved in its high affinity interaction with Rh*; 2) elucidation of the conformational changes of G_t induced by the tight interaction with Rh*; and 3) resolving how the conformational changes induced by guanine nucleotide binding to the  subunit decrease the affinity of G_t for Rh* and cause Rh* decay. The tryptic digestion pattern of G_t is well known (28, 29). The GTP-induced conformational switch leads to a changed proteolytic digestion at Arg²⁰⁴, in -helix 2 (28, 30). A proteolytic site at the amino terminus of the _t subunit is partially protected by the presence of the _t subunit (31). In this work, we examine whether the high affinity interaction between G_t and activated rhodopsin (Rh*) affects the _t and _t subunit proteolytic digestion pattern. The effects of phospholipid vesicles (PL), rod outer segment (ROS) membranes kept in the dark, or ROS membranes containing Rh* on tryptic digestion patterns of G_t were investigated. In particular, we focused on the changes in the tryptic digestion pattern and time course of the _t subunit induced by Rh* interaction. Our results indicate that the presence of ROS membranes containing Rh, Rh*, or phospholipid vesicles each uniquely affects G_t proteolysis but that interaction with Rh* has the most dramatic effects.

EXPERIMENTAL PROCEDURES

TPCK-treated trypsin was purchased from Worthington. GDPS and TLCK were products of Boehringer Mannheim. The LumiGLO substrate kit and peroxidase-labeled antibodies to rabbit or mouse IgG were purchased from Kirkegaard and Perry Laboratories, Inc. Fluorescent and standard molecular weight markers were obtained from Sigma. All other chemicals and reagents were of the highest purity available.

Monoclonal antibody 4A was prepared and purified as described by Hamm and Bownds (32) and Witt et al. (33). The  subunit antisera 116, 1398, and 8645 were a generous gift of Dr. D. Manning (Department of Pharmacology, University of Pennsylvania, Philadelphia) (34, 35). Antiserum 116 recognizes the synthetic peptide IKNNLKDCGLF, which corresponds to _i1/2-(344-354). This antiserum also recognizes the _t subunit (35). Antiserum 1398 recognizes _t-(36-47) (GAGESGKSTIVK). Antiserum 8645 recognizes the synthetic peptide FDVGGQRSERKK, which corresponds to _t-(195-206).

SDS-polyacrylamide gel (12.5%) electrophoresis was carried out according to the method of Laemmli (36). Proteins and peptides were electroblotted from the SDS-polyacrylamide gel to nitrocellulose (0.1 µm, Schleicher & Schuell) essentially as described by Towbin et al. (37). After an overnight transfer, the nitrocellulose was examined under UV light to detect the fluorescent molecular weight markers. The nitrocellulose was then incubated in PBS (10 mM NaH₂PO₄, pH 7.4, 0.9% NaCl), containing 3% lowfat dried milk and 0.2% Tween 20 (PBS/milk) at room temperature. After 30 min, the nitrocellulose was incubated in PBS/milk buffer containing monoclonal antibody 4A (50 µg/ml) or an anti-_t antiserum (1:8,000 dilution) for 1 h at room temperature. The immunoblots were washed four times (10 min each) with PBS/milk buffer and incubated in PBS/milk buffer containing peroxidase-labeled second antibody (1:10,000 dilution) for 1 h at room temperature. The washing step was repeated as described above, followed by two washes with PBS and one with distilled water. The immunoblots were incubated in LumiGLO substrate for 1 min at room temperature in the light and then exposed to Kodak XAR-2 film for a few seconds.

For some experiments, the immunoblots were stripped by incubation in stripping buffer (62 mM Tris-HCl, pH 6.5, 100 mM -mercaptoethanol, 2% SDS) for 30 min at 50 °C, followed by several washes with PBS.

Neutral phosphatidylcholine (PC) vesicles were a generous gift of Dr. J. Malinski (Department of Biochemistry, Baylor College of Medicine, Houston). PC vesicles were prepared essentially as described by Kim et al. (38) and resuspended in buffer A (10 mM MOPS, pH 7.5, 200 mM NaCl, 2 mM MgCl₂) containing 0.1 mM EDTA. The phospholipid concentration was determined as described (39). In the final vesicle preparation, the total phospholipid concentration was 4.9 mM, which corresponded to that present in ROS membranes containing 82 µM rhodopsin.

Bovine ROS membranes were prepared as described previously (29). ROS membranes stripped with 4 M urea (urea-washed ROS membranes) were prepared as described (40). Aliquots of both ROS membrane preparations were stored in the dark at 80 °C until needed. Rhodopsin concentration was determined by measuring the absorbance of solubilized ROS membrane suspensions at 500 nm before and after bleaching. G_t was extracted from ROS membranes as described previously (29) and stored in 40% glycerol at 20 °C. G_t concentration was determined by the Coomassie Blue binding method (41), using bovine serum albumin as a standard (Pierce).

Limited tryptic digestion of G_t in the absence and presence of urea-washed ROS membranes kept in the dark, containing Rh* or PL, was performed essentially as described by Mazzoni et al. (29). G_t was incubated in buffer A (10 mM MOPS, pH 7.5, 200 mM NaCl, 2 mM MgCl₂, 1 mM dithiothreitol) with urea-washed ROS membranes either in the dark or light for 10 min at room temperature. The rhodopsin concentration was 36 µM, while the G_t:rhodopsin molar ratio was 1:5. For some experiments performed in ambient light, 100 µM GDPS was present in the buffer throughout the assay. All of the manipulations of samples containing urea-washed ROS membranes kept in the dark were performed under dim red light. After incubation, samples were centrifuged at 13,000 × g for 5 min at room temperature.

G_t (0.55 mg/ml) was similarly incubated in buffer A with PL (2.2 mM PC). The samples containing ROS membranes or PL were centrifuged at 50,000 rpm for 30 min at 4 °C in a Beckman fixed angle TL-100.2 ultracentrifuge rotor. Pellets were resuspended in buffer A containing 25% glycerol at a G_t concentration of 1.6 mg/ml and kept in ice. After removing an aliquot for the control at time 0, trypsin solution (0.06 mg/ml in buffer A containing 25% glycerol) was added. The final concentration of G_t was 0.8 mg/ml, and the trypsin:G_t (w/w) ratio was 1:25. As a control, G_t was subjected to identical treatment in the absence of urea-washed ROS membranes. The reaction was terminated by incubating an aliquot of the reaction mixture with TLCK at a final concentration of 40 µg/ml. After 5 min, 2 × electrophoresis sample buffer (37) was added, and the samples were immediately frozen and stored at 80 °C. Proteolytic fragments were separated by SDS-polyacrylamide (12.5%) gel electrophoresis as described above.

For both Coomassie Blue-stained gels and immunoblots, the density of polypeptide bands was measured using a Molecular Dynamic densitometer (Personal Densitometer SI). Curve fitting of the data was performed with nonlinear least square criteria using GraphPad Prism software.

RESULTS

The recent determination of the high resolution crystal structure of the heterotrimeric G protein, G_t, provided some indication of the orientation of the molecule with respect to the membrane and the visual receptor rhodopsin. To further understand the structural basis of the high affinity interaction between rhodopsin and G_t we determined the effect of the rhodopsin-G_t complex formation on the accessibility of tryptic cleavage sites on the _t and _t subunits.

The sequential appearance of proteolytic fragments during the time course of G_t-limited tryptic digestion has been widely studied, and the origin of these fragments is well characterized (28, 29, 30) (Fig. 1). As indicated, the _t subunit has three regions that are readily available for limited tryptic digestion: Lys¹⁸, Arg²⁰⁴, and Arg³¹⁰. Five transient proteolytic fragments of the _t subunit appear (₃₈, ₃₄, ₃₂, ₂₃, and ₁₅), while the final fragments are ₂₁, ₁₂, ₅, and ₂ (29). The _t subunit is rapidly converted to two stable fragments (₂₃ and ₁₄) (Fig. 1), while the _t subunit remains intact (28). Under nondenaturing conditions, the two tryptic fragments of the _t subunit and the _t subunit remain tightly associated (28).

Distribution of proteolytic fragments from the _t and _t subunits.

Purified G_t was incubated with PL, ROS membranes containing inactive (Rh), or light-activated rhodopsin (Rh*), and after centrifugation to remove any soluble G_t the membrane fractions were incubated with trypsin. The rate of cleavage was followed for 2 h. Potentially, interaction of G_t with membranes or Rh* could either protect or increase accessibility of known cleavage sites or expose new cleavage sites.

The time course of limited digestion of G_t in the presence and absence of urea-washed ROS membranes is shown in Fig. 2. When soluble G_t was digested with trypsin, the sequential appearance of proteolytic fragments that migrated in the Coomassie Blue-stained polyacrylamide gels with apparent molecular weights of 38, 34, 32, 23, 21, 14, and 12 kDa was evident (Fig. 2A). According to our previous study (29) (Fig. 1) these peptide bands corresponded to five _t (₃₈, ₃₄, ₃₂, ₂₁, and ₁₂) and two _t fragments (₂₃ and ₁₄) A dramatically different result was obtained when G_t was incubated with ROS membranes containing Rh* and allowed to form a high affinity complex. In this complex, the _t subunit remained relatively uncleaved during the 2 h of digestion time course (Fig. 2B). To determine whether this protection occurred specifically because of the high affinity interaction with Rh*, parallel experiments were performed using G_t bound to ROS membranes kept in the dark (Fig. 2D) or G_t-Rh* incubated with GDPS, which dissociates the high affinity complex (26, 27) (Fig. 2C). When G_t was in the presence of inactive Rh, the time course of tryptic digestion of _t was much more similar to that observed for soluble G_t, while in the presence of Rh* and GDPS, the tryptic pattern was intermediate. Since rhodopsin was also partially cleaved by trypsin, the proteolytic digestion pattern of the _t subunit was difficult to analyze in the Coomassie Blue-stained gels due to the presence of numerous peptide bands (Fig. 2) (see below).

Time course of G_t limited tryptic digestion in the presence and absence of ROS membranes containing Rh, Rh*, or Rh* plus GDPS.

In the presence of urea-washed ROS membranes, the fragments originating from the cleavage of the _t subunit were visible in the gels (Fig. 2). Although under all three conditions a small amount of ₂₃ and ₁₄ fragments was present during the time course, it appeared that there was less of each fragment throughout the time course (Fig. 2). This was different from the digestion time course of soluble G_t, in which the two fragments were generated immediately. Thus, it would appear that the presence of membranes of any kind partially protects the _t subunit from tryptic cleavage at Arg¹²⁹.

To examine the presence of _t fragments more clearly, immunoblotting with antipeptide antibodies of defined specificity was carried out. The anti-_t-(195-206) antiserum 8645 recognized several _t fragments, which were easily identifiable in the immunoblot (Fig. 3A), ₃₈, ₃₄, ₃₂, ₂₃, and ₂₁. In the presence of Rh*, the _t subunit was a very poor substrate for tryptic cleavage (Fig. 3B). The ₃₄ and ₃₂ fragments were not generated, while low amounts of ₂₃ and ₂₁ appeared very late in the time course. Similar experiments performed with monoclonal antibody 4A, which binds to the amino-terminal region of the _t subunit (29); antiserum 1398, which binds to the phosphate binding loop (34); and antiserum 116, which binds to the carboxyl-terminal region (35) confirmed that under conditions of high affinity interaction between G_t and Rh*, the three major tryptic sites on the _t subunit were protected from cleavage (data not shown).

Identification of _t proteolytic fragments by immunoblot using anti-_t-(195-206) antiserum.

In the presence of ROS membranes that had not been light-activated, (Fig. 3D) or phospholipid vesicles (Fig. 3C), the time courses of limited tryptic digestion of the _t subunit were more similar to that obtained in the absence of membranes. However, the relative production of ₃₂ was lower in the presence of membranes or PL, probably due to protection of the cleavage site at Lys¹⁸. In addition, in the presence of Rh less ₃₄ was produced, suggesting a modest protection of Arg³¹⁰, and the ₂₃ fragment was somewhat stabilized.

To quantitate the data, we analyzed the immunoblots by densitometry (Fig. 4). Both the disappearance of _t and the appearance of the ₃₄, ₃₂, and ₂₁ fragments were measured. Fig. 4 shows clearly that the _t subunit was protected in the presence of Rh*. This implies that all three major proteolytic sites of _t are protected when G_t is in high affinity interaction with Rh* (Table I).

Quantitative analysis of _t subunit and proteolytic fragments detected using anti-_t-(195-206) antiserum.

Table I. Protected tryptic cleavage sites of the 1 subunit by incubation of Gt with PL, Rh, or Rh* Lys18 Arg264 Arg310 Gt +/a     Gt + Rh* ++ ++ +++ Gt + Rh* + GDPS ++ +/ + Gt + Rh ++   +/ Gt + PL ++     a  +++, complete protection; ++, protection; +, mild protection; +/, partial protection; , no protection.

In the presence of PL, there was an increased generation of ₃₄ (Fig. 4B) and decreased generation of the ₃₂ fragment (Fig. 4C). This suggests that interaction of G_t with phospholipids leads to protection of the cleavage site at Lys¹⁸ (see also Fig. 5). In the presence of Rh, there was also a decreased production of the ₃₂ fragment (Fig. 4C) but no compensating increase of ₃₄ (Fig. 4B). In fact, there was a decreased production of the ₃₄ fragment. This suggests that Rh is capable of weakly protecting the cleavage site at Arg³¹⁰. In the presence of either PL or Rh, the production of ₂₁ was increased with respect to G_t in solution (Fig. 4D). It thus appears that proteolytic cleavage at Arg²⁰⁴ is accelerated when Lys¹⁸ is protected.

Quantitative analysis of the ₃₈ fragment produced during limited tryptic digestion of G_t in the presence and absence of phospholipid vesicles.

GDPS is known to reverse the high affinity state between G_t and Rh* (26, 27). The data (Fig. 4, Table I) show that the high affinity interaction with Rh* was needed for its protection of proteolytic sites on _t, because there was a partial reversal of the protection of Arg³¹⁰ and Arg²⁰⁴. Comparison of the amounts of ₃₄ and ₃₂ shows that even in the low affinity state, the cleavage site at Lys¹⁸ was protected, consistent with the ability of PL or Rh to protect this site.

To examine in more detail the ability of phospholipids to protect the cleavage site at Lys¹⁸, we studied the appearance of the transient ₃₈ fragment in the presence or absence of PL. Fig. 5 shows that PL completely inhibited the production of the ₃₈ fragment. In addition, in the presence of PL, the formation of ₃₂ was decreased at the expense of an increased formation of the ₃₄ fragment.

Table I summarizes the findings, while the locations of the various cleavage sites on the G_t heterotrimer are shown in the stereoview in Fig. 6. It is interesting to note that all _t tryptic cleavage sites are located on the same side of the molecule (Fig. 6), in close proximity to the receptor, _t subunit, and effector binding regions. This observation may suggest that protein regions that are involved in protein-protein interaction are flexible.

Stereo view of heterotrimeric G_t showing the tryptic cleavage sites on _t and _t.

DISCUSSION

The studies described here provide insight into the regions of a G protein that bind to an activated receptor. The accessibility of G_t to tryptic proteolysis is changed dramatically when it is in high affinity interaction with Rh*. We have dissected the effects of phospholipids, ROS membranes containing inactive Rh, and the low affinity interaction of GDPS-bound G_t with Rh* on the availability of the tryptic cleavage sites.

It is well known that heterotrimeric G proteins are membrane-associated even in the absence of activated receptors. This association is thought to be mediated mainly by interactions of fatty acyl and prenyl groups of  and  subunits with membranes (42, 43). All G protein  subunits are modified at or near their amino termini by covalent attachment of the fatty acid myristate and/or palmitate, while the  subunits are prenylated at a cysteine residue located in their carboxyl termini (43). The _t subunit is heterogenously modified at its amino terminus by myristate and three other less hydrophobic fatty acids (44, 45), but it does not contain palmitate. The presence of membranes is important for _t-_t subunit interaction (46), so it might be expected that their presence affects the proteolytic cleavage pattern of G_t. In fact, we found that both phospholipid vesicles and ROS membranes protect the cleavage site in the amino terminus of _t at Lys¹⁸.

Numerous biochemical (29, 31, 47, 48) and mutational (49, 50) studies have implicated the amino terminus of G protein  subunits in the interaction with subunits. The resolution of the crystal structure of heterotrimeric G_t (14) has confirmed that the amino-terminal helix of the _t subunit is involved in forming a binding site for the _t subunit. Numerous potential proteolytic cleavage sites are located in this amino-terminal helix of _t (29). Navon and Fung (31) have shown that the presence of _t reduces the cleavage rate of the _t subunit at Glu²¹ by Staphylococcus aureus V8 protease. The chymotryptic cleavage sites at Leu¹⁵ and Leu¹⁹ are partially protected (51) by the presence of _t that directly contacts these residues (14). It is likely that in the heterotrimeric G_t the tryptic cleavage site at Lys¹⁸ is also partially protected. The presence of membranes of any kind determines an increase of the protection. The effect may be direct or the consequence of a conformational change of the amino-terminal -helix of _t induced by the concomitant interaction with _t and membranes. Overall, the protection suggests that this region is in contact with the membrane, consistent with the crystal structure of the heterotrimer (14).

Our results indicate that in the presence of Rh* the _t subunit of heterotrimeric G_t is almost completely protected from tryptic hydrolysis. In particular, the cleavage site at Arg³¹⁰ is not available for digestion. We cannot determine whether this protection is the consequence of Rh* binding to that region of _t or a change of _t conformation upon Rh* binding. However, the former possibility is in agreement with previous observations. Data reported by Hamm et al. (4) indicate that the _t residues 311-323 participate with the carboxyl-terminal residues 340-350 to form the binding site for Rh*. Resolution of the crystal structure of the _t subunit in the GTPS and GDP-bound forms (8, 11) has shown that _t residues 320-323 (TCAT) contribute directly to nucleoside binding through a van der Waals contact with the guanine ring. A receptor-stimulated movement of -helix 5 may allow for the release of GDP. Consistent with this suggestion, disturbing the interactions between the conserved TCAT region (6-5 loop) of the  subunit and the guanine ring of GDP has been shown to decrease GDP affinity (16, 17). The crystal structures of _t-GDP and heterotrimeric G_t also show that Arg³¹⁰ is located in a surface-exposed loop, the 4/6 region, which is spatially close to the carboxyl-terminal residues, 340-350 (Fig. 6). The activated receptor could trigger the release of bound guanine nucleotide by interacting with this region in concert with the COOH-terminal region. Binding of GDPS interacting through the guanine ring with the 6-5 loop could decrease the affinity of the _t subunit for Rh* via 6 or 5. This effect is demonstrated here by the partial reversal of the protective effect of Rh* at Arg³¹⁰ by GDPS (Table I). Both guanine nucleoside diphosphate and triphosphate are able to modulate the interaction of the _t subunit with Rh* and cause Rh* decay as reported by several observations (26, 27). The mechanism of this communication between the Rh* and nucleotide binding sites is not known.

In the presence of dark-adapted rhodopsin, the _t cleavage site at Arg³¹⁰ is not completely available for tryptic digestion as indicated by the production of lower amounts of the 34-kDa fragment. A larger amount of the fragment is generated when G_t is digested in the presence of phospholipid vesicles. These observations suggest that there is a low affinity interaction of this region with inactive Rh.

Our data also indicate that in the presence of Rh* another tryptic cleavage site of _t, Arg²⁰⁴, is less available to digestion. In the presence of GDPS, the accessibility to this cleavage site is partially restored. Arg²⁰⁴ is located in the switch II region, residues 198-215 (8), and is protected from tryptic hydrolysis in the _t·GTPS complex but not in the _t·GDP complex (28). The switch II region undergoes a distinct transition from a distorted helical structure with the Arg²⁰⁴ side chain in a solvent-exposed conformation when _t is in the GDP-bound form to a well ordered helix (2-helix) interacting with residues from the 3-helix and switch III in the _t·GTPS complex (9). Structural studies have shown that this region forms an important part of the _t contact interface (14) (Fig. 6). In the presence of Rh*, protection of the cleavage site of _t at Arg²⁰⁴ can result from a direct interaction of Rh* with this region or as a consequence of a conformational change induced by the activated receptor. Alternatively, Rh* can directly or indirectly cooperate with _t to determine this protection. In the presence of GDPS, this protective effect is reduced. Thus, after release of GDP as a consequence of _t activation by Rh*, the binding affinity of the switch II region for _t may increase, and the accessibility of the protease to Arg²⁰⁴ is almost completely blocked. Our results suggest that in the "empty pocket" state the _t subunit forms a tight complex with _t and Rh*.

In conclusion, using limited tryptic digestion as an experimental approach, we have defined the regions of G_t involved in the interaction with the activated receptor. In the presence of Rh*, two major cleavage sites of the _t subunit, Arg²⁰⁴ and Arg³¹⁰, are protected, while the cleavage rate at Lys¹⁸ is substantially decreased in the presence of either ROS membranes or PL.