The Ca2+-dependent Binding of Calmodulin to an N-terminal Motif of the Heterotrimeric G Protein β Subunit*

Ca2+ ion concentration changes are critical events in signal transduction. The Ca2+-dependent interactions of calmodulin (CaM) with its target proteins play an essential role in a variety of cellular functions. In this study, we investigated the interactions of G protein βγ subunits with CaM. We found that CaM binds to known βγ subunits and these interactions are Ca2+-dependent. The CaM-binding domain in Gβγ subunits is identified as Gβ residues 40–63. Peptides derived from the Gβ protein not only produce a Ca2+-dependent gel mobility shifting of CaM but also inhibit the CaM-mediated activation of CaM kinase II. Specific amino acid residues critical for the binding of Gβγ to CaM were also identified. We then investigated the potential function of these interactions and showed that binding of CaM to Gβγ inhibits the pertussis toxin-catalyzed ADP-ribosylation of Gαo subunits, presumably by inhibiting heterotrimer formation. Furthermore, we demonstrated that interaction with CaM has little effect on the activation of phospholipase C-β2 by Gβγ subunits, supporting the notion that different domains of Gβγ are responsible for the interactions of different effectors. These findings shed light on the molecular basis for the interactions of Gβγ with Ca2+-CaM and point to the potential physiological significance of these interactions in cellular functions.

In response to external stimuli or changes in ligand concentration, the activated seven-transmembrane receptors of the cell interact with heterotrimeric G proteins. These in turn activate or inhibit a variety of intracellular proteins through either the G␣ subunits with GTP bound or the free G␤␥ subunits or both, leading to the generation of second-messenger molecules and changes in the patterns of cellular metabolic activity and growth. Whereas direct interaction of G protein ␣ subunits with effectors has been known for a long time, recently it has been demonstrated that G␤␥ subunits also play critical roles in effector activation, in modulating the interaction among various G protein pathways, and in regulating cellular functions. For example, the G␤␥ dimers have been shown to participate in interactions with adenylyl cyclases (1), phospholipases (2,3), phosducin (4), ␤-adrenergic receptor kinases (5-7), Bruton tyrosine kinase (8), inositol trisphosphate kinases (9,10), mitogen-activated protein kinase (11), and a number of ion channels (12)(13)(14)(15)(16). The activation of phospho-lipase C-␤2 and ␤3 by G␤␥ has been suggested to account for the PTX 1 -sensitive increase in intracellular Ca 2ϩ in response to a number of chemoattractants (17,18) and other ligands, thus regulating intracellular Ca 2ϩ concentration.
Calmodulin (CaM) has been known to act as an intracellular calcium sensor protein. When the intracellular Ca 2ϩ concentration increases, CaM can bind up to four Ca 2ϩ ions, changing its conformation and regulating cellular functions such as activation or inhibition of a large number of enzymes (19,20), ion channels (21), and receptors (22). These Ca 2ϩ -dependent interactions of CaM with its target proteins have played an important role in intracellular Ca 2ϩ signaling and in various cellular functions including cell growth and differentiation. There are reports demonstrating that G protein ␤␥ complex can bind to CaM when passed through the CaM-agarose column (23,24); however, the precise nature of the interaction is not clear.
In this report, we investigated the interaction of Ca 2ϩ -CaM with G␤␥ subunits and the potential physiological significance of this interaction. We found that CaM not only binds to G␤␥ subunits purified from brain but also to the most diverse G␤ 5 L␥ complex from retina. We then identified and characterized the CaM-binding domain of the G␤ subunit by using synthetic peptides and site-specific mutation. Furthermore, we showed that binding of CaM to ␤␥ inhibits the ␤␥-dependent PTXcatalyzed ADP-ribosylation of G␣ subunits. Using both brain G␤␥ subunits and the CaM-binding peptide derived from the ␤ subunit, we demonstrated that the Ca 2ϩ -CaM-dependent activation of CaM kinase II could be inhibited at a molar ratio of 1 (peptide/CaM). These studies provide insight into the molecular basis for the interactions of ␤␥ with Ca 2ϩ -CaM and the potential functions of these interactions in modulating cellular signaling and other functions.

Materials
Synthetic peptides were obtained from the Peptide Synthesis Facility (Beckman Institute, California Institute of Technology). Peptides were purified by high performance liquid chromatography and verified by mass spectrometry. Bovine brain CaM was purchased from Calbiochem.

CaM Binding Assay of G Protein ␤5L␥ Subunits
Bleached bovine rod outer segment membranes were prepared from bovine retina as described (25). Transducin was removed from the membranes by 6 ϫ hypotonic elution with 100 M GTP␥S in the presence of 3 mM MgCl 2 (26). The remaining membranes were solubilized for 12 h at 4°C in 20 ml of a buffer containing 50 mM Hepes, pH 8.0, 100 * This work was supported by a grant from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  mM NaCl, 3 mM MgCl 2 , 10 mM 2-mercaptoethanol, and 1% (by mass) sodium cholate. Unsolubilized material was pelleted for 40 min at 100,000 ϫ g, and the supernatant was diluted with 30 ml of a buffer containing 50 mM Hepes, pH 8.0, 100 mM NaCl, 3 mM MgCl 2 , 10 mM 2-mercaptoethanol, and 1.67 mM CaCl 2 (final concentration: 1 mM CaCl 2 , 0.4% sodium cholate). The diluted supernatant was then applied at a flow rate of 0.3 ml/min to a CaM-Sepharose column (10-ml bed volume, Pharmacia Biotech Inc.) equilibrated with the above described buffer. The column was washed with 50 ml of the above buffer. Proteins bound to the CaM-Sepharose column were then isocratically eluted with a buffer containing 50 mM Hepes, pH 8.0, 100 mM NaCl, 3 mM MgCl 2 , 10 mM 2-mercaptoethanol, 10 mM EDTA, and 0.4% sodium cholate. Fractions of 600 l were collected. 20 l of the indicated fractions were separated by 10% SDS-PAGE and immunoblotted using the G␤ 5 -specific antiserum CT215 as described (27).

Gel Mobility Shifting and Gel Overlay Assays
High affinity binding of G protein ␤ peptides to CaM was demonstrated by the gel mobility shifts of CaM in 12.5% nondenaturing polyacrylamide gels in the presence or absence of 4 M urea. Different concentrations of G␤ peptide were incubated with 1 M CaM at room temperature for 30 min. To determine the Ca 2ϩ dependence and binding stoichiometry, gels were run in the presence of either 0.5 mM CaCl 2 or 2 mM EGTA. Proteins were visualized by Coomassie Brilliant Blue staining.

Fluorescence Measurements and Determination of Dissociation Constants
Fluorescence emission spectra were obtained using SLM 4800 spectrofluorimeter. Excitation was at 295 nm. Excitation and emission band-passes were both 10 nm. Total fluorescence was determined by integration of emission spectra. The fluorescence titration data was used to determine the dissociation constant as described previously (28). By plotting the fraction of bound peptide as a function of free CaM concentrations, we obtained the peptide-CaM titration curve. The dissociation constant for the peptide-CaM binding was obtained from the curve fitting.

Site-specific Mutagenesis
Mutations in the putative CaM-binding site of G␤ 1 were generated by polymerase chain reaction with the high fidelity DNA polymerase, Pfu (Stratagene). The mutations were confirmed by DNA sequencing done in the DNA sequencing facility at Caltech. The ␤ constructs were subcloned into pCDNA3.1 vector (Invitrogen, San Diego, CA).

In Vitro Translation and Binding Assays
Transcription and translation were carried out using a rabbit reticulocyte lysate system from Promega at 30°C for 90 min with amino acid mixtures without methionine. 5 Ci of [ 35 S]methionine (Ͼ1000 Ci/ mmol, DuPont NEN) was added to the reaction mixture to monitor the synthesis of new proteins. 1-2 g of cDNAs were used in each translation. 2.5 l aliquots of the translation mixture were separated on 12% SDS-PAGE. The newly synthesized proteins were identified by autoradiography. After in vitro translation, the protein mixture was incubated with CaM-agarose beads in the presence of Ca 2ϩ for 30 min at room temperature. Then the CaM-agarose beads were washed extensively to get rid of nonspecific binding. Proteins bound to the beads were eluted by a buffer containing 10 mM EGTA. The eluted G␤ proteins were separated by SDS-PAGE and visualized by autoradiography.

ADP-ribosylation of G␣o Subunits by PTX
Pertussis toxin-catalyzed ADP-ribosylation of G␣o was performed as described previously (29). Briefly, 0.1 g of recombinant G␣o was mixed with 0.1 g of purified rat brain ␤␥ subunits in the absence or presence of Ca 2ϩ -CaM and incubated for 10 min at room temperature before the addition of the reaction mixture (20 mM Tris-HCl, pH 8.0, 1 mM EDTA, 2 mM MgCl 2 , 2 mM dithiothreitol, 0.5 M 32 P-labeled NAD (20,000 cpm/pmol), and 10 g/ml pertussis toxin). Reactions were incubated at room temperature for 30 min and terminated by adding 5 ϫ SDS-PAGE sample buffer. Samples were resolved on SDS-polyacrylamide gels and stained with Coomassie Blue. Labeling of proteins was detected by autoradiography.

Enzymatic Activity Assays of CaM Kinase II
The Ca 2ϩ -CaM-dependent activity of CaM kinase II was assayed essentially as described previously (30). Briefly, different concentrations of G␤ peptide were preincubated with Ca 2ϩ -CaM for 30 min at room temperature. The phosphorylation reactions contained 50 mM Hepes buffer, pH 7.5, 10 mM magnesium acetate, 1 mM CaCl 2 , 0.1 mM [␥-32 P]ATP, 1 M CaM, 5-10 g of CaM kinase II peptide substrate (Chemicon International, Inc., Temecula, CA), and purified rat brain CaM kinase II (Calbiochem). All assays were initiated by the addition of kinase.

COS-7 Cell Transfection and Phosphoinositol Phospholipase C Activity Assays
In Vivo Transfection Assay-cDNAs encoding PLC-␤2, G␤␥, and CaM were cotransfected into COS-7 cells with LipofectAMINE (Life Technologies, Inc.). Then the cells were labeled with 10 Ci/ml myo- [2-3 H]inositol the following day. 48 h after transfection, the activity of PLC-␤2 was assayed by determining the levels of inositol phosphates as described previously (31,37). In

RESULTS
Binding of G␤␥ Subunits to Ca 2ϩ -CaM-To demonstrate the direct interaction of G␤␥ subunits with CaM, purified bovine brain ␤␥ subunits were incubated with CaM-agarose gel in the presence of 1 mM CaCl 2 . After extensively washing with Ca 2ϩcontaining buffer, the bound ␤␥ subunits were eluted by buffer containing EGTA. As shown in Fig. 1A, brain ␤␥ binds to CaM, and this binding is Ca 2ϩ -dependent. Direct binding was also observed in a gel overlay assay with 125 I-labeled CaM (data not shown). There are five different G protein ␤ subunits that have been characterized in mammalian systems. Among them, G␤ 5 is the most diverse form and is expressed predominately in neuronal cells (27). A splice variant of G␤ 5 , ␤ 5 L, is found only in rod outer segment membrane. To examine whether CaM binding is a common feature in different ␤␥ subunits, G␤ 5 L␥ extracts were obtained from bovine retina and passed through a CaM-agarose fast protein liquid chromatography column in the presence of Ca 2ϩ . Proteins bound to the CaM-agarose column were eluted with buffer containing 10 mM EDTA and detected by specific antibodies against the G␤ 5 subunit. Like other ␤␥ subunits, ␤ 5 L␥ also binds to Ca 2ϩ -CaM (Fig. 1B), suggesting that the CaM-binding domain is conserved in all G␤␥ subunits.
Identification of a CaM-binding Domain of G␤␥ Subunits-To identify the CaM-binding domain on G␤␥ subunits, the amino acid sequences of ␤␥ were compared with those of known high affinity CaM-binding proteins. We found that the N terminus of all G␤ subunits contains a putative CaM-binding domain that exhibits the characteristics typical of CaM-binding peptides (32). Fig. 2 shows the alignment of the N-terminal domain sequences of G␤ subunits with well known CaM-binding domains of rat olfactory cyclic nucleotide-gated channel (RAT OCNC), skeletal muscle myosin light chain kinase (SK-MLCK), Ca 2ϩ pump, calcineurin, Ras-like GTPase Kir/Gem, murine inducible nitric oxide synthase (iNOS), and CaM kinase II. To demonstrate that the small domain from amino acid residues Val-40 -Trp-63 is indeed the CaM-binding domain of G␤ subunits, synthetic peptides representing the putative CaM-binding domain were prepared and tested for their abilities to bind CaM in the presence of Ca 2ϩ using both gel mobility shifting assay and tryptophan fluorescence assays.
Binding of peptides to CaM was first assayed by nondenaturing polyacrylamide gel shifting assay. Depending on the charges and hydrophobicity of the peptide, high affinity binding of the peptide to CaM was detected as a complex band with increased or decreased mobility compared with the unbound CaM band. In the presence of Ca 2ϩ , the mobility of CaM was decreased by the peptide corresponding to G␤ residues 40 -63 (Fig. 3A). Several ratios of the peptide to CaM were used. In the absence of peptide, there is a single, fast-moving band reflecting pure Ca 2ϩ -CaM (Fig. 3A, CaM alone). At ratios of 0.25 and 0.75 of peptide to CaM, two bands were visible on the gel, the fast-moving CaM band and the lower mobility peptide-Ca 2ϩ -CaM complex. At a 1:1 molar ratio of peptide to CaM, all of the CaM was gel-shifted, and the intensity of the peptide-Ca 2ϩ -CaM band was increased as the result of complex formation (Fig. 3A). At still higher molar ratios, no new band was detected on the gel nor did the peptide-CaM complex band change in intensity, indicating a 1:1 binding stoichiometry of peptide to CaM and the absence of multivalent peptide-CaM complexes (Fig. 3A). Similar results were obtained when gel shift assays were performed in the presence of 4 M urea. However, in the presence of Ca 2ϩ chelator, EGTA, only the pure CaM bands were detected on the gel and no peptide-CaM complex band was visible (Fig. 3B), indicating the complex formation is Ca 2ϩ -dependent. Peptides from other regions of the G␤ subunit had no effect on the mobility of CaM (data not shown).
Since the synthetic peptide contains a tryptophan residue, whereas CaM contains none, binding of G␤ peptide to CaM was directly studied by monitoring the changes in tryptophan fluorescence (Fig. 4). When the peptide was excited at 295 nm, it exhibited an intrinsic tryptophan fluorescence emission peak  . 1. Binding of G protein ␤␥ subunits to Ca 2؉ -CaM. A, interaction of rat brain G␤␥ subunits with Ca 2ϩ -CaM. Purified rat brain ␤␥ subunits were incubated with CaM-agarose in the presence of 1 mM CaCl 2 for 30 min at room temperature. Then, the CaM-agarose complexes were washed extensively with phosphate-buffered saline containing CaCl 2 . Proteins bound to CaM-agarose were eluted with buffer containing EGTA. Eluted G␤ protein was visualized by Western blot using specific antibodies against G␤ subunits. Con, control. B, binding of G␤ 5 L␥ subunits to Ca 2ϩ -CaM. G␤ 5 L␥ extracts were obtained from bovine retina and were then applied to a CaM-Sepharose fast protein liquid chromatography column (Pharmacia) equilibrated with Ca 2ϩcontaining buffer. Proteins bound to the CaM-Sepharose column were then eluted with an EDTA-containing buffer. 20 l of the indicated fractions were separated by 10% SDS-PAGE and immunoblotted using the G␤ 5 -specific antiserum CT215. at 353 nm. Addition of CaM in the presence of Ca 2ϩ not only caused a blue shift of the fluorescence peak but also increased the maximal fluorescence intensity (Fig. 4A). Since Ca 2ϩ alone has no effect on the emission spectrum, the observed change must have resulted from the binding of Ca 2ϩ -CaM to the peptide. The blue shift in fluorescence emission and the increase in total intensity indicate that the environment of the G␤ peptide became hydrophobic, presumably due to interactions with the hydrophobic domains of CaM upon formation of complex, further confirming the CaM binding property of the peptide. The changes in fluorescence property upon formation of the peptide-CaM complex were fully reversed by the addition of EGTA (Fig. 4A), indicating that the binding of CaM to the peptide is Ca 2ϩ -dependent. Titration of fixed amounts of peptide with different concentrations of CaM showed a saturation pattern. As shown in Fig. 4B, when a 0.1 M concentration of peptide was used, the fluorescence intensity increased almost linearly with increasing CaM concentrations and reached a plateau at 0.1 M CaM, confirming a 1:1 tight binding between the peptide and CaM. When the fraction of bound peptide was plotted as a function of free CaM concentration, the dissociation constant (K d ) obtained was ϳ40 nM (Fig. 4C), indicating high affinity binding to Ca 2ϩ -CaM. The actual binding affinity could be higher than 40 nM; however, the limitation of this fluorescence measurement does not allow us to test low peptide concentration.
Mutation Analysis of the Putative CaM-binding Domain of G␤ Subunits-To identify the key amino acids essential for CaM binding, two types of mutations were carried out in the putative CaM-binding domain by site-directed mutagenesis. One set contains a number of basic amino acid residues (Arg-42, Arg-48, Arg-49), and the other group involves hydrophobic residues (Ile-43, Leu-51, Leu-55). These specific amino acids were individually replaced by alanine residues. Mutant proteins were then compared with wild type protein for their ability to bind Ca 2ϩ -CaM. Both wild type and mutant cDNA were translated in vitro together with ␥2 subunit in rabbit reticulocyte lysates in the presence of 35 S-labeled methionine. Fig. 5A shows the in vitro translated wild type and mutant G␤ proteins separated by SDS-gel and detected by autoradiography. The amount of wild type protein and mutant proteins synthesized are similar under these assay conditions. The in vitro translated protein mixtures were incubated with CaMagarose beads to assess the CaM binding ability. After extensive washing with Ca 2ϩ -containing buffer (100 -200 bead volume), the bound protein was eluted with EGTA buffer. As shown in Fig. 5B, like the wild type G␤ protein, one of the mutants (R42A) retained its ability to bind CaM. However, the binding affinities of other mutants (I43A, L55A, R48A/R49A, and L51A) for CaM were reduced by varying degrees under the same assay conditions. For instance, mutations at residues Arg-48, Arg-49, and Leu-51 dramatically reduced the CaM binding ability of the protein (Fig. 5B), suggesting that these residues are critical in the interactions. To further probe the conformation and activity of mutant G␤ subunits, we analyzed the formation of ␤␥ dimers using in vitro translated proteins and coimmunoprecipitation with specific anti-G␥ 2 antibodies. The wild type G␤ and its mutant proteins can be precipitated by ␥ 2 antibodies, and there is little difference in their ability to interact with ␥ subunits (data not shown), indicating that mu-  1 M) was excited at 295 nm, the peptide exhibited a fluorescence emission spectrum with a maximal emission peak at 354 nm (peptide w/Ca 2ϩ ). In the presence of Ca 2ϩ , addition of CaM not only caused a blue shift of the fluorescence peak but also increased the total fluorescence intensity (peptide ϩ Ca 2ϩ -CaM). Addition of EGTA completely reversed these fluorescence changes (peptide ϩ Ca 2ϩ -CaM ϩ EGTA), indicating that the interactions between the G␤ peptide and CaM is Ca 2ϩ -dependent. B, the G␤ peptide was tant ␤ subunits could still fold into the native conformation and retain affinity for ␥ subunits.
Effects of CaM Binding on ADP-ribosylation of G␣o Subunits-To explore the possible physiological functions of CaM binding to G␤␥ subunits, we first examined the effects of CaM binding on ␤␥-dependent pertussis toxin-catalyzed ADP-ribosylation of G␣o subunit. PTX is a bacterial toxin that catalyzes the ribosylation of the C-terminal cysteine residues of G␣o, G␣i, and G␣t subunits. PTX modification blocks the interactions of G␣ subunits with receptors and thus blocks the ligandmediated signal transduction. The labeling of G␣o by PTX requires the G␣o␤␥ heterotrimer rather than the free G␣o subunit (33). Thus, ADP-ribosylation of ␣o can be used as a sensitive indicator of formation of ␣o␤␥ heterotrimers. Incubation of CaM with bovine brain G␤␥ subunits in the presence of 0.1 mM Ca 2ϩ inhibited the ribosylation reaction, as shown in Fig. 6. The reduction of labeling is concentration-dependent. At equal molar concentrations of CaM and ␤␥ subunits, the PTXcatalyzed ribosylation of G␣o was almost completely blocked by Ca 2ϩ -CaM. Since Ca 2ϩ alone had no effect on the reaction, the simplest interpretation of this observation is that Ca 2ϩ -CaM formed a complex with G␤␥ subunits and this complex lost its ability to interact with G␣o subunit; therefore, CaM inhibited the ␤␥-dependent ADP-ribosylation of G␣o.
Binding of CaM to G␤␥ Has Little Effect on ␤␥-activated PLC-␤2 Activity-To further investigate the nature of CaM binding on G␤␥ subunits, we examined its effects on ␤␥-stimulated PLC-␤2 activity. As shown in Fig. 7A, CaM has little effect on PLC-␤2 activity stimulated by ␤␥ subunits in an in vitro reconstituted system. To confirm the observations obtained in the reconstitution experiments, cDNAs encoding PLC-␤2, G␤␥, and CaM were transfected into COS-7 cells. Coexpression of PLC-␤2 and ␤␥ subunits increased inositol 1,4,5-trisphosphate release 3-5 fold (Fig. 7B, column 3). However, cotransfection of CaM expression plasmids had little ef-fect on the ␤␥-stimulated PLC-␤2 activity (Fig. 7B, column 4). These results indicated that in the G␤ subunit, the domain responsible for PLC-␤2 activation is distinct from the domain binding to CaM, suggesting that different domains of the ␤ subunit are involved in interactions with different effector proteins.
Inhibition of Ca 2ϩ -CaM-dependent Enzyme, CaM Kinase II, by G␤␥ Subunits and Its Peptide-The putative CaM-binding domain of G␤ subunits was further characterized by using the synthetic peptide derived from G␤ to inhibit Ca 2ϩ -CaM-dependent activation of CaM kinase II. Concentration-dependent effects of the peptide on CaM kinase II activity were examined. As shown in Fig. 8A, the peptide was a potent inhibitor of CaM kinase II activity. At a 1:1 binding ratio of G␤ peptide to CaM, the peptide totally inhibited the Ca 2ϩ -CaM-dependent activation of CaM kinase II, suggesting that the binding affinity of G␤ peptide is comparable to the affinity of CaM kinase II toward Ca 2ϩ -CaM and should be in the nM range. Other peptides from the N-terminal regions had no effect on the Ca 2ϩ -CaM-stimulated enzymatic activity (data not shown). To examine whether G␤␥ subunits can also inhibit the Ca 2ϩ -CaMdependent CaM kinase II activity, Ca 2ϩ -CaM was incubated with brain G␤␥ subunits for 30 min at room temperature. Then the Ca 2ϩ -CaM-stimulated CaM kinase II activity was assayed. Fig. 8B shows that brain G␤␥ subunits inhibited 70 -80% of Ca 2ϩ -CaM-stimulated CaM kinase II activity, indicating that ␤␥ can competitively bind to Ca 2ϩ -CaM. DISCUSSION Previous studies have demonstrated that G protein subunits can act as potent inhibitors of the Ca 2ϩ -CaM-stimulated phosphodiesterase activity (23,24), probably through interactions of G␤␥ with CaM. In this report, we show direct binding of G␤␥ subunits with CaM. This CaM-binding property of G␤ is Ca 2ϩdependent and conserved in known G␤ subunits, including the most diverse ␤5 subunit. We also identified and characterized the CaM-binding domain in ␤␥ subunits using three different methods. In the gel mobility shifting assays and tryptophan fluorescence assay, a conserved 25-amino acid peptide in the N-terminal region of G␤ subunits was found to interact with CaM. By modifying specific amino acids in this region, we identified some key residues (Arg-48, Arg-49, Leu-51, Ile-43, Leu-55) that play an important role in the binding of ␤␥ to CaM.
The interaction between G␤␥ subunits and CaM is Ca 2ϩ -dependent. Ca 2ϩ is an important intracellular messenger in many cellular functions including cell growth and development (34,35). Many seven-transmembrane receptors activate the G protein subunits, which in turn activate PLC-␤, resulting in the production of inositol 1,4,5-trisphosphate and diacyglycerol from PtdInsP 2 . Inositol 1,4,5-trisphosphate acts as an intracellular second messenger by binding to the inositol 1,4,5trisphosphate receptors in the endoplasmic reticular mem- brane, triggering the release of Ca 2ϩ from the endoplasmic reticulum and therefore increasing the intracellular Ca 2ϩ concentration. Ca 2ϩ is also believed to organize and stabilize CaM domain structure in a conformational state that can bind target proteins (32,36). The calmodulin concentration in some cells, e.g. neuronal cells, is in the order of micromolar. Calcium release can thus convert 10 -50% of these molecules to the Ca 2ϩ -bound form. The affinity of G␤␥ for Ca 2ϩ -CaM is sufficiently high so that under these conditions much of the free ␤␥ in the cells should be in the CaM-bound form. Therefore, an increase in intracellular Ca 2ϩ concentration could selectively regulate the interactions of G␤␥ with a number of other proteins through binding with CaM.
The interaction of G␤␥ with Ca 2ϩ -CaM could play an important role in the cross-talk mechanism between different G protein pathways (37). Ca 2ϩ -CaM has been shown to regulate the formation and hydrolysis of cAMP. In the G␣s-coupled pathway, both adenylyl cyclases and phosphodiesterases are Ca 2ϩ -CaM-dependent, which could serve as the convergence point for Ca 2ϩ -dependent and G␣s-dependent stimuli. The primary structure of the identified CaM-binding domain of G␤␥ shows features similar to some other CaM-binding proteins and inhibitors (21, 28, 38 -45). For instance, the identified ␤ CaM-binding domain contains a high percentage of hydrophobic residues and an excess of positively charged residues, a property found in all CaM-binding peptides. It is believed that the basic residues contribute to CaM binding via electrostatic interactions with acidic residues in CaM, whereas the hydrophobic amino acids seem to play a more important role in CaM binding through interactions with the hydrophobic patches of the globular domains of CaM (38,39). From x-ray structural analysis, 80% of all contacts are van der Waals interactions, and the interaction appears to involve two key hydrophobic amino acids separated by eight residues, although flexibility does exist in the interaction because of the flexible central helix of CaM (36). CaM contains two globular domains, and these two domains are connected by a long alpha helical segment (32,36). When different CaM binding targets are recognized, this long central segment allows different relative positioning of the two lobes of CaM. Therefore, although the secondary structure of the G␤ CaM-binding domain is not a typical amphipathic ␣ helix (46), CaM could still position itself to bind to the G␤ protein, possibly by stabilizing a change in the conformation of the N-terminal domain of G␤ protein. On the other hand, a number of CaM-binding proteins have been identified and shown to contain sites that are not typical amphipathic helices but consist of basic amino acids interspersed with nonpolar amino acids (52)(53)(54).
The binding of CaM to G␤␥ interfered with the formation of G␣␤␥ trimers as assayed by the inhibition of PTX-catalyzed ADP-ribosylation of G␣o. Based on the secondary structure of the ␤ subunit, the N-terminal region of the ␤ subunit is in close proximity to the N-terminal portion of G␣ subunit (47)(48)(49). It has been reported that effector activation by the ␤␥ subunits is blocked upon the addition of the G␣ subunit presumably by heterotrimerization (1,13). Thus, by interacting with the Nterminal domain of ␤ subunit, the Ca 2ϩ -CaM complex could affect the heterotrimer formation of G␣␤␥.
Interaction of CaM with ␤␥ has little effect on the G protein ␤␥ subunit-activated PLC-␤2 activity. These results support the notion that interaction of different ␤␥-responsive effectors is mediated by distinct domains of G␤␥ (50,51). By using a series of chimeras between Dictyostelium and mammalian ␤ subunits, a small C-terminal segment of G␤ was identified as responsible for the activation of PLC-␤2 (50). The CaM-binding domain of G␤ identified in this report is located in the Nterminal region. This region of G␤ was suggested to play a role in the interaction of adenylyl cyclase type 2, the muscarinic receptor-gated atrial inwardly rectifying potassium channel (GIRK1), and in the activation of mitogen-activated protein kinase pathways (50,51). It will be of great interest to determine whether binding of CaM to G␤␥ can affect the activation of the potassium channels (GIRK) and mitogen-activated protein kinase pathways.