Identification of critical residues controlling G protein-gated inwardly rectifying K(+) channel activity through interactions with the beta gamma subunits of G proteins.

G protein-sensitive inwardly rectifying potassium (GIRK) channels are activated through direct interactions of their cytoplasmic N- and C-terminal domains with the beta gamma subunits of G proteins. By using a combination of biochemical and electrophysiological approaches, we identified minimal N- and C-terminal G beta gamma -binding domains responsible for stimulation of GIRK4 channel activity. Within these domains one N-terminal residue, His-64, and one C-terminal residue, Leu-268, proved critical for G beta gamma-mediated GIRK4 activity. Moreover, mutations at these GIRK4 sites reduced significantly binding of the channel domains to G beta gamma . The corresponding residues in GIRK1 also showed a critical involvement in G beta gamma sensitivity. In GIRK4/GIRK1 heteromers the GIRK4 His-64 and Leu-268 residues showed greater contributions to G beta zeta sensitivity than did the corresponding GIRK1 His-57 and Leu-262 residues. These results identify functionally important channel interaction sites with the beta gamma subunits of G proteins, critical for channel activity.

Four mammalian G protein-sensitive Inwardly Rectifying K ϩ (GIRK1-4 or Kir3.1-3.4) channels have been identified thus far (5)(6)(7)(8)(9). Both the N-(ϳ90 amino acids) and C-terminal (ϳ240 amino acids or more) ends of these channels are cytoplasmic (5). Two transmembrane domains (ϳ90 amino acids) surround the P-region that harbors the potassium ion "selectivity filter" (10). GIRK channels can function as highly active heteromultimers (pairing of GIRK1 with any of the other subfamily members) or low to moderately active homomultimers ((GIRK2-4) reviewed in Refs. 11 and 12). Mutations at a specific position within the P-region of these channels (i.e. GIRK4-S143T or GIRK1-F137S) greatly enhance the activity of homomultimers (13,14). Use of these highly active point mutants simplifies the experimental design of structure-function studies and allows assessment of the relative contributions of each of the two subunits in the heteromultimeric complex (14).
A number of studies have demonstrated direct binding of G␤␥ subunits to the full-length GIRK proteins (15) or to segments of channel subunits (16 -21). Although G␤␥ subunits can interact directly with both the N and C termini of GIRK1, interactions with the C terminus of this channel were shown to be the strongest (16,17). Moreover, the N terminus of GIRK1 but not the C terminus interacts with the G␣␤␥ heterotrimer (16,20). Functionally, both native and recombinant hetero-or homomultimeric GIRK channels are activated by the G␤␥ subunits (4,8,9,22). No qualitative difference in the G␤␥ sensitivity of active homomultimeric mutants versus heteromultimeric channels was observed (14). In contrast, the inwardly rectifying K ϩ channel IRK1 (5) is G␤␥-insensitive (23,24), despite its high degree of similarity in protein sequence to the four members of the GIRK subfamily. By using a chimeric approach between GIRK4 and IRK1, a critical C-terminal Leu residue was identified (GIRK4-L339 or GIRK1-L333), mutation of which abolished the G␤␥-mediated agonist-induced GIRK channel activation but not the G␤␥-mediated agonist-independent (basal) GIRK channel activity (24). This result suggested that distinct channel sites may be involved in basal versus agonist-induced channel activation through the G␤␥ subunits.
In the present study we sought to identify additional specific residues of the GIRK subunits, critical for G␤␥-mediated activity. Two amino acid residues, the N-terminal GIRK4(H64) and the C-terminal GIRK4(L268), are shown to be critically involved in G␤␥-mediated currents. Moreover, mutation of each of these GIRK4 residues is shown to reduce binding to the G␤␥ subunits. The corresponding GIRK1 residues are also found to be involved in the G␤␥ sensitivity of GIRK1 subunits (GIRK1(H57) and GIRK1(L262)). In contrast to the mutation of GIRK4(L339) or GIRK1(L333), which selectively controlled ag-onist-induced but not agonist-independent G␤␥-mediated currents (24), the two sites reported here are found to abolish all G␤␥-mediated currents.

EXPERIMENTAL PROCEDURES
Human homologs of GIRK1 and GIRK4 (GenBank TM accession numbers U39195 and U39196 (9)) or their point-mutated active counterparts (GIRK1-F137S or GIRK1* and GIRK4-S143T or GIRK4*), subcloned in the pGEMHE plasmid vector (25), were used as described previously (9,14,24). The chimeric cDNA constructs were produced by splicing by overlapping extension PCR (26). PCRs, using Vent DNA polymerase, were performed for only 15 cycles to avoid errors. Point mutations were generated using the Quickchange site-directed mutagenesis kit (Stratagene). The sequence of all constructs was confirmed by automated DNA sequencing (Sequencing facility, Cornell, Ithaca, NY). The ␤ARK-PH construct (amino acids 452-689) (referred to as ␤ARK) was used to sink G␤␥ subunits away from the channel. For confocal studies GIRK1-GFP was used as described previously (27).
All constructs were linearized with Nhel, and cRNAs were transcribed in vitro using the "message machine" kit (Ambion). RNAs were electrophoresed on formaldehyde gels, and concentrations were estimated from two dilutions using RNA marker (Invitrogen) as a standard.
Xenopus oocytes were surgically extracted, dissociated, and defolliculated by collagenase treatment and microinjected with 50 nl of a water solution containing the desired cRNA. Unless otherwise indicated, we used the following approximate quantities: GIRK channel subunits, 1.0 ng per species; IRK1 channel, 0.25 ng; M 2 receptor, 1.0 ng; G␤ subunits, 1.0 ng; G␥ subunit, 1.0 ng; ␤ARK-PH, 1.0 ng. G␤ 2 or G␤ 1 with G␥ 2 were used in all G␤␥ coexpression experiments.
Oocytes were incubated for 3 days at 19°C. Whole oocyte currents were then measured by conventional two-microelectrode voltage clamp with a GeneClamp 500 amplifier (Axon Instruments). Agarose cushion microelectrodes were used with resistances between 0.1 to 1.0 megohms (28). Oocytes were constantly superfused with a high potassium solution containing (in mM) 91 KCl, 1 NaCl, 1 MgCl 2 , 5 KOH/HEPES (pH 7.4). To block or activate currents, the oocyte chamber was perfused with solutions of the same composition with 3 mM BaCl 2 or 5 M ACh. Typically oocytes were held at 0 mV (E K ), and currents were constantly monitored by 500-ms pulses to a command potential of Ϫ80 mV for 200 ms followed by a step to ϩ80 mV for another 200 ms, and the cycle was repeated every 2 s. Periodically, a protocol was applied with a command potential from Ϫ100 to ϩ100 mV with 10-mV increments. Current amplitudes were measured at the end of the 200-ms pulse at each potential. Control currents were evaluated in this manner 2-5 min after impaling the oocytes just before application of ACh. ACh-activated currents were evaluated at the peak of the response to ACh, and Ba 2ϩ -insensitive currents were evaluated once steady-state inhibition was achieved, 1-3 min after application of 3 mM Ba 2ϩ . Basal current is the difference between control and Ba 2ϩ -insensitive currents, and AChinduced (or ACh-sensitive) current is the difference between ACh-activated and control currents. Error bars in the figures represent S.E. Each experiment shown or described was performed on 3-5 oocytes of the same batch. A minimum of 2-3 batches of oocytes were tested for each experiment shown.
Single channel recording was performed, as described previously (29 -31). Single channel currents were filtered at 1 kHz with a 6-pole low-pass Bessel filter, sampled at 5 kHz. Single channel data were analyzed with pCLAMP8 software supplemented with some of our own programs. Base-line drifts were carefully adjusted with Clampfit8 before idealization. Only those records containing a few or no double openings were used for open time histogram fitting. Recordings were performed 2 days after injection.
Recombinant bovine G␤ 1 ␥ 2 subunits were purified from Sf9 cells infected with baculoviruses encoding for ␤ 1 , ␥ 2 , and His 6 -␣ i1 as described (32). cDNAs encoding fragments of GIRK4 or IRK1 and the PH domain of ␤ARK were generated by PCR and cloned in frame with the GST coding sequence in pGEX-4T-3 (Amersham Biosciences). Expression of fusion proteins was induced by 0.1 mM isopropyl-1-thio-␤-Dgalactopyranoside at 37°C for 2 h, and the fusion proteins were purified using glutathione 4B-Sepharose beads. The binding assay of G␤␥ to the fusion proteins was performed as described by Huang and colleagues (16). Briefly, 1.0 M GST fusion protein and 0.1 M G␤␥ were incubated in phosphate-buffered saline, 0.1% Lubrol, and glutathione-Sepharose beads at 4°C for 30 min. After washing three times with 200 volumes of phosphate-buffered saline with 0.01% Lubrol, the bound proteins were eluted from beads by heating in protein sample buffer at 70°C for 10 min and were then electrophoresed in a 12% SDS-PAGE. GST fusion proteins were visualized by Coomassie staining, and G␤ 1 was detected by immunoblotting using a G␤ antibody (Santa Cruz Biotechnology) and visualized with ECL (Amersham Biosciences). Densitometry was used to quantify the relative amounts of bound G␤␥.
Detection of GFP-tagged channels in oocytes has been described previously in detail (33). Briefly, 2-4 days after injections, Xenopus oocytes were fixed in 4% paraformaldehyde overnight at room temperature. Fixed oocytes were embedded in 3% agarose, and 50-m sections were cut using a Vibratome. The cut sections were mounted on coverslips and imaged using a Leica TCS confocal microscope. To compare fluorescence intensities between different oocyte sections, image acquisition parameters such as pinhole size, intensity, and offset were kept constant.

RESULTS
Minimal G␤␥-binding Domains in the N and C Termini of GIRK4 Channels-To identify minimal G␤␥-binding domains within the N and C termini of the GIRK4 channels, we constructed and purified GST fusion proteins and tested their binding to purified G␤␥ subunits. We tested G␤␥-binding of the full N and C termini of GIRK4 and of fragments harboring incremental deletions of ϳ25 amino acids, until G␤␥ binding was no longer detected. G␤␥ binding of GIRK4 channel fragments was compared with that of the following: (a) the PH domain of the ␤-adrenergic receptor kinase (␤ARK) (e.g. Ref. 24); (b) the corresponding full N-and C-terminal fragments of the IRK1 (or Kir2.1) channel, which is related to GIRK4 but is G␤␥-insensitive (e.g. Ref. 24); and (c) the GST protein alone, which was common to all purified fusion proteins. Fig. 1A shows G␤␥ binding for the full N terminus, GIRK4-(1-92) (denoted as G4-(1-92)), and for a minimal N-terminal G␤␥-binding fragment G4-(41-92). Fig. 1B shows G␤␥ binding for the full C terminus G4-(184 -419), and no binding for fragments smaller than G4-(253-322). In fact, maximal G␤␥ binding was obtained FIG. 1. Identification of minimal G␤␥-binding regions on the N and C termini of GIRK4. A, GST fusion constructs were made with fragments of the N terminus of GIRK4. The full GST and GST-IRK1-(1-86) were used as negative controls, whereas GST-␤ARK-PH was used as a positive control for G␤␥ binding. Proteins were purified using glutathione 4B-Sepharose beads and were detected by Coomassie staining (top). Purified GST fusion proteins were incubated with G␤␥ and glutathione-Sepharose beads. Following wash, the bound proteins were released from the beads by heating in protein sample buffer and were separated by SDS-PAGE. G␤ was detected by immunoblotting with anti-G␤ antibody. The position of G␤ is indicated by the arrow. The representative gel shown is one of two experiments performed. B, GST fusion proteins were made with fragments of the C terminus of GIRK4. The proteins were purified and detected, and G␤␥ interactions were determined as shown in A. The representative gel shown is one of three experiments performed.
Regions within Minimal G␤␥-binding Domains Responsible for Stimulation of GIRK4 Channel Activity-To identify regions within the two minimal G␤␥-binding domains of GIRK4 involved in stimulating channel activity, we constructed chimeras between the GIRK4 and IRK1 channels. GIRK4 regions of ϳ25 amino acids, within the minimal N-and C-terminal G␤␥-binding domains, were replaced by the corresponding IRK1 regions. These chimeras were constructed in the background of GIRK4(S143T) (referred to as G4*), which has been shown previously to form functional homomeric channels (14). Heterologous coexpression of control and mutant channels with muscarinic M 2 receptors was carried out in Xenopus laevis oocytes. Measurement of K ϩ currents was performed using two-electrode voltage clamp. Fig. 2A shows that the N-terminal chimera G4*(IRK1 S35-V56 ), which replaced the GIRK4 region between 40 and 61 with the corresponding IRK1 region between Ser-35 and Val-56, showed normal agonist-independent (basal) and agonist-induced G4* currents. Similarly, the Cterminal chimera G4*(IRK1 V270-D291 ) replacing the GIRK4 region between 277 and 298 also showed normal basal and agonist-induced currents. These results suggested that neither the 40 -61 nor the 277-298 GIRK4 regions possessed any key residues responsible for the differences in G␤␥ sensitivity between GIRK4 and IRK1. Chimera G4*(IRK1 L316-Y341 ) showed basal but not agonist-induced currents. This result agreed with previous work showing the corresponding GIRK region (e.g. G4-(323-348)), and in particular a Leu residue (G4 Leu339 or G1 Leu333 ) was critical in G␤␥-mediated agonist-induced activa-FIG. 2. Chimeras between GIRK4* and IRK1 channels reveal regions important in activation of K ؉ currents. A, left, schematic illustration of chimeric constructs between GIRK4* and IRK1. Specific segments of IRK1 (black regions) were used to replace corresponding segments of GIRK4* (white regions) within the minimal binding domains (see Fig. 1). Each chimera was named according to the IRK1 segment (identified by the first and last IRK1 amino acid of the segment), which replaced the corresponding segment of GIRK4*. Note that the corresponding numbers, defining the limits of the grafted segments, are shifted (6 -7 numbers) as determined by alignment of the two sequences. Right, basal and agonist-induced currents at Ϫ80 mV of GIRK4*, IRK1, and the seven chimeras. Although two of the chimeric channels showed similar function to the wild-type GIRK4*, four other chimeras were non-functional. Chimera G4*(IRK1 L316-Y341 ), which displays normal basal currents but lacks ACh-induced currents, has been characterized previously (24). Summary data are from one representative batch of oocytes, n ϭ 3-5 oocytes tested per channel. B, confocal images from oocyte sections expressing GFP-tagged GIRK1 subunits (G1-GFP) alone or in combination with G4* and G4* chimeras. When G1-GFP was expressed alone it showed a predominantly intracellular distribution. Coexpression with G4* induced translocation of G1-GFP to the membrane as detected by the fluorescence signal in the membrane. Representative images are from one of three batches of oocytes tested. tion as well as for maximal binding to G␤␥ (Ref. 24; also compare in Fig. 1B G␤␥ binding to GST-G4-(253-322) versus GST-G4-(253-348)). In contrast, chimeras G4*(IRK1 Q57-V86 ), G4*(IRK1 D246-I269 ), and G4*(IRK1 F292-Y315 ) replacing, respectively, the G4* regions 62-91, 253-276, and 299 -322 were not functional. GIRK1 channels tagged with the green fluorescence protein (G1-GFP) showed minimal distribution on the cell surface (Fig. 2B). Because GIRK4 or GIRK4* localize efficiently to the plasma membrane (12), untagged GIRK4* was coexpressed with G1-GFP causing translocation of fluorescence signal to the cell surface. This indicated association of G4* with G1-GFP and successful targeting of G1-GFP to the plasma membrane (Fig.  2B). To test whether the non-functional chimeras expressed proteins capable of associating with GIRK1 and localizing to the cell surface, the three non-functional G4* chimeras were coexpressed with G1-GFP. Indeed, the non-functional G4* chimeras were able to associate with and localize G1-GFP to the cell surface.
Single Residue Differences between GIRK4 and IRK1 Critical for Channel Function-Within the N-terminal GIRK4-(62-91) region that proved critical for the functional integrity of G4*, 13 residues are different from the corresponding amino acids in IRK1. Mutation of each G4* N-terminal residue within this region to the corresponding IRK1 amino acid revealed three differences critical for channel function. Each of the G4*(H64F), G4*(Y71Q), and G4*(L84I) mutants abolished completely channel activity (Fig. 3A). Similarly, within the C-terminal GIRK4-(253-276) and GIRK4-(299 -322) regions that produced non-functional chimeras, there are 11 amino acid differences from IRK1. Two of the 11 mutants, G4*(L268I) and G4*(A318C), also abolished completely K ϩ currents. The effects of these mutations were more severe than those in the previously identified mutant G4*(L339E), which resulted in loss of agonist-induced currents but left intact G␤␥-mediated basal currents (24).
Rescue of Non-functional GIRK4 Mutants Allows Identification of G␤␥-sensitive Residues-Inwardly rectifying K ϩ channels, including GIRK subfamily members, depend on phosphatidylinositol-bisphosphate (PIP 2 ) for their activity. In fact, gating molecules, such as the G␤␥ subunits or Na ϩ ions, have been shown to strengthen channel-PIP 2 interactions (34 -37). It has been shown previously that amino acids in the GIRK4 region (214 -252) and the corresponding IRK1 region (207-245) are critical determinants of interactions of these channels with PIP 2 . One mutant, G4*(I229L), exchanging a GIRK4 residue for the corresponding one in IRK1, strengthened channel PIP 2 interactions sufficiently enough to allow gating of the mutant channel by PIP 2 alone in the absence of gating molecules, such as G␤␥ or Na ϩ ions. Yet this mutation preserved G␤␥ stimulation of K ϩ currents. Each of the three N-terminal and two C-terminal single point mutants that showed no K ϩ currents was paired with the I229L mutation, and each of the double  Fig. 2A, distinct amino acids in these regions of GIRK4* were replaced with the corresponding IRK1 residues. One representative batch of 3-5 oocytes is shown for each series of point mutants. A, three amino acids, His-64, Tyr-71, and Leu-84, in GIRK4* N terminus were found to be critical for GIRK4* function. Currents were recorded at Ϫ80 mV. B, in the GIRK4* C terminus two amino acids, Leu-268 and Ala-318, were also found to be critical for GIRK4* function. Currents were again recorded at Ϫ80 mV. mutant channels was expressed in Xenopus oocytes and tested for activity. Fig. 4A shows that all three N-terminal mutants G4*(I229L,H64F), G4*(I229L,Y71Q), and G4*(I229L,L84I) showed convincing inwardly rectifying basal K ϩ currents. In fact, all of these mutants that now displayed basal currents also showed intact agonist-induced currents (Fig. 4, A right, and C right). Because activity of each of the "functionally lethal" point mutants H64F, Y71Q, and L84I was rescued by the I229L mutation, we proceeded to test whether G␤␥ subunits were contributors to the rescued basal K ϩ currents. We have shown previously that molecules that bind G␤␥ can serve as a "sink" to sequester G␤␥ subunits away from the channel, inhibiting basal channel activity (24). Fig. 4B shows ␤ARK-PH coexpression with each of the rescued N-terminal double mutants and with the control single mutant, G4*(I229L). Of the three N-terminal mutants, only G4*(I229L, H64F) was insensitive to ␤ARK, indicating that with the H64F mutation the G␤␥ subunits were no longer contributing to the rescued basal currents. Fig. 4C shows a similar rescue approach for the two C-terminal "functionally lethal" mutants. Again, the rescued G4*(I229L,L268I) and G4*(I229L,A318C) currents were convincingly above background (ϳ4 A versus Ͻ0.5 A) and clearly inwardly rectifying (Fig. 4C). Again, these rescued C-terminal mutants showed intact agonist-induced currents. Of these two C-terminal rescued mutants, the L268I currents were not significantly affected by ␤ARK coexpression (Fig. 4D). Thus, of the five non-functional mutants, one in the N terminus, G4*(H64F), and one in the C terminus, G4*(L268I), did not show inhibition by the G␤␥ sink, ␤ARK-PH, suggesting that the His-64 and Leu-268 residues are critical for G␤␥ sensitivity.
Effects of Mutations of the His-57 and Leu-262 Residues on the G␤␥ Sensitivity of GIRK1-The two identified G␤␥-sensitive residues in GIRK4 are identical in GIRK1. We proceeded to test their importance in GIRK1 in the background of GIRK1(F137S) (referred to as G1*), which has been shown previously to yield functional homomers (13,14). Fig. 5A shows that G1*(L262I), the mutant corresponding to G4*(L268I), greatly reduced but did not abolish basal G1* currents. ␤ARK-PH inhibited the remaining basal currents, suggesting that in GIRK1 additional residues may contribute to G␤␥-mediated basal currents. In contrast, the N-terminal GIRK1* mutant G1*(H57F) behaved as the corresponding G4*(H64F), abolishing completely K ϩ currents. As expected, the double mutant G1*(H57F,L262I), containing the non-functional H57F mutation, abolished all G1* currents.
Relative Contributions of the His-64/57 and Leu-268/262 Residues of the GIRK4/1 Subunits to the G␤␥ Sensitivity of the GIRK4/1 Heteromer-To examine the relative contributions of the GIRK1 and GIRK4 subunits to the heteromeric GIRK1/ GIRK4 channels, we expressed each double mutant subunit with the corresponding wild-type subunit and recorded basal currents. Fig. 5B shows that the G1/G4(H64F,L268I) heteromeric channels showed no activity, in contrast to the G1(H57F,L262I)/G4 channels that exhibited small inwardly rectifying K ϩ currents. Heteromeric assembly of mutant and wild-type subunits did contribute to these small currents (see below). When both mutant subunits were coexpressed, G1(H57F,L262I) and G4(H64F,L268I), no K ϩ currents could be measured. These results suggest a dominant role for the two identified residues in GIRK4 over GIRK1 subunits in a heteromeric channel. We further explored the relative contributions of each of the N-and C-terminal mutations to the heteromeric channels. Fig. 5C shows expression of each of the N-and C-terminal mutants alone or in combination with their wildtype counterpart. The G4(H64F) was the only single mutant that showed no currents when expressed with its G1 wild-type counterpart. This result suggests again a dominant role for the N-terminal G␤␥-sensitive site of GIRK4. The G4(L268I)/G1 showed greatly reduced currents. Of the G1 mutants, the G1(H57F)/G4 heteromers also showed greatly attenuated K ϩ currents, whereas the G1(L262I)/G4 showed attenuated but larger heteromeric currents than any of the other mutant/wildtype combinations. To test whether currents resulting from the G1(mutant)/G4(wild-type) combinations reflected activity from heteromeric channels or homomeric wild-type G4 channels, we examined the single channel characteristics of each of the heteromeric combinations exhibiting basal currents.
As shown in Fig. 6, G4 homomeric activity is characterized by short lived and not clearly resolved transitions at any particular level (see also Ref. 9). In contrast, G1/G4 heteromeric activity exhibits a clear peak in the single channel conductance amplitude histogram. Each of the G1 mutant combinations with G4 wild-type, G1(H57F)/G4, G1(L262I)/G4, and G1(H57F, L262I)/G4, showed clear peaks in their amplitude histograms, indicating that functional heteromeric channels were formed. Similarly, the only G4 mutant exhibiting channel activity when paired with the wild-type G1 (i.e. G1/G4(L268I)) also showed clear heteromeric channel amplitude transitions (not shown). Open time histograms could be fitted by two exponentials 1 and 2 yielding a mean open time (Fig. 6B). The predominant change in the open time kinetics due to heteromerization occurred on the second open time component, 2 . Thus, all heteromeric combinations that displayed activity indeed reflected contributions from heteromeric channel assemblies. These results, taken together with the results of Fig. 5C, suggest that the G4(H64F) mutant acted as a dominant negative in the heteromeric G1/G4(H64F) channel, whereas the G4(L268I) mutant greatly reduced heteromeric G1/G4(L268I) currents. Because the G4(L268I) mutant did not but the G4(H64F) mutant did abolish basal currents when coexpressed with wild-type GIRK1 channels, it seems that the role of His-64 is more critical than Leu-268 in controlling heteromeric channel activity. G1(H57F) also greatly reduced but did not abolish heteromeric G1(H57F)/G4 currents. The G1(L262I) caused even less inhibition of heteromeric G1(L262I)/G4 currents, consistent with its effects in G1* homomers (Fig. 5A). Overall, these results underscore the dominant role of GIRK4 in heteromeric G␤␥ sensitivity through the two identified N-and C-terminal residues.
N-and C-terminal Point Mutations Reduce GIRK4 Channel Binding to the G␤␥ Subunits-To test whether the H64F and L268I mutations affected channel binding to the G␤␥ subunits, we constructed and purified GST fusion proteins of the minimal G␤␥-binding domains with or without the point mutations of interest, and then we tested their binding to purified G␤␥ subunits. Both the H64F and the L268I mutations significantly reduced binding to G␤␥ of the N-and C-terminal binding fragments, respectively (Fig. 7). DISCUSSION A number of attempts have been made prior to this study to localize the N-and C-terminal channel regions responsible for direct binding with the G␤␥ subunits (16 -18, 38). Specifically, Huang et al. (17), using a deletion mutagenesis approach, have localized the G␤␥-binding region of GIRK1 to the N-terminal domain including amino acids 34 -86 and the C terminus involving two separate fragments, 318 -374 and 390 -462. Fragments corresponding to GIRK1-(34 -86) and GIRK1-(318 -374) from GIRK-(2-4) also displayed G␤␥ binding. Kunkel and Peralta (18), with a combination of chimeras and deletion mutations, reported the GIRK1-(290 -356) region to be important in interactions with G␤␥. Krapivinsky et al. (38) used peptides derived from GIRK1 or GIRK4 amino acid sequences to compete direct G␤␥ binding to atrial purified K ϩ channels. Effective peptides included the GIRK1-(364 -383) region and the GIRK4-(209 -225) and -(226 -245) regions. It is difficult to discern commonly identified channel regions of interactions with G␤␥ from these studies.
Our results indicate that in GIRK4 minimal G␤␥-binding domains involve the N-terminal 41-92 residues and the Cterminal 253-348 residues. The N-terminal region GIRK4-(41-92) indeed corresponds closely to the GIRK1(34 -86) region, as identified by Huang et al. (17). Similarly, the C-terminal GIRK4-(253-348) region overlaps with the GIRK1 regions identified by Kunkel and Peralta (18) and by Huang et al. (16,17). Moreover, comparisons of these studies together suggest FIG. 5. Mutation of the critical His and Leu residues in GIRK1* or wildtype heteromultimers GIRK1/GIRK4 also reduce basal G␤␥-mediated currents. A, His-57 and Leu-262 in GIRK1* corresponding to the His-64 and Leu-268 in GIRK4* were also found to be critical for G␤␥-mediated basal current. Data from a representative batch of oocytes are shown (n ϭ 3-5 oocytes). B, His and Leu double mutations of the wild-type GIRK1 or GIRK4 were each tested in heteromultimers with the wild-type counterpart or with each other. Currents were normalized relative to those from the wild-type GIRK1/GIRK4 (p Ͻ 0.001, n ϭ 4). C, single His or Leu mutation of the wild-type GIRK1 or GIRK4 were expressed alone or in combination with their wild-type counterparts. Currents were normalized to those from GIRK1/GIRK4 (as in B).
that GIRK1 may contain additional G␤␥-binding sites in its unique C-terminal region, which are distinct from other subfamily members.
In our previous study (24) we determined that when the 323-348 region of GIRK4 was replaced with the corresponding region from IRK1, it supported G␤␥-mediated basal currents but not agonist-induced currents. In our present study, deletion of this region from a C-terminal fragment showing maximal binding decreased binding to G␤␥ (compare the G␤␥ binding to the GST-G4-(253-348) versus the GST-G4-(253-322)). These results suggest that the difference in binding reflects channel-G␤␥ interactions occurring during agonist-induced stimulation. The difference in binding contributed by the 322-348 region was similar to that obtained upon mutation of a single residue within this region (i.e. L339E; see Ref. 24). In addition, GIRK4*(L339E) lacked ACh-sensitive currents but displayed G␤␥-sensitive basal currents (24). Because G␤␥ binding and G␤␥-mediated basal currents were intact in GIRK4*(L339E), it was concluded that additional G␤␥-binding sites contribute to basal channel activity.
In the present study we have identified two channel sites critical for G␤␥ interactions and overall channel activity. Mutation of GIRK4(H64) and GIRK4(L268) reduced binding to G␤␥ and abolished K ϩ currents. Interestingly, mutation of the corresponding residue His-57 in GIRK1 also abolished K ϩ currents. This result is in contrast to that obtained from mutation of the corresponding C-terminal residue Leu-262, which significantly reduced K ϩ currents but did not abolish them. This result is consistent with the idea that the GIRK1 C terminus contains additional G␤␥ interaction sites that contribute to G␤␥-mediated activity (17).
Biochemical and electrophysiological evidence from several studies, including ours, have suggested multiple binding sites in the C-and N-terminal segments of GIRK channels (16,17,21,38). The multiplicity of G␤␥-binding sites with effector proteins is in agreement with a number of studies including the crystal structure of the G␤␥-phosducin complex, where multiple sites of interaction are seen between the two proteins (39). We have provided evidence suggesting distinct functional roles to specific binding sites. Thus, the Leu-339 site interacts with G␤␥ released from receptor stimulation (24), whereas the His-64 and Leu-268 sites interact with G␤␥ to yield basal activity and enable overall G␤␥-mediated stimulation. Our results suggest that for the K ϩ channel the multiplicity of interactions subserves distinct functional roles.
The N but not the C terminus of GIRK1 has been shown to support interactions with the heterotrimeric G␣␤␥ protein. It is intriguing to hypothesize that such interactions of the channel with the heterotrimeric G protein confer specificity to G␤␥ signaling by hardwiring the signaling molecules into a receptor-channel-effector macromolecular complex. The N-terminal GIRK channel mutations (GIRK4-H64F and GIRK1-H57F) abolished all K ϩ currents. The mechanism by which the Nterminal mutation resulted in a complete loss of function is not known. The GIRK4(H64F) mutation decreases binding to G␤␥, and consequently it could result in a decrease in the binding of the heterotrimeric G protein to the N terminus and disruption of the presumed macromolecular complex.
In a recent study we identified three functionally important G␤ 1 residues that interact with GIRK channels. 2 One of these, Ser-98, is shared with G␣ subunits and is thought to be involved in agonist-induced interactions. The other two G␤ 1 residues, Ser-67 and Thr-128 that do not interact with G␣ subunits, affect basal K ϩ currents when mutated. Possible interactions of GIRK4/1 residues His-64/His-57 and/or Leu-268/Leu-262 with the G␤ 1 residues Ser-67 and Thr-128 remain to be tested.
Mutation of the two G␤␥-interacting residues, GIRK4* (H64F) and GIRK4*(L268I), abolished both basal and agonistinduced activities. The activity of each of these two mutants could be rescued by strengthening channel-PIP 2 interactions with the mutation GIRK4*(I229L) (35). Thus, although agonist-induced activation requires basal activity, intact basal channel-G␤␥ interactions do not seem to be a prerequisite for agonist stimulation of the channel. We have shown previously (24) that agonist stimulation of channel activity involves G␤␥ interactions with residues GIRK4(L339) and GIRK1(L333). In the present study we show that G␤␥-mediated basal activity depends on interactions with sites GIRK4(H64,L268) and GIRK1(H57,L262). Because G␤␥ stimulation has been shown to lead to stabilization of channel-PIP 2 interactions (34, 25), we previously proposed that G␤␥ gating of the channel may proceed through modulation of channel-PIP 2 interactions (37). Our FIG. 7. Mutants H64F and L268I reduce G␤␥ binding to the channel. GST fusion constructs were made using wild-type and mutant minimal binding domains from the N and C termini of GIRK4. GST was used as negative and ␤ARK-PH as positive controls for G␤␥ binding. Fusion proteins were purified and detected, and G␤␥ interactions were determined as described in Fig. 1. The position of G␤ is indicated by the arrow. Density scanning was used to quantify the relative amounts of bound G␤␥. The bar graph summarizes binding data for the mutant and wild-type constructs. The results shown represent mean Ϯ S.E. for 5-6 separate experiments. *, p Ͻ 0.01 compared with the corresponding wild-type group.
previous (24) and present work has implicated distinct channel sites in interactions with G␤␥ to produce basal or agonistinduced currents. Thus, it is possible that each of the channel-G␤␥ interactions (i.e. basal versus agonist-dependent) stabilize distinct channel-PIP 2 interactions. Precise identification of specific channel-PIP 2 interaction sites modulated by each of the channel-G␤␥ interaction sites remains to be determined.