Gαi and Gβγ Jointly Regulate the Conformations of a Gβγ Effector, the Neuronal G...

G�i and G�� Jointly Regulate the Conformations of a G��Effector, the Neuronal G Protein-activated K� Channel (GIRK)*□S

Received for publication, November 17, 2009, and in revised form, December 15, 2009 Published, JBC Papers in Press, December 16, 2009, DOI 10.1074/jbc.M109.085944

Shai Berlin‡1,2, Tal Keren-Raifman‡, Ruth Castel‡3, Moran Rubinstein‡1, Carmen W. Dessauer§, Tatiana Ivanina‡,and Nathan Dascal‡4

From the ‡Department of Physiology and Pharmacology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel andthe §Department of Integrative Biology and Pharmacology, University of Texas Health Science Center, Houston, Texas 77030

Stable complexes among G proteins and effectors are anemerging concept in cell signaling. The prototypical G�� effec-tor G protein-activated K� channel (GIRK; Kir3) physicallyinteracts with G�� but also with G�i/o. Whether and how G�i/osubunits regulate GIRK in vivo is unclear. We studied tripleinteractions among GIRK subunits 1 and 2, G�i3 and G��. Weused in vitro protein interaction assays and in vivo intramolec-ular Forster resonance energy transfer (i-FRET) between fluoro-phores attached to N and C termini of either GIRK1 or GIRK2subunit. We demonstrate, for the first time, that G�� and G�i3distinctly and interdependently alter the conformational statesof the heterotetrameric GIRK1/2 channel. Biochemical experi-ments show that G�� greatly enhances the binding of GIRK1subunit to G�i3

GDP and, unexpectedly, to G�i3GTP. i-FRET

showed that both G�i3 and G�� induced distinct conforma-tional changes in GIRK1 and GIRK2. Moreover, GIRK1 andGIRK2 subunits assumed unique, distinct conformations whencoexpressedwith a “constitutively active”G�i3mutant andG��together. These conformations differ from those assumed byGIRK1 or GIRK2 after separate coexpression of either G�i3 orG��. Both biochemical and i-FRET data suggest that GIRK actsas the nucleator of the GIRK-G�-G�� signaling complex andmediates allosteric interactions between G�i

GTP and G��. Ourfindings imply that G�i/o and the G�i�� heterotrimer can reg-ulate a G�� effector both before and after activation byneurotransmitters.

It is believed that signaling via G protein-coupled receptors(GPCRs)5 occurs within multiprotein complexes that include

GPCRs, G proteins, and effectors (1–3). The G protein-acti-vated K� channel (GIRK, Kir3), an important mediator of neu-ronal inhibition (4), is activated by the binding of G�� to thecytosolic N and C termini (NT and CT, respectively) of GIRK.G�� associates with the channel before and after receptor acti-vation (5, 6). GIRK NT and CT segments also bind G�i/o (seeRef. 7), but it is not clearwhether and howG�i/o regulatesGIRKin vivo. G� clearly plays a role in determining specificity ofsignaling from GPCR to GIRK (8, 9). In heterologous systems,expression of G�i reduces GIRK basal activity (Ibasal) andincreases the relative extent of activation by added or coex-pressedG�� (10–12). Recently, we have demonstrated that thisregulation is exerted by the GDP-bound form of G�i, G�i

GDP

(12); no role for G�iGTP inGIRK gating could be assigned so far.

We proposed that regulation of GIRK by G�i relies upon theformation of theG�i

GDP-G�� heterotrimer, which forms a per-sistent, dynamic signaling complexwithGIRK to ensure propergating with low Ibasal and high signal-to-background ratio uponG�� activation (11, 12).

We hypothesized that within a multiprotein signaling com-plex, the partner proteins alter each other’s conformation andactivity at various stages of the signaling process. Thus, G�i andG�� may regulate the conformation of the channel both beforeand after activation.We explored this hypothesis in vitro and invivo using biochemical, functional, and intramolecular Forsterresonance energy transfer (i-FRET) methods. Our data revealinterdependent triple interactions amongG�i3, G��, andGIRKsubunits, which correlate with distinct conformational and gat-ing states of the GIRK channel.

EXPERIMENTAL PROCEDURES

Additional details on all methods are available in the supple-mental methods.cDNA Constructs and Electrophysiology—The cDNAs used

in this study were obtained or prepared using standard PCR-based procedures. G1NC was constructed as described (12),with an 8-amino acid linker GSTASGST replacing the trans-membrane segment (amino acids 85–184). Doubly labeled(DL)-GIRK1 and DL-GIRK2 were created by fusing CFP to theNT and YFP to the CT of the GIRK subunit, via linkers (seesupplementalmethods).Other cDNAconstructswere as inRef.

* This work was supported, in whole or in part, by National Institutes of HealthGrants GM68493 (to N. D.) and GM60419 (to C. W. D.). This work was alsosupported by the United States Israel Binational Science Foundation Grant01-122 (to N. D. and C. W. D.), Israel Science Foundation Grants 49/08 and1396/05 (to N. D.), and a Segol Fellowship (to M. R.).

□S The on-line version of this article (available at http://www.jbc.org) containssupplemental methods and supplemental Figs. S1–S5.

1 The work was performed in partial fulfillment of the requirements for Ph.D.degrees at Tel Aviv University.

2 To whom correspondence may be addressed. Tel.: 972-3-6405743; Fax: 972-3-6409113; E-mail: [email protected].

3 The work was performed in partial fulfillment of the requirements for theM.Sc. degree at Tel Aviv University.

4 To whom correspondence may be addressed. Tel.: 972-3-6405743; Fax: 972-3-6409113; E-mail: [email protected].

5 The abbreviations used are: GPCR, G protein-coupled receptor; GIRK, G pro-tein-activated K� channel; ivt, in vitro translated; CT, C terminus; NT, Nterminus; GST, glutathione S-transferase; i-FRET, intramolecular Forsterresonance energy transfer; CFP, cerulean fluorescent protein; YFP, yellow

fluorescent protein; GTP�S, guanosine 5�-3-O-(thio)triphosphate; ACh,acetylcholine; m2R, m2 receptor; DL, doubly labeled; PM, plasmamembrane.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 285, NO. 9, pp. 6179 –6185, February 26, 2010© 2010 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.

FEBRUARY 26, 2010 • VOLUME 285 • NUMBER 9 JOURNAL OF BIOLOGICAL CHEMISTRY 6179

http://www.jbc.org/cgi/content/full/M109.085944/DC1





12. Xenopus oocytes were injected with RNAs, and whole-cellcurrents were measured using standard two-electrode voltageclamp procedures at 20–22 °C, in the ND96 (low K�) solution(96 mM NaCl, 2 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 5 mM

HEPES, pH 7.5) or in a high K� solution (see Fig. 3, hK) (24 mM

K�, isotonically replacing NaCl in ND96) as described (12) (seesupplemental methods). Differences in current amplitude,resulting from variations inGIRK expression, were corrected tochannel expression,determinedby thequantitationofYFP fluo-rescence as described (12). All currents were normalized to thecontrol group.PulldownAssay—GST-fusedG�i3 and itsmutantswere puri-

fied as described (13). [35S]Methionine-labeled segments ofGIRK1 (G1NC (G1N1–84C183–501); G1NYFP (YFP-N1–84); G1C(C184–501); G1C363–501 (C363–501)) were synthesized in rabbitreticulocyte lysate. Interaction among GST-fused G�i3 or itsmutants with the in vitro translated (ivt), [35S]methionine-la-beled segments of GIRK1 was studied by pulldown on glutathi-one affinity beadswith eitherGDPorGTP�S (see supplementalmethods). The eluted proteins were separated on 12% poly-acrylamide-SDS gels. The radioactive signals from proteinbands of the gels were imaged and quantitated using Phospho-rImager and the software ImageQuant (GE Healthcare). West-ern blots were performed using G� antibody (Santa Cruz Bio-technology) and ECL reagents from Pierce.Imaging and Spectral FRET Analysis—Fluorescent signals

were collected with the Zeiss 510 META confocal microscopeand analyzed as described (14), with modifications (see supple-mental methods). Briefly, two spectra were collected from theanimal hemisphere of each oocyte, with 405-nm (CFP excita-tion) and 514-nm (YFP excitation) laser lines. Net FRET signal

of the CFP/YFP-labeled channelswas calculated in the YFP emissionrange (with the 405 nm excitation)by consecutive subtraction of ascaled CFP-only spectrum (givingthe A ratio parameter) and then ofthe ratio A0, which reports thedirect excitation of YFP by the405-nm laser. Therefore,

Ratio A �F405

F514�

Fdirect

F514�

FFRET

F514

(Eq. 1)

where

Fdirect

F514� Ratio A0 (Eq. 2)

Then

� A � A0� �FFRET

F514(Eq. 3)

Dynamic i-FRET—Ionic currentswere recorded in ND96 solution at�80 mV (see supplemental meth-ods). Oocytes were repetitivelyexcited with the 405-nm laser every

40 s. An �150 � 100-�m region of the membrane area wasimaged. The CFP and YFP fluorescence was collected using470–500- and 505–550-nm band pass filters, respectively, andbackground signals were subtracted. YFP andCFP intensities ateach point were normalized to an initial four measurements ineach cell. This configuration allows leakage of CFP into the YFPrecording window of�0.30. Because this leakage is purely opti-cal and constant irrespective of the FRET changes, we did notcorrect for it (15). The ratio of normalized intensities wasdenoted FYFP/FCFP (16). Some photobleaching occurred duringrepetitive excitations, but it had negligible effects on the FYFP/FCFP (see Fig. 4). Cells showing �15% bleaching of YFP or CFPwere discarded.Statistical Analysis—Results are shown as mean S.E. (see

also supplemental methods).Asterisks or pound signs in Figs. 1,3, and 4 indicate statistically significant differences as follows: *or #, p 0.05; ** or ##, p 0.01; *** or ###, p 0.001, ascompared with the control group, unless specified differently.

RESULTS

GIRK-G�-G��Triple Interactions in Vitro—Weconstructeda cDNA encoding the complete cytosolic domain of GIRK1,G1NC, in which the NT and CT were fused whereas the trans-membrane region was replaced by a short linker (Fig. 1A), totest its interaction with the G proteins. Similar constructs werepreviously shown to have a strong propensity to form stabletetramers in solution (17, 18). The ivt 35S-labeled G1NC gave asingle protein band on SDS gel, as expected (Fig. 1B).Moreover,G1NC bound purified G�1�2 (12). We have recently reported(12) that G�� enhances the interaction of G�i3

GDP with the

FIGURE 1. G�� enhances the interaction between G�i3 and GIRK1. A, schematic depiction of the ivt synthe-sized GIRK1 fragments. B, pulldown of 35S-labeled ivt proteins by GST-G�i3 in the presence of GDP or GTP�S,with or without G��. Autoradiograms of one-sixtieth of the total loaded protein (upper panel, Input) and of thebound protein (lower panel, Binding) are shown. C, summary of binding of the ivt synthesized GIRK1 proteins(normalized to that of G1NC to G�i3

GDP) to GST-G�i3, in the presence of either GDP or GTP�S, with or withoutthe addition of 3 �g of purified G�1�2. D, the interaction of G�i3GA or G�i3QL with purified G�1�2, measured byWestern blots as in E, in the presence of GDP or GTP�S. E, the interaction of G1NC with G�i3GA or G�i3QL, in thepresence of either GDP or GTP�S, with or without G��. The upper panel shows the summary of all experiments.Below the summary is a representative autoradiogram of pulled-down G1NC, and the bottommost panel showsWestern blot of bound G�. In D and E, in each experiment, binding of G�� or G1NC was normalized to thatmeasured with G�i3QLGDP. Results are shown as mean S.E. *, p 0.05; **, p 0.01, as compared with thecontrol group.

G�i and G�� Control Conformation of a G�� Effector

6180 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 285 • NUMBER 9 • FEBRUARY 26, 2010









G1NC.Here, we further characterized the interaction ofGIRK1with G�i3

GDP and also with G�i3GTP.We similarly produced ivt

CT of GIRK1 (G1C), the distal half of the CT (G1C363–501) andYFP-fused NT of GIRK1, G1NYFP (Fig. 1A). A GST-fused G�i3bound ivt G1NC as well as separate G1C and G1NYFP, whereasYFP or the distal CT (G1C363–501) did not show detectablebinding (Fig. 1, B and C). The latter observations confirmresults obtained in a reciprocal configuration, with GST-fusedchannel parts and ivt G�i3 (9).

The addition of purified G�1�2 had no effect on the bindingof GST-G�i3 to separate GIRK1 N and C termini, in the pres-ence of either GDP or GTP�S (Fig. 1B, summary in panel C). Incontrast, the binding of G1NC to GST-G�i3

GDP was enhanced�9-fold by the addition of G�1�2, corroborating the previousreport (12). Unexpectedly, G�� also strongly enhanced thebinding ofG1NC toGST-G�i3 in the presence ofGTP�S (Fig. 1,B and C), a condition where GST-G�i3 poorly binds G�1�2(supplemental Fig. S1). To alleviate the concern that a residualamount of G�i3�� heterotrimers remaining in GTP�S couldbias the results, we constructed GST-fused versions of two wellcharacterized G�i3 mutants that mimic G�i3

GTP and G�i3GDP,

respectively: constitutively active G�i3Q204L (G�i3QL) andconstitutively inactive G�i3G203A (G�i3GA) (see Ref. 12).GST-G�i3GA showed the expected strong interaction withG�1�2 in GDP andmuch less in GTP�S, as reported (19). GST-G�i3QL showed some interaction with G�1�2 in GDP (see alsoRef. 20) but not in GTP�S (Fig. 1D, example in panel E).Both GST-G�i3GA and GST-G�i3QL bound G1NC in the

absence of G�� (Fig. 1E). G�� enhanced the binding of bothG�i3GA and G�i3QL to G1NC, in GDP as well as in GTP�S,despite the great differences in their interaction with G�� (Fig.1E). Thus, GIRK1 likely binds the active G�i3 (G�i3

GTP�S andG�i3QL) directly and not through G��; by binding to GIRK1,G�� enhances this interaction.G Protein Subunits Induce Non-identical Conformational

Rearrangements in GIRK1 and GIRK2—i-FRET was previouslyused to detect conformational changes in membrane proteins(16). We created a doubly labeled GIRK1 subunit (DL-G1) byfusing CFP (the donor) and YFP (the acceptor) to the extremesof NT and CT of GIRK1, respectively. DL-G1 was coexpressedwith the neuronal GIRK2 subunit in Xenopus oocytes to giveDL-G1/G2 channels, and visualized in the plasma membrane(PM) using a confocal microscope (Fig. 2). A linear relation ofYFP to CFP fluorescence in DL-G1 confirmed the expectedCFP/YFPmolar ratio of 1:1 (supplemental Fig. S2C). We testedthe functionality of all our tagged clones (Fig. 3) and found thatDL-G1/G2 displayed similar function to wild-type G1/G2. TheDL-G1/G2 channel had the characteristic basal activity (Ibasal),was readily activated by acetylcholine (ACh) via a coexpressedmuscarinic m2 receptor (m2R) and blocked by Ba2� (Fig. 3Aa).Moreover, DL-G1/G2 exhibited the typical inward rectification(Fig. 3Ab). Importantly, the regulation of the channel by coex-pressed G�i3GA, G�i3QL and the G��-scavenging proteinm-phosducin, was also identical to that of the wild-type G1/G2channel (Fig. 3B) (12).We found that G�i3GA strongly reducedIbasal and when coexpressed with G��, G�i3GA did not reducethe total G��-dependent current, I�� (as also observed for thewild-type channel). Thus, G�i3GA enhanced the relative acti-

vation by coexpressed G�� (R��) (supplemental Fig. S3A).G�i3QL affected neither Ibasal, I��, nor R��. m-Phosducinreduced Ibasal but, in sharp contrast to G�i3GA, also substan-tially diminished I��, acting as a typical G�� scavenger (Fig. 3Band supplemental Fig. S3A) (11, 12).We quantified the i-FRET signal using the spectral FRET

technique (14) (Fig. 2C and supplemental Fig. S2E). Each oocytewas excited at twowavelengths, 405 and 514nm, emission spec-tra were collected (Fig. 2, B and C, and supplemental Fig. S2E),and parameters characterizing bleed-through and energytransfer (ratios A0 and A, respectively) were calculated. Theextent of FRET is proportional to (A � A0). Ratios A and A0were linear over a broad range of wavelengths, a testimony tothe absence of optical and calculation artifacts (14) (Fig. 2D).We found a strong basal i-FRET signal ((A � A0) � 0.28

0.02) in DL-GIRK1/2 (Figs. 2D and 4A), indicating proximitybetween the CFP and YFP moieties. Coexpression of G�� sig-nificantly increased i-FRET (by �17%; Fig. 4A) concomitantlywith the increase in channel activity (Fig. 3B). Inversely, coex-pression of G�i3GA or m-phosducin markedly decreased bothi-FRET (by up to 20%) and Ibasal. Thus, low i-FRET inDL-G1/G2 correlates with non-conductive conformation(s),and high i-FRET corresponds toG��-activated, open state(s) ofthe channel.

FIGURE 2. Imaging DL-GIRK1/2 at PM and assessing i-FRET by spectralFRET technique. A, schematic presentation of the DL-G1/G2 channel (thecytosolic domains of GIRK2 are not shown). Bottom insets, color-coded images(averaged emission spectrum) of three oocytes expressing DL-G1/G2obtained with 514 nm (panel a) and 405 nm excitation (panel b). Objective �5,zoom 0.7. B, spectral FRET. Shown are four images of oocytes expressing: 1,m2R-YFP excited with 514 nm; 2, m2R-YFP excited with 405 nm (white arrowpoints to the PM of the oocyte); 3, m2R-CFP excited with 405 nm; 4, DL-G1excited with 405 nm. Objective �20, zoom 2. White bar � 50 �m. C, normal-ized emission spectra of m2R-CFP (blue trace, 3) and DL-G1/G2 (green trace, 4)and the FFRET subtracted spectrum (red trace, 4-3). Numbers in C and D matchimages in panel B. The dichroic mirror is marked by a light gray area. FL, fluo-rescent. D, ratio A (black) of DL-G1/G2 and ratio A0 (gray) plotted against theemission wavelengths. Results are shown as mean S.E.









Coexpression of G�� with G�i3-GA restored the active statewith large I�� and with high i-FRET. In contrast, coexpressionof G�� did not reverse the m-phosducin-induced reduction in

i-FRET and in I�� (Figs. 3B and 4A), supporting the differenti-ation of the actions of G�i3GA from those of a simple G��scavenger (11). Expression of G�i3QL did not significantly alterthe basal i-FRET and currents, but further coexpression of G��led to a strong increase in i-FRET (�30%), significantly greaterthan under all other conditions (Fig. 4A; see supplemental Fig.S4 for complete statistical analysis).Unlike GIRK1, G�� does not enhance G�i3-GIRK2 interac-

tion, and G�i3 does not regulate Ibasal of homotetramericGIRK2 (12). To assess whetherG�i3 andG�� confer conforma-tional changes upon GIRK2 within the GIRK1/2 heterotet-ramer, we constructed a doubly labeled GIRK2 (DL-G2) as inGIRK1 (see “Experimental Procedures”). DL-G2 was expressedwith wild-type GIRK1 to produce G1/DL-G2 heterotetramers,which showed adequate regulation by G�i3 and G�� (supple-mental Fig. S3B) and showed basal i-FRET of 0.25 0.02 (n �44). Similarly to DL-G1, i-FRET of DL-G2 was reduced byG�i3GA (by �20%), an effect reversed by G�� (Fig. 4B). How-ever, other parameters of i-FRET in the G1/DL-G2 heterotet-ramer were differently affected by the coexpressed G proteins.Coexpression of G�� did not increase i-FRET above basal leveleither in the presence or in the absence of G�i3. Most interest-ingly, concomitant expression of G�i3QL with G�� signifi-cantly reduced the basal FRET signal (by�26%), opposite to itseffect on DL-G1.As a control, we engineered a doubly labeled IRK1 (Kir2.1)

channel, DL-IRK1 (Fig. 3C and supplemental Fig. S5). Thisinwardly rectifying K� channel does not directly interact with,and is not regulated by, G proteins (21). DL-IRK1 exhibited PMlocalization, constitutively active gating, the typical stronginward rectification (Fig. 3, C and D) and strong basal i-FRET((A � A0) � 0.35 0.01). Coexpression of G��, G�i3GA,G�i3QL, or m-phosducin caused no significant changes ini-FRET (Fig. 4C) or in channel currents. These results confirmthe specificity of G protein regulation of conformational statesof GIRK1/2 by G protein subunits, revealed by i-FRET.Dynamic Structural Rearrangements inGIRK1/2 uponGPCR

Activation—The changes in i-FRETcaused by the presence of G proteinsubunits probably reflect the con-formations of the channel at differ-ent activation steps. To verify thatthese static measurements reflectthe conformations adopted by thechannel during its physiologicalactivation process, we monitoreddynamic changes in i-FRET causedby activation of a GPCR (m2R). Inthese experiments, the electro-physiological response (K� current)and fluorescence signals were col-lected simultaneously from individ-ual oocytes (Fig. 5A).Oocytes expressingDL-GIRK1/2,

m2R, and the wild-type G�i3 werevoltage-clamped at �80 mV andcontinuously perfused with ND96solution. Concomitantly, a region of

FIGURE 3. GIRK1 and IRK1 functionality. Aa, representative DL-G1/G2 cur-rents measured in Xenopus oocytes. Horizontal bars indicate the exchange ofthe external solutions. Shift from low K� ND96 (ND) to 24 mM K� (hK) revealedthe inward Ibasal, and the addition of 10 �M ACh induced additional current,IACh. The addition of 5 mM Ba2� inhibited all GIRK currents. Ab, current-voltagecurve obtained using a voltage ramp from �120 to �50 mV. The curve dem-onstrates the typical mild inward rectification of the GIRK1 channel. B, sum-mary of GIRK currents (five experiments, n � 13–25 oocytes/group) correctedfor the DL-G1 PM expression and normalized to Ibasal of DL-G1/G2. Bars withbold margins represent groups with coexpressed G��. Ca, schematic presen-tation of the tetrameric DL-IRK1 channel (CT and NT of only two out of foursubunits are shown). Cb, current-voltage curve obtained using a voltageramp from �120 to �20 mV. Results are shown as mean S.E. *, p 0.05, ascompared with the control group.

FIGURE 4. i-FRET reveals distinct G protein-induced conformational changes in GIRK1 and GIRK2 but notin IRK1. The experiments were performed on heterotetrameric G1/G2 channels containing either DL-G1 (A) orDL-G2 (B) and homotetrameric IRK1 (C). For illustration, one double-labeled subunit (within the tetramericchannel) is depicted above each graph. A, summary of five independent experiments showing changes ini-FRET in DL-G1/G2 caused by G�i3GA, G�i3QL, G��, and m-phosducin (�m-phosd). In each experiment, the(A � A0) parameter in each oocyte was normalized to the mean (A � A0) of the DL-GIRK1 (control) group. (A0denotes the average A0 measured from more than seven oocytes in the same experiment.) B, summary of threeindependent experiments showing changes in i-FRET in G1/DL-G2 caused by G�i3GA, G�i3QL and G��.C, summary of two experiments showing i-FRET in IRK1. Note: The G protein subunits or m-phosducin do not induceany changes in the i-FRET of IRK1. n.s., not significant. The number of oocytes is shown within bars. Statisticalcomparisons of multiple groups were done using one-way analysis of variance followed by pairwise Student-New-man-Keuls test. Results are shown as mean S.E. *, #, p 0.05, **, ##, p 0.01. n.s.: non-significant.








the oocyte membrane was excited with the excitation wave-length of the donor (405 nm) every 40 s. The emission signals ofboth donor and acceptor were recorded separately, and theFYFP/FCFP ratio was calculated. This ratiometric imagingmethod allows the monitoring of changes in the relative posi-tion of donor and acceptor fluorophores. For instance, reduc-tion of the distance between donor and acceptor increases theenergy transfer, causing an increase in YFP emission intensityconcurrently with a decrease in CFP emission intensity andresulting in an increase in FYFP/FCFP ratio (16).

Under the recording conditions used, DL-G1/G2 showedconsiderable basal activity (11), corresponding to a mixture of“preactivated” states (see “Discussion”). The addition of 10 �M

ACh (Fig. 5B, gray area) caused additional activation of GIRK(Fig. 5B, top red trace) and a concomitant increase in FYFP/FCFPratio (Fig. 5B). When ACh was washed from the system, thechannel reassumed its basal, preactivated FYFP/FCFP ratio andbasal activity. These dynamic changes match our static i-FRETdata (see “Discussion”). No change in FYFP/FCFP was seen in

control experiments, where oocytes were exposed to 30 �M

atropine, which completely inhibited the ACh-evoked K� cur-rent (Fig. 5C). Importantly, pretreatment with pertussis toxin(PTX) completely inhibited both ACh-evoked currents and theACh-induced change in i-FRET (Fig. 5D). The use of atropinerules out the involvement of the receptor in the induction of thechange in conformation of the GIRK, and the latter is mostprobably caused by the direct interaction of the G proteinsthemselves. Finally, ACh did not induce any ionic currents orchanges in FYFP/FCFP in DL-IRK1 channel (Fig. 5E).

DISCUSSION

G�� Is Crucial for Strong Interaction betweenGIRK1andG�i—Our biochemical results strongly support the idea of a pre-formed signaling complex of heterotrimeric G�i�� withGIRK1, initially based upon the finding of a strong interactionof G�i1

GDP�� with the NT of GIRK1 in vitro (5). Indeed, G��greatly enhanced the interaction of GIRK1 with G�i3

GDP underconditions favoring the formation of G�� heterotrimers (Fig.1) (12). Somewhat at oddswithHuang et al. (5), our data suggestthat both NT and CT are necessary for the formation of thestrong GIRK1-G�i�� complex as the effect of G�� was notpresent in separate N and C termini. Further quantitative bind-ing studies may be needed to understand the reason for thisdiscrepancy.A novel and unexpected finding was that G�� also enhanced

the interaction ofGIRK1withG�i3 and the constitutively activeG�i3QL in GTP�S, a condition when no G�i3QL-G�� interac-tion was detected by pulldown (supplemental Fig. S1). There-fore, we suggest that the conformational change in GIRK1induced by the binding ofG�� allosterically improves the directbinding of G�i3

GTP to GIRK1. Alternatively, GIRK1 may stillbind G�i3

GTP via G��, utilizing a second G��-binding site onG�i3, distinct from the classical high affinity interaction site(22). In all, the biochemical data support the persistence ofGIRK-G�i/o-G�� signaling complexes, both before and afteractivation by agonist-bound GPCR. The continuous attach-ment of G�i to GIRK1/2 would ensure a high local concentra-tion of G� and a diffusion-independent reassociation withG��, which allows for fast termination of the physiologicalresponse upon removal of the agonist because of the fast kinet-ics of G�GDP-G�� binding (23).i-FRET and GIRK Conformation—The available partial crys-

tal structures of GIRK tetramers indicate proximity betweentheNTof one subunit to themid-CTof an adjacent subunit (17,18, 24). In a tetrameric DL-G1/G2 channel, i-FRET couldpotentially arise from proximity of CFP-NT of one DL-G1 sub-unit to the CT-YFP of an adjacent DL-G1. This would require a1-1-2-2 arrangement of subunits in the tetramer, which is prob-ably viable, by analogy with 1-1-4-4 in GIRK1/GIRK4, although1-4-1-4 is preferable (25, 26). However, it is important to notethat the proximal NTs (�40 amino acids) and the distal CT(�130 amino acids in GIRK1, amino acids 370–501; �45amino acids in GIRK2) are omitted from the above crystalstructures. The unique distal CT segment of GIRK1 is essentialfor the strong triple GIRK-G�i-G�� interaction and for thedifferential regulation of GIRK1 and GIRK2 by G�i and G��(12). The flexible proximal NT and distal CT are probably long

FIGURE 5. GPCR activation induces specific dynamic changes in i-FRET ofDL-G1/G2. A, scheme of the experimental setup, where oocyte is bothrepeatedly imaged and electrophysiologically monitored by the two elec-trode voltage clamp method. B, top trace, representative electrophysiologicalrecording made in a low K� solution. The shaded area denotes ACh applica-tion. The middle trace shows averaged intensities in CFP (blue) and YFP (yel-low) windows (n � 14 oocytes from four donor frogs). The bottom trace showsaverage FYFP/FCFP. C–E, control experiments. Each subfigure shows currenttraces recorded in representative oocytes (top) and mean FYFP/FCFP (bottom).C, oocytes expressing DL-G1/G2 were briefly perfused with ACh to verify theexpression of the channels and receptor and then with 30 �M atropine (n � 8).D, the ACh-induced current and changes in i-FRET are abolished in oocytesexpressing DL-G1/G2, pretreated with pertussis toxin (�PTX, n � 6).E, oocytes expressing DL-IRK1 show no electrophysiological and i-FRETresponse to 10 �M ACh (n � 10).




enough to allow for intrasubunit interactions. Such intrasu-bunit contacts were proposed to play an essential role in GIRKgating by G�� (27). Therefore, we interpret an increase ini-FRET as nearing of the NT and CT of either the same oradjacent GIRK subunits. The exact structural rearrangementsremain to be determined. However, it is certain that changes ini-FRET reflect G protein-induced changes in conformation ofthe channel (Figs. 3C and 5B).G Protein-regulated Conformational States of the GIRK

Channel—Measurements of single-channel or population cur-rents in native and chimeric GIRK channels demonstrated thatheterotetrameric GIRK channels assume a number of closedand open conformations with a variable number of G�� mole-cules bound, from a closed, non-conducting state (interpretedas G��-devoid), via intermediate states with one, two, or threeG�� molecules bound and correspondingly increasing openprobability (Po), to a fully G��-activated state with four G��bound and the highest Po (27–30). Analysis of i-FRET, whencombined with biochemical and functional assays, extends ourunderstanding of the rules andmodes of G protein-GIRK inter-actions and of the regulation of channel conformation. A sim-plified schematic of some of the interactions and conforma-tions is presented in Fig. 6.Preactivated (resting) states of GIRK1/2 probably reflect a

mixture of conformations of GIRK1/2 channels, associatedwith a variable number of G�i�� heterotrimers and/or G��that lack matching G�i

GDP (Fig. 6, b and d), as suggested byfunctional data (11). Correspondingly, DL-G1/G2 and G1/DL-G2 expressed alone at high levels yield high Ibasal (10) and anintermediate i-FRET signal. A substantial basal GIRK activityhas been reported in hippocampal and cortical neurons (31, 32);therefore, the preactivated states are physiologically relevant.Overexpression of high doses of the G protein subunits or

G�� scavengers shifts the channel population toward morehomogenous states, some of which are detected by our i-FRETassay. 1) The first is G��-free state, with low Ibasal (Fig. 6a). Thisstate is promoted by m-phosducin, which removes G�� awayfrom the channel (6, 11). It is characterized by low i-FRET andis depicted in Fig. 6 as having a large distance between NT and

CT of GIRK1. 2) The second is G��-activated state(s), withhigh i-FRET andGIRK currents, presumablywith several (up tofour in the full tetramer) G�� dimers bound at activation sites(Fig. 6, d and e). 3) Joint expression of G�� and G�i3QL confersa conformation (Fig. 6c) characterized by the highest i-FRET inGIRK1 but the lowest i-FRET in GIRK2 (Fig. 3G). These oppo-site changes in i-FRET further point toward a difference in reg-ulation of GIRK1 and GIRK2 subunits by G�i and G�� (12).Neither G�iQL nor G�� alone confer such changes in i-FRET,indicating concomitant binding ofG�i3

GTP andG�� to the het-erotetrameric GIRK1/2.Cells overexpressing G��, G��G�iGA, or G��G�iQL

show high constitutive, non-desensitizing GIRK currents (11,12), clearly reporting the open states of the channel. Despite thedifferent relative positions of GIRKN and C termini, these var-ious open state(s) give similar whole-cell currents (but itremains to be seen whether they have distinct single-channelproperties). The overall conformation of the channel in thepresence of ACh should resemble that seen in static FRETexperiments upon coexpression of G�� or, even more closely,G��G�i3QL. Accordingly, dynamic activation of GIRK byACh via coexpressed m2R caused an increase in both GIRKcurrent and i-FRET inDL-G1/G2. The increase in i-FRETmostprobably corresponds to the opening of the channel becauseGIRK channels expressed in Xenopus oocytes show little ACh-induced desensitization over time periods of several minutes(33). Control experiments showed no changes in i-FRET uponexposure to the muscarinic antagonist atropine and after treat-ment with pertussis toxin, which prevents the activation of G�iby GPCRs. Finally, no G protein, or GPCR, changes in i-FRETwere observed in the double-labeled G��-insensitive IRK1channel. These control experiments demonstrate the authen-ticity and specificity of i-FRET changes caused by G�i and G��in GIRK1/2.The low i-FRET and small GIRK currents observed after the

expression of G�i3GA are indistinguishable from those seen inG��-free channels in the presence of m-phosducin, as ifG�i3GA removed G�� away from the channel. However, giventhe compelling evidence for a strong triple association betweenGIRK1, G�i

GDP and G�� (Fig. 1) (5–7, 34), we hypothesize thatthe G�i3GA-G�� heterotrimers, formed after expression ofG�iGA, remain mostly associated with GIRK1/2. In support,we observe specific differences in the effects ofG�i3GA (but notthe QL mutant) on i-FRET and the total GIRK current whencoexpressed with G��. This corroborates the hypothesis that itis the GDP-bound G�i (or G�i�� heterotrimer) that regulatesthe basal activity of GIRK and “primes” the channel for activa-tion by G�� (11, 12).The new biochemical and i-FRET results, as well as previous

functional observations, are compatiblewith a “two-sitemodel”in which GIRK possesses an anchoring site for G�i/o (and/orG�i/o��) and a separate activation site, where the binding ofG�� leads to channel opening (12) (Fig. 6). The assumption ofseparate binding sites for G�i (or G��) and G�� is supported,although not proved, by the absence of competition amongG�iandG�� for binding toG1NCandG2NCunder conditions thatfavor either the formation or the dissociation of G�i�� hetero-trimers and by the more-than-additive character of i-FRET

FIGURE 6. The two-site model of GIRK regulation by G�i and G��. Forsimplicity, the GIRK channel is shown as if composed of only one GIRK1 sub-unit. The locations of anchoring and activation sites are shown arbitrarily.Distances between NT and CT are depicted schematically, where high i-FRETis interpreted to indicate proximity between the N and C termini (of the sameor of adjacent subunits). Inversely, low i-FRET correlates with distancing of thetermini. �phos., m-phosducin.



changes induced by G�iQL and G��. The two sites may lie inproximity (or even partly overlap). According to the two-sitemodel, the high basal activity of overexpressed GIRK1/2 is dueto an excess of G�� associated with the channel, over G�i (11,12, 35). Under these conditions, some channels have G��“uncompensated” by G� and thus bound to the activation sites(Fig. 6d). The coexpressed G�i3GA associates with G�� sub-units, obstructing their interaction with the activation sites(hence the decrease in Ibasal and i-FRET). The newly formedG�i3�� heterotrimers are docked at the anchoring site (Fig. 6b).When more G�� is coexpressed, free G�� subunits bind at theactivation sites, hence the high i-FRET and currents (Fig. 6e).The two-site hypothesis provides an economical description ofa normal physiological situation; upon activation of GPCR,G�� detaches from G�i and interacts with the activation site,whereas G�i

GTP stays at the anchoring site (Fig. 6c). (Alterna-tively, G�� may stay at the same site, and it is G�i that shifts toanother location. Also, G�i and G�� do not have to fully disen-gage (16, 36); a rearrangement followed by an exposure of a fewcrucial residues onG��may be sufficient). OnceGTP is hydro-lyzed by G�, G�� rebinds G� and becomes docked again.Conclusions—Coordinated biochemical, functional, and

i-FRET data support the existence of a persistent, dynamicGIRK1/2-G�i/o-G�� complex, both before and after activationby agonist-bound GPCRs. We propose that GIRK1 acts as thenucleator of the complex and that both active and inactive G�i3remain bound to the channel, ensuring fast and specific activa-tion and termination of the signal. New in vitro and in vivo datareveal intricate triple interactions among GIRK1, G�i3, andG�� and show that the presence of G�� is crucial to ensurestrong GIRK1-G�i interactions. At least some of the effects ofG�� onGIRK1-G�i interaction appear to be allosteric. G�i andG�� induce mutually dependent conformational rearrange-ments in GIRK1/2, characterized by distinct changes in i-FRETin the GIRK subunits and often, but not always, by changes inchannel activity.

REFERENCES1. Dupre, D. J., Robitaille, M., Rebois, R. V., and Hebert, T. E. (2009) Annu.

Rev. Pharmacol. Toxicol. 49, 31–562. Doupnik, C. A. (2008) J. Recept. Signal Transduct. Res. 28, 83–913. Dowal, L., Provitera, P., and Scarlata, S. (2006) J. Biol. Chem. 281,

23999–240144. Yamada, M., Inanobe, A., and Kurachi, Y. (1998) Pharmacol. Rev. 50,

723–7605. Huang, C. L., Slesinger, P. A., Casey, P. J., Jan, Y. N., and Jan, L. Y. (1995)

Neuron 15, 1133–11436. Riven, I., Iwanir, S., and Reuveny, E. (2006) Neuron 51, 561–5737. Clancy, S. M., Fowler, C. E., Finley, M., Suen, K. F., Arrabit, C., Berton, F.,

Kosaza, T., Casey, P. J., and Slesinger, P. A. (2005)Mol. Cell. Neurosci. 28,375–389

8. Leaney, J. L., and Tinker, A. (2000) Proc. Natl. Acad. Sci. U.S.A. 97,5651–5656

9. Ivanina, T., Varon, D., Peleg, S., Rishal, I., Porozov, Y., Dessauer, C. W.,Keren-Raifman, T., andDascal, N. (2004) J. Biol. Chem. 279, 17260–17268

10. Peleg, S., Varon, D., Ivanina, T., Dessauer, C. W., and Dascal, N. (2002)Neuron 33, 87–99

11. Rubinstein, M., Peleg, S., Berlin, S., Brass, D., and Dascal, N. (2007)J. Physiol. 581, 17–32

12. Rubinstein,M., Peleg, S., Berlin, S., Brass, D., Keren-Raifman, T., Dessauer,C. W., Ivanina, T., and Dascal, N. (2009) J. Physiol. 587, 3473–3491

13. Rishal, I., Keren-Raifman, T., Yakubovich, D., Ivanina, T., Dessauer, C.W.,Slepak, V. Z., and Dascal, N. (2003) J. Biol. Chem. 278, 3840–3845

14. Zheng, J., Varnum, M. D., and Zagotta, W. N. (2003) J. Neurosci. 23,8167–8175

15. Okamoto, K., Nagai, T., Miyawaki, A., and Hayashi, Y. (2004)Nat. Neuro-sci. 7, 1104–1112

16. Vilardaga, J. P., Bunemann, M., Feinstein, T. N., Lambert, N., Nikolaev,V. O., Engelhardt, S., Lohse, M. J., and Hoffmann, C. (2009)Mol. Endocri-nol. 23, 590–599

17. Nishida, M., and MacKinnon, R. (2002) Cell 111, 957–96518. Pegan, S., Arrabit, C., Zhou, W., Kwiatkowski, W., Collins, A., Slesinger,

P. A., and Choe, S. (2005) Nat. Neurosci. 8, 279–28719. Ogier-Denis, E., Houri, J. J., Bauvy, C., and Codogno, P. (1996) J. Biol.

Chem. 271, 28593–2860020. Majumdar, S., Ramachandran, S., and Cerione, R. A. (2006) J. Biol. Chem.

281, 9219–922621. Stanfield, P. R., Nakajima, S., and Nakajima, Y. (2002) Rev. Physiol. Bio-

chem. Pharmacol. 145, 47–17922. Wang, J., Sengupta, P., Guo, Y., Golebiewska, U., and Scarlata, S. (2009)

J. Biol. Chem. 284, 16906–1691323. Shea, L. D., Neubig, R. R., and Linderman, J. J. (2000) Life Sci. 68, 647–65824. Inanobe, A.,Matsuura, T., Nakagawa, A., and Kurachi, Y. (2007)Channels

1, 39–4525. Silverman, S. K., Lester, H. A., and Dougherty, D. A. (1996) J. Biol. Chem.

271, 30524–3052826. Tucker, S. J., Pessia, M., and Adelman, J. P. (1996) Am. J. Physiol. 271,

H379–38527. Sadja, R., Alagem, N., and Reuveny, E. (2002) Proc. Natl. Acad. Sci. U.S.A.

99, 10783–1078828. Yamada,M., Jahangir, A., Hosoya, Y., Inanobe, A., Katada, T., andKurachi,

Y. (1993) J. Biol. Chem. 268, 24551–2455429. Ivanova-Nikolova, T. T., Nikolov, E. N., Hansen, C., and Robishaw, J. D.

(1998) J. Gen. Physiol. 112, 199–21030. Nemec, J., Wickman, K., and Clapham, D. E. (1999) Biophys. J. 76,

246–25231. Chen, X., and Johnston, D. (2005) J. Neurosci. 25, 3787–379232. Wiser, O., Qian, X., Ehlers, M., Ja,W.W., Roberts, R.W., Reuveny, E., Jan,

Y. N., and Jan, L. Y. (2006) Neuron 50, 561–57333. Vorobiov, D., Levin, G., Lotan, I., andDascal, N. (1998) Pflugers Arch. 436,

56–6834. Rebois, R. V., Robitaille, M., Gales, C., Dupre, D. J., Baragli, A., Trieu, P.,

Ethier, N., Bouvier,M., andHebert, T. E. (2006) J. Cell Sci. 119, 2807–281835. Rishal, I., Porozov, Y., Yakubovich, D., Varon, D., and Dascal, N. (2005)

J. Biol. Chem. 280, 16685–1669436. Gales, C., Van Durm, J. J., Schaak, S., Pontier, S., Percherancier, Y., Audet,

M., Paris, H., and Bouvier, M. (2006) Nat. Struct. Mol. Biol. 13, 778–786



Gαi and Gβγ Jointly Regulate the Conformations of a Gβγ Effector, the Neuronal G...

Documents

Transcript of Gαi and Gβγ Jointly Regulate the Conformations of a Gβγ Effector, the Neuronal G...