Activity of several ion channels is controlled by heterotrimeric GTP-binding proteins (G proteins) via a membrane-delimited pathway that does not involve cytoplasmic intermediates. The best studied example is the K+ channel activated by muscarinic agonists in the atrium, which plays a crucial role in regulating the heartbeat. To enable studies of the molecular mechanisms of activation, this channel, denoted KGA, was cloned from a rat atrium cDNA library by functional coupling to coexpressed serotonin type 1A receptors in Xenopus oocytes. KGA displays regions of sequence homology to other inwardiy rectifying channels as well as unique regions that may govern G-protein interaction. The expressed KGA channel is activated by serotonin 1A, muscarinic m2, and b-opioid receptors via G proteins. KGA is activated by guanosine 5'-[ythio]triphosphate in excised patches, confirming activation by a membrane-delimited pathway, and displays a conductance equal to that of the endogenous channel in atrial cells. The hypothesis that similar channels play a role in neuronal inhibition is supported by the cloning of a nearly identical channel (KGB1) from a rat brain cDNA library.A major signal transduction mechanism in cardiac physiology and neurobiology is the direct coupling of neurotransmitter receptors to ion channels by a membrane-delimited pathway that does not involve cytoplasmic intermediates (for reviews, see refs. 1 and 2). The best studied member of this group is the G protein-activated K+ channel found in atria of all vertebrates. This channel, which we denote KGA, figured in the original discovery of chemical synaptic transmission (3,4) and plays a crucial role in regulating the heartbeat. The KGA channel rectifies at the single-channel level, allowing much larger inward than outward currents (5, 6). It is activated by acetylcholine acting on muscarinic m2 receptors via a pathway that includes a pertussis toxin (PTX)-sensitive G protein, probably of the G, family (1,(7)(8)(9)(10)(11)(12)

The frog Xenopus laevis has provided significant insights into developmental and cellular processes. However, X. laevis has an allotetraploid genome precluding its use in forward genetic analysis. Genetic analysis may be applicable to Xenopus (Silurana) tropicalis, which has a diploid genome and a shorter generation time. Here, we show that many tools for the study of X. laevis development can be applied to X. tropicalis. By using the developmental staging system of Nieuwkoop and Faber, we find that X. tropicalis embryos develop at similar rates to X. laevis, although they tolerate a narrower range of temperatures. We also show that many of the analytical reagents available for X. laevis can be effectively transferred to X. tropicalis. The X. laevis protocol for whole-mount in situ hybridization to mRNA transcripts can be successfully applied to X. tropicalis without alteration. Additionally, X. laevis probes often work in X. tropicalis-alleviating the immediate need to clone the X. tropicalis orthologs before initiating developmental studies. Antibodies that react against X. laevis proteins can effectively detect the X. tropicalis protein by using established immunohistochemistry procedures. Antisense morpholino oligonucleotides (MOs) offer a new alternative to study loss of gene activity during development. We show that MOs function in X. tropicalis. Finally, X. tropicalis offers the possibility for forward genetics and genomic analysis.

Xenopus oocytes injected with GIRK1 mRNA express inwardly rectifying K+ channels resembling IKACh. Yet IKACh, the atrial G protein-regulated ion channel, is a heteromultimer of GIRK1 and CIR. Reasoning that an oocyte protein might be substituting for CIR, we cloned XIR, a CIR homolog endogenously expressed by Xenopus oocytes. Coinjecting XIR and GIRK1 mRNAs produced large, inwardly rectifying K+ currents responsive to m2-muscarinic receptor stimulation. The m2-stimulated currents of oocytes expressing GIRK1 alone decreased 80% after injecting antisense oligonucleotides specific to the 5' untranslated region of XIR, but GIRK1/CIR currents were unaffected. Thus, GIRK1 without XIR or CIR only ineffectively produces currents in oocytes. This result suggests that GIRK1 does not form native homomultimeric channels.

Injection of rat atrial RNA into Xenopus oocytes resulted in the expression of a guanine nucleotide binding (G) protein-activated K+ channel. Current through the channel could be activated by acetylcholine or, if RNA encoding a neuronal SHT1A receptor was coinjected with atrial RNA, by serotonin (5HT). A 5HT-evoked current (ISHT) Xenopus oocytes may be employed for cloning of the G-proteinactivated K+ channel cDNA and for studying the coupling between this channel and G proteins.Parasympathetic regulation of the rate of heart contraction is exerted through the release of acetylcholine (ACh), which opens a K+ channel in the atrium and thus slows the rate of depolarization that leads to initiation of the action potential (1, 2). The coupling between binding of ACh to a muscarinic receptor and opening ofthe K+ channel occurs via a pertussis toxin (PTX)-sensitive heterotrimeric guanine nucleotide binding (G) protein, Gk (3)(4)(5), probably belonging to the inhibitory G-protein Gi family (6,7). Activation of this G-protein-activated K+ (KG) channel by Gk does not require cytoplasmic intermediates (for review, see refs. 8 and 9). However, a long-standing controversy exists as to which G-protein subunit couples to the KG channel. Purified fBy subunit complexes (10,11) and a subunits of the Gi family (6,7,12) activation by G proteins (18-22). The similarity of the channels and of the signaling pathways in atrium and some nerve-cell preparations was strengthened by the demonstration of the coupling of a neuronal 5HT1A receptor (5HT1A-R), transiently expressed in atrial myocytes, to the atrial KG channels (23). By electrophysiological and pharmacological criteria, the atrial KG channel belongs to a family of inward rectifiers that conduct K+ much better in the inward than the outward direction, are blocked by extracellular Na+, Cs+, and Ba2+, and are believed to possess a single-file pore with several permeant and blocking ion binding sites (24). Many inward rectifiers are not activated by transmitters or voltage but seem to be constitutively active. Inward rectification of the atrial KG channel is due to a block of K+ efflux by intracellular Mg2+ (25), but for some channels of this family inward rectification may not depend on the Mg2+ block (26, 27). The molecular structures of atrial and neuronal KG channels are unknown. Inwardly rectifying K+ channels structurally similar to voltage-activated K+ channels have been cloned from plant cells (28,29). Recently, the primary structures of two mammalian inward-rectifier channels have been elucidated by molecular cloning of their cDNAs via expression in Xenopus oocytes: an ATP-regulated K+ channel from kidney, ROMK1 (30), and an inward rectifier from a macrophage cell line, IRK1 (31). Both appear to belong to a superfamily of K+ channels, with only two transmembrane domains per subunit and a pore region homologous to that of K+, Ca2 , and Na+ voltage-dependent channels (see ref. 32). It has been hypothesized that the structure of G-protein-activated inwardly rectify...

G protein-activated inwardly rectifying K ÷ channel subunits GIRK1 (Kit 3.1), GIRK2 (Kit 3.2), and CIR (Kir 3.4) were expressed individually or in combination in Xenopus oocytes and CHO cells. GIRKI coexpressed with CIR or GIRK2, produced currents up to 10-fold larger than any of the subunits expressed alone. No such clear synergistic effects were observed upon coexpression of CIR/GIRK2 under the same conditions. Coexpression of G protein i~/ (Gm~2) increased the current through GIRKllGIRK2 and GIRK2 channels. G~ subunits purified from bovine brain, increased channel activity 50-1000-fold in patches from cells expressing GIRKllGIRK2 or GIRK2 alone. The single GIRKIIGIRK2 channels resembled previously described neuronal G protein-gated K ÷ channels. In contrast, single GIRK2 channels were short-lived and unlike any previously described neuronal K ÷ channel. We propose that some neuronal G protein-activated inward rectifier K ÷ channels may be formed by a GIRKllGIRK2 heteromultimer and that G,~ activation may involve both subunits.