Discovery at the VICB
By: Carol A. Rouzer, VICB Communications
Published: July 21, 2014
A selective agonist of G protein-gated inwardly rectifying potassium channels has anxiolytic effects in mice.
The G protein-gated inwardly rectifying potassium (GIRK) channels play an important signaling role in the heart and nervous system. Comprising tetramers of four subunits (GIRK1-4), these channels tonically conduct potassium ions out of the cell, reducing excitability. When activated by the Gβγ subunits of the Gi/o class of G proteins (Figure 1), K+ flow increases, generating a slow inhibitory post synaptic potential (siPISP), which suppresses future synaptic transmission (Figure 2). In the nervous system, the pysiological activators of GIRK channels are primarily G protein-coupled neurotransmitter receptors; however, the channels can be pharmacologically activated by ethanol, volatile anesthetics, and naringin. Mounting data support a role for GIRK channels in a range of disorders as disparate as epilepsy, addiction, Down’s syndrome, and Parkinson’s disease (Figure 3). However, a complete understanding of the role of GIRK channels in neurophysiology and neuropathology requires molecular probes of GIRK function. In response to this need, Vanderbilt Institute of Chemical Biology member Dave Weaver, his colleague Kevin Wickman (University of Minnesota), and their laboratories explored the mechanism of action and in vivo activities of ML297, a selective agonist of channels containing the GIRK1 subunit [N. Wydeven et al., (2014) Proc. Natl. Acad. Sci. U.S.A., published online July 7, DOI:10.1073/pnas.1405190111].
Figure 1. Schematic diagram of GIRK channel regulation. GIRK channels (purple and light blue) are believed to be closely associated with G protein-coupled neurotransmitter receptors (GPCR shown in green) in the plasma membranes of neuronal cells. The receptors that modulate GIRK channel function are associated with G proteins of the Gi/o class. Activation of the GPCR leads to binding of the Gβγ subunits of the G protein to the GIRK channel. The result of Gβγ binding is an increase in affinity for phosphatidylinositol 4,5-bisphosphate (PtdIns-(4,5)P2), which induces a conformation change in the channel, increasing the flow of K+ out of the cell. Note that K+can flow through the channel from the outside to the inside of the cell if the membrane potential is lowered below the equilibrium potential for K+. This is the basis for the name “inwardly rectifying.” The schematic also shows the wide range of additional proteins that modulate GIRK channel function. Image reproduced by permission from Macmillan Publishers Ltd, from C. Lüscher and P. A. Slesinger (2010) Nat. Rev. Neurosci., 11, 301, copyright 2010.
Figure 2. Diagram of the function of a GIRK channel in the nervous system. One kind of G protein-coupled receptor that signals through GIRK channels is the γ-amino butyric acid-B (GABAB) receptor. Under conditions of low stimulation of GABA neurons (left) small amounts of GABA are released, and these directly stimulate GABAA receptors located across from the nerve terminal. These receptors generate a fast inhibitory post synaptic potential (fIPSC). When stimulation of GABA neurons is high (right), large amounts of GABA diffuse away from the synapse and bind to GIRK-linked GABAB receptors. These generate a slow inhibitory post synaptic potential (sIPSP). Image reproduced by permission from Macmillan Publishers Ltd, from C. Lüscher and P. A. Slesinger (2010) Nat. Rev. Neurosci., 11, 301, copyright 2010.
Figure 3. Diagram of the sagital view of a rat brain indicating regions where GIRK channel signaling is believed to play a role in disordered function. Disorders can result from changes in excitability (labeled in yellow) or cell death (labeled in gray). VTA, ventral tegmental area; PAG, periaqueductal gray; LC, locus coeruleus; SNc, substantial nigra pars compacta. Image reproduced by permission from Macmillan Publishers Ltd, from C. Lüscher and P. A. Slesinger (2010) Nat. Rev. Neurosci., 11, 301, copyright 2010.
Although the pharmacology and anticonvulsant effects of ML297 had been previously described (K. Kaufmann, et al., (2013) ACS Chem. Neurosci., 4, 1278), the basis for its potency and selectivity as an agonist for GIRK1-containing channels was not understood. To address these questions the investigators began their studies with HEK-293 cells transfected with the gene for the γ-aminobutyric acid-B (GABAB) receptor along with the genes for the GIRK1 and GIRK2 subunits. They expressed GIRK1 subunits because ML297 is GIRK1-selective and GIRK2 subunits because GIRK1 cannot form a functional homotetrameric channel. Electrophysiological studies demonstrated that stimulation of the cells with the GABAB agonist baclofen increased K+ flow from the cells, confirming that the GABAB receptors had formed a functional complex with GIRK1/2 channels in the cells. Exposure of the cells to ML297 also stimulated K+ flow, through its ability to elicit a concentration-dependent increase in the gating of the GIRK1/2 channels. Closer examination revealed that ML297 induced the channels to stay open longer while having no effect on the conductance of the individual channels. In contrast, baclofen increased gating by causing the channels to open more frequently for a short duration. ML297 stimulated GIRK1/2 channel K+ gating a maximum of 8-fold, and the concentration producing 50% of the maximal effect (EC50) was 233 nM. The agonist also stimulated K+ gating in cultured hippocampal neurons, which naturally express GIRK1/2. The EC50 in these cells was 377 nM. On the other hand, ML297 had no effect on K+ gating in HEK-293 cells expressing only GIRK2, even though GIRK2 forms functional homotetrameric channels.
To determine if the effects of ML297 on GIRK1/2 channel function require the presence of Gβγ, the researchers used GRK3ct, the C-terminal fragment of G protein-coupled receptor kinase 3, which acts as a scavenger of Gβγ subunits. Addition of GRK3ct to transfected HEK-293 cells markedly suppressed the response to baclofen while having no significant effect on the response to ML297. These results confirmed that the GABAB receptor couples to GIRK1/2 through Gβγ while also demonstrating that ML297 activates the channels directly, without a requirement for Gβγ binding.
Phosphatidylinositol 4,5-bisphosphate (PIP2) is required for GIRK-mediated K+ conductance under physiological conditions. To investigate the role of PIP2 in ML297-activated GIRK1/2 channel function, the investigators expressed the gene for the Danio rerio voltage-sensitive phosphatase (Dr-Vsp) in their HEK-293 cells. Dr-Vsp hydrolyzes PIP2 in response to a strong depolarization, and the investigators discovered that cells expressing Dr-Vsp exhibited a marked decrease in their responsiveness to ML297 following exposure to depolarization. In contrast, depolarization had no effect on the ability of cells not expressing Dr-Vsp to respond to ML297. The results confirmed that PIP2 must be present for ML297 to stimulate GIRK1/2 channel function.
Having established that ML297 stimulates GIRK1/2 channel function through a PIP2-dependent, G protein-independent mechanism, the investigators wished to determine how ML297 interacts with GIRK1. To achieve this goal, they used HEK-293 cells expressing GIRK2 and a series of chimeras of GIRK1 subdomains on a GIRK2 background. They then evaluated ML297-evoked GIRK activation in the cells using a high-throughput thallium flux assay. All of the cells formed functional GIRK channels, but the only ones that responded to ML297 contained chimeras that included the P-M2 domain of GIRK1. This domain includes the pore helix/potassium selectivity filter and the second membrane-spanning domain. Next, they used site-directed mutagenesis to explore the role of individual P-M2 domain amino acids in the response to ML297. The results revealed that phenylalanine-137 and aspartic acid-173 are both required for sensitivity to ML297 (Figure 4). Mutation of either of these amino acids in GIRK1 produced a protein that was nonresponsive to ML297 stimulation. In contrast, mutation of the two comparable amino acids in GIRK2 to phenylalanine and aspartic acid, respectively, generated a GIRK2 subunit that could pair with wild-type GIRK2 to create an ML297-responsive channel.
Figure 4. (a) Crystal structure of the GIRK2 homodimer bound to Gβγ. Binding sites for PIP2, alcohol, and Na+ are indicated. (b) Enlargement of the region of the GIRK channel transmembrane domain identified as the interaction site for ML297. The locations of the two key amino acids, D173 and F137, are indicated in yellow. The pore helix (containing F137) and second membrane-spanning domain (M2, containing D173) are shown. Three potassium ions (green) mark the location of the channel through which the ions cross the membrane. Image reproduced by permission from N. Wydeven et al., (2014) Proc. Natl. Acad. Sci. U.S.A., published online July 7, DOI:10.1073/pnas.1405190111. Copyright 2010, N. Wydeven, et al.
Prior work had shown that ML297, at doses of 60 mg/kg, suppresses seizure activity in mouse models of epilepsy. The researchers extended their evaluation of the pharmacologic effects of the GIRK1-selective agonist using multiple behavioral tests in mice. They determined that ML297 suppresses motor activity at the 60 mg/kg dose, but not at doses of 30 mg/kg and lower. At these doses, the compound also exhibited no reinforcing effects or anti-depressant activity. However, it did demonstrate anti-anxiety activity in an elevated plus maze test (Figure 5) and in a stress-induced hyperthermia test. Mice genetically lacking GIRK1 (Girk1-/-) did not respond to ML297 in these tests. The results suggest that ML297 is a promising lead molecule for an anti-anxiety drug that lacks addictive liability and does not suppress motor function. Further work will clearly be required to test this hypothesis. In the interim, ML297 will continue to serve as a valuable research tool to define the role of GIRK1-containing channels in health and disease.
Figure 5. Schematic of the elevated plus maze. The structure contains both open and enclosed arms. Mice are naturally averse to open spaces and will spend most of their time in the enclosed arms. Under the influence of an effective anti-anxiety drug the mouse spends more time in the open arms of the maze. Image reproduced from Wikimedia Commons under the Creative Commons Attribution-ShareAlike 3.0 Unported licence.