Lester/Davidson Lab 1997 Annual Report

Examining the Physiological Role of Ion Channels: Generation of a "Knock-In" Mouse with Point Mutations in Neuronal Nicotinic Acetylcholine Receptors

Mark W. Nowak, Melihat Fidan-Nowak, Purnima Deshpande, Cesar Labarca, Hong Dang and Paulo Kofuji

In an effort to understand the physiological role of the neuronal nicotinic acetylcholine receptors (nAChR) we will generate "knock-in" mice containing well-characterized point mutations in the various nAChR. We have chosen as our first target the 4 subunit. Frontal lobe autosomal dominant nocturnal epilepsy has, in part, been shown to result from a point mutation in the 4 subunit such that the receptor becomes less active. In addition, the 4 subunit in association with the 2 subunit forms the high affinity nicotinic binding site in the brain, and most likely is involved in nicotine addiction.

The type of point mutation that we have generated will render the neuron expressing the mutant 4 subunit "hyperexcitable". That is, the mutant receptor has an increased open time upon activation leading to prolonged depolarization of the neuron. Mutations of this type are found in the muscle nAChR in the degenerative disease Slow Myasthenia Channel Syndrome. In this disease mutations are primarily found in the M2 transmembrane domain which forms the ion channel pore. We have extensively examined the effects of ion channel pore mutations in the mouse muscle nAChR in the Xenopus oocyte expression system.

Using the information from our studies on the mouse nAChR we have generated the following mutants in the ion channel pore of the rat 4 subunit: Leu9'Ser, Leu9'Ala, Thr12'Pro, Val13'Met, Leu17'Ser and Leu17'Ala. Oocytes were coinjected with the mutant 4 subunit along the 2 subunit, and the effects of the mutations examined. Mutant 4:2 receptors displayed increases in acetylcholine (ACh) sensitivity 3-100-fold.

We focused on the Leu9'Ala and Leu9'Ser mutations which increased ACh sensitivity 10- and 100-fold, respectively. From our studies on the mouse nAChR we know that the increased ACh sensitivity results, in part, from an increase in the open time of the receptor. Voltage-jump relaxation measurements for these two mutant receptors indeed showed pronounced increases in the relaxation time constants. In addition, these mutations increased the oocyte leak current which was blocked by open-channel blockers. This observation is consistent with spontaneous receptor activity. Based on these findings the 4 Leu9'Ala and Leu9'Ser mutations should render a neuron "hyperexcitable".

The M2 transmembrane domain is located in exon 5 of mouse 4. Following the sequencing of a genomic clone containing exon 5 the Leu9'Ala and Leu9'Ser mutations were made. We are currently inserting into this construct a cassette containing a neomycin resistance gene and the Herpes thymidine kinase gene for selection purposes.

Hyper-active 5HT3 receptor knock-in mice

Hong Dang, Cesar Labarca, Purnima Deshpande, Nicolas Guy¹, David Julius¹

¹University of California at San Francisco

To study how ion channels function in the brains of intact animals, we are introducing a hyper-active 5HT3 channel into the native genetic environment of mice by gene-targeting (a "knock-in" strategy, in contrast to a "knock-out" strategy where a native gene is inactivated). The choice of this particular ligand-gated ion channel was based on (1) the likelihood that this receptor is encoded by a single gene (there is no genetic evidence for other subunits, thus little chance for compensation from other subunits) and (2) the low calcium selectivity in vitro of the 5HT3 receptor, which indicates that it is unlikely to be cytotoxic. Heterologous expression in Xenopus oocytes were used to examine shifts of the receptor's sensitivity to serotonin. Two point mutations (a valine to either serine or alanine) in the M2 channel-lining domain were found to decrease the EC₅₀ values by both 100- and 60-fold.

The same mutations were made in a piece of genomic DNA containing the exon coding for the M2 region. A Neo-TK (neomycin resistant gene-thymidine kinase gene) cassette flanked by two loxP sequences was inserted in the nearby intron to provide both positive (Neo) and negative (TK) selection markers. This DNA will then be introduced into mouse embryonic stem (ES) cells. One copy of the wild-type exon replaced by the mutated DNA through homologous recombination, will be selected by G418 resistance and screened by PCR and Southern blotting. The removal of the selection cassette from the genomic DNA will be achieved through negative selection (FIAU) against the TK gene after transient transfection with the Cre gene, whose product, CRE recombinase, deletes DNA sequences between two similarly oriented loxP units, leaving a foot-print of 34 bps in the intron.

Clones of ES cells containing the intended mutations will be used to generate mice carrying the above mutations. These animals may then be analyzed to identify specific alterations associated with a mutation.

Couplings Between Distant Regions in the Nicotinic Acetylcholine Receptor: Acidic Residues Near the Agonist Binding Site and Hydrophobic Residues in the Pore

Wenge Zhong^*, Haiyun Zhang¹, Dennis A. Dougherty^*

^*Division of Chemistry and Chemical Engineering

¹Division of Physics, Mathematics and Astronomy

Conventional and nonsense suppression mutagenesis methods and double mutant cycles were employed to study functional interactions within the muscle nicotinic acetylcholine receptor expressed in Xenopus oocytes. We measured macroscopic dose response relations for receptors with natural and unnatural mutations at tyrosine 93 in the a subunit (aY93) and/or D (aspartic acid) to N (asparagine) mutations at 174 in the g and 180 in the d subunits. Comparison among the EC₅₀ values for these receptors leads to the conclusion that there is no hydrogen bond between aY93 and gD174 or dD180. However, these D-to-N mutations decrease the receptor's sensitivity to acetylcholine (ACh). Single-channel analysis indicates that the D-to-N mutations affect gating of the channel primarily by a decrease in the opening rate, in addition to a moderate increase in the channel closing rate. The D-to-N mutations were also studied in combination with mutations from wild-type leucine to serine at the 9' position of the M2 pore-lining region. The opposing effects on EC₅₀ values are not all multiplicative, suggesting remote energetic couplings between the acidic residues and the 9' leucine residues, particularly in the b subunit. These acidic residues participate in the intramolecular transduction pathway that couples agonist binding to channel gating.

Leucine Mutations at the 9' Position of the Acetylcholine Receptor M2 Domain: Kinetic Bases of the Changes in Channel Gating

Haiyun Zhang^*, Purnima Deshpande, Cesar Labarca

^*Division of Physics, Mathematics and Astronomy

In nicotinic acetylcholine receptors (nAChR), a highly conserved leucine residue occurs in the pore-lining M2 region at the 9' position and is thought to comprise at least part of the gating mechanism. Previous results show that mutation of this residue to serine produces a leftward shift of ~ tenfold in the ACh dose-response relation for each subunit mutated. We expressed nAChRs with varying numbers (m*_s) of Leu9'Ser mutant subunits in Xenopus oocytes and studied both macroscopic voltage-jump relaxations and single-channel properties in outside-out patches. Mutated AChRs have (1) slower voltage-jump relaxations (~ 200-fold greater time constant than wild type), (2) longer ACh-induced openings and bursts, (3) briefer closed times, and (4) more frequent spontaneous openings. These effects increase with m*_s. For example, average channel open times for wild-type a₂bgd and a^*₂b^*gd^* (m*_s = 4) receptors are 1.8 0.1 ms and 28 3 ms, respectively. The spontaneous openings, however, have an average open time of 0.3 0.1 ms, which remains constant for all m*_s. In addition, the spontaneous opening frequency increases ~ 10²-fold from a₂bgd^* (m*_s = 1) to a^*₂b^*gd^* (m*_s = 4). In order to study the consequence of the Leu9'Ser mutations for synaptic function, synthesized postsynaptic currents were produced with brief ACh pulses delivered to outside-out patches. The currents decay as expected from channel burst duration. For instance, the a^*₂b^*gd^* receptor (m*_s = 4) displays a time constant of 1340 200 ms ( 600-fold greater than for the wild-type receptor). Thus, (1) both longer and more frequent openings contribute to the 10⁴-fold shift in EC₅₀ between the wild-type receptor and receptor with m*_s = 4, and (2) the highly conserved 9' leucine is crucial for appropriately brief synaptic events.

The Role of Hydrophobic Residues in the M2 Domain of the AChR in Channel Gating

Cesar G. Labarca, Haiyun Zhang, Purnima Deshpande

The M2 transmembrane domain of the AChR has numerous hydrophobic residues, mainly Leu and Val. We previously showed that mutation of Leu 9' to Ser altered the gating of the channel. The presence of each mutated subunit decreased the EC50 by roughly a factor of 10, and the effect was nearly independent of the subunit mutated. The frequency of spontaneous openings increased with the number of mutated subunits. We have now examined the effect of mutating hydrophobic residues to Ser in M2 at positions other than 9'. In receptors with a subunit mutations at positions 7', 11', 13', 15', 16' and 17', the dose response relation for ACh shifts to the left; the largest shift occurs at position 13', with an EC50 over three orders of magnitude lower than wild type. In a helical wheel representation, most of these residues cluster on two opposing faces and orthogonal to the polar 6' and 10' residues that face the open lumen, as though these hydrophobic positions face the flanking M2s of neighboring subunits. In contrast, receptors with 8' mutations show a rightward dose-response shift; these residues would face away from the open pore. Thus, gating is influenced by multiple hydrophobic interactions (a) between the M2 transmembrane helices surrounding the ion pore and (b) between the M2 helices and the distal surrounding domains of the receptor. We propose a model for AChR gating.

Site-Specific, Nitrobenzyl-Induced, Photochemical Proteolysis Applied to Ion Channels In Vivo

Pamela M. England, Dennis A. Dougherty^*

^*Division of Chemistry and Chemical Engineering

A method for site-specific, nitrobenzyl-induced photochemical proteolysis (SNIPP) of diverse proteins expressed in living cells has been developed based on the chemistry of the unnatural amino acid (2-nitrophenyl)glycine (Npg). Using the in vivo nonsense codon suppression method for incorporating unnatural amino acids into proteins expressed in Xenopus oocytes, Npg has been incorporated into two ion channels: the Drosophila Shaker B (ShB) K⁺ channel, and the nicotinic acetylcholine receptor (nAChR). Functional studies in vivo show that photolysis of a protein containing an Npg residue does lead to peptide backbone cleavage at the site of the novel residue. Using this method, evidence is obtained for an essential functional role of the ÒsignatureÓ Cys128-Cys142 disulfide loop of the nAChR a subunit.

Unnatural amino acid mutagenesis: incorporation of a caged tyrosine in the a subunit of the mouse nicotinic acetylcholine receptor

Jeffrey C. Miller and Scott K. Silverman¹

¹Division of Chemistry and Chemical Engineering

The ability to incorporate unnatural amino acids in proteins expressed in Xenopus oocytes provides a unique opportunity to specifically place sidechains with novel behaviors within proteins in living cells. We have placed a photolabile nitrobenzyl tyrosine derivative, tyr(ONB), in the mouse nicotinic acetylcholine receptor a subunit. Wild-type tyrosine was replaced with this residue at a subunit positions 93, 127, 190, and 198. With the aid of modified electrophysiological recording equipment, channel currents from oocytes expressing the mutated receptors were recorded in real time as the tyrosines were uncaged by 1 ms pulses from a 250 J per pulse flash lamp. Before the tyr(ONB) is uncaged the oocytes show only small responses to acetylcholine; they show much larger responses after light-induced uncaging of tyr(ONB). When flashes are delivered to the oocytes in the presence of acetylcholine, the time constant for the current increases may be measured. The time constant varies both with the site at which the caged tyrosine was incorporated and with the pH of the recording solution. Time constants ranging from 20 ms to several seconds have been observed. The underlying explanation for this variation is under investigation.

The possible uses of the tyr(ONB) residue and similar residues such as caged serine and caged phosphoserine include probing the time course of signaling events in neuroreceptors with millisecond time resolution. In favorable cases, caged sidechains might be used to probe the pKa of residues within some membrane proteins.

Unnatural Amino Acids for In Vivo Site-Specific Labeling of Membrane Proteins

Justin P. Gallivan¹, Dennis A. Dougherty¹

A key structural issue for all membrane proteins is the transmembrane topology. Defining which regions of a protein are intracellular, extracellular or embedded in the membrane is not only essential for determining how a protein folds, but also how it functions. There are relatively few methods to determine the transmembrane topology of functional proteins expressed in living cells, and most involve the introduction of major structural perturbations. To address this issue, we have developed a new experimental method to determine the surface accessibility, and ultimately the transmembrane topology, of membrane proteins expressed in Xenopus oocytes.

We have used the in vivo nonsense suppression technique to incorporate biotinylated unnatural amino acids into functional ion channels expressed in Xenopus oocytes. Binding of ¹²⁵I-streptavidin to biotinylated receptors was used to determine the surface accessibility and thus the topology of individual amino acids. We have incorporated biotinylated amino acids into sites in the main immunogenic region (MIR) of the nicotinic acetylcholine receptor and assayed their accessibility to streptavidin. Biocytin is efficiently incorporated into five sites in the MIR, and extracellular streptavidin binds to one residue in particular, 70, thus establishing its position on the receptor surface.

We believe that the incorporation of biotinylated amino acids represents a potentially general method for determining the surface accessibility and topology of membrane proteins. In work over the next year, we hope to use the method to experimentally determine the transmembrane topology of other ion channels and neurotransmitter transporters.

Role of Conserved Proline in Transmembrane Domain M1 of Ligand-Gated Ion Channels

Hong Dang and Justin Gallivan

A conserved proline residue is found in the first transmembrane domain of every subunit of the ligand-gated ion channel super-family. The location of this conserved proline residue, between the extracellular ligand-binding domain and the main channel-lining domain (M2) in the primary sequence, hints at its involvement in the transduction of ligand-binding signals to the gating of the ion channel. Preliminary study--by conventional mutagenesis of the a7 homomeric nicotinic receptor expressed in Xenopus oocytes--indicates that replacing the conserved proline residue with either glycine, threonine, or phenylalanine results in a reduced number of surface antagonist (alpha-bungarotoxin) binding sites and a complete inactivation of nicotine currents. Proline is the only imino acid naturally incorporated de novo into proteins. Its a-nitrogen is part of a pyrrolidine ring, which places unique constraints on the polypeptide backbone around it. The unique nature of proline and the degree of conservation of this residue in ligand-gated ion channels suggest that replacement with any of the other 19 amino acids by conventional mutagenesis would represent too drastic a change from the native conformations that would render the receptor inactive.

The stop codon suppression method is well-suited for introducing subtle changes into this conserved proline residue. This approach allows the incorporation of synthetic amino acid analogs at a specific site on a given protein The main problem we encountered in using the stop codon suppression approach was that the injected aminoacyl-tRNA was inefficient in suppressing the normal termination process at a particular site. To evaluate the feasibility of such an approach for the homomeric receptors expressed from oocytes, a stop codon (TAG) (replacing the one for M1 proline) was introduced into cDNA clones of the a7 nicotinic and the 5HT3 receptors. Co-injection of the TAGed mRNA and prolyl-tRNA into oocytes produced no detectable nicotinic current for the a7 receptor, but yielded a whole cell current of about 200-400 nA for the 5HT3 receptor at a saturating serotonin concentration. By introducing proline analogs and characterizing the resultant receptor physiology in oocytes, the role of this conserved residue in the function of the ligand-gated receptors may be elucidated.

Studying modulation of the voltage-gated potassium channel Kv1.3 with unnatural amino acid probes

Yanhe Tong

Potassium channels play a significant role in setting the membrane potential as well as in setting the frequency and duration of action potentials in living cells. Voltage-gated potassium channels are a subfamily of potassium channels which open following membrane depolarization. Recent work has demonstrated that the voltage-gated potassium channel Kv1.3 can be phosphorylated at multiple sites not only by serine/threonine protein kinases, but also by tyrosine kinases. Using the newly developed nonsense codon suppression method, "caged" tyrosine, serine and threonine can be incorporated into putative phosphorylation sites of the channels expressed in Xenopus oocytes. In short, the specific phosphorylation sites will be protected from kinases by a photolabile protecting group. The phosphorylation of the channel will begin only after photolytic deprotection. The rate of deprotection can be controlled by the light intensity, and it is possible to achieve deprotection in ~2 s. Therefore, modulation of the channel kinetics can be measured during the phosphorylation process. Coexpression of Kv1.3 containing the caged side chain together with various protein kinases or lipases or activating endogenous kinases in oocytes expressing unnatural amino acid incorporated Kv1.3 will provide information about the phosphorylation profile of the channel. This study not only will help us to understand the mechanisms of channel regulation by neurotransmitters, hormones and the second messenger system, but also will establish a new method for studying phosphorylation mechanisms of ion channels, receptors, and other cytosolic proteins. The system generated in this study will also serve as a set of valuable tools for studying various signaling pathways.

Expression and Circular Dichroism Studies of the Extracellular Domain of the a Subunit of the Nicotinic Acetylcholine Receptor

Anthony P. West, Jr.*, Pamela J. Bjorkman, Dennis A. Dougherty*

*Division of Chemistry and Chemical Engineering

To provide material suitable for structural studies of the nicotinic acetylcholine receptor, we have expressed and purified the Nterminal extracellular domain of the mouse muscle a subunit. Several constructs were initially investigated using Xenopus oocytes as a convenient small scale expression system. A fusion protein (a210GPI) consisting of the 210ÊNterminal amino acids of the a subunit and a glycosyl-phosphatidylinositol anchorage sequence conferred surface abungarotoxin binding in oocytes. Coexpression of a210GPI with an analogous construct made from the d subunit showed no evidence of heterodimer formation. The a210GPI protein was chosen for large scale expression in transfected Chinese hamster ovary cells. The a210GPI protein was cleaved from these cells and purified on an immunoaffinity column. Gel and column chromatography show that the purified protein is processed as expected and exists as a monomer. The purified protein also retains the two distinct, conformation-specific binding sites expected for the correctly folded a subunit. Circular dichroism studies of a210GPI suggest that this region of the receptor includes considerable bsheet secondary structure, with a small proportion of ahelix.

Deactivation Kinetics of GIRK Currents are Accelerated by RGS Proteins

P. Kofuji and C. A. Doupnik

One well described aspect of native G protein-activated inwardly rectifying K⁺ channels (GIRK) is their fast activation and deactivation kinetics with step applications of agonists for G protein-coupled seven helix receptors. In atrial cells, acetylcholine (ACh)-evoked GIRK currents deactivate in < 1s with rapid removal of ACh, presumably due to the rapid dissociation or removal of G dimers from the channel. The off rate of the muscarinic response is much faster than the intrinsic GTP hydrolysis rate of Gi or Go subunits measured in vitro (2-5/min), suggesting GIRK channels may possess an intrinsic GTPase activating protein (GAP) activity similar to other G protein effector molecules. We investigated the molecular determinants of GIRK deactivation by coexpression of various components of this signalling pathway in oocytes and CHO cells. In CHO cells transfected with plasmids encoding GIRK1, GIRK2 and the m2 type muscarinic receptor, application of 1 M ACh induced GIRK currents that activated rapidly (_act ~ 500 ms), but deactivated slowly (_deac = 5-13 s) compared to atrial cells (though this rate is similar to the GTPase rate of Ga subunits). The ACh-evoked currents were abolished by pertussis toxin pretreatment, indicating the involvement of Gi or Go proteins. Similar slow deactivation of GIRK currents is seen in oocytes. We investigated the effect of the newly described class of proteins, "regulators of G protein signalling" (RGS), which accelerate the in vitro GTPase activity of Gi and Go proteins via a direct interaction with the G proteins (Berman et al, Cell 86: 445, 1996). Coexpression of the RGS4 isoform with GIRKs and the m2R in CHO cells, led to expression of GIRK currents that now deactivated with a time constant (_deac = 0.4 to 1.5 s) comparable to that measured for atrial cells (_deac = 0.5 to 1 s). We conclude that GIRK channels do not act as GAPs as previously suggested; instead, RGS proteins exert this rolein vivo.

A C-terminal peptide of the GIRK1 subunit directly blocks the G protein-activated K⁺ channel (GIRK) expressed in Xenopus oocytes.

Tudor Luchian¹, Nathan Dascal², Carmen Dessauer³ , Dieter Platzer¹, Alfred Gilman³, & Wolfgang Schreibmayer.

¹Department of Medical Physics and Biophysics, University of Graz, Harrachgasse 21/4, A-8010 Graz, Austria;

²Department of Physiology and Pharmacology, Sackler School of Medicine, Tel Aviv University, Ramat Aviv, Israel;

³Department of Pharmacology, Uiniversity of Texas-Southwestern Medical School, Dallas TX

The cytosolic C-terminal region of the G-protein activated, inwardly rectifying potassium channel subunit GIRK1 has been proposed to form a blocking gate that may keep the channel closed in the absence of activation by the G_bg subunits. As an initial test of this hypothesis, we expressed GIRK channels (of GIRK1/GIRK5 and GIRK1/GIRK4 subunit composition) in Xenopus laevis oocytes, and we studied effects of two synthetic peptides in isolated inside-out membrane patches. The GIRK channel was activated by bath application of a non-hydrolysable analogue of GTP, GTP-g-S, or of the purified recombinant G_b1g2 subunits. A peptide (DS6) derived from a sequence very near the end of the C terminus of GIRK1 reversibly blocked GIRK activity by 83.4%3.1% (n=10) at 100 mg/ml. IC₅₀ of the DS6 block was 7.92.0 or 3.50.5 mg/ml, and the Hill coefficient was 1.20.3 or 0.750.08, for GIRK channels of GIRK1/GIRK5 or GIRK1/GIRK4 composition, respectively. Application of a control peptide derived from the cytoplasmic N-terminal part of GIRK1 (DS4) at a concentration of 100 mg/ml was without effect. Dose dependency of GIRK activation by purified G_bg showed that DS6 significantly reduced the maximal achievable open probability but did not significantly alter the EC₅₀ for activation by G_bg. Thus, DS6 block of GIRK channels does not result from competition between the peptide and functional GIRK channels for the available G_bg. Single-channel analysis of DS6 block revealed that burst duration was reduced whereas long closed times between bursts were markedly increased, accounting for the observed inhibition, while other gating parameters remained virtually unchanged. Analysis of voltage dependence of GIRK block by the DS6 peptide revealed that inhibition depends slightly on voltage, increasing at more negative potentials. Our data support the hypothesis that the distal part of the carboxy-terminus of GIRK1 is a part of the intrinsic gate that keeps GIRK channels closed in the absence of G_bg.

A Regenerative Link in the Ionic Fluxes through the weaver GIRK2 and GIRK4 Channels Underlies the Pathophysiology of the Mutation

Scott K Silverman*, Paulo Kofuji, Dennis A. Dougherty*

*Division of Chemistry and Chemical Engineering

The homozygous weaver mouse displays neuronal degeneration in several brain regions. Previous experiments in heterologous expression systems showed that the GIRK2 (G protein-gated inward rectifier K⁺) channel bearing the weaver pore-region GYG to SYG mutation (1) is not activated by G_bg subunits, but instead shows constitutive activation; and (2) is no longer a K⁺-selective channel, but conducts Na⁺ as well. The present experiments on weaverGIRK2 (wvGIRK2) expressed in Xenopus oocytes show that the level of constitutive activation depends on intracellular Na⁺ concentration; in particular, manipulations that decrease intracellular Na⁺ produce a component of Na⁺-permeable current activated via a G protein pathway. Therefore, constitutive activation may not arise because the weaver mutation directly alters the gating transitions of the channel protein; instead, there may be a regenerative cycle of Na⁺ influx through the wvGIRK2 channel, leading to additional Na⁺ activation. We also show that the wvGIRK2 channel is permeable to Ca²⁺, providing an additional mechanism for the degeneration that characterizes the weaver phenotype. We also demonstrate that the GIRK4 channel bearing the analogous weaver mutation has properties similar to those of the wvGIRK2 channel, providing a glimpse of the selective pressures that have maintained the GYG sequence in nearly all known K⁺ channels.

Integrin Interactions with GIRK Channels

Jancy C. McPhee and Yan Dang

A three amino acid motif, arginine-glycine-aspartate (RGD), just extracellular to the first transmembrane domain, is well-conserved on all known G-protein gated inward rectifier channel (GIRK) subunits. This sequence is the same as the ligand frequently recognized by another family of transmembrane proteins called integrins. Integrins often recognize an RGD sequence found on extracellular matrix molecules; accordingly, they have been associated with functions such as cell-substrate adhesion or cell-cell adhesion. Interestingly, they have also been implicated in neuronal excitability in the hippocampus, neurite outgrowth, and stretch-induced neurotransmitter release. It is conceivable, therefore, that integrins may regulate ion channel activity. Because of the presence of the classic integrin ligand recognition site, RGD, on all GIRK subunits, we decided to examine whether GIRK activity, localization or expression is affected directly by interaction with integrin.

For our first approach, we injected RNA into Xenopus laevis oocytes to express the muscarinic acetylcholine receptor and a combination of GIRK1 and GIRK4 subunits. We compared the expression of WT and mutant subunits with the extracellular GIRK RGD site mutated to RGE. This type of mutation should disrupt GIRK channel-integrin interactions, if they exist. The GIRK1-RGE/GIRK4-RGE double mutant complex produced no detectable current. The GIRK1-RGE/GIRK4 complex consistently expressed at a level 3-4 times lower than that produced from the same amount of WT GIRK1/GIRK4 RNA. The GIRK1/GIRK4-RGE complex expressed at about 10 fold lower levels than the double WT GIRK subunit complex. Voltage jump relaxations steps clarify that, in the one mutant/one WT subunit combinations described above, the mutant subunits were actually being expressed and that observed current was not merely due to the expression of the WT subunit. We were intrigued that a direct interaction of GIRK and integrin may occur and that this interaction might regulate the amount or timing of GIRK expression or localization in oocytes. Comparison of the total protein level of the WT and mutant channel constructs using Western blot analysis revealed that the mutant constructs were expressed at similar levels to the WT. Preliminary experiments isolating plasma membrane fractions from oocytes expressing these channels suggested that the mutant channels were located in the plasma membrane to the same extent as WT. Therefore, the mutation of the RGD sight seems to decrease GIRK current without eliminating channel protein synthesis or localization to the plasma membrane in oocytes. To directly show that the channel interacts with integrin, we are currently performing co-immunoprecipitation experiments to show that expressed WT GIRK channel and the endogenous Xenopus integrin form a complex.

In the near future, we plan to shift our analysis to native cells that contain GIRK channels, such as atrial cells. Co-immunoprecipitation experiments can be performed with these cells, and whether application of an RGD containing peptide can compete for integrin-GIRK channel interactions and alter GIRK activity will also be explored. Expressing WT and mutant channels in mammalian cells by transfection of CHO cells is also in our future plans. Using this system, we hope to get a more detailed analysis of GIRK function and localization than we could in oocytes, and, hopefully, elucidate the role of integrin in the regulation of GIRK activity.

1. Cyclic nucleotide-gated channels: structural basis of ligand efficacy

Jun Li, William N. Zagotta¹

¹. Department of Physiology and Biophysics

Howard Hughes Medical Institute

University of Washington School of Medicine

Most proteins exhibit allosteric coupling, which can be defined as indirect interactions between physically remote sites. Hemoglobin and nicotinic acetylcholine receptors are the exemplar proteins traditionally used in the study of allosteric mechanisms. Recently, however, the cyclic nucleotide-gated (CNG) channels have emerged as another excellent model system. This is because CNG channels do not desensitize and the atomic structures of some of their homologous proteins, such as E.coli CAP, are readily available. In a reexamination of structure-function studies concerning binding and gating in CNG channels, we draw attention to some mechanistic implications not yet seriously considered before.

We showed that each of the three parameters in the Monod-Wyman-Changeux (MWC) model, K, f, and L₀, specifies one of the three basic processes that comprise the ubiquitous, ligand-mediated conformational changes in proteins: the initial binding (K), the spontaneous conformational changes (L₀), and the amount of free energy transferred from additional binding to the stabilization of conformational changes (f). f thus represents ligand efficacy. It is a measure, on the one hand, of the increase in binding strength due to the conformational changes and, on the other hand, of how much of the free energy released by the additional binding contributes to the conformational changes. We then reviewed studies which have revealed structural motifs responsible for these three functional aspects. Finally, we proposed that most of the documented functional modifications can be classified into the K types, f types, or L₀ types. Using CNG channels as an example, we applied this classification scheme to natural and recombinant mutations, as well as to many modes of channel modulation.

The MWC model is a basic model for CNG channels: it captures key features of allosteric transitions. In applying this model, however, one almost always has to elaborate on it in order to accommodate additional properties, much in the same way that the ideal gas laws must be modified to better characterize real-life gaseous matters.

2. Single-channel analysis of the activation mechanism of CNG channels

Jun Li

According to one hypothesized mechanism for the activation of cyclic nucleotide-gated (CNG) channels, based on the Monod-Wyman-Changeux model, in the absence of ligand there is a small but finite opening probability (L₀); each additional binding stabilizes the open state by an additional factor (f), such that a fully liganded channel is favored to open (L₀f⁴>>1). This simple model assumes complete symmetry among binding sites. But an actual CNG channel may not display such ideal symmetry: the binding events may have a range of affinities or contribute a range of stabilization factors. We examined the openings of a single, heterologously expressed, homomeric CNG channel under steady-state conditions, with the hope that the statistical properties of the channelÕs behavior may help to distinguish among these possibilities.

We used the cloned rat olfactory CNG channel expressed in Xenopus oocytes. The single- channel openings show neither subconductance nor severely flickery openings. Most of the openings last between 1 and 500 ms. At higher ligand concentrations longer openings occur more frequently and briefer openings decrease in frequency. These properties simplified the analysis, especially in comparison with other CNG channels which show fast flicker (<<1 ms) or more than one conductance level. However, several technical/analytical difficulties did arise: (1) endogenous stretch-activated channels are very similar to the expressed CNG channels and were present in most patches, (2) the openings do not show obvious bursting patterns, which are predicted by some common models and widely observed among ligand-gated channels, and (3) even at a steady level of ligand the opening probabilities often change with time (i.e., 20%-80%), suggesting spontaneous shifts in binding affinity, gating mode, or both. Preliminary analysis indicates that some asymmetries among binding events definitely exist. Thus, the original, three-parameter MWC model must be extended to describe the activation of the rat olfactory CNG channel.

Second Messengers, Trafficking-Related Proteins, and Amino Acid Residues that Contribute to the Functional Regulation of the Rat Brain GABA Transporter GAT1

Michael W. Quick¹, Janis L. Corey²

¹Department of Neurobiology, University of Alabama at Birmingham, Birmingham AL 35294-0021

²SIBIA, 505 Coast Boulevard South, La Jolla CA 92037-4641

Recent evidence indicates that several members of the Na⁺-coupled transporter family are regulated and this regulation in part occurs by redistribution of transporters between intracellular locations and the plasma membrane. We elucidate components of this process for both wild-type and mutant GABA transporters (GAT1) expressed in Xenopus oocytes using a combination of uptake assays, immunoblots, and electrophysiological measurements of membrane capacitance, transport-associated currents, and GAT1-specific charge movements. At low GAT1 expression levels, activators of PKC induce redistribution of GAT1 from intracellular vesicles to the plasma membrane; at higher GAT1 expression levels, activators of PKC fail to induce this redistribution. However, co-injection of total rat brain mRNA with GAT1 permits PKC-mediated modulation at high transporter expression levels. This effect of brain mRNA on modulation is mimicked by co-injection of syntaxin 1a mRNA, and is eliminated by injecting synaptophysin or syntaxin antisense oligonucleotides. Additionally, botulinum toxins, which inactivate proteins involved in vesicle release and recycling, reduce basal GAT1 expression and prevent PKC-induced translocation. Mutant GAT1 proteins, in which most or all of a leucine heptad repeat sequence was removed, display altered basal distribution and lack susceptibility to modulation by PKC, delineating one region of GAT1 necessary for its targeting. Thus, functional regulation of GAT1 in oocytes occurs via components common to transporters and to trafficking in both neural and non-neural cells, and suggests a relationship between factors that control neurotransmitter secretion and the components necessary for neurotransmitter uptake.

Amino Acid Residues Responsible for Low-pH Potentiation of the Transport-Associated Current in two Mammalian Serotonin Transporters

Y-W. Cao, S. Mager

Mammalian serotonin transporters (SERT) expressed in Xenopus oocytes display at least two currents uncoupled from substrate flux: the leakage current in the absence of 5-HT and the transport-associated current in the presence of 5-HT. Both currents are due to "channel-like" behaviors in serotonin transporters. We previously reported that acidic pH strongly potentiates the transport-associated current in rat SERT (rSERT) but not in human SERT (hSERT). To understand the mechanism of this low pH potentiation and its possible relevance to channel gating, we performed domain-swap analysis between rSERT and hSERT, followed by site-directed mutagenesis. We found that a single amino acid difference at position 490 (threonine in rSERT and lysine in hSERT, in the extracellular loop between TM9 and TM10) is responsible for the difference in response to low pH. However, since threonine is predominantly protonated already at neutral pH, we propose that the low-pH potentiation of the transport-associated current is due to the protonation of the two neighboring glutamate residues (at position 493 and 494). The positively charged lysine in hSERT would prevent the protonation of these two glutamate residues and thus abolish the low-pH potentiation in hSERT. Consistent with this model, mutations at these two glutamate residues in rSERT also abolished the low-pH potentiation. These results indicate that one or both of the negatively charged glutamate residues may play an important role in a gating process at the serotonin transporter.

Rethinking Neurotransmitter Transporter Topology

N. Yu, Y-W. Cao, S. Mager.

Covalent cysteine-modification was used along with site-directed mutagenesis to study the membrane topology of the GABA transporter GAT1. Treatment of Xenopus oocytes expressing GAT1 with either (2-sulfonatoethyl)methane-thiosulfonate (MTSES) or [2-(trimethylammonium)ethyl]methanethio-sulfonate (MTSET) (3.75 mM, 30 min) significantly decreased the GABA transport current (to 22% and 10% of control, respectively). Substitution of cysteine at position 74 to alanine (C74A) eliminated the sensitivity to these membrane-impermeant compounds, indicating that C74 is exposed to the extracellular side. Furthermore, wild-type (WT) GAT1 treated with MTSET generated large Na⁺ leakage currents that were undetectable with unmodified WT transporters; Li⁺ leakage currents also increased several-fold. MTSET had no effect on either Na⁺ or Li⁺ leakage currents in the C74A mutant. Interestingly, substitution of tryptophan to leucine at position 68 (W68L), a mutation that affects ion binding and permeation, decreased the transporter's sensitivity to both reagents. Effects of C74 modification on ion conductance and altered reactivity observed with the W68L mutant suggest that C74 lies in or near the permeation pathway. These findings, together with recent evidence placing the TM2-TM3 loop on the extracellular side, suggest that the TM2 region does not traverse the membrane. We speculate that TM2 may form a loop structure reminiscent of the cation channel P-loop.

Use of and Precautions with a Codon-Optimized Green Fluorescent Protein as a Reporter Gene in Co-Transfection Experiments

B. M. Sullivan, M. Lanzrein, N. Davidson, M. C. Jasek.

We describe a method of co-transfecting a reporter gene for efficient recognition of cells for single cell electrophysiology. In cultured mammalian cells, a codon-optimized form of GFP, denoted plasmid Green Lantern-1 by the supplier (for which we use the abbreviation GL or pGL), was co-transfected with three additional plasmids, two coding for G-protein gated inwardly rectifying potassium (GIRK) channel subunits and one coding for the muscarinic m2 receptor. Electrophysiological recordings demonstrated a 98% correlation between GL fluorescence and the presence of GIRK currents. The highest expression levels of the gene(s) of interest were obtained when small amounts of GL plasmid were used. The inhibitory effect of coexpression of large amounts of GL plasmid was observed for each of several co-transfected test genes, including two isoforms of nitric oxide synthase and lac-Z. Furthermore, the inhibitory effect was not observed when an equimolar amount of the lac-Z plasmid was substituted for the GL plasmid. Our data show that GL provides most effective, non-interfering recognition of single cells for physiological experiments if GL is used in the lowest amount that gives a signal in a sufficient fraction of cells.

Functionally significant differences between the two isoforms of the nitric oxide synthase that are constitutively expressed in the brain

Brian M. Sullivan and Yong-Xin Li

Nitric oxide (NO) is an important signaling molecule in multiple brain regions. The "neuronal" and "endothelial" isoforms of nitric oxide synthase, which produce nitric oxide in response to calcium/calmodulin binding, are each constitutively expressed in multiple partially-overlapping cell types and regions within the brain. Using adenoviral vectors to induce high-level expression of the neuronal and endothelial isoforms (nNOS and eNOS, respectively) of the nitric oxide synthase in cell lines and primary neuronal cultures, we are addressing the question "Why have two enzymes in the brain that catalyze the same reaction?" Previous work by others has shown that the subcellular localizations of the enzymes are determined by different mechanisms; the eNOS is fatty acylated and the nNOS contains a PDZ domain. We are using immunocytochemistry combined with fluorescence tagging to determine if there are differences in the subcellular localizations of the enzymes in infected cultured hippocampal neurons. We are also using a nitric oxide-sensitive electrode, which is able to discern concentrations of a few nM NO, to monitor nitric oxide production in infected Chinese Hamster Ovary (CHO) cells. This assay will allow us to determine whether the enzymes differ in their ability to release nitric oxide into the extracellular space and whether they have different calcium sensitivities. We have measured NO production in both intact and homogenized infected CHO cells using this electrochemical device. We detect ~ 100 nM NO produced by a 35 mm dish of infected CHO cells in the presence of a calcium ionophore. NO production, as assayed by the electrode, is sensitive to NO-synthase inhibitors. We are using the system to analyze differences between the nNOS and the eNOS. Early results suggest that the enzymes release NO to the extracellular space differently; they may also have different calcium sensitivities.

Activation of GIRKs overexpressed by recombinant adenovirus blocks hippocampal neuron excitability

Markus U. Ehrengruber, Craig A. Doupnik, Youfeng Xu, Justine Garvey, Mark C. Jasek, Norman Davidson

G protein-gated inward rectifier K⁺ channel subunits (GIRK14) have been cloned from neuronal and atrial tissue and function as heterotetramers. To examine the inhibition of neuronal excitation by GIRKs, we overexpressed GIRKs in cultured hippocampal neurons from 18 d rat embryos which normally lack or show low amounts of GIRK protein and currents. Adenoviral recombinants containing the cDNAs for GIRK1, GIRK2, GIRK4, and the serotonin 5HT_1A receptor were constructed. Typical GIRK currents could be activated by endogenous GABA_B, serotonin 5HT_1A, and adenosine A1 receptors in neurons coinfected with GIRK1+2 or GIRK1+4. Under current clamp, GIRK activation increased the cell membrane conductance by 12-fold, hyperpolarized the cell by 1114 mV, and inhibited action potential firing by increasing the threshold current for firing by 23-fold. These effects were not found in non- and mock-infected neurons, and were similar to the effects of muscarinic stimulation of native GIRK currents in atrial myocytes. Two inhibitory effects of GIRK activation, hyperpolarization and diminution of depolarizing pulses, were simulated from the experimental data. These inhibitory effects are physiologically important in the voltage range between the resting membrane potential and the potential where voltage-gated Na⁺ and K⁺ currents are activated, that is where GIRK currents are outward. Adenoviral GIRK overexpression can now be tested for its effect on (hippocampal) synaptic transmission and in model systems of certain biological malfunctions, including epilepsy.

Analysis of gene expression induced by BDNF in the hippocampus

Markus Lanzrein, W. Bryan Smith, Hyejin Kang, Erin M. Schuman, Norman Davidson

Long lasting synaptic potentiation (LTP) in the hippocampus can be divided into an initial phase (1-3 hrs) that relies upon modification of existing proteins and a late phase (>3 hrs) that is dependent on transcription and translation. Thus, new genes are induced during the late phase of LTP which presumably function in the maintenance of the synaptic potentiation. The neurotrophins BDNF and NT3 can induce a long-lasting enhancement in synaptic transmission in the Schaffer collateral - CA1 synapses in the rat hippocampus. This enhancement is also sensitive to inhibitors of protein synthesis and therefore requires de novo synthesis of proteins. For a better understanding of neurotrophin action and synaptic plasticity, it would be interesting to know which genes are induced in the hippocampus upon neurotrophin treatment. We are approaching this problem by constructing cDNA libraries from acute hippocampal slices that have been potentiated by BDNF treatment and from unpotentiated control slices. Due to the limited amount of tissue available, we are employing PCR-based methods to generate library cDNA. Subtractive hybridization is then used to obtain cDNA libraries that are specifically enriched in genes that are induced by BDNF. As a second approach, we are utilizing serial analysis of gene expression (SAGE). This novel technique is based on the analysis of small (9bp) sequence tags that are derived from a given mRNA population. SAGE allows the quantitation of the expression level of each individual mRNA and differences in expression induced by BDNF treatment can be determined.

Presynaptic Modulation of Neurotransmitter Release by BDNF

Yong-Xin Li, Yinong Zhang, Sheri McKinney, Erin Schuman and Norman Davidson

Brain-derived neurotrophic factor (BDNF) enhances neurotransmitter release at central synapses. We investigated the effects of BDNF on E18 hippocampal neurons after ten days in culture. BDNF (100 ng/ml) increased the frequency of excitatory postsynaptic current bursts (EPSCs) to 205 ± 20.8% of the levels found for the control samples and increased the amplitudes of evoked synaptic currents by 45 ± 10% in four out of seven pairs of neurons. However, there were no detectable effects of BDNF on the resting membrane potential (Vm), input resistance (Rin), action potential threshold (AD), action potential amplitude (AP), or amplitude of the current induced by an application of glutamate (I). BDNF also increased the frequency of miniature EPSCs (mEPSCs) (268.0 ± 46.8 % of the levels found for the control samples) in the presence of 50 nM TTX without affecting the mEPSC amplitude. The effect of BDNF on mEPSC frequency was blocked by the tyrosine kinase inhibitor K_252a and also by removal of [Ca²⁺]_OUT. In addition, Fura-2 recordings showed that BDNF elicited an increase in [Ca²⁺]_IN; this effect was also blocked by K_252a and dependent on [Ca²⁺]_OUT.The results demonstrate that BDNF affects synaptic transmission at a presynaptic locus and that this effect is accompanied by a rise in [Ca²⁺]_IN in the signal transduction pathway.

Use of organotypic hippocampal cultures and antisense suppression methods to examine synaptic physiology.

Mark C. Jasek

Examinations of receptor physiology often rely heavily on the use of agonist, antagonist, and other pharmacological agents. In this day of molecular characterization of various classes of receptors and their numerous subtypes, not to mention subunits, it is often the case that selective pharmacological tools are unavailable. This presents a large hurdle in understanding the roles that select receptors and or their subunits play in synaptic physiology. In order to circumvent this problem, we are developing methods to delete functional proteins from functional synapses. Currently, we are testing two basic antisense methodologies in mammalian cell line cultures. The first method uses mammalian expression vectors which generate antisense RNA. This antisense RNA binds to the select RNA of interest and prevents translation of the protein. Once a suitable antisense RNA expressing vector has been developed, this will then be transferred into a viral expression system (Adenovirus) allowing efficient infection of various cells or tissue preparations. The second approach utilizes antisense oligodeoxynucleotides, containing phosphorothioate bonds between bases for increased stability, to hybridize with the native DNA and prevent the transcription of RNA of the protein of interest.

The model of synaptic transmission we will be using is the CA3 projection to CA1 pyramidal neurons in organotypic hippocampal slice cultures. This method, originally developed by B. Gähwiler, involves the mounting of hippocampal slices onto coverslips which are placed into tubes containing culture media and are subsequently arranged in a rotating roller bin. These cultures survive for several weeks, providing us with the time to reduce protein levels of interest using our antisense methods.

Molecules we plan to examine initially include Trk B, metabotropic glutamate, and GIRK receptors.

Back to Table of Contents | Back to the Lester/Davidson Lab Homepage