JBC INTERFERin siRNA transfection reagent

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M306978200 on August 25, 2003

J. Biol. Chem., Vol. 278, Issue 44, 42821-42828, October 31, 2003
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
278/44/42821    most recent
M306978200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Chang, C.-s. S.
Right arrow Articles by Olsen, R. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Chang, C.-s. S.
Right arrow Articles by Olsen, R. W.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

A Single M1 Residue in the {beta}2 Subunit Alters Channel Gating of GABAA Receptor in Anesthetic Modulation and Direct Activation*

Chang-sheng S. Chang{ddagger}, Riccardo Olcese§, and Richard W. Olsen{ddagger}§

From the {ddagger}Department of Molecular and Medical Pharmacology and the §Department of Anesthesiology, David Geffen School of Medicine, UCLA, Los Angeles, California 90095-1735

Received for publication, July 1, 2003 , and in revised form, August 22, 2003.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
General anesthetics allosterically modulate the activity of neuronal {gamma}-aminobutyric acid, type A (GABAA), receptors. Previous mutational studies from our laboratory and others have shown that the regions in transmembrane domain 1 (M1) and pre-M1 of {alpha} and {beta} subunits in GABA receptors are essential for positive modulation of GABA binding and function by the intravenous (IV) general anesthetics. Mutation of {beta}2Gly-219 to Phe corresponded in {rho} nearly eliminated the modulatory effect of IV anesthetics in {alpha}1/{beta}2/{gamma}2S combination. However, the general anesthetics retained the ability to directly open the channel of mutant G219F, and the apparent affinity for GABA was increased, and desensitization rate was reduced. In this study, we made additional single mutations such as 219 Ser, Cys, Ile, Asp, Arg, Tyr, and Trp. The larger side chains of the replacement residues produced the greatest reduction in enhancement of GABA currents by IV anesthetics at clinical concentrations (Trp > Tyr = Phe > Arg > Asp > Ile > Cys > Ser = wild type). Compared with a 2–3-fold response in wild type, pentobarbital and propofol enhanced less than 0.5-fold; etomidate and alphaxalone modulation was reduced from more than 4- to 1-fold in G219F, G219Y, and G219W. A linear correlation was observed between the volume of the residue at position 219 and the loss of modulation. An identical correlation was found for the effect of modulation on left-shift in the GABA EC50 value; furthermore, the same rank order of residues, related to size, was found for reduction in the maximal direct channel-gating by pentobarbital (1 mM) and etomidate (100 µM) and for increased apparent affinity for direct gating by the IV anesthetics.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 
Considerable evidence supports the hypothesis that general anesthesia can be produced, at least in part, by enhancing neuronal inhibition mediated by GABAA receptors (GABAR).1 GABA and most volatile and general intravenous (IV) anesthetics bind to GABAR. These are ligand-gated channels that enhance the flow of chloride ions (Cl–) into the postsynaptic neuron, thus increasing its resting potential and making it less likely to fire, i.e. the inhibitory effect in GABA synapses.

GABAR are modulated positively by a wide variety of structurally diverse general anesthetics (1). In particular, volatile anesthetics (2, 3) and IV anesthetic agents such as the barbiturates (4, 5), propofol (6), etomidate (7), and neuroactive steroid anesthetics, like alphaxalone (8–10), all enhance the function of GABAR at clinically relevant concentrations. In addition, IV and volatile anesthetics can activate GABAR directly in the absence of GABA (11). Anesthetics appear to allosterically modulate ligand-gated ion channel (LGIC) receptor function by binding to the receptor itself, but not at the agonist-binding site, and to affect the conformation of the membrane protein in a functional manner. Exactly how such modulatory sites affect agonist-regulated channel opening is a major contemporary question in pharmacology and neurobiology. Obviously, the structure and function of these integral membrane proteins, including GABAR, are sensitive to the membrane environment and its molecular constituents.

The mammalian GABAR is a heteropentameric complex formed by 19 different glycoprotein subunits ({alpha}1–6, {beta}1–3, {gamma}1–3, {delta}, {epsilon}, {theta}, {pi}, and {rho}1–3) that co-assemble to form a chloride channel (12–14). Most GABAR in vivo are composed of pentameric complexes of 2{alpha},2{beta}, and 1{gamma} subunits. GABAR are members of the LGIC receptor superfamily and share significant amino acid sequence homology with other members (15, 16). Each subunit in the superfamily has a similar proposed three-dimensional structure and membrane topology, with four membrane-spanning {alpha}-helical domains, M1–M4. The M2 transmembrane domain of all five subunits contributes to the channel wall and determines ion permeability, whereas the N-terminal extracellular domain determines the ligand binding specificity. In the pentameric complex of GABAR, there are two putative GABA-binding pockets at the interfaces of {alpha} and {beta} subunits. Amino acids that participate in ligand binding have been identified by photoaffinity labeling and sequencing, as well as by site-directed mutagenesis (17–23). Loops of amino acids that form an acetylcholine-binding pocket were identified in the {alpha} subunit of nAChR (24, 25), and homologous residues involved in GABA binding were identified in {beta} subunits of GABAR (26, 27). Additional amino acids involved in the GABA-binding site in the {alpha} subunit were homologous to residues involved in binding acetylcholine in the {gamma} subunit of nAChR (21). The predicted three-dimensional structure and alignments with the molluscan acetylcholine-binding protein (28) provide models of the domains of ligand binding and the activation of channel gating in GABAR (29), possibly relevant to sites of anesthetic action. In addition, cryo-electron microscopy of crystalline membranes of electric organ containing nAChR (30) have obtained structural information at 4-Å resolution that is totally consistent with previous conclusions about LGIC functional domains including ligand-induced channel gating by agonists and modulators.

Pharmacological subtypes of GABAR exist due to multiple isoforms of varying subunit composition (10, 31, 32). Heterogeneity in modulation of GABAR [3H]flunitrazepam and [3H]muscimol binding is observed for PB, propofol, etomidate, and steroids (8, 10, 14, 33), and similar differences are seen for receptor subtypes in electrophysiological assays (31, 34, 35). In GABAR, a {beta} subunit is essential for receptor membrane expression and anesthetic modulation (1, 8, 14). Wafford and co-workers (31, 34, 36) identified a {beta} subunit-specific interaction, whereas isoforms containing the {beta}2 and {beta}3 subunits are sensitive to the anxiolytic non-benzodiazepine drug loreclezole, the {beta}1 is not. Furthermore, the same {beta} selectivity applied to the anesthetic etomidate, a structural analog of loreclezole, leading to identification of a residue in the {beta} M2 domain important for etomidate action (36, 37).

In contrast to other members of the LGIC superfamily, {rho} subunit GABA receptors are insensitive to nearly all anesthetic compounds, including propofol, barbiturates, volatile anesthetics, and steroid anesthetics (38, 39). The dissimilar pharmacology of {rho} subunit helped identify residues crucial for positive modulation by general anesthetics (40–43). Chimeras, domain-swapping, and point mutations comparing {rho} and non-{rho} subunits allowed us to identify areas required for IV anesthetic modulation within pre-M1 and M1 regions and the identification of specific amino acid residues within {alpha} and {beta} that seem to be necessary for modulation of GABAR by IV general anesthetics. The current study further examines our demonstration (43, 44) that mutation of one specific residue, Gly-219 in the mouth of the M1 region in the {beta}2 subunit, nearly eliminated the enhancement of submaximal GABA-activated currents by IV anesthetics. Additional experiments were performed to test what sort of alterations these specific mutations at {beta}2Gly-219 produced on modulation and direct activation by IV general anesthetics. A preliminary report of this work has appeared (45).


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS
 DISCUSSION
 REFERENCES
 
Molecular Cloning—Rat {alpha}1, {beta}2, and {gamma}2S GABA receptor subunit cDNAs are in the expression vector pBlueScript II SK. For mutants of the {beta}2 subunit in the mature rat protein sequence used in this study, a glycine at position 219 substituted by serine, cysteine, isoleucine, aspartate, arginine, tyrosine, and tryptophan were generated by recombinant PCR using overlapping complementary oligonucleotide methods designed to create a 20-bp overlap at the desired amino acid mutation region. Using a {beta}2 sense oligonucleotide and an antisense oligonucleotide 5'-TGTTTCTTTTCAGCTTAAAACTTAGGGA-3' with {beta}2 cDNA as a template, upstream PCR fragments with deletion of a HindIII site without changing the amino acid residue at position 219 were generated. Simultaneously, by using a sense junction oligonucleotide with a single mutation 5'-TCCCTAAGTTTTAAGCTGAAAAGAAACATTTTCTACTTCATCCTG-3' (for example, G219F) and an antisense oligonucleotide after the termination codon with {beta}2 cDNA as template, downstream PCR fragments with a single mutation were also prepared in separate PCRs. Upstream and downstream PCR fragments were then combined and amplified to create the full-length {beta}2 cDNA fragments that were subcloned into pBlueScript II SK vector for expression in Xenopus oocytes. All mutants were verified by restriction digest (Hind-III used for the first mutation screen, XhoI and NotI used for the orientation of ligation) and double-stranded DNA sequencing using standard techniques.

Expression of Rat GABAA Receptors in Xenopus Oocytes—Capped mRNA (cRNA) transcribing for the wild type and mutant subunits was synthesized by in vitro transcription from ApaI-linearized cDNA construct in pBlueScript II SK using the mMessage mMachine kit (Ambion). cRNA concentrations were calculated by UV absorption and corroborated by comparison to RNA standards on 1% agarose RNA HEPES gels. Oocytes from Xenopus laevis were prepared as described previously (46) and injected with a total volume of 50 nl of cRNA (100–200 pg/nl/subunit) mixed in a ratio of 1:1:2 ({alpha}1:{beta}2:{gamma}2S). Oocytes were maintained in 6-well plates at 17–19 °C in SOS solution (in mM, 100 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, and 5 HEPES), supplemented with 50 µg/ml gentamycin and 100 µg/ml streptomycin and penicillin, and were used for electrophysiological experiments 3–7 days (5–10 days for the better expression of G219R mutant) following injection. The total amount of cRNA was scaled to yield maximal GABA-induced currents of 1.5–2 µA for wild type {alpha}1{beta}2{gamma}2S.

Two-electrode Voltage Clamp Analysis—Oocytes under a two-electrode voltage clamp (47) (voltage hold at –70 mV) were gravity-perfused continuously with ND96 recording solution containing (in mM) 96 NaCl, 2 KCl, 1.8 CaCl2, 1 MgCl2, and 5 HEPES, pH 7.4, at an approximate rate of 5 ml/min. In general, drugs and reagents (GABA, Sigma) were dissolved in ND96. Stock drug solutions (50–100 mM) were made in 10% dimethyl sulfoxide and further diluted in the same medium to the desired concentrations. Maximal current amplitude (1.5–2 µA) for all receptors was elicited by a 30-s application of a saturation concentration of 1 mM GABA, which was measured to be a lower concentration (100 µM) in some mutant receptor combinations. A standard two-electrode voltage clamp recording was carried out using an Axoclamp-2A amplifier (Axon Instruments, Foster City, CA) interfaced to a computer with a DigiData 1322-A device (Axon Instruments). Electrodes were filled with 3 M KCl and had a resistance of 0.5–1.5 megohms. By using a standard similar Hill equation to model the GABA dose responses, data acquisition and analyses were performed using pClamp 8.2, Axo-Data, Axo-Graph (Axon Instruments), Excel (Microsoft), and Prism software (GraphPad). Statistical comparisons of potentiation data employed non-linear least square Excel for multiple independent samples and sigmoid curve fitting of dose-response curves using Equation 1,

(Eq. 1)
where I is the peak amplitude of the current elicited by a given concentration of agonist [A]; Imax is the maximum amplitude of the current; EC50 is the concentration required for half-maximal inhibition, and n is the Hill coefficient. The normalized data are presented as % control to generate concentration-response curves. We calculated that the mutant receptors (G219F: EC50, 15.9 µM; and G219W: EC50, 21.8 µM) show a 3-fold shift in GABA EC50 from wild type receptors (EC50, 45.8 µM).

Anesthetic potentiations of GABA were recorded at EC10 GABA. Potentiation is defined as (I(Anes.+GABA) – I(GABA))/I(GABA) in terms of enhancement fold, where I(Anes.+GABA) is the current response in the presence of the anesthetic pre-perfused for 15 s, followed by anesthetics containing EC10 GABA for 30 s by perfusion or by direct application, and washing out the applied drugs for 1.5–2.5 min (a longer washout time is required for etomidate and alphaxalone application). I(GABA) is the control EC10 GABA-induced current.

The correlation between anesthetic modulation and residue volume were fitted to the one-phase exponential Equation 2,

(Eq. 2)
where Ef is enhancement fold; E0 represents the enhancement for residue-occupied "zero" volume; ER is the correlation versus residue volume; V is the residue volume; and t is rate of enhancement.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
DISCUSSION
 REFERENCES
 
GABA-sensitive Point-mutated Recombinant GABAR—We introduced point mutations Gly-219* located at the entrance of transmembrane domain 1 (M1) into the rat GABAR {beta}2 subunit cDNA (Fig. 1a), including the residue Phe found in the anesthetic-insensitive homomeric {rho}1 subunit (GABAC receptor) (Fig. 1b).



View larger version (36K):
[in this window]
[in a new window]
  		FIG. 1.
a, schematic diagram of a subunit of the GABAA receptor. Topology of a single subunit is highlighted showing the four transmembrane domains (M1–M4) and the location of the mutated amino acids in M1 and (b). Alignment of the amino acid sequences of the pre-M1 and the entrance of M1 domains of rat {alpha}1, {beta}2 GABAA receptor subunits, and human {rho}1 GABAC receptor subunits are shown. The arrow points to the glycine residue (position 223 in {alpha}1 and position 219 in {beta}2), which has been mutated in the {beta}2 subunit to phenylalanine, the corresponding residue found in human {rho}1 subunit (position 261), that is anesthetic-insensitive.

 

The Concentration-dependent Responses of Wild Type and Mutant Receptors—The wild type and point-mutated {beta}2 subunit capped mRNAs (cRNAs) were mixed and co-injected into Xenopus oocytes with rat {alpha}1 and {gamma}2S (short form) subunit cRNAs. The GABA-induced current could be measured after a 3-day incubation. Remarkably, there were no GABA-induced currents for wild type and G219I in response to 100 nM GABA, but there were for G219R, G219F, and G219W. Compared with 100 µM GABA application, the 1 mM GABA could not induce greater currents for G219F and G219W (in this experiment, rapid desensitization was elicited at a high concentration of 1 mM GABA for G219F and G219W); however, we could observe greater 1 mM GABA-induced currents than 100 µM for wild type, G219I, and G219R (Fig. 2a). The EC50 value of GABA for the wild type receptor {alpha}1{beta}2{gamma}2S was 45.8 µM (Fig. 2b), which is similar to the value previously obtained in our laboratory on oocyte expression (46). The GABA dose-response curve for the {alpha}1{beta}2(G219F){gamma}2 receptor, which has larger aromatic side chain volume, was shifted to the left, with a decreased EC50 value of 15.9 µM, significantly different (a 3-fold shift) from wild type (**, p < 0.001) (Fig. 2b). The Hill slopes of the GABA concentration-response relationships for the point-mutated {beta}2 containing receptors were assumed similarly and obtained the value of 0.72 under running the non-linear least square curve fitting by Excel software. The maximal GABA-induced currents for mutant receptor combinations were less. Theoretically, the fully open probability for maximal channel activation could not be elicited by GABA application alone, so we let the concentration-response curves float to a higher value of 100% to generate the best curve fitting. No spontaneous channel opening was observed in the mutant receptors when applying 50 µM picrotoxin to block the GABA channel (data not shown). Fig. 2b, inset, shows the EC50 values for the wild type and mutant GABAR.



View larger version (22K):
[in this window]
[in a new window]
  		FIG. 2.
a, the concentration-dependent responses of wild type, {alpha}1{beta}2{gamma}2S, and mutant receptors, {alpha}1{beta}2(G219I){gamma}2S, {alpha}1{beta}2(G219R){gamma}2S, {alpha}1{beta}2(G219F){gamma}2S, and {alpha}1{beta}2(G219W){gamma}2S. The wild type and point-mutated rat {beta}2 subunit cRNAs were mixed with rat {alpha}1 and {gamma}2S subunit cRNAs at 1:1:2 ratio and co-injected into Xenopus oocyte. The GABA-induced current could be measured from the GABAA receptor chloride channel on oocytes using two-electrode voltage clamp at –70 mV, after the 3–4-day incubation at 17–19 °C. b, GABA dose-response curves for wild type and point-mutated GABAA receptor subtypes expressed in oocytes injected with the cRNAs encoding for the wild type or mutated {beta}2 subunit in combination with the {alpha}1 and {gamma}2S subunits. The data points and error bars represent the means ± S.D. of 15 ± 9 experiments for the different GABA concentrations. The Hill slopes of the GABA concentration-response relationships for the point-mutated {beta}2 containing receptors were assumed similarly, obtaining the value of 0.72 under running the non-linear least square curve fitting by Excel software. {square}, WT(G), wild type {alpha}1{beta}2{gamma}2S, EC50, 45.8 µM; {circ}, GI, {alpha}1{beta}2(G219I){gamma}2S, EC50, 31.9 µM; {Delta}GR, {alpha}1{beta}2(G219R){gamma}2S, EC50, 24.7 µM; {blacksquare} GF, {alpha}1{beta}2(G219F){gamma}2S, EC50, 15.9 µM; • GW, {alpha}1{beta}2(G219W){gamma}2S, EC50, 21.8 µM.

 

The Correlation of EC50 GABA Relative to the Substituted Side Chain Residue Volume at Position 219 of {beta}2 Subunit— Mutation of Gly-219 to large amino acid residues produced a leftward shift in the curves, consistent with an increased apparent affinity for GABA in the mutant receptors (Fig. 2b). The mutant receptors {alpha}1{beta}2(Gly-219*){gamma}2S had EC50 GABA values lower than that for the wild type receptor. In receptors mutated at {beta}2 Gly-219, a strong negative correlation between the EC50 value and the molecular volume size was observed (Fig. 3) (48). The EC50 value for GABA varied over ~3-fold magnitude. For EC50 GABA, normalized maximal current amplitude was generated as mentioned above, and individually, the variations of EC50 for the wild type and mutant receptors were generated by Prism software (GraphPad). Data for all receptors tested are summarized in Table I. Interestingly, when {beta}2Gly-219 was replaced by negatively charged aspartate which showed a slower desensitization than wild type, conversely, the positively charged arginine showed a much faster desensitization and a slower deactivation than wild type (data not shown).



View larger version (11K):
[in this window]
[in a new window]
  		FIG. 3.
The correlation for the EC50 GABA within the mutation series of {alpha}1{beta}2 {gamma}2S receptors at position 219 and side chain volume change relative to glycine. The parameters of the amino acids side chain volume were from Harpaz et al. (58). The variations of EC50 were generated by Prism software (GraphPad).

 

View this table:
[in this window]
[in a new window]
  		TABLE I
The characteristics of wild type and mutant GABAA receptor from GABA-induced current data

From curve fitting of concentration response data, values indicated the EC50, calculated EC10 for the experiments of anesthetic positive modulation, and Emax. % of control wild type {alpha}1{beta}2{gamma}2S was at the same injection day (n = 20-30 experiments). In general, the expression of mutants were less than that of wild type except for G219S and G219I.

 

The Effects of Point Mutations on Potentiation of GABA Responses by Pentobarbital and Propofol—To determine whether the mutated amino acid residues influenced the ability of IV general anesthetics at a constant concentration, which would not elicit more than EC5 of maximal GABA current on direct activation by itself, to potentiate GABAR function, the effects of PB and propofol on potentiation of EC10 GABA responses (wild type {alpha}1{beta}2{gamma}2S, EC10, 5.03 µM; {alpha}1{beta}2(G219I) {gamma}2S, EC10, 4.52 µM; {alpha}1{beta}2(G219R){gamma}2S, EC10, 3.55 µM; {alpha}1{beta}2(G219F){gamma}2S, EC10, 0.98 µM; {alpha}1{beta}2(G219W){gamma}2S, EC10, 1.15 µM) were examined. Fig. 4a, a–e shows the potentiation by PB (10 µM) and propofol (5 µM) of EC10 GABA-induced currents in oocytes expressing the wild type {alpha}1{beta}2{gamma}2S receptor (calculated by the following equation: (I(Anes.+GABA) – I(GABA))/I(GABA) as the enhancement fold). GABA responses were greatly enhanced by co-application of either PB (1.9 ± 0.6-fold, n = 32 versus control) or propofol (2.3 ± 0.4-fold, n = 29). When {beta}2Gly-219 was substituted by a large amino acid such as Trp, the mutations at position Gly-219 of {beta}2 subunit nearly eliminated positive modulation (less than 0.5 enhancement fold) by PB and propofol (0.4 ± 0.03- and 0.28 ± 0.08-fold of control value, respectively, Fig. 4b), suggesting that {beta}2Gly-219 is one of the key residues for the modulation of GABAR by IV general anesthetics. On the other hand, greater potentiation by the anesthetics was observed in some {beta}2Gly-219 mutated receptors. For example, {alpha}1{beta}2(G219I){gamma}2S, in which glycine was substituted by the amino acid isoleucine that is smaller than an aromatic group Phe or Trp, both PB and propofol elicited a greater and significant potentiation of EC10 GABA currents (1.2 ± 0.25- and 1.1 ± 0.31-fold of control value, respectively, Fig. 4b).



View larger version (17K):
[in this window]
[in a new window]
  		FIG. 4.
a, a–e, enhancement of EC10 GABA responses by pentobarbital (PB, 10 µM) and propofol (Pro, 5 µM). EC10 GABA currents in wild type GABAA {alpha}1{beta}2{gamma}2S receptors are strongly enhanced by pre-perfusion for 15 s before co-application with EC10 GABA (wild type {alpha}1{beta}2{gamma}2S, EC10, 5.03 µM; {alpha}1{beta}2(G219I){gamma}2S, EC10, 4.52 µM; {alpha}1{beta}2(G219R){gamma}2S, EC10, 3.55 µM; {alpha}1{beta}2(G219F){gamma}2S, EC10, 0.98 µM; {alpha}1{beta}2(G219W){gamma}2S, EC10, 1.15 µM) at clinically relevant concentrations. Note that the constant concentrations of all anesthetics that were examined would not elicit more than EC5 of maximal GABA current in all receptor combinations of pentobarbital (PB, 10 µM) and propofol (Pro, 5 µM). In contrast, EC10 GABA-induced currents in {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant with substituted aromatic group receptors are less enhanced by co-application of PB (10 µM) or propofol (5 µM). The statistics panels of EC10 GABA modulation by {blacksquare} PB or {square} Pro are shown as the enhancement fold in b. Individual recordings are from oocytes injected with cRNAs encoding the indicated subunit combination. GW means, for example, {alpha}1{beta}2(G219W){gamma}2S receptor combination.

 

The Effects of Point Mutations on Potentiation of GABA Responses by Etomidate and Alphaxalone—To determine whether the mutated amino acid residues influenced the ability of other IV general anesthetics to potentiate GABAR function, the effects of etomidate and alphaxalone on potentiation of EC10 GABA responses were examined. Fig. 5a, a–d shows the potentiation by etomidate (5 µM) and alphaxalone (0.5 µM) of EC10 GABA-induced currents in oocytes expressing the wild type {alpha}1{beta}2{gamma}2S receptor. GABA responses are greatly enhanced by co-application of either etomidate (3.1 ± 0.35-fold, n = 21 versus control) or alphaxalone (3.3 ± 0.4-fold, n = 18). When {beta}2Gly-219 was substituted by a large amino acid such as Trp, the mutations at position Gly-219 of {beta}2 subunit decreased the positive modulation down to less than 1-fold enhancement by etomidate and alphaxalone (0.75 ± 0.09- and 0.7 ± 0.11-fold of control value, respectively, Fig. 5b). On the other hand, greater potentiation by etomidate and alphaxalone such as by PB and propofol was also observed in some {beta}2Gly-219 mutated receptors. Noteworthy, {alpha}1{beta}2(G219R){gamma}2S, in which glycine was substituted by a positively charged Arg, both etomidate and alphaxalone elicited a significantly lower potentiation of EC10 GABA currents (1.1 ± 0.13- and 1.3 ± 0.06-fold of control value, respectively, Fig. 5b), compared with wild type receptor and other non-aromatic group substituted residue combinations.



View larger version (17K):
[in this window]
[in a new window]
  		FIG. 5.
a, a–d, enhancement of EC10 GABA responses by etomidate (5 µM) and alphaxalone (0.5 µM). EC10 GABA currents in wild type GABAA {alpha}1{beta}2{gamma}2S receptors are strongly enhanced by pre-perfusion for 15 s before co-application with EC10 GABA (wild type {alpha}1{beta}2{gamma}2S, EC10, 5.03 µM; {alpha}1{beta}2(G219I){gamma}2S, EC10, 4.52 µM; {alpha}1{beta}2(G219R){gamma}2S, EC10, 3.55 µM; {alpha}1{beta}2(G219F){gamma}2S, EC10, 0.98 µM; {alpha}1{beta}2(G219W){gamma}2S, EC10, 1.15 µM.) at clinically relevant concentrations of etomidate (Eto, 5 µM) and alphaxalone (Alpha, 0.5 µM). In contrast, EC10 GABA-induced currents in {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant with substituted aromatic group receptors are less enhanced by co-application of etomidate (5 µM) and alphaxalone (0.5 µM). The statistics panels of EC10 GABA modulation by etomidate ({blacksquare} Eto) or alphaxalone ({square} Alpha) are shown as the enhancement fold in b. Individual recordings are from oocytes injected with cRNAs encoding the indicated subunit combination. GW means, for example, {alpha}1{beta}2(G219W){gamma}2S receptor combination.

 

The Correlation of Anesthetic Modulation Relative to the Substituted Side Chain Residue Volume at Position 219 of {beta}2 Subunit—When the {alpha}1{beta}2{gamma}2S GABAR {beta}2Gly-219 mutants were tested for modulation by the IV general anesthetics PB and propofol, there appeared to be a limiting volume for the side chains, such that amino acid residues with small side chains such as Gly and Ile demonstrated greater anesthetic potentiation, whereas very large residues such as Trp dramatically decreased the anesthetic potentiation by pentobarbital and propofol (10 and 5 µM), respectively, Fig. 4, a and b. Similar results were obtained in the modulation of etomidate and alphaxalone (5 and 0.5 µM) respectively, Fig. 5, a and b. The potentiation of EC10 GABA responses by anesthetics in {alpha}1{beta}2(Gly-219*){gamma}2S mutant GABAR declines with increasing side chain volume of the amino acid at position Gly-219 using the one-phase exponential equation (Equation 2) (Fig. 6), suggesting that Gly-219 of the {beta}2 subunit is a key residue for anesthetic potentiation of GABAR. There was a strong correlation between the substituted side chain volume and the degree of anesthetic potentiation.



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 6.
Potentiation of EC10 GABA responses by pentobarbital (PB, 10 µM) (a), propofol (5 µM) (b), etomidate (5 µM) (c), and alphaxalone (0.5 µM) (d) in {alpha}1{beta}2{gamma}2S mutant GABAA receptors relative to side chain volume of the amino acid at position {beta}2Gly-219. There was a strong positive correlation between the side chain volume of substituted residue and the degree of potentiation by IV anesthetics. Data are presented as mean ± S.E. from at least 15 experiments for each receptor combination by PB and propofol (Pro) and at least 8 experiments by etomidate (Eto) and alphaxalone (Alpha). The line represents a best fitting by one-phase exponential equation to the data.

 

Effects of Mutations in {beta}2 Subunit at Gly-219 on Direct Activation of GABAR by Intravenous Anesthetics—The direct activation of wild type receptors by PB (1 mM), propofol (50 µM), etomidate (100 µM), and alphaxalone (50 µM) was the same or less evident in mutant {alpha}1{beta}2(G219F){gamma}2S and {alpha}1{beta}2(G219W){gamma}2S receptors (Fig. 7 and Table II). Direct activation of wild type GABAA {alpha}1{beta}2{gamma}2S receptors by IV anesthetics was described previously (35). High concentrations of IV anesthetics can produce a profound inhibition on the effect of GABA responses (such as propofol >50 µM) (6), and there is an immediate "rebound" current upon washout, i.e. the current did not immediately return to the control level. In these experiments, inhibitory and rebound effects were most evident at concentrations more than PB (1 mM), propofol (50 µM), etomidate (100 µM), and alphaxalone (50 µM), respectively, although small rebound currents immediately after washout of applied anesthetics were occasionally observable at the concentrations listed above (Fig. 7a). An IV anesthetic concentration chosen in our experiment thus was suitable to compare near-maximal direct activation in different receptor combinations with minimal interference from inhibition. Etomidate (100 µM) produced significant direct effects in wild type {alpha}1{beta}2{gamma}2S receptors in the absence of GABA but less efficacy in {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant receptor (Fig. 7a). At high concentrations, alphaxalone produces substantial direct activation of GABAR. Alphaxalone (50 µM) was chosen for the direct activation experiments, because at this concentration, alphaxalone elicits a large direct effect with minimal blocking effects and minimal washout time (to remove neurosteroid from saturated GABAR). Thus, a very slow deactivation during washout was observed in Fig. 7a. Direct activation of all the mutant receptors by alphaxalone was similar to that for the wild type {alpha}1{beta}2{gamma}2S receptor (Fig. 7a).



View larger version (19K):
[in this window]
[in a new window]
  		FIG. 7.
a, direct activation by supra-anesthetic concentrations of pentobarbital (PB), propofol (Pro), etomidate (Eto), and alphaxalone (Alpha). Wild type and mutant GABAA {alpha}1{beta}2{gamma}2S receptors are activated directly by pentobarbital (1 mM), propofol (50 µM), etomidate (100 µM), and alphaxalone (50 µM). Horizontal scale bar, 15 s for direct activation by pentobarbital, propofol, etomidate, and alphaxalone and 15 s for the maximal GABA (1 mM) applications. Individual recordings are from oocytes injected with cRNAs encoding the indicated subunit combinations. b and c, concentration-response relationships for direct activation of GABAA {alpha}1{beta}2(Gly-219*){gamma}2S receptors by pentobarbital and etomidate. For wild type (WT) {alpha}1{beta}2{gamma}2S receptors, PB produces significant direct activation at a concentration of 1 mM (p < 0.05, 15–25 experiments). The direct activation of wild type {alpha}1{beta}2{gamma}2S ({blacksquare} WT(G)) receptor by PB has an Emax value of 34.1% of the maximal GABA-evoked current. In contrast, the {alpha}1{beta}2(G219F){gamma}2S ({diamondsuit}GF) or {alpha}1{beta}2(G219W){gamma}2S ({diamond}GW) receptor is more sensitive than the wild type {alpha}1{beta}2{gamma}2S receptor to the direct actions of PB with lower efficacies ({alpha}1{beta}2(G219F){gamma}2S, 25.8%, n = 18; {alpha}1{beta}2(G219W){gamma}2S, 22.2%, n = 12; of control, respectively). PB produced a moderate direct activation at a concentration of 1 mM in {alpha}1{beta}2(G219I){gamma}2S receptor (31.2% of control, n = 7). c, the direct activation of {alpha}1{beta}2{gamma}2S ({blacksquare} WT(G)) receptors produced by etomidate has a significant high Emax value of 64.1% at a concentration of 100 µM. Similar to PB, the {alpha}1{beta}2(G219F){gamma}2S ({diamondsuit}GF) or {alpha}1{beta}2(G219W){gamma}2S ({diamond}GW) receptor is more sensitive to the direct activation of etomidate at a concentration of 100 µM with lower efficacies (p < 0.001 for each concentration, Emax value of 35.2 and 29.4%, 8–15 experiments, respectively).

 

View this table:
[in this window]
[in a new window]
  		TABLE II
Direct activation data of GABAA receptor by anesthetics (% of maximal GABA response)

The best fitting concentrations of anesthetics (PB, 1 mM; propofol, 50 µM; etomidate, 100 µM; alphaxalone, 50 µM) were applied for maximal direct response of GABAA receptor (% of control maximal GABA responses in wild type {alpha}1{beta}2{gamma}2S and mutant receptors at 1 mM to 100 µM, respectively).

 

Direct Activation Curve of GABAR by PB or Etomidate— Further examination of the direct activation by PB revealed that the {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant receptor has a 3–5-fold higher apparent affinity (from 140 to 30 and 51 µM, for G219F and G219W, respectively) than the wild type {alpha}1{beta}2{gamma}2S receptor, as shown by the leftward shift in the concentration-response relationship for direct activation by PB (Fig. 7b). The GABA concentration-response curve for the {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant receptor is shifted to the left in a similar manner. PB (1 mM) activated the {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant receptor directly but elicited significantly smaller maximal currents compared with wild type {alpha}1{beta}2{gamma}2S receptors (p < 0.05 compared with wild type; Table II). As with PB, etomidate has an increased apparent affinity in the {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S mutant receptor, as indicated by a leftward shift in the etomidate direct activation concentration-response curve (Fig. 7c). Direct activation by 50 µM etomidate of the {alpha}1{beta}2(G219F){gamma}2S or {alpha}1{beta}2(G219W){gamma}2S also was reduced slightly compared with the wild type {alpha}1{beta}2{gamma}2S receptor (Table II). Although high concentrations of etomidate can produce channel inhibition and rebound currents at GABAR (49), these were not seen during our experiments at etomidate concentrations up to 100 µM. Similarly, the direct activation of GABAR by propofol and alphaxalone was more potent, however, with slightly reduced maximal response currents of both at 50 µM (Table II).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
REFERENCES
 
Little is known to date about the domains within GABAR that participate in the action of anesthetics. Mutation of certain amino acid residues within the M2 region reduced anesthetic sensitivity of GABAR (50), i.e. reduced enhancement of GABA but not direct agonism (51).

Channel residues can affect the apparent EC50 and binding of agonists and modulators by allosteric coupling (29, 40). Although the residue in M2 that forms a cavity with residues within M3 for the modulation of volatile anesthetics and n-alcohols (40) is the most implicated in anesthetic action to date, nevertheless, it may not be the anesthetic-binding site but merely conformationally coupled to anesthetic action. However, modeling of the {alpha}-helix of M2 suggests that the residue actually is not within the channel pore but may be on the opposite side of the helix, related to the rotation and movement in an activated state of channel opening (52) and thus in contact with lipids or other membrane-spanning domains of the same or other subunits that may represent a pocket, a site for attachment.

By using sequence scanning of {rho} versus non-{rho}, by the approach of mutating {beta}2Gly-219 to Phe found in {rho} subunit, we identified a point mutation in this region that is required for IV anesthetic modulation. Glycine and proline are conformationally important amino acids in that they appear to influence the conformation of the polypeptide (53). Glycine essentially lacks a side chain and therefore can adopt conformations that are sterically forbidden for other amino acids. This confers a high degree of local flexibility on the polypeptide. Accordingly, glycine residues are frequently found in turn regions of proteins where the backbone has to make a sharp turn. Glycine occurs abundantly in certain fibrous proteins due to its flexibility and because its small size allows adjacent polypeptide chains to pack together closely. These studies confirm our previous findings that mutation {beta}2G219F in {alpha}1{beta}2{gamma}2S can ablate potentiation by the IV anesthetics (43), suggesting that Gly-219 of the {beta} subunit is a key residue for anesthetic potentiation of GABAR.

We further characterized this site located at the entrance of M1 in the {beta}2 subunit, as to whether it works as a modulation site or a binding site, by replacing Gly with residues of various sizes and hydrophobicity. GABA-induced currents of receptors mutated at {beta}2Gly-219 were somewhat potentiated by IV general anesthetics at constant concentration, and the degree of potentiation was decreased in a manner correlated with the molecular volume of substituted amino acids (Fig. 6). By comparing the substitutions, the largest potentiation by anesthetics was observed with wild type {alpha}1{beta}2{gamma}2S, which showed the largest EC50 value for GABA in the absence of modulator, whereas the smallest potentiation by anesthetics was seen in aromatic residues, {beta}2 (G219F, G219Y, and G219W), which also showed the biggest left-shift for GABA EC50 values. There was a strong negative correlation between the left-shift for GABA EC50 value and the degree of anesthetic potentiation.

Similar results, i.e. a correlation between left-shift either for GABA or for anesthetics and for anesthetic modulation with volume of residue, were also found in mutation at the residues within M2 and M3, Ser-270 and Ala-291, of GABA {alpha} subunit and were interpreted as evidence for an actual site of contact with ligand (40, 54). The authors suggested (42, 54) that large amino acid residues already increased the probability of channel opening resulting in a correspondingly lower opportunity for additional enhancement by binding of an anesthetic ligand in the same space.

The correlation with the volume of the residue at position {beta}2Gly-219 and IV anesthetic enhancement of GABA observed here also is consistent with this residue being in the binding pocket for anesthetics. However, mutation at {beta}2Gly-219 failed to prevent direct gating by the same IV anesthetics, and thus this residue is not in the binding site involved in this action of the ligands. The two actions of the IV anesthetics could involve two separate sites of binding, although several models suggest a single binding site could mediate both (55). Whether this position is or is not a contact point for anesthetics needs further investigation, e.g. on ligand "cut-off" analysis (42) or using irreversible cysteine-modifying reagents to covalently label this site by selectively mutating to {beta}2G219C (56, 57).

The observation that the size of residue at {beta}2Gly-219 affects the apparent affinity (left-shift) for direct gating by both GABA agonists and IV anesthetics suggests increased probability of opening, a finding more consistent with involvement of this residue in an allosteric coupling domain rather than an anesthetic contact site. This possibility will require examination of channel kinetics using a system other than oocyte expression. Furthermore, a report (30) using cryo-electron microscopy to examine the structure of the nAChR, which like GABAR is a member of the LGIC family, showed features of possible relevance to these results on GABAR {beta}2Gly-219 position. Whereas M2 forms the channel pore wall, M1, M3, and M4 are bundled together near M2. Hydrophobic residues near Gly-219 at the extracellular end of M1 were found to come in contact with the back side of M2 and thus potentially interact. Thus the latest structural information supports the importance of the top of M1 including Gly-219 in possible interactions with the pore-lining M2 and an effect on channel gating plus modulation by allosteric ligands.


   FOOTNOTES
 
* This work was supported by National Institutes of Health Grant GM08652. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

To whom correspondence should be addressed: UCLA School of Medicine, Dept. of Molecular and Medical Pharmacology, 650 Young Dr. South, CHS 23-338, CA 90095-1735. Tel.: 310-454-5606; Fax: 310-267-2003; E-mail: Rolsen{at}mednet.ucla.edu.

1 The abbreviations used are: GABAR, {gamma}-aminobutyric acid type A receptor; GABA, {gamma}-aminobutyric acid; GABAA, {gamma}-aminobutyric acid type A; LGIC, ligand-gated ion channel; M1–4, transmembrane domain 1–4; PB, Pentobarbital; cRNA, capped mRNA; nAChR, nicotinic acetylcholine receptor; IV, intravenous. Back


   ACKNOWLEDGMENTS
 
We thank Tarek Abouelleil for assistance with the technical support of pClamp software and Digidata installation.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Olsen, R. W. (1998) Toxicol. Lett. 100, 193–201[CrossRef][Medline] [Order article via Infotrieve]
  2. Harris, B. D., Wong, G., Moody, E. J., and Skolnick, P. (1995) Mol. Pharmacol. 47, 363–367[Abstract]
  3. Longoni, B., and Olsen, R. W. (1992) Adv. Biochem. Psychopharmacol. 47, 365–378[Medline] [Order article via Infotrieve]
  4. Barker, J. L., and Ransom, B. R. (1978) J. Physiol. (Lond.) 280, 355–372[Abstract/Free Full Text]
  5. Steinbach, J. H., and Akk, G. (2001) J. Physiol. (Lond.) 537, 715–733[Abstract/Free Full Text]
  6. Hales, T. G., and Lambert, J. J. (1991) Br. J. Pharmacol. 104, 619–628[Medline] [Order article via Infotrieve]
  7. Uchida, I., Kamatchi, G., Burt, D., and Yang, J. (1995) Neurosci. Lett. 185, 203–206[CrossRef][Medline] [Order article via Infotrieve]
  8. Olsen, R. W., and Sapp, D. W. (1995) Adv. Biochem. Psychopharmacol. 48, 57–74[Medline] [Order article via Infotrieve]
  9. Harrison, N. L., and Simmonds, M. A. (1984) Brain Res. 323, 287–292[CrossRef][Medline] [Order article via Infotrieve]
  10. Olsen, R. W., Cagetti, E., and Waller, M. (2003) in Neurosteroid Effect in the Central Nervous System: The Role of GABAA Receptor (Smith, S. S., ed) pp. 47–61, CRC Press, Inc., Boca Raton, FL
  11. Yang, J., Isenberg, K. E., and Zorumski, C. F. (1992) FASEB J. 6, 914–918[Abstract]
  12. Whiting, P. J., McKernan, R. M., and Wafford, K. A. (1995) Int. Rev. Neurobiol. 38, 95–138[Medline] [Order article via Infotrieve]
  13. Macdonald, R. L., and Olsen, R. W. (1994) Annu. Rev. Neurosci. 17, 569–602[Medline] [Order article via Infotrieve]
  14. Olsen, R. W., and Sawyer, G. W. (2003) in Encyclopedia of Biological Chemistry (Lennarz, W., and Lane, M. D., eds) Elsevier, San Diego, in press
  15. Ortells, M. O., and Lunt, G. G. (1995) Trends Neurosci. 18, 121–127[CrossRef][Medline] [Order article via Infotrieve]
  16. Corringer, P. J., Le Novere, N., and Changeux, J. P. (2000) Annu. Rev. Pharmacol. Toxicol. 40, 431–458[CrossRef][Medline] [Order article via Infotrieve]
  17. Smith, G. B., and Olsen, R. W. (1994) J. Biol. Chem. 269, 20380–20387[Abstract/Free Full Text]
  18. Duncalfe, L. L., Carpenter, M. R., Smillie, L. B., Martin, I. L., and Dunn, S. M. (1996) J. Biol. Chem. 271, 9209–9214[Abstract/Free Full Text]
  19. Sigel, E., Baur, R., Kellenberger, S., and Malherbe, P. (1992) EMBO J. 11, 2017–2023[Medline] [Order article via Infotrieve]
  20. Sawyer, G. W., Chiara, D. C., Olsen, R. W., and Cohen, J. B. (2002) J. Biol. Chem. 277, 50036–50045[Abstract/Free Full Text]
  21. Chiara, D. C., Xie, Y., and Cohen, J. B. (1999) Biochemistry 38, 6689–6698[CrossRef][Medline] [Order article via Infotrieve]
  22. Nishikawa, K., Jenkins, A., Paraskevakis, I., and Harrison, N. L. (2002) Neuropharmacology 42, 337–345[CrossRef][Medline] [Order article via Infotrieve]
  23. Changeux, J. P., Galzi, J. L., Devillers-Thiery, A., and Bertrand, D. (1992) Q. Rev. Biophys. 25, 395–432[Medline] [Order article via Infotrieve]
  24. Chiara, D. C., and Cohen, J. B. (1997) J. Biol. Chem. 272, 32940–32950[Abstract/Free Full Text]
  25. Daubas, P., Devillers-Thiery, A., Geoffroy, B., Martinez, S., Bessis, A., and Changeux, J. P. (1990) Neuron 5, 49–60[CrossRef][Medline] [Order article via Infotrieve]
  26. Wagner, D. A., and Czajkowski, C. (2001) J. Neurosci. 21, 67–74[Abstract/Free Full Text]
  27. Amin, J., and Weiss, D. S. (1993) Nature 366, 565–569[CrossRef][Medline] [Order article via Infotrieve]
  28. Brejc, K., van Dijk, W. J., Klaassen, R. V., Schuurmans, M., van Der Oost, J., Smit, A. B., and Sixma, T. K. (2001) Nature 411, 269–276[CrossRef][Medline] [Order article via Infotrieve]
  29. Jenkins, A., Andreasen, A., Trudell, J. R., and Harrison, N. L. (2002) Neuropharmacology 43, 669–678[CrossRef][Medline] [Order article via Infotrieve]
  30. Miyazawa, A., Fujiyoshi, Y., and Unwin, N. (2003) Nature 423, 949–955[CrossRef][Medline] [Order article via Infotrieve]
  31. Wafford, K. A., Bain, C. J., Quirk, K., McKernan, R. M., Wingrove, P. B., Whiting, P. J., and Kemp, J. A. (1994) Neuron 12, 775–782[CrossRef][Medline] [Order article via Infotrieve]
  32. Sieghart, W., and Sperk, G. (2002) Curr. Top. Med. Chem. 2, 795–816[CrossRef][Medline] [Order article via Infotrieve]
  33. Carlson, B. X., Mans, A. M., Hawkins, R. A., and Baghdoyan, H. A. (1992) J. Pharmacol. Exp. Ther. 263, 1401–1414[Abstract/Free Full Text]
  34. Thompson, S. A., Whiting, P. J., and Wafford, K. A. (1996) Br. J. Pharmacol. 117, 521–527[Medline] [Order article via Infotrieve]
  35. Jones, M. V., Harrison, N. L., Pritchett, D. B., and Hales, T. G. (1995) J. Pharmacol. Exp. Ther. 274, 962–968[Abstract/Free Full Text]
  36. Belelli, D., Lambert, J. J., Peters, J. A., Wafford, K., and Whiting, P. J. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 11031–11036[Abstract/Free Full Text]
  37. Moody, E. J., Knauer, C., Granja, R., Strakhova, M., and Skolnick, P. (1997) J. Neurochem. 69, 1310–1313[Medline] [Order article via Infotrieve]
  38. Mihic, S. J., and Harris, R. A. (1996) J. Pharmacol. Exp. Ther. 277, 411–416[Abstract/Free Full Text]
  39. Harrison, N. L., Kugler, J. L., Jones, M. V., Greenblatt, E. P., and Pritchett, D. B. (1993) Mol. Pharmacol. 44, 628–632[Abstract]
  40. Mihic, S. J., Ye, Q., Wick, M. J., Koltchine, V. V., Krasowski, M. D., Finn, S. E., Mascia, M. P., Valenzuela, C. F., Hanson, K. K., Greenblatt, E. P., Harris, R. A., and Harrison, N. L. (1997) Nature 389, 385–389[CrossRef][Medline]</