Functional Modulation of Human Recombinant γ-Aminobutyric Acid Type A Receptor by Docosahexaenoic Acid*

Human γ-aminobutyric acid type A (GABAA) receptors were expressed in the baculovirus/Sf-9 insect cell expression system using recombinant cDNA of α1β2γ2s subunits. The effect of unsaturated fatty acids on GABAA receptor complexes was investigated electrophysiologically using conventional whole cell recording under voltage clamp. Three distinct effects of docosahexaenoic acid (DHA) on the GABA responses were observed. First, DHA, at a concentration of 10−7 m or greater, accelerated the desensitization after the peak of the GABA-induced current. Second, DHA (10−6 m) potentiated the peak amplitude of GABA response. This potentiation by DHA was inhibited in the presence of Zn2+ (10−5 m); Cu2+ and Ni2+ mimicked the action of Zn2+. Zn2+ (10−5 m) did not block the GABA response on α1β2γ2s receptor complexes. Third, DHA, at a concentration of 3 × 10−6 m or higher, gradually suppressed the peak amplitude of GABA response. A protein kinase A inhibitor, a protein kinase C inhibitor, and a Ca2+ chelator did not modify the effects of DHA on GABA-induced chloride ion current. Six unsaturated fatty acids other than DHA were examined. Arachidonic acid mimicked the effect of DHA while e.g. oleic acid had no effect. The inhibition of the GABA response in the presence of DHA was also observed in cells expressing GABAA receptors of α1 and β2 subunit combinations. The data show that the γ subunit is essential for DHA and arachidonic acid to potentiate the GABA-induced Cl− channel activity and to affect the desensitization kinetics of the GABAAreceptor.

Polyunsaturated fatty acids are essential fatty acids that are abundantly bound in the phospholipids of brain membranes. The fatty acids influence the membrane fluidity of the neuronal plasma membrane as well as the functional properties of integral membrane protein (1). Arachidonic acid (AA) 1 and docosahexaenoic acid (DHA) are polyunsaturated fatty acids localized in the 2-position of membrane phospholipids and are liberated by various intracellular enzymes such as phospholipase A 2 and diacylglycerol lipase (2).
Recently, various actions of these free fatty acids on the functional properties of neuronal plasma membrane neurotransmitter receptors have been reported, such as N-methyl-Daspartic acid receptor (3,4), kainate receptor (4), and potassium channel receptor (5,6). As for the effect of unsaturated fatty acids on ␥-aminobutyric acid type A (GABA A ) receptors, an electrophysiological study demonstrated that AA and DHA potently suppressed the GABA A receptor-mediated current response in the rat substantia nigra neurons (7). Receptor binding studies have shown marked effects of unsaturated fatty acids on brain GABA A receptor complexes in vitro (8 -11), as well as on human recombinant GABA A receptors expressed in Sf-9 insect cells (12). These findings suggest a variety of actions of fatty acids on the GABA A receptor-channel complex. The diversity of effects induced by free fatty acids might be due to ontogenetic, phylogenetic, and regional differences of the GABA A receptor subunit composition in the brain regions studied (13). Furthermore, a recent study on recombinant human GABA A receptors (12) reported that the effect of unsaturated fatty acids was dependent on the subunit composition of the GABA A receptor complexes. In addition, various combinations of GABA A receptor subunits might be expressed in the same neuron, complicating the interpretation of data on the effect of free fatty acids on the GABA A receptor complex obtained at the level of single cells, as, for example, by whole cell patch clamp techniques.
The insect cell Sf-9/baculovirus expression system in combination with electrophysiological techniques provide the means for the investigation of functional difference between GABA A receptors consisting of various subunits. For this purpose, we employed recombinant human GABA receptor-channel complexes composed of ␣ 1 ␤ 2 ␥ 2s subunits expressed in the Sf-9 insect cell system. GABA A receptor complexes composed of ␣ 1 , ␤ 2 , and ␥ 2s subunits have been suggested to be highly abundant in the vertebrate brain (13,14) and in most brain areas, e.g. hippocampus (15), cortex (16), and cerebellum (17).
In this study, the effects of DHA and other fatty acids on GABA-induced chloride currents were investigated in recombinant human GABA A receptor-channel complexes composed of ␣ 1 ␤ 2 ␥ 2s subunits expressed in the Sf-9 insect cell system.

EXPERIMENTAL PROCEDURES
The construction of expression vectors for the human GABA A receptors subunits ␣ 1 , ␤ 2 , and ␥ 2s has been described previously (18). Solidphase DNA sequencing (Dynal®) combined with sequenase version 2.0 DNA sequencing kit (U. S. Biochemical Corp.) of each insert DNA verified that the amino acid sequences of the ␣ 1 and ␤ 2 subunits were identical to those previously reported (19,20) and that the ␥ 2s subunit varied at amino acid residue 81 (threonine instead of serine) and amino acid residue 142 (threonine instead of serine) compared with the reported amino acid sequences (21).
Expression of GABA A Receptor in Sf-9 Cells-Sf-9 insect cells were grown in spinner flask cultures at 27°C in serum-free medium (Sf900-II-SFM, Life Technologies, Inc.). Insect cells at a density of 7.5 ϫ 10 5 cells/35-mm Petri dish (Falcon) were infected with baculovirus contain-ing human cDNA of ␣ 1 , ␤ 2 , and ␥ 2s subunits at a multiplicity of infection of 3:1:2 or ␣ 1 and ␤ 2 subunits at a multiplicitity of infection of 3:1. The infected cells were incubated at 27°C for 48 h.
Electrophysiological Experiments-At approximately 48 h postinfection, Sf-9 cells were constantly perfused with extracellular solution at a flow rate of 3-4 ml/min. The composition of the standard external solution was (in mM): NaCl 150, KCl 5, MgCl 2 1, CaCl 2 2, glucose 10, HEPES 10. The pH was adjusted to 7.4 by Tris base. Patch pipettes contained a solution of (in mM): Cs 2 SO 4 50, CsCl 78, MgCl 2 6, EGTA 5, ATP 5, HEPES 10 (pH 7.2 by Tris base). Series resistances in whole cell voltage clamp experiments were calculated from the capacitative current peak in a 10 mV voltage step and were in range of 10 -25 megohms.
Electrical recordings were carried out using conventional whole cell patch recording. Patch pipettes were made of glass capillaries with an outer diameter of 1.5 mm using a vertical puller (Narishige, PB-7, Japan). The cells were voltage-clamped using a voltage-clamp amplifier (Nihon Koden, CEZ-2300, Japan). All signals were filtered with a low pass filter with a cut-off frequency of 1 KHz, monitored on a syncroscope (Iwatsu, MS-SlOOA, Japan) and a pen recorder (Sanei, Recti Horiz 8K, Japan), then digitized at a rate of 44 KHz (Sony, PCM 501 ESN, Japan). The data were stored on video tape (Mitsubishi, HV-F73, Japan). All experiments were carried out at room temperature (23-25°C).
Drug Application-Rapid application of drugs was achieved by the "Y tube" method as described previously (22). In the present study, each drug was applied at an interval of more than 2 min unless otherwise stated. The drugs used for current recording were GABA, H-89, chelerythrine, and BAPTA (Sigma). DHA, AA, docosahexapentaenoic acid, docosatetraenoic acid, docosatrienoic acid, docosadienoic acid, and oleic acid were from Funakoshi, Tokyo, Japan. Free fatty acids were dissolved in dimethyl sulfoxide and diluted into the extracellular solution just before use. The final concentration of dimethyl sulfoxide was less than 0.1% (v/v).
Statistical Analysis-The experimental values are presented as mean Ϯ S.D. Student's t test was used when two groups were compared. When relationships between the peak current amplitude and the GABA concentration were examined, continuous lines were fitted according to the following equation where I is the normalized value of the current, I max the maximal response, C the GABA concentration, EC 50 the concentration corresponding to the half-maximal response, and n the apparent Hill coefficient. The concentration-inhibition curves were drawn using a mirror image of the modified Michael-Menten equation in combination with a least-square fitting routine after normalizing the amplitudes of the responses unless otherwise stated.
I indicates the normalized value of the current obtained with the concentration (C) of the drug, where the control response is defined as 1. Then, IC 50 and n denote the concentration giving half-maximal response and the Hill coefficient, respectively.

RESULTS
In Sf-9 cells transfected with homomeric ␣ 1 , ␤ 2 , or ␥ 2s or dimeric ␣ 1 ␥ 2s or ␤ 2 ␥ 2s receptor subunit combinations, GABA (up to 10 Ϫ3 M) did not induce any detectable current (less than 5 pA, n ϭ 10 -12 in each group) at a holding potential (V H ) of Ϫ40 mV (data not shown). On the other hand, in cells transfected either with dimeric ␣ 1 and ␤ 2 , or with trimeric ␣ 1 , ␤ 2 , and ␥ 2s receptor subunit combinations, 3 ϫ 10 Ϫ5 M GABA produced an initial peak inward current followed by a gradual desensitization at a V H of Ϫ40 mV ( Figs. 1 and 6). Immediately after the removal of GABA from the perfusate, the inward current returned to the current level prior to the application of GABA. The respective concentrations of GABA for threshold and EC 50 were 10 Ϫ6 M and 3.1 ϫ 10 Ϫ5 M in ␣ 1 ␤ 2 , and 10 Ϫ6 M and 4.7 ϫ 10 Ϫ5 M in ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations.
Effects of DHA on GABA Responses of Sf-9 Cells Expressing ␣ 1 ␤ 2 ␥ 2s Receptor Subunit Combinations-In ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations, three distinct time and concentrationdependent effects of DHA were observed. First, DHA (10 Ϫ6 M) gradually accelerated the decay of I GABA following the peak amplitude as shown in the Fig. 1A, a. In the absence of DHA, the decay time to 80% of the initial peak amplitude (T 0.8 ) was 2.17 Ϯ 0.09 s (mean Ϯ S.D., n ϭ 7). After continuous application of DHA (10 Ϫ6 M) for 3 and 10 min, the T 0.8 values were reduced to 1.26 Ϯ 0.08 s (n ϭ 7) and 0.34 Ϯ 0.06 s (n ϭ 6), respectively. Both T 0.8 values obtained with DHA are significantly different from the control T 0.8 value (p Ͻ 0.05, t test). After the removal of DHA from the perfusate, the T 0.8 values returned gradually to the control value. This process usually required more than 20 min.
The effect of DHA on the peak amplitude induced by GABA in ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations was concentration dependent. The peak I GABA was potentiated by 10 Ϫ6 M DHA (n ϭ 8) (Fig. 1A, a). This potentiation was apparent even if 10 Ϫ6 M DHA was applied simultaneously with GABA and faded rapidly upon removal of DHA from the perfusate. The effect of 10 Ϫ6 M DHA on the GABA response was also dependent on the concentration of GABA used, since 10 Ϫ6 M DHA significantly potentiated the current responses to 3 ϫ 10 Ϫ5 M GABA or lower. However, at concentrations of 10 Ϫ4 M or higher, the potentiation was not evident (Fig. 2B, bar graph inset).
DHA, at concentrations higher than 3 ϫ 10 Ϫ6 M, invariably suppressed the peak I GABA in a concentration and time-dependent manner (Fig. 1A, b, and B). The time interval to the maximal suppression of I GABA was dependent on the DHA concentration applied (Fig. 1B). The peak I GABA suppressed by DHA gradually returned to the original value after the washing out of DHA (Fig. 1A, b, and B). The concentration-response relationships for GABA in the absence or presence of DHA demonstrate that DHA affects the peak I GABA without affecting the  (Fig. 2). These results indicate that DHA at a concentration higher than 10 Ϫ6 M inhibits the GABA response in ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations in a noncompetitive manner.
DHA and other fatty acids affect the mobilization of intracellular effector molecules such as protein kinase C (23) and intracellular Ca 2ϩ (24,25). Since several different putative phosphorylation sites have been suggested in GABA A receptor complexes consisting of ␣ 1 ␤ 2 ␥ 2s subunit combinations (26), we investigated the possible involvement of several intracellular modulators in the effects of DHA on the GABA response of ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations. Addition of H-89, a protein kinase A inhibitor, chelerythrine, a protein kinase C inhibitor and intracellular application of BAPTA, a Ca 2ϩ chelator, did neither affect the potentiation of the 3 ϫ 10 Ϫ5 M GABA response by 10 Ϫ6 M DHA nor the inhibition of the GABA responses induced by 3 ϫ 10 Ϫ6 M DHA (Fig. 3). Furthermore, the effects of DHA on the desensitization of I GABA persisted in the presence of these inhibitors. The  4). These results suggest that neither protein kinase A or C or intracellular Ca 2ϩ are involved in the modulatory actions of DHA on the GABA response of ␣ 1 ␤ 2 ␥ 2 s receptor subunit combinations.
Surprisingly, the facilitatory effect of DHA (10 Ϫ6 M) on the peak I GABA was blocked by adding Zn 2ϩ (10 Ϫ5 M) (Fig. 4A), a concentration at which Zn 2ϩ only minimally suppressed the peak of GABA response in the ␣ 1 ␤ 2 ␥ 2s GABA A receptor subunit combinations (Fig. 4A). The DHA concentrations for threshold and IC 50 values for the response induced by GABA (3 ϫ 10 Ϫ5 M) were 3 ϫ 10 Ϫ7 M and 3 ϫ 10 Ϫ6 M with 10 Ϫ5 M Zn 2ϩ , respectively. Other divalent cations such as Ni 2ϩ (Fig. 4C, n ϭ 4) and Cu 2ϩ (n ϭ 4, data not shown) mimicked the antagonistic effect of Zn 2ϩ on the potentiation of the GABA response by DHA. Removal of both Ca 2ϩ and Mg 2ϩ from the perfusate did not affect the potentiating ratio of the response induced by 10 Ϫ6 M DHA (19.2 Ϯ 4.8%, n ϭ 4). However, the changes in the desensitization kinetics of I GABA induced by DHA were unaffected by these divalent cations (Fig. 4, A and C).
Unsaturated fatty acids other than DHA, e.g. AA have been reported to inhibit the GABA A response in rat substantia nigra neurons (7). Therefore, we examined the effects of various unsaturated fatty acids on GABA A ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations with the surprising result that only AA was able to mimic the effects of DHA while compounds structurally closely related to DHA, e.g. docosapentaenoic acid or an endogenous, relatively common unsaturated fatty acid, oleic acid, were inactive (Fig. 5).
Effect of DHA on GABA A Receptors Composed of ␣ 1 ␤ 2 Subunit Combinations-In Sf-9 cells expressing GABA A receptors composed of ␣ 1 ␤ 2 subunit combinations, DHA, in a concentration-dependent manner, suppressed the response to GABA (3 ϫ 10 Ϫ5 M) (Fig. 6A). The effect of DHA was restricted to the inhibitory effect, since neither a potentiation of the GABA response (Fig. 6A) or an alteration of the desensitization of I GABA (Fig. 6C) as in the case of ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations could be observed in ␣1␤2 subunit combinations.
The concentration-response relationship for GABA indicates that GABA concentrations for threshold and EC 50 values were 3 ϫ 10 Ϫ6 M and 4.35 ϫ 10 Ϫ5 M in the absence, and 3 ϫ 10 Ϫ6 M and 1.15 ϫ 10 Ϫ5 M in the presence of DHA (10 Ϫ6 M) (Fig. 6B). These data suggest that DHA is a noncompetitive inhibitor of GABA in ␣ 1 ␤ 2 subunit combinations. DISCUSSION In the present study it was shown that 1) in GABA A receptor complexes consisting of ␣ 1 ␤ 2 ␥ 2s and ␣ 1 ␤ 2 receptor subunit combinations, Zn 2ϩ blocks the GABA-induced Cl Ϫ channel activity. The presence of a ␥ 2s subunit in the GABA A receptor complexes markedly diminished the inhibitory effect of Zn 2ϩ . 2) The presence of the ␥ 2s subunit in the receptor complex is essential for DHA to potentiate the GABA response of ␣ 1 ␤ 2 ␥ 2s receptor subunit combinations, this response is blocked by Zn 2ϩ , Cu 2ϩ and Ni 2ϩ . Furthermore, DHA accelerates the desensitization of I GABA in ␣ 1 ␤ 2 ␥ 2s combinations, this effect was not blocked by the divalent cations. 3) DHA and AA, at concentrations higher than 10 Ϫ6 M, block the GABA-induced Cl Ϫ channel activity in ␣ 1 ␤ 2 ␥ 2s and ␣ 1 ␤ 2 receptor subunit combinations, while other unsaturated fatty acids, e.g. oleic acid, had no effect. These result suggest that DHA has a number of molecular target sites within the GABA A receptor channel complex. One site might be related to the Cl Ϫ channel domain. Two additional sites relate functionally to the presence of the ␥ 2 subunit in the receptor complex. One site is involved in the potentiation of the GABA response and is sensitive to antagonism by Zn 2ϩ . The other site may participate in the acceleration of the desensitization of the GABA-gated current, an effect insensitive to Zn 2ϩ .
Inhibitory Effect of DHA on the GABA-induced Chloride Ion Current of GABA A Receptor Complexes-At a concentration of 3 ϫ 10 Ϫ6 M or higher, DHA is a noncompetitive antagonist of GABA A receptor responses in ␣ 1 ␤ 2 ␥ 2s (Fig. 1) and ␣ 1 ␤ 2 (Fig. 6) receptor subunit combinations. A similar inhibitory effect of DHA has been reported on native GABA A receptor complexes of acutely dissociated neurons from rat substantia nigra (7). On the other hand, DHA has been shown to potentiate the Nmethyl-D-aspartic acid responses of dissociated rat nigra neurons (7) as well as cortical neurons (4). The present results suggest that DHA selectively inhibits the function of the GABA A receptor-chloride ion channel complexes.
The inhibitory action of DHA is time-and concentration-dependent (Figs. 1 and 6). In addition, the DHA-induced inhibition of I GABA ceases only slowly after the removal of DHA from the extracellular solution. There are several possible explanations for the gradual action of DHA on the GABA A receptorchannel complex. First, the suppression of the GABA response by fatty acids might be mediated through an intracellular second messenger system, since fatty acids have been shown to enhance the diacylglycerol-dependent activation of protein kinase C (23) and to increase the level of free intracellular Ca 2ϩ (24). Furthermore, the GABA A receptor-Cl Ϫ channel complex has been reported to be modulated by intracellular Ca 2ϩ (27), by protein kinase A (28) or by protein kinase C (13,26). However, the present results show that neither chelerythrine, a protein kinase C inhibitor, H-89, a protein kinase A inhibitor, nor intracellular BAPTA interfered with the effects of DHA on the GABA response (Fig. 3), suggesting that these intracellular signaling pathways are not involved in the effects of DHA on the GABA A receptor complex.
Second, since fatty acids have been shown to be able to modify the membrane fluidity and function of integral membrane proteins (1), the slow onset of the action of DHA might be explained by changes induced in the lipid microenvironment of the GABA A receptor-chloride ion channel complex. However, this explanations seem improbable since among various structurally closely related fatty acids examined, only DHA and AA affect the GABA responses (Fig. 5). AA and DHA have been shown to inhibit muscimol-induced Cl Ϫ uptake in rat cerebral cortical synaptosomes (29). Likewise, free fatty acids have been shown to decrease the binding of tert-butylbicyclophosphorothionate, which has been proposed to bind at the Cl Ϫ channel domain (10,30). Therefore, it might be concluded that DHA and AA bind to a site close to the Cl Ϫ channel, acting as a Cl Ϫ channel blocker.
Potentiation of the Peak Amplitude and Acceleration of the Desensitization of the GABA Response in the ␣ 1 ␤ 2 ␥ 2s GABA A Receptor Subunit Combinations by DHA-A distinct action of DHA on the GABA response is to accelerate the desensitization of I GABA after the peak response in ␣ 1 ␤ 2 ␥ 2s GABA A receptor subunit combinations. This change in the desensitization kinetics was observed at DHA concentration as low as 10 Ϫ7 M, and showed both slow development and recovery (Fig. 1). In contrast, DHA did not change the desensitization of ␣ 1 ␤ 2 GABA A receptor subunit combinations (Fig. 6).
A striking action of DHA on the GABA A response is that DHA (10 Ϫ6 M) potentiated the peak I GABA in ␣ 1 ␤ 2 ␥ 2s GABA A receptor combinations, this effect was absent in ␣ 1 ␤ 2 GABA A receptor combinations. The potentiating effect of DHA is obvious when GABA concentrations lower than 10 Ϫ4 M was applied, no potentiation was observed at GABA concentrations of 10 Ϫ4 M or higher (Fig. 2B). Although this might be due to the limitation of the Y-tube application system (22), another plausible explanation might be a "ceiling effect," since the GABA concentration of 10 Ϫ4 M to 10 Ϫ3 M induces the maximal current response achievable in a given cell and therefore masks potentiating effects of DHA. The very rapid desensitization of I GABA induced by higher concentrations of GABA itself in addition to the accelerated desensitization induced by DHA might reduce Current traces were obtained prior to (left trace) and 4 min after the addition of DHA (right trace) in the same cell. The peak amplitudes were arbitrary manipulated for a convenience to compare the decays of GABA responses in the presence and absence of DHA. The holding potential was Ϫ40 mV. the true peak current. Unlike the slow inhibition of GABA response by DHA (Fig. 1), the potentiation of I GABA was apparent even if DHA (10 Ϫ6 M) was applied simultaneously with GABA.
The high affinity benzodiazepine binding within the GABA A receptor complex is dependent on the presence of the ␥ 2 (or ␥ 3 ) subunit (31). Unsaturated free fatty acids increase the benzodiazepine binding on the GABA A receptors in native (11) and recominant (12) GABA A receptor chloride channel complexes. Free fatty acids interact with a GABA A receptor subunit configuration that requires the presence of a ␥ 2 subunit. The potentiation of the GABA response on one hand and the increased desensitization of GABA responses on the other seem to be independent of each other since: 1) the development of the acceleration of desensitization is a slow, while the potentiation of the GABA response by DHA is a rapid phenomenon (Fig. 1A,  a); and 2) divalent cations blocked the potentiation of GABA response by DHA, but the desensitization was not affected (Fig. 4).
The existence of a fatty acid binding domain has been reported for the N-methyl-D-aspartic acid receptor channel complex (32) and the binding of fatty acids to the specific binding domain induces a conformational change of the protein (33). The present data suggest that the binding of unsaturated fatty acids to a specific binding site near or at the ␥ 2 subunit results in a conformational change of the GABA A receptor complex which leads to an alteration of the response to GABA, e.g. the potentiation of the I GABA and the acceleration of the desensitization of the GABA A receptor composed of ␣ 1 ␤ 2 ␥ 2s subunit combinations.
The Effect of Zn 2ϩ on the GABA Response of ␣ 1 ␤ 2 ␥ 2s GABA A Receptor Subunit Combinations and Their Modulation by DHA-In accordance with previous findings (34), Zn 2ϩ blocked the GABA A receptor response. Our experiment demonstrated that ␣ 1 ␤ 2 GABA A receptor complexes were 3 orders of magnitude more sensitive to the inhibitory effect of Zn 2ϩ than at ␣ 1 ␤ 2 ␥ 2 GABA A receptor subunit combinations. This result indicates that the presence of ␥ 2s subunits render the GABA A receptor-channel complex less sensitive to Zn 2ϩ . The binding site for Zn 2ϩ on the GABA A receptor-channel complex has been reported to be located at the extracellular part of the Cl Ϫ channel complex, resulting in a reduction of the opening frequency of the Cl Ϫ channel (35). Interestingly, Zn 2ϩ (10 Ϫ5 M) inhibited the potentiation of the GABA A receptor response induced by 10 Ϫ6 M DHA in ␣ 1 ␤ 2 ␥ 2s GABA A receptor subunit combinations (Fig. 4). At least four possible mechanisms of action for this effect can be suggested. 1) Zn 2ϩ could inhibit the binding of GABA to the receptor in ␣ 1 ␤ 2 ␥ 2s receptor combinations. However, since a similar concentration-response relationship for GABA in the absence or presence of 10 Ϫ5 M Zn 2ϩ is observed 2 this explanation is unlikely. 2) Zn 2ϩ could reduce the bioactivity of DHA. However, since the desensitization of I GABA is equally affected by DHA (10 Ϫ6 M) in the absence or presence of 10 Ϫ5 M Zn 2ϩ (Fig. 4A), this explanation is improbable. 3) Zn 2ϩ could influence the binding of DHA to a site at the ␥ 2s subunit. 4) Zn 2ϩ could interfere with the interaction between ␥ subunits and other subunits of the GABA A receptor.
Possible Involvement of DHA in Neuronal Function-DHA and AA are abundant fatty acids in the brain tissue. The respective concentration of DHA and AA are 17 and 12% (by weight) of total fatty acids in the adult rat brain (36). These fatty acids can be released from membrane phospholipids by the action of phospholipase A 2 (2). The physiological relevant concentration of these fatty acids in the neuron has been estimated to be 1-10 M (37). An increase in phospholipase activity occurring in, e.g. ischemia and epilepsy, will cause a transient rise of the concentration of free fatty acids by more than several times (37). Zn 2ϩ is found throughout the brain, being concentrated in the particular areas such as cortex and hippocampus (38). It is of particular interest that Zn 2ϩ is concentrated in the synaptosomes (39) and Zn 2ϩ has been shown to be released from the nerve terminal (40).
The present result suggest a variety of actions of DHA and AA on the activity of GABA A receptor channel complex and emphasize the importance of the lipid microenvironment for the activity of ligand-gated ionic channels. The combination of three factors, namely, the concentration of DHA and AA, the subunit composition of the GABA A receptor complexes and the presence of Zn 2ϩ induce a whole spectrum of potential modulatory mechanisms affecting GABA A response complexes in the central nervous system.