GABAa receptor gating and the IPSC

GABAa receptors undergo considerable spontaneous motion while in the closed state (Bera and Akabas, 2005), suggesting a level of flexibility appropriate for rapid gating (Maconochie et al., 1994; Jones and Westbrook, 1995; McClellan and Twyman, 1999; Burkat et al., 2001; Chakrapani and Auerbach, 2005). In theory, the channel can open in the absence of agonist, albeit with an extremely low probability (Chang and Weiss, 1999a; Campo-Soria et al., 2006). Of note, some recombinant (Sigel et al., 1989; Maksay et al., 2003; Lindquist et al., 2004; Wagner et al., 2005; Ranna et al., 2006) and native (Birnir et al., 2000;

Jones et al., 2006; McCartney et al., 2007) GABAa receptors do exhibit measurable spontaneous activity, but most allow ion flux only after they have been "occupied" by agonist, two molecules of GABA being needed for efficient gating (Baumann et al., 2003). In such a scheme — shared with other LGICs, and based on an extension of the del Castillo-Katz (del Castillo and Katz, 1957) mechanism (Fig. 1) — the agonist can be seen as activating the receptor by producing a massive (~100K-10M fold) increase in the basal probability of opening (Campo-Soria et al., 2006; Sine and Engel, 2006). Thus, exposure to GABA drives receptors through mono- and di-liganded closed states to the open state.

In order to describe observed channel behaviour (and hopefully to reflect physical events (Colquhoun, 2006)), extensions of such reaction schemes have been developed by many groups (Macdonald et al., 1989; Weiss and Magleby, 1989; Twyman et al., 1990; Jones and Westbrook, 1995; Haas and Macdonald, 1999; Jayaraman et al., 1999; Burkat et al., 2001; Akk et al., 2004; Celentano and Hawkes, 2004; Lema and Auerbach, 2006). All these models contain multiple open and closed states (Figs. 1d, e). A key difference among them is the presence of additional agonist-bound closed states, which can be entered instead of the open states (desensitized states; (Jones and Westbrook, 1995)), or prior to the open states, so-called "pre-gateway non-conducting" (Lema and Auerbach, 2006) or "flip" states (Colquhoun, 2006).

At the synapse, following the release of a single vesicle, GABA reaches a peak concentration that would, at equilibrium, produce maximal receptor activation. However, because the on-rate of binding of GABA is slower than the rate of diffusion (Jones et al., 1998), the short exposure means that only a fraction of the ten to a few hundred receptors clustered opposite the single release site (Edwards et al., 1990; Mody et al., 1994; Nusser et al., 1997; Brickley et al., 1999) will be fully occupied. The degree of receptor occupancy varies between synapses on different neurons and even between those on a single cell (Nusser et al., 1997; Perrais and Ropert, 1999; Hajos et al., 2000; Overstreet et al., 2002; Mozrzymas et al., 2003), and will, of a del Castillo-Katz occupied ka e a+r , - ar , " ar*

c Simplified two-site k, ar a2r a,r*

b Two-site allosteric k\ k2

e2 a2r a2r'

deactivation

- channel 1 channel 2 channel 3

Fig. 1. Reaction mechanisms for cl-LGICs and the generation of a postsynaptic current. Panel (a) The del Castillo-Katz scheme (del Castillo and Katz, 1957), in which the agonist (A) binds to the receptor (R) forming a complex (AR) that changes conformation to yield a receptor with an open channel (AR*). KA is the equilibrium dissociation constant for binding and E is the equilibrium constant for the gating step (opening rate constant/shutting rate constant; b/a, panel c). Panel (b), Extension of this reaction mechanism, in line with the Monod-Wyman-Changeux scheme (see Colquhoun, 1998 for discussion), to include two binding sites and the potential for constitutive (un-liganded) channel opening. E0 is the equilibrium constant for gating in the absence of agonist. The progressively darker shading moving from E0 to E2 is designed to indicate the massively increased likelihood of conformational change in the presence of two agonist molecules (see also Downing et al., 2005; Campo-Soria et al., 2006; Sine and Engel, 2006). Panel (c) A simplified two-site model, in which un-liganded and mono-liganded openings are ignored. The microscopic rate constants are shown for binding (k1 and k2), unbinding (k_1 and k_2), opening (b) and closing (a). Shown beneath the reaction scheme are idealized state transitions (after Sine and Engel, 2006) for three channels activated by a brief pulse of agonist. Rate constants k1, k2, k_1, k_2 and b determine the rising phase of the response (activation), while k_2, b and a govern the decay (deactivation). Open states of the same three events (1-3) are shown below, together with 17 other simulated openings. When summed, these produce the exponential decay characteristic of the synaptic current. The continuous orange line is the numerical integration of the same simulation. Panels (d) and (e) Examples of reaction schemes that have been used to describe the behaviour of GABAa receptors. The widely used model of Jones and Westbrook (Jones and Westbrook, 1995) (d) incorporates two liganded closed (desensitized) states (Dslow and Dfast), while the model of Lema and Auerbach (Lema and Auerbach, 2006) (e) contains an additional "pre-open closed'' or gateway state. In all schemes, the open (ion conducting) states of the receptors are shown in red.

course, be sensitive to factors that modify the time course of the GABA concentration transient. As outlined above, for those receptors that are occupied by GABA, a conformational change is elicited that may ultimately lead to channel opening. The behaviour of the receptors can be envisaged as a series of transitions through various closed, open and desensitized states. The time spent in each of the various states, and thus the time course of the postsynaptic current, is determined by the properties of the receptors and the profile of GABA exposure (Mozrzymas, 2004).

GABAergic miniature inhibitory postsynaptic currents (mlPSCs), resulting from the release of a single vesicle, have a rapid onset with rise times of only a few hundred microseconds (Bier et al., 1996; Nusser et al., 1997; Brickley et al., 1999). This reflects the proximity of the receptors to the site of GABA release, the rapid binding of GABA and the speed of the transition between closed and open states (i.e., a high rate constant for opening). As the GABA concentration transient is brief, the decay of the IPSC reflects the closure of channels following removal of ligand (deactivation). The duration of this process depends on the rate constant for channel closing, and also on various transitions of the receptor, notably entry into and exit from desensitized states that have been viewed as trapping GABA on the receptor prior to the final unbinding (Jones and Westbrook, 1995; Chang and Weiss, 1999b; Haas and Macdonald, 1999; Bianchi and Macdonald, 2001). Because the rates of these transitions differ for GABAA receptors of different subunit composition, the expression of different receptor subtypes contributes to differences in IPSC decay observed at different stages of development (Okada et al., 2000; Vicini et al., 2001) and in different cell types (Nusser et al., 1999; Bacci et al., 2003; Ramadan et al., 2003). Typically, IPSCs have a much longer duration in immature than in mature neurons.

As intimated above, discrete point-to-point communication, mediated by GABA release from an axon terminal onto closely apposed receptors, is not the only form of GABAergic synaptic signalling. The factors that determine the transmitter exposure of any particular receptor include its location relative to the GABA release site(s), the geometry and arrangement of the neighbouring cellular elements, the presence of diffusional barriers and the proximity of GABA transporters (see above; and Barbour and Hausser, 1997; Kullmann, 2000; Overstreet et al., 2002). Particularly under conditions favouring the release of multiple vesicles, GABA can escape the synaptic cleft at a sufficiently high concentration to activate additional receptors that may be located adjacent to the immediate postsynaptic receptors ("perisynaptic"), at other postsynaptic clusters within the same synaptic contact, outside the synapse ("extrasynaptic"), or at nearby synapses. Such "spillover" of GABA onto extrasynaptic or perisynaptic (a6- and a4-containing) receptors has been shown to contribute a slow component to IPSCs in granule cells of the cerebellum (Rossi and Hamann, 1998) and dentate gyrus (Wei et al., 2003). A more unusual form of spillover (possibly with activation of receptors only at sites remote from the point of release) is thought to contribute to the generation of slowly rising and decaying IPSCs ("GABAa, slow") seen in hippocampal CA1 pyramidal neurons (Pearce, 1993; Banks et al., 2000). These events originate at distal dendritic sites (Banks et al., 1998) and appear to reflect the activity of a unique population of interneurons and the activation of receptors with a different subunit composition (a5-containing) to those that underlie fast IPSCs from somatic and dendritic sites (a2- and a1-containing, respectively (Prenosil et al., 2006)). Of note, similar slow IPSCs are also present in pyramidal neurons of the subiculum, which lack a5 subunits (Prenosil et al., 2006).

Tonic activity of GABAA receptors

GABAA receptors are not activated only during "phasic", i.e., synaptic signalling. Persistent or "tonic" activation of receptors, independent of any identified release event, occurs prior to synapse formation in embryonic (Valeyev et al., 1993; LoTurco et al., 1995; Owens et al., 1999; Demarque et al., 2002) and immature (Sipila et al., 2007) as well as newly derived postnatal neurons (Nguyen et al., 2003; Liu et al., 2005; Ge et al., 2006), as well as in a variety of mature (synaptic-ally connected) neurons (Mody, 2001; Semyanov et al., 2004; Farrant and Nusser, 2005). Tonic GABAergic signalling in mature neurons was first identified in granule cells of the cerebellar cortex (Kaneda et al., 1995; Brickley et al., 1996; Tia et al., 1996; Wall and Usowicz, 1997), where, in addition to blocking IPSCs, GABAa receptor antagonists were shown to decrease resting membrane conductance (see Farrant and Nusser, 2005). To date, similar GABA-mediated tonic conductances have been demonstrated in granule cells of the dentate gyrus (Nusser and Mody, 2002; Stell and Mody, 2002; Stell et al., 2003; Mtchedlishvili and Kapur, 2006), pyramidal cells and inhibitory interneurons in the CA1 region of the hippocampus (Semyanov et al., 2003; Scimemi et al., 2005; Shen et al., 2005), pyramidal neurons and inter-neurons in the somatosensory cortex (Yamada et al., 2006; Keros and Hablitz, 2005), thalamocortical relay neurons of the ventral basal complex and dorsal lateral geniculate nucleus (Porcello et al., 2003; Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005), magnocellular neurosecretory neurons of the supraoptic nucleus (Park et al., 2006), and at axon terminals of retinal bipolar cells (Hull et al., 2006).

While some GABAA receptors exhibit a low but measurable probability of opening in the absence of agonist (see above), a variety of evidence suggest that the widely-observed tonic activity of GABAA receptors reflects, in some neurons atleast, the persistent ligand-induced activation of specific receptor populations by low concentrations (nM to mM) of ambient GABA (see Cavelier et al., 2005; Farrant and Nusser, 2005; Santhakumar et al., 2006; but see McCartney et al., 2007). For different neurons, the concentration of ambient GABA to which they are continuously exposed may differ, but it is ultimately dictated by the effectiveness of Na+- and CP-dependent GABA uptake carriers. Often, such transporters are viewed as simply removing GABA from the extracellular space. In so doing, they act to limit spillover of GABA molecules from the synapse, alter the GABA concentration waveform in the synaptic cleft and ensure that "resting" GABA concentration remains sufficiently low to limit de-sensitization of synaptic receptors (Roepstorff and Lambert, 1994; Overstreet et al., 2002; Keros and Hablitz, 2005). However, transporters also have a significant and dynamic influence on ambient GABA (Richerson and Wu, 2003). This is because the extracellular concentration of GABA at which they are at equilibrium is dictated by the membrane potential and the transmembrane gradients for the transporter substrates (GABA, Na+, Cl"), each of which can vary (Overstreet and Westbrook, 2001; Richerson and Wu, 2003; Wu et al., 2003, 2006; Allen et al., 2004). In the extreme case, transporters can operate in reverse, and could, therefore, act as a source of GABA (Attwell et al., 1993). In general, such reversal does not contribute to the generation of tonic GABAa receptor activity. Rather, blockade of transport increases the magnitude of the tonic current (Wall and Usowicz, 1997; Nusser and Mody, 2002; Jensen et al., 2003; Rossi et al., 2003; Semyanov et al., 2003; Sipila et al., 2004; Keros and Hablitz, 2005; Scimemi et al., 2005; Sipila et al., 2007), suggesting that, in some cases at least, normally functioning uptake fails to reduce the ambient GABA concentration below that capable of activating certain GABAA receptors.

The identity of GABAA receptors that generate tonic activity has been established in several neuronal populations (see Farrant and Nusser, 2005). Of central importance in many cases is the 8 subunit. Unlike the al, a2, a3, a6, ß2/3 and g2 subunits, which are found throughout the neuronal membrane but are highly enriched in GABAergic synapses (Craig et al., 1994; Nusser et al., 1995a, b, 1998; Somogyi et al., 1996; Fritschy et al., 1998; Brunig et al., 2002), the 8 subunit has been shown to be present exclusively in extrasynaptic and perisynaptic locations (Nusser et al., 1998; Wei et al., 2003). This finding is consistent with the idea that a g subunit (most commonly the g2) is indispensable for synaptic clustering of GABAa receptors (Essrich et al., 1998; Baer et al., 1999; Schweizer et al., 2003; Alldred et al., 2005; see also, Fritschy and Brunig, 2003; Luscher and Keller, 2004). The 8 subunit co-assembles with the a6 subunit in cerebellar granule cells and with the a4 subunit in several areas, including the thalamus and dentate gyrus (Barnard et al., 1998; Sur et al., 1999). The initial suggestion that extrasy-naptic a6- and 8-containing receptors might mediate the tonic conductance in cerebellar granule cells (Brickley et al., 1996; Wall and Usowicz, 1997; Nusser et al., 1998) followed the recognition that the postnatal development of the conductance mirrored the delayed expression of these subunits, and that a6- and 8-containing receptors display a high affinity for GABA and desensitize slowly and less extensively than aßg receptors (Saxena and Macdonald, 1996; see also Farrant and Nusser, 2005). Subsequent studies showed that the tonic conductance in cerebellar granule cells is reduced by the a6-selective antagonist furosemide (Hamann et al., 2002) and abolished in mice lacking either the a6 (Brickley et al., 2001) or 8 subunits (Stell et al., 2003). Likewise, deletion of the 8 subunit (and concomitant loss of a4 expression (Peng et al., 2002)) reduces the tonic receptor activation in granule cells of the dentate gyrus (Stell et al., 2003). The 8 subunit (in combination with a4 and ß2 subunits) is also implicated in the generation of the tonic conductance seen in thalamocortical neurons of the dorsal lateral gen-iculate and ventral basal thalamus (Belelli et al., 2005; Cope et al., 2005; Jia et al., 2005).

Although 8-containing receptors, by virtue of their high affinity for GABA and limited desensitization, would appear to underlie most of the tonic GABAA receptor activity seen in mature neurons, other receptor subtypes have been suggested to sustain a similar role. Although it seems obvious, it should be remembered that all GABAA receptors (even those exhibiting significant desensitization) could contribute to the generation of a tonic conductance if the ambient concentration of GABA were high enough to activate them. As discussed by Scimemi and colleagues (Scimemi et al., 2005), this is an important consideration when comparing different in vitro studies, some of which have measured tonic receptor activity in the presence of added GABA (Stell and Mody, 2002; Stell et al., 2003; see also McCartney et al., 2007), blockers of GABA uptake (Semyanov et al., 2003), or blockers of GABA metabolism (Wu et al., 2003; Caraiscos et al., 2004). It is also important with regard to the interpretation of the effects on tonic receptor activity of (potentially selective) positive allosteric modulators. While a lack of effect may be used to rule out the contribution of specific receptor subtypes, enhancement of the tonic conductance could reflect either a selective increase in the affinity of those receptors responsible for the conductance or a recruitment of receptor populations of lower affinity that do not ordinarily exhibit tonic activity. In hippocampal pyramidal neurons, for example, a5-containing receptors would appear to contribute to the tonic conductance only when the concentration of extracellular GABA is elevated, while in mature tissue under "normal" conditions (without blockade of GABA transaminase (Caraiscos et al., 2004)) the conductance is mediated by 8-containing receptors (Scimemi et al., 2005). These observations also point to a tonic activity that is potentially dynamic, with different GABAA receptor populations playing a role under different conditions, as might result from developmental changes, pharmacological intervention, changes in the efficacy of GABA uptake, physiologically relevant changes in network activity or pathological changes such as epilepsy (see Scimemi et al., 2005).

Neuronal ion regulation and the driving force for GABAa receptor-mediated currents

As noted above, GABAA receptors are permeable to two physiologically relevant anions, Cl" and HCO", with a HCO"/Cl" permeability ratio (PHCO" /PCl") of around 0.2-0.4. A frequent and erroneous assumption in the GABA literature is that this permeability ratio translates directly into identical quantitative relations of the respective anion currents during synaptic and tonic responses. With an intracellular pH (pH;) of about 7.1-7.2 and the above PHCO" /PCr ratio, the intraneuronal HCO" concentration ([HCO"]; ~15mM) produces an influence on EGABA that is equal to about 3-6 mM Cl". Hence, the role of HCO" in setting the value for IPSP reversal can be significant, not only quantitatively but also qualitatively. In neurons with a rather hyperpolarized resting membrane potential and low [Cl"];, the HCO"-mediated current component can, in fact, exceed the Cl"-me-diated current, which leads to HCO3"-dependent depolarizing IPSPs as shown in experiments on pyramidal neurons in adult neocortical slices (Kaila et al., 1993; see also Connors et al., 1982; Gulledge and Stuart, 2003).

The above observations have important consequences. They show that modestly depolarizing

IPSPs can be generated in certain neurons under resting, steady-state conditions (i.e., in the absence of activity-dependent redistribution of Ci"; see below) despite the presence of a functional Ci" extrusion mechanism and a Ci" equilibrium potential (ECi) that is more negative than the resting membrane potential. More generally, it is obvious that Egaba is always more positive than ECl, and this difference is accentuated when [Cl"]; is low. A corollary of this is that in immature neurons and other neurons that have a high [d"]i, the contribution of HCO" to EIPSP is negligible (Fig. 2).

Hyperpolarization of the membrane potential is a key property of the "classical" IPSPs described in the original work by Eccles and colleagues on the glycinergic mechanisms in spinal cord motoneurons (Burke, 2006) and in most textbooks of neuroscience in relation to both glycinergic and GABAa receptor-mediated responses. However, from the above it is evident that robustly hyperpolarizing IPSPs in resting neurons are probably not as common as is generally thought. Because of the HCO" permeability, the IPSP reversal potential can be very close to resting Vm in a number of cell types. This, however, does not imply an absence of an inhibitory effect of such IPSPs because the opening of the GABAa receptors leads to what is known as ''shunting inhibition": a decrease in the input resistance of the neuron with a consequent decrease in the membrane time constant and space constant of the target cell. Here, it is important to note that shunting inhibition is local and lasts only for the duration of the synaptic conductance change, whereas hyperpolarizing or depolarizing synaptic potentials outlast the originating conductance change, with their duration and spread dictated by passive membrane properties and voltage-gated channels. Thus, the decrease in time constant caused by shunting inhibition will suppress the temporal summation of simultaneously occurring excitatory inputs, while the decrease in space constant will suppress the spatial summation of such inputs. In fact, slightly depolarizing IPSPs such as those seen in hippocampal dentate granule cells may exert a more effective inhibitory action than hyperpolarizing ones. This is because the GABAa receptor current shows outward rectification, and the slope conductance of the current-voltage (I-V)

Intracellular [CI"] (mM)

Fig. 2. The effect of bicarbonate on the GABAA reversal potential. The relationship between the intracellular concentration of Cl— ([Cl—]i) and the reversal potential (E) for a purely Cl—-mediated event and one mediated by GABAA receptors (with permeability to both Cl— and HCO—) was calculated for different permeability ratios (PHCO— /Per) according to:

Intracellular [CI"] (mM)

Fig. 2. The effect of bicarbonate on the GABAA reversal potential. The relationship between the intracellular concentration of Cl— ([Cl—]i) and the reversal potential (E) for a purely Cl—-mediated event and one mediated by GABAA receptors (with permeability to both Cl— and HCO—) was calculated for different permeability ratios (PHCO— /Per) according to:

RT l Pa-[Cl—],- + Phco—[HCO—],-" F n Pcl—[Cl—]0 + Phco— [HCO—]0 , where R is the ideal gas constant, Tis the absolute temperature at 37°C and F is Faraday's constant. The extracellular concentration of Cl— ([Cl—]o) was assumed to be 130 mM, the extracellular concentration of HCO— ([HCO—]o) was assumed to be 25 mM, and the intracellular concentration of HCO— ([HCO—]i) was calculated to be 14.1 mM (assuming 5% ambient CO2 and an intracellular pH of 7.15). Note: Light grey (0.2), mid grey (0.3) and black (0.4).

relation is enhanced at depolarized values of Vm. In addition, the absence of a hyperpolarizing component also prevents the generation of "rebound" excitation and the associated synchronization in the firing of cortical principal neurons that is often seen after hyperpolarizing IPSPs (Bevan et al., 2002; Somogyi and Klausberger, 2005). Interestingly, a recent study on subthalamic nucleus (STN) neurons (Baufreton et al., 2005) has shown that hyperpo-larizing IPSPs increase the availability of voltage-gated Na+ channels that leads to a negative shift in the action potential (AP) firing threshold and to a dramatic enhancement of EPSP-AP coupling. This implies that GABAergic inhibition can prime STN neurons to respond more efficiently to excitatory input, which adds a new dimension to the conventional view that GABAergic transmission from the globus pallidus (GP) to the STN restrains the activity of the STN (Albin et al., 1989).

Tonic GABAA receptor activity results in a conductance increase that (just like a synaptically evoked increase) reduces the magnitude and duration of the voltage response to an injected current and increases the decrement of voltage with distance. The obvious difference between the two is that while one is spatially and temporally discrete, the other is distributed and persistent and leads to a maintained shunting inhibition. Predictably, therefore, tonic GABAA receptor activation makes a cell less easily excited in response to a depolarizing input, and shifts the relationship between excitation current and output firing rate to the right (i.e., to higher levels of current) (Brickley et al., 1996; Hamann et al., 2002; Chadderton et al., 2004; Cope et al., 2005). In the case of excitatory synaptic conductances (as opposed to step current inputs), the effect of shunting inhibition is to decrease the slope of the input-output relationship, producing a change in "gain" (Chance et al., 2002; Mitchell and Silver, 2003). Again, just as for phasic events, the polarity of currents evoked by tonic receptor activation is not fixed and will depend on the membrane voltage and the net movement of Cl" and HCO", and thus on the Cl" and pH regulation of the cell (see below). In thalamocortical neurons, for example, persistent GABAA receptor activation has been shown to produce a hyperpolarization of 1-2 mV, and block of the tonic receptor activity causes a depolarization that promotes a shift from a burst firing towards a tonic firing mode (Cope et al., 2005). By contrast, in immature neurons, tonic GABAA receptor activity can cause depolarization, as described below.

Neuronal chloride regulation

Neuronal Cl" homeostasis is mainly controlled by the SLC12A family of cation-chloride co-transporters (CCCs; Fig. 3). The CCCs are composed of glycoproteins (MW of monomers in the range of 120-200 kDa) with 12 putative transmembrane segments flanked on one side by a small intracellu-lar amino-terminus and by a large intracellular carboxy-terminus on the other (Payne et al., 2003). CCCs are secondary active transporters that do not directly consume ATP, but derive the energy for Cl" translocation from the Na+ and K+ gradients generated by the Na-K-ATPase. The Na-K-2Cl co-transporters (NKCCs) mediate Cl" uptake fuelled by the plasmalemmal Na+ gradient, and K-Cl co-transporters (KCCs) extrude Cl" under physiological conditions using the K+ gradient. Both types of transporters are electrically neutral and hence do not have direct effects on the membrane potential. So far, two NKCC (NKCC1 and NKCC2) and four KCC isoforms (KCC1-4) have been identified (Delpire and Mount, 2002; Payne et al., 2003; Gamba, 2005). From these, NKCC1 as well as KCC2 and KCC3 have been shown to be expressed and functional in neurons (Payne et al., 1996; Plotkin et al., 1997; Rivera et al., 1999; Boettger et al., 2003). Some immature neurons express KCC4, at least at the mRNA level (Li et al., 2002).

NKCCs are kinetically activated by phosphory-lation while KCCs are activated by dephos-phorylation (Russell, 2000; Gamba, 2005). In neurons, much work remains to be done to identify the specific signalling cascades that modify the kinetics of Cl" transporters (Kelsch et al., 2001; Payne et al., 2003; Khirug et al., 2005). The phos-phorylation cascade that targets NKCC is suppressed by intracellular Cl" that leads to a negative feedback loop that inhibits NKCC1 activity when a sufficiently high [Cl"]i is achieved (Lytle and Forbush, 1996). As for KCC2, an increase in [Cl"]; leads to an immediate increase in net Cl" efflux that is caused by the increase in the outward driving force for K-Cl co-transport and not by a change in the kinetics of the transporter (Payne et al., 2003; Khirug et al., 2005). Recent work points to a role for WNK3, a member of the WNK family of serine-threonine kinases, in the control of [Cl"]; by upregu-lation of NKCC1 and down-regulation of KCC2 activity (Kahle et al., 2005).

The KCC2 molecule has gained much attention since it has an exclusively neuron-specific expression and is responsible for the generation of "classical" hyperpolarizing IPSPs (Rivera et al., 1999; Hubner et al., 2001). KCC2 is operational under isotonic conditions (Mercado et al., 2006) and in cortical principal cells it is constitutively active

Cl~ extruding

Cl"

0 0

Post a comment