Metabotropic, but not allosteric, effects of neurosteroids on GABAergic inhibition depend on the phosphorylation of GABAA receptors

Neuroactive steroids (NASs) are synthesized within the brain and exert profound effects on behavior. These effects are primarily believed to arise from the activities of NASs as positive allosteric modulators (PAMs) of the GABA-type A receptor (GABAAR). NASs also activate a family of G protein–coupled receptors known as membrane progesterone receptors (mPRs). Here, using surface-biotinylation assays and electrophysiology techniques, we examined mPRs' role in mediating the effects of NAS on the efficacy of GABAergic inhibition. Selective mPR activation enhanced phosphorylation of Ser-408 and Ser-409 (Ser-408/9) within the GABAAR β3 subunit, which depended on the activity of cAMP-dependent protein kinase A (PKA) and protein kinase C (PKC). mPR activation did not directly modify GABAAR activity and had no acute effects on phasic or tonic inhibition. Instead, mPR activation induced a sustained elevation in tonic current, which was blocked by PKA and PKC inhibition. Substitution of Ser-408/9 to alanine residues also prevented the effects of mPR activation on tonic current. Furthermore, this substitution abolished the effects of sustained NAS exposure on tonic inhibition. Interestingly, the allosteric effects of NAS on GABAergic inhibition were independent of Ser-408/9 in the β3 subunit. Additionally, although allosteric effects of NAS on GABAergic inhibition were sensitive to a recently developed “NAS antagonist,” the sustained effects of NAS on tonic inhibition were not. We conclude that metabotropic effects of NAS on GABAergic inhibition are mediated by mPR-dependent modulation of GABAAR phosphorylation. We propose that this mechanism may contribute to the varying behavioral effects of NAS.

In addition to their accepted roles as GABA A R PAMs, there is an accumulation of evidence that NAS can exert metabotropic effects on the efficacy of GABAergic inhibition (7,8). Specifically, NAS enhance the phosphorylation of key regulatory residues within GABA A Rs, including serines 408/409 in the ␤3 subunit. Enhanced phosphorylation of the ␤3 subunit correlates with increased receptor membrane accumulation and a sustained potentiation of GABAergic inhibition (9 -11). Some studies have suggested that in addition to binding to GABA A Rs, NASs such as allopregnanolone (ALLO), bind to and activate a family of membrane progesterone receptors (mPRs) (12,13). These are a family of G protein-coupled receptors, of which some subtypes are highly expressed in the brain and have been shown to couple to G i/o or G s signaling pathways (12)(13)(14)(15)(16)(17). However, the role that these receptors play in mediating the effects of NAS on GABAergic inhibition is unknown.
Here we examine how NASs produce distinct metabotropic and allosteric effects on GABAergic inhibition. Our results demonstrate that NASs enhance tonic inhibition via mPR-dependent phosphorylation of Ser-408/9 in the GABA A R ␤3 subunit. Importantly, this mechanism is independent from the allosteric effects of NAS at the GABA A R.

mPR agonists potentiate GABA A R phosphorylation and plasma membrane accumulation
Previous studies have shown that endogenous and selected synthetic neurosteroids, including ALLO, tetrahydrodeoxycorticosterone (THDOC), and SAGE-516, exert sustained effects on GABAergic inhibition. In parallel, these respective agents all enhance phosphorylation of key regulatory sites within the intracellular domain of the ␤3 subunit of the receptor, serine residues 408 and 409 respectively (Ser-408/9) (9 -11).
Neurosteroids have been reported to activate a putative family of G protein-coupled mPRs. Five mPRs (␣, ␤, ␥, ␦, and ⑀) that couple to G i (␣, ␤, ␥) or G s (␦, ⑀) are expressed in mammals and at high levels in the brain (12)(13)(14)(15)(16)(17). To examine the role that mPRs may play in mediating the effects of neurosteroids on GABAergic inhibition, we exposed hippocampal slices to ORG OD 02-0 (ORG), which selectively activates mPRs but not nuclear progesterone receptors (14). Hippocampal slices were exposed to 300 nM ORG (its estimated EC 50 concentration for mPRs) or vehicle for 10 min (14), and its effects on the phosphorylation of Ser-408/9 were determined. To do so, SDS-soluble lysates were immunoblotted with a phospho-specific antibody against phosphorylated Ser-408/9 (pS408/9), 4 a phospho-independent antibody selective for the ␤3 subunit, together with an antibody against actin to control for loading. The ratio of pS408/9/␤3 immunoreactivity was then determined and compared between treatments as detailed previously (10, 18 -20). Under these conditions, ORG significantly increased Ser-408/9 phosphorylation to 138.9 Ϯ 20.0% of control ( Fig. 1A; p Ͻ 0.05; n ϭ 6 mice). Published studies have shown that Ser-408/9 within ␤3 are phosphorylated by a number of protein kinases, including both PKA and PKC (21). Thus, we examined the effects of ORG on GABA A R phosphorylation in the presence of 10 M GF 109203X (GFX), a selective inhibitor of PKC, and 1 M KT5720 (KT), a selective inhibitor of PKA, for 10 min. Incubation with GFX or KT alone did not significantly modify ORG-induced phosphorylation. However, their co-application prevented ORG-induced phosphorylation ( Fig. 1A; 94.7 Ϯ 17.85% of control, p ϭ 0.2811, n ϭ 6). Under similar experimental conditions, exposure of slices to 100 nM ALLO significantly increased Ser-408/9 phosphorylation to 120.5 Ϯ 3.2% of control ( Fig. 1B; n ϭ 3 mice, p Ͻ 0.01). Consistent with our studies using ORG, ALLO-induced potentiation of Ser-408/9 phosphorylation was abrogated using GFX/KT ( Fig. 1B; 99.4 Ϯ 2.8% of control, n ϭ 3 mice, p ϭ 0.7458). To further evaluate the role that mPRs may play in regulating GABA A R phosphorylation, slices were exposed to progesterone (P4), which activates both mPRs and nuclear progesterone receptors. In keeping with our results, P4 significantly increased Ser-408/9 phosphorylation to 146.6 Ϯ 21.9% of control (p Ͻ 0.05, n ϭ 3).
Using biotinylation, we determined whether, in parallel with increasing Ser-408/9 phosphorylation, ORG treatment modifies accumulation of GABA A Rs on the plasma membrane. Immunoblotting revealed that exposure to ORG significantly increased the plasma membrane levels of the ␤3 subunit to 141.4 Ϯ 11.52% of control ( Fig. 2; p Ͻ 0.01, n ϭ 6 mice) and the ␣4 subunit to 114.1 Ϯ 4.2% of control ( Fig. 2; p Ͻ 0.05, n ϭ 5 mice), without affecting membrane levels of the transferrin receptor (TfR). Importantly, our surface fractions were free of the cytosolic protein actin, demonstrating the robustness of our surface protein biotinylation procedure. Collectively, these experiments suggest that ORG and P4 mimic the effects of NAS on ␤3 subunit phosphorylation and induce a selective increase in the plasma membrane accumulation of GABA A Rs.

ORG does not allosterically modulate GABAergic inhibition
To examine the possible effects of mPR activation on GABAergic inhibition, we first assessed whether ORG directly modifies GABA A R activity. To test this, we expressed receptors composed of ␣4␤3 subunits in HEK293 cells. Using whole-cell patch-clamp recordings, the effects of ORG on the magnitude of GABA-induced currents (I GABA ) using an EC 20 concentration of GABA (1 M) were assessed. In contrast to ALLO, 100 nM ORG did not potentiate I GABA (Fig. 3A; 116 Ϯ 17% of control, p ϭ 0.578, n ϭ 3). Consistent with published studies, 100 nM ALLO potentiated I GABA by 647 Ϯ 118% of control ( Fig. 3A; p Ͻ 0.01, n ϭ 4). To extend our experiments using HEK293 cells, we used cells that stably express GABA A Rs allowing the use of automated Q-patch technique to further examine the pharmacological effects of ORG on differing receptor subtypes as described previously (22). Here, we used LTK cells that 4 The designations used are: pS408/9, phosphorylated Ser-408/9; S408/9A, mutant S408A/S409A.  Figure 1. ORG modulates GABA A R phosphorylation. A, hippocampal slices were exposed to vehicle, 100 nM ORG alone, or 100 nM ORG with 10 M GFX, 1 M KT5720 for 20 min. Slices were then lysed and immunoblotted with pS408/9, ␤3, and actin antibodies. The ratio of pS408/9 to ␤3 immunoreactivity was then determined and normalized to control (100%). *, significantly different from control, p Ͻ 0.05 (t test; n ϭ 6 mice). B, hippocampal slices were exposed to vehicle, 100 nM ALLO, or 100 nM ALLO in the presence of PKA/PKC inhibitors (GFX/KT). *, significantly different from control, p Ͻ 0.05 (t test; n ϭ 3 mice).

Metabotropic neurosteroid signaling
express receptors composed of ␣1␤2␥2 and CHO cells that express ␣4␤3␦ subunits, combinations that mimic the properties of receptor subtypes that mediate phasic and tonic inhibition, respectively. ORG from 0.01-1 M did not significantly modify I GABA for either combination ( Fig. 3B; n ϭ 3). Thus, these experiments suggest that, in contrast to ALLO, ORG does not directly modify GABA A R activity.
To further analyze the effects of ORG, we tested its allosteric effects on GABAergic inhibition in dentate gyrus granule cells using whole-cell patch-clamp recording. To measure effects on phasic inhibition, the effects of ORG on the properties of spontaneous inhibitory synaptic currents (sIP-SCs) was evaluated. To do so, the effects of vehicle or 300 nM ORG on the amplitude, decay, and frequency of sIPSCs were examined. sIPSC amplitude (83.5 Ϯ 7.9 pA for vehicle, 80.7 Ϯ 6.9 pA for ORG, n ϭ 7 cells), decay (8.7 Ϯ 0.6 ms for vehicle, 8.9 Ϯ 0.5 ms for ORG, n ϭ 7 cells), and frequency (1 Ϯ 0.2 Hz for vehicle, 0.9 Ϯ 0.2 Hz for ORG, n ϭ 7) were not significantly altered by ORG (Fig. 4A).
To measure possible effects on tonic inhibition, slices were exposed to drugs as outlined above, and the change in holding current was measured in 1 M GABA in response to the GABA A R antagonist, picrotoxin. Tonic current in the presence of 300 nM ORG was 75.4 Ϯ 10.4 pA (n ϭ 10 cells) compared with 67 Ϯ 13.2 pA (n ϭ 8 cells) in the presence of vehicle (DMSO) control (Fig. 4B). These results suggest that ORG does not have any direct or allosteric effects on GABAergic inhibition, which is consistent with our studies on recombinant receptors. In contrast to our results with ORG, published studies have shown that acute exposure of slices to NAS, including ALLO, slows sIPSC decay and increases tonic current (1,8).

Sustained exposure of hippocampal slices to ORG selectively increases tonic current
To assess the effects of sustained exposure of ORG on GABAergic inhibition, slices were incubated with 300 nM ORG for 15 mins and then washed for a further 30 -50 min before experimentation (Fig. 5A). Under these conditions, ORG significantly increased tonic inhibition in DGGCs from 52.6 Ϯ 13 pA (n ϭ 8 cells) in control to 91 Ϯ 12 pA ( Fig. 5B; p Ͻ 0.01, n ϭ 10 cells) following ORG exposure. Consistent with our biochemical studies, the effects of ORG on tonic current were blocked by co-incubating slices with GFX/KT5720. (Fig. 5B; 54.2 Ϯ 11 pA, n ϭ 9 cells). In contrast to these effects on tonic current, sustained exposure to ORG did not affect the amplitude (106.9 Ϯ 9.3 pA for vehicle, 105.1 Ϯ 13 pA for ORG, n ϭ 10 cells), frequency (1 Ϯ 0.4 Hz for vehicle, 1.1 Ϯ 0.4 Hz for ORG, n ϭ 10 cells), or decay times (10.7 Ϯ 0.7 ms for vehicle, 11 Ϯ 0.8 ms for ORG) of sIPSCs ( Fig. 5C; n ϭ 10 cells).

The metabotropic effects of ORG on tonic inhibition are dependent upon Ser-408/9
Our electrophysiological and biochemical studies suggest that ORG exposure results in elevated phosphorylation of the ␤3 subunit, elevations in plasma membrane accumulation of GABA A Rs, and a sustained increase in tonic inhibition. To further assess whether Ser-408/9 directly contribute to the effects of ORG on tonic current, we prepared hippocampal slices from S408/9A homozygote mice. These mice have been shown to have reductions in tonic but elevated phasic inhibition in DGGCs. Thus, we examined the effects of ORG exposure on tonic inhibition in DGGCs from S408/9A mice. In contrast to WT mice (Fig. 5), exposure of hippocampal slices to ORG did Hippocampal slices were exposed to vehicle or 300 nM ORG and then labeled on ice with NHS-Biotin. After lysis and purification on immobilized avidin, surface and total fractions were immunoblotted with antibodies against the GABA A R ␣4, and ␤3 subunits, TfR, or actin. The surface levels of ␣4, ␤3, and TfR were then determined and normalized to vehicle-treated controls. *, significantly different from control (p Ͻ 0.05, t test, n ϭ 6 -7 mice). The lines represent where individual lanes gel lanes were spliced together to make the respective image. . The magnitude of I GABA was then normalized to that seen in cells exposed to GABA alone. B, LTK or CHO cells expressing GABA A Rs composed of ␣1␤2␥2 or ␣4␤3␦ receptors, respectively, were subjected to Q-patch recording. The magnitude of I GABA induced by EC 20 GABA alone or in the presence of 1-1000 nM ORG was determined. Responses were then normalized to those seen with GABA alone.

Metabotropic neurosteroid signaling
not significantly modify tonic current in DGGCs from S408/9A mice. Tonic current was 71.4 Ϯ 6 pA (n ϭ 10 cells) in control compared with 67.5 Ϯ 10 pA (n ϭ 10 cells) following exposure to ORG (Fig. 6A). This suggests that Ser-408/9 in the ␤3 subunit are critical determinants for the metabotropic effects of ORG on tonic current.

The metabotropic effects of neurosteroids on tonic current are mediated by Ser-408/9
In common with ORG, THDOC and ALLO have been shown to induce sustained increases in tonic inhibition that parallel increased Ser-408/9 phosphorylation (11). Therefore, we examined whether these metabotropic effects are dependent upon Ser-408/9. First, we examined the effects of vehicle or THDOC on tonic current as detailed above. In DGGCs from S408/9A mice, 100 nM THDOC did not significantly modify tonic current ( Fig. 6B; control ϭ 68.7 Ϯ 9 pA, THDOC ϭ 72.6 Ϯ 13 pA, n ϭ 9 cells). Likewise, 100 nM ALLO was without effect on this parameter in the mutant mice ( We also assessed the role that Ser-408/9 play in mediating the effects of NAS on GABA A R plasma membrane accumulation. To do so, slices from WT and S408/9A mice were incubated with 100 nM ALLO for 20 min, and slices were subject to biotinylation. Surface and total fractions were then immunoblotted with antibodies against the ␣4 and ␤3 subunits. In WT mice, ALLO significantly increased the plasma membrane levels of both subunits ( Fig. 7; 142.5 Ϯ 18.3 and 121 Ϯ 8.5% of control, respectively; p Ͻ 0.005, n ϭ 5 mice). In contrast to this, ALLO was without effect on these parameters in slices from S408/9A mutant mice ( Fig. 7; 101.2 Ϯ 8.3 and 105 Ϯ 11.5% of control, respectively; p ϭ 0.350, n ϭ 6 mice). Collectively, these results suggest that the ability of NAS and mPR agonists to induce sustained effects on tonic current is dependent upon Ser-408/9 in the ␤3 subunit.

Ser-408/9 are not primary determinants for the effects of NAS on sIPSC
The role that Ser-408/9 plays in mediating the allosteric effects of ALLO on GABAergic inhibition was also tested. To do so, we compared the effects of ALLO on sIPSC properties measured in DGCCs from both genotypes (Table 1). sIPSC decay was significantly slowed to similar degrees in WT and S408/9A mice. sIPSC amplitude in S408/9A mice was reduced by ALLO,  A, the protocol used to measure the metabotropic effects of ORG on GABAergic inhibition is illustrated. B, sample traces are shown for tonic currents recorded from slices exposed to vehicle, 300 nM ORG, or 300 nM ORG ϩ 10 M GFX ϩ 1 M KT. These data were then used to determine tonic current under the varying conditions, and levels were compared with those seen in vehicletreated controls. C, the effect of ORG or vehicle on sIPSCs was determined. Sample traces and individual events are shown in the right-hand panel. These data were then used to determine sIPSC amplitude, decay, and frequency. *, significantly different from control, p Ͻ 0.01 (ANOVA; n ϭ 10 cells).

Metabotropic neurosteroid signaling
and sIPSC frequencies were unaffected by ALLO in either genotype. Collectively, these results suggest that the ability of ALLO to induce sustained effects on tonic current is mediated via Ser-408/9 in the GABA A R ␤3 subunit but that these residues are dispensable for the allosteric efficacy of ALLO.

The NAS antagonist GR3027 abrogates the allosteric but not metabotropic effects of NAS on GABAergic inhibition
To further explore the mechanisms underlying the metabotropic and allosteric effects of NAS on GABAergic inhibition, we used GR3027, a compound that has been shown to block allosteric effects of NAS on GABA A R receptor activity but not compromise receptor agonist activation (23,24). Consistent with this notion, 10 M GR3027 blocked the effects of 100 nM ALLO on sIPSC decay, 8.9 Ϯ 0.6 ms before and 8.1 Ϯ 0.8 ms after (n ϭ 6 cells) the acute co-application of ALLO and GR3027 (Fig. 8A). However, GR3027 did not impact the metabotropic effects of ALLO on tonic inhibition. Co-incubation of hippocampal slices with ALLO plus GR3027 increased tonic current to 97 Ϯ 10.6 pA (n ϭ 11 cells) compared with 60.6 Ϯ 7.5 pA (n ϭ 9 cells) in vehicle control (Fig. 8B). This result provides further evidence that distinct molecular mechanisms mediate the metabotropic and allosteric effects of NAS on GABAergic inhibition.

mPRs mediated the effects of ORG and NAS on GABA A R phosphorylation
Our results suggest that ORG and NAS exert sustained effects on tonic inhibition by promoting the phosphorylation of GABA A Rs on residues that include Ser-408/9 in the ␤3 subunit. Due to the lack of specific mPR antagonists, we examined the role these GPCRs play in mediating the effects of NAS on GABA A R phosphorylation using recombinant expression in HEK293 cells. First, we used reverse transcriptase quantitative  Figure 7. The ability of NAS to modulate plasma membrane accumulation of GABA A Rs depends upon Ser-408/9. Hippocampal slices from WT and S408/9A mice were incubated with 100 nM ALLO and subject to biotinylation followed by immunoblotting with ␣4 and ␤3 subunit antibodies. Surface levels of both subunits were then determined and normalized to those in vehicle-treated controls. *, significantly different from control, p Ͻ 0.05 (t test; n ϭ 6 mice).

Metabotropic neurosteroid signaling
PCR (RT-qPCR) to examine which mPRs were endogenously expressed in this cell line. The relative levels of mPR mRNAs were then compared by reference to a tubulin standard, and the results were then expressed as arbitrary units, as described previously (25). 2.2 Ϯ 0.4 and 5.9 Ϯ 0.8 arbitrary units of mPR␣ and mPR␤ mRNAs were detected in HEK293 cells, whereas those corresponding to ␥, ␦, and ⑀ were not. The effect of ALLO on phosphorylation of GABA A Rs composed of ␣4 and ␤3 subunits was then tested when transiently expressed in HEK293 cells. 100 nM ALLO was without effect on the phosphorylation of Ser-408/9, as measured using immunoblotting ( Fig. 9A; 105.4 Ϯ 8.4% of control, p ϭ 0.488, n ϭ 5). The subcellular localization of mPRs has been widely debated due to studies describing both intracellular (26,27) and surface expression of mPR (28,29). mPR␣ was found in intracellular compartments in HEK293 cells and COS-1 cells. Thus, to investigate whether the lack of ALLO-mediated phosphorylation of ␤3 subunits was due to the absence of mPRs at the plasma membrane, we used a murine mPR␣ construct modified at its C terminus with the FLAG epitope (mPR␣-FLAG) that demonstrated localization to the plasma membrane (Fig. 9B). Exposure of ALLO to HEK293 cells transfected with GABA A Rs ␣4/␤3 and mPR␣-FLAG resulted in increased phosphorylation of Ser-408/9 ( Fig. 9; 146 Ϯ 21.0% of control, p Ͻ 0.01, n ϭ 5). 300 nM ORG also increased phosphorylation of these residues to 139.2 Ϯ 11.6% of control ( Fig. 9D; p Ͻ 0.01, n ϭ 3), an effect that was occluded by inhibitors of PKA and PKC ( Fig. 9D; 92.06 Ϯ 17.1% of control, p ϭ 0.460, n ϭ 3). Consistent with the effect of ORG on ␤3 surface expression in hippocampal slices, ORG increased GABA A R plasma membrane accumulation of ␤3 subunits in HEK293 cells expressing GABA A Rs ␣4/␤3 and mPR␣-FLAG (Fig 10; 117.2 Ϯ 8.0% of control, p Ͻ 0.05, n ϭ 3).
Thus, these results suggest that NAS and ORG modulate GABA A R phosphorylation dependent upon the activation of mPRs.

Discussion
NAS are accepted to play key roles as endogenous modulators of GABAergic inhibition, in part by acting as GABA A R PAMs (1,8). Recent studies have suggested that, in addition to their role as PAMs, NASs exert sustained metabotropic effects on GABAergic inhibition via an underlying mechanism that remains ill-defined (10,30). In addition to its role as a GABA A R PAM, ALLO has been shown to bind and activate mPRs, which have been proposed to be a novel family of GPCRs that are highly expressed in the brain (12)(13)(14)(15)(16)(17). Although tools to manipulate the activity of this family of G protein-coupled receptors are limited, we made use of the only available selective agonist, ORG (14). ORG increased Ser-408/9 phosphorylation dependent upon PKA/PKC in acute hippocampal slices. Under the same experimental conditions, ALLO induced similar PKA/ PKC-dependent enhancement on Ser-408/9 phosphorylation. Moreover, ORG also increased the plasma membrane accumulation of GABA A Rs. However, in contrast to ALLO, ORG did not exhibit any efficacy as a GABA A R PAM. Consistent with these results, acute exposure of hippocampal slices to ORG did not modify tonic current or the properties of sIPSCs.
Next, we assessed whether ORG exerts any metabotropic effects on GABAergic inhibition by exposing slices to this agent, followed by extensive washing to measure effects on tonic and phasic inhibition. Consistent with our biochemical analysis, ORG increased tonic current in a mechanism that depended upon PKA and PKC. In contrast to this sustained exposure, ORG did not have a significant impact on sIPSCs. To assess whether the ability of ORG to exert its metabotropic effects is dependent upon its ability to modulate receptor phosphorylation, we made use of S408/9A mice. In contrast to WT mice, the ability of ORG to exert sustained effects on tonic current was lost in these mutant mice. Similar experiments revealed that Ser-408/9 in the GABA A R ␤3 subunit are critical in mediating the metabotropic effects of ALLO on tonic inhibition and their ability to modulate receptor trafficking. In contrast to this, and its ability of ALLO to slow sIPSC decay (1), a hallmark of their efficacy as GABA A R PAMs, was maintained in the mutant mice. To further evaluate the relationship between the metabotropic and allosteric signaling mechanism of NAS, we used GR3027, a neurosteroid antagonist that ablates their efficacy as PAMs (23,24). As predicted, this agent blocked the effects of ALLO on sIPSC kinetics but not on its ability to induce metabotropic effects on tonic inhibition.
Finally, we directly tested whether the ability of ORG and ALLO to modulate phosphorylation of GABA A Rs is actually dependent upon mPRs, a task that is complicated by the lack of specific mPR antagonists. Given the heterogeneity of mPR structure and the lack of knockout mice, we tested this hypothesis using recombinant expression. We found that HEK293 cells express mPRs, but these proteins are largely restricted to intracellular compartments. Accordingly, NAS did not significantly modify GABA A R phosphorylation in these cells. Previous studies exploring the subcellular localization of mPRs have  Table 1. B, hippocampal slices were exposed to vehicle, 100 nM ALLO, 10 M GR3027 for 15 min and washed extensively before tonic current was measured (see Fig. 5). *, significantly different from control, p Ͻ 0.05 (t test, n ϭ 5-9 cells). The red box alone represents data for 100 nM alone (see Fig. 6).

Metabotropic neurosteroid signaling
proven controversial, with some studies suggesting that their expression is restricted to intracellular compartments (26,27), whereas others suggest plasma membrane expression (28,29). However, in cells expressing mPR␣-FLAG, accumulation of mPR immunoreactivity was evident on the plasma membrane and facilitated NAS-and ORG-dependent phosphorylation of the ␤3 subunit on Ser-408/9, via a PKA-and PKC-dependent mechanism, similar to events seen in brain slices. Clearly, further experimentation is required to determine the mechanisms that regulate mPR signaling and their roles in determining the effects of NAS on neuronal excitability.
Collectively, our experiments suggest that neurosteroids exert profound metabotropic effects on tonic inhibition by promoting the phosphorylation of Ser-408/9 in the ␤3 subunit via the activation of mPRs via a mechanism distinct from their actions as GABA A R PAMs. NASs also potentiate the phosphorylation of Ser-443 in the ␣4 subunit, which at least in expression systems enhances the plasma membrane accumulation of GABA A Rs (10,31). It will be of interest to examine the role that Ser-443 plays in mediating the metabotropic effects of NAS on tonic inhibition. However, such studies await the production of suitable phospho-specific antibodies and mutant mice.
Besides modulating GABAergic inhibition, NASs have also been suggested to exert broader effects on neuronal survival, neuroprotection, and promoting neurogenesis (12,32,33). It is tempting to speculate that in part, these effects may be mediated via the ability of NAS to modulate mPR activation and perhaps the recently described efficacy of these agents to alleviate mood disorders (34). Finally, alterations in NAS signaling may also contribute to altered neuronal excitability seen in Fragile X-syndrome, where compromised tonic inhibition and modified GABA A R phosphorylation has been demonstrated (35). . mPRs mediate the effects of NAS and ORG on GABA A R phosphorylation. A, HEK293 cells transfected with ␣4/␤3 GABA A Rs were treated with either vehicle or 100 nM ALLO for 5 min and immunoblotted with pS408/9, ␤3, and actin antibodies. The ratio of pS408/9/␤3 immunoreactivity was then determined and normalized to control (100%). B, HEK293 cells transfected with mPR␣-FLAG or empty plasmid were labeled on ice with NHS-Biotin. After cells were lysed and protein was purified on immobilized avidin, surface and total fractions were immunoblotted with antibodies against FLAG, tubulin, or TfR. C, HEK293 cells co-transfected with ␣4/␤3 GABA A Rs and mPR␣-FLAG were treated with 100 nM ALLO (D) or 300 nM ORG (E) for 5 min or PKA and PKC inhibitors (10 M GFX, 1 M KT) with vehicle or 300 nM ORG for 15 min (10-min pretreatment with inhibitors). The ratio of pS408/9/␤3 immunoreactivity was then determined and normalized to control or GFX/KT alone (100%). *, significantly different from control, p Ͻ 0.05, n ϭ 3-5. and mPR␣-FLAG were treated with vehicle or 300 nM ORG for 10 min. Cells were fixed with PFA (no permeabilization), and surface ␤3 subunits were labeled using ␤3 antibodies conjugated to GFP secondary antibodies. In each panel the scale bar represents 20 m. Cells were refixed and permeabilized, and intracellular ␤3 subunits were labeled with ␤3 antibodies conjugated to RFP secondary antibodies. Surface ␤3 levels were quantified by dividing surface levels by the sum of surface and intracellular levels of ␤3. *, significantly different from control p Ͻ 0.05 (t test, n ϭ 20 cells from four independent transfections).

Metabotropic neurosteroid signaling
libitum. All animal protocols were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the institutional animal care and use committee of Tufts University. S408/9A mice were maintained on the C57BL/6J background and have been back-crossed for 15 generations (20).
LTK cells were stably transfected with the ␣1␤2␥2 subunits of the GABA receptor, and CHO cells were transiently transfected with the ␣4␤3␦ subunits via the Lipofectamine method. Cells were passaged at a confluence of about 50 -80% and then seeded onto 35-mm sterile culture dishes containing 2 ml of culture complete medium without antibiotics or antimycotics. Cells were cultivated at a density that enabled the recording of single cells without visible connections to other cells (22).

Western blotting
Slices treated with neurosteroids were snap-frozen, briefly thawed on ice, and homogenized using a hand-held motorized tissue grinder (Fisher, catalogue no. 12-1413-61) in lysis buffer containing the following: 20 mM Tris-Cl, pH 8.0, 150 mM NaCl, 5 mM EDTA, pH 8.0, 10 mM NaF, 2 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 0.1% SDS. 2% Triton X-100 (v/v) was added to the samples before solubilizing on a rotating wheel at 4°C for at least 2 h. HEK293T cells treated with neurosteroids were solubilized in lysis buffer (as described above) with no SDS and 1% Triton X-100. Samples were centrifuged at 15,000 rpm for 15 min, and protein concentration was determined using the Bradford assay. Samples were boiled in 4ϫ sample buffer for 1-3 min at 95°C. 25 g of protein was loaded on 8 -10% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes overnight at 4°C. Membranes were incubated with 6% milk in PBS-T for all antibodies except antiphospho-Ser-408/9 GABA A ␤3, for which 6% BSA in PBS-T was used. Blots were incubated with primary antibodies overnight at 4°C and secondary antibodies at room temperature for 1.5 h. Blots were developed using chemiluminescence on the charge-coupled device-based ChemiDoc XRS system (catalogue no. 1708265) and quantified using densitometric analysis on ImageJ. After controlling for gel loading with reference to actin immunoreactivity, the ratio of pS408/9/␤3 immunoreactivity was then determined and compared between treatments as detailed previously (10, 18 -20) These values were then used for statistical analysis (Student's t test or one-way ANOVA) on GraphPad Prism 7.

Biotinylation
Cells or slices were incubated on ice for 20 -30 min with 1 mg/ml EZ-link Sulfo-NHS-SS-Biotin (Thermo Fisher Scientific, catalogue no. 21331) in 1ϫ PBS or nACSF, respectively. Excess biotin was quenched in 2 ml of 50 mM glycine and 0.1% BSA in PBS/nACSF three times, followed by three washes in ice-cold PBS or nACSF. Cell and tissue lysates were prepared as described above. 350 -500 g of protein lysate were incubated with Pierce NeutrAvidin agarose (Thermo Fisher Scientific, catalogue no. 29200) for 18 -24 h at 4°C. Agarose was washed twice with 150 mM NaCl, 1% Triton lysis buffer and twice with 500 mM NaCl, 1% Triton lysis buffer, followed by a final wash with 150 mM NaCl 1% Triton lysis buffer. Bound material was eluted with sample buffer and subjected to SDS-PAGE and then immunoblotted with antibodies as described above.

RT-qPCR
RNA was extracted from HEK293T cells using the RNeasy Plus Mini Kit (Qiagen). cDNA was synthesized using the Super-

Metabotropic neurosteroid signaling
Script IV First-Strand Synthesis kit (Life Technologies) with random hexamers following the manufacturer's instructions. The gene dose was calculated based on the standard curve method relative to tubulin as described previously (25).

Immunocytochemistry
Coverslips with transfected HEK293T cells were treated with ORG OD 02-0 for 5 min and fixed in 4% PFA/PBS for 20 min. PFA was quenched with 0.1 M glycine for 5 min and washed three times with 1ϫ PBS. Surface and intracellular staining was then performed as described by Connolly et al. (38) using anti-GABA A ␤2,3 clone BD17 antibody (Millipore).

Image analysis
The outline of cells was drawn using the drawing tool in ImageJ, and the average integrated density of fluorescence for each channel (emission 488 and 568 nm) was determined. Surface levels were quantified as follows: surface fluorescence/ (surface fluorescence ϩ intracellular fluorescence). Surface levels were normalized to those in control-treated cells (38).

Brain slice preparation for electrophysiology
Brain slices were prepared from 3-5-week-old male C57BL/6J mice. Mice were anesthetized with isoflurane and decapitated, and brains were rapidly removed and submerged in ice-cold cutting solution containing 126 mM NaCl, 2.5 mM KCl, 0.5 mM CaCl 2 , 2 mM MgCl 2 , 26 mM NaHCO 3 , 1.25 mM NaH 2 PO 4 , 10 mM glucose, 1.5 mM sodium pyruvate, and 3 mM kynurenic acid. Coronal 310-m-thick slices were cut with the vibratome VT1000S (Leica Microsystems, St. Louis, MO). The slices were then transferred into an incubation chamber filled with prewarmed (31-32°C) oxygenated nACSF of the following composition: 126 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 26 mM NaHCO 3 , 1.25 mM NaH 2 PO 4 , 10 mM glucose, 1.5 mM sodium pyruvate, 1 mM glutamine, 3 mM kynurenic acid, and 0.005 M GABA bubbled with 95% O 2 , 5% CO 2 . Exogenous GABA was added to standardize ambient GABA in the slice and provide an agonist source for newly inserted extrasynaptic GABA A Rs. Slices were allowed to recover at 32°C for at least 30 min before exposure to neurosteroids. Hippocampal slices were incubated for 15 min in a chamber containing either control or neurosteroids dissolved in nACSF that did not contain kynurenic acid. Following this incubation, slices were transferred to a submerged, dual-perfusion recording chamber (Warner Instruments, Hamden, CT) on the stage of an upright microscope (Nikon FN-1) with a ϫ40 water immersion objective equipped with differential interference contrast/infrared optics. Slices were maintained at 32°C and gravity-superfused with nACSF solution (with kynurenic acid) throughout experimentation and perfused at a rate of 2 ml/min with oxygenated (O 2 /CO 2 , 95%/5%) nACSF. Slices were perfused for 30 -60 min before recordings were started.

Electrophysiological recordings
Whole-cell currents were recorded from the dentate gyrus granule cells in 310-m-thick coronal hippocampal slices. Patch pipettes (5-7 megaohms) were pulled from borosilicate glass (World Precision Instruments) and filled with intracellu-lar solution of the following composition: 140 mM CsCl, 1 mM MgCl 2 , 0.1 mM EGTA, 10 mM HEPES, 2 mM Mg-ATP, 4 mM NaCl, and 0.3 mM Na-GTP (pH 7.2 with CsOH). A 5-min period for stabilization after obtaining the whole-cell recording conformation (holding potential of Ϫ60 mV) was allowed before currents were recorded using an Axopatch 200B amplifier (Molecular Devices), low-pass-filtered at 2 kHz, digitized at 20 kHz (Digidata 1440A; Molecular Devices), and stored for offline analysis. For tonic current measurements, an all-points histogram was plotted for a 10-s period before and during 100 M picrotoxin application, once the response reached a plateau level. Recordings with unstable baselines were discarded. Fitting the histogram with a Gaussian distribution gave the mean baseline current amplitude, and the difference between the amplitudes before and during picrotoxin was considered to be the tonic current. The negative section of the all-points histogram that corresponds to the inward inhibitory synaptic currents was not fitted with a Gaussian distribution (39,40). Series resistance and whole-cell capacitance were continually monitored and compensated throughout the course of the experiment. Recordings were eliminated from data analysis if series resistance increased by Ͼ20%. sIPSCs were analyzed using the mini-analysis software (version 5.6.4; Synaptosoft, Decatur, GA). Minimum threshold detection was set to 3 times the value of baseline noise signal. To assess sIPSC kinetics, the recording trace was visually inspected, and only events with a stable baseline, sharp rising phase, and single peak were used to negate artifacts due to event summation. Only recordings with a minimum of 200 events fitting these criteria were analyzed. sIPSC amplitude and frequency from each experimental condition were pooled and expressed as mean Ϯ S.E. To measure sIPSC decay, we averaged 100 consecutive events and fitted the decay to a double exponential and took the weighted decay constant (). Statistical analysis was performed by using Student's t test (paired and unpaired where appropriate), where p Ͻ 0.05 is considered significant.

Analysis of GABA-induced currents in HEK293T cells
HEK293T cells expressing GABA A Rs composed of ␣4␤3 subunits were superfused, at a rate of 2 ml/min at 32-33°C, with an extracellular solution containing 140 mM NaCl, 5 mM KCl, 1.2 mM MgCl 2 , 2.5 mM CaCl 2 , 10 mM HEPES, and 11 mM glucose and adjusted to pH 7.4 with NaOH as detailed previously (31). Experiments were started 3-5 min after achieving the whole-cell configuration at Ϫ60 mV. GABA (1 M, ϳEC 50 ) and other drugs were applied via a fast-step perfusion system (Warner Instruments).