Synaptic GABA_A receptors are directly recruited from their extrasynaptic counterparts

Characterizing the properties of GABA_A receptors engineered with the BBS and ecliptic pHluorin

Studies on the dynamics of cell-surface GABA_A receptor populations have been compromised by the lack of suitable high-affinity probes. To measure GABA_A receptor endo- and exocytosis directly, we have engineered the receptor β3 subunit with the bungarotoxin binding site (BBS) peptide WRYYESSLEPYPD. The BBS is derived from the α subunit of the muscle nicotinic acetylcholine receptor and has been established to bind Bgt with an affinity of approximately 3 nM (Scherf et al, 2001; Katchalski-Katzir et al, 2003). The BBS together with a pHluorin reporter separated by a 13-amino-acid alanine/proline linker was added between amino acids 4 and 5 of the mature β3 subunit (Figure 1A). The resulting construct, ^pHBBSβ3, was then expressed in HEK-293 cells and 48 h after transfection SDS-soluble lysates were subjected to immunoblotting with anti-β3 antibody. This revealed the presence of a major band of 87 kDa only in cells expressing ^pHBBSβ3, in agreement with the predicted molecular mass of this subunit (Figure 1B).

We further assessed the ability of ^pHBBSβ3 subunits to assemble with receptor α1 and γ2 subunits to form functional heteromeric α1β3γ2 receptors using patch clamp recording (Kittler et al, 2000b; Bedford et al, 2001; Jacob et al, 2005). Fast application of GABA produced inward currents (I_GABA) for cells expressing receptors composed of α1β3γ2, α1^pHβ3γ2 or α1^pHBBSβ3γ2 subunits (Figure 1C). Using dose–response analysis, it was evident that these differing receptors exhibited similar EC₅₀ values for GABA of 9.89±2.30, 2.38±0.4 and 6.12±1.94 μM (mean±s.e.m.; n=6–8) respectively in addition to maximal currents, strongly suggesting similar receptor densities on the plasma membrane for these varying subunit combinations (Figure 1D). To measure the incorporation of the γ2 subunit into functional receptors, we analyzed modulation by benzodiazepines, a property critically dependent upon incorporation of this subunit into heteromeric αβγ2 receptors (Rudolph et al, 2001; Sieghart and Sperk, 2002). Flurazepam at a concentration of 100 nM produced similar robust enhancements of I_GABA at EC₂₀ GABA concentrations of approximately 220% for receptors composed of α1β3γ2, α1^pHβ3γ2 or α1^pHBBSβ3γ2 subunits (Figure 1E and F). Together, these observations suggest that the addition of the BBS and pHluorin reporters to the N-terminus of the β3 subunit does not compromise receptor functional expression, a finding consistent with prior studies using GABA_A receptors engineered with N-terminal reporters GFP, pHluorin or myc reporters (Connolly et al, 1996; Kittler et al, 2000b; Taylor et al, 2000; Jacob et al, 2005).

GABA_A receptors modified with BBS bind Bgt with high affinity

The ability of GABA_A receptors incorporating ^pHBBSβ3 subunits to bind Bgt was analyzed by exposing live transfected HEK-293 cells to varying concentrations of ¹²⁵I-Bgt at 37°C for 10 min. After extensive washing, the level of bound Bgt was measured (Figure 2A). This revealed that cells expressing GABA_A receptors incorporating ^pHBBSβ3 subunits bound Bgt with a K_D of 7.85±2.41 × 10⁻⁹ M (mean±s.e.m.; n=3), consistent with previous studies on the BBS or AMPA receptors modified with this sequence (Scherf et al, 2001; Katchalski-Katzir et al, 2003; Sekine-Aizawa and Huganir, 2004). No significant binding was observed in cells expressing GABA_A receptors containing wild-type β3 subunits (Figure 2A).

To visualize cell-surface GABA_A receptor populations, live expressing HEK-293 cells were stained with rhodamine-conjugated Bgt (Rd-Bgt) and then visualized using confocal microscopy. This imaging revealed robust labelling of the surface of cells exhibiting endogenous green fluorescence, indicating expression of GABA_A receptors incorporating ^pHBBSβ3 subunits (Figure 2B). No binding was observed in non-transfected cells. To examine the stability of Rd-Bgt binding, cells expressing receptors incorporating ^pHBBSβ3 subunits were labelled for 10 min with Rd-Bgt, washed and incubated at 15°C to block endo- and exocytosis and subjected to live confocal imaging. The intensity of Bgt staining was then measured at the periphery of expressing cells over time (Figure 2B and C) and compared to the signal seen at zero time, which was given a value of 100%. Rd-Bgt labelling at the surface of HEK-293 cells remained unaltered over a 200 min time course, suggesting that Rd-Bgt complexes are stable (Figure 2D).

Next we analyzed the ability of Rd-Bgt to detect GABA_A receptors incorporating ^pHBBSβ3 subunits when expressed in hippocampal neurons. 8–10-Div neurons were transfected with ^pHBBSβ3 and 4–7 days later live neurons were exposed to Rd-Bgt, fixed and subjected to confocal microscopy. Neurons expressing ^pHBBSβ3 subunits exhibited endogenous green fluorescence and also significant levels of cell-surface Rd-Bgt binding, whereas the background labelling to non-fluorescent untransfected neurons was negligible (Figure 3A). Rd-Bgt staining in transfected neurons was blocked by excess of unlabelled Bgt (Figure 3B). To further control for possible binding of Bgt to endogenous neuronal AChRs, we examined the effects of curare on Rd-Bgt staining in neurons expressing ^pHBBSβ3 subunits. Although curare blocks Bgt binding to neuronal AChRs containing α7 subunits, it does not block binding to the BBS (Schoepfer et al, 1988; Sekine-Aizawa and Huganir, 2004). In agreement with this, 100 μM curare did not modify Rd-Bgt binding to neurons expressing ^pHBBSβ3 subunits, suggesting that fluorescent Bgt binding sites in our experiments represent GABA_A receptors and not endogenous AChRs (Figure 3E and F). We also examined the stability of Rd-Bgt/GABA_A receptor complexes in hippocampal neurons by incubating Rd-Bgt-stained neurons at 15°C. The level of Rd-Bgt staining in 1–2 μm² areas of interest was measured over a time course of 200 min (Figure 3E and F) with levels evident at zero time being given a value of 100%. No statistically significant loss of Rd-Bgt staining was seen over this time period (Figure 3G), suggesting that Rd-Bgt/GABA_A receptor complexes are stable on the surface of neurons.

The accumulation of GABA_A receptors incorporating ^pHBBSβ3 at synaptic sites in hippocampal neurons was also addressed. Expressing neurons were labelled with Rd-Bgt and fixed. Presynaptic terminals were then stained with antibodies against synapsin. Consistent with published studies analyzing the distribution of endogenous β3 subunits, Rd-Bgt staining on the surface of hippocampal neurons was evident as distinct clusters in addition to diffuse staining (Danglot et al, 2003; Jacob et al, 2005). We next calculated the number of these receptor clusters that were apposed to synapsin staining and therefore judged to be synaptic, as it has been documented that clusters containing the β3 subunit are found at both synaptic and extrasynaptic sites (Danglot et al, 2003; Jacob et al, 2005). A total of 24.6±6.4% of these clusters were apposed to synapsin staining and were therefore judged to be synaptic (Figure 3H; 250 clusters counted, mean±s.e.m.; n=5). Similar levels of synaptic targeting for endogenous GABA_A receptors containing the β3 subunit have been previously reported in 12–14-Div hippocampal neurons (Danglot et al, 2003; Jacob et al, 2005).

Extrasynaptic GABA_A receptors exhibit short residence times on the cell surface

We next compared the relative residence times on the cell surface of synaptic and extrasynaptic GABA_A receptors. Neurons expressing receptor ^pHBBSβ3 subunits were labelled with Rd-Bgt and incubated at 37°C for up to 90 min in the presence of 50 nM unlabelled Bgt in the external media. At differing time points, neurons were fixed and then incubated with synapsin antibodies to differentiate synaptic and extrasynaptic GABA_A receptors (Kittler et al, 2001; Jacob et al, 2005). At zero time, abundant diffuse extrasynaptic Rd-Bgt staining was seen in addition to labelled synaptic receptors (Figure 4, panels 1 and 2). However, a consistent finding after 90 min was an extensive loss of extrasynaptic Rd-Bgt labelling (Figure 4, panels 3 and 4).

To quantify this apparent loss of extrasynaptic receptors, we measured Rd-Bgt staining in 1–2 μm² areas of interest representing synaptic receptors and extrasynaptic receptors that were at least 2 μm from the nearest synapses and within the same neuron (Figure 4, panels 3 and 4). The ratio of extrasynaptic-to-synaptic Rd-Bgt staining (Bgt-ES/Bgt-S) was then determined and ratios at zero time were given an arbitrary value of 1.0. This ratio decreased rapidly with time of incubation at 37°C, reaching statistical significance after 10 min and steady state at 60 min (Figure 5A and B). At 90 min, this ratio had significantly decreased to 0.42±5.2% (Figure 5B; P<0.01; Student's t-test; n=6). Given the high stability of Bgt/GABA_A receptors complexes (Figures 2 and 3), our results demonstrate that extrasynaptic GABA_A receptors containing ^pHBBSβ3 have much shorter residence times on the cell surface as compared to their synaptic counterparts.

Extrasynaptic GABA_A receptors undergo clathrin-dependent endocytosis

A number of studies have illustrated that GABA_A receptors undergo constitutive clathrin-dependent endocytosis. We therefore sought to determine whether the time-dependent loss of extrasynaptic receptors seen in Figure 4 represents their preferential endocytosis. To do so, hippocampal neurons expressing ^pHBBSβ3 subunits were preincubated for 1 h with 50 μM myristoylated membrane-permeable p4 peptide (M-p4), a potent inhibitor of GABA_A receptor endocytosis (Kittler et al, 2000a, 2005; Chen et al, 2005) before labelling with Bgt. Pretreatment of neurons with M-p4 appeared to block the loss of extrasynaptic receptors seen in untreated neurons (Figure 4, panels 5 and 6). After quantifying these results, it was evident that M-p4 blocked the time-dependent decrease in Bgt-ES/Bgt-S ratio seen in controls, an effect that was not replicated with a scrambled myristoylated version of p4 peptide (M-Sc) or non-myristoylated (non-cell permeable) p4 peptide (P4) (Figure 5A and B).

To further analyze the intracellular fate of internalized GABA_A receptors, neurons were labelled for 5 min with Rd-Bgt, washed extensively and then incubated at 37°C for 120 min. After this time period, neurons were finally stained for 5 min with Alx-Bgt before fixation and analysis by confocal microscopy. The majority of the Rd-Bgt label over the incubation period at 37°C accumulated in perinuclear structures with endosomal-like morphology. These intracellular organelles also contained pHluorin fluorescence (Figure 5C–F). In contrast, Alx-Bgt staining was found extensively over the entire plasma membrane as diffuse staining in addition to clusters (Figure 5C–F). Remaining cell-surface Rd-Bgt staining was also found within these clusters, consistent with its stabilization at synaptic sites, as illustrated in Figure 4. Given the high stability of GABA_A receptor Bgt complexes at 15°C, conditions at which both endo- and exocytosis (Figure 3) are blocked, our results suggest that internalized Bgt-GABA_A receptors are specifically targeted to endosomal structures, where they are capable of rapid recycling back to the plasma membrane or are targeted for lysozomal degradation (Kittler et al, 2000a, 2004).

We further analyzed the selective endocytosis of extrasynaptic GABA_A receptors by comparing their distribution with that of the clathrin AP2 adaptin, which directly binds to motifs within the intracellular domains of receptor subunits, facilitating their endocytosis (Kittler et al, 2000a, 2005; Chen et al, 2005). Neurons expressing ^pHBBSβ3 subunits were stained with antibodies against AP2 adaptin and presynaptic terminals were labelled with synapsin antibodies and then imaged using confocal microscopy. Punctate staining that incorporated both GABA_A receptor ^pHBBSβ3 subunits and AP2 immunoreactivity was found on both neuronal cell bodies and processes. Synaptic GABA_A receptors were also evident as measured by the colocalization of ^pHBBSβ3 fluorescence and synapsin staining (Figure 5C and D). We then calculated the number of synaptic GABA_A receptors that also contained immunoreactivity for the AP2 adaptin. Only 19.2±1.2% of synaptic GABA_A receptors exhibited detectable levels of AP2 staining (mean±s.e.m.; 165 synapses analyzed, n=4). This suggests that GABA_A receptor AP2 complexes are found primarily at extrasynaptic sites, consistent with their preferential endocytosis.

Imaging newly inserted GABA_A receptors using Bgt staining

To selectively measure GABA_A receptor insertion into the plasma membrane, we first blocked constitutive receptor endocytosis. Previous biochemical experiments have revealed that in excess of 25% of the total cell-surface receptor population are internalized within 15 min at steady state (Kittler et al, 2004). To do so, neurons were first treated with M-p4 and incubated with Rd-Bgt at 37°C. At varying time points, neurons were fixed and stained with synapsin antibodies to differentiate synaptic and extrasynaptic GABA_A receptor populations. Using this procedure, Rd-Bgt binding could be detected within 5 min of incubation and appeared to increase dramatically with further incubation at 37°C (Figure 6A–F). The specificity of this new Rd-Bgt staining was measured by coincubating neurons with Rd-Bgt and a large excess of unlabelled toxin. This treatment completely blocked the appearance of new Rd-Bgt staining compared to neurons incubated with Rd-Bgt alone (Figure 6G and H). Newly inserted GABA_A receptors were evident at both synaptic and extrasynaptic sites at 5 min, suggesting that insertion can occur throughout the entire surface (Figure 6B). However, as inhibitory synaptic sites represent only a small fraction of the neuronal surface, the majority of newly inserted receptors will be delivered to extrasynaptic sites.

To provide quantitative information on the delivery of newly inserted receptors to synaptic sites over time, we measured the ratio of newly inserted Rd-Bgt binding sites at synaptic and extrasynaptic sites (S-Bgt:ES-Bgt), with values evident at 5 min (zero time) given a value of 100%. In excess of 85% of all synapses analyzed in our experiments showed increased levels of Rd-Bgt staining, and these were chosen for further analysis, with the ratios seen at zero time being given a value of 1.0. This ratio increased slowly and reached a level of 1.41±0.08 at 30 min (Figure 6J), which was significantly different from zero time (P<0.001; n=4). This suggests that newly inserted receptors accumulate slowly at postsynaptic inhibitory synaptic sites.

Newly inserted extrasynaptic receptors are able to access synaptic sites

The slow accumulation of newly inserted GABA_A receptors at synaptic sites may reflect either slower direct insertion of GABA_A receptors at synaptic sites as compared to extrasynaptic sites or the diffusion of newly inserted extrasynaptic receptors to synaptic sites and their subsequent stabilization. To address these possibilities, we utilized pulse-chase analysis with Bgt conjugated to distinct fluorophores. Neurons were first pretreated with M-p4 peptide to block endocytosis and existing populations of receptors containing ^pHBBSβ3 subunits were blocked with unlabelled toxin. All GABA_A receptors inserted into the plasma membrane over a 5 min time period were then labelled with Rd-Bgt toxin, washed extensively and then incubated at 37°C in toxin-free media. At differing time periods, neurons were fixed and stained with VIAAT antibodies to visualize presynaptic inhibitory terminals and imaged by confocal microscopy. Newly inserted receptors were clearly visible after 5 min incubation with Rd-Bgt (zero time) and also after incubation at 37°C, but there appeared to be increased labelling of synaptic sites at later time points (Figure 7C and D). To quantify these results, we compared the level of Rd-Bgt staining at synaptic and extrasynaptic sites over time as outlined in Figures 4, 5 and 6, with ratios evident at zero time being given a value of 1.0. This ratio significantly increased with time of incubation at 37°C (Figure 7D), suggesting that newly inserted GABA_A receptors at extrasynaptic domains can over time move to synaptic sites.

DNA constructs

The ^pH/BBSβ3 subunit construct was derived from ^pHβ3 (Jacob et al, 2005). Briefly, ^pHβ3 was subjected to two rounds of mutagenesis by PCR to introduce the following changes: to introduce the 13-amino-acid BBS tag directly after the pHluorin tag and to add 12 alanines and one proline to separate the BBS tag from the pHluorin tag. The final construct contains the pHluorin tag between amino acids 3 and 4 of the mature β3 subunit. The pHluorin tag is followed by 13-amino-acid alanine/proline spacer.

Cell culture and transfection

Hippocampal neuron cultures were prepared as described previously (Kittler et al, 2000a, 2000b). Neurons were transfected with Effectene according to the manufacturer's specifications (Qiagen, Valencia, CA, USA). Briefly, 1 μg of DNA was used for every two 35 mm cell culture dishes containing hippocampal neurons. DNA was added to 150 μl of EC buffer and 8 μl of Enhancer. The solution was incubated for 5 min at room temperature and 75 μl of Effectene was added. After 10 min incubation at room temperature, the final solution was added to hippocampal neurons. The medium was changed after 6 h. HEK-293 cells were transfected using electroporation, as detailed earlier (Connolly et al, 1996, 1999).

Electrophysiology

HEK-293 cells were transfected with equimolar mixes of the relevant expression construct. After 18–48 h, coverslips containing the transfected HEK-293 cells were transferred to a recording chamber mounted on the stage of an inverted microscope and transfected cells identified by their endogenous fluorescence (Kittler et al, 2000b; Jacob et al, 2005). The external solution contained (in mM) NaCl 140, KCl 4.7, HEPES 10, NaHCO₃ 23, glucose 11, MgCl₂ 1.2 and CaCl₂ 2.5 (adjusted to pH 7.4 with NaOH) continuously oxygenated with a mixture of 95%O₂/5%CO₂. The recording chamber was perfused at a rate of 5–10 ml/min and maintained at 32°C (Kittler et al, 2000a, 2005; Chen et al, 2005; Jacob et al, 2005). The internal solution contained (in mM) KCl 140, CaCl₂ 1.0, HEPES 10, MgCl₂ 2 and ATP (adjusted to pH 7.2 with KOH). Pipettes had a resistance of 2–4.5 MΩ when filled with this internal solution. Patch clamp experiments were performed in the whole cell configuration using an Axopatch 200B amplifier. Series resistance and membrane capacitance were partially compensated (70–80%) and current traces were low-pass filtered at 2 kHz using a four-pole Bessel filter. The holding potential in all experiments was −50 mV. Drugs were rapidly applied to single cells using a modified U-tube, placed 50–100 μm away from the cell of interest (Kittler et al, 2000a, 2005; Bedford et al, 2001; Chen et al, 2005; Jacob et al, 2005).

Radiolabelled Bgt binding

Heteromeric GABA_A receptors containing ^pHBBSβ3 subunits were expressed in HEK-293 cells via electroporation (Connolly et al, 1996, 1999). After 48 h, cells were labelled with varying concentrations of ¹²⁵I-Bgtx (1000 Ci/mmol) at 37°C for 10 min before extensive washing, as detailed previously (Wells et al, 1998). After correcting for nonspecific binding using HEK-293 cells expressing heteromeric receptors containing wild β3 subunits, the level of bound Bgt was measured in a γ counter and then analyzed using the Origin software package to estimate the affinity of Bgt for ^pH/BBSβ3 subunits.

Fluorescent Bgt labelling

HEK-293 cells or hippocampal neurons were incubated with Rd-Bgt at a final concentration of 10 nM for 5 min at 37°C. Nonspecific binding was evident in the presence of 100 nM/ml unlabelled Bgt. The cells were then washed extensively and used for immunocytochemistry or live imaging. Blocking of existing cell-surface receptors was performed using 50 μg/ml of unlabelled Bgt for 10 min at 37°C.

GABA_A receptor endocytosis

14-Div hippocampal neurons expressing ^pH/BBSβ3 subunits were pretreated with 50 μM myristoylated p4 peptide, scrambled myristoylated p4 peptide, non-myristoylated p4 peptide or vehicle for 1 h. Neurons were then labelled with 10 nM Rd-Bgt for 5 min, washed extensively and placed in the fresh medium containing the peptides supplemented with 50 nM unlabelled Bgt. The neurons were then fixed at varying time points, subjected to immunocytochemistry with anti-synapsin antibodies and examined using confocal microscopy.

Visualizing GABA_A receptor membrane insertion

14-Div hippocampal neurons expressing recombinant GABA_A receptors were pretreated with M-p4 peptide for 1 h. All cell-surface ^pH/BBSβ3-containing GABA_A receptors were blocked with unlabelled Bgt for 10 min. The cells were then washed extensively and left to recover in the medium containing M-p4 peptide for 5 min. Newly inserted receptors were labelled with Rd-Bgt and fixed immediately after the recovery period (time point 0) or 15 and 30 min after the recovery period. Fixed cells were stained with antibodies against synapsin.

Immunocytochemistry

HEK-293 cells and transfected hippocampal neurons were processed for immunocytochemistry under either permeabilized or non-permeabilized conditions as described previously (Jacob et al, 2005; Kittler et al, 2000a, 2000b). The following primary antibodies were used: anti-vesicular polyclonal anti-synapsin 1 at 1:500 (Affinity BioReagents) and anti-β-adaptin (BD Biosciences).

Image acquisition and data analysis

Confocal images of immunostained neurons were taken using a × 60 objective, acquired with Bio-Rad software and analyzed with MetaMorph. A receptor cluster was defined as being approximately 0.5–2 μm in length and approximately two- to three-fold more intense than background diffuse fluorescence. Synaptic clusters were colocalized with or directly apposed to ynspasin staining. Clusters further than 1 μm from presynaptic marker staining were considered extrasynaptic (Jacob et al, 2005). All channels of an image were first background subtracted, then thresholded and stacked to determine apposition of postsynaptic receptor subunits with a presynaptic marker (Jacob et al, 2005). The threshold value was determined for each culture and used for all images from that culture.