We employed plasmid DNAs encoding the human α1 (pCIS), rat α3L (pcDNA3.1) and human β (pcDNA3.1) GlyR subunits, plus mouse neuroligin 2A (pNice) and rat gephyrin P1 (pCIS). Empty pEGFP plasmid was also transfected as an expression marker. Site-directed mutagenesis was performed using the QuikChange mutagenesis kit (Agilent Technologies) according to manufacturers’ instructions and the successful incorporation of mutations was confirmed by DNA sequencing. The plasmid DNAs were transfected into HEK293 cells via a calcium phosphate-DNA co-precipitation protocol. When expressing α1β GlyRs, we transfected the following DNA quantities into each 35 mm culture dish of HEK293 cells: 0.03 μg α1 GlyR, 1.5 μg β GlyR, 0.2 μg neuroligin 2A, 0.2 μg gephyrin, 0.1 μg EGFP. When expressing α3β GlyRs, we transfected the following DNA quantities into each 35 mm culture dish: 0.15 μg α3L GlyR, 1.5 μg β GlyR, 0.2 μg neuroligin 2A, 0.2 μg gephyrin, 0.1 μg EGFP. We have previously validated that these transfection conditions result in a high level of expression of heteromeric α1β and α3β GlyRs, respectively (Zhang et al., 2015).

Details of this method have recently been published (Dixon et al., 2015; Zhang et al., 2015). Briefly, E15 timed-pregnant rats were euthanized via CO2 inhalation in accordance with procedures approved by the University of Queensland Animal Ethics Committee (approval number: QBI/203/13/ARC). Embryos were surgically removed and placed into ice cold Ca²⁺-Mg²⁺-free Hank’s Balanced Salt Solution under sterile conditions. The spinal cords were then dissected out and pinned at the wider proximal end while meninges were carefully detached. The dissected neurons were then triturated, centrifuged and resuspended in Dulbecco’s Modified Eagles Medium supplemented with 10% fetal bovine serum. Between 40,000 and 80,000 neurons were plated onto each 12 mm poly-D-lysine-coated coverslip in 4-well plates. Neuronal cultures were always maintained in a 5% CO₂ incubator at 37°C. After incubation for 24 h the entire Dulbecco’s Modified Eagles Medium supplemented plus 10% fetal bovine serum medium was replaced with Neurobasal medium including 2% B27 and 1% GlutaMAX supplements. A second (and final) feed 1 week later replaced half of this medium. Neurons were used in co-culture experiments 1–4 weeks later. Heterosynaptic co-cultures were prepared by directly introducing transfected HEK293 cells onto the primary neuronal cultures 1-3 days prior to electrophysiological recording.

All experiments were performed at 20–22°C at a holding potential of -60 mV. Whole-cell recordings were obtained with a HEKA EPC10 amplifier (HEKA Electronics, Lambrecht, Gremany) and Patchmaster software (HEKA), with currents filtered at 4 kHz and sampled at 10 kHz. Series resistance was compensated to 60% and monitored throughout the recording. Patch pipettes (4–8 MΩ) made from thick-walled borosilicate glass (GC150F-7.5, Harvard Apparatus) were filled with an internal solution comprising (in mM): 145 CsCl, 2 CaCl₂, 2 MgCl₂, 10 HEPES, 10 EGTA and 2 MgATP, pH 7.4 with NaOH. Cells were perfused with extracellular solution comprising (in mM): 140 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 D-glucose, pH 7.4 with NaOH.

For outside-out recordings, pipettes were fired-polished to a resistance of ∼10 MΩ and filled with the same internal solution. Macroscopic currents in outside-out patches pulled from transfected HEK293 cells were activated by brief (<1 ms) exposure to agonists using a piezoelectric translator (Siskiyou). Currents were recorded using a Multiclamp 700B amplifer and pClamp10 software (Molecular Devices), filtered at 4 kHz and sampled at 10 kHz.

In accordance with standard practice in the field, solutions containing buffered concentrations of free Zn²⁺ were made by adding 10, 100, 500, and 1000 μM free Zn²⁺ to the standard extracellular solution + 10 mM tricine (Sigma-Aldrich) to produce buffered Zn²⁺ concentrations of 0.1, 1, 5, and 10 μM, respectively (Paoletti et al., 1997; Low et al., 2000; Miller et al., 2005b; Vergnano et al., 2014). The added Zn²⁺ concentration was aliquoted from a 10 mM aqueous ZnCl₂ stock solution on the day of the experiment. ZX1 (Strem Chemicals) was dissolved as a 100 mM stock in water and added to the control solution on the day of the experiment.

Analyses of IPSC amplitudes, 10-90% rise times and decay time constants were performed using Axograph X (Axograph Scientific). Single peak IPSCs with amplitudes of at least three times the background noise were detected using a semi-automated sliding template. Events were visually inspected and only well-separated IPSCs with no inflections in the rising or decay phases were included. Current decay phases were fitted with double-exponential functions and a weighted time constant was calculated from individual time constants (τ1, τ2) and their relative amplitudes (A1,A2) as follows: τ_weighted= (τ1 × A1+τ2 × A2)/(A1+A2). Statistical comparisons were performed using non-parametric tests (Prism 6, GraphPad Software, Inc.). For multiple comparisons we used either the Friedman test or the Wilcoxon matched-pairs signed rank test and for single comparisons we used the Mann–Whitney U-test. In all tests we took p < 0.05 to be statistically significant. All displayed error bars represent SEM.

The α1β GlyR is the main synaptic isoform (Lynch, 2009). Zn²⁺ potentiates EC₅₀ glycine-gated currents in recombinant α1β GlyRs with an EC₅₀ near 40 nM (Miller et al., 2005b) and inhibits with an IC₅₀ near 13 μM (Miller et al., 2005a). We also investigated the α1^H107Nβ and the α1^W170Sβ GlyRs. The H107N mutation eliminates Zn²⁺ inhibition and increases the Zn²⁺ potentiation EC₅₀ to 0.8 μM (Miller et al., 2005b). We reasoned that this receptor should respond only to high nanomolar or micromolar concentrations of Zn²⁺. The W170S mutation, which occurs naturally in a sporadic case of human hyperekplexia (Al-Futaisi et al., 2012), eliminates Zn²⁺ potentiation but leaves Zn²⁺ inhibition intact with an IC₅₀ near 5 μM (Zhou et al., 2013). We employed this mutant as both a negative control for the potentiating effect of Zn²⁺ and to determine whether Zn²⁺ inhibition (when isolated from potentiation) is capable of affecting IPSC parameters.

In order to precisely control the extracellular Zn²⁺ concentration, we buffered it with 10 mM tricine as described in Section “Materials and Methods.” We also applied 10 mM tricine with no added Zn²⁺ to buffer the free Zn²⁺ concentration to virtually zero, although as recently demonstrated for NMDARs, this may not remove tonic Zn²⁺ that is bound constitutively to low nanomolar affinity Zn²⁺ binding sites (Anderson et al., 2015). Thus, our starting hypothesis is that 10 mM tricine ablates freely diffusing Zn²⁺ but not constitutively bound Zn²⁺.

Sample recordings of spontaneous IPSCs from HEK293 cells expressing the indicated constructs are shown in Figure 1A. As these IPSCs were almost entirely abolished by 1 μM tetrodotoxin (not shown), we infer they were induced by spontaneous action potentials. The remaining miniature IPSCS were too small to analyze. Figure 1A (left) shows sample recordings of spontaneous IPSCs recorded from single cells expressing either the α1β GlyR (top), the α1^H107Nβ GlyR (middle) or the α1^W170Sβ GlyR (bottom). The left panels show control recordings in tricine-free extracellular solution, the center panels show the effect of 10 mM tricine and the right panels show the effect of adding 5 μM free Zn²⁺. Individual IPSCs from the recordings in A were normalized and digitally averaged to produce the global mean IPSC waveforms as displayed in Figure 1B. The mean IPSC amplitudes, decay time constants and 10–90% rise times for each construct recorded in tricine-free extracellular solution are presented in Figure 1C. We found that IPSCs mediated by α1^W170Sβ GlyRs exhibited significantly larger amplitudes, longer decay time constants and slower rise times relative to those mediated by α1β GlyRs. This indicates that the W170S mutation affects intrinsic receptor gating or synaptic clustering properties in addition to ablating Zn²⁺ potentiation.

Figures 1D–F summarizes the effects of adding either 10 mM tricine alone or 5 μM free Zn²⁺ to IPSCs mediated by the three constructs. We observed no effect 10 mM tricine on the IPSC amplitude mediated by any construct (Figure 1D). However, the removal of free Zn²⁺ significantly accelerated the decay rate of α1β GlyRs (Figure 1E) indicating that the free Zn²⁺ concentration is high enough to positively modulate these receptors. In contrast, 10 mM tricine had no effect on the IPSC decay rate mediated by α1^W170Sβ GlyRs. This not only provides a control for the GlyR-specificity of the Zn²⁺ potentiating effect, but also indicates that the low affinity inhibitory effect of Zn²⁺ does not appreciably affect IPSC parameters. All these findings are reinforced by the converse experiment whereby increasing free Zn²⁺ to 5 μM significantly slowed the decay rate in α1β GlyRs and α1^H107Nβ GlyRs but had no effect on α1^W170Sβ GlyRs (Figure 1E). The removal free Zn²⁺ also accelerated the rise times of IPSCs mediated by α1β GlyRs only (Figure 1F).

In an attempt to generalize our findings, we investigated the Zn²⁺-sensitivity of the other major adult synaptic GlyR isoform, the α3β, which mediates synaptic inhibition onto nociceptive neurons in superficial laminae of the adult spinal cord dorsal horn (Zeilhofer et al., 2012). Despite their importance in pain signal processing mechanisms and their relevance as novel therapeutic targets for inflammatory pain (Lynch, 2009; Zeilhofer et al., 2012), the Zn²⁺ sensitivity of α3β GlyRs has never been investigated. We thus repeated the experiment of Figure 1A on spontaneous IPSCs mediated by α3β GlyRs. Examples of normalized, averaged IPSCs recorded from one cell exposed sequentially to standard extracellular solution containing no tricine, 10 mM tricine, or 10 mM tricine plus 5 μM free Zn²⁺ are shown in Figure 2A, with averaged results summarized in Figures 2B–D. According to the non-parametric Friedman test, these data reveal that the removal of free Zn²⁺ by 10 mM tricine had no significant effect on any IPSC parameter. However, as all data distributions in Figure 2B satisfied the D’Agostino-Pearson omnibus test for normality (using p < 0.05 as cutoff), we performed a repeated measures one-way ANOVA on decay time constants of the control versus 10 mM tricine and 10 mM tricine plus 5 μM free Zn²⁺ experimental conditions. As 10 mM tricine significantly accelerated the IPSC decay rate according to this test (p < 0.05, n = 8), we conclude that the endogenous levels of Zn²⁺ in artificial synapses significantly prolong the decay time constant of α3β GlyR-mediated IPSCs.

The α1β GlyR Zn²⁺ EC₅₀ and IC₅₀ values cited above were obtained under steady-state applications of Zn²⁺ and EC₅₀ glycine (Miller et al., 2005a,b). As these conditions do not apply to the synapse, we calibrated the Zn²⁺ sensitivity of the GlyR constructs using fast solution exchange to mimic synaptic activation conditions. The glycine concentration in the synaptic cleft has been estimated to reach a peak of 1–3 mM in embryonic zebrafish hindbrain neurons (Legendre, 1998) and 2.2–3.5 mM in adult rat spinal motor neurons (Beato, 2008). Glycine is cleared from the cleft with a time constant of 0.6–0.9 ms (Beato, 2008). Considering these parameters, we simulated synaptic activation conditions by applying 1 mM glycine with or without free Zn²⁺ for 0.5–1 ms to outside-out macropatches expressing the GlyR isoform of interest. To calibrate and optimize the solution application system, we rapidly switched the solution perfusing an open patch pipette between standard extracellular solution and an extracellular solution that had been diluted by 50% with distilled water. Figure 3A shows an example of an open pipette response, indicating the typical solution exposure profile. We performed this control regularly to ensure that the solution switching rate remained constant within and between experiments.

Contaminating Zn²⁺ released into solution from plastic and glass laboratory ware is thought to reach concentrations as high as 100 nM (Kay, 2004), which is high enough to modulate steady-state GlyR currents activated by low glycine concentrations (Miller et al., 2005b; Cornelison and Mihic, 2014). We determined whether Zn²⁺ contamination might have contributed to the results of Figures 1 and 2 via two sets of control experiments. The first involved switching rapidly between a control (putatively Zn²⁺-contaminated) solution and a Zn²⁺-chelating solution. In this experiment the control and test solutions both comprised our standard extracellular solution plus 1 mM glycine, although the test solution also contained either 10 mM tricine or 100 μM ZX1 (Perez-Rosello et al., 2015). A sample recording from a single macropatch containing α1β GlyRs suggests the chelators had little effect on decay rate (Figure 3B) and the averaged result confirms the lack of a significant effect on either the decay or rise times (Figures 3C,D). It is evident, however, that ZX1 showed a tendency to decrease the decay time constant that may become significant with a larger number of observations. If so, this effect could be due to either (1) a direct allosteric inhibitory effect on the GlyR or (2) the removal of tonically bound Zn²⁺ from the GlyR. The second possibility will be investigated below.

In a second control experiment, we initially exposed macropatches to a buffered 100 nM Zn²⁺ solution and switched this rapidly to either a control (putatively Zn²⁺-contaminated) solution (n = 10 macropatches) or to a Zn²⁺-chelating solution (i.e., control + 10 mM tricine, n = 3 macropatches). Again, no significant differences in rise times or decay rates were observed in either experiment (p > 0.05 by Mann–Whitney U-test). Together, these results strongly suggest that contaminating Zn²⁺ in our extracellular solution had no significant effect on IPSC rise or decay rates.

To mimic synaptic activation, we employed standard extracellular solution without added tricine as the control solution to replicate the conditions employed in Figures 1 and 2. Test solutions had the same composition except that we added 1 mM glycine either alone or together with 10 mM tricine and 100 nM, 1 or 10 μM free Zn²⁺. Examples of current responses to the sequential application of each of these four solutions to α1β, α1^H107Nβ, and α1^W170Sβ GlyRs are shown in Figure 4A, with averaged weighted deactivation time constants shown in Figure 4B. The averaged individual time constants (τ1, τ2) and their relative amplitudes (A1, A2) for each experimental condition are presented in Supplementary Table S1. For the α1β GlyR, we observed no change in the deactivation rate or 10–90% rise time at 100 nM Zn²⁺. We observed a statistically significant slowing in the deactivation rate at 1 μM and a dramatic slowing in deactivation rate at 10 μM free Zn²⁺. We also observed a statistically significant slowing in the deactivation rate at 10 μM free Zn²⁺at the α1^H107Nβ GlyR. At 10 μM Zn²⁺, the percentage increase in the macropatch current deactivation time constant at both receptors (α1β: 76 ± 41%; α1^H107Nβ: 76 ± 19%) was comparable to the magnitude of the IPSC decay time constant increase from 10 mM tricine-containing to standard extracellular solution (α1β: 74 ± 13%, α1^H107Nβ: 50 ± 24%). In contrast, at the α1^W170Sβ GlyR we observed no change in IPSC magnitude, decay rate or rise time at free Zn²⁺ concentrations up to 10 μM (Figures 4A–C). Together, these results suggest that the free synaptic Zn²⁺ concentration reaches a concentration of at least 1 μM following a single synaptic stimulation. The lack of an inhibitory effect of 10 μM Zn²⁺ on α1^W170Sβ GlyRs is most likely due to a combination of the slow onset of Zn²⁺ inhibition and the brief exposure time to Zn²⁺ under synaptic stimulation conditions.

We next considered the possibility that the magnitude of the Zn²⁺ effect may depend on the synaptic glycine concentration. We therefore repeated the experiment of Figure 4A at α1β GlyRs using 3 mM glycine, in line with the maximum predicted synaptic glycine concentration (Legendre, 1998; Beato, 2008). Sample recordings reveal that 10 μM Zn²⁺ exerts a diminished prolonging effect on deactivation rate of currents activated by brief applications of 3 mM glycine (Figure 4C). This trend is supported by the averaged data shown in Figure 4D. Thus, if the concentration of glycine in the synaptic cleft is 3 mM or higher, our results imply that Zn²⁺ must rise to concentrations greater than 10 μM during synaptic transmission to account for the effect of 10 mM tricine on the IPSC decay rate.

ZX1 is a fast, high affinity Zn²⁺ chelator (Radford and Lippard, 2013). When applied at a concentration of 100 μM, it has been shown to efficiently remove tonic Zn²⁺ bound constitutively to NMDARs (Anderson et al., 2015). Given that GlyRs and NMDARs both have high affinity Zn²⁺ sites with similar nanomolar affinities, it is reasonable to hypothesize that ZX1 might act similarly on GlyRs. Indeed, a recent study suggested that ZX1 removes tonic Zn²⁺ bound to synaptic GlyRs (Perez-Rosello et al., 2015). We investigated the effects of 100 μM ZX1 on IPSCs mediated by α1β GlyRs both as a control for possible non-specific (e.g., pharmacological) effects of tricine and to investigate the possibility that tonic Zn²⁺ might be bound constitutively to synaptic GlyRs. Figure 5A shows examples of normalized, averaged IPSCs recorded from one cell before, during and after exposure to 100 μM ZX1. Results averaged from nine cells recorded under identical conditions reveal that 100 μM ZX1 had no significant effect on IPSC amplitude (Figure 5B) or 10–90% rise time (Figure 5C) although it significantly reduced the IPSC decay time constant (Figure 5D). However, there was no significant difference in the magnitude of the effect of 10 mM tricine and 100 μM ZX1 on the IPSC decay time constant (Figures 5E–G). Moreover, in five cells where we quantitated the effects of 10 mM tricine and 100 μM ZX1 sequentially in the same cell, we found the ZX1 exerted no significant additional effect on the IPSC decay time constant (not shown). Thus, if tonic Zn²⁺ is bound constitutively to synaptic α1β GlyRs in our artificial synapses, then its removal has no functional consequence. We thus infer that the effect of tricine and ZX1 on glycinergic IPSCs is mediated by freely diffusing Zn²⁺.

In a final experiment, we quantitated the recovery time course of IPSC decay time constant following ZX1 washout (Figure 5H). In this experiment we digitally averaged all IPSCs that occurred within each 30 s interval and fitted a single decay time constant to the averaged waveform. Results were averaged from five cells. As seen in Figure 5H, complete recovery from ZX1 application required around 1 min, although the decay time constant eventually stabilized at a value significantly higher than the initial control IPSC decay time constant (Figures 5D,H). The recovery from ZX1 inhibition suggests that the concentration of free synaptic Zn²⁺ is not depleted over time under our experimental conditions.