All chemicals including N-oleoyl glycine (NOGly) were obtained from Sigma (Sydney, Australia) unless otherwise stated. 2-[8-(2-octylphenyl)octanoylamino] acetic acid (8–8 OPGly) was synthesized as previously described (Gallagher et al., 2023). For all experiments zinc refers to the use of zinc chloride (ZnCl₂). Zinc and metal chelators; tricine and Ca-EDTA were stored as stock solutions dissolved in frog Ringer’s solution (ND96: 96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM hemi-Na + -HEPES (Genscript, NJ, U.S.A.), pH 7.5) at 4°C. Stock solutions of both chelators were pH adjusted to 7.5, such that addition of tricine and Ca-EDTA would not alter the pH of recording solutions (Supplementary Table S2). Lipid modulators were stored at −20°C in DMSO and thawed once, on the day of use.

pGEMHE-h plasmids containing the cDNA of GlyR subunits α₁b, and α₃L (herein referred to as α₁, and α₃) were provided by Mary Collins (Faculty of Pharmacy, University of Sydney). Mutations were inserted into wild-type (WT) cDNAs through site directed mutagenesis, performed using oligonucleotide primers obtained from Sigma Aldrich (Sigma, Australia), the Q5 site-directed mutagenesis kit (New England Biolabs, Victoria, Australia) and MJ mini personal thermal cycler (Bio-rad, Hercules, CA, United States). Plasmids and PCR products were transformed into DH5α component Escherichia coli (E. coli) cells and DNA was extracted and purified using the Pure Link Quick Plasmid Purification kit (Life Technologies, Löhne, Germany). Sequences were confirmed by the Australian Genome Research Facility (Sydney, Australia).

For use in electrophysiology experiments, WT and mutant DNA were linearized using the NheI restriction enzyme (New England Biolabs, MA, USA) and RNA was transcribed with the T7 RNA polymerase mMessage mMachine Kit (Ambion, TX, USA). DNA and RNA concentrations were determined using a microvolume spectrophotometer (Thermo Scientific NanoDrop, Sydney, NSW, Australia) and stored at −20°C.

Xenopus laevis (RRID:NCBITaxon_8355) were obtained from NASCO (Fort Atkinson, WI, U.S.A.). All surgical procedures were conducted as described previously (Vandenberg et al., 1998), in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Use and approved by the University of Sydney Animal Ethics Committee. Briefly, female Xenopus laevis frogs were anesthetised using 0.17% 3-aminbenzoic acid ethyl ester (Tricaine) for 12 min. Small incisions (1–1.5 cm) were made in the skin and muscle layers, in the lateral portion of the lower abdomen and ovarian sacs were extracted using forceps. Ovarian sacs were disrupted through mechanical agitation with surgical scissors and individual oocytes were detached from follicular cells by enzymatic digestion with 3 mg/mL collagenase A (Boehringer, Mannheim, Germany) at 18°C for 45–90 min. Oocytes were then rinsed thoroughly with OR-2 solution (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl•6H2O, 5 mM hemi-Na + -HEPES, pH 7.5) and subsequently rinsed in storage solution which consisted of ND96 supplemented with 50 μg/mL gentamycin, 2.5 mM sodium pyruvate, 100 μM/mL tetracyclin and 0.5 mM theophylline.

Defolliculated stage IV-V oocytes were microinjected with 4.6 ng of GlyR RNA in a 23 nL volume (Drummond Nanoject, Drummond Scientific Co., Broomall, PA, U.S.A.) and stored on a platform shaker (Ratek, Victoria, Australia) at 18°C in ND96 storage solution. Oocytes were rinsed daily with ND96 storage solution. GlyR expression levels were sufficient to measure receptor activity 2–6 days following RNA injection.

During recording, oocytes were continually perfused with ND96 solution within a 500 μL recording chamber. All recording was conducted within ND96 solution unless otherwise stated. Microelectrodes were fabricated from 1 mm borosilicate capillary tubes (Harvard Apparatus, Holliston, MA, United States) using a single-stage glass micropipette puller (Narasihge Japan, Tokyo, Japan) and filled with 3 M KCl. Whole cell currents were measured at −60 mV using a Geneclamp 500B amplifier (Axon Instruments, Foster City, CA, United States) interfaced with a Powerlab 2/26 chart recorder (ADI Instruments, Sydney, Australia) and analysed using LabChart software (ADI Instruments). All data was analysed using Graph Pad Prism 9.0 (GraphPad Software, San Diego, CA, United States).

Buffers were diluted 1:10 with 2% HNO₃ and analyzed on an Agilent 7,700 s ICP-MS (Mulgrave, Vic, Australia). Samples, standards and a 50 ng mL⁻¹ Rhodium internal standard were carried via a three-channel peristatic pump and introduced into the ICP-MS via a MicroMist nebulizer and Scott type double pass spray chamber, cooled to 2°C. External calibration was performed using a 5-point (0.6–60 ppb) calibration curve. Agilent Technologies ICP-MS Chemstation software was used for all instrument control and data analysis. All acids and standards used were trace-metal grade and purchased from Choice Analytical (Thornleigh, NSW, Australia).

For all recording methods, example traces can be found in the figures associated with the results. For all statistical tests throughout the paper, a p-value of <0.05 was taken to be statistically significant. Statistical significance is represented as ns or not shown for non-significant results, and * = p ≤ 0.05, ** = p ≤ 0.01, *** = p ≤ 0.001 and **** = p ≤ 0.0001 for significant result.

It has been reported that zinc is a common contaminant of biological solutions and laboratory plasticware may be a source of contamination. We tested the stock solutions of glycine, tricine and Ca-EDTA as well as the oocyte storage solution. In addition we sampled the ND96 recording solution after it had passed through plastic tubing and entered the oocyte recording chamber. Zinc concentrations in stock solutions, as determined by ICP-MS, were as follows: Tricine = 336 nM; Ca-EDTA = 581 nM; Glycine = 433 nM. Oocyte storage solution = 4.58 μM; ND96 = 489 nM; These values are consistent with previously reported values for similar solutions (Kay, 2004; Cornelison and Mihic, 2014), and provide an indication of what may be expected in the recording solutions. While it is clear that zinc is present at low concentrations in the solutions used in this study, it is difficult to estimate the precise concentration that will be present under our recording conditions because of the likely low rate of washout of zinc when tightly bound to the receptor. Thus, in the following experiments we assume that zinc is present but have not made further adjustments to zinc concentration estimates.

Mutant GlyRs that are insensitive to zinc would make useful tools for investigating the actions of PAMs. Previous studies have explored mutations within the potentiating binding site and identified a H215A mutation within GlyRα₁ that reduces zinc potentiation, and similar mutations of aspartate (D194A) and glutamate residues (E192A) also abolish zinc potentiation (Miller et al., 2005). As E192 is a conserved residue across all GlyR subunits, we explored the E192A mutation in GlyRα₁ and GlyRα₃. The E192A mutation in GlyRα₁ produced functional receptors, however its sensitivity to zinc was not significantly different to WT GlyRα₁ at any zinc concentration tested (Supplementary Figure S1). For GlyRα₃, the E192A mutant had significantly reduced zinc potentiation for concentrations up to 30 μM but significantly increased zinc potentiation at 100 μM (Supplementary Figure S2). Thus, the E192A mutation does not abolish zinc potentiation in either receptor, making this mutation unsuitable for further study.

The hyperekplexic W170S mutation has also been shown to abolish zinc potentiation of GlyRα₁ expressed in HEK239T cells (Zhou et al., 2013) and oocytes (Cornelison et al., 2017), and has been used in functional studies to test for interactions between zinc and PAMs of GlyR (Cornelison et al., 2017). The W170 residue is conserved across all GlyR subunits (Figure 1) however, its ability to disrupt zinc potentiation has only been assessed in GlyRα₁. We therefore explored the W170S mutation in the α₁ and α₃ subunits. Unlike GlyRα₁ W170S which produced robust concentration-dependent currents in response to glycine, GlyRα₃ W170S did not produce any current in response to glycine up to 1 mM. This suggests the mutation greatly reduces the expression of the receptor or renders it non-functional. In GlyRα₁, the W170S mutation did not significantly impact maximal currents but it significantly reduced glycine affinity, increasing the EC₅₀ concentration from 17 μM to 48 μM (Supplementary Figure S3).

The zinc sensitivity of GlyRα₁ W170S was compared to WT GlyRα₁ by measuring zinc concentration-responses in the presence of a low glycine concentration (EC₅), with examples traces for WT and W170S shown in Figure 2. The W170S mutation in GlyRα₁ abolished zinc potentiation and incurred significantly greater inhibitory activity, causing the zinc-induced modulation to be significantly different to GlyRα₁ WT at every concentration tested (Figure 2). This is consistent with previous literature (Cornelison et al., 2017) and permits this receptor to be utilized as a ‘zinc-insensitive’ GlyR with respect to contaminating zinc concentrations, which are likely to be in the nanomolar range.

Metal chelators are commonly added to buffers to remove metal ions such as zinc. A common chelator used for functional GlyR studies is tricine (Kirson et al., 2013; McCracken et al., 2013b; Cornelison et al., 2017). To determine the impact of tricine on GlyRα₁, increasing concentrations of tricine were applied in the presence of a low glycine concentration (EC₅) (Figure 3).

Tricine was found to inhibit glycine induced currents by up to 84.6 ± 2.8% with an IC₅₀ of 1.0 ± 0.2 mM (Figure 3). To determine if this inhibition is due to zinc chelation or non-zinc mediated effects, the same experiment was also conducted on GlyRα₁ W170S. Tricine was found to inhibit glycine induced currents on GlyRα₁ W170S by up to 96.7 ± 10.4% with an IC₅₀ of 16.6 ± 5.5 mM (Figure 3), indicating that there is a non-zinc mediated component to tricine inhibition at GlyRs.

Another chelator used to sequester divalent cations is EDTA. However, it is a non-specific chelator and can sequester other divalent cations in solution which may induce artifacts. Ca²⁺ in particular is present in recording solutions such as ND96 and can impact the stability of oocyte membranes (Wozniak et al., 2017). To minimize the sequestering of Ca²⁺, a pre-loaded form of EDTA, Ca-EDTA, can be used. Ca-EDTA inhibited glycine induced currents on WT GlyRα₁ by 52.9 ± 4.6% with an IC₅₀ of 1.6 ± 0.6 μM (Figure 3), however the inhibitory effects of Ca-EDTA were abolished in GlyRα₁ W170S (max inhibition of 7.3 ± 2.1%) (Figure 3). This suggests that the inhibitory effects of Ca-EDTA on WT GlyRα₁ are due to chelation of zinc ions and will be a more useful chelator than tricine when looking at other modulators of GlyRs.

Zinc has previously been shown to impact the activity of PAMs and may synergistically enhance their modulatory effects (McCracken et al., 2013a,b). The activity of PAMs which have been assessed in experimental set-ups that have contaminating zinc could therefore be inadvertently over-estimated. Our lab has recently published studies on a series of acyl- and phenylene-containing bioactive lipids which act as PAMs at GlyRs (Gallagher et al., 2020, 2023). To determine if the activity of these PAMs were impacted by contaminating zinc, the activity of the most efficacious modulator 8–8 OPGly and NOGly were compared between WT GlyRα₁ and the W170S mutant. The degree of potentiation induced by 8–8 OPGly or NOGly at GlyRα₁ W170S were not significantly different from the potentiation induced on the WT receptor (Figure 4), which suggests that the W170S mutation does not influence the actions of 8–8 OPGly or NOGly.

To further examine the impact of zinc and zinc-chelators on lipid-mediated potentiation of GlyRs, the potentiating effects of 8–8 OPGly were assessed in the absence and presence of metal chelators on the same cell expressing the W170S mutant (Figure 5). 1 mM tricine significantly reduced the raw currents induced by the EC₅ concentration of glycine by over 50% however, when 8–8 OPGly was co-applied the degree of potentiation was very similar for both receptors. In comparison, 10 μM Ca-EDTA, which causes ~45% inhibition of WT GlyRα₁, had no effect on the W170S mutant and the level of 8–8 OPGly stimulation also remained the same. Thus, we can conclude that the stimulatory effects of 8–8 OPGly are independent of the actions of any contaminating zinc.

Zinc synergistically enhances the positive allosteric effects of ethanol and anesthetics on GlyRs (Cornelison et al., 2017) and while the basal, or contaminating, concentration of zinc in our recording system does not impact 8–8 OPGly activity, we wished to investigate whether higher concentrations of zinc will influence 8–8 OPGly activity. 1 μM zinc was used to test for possible synergistic interactions as synaptic zinc concentrations have been shown to rise to at least 1 μM following synaptic simulation, when conducted in an artificial synapse (Zhang et al., 2016). Both modulators (zinc, 8–8 OPGly) were applied independently, and the sum of their individual activities were used to calculate an estimate of the modulatory activity that would be expected if there were no synergism between the two. Both modulators were then co-applied, and the degree of potentiation achieved by their co-application was compared to the sum of the individual applications. 1 μM Zinc significantly enhances the potentiation induced by 1 μM 8–8 OPGly at GlyRα₁ when analysed using a paired t-test (Figure 6). The degree potentiation elicited by the co-applications of zinc and 8–8 OPGly was 90% greater than the sum of the individual components. This ranged from 24 - 246% for the 6 cells.

Zinc chelators have previously been used to remove contaminating zinc ions from recording solutions. One of the most commonly used chelators is tricine (Kirson et al., 2013; McCracken et al., 2013a; Cornelison et al., 2017), which is a derived from tris and glycine and was originally synthesized as one of Good’s buffers (Good et al., 1966). In this study we have demonstrated that tricine significantly reduces the glycine-induced currents of GlyRα₁ by over 80%. This could be due the chelation of contaminating zinc ions present within the ND96 solution which would produce inhibition by removing zinc-mediated potentiation of the receptor. To explore this, the effect of tricine was also assessed using the W170S GlyRα₁, which abolishes zinc potentiation. The mutant receptor was similarly inhibited by tricine which demonstrates that tricine reduces receptor activity via a non-zinc mediated mechanism.

As tricine is an analog of glycine, its structural similarity may allow it to access and obstruct the agonist binding cavity. The tris-moiety is significantly larger than the amine group of glycine and if bound within the orthosteric binding cavity, would prevent the closure of loop-C which is required for receptor activation (Kumar et al., 2020; Yu et al., 2021). Tricine could therefore inhibit receptor activity by acting as a competitive antagonist. Tricine has also been shown to form complexes with glycine (Zayed and Ammar, 2014) which would reduce the concentration of free glycine in recording solutions and would reduce receptor activation. Additional studies which explore these possible inhibitory mechanisms are required to understand the impact tricine has on the GlyR.

It is also possible that tricine impacts GlyR activity through non-selective chelation of divalent cations. Tricine can also bind strongly to magnesium, calcium, cobalt, copper and nickel (Ferreira et al., 2015). While cobalt, copper and nickel are not commonly used in biological buffers or significantly impact GlyR activity, the recording solution ND96 contains 1 mM magnesium and 1.8 mM calcium. Chelation of these ions would alter the ionic composition of the buffer which could significantly impact on the function of the receptor. Previous studies have used other chelating agents such as EDTA (Kay, 2004; Trombley et al., 2011) and ZX1 (Perez-Rosello et al., 2015; Zhang et al., 2016), however they too have varying levels of selectivity. As a way to address this, we also explored the activity of Ca-EDTA which is pre-loaded with calcium and would not be expected to sequester calcium ions from ND96. 10 μM Ca-EDTA was found to inhibit the WT GlyRα1 by over 50%, but when applied to W170S GlyRα₁ the inhibitory activity was abolished. This suggests the inhibitory effects of Ca-EDTA on the WT receptor are solely due to the chelation of contaminating zinc ions. The inhibition of the WT receptor also confirms the presence of contaminating zinc ions in ND96 solution which are causing a basal level of potentiation during recording.

An alternative approach to negate zinc contamination is the use of zinc-insensitive mutations. The W170S mutation within GLRA1 produces a hyperekplexic phenotype by removing zinc potentiation (Al-Futaisi et al., 2012; Zhou et al., 2013). W170 does not form part of the zinc binding site, but instead occurs on the β8-strand and sits parallel in between the β1 and β9-strands. The aromaticity of W170 likely influences the positioning of the β9-strand which binds zinc, and therefore indirectly contributes to the conformation of the zinc binding site responsible for potentiation. Within this study, the W170S mutation was found to completely abolish zinc potentiation in GlyRα₁. This is consistent with previous literature (Cornelison et al., 2017) and allows this receptor to be used as a tool to assess zinc-contamination. However, it is important to note the mutation was found to reduce the affinity of glycine compared to the WT receptor. This could suggest the mutation impacts receptor functionality or may be a result of removing basal zinc potentiation caused by contaminating zinc, similar to the addition of Ca-EDTA.

Overall, these results indicate that the W170S mutation and the chelator Ca-EDTA are suitable methods for removing the impact of zinc contamination, and that tricine is not a suitable chelator for functional studies of GlyRα₁. To further explore the utility of these methods in practice, all three methods were used to assess the impact of bioactive lipid PAMs of GlyRs. When the bioactive lipid PAMs NOGly and 8–8 OPGly were tested on W170S GlyRα₁ they induced similar levels of potentiation as WT GlyRα₁. This suggests that lipid modulation of GlyRs is independent from zinc modulation and does not require zinc to illicit potentiation. This is important for modulators which are being developed as possible pain therapeutics. 8–8 OPGly was also found to illicit similar levels of potentiation at GlyRα₁ in the absence and presence of both chelators, despite the reduction in glycinergic current from zinc-chelation and tricine-induced inhibition. This further supports that lipid modulation of GlyRs is not impacted by contaminating zinc or by the inhibitory mechanism induced by tricine. Furthermore, this suggests that the degree of lipid modulation is not influenced by nanomolar concentrations of zinc which are likely to occur in biological solutions, and has not impacted the results of previous studies (Gallagher et al., 2020, 2023).

It is not only crucial to determine the activity of PAMs in the absence of zinc, but also beneficial to test novel PAMs in the presence of physiologically relevant zinc concentrations to mimic endogenous conditions. Endogenous zinc concentrations have previously been shown to synergistically enhance the activity of some PAMs. One example of this is ethanol, which was found to act synergistically with zinc at GlyRα1 when expressed in oocytes (McCracken et al., 2010). Additionally, when ethanol and volatile an aesthetics such as isoflurane, chloroform and halothane are tested on W170S GlyRα1, their potentiating activity is reduced, which does not occur for intravenous anesthetics such as toluene or trichloroethane (Cornelison et al., 2017). Developing modulators which interact synergistically with zinc could enhance their activity in vivo. Within this study a 1 μM concentration of zinc was used to test for possible synergistic interactions as synaptic zinc concentrations have been shown to rise to at least 1 μM following a single synaptic simulation, when conducted in an artificial synapse (Zhang et al., 2016). Similar to ethanol, the co-application of zinc and 8–8 OPGly was greater than the sum of their individual effects, which suggests a synergistic interaction is occurring (McCracken et al., 2010). It is unclear how zinc binding within the ECD mechanistically enhances the activity of lipid modulators which are expected to bind within the TM domains. Zinc binding may induce conformational changes that are transduced down into the TM domains via intermediate loops, which could enhance the efficacy of the lipid binding cavity. Zinc has also been suggested to act at an intracellular binding site within the ICD (Trombley et al., 2011) which could be partially responsible for the synergistic interaction. Regardless of its mechanisms, the synergistic interaction between the two modulators may be advantageous for analgesic drug development, as it suggests the endogenous activity of bioactive lipid PAMs will be enhanced in vivo at nociceptive synapses which contain zinc (Birinyi et al., 2001).