Characterization of zebrafish GABAA receptor subunits

γ-Aminobutyric acid (GABA), the major inhibitory neurotransmitter in the central nervous system, exerts its effect through the activation of GABA receptors. GABAA receptors are ligand-gated chloride channels composed of five subunit proteins. Mammals have 19 different GABAA receptor subunits (α1–6, β1–3, γ1–3, δ, ε, π, θ, and ρ1–3), the physiological properties of which have been assayed by electrophysiology. However, the evolutionary conservation of the physiological characteristics of diverged GABAA receptor subunits remains unclear. Zebrafish have 23 subunits (α1, α2a, α2b, α3–5, α6a, α6b, β1–4, γ1–3, δ, π, ζ, ρ1, ρ2a, ρ2b, ρ3a, and ρ3b), but the electrophysiological properties of these subunits have not been explored. In this study, we cloned the coding sequences for zebrafish GABAA receptor subunits and investigated their expression patterns in larval zebrafish by whole-mount in situ hybridization. We also performed electrophysiological recordings of GABA-evoked currents from Xenopus oocytes injected with one or multiple zebrafish GABAA receptor subunit cRNAs and calculated the half-maximal effective concentrations (EC50s) for each. Our results revealed the spatial expressions and electrophysiological GABA sensitivities of zebrafish GABAA receptors, suggesting that the properties of GABAA receptor subunits are conserved among vertebrates.

www.nature.com/scientificreports/ α2 subunit perturbed the expression of the proneural gene neurod and a GABA-synthesizing enzyme gene gad1b within 1 day of development 16 . Zebrafish larvae lacking the β3 subunit showed reduced sensitivity to anesthetic drugs such as etomidate and propofol 17 . Patch-clamp recordings of GABA A receptor-mediated miniature inhibitory postsynaptic currents from zebrafish Mauthner cells revealed three different types of gating kinetics, suggesting that zebrafish also have multiple GABA A receptor subtypes comprising different subunit combinations 18 . However, the electrophysiological characteristics of the zebrafish GABA A receptor subunit have not yet been explored.
In this study, we performed phylogenetic analysis and cloned cDNAs for zebrafish GABA A receptor subunits. Our whole-mount in situ hybridization revealed the spatial expression patterns of GABA A receptor subunit genes in 5 dpf larvae. We also assayed GABA-mediated gating of zebrafish GABA A receptors composed of various combinations of receptor subunits in Xenopus oocytes. These attempts provide useful information on the spatial expressions and electrophysiological GABA sensitivities of zebrafish GABA A receptors and suggest that the properties of GABA A receptor subunits are conserved among vertebrates.

Results
Phylogenetic analysis and cloning of zebrafish GABA A receptor subunits. Nineteen GABA A receptor subunits/genes have been identified in mammals (α1-6/gabra1-6, β1-3/gabrb1-3, γ1-3/gabrg1-3, δ/gabrd, ε/gabre, π/gabrp, θ/gabrq, and ρ1-3/gabrr1-3) 19 . Previous searches for GABA A receptor subunits in the zebrafish genome database have suggested 23 GABA A receptor subunits/genes comprising eight α (α1/gabra1, α2a/gabra2a, α2b/gabra2b, α3-5/gabra3-5, α6a/gabra6a, and α6b/gabra6b), four β (β1-4/gabrb1-4), three γ (γ1-3/gabrg1-3), one δ/gabrd, one π /gabrp, and five ρ (ρ1/gabrr1, ρ2a/gabrr2a, ρ2b/gabrr2b, ρ3a/gabrr3a, ρ3b/gabrr3b) as well as additional ζ/gabrz subunits, but neither ε nor θ subunits 12,13 . Some subunits that have a or b at the end of the subunit/gene name are paralogs generated by a suspected duplication of the whole genome during fish evolution 20 . We recapitulated in silico analysis using human and mouse GABA A receptor protein sequences as queries to obtain zebrafish GABA A receptor protein sequences. We successfully cloned cDNAs for all zebrafish GABA A receptor subunits except for α2b from an RNA mixture extracted from a pool of 1-5 dpf zebrafish embryos/larvae. The previously annotated zebrafish α2b subunit (XP_017214538.1) showed 86% amino acid identity to the zebrafish α2a subunit in the N-terminus (exons 1-8) and only 10% identity in the C-terminus (exon 9). Therefore, the α2b information has been removed from the National Center for Biotechnology Information (NCBI) annotation as it was presumably caused by an incorrect annotation of the last exon. We then searched for another exon encoding the C-terminus of α2b in the genome database using the C-terminus protein sequence of zebrafish α2a as a query and identified the other last exon encoding a possible α2b C-terminus that showed 76% amino acid identity to zebrafish α2a. We successfully cloned the intact coding sequence of this newly annotated subunit and named zebrafish GABA A receptor α2b subunit (LC596832), which differed from the previous annotation only in the last exon. We then updated the phylogenetic tree of human, mouse, and zebrafish GABA A receptor subunits (Fig. 1). Our amino acid alignments of the GABA A receptor subunits showed that each subunit is conserved among vertebrates, especially in four transmembrane domains (Supplementary Figs. 1-17; Supplementary Table 1). We also confirmed that the ζ subunit, which is found in zebrafish but not in mammals, belongs to the π subfamily, with the highest similarity to the zebrafish π subunit, indicating that the ζ subunit is a paralog of the π subunit. Thus, we suggest renaming the π and ζ subunits to πa and πb subunits, respectively. We hereafter refer to π and ζ as π/πa and ζ/πb, respectively. Spatial expression of zebrafish GABA A receptor genes. A previous RT-PCR analysis described the expression of some, but not all, GABA A receptor genes in the brain and eye of adult zebrafish 12 . Another wholemount in situ hybridization study reported the spatial expression patterns of eight subunit genes in zebrafish embryos at 1, 2, and 4 dpf 13 . Since zebrafish larvae with a defect in the α1 gene showed seizure-like motor activity as early as 4 dpf 15 and GABA A receptor antagonist-induced zebrafish seizure can be assayed at 7 dpf 21 , we investigated the spatial expression of GABA A receptor subunit genes by whole-mount in situ hybridization in zebrafish larvae at 5 dpf, when the deficiency of GABA A receptor is likely correlated with seizure. The α1 gene was predominantly expressed in the forebrain, midbrain, hindbrain, and eye, while the other subunit genes were expressed by different but restricted patterns in the olfactory bulb, forebrain, midbrain, and eye at low levels ( Fig. 2a-g,w,x). Probes for β subunits showed broad labeling in the whole brain ( Fig. 2h-k). Expression of all three γ subunit genes was observed in broad brain regions including the olfactory bulb, forebrain, midbrain, hindbrain, and eye ( Fig. 2l-n). The δ and ζ/πb genes were also expressed in the broad brain regions, while the π/πa gene was expressed in the restricted pattern in the midbrain and eye (Fig. 2o-q). Among the five ρ subunits, the ρ2a gene was predominantly expressed in the olfactory bulb, forebrain, midbrain, hindbrain, and eye (Fig. 2s). Expression of the other ρ subunit genes was also observed at low levels in the broad brain regions (Fig. 2r-v). We have summarized the spatial expressions with staining intensities indicated by +++, ++, or + in Table 1. These different but overlapping expressions of GABA A receptor subunit genes suggest the formation of various GABA A receptor subtypes comprising different subunit combinations.
GABA concentration-response of zebrafish GABA A receptor subtypes. To assess the electrophysiological properties of zebrafish GABA A receptor subunits, we employed two-electrode voltage-clamp recordings and recorded GABA-evoked currents from Xenopus oocytes expressing single or multiple GABA A receptor subunits. We first recorded GABA currents from oocytes injected with one type of subunit cRNAs. The expression of single α, γ, δ, π/πa, or ζ/πb subunits did not generate GABA-evoked currents at any GABA concentration, while that of either single β subunit yielded small currents (β1: 27.5 ± 10.7 nA, n = 4; β2: 29.4 ± 6.3 nA, n = 5; β3: 81.7 ± 7.5 nA, n = 5; β4: 28.3 ± 3.6 nA, n = 5) only in the presence of GABA at 10 mM, which is a Bootstrap : 100 Tree scale : 0.1 Figure 1. A phylogenetic tree of GABA A receptor subunits. Amino acid sequences of GABA A receptor subunits from humans, mice, and zebrafish were used to create a phylogenetic tree. This phylogenetic tree detected the subfamilies of α, γ/ ε, β/ θ, δ/π, and ρ. H: human; M: mouse; Z: zebrafish. www.nature.com/scientificreports/   www.nature.com/scientificreports/ either β1 or β2 with any α subunit in the absence or presence of the γ2 subunit failed to elicit currents following exposure to GABA. We also tested all γ, δ, π/πa, and ζ/πb subunits to determine whether their co-expression changed the EC50 of α1β3 GABA A receptors, implying the incorporation of these subunits into the functional heteropentameric channels. Co-expression of the α1 and β3 subunits with the γ1, γ2, or γ3 subunit generated GABA-evoked currents with higher EC50 values compared to those in α1β3 GABA A receptors, while those with the π/πa subunit yielded currents with lower EC50 values (Fig. 3f). Interestingly, co-expression of α1 and β3 with either the δ or ζ/πb subunit eliminated GABA-dependent currents. These results showed that the electrophysiological sensitivity of zebrafish GABA A receptors to GABA differed according to the subunit combination, providing the functional diversity of GABA A receptor subtypes in zebrafish, as observed in mammals.

Discussion
In this study, we investigated the phylogeny, expression, and electrophysiology of zebrafish GABA A receptor subunits using in silico analysis, in situ hybridization, and in vitro current recording, respectively. These analyses revealed differences in the spatial expression and electrophysiological properties of GABA A receptors in zebrafish and suggested the conservation of receptor characteristics with minor differences in vertebrates.

Conservation of GABA receptor genes in vertebrates.
Previous database searches have suggested the presence of 23 GABA A receptor subunits/genes in zebrafish 12 . However, one of the annotated α2b/gabra2b exons was removed from the database as a result of standard genome annotation processing in NCBI (https ://www. ncbi.nlm.nih.gov/prote in/10406 62547 ). In this study, we identified a new exon and corrected the α2b/gabra2b annotation. Our cloning of 23 cDNAs of zebrafish GABA A receptor subunits confirmed that all of the exonintron junctions were correct for the 22 previously suggested and 1 newly identified GABA A receptor subunit. We noticed that the β4 subunit is found in zebrafish, amphibians, reptiles, and birds but not in mammals. Interestingly, the spatial expression pattern and electrophysiological properties of the β4 subunit were similar to those of the β3 subunit in zebrafish. Thus, β4 may serve as a reserve of β3 to form functionally indistinguishable GABA A receptor subtypes. Our phylogenetic analyses also suggested that the zebrafish-specific ζ subunit is a paralog of the π subunit, presumably generated by gene duplication in teleosts 20 . Thus, our nomenclature of changing the π and ζ to πa and πb, respectively, is reasonable. We also noted that neither ε nor θ subunit is found in zebrafish, while the ε subunit is found only in mammals and birds and the θ subunit is found in mammals, birds, and reptiles. A recent study proposed that a subfamily of ρ subunits is phylogenetically close to a subfamily comprising α, γ, and ε subunits 13 . However, our phylogenetic tree suggested that the ρ subfamily is instead close to a subfamily comprising the β, θ, δ, and π subunits, consistent with an old phylogenetic study 24 . This discrepancy could be caused by a difference in phylogenetic methods and, thus, will be solved in future development of phylogenetic methods. www.nature.com/scientificreports/ Spatial distributions of GABA A receptor genes. A previous in situ hybridization study reported the expression patterns of GABA A receptor subunit genes in zebrafish at 1, 2, and 4 dpf 13 . Here, we assayed the spatial expressions of α and the other subunit genes in zebrafish at 5 dpf. The nucleotide sequence identity of the GABA A receptor coding regions was 41.6-75.0%, with the highest between ρ3a and ρ3b paralogs and the second highest (72.1%) between ρ2a and ρ2b paralogs ( Supplementary Fig. 18). Although the expression patterns of ρ3a and ρ3b were almost identical, those of ρ2a and ρ2b were different at least in the olfactory bulb. Thus, we assume that each antisense probe presumably recognizes its specific target. The results of in situ hybridization showed that  www.nature.com/scientificreports/ α1 was predominantly expressed in the larval brain at 5 dpf. Depletion of the α1 subunit in zebrafish affected spontaneous behavior at 5 dpf 14,15 . Thus, the α1 appears to be the major isoform in zebrafish, similar to that in mammals 22 . We also confirmed the overlapping expression patterns of the β3 and the γ2 with the α1 subunit in the larval forebrain, midbrain, and hindbrain, suggesting that the α1β3γ2 GABA A receptor, which is the prevalent subtype in mammals, may be formed in zebrafish. However, our RNA labeling does not provide insights at the cellular level and thus, we cannot determine the actual co-expression.

Electrophysiological properties of zebrafish GABA A receptor subunits. Previous electrophysi-
ological recordings of GABA-evoked currents from Xenopus oocytes expressing the human β3 subunit with the human α (α1-6) subunit enabled EC50 comparisons of different GABA A receptor subtypes 23 . The order of the EC50 values for the human GABA A receptor was α4β3 < α6β3 < α5β3<α1β3, α2β3, α3β3. Co-expression of the human γ2 subunit increased the EC50 values and mostly maintained the order: α6β3γ2, α5β3γ2 < α4β3γ2, α1β3γ2, α2β3γ2, α3β3γ2. This finding is consistent with the fact that α1/2/3 subunit-containing phasic GABA A receptors localize at synaptic sites where the GABA concentration increases to more than 1 mM during inhibitory transmission, while α4/5/6 subunit-containing tonic GABA A receptors function at extrasynaptic sites where the GABA concentration is low (~ 0.5 µM) 2,7 . Similar oocyte recordings using zebrafish GABA A receptor subunits in this study revealed an order of EC50 values of α6bβ3 < α5β3, α1β3, α2aβ3 < α3β3 in the absence of the γ2 subunit and α6bβ3γ2 < α5β3γ2, α2aβ3γ2, α1β3γ2 < α3β3γ2 in the presence of the zebrafish γ2 subunit. Thus, the electrophysiological properties of each subunit appeared to be conserved among vertebrates, suggesting that α1/2/3-and α5/6b-containing GABA A receptors are synaptic and extrasynaptic, respectively, in zebrafish. Co-expression of the γ2 subunit increased the EC50 values in both human and zebrafish, further supporting the notion that the electrophysiological properties of GABA A receptor subunits are conserved in vertebrates. Previous studies of oocyte electrophysiology have shown that β3 homopentameric GABA A receptors produce small currents when exposed to 10 mM GABA and much larger currents when exposed to 100 µM etomidate 17,25 . We recapitulated that zebrafish β3 homopentameric GABA A receptors elicited small currents upon exposure to 10 mM GABA and large currents upon exposure to 100 µM etomidate, with the latter showing an EC50 of 30.3 ± 5.7 µM (data not shown). Zebrafish homopentameric GABA A receptors comprising the β1, β2 or β4 also produced small currents following exposure to 10 mM GABA, suggesting that either these β subunits can be expressed in Xenopus oocytes. However, heteropentameric zebrafish GABA A receptors containing β3 or β4 yielded GABA-evoked currents, whereas those containing β1 or β2 did not. We were unable to determine why zebrafish β1 and β2 subunits failed to function in heteropentameric GABA A receptors. A triple-knockout study of the GABA A receptor β subunit in mice suggested that β3 plays an indispensable role in inhibitory synaptic transmission in mammals 26 . CRISPR-mediated disruption of the β3 gene in zebrafish increased spontaneous larval movements, also implying an essential physiological function of the β3 subunit in zebrafish 17 . Thus, β3 is presumably the primary β isoform not only in mammals but also in zebrafish.
The mammalian ρ subunit can form homopentameric GABA receptors, which were initially referred to as GABA C receptors 27 and eventually recategorized as GABA A receptors 19 . Our electrophysiology also showed that zebrafish ρ2a subunit forms functional homopentameric GABA A receptors. The expression of the ρ2a subunit was observed in zebrafish eyes, similar to the expression of the ρ2 subunit in mice 28 .
Although heteropentameric zebrafish α1β3 GABA A receptors elicited GABA-evoked currents in Xenopus oocytes, the additional expression of either the δ or ζ/πb subunit eliminated the currents inconsistent with the findings in mammalian GABA A receptor cases 23 . The zebrafish δ and ζ/πb subunits may suppress the formation of α1β3 heteropentameric channels. We also failed to record GABA-mediated currents from oocytes injected with α4, α6a, β1, β2, ρ1, ρ2b, ρ3a, or ρ3b cRNA inconsistent with the previous electrophysiology using mammalian orthologs of these subunits. The efficiency of zebrafish protein synthesis may differ among receptor subunits in Xenopus oocytes.
Taken together, our current study provides basic information on the expression and gating properties of zebrafish GABA A receptors. Since recent development in CRISPR/Cas9 technology have enabled easy and multiple targeted gene disruption, future studies of GABA A receptor knockout in zebrafish will clarify the physiologically relevant function of each GABA A receptor subunit and unveil the significance of GABA A receptor diversity. Cloning of GABA A receptor subunits. Total RNA was extracted from mixtures of 1, 2, and 3 dpf zebrafish embryos and 4 and 5 dpf larvae using Sepasol RNA II Super (Nacalai Tesque) as described previously 30 . Oligo(dT)18 Primer (Thermo Fisher Scientific), SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific), and Phusion DNA polymerase (New England Biolabs) were used for RT-PCR as described previously 31 .

Materials and methods
The primer sequences are listed in Supplementary Whole-mount in situ hybridization. In situ hybridization of whole-mount zebrafish embryos with a digoxigenin-labeled antisense RNA probe was performed as described previously 33  In vitro synthesis of capped cRNAs. Capped zebrafish GABA receptor mRNAs were synthesized from pCS2 + based plasmids using the mMessage mMachine SP6 Transcription Kit (Thermo Fisher Scientific) as described previously 34 .
Electrophysiology. Electrophysiology was performed as described previously 35 . In brief, oocytes were injected with five femtomoles of cRNAs using a Nanoject II (Drummond Scientific) and incubated in Barth's solution (88 mM NaCl, 1 KCl, 2.4 mM NaHCO 3 , 0.33 mM Ca(NO 3 ) 2 , 0.41 mM CaCl 2 , 0.82 mM MgSO 4 , and 10 mM HEPES at pH 7.5 with NaOH supplemented with gentamicin at 50 µL/mL and penicillin/streptomycin at 100 units/mL) at 17 °C for 24-72 h before recording. Oocyte recording solution (90 mM NaCl, 1 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , and 10 mM HEPES at pH 7.5 with NaOH) and GABA solutions of different concentrations were applied to oocytes using a BPS-8 solution switcher (ALA Scientific). The borosilicate electrodes had resistances of ~ 0.5 MΩ when filled with 3 M KCl. Two-electrode voltage-clamp recordings were made from oocytes held at -50 mV using pClamp 10.2 to control GeneClamp 500B amplifier via Digidata 1440A digitizer (Molecular Devices). Signals were low-pass filtered at 10 Hz and sampled at 100 Hz. The recordings were analyzed using Clampfit 10.7 (Axon Instruments) and SigmaPlot 11.0 (Systat Software). The sample numbers are indicated in the figures. The EC50s and Hill coefficients were calculated using the sigmoid standard curve as below. x: GABA concentration (EC50). y: normalized current.
Statistics. Quantitative data are presented as means ± SEM. All error bars in the graphs represent the SEM values. Statistical significance was determined by pairwise analysis of variance.

Ethics statement. This study was approved by the Animal Care and Use Committee of Aoyama Gakuin
University (A9/2020) and carried out according to the Aoyama Gakuin University Animal Care and Use Guidelines and the Animal Research of in vivo Experiments (ARRIVE) guidelines.