Synaptic transmission-dependent regulation of neurotransmitter receptor accumulation at postsynaptic sites underlies the formation, maintenance and maturation of synaptic function. Previous in vitro studies showed that glycine receptor (GlyR) clustering requires synaptic inputs. However, in vivo GlyR regulation by synaptic transmission is not fully understood. Here, we established a model system using developing zebrafish, in which GlyRs are expressed in Mauthner cells (M-cells), a pair of giant, reticulospinal, hindbrain neurons, thereby enabling analysis of GlyR clusters over time in identifiable cells. Bath application of a glycinergic blocker, strychnine, to developing zebrafish prevented postsynaptic GlyR cluster formation in the M-cells. After strychnine removal, the GlyR clusters appeared in the M-cells. At a later stage, glycinergic transmission blockade impaired maintenance of GlyR clusters. We also found that pharmacological blockade of either L-type Ca2+ channels or calcium-/calmodulin-dependent protein kinase II (CaMKII) disturbed GlyR clustering. In addition, the M-cell-specific CaMKII inactivation using the Gal4-UAS system significantly impaired GlyR clustering in the M-cells. Thus, the formation and maintenance of GlyR clusters in the M-cells in the developing animals are regulated in a synaptic transmission-dependent manner, and CaMKII activation at the postsynapse is essential for GlyR clustering. This is the first demonstration of synaptic transmission-dependent modulation of synaptic GlyRs in vivo.
Clustering of neurotransmitter receptors at postsynaptic sites is crucial for synaptogenesis and establishment of proper synaptic function. After apposition of a presynaptic axon and a postsynaptic dendrite or soma, postsynaptic components such as neurotransmitter receptors and their scaffolding proteins accumulate at appropriate sites (McAllister 2007). After synapses are established, postsynaptic receptors can still move between synaptic and extrasynaptic sites and from the plasma membrane to the intracellular compartments (Cognet et al. 2006; Gerrow & Triller 2010). The regulation of neuronal postsynaptic receptors has been extensively studied in excitatory glutamatergic synapses (Craig 1998; Specht & Triller 2008). Localization of glutamate receptors is regulated by synaptic activity (Rao & Craig 1997; O'Brien et al. 1998). For example, AMPA receptors accumulate at transmission-active postsynaptic sites by diffusional trapping of receptors (Ehlers et al. 2007). Synaptic activity-dependent regulation of the number of postsynaptic receptors mediates synaptic plasticity, which occurs in long-term potentiation and depression (Malenka & Bear 2004).
Synaptic clustering of inhibitory neurotransmitter receptors such as glycine and GABAA receptors also depends on neuronal activity (Choquet & Triller 2003; Renner et al. 2008; Specht & Triller 2008). These receptors bind to the scaffold protein gephyrin and are anchored at postsynaptic sites (Kirsch et al. 1993; Feng et al. 1998; Dumoulin et al. 2000; Lévi et al. 2004; Jacob et al. 2005). Time-lapse imaging of glycine and GABAA receptors (GlyRs and GABAARs) in cultured neurons has showed that GlyRs and GABAARs continuously move on the plasma membrane by lateral diffusion and stay for a longer time at gephyrin-rich synaptic sites than at extrasynaptic regions (Meier et al. 2001; Dahan et al. 2003; Bannai et al. 2009). Excitatory activation of cultured hippocampal neurons increased GABAAR motility and reduced the synaptic GABAAR cluster size (Naylor et al. 2005; Bannai et al. 2009). Diffusional GlyR mobility is also modified by neural activity (Lévi et al. 2008), and GlyR clusters are not formed without glycinergic transmission in cultured spinal neurons (Kirsch & Betz 1998; Lévi et al. 1998). These studies were carried out using dissociated neurons in vitro. However, it remains unclear whether formation and maintenance of inhibitory receptor clusters are regulated by synaptic activity in vivo. The postsynaptic mediator of the synaptic transmission-dependent GlyR clustering also remains to be elucidated.
In this study, we investigated receptor clustering in vivo by focusing on GlyRs in zebrafish Mauthner cells (M-cells), which are identifiable reticulospinal neurons in the hindbrain. In zebrafish, a single pair of M-cells is formed at 7.5 h postfertilization (hpf), and this single pair is maintained to adulthood (Mendelson 1986). M-cells receive numerous glycinergic inputs on their soma-dendritic membrane (Faber & Korn 1978; Koyama et al. 2011; Moly & Hatta 2011). We carried out GlyR cluster immunolabeling and in vivo whole-cell recording of glycinergic miniature postsynaptic currents (mPSCs) in M-cells in the presence or absence of strychnine, a specific GlyR blocker. We found that glycinergic transmission is necessary for the formation and the maintenance of GlyR clusters. In addition, we applied GAL4-mediated gene expression in M-cells and showed that CaMKII activation at the postsynapse is essential for synaptic GlyR clustering in developing M-cells.
Glycinergic transmission blockade leads to a lasting functional deficit in the glycinergic synapse in vivo
To assess whether the lack of glycinergic transmission affects glycinergic synapses, we recorded mPSCs in vivo. Whole-cell recordings of M-cells at 3 dpf in the presence of TTX, bicuculline, DNQX and APV showed spontaneous inward currents at 6.8 ± 1.2 Hz (n = 7; Fig. 1A, C), which is similar to a previous report that found mPSCs in M-cells of 7–16 Hz at 52 hpf (Legendre & Korn 1994). These mPSCs are mediated by the glycinergic synapse, because they were eliminated by application of strychnine (0.3 ± 0.2 Hz, n = 3; Fig. 1A, C). However, when zebrafish embryos were raised in the presence of strychnine from 22 hpf to 3 dpf, glycinergic mPSCs were not observed after strychnine washout for more than 10 min (0.01 ± 0.01 Hz, n = 4; Fig. 1B, C). It is unlikely that the remaining strychnine disturbed the glycinergic mPSCs, because glycinergic mPSCs were not observed in the larvae after 6–10 h of washout before recording (0.0 ± 0.0 Hz, n = 3, Fig. S1). Thus, glycinergic transmission blockade results in a lasting functional deficit in the glycinergic synapse.
Glycinergic input is required for GlyR cluster formation during zebrafish development
It has been reported that GlyR clusters are formed at postsynaptic sites in a synaptic transmission-dependent manner in cultured spinal neurons (Kirsch & Betz 1998; Lévi et al. 1998). To examine whether the loss of glycinergic transmission in strychnine-treated larvae is attributable to defects in GlyR cluster formation, we next examined the formation of GlyR clusters in developing M-cells using anti-GlyR antibody immunolabeling. GlyR immunoreactivity was detected as dots at the M-cell surface beginning at 2 dpf (Fig. 2Aa-e). The number of GlyR clusters at the surface of a single M-cell soma increased until at least 5 dpf (Fig. 2B, Table S1). The membrane surface area of the M-cell somata, which was calculated by integration of the soma circumferences, also increased over time during development (Fig. 2C, Table S1). The GlyR cluster density (cluster number/surface area) reached a plateau by 3 dpf and was maintained thereafter to adulthood (Fig. 2Aa-i, D; Table S1). Double labeling with anti-synaptophysin, a marker for presynaptic terminals, showed that GlyR clusters were apposed to synaptophysin, indicating that GlyR clusters are formed at postsynaptic sites (Fig. 2E).
To address whether blocking glycinergic transmission affects GlyR clusters, strychnine was bath-applied to embryos from 22 hpf (before GlyR clusters had formed). This strychnine treatment significantly impaired the formation of GlyR clusters on M-cells both in number (P <0.05 at 2 dpf, P <0.01 at 3 dpf) and in density (P <0.05 at 2 dpf, P <0.001 at 3 dpf, Fig. 2Aj-l, B, D; Table S1). The M-cell surface area was unaffected by strychnine treatment (Fig. 2C, Table S1). Strychnine-mediated impairment of GlyR clustering was observed at various strychnine concentrations [control (0 μm): 0.36 ± 0.02 μm−2, n = 3; 30 μm: 0.02 ± 0.004 μm−2, n = 9; 100 μm: 0.03 ± 0.005 μm−2, n = 6; 400 μm: 0.02 ± 0.003 μm−2, n = 9; at 3 dpf; Fig. S2]. These results show that glycinergic transmission is necessary for GlyR cluster formation in vivo. The lack of normal GlyR cluster formation in the presence of strychnine might be attributable to the withdrawal of presynaptic terminals. To test this possibility, we examined the projection of glycinergic terminals onto the M-cells using glyt2:GFP transgenic zebrafish after application of strychnine from 22 hpf to 3 dpf. GFP-positive presynaptic boutons were observed on the soma-dendritic membranes in strychnine-treated larvae (Fig. S3). Thus, blocking glycinergic input disturbed postsynaptic GlyR cluster formation without withdrawal of presynaptic terminals.
To furthermore investigate the role of glycinergic transmission in GlyR clustering, we examined whether GlyR clusters in the M-cells are restored after removal of strychnine. Zebrafish were treated with strychnine from 22 hpf to 2 dpf, and the embryos were then raised in breeding water and examined for immunolabeling at the indicated stages. After strychnine washout, GlyR clusters were observed on M-cells beginning at 3 dpf and increased in density over time (Fig. 3Aa-e, B; Table S2). Similarly, when embryos were treated with strychnine until 3 dpf, the formation of GlyR clusters appeared within 2 days of strychnine washout (Fig. 3Af-i, B; Table S2). The GlyR clusters were apposed to synaptophysin, confirming that they were formed at postsynaptic sites (Fig. 3C). Therefore, GlyR clustering at synaptic sites is regulated in a synaptic transmission-dependent manner, even if the initiation of glycinergic transmission is delayed during the early developmental stages. Taken together, our labeling and electrophysiological recordings showed that glycinergic transmission is required for the in vivo formation of GlyR clusters, and thus, of functional glycinergic synapses.
Glycinergic transmission is required to maintain previously formed GlyR clusters
To assess whether glycinergic transmission is necessary for the maintenance of GlyR clusters at synapses, we labeled GlyR clusters after a 24-h application of strychnine at later stages when GlyR clusters had been already formed. We were surprised to find that the previously formed GlyR clusters disappeared from the surface of M-cells after a 24-h strychnine treatment in the early larval stages (2–3 dpf, Fig. 4Aa, B; 3–4 dpf, Fig. 4Ab, B; 4–5 dpf, Fig. 4Ac, B; Table S3). Similarly, GlyR clusters were markedly decreased after a 24-h application of strychnine in late larvae and juveniles (10–11 dpf, Fig. 4Ad, B; 15–16 dpf, Fig. 4Ae, B; 30–31 dpf, Fig. 4Af, B; Table S3). Interestingly, the extent of the strychnine-induced reduction in the GlyR cluster density decreased as the fish aged (95.6 ± 2.2% at 5 dpf; 82.7 ± 5.7% at 11 dpf; 73.4 ± 5.7% at 16 dpf; 51.3 ± 6.6% at 31 dpf; Fig. 4C). Thus, glycinergic transmission is required not only for the formation but also for the maintenance of GlyR clusters during larval and juvenile periods. The importance of glycinergic input in maintaining GlyR clusters seems to attenuate in an age-dependent manner.
To investigate whether glycinergic transmission could dynamically regulate GlyR clusters, we examined effect of short-term strychnine treatment on GlyR clusters. Zebrafish were treated with strychnine for 1 h at 3 dpf. Surprisingly, strychnine treatment for only 1 h almost completely eliminated the GlyR clusters at the M-cell surface (Fig. 5Ab, B; Table S4).
We then examined the recovery of GlyR clusters at the surface of M-cells after the 1-h strychnine treatment. After strychnine washout, the GlyR clusters started to be reformed on M-cells in 10 h and increased thereafter (Fig. 5Ab–f, B; Table S4). It took less than 1 day for M-cells to recover a significant level of GlyR clusters on their surface. The GlyR clusters recovered much faster than did those in the longer-term (22 hpf to 3 dpf) strychnine treatment experiment (Fig. 3Af-i, B; Fig. 5B). These data suggest that the GlyR clusters are dynamically and continuously regulated by GlyR activities.
CaMKII activation in the postsynaptic cell is required for GlyR clustering
To elucidate the molecular basis underlying synaptic transmission-dependent GlyR clustering in vivo, we focused on Ca2+-mediated events. It has been reported that pharmacological application of either nifedipine or KN-93, which inhibits the L-type Ca2+ channels and CaMKII, respectively, compromises the formation of GlyR clusters in cultured spinal neurons (Kirsch & Betz 1998; Charrier et al. 2010). We applied nifedipine or KN-93 to zebrafish embryos from 24 hpf to 3 dpf and immunolabeled GlyRs at 3 dpf. Application of either nifedipine or KN-93 to zebrafish embryos significantly impaired GlyR clustering similar to the strychnine treatment (Fig. 6Aa, b, d, e, B; Table S5). In contrast, application of cyclosporine A, a calcineurin inhibitor, did not affect GlyR clusters (Fig. 6Ac, B; Table S5). The impairment of GlyR clustering in the presence of nifedipine or KN-93 is not caused by the elimination of presynaptic terminals, because glycinergic presynaptic boutons, which can be labeled by GFP in glyt2:GFP transgenic zebrafish, were not morphologically affected in M-cells after 2 days of the drug application (data not shown). Glycinergic presynaptic terminals were also unchanged after treatment with cyclosporine A. Thus, the L-type Ca2+ channels and CaMKII, but not calcineurin, play important roles in GlyR clustering in vivo.
Because pharmacological application of drugs to developing zebrafish embryos affects presynaptic and postsynaptic cell targets, we cannot simply conclude that postsynaptic CaMKII regulates GlyR clustering. Indeed, CaMKII is involved in the release of neurotransmitters at the presynaptic terminals (Nichols et al. 1990). To clarify this issue, we inhibited CaMKII specifically in M-cells using the Gal4-UAS system (Asakawa et al. 2008). We established a Gal4 driver zebrafish line, hspGFF62A, in which a modified GAL4 transcription activator is expressed in the M-cells (Fig. 6C). We also generated UAS transgenic lines expressing mCherry and either autocamtide-2-related inhibitory peptide II (AIP2) or its control peptide 2 (ACP2), using the auto-cleavage peptide sequence 2A in between the two genes. AIP2 originates in the autoinhibitory domain of CaMKII and, thus, is used to inhibit CaMKII activation, whereas ACP2 is designed as a control peptide by changing some critical amino acid residues in AIP2 that inhibits CaMKII (Ishida et al. 1998; Khoo et al. 2006). We confirmed that mCherry was expressed in M-cells in transgenic larvae that carried both hspGFF62A and UAS:mCherry-2A-ACP2 and in those that carried hspGFF62A and UAS:mCherry-2A-AIP2 (Fig. S4). In ACP2-expressing M-cells, GlyR clusters were normally formed at the surface of the M-cells (Fig. 6Da, E; Table S5). In contrast, GlyR cluster formation was significantly impaired in AIP2-expressing M-cells (P <0.001; Fig. 6Db, E; Table S5). These results show that CaMKII activation in postsynaptic cells is essential for the GlyR clustering in vivo.
In this study, we present the first in vivo evidence that the formation and maintenance of GlyR clusters in zebrafish M-cells require glycinergic synaptic transmission and that L-type Ca2+ channels and the postsynaptic CaMKII are involved in the process. We found three developmental features of GlyR clustering in M-cells. First, GlyR clusters appear by 2 dpf and subsequently increase in number. This is consistent with previous physiological findings that the frequency of mPSCs in zebrafish M-cells increases from 26 to 52 hpf (Ali et al. 2000). Second, blockade of glycinergic transmission impairs the formation of GlyR clusters, but termination of the blockade enables initiation of the cluster formation. Thus, GlyRs accumulate at synaptic sites in a synaptic transmission-dependent manner, even if the initiation of glycinergic transmission is delayed during the early developmental stages. Third, glycinergic transmission is necessary for the maintenance of GlyR clusters, at least until 30 dpf, and the importance of glycinergic input for the maintenance appears to decrease in an age-dependent manner. In addition, short-term strychnine treatment (1 h) experiments have suggested that continuous glycinergic transmission is crucial for the maintenance of GlyR clusters. Thus, our findings show that GlyR clusters are dynamically regulated in a synaptic transmission-dependent manner during development.
We found that application of strychnine eliminated GlyR clusters within 1 h, which is a very fast response. What is the cellular mechanism by which postsynaptic GlyRs are rapidly eliminated in vivo by strychnine treatment? There are at least two potential possibilities to explain the disappearance of GlyR clusters after strychnine treatment. One possibility is that glycinergic transmission blockade accelerates GlyR lateral diffusion by unanchoring them from the postsynaptic structure, which is similar to the regulation of glutamate receptors (Ehlers et al. 2007). Indeed, GlyR mobility is regulated by lateral diffusion in cultured neurons (Dahan et al. 2003). Another possibility is that strychnine promotes degradation of surface GlyRs. It has been reported, however, that the half-life of plasma membrane GlyRs is approximately 14 h in cultured neurons, and that is not affected by strychnine-induced blockade of glycinergic transmission (Rasmussen et al. 2002). The reported half-life of the GlyR protein is much longer than 1 h within which the GlyR clusters were eliminated by strychnine, suggesting that degradation of cell surface GlyR is not primarily responsible for the rapid elimination of GlyR clusters. We therefore suggest that blockade of glycinergic transmission facilitates the lateral diffusion of GlyR that results in the rapid disappearance of GlyR clusters. We also found that after washout of strychnine at 3 dpf, the GlyR re-clustering occurred in less time in the 1-h strychnine treatment experiment than in the experiments with the longer (22 hpf to 3 dpf) strychnine treatment (Fig. 5B). The difference in recovery time suggests that long-term strychnine treatment affects regulation of GlyR more profoundly than does the acute blockade of glycinergic transmission. The changes caused by the long-term strychnine treatment might involve expression and/or intracellular transport of GlyR, in addition to the presumed facilitation of GlyR lateral diffusion. Indeed, it was shown in cultured neurons that strychnine treatment suppresses intracellular transport of newly synthesized gephyrin, which binds to GlyR during intracellular transport and at synaptic sites (Maas et al. 2009). Further studies are needed to clarify whether the GlyR supply is affected in the M-cells after long-term treatment with strychnine.
How does glycinergic transmission regulate GlyR clustering? Strychnine blocks glycinergic inputs to the M-cells directly. At the same time, it might induce excessive excitation of M-cells and connected neurons, because glycinergic inputs can inhibit action potentials through their shunting effects (Takahashi et al. 2002), Therefore, strychnine-induced disappearance of GlyR clusters can be caused directly by the elimination of glycinergic inputs or, alternatively, indirectly by the induction of excessive excitation. The possibilities might be assessed by treatment with TTX, which eliminates evoked transmission at both the excitatory and inhibitory synapses. A previous in vitro study showed that TTX treatment of cultured neurons disturbs GlyR cluster formation (Kirsch & Betz 1998). This result supports the notion that the strychnine-induced disappearance of GlyR clusters can occur without the elevation of excitation and thus might be elicited directly by the lack of glycinergic transmission. However, another in vitro study using TTX indicates contradictory results (Lévi et al. 1998). It is not feasible to test the effect of TTX in vivo, because bathing zebrafish in a TTX solution would kill the fish, leaving this issue undetermined. However, it is shown here that blockade of voltage-gated L-type Ca2+ channels inhibited GlyR clustering in vivo. This result is difficult to explain using the model of increased excitation in which the Ca2+ channels should be activated. In contrast, possible glycinergic transmission-induced postsynaptic depolarization appears to account for the result.
The synaptic transmission-dependent maintenance of GlyR clusters observed in larvae and juveniles decreased as the fish aged. In agreement with this observation, it has been reported that glycinergic transmission blockade does not affect GlyR clusters in adult goldfish M-cells (Seitanidou et al. 1992). Thus, the maintenance of GlyR clusters appears to be age dependent. The age dependency might be partly attributable to a reduction of intracellular Cl− concentration in developing neurons. It has been shown that glycinergic transmission causes hyperpolarization in mature neurons, but that it can induce depolarization in immature and young neurons, in which the concentration of intracellular Cl− is high (Tapia & Aguayo 1998). The intracellular Cl− concentration decreases due to the late onset of the K+-Cl−-coupled cotransporter (KCC2) expression, which exports cytosolic Cl− (Tapia & Aguayo 1998; Ben-Ari 2002; Reynolds et al. 2008; Zhang et al. 2010). Although KCC2 expression increases during zebrafish development (Reynolds et al. 2008), it remains to be elucidated whether or when the Cl− reversal potential changes in the M-cells.
Previous studies using cultured neurons have suggested that L-type Ca2+ channels and CaMKII are involved in GlyR clustering (Kirsch & Betz 1998; Charrier et al. 2010). Our in vivo results suggest that L-type Ca2+ channels and CaMKII are essential for GlyR clustering. In addition, we found that targeted expression of CaMKII inhibitor peptides in the M-cells disturbed GlyR clusters at the surface of the cells. This is the first demonstration that CaMKII activation in the postsynaptic cells is required for GlyR clustering. In adult goldfish M-cells, CaMKII is abundant at postsynaptic sites in the soma and lateral dendrites (Pereda et al. 1998; Flores et al. 2010). CaMKII association with the L-type Ca2+ channel has also been reported in cardiac muscles (Hudmon et al. 2005). In addition, glycinergic currents can be enhanced by intracellular application of activated CaMKII in cultured spinal neurons (Wang & Randić 1996). Thus, CaMKII might be a crucial mediator that links Ca2+ and GlyR clustering. Our research provides a model where glycinergic transmission causes the local depolarization of postsynaptic membranes that activates L-type Ca2+ channels and then increases Ca2+ permeability. The intracellular Ca2+ activates CaMKII, thereby facilitating the GlyR clustering at postsynaptic sites (Fig. 7).
If this is the case, how does CaMKII mediate GlyR clustering? It has been shown that gephyrin has several serine residues that can be phosphorylated by CaMKII in vitro (Charrier et al. 2010). Phosphorylation of several specific gephyrin serine residues promotes interaction with peptidyl–prolyl isomerase Pin1, a molecular chaperone, and causes a conformational change in gephyrin that enhances gephyrin binding to GlyRs (Zita et al. 2007). However, these serine residues of gephyrin are different from the actual phosphorylation targets identified by an in vitro CaMKII reaction (Charrier et al. 2010). Although the gephyrin phosphorylation site that is critical for GlyR clustering in vivo is still unknown, CaMKII or CaMKII-dependent protein kinase likely promotes gephyrin phosphorylation, which accelerates the GlyR clusters formation at the postsynaptic sites. Protein kinase C (PKC) might also be involved in the process; a recent study has showed that PKC-induced phosphorylation of the GlyRβ subunit regulates GlyRβ–gephyrin interaction (Specht et al. 2011).
Our GlyR clustering assay in zebrafish M-cells provides an excellent model for analyzing synaptic transmission-dependent receptor accumulation and its underlying mechanisms in a developing single cell. In this study, we also succeeded in inducing the expression of certain genes in M-cells using the Gal4-UAS system. Zebrafish embryos and larvae are transparent and are thus suitable for visualizing fluorescence-tagged receptor behavior in vivo, which will be useful in future studies.
Zebrafish (Danio rerio) embryos, larvae, juveniles and adults were maintained at 28.5 °C. Experiments were carried out at room temperature (25–28 °C). Wild type and five of transgenic lines were used: glyt2:GFP (McLean et al. 2007), Tol-056 (expressing GFP in the M-cells; Satou et al. 2009; Tanimoto et al. 2009), hspGFF62A (Asakawa et al. 2008), UAS:mCherry-2A-AIP2 (this study) and UAS:mCherry-2A-ACP2 (this study). From 26 hpf, the hspGFF62A fish expresses Gal4FF (Asakawa et al. 2008) in the M-cells, other neurons and cardiac muscles. For CaMKII inhibition, UAS:mCherry-2A-AIP2 was used to express Autocamtide-2-related CaMKII inhibitory peptide (AIP2, KKKLRRQEAFDAL) (Ishida et al. 1998). As a control, UAS:mCherry-2A-ACP2 was similarly used to drive expression of Autocamtide-2-related control peptide II (ACP2, KKKGRAQERFDCL) (Khoo et al. 2005). These two UAS transgenic lines were generated using a Tol2-mediated transgenesis method as described previously (Kawakami et al. 2000). All procedures were carried out in compliance with the guidelines set by Nagoya University.
Zebrafish were bathed in a solution containing strychnine hydrochloride (Sigma-Aldrich, St Louis, MO, USA) at 20 ~ 800 μm. Zebrafish were bathed in the strychnine solution for various lengths of time: 22 hpf–48 hpf; 22 hpf–72 hpf; 22 hpf–96 hpf; 22 hpf–120 hpf; 2 dpf–3 dpf; 3 dpf–4 dpf; 4 dpf–5 dpf; 10 dpf–11 dpf; 15 dpf–16 dpf; or 30 dpf–31 dpf. Similarly, nifedipine (Nacalai Tesque), KN-93 (Biomol, Plymouth Meeting, PA, USA) and cyclosporine A (Biomol) were bath-applied at 100 to 200 μm, 5 μm and 42 μm, respectively, in breeding water starting at 24 hpf.
In the experiment with cyclosporine A, we used a higher concentration than those in a previous study (Ponnudurai et al. 2012), where cyclosporine A was shown to have clear effects on zebrafish embryos. Therefore, although we have not confirmed directly whether cyclosporine A actually inhibited calcineurin, cyclosporine A is likely to be effective in our experiments.
GlyRs were immunolabeled as described previously (Hirata et al. 2005). Briefly, horizontal sections (20–30 μm) and whole-body zebrafish embryos/larvae were fixed in 4% paraformaldehyde at room temperature for 30 min and 7 h, respectively, and then washed several times in phosphate-buffered saline (PBS, pH 7.4) containing 0.1% Tween 20. The samples were incubated with PBS containing 2% BSA, 5% goat serum and 0.5% Triton X-100 to block nonspecific reactions and then with the following primary anti-bodies: mAb4a (1 : 1000, anti-GlyR; Synaptic Systems, Goettingen, Germany), anti-synaptophysin 1 (1 : 50; Synaptic Systems) or 3A10 (1 : 50, anti-neurofilament; Developmental Studies Hybridoma Bank, University of Iowa). The signals were visualized with the secondary fluorescent antibodies Alexa 488- or 555-conjugated anti-mouse IgG (1 : 1000) and anti-rabbit IgG (1 : 150). The fluorescent images were captured using a confocal microscope (FV300; Olympus, Tokyo, Japan).
To count the number of GlyR clusters on M-cell somata, 20- to 30-μm-thick horizontal sections that included the whole M-cell soma were immunolabeled with mAb4a and 3A10. Serial optical sections were acquired at 1-μm intervals through the whole cell body. The contours of the M-cell soma on 3A10 immunoreactive images were outlined using Adobe Photoshop CS3. The surface area of each M-cell soma was calculated by integration of the soma circumferences in each optical section. GlyR clusters were defined as puncta with a diameter of 0.5–2.0 μm according to the established protocol (Triller et al. 1985). Statistics are represented by mean ± SEM, and a t-test was used except where indicated otherwise.
Whole-cell recording of M-cells in Tol-056 larvae, in which the M-cells are identifiable as GFP-positive neurons, was carried out at 26–28 °C as described previously (Tanimoto et al. 2009). Larvae at 3 dpf (3.0–3.6 dpf) were temporarily anesthetized and immobilized in 0.02% tricaine methanesulfonate (MS-222; Sigma-Aldrich) and 1 mm D-tubocurarine (Sigma-Aldrich) for approximately 20 min. The larvae were then rinsed and pinned on a silicone-coated dish with fine tungsten pins. The larvae were soaked in extracellular solution containing (in mm) 134 NaCl, 2.9 KCl, 1.2 MgCl2, 2.1 CaCl2, 10 HEPES and 10 glucose; Tetrodotoxin (TTX, 1 μm; Wako, Osaka, Japan) was used as a voltage-gated sodium channel blocker, and the pH was adjusted to 7.8 with NaOH. In some experiments, d-2-amino-5-phosphonopentanoic acid (APV, 50 μm; Sigma-Aldrich, an NMDA receptor blocker), 6,7-dinitroquinoxaline-2,3-dione (DNQX, 50 μm; Tocris Bioscience, Bristol, UK, an AMPA receptor blocker), bicuculline (10 μm; Sigma-Aldrich, a GABAAR blocker) or strychnine hydrochloride (5 μm; Sigma-Aldrich, a GlyR blocker) was added to the extracellular solution. Miniature postsynaptic currents (mPSCs) were recorded at a holding potential of −60 mV and sampled at 100 kHz with a MultiClamp 700B amplifier controlled by Clampex 10.2 (Molecular Devices, Sunnyvale, CA, USA). Recording pipettes with a resistance of 2.3–6.5 MΩ were filled with an intracellular solution containing (in mm) 130 CsCl, 2 MgCl2, 10 HEPES, 10 EGTA and 4 Na2ATP at 290 mOsm, and adjusted to pH 7.2 with KOH. Series resistance was compensated by 70%–80%. Patched neurons were labeled with 0.005% Alexa Fluor 594 hydrazide (Invitrogen, Carlsbad, CA, USA) in the intracellular solution. Data were filtered at 3 kHz. Synaptic events were detected using the template function for events more than three standard deviations above the basal noise with Clampfit 10.2 (Molecular Devices).
We wish to thank Drs S. Takagi, M. Tanimoto and K. Ogino for their discussions and comments; J. R. Fetcho, S. Higashijima and K. Horikawa for supplying fish and Ms Y. Matsutani for fish care. This work was supported by the National BioResource Project, Zebrafish. This work was also supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan to H.H. and Y.O.; by the Takeda Science Foundation; and by a Career Development Award from the Human Frontier Science Program to H.H. Finally, I.Y. was supported by a fellowship from the Japan Society for the Promotion of Science.