Taurine-containing neurons in the amphibian retina
The taurine-positive neurons in tiger salamander retina were labeled by the specific antibody of taurine in frozen vertical sections. Figure 1A shows the immunocytochemical results, indicating that the anti-taurine labeling was present in the somas and processes of rods and a few cones, as well as in the displaced bipolar cells located in the outer nuclear layer (see asterisks). Displaced bipolar cells in the outer nuclear layer (ONL) of salamander retina have been characterized as Off-type bipolar cells in a previous study [56]. The anti-taurine labeling was also present in the bipolar cell somas in the inner nuclear layer (INL) and the processes at the distal half of the inner plexiform layer (IPL). In contrast, no taurine antibody labeling was observed in the proximal half of the IPL. It is known that the axon terminals of On- and Off-bipolar cells in salamander retina are located separately at the proximal and distal half of the IPL respectively, therefore the taurine-positive labeling in the distal IPL indicates that Off-bipolar terminals may contain taurine. The taurine antibody-labeling results indicated that both amacrine and ganglion cell somas were taurine negative in salamander retina.
To further characterize taurinergic neurons, a specific taurine transporter antibody was used. Figure 1B shows transporter antibody labeled photoreceptor axones and some processes at the distal sublamina of the IPL in retinal sections (see arrows). The result of taurine transporter labeling was consistent with taurine labeling further indicating that Off-bipolar cells might be taurinergic neurons releasing taurine. As Off-bipolar cells are also glutamatergic neurons, our results suggest that these neurons might release both glutamate and taurine. Therefore, amacrine and ganglion cells as the third-order neurons receive juxtaposed inputs of glutamate and taurine in the IPL.
Taurine dose responses in isolated third-order neurons
To study the sensitivity of taurine, taurine dose-dependent currents were studied in dissociated third-order neurons in whole-cell voltage-clamp mode. The cells were identified by their electrical properties with voltage-dependent Na+ currents in multiple depolarizing voltage steps. Photoreceptors, bipolar cells and horizontal cells were lacking Na+ currents, enabling the third-order neurons in isolated cell preparations to be distinguished. Third-order neurons can be further characterized as Na+ currents are much smaller in amacrine cells compared to those in ganglion cells. The cells were voltage-clamped at -70mV, near the resting potential of the cells. Inward currents were recorded from the cells elicited with taurine concentrations of 10µM, 20µM, 50µM, 100µM, 500µM, 1mM. 2mM and 4mM puffed locally on the neurons. Figure 1C shows typical taurine currents elicited by various concentrations recorded from ganglion cells. The amplitudes of taurine currents were measured and plotted as a dose response relationship curve (Fig. 1D), showing that 10µM taurine elicited a negligible current with maximum currents elicited by 2mM taurine. A taurine concentration of 330µM elicited a half maximum response (EC50) determined from the dose response relationship curve (see dot-lines, Fig. 1D). Taurine dose response currents also indicate that concentrations of taurine below the EC50 value elicit sustained inward currents with limited desensitization. Transient currents with fast desensitization were observed when the concentration was higher than the EC50, as shown in figure 1C.
Taurine suppressed glutamate-induced [Ca2+]i in the third-order neurons
The effect of taurine on glutamate response was studied in the third-order neurons using Ca2+ imaging. The isolated cells were loaded with a membrane permeable Ca2+ indicator Fluo-4-AM. The third-order neurons were identified by their morphology (see the Methods section). Application of 200µM glutamate evoked a large [Ca2+]i increase from the third-order neurons. When taurine was co-applied with glutamate in 10µM, 350µM and 2mM, the minimum, EC50 and maximum concentrations respectively, the [Ca2+]i was suppressed in a dose-dependent manner. Figure 2A shows the [Ca2+]i changes in the third-order neurons activated by glutamate and glutamate with varying concentrations of taurine. On average, 10µM taurine reduced 23% of glutamate-elicited free [Ca2+]i levels, with the EC50 and maximum concentration of taurine blocking 40% and 67% of the glutamate-elicited free [Ca2+]i respectively in the third-order neurons (n=40). This indicates that the effect of taurine effectively suppressed glutamate-elicited [Ca2+]i in the neurons.
As glutamate activates kainate-, AMPA- and NMDA-sensitive ionotropic receptors, we used the receptor agonists to separate activity at each single receptor subtype. Previous Ca2+ imaging studies indicate that low concentrations of kainate and AMPA specifically activate their own receptors that are highly permeable to Ca2+ in the third-order neurons [17]. Kainate and AMPA at concentrations of 15µM increased [Ca2+]i in the neurons (Fig. 2B &2C). 10µM taurine reduced AMPA-induced [Ca2+]i to 80% of the control (n=72), but had no significant effect on kainate-induced [Ca2+]i with no reduction in the level of [Ca2+]i (n=23). This indicates that the low concentration of taurine selectively suppressed AMPA-sensitive receptor response in the neurons. Higher taurine concentrations gradually suppressed AMPA-induced [Ca2+]i, but acutely reduced kainate-induced [Ca2+]i in the neurons (Fig. 2B and 2C). Kainate-induced [Ca2+]i, but not AMPA-induced [Ca2+]i, was completely blocked by 2mM taurine. Since AMPA receptors, but not kainate receptors, are synaptic in the third-order neurons [18], the different effects of taurine on AMPA and kainate responses suggest that taurine might have alternate functions on synaptic and non-synaptic glu tamate receptors. The effect of taurine on NMDA receptors was also examined in the third-order neurons. On average, 10µM and 350µM taurine suppressed 29% and 77% of NMDA-induced [Ca2+]i in the neurons, respectively (n=27). NMDA-induced [Ca2+]i was completely blocked by 2mM taurine (Fig. 2D).
Taken together, these results indicate that taurine at higher concentrations from 350µM to 2mM could effectively block Ca2+-permeable kainate and NMDA receptors but have a lesser effect on Ca2+ -permeable AMPA receptors.
The effect of taurine was sensitive to strychnine and picrotoxin, but not GABA receptor antagonists
As the third-order neurons possess both glycine and GABA receptors, the effect of taurine on suppression of glutamate-induced [Ca2+]i was examined with the addition of various glycine and GABA receptors antagonists. Figure 3A shows that 2mM taurine suppressed glutamate-induced [Ca2+]i by 55% (t=17.21, df=136, p<0.0001), which was significantly reversed by 2µM strychnine to 75% of the control (t=6.44, df=136, p<0.0001). The remaining strychnine-insensitive effect indicates that taurine might activate different types of receptors. The effect of taurine on AMPA- and kainate-induced [Ca2+]i in the third-order neurons was also tested. In the control, 2mM taurine significantly suppressed both 15µM AMPA- by 65% (t=15.78, df=82, p<0.0001) and 15µM kainate-induced [Ca2+]i by 85% (t=19.65, df=52, p<0.0001) in the third-order neurons, as shown in figure 3B and 3C. Though strychnine appeared to be more effective on blocking the effect of taurine on AMPA-induced [Ca2+]i compared to kainate-induced [Ca2+]i, the reversal was significant for both AMPA (from 35% to 85%; t=9.4488, df=82, p<0.0001) and kainate (from 15% to 60%; t= 5.52, df=52, p<0.0001). These results suggest that the action of taurine on glutamate receptors could therefore be mediated by strychnine-sensitive receptors.
The effect of strychnine could be duplicated by 100µM picrotoxin, a non-selective blocker for Cl- permeable receptors, including both glycine and GABA receptors. On average, picrotoxin significantly reversed the effect of 2mM taurine on glutamate-induced [Ca2+]i from 55% to 30% (t=3.93, df=96, p = 0.0002), showing in figure 3D. Picrotoxin was more effective in blocking the taurine effect on AMPA-induced [Ca2+]i. Figure 3E shows that in control taurine suppressed about 70% of AMPA-induced response; with picrotoxin taurine having 20% suppression on the AMPA-induced response (t=12.18, df=64, p<0.0001). This indicates that the effect of taurine on glutamate-induced [Ca2+]i is active through a Cl- permeable receptor blocked by picrotoxin.
The effect of picrotoxin and strychnine was not replicated by GABA receptor antagonists that had no significant effect on taurine suppression of glutamate-induced [Ca2+]i . Figure 4A and 4B show that both GABAA receptor antagonists, 10µM bicuculline (t=0.89, df=98, p=0.375) and 20µM SR95531 (t=0.59, df=80, p=0.5642) had no significant effect on reversing taurine-produced suppression on glutamate-induced [Ca2+]i . In addition the GABAC receptor antagonist 50µM TPMPA (t=0.26, df=70, p=0.7913) and the GABAB receptor antagonist 10µM CGP55845 (t=0.77, df=82, p=0.4422) were also unable to reverse taurine-produced suppression (Fig. 4C and 4D). These results clearly show that the effect of taurine was unrelated to all GABA receptor subtypes.
Taurine suppressed glutamate-induced Ca2+ influx through both glutamate receptors and voltage-gated Ca2+ channels
Glutamate elicited [Ca2+]i in the third-order neurons could be the result of activation of both metabotropic and ionotropic receptors. Therefore to separate their individual effects, CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) and AP-7 (D,L-2-amino-7-phosphonoheptanoic acid) were used to block all ionotropic glutamate receptors. The glutamate-induced [Ca2+]i was completely blocked by the presence of these ionotropic glutamate receptor antagonists, and in addition when 2mM taurine was co-applied it had no effect on [Ca2+]i in the neurons (Fig. 5A), suggesting that activation of ionotropic glutamate receptors causes a raise of [Ca2+]i in the third-order neurons.
Glutamate depolarizes the third-order neurons leading to activation of voltage-gated Ca2+ channels, therefore glutamate-induced [Ca2+]i could be the result of Ca2+ influx through both glutamate receptors and voltage-gated Ca2+ channels. To separate [Ca2+]i influx via glutamate receptors from voltage-dependent Ca2+ channels, 1mM cobalt was used to block voltage-gated Ca2+ channels. On average, cobalt reduced about 55% of the glutamate-induced [Ca2+]i (n=42, Fig. 5B, see asterisk). The remaining cobalt-insensitive [Ca2+]i was further suppressed by 2mM taurine, indicating that taurine suppressed Ca2+ influx through ionotropic glutamate receptors.
Glutamate had a limited effect on triggering internal Ca2+ release in isolated neurons
In many cases, Ca2+ influx could stimulate Ca2+ release from the intracellular Ca2+ stores. To assess the contribution of Ca2+ release from internal stores, 1.5µM ruthenium red was used to block ryanodine receptors in the intracellular organelles, which blocked Ca2+-sensitive release from internal stores in the third-order neurons. The effects of taurine on glutamate-induced Ca2+ increase were examined in the control and with ruthenium red. In the control, taurine suppressed 55% of glutamate-induced [Ca2+]i in the neurons, ruthenium red approximately reduced 15% of glutamate-induced [Ca2+]i (Fig. 5C), glutamate had a limited effect on triggering ryanodine receptor -sensitive internal Ca2+ release. With the blockage of Ca2+ release from internal stores, taurine suppressed 35% of the total glutamate-induced [Ca2+]i. Since the taurine suppression was reduced in the presence of ruthenium red, this indicates that the effect of taurine might reduce glutamate-induced internal Ca2+ release via ryanodine receptors.
Another approach was to use thapsigargin to block the Ca2+-uptake pump in intracellular Ca2+ release organelles resulting in depletion of internal Ca2+ release stores. The same experimental protocol was applied for the control and with the presence of 1.5µM thapsigargin. Figure 5D demonstrates that thapsigargin pre-perfusion only caused 5% reduction of glutamate-induced [Ca2+]i compared to the control. Depleting internal Ca2+ release stores had no significant effect on taurine suppression of glutamate-induced [Ca2+]i . With thapsigargin, taurine still reduced approximately 45% of glutamate-induced [Ca2+]i similar to the effect of taurine in control.
The above results confirm that Ca2+ influx stimulated internal Ca2+ release played a minor role in glutamate-induced [Ca2+]i, therefore the major cause of glutamate-induced [Ca2+]i was Ca2+ influx through receptors in isolated neurons. Possibly, the effect of taurine was via regulation of Ca2+ permeability of glutamate receptors and voltage-gated Ca2+ channels.
Taurine regulation of glutamate-induced [Ca2+]i was sensitive to CaMKII and PKA inhibitors
If intracellular signal transduction pathways are involved in taurine regulation, then their inhibitors should reverse the suppressive potential of taurine. Several intracellular protein inhibitors were used to block taurine produced effect and among these inhibitors the selective cell-permeable CaMKII inhibitor KN-62 and protein kinase A inhibitor PKI (14-22) amide (both 1µM) could partially reverse the taurine-produced suppressive effect on glutamate responses. Figure 6A shows the statistics of the reductions of glutamate-induced [Ca2+]i by taurine in control and in the presence of KN-62 or PKI (14-22). In the control, taurine suppressed glutamate-induced [Ca2+]i. by 62%+3% (n=40). Application of KN-62 had no effect on glutamate-induced [Ca2+]i., however with KN-62, taurine suppression was reversed significantly to 47%+3% (n=40, t=4.44, df=78, p<0.0001). In the case of PKI (14-22) amide, taurine suppressed glutamate-induced [Ca2+]i by 63%+5% (n=38) in the control, which was reduced to 52%+4% (t=1.99, df=74, p=0.0498) with blockade of PKA. This indicates that both CaMKII and PKA are the intracellular proteins involved in the taurine signaling transduction.
As shown in figure 3A the taurine suppression of glutamate-induced [Ca2+]i was via both the strychnine-sensitive and -insensitive receptors, CaMKII and PKA and could involve the downstream pathways of these taurine-activated receptors. To specify which type of taurine-activated receptors triggers the CaMKII-sensitive regulation in glutamate-induced [Ca2+]I, we performed the experiments with strychnine to block strychnine-sensitive taurine receptors, isolating strychnine-insensitive taurine receptor response. Figure 6B is the statistical results from 48 cells, showing that 2µM strychnine blocked 67% of taurine-caused suppression on glutamate-induced [Ca2+]i. With strychnine, KN-62 further blocked about 11% of the taurine suppression on glutamate-induced [Ca2+]i (t=3.60, df=88, p=0.0005), indicating that CaMKII mediates strychnine-insensitive taurine regulation in the third-order neurons.
Evidence of Ca2+ -dependent regulation of taurine response
We found that while taurine regulates glutamate response, Ca2+ influx via glutamate receptors also regulates taurine response. This was studied in a whole-cell voltage-clamp recording mode. Taurine-elicited currents were measured in the presence and absence of glutamate receptor agonists. As the results in figure 2 show that activation of AMPA and kainate receptors could effectively raise [Ca2+] levels, both 15µM AMPA and kainate were used to elevate [Ca2+]i levels in the neurons. A taurine dose of 350µM was applied before the application of AMPA or kainate. Taurine elicited a large inward current with fast desensitization at a holding voltage -70mV, and adding AMPA or kainate had no significant change on the substandard taurine currents (the left panels, Fig. 7A &7B). After pre-application of AMPA and kainate however, the amplitudes of taurine currents were reduced in the same cells, and the kinetics of the onset currents was much slower when compared to the control (the right panels, Fig. 7A &7B). Although 15µM AMPA and kainate alone could elicit large [Ca2+]i increases (see Fig. 2B and 2C), the currents elicited by these agonists were small at -70mV. Therefore, the effect of AMPA and kainate on taurine-elicited currents might be mainly caused by an increase in [Ca2+]i levels, rather than conductance change of the neurons.
To further identify that increasing [Ca2+]i is the key for altering taurine response, a fast Ca2+ chelator BAPTA was applied to buffer [Ca2+]i in the micro-domain regions under the plasma membranes of the cells. BAPTA (10mM) was applied into the cytosol through recording electrodes during whole-cell recording. It was determined that BAPTA reduced the effects of AMPA and kainate on taurine-elicited currents (6 cells out of 8 recordings, Fig. 7C &7D). With BAPTA buffer, both kinetics and amplitudes of taurine currents increased suggesting that intracellular free Ca2+ is a key element for regulation of taurine currents by activation of glutamate receptors in the third-order neurons.
Taken together, our study indicates that taurine and glutamate reciprocally inhibit each other in the third-order neurons. Figure 8 depicts the intracellular regulation between taurine and glutamate, which summarizes our findings. Taurine inhibits glutamate-induced [Ca2+]i via a CaMKII pathway. Meanwhile rapid a increase of intracellular free Ca2+ by activation of AMPA and kainate receptors in turn negatively controls taurine response in the retinal third-order neurons. However, internal Ca2+ release from the cellular organelles seemed to be less effective in the intracellular regulation pathways between taurine and glutamate in isolated third-order neurons.