Onset of macroscopic currents mediated by α2 homomeric GlyRs
The activation phase of currents evoked by concentration steps of an agonist gives information on the activation kinetics of the receptor channel. The rising phase of currents evoked by glycine on α2 homomeric GlyRs was analysed on outside-out patches (see Methods). For each concentration of glycine, a series of 15–50 trials evoked with a ≥ 10 s interval was used to generate macroscopic averaged traces, as exemplified in Fig. 3A.
Figure 3. Activation time course of glycine-evoked responses
A, averaged traces of currents (n= 15–50) obtained in different patches showing the activation phase of the responses evoked by the application of 0.1, 0.3, 1, 3, 10 and 30 mm glycine. Traces were normalised to their maximum amplitudes. B, normalised averaged current from 20 responses evoked by step applications of 1 mm glycine. Note that the activation phase has two components with a fast time constant τfast= 7 ms (69 %) and a slow time constant τslow= 74 ms (31 %). C, the plot of fast rising rates (1/τfast) versus glycine concentrations was fitted with the equation 1/τfast=α+β([glycine]n/([glycine]n+rEC50n)), giving: α= 20 s−1, β= 5087 s−1, rEC50= 13 mm and n= 1.35. Insert shows a similar plot for the slow rising rate (1/τslow). Note that the slow rising rate was equally concentration dependent. Each point of both plots represents the average of 5–19 experiments. D, plot of the relative proportion of the fast (•) and the slow (□) rising phase components versus glycine concentration. The relative proportion of the slow component decreased when the concentration of glycine increased. Each point is the average of 5–19 experiments. Standard deviations are identical for both component areas and are thus only indicated for the fast component area.
Download figure to PowerPoint
As shown in Fig. 3A, the rise time of the glycine-evoked currents decreased when the agonist concentration increased, to reach a minimum at a glycine concentration of 100 mm (Fig. 3C). For each concentration, the pulse duration was adjusted (0.4–3 s) in order to obtain a steady-state current. The rising phases of the outside-out currents evoked by the application of ≤ 10 mm glycine were best fitted with the sum of two exponential curves. In three out of the seven patches tested with 30 mm glycine concentration steps, and in four out of the seven patches tested with 100 mm glycine, the outside-out current exhibited an activation phase that was best fitted by a single exponential function. Figure 3B shows a representative onset of an averaged patch current evoked by a 1 mm glycine concentration step. In this example, the rising phase was fitted with the sum of two exponential curves giving time constants, τfast= 7 ms (69 %) and τslow= 74 ms. As shown in Fig. 3C and D, the time constant values and relative areas of the fast and slow components were dependent on the agonist concentration. Plot of the τfastversus concentration was fitted with the following equation (Legendre, 1998):
where α is an approximation of the closing rate constant, β is an approximation of the opening rate constant, n is the number of binding sites, and rEC50 is the concentration of glycine that give half of the maximum opening rate constant (Colquhoun & Hawkes, 1995). The fast component increased with a slope factor n of 1.35 from a minimal rate constant of 20 s−1 to a maximal rate constant of 5087 s−1 suggesting that the opening rate constant β of the channel is considerably higher than the closing rate constant α. rEC50 was 13 mm. The efficacy of the receptor (E) could therefore be estimated ≈255 since E=β/α (Colquhoun, 1998). Accordingly, the efficacy of the α2 homomeric GlyRs could be > 20 higher than the E value estimated for synaptic heteromeric α1/β GlyRs (E= 11; Legendre, 1998).
To confirm this point, we have analysed the open and close time distribution in single receptor bursts of openings in response to short (1 ms) concentration pulses of glycine near GlyR saturation (30 mm). To perform this analysis, patches with a single functional GlyR were selected. As shown in Fig. 4A, a GlyR opens in bursts of long openings interrupted by very short closures. Single openings and closures were manually detected and measured using a filter cut-off frequency of 5 kHz. Opening and closing time constants were estimated by pooling measurements made on 110 sweeps from three patches. The open time histogram was best fitted by a single exponential curve (Fig. 4Ba) giving a open time constant value of τo= 49.6 ms. The closed time histogram was also best fitted with a single exponential curve (Fig. 4Bb) with time constant τc= 0.19 ms. This analysis gave an estimation of the closing rate constant α as α= 1/τo. Accordingly, α= 20.1 s−1, a value closely similar to the value obtained by analysing the activation rate constant of averaged outside-out currents.
Figure 4. Open time and closed time distribution of the α2 homomeric GlyR
A, example of single-channel recordings obtained in response to 1 ms step application of 30 mm glycine (cut-off filter frequency = 2 kHz for display purposes; VH=−50 mV). Ba and b, open time (Ba) and closed time (Bb) distribution obtained by pooling single-channel currents (110 trials) obtained in response to 1 ms step application of 10–30 mm glycine in three different experiments. Distribution histograms are shown as a function of log interval with the ordinates on a square root scale. Both distribution were best fitted with a single exponential function giving an open time constant τo= 49.6 ms and a closed time constant τc= 0.19 ms.
Download figure to PowerPoint
Closed time histogram reveals the presence of a single closed time constant suggesting a simple mechanism underlying channel reopening within a burst. Assuming a simple Markov model with several liganded closed states, the number of openings per burst will depend both on the dissociation rate constant (koff) linking the liganded closed state near the open state to another liganded closed state and on the opening rate constant β linking the liganded closed state to the open state. Accordingly, the number of openings per burst No= 1+β/koff. On 110 trials analysed, the average number of openings per bursts was 2.7 indicating that koff is ≈2 times slower than β. Therefore, contrasting with its complex activation behaviour, the deactivation of the α2 homomeric GlyR seems to be governed by a very simple mechanism which can be reduced to a single closing/opening rate equilibrium governed by the open rate and the dissociation rate ratio.
To further characterise the kinetic properties of the α2 homomeric GlyR, we have investigated the deactivation properties of this receptor when activated by the application of different concentrations of glycine. The decay phase of outside-out currents evoked by short pulses (1 ms) of 10 mm of glycine (Fig. 5) could be systematically fitted with a single exponential function, giving a mean time constant of 159 ± 70 ms (n= 13). No significant differences were observed between experiments using 1, 10 and 30 mm glycine (P > 0.05 using one-way ANOVA). Because of the poor opening probability of GlyRs in response to short pulses of glycine at concentration ≤ 1 mm (see below), we could not test short pulses with lower concentrations of glycine. Nevertheless, the decay phases of currents evoked by long pulses (0.4–3 s) of 0.1–30 mm were equally fitted by single exponential curves with time constant values of ≈160 ms, which were not significantly different between long and short pulses or between concentrations (P > 0.05 using two-way ANOVA). These results suggest that the deactivation of α2 homomeric GlyR opening is underlied by bursts arising from a single open state. However, we cannot completely exclude the presence of other open states linked to other liganded closed state if they have nearly similar opening and closing rate constants.
Figure 5. Slow decay time constant of currents mediated by α2 homomeric GlyRs
Superimposed traces of 10 responses obtained from a set of 50 individual currents evoked by identical step applications of 10 mm glycine (1 ms, VH=−50 mV). The lower trace represents the average of the 50 responses and was fitted with a single exponential function with a time constant value τdecay= 153 ms.
Download figure to PowerPoint
Single-channel first latency analysis reveals a slow and bimodal activation of α2 homomeric GlyRs
To determine the microscopic determinants of the slow and biphasic onset of macroscopic GlyR currents, we analysed the activation of the α2 homomeric GlyR in outside-out patches containing 1–3 active GlyRs (see Methods). Figure 6 shows the activation of a single receptor in response to repetitive step applications at different concentrations of glycine and in different patches. Averaged single-channel responses (80–200 trials) produce ensemble currents with time courses similar to the ones observed for macroscopic currents previously described (see Fig. 3). At all concentrations tested (0.1–30 mm), the ensemble averaged currents exhibit a biphasic rising phase with fast and slow components similar to macroscopic currents. To determine if the biphasic component of the activation phase reflects a complex GlyR behaviour occurring before the GlyR channel opens, we have analysed the distribution of initial closed times leading to the first opening (first latencies). As shown in Fig. 6, the increase in the activation time of the averaged macroscopic current obtained when the glycine concentration was decreased appeared to be related with an increase in the first latency (FL) duration. The FL cumulative distribution was clearly bimodal as shown in the example of Fig. 7A. In this example, the activation of a single GlyR was evoked by 1 mm glycine application and the FL distribution was fitted by the sum of two exponential functions with time constants τfast= 4.9 ms (68 %) and τslow= 71 ms. The corresponding averaged ensemble current exhibited a rising phase with τfast= 4.9 ms (67 %) and τslow= 75 ms, suggesting that the biphasic activation phase of the ensemble current could be determined by changes in GlyR conformation leading to channel openings. Accordingly, FL distribution of single-channel openings evoked by other concentrations of glycine (0.1–30 mm) equally exhibited two components (Fig. 7B and C). As previously observed for the analysis of activation rate constants, the two FL rate constants decreased when glycine concentration increased, and the fast component became dominant for glycine concentrations ≥ 1 mm (Fig. 7D).
Figure 6. The first latency of activation decreases when the glycine concentration increases
A and B, representative, non-consecutive, single-channel openings of a single α2 homomeric GlyR evoked by a 400 ms step application of 10 mm glycine (A) and 3 mm glycine (B) on the same patch. Ensemble average currents (lower traces, n= 150 for each concentration) were best fitted with a bi-exponential function (smooth lines). Fast and slow time constant and their relative areas are indicated for both concentrations. C, representative, non-consecutive single-channel openings evoked by a 3 s step application of 0.3 mm glycine. The ensemble average current (lower trace, n= 80) was best fitted with a bi-exponential function (smooth line) with time constants and relative weights as indicated.
Download figure to PowerPoint
Figure 7. First latency analysis of α2 homomeric GlyRs
A, example of a first latency distribution obtained at 1 mm glycine in a single-channel experiment. This first latency distribution was best fitted by the sum of two exponential functions (smooth line). Broken lines represent each exponential function with their time constant values and their relative areas indicated on the right. B, example of first latency distributions obtained at different glycine concentrations (0.1–30 mm) from different patches. The number of first latencies pooled ranges from 100 to 200 events for each concentration. All first latency distributions were best fitted by sums of two exponential functions (smooth lines). C, plot of the fast (1/τfast) and slow (1/τslow) rate constants versus glycine concentration. The fast rate was fitted with the equation 1/τfast=β([glycine]n/([glycine]n+rEC50n)), with β= 3835 s−1, rEC50= 10.2 mm and n= 1.56. D, plot of the relative proportion of the fast (•) and the slow (□) rising phase components versus glycine concentration. The relative proportion of the slow component decreased when the concentration of glycine increased. Each point is the average of 2–7 experiments. Standard deviations apply to both slow and fast components and are indicated only for the fast component for all concentrations analysed except for 0.1 mm (n= 2).
Download figure to PowerPoint
The plot of the fast FL rate versus glycine concentration was fitted with the following equation:
This equation only differs from the equation previously used for macroscopic current onset by the absence of the added constant α since the closing rate constant (channel opening time constant) was not involved in first latency. This fit gave an opening rate constant β of 3853 s−1, a rEC50 of 10.2 mm and a slope factor n of 1.56. These values are very similar to the ones obtained with macroscopic currents. This further suggests that the rising phase of the currents evoked by glycine concentrations ≥ 0.03 mm is likely to be governed by conformational changes between closed state before channel openings.
The presence of a slow FL component could suggest the existence of a desensitised state with a fast desensitisation rate and a fast recovery rate linked to a partially liganded closed state. As the slow FL rate constant was concentration-dependent, this partially liganded closed state is unlikely to be directly linked to the single open state we determined. Indeed, if this closed state was directly linked to the putative desensitised state and to the open state, the slow FL rate constant would be independent of agonist concentration (Burkat et al. 2001).
To further determine if a desensitised state is linked to a liganded closed state, distally located to the fully liganded closed state leading to channel openings, we used a protocol derived from experiments demonstrated by Mozrzymas et al. (2003) to show the existence of a mono-liganded desensitised state in the GABAA receptor. This experiment consists of analysing the consequences of the outside-out patch pre-incubation using an infra-liminar concentration of agonist on the responses evoked by the application of a near-saturating concentration of glycine. Pre-incubation with an infra-liminar concentration of agonist should accumulate receptors in partially liganded states. If a desensitised state is linked to a partially liganded closed state, pre-incubation will result in a decrease in the amplitude of the responses evoked by a high concentration of agonist (Mozrzymas et al. 2003). To perform these experiments, patches were pre-incubated with 10 µm glycine. At this concentration, we did not observed any channel openings for a period of time > 1 min (data not shown), which confirms the weak ability of GlyRs to reach a liganded closed state linked to an open state at such concentrations. As shown in Fig. 8, the pre-incubation with 10 µm glycine depresses the current evoked by 30 mm of glycine by 47 ± 15 % (n= 6) when compared to control responses evoked in the absence of pre-incubation with 10 µm glycine. This result further suggests the presence of a desensitised state linked to a partially liganded closed state.
Figure 8. Pre-equilibration of receptors at a low glycine concentration depresses the current responses evoked by a saturating glycine concentration
Ensemble averaged currents (n= 20) obtained in response to 400 ms step application of 30 mm glycine in normal saline solution (A), after a pre-equilibration with 10 µm glycine for ≥ 1 min (B) and after washing with normal saline solution (C) in the same patch.
Download figure to PowerPoint
α2 homomeric GlyRs have a low open probability when activated by a synaptic-like application
According to the opening (β) and the closing (α) rate constant values that we obtained (α≈ 20 s−1 and β≈ 5000 s−1), the maximum open probability (PO,max) of the homomeric α2 GlyR should be very high. Indeed, PO,max=β/(α+β), which predicts a PO,max close to 1 (0.996). To confirm this point, non-stationary variance-amplitude analysis (Sigworth & Sine, 1987) was used to estimate the maximal open probability of the receptor (see Methods). Maximal open probability of the receptor was analysed on responses evoked by 30 mm glycine applications. Figure 9B shows an example of a variance-amplitude plot computed from 50 responses obtained from a single channel in response to 400 ms step applications of 30 mm glycine. The duration of the application was set to obtain a maximum response amplitude. The variance-amplitude plot was fitted with the following equation (Sigworth & Sine, 1987):
where i is the elementary current, I the averaged macroscopic current and N the total number of available receptors. The maximal open probability (PO,max) of the receptor corresponds to the maximal amplitude observed for the averaged macroscopic current divided by the theoretical maximal current iN. At 30 mm glycine, the mean PO,max calculated from four different experiments was 0.94 ± 0.04. This is close to the predicted value calculated with α≈ 20 s−1 and β≈ 5000 s−1.
Figure 9. Maximal open probability of the α2 homomeric GlyR
A, traces illustrating six representative recordings from 50 responses used to construct the variance–amplitude plot in B. B, variance-amplitude plot computed from 50 responses obtained from a single channel activated by 400 ms step application of 30 mm glycine. The black curve represents the fitted model σ2=iI− (I2/N), with i= 5.29 pA and N= 1. The estimated maximal open probability is PO,max=Ipeak/(iN) = 0.97. C, bar graph of the open probability at the peak of GlyR currents obtained in response to long application (0.4–3 s) at 0.3 mm glycine (n= 4), 1 mm glycine (n= 4), 3 mm glycine (n= 3), 10 mm glycine (n= 4) and 30 mm glycine (n= 4).
Download figure to PowerPoint
Agonist clearance at the synaptic cleft is usually fast (< 0.2 ms; Clements, 1996) but despite a high maximal open probability, the ability of a synaptic release to fully activate postsynaptic receptors also depends on the receptor activation properties and on the concentration of neurotransmitter released in the synaptic cleft. The activation kinetics of the α2 homomeric GlyR current appears to be slow at concentrations ≤ 10 mm, which may suggest that those GlyRs will be inefficient in response to a brief synaptic release. To address this question, we have mimicked the synaptic release of glycine by applying short pulses (1 ms) of 1 mm glycine on outside-out patches assuming that the peak concentration of glycine released in the synaptic cleft is ≤ 1 mm (Legendre et al. 1998; Suwa et al. 2001). As shown in Fig. 10Ab, the amplitude of the current evoked by 1 ms step application of 1 mm glycine (black trace) is considerably reduced compared to the amplitude of the response evoked by a 400 ms step application (grey trace). When a 1 ms concentration pulse of 30 mm was applied, we were not able to fully activate the receptor (Fig. 10Aa). To estimate the open probability of the channel in response to a 1 ms pulse of 1 mm glycine, we first performed non-stationary noise analysis on outside-out responses evoked by 1 ms application of 30 mm glycine. By fitting the plot of the variance versus the amplitude of the responses (see above) (Fig. 10Ba), we were able to estimate a Po= 0.75 ± 0.05 (n= 5) for responses evoked by 1 ms concentration pulse of 30 mm glycine. In three of these experiments, we could also estimate in the same patch the GlyR Po in response to a 1 ms pulse of 1 mm glycine by using the parabolic fit obtained from the 30 mm glycine application. In these cases, the Po was estimated by dividing the maximum amplitude current, obtained experimentally when GlyRs were activated by a 1 ms pulse of 1 mm glycine, by the maximum estimated current given by the parabolic fit (Fig. 10Bb). This approach gave a mean Po of 0.1 ± 0.03 for GlyR activated by 1 mm synaptic-like pulses, which suggests that homomeric α2 GlyR would be inefficient if they were postsynaptically located and activated by a single vesicle release.
Figure 10. Activation of α2 homomeric GlyRs by brief pulses of glycine mimicking synaptic transmission
Aa and b, average of currents obtained in response to brief (1 ms; black traces; 50 trials averaged) and long (400 ms; grey traces; 15 trials averaged) step applications of 30 mm glycine (Aa) and 1 mm glycine (Ab) in the same patch. Ba and b variance-amplitude plot computed from 50 current transients obtained in response to brief pulses of 30 mm glycine (Ba) and 1 mm glycine (Bb) in the same patch, corresponding to traces in Aa and A2, respectively. The amplitude and the variance were computed for a period of 750 ms starting at the peak of the averaged response. The black curve (Ba) represents the fitted model σ2=iI− (I2/N), with i= 3.52 pA and N= 9.5. The open probability at the peak of the response was: Po= 0.75. The fit obtained in Ba was used in Bb (grey curve) to estimate the open probability of the receptor in response to brief pulse of 1 mm glycine giving Po= 0.07.
Download figure to PowerPoint