Single-channel study of the spasmodic mutation alpha1A52S in recombinant rat glycine receptors

Abstract

Inherited defects in glycine receptors lead to hyperekplexia, or startle disease. A mutant mouse, spasmodic, that has a startle phenotype, has a point mutation (A52S) in the glycine receptor alpha1 subunit. This mutation reduces the sensitivity of the receptor to glycine, but the mechanism by which this occurs is not known. We investigated the properties of A52S recombinant receptors by cell-attached patch-clamp recording of single-channel currents elicited by 30-10000 microM glycine. We used heteromeric receptors, which resemble those found at adult inhibitory synapses. Activation mechanisms were fitted directly to single channel data using the HJCFIT method, which includes an exact correction for missed events. In common with wild-type receptors, only mechanisms with three binding sites and extra shut states could describe the observations. The most physically plausible of these, the 'flip' mechanism, suggests that preopening isomerization to the flipped conformation that follows binding is less favoured in mutant than in wild-type receptors, and, especially, that the flipped conformation has a 100-fold lower affinity for glycine than in wild-type receptors. In contrast, the efficacy of the gating reaction was similar to that of wild-type heteromeric receptors. The reduction in affinity for the flipped conformation accounts for the reduction in apparent cooperativity seen in the mutant receptor (without having to postulate interaction between the binding sites) and it accounts for the increased EC50 for responses to glycine that is seen in mutant receptors. This mechanism also predicts accurately the faster decay of synaptic currents that is observed in spasmodic mice.

Figure 10. Close-up of the subunit interface at the base of the apo Aplysia-AChBP crystal structure

This region would be immediately above the membrane domains, which begin after the β10 strand, in the full-length receptor. Two out of the five protomers are shown (B in pink and C in blue). The probable position of the A52S mutation is marked, as an alanine, in protomer B. The adjacent intersubunit hydrogen bond network between loops 2 and F is shown as dotted lines, with water molecules in red (W1 and W2). Residues are numbered according to the Aplysia AChBP sequence. According to published structure-based sequence alignments (Brejc et al. 2001; Hansen etal. 2004), the correspondence between residues is as follows: Asn48 in Aplysia-AChBP is Met in the glycine receptor α1 subunit, Lys173 is Gln and Tyr174 is Phe. The bulky aromatic residue at the beginning of the β9 strand is conserved across the superfamily. Bond lengths are in Angstroms.

Figure 1. Heterogeneous and homogeneous clusters of α1 A52S-β heteromeric receptor activations recorded in 1 mM glycine

A, when expression of the receptor was low enough to see only isolated single clusters, the P_open of each cluster was very consistent within patches and between patches. B, two clusters separated by about 25 s, and indicated by boxes in A, are shown on an expanded scale. The other clusters in the patch were similar. No double activations were seen in this patch. C, in a patch where more than one channel was typically active, which was typical of patches recorded more than two days after transfection of the cells, large variability of cluster P_open was observed, with two or more apparent populations. D, two representative clusters, indicated by boxes in (C), with very different open probabilities. One population (upper trace) seemed to display an open probability similar to that observed in the homogeneous recordings. Patches that demonstrated this kind of heterogeneous mixture of clustered activations were discarded. The data were filtered at 3 kHz.

Figure 2. Activation of heteromeric α1 A52S-β receptors by increasing concentrations of glycine

A, raw data traces are at five concentrations of glycine on heteromeric α1 A52S-β receptors. Bursts of openings occur at 30 μm glycine, but at higher concentrations, these activations group into clusters. B, the empirically fitted dwell-time histograms show that at low concentration, an appreciable proportion of openings are too fast to be observed. At higher concentrations of glycine, the apparent open time lengthens, mainly because the number of short shuttings that are missed increases progressively with glycine concentration. The predominant fast shut time observed in wild-type receptors is similar for A52S, but less pronounced. The longer intracluster shut times that are observed at low concentrations probably represent unbinding and rebinding of agonist, because they gradually shorten as the receptor becomes progressively more heavily liganded at higher concentrations of glycine. At the highest concentration, when the receptor is saturated with agonist, these longer shut times all but disappear. Dwell-time distributions were fitted with mixed exponential densities; the number of components is the same as summarized in Tables 1 and 2. The number of components required to fit the open and shut times observed at 100 μm glycine were three and five, respectively. Open-time distributions at 30, 100, 300, 1000 and 10000 μm glycine include 10480, 7561, 8430, 5582 and 4693 openings, respectively. Shut-time distributions include (in the same order): 10480, 7562, 8431, 5583 and 4694 shut times.

Figure 4. The probability of the channel being open during clusters of α1A52S-β mutant receptor activations

A, expanded view of the beginning of representative clusters for α1 A52S-β heteromeric receptors at five concentrations of glycine. The consistent amplitude of the activations is obvious. Openings are upward. The shortening (and eventual disappearance) of long shuttings within the clusters as the concentration increases is obvious. B, the single-channel P_open–concentration response relation for A52S heteromeric receptors is shifted to the right, compared with wild-type receptors. The dotted line is a Hill equation fitted to data from wild-type receptors (data from Burzomato et al. 2004). Note that the maximum fitted P_open is very similar for A52S receptors (97% for A52S, and 98% for wild type), but the Hill slope for the mutant receptor (2.2) is lower than for wild type. Both curves are plotted on an absolute scale, not normalized. Simultaneous fits to wild-type and mutant data with the Hill slope constrained to be the same for each curve did not describe either set of data satisfactorily (data not shown).

Figure 3. Properties of bursts for wild-type and α1A52S mutant heteromeric glycine receptors

A, a typical burst-length distribution for α1-A52S-β heteromeric glycine receptors at 30 μm glycine. B, the same distribution for wild-type heteromeric receptors, using the data of Burzomato et al. (2004) is plotted for comparison at 10 μm glycine, a concentration equi-effective in terms of P_open. The burst-length distributions are fitted with a mixture of exponential densities, with three components in each case (see Table 3). In comparison with wild type, A52S records show a much smaller proportion of long bursts (that is, those longer than 10 ms) and many more short bursts, the fastest of which arise from isolated single apparent openings. The critical time for dividing the record into groups of openings arising from a single channel was 4 ms for both datasets. This choice was unambiguous in the case of the wild type, but not for A52S (see Fig. 2 and text). This certainly resulted in a large number of misclassified bursts for A52S. The histogram for αA52S includes 3316 bursts and the parameters were τ₁ = 0.06 ms (area 27%), τ₂ = 0.4 ms (area 25%), τ₃ = 3.4 ms (area 48%). The wild-type histogram contains 1598 bursts, and the fitted parameters were τ₁ = 0.4 ms (area 32%), τ₂ = 3.4 ms (area 44%), τ₃ = 17 ms (area 24%). The burst P_open distribution for αA52S (C) is unusually flat, which was not predicted by any mechanism we fitted (see Results).Wild-type bursts of openings, contrastingly, had P_open (D) that was strongly skewed towards the maximum value, much as predicted from simulations. It is not possible to calculate the P_open for bursts that consist of a single opening, so these were excluded. The total number of bursts plotted in these histograms are (for A52S, C) 1702 and (for wild-type, D) 898.

Figure 9. The flip mechanism describes the observed data for αA52S well on all the criteria

The flip mechanism (A, Scheme 3 of Fig. 5) has only 14 free parameters, yet predicts the observed P_open–concentration relation (B) very well. Although the dwell-time distributions (C) were not perfectly predicted by the fitted rates, the errors were quite minor and tended to be in the fastest shut times, of which many are missed. Only shut times shorter than t_crit (arrow) are used for fitting (see Fig. 6 legend).

Figure 6. A simple mechanism with three binding sites fails to predict the observed data for the α1A52S-β mutant

A, a simple model for glycine receptor activation, with three binding sites and an open state corresponding to each bound state (Scheme 1 from Fig. 5). B, experimental P_open values are plotted as filled circles against the glycine concentration. The solid line is the apparent P_open–concentration curve predicted by the fitted scheme and rate constants taking into account the effect of missed events. The dashed line is the ideal curve expected if no events were missed. The predicted P_open–concentration curve does not describe the observed data, mainly because its slope is too shallow (on average 1.5). These plots indicate that this mechanism describes the data poorly for the αA52S mutant. C, all the plots show a comparison of the predictions of the fit with the experimental data. The mechanism was fitted simultaneously to four sets of data at three different glycine concentrations; one of the four sets is shown in this and the other figures. The first two rows of plots show the apparent open- and shut-times distributions. The histograms are the experimental distributions (note that only shut times below t_crit are fitted, see Methods). These are the same in Figs 7–9. The open-time histograms at 300, 1000 and 10000 μm include 9034, 6448 and 8002 openings, respectively, and the shut time histograms include (in the same order) 9033, 6447 and 7998 shuttings. The solid lines are predicted (HJC) distributions calculated from the mechanism, the resolution and the values of the rate constants that were found to maximize the likelihood of the experimental sequences of single-channel openings and shuttings. These distributions allow for missed events on the basis of the imposed resolution, while the dashed lines are the distributions expected if no events were missed. In the third row, the mean durations of openings that are adjacent to shut times in a specified range of duration are plotted against the mean durations of the shut times in each chosen range. The ranges are contiguous; for example in the bottom left panel, the range boundaries are: 30 (the resolution), 100, 300, 2000 and 10000 μs. These plots illustrate the negative correlation between the duration of adjacent open and shut times. Experimental points are shown as open diamonds (± s.d. of the mean) joined by a solid line, predicted points as filled circles, and the theoretical continuous relationship between open time and adjacent shut time as a dashed line. In this mechanism, the affinity of the receptor for glycine varies with the number of binding sites that are occupied. The receptor can open from each bound state, and the resulting three different open states predict the observed open dwell times adequately. But the shut dwell times are not at all well described by this mechanism, and errors in the correlation plots are apparent. In particular, the fastest shuttings are not represented properly, suggesting that additional shut states are required to describe the behaviour of the receptor. For the distributions of apparent shut times (C, 2nd row), the value of t_crit is shown by a vertical arrow. Only shut times shorter than t_crit were used for fitting (see Methods and Results; arrow). The observations, and the predicted fit, are shown for longer shut times to show the small number of 1–10 ms shut times at 10 mm glycine that are not predicted by the fit.

Figure 7. A mechanism that includes three extra distal shut states fits the data well for the αA52S mutant, apart from P_open

A, this mechanism (Scheme 2 from Fig. 5) has 14 free parameters, as each binding site is represented as equal and independent. Although some errors are found in the fast components of the shut-time distribution at the lower concentrations (C), the dwell times and correlation plots were well described. However, the P_open–concentration response curve that this mechanism predicted (B) was too shallow to be satisfactory. Only shut times shorter than t_crit (arrow in C) were used for fitting (see text and Fig. 6 legend) and the (few) 1–10 ms shut times at 10 mm glycine are not correctly predicted by the fit. Of the models that we describe here, this model predicted these shut times somewhat better than the others (perhaps simply because it has the largest number of free parameters), but still not well.

Figure 8. When the binding of glycine is allowed to vary with the level of liganding, the description of the data for the αA52S mutant is slightly improved

This mechanism (A, Scheme 2 of Fig. 5) had 18 free parameters. On average, the P_open–concentration response data (B) were better predicted by the steeper curve that this mechanism produced, than when the binding affinities were constrained to be the same for each step. There was no noticeable improvement in the description of dwell times or correlation plots (C), which were fitted very well. Only shut times shorter than t_crit (arrow) were used for fitting (see legend to Fig. 6).

Figure 5. Some of the kinetic schemes that were tested for the αA52S heteromeric glycine receptor

These schemes (and other variants) were previously tested on wild-type α1β glycine receptors (Burzomato et al. 2004). Agonist molecules (glycine) are indicated by A, and the number bound to each state by a subscript. The resting (shut) states of the receptor are denoted R, and additional shut states either D (desensitized) or F (flipped, i.e. the altered pre-open conformation; see Results). Open states of the channel are indicated by an asterisk (e.g. A₃F*). The names of the rate constants for the different steps of the reactions are shown, and the statistical factors for the binding rate constants have been included (e.g. the association rate for the vacant receptor is 3k₊₁ because any of the three identical binding sites can be occupied).