Interpretation of MTS modification results
The aim of this study was to investigate the differential involvement of the Ala101–Thr112 domain in contributing to the binding and signal transduction mechanisms of glycine and taurine. Conclusions are based partly on the reaction rates of charged MTS derivatives with cysteine-substituted mutant GlyRs. However, several limitations apply to the interpretation of such data. The first concerns the fact that some cysteine-substitution mutations resulted in large agonist sensitivity changes, implying that the mutations themselves may have induced local structural alterations. The implications of this for the interpretation of results are considered below for each affected mutant GlyR. A second limitation concerns the fact that each functional GlyR contains 5 introduced cysteines. This study has not attempted to ascertain how many cysteines need to be modified to give the entire effect. Thus, the true cysteine modification rates are not known, and hence, conclusions can only be drawn on the relative changes in the reaction rates, not on their absolute magnitudes. Finally, a lack of a measurable effect of an MTS reagent at a given substituted cysteine does not imply a lack of reactivity.
The reactivity of A101C with MTSET and MTSES indicates that Ala101 is accessible to the protein surface. The MTSET modification rate of A101C is significantly increased in the glycine-bound state relative to both the unbound and taurine-bound states (Fig. 4b). The MTSES modification rate is also significantly increased in the glycine-bound state relative to the unbound state (Fig. 6b). Therefore, electrostatic potential changes are eliminated as a major determinant of reactivity rate changes at this residue. Thus, the water accessibility of this residue is increased in the glycine-bound state. Possible causes of this effect are considered below. The reactivity rate of A101C with both MTSES and MTSET is lower than for any other residue measured in this study. In addition, DTT did not reverse the effects of MTSES or MTSET modification at A101C. This was most likely due to hindered DTT access to the disulfide bond. These observations suggest that Ala101 lies in a location of restricted surface access.
The MTS reactivity of the E103C, G105C, F108C and T112C mutant GlyRs indicates that these residues are also exposed at the protein aqueous interface. As none of these substituted cysteines responded to both MTSET and MTSES, it is not possible to draw definitive conclusions as to the roles of electrostatic forces and surface accessibility changes in contributing to differences in MTS reactivity rates. However, due to their close proximity, it is unlikely that the surface electrostatic potential differs significantly from Ala101 to Glu103. Therefore, as the MTSET modification rate of G103C is not state-dependent, it is unlikely that the surface accessibility of this residue is changed between the unbound, glycine-bound and taurine-bound states.
F108C is characterized by a drastically different MTSES reactivity rate in the unliganded state relative to the glycine- or taurine-bound states (Fig. 6b). A difference of this magnitude is unlikely to be solely the result of an electrostatic potential shift (cf. Pascual and Karlin 1998). Thus, the surface accessibility of Phe108 must change as the channel transitions from the closed to the open channel states, implying that either the residue itself or an overlying domain moves during channel activation.
A taurine binding site at Asn102
Asn102 has previously been identified as a glycine and taurine binding site, based on the effects of the N102A mutation (Vafa et al. 1999). The N102A mutation had a proportionately greater disruptive effect on glycine relative to taurine sensitivity. The present study demonstrates that N102C has a proportionately greater effect on taurine relative to glycine sensitivity (Fig. 1b). Indeed, as taurine-activated currents were very small, it was possible to demonstrate that the mutation disrupted both the agonist and antagonist effects of taurine (Fig. 2). This indicates that Asn102 forms a contact site for taurine regardless of whether it binds as an agonist or an antagonist.
Although A101H also disrupted agonist sensitivity (Vafa et al. 1999), the present study shows that the more conservative A101C mutation had no significant effect on the apparent affinities of glycine or taurine (Fig. 1b). Thus, Ala101 is not an agonist-binding site. As discussed above, the water accessibility of this residue is increased in the glycine-bound state, relative to the unbound and taurine-bound states. As the accessibility is not increased in the taurine-bound state, it is possible that this residue either undergoes a different allosteric conformational change following taurine binding or that the larger taurine molecule physically impedes the access of MTS agents to A101C. The later alternative is consistent with the earlier conclusion that Ala101 lies in a location of restricted surface access, possibly in an agonist binding cleft.
The E103C mutation also strongly reduced the apparent affinities of glycine and taurine (Fig. 1b). As the affinities of both agonists were disrupted to a similar extent, it is possible that Glu103 also contributes to a common glycine and taurine binding site. An alternate possibility is that the E103C mutation disrupts the adjacent binding site by a local structural disturbance. As no other mutations to residues in the Ala101–Thr112 domain resulted in large increases in the EC50 values for glycine or taurine, it is highly unlikely that any other domain residues form agonist-binding sites.
Control of taurine gating by the Lys104–Thr112 domain
When recombinantly expressed in mammalian cell lines, human homomeric α1 GlyRs exhibit a glycine EC50 of about 20 µm, a taurine EC50 of about 120 µm and taurine behaves a full agonist relative to glycine (e.g. Figure 1; Table 3). In contrast, when the same GlyR is recombinantly expressed in Xenopus oocytes, the glycine EC50 is increased to 200 µm, the taurine EC50 is around 2 mm and the ratio of the peak taurine- to glycine-gated currents is about 0.3 (Schmieden et al. 1995). Using the Xenopus oocyte expression system, Schmieden et al. (1995) found that β-amino acids, which rotate spontaneously between cis- and trans- isomers, act as partial agonists, and that molecules which are locked in a trans-like configuration act purely as antagonists. As glycine, which exists in the cis-conformation, acts as a full agonist, it was proposed that β-amino acid partial agonism results from the ligands binding as agonists in the cis- conformation and as antagonists in the trans- conformation. This model necessarily predicts the existence of a trans- conformation antagonist-binding site. Although this model reconciles many observations, it does not provide direct evidence for the existence of such a site.
In a subsequent study using the Xenopus oocyte expression system, Schmieden et al. (1999) demonstrated that the K104A, F108A or T112A mutations dramatically increased the magnitude of currents activated by β-amino acid agonists relative to glycine, while simultaneously causing modest reductions in β-amino acid EC50 values. On the basis of these results, it was proposed that the β-amino acid antagonistic potency was reduced by either a disruption to the β-amino acid inhibitory gating mechanism or a disruption to the putative β-amino acid antagonist-binding site.
The present study found that the K104C mutation caused modest (2–3 fold) increases in the apparent affinities of glycine and taurine (Table 3). These findings are in agreement with those of Schmieden et al. (1999). However, Schmieden et al. also found that the mutation increased the magnitude of taurine-gated currents relative to glycine-gated currents. Such an effect was not seen in the present study, presumably because taurine was already a full agonist of the WT GlyR. Hence, although no evidence for a β-amino acid-specific mechanism was found in the present study, the results do not necessarily refute the conclusion of Schmieden et al. (1999) that Lys104 is a candidate as a putative taurine antagonist-binding or gating element.
The present study also found that the F108C mutation has no effect on the affinity or relative peak current magnitude of taurine, although it did induce a significant (2.5-fold) increase in glycine sensitivity (Table 3). Hence, contrary to its effect on Xenopus oocyte-expressed GlyRs (Schmieden et al. 1999), this residue is unlikely to mediate β-amino acid-specific responses in mammalian cell-expressed GlyRs.
In an early study using Xenopus oocyte-expressed Schmieden et al. (1992) demonstrated that Ile111 is responsible for the difference in the ratio of taurine- to glycine-gated currents between the GlyR α1 and α2 homomers. As the present study indicates that I111C significantly increases both glycine- and taurine-gated current magnitudes, but does not affect the apparent affinity of either agonist (Table 3), Ile111 does not mediate β-amino acid-specific responses in the present system.
The T112C mutation induced a significant (four-fold) increase in taurine sensitivity, but had no effect on glycine sensitivity or on the relative magnitude of taurine- to glycine-gated currents (Table 3). This implies that Thr112 is a taurine-specific gating element. Contrary to the present study, Schmieden et al. (1999) found that the T112A mutation had little effect on the EC50 values of either agonist.
It is relevant to note that Thr112 is involved in mediating the inhibitory effects of other antagonists. Based on the effects of alanine substitution mutations, His107, His109 and Thr112 were identified as important determinants of zinc inhibition (Harvey et al. 1999; Laube et al. 2000). As a histidine-specific agent abolishes zinc inhibition, the zinc binding site probably formed by His107 and His109 (Harvey et al. 1999), with Thr112 more likely to be involved in the zinc inhibitory gating process. The T112A mutation also dramatically decreased the inhibitory potency of the competitive antagonist, strychnine (Schmieden et al. 1999) and induced relatively smaller decreases in the inhibitory potencies of the apparent competitive antagonists, nipecotic acid, isobutyric acid, tropisetron and atropine (Maksay et al. 1999; Schmieden et al. 1999). However, these studies have presented no evidence that Thr112 forms an antagonist-binding site. The results of the present and previous studies suggest that Thr112 may be a common mediator of inhibitory gating signals.