The TM2-TM3 loop of the α10 subunit renders a Ca2+ potentiation phenotype in α9α10 nAChRs
Extracellular Ca2+ ions modulate the activity of several nAChRs [14–17, 71]. As previously reported [18, 47], Fig. 1 shows the differential responses of rat homomeric α9 and heteromeric α9α10 nAChRs expressed in Xenopus laevis oocytes to changes in extracellular Ca2+ concentrations. Thus, as previously reported [47] and depicted with modifications in Fig. 1a for the sake of comparison, responses of homomeric α9 nAChRs to 100 µM ACh have a maximal amplitude in the absence of extracellular Ca2+ (182 ± 20 nA, n = 4 [47]), indicating that channel gating by ACh does not require the presence of this cation. In addition, α9 nAChRs responses are reduced in the presence of increasing (0.05–1.8 mM) Ca2+, being smallest (5.63 ± 0.58%, n = 3) at a 1.8 mM physiological extracellular Ca2+ concentration (Fig. 1a and [47]). At -70 mV the concentration of Ca2+ that produces a 50% of response block (IC50) is 0.10 ± 0.01 mM [47]. Moreover, Ca2+ block of α9 receptors is voltage-dependent, since it is more pronounced at hyperpolarized than at depolarized potentials [47]. On the contrary, when α9 assembles with α10, rendering α9α10 heteromeric nAChRs, a bell-shaped profile of ACh-evoked responses at different Ca2+ concentrations is observed (Fig. 1b and [18]). Thus, responses to 10 µM ACh are small in the presence of nominally zero Ca2+ (16.79 ± 6.47% of responses at 1.8 mM, n = 7), are potentiated by sub-millimolar Ca2+ concentrations reaching a maximum at 0.5 mM (142.98 ± 10.12% of responses at 1.8 mM, n = 10) and then onwards blocked by further increases in extracellular Ca2+ within the millimolar range. Whereas potentiation of α9α10 nAChRs by Ca2+ is voltage independent, blockage is more pronounced at hyperpolarized than at depolarized potentials [18]. Taken together, these results suggest that it is the α10 subunit the one that provides structural determinants that subserve a Ca2+ potentiation phenotype of α9α10 nAChRs.
In order to dissect the domains of the α10 subunit that are critical for the Ca2+ potentiation phenotype of rat α9α10 nAChRs, we engineered multiple α10α9X subunit chimeras constructed by sequentially exchanging amino acidic regions of α10 by their corresponding ones from the α9 nAChR subunit as indicated in Fig. 2. All chimeras were heterologously expressed in X. laevis oocytes, and their function tested by two-electrode voltage-clamp recordings. None of the chimeric subunits elicited any response to ACh when expressed alone or when co-expressed with the α10 subunit, but were functional when co-expressed with rat α9 (Table 1). The amplitude of responses observed for the chimeric receptors at 1.8 mM Ca2+ (Table 2) were ~ 10–30 times higher than those reported for homomeric rat α9 receptors (range of responses from 5 to 20 nA, with some exceptional responses of 50 and 70 nA, [23, 47]), indicating that, when co-expressed, rat α9 and chimeric subunits form heteromeric assemblies. To perform a comparative analysis of Ca2+ modulation on the generated chimeras, responses to ACh for each chimeric receptor were recorded in normal Ringer’s solution at a range of Ca2+ concentrations (nominally zero to 3 mM Ca2+). A near-EC50 ACh concentration (Table 1) was chosen for each chimeric receptor, in order to provide an ample dynamic range for Ca2+ modulation, allowing for both potentiation and block [50]. Since amplitude responses to ACh vary across different experiments and oocyte batches (even under the same conditions), current amplitudes at different Ca2+ concentrations were normalized to the response at 1.8 mM Ca2+ for each individual oocyte.
Table 1
Parameters of the concentration-response curves of WT, chimeric and mutant receptors.
Receptor | EC50 (µM) p | Hill p | n | R2 |
α9α10 | 17 ± 4 | | 1.1 ± 0.2 | | 4 | 0.9406 |
α9α10 α9Nt | 16 ± 1 | 0.5238 | 1.2 ± 0.2 | 0.6825 | 5 | 0.9754 |
α9α10 α9Nt −TM1TM2loop | 46 ± 3 | 0.0159* | 0.70 ± 0.03 | 0.0285* | 4 | 0.9747 |
α9α10 α9Nt −TM2 | 34 ± 9 | 0.3333 | 0.9 ± 0.2 | 0.2571 | 6 | 0.8571 |
α9α10 α9Nt −TM2TM3loop | 4.0 ± 0.5 | 0.0159* | 1.6 ± 0.2 | 0.1111 | 5 | 0.9117 |
α9α10 α9extra | 11 ± 3 | 0.2857 | 0.90 ± 0.07 | 0.2857 | 5 | 0.9545 |
α9α10 α9TM2TM3loop α9α10E71Q | 78 ± 18 104 ± 20 | 0.0943 0.0286* | 0.70 ± 0.05 0.80 ± 0.07 | 0.3727 0.2000 | 7 4 | 0.9378 0.9538 |
α9α10E201Q | 135 ± 23 | 0.0159* | 1.10 ± 0.14 | > 0.9999 | 5 | 0.9467 |
α9α10E287Q | 20 ± 4 | 0.5333 | 0.97 ± 0.08 | 0.8571 | 3 | 0.9631 |
The EC50 values are shown as mean ± S.E.M. The p value reported corresponds to a Mann-Whitney test when comparing to the value obtained for the α9α10 receptor (*p < 0.05). |
Table 2
Maximum Currents (Imax) evoked by 1000 µM ACh at 1.8 mM Ca2+.
Receptor | I max (nA) p | n |
α9α10 | 195 ± 42 0.0005*** | 12 |
α9α10 α9Nt | 271 ± 146 0.0020** | 10 |
α9α10 α9Nt −TM1TM2loop | 139 ± 37 0.0095** | 6 |
α9α10 α9Nt −TM2 | 410 ± 119 0.0020** | 10 |
α9α10 α9Nt −TM2TM3loop | 303 ± 101 0.0040** | 8 |
α9α10α9 extra | 246 ± 56 0.0159* | 5 |
α9α10 α9TM2TM3loop | 211 ± 55 0.0061** | 7 |
α9α10E71Q | 122 ± 55 0.0159* | 5 |
α9α10E201Q | 53 ± 16 0.0240* | 10 |
α9α10E287Q | 256 ± 50 0.0159* | 5 |
Imax values are shown as mean ± S.E.M. The p values reported correspond to Mann-Whitney test when comparing to the value for the α9 receptor (*p < 0.05, **p < 0.01 and ***p < 0.001). |
Figure 3 shows the results obtained in α9α10α9x receptors in which the α10 chimeric subunits were engineered by the exchange of either the extracellular N terminal domain (α10α9Nt, left panel) or the extracellular N terminus, the TM1 and TM1-TM2 loop (α10α9Nt–TM1TM2 loop, right panel) of α10 by those of the α9 nAChR subunit. Similar to that described for α9α10 nAChRs (Fig. 1 and [18]), Ca2+ exerted a dual effect and responses to a near-EC50 ACh concentration (10 µM and 40 µM, respectively) were potentiated by Ca2+ up to 0.2 mM and blocked by higher concentrations of this ion. Thus, ACh-evoked currents at 0.2 mM Ca2+ were significantly higher than those observed at nominally zero Ca2+: α9α10α9Nt, 115.2 ± 27.3% and 225.8 ± 18.9%, n = 6, (p = 0.0373); α9α10α9Nt–TM1TM2 loop, 95.2 ± 18.8% and 358.3 ± 32.4%, n = 4–6, (p = 0.0022), (Kruskal-Wallis followed by Dunn’s multiple comparison-test) in 0 and 0.2 mM Ca2+, respectively. On the other hand, responses to 3 mM were significantly smaller than those observed at 0.2 mM Ca2+: α9α10α9Nt, 225.8 ± 18.9% and 64.9 ± 5.9%, n = 6, (p = 0.0010); α9α10α9Nt–loopTM1TM2, 358.3 ± 32.4% and 73.5 ± 9.3%, n = 4–6, (p = 0.0011), (Kruskal-Wallis followed by Dunn’s multiple comparison-test) in 0.2 and 3 mM Ca2+, respectively. As described in Fig. 1 and [18] for wild-type α9α10 nAChRs, potentiation of chimeric receptors by Ca2+ was voltage independent (Fig. 3c). Thus, the 0.2 / 0 Ca2+ ratios of response amplitudes were similar (α9α10α9Nt, 2.1 ± 0.4 and 2.5 ± 0.3, n = 5 (p = 0.076, paired t test); α9α10α9Nt–TM1TM2 loop, 5.4 ± 2.4 and 5.9 ± 2.6, n = 5 (p = 0.223, paired t test), at -90 and + 40 mV, respectively.
We subsequently extended the α10 for α9 subunit amino acid replaced region by adding the TM2 of α9 to α10α9Nt–TM1TM2 loop, rendering the α10α9Nt–TM2 chimeric subunit and co-expressed it with α9 (Fig. 4A, left panel). Similar to that observed in α9α10 (Fig. 1 and [18]) and in α9α10α9Nt and α9α10α9Nt–TM1TM2loop (Fig. 3) chimeric receptors, Ca2+ exerted a dual effect on α9α10α9Nt–TM2 and responses to 30 µM ACh were potentiated by Ca2+ up to 0.2 mM and blocked by higher concentrations of this ion. Thus, ACh-evoked currents at 0.2 mM Ca2+ (322.0 ± 40.0%) were significantly higher than those observed at nominally zero Ca2+ (132.4 ± 27.5%, n = 5–8, p = 0.039), (Fig. 4b, left panel). On the contrary, responses to 3 mM (69.5 ± 7.6%) were significantly smaller than those observed at 0.2 mM Ca2+ (322.0 ± 40.0%, n = 5–8, p = 0.0003, Kruskal-Wallis followed by Dunn’s multiple comparison-test). Moreover, potentiation was once again voltage-independent (Fig. 4c, left panel). Thus, the 0.2 / 0 Ca2+ ratio of response amplitudes was similar at -90 mV (4.1 ± 0.2) and + 40 mV (2.8 ± 0.4), (p = 0.1424, n = 5, paired t test). Taken together, experiments in Fig. 3 and Fig. 4 (left panel) indicate that the replacement of the N terminal region until the entire TM2 of α10 by the corresponding amino acid regions of α9, do not recover a Ca2+ modulation phenotype similar to that observed in α9 homomeric receptors. Thus, Ca2+ potentiation was still observed at sub-millimolar concentrations and was voltage-independent. However, it should be noted that different to α9α10 wild-type receptors, where responses at zero Ca2+ were only 16.79 ± 6.47% of responses at 1.8 mM (n = 7), chimeric receptors exhibited larger current amplitudes under similar conditions: 115.2 ± 27.3% (n = 6) p = 0.014, 95.2 ± 18.8% (n = 6), p = 0.0047 and 132.4 ± 27.5% (n = 6) p = 0.0012, for α10α9Nt, α10α9Nt–TM1TM2loop and α10α9Nt–TM2, Mann-Whitney test. Moreover, maximal responses were obtained at a lower 0.2 mM Ca2+ concentration, compared to 0.5 mM in the case of α9α10 wild-type receptors.
As shown in Fig. 4a (right panel) the α10α9Nt–TM2TM3loop chimeric subunit was constructed by introducing the TM2-TM3 loop of α9 in replacement of the corresponding region of α10α9Nt–TM2 and co-expressed with α9 in Xenopus oocytes. Different from that observed for α9α10 wild-type receptors and the above chimeras, responses were not potentiated by Ca2+, indicating that ACh responses became independent of the presence of extracellular Ca2+. Similar to homomeric α9 receptors, responses to 5 µM ACh of α9α10α9Nt–TM2TM3loop were only blocked by increasing Ca2+ concentrations (Fig. 4b, right panel). Thus, responses at zero Ca2+ (611.3 ± 91.2%) were significantly higher than those at 3 mM Ca2+ (108.1 ± 10.7%), n = 5–8 (p = 0.0020, Kruskal-Wallis followed by Dunn’s multiple comparison-test). As for α9 receptors [47], block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.20 ± 0.03 and 1.2 ± 0.3 (p = 0.0323, n = 7, paired t test), at -90 and + 40 mV, respectively. Thus, whereas the allosteric potentiation by Ca2+ is lost in α9α10α9Nt–TM2TM3loop nAChRs, its blocking effect is still maintained. It should be noted that the Ca2+ inhibition curve was slightly shifted to the right in the α9α10α9Nt–TM2TM3loop chimeric receptor, presenting a higher IC50 (0.5 ± 0.1 mM, n = 6), that differed significantly from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, from [47], Mann-Whitney test, p = 0.0238). Taken together, these results suggest that amino acids located within the α10 TM2-TM3 loop are required for potentiation of α9α10 receptors by external Ca2+. However, the results so far do not preclude the possibility that the loss of Ca2+ potentiation, is due to the interchange of the α10 TM2-TM3 loop by the corresponding one from α9, within the overall amino acid context of a α9 N terminal-TM2 domain, which is also present in the α10α9Nt–TM2TM3loop chimeric subunit. Thus, it has been described that the consequences of a given amino acid substitution greatly depends upon the overall sequence (and hence structure) of the protein, a phenomenon that can be referred to as epistasis [72].
In order to dissect the role of the α10 TM2-TM3 loop in Ca2+ potentiation, chimeric α10α9x subunits which included reduced α9 domains were engineered. The fact that potentiation by Ca2+ in all chimeric receptors was always voltage-independent, indicates that Ca2+ modulation is not subject to the electric field of the membrane. Therefore, as suggested previously for other nAChRs [8], including α9α10 [18], it most likely depends on an extracellular binding site that allosterically modulates coupling between ligand binding and gating, rather than on transmembrane regions. Consequently, we constructed the chimeric α10α9extra subunit, where only the extracellular domains of α10 (the Nt, preTM1 and TM2-TM3 loop) involved in coupling agonist binding to channel gating [73] were replaced by the corresponding ones from α9 (Fig. 5a, left panel). Similar to that described for murine α9 nAChRs (Fig. 1 and [47]) and the α9α10α9Nt–TM2TM3 loop chimeric receptor (Fig. 4b, right panel), responses to 10 µM ACh of the α9α10α9extra were not potentiated by Ca2+, and only blocked by increasing Ca2+ concentrations (Fig. 5b, left panel). Thus, responses were significantly higher at zero Ca2+ (217.3 ± 6.3%) than at 3 mM Ca2+ (72.3 ± 5.3%, n = 7, p = 0.020, Kruskal-Wallis followed by Dunn’s multiple comparison-test). It should be noted that the Ca2+ inhibition curve was slightly shifted to the right in α9α10α9extra with a Ca2+ IC50 (0.7 ± 0.2 mM, n = 6) that significantly differed (p = 0.0238, Mann-Whitney test,) from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, from [47]). Moreover, as for α9 receptors [47] and α9α10α9Nt–TM2TM3loop (Fig. 5c, left panel), block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.20 ± 0.05 and 0.60 ± 0.08 (p = 0.0016, n = 4, paired t -test), at -90 and + 40 mV, respectively. Taken together, these results suggest that amino acids located within the extracellular domains of α10 (the Nt, preTM1 and TM2-TM3 loop) are required to render a potentiation phenotype of α9α10 receptors by external Ca2+.
In order to provide further insights into the role of the TM2-TM3 loop in Ca2+ potentiation of α9α10 nAChRs, we generated the α10α9TM2TM3loop chimeric subunit, in which only the TM2-TM3 loop of α10 was replaced by the corresponding region of α9 (Fig. 5a, right panel). Figure 5b, right panel, shows that similar to α9α10α9Nt–TM2TM3loop nAChRs, α9α10α9TM2TM3loop receptors were not potentiated by extracellular Ca2+, but instead blocked by this ion. Thus, responses were significantly higher at zero Ca2+ (401.0 ± 103.4%) than at 3 mM Ca2+ (71.3 ± 5.2%, n = 6, p = 0.0008, Kruskal-Wallis test followed by Dunn’s multiple comparison-test). It should be noted that the Ca2+ inhibition curve was slightly shifted to the right in α9α10α9TM2TM3loop with a Ca2+ IC50 (0.8 ± 0.3 mM, n = 5), that differed significantly from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, p = 0.0357, Mann-Whitney test). Moreover, similar to α9 [47] and α9α10α9Nt–TM2TM3loop (Fig. 5c, left panel) receptors, block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.14 ± 0.02 and 0.74 ± 0.14 (p = 0.0067, n = 6, paired t -test), at -90 and + 40 mV, respectively. Since potentiation by Ca2+ is more pronounced at low agonist concentrations ([18] and Fig. 2), modulation by this ion was further assessed in α9α10α9TM2TM3loop at a sub-EC50 ACh concentration (25 µM) (Fig. 6). Once again, responses were not potentiated by Ca2+, and only block was observed (Fig. 6a), indicating that channel gating by ACh is independent of the presence of extracellular Ca2+, even at a sub-EC50 ACh concentration. Thus, responses at zero Ca2+ (668.7 ± 340.2%) were significantly higher than those at 3 mM Ca2+ (46.3 ± 20.9%, n = 5–6, p = 0.0412, Kruskal-Wallis followed by Dunn’s multiple comparison-test). It should be noted that responses of α9α10α9TM2TM3loop are the result of the co-assembly of the chimera with α9 and not that of α9 homomeric receptors, since the EC50 of homomeric α9 (11.4 ± 0.8 µM, [46]) differs from that of α9α10α9TM2TM3loop (78 ± 18 µM, p = 0.0106, Mann-Whitney test) receptors. Moreover, maximal responses to ACh at 1.8 mM α9α10α9TM2TM3loop (211 ± 55 nA, n = 7, p = 0.0061, Table 2) were significantly higher than those reported for α9 homomers (range of responses from 5 to 20 nA, with some exceptional responses of 50 and 70 nA [23, 47]). In addition, whereas α9 homomeric receptors have a slow decay rate of ACh evoked currents upon a prolonged application of a high ACh concentration (94.9 ± 1.6% of remaining current after 20 s of the peak response to ACh [23], α9α10α9TM2TM3loop exhibited a significant decay rate (61.9 ± 4.6% of remaining current after 20 s of the peak response to ACh, p < 0.0001, One-Way ANOVA followed by Dunnett´s multiple comparison test), similar to that observed for α9α10 receptors (57.1 ± 4.3% of remaining current after 20 s of the peak response to ACh, from [23], p ≤ 0.0001, one-way ANOVA followed by Dunnett´s multiple comparison test). The reverse chimera, α9α10TM2TM3loop, in which the TM2-TM3 loop of α9 was replaced by the corresponding region of α10 did not render functional receptors even at a high 1 mM ACh concentration, either when injected alone or co-injected with α9 subunits.
Taken together, the results so far strongly support the hypothesis that the TM2-TM3 loop of α10 contains structural determinants that are critical to render the Ca2+ potentiation phenotype of α9α10 nAChRs.
Molecular dynamics simulations
The data derived from chimeric receptors revealed that the TM2-TM3 loop of α10 is a key player in the Ca2+ potentiation phenotype of the α9α10 nAChR. This finding prompts the hypotheses that Ca2+ potentiation of α9α10 nAChRs arises as a consequence of the differential binding of this cation to the environment of the TM2-TM3 loops of α9 and α10 subunits, perturbing the energy landscape of the closed to open state transition.
To test this hypothesis, we performed MD simulations. First, we evaluated Ca2+ interaction sites along the structure of α9 and α9α10 nAChRs. Given the lack of any experimental structure of these nAChRs, we used AlphaFold to build their theoretical structural models, CHARMM-gui to embed receptors in a POPC membrane and GROMACS for MD simulations, as described in Methods. With this setup, a series of simulations of the different receptors were performed, in the presence and absence of Ca2+, using a newly developed multisite calcium model [70]
The MD simulation for the α9 pentamer revealed two Ca2+ binding sites in the environment of the TM2-TM3 loops (Fig. 7a and d). Each site presented the cation lodged at an interface between subunits and coordinated by anionic side chains. In site 1, Ca2+ was coordinated by E288, from the TM2-TM3 loop, at both the (+) and (-) faces. Although there are five potential Ca2+ binding sites 1, only two Ca2+ ions were simultaneously observed alternating between these sites. Ca2+ binding site 2 occurred in the exterior side of the interchain interface, where Ca2+ was coordinated by E294, from the TM2-TM3 loop, at the (+) face, and E204 and E199, from loop 9, and R236 of the pre TM1 domain, at the (–) face. In addition, a third Ca2+ binding site coordinated by residues D69, D71, E72, located in the β1–β2 loop, and E202, from the β8–β9 loop, from the same subunit, was revealed (Fig. 7i). This latter finding is in line with previous studies in α7 nAChRs [49, 62] that have proposed the existence of a binding site for Ca2+ that involves residues E44 (E72 rat α9 numbering) and E172 (E202 rat α9 numbering). Figure 8 shows a global view of the location of the three Ca2+ sites in relation to the ECD and the TM.
The MD simulations for α9α10 and α9α10α9TM2TM3loop receptors revealed Ca2+ binding coordinated by contiguous α9 E288 and α10 E287 side chains in all subunit interfaces, in direct homology to the Ca2+ binding site 1 observed in α9 nAChRs (Fig. 7b and c). On the other hand, Ca2+ binding at site 2 in α9α10 and α9α10α9TM2TM3loop receptors was only observed at the α10(+)/α9(-) interfaces (Fig. 7e and f, respectively), coordinated by α10 E293 and α9 E199, α9 E204 and α9 R236 residues. No Ca2+ binding was observed in the region homologous to site 2 in the α9(+)/α10(-) (Fig. 7g and h, respectively) and α10(+)/α10(-) interfaces, where the negative charge of residue α9 E204 is replaced by a positive charge of residue R203 when α10 is in the (-) face. A third Ca2+ binding site, homologous to the Ca2+ binding site 3, was also observed in α9α10 and α9α10α9TM2TM3loop receptors, coordinated by residues α9D69, α9D71, α9 E72 and α9 E202, and α10D68, α10D70, α10 E71 and α10E201 (Fig. 7j-m).
To gain further insight into the interaction of Ca2+ with binding sites 1, 2 and 3, we calculated the Ca2+ occupancy in these sites for the α9, α9α10 and α9α10α9TM2TM3loop receptors. Calcium occupancy was calculated as the fraction of time this ion is observed within a radius of 6 Å of specific residues that underlie the binding sites, during the simulations. To measure Ca2+ occupancy for site 1, α9E288 and α10E287 were evaluated. For site 2, α9 E294 or α10 E293, and α9 E204, E199 and R236 were analyzed. For site 3, α9D69, α9D71, α9 E72, α9 E202, α10D68, α10D70, α10 E71 and α10E201 were analyzed. Figure 9a and b show that the averaged Ca2+ occupancy of sites 1 and 2 were similar for the α9, α9α10 and α9α10α9TM2TM3loop nAChRs (Fig. 9a: α9, 0.65 ± 0.12; α9α10, 0.55 ± 0.11; α9α10α9TM2TM3 loop, 0.60 ± 0.13, n = 4; Fig. 9b: α9, 0.78 ± 0.01; α9α10, 0.86 ± 0.07; α9α10α9TM2TM3 loop, 0.78 ± 0.05, n = 4; Kruskal Wallis followed by Dunn’s multiple comparison-test, α9 vs α9α10 = α9 vs α9α10α9TM2TM3 loop = α9α10 vs α9α10α9TM2TM3 loop, p > 0.9999). In the case of site 3, the average Ca2+ occupancy of α9 receptors was higher than those of α9α10 and α9α10α9TM2TM3loop nAChRs, but similar for the α9α10 and α9α10α9TM2TM3loop receptors (Fig. 9c: α9, 0.89 ± 0.01; α9α10, 0.45 ± 0.07; α9α10α9TM2TM3 loop, 0.46 ± 0.06, n = 3–4 Kruskal Wallis followed by Dunn’s multiple comparison-test, α9 vs α9α10, p = 0.0298 α9 vs α9α10α9TM2TM3 loop, p = 0.0660 α9α10 vs α9α10α9TM2TM3 loop, p > 0.9999).
In summary, the MD simulations revealed that Ca2+ is coordinated by the same key anionic residues and with similar Ca2+ occupancy in the environment of the TM2-TM3 loops of α9, α9α10 and α9α10α9TM2TM3loop receptors. This result suggests that the TM2-TM3 loop of α10 does not contribute to the Ca2+ potentiation phenotype of the α9α10 nAChR through the formation of novel Ca2+ binding sites that are not present in the α9 homomeric receptor.
To test the predictions derived from the MD simulations we generated E72Q and E202Q (site 3) mutants in rat α9, and E71Q, E201Q (site 3) and E287Q (site 1) in rat α10 subunits, and determined their contribution to Ca2+ potentiation of α9α10 nAChRs. Each mutant subunit was co-expressed with its wild-type counterpart in Xenopus oocytes. Notably, α9E72Qα10 and α9E202Qα10 mutant receptors failed to respond to 1 mM ACh (n = 6). On the contrary, α9α10E71Q, α9α10E201Q and α9α10E287Q complexes formed functional channels (Table 1). As reported in Table 1, both α10E71Q and α10E201Q substitutions produced a shift of the ACh concentration-response curve to the right and an increase in the ACh EC50 (104 ± 20 µM and 135 ± 23 µM, respectively). The significant increase in the EC50 of mutant receptors compared to wild-type α9 (p = 0.0013 and 0.0007, student t test, Table 3) and α9α10 (p = 0.0286 and 0.0159, Mann-Whitney test, Table 1) receptors indicate that both α9α10E71Q and α9α10E201Q receptors are assembly competent and that responses do not derive from homomeric α9 wild-type receptors. In the case of α10E201Q receptors, the significant increase in the Imax compared to wild-type α9 (p = 0.0159, Mann-Whitney test, Table 2) receptors indicate that α9α10E287Q receptors are assembly competent and that responses do not derive from homomeric α9 wild-type receptors. Figure 10 show the modulation profile of α9α10E71Q, α9α10E201Q and α9α10E287Q receptors obtained at a near EC50 concentration of ACh (100 µM for α9α10E71Q and α9α10E201Q, and 20 µM for α9α10E287Q) and increasing concentrations of extracellular Ca2+. Different to that observed for α9α10 wild-type receptors, responses to ACh of α9α10E71Q, α9α10E201Q and α9α10E287Q receptors were not potentiated by Ca2+, indicating that channel gating by ACh became independent of the presence of extracellular Ca2+. Similar to homomeric α9 receptors, responses to ACh of α9α10E71Q, α9α10E201Q and α9α10E287Q receptors were only blocked by increasing Ca2+ concentrations (Fig. 10d-f). Thus, α9α10E71Q responses were highest at zero Ca2+ (211.0 ± 35.6%) and lowest at 3 mM Ca2+ (64.8 ± 2.7%, n = 4, p = 0.0069, Kruskal-Wallis followed by Dunn’s multiple comparison-test). It should be noted that the inhibition curve was shifted to the right with a Ca2+ IC50 (0.8 ± 0.1 mM, n = 6) that differed significantly (Mann-Whitney test, p = 0.0238) from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, from [47]). In addition, like α9 nAChRs [47], block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.25 ± 0.05 and 1.05 ± 0.29 (Fig. 10g, p = 0.0302, n = 5, paired t -test), at -90 and + 40 mV, respectively. Similarly, responses to ACh of α9α10E201Q were highest at zero Ca2+ (201.6 ± 39.7%) and lowest at 3 mM Ca2+ (74.3 ± 11.2%, n = 7, p = 0.0071, Kruskal-Wallis followed by Dunn’s multiple comparison-test), with a Ca2+ IC50 (0.9 ± 0.3, n = 6) that differed significantly (Mann-Whitney test, p = 0.0238) from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, from [47]). Moreover, as for α9 receptors [47], block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.18 ± 0.05 and 0.53 ± 0.12 (Fig. 9h, p = 0.0358, n = 6, paired t-test), at -90 and + 40 mV, respectively. Similarly, α9α10E287Q responses were highest at zero Ca2+ (320.7 ± 97.4%) and lowest at 3 mM Ca2+ (67.6 ± 5.6%, n = 5, p = 0.0010, Kruskal-Wallis followed by Dunn’s multiple comparison-test) with a Ca2+ IC50 (0.3 ± 0.1, n = 4) that did not differ significantly (Mann-Whitney test, p = 0.400 ) from that of homomeric α9 receptors (0.10 ± 0.01 mM, n = 3, from [47]). Like α9 nAChRs [47], block of ACh evoked responses by Ca2+ was voltage-dependent and more pronounced at hyperpolarized than at depolarized potentials, exhibiting a 3 / 0 Ca2+ ratio of response of 0.26 ± 0.02 and 0.48 ± 0.04 (Fig. 10i, p = 0.0031, n = 10, paired t -test), at -90 and + 40 mV, respectively. Taken together, these results suggest that α10E71Q, α10E201Q and α10E287 are required for potentiation of α9α10 receptors by external Ca2+ and are compatible with those obtained in MD simulations, indicating that α10E71Q, α10E201Q and α10E287 form part of Ca2+ binding sites.
Table 3
EC50 values of WT, chimeric and mutant receptors.
Receptor | EC50 (µM) p |
α9 | 11.4 ± 0.8 | |
α9α10 | 17 ± 4 | 0.176 |
α9α10α9Nt | 16 ± 1 | 0.016* |
α9α10 α9Nt −TM1TM2loop | 45 ± 3 | <0.0001**** |
α9α10 α9Nt −TM2 | 34 ± 9 | 0.0699 |
α9α10 α9Nt −TM2TM3loop | 4.0 ± 0.5 | <0.0001**** |
α9α10 α9extra | 11 ± 3 | 0.9282 |
α9α10 α9TM2TM3loop | 78 ± 18 | 0.0106* |
α9α10E71Q | 104 ± 20 | 0.0013** |
α9α10E201Q | 135 ± 23 | 0.007*** |
α9α10E287Q | 20 ± 4 | 0.0244* |
The EC50 values are shown as mean ± S.E.M. The p- value reported corresponds to t-student test comparing against the value obtained for the corresponding α9 receptor from Verbitsky et. al 2000 [46]. (*p < 0.05 **p < 0.01,***p < 0.001 and ****p < 0.0001). |
Online resource 1 |
Previous work applied MD simulations to a model of the full-length α7 nAChR and evidenced conformational changes in the TM2-TM3 loops, in the presence of Ca2+ [21]. Since conformational changes in the TM2-TM3 loops might have implications for their role in channel gating, we examined the potential impact of Ca2+ on the structure and motions of the TM2-TM3 loops of α9, α9α10 and α9α10α9TM2TM3 loop receptors- For that purpose we measured the root-mean-square-fluctuation (RMSF) profile of the TM2-TM3 loops of α9 and α10 subunits from α9, α9α10 and α9α10α9TM2TM3 loop receptors, in the presence and absence of Ca2+ (Fig. 11). No changes in the flexibility of the TM2-TM3 loops of α9 (Fig. 11a), α9α10 (Fig. 11b and c) and α9α10α9TM2TM3 loop (Fig. 11d and e) receptors, were observed in the presence of Ca2+. These results suggest that changes in structural flexibility of the α9 and α10 TM2-TM3 loops most likely do not account for the differential calcium potentiation phenotype between homomeric α9 and α9α10 heteromeric nAChRs.
The electrophysiology results show that α10E71 and α10E201 are required for potentiation of α9α10 receptors by external Ca2+. MD simulations of the α9α10 pentamer revealed Ca2+ binding sites in the environment of α10E71, α10E201, α9E72 and α9E202. Previous work has shown that in α7 receptors these residues establish electrostatic interactions with the invariant residue R209 [62, 74–76], a key component of the pathway linking agonist binding to channel gating [77, 78]. Similarly, our MD data show that α9E72 and α9E202 establish electrostatic interactions with the conserved residue α9R235, and α10E71 and α10E201 with α10R234 (Fig. 12). This result might suggest that these glutamates contribute to coupling of Ca2+ binding to potentiation in addition of being part of a Ca2+ binding site.