**2. Results**

*2.1. TOA+ Binds to Wild-Type KcsA with High Affinity and Diminishes the Thermal Stability of the Channel*

An assay based on the thermal denaturation of the KcsA protein was previously developed to study the binding of different channel ligands to KcsA, including permeant or nonpermeant cations, membrane lipids, and others [35,37–40]. Figure 1A,B shows examples of thermal denaturation curves to illustrate that the addition of TOA+ at micromolar

concentrations to the detergent-solubilized channel leads to a concentration-dependent decrease in the thermal stability of the wild-type protein, both at pH 7.0 and pH 4.0. Because of the destabilizing effects of TOA+ on KcsA, these experiments were always conducted at a constant 10-mM concentration of Na+ to ensure that the tetrameric channel does not dissociate into monomers at the different TOA+ concentrations tested (Figure 1A,B insets).

**Figure 1.** Effects of tetraoctylammonium (TOA+) on the thermal denaturation of wild-type (WT) KcsA. The intrinsic fluorescence at pH 7.0 (panel (**A**)) and pH 4.0 (panel (**B**)) and in the absence of TOA+ (full and empty circles, respectively) or upon the addition of 1 μM (full and empty squares, respectively) or 100 μM (full and empty triangles, respectively) of TOA+. The insets show SDS-PAGE analysis of the different samples to illustrate the integrity of tetrameric WT KcsA at room temperature. Panel (**C**) shows the dependence of the midpoint temperature of the protein denaturation process (*t*m) of WT KcsA with increasing concentrations of TOA+, either at pH 7.0 (•) or pH 4.0 (-). The results are the average *t*m (in Celsius) ± S.D. from three independent titrations. Panel (**D**) illustrates the fitting of Equation (1) to the experimental data from Panel (**C**) (see Methods). The apparent dissociation constants for the TOA+-channel complexes and their 95% confidence intervals estimated at pH 7.0 and pH 4.0 were 1.8 (0.54–6.02) × 10−<sup>7</sup> M and 3.6 (1.89–6.98) × 10−<sup>8</sup> M, respectively. The observed differences between such dissociation constant (KD) values were not statistically significant.

The midpoint temperatures (*<sup>t</sup>*ms) from the thermal denaturation curves at different TOA+ concentrations were used to build titration binding curves, such as those shown in Figure 1C. The fitting of such curves to a simple two-state binding equilibrium (see Methods) allows us to estimate the apparent dissociation constants (KDs) for the binding of TOA+ to WT KcsA, which were in the 10−<sup>7</sup> and 10−<sup>8</sup> M range at pH 7.0 and pH 4.0, respectively (Figure 1D). These indicate that TOA+ has a very high affinity for binding to the channel under either experimental condition. This is in contrast with previous observations on the binding to KcsA of tetrabutylammonium (TBA+), a shorter acyl chain QA, for which the binding affinity at pH 7.0 (KD ~5 × 10−<sup>9</sup> M) is similar to that reported here for TOA+ but drops five orders of magnitude at pH 4.0 (KD~3.5 × 10−<sup>4</sup> M) [35]. Such a dramatic difference in binding affinities could be rationalized based on the crystallographic information on the binding site for these compounds in the channel protein [12,32,33,41] As indicated in our introduction, the channel-bound QAs are further stabilized at the cavity through the interaction of their alkyl chains with the hydrophobic channel wall. In fact, the long alkyl chains of TOA+ completely traverse the channel protein wall [33], so that hydrophobic interactions become particularly important to stabilize the TOA+–KcsA complex.

#### *2.2. The Presence of TOA+ in the Cavity Diminishes the Affinity of WT KcsA to Bind K+ in the Closed Channel State*

The availability of cation binding sites in the SF of the TOA+–KcsA complex has been explored by thermal denaturation experiments of such complexes in the presence of increasing concentrations of either permeant K+ or nonpermeant Na+. In these experiments, an excess TOA+ concentration of 100 μM, identical to that used in the crystallographic studies referenced above, was maintained constant throughout the titrations with the cations. Figure 2A,C shows representative binding curves for Na+, at both pH 7.0 and pH 4.0, respectively. Binding curves for Na+ to WT KcsA in the absence of TOA+ [37] are also included in all panels to facilitate comparison.

In these conditions, it is known that Na+ binds to a single set of sites provided by the S1 and S4 crystallographic sites in a nonconductive pore conformation [23] (PDB 2ITC). This binding process has a KD in the millimolar range, slightly higher at pH 4.0 than at pH 7.0 (Table 1), which should correspond to the overall KD for Na+ binding to the alluded S1 and S4 sites. In the TOA+–KcsA complex, access to the S4 site is blocked by the presence of TOA+; however, there is still Na+ binding to the available S1 site (Figure 2A,C), although the extent of thermal stabilization (on top of the thermal destabilization caused by TOA+) is lower and its KD decreases an order of magnitude compared to the samples in the absence of TOA+ (Figure 2B,D and Table 1). These observations on Na+ binding to the available S1 site in the QA–KcsA complex are quite similar at pH 7.0 and pH 4.0 and analogous to the results previously seen in the presence of TBA+ [35,39].



**Figure 2.** Effect of TOA+ on Na+ binding to WT KcsA. Panels (**A**,**C**) illustrate Na+ binding to the WT KcsA channel at pH 7.0 (panel (**A**)) or pH 4.0 (panel (**C**)), in the absence (circles) and presence (squares) of 100 μM TOA+, monitored through the Na+-concentration-dependence of the midpoint temperature of the protein denaturation process (*t*m). Each experimental point is the average *t*m (in Celsius) ± S.D. (*n* = 3). Panels (**B**,**D**) show the fitting of Equation (1) to the experimental data from panels (**A**,**C**), respectively (see Methods). The apparent KD values estimated for the above binding events are given in Table 1.

Experiments similar to the above were also conducted at increasing K+ concentrations instead of Na+. As reported previously in the absence of QAs [37], K+ binds to two different sets of binding sites in WT KcsA at pH 7.0 (Figure 3A), which is consistent with crystallographic evidence on the ability of permeant cations to induce concentration-dependent transitions between nonconductive and conductive conformations of the SF [12,13]. The first set of such sites, assigned to the crystal S1 and S4 sites, shows a high affinity for K+ (micromolar KD), thus securing displacement of potentially competing nonpermeant cations. The second set of sites results from the contribution of all S1 to S4 crystallographic sites, is available only to permeant cations when the filter is in the conductive conformation and shows low affinity (millimolar KD) to favor cation dissociation and permeation. Figure 3B,C also shows that in the presence of a saturating concentration of TOA+, two different sets of sites are still available for K+ binding in the TOA+–KcsA complex, which, therefore, is clearly noncollapsed. This indicates that the TOA+-bound channel retains the ability to undergo K+ concentration-dependent transitions between different conformations of its SF. Despite such similarity, it is observed that the binding curve, mainly in the lower K+ concentration range, shows a lower slope than in the absence of TOA+, indicating a loss in K+ binding affinity. Indeed, Table 1 shows that the KD values for the two K+ binding events in the TOA+–KcsA complex are increased with respect to those in the absence of the QA. This is particularly noticeable in the first, high affinity K+ binding event in which the KD values differed by approximately two orders of magnitude. It should be noted that in the presence of TOA+, the entrance to the channel's SF through the S4 site is permanently blocked by the bound TOA+. In the first K+ binding event, which takes place in the low

K+ concentration range, the SF is in a nonconductive state, collapsed at the S2/S3 sites. This, along with the TOA+ blockade of the S4 site, indicates that the decreased affinity for K+ under the TOA+ blockade should be attributed to the binding of K+ to the S1 site in the nonconductive channel state. Nonetheless, as the K+ concentration increases, a second K+ binding event takes place, causing the filter to undergo a conformational transition in which the permeant cation reaches internal binding sites within the pore to provide the characteristic increase in thermal stability to the protein. Figure 3D shows the results from experiments similar to those described in the previous paragraph but now conducted at pH 4.0 to induce channel inactivation. These experiments are further complicated because pH 4.0 causes an additional thermal stabilization of the channel protein compared to pH 7.0 [35]. Still, as described above for the resting channel at pH 7.0, both in the WT KcsA alone or when complexed to TOA+, the two sets of K+ binding sites are also detected in the K+ titrations of all samples at pH 4.0, when the channel is inactivated (Figure 3E,F). This was previously reported as a common feature in three different models of inactivated KcsA channels [35]. As expected for an inactivated state, all WT KcsA samples at pH 4.0 show a decrease in the affinity for K+. Indeed, Table 1 shows that when compared to the pH 7.0 samples, the KDs for K+ binding at pH 4.0 increase approximately three and two orders of magnitude, respectively, for the first and second K+ binding events. Such effects on the binding affinity caused by pH-induced channel inactivation are comparable to those caused by the presence of TOA+ bound to the resting state of the channel at pH 7.0. Interestingly, in contrast to TOA+, the shorter chain-length TBA+ does not critically change the interaction between the SF and the permeant cations [32,33,39].

**Figure 3.** Effect of TOA+ on K+ binding to WT KcsA. Panels (**A**,**D**) illustrate K+ binding to the WT KcsA channel in the absence (circles) and presence of 100 μM TOA+ (squares) at pH 7.0 (panel (**A**)) or pH 4.0 (panel (**D**)), monitored through the K+-concentration-dependence of the midpoint temperature of the protein denaturation process (*t*m). Each experimental point is the average *t*m (in Celsius) ± S.D. (*n* = 3). When the low (panels (**B**,**E**)) and high (panels (**C**,**F**)) K+ concentration ranges are analyzed separately, Equation (1) adequately fits the data from panels (**A**,**D**), indicating that two different sets of K+ binding sites are present in the WT KcsA. The apparent KD values estimated for the above binding events are given in Table 1.

#### *2.3. The Homo-FRET Process in the W67 KcsA Mutant Reports Differences between the Nonconductive, Conductive, and Inactivated Conformations of the Selectivity Filter*

We recently reported an analytical framework to analyze the homo-Förster resonance energy transfer (homo-FRET) within a quadruple mutant W26, 68, 87, 113F KcsA channel bearing a single tryptophan residue (W67) per subunit (hereby called W67 KcsA) to characterize the interplay between the pore helix conformation and the cation occupancy at the SF [34] (Figure 4A). The ion channel activity of such W67 KcsA mutant was previously shown to be very similar to that of the WT KcsA channel [34].

The time-resolved anisotropy decays of this mutant protein at pH 7.0 (Figure 4B) show that the rate of the homo-FRET process (*k*1) increases in the presence of increasing K+ concentrations and, consequently, the steady-state anisotropy (<r>SS) and the W67–W67 intersubunit distance decrease along with it, further defining the characteristic two consecutive K+ binding events in the mutant channel (Figure 4C,D).

**Figure 4.** Influence of pH and K+ concentration on the homo-Förster resonance energy transfer (homo-FRET) process among the tryptophan residues in the W67 KcsA channel. Panel (**A**) shows a schematic top view of the tetrameric structure of KcsA (PDB: 1K4C; each monomer numbered 1 to 4) and the location of the W67 residues. Panel (**B**) illustrates representative fluorescence anisotropy decays obtained for W67 KcsA in the presence of 0.1, 1, and 200 mM K+ at both pH 7.0 (inner gate closed) and pH 4.0 (inner gate open). Equation (3) was then fitted to each data set, and the W67–W67 intersubunit distances were calculated from Equation (4). Panels (**C**,**D**) show the changes in the steady-state anisotropy <r>SS (panel (**C**)) and in the W67–W67 intersubunit lateral distances calculated from the time-resolved anisotropy decays (panel (**D**)) at different K+ concentrations at both pH 7.0 and pH 4.0. Each titration curve is built from duplicate measurements from independent experiments (average ± S.D.) at each of the 20 to 25 different K+ concentrations covering the indicated concentration range. Panel (**E**) shows the linear correlation between the <r>SS and W67–W67 lateral distances (distance = 92.165 × <r>SS + 3.015; r2 = 0.960).

Different from the samples at pH 7.0, we found that the first K+ binding event at K+ < 0.1 mM was no longer detected at pH 4.0, where the channel's inner gate is open and the SF is inactivated. To test whether or not this first K+ binding event is present at pH 4.0, we ran complementary thermal denaturation experiments on the mutant W67 channel. In contrast to the observations from steady-state or time-resolved anisotropy fluorescence measurements, the thermal denaturation experiments showed that the W67 KcsA mutant behaves just like the WT-like channel as both K+ binding events were detected at either pH 7.0 or pH 4.0 (Figure 5).

**Figure 5.** Effect of TOA+ on K+ binding to W67 KcsA. Panels (**A**,**D**) illustrate K+ binding to the W67 KcsA channel in the absence (circles) and presence of 100 μM TOA+ (squares) at pH 7.0 (panel (**A**)) or pH 4.0 (panel (**D**)), monitored through the K+-concentration-dependence of the midpoint temperature of the protein denaturation process (*t*m). Each experimental point is the average *t*m (in Celsius) ± S.D. (*n* = 3). As to the WT KcsA from Figure 3, when the low (panels (**B**,**E**)) and high (panels (**C**,**F**)) K+ concentration ranges are analyzed separately, Equation (1) adequately fits the data from panels (**A**,**D**), which is evidence that the two different sets of K+ binding sites are present in W67 KcsA. The apparent KD values estimated for the above binding events are given in Table 1.

Such apparent discrepancy between the two techniques could be simply explained by assuming that when the inner gate is opened by the acidic pH, the conformational changes involved in thermal stabilization at the low μM K+ range take place without significant modifications at the pore helix level, where the W67 reporter is located, as reflected by the maintenance of the steady-state anisotropy <r>SS (~0.165) or the intersubunit distances (~18 Å), characteristics of the collapsed SF at low K+ concentrations and pH 7.0.

The comparison of anisotropy results from W67 KcsA at pH 7.0 and pH 4.0 at intermediate cation concentrations (the second K+ binding event) shows a clear decrease in K+ binding affinity in the inactivated, pH 4.0 state (Figure 4C,D). This is confirmed by parallel observations from thermal denaturation experiments using either WT or W67 KcsA channels (see Figures 3 and 5 and Table 1). Additionally, when the K+ concentration is raised above 0.1 mM at pH 7.0 (channel in the closed-conductive conformation) or

5–10 mM at pH 4.0 (channel in the open-inactivated form), the <r>SS and W67–W67 intersubunit distances decrease progressively to similar values of ~ 0.130 and 15 Å, respectively, in both cases. This indicates that the final conformations adopted by the SF at high K+ are indistinguishable in terms of intersubunit W67 distances (see Table 2), regardless of pH.

**Table 2.** Comparison of the W67–W67 intersubunit distances calculated from the available X-ray crystallography data and from the homo-FRET analysis of the W67 KcsA mutant channel.


(a) PDB accession numbers of each crystallographic structure are given in parentheses. (b) The calculated distances from the homo-FRET analysis in the W67 mutant channel represent the mean ± S.D. of at least two independent experiments performed at pH 7 (closed state) or pH 4.0 (open state). (c) n.a., not available. (d) The X-ray data from the open states were collected at pH 7.0 using constitutively open mutant channels. (e) The opening of the inner gate was induced by acidic pH.

> Interestingly, at both pH conditions, the K+-dependent changes in the steady-state anisotropy values and the W67–W67 intersubunit distances calculated from the timeresolved anisotropy decays (Figure 4C,D) show an excellent linear correlation. Therefore, the intersubunit distance for a given experimental condition can be easily estimated by simply interpolating the steady-state anisotropy value in the correlation curve shown in Figure 4E. Such a correlation curve was built with the data presented in this work, as well as data previously determined for saturating concentrations of different cations [34].

#### *2.4. The Presence of TOA+ Bound at the Cavity Promotes the Inactivated State of the Selectivity Filter and Allosterically Changes the Pore Helix Conformation*

Once the characterization of the K+ binding behavior of the W67 KcsA mutant channel at pH 7.0 and 4.0 was completed, we ran similar experiments but now in the presence of TOA+ in order to test the effect of this channel blocker on the SF and pore helix conformation and dynamics. Prior to that, control thermal denaturation assays showed that TOA+ binds to the W67 mutant channel in a manner similar to that observed when WT KcsA was used in the experiments. Indeed, the apparent dissociation constants and the corresponding confidence intervals for the TOA+–W67 KcsA complex, estimated at pH 7.0 and pH 4.0, were 10.7 (10.1–11.4) × 10−<sup>8</sup> M and 5.2 (3.6–7.4) × 10−<sup>8</sup> M, respectively. Furthermore, thermal denaturation assays of K+ binding to the W67 mutant channel in the presence of 100 μM TOA+ showed that binding of K+ to two different sets of sites was detected at either pH 7.0 or pH 4.0 (Figure 5). This is also similar to that previously observed in the WT protein and indicates that the stack of K+ binding sites at the SF remains available upon binding of TOA+ to the W67 mutant channel cavity.

In the homo-FRET experiments, two main effects of TOA+ binding to the W67 KcsA channel were detected: first, on the <r>SS (and, thus, on the W67–W67 intersubunit distance) and, second, on the affinity of the SF to bind K+. As to the former, Figure 6A,C shows that the presence of TOA+ promotes an increase in the <r>SS values at all the K+ concentrations tested compared to samples prepared in the absence of TOA+. Such an effect is much more noticeable at pH 7.0 than at pH 4.0 when the channel is already inactivated by the acidic pH. Indeed, the titration curve from the TOA+–containing samples at pH 7.0 closely resembles those obtained at pH 4.0, with or without TOA+, suggesting that similarly to the acidic pH, TOA+ by itself causes inactivation of the channel SF at pH 7.0. The <r>SS values from above were interpolated from the correlation curve shown in Figure 4E to yield the corresponding W67–W67 intersubunit distances. As concluded from the <r>SS values, it is observed that the intersubunit distances are increased in the TOA+–containing samples throughout the K+ titration curve and that such effects are more noticeable at pH 7.0 than at pH 4.0 (Figure 6B,D). This increase in the intersubunit distance at the SF and the resulting loosening of its structure seems consistent with the decreased thermal stability induced by TOA+ in these samples (Figures 3 and 5). Finally, the intersubunit distances of several samples prepared in the presence of TOA+ were also determined experimentally from the time-resolved anisotropy decays. Such experimental values showed an excellent agreemen<sup>t</sup> with those calculated by interpolation from the steady-state anisotropy (black symbols in Figure 6B,D), attesting to the reliability of the simpler interpolation approach.

**Figure 6.** Effect of TOA+ on the conformation of the W67 KcsA channel. Panels (**A**) and (**C**) show a representative experiment of the dependence of the W67 <r>SS on K+ concentration at both pH 7.0 and pH 4.0, in the absence or presence of 100 μM TOA+. The anisotropy values correspond to the average of ten measurements ± S.D. The solid lines represent the best fit of Equation (10) to the data. The apparent KDs for the above binding events were calculated from two independent experiments, and their average values, along with the confidence intervals, are given in Table 1. Panels (**B**,**D**) illustrate the dependence of the W67–W67 intersubunit distances on K+ concentration. The results from the samples without TOA+ were obtained experimentally from time-resolved anisotropy decays, as explained in Figure 4. In the samples containing 100 μM TOA+, such distances were obtained by simply interpolating the <r>SS values using the correlation plot from Figure 4E. Nonetheless, the intersubunit distances in several samples containing 100 μM TOA+ were also determined from time-resolved anisotropy decays (black symbols) to test the goodness of the interpolation procedure.

As to the effects of TOA+ on the affinity of K+ to bind to the SF, the titration curves in Figure 6A,B show that at pH 7.0, the sigmoidal decrease in <r>SS or in the intersubunit distance in the TOA+-containing samples occurs at higher K+ concentrations than in its absence. This attests to a TOA+-induced loss in the affinity of the SF to bind K+. Again, this is not as noticeable in the samples at pH 4.0, in which the channel is already inactivated.

As indicated previously, the homo-FRET experiments monitor quite precisely the second K+ binding event to the KcsA channel, which corresponds to the K+-induced transitions from the nonconductive to either the conductive (pH 7.0) or the inactivated (pH 4.0) states. Therefore, in an attempt to quantitate the observed loss in affinity from the anisotropy changes, an analysis was performed by fitting the model described in Equation (10) (see Methods) to the data corresponding to the second K+ binding event in Figure 6A,C. At pH 7.0, the KD value increases by an order of magnitude when TOA+ is present (0.45 to 4.3 mM), whereas at pH 4.0, it remains fairly constant (44 to 19 mM). Such KD values derived from homo-FRET are in qualitative agreemen<sup>t</sup> with the KD values estimated from thermal denaturation (Table 1 and Figure 5). Again, these observations further sugges<sup>t</sup> that the binding of TOA+ causes inactivation of the channel SF at pH 7.0.

In addition to the main observations on the KDs for K+ binding, it was observed that the fit of Equation (10) to the data required a cooperativity parameter (*h*) higher than 1 only in the case of the pH 4.0 samples in the absence of TOA+ (a "*h*" parameter of 2, with a 95% confidence interval of 1.8 to 2.2, was used in those samples). This finding adds an additional feature to the pH-induced inactivated state of the channel since a cooperative binding process seems involved in facilitating the final inactivated conformation. A similar cooperative effect on the K+ binding to the open channel was previously described by NMR [21]. Such a presumed cooperative process, however, occurs at fairly high K+ concentrations and disappears when TOA+ is bound to the channel cavity.
