Differential Roles of Blocking Ions in KirBac1.1 Tetramer Stability

Abstract

Potassium channels are tetrameric proteins that mediate K⁺-selective transmembrane diffusion. For KcsA, tetramer stability depends on interactions between permeant ions and the channel pore. We have examined the role of pore blockers on the tetramer stability of KirBac1.1. In 150 mm KCl, purified KirBac1.1 protein migrates as a monomer (∼40 kDa) on SDS-PAGE. Addition of Ba²⁺ (K_1/2 ∼ 50 μm) prior to loading results in an additional tetramer band (∼160 kDa). Mutation A109C, at a residue located near the expected Ba²⁺-binding site, decreased tetramer stabilization by Ba²⁺ (K_1/2 ∼ 300 μm), whereas I131C, located nearby, stabilized tetramers in the absence of Ba²⁺. Neither mutation affected Ba²⁺ block of channel activity (using ⁸⁶Rb⁺ flux assay). In contrast to Ba²⁺, Mg²⁺ had no effect on tetramer stability (even though Mg²⁺ was a potent blocker). Many studies have shown Cd²⁺ block of K⁺ channels as a result of cysteine substitution of cavity-lining M2 (S6) residues, with the implicit interpretation that coordination of a single ion by cysteine side chains along the central axis effectively blocks the pore. We examined blocking and tetramer-stabilizing effects of Cd²⁺ on KirBac1.1 with cysteine substitutions in M2. Cd²⁺ block potency followed an α-helical pattern consistent with the crystal structure. Significantly, Cd²⁺ strongly stabilized tetramers of I138C, located in the center of the inner cavity. This stabilization was additive with the effect of Ba²⁺, consistent with both ions simultaneously occupying the channel: Ba²⁺ at the selectivity filter entrance and Cd²⁺ coordinated by I138C side chains in the inner cavity.

Potassium channels are expressed in many cell types and are key players in a wide range of physiological processes. One subset of potassium channels, the inward-rectifying potassium (Kir) channels, are functionally blocked by cytosolic cations such as Mg²⁺ and polyamines and contribute to the regulation of membrane excitability, cardiac rhythm, vascular tone, insulin release, and salt flow across epithelia (1–3). There are seven subfamilies of eukaryotic Kir channel genes. Among them, Kir1 encodes weak rectifiers, whereas Kir2 and Kir5 encode strong rectifiers; Kir3 encodes G-protein-regulated channels; and Kir6 encodes ATP-sensitive channels (4). Recently, a related bacterial family of genes (KirBac) has been identified (5, 6), and in 2003, the first member (KirBac1.1) was crystallized (7), providing a structural model for eukaryotic channels.

The crystal structure of KirBac1.1 revealed a tetrameric pore structure similar to that seen in KcsA and a novel cytoplasmic domain (7, 8). The selectivity filter of both KirBac1.1 and KcsA consists of an extremely conserved pore loop followed by a central cavity, forming a transmembrane ion-selective permeation pore (7, 8). The linear arrangement of five oxygen rings (four from carbonyl oxygens and one from a Thr side chain) in the selectivity filter coordinates with ions, compensating for the energy barrier caused by K⁺ dehydration, thereby facilitating the rapid diffusion of K⁺ across the membrane (8–12). Two-thirds of the KirBac1.1 amino acid residues constitute the cytosolic domain that is highly conserved among the Kir subfamilies and form the cytosolic vestibule (13–16), which, together with the transmembrane pore, generates an 88-Å-long ion conduction pore (7).

The prototypic potassium channel KcsA exists very stably as a tetramer, even in the harsh conditions of SDS-PAGE (17). In addition to protein-protein interaction between monomers, protein-lipid and protein-ion interactions play important roles in stabilizing the KcsA tetramer (17–20). The selectivity filter of KcsA, coordinated with K⁺ ions, can serve as a bridge between the four monomers to maintain the structure of the selectivity filter and the tetrameric architecture of the channel as a whole (11, 21). Blocking ions, such as Ba²⁺, also act as strong stabilizers (17). In the crystal structure of KcsA, Ba²⁺ occupies a site equivalent to the S4 K⁺-binding site within the selectivity filter (22). Other permeant ions (Rb⁺, Cs⁺, Tl⁺, and NH⁺₄) and strong blockers (Sr²⁺) can also contribute to the thermostability of the KcsA tetramer in SDS-PAGE (17). In contrast, impermeant ions such as Na⁺ and Li⁺ or weak blockers such as Mg²⁺ tend to destabilize the KcsA tetramer (17, 19).

Like KcsA, KirBac1.1 purified using decylmaltoside or tridecylmaltoside is active and presumably stable as a tetramer in mild detergent solutions. However, in SDS-PAGE, KirBac1.1 migrates exclusively as a monomer (23). Because KcsA and KirBac1.1 are structurally similar in the transmembrane region of the pore, we hypothesized that permeant and blocking ions would also affect KirBac1.1 tetramer stability in SDS-PAGE. In the present work, the effects of blocking ions such as Ba²⁺ and Mg²⁺ on KirBac1.1 tetramer stability were examined to provide insight to the physical nature of their interaction with KirBac1.1, particularly in the selectivity filter and TM2 cavity. The data reveal important differences in the nature of the interaction of Mg²⁺ and Ba²⁺ with the channel as well as provide previously unavailable evidence for the nature of Cd²⁺ coordination within the channel.

MATERIALS AND METHODS

Protein Preparation and Purification—Protein purification and ⁸⁶Rb⁺ flux assays were performed as previously described (23, 24). Mutants of KirBac1.1 were constructed using the QuikChange site-directed mutagenesis kit (Stratagene) following standard instructions. All mutants were confirmed by DNA sequencing. For protein purification, Escherichia coli BL21-Gold (DE3)pLysS competent cells were transformed and inoculated into LB medium and incubated in a shaker at 37 °C until an A₆₀₀ of 1.0 was reached. Protein expression was induced with 1 mm isopropyl d-thiogalactopyranoside and grown for 3 h at 37 °C. E. coli cells were harvested, lysed by a freeze-thaw cycle, and resuspended into suspension buffer (50 mm Tris-HCl, pH 8.0, 150 mm KCl, 10 mm imidazole, 30 mm decylmaltoside, and one EDTA-free protease inhibitor mixture tablet). The suspension was incubated for 2–4 h at room temperature and then centrifuged at 30,000 × g for 45 min. The supernatant was mixed with cobalt affinity beads and incubated at 4 °C for 2 h. The supernatant/bead mixture was moved to an empty column, and the beads were extensively washed with 20–30 bed volumes of washing buffer (50 mm Tris-HCl, pH 8.0, 150 mm KCl, 10 mm imidazole, and 5 mm decylmaltoside). Target protein was eluted with 1–2 ml of washing buffer containing 500 mm imidazole. Purified protein was concentrated using 30-kDa cut-off centrifugal ultrafilters (Millipore) and then loaded onto a Superdex 200 10/300 GL column with running buffer (50 mm Tris-HCl, pH 8.0, 150 mm KCl, and 1 mm tridecylmaltoside). The fraction corresponding to the major peak at ∼10.8 ml was collected and used in ⁸⁶Rb⁺ flux assays or for tetramer stability assays in SDS-PAGE.

Functional Assay of KirBac1.1 Activity—For ⁸⁶Rb⁺ flux assays, purified protein was mixed with lipids (phosphatidylethanolamine/phosphatidylglycerol at a ratio of 9:1; Avanti) solubilized in Buffer A (450 mm KCl, 10 mm HEPES, and 4 mm N-methyl-d-glucamine, pH 7.4) containing 37 mm CHAPS2 for a final lipid/protein ratio of 100:1 and incubated for 30 min at room temperature. The mixture was loaded onto a Sephadex G-50 column pre-equilibrated with Buffer A to remove detergents and obtain reconstituted proteoliposomes. Extraliposomal solution was changed into Buffer B (450 mm sorbitol, 10 mm HEPES, 4 mm N-methyl-d-glucamine, and 50 μm KCl, pH 7.4) by passing the proteoliposomes through a Sephadex G-50 column pre-equilibrated with Buffer B and then utilized for ⁸⁶Rb⁺ flux assay immediately. Blocking ions at different concentrations were added to 50-μl aliquots of reconstituted proteoliposomes. ⁸⁶Rb⁺ uptake was initiated by mixing proteoliposomes with 4 volumes of Buffer B containing 1–5 μm ⁸⁶Rb⁺. Extraliposomal ⁸⁶Rb⁺ was removed by passage over a 0.5-ml Dowex cation exchange column in the protonated N-methyl-d-glucamine form. Samples were mixed with scintillation fluid and counted in a liquid scintillation counter. Valinomycin was used to assay maximal ⁸⁶Rb⁺ uptake.

Electrophoretic Analysis of KirBac1.1—For SDS-PAGE, 4∼15% Tris-HCl Ready Gels (Bio-Rad) were loaded with 10 μg of protein mixed with loading buffer (final concentration of 62 mm Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, and 0.01% bromphenol blue). The SDS-PAGE running buffer was 25 mm Tris-base and 192 mm glycine, pH 8.3, containing 0.1% SDS. The resulting gels were stained by PhastGel Blue (Sigma). Densitometry measurements were done using ImageJ software, and the fraction of tetramer was normalized by taking the maximal tetramer band in the gel as 1.0. All gels were done in triplicate. No adjustments were made to scanned images except for cropping.

Data Analysis—In some figures, the empirical Hill equation F = 1/(1 + ([X]/K_1/2)^H) was fitted to data using the least-squares method in Excel (Solver add-in, Microsoft), where F = fraction of tetramer, [X] = concentration of cation, K_1/2 = concentration of cation at half-maximal effect, and H = Hill coefficient.

A Fourier transform power spectrum, P(ω), was calculated using Visual Basic (Microsoft) for the sequence of K_1/2 values of residues 135–142 using Equation 1 (25),

(Eq. 1)

where ω = frequency, l = number of residues in the segment, h_k = K_1/2 of the corresponding residue, and h̄ = mean K_1/2 of the segment.

As a quantitative measure of periodicity, an α-periodicity index (α-PI) was calculated using Equation 2.

(Eq. 2)

Large α-periodicity index values (>2) along with a peak in the power spectrum near 100° have been regarded to signify α-helical structures (25–27).

In all figures, experiments were performed at an ambient temperature of 22 °C, including size exclusion chromatography, ⁸⁶Rb⁺ flux assays, incubation of purified protein with blocking cations, and SDS-PAGE. Values are represented as mean ± S.E.

RESULTS

Purified KirBac1.1 Exists as a Tetramer in Detergent Solution—KcsA exists as a stable tetramer, even in the harsh conditions of SDS-PAGE (8, 17). KirBac1.1 was less stable and migrated as a single major band at ∼40 kDa in SDS-PAGE (Fig. 1A), corresponding to the monomer form. In addition, there was a minor band at a slightly lower molecular mass that also gave a positive signal in Western blotting using anti-His₆ tag antibodies, suggesting that it may be an N-terminal degradation product of KirBac1.1 (Fig. 1A) (23). In examining a series of KirBac1.1 mutations with introduced cysteines, we found that substitution of cysteine at Ile-131 greatly enhanced the tetramer stability of KirBac1.1, as indicated by a major band at ∼160 kDa, corresponding to the expected tetramer molecular mass, in SDS-PAGE (Fig. 1A). In contrast, there was no tetramer present in SDS-PAGE when cysteine was substituted at Ala-109 (Fig. 1A), which is physically located very close to residue 131 (Fig. 1B). However, both mutants showed an additional faint band at the expected dimer molecular mass (∼80 kDa), indicating the formation of disulfide-linked dimers during the purification process or electrophoresis. Importantly, although these wild type (WT) KirBac1.1 and cysteine substitution mutants showed different stability in SDS-PAGE, gel filtration analysis revealed that they all ran as a single peak, with almost identical retention volumes, which presumably represents the tetramer form in solution (Fig. 1C).

FIGURE 1.

Purification and tetramer stability of WT KirBac1.1 and mutants A109C and I131C. A, representative SDS-PAGE of purified WT KirBac1.1, A109C, and I131C proteins. All experiments in this figure and the following figures were performed at an ambient temperature of 22 °C. Bands consistent with the dimer are present at ∼70–80 kDa for A109C and I131C. A high molecular mass band at ∼160 kDa, consistent with the tetramer, is seen for only I131C. Tetramer (T), dimer (D), and monomer (M) are shown to the right of the gel. B, ribbon diagram of the KirBac1.1 crystal structure showing only the distal portion of the pore helix, the selectivity filter, and the second transmembrane α-helix. Ala-109 and Ile-131 are highlighted. C, size exclusion chromatography of purified WT KirBac1.1, A109C, and I131C using a Superdex 200 column. The running buffer contained 50 mm Tris-HCl, pH 8.0, 150 mm KCl, and 1.0 mm tridecylmaltoside. The position of KcsA, presumed tetrameric, is marked. mAU, milliabsorbance units.

Ba²⁺ Stabilizes the KirBac1.1 Tetramer—Various permeant ions stabilize the KcsA tetramer, with the permeant blocker Ba²⁺ being particularly effective (17). Preincubation of the WT KirBac1.1 protein with K⁺, Rb⁺, and Cs⁺ at concentrations up to 900 mm did not alter the banding pattern seen in SDS-PAGE. However, preincubation of KirBac1.1 with Ba²⁺ as low as 20 μm led to the appearance of a KirBac1.1 tetramer band in SDS-PAGE, together with smearing of the monomer band toward higher molecular mass (Fig. 2A). The tetramer stabilization was concentration-dependent, with a saturating effect of >90% protein being tetrameric at ∼500 μm Ba²⁺ (Fig. 2B). The tetramer stability of I131C was accompanied by enhanced Ba²⁺ sensitivity (Fig. 2B), whereas the mutation A109C resulted in a right shift of the Ba²⁺ sensitivity, with the tetramer band first appearing in the presence of 200 μm Ba²⁺.

FIGURE 2.

Barium stabilization of the KirBac1.1 tetramer. A, representative SDS-polyacrylamide gels of purified WT KirBac1.1, A109C, or I131C in the presence of varying concentrations of BaCl₂. Purified KirBac1.1 proteins were preincubated with BaCl₂ for 10 min, mixed with 5× protein loading buffer without dithiothreitol, and loaded onto a 4–15% gradient gel. T, tetramer; M, monomer. B, tetramer formation as a function of [Ba²⁺]. The relative density of the tetramer band was estimated by densitometry. Each data point represents an average of three separate gels. C, barium block of KirBac1.1 channel activity. Purified KirBac1.1 channels were reconstituted into liposomes (1-palmitoyl-2-oleoylphosphatidylethanolamine/1-palmitoyl-2-oleoylphosphatidylglycerol at a ratio of 9:1), and ⁸⁶Rb⁺ uptake was measured at 45 s in the presence of varying concentrations of external BaCl₂ (n = 3–5 for each point). Uptake was normalized to maximal uptake as assessed by valinomycin.

Ba²⁺ is an effective blocker of potassium channels because the diameter of its hydrated form is quite close to that of potassium. As shown previously (23), KirBac1.1 activity is sensitive to Ba²⁺ with affinity in the low micromolar range. Even though the A109C and I131C mutations significantly affected the tetramer-stabilizing actions of Ba²⁺, their sensitivities to Ba²⁺ block were not significantly affected (Fig. 2C).

Inner Cavity Blocker Mg²⁺ Does Not Stabilize the KirBac1.1 Tetramer—Mg²⁺ is an important intracellular blocking ion in eukaryotic Kir channels. Krishnan et al. (17) reported that Mg²⁺ does not contribute to the thermostability of tetrameric KcsA, despite the fact that Mg²⁺ is a weak KcsA blocker. Mg²⁺ can strongly block KirBac1.1, and ⁸⁶Rb⁺ flux through KirBac1.1 was blocked with K_1/2 ∼ 80 μm (Fig. 3C). However, preincubation with even 10 mm Mg²⁺ failed to cause a detectable increase in KirBac1.1 tetramer stability (Fig. 3A). Moreover, the Ba²⁺ dependence of tetramer stabilization of WT was unaffected by the presence of 10 mm Mg²⁺ (Fig. 3B). These data indicate fundamental differences between the interaction of Mg²⁺ and Ba²⁺ ions within the pore. First, that Mg²⁺ does not stabilize the tetramer is consistent with a binding site outside the selectivity filter, probably in the inner cavity (1). Second, the lack of effect of Mg²⁺ on Ba²⁺ stabilization suggests that both ions might occupy the pore simultaneously, potentially one in the filter and one in the inner cavity.

FIGURE 3.

Magnesium blocks KirBac1.1 but does not stabilize the tetramer. A, shown is a representative SDS-polyacrylamide gel of WT KirBac1.1 protein preincubated with MgCl₂ for 10 min, mixed with 5× protein loading buffer without dithiothreitol, and loaded onto a 4–15% gradient gel. M, monomer. B, tetramer stabilization by Ba²⁺ is not sensitive to the presence of Mg²⁺. WT KirBac1.1 protein was preincubated in 10 mm MgCl₂ and varying [BaCl₂] for 10 min prior to loading onto SDS-polyacrylamide gel. Data points represent fractional tetramer formation as a function of [Ba²⁺]. C, purified KirBac1.1 channels were reconstituted into liposomes (POPE/POPG at a ratio of 9:1), and ⁸⁶Rb⁺ uptake was measured at 45 s in the presence of varying concentrations of external MgCl₂ (n = 3 for each point and normalized to maximal uptake as assessed by valinomycin). The smooth line is a fitted Hill equation (see “Materials and Methods”) with H = 0.98 and K_1/2 = 78 μm.

Cd²⁺ in the Inner Cavity Can Stabilize the KirBac1.1(I138C) Tetramer—Cd²⁺ ions can specifically coordinate with the sulfhydryl group of cysteine and have been widely used to test the accessibility of residues to a hydrophilic environment in ion channels (28–31). In such experiments, it is implicitly assumed that coordination of Cd²⁺, either in an octahedral pattern with four sulfhydryl groups and two waters or in a tetrahedral pattern with four sulfhydryl groups at the vertices of a pyramid, results in block of the central ion permeation pathway. The above assumptions predict that Cd²⁺ ions should, by coordinating residues from all four subunits of the K⁺ channel, act like Ba²⁺ in the selectivity filter to stabilize the tetramer. Direct biochemical evidence for such an effect in ion channels is absent, but the present approach provides an opportunity to specifically test this prediction.

According to the crystal structure of KirBac1.1 in a presumably closed state, the side chains of Ala-109, which lies immediately beneath the selectivity filter, and of several residues in M2 between Met-135 and Ala-147 are predicted to face the inner cavity. Cd²⁺ was without effect on WT tetramer stability (Fig. 4A) and without obvious effect on the tetramer stability of all mutants examined (see Fig. 5) except I138C. In this mutant, there was a very clear tetramer-stabilizing effect of Cd²⁺, with K_1/2 ∼ 700 μm (Fig. 4, B and C). In the presumably closed channel crystal structure (Fig. 4D) (7), in addition to I138C, there are other residues that face the pore, such as Thr-142 and Val-145, one and two turns below on the M2 helix. It is conceivable that the helices in the Cd²⁺-blocked inner cavity are displaced relative to the crystal structure such that I138C generates the most stable coordination site, but nevertheless, the present result does indeed provide strong physical evidence for the long-held assumption that Cd²⁺ blocks such substituted channels by coordination of cysteines from all four subunits.

FIGURE 4.

Cd²⁺ stabilizes the KirBac1.1(I138C) tetramer. A, SDS-PAGE of WT KirBac1.1 protein (left lane) or WT KirBac1.1 preincubated in 10 mm Cd²⁺ prior to loading onto the gel (right lane). B, representative SDS-PAGE of KirBac1.1(I138C). Mutant protein was preincubated with varying [Cd²⁺] for 10 min prior to loading onto the gel. Tetramer (T), dimer (D), and monomer (M) are shown to the right of the gel. C, [Cd²⁺] dependence of tetramer formation. Data points represent average densitometry measurements of the tetramer band from gels as in B (n = 3 for each point). The curve represents a fitted Hill equation (see “Materials and Methods”) with H = 1.7 and K_1/2 = 687 μm. D, ribbon diagram of KirBac1.1, with Ile-138 highlighted.

FIGURE 5.

Cd²⁺ block of cysteine-substituted KirBac1.1 mutants. A, ⁸⁶Rb⁺ uptake into liposomes reconstituted with KirBac1.1 cysteine mutants was measured at 60 s at varying external Cd²⁺ concentrations. Bars represent the [Cd²⁺] needed to block 50% of maximal uptake (K_1/2) for each mutant (n = 3–7). Mutants labeled with asterisks were nonfunctional. NE, no bacterial expression of protein. Inset, Fourier transform power spectrum of the K_1/2 data (see “Materials and Methods”). The major peak occurs at 92° with an α-periodicity index of 3.2, consistent with the α-helical structure of TM2. The peak of an ideal α-helix is marked by an arrow. B, a helical wheel plot of residues in the second transmembrane α-helix of KirBac1.1 is shown. Cysteine substitution of residues that increased or decreased sensitivity to Cd²⁺ block as compared with WT is colored red or blue, respectively. Cysteine substitution of residues that resulted in nonfunctional channels or no expression are in white. C, pore-facing cysteine-substituted residues confer increased sensitivity to Cd²⁺ block. The space-filled structure of the KirBac1.1 channel is shown, with one monomer moved to the left to show the central pore. The monomer on the left has been rotated 180° to show pore-facing residues. The cytoplasmic domains have been removed for clarity. In each view, the arrow represents the central axis of the channel for reference. Residues that confer increased sensitivity to Cd²⁺ block are colored red.

Cd²⁺ Ions Block KirBac1.1 Channel Activity by Interaction with Cysteines Lining the Inner Cavity—The Cd²⁺ block of KirBac1.1 M2 cysteine substitution mutants was also assessed. As shown in Fig. 5A, cysteine-less WT KirBac1.1 was blocked by Cd²⁺, with K_1/2 ∼ 60 μm. In addition to strong block of the I138C mutant, substitution of cysteine at Met-135, Ala-139, Thr-142, and Val-145 also conferred marked enhancement of sensitivity to Cd²⁺ block (Fig. 5A). A Fourier transform of the Cd²⁺ blocking potency revealed a major peak in the power spectrum at 92° (Fig. 5A, inset), with an α-periodicity index of 3.2, strongly consistent with an α-helical structure that is supported by a helical wheel projection (Fig. 5B). Additionally, the potency of Cd²⁺ block matched very closely the predicted pore-facing residues in the crystal structure (Fig. 5C).

Even though residue 109 faces the inner cavity, the potency of Cd²⁺ block of this mutant was not increased compared with WT KirBac1.1. In contrast to the other tested residues, residue 109 is located on the end of the pore helix at the bottom of the selectivity filter and is conceivably too rigid to have sufficient flexibility to coordinate with Cd²⁺.

Competitive and Additive Effects of Cd²⁺ and Ba²⁺ on WT and I138C Tetramer Stability—The crystal structure of KcsA in the presence of Ba²⁺ demonstrates that Ba²⁺ occupies a single site equivalent to the S4 K⁺-binding site in the selectivity filter (22). Mutation of Thr-75 in KcsA (equivalent to Thr-110 in KirBac1.1), which is involved in Ba²⁺ coordination, led to a dramatic decrease in Ba²⁺ sensitivity of tetramer stabilization, interpreted as a reduction of Ba²⁺ affinity at the S4 site (32). Because Cd²⁺ coordination of I138C occurs in the middle of the inner cavity, it is an interesting question whether Cd²⁺ and Ba²⁺ may simultaneously occupy the pore and what the effect of such double occupancy may have on tetramer stability. For WT channels, the addition of a high concentration of Cd²⁺ (10 mm) resulted in a weak right shift of the sensitivity of the channel to the tetramer-stabilizing effect by Ba²⁺ (Fig. 6, A and B). This result suggests that, in the absence of cysteine thiol side chains, Cd²⁺ interacts mainly with KirBac1.1 at or near the selectivity filter region, thus showing a more or less competitive effect with Ba²⁺. However, preincubation of I138C with Cd²⁺ resulted in a qualitatively different effect on the tetramer-stabilizing effect by Ba²⁺, with a significant leftward and upward shift in the presence of Cd²⁺ at 0.2 mm that is further enhanced at 1 mm (Fig. 6, C and D). Interestingly, at 10 mm Cd²⁺, the concentration dependence of tetramer stability by Ba²⁺ is shifted back to the right, presumably due to the competitive effect of Cd²⁺ with Ba²⁺ at the selectivity filter, as seen in WT KirBac1.1.

FIGURE 6.

Competitive and additive effects of Cd²⁺ and Ba²⁺ on WT KirBac1.1 and I138C tetramer stability. A and C, representative SDS-PAGE of WT KirBac1.1 and I138C proteins, respectively, preincubated with Cd²⁺ and Ba²⁺ at the concentrations indicated prior to loading onto the gel. B and D, Ba²⁺ and Cd²⁺ dependence of tetramer formation. Data points represent average densitometry measurements of the tetramer band from gels as in A and C for WT KirBac1.1 (B) or I138C (D) protein (n = 3 for each point).

DISCUSSION

Tetramer Stabilization at the Selectivity Filter—K⁺ channels from different origins have a conserved “TVGYG” motif, which generates the selectivity filter that underlies rapid and selective K⁺ transmembrane diffusion. The crystal structures of KcsA, MthK, KvAP, and KirBac1.1 indicate that all these K⁺ channels share almost identical selectivity filter structures and very similar inner cavities lined by the M2 (S6) helices (7, 8, 33, 34). It has long been recognized that complete removal of K⁺ will cause K⁺ channels to permanently lose their K⁺ selectivity or ion permeation capacity as a result of collapse or distortion of the selectivity filter (35–37), and molecular dynamics simulations suggest that the residence of K⁺ ions plays a critical role in stabilizing the structure of the selectivity filter of KcsA (9, 12, 21). Such a stabilizing effect conferred by the permeant ion-oxygen ring coordination is likely a property of all K⁺ channels (11, 21). Krishnan et al. (17) have shown that multiple permeant ions not only stabilize the selectivity filter but also stabilize the channel tetramer itself in KcsA.

KirBac1.1 is one member of a family of prokaryotic genes that contain significant primary sequence homology to the eukaryotic Kir channel gene family (5). KirBac1.1 has been crystallized (7) and has now been extensively used as a model for eukaryotic Kir channel structures. Both ion flux experiments (23) and voltage-clamp experiments3 confirm that KirBac1.1 functions as a K⁺-selective inward rectifier channel. As with KcsA, the availability of high levels of pure KirBac1.1 protein permits multiple previously impossible biochemical analyses of channel structure and function, in addition to crystallography. In the present study, we have taken advantage of this to examine control of KirBac1.1 tetramerization. We have shown that KirBac1.1 is less stable than KcsA as a tetramer, running exclusively as a monomer in SDS-PAGE, but that Ba²⁺ confers tetramer stability. Our results indicated that the Ba²⁺ concentration dependence of KirBac1.1 tetramer stability is influenced by a mutation (A109C) located at the predicted Ba²⁺-binding site (22), consistent with tetramer stabilization resulting from a selectivity filter interaction. The permeant ions K⁺, Rb⁺, and Cs⁺, however, showed no obvious increase in KirBac tetramer stability at high concentrations, in distinct contrast to their stabilizing effect seen in KcsA. This may be due to the assay not being sensitive enough to report the increased stability conferred by these cations, and further studies are warranted.

It is unclear how the tetramer is stabilized by the I131C mutation. However, several residues in TM1 and TM2 of KcsA that are adjacent to the pore helix are involved in tetramer stability (38), including Met-96, a residue homologous to Ile-131 in KirBac. Interestingly, mutation of Met-96 in KcsA resulted in decreased tetramer stability, in contrast to the increased stability seen with the I131C mutation in KirBac.

Selectivity Filter Versus Inner Cavity Blocker Interactions—Mg²⁺ is a well characterized intracellular blocker of Kir channels (2, 39, 40). Other cations such as Ba²⁺ are less physiologically relevant but can also block Kir channels effectively (41–44). However, their structural mechanisms of blocking appear to be different. For Mg²⁺ blocking, Ser-165 and Asp-172 in the inner cavity of Kir2.1 (45–48) are important, implying an inner cavity site. Ba²⁺, with its radius of the hydrated form quite close to that of hydrated K⁺, is likely to block the Kir channel at the selectivity filter (41, 49), as in KcsA and other K⁺ channels. Whereas both Ba²⁺ and Mg²⁺ are group IIA cations and have similar chemical properties, Ba²⁺ greatly enhances the tetramer stability of KirBac1.1 in SDS-PAGE, whereas no obvious stabilizing effects are observed for Mg²⁺. Lacking d electrons, Mg²⁺ coordinates rigidly with oxygen, always in an octahedral geometry with very little flexibility in bond length and angle (50). One possible interpretation is that inflexible coordination patterns of Mg²⁺ make it difficult for this ion to coordinate with carbonyl atoms from four KirBac1.1 monomers at the same time such that Mg²⁺ could be a high affinity blocker of KirBac1.1 but not competent for strong tetramer stabilization. However, there are no obvious competitive or additive effects of Mg²⁺ on Ba²⁺-induced tetramer stabilization, consistent with different binding sites with little or no interaction, i.e. at the selectivity filter S4 site and in the inner cavity.

Tetrameric Coordination of Cysteine Side Chains by Cadmium Ions in the Inner Cavity—Cadmium can specifically coordinate with the sulfhydryl side chains of cysteine and has been widely used as a probe to test the accessibility of a residue in hydrophilic environments. In particular, many studies have utilized cadmium accessibility to assess pore-lining residues in different ion channels (28–30). Our cadmium blocking assay demonstrated that Met-135, Ile-138, Ala-139, Thr-142, and Val-145 of TM2, which lie on one face of the M2 helix of KirBac1.1, tilted at ∼40° to the helix axis, are highly sensitive to Cd²⁺ block when substituted with cysteines. These are exactly equivalent to the residues previously identified by this approach as pore lining in eukaryotic Kir2.1 and Kir6.2 channels, an important confirmation of the similarity of eukaryotic and bacterial Kir channel structures and hence the relevance of KirBac1.1 as a structural model of eukaryotic Kir channels. In the present experiments, we were also able to test the previously implicit assumption that Cd²⁺ block of such channels requires coordination with four cysteines in the central axis of the channel; indeed, I138C shows a very prominent tetramer band in the presence of Cd²⁺. However, despite the fact that all cysteine substitution mutants at these positions led to a remarkable increase in cadmium blocking, I138C was the only residue to show such a clear tetramer band. The reasons for this are not clear, but potentially, block of KirBac1.1 might occur through weaker interaction of Cd²⁺ with less than four cysteines of KirBac1.1 at sites other than residue 138, which could thereby interrupt K⁺ permeation without stabilizing the tetramer.

The coordination of Cd²⁺ at I138C could occur in two forms: a tetrahedral form with the four sulfhydryl groups at the vertices of a triangular pyramid and an octahedral form with four sulfhydryl groups and two waters at the vertices of an octahedron. We cannot determine which occurs, although the octahedral form would require less movement from the closed crystal structure.

The data are consistent with the notion that residues facing the pore in the closed crystal structure also face the pore in the Cd²⁺-blocked state. Structural changes in the channel might occur as it transitions from a conducting state to a Cd²⁺-bound state. In Kir6.2, increasing the open probability lowered the potency of Cd²⁺ block by decreasing the on-rate with no change in the off-rate, suggesting that Cd²⁺ enters the open channel, but movement in TM2 is required to bring the side chains close enough to coordinate the Cd²⁺ ion (30). Given the similarities between KirBac1.1 and eukaryotic Kir channels (including Kir6.2), we might suggest that KirBac1.1 behaves similarly.

Conclusions—The stability of KcsA and KirBac1.1 estimated by their tetramer band appearance in SDS-PAGE may not be a direct indication of their real oligomeric state in a membrane; however, it is a reasonable way to evaluate contributions of ion-protein interactions in stabilizing the tetramer. Examination of conformational changes in KcsA and KirBac1.1 following the coordination and disassociation of permeant or blocking ions provides complementary data to that of crystallographic (9, 10, 51) and other (52) approaches that will lead to further advances in our understanding of the structural basis of K⁺ selection and channel blockade.