A cyclic nucleotide modulated prokaryotic K+ channel

Abstract

A search of prokaryotic genomes uncovered a gene from Mesorhizobium loti homologous to eukaryotic K(+) channels of the S4 superfamily that also carry a cyclic nucleotide binding domain at the COOH terminus. The gene was cloned from genomic DNA, and the protein, denoted MloK1, was overexpressed in Escherichia coli and purified. Gel filtration analysis revealed a heterogeneous distribution of protein sizes which, upon inclusion of cyclic nucleotide, coalesces into a homogeneous population, eluting at the size expected for a homotetramer. As followed by a radioactive (86)Rb(+) flux assay, the putative channel protein catalyzes ionic flux with a selectivity expected for a K(+) channel. Ion transport is stimulated by cAMP and cGMP at submicromolar concentrations. Since this bacterial homologue does not have the "C-linker" sequence found in all eukaryotic S4-type cyclic nucleotide-modulated ion channels, these results show that this four-helix structure is not a general requirement for transducing the cyclic nucleotide-binding signal to channel opening.

Figure 1.

MloK1 in the S4 ion channel superfamily. Predicted transmembrane segments shown as cylinders, CNBD in yellow, pore region in cyan. Alignments are colored to correspond with transmembrane topology diagram. In CNBD, the completely conserved Arg is shown in green and the adjacent conserved hydroxy amino acid in red. Asterisk marks the position of acidic residues in CNG channels (Varnum et al., 1995) responsible for selectivity against cAMP. Channel abbreviations, with NCBI accession nos., are as follows: HCN1, mouse HCN (O88704), HCN2, human HCN (Q9UL51), CNGa1, bovine CNG (Q00194), CNGa2, bovine CNG (Q03041), SPIH, sea urchin HCN (NP999729).

Figure 2.

Overexpression and purification of MloK1. Protein samples were run on a 10% polyacrylamide SDS gel stained with Coomassie blue as follows: lane 1, protein ladder with molecular weights indicated; lanes 2 and 3, pre- and post-induction whole E. coli lysates; lane 4, clarified membrane extract of induced cells; lane 5, Ni²⁺ column flow-through of membrane extract; lane 6, nonspecifically bound protein wash of Ni²⁺ column; lane 7, Ni²⁺ column–eluted MloK1 (7 μg); lane 8, gel filtration column-purified MloK1 (7 μg). Arrow indicates position of purified MloK1 protein.

Figure 3.

Gel filtration chromatography of MloK1. Preparations of Ni²⁺ column–purified (solid curve) or gel filtration–purified (dashed curve in B) MloK1 were applied to a Superdex gel filtration column monitored at 280 nm in the absence (A) or presence (B) of 50 μM cAMP. Dashed lines mark the tetrameric MloK1 peak and arrows indicate the void volume and molecular weights of two calibrated membrane proteins run under identical conditions: MthK K⁺ channel tetramer (250 kD) and CLC-ecl Cl⁻ transporter dimer (92 kD). According to a calibration using five membrane proteins of known sizes run under identical conditions, the elution volume of MloK1 corresponds to ∼150 kD. Scale bar for absorbance at 280 nm represents 20 mAU for A and 100 mAU for B.

Figure 4.

Concentrative ⁸⁶Rb⁺ uptake by MloK1. (A) Representative time course of accumulation of ⁸⁶Rb⁺ into liposomes reconstituted with MloK1 (10 μg/mg lipid, 200 μM cAMP). Each point represents radioactivity of a single 100-μl sample removed from the reaction mix at the time point indicated. After the 2-h sample was collected, valinomycin was added (arrow) and a sample was taken 2 min later. (B) Valinomycin-normalized ⁸⁶Rb⁺ uptake in liposomes reconstituted with MloK1 (10 μg/mg lipid) in the presence of cAMP (200 μM, black circles) or in its absence (gray circles). Unfilled circles represent background ⁸⁶Rb⁺ uptake in protein-free liposomes. Symbols represent mean ± SEM of at least three different experiments. Smooth curves carry no theoretical meaning. Inset, cartoon of ⁸⁶Rb⁺ uptake assay, with liposomes loaded with high K⁺ (400 mM) immersed in low K⁺ solution (50 μM) containing tracer Rb⁺; this leads to a high negative intraliposomal potential, which the Rb⁺ tracer follows (Heginbotham et al., 1998).

Figure 5.

⁸⁶Rb⁺ uptake is dependent on protein concentration. (A) Valinomycin-normalized ⁸⁶Rb⁺ uptake time course for 0 (white), 0.1 (light gray), 1 (dark gray), and 5 (black) μg MloK1/mg lipid in the presence of 50 μM cAMP. (B) Protein concentration dependence of ⁸⁶Rb⁺ uptake. Each time point represents ⁸⁶Rb⁺ accumulation after 45 min in liposomes containing MloK1 at the indicated concentrations and in the presence of 50 μM cAMP. Symbols represent mean ± SEM for three different experiments.

Figure 6.

Ionic selectivity of MloK1-mediated ⁸⁶Rb⁺ uptake. Each bar represents ⁸⁶Rb⁺ uptake into liposomes reconstituted with 5 μg protein/mg lipid measured after 45 min of uptake and normalized to the total amount of radioactivity in the sample. Reconstituted liposomes were formed in the presence of the 400 mM of the Cl⁻ salts of the indicated cations, along with of 50 μM cAMP. Bars represent mean ± SEM for 7–12 different experiments.

Figure 7.

Cyclic nucleotide dependence of ⁸⁶Rb⁺ uptake. Flux time courses into liposomes reconstituted with 5 μg MloK1/mg lipid in the presence of cAMP (A) and cGMP (B), at 0 (white), 0.1 (light gray), 1 (dark gray), and 10 (black) μM concentration. (C) Cyclic nucleotide dose dependence of ⁸⁶Rb⁺ uptake. Uptake values at 15 min (squares) and 45 min (circles) are normalized to both maximum (saturating concentration of cNMP) and minimum (no cNMP) values. Smooth curves are drawn according to single-site binding functions, with K_1/2 = 60 nM for cAMP (black) and 600 nM for cGMP (gray). Each symbol represents mean ± SEM for three different experiments.

Figure 8.

MloK1-mediated K⁺ flux is conductive. (A) Cartoon of the method. Liposomes are loaded with high K⁺ and suspended in low K⁺, with a K⁺-specific conductance present in the membrane. Under these conditions, the membrane becomes polarized, inside negative. If a proton ionophore is present to circumvent the impermeability of the lipid bilayer, protons will be drawn into the liposomes to “follow” the electrical potential set up by the K⁺ gradient. (B) Demonstration of electrically driven proton uptake in MloK1-reconstituted liposomes. Vesicles (20 mg/ml) loaded with high K⁺ (450 mM) were diluted 20-fold into lightly buffered low K⁺ (10 mM) medium, and pH was recorded. At asterisk, FCCP (0.05–0.5 μg/ml), and at arrow, valinomycin (0.5 μg/ml) were added. Top, protein-free liposomes. Bottom, liposomes reconstituted with 5 μg protein/mg lipid. Horizontal scale bar, 30 s; vertical scale bar, 13 nmol H⁺, 0.003 pH units. Upward deflection represents pH increase in external solution.