A short intrinsically disordered region at KtrB's N-terminus facilitates allosteric regulation of K+ channel KtrAB

Abstract

K⁺ homeostasis is crucial for bacterial survival. The bacterial K+ channel KtrAB is regulated by the binding of ADP and ATP to the cytosolic RCK subunits KtrA. While the ligand-induced conformational changes in KtrA are well described, the transmission to the gating regions within KtrB is not understood. Here, we present a cryo-EM structure of the ADP-bound, inactive KtrAB complex from Vibrio alginolyticus, which resolves part of KtrB's N termini. They are short intrinsically disordered regions (IDRs) located at the interface of KtrA and KtrB. We reveal that these IDRs play a decisive role in ATP-mediated channel opening, while the closed ADP-bound state does not depend on the N-termini. We propose an allosteric mechanism, in which ATP-induced conformational changes within KtrA trigger an interaction of KtrB's N-terminal IDRs with the membrane, stabilizing the active and conductive state of KtrAB.

Fig. 1. High-resolution structure of KtrAB in its intact KtrB₂A₈ assembly compared to the nonphysiological KtrB₂A₈B₂ assembly.

Cryo-EM density maps of a KtrB₂A₈ with an overall resolution of 2.8 Å and b KtrAB in the nonphysiological KtrB₂A₈B₂ assembly with an overall resolution of 2.5 Å. c Side-view and d bottom-view of the overlay of both structures (KtrB₂A₈B₂ colored, KtrB₂A₈ pink). The assembly of an additional KtrB dimer to the free side of KtrA does not change the overall structure of the KtrAB complex. Small rotations were identified in two KtrA subunits where the extra KtrB dimer binds, with no significant consequence on the remaining protein. KtrB: gray, D1M2 helix: goldenrod, intramembrane loop: dark cyan, R427: magenta, N-terminus of KtrB: green, KtrA: blue (navy blue, cornflower blue), ADP: orange. If not stated otherwise, coloring is maintained in all figures.

Fig. 2. Structure of the KtrAB complex in the ADP-bound conformation.

a Side-view of KtrAB with all features colored as described in Fig. 1, adopting the ADP-bound conformation. The D1M2 helices (goldenrod) extend into the KtrA ring. b View of KtrAB from the cytosol. The KtrA ring adopts the oval shape with a short axis of 6.3 nm and a long axis of 7.8 nm, as described previously for KtrAB in the ADP-bound conformation. An ADP molecule is resolved in each KtrA subunit. c R40 in KtrA coordinates the β-phosphates of the ADP molecules. A non-protein density was assigned to a Mg²⁺ (green) coordinated to the α-phosphate. d In alternating KtrA subunits, residue R15 coordinates the phosphate.

Fig. 3. Binding affinities of nucleotides to KtrA and nucleotide-dependent protein stabilization.

KtrA was purified via SEC and the affinities of nucleotides were determined using ITC. a The upper panel shows the raw heat exchange data of ATP and ADP binding to KtrA, respectively. The lower panels present the integrated injection heat pulses, normalized per mole of injection. The binding curves were fitted by a one-site binding model, resulting in the indicated dissociation constants (K_D). Respective K_D values and SDs are noted for triplicates of each titration. b, c Thermostability of KtrA was investigated by differential scanning fluorimetry. Purified KtrA was mixed with different concentrations of ATP (red) or ADP (blue) and changes in SyproOrange fluorescence (λ_ex 470 nm; λ_em 555 nm) were detected over a temperature increase from 25 to 80 °C. The peak of the raw data's first derivative reflects the protein's melting point. d Melting temperatures of purified KtrA (black) and in the presence of different concentrations of ATP (red) or ADP (blue) (10 µM to 1000 µM). ADP binding induces higher thermal stability at significantly lower concentrations. e To test ATP/ADP displacement, competitive binding experiments were performed. 5 µM KtrA were pre-incubated with either 500 µM ATP (black-framed blue point) or 500 µM ADP (black-framed red point) and supplied with different concentrations of ADP (blue) or ATP (red), respectively. Displacement of both nucleotides was observed. Data points in d and e show the means and SDs of technical triplicates.

Fig. 4. Structural details of the selectivity filter and stabilization of the intramembrane loop of KtrB.

a Close-up view of the selectivity filter. Coordination of a potassium (purple density) in the selectivity filter (SF) at the S3 binding site by backbone carbonyl oxygens. Two additional non-protein densities in the SF were assigned as water molecules. D1 and D3 have been removed for better visualization. b Close-up view of the intramembrane loop (cyan). It is stabilized in the closed conformation mainly by interactions with the backbone of the extended D1M2 helix and the D4M2 helix, together with a salt bridge between K325 and the backbone carboxylate of the C-terminal G455 from the neighboring KtrB protomer. The conserved R427 (magenta) of helix D4M2, which is stabilized by Q99, forms an additional electrostatic barrier for potassium flux. c Calculated pore radii in a KtrB monomer along the central axis. For reference, the radius of a dehydrated K⁺ is shown as a dashed line. The pathway is clearly restricted by the intramembrane loop. Radii were calculated using HOLE. d K⁺ uptake by E. coli LB2003 cells producing KtrAB_K325A was performed to determine uptake kinetics. After protein production, cells were depleted of K⁺ and, after 10 min incubation at room temperature, different K⁺ concentrations (0.1, 0.2, 0.4, and 1 mM, different shades of gray) were added. 1 ml samples were taken at different time points, and cells were separated from medium via centrifugation through silicone oil. Intracellular K⁺ concentration was determined by flame photometry. Uptake experiments were performed in biological triplicates with one representative graph shown. e Slopes from initial uptake velocities are plotted against the used KCl concentration with error bars resulting from the linear fitting. A Michaelis-Menten fit was performed to determine V_max and K_m. The indicated error is of the individual fit. f Uptake experiments were also performed with the KtrAB_Q99A variant and KtrAB WT for comparison (Supplementary Fig. 6). Average Michaelis-Menten kinetics for WT KtrAB, KtrAB_Q99A, and KtrAB_K325A derived from biological triplicates. Both variants revealed increased V_max compared to WT. The K_m values remained similar.

Fig. 5. Stabilization of extended D1M2 helices within KtrAs and identification of KtrB’s unstructured N-termini at the interface of KtrA and KtrB.

a Extended helices of KtrB (yellow) are stabilized by several hydrophobic interactions and hydrogen bonds with KtrA (blue), including a salt bridge between R117 (KtrB) and E38 (KtrA) (distance 3.1 Å). b Cryo-EM density (gray mesh) and c model for the N-terminus of KtrB (green) at the interface of KtrB and KtrA, which is resolved from residue R7 to D14. R7 forms a stabilizing salt bridge with E64 from KtrA. A hydrophobic patch from V9 to P13 interacts with the extended helix and KtrA. A motif of three positively charged residues (K16, R17, K19) was not resolved, indicating high flexibility in this region.

Fig. 6. Effect of N-terminus deletion on the KtrAB complex.

a Cryo-EM map at 3.6 Å resolution of the KtrBΔ2-19₄A₈ assembly. The absence of the N-terminus has no effect on the ADP-bound conformation of the system. b While the system adopts the ADP-bound conformation with extended D1M2 helices and an oval-shaped KtrA ring, density for the N-termini is lacking, as expected. c For comparison, a close-up view of full-length KtrAB low-pass filtered to the same resolution (3.6 Å) is shown, presenting the density of the N-terminus (green).

Fig. 7. All-atom MD simulations of KtrB dimer and the KtrAB complex in a lipid environment.

a Snapshot of the membrane and the KtrB dimer after an initial equilibration of the system. The disordered N-termini (blue and green) hover below the membrane. b After 1 µs of MD simulation at 310 K, the N-terminus of one protomer (blue) interacts with the membrane surface, primarily via its positively charged residues. c, d The N-terminus of the other protomer (green) shows a strong interaction with the lipid membrane after 1000 ns of simulation time at 310 K, with the hydrophobic patch (pink in d) interacting with the acyl chains while the basic residues (blue in d) interact with negatively charged head groups. e Dynamics of the N-termini during the MD simulation of KtrB at different temperatures. Snapshots of the N-termini (blue) taken every 1 ns from the 1 µs simulation after superimposition of the protein core. f Dynamics of KtrB’s N-termini (blue) in the KtrAB complex during MD simulations at 310 K. Snapshots of the N-termini taken every 1 ns from the 1 µs simulation after superimposition of the protein. g The fraction of native contacts between KtrA and the N-termini of KtrB remains high in the KtrAB complex even at elevated temperatures. h The length of KtrB’s N-termini varies in the KtrB dimer, showing their unstructured nature. In the KtrAB complex, the length is stable as the termini remain in the pocket between the subunits. i Distributions of the distance of the KtrB N-termini from the membrane surface in the $z$ direction (in Å) in the MD simulations of the KtrB dimer and the full KtrAB complex. The membrane is indicated by the gray shading, $z=0$ corresponding to the average phosphate positions. j Distribution of the number of contacts formed between lipids and N-terminal residues of KtrB in the MD simulations of the KtrB dimer and the full KtrAB complex. The time series of the results of panels g–j are shown in Supplementary Fig. 8.

Fig. 8. Dependency of the intramembrane loop flexibility of KtrAB on nucleotides and KtrB N-termini.

a DEER measurements of spin-labeled natively assembled KtrAB_T318C and KtrAB_Δ2-19__T318C in the presence of 10 mM ADP (blue) or 10 mM ATP (red) reconstituted in E. coli polar lipid liposomes with a lipid:protein ratio of 1:10. Left panels: Experimental raw data V(t) with fitted background function of individual measurements; middle panels: Background-corrected dipolar evolution functions F(t); right panels: Interspin distance distributions P(r) obtained by Tikhonov regularization. Light gray and black striped areas, respectively, represent the full variation of possible distance distributions. The lower and upper error estimates (light gray and black lines, respectively) represent the respective mean values plus and minus twice the standard deviation. On the x-axis, the green range corresponds to a reliable shape of the distribution, the yellow range to reliable mean distance and width, the orange range to reliable mean distance (but not width), and the red range to recognition of a long-distance contribution that cannot be quantified. DeerAnalysis2022 was used for the analysis. Distances are measured between the labeled residues in both protomers, the indicated distance in the model is the mean distance resulting from a rotamer library analysis using MMM2020.1. While in the presence of ADP the interspin distances resemble each other, the gating loops must be significantly delocalized in KtrAB_Δ2-19__T318C in the presence of ATP when compared to full-length KtrAB_T318C. b Comparison of distance distributions from KtrB_T318C (black), KtrAB_T318C + ATP (blue), and KtrAB_Δ2-19__T318C + ATP (red). The corresponding DEER measurement of KtrB_T318C is shown in Supplementary Fig. 10.

Fig. 9. Deletion and mutations of the KtrB N-terminus affect K⁺ uptake by the KtrAB complex.

a, b V_max (a) and K_M (b) for KtrB IDR variants. Uptake experiments were performed as described above. E. coli LB2003 cells produced KtrAB WT and variants thereof, as well as KtrB in the absence of KtrA. K⁺-depleted cells were prepared, and K⁺ uptake was measured in the presence of 0.1, 0.2, 0.4, and 1 mM KCl. All measurements were performed several times as indicated in the figure (n = x) and the Michaelis-Menten kinetics were determined using initial uptake velocities. Data are presented as a scatter plot graph with gray bars representing mean values +/− SDs. Exemplary uptake data are shown in Supplementary Fig. 11.

Fig. 10. Suggested gating mechanism of the KtrAB complex.

Cartoon representing an allosteric mechanism in which the interaction of KtrB’s N-termini with the membrane, together with ATP- and Na⁺-triggered conformational changes within KtrA, induce the activation of the KtrAB system. In the ADP-bound conformation (left, ADP in orange), the RCK ring (blue) has an oval shape, and the D1M2 helices are extended (gold), reaching into the KtrA ring. The N-termini of KtrB (green) lie flat on the KtrA ring, and the intramembrane loop (cyan) blocks potassium flux. Middle: Upon hyperosmotic stress, which goes along with an increased ATP concentration, the ADP-bound conformation gets destabilized by nucleotide exchange, and the N-termini relocate towards the membrane, establishing electrostatic contacts with the negatively charged phospholipid headgroups via several positive charges. Subsequently, a hydrophobic interaction of the hydrophobic patch with the acyl chains locks the interaction. This induces conformational rearrangements in KtrB: the D1M2 helices break, and helices D4M2 (pink) containing R427 relocate away from the pore. In parallel, in KtrA, further replacement of ADP by ATP (dark red) and Na⁺ results in a square-shaped KtrA ring, pulling open the unlocked gate. Subsequent potassium uptake increases the intracellular positive charge (right), which locally shields the negatively charged membrane surface, leading to a repulsion of the N-termini. This is the first step of restoring the inactive, ADP-bound conformation to avoid K⁺-mediated cytotoxic processes by unhindered uptake. Adapted from ref. .