Molecular Biology and Cell Culture

As described previously [12], human NKA α1, α2, and α3 isoforms were tagged on their N-termini by Cerulean (mCer3) and eYFP for progressive photobleaching experiments and by mCyRFP1 [46] or mMaroon1 [45] for FLIM and TCSPC experiments. The single amino acid mutations were prepared using a Quick Change Lightning II Site-Directed Mutagenesis Kit (Agilent Technologies, USA) and verified by single-pass primer extension analysis (ACGT Inc., USA). To improve the localization of the NKA α subunit, we coexpressed the unlabeled β subunit, which helped translocate and anchor the α subunit in the plasma membrane. HEK AAV-293 cells were transiently transfected using a Lipofectamine 3000 transfection kit (Invitrogen, USA) and incubated for 48 h post-transfection. Cells were trypsinated and plated at a density of 200,000 cells/well into two-well poly-d-lysine (PDL)-coated glass-bottom chamber slides. Cells were allowed to adhere for at least 1 h, washed with PBS, and imaged.

For confocal microscopy, HEK AAV-293 cells stably expressing N-terminal eCFP-labeled NKA [59, 60] were transfected with 5.0 µg of N-terminus eYFP-labeled NKA WT or G301 A/R mutants and 5.0 µg of unlabeled NKA β subunit. Cells were cultivated in DMEM supplemented with 10% FBS and 2% DMSO. Twenty-four hours post-transfection, cells were washed with sterile PBS, trypsinated, and seeded at 100,000 cells/well into a PDL-coated four-well glass-bottom chamber slide. Cells were allowed to adhere for 8–10 h after seeding, and subsequently, the media was replaced with serum-starvation media without FBS (DMEM + 2% DMSO) and incubated overnight.

Confocal Microscopy

Localization experiments were performed using a Leica SP5 laser scanning confocal microscope with a 63 × water immersion objective. To observe the fluorescence of eCFP-labeled α subunit expressed by the stable cell, it was excited by a 458-nm line of an argon laser, and fluorescence was detected in the range between 465 and 511 nm. eYFP was excited by a 514-nm line of an argon laser, and emission was detected between 520 and 557 nm. Images were acquired in sequential line scans with an average of 64 scans.

Images were analyzed by manually cropping the membrane and cell interior area, measuring the average intensity of eCFP and eYFP fluorescence in ImageJ software [61]. Membrane localization was calculated as the ratio between the average intensity in the membrane divided by the average intensity measured in the cell interior for both fluorescent proteins.

Progressive Acceptor Photobleaching

Twenty-four hours post-transfection, cells were trypsinized and plated at 100,000 cells/well into four-well glass-bottom chamber slides and incubated overnight at 37 °C. Immediately before imaging, cells were washed with PBS, and images were acquired by a Nikon Eclipse Ti2 inverted microscope with a 1.49 Apo TIRF 100 × objective with a Nikon PFS system. Each photobleaching step consists of CFP and YFP images taken every 10 s followed by 10 s of YFP photobleaching. There were a total of 75 photobleaching steps.

Fluorescence Lifetime Measurements

Fluorescence lifetime measurements were performed using time-correlated single-photon counting (TCSPC) as previously described [12, 44, 50]. TCSPC histograms were acquired for HEK AAV-293 cells expressing mCyRFP1-α1 WT, G301A or G301R mutants, and mMaroon1-α1 WT or G301 mutants as described in the “Molecular Biology and Cell Culture” section. Cells expressing both mCyRFP1 and mMaroon1 proteins were manually selected using a defocused excitation supercontinuum laser beam (Fianium) to excite the whole cell. Donor (mCyRFP1) signal was acquired using an excitation filter 482/18-nm bandpass filter with 0.3 neutral density filter. Fluorescence emission was detected using a 640/50-nm emission bandpass filter. After cell selection using the defocused laser beam, the defocusing lens was removed from the light path, the laser intensity was attenuated with a 1.0 neutral density filter, and the focused laser was placed on the cell membrane yielding 100,000 photons/s count rate. Under those conditions, we observed less than 10% photobleaching during the 60-s acquisition period. Fluorescence was detected through 1.2 N. A. water immersion objective with an avalanche photodiode connected to a photon-counting module (PicoHarp 300, PicoQuant, Germany) with a time channel width of 16 ps. As previously described [12], mCyRFP1-labeled α subunit exhibits a minor second component (relative amplitude below 5%) and can be considered a single-exponential decay. The cells expressing mCyRFP1- and mMaroon1-labeled α subunits showed multiexponential decays with two components. To quantify these components, fluorescence decay histograms from four independent transfections were analyzed in SymPho Time 64 software using global analysis. The intensity decays were fitted with a multiexponential model using the formula \(F=_^_}}+_^_}}\) where F is fluoresence intensity, A1 and A2 are amplitudes of two components, and τ1 and τ2 are lifetimes of each component. Specifically for FRET, A1 and A2 are amplitudes of two components and τ1 and τ2 are lifetimes of donor alone and lifetime of FRET species. We assumed the presence of a population of donor-labeled α subunits does not interact with acceptor-labeled α subunits. This population is accounted for in the global fitting with a fixed mCyRFP1 lifetime of 3.48 ns, which was the value obtained from cells expressing donor-labeled α subunit with no acceptor-labeled binding partners. The second shared lifetime was determined without any additional restrictions for all histograms in the dataset. We did not detect a significant difference between the second shared lifetime for WT, G301A, or G301R pairs, where the second shared lifetime was ranging from 0.922 to 1.23 ns. The average FRET efficiency was calculated using the following formula: \(E\left(\%\right)=100\times \left(1-\frac_}}_}}\right)\), where τD is the lifetime of donor alone and τDA is amplitude-weighted average lifetime. To calculate FRET efficiency for FRET component, it can be calculated with the formula above, but τDA is the lifetime of FRET component. The distance between donor and acceptor can be calculated as \(r=_^\left(\frac\right)\right)}^\), where E is the FRET efficiency and R0 is the Förster distance of selected FRET pair. R0 for mCyRFP1 and mMaroon was taken to be 63.34 Å [62].

Plasma Membrane Microsome Preparation

Plasma membrane microsomal fractions from HEK cells were prepared as described previously [63]. HEK293 T cells transiently transfected with the Lipofectamine 3000 (Invitrogen) kit were collected 48 h post-transfection and homogenized using a glass Dounce homogenizer in the homogenization buffer supplemented with a protease inhibitor cocktail (250 mM sucrose, 5 mM MgCl2, 10 mM HEPES, 1 mM PMSF; pH 7.4). Subsequently, the homogenate was centrifuged at 500 × g for 10 min at 4 °C. The supernatant was kept on ice while the pellet was dissolved in homogenization buffer, and homogenization and centrifugation steps were repeated. Both supernatants were combined and layered on top of a discontinuous sucrose gradient prepared by layering 4.5 ml of 20% sucrose in homogenization buffer on top of 4.5 ml of 50% sucrose in homogenization buffer and centrifuged at 30,000 × g for 1 h at 4 °C. After the centrifugation, two major bands between sucrose interfaces were collected, diluted twofold, and subjected to another centrifugation at 100,000 × g for 30 min at 4 °C. The pellet was resuspended in the homogenization buffer, and protein concentration was determined using a BCA Protein Assay Kit (Thermo Fisher Scientific).

Western Blotting

Samples (10 µg of total protein/lane) were denatured in Laemmli sample buffer (Bio-Rad) with β-mercaptoethanol, run on a 4–20% polyacrylamide gradient gel (Bio-Rad), and transferred to polyvinylidene difluoride membrane. Following the transfer, membranes were blocked in the blocking solution (Intercept (PBS) Blocking Solution, Licor) for 1 h at room temperature. Afterward, membranes were incubated with primary antibody, anti-pan NKA from Development Studies Hybridoma Bank, University of Iowa (1:5000, cat. no. AB2166869), overnight in PBS-T at 4 °C with gentle rocking. Blots were incubated with a 1:20,000 dilution of anti-mouse secondary antibody IRDye800 CW goat anti-mouse (Licor) for 1 h at room temperature and visualized using the Infrared Odyssey Imaging System (Licor). All uncropped western blots are included in the supporting information of this manuscript.

Statistics

Data acquired for Progressive acceptor photobleaching are represented as mean ± standard error of the mean (SEM) of N measurements. Data for NKA localization are represented by individual data points with a highlighted mean line. All data were tested for normality of distribution using the Shapiro–Wilk normality test in OriginPro2024b software (OriginLab). The statistical significance of localization data was evaluated using one-way ANOVA with the Dunn-Sidak post hoc test. The differences considered statistically significant had p < 0.05. TCSPC data did not show a normal distribution. The variable sample size disregarded data transformation; therefore, we tested the statistical significance using nonparametric Kruskal–Wallis ANOVA with p < 0.05.

Molecular Dynamic Simulations

We used our previously published model of the NKA dimer as an initial structure for the simulation of the G301A and G301R mutants [12]. We modeled NKA residues Glu334, Glu787, Asp811, and Glu961 (in α1 numbering) as protonated [64]; we further kept two 1.2-diacyl-sn-glycero-3-phosphocholine molecules and cholesterol embedded between the α and β subunits to preserve the structural stability of the E1 state (PDB: 3 WGU [47]); it is required for NKA function [70, 71]. For simplicity, we generated a symmetric bilayer without cholesterol. The complexes were inserted in a pre-equilibrated 150 × 150 Å bilayer of POPC:POPE:POPS lipids (2:1:1) to mimic the composition of the plasma membrane. POPS lipids and cholesterol were included because there is evidence that both are required to maintain the structural integrity of the C-terminal region of NKA [72]. Therefore, we expect that both symmetric and asymmetric membranes would yield similar results in terms of structural stability of NKA in the homooligomer form, but the reviewer is correct that the simple membrane structure is a limitation of this work. We used the replacement method to generate lipid packing around the αβ-FXYD2 dimers. For lipids, we used the LIPID21 force field [65] and a cutoff distance of 10 Å for non-bonded interactions. The initial systems were solvated using TIP3P water molecules with a minimum margin of 25 Å in the z-axis between the edges of the periodic box and the protein. Na+, K+, and Cl− ions were added to neutralize the system and to produce Na+ and K+ concentrations of 140 and 5 mM, respectively [66, 67]. Molecular simulations were performed with AMBER20 on Tesla V100 GPUs [68] using the AMBER ff19SB force field [69]. We maintained a temperature of 310 K with a Langevin thermostat and a pressure of 1.0 bar with the Monte Carlo barostat. We used the SHAKE algorithm to constrain all bonds involving hydrogens and allow a time step of 2 fs. The systems were subjected to 5000 steps of steepest-descent energy minimization followed by equilibration as follows: two 25-ps MD simulations using a canonical ensemble (NVT), one 25-ps MD simulation using an isothermal–isobaric ensemble (NPT), and two 5-ns MD simulations using the NPT ensemble. The equilibrated structures were used as a starting point to perform a triplicate MD trajectory of each complex for a total simulation time of 4.6 µs using an NPT ensemble. The equilibrated monomers were used to perform a single MD trajectory with a total simulation time of 1.9 µs. Table 1 summarizes the simulations replicates and simulation times.

Table 1 Summary of simulations and simulation timesMethod Limitations

The simulations used a pre-equilibrated membrane, so we did not rearrange lipid placements, but we randomized velocities and re-ran the trajectories. Our primary focus was the comparison of the WT and mutant structures, however, starting conditions, including lipid and ion positions, might influence the outcome.

Molecular Dynamic Simulations Data Analysis

To investigate structural changes upon mutations of G301, we performed quantitative analyses on the molecular dynamics simulations. RMSD, RMSF, tilt angle, Na+ binding, N-/N- distance, and water penetration were quantified using CPPTRAJ [73]. The tilt angle was defined as the angle between TM3 and the normal of the lipid membrane. Na+ binding was measured as the distance of individual Na+ to their respective binding sites (position I: A327, N329, E331, D809, D813; position II: A327, T777, N781, D813; position III: Y776, Q931, D934). N–N- distance was defined as the distance between the centers of masses of the cytosolic domain of the α subunit. Lastly, water penetration was quantified by computing the cylindrical distribution of water molecules. For this quantification, the center of the lipid bilayer was placed at the origin and all water molecules in the trajectories were calculated. Subsequently, we constructed a 2D histogram of water density along both the z-axis and the radial axis (axis perpendicular to the z-axis). We focused on the − 10 to + 10 Å region, where the most significant differences between the variants were observed. The lipid bilayer was positioned from − 20 to + 20 Å. To further condense the information, we defined a threshold of 30 Å along the radial axis and computed the total water density along the radial axis. The data were visualized and analyzed using PyMOL [74] and VMD software [75].