Sodium binding and GAT1 electrostatics

a) Averaged density of sodium is shown in close proximity to GAT1 in a side view. The GAT1 surface is shown in white, the transmembrane helices TM1, TM2, and TM6 are highlighted in grey. The residues (R69 and D451) forming the salt bridge of the extracellular gate (EC gate) are shown as grey sticks, the recruiting site residues D281 and E283 as cyan sticks. The sodium density is represented as isovolume and coloured according to the legend in panel c. The trajectories were fitted to the Cα atoms of the transmembrane domain b) Top view of GAT1, showing TM1, TM2 and TM6, the regions of high density of sodium and the recruiting site residues, coloured according to the legend in panel c. Sodium density at the temporary site is indicated by TMP. c) Same as in panel b, but data are from the double alanine (D281A and E283A) mutant simulations. The scale bar to the right explains the colour code of sodium density. d) Positions of sodium ions binding directly to NA1 (green), e) directly NA2 in purple or f) first residing in the temporary binding site (orange) before reaching NA1 or NA2. g) Quantification of sodium binding events, calculated as cumulative sodium ions entering the vestibule. h) Top view and i) side view of the electrostatic potential of GAT1 wild-type mapped onto its surface. j) Top and k) side view of the electrostatic potential of the double alanine mutant (D281A and E283A).

Effect of D281A-E283A mutation:

a) Representative confocal microscopy images of GAT1 wild-type and double alanine mutant (D281A-E283A) expressed in HEK293 cells. Transporters were N-terminally tagged with YFP, the cell membrane was stained by trypan blue. Composite images display the superimposition of tagged GAT1 and trypan blue stained membranes. The inserts in the lower left corner show a zoom-in on the cell membrane. b) Mean fluorescence of cytosolic and membrane resident YFP tagged GAT1. c) Concentration-dependent [3H]GABA uptake in HEK293 cells stably expressing GAT1 wild-type and D281A-E283A. d) Quantification of the Km (µM) and the Vmax (pmol/106 cells/min) of the GABA uptake. Data is displayed as mean ± SD. Concentration dependent uptake was measured in triplicates in six biologically independent experiments. Single data points for Km and Vmax are the averages for individual experiments.

Effect of D281A-E283A mutation:

a) Representative traces of whole-cell patch clamp recordings showing GAT1 wild-type mediated currents in response to different GABA concentrations. b) Concentration dependent average steady state currents. c) Representative traces of whole-cell patch clamp experiments showing concentration-dependent GAT1 D281A-E283A mediated currents (data points are shown as running average (n=11) traces). d) Average steady state currents plotted against the used GABA concentration. e) Comparison of concentration dependent average steady state currents of GAT1 wild-type and double mutant, normalised to the respective currents observed at 300 μM GABA. f) Normalised sodium-dependent [3H]GABA uptake curve. g) Sodium EC50 values for GAT1 wild-type and the double mutant D281A-E283A observed in panel f. Data is shown as mean ± SD from 5 biological independent replicas of [3H]GABA uptake experiments measured in triplicates. In case of electrophysiological measurements, at least 7 independent recordings from 7 independent cells were recorded.

Short-range effect of sodium binding:

a) GAT1 structure represented as a cartoon, highlighting the ion binding sites b) Visualisation of the sodium binding sites NA1 (green) and c) NA2 (purple). The dashed lines indicate the distances used for determining the compactness NA1 and NA2. d) Compactness of NA1 and e) NA2 before (in grey) and after sodium binding (coloured according to panels a and b) The vertical dashed lines indicate the measure in the cryo-EM structure (PDB ID: 7Y7W), where GAT is pictured in a holo inward-occluded conformation. f) Reference orientation of GAT1 as used for panel g. g) Top view projection of sodium ion positional density after binding to NA1 and NA2. The canonical binding positions of Na1 and Na2 are indicated by stars. h) Free energy landscape of sodium binding, where coordinates are projected along IC1 and IC2, respectively.

Global structural effects of sodium binding to NA2:

Scatter plot of principal component 1 (PC1) against PC2 of all simulations that showed NA2 binding. Trajectories are divided into a) before and b) after Na2 binding. The pink triangle corresponds to the outward-open AlphaFold model of GAT1. All trajectories are projected to the same PC1 and PC2 axis. (c,d) Root Mean Square Fluctuations (RMSF) of motions along PC1 mapped as β-factors on the TM helices of GAT1 for the frames c) before Na2 binding and d) after Na2 binding. (e, f) Using the same orientation as in panel c, superpositioning of 1 ns separated frames with the maximum likelihood method. Structures are coloured according to the degree of correlated motions along PC1 for frames e) before Na2 binding and f) after Na2 binding. g) Density distribution of the distances Y241-F51 and Y241-F227 visualised in panel h, where values before Na2 binding are shown in grey, and after Na2 binding are coloured according to the legend. The vertical dashed lines indicate the distance measured in the cryo-EM structures of the GAT1 holo occluded and apo inward-open conformations (PDB ID: 7Y7W, 7Y7V) (Zhu et al., 2023) and in the AlphaFold outward-open model h) The distances shown in panels g are indicated in orange between TM1a (F51) and TM5a (Y241), and in yellow between TM4 (F227) and TM5b (Y241).

The substrate GABA completes the sodium binding site NA1:

a-c) Comparison of the sodium binding site NA1 in a) the cryo-EM holo inward-occluded GAT1 (PDB ID: 7y7w (Zhu et al., 2023), b) AlphaFold holo outward-open GAT1, c) and SERT. The group donating the sodium-coordinating carboxylate moiety is GABA in GAT1 (a, b) and D98 in SERT c). d) Zoom onto S1 showing as isomesh surfaces the regions in which sodium ions Na1 and Na2 are found in the four 0.5 μs long simulations in the absence of GABA. e) Time course and averaged histogram of the compactness of NA1. The vertical dashed line highlights the measured NA1 compactness in the holo inward-occluded GAT1 cryo-EM structure (PDB ID: 7Y7W (Zhu et al., 2023) f,g) Same as in panels d, and e, but in the presence of GABA.

The compactness of the binding sites.

The compactness was quantified by first determining the center of mass (CoM) between coordinating atoms of the sodium binding sites NA1 and NA2, respectively, and by measuring the average distances between the CoM and the coordinating atoms. (a) Represents the time evolution of NA1 and NA2 compactness, where the time point of sodium entering NA1, NA2, or the temporary site (TMP) is indicated by a vertical line, colored according to the legend.

Principal components analysis of NA1.

We concatenated the trajectories and conducted a PCA analysis. The covariance matrix was calculated using all trajectories, focusing on the atoms of the transmembrane helices (49:75, 64:76, 80:109, 121:155, 212:230, 282:309, 316:348,321345, 377:411, 420:439). The transmembrane helices C-alpha atoms were used as the fitting group. Subsequently, we divided the trajectories into two groups based on Na1 or Na2 binding and projected them onto the principal components previously calculated. these are shown as scatterplots in (a,b), where the pink triangle corresponds to the NaCl-bound protein conformation.

Markovian behavior validation.

(a) show the result of the Chapman-Kolmogorov (CK) test using a lag time equal to 25 ns and assuming four macrostates. The predicted values obtained from our MSMs are shown by the solid lines while the estimated values for longer lag time are shown by the dashed traces. The superposition of predicted and estimated values indicates that the MSM assures the Markovian behavior. (b) displays the time traces of the relation between the lag time and the implied timescales (or relaxation time) associated with the ten slowest processes, with the top blue trace indicating the slowest process. The solid lines refer to the maximum likelihood result and the dashed lines show the ensemble mean computed with Bayesian sampling procedure. The black line with the grey area indicates the timescale threshold where the MSM cannot resolve processes. In both panels, the non-grey areas indicate 95% confidence intervals computed using the Bayesian sampling procedure mentioned above.