A novel proton transfer mechanism in the SLC11 family of divalent metal ion transporters

Abstract

In humans, the H⁺-coupled Fe²⁺ transporter DMT1 (SLC11A2) is essential for proper maintenance of iron homeostasis. While X-ray diffraction has recently unveiled the structure of the bacterial homologue ScaDMT as a LeuT-fold transporter, the exact mechanism of H⁺-cotransport has remained elusive. Here, we used a combination of molecular dynamics simulations, in silico pK _a calculations and site-directed mutagenesis, followed by rigorous functional analysis, to discover two previously uncharacterized functionally relevant residues in hDMT1 that contribute to H⁺-coupling. E193 plays a central role in proton binding, thereby affecting transport properties and electrogenicity, while N472 likely coordinates the metal ion, securing an optimally "closed" state of the protein. Our molecular dynamics simulations provide insight into how H⁺-translocation through E193 is allosterically linked to intracellular gating, establishing a novel transport mechanism distinct from that of other H⁺-coupled transporters.

Figure 1

Selected residues based on pK _a analysis and the charged cluster around E127. (A) The side-chain pK _a values of titratable residues were estimated in silico in ScaDMT, both in the presence and absence of the bound Mn²⁺ ion. Residues having a pK _a shift of more than 1.5 units from their canonical pK _a values in solution are shown. Additionally, H233, mutations of which have been shown to cause pH-independent transport in rat DMT1, is also included for reference. Residues highlighted in bold are shown in the close-up view on the lower half of panel B. (B) All residues on panel A are highlighted in gold on the ScaDMT structure. The close-up view of the substrate binding site and a charged residues cluster are shown below. The following regions are color-coded: TMH 1a stub (truncated N-terminus; orange), TMH 6b (green), TMH 8 (red). TMH 8 is shown as a transparent helix in the close-up view.

Figure 2

Helicity changes upon protonation in ScaDMT. (A) Time average of helicity of each residue as observed in simulations of ScaDMT in various protonation states, quantified using STRIDE. The indicated transmembrane regions were taken from the PDBTM database. Regions with most differences in helicity are shown, which are TMH 6 (top) and TMH 8 (bottom). (B) Helicity changes in each MD trajectory were also quantitated by calculating the time average of the sum of dihedral angles ψ _i + ϕ _i+1 along the protein chain. For an ideal α-helix, this value is ≈−105°. Large values around 200° correspond to local peptide bond flips.

Figure 3

Metal ion coordination and intracellular gate opening in ScaDMT. (A) A closed coordination shell of the metal ion substrate evolved in simulations. Compared to the conformation found in the X-ray structure, the backbone carbonyl oxygen of G46 and the side-chain of Q389 fill the remaining two positions in an octahedral coordination shell. (B) In the closed conformation (left panel), the side-chain of Y47 fills a cavity between TMHs 6 and 7. When gate opening is observed, M48 enters a hydrophobic pocket formed mostly by TMHs 5 and 7. (C) Intracellular gate opening is observed in 3 of 5 “canonical” simulations (left), but not when E127 (right) or any of the residues D49 and H233 is protonated. For clarity, the right panel shows only the results of simulations where E127 was protonated. Gate opening is also observed in 1 of 5 simulations where H228 is protonated (not shown in figure), evolving into a conformation similar to that of canonical #3 and #5.

Figure 4

Plasma membrane expression and substrate uptake by hDMT1 and mutants. (A) Representative blot of the plasma membrane expression of hDMT1 and the indicated mutants in transiently transfected HEK293 cells. Expression was assessed by immunoblotting of the labeled plasma membrane protein by sulfo-NHS-LC-biotin isolated on streptavidin agarose beds. (B) Biotin was used as equal loading control (the absence of actin confirms the purity of the membrane fraction). (C) Mean plasma membrane expression of each protein was obtained by averaging the optical density (OD) of the corresponding bands obtained form 3 independent blots. These values were normalized to the OD of its corresponding loading control. (D) Radiolabeled iron uptake (⁵⁵Fe²⁺) (1 μM) by HEK293 cells expressing the indicated proteins was measured in uptake buffer at pH 5.5 after 15 min incubation. (E) Cd²⁺ (1 μM) transport activity by HEK293 cells expressing the indicated proteins. Changes of fluorescence intensity of Calcium-5-dye were continuously recorded at extracellular pH 6.5 during 10 min incubation with Cd²⁺ and the uptake was determined as the area under the curve. The radioactive iron uptake (D) and the fluorescence intensity (E) of each isoform were corrected for the unspecific uptake by empty vector transfected HEK293 cells. Results from 3 independent experiments were normalized to Fe²⁺ (D) (N = 41–56) or Cd²⁺ (E) (N = 18–24) uptake by hDMT1 WT, corrected according to their plasma membrane expression and represented as the mean ± S.D. (F) Uptake of 20 μM ⁵⁵Fe²⁺ (pH 5.5, 10 min) in oocytes expressing hDMT1 and the indicated mutants. Uptake values were corrected for unspecific iron uptake in non-injected oocytes. Data from 6 different batch of oocytes were normalized to the mean iron uptake by hDMT1 WT (0.83 ± 0.3 to 2.79 ± 0.4 pmol∙oocyte⁻¹∙min⁻¹) and are represented as the mean ± S.D. (25 to 74 oocytes). Statistical differences (D–F) were assessed using T-test (Fe²⁺-uptake by hDMT1 WT vs each mutant); ns p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 5

Fe²⁺ and H⁺ saturation kinetics of WT hDMT1 and E193 and M294 mutants in stably transfected HEK293 cells. (A) Representative experiments of ⁵⁵Fe²⁺ uptake kinetics by hDMT1 and E193 and M294 mutants as a function of increasing Fe²⁺ concentrations (0.1–20 µM). (B) Representative experiment of ⁵⁵Fe²⁺ uptake kinetics by hDMT1 and E193 mutants as a function of increasing H⁺ concentration (pH 5.5–7.5). For each concentration of the substrate, the background uptake measured for the empty vector transfected HEK control cells was subtracted. Uptake values were also corrected according to their plasma membrane expression and represented as the mean ± S.D. A Michaelis-Menten equation (connecting lines) was fit to the data and the obtained kinetic parameters are summarized in a table below each graph. (C) Summary table of the apparent affinities for Fe²⁺ and H⁺ of hDMT1 and the indicated mutants. K _m values were calculated from 2–3 independent experiments and are presented as means ± S.D.

Figure 6

Steady-state and pre-steady state currents for WT hDMT1 and E193 mutants expressed in X. laevis oocytes. (A) Current-voltage relationship (V _h = −50 mV; pH 5.5; 20 μM Fe²⁺). (B) Q/V _m relationship (V _h = −50 mV; pH 5.5). Pre-steady state currents were integrated with time to obtain the charge Q and depicted at test potentials (V _t) ranging from −110 mV to +90 mV. Corresponding curves on panels A and B were recorded in the same oocytes, which were obtained from the same batch. Similar results were observed in 7–10 oocytes from 3 different batches. (C) The maximal current evoked by 20 µM Fe²⁺ (Fig. 1A) was taken as I _max. Pre-steady state parameters (Q _max, z, V _0.5 and r ²) were derived from data in panel B by fitting a Boltzmann equation (connecting lines). The number of functional transporters (N _T) expressed in the plasma membrane was determined as N _T = Q _max/z∙e (e = 1.6 × 10⁻¹⁹ C). Turnover rate was determined as −I _max/Q _max.

Figure 7

Fe²⁺ and H⁺ saturation kinetics of the steady-state currents of hDMT1 and E193D mutant expressed in X. laevis oocytes. (A) Representative experiment of Fe²⁺-evoked steady-state currents (V _h = −50 mV; V _t = −70 mV; pH 5.5) as a function of [Fe²⁺]. (B) Fe²⁺-transport K _0.5 as a function of membrane potential (V _m), each K _0.5 value corresponds to the mean ± S.D. calculated from 4–5 oocytes obtained from 2 different batches. (C) Representative experiment of Fe²⁺-evoked steady-state currents (V _h = −50 mV; V _t = −70 mV; 20 µM) as a function of [H⁺]. (D) Fe²⁺-transport K _0.5 as a function of membrane potential (V _m), each K _0.5 value corresponds to the mean ± S.D. calculated from 4–5 oocytes obtained from 2 different batches. Kinetic parameters of Fe²⁺-transport of the representative panels A and C were calculated by fitting a 4-parameter sigmoidal equation (connecting lines) to the data and are summarized below the corresponding graph.

Figure 8

pH-dependence of the pre-steady state currents of hDMT1 and E193D mutant expressed in X. laevis oocytes. Dependence of the Q/V _m relationship of hDMT1 WT (A) and E193D (B) mutant on [H⁺]. Pre-steady state currents were integrated with time to obtain the charge Q, and depicted at test potentials V _t (−10 mV to +90 mV) from the holding potential V _h = −50 mV at the indicated pH values. A Boltzmann equation was fit (connecting lines) to data in panels A and B to obtain the kinetic parameters (Q _max and V _0.5). (C) Dependence of Q _max on [H⁺]. For each individual oocyte, Q _max values were normalized to the Q _max determined at pH 5.5 (WT Q _max 81.6 to 97.7 nC; E193D Q _max 41.55 to 81.6 nC). Represented Q _max values correspond to the mean ± S.D. calculated from 5–13 oocytes obtained from 4 different batches. (D) V _0.5 as a function of [H⁺], each V _0.5 value corresponds to the mean ± S.D. calculated from 3–10 oocytes obtained from 4 different batches. Connecting lines in panels C and D were drawn only for illustration purposes.

Figure 9

Accessible rotameric states of E127 in ScaDMT. Cluster centroid conformations of the E127 side-chain observed in the “canonical” simulations (A) and when E127 is protonated (B). Accessibility of the extracellular and intracellular vestibule, cluster numbers, the χ ₁ and χ ₂ angles of the side-chain, and side-chain pK _a values in each conformation are indicated. The five simulation trajectories are shown color-coded for each protonation state (C, D), mapped into the space of the χ ₁-χ ₂ dihedral angles.

Figure 10

Proposed mechanistic model of the transport cycle of hDMT1. Transmembrane helices are shown color-coded: TMH 1a (orange), TMH 1b (dark blue), TMH 3 (light blue), TMH 8 (red), with TMH 3 located behind TMH 8. Divalent metal ion substrate is shown as yellow disc; proton as black disc. Water-accessible vestibules are indicated with dashed lines. Arrows indicate movements of components. (A) Substrate binding happens in the high-pK _a state of E193, likely in an ordered manner. After substrates bind, a conformational switch is proposed to occur, bringing the protein into an inward-open-like occluded state, where E193 flips over to a low-pK _a conformation. (B) In the low-pK _a conformation, proton dissociation occurs possibly along a pathway flanked by D221, D190 and R445 (D153, D124 and R360 in ScaDMT, respectively). (C) The loss of the proton induces three key conformational changes: the partial unwinding of TMH 8 near residues 410–417 (corresponding to residues 325–330 in ScaDMT), increased flexibility of the E193 side-chain and the opening of the intracellular gate. At this point, substrate-dependent uncoupled proton leak could happen at high proton concentrations by jumping back to state A, if transition to state D is slow enough. (D) After gate opening, the metal ion substrate is solvated and leaves the binding site. (E) Local helix unwinding in TMH 8 facilitates the recovery of E193 into the high-pK _a state, and closure of the intracellular gate happens. (F) The high-affinity outward-open state is recovered. Additionally, this mechanism also allows for the uniport of metal ions in the absence of protons through cycle A-D-E-F-A.