Calcium regulation of the human mitochondrial ATP-Mg/Pi carrier SLC25A24 uses a locking pin mechanism

Mitochondrial ATP-Mg/Pi carriers import adenine nucleotides into the mitochondrial matrix and export phosphate to the cytosol. They are calcium-regulated to control the size of the matrix adenine nucleotide pool in response to cellular energetic demands. They consist of three domains: an N-terminal regulatory domain containing four calcium-binding EF-hands, a linker loop domain with an amphipathic α-helix and a C-terminal mitochondrial carrier domain for the transport of substrates. Here, we use thermostability assays to demonstrate that the carrier is regulated by calcium via a locking pin mechanism involving the amphipathic α-helix. When calcium levels in the intermembrane space are high, the N-terminus of the amphipathic α-helix is bound to a cleft in the regulatory domain, leading to substrate transport by the carrier domain. When calcium levels drop, the cleft closes, and the amphipathic α-helix is released to bind to the carrier domain via its C-terminus, locking the carrier in an inhibited state.

up a net exchange of charge using an ATP/ADP hetero-exchange to characterise the effects of calcium, magnesium and membrane potential on substrate transport carried out by the carrier.
The ∆1-191_APC1 was expressed, purified and reconstituted into liposomes and ATP uptake was measured in saturating conditions (5 mM) of magnesium chloride, calcium chloride or EGTA on the outside, in exchange for ADP loaded on the inside ( Supplementary Fig. 6). Without the addition of divalent cations to the outside of liposomes, the specific initial uptake rate of ATP was low (~5 nmol min -1 mg -1 protein). This low rate of ATP for ADP exchange was not due to the build up of membrane potential by the exchange of a minus 3 species for minus 2 species, as there was no significant difference in rate in the presence of valinomycin and potassium salts, which will dissipate the membrane potential ( Supplementary Fig. 6). The specific initial uptake rate of ATP was highest in the presence of externally added magnesium, increasing six-fold to ~30 nmol min -1 mg -1 protein ( Supplementary Fig. 6). The addition of calcium also increased the specific initial uptake rate of ATP three-fold to ~15 nmol min -1 mg -1 protein ( Supplementary Fig. 6). These results confirm that uptake of ATP by APC has a requirement for ATP chelated to magnesium, but also suggest that ATP chelated to calcium is also preferably transported over free ATP. Supplementary Fig. 1 Purification and reconstitution of APC1 a) 12 % acrylamide gel of APC1 purification: 1, Isolated yeast mitochondria; 2, pellet after ultracentrifugation (unsolubilised protein); 3, soluble protein in detergent; 4, proteins that do not bind to Ni-sepharose; 5, protein bound to Ni-sepharose; 6, protein and Ni-sepharose resin after factor Xa treatment; 7, protein that remains bound to resin after separation through a centrifugation filter; 8, protein that is eluted after factor Xa treatment (purified carrier). b) Quantification of APC1 by SDS-PAGE (inset): 1, total protein at the start of reconstitution; 2, protein after the removal of detergent and desalting; 3, protein not associated with liposomes after high speed spin; 4, protein associated with liposomes after high speed spin, indicating that approximately 80 % of the protein was incorporated. Error bars represent the standard deviation from three independent gel lanes. Error bars represent the standard deviation for measurements taken from three independent liposome preparations. b) The relationship between initial ATP uptake rate and the calcium concentration was plotted using the specific initial rates from panel A and fitted with a curve describing hyperbolic saturation kinetics. c) Uptake curves from APC1 proteoliposomes loaded with 2.5 mM ATP, in the presence of 1 mM calcium chloride and increasing concentrations of EGTA (0, 0.5, 1, 2, 2.5, 3 and 5 mM). Measurements were taken from a single liposome preparation. d) Specific initial rates of ATP uptake plotted as a function of EGTA concentration. Specific initial rates and the EC 50 value were estimated using least-squares curve fitting of the data by Prism (GraphPad). e) Standard curve of fluorescent signal (relative fluorescent units; RFU) from the low affinity calcium probe fluo-5N at 0.2, 1, 5, 10, 30, 200, 1000 µM calcium chloride. f) Data from panel b corrected to free-calcium concentration by interpolation of fluorescent signal against standard curve in panel e. Results from three independent liposome preps have been plotted individually and fitted with a curve describing hyperbolic saturation kinetics. The estimated V max for ATP transport plotted with circles, squares and triangles was 214, 218 and 191 (mol min -1 mg -1 ), respectively, and the estimated EC 50 for calcium was 239, 191 and 107 (µM), respectively.

Supplementary Fig. 4 Kinetics of ATP transport at two calcium concentrations.
Uptake from APC1 proteoliposomes loaded with 2.5 mM ATP, with increasing concentrations of ATP (1.5, 6, 12, 20, 38, 60, 80 and 100 µM) added to the outside and in the presence of either a) 0.25 mM or b) 4 mM calcium chloride. All liposomes were loaded with 1 mM DTT and 20 mM tris-HCl pH 7.4. For ATP concentrations above 10 μM, [ 14 C]-ATP was mixed in a decreasing ratio with unlabelled ATP and the uptake rate was adjusted accordingly. c) The Michaelis-Menten relationship between initial ATP uptake rate and the ATP concentration was plotted for 0.25 mM (filled squares) and 4 mM (open squares) concentrations of external calcium using the specific initial rates from panel a and b, respectively. Error bars represent the standard deviation for measurements taken from three independent liposome preparations. Specific initial rates, K m and V max values were estimated using least-squares fitting of the data by Prism (GraphPad). SDS-PAGE gel analysis of PEG-maleimide 2K reacted A) wild-type APC1 and B) C15S mutant APC1, in the presence or absence of SDS. Samples were taken and quenched using DTT, before being mixed with SDS-PAGE gel loading buffer. Samples were run on an SDS-PAGE gel. For the PEG-maleimide reacted wild-type protein treated with SDS, 4 PEG modified cysteine residues can be observed, whereas in the C15S mutant, only 3 react as expected. In the non-SDS treated samples, one cysteine residue is modified, whereas in the C15S mutant, no cysteine residues are modified.

Supplementary Fig. 6 Effect of magnesium and calcium on ATP/ADP hetero-exchange by the carrier domain of APC1 alone.
The N-terminal truncation (∆1-191_APC1) construct was reconstituted into liposomes and the amount of 1.5 µM [ 14 C]-ATP taken up, measured. All liposomes were loaded with 2.5 mM ADP, 1 mM DTT and 20 mM tris-HCl pH 7.4. Either 5 mM magnesium chloride, 5 mM calcium chloride or 5 mM EGTA was added to the outside of the liposome, and hetero-exchange between ATP and ADP was monitored over a 15-minute time course. The KCl concentration was 50 mM on the inside and outside of the liposomes and uptake in the presence and absence of valinomycin was tested in order to see what effect membrane potential had on the rate of the ATP/ADP exchange. Specific initial rates were estimated using least-squares fitting of the data by Prism (GraphPad). Statistical tests were calculated using unpaired, two-tailed Student's t-tests; NS P > 0.05; ** P ≤ 0.01. Alignments of APC orthologues from mammal, invertebrates, plants and fungi were prepared and the linker regions in the regulatory domain and the cytoplasmic loops of the carrier domain in APC orthologues are presented. In order to identify conserved residues that might mediate interactions between the regulatory domain and the carrier domain, APC sequences were compared with either human calmodulin or human ADP/ATP carrier isoform-1 (AAC-1), which highlight residues that are important for calcium binding in the regulatory domain or substrate transport in the carrier domain, respectively, and therefore not part of the interaction between the regulatory and carrier domains.
Positions where charged residues are conserved, but not indicated to be involved with calcium binding or substrate transport and have been mutated to alanine in this study have been labelled according to the APC1 sequence. Residues are coloured according to the percentage of the residues in each column that agree with the consensus sequence as in the key. Only the residues that agree with the consensus residue for each column are coloured. Image of alignment prepared using Jalview 4 .

Supplementary Fig. 10 Effect of mutations on the uptake activity of APC1.
Specific initial uptake rate of 1.5 μM [ 14 C]ATP into proteoliposomes was measured. Proteoliposomes contained either wild-type or one of five mutant APC1. Each proteoliposome contained 2.5 mM ATP, 1 mM DTT and 20 mM tris-HCl pH 7.4, whereas the external buffer contained either 2.5 mM calcium chloride (light grey) or 2.5 mM EGTA (dark grey). Results were from three independent liposome preparations. Specific initial rates were fitted and significance values were calculated using unpaired, two-tailed Student's t-tests, where; NS p > 0.05; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001. The significance between mutant and wild-type specific initial uptake rates in the presence (light grey dashed line) or absence of calcium (dark grey dashed line) are displayed above each bar. The significance between the specific initial uptake rate in the presence or absence of calcium is displayed above half-tick lines in each case. The final concentration of lipid in the assay was 0.5 mg mL -1 . The final protein concentration in each liposome sample was quantified by SDS-PAGE analysis, and compared to a standard curve of known protein concentration.