A Photo-clickable ATP-Mimetic Reveals Nucleotide Interactors in the Membrane Proteome

Abstract

ATP is an important energy metabolite and allosteric signal in health and disease. ATP-interacting proteins, such as P2 receptors, control inflammation, cell death, migration, and wound healing. However, identification of allosteric ATP sites remains challenging, and our current inventory of ATP-controlled pathways is likely incomplete. Here, we develop and verify mipATP as a minimally invasive photoaffinity probe for ATP-interacting proteins. Its N⁶ functionalization allows target enrichment by UV crosslinking and conjugation to reporter tags by "click" chemistry. The additions are compact, allowing mipATP to completely retain the calcium signaling responses of native ATP in vitro and in vivo. mipATP specifically enriched for known nucleotide binders in A549 cell lysates and membrane fractions. In addition, it retrieved unannotated ATP interactors, such as the FAS receptor, CD44, and various SLC transporters. Thus, mipATP is a promising tool to identify allosteric ATP sites in the proteome.

Keywords: ATP; ATP photoaffinity probe; calcium signaling; chemical proteomics; photoaffinity labeling; purinergic signaling; quantitative proteomics; target discovery.

Figure 1.. mipATP reconstitutes purinergic Ca²⁺ responses in vitro and in vivo

(A) Structure of the minimally invasive photoaffinity ATP probe (mipATP). Minimalist photo-clickable handle highlighted in blue and red. (B) Single cell measurements of extracellular ATP-mediated Ca²⁺ signals in HeLa cells expressing a nuclear targeted, ratiometric GCaMP6s transgene (also see Figure S1). Left, boxplot indicating the median frequency of Ca²⁺ transients per cell over 20 minutes after exposure to 10 µM ATP or ATP analog in HeLa cells expressing a GCaMP6s-3xNLS-P2A-mBeRFP-3xNLS transgene. On each box, the central mark (red) indicates the median, and the bottom and top edges of the box (blue) indicate the 25^th and 75^th percentiles, respectively. The whiskers extend to the most extreme data points not considered outliers, and the outliers are plotted with the ‘+’ symbol (red). Right, average Ca²⁺ signal of HeLa cells after exposure to 10 µM ATP or ATP analog as determined by the nuclear GCaMP6s/mBeRFP fluorescence ratio. Data represent the mean GCaMP6s/mBeRFP fluorescence ratio and the bottom and top edges of the shaded region indicate the 25^th and 75^th percentiles, respectively. Data was compiled from (n) number of cells per condition from at least two independent experiments for each nucleotide: ATP n=211, ATPγS n=198, mipATP n=336, AMP-PNP n=104, and Vehicle n=308. Asterisks, unpaired, two-tailed t tests between the ATP treated cells versus the indicated ATP analog; ****p < 0.0001, ns = not significant. (C) Ratiometric imaging of ATP-mediated Ca²⁺ transients in zebrafish larvae tail fins (also see Movie S1). Left, scheme of experimental setup. Tail fin tips of 2.5–4 days post-fertilization (dpf) transgenic Ca²⁺ reporter zebrafish larvae Tg(krt4:QF2; QUAS:GCaMP6s-3xNLS-P2A-mKate2–3xNLS) are amputated and wounded tail fins are subsequently perfused with a bolus of 1 mM ATP or ATP analog at 2 minutes into imaging. Right, ratiometric imaging of Ca²⁺ transients induced by perfusion of wounded tail fins with 1 mM ATP or ATP analog. Nuclear GCaMP6s signals are normalized by nuclear mK2 (mKate2) fluorescence. Scale bars, 50 µm. Time in minutes. (D) Average spatiotemporal Ca²⁺ signal profiles of the indicated (n) number of transgenic Ca²⁺ reporter larvae Tg(krt4:QF2; QUAS:GCaMP6s-3xNLS-P2A-mKate2–3xNLS) after perfusion with 1 mM ATP or ATP analog at 2 minutes from the start of imaging. Please also see Figure S1.

Figure 2.. mipATP reversibly interacts with A549 cell membrane proteins

(A) A549 cell membrane fractions were incubated with 50 µM mipATP or vehicle ± UV-irradiation (365 nm) prior to click chemistry with TAMRA-azide. Samples were then separated by SDS-PAGE and gels were scanned for TAMRA fluorescence (25 µg sample per lane). Representative SDS-PAGE in-gel fluorescence scan of three individual experiments. (B) A549 membrane fractions were incubated with increasing concentrations (1–50 µM) of mipATP then UV-irradiated (365 nm) prior to click chemistry with TAMRA-azide. Samples were separated by SDS-PAGE and gels were scanned for TAMRA fluorescence (25 µg sample per lane). Left, representative SDS-PAGE in-gel fluorescence scan. Right, analyses of photo-crosslinking efficiencies of mipATP to the selected proteins based on normalized TAMRA-fluorescence signal from the isolated bands. Data represent mean ± SEM from three independent experiments. EC₅₀ values (µM) were determined by fitting the data to a dose-response curve. Parenthesis, 95% confidence interval (CI). Band 1 EC₅₀ = 6.18 (0.55–36.47) µM, Band 2 EC₅₀ = 6.32 (2.23–16.57) µM, Band 3 EC₅₀ = 9.67 (1.12–146.2) µM. (C) A549 membrane fractions were pre-incubated with increasing concentrations of ATP or ATPγS (0, 20, 200, or 2,000 µM) and then probed with 20 µM mipATP. Next, samples were UV-irradiated (365 nm) prior to click chemistry with TAMRA-azide. Samples were separated by SDS-PAGE and gels were scanned for TAMRA fluorescence (25 µg sample per lane). Left panels, representative SDS-PAGE in-gel fluorescence scans. Right panels, analyses of selected protein TAMRA-fluorescence. The fluorescence intensity value of each band was normalized to the value of the corresponding no competition (0 µM) band within the same experiment. Data represent the mean ± SEM from three independent experiments. Asterisks, unpaired, two-tailed t tests between the indicated sample and the corresponding no competition (0 µM) sample; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Please also see Figure S2.

Figure 3.. Mass spectrometry maps the mipATP membrane interactome

(A) Intact A549 cells were probed with 50 µM mipATP alone or in the presence of 100 times excess of unmodified ATP. Labeled proteins were identified by SILAC-based quantitative proteomics. ATP-quenchable mipATP-protein interactions were ranked according to SILAC ratios. Plot of log₂ values of the SILAC ratios for proteins identified in the “forward” and “reverse” experiments. Dashed lines mark the 90^th percentile threshold ratio values. Proteins with SILAC ratios in the 90^th percentile of both “forward” and “reverse” experiments were considered high confidence mipATP interactors and are indicated in the top right quadrant highlighted as black dots in a purple background. Data are representative of one “forward” and one “reverse” experiment. Please also see Figures S3D, S3E and Table S1. (B) Left, 73 proteins were identified as high confidence mipATP interactors in at least two of three mass spectrometry experiments (Micallef and Rodgers, 2014). Right, classification of the high confidence mipATP interactors based on their predicted likelihood of binding mipATP according to GO and Uniprot database analyses. Please also see Figure S3F and Table S1. (C) High confidence interactors identified in ≥5% of CRAPome database experiments and/or similar proteomic studies using fatty acid and small molecule photo-clickable probes (Niphakis et al., 2015; Parker et al., 2017a) were filtered out resulting in 32 high confidence mipATP binders. Please also see Table S1.

Figure 4.. Verification of known and unannotated ATP interactors

(A) Verification of mipATP interactions with high confidence mipATP targets known to bind ATP. Isolated A549 membrane fractions were probed with 10 µM mipATP after preincubation with 100 µM ATPγS or ATP competitor or vehicle ± UV-irradiation (365 nm) prior to click chemistry with biotin-azide. Biotin-conjugated proteins were purified with streptavidin agarose and separated by SDS-PAGE. MRP1- (left), CK1δ- (middle), and CD73-mipATP (right) complexes were detected by western blot with anti-MRP1, anti-CK1δ, and anti-CD73 antibodies, respectively. Left panels, representative western blots. Inputs are from 5 µg of membrane protein extract prior to pulldown. Right panels, analyses of protein-mipATP complex signal intensities normalized to the value of the corresponding complex in the absence of competition + UV-irradiation within the same experiment. Data represent mean ± SEM of at least three independent experiments (four for MRP1 and CD73). (B) Western blot with anti-P2X4 antibody detects P2X4-mipATP complexes purified from isolated A549 membrane fractions probed with mipATP (10 µM) in the presence of the indicated ATP derivative (100 µM) or vehicle ± UV. Right, analysis of P2X4-mipATP signal intensities. P2X4-mipATP signal values were normalized to the value of the complex in the absence of competition + UV within the same experiment. Data represent mean ± SEM of two independent experiments. (C) Western blot with anti-CD44 antibody detects CD44-mipATP complexes purified from isolated A549 membrane fractions probed with mipATP (10 µM) in the presence of the indicated competitor (100 µM) or vehicle ± UV. Right, analysis of CD44-mipATP signal intensity values normalized to the value of the complex in the absence of competition + UV within the same experiment. Data represent mean ± SEM of two independent experiments. (D) Membrane fractions from A549 cells stably expressing recombinant MFSD5-FLAG were probed with mipATP (10 µM) in the presence of the indicated ATP derivative (100 µM) or vehicle ± UV. Purified MFSD5-mipATP complexes were detected by western blot with an anti-FLAG antibody. Right, analysis of MFSD5-mipATP signal intensity values normalized to the value of the complex in the absence of competition + UV within the same experiment. Data represent mean ± SEM of three independent experiments. (E) Membrane fractions from A549 cells stably expressing recombinant MFSD5-FLAG were probed with mipATP (10 µM) in the presence of the indicated competitor (100 µM) ± UV. Purified MFSD5-mipATP complexes were detected by western blot with an anti-FLAG antibody. Right, analysis of MFSD5-mipATP signal intensity values normalized to the value of the complex in the absence of competition + UV within the same experiment. Data represent mean ± SEM of three independent experiments. (F) Western blot with anti-FAS antibody detects FAS-mipATP complexes purified from isolated A549 membrane fractions probed with mipATP (10 µM) in the presence of the indicated ATP derivative (100 µM) or vehicle ± UV. Right, analysis of FAS-mipATP signal intensities normalized to the value of the complex in the absence of competition + UV within the same experiment. Data represent mean ± SEM of three independent experiments. (G) Western blot with anti-FAS antibody detects FAS-mipATP complexes purified from isolated A549 membrane fractions probed with mipATP (10 µM) in the presence of the indicated nucleotide/nucleoside (100 µM) ± UV. Right, analysis of FAS-mipATP signal intensities normalized to the value of the complex in the absence of competition + UV within the same experiment. A lane for molybdate competition (right of the UTP competition lane) was removed from the blot image because it is irrelevant to this dataset. Data represent mean ± SEM of three independent experiments All inputs are from 5 µg of solubilized membrane protein extract prior to pulldown. Please also see Figure S3G. Asterisks, unpaired, two-tailed t tests between the indicated sample and the corresponding no competition + UV-irradiation sample; *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns = not significant.