Affinity-guided labeling strategy for P2X7.

(A) Schematic representation of the P2X7 labeling strategy using ligand-directed N-cyanomethyl NASA chemistry. A P2X7 ligand (Lg) binds to allosteric sites on the receptor, bringing the N-cyanomethyl NASA warhead into close proximity with endogenous lysine (K) residues. This spatial arrangement enables the covalent transfer of a biotin moiety via amide bond formation, resulting in highly specific biotinylation of P2X7. The biotin tag allows super-resolution imaging of nanoscale P2X7 localization using a Streptavidin-Alexa 647 probe (Strept-A 647). Notably, following biotin transfer, the ligand can dissociate from P2X7, leaving the allosteric sites unoccupied. (B) Crystal structure of panda P2X7 (pdP2X7) shown in ribbon representation, bound to AZ10606120 depicted as spheres (PDB:5U1W) (Karasawa & Kawate, 2016). One of the three ATP-binding sites and the approximate location of the membrane are also indicated. Inset, enlarged view of the AZ10606120-binding pocket, rotated 180°. Distances (in Å) between the α-carbon of selected lysines and the hydroxyl group of AZ10606120 are displayed. Note that K300 is not visible in this view. (C) Chemical structure of X7-uP.

X7-uP is a potent P2X7 inhibitor that rapidly labels ectopically expressed P2X7 in HEK293T cells.

(A) Whole-cell currents evoked by 10 μM BzATP are reversibly inhibited by co-applying 1 μM X7-uP (upper trace) or 1 μM AZ10606120 (middle trace) in cells transiently transfected with rP2X7. Inhibitors were pre-applied alone for 10 s before 2 seconds of co-application. In the control (absence of inhibitors), BzATP-induced currents further increased, demonstrating current facilitation, which is an expected feature of P2X7 activation (Surprenant et al, 1996). (B) Summary of whole-cell inhibition at the indicated concentrations (n = 4-8 cells for each condition). Bars represent mean ± s.e.m. (C) Western blot analysis of P2X7 labeling by X7-uP. Cells transiently transfected with P2X7c-myc were treated with 1 μM X7-uP for 0-60 minutes, in the absence or presence of 10 μM AZ10606120 or 10 μM A740003 (as indicated), followed by extensive washing. After cell lysis, biotinylated proteins were pulled down, separated on SDS-PAGE, and Western blotting was revealed using an anti-c-myc antibody (@c-myc). Molecular mass markers are shown on the right. Control of P2X7c-myc expression is presented in the corresponding input. β-Actin was used as a loading control. (D) Time course plot of P2X7 labeling with 1 μM X7-uP. Data (mean ± s.e.m., n = 3 independent transfections) were fitted with Eq. (1) to determine the pseudo-first-order reaction rate kapp (mean ± s.e.m.).

X7-uP labeling is highly selective for P2X7.

(A-B) Confocal images of HEK293T cells transiently transfected with either P2X7-mScarlet (A) or various P2X subunits tagged with GFP (P2X1-GFP, P2X2-GFP, P2X3-GFP, P2X4-GFP, P2X5-GFP, or P2X6-GFP) (B) were labeled with X7-uP and revealed using Strept-A 647 (red) in FBS-free DMEM. Labeling was performed in the presence of 10 μM AZ10606120 or 10 μM A740003 (A). Nuclei were stained with Hoechst (blue). For clarity, mScarlet and GFP signals are displayed in orange and green, respectively. Scale bars, 10 μm. (C) Quantification of Alexa 647 fluorescence. Bars represent mean ± standard deviation (s.d.) (n = 75-129 cells, t-test comparisons to P2X7-mScarlet, ****P < 0.0001).

X7-uP labels K82 and K117 in rat P2X7.

(A) Molecular docking of 1 (same pose as shown in Figure 1 — figure supplement 1C) in pdP2X7, showing distances (in Å) between the reactive carbonyl of 1 (stick representation) and selected α-carbons of nearby residues (blue). Residues shown in parentheses correspond to equivalent rP2X7 residues. (B) Confocal images of HEK293T cells transiently transfected with different P2X7 constructs: P2X7-mScarlet, K82A, K117A, and K82A/K117A. Refer to the legend of Fig. 3 for additional details. Scale bars, 10 μm. (C) Quantification of Alexa 647 fluorescence. Bars represent mean ± s.d. (n = 90-190 cells, t-test comparisons to indicated conditions, ****P < 0.0001). (D) Whole-cell currents evoked by 10 μM BzATP are reversibly inhibited by co-application of 0.5 μM X7-uP (upper trace) to BzATP in a cell transiently transfected with the double mutant K82A/K117A. The control (absence of X7-uP) is shown in the bottom trace. (E) Summary of whole-cell inhibition for K82A/K117A (n = 7 cells for X7-uP and 5 cells for control). Bars represent mean ± s.e.m.; Mann-Whitney test (**P < 0.005).

dSTORM data revealed nanoscale P2X7 plasma membrane localization in BV2 cells.

(A) Cartoon and experimental timeline of BV2 cell treatments. IL-1β release was assessed in the supernatant (sup), and the same cells were labeled with 1 mM X7-uP after extensive washout. (B) Quantification of IL-1β release by ELISA following the indicated treatments: LPS (1 μg/mL for 24 hours), ATP (1 mM for 30 minutes), BzATP (300 μM for 30 minutes), and MβCD (15 mM for 15 minutes). Bars represent mean ± s.e.m. (n = 12 samples from 3 independent experiments). Data were compared using Kruskal-Wallis followed by Dunn’s multiple comparisons (*P = 0.0208, #P = 0.0362, **P = 0.0014, ****P < 0.0001). (C) Normalized quantification of IL-1β release induced by LPS+ATP or LPS+BzATP in the presence of P2X7 inhibitors AZ10606120 or A740003. Bars represent mean ± s.e.m. (n = 6 samples from 6 independent experiments). One-way ANOVA with Dunnett’s multiple comparisons to control condition for ATP data (****P < 0.0001). Kruskal-Wallis followed by Dunn’s multiple comparisons to control condition for BzATP data (*P = 0.0414, *** P = 0.0006). (D) Bright-field and dSTORM images of X7-uP-labeled BV2 cells revealed with Strept-A 647 corresponding to experiments shown in panel B. Scale bars, 10 μm. Insets: Magnified dSTORM images. Scale bars, 1 μm. (E) Quantification of single P2X7 localization density. Bars represent mean ± s.e.m. (each data point represent a cell, n = 3 independent experiments). One-way ANOVA with Tukey’s multiple comparisons (*P < 0.019, #P < 0.0194, @P < 0.0477, ***P < 0.0002, ###P < 0.0008, ****P < 0.0001). (F) Relative frequency of cluster size. Inset: percentage of clusters larger than 0.025 mm2. (G) Number of detections per cluster. Bars represent mean ± s.e.m. One-way ANOVA with Tukey’s multiple comparisons (*P < 0.0197, #P < 0.0277, @P < 0.0389, ****P < 0.0001). (H) Images showing tessellation analysis of cells treated either with LPS+ATP or left untreated. Inset: magnification. Scale bars, 200 nm. (I) Number of fluorophores per cluster. Bars represent mean ± s.e.m. One-way ANOVA with Tukey’s multiple comparisons (****P < 0.0001).

Nanoscale redistribution of individual P2X7 receptors in microglia under pro-inflammatory conditions at the plasma membrane.

The cartoon illustrates two distinct clusters of P2X7 receptors (blue), each adorned with one, two, or three fluorescently tagged tetrameric biotin-bound streptavidin (red). In untreated cells, each cluster contains an average of 1.5 fluorophores per P2X7 receptor. Treatment with LPS and ATP promotes P2X7 clustering by increasing the average number of fluorophores per cluster to between 4 and 5, resulting in an increased number of P2X7 receptors per cluster, from one to three. This redistribution synergistically triggers IL-1β release.

Mapping the X7-uP labeling site.

(A) Sequence alignment of selected regions from P2X receptors: panda P2X7 (pdP2X7), rat P2X7 (rP2X7), mouse P2X7 (mP2X7), human P2X7 (hP2X7), and rat and human P2X1-6 (rP2X1-6, hP2X1-6). Lysine residues are highlighted in red. Secondary structure elements are depicted in carton representation and are labeled as in the crystal structure of pdP2X7 (Karasawa & Kawate, 2016). (B) Chemical structures of compound 1 and AZ10606120. Red star indicates electrophilic carbon of the N-cyanomethyl NASA reacting group. (C) Right, crystal structure of AZ10606120 (stick representation) bound to pdP2X7 (PDB:5U1W) (Karasawa & Kawate, 2016). Left, same view with 1 docked to pdP2X7 (left). Distances (in Å) separating the reactive carbonyl of 1 and selected α-carbons of nearby residues (red stick) are shown, along with residues (blue stick) important for AZ10606120 binding (Karasawa & Kawate, 2016). Residues shown in parentheses are the equivalent rat P2X7 residues.

Multi-step synthesis of X7-uP.

1,2-DCE: 1,2-dichloroethane; EDC: 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide; HOBt: 1-Hydroxybenzotriazole; DIEA: N,N-Diisopropylethylamine; DMAP: 4-Dimethylaminopyridine; DMTMM: 4-(4,6-dimethoxy-1,3,5-triazin-2-yl)-4-methyl-morpholinium chloride.

1H NMR spectrum of compound 3.

1H NMR spectrum of compound 5.

1H NMR spectrum of compound 6.

1H NMR spectrum of compound 7.

1H NMR (A) and 13C NMR (B) spectra of compound 8.

1H NMR (A) and 13C NMR (B) spectra of compound 10.

1H NMR (A) and 13C NMR (B) spectra of compound 11.

1H NMR (A) and 13C NMR (B) spectra of compound X7-uP.

Specificity and kinetic analysis of P2X7 labeling by X7-uP.

(A) Western blot analysis showing the specificity of P2X7 labeling by X7-uP in untransfected (-) or P2X7c-myc-transfected (+) HEK293T cells (refer to Fig. 2C for details). Molecular mass makers are shown on the right. (B) Western blot analysis of P2X7 labeling using different concentrations of X7-uP at various time points in HEK293T cells expressing P2X7c-myc. β-Actin was used as a loading control. Molecular mass makers are shown on the right. (C) Time course of P2X7 labeling using 0.5 μM and 2.5 μM X7-uP. Data (mean ± s.e.m., n = 3 independent transfections) were fitted using Eq. (1), providing pseudo-first-order reaction rate constants (kapp) (mean ± s.e.m.). (D) Concentration-dependence of kapp values for X7-uP. Data were fitted with Eq. (2), yielding the labeling rate constant (kL = 0.011 ± 0.003 s-1, mean of triplicate ± standard deviation (s.d.)) and the dissociation constant (Kd = 7.3 ± 2.7 μM, mean ± s.d.).

Determination of X7-uP-induced P2X7 labeling yields in HEK293T and BV2 cells.

(A) Western blot analysis of P2X7 in the supernatants of cell lysates from P2X7c-myc-transfected HEK293T cells (left) or BV2 cells (right), either untreated (-) or treated with X7-uP (+) for 60 minutes (HEK293T cells) or 10 minutes (BV2 cells). P2X7 was detected using either an anti-c-myc antibody (for HEK293T cells) or a mouse anti-P2X7 antibody (for BV2 cells). β-Actin was used as a loading control. Molecular mass makers are shown on the right. (B) Quantification of P2X7 labeling yield, calculated using the following equation: (𝟏 − 𝒂⁄𝒃) × 𝟏𝟎𝟎, where 𝒂 and 𝒃 represent the amounts of P2X7 in lanes 2 and 1 of the supernatants, respectively.

Cell-surface labeling assay of HEK293T cells transfected with various GFP-tagged P2X subunits.

(A) Western blot analysis of cell-surface expression of GFP-tagged P2X subunits. After transient transfection, cells were treated with the cell-impermeant reagent sNHS-SS-biotin. Following extensive washing, cells were lysed, and biotinylated proteins were pulled down and separated on SDS-PAGE. Western blot was revealed using an anti-GFP antibody (@GFP). Molecular mass makers are shown on the right. NT: non-transfected cells (B) Corresponding input control for the blot shown in panel A. (C) Quantification of cell-surface P2X expression. (D) Quantification of total P2X expression. Band intensities of P2X-GFP monomers were normalized to their respective β-actin controls. Mean ± s.e.m., n = 5 independent transfections.

dSTORM images of HEK293T cells transfected with P2X7-mScarlet.

(A-C) Wild-field fluorescence images (left) and corresponding dSTORM super-resolution images (right) of HEK293T cells labeled with X7-uP and revealed with Strept-A 647. Images were acquired either in the absence (A) or presence (B) of 10 μM AZ10606120. (C) Control condition in the absence of X7-uP. Scale bars, 5 μm. Scale bars in insets, 0.5 μm.

Additional data related to Fig. 4B.

Confocal images of HEK293T cells transiently transfected with different P2X7 constructs (P2X7-mScarlet, K110A and K300A). Refer to the legend of Fig. 3 for additional details. Scale bars, 10 μm.

X7-uP-induced P2X7 labeling in BV2 cells.

(A) Western blot analysis of P2X7 labeling in BV2 cells treated with 1 μM X7-uP for 10 minutes in the absence or presence of 10 μM AZ10606120 or 10 μM A740003. After extensive washing, cells were lysed, and biotinylated proteins were pulled down and separated on SDS-PAGE. Western blot was revealed using a mouse anti-P2X7 antibody (@mP2X7). Molecular mass makers are shown on the right. (B) Corresponding input control for the blot shown in panel A. β-Actin was used as a loading control. (C) Confocal images of BV2 cells labeled with X7-uP and visualized with Strept-A 647 in the absence or presence of AZ10606120 or A740003, or in the absence of X7-uP (control, Strept-A 647 alone). Alexa 647 signals are shown in red, and nuclei are stained with Hoechst (blue). Scale bars, 10 μm. (D) Quantification of Alexa 647 fluorescence. Bars represent mean ± s.d. (n = 46-220 cells, t-test comparisons to control, ****P < 0.0001).

dSTORM analysis of X7-uP-induced P2X7 labeling in BV2 cells.

(A) Bright-field and dSTORM images of BV2 cells labeled with Strept-A 647 in the presence (+X7-uP) or absence (w/o X7-uP) of X7-uP. Scale bars, 10 mm. Insets: Magnified dSTORM images. Scale bars, 1 mm. (B) Quantification of the number of localizations in control and X7-uP-treated cells (n = 3 independent experiments). Bars represent mean ± s.e.m. Unpaired t-test, ****P < 0.0001. (C) Bright-field and dSTORM images of BV2 cells labeled with X7-uP and revealed with Strept-A 647 in untreated or MβCD-treated cells. (D) dSTORM Metamorph pixelized global images of X7-uP-labeled BV2 cells, revealed with Strept-A 647, showed clustering in the LPS, LPS+ATP, and LPS+BzATP conditions compared to the untreated control. Scale bars, 5 μm. Insets: Magnified images. Scale bars, 1 μm. (E) Number of detections per cluster (same data set as Fig. 5G, shown in log scale). (F) Average intensity per cluster. Bars represent mean ± s.e.m. (G) Distance to the closest neighbor. Bars represent mean ± s.e.m.

Blink detection features.

Detection of single-molecule blinking (localization) in untreated (NT) and LPS+ATP-treated BV2 cells across several regions of interests (ROIs), as a function of acquisition frames. The first 2,000 frames correspond to stack 1 (stk1), and the subsequent 2,000 frames correspond to stacks 2, 3, and 7 (stk2, stk 3 and stk7). All single blinks originated from isolated detections, except for ROIs 2, and 3 in the LPS+ATP condition, where they arose from clusters.

Confocal data of X7-uP-induced P2X7 labeling in BV2 cells.

(A) Confocal images of BV2 cells, untreated or treated with the indicated conditions, labeled with X7-uP and visualized with Strept-A 647. Alexa 647 signals are shown in red, and nuclei are stained with DAPI (blue). Scale bars, 10 μm. (B) Quantification of Alexa 647 fluorescence. Bars represent mean ± s.e.m. (n = 76-376 cells) Data were compared using Kruskal-Wallis followed by Dunn’s multiple comparisons (*P = 0.0191, **P = 0.0099, ##P = 0.0027, ****P < 0.0001).