A genetically encoded single-wavelength sensor for imaging cytosolic and cell surface ATP

Abstract

Adenosine 5' triphosphate (ATP) is a universal intracellular energy source and an evolutionarily ancient, ubiquitous extracellular signal in diverse species. Here, we report the generation and characterization of single-wavelength genetically encoded fluorescent sensors (iATPSnFRs) for imaging extracellular and cytosolic ATP from insertion of circularly permuted superfolder GFP into the epsilon subunit of F₀F₁-ATPase from Bacillus PS3. On the cell surface and within the cytosol, iATPSnFR^1.0 responds to relevant ATP concentrations (30 μM to 3 mM) with fast increases in fluorescence. iATPSnFRs can be genetically targeted to specific cell types and sub-cellular compartments, imaged with standard light microscopes, do not respond to other nucleotides and nucleosides, and when fused with a red fluorescent protein function as ratiometric indicators. After careful consideration of their modest pH sensitivity, iATPSnFRs represent promising reagents for imaging ATP in the extracellular space and within cells during a variety of settings, and for further application-specific refinements.

Fig. 1

Design and optimization of a single-wavelength ATP sensor. a Schematic showing the design and workflow used to optimize QUEEN-7µ into a single-wavelength ATP sensor with the goal of displaying the sensor on the surface of cells. b Dose–response curves of iATPSnFR over several successive rounds of mutagenesis (Ex: 488 nm, Em: 515 nm). Fluorescence quenching at very high ATP concentrations can be observed in addition to binding-dependent increases. c Dose–response curves for purified ATeam, QUEEN-7µ, iATPSnFR^1.0, and iATPSnFR^1.1. ATeam dose–response curves were acquired with Ex: 435 nm and Em: 530 nm. The other constructs were with Ex: 488 nm, Em: 515 nm. d Dose–response curves of purified iATPSnFR^1.1 to ATP, ADP, AMP, and adenosine. e, f Excitation and emission spectra for iATPSnFR^1.0 and iATPSnFR^1.1 in solution (the traces are the average from 48 replicates each in a 96-well plate). The error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. When greater than one (in the case of exemplar traces and graphs), n is provided in the figure panels and refers to the number of independent evaluations

Fig. 2

Representative images and traces for responses measured with cell surface iATPSnFRs. a The upper panels show three examples of ATP responses for HEK293 cells expressing iATPSnFR^1.0. The lower panels show representative traces and an average dose–response curve from 290 cells. b As in (a), but for iATPSnFR^1.1. The dose–response curve represents an average from 191 cells. In the dose–response curves, the error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells

Fig. 3

Characterization of cell surface iATPSnFRs in HEK293 cells. a Single planes from a confocal z-stack of transiently transfected HEK293 cells expressing cytoplasmic mCherry and membrane-displayed iATPSnFR^1.0. b As in (a), but for iATPSnFR^1.1 transiently expressed in HEK293 cells. c Average traces for iATPSnFR^1.1 (green) and iATPSnFR^1.0 (red) over a range of concentrations from 10 µM to 3 mM measured with wide field epifluorescence imaging. d, e Dose–response curves for iATPSnFR^1.0 and iATPSnFR^1.1 displayed on the surface of HEK293 cells. Sensitivity (EC₅₀), total dynamic range (Max) and cooperativity (nH) shown. f Representative Western blot from HEK293 cells transfected with iATPSnFR^1.0 or iATPSnFR^1.1 probed for GFP (iATPSnFRs) and GAPDH (loading control). The full-uncropped gel is shown in Supplementary Figure 9. In panel (f), the graph shows the quantification of iATPSnFR expression from four Western blots normalized to GAPDH loading control. The circle represents the mean, the box the s.e.m., the whiskers the s.d., and the horizontal line the median. The data were compared with an un-paired Students t test. In the other panels, the error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells for panels (c and d) and independent evaluations for panel (f)

Fig. 4

Assessment of the kinetics and ligand selectivity of cell surface ATP sensors. a Traces for cell-displayed iATPSnFR^1.0, iATPSnFR^1.1, and iGluSnFR kinetics from fast-solution change experiments. We could change solutions in under ~10 ms. b Shows the τ_on and τ_off for iATPSnFR^1.0 and iATPSnFR^1.1 at various ATP concentrations. c Shows the response of iATPSnFR^1.0 and iATPSnFR^1.1 to ATP, ADP, AMP, and adenosine (1 mM applications). d The response of iATPSnFR^1.0 and iATPSnFR^1.1 to ATP, GTP, and UTP (1 mM applications). e Traces of iATPSnFR^1.0 and iATPSnFR^1.1 with three different concentrations of ATP (1, 0.3, and 0.1 mM) with and without 1 mM ADP. ADP has little effect on iATPSnFR responses. The error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells

Fig. 5

Characterization of cell surface and cytosolic iATPSnFR^1.0 in astrocytes and neurons. a A single plane from a confocal z-stack of a U373MG cell expressing cytoplasmic mCherry and membrane-displayed iATPSnFR^1.0. b Confocal images of U373MG cells expressing iATPSnFR^1.0 before ATP application (control) and during a 300 µM ATP application (+ATP), as well as the change in fluorescence (dF). c Confocal images of hippocampal neurons before ATP application (control) and during 1 mM ATP application, as well as the change in fluorescence (dF). d, e Dose–response curves for iATPSnFR^1.0 when displayed on the surface of U373MG astroglia and hippocampal neurons. f A confocal image of a U373MG astroglia cell transfected with iATPSnFR^1.0 before a 3 mM puff of ATP (control), during the puff (ATP), as well as the change in fluorescence (dF). An average trace from all cells with the s.e.m. is also shown. g Representative hippocampal astrocyte in acute brain slice expressing iATPSnFR^1.0 before the application of ATP (control), during the application of 3 mM ATP (ATP), and the dF. An average trace from all cells with the s.e.m. is also shown. h Scatter graph summary of the effect of 3 mM ATP on iATPSnFR^1.0 expressed in the cell types indicated. In panel (h), the circle represents the mean, the box the s.e.m., the whiskers the s.d., and the horizontal line the median. p Values are shown for statistical comparisons to the buffer puff control using un-paired Student’s t tests. Abbreviations: hip hippocampus, SC spinal cord, MSNs medium spiny neurons. Note broken y-axis. In other panels, the error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells

Fig. 6

Characterization of cytosolic mRuby-iATPSnFR^1.0. a Cartoon schematic of mRuby-iATPSnFR^1.0: ATP binding is schematized to show an increase in cpSFGFP fluorescence, whereas mRuby fluorescence remains unchanged. b Dose–response curves for iATPSnFR^1.0 in relation to mRuby-iATPSnFR^1.0 in HEK293 cell lysates. c Changes in fluorescence of mRuby-iATPSnFR^1.0 (green and red channels) before, during, and after inhibition of glycolysis with 2DOG in HEK293 cells. d Changes in fluorescence of mRuby-iATPSnFR^1.0 (green and red channels) during oxygen-glucose deprivation in hippocampal brain slices. Data from astrocytes as well as CA1 pyramidal neuronal cell bodies and dendrites are shown. The error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells

Fig. 7

Assessments of how iATPSnFR^1.0 was affected by pH. a–d Effect of pH on the unbound and ATP bound fluorescence of the indicated constructs. e, f Comparison of the effect of ATP on the fluorescence of the indicated constructs at pH 6.4 and 7.4. The error bars represent the s.e.m. and in some cases are smaller than the symbols used for the mean. n numbers are provided in the figure panels and refer to the number of cells in panel (a), and number of independent evaluations in (b–f)