Protein engineering and in vitro characterization. (A) Structure of the binding pocket of glutamate-bound GltI (2VHA.pdb) with residues S72 (left) and S27 (right) shown as sticks and bound glutamate as sticks inside transparent spheres. (B) Fluorescence response of purified jAspSnFR3-mRuby3 when titrated with aspartate (black) or glutamate or asparagine (grey tones). Ex. 485 nm (20 nm bandpass), Em. 535 nm (20 nm bandpass). Error bars are s.d. of three technical replicates. (C) Live cell imaging in the phase contrast, GFP and RFP channels of H1299 Nuclear-RFP cells expressing jAspSnFR3 after 24 hours with/without glutamine.

Figure 1—figure supplement 1. Aspartate specificity and excitation/emission spectra.

Figure 2—figure supplement 2. Decoy, temperature and pH sensitivity.

jAspSnFR3 resolves temporal aspartate changes in live cells. (A) Overview of aspartate metabolism and the effect of glutamine depletion, mitochondrial inhibition, GOT1/2 knockout and pyruvate/asparagine supplementation. GLN, glutamine. ETC, electron transport chain. LDH, lactate dehydrogenase. OAA, oxaloacetic acid. ASP, aspartate. ASN, asparagine. For (B), (C), (E) and (F), sensor signal over time shown as RFP normalized jAspSnFR3 signal following various perturbations of live cells. All experiments shown are normalized to a pre-treatment scan, then treated with the specified drug or amino acid and scanned 30 min following treatment. Grey dashed lines indicate the time of treatment. (B) H1299 cells treated with a rotenone titration and rescued by co-treatment with pyruvate. (C) H1299 GOT1/2 double knockout cells grown in media with 40 mM aspartate, washed thrice in media without aspartate and then changed into media with a titration of aspartate. (D) Western blot verification of H1299 GOT1/2 double knockout. (E) H1299 cells treated with either rotenone or metformin to compare inhibitor kinetics. (F) H1299 cells changed into media with a titration of glutamine with or without 1 mM asparagine. Markers indicate the average using available well replicates and are superimposed on a bootstrapped 95% confidence interval colored using the same color code as the markers. AU, arbitrary unit.

Figure 1—figure supplement 1. Rotenone titration in different cell lines.

Figure 2—figure supplement 2. Narrow range glutamine limitation.

jAspSnFR3 signal predicts LCMS measured intracellular aspartate concentration. A Hill equation with top and bottom asymptotes, midpoint and slope as free variables is fitted to the datapoints and shown by the black line. The intracellular aspartate concentration at the inferred half maximum of RFP normalized jAspSnFR3 signal is reported in the red inserts. (A) Rotenone, metformin and antimycin A titrations in H1299 cells. (B) Rotenone titration in HEK293t cells. (C) Rotenone and metformin titrations in HT1080 cells. Markers indicate a single well from which both LCMS and jAspSnFR3 data was collected. Replicate wells have identical color and marker shape. AU, arbitrary unit.

Figure 1—figure supplement 1. jAspSnFR3 signal does not correlate with glutamate concentration.

CRISPR guides.

(A) Switching specificity of the iGluSnFR3 precursor from glutamate to aspartate using S72X library (left) and S72P, S27X library (right). Titrations with aspartate (solid lines) and glutamate (dashed lines) in bacterial lysate. (B) Excitation and emission spectra of jAspSnFR3-mRuby3. Left, 1-photon spectra. Excitation wavelength was varied from 400 nm to 520 nm (7.5 nm bandpass) while observing emission at 535 nm (10 nm bandpass). Emission wavelength was varied from 535 nm to 600 nm (10 nm bandpass) while exciting at 510 nm (7.5 nm bandpass). Fluorescence was measured both in the absence (dashed lines) and presence of 10 mM aspartate (solid lines). Right, 2-photon cross-sections, also ± 10 mM aspartate, with an overlay of calculated F/F (green). Vertical bar indicates 1040 nm.

(A) jAspSnFR3-mRuby3 does not appreciably change its green fluorescence in response to other amino acids (alanine, phenylalanine, glycine, histidine (red line), isoleucine, leucine, methionine, proline, glutamine, arginine, serine, threonine, valine, or tryptophan). Insert with aspartate in black and glutamate/asparagine in grey for comparison. Ex. 485 nm (20 nm bandpass), Em. 535 nm (20 nm bandpass), 0.2 µM purified protein in PBS. (B) jAspSnFR3-mRuby3 shows increased red fluorescence at millimolar concentrations of all amino acids, with apparent responses to histidine at 100 µM (red trace). Ex. 587 nm (20 nm bandpass), Em. 662 nm (20 nm bandpass). (C) jAspSnFR3-mRuby3 does not respond to other decoys: citrate, lactate, pyruvate, malate, alpha-ketoglutarate, cis-aconitate, succinate, fumarate, or oxaloacetate (orange squares); nor to relevant pharmacological treatments: rotenone (green squares) or metformin. The small increase in fluorescence from rotenone is likely due to the scattering of a visibly turbid solution; rotenone has very low solubility in water. Ex. 485 nm (20 nm bandpass), Em. 535 nm (20 nm bandpass). (D) jAspSnFR3-mRuby3 is not adversely affected by temperature. Fluorescence as a function of aspartate titration at 23°C (light grey), 30°C (medium grey), and 37°C (black). Error bars are standard deviation of three technical replicates. (E) pH sensitivity of jAspSnFR3-mRuby3 (green component). Ex 485 nm (5 nm bp), Em 515 nm (10 nm bp). Error bars are standard deviation of 5 technical replicates. Solid line is with 3 mM aspartate, dashed line is without aspartate.

jAspSnFR3 temporal response after rotenone treatment. RFP normalized jAspSnFR3 signal change over time following rotenone treatment of live cells. Related to Figure 2, panel B; however, these experiments are not normalized to a pre-treatment scan. Grey dashed lines indicate the time of treatment and the first scan occurs 30 min after this. (A) HT1080 cells using nuclear RFP to normalize the jAspSnFR3 signal. (B) HT1080 cells using an RFP fused to jAspSnFR3 (jAspSnFR3-mRuby3) for normalization. (C) Comparison between the steady-state signal of (A) and (B) with a linear regression shown as a red dashed line to show that nuclear RFP and RFP fusion normalizations are equivalent. (D) HEK293t cells using an RFP fused to jAspSnFR3 for normalization. For plots (A), (B) and (D) markers indicate the average using available well replicates and are superimposed on a bootstrapped 95% confidence interval colored using the same color code as the markers. For plot (C) markers indicate the average using available well replicates and errorbars are drawn as +/-the standard deviation of the replicates. AU, arbitrary unit.

Aspartate depletes when glutamine is limiting. RFP normalized jAspSnFR3 signal change over time following glutamine depletion in H1299 cells. Identical to Figure 2, panel F but with fewer glutamine concentrations and more well replicates. AU, arbitrary unit.

RFP normalized jAspSnFR3 signal, following various perturbations to live cells, is not correlated with the LCMS measured intracellular glutamate concentration. Datapoints are fitted to a local linear regression, shown by the black line, otherwise, these plots are identical to those in Figure 3. AU, arbitrary unit.