Field stimulation screen for improved iGABASnFR.

a. Top: Schematic representation of iGABASnFR1 and the amino acid substitutions that gave rise to iGABASnFR2 and iGABASnFR2n. White, IgG secretion signal (cleaved off during trafficking to cell surface); blue, GABA binding protein Pf622; green, cpSFGFP; dark green, Myc epitope tag; red, PDGFR transmembrane domain. Numbering is relative to each of the constituent protein domains; the relationship to sequential numbering of the entire polypeptide is shown in Figure 1 -figure supplement 1. Bottom: Crystal structure of a preliminary version of iGABASnFR (PDB ID 6DGV) with the 39 sites targeted for mutagenesis. Gray spheres indicate the approximate position of GABA, based on homology to the liganded structure of GABA-binding protein Atu4243 (Planamente et al. 2012).

b. Mutagenesis and screening strategy. Variants with single site mutations are screened for ΔF/F and expression in an initial field stimulation assay, strong performers identified and then combined and re-screened in a second round. The top-performing mutants (positive-going quadruple mutant iGABASnFR2 and negative-going triple mutant iGABASnFR2n) are characterized further.

c. Joint optimization of sensitivity (ΔF/F, x-axis) and expression, measured with responsive pixels (y-axis). ΔF/F is normalized to in-plate iGABASnFR1 controls. Black circle: (1,1) position represents iGABASnFR. Plus signs: mutants from the first round that were put together to form combos in the second round. iGABASnFR2 and 2n exhibited increased ΔF/F (4.3-fold and -2.2-fold of iGABASnFR, respectively) and greater numbers of responsive pixels (13.1-fold and 10.3-fold). Illumination: 0.34 mW/mm2, imaging framerate: 50 Hz.

d. Performance measures of sensor variants from the screen. Single site variants shown in gray, double site combinations shown in black. Variants are ranked according to the ΔF/F0 values measured for 40 AP, and those rankings maintained for the lower displays of sensor F0, tau on, the time constant of the rising phase of the response, and tau off for the decay. All values are normalized to in-plate iGABASnFR1 controls.

Characterization of iGABASnFR variants in cultured neurons.

a. Fluorescence images of primary neurons expressing iGABASnFR1 (orange), iGABASnFR2 (green), iGABASnFR2n (blue) under the CAG promoter at baseline, at peak brightness after field stimulation with 40 APs, and the corresponding ΔF/F0. Scale bar, 20 μm.

b. Time courses of the ΔF/F0 response to 1, 10 and 40 APs delivered at 83 Hz. iGABASnFR2n signals are inverted for display. Traces and error bars denote mean ± s.e.m., n = 20 culture wells for each variant.

c. 10–90% rise time of the three sensor variants for the 1 AP condition.

d. Peak ΔF/F0 of the three variants over different levels of stimulation.

e. Signal-to-noise (d′) of the three variants over different levels of stimulation.

Crystal structure of iGABASnFR2 in complex with GABA.

a. Superposition of the structure of iGABASnFR2 in complex with GABA (cyan) and unliganded iGABASnFR precursor (magenta). The bound GABA molecule is represented as colored spheres.

b. Cartoon representation of iGABASnFR2 with key mutations shown as yellow sticks. Mutations in the original iGABASnFR1 shown in orange.

c. GABA binding site of iGABASnFR2. Residues involved in GABA binding are shown as lines and GABA shown as stick.

Biophysical properties of iGABASnFR variants.

a. GABA titrations with purified iGABASnFR protein. Lines indicate fits to mean of n = 5 titration series, error bars are s.e.m.

b. GABA titrations with sensors expressed on the surface of cultured neurons. Fits (bottom panel) show ΔF/F0 response to increasing concentrations of GABA measured with ROIs placed on individual cell bodies (top panel). In these conditions iGABASnFR2 shows a greater dynamic range than iGABASnFR1, in contrast to results with purified protein. Color lookup table is the same for all images. Scale bar: 50 µm. Illumination: 5.6 mW/mm2, imaging at 1 frame per sec.

c. Observed reaction rate constant (Kobs) values from stopped-flow measurements for the three sensors.

d. Stopped flow kinetics of iGABASnFR variants. Lines indicate fits to mean of n = 3 replicates from three separate batches of purified protein.

e. One-photon excitation and emission spectra of soluble iGABASnFR protein in the presence (10 mM) and absence of GABA.

f. Two-photon excitation spectra and the computed ΔF/F0 of iGABASnFR variants in the presence (10 mM) and absence of GABA.

Photophysical properties of iGABASnFR variants as purified proteins

iGABASnFR2 reliably reports direction selectivity in the retina.

a. Schematic illustration of centrifugal direction selectivity in starburst cell’s motion responses.

b. Spatially asymmetric inhibitory connections (red dots) between starburst cells (SACs) and direction-selective ganglion cells (DSGCs), which are proposed to generate direction selectivity in DSGCs.

c. Example field of view of SAC processes expressing iGABASnFR1. Yellow region of interest (ROI) is analyzed to evaluate responses to visual stimuli.

d. Responses to static flash from ROI in c but with SACs expressing iGABASnFR2.

e. Responses to motion stimulus.

f-h. Results of imaging using iGABASnFR2.

i. Histograms of response amplitude index (left) and response reliability (right) from SACs expressing iGABASnFR. n = 147 ROIs collected across 5 retinae.

j. As in i but with SACs expressing iGABASnFR2. n = 346 ROIs collected from 3 retinae. Responses to motion stimulus.

k. Average signals during preferred direction motion with iGABASnFR1 (top, gray) and iGABASnFR2 (bottom, cyan). Line and shading indicate mean ± s.d (n = 147 for iGABASnFR, n = 346 for iGABASnFR2).

l. Comparison of SNR of the motion response detected with the two sensor versions. p = 0. Two-tailed Mann-Whitney-Wilcoxon test.

m. Comparison of direction selectivity (CV, circular variance) for the two sensor versions. p = 7.812 x 10-6. Two-tailed Mann-Whitney-Wilcoxon test.

iGABASnFR2 reliably detects synaptic GABA release in slices and sensory-evoked GABA in vivo.

a. Top: Image from a whole-cell recording of an iGABASnFR1-expressing hippocampal interneuron in area CA3 of an acute brain slice. To identify axonal boutons, morphology was visualized using Alexa Fluor 594 (red channel), included in the internal solution. The image shown is from the red channel. The axonal segment (dotted rectangle) is shown magnified in the inset, with the position of a schematized 1.5 µm-wide Tornado scan path indicated. Image is the average projection of a z-stack covering 50 µm.

Bottom: iGABASnFR1 fluorescence signal acquired using a 0.5 kHz Tornado scan at the bouton shown above. Horizontal axis: time (700 ms scan duration); vertical axis: spiral turn angle. Trace shows mean ΔF/F₀ response (±SEM) across 15 trials of five action potentials at 50 Hz (arrows).

b. Top: As in (a), except the image shows iGABASnFR2 fluorescence from an interneuron in area CA1 in an organotypic slice. The image is the average projection of a 30 µm z-stack.

Bottom: As in (a), but showing iGABASnFR2 signal from the Tornado scan. Trace shows the mean ΔF/F₀ response (±SEM) across 11 trials of four action potentials delivered at 20 Hz (arrows).

c. Brief rhythmic whisker stimulus triggers volume-transmitted extracellular elevations of GABA in the barrel cortex detected by iGABASnFR2 fluorescence. Left: Schematic of experimental arrangement, with whisker stimulation via 4 air-puffs at 20 Hz while imaging the contralateral barrel cortex. Right: Image panel shows barrel cortex area (∼300 µm depth) and position of a 1 kHz Tornado-scan. Trace is the single-trial fluorescence response of iGABASnFR2 to a contralateral 200 ms whisker stimulation, as indicated.

a. Annotated amino acid sequence of iGABASnFRs. iGABASnFR1 sequence shown in gray text, with sites targeted for mutagenesis in black. Shading indicates domains according to: IgG secretion signal Pf622 2-276 SFGFP 147-238 Junction SFGFP 1-146 Pf622 277-320 Myc epitope PDGFR transmembrane domain 513-561 b. iGABASnFR2 sequence, with mutations shown in red text. Mutations are listed according to domain numbers, with sequential numbering through the polypeptide indicated at the left. Sequential numbering through the polypeptide is indicated on the left, while mutations in the text are listed according to domain numbers: S99PfA.F102PfY.F104PfY.L178gfpS. S99PfA.F102PfY.F104PfY.L178gfpS, which we subsequently refer to as iGABASnFR2, and the best variant with inverted signal, iGABASnFR.S99PfA.F104PfH.R168gfpP was chosen to be iGABASnFR2n. c. iGABASnFR2n sequence, with mutations shown in red text. Sequential numbering through the polypeptide is indicated on the left, while mutations in the text are listed according to domain numbers: S99PfA.F104PfH.R168gfpP. d. Schematic showing the relationship between domain-based numbering scheme and sequential numbering through the entire polypeptide (top) and the domain-based numbering scheme. In both schemes numbering starts at the residue exposed after signal sequence cleavage.

Responses of iGABASnFR variants to GABA- related compounds.

Different iGABASnFR variants (200 nM) were titrated with various concentrations of ligands of interest. Estimated EC50 values shown in insets. Each titration has a minimum of n=3 replicates.

Competition of GABA-related compounds for binding to iGABASnFR variants.

Different iGABASnFR variants (200 nM) were titrated with increasing concentrations of GABA in the presence of different potential competing compounds at 1 mM. Lines indicate fits to mean of n = 3 titration series, error bars are s.e.m. Estimated EC50 values shown in insets.

pH titrations of iGABASnFR variants.

Fluorescence response of the three iGABASnFR variants across a pH titration, under GABA- saturated (10 mM, solid lines) and GABA-free (dashed lines) conditions. Fluorescence values are normalized to the peak fluorescence of the GABA-saturated form of each sensor. Measurements were performed using 200 nM sensor concentration. Data points represent the mean of n = 3 technical replicates; error bars indicate s.e.m.

Data collection and refinement statistics of iGABASnFR2 in complex with GABA.