Simultaneous recording of multiple cellular signaling events by frequency- and spectrally-tuned multiplexing of fluorescent probes

Abstract

Fluorescent probes that change their spectral properties upon binding to small biomolecules, ions, or changes in the membrane potential (V_m) are invaluable tools to study cellular signaling pathways. Here, we introduce a novel technique for simultaneous recording of multiple probes at millisecond time resolution: frequency- and spectrally-tuned multiplexing (FAST^M). Different from present multiplexing approaches, FAST^M uses phase-sensitive signal detection, which renders various combinations of common probes for V_m and ions accessible for multiplexing. Using kinetic stopped-flow fluorimetry, we show that FAST^M allows simultaneous recording of rapid changes in Ca²⁺, pH, Na⁺, and V_m with high sensitivity and minimal crosstalk. FAST^M is also suited for multiplexing using single-cell microscopy and genetically encoded FRET biosensors. Moreover, FAST^M is compatible with optochemical tools to study signaling using light. Finally, we show that the exceptional time resolution of FAST^M also allows resolving rapid chemical reactions. Altogether, FAST^M opens new opportunities for interrogating cellular signaling.

Keywords: Arbacia punctulata; cell biology; fluorescence multiplexing; signal transduction; voltage-sensitive dye.

Figure 1.. Strategies for multiplexing of fluorescent probes, and outline of the chemosensory signal transduction in the flagellum of sea urchin sperm.

(A) Spectrally separable emission spectra (dashed) of probes allow their simultaneous recording using optical filtering. (B) Spectrally separable excitation spectra (outlined) allow quasi-simultaneous recording of probes using excitation-switching. (C) Frequency-tagging and phase-sensitive detection of fluorescence combined with optical filtering using frequency- and spectrally-tuned multiplexing (FAST^M) allow simultaneous recording of probes based on separable excitation and/or emission spectra. (D) Schematic of the chemosensory signaling pathway and (E) illustration of the time course of the signaling events in sea urchin sperm (reviewed in Strünker et al., 2015). Resact, the chemoattractant peptide released by the egg, triggers the synthesis of cGMP by activating a receptor guanylyl cyclase (GC). The rise in cGMP elicits a pulse-like V_m hyperpolarization mediated by a cyclic nucleotide-gated K⁺ channel (CNGK). The hyperpolarization activates a voltage-gated Na⁺/H⁺ exchanger (sNHE) and a hyperpolarization‐activated and cyclic nucleotide‐gated (HCN) channel. The Na⁺/H⁺ exchange increases [Na⁺]_i and pH_i. In turn, the increase in pH_i primes pH_i-controlled CatSper Ca²⁺ channels to open during the recovery from hyperpolarization driven by HCN channels. The resulting Ca²⁺ influx drives chemotactic steering towards the egg. (F) Schematic of the stopped-flow setup: one syringe is filled with a suspension of probe-loaded sperm, and a second syringe is filled with a solution of resact. The syringe pistons move synchronously to rapidly mix sperm with resact in a micromixer and subsequently push this mixture into an observation cuvette, where spectroscopic measurements are performed (see Hamzeh et al., 2019).

Figure 2.. Experimental configuration for frequency- and spectrally-tuned multiplexing (FAST^M) of Fura-2, BCECF, and RhoVR in a stopped-flow device.

(A) Superposition of excitation (outlined) and emission (filled) spectra of fluorescent probes for Ca²⁺ (Fura-2), pH (BCECF), and V_m (RhoVR). Bandpass filters used for excitation (filled) and emission (outlined) are shown above the spectra. Inset: excitation and emission spectra depicted individually with respective filters (black bars). (B) Schematic of FAST^M: each probe is excited by an LED modulated at a different frequency. The modulated emission is optically filtered and collected by two photomultipliers (PMTs). The PMT signals are demodulated by lock-in amplifiers in a phase-sensitive fashion to recover in real time [Ca²⁺]_i, pH_i, and V_m signals.

Figure 3.. Resact-induced pH_i, [Ca²⁺]_i, and V_m signals recorded in sperm loaded with BCECF, Fura-2, or RhoVR using optical filtering alone or frequency- and spectrally-tuned multiplexing (FAST^M).

Time course of the fluorescence signals recorded from BCECF- (A–C), Fura-2- (D–F), or RhoVR-loaded sperm (G–I) after mixing with resact (50 pM). Fluorescence was recorded in the BCECF, Fura-2, and RhoVR channels using optical filtering alone (gray traces) or FAST^M (colored traces). (A, D, G) Fluorescence signals in arbitrary fluorescence units (AFU); to ease the comparison, signals in (A), (D), and (G) were normalized (set to 1) to the baseline fluorescence (F₀) in the BCECF, the Fura-2, and the RhoVR channel, respectively, recorded immediately after mixing with resact. (B, E, H) Resact-evoked change in fluorescence (ΔF) with respect to the baseline fluorescence (F₀), that is, ΔF/F₀ (%); ^#signals smoothed with a sliding average of 80 ms. (C, F, I) First 2 s of the fluorescence signal recorded in the BCECF channel plotted against that recorded in the Fura-2 or the RhoVR channel using either optical filtering (top panel) or FAST^M (bottom panel). Gray line: linear fit of the plots to quantify the crosstalk between the channels (see explanation in the text). (J) Percent crosstalk between the channels according to the analysis shown in (C), (F), and (I).

Figure 4.. Simultaneous recording of resact-evoked pH_i, [Ca²⁺]_i, and V_m signals in sperm loaded with BCECF, Fura-2, and RhoVR.

Relative changes in fluorescence ∆F/F₀ evoked by 50 pM resact. The respective control signal evoked by mixing with artificial sea water (ASW) was subtracted, setting the control-signal level to ΔF/F₀ (%) = 0 (dotted line). Signals were recorded using optical filtering alone (A) or frequency- and spectrally-tuned multiplexing (FAST^M) (B). (C) Simultaneous FAST^M recording of resact-evoked signals from pooled sperm loaded separately with either BCECF, Fura-2, or RhoVR.

Figure 5.. Interrogating putative probe-related perturbations of signaling.

Resact-evoked V_m, pH_i, and [Ca²⁺]_i signals recorded individually from different sperm samples loaded with one probe only (A) or recorded simultaneously from triple-loaded sperm (B); to facilitate direct comparison, Fura-2 fluorescence was multiplied by –1 to depict an increase of [Ca²⁺]_i as an increasing signal. (C) Comparison of V_rest of sperm loaded with RhoVR (single-loaded) or RhoVR, BCECF, and Fura-2 (triple-loaded). (D) Calibrated resact-induced (50 pM) V_m response and accompanying pH_i and [Ca²⁺]_i signals. The artificial sea water (ASW) control was subtracted, and the dotted black line indicates ΔF/F₀ = 0 and V_rest. The V_m at the onset of the pH_i and [Ca²⁺]_i signals was deduced from the signal latencies. (E) V_m at the onset of pH_i and [Ca²⁺]_i signals in single- versus triple-loaded sperm. With increasing resact concentrations, the rise in pH_i and [Ca²⁺]_i commenced at increasingly negative V_m (Seifert et al., 2015).

Figure 6.. Simultaneous recording of resact-induced signaling events in sperm loaded with various triple combinations of probes for Ca²⁺, pH, Na⁺, and V_m using frequency- and spectrally-tuned multiplexing (FAST^M).

Superposition of excitation (outlined) and emission (filled) spectra of (A) Fura-2, VF2.1.Cl, and pHrodo; (B) Fura-2, RhoVR, and ANG-2; (C) BeRST, pHrodo, and ANG-2. Bandpass filters used for excitation (filled) and emission (outlined) are depicted above the spectra. Inset: individual excitation and emission spectra with respective filters (black bars). (D–F) Signals (∆F/F₀) evoked by 500 pM resact corrected for the artificial sea water (ASW) control and normalized to their respective peak values (set to 1) for easier illustration.

Figure 6—figure supplement 1.. Spectral utility of voltage-sensitive probes.

Calibrated resact-evoked (500 pM) V_m response in sea urchin sperm reported by VF2.1.Cl (A), RhoVR (B), and BeRST (C) upon excitation at the principal peak (gray) versus excitation at the secondary peak (blue). Inset: excitation and emission spectra of probes in sea urchin sperm. Bandpass filters used for excitation (filled) and emission (outlined) are depicted above.

Figure 7.. Simultaneous ratiometric recording of [Ca²⁺]_i and pH_i signals in human sperm.

(A) Superimposed excitation (outlined) and emission (filled) spectra of Fura-FF and BCECF. Inset: individual spectra with respective filters (black bars). (B) Schematic of frequency- and spectrally-tuned multiplexing (FAST^M) configuration for simultaneous ratiometric dual-excitation recording of Fura-FF and BCECF in human sperm. (C) Crosstalk among channels based on the analysis shown in Figure 7—figure supplement 1. ^#Under these particular conditions, the approach to quantify crosstalk yielded an erroneously inflated value (for details, see Figure 7—figure supplement 1). (D) Left panels: ratiometric [Ca²⁺]_i and pH_i signals (ΔR/R₀) in human sperm evoked by NH₄Cl (10 mM) or progesterone (100 nM) corrected for the buffer control. Right panel: fluorescence signals in the individual Fura-FF and BCECF channels underlying the ratio.

Figure 7—figure supplement 1.. Analysis of crosstalk between BCECF and Fura-FF channels recorded with frequency- and spectrally-tuned multiplexing (FAST^M).

Human sperm loaded with BCECF or Fura-FF were mixed with either NH₄Cl or progesterone, respectively. The first 5 s (BCECF-loaded sperm) and 20 s (Fura-FF-loaded sperm) of the fluorescence signals recorded in the different channels were plotted against each other and crosstalk was evaluated by linear fitting of the data. The steepness of the linear fit (gray line) is a measure of the crosstalk between channels. Of note, Fura-FF fluorescence upon excitation at the isosbestic point at 340 nm does not change with [Ca²⁺]_i. Fura-FF is also excited by 445 nm, and the emission changes with [Ca²⁺]_i. Consequently, the linear fit of the plot of fluorescence signals at 445 vs. 340 nm excitation indicates severe crosstalk among channels. The absolute fluorescence values are, however, an order of magnitude lower compared to that recorded in BCECF-loaded sperm and do, thus, not affect simultaneous recording of Fura-FF and BCECF.

Figure 8.. Simultaneous recording of [Ca²⁺]_i, pH_i, and V_m signals in sea urchin sperm evoked by flash photolysis of caged cGMP.

(A) Superimposed absorbance spectrum of BECMCM-cGMP and excitation (outlined) and emission (filled) spectra of Fluo-4, pHrodo, and BeRST. Bandpass filters used for excitation (filled) and emission (outlined) are depicted above the spectra. Inset: individual spectra and respective filters (black bars). (B) Schematic of frequency- and spectrally-tuned multiplexing (FAST^M) configuration for uncaging experiments. (C) V_m, pH_i, and [Ca²⁺]_i signals evoked by uncaging intracellular cGMP with a 50 ms UV-flash (gray bar).

Figure 9.. Simultaneous recording of the kinetics of Ca²⁺ dissociation from Fura-2, Fluo-4, and Calbryte 630.

(A) Superimposed excitation (outlined) and emission (filled) spectra of Fura-2, Fluo-4, and Calbryte 630. Inset: individual spectra depicted with respective filters (black bars). (B) Schematic of the frequency- and spectrally-tuned multiplexing (FAST^M) configuration for simultaneous recording of Ca²⁺ dissociation from Fura-2, Fluo-4, and Calbryte 630. (C) Crosstalk between channels according to Figure 9—figure supplement 1. (D) Changes in Fura-2, Fluo-4, and Calbryte 630 fluorescence and 370 nm/340 nm emission ratio of Fura-2 upon mixing of the Ca²⁺-bound probes with the Ca²⁺ chelator BAPTA. (E) K_off values determined by exponential fitting of the individual fluorescence traces and the ratio of Fura-2 (370/340 nm): Fura-2, 340ex: 115 ± 2 s^–1; Fura-2, 370ex: 122 ± 3 s^–1; Fura-2 ratio: 84 ± 2 s^–1; Fluo-4: 354 ± 3 s^–1; Calbryte 630: 178 ± 2 s^–1 (n = 4).

Figure 9—figure supplement 1.. Analysis of crosstalk between Fura-2, Fluo-4, and Calbryte 630 channels simultaneously recorded with FAST^M.

The first 18 ms of the fluorescence signals recorded in the different channels were plotted against each other, and crosstalk was evaluated by a linear fit to the data. The steepness of the linear fit (gray line) is a measure of the crosstalk between channels.

Figure 9—figure supplement 2.. Signal-to-noise (S/N) ratio of fluorescence signals recorded upon mixing of Ca²⁺-bound Fura-2 with BAPTA.

(A) Kinetics of Ca²⁺ dissociation from Fura-2 recorded upon excitation at 340 nm. The kinetics were recorded with frequency- and spectrally-tuned multiplexing (FAST^M) using either different lock-in amplifier time constants (100 µs and 1 ms) or without modulation of the excitation light and using a conventional amplifier (no lock-in). (B) S/N ratio determined by dividing the mean signal amplitude by the standard deviation of 200 data points. The gray line indicates the S/N ratio in recordings performed without modulation of the excitation light.

Figure 10.. Single-cell frequency- and spectrally-tuned multiplexing (FAST^M) fluorescence microscopy for simultaneous recording of cAMP and [Ca²⁺]_i.

(A) Octopamine-signaling pathway in HEK cells coexpressing the DmOCTβ1 receptor, CNGA2-TM channel, and a FRET-based cAMP biosensor. (B) Superimposed excitation (outlined) and emission (filled) spectra of the FRET donor-acceptor pair cerulean-citrine, and the Ca²⁺-probe Calbryte 630. Bandpass filters used for excitation (filled) and emission (outlined) are depicted above the spectra. Inset: individual spectra and filters (black bars). (C) Schematic of the FAST^M configuration for single-cell microscopy. (D) Changes in fluorescence (ΔF/F₀ (%)) of the FRET donor and acceptor as well as Calbryte 630 evoked by octopamine (20 µM). Inset: image of the field of view with a single cell enclosed by an aperture.

Figure 11.. Simultaneous recording of cAMP and [Ca²⁺]_i signals in single cells using frequency- and spectrally-tuned multiplexing (FAST^M).

(A) Octopamine-induced (20 µM) [Ca²⁺]_i signals and changes in the FRET ratio (donor/acceptor), that is, cAMP signals, in the absence or presence of a non-binding (R307Q), lower (M329C), or higher-affinity (wildtype) FRET-based cAMP biosensor. Data points corresponding to the representative traces are labeled with a red asterisk in (B) and (C). (B) Onset of the octopamine-induced cAMP and [Ca²⁺]_i signals. (C) Comparison of the time to saturation of the cAMP signal and the onset of the [Ca²⁺]_i signal. The gray line depicts the theoretical perfect correlation.

Figure 11—figure supplement 1.. Affinity of mlCNBD-M329C for 8-NBD-cAMP.

Increase of 8‐NBD‐cAMP fluorescence (emission at 550 nm) upon binding to mlCNBD-M329C (1 μM). 8‐NBD‐cAMP fluorescence in the absence of mlCNBD-M329C was subtracted. The solid line represents a nonlinear least‐squares fit to ΔF = RL · x, with the receptor-ligand complex RL and the normalization factor x, that relates the concentration of bound 8-NBD-cNMP to the change in fluorescence (ΔF) (Cukkemane et al., 2007). The KD value was 0.12 µM.