Submillisecond optical reporting of membrane potential in situ using a neuronal tracer dye

Abstract

A major goal in neuroscience is the development of optical reporters of membrane potential that are easy to use, have limited phototoxicity, and achieve the speed and sensitivity necessary for detection of individual action potentials in single neurons. Here we present a novel, two-component optical approach that attains these goals. By combining DiO, a fluorescent neuronal tracer dye, with dipicrylamine (DPA), a molecule whose membrane partitioning is voltage-sensitive, optical signals related to changes in membrane potential based on FRET (Förster resonance energy transfer) are reported. Using DiO/DPA in HEK-293 cells with diffraction-limited laser spot illumination, depolarization-induced fluorescence changes of 56% per 100 mV (tau approximately 0.1 ms) were obtained, while in neuronal cultures and brain slices, action potentials (APs) generated a Delta F/F per 100 mV of >25%. The high sensitivity provided by DiO/DPA enabled the detection of subthreshold activity and high-frequency APs in single trials from somatic, axonal, or dendritic membrane compartments. Recognizing that DPA can depress excitability, we assayed the amplitude and duration of single APs, burst properties, and spontaneous firing in neurons of primary cultures and brain slices and found that they are undetectably altered by up to 2 microm DPA and only slightly perturbed by 5 microm DPA. These findings substantiate a simple, noninvasive method that relies on a neuronal tracer dye for monitoring electrical signal flow, and offers unique flexibility for the study of signaling within intact neuronal circuits.

Figure 1.

Fluorescence changes of DiO/DPA and eGFP/DPA FRET pairs in response to voltage steps. A, Confocal images of a HEK-293 cell labeled by incubation with DiO-C₁₈ (left), and a HEK-293 cell transiently expressing eGFP-F (right). Yellow cross is an example of a typical spot detection site. B, Fluorescence signals in response to voltage steps (indicated above) from cells in whole-cell patch-clamp mode labeled with DiO-C₁₈ (top), eGFP-F (middle), and DiO-C₁₆ (bottom), and superfused with an extracellular solution containing 5 μm DPA (top and middle) or 1 μm DPA (bottom). All traces are averages of five sweeps. Dotted line indicates zero change in fluorescence. C, Summary plots of peak amplitude fluorescence changes within 2 ms of the start of the voltage step. D, Summary bar graph showing average maximal ΔF/F for a 100 mV step from a holding potential (hp) of −70 mV. Black bars, DiO-labeled cells; 5 μm DPA, n = 6; 1 μm DPA, n = 7. Gray bars, eGFP-F cells; 5 μm DPA, n = 5; 1 μm DPA, n = 2.

Figure 2.

Quantification of voltage-dependent relaxation and rebound under sustained voltage steps. Summary plots of fluorescence amplitudes measured at three time points in Figure 1 B: the initial peak signal within 2 ms of voltage step (green), relaxation at the end of the voltage step (cyan), and rebound just following repolarization of the membrane (gray). Shaded gray area indicates the physiological voltage range from −70 mV to +40 mV.

Figure 3.

Temporal fidelity of DiO/DPA FRET pair. A, Fluorescence quenching responses to brief depolarizations from a HEK-293 cell bath labeled with DiO and incubated with 1 μm DPA. Membrane potential was stepped to +100 mV from an hp of −70 mV for durations of 0.1, 0.3, 1, 2, and 5 ms, yielding peak ΔF/F amplitudes of −14, −25, −28, −28, and −29%, respectively. Gray trace (τ rise) is a single-exponential fit to the initial quenching response of the 5 ms pulse. Black dashed traces (τ decay) are double-exponential fits of 0.17, 0.29, and 0.38 to the recovery of the quenched fluorescence response following termination of the 0.3, 1, and 5 ms step pulses, respectively. Each trace is an average of 10 sweeps and was filtered offline to 2 kHz. B, Correlation plot of the τ rise of fluorescence (+100 mV pulse for 5 ms) versus the weighted time constant of a double-exponential fit of the capacitance transients in response to a 5 mV step. Dashed line is a linear fit to all the data (Pearson correlation). The y-intercept is 0.12 ms and approximates the intrinsic time constant of the voltage-dependent quenching of the DiO fluorescence. C, Fluorescence quenching responses to trains of 1 ms, 100 mV depolarizations (hp = −70 mV) delivered at various frequencies (1000 Hz stimuli used 0.5 ms pulses). Traces were filtered offline to 3.5 kHz, and are averages of 10 and 15 trials for the 100 Hz and 300 Hz trains, respectively, and 17 trials for both 666 and 1000 Hz. DPA concentration was 1 μm. Inset is an expanded time scale of the 1000 Hz train (calibration: 4 ms, 12% ΔF/F).

Figure 4.

Optical responses of the DiO/DPA FRET pair faithfully follow AP voltage waveforms in single sweeps. A, Averaged fluorescence response (black trace; n = 17 sweeps) from a voltage-clamped HEK-293 cell labeled with DiO-C₁₆ and in the presence of 1 μm DPA to an AP waveform used as command voltage (gray trace). The record was filtered offline to 2 kHz and inverted for comparison to AP waveform. B, 100 Hz AP train waveform used as the command voltage for C and D. C, Seven consecutive fluorescence quenching responses to the train of AP waveforms (C), in the presence of 1 μm DPA. Right, The averaged trace (n = 11 sweeps). D, Seven consecutive eGFP-F fluorescence quenching responses to the train of AP waveforms (B), in the presence of 1 μm DPA. Right, The averaged trace (n = 11 sweeps). Single sweeps were filtered offline to 1 kHz. Scale bar in the bottom right of D corresponds to averaged traces in C and D.

Figure 5.

The DiO/DPA FRET pair reports high-frequency APs in cultured hippocampal neurons. Properties of APs recorded under current clamp in the presence of DPA are shown. A, Records of APs from two different hippocampal neurons 4 min after the formation of the whole-cell configuration, either in the absence of DPA (left, current injection 2200 pA) or after preincubation in 1 μm DPA for 19 min (right, current injection 1600 pA). Far right, Summary bar graphs of AP properties recorded in the absence of DPA (n = 12 cells) or after the 1 μm DPA preincubation (≥10 min, mean = 20 ± 4 min, n = 10). “WC time” is the average time at which APs were recorded following whole-cell formation, “AP height” is the difference between the peak amplitude and resting potential, and “AP width” is the width of the AP at half its maximal amplitude. No significant differences were observed (p > 0.1). DiO/DPA reports APs in neurons. B, DIC image of the cultured hippocampal neuron and recording pipette (top). Inset, Confocal image of the same neuron and the location of the optical recording (yellow cross) made during a 100 Hz train of current injections (1400 pA) to initiate APs. Shown are electrical recordings (middle) and corresponding optical recordings (bottom). Red traces are averages of six sweeps where there was a failure of AP generation in response to the first current injection. Blue traces are averages of 12 sweeps, where each current pulse successfully generated an AP. C, DiO/DPA reports APs in single trials. APs (red) were evoked by a train of five 1900 pA current pulses at 100 Hz (top). Ten consecutive single sweeps are shown with the average below. Individual sweeps were filtered offline to 1 kHz.

Figure 6.

The DiO/DPA FRET pair reports simple and complex cerebellar Purkinje neuron activity in brain slices. A, B, DIC (A) and confocal fluorescence (B) images of a Purkinje neuron located in a brain slice preincubated and superfused with 1 μm DPA and 1 min after cell-attached labeling with DiO-C₁₆. C, Single sweep AP-associated DiO/DPA responses (left) and their average (n = 19; bottom right) obtained from the Purkinje neuron displayed in A and B. Single APs were elicited by whole-cell current injection (2500 pA, 2 ms). Red traces are the current-clamp voltage recordings. Confocal image shows decreased membrane staining 24 min after DiO labeling, and yellow cross indicates the location of optical recordings. Right middle, Two plots of the amplitudes and widths, as a function of time, of the individual electrical (gray) and optical (black) AP recordings (scale of bottom plot is in milliseconds). D, DiO/DPA optical responses to single APs elicited by whole-cell current injection in the presence of 2 μm DPA. Bottom right, An average of 19 responses superimposed on the average electrical recording (red). E, Single and averaged (n = 10) optical responses to complex spike-like bursts elicited by larger current injections (3500 pA, 10 ms) from a different cell. All optical traces shown were filtered offline at 3.5 kHz.

Figure 7.

DPA does not disrupt firing properties of Purkinje neurons. A, Superimposed whole-cell current-clamp recordings of Purkinje neuron action potentials in response to a 2 ms step current injection. Displayed are single sweeps from three different neurons (black, control; blue, 2 μm DPA; red, 5 μm DPA). Note that the AP duration is prolonged in 5 μm DPA. B, Summary plot of AP duration measurements (full width at half-maximal amplitude) as a function of DPA concentration. Each data point corresponds to a different Purkinje neuron. Horizontal lines indicate mean values. C, Complex spike-like responses elicited with longer/larger current injections (10 ms, 3.5 nA) recorded in the indicated concentrations of DPA. Note that 5 μm DPA is associated with a loss of spikelets in the burst (see also Table 1). D, Extracellular recordings made from Purkinje neurons in slices incubated in control solutions or indicated DPA concentrations. Note that regular spontaneous firing rates are maintained in slices equilibrated in DPA. Vertical scale bar represents 45 pA for control, 15 pA for 1 μm DPA, and 30 pA for 2 and 5 μm DPA; horizontal scale bar is 2 s. E, Cumulative probability plots of Purkinje neuron spontaneous firing rates measured with extracellular electrophysiological methods in control (n = 82 cells) and DPA-treated conditions (1 μm, n = 34; 2 μm n = 34; 5 μm n = 45). All measurements have been made on Purkinje neurons in slices equilibrated in the indicated concentration of DPA for >45 min. No significant differences were observed (pairwise Kolmogorov–Smirnov tests, p > 0.05).

Figure 8.

Spot detection of soma, dendrite, and axonal voltage responses. A, Averaged optical responses (black traces) to complex spike-like bursts (red traces) recorded from the soma (n = 16 sweeps) and dendrite (n = 13 sweeps) of a Purkinje neuron indicated in the confocal image (yellow crosses). B, Optical responses to single APs recorded from different locations (yellow crosses) within the same cell. Averaged responses from the axon (blue, n = 9 sweeps), soma (black; n = 14), and dendrite (red, n = 10) are shown to the left. Traces have been temporally aligned with respect to the electrically recorded AP and superimposed below the micrograph. Responses were recorded in 2 μm DPA.