Imaging voltage in neurons

Abstract

In the last decades, imaging membrane potential has become a fruitful approach to study neural circuits, especially in invertebrate preparations with large, resilient neurons. At the same time, particularly in mammalian preparations, voltage imaging methods suffer from poor signal to noise and secondary side effects, and they fall short of providing single-cell resolution when imaging of the activity of neuronal populations. As an introduction to these techniques, we briefly review different voltage imaging methods (including organic fluorophores, SHG chromophores, genetic indicators, hybrid, nanoparticles, and intrinsic approaches) and illustrate some of their applications to neuronal biophysics and mammalian circuit analysis. We discuss their mechanisms of voltage sensitivity, from reorientation, electrochromic, or electro-optical phenomena to interaction among chromophores or membrane scattering, and highlight their advantages and shortcomings, commenting on the outlook for development of novel voltage imaging methods.

Figure 1. Biophysics of the plasma membrane

Illustration of the plasma membrane showing a simplified model (Gouy-Chapman-Stern) of the relevant structures, potentials, and distances involved in membrane voltage sensing (Olivotto et al., 1996). The cell’s overall potential, V_cell, is the difference in voltage between the bulk external media and the bulk internal media, as is governed by the concentration differences of ions in the two solutions. The local environment of the membrane has different potentials, however, reflecting electrical structures in the membrane. The negatively charged phosphate heads lead to strong polarization and alignment of water and ions immediately adjacent to the membrane. As one extends further into the bulk, the concentration of ions and water gradually transitions to that of the bulk and the field drops exponentially- the distance where the field drops to 1/e of the initial value is called the Debye length, and this region called the Debye Layer (d_D). In the figure, V_Esurf and V_Isurf represent the external and internal surface potential surrounding the bilipid layer. The field inside the membrane, V_mem is illustrated as homogeneous, an assumption that is clearly not true locally, in the presence of transmembrane proteins and pores.

Figure 2. Mechanisms of voltage sensitivity

Schematic of the physical mechanisms leading to voltage sensitivity in plasma membrane measurements, along with typical spectral signatures. Starting from top left, going across the figure: A) repartitioning, where the dye molecules move in and out of the membrane with voltage changes (see Table 1A); B) reorientation, where the electric field acting on the chromophore’s dipole produces a torque changing the relative alignment with respect to the membrane (see Table 1B); C) electrochromism, where the membrane potential changes the relative energy of the ground and excited states of the chromophore altering the excitation and emission wavelength (see Table 1C); D) FRET; where voltage induced conformational or spectral changes alter the efficiency of energy transfer (see Table 1D); E) collisional quenching is used in some hybrid schemes, where voltage induced motions lead to energy transfer, altering the fluorescence quantum yield and lifetime (see Table 1E); F) voltage induced dimerization/aggregation, where changing voltage induces aggregation of chromophores, altering the spectra; G) intrinsic imaging (complex refractive index changes due to action potential activity); H) SHG (electro-optic), where changes in voltage alter the effective χ⁽²⁾, modulating the SHG signal (see Table 1F); I) nanoparticles, not a mechanism per se, but used as a novel chromophore, or as an sensitivity “amplifier” for existing nearby chromophores.

Figure 3. Voltage imaging in mammalian preparations with one photon and two photon microscopy

(A) One-photon voltage imaging from neuronal ensembles from cat neocortex in vivo. A photodiode array was used to monitor the responses over a ~2 mm by 2mm patch of cortical area stained with RH795. Pseudo color images represented as averages over all of the captured frames. a) Visual stimulation (eyes open) evoked a change in membrane potentials. b) Spontaneous activity (eyes closed) of the same cortical territory. Note how the optical recording in both cases revealed almost identical patterns. Reprinted from (Tsodyks et al., 1999) with permission. (B) One-photon voltage imaging of back propagating action potentials in individuals dendritic spines of rat neocortical neurons in vitro. a) raw confocal image is shown on left, and deconvolved reconstructed image is show in the right. b) left shows the individual signals recorded at positions 1–3 as indicated in panel a, along with the electrical signal measured at the soma, while b) right shows the averaged result from 4 measurements. Reprinted from (Holthoff et al., 2010) with permission. (C) Two-photon voltage imaging in vivo. a) Colored areas in the barrel cortex mark regions were intrinsic imaging showed reflectivity changes of >0.1% following whisker stimulation with white line marking area of two-photon voltage responses. b) Averaged (n=400) trials of the VSD response for three different focal depths (40, 200, 400 μm), with dashed line indicating onset of stimulation. Reprinted from (Kuhn et al., 2008), with permission. (D) Two photon voltage imaging in vitro. a) Transmitted light image of a neuron in acute rat brain slice. b) Fluorescence image of a cortical pyramidal cell filled with the fluorescent dye FM4–64. Inset: zoom onto the 10 X 10 μm area outlined in the image showing dendritic spines. c, d) Point-dwelling ability and photon counting permit optical recording of fast events with while maintaining significant signal-to-noise ratios. Electrical (c) and unfiltered optical (d) traces of an action potential in a rat cortical pyramidal neuron loaded with the potentiometric dye di-2-ANEPEQ. Traces were averages of four recordings. Reprinted from (Vučinić and Sejnowski, 2007) with permission.

Figure 4. Novel modalities of voltage imaging

(A) Genetically-expressed voltage sensitive proteins can optically report membrane voltage of mammalian neurons. a) VSFP3.1_mOrange2 transfected into a cultured hippocampus neuron and expressed in the soma, axon, and dendrites (overview). High magnification image that demonstrated that VSFP3.1_mOrange2 is largely expressed in the plasma membrane with some fluorescence in perinuclear areas (insets). Scale bars are 40 μm in the overviews and 10 μm in the insets. b) Electrical (upper traces) and optical (bottom traces) recordings from a neuron expressing VSFP3.1_mOrange2. The voltage sensitive protein was able to sense evoked action potential bursts (upper right panel) and also spontaneous spikes (asterisks in the bottom panel) in a single sweep. Reprinted from (Perron et al., 2009), with permission. (B) The hybrid chemical sensor pair of DiO/DPA gives high fidelity for high-frequency bAPs. a) DIC and confocal (inset) image of cultured hippocampal neurons incubated with DiO and DPA. The region of interest is shown as a yellow cross in insets. b) Current injections at 100 Hz (top) evoked action potentials at the soma (middle), which induced voltage-dependent fluorescence changes (bottom). Optical traces were averaged: blue 6 trials, red 12 trials. The trial to trial fluctuations of the first episode during current injection led to successful (blue) and unsuccessful (red) firing of action potentials, which was accurately reported by the dye combination. Reprinted from (Bradley et al., 2009) with permission. (C) SHG signal captured membrane voltage transients in a hippocampal slice. a) Membrane restricted SHG signal was obtained by intracellularly loaded FM4-64 via a recording pipette. b) Line-scan recordings (red line in a) of SHG along the somatic plasma membrane revealed action potentials with high fidelity (S/N ~7–8) after averaging (n=55). c) The somatic membrane potential was monitored in current clamp mode. Super threshold depolarization elicited action potentials. Reprinted from (Dombeck et al., 2005), with permission. (D) Quantum dots (Q-dots) can sense electric field changes in mammalian cells. a) Q-dots successfully targeted the plasma membrane of cultured hippocampal neurons. b) Q-dots showed strong modulations in fluorescent intensity to membrane voltage changes induced by altering the potassium concentration surrounding the cells. Reprinted from (Fan and Forsythe, 2008), with permission