• absolute membrane potential;
  • calibration;
  • flow cytometry;
  • image cytometry;
  • Nernstian dye

Membrane potential is an important physiological parameter of the living cell, which besides its traditional role in neurosciences, gains increasing attention also in the context of circulating cells by recognizing its role in different transmembrane-signaling processes such as immune-recognition [1], and in apoptosis [2, 3]. Besides its direct measurement with electrophysiological techniques (electrodes and patch-clamp), indirect techniques also exist. These are based on measuring fluorescence of either the ionic dyes—such as the cationic 3,3′-dihexyloxacarbocyanine (DiOC6(3)) and the anionic bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4(3)) or bis-oxonol—responding with altered translocation or ratiometric dyes—such as the merocyanine 540 (MC540), the 5,5′,6,6′-tetrachloro-1,1′,3,3′ tetraethylbenzimidazolocarbocyanine iodide (JC-1), the 1-(3-sulfonatopropyl)−4-{β[2-(di-n-octylamino)−6-naphthyl]vinyl}pyridinium betaine (di-4-ANEPPS), and the 3-hydroxiflavones (e.g. F2N12S)—responding with spectral shifts (electrochromism) to the changes of membrane electric fields [2-6].

In spite of the long time elapsed since the introduction of the first membrane potential indicating dyes, the details of their working mechanisms remained largely unknown [6]. The complexity arises partly from the heterogeneity of fluorescence of the dyes taken up by the cells, due to the strongly heterogeneous environment—implying a need for separating fluorescence contributions of cell organelles, e.g. mitochondria—, the presence of bound and unbound (free) dye states, and partly from the composed nature of the membrane electric field itself (Fig. 1). The membrane electric field has also surface potential and dipole potential components caused by the charged lipid headgroups and protein modifications in addition to the transmembrane potential component caused by the asymmetric bulk ion distributions at the two sides of the membrane [4, 5]. The complexity is further enhanced by the fact that all these factors are capable of modulating the translocation, accumulation, and emitted fluorescence of a potential indicator dye in a cell-type dependent manner (e.g., via the relative amount of the free and bound dye states dictated by the pool of intracellular proteins) precluding the existence of a universal calibration curve between fluorescence intensity and membrane potential [7-9].


Figure 1. Schematic representation of the electrostatic potential profile across the cell membrane, made in coherence with Ref. [4]. Surface potential Ψs generated by the negatively charged lipid headgroups is the potential drop between the bulk extra- (or intra-) cellular space and the membrane surface, dipole potential Ψd generated by the dipolar lipid groups just under the membrane surface is the potential drop between the membrane surface and the middle of the bilayer, transmembrane potential Ψ generated by the asymmetric ion distribution at the two sides of the membrane is the potential drop between the extra- and intra-cellular bulk fluid compartments. Es, Ed, and E (symbolized by filled arrows) are the corresponding electric field strengths spanning only those regions where potential changes occur, pointing in the directions of potential drops, and having magnitudes (symbolized by proportional arrow widths) of the rates of potential changes (“negative potential gradients”). Remarkable features are that: (i) Although the dipole and surface potential fields occur in a pairwisely counterbalancing manner at the two surfaces of the membrane, they might influence membrane permeability and translocation kinetics selectively for ionic dyes such as oxonol. In this respect it should be mentioned that because the dipole potential cannot be measured directly with electrodes (“hidden” membrane potential component), its existence might be inferred only via indirect observations, such as the experience that the membrane permeability for negative ions is larger than for positive ones. (ii) The sign of the dipole potential is positive and its magnitude can be much larger than that of the transmembrane potential (e.g., +300 mV compared to −100 mV), indicated also by the associated field vectors (Ed) having width much thicker than that of the transmembrane field (E). [Color figure can be viewed in the online issue which is available at]

Download figure to PowerPoint

In their recent works Klapperstück et al. (9, and in this issue, page 612), by following the direction started by Krasznai et al. [7, 8], show an elegant and efficient way of determination of absolute membrane potential by using the negatively charged translocating dye bis-oxonol (DiBAC4(3)) in a flow cytometer. The essence of the method is that the same oxonol fluorescence vs. extracellular dye concentration calibration curve recorded for any known value (e.g., zero) of the membrane potential can be used for determining the membrane potential of any sample of the same cell type. For a practical example of oxonol fluorescence calibration and details of the principle of absolute membrane potential determination, please see Figures 2 and 3. The only prerequisite is that a known membrane potential (Ψ0 in Fig. 3) version of the sample as a reference for the extracellular dye concentration (De,0 in Fig. 3) should also be available, accomplished, e.g. by nulling of membrane potential with the pore-forming ionophore gramicidin. The method is based on the following assumptions: (i) Factors determining fluorescence of the uptaken dye—i.e. the ratio of free and bound intracellular dye concentrations—are independent of membrane potential, or equivalently, the sole role of membrane potential is in determining the amount of the dye taken up by the cells (“Nernstian behavior”) [7-9]. (ii) The extracellular free dye concentrations are not influenced by the membrane potential and possible changes in cell concentration levels.


Figure 2. 3-D surface plots representing the two phases of indication of membrane potential by the anionic oxonol dye: the potential dependent dye-uptake by the cells (Panel A), and fluorescence emission by the uptaken dye (Panel B). Designations: external dye concentration De, internal dye concentration Di, membrane potential Ψ, fluorescence intensity I. The 1st phase, in Panel A, indicates the potential dependent dispersion of dye-uptake (“Nerstian distribution”). While the oxonol dye is readily taken up at small magnitude (close to zero) membrane potentials (slopes of the Di vs. De curves are close to unity), it is effectively excluded at large magnitude (considerably smaller than zero) membrane potentials (slopes of the Di vs. De curves are close to zero). In practice, in flow cytometry, the dispersion of the measured dye-uptake is caused partly by the dispersion of membrane potential (intracellular ion concentrations) of the cells, and—because the measured parameter is not the dye concentration but the total amount of dye incorporated by the cell—partly by the dispersion of cell volume. The range of Di values is always compressed relative to that of De (i.e., Di<De) due to the exclusion of the anionic dye by the negative potentials. The 2nd phase, in Panel B, refers to the fluorescence emission of the uptaken dye. The main features are that while at low dye-concentrations fluorescence depends approximately linearly on the concentration, at high concentrations the fluorescence shows saturation. The plots were computed based on the binding saturation model of oxonol fluorescence and experimentally observed parameters of resting JY B-lymphoblast cells (type “A”) taken from Emri et al. (8).The surface plot of Di in Panel A was calculated as inline image with Ψ° = 25.6 mV (calculated as inline imagewith R universal gas constant, T = 293 K room temperature, and F the Faraday-constant). The surface plot of fluorescence intensity I was calculated by using inline image The parameters inline image, inline image, and inline image describe the contributions to the emitted fluorescence of the free and bound forms of intracellular oxonol (saturating binding), obtained from the corresponding a1, a2 and a3 parameters in Table 2 of Ref. [8] by eliminating the potential-dependent exponential factors via inline image = a1/exp(−63.4/25), inline image = a2/exp(63.4/25), inline image = a3/exp(63.4/25) resulting in inline image = 8.3 µM−1, inline image = 92.9 µM, inline image = 4.5 µM. It was also assumed in the computation that these binding parameters are valid not only at the experienced resting potential (−63.4 mV) of the type “A” JY cells, but also in the whole 0-(−100) mV range. [Color figure can be viewed in the online issue which is available at]

Download figure to PowerPoint


Figure 3. Schematic representation of the steps of absolute membrane potential determination with oxonol in “3-ways coordinate systems.” Horizontal right axes designate oxonol fluorescence intensity I, horizontal left axes extracellular dye concentration De, vertical axes intracellular dye concentration Di. The straight lines at the left describe the “Nerstian dye-uptake” of the cells (strictly valid for small concentrations), the increasing curves at the right are the emitted fluorescence I vs. intracellular dye concentration Di calibration curves. The “3-ways coordinate axes” have been set up to follow the logic of membrane potential indication by oxonol. Namely, that it can be decomposed into two independent phases: a dye-uptake dictated by the amount of extracellular dye, membrane potential and cell volume and light emission dictated by the amount of intracellular dye and its distribution according to the free and bound forms, or possibly according to the monomeric and oligomeric states at large concentrations. As to the calibration phase (Panel A), it is noted that due to the different membrane potentials (Ψ1 and Ψ2) the same extracellular dye concentration (De,1) can lead to different fluorescence intensities (I1, I2), and vice versa, the same fluorescence intensity (I1) can be the result of two different extracellular dye concentrations (De,1, De,2), precluding the one-to-one correspondence also between membrane potential and fluorescence. Not the shape (the functional form) but the range of the I–Di calibration curve is the characteristic which is affected by the membrane potential at which it was recorded. As an illustration, the De axis is mirrored onto the Di axis by the line marked by potential Ψ1 closer to the origin than by the line marked by a more positive potential Ψ2, i.e. the Di range of calibration is larger at more positive membrane potentials. Although calibration can be done at any membrane potential (e.g., via adjusting extracellular ionic milieu and/or adding ionophores), nulling of potential with gramicidin is favored, because both the accuracy of the potential value adjustment and the range of calibration are here the largest ones (the slope of the DiDe straight line is 1 at the maximum for non-positive membrane potentials). In contrast, the shape of the corresponding I–De curves does depend on the membrane potential: a smaller or larger portion of the I–Di curve in the direction of the origin of the I–Di coordinate system is stretched to the whole De axis proportionally to the membrane potential value. As to the potential read-off phase (Panel B), it is noteworthy that taking only a single Ix intensity reading of a sample having the unknown membrane potential Ψx at extracellular dye concentration De,0, it can be interpreted as being due to both the De,x(false)-Ψ0 and the De,x(true)-Ψx value pairs. However, by taking also a second fluorescence reading I0 of another sample measured at a known membrane potential Ψ0—e.g. via potential nulling with gramicidin (Ψ0 = 0 mV), or via adjusting the extracellular ionic milieu—and at the same extracellular dye concentration as for Ix, De,0, uniquely fixes the extracellular dye concentration at De,0. Consequently, the unknown membrane potential Ψx for Ix, is De,x(true)(=De,0) also fixed (see also the formulae for Ψx and Ψ0 in Panel B, where Ψ° is the room temperature reference value of membrane potential, 25.6 mV). For the applied extracellular dye concentrations, small values (< 100 nM) are favored, because in this regime ratios of intracellular dye concentrations (Di,x/Di,0) in the formula for Ψx (in Panel B) can be replaced by the ratios of the corresponding fluorescence intensities (Ix/I0) due to the approximate linearity of the I–Di calibration curve. This implies also that at low extracellular dye concentrations there is no need for a calibration curve. Actually, the need for an I–De (or I–Di) calibration curve is posed by its possible nonlinear nature and zero offset. [Color figure can be viewed in the online issue which is available at]

Download figure to PowerPoint

Emri et al. has demonstrated that the intracellular dye partition can be taken as independent from the membrane potential for the case of JY cells [8], by showing that the a2/a3 ratio of their dye-partition model stays constant. The constant concentration of extracellular dye is nicely demonstrated by Klapperstück et al., (in this issue, page 612), to be valid for small cell and extracellular dye concentrations (<0.5 × 106 cells/ml and <100 nM, respectively). They also showed that at low extracellular dye concentrations (<100 nM), the fluorescence vs. extracellular dye concentration calibration curves can be taken as fairly linear, with the consequence that the ratios of intracellular dye concentrations necessary for the calculation of membrane potential can be replaced by the ratios of the corresponding fluorescence intensities, indicating the robustness and simplicity of the method at these dye concentrations (see also Fig. 3, Panel B).

Although, the elaborated method is already useful as it stands at the level of cell populations, it has the shortcomings that it does not lead to increasing the accuracy of membrane potential determination at the single cell level, i.e. it does not reduce the width of the distribution of the membrane potential-related fluorescence signal [6]. To achieve this goal, ratio mode detection could be mentioned. By simultaneously measuring and forming the ratio, a suitable potential-dependent and a potential-independent signal, the volume fluctuation of the amount of uptaken dye drops out considerably reducing uncertainty of potential indication [6]. A further possibility may also be offered by the detection of fluorescence anisotropy of the oxonol dye, which is inherently ratiometric. Free state–bound state discrimination of the uptaken dye (i.e., resolving the calibration curve according to the contributions of the free and bound dye forms) could be accomplished, e.g. based on the different degrees of rotational mobility and different lifetimes of the free and bound forms, manifested in different degrees of fluorescence anisotropy (governed by the ratio of the lifetime and rotational correlation time, τ/ϕrot), a parameter, the measurement of which is feasible in most commercially available flow cytometers [10]. Although more complicated, the same resolution of the calibration curve could be achieved by direct fluorescence lifetime measurements of the oxonol dye in a fluorescence lifetime imaging microscope (FLIM) or also in flow cytometry, in the phase-modulation measuring scheme of lifetime realized in some laboratories.

In addition to the practical value of measuring membrane potential more accurately, the importance of research on the fine details of membrane–dye interactions is underscored by the facts that: (i) Changes in membrane potential may be coupled to mechanical alterations of the cell membrane, e.g. changing bilayer thickness with the consequence of changing lipid packing [11] and protein accommodation via “hydrophobic mismatch”. (ii) The dipole potential, a “hidden component” of the membrane potential [4, 5] representing an energy valley for the membrane incorporated dye, may also influence the free–bound dye partition. In this respect, the multiplexing capability of flow cytometry can be exploited enabling correlated detection of membrane mechanics and oxonol fluorescence. The intrinsic optical property for monitoring membrane mechanics is birefringence measurable in the form of 90° light scattering (side or perpendicular light scattering) and/or its degree of polarization [11, 12]. Membrane mechanics can be further coupled to the dipole potential via the degree of lipid packing, the possibility of which can be checked by measuring fluorescence of a suitable polarity sensitive dye, e.g. F2N12S from the 3-hydroxiflavone family [5]. The ultimate solution would be multiparameter flow cytometry with 2–3 different signal channels depending on the spectral separability, for monitoring bilayer mechanics (via birefringence), transmembrane potential (via e.g. fluorescence of oxonol, or other “Nernstian dye”) and dipole potential (via fluorescence of a suitable polarity sensitive dye). This kind of a measuring arrangement, through separating the effects of dipole potential and transmembrane potential would give a more thorough description of the effects of membrane electric fields exerted on the different transmembrane signaling processes. There is a growing demand for this approach also from a theoretical viewpoint of describing membrane potential-related effects based not only on pure electricity but involving also lipid mechanics (electromechanics) [11].

Literature Cited

  1. Top of page
  2. Literature Cited