Noninvasive measurement of hydrogen and potassium ion flux from single cells and epithelial structures

P. J. S. Smith, J. Trimarchi


This review introduces new developments in a technique for measuring the movement of ions across the plasma membrane. With the use of a self-referencing ion-selective (Seris) probe, transport mechanisms can be studied on a variety of preparations ranging from tissues to single cells. In this paper we illustrate this versatility with examples from the vas deferens and inner ear epithelium to large and small single cells represented by mouse single-cell embryos and rat microglia. Potassium and hydrogen ion fluxes are studied and pharmacological manipulation of the signals are reported. The strengths of the self-referencing technique are reviewed with regard to biological applications, and the expansion of self-referencing probes to include electrochemical and enzyme-based sensors is discussed.

  • ion-selective electrodes
  • self-referencing probes
  • vas deferens
  • embryos
  • microglia
  • inner ear

this review introduces the application of a new technique for studying tissues and cells by monitoring the movement of molecules in close proximity to the plasma membrane with the use of self-referencing probes. The principle underlying this approach has long been established, but the recent incorporation of different sensors provides novel approaches to cell physiology. Here we aim to draw attention to these new developments by presenting selected examples where detection of hydrogen and potassium ion flux is providing biological data from systems as diverse and complex as epithelia and single cells. We conclude by discussing opportunities for self-referencing electrochemical and biosensor probes making possible detection of molecules such as oxygen, nitric oxide, and glucose moving across the membrane of single cells. In this review the focus is on applications of this methodology. The technical details have recently been reviewed elsewhere (61).

There is a clear discontinuity between the ionic composition of the inside and outside of cells. In the case of potassium, the internal activity of this ion ([K]i), is more than 50 times that of the bulk medium. However, the boundary between a living cell and the surrounding medium is far from simple. It is easy to lose sight of the fact that there can be a considerable disparity in the ion activities in close proximity to the cell membrane and those in the bulk medium. Boundary conditions and surface charges immediately at the interface between the cell membrane and the medium can, in theory, drastically alter the ionic composition of the cell-to-medium boundary (46,47). Surface and boundary voltages, and their related effects on ion accumulation, have been extensively modeled (12) and are largely confined within the Debye length (∼8 Å in saline). This is a very difficult region in which to make direct measurements. However, there are other disparities between the immediate environment of a cell and those of the bulk medium that are amenable to direct measurement. As ions, or molecules in general, move between a cell and its immediate surroundings, there inevitably exists a diffusion gradient within micrometers of the cell surface. If transport is steady and maintained over the long term, a chemical gradient will be established (Fig. 1). For example, in the case of a net influx, there will be a depletion surrounding the site of transport. This gradient can be measured noninvasively with a suitable probe positioned within its boundaries, revealing new facets of cellular transport mechanisms that complement the more familiar approaches to cellular electrophysiology.

Fig. 1.

A diagrammatic and simplistic representation of the conditions in close proximity to the medium exposed surface of a cell. Within angstroms of the membrane there is a complex microenvironment influenced by surface voltages, boundary conditions, and molecular structure, represented here by the solid gray zone. Beyond that zone, ions exist in solution with an activity that is influenced by the net transport across the plasma membrane. Here, the stippling density is meant to impart the conditions of efflux, where a concentration of ions has built up near the membrane. As the distance increases, the activity diminishes as ions diffuse into the bulk medium. Conversely, there would be a rarefaction of a substance being absorbed. Seris measurements are within the limits of diffusion gradients but away from the complex boundary zone. A Seris probe is shown at the two points of measurement, which would normally be ∼10 μm apart. The tip diameter of the probe is between 2 and 4 μm, and the plasma membrane is measured in nanometers. Clearly, this drawing is not to scale.Inset: an expansion of the membrane illustrating the local changes in ion activity that would be expected from the activity of a transporting molecule. Note that the technique does not monitor these individual events but the average of their activities (62).

One approach to assessing the movement of ions in proximity to the cell membrane requires the sampling of the local activity at more than one point with the use of multiple electrodes. This is difficult, not because there is a lack of suitably selective sensors but because all electrodes suffer from pronounced drift. This drift is of such a magnitude that it reduces the application of ion sensors to comparatively large changes in ion activity, such as between the experimental medium and within the cytosol of the cell. One solution that greatly enhances sensitivity for the extracellular detection of ion gradients is to borrow an older approach for extracellular detection of current densities and use a single self-referencing probe. As previously published for self-referencing calcium ion flux detection (61), a glass microprobe, front-filled with an ion-selective liquid membrane, is moved in a square-wave translation such that the sensor rests for 1 s at each of two positions ∼10 μm apart. Via this method, ion activity can be locally sampled at known positions from each other and the cell membrane. One of these positions can be within a micrometer of the cell membrane, but both should lie within the diffusion gradient, thereby allowing the calculation of ion flux. This is the basis of a noninvasive, self-referencing, chemically selective probe and is the subject of this review.

Self-referencing probes have a relatively long history, particularly with regard to voltage detection. The first example of a noninvasive, self-referencing voltage probe, applied to a biological system, was published in 1950 (5), when an ingenious vibrating platinum probe was applied to the measurement of surface voltages from plants. This pioneering study preceded the subsequent development of other aerial probes achieving similar voltage resolutions (3, 20,53). Most importantly, these first authors foresaw the use of this device in conductive media in which diffusion potentials could be measured. Despite an early report on the measurement of external voltage fields in a liquid medium (15), it was not until the seventies that a complete description of a self-referencing voltage probe appeared (24). In that design the probes had a sensitivity of low nanovolts over sampling distances of tens of micrometers. That device, commonly referred to as the “vibrating probe,” achieved its voltage resolution by minimizing the impact of drift through the simple expedient of using a single electrode to compare voltages at two positions micrometers apart. The probe capacitively coupled to the external voltage field and used phase detection to isolate signals coherent with the frequency of vibration. The differential voltage acquired through this technique could be converted into a current through Ohm's law. The technique has been reviewed on several occasions, most recently in 1990 (48).

There were indications that the same principle of self-referencing could be applied to the selective detection of ion activity and, therefore, ion flux across the plasma membrane. For example, pH gradients can be observed along the length of fungal hyphae by using a pH electrode to locally sample the activity of hydrogen ions and comparing these values with those in the bulk medium (19). The electrode was positioned at the recording site by manually operated positioners. In 1987 a significant step was made in producing an automated self-referencing, ion-selective (Seris) probe when an artificial gradient was accurately followed beyond the levels possible with the use of static electrodes (23), a concept that found its first biological application in 1990 (33). Both of these early studies (23, 33) focused on calcium and established the Seris probe as a powerful new tool for monitoring the net trans-plasma membrane movement of this important ion. This experimental approach has subsequently found numerous and compelling applications in calcium detection, where results not only support and reinforce data acquired through other methodologies but offer new insights into cellular regulation of calcium transport and the role this ion plays in cell physiology (see Refs. 60-62for review). The ability to detect the activity of an ion, inside or outside a cell, depends on the suitability of the sensor. In the application discussed here, ion-selective liquid membranes are used. These membranes have seen extensive application in cell biology but can be restricted by complicated additive voltages not dependent on the ion being targeted. The complications, for example, can come from junction potentials, direct voltage detection, and nonspecificity of the ionophores (1). The problems are encapsulated in the Nikolsky-Eisenman and Nernst equations, which define the selectivity and performance of a liquid membrane.

The application of these two pivotal equations and how a Seris probe can be built and applied to the local detection of ion activity was dealt with in depth in a previous paper (61), in which the subject was the detection of calcium ion movements. Several important features of Seris probes were dealt with in that paper, notably drift characteristics, response times, and the problem of contaminating chemicals. Most features involved in calcium detection apply generally to the measurement of all ions and are not repeated here. The purpose of the current review is to draw attention to the application of the Seris approach to the detection of hydrogen and potassium ions in the study of single cells and epithelial structures. It is worth noting that the application of this technique can be expanded to include any ion where a suitable ion-selective membrane and physiological subject coincide [see, for example, cadmium detection (51)].

The diversification of the self-referencing ion probes from calcium to other molecules was not led by demand from the animal sciences but from botany (31). The attraction lies in an inherent advantage of a noninvasive technique in plant studies because it avoids the need to penetrate the relatively rigid cell wall, where the contents are often under pressure. Plant sciences have made considerable use of self-referencing technology, and a list of probe-related publications illustrating the diversity of biological applications, including uses in mycology and botany, can be found The first Seris plant study (31) demonstrated that the commercially available hydrogen and potassium ion liquid membranes [such as Fluka hydrogen ionophore I (tridodecylamine) in cocktail B and Fluka potassium ionophore I (valinomycin) in cocktail B] could be used in a manner similar to the liquid membrane for the detection of calcium. The results in this case showed pronounced hydrogen and potassium ion fluxes around roots. Although the probe was originally conceived for single cell studies, these results foreshadowed an unexpected strength of the Seris probes: their ability to aid in the characterization of ion-transporting mechanisms embedded within epithelial structures. Our first two examples, the mammalian vas deferens and the cation-transporting epithelia of the inner ear, illustrate epithelial applications. Our second two examples move from the tissue to smaller targets, individual cells, represented by the mouse single-cell embryo and primary cultures of single microglia. These examples illustrate the ability of a Seris probe to “microsample” the ion activities in close proximity to cellular structures, measuring the diffusion gradient and ion fluxes.


Mammalian spermatozoa are maintained in a quiescent state as they pass through the caudal epididymis and vas deferens. An acidic pH, ∼6.5 (11, 35), is essential to achieve this immobility (22, 25), but the mechanism for achieving this acidity has been unclear. Studies performed before those under review here used perfused epididymis and cultured cells to implicate a role for an apical Na+/H+ exchanger in establishing the acidic environment (2). However, studies in which antibodies were used against the vacuolar (V-type) hydrogen pump, a nonphosphorylating H+-ATPase, beautifully demonstrated the presence of this pump in the region of the apical membrane of cells within the proximal to middle part of the vas deferens (Fig.2, A and D) (9). This finding raised the possibility that this pump plays a role in luminal acidification.

Fig. 2.

Immunohistochemical localization of the vacuolar-type proton pumps and carbonic anhydrase II (CAII) in the cells of the proximal section of the rat vas deferens and caudal epididymis. A–C: localization of the vacuolar H+-ATPase (A) in a subpopulation of cells in the proximal vas deferens with the colocalization of CAII (B). A and Bare merged in C. The scale bar in A (10 μm) is common to images in B and C. Images are adapted from Breton et al. (7). D–F: effect of treatment with colchicine on the pump distribution in the caudal epididymis. The normal distribution (D) of vacuolar H+-ATPases changes (E and F) as colchicine disrupts the microtubules and prevents the organized delivery of the pumps to the apical membrane. Images are adapted from Breton et al. (8).

By taking a proximal section of the vas deferens and splitting it open along its axis, it is possible to scan over the surface of the apical epithelium with a hydrogen-selective Seris probe (H+probe). A substantial flux of hydrogen ions is measured, showing regions of higher flux thought to occur over patches of pump-containing cells (9). Application of bafilomycin A1, a selective blocker for the V-type pump, drastically reduces the hydrogen ion flux (Fig. 3), producing direct evidence that the V-type pumps are responsible for most of the acidification in the proximal region of the vas deferens.

Fig. 3.

Seris H+ probe measurements of proton activity in the medium within 20 μm of the apical membrane of the rat proximal vas deferens. A: summary of results from studies targeting the efficiency of proton pumping through the vacuolar-type pump. Bafilomycin A1 (Bafilo A1) is a relatively specific blocker of the proton pump. Tetanus toxin (TT) cleaves cellubrevin and prevents the vesicles trafficking the pumps from docking with the membrane. Acetazolamide (Actz) inhibits CAII, reducing the production of H+ for pumping. All these compounds cause a significant reduction in acidification above the apical membrane. B: schematic of the proposed mechanisms underlying this process in the proton pump (PP)-rich cells. DPC, diphenylamine-2-carboxylate; SITS, 4-acetamido-4′-isothiocyanatostilbene-2,2′-disulfonic acid.C: results from experiments targeting the movement of bicarbonate ions out of the cell on the basolateral surface using compounds such as SITS, which targets a transporter, and DPC, which targets a proposed Cl/HCO3 channel. DPC on its own does not cause a significant change in proton transport, but when combined with Cl-free medium, it significantly reduces H+ flux. Data are presented as percentages of control values, thus allowing different experiments to be graphed together.

Hydrogen ion pumping requires a hydrogen ion source. In the mammalian testis and epididymis, studies have demonstrated the presence of carbonic anhydrase (CA) (13, 17, 18). At least two isoenzymes are found in the epididymis, the membrane-associated CAIV and the cytosolic CAII, with colocalization in the vas deferens between the V-type pumps and CAII (Fig. 2, B and C) (7, 9). Inhibition of CAII with acetazolamide eliminated the bafilomycin-sensitive component of the hydrogen ion flux (Fig. 3) (7), confirming a role for this enzyme in hydrogen ion generation in these cells. Of interest is the inability of compounds targeting the V-type proton pump or its performance to shut down the hydrogen flux entirely; only 60–80% of the signal is lost. A possible, but as yet unproven, explanation is that the residual flux is an indirect result of CO2 production by cellular respiration reacting with water outside the cell to produce hydrogen ions and bicarbonate. The probe will detect the generated hydrogen ions as a directional flux of protons coming from the cell surface.

By analogy with the kidney type A intercalated cells, we might expect an electroneutral Cl/HCO3 exchanger (AE1), coupled to basolateral Cl channels and CAII, to aid in the generation of hydrogen ions for apical pumping. However, this exchanger could not be identified in the basolateral membranes with an antibody that detects both the AE1 and AE2 isoforms (9). This left the puzzling question as to the role of Cl and HCO3 transporters in the process of vas deferens apical membrane proton transport. With the use of a combination of chloride ion removal as well as the application of 4-acetamido-4′-isothiocyanatostilbene-2,2′-disulfonic acid (SITS) and diphenylamine-2-carboxylate (DPC), further studies with Seris H+ probes have concluded that a Cl/HCO3 exchanger is not involved, but the nature of the HCO3 transporter is not known (7). The action of DPC in the absence of chloride indicates the possible presence of a Cl/HCO3 channel (7). Figure 3 presents a model of the proton pump (PP)-rich cell and a summary of these results.

The difference in chloride dependence between the hydrogen ion-pumping cells of the proximal vas deferens and the intercalated cells of the kidney shows that there are differences in the mechanisms for generating an apical hydrogen ion flux (10). However, some processes and control mechanisms might still be common to the two tissues. An obvious comparison would be whether in the vas deferens, as in the kidney, the V-type pumps are trafficked to the apical membrane in a soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE)-dependent manner. This question has been examined in Seris studies in conjunction with immunohistochemistry and Western blotting before and after treatment with tetanus toxin (8). This toxin is known to cleave the vesicle-bound SNARE cellubrevin, preventing vesicle docking with the membrane. Cellubrevin colocalizes with the V-type hydrogen ion pumps (8), and tetanus toxin treatment inhibits the bafilomycin-sensitive proton secretion by 64% (Fig. 3). Western blotting demonstrated the expected cleavage of the cellubrevin in the intact preparation. Furthermore, treatment of the tissue with colchicine, a disrupter of the microtubule structure, causes a marked redistribution of the V-type H+- ATPase from the apical membrane to the cytosol (Fig.2, E and F). These results strongly support the model of SNARE-dependent vesicle trafficking to the apical membrane.

One complication, unique to the measurement of hydrogen ion flux, is the almost ubiquitous presence of a buffer in the bulk medium. This will inevitably reduce the activity of hydrogen ions as they diffuse from the site of transport, thus maintaining the bulk pH. Two attempts have been made to accommodate the buffering effect, deriving a valid and quantifiable flux measurement. The first attempt at this correction (16) was used in the data discussed below for microglia (59). This approach required an estimate for the diffusion constant of the buffer. The second equation, derived by D. M. Porterfield (unpublished observations), reads as followsJ=D(Δ[H+]+[Buffer]·0.25Δ[H+]·Ka1)Δr1 where J is the flux, D is the diffusion coefficient for hydrogen ions, Δ[H+] is the change in hydrogen ion activity between the two poles of measurement (calculated as previously described in Ref. 61), [Buffer] is the buffer concentration (expressed in mol/cm3),K a is the pK a of the buffer (expressed in cm−3), and Δr is the distance (in cm) between the two measuring positions of the Seris probe. The important factor in this equation is the constant 0.25, which relates the buffering slope to the pK a. The buffering slope is the relationship between pH and the buffer concentration. The advantage of this approach is the ability to use this equation with different buffers. A practical example is given in Fig. 4, for which data were collected from the isolated retinal cells of the skate (37). Raw data are shown (Fig. 4 B) and then converted to flux values before (Fig. 4 C) and after correction for the buffer (Fig.4 D). As would be expected, the flux is the same in both 2 and 25 mM HEPES, once corrected.

Fig. 4.

Correction of H+ flux data by taking into account the concentration of the pH buffer in the bulk medium.A: Seris H+ probe photographed in the near pole position during measurements from a single horizontal cell isolated from the retina of the skate. In this experiment the position of closest approach was ∼1 μm from the membrane, and the far pole was 50 μm away. Background measurements of the bulk H+activity were made at a position 160 μm above the cell. Scale bar = 50 μm. B: recording acquired in close proximity to an external horizontal cell. Starting from left, the trace indicates the size of the differential signal recorded from 1 cell bathed in a skate-modified Ringer solution containing 2 mM of the pH buffer HEPES. Under these conditions, a differential signal of ∼85 μV was recorded. At background (Bkg) no differential signal could be detected. Returning the electrode to its initial position (at cell) restored the signal. Replacing the bulk solution with another containing 25 mM HEPES reduced the signal to ∼10 μV. After the solution with 2 mM HEPES was returned, the value was restored.C: calculation of proton flux ignoring the presence of the buffer. Values represent means ± SE of fluxes for the cell shown in B. Flux was calculated as the diffusion constant multiplied by the calculated change in H+ concentration, divided by the excursion of the electrode (62).D: calculation of proton flux accounting for the presence of the pH buffer HEPES. Note that the flux of protons out of the cell is the same in both 2 and 25 mM HEPES. Values represent means ± SE of fluxes for the cell shown in B. A correction made using this method does not discriminate between variation from the biological preparation and noise from the electronics and overall system. This explains the large SD of the mean after correction in D. The original data (39) and corrected flux values were kindly provided by Paul Malchow (University of Illinois, Chicago, IL).


Sensing static position and body motion involves the modulation of a transepithelial current through the inner ear neuroepithelium. This current is carried by potassium ions moving down the electrochemical gradients across the basolateral and apical membranes of the epithelium; the concentration of potassium in the inner ear lumen is 25–30 times that in the perilymph bathing the basolateral epithelial surface (56). Indirect evidence suggests that a specific cell type within the epithelium, the vestibular dark cells, establishes and maintains the transepithelial potassium gradient by transporting potassium across the basolateral surface into the cell cytoplasm and from there to the lumen of the inner ear (4, 30,55). By employing vibrating voltage probe technology to derive near-field currents and Seris probes to measure the ionic composition in the proximity of epithelial cells, it was possible to conclusively identify vestibular dark cells as being responsible for transporting and concentrating potassium within the inner ear lumen (42). The molecules participating in the potassium transport across the vestibular epithelium were characterized (42, 43, 57, 65) by combining Seris technology with the use of a micro-Ussing chamber for measuring transepithelial voltage, resistance, and current and with the use of patch-clamp techniques for monitoring whole cell currents from individual epithelial cells, as well as with pharmacology. Seris technology has identified similar mechanisms underlying potassium transport across cochlear (acoustic) epithelium (65).

Initial electrophysiological investigation of the inner ear epithelium identified a transepithelial current by sealing the epithelium over the aperture of a micro-Ussing chamber. This current could be modulated by shifts in the ionic composition of solutions on either side of the epithelium (39, 40, 67). Although the micro-Ussing chamber (80-μm diameter) allowed determination of the electrogenic properties of the epithelium, the specific ionic species accounting for the electric current remained only an assumption until self-referencing technology was employed and a direct correlation to potassium ion secretion was demonstrated (42).

Relative current densities over the apical surface of epithelial cells were measured with the use of the vibrating voltage probe, and a large current (40–60 μA/cm2) was detected emanating from vestibular dark cells (42, 65). Inhibition of the Na+-K+-Cl cotransporter with bumetanide or the Na+-K+-ATPase with ouabain reduced the current, whereas increased basolateral potassium concentrations enhanced the current. These data suggested that potassium transport into the basolateral surface of vestibular dark cells was critical in maintaining the apically directed current and that perhaps potassium ions carried the apically directed current as they become concentrated in the inner ear lumen (42). Employment of potassium-selective Seris probes (K+ probes) directly demonstrated that vestibular dark cells were indeed secreting potassium into the inner ear lumen (42). The activity of potassium 20–240 μm above that of the epithelium was shown to be elevated over that of the bulk medium, and positioning the Seris probe in close proximity of vestibular dark cells revealed these cells to be the source of the potassium efflux. Both bumetanide and ouabain caused a decrease in the potassium activity near vestibular dark cells (Fig.5) and established a model by which potassium is transported into vestibular dark cells by the Na+-K+-Cl cotransporter and Na+-K+-ATPase and subsequently secreted into the inner ear lumen, where it becomes concentrated.

Fig. 5.

Seris K+-probe measurement of potassium activity in the proximity of the gerbil vestibular dark cell epithelium. A: K+ activity in the proximity of the apical surface is decreased by basolateral bumetanide (Bumet), a blocker of the Na+-K+-Clcotransporter, and basolateral ouabain, a blocker of the Na+-K+-ATPase. Basolateral exposure of the epithelium to hyposmotic solutions enhances K+ activity in the proximity of the apical surface. B: schematic of ion transport by a vestibular dark cell and the sites of action of pharmacological agents. PLC, phospholipase C; PKC, protein kinase C;I Ks, slowly activating potassium channel; DIDS, disulfonic stilbene. C: apical DIDS, anI Ks channel opener, evoked a decline in K+ activity in the proximity of the basolateral surface, indicating an enhanced K+ influx. Apical ATP, an activator of the PLC/PKC pathway that inhibits I Kschannels, evoked an increase in basolateral K+ activity indicative of decreased K+ efflux. Data are presented as percentages of control K+ activity after normalization of control to ±100%.

Three classes of potassium channels have been described in the apical surface of vestibular dark cells, and each can contribute to the potassium efflux from vestibular dark cells into the inner ear lumen. Apical potassium channels include the slowly activating channel (I Ks) (41), a nonselective cation channel (44), and a maxi-K+ channel (63). Modulation of the transepithelial voltage, resistance, current, and potassium efflux by disulfonic stilbenes (Fig.5) suggested that the primary mechanism by which potassium moves across the apical surface of vestibular dark cells is throughI Ks channels (41, 57). Subsequent studies demonstrated that phospholipase C and protein kinase C are capable of modulating I Ks channels and potassium efflux through the apical neuroepithelium (43). Interestingly, potassium efflux through the apical membrane also participates in cell volume regulation (66). Hyposmotic challenge of inner ear neuroepithelium resulted in an increase in the volume of the epithelial cells, followed by a compensatory volume decrease that returned cells to their original volume (67). The Seris K+ probe measured an elevation in potassium activity in the proximity of vestibular dark cells during compensatory volume decreases and established that the primary mechanism underlying this decrease is a change in osmotic pressure induced by potassium efflux through I Ks channels in the apical surface (66).


The studies on the inner ear neuroepithelium illustrate the usefulness of the Seris technique for understanding the physiology of cellular layers by measuring ionic activities in the proximity of cell aggregates. The role of potassium in volume regulation of individual cells undergoing cell death has also been investigated with the use of Seris K+ probes (64). One characteristic of apoptosis, a particular class of cell death, is pronounced cell shrinkage (6, 28). Indeed, single-cell mouse embryos treated with agents that induce apoptosis undergo rapid cell shrinkage (2% decrease in volume per minute) (27, 64). With the use of the Seris K+ probe, it was determined that, coincident with shrinkage, the activity of potassium in the proximity of embryos became elevated, and pharmacological dissection suggested that potassium was fluxing through tetraethylammonium-sensitive potassium channels (Fig. 6) (64).

Fig. 6.

Seris K+-probe measurement of K+activity ([K+]) in the proximity of an individual mouse embryo undergoing cell death. Exposure of mouse embryos to oxidative stress (H2O2) evoked an increase in K+ activity near embryos, indicating an enhanced K+ efflux. Pretreatment (15 min) with the broad spectrum K+ channel blocker tetraethylammonium (TEA) decreased the K+ activity evoked by H2O2, suggesting that K+ efflux during oxidation-induced cell death occurs through K+ channels in the plasma membrane.Inset: image of a mouse embryo surrounded by its extracellular covering, the zona pellucida, with an adjacent K+-selective probe.

Movement of potassium across cell plasma membranes is critical to a wide variety of cellular processes, and the Seris technology is uniquely suited to noninvasively investigate the physiology of small tissues and individual cells by monitoring changes in the ionic composition of the media in their proximity. Next, we describe how single cells need not be relatively large, as are mouse embryos, but can be as small as 10 μm in diameter.


The microglia are an interesting group of neural cells with features that set them apart from the other glia and neurons. Central to this is the developmental origin of microglia. Although still a matter of ongoing debate, the body of evidence now favors an ontogenetic relationship of microglia with mononuclear phagocytes (36, 21), a distinctly unique lineage for neural tissues. Microglia respond in specific manners to several pathological conditions and injury (32). Activated microglia can undergo oxidative bursts, facilitated oxygen radical production, and phagocytosis of pathogens and cellular debris. Microglia also interact with other cells of the brain, notably astrocytes, where inflammation initiates astrocytic release of cytokines such as granulocyte/macrophage colony-stimulating factor and macrophage colony stimulating factor, which in turn activate microglia. Potassium appears to be a key player in several microglial responses such that cytokine-induced proliferation and differentiation in microglia, for example, involve activation of an inwardly rectified potassium channel (Kir) (54, 58). Furthermore, injury to brain tissue can radically alter the microenvironment, raising external potassium concentrations and influencing transport mechanisms (14, 29). The potential significance to the microglia of potassium activity led to the study of the cellular mechanism behind potassium transport (59).

Initial application of a potassium-selective Seris probe to isolated microglia from rat brain showed a clear K+ influx, registered as a lower activity in proximity to the plasma membrane. Close to the membrane, a drop of −9.43 ± 4.2 μM in the external potassium activity (Δ[K+]o) was recorded, but the values within the sample showed an apparent bimodal distribution with peaks at −6 and −15 μM. This Δ[K+]o is referred to by the authors as a differential diffusion potential. Clearly, one possible source of the change in the measured activity of potassium is influx through the Kir channel. The presence of this channel was confirmed in these isolated cells, but blocking the activity with 1 mM Cs+ and 2 mM Ba2+ did not affect the [K+]o gradient.

Kir presence was also examined by activating this channel with voltage clamp and comparing the elicited currents with changes in potassium ion gradients in the proximity of microglia recorded with the use of the Seris technique. The clamp protocol included a two-step hyperpolarization to −100 and −110 mV from a holding potential of −70 mV. Under these conditions the current density per driving force was 76 × 10−4pA · μm−2 · mV−1. The voltage-sensitive inward potassium current generated a small but measurable [K+]o depletion gradient. On the basis of this Seris-derived [K+]o gradient, a potassium flux can be calculated (61, 62) and converted into the equivalent current. This can then be compared directly with the concurrently measured clamp values. The Seris probe measured 0.61–0.77 of the current recorded through the whole cell patch. This is in good agreement with the expected value, but there is still a clear underestimation that may be attributable to the position of the probe in relation to the membrane. A position-dependent response is expected if a diffusion potential is being measured (61) and has been demonstrated for the microglial data (59). It should also be noted that the Seris probe is measuring a local flux with a sensor of tip diameter between 2 and 4 μm. It is assumed that the whole cell current is evenly distributed across the cell surface, but this need not be the case. Clustering of channels may offer an alternative explanation to the mismatch between the Seris probe estimates of potassium ion influx and the monitoring with voltage-clamp techniques. What is clear from these results is that noninvasive measurement of the flux with a Seris probe closely follows the expected value from voltage clamp. Coupling of these techniques in the future can be expected to continue validation of the Seris measurements, not only for potassium ions but for other ions where voltage-dependent currents can be isolated.

The experiments described illustrate two important points. First, the Seris probes can measure the expected whole cell current with an acceptable level of agreement. Second, it is clear that Kirdoes not contribute significantly to the standing potassium influx measured in the cultured microglia. Because Kir does not contribute to the potassium flux measured in 64% of the cells examined, the question arises as to what does. The Na+-K+-ATPase is an obvious candidate, but in blocking experiments with both ouabain and strophanthidin, the latter targeted ouabain-insensitive P-type ATPases (68, 69) and the potassium influx gradient was undiminished. This result was surprising given the neural environment of microglia, but in considering their developmental origin, it seemed possible that these cells were using a mechanism for potassium transport common to other cell lineages. The H+-K+-ATPase was shown to be responsible for the K+-influx gradient because it was rapidly shut down by SCH-28080 and omprazole, compounds selective for that pump (Fig. 7). A counter efflux of hydrogen ions was also demonstrated with the use of a Seris H+ probe and was inhibited by SCH-28080. Furthermore, an antibody raised against the gastric form of the H+-K+-ATPase bound to microglia in culture (59).

Fig. 7.

Effect of an H+-K+-ATPase pump blocker on the K+ activity in proximity to an isolated rat microglial cell (59). Changes in K+ activity were monitored with a Seris probe before and after the application of the blocker SCH-28080 at concentrations of 1 and 5 μM. The compound was washed out between the trials, and the signal was recovered. A continuous recording is shown as the Seris probe was moved from background positions far from the cell (shaded bar) to within the K+ diffusion layers surrounding the cell (solid bar). Because data are acquired through a running average (62), new values take some time to be established. In this case, the buffer is ∼20 s. With the use of known background K+ activities, changes in K+ activity immediately outside the cell (Δ[K+]o) were calculated from the measured differential diffusion potential between two positions 10 μm apart.

The physiological experiments performed with the Seris probe demonstrated that there is an ATPase, normally associated with the intestinal system, present in the microglia of the central nervous system. However, in some parameters the neural H+-K+- ATPase differs from the gastric variety despite sharing the common pharmacology described. Examining the regulation of the diffusion potential by this pump to changes in potassium concentration shows a maximal pumping activity at a [K+]o of 7 mM with a Michaelis-Menten constant of 3.67 mM. The transporter dependence on extracellular pH did not exhibit the expected Michaelis-Menten passive ion availability behavior. Saturation of the transporter by hydrogen ions was not achieved at pH values as low as 6.63.


The study of the microglia revealed for the first time the involvement of an H+-K+-ATPase in these important reactive cells of the brain. The pump may be a new member of the Na+-K+-ATPase superfamily. As with the other studies described, the Seris technique was used to approach a problem in cell physiology more effectively than could have been achieved with other techniques. All studies produced unique data of considerable biomedical importance. Each of these studies used sensors available commercially in the form of liquid membranes incorporating specific ionophores, and, therefore, the characteristics of these ionophores limit the molecules that can be detected. However, the principle of the self-referencing technique has the potential for further applications by expanding the types of molecules detected. For example, a self-referencing electrochemical probe, in this case developed to measure local oxygen gradients, has been successfully used with a single neuron and a plant cell (34). Subsequently, the same methodology has been applied to single pancreatic β-cells (52) and diversified to include the detection of nitric oxide and ascorbate from single cells (Pepperell J, Porterfield DM, and Smith PJS, unpublished observations). Currently under development is the incorporation of biosensors (26, 49) onto the reactive surface of a self-referencing electrochemical probe. Preliminary results have demonstrated the feasibility of this approach for glucose detection (50), and we look forward to the rapid diversification of self-referencing technologies to other enzyme reactions with redox products.

In conclusion, the Seris technique has shown great versatility, noted here by the physiologically diverse types of cells studied and molecular components revealed. From a complex and heterogeneous cellular pavement to individual cells, nonelectrogenic events, as produced by the “gastric” H+-K+-ATPase, can be investigated. With the continued incorporation of new sensors, as well as the combination of the self-referencing approach with other methodologies, we can look forward to exciting observations in cellular transport physiology.


The authors are supported by National Center for Research Resources (NCRR) Grant P41 RR-101395 (P. J. S. Smith), National Institutes of Health Grant KO-81099, and the Lalor Foundation (J. Trimarchi). The BioCurrents Research Center is a resource of the NCRR, which specializes in the design, development, and application of methods for the study of cell transport phenomena. As such, we encourage biomedical researchers who want to take advantage of the technology available to contact P. J. S. Smith. Information on the resource can be found at


  • Address for reprint requests and other correspondence: P. J. S. Smith, BioCurrents Research Center, 7 MBL St., Woods Hole, MA 02543 (E-mail: psmith{at}


View Abstract