Vibrating probe analysis of teleost opercular epithelium: correlation between active transport and leak pathways of individual chloride cells

Paracellular pathway remodeling enhances sodium secretion by teleost fish in hypersaline environments

In vertebrate salt-secreting epithelia, Na(+) moves passively down an electrochemical gradient via a paracellular pathway. We assessed how this pathway is modified to allow Na(+) secretion in hypersaline environments. Mummichogs (Fundulus heteroclitus) acclimated to hypersaline [2× seawater (2SW), 64‰] for 30 days developed invasive projections of accessory cells with an increased area of tight junctions, detected by punctate distribution of CFTR (cystic fibrosis transmembrane conductance regulator) immunofluorescence and transmission electron miscroscopy of the opercular epithelia, which form a gill-like tissue rich in ionocytes. Distribution of CFTR was not explained by membrane raft organization, because chlorpromazine (50 μmol l(-1)) and filipin (1.5 μmol l(-1)) did not affect opercular epithelia electrophysiology. Isolated opercular epithelia bathed in SW on the mucosal side had a transepithelial potential (Vt) of +40.1±0.9 mV (N=24), sufficient for passive Na(+) secretion (Nernst equilibrium voltage≡ENa=+24.11 mV). Opercular epithelia from fish acclimated to 2SW and bathed in 2SW had higher Vt of +45.1±1.2 mV (N=24), sufficient for passive Na(+) secretion (ENa=+40.74 mV), but with diminished net driving force. Bumetanide block of Cl(-) secretion reduced Vt by 45% and 29% in SW and 2SW, respectively, a decrease in the driving force for Na(+) extrusion. Estimates of shunt conductance from epithelial conductance (Gt) versus short-circuit current (Isc) plots (extrapolation to zero Isc) suggested a reduction in total epithelial shunt conductance in 2SW-acclimated fish. In contrast, the morphological elaboration of tight junctions, leading to an increase in accessory-cell-ionocyte contact points, suggests an increase in local paracellular conductance, compensating for the diminished net driving force for Na(+) and allowing salt secretion, even in extreme salinities.

Tight junction pore and leak pathways: a dynamic duo

Tissue barriers that restrict passage of liquids, ions, and larger solutes are essential for the development of multicellular organisms. In simple organisms this allows distinct cell types to interface with the external environment. In more complex species, the diversity of cell types capable of forming barriers increases dramatically. Although the plasma membranes of these barrier-forming cells prevent flux of most hydrophilic solutes, the paracellular, or shunt, pathway between cells must also be sealed. This function is accomplished in vertebrates by the zonula occludens, or tight junction. The tight junction barrier is not absolute but is selectively permeable and is able to discriminate between solutes on the basis of size and charge. Many tight junction components have been identified over the past 20 years, and recent progress has provided new insights into the proteins and interactions that regulate structure and function. This review presents these data in a historical context and proposes an integrated model in which dynamic regulation of tight junction protein interactions determines barrier function.

Tight junction pore and leak pathways: a dynamic duo

Tissue barriers that restrict passage of liquids, ions, and larger solutes are essential for the development of multicellular organisms. In simple organisms this allows distinct cell types to interface with the external environment. In more complex species, the diversity of cell types capable of forming barriers increases dramatically. Although the plasma membranes of these barrier-forming cells prevent flux of most hydrophilic solutes, the paracellular, or shunt, pathway between cells must also be sealed. This function is accomplished in vertebrates by the zonula occludens, or tight junction. The tight junction barrier is not absolute but is selectively permeable and is able to discriminate between solutes on the basis of size and charge. Many tight junction components have been identified over the past 20 years, and recent progress has provided new insights into the proteins and interactions that regulate structure and function. This review presents these data in a historical context and proposes an integrated model in which dynamic regulation of tight junction protein interactions determines barrier function.

Mechanosensitive signalling in fish gill and other ion transporting epithelia

Epithelia involved in vectorial salt transport respond to apical and basolateral changes in osmotic activity by moderating the transmural solute transport rate simultaneously with underlying volume regulatory mechanisms involved in regulatory volume increase (RVI) and decrease (RVD). This review examines rapid osmotic responses in salt secreting epithelia of marine and euryhaline teleost fish, with inclusion of recent results from other ion transporting epithelia that also respond rapidly to osmotic shock. Mitochondrion-rich chloride secreting cells of marine teleost fish gills and skin, when exposed to hypertonic shock, activate NaCl secretion via phosphorylation of Na(+), K(+), 2Cl(-) cotransporter (NKCC1) in the basolateral membrane and activation of anion channels in the apical membrane. Conversely, NaCl secretion is inhibited when chloride secreting cells are swollen osmotically. Mammalian airway epithelial cells also possess NKCC1 basally and apical anion channels [Cystic Fibrosis Transmembrane conductance Regulator (CFTR)]; with hypotonic shock, this epithelium releases ATP and NaCl secretion is stimulated via purinergic receptors, while hypertonic shock inhibits Na(+) uptake. In the eye, the ciliary epithelium activates Cl(-) channels in response to hypotonic shock as RVD, an effect that modulates transepithelial fluid transport rates. In the renal A6 cell line, K(+) and Cl(-) effluxes activate during RVD and RVI Na(+) transepithelial absorption. A common theme in these systems is ATP release in hypotonic shock with subsequent RVD-effective mechanisms such as NKCC1 inhibition and K(+) and Cl(-) efflux, but there are different effects of osmotic changes on transepithelial transport, apparently depending on the role of the epithelial system.

Fish louseArgulus funduli(Crustacea: Branchiura) ectoparasites of the euryhaline teleost host,Fundulus heteroclitus, damage the ion-transport capacity of the opercular epithelium

Killifish ( Fundulus heteroclitus (L., 1766)) collected in the wild and kept in full-strength seawater were naturally parasitized by the ectoparasite Argulus funduli Krøyer, 1863, a copepod fish louse that creates inflamed skin lesions on the opercular epithelium and host gills. We assessed the damage done by lesions by counting the density of mitochondria-rich cells by fluorescence microscopy and by measuring Cl–secretion rate electrophysiologically using control (no lesions) and affected isolated opercular epithelia, often as paired left and right membranes from a single fish. Epithelia with lesions had a significantly reduced Cl–secretion rate, and in the lesions, the density of chloride cells was near zero. Contralateral membranes without lesions from infested fish had transport rates not significantly different from membranes taken from uninfected control animals, indicating no overcompensation on the contralateral membranes. Healthy control and infested animals were transferred to hypersaline conditions (twice seawater). Infested and control animals all survived transfer and had elevated plasma Na+and plasma osmolality. Infested animals failed to significantly elevate Imto the same level as healthy animals and there was a difference in hematocrit. Happily, the hypersaline challenge also resulted in detachment and death of adult A. funduli. We conclude that Argulus lesions impair salt transport in affected membranes but do not significantly affect survival on hypersaline challenge, and that hypersaline exposure is a successful treatment for A. funduli infestation in these strongly euryhaline teleosts.

The Multifunctional Fish Gill: Dominant Site of Gas Exchange, Osmoregulation, Acid-Base Regulation, and Excretion of Nitrogenous Waste

The fish gill is a multipurpose organ that, in addition to providing for aquatic gas exchange, plays dominant roles in osmotic and ionic regulation, acid-base regulation, and excretion of nitrogenous wastes. Thus, despite the fact that all fish groups have functional kidneys, the gill epithelium is the site of many processes that are mediated by renal epithelia in terrestrial vertebrates. Indeed, many of the pathways that mediate these processes in mammalian renal epithelial are expressed in the gill, and many of the extrinsic and intrinsic modulators of these processes are also found in fish endocrine tissues and the gill itself. The basic patterns of gill physiology were outlined over a half century ago, but modern immunological and molecular techniques are bringing new insights into this complicated system. Nevertheless, substantial questions about the evolution of these mechanisms and control remain.

Cellular composition and ultrastructure of the gill epithelium of larval and adult lampreys

Lampreys, one of the only two surviving groups of agnathan (jawless)vertebrates, contain several anadromous species that, during their life cycle,thus migrate from fresh to seawater and back to freshwater. Lampreys have independently evolved the same overall osmoregulatory mechanisms as the gnathostomatous (jawed) and distantly related teleost fishes. Lamprey gills thus likewise play a central role in taking up and secreting monovalent ions. However, the ultrastructural characteristics and distribution of their epithelial cell types [ammocoete mitochondria-rich (MR) cell, intercalated MR cell, chloride cell and pavement cell] differ in several respects from those of teleosts. The ultrastructural characteristics of these cells are distinctive and closely resemble those of certain ion-transporting epithelia in other vertebrates, for which the function has been determined. The data on each cell type, together with the stage in the life cycle at which it is found, i.e. whether in fresh or seawater, enable the following proposals to be made regarding the ways in which lampreys use their gill epithelial cells for osmoregulating in hypo- and hypertonic environments. In freshwater, the intercalated MR cell takes up Cl– and secretes H+,thereby facilitating the uptake of Na+ through pavement cells. In seawater, the chloride cell uses a secondarily active transcellular transport of Cl– to provide the driving force for the passive movement of Na+ through leaky paracellular pathways between these cells.

Rapid regulation of NaCl secretion by estuarine teleost fish: coping strategies for short-duration freshwater exposures

This review summarizes the mechanism of Cl(-) active secretion and its regulation in estuarine teleost fish. Small estuarine fish such as the killifish, Fundulus heteroclitus, forage in shallow water following advancing tides and are exposed regularly to very dilute microenvironments. Using the killifish opercular epithelium and related teleost membranes containing mitochondria-rich cells, the regulation includes a reduction of active Cl(-) secretion and passive diffusive ion loss in a three-stage process spanning approximately 30 min. There is a combination of sympathetic neural reflex mediated by alpha(2)-adrenoceptors operating via intracellular inositol tris phosphate and intracellular Ca(2+) and a cellular hypotonic shock response, followed by covering over of ion-secreting cells by pavement cells. This effectively minimizes salt loss in dilute media. The upregulation of salt secretion on return to full strength seawater may be via hormones (arginine vasotocin and urotensin I) and neurotransmitter (vasoactive intestinal polypeptide) in combination with hypertonic shock. A hypothetical model includes involvement of protein kinase A and C and protein phosphatases 1 and 2A in regulation of the NKCC1 cotransporter on the basolateral side and protein kinase A regulation of the CFTR-like apical anion channel.

The route of passive chloride movement across amphibian skin: localization and regulatory mechanisms

Transepithelial Cl(-) conductance (G(Cl)) in amphibian skin can be activated in several species by serosa positive potentials. Mitochondria-rich cells (MRC) or tight junctions (TJ) between the epithelial cells are possible sites for this pathway. The properties and the techniques used to investigate this pathway are reviewed in the present paper. In situ techniques are preferable, since specific properties of the MRC are apparently not maintained in isolated cells. Volume measurements and electronprobe microanalysis of intracellular ions suggest the localization of voltage-activated G(Cl) to MRC. G(Cl) correlates poorly with the density of MRC. The vibrating voltage probe allows quantitative correlation of the local Cl(-) current through morphologically identified structures and the transepithelial Cl(-) current. Our analysis shows that 80% of the voltage-activated Cl(-) current is accounted for by current through MRC or their immediate vicinity. The activation patterns of this current and the inhibition by the alpha(1)-adrenergic agonist, epinephrine, conform to those of the transepithelial current. However, less than 20% of the MRC are active at a certain moment and the activity is spontaneously variable with time. The molecular nature of this pathway, physiological control mechanisms and their relation to the temporal activity of MRC remain to be studied.

Natriuretic peptides in fish physiology

Natriuretic peptides exist in the fishes as a family of structurally-related isohormones including atrial natriuretic peptide (ANP), C-type natriuretic peptide (CNP) and ventricular natriuretic peptide (VNP); to date, brain natriuretic peptide (or B-type natriuretic peptide, BNP) has not been definitively identified in the fishes. Based on nucleotide and amino acid sequence similarity, the natriuretic peptide family of isohormones may have evolved from a neuromodulatory, CNP-like brain peptide. The primary sites of synthesis for the circulating hormones are the heart and brain; additional extracardiac and extracranial sites, including the intestine, synthesize and release natriuretic peptides locally for paracrine regulation of various physiological functions. Membrane-bound, guanylyl cyclase-coupled natriuretic peptide receptors (A- and B-types) are generally implicated in mediating natriuretic peptide effects via the production of cyclic GMP as the intracellular messenger. C- and D-type natriuretic peptide receptors lacking the guanylyl cyclase domain may influence target cell function through G(i) protein-coupled inhibition of membrane adenylyl cyclase activity, and they likely also act as clearance receptors for circulating hormone. In the few systems examined using homologous or piscine reagents, differential receptor binding and tissue responsiveness to specific natriuretic peptide isohormones is demonstrated. Similar to their acute physiological effects in mammals, natriuretic peptides are vasorelaxant in all fishes examined. In contrast to mammals, where natriuretic peptides act through natriuresis and diuresis to bring about long-term reductions in blood volume and blood pressure, in fishes the primary action appears to be the extrusion of excess salt at the gills and rectal gland, and the limiting of drinking-coupled salt uptake by the alimentary system. In teleosts, both hypernatremia and hypervolemia are effective stimuli for cardiac secretion of natriuretic peptides; in the elasmobranchs, hypervolemia is the predominant physiological stimulus for secretion. Natriuretic peptides may be seawater-adapting hormones with appropriate target organs including the gills, rectal gland, kidney, and intestine, with each regulated via, predominantly, either A- or B-type (or C- or D-type?) natriuretic peptide receptors. Natriuretic peptides act both directly on ion-transporting cells of osmoregulatory tissues, and indirectly through increased vascular flow to osmoregulatory tissues, through inhibition of drinking, and through effects on other endocrine systems.

Apical structures of "mitochondria-rich" alpha and beta cells in euryhaline fish gill: their behaviour in various living conditions

BACKGROUND

One of the characteristic features of the two types (alpha and beta) of "mitochondria-rich" (chloride) cells in the gill epithelium of freshwater fishes is the presence in their apical region of tubulovesicular structures. A further analysis of the ultrastructural features of these apical elements as well as that of their modifications under various living conditions should help to understand better the respective rôle of both alpha and beta cells in these conditions.

METHODS

Atlantic salmon (Salmo salar) maintained in fresh water as well as tilapia (Oreochromis niloticus) maintained either in fresh water or in deionized water or in 20% saltwater were examined. Measurements of surface areas of apical structures in the various living conditions were also performed.

RESULTS

In the alpha cells of freshwater fishes, the apical structures consisted of isolated vesicles containing a filamentous material resembling that coating the apical surface. They were closely related to the apical plasma membrane and did not penetrate the region containing the tubular system. When fishes were transferred to deionized water, the number of the apical membrane folds increased significantly, as did the number and size of apical structures which became elongated. In saltwater-adapted fishes, the apical structures showed a tendency to collapse and took the appearance of flattened and slightly curved elements. These observations tended to indicate that in alpha cells the apical structures were extensions of the apical plasma membrane and thereby might be implicated in sodium uptake when fishes are placed in fresh or deionized water and in chloride excretion when they are transferred to salt water. In beta cells, the apical structures were usually separated from the apical plasma membrane by a zone rich in cytoskeleton elements. They penetrated deeply into the supranuclear region, where they intermingled with the elements of the tubular system. They consisted mainly of tubular elements that contained a material resembling that present in the trans tubular Golgi network from which they might originate. The apical structures remained unaltered in beta cells whatever the medium (fresh or deionized water) in which the fish was placed.

CONCLUSIONS

The alpha cells which are usually thought to be mainly involved in chloride excretion when fishes are transferred into seawater might also be implicated in sodium uptake in freshwater living conditions. The rôle of beta cells, in contrast, still remains to be established.

Endogenous ionic currents and DC electric fields in multicellular animal tissues

Through the use of the non-invasive vibrating probe technique for detecting extracellular ionic currents developed in 1974 [Jaffe and Nuccitelli: J Cell Biol 63:614-628, 1974], embryonic currents have been detected in a wide range of animal systems (recently reviewed in [Nuccitelli, Noninvasive Techniques in Cell Biology. New York: Wiley-Liss, 1990, pp 273-310]. In four of these studies, the corresponding electric field has been measured within the animal tissue. Such measurements of internal electric fields are quite challenging because they involve the insertion of microelectrodes into the developing tissue along specific regions of current flow. This paper reviews the evidence for endogenous transembryonic currents and dc electric fields in animal systems and provides the range of values for such physiological fields. These data should provide a guide to the range of imposed electric field strengths that could influence normal biological functions in living organisms.

Sites of iontophoretic current flow into the skin: identification and characterization with the vibrating probe electrode

The routes taken by charged substances (e.g., peptides) through the skin during iontophoretic drug delivery are not well characterized. We have used a vibrating probe electrode to reproducibly identify and vectorize site-specific (spatial resolution = 20 microns) ionic flows as they were occurring in hairless mouse skin clamped at clinically relevant current densities. These iontophoretic currents were primarily appendageal, and certain appendages (e.g., small hairs) appeared to carry most of the current. This finding may have important ramifications with respect to irritation, allergic reaction, and electrical current damage in iontophoretic drug delivery. The size and direction of the current vectors could change under certain conditions (e.g., in an unbuffered preparation, where pH changes occurred during the experiment). The vibrating probe can operate in (and is not adversely affected by) the ranges of pH, tonicity, and current required for the study of iontophoretic currents.

Localization of sodium absorption and chloride secretion in an intestinal epithelium

Hen coprodeum absorbs sodium electrogenically and, when stimulated by theophylline, secretes chloride. In this study the vibrating microprobe technique was used to localize the transport of these ions to intestinal villi/folds and crypts. With the isolated, stretched epithelium, controlled by light microscopy and scanning electron microscopy, in open circuit, currents were inward, 40 +/- 7 microA/cm2, 50 microns vertically above villi, and outward, 36 +/- 7 microA/cm2 above crypts. The currents decayed exponentially to near zero at 300 microns with the same length constant. A physical model simulating the observed loci of current sources and sinks predicts potential profiles consistent with our data. Extrapolation of the currents gives a surface potential of 45 microV, negative on villi and positive above crypts. Short circuiting increased villus current to 86 +/- 27 microA/cm2 at 50 microns, and amiloride treatment reduced it to -8 microA/cm2; in both cases crypt currents were abolished. The inward currents are compatible with sodium absorption. Induction of chloride secretion after amiloride treatment, resulted in current circuits similar to those induced by sodium absorption, with villus currents of 23 +/- 7 microA/cm2. This is in accord with chloride secretion at the villi. Quantitative estimates of crypt number (860/cm2) and opening diameter (15 microns), in conjunction with isotopic measurements of active and electrical potential-driven ion fluxes demonstrate, however, that only 4% of the potential-driven co-ion transport occurs through the crypts. This indicates that nearly all chloride secretion comes from the sodium-absorbing villar area. Were the chloride secretion to occur solely from the crypts, the current should have been in the opposite direction and 10,000-fold larger.

Intercellular junctions in the gill epithelium of the Atlantic hagfish, Myxine glutinosa

The intramembrane organization of the occluding junctions in the gill epithelium of the Atlantic hagfish, Myxine glutinosa, was studied by means of freeze-fracture electron microscopy. Mitochondria-rich cells, characterized by assemblies of rod-shaped particles in the luminal plasma membrane and by an extensive intracellular amplification of the basolateral plasma membrane, are singly distributed between the pavement cells in the gill epithelium of this marine and stenohaline cyclostome. The occluding junctions between mitochondria-rich cells and pavement cells do not differ from those between adjacent pavement cells, concerning the number of superimposed strands (median 6, range 4-9) and their geometrical organization. These observations suggest that, in contrast to marine teleosts, the paracellular pathway plays a minor role in transepithelial ion movements in the hagfish gill epithelium. The findings are in agreement with the absence of hypoosmoregulatory mechanisms in hagfish, as have been evolved in various marine vertebrates. In addition, small communicating junctions are demonstrated between pavement cells; they possibly serve for a coordinated synthesis and secretion of mucus by the pavement cells.

Ionic currents in morphogenesis

Morphogenetic fields must be generated by mechanisms based on known physical forces which include gravitational forces, mechanical forces, electrical forces, or some combination of these. While it is unrealistic to expect a single force, such as a voltage gradient, to be the sole cause of a morphogenetic event, spatial and temporal information about the electrical fields and ion concentration gradients in and around a cell or embryo undergoing morphogenesis can take us one step further toward understanding the entire morphogenetic mechanism. This is especially true because one of the handful of identified morphogens is Ca2+, an ion that will not only generate a current as it moves, but which is known to directly influence the plasma membrane's permeability to other ions, leading to other transcellular currents. It would be expected that movements of this morphogen across the plasma membrane might generate ionic currents and gradients of both electrical potential and intracellular concentration. Such ionic currents have been found to be integral components of the morphogenetic mechanism in some cases and only secondary components in other cases. My goal in this review is to discuss examples of both of these levels of involvement that have resulted from investigations conducted during the past several years, and to point to areas that are ripe for future investigation. This will include the history and theory of ionic current measurements, and a discussion of examples in both plant and animal systems in which ionic currents and intracellular concentration gradients are integral components of morphogenesis as well as cases in which they play only a secondary role. By far the strongest cases for a direct role of ionic currents in morphogenesis is the polarizing fucoid egg where the current is carried in part by Ca2+ and generates an intracellular concentration gradient of this ion that orients the outgrowth, and the insect follicle in which an intracellular voltage gradient is responsible for the polarized transport from nurse cell to oocyte. However, in most of the systems studied, the experiments to determine if the observed ionic currents are directly involved in the morphogenetic mechanism are yet to be done. Our experience with the fucoid egg and the fungal hypha of Achlya suggest that it is the change in the intracellular ion concentration resulting from the ionic current that is critical for morphogenesis.

Intracellular voltage recordings in the opercular epithelium of Fundus heteroclitus

Opercular epithelial cells of Fundus heteroclitus were investigated using conventional microelectrodes. The area of interest was the cells lining the inside of the opercular epithelium closest to the gill arches, an area with a high density of chloride cells. Only one cell type could be discerned from the values of 60 opercular cells measured with the opercular epithelium in open circuit conditions. A mean apical voltage of -18.0 +/- 0.6 mV was observed with intracellular values ranging from -10 to -30 mV. The predicted intracellular chloride content was 59 mM/liter. Apical fractional resistance (faR) was 0.78 +/- 0.02. The intracellular potential measurements were typically difficult to maintain for extended periods (longer than 3 min). The opercular cells depolarize with serosal isoproterenol treatment (10(-6) M) corresponding to the increase in opercular transepithelial potential. The opercular cell apical fR decreased with isoproterenol treatment. These data indicate the observed opercular cells were involved in opercular chloride transport.

Segregation of gastric Na and Cl transport: a vibrating probe and microelectrode study

The short-circuit current (Isc) of resting Necturus gastric mucosa (approximately 20 microA/cm2) can be attributed to the algebraic sum of the net Cl- secretion and amiloride-inhibitable net Na+ absorption. We have attempted to identify the cell types [surface epithelial cells (SCs) or oxyntic cells (OCs)] responsible for the transport of these ions in Necturus gastric mucosa using microelectrodes (ME) and a vibrating probe (VP). Mucosae were mounted horizontally in an open-topped Plexiglas chamber either serosal side up for basolateral ME impalements of OCs or mucosal side up for apical impalements of SCs and VP measurements. Cell impalements were made under open-circuit conditions, and VP measurements were performed under short-circuit conditions. Impalements of OCs indicate that neither the ratio of their apical to basolateral cell membrane resistances (Ra/Rb = 1.3 +/- 0.2) nor their cell membrane potentials were affected by 10(-6) M mucosal amiloride. In contrast, impalements of SCs indicate that amiloride increased their Ra/Rb from 3.5 +/- 0.2 to 15.6 +/- 1.8 and hyperpolarized both cell membrane potentials by greater than 20 mV. VP measurements showed that the amiloride-induced change in the current from SCs (5.6 microA/cm2) accounted for the amiloride-induced change in the Isc (5.5 microA/cm2). A non-zero current (4.4 +/- 1.0 microA/cm2) measured over SCs in the presence of amiloride was due to contamination from current arising from the gastric crypts that contain the OCs.(ABSTRACT TRUNCATED AT 250 WORDS)

Localization of chloride conductance to mitochondria-rich cells in frog skin epithelium

Cell volume determinations and electrophysiological measurements have been made in an attempt to determine if mitochondria-rich (MR) cells are localized pathways for conductive movements of Cl across frog skin epithelium. Determinations of cell volume with video microscope techniques during transepithelial passage of current showed that most MR cells swell when the tissue is voltage clamped to serosa-positive voltages. Voltage-induced cell swelling was eliminated when Cl was removed from the mucosal bath solution. Using a modified vibrating probe technique, it was possible to electrically localize a conductance specifically to some MR cells in some tissues. These data are evidence supporting the idea that MR cells are pathways for conductive movements of Cl through frog skin epithelium.