Action of palytoxin on apical H + /K + -ATPase in rat colon Georgios Scheiner-Bobis 1 , Thomas Hu¨ bschle 2 and Martin Diener 2 1 Institute for Biochemistry and Endocrinology, 2 Institute for Veterinary Physiology, Justus-Liebig-University Giessen, Germany Palytoxin stimulated a cation-dependent short-circuit cur- rent (Isc) in rat distal and proximal colon in a concen- tration-dependent fashion when applied to the mucosal surface of the tissue. The distal colon exhibited a higher sensitivity to the toxin. The palytoxin-induced Isc was blocked by vanadate but was resistant to ouabain or scilliroside, suggesting the conversion of a vanadate-sen- sitive H + /K + -ATPase into an electrogenic cation trans- porter. Cation substitution experiments with basolaterally depolarized tissues suggested an apparent permeability of the palytoxin-induced conductance of Na + >K + >Li + . Immunohistochemical control experiments confirmed the absence of the Na + /K + -ATPase in the apical membrane. Consequently, the pore-forming action of palytoxin is not restricted to Na + /K + -ATPase but is also observed with the colonic H + /K + -ATPase. Keywords: ATPase; colon; palytoxin; Isc; ion channel. P 2C -type ATPases are oligomeric enzymes consisting of a and b subunits [1]. The sodium pump from the plasma membranes of animal cells, a member of this group, generates a sodium gradient by pumping three Na + ions out of the cell and two K + ions into the cell for each ATP hydrolyzed [2]. This sodium gradient is the driving force of all secondarily active transporters and a presupposition for neuronal conduction of signals. The closest relatives of the sodium pump are the proton pumps from gastric and colon epithelial cells [3]. Although these pumps are not identical, they both catalyze an active secretion of protons driven by ATP hydrolysis. Unlike the sodium pump, however, both proton pumps are electro- neutral: each transports one K + ion from the luminal side into the cytosol for each H + secreted. Several naturally occurring toxins have been identified as specific inhibitors of the sodium pump. Among them, the so-called cardioactive steroids or cardiac glycosides are not only known for their ability to selectively target the sodium pump but are widely used as effective medication for patients with heart failure or heart insufficiency [4]. Paly- toxin, a toxin isolated from corals of the family Palythoa (e.g. Palythoa caribaeorum), is also a highly specific inhibitor of the sodium pump [5,6]. This most potent toxin (for rodents the LD 50 is 10–250 ng per kg of body weight) of animal origin can also be found together with ciguatoxin, maitotoxin, or gambierol in fishes that contribute to ciguatera poisonings [7,8]. Palytoxin is a rather unique and large molecule with the structural formula C 129 H 223 N 3 O 54 . The molecule can be divided into three subdomains, each connected by peptide bonds: a large N-terminal polyhydroxy x-amino acid followed by a dehydro-b-alanine residue and an aminopropanol group. The number of free hydroxyl groups is 42 [9,10]. Unlike the cardioactive steroids, however, which inhibit both ATP hydrolysis and ion conduction, palytoxin acts by arresting the ionophore of the pump into a permanently open state. Thus, in this case, inhibition of ATP hydrolysis is no longer associated with inhibition of ion conductivity. Yeast cells, which are usually insensitive to palytoxin, display a palytoxin-induced K + efflux when they hetero- logously express a and b subunits of the mammalian sodium pump [11,12]. This flux is sensitive to ouabain (g-strophan- thine), the most well known inhibitor of the pump. Based on these and other experiments showing the formation of palytoxin-induced ion channels in membranes containing in vitro-translated Na + /K + -ATPase [13], it is widely accepted that palytoxin specifically targets the sodium pump and inhibits its catalytic activity by converting the ATPase into an ion channel. Palytoxin action on other P 2C -type ATPases has not yet been demonstrated. Thus, in the current investigation, we describe the action of palytoxin on the H + /K + -ATPase from the rat colon and demonstrate that the interaction of this toxin with the enzyme results in specific currents that are similar to those observed from its action on sodium pumps. EXPERIMENTAL PROCEDURES Solutions The Ussing chamber experiments were carried out in a bathing solution containing according to Parsons & Paterson [14] (mmolÆL )1 ): NaCl, 107; KCl, 4.5; NaHCO 3 , 25; Na 2 HPO 4 ,1.8;NaH 2 PO 4 ,0.2;CaCl 2 ,1.25;MgSO 4 ,1; andglucose,12.Thesolutionwasgassedwithamixtureof 5% CO 2 and 95% O 2 ; the pH was 7.4. For depolarization of the basolateral membrane, a modified bathing solution was used in which NaCl was replaced by 111.5 mmolÆL )1 KCl. In the LiCl bathing solution, NaCl was replaced equimo- larly by LiCl. Correspondence to G. Scheiner-Bobis, Institut fu ¨ r Biochemie und Endokrinologie, Justus-Liebig-Universita ¨ t Gießen, Frankfurter Str. 100, D-35392 Gießen, Germany. Fax: + 49 641 99 38189, Tel.: + 49 641 99 38180, E-mail: Georgios.Scheiner-Bobis@vetmed.uni-giessen.de Abbreviations:NMDG,N-methyl- D -glucamine; Gt, tissue conductance; Isc, short-circuit current. (Received 11 March 2002, revised 14 May 2002, accepted 18 June 2002) Eur. J. Biochem. 269, 3905–3911 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03056.x Tissue preparation Wistar rats were used with a weight of 180–220 g. The animals had free access to water and food until the day of the experiment. Animals were stunned by a blow on the head and killed by exsanguination (approved by Regi- erungspra ¨ sidium Giessen, Giessen, Germany). The serosa and muscularis propria were stripped away by hand to obtain the mucosa-submucosa preparation of the distal part of the colon descendens. Two distal and two proximal segments of the colon of each rat were prepared. Short-circuit current measurement The tissue was mounted in a modified Ussing chamber, bathed with a volume of 3.5 mL (see above) on each side of the mucosa and short-circuited by a voltage clamp (Ing. Buero Mussler, Aachen, Germany) with correction for solution resistance as described previously [15]. The exposed surface of the tissue was 1 cm 2 . Short-circuit current (Isc) was continuously recorded and tissue conductance (Gt) was measured every min. Isc is expressed as lAÆh )1 Æcm )2 , i.e. the flux of a monovalent ion per time and area with 1 lEqÆh )1 Æcm )2 ¼ 26.9 lAÆcm )2 . Tissues were left for 1 h to stabilize the Isc before the effect of drugs was studied. The baseline electrical parameters were determined as the mean obtained during 3 min just before administration of a drug. Immunohistochemical detection of the Na + /K + -ATPase in colonic epithelium Wistar rats (n ¼ 2) were anesthetized with sodium pento- barbital (60 mgÆkg )1 body weight; Narcoren, Merial GmbH, Hallbergmoos, Germany) and transcardially per- fused with 4% paraformaldehyde in 100 mmolÆL )1 phos- phate buffer (pH 7.2). The distal colon was removed and postfixed in the same fixative for 1 h at room temperature and then the tissue was cryoprotected in 20% sucrose in phosphate buffer overnight at 4 °C. Tissue was cut the following day. Coronal 10–12 lm colonic sections were cut on a cryostat (model HM 500, Microm, Walldorf, Germany). To detect Na + /K + -ATPase immunoreactivity, a commercial tyra- mide amplification kit (NEL700, NEN Life Science Prod- ucts GmbH, Cologne, Germany), based on the catalyzed reporter deposition method, was used. Tyramide amplifica- tion staining was performed according to the kit description in a phosphate buffer system (pH 7.2). In detail, sections were placed in 10% fetal bovine serum containing 0.3% Triton X-100 for 1 h at room temperature. Incubation with the primary anti-(Na + /K + -ATPase) Ig (MA3-929, mon- oclonal, mouse, a 1 -subunit, Affinity BioReagents, Golden, CO, USA) was performed for 24–36 h at 4 °C at a dilution of 1 : 150 to 1 : 5000). The primary antibody was then detected with a secondary biotinylated anti-(mouse IgG) Ig (1 : 200, Vector BA-2001, Linaris Biologische Produkte, Wertheim-Bettingen, Germany) for 1 h at room tempera- ture. After amplification, the immunohistochemical processing was finished with 1 : 200 fluorescein (FITC)- conjugated avidin D (Vector, Linaris Biologische Produkte, Wertheim-Bettingen, Germany). In order to demonstrate the overall morphology of the colonic epithelium, parallel series of sections adjacent to cryosections of the immuno- fluorescent-stained series were cut for light-microscopic analysis and consequently counterstained using cresylviolet. Finally, these sections were cover slipped with Entellan (Merck, Darmstadt, Germany) while immunofluorescent sections were cover slipped with crystal/mount (Biomedia, FosterCity, USA). Microscopic analysis Sections were analyzed using a an Olympus BX50 light/ fluorescent microscope (Olympus Optical Co., Hamburg, Germany). For light microscopy, digital images were taken with an Olympus Camedia 3030 camera using the Olympus CAMEDIA MASTER software package (Olympus Optical Co., Hamburg, Germany). For fluorescent microscopy, digital images were taken with a Visicam (PCO Computer Optics, Kehlheim, Germany) using the METAMORPH/METAFLUOR software package (Visitron Systems, Puchheim, Germany). Image editing software ( ADOBE PHOTOSHOP ) was used to adjust brightness and contrast and to combine the individual images into the greyscale mode figure plate. Drugs Palytoxin (purchased from L. Be ´ ress, Institute for Toxicol- ogy, University of Kiel, Germany) was dissolved in 10 mmolÆL )1 Hepes, 0.5 mmolÆL )1 Tris, 1 mmolÆL )1 CaCl 2 and 1 gÆL )1 BSA. Sodium orthovanadate (Calbiochem, Bad Soden, Germany) was dissolved in an aqueous stock solution. Ouabain was dissolved in dimethylsulfoxide (final concentration 2.5 lLÆmL )1 ), scilliroside (Sandoz, Basel, Switzerland) was dissolved in methanol (final concentration 2.5 lLÆmL )1 ). If not indicated differently, drugs were from Sigma, Deisenhofen, Germany. Statistics Results are given as means ± SEM. When the means of several groups were compared, an analysis of variances was first performed. If the analysis of variances indicated significant differences between the groups investigated, further comparison was carried out by a Student’s t-test (paired or unpaired as appropriate) or by the Mann– Whitney U-test. An F-test was applied to decide which test method was to be used. RESULTS Basal effects of palytoxin Palytoxin (10 )8 molÆL )1 on the mucosal side) induced an increase in short-circuit current (Isc) in rat distal and proximal colon (Fig. 1A). The response started immediately after administration of the toxin and was stable at least for 30 min The effect was concentration-dependent (Fig. 1B). A first, significant increase in Isc occurred at a concentration of 10 )10 molÆL )1 . Similar effects were observed in the distal and proximal colon, although the potency of palytoxin appeared to be higher in the distal than in the proximal colon. The increase in Isc was concomitant with a rise in tissue conductance (Gt). At the highest concentration of palytoxin tested (5 · 10 )8 molÆL )1 ), Gt increased by 3906 G. Scheiner-Bobis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 7.1 ± 1.7 msÆcm )2 in the distal and by 8.9 ± 2.7 msÆcm )2 in the proximal colon (n ¼ 6–8, p < 0.05 for both colonic segments). The effect of palytoxin was enhanced in the presence of mucosal borate (0.5 mmolÆL )1 ), especially in the proximal colon (Table 1). Similar observations were made in the past concerning palytoxin effects on erythrocytes, neurosyna- ptosomes or yeast cells that express mammalian sodium pumps [6,11]. Although no real evidence exists about the role of borate, borate alone does not induce any cation fluxes from erythrocytes [6] or from yeast expressing the mammalian sodium pump [11] or from the colon tissues investigated here. It is possible that borate interacts with some of the 42 free hydroxyl groups of palytoxin, similarly to the way it interacts with carbohydrates. It also might be that it interacts with the carbohydrates of the strongly glycosylated b subunits of the P 2C -type ATPases. These possible complexes might induce a particular conformation of the palytoxin molecule or of the enzyme that favors mutual interaction between the two reactants. Therefore, all subsequent experiments were carried out with borate in the mucosal solution using a palytoxin concentration of 10 )8 molÆL )1 . For theoretical reasons it is not possible that admin- istration of palytoxin to the basolateral side of an epithelium can induce an Isc. If the toxin converts the Na + /K + -pump into a cation channel, the cytosolic Na + concentration will increase and finally reach the extracel- lular concentration, whereas the cytosolic K + concentra- tion will fall to the level at the extracellular side. Thus there is no more driving force for any active ion movement, i.e. there will be no short-circuit current response. Therefore, as expected, when applied at the serosal side, in six independent experiments the toxin had no effect on Isc (data not shown). Sensitivity against inhibitors of ATPases The effect of palytoxin (10 )8 molÆL )1 at the mucosal side) was resistant to mucosal ouabain (10 )3 molÆL )1 ) or scill- iroside (10 )4 molÆL )1 at the mucosal side) (Table 1), a potent blocker of the Na + /K + -pumpinrattissue[16].All pump inhibitors were administered 1 h prior to palytoxin; for effects of the blockers on baseline Isc, see Table 2. In contrast, pretreatment with sodium orthovanadate (10 )4 molÆL )1 at the mucosal side) nearly suppressed the action of palytoxin (Table 1). Table 1. Effect of palytoxin on Isc under different conditions. The increase in Isc evoked by palytoxin (10 )8 molÆL )1 at the mucosal side) was measured in the absence of any drugs or in the presence of Tris borate (10 )4 molÆL )1 at the mucosal side; pretreated for 15 min), ouabain (10 )3 molÆL )1 at the mucosal side; pretreated for 1 h), scilliroside (10 )4 molÆL )1 at the mucosal side; pretreated for 1 h), or vanadate (10 )4 molÆL )1 at the mucosal side; pretreated for 1 h), or after replacement of NaCl by NMDG chloride (107 mmolÆL )1 NMDG chloride buffer at the mucosal side). *p < 0.05 vs. baseline, p < 0.05 vs. response to palytoxin in the absence of any drugs. Distal colon D Isc (lEqÆh )1 Æcm )2 ) Proximal colon D Isc (lEqÆh )1 Æcm )2 ) n Palytoxin 2.7 ± 0.6* 0.3 ± 0.2 5–8 Palytoxin + borate 4.2 ± 0.7* 1.9 ± 0.4*  5–9 Palytoxin 2.8 ± 0.8* 2.0 ± 0.9 5–9 Palytoxin + ouabain 2.3 ± 0.7* 2.2 ± 0.6* 6–9 Palytoxin + vanadate 0.4 ± 0.1* 0.3 ± 0.1 6 Palytoxin + scilliroside 2.1 ± 0.6* 1.1 ± 0.5 6–8 Palytoxin, Na ± free 0.1 ± 0.1 0.1 ± 0.1 6 Basolateral depolarization 0.5 ± 0.1* 0.1 ± 0.1 5–6 (apical Na + ) Basolateral depolarization )1.7 ± 0.5* )0.9 ± 0.2* 6–8 (apical Li + ) Fig. 1. Induction of a short-circuit current in rat distal and proximal colon by palytoxin. (A) Typical Isc response evoked by palytoxin (10 )8 molÆL )1 at the mucosal side in the presence of 0.5 mmolÆL )1 Na borate at the mucosal side). (B) Concentration-dependent increase in Iscabovebaseline(D Isc) evoked by palytoxin in the distal (closed circles) and proximal (open rectangles) rat colon. Palytoxin was administered cumulatively at the mucosal side in the presence of 0.5 mmolÆL )1 Na borate. Values are means ± SEM, n ¼ 6–8. Ó FEBS 2002 Palytoxin action on colonic H + /K + -ATPase (Eur. J. Biochem. 269) 3907 Ionic selectivity of the palytoxin-induced pore Assuming that palytoxin might induce cation-permeable pores in the apical membrane, the Isc evoked by the toxin should consist of an influx of Na + , the prevalent cation in the mucosal solution, into the cell with the consequence of the stimulation of a pump current generated by the basolateral Na + /K + -ATPase [17]. In order to test this hypothesis, NaCl was replaced by NMDG chloride in the buffer solution. Under these conditions, palytoxin (10 )8 molÆL )1 at the mucosal side) no longer had any significant effect on Isc (Table 1). In order to elucidate the cationic selectivity of the palytoxin-induced pore, a protocol was used in which the basolateral membrane was electrically eliminated by a basolateral depolarization. The basolateral membrane was depolarized by a high K + solution (111.5 mmolÆL )1 KCl at the serosal side). Due to the high basolateral K + perme- ability, the electrical properties of the tissue, which are normally characterized by two batteries in series, are then expected to be dominated by the apical membrane [18]. Consequently, the current evoked by palytoxin in the presence of different monovalent cations should not be affected by the ionic selectivity of the basolateral Na + /K + - ATPase. Depolarization of the basolateral membrane induces a negative current across the tissue (Fig. 2A) due to diffusion of K + across the apical membrane into the mucosal compartment, which is driven by the applied K + gradient as reported previously [17]. When palytoxin (10 )8 molÆL )1 at the mucosal side) was administered in the presence of mucosal Na + (107 mmolÆL )1 ), the toxin induced a prompt increase in Isc, especially in the distal colon (Fig. 2A, Table 1). This suggests that the pore induced by palytoxin has a permeability for Na + that is higher compared with that for K + , leading to a flux of Na + from the mucosal to the serosal compartment and an increase in Isc. An opposite effect was observed when LiCl was present in the mucosal solution. Under these conditions, palytoxin (10 )8 molÆL )1 at the mucosal side) stimulated a negative current (Fig. 2B), suggesting that the pores formed by palytoxin in the apical membrane have a permeability for K + that is higher than that for Li + , thereby mediating a flux of K + from the serosal into the mucosal compartment driven by the chemical gradient. Similar experiments were performed with CsCl in the mucosal compartment; however, this solution proved to be toxic for the tissue as indicated by a massive increase in Gt. Nevertheless, this series of experiments suggests a permeability sequence of the palytoxin-formed pore of Na + >K + >Li + . Morphological control experiments ANa + /K + -stimulated ATPase activity has been found in vesicles isolated from the apical membrane of rat distal colon [19]. Therefore, using an immunohistochemical tech- nique we investigated whether we could find evidence for a Na + /K + -ATPase in the apical membrane using a mono- clonal antibody against the murine a 1 -subunit of this pump. The overall morphology of rat distal colonic crypts was demonstrated in cresylviolet counterstained sections (Fig. 3A). In adjacent immunhistochemically processed sections immunoreactivity of the a 1 -subunit of the Table 2. Effect of inhibitors/activators on baseline Isc. Concentration of drugs were: Tris borate (10 )4 molÆL )1 at the mucosal side), ouabain (10 )3 molÆL )1 at the mucosal side), vanadate (10 )4 molÆL )1 at the mucosal side). *p < 0.05 vs. baseline. Distal colon D Isc (lEqÆh )1 Æcm )2 ) Proximal colon D Isc (lEqÆh )1 Æcm )2 ) n Borate 0.4 ± 0.2 0.7 ± 0.7 5–9 Ouabain )0.2 ± 0.3 )0.6 ± 0.4 6–9 Vanadate )0.6 ± 0.4 )1.0 ± 0.2* 6 Fig. 2. Action of palytoxin (10 )8 molÆL )1 at the mucosal side) under conditions in which the basolateral membrane was depolarized by a high concentration of K + (111.5 mmolÆL )1 KCl solution at the serosal side; black bar) either in the presence of mucosal Na + (107 mmolÆL )1 ;A)or mucosal Li + (107 mmolÆL )1 ;B).The two schematic drawings sum- marize the experimental conditions. The line tracings are typical for 5–8 experiments with similar results; for statistical evaluation, see Table 1. 3908 G. Scheiner-Bobis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Na + /K + -ATPase was restricted to the basolateral mem- brane of rat colonic epithelial cells, as shown for the sagital (Fig. 3C) and coronal plane (Fig. 3D). The specificity of this immunohistochemical signal was further analyzed in con- trol experiments. Omission of the primary antibody led to a dramatically reduced immunoreactivity in particular at the basolateral membrane of the colonic epithelial cells (com- pare Fig. 3B with Fig. 3C). DISCUSSION Palytoxin is produced from corals of the genus Palythoa.It is the strongest toxin produced by animals, with an LD 50 for rodents of 10–250 ngÆkg )1 body weight. The toxin is without any effect on bacterial or native yeast cells. In erythrocytes and other animal cells, however, it induces an efflux of K + ions from the cytosol and a series of secondary effects that are most likely associated with the cell depolar- ization induced by the K + loss. Using yeast as a heterologous expression system, the sodium pump was shown to be the target of palytoxin [11,12], which, as verified in experiments involving in vitro translation and integration of the expressed proteins in membranes [13], is converted by the toxin into an ion channel [20]. This channel, very much like natural ion channels, allows ions to flow through it following their electrochemical gradients. Apparently, nature has devel- oped here a highly effective toxic principle; the conversion of a pump into a channel, most likely by arresting the natural ionophore of the pump into a permanently open state. Palytoxin acts by binding to extracellular sites [6,21] of the sodium pump. Considering the high degree of homology between the various P 2C -type ATPases, which approaches 65% [3], we were interested in investigating the interactions of the toxin with the H + /K + -ATPase from rat colon. The experiments were carried out in an Ussing chamber, which, compared to the single-cell paradigm, has the advantage of allowing one to asses at any time of the measurement either the apical or the basolateral surfaces of the epithelial cell membrane. Administration of palytoxin on the apical membrane results in the generation of a current (Fig. 1A), which is dependent on the presence of Na + ions (Table 1). The most plausible explanation for this observation is that palytoxin induces the formation of cation channels in the apical membrane mediating the influx of Na + ,which,when extruded by the electrogenic basolateral Na + /K + -ATPase, leads to the generation of a transepithelial current. The Fig. 3. Immunohistochemical detection of the a 1 -subunit of the Na + /K + -ATPase in rat colonic crypts. Photomicrographs of rat colonic crypts are shown in the sagital (A–C) and coronal (D) plane. The morphology of sagital colonic crypts can be detected from a cresylviolet counterstained cryosection (10–12 lm) as shown in A. The immunofluorescent detection of the primary antibody MA3-929 (Affinity BioReagents, dilution 1 : 500) is shown in (C) and (D). Note the intense immunoreactivity at the basolateral membrane of the proliferating cryptic cells [see white arrowheads in (C) and inset in (B,D)] as compared to the basal immunohistochemical signal detected in an adjacent control cryosection (B) in which the primary antibody was omitted [see white arrowheads in (B)]. The dotted frame area in (D) shows a coronal view of a single colonic crypt, which is also shown at higher magnification in the inset (B,D). Scale bar (A–D), 50 lm. Scale bar inset (B,D), 20 lm. Ó FEBS 2002 Palytoxin action on colonic H + /K + -ATPase (Eur. J. Biochem. 269) 3909 pores induced by palytoxin have an apparent permeability of Na + >K + >Li + as suggested by cation experiments in which the basolateral membrane was by-passed by a basolateral depolarization (Fig. 2). The inhibition of palytoxin-induced currents by vanadate, a known blocker of apical H + /K + -ATPase [22], suggests that this enzyme is involved inthe formation of the palytoxin- induced channels. The enzymatic activity [19], the amount of mRNA, and the protein expression of the H + /K + -ATPase are higher in the distal compared to the proximal colon [23]. Therefore, it seems reasonable to assume that this segmental heterogeneity might be responsible for the higher sensitivity of the distal colon to palytoxin when compared with the proximal part of this organ (Fig. 1B). In the experiments described here, ouabain did not have any effect on the palytoxin-induced current across the apical membrane of colon epithelial cells. Ouabain inhibition of the colonic H + /K + -ATPase has been a matter of contro- versy in many investigations. Early experiments involving ATPase measurements on apical membrane preparations appeared to indicate the presence of at least two types of H + /K + -ATPase, denoted ouabain-sensitive and ouabain- resistant. These results were supported by expression cloning experiments showing a ouabain-sensitive form when the a subunit of colonic H + /K + -ATPase was expressed together in Xenopus oocytes with the bsubunit of the H + / K + -ATPase from toad bladder or with either the b subunit of gastric H + /K + -ATPase or the b subunit of Na + /K + - ATPase in HEK293 cells [24,25]. In Sf9 cells, however, expression of H + /K + -ATPase a subunit without a corre- sponding b subunit produces a ouabain-insensitive H + /K + - ATPase [26]. Finally, coexpression of cDNAs encoding the H + /K + -ATPase a and b subunits also results in ouabain- resistant enzymes [22]. These, however, remain highly sensitive to orthovanadate [22]. Our measurements indi- rectly support these latter findings, as the palytoxin-induced conductance was insensitive to 1 m M ouabain while being inhibited at low concentrations of orthovanadate. Similar findings showing a palytoxin-induced K + efflux that is resistant to ouabain have also been demonstrated using rat erythrocytes, which contain Na + /K + -ATPase but not H + /K + -ATPase [21]. Thus, in order to exclude a possible involvement of the sodium pump in the currents observed here, the localization of the sodium pump a subunit was investigated by an immunohistochemical method applied to sagital and coronal segments of the rat distal colonic crypts. As shown in Fig. 3, the presence of Na + /K + -ATPase a 1 subunit in the apical membranes can be excluded, allowing us to conclude that the previously measured Na + -andK + -stimulated ATPase activity found in the apical membranes of the colon is not due to Na + / K + -ATPase, but could be associated with some form(s) of the colonic H + /K + -ATPase, as suggested previously [27,28]. The overall conclusion of the investigation presented here is that palytoxin targets not only the sodium pump but also the colonic H + /K + -ATPase. Apparently, as with the Na + / K + -ATPase, the toxin converts the H + /K + -ATPase into an ion channel that allows cations to pass down their electrochemical gradients. This ion channel may be the ionophore of the pump, which under physiological condi- tions is only accessible from one side of the plasma membrane and becomes arrested in a permanently open state upon interaction with the toxin. This conclusion is indirectly supported by the fact that monovalent cations can penetrate the channel, whereas large or divalent cations fail to be conducted. Taking into consideration the relatively high homology of the Na + /K + -ATPase and colonic H + /K + -ATPase with the gastric H + /K + -ATPase, one might expect to obtain similar palytoxin effects with the latter enzyme. Although this has yet to be investigated, the fact that other P 2 -type cation pumps such as the yeast Na + -ATPase or Ca 2+ - ATPase are not sensitive to palytoxin suggests that P 2C -type ATPases are exclusive targets of the toxin. Furthermore, because the b subunit has been shown to influence the enzyme kinetic properties of these cation pumps, this subunit might be required for and closely associated with the formation of the palytoxin-induced cation channel. ACKNOWLEDGEMENTS We wish to acknowledge the diligent care of B. Bru ¨ ck, A. Metternich, B. Schmidt and E. Haas. G. S. B. was supported through the Deutsche Forschungsgemeinschaft grant Sche 307/5-1 and M. D. through the Deutsche Forschungsgemeinschaft grant Di 388/3-4. REFERENCES 1. Axelsen, K.B. & Palmgren, M.G. (1998) Evolution of substrate specificities in the P-type ATPase superfamily. J. Mol. Evol. 46, 84–101. 2. Scheiner-Bobis, G. (2002) The sodium pump: its molecular properties and mechanics of ion transport. Eur. J. Biochem. 269, 2424–2433. 3. Jaisser, F. & Beggah, A.T. (1999) The nongastric H + -K + - ATPases: molecular and functional properties. Am.J.Physiol. 276, F812–F824. 4. Hansen, O. (1984) Interaction of cardiac glycosides with (Na + ,K + )-activated ATPase. Biochemical link to digitalis-in- duced inotropy. Pharmacol. Rev. 36, 143–163. 5. Ishida, Y., Tagaki, K., Takahashi, M., Satake, N. & Shibata, S. (1983) Palytoxin isolated from marine coelenterates: the inhibitory action on (Na,K)-ATPase. J. Biol. Chem. 258, 7900–7902. 6. Habermann, E. (1989) Palytoxin acts through Na + /K + -ATPase. Toxicon 27, 1171–1187. 7. Kodama, A.M., Hokama, Y., Yasumoto, T., Fukui, M., Manea, S.J. & Sutherland, N. (1989) Clinical and laboratory findings implicating palytoxin as cause of ciguatera poisoning due to Decapterus macrosoma (mackerel). Toxicon 27, 1051–1053. 8. Wachi, K.M., Hokama, Y., Haga, L.S., Shiraki, A., Takenaka, W.E., Bignami, G.S. & Levine, L. (2000) Evidence for palytoxin as one of the sheep erythrocyte lytic in lytic factors in crude extracts of ciguateric and non-ciguateric reef fish tissue. J. Nat. Toxins 9, 139–146. 9. Moore, R.E. & Bartolini, G.J. (1981) Structure of palytoxin. J. Am. Chem. Soc. 103, 2491–2494. 10. Cha, J.K., Christ, W.J., Finan, J.M., Fuijoka, M., Kishi, J., Klein, L.L., Ko, S.S., Leder, J., McWhorter, W.W., Pfaff, K P., Yonaga, M., Uemura, D. & Hirata, Y. (1982) Stereochemistry of palytoxin. 4. Complete structure. J. Am. Chem. Soc. 104, 7369–7371. 11. Scheiner-Bobis, G., Meyer zu Heringdorf, D., Christ, M. & Habermann, E. (1994) Palytoxin induces K + efflux from yeast cells expressing the mammalian sodium pump. Mol. Pharmacol. 45, 1132–1136. 12. Redondo, J., Fiedler, B. & Scheiner-Bobis, G. (1996) Palytoxin- induced Na + influx into yeast cells expressing the mammalian sodium pump is due to the formation of a channel within the enzyme. Mol. Pharmacol. 49, 49–57. 3910 G. Scheiner-Bobis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 13. Hirsh, J.K. & Wu, C.H. (1997) Palytoxin-induced single-channel currents from the sodium pump synthesized by in vitro expression. Toxicon 35, 169–176. 14. Parsons, D.S. & Paterson, C.R. (1965) Fluid and solute transport across rat colonic mucosa. Quart. J. Exp. Physiol. 50, 220–231. 15. Reuss, L. (2001) Ussing’s two-membrane hypothesis: the model and half a century of progress. J. Membr. Biol. 184, 211–217. 16. Robinson, J.W. (1970) The difference in sensitivity to cardiac steroids of (Na + +K + )-stimulated ATPase and amino acid transport in the intestinal mucosa of the rat and other species. J. Physiol. 206, 41–60. 17. Schultheiss, G. & Diener, M. (1997) Regulation of apical and basolateral K + conductances in rat colon. Br. J. Pharmacol. 122, 87–94. 18. Fuchs, W., Larsen, E.H. & Lindemann, B. (1977) Current-voltage curve of sodium channels and concentration-dependence of sodium permeability in frog skin. J. Physiol. 267, 137–166. 19. Del Castillo, J.R., Rajendran, V.M. & Binder, H.J. (1991) Apical membrane localization of ouabain-sensitive K + -activated ATPase activities in rat distal colon. Am. J. Physiol. 261, G1005– G1011. 20. Scheiner-Bobis, G. (1998) Ion-transporting ATPases as ion channels. Naunyn-Schmiedeberg’s Arch. Pharmacol. 357, 477–482. 21. Habermann, E., Hudel, M. & Dauzenroth, M E. (1989) Palytoxin promotes potassium outflow from erythrocytes, Hela and bovine adrenomedullary cells through its interaction with Na + /K + - ATPase. Toxicon 27, 419–430. 22. Sangan, P., Thevananther, S., Sangan, S., Rajendran, V.M. & Binder, H.J. (2000) Colonic H-K-ATPase alpha- and beta-sub- units express ouabain-insensitive H-K-ATPase. Am.J.Physiol. 278, C182–C189. 23. Sangan, P., Rajendran, V.M., Mann, A.S., Kashgarian, M. & Binder, H.J. (1997) Regulation of colonic H-K-ATPase in large intestine and kidney by dietary Na depletion and dietary K depletion. Am. J. Physiol. 272, C685–C696. 24. Cougnon, M., Planelles, G., Crowson, M.S., Shull, G.E., Rossier, B.C. & Jaisser, F. (1996) The rat distal colon P-ATPase alpha subunit encodes a ouabain-sensitive H + ,K + -ATPase. J. Biol. Chem. 271, 7277–7280. 25. Asano, S., Hoshina, S., Nakaie, Y., Watanabe, T., Sato, M., Suzuki, Y. & Takeguchi, N. (1998) Functional expression of putative H + -K + -ATPase from guinea pig distal colon. Am. J. Physiol. 275, C669–C674. 26. Lee, J., Rajendran, V.M., Mann, A.S., Kashgarian, M. & Binder, H.J. (1995) Functional expression and segmental localization of rat colonic K-adenosine triphosphatase. J. Clin. Invest. 96, 2002– 2008. 27. Cougnon, M., Bouyer, P., Planelles, G. & Jaisser, F. (1998) Does the colonic H,K-ATPase also act as an Na,K-ATPase? Proc. Natl Acad.Sci.U.S.A.95, 6516–6520. 28. Rajendran, V.M., Sangan, P., Geibel, J. & Binder, H.J. (2000) Ouabain-sensitive H,K-ATPase functions as Na,K-ATPase in apical membranes of rat distal colon. J. Biol. Chem. 275, 13035– 13040. Ó FEBS 2002 Palytoxin action on colonic H + /K + -ATPase (Eur. J. Biochem. 269) 3911 . yet been demonstrated. Thus, in the current investigation, we describe the action of palytoxin on the H + /K + -ATPase from the rat colon and demonstrate that. borate in the mucosal solution using a palytoxin concentration of 10 )8 molÆL )1 . For theoretical reasons it is not possible that admin- istration of palytoxin