Role of the Na+/K+/Cl- transporter in the positive inotropic effect of ouabain in cardiac myocytes

Investigation of Differentially Expressed Genes in the Ventricular Myocardium of Senescent Rats

Aging alters a variety of physiological functions of the heart. The molecular basis of the age-related functional changes has not been fully understood. Differential gene expression provides the basis for many fundamental cellular processes associated with development and aging. The identification and cloning of genes whose expression is modulated by aging can be of importance for our better understanding of these age-related phenomena. In order to isolate and characterize gene products differentially expressed in senescent hearts, we applied a differential display method for screening those genes in rat ventricular myocardium. Total RNAs were isolated from 2-month-old (young) and 24-month-old rat (senescent) ventricles by the acid-guanidium-phenol-chloroform method. The first-strand synthesis of the cDNAs from each RNA was carried out with oligo-d(T) primers. The differential display screening was performed with three arbitrary primers and eight anchor primers, and the products were isolated on a 6% denaturing polyacrylamide gel. The bands showing differential expression were excised and subcloned into T-vector. We selected 19 upregulated clones and 66 downregulated clones in aged rat hearts. The differential expression of those candidate genes was confirmed by reverse Northern blot analysis. The selected genes were sequenced by dye-terminator methods. Among 31 clones, 15 clones were unknown. The known products included alpha-myosin heavy chain, cytochrome oxidase subunit, H(+)-transporting ATP synthase F0 complex subunit c isoform 3 (ATP5G3), and Na(+)-K(+)-Cl(-) cotransporter. The RT-PCR differential display method effectively identified genes differentially expressed in senescent hearts, and may be a useful tool for investigating factors responsible for age-related physiological changes.

α1-AR-mediated activation of NKCC in rat cardiomyocytes involves ERK-dependent phosphorylation of the cotransporter

We studied molecular and functional characteristics as well as hormonal regulation of the Na-K-2Cl cotransporter (NKCC) in the isolated rat heart and cardiomyocytes. NKCC activity was measured as bumetanide-sensitive86Rb+influx in isolated perfused rat hearts and isolated cardiomyocytes. Stimulation of α1-adrenoceptors (AR) by phenylephrine (30 μM) increased86Rb+influx. The NKCC inhibitor bumetanide (50 μM) reduced the response to phenylephrine by 45 ± 13% ( n = 12, P < 0.01). PD-98059 (10 μM), an inhibitor of the activation of the mitogen-activated protein kinases extracellular signal-regulated protein kinase 1 and 2 (ERK1/2), reduced the total response to phenylephrine by 51 ± 13% ( n = 10, P < 0.01) and eliminated the bumetanide-sensitive component, indicating that α1-AR mediated stimulation of NKCC is dependent on activation of ERK1/2. Inhibitors of protein kinase C or phosphatidylinositol 3-kinase had no effect. The presence of NKCC mRNA and protein was demonstrated in isolated rat cardiomyocytes. Phosphorylation of NKCC after α1-AR stimulation was shown by immunoprecipitation of the phosphoprotein from32Piprelabeled cardiomyocytes. Increased phosphorylation of the NKCC protein was also abolished by PD-98059. We conclude that the NKCC is present in rat cardiomyocytes and that ion transport by the cotransporter is regulated by α1-AR stimulation through phosphorylation of this protein involving the ERK pathway.

Decreased blood pressure and vascular smooth muscle tone in mice lacking basolateral Na(+)-K(+)-2Cl(-) cotransporter

The basolateral Na(+)-K(+)-2Cl(-) cotransporter (NKCC1) functions in the maintenance of cellular electrolyte and volume homeostasis. NKCC1-deficient (Nkcc1(-/-)) mice were used to examine its role in cardiac function and in the maintenance of blood pressure and vascular tone. Tail-cuff measurements demonstrated that awake Nkcc1(-/-) mice had significantly lower systolic blood pressure than wild-type (Nkcc1(+/+)) mice (114.5 +/- 2.2 and 131.8 +/- 2.5 mmHg, respectively). Serum aldosterone levels were normal, indicating that extracellular fluid-volume homeostasis was not impaired. Studies using pressure transducers in the femoral artery and left ventricle showed that anesthetized Nkcc1(-/-) mice have decreased mean arterial pressure and left ventricular pressure, whereas myocardial contraction parameters were not significantly different from those of Nkcc1(+/+) mice. When stimulated with phenylephrine, aortic smooth muscle from Nkcc1(+/+) and Nkcc1(-/-) mice exhibited no significant differences in maximum contractility and only moderate dose-response shifts. In phasic portal vein smooth muscle from Nkcc1(-/-) mice, however, a sharp reduction in mechanical force was noted. These results indicate that NKCC1 can be important for the maintenance of normal blood pressure and vascular tone.

Na(+)/K(+)/Cl(-) cotransporter activates mitogen-activated protein kinase in fibroblasts and lymphocytes

In a previous work, we have shown that overexpression of the Na(+)/K(+)/Cl(-) cotransporter (NKCC1) induces cell proliferation and transformation. We investigate in the present study the role of the NKCC1 in the mitogenic signal transduction. We show that overexpression of the cotransporter gene (NKCC1) in stablely transfected cells (Balb/c-NKCC1), resulted in enhanced phosphorylation of the extracellular regulated kinase (ERK) to produce double phosphorylated ERK (DP-ERK). Furthermore, the level of DP-ERK was reduced by 50-80% following the addition of bumetanide, a specific inhibitor of the Na(+)/K(+)/Cl(-) cotransporter, in quiescent as well as in proliferating cultures of the Balb/c-NKCC1 clone. In order to explore further the role of the Na(+)/K(+)/Cl(-) cotransporter in mitogenic signal transduction, we measured the effect of the two specific inhibitors of the cotransporter; bumetanide and furosemide, on DP-ERK level in immortalized non-transformed cells. In Balb/c 3T3 fibroblasts stimulated with FGF, bumetanide, and furosemide inhibited 50-60% of the ERK 1/2 phosphorylation. The inhibitor concentration needed for maximal inhibition of ERK 1/2 phosphorylation was similar to the concentration needed to block the K(+) influx mediated by the Na(+)/K(+)/Cl(-) cotransporter in these cells. To analyze whether the Na(+)/K(+)/Cl(-) cotransporter has a role in the mitogenic signal of normal cells, we measured the effect of bumetanide on ERK phosphorylation in human peripheral blood lymphocytes. The phosphorylation of ERK 1/2 in resting human lymphocytes, as well as in lymphocytes stimulated with phytohemagglutinin (PHA) was inhibited by bumetanide. The effect of bumetanide on ERK 2 phosphorylation was much lower than that of ERK 1 phosphorylation. The finding that the Na(+)/K(+)/Cl(-) cotransporter controls the ERK/MAPK (mitogen-activated protein kinase) signal transduction pathway, support our hypothesis that Na(+) and K(+) influxes mediated by this transporter plays a central role in the control of normal cell proliferation. Exploring the cellular ionic currents and levels, mediated by the Na(+)/K(+)/Cl(-) cotransporter, should lead to a better comprehension of cell proliferation and transformation machinery.

Overexpression of the Na(+)/K(+)/Cl(-) cotransporter gene induces cell proliferation and phenotypic transformation in mouse fibroblasts

Na(+)/K(+)/Cl(-) cotransporter activity is stimulated in early G(1) phase of the cell cycle and this stimulation was shown to be an essential event in fibroblast cell proliferation. In order to elucidate further the role of the Na(+)/K(+)/Cl(-) cotransporter in cell proliferation, we overexpressed the gene encoding the Na(+)/K(+)/Cl(-) cotransporter in mouse fibroblasts, and analyzed cellular phenotypic changes. Mouse Balb/c 3T3 cells were stably transfected with the cDNA of the shark rectal gland Na(+)/K(+)/Cl(-) cotransporter gene (NKCC1), and expressed in a mammalian vector under the cytomegalovirus promoter (Balb/c-NKCC1 cells). The transfected cells exhibited up to 10-fold greater bumetanide-sensitive Rb(+) influx compared to the control cells. The Balb/c-NKCC1 cells have acquired a typical transformation phenotype indicated by: (1) Loss of contact inhibition exhibited by growth to a higher cell density in confluent cultures, and formation of cell foci; (2) proliferation in low serum concentrations; and (3) formation of cell colonies in soft agar. The control cells transfected with the NKCC1 gene inserted in the opposite orientation in the vector retained their normal phenotype. Furthermore, the two specific inhibitors of the Na(+)/K(+)/Cl(-) cotransporter activity; bumetanide and furosemide inhibited the clonogenic efficiency in the NKCC1 transfected cells. These control experiments indicate that the apparent transformation phenotype acquired by the Balb/c-NKCC1 cells was not merely associated with the process of transfection and selecting for the neomycin-resistant clones, but rather with the overexpression of the Na(+)/K(+)/Cl(-) cotransporter gene. In order to ascertain that the regulated and normal expression of the Na(+)/K(+)/Cl(-) cotransporter control cell proliferation, the effect of bumetanide a specific inhibitor of the cotransporter, was tested on Balb/c 3T3 cell proliferation, induced by fibroblasts growth factor (FGF) and fetal calf serum (FCS). Bumetanide inhibited synchronized Balb/c 3T3 cell exit from the G(0)/G(1) arrest and entering S-phase. The inhibition was reversible, as removal of bumetanide completely released cell proliferation. Taken together, these results propose that the NKCC1 gene is involved in the control of normal cell proliferation, while its overexpression results in apparent cell transformation, in a manner similar to some protooncogenes.

Loop diuretics inhibit detubulation and vacuolation in amphibian muscle fibres exposed to osmotic shock

The effect of loop diuretics at concentrations known to influence cellular water entry coupled to Na-K-Cl co-transport, upon the vacuolation and detubulation following osmotic shock, was investigated in amphibian skeletal muscles. These were exposed to a glycerol-Ringer solution (18 min), an isotonic Ca2+/Mg2+ Ringer solution and cooling. Adding bumetanide (1.0 and 2.0 microM) to these solutions sharply reduced the incidence of detubulation, assessed by abolition or otherwise of action potential after-depolarisations, from 93.9 +/- 4.7% (n = 6) to 5.0 +/- 1.1% (n = 4: mean +/- SEM: 2.0 microM bumetanide). It dramatically reduced the number and fraction of muscle volume occupied by tubular vacuoles, measured using confocal microscopy, from 60.3 +/- 4.3% (n = 10) to 9.0 +/- 1.1% (n = 35). The incidence of large horseradish peroxidase-lined tubular vacuoles, viewed using electronmicroscopy, similarly was reduced with 2 microM bumetanide in the glycerol-Ringer solution. Bumetanide acted through cellular volume adjustments early in the detubulation protocol. Thus, it exerted its maximum effect when added to the glycerol-Ringer, rather than the Ca2+/Mg2+ Ringer solution. Furthermore, whereas fibre diameters measured using scanning electron microscopy returned to normal during glycerol treatment relative to those of control fibres left in isotonic Ringer, addition of 2.0 microM bumetanide in the glycerol Ringer left markedly smaller fibre diameters. Finally equipotent concentrations of the chemically distinct loop diuretics. furosemide and ethacrynic acid similarly influenced detubulation. These findings implicate Na-K-Cl co-transport in the water entry into muscle fibres that would be expected following introduction of extracellular glycerol. This might then enable the subsequent Na-K-ATPase dependent water extrusion that produces the tubular distension (vacuolation) and detachment (detubulation) following glycerol withdrawal, phenomena also observed in muscular dystrophy.

Alpha1-adrenergic activation of myocardial Na-K-2Cl cotransport involving mitogen-activated protein kinase

The translocation mechanisms involved in the alpha1-adrenoceptor-stimulated efflux of the potassium analog 86Rb+ were studied in isolated rat hearts. Phenylephrine (in the presence of a beta-blocker) increased the efflux of 86Rb+ and 42K+, and the Na-K-2Cl (or K-Cl) cotransport inhibitor bumetanide reduced the response by 42 +/- 11%. Furosemide inhibited the response with a lower potency than that of bumetanide. The bumetanide-insensitive efflux was largely sensitive to the K+ channel inhibitor 4-aminopyridine. Inhibitors of the Na+/H+ exchanger or the Na+-K+ pump had no effect on the increased 86Rb+ efflux. The activation of the Na-K-2Cl cotransporter was dependent on the extracellular signal-regulated kinase (ERK) subgroup of the mitogen-activated protein (MAP) kinase family. Phenylephrine stimulation increased ERK activity 3.4-fold. PD-98059, an inhibitor of the ERK cascade, reduced both the increased 86Rb+ efflux and ERK activity. Specific inhibitors of protein kinase C and Ca2+/calmodulin-dependent kinase II had no effect. In conclusion, alpha1-adrenoceptor stimulation increases 86Rb+ efflux from the rat heart via K+ channels and a Na-K-2Cl cotransporter. Activation of the Na-K-2Cl cotransporter is apparently dependent on the MAP kinase pathway.

Improved hypothermic preservation of rat hearts by furosemide

The effect of furosemide, a blocker of the Na+/K+/Cl- cotransporter, on hypothermic preservation of rat hearts was studied with use of the Langendorff perfusion system and electron microscopy. Furosemide significantly improved the mechanical recovery and the coronary flow of the hearts preserved for 8 hours in St. Thomas' Hospital cardioplegic solution at a temperature of 4 degrees C. Furosemide at the concentration of 100 mumol/L was found to have an optimal effect, whereas at high concentrations (1000 mumol/L) it was found to have toxic effects. In addition, furosemide reduces the time elapsed between the end of the preservation time and the resumption of myocardial contractions. Ultrastructural evaluations were done in which the presence of swollen mitochondria was chosen as a criterion of hypothermic ischemic damage to the myocardium. Morphometric analysis indicated that the mitochondrial volume of hearts stored for 8 hours in St. Thomas' Hospital cardioplegic solution increased by 72% as compared with the mitochondrial volume of hearts that were not exposed to the hypothermic ischemic conditions (control group). The addition of 100 mumol/L furosemide to the cardioplegic solution resulted in a significant reduction of mitochondrial swelling during the period of 8 hours' storage, which amounted only to 28% as compared with the figure for the control group. The reduction of mitochondrial swelling by furosemide and the improved mechanical and coronary flow recoveries are thought to be related to the blocking of the sarcolemmal Na+/K+/Cl- cotransporter and consequently the reduction of the Na+ influx during hypothermic ischemic storage.

The effect of furosemide on calcium ion concentration in myocardial cells

The effect of furosemide and ouabain on the intracellular concentration of Ca2+ was studies in myocardial cell cultures using Fura-2, a fluorescent agent, as an intracellular Ca2+ indicator. Introduction of 200 microM ouabain to the cultured cells increased the intracellular calcium concentration from an average of 236 nM up to an average of 833 nM. Introduction of 100 microM furosemide, prior to the administration of ouabain, decreased the ouabain induced Ca2+ elevation to an average of only 473 nM. Introduction of 2.5 mM EGTA prior to the administration of ouabain abolished the ouabain induced Ca2+ increase.

Ca(2+)-dependent heat production by rat skeletal muscle in hypertonic media depends on Na(+)-Cl- co-transport stimulation

1. The rate of energy dissipation (E) in isolated, superfused soleus muscles from young rats was continuously measured under normosmotic and 100-mosM hyperosmotic conditions. The substantial increase of E with respect to basal level in hyperosmolarity (excess E), which is entirely dependent on the presence of extracellular sodium, was largely prevented or inhibited by bumetanide, a potent inhibitor of Na(+)-Cl- co-transport system, or by the removal of chloride from the superfusate (isethionate substitution). Bumetanide or the removal of chloride also acutely decreased basal E, by about 7%. 2. Bumetanide almost entirely suppressed the major, Ca(2+)-dependent part of excess E in hyperosmolarity, as well as the concomitant increase of 45Ca2+ efflux and small increase in resting muscle tension; in contrast, the part of excess E associated with stimulation of Na(+)-H+ exchange in hyperosmolarity was left unmodified. 3. Reduction of 22Na+ influx by bumetanide was more marked in hyperosmolarity than under control conditions, although stimulation of total 22Na+ influx by a 100-mosM stress was not statistically significant. Inhibition of Ca2+ release into the sarcoplasm using dantrolene sodium did not prevent the stimulation of bumetanide-sensitive 22Na+ influx, but rather increased it about fourfold. 4. It is concluded that the largest part of excess E in hyperosmolarity, which is Ca(2+)-dependent energy expenditure, is suppressed when steady-state stimulation of a Na(+)-Cl- co-transport system is inhibited either directly by bumetanide or the removal of extracellular chloride, or indirectly by the blocking of active Na(+)-K+ transport. How the stimulation of Na(+)-Cl- co-transport, by as little as 1 nmol s-1 (g wet muscle weight)-1 during a 100-mosM stress, enhances Ca(2+)-dependent heat by as much as 2.5 mW (g wet muscle weight)-1 remains to be clarified.

In vivo assessment of the inotropic and toxic effects of oxidized ouabain

Oxidized ouabain, a product of the oxidative cleavage of the rhamnose ring in ouabain has been found to have a higher inotropic toxic ratio in cultured cardiac myocytes. The purpose of our study was to evaluate the efficacy and toxicity of oxidized ouabain in comparison with ouabain in intact animals. Drugs were infused to healthy cats; the positive inotropic effect, and the time-course of development of arrhythmia were followed and recorded until death. Oxidized ouabain was associated with a higher increase in arterial blood pressure, a mean increase of 41 +/- 19% as compared with 21 +/- 8% in the ouabain group (p < 0.10). There were no significant differences in maximal increases of dP/dt or dP/dt/P (65 +/- 29%, 28 +/- 10% for oxidized ouabain and 49 +/- 16%, 27 +/- 11% for ouabain, respectively). The mean doses causing persistent arrhythmia (toxic dose) were 93 +/- 23 micrograms/kg of oxidized ouabain vs 39 +/- 14 micrograms/kg of ouabain. Lethal arrhythmias were produced by 215 +/- 46 micrograms/kg of oxidized ouabain and 62 +/- 16 micrograms/kg of ouabain. The ratio of toxic to lethal doses was 0.62 +/- 0.11 for ouabain vs 0.45 +/- 0.09 for oxidized ouabain (p < 0.05), but the inotropic to toxic dose ratios were not different. We conclude that oxidized ouabain acts similarly to the known cardiac glycosides in doses which produce inotropic effects in cats, has a lower potency as compared to ouabain, and appears to have a more benign course of intoxication.

The regulation of the intracellular sodium ion concentration in cultured chick embryo heart cells

Using digital imaging microscopy with the fluorescent indicator sodium-binding benzofuran isophtalate, we examined the cytosolic Na+ concentration ([Na+]i) in individual chick embryo heart cells. Inhibition of the Na(+)-H+ exchanger using Na(+)-free (Li+ substituted) medium and inhibition of the Na(+)-efflux through the Na(+)-Ca2+ exchanger using Ca(2+)-free medium didn't change the [Na+]i. The opening of voltage-dependent Na+ channels with veratridine (150 micrograms/ml) and inhibition of the Na(+)-K(+)-Cl(-)-cotransporter with bumetanide (10 microM) led to an increase in [Na+]i by 107% and 86%, respectively, suggesting that the Na+ channels and the Na(+)-K(+)-Cl- cotransporter predominantly regulate the [Na+]i in cultured chick embryo heart cells.

Calcium-induced net potassium uptake of pig hearts in vivo

To determine whether the catecholamine-induced myocardial potassium uptake could be mimicked by increasing extracellular and intracellular calcium concentrations in vivo, we measured changes in myocardial potassium balance in nine anaesthetized open-chest pigs with PVC-valinomycin electrodes in arterial and coronary sinus blood. CaCl2 infusion (200-400 mumol min-1) into the left coronary artery increased coronary sinus blood calcium concentration from 2.29 (2.19-2.42) to 4.63 (3.76-5.67) mmol l-1 (median, 95% confidence interval, P = 0.01) indicating a similar increment in myocardial extracellular calcium concentration. The contractility measure LV dP/dt increased 95 (76-147) %, indicating a substantial increment in intracellular calcium concentration. During the CaCl2 infusion coronary sinus potassium concentration declined to a nadir 0.12 (0.09-0.17) mmol l-1 below baseline (P = 0.008) whereas arterial concentration remained unchanged. Peak myocardial potassium uptake was 18 (7-32) mumol min-1 100 g-1 and occurred 150 (110-195) s after start of infusion. The response remained unaltered after adrenoceptor blockade by prazosin and propranolol. Prolonged CaCl2 infusion caused a net myocardial potassium loss which was accompanied by metabolic and haemodynamic indications of myocardial ischaemia. These findings are consistent with enhanced Na-K pump activity in the intact beating pig heart in response to increased extracellular and intracellular calcium concentrations.

Binding properties and biological effects of oxidized-ouabain on cultured neonatal-rat cardiac myocytes. Implications on the mechanism of action of the digitalis-glycosides

Mild oxidation of ouabain with NaIO4, causes the cleavage of the bond between C2' and C3' of the rhamnose ring, leaving the steroid moiety intact. The oxidized ouabain (ox-ouabain) was examined on spontaneously contracting cultured rat-cardiac myocytes. Two classes of binding sites, with high and low affinities, were detected for both ox-ouabain and unmodified ouabain. The dissociation constants (KD) were found to be similar for both compounds, but the rate constants of association (ka) and dissociation (kd) of the low affinity sites were higher for ox-ouabain as compared with ouabain. Displacement experiments showed that ox-ouabain and ouabain bind to the same sites. The effects of ox-ouabain and ouabain on the activity of Na+, K(+)-ATPase were determined in microsomal preparations. Similar dose-response curves for the inhibition of the enzyme activity were determined for both drugs. Inhibition was observed only at concentrations above 10(-6) M. The biological effects of the drugs were examined by their capacity to induce positive inotropic or toxic effects. Concentrations of ox-ouabain which induced positive inotropic effects (increase in amplitude of systolic cell motion), ranged from 5 x 10(-8) M to 5 x 10(-6) M, as compared with 10(-7) M to 5 x 10(-7) M with ouabain. "Toxic" effects (decrease in the amplitude of systolic motion, increased beating frequencies and elevation in the position of maximal relaxation) was observed only with 10(-5) M ox-ouabain as compared with 10(-6) M ouabain. The mechanism of the inotropic action of ox-ouabain at the lower concentration range was investigated by measuring the effect of the drugs on 86Rb+ (analogue of K+) influx. Dose-response curves of effects of ouabain and ox-ouabain on 86Rb+ influx were bi-phasic. At low concentrations stimulation was observed, whereas at high concentrations 86Rb+ influx was inhibited. Ox-ouabain stimulated 86Rb+ influx by lower concentrations and to a greater extent than ouabain. A part of 86Rb+ influx into cardiac myocytes is mediated by the K+/Na+/Cl- cotransporter, which can be inhibited by loop diuretic drugs such as bumetanide. We have previously shown that ouabain, at low concentrations, stimulates the activity of the cotransporter. It is shown in the present work that ox-ouabain stimulates the activity of the cotransporter by lower concentrations and to a greater extent than ouabain.(ABSTRACT TRUNCATED AT 400 WORDS)

Effects of aluminium on electrical and mechanical properties of frog atrial muscle

1. The effects of aluminium on membrane ionic currents were studied in single cardiac myocytes. Most of the work was done on frog atrial cells, but some experiments were also carried out on single cells isolated from rabbit ventricles and atria. 2. The effects of aluminium on the force of contraction of frog atrial trabeculae were also investigated. 3. Aluminium was prepared from AlCl3 as a stock 0.5 M solution which has a pH of 3.5. Before each experiment, this solution was added to the control solution, to give a final concentration of 20-100 micrograms ml-1 aluminium (0.75-3.75 mM AlCl3). The solutions were brought to a pH of 7.4 or 7.6. at which they consist of a mixture of amorphous aluminium hydroxides and a very small amount of soluble ionic aluminium complexes: free aluminium cations (less than 10 pM), aluminohydroxide anions (less than 8 microM). The addition of this suspension reduced the peak inward calcium currents in single rabbit atrial and ventricular cells and in frog atrial cells. In the latter, the peak current was reduced (at + 10 mV) to 45% of control (mean of 9 cells). This effect was reversible upon washout, and was obtained at all membrane potentials, with no shift of the calcium current voltage relationship along the voltage axis. 4. Aluminium also reduced the time-dependent potassium current IK. This reduction was observed at all membrane potentials. For example, at + 10 mV, the mean reduction of IK (n = 9) was to 69% of the control amplitude. This effect, which was very difficult to reverse, was not due to IK rundown. The fully activated current-voltage relationships (obtained by standard 'tail' analysis) showed that the effect of aluminium was due mainly to a decrease in conductance and not to a shift in the activation range of IK. The mean voltage of half activation was shifted by 8 mV in the depolarizing direction (n = 5). 5. The background potassium current IK1 was also slightly but consistently changed in a complex fashion, with an outward shift at membrane potentials positive to -60 mV. For example, at a membrane potential of -40mV, the mean shift was by 22 + 4pA. At more negative potentials, there was an inward shift in the current amplitudes. For example, for steps to -I00 mV the current elicited was larger (more inward) by 53 pA (mean value, n = 10). The reversal potential was slightly shifted (<10 mV) in the hyperpolarizing direction. 6. The force of contraction of frog atrial trabeculae was altered by aluminium in a complex manner, which showed marked seasonal variation. During most of the year, 50-100,ug ml-1 aluminium caused a biphasic change, with an early small and consistent decrease, followed by a large increase in twitch amplitude. For a short period corresponding to the (local) winter months the sensitivity to aluminium was greatly enhanced. Aluminium lOOupgml-1 totally abolished contraction (n = 5), while a lower concentration (20,ug ml- 1) produced a sustained reduction in the force of contraction. Similar biphasic and seasonal responses have been reported to be induced by lanthanum. 7. The biphasic changes in twitch amplitude were independent of the transmembrane sodium gradient. Aluminium produced the same effects when 90% of the extracellular sodium was replaced by lithium. Caffeine (5 mM) attenuated or even inverted the positive inotropic effect of aluminium. These results imply that aluminium alters the release of calcium from intracellular, caffeine-sensitive stores. This could be effected either by augmenting the amount released during each activation, and/or by increasing the loading of stores prior to release.