INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE RUMEN IS THE MOST important site of Mg²⁺ absorption in the gastrointestinal tract of sheep and maintains Mg²⁺ homeostasis (35). Net Mg²⁺ absorption across the rumen epithelium occurs against an electrochemical gradient (3, 4, 22) and is mediated by an active transport mechanism (24).

It has long been known that an increase in K⁺ intake and, consequently, in ruminal K⁺ concentration ([K⁺]) decreases the apparent digestibility of Mg²⁺ (9, 14, 30). This effect of K⁺ is restricted to the forestomachs (35). The underlying mechanisms of this impaired Mg²⁺ absorption from the rumen have been studied intensively by in vivo and in vitro methods (4, 21-23), and there is now a growing body of evidence that the active and transcellular component of net Mg²⁺ uptake is significantly decreased by high ruminal [K⁺] (23). Furthermore, the reduced net Mg²⁺ transport has been shown to be closely correlated with electrophysiological changes of the rumen epithelium at high ruminal [K⁺]. There is a positive correlation between the [K⁺] of the ruminal fluid and the transepithelial potential difference (PD_t) (8, 18, 32). This K⁺-dependent increase of PD_t (blood-side positive) causes a small and passive backflow of Mg²⁺, probably via the paracellular pathway from the blood side into the lumen (4, 21, 23). Studies with microelectrodes have demonstrated that an increase in the mucosal [K⁺] leads to a reversible concentration-dependent depolarization of the potential difference across the apical membrane of the rumen epithelium (PD_a); this depolarization is accompanied by a decrease in the unidirectional mucosal-to-serosal transcellular Mg²⁺ flux (J^Mg_ms) (21). The K⁺-induced changes of Mg²⁺ fluxes, that is, a slight increase in the serosal-to-mucosal Mg²⁺ flux (J^Mg_sm) and a marked reduction in the J^Mg_ms, can be perfectly simulated by corresponding alterations of PD_t by a voltage clamp (23).

These observations have led to a tentative model of luminal uptake of Mg²⁺ in the rumen epithelium (21, 23), suggesting that Mg²⁺ enters the apical membrane by a PD-dependent transport mechanism, or, because K⁺ influences PD_a, by a K⁺-sensitive transport mechanism via a conductance or a carrier. The predominant driving force for this Mg²⁺ uptake is probably PD_a and, to a small extent, the chemical gradient between the luminal and cytosolic Mg²⁺ concentrations ([Mg²⁺]). Consequently, an ideal experiment for establishing this model of Mg²⁺ transport should include the simultaneous measurement of PD_t, PD_a, intracellular [Mg²⁺] ([Mg²⁺]_i), and Mg²⁺ fluxes. Unfortunately, it is not possible to measure Mg²⁺ transport and the other parameters directly and simultaneously with the squamous, stratified, and keratinized rumen epithelium. However, an indirect approach could give us the required information regarding Mg²⁺ transport and PD_a. Any change of PD_t, which can be easily manipulated in vitro in conventional Ussing chambers, induces alterations of PD_a and/or the potential difference of the basolateral membrane (PD_b). Hence, we have measured unidirectional ²⁸Mg²⁺ fluxes (J^Mg_sm and J^Mg_ms) at different PD_t values (blood-side positive) and correlated the flux rates with PD_a as determined by mucosal impalement with microelectrodes. To test the hypothesis of a PD-dependent Mg²⁺ uptake in a more direct way, we have performed experiments with cultured ruminal epithelial cells (REC). With the aid of the fluorescence probe mag-fura 2, we have measured the free [Mg²⁺]_i of isolated REC under basal conditions and after alterations of the transmembrane voltage (E_m). The obtained results clearly show PD-dependent changes of Mg²⁺ transport across the sheep rumen epithelium and of the [Mg²⁺]_i and support the assumption of electrodiffusive Mg²⁺ uptake across the luminal membrane.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. Medium 199, trypsin, glutamine, antibiotics (gentamicin, nystatin, and kanamycin), and FCS were purchased from Sigma (St. Louis, MO). Dulbecco's phosphate-buffered saline (DPBS) and collagen were obtained from Biochrom (Berlin, Germany). Mag-fura 2-AM, fura 2-AM, 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein-AM, and pluronic acid were from Molecular Probes (Eugene, OR). All other chemicals were purchased from Sigma.

Epithelia. Epithelia were taken from the ventral rumen sac of freshly slaughtered sheep. Pieces of rumen wall were immediately immersed in the control buffer solution (maintained in 38°C, gassed with 95% O₂-5% CO₂), and the epithelia were stripped of the underlying muscle layers and the serosa.

Incubation. Pieces of epithelia were mounted between the two halves of the incubation chamber with an exposed area of 3.14 cm² (for flux studies) or 0.79 cm² (for microelectrode studies). We minimized edge damage by using rings of silicone rubber on both sides of the tissues. Flux chambers were connected to reservoirs containing 16 ml buffer solution on each side. The solutions (standard Ringer solution that contained in mM: 95 NaCl, 5 KCl, 25 NaHCO₃, 1 Na₂HPO₄, 2 NaH₂PO₄, 10 glucose, 1 CaCl₂, 2 MgCl₂, 9 sodium-acetate, 9 sodium-propionate, 9 butyric acid, and 5 NaOH) were kept at 38°C and were continuously stirred by the use of a gas lift system that supplied 95% O₂-5% CO₂. The microelectrode chamber had a volume of 0.7 ml on the mucosal side and 0.4 ml on the serosal side. Both chamber halves were perfused with solution from gassed reservoirs, and the solutions were heated to 38°C immediately before being added to the chamber.

Electrical measurements. Chambers were connected to a computer-controlled voltage-clamp device (AC Microclamp, Aachen, Germany) or to a voltage-clamp and microelectrode device (Biomedical Instruments, Munich, Germany). PD_t was measured via KCl (0.5 M) agar bridges and calomel electrodes with reference to the mucosal solution. Tissue conductance (G_t) was determined from the change in PD_t caused by bipolar current pulses of 100 µA of 100-ms duration. The current was passed either through Ringer-agar bridges connected to Ag-AgCl electrodes in 3 M KCl (flux chambers) or through rings of Ag-AgCl electrodes (placed in each half of the microelectrode chamber). In each setup, fluid resistance and junction potentials were measured before mounting the epithelia and were corrected for during the experiments.

Microelectrodes. Conventional microelectrodes were pulled from borosilicate glass (1.2 or 1.5 mm OD) and filled with 0.5 M KCl, yielding resistances of 15-30 M. We impaled rumen epithelial cells across the apical membrane by means of a motorized micromanipulator with a Piezzo element and measured PD_a with reference to the mucosal solution. PD_a and PD_t were observed on an oscilloscope and displayed on a chart recorder. Impalements were accepted if 1) the change in PD_a was abrupt while advancing into the tissue, 2) PD_a remained stable for at least 1 min, and 3) PD_a returned to 0 ± 3 mV on withdrawal of the electrode.

Flux studies. ²⁸MgCl₂ (Department of Physics, Munich-Garching) was added to one side of the epithelium to yield a specific activity of 2 kBq/µM in the buffer solution. After an equilibrium period of 45 min, aliquots were sampled at 30-min intervals and replaced by aliquots of unlabeled solution. ²⁸Mg²⁺ was counted in a well-type crystal scintillation counter (Gammaszint 5300, Berthold). The counts were corrected for physical decay. Unidirectional fluxes were calculated from the rate of tracer appearance on the other side of the tissue. Paired determinations of fluxes were accepted if G_t differed by <25%.

Whole cell patch-clamp experiments. To determine E_m of REC and to verify the effect of the manipulations that changed it, we used standard electrophysiological methods. REC were routinely grown as described under cell culture. Before the patch-clamp experiments, cells were seeded onto uncoated glass coverslips for 1 day and then transferred to a chamber mounted on the stage of an inverted microscope (Axiovert, Zeiss) for recording. Resting membrane potential and cell currents were recorded from single cells by using the conventional patch-clamp technique in the whole cell mode under current clamp (I = 0) or voltage clamp (V = 0) with an EPC-7 patch-clamp amplifier (HEKA Elektronik). Computer-controlled voltage-clamp protocols were used to generate current-voltage relationships (winTIDA 2.1, HEKA Elektronik). After low-pass filtering (0.8 kHz, four-pole Bessel), data were sampled at 2 kHz and stored on hard disc. The standard bath solution was the NaCl solution (in mM: 140 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, and 10 HEPES). When the K⁺ content of the bath solution was increased, an appropriate amount of NaCl was replaced by KCl. Pipettes were filled with K⁺ solution (in mM: 110 potassium-gluconate, 25 KCl, 10 NaCl, 1 EGTA, 2 MgCl₂, 2 ATP, and 10 HEPES) in all experiments.

Cell culture. Primary cultures of REC were prepared as described by Galfi et al. (12). Briefly, REC were isolated by fractional trypsination and grown in medium 199 containing 10% FCS, 1.36 mM glutamine, 20 mM HEPES, and antibiotics (gentamicin 50 mg/l and kanamycin 100 mg/l) under an atmosphere of humidified air-5% CO₂ at 38°C. Experiments were performed between 6 and 12 days after seeding.

Measurement of cytoplasmic Mg²⁺ and Ca²⁺ by spectrofluorometry. Cells were loaded with either 5 µM mag-fura 2-AM or 10 µM fura 2-AM for the determination of [Mg²⁺]_i and intracellular Ca²⁺ concentration ([Ca²⁺]_i), respectively. Cells were subsequently washed twice in DPBS. The REC were incubated for a further 30 min to allow for complete deesterfication and washed twice before measurement of fluorescence. Intracellular ion concentrations were determined by measuring the fluorescence of the probe-loaded REC in the spectrofluorometer LS-50 B (Perkin-Elmer) by using the fast filter accessory, which allowed the measurement of fluorescence at 20-ms intervals with excitation for mag-fura 2 and fura 2 at 340 and 380 nm and emission at 515 nm. All measurements were made at 37°C in a 3-ml cuvette containing 2 ml cell suspension under stirring. The [Mg²⁺]_i and [Ca²⁺]_i were calculated from the 340/380-nm ratio according to the formula of Grynkiewicz et al. (16), with a dissociation constant of 1.5 mM and 224 nM, respectively, for the mag-fura-2/Mg²⁺ and fura-2/Ca²⁺ complexes. The minimum (R_min) and maximum (R_max) ratios were determined at the end of each experiment by using digitonin. R_max for mag-fura 2 was obtained by the addition of 25 mM MgCl₂ in the absence of Ca²⁺ and R_min by addition of 50 mM EDTA, pH 7.2, to remove all Mg²⁺ from the solution. R_max for fura 2 was obtained in solutions with 2 mM Ca²⁺ and R_min by addition of 20 mM EGTA, pH 8.0.

Statistical analysis. If not otherwise stated, data are presented as means ± SD. Significance was determined by Student's t-test or Tukey's ANOVA. Correlations between variables were tested by calculating Pearson's product moment correlation coefficients. P < 0.05 was considered significant. All statistical calculations were performed using SigmaStat (Jandel Scientific).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PD_t and unidirectional ²⁸Mg²⁺ fluxes. The unidirectional flux rates of ²⁸Mg²⁺ are shown in Fig. 1. Increasing the PD_t (blood-side positive) induced a decrease in the unidirectional J^Mg_ms from 65.9 ± 13.8 nmol · cm² · h¹ under short-circuit conditions (PD_t = 0 mV) to 48.5 ± 8.4 (PD_t = 15 mV), 40.6 ± 7.8 (PD_t = 30 mV), and 31.5 ± 6.1 nmol · cm² · h¹ (PD_t = 45 mV). The ²⁸Mg²⁺ flux in the opposite direction (J^Mg_sm) was slightly enhanced from 6.9 ± 1.2 nmol · cm² · h¹ (PD_t = 0 mV) to 9.6 ± 2.2 (PD_t = 15 mV), 14.1 ± 2.8 (PD_t = 30 mV), and 17.5 ± 1.9 nmol · cm² · h¹ (PD_t = 45 mV). These changes in the unidirectional ²⁸Mg²⁺ fluxes caused a significant decrease in net ²⁸Mg²⁺ absorption (J^Mg_net) from 58.9 ± 12.8 nmol · cm² · h¹ to 38.9 ± 6.3, 26.5 ± 5.1, and 14.0 ± 4.5 nmol · cm² · h¹ at 0, 15, 30, or 45 mV, respectively.

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Unidirectional 28Mg2+ fluxes (JMg2+) at different transepithelial potential difference (PDt) values (serosal-side positive) determined in conventional Ussing chambers under voltage-clamp conditions. Mucosal and serosal concentrations of Mg2+ were identical (2 mM). JMgms, mucosal-to-serosal Mg2+ flux; JMgsm, serosal-to-mucosal Mg2+ flux. Values are means ± SD; n = 9-11.

PD_t and PD_a. Because simultaneous determinations of ²⁸Mg²⁺ fluxes and PD_a were not possible, we tested the effect of applied changes in PD_t on PD_a in accompanying microelectrode studies. During cell impalements, epithelia were voltage clamped to various PD_t values between 0 mV and 80 mV, with the mucosal side as the reference. The applied PD_t included the physiological range of 20-60 mV (blood-side positive; Ref. 7). Under short-circuit conditions (PD_t = 0 mV), PD_a was 54 ± 5.5 mV. Increasing PD_t from 0 mV to 60 mV induced a reversible depolarization of PD_a to 10 ± 7 mV (Fig. 2). The relationship between PD_t and PD_a was linear: PD^a = 0.72 PD_t  53.4 mV (r² = 0.9; n = 13, P < 0.01).

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Potential difference of apical membrane (PDa) as a function of PDt (serosal-side positive) Regression equation is given in PDt and PDa. Values are means ± SD; n = 15.

View larger version (10K): [in this window] [in a new window]   Fig. 3.   JMgms measured at defined PDt values (0-45 mV, serosal-side positive; see Fig. 1) as a function of PDa. PDa was calculated by means of regression equation (PDa = 0.72 PDt  53.4) determined from preceding microelectrode studies (see Fig. 2). JMgms is related to PDa in an exponential fashion: JMgms = 18.7 + 6.08 · e(0.038 · PDa). Line is fitted to data by nonlinear regression analysis. There is a strong negative correlation between PDa and JMgms; correlation coefficient = 0.99. Values are means ± SD; n = 9-11 for each equation.

Free [Mg²⁺]_i. The determination of ²⁸Mg²⁺ fluxes of the whole epithelium clearly revealed that variations of PD_a (and PD_t) influenced transcellular Mg²⁺ transport. It was hypothesized that depolarization of PD_a reduces Mg²⁺ uptake, which should cause corresponding changes in [Mg²⁺]_i. Because measurements of [Mg²⁺]_i have not been made previously with isolated REC, the first series of experiments was designed to study [Mg²⁺]_i as a function of the extracellular [Mg²⁺] ([Mg²⁺]_e) without manipulating the membrane potential. The basal cytosolic free [Mg²⁺]_i of REC (cultured in the presence of Ca²⁺ and Mg²⁺ as described in MATERIALS AND METHODS) was 0.54 ± 0.08 mM (n = 12), measured in Ca²⁺-free and Mg²⁺-free Hanks' balanced salt solution. REC showed a rapid concentration-dependent increase in [Mg²⁺]_i when [Mg²⁺]_e was increased (Fig. 4). Figure 5 shows the effect of varying the bath [Mg²⁺] on the cytosolic [Mg²⁺] of REC with a physiological Mg²⁺ content. [Mg²⁺]_i of REC increased from 0.59 ± 0.13 mM (n = 3) in media with 0.5 mM Mg²⁺ to a maximum level of 1.26 ± 0.1 mM (n = 3) in solutions with a [Mg²⁺]_e of 7.5 mM. The changes in [Mg²⁺]_i showed sigmoid saturation kinetics with a half-maximal saturation point of 1.2 mM.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Original tracings of measurements of intracellular Mg2+ concentration ([Mg2+]i) () and intracellular pH (pHi) () in ruminal epithelial cells (REC). Extracellular [Mg2+] ([Mg2+]e) was increased stepwise from 0 mM to a final [Mg2+] of 1 mM; 0.5 mM MgCl2 was added at time indicated by arrows. [Mg2+]i and pHi were measured simultaneously in mag-fura 2- and 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF)-loaded REC. BCECF signals were calibrated to pHi values with 10 µM nigericin in a high-K+ medium between pH 6.0 and 8.0. The pHi of REC was 7.1 ± 0.3 in time course of experiments. Inset: original tracing of a measurement of intracellular Ca2+ concentration in REC; there are no changes in intracellular Ca2+ with increasing [Mg2+]e.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   [Mg2+]i as a function of [Mg2+]e. [Mg2+]e was stepwise increased from 0 mM to 7.5 mM. Extracellular Ca2+ concentration was 1 mM throughout experiment; n = 3-6. Line is drawn assuming sigmoidal kinetics (Hill Method) and is fitted to data by nonlinear regression analysis. Parameters used are: y0 = 0.54, a = 0.72, b = 2.84, c = 1.26, correlation coefficient = 0.997.

Effect of high extracellular K⁺ and of quinidine on the E_m of REC. The elevation of the extracellular [K⁺] ([K⁺]_e) is a simple and often-applied method for depolarizing the E_m by reducing K⁺ efflux via K⁺ channels. Whole cell patch-clamp experiments were carried out to verify the effects of increased [K⁺]_e on the E_m of isolated REC (Fig. 6). The REC resting potential determined by conventional patch-clamp techniques was 24 ± 3 mV (n = 40 single cells; maximum 70 mV, minimum 5 mV). An increase of [K⁺]_e from 5 mM in control solutions to either 40 or 80 mM resulted in a reversible depolarization of the REC E_m by ~15 and 32 mV, respectively (Fig. 6B). Like high K⁺, the application of the K⁺ channel blocker quinidine depolarized the E_m of REC by ~30 mV.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of high extracellular K+ concentration ([K+]e) on membrane potential (Em) of REC. As in fluorescence measurements, standard bath was NaCl solution. When K+ content of bath solution was increased, an appropriate amount of NaCl was replaced by KCl. Pipettes were filled with K+ solution in all experiments. A: original recording showing depolarization after changing bath from a control solution with 5 mM K+ to a solution with 80 mM K+. B: summary of effects of high [K+]e on transmembrane voltage (Em) of REC. Values are means ± SD; n = 40, 7, and 3 for [K+]e = 5, 40, and 80 mM, respectively. Inset: whole cell currents from a typical cell bathed in solutions with either 5, 40, or 80 mM K+.

Effect of high [K⁺]_e and of quinidine on [Mg²⁺]_i of REC. The effects of increasing [K⁺]_e on [Mg²⁺]_i are summarized in Fig. 7. In all experiments, high [K⁺]_e led to a reduction of [Mg²⁺]_i. As Fig. 7 shows, an increase in the [K⁺]_e from 5 mM to 40, 80, or 140 mM decreased [Mg²⁺]_i (P < 0.01) by 7 ± 2.7, 11 ± 2.6, and 42 ± 6.2%, respectively.

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Summary of effects of high [K+]e on [Mg2+]i of REC. Results are expressed as percent decrease from control [Mg2+]i ([Mg2+]i in media with a [K+]e of 5 mM = 100%). [K+]e was increased from 5 mM to 40, 80, or 140 mM. An increase in [K+]e decreased [Mg2+]i in all cases. Number of single experiments is shown in parentheses. ** P < 0.01 vs. control. Inset: original recordings of 2 measurements showing effect of increasing bath [K+] from 5 to 40 mM.

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Effect of K+ channel blocker quinidine on free [Mg2+]i of REC. Curves show results of single experiments (n = 6), and each point represents mean of repetitive measurements of [Mg2+]i under same conditions. Columns give mean values obtained without quinidine (control) and with increasing quinidine concentrations in bath.

Effect of membrane hyperpolarization on [Mg²⁺]_i. To test the effect of membrane hyperpolarization on [Mg²⁺]_i, we enhanced the K⁺ diffusion by application of 100 µM valinomycin, a K⁺ ionophore, in the presence of an outward K⁺ gradient {intracellular [K⁺] ([K⁺]_i) > [K⁺]_e}. With 2 mM Mg²⁺ in the extracellular solution, the resulting hyperpolarization of the REC E_m induced a rapid 15% increase in [Mg²⁺]_i (P < 0.01). Cells treated with valinomycin and equal [K⁺] on both membrane sides ([K⁺]_i = [K⁺]_e = 140 mM) served as controls. Without a K⁺ gradient, valinomycin application resulted in a slow, not significant, elevation of [Mg²⁺]_i by only 5% (Fig. 9).

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Effect of hyperpolarization of Em generated by outward K+ gradients ([K+]i = 140 mM > [K+]e = 5 mM) in presence of 100 µM valinomycin (left). Right: increase of [Mg2+]i after application of 100 µM valinomycin, but without a K+ gradient ([K+]i = [K+]e = 140 mM) across cell membrane. Number of single experiments are shown in parentheses.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

²⁸Mg²⁺ fluxes. A net absorption of ²⁸Mg²⁺ is observed in the absence of transepithelial electrochemical gradients, since J^Mg_ms is much greater than J^Mg_sm, which may indicate that the passive and probably paracellular permeability of the rumen epithelium is low for Mg²⁺. These data are in agreement with previously published results (21, 23). In addition, our findings confirm the long-known negative correlation between PD_t and net Mg²⁺ flux (J^Mg_net) (4, 21-23, 35).

PD_t/PD_a and J^Mg_ms. PD_t-dependent Mg²⁺ absorption has been previously reported in different parts of the kidney (1, 7, 33) and in the gastrointestinal tract (19) but was mainly explained as paracellular movement in these epithelia. In contrast, it has been shown by Leonhard-Marek and Martens (21) that the majority of PD_t-dependent J^Mg_ms in rumen epithelium is electrogenic transport through the cell. In our study the reduction of J^Mg_ms after an increase of PD_t is remarkably higher than the increase in J^Mg_sm, which implies that predominantly the cellularly mediated Mg²⁺ transport is affected.

Concentration dependence of [Mg²⁺]_i. [Mg²⁺]_i of REC was determined for the first time in this study. Typically, [Mg²⁺]_i is between 0.5 and 1.0 mM (5, 11, 13, 15, 17, 29, 34) and remains constant, even when the cells are incubated in Mg²⁺-rich media. In contrast, as our results show, REC exhibits another behavior. [Mg²⁺]_i rises from 0.59 ± 0.13 mM to 1.26 ± 0.1 mM as the [Mg²⁺]_e is stepwise increased from 0.5 mM to 7.5 mM. It must be taken into consideration that the experiments were performed with cell suspensions, where apical and serosal membranes are no longer separated by tight junctions, and therefore the whole membrane was exposed to the different [Mg²⁺]. In the physiological situation only the apical membrane is exposed to high [Mg²⁺] (up to 8 mM), whereas the serosal or blood [Mg²⁺] is lower and relatively constant (0.8-1.2 mM). The rise in [Mg²⁺]_i is obviously dependent on [Mg²⁺]_e and exhibits a saturable component. Leonhard-Marek et al. (20) demonstrated at the tissue level that the process of Mg²⁺ transport across the apical membrane of the rumen epithelium involves a PD- or K⁺-sensitive mechanism and a second, parallel working PD-independent mechanism, which could result from apical Mg²⁺/H⁺ exchange or from cotransport of Mg²⁺ with anions (20, 31). The latter may become saturated at [Mg²⁺] in the rumen liquid above 4 mM (4, 24). Even under basal conditions, these influx pathways seem to be active and may therefore reflect the Mg²⁺-absorbing ability of these cells.

Effect of E_m depolarization and hyperpolarization on [Mg²⁺]_i. The results of the patch-clamp studies confirm that high [K⁺]_e or application of quinidine, which has been shown to be an effective blocker of K⁺ channels in rumen epithelium (2), led to a depolarization of the E_m of isolated REC. These effects are consistent with the existence of a K⁺ conductance in the cell membrane of REC, as has been proposed from flux studies (2, 21). An increase of the bath [K⁺] or application of quinidine also decreased [Mg²⁺]_i in a dose-dependent manner, which implies an impairment of electrodiffusive Mg²⁺ uptake mechanisms by a reduction of the transmembrane electrical gradient. Quinidine has been demonstrated to inhibit both the Na⁺/H⁺ exchanger and the Na⁺/Mg²⁺ exchanger (31), but the resulting effects (acidification of the cytosol and reduction of Mg²⁺ efflux) should increase, not decrease, [Mg²⁺]_i.