Cerebellar granule cell GABA(A) receptors studied at the single-channel level: modulation by protein kinase G

Rat cerebellar granule cells GABA(A) receptors were studied at the single-channel level in outside-out patches. Three conductance levels were detected as activated by 0.1 microM GABA: 11, 20 and 30 pS. Single-channel I-V relationships were linear. The probability of opening did not vary over time within single patches. Kinetic analysis brought to a mean open time constant of 3.2, 2.9 and 2.8 ms respectively for each conductance level and a closed time histogram fitted by the sum of two exponential functions (tau c1 = 2.1 ms, 43%; tau c2 = 18.2 ms, 57%). Protein kinase G (PKG) activation did not affect single-channel conductances, but resulted in a reduction over time of single-channel open probability for all the conductance levels. Kinetically, protein kinase G modified the mean open time constants and the relative areas of the two components of the closed state distribution whereas the mean closed time constants remained unaffected. These results confirm and add details about cerebellar granule GABA(A) receptors down regulation by PKG.

Prion protein fragment 106-126 induces apoptotic cell death and impairment of L-type voltage-sensitive calcium channel activity in the GH3 cell line

The prion diseases are transmissible neurodegenerative pathologies characterized by the accumulation of altered forms of the prion protein (PrP), termed PrP(Sc), in the brain. Previous studies have shown that a synthetic peptide homologous to residues 106-126 of PrP (PrP 106-126) maintains many characteristics of PrP(Sc), i.e., the ability to form amyloid fibrils and to induce apoptosis in neurons. We have investigated the intracellular mechanisms involved in the cellular degeneration induced by PrP 106-126, using the GH3 cells as a model of excitable cells. When assayed in serum-deprived conditions (48 hr), PrP 106-126 (50 microM) induced cell death time-dependently, and this process showed the characteristics of the apoptosis. This effect was specific because a peptide with a scrambled sequence of PrP 106-126 was not effective. Then we performed microfluorimetric analysis of single cells to monitor intracellular calcium concentrations and showed that PrP 106-126 caused a complete blockade of the increase in the cytosolic calcium levels induced by K+ (40 mM) depolarization. Conversely, the scrambled peptide was ineffective. The L-type voltage-sensitive calcium channel blocker nicardipine (1 microM) also induced apoptosis in GH3 cells, suggesting that the blockade of Ca2+ entry through this class of calcium channels may cause GH3 apoptotic cell death. We thus analyzed, by means of electrophysiological studies, whether Prp 106-126 modulate L-type calcium channels activity and demonstrated that the apoptotic effect of PrP 106-126 was due to a dose-dependent inactivation of the L-type calcium channels. These data demonstrate that the prion protein fragment 106-126 induces a GH3 apoptotic cell death inducing a selective inhibition of the activity of the L-type voltage-sensitive calcium channels.

Involvement of phosphatase activities in the run-down of GABA(A) receptor function in rat cerebellar granule cells in culture

Run-down of GABA activated Cl- currents was found when rat cerebellar granule cells in culture were studied by the whole-cell patch-clamp technique in the absence of ATP in the pipette medium. This event could be prevented, even in the absence of ATP, by using the perforated-patch technique or by adding to the pipette medium either a blocker of protein tyrosine phosphatase, sodium vanadate, or deltamethrin, a blocker of the protein serine/threonine phosphatase calcineurin. Conversely, run-down could be partially induced, even in the presence of ATP, by blockers of tyrosine kinases. A reduction of GABA(A) receptor activity was also found in outside-out membrane patches when ATP was not on the membrane inside. The run-down phenomenon involved all three conductance levels found in these patches: 11, 20 and 30 pS. In all three cases it was due to a reduction of channels' open probability. The single-channel experiments showed that also in this case run-down was prevented by either sodium vanadate or deltamethrin on the membrane cytoplasmic side. Overall, through relatively unphysiological conditions (cells in culture and patch-clamp techniques), the study of the run-down phenomenon shows that the tyrosine phosphorylation state of GABA(A) receptors is of importance in maintaining it in a proper functional state. The data also show that tyrosine phosphorylation state is controlled by a protein tyrosine phosphatase, whose activity in turn is blocked via serine/threonine phosphorylation.

Effect of nitric oxide donors on GABA uptake by rat brain synaptosomes

The effect of nitric oxide donors and L-arginine on the uptake of GABA was studied in synaptosomes purified from rat brain. The neurotransmitter uptake was significantly reduced by S-nitrosoacetylpenicillamine and by sodium nitroprusside, although in this case to a lesser extent. A slight inhibitory effect was found preincubating rat brain synaptosomes with 1 mM L-arginine as well. The S-nitrosoacetylpenicillamine effect gradually disappeared with decomposition of the substance by exposure to light. The nitric oxide effect appears to be mainly due to a decrease in the V for synaptosomal GABA uptake and seems to be related to a partial collapse of nerve endings ionic gradients. Functionally, it could result over time in a reduced availability of GABA at the synapses involved.

A dual mechanism for impairment of GABAA receptor activity by NMDA receptor activation in rat cerebellum granule cells

The function of the GABAA receptor has been studied using the whole cell voltage clamp recording technique in rat cerebellum granule cells in culture. Activation of NMDA-type glutamate receptors causes a reduction in the effect of GABA. Full GABAA receptor activity was recovered after washing out NMDA and NMDA action was prevented in a Mg+2 containing medium. The NMDA effect was also absent when extracellular Ca+2 was replaced by Ba+2 and when 10 mM Bapta was present in the intracellular solution. Charge accumulations via voltage activated Ca+2 channels greater than the ones via NMDA receptors do not cause any reduction in GABAA receptor function, suggesting that Ca+2 influx through NMDA receptor channels is critical for the effect. The NMDA effect was reduced by including adenosine-5'-O-3-thiophosphate (ATP-gamma-S) in the internal solution and there was a reduction in the NMDA effect caused by deltamethrin, a calcineurin inhibitor. Part of the NMDA induced GABAA receptor impairment was prevented by prior treatment with L-arginine. Analogously, part of the NMDA effect was prevented by blockage of NO-synthase activity by N omega-nitro-L-arginine. A combination of NO-synthase and calcineurin inhibitors completely eliminated the NMDA action. An analogous result was obtained by combining the NO-synthase inhibitor with the addition of ATP-gamma-S to the pipette medium. The additivity of the prevention of the NMDA impairment of GABAA receptor by blocking the L-arginine/NO pathway and inhibiting calcineurin activity suggests an independent involvement of these two pathways in the interaction between NMDA and the GABAA receptor. On the one hand Ca+2 influx across NMDA channels activates calcineurin and dephosphorylates the GABAA receptor complex directly or dephosphorylates proteins critical for the function of the receptor. On the other hand, Ca+2 influx activates NO-synthase and induces nitric oxide production, which regulates such receptors via protein kinase G activity.

Nitric oxide and GABAA receptor function in the rat cerebral cortex and cerebellar granule cells

The aim of the present work was to investigate the mechanism by which the diffusible factor nitric oxide regulates GABAA receptor function in the brain. The effect of nitric oxide on GABAA receptor function has been studied in two different neuronal preparations: rat cerebral cortex microsacs and rat cerebellum granule cells in culture. In the first case, GABA-stimulated 36Cl-accumulation was studied as an index of GABAA receptor function. The maximal rate of GABA-stimulated 36Cl- accumulation (Vmax) was reduced by treatment of microsacs with nitric oxide chemical donors such as sodium nitroprusside (-26%) and S-nitroso-acetyl-penicillamine (-11%). The greater effect of the former agent is due to an additional interference by its breakdown products. The biochemical precursor L-arginine (1 mM) produced the same Vmax decrease as S-nitroso-acetyl-penicillamine. This effect was reversed by a nitric oxide synthase blocker and appears truly nitric oxide mediated. The action of nitric oxide in this system does not seem to imply cyclic GMP formation. GABAA receptor function was studied by whole-cell patch-clamp in rat cerebellum granule cells in culture. In this case, L-arginine (100 microM) profoundly reduced the Cl- current elicited by 10 microM GABA and its effect subsided following washing out. The effect of L-arginine was observed almost exclusively on the rapidly desensitizing component of the GABA-activated current. The action of L-arginine was blocked by a protein kinase G inhibitor and mimicked by its activators. Thus, it appears that this effect in these cells involves nitric oxide formation, cyclic GMP accumulation and protein kinase G-catalysed phosphorylation of GABAA receptor.

Valinomycin acts as a channel in ultrathin lipid membranes

When the thickness of monolayer membranes formed by bolaform archaeal lipids is reduced to the approximate length of two valinomycin molecules, the zero-current conductance does not show any more a linear dependence on valinomycin concentration; instead, a quadratic behaviour is observed. This suggests that a dimer permeation pore is formed and therefore the conduction mechanism changes from carrier to channel.

Modulation by nitric oxide of rat brain GABAA receptors

The effect of nitric oxide (NO) on the function of GABAA receptors was studied in two different rat brain neuron populations. Cerebral cortex neuronal GABAA receptors were studied by preparing microsacs and evaluating 36Cl- accumulation. Whether nitric oxide was provided by sodium nitroprusside (SNP) or by the metabolic precursor precursor arginine there was a 15-25% reduction in the Vmax for GABA-stimulated 36Cl- accumulation. The arginine effect could be reversed by the NO synthase (NOS) inhibitor N omega-nitro-L-arginine. GABAA receptor mediated Cl- currents were studied in rat cerebellar granule cells by whole-cell patch clamp. S-Nitroso-N-acetylpenicillamine (SNAP), sodium nitroprusside and L-arginine reduced the Cl- current elicited by 10 microM GABA. The L-arginine effect was reversible upon its washing out. This circumstance indicates that NO produced by endogenous NOS can inhibit GABAA receptor function in cerebellar granule cells.

Modulation by extracellular pH of the activity of GABAA receptors on rat cerebellum granule cells

The Cl- currents activated by GABA via GABAA receptors in rat cerebellum granule cells in culture were studied by whole-cell patch-clamp. These currents were measured at various extracellular pH. The currents activated by 100 microM GABA, both the peak and the steady-state component, increase at acidic pH's and decrease at basic pH's. The transition point being at around 7.7. Interestingly, passing from pH 7.4 to 6.4 the GABA dose-response curve indicates that the increases in the peak current are related to an augmented maximal current. The increases in the steady-state component are mainly due to a higher affinity of the receptors for the neurotransmitter and disappear at saturating [GABA]. The study of the I-V curves for the GABA activated peak Cl- currents at pH 6.4, 7.4 and 8.4 reveals linearity in the latter instance. However, an outward rectification is present at the two more acidic pH's. This fact suggests that the protonation of basic amino acids at the acid pH does involve rectification of Cl- channel conductance. Overall, the data indicate that slight changes in in situ extracellular pH may have profound influences on GABAA receptor function.

Inactivation of voltage-dependent calcium current in an insulinoma cell line

We have studied the mechanism of Ca current inactivation in the beta-cell line HIT-T15 by conventional and perforated patch recording techniques, using two pulse voltage protocols and a combination of current and tail current measurements. In 5 mM Ca, from a holding potential of -80 mV, the maximum current showed a complex time course of inactivation: a relatively fast, double exponential inactivation (tau h1 approximately 12 ms and tau h2 approximately 60 ms) and a very slowly inactivating component (tau > 1 s). The faster component (tau h1) was due to the voltage-dependent inactivation of a low-threshold-activated (LVA), T-type current, which deactivates more slowly (tau approximately 3-5 ms) than the other components (tau approximately 0.2-0.3 ms). The intermediate component (tau h2) was due to the Ca-dependent inactivation of a portion of the high-threshold-activated (HVA) current. A saturating dose of the dihydropyridine (DHP) nifedipine (10 microM) did not affect the LVA current, but inhibited by 68 +/- 5% the transient, Ca-sensitive portion of the HVA current and by 33 +/- 12% the long lasting component. We suggest that three components of the calcium current can be resolved in HIT cells and the main target of DHPs is a HVA current, which inactivates faster than the DHP-resistant HVA component and does so primarily through calcium influx.

Asymmetric black membranes formed by one monolayer of bipolar lipids at the air/water interface

In this work a new technique is presented for the formation of black lipid membranes from a single monolayer of bipolar lipids at the air/water interface. The lipid, extracted from the thermophilic archaeobacterium Sulfolobus solfataricus, is characterized by two different polar heads. The membrane is formed with a technique similar to that introduced by Montal and Mueller; however, the lipid is spread only on one side of the teflon partition. Conductance in the presence of valinomycin, voltage-dependent capacitance, current-voltage measurements and electroporation indicate that, as expected, the membrane is asymmetric.

Electric shock convulsions in the rabbit and brain cortex GABAA receptor function

The effect of electric shock convulsions (ESC) on the function of brain cortex GABAA receptors has been studied in the rabbit. Three single electroconvulsive shocks (ECS) were given at intervals of 48 hours and the brain cortex was sampled 36 hours after the last shock. The dose-response curve was determined for GABA-stimulated 36Cl-accumulation into brain cortex microsacs. The parameters of the curve (maximal accumulation rate, Ka and Hill coefficient, n) were constant when determined in two different series of experiences. Animals handled in the same way as the animals from the electric shock group but which did not receive the ECSs (sham ECS group) showed similar maximal accumulation rate and Ka. However, the average n coefficient was significantly higher in the electric shock group. Naive animals, taken from their cages just before the sacrifice, showed dose-response curves which varied from one experimental series to another. This last result (confirming previous observations) shows modifications and inconsistencies in the evaluation of GABAA receptor function in stressed handling-naive animals.

Electroporation in symmetric and asymmetric membranes

We present results of electrical measurements performed both on symmetric and asymmetric membranes in current-clamp conditions. The current-voltage characteristic curve of the membranes shows a reversible conductance transition to a higher level above a critical potential Vc. The experimental results are interpreted in the light of the electroporation theory, which allows estimates of the line tension to be made. These estimates are compared to previous experimental findings or theoretical calculations. The behaviour of symmetric membranes of different chain lengths or consisting of mixtures of short and long chains indicates a strong dependence of Vc on the chain composition and on the presence of charges on the polar head. The electroporation process is also analyzed in asymmetric bilayers consisting of a charged and an uncharged monolayer, a condition which mimics that of natural membranes. Therefore it is possible to analyze the electrical forces acting on the uncharged monolayer due to the presence of charges on the other one, under several ionic-strength conditions. It is shown that the instability arises in the uncharged monolayer, while the coupling between the two monolayers triggers the electroporation process.

Effect of protein kinase C activators on the uptake of GABA by rat brain synaptosomes

Synaptosomes were prepared from rat brain by a discontinuous Ficoll gradient method and used for studying the uptake of labelled GABA. Two GABA uptake components were evidenced, a high (Km = 3.13 microM) and a low (Km = 92.4 microM) affinity one. Preincubation of synaptosomes with two different activators of protein kinase C, phorbol 12, 13-diacetate (PDAc) and oleyl-acetyl glycerol (OAG), resulted in a change of GABA uptake. In particular, the low affinity component increased its Vmax by 58-74%, with no change in the Km. No statistically significant modification was detected for the high affinity component.

Regulation of GABAA receptor in cerebellar granule cells in culture: differential involvement of kinase activities

GABAA receptor function was studied in rat cerebellar granule cells in culture, by the whole-cell patch-clamp approach. The data show that GABA activates Cl- currents in these neurons which reverse at the appropriate membrane potential and are blocked by picrotoxin. The GABA-activated currents desensitize with time of application of the neurotransmitter at concentrations > or = 10(-6) M. The dose-response curve for the peak Cl- current gives a Ka value of 2.3 microM with a Hill coefficient of 1.2. The peak Cl- current elicited by GABA decreases with time of cell registration, with a time-constant of 7.3 min. Residual responsiveness though is maintained thereafter. This "run-down" phenomenon can be completely prevented by adding adenosine-5'-triphosphate + Mg2+ in the pipette solution. Treatments which directly (8-bromoadenosine-3',5'-cyclic-monophosphate; adenosine-3', 5'-cyclic-monophosphate) or indirectly (forskolin, isobutylmethylxanthine) increase the adenosine-3',5'-cyclic-monophosphate intracellular content reduce the GABA-induced Cl- current. Conversely, treatment with the protein kinase A and C inhibitor 1-(5-isoquinolinylsulphonyl)-2-methylpiperazine potentiates the effect of GABA. On the whole, the data indicate that different protein kinase activities modulate the functional state of the GABAA receptors on granule cells from the rat cerebellum.

A transient outward current dependent on external calcium in rat cerebellar granule cells

The outward potassium current of rat cerebellar granule cells in culture was studied with the whole-cell patch-clamp method. Two voltage-dependent components were identified: a slow current, resembling the classical delayed rectifier current, and a fast component, similar to an IA-type current. The slow current was insensitive to 4-aminopyridine and independent of external Ca2+, but significantly inhibited by 3 mM tetraethylammonium. The fast current was depressed by external 4-aminopyridine, with an ED50 = 0.7 mM, and it was abolished by removal of divalent cations from the external medium. The sensitivity of the transient outward current to different divalent cations was investigated by equimolar substitution of Ca2+, Mn2+ and Mg2+. In 2.8 mM Mn2+, the transient potassium conductance was comparable to that in 2.8 mM Ca2+, while in 2.8 mM Mg2+ the transient component was drastically reduced, as in the absence of any divalent cations. However, when Ca2+ was present, Mg2+ up to 5 mM had no effect. The transient current increased with increasing concentrations of external Ca2+, [Ca2+]o, and the maximum conductance vs. [Ca2+]o curve could be approximated by a one-site model. In addition, the current recorded with 5.5 mM BAPTA in the intracellular solution was not different from that recorded in the absence of any Ca2+ buffer. These results suggest that divalent cations modulate the potassium channel interacting with a site on the external side of the cell membrane.

Voltage-dependent calcium currents in dissociated granule cells from rat cerebellum

Voltage-dependent calcium currents were investigated by the patch-clamp technique in whole-cell recording configuration in cultures from 8-day-old rat cerebella, which contained greater than or equal to 90% granule cells. In solutions designed to minimize the sodium and potassium conductances and in 20 mM barium, an inward current activated around -25 mV, reached a peak amplitude at +20 mV and reversed around +80 mV. In 20 mM calcium, this current was approximately 50% of that in barium. From one to three days in vitro only 16% of the cells tested (n = 20) had a current exceeding 50 pA in maximum amplitude, while after four days in vitro the current reached 100 pA in all neurons tested (n greater than 70). Verapamil (50-100 microM) reversibly depressed this current. The dihydropyridine agonist Bay K 8644 (1 microM) enhanced the maximum conductance by 25 +/- 8% (n = 4), caused a negative shift in the activation of 21 +/- 5 mV and a prolongation of the deactivation time course as the voltage was stepped back from +20 to -80 mV. The GABAB agonist baclofen (50 microM) reversibly depressed the current by 27 +/- 8% in 80% of the cells. The effect was similar to that of GABA (10 microM), when the GABAA response (chloride current) was partially blocked by bicucculine. This current can be classified as a dihydropyridine-sensitive high-voltage-activated calcium current. The modulation by GABAB agonists is likely to be significant for presynaptic inhibition.

Adsorption of gamma-aminobutyric acid to phosphatidylserine membranes

The interaction of the negatively-charged phosphatidylserine (PS) and gamma-Aminobutyric acid (GABA) is examined in black lipid membranes (BLM) and inverse micelles. GABA does not permeate through PS membranes and, in concentrations of 10(-5)-10(-4) M, it reduces the negative potential at the membrane-aqueous solution interface. The effect is owing to the adsorption of the GABA cationic species and the consequent decrease of the negative surface charge density of the membrane. When the intrinsic pH of the membrane-solution interface is considered, the Gouy-Chapman-Stern theory describes the GABA screening effect and makes it possible to calculate the GABA-PS binding constant. This value is compared with that obtained measuring the partition of 14C-GABA between an organic phase containing PS and the aqueous solution. The results presented strongly suggest that the electrostatic force plays a major role in GABA-PS interaction.

Guanosine-5'-O-(3-thiotriphosphate) modifies kinetics of voltage-dependent calcium current in chick sensory neurons

Internal perfusion with the G-protein activator guanosine-5'-O-(3-thiotriphosphate) (GTP-gamma S) mimics the effect of noradrenaline and dopamine on the voltage-dependent calcium current in chick dorsal root ganglion (DRG) cells. With 100 microM GTP-gamma S in the pipette, the current at +10 mV was depressed by approximately 50%, with a 10-fold increase of its time to peak. The activation time course of the control calcium current could be approximated with a single exponential curve, whereas with GTP-gamma S the activation time course was double exponential, with time constants tau 1 and tau 2. 2 mM Mg-ATP in the pipette prevented the GTP-gamma S-induced current decrease in 70% of the cells, but the time course of the current was always double exponential. From -50 mV, the current at +10 mV was best fitted with tau 1 = 1.7 +/- 0.5 and tau 2 = 25.6 +/- 5.5 in seven cells. Both time constants decreased with increasing depolarizations. In the first 2 min of recording, the current changed with time. However, both tau 1 and tau 2 were constant, whereas the relative contribution of the slow component increased from 10 to 70%. In addition, the effect was independent of the holding potential in the range from -100 to -30 mV. These results suggest that the activation of a G-protein causes a fraction of the high-threshold calcium channels to switch to a new closed state, with slower opening kinetics.

A transient voltage-dependent outward current in cultured cerebellar granules

Granule cells were dissociated from rat cerebella with a procedure that yields a 98% pure cell population. Potassium currents in these cells were studied using the patch-clamp technique. Depolarizing pulses of 10 mV step and 100 ms duration from a holding potential of -80 mV elicited two different potassium outward currents: a transient, low-voltage activated component and a long lasting, high-voltage activated component. At +30 mV, the total current reached an amplitude of 2 nA (mean value of 15 experiments). The reversal potential of the transient current, estimated by measuring tail currents, was -77 mV, close to that predicted by the Nernst equation. The transient current was half inactivated with a holding potential of -78 mV and completely inactivated with -50 mV or more positive holding potentials. Finally, the current decay could be fitted by the sum of two exponentials with time constants of about 20 and 250 ms.

Conductance transition induced by an electric field in lipid bilayers

A cooperative phenomenon showing a structural change in the organization of bilayer lipid membranes at a critical value of the applied electric field is presented. The transition is characterized by a sharp increase in conductance. The phenomenon can be observed under current-clamp conditions (rather than the usual voltage-clamp conditions) to avoid rupturing the membrane. At a critical potential value the conductance increases and therefore the potential decreases to keep the current constant. Results refer to membranes made of egg phosphatidylcholine (PC), diphytanoylphosphatidylcholine and cholesterol/egg PC. It is found that the critical potential at which the transition occurs depends dramatically on pH and ionic concentration, indicating that the electrical properties of the external surface determine the major characteristics of such a transition.

Dependence of the conductance of the alpha-latrotoxin channel on applied potential and potassium concentration

alpha-Latrotoxin, with a molecular weight 130,000, is the main component of black widow spider venom, and acts at the presynaptical level, inducing a notable release of neurotransmitters in the synapses of all vertebrates. In artificial lipid membranes, this neurotoxin induces the formation of cation-selective ionic channels, whose conductance depends on the intensity and direction of the applied potential. In fact, also in the presence of symmetrical solutions of potassium chloride, the voltage-current characteristics of the single channel strongly rectify. Such rectification, which depends on the concentration of the ions in solution, can be described by a one-site, one-ion model for a channel. The data fit provides the values of the three parameters describing the model. Moreover, a statistical analysis of the amplitude of the single channel, as a function of the concentration of potassium chloride, has made it possible to verify the consistency of the model used.

Permeation of divalent cations through alpha-latrotoxin channels in lipid bilayers: steady-state current-voltage relationships

alpha-Latrotoxin, a polypeptide neurotoxin known to cause massive release of transmitter from vertebrate nerve terminals, is thought to act by forming cation-selective channels in plasma membranes. This paper describes the steady-state current carried by Ca2+, Sr2+ and Ba2+ through pores of alpha-LaTx molecules incorporated in artificial bilayer membranes made of neutral lipids. Even when the solutions separated by the membrane are identical, the I-V relations rectify strongly, the current being higher when the side to which the toxin is added is positive. The polarity of the rectification is consistent with the hypothesis that the mechanism of action of the toxin is, at least in part, that of promoting inwardly directed flow of cations, and thus, accumulation of Ca2+ and other ions in the intracellular spaces. The dependence of the I-V characteristics on voltage and Ca2+ concentration is well described by a one-site, one-ion model for a channel. Three parameters of the model are deduced: the binding constant of the site for Ca2+, K = 1.5 M-1 (or K = 7 M-1 when activities are used instead of concentrations); the "electrical" distance of the site from the toxin-containing solution, alpha = 0.3; the free energy difference between the two barrier peaks, delta F = 0.26 kT. The values of the parameters deduced by studying the channel in the presence of Ca2+ give theoretical curves that also fit the data with Sr2+ and Ba2+, indicating a low level of discrimination among these three cations.