Permeation of internal and external monovalent cations through the catfish cone photoreceptor cGMP-gated channel

The permeation of monovalent cations through the cGMP-gated channel of catfish cone outer segments was examined by measuring permeability and conductance ratios under biionic conditions. For monovalent cations presented on the cytoplasmic side of the channel, the permeability ratios with respect to extracellular Na followed the sequence NH4 > K > Li > Rb = Na > Cs while the conductance ratios at +50 mV followed the sequence Na approximately NH4 > K > Rb > Li = Cs. These patterns are broadly similar to the amphibian rod channel. The symmetry of the channel was tested by presenting the test ion on the extracellular side and using Na as the common reference ion on the cytoplasmic side. Under these biionic conditions, the permeability ratios with respect to Na at the intracellular side followed the sequence NH4 > Li > K > Na > Rb > Cs while the conductance ratios at +50 mV followed the sequence NH4 > K approximately Na > Rb > Li > Cs. Thus, the channel is asymmetric with respect to external and internal cations. Under symmetrical 120 mM ionic conditions, the single-channel conductance at +50 mV ranged from 58 pS in NH4 to 15 pS for Cs and was in the order NH4 > Na > K > Rb > Cs. Unexpectedly, the single-channel current-voltage relation showed sufficient outward rectification to account for the rectification observed in multichannel patches without invoking voltage dependence in gating. The concentration dependence of the reversal potential for K showed that chloride was impermeant. Anomalous mole fraction behavior was not observed, nor, over a limited concentration range, were multiple dissociation constants. An Eyring rate theory model with a single binding site was sufficient to explain these observations.     

Conduction properties of the cloned Shaker K+ channel.

The conduction properties of the cloned Shaker K+ channel were studied using electrophysiological techniques. Single channel conductance increases in a sublinear manner with symmetric increases in K+ activity, reaching saturation by 0.6 M K+. The Shaker K+ channel is highly selective among monovalent cations; under bi-ionic conditions, its selectivity sequence is K+ > Rb+ > NH+4 > Cs+ > Na+, whereas, by relative conductance in symmetric solutions, it is K+ > NH+4 > Rb+ > Cs+. In Cs+ solutions, single channel currents were too small to be measured directly, so nonstationary fluctuation analysis was used to determine the unitary Cs+ conductance. The single channel conductance displays an anomalous molefraction effect in symmetric mixtures of K+ and NH+4, suggesting that the conducting pore is occupied by multiple ions simultaneously.     

Conduction properties of the M-channel in rat sympathetic neurons.

We have investigated the conduction properties of the M-channel in rat superior cervical ganglion neurons. Reversal potentials measured under bi-ionic conditions yielded a permeation sequence of Tl > K > Rb > Cs > NH4 > Na. Slope conductances gave a conductance sequence of K > Tl > NH4 > Rb > Cs. M-current was shown to exhibit a number of features atypical of potassium channels. First, the conduction of monovalent cations relative to K was very low. Second, the nature of the permeant ion did not affect the deactivation kinetics. Third, M-current did not exhibit anomalous mole-fraction behavior, a property suggestive of a multi-ion pore. Finally, external Ba, which is a blocker of M-current, showed a preferential block of outward current and had much less effect on inward current. The permeability sequence of the M-channel is very similar to other K-selective channels, implying a high degree of conservation in the selectivity filter. However, other conduction properties suggest that the pore structure outside of the selectivity filter is very different from previously cloned potassium channels.     

The relation between ion permeation and recovery from inactivation of ShakerB K+ channels.

We have studied the relation between permeation and recovery from N-type or ball-and-chain inactivation of ShakerB K channels. The channels were expressed in the insect cell line Sf9, by infection with a recombinant baculovirus, and studied under whole cell patch clamp. Recovery from inactivation occurs in two phases. The faster of the two lasts for approximately 200 ms and is followed by a slow phase that may require seconds for completion. The fast phase is enhanced by both permeant ions (K+, Rb+) and by the blocking ion Cs+, whereas the impermeant ions (Na+, Tris+, choline+) are ineffective. The relative potencies are K+ > Rb+ > Cs+ > NH4+ >> Na+ approximately choline+ approximately Tris+. Ion permeation through the channels is not essential for recovery. The results suggest that cations influence the fast phase of recovery by binding in a site with an electrical distance greater than 0.5. Recovery from fast inactivation is voltage-dependent. With Na+, choline+, or Tris+ outside, about 15% of the channels recover in the fast phase (-80 mV), and the other 85% apparently enter a second inactivated state from which recovery is very slow. Recovery in this phase is not influenced by external ions, but is speeded by hyperpolarization.     

Conductance and permeation of monovalent cations through depletion-activated Ca2+ channels (ICRAC) in Jurkat T cells.

We studied monovalent permeability of Ca2+ release-activated Ca2+ channels (ICRAC) in Jurkat T lymphocytes following depletion of calcium stores. When external free Ca2+ ([Ca2+]o) was reduced to micromolar levels in the absence of Mg2+, the inward current transiently decreased and then increased approximately sixfold, accompanied by visibly enhanced current noise. The monovalent currents showed a characteristically slow deactivation (tau = 3.8 and 21.6 s). The extent of Na+ current deactivation correlated with the instantaneous Ca2+ current upon readdition of [Ca2+]o. No conductance increase was seen when [Ca2+]o was reduced before activation of ICRAC. With Na+ outside and Cs+ inside, the current rectified inwardly without apparent reversal below 40 mV. The sequence of conductance determined from the inward current at -80 mV was Na+ > Li+ = K+ > Rb+ >> Cs+. Unitary inward conductance of the Na+ current was 2.6 pS, estimated from the ratios delta sigma2/delta Imean at different voltages. External Ca2+ blocked the Na+ current reversibly with an IC50 value of 4 microM. Na+ currents were also blocked by 3 mM Mg2+ or 10 microM La3+. We conclude that ICRAC channels become permeable to monovalent cations at low levels of external divalent ions. In contrast to voltage-activated Ca2+ channels, the monovalent conductance is highly selective for Na+ over Cs+. Na+ currents through ICRAC channels provide a means to study channel characteristics in an amplified current model.     

Functional role of a conserved aspartate in the external mouth of voltage-gated potassium channels.

Mutation of the glycines in a conserved Gly-Tyr-Gly-Asp sequence in the P-region of voltage-gated K channels has identified determinants of Na/K selectivity. But the function of the negatively charged Asp is not known because mutations at this position are not tolerated, owing to the fourfold replication of mutations in a tetrameric channel. We have successfully mutated Asp378-->Thr in a tandem dimer Kv2.1 construct to yield a twofold neutralization of charge at this site. When expressed in Xenopus oocytes, the mutated channels showed markedly altered ion conduction and blockade. Potassium conduction in the inward direction was selectively reduced, so that the instantaneous current-voltage relationship obtained in isotonic KCl became strongly outwardly rectifying. The relative permeability to Na+, PNa/PK, increased from 0.02 to 0.10 without changing the ion selectivity sequence K > Rb >> Cs >> Na. The IC50 for block by external tetraethylammonium (TEA) increased more than 100-fold without affecting block by internal TEA. We conclude that Asp378 is an essential part of a potassium ion binding site associated with the Na/K selectivity filter at the external mouth of the pore.     

Mechanisms of Anion and Cation Permeations in the Resting Membrane of a Barnacle Muscle Fiber

The resting membrane of a barnacle muscle fiber is mostly permeable to cations in a solution of pH 7.7 whereas it becomes primarily permeable to anions if the pH is below 4.0. Mechanisms of ion permeation for various monovalent cations and anions were investigated at pH 7.7 and 3.9, respectively. Permeability ratios were obtained from the relationship between the membrane potential and the concentration of the test ions, and ionic conductances from current-voltage relations of the membrane. The permeability sequence for anions (SCN > I > NO₃ > Br > ClO₃ > Cl > BrO₃ > IO₃) was different from the conductance sequence for anions (Br, Cl > ClO₃, NO₃ > SCN). In contrast, the permeability and conductance sequences were identical for cations (K > Rb > Cs > Na > Li). The results suggest that anion permeation is governed by membrane charges while cation permeation is via some electrically neutral mechanism.     

The Effects of Alkali Metal Cations and Common Anions on the Frog Skin Potential

The effects on the potential difference across isolated frog skin (R. catesbeiana, R. pipiens) of changing the ionic composition of the bathing solutions have been examined. Estimates of mean values and precision are presented for the potential changes produced by substituting other alkali metal cations for Na at the outside border and for K at the inside border. In terms of ability to mimic Na at the outside border of bullfrog skin, the selectivity order is Li > Rb, K, Cs; at the outside border of leopard frog skin, Li > Cs, K, Rb. In terms of ability to mimic K at the inside border of bullfrog and leopard frog skin: Rb > Cs > Li > Na. Orders of anion selectivity in terms of sensitivity of the potential for the outside border of bullfrog skin are Br > Cl > NO₃ > I > SO₄, isethionate and of leopard frog skin are Br, Cl > I, NO₃, SO₄. An effect of the solution composition (ionic strength?) on the apparent Na-K selectivity of the outside border is described. The results of the investigation have been interpreted and discussed in terms of the application of the constant field equation to the Koefoed-Johnsen-Ussing frog skin model. These observations may be useful in constructing and testing models of biological ionic selectivity.     

KINETICS OF PENETRATION : VII. MOLECULAR VERSUS IONIC TRANSPORT

In some living cells the order of penetration of certain cations corresponds to that of their mobilities in water. This has led to the idea that electrolytes pass chiefly as ions through the protoplasmic surface in which the order of ionic mobilities is supposed to correspond to that found in water. If this correspondence could be demonstrated it would not prove that electrolytes pass chiefly as ions through the protoplasmic surface for such a correspondence could exist if the movement were mostly in molecular form. This is clearly shown in the models here described. In these the protoplasmic surface is represented by a non-aqueous layer interposed between two aqueous phases, one representing the external solution, the other the cell sap. The order of penetration through the non-aqueous layer is Cs > Rb > K > Na > Li. This will be recognized as the order of ionic mobilities in water. Nevertheless the movement is mostly in molecular form in the nonaqueous layer (which is used in the model to represent the protoplasmic surface) since the salts are very weak electrolytes in this layer. The chief reason for this order of penetration lies in the fact that the partition coefficients exhibit the same order, that of cesium being greatest and that of lithium smallest. The partition coefficients largely control the rate of entrance since they determine the concentration gradient in the non-aqueous layer which in turn controls the process of penetration. The relative molecular mobilities (diffusion constants) in the non-aqueous layer do not differ greatly. The ionic mobilities are not known (except for K⁺ and Na⁺) but they are of negligible importance, since the movement in the non-aqueous layer is largely in molecular form. They may follow the same order as in water, in accordance with Walden''s rule. Ammonium appears to enter faster than its partition coefficient would lead us to expect, which may be due to rapid penetration of NH₃. This recalls the apparent rapid penetration of ammonium in living cells which has also been explained as due to the rapid penetration of NH₃. Both observation and calculation indicate that the rate of penetration is not directly proportional to the partition coefficient but increases somewhat less rapidly. Many of these considerations doubtless apply to living cells.     

Binding of alkaline cations to the double-helical form of gramicidin.

Gramicidin is a polypeptide antibiotic that forms monovalent cation-specific channels in membrane environments. In organic solvents and in lipids containing unsaturated fatty acid chains, it forms a double-helical "pore" structure, in which two monomers are intertwined. This form of gramicidin can bind two cations inside its lumen, and the crystal structures of both an ion complex and an ion-free form have been determined. In this study, we have used circular dichroism (CD) spectroscopy to examine the binding mechanism and the binding constants (K1 and K2) of cations to gramicidin in the double helical form in methanol solution. The dramatic change in optical rotation in the far-ultraviolet CD spectrum of gramicidin provides a useful tool for monitoring the binding. The binding mechanism appears to involve a large conformation change associated with the binding of ions to the first of the two sites. The calculated values for the K1 binding constants for alkaline cations are considerably smaller than the K2 binding constants. The order of binding affinity for alkaline cations is similar to that for the helical dimer "channel" form of gramicidin, i.e., Cs+ approximately Rb+ > > K+ > Li+, but in comparison to the helical dimer form, the binding to double-helical dimers is dominated by a cation size-dependent conformational change in the gramicidin structure.     

Cation Activation of the Basolateral Sodium-Potassium Pump in Turtle Colon

The current generated by electrogenic sodium-potassium exchange at the basolateral membrane of the turtle colon can be measured directly in tissues that have been treated with serosal barium (to block the basolateral potassium conductance) and mucosal amphotericin B (to reduce the cation selectivity of the apical membrane). We studied the activation of this pump current by mucosal sodium and serosal potassium, rubidium, cesium, and ammonium. The kinetics of sodium activation were consistent with binding to three independent sites on the cytoplasmic side of the pump. The pump was not activated by cellular lithium ions. The kinetics of serosal cation activation were consistent with binding to two independent sites with the selectivity Rb > K > Cs > NH₄. The properties and kinetics of the basolateral Na/K pump in the turtle colon are at least qualitatively similar to those ofthe well-characterized Na/K-ATPase of the human red blood cell .     

Interactions in cation permeation through the gramicidin channel. Cs, Rb, K, Na, Li, Tl, H, and effects of anion binding.

As a prototype for binding and interaction in biological Na and K channels, the single channel conductances for Li, Na, K, Rb, Cs, H, and Tl and the membrane potentials for Tl-K mixtures are characterized for gramicidin A over wider concentration rangers than previously and analyzed using an "equilibrium domain" model that assumes a central rate-determining barrier. Peculiarities in the conductance-concentration relationship for TlF, TlNO3, and TlAc suggest that anions bind to Tl-loaded channels, and the theory is extended to allow for this. For concreteness, the selectivity of cation permeation is characterized in terms of individual binding and rate constants of this model, with the conclusions that the strongest site binds Cs greater than Rb greater than K greater than Na greater than Li, while the next strongest binds Na greater than K greater than Li greater than Rb greater than Cs. However, because Schagina, Grinfeldt, and Lev''s recent finding of single filing (personal communication) indicates that the channel sites in gramicidin cannot be at equilibrium with the solution, and work in progress with Hägglund and Enos (Biophys. J. 21:26a. [Abstr.]) indicates that the simplest model adequate to account for the observed concentration-dependences of flux-ratio, conductance, I--V characteristic, and permeability has three barriers and four sites, some implications of additional rate-determining barriers at the mouth of the channel are discussed. The results are summarized using phenomenological "experimental" parameters that provide a model-independent way to represent that data concisely and which can be interpreted physically in terms of any desired model.     

Ion selectivity of porcine skeletal muscle Ca2+ release channels is unaffected by the Arg615 to Cys615 mutation.

The Arg615 to Cys615 mutation of the sarcoplasmic reticulum (SR) Ca2+ release channel of malignant hyperthermia susceptible (MHS) pigs results in a decreased sensitivity of the channel to inhibitory Ca2+ concentrations. To investigate whether this mutation also affects the ion selectivity filter of the channel, the monovalent cation conductances and ion permeability ratios of single Ca2+ release channels incorporated into planar lipid bilayers were compared. Monovalent cation conductances in symmetrical solutions were: Li+, 183 pS +/- 3 (n = 21); Na+, 474 pS +/- 6 (n = 29); K+, 771 pS +/- 7 (n = 29); Rb+, 502 pS +/- 10 (n = 22); and Cs+, 527 pS +/- 5 (n = 16). The single-channel conductances of MHS and normal Ca2+ release channel were not significantly different for any of the monovalent cations tested. Permeability ratios measured under biionic conditions had the permeability sequence Ca2+ >> Li+ > Na+ > K+ > or Rb+ > Cs+, with no significant difference noted between MHS and normal channels. This systematic examination of the conduction properties of the pig skeletal muscle Ca2+ release channel indicated a higher Ca2+ selectivity (PCa2+:Pk+ approximately 15.5) than the sixfold Ca2+ selectivity previously reported for rabbit skeletal (Smith et al., 1988) or sheep cardiac muscle (Tinker et al., 1992) Ca2+ release channels. These results also indicate that although Ca2+ regulation of Ca2+ release channel activity is altered, the Arg615 to Cys615 mutation of the porcine Ca2+ release channel does not affect the conductance or ion selectivity properties of the channel.     

Ion permeation and block of M-type and delayed rectifier potassium channels. Whole-cell recordings from bullfrog sympathetic neurons

Ion permeation and conduction were studied using whole-cell recordings of the M-current (I(M)) and delayed rectifier (IDR), two K+ currents that differ greatly in kinetics and modulation. Currents were recorded from isolated bullfrog sympathetic neurons with 88 mM [K+]i and various external cations. Selectivity for extracellular monovalent cations was assessed from permeability ratios calculated from reversal potentials and from chord conductances for inward current. PRb/PK was near 1.0 for both channels, and GRb/GK was 0.87 +/- 0.01 for IDR but only 0.35 +/- 0.01 for I(M) (15 mM [Rb+]o or [K+]o). The permeability sequences were generally similar for I(M) and IDR: K+ approximately Rb+ > NH4+ > Cs+, with no measurable permeability to Li+ or CH3NH3+. However, Na+ carried detectable inward current for IDR but not I(M). Nao+ also blocked inward K+ current for IDR (but not IM), at an apparent electrical distance (delta) approximately 0.4, with extrapolated dissociation constant (KD) approximately 1 M at 0 mV. Much of the instantaneous rectification of IDR in physiologic ionic conditions resulted from block by Nao+. Extracellular Cs+ carried detectable inward current for both channel types, and blocked I(M) with higher affinity (KD = 97 mM at 0 mV for I(M), KD) approximately 0.2 M at 0 mV for IDR), with delta approximately 0.9 for both. IDR showed several characteristics reflecting a multi-ion pore, including a small anomalous mole fraction effect for PRb/PK, concentration-dependent GRb/GK, and concentration- dependent apparent KD's and delta's for block by Nao+ and Cso+. I(M) showed no clear evidence of multi-ion pore behavior. For I(M), a two- barrier one-site model could describe permeation of K+ and Rb+ and block by Cso+, whereas for IDR even a three-barrier, two-site model was not fully adequate.     

Ion effects on gating of the Ca(2+)-activated K+ channel correlate with occupancy of the pore.

We studied the effects of permeant ions on the gating of the large conductance Ca(2+)-activated K+ channel from rat skeletal muscle. Rb+ blockade of inward K+ current caused an increase in the open probability as though Rb+ occupancy of the pore interferes with channel closing. In support of this hypothesis, we directly measured the occupancy of the pore by the impermeant ion Cs+ and found that it strongly correlates with its effect on gating. This is consistent with the "foot-in-the-door" model of gating, which states that channels cannot close with an ion in the pore. However, because Rb+ and Cs+ not only slow the closing rate (as predicted by the model), but also speed the opening rate, our results are more consistent with a modified version of the model in which the channel can indeed close while occupied, but the occupancy destabilizes the closed state. Increasing the occupancy of the pore by the addition of other permeant (K+ and Tl+) and impermeant (tetraethylammonium) ions did not affect the open probability. To account for this disparity, we used a two-site permeation model in which only one of the sites influenced gating. Occupancy of this "gating site" interferes with channel closing and hastens opening. Ions that directly or indirectly increase the occupancy of this site will increase the open probability.     

Intracellular Methamphetamine Prevents the Dopamine-induced Enhancement of Neuronal Firing

The dysregulation of the dopaminergic system is implicated in multiple neurological and neuropsychiatric disorders such as Parkinson disease and drug addiction. The primary target of psychostimulants such as amphetamine and methamphetamine is the dopamine transporter (DAT), the major regulator of extracellular dopamine levels in the brain. However, the behavioral and neurophysiological correlates of methamphetamine and amphetamine administration are unique from one another, thereby suggesting these two compounds impact dopaminergic neurotransmission differentially. We further examined the unique mechanisms by which amphetamine and methamphetamine regulate DAT function and dopamine neurotransmission; in the present study we examined the impact of extracellular and intracellular amphetamine and methamphetamine on the spontaneous firing of cultured midbrain dopaminergic neurons and isolated DAT-mediated current. In dopaminergic neurons the spontaneous firing rate was enhanced by extracellular application of amphetamine > dopamine > methamphetamine and was DAT-dependent. Amphetamine > methamphetamine similarly enhanced DAT-mediated inward current, which was sensitive to isosmotic substitution of Na⁺ or Cl⁻ ion. Although isosmotic substitution of extracellular Na⁺ ions blocked amphetamine and methamphetamine-induced DAT-mediated inward current similarly, the removal of extracellular Cl⁻ ions preferentially blocked amphetamine-induced inward current. The intracellular application of methamphetamine, but not amphetamine, prevented the dopamine-induced increase in the spontaneous firing of dopaminergic neurons and the corresponding DAT-mediated inward current. The results reveal a new mechanism for methamphetamine-induced dysregulation of dopaminergic neurons.     

Freeze-thaw injury to isolated spinach protoplasts and its simulation at above freezing temperatures

Possibilities to account for the mechanism of freeze-thaw injury to isolated protoplasts of Spinacia oleracea L. cv. Winter Bloomsdale were investigated. A freeze-thaw cycle to −3.9 C resulted in 80% lysis of the protoplasts. At −3.9 C, protoplasts are exposed to the equivalent of a 2.1 osmolal solution. Isolated protoplasts behave as ideal osmometers in the range of concentrations tested (0.35 to 2.75 osmolal), arguing against a minimum critical volume as a mechanism of injury. Average protoplast volume after a freeze-thaw cycle was not greatly different than the volume before freezing, arguing against an irreversible influx of solutes while frozen. A wide variety of sugars and sugar alcohols, none of which was freely permeant, were capable of protecting against injury which occurred when protoplasts were frozen in salt solutions. The extent of injury was also dependent upon the type of monovalent ions present, with Li = Na > K = Rb = Cs and Cl ≥ Br > I, in order of decreasing protoplast survival. Osmotic conditions encountered during a freeze-thaw cycle were established at room temperature by exposing protoplasts to high salt concentrations and then diluting the osmoticum. Injury occurred only after dilution of the osmoticum and was correlated with the expansion of the plasma membrane. Injury observed in frozen-thawed protoplasts was correlated with the increase in surface area the plasma membrane should have undergone during thawing, supporting the contention that contraction of the plasma membrane during freezing and its expansion during thawing are two interacting lesions which cause protoplast lysis during a freezethaw cycle.     

Ion transport mediated by the valinomycin analogue cyclo(L-Lac-L-Val-D- Pro-D-Val)3 in lipid bilayer membranes

Cyclo(L-Lac-L-Val-D-Pro-D-Val)3 (PV-Lac) a structural analogue of the ion-carrier valinomycin, increases the cation permeability of lipid bilayer membranes by forming a 1:1 ion-carrier complex. The selectively sequence for PV-Lac is identical to that of valinomycin; i.e., Rb+ greater than K+ greater than Cs+ greater than or equal to NH+4 greater than Na+ greater than Li+. The steady-state zero-voltage conductance, G(0), is a saturating function of KCl concentration. A similar behavior was found for Rb+, Cs+, and NH+4. However, the ion concentration at which G(0) reaches a plateau strongly depends on membrane composition. The current-voltage curves present saturating characteristics, except at low ion concentrations of Rb+, K+, or Cs+. The ion concentration at which the saturating characteristics appear depends on membrane composition. These and other results presented in this paper agree with a model that assumes complexation between carrier and ion at the membrane-water interface. Current relaxation after voltage-jump studies were also performed for PV-Lac. Both the time constant and the amplitude of the current after a voltage jump strongly depend on ion concentration and membrane composition. These results, together with the stationary conductance data, were used to evaluate the rate constants of the PV-Lac-mediated K+ transport. In glycerolmonooleate they are: association rate constant, 2 x 10(6) M-1 s-1; dissociation rate constant, 4 x 10(5) s-1; translocation rate constant for complex, 5 x 10(4) s-1; and the rate of translocation of the free carrier (ks), 55 s-1. ks is much smaller for PV-Lac than for valinomycin and thus limits the efficiency with which the carrier is able to translocate cations across the membrane.     

Cationic selectivity and competition at the sodium entry site in frog skin

The cation selectivity of the Na entry mechanism located in the outer membrane of the bullfrog (Rana catesbeiana) skin epithelium was studied. This selectivity was determined by measuring the short-circuit current when all of the external sodium was replaced by another cation and, also, by noting the relative degree of inhibition that the alkali metal cations produced on Na influx. The ability of the Group Ia cations to permeate the apical membrane was determined from the tracer uptake experiments. The results demonstrate that (a) only Li and Na are actively transported through the epithelium; (b) the alkali cations K, Rb, and Cs do not enter the epithelium through the apical border and, therefore, Na and Li are the only alkali cations translocated through this membrane; (c) these impermeable cations are competitive inhibitors of Na entry; (d) the cations NH4 and Tl exhibit more complex behavior but, under well-defined conditions, also inhibit Na entry; and (e) the selectivity of the cation binding site is in the sequence Li congruent to Na > Tl > NH4 congruent to K > Rb > Cs, which corresponds to a high field strength site with tetrahedral symmetry.     

Ion Absorption and Retention by Chlorella pyrenoidosa. III. Selective Accumulation of Rubidium, Potassium and Sodium

The selective preference of Chlorella pyrenoidosa for alkali metal cations was found to have the order Rb > K > > > Na. It was demonstrated that a cation of higher preference can replace in the cell cations of lower preference by an ion interchange process.

The replacement of Na from the cell by K or Rb occurred against high external Na concentrations up to 450 meq/1 Na.

It is suggested that the structural selectivity of the Chlorella cell may be amplified by a chromatographic type repetitive selection process, driven by metabolically dependent unsymmetric shape changes of cellular membranes.