Axonal surface charges: evidence for phosphate structure.

The effect of different extracellular alkaline-earth cations (Ca2+, Mg2+, Sr2+, Ba2+) upon the threshold membrane potential for spike initiation in crayfish axon has been studied by means of intracellular microelectrodes. This was done at the following extracellular concentrations of the divalent uranyl ion (UO2/2+): 1.0 X 10(-6) M, 3.0 X 10(-6) M, and 9.0 X 10(-6) M. At each concentration employed, extensive neutralization of axonal surface charges by UO2/2+ was evidenced by the fact that equal concentrations (50 mM) of alkaline-earth cations did not have the same effect on the threshold potential. The selectivity sequences observed at the different uranyl-ion concentrations were: 1.0 X 10(-6) M UO2/2+, Ca2+ greater than Mg2+ greater than Sr2+ greater than Ba2+; 3.0 X 10(-6) M UO2/2+, Ca2+ greater than Mg2+ greater than Ba2+ larger than or equal to Sr2+; 9.0 X 10(-6) M UO2/2+, Ca2+ approximately Ba2+ greater than Sr2+ greater than Mg2+. These selectivity sequences are in accord with the equilibrium selectivity theory for alkaline-earth cations. At each of the concentrations used, uranyl ion did not have any detectable effect on the actual shape of the action potential itself. It is concluded that many (if not most) of the surface acidic groups in the region of the sodium gates represent phosphate groups of membrane phospholipids, but that the m gates themselves are probably protein-aceous in structure.     

Activation of three types of membrane currents by various divalent cations in identified molluscan pacemaker neurons

We investigated membrane currents activated by intracellular divalent cations in two types of molluscan pacemaker neurons. A fast and quantitative pressure injection technique was used to apply Ca2+ and other divalent cations. Ca2+ was most effective in activating a nonspecific cation current and two types of K+ currents found in these cells. One type of outward current was quickly activated following injections with increasing effectiveness for divalent cations of ionic radii that were closer to the radius of Ca2+ (Ca2+ greater than Cd2+ greater than Hg2+ greater than Mn2+ greater than Zn2+ greater than Co2+ greater than Ni2+ greater than Pb2+ greater than Sr2+ greater than Mg2+ greater than Ba2+). The other type of outward current was activated with a delay by Ca2+ greater than Sr2+ greater than Hg2+ greater than Pb2+. Mg2+, Ba2+, Zn2+, Cd2+, Mn2+, Co2+, and Ni2+ were ineffective in concentrations up to 5 mM. Comparison with properties of Ca2(+)-sensitive proteins related to the binding of divalent cations suggests that a Ca2(+)-binding protein of the calmodulin/troponin C type is involved in Ca2(+)-dependent activation of the fast-activated type of K+ current. Th sequence obtained for the slowly activated type is compatible with the effectiveness of different divalent cations in activating protein kinase C. The nonspecific cation current was activated by Ca2+ greater than Hg2+ greater than Ba2+ greater than Pb2+ greater than Sr2+, a sequence unlike sequences for known Ca2(+)-binding proteins.     

Interaction of calcium with bovine plasma protein C

The binding of 45Ca2+ to bovine plasma protein C (PC) and to activated bovine plasma protein C (APC) has been examined by equilibrium ultrafiltration at pH 7.4 and 25 degrees C. Under these conditions, PC possesses 16.0 plus or minus 2.0 equivalent Ca2+ binding sites, of average KD (8.7 plus or minus 1.5) x 10(-4) M, and APC contains 9.0 plus or minus 1.0 equivalent Ca2+ binding sites, with an average KD of (4.3 plus or minus 1.1) x 10(-4) M. Both Mn2+ and Sr2+ were capable of ready displacement of Ca2+ from a Ca2+-PC complex, while Mg2+ was less effective in this regard. The alpha-thrombin-catalyzed activation of PC was inhibited by the presence of Ca2+. A kinetic analysis of this effect demonstrated that it was, in large part, due to an increase in the Km of the reaction. Addition of other divalent cations, e.g. Mn2+, Sr2+, and Mg2+, in place of Ca2+ also resulted in inhibition of the alpha-thrombin-catalyzed activation of PC in a manner which paralleled their ability to displace Ca2+ from a Ca2+-PC complex. On the other hand, the activation of PC by the coagulant protein from Russell's Viper venom was augmented by the presence of Ca2+. Other divalent metal ions, such as Sr2+ and Mn2+, in the absence of Ca2+, also weakly stimulated this reaction. Mg2+ was without notable effect.     

Role of bivalent and monovalent cations in the functioning of myosin ATPase

The effects of bivalent (Mg2+, Ca2+, Sr2+) and monovalent (K+, Na+, NH4+) cations on the ATPase activity of subfragment 1 of myosin (SI) with a decreased Mg2+ content (EDTA-SI) were studied. Mg2+ activate the EDTA-SI ATPase, but only in the absence of other activating cations. K+, NH4+, a2+ and Sr2+ have a much stronger activating effect on EDTA-SI ATPase than on Mg-SI (SI enriched with Mg2+) ATPase. Monovalent cations inhibit Mg2+-ATPase and Ca2+-ATPase of EDTA-SI, while K+ and NH4+ activate Sr2+-ATPase of EDTA-SI. Based on experimental results and literary data, a hypothesis on the participation of the cations in the functioning of myosin ATPase was postulated. This hypothesis entails the existence of two closely interconnected cation-binding sites in the vicinity of the myosin active center (one for bivalent and one for monovalent cations); the ATPase activity of myosin is at any moment dependent on the nature of cations present in these two sites. An attempt to explain the role of the cations in the accomplishment of the ATPase reaction by myosin was made.     

Studies of gastric Ca2+-stimulated adenosine triphosphatase. I. characterization and general properties

Gastric microsomes do not contain any significant Ca2+-stimulated ATPase activity. Trypsinization of pig gastric microsomes in presence of ATP results in significant (2-3 fold) increase in the basal (with Mg2+ as the only cation) ATPase activity, with virtual elimination of the K+-stimulated component. Such treatment causes unmasking of latent Mg2+-dependent Ca2+-stimulation ATPase. Other divalent cations such as Sr2+, Ba2+, Zn2+, and Mn2+ were found ineffective as a substitute for Ca2+. Moreover, those divalent cations acted as inhibitors of the Ca2+-stimulated ATPase activity. The pH optimum of the enzyme is around 6.8. The enzyme has a Km of 70 microM for ATP and the Ka values for Mg2+ and Ca2+ are about 4 x 10(-4) and 10(-7) M, respectively. Studies with inhibitors suggest the involvement of sulfhydryl and primary amino groups in the operation of the enzyme. Possible roles of the enzyme in gastric H+ transport have been discussed.     

Divalent cation conduction in the ryanodine receptor channel of sheep cardiac muscle sarcoplasmic reticulum

The conduction properties of the alkaline earth divalent cations were determined in the purified sheep cardiac sarcoplasmic reticulum ryanodine receptor channel after reconstitution into planar phospholipid bilayers. Under bi-ionic conditions there was little difference in permeability among Ba2+, Ca2+, Sr2+, and Mg2+. However, there was a significant difference between the divalent cations and K+, with the divalent cations between 5.8- and 6.7-fold more permeant. Single-channel conductances were determined under symmetrical ionic conditions with 210 mM Ba2+ and Sr2+ and from the single-channel current-voltage relationship under bi-ionic conditions with 210 mM divalent cations and 210 mM K+. Single-channel conductance ranged from 202 pS for Ba2+ to 89 pS for Mg2+ and fell in the sequence Ba2+ greater than Sr2+ greater than Ca2+ greater than Mg2+. Near-maximal single-channel conductance is observed at concentrations as low as 2 mM Ba2+. Single-channel conductance and current measurements in mixtures of Ba(2+)-Mg2+ and Ba(2+)-Ca2+ reveal no anomalous behavior as the mole fraction of the ions is varied. The Ca(2+)-K+ reversal potential determined under bi-ionic conditions was independent of the absolute value of the ion concentrations. The data are compatible with the ryanodine receptor channel acting as a high conductance channel displaying moderate discrimination between divalent and monovalent cations. The channel behaves as though ion translocation occurs in single file with at most one ion able to occupy the conduction pathway at a time.     

Effects of bivalent cations on prostaglandin biosynthesis and phospholipase A2 activation in rabbit kidney medulla slices.

The bivalent cations Ca2+, Mg2+, Co2+, Mn2+, Sr2+ and Ba2+ were compared for their stimulatory or inhibitory effect on prostaglandin formation in rabbit kidney medulla slices. Ca2+, Mn2+ and Sr2+ ions stimulated prostaglandin generation up to 3--5-fold in a time- and dose-dependent manner (Ca2+ greater than Mn2+ congruent to Sr2+). The stimulation by Mn2+ (but not by Sr2+) was also observed in incubations of medulla slices in the presence of Ca2+. Mg2+ and Co2+ ions were without significant effects on either basal or Ca2+-stimulated prostaglandin synthesis. The stimulatory effects of Ca2+, Mn2+ and Sr2+ on medullary generation of prostaglandin E2 were found to correlate with their stimulatory effects on the release of arachidonic acid and linoleic acid from tissue lipids. The release of other fatty acids was unaffected, except for a small increase in oleic acid release. As both arachidonic acid and linoleic acid are predominantly found in the 2-position of the glycerol moiety of phospholipids, the stimulation by these cations of prostaglandin E2 formation appears to be mediated via stimulation of phospholipase A2 activity.     

Thermodynamic data on the binding of six M2(+)-ions to bovine, goat, and human alpha-lactalbumin.

By batch microcalorimetry we titrated the apo-forms of bovine, goat, and human alpha-lactalbumin with Mg2+, Ca2+, Sr2+, Ba2+, Mn2+, and Cd2+ ions at pH 7.5 and 25 degrees C. The titration curves enabled us to calculate the apparent enthalpy changes and binding constants and thus, also the free energy and the entropy changes of the binding. CD-spectra showed that all cations induce the same conformational change to the native form of the protein. The calorimetric and spectroscopic results, as well as sequence comparisons confirm the hypothesis that all these ions occupy the very same site on the molecule. The thermodynamic parameters, plotted vs the ionic radii, run parallel for the three proteins, which illustrates the earlier proposed "rigid site" model.     

Stringent requirement for Ca2+ in the removal of Z-lines and alpha-actinin from isolated myofibrils by Ca2+-activated neutral proteinase

Treatment of isolated myofibrils with Ca2+-activated neutral proteinase (CANP) results in specific removal of Z-line and of alpha-actinin. To investigate the ionic requirement for these processes, we measured Z-line removal by phase-contrast and interference microscopy and alpha-actinin removal by sodium dodecyl sulphate/polyacrylamide-gel electrophoretic analysis of myofibrillar proteins. The proteolytic digestion of native purified proteins was measured directly on polyacrylamide gels and by the fluorescamine technique. We found that the removal of Z-line and alpha-actinin as well as the release of proteolytic degradation products from isolated myofibrils by CANP occur only in the presence of Ca2+; Sr2+, Ba2+, Mn2+, Mg2+, Co2+ and Zn2+ are all ineffective. In contrast with this stringent requirement for Ca2+, the proteolytic activity of CANP measured with denatured casein, native and denatured haemoglobin, native actin and tropomyosin also occurs in the presence of other bivalent cations, in the following order: Ca2+ greater than Sr2+ greater than Ba2+. These data suggest that only Ca2+ can produce the conformational change in myofibrils that renders them susceptible to the action of CANP, whereas its proteolytic activity is stimulated by several bivalent ions.     

Tests of an electrostatic screening hypothesis of the inhibition of neurotransmitter release by cations at the frog neuromuscular junction.

We have investigated an electrostatic screening hypothesis of cationic inhibition of quantal release at the neuromuscular junction of the frog (Rana pipiens). According to this hypothesis, increasing the extracellular concentration of an inhibitory cation reduces the quantal content (m) of the end-plate potential by reducing the ability of negative surface charge to attract Ca2+ to the external surface of the presynaptic membrane. The inhibitory power of various cations should depend only on their net ionic charge and should increase strongly with increasing charge. We have demonstrated, in Ringer's solutions containing modified concentrations of Na+, Ca+, and Mg2+, that at fixed concentrations of Ca2+ and Na+ (a) the dependence of m on [Mg2+]0 is satisfactorily accounted for by electrostatic theory and (b) the dependence of m on the univalent cation concentration of the modified Ringer's solution is satisfactorily predicted from the Mg2+ inhibition of m. (Glucosamine or arginine was used to replace a fraction of the Na+ content of Ringer's solution in the latter experiments.) These results are consistent with electrostatic screening actions of Mg2+ and univalent cations in the inhibition of m. We have also re-examined the inhibition of m caused by the addition to Ringer's solution of two trace concentration divalent cations, Mn2+ and Sr2+. Our data suggest that the inhibition of m by Sr2+ at high quantal contents may also be due to surface charge screening, while the potent inhibitory actions of Mn2+ may be due to its ability to bind negative surface charge.     

Two Ca2+ functions are demonstrated by the substitution of specific divalent and lanthanide cations for the Ca2+ required by the aggregation factor complex from the marine sponge, Microciona prolifera

Multivalent cations were tested for their ability to replace the Ca2+ requirements of aggregation factor (AF) complex in activity, stability, and integrity assays. The ability of each cation to replace the Ca2+ required for the cell aggregation-enhancing activity of AF was examined by replacing the usual 10 mM Ca2+ with the test cation at various concentrations in the serial dilution assay of the AF. The other alkaline earth cations, Mg2+, Sr2+, and Ba2+, could not replace Ca2+; two transition elements, Mn2+ and Cd2+, partially replaced calcium. All 15 of the available lanthanides (including La3+ and Y3+) produced normal activity but only at 10-400-fold lower cation concentrations than Ca2+. An AF preparation is stable and remains active for months in 1 mM Ca2+ but decays rapidly when Ca2+ is lowered. Sr2+ and Ba2+ at 20 mM but not at 1 mM could replace 1 mM Ca2+ and give long term stability. AF was not stable in the presence of Mg2+, even at 100 mM. High Mn2+ concentrations did not stabilize AF even though AF was partially active in Mn2+. Cd2+ gave full stability at 75 mM and La3+ at about 0.1 mM. When Ca2+ is chelated, the macromolecular subunits of the AF slowly dissociate. Permeation chromatography and analytical ultracentrifugation showed that the cations that stabilized activity maintained the integrity of AF complex while those that failed to stabilize activity allowed the complex to dissociate into subunits, indicating that these two Ca2+ requirements are related. The cation specificities for activity and for stability-integrity are different indicating that these are separate Ca2+-dependent functions.     

Formation of a hexagonal lattice structure by an R-form lipopolysaccharide of Klebsiella: effect of various divalent cations on the lattice formation

The R-form lipopolysaccharide from Klebsiella pneumoniae strain LEN-111 (O3-:K1-), from which cationic material had been removed by electrodialysis, was previously shown to form a hexagonal lattice structure with the lattice constant of 14 to 15 nm when suspended in 50 mM tris(hydroxymethyl)aminomethane buffer at pH 8.5 containing 10 mM Mg2+. Under this experimental condition, effects of other divalent metal cations on the hexagonal assembly of the electrodialyzed LPS were compared with that of Mg2+. The Zn2+, Hg2+, Cu2+, and Ni2+ could produce essentially the same hexagonal lattice structure with the lattice constant of 14.5 to 15.0 nm as that formed with Mg2+. The Cd2+, Co2+, and Fe2+ produced the hexagonal lattice structure with the lattice constant of 15.5 to 16.0 nm, and Ba2+, Sr2+, and Ca2+ produced that with the lattice constant of 18 to 19 nm. In addition, the hexagonal lattice structures formed with the latter three cations were less orderly than those formed with the other cations. When the higher concentrations of Ba2+, Sr2+, and Ca2+ were used, the lattice constants were not shortened. The length of lattice constants of the hexagonal lattice structures formed with the divalent cations did not relate to the quantity of the cations bound to the LPS. Among the divalent cations tested, Hg2+ was bound to the LPS in the smallest amount (its atomic ratio to P, 0.07), and Zn2+ and Fe2+ were bound in very large amounts (their atomic ratios to P, 2.94 and 8.28, respectively).     

Comparative structural aspects of cation binding to phosphatidylserine bilayers

X-ray diffraction data recorded for monovalent and divalent cation complexes of a series of phosphatidylserines (PS) varying in chain length reveal a simple structural pattern. Only two bilayer structural types differing in hydrocarbon chain tilt but with similar polar group conformations are observed for (i) anhydrous acidic PS, (ii) anhydrous K+-PS, and (iii) Li+, Mg2+, Ca2+, Sr2+, Ba2+, and Pr3+ complexes of 'hydrated' PS. The X-ray diffraction data suggest that PS becomes dehydrated on complexing with Li+, Mg2+, Ca2+, and other divalent cations and adopts either the chain untilted (form I) or tilted (form II) bilayer structure.     

Effect of cations and temperature on kinetics of desmin assembly.

Smooth-muscle desmin, which was isolated from avian gizzard, was purified and used to form reconstituted intermediate filaments. Filament assembly was done in the presence of physiological cations, Ca2+, Mg2+, Na+, and Na+ plus Mg2+, and with non-physiological cations Cu2+ and Ni2+. Assembly was done at 2 degrees, 22 degrees and 37 degrees C, and was monitored by absorbance and by electron microscopy. Absorbance increased most rapidly during the first 2-5 min and then increased at a slower rate with the physiological cations, but decreased after that time with the non-physiological cations. For each physiological cation, absorbance increased with increasing temperature. This was particularly evident with Ca2+, which produced the lowest absorbance at 2 degrees C and the highest at 37 degrees C. When ionic strength was comparable, filament-forming buffers that contained bivalent cations were associated with higher absorbance values. Filament diameters were significantly smaller 60 min after assembly initiation than after 5 min. Average filament diameters, when formed in the presence of Cu2+ or Ni2+, were 10% greater than in the presence of the physiological cations and did not show a consistent tendency to decrease as time increased. These results demonstrate the importance, not only of the pH and ionic composition of the filament-forming buffer, but also of the temperature and duration of dialysis for reconstitution of desmin filaments.     

Characterization of calcium binding to brush-border membranes from rat duodenum.

The Ca2+-binding properties of isolated brush-border membranes at physiological ionic strength and pH were examined by rapid Millipore filtration. A comprehensive analysis of the binding data suggested the presence of two types of Ca2+-binding sites. The high-affinity sites, Ka = (6.3 +/- 3.3) X 10(5) M-1 (mean +/- S.E.M.), bound 0.8 +/- 0.1 nmol of Ca2+/mg of protein and the low-affinity sites, Ka = (2.8 +/- 0.3) X 10(2) M-1, bound 33 +/- 3.5 nmol of Ca2+/mg of protein. The high-affinity site exhibited a selectivity for Ca2+, since high concentrations of competing bivalent cations were required to inhibit Ca2+ binding. The relative effectiveness of the competing cations (1 and 10 mM) for the high-affinity site was Mn2+ approximately equal to Sr2+ greater than Ba2+ greater than Mg2+. Data from the pH studies, treatment of the membranes with carbodi-imide and extraction of phospholipids with aqueous acetone and NH3 provided evidence that the low-affinity sites were primarily phospholipids and the high-affinity sites were either phosphoprotein or protein with associated phospholipid. Two possible roles for the high-affinity binding sites are suggested. Either high-affinity Ca2+ binding is involved with specific enzyme activities or Ca2+ transport across the luminal membrane occurs via a Ca2+ channel which contains a high-affinity Ca2+-specific binding site that may regulate the intracellular Ca2+ concentration and gating of the channel.     

Na+/Ca2+ antiport in cultured arterial smooth muscle cells. Inhibition by magnesium and other divalent cations

Cultured smooth muscle cells from rat aorta were loaded with Na+, and Na+/Ca2+ antiport was assayed by measuring the initial rates of 45Ca2+ influx and 22Na+ efflux, which were inhibitable by 2',4'-dimethylbenzamil. The replacement of extracellular Na+ with other monovalent ions (K+, Li+, choline, or N-methyl-D-glucamine) was essential for obtaining significant antiport activity. Mg2+ competitively inhibited 45Ca2+ influx via the antiporter (Ki = 93 +/- 7 microM). External Ca2+ or Sr2+ stimulated 22Na+ efflux as would be expected for antiport activity. Mg2+ did not stimulate 22Na+ efflux, which indicates that Mg2+ is probably not transported by the antiporter under the conditions of these experiments. Mg2+ inhibited Ca2+-stimulated 22Na+ efflux as expected from the 45Ca2+ influx data. The replacement of external N-methyl-D-glucamine with K+, but not other monovalent ions (choline, Li+), decreased the potency of Mg2+ as an inhibitor of Na+/Ca2+ antiport 6.7-fold. Other divalent cations (Co2+, Mn2+, Cd2+, Ba2+) also inhibited Na+/Ca2+ antiport activity, and high external potassium decreased the potency of each by 4.3-8.6-fold. The order of effectiveness of the divalent cations as inhibitors of Na+/Ca2+ antiport (Cd2+ greater than Mn2+ greater than Co2+ greater than Ba2+ greater than Mg2+) correlated with the closeness of the crystal ionic radius to that of Ca2+.     

Role of cations in the flocculation of Saccharomyces cerevisiae and discrimination of the corresponding proteins

Asexual yeast flocculation was studied using strong flocculents of Saccharomyces cerevisiae. The inhibitory effect of cations on flocculation is considered to be caused by competition between those cations and Ca2+ at the binding site of the Ca(2+)-requiring protein that is involved in flocculation. Inhibition of flocculation by various cations occurred in the following order: La3+, Sr2+, Ba2+, Mn2+, Al3+, and Na+. Cations such as Mg2+, Co2+, and K+ promoted flocculation. This promoting effect may be based on the reduction of electrostatic repulsive force between cells caused by binding of these cations anionic groups present on the cell surface. In flocculation induced by these cations, trace amounts of Ca2+ excreted on the cell surface may activate the corresponding protein. The ratio of Sr2+/Ca2+ below which cells flocculated varied among strains: for strains having the FLO5 gene, it was 400 to 500; for strains having the FLO1 gene, about 150; and for two alcohol yeast strains, 40 to 50. This suggests that there are several different types of cell surface proteins involved in flocculation in different yeast strains.     

Chelation of divalent cations by ATP, studied by titration calorimetry.

Thermodynamic parameters and stoichiometry for the formation of complexes of ATP with Mg2+, Ca2+, and Sr2+ were determined by titration calorimetry. In each case, 1:1 stoichiometry was observed and complex formation was entropy driven. Binding constants for formation of complexes decreased in the order of Mg2+ greater than Ca2+ greater than Sr2+, as expected from charge density considerations. Monovalent cations hindered complex formation with Mg2+, apparently by competing with the divalent cation for complexation with ATP. Analysis of this competitive effect provided estimates of the binding constants for complexes of ATP with monovalent cations, which decreased in the order expected from charge density considerations (Li+ greater than Na+ greater than K+).     

A cation channel for K+ and Ca2+ from Tetrahymena cilia in planar lipid bilayers

The single-channel properties for monovalent and divalent cations of a voltage-independent cation channel from Tetrahymena cilia were studied in planar lipid bilayers. The single-channel conductance reached a maximum value as the K+ concentration was increased in symmetrical solutions of K+. The concentration dependence of the conductance was approximated to a simple saturation curve (a single-ion channel model) with an apparent Michaelis constant of 16.3 mM and a maximum conductance of 354 pS. Divalent cations (Ca2+, Ba2+, Sr2+, and Mg2+) also permeated this channel. The sequence of permeability determined by zero current potentials at high ionic concentrations was Ba2+ greater than or equal to K+ greater than or equal to Sr2+ greater than Mg2+ greater than Ca2+. Single-channel conductances for Ca2+ were nearly constant (13.9 pS-20.5 pS) in the concentrations between 0.5 mM and 50 mM Ca-gluconate. In the experiments with mixed solutions of K+ and Ca2+, a maximum conductance of Ca2+ (gamma Camax) and an apparent Michaelis constant of Ca2+ (K Cam) were obtained by assuming a simple competitive relation between the cations. Gamma Camax and K Cam were 14.0 pS and 0.160 mM, respectively. Single-channel conductances in mixed solutions were well-fitted to this competitive model supporting that this cation channel behaves as a single-ion channel. This channel had relatively high-affinity Ca2+-binding sites.     

Association-dissociation equilibria of Octopus hemocyanin

The equilibria between the native (decameric) Octopus hemocyanin and its subunits were studied by analytical sedimentation. Equilibrium is obtained slowly, but the reaction is thermodynamically reversible. The mass action law for a monomer-decamer reaction is obeyed. The reassociated hemocyanin is virtually identical in its sedimentation behavior and oxygen binding with the native protein. The association-dissociation equilibria are mediated by cations; Mg2+, Ca2+, Na+, and H+ are all effective in stabilizing the decameric form at appropriate concentrations. About three to four cations per monomer must be bound for association to occur. Under some conditions, dimers of the subunits can be observed, but formation of this dimer does not depend on cation concentration, and it does not appear to be an obligate intermediate in the association to decamer.