ABSTRACT
The large axons in the cerebro-visceral connective have been shown to function for appreciable periods in preparations bathed in sodium-free non-electrolyte solutions.
The results of experiments on the effects of organic monovalent cations and anions, together with observations on the effects of tetrodotoxin, procaine and manganous ions and the changes in conduction velocity in tris chloride and dextran solutions indicate that the action potentials are, nevertheless, mediated by conventional sodium-dependent mechanisms.
Radioisotope experiments show that there is a small fraction, of approx. 0.5 mM/kg. tissue, which does not exchange rapidly with the 22Na in the bathing medium and which can be depleted by stimulation in sodium-free solutions.
On the basis of these observations it is suggested that there is sequestered extra-axonal sodium fraction which can be utilized by the large axons to maintain action potentials in preparations bathed in sodium-free solutions.
INTRODUCTION
Previous communications from this laboratory have been concerned with an attempt to elucidate the ionic basis of action potential production in the central nervous system of Anodonta cygnea. This eulamellibranch mollusc was chosen for these studies as it appears to be the possessor of the most dilute blood so far described in the Animal Kingdom (Potts, 1954) and, for this reason, presents some intriguing problems in the interpretation of axonal function. Despite the low sodium concentration of the blood it was found, nevertheless, that the action potentials appeared to be sodium-dependent (Treherne, Mellon & Carlson, 1969), an observation which can be correlated with the extremely low intracellular concentration of this cation in the central nervous tissues (Mellon & Treherne, 1969). It was also concluded, on the basis of both electrophysiological and radioisotopic evidence, that the axon surfaces were relatively accessible to small water-soluble ions and molecules in the bathing medium. These conclusions accorded with ultrastructural observations which indicated that there were no visible structures which would be likely to restrict the access of small inorganic ions to the axon surfaces (Gupta, Mellon & Treherne, 1969).
The earlier experiments on the cerebro-visceral connective of Anodonta indicated that there was little regulation of the ionic composition of the fluid bathing the surfaces of the small axons. Exposure to sodium-free solutions, for example, resulted in the rapid development of a conduction block in the small axons (Treherne, Mellon & Carlson, 1969) in much the same way as in the connectives in the leech central nervous system (Nicholls & Kuffler, 1964) where there appears to be a ready access to the axon surfaces via the narrow intercellular channels formed by the close apposition of adjacent nerve cell membranes (Kuffler & Potter, 1964). With the larger axons in the connective of Anodonta (i.e. in the range of 2·0 to 6·0 μ) there was, however, abundant evidence of their ability to function in sodium-free solutions (Treherne, Mellon & Carlson, 1969). It was not clear, however, from the previous investigations whether this resulted from a regulation of the extra-axonal sodium level or from the fact that the larger axons were able to utilize some other cation or anion species to carry the inwardly directed action current when bathed in sodium-free solutions. The present investigation was, therefore, initiated in an attempt to elucidate the ionic basis of action potential production in the large axons of the cerebro-visceral connectives.
METHODS
The techniques of external recording of the compound action potential were essentially similar to those reported in Treherne, Mellon & Carlson (1969) in which the isolated connective was immersed in a chamber of five compartments. Only those techniques which differed from ones described previously will be outlined here.
Isolated connectives were ligatured at each end with short lengths of hair, immersed in various solutions and weighed on a 5 mg. torsion balance. Prior to weighing, the connective was carefully blotted on filter paper.
In order to study sodium exchange, isolated, paired connectives were immersed in solutions containing 5 mM-CaCl2 and varying concentrations of NaCl. The solutions were adjusted to 42 m-osmoles with sucrose. One connective of each pair was soaked in a solution containing unlabelled Na. The other connective was placed in a solution containing 22Na. After soaking, the connectives were weighed and then ashed in a muffle furnace on platinum foil at 460−8°C. for over 6 hr. Total sodium concentration in the unlabelled connective was measured using the Unicam SP 900 flame spectrophotometer. A check was made to ensure that quantitative recoveries of added amounts of sodium were obtained. The radio-activity of the labelled connective was determined using an end-window counter (Mullard MX 123).
RESULTS
The effects of monovalent cations
Figure 1 shows the effects of four different cations on the fast action potentials recorded in isolated cerebro-visceral connectives. The cations were added as chloride salts, at a concentration of 21 mM/l., with no other inorganic ions or molecules in the solutions. It will be seen that the organic cations, tris and choline, caused a rapid loss of function. There was, however, a partial recovery on replacement of these cations with an isotonic sucrose solution. A full recovery was achieved when the connectives were bathed in normal Ringer solution.
Action potentials were recorded for extended periods in connectives bathed in a solution of 21 mM/l. lithium chloride. Tetra-ethyl-ammonium (TEA), on the other hand, resulted in a conduction block. Axonal function, in preparations bathed with TEACI, did, however, persist for longer periods than with either tris or choline chloride solutions.
The effects of divalent cations
In preparations maintained in a solution of 4·7 mM/l. CaCl2, in which isotonicity was maintained by the addition of sucrose, the fast action potentials persisted, but with an appreciable decline in conduction velocity (Fig. 2). Addition of 9·3 mM/l. MgCl2 to the CaCl2, in place of sucrose, resulted in some decline in spike amplitude and conduction velocity. With 9·3 mM/l. MnCl2, added to 4·7 mM/l. CaCl2, there was an appreciable decline in conduction velocity which was also associated with a decrease in spike height. Addition of 14·0 mM/l. tris chloride to a solution containing 4·7 mM/l. CaCl2 resulted in a rapid loss of function, an effect which was reversed by the substitution of normal Ringer solution.
Experiments were carried out to test the effects of the substitution of anions on the fast compound action potentials in isolated cerebro-visceral connectives. The results showed that axonal function was maintained in 14·0 mM/l. Na2SO4, but rapidly declined in tris sulphate or acetate (Fig. 3).
In a previous investigation it was shown that the fast action potentials in the cerebro-visceral connective were rapidly abolished by the addition of dilute tetrodotoxin to the bathing solution (Treherne, Mellon & Carlson, 1969). The possibility existed, however, that the sodium specificity might be altered in sodium-free solutions, an effect which might be reflected in a reduced sensitivity to tetrodotoxin. To test this possibility the effects of tetrodotoxin on fast action potentials were observed in preparations maintained in normal Ringer solution and in isotonic sucrose solution.
It was found that there was an increased sensitivity of the fast axons to tetrodotoxin in preparations maintained in isotonic sucrose solution, a blocking effect being observed at a concentration 5×10-8 M TTX (Fig. 4). In all cases there was a rapid return of axonal function on return of the axons to normal isotonic sucrose solution.
Effects of procaine
The fast action potentials showed a decline, in preparations bathed in isotonic sucrose containing 10 mM/l. procaine, with the development of a conduction block within 60 sec. (Fig. 5). The decline in the amplitude of the compound action potentials was also accompanied by an appreciable drop in conduction velocity. Substitution of normal Ringer for sucrose, in the presence of procaine, resulted in a rapid return of axonal function.
The effects of the composition of the bathing fluid on conduction velocity
Figure 6 shows the effects of the substitution of tris and choline chloride for NaCl on the conduction velocity of the fast action potentials. It will be seen that in the absence of sodium in the Ringer solution there was an appreciable drop in conduction velocity. This effect was more marked than that observed when the preparation was bathed with isotonic sucrose solutions.
The rapid decline in conduction velocity of the fast action potentials observed in isotonic sucrose solution was not paralleled by a similar effect in isotonic dextran (10,000 M.W., Fig. 7).
The conduction velocity of the fast action potentials was measured at several sodium concentrations in solutions in which isotonicity was maintained either with dextran or tris chloride. The results of these experiments are summarized in Fig. 8. In this graph the relative conduction velocity (θtest/θnormal) is plotted against the square root of the relative sodium concentration . This graphical form enables a comparison to be made with data for Carcinus (Katz, 1947) and Loligo axons (Hodgkin & Katz, 1949). The data for Loligo axons was calculated from measurements made on the maximum rate of rise of the action potentials in test solutions of varying sodium concentration. These calculations were based on the assumption that, in a simplified theoretical system, the conduction velocity can be related to the square root of the rate of rise of the action potential (Hodgkin & Katz, 1949). In both the experiments on Carcinus and Loligo the sodium concentrations of the test solutions were adjusted by dilution with sucrose.
It will be seen that there is a profound difference between the relation of relative conduction velocity and the square root of the relative sodium concentration for the Anodonta fast axons depending upon whether the sodium was diluted with dextran or tris chloride. When isotonicity was maintained using dextran there was very little decline in conduction velocity over the entire range of sodium concentrations. With tris chloride, on the other hand, the conduction velocity showed a steep decline at reduced concentration of external sodium ions. The form of the curve obtained for sodium diluted with tris chloride does, in fact, more closely correspond to those obtained with Carcinus and Loligo axons.
A comparison was also made of the conduction velocities, measured in fast axons, in preparations maintained in isotonic sucrose solution and in a sucrose solution containing 4·7 mM/l. CaCl2 (Fig. 9). It will be seen that there was a decline in conduction velocity in both sucrose and sucrose-calcium solutions, the reduction being more marked in the former solution.
Weight changes observed in various experimental solutions
Experiments were carried out to determine the extent of any changes in weight of ligatured isolated connectives which would result from any alterations in water content produced by the various experimental solutions used in this investigation.
Figure 10 shows that significant changes in weight could be measured in preparations maintained in normal Ringer solution or in isotonic sucrose. There was a significant reduction in weight, to about 85 % of that of connectives maintained in blood, in normal Ringer solution. There was also an increase in the weight of connectives bathed in isotonic sucrose solution. The latter effect was, however, abolished if the sucrose-calcium solution was used.
With 21·0 mM/l. tris chloride there was no appreciable weight change in isolated ligatured connectives, although a slight reduction was observed if 4·7 mM/l. CaCl2 was used to replace part of the tris chloride (Fig. 11). Transfer to sucrose solution resulted in an appreciable increase in weight of the connectives.
Weight changes were also observed in solutions of varying dextran concentration. Figure 12 shows the changes resulting from the bathing of isolated connectives in dextran solutions at three different concentrations. It will be seen that at 42·0 mM/l. there was a dramatic reduction in weight, with a smaller reduction at 21·0 mM/l. and a slight increase at 10·5 mM/l. Figure 13 shows the relation between the weight and the concentration of dextran in the bathing solution. It is apparent from these results that the normal weight (i.e. that measured in preparations in blood) is maintained at a dextran concentration of approx. 14·0 mM/l.
The weight changes in sucrose solution obtained with intact ligatured connectives were also compared with those in split connectives. Figure 14 shows that, as in intact connectives, the split connectives showed an initial rise on being placed in 42·0 mM/l. sucrose solution. After approx. 5 min., however, there was a progressive reduction in weight which continued on return of these connectives to blood.
Axonal function in split connectives
Experiments were carried out to determine whether the ability of the fast axons to function in sodium-free solutions depended upon the presence of the peripheral nerve sheath. As it was found to be impossible to de-sheath the connectives in a conven-tional way, by pulling the sheath away using finely ground watchmaker’s forceps, a preparation was used in which connectives were split longitudinally using a sharpened needle. This preparation corresponds to the one referred to in the preceding section. It was found that the fast axons continued to function for appreciable periods in these preparations when soaked in 42·0 mM/l. dextran solution (Fig. 15).
Effect of stimulation régime on axonal function
In the previous experiments the preparations were only stimulated intermittently when action potentials were being actually recorded. Additional experiments were, however, also carried out to test the effects of continuous stimulation on axonal function. Figure 16 A shows the whole compound action potential recorded from an isolated cerebro-visceral connective recorded at slow sweep speed. Figure 16 B and C show the same action potential, at twice the gain, following stimulation for 90 min. and 180 min. respectively. The effect of stimulation (0·5 sec.-1; 3·0 msec, duration) was to cause a reduction of the large slowly conducting component of the compound action potential. The rapidly conducting fibres did not, however, show any appreciable diminution following continuous stimulation in blood.
Extent of 22Na exchange in isolated connectives
Earlier experiments, with 22Na and 24Na, indicated that there was a fairly complete exchange of labelled sodium with that contained in the tissues of the connective (Mellon & Treherne, 1969). However, the extent of individual variation in these earlier experiments made it difficult to identify any relatively small, slowly-exchanging fractions. An additional series of experiments was, therefore, carried out including some in which the labelled sodium was present at relatively low concentration in the bathing medium. The results of these experiments are summarized in Table 1. It will be seen that, after 3 hr. soaking, the specific activity of the tissue sodium was lower than that in the bathing medium. This is particularly apparent at the lowest concentration, 1·23 mM/l. Na, where the ratio of the specific activities in the tissue and the bathing medium averaged 0·58. Using the values for sodium concentrations of 1·23 and 4·12 mM/l. it can be calculated that the non-exchanging sodium, after a 3 hr. period, amounted to 0·51 and 0·54 mM/kg. (wet weight) respectively.
Effects of stimulation on sodium content of tissues
Experiments were carried out to determine the effects of continuous stimulation on the sodium content of connectives maintained in sucrose-calcium solutions. Paired connectives, taken from the same animals, were used in these determinations. In each case one ligatured connective was stimulated in the conventional nerve chamber whilst the other was soaked for an equivalent period in sucrose-calcium solution without stimulation. The results of this experiment are summarized in Table 2. It is apparent from these data that the sodium content of unstimulated nerve was relatively low, but that stimulation caused a further depletion for in each case the sodium content of the stimulated connective was lower than that of the control one. By subtraction of the mean values of the stimulated from the unstimulated connectives it is possible to arrive at a figure for the depletion of sodium of 0·17 mM/kg. tissue. This depletion was produced by a stimulus régime of 0·5 sec.-1 for a period of 45 min. Conduction processes in the connective had effectively ceased, however, within 30 min. (see Fig. 17 B).
DISCUSSION
A primary aim of the present investigation was to contribute to an understanding of the ionic basis of the action potentials in the rapidly conducting axons of the cerebrovisceral connective. In a previous paper it was shown that, in contrast to the small axons, the large ones (in the range of approx, 2μ0−6μ0 μ) were capable of functioning for appreciable periods in preparations bathed in sodium-free solutions (Treherne, Mellon & Carlson, 1969). The ability to function in the absence of external sodium ions did not appear to result from any significant restriction in access of small watersoluble ions and molecules to the axonal membranes. This was shown in both electrophysiological experiments (Treherne, Mellon & Carlson, 1969) and in those using radiosodium to follow the exchanges of this cation with that in the bathing medium (Mellon & Treherne, 1969). This physiological evidence was in essential agreement with the electron microscopic picture of the central nervous tissues, which showed that there were no visible structures which would be likely to restrict the intercellular movements of inorganic cations between the external medium and the extra-axonal fluid (Gupta, Mellon & Treherne, 1969).
Despite their ability to function in sodium-free solutions the rapidly-conducting axons appeared, nevertheless, to be conventional in that the action potentials appeared to be mediated by sodium-dependent mechanisms. This was shown in experiments in which a conduction block was produced by the addition of dinitrophenol or ouabain to the normal Ringer solution bathing the connectives. A rapid return of axonal function in these poisoned preparations was produced by elevation of the sodium concentration, but not by an increase in the calcium level in the bathing medium (Treherne, Mellon & Carlson, 1969). Furthermore, it was also demonstrated that the fast action potentials were blocked by dilute tetrodotoxin, a substance which has been shown to be a specific inhibitor of the sodium conductance channels in the active axon membranes of crustaceans (Narahashi, Moore & Scott, 1963) and cephalopods (Nakamura, Nakajima & Grundfest, 1964).
The above results do not, however, constitute conclusive evidence that the large axons continue to utilize sodium ions in action potential production in the absence of this cation in the bathing medium. It could be postulated, for example, that in the active state the axons switch from a sodium-dependent mechanism to one involving a net efflux of anions from the axoplasm in preparations bathed in sodium-free solutions of non-electrolytes. In such a situation it would be expected that axonal function would be blocked by the presence of the particular anion at the axon surfaces under sodiumfree conditions. The observations, from the present investigation, that solutions of tris or choline chloride failed to sustain the fast action potentials would accord with this hypothesis, if the action current was mediated by chloride ions. However, it was found that these action potentials were also blocked by solutions of tris acetate and sulphate, but were sustained by a bathing solution of 14·0 mM/l. sodium sulphate. It seems reasonable, therefore, to reject the hypothesis that the action current in the fast axons results from an anion efflux in sodium-free media.
An alternative hypothesis, to account for the ability of the fast axons to function in the absence of external sodium ions, would be that the active axonal membranes alter their cation specificity in sodium-free solutions so as to utilize some other cation species to carry the inwardly directed action current. By analogy with suggestions which have been made for some gasteropod neurones (Gerasimov, Kostyuk & Maiskii, 1965; Kerkut & Gardner, 1967) calcium would be the most likely cation to be so involved. This latter hypothesis does not, however, accord with the experimental results obtained in this investigation. The fact that the fast action potentials fail, at normal calcium levels, in the presence of tris chloride does not support this hypothesis. Furthermore, if axonal function in sodium-free media was produced by a change in cation specificity then it would be reasonable to expect that the axons would become less sensitive to tetrodotoxin molecules. The present results showed, however, that the tetrodotoxin sensitivity increased in sucrose solutions which again militates against the hypothesis that the fast axons switch from a sodium-dependent mechanism to one involving some other cation such as calcium.
As in the previous investigation (Treherne, Mellon & Carlson, 1969) the balance of the evidence accumulated here indicates that the action potentials associated with the fast axons are sodium-dependent, even in preparations bathed in sodium-free solutions of non-electrolytes. Further evidence for sodium-mediated fast action potentials was obtained in the experiments which showed that lithium would support action potentials, for this cation is known to be able to substitute for sodium in conventional excitable systems (c.f. Hodgkin, 1951; Keynes & Swan, 1959). Similarly, the difference obtained between the behaviour of the axons in tris and TEA solutions would seem to support the concept of a sodium dependency of the action potential. Thus TEA, which is known to act as a sodium substitute in some excitable cells (cf. Tasaki, 1968) maintained the fast action potentials for longer than either tris or choline. The eventual decline in excitability in TEA can be attributed to the effect which this cation has in other neurones of reducing the resting potential (Tasaki, 1968).
The observations on the changes in conduction velocity of the fast action potentials are of interest in interpreting the ionic basis of excitability in the larger axons in the connective of Anodonta. For example, in a previous communication (Treherne, Mellon & Carlson, 1969) and also in the present investigation it was noted that there was a rapid reduction in conduction velocity in the fast axons in preparations bathed in sucrose solutions. Now there are two fairly obvious ways by which such an effect could have been produced. First, with axons in which the volume of the external conducting fluid is small, as would be the case with the extracellular fluid in the connectives, the conduction velocity would be expected to be proportional to , where ri is the axoplasmic resistance and re the resistance of the extracellular fluid per unit length (Hodgkin, 1954). Thus any reduction in re, caused by dilution of the extracellular ions, would produce a decline in conduction velocity. Secondly, it would be reasonable to suppose that, as in the squid axon, the conduction velocity could be related to the square root of the rate of rise of the action potential (Hodgkin & Katz, 1949). As the rate of rise of the action potential decreases with decreasing sodium concentration, then it follows that the conduction velocity would decline with decreasing extraaxonal sodium levels.
The present results have shown that there is a rapid increase in weight in connectives maintained in 42 · 0 mM/l. sucrose solution. This effect appeared to result from an extracellular uptake of water, for no sustained increase in weight was obtained in split connectives. It is clear, therefore, that such an extracellular swelling, with consequent dilution of the extra-axonal ions, would cause a reduction in conduction velocity by either of the mechanisms outlined in the preceding paragraph. An appreciable reduction in conduction velocity was, in fact, observed in 42·0 mM/l. sucrose (Treherne, Mellon & Carlson, 1969). In 42·0 mM/l. dextran, on the other hand, in which there was a slight reduction in fluid volume within the nervous tissues, there was no reduction in conduction velocity of the fast axons. The correlation between conduction velocity and extracellular volume was, however, not complete, for in sucrose-calcium solution (in which there was no appreciable increase in fluid volume within the connectives) there was, nevertheless, a decrease in conduction velocity.
It is apparent from the above results that there was a profound difference in the effects of sodium-free solutions of electrolytes and non-electrolytes on the excitability of the fast axons. Thus solutions of tris or choline chloride caused the development of a conduction block within a few seconds, whereas in dextran solution the fast axons continued to function for several hours. A difference was also evident in the relation between conduction velocity and the relative sodium concentration of the bathing medium according to whether isotonicity was maintained with tris chloride or dextran. In solutions in which the osmotic concentration was maintained with dextran there was little reduction in conduction velocity with decreasing sodium concentration in the bathing medium. In solutions containing tris chloride, on the other hand, there was a decline in conduction velocity with decreasing sodium concentration. With the latter solutions the fast axons behaved more nearly like conventional axons such as those of Carcinus (Katz, 1947) and Loligo (Hodgkin & Katz, 1949).
It seems clear from the above evidence that in dextran solution sodium ions must have been present in the fluid bathing the axon surfaces in sufficient concentration to maintain axonal function. It also seems reasonable to suppose that these sodium ions must have been displaced by the organic cations in solutions of tris and choline chloride. It would follow from this that the sodium ions in the immediate vicinity of the fast axons were likely to have been associated with indiffusible or very slowly diffusing anions, from which they were not displaced in non-electrolyte solutions. It may be significant in this respect that there appears to be an osmotically active fraction present in the extracellular fluid which does not readily diffuse into the bathing medium. This supposition is made on the basis of the observed swelling obtained in the sucrose solution and from the fact that an apparently normal fluid volume can be maintained in 14·0 mM/l. dextran solution. It also seems clear from the above results that the sodium present at the surfaces of the fast axons, in connectives bathed in sodium-free non-electrolyte solutions, does not represent the total sodium available to maintain the activity of these axons. It was shown, for example, that on replacement of tris chloride by a non-electrolyte there was a rapid return of the fast action potentials. It follows, therefore, that the sodium which was available to maintain the activity of the fast axons in non-electrolyte solutions must have been present in two forms: an immediately accessible fraction, which could be displaced by small diffusible cations such as tris, and an additional fraction, which was not immediately available but could be mobilized to sustain axonal function.
According to the latter hypothesis it might be expected that there would be a sequestered sodium fraction close to the membranes of the fast axons. Such a fraction might also be expected to form a relatively small proportion of the total tissue sodium in view of the relatively small number of fast axons in the connective (Gupta, Mellon & Treherne, 1969; Treherne, Mellon & Carlson, 1969). Previous experiments using 22Na and 22Na showed, however, that there was a fairly complete exchange of labelled sodium with that contained in the tissues of the connectives (Mellon & Treherne, 1969). These latter experiments were carried out using bathing solutions containing the labelled sodium at the normal blood concentration of 15·0 mM/l. (Potts, 1954). When, as in the present investigation, the sodium exchange was followed at relatively low external sodium levels it was found that there was a small fraction, of around 0·51 mM/kg. tissue, which did not exchange with the cations in the bathing medium after 3 hr. There is then some evidence for a small sequestered sodium fraction within the central nervous tissues which might be available to maintain action potential in the fast axons in sodium-free solutions. In postulating such a system it would be necessary to demonstrate that such a sodium fraction could be depleted by stimulation of the fast axons in non-electrolyte solutions. Such an effect was, in fact, demonstrated in the present investigation when it was shown that a continuous stimulation régime resulted in a significant decline in the sodium content, as compared with unstimulated connectives maintained in sodium-free solutions. Stimulation, at 0 · 5 sec.-1, for 30 min. resulted in a mean decline of 0·17 mM/kg. tissue.
It was also shown in this investigation that although fast action potentials could be elicited for several hours in non-electrolyte solutions, with intermittent stimulation, axonal function ceased after approx. 30 min. with continuous stimulation. Such a result can be interpreted in terms of the depletion of the sodium store postulated in the preceding paragraph. It should also be noted that continuous stimulation, in preparations bathed in blood, resulted in a rapid decline in the slow action potentials associated with the small axons, whereas the fast action potentials persisted for extended periods. This state of affairs could have resulted from a proportionally greater rise in the intracellular sodium level resulting from the stimulation of the small axons which possess a very much larger surface/volume ratio than the large fast axons.
It is now relevant to consider the question of the amount of sodium which would be required to maintain the fast action potentials in non-electrolyte solutions. It was shown that the sequestered sodium store amounted to approx. 0·51 mM/kg. tissue. It was also demonstrated that with continuous stimulation, at 0·5 sec.-1, the fast action potentials persisted for 30 min. and resulted in a depletion, of 0·17 mM/kg., of the sequestered sodium fraction. Now the majority of the large axons in the cerebro-visceral connective were found to be in the range 2·0−4·0μ in diameter. With a sodium influx of 3·5 × 10-12 M cm-2 impulse-1 (cf. Hodgkin, 1958) this would correspond to a net influx of 2·2−4·4×10-15 M impulse-1 per unit length of axon. Given that a 1·0 cm. length of connective averages 1·13 mg. and that there are approx. 70 large axons in a connective (Treherne, Mellon & Carlson, 1969) then this would correspond to a total sodium utilization of 0·12−0·23 mM/kg. tissue for a stimulation régime of 30 min. at 0·5 sec-1. This estimated figure is less than the value of 0·51 mM/kg. for the sequestered sodium in the connective. Comparison of the estimated value, of 0·12−0·23 mM/kg., with that for the total sequestered sodium would suggest that the utilization of the released sodium would be extremely efficient. In fact, the figure of 0·17 mM/kg. (obtained for the depletion of sodium from the connectives following stimulation for 30 min.) would imply that the amount of the released sodium diffusing into the bathing medium via the extracellular spaces was approximately equivalent to that entering the axons themselves. However, it should be borne in mind that the calculated value for sodium uptake by the large axons is likely to represent an outside estimate. The above calculations assumed, for example, that the whole population of large axons functioned for the entire period of stimulation and that there was a maximum sodium influx with each action potential for the whole of the period of stimulation. Furthermore, any local resorption of released sodium ions would affect the above comparisons and would lead to an incorrectly high estimate of the proportion of the released sodium utilized in carrying the action current in the large axons.
Finally, it remains to consider the possible localization and mobilization of the sequestered sodium fraction postulated above. One possibility would be that the extracellular sodium is associated with the anion groups of relatively large molecules which are retained by the peripheral nerve sheath. This possibility would not, however, accord with the observation that the fast action potentials persisted in split connectives maintained in dextran solution. An alternative hypothesis would be that the sequestered sodium fraction is contained in the glia. However, electronmicroscopic investigation has shown that there is an extremely sparse distribution of glial elements in the cerebro-visceral connectives of Anodonta and that the axon surfaces are, for the most part, not separated by glial folds (Gupta, Mellon & Treherne, 1969). This state of affairs largely applies to the small axons (less than 0·5 μ in diameter) which form the greater part of the volume of the connective. The large axons on the other hand show a much more frequent association with glial elements (Treherne, Carlson & Gupta, 1969). This association takes the form of a close apposition (i.e. to within 100−200 Å) of a glial membrane with a portion of the axon surface. The remainder of the axon surface is surrounded by the closely applied membranes of other axons which also delimit a 100−200 Å space around the axon surface. This glial association would seem to constitute the most likely system which could maintain a small sodium store in the regions of the surfaces of the larger axons. Thus, if the sequestered sodium fraction was contained in the glial elements, then any release of the cations would take place into the extremely restricted space formed by the closely applied membranes of the adjacent glial cell and axons, so as to lead to a maximal sodium concentration in the extra-axonal fluid.
It has been suggested in a preceding paragraph that the sodium in the immediate vicinity of the large axons may be present in two fractions : an immediately available one, which can be displaced by organic cations, and a sequestered one, which is not immediately depleted by such cations. If this suggestion is correct then it would seem reasonable to identify the latter sodium fraction with that contained in the glial elements associated with the large axons. The immediately available sodium fraction would, according to this hypothesis, be most likely to be associated with indiffusible or slowly diffusing extracellular anions in the region of the axon surfaces.
The hypothetical mechanism proposed above is strikingly similar to the revised hypothesis which has been recently advanced by Chamberlain and Kerkut (1967, 1969) to explain the apparent absence of any involvement of calcium ions in carrying the inwardly-directed action current in the neurones of Helix aspersa. The following passage, taken from the paper by Chamberlain and Kerkut, could be equally well applied to the situation in Anodonta cygnea- ‘There may be a small store of sodium ions closely associated with the nerve membrane (either trapped in the glia or in the folds of the membrane, or even actually in the membrane structure) which cannot be easily removed by washing in sodium-free Ringer but which is affected by the calcium concentration’.
The conclusions which have been advanced in the present study on Anodonta differ somewhat from those which have been suggested in some previous studies on the ionic basis of axonal function in an insect species, Carausius morosus (Treherne & Maddrell, 1967; Treherne, 1967). The axons in the central nervous system of C. morosus, like the large ones in A. cygnea, can function for appreciable periods in sodium-free media, although the action potentials appear to be mediated by sodiumdependent mechanisms. On the basis of the apparently complete and rapid exchange of the sodium in the central nervous tissues with labelled cations in the blood it was suggested that the regulation of the extra-axonal sodium level might be achieved by a local re-cycling of the cation by the glial elements. The glial elements in the insect species which have been studied (cf. Smith & Treherne, 1963) form a much more complete covering of the axons than in Anodonta. However, the possibility cannot be eliminated that in the insect central nervous tissues the regulation of the extraaxonal sodium level might also be achieved by a similar mechanism to that which has been tentatively proposed for the fast axons of Anodonta in the present study.