The Relations Between Yolk and White in the Hen’s Egg: V. The Osmotic Properties of the Isolated Vitelline Membrane

In order to support the avian vitelline membrane between the two chambers of a dialysing apparatus so that no leaks should take place, it was necessary to devise a special apparatus. This is shown in Fig. 1. It consists of two glass chambers, A and B, which are held in position with their end-pieces and the diaphragm separating them, by a large G-clamp, C, C. This clamp, when fully screwed up, presses on the ends of the apparatus through the pieces of cork L, M, and the ends themselves are formed of two pieces of glass, N and O. These make contact with the chambers by means of the circular rubber washers P and Q, and have, projecting from their inner surfaces, a number of discs, RR, which are cemented to them and serve the purpose of reducing the total capacity of the chambers. The main openings of the chambers, S and T, are situated on the top ; through these enter the four conductivity electrodes, U₁-U₄ and the two air tubes, V_1,V₂, which can be used for stirring the solutions by bubbling. In addition to the main openings, N and T, there are two side openings, S₁ and T₁ which are invisible on the elevation, but one of which, T₁ appears on the transverse section. These side openings serve for introducing and removing solutions without interfering with the position of the electrodes, so that once the factor characteristic of each chamber has been found, the absolute conductivity can be calculated directly from the observed resistance. The side openings also serve for attachment to an aspirator system if it is desired to bubble a stream of air through the chamber.

The chambers are separated from one another (a) by two circular rubber washers, P₁ and Q₁ (6) by two glass discs, W and X, in which central holes have been cut, and (c) by two thick rubber discs, Y and Z, in which central holes have been cut, these holes being slightly smaller than those of the corresponding glass discs. The membrane itself is held between the two thick rubber discs, as indicated in Fig. I, and at its edges receives the full pressure of the clamp, transmitted along the sides of the chambers and through the rubber washers. After a little practice it was quite easy to obtain a degree of pressure of the clamp which was sufficient to avoid any leakage out of the chambers, and that leakage from one chamber to another, other than through the pores of the membrane, did not take place was easily shown by adding a small amount of a dye to one side. Thus in all the experiments recorded in this paper the absence of gross leaks was ascertained by adding a few drops of Congo red to one chamber at the end of the experiment. This dye exists in solution in the form of colloidal molecular aggregates too large to pass rapidly through the membrane itself, but yet able to indicate at once the presence of leaks due to the imperfect closure of the diaphragm.

It may be observed that the air tubes, V₁ and V₂, were bent round on entering the chambers so as not to pass between the electrodes, and not to deliver their stream between the electrodes. The electrodes themselves were of platinum and as nearly as possible 1 cm. square. They were coated with platinum black in the presence of a trace of lead until a uniform velvety surface was obtained. From the mercury in the electrode tubes, thick copper wires were led off to a reversing switch of the mercury cup type and so to an apparatus for determining conductivity, comprising Wheatstone bridge, variable resistance, buzzer, potentiometer, and telephone.

In assembling the apparatus for an experiment the following procedure was adopted. A fresh infertile egg was broken and the contents turned out into an evaporating basin. The white was removed and replaced by o-8 per cent. Ringer solution. The small germinal area was removed by scissors and forceps, and through the hole thus made a pipette, connected to a filter pump by means of a Buchner flask, was introduced. With this assistance, the greater part of the yolk was removed on the vacuum-cleaner principle, a procedure which was advantageous in that the surrounding solution remained clean and the manipulation of the yolk was consequently easier. The whole of the yolk cannot, however, be removed by this method as the vitelline membrane is liable to be torn if it is sucked up into the pipette, so the final cleaning stages were accomplished by directing a stream of Ringer solution into the interior through a wash-bottle delivery tube with a sharply curved end. Meanwhile a bowl of Ringer solution had been prepared, within which was placed a pile of Petri dishes and small crystallising dishes, so arranged that when the glass ring W and the rubber ring Y were placed on the top, the upper surface of the rubber ring was just flush with the surface of the Ringer solution. The cleaned vitelline membrane was now transferred carefully to the bowl and floated on to the top of the rubber ring, after which a few cuts with the scissors would suffice to make it lie down over the hole of the diaphragm in a star-shaped figure. Before any drying could take place, the rubber ring Z and the glass ring X were placed in position, and the whole system, W, Y, Z, X, could now be lifted up. Meanwhile the clamp C had been fixed in a wooden holder in a vertical position, and the lower parts of the apparatus (L, N, P, A and Pj) had been assembled. It was therefore a fairly simple matter to add the central pieces containing the membrane and the remainder of the apparatus (Q₁B, Q, O and M) before screwing down the clamp. As the clamp was provided with a universal ball-and-socket joint next to the cork pad M, any inequalities of axis due to imperfect grinding of the ends of the chambers, were automatically corrected. The assembled apparatus was then detached from the wooden support and immediately filled with the solutions which were to be tested, after which readings with the conductivity apparatus were at once begun.

It was found that the observed resistance of normal sodium chloride solution (5.8 per cent.) was 4.3 ohms in chamber A and 4.7 ohms in chamber B instead of the theoretical (13.5 ohms). The high conductivity (due to the wires holding the platinum foil, the size of the chambers, etc.) was therefore corrected for by multiplying the observed resistances of chamber A by the factor 3.14 and those of chamber B by the factor 2.87. This correction was tested by means of normal sodium hydroxide and normal hydrochloric acid solutions, and found to be adequate in each case. In the case of the solutions most used, 0.9 per cent. Ringer gave a resistance of 21.5 ohms, i.e. 67.5 ohms absolute resistance, or conductivity, k, 148 × 10⁻⁴; 0.7 per cent. Ringer gave absolute resistance 83.5 ohms or absolute conductivity, k, 120 × 10⁻⁴. Bellini’s figures (1) were, for yolk, k, 172 × 10⁻⁴ and for white, k, 500 × 10⁻⁴, but a comparison between salt solutions and such highly colloidal mixtures is difficult.

Figs. 2 and 3 show the effect obtained when the apparatus was set up with collodion membranes. Fig. 2 illustrates the results of Exp. 2 in which a Zsigmondy “ultra-fine” collodion membrane was used (pore-size < 0.3p.), chamber A containing 0.9 per cent. Ringer and chamber B 0.7 per cent. Ringer, corresponding in freezing-point depression to yolk and white respectively. The former began at absolute conductivity k, 156 × 10⁻⁴, the latter at k, 119.5 × 10⁻⁴, but after 1300 minutes equilibrium had been reached at k, 134 × 10^-4. Fig. 3 illustrates the results of Exp. 3 in which a Zsigmondy “Membran-filter” collodion membrane was used (pore-size > 0.3p.), the chambers being filled with the same solutions. Again, equilibrium was reached after 1300 minutes. The difference of pore size, as was expected, made no appreciable difference, and was evidently not a limiting factor in the attainment of equilibrium.

After a few other similar trials of the apparatus, the vitelline membrane itself was used, and, as Fig. 4 shows, when placed between Ringer solutions of the same freezing-point depression as white and yolk, made absolutely no attempt to resist the dilution of the stronger and the concentration of the weaker. In Exp. 14 the two solutions, beginning at k, 153.5 × 10^-4 and 122 × 10^-4 respectively, attained equilibrium at 134.5 ^{x 10.4} after¹²⁰⁰ minutes. In this experiment the yolk side, i.e. the internal side of the membrane, was duly turned towards the stronger of the two solutions. When equilibrium had been attained, the resistance across the membrane was taken, only one electrode in each chamber being used ; this gave an average value of 79 ohms (uncorrected) or 237 ohms (corrected). But as the distance between the two electrodes used (see Fig. 1) was 4.9 cm. the absolute resistance was only 48.5 ohms, i.e. even less than that found for the spaces between the electrodes. Evidently the membrane itself offers no extra resistance to the passage of the current.

Some other results obtained with the vitelline membrane are plotted in Fig. 5, which contains the data of Exps. 7 and 13. In both these cases the position of the membrane relative to the Ringer solutions was reversed, i.e. the side which had originally been in contact with the yolk was turned towards the weaker of the two and vice versa. However, this had no effect on the attainment of equilibrium, although in Exp. 7 the time taken was rather shorter than in the other experiments.

In this connection, a precise determination of the thickness of the vitelline membrane is of some interest. In view of its tenuous and fragile nature, the best way to accomplish this seemed to be the calibration of the fine adjustment of a microscope and the observation of the number of turns required to focus the upper and under surface of the membrane. The calibration was carried out by measuring the thickness of several transparent membranes by means of a “P-gauge” with an accurate micrometer screw. Particulars are shown in Table 1.

The average number of turns of the fine adjustment necessary for a number of vitelline membranes was 3.5, from which it may be concluded that the average thickness of the membrane is 0.024 mm-It may therefore not be surprising that osmotic equilibrium should be attained more quickly in the case of the vitelline membrane than in the case of collodion membranes.

It was next thought advisable to try the effect of using dialysates of yolk and white instead of the Ringer solutions. For this purpose a thin collodion bag filled with distilled water was suspended in yolk or white for 24 hours, and the osmotic concentration of the dialysate then adjusted by trial and error to the osmotic concentration of the original yolk or white. Thus the freezing-point depression of the yolk dialysate was—0.318° C. initially; it was concentrated on a warm water bath until its freezing-point was—0.606° C., and finally diluted so as to come to—0.55° C. Both dialysates were filtered through a Berkefeld candle in order to avoid possible development of yeasts and bacteria. In concentrating the dialysates, a small amount of protein separated out, and in addition to traces of protein there were also present small amounts of glucose and probably other small organic molecules. By drying the collodion bags well before use, it might have been possible to avoid this, but dialysis would then have been so slow as to require the addition of some preservative which might have exerted unknown effects on the membrane permeability.

The final absolute conductivities before the beginning of the experiment were : yolk 95.2 × 10⁻⁴, white 84.8 × 10⁻⁴, and the final pH was: yolk 6.2, white 9.1. It was soon obvious that the system was not behaving like those of the previous experiments, for, as is shown in Fig. 6 (Exp. 8), the conductivity of the white dialysate rapidly rose, while that of the yolk dialysate remained stationary, until virtual equilibrium was reached before 150 minutes. The pH, colorimetrically determined, was found to be the same in both chambers after 400 minutes, the white having become more acid, the yolk more alkaline. This was, of course, in striking contrast with the behaviour of the vitelline membrane in the intact egg, where, as we know from the work of many investigators (the most recent being Gaggermeier (2)), the white assumes very shortly after laying pH of 9.1 and retains this high alkalinity for as long as 250 days, the yolk meanwhile continuing at pH 6-o, 5.0 or lower.

Another experiment of the same kind (Exp. 9) is shown in Fig. 7. Here there seems to have been some impediment to the attainment of equilibrium, and the curves did not come together until some 400 minutes had passed, but the general appearance of the graph is quite the same as Fig. 6, the white dialysate changing considerably in conductivity, and the yolk dialysate remaining unchanged. The explanation of this result must lie in the fact that where such complicated mixtures of salts are concerned, it is impossible to assume that all the constituents are completely dissociated at the beginning of the experiment. Thus in Figs. 6 and 7 some constituent, possibly phosphate, must be dissociating on the yolk side just as fast as other ions are passing through to the side of the white. These experiments do not therefore tell much about the properties of the membrane, but in Exp. 9 (Fig. 7) a subsequent drift not shown in the figure took place between 450 and 1410 minutes, leading to a final average value for both chambers of absolute conductivity, k 92 × 10⁻⁴, which is almost exactly half way between the startingpoints. Evidently the salts which dissociate on the yolk side during the experiment are dialysable, and true equilibrium is eventually attained.

All the experiments so far described go to show that when isolated the vitelline membrane of the hen’s egg behaves like a dead collodion membrane, offering no resistance to the passage of water and ions other than the factors which affect any swarm of particles passing through the pores of a sieve. But this leads directly to a dilemma. These experiments in clearing the membrane of responsibility, throw the burden of the steady state in the intact egg on to the yolk and white so that the explanation would seem to lie in some such factor as the difficulty of water transport through these media, etc. Yet this cannot be the case, for if it were, the yolk and white, when separated by a collodion membrane, would not come rapidly into osmotic equilibrium, as they do. Only three ways out of this dilemma appear to be possible: (1) it may be said that the steady state in the intact egg is maintained by the performance of work at the membrane, the energy being derived from the mechanism of glycolysis in the yolk. If this were so, the isolated vitelline membrane might be expected to be powerless, and the yolk glycolysis, in the absence of the normal membrane, might not be of any value osmotically. But this is not the only possibility, for (2) it may be said that the vitelline membrane is not the efficient structure in any case, but that some film, either (a) protoplasmic, or (b) formed of fatty substances or of denatured protein, exists at the surface of the yolk underneath the vitelline membrane, and that this film is the responsible agent in the intact egg. Finally, (3), it may be maintained that the vitelline membrane itself is the efficient structure and possesses peculiar properties other than the capacity to utilise energy provided by the yolk, but that these properties are instantly destroyed when the membrane comes into contact with salt solutions, owing to denaturation of the keratin or to some other process. This view is difficult to disprove since it precludes the study of the vitelline membrane under anything approaching simplicity of conditions.

Some reasons will now.be given for deciding against hypothesis (2). If it were true, the conductivity of a salt solution suspended above a mass of yolk should rise with exactly the same speed whether any membrane (vitelline, collodion) were present between the yolk and the salt solution or not, for the essential membrane, the surface film, would always be there. That there does exist some such film is shown readily if drops of yolk be allowed to fall into Ringer solution ; they do not mix with it, but retain each their own boundaries, and can only be dispersed into a uniform emulsion by vigorous stirring. But that this film does not have the properties which it ought to have on hypothesis (2) is shown by Fig. 8 (Exp. 18). Here the dialysing apparatus was set up vertically, with the lower chamber the same as usual only containing no electrodes. The lower chamber was filled with yolk, and the upper with 0.7 per cent. Ringer solution. When collodion or vitelline membranes were used to separate the yolk from the Ringer solution, the resistance of the upper chamber fell very regularly, but when no membrane was used, and the saline solution was floated carefully on to the surface of the yolk, the fall was much more rapid. This does not favour the view that the yolk-Ringer interface forms an effective barrier for ions or water.

Again, if a surface film exists on the yolk, it would be of interest to see whether breaking it had any effect on the relations between yolk and white. It was not possible to take the yolk out of an intact egg, shake it thoroughly, and replace it in an intact vitelline membrane surrounded with white, but it was possible to place well-mixed yolk on one side of a vitelline membrane in the dialysing apparatus and well-mixed white on the other, and this was done in Exp. 22 (Fig. 9). The result, as shown by freezing-point determinations every few days, was an attainment of equilibrium so slow as to imitate the intact egg. It could only be concluded that either (a) the surface film had reformed itself at once in the glass chamber at the yolk-membrane interface, or (b) that work was being performed just as suggested for the intact egg, both membrane and glycolytic mechanism being present together. One possibility seemed definitely excluded, namely that the surface film was protoplasmic in nature, for such a structure could hardly reform itself automatically after the thorough mixing which the fresh yolk had undergone.

Next the experiment of Straub (3) was once more repeated, but with an important difference. Straub’s conclusion that yolk and white, separated only by a collodion bag, came quickly into equilibrium, was supported by experiments in which the bag was rotated by a clockwork mechanism during the dialysis. This rotation might be expected to have the effect of shearing away the protoplasmic or fatty surface film as soon as it was formed, so it was important to repeat the experiment using a stationary bag. As Fig. 9 shows this modification was quite ineffectual, equilibrium being attained just as quickly whether the bag was rotated or not. This again is fairly strong evidence against the existence of an effective barrier in the shape of a surface film.

When a thicker collodion membrane (a Zsigmondy “Membran-filter”) was placed in the dialysing apparatus with yolk in one chamber and white in the other, however, a rather longer time was taken to attain equilibrium, namely, 13 days (Exp. 23). Yet, as Fig. 9 shows, this was not so long as the time taken when the vitelline membrane was in the apparatus. Here it is necessary, in comparing these membranes, to take their thickness, surface, and the amount of yolk, etc., in the system into consideration. The following data are therefore given :

In order to obtain a factor for each system, it may be assumed that the rate of passage of water through the membrane is proportional to the surface of the membrane and the reciprocal of its thickness. It may also be assumed that the effect of the passage of a given amount of water through the membrane will be proportional to the reciprocal of the mass of material on each side. The rate of equilibration should be proportional, then, to where s is the surface, t the thickness of the membrane, and m the mass of material on each side. The factors, derived in this way for each experiment, were as follows :

1. The vitelline membrane of the hen’s egg, when isolated and placed between two salt solutions of different strengths, opposes no resistance to the dilution of the stronger and the concentration of the weaker.

2. Yolk and white, when separated by a thin collodion membrane, rapidly attain osmotic equilibrium.

3. Yolk and white, when separated by the vitelline membrane itself in the dialysis apparatus, equilibrate much more slowly, though not as slowly as in the intact egg.

4. It follows that neither the vitelline membrane, nor the yolk, nor the white, alone, is responsible for the maintenance of the steady state which exists in the intact egg. The phenomenon arises out of some collaboration of the three. It may be due (a) to the performance of thermodynamic work at the membrane, the glycolytic mechanism of the yolk providing the energy and the membrane utilising it, or (b) to the possession of osmotic properties by the vitelline membrane which are instantly lost when it is removed from the egg and brought into contact with salt solutions, or (c) to some feature of the physical structure of the yolk and white which retards the attainment of equilibrium, and which is not wholly destroyed by artificial mixing.

5. Experimental evidence is adduced in this paper which makes the second of these possibilities unlikely. Grave objections to the first have already been raised in the preceding papers of this series. But, on the other hand, there is as yet no positive evidence in favour of the third.

The author wishes to acknowledge his indebtedness to Dr Michael Smith, and to Mr Shepherd, who carried out the freezing-point measurements. The kind assistance of Dr Malcolm Dixon and Mr Leese was of much value in the rest of the work. The cost of these researches was partly defrayed by a subvention from the Government Grant Committee of the Royal Society, to whom the thanks of the author are due.