Protoplasmic Structure and Mitosis: II. the Nature and Cause of Birefringence Changes in the Sea-Urchin Egg At Anaphase

It is well known from the work of Schmidt (1937, 1939) that the birefringence of the mitotic figure of the sea-urchin egg decreases during anaphase. The same effect was also noticed in chick tissue culture cells by Hughes & Swann (1948). By analogy with the contraction of muscle fibres and protozoan myonemes, Schmidt interpreted this decrease as evidence for the contraction of spindle fibres. Whether or not his interpretation is correct, however, it is clear that extensive changes in protoplasmic structure do take place during anaphase. The present paper describes one aspect of these changes.

In the first paper of this series (Swann, 1951a), the technique of quantitative birefringence analysis for spindles and asters was described. The retardation as measured at any point in the mitotic figure is the sum of the retardations of all the elements of the spindle or aster traversed by the illuminating beam at that point. To interpret structure, however, it is preferable to know the coefficient of birefringence within the spindle or aster, rather than the total retardation. In the case of the aster, because of its radial symmetry, a curve for coefficient of birefringence can be calculated exactly from the retardation data. In the case of the spindle, however, a curve for coefficient of birefringence can only be calculated by assuming a definite shape; but as the spindle is neither clearly defined, nor constant in length, it is usually more satisfactory to compare curves of retardation. The two types of curve, in any case, are rather similar. With the asters, of course, there is no difficulty about comparing curves for coefficient of birefringence.

In comparing the curves for different stages in anaphase, both increases and decreases of birefringence are apparent. The molecular interpretation of these changes is an interesting problem, but not, as will become evident, immediately relevant to the question of what is causing the changes. It can, in fact, only be considered in relation to the form and intrinsic birefringence of the mitotic figures, and this raises further difficulties, since fixatives and imbibition agents all affect the labile protoplasmic structures of the egg. A detailed examination of these questions is left to a later paper.

The photographs in Pl. 9, figs. 1–4, are taken from a time lapse film (series A) of an egg of Psammechinus miliaris. They show the egg in metaphase, and at three successive stages in anaphase. It is evident that the birefringence of the whole mitotic figure decreases steadily, while the asters grow and the spindle lengthens. Because of this lengthening, complete spindle retardation curves cannot be superimposed, and the polar and equatorial regions have to be compared separately. This is done in Text-figs. 1 and 2, the curves of which are derived from the photographs in Pl. 9, figs. 1–4.

The first retardation curve, 56 min. after fertilization, shows the spindle in metaphase. It should be compared with the 58 min. curve. At the poles, both curves fit precisely, and there has evidently been no change in structure there during the interval of 2 min. At the equator, however, there has been a considerable decrease in retardation, and therefore in coefficient of birefringence. The next curve, at 60 min., shows a further decrease at the equator, and a slight decrease near the poles but a greater decrease midway between the two. The fourth curve, at 62 min. (shortly before cleavage), shows little change at the equator, but a further decrease in the polar and subpolar regions.

These changes are characteristic. They can be followed in more detail in series B, where the time intervals are shorter. Photographs from this series are shown in Pl. 9, figs. 5–10, and the corresponding retardation curves are given in Text-figs. 3 and 4. Four photographs of an egg in the two-cell stage during anaphase, showing similar changes, are given in Pl. 9, figs. 11–14.

From these curves and photographs, it is evident that the decrease in coefficient of birefringence first appears at the equator of the spindle. Two regions of decreasing birefringence then move away from the equator towards either pole. In the final stages, the equatorial birefringence remains unchanged, and the decrease is confined to the region near the poles. This sequence of events is shown diagrammatically in Text-fig. 5, where the region of the spindle that shows a decrease in birefringence is plotted against time. The region in question is shown as a vertical line stretching the appropriate distance, the scale is being measured in microns from the equator. The gradual movement of the region of decreasing birefringence from the equator to the poles is obvious. These changes have been photographed in many dozens of eggs, and analysed quantitatively in seven cases. The same sequence of events is always apparent.

The pattern of birefringence decrease during anaphase closely resembles the movements of the chromosomes. They also start at the equator, and move apart in two groups towards either pole. The next step, therefore, is clearly to find out whether the region of decreasing birefringence actually corresponds to the position of the chromosome groups. In the sea-urchin egg unfortunately, this is not easily decided, for the chromosomes are invisible by ordinary light, by phase contrast, or by polarized light. They can, of course, be seen in fixed and stained specimens, but the birefringence of the spindle is then drastically altered.

The problem was finally solved by putting single living eggs in a drop of sea water on a slide, and following their development under the polarizing microscope with the time-lapse camera. When the eggs reached particular stages they were fixed, dehydrated, and embedded in a drop of celloidin. The drops of celloidin were hardened, embedded in wax, and then sectioned and stained in the ordinary way. The position of the chromosomes in the spindle was then determined, and correlated with the appearance of the egg in polarized light just before fixation, and with the region of birefringence decrease as measured from the time-lapse film.

About thirty single eggs were treated in this way, and the results leave no doubt that the region of birefringence decrease does correspond to the position of the chromosome groups. A few of the results of this investigation are shown in Pl. 10, where a diagrammatic representation of the position of the chromosomes in the spindle, determined from sections, is shown beneath the photographs of the egg just before fixation. The approximate positions of the chromosomes, estimated from these data, have been inserted in the spindle retardation curves of Text-figs. 1 and 3.

The fact that the region of birefringence decrease is extensive, while the chromosome groups are compact, suggests that the decrease is not due directly to the chromosomes, as it might be if they were large and possessed a sign of birefringence opposite to the spindle material. Sea-urchin chromosomes in fact, appear to be completely isotropic. The conclusion is inevitable that the decrease in birefringence is due to structural changes in the spindle material. The fact that the decrease occurs only in the regions round the chromosomes, suggests that the structural changes are actually initiated by them.

Curves for coefficient of birefringence in the asters of series A are given in Text-figs. 6 and 7. The form of the curves at metaphase (56 min.) has already been discussed (Swann, 1951 a). At this stage the asters only extend out to about 15 μ. By 58 min., however, there has been a marked increase in coefficient of birefringence in all but the most central regions of the asters, as well as an increase in radius of about 3 μ. By contrast, the spindle at the same stage shows only a slight drop in the region of the equator (Text-fig. 1). By 60 min., however, the familiar decrease in birefringence has begun, the main region of decrease being between 3 and 6μ. Further out however, beyond about 9 μ, there has been a slight increase in birefringence, and the asters have grown to about 21 μ. By 62 min. there has been a further decrease in birefringence, this time between 6 and 12 μ, and a further slight increase beyond 15μ. The asters have grown to nearly 30μ in radius.

The anaphase changes in the spindle appear to consist mainly of a shifting pattern of birefringence decrease. In the asters, on the other hand, there are clearly two separate processes at work. The first of these, which starts at the very beginning of anaphase, is a gradual increase in the coefficient of birefringence of all but the innermost regions of the asters. This leads to a steady growth in actual size. The second process, which does not become apparent until at least 2 min. after the start of anaphase, is a decrease in the coefficient of birefringence. Unlike the increase just discussed, this decrease does not affect the whole aster at the same time. It starts from the centre and moves outwards. If affects not merely the more biréfringent regions of the original metaphase asters, but also the birefringence built up in the first few minutes of anaphase. It does not appear to affect the rather weak birefringence built up in the outer regions of the aster during anaphase, though there may of course be a decrease in these regions that is masked by a more rapid increase.

The final state of affairs is an aster of about double the radius and one-third the coefficient of birefringence of metaphase. The strength of birefringence is actually rather variable, and tends to be higher in two-cell than in one-cell stages. The final curve differs in various respects from the metaphase curve; its maximum is more flat-topped, and occurs rather further out from the centre.

These changes can be followed in more detail in Text-figs. 8 and 9, taken from series B. The shift of the regions of birefringence decrease with time, is shown diagrammatically in Text-fig. 10, which should be compared with the corresponding diagrams for the spindle (Text-fig. 5).

The most striking fact about the decrease of birefringence in the asters is that it does not begin until some 2 min. after decrease has begun in the spindle. By examining the various text-figures, however, it can be seen that this interval corresponds in fact to the time between the beginnings of birefringence decrease at the equator and at the poles of the spindle. The decrease of birefringence in the asters would seem therefore to be simply a part of the same process that is going on in the spindle.

Before considering the other aspect of the changes in the asters during anaphase, that is their growth in size, it is necessary to examine their behaviour in the earlier stages of the mitotic cycle. A cursory examination in ordinary light suggests that they reach a diameter of about 15 μ quite early in prophase, and then remain constant in size until anaphase. A more detailed examination in polarized light bears out this conclusion. In Text-fig. 11, for example, the radius of the asters in series B is plotted against time. On the same graph the retardation of the spindle equator is plotted to show the start of anaphase. The aster radius remains more or less constant at 15 μ up to the start of anaphase, when it begins to increase rapidly. The same effect is shown even more clearly in Text-fig. 12, where aster radius, measured on fixed and stained cells, is correlated with the state and position of the chromosomes. Throughout prophase and metaphase the aster radius remains constant, but as soon as the chromosomes start to separate it rises sharply, reaching, by the end of anaphase, a value about double that at metaphase. It is evident that the growing aster very soon reaches some sort of equilibrium, which is maintained throughout prophase and metaphase. With the onset of anaphase, however, this equilibrium is upset, and the aster starts once more to grow.

At first sight there appears to be no counterpart in the spindle, to the growth of the asters during anaphase. Curves of retardation across the spindle equator show no increase in diameter during anaphase (Swann, 1951a), nor is there at any time, a marked increase in coefficient of birefringence. On the other hand, there must be a certain intake of new material since the spindle as a whole, elongates. Why the intake of extra material should lead to elongation has been discussed elsewhere (Swann, 1951 b), and need not concern us here. It is interesting, however, to find that elongation, like aster growth, starts at the beginning of anaphase. In Text-fig. 11, the length of the spindle in series B is plotted against time. Although there is a slow elongation during prophase and metaphase, it is evident that the rate of elongation rises markedly at the beginning of anaphase. The same effect is seen more strikingly in the cells of the chick, where spindles may remain static for many minutes, but elongate suddenly as the chromosomes start to move apart (Hughes & Swann, 1948). Similar effects have been noticed by Ris (1943, 1949) and Bonnevie (1947). Evidently the spindle, like the asters, is in a condition of equilibrium during metaphase, but with the start of anaphase becomes capable of absorbing fresh material. Whereas, however, the aster simply increases in coefficient of birefringence, and grows in diameter, the spindle, because of its particular shape and structure, grows in length.

In short, while metaphase is a static period, there are two quite distinct types of structural change that affect the mitotic figure in anaphase. The first of these is a decrease of birefringence, which starts round the chromosomes as they leave the metaphase plate, moves with them to either spindle pole, and then spreads slowly outwards in the asters. The second change is a gradual growth that affects the whole mitotic figure at the same time, appearing in the spindle as an increase in length, and in the asters as an increase in size.

The metaphase spindle and asters both show a fall in coefficient of birefringence with distance from the centrosome. The significance of this was discussed in a preliminary way in the first paper of the present series (Swann, 1951 a). It was argued that since there are no marked variations of density within the asters and spindle, the fall in coefficient of birefringence must be due either to changes in the proportion of oriented material, or to changes in molecular and micellar arrangement. The fall in coefficient of birefringence with distance, however, is very approximately an inverse square. It may possibly be explained therefore by a structure of discrete fibrils radiating from the centrosome, which, for geometrical reasons, would give an inverse square fall in proportion of oriented material. It is also possible, however, that the aster and spindle are homogeneous structures, in which case the fall in coefficient of birefringence must be accounted for solely in terms of changes in molecular arrangement. Further discussion of these two structures is not possible without a consideration of form and intrinsic birefringence, which raises a number of difficult problems. The whole question was therefore left to a later paper.

In view of this uncertainty about the significance of the variation of coefficient of birefringence with distance, it is clearly useless to attempt a detailed interpretation of the changes in coefficient of birefringence with time. A preliminary discussion, however, is not perhaps without value.

The fact that, throughout anaphase, the asters are growing in diameter, must mean that fresh material is being oriented. The proportion of oriented material, in other words, is increasing at 15μ from the centre and beyond, from nil to some finite value. The coefficient of birefringence, however, is also increasing within the limits of the original metaphase aster. This may be due to changing molecular and micellar arrangements, but it is tempting to suppose that it also is due to an increase in the proportion of oriented material. This is borne out by the fact that the general shape of the curve for coefficient of birefringence does not appear to be altered at 1 or 2 min. after the start of anaphase; it is simply shifted outwards. It is perhaps significant also that the maximum value of the metaphase curve is not markedly exceeded during anaphase; a rather larger region of the aster merely attains the maximum value, suggesting perhaps that over this region all the available material is oriented.

The decrease in coefficient of birefringence, like the increase, might be due either to changes in the proportion of oriented material or to changes in molecular and micellar arrangement. The inner regions of the aster, however, suffer as much as a tenfold decrease in coefficient of birefringence, and the spindle even more. Yet the asters and spindle still retain their fibrillar appearance. It seems likely, therefore, that the decrease in coefficient of birefringence is due to molecular and micellar changes rather than an extensive decrease in the proportion of oriented material. This is essentially the conclusion of Schmidt (1937,1939), who supposed the decrease to be due to a contraction of the fibres. Contraction necessarily involves molecular and micellar rearrangements, in which the various participating elements move from a predominantly longitudinal to a more random or even predominantly transverse arrangement. Such a change is bound to lead to a reduction in positive birefringence.

In an earlier section, the close correlation between the position of the chromosomes and the regions of decreasing birefringence was pointed out. It was shown that the decrease could not be due directly to the chromosomes, which are small, and apparently isotropic. The region of decrease, on the other hand, is large, and spreads out on either side of the chromosome groups; moreover, when the chromosomes reach the neighbourhood of the centrosome, the birefringence decrease spreads out into the asters. It seems possible, therefore that the change is brought about indirectly by the chromosomes themselves.

Very little is known of how such structural changes as have been described in the present paper might be brought about. There would seem to be essentially two possible mechanisms. The changes might be the result of a self-propagating process, either physical or chemical, initiated by the chromosomes ; or they might be caused by the release and diffusion from the chromosomes of substances, acting directly or indirectly on the protoplasm. In spite of the small amount of evidence available, it is possible to decide between these two alternatives with some certainty.

A self-propagating mechanism, no matter of what type, continues after it is initiated, without regard to the original stimulus. Once started, for instance, a nerve-action potential is unaffected by whether the original stimulus is repeated or removed. The pattern of movement of a propagated change is therefore unaffected by any subsequent movement of the stimulating mechanism. This, however, is conspicuously not the case with the pattern of birefringence decrease in the spindle. The region of decrease moves with the chromosome groups, whereas, if the mechanism were a self-propagating one, it should spread outwards, if not concentrically, at least without regard to the movements of the chromosomes.

We are led therefore to the idea that the decrease in birefringence is brought about by the release of a substance which, it may be supposed, diffuses away from the chromosomes, and produces its effect as it goes. A difficulty about such an explanation, however, is that the observed rates of movement are so slow. The spread of the decrease in birefringence is not easily measured in the spindle, because of the complicated pattern of change, though it is clearly of the order of only a few microns per minute. In the asters the rate can be measured more exactly, and seems to be not more than 5 or 6μ per minute. As an example of the expected rate of diffusion in a sea-urchin egg, we can take the calculations of Rothschild (1949), although he was concerned with a slightly différent problem, namely the time course of diffusion from a point on the periphery of the egg. From his curves it can be seen that a point in the centre of the egg 50 μ. from the periphery, would reach 80% of the final concentration in 5 sec., for a diffusion constant of about 0·07 cm.²/day corresponding to a molecular weight of about 30,000.

This discrepancy might be overcome if the diffusing substance were of virus-like dimensions, though this seems improbable. A more likely explanation is that the oriented structure of the spindle and aster is retarding diffusion. Rothschild’s calculations were based on the assumption that the viscosity of cytoplasm is only a little above that of water, and for the unfertilized sea-urchin egg this is known to be the case. In the spindle and asters, on the other hand, there is an oriented structure, and it has long been realized that the resistance of such structures to the passage of molecules rises very sharply as the molecule approaches the pore size of the gel. If the diffusing substance in question consisted of relatively large molecules, diffusion might well be enormously slowed down. Yet another possibility is that ordinary diffusion is not operative, but that there is diffusion with a trap action. In such a system, the molecules of diffusing substance are caught and bound by the protoplasmic structure, so that the diffusion front does not advance until all the traps in its way are filled. Under these conditions, the apparent rate of diffusion is governed by the rate of release of the diffusing substance.

The increase in coefficient of birefringence during anaphase is a rather different case from the decrease discussed in the previous paragraphs, since it takes place simultaneously throughout the mitotic figure, leading to spindle elongation and aster growth. It would seem, therefore, that although the change in question is not completed at once, the whole cytoplasm is affected in some way in a short space of time after the start of anaphase. This is borne out to some extent by birefringence and light scattering changes in the cortex which also appear immediately after the start of anaphase. These changes are described by Mitchison & Swann (1952).

As there is no sign of a gradually moving pattern of change in this case, it is not possible to decide between a self-propagating mechanism and diffusion. Since, however, we have already postulated a diffusion mechanism of some sort for the decrease in birefringence, it is perhaps simplest to think in terms of the release of a second substance, controlling the growth of the spindle and asters, and possibly also the structure of the cortex. Since this change spreads throughout the cell much more rapidly than the decrease in birefringence, the substance in question cannot possess a very large molecule, nor can it be subject to a trapped diffusion mechanism.

From the evidence of the present paper, a hypothesis can be constructed to account for a small part of mitosis. The essential facts are that a pattern of changes in birefringence can be observed in the mitotic figure during anaphase. One type of change, which can be related to the position of the chromosomes, takes the form of a progressive decrease in coefficient of birefringence, probably due to molecular and micellar re-arrangements. The second type of change, which cannot be related to the position of the chromosomes, takes the form of an increase in coefficient of birefringence, probably due to the orientation of fresh material. The hypothesis suggests that these changes are brought about by the release of two different substances from the chromosomes during anaphase. The interpretation of the birefringence changes, though an interesting problem, is not relevant to the hypothesis.

It is not proposed to examine this hypothesis in detail, since observations to be described in later papers have also to be taken into account. There are, however, a few points of some interest that are relevant to the idea. The first of these is the bearing of the hypothesis on theories of anaphase movement. Since chromosomes normally all start to move at the same time, it has usually been supposed that they are controlled by some unspecified external agency. A few cases are known, however, where some chromosomes move apparently autonomously (Schrader, 1944). The present hypothesis can perhaps explain such cases, since it supposes the chromosomes to be responsible for their own movement, by controlling the structural changes in the spindle. It should be added that the birefringence evidence suggests, as Schmidt (1937, 1939) has already concluded, that the mechanism of movement is a contractile one.

A further interesting point is the parallel between this hypothesis, and the suggestion put forward by Rothschild & Swann (1949) and Rothschild (1949) to account for the structural changes in the sea-urchin egg at fertilization. They observed and measured a cortical change spreading round the egg from the point of sperm entry, taking some 20 sec., and concluded that it could best be explained by supposing the sperm head to release some active substance which diffused through the egg. The parallel between the release of substances from the sperm nucleus and the chromosomes is obvious.

The hypothesis may also be relevant to the problem of cell division. Since the decrease of birefringence spreads outwards at about 5 or 6 μ per minute, it should take about 10 min. to reach the cell surface. This is in fact about the time after the beginning of anaphase at which cleavage begins. Moreover, since the chromosomes end up at either pole of the spindle, the change would be expected to reach the poles of the cell first ; and there is clear evidence, both from birefringence and dark ground observations, that this is the case (Mitchison & Swann, 1952). These ideas are expanded into a hypothesis of cell division by Swann (1952) and Mitchison (1952).

If the chromosomes do in fact release active substances, it is clearly of the greatest importance to discover what they are. Since the ratio of total chromosome volume to total cell volume in the sea-urchin egg is only about 1 : 10⁴, and as only a part of the chromosomes could presumably consist of the substances in question, it is hardly to be expected that ordinary cytochemical methods would be sensitive enough to detect substances being released. A number of such tests have in fact been tried, but without success. The inference is that these hypothetical substances must be catalytic in nature, and the term ‘Structural agents’ has been suggested for them (Swann, 1951c).