The Fertilization Reaction in the Sea-Urchin: The Block to Polyspermy

The experiments described in this paper are concerned with the reactions of sea-urchin eggs to high concentrations of spermatozoa of the same species. Several papers on this and allied subjects have recently been published, and it may therefore be convenient to summarize the results so far obtained. The first series of observations involved the measurement of the time taken for a visible change in the cortical structure of the egg, which occurs at fertilization, to pass completely over the egg surface (Rothschild & Swann, 1949). It has of course been known since the nineteenth century that the surface of the sea-urchin egg changes at fertilization, and the method of observing this change, whether with polarized light, darkground illumination or ordinary fight, all of which have been used at one time or another, is of no great interest in this context. The new feature in our work was to measure the conduction velocity of this change. The total conduction time was about 20 sec. at 18° C. As no faster post-insemination change had been observed, it was natural to consider whether this one could be the block to polyspermy. There were two methods of trying to answer this question. The first was to investigate the effect of polyspermy-inducers, such as nicotine, on the conduction time of the cortical change. If this change were the block to polyspermy, polyspermy-inducers might be expected to make the conduction time longer, unless they acted by increasing the chances of a successful sperm-egg union. Nicotine had no effect on the conduction time of the cortical change (Rothschild & Swann, 1950) ; but a consideration of its alternative mode of action led to the second method of investigation, an examination of the probability of a collision between a spermatozoon and an egg being successful (Rothschild & Swann, 1951 b). If we think in terms of sperm-egg collisions and the probability of their being successful, an assumption must be made about the way such collisions take place; it is that suspensions of eggs and spermatozoa can be treated as spheres in assemblages of gas molecules, in which case an estimate of the probability of a successful collision can be made. The concept of the sperm-egg collision is not essential for a quantitative description of what happens when eggs and spermatozoa are mixed together. The relationship between the number of eggs fertilized and the time of contact between the eggs and the spermatozoa can be described in terms of a fertilization parameter a (Rothschild, 1952), the value of which depends on the condition of the eggs and spermatozoa, the density of the latter and the speeds at which they swim. This fertilization parameter does not involve any assumptions about sperm-egg collisions or the probability of their being successful, though it can be described in terms of these two variables.

These experiments depended on developing a technique whereby eggs and spermatozoa could be brought into contact with each other for known periods of time and then functionally separated. This technique suggested a method of investigating the conduction time of the block to polyspermy. Eggs were placed in contact with spermatozoa for a known time, 25 sec., at the end of which the spermatozoa were killed; later, the number of monospermic, polyspermie and unfertilized eggs were counted. Simultaneously, an identical experiment was done, but instead of killing the spermatozoa at the end of the 25 sec., more were added. This increased the number of polyspermie eggs, showing that during the original sperm-egg interaction time, all the blocks to polyspermy were not completed ; from which it was concluded that the conduction time of the block to polyspermy was relatively long, being of the order of ten or more seconds (Rothschild & Swann, 1951a). These results were of a preliminary nature, and in a later paper (Rothschild & Swann, 1951 b) we said they would be systematically repeated. At the same time, certain difficulties in there being a low-speed block to polyspermy but a comparatively small incidence of polyspermy at normal sperm densities were discussed in detail.

Since that time, a simpler method of investigating the conduction time of the block to polyspermy in the sea-urchin egg has been devised, which is the subject of this paper.

Eggs, with their jelly-coats intact, and spermatozoa of Psammechinus miliaris were used. The room temperature varied from 16 to 18 ° C. during the season.

The experimental procedure depended on the following self-evident thesis. Suppose, for example only, that a number of eggs are all fertilized at t = 0 and that the block to polyspermy is complete at t = 5 sec. After 5 sec. there will be a certain number of polyspermie eggs in the egg population. The eggs will not all be polyspermie because some of them will not have sustained more than one successful sperm-egg collision during that 5 sec. As, however, all the blocks to polyspermy are complete by 5 sec., the number of polyspermie eggs will never be greater than it is at 5 sec. If the proportion of polyspermie eggs is 0·5 after 5 sec., the proportion must be 0·5 after 6, 10, 30, or 180 sec. It follows from this argument that if we take a series of egg suspensions and fertilize all the eggs in them at t = 0, and if we ‘remove’ the spermatozoa from these egg suspensions at various times after t = 0, say, 5, 10, 15 and 40 sec., then the time after which there is no increase in the incidence of polyspermy will be the conduction time of the block to polyspermy. Conversely, any decline in the incidence of polyspermy at t = r, as compared with the incidence of polyspermy at t=s, where r> s, will be due to sampling errors or mistakes in deciding whether an egg is polyspermie or monospermic. The first problem to be resolved in an experiment of this type is how to fertilize all the eggs at t = 0. In a previous paper (Rothschild & Swann, 1951b) it was shown that when eggs and spermatozoa are mixed together and the number of spermatozoa per ml. is 3·67×10⁵, some 12 sec. must elapse before half the eggs in the suspension are fertilized. The incidence of polyspermy at such sperm densities is too low for accurate calculations based on that incidence. If the sperm density in contact with the eggs is increased by a factor of 300, however, both difficulties are resolved, for a significant number of eggs become polyspermie and all eggs in the suspension are fertilized, for the first time, in the neighbourhood of t = o.

The experimental method requires first, that the eggs and spermatozoa be rapidly mixed together, and secondly, that after they have been in contact with each other for a particular time, the reaction should be suddenly stopped. This was achieved by adding 20 ml. of sperm suspension to 2 ml. of egg suspension and, after the required sperm-egg interaction time, transferring the 22 ml. of eggs and spermatozoa into 1000 ml. of sea water. This reduces the density of the spermatozoa in contact with the eggs by a factor of about 45, and as the original density was always about 10⁸/ml., the density after dilution was 2−3 ×10⁶/ml. At this sperm density, the incidence of polyspermy is negligible, being less than 2 % (Rothschild & Swann, 1950).

The results and their interpretation are easier to follow if prefaced by a brief outline of the theoretical considerations upon which the experiments and analysis are based. Each experiment provides information showing how the proportion of polyspermie eggs in a suspension varies with the time of contact between eggs and spermatozoa. Five conditions of a fertilized egg are shown in Fig. 1,a—e. In a the egg has just been fertilized; b is a little later, the block to polyspermy having travelled part of the way round the upper half of the egg; in c half the egg is covered; in d three-quarters of the egg are covered; and in e, when t=T, the block to polyspermy has passed completely over the egg surface. It is clear from these diagrams that the chances of polyspermy in unit time are greatest immediately after the first fertilization, since from then on the area of egg surface available for re-fertilization gets smaller and smaller. Theoretically, there must therefore be a change in the form of the curve relating the proportion of polyspermie eggs to the time of contact with the spermatozoa, because at t = T, the conduction time of the block to polyspermy, there can be no further increase in the number of polyspermie eggs. The form of the theoretical curve is shown in Fig. 3. The abscissa at the ‘point of intersection ‘between the curved and horizontal portions of this curve will be the conduction time of the block to polyspermy, T. To determine T and its precision, a curve must be fitted to the experimental points. The form of this curve cannot be determined from the experimental data; it must be based on some hypothesis. The hypothesis, which leads to equ. (1) on p. 480, is that when an egg is fertilized, a change in surface structure which makes the egg ‘impermeable’ to further spermatozoa passes over the egg at a uniform rate, unaffected parts of the egg surface remaining equally receptive to spermatozoa. We cannot prove that such a hypothesis is correct, but the experimental data may either show that it should be dismissed, or that there are insufficient grounds for rejecting it. The methods of fitting curves to the experimental points, of estimating the ‘intersection point’ T and its precision, and of testing the original hypothesis present problems of considerable difficulty. Details of the methods used will be found at the end of this paper in an appendix by Dr H. Ruben, to whom we are much indebted for devising a rigorous way of analysing the experimental data.

Fertilization at t = 0. It was first necessary to establish that when eggs were mixed with high concentrations of spermatozoa, all the eggs were in fact fertilized in the neighbourhood of t = 0. An experiment to verify that this does happen is shown in Fig. 2. This experiment was done by the ‘hypo-hypertonic method’ described in a previous paper (Rothschild & Swann, 1951 b). The sperm density was somewhat lower than that used in the block to polyspermy experiments to be described ; nevertheless, 50% of the eggs were fertilized in 1·25 sec., 94% in 2·5 sec. and 100 % in 3·75 sec. It is therefore safe to assume that in all the block to polyspermy experiments, 100 % fertilization occurred within 2 sec. of t = 0, the time at which the eggs and spermatozoa were mixed together.

Conduction time of the block to polyspermy T. Fig. 3 a and b are graphs showing the variations in the proportions of polyspermie eggs after different sperm-egg interaction times ranging from o sec., when all the eggs were fertilized for the first time, to 180 sec. Each graph consists of the experimental points, the fitted curve based on equ. (1), p. 480, and the point on the curve at the ordinate at t=T, the conduction time of the block to polyspermy. These two experiments were deliberately selected to show a ‘good’ and an ‘indifferent’ result, from the point of view of correspondence between theory and experiment. The words ‘good ‘and ‘indifferent ‘will be quantitatively defined in the next section.

The results of all experiments are summarized in Table 1. The second column in this table gives the conduction times, T, of the block to polyspermy in all experiments ; these are also shown as a histogram in Fig. 4. The third column gives the standard error of T. It will be observed that the block times, whose arithmetic mean is 63 sec., vary over a wide range from experiment to experiment. This is to be expected, as suspensions of eggs from different females often show marked variations in their properties (Rothschild, 1949 a). One of us (R.) found two lots of eggs of Echinus esculentus which appeared to have no block to polyspermy at all, or an extremely slow one, in the same year as these experiments were done. The fourth column in Table 1 gives the values of a. for each experiment. This figure is of great importance in the analysis of the results ; it is a measure of the condition of the egg surface during the propagation of the block to polyspermy, and it can be directly compared with the a values obtained in the previous experiments on monospermic eggs (Rothschild & Swann, 1951b). The fifth column gives the standard error of a. The sixth column is identical with the fourth except that the a’s have been divided by n, the sperm density relevant to the experiment in question. The values of χ² are tabulated in the seventh column, the number of degrees of freedom being given in brackets. Three of the fourteen experiments produced values of χ² which were sufficiently high to warrant rejection of the original hypothesis. These experiments are marked with an asterisk. One of them, Exp. 13, is also shown in Fig. 3 b. All other experiments were ‘good ‘in the sense that the results provided insufficient grounds for rejecting the original hypothesis.

The probability of a successful sperm-egg collision, p, is related to a by the equation α = Zp, where Z is the number of spermatozoa colliding with the egg surface in unit time. The evaluation of Z requires an estimate of the mean speed of the sperm suspension. Although this can be made with some confidence at low sperm densities, all the experiments described in this paper were carried out at high sperm densities, in which circumstances mean speeds cannot be accurately measured. For this reason, and because a is an ‘observable’ quantity, being the slope of the curve relating the logarithm of the proportion of fertilized eggs to the sperm-egg interaction time (first fertilization),^* the results are given in terms of a rather than p in this paper.

In the previous paper (Rothschild & Swann, 1951b, p. 407), values of p, the probability of a successful sperm-egg collision, were tabulated for different sperm densities. These values have been converted into the corresponding α/n values ^† and plotted against log₁₀ n in Fig. 5. Even if the first point, wthis paper, though reference tohich had a large standard error, is ignored, there is a tendency for the points to flatten out at higher sperm densities. To compare these α/n values with those obtained in the experiments tabulated in Table 1, the α/n, n plot must be logarithmic, which may tend to reduce the curvature of the relationship between a/n and log₁₀ αn in Fig. 5. In such circumstances any discontinuity between a/n before and after the first fertilization might be difficult to see. The comparison is made in Fig. 6, and, in spite of the adverse effects of the logarithmic plot, it is clear that α/n during the propagation of the block to polyspermy is markedly lower than α/n before fertilization. This is not due to differences between the eggs in successive years, because α/n for the experiment in Fig. 2, which is shown as a triangle in Fig. 6, is precisely where it should be for the pre-fertilization α/n level.

The conclusions from these results are first, that there is a low-speed block to polyspermy, which takes some 63 sec. (arithmetic mean) to pass over the egg surface. Secondly, that the fall in a after fertilization and before the block to polyspermy is complete strongly suggests that there is a higher speed but partial block to polyspermy, which covers the egg surface in a shorter time than the final block which makes the egg completely impermeable to spermatozoa. Elsewhere, attention has been drawn to a feature of previous experiments which is suggestive in this

The original hypothesis, first mentioned on p. 472, involves several assumptions and simplifications which might make the fit between theory and experiment less good, or in other words make χ² rather high. One simplifying assumption is that the conduction rate of the block to polyspermy is uniform. The only change that has so far been observed, the cortical change, is not propagated over the egg surface at a uniform rate (Rothschild & Swann, 1949), and there is some evidence for believing that this non-uniformity is real (Rothschild, 1949 b). A further simplifying assumption is that all eggs in the suspension are fertilized exactly at t = o. Finally, we have assumed that no re-fertilization takes place after the dilution of the original sperm-egg mixture. Although these assumptions are individually justifiable and necessary as approximations, they may collectively produce deviations from theory. Any attempt to take these factors into consideration in devising an analytical procedure would make the already difficult calculations too complicated to warrant serious consideration.

The data tabulated in Table 1 make it reasonably certain that under the conditions of these experiments, the block to polyspermy takes between 17 and 94 sec. to pass over the egg surface, the variation being probably due to the varying properties of eggs from different females. This conclusion can also be reached, though not quantitatively, by an examination of the raw experimental data, as in Fig. 3 a and b.

The position is not quite the same in regard to the high-speed block to poly-spermy, the existence of which, as was said earlier, is strongly suggested by the fall in a during the conduction of the block to polyspermy. No high-speed block to polyspermy has been observed, from which it follows that its existence can only be the subject of surmise. The observations on the fall in a do however agree, in an interesting way, with our previous observations on the incompatibility of the observed incidence of polyspermy at normal sperm densities with a block to polyspermy which takes ten or more seconds to pass over the egg surface. Examination of Fig. 6 shows that if the pre-fertilization α/n values are extrapolated to n = 10⁸, the pre- and post-fertilization α/n values are in the ratio of 1:15. Bearing in mind that the experiments to determine α/n for unfertilized eggs were done on jelly-free eggs while those described in this paper were not (see Rothschild & Swann, 1951b), the factor for the fall in α/n may be of the order of 20. This means that, taking a value of 60 sec. for the complete block to polyspermy, the proportion of polyspermie eggs in an experiment in which n equalled 10⁸/ml. would have been reached in about 60/20 = 3 sec., if α/n had not fallen by a factor of 20. This suggests that the first, incomplete, block to polyspermy must pass over the egg surface in less than 3 sec. If the conduction time of the first incomplete block to polyspermy were as much as 3 sec., the experimental 5 sec. points in the curves relating the proportions of polyspermie eggs to the sperm-egg interaction times should lie well above the theoretical 5 sec. points ; while if this fast block were complete in a small fraction of a second, the experimental points should be evenly distributed about the theoretical curve. This is because α/n would be twenty times higher during the propagation of the fast block to polyspermy than after its completion. Table 2 is a comparison of the observed and theoretical 5 sec. values ; there is a clear tendency for the experimental values to be higher than the theoretical ones. Because the experimental 5 sec. points are not much higher than the theoretical points, the conduction time of the fast but incomplete block to polyspermy must probably be less than 1 sec. In view of the simplifying assumptions referred to above, this argument about the conduction time of the incomplete block to polyspermy cannot at present be more than tentative.

The ‘conditions of the experiment’ were mentioned earlier on. The high sperm density used, which is essential in these experiments, may be responsible for two abnormal conditions. First, there may be an abnormally high concentration of the lytic substance A. Ill in contact with the eggs. This might make them unusually receptive to spermatozoa. In that case one would not expect α/n to be lower than in the pre-fertilization experiments, since, for a given sperm suspension, a is a direct measure of the ‘receptiveness ‘of the eggs to spermatozoa. Furthermore, the experiment shown in Fig. 2, which occurs again as a triangular point in Fig. 6, is inconsistent with this explanation of the facts. Secondly, the high concentration of spermatozoa might cause a fall in the pH of the medium which in turn might produce an abnormal amount of polyspermy. The arguments against the A. Ill hypothesis probably apply in this case as well. Neither of these factors are easy to control in experiments of this type, though they can and should be the subject of further experiments. To sum up, these experiments suggest that the adhesion of the spermatozoon to the egg surface initiates a change which probably takes a second or somewhat less to pass over the egg surface. This does not make the egg completely impermeable to spermatozoa but reduces the probability of re-fertilization by a factor which may be of the order of 20. Simultaneously or immediately afterwards, what might be called a slow mopping-up process is initiated. The latter may well be intimately associated with the cortical change and even with the elevation of the fertilization membrane. The fact that eggs whose fertilization membranes have been removed cannot normally be re-fertilized does not necessarily conflict with this view, as by the time this can be attempted, far-reaching changes have started in the egg as a whole. In such circumstances successful re-fertilization would be improbable, though under certain special conditions re-fertilization can be achieved (Sugiyama, 1952).

The preceding arguments might be thought to imply that the rapid partial block and the slower complete block to polyspermy are distinct processes involving different mechanisms. They may, however, be different aspects of one continuous process—the rapid partial block representing a steeply falling part of the curve of a against time, and the slow complete block, a later and more gradual decline in the same curve. At present there is no evidence upon which to decide between these alternatives, though the earlier results on the effect of nicotine (Rothschild & Swann, 1950) require consideration in this context. Nicotine was found to increase the probability of a successful collision, during the period of the block to polyspermy, by a factor of about 20 at a sperm density of 10⁸/ml. As nicotine did not increase sperm speeds or the conduction time of the cortical change, it must simply have raised α by a factor of about 20. The amount by which α normally falls, after the initiation of the block to polyspermy, has been shown in this paper to be 15−20. This raises the interesting possibility that there are two separate or separable blocks to polyspermy, the first but not the second of which is suppressed by nicotine.

By H. RUBEN

From the Department of Genetics, University of Cambridge

t = sperm-egg interaction time,

α = fertilization parameter (measuring the sperm-egg interaction rate),

and T=conduction time of the block to polyspermy.

p = number of polyspermie eggs,

and r = m+p.

The sampling variances of and in their marginal distributions are I_{α α} /|I/|and I_TTI| 1|, and if ^we approximate by substituting the estimated values of α and T for their true values in the latter expressions, we obtain intervals within which α and T may be said to lie, corresponding to any level of fiducial probability.

One of us (R.) is indebted to the Medical Research Council for provision of a laboratory assistant.