ABSTRACT
Eggs from females homozygous for a sex-linked mutant factor, fsnasA, will not support the development of a viable embryo.
Approximately half of the eggs produced by fsnasA females are abnormal, containing little or no yolk.
Twenty per cent, of the embryos die during the first 2 hours of development. The remaining embryos develop into monster larvae which are not capable of hatching from the eggs.
The major effect of the factor is on the synchrony of cleavage and blastoderm mitoses. Moreover, the regular pattern in which disturbances occur suggests stratification of the egg cytoplasm.
Gastrulation and all subsequent developmental processes are abnormal. Germ-band formation and the development of mesodermal elements are strongly affected. The posterior mid-gut invaginates from the ventral surface of the embryo, and the proctodaeum and Malpighian tubules never form.
The possible effects of the factor on the processes of chromosome duplication during cleavage are discussed in connexion with Painter’s theory.
The developmental effects of fsnasA are compared with those of a sex-linked recessive lethal factor X2. The basic similarities of the two are probably the result of abnormal rates of mitotic division during blastoderm stages. Differences between the two suggest there may be a posterior to anterior gradient in regional differentiation of the embryo.
Gonad formation in fsnasA embryos from pole cells which migrate through the blastoderm wall during the early stages of gastrulation supports Poulson’s theory of gonad formation.
The broader implications of studies on female-sterility genes in Drosophila are briefly discussed.
INTRODUCTION
WADDINGTON (1956) has called attention to the importance of the relation between the structure of the egg and hereditary factors which determine characters in the developing organism: ‘When we discuss the eggs of the different kinds of animals, we… find that the eventual origin from which the whole later development springs is the orderly arrangement of essential parts of the ovum. We must therefore enquire a little more deeply how this arrangement is brought about. In particular, what is the relation between it and the hereditary factors or genes which determine the detailed character of the adult organism?’
Although, in most animals, no such relations have been carefully studied (with the exception of direction of coiling in Limned) there exists in Drosophila a class of female-sterility genes or factors in which females are sterile because the egg cytoplasm will not support the development of a viable zygote (Lynch, 1919; Merrell, 1947; Counce, 1956 a, b, c); e.g. when a female heterozygous for the gene is mated to a mutant male, females homozygous for the factor develop into adults; however, when these homozygous mutant females are mated, their offspring never develop to an adult stage. Apparently many such factors exist in Drosophila—an analysis of eight II chromosome female-sterility factors with no visible effects and chosen at random, produced three which were of the above type (Counce, unpublished). It should therefore be possible to collect considerable information as to the role of the egg cytoplasm and its relation to hereditary factors.
An advantage of such mutants is the existence of a viable homozygote for studies in developmental physiology, such as processes of ovary formation and egg development, and the developmental history of phenotypic differences when they exist. There is also a technical advantage over recessive lethal factors causing embryological death, for from female-steriles one can obtain 100 per cent, zygotes or eggs which are affected, rather than 25 per cent. This can be of great value in experimental studies of mutants, such as respiration rates, chromatographic differences, electrophoretic studies, &c., where it is sometimes necessary to utilize large numbers of eggs.
This paper deals with the effects of a sex-linked female-sterility factor, nasratA(fsnas A) on the development of the embryo in eggs from homozygous mutant females mated to mutant males.
MATERIALS AND METHODS
The sex-linked sterility factor, fsnasA, was found by Dr. Gamal Nasrat during the course of experiments on chemical mutagenesis carried out at the Institute of Animal Genetics, Edinburgh. There are no visible characters which distinguish it from the wild type, and it is associated with no gross chromosomal rearrangements (Dr. H. Slizynska, private communication). Linkage tests show that the factor is closely linked to scute (10 ±). All homozygous females are sterile; mutant males are fertile with non-mutant females. No analysis has been made of possible effects of the factor in heterozygotes. The stock is maintained against the Müller-5 chromosome.
The following description of embryogenesis is based on a study of serial sections of 647 embryos and eggs, from matings of fsnasA homozygous females to fsnasA males. At a given temperature, the rate of development in Drosophila embryos is quite constant, and it is possible to obtain embryos at various developmental stages by making timed collections of eggs and allowing them to develop to the desired stage. All developmental times and stages correspond to those given in Sonnenblick (1950) and Poulson (1950).
Detailed descriptions of methods of egg collection and histological techniques may be found elsewhere (Counce, 1956 a, b, c). Briefly, eggs are collected on agar lids over a 1-hour period from rapidly laying females 4-8 days of age. They develop at 25 ° C. until the desired stage is reached, and are then fixed. Routine methods of dehydration and embedding are used. Sections are routinely cut at 3-5 p and are stained with Heidenhain’s iron haematoxylin without counterstain.
RESULTS
Only offspring from crosses of mutant females to mutant males were studied.
Sectioned eggs could be divided into three categories: abnormal eggs which contained little or no yolk; early lethals—embryos which ceased development before the beginning of primary differentiation; and late lethals—embryos which developed beyond the first critical period and formed recognizable larval structures.
A. Abnormal eggs
Approximately half of the eggs produced by the females used in this study were yolkless or nearly so. All females laid normal and abnormal eggs. Such eggs are very white and extremely opaque when seen on the agar lids used for egg collection, and tend to collapse even when gently touched with a brush or dissecting needle; frequently their shape is abnormal. They are almost impossible to prick for fixation without severe damage. If they are left on the agar lids for more than 4 hours after deposition they are usually distinguishable from eggs which contain normal amounts of yolk. Recognizably yolkless eggs were not collected for fixation, but at early stages they could not always be distinguished from embryos in early developmental stages which are also opaque, and were therefore included among the sectioned material.
One hundred and nineteen yolkless or almost yolkless eggs were sectioned. Of these, 90, or about 75 per cent., were completely devoid of yolk, while the others had a few scattered yolk globules. In sectioned material the cytoplasm of such eggs is usually stringy and vacuolated, the size of the vacuoles varying from egg to egg. In some, small round objects are discernable which are not nuclei but which do not take the dark stain which Heidenhain’s iron haematoxylin imparts to normal yolk globules, suggesting that abnormal yolk may be deposited.
Fertilization and subsequent nuclear divisions can occur in these abnormal eggs. Nuclei are found in approximately half of the sectioned eggs. Although there may be several thousand nuclei present in the egg, they never migrate to the surface to form a blastoderm, and they soon become pycnotic.
The inner membrane (or plasma membrane) of fsnasA eggs may be abnormal or weak in the region of the micropyle, for frequently the ooplasm may ‘leak’ out at the anterior end and flow into the micropyle. In some, so much material flows out that it fills part of the space between the vitelline membrane and the chorion.
It is probably this same weakness which accounts for the difficulties encountered when we attempted to make films of the development of fsnasA embryos. All our attempts failed because the eggs always collapsed within an hour or two of being placed either in water (in which normal and other mutant embryos develop without difficulty) or in insect saline (0·75 per cent.). Leaks were also found in sectioned eggs in which development was proceeding.
Out of the total number of eggs examined, 48, or 7·4 per cent., were classified as unfertilized. This figure is probably exaggerated, for some early lethals are no doubt also included in the group. It is not always possible to determine after 6 hours whether the gamete nuclei have failed to unite or whether development has ceased shortly after this has occurred. The difficulty is compounded because certain cytological abnormalities typical of these eggs (see below) make it even more difficult to ascertain if there are nuclear structures present.
B. Early lethals
Of the 480 eggs in which there was recognizable development, 94, or about 20 per cent., ceased to develop before the beginning of primary tissue differentiation, the majority (90/94) ceasing to develop during the first 2 hours (nuclear cleavage stages). These were classified as ‘early lethals’. Their distribution was as follows:
It is probable that this is an underestimate of the number of early lethals because some of the cleavage and blastoderm stage embryos might actually have ceased development in a short time had they not been fixed, and those embryos which die during maturation, syngamy, or early cleavage and were not fixed until late in development will probably be classed as unfertilized eggs.
Early lethals are characterized by a vast array of cytological abnormalities and of abnormalities in nuclear distribution. Because the cytological abnormalities found in fsnasA embryos are of special interest, they will be discussed in a separate section.
Micropylar leaks were found in nine of the early lethals.
C. Late lethals
Cleavage 
Those fsnasA embryos which successfully pass the first 2 hours of development usually continue to develop until the end of embryonic life. However, they become abnormal early in development, and as development continues, these abnormalities become magnified.
Although the existence of abnormalities in early cleavage stages indicate that development may become abnormal very early in embryogenesis, it can remain normal through at least the 4th cleavage division.
By the 6th-8th divisions, however, development has become abnormal in almost all embryos. There were available only a limited number of embryos at this stage and all were abnormal, but the occurrence of some normal blastoderm stages indicate that development is not invariably abnormal by this stage. Abnormal embryos are characterized by abnormalities in migration of nuclei, lack of mitotic synchrony, and aberrant cytological features (see below).
Blastula stages 
In the normal embryo, nuclei migrate to the surface at the 9th cleavage division and divide there in synchrony three more times. The cell-membranes then form from the outer surface inward creating a cellular blastoderm. During this time the yolk granules are becoming concentrated in the centre. The pole cells are differentiated at the posterior end, usually during the 9th and 10th divisions.
One hundred and seventeen fsnasA embryos at this stage of development were studied. These included embryos from the time of the 1st divisions at the periphery (blastema) to the formation of the cell-membranes (cellular blastoderm). Classification is not always simple in early blastema stages, for abnormalities in nuclear distribution and mitotic synchrony result in embryos which are at the 1st or 2nd blastema divisions at the anterior end, but which have the typical appearance of late cleavage stage embryos at the posterior end. About 20 per cent, of the embryos at this stage are apparently developing normally.
Nuclear divisions during cleavage and the syncytial blastoderm stage are synchronous in normal embryos, with the exception of those of the yolk nuclei and the pole cells. In fsnasA embryos over half of the embryos in these stages show departures from this synchrony. Moreover, there is a general pattern of differences in synchrony which occurs in 90 per cent, of the affected blastula stages. The yolk nuclei tend first of all to depart from synchrony earlier than is normal, and frequently show many cytological disturbances. The surface of the embryo shows three (sometimes four) different areas which are clearly delimited from one another (Plate 1, fig. A). The posterior is usually most severely affected, the area being always somewhat broader ventrally, and in 40 per cent, of the embryos which show abnormal synchrony, the posterior region is the only visibly affected surface portion. This region is frequently in late stages of cleavage while the rest of the embryo is in the 1st or 2nd stage of blastoderm cleavage. Later this posterior region may not form a complete blastoderm. This is due to two different processes: in some embryos nuclei do not migrate to all areas, while in others there is an indication of grave nuclear disturbance and of nuclear breakdown after they reach the surface. The anterior tip, when it is affected, is usually one cleavage stage behind the adjacent sections, although one also finds embryos in which nuclear migration is not completed. Nuclear breakdown may also occur. The two central regions may or may not show differences from each other. Differences, when they do exist, are usually in nuclear size; in some instances they are in different phases of cleavage, the anterior of the two apparently the more advanced.
The boundaries between the different regions are very sharp; usually the first one or two bordering cells are markedly abnormal. This is important since normal embryos which are fixed in stages when synchronous mitoses are occurring show a fixation gradient, and one must be able to distinguish between these gradients and actual differences. The borders of these regions also roughly coincide with the positions of the abnormal furrows which form during gastrulation.
Many of the blastoderm embryos showed cytological abnormalities. Abnormalities in mitotic phase, division figures, and abnormal distribution of nuclei may be related and are found in different combinations. Table 1 shows the relations between these three abnormalities based on a detailed analysis of eighty-three embryos.
Because of irregularities in division and distribution of nuclei, cell-membrane formation is usually abnormal in fsnasA. The blastoderm wall varies in thickness, usually being narrower at both ends than in the central regions. When the blastoderm wall is incomplete at either end there is usually a broad band of cytoplasm containing a mixture of yolk globules and nuclei which runs to the tip (Plate 1, fig. B). The size of nuclei making up the cellular blastoderm also varies. In some regions the nuclei are much smaller and more closely packed than in normal embryos, suggesting that in some regions an abnormally large number of nuclei may form prior to cell-membrane formation.
In normal embryos the posterior polar plasm is visibly differentiated from the rest of the periplasm by the presence of discrete granules which were called ‘the germ-cell determinants’ by Hegner. Their exact function is not clear; however, most of them are included in the cytoplasm of the pole cells during their formation. These discrete bodies are only rarely observed in fsnasA embryos.
In approximately half of the abnormal blastula embryos, the pole cells do not form; they form in the remainder but there is always some breakdown of pole cells rather soon after they form.
After the formation of cell-walls at about 3 hours in the normal embryo, some of the pole cells migrate into the interior of the embryo; those which remain at the posterior end are later carried passively into the embryo by the developing posterior mid-gut rudiment. In fsnasA embryos, pole cells migrate into the interior at 3 hours; apparently all of the pole cells which are not too abnormal may enter at this time (Plate 1, fig. C). The second movement of pole cells never occurs in fsnasA because of abnormalities in the formation of the posterior mid-gut rudiment.
Distribution of yolk is rather abnormal, the globules being concentrated peripherally while large vacuoles in the centre contain only tiny droplets connected by cytoplasmic threads (cf. Counce, 1956a). Frequently large clumps of highly abnormal yolk nuclei are found within these vacuoles.
Eighteen of the embryos had material leaking into the micropyle. In one embryo, it was apparently the only abnormality.
Gastrulation and germ-band extension 
Gastrulation in the normal embryo consists of the formation of a long furrow on the ventral surface (the germ-band), followed shortly by the formation of an anterior cleft, the cephalic furrow. The posterior mid-gut rudiment then moves from the posterior region of the embryo around to the dorsal surface as the germ-band extends around to the dorsal side, the posterior mid-gut rudiment carrying the pole cells in its deepening cavity. The anterior mid-gut invaginates from the anterior ventral surface at about the same time. During this period the rudimentary embryonic membranes are also formed.
At this stage of development all fsnasA embryos are characterized by the development of a deep cephalic furrow in the anterior region; a deep furrow, which forms from the ventral surface about one-fifth of the way from the posterior end, and, anterior to it, two slightly less pronounced lateral furrows running in a dorso-ventral direction. There are also several lateral subsidiary furrows, usually running from the dorsal to the ventral surface, and there may be a furrow anterior to the cephalic furrow.
Germ-band formation is delayed (or formation of the cephalic furrow is precocious); its formation may also be abnormal, and its shape is always severely affected by the presence of the many deep furrows in the embryo (Plate 1, fig. D). In some instances the connexion between the various regions of the embryo created by the furrows is so narrow that the germ-band may be only one cell layer in thickness. The furrows can apparently prevent the formation of the germ-band in some areas. In rare instances the germ-band may never form.
The posterior mid-gut invagination never moves from the ventral surface. Part of the major posterior furrow is probably the posterior mid-gut rudiment, the two lateral folds just anterior to it representing the lateral arms of the posterior mid-gut which are apparent on the dorsal surface of the living embryo as the rudiment moves forward. Its position on the ventral surface is anterior to that of the normal rudiment which also forms first ventrally and then moves over the posterior pole. In fsnasA embryos this may be due to the failure of certain gross movements of the egg during cleavage and blastoderm formation (Imaizumi, 1954). This interpretation is given added support by the fact that the stomo-daeum in fsnasA embryos forms further anterior than is normal.
No evidence of proctodaeal formation has been found at any stage in the development of fsnasA embryos.
Cell disintegration has already begun by this time, usually at the posterior end among those nuclei involved in large masses of cytoplasm when blastoderm formation is not completed, and also in the anterior end where cells have remained in the non-cellular cytoplasm.
Pole cells were identified in only 8 out of 64 embryos, undoubtedly because it is difficult to identify them positively at this stage when they may be isolated single cells. None are ever included in the posterior mid-gut invagination.
The embryonic membranes, or elongate cells which closely resemble them, form in the abnormal embryos at this time, covering the entire region from the cephalic furrow back to the posterior lateral folds, a region more extensive, however, than that usually covered by these membranes.
In fsnasA embryos the posterior tip is frequently attenuated, and this abnormal shape first becomes obvious during the late stages of gastrulation.
Leaks were observed in 10 out of 64 embryos.
Histogenesis and primary organogenesis (5–12 hours)
Between 5 and 12 hours in the normal embryo the basic tissues become differentiated and the primary organs formed. At the end of the period three major morphogenetic movements occur—shortening of the germ-band, dorsal closure, and head involution. These developmental stages in fsnasA embryos carry the strong imprint of earlier abnormalities.
The formation of the stomodaeum is always affected; it forms at the anterior tip instead of on the ventral surface and is usually very short. No stomodaeum was formed in 13 out of the 96 embryos studied at these stages.
Formation of the anterior mid-gut may also be affected, the rudiment frequently consisting of fewer cells than normal, and in six embryos it failed to form at all. Its position in regard to the stomodaeum is frequently aberrant.
The posterior mid-gut could be definitely identified in only twenty embryos. However, it is not easy to identify during this period for its structure may be modified considerably, and the cells at this time bear a strong resemblance to some of the mesodermal elements.
The results of earlier abnormalities in the formation of the germ-band are now clearly apparent. It was lacking in 6 embryos, incomplete in 26, and showed little or no differentiation into its component parts in 21. Degeneration of the mesoderm may begin by this time; pycnotic or disorganized mesodermal cells appeared in thirteen embryos. In less severely affected embryos the differentiation of the mesoderm into somatic and visceral musculature and fat-bodies may be detected.
Although the differentiation of the hypodermis is very good, its development reflects abnormal blastoderm formation. It was incomplete in of the embryos at these stages: of these, 57·6 per cent, were incomplete on the posterior ventral surface, 16·7 per cent, at both anterior and posterior ends, 19·7 per cent, at the anterior end only, and the remaining 6 per cent, in other regions. Tracheal pits form (perhaps earlier than in normals) sometimes on the inner edges of the deep clefts which continue to divide the embryo. It is apparently impossible for the tracheal rudiments to elongate to any considerable extent, or to join to form the long main trunks. The salivary glands, which are also ectodermal in origin, may be identified in older embryos in this group, but their distortion and displacement resulting from the abnormal furrows makes it difficult to identify them accurately before they enter their secretory phase at 12 hours. The rudiments apparently originate very near the boundaries of the cephalic furrow.
In some embryos segmentation is superimposed over the major furrows, but segments may appear slightly later than normal.
Pole cells were identified in 38 of the 96 embryos. As the embryos increase in age there is aggregation of the pole cells during gonad formation and they are more easily identified: e.g. at 5 hours they were identified in only 2 out of 10 embryos, as compared with 7 out of 14 at 10 hours.
The major morphogenetic movements which take place during this period are all affected in the abnormal embryos. It is doubtful whether any shortening of the very abnormal germ-band occurs. Dorsal closure of sorts takes place, probably by the transformation of the embryonic membranes or the embryonic membrane-like cells which form over the long central portion. The head segment —even the head ectoderm—may fail to develop, and involution is usually abortive or does not occur. Only a few embryos have been observed in which considerable involution has taken place, and there is never the extension of the segments at the end of involution which there is in normal embryos (Ede & Counce, 1956).
Almost 9 out of 10 embryos between the ages of 5 and 12 hours show signs of advancing deterioration, although rarely embryos up to the age of 10 hours may have no disintegrating cells. Disintegrating cells are most frequently found in the posterior region or the anterior region where nuclei are involved in masses of cytoplasm. It has been mentioned that breakdown of the mesoderm may be apparent by this time.
Only two embryos of this stage were observed with leaks at the anterior end.
Final pattern of damage
Following the major morphogenetic movements in the normal embryo, the period from 12 to 18 hours is largely devoted to cellular differentiation in the various organs and tissues until they reach the stage of differentiation characteristic of early first instar larvae. Hatching occurs at about 20 hours.
Although development first becomes abnormal in fsnasA embryos at a very early stage, and cells begin to degenerate almost immediately after their formation, monster embryos with many recognizable larval characteristics may be formed at the end of embryonic development. None of these ever emerge completely from the embryonic envelopes, although some may succeed in rupturing the membranes and partially emerging.
These monster embryos may be divided roughly into two groups—those in which segmentation has been at least partially superimposed over the major furrows (‘segmented’ embryos) (Plate 1, fig. E), and those in which only the furrows are still evident, resulting in an embryo consisting of four or five sections (‘compartment’ embryos) (Plate 1, fig. F). These two groups gradually grade into each other, but as a rule those with only the major furrows are more abnormal in all respects: for example, only 3 / 24 (12·5 per cent.) of the ‘compartment’ embryos showed differentiation of mesodermal tissues, and mesodermal tissues were completely lacking in 4/24 (16·7 per cent.), while the comparative figures for ‘segmented’ embryos were 35/58 (60·3 per cent.) and 0/58 (0·0 per cent.).
Formation of the gut is always affected. No embryo has ever been observed which had either a hind-gut or Malpighian tubules. Owing to the abnormal origin of the posterior mid-gut invagination, the structure of the mid-gut is always abnormal. However, in 18 /56 of the ‘segmented’ embryos, the two rudiments of the mid-gut formed a more or less complete structure around the main yolk mass; this never occurred in any of the ‘compartment’ embryos, where the middle regions of the embryo usually consist of a mass of yolk and undifferentiated or disorganized and pycnotic cells covered by hypoderm-like cells. Rudiments of the posterior mid-gut were present in 41·6 per cent, of the ‘compartment’ embryos and in 76·8 per cent, of the ‘segmented’ embryos. In many embryos the anterior mid-gut developed certain characteristic structures, such as the gastric portion of the proventriculus and the gastric caecae. Union with the abnormal stomodaeum or oesophagus did not always occur, but a recognizable pro ventricular structure sometimes formed. The anterior mid-gut was completely lacking in 12·5 per cent, of the ‘compartment’ embryos but in only 2 per cent, of the ‘segmented’ embryos.
The mesoderm is severely affected. It is capable of differentiation into its major components—fat-bodies, somatic and visceral musculature—in some embryos. The spatial displacements which occur in all embryos make it impossible for complete normal differentiation of either the somatic or visceral musculature to occur, and usually these elements consist of large clumps of partially differentiated cells with slightly different staining reactions and morphology (Plate 2, fig. G). The somatic elements of the gonad can develop.
In two embryos in which head involution was almost normal the ring gland and aorta were formed as well as the frontal sac and Guirlanden cells, indicating that the potential for differentiation into these structures exists, but the circumstances which permit it to be realized occur only rarely in fsnasA embryos.
Development of the nervous system is also affected by the abnormal furrows, and it usually consists of isolated clumps of cells, differentiation usually being best in the brain region. In ‘compartment’ embryos, recognizable ganglion cells are never found in the central yolk-filled compartment, but probably contribute, along with mesoderm cells, to the mass of undifferentiated and frequently pycnotic cells found there.
Hypoderm formation is almost never complete (mirroring earlier developmental events) although its differentiation is usually good. It is most frequently incomplete posteriorly and ventrally where nervous tissue and clumps of somatic mesoderm are exposed to the surface. In the head region the brain is often exposed, a common event in Drosophila mutants when head involution is not normal (see discussion in Counce, 1956b). The main tracheal stems are never formed but isolated bits of trachea with larval differentiation are found in almost all embryos, and posterior spiracles differentiate in a good number. At least one salivary gland formed in most embryos, although the position of the gland or glands was usually distorted. Many contained the normal secretion which stains with Heidenhain’s, indicating that their functional capacity is probably not impaired. Rarely cells which had not invaginated to form the sac-like gland also showed secretory products.
Pole cells or gonads were identified in over half of the ‘segmented’ embryos and in 37-5 per cent, of the ‘compartment’ embryos; this is approximately equal to the number of embryos in which pole cells are formed. In a few instances three small gonads were formed, while in others a single pole cell apparently organized a ‘gonad’ around itself (Plate 2, fig. G). Frequently, only a single gonad formed, but isolated pole cells might be apparent in other regions of the same embryo. Usually the gonad or gonads were in close contact with cells of the posterior mid-gut rudiment.
Cell disintegration is found at this stage in all ‘compartment’ embryos and in over 80 per cent, of the ‘segmented’ embryos. In some ‘compartment’ embryos either the posterior or anterior end may be devoid of cells; these areas resemble the appearance of unfertilized and disintegrating eggs. These are undoubtedly embryos in which large areas of the blastoderm were left uncompleted by the failure of cells to migrate or by the disintegration of cells before a blastoderm formed.
No leaks were observed in these embryos, but at this stage their presence could be masked by the increased breakdown of cellular material.
Cytological abnormalities
Cytological abnormalities occur with a high frequency in fsnasA embryos (cf. Table 1). They first appear during cleavage stages. Nuclei of various sizes form; polyploid mitotic figures are common, and polyploid and multinucleate cells occur in later stages. Nuclei may also clump together in single cytoplasmic islands. This and the common occurrence of polyploidy, spindle interference, and multipolar spindles may indicate that nuclei do not separate far enough at telophase. This is also suggested by the frequency with which nuclei in tandem are observed.
Many embryos appear to cease development in an active mitotic stage. In these embryos spindles soon take on a cloudy appearance, and although they retain their shape, the entire structure stains uniformly, no fibres being discernable. Large multipolar spindles are common; sometimes only clumps of chromatin can be detected distributed randomly within the spindle. In late stages (early death—late fixation) one may find only clumps of chromatin in the cytoplasm of eggs which have the same general appearance as unfertilized eggs. One embryo was found in which at least twelve spindles could easily be identified; however, there were no staining chromosomes with the exception of those in the metaphase plate formed by the polar body chromosomes (Plate 2, fig. H). The number of cytoplasmic islands in the embryo would indicate that more mitotic figures should have been present. It is highly probable that this is not an isolated event. In some of the early lethals it was only possible to make out faint grey areas where spindles had formed; moreover, some embryos which were classified as unfertilized had well-developed discrete cytoplasmic islands (which ‘normal’ unfertilized eggs do not) although no mitotic figures or nuclei could be identified.
Centriole division may also be impaired, for in one embryo a single centriole was apparently functioning as a common pole for three spindles, while in another a nucleus with only a single centriole organizing spindle-fibres was found (Plate 2, fig. I).
Pole-cell nuclei frequently become pycnotic within a few minutes of their formation. This may apply to all the pole cells of an embryo or to isolated cells. The lack of granules in the pole cells has been discussed previously.
Pycnotic nuclei (in addition to those of the pole cells) are found in mesodermal tissues, and in nuclei in abnormal regions at the anterior and posterior ends. Pycnotic nuclei are also found in embryos when development ceases at late cleavage or during blastoderm formation. These embryos also have many other abnormal cell types.
The nuclear and mitotic abnormalities described above have frequently been observed in abnormal Drosophila embryos (see discussion in Counce, 1956a). In addition there is apparently another type of nuclear breakdown which has not been described in earlier publications. In the early stages of the development of these abnormal nuclei, prophase nuclei have large clumps of chromatin in the centre of the nucleus, some of it staining faintly (Plate 2, fig. J); unlike normal nuclei the peripheral areas are apparently devoid of chromatin. There is also a rather indefinable ‘raggedness’ about the chromatin or condensed chromosomes. Later, the chromatin appears to divide into two clumps, one which contains most of the darkly staining material while the other is made up of more faintly staining material (Plate 2, fig. J). These nuclei may undergo division but probably soon break down as figures of abnormal metaphase and anaphase would indicate (Plate 2, fig. K). (Figs. J and K are taken from the same embryo which is hours after fertilization judged both from collection time and stage of development). At a later stage the nuclei take on a cloudy appearance, the nuclear sap apparently giving a positive reaction with Heidenhain’s haematoxylin (Plate 2, fig. L). Such nuclei may be the forerunners of the faintly staining nuclei which are characteristic of the posterior end (Plate 2, fig. M), especially in the region directly above and to the sides of the pole cells. They are often so faint that only very careful study reveals their presence. Single cloudy nuclei may also be found in other regions of the embryo, especially during blastoderm stages. No cloudy nuclei have been observed later than early gastrula stages.
Yolk nuclei also behave very abnormally; they soon become very large, have an unusual staining reaction, and tend to clump together in long strips in the central regions of the embryo (Plate 2, fig. M).
DISCUSSION
Although the mutant factor fsnasA has no visible effects in the adult, it has a very marked effect on the formation of the egg, and on the embryos developing in eggs from mutant mothers. Over half the eggs from fsnasA females contain abnormal amounts of yolk, and the structure of the cytoplasm is also altered. Abnormal eggs have been observed in other female-sterile mutants (Counce, 1956c), although there is a relationship between the type of abnormality and the genotype of the mother. The formation of such eggs is no doubt related to physiological disturbances in the mutant females. There is a possibility that the ring gland is in some manner involved, since it is known that a hormone produced by the gland is involved in the process of yolk deposition.
In the developing embryo the lethal effect offsnasAmay be expressed either before blastoderm formation (early lethals) or, if this early epigenetic crisis is survived, development continues until a monster larva is produced which is not able to survive a second epigenetic crisis, that of hatching. Lethals with more than one critical period are not uncommon (Hadorn, 1955; Counce, 1956a). In the present instance all lethal individuals have a basic similarity in that all are characterized by abnormalities in division rate, in mitotic division and in cell structure. It is probable then that the two classes are closely related, although it is not clear why some survive the first critical period while others do not. The early lethals may develop in a cytoplasmic environment which is more abnormal —it may be that a threshold phenomenon of some sort is involved. Differences in development may be the result of variance between eggs from the same ovary, or perhaps of differences in age, environmental conditions in bottles, or the presence of modifying genes (see also discussion in Ede, 1957).
Whether early or late, the main effect of the gene is probably on the control of cleavage synchrony. Moreover, there appears to be a regular pattern in which disturbances of synchrony occur, suggesting the existence of stratification within the egg cytoplasm. It is not possible to determine whether loss of synchrony is the result of delay in some regions or an acceleration in others or a combination of the two. It is probable that it is the last mentioned which most frequently happens; it is also likely that the differences become magnified in successive divisions, for they stand out more sharply in the late blastema and cellular blastoderm than in the early blastema. At any rate, development of the embryo is severely altered in the 4th dimension, the odd pattern of delay and precocity occurring again in gastrulation and even during the period of histogenesis.
The rapidity and synchrony of cleavage divisions in Drosophila has been frequently discussed (e.g. Sonnenblick, 1950; Goldschmidt, 1955). Painter (1940) proposed that rapid division is possible because the nurse-cells furnish partially assembled ‘chromosomal material’ in the cytoplasm of the egg. During cleavage these partially assembled materials would be utilized rather than a complete synthesis occurring. It is not difficult to invent explanations, with this theory as a basis, as to how this process of reassembly might go awry in female-sterile eggs and seriously affect cleavage. The nurse-cell nuclei are of course homozygous for the mutant factor and may provide in the egg cytoplasm abnormal amounts or kinds of ‘building blocks’; alternatively, some substance which is necessary for the reassembly to occur may be affected.
The cytological evidence shows that the chromosomal cycle is severely affected. ‘Chromosome-less’ spindles found in one embryo, and the abnormal staining reactions of resting stage and prophase nuclei, as well as the ‘raggedness’ of some mitotic figures, may indicate that some stage in the cycle of chromosome duplication itself is basically affected. Other cytological evidence such as multipolar spindles, polyploidy, &c., although feasibly related to abnormalities in chromosome duplication, could also be related to abnormalities in function or structure of the spindle or centrioles. A third possibility is that the cytoplasm is in some way abnormal and prevents a potentially normal division from taking place (cf. Counce, 1956a).
It is important that we have in this mutant some possibility of carrying out further investigations on a cytochemical and physiological level which may give us some insight as to how these eggs differ from normal eggs. This in turn may help us to understand some aspects of the cleavage cycle in Drosophila and perhaps some of the processes involved in mitosis in general.
It is of interest that in addition to these abnormalities in synchrony and their fairly definite pattern in the egg, the polar granules which differentiate the posterior polar plasm from the rest of the periplasm are usually absent. What—if any—connexion there exists between them it is not possible to say, since the function and significance of the polar granules are not understood. One of us (S. J. C.) has studied a II chromosome female-sterility factor in which cleavage synchrony in the zygote becomes abnormal at a much earlier stage (3rd to 4th divisions) and in these eggs the polar granules are apparently normal.
The sex-linked lethal factor X2 (Ede, 1956) located near forked (56·7) has many striking similarities to the pattern of damage in fsnasA embryos. In X2 the departure from normal development is usually first apparent in the gastrula stage, but in two blastoderm embryos Ede observed a striking abnormality: in both a complete blastoderm was formed, but the nuclei were of various sizes and in different stages of mitosis. As in fsnasA embryos, the areas were sharply delimited (Plate 1, fig. A and Ede, Fig. 1). Gastrulation abnormalities are similar in that many deep abnormal furrows form, but in X2 they form only after extension of the germ-band begins. Invagination of the posterior mid-gut is affected in both but, again, the later effect of X2 is apparent, for the rudiment moves as far as the posterior pole before it invaginates. Subsequent developmental abnormalities are similar in the two, no doubt the consequence of early similarities. It should be pointed out that although there are marked similarities, the effects of the two factors vary both in time and space. X2 embryos usually show their first abnormalities in early gastrulation, and the anterior end of the embryo is more severely affected than the posterior end—the converse of the effects of fsnasA in which the posterior end is most severely affected (suggesting the possible existence of a posterior to anterior gradient in differentiation ?). The basic similarities of the two mutants in embryogenesis are probably based on abnormalities in mitotic rates which produce more surface cells than is normal.
The formation of gonads in fsnasA embryos is of interest because of the questions concerning the mode of origin of the gonads in normal embryos. Earlier (reviewed by Sonnenblick, 1950 and Poulson, 1950) it was assumed that pole cells from the first period of migration were converted into yolk nuclei and those which were carried into the embryo in the posterior mid-gut later migrated out of the gut into the body-cavity and there formed the gonads. Poulson (1947, 1950) now believes that the latter cells contribute only to the formation of the middle mid-gut, and that the future germ-cells are pole cells from the first period of migration. These lethal fsnasA embryos, admittedly very abnormal in development, support the interpretation of Poulson. In fsnasA embryos, re-migration of pole cells (frequently all that form) at the first period has been observed in sectioned material; no pole cells are carried into the embryo during the second period because of abnormalities in posterior mid-gut formation. Yet gonads form in some fsnasA embryos, showing that pole cells which enter during the first phase are capable of developing into gonads.
Waddington (1951,1956) and Dalcq (1951) have discussed at some length the importance, for evolutionary processes, of mutations which alter the structure of the egg. Waddington has also pointed out that female-sterility genes in Drosophila provide favourable material for the investigation of such processes. Beatty (1949) has shown that such genes may alter the structure of the egg at any stage, and embryological studies of those affecting development in the egg (Counce, 1956 a, b, c) show further that development within the egg may be affected as early as the processes of fertilization or as late as the stage of larval differentiation of tissues.
Waddington (1951) voiced regret that in Drosophila no mutations ‘… had produced a change [in egg determination] which would allow a viable but abnormal development’. However, it was later shown (Counce, 1956b) that exactly such an effect was connected with the female-sterility gene fused. This gene is lethal to the zygote only if the somatic cells of the mother or the zygote nucleus do not contain the normal counterpart of the locus (Lynch, 1919; Counce, 1956a). It is possible, therefore, to obtain heterozygous females from (1) homozygous fused females mated to non-fused males, and (2) from non-fused females mated to fused males. In such heterozygotes the nuclear constitutions will be identical, the egg cytoplasms genetically different. Examination of adult female heterozygotes of these two kinds showed that abnormalities in abdominal segmentation were 30–40 times more frequent in females which developed from fused ooplasm; further, these abnormalities were connected with abnormalities in musculature and segmentation patterns in the embryo. Undoubtedly, more examples will be found if someone looks for them, and it would seem that genes affecting the ooplasm and the establishment of predetermined patterns in the egg will be the most fruitful area for such a search. So many female-sterile mutants are known in Drosophila that the material for study is almost limitless, and it should be possible to provide a vast array of effects in time and space. Such mutations also have technical advantages for study by such techniques as paper chromatography (Hadorn & Mitchell, 1951), microrespiration studies (Chen, 1951), and cytochemistry of development (Yao, 1949,1950) which have recently been used so successfully in studies of Drosophila development.
ACKNOWLEDGEMENTS
We are indebted to Prof. C. H. Waddington, F.R.S., for encouraging this work and for providing laboratory facilities. One of us (S. J. C.) was the recipient of a Macaulay Fellowship from the University of Edinburgh. We should like to express our gratitude also to Dr. Gamal Nasrat for mutant stocks, to Dr. H. Slizynska for salivary gland analyses, and to Dr. Charlotte Auerbach for help with linkage tests. Valuable technical assistance was provided by the Institute photographer, Mr. Donald Pinkney, and by Mr. James Nelson and Miss Sigrid Seidel.
REFERENCES
EXPLANATION OF PLATES
All figures are of fsnasA embryos from mutant females mated to mutant males. Sections are 5-7 μ in thickness, and are stained with Heidenhain’s iron haematoxylin. Photographs of entire sections are oriented so the anterior end is to the right.
Abbreviations: A, anterior; BLN, blastoderm nucleus; BMUS, somatic musculature; BR, brain; CF, cephalic furrow; co, centriole; CYT, cytoplasm; FB, fat-body; FN, faintly staining nucleus; GB, germ-band; GO, gonad; HEC, head ectoderm; HY, hypodermis; MG, mid-gut; p, posterior; PC, pole cell; PH, pharynx; PMG, posterior mid-gut; PV, proventriculus; PYC, pycnotic nucleus; SM, segment; IR, trachea; VMUS, visceral musculature; VNS, ventral nervous system; YN, yolk nucleus.
FIG. A. Frontal section of blastoderm stage showing differences in cleavage synchrony. Note size differences between nuclei in the anterior and middle regions. Cell-membranes are forming.
FIG. B. Frontal section of posterior tip, showing incomplete blastoderm formation. A broad band of cytoplasm containing yolk and nuclei extends to the tip while at the sides a cellular blastoderm has formed.
FIG. C. Migration of the pole cells through the blastoderm wall ( hours). AU pole cells have migrated in this embryo.
FIG. D. Frontal section of gastrula (−5 hours). The shape of the germ-band has been distorted by deep lateral and ventral folds, and is exposed to the surface at one side (arrow).
FIG. E. Longitudinal section of ‘segmented’ embryo at end of development. Partial head involution has occurred. In the posterior region, segments have formed. The brain is exposed to the anterior surface. The posterior and anterior mid-gut rudiments have fused, and a well-developed proventriculus is present. Note the attenuated shape of the most posterior segment.
FIG. F. Longitudinal section of ‘compartment’ embryo at end of development. The posterior sections consist mainly of yolk and undifferentiated cells bordered by the hypodermis. Some nervous tissue and head ectoderm have formed in the anterior-most section.
FIG. A. Frontal section of blastoderm stage showing differences in cleavage synchrony. Note size differences between nuclei in the anterior and middle regions. Cell-membranes are forming.
FIG. B. Frontal section of posterior tip, showing incomplete blastoderm formation. A broad band of cytoplasm containing yolk and nuclei extends to the tip while at the sides a cellular blastoderm has formed.
FIG. C. Migration of the pole cells through the blastoderm wall ( hours). AU pole cells have migrated in this embryo.
FIG. D. Frontal section of gastrula (−5 hours). The shape of the germ-band has been distorted by deep lateral and ventral folds, and is exposed to the surface at one side (arrow).
FIG. E. Longitudinal section of ‘segmented’ embryo at end of development. Partial head involution has occurred. In the posterior region, segments have formed. The brain is exposed to the anterior surface. The posterior and anterior mid-gut rudiments have fused, and a well-developed proventriculus is present. Note the attenuated shape of the most posterior segment.
FIG. F. Longitudinal section of ‘compartment’ embryo at end of development. The posterior sections consist mainly of yolk and undifferentiated cells bordered by the hypodermis. Some nervous tissue and head ectoderm have formed in the anterior-most section.
FIG. G. Tissue differentiation. The mesoderm has separated into its component elements. The visceral musculature has not become attached to the gut, and has clumped together. Some of the somatic muscle-cells have fused and elongated. A gonad has formed and contains a single germcell; the pycnotic cells in the gonad are probably derived from dying pole cells and mesodermal cells of the gonad.
FIG. H. ‘Chromosomeless’ spindle. The shape of the spindle and the clear area in its centre suggest the presence of some substance in the equatorial region.
FIG. I. Spindle organized by single centromere. The nuclear membrane is still intact at the opposite pole. Normal metaphase at right.
FIG. J. Stages in development of abnormal nuclei, N, normal; spindle forming; B, chromatin tending to clump; staining reaction not normal; c, abnormal; chromatin separated into lightly stained clump and darkly stained clump.
FIG. K. Abnormal polyploid metaphase, polar view. The fuzzy appearance of the chromosomes is apparent.
FIG. L. ‘Cloudy’ nucleus (arrow). Normal blastoderm nucleus at left.
FIG. M. Types of abnormal nuclei, including a typical pycnotic nucleus, and several faintly staining peripheral nuclei. At the right are abnormal yolk nuclei which have clumped together.
FIG. G. Tissue differentiation. The mesoderm has separated into its component elements. The visceral musculature has not become attached to the gut, and has clumped together. Some of the somatic muscle-cells have fused and elongated. A gonad has formed and contains a single germcell; the pycnotic cells in the gonad are probably derived from dying pole cells and mesodermal cells of the gonad.
FIG. H. ‘Chromosomeless’ spindle. The shape of the spindle and the clear area in its centre suggest the presence of some substance in the equatorial region.
FIG. I. Spindle organized by single centromere. The nuclear membrane is still intact at the opposite pole. Normal metaphase at right.
FIG. J. Stages in development of abnormal nuclei, N, normal; spindle forming; B, chromatin tending to clump; staining reaction not normal; c, abnormal; chromatin separated into lightly stained clump and darkly stained clump.
FIG. K. Abnormal polyploid metaphase, polar view. The fuzzy appearance of the chromosomes is apparent.
FIG. L. ‘Cloudy’ nucleus (arrow). Normal blastoderm nucleus at left.
FIG. M. Types of abnormal nuclei, including a typical pycnotic nucleus, and several faintly staining peripheral nuclei. At the right are abnormal yolk nuclei which have clumped together.