Studies on mitochondria in the early development of the slug Arion ater rufus L

Recent research has demonstrated unequivocally that mitochondria are the most vital organelles of the cell (Green, 1964; Lehninger, 1964). Apart from producing energy and metabolites for all the life processes of the cell, it is now becoming increasingly apparent that they may have other specialized functions to perform.

There is considerable evidence to show that mitochondria may play a morphogenetic role in development (Brachet, 1960; Novikoff, 1961; Gustafson, 1965). Their fine structure is also known to change during embryogenesis (Weber, 1962; Yamamoto, 1964; Berg & Long, 1964).

This investigation deals with certain aspects of mitochondrial structure, distribution, cytochemical composition and functional significance with relation to embryonic development.

The slugs were obtained from Aberystwyth, Wales, and were bred outside the Zoology Laboratory, University of Reading. The eggs were gently decapsulated in tap-water using a binocular microscope and the embryos were rinsed in isotonic saline (0·28 % NaCl) and examined by various methods. Mitochondria were studied in ova and in a closely knit series of early embryos up to the postgastrula stage. The following methods were used: (a) classical cytological techniques; (b) phase-contrast microscopy and vital staining; (c) cytochemical methods including enzyme tests on living embryos; (d) electron microscopy.

Centrifuged embryos were also examined by the first three methods. These were obtained by subjecting the entire capsules to a centrifugal force of 950 g for 6 min and then decapsulating them immediately afterwards.

Recent modifications of traditional mitochondrial techniques were used successfully. Embryos were fixed in various fluids like Flemming, Altmann, Champy and Helly and then postchromed in 3 % potassium dichromate for 3–5 days in most cases. After fixation the embryos were washed overnight and quickly dehydrated, cleared and stored in 1 % celloidin in methyl benzoate according to Peterfi’s method (Pantin, 1960). Each embryo was then rapidly embedded in 56 °C ceresin-wax and serial sections (4μ thick) were cut with a Cambridge rocking microtome. Acid fuchsin was widely used for staining mitochondria. Altmann’s method modified by Metzner & Krause (1928) and Kull’s method after Baker (1950) were used. The best results were obtained by the Helly-Metzner combination. Heidenhain’s haematoxylin used in routine work in conjunction with Champy, Helly, Perenyi and Ciaccio was also very useful in the study of mitochondria. Careful differentiation with picric acid, instead of iron alum, produced good results. With acid fixatives, however, mitochondrial form was not so well preserved.

Embryos were squashed in isotonic saline and examined under oil immersion with a Wild M 20 phase-microscope.

This was carried out according to Baker (1958). Embryos were stained in Janus green B, Grubler’s neutral red, Nile blue, toluidine blue and methylene blue for 5–40 min using a 0· 01 % solution of the stain in isotonic saline. These were then rinsed and squashed in saline and examined with an apochromatic oil-immersion objective.

A variety of cytochemical tests were carried out on both living and fixed embryos (Table 1).

For cytochrome oxidase whole embryos, decapsulated without injury and with polar bodies intact, were incubated vitally for 15–45 min in Burstone’s medium (Pearse, 1961), and then washed, compressed and examined in saline, under apochromatic oil-immersion, just as vitally stained preparations are studied—‘vital enzymology’. This vital method proved to be excellent for the study of cytoplasmic inclusions, not only for this enzyme, but for succinic dehydrogenase and other enzymes as well. Localization is very sharp and well defined and the inclusions could be differentiated easily, as in vitally stained preparations, for Brownian movement is still evident. Further the results are highly reproducible. Loss of enzyme activity and artifacts due to fixation and diffusion are virtually non-existent.

For succinic dehydrogenase, Nachlas’s nitro-BT method (Pearse, 1961) and Seligman and Rutenbuig’s néotétrazolium method (Gomori, 1952) were used. Embryos were incubated for periods of (optimum time, h) to avoid diffusion artifacts and examined as for cytochrome oxidase. It was found necessary to use accelerators (Ca²⁺ and Al³⁺ salts) to obtain quicker results.

For other cytochemical tests appropriate fixatives were used and serial sections (4μ thick) were obtained using the same technique employed in the classical methods. All colour reactions were visualized with a Zeiss apochromatic oil-immersion objective.

Illustrations were drawn with a Zeiss camera-lucida and photomicrographs were taken with a Zeiss photo-microscope using Ilford Pan-F film.

Embryos were fixed in plain aqueous 1 % OsO₄ for 1 h at about 2 °C, and then for | h at room temperature (Baker, 1965). These were then embedded in Araldite and sectioned with a Huxley ultramicrotome. Sections were mounted on carbon- and formvar-coated copper grids, stained with uranyl acetate and examined with an Akashi TRS-50 electron microscope.

Mitochondria stain bright red with Metzner’s method while the rest of the cytoplasm is a transparent pale yellow. With Kull’s method the mitochondria stain crimson red, the ground cytoplasm pale yellow, but the yolk has a bluish tinge. Albumen vesicles are pale yellow in both cases. Larger, refractile Golgi bodies, nucleoli and chromosomes are also stained very intensely while asters stand out very clearly.

In general, the mitochondria appear as minute, distinct, rounded bodies, apparently of uniform size, about 0·5μ in diameter. They are found distributed in the cytoplasm forming circles around almost colourless yolk granules, or are seen in the interstices between the yolk and other inclusions. They are sometimes seen in small groups or more rarely arranged in linear rows or chains. In Limnaea, Bretschneider & Raven (1951) describe the formation of small alphagranules (0·5μ in diameter) from more elongate, filiform mitochondria during oogenesis. These perhaps correspond to the rounded mitochondria seen in Arion.

In densely populated areas mitochondria are very closely packed and seem to touch one another. A few appear to be somewhat larger. There are more mitochondria around nuclei situated very close to the nuclear membrane. In the region of the mitotic apparatus, however, there are comparatively fewer mitochondria (Fig. 12). The spindles are almost devoid of mitochondria and very few are seen between astral rays. Mitochondria are arranged to form a distinct ring around the centrioles. All these observations have been confirmed in electron micrographs. Soon after each division a finely granular zone of cytoplasm appears around each newly formed nucleus (Figs. 2, 11). The fine granules are RNA-positive and are probably microsomes. Mitochondria are evenly distributed among these granules but they are more numerous towards the periphery of these masses. The polar bodies too have their due share of mitochondria which have presumably found their way there during the maturation divisions. This may also explain why the tip of the animal pole (AP), soon after the maturation divisions, is devoid of mitochondria.

FIGURES 1-8

Illustrations showing distribution of mitochondria in early embryos.

The pattern of distribution of mitochondria varies somewhat in different stages of development.

Ova. Mitochondria are more or less evenly distributed except in the nuclear region at the AP and in the cortical zone (Figs. 1, 9, 13). The most striking feature is a thin densely packed zone of mitochondria just inside an agranular cortical region within the plasma membrane (Figs. 8, 14–16). Mitochondria then show a gradual decrease in number towards the centre of the ovum. Under high magnification, starting from the periphery, one sees the vitelline and plasma membranes, which may be closely apposed to one another, and then a clear cortical zone (about 1–2μ in thickness) devoid of any granular inclusions. Just beneath this ‘cortex’ is the dense zone of mitochondria (3–4 layers thick) and then the general cytoplasm with fewer mitochondria, yolk and other granules. In the general cytoplasm there is a very gradual decrease in the number of mitochondria as the centre is reached. Very exceptionally one or two yolk granules may be found within the subcortical mitochondrial zone.

Thus there is a fairly well defined but subtle gradient distribution of mitochondria of a centripetal nature, showing a more or less concentric arrangement of these granules in the cytoplasm. One has to look at perfectly flattened sections without asters and nuclei to appreciate this gradient fully. Complete series of sections of six ova at various stages of development, from second maturation division to very early cleavage, have been studied and in all cases this gradient is evident. In later ova, however, there are slightly fewer mitochondria in the subcortical zone. Their earlier distribution suggests that mitochondria originate in the subcortical region and later migrate to other regions as development proceeds. In the cleaving ovum the mitochondria show the same arrangement around asters as described before.

There seems to be hardly any difference between this stage and the late ovum stage. The same gradient is still evident even though there are inter-blastomeric cell membranes. Mitochondria do not crowd on either side of the latter as in the subcortical zone so that the general distribution from the periphery to the centre of the embryo is very much the same as before. There is, however, a slight increase in the number of mitochondria in the subcortical zone of the vegetal halves of the two blastomeres (Figs. 2, 11).

An early 4-cell stage is very similar to a 2-cell stage. But as the third cleavage is approached there is a marked crowding of mitochondria in the subcortical zone at the vegetal pole (Fig. 3). As the third cleavage asters appear the crowding becomes more and more apparent in the vegetal parts of the four blastomeres (Fig. 12).

This is the crucial stage in the early development of Arion. Its most striking and obvious feature is an intensely staining, densely packed, pyramid-shaped aggregation of mitochondria at the vegetal pole (VP) occupying about a fourth of the volume of the megameres (Figs. 4, 17–19). This enormous massing of mitochondria is the most significant event in the process of cytodifferentiation in Arion and has profoundly affected the general pattern of distribution of these granules seen in the earlier stages, especially in the ovum. The appearance of this aggregate can be said to herald a vegetalizing influence. As to its final fate or localization in future embryos, a remarkable series of morphogenetic movements occur whereby these mitochondria come to lie in the cells that will differentiate into future mesoderm and endoderm.

The mitochondrial aggregation occupies a sector of over 60°. The mitochondria are so densely packed that they form a compact mass leaving hardly any spaces in between. Towards the AP they gradually become less and less compact and then merge with the general cytoplasm rather abruptly. A few yolk granules are seen to invade this zone. The vegetal parts of all four megameres (1A–1D) contribute towards the whole mitochondrial mass. This can be easily seen in serial sections and in whole embryos incubated for cytochrome oxidase. These vegetal mitochondria stain very intensely with acid fuchsin and Heidenhain’s haematoxylin and appear as intense chocolate purple masses in oxidase preparations. In sections incubated for phosphatases, they appear as colourless, well-defined areas at the VP. The vegetal mitochondria are also appreciably eosinophil, and phospholipid- and RNA-positive. The aggregation has been very useful in cytochemical studies of mitochondria.

In other regions the distribution of mitochondria appears to conform to that of the earlier stages except that the micromeres and the lateral regions of the megameres appear denser, there being more mitochondria. The original centripetal pattern is, however, still maintained.

The distribution is very similar to that of the 8-cell stage. The vegetal mass is in the same position in the macromeres 2A–2D as it was in the 8-cell stage. The uppermost quarter of micromeres (1a–1d) appear to be more densely-packed while the outer lateral regions of the other micromeres and the megameres, as usual, have more mitochondria so that the general pattern is still the same. In the 16-cell stage the micromeres (1 a–1 d) have divided and the distribution is very similar to that seen in the 12-cell stage.

The mitochondrial aggregation is now in the vegetal parts of the macromeres 3A–3D. The uppermost quartet of micromeres appears to be slightly more densely packed. There is always a tendency towards the accumulation of more mitochondria at the AP, especially in the blastula. This, perhaps, has a functional significance in that the contents of the cleavage cavity are more often expelled at the AP and the cells there are pseudopodially very active. At this stage the cleavage cavity is seen to disappear completely for a time, bringing all the blastomeres in close contact with one another. More mitochondria are seen in the innermost regions of all cells except blastomere 3D, along the margins of contact.

This is a very remarkable stage showing the primary mesoderm cell (4d) being formed and drawn into the cleavage cavity. The vegetal mitochondria are now distinctly in the megameres and a good portion of the mitochondria that were originally in 3D is cut off into 4d. This division of 3D into 4D and 4d occurs at about the 32-cell stage and is an oblique, vertical, unequal division whereby 4d gets about three-quarters of the cytoplasm of 3D. As a result a massive mesodermal cell is formed and then drawn into the cleavage cavity with the help of pseudopodial processes extended by the micromeres (Figs. 5,20). Almost the whole vegetal half of the cell 4d is densely packed with mitochondria with the nucleus holding a prominent and significant position in the centre separating the mitochondrial zone from the inner, more yolky cytoplasm. The free inner boundaries of all the other cells lining the cleavage cavity and the pseudopodial process have more mitochondria. This suggests that mitochondria also abound in those parts of cells that are physiologically active. The cell 4D and the other three megameres which have still not divided are all full of the vegetal mitochondria and they will eventually form endoderm.

In a later stage (40+ cells) when 4d is lying freely in the blastocoele the mitochondria distribute themselves more evenly in its cytoplasm and still later when 4d divides mitochondria are equally divided among the two mesodermal cells.

The mitochondrial aggregation is more dispersed among the future endodermal cells at the VP. These are the derivatives of 4A–4D and 4a–4c. These cells still give an intense reaction for cytochrome oxidase.

Only the cells containing the vegetal mitochondria seem to invaginate to form endoderm (Figs. 6, 21). The invaginating flask-shaped cells have their outer, attenuated, vegetal borders packed with mitochondria, but not so densely as in the earlier stages. Inside the mitochondrial zone is a region corresponding to the neck of the flask, with linear arrays of albumen vesicles with fewer mitochondria in between them. Further inside, occupying the flask, are the nuclei and the denser cytoplasm filled with yolk, mitochondria and other granules. The ectodermal cells have more mitochondria and ingested albumen filling their peripheral cytoplasm (‘ectoplasm’), and yolk and mitochondria in their inner regions (‘endoplasm’), with the nuclei intervening. In general, the outer zones, which were once denser in early embryos, appear less dense than the inner yolky regions, owing to the albumen. Mitochondria are, however, found everywhere around and in between yolk granules and albumen vesicles.

The most significant feature is then the vegetal aggregation in the invaginating endodermal cells. After invagination, as the blastopore closes up, all visible signs of this remarkable mass of mitochondria disappear. Late gastrulae incubated for cytochrome oxidase show no intense reaction at the VP. Only a vague rim around the closed blastopore may be seen. In sections, the vegetal mitochondria, now in the endodermal cells lining the archenteron, are gradually seen to disperse to other parts of the cytoplasm, as they did in the primary mesodermal cell. Once the aggregation has played its part in the differentiation of the germ layers, its functions seem to be over and it gradually fades away into obscurity.

The body and mantle regions are more densely packed while the cephalic vesicle has comparatively fewer mitochondria per unit area, owing to the presence of albumen in both its ectoderm and endoderm (Fig. 22). The vegetal aggregation has completely dispersed in the endoderm and is no longer evident. Most of the endoderm now forms the inner lining of the cephalic vesicle. The cells are massive and have developed huge albumen vacuoles, so much so that their nuclei and cytoplasm with mitochondria have been pushed towards the inside and occupy but a tiny portion of the cell. Mitochondria, however, may be seen right around albumen vacuoles, especially on the outer borders of the cells. On the radial or lateral walls they may be found in single file or are totally absent. Some mitochondria appear to be slightly larger than is usual. The flattened ectodermal cells of the cephalic vesicle have numerous albumen vesicles of assorted sizes, with more mitochondria in between and around them than in the endodermal cells.

The endodermal epithelium which is closely apposed to the invaginating shell gland needs special mention. Cells here are small and very dense and have slightly more mitochondria than even the overlying cells of the shell gland and the rest of the ectoderm. Closer examination (Figs. 7, 23) shows that these endodermal cells have little or no albumen. They have a few yolk granules and more mitochondria, especially towards their outer borders, which are in intimate contact with the shell-gland epithelium. The shell-gland cells have ingested albumen in their outer regions like all other ectodermal cells of the body and they are fairly densely packed with mitochondria. The rest of the body and mantle ectoderm is similar to that of the shell gland in mitochondrial distribution. The mesodermal cells have no albumen but have yolk and as usual they look denser, there being more mitochondria. The ectodermal cells lining the oral invagination have numerous albumen vacuoles but they have also more mitochondria than the other cells of the cephalic ectoderm.

With positive phase the mitochondria appear as rounded, homogenous, fairly dark (brownish) objects, some showing limited Brownian movement. They are very numerous in the vegetal regions especially of 8-cell and 16-cell stages and vary slightly in size. Their distribution conforms to that seen by classical methods. In centrifuged embryos they stratify mainly in the hyaloplasm zone, being more numerous towards the equator. A few mitochondria are also seen in the yolky zone. This has been confirmed by classical and other methods (see centrifugation studies).

The mitochondria are not stainable in early embryos. Staining for up to 2 h does not colour the granules with the dyes used. Many of the other cellular inclusions like Golgi bodies and lysosomes take up most of the dyes very readily. In post-gastrula stages, however, some of the mitochondria are stained with Janus green B, while others remain colourless. This indicates that mitochondria may change appreciably in their physiological state during certain stages of development and also reflects a possible heterogeneity.

In Arion, cytochrome oxidase and succinic dehydrogenase activities were not found exclusively in the mitochondria. Apart from the mitochondria which show intense activity, larger rounded and bean-shaped Golgi bodies are also as intensely positive (Sathananthan, 1970).

Mitochondria and Golgi bodies appear intense chocolate purple while the rest of the cytoplasm is almost colourless. The mitochondria are immobile while the Golgi bodies are active and show vigorous movement and can be easily distinguished from the former. The mitochondria sometimes appear as dense clouds at the VP and around nuclei. Whole embryos are uniformly purple brown with slightly darker rings around the nuclei. The vegetal aggregation appears as an intense chocolate brown area in the megameres at the 8-cell and later stages.

The same inclusions that show oxidase activity are dehydrogenase-positive. But there is apparent a slight over-all decrease in mitochondrial positivity and comparatively fewer Golgi bodies are positive. With néotétrazolium the cytoplasm is colourless to light purple brown, while the mitochondria and Golgi are dark crimson red. With nitro-BT these granules are intense purple while the cytoplasm is light purple to colourless. Nitro-BT is definitely superior as there are no diffusion artifacts and localization is excellent. Even subtle variations in mitochondrial shape can be appreciated.

Whole embryos are uniformly reddish brown with a purple tinge. The vegetal aggregation in the 8-cell stage is not intensely positive although, in later embryos, the mesodermal and endodermal cells appear a shade more intense.The regions around the blastopore and the mouth are, however, intensely positive.

Mitochondria stain intensely with Heidenhain’s haematoxylin and are feebly eosinophil. They are stained fairly intensely with Sudan black B, which can be attributed to the presence of lipids, chiefly their membrane-bound phospholipids. The vegetal mitochondria are appreciably pyroninophil. Most of the pyroninophil reaction could be removed by pretreating with ribonuclease. Mitochondria give negative reactions when tested for peroxidase, phosphatases, polysaccharides, muco-substances, DNA, calcium, iron and melanin (Table 1).

In centrifuged ova, the mitochondria stratify chiefly in the more centripetal hyaloplasm zone forming a very broad band (Fig. 10). The dividing line at the equator between the heavier yolky zone and the hyaloplasm is very sharply defined. The mitochondria are crowded more towards the equator and decrease gradually in number as the oil cap zone is approached. There is invariably an almost clear region apparently devoid of granules, just below the fat cap. A few mitochondria, however, are also seen in the heavier yolky zone, where they are found evenly distributed between the yolk granules, or form rows or chains round the yolk granules. Thus the stratification of these granules is incomplete.

When stratified eggs are incubated for cytochrome oxidase, the sites of enzyme actively correspond closely to the distribution of mitochondria. The hyaline zone gives an intensely positive reaction, the region above the equator being most intense. On closer examination the mitochondria in the hyaloplasm zone are the chief sites of enzyme activity. Those in the yolky region and the Golgi bodies there are also positive. Similar, though less striking results are obtained on incubation for succinic dehydrogenase.

The mitochondria were found to show remarkable differences in fine structure during very early development. The mitochondria in early ova (just prior to the second maturation division) are fairly dense and their internal structure is rather indistinct (Figs. 25, 26). Those in the astral region are particularly dense (Fig. 24), and somewhat resemble the dense ‘cytoplasmic bodies’ seen in mammalian eggs (Hadek, 1965). The distribution of mitochondria in the astral zone closely corresponds to that seen under the light microscope. Mitochondria and other granules are generally excluded from the inner astral zone. There are, however, a few mitochondria forming a ring round the centre of the aster, while a few others are seen between the inner astral rays. The outer astral rays extend into the granular cytoplasm, where they are seen to disappear between many mitochondria and yolk granules. The mitochondria associated with the aster are also smaller (about 0·3μ than those in other parts of the egg. In those mitochondria which are larger and more electron-lucid a few cristae are seen projecting into the matrix or extending right across the organelle rather haphazardly. Dense intramitochondrial granules are rarely seen. In maturing ova the mitochondria are mostly spherical or oval in shape, measuring about 0·5–0·6μ in maximum diameter. Very rarely, elongate forms are seen reaching about 1μ in length. The mitochondria in mature ova have a better-defined internal structure and are not so electron-dense (Fig. 27). The cristae are more abundant, fairly distinct and often extend right across the granule, transversely, diagonally or very rarely longitudinally. Intramitochondrial granules are seen in a few cases. There is an abundance of mitochondria in the subcortical zone confirming the lightmicroscope observations. A few are seen very close to the plasma membrane (about 0·1–0·2 μ from it). The yolk granules, Golgi bodies and lipid droplets are seen some distance away from the surface.

Tn the 8-cell stage (Figs. 28–31) mitochondria are highly polymorphic and very queer in appearance. This stage, as we have already seen, is of great morphogenetic importance in the development of Arion, when an increase in mitochondrial number and an extraordinary segregation of mitochondria takes place at the vegetal pole. These polymorphic mitochondria could well be those from the vegetal aggregation. Mitochondria are extremely abundant in certain regions and some are enormous, reaching a maximum diameter of about 1μ. The smaller mitochondria are roughly rounded or oval and resemble those seen in mature ova, while the larger ones are mostly irregular in shape, there being spherical, triangular, club-shaped, rod-like and dumb-bell shaped forms. On the whole they are very irregular in size and shape and are comparatively larger, almost twice or thrice as large as those seen in the ovum. The presence of elongate, dumb-bell-shaped forms (Fig. 29) may strongly suggest that they divide by a process akin to binary fission. Most mitochondria have an elaborate system of well-defined cristae, often irregularly spaced within their matrices. In others the cristae are more compacted and mote regularly arranged—transversely, diagonally or longitudinally—and stretch across the entire organelle, leaving little space for matrix. Denser granules are found in the matrix or are sometimes seen attached to the cristae or inner membrane.

Another salient feature is that ergastoplasmic cisternae and vesicles are often intimately associated with mitochondria and yolk granules. Certain mitochondria (Figs. 29, 30) have vesicular elements of the endoplasmic reticulum closely apposed to their outer membranes, sometimes enveloping the whole granule. In others, elongate cisternae seem to extend from one granule to another. Such an association was also seen in a few cases in ova, but was not so evident.

From the foregoing detailed survey of mitochondria, where over fifty embryos have been studied in serial sections by classical methods and numerous others by vital, cytochemical and ultrastructural methods one striking and illuminating fact emerges, namely that the mitochondria seem to play a very dramatic morphogenetic role in the early development of Arion.

In invertebrate development there are many instances where mitochondria have been found to be involved with the process of early development (Raven, 1958a; Brachet, 1960; Novikoff, 1961; Gustafson, 1965). Mitochondrial segregation, in particular, seems to be an important morphogenetic factor in most cases.

In molluscs, segregation of mitochondria has been reported in a few cases. Reverberi (1958) followed the distribution of mitochondria in Dentalium and demonstrated an accumulation of these granules in the polar lobe. Polar lobes, as we know, play an important part in morphogenesis and if removed cause abnormalities. The polar lobe mitochondria are always retained in the D quadrant and finally pass into the cell 4d which forms the mesoderm. These mitochondria may be likened to the vegetal mitochondria in Arion. A distinct segregation of mitochondria during early development was also reported in Sphaerium (Woods, 1932) and this was later traced to cell 4d. The distribution of mitochondria in cleavage and gastrula stages of Limnaea (Raven, 1945, 1946) is somewhat similar to that seen in Arion. That in the ovum, however, differs very markedly, for the animal pole plasm was not seen in Arion. A vegetal aggregation was also absent in Limnaea.

In the egg of the annelid Tubifex mitochondria accumulate in the polar plasm, which was shown to have great morphogenetic significance by centrifugation experiments. These mitochondria were traced later to somatoblasts 2d and 4d (Lehmann, 1956, 1958; Lehmann & Mancuso, 1957).

Considerable work has also been done concerning mitochondria in ascidians (Ries, 1937, 1939; Berg, 1956, 1957; Reverberi, 1956, 1957). From all their observations it is evident that the posterior blastomeres (megameres) of 4-and 8-cell stages require the presence and activity of mitochondria. These mitochondria seem to play a very important role in differentiation—in this case the differentiation of mesoderm and then muscle.

Recently Gustafson (1965) has reviewed the morphogenetic significance of mitochondrial and other gradients in sea-urchin embryos at a biochemical level. In the light of his work and that of Hôrstadius (1955) on reduction gradients, and of many others, an interesting parallel could be drawn in the case of Arion. Although the morphogenetic pattern is different in the sea urchin certain general comparisons could be made with regard to mitochondrial distribution.

In the sea urchin there are evidently two gradients in mitochondrial distribution appearing at different stages of development. The first is an animal-vegetal (basipetal) gradient which manifests itself at the mesenchyme-blastula stage and the second is a stronger vegetal-animal (acropetal) gradient which appears in the gastrula. It was concluded that development is controlled by the interaction of these two oppositely directed and mutually antagonistic gradients—a concept outlined by Runnstrom as early as 1928.

In the ovum, mitochondria are distributed along a centripetal gradient with more mitochondria in the periphery and gradually decreasing numbers towards the centre (Fig. 32A). The condition is much the same in the 2-cell and 4-cell stages, although one sees a gradual increase in the number of mitochondria in the vegetal hemisphere just below the cortical region. Then there is the dramatic segregation of mitochondria at the VP soon after the third division which profoundly alters the existing pattern.

If an 8-cell stage is examined very closely, forgetting for a moment that the vegetal segregation ever existed, it is seen that the original gradient pattern is virtually unaffected. As in the earlier stages, there are more mitochondria towards the periphery, decreasing in number as one approaches the centre of the embryo. This is true irrespective of the presence of a cleavage cavity and inter-blastomeric cell membranes. The appearance of the vegetal mitochondria merely seems to emphasize the pattern already existing in that region of the embryo. Owing to the formation of the cleavage cavity nearer the AP the micromeres have more mitochondria as they have inherited most of the densely packed cytoplasm that was originally in the outer upper half of the ovum. The macromeres have inherited most of the inner core of cytoplasm with fewer mitochondria, and the rest of the dense outer cytoplasm and of course the vegetal aggregation. The eccentric position of the cleavage cavity, then, would appear to suggest that there is now an animal-vegetal gradient, whereas it is in reality a part of the original centripetal gradient field. The lateral outer parts of the macromeres also show more mitochondria and are as densely packed as the micromeres. With the appearance of the vegetal mass of mitochondria one may now visualize a stronger vegetal-animal gradient opposing the original centripetal gradient field, whose centre of focus has now shifted very slightly towards the AP and seems to be exerting its influence as if it were an animal-vegetal gradient. This interpretation brings Arion more or less in line with the echinoderms.

It must, however, be reiterated that, in this context, the so-called animalvegetal gradient is really a centripetal gradient field with its centre of activity extending throughout a curved surface, which is the subcortical zone of the animal seven-eighths of the embryo (Fig. 32B). This indeed is the region that will eventually differentiate for the most part into ectoderm. So regarding this as an animal-vegetal gradient is justified. The vegetal-animal gradient, however, is more axial in nature.

In later stages, especially in the invaginating gastrula, the two gradients can be visualized in the more familiar manner (Fig. 32C). There are as before the two mutually antagonistic gradients, animal and vegetal, represented as in the case of the echinoderms.

Raven (1958 b) has also elucidated the influence of gradient fields in the development of Limnaea and he concludes that the organization of the egg is governed by the interaction of axial and cortical gradient fields.

Whatever the interpretation may be, the mitochondrial aggregation which manifested itself after the third cleavage has obviously a vegetalizing influence and undoubtedly has a precise morphogenetic function in Arion. It is very likely that animalizing and vegetalizing influences interact with one another to bring about differentiation of the germ layers, as in Limnaea and in the echinoderms.

The number of mitochondria seems to be more or less constant during very early cleavage, although a slight increase may be noted prior to the third division. At the 8-cell stage there is a dramatic rise in the population due mainly to the appearance of the vegetal mass. An increase in mitochondrial number during early cleavage has also been reported in Limnaea (Raven, 1958a). As to the origin of the vegetal mitochondria, many of them seem to have originated at the VP and appear to radiate upwards, towards the AP. It is also likely that at least some of them arose in the subcortical zone of the vegetal hemisphere in the earlier stages and later migrated towards the VP. As elongate dumb-bell-shaped forms were seen in electron micrographs of mitochondria, it is probable that they arose from pre-existing mitochondria by growth and fission. The recent demonstration of DNA in mitochondria in a variety of cells, both adult and embryonic (Nass, Nass & Afzelius, 1965), lends support to this view. However, the possibility that at least some mitochondria could have originated from the cell membrane (Robertson, 1964) needs to be carefully examined.

Between the 8-cell and the blastula stage there seems to be hardly any rise in the mitochondrial population. The general crowding beneath the cortical zone, seen in the earlier stages, gradually disappears in blastulae and gastrulae. The vegetal mitochondrial population does not seem to change substantially and these mitochondria are distributed among the future mesoderm and endoderm cells. At gastrulation there is considerable growth, and tests for RNA suggest that there is an active period of protein synthesis during invagination. Although the embryo has taken in appreciable amounts of albumen the cells are still fairly densely packed. So it is possible that there is a slight over-all increase in mitochondrial number at this stage. A rise in the mitochondrial population is also seen in the sea urchin at the onset of gastrulation (Gustafson, 1965). After gastrulation copious amounts of albumen are taken in, and although the embryo grows in size the mitochondrial population seems to be much the same as in the gastrula stage. The mitochondria are more densely distributed only in the visceral and body regions of the post-gastrula.

Recent work indicates that cytochrome oxidase and succinic dehydrogenase are almost exclusively intramitochondrial and closely bound to the structure of mitochondria (Lehninger, 1964). In Arion, however, Golgi bodies also show intense activity of these enzymes (Sathananthan, 1970).

Almost all the mitochondria in Arion show cytochrome oxidase activity. This is understandable as this enzyme is the final common pathway of most oxidative processes in the cell. There was an increase in cytochrome oxidase activity at the VP, coinciding with an increase in mitochondrial number in that area. Whether or not this reflects a higher rate of respiration should be determined quantitatively. In sea urchins, respiration increases during cleavage and it rises appreciably at the onset of gastrulation when an increase in mitochondrial number is observed at the VP. Berg & Long (1964) have shown that the vegetal mitochondria are larger, have more cytochrome oxidase activity and therefore have a more intense energy metabolism. The condition in Arion could well be similar to that seen in echinoderms. Some of the mitochondria in the 8-cell stage are enormous and have a complicated internal structure. This may also account for the increase in cytochrome oxidase activity at the VP.

When 8-cell stages are incubated for succinic dehydrogenase the vegetal aggregation of mitochondria does not show a very intense differential staining reaction as in the case of cytochrome oxidase. There is, then, a possible heterogeneity, as was indicated by Novikoff (1961) in liver cells. Nachlas, Walker & Seligman (1958) have found differences in mitochondrial stainability in kidney cells using nitro-BT. Colourless mitochondria-like granules have been seen, especially in later embryos of Arion. These observations suggest that mitochondria may not be all alike biochemically, and may show differences in staining reactions at different times. This seems to be more likely in the case of developing embryos where changes in number, form and fine structure have been noted.

There seems to be some correlation between dehydrogenase activity and Janus green stainability. This stain gives poor results with sea-urchin embryos. The mitochondria seem to be virtually unstainable in the early embryos of Arion. In later stages, however, some mitochondria are stainable. This again reflects a possible heterogeneity. Gustafson (1965) has clearly shown that mitochondrial stainability is related to their physiological state.

In Arion, mitochondria are frequently located near a supply substrate such as lipid droplets or glycogen granules or around albumen vesicles and yolk granules. This is undoubtedly associated with their metabolic functions. Lipid droplets have been shown to be closely associated with mitochondria by cytochemical methods (Sathananthan, 1966). This has been partly confirmed by electron microscopy (Fig. 28). More intimate associations between mitochondria and lipid droplets have been reported in the sea urchin (Brachefi 1960, fig. 65). It is now known that mitochondria are capable of synthesizing and oxidizing fatty acids to completion (Lehninger, 1964). Free glycogen is found evenly distributed in the cytoplasm but accumulations are seen around nuclei and asters and in other parts where mitochondria are abundant. Mitochondria are also known to be involved in the metabolism of other carbohydrates. This may explain why they are seen around albumen vesicles and yolk spherules. Albumen vesicles contain lipid, galactogen and various mucopolysaccharides (Sathananthan, 1968), while the yolk granules are also chemically complex, being glyco-lipo-protein in nature (Sathananthan, 1970). The mitochondria do not seem to play a direct role in yolk formation in Arion, but together with the ergastoplasm they may be involved in its breakdown and absorption.

Mitochondria are also known to be abundant in cells or parts of cells, where activity is intense. They were found to be more numerous in the innermost parts of the blastomeres lining the cleavage cavity in the blastulae of Arion. An interesting parallel is seen in the contractile vacuoles of Protozoa, which are surrounded by a cluster of mitochondria (Mercer, 1965). The subcortical aggregation of mitochondria in early stages is also significant where the activities of the cell membrane are concerned. This membrane is physiologically very active in that molecules are constantly passing in and out through it. It is also involved in a process akin to pinocytosis (Sathananthan, 1968). This partly explains why mitochondria are found crowded below the cortex. It is also known that in muscle, mitochondria form tightly packed columns between the muscle fibrils and provide energy for muscle contraction and movement. In Arion there is considerable pulsatory and pseudopodial activity in the micromeres due partly to the active rhythmic expulsion of the contents of the cleavage cavity at the AP. There are also precise morphogenetic movements during gastrulation, where micromeres were found to extend pseudopodia to the macromeres and some of these pseudopodia are involved in the actual withdrawal of the primary mesodermal cell into the cleavage cavity. Mitochondria are found to be abundant in all parts of the micromeres which show pseudopodial activity. There seems little doubt that the activities of the cleavage cavity, plasma membrane and the pseudopodia all require energy and it is obvious that the ATP produced by the mitochondria provides this energy.

Electron-microscope studies of embryos of Arion seem to indicate that mitochondria do undergo changes in form and fine structure during early development. The mitochondria in the early embryos show a peculiar fine structure differing markedly from those of many other cell types. They also seem to have undergone progressive changes during very early cleavage and are the only organelles that have profoundly altered during early differentiation. An over-all increase in mitochondrial number and size and an increase in the number of cristae have taken place by the time the 8-cell stage is reached. Regional differences in size and fine structure of mitochondria seem to exist in ova. Those in the central astral zone are smaller and rather dense and have a poor internal structure, while those in other regions, including the cortical zone, have better internal organization. So it seems likely that there are also structural and size differences along a centripetal gradient in Arion.

Changes and regional differences in the ultrastructure of mitochondria have been reported in certain animals. In Tubifex, Weber (1962) found an abundance of mitochondria in the polar plasm and cortical regions. These results are similar to those obtained in Arion. The polar plasm is of great morphogenetic significance and could be likened to the vegetal aggregation of mitochondria in Arion. In the sea urchin Berg & Long (1964) showed that larger mitochondria with more cristae were found in the invaginating vegetal cells of the early gastrula. The numerous large, polymorphic mitochondria seen in some micrographs of the 8-cell stage could well be from the vegetal region. Yamamoto (1964) has also found marked alterations in the structure of mitochondria in fish embryos.

The enlargement of mitochondria, their increase in numbers, and the increase in complexity of their internal organization, no doubt reflect higher levels of mitochondrial activity. These events seem to coincide with their morphogenetic segregation in Arion. The increase in cytochrome oxidase activity at the 8-cell stage is also perhaps due to a higher level of energy metabolism. The intimate association of mitochondria with vesicles and cisternae of the ergastoplasm is further evidence of their increased activity.

Recent studies have shown that mitochondria contain appreciable amounts of intrinsic RNA (De Robertis, Nowinski & Saez, 1965). This has been observed in Arion also, especially in the vegetal aggregation of mitochondria. Electron micrographs show well-developed cisternae and vesicular elements of the ergastoplasm associated with mitochondria, which are, no doubt, partly responsible for this RNA-positivity. The presence of RNA in mitochondria may be very significant in the context of protein synthesis. The generally accepted view is that mitochondrial enzyme proteins are synthesized by the ribosomes attached to the endoplasmic reticulum and later incorporated in mitochondria. Recent biochemical work, however, suggests that mitochondria may be able to synthesize some of their own membrane proteins (Lehninger, 1964). The discovery of intramitochondrial DNA in many types of cells with the electron microscope (Nass et al. 1965) makes it seem more likely now that mitochondria may be in fact sites of protein synthesis. Studies on intramitochondrial DNA were not made in Arion. However, if mitochondria prove to be universally DNA-positive they could be self-reproducing organelles and they could also play a part in the hereditary process. It is then conceivable that they are even better equipped to perform a morphogenetic function in developing embryos.

On account of their probable morphogenetic significance there is reason to suppose that mitochondria, together with nucleic acids, may play a part in induction. In gastropods, the formation of the shell gland is thought to be due to the inductive action exerted by the tip of the archenteron on the ectoderm, with which it makes intimate contact (Raven, 1964). In Arion, the small-celled endoderm at the tip of the archenteron has comparatively more mitochondria than the invaginating shell-gland epithelium. Mitochondria were found to be slightly more numerous on either side of the margins of contact of the two epithelia especially in the endoderm cells. Although little is definitely known about the mechanism of induction it is thought that certain chemically active substances (inducing agents) diffuse from the endoderm cells into the ectoderm cells. The mitochondria may perhaps contribute some of these substances in the form of enzymes or they could produce the energy for the inductive process in the form of ATP and play a part in active transport of these substances from cell to cell.

In this context it is also worth mentioning certain observations on the early blastula of Arion. The 24-cell stage is unique in that the cleavage cavity completely disappears for a time, bringing all the cells together in the centre of the embryo. There is intimate contact between the micromeres and macromeres and this happens just before gastrulation. Curiously enough, mitochondria are found to be more numerous in the innermost regions of all the blastomeres (except 3D), where there is mutual contact. In addition there is an accumulation of free RNA in the zones of mutual contact (Sathananthan, 1966). It appears that all the blastomeres, so to speak, meet in the centre as if to decide the fate of one of their fellows, which presumably is cell 3D, which soon buds off the primary mesoderm cell. It is plausible that all this points towards an additional chemical inductive process or interaction between cells in the central zone whereby the fate of 3D and perhaps that of the other macromeres are decided. Of course, the main morphogenetic influence is exerted by the vegetal aggregation of mitochondria, which interacts with other animal influences in the gradient system. In fact, when the cells come together the animal and vegetal influences can interact with each other more effectively than when there is a fluid-filled cleavage cavity between them. Thus both intracellular and intercellular interactions could well be finally responsible for the differentiation of mesoderm and endoderm.

In conclusion, from all the above considerations it is evident that in Arion there seems to be a straightforward case for the morphogenetic significance of mitochondria. In addition to their energy-linked functions they seem to play specific directive roles in development. Segregation of mitochondria in particular seems to be an important factor in morphogenesis. Even if differential distributions of mitochondria are normally attributed to levels of general metabolic activity, it has been shown beyond reasonable doubt that the vegetal aggregation of mitochondria plays a precise directive role in Arion. Many avenues available to a better understanding of their morphogenetic significance have been explored. But a quantitative biochemical approach and more experimental studies are necessary before the exact significance of mitochondria in morphogenesis is fully appreciated.

La structure, la distribution, la nature cytochimique et la signification physiologique des mitochondries ont été étudiées au cours de l’embryogénèse de la limace Arion depuis l’œuf jusqu’au stade postgastrulaire.

Les mitochondries semblent subir des changements progressifs mais profonds dans leur forme et leur structure fine au cours des premières phases du développement et leur distribution s’établit suivant un gradient qui ressemble plus ou moins à celui de l’Oursin.

L’événement le plus significatif de la cyto-différenciation est l’apparition d’une extraordinaire aggrégation de mitochondries au stade des 8 blastomères et suggèrent une influence végétalisante témoignant sans nul doute d’une fonction morphogénétique très précise. Cette aggrégation végétative est entraînée dans une remarquable série de mouvements morphogénétiques et ses mitochondries sont ultérieurement incorporées dans les cellules qui se différencient en mésoderme et endoderme. Donc, à côté de leur rôle énergétique, les mitochondries semblent jouer un rôle directeur, de façon dramatique, dans le développement et, en particulier, la ségrégation mitochondriale semble être un facteur important dans la différenciation.

Certains aspects de leur distribution et de leur composition enzymatique sont discutés en relation avec leurs fonctions métaboliques normales.

La signification du RNA et du DNA mitochondrial pour la synthèse des protéines et la duplication des mitochondries, ainsi que les rôles possibles des mitochondries dans les phénomènes d’induction sont également discutés à la lumière des travaux récents.

 
• AV

  albumen vesicles

 
• C

  cortical region

 
• Ect.

  Ectoderm

 
• End.

  Endoderm

 
• ER

  ergastoplasm

 
• G

  Golgi

 
• GC

  granular cytoplasm

 
• HZ

  hyaloplasm zone

 
• L

  lysosome

 
• LG

  lipid globules

 
• M

  mitochondria

 
• Mes.

  mesoderm

 
• OC

  oil cap

 
• P

  pseudopodium

 
• PM

  plasma membrane

 
• SCM

  subcortical mitochondria

 
• SG

  shell gland

 
• VM

  vegetal mitochondria

 
• ViM

  vitelline membrane

 
• Y

  yolk granule

 
• YZ

  yolky zone

I am deeply indebted to Professor Alastair Graham, D.Sc., and Dr Vera Fretter, D.Sc., for providing the facilities for my work at the University of Reading, England, and for their invaluable guidance, constructive criticism and encouragement.

I am also grateful to Dr John R. Baker, F.R.S., for helping me a great deal with electron microscopy at the University of Oxford; and to Professor T. Gustafson of the University of Stockholm, Sweden, for his valuable advice.

My thanks are also due to Mr J. M. McCrae, Cytological Laboratory, Oxford University, for technical assistance in electron microscopy; and to the Department of Sedimentology, University of Reading, and to Mr C. H. Chang of the Science Faculty, University of Ceylon, Colombo, for photographic assistance.