J. Embryol. exp. Morph. Vol. 19, 2, pp. 203-15, April 1968 2 0 3With 7 platesPrinted in Great Britain

Yolk distributionand utililization during early development of a

teleost embryo {Brachydanio rerio)

By ROBERT J. THOMAS1

From the Department of Biochemistry and Biophysics,Iowa State University

The work of Oppenheimer (1934, 1936) and Tung (1955, see also Devillers,1961) suggests that an embryo such as Fundulus, a marine teleost, explanted atearly cleavage (two-cell stage) is more dependent upon the amount of periblastyolk included in an explant than is an embryo that is more advanced prior toexplanation. In other words, 'if it (the embryo) is explanted before a so-calledcritical stage, the blastoderm turns into a hyperblastula (a non-differentiatedmass of cells); if it is explanted after that, it undergoes differentiation. Thecritical stage corresponds to eight blastomeres in Carassius, thirty-two inFundulus, and a young blastula in Salmo' (Devillers, 1961, p. 391). The zebrafish,Brachydanio rerio, with a smaller egg than these fish, has not been characterizedby explanation. Devillers, in summary, states that 'these results mean that asubstance indispensable for differentiation exists in the yolk sphere' (p. 392).He continues,' On the other hand, how this hypothetical material may reach theblastoderm needs to be explained. The base of the blastoderm is in direct contactwith the periblast in the early stages; later on, this syncytium 'buds' offblastomeres that add themselves to the embryonic disc' (p. 397). He then asksthe question 'but then how can one explain that in later stages the syncytiumcan still impose an orientation on the germ? Are diffusing organizing substancesinvolved?' (p. 397).

Several questions may be further suggested from Devillers' initial question.After the first horizontal cleavage, is more yolk being added to the cells of theblastoderm? How is the yolk added to the cells of the blastoderm? What is inthe yolk that may be the indispensable substance suggested by Devillers ? Thisstudy will attempt to answer these questions.

1 Author's address: Department of Biological Sciences, State University of New York atAlbany, New York 12203, U.S.A.

204 R. J. THOMAS

MATERIALS AND METHODSThe developing embryo

Mature zebrafish, Brachydanio rerio (Hamilton-Buchanan), were obtainedfrom commercial hatcheries,. In the laboratory, the fish were kept in balancedaquaria at 25 °C with a day-period of 14 h (tungsten illumination). Both livebrine shrimp (Artemia salina), hatched in the laboratory from dry eggs, and dryfishfood were fed to the fish daily.

For breeding, a gravid female and several mature males were confined in aplastic breeding trap kept at 26 °C in an aquarium without sand or vegetation.Breeding usually occurred, if at all, within the first hour of 'day' illumination(Hisaoka & Firlit, 1962 a). Fifty to 300 closely developing eggs were found in thebottom of the spawning tank and placed in a medium-sized dish for observationand use. The use of very closely developing populations from a given batch ofeggs was accomplished by further selection of subspawns chosen by taking eggswhich progress from the 1-cell to a 2-cell stage within a 60 s period. The develop-mental stages of B. rerio enumerated and timed by Hisaoka, Ott & Marchese(1957) and Hisaoka & Battle (1958) for 26 °C. incubation were used. Embryosfrom early high blastula (stage 8) to closure of the blastopore (stage 17) werefixed for studies of normal morphology or were exposed to experimental milieuand subsequently fixed.

Experimental treatments

Developing embryos at selected stages were exposed to either 49-5 % deuteriumoxide or 10~3M colchicine (Calbiochem) in aquarium water for various timeperiods. Parallel subspawn samples of eggs with chorions torn to provide freeraccess of solutions were exposed to the abnormal milieu to determine theavailability of the milieu to the embryo. For the materials used, the chorion didnot hinder availability to the embryo. In all cases, some of the controls werepermitted to hatch to test the viability of the spawn.

Fixation of embryo

During the study of normal events, eggs were fixed for 30 min with 1 %osmium tetroxide with either 3 x 10~3M or 10~2M calcium chloride buffered withveronal acetate at pH 7-5; or 1 % hydroxyadipaldehyde buffered with phosphateat pH 7-5, rinsed in phosphate buffer at pH 7-5 and postfixed in 1 % osmiumtetroxide buffered with phosphate at pH 7-5 (Sabatini, Bensch & Barnett, 1963);or with 1 % hydroxyadipaldehyde buffered with sodium cacodylate at pH 7-6,rinsed in sodium cacodylate buffer at pH 7-6 and postfixed in 1 % osmiumtetroxide buffered with veronal acetate at pH 7-6. Experimentally treatedmaterial was fixed only in 1 % osmium tetroxide with 10~2M calcium chloridebuffered with veronal acetate at pH 7-5. All fixatives were adjusted to a finalosmolarity of 172 m-osmoles (calculated Powell, Philpott & Maser, 1964;

Teleost yolk 205Maser, Powell & Philpott, 1967). The fixed eggs were dehydrated through anascending series of graded ethanols. Eggs were embedded in Epon resin (Luft,1961). Thick sections were hand-cut with razor blades or with glass knives on aReichert Ultramicrotome OMU-2 for preliminary observations and correlationof stage morphology to the in vivo observations just prior to fixation. Hisaoka &Firlit's study (1960) of sectioned paraffin-embedded B. rerio eggs was used as abasis of comparison. Thick sections were observed and photographed underphase-contrast optics. Selected embryos were sectioned at 40-65 m/i with anLKB Ultrotome with a Dupont diamond knife or a Reichert UltramicrotomeOMU-2 with glass knives and mounted on 75 x 300-mesh unfilmed copper grids.All sections for electron microscopy were stained with lead citrate (Reynolds,1963) prior to observation on an RCA EMU-3F electron microscope operated ateither 50 or 100 kV.

RESULTSYolk distribution: normal

At early high blastula (stage 9) the blastomeres are situated on the yolk mass(Plate 1, fig. A). Yolk, as round opaque particles, is visible at low magnificationswith phase-contrast optics and is located at the periphery of all cells (Plate 1,fig. A). After the first horizontal cleavage, large yolk particles (4/A maximumdiameter) as well as yolk particles equal in size or smaller than the 0-5 /i mito-chondria are found in the cytoplasm.

Below the blastomeres in the periblast, cytoplasm that is contiguous with theyolk mass (Plate 2, fig. A) contains large closely packed particles of yolk.Ribosome-like particles can be seen in high numbers in the blastomeres andbetween the yolk particles of the periblast. Each yolk particle is surroundedby a membrane (M) if an appropriate cross-section is observed (Plate 2, fig. B).The interior of the yolk particle appears uniform under phase-contrast opticsPlate 1, fig. A). At the electron-microscopic level of observation, membranousvesicles (V) can occasionally be seen within the yolk (Plate 2, fig. B).

In cells of the late high blastula, yolk particles are located peripherally to thenucleus or the mitotic apparatus (Plate 1, fig. B). The number of yolk particlesin cells near the periblast is higher than in cells toward the surface of the embryo.The periblast contains a large variety of yolk particle sizes (Plate 3, fig. A) ascompared to those in the periblast at early high blastula (Plate 2, fig. A). Infortuitous section, contact of the blastomeres (B) with the periblast syncytium(P) shows no membrane-membrane contact but rather a row of vesicles separatingperiblast from blastomeres (Plate 3, fig. B).

Yolk particles are found to be more numerous in the cells close to theperiblast from late high blastula (Plate 1, fig. C) to the one-third epiboly stage(Plate 1, fig. F). Various sizes of yolk particles are in the periblast syncytiumin these same stages. However at one-half epiboly (Plate 1, fig. G), only randomlyscattered cells have definite yolk particles. The periblast yolk of the one-half

14 JEEM 19

206 R. J. THOMAS

epiboly embryo (Plate 1, fig. G) appears similar to that of the early high blastulaembryo (Plate 1, fig. A). Various-sized yolk particles present in the periblast ofembryos from late high blastula to one-third epiboly stage are absent at the one-half epiboly stage.

Throughout the entire development period from late high blastula to one-halfepiboly stage, yolk particles that were first seen in all cells of the early highblastula embryo are missing from cells furthest from the periblast.

Yolk distribution: experimental

Eggs at the eight-cell stage were exposed to either 10~3M colchicine or 49-5 %deuterium oxide. In a period of 3 h the control embryos progressed to the flat/very-late blastula stage (Plate 4, fig. A). Their gross morphology is indistinguish-able from other material at the same stage used in this study (Plate 1, figs. C, D).

Colchicine exposure results in immediate mitotic arrest (Plate 4, fig. B).Cleavage did not occur in the exposed eggs. Some yolk particles are seenwithin the cells but are fewer and smaller than those seen in an early highblastula embryo. The periblast is enlarged and small yolk particles are near theyolk mass (Plate 4, fig. B).

After 3 h exposure to deuterium oxide, the gross morphology of the blastulais similar to that of the flat/very-late blastula, but the cells have a diameter morelike late high/flat blastula embryos (compare Plate 4, fig. C to the normal cellsize in Plate 1, fig. C). Intercellular spaces are absent (Plate 4, fig. C). Manymitotic figures with prominent poles are present. The quantity of yolk particlesin cells near the periblast appears the same as yolk particles in the controlembryos. The periblast is greatly enlarged with more small yolk particles in theperiblast than is seen in the control embryos (compare Plate 4, fig. C to control,Plate 4, fig. A).

PLATE 1

Figs. A-G, x 125.

Fig. A. Early high blastula embryos have uniformly distributed yolk after bipolar differentia-tion. Note the absence of small yolk particles in the periblast syncytium.Fig. B. Late high blastula embryos have smaller volume cells. Yolk particles in cells near theperiblast are more numerous than in cells away from the periblast.Fig. C. Flat blastula embryos have smaller volume cells.Fig. D. Very late blastula cells near the periblast contain the majority of yolk particles.Mitotic figures in various phases are frequent.Fig. E. At early gastula stage, the blastoderm is lower. Small yolk particles in the periblast andin cells near the periblast are higher in number than in cells away from the periblast.Fig. F. One-third epiboly stage embryos still show small yolk particles in the periblast and incells near the periblast.Fig. G. At one-half epiboly stage, the marginal periblast has several prominent nuclei. Smallyolk particles in the periblast are absent. Only scattered cells above the periblast contain yolkparticles.

/. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 1

R. J. THOMAS facing p. 206

J. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 2

R. J. THOMAS

J. Embryo!, exp. Morph., Vol. 19, Part 2 PLATE 3

Fig. A. Membrane-bound yolk particles within yolk particles are found in the periblast ofthe late high blastula embryo. Hydroxyadipaldehyde buffered with sodium cacodylate fixation;lead citrate stained, x 14600.Fig. B. Blastomeres (B) are in 'intimate' contact with the periblast syncytium (P). There isno definitive membrane-membrane separation but rather a row of vesicles. Hydroxyadipalde-hyde buffered with sodium cacodylate fixation; lead citrate stained, x 8800.

R. J. THOMAS

J. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 4

R. J. THOMAS

J. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 5

Figs. A, C, E, x 125. Figs. B, D, F, x 320.Figs. A, B. Control embryos after 12 h from late high blastula were all at one-half epibolystage. Small yolk particles are absent from the periblast cytoplasm. Nuclei are easily seen inthe periblast (arrow, fig. B).Figs. C, D. Colchicine-exposed late high blastula embryos are arrested with cell diameters ofcells of the late high to flat blastula. Nuclei are absent. Note the almost complete absence ofyolk particles in cells near the periblast.Figs. E, F. Deuterium oxide-exposed embryos have progressed to only the one-third epibolystage. Little yolk is present in the cells of the embryo.

R. J. THOMAS

J. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 6

. - . ,. •*•

R. J. THOMAS

J. Embryol. exp. Morph., Vol. 19, Part 2 PLATE 7

R. J. THOMAS

Teleost yolk 207Six hours after initial exposure to the experimental milieu, the control embryos

(Plate 4, fig. A) are at one-third epiboly stage. The colchicine-exposed embryosdid not advance morphologically (not shown), and deuterium oxide-exposedembryos (Plate 4, figs. E, F) showed cytolysis (breakdown of the cell membrane).The tangential section of a deuterium oxide-exposed embryo shows numerousperiblast nuclei and broken cell membranes (arrow, Plate 4, fig. F). Deuteriumoxide exposure was terminated because of disintegration of the embryos.

As the control embryos continued to closure, 12 h after initial exposure of theexperimental embryos, the colchicine-exposed embryos remained morphologi-cally unchanged. There was no indication of cell movement or epiboly.

Other eggs were initially exposed at a later developmental stage, early highblastula (stage 9), to either colchicine or deuterium oxide for a period of 4 h.During the period of exposure the control eggs progressed to the very lateblastula stage (Plate 4, fig. G).

Colchicine-exposed blastulae (Plate 4, fig. H) have cells of varying diameters,and mitotic figures are absent. Many small yolk particles are found in theperiblast yet they are absent in the larger volume cells. Both intact nuclei andfree chromosomes are found throughout the blastula. The general cytoplasmappears unchanged from the controls.

Deuterium oxide-exposed eggs progressed to the late high/flat blastula stageas judged by cell diameters (Plate 4, fig. I). Intercellular spaces are absent. Yolkparticles are less frequent in all cells (Plate 4, fig. I, compared to control, Plate 4,fig. G). However, the amount of yolk is less in the deuterium oxide-exposed cellswhen compared to an embryo with a similar cell diameter (compare Plate 4,fig. I, to Plate 1, figs. B, C).

Late high blastula embryos were exposed to either colchicine or deuteriumoxide for a period of 9 h. During the period of exposure, the control embryosprogressed to the one-half epiboly stage (Plate 5, figs. A, B). The marginal peri-blast is at the equator of the embryo and an enveloping layer covers the loosecells of the embryo. Yolk particles are in scattered cells and the periblast doesnot have any small yolk particles. Only periblast nuclei (arrow, Plate 5, fig. B)and large yolk particles of the main yolk mass comprise the periblast.

Colchicine-exposed embryos treated as above (Plate, 5, figs. C, D) arearrested at late high/flat blastula stage (stage 10/11; Plate 1, figs. B, C). Mitoticfigures are absent; cells are almost completely free of yolk particles.

PLATE 2

Fig. A. Periblast yolk of the early high blastula embryo is mainly large membrane-bound yolkparticles. Cytoplasm can be seen between some yolk particles in the periblast. Osmiumtetroxide with calcium fixation; lead citrate stained, x 5350.Fig. B. At a higher magnification, the membranes (M) around the yolk particles are readilyapparent. Some yolk particles are within others but separated by their own membranes.Vesicles (F) can be seen in some yolk particles. Osmium tetroxide with calcium fixation; leadcitrate stained, x 16600.

14-2

208 R. J. THOMAS

After 9 h exposure to deuterium oxide, an embryo exposed at late high blastulastage has progressed only to the one-third epiboly stage (compare Plate 5, fig. E,to Plate 1, fig. F). Cells contain significantly less yolk than a similarly advancedembryo. The periblast contains a large number of small yolk particles.

Yolk fate

In the periblast some membrane-bound yolk particles nestled within others(Plate 2, fig. A; Plate 3, fig. A) are suggestive of large yolk particle fragmentation.

In cells of all stages from late high blastula to one-half epiboly, some yolkparticles (0-5-1-5 fi in diameter) are seen with various ratios of yolk and cen-trally located membranous elements and ribosome-like particles (Plate 6, figs.A, B, Plate 7, figs. A, B).

A cytoplasmic structure (4-7^ average diameter) found in a single cell(Plate 7, fig. B) suggested the ultimate fate of yolk particles. The interior of thiscomplex structure contains numerous membranous vesicles and ribosome-likeparticles. Separated from the cytoplasm by a membrane, a thin layer of materialsimilar in density and structure to a yolk particle (arrows, Plate 7, fig. B) isdelineated from the interior by a second membrane.

DISCUSSION

Bipolar differentiation, continuing into late cleavage stages of B. rerio(Roosen-Runge, 1938), results in all cells having a complement of yolk particles.The film of B. rerio cleavage by Hisaoka et al. (1957) shows streaming of materialtoward the blastomeres until early blastula. At early high blastula only largeyolk particles (4 /i) and yolk particles the size of mitochondria or smaller werecarried into the blastomeres. The first horizontal cleavage of the fertilized eggseparates the blastomeres from the yolk within the periblast. The cell membraneof the fertilized egg after the horizontal cleavage encompasses the periblastnuclei and cytoplasm. The periblast cytoplasm contains membrane-bound yolkparticles and is contiguous with cytoplasm between yolk particles of the yolkmass. The space between the yolk mass and the periblast syncitium observed inFundulus by Lentz & Trinkaus (1967) is also apparent in the phase micrographs(Plate 1, figs. B-F). However, electron micrographs of the periblast cytoplasm andyolk (Plate 2, fig. A) show no such space. The space present between the periblastcytoplasm and the yolk mass should be considered' halo' from the phase-contrastoptical system.

We may assume that the yolk in any given cell will be utilized almost as readilyas in a neighboring cell. In this study, mitotically active cells with peripherallylocated yolk were found on the central periblast of the yolk mass. Marrable'sdata (1965) shows that B. rerio has a high mitotic index (83 %) and a uniformdistribution of mitosis at blastula stages. These data would suggest, excludingthe occurrence of yolk transfer from the periblast, that uniform distribution andusage of intracellular yolk should be expected. However, yolk is diminished in

Teleost yolk 209cells away from the periblast, and yolk particles are consistently seen in highconcentration in cells near the periblast.

Two possibilities may be considered for the occurrence of yolk in cells nearthe periblast. Yolk particles were never found extracellularly. The direct transferof yolk by phagocytosis is excluded as a possible mechanism by Lentz &Trinkaus (1967), who have reported that pinocytosis and phagocytosis are'apparently absent in Fundulus\ The occurrence of vesicle boundaries (Plate 3,fig. B) between blastomeres and periblast is suggestive of either transfer ofcytoplasm across the boundary or vesicle cleavage at the end of periblast mitosis.In the blastoderm, both furrow cleavage and vesicle cleavage have been seenwith either osmium tetroxide with calcium or hydroxyadipaldehyde fixation.Humphreys (1964) has shown similar vesicle cleavage boundaries with Mytilusedulis polar bodies when polar body material is drawn into the cytoplasm of thecell. Therefore, the other possible mechanism is transfer of yolk during mitosisfrom the periblast syncytium.

Deuterium oxide, with a very large 'isotope effect', causes mitotic cellsimmersed in deuterium oxide (70-96 %) to be immediately arrested regardlessof progress through the mitotic cycle (Gross & Spindel, 1960 a, b; Marsland &Zimmerman, 1963, 1965). In this study, exposure of B. rerio embryos to con-centrations of deuterium oxide (49-5 %), which would retard but not arrestmitosis, allowed gross normal development of all embryos. Developmentaladvancement, when cell diameters are compared (Marrable, 1965), was alwaysretarded; deuterium oxide exposure resulted in blastomeres that were swollen and

PLATE 4

Figs. A-E, G-I, x 125. Fig. F, x 320.

Fig. A. A flat to very late blastula embryo is developed from the eight-cell stage after 3 h.Note the yolk particles in cells near the periblast. This embryo is a control for embryos shownin fig. B and C.Fig. B. An eight-cell embryo exposed to colchicine for 3 h does not advance developmentally.Yolk particles are infrequent and smaller in size.Fig. C. Embryos exposed to deuterium oxide for 3 h advance to late high to flat blastulastages when cell size is compared to normally developed eggs. Intercellular spaces are absentand the periblast is enlarged. Mitotic figures have prominent aster-like structures.Fig. D. The control embryo for 6 h exposure to deuterium oxide proceeds to the one-thirdepiboly stage.Figs. E, F. A tangential section of an embryo exposed to deuterium oxide for 6 h shows brokencell membranes (arrows, fig. F) and many nuclei in the periblast.Fig. G. The control embryo for embryos exposed to colchicine or deuterium oxide at earlyhigh blastula is at the very-late blastula stage after 4 h.Fig. H. A colchicine-exposed embryo has a blastoderm of varying cell sizes. Yolk particlesare absent in the large cells near the free surface of the embryo.Fig. I. Deuterium oxide-exposed embryos have progressed to the late high to flat blastulastage embryos. Yolk in cells away from the periblast is diminished.

210 R. J. THOMAS

had reduced or missing intercellular spaces, and mitotic apparatuses that hadprominent astral regions. The effect of deuterium oxide is therefore similar tothat described by Gross & Spindel (1960 a, b).

Exposure of eight-cell embryos to deuterium oxide resulted in otherwisenormal development, even yolk utilization, until cytolysis occurred. Early highblastula embryos exposed to deuterium oxide lagged developmentally and showedfewer yolk particles in the blastoderm as compared with the control. Late highblastula embryos proceeded to the one-third epiboly stage as the control embryosproceeded to the one-half epiboly stage. Significantly, the cells of deuteriumoxide-exposed embryos did not contain as many yolk particles in the cells as insimilarly advanced embryos.

The developmental lag caused by exposure to deuterium oxide, althoughpossibly due to slowing of the mitotic cycle, would probably not cause cessationof yolk utilization. If yolk utilization did continue in the blastomeres withouta contribution of yolk-containing cells to the blastoderm by the periblast, thefinding of cells with less yolk in the embryo would be expected. Indeed, thedeficiency of yolk in deuterium oxide-exposed eggs did occur. Therefore, theconclusion may be advanced that mitosis from the periblast does occur andresults in yolk appearing in cells near the periblast. However, this hypothesiscannot be confirmed by exposure of the embryos to deuterium oxide because ofcytolysis in some embryos exposed for 4 h or longer.

Some colchicine-exposed blastulae with a thick blastoderm are found withlarge, surface cells but smaller deep cells. A surface cell is immediately exposedto colchicine, but a cell deep within the blastoderm may be affected by colchicinelater due to penetration difficulties. This phenomenon of rapid exposure is mostapparent at the eight-cell stage when all cells are immediately exposed or atlate high blastula when the blastoderm is thin; early high blastulae have asignificantly thicker blastoderm which would present a problem of penetration.In embryos which have cells that are considered rapidly exposed to colchicine,no interphase nuclei are evident. However, in the early high blastula embryowhere exposure is considered variable because of penetration, interphase nucleiare found.

Rapid arrest at mitosis by colchicine could be expected to affect any mitosisthat might occur from the periblast but not affect yolk utilization. Thereforethe question of yolk transfer from the periblast syncytium to the blastoderm canbe experimentally tested. If yolk is transferred to the cells of the blastoderm byany means other than mitosis, yolk could be expected in some blastomeres inspite of mitotic arrest. However, in embryos rapidly affected by colchicine, theeight-cell and late high blastula embryos, yolk is relatively absent.

Early high blastulae exposed to colchicine, however, have reduced quantitiesof yolk at the blastula surface where cells are arrested mitotically but normalamounts of yolk in cells close to the periblast which are unaffected by colchicine.If yolk is transferred by periblast mitosis, colchicine-arrested cells near the

Teleostyolk 211surface should contain a reduced amount of yolk as compared to untreated cellsnear the periblast.

However, if a yolk complement was given to cells during bipolar differentia-tion and subsequently utilized and new cells with yolk were formed from theperiblast, we would expect to find cells without yolk in untreated cells furthestfrom the periblast and cells close to the periblast with a complement of yolk.This situation is observed in normally developing embryos but is not seen inembryos where mitosis has been uniformly stopped by colchicine.

As suggested by Oppenheimer (1934,1936), Tung (1955) and Devillers (1961),some substance which carries information (or material) necessary for thedifferentiation processes after a 'critical' stage is within the yolk. Identificationof the components of yolk should provide important information on the natureof the 'yolk-substance' necessary for differentiation.

Yolk particles as seen in this study in one-third epiboly cells and earlier havebeen observed initially to have only amorphous yolk material surrounded by aunit membrane, then to have membranous elements and ribosome-like particlesin the center of membrane-bound yolk material, and to appear finally as ahighly complex particle appearing as cytoplasm surrounded by little yolk and aunit membrane. These yolk particles containing various quantities and types ofinclusions could be arranged as suggested above in a sequential pattern of utiliza-tion. Bellairs (1958) has seen a similar, but more complex, sequential ordering ofyolk particles in the chick embryo at gastrulation and later. Strikingly, the finalmember of the suggested sequence appears as an isolated portion of cytoplasmsurrounded by a thin layer of yolk and an external unit membrane. Thus theproposal that a supply of maternal membranous elements and ribosomes be thedetermining factor from the periblast necessary for permitting differentiationat later development stages is a highly suggestive hypothesis.

Hisaoka & Firlit's work (19626) suggests the occurrence of RNA in a maskedform in yolk of the mature oocyte. After fertilization, RNA concentrationremains constant histochemically in all cells during cleavage and blastulastages except for regional decreases during interphase and prophase (Hisaoka &Firlit, 1961).

In Rana pipiens, Brown & Caston (1962) report the binding and masking ofexogenous isotope-labeled ribosomes to yolk in studies of in vitro proteinsynthesis during early development. In R. pipiens, Karasaki (1963) notes the

PLATE 6

Fig. A. Yolk particles have a uniform granular appearance at high magnification exceptwhere membranous vesicles are within the yolk. Osmium tetroxide with calcium fixation; leadcitrate stained, x 42100.Fig. B. Yolk particles are seen with varying amounts of membranous material and ribosome-like particles. A sequence of these yolk particles may be used to describe the order of yolkutilization. Osmium tetroxide with calcium fixation; lead citrate stained, x 13700.

212 R. J. THOMAS

decrease in yolk in the embryo, but he does not report ribosome-like particleswithin Rana yolk platelets. Rounds & Flickinger (1958) conclude, however, fromtheir study of neural induction in R.pipiens that yolk breaks down in chordames-oderm with a subsequent release of RNA. Bellairs (1958) reports micro-particlesin yolk similar to the cytoplasmic micro-particles now known as ribosomes. Inthe insect fat body, Locke & Collins (1965) have found isolation of membranesand ribosomes in 'Protein + RNA granules' similar in appearance to B. rerioyolk particles. Also in the sequestering insect fat body, granules fuse to formlarger 'fat' particles within cells; in contrast, in the periblast of B. rerio duringyolk utilization, large yolk particles appear to 'pinch, off' into smaller yolkparticles. These developing systems are considered diverse by Williams (1967)and cannot be directly compared to teleost embryos.

The recent work of Perry (1967) on the morphological distinction betweenglycogen and ribosomes shows the possibility of confusion between the twomolecules. The ribosome-like particles in this study are about 16-19 /i indiameter, within the range for ribosomes characterized by Perry.

The conclusion is strongly suggested that membranous material and ribosome-like particles in the yolk of B. rerio originate maternally and are necessary fordifferentiation. In addition, protein (yolk ?)-masked m-RNA as found recentlyby Stavy & Gross (1967) in Echinoderm embryos might be also expected inyolk of B. rerio.

Therefore, the several questions initially proposed from Devillers' questionmay be answered as: yes, more yolk is added to the cells of the blastoderm byperiblast mitosis. The 'indispensable' substance contained in yolk necessary fordifferentiation may be the maternal contribution of membranes and ribosome-like particles within yolk.

SUMMARY

1. Yolk is found evenly distributed in all cells of the blastoderm after bipolardifferentiation at the early high blastula stage. Small yolk particles are commonin the periblast until the one-third epiboly stage. Cells farther away from theperiblast during this period contain yolk particles which are decreased in numberand smaller in size. At one-third epiboly to one-half epiboly stage, small yolk

PLATE 7

Fig. A. The yolk particle in the center of the micrograph contains ribosome-like particlessimilar to ribosomes of the cytoplasm. Osmium tetroxide with calcium fixation; lead citratestained, x 16500.Fig. B. The ultimate fate of a yolk particle may be the structure which dominates this micro-graph. Membrane-bound yolk, similar to yolk in a smaller particle (arrows) is separated fromthe interior of the particle by a second membrane. The interior of the particle appears similarto the cytoplasm outside the yolk particle. Osmium tetroxide with calcium fixation; leadcitrate stained, x 18700.

Teleostyolk 213particles in the periblast are absent, and only scattered cells of the embryocontain yolk. The hypothesis is suggested that mitosis from the periblast providesnew blastomeres with a yolk complement and is the mechanism of yolk transferto the blastoderm.

2. Exposure of B. rerio to 49-5% deuterium oxide solution causes a slowingof normal development, a swelling of blastula cells, and ultimately cytolysis.Although development and mitosis are slowed down, yolk utilization continues.Cells away from the periblast contain smaller and fewer yolk particles thancells of a control embryo with similar cell diameters. Mitosis from the periblastis slowed down by deuterium oxide and could not then provide yolk-containingcells to the blastoderm as rapidly. From this observation it is inferred thatmitosis from the periblast is providing cells to the blastula with a yolk com-plement.

3. Colchicine causes immediate cessation of mitosis and uniform develop-mental arrest if penetration is rapid. Arrested cells are found with much less yolkthan control cells of equal size. Cessation of mitosis and the absence of yolktransfer is consistent with the hypothesis that cells with a yolk complementoriginate from the periblast by mitosis.

4. Yolk particles found in B. rerio cells are seen initially to have only amor-phous yolk material surrounded by a unit membrane, then to have membranouselements and ribosomes in the center of membrane-bound yolk material, andto appear finally as a highly complex particle that appears to have cytoplasmsurrounded by a little yolk and a unit membrane. The above order is suggestedas a sequential pattern of yolk utilization.

5. Yolk, besides providing soluble nutrients to the embryonic cells, con-tributes maternal membranous and ribosome-like material to the blastomeres.

RESUME

Repartition et utilisation du vitellus au cours des premiers stadesdu developpement d'un embryon du Teleosteen Brachydanio rerio

1. Le vitellus se trouve reparti egalement entre toutes les cellules du blasto-derme apres la differentiation bipolaire au premier stade de la blastula avance.De petites particules vitellines se trouvent communement dans le periblastejusqu'au stade du tiers de l'epibolie. Des cellules plus eloignees du periblaste aucours de cette periode contiennent des particules vitellines moins nombreuses etplus petites. Du stade du tiers a celui de la moitie de l'epibolie, il n'y a pas depetites particules de vitellus dans le periblaste et seules des cellules disperseesdans l'embryon contiennent du vitellus. On suggere l'hypothese que les mitosesa partir du periblaste fournissent de nouveaux blastomeres charges de vitelluset qu'il s'agit la du mode de transfert du vitellus au blastoderme.

2. L'exposition de B. rerio a une solution a 49,5% d'eau lourde (D2O)provoque un ralentissement du developpement normal, un gonrlement des

214 R. J. THOMAS

cellules de la blastula et finalement la cytolyse. Bien que le developpement et lesmitoses soient tres ralentis, l'utilisation du vitellus se poursuit. Les celluleseloignees du periblaste contiennent des particules de vitellus plus petites etmoins nombreuses que celles des cellules d'un embryon temoin a diametrescellulaires semblables. Les mitoses a partir du periblaste sont tres ralenties parl'eau lourde et ne pourraient done pas fournir de cellules vitellines au blastodermeaussi rapidement. De cette observation, on deduit que les mitoses du periblastefournissent a la blastula des cellules chargees de vitellus.

3. La colchicine fait cesser immediatement les mitoses et provoque un arretgeneral du developpement si sa penetration est rapide. Les cellules bloqueesrenferment beaucoup moins de vitellus que les cellules temoins de taille egale.L'arret des mitoses et l'absence de transfert de vitellus sont en accord avec l'hypo-these selon laquelle les cellules chargees de vitellus prennent naissance parmitose a partir du periblaste.

4. Les particules vitellines observees dans les cellules de B. rerio n'ont d'abordqu'une substance amorphe entouree par une membrane unitaire, puis renfer-ment des elements membranaires et des ribosomes au centre du materielvitellin limite par une membrane, et apparaissent finalement comme des particuleshautement complexes contenant du cytoplasme entoure par un peu de vitelluset une membrane unitaire. On suggere que la succession precedente representele processus sequentiel d'utilisation du vitellus.

5. Le vitellus fournit des materiaux nutritifs solubles aux cellules embryon-naires, et apporte en outre aux blastomeres du materiel membranaire et dumateriel de type ribosomique tous deux d'origine maternelle.

This work was supported by Research Grants HD-2585 and HD-3762 to Dr L. EvansRoth from the National Institute of Child Health and Human Development, United StatesPublic Health Service.

REFERENCES

BELLAIRS, R. (1958). The conversion of yolk into cytoplasm in the chick blastoderm as shownby electron microscopy. / . Embryol. exp. Morph. 6, 149-61.

BROWN, D. D. & CASTON, J. D. (1962). Biochemistry of animal development. I. Ribosomesand protein synthesis in early development of Rana pipiens. Devi Biol. 5, 412-34.

DEVILLERS, C. (1961). Structural and dynamic aspects of the Teleostean egg. In Advances inMorphogenesis (ed. Abercrombie and Brachet), vol. i, pp. 379-428. New York: AcademicPress.

GROSS, P. R. & SPINDEL, W. (1960a). Mitotic arrest by deuterium oxide. Science, N.Y. 131,37-9.

GROSS, P. R. & SPINDEL, W. (19606). Heavy water inhibition of cell division: an approach tomechanism. Ann. N. Y. Acad. Sci. 90, 500-22.

HISAOKA, K. K. & BATTLE, H. I. (1958). The normal developmental stages of the zebrafish,Brachydanio rerio (Hamilton-Buchanan). / . Morph. 102, 311-28.

HISAOKA, K. K. & FIRLIT, C. F. (1960). Further studies on the embryonic development of thezebrafish, Brachydanio rerio (Hamilton-Buchanan). / . Morph. 107, 205-25.

HISAOKA, K. K. & FIRLIT, C. F. (1961). The distribution and localization of ribonucleicacid in the zebrafish embryo as determined by histochemical methods. Am. Zoologist1, 359.

Teleost yolk 215HISAOKA, K. K. & FIRLIT, C. F. (1962 a). Ovarian cycle and egg production in the zebrafish

Brachydanio rerio. Copeia 4, 788-92.HISAOKA, K. K. & FIRLIT, C. F. (19626). The localization of nucleic acids during oogenesis

in the zebrafish. Am. J. Anat. 110, 203-15.HISAOKA, K. K., OTT, J. N. & MARCHESE, A. N. (1957).Time lapse studies on the embryology

of the zebrafish. Anat. Rec. 128, 565.HUMPHREYS, W. J. (1964). Electron microscope studies of the fertilized egg and the two-cell

stage of Mytilus edulis. J. Ultrastruct. Res. 10, 244-62.KARASAKI, S. (1963). Studies on Amphibian yolk. 5. Electron microscopic observations of

yolk platelets during embryogenesis. / . Ultrastruct. Res. 9, 225-47.LENTZ, T. L. & TRINKAUS, J. P. (1967). A fine structural study of cytodifferentiation during '

cleavage, blastula, and gastrula stages of Fundulus heteroclitus. J. Cell Biol. 32, 121-38.LOCKE, M. & COLLINS, J. V. (1965). The structure and formation of protein granules in the fat

body of an insect. / . Cell Biol. 26, 857-84.LUFT, J. H. (1961). Improvements in epoxy resin embedding methods. / . biophys. biochem.

Cytol. 9, 409-14.MARRABLE, A. W. (1965). Cell numbers during cleavage of the zebra fish egg. J. Embryol. exp.

Morph. 14, 15-24.MARSLAND, D. & ZIMMERMAN, A. M. (1963). Cell division: differential effects of heavy water

upon the mechanisms of cytokinesis and karyokinesis in the eggs of Arbacia punctulata.Expl Cell Res. 30, 23-35.

MARSLAND, D. & ZIMMERMAN, A. M. (1965). Structural stabilization of the mitotic apparatusby heavy water in the cleaving eggs of Arbacia punctulata. Expl Cell Res. 38, 306-13.

MASER, M. D., POWELL, T. E. Ill & PHILPOTT, C. W. (1967). Relationships among pH,osmolality, and concentration of fixative solutions. Stain Technol. 42, 175-82.

OPPENHEIMER, J. M. (1934). Experiments on early developing stages of Fundulus. Proc. natn.Acad. Sci. U.S.A. 20, 536-8.

OPPENHEIMER, J. M. (1936). The development of isolated blastoderms of Fundulus heteroclitus.J. exp. Zool. 72, 247-69.

PERRY, M. M. (1967). Identification of glycogen in thin sections of amphibian embryos. / . CellSci. 2, 257-64.

POWELL, T. E. Ill, PHILPOTT, C. W. & MASER, M. D. (1964). On the pH and osmolality offixative components. / . Cell Biol. 23, 110 A.

RAVEN, C. P. (1958). Morphogensis: The Analysis of Molluscan Development. New York:Pergamon Press.

REYNOLDS, E. S. (1963). The use of lead citrate at high pH as an electron opaque stain inelectron microscopy. / . Cell Biol. 17, 208-12.

ROOSEN-RUNGE, E. C. (1938). On the early development—bipolar differentiation and cleavage—of the zebrafish, Brachydanio rerio. Biol. Bull. mar. biol. Lab., Woods Hole 75,119-33.

ROUNDS, D. E. & FLICKINGER, R. A. (1958). Distribution of ribonucleoprotein during neuralinduction in the frog embryo. / . exp. Zool. 137, 479-99.

SABATINI, D. D., BENSCH, K. & BARRNETT, R. J. (1963). Cytochemistry and electron micro-scopy. The preservation of cellular ultrastructure and enzymatic activity by aldehydefixation. / . Cell Biol. 17, 19-58.

STAVY, L. & GROSS, P. R. (1967). The protein-synthetic lesion in unfertilized eggs. Proc. natn.Acad. Sci. U.S.A. 57, 735-^2.

TUNG, T. C. (1955). Acta Biol. exp. sin. 4, 129. Cited in Devillers (1961).WILLIAMS, J. (1967). Yolk utilization. In The Biochemistry of Animal Development (ed. R.

Weber), pp. 343-82. New York: Academic Press.

{Manuscript received 31 August 1967, revised 17 October 1967)