Morphogenetic disturbances from timed inhibitions of protein synthesis in Fundulus

Oviparous teleost fish such as Fundulus heteroclitus have many advantages for the analysis of macromolecular events in embryogenesis. Eggs are available in large numbers. Their development can easily be made isochronous. Thus they constitute an appropriate vertebrate system upon which molecular biological studies can be carried out quantitatively.

In a wide variety of organisms development from fertilization to the high blastula appears to be controlled by informational RNA previously synthesized in the ovum and thus on maternal templates (reviewed by Davidson, 1968 and Tyler & Tyler, 1970). The teleosts are no exception to this rule (Wilde & Crawford, 1966, 1968; Kafiani, 1970). However, in certain teleosts an RNA synthesis essential to normal morphogenesis begins immediately upon fertilization (Crawford & Wilde, 1966). Its expression in the control of morphogenesis begins at gastrulation and continues from that time. The integrity of the temporal control of the serial order of morphogenetic transcription is essential for normogenesis.

The temporal control of protein synthesis and the integrity of its serial order in early embryogenesis may also be an essential mechanism in morphogenesis. We report here the effects on morphogenesis of interference with protein synthesis in the early zygote.

Fundulus heteroclitus (Linnaeus) was obtained in very large numbers by seine in estuarine waters of Frenchman Bay and kept in floating live cars in sea water. Ripe eggs were obtained by manual stripping into membrane filtered (0·45 μm pore size) 50% sea water, sperm by mincing dissected testes in filtered 50% sea water. Gametes were mixed for 60 sec as measured by stop watch. Shortly before the end of this period zygotes were washed quickly several times. During the breeeding season gametes are uniformly functional. At 30 sec over 90% of eggs are fertilized as tested by previous experiments and by contemporary controls. Embryos developed at 16°C in an incubator in filtered 50% sea water which was changed daily.

Development was regular in controls. When care is taken, with proper culture methods in clean water and glassware, anomalies are extremely rare. Hatching takes place at 40 ± 2 days. Stage references are to the normal tables of Oppenheimer (1937), Armstrong & Child (1965) and the resolution of differences by New (1966). Stage reference numbers are always referred to Oppenheimer, which correlated with the unpublished color photographs of Wilde.

Embryos were fixed in full strength formalin directly from unmanipulated specimens. Fixation in this manner insured rapid penetration of the chorion. Fixed specimens were then dissected free from their chorions, dehydrated in the usual ethanol series, doubly imbedded in methyl benzoate celloidin and paraffin and sectioned at 6μm. Sections were stained in Delafield’s hematoxylin and fast green and mounted in the usual manner.

This method of fixation was adopted since dissection in vivo of the chorion is difficult. Prior fixation permits dissection after very delicate cellular structures have been hardened. This tends to protect anomalous embryos from accidental distortion and artifact production. The yolk itself still presents a problem as it resists infiltration and has a tendency to ‘shingle’ or crack during sectioning. Often such shards are deposited on top of the cellular masses, organs and parts under study.

Photographs of living specimens were taken using an MP-3 camera (Polaroid-Land Corporation) and either Pola-color or ASA 3000 polaroid film with a Wild-Heerbrugg dissecting microscope and a monocular tube using reflected illumination against a black background.

Histological sections were photographed on Wratten Metallographic plates (Eastman-Kodak) with a Zeiss Ultraphot using apochromatic objectives and an aplanat-achromatic condenser.

Zygotes or embryos were washed six times with a vigorous jet of filtered 50% sea water, and incubated in 3 ml of 50% sea water (incubation medium) containing 3·0 μcuries of ¹⁴C-labeled amino acids (New England Nuclear). For any individual assay, 50 –100 zygotes or embryos were used and for the bulk of the experiments the exogenous ¹⁴C-amino acid was either uniformly labeled valine or lysine (spec. act. 200 mCi/mmole, New England Nuclear). Incubations were for 2 h in covered 50 ml beakers, shaking gently on a Dubnoff shaker at 25°C. At the end of this period, the embryos were vigorously washed six times with 50 ml of 50% filtered sea water and quickly transferred to an ice cold Potter-Elvehjem homogenizer to which 5·0 ml of cold 5% trichloroacetic acid (TCA) containing cold carrier amino acid was added. The embryos were homogenized and the chorions separated and discarded. After standing at 0°C for 15 min with frequent mixing, the homogenate was centrifuged and the supernatant discarded. This procedure was continued until there was no radioactivity (usually four washes) in the supernatant. The precipitate was resuspended in 5·0 ml of 5% TCA and heated at 90°C for 15 min. Upon cooling and centrifugation, the precipitate was further washed serially by resuspension in and centrifugation from, 5·0 ml of distilled H₂O, ethanol-ether (1:1 v/v) and ether.

The final pellet was dried. Solutions of approximately 1 mg/ml were made from the protein pellet in 0·1 M-NaOH. Aliquots from these solutions were counted in an automatic gas flow counter at 20% efficiency. Other aliquots were analyzed for protein concentration by the modified biuret test of Itzhaki & Gill (1964).

Fundulus embryos are endowed with a very tough chorion which rapidly rises from the egg at fertilization. Microbes presumably can be found attached to its exterior but in the living zygote are prevented from contact with the blastomeres. Whatever microbiota are present external to the chorion are removed prior to the assay by the addition of TCA and homogenization in this medium. The chorions are split into large pieces and are removed from the homogenate by a very mild initial centrifugation. Consequently the probability that exogenous microbial incorporation of amino acids into protein would complicate the results is minimal.

A variety of exogenous amino acids are readily but differentially incorporated into protein in normal Fundulus zygotes. In all cases the rate of incorporation during cleavage is low but increases dramatically at the onset of gastrulation (stage 10 –11). Phenylalanine is an exception requiring further study. Pertinent data are presented in Table 1. In the present experiments, either lysine or valine was used as the radioactively labeled precursor supplied in the incubation medium.

Preliminary studies had indicated that pactamycin was an excellent reversible inhibitor of protein synthesis in Fundulus embryos. In the present work, concentrations of 20 μg/mI were used in all cases since this gave maximum inhibition and reasonable survivorship as determined previously by dose response studies in this system. The results are recorded in Table 2. Incorporation of the labeled precursor was inhibited at all stages. Permanent defects of morphogenesis are brought about by this inhibitor, as will be presently described. Using both of these criteria we conclude that the drug acts intracellularly and directly on protein synthesis, in a manner similar to its action as reported in other systems (for review see Cohen, Herner & Goldberg, 1969).

Embryos of a wide spectrum of stages were tested and because all results were similar we report here representative data of embryos of stage 22 (midway in development, eyes unpigmented). These were given a 2 h pulse of pactamycin (20 μg/ml). At the end of the pulse time, the embryos were washed three times with 50 ml of 50% sea water and 15 min later this washing procedure was repeated. At different times after removal of pactamycin the embryos were tested for their ability to incorporate [¹⁴C]valine into protein. The results are in Table 3. ft appears that the normal rate of protein synthesis is restored within 4 h after pactamycin is removed.

The incorporation of [¹⁴C]uridine into Fundulus RNA was measured in the presence of the inhibitor. Embryos of stage 26 –28 were typical of a wide variety of stages tested. The results are shown in Table 4 which also contains accompanying data for the effect of pactamycin on protein synthesis at this particular stage. It is clear that during the experimental period, RNA synthesis in Fundulus embryos is not significantly affected by pactamycin.

When zygotes or early embryos are maintained in pactamycin continuously, mortality is 100% within 2–4 days. However, depending upon the time after fertilization when the inhibitor is added, certain significant observations can be made. Generally, there is a delay in cleavage time compared to the control, irregularity of blastomere size and irregularity of blastomere distribution over the yolk. With an early onset of the inhibitor treatment (within 1 min of fertilization: (0+1)), no cleavage occurs although the blastodisc is raised in an apparently normal manner. With a delay of onset of treatment varying from (0+ 1) to (0 + 2) initial cleavages occur although they are greatly delayed. Thus at (0+1) 15% of the zygotes underwent the first cleavage. This was increased to 40% at (0+10). Delay of onset of treatment until (0 + 20) led to the accomplishment of three cleavages in 25% of zygotes so treated.

Initiation of inhibitor treatment at any later time (i.e. beyond 0 + 20) during cleavage permitted two or three more cleavages but the resultant ‘morulae’ or ‘blastulae’ were very abnormal. Common attributes of all were irregularity in cell distribution and inequality in cell size.

It should be stressed that continuous incubation in pactamycin is ultimately lethal although increase in ‘free time’ prior to onset of treatment tends to permit a small number of cleavage cycles and a delay in mortality. Since the pactamycin treatment at this time (stage 2) leads to 74% inhibition of protein synthesis, rising to 99% at stage 6–7 (16–32 cells), we conclude that a certain amount of immediately preceding, yet ongoing protein synthesis is required for cleavage processes (Table 2).

Morphogenetic defects associated with variation of pactamycin pulse initiation time and length were investigated. The pulse lengths selected were 15 min, 30 min, 60 min, 120 min and one day (1440 min). Incubations in pactamycin were begun at (i.e. 0+15 sec), 0 +1, 0 + 2, 0 + 3,0 + 4, 0 + 5,0 +10,0 + 30 and 0 + 60 min for each pulse length. Approximately 30 embryos were used in each case and the fluid volume in each dish was 15 ml. Daily observations were made and survivors were fixed for histological examination at 0 + 40 days, the onset of hatching of the controls.

The matrix of this large experiment can be followed by reference to Table 5 which will also be useful in the section on histological analysis.

During the experiment, living embryos were examined for success or failure and degree of normality of cleavage, of formation of the high blastula, of gastrulation and epiboly, of body axis formation and of the degree of normality expressed in the formation of the head and head structures. These observations were correlated with histological analyses which will be reported in the following section.

In general the developmental failures followed the serial order and time dependence previously reported by us for precisely timed initiations of inhibition of RNA synthesis by actinomycin D (Wilde & Crawford, 1966). The type of failure of morphogenesis was primarily dependent upon time of pulse initiation with pactamycin while the severity of the defect was in part dependent upon the length of the pulse period.

1. Pulse initiation at 0 + 0 or led to complete failure of cleavage (Fig. 1) in most cases, while a few showed accumulation of a ‘pile’ of cells.

2. Pulse initiation at 0+1 led to cleavage and the accumulation of a ‘pile’ of cells resembling somewhat the blastula (Fig. 2). No further morphogenesis ensued. Some embryos suffering pulse durations of 15 and 30 min survived to hatching of the controls while those exposed to pulse duration of 60, 120 and 1440 min succumbed.

3. Pulse initiation at 0 + 2 led to results similar in all observable detail to those at 0+ 1.

4. Pulse initiation at 0 + 3, with a duration of 15 min led to very slightly improved morphogenesis as expressed by irregular cellular masses which were somewhat longer than wide (Fig. 3). However, it is questionable whether this result should be taken as evidence of body axis formation in view of the histological chaos to be described. Longer pulse durations were lethal with a gradual mortality prior to hatching of the controls.

5. Zygotes subjected to pulse initiation at 0 + 4 gave rise to anomalous embryos, some of which were, however, axiated, but barely so (Fig. 4). Survivors in this class were from a pulse duration of 15 min. Longer pulses led to an increase in mortality prior to hatching of the controls.

6. Embryos subjected to pulse initiation at 0 + 5 were axiate but anencephalic, when the pulse was of 15 min duration (Fig. 5). Longer pulse times (e.g. 60 min) led to survivors which exhibited non-axiate cell masses similar to 0 +1 and 0 + 2 (Fig. 2).

7. Survivors at pulse initiation time 0+10 were axiate-anencephalic embryos (Fig. 6) when duration was 15 min. Durations of 30 and 60 min led to survivors which were non-axiate, chaotic, cellular masses (Fig. 7).

8. Pulse initiation at 0 + 30 with duration times of 15 and 30 min led to embryos whose development was normal (Fig. 8). Pulse duration of 60 min led to abnormal embryos of the axiate-anencephalic type (Fig. 9).

9. The final pulse initiation time studied (0 + 60), gave rise to embryos of normal development at a pulse duration of 15 min (Fig. 10). Pulse durations of 30 and 60 min led to the development of microcéphalie embryos (Fig. 11). The severity of microcephaly was greater with the longer pulse (Control = Fig. 12).

It should be emphasized that all of the pulse initiation times had their onset prior to the normal time of first cleavage (90–110 min).

The increase in severity with increase in pulse duration is compatible with the concept that the synthesis of macromolecules important to morphogenesis is bimodal. That is, initiation time delimits the signal type in an ‘informative’ and labile mode while the increased pulse time covers and overwhelms a ‘confirming’ or permanent mode. This hypothesis will be developed further in subsequent publications.

Analysis of longitudinal sections of embryos of this series corroborates and extends the in vivo studies reported above. The serial order of time dependent failures of morphogenesis is exemplified by Fig. 13, 0 + 0 (120 min), cleavage failure; Figs. 14 and 15 0+1 (15 min) and 0 + 2 (15 min), amorphous cell mass, gastrulation failure; Figs. 16 and 17, 0 + 3 (15 min) faulty axis formation; Fig. 18, 0 + 4 (15 min) barely axiate; Fig. 19, 0 + 5 (15 min) axiate-anencephalic embryos; Fig. 20, 0+10 (15 min) improved axiation but retained anencephaly; Fig. 21, 0 + 30 (15 min), 0 + 30 (30 min), 0 + 60 (15 min), control, normal morphogenesis.

1. Cleavage failure (Figs. 1 and 13; exp. series 0+0, 0+1/4). Examination of serial sections of these specimens indicates that they consist of a single mass of material with many vacuoles. There is no evidence of cleavage furrows or membranes. Interior to the large vacuole layer the mass is more granular. These granules are sparse until a still more central ‘nuclear area’ is reached. This consists of a somewhat more dense ring of larger granules. However, no nuclear membrane can be observed. The picture is compatible with that of a dead or dying single cell.

2. Amorphous cell mass-gastrulation failure (Figs. 2, 14 and 15; exp. series 0 + 1 and 0 + 2). Fig. 14 is a median section through a mass of cells resting as a ‘high blastula’ upon the yolk surface. The arrangement of cells shows no pattern nor can any tissue structure be seen. Peripherally there are scattered cells with greater amounts of cytoplasm and larger nuclei which are, perhaps, representative of periblast. There are also a few irregular internal spaces which, however, do not show cell profiles conforming to the space. There is, therefore, no evidence of any organ or tissue differentiation although there has obviously been successful cleavage to form the multicellular mass. Fig. 15 represents a median section through a mass which has sunk into the yolk. The cellular pattern is quite similar to that seen in Fig. 14. No tissue or organ patterns can be seen. There are scattered large cells probably referable to periblast. Within the chaotic pattern however, four melanocytes (M) have differentiated. There is no evidence of morphogenesis. In this group (0 + 2) there is one specimen which appears to be somewhat longer than it is wide. However, in detail it is identical to Fig. 15 in lacking any semblance of morphogenesis, tissue or organ formation. Its dimensions are, therefore, interpreted as being the casual effects of random growth and not of the initiation of axiation.

3. Attempts at abortive axis formation (Figs. 3, 16 and 17; exp. series 0 + 3). Longitudinal sections near the mid-line of these specimens show certain structural and tissue-like differentiations within the prevailing histological chaos. The first to appear is an irregular cellular rod, the cells of which are often vacuolated in a manner reminiscent of early notochordal cells (Fig. 16). The rod is quite irregular with constrictions, bends and twists. It is bounded by a well defined ‘membrane’ again reminiscent of notochord sheath. It lies in the long axis of the cellular mass. Nearby a second organoid structure is present (Fig. 17). This represents a rather regular lumen surrounded by somewhat columnar cells of an epithelial appearance. By comparison with control specimens, this structure appears to represent poorly differentiated alimentary tract. There is no evidence of somites or nervous tissue. Aside from occasional pigment cells, no other identifiable cell types or tissues are to be seen.

4. Embryos definitively but irregularly and barely axiate (Figs. 4 and 18; exp. series 0 + 4). Sections from embryos of this series show definite axial structures but they remain irregular and their pattern is obscure. The following tissues can be definitely identified. Notochord (N) and sheath with vacuolated cells, alimentary tract with columnar, epithelial cells, well matured and polarized, blood vessels and erythrocytes. There are many other organoid structures of a glandular nature and perhaps some very poorly organized nervous tissue. In longitudinal sections one still cannot identify anterior and posterior. No somites or muscle blocks are to be found. Pigment cells are numerous. The cellular material between the lobes and identifiable tissues remains irregular and randomly disposed.

5. Embryos axiate but anencephalic (Figs. 5 and 19; exp. series 0 + 5). In examining longitudinal sections such as shown in Fig. 19 it is tempting to identify the ‘head’ of the embryo at the bottom and the ‘tail’ at the top. Caution is required however. Such embryos have a well developed notochord and sheath (A), an alimentary tract (A) with columnar epithelium and blood vessels containing erythrocytes. Nervous tissue is not easily identifiable but, perhaps, is represented by dense aggregations such as NV. Branchial arches cannot be made out nor can regularly arranged somitic muscles. Notochord is present throughout the bulk of the section and is more highly differentiated (vacuolated) toward the top. Very tentatively we identify the top of the figure as anterior. There is no brain and no skull. The embryo is completely anencephalic.

6. Embryos axiate with improved organogenesis but relatively anencephalic (Figs. 6 and 20; exp. series 0+10). The longitudinal section represented by Fig. 20 illustrates a much less abnormal embryo. A well differentiated notochord (N) is present which in other sections can be demonstrated to run from the large vesicle (V) throughout the embryo. The vesicle area with poorly differentiated nervous tissue to the left (NV) and a similar mass to the right constitutes the anterior end. No skeletal elements of the skull are present. There is a well differentiated alimentary tract (A) with polarized columnar epithelium. The ventral diverticulum of glandular cells (L) may represent the liver while the dorsal small glandular mass (P) may represent pancreas. Posteriorly and dorsally, poorly formed muscles of somitic origin (S) are present. The vesicle (V) is abnormal and cannot be identified. Posteriorly, just above the notochord, ependyma of the ventral aspect of the spinal cord appears (E). There are numerous pigment cells and fragile blood vessels of various sizes containing erythrocytes. In the ‘head’ region are numerous clusters of large cells with abundant cytoplasm and large nuclei which resemble periblast. If they indeed represent periblast, such invasion of the embryo is a surprising phenomenon. However, further studies on abnormal periblast are required.

7. Embryos with normal morphogenesis (Figs. 8, 10, 12 and 21; exp. series 0 + 30, 0 + 60 and controls). Fig. 21 represents a longitudinal section of a control embryo fixed just prior to hatching and is representative of specimens of 0 + 30 (15 min), 0 + 30 (30 min) and 0 + 60 (15 min). The organs and tissues which result from conditions of normal morphogenesis may be usefully compared with the previous figures in assessing the sériation of developmental defects in the previously described experimental series.

By reference to Table 5 it will be recalled that increase in pulse duration of pactamycin tends to shift the resultant anomalous embryos toward the beginning of the serial order of defects with a similarity to earlier pulse initiation time. Thus, 0 + 5 (60 min) is like the 0 + 2 series; 0 +10 (30 min) and 0 +10 (60 min) are similar to the 0 + 2 series. The class 0 + 30 (60 min) is similar to the 0+10 series, while 0 + 60 (30 min) and 0 + 60 (60 min) are microcéphalie rather than anencephalic.

Pactamycin has previously been demonstrated to be an inhibitor of protein synthesis in mammalian (Colombo, Felicetti & Baglioni, 1966 and Felicetti, Colombo & Baglioni, 1966) and bacterial (Cohen, Herner & Goldberg, 1969) systems. The site of action has been shown to be at the binding of aminoacyl-transfer RNA to ribosomal subunits (Cohen, Goldberg & Herner, 1969). The inhibitor has now been shown to be a useful and effective agent for studies of the control of protein synthesis in embryos of Fundulus heteroclitus. It inhibits protein synthesis at all stages of development at the level of 75% or greater as expressed by a decrease in incorporation of radioactive amino acids into trichloroacetic acid-insoluble protein. Furthermore, its effect on protein synthesis is reversible, which allows studies of the effect of inhibitions at particular periods of synthesis in relation to morphogenesis. Since a pulse period of two hours or less of pactamycin does not effect RNA synthesis, a clear distinction between effects referable to translation rather than transcription can be made.

Failures of morphogenesis due to precisely timed pulses of pactamycin fall in a serial order whose regular sequence is predictable and closely related to a similar serial order elucidated by inhibition of RNA synthesis with actinomycin D.

It is this predictable sequence of morphogenetic failures due to precisely timed additions of pactamycin to the incubation medium of the embryos that we wish to stress at this time in correlation with our previous data on the morphogenetic defects related to RNA synthesis inhibition (Wilde & Crawford, 1966, 1968).

Development to the high blastula (stage 9–10) in Fundulus is independent of contemporary RNA synthesis on zygotic templates although in the zygote there is initiated synthesis of RNA meaningful for the post-blastula period beginning within minutes following fertilization. The data for Fundulus are consistent with the findings in a broad range of embryonic systems (Davidson, 1968). The conclusion generally held is that genomic transcription of informational molecules required for cleavage occurs prior to fertilization and thus on maternal templates. This would include active templates for proteins concerned with cleavage spindles, chromosome replication and increment of cell surface. This has recently been established in echinoderms (Raff, Colot, Selvig & Gross, 1972).

Inhibition of protein synthesis through administration of pactamycin immediately upon fertilization aborts the first cleavage. If the pulse initiation is delayed an effect decreasing in intensity is observed up to a pulse initiation time of 20 min following fertilization. In the latter circumstances, sparing the first 10 to 20 min permits some zygotes to undergo two or three cleavages. Therefore during this time anticipatory syntheses of cleavage associated proteins are taking place presumably upon maternal templates. Anticipatory syntheses appear to be required for cleavages two or three cycles in the future. Therefore, an ongoing protein synthesis is required in Fundulus for normal development in the cleavage period. Actinomycin D has no such effect. In the presence of this drug, cleavage to the high blastula is normal in form and in timing.

We have previously reported (Wilde & Crawford, 1966) a serial order of defects in morphogenesis of high predictability dependent strictly upon the time of initiation of actinomycin D inhibition of RNA synthesis. The data presented in this paper demonstrate the similarity of morphogenetic defects beyond the high blastula conferred by pulses of pactamycin and actinomycin D. These correlations are shown in Table 6. The morphogenetic defects are expressed morphologically 40 or more hours after the cessation of the inhibitory pulse which led to the defect. While protein synthesis as a general phenomenon is much more rapidly restored following pulse termination, the morphogenetic defects are permanent.

We are thus drawn to the conclusion that specific protein syntheses, required for normal morphogenesis beyond the blastula, are initiated prior to the second minute following fertilization and thus upon zygotic templates (or alternatively on certain maternal templates stimulated to function by the act of fertilization). Furthermore, in dealing with the minimal pulse times in these experiments (15 min), the effect on any particular synthesis must be complete by the termination of the pulse or shortly thereafter since protein synthesis as measured by standard methods will be resumed.

Protein synthesis is low during early zygotic periods as reflected in the data. Much of this synthesis must necessarily be concerned with spindle protein and histone synthesis, to name the most obvious. It would appear therefore that the concurrent protein synthesis required for post-blastula morphogenesis takes place at a very low level. If informational RNA is presumed to be attached to appropriate ribosomal configurations and the presence of potentially functional ribosomal units is assumed, upon relief of the inhibition, why is subsequent morphogenesis disturbed? Perhaps pactamycin firmly enters the functional unit and renders it inactive on a relatively permanent basis. Consequently, since a serial order of RNA synthesis essential to post-blastula morphogenesis has been demonstrated to be time dependent, the essential first morphogenetically meaningful proteins are never given the morphogenetically correct, temporal opportunity to be synthesized, their templates having been passed by in time and, perhaps, degraded. Relief of the inhibition may allow for all the essential protein syntheses, certain of which are now out of proper sequence and relationship to the ongoing development. Under these conditions they cannot play their normal morphogenetic role.

It follows, in the major axial systems studied here in Fundulus, that early syntheses are successionally dependent upon antecedent ones. Only after the fifth minute is there any expression of reasonably normal morphogenesis of one part (viz. gut) while another part (the brain) remains utterly defective. Morphogenetically meaningful macromolecular syntheses appear to fall into a ramifying scheme which is initially one tract prior to branching. Organogenesis of a particular structure would be inhibited by failure of antecedent syntheses at a branch point. Under such circumstances, after the fifth minute, inhibition along one ramus would lead to the failure of morphogenesis of subsequent dependent development while along an uninhibited branch morphogenesis would continue more or less normally. Such a scheme however, must be broad enough conceptually to include connecting links at varying levels of organogenesis.

These data are consistent with the conclusion that morphogenesis in Fundulus is under genomic control and that the control functions through the well established mechanisms of transcription and translation. They further indicate that the macromolecular chemistry of morphogenesis begins within seconds of fertilization at least as expressed in the primary and major morphogenetic activities of the zygote. The expression of genomic control as reflected in the time dependent serial order of developmental failures here analyzed indicates that morphogenetic chemistry proceeds rapidly and stepwise. Failure at early steps causes abnormal orientation of any further development.

Yet it is of intense interest that development does in fact continue, carrying the immutable defects, notochord without nervous system, nervous system without somites, etc. The morphogenetic program entered into at fertilization behaves as though all cells and their progeny were committed and cognizant of time flow and position; only those cells which were the open targets of the inhibitor during, and only during, the pulse failed in morphogenetic commitment and behaviour.

It has long been considered that morphogenesis is initiated at gastrulation via inductive processes. We wish to emphasize that the classic phenomenology is, in Fundulus, preceded by 40 h of essential macromolecular synthesis without which morphogenesis is aborted. Indeed the most important initiating and controlling events appear to be confined to the period immediately following fertilization and well within the first cleavage period while the zygote is a single cell.

The specific protein syntheses characteristic of cellular differentiation are expressed in many of the terata discussed here. These are also presumably under genomic control. However, the transcription and translation processes for cellular differentiation appear to be initiated later in embryogenesis beyond the period of morphogenetic determination. The relationship of primary morphogenetic processes to the chemistry of cellular differentiation remains to be explored.

The authors wish to thank Mrs Diane Zucker and Mrs Michele Koppelman for their excellent technical assistance. Gift of pactamycin from the Upjohn Co. was very much appreciated.

The work reported here was supported by N.S.F. Grant GB-6766 and an NSF Grant to the Mount Desert Island Biological Laboratory GB-28139.

FIGURES 1–12

All figures are of living eggs and embryos. Initial magnification 50 x . Bar indicates 1 mm. All photographs taken at 30 days after fertilization.