Although much is known about the specification and determination of the two primary axes (animal/vegetal and dorsoventral or oral/aboral) in a number of embryos, little is understood about bilaterality. In the sea urchin, left/right asymmetry is crucial to normal development as the echinus or adult rudiment is positioned on the left side of the larva. We examined the establishment of bilateral asymmetry in Lytechnis variegatus embryos by determining the relationship of the first cleavage planes to the left/right axis. Embryos were bisected at different times to determine when the bilateral axis is committed. These lineage tracing and cell separation experiments demonstrated that the first cleavage plane divides the embryo into left and right halves, although this is conditional until after late blastula stage. The relationship between the specification of the dorsoventral axis and the bilateral axis was examined experimentally. In other species when the dorsal and ventral halves of early echinoderm embryos (preblastula) are separated, the dorsal half often reverses (180°) its dorsoventral axis. We asked whether those larvae with an inverted dorsoventral axis would shift the position of the echinus rudiment from the original left side to the new left side. If so, it would demonstrate that the larval asymmetry is dependent upon specification of the dorsoventral axis. Using a combination of lineage tracing and cell separation techniques, we show that the left/right asymmetry is specified with respect to the dorsoventral axis.

Bilateral animals have three axes: anteroposterior, dorsoventral and left/right. Although a considerable amount is known of the establishment of the anteroposterior and dorsoventral axes, very little is understood about the developmental mechanisms distinguishing left from right. One outstanding question concerning the specification and determination of the bilateral axis is: when and how is it established with respect to the other axes? In other words, are the spatial relationships between the three axes dependent upon one another so that, if the orientation of one axis changes, the other axes are also affected, or can the specification of any of the three axes be uncoupled from the others? There are few systems in which this can be experimentally tested. The echinoderm embryo, however, can be used to examine the interdependency of axial specification since the dorsoventral axis can be artificially reversed at an early stage of development. We used this experimental system to examine the relationship between the specification of the dorsoventral axis and left/right asymmetry.

Unlike vertebrates, which have multiple asymmetries on both sides of their bilateral bodies, most echinoderm larvae have only one readily observed asymmetrical structure; a group of coelomic cells on the left side of the gut develops into the echinus rudiment which produces most of the future adult. The first sign of this asymmetry is the formation of the hydropore, a portion of the adult water vascular system (Fig. 1A). Eventually, when the echinus rudiment is fully developed (Fig. 1B), the larva metamorphoses into an adult. Although the specification of the left/right asymmetry and, correspondingly, the echinus rudiment have not been examined in any species of sea urchins, a few starfish have been looked at with respect to that issue. When starfish embryos were separated into left and right halves at an early stage (between blastula and gastrula), the larvae developed differentially. The left halves, which had higher viability, formed a hydropore on the left side, while the right halves were less viable and had a hydropore on the left or right side, or on both sides. From these data, it was concluded that, in starfish embryos, the position of the hydropore is fixed and stable in the left half, but not in the right half (Runnström, 1920; Hörstadius, 1928, 1973).

Fig. 1.

L. variegatus larvae. (A) 7-day-old pluteus larvae, from the dorsal (aboral) view. The hydropore (arrow) is on the left side of the esophagus (e), anterior to the stomach (s). The inset is a magnification of the hydropore (arrow). (B) 24-day-old echinopluteus with well-developed echinus rudiment (r) on the left side of the stomach. The folded tube feet are visible (arrows). Bar, 13.1 μm.

Fig. 1.

L. variegatus larvae. (A) 7-day-old pluteus larvae, from the dorsal (aboral) view. The hydropore (arrow) is on the left side of the esophagus (e), anterior to the stomach (s). The inset is a magnification of the hydropore (arrow). (B) 24-day-old echinopluteus with well-developed echinus rudiment (r) on the left side of the stomach. The folded tube feet are visible (arrows). Bar, 13.1 μm.

Similar experiments investigating the determination of the dorsoventral axis in echinoderm embryos demonstrated that the echinoderm embryo is capable of reversing its dorsoventral axis. Hörstadius (Hörstadius and Wolsky, 1936; Hörstadius, 1973) discovered that, by cutting a 16-cell stage Para-centrotus lividus embryo into dorsal and ventral halves, the dorsal half reversed its dorsoventral polarity so that what was originally dorsal became ventral, and ventral became dorsal (experiment reviewed in Fig. 2A). Unfortunately, the experiment did not reveal the effect of this reversal upon the orientation of the hydropore and echinus rudiment. If the dorsal halves had survived longer, there would have been two possible outcomes with respect to the reversal of the left-right asymmetry (reviewed in Fig. 2B): (1) the echinus rudiment could have been on the left side, indicating that its location was dependent upon the (re)specification of the dorsoventral axis, or (2) the echinus rudiment could have been on the right side, indicating that the location of the specified asymmetrical structure had not switched when the dorsoventral axis reversed (Fig. 2B). Reversal of the dorsoventral axis in dorsal half embryos was also observed with Heliocidaris erythrogramma, a direct developing sea urchin, but the effect on left/right asymmetries was not noted (Raff, 1992; Henry and Raff, 1993). We used this same basic experimental manipulation, of cutting embryos into dorsal and ventral halves, in order to learn whether bilaterality is established with respect to the dorsoventral axis, or whether bilateral pattern information is specified by a mechanism distinct from that of the dorsoventral axis.

Fig. 2.

Description of a manipulation experiment and the predicted results. (A) Hörstadius (1973) cut P. lividus embryos into dorsal and ventral halves at the 16-cell stage, then marked the cut edge with Nile Blue. The position of the dye on the dorsal side of both halves indicated that the dorsal half had reversed its dorsoventral axis. (B) The two possible outcomes of the experiment with respect to the position of the echinus rudiment in the dorsal halves that had reversed their dorsoventral axis. If the polarity of the left/right asymmetry reversed in those particular dorsal halves, the echinus rudiment would be located on the left side. If the left/right asymmetry did not invert its original polarity in the dorsal halves then the echinus rudiment would be on the new right side. The goal of the studies reported in this paper will resolve this issue.

Fig. 2.

Description of a manipulation experiment and the predicted results. (A) Hörstadius (1973) cut P. lividus embryos into dorsal and ventral halves at the 16-cell stage, then marked the cut edge with Nile Blue. The position of the dye on the dorsal side of both halves indicated that the dorsal half had reversed its dorsoventral axis. (B) The two possible outcomes of the experiment with respect to the position of the echinus rudiment in the dorsal halves that had reversed their dorsoventral axis. If the polarity of the left/right asymmetry reversed in those particular dorsal halves, the echinus rudiment would be located on the left side. If the left/right asymmetry did not invert its original polarity in the dorsal halves then the echinus rudiment would be on the new right side. The goal of the studies reported in this paper will resolve this issue.

We address several specific questions in an attempt to answer when and how bilateral asymmetry is established in the sea urchin, Lytechinus variegatus. (1) What is the relationship of the left/right axis to the first cleavage planes? (2) When is the difference between the left and right sides, or the position of the echinus rudiment, committed? (3) Is the left/right asymmetry dependent upon the dorsoventral axis such that if the dorsoventral axis is reversed, asymmetric bilaterality will also be reversed?

Gamete and embryo preparation

Lytechinus variegatus adults were acquired from North Carolina and Florida. Gametes were obtained by intracoelomic injection of 0.5 M KCl. The eggs were fertilized with dilute sperm in artificial sea water (ASW) containing 8 mM p- aminobenzoic acid (Sigma, Inc). Before first cleavage, the embryos were washed, poured through a 0.75 μm nylon mesh to remove the fertilization envelope, then washed again. Embryos were maintained in glass dishes at 23 °C or in a 15 °C water bath until the 2-cell stage, when they were prepared for micromanipulation.

Lineage tracing techniques

To determine the position of the first cleavage plane relative to the larval axis, one cell of a 2- cell stage embryo was labeled with either Nile Blue or 1,1′-dihexadecyl-3,3,3′3′-tetramethylin- docarbocyanine perchlorate {DiI C16(3)} (D-384, Molecular Probes, Eugene, Oregon), a fluorescent lipophilic dye.

DiI C16(3)

In order to label cells for several days, DiI C16(3) was injected into the cell. It was first diluted to a 10% stock solution in 100% ethanol. From this, a 0.3% DiI solution was made in soybean oil (Wesson cooking oil). Polyethylene tubing (76 mm inner diameter) attached to a 0.2 ml Gilmont syringe was backfilled filled with approximately 5 μl of mercury and then 50 μl of the 0.3% DiI solution. A siliconized glass needle (0.85 mm outer diameter, 0.55 mm inner diameter, Drummond Scientific Company, Broomall, PA), whose tip had been pulled to a tip diameter of approximately 2 μm, was attached to the end of the tubing and the DiI solution was front-filled into the needle. The needle was mounted onto a Narishege micromanipulator and positioned adjacent to a Leitz Dialux 12 microscope. The DiI solution, in the form of a dye/oil droplet, was microinjected into one cell of the 2-cell stage embryos, which were held in a Kiehart chamber (Kiehart, 1982). DiI rapidly diffused throughout the cell, but the oil droplet and some DiI were retained throughout development. For the lineage tracing experiment, one cell at the 2-cell stage was injected with one DiI droplet. Another experiment, designed to test for axis reversal, required that two DiI droplets be injected into one cell at the 2-cell stage. Under this condition, the DiI droplets were positioned so that the second cleavage would partition one droplet into each of the two daughter cells.

Following microinjection, the embryos were removed from the Kiehart chamber and washed with ASW. They were cultured in the dark until prism stage (24 hours old) or pluteus stage (48 hours old), when they were monitored with a fluorescent microscope. The position of the dye and the droplet were noted.

Nile blue

For short-term cell surface labeling, a solution of 2% low melting point agarose and 0.5% Nile Blue was backfilled into a pulled capillary tube (tip diameter approximately 30 μm). After the agarose had cooled, the tip of the capillary tube was applied, freehand, to the cell membrane of a 2-cell embryo. A dissecting microscope facilitated the process. Only those embryos that were half-labeled and continued to develop normally were set aside for cell separation experiments.

Cell separation

Embryos were cut in half at the 2-cell, 4-cell or blastula stages. Individual embryos were placed in Ca2+/Mg2+-free ASW for one to three minutes and a glass needle was gently forced down between the cells that were to be separated. Once the two halves were apart, they were briefly flushed with ASW and then moved to a 2% agarose-coated dish which contained Millipore-filtered ASW. The half embryos from a single separation experiment were raised as a pair in a dish. After 3–24 hours, each pair of half embryos was transferred to an agarose-free dish. The embryos were examined every other day with a compound microscope for the position of the hydropore (on about day 7) and the echinus rudiment (on about day 20), as well as overall larval characteristics.

Embryos were cut in half along the first or second cleavage plane. To test for the fixation of the left/right axis, embryos were cut into left and right halves (along the first cleavage plane) at the 2-cell stage, early blastula (4–5 hours postfertilization) stage and the late blastula, premesenchyme blastula stage (6.5–7.5 hours postfertilization). Embryos were also separated into dorsal and ventral halves (along the second cleavage plane) at the 4-cell stage, after having been marked with Nile Blue or DiI at the 2-cell stage.

Maintenance of embryos and larvae

All embryos were raised at 23 °C in 35 mm Falcon plastic dishes filled with Millipore-filtered ASW. Every other day, larvae were transferred to clean dishes filled with Millipore-filtered ASW. The larvae were fed Dunaliella tertiolecta (8,000/larva) every other day after the fourth day of development.

Axis nomenclature

Discussions related to defining the positions of the dorsal, ventral, left and right sides of an organism can be confusing. This stems from the fact that, although the animal/vegetal and dorsoventral axes are defined by their specific properties, the left and right sides are characterized not by traits unique to them, but are defined geometrically by their position relative to the other two axes. In sea urchins, the animal/vegetal axis is fixed during oogenesis and, thus, can not be manipulated. In contrast, the dorsoventral axis can be reversed experimentally. By definition, the left and right sides would also switch in such embryos. But this poses a problem. If the dorsoventral axis reversed but the property that specified the left/right asymmetry did not, then the organs that were once on the left side would now be on the newly defined right side (Fig. 2B). Thus, left/right characteristics would be fixed independent of the determination or commitment of the dorsoventral axis. If the specification of bilateral asymmetry was inextricably bound to the dorsoventral axis, however, the reversal of the dorsoventral axis would require that the position of the bilateral organs also reverse. It is this relationship of axis specification that we are examining.

The relationship between the left/right axis and the first two cleavage planes of L. variegatus

DiI, in an oil-based media, was injected into one cell at the 2-cell stage of 105 embryos (Fig. 3A). The location of the dye relative to the larval axis was examined at the prism or pluteus stages, when the dorsal, ventral, left, right, anterior and posterior regions were distinguishable. Three patterns were observed; labeling was on the right side, the left side, or all over. In 75 (75/105) larvae, the dye was distributed bilaterally, with 30 (30/75) marked on the right side and 45 (45/75) marked on the left (Table 1). In most of the embryos, the division between labeled and unlabeled tissue equally bisected the larva and was parallel to the plane of bilaterality (Fig. 3B-D). The angle of the division was sometimes tilted 5–10 ° to the left or the right of the bilateral plane. In the 30 embryos where the dye was spread uniformly, it was likely that the original dye injections were made prior to completion of the first cleavage.

Table 1.

Results from the DiI injection experiment

Results from the DiI injection experiment
Results from the DiI injection experiment
Fig. 3.

Lineage tracing of 2-cell L. variegatus. The left-hand column is bright field and the right-hand column is fluorescence. (A) 2-cell stage embryo after injection of DiI/oil droplet (arrow). (B) Prism stage embryo (ventral or oral view) with fluorescent label on its right side. The anus (a) is in focus. (C-D) Two plutei (ventral or oral view) with fluorescent label on their right side. The DiI/oil droplet (arrow) is visible in both the bright-field and fluorescent light. The fluorescent images (C and D) show several labeled mesenchyme cells on the unlabeled side of the larvae (arrowhead). Thus, pigment or skeletogenic cells migrate all over the larvae, regardless of which side they originated from. Bar, 20.0 μm.

Fig. 3.

Lineage tracing of 2-cell L. variegatus. The left-hand column is bright field and the right-hand column is fluorescence. (A) 2-cell stage embryo after injection of DiI/oil droplet (arrow). (B) Prism stage embryo (ventral or oral view) with fluorescent label on its right side. The anus (a) is in focus. (C-D) Two plutei (ventral or oral view) with fluorescent label on their right side. The DiI/oil droplet (arrow) is visible in both the bright-field and fluorescent light. The fluorescent images (C and D) show several labeled mesenchyme cells on the unlabeled side of the larvae (arrowhead). Thus, pigment or skeletogenic cells migrate all over the larvae, regardless of which side they originated from. Bar, 20.0 μm.

From these data, it was concluded that the first cleavage plane divided the embryo into left and right halves. The second cleavage plane, therefore, bisected the future dorsoventral axis. This axial information was used in the following cell separation experiments.

The timing of the commitment of left/right asymmetry

At various stages, embryos were cut into left and right halves along the plane of the first cleavage. To orient embryos that were bisected after the 2-cell stage, one cell at the 2-cell stage was marked with Nile Blue. Later, the marked embryos were cut in half along the dye line. Left/right asymmetry was assessed by examining the overall symmetry of the larva and, when the larva was older, the position of the hydropore and the echinus rudiment was noted. Pairs of larvae were followed.

Each member of the twelve surviving pairs derived from a bisected 2-cell stage embryo had a left hydropore and was normal in all other respects (Table 2A). Six of the pairs survived long enough as a pair to produce an echinus rudiment and this was on the left side in all cases. Only one individual of each of the other 6 pairs survived to the time of rudiment formation; all had the echinus rudiment on the left side. There were 18 pairs in which only one of the pair survived long enough to observe the position of the hydropore. Of these, all but two had a left hydropore (Table 2A). Nine of the 16 with a left hydropore survived to produce a left echinus rudiment. One single larva formed a hydropore on both the left and right sides (Table 2A) and eventually had an echinus rudiment on the right side. The other larva did not form a definitive hydropore (Table 2A) or rudiment.

Table 2.

Results from cell separation along the first cleavage plane

Results from cell separation along the first cleavage plane
Results from cell separation along the first cleavage plane

Of 17 pairs of early blastula halves that survived one week, 15 produced two larvae each of which had left hydropores. The other two pairs consisted of one normal larva with a left hydropore and one that lacked coelom development (Table 2B). One complete pair and four single larvae survived to produce left echinus rudiments. All 16 of the single half blastula produced left hydropores. Eight of those larvae survived long enough to produce left echinus rudiments. 14 pairs of late blastula half embryos and 13 single late blastula half embryos survived to the age when the hydropore could be discerned. Both larvae in nine of the pairs had a left hydropore. For two pairs (2/14), the hydropore was on the left of one half of a pair and in the right half of the other. The remaining three pairs (3/14) had one larva with a left hydropore and the other with both left and right hydropores. Of the 13 single larvae, the majority had left hydropores, but two had a right hydropore, and two had a left and right one (Table 2C). In general, when one half had a left hydropore, it would eventually produce a left echinus rudiment (Fig. 4A,B). If the other half had a left and a right hydropore, two rudiments would result (Fig. 4C,D). From these experiments, we observed that, prior to the late blastula stage, both halves of a bisected embryo place the hydropore on the left side. Beginning at the late blastula stage, bisection sometimes resulted in embryos with hydropores on the right side or with both a left and a right hydropore.

Fig. 4.

A pair of larvae, which were produced by cutting one late blastula (6.5 hours old) into a left half and a right half: (A-B) one half and (C-D) the other half. (A) One half larva at 12 days old with a left hydropore (arrow). (B) By 28 days this half larvae develops an echinus rudiment (arrows) on the left side. (C) At 12 days of age the larva is less developed than its counterpart (in Fig. 4A). The inset shows that the larva has left and right hydropores (arrow), which are joined by a continuous hydroporic canal (arrowhead). (D) The larva at 28 days old with two bilateral early echinus rudiments. The hydrocoel has formed on both sides (arrow). Bar, 21.3 μm.

Fig. 4.

A pair of larvae, which were produced by cutting one late blastula (6.5 hours old) into a left half and a right half: (A-B) one half and (C-D) the other half. (A) One half larva at 12 days old with a left hydropore (arrow). (B) By 28 days this half larvae develops an echinus rudiment (arrows) on the left side. (C) At 12 days of age the larva is less developed than its counterpart (in Fig. 4A). The inset shows that the larva has left and right hydropores (arrow), which are joined by a continuous hydroporic canal (arrowhead). (D) The larva at 28 days old with two bilateral early echinus rudiments. The hydrocoel has formed on both sides (arrow). Bar, 21.3 μm.

Dorsoventral axis reversal in the dorsal half embryos: what happens to bilaterality?

Two separate experiments tested the hypothesis that, when the dorsoventral axis is reversed, the position of bilaterally asymmetric structures is also respecified. First, we observed the location of the hydropore and echinus rudiment that developed in dorsal half L. variegatus embryos (See Fig. 2B). One cell at the 2-cell stage was marked with Nile Blue and at the 4-cell stage the embryos were cut along the second cleavage plane into dorsal and ventral halves. The Nile Blue was simply used to mark the orientation of the first cleavage plane. The pairs were followed and examined for the position of the hydropore and echinus rudiment. Since previous experiments have shown that ventral halves do not invert the dorsoventral axis (Fig. 2A), we anticipated that at least one member of a dorsoventral pair would have a hydropore on the left side. Since the other larva of the pair (the dorsal half) could have a hydropore on the left or the right side, there are three possible interpretations for each pair of dorsal and ventral larvae (outlined in Fig. 5). If the dorsal half larva has a hydropore on the left side, either no axis reversal occurred or both the dorsoventral axis and the position of the hydropore reversed 180°. If, on the other hand, the hydropore were on the right side of the dorsal half larva, then only the dorsoventral axis reversed and the position of the hydropore was not respecified. (Fig. 5).

Fig. 5.

The two outlined experiments demonstrate which axis or axes reversed when 4-cell stage embryos were cut into dorsal (D) and ventral (V) halves. The experiment described on the left required that one cell at the 2-cell stage be marked with Nile Blue so that a cut could consistently be made along the second cleavage plane at the 4-cell stage. For these half embryo pairs, the position of the hydropore was noted. Since it is recognized that the ventral half does not reverse its dorsoventral axis (Hörstadius and Wolsky, 1936; Hörstadius, 1973; Raff, 1992; Henry and Raff, 1993), the hydropore would always be on the left side of at least one half embryo, the ventral half. This embryo is described at the top of the figure. If the hydropore was on the left side of both the dorsal and ventral halves then there were two possible interpretations. (1) Neither the dorsoventral axis or the left/right asymmetry associated with the position of the hydropore reversed, or (2) both axes inverted. This was what was experimentally observed, indicated by the arrows. If the hydropore had been on the right side of one half embryo (dorsal half) of a pair, a single axis reversal occurred. This was not observed, as indicated by the stop. The other cell separation experiment, described in the right of the figure, required that one cell at the 2-cell stage, left (L) or right (R), be injected with two DiI droplets. (This example has the dye on the left side.) At the 4-cell stage the embryo was separated along the second cleavage plane so that each half, ventral and dorsal, contained one DiI droplet. The position of the DiI droplet was noted in the resulting pair of larvae. If the DiI droplet were on the same side in both halves (the left side in this example), then no axes reversed. This occurred in a small number of cases (indicated by the smaller arrow). If the DiI droplet were on different sides of each half, then either the dorsoventral axis reversed and the position of the echinus rudiment polarity of the left/right axis) did not invert, or both axes reversed. Although the DiI injection/separation data support both of these interpretations, the Nile Blue experiment described above indicated that the former situation could not occur (indicated by the stop). Therefore, we can conclude that both the dorsoventral axis and the left/right asymmetry inverted in the dorsal half embryos (indicated by the larger arrow). Information from these two experiments allowed us to conclude that when the dorsoventral axis reversed in the dorsal half embryo, the position of the echinus rudiment would also change relative to its original location; i.e. the dorsoventral axis could not reverse without affecting the polarity of the left/right axis. The plutei are shown from the dorsal (aboral) view so that left and right are in the same position relative to the cleavage stage embryos from which they were derived.

Fig. 5.

The two outlined experiments demonstrate which axis or axes reversed when 4-cell stage embryos were cut into dorsal (D) and ventral (V) halves. The experiment described on the left required that one cell at the 2-cell stage be marked with Nile Blue so that a cut could consistently be made along the second cleavage plane at the 4-cell stage. For these half embryo pairs, the position of the hydropore was noted. Since it is recognized that the ventral half does not reverse its dorsoventral axis (Hörstadius and Wolsky, 1936; Hörstadius, 1973; Raff, 1992; Henry and Raff, 1993), the hydropore would always be on the left side of at least one half embryo, the ventral half. This embryo is described at the top of the figure. If the hydropore was on the left side of both the dorsal and ventral halves then there were two possible interpretations. (1) Neither the dorsoventral axis or the left/right asymmetry associated with the position of the hydropore reversed, or (2) both axes inverted. This was what was experimentally observed, indicated by the arrows. If the hydropore had been on the right side of one half embryo (dorsal half) of a pair, a single axis reversal occurred. This was not observed, as indicated by the stop. The other cell separation experiment, described in the right of the figure, required that one cell at the 2-cell stage, left (L) or right (R), be injected with two DiI droplets. (This example has the dye on the left side.) At the 4-cell stage the embryo was separated along the second cleavage plane so that each half, ventral and dorsal, contained one DiI droplet. The position of the DiI droplet was noted in the resulting pair of larvae. If the DiI droplet were on the same side in both halves (the left side in this example), then no axes reversed. This occurred in a small number of cases (indicated by the smaller arrow). If the DiI droplet were on different sides of each half, then either the dorsoventral axis reversed and the position of the echinus rudiment polarity of the left/right axis) did not invert, or both axes reversed. Although the DiI injection/separation data support both of these interpretations, the Nile Blue experiment described above indicated that the former situation could not occur (indicated by the stop). Therefore, we can conclude that both the dorsoventral axis and the left/right asymmetry inverted in the dorsal half embryos (indicated by the larger arrow). Information from these two experiments allowed us to conclude that when the dorsoventral axis reversed in the dorsal half embryo, the position of the echinus rudiment would also change relative to its original location; i.e. the dorsoventral axis could not reverse without affecting the polarity of the left/right axis. The plutei are shown from the dorsal (aboral) view so that left and right are in the same position relative to the cleavage stage embryos from which they were derived.

Each member of the 10 pairs derived from a 4-cell stage embryo had a left hydropore and was normal (Table 3). Of those that survived longer, the echinus rudiment was also on the left side. Of the 5 single embryos that came from the 4-cell stage separation, i.e. the other half of the pair did not survive, all had a left hydropore (Table 3). These results indicate that the dorsoventral axis did not reverse by itself (Fig. 5). If we assume that the dorsoventral axis reversed in the dorsal half, as it does in other species, then left/right asymmetry also reversed. However, the same results could occur if neither axis reversed (Fig. 5). Therefore, in order to know whether or not the dorsoventral axis did reverse, the following experiment was done.

Table 3.

Results from cell separation along the second cleavage plane

Results from cell separation along the second cleavage plane
Results from cell separation along the second cleavage plane

The second experiment directly addressed the question: does the left/right axis reverse when the dorsoventral axis reverses? Two small DiI/Wesson Oil droplets were injected into one cell at the 2-cell stage. At the 4-cell stage, only those embryos with one DiI droplet in two adjacent cells were cut into dorsal and ventral halves, along the second cleavage plane. There are only two possible outcomes to this experiment since the previous experiment showed that left/right asymmetry does not reverse unless the dorsoventral axis does too. If neither the dorsoventral or left/right axes reversed, then the DiI droplet would be on the same bilateral side in both half embryos. However, if both of these axes reversed, then the oil droplet would be on the right side of one larva and on the left side of the other (Fig. 5). Fig. 6 shows two pairs of embryos in which this double axis reversal occurred. One larva for each pair is labeled on the left (Fig. 6A,C) and the other on the right (Fig. 6B,D). Out of ten embryo pairs that appeared to develop normally, eight had the left/right pattern of dye labeling, demonstrating that the dorsal half larvae usually inverts both the dorsoventral axis and the position of the hydropore. The other two pairs of dorsal and ventral halves had dye on the same side in each half; one pair had a DiI droplet on the right side of both larvae and another pair were both labeled on the left side (Table 4). These two pairs of larvae clearly indicate that the neither the dorsoventral axis or the position of the hydropore reverse in some dorsal halves.

Table 4.

Results from cell marking and cell separation experiment

Results from cell marking and cell separation experiment
Results from cell marking and cell separation experiment
Fig. 6.

2-day old dorsal and ventral half larvae. These larvae were produced by injecting one cell at the 2-cell stage with two DiI/oil droplets and then cutting the embryo along the second cleavage plane at the 4-cell stage, so that each half (dorsal and ventral) retained an oil droplet. The two half larvae were raised as a pair and when they were 2 days old, they were examined for the position of the DiI/oil droplet (arrow). All of the photographs were taken from the ventral (oral) view. The bright-field images are in the left-hand column while the fluorescent images are in the right-hand column. (A,B) One dorsal and ventral pair. (A) The larva is labeled on the left side. (B) The larva is labeled on the right side. (C,D) Another dorsal and ventral pair. (C) The larva is labeled on the left side. (D) The larva, which is tilted slightly onto its right side, is labeled on the right side. Bar, 21.3 μm.

Fig. 6.

2-day old dorsal and ventral half larvae. These larvae were produced by injecting one cell at the 2-cell stage with two DiI/oil droplets and then cutting the embryo along the second cleavage plane at the 4-cell stage, so that each half (dorsal and ventral) retained an oil droplet. The two half larvae were raised as a pair and when they were 2 days old, they were examined for the position of the DiI/oil droplet (arrow). All of the photographs were taken from the ventral (oral) view. The bright-field images are in the left-hand column while the fluorescent images are in the right-hand column. (A,B) One dorsal and ventral pair. (A) The larva is labeled on the left side. (B) The larva is labeled on the right side. (C,D) Another dorsal and ventral pair. (C) The larva is labeled on the left side. (D) The larva, which is tilted slightly onto its right side, is labeled on the right side. Bar, 21.3 μm.

Axis specification and determination in a variety of organisms has been well studied, particularly with reference to the dorsoventral and anteroposterior (animal/vegetal) axes. Nevertheless, the process by which the third, bilateral, axis is established has been a ‘deep and neglected problem’ (Brown and Wolpert, 1990). This is partly due to the fact that, unlike the two other axes, the left/right axis often superficially results in a mirror image and there are no molecular markers for molecules found on only one side of the embryo. Since a bilaterally symmetric animal does not require any differential left-right positional information (only information on the distance from the midline), the differentiation of bilaterality often is considered uninteresting. However, bilateral animals are rarely completely symmetric; thus, there must be mechanisms by which left-right asymmetries are aligned with respect to the other two axes. Only recently have researchers begun to investigate these mechanisms of bilateral asymmetry specification (for reviews, see Berg, 1991; Brown et al., 1991; Brueckner et al., 1991; Morgan, 1991; Wood and Kershaw, 1991; Yost, 1991). We chose the sea urchin embryo as a model system because the position of the echinus rudiment is bilaterally asymmetric and the embryos can be experimentally manipulated so that left/right specification and determination can be examined.

The study of left-right pattern formation requires knowledge of when the bilateral axis is first specified. In nine species of sea urchins, the relationship between the first cleavage plane and the dorsoventral or left/right axes has been characterized with lineage tracing markers. Four different spatial relationships were found (Hörstadius and Wolsky, 1936; Kominami, 1988; Cameron et al., 1989; Henry et al., 1990, 1992). One of these patterns, found in Lytechinus pictus, Strongylocentrotus droebachiensis and Heliocidaris tuberculata, is that the first cleavage plane marks the dorsoventral (aboral-oral) axis, thus bisecting the embryo into left and right halves. This was true for Lytechinus variegatus, as demonstrated with a novel lineage tracing technique. DiI, mixed with oil to create a longer lasting intracellular microdroplet, was followed along with the fluorescent dye which diffused throughout the daughter cells. It was found that the initial cleavage conditionally divides the embryo into right and left halves. The term ‘conditional’ is used since experimental manipulations show that this normal fate is not yet fixed. Since a switch in the dorsoventral axis also switches the bilateral location of the left/right structures, their position is not fixed by cytoplasmic determinants present in the egg. An unanticipated observation reinforces this conclusion. In our lineage experiment there was a statistically significant (Chi square P<0.001) bias towards embryos being labeled on the right side. An asymmetric distribution of lineage tracer was also observed in S. droebachiensis when Nile Blue was the lineage tracer, in L. pictus when a fluorescent dye was used (Henry et al., 1992), and in phoronids who were marked with either Nile Blue or a fluorescent dye (Freeman, 1991). Thus, since the marking of a cell at the 2-cell stage can somehow bias this cell towards becoming the right side, this reinforces the conclusion that neither the bilateral nor the dorsoventral axes are prespecified via cytoplasmic determinants.

Only a few incomplete studies examined the left-right asymmetry in sea urchins prior to this study. When P. lividus embryos were separated between the swimming blastula and mesenchyme blastula stages, it was observed that the left and right pairs had asymmetric arm deficiencies, narrowing the time frame of bilateral determination (Hörstadius and Wolsky, 1936; Hörstadius, 1973). Unfortunately, the left/right pairs were not reared to an age where echinus rudiment differentiation could be detected, so it was not learned whether this singular bilaterally asymmetric feature was fixed at the time of the operation. In our study, when L. variegatus embryos were cut into left and right halves at different times after fertilization and raised as pairs, the resulting larvae suggest that the left/right asymmetry still is not fixed by the time of the early blastula stage (5.5–6.5 hours), since all pairs resulted in normal larvae. If the left/right halves were separated during late blastula (6.5–7.5 hours old), one half (presumably the left half) retained normal left coelom development but the other half sometimes had left or right hydropores, or a pair of hydropores. This is similar to what was observed in starfish embryos (Runnström, 1920; Hörstadius, 1928, 1973). Since abnormalities in left/right patterning arise at this time in these experimental half larvae, the late blastula stage appears to be the period during which the left/right axis becomes committed. This time frame also corresponds to when the dorsoventral axis is fixed in sea urchins and starfish (Runnström, 1920; Hörstadius, 1936, 1973, Hardin et al., 1992), indicating that the commitment of both of these axes is temporally coincidental.

Earlier, it was demonstrated that when 2-cell or 16-cell stage embryos were separated into dorsal and ventral halves, the dorsal half could invert its dorsoventral axis (Hörstadius and Wolsky, 1936; Hörstadius, 1973; Raff, 1992; Henry and Raff, 1993). We took advantage of this observation and manipulation technique to study how the left-right asymmetry is established with respect to the dorsoventral axis. By simply cutting Lytechinus embryos into dorsal and ventral halves at the 4-cell stage and raising them to an advanced larval stage, it was found that, when the dorsoventral axis reversed, so did the left-right polarity. This was confirmed by performing the double DiI injection and cell separation experiment. Also, when the dorsoventral axis did not reverse in these dorsal half embryos (in two cases) the left/right axis did not reverse either. This supports the hypothesis that the specification of these two axes is intimately coupled.

The marking and cell separation experiments indicate that in every case the sea urchin embryo specifies left/right asymmetry coincident with the dorsoventral specification. Although this type of pattern formation is suspected to be true for all bilaterally asymmetric organisms (see discussion section in Wood and Kershaw, 1991), this is the first time that it was tested with lineage tracing and under experimental conditions that allowed for reversal of the dorsoventral axis. One other animal has been examined after dorsoventral axis reversal. In the nematode Caenorhabditis elegans, two cells at the 3-cell stage were manipulated so they exchanged positions, resulting in the reversal of the dorsoventral axis. Since the embryos with a reversed dorsoventral axis had normal asymmetry of the left/right axis with respect to the new dorsoventral axis orientation, the left/right polarity had inverted too (Priess and Thomson, 1987). Thus, the left/right axis is not committed until the dorsoventral axis is established, which occurs between the 3-cell and 6-cell stage (Wood, 1991). Thus, it was concluded that in C. elegans the asymmetries along the left/right axis are not specified at least until the dorsoventral axis is specified and the differences between left and right are due to cell interactions (Wood, 1991). Like C. elegans, sea urchin embryos have a period during which reversal of the dorsoventral axis results in reversal of the left/right axis. At some point after this time, the asymmetric bilateral axis is fixed. The experiments that we have described, as well as those with C. elegans, again demonstrate that at the very least we know the bilateral axis and the polarity of that axis is dependent upon cell interactions, not segregation of cytoplasmic determinants.

A most perplexing aspect of axis specification in sea urchins is that when the dorsoventral axis is reversed it is inverted by 180 ° and only in the dorsal half. Why it happens only in one half and how the exact hemispherical inversion is controlled is not understood. It was previously hypothesized that there is an gradient from ventral to dorsal and the gradient is stabilized in the ventral half, thus preventing axis inversion (Hörstadius, 1973). Presumably, in the dorsal half the molecular machinery orients the position of this axis. The polarity of the axis, on the other hand, is less stable than the orientation; thus, the act of cutting the embryo in half disrupts the polarity but not the position of the dorsoventral axis. We hypothesize that the polarity is randomized when the dorsal half is removed from the ventral half. This would explain the observation that some dorsal half embryos (from the DiI injection/separation experiment) did not reverse this axis, while others did rotate the axis 180°. Our results agree with previous observations in H. erythrogramma embryos, where not all dorsal half embryos reverse their dorsoventral axis (Raff, 1992; Henry and Raff, 1993). Whether first cleavage fixes the axis orientation but not the polarity is not known.

How might left/right polarity become specified with respect to the dorsoventral and anteroposterior axes? Recently, a model for the establishment of handedness in animals was proposed that consisted of three components: conversion, random generation of asymmetry and interpretation. Conversion is the first step by which a molecular asymmetry is translated into handedness at a cellular level. Most importantly, it requires that the handed molecules orient with respect to the anteroposterior and dorsoventral axes (Brown and Wolpert, 1990). Therefore, if one of these primary axes reverses, two changes are made: the handed molecules reverse polarity and, as a result, the left/right axis also inverts. Our micromanipulation and lineage tracing experiments support this hypothesis. At a molecular level, this interdependency of the dorsoventral and left/right axes is not well understood. However, there are two genes in mice which, when mutated, have an effect on bilateral asymmetry. The iv gene effects the random generation of asymmetry, the second step of the Brown and Wolpert (1990) model, (Layton, 1976) while the inv gene results in complete inversion of left/right polarity (Yokoyama et al., 1993). It is also known that the extracellular matrix in frog embryos provides cues for bilateral axis asymmetry (Yost, 1990, 1992). In its simplest form, the establishment of the anteroposterior and dorsoventral axes must create a differential, bilateral expression of factors, to the left side or to the right. At a later stage, the embryo must be able to ‘read’ these specified asymmetric differences and respond by generating an asymmetric left/right pattern. This model must involve cell interactions since, experimentally, axial specification is shown to occur in the multicellular embryo.

We are particularly indebted to Norris Armstrong for the insight and enthusiasm that he brought to the conversations between he and E. R. M. concerning this research. Support provided by HD14483.

Berg
,
H. C.
(
1991
).
Bacterial motility: handedness and symmetry
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
58
72
.
England
:
Wiley, Chichester (Ciba Foundation Symposium 162
).
Brown
,
N. A.
,
McCarthy
,
A.
and
Wolpert
,
L.
(
1991
).
Development of handed body asymmetry in mammals
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
182
-
201
. Wiley, Chichester (Ciba Foundation Symposium 162).
Brown
,
N. A.
and
Wolpert
,
L.
(
1990
).
The development of handedness in left/right asymmetry
.
Development
109
,
1
9
.
Brueckner
,
M.
,
McGrath
,
J.
,
D’Eustachio
,
P.
and
Horwich
,
A. L.
(
1991
).
Establishment of left-right asymmetry in vertebrates: genetically distinct steps are involved
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
202
-
218
. Wiley, Chichester (Ciba Foundation Symposium 162).
Cameron
,
R. A.
,
Fraser
,
S. E.
,
Britten
,
R. J.
and
Davidson
,
E. H.
(
1989
).
The oral-aboral axis of a sea urchin embryo is specified by first cleavage
.
Development
106
,
641
647
.
Freeman
,
G.
(
1991
).
The bases for and timing of regional specification during larval development in Phoronis
.
Development
147
,
157
173
.
Hardin
,
J.
,
Coffman
,
J. A.
,
Black
,
S. D.
and
McClay
,
D. R.
(
1992
).
Commitment along the dorsoventral axis of the sea urchin embryo is altered in response to NiCl2
.
Development
116
,
671
685
.
Henry
,
J. J.
,
Klueg
,
K. M.
and
Raff
,
R. A.
(
1992
).
Evolutionary dissociation between cleavage, cell lineage and embryonic axes in sea urchin embryos
.
Development
114
,
931
938
.
Henry
,
J. J.
and
Raff
,
R. A.
(
1993
).
Progressive determination of cell fate along the dorsoventral axis in the sea urchin Heliocidaris erythrogramma. In press
Henry
,
J. J.
,
Wray
,
G. A.
and
Raff
,
R. A.
(
1990
).
The dorsoventral axis is specified prior to first cleavage in the direct developing sea urchin Heliocidaris erythrogramma
.
Development
110
,
875
884
.
Hörstadius
,
S.
(
1928
).
über die Determination des Keimes bei Echinodermen
.
Acta zool., Stockholm
9
,
1
39
Hörstadius
,
S.
(
1936
).
über die zeitliche Determination im Keim von Paracentrotus lividus LK
.
Wilhelm Roux’s Arch. EntwMech. Org
.
135
,
1
39
.
Hörstadius
,
S.
(
1973
).
Experimental Embryology of Echinoderms
.
Oxford
:
Clarendon Press
.
Hörstadius
,
S.
and
Wolsky
,
A.
(
1936
).
Studien über die Determination der Bilateralsymmetrie des jungen Seeigelkeimes
.
Wilhelm Roux’s Arch. EntwMech. Org
.
135
,
69
113
.
Kiehart
,
D. P.
(
1982
).
Microinjection of echinoderm eggs: Apparatus and procedures
.
Method Cell Biol
.
25
,
13
31
.
Kominami
,
T.
(
1988
).
Determination of dorso-ventral axis in early embryos of the sea urchin, Hemicentrotus pulcherrimus
.
Dev. Biol
.
127
,
187
196
.
Layton
,
W. M.
(
1976
).
Random determination of a developmental process
.
J. Hered
.
67
,
336
338
.
Morgan
,
M. J.
(
1991
).
The asymmetrical genetic determination of laterality: flatfish, frogs and human handedness
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
234
-
250
. Wiley, Chichester (Ciba Foundation Symposium 162).
Priess
,
J. R.
and
Thomson
,
J. N.
(
1987
).
Cellular interactions in early C. elegans embryos
.
Cell
48
,
241
250
.
Raff
,
R. A.
(
1992
).
Direct-developing sea urchins and the evolutionary reorganization of early development
.
BioEssays
14
,
211
218
.
Runnström
,
J.
(
1920
).
Entwicklungsmechanische Studien and Henricia Sanguinolenta Forbes und Solaster spec
.
Wilhelm Roux’s Arch. EntwMech. Org
.
40
,
459
484
.
Wood
,
W. B.
(
1991
).
Evidence from reversal of handedness in C. elegans embryos for early cell interactions determining cell fates
.
Nature
349
,
536
538
.
Wood
,
W. B.
and
Kershaw
,
D.
(
1991
).
Handed asymmetry, handedness reversal and mechanisms of cell fate determination in nematode embryos
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
143
164
.
Wiley, Chichester (Ciba Foundation Symposium 162
).
Yokoyama
,
T.
,
Copeland
,
N. G.
,
Jenkins
,
N. A.
,
Montgomery
,
C. A.
,
Elder
,
F. F. B.
and
Overbeek
,
P. A.
(
1993
).
Reversal of left-right asymmetry: A situs inversus mutation
.
Science
260
,
679
682
.
Yost
,
H. J.
(
1990
).
Inhibition of proteoglycan synthesis eliminates left-right asymmetry in Xenopus laevis cardiac looping
.
Development
110
,
865
874
.
Yost
,
H. J.
(
1991
).
Development of the left-right axis in amphibians
. In
Biological Asymmetry and Handedness
(ed.
G. R.
Bock
and
J.
Marsh
), pp.
165
181
.
Wiley
,
Chichester (Ciba Foundation Symposium 162
).
Yost
,
H. J.
(
1992
).
Regulation of vertebrate left-right asymmetries by extracellular matrix
.
Nature
357
,
158
161
.