INTRODUCTION

In the unfertilized eggs of many kinds of animal, there is asingle axis, namely, the animal-vegetal axis, and eggs exhibitradial symmetry along the axis. The second axis is establishedjust after fertilization or during early embryogenesis, andembryos become bilaterally symmetrical. In the ascidian egg,the second axis, which corresponds to the future anterior-posterior axis, is generated during ooplasmic segregation.Movements of the ooplasm progress in two phases betweenfertilization and the first cleavage (Conklin, 1905; Hirai, 1941;Sawada and Osanai, 1981; Jeffery and Meier, 1983; Sardet etal., 1989). During the second phase of ooplasmic segregation,vegetally located cytoplasm moves towards the future posteriorpole (Fig. 1). This movement is the first observable evidenceof an anterior-posterior axis during ascidian embryogenesis.

The study of ascidian embryogenesis has provided varioustypes of evidence for the presence of localized maternalfactors, in specific regions of the egg cytoplasm, that playimportant roles in the determination of developmental fatesduring early embryogenesis (reviewed by Venuti and Jeffery,

1986; Meedel, 1992; Nishida, 1992b; Satoh, 1993). Indeed,experiments involving redistribution of cytoplasm and transferof cytoplasm have revealed the presence of cytoplasmic deter-minants that confer on muscle-, endoderm- and epidermis-lineage cells the ability to become muscle, endoderm, andepidermis, respectively (Whittaker, 1980, 1982; Deno andSatoh, 1984; Nishida, 1992a, 1993, 1994). These three kindsof cytoplasmic determinant move in different directions duringooplasmic segregation. Prior to the onset of the first cleavage,three kinds of determinant settle at sites corresponding tospecific regions of the future bilateral fate map (Nishida, 1994).

In addition to determinants of cell type, cytoplasmic deter-minants for gastrulation movements are also present in theooplasm. Ortolani (1958) bisected fertilized eggs near theequator and showed that the animal fragments developed intopermanent blastulae. Bates and Jeffery (1987) showed thatwhen small region of vegetal-pole cytoplasm was removedfrom the zygote between the first and second phase of segre-gation, the embryos fail to gastrulate. The removal ofcytoplasm at the vegetal pole could be mimicked by UV irra-diation of the fertilized egg near the vegetal pole (Jeffery,

3093Development 120, 3093-3104 (1994)Printed in Great Britain © The Company of Biologists Limited 1994

Unfertilized eggs of the ascidian

Halocynthia roretzi areradially symmetrical along the animal-vegetal axis. Afterfertilization, ooplasmic segregation results in formation ofan anterior-posterior axis horizontally, and eggs becomebilaterally symmetrical. When 8-15% of the cytoplasm ofthe posterior-vegetal region of the egg was removed afterthe second phase of ooplasmic segregation, most of theembryos completed gastrulation but developed into radial-ized larvae along the animal-vegetal axis with no apparentanterior-posterior axis. Removal of cytoplasm from otherregions did not affect formation of this latter axis. Thecleavage pattern of the embryos that were deficient inposterior-vegetal cytoplasm (PVC) exhibited radialsymmetry instead of the complicated bilateral symmetry ofnormal embryos. Detailed comparisons of cleavagepatterns revealed the duplication of the anterior cleavagepattern in the originally posterior halves of the PVC-deficient embryos. The PVC-deficent larvae lacked musclecells, which are normally derived from the posterior blas-tomeres. Examination of the developmental fates of theearly blastomeres of the PVC-deficient embryos revealed

that all of the vegetal blastomeres had assumed anteriorfates. These results suggest that the PVC-deficient embryosare totally anteriorized.

When posterior-vegetal cytoplasm was transplanted tothe anterior-vegetal position of PVC-deficient eggs, theaxial deficiency was overcome, and reversal of the anterior-posterior axis was observed. The results of transplantationof posterior-vegetal cytoplasm to the anterior-vegetalposition in normal eggs demonstrated that formation of theanterior structure is suppressed by posterior-vegetalcytoplasm. These results suggest that posterior fate isspecified by the presence of posterior-vegetal cytoplasm,while anterior fate is specified by the absence of posterior-vegetal cytoplasm. Thus, posterior-vegetal cytoplasmdetermines the anterior-posterior axis by generating theposterior cleavage pattern and conferring posterior fateson cells, as well as by inhibiting anterior fates that wouldotherwise occur by default.

Key words: ascidian embryogenesis, axis determination, cytoplasmicdeterminants, cleavage pattern, cytoplasmic transfer

SUMMARY

Localization of determinants for formation of the anterior-posterior axis in

eggs of the ascidian

Halocynthia roretzi

Hiroki Nishida

Department of Life Science, Tokyo Institute of Technology, Nagatsuda, Midori-ku, Yokohama 227, Japan

3094

1990). These results suggest that factors that specify the site ofgastrulation, known as axial determinants, are localized in thevegetal-pole region after the first phase of segregation. Incontrast to our knowledge of the determinants that are relatedto animal-vegetal axis, little is known about the specificationof the anterior-posterior axis of ascidian embryos. In this study,localized cytoplasm that is responsible for determination of theanterior-posterior axis was identified by the removal and trans-plantation of cytoplasm from specific regions of the eggs.

MATERIALS AND METHODS

EmbryosNaturally spawned eggs of

Halocynthia roretzi were artificially fer-tilized and then manually dechorionated with sharpened tungstenneedles. To facilitate normal development, dechorionated embryoswere reared in the supernatant of a homogenate of cleaving embryosthat contained 50

µg/ml streptomycin sulfate and 50 µg/ml kanamycinsulfate (Nishida and Satoh, 1985). At 13°C, tadpole larvae hatchedabout 35 hours after fertilization. The temperature was lowered to 9°Cto lengthen the duration of the stage required for microsurgery.

HistologyEggs undergoing ooplasmic segregation were fixed in Bouin’s fluid.The fixed specimens were dehydrated through an ethanol series andcleared with xylene. The cleared specimens were embedded inparaffin, sectioned at 8 µm and attached to glass slides. Milligan’strichrome staining was performed essentially as described by Jeffery(1989). After treatment with xylene to remove paraffin, sections wererinsed successively in 100% and 95% ethanol, allowed to mordant inpotassium dichromate-HCl for 5 minutes, and stained with acidfuchsin for 8 minutes. The stain was fixed in phosphomolybdic acidfor 3 minutes, and then slides were counterstained with orange G for10 minutes and with fast green for 15 minutes. They were then washedwith 1% acetic acid for 3 minutes. The stained sections were dehy-drated in 95% and 100% ethanol, cleared in xylene and mounted inParmount (Fisher Scientific, New Jersey).

Removal of egg fragmentsFertilized eggs after the second phase of ooplasmic segregation wereoriented for microsurgery using the positions of the polar bodies andthe transparent myoplasm as markers (Fig. 1; Nishida, 1992a). Then,fragments of various regions were severed from eggs with a glassneedle under a stereomicroscope (SZH-10; Olympus). The volumesof the egg fragments that had been removed were calculated from theirdiameters, which were measured with an ocular micrometer.

Markers of the differentiation of specific tissuesDifferentiation of epidermis was evaluated by monitoring theexpression of the Epi-2 antigen, which is specifically recognized bya monoclonal antibody (Nishikata et al., 1987b). Formation ofendoderm was monitored by histochemical detection of alkaline phos-phatase (AP) activity by the method of Whittaker and Meedel (1989),with 5-bromo-4-chloro-3-indolyl phosphate (BCIP) as the substrate.The reaction results in the formation of brownish-purple deposits.Differentiation of muscle was monitored by following the expressionof myosin heavy chains and acetylcholinesterase (AchE). Myosin wasdetected with a monoclonal antibody (Nishikata et al. 1987a; Makabeand Satoh, 1989). Histochemical detection of AchE activity wascarried out as described by Karnovsky and Roots (1964) with acetyl-choline iodide as the substrate. The reaction results in deposition of abrown precipitate. Differentiation of notochord was evaluated fromobservations of cellular morphology or by monitoring the expressionof the Not-1 antigen, which is specifically recognized by a monoclonal

antibody (Nishikata and Satoh, 1990). All the monoclonal antibodieswere generously provided by Dr T. Nishikata (Konan University,Kobe, Japan). Indirect immunofluorescence staining with monoclonalantibodies was carried out by standard methods using FITC-conju-gated secondary antibody.

Scanning electron microscopy (SEM)Embryos were fixed in 2.5% glutaraldehyde in seawater for 1 hour atroom temperature, and then they were dehydrated through a gradedethanol series. Critical-point drying was performed with CO2 in amodel HCP-1 (Hitachi) apparatus. The dried embryos were orientedon aluminum stubs and coated with a thin layer of gold using a an ioncoater (model IB-2; Eiko Engineering). Samples were observed andphotographed with a MINI-SEM (Hitachi-Akashi).

Isolation of blastomeres and inhibition of cell divisionEach blastomere was isolated with a fine glass needle at the 8-cellstage. Blastomeres were cultured separately in agar-coated plasticdishes until controls hatched. To inhibit cell division, 110-cellembryos were cultured in seawater that contained cytochalasin B(Aldrich Chemical Co.) at a final concentration of 2.5 µg/ml. Arrestedembryos were reared until control embryos reached the middle-tailbudstage and then they were processed for immunostaining with anotochord-specific monoclonal antibody.

Transplantation of egg cytoplasm by cell fusionEgg fragments that had been severed from eggs were fused to anterioror posterior regions of egg cells by polyethylene glycol- and electricfield-mediated fusion (PGEF-mediated fusion), as described previ-ously (Nishida, 1994). In brief, an egg fragment was allowed to adherefirmly to an egg cell at the desired position as a result of treatmentwith 30% (w/v) polyethylene glycol in water. Then a single rectan-gular electrical pulse of 800 V/cm was applied to the adhering eggfragment and egg cell for 10-20 µseconds in fusion medium (0.77 MD-mannitol in 0.25% Ca2+-free artificial seawater). Then thespecimen was immediately transferred to seawater. Fused egg cellsdivided with a normal schedule of cleavage. In some cases, eggfragments were stained with 0.025% Nile Blue B in seawater for 2minutes. After fusion, the fate of the region that originated from thefused fragment was traced with the aid of vital blue staining.

RESULTS

Ooplasmic segregation and the embryonic axis in H.roretziThe ooplasm of ascidian eggs moves extensively after fertil-ization, and this process is known as ooplasmic segregation.Fig. 1 shows ooplasmic segregation in eggs of H. roretzi.Movements of the ooplasm occur in two phases between fer-tilization and the first cleavage. In unfertilized eggs, a firstmeiotic spindle is located at the animal pole. Myoplasm, whichis segregated to muscle-lineage cells during embryogenesis, islocated in the cortical region of the egg. It can be recognizedas clear cytoplasm in living eggs and as a deep red, yolklessregion in sections stained by Milligan’s trichrome method (Fig.1A,D), as is also the case in Styela (Jeffery, 1989). The firstphase (0-10 minutes after insemination at 9°C) involves con-traction of the plasma membrane and cortex in the direction ofthe vegetal pole and results in segregation of the myoplasm tothe vegetal region. Eggs show radial symmetry along theanimal-vegetal axis. During the second phase (85-110minutes), the myoplasm moves towards the future posteriorpole together with a sperm aster and forms a myoplasmic

H. Nishida

3095Ascidian A-P axis determinants

domain just vegetal of the equator. After the male pronucleushas encountered the female pronucleus at the posterior pole,they move together to the center of egg. The trail of thepronuclei is visible as a clear region (Fig. 1C, arrow) in livingeggs. The myoplasm and the trail of the pronuclei were usedas posterior markers for the orientation of eggs. During thesecond phase of ooplasmic segregation, the bilateral symmetryof the egg becomes established. The first cleavage occurs 160minutes after fertilization.

In ascidians, the anterior-posterior axis (A-P axis) of eggsand embryos does not correspond precisely to the A-P axis oflarvae because cells do not remain stationary during morpho-genetic movements. In the animal hemisphere, the A-P axis ofeggs and early embryos corresponds precisely to the A-P axisof larvae because positional rearrangement does not take placein the animal hemisphere. Anterior-animal blastomeres giverise to head epidermis of larvae and posterior-animal blas-tomeres develop into tail epidermis. In the vegetal hemisphere,anterior blastomeres, which are designated A-line cells(Conklin, 1905), mainly give rise to anterior endoderm,notochord, and spinal cord. Notochord and spinal cord cells arelocated in the larval tail. Posterior blastomeres, which are des-

ignated B-line cells, mainly develop into posterior endoderm,mesenchyme, and muscle. Muscle cells are found in the tail,while posterior endoderm and mesenchyme cells are found inthe trunk region (Conklin, 1905; Nishida, 1987). In the presentpaper, ‘anterior’ and ‘posterior’ are used in the way that theseadjectives have been traditionally applied to eggs and earlyembryos. Thus, in the present context, ‘anterior fate’ does notmean the fate that gives rise to the anterior structures of larvae.

Removal of cytoplasmic fragments after the secondphase of ooplasmic segregationCytoplasmic fragments with volume equal to 8-15% of that ofan entire egg were removed from various regions of fertilizedeggs with a fine glass needle and the development of themanipulated eggs was examined. Between the first and secondphase of ooplasmic segregation, removal of vegetal cytoplasmresulted in failure of gastrulation, as previously reported inStyela (Bates and Jeffery, 1987). Removal of cytoplasm fromother regions did not affect normal development. After thesecond phase of ooplasmic segregation, removal of animal(n=59), anterior (n=66), lateral (n=56), and vegetal (n=49)cytoplasm had no effect on normal development (Fig. 2A). By

Fig. 1. Ooplasmic segregation in the eggs of Halocynthia roretzi. (A-C) Living eggs. (D-F) Paraffin sections of eggs stained by Milligan’strichrome method. The animal pole is up and the vegetal pole is down. Yolk granules are stained green. (A,D) Unfertilized eggs. A first meioticspindle is visible at the animal pole in the section. (B,E) Fertilized eggs after the first phase of ooplasmic segregation. Myoplasm, which istransparent in the living egg and is stained deep red in the section, has concentrated in the vegetal-pole region. (C, F) Eggs after the secondphase of ooplasmic segregation. The anterior pole is to the left. The second polar body is present at the animal pole. The myoplasm has movedtowards the future posterior pole. Arrow indicates the trail of the pronuclei. Solid line indicates the cutting plane for removal of a posterior-vegetal fragment of the egg. In the sectioned specimen, an aster surrounding the pronuclei is visible. Scale bar, 100 µm.

3096

contrast, when posterior cytoplasm was removed, the eggsdeveloped into malformed larvae in 43% (33 out of 76) of cases(see Fig. 2B,F). The malformed larvae did not have a distincthead and tail, and they seem to be radially symmetrical alongone axis.

To achieve reproducibility of malformation, we carried outthree series of experiments to determine the most effectivetiming of removal, as well as the position and volume ofcytoplasm that was removed. Removal of the posterior-vegetalcytoplasm (Fig. 1C) was more effective than removal of theposteriormost cytoplasm. Removal of the posterior-animalcytoplasm did not result in malformation. The occurrence ofthe malformation was achieved with gradually increasingfrequency when removal of posterior-vegetal cytoplasm wascarried out at 95-125 minutes after fertilization at 9°C, and thefrequency was constant for removal between 125 and 145minutes. After 145 minutes, it was difficult to cut eggs,probably because of formation of the mitotic apparatus. Whenthe posterior-vegetal cytoplasm was removed at 125-145minutes, deletion of 8-15% of the total volume of the eggresulted in malformation in approximately 80% of cases, whileremoval of 5-8% of the egg volume resulted in malformationin about 35% of cases. Removal of ooplasm equivalent to morethan 15% of the egg volume was associated with the possibleremoval of pronuclei. Therefore, in the experiments describedhereafter, posterior-vegetal cytoplasm equivallent to 8-15% ofthe egg volume was removed 125-145 minutes after fertiliza-tion. We designated such eggs as eggs deficient in posterior-vegetal cytoplasm or PVC-deficient eggs.

The possibility that malformation was caused merely by dis-turbance of posterior cytoplasmic architecture during themicrosurgery, was examined. When the cytoplasmic bridgebetween the egg and the fragment was extended until the eggand its fragment were connected by only a thin string, then theglass needle was withdrawn without actual cutting of thebridge, the fragment was retracted into the egg and the eggbecome spherical once again before cleavage occurred. In 81%of such cases, normal larvae developed. In the remaining 19%of cases, larvae showed some abnormalities. However, they allhad a distinct head and tail. These results suggest thatposterior-vegetal cytoplasm must, indeed, be removed from theegg to obtain a radialized larva.

PVC-deficient eggs developed into two morphologicallydistinct types of radialized larva. The larvae with the mor-phology shown in Fig. 2B were designated type A (70% ofradialized larvae) and those shown in Fig. 2F were designatedtype B (30% of radialized larvae). Tissue differentiation wasexamined in these radialized larvae by monitoring theexpression of various molecular markers. Expression of theEpi-2 antigen, which is a marker of epidermis, was observedin all type A and type B larvae. The Epi-2 antigen was localizedin the outer epithelium (Fig. 2C,G). Alkaline phosphatase (AP)activity, which is a marker of endoderm, was observed withinall of the type A and type B larvae (Fig. 2D,H). The Not-1antigen, which is a marker of notochord, was also detected inall radialized larvae. In type A larvae, the innermost cell massexpressed the Not-1 antigen; in type B larvae, some notochordcells protruded from the larvae (Fig. 2E,I). Such protrusion ofcells was the main difference between type A and type Blarvae. Muscle cell was scarcely observed when the expressionof myosin (4% of 72 cases) and of acetylcholinesterase (0%

of 21 cases) was examined. Thus, the radialized larvaeconsisted of an outer epidermal layer, an intermediate endo-dermal layer, and an inner mass of notochord cells. Theylacked muscle cells. Such organization of the three kinds oftissue was also recognized in specimens observed withNomarsky optics (Fig. 2B,F). There were some unidentifiedcells in the larvae because markers for all cell types are not yetavailable. However, it is likely that these unidentified cellswere cells of the spinal cord for the reason given in a latersection.

Cleavage pattern and the development of radializedembryosThe cleavage pattern of ascidian embryos is unique andinvariant (Conklin, 1905; Satoh, 1979). Cleavages progress ina bilaterally symmetrical manner. The cleavage pattern differssignificantly between the anterior half and the posterior half ofthe embryo, especially in the vegetal hemisphere (Fig. 3). Atthe 8-cell stage, the cells of the posterior-vegetal blastomerepair, the B4.1 pair, protrude posteriorly from the embryos (Fig.3B). Successively, only the cells of the most posterior blas-tomere pair at each stage undergo unequal cleavage. Thesecleavages occur three times prior to the 64-cell stage,producing smaller cells posteriorly (Fig. 3E,G).

In most of the PVC-deficient embryos, the cleavage patternwas totally changed and cleavages were radially symmetrical.At the 8-cell stage, there were no protruded blastomeres thatresembled the B4.1 cells (Fig. 3C,D). At the 16- and 32-cellstages, the blastomeres were arranged radially. Therefore, noanterior-posterior axis was recognized in either the animal orthe vegetal hemisphere (Fig. 3F,H), and no unequal cleavagewas observed. Detailed comparisons of SEM images ofembryos at the 76-cell stage (initial gastrula; Fig. 3I,J) and the118-cell stage (middle gastrula; Fig. 3K,L,L′) revealed that thecleavage pattern had been anteriorized in the PVC-deficientembryos. In normal 76- and 118-cell embryos, the arrangementof anterior blastomeres is somewhat radial, while that ofvegetal blastomeres is irregular (Conklin, 1905; Nishida,1986). There is a row of eight spinal cord precursor blas-tomeres in the anterior margin of the vegetal hemisphere(arrowheads in Fig. 3I,K). Inside, there is a row of eightnotochord precursors (arrows). In the center, the cells are endo-dermal precursors (open circle) (Nishida, 1987). In the PVC-deficient radial embryos, the arrangement of all of the vegetalblastomeres resembled that of the anterior part of a normalembryo (Fig. 3J,L). It appeared that a row of spinal cord pre-cursors (arrowheads) and a row of notochord precursors(arrows) encircled the central endodermal precursors (opencircle).

The PVC-deficient embryos gastrulated (Fig. 3L,L′), butformation of a neural tube was not observed. When all of thevegetal cells had gastrulated, type A embryos were formed(Fig. 3N). By contrast, when vegetal cells gastrulated onlypartially, some notochord cells differentiated outside theembryos (Fig. 3O), with resultant formation of type Bembryos.

Developmental fates in the radialized embryosThe cleavage pattern of the PVC-deficient embryos was ante-riorized. In the next set of experiments, we investigatedwhether or not the developmental fate of each blastomere of

H. Nishida

3097Ascidian A-P axis determinants

the PVC-deficient embryo was also anteriorized. In the firstexperiment, cells of 8-cell embryos were manually dissociatedwith a fine glass needle, and all eight blastomeres from a singleembryo were cultured together as a single group of eight partialembryos. Then the number of partial embryos in each groupthat contained notochord cells was determined. Notochordcells were identified by their morphological features. Theyelongated and protruded from the partial embryos and avacuole formed on one side of each notochord cell (Fig. 4A,arrow; Nishida 1991). During normal embryogenesis, 32 of atotal of 40 notochord cells originate from the anterior-vegetalA4.1 blastomere pair of 8-cell embryos. Only the 8 notochordcells of the tip of the larval tail originate from the posterior-vegetal B4.1 blastomere pair (Nishida, 1987). Recent experi-ments involving isolation of blastomeres showed that A4.1cells autonomously give rise to notochord cells in isolation,

while B4.1 cell do not (Nakatani and Nishida, 1994). Fifteennormal, 8-cell embryos were dissociated into single cells andcultured (Fig. 4A). One of eight partial embryos in a groupdeveloped notochord cells in 7 cases. Two of eight partialembryos contained notochord cells in 8 cases. In no case didthree or more of the eight partial embryos in a group developnotochord cells. The partial embryos that contained notochordcells were always derived from the isolated A4.1 cells. Thus,the maximal number of notochord-containing embryos in agroup was two of eight. Twenty PVC-deficient, radial 8-cellembryos were processed similarly (Fig. 4B). The numbers ofcases in which 1, 2, 3, and 4 of 8 partial embryos developednotochord cells were 0, 7, 7, and 6, respectively. Thenotochord-containing partial embryos were derived exclu-sively from the vegetal isolates. In no case did five or morepartial embryos develop notochord cells. The maximum

Fig. 2. (A) Normal larva derived from an egg from which anterior cytoplasm had been removed. (B-E) Type A larvae that were derived fromeggs from which posterior-vegetal cytoplasm had been removed (PVC-deficient eggs). (F-I) Type B larvae that originated from PVC-deficienteggs. Type A and type B larvae show radial symmetry along the vertical axis with no apparent anterior-posterior axis. There is no distinct headand tail. These larvae consist of the outer epidermal layer, which was stained immunohistochemically for the Epi-2 antigen (C,G); anintermediate endodermal layer, which was stained histochemically for alkaline phosphatase (D,H); and an inner mass of notochord cells, whichwas stained immunohistochemically for the Not-1 antigen (E,I). Scale bar, 100 µm.

3098 H. Nishida

3099Ascidian A-P axis determinants

number was four. These vegetal partial embryos nevercontained muscle cells (n=70), as determined with a mono-clonal antibody against myosin. These results suggest that, inextreme cases, all four vegetal cells in radialized embryosassumed the anterior A4.1 fate.

In the next experiments, cleavage was arrested by treatmentwith cytochalasin B after the 110-cell stage, and thennotochord differentiation was monitored by following theexpression of the Not-1 antigen. Nishikata and Satoh (1990)reported that, when normal 110-cell embryos are arrested,some of the notochord precursor cells express the Not-1antigen. In their experiments, the arrangement and maximumnumber of Not-1-positive cells coincided with those of thenotochord-lineage cells at the 110-cell stage. These earlierresults were confirmed in this investigation. In the anterior-vegetal quarter of the 110-cell embryo there is a row of 8 A-line notochord precursors (A8.5, A8.6, A8.13, and A8.14pairs), and in the posterior-vegetal quarter there are twosmaller B-line precursors (B8.6 pair). This arrangement wasfrequently observed in the cleavage-arrested 110-cell embryos,

as shown in Fig. 4C. The numbers of Not-1-positive cells areshown in Fig. 5. The maximum number was ten. The radial-ized 110-cell embryos were cleavage-arrested and analyzed ina similar manner. In most cases, the number of Not-1-positivecells exceeded ten and reached as many as 17 in some cases(Fig. 5). The distribution of Not-1-positive cells in the arrestedembryos was different from that in the control, and Not-1-positive cells encircled the embryos (Fig. 4D). When normaland radialized embryos were cleavage-arrested and stained foralkaline phosphatase, the activity was detected in severalvegetal-pole cells. These observations suggest that thenotochord-precursor-like cells of the radialized embryosobserved by SEM, as described in the previous section, wereindeed notochord-precursor cells. They support the hypothesisthat the PVC-deficient embryos are anteriorized and consist ofduplicated anterior parts of the normal embryo.

Axis reversal by transplantation of posteriorcytoplasm to an anterior positionThe posterior-vegetal cytoplasm is necessary for formation ofposterior structures of embryos. In our next experiments, weasked whether posterior-vegetal cytoplasm is sufficient for theformation of posterior structures. Egg fragments containingposterior-vegetal cytoplasm were transplanted to heterotopicpositions by fusing the fragments to the anterior-vegetal regionof PVC-deficient eggs by PGEF-mediated fusion. The resultsare shown in Table 1. A total of 82% of the PVC-deficient eggsdeveloped into radialized larvae. When the posterior-vegetalcytoplasm was transplanted to the anterior-vegetal position,68% of specimens developed into tailed larvae with a distincthead and tail (Fig. 6A), while 32% of the specimens gave riseto radialized larvae. The tailed larvae were always found tocontain muscle cells in the tail region when examined byspecific staining for myosin (Fig. 6B). Thus, the deficiency inthe formation of the A-P axis was overcome to a significantextent. In control experiments, anterior-vegetal cytoplasm wastransplanted in the same way, but anterior-vegetal cytoplasmhad no analogous activity (Fig. 6C).

In 32 cases, posterior-vegetal fragments were vitally stainedwith Nile Blue and then fused to anterior-vegetal regions inorder to examine whether the original A-P axis was restored

Fig. 3. Cleavage patterns and development of normal and PVC-deficient embryos. (A) A posterior-vegetal cytoplasmic fragment hasbeen removed from an egg after the second phase of ooplasmicsegregation. (B) Lateral view of a normal 8-cell embryo. The animalpole is up; anterior is to the left. The posterior-vegetal B4.1blastomere pair protrudes from the embryo in the posterior direction.(C,D) Two orthogonal lateral views of the same 8-cell PVC-deficientembryo. The embryo has no apparent anterior-posterior axis.(E) Animal view of a normal 16-cell embryo with the anterior poleup. (F) Animal view of a radialized 16-cell embryo that was derivedfrom a PVC-deficient egg. (G,H) Animal views of 32-cell embryosthat were derived from normal and PVC-deficient eggs, respectively.(I,J) SEM images of vegetal views of normal and radialized embryosrespectively, at the 76-cell stage. (K) Vegetal view of a normal 118-cell embryo. (L,L′) Stereogram of a vegetal view of a radializedembryo at the 118-cell stage. Gastrulation has progressed. In E,G,Iand K, anterior is up. In I,J,K,L and L′, arrowheads showblastomeres that are precursors to spinal cord. Arrows indicatenotochord precursors, and open circles show endodermal precursors.(M) A normal tailbud embryo. (N,O) Type A and type B radializedembryos at the tailbud stage. Scale bar, 100 µm.

Fig. 4. Developmental fates in PVC-deficient embryos. (A) A partial embryos that was derived from an isolated A4.1 blastomere of a normal 8-cell embryo. A notochord cell (arrowhead) is visible. (B) A partial embryos that was derived from a vegetal blastomere of a radialized 8-cellembryo that originated from a PVC-deficient egg. (C) Cell division of a normal embryo had been inhibited from progressing beyond the 110-cell stage. Expression of the notochord-specific antigen, Not-1, was detected by immunostaining. There is a row of eight Not-1-positive cells,which are probably A-line notochord precursors, in addition to two smaller, positive cells, which are B-line precursors (arrowheads). (D) Aradialized 110-cell embryo, which was derived from a PVC-deficient egg and was treated similarly to the embryo in C. In this specimen, 17Not-1-positive cells encircled the embryo. Scale bar, 100 µm.

3100

or reversed. In 22 out of 32 cases (68%), tailed larvaedeveloped. Among the 22 tailed larvae, 18 specimens (82%)had the blue staining in their tail and mesenchyme, which isderived from the posterior (B-line) blastomeres. Fig. 6D-Hshows the cleavage pattern in a typical case. At the 8-cell stage,one blastomere pair protruded in the original anterior direction,resembling the B4.1 blastomere pair (compare Fig. 6G withFig. 3B). These B4.1-like cells divided unequally to yieldlarger B5.1-like cells and smaller B5.2-like cells at the 16-cellstage. These cells inherited labelled posterior-vegetalcytoplasm. These results together indicate that the original A-P axis was reversed by transplantation of posterior-vegetalcytoplasm to an anterior-vegetal position in PVC-deficienteggs. It is suggested that posterior-vegetal cytoplasm is suffi-cient for ectopic formation of posterior structures.

Inhibition of anterior fates by posterior cytoplasmBecause the A-P axis was reversed by transplantation ofposterior-vegetal cytoplasm, one might suppose that genera-tion of anterior structures from the original anterior part of theegg must be inhibited and the original anterior fate must bechanged to posterior fate by the introduced posteriorcytoplasm. This possibility was examined by transplantingposterior-vegetal cytoplasm to the anterior-vegetal position ofnormal eggs with posterior-vegetal cytoplasm in the original

position. Thus, the operated eggs had posterior-vegetalcytoplasm on both sides of the egg. Most of them developedinto malformed larvae, in which discrimination of head and tailwas difficult. Formation of anterior structures was monitoredby examining for the presence of notochord cells (Table 2). Inapproximately half of the cases, the number of cells thatexpressed the Not-1 antigen was significantly reduced (Fig.7A) and, in 7% of cases, no expression of the Not-1 antigenwas observed at all. By contrast, the number of muscle cellsthat expressed myosin increased in 60% of cases (n=20; Fig.7B). These results suggest that posterior-vegetal cytoplasmsuppresses the expression of anterior fate.

In control experiments, anterior-vegetal cytoplasm wastransplanted to the anterior-vegetal position of normal eggs.There was no significant reduction in number of notochordcells. Approximately 70% of larvae had a distinct head and tailand, in 20% of cases, a morphologically normal tailbuddeveloped (Fig. 7C). In another control experiment, anterior-vegetal cytoplasm was transplanted to the posterior-vegetalposition of normal eggs. Of these eggs, 81% (n=47) had adistinct head and tail (Fig. 7D).

DISCUSSION

The results of this investigation suggest that the posterior-vegetal cytoplasm, after the second phase of ooplasmic segre-gation, is required for formation of posterior components of theascidian embryo. The PVC-deficient embryos showed a radi-alized pattern of cleavage, which was equivalent to a duplica-tion of the pattern of cleavage in the anterior half of the normalembryo. The developmental fates of early blastomeres werealso anteriorized. Thus, removal of posterior-vegetal cyto-plasm caused total duplication of the anterior half. Whenposterior-vegetal cytoplasm was transplanted to the anterior-vegetal region, the anterior-posterior axis of the embryos wasreversed. This result suggests that the posterior-vegetalcytoplasm is sufficient for formation of posterior componentseven when transplanted to a non-standard location. Theposterior-vegetal cytoplasm contains most of the myoplasm(Fig. 1C,F), which is known to play an important role in thedetermination of muscle fate. Jeffery and Swalla (1990)pointed out that myoplasm, after the first phase of ooplasmicsegregation, has multiple roles. The results of the presentinvestigation suggest that myoplasm, after the second phase ofooplasmic segregation, is important not only for the determi-nation of muscle, but also for specification of the embryonicaxis by directing posterior fate.

The roles of the posterior-vegetal cytoplasmThe posterior-vegetal cytoplasm appears to include factorsinvolved in four distinct processes, as follows. (1) Muscleformation. The PVC-deficient larvae lacked muscle cells. Itwas shown previously by cytoplasmic-transfer experimentsthat muscle determinants are localized in the posterior regionafter the second phase of ooplasmic segregation (Nishida,1992a). (2) Suppression of anterior fate. When posterior-vegetal cytoplasm was transplanted to the anterior position ofPVC-deficient eggs, reversal of the embryonic axis occurred.When posterior-vegetal cytoplasm was transplanted to theanterior position of normal eggs, formation of the anterior

H. Nishida

Fig. 5. Histogram showing the distribution of numbers of Not-1-positive blastomeres in cleavage-arrested normal 110-cell embryos(upper) and radialized 110-cell embryos that were derived from thePVC-deficient eggs (lower). In normal arrested embryos the numbernever exceeded ten (broken vertical line).

Table 1. Results of transplantation of egg fragments toPVC-deficient eggs

Position ofCytoplasm transplant n Tailed Radialized Significance

– – 65 18% 82%Posterior- Anterior- 50 68% 32% P<0.001vegetal vegetalAnterior- Anterior- 55 15% 85% 0.7>P>0.5vegetal vegetal

The resultant embryos were divided into two categories based on theirmorphology, namely, tailed larvae and radialized larvae. Probability wascalculated by the χ2 test when results were compared to the results in the firstrow, namely, results when no transplantation was carried out.

3101Ascidian A-P axis determinants

Table 2. Results of transplantation of egg fragments to normal eggsNormal Reduced

Position of amount of amount of NoCytoplasm transplant n notochord notochord notochord Significance

Posterior-vegetal Anterior-vegetal 46 41% 52% 7% P<0.001Anterior-vegetal Anterior-vegetal 49 84% 16% 0%Anterior-vegetal Posterior-vegetal 47 81% of embryos developed into tailed larvae.

Numbers of notochord cells were estimated from the expression of the Not-1 antigen. The embryos were tentatively divided into three categories on the basisof the number of notochord cells. Probability was calculated by the χ2 test when the results in the first and second rows were compared.

Fig. 6. Transplantation of posterior-vegetal cytoplasm to PVC-deficient eggs. (A) Posterior-vegetal cytoplasm was transplanted to the anterior-vegetal position of a PVC-deficient egg. A larva with a distinct head and tail developed. (B) The same larva in which myosin was visualized byimmunostaining. Muscle cells can be seen in the tail. (C) Radialized larva in which anterior-vegetal cytoplasm had been transplanted to ananterior-vegetal position. (D-I) Posterior-vegetal cytoplasm was labeled with Nile Blue and then fused to an anterior-vegetal position of a PVC-deficient egg. The original anterior pole is to the left. (D) Just after fusion. Lateral view. The large arrow shows the position of the polar body.The small arrow indicates the trail of pronuclei, which is a marker of the posterior pole of the egg. (E) 2-cell stage. Animal view. (F) 4-cellstage. (G) 8-cell stage. Lateral view. A blastomere pair that has blue labelling protrudes anteriorly from the embryo. (H) The labelled cells havedivided unequally at the 16-cell stage. Vegetal view. The embryo resembles a normal 16-cell embryo but with its axis reversed. (I) Theresultant larva has blue label in its tail and mesenchyme cells (arrowhead). Scale bar, 100 µm.

3102

components of each embryo was inhibited. Therefore, trans-planted posterior-vegetal cytoplasm suppressed anterior fate.(3) Generation of a posterior cleavage pattern. The PVC-deficient embryos showed a radialized cleavage pattern thatwas a duplication of the anterior pattern. Transplantation ofposterior-vegetal cytoplasm to an anterior-vegetal positioncaused reversal of the anterior-posterior polarity of thecleavage pattern. Therefore, posterior-vegetal cytoplasm isnecessary and sufficient for generation of a posterior pattern ofcleavage. (4) Morphogenesis for tail formation. This propertyis less well defined than the above-described three properties.However, the PVC-deficient larvae did lack a distinct tail.There may be factors that control morphogenetic movementsfor formation of the tail. With respect to similar morphogeneticfactors, Bates and Jeffery (1987) and Jeffery (1990) suggestedthe existence of an ooplasmic factor that is responsible for gas-trulation movements in ascidian embryos. When a small regionof vegetal-pole cytoplasm is removed between the first andsecond phase of ooplasmic segregation, or the vegetal pole isirradiated with UV light, the embryos do not gastrulate. In spiteof the deficiency in morphogenesis, such embryos retain theirnormal mechanisms for cell-type specification and a normalcleavage pattern.

It is likely that the PVC-deficient larvae lacked mesenchymecells. This absence was not directly proved because no goodmarkers are available for differentiation to mesenchyme.However, our observations of the cleavage patterns of PVC-deficient embryos at the 76-cell stage and examination ofdevelopmental fates at the 110-cell stage with cleavage-arrested PVC-deficient embryos suggest that the PVC-deficientembryos were totally anteriorized and, therefore, they lackedmesenchyme-lineage cells. Thus, the posterior-vegetalcytoplasm may also be involved in formation of mesenchyme.

It is not clear whether different factors are responsible forthe distinct roles described above, or whether multiple rolesshould be attributed to a single factor. It is possible that allthe properties are attributable to one single factor. InDrosophila, a graded signal that is provided by a concentra-tion gradient of bicoid protein, exerts total control over thedevelopment of the anterior portion of the fly embryo (Drieverand Nusslein-Volhard, 1988). In C. elegans, the activity of theproduct of the skn-1 gene determines the identity of the EMSblastomere of the 4-cell embryo. The EMS cell gives rise tovarious types of cell and generates an invariant cleavagepattern after the 4-cell stage (Bowerman et al., 1992, 1993).In the ascidian, we are only just starting to define the way inwhich the posterior half of the embryos is specified bylocalized maternal factors.

Posterior dominance of developmental fateRemoval of posterior cytoplasm caused posterior blastomeresto adopt the fate of anterior blastomeres. When posteriorcytoplasm was transplanted to an anterior position, posteriorfate overcame anterior fate. By contrast, transplantation ofanterior cytoplasm to a posterior position did not affect devel-opment. Therefore, posterior fate is dominant over anteriorfate. In other words, anterior fate may occur by default. Thepresence of posterior cytoplasm specifies posterior fate, whileanterior fate is directed by the absence of posteriorcytoplasm.

It was shown recently that inductive interactions at the 32-cell stage are involved in formation of notochord duringascidian embryogenesis (Nakatani and Nishida, 1994).Inductive influences are generated from endoderm precursorsand adjacent notochord precursors. Anterior to the notochordprecursors, there are spinal cord precursors (Fig. 3I). We donot yet know whether spinal cord is induced by notochord.Anterior to the spinal cord precursors, there are precursors ofbrain and sensory pigment cells. These cells are induced duringgastrulation (Rose, 1939; Reverberi and Minganti, 1946;Reverberi et al., 1960; Okado and Takahashi, 1988). In H.roretzi, sensory pigment cells are induced by spinal cord pre-cursors at the 180-cell stage (Nishida, 1991). As the notochordand brain are induced, it is likely that anterior components ofascidian embryos are induced sequentially during embryogen-esis. In contrast, endoderm precursors, which are in the vegetalpole region, and muscle precursors, which reside in theposterior region are determined by cytoplasmic factors in theegg (Nishida, 1992a, 1993). Provided that posterior fate isdetermined by cytoplasmic factors present in the egg and, inaddition, that anterior fate is determined by sequential induc-tions during embryogenesis, it is easy to explain the posteriordominance that was observed in the present study. Withoutposterior cytoplasm, sequential inductions also proceed in theoriginal posterior direction, with formation of anterior compo-nents. When posterior cytoplasm is present, sequential induc-tions cannot occur because the developmental fate of cells hasalready been determined by maternal cytoplasmic factors.There may be no maternal determinants unique to the anteriorpart of the zygote.

It is not clear whether a fixed anterior-posterior axis existsin the animal hemisphere at early stages because the major fateof both anterior and posterior animal cells is epidermis, inaddition to brain, which is induced at the anterior margin.

H. Nishida

Fig. 7. Transplantation of egg fragments to a normal egg.(A,B) Posterior-vegetal cytoplasm was transplanted to an anterior-vegetal position of normal eggs. The resultant embryos had reducednumbers of notochord cells that expressed the Not-1 antigen (A) andincreased numbers of muscle cells that expressed myosin (B).(C,D) Anterior-vegetal cytoplasm was transplanted to an anterior-vegetal (C) and to a posterior-vegetal (D) position of normal eggs. Inboth cases, normal larvae developed. Scale bar, 100 µm.

3103Ascidian A-P axis determinants

Mechanisms for generation of a unique cleavagepatternAscidian embryos exhibit a complicated but invariant cleavagepattern (see, for example, Fig. 3I,K). The pattern of cleavagein the animal hemisphere and anterior-vegetal region is ratherregular. By contrast, that in the posterior-vegetal region isextremely irregular, mainly because of successive unequalcleavages. To gain some insight into the mechanisms, thepresent author proposes the hypothesis that the posterior poleof the vegetal hemisphere attracts a spindle pole.

The third cleavage occurs horizontally, but the cleavageplane is slightly inclined and, consequently, the B4.1 cell pairprotrudes posteriorly from the embryo (Fig. 3B). This phe-nomenon may be explained if, within B3 cells (posterior pairat the 4-cell stage), the vegetal pole of the third spindle isdrawn towards the posterior, with resultant inclination of thespindle axis. After the 8-cell stage, three successive, unequalcleavages occur, such that only the cells of the most posteriorblastomere pair at each stage undergo unequal cleavage,always producing smaller cells posteriorly. All the othercleavages are equal cleavages in terms of cell size (for detailsof the cleavage pattern of ascidians, see Conklin, 1905, andSatoh, 1979). The smallest blastomere of the 64-cell embryo,B7.6, never divides during embryogenesis. It gives rise to asingle cell in the endodermal strand (Nishida, 1987). Duringthese unequal cleavages, the posterior pole may attract thespindle pole, with resultant production of smaller cells at theposterior pole. The present investigation suggests that factorsresponsible for the attraction of the spindle pole are transfer-able together with the posterior-vegetal fragments of eggs.These hypotheses must be examined by observations of theaxis and the position of each spindle during normal embryo-genesis, as well as in experimentally manipulated embryo.

Similarity to the dorsoventral axis formation inamphibian eggsIn unfertilized eggs of many kinds of animal, there is ananimal-vegetal axis. The second axis is specified just after fer-tilization or during early embryogenesis. The two axes are thebasis for the bilateral body plan. As pointed out by Bates andJeffery (1987), there are remarkable similarities betweenascidians and amphibians in the process that is involved inestablishment of the second axis of the egg, which is theanterior-posterior axis in ascidians and the dorsal-ventral axisin amphibians. In Xenopus, the specification of the dorsal-ventral axis begins after fertilization with a shift of vegetalcytoplasm to the future dorsal side (Scharf and Gerhart, 1980).This rearrangement of ooplasm may correspond to ooplasmicsegregation in ascidian eggs. In Xenopus, it was shown recentlythat transplantation of dorsal cytoplasm to a ventral positioncauses formation of a secondary axis (Yuge et al., 1990).Dominance of dorsal fate is observed. Hainski and Moody(1992) reported that mRNA from dorsal blastomeres can alterventral blastomeres and yield a secondary axis, an indicationthat dorsal determinants are maternal mRNAs.

Abbate and Ortolani (1961) and Cammarata (1973) removeda small portion of cortical cytoplasm from uncleaved eggs ofCiona and Ascidia by sucking up cytoplasm with amicropipette. They reported that no identifiable developmen-tal defects were observed in such embryos. Bates and Jeffery

(1987) removed egg fragments equal to 10% of the totalvolume from the posterior-vegetal region of eggs of Styelaafter the second phase of ooplasmic segregation. They alsoreported that no significant defects arose. They removed eggfragments by extruding a portion of the egg through a hole inthe chorion and severing it from the remaining ooplasm. Weremoved posterior-vegetal cytoplasm in the same way as theydid using eggs of Halocynthia, and we observed the radializa-tion of embryos. There is a species difference between Styelaand Halocynthia. Bates and Jeffery (1987) state that only 50%of the myoplasm was removed. Because Halocynthia eggshave twice the diameter of Styela eggs, it is possible thatposterior determinants are localized in a relatively restrictedregion. Therefore, we can remove most of the relevant factorswith ease. It is implied that there may be no difference inmechanism between two species.

There appear to be two foci of factors that are developmen-tally important for axis formation in uncleaved ascidian eggs.One is the vegetal-pole region between the first and secondphase of ooplasmic segregation and the other is the posterior-vegetal region after the second phase.

The author thanks Dr N. Satoh (Kyoto University) and Dr R.Kuraishi (Tohoku University) for their critical comments on the man-uscript and all members of my laboratory for useful discussions.Thanks are also due to Dr T. Nishikata (Konan University) forproviding the monoclonal antibodies, Dr T. Numakunai for supplyingliving material, and all other members of the Asamushi Marine Bio-logical Station for facilitating the author’s studies there. The authoralso thanks Dr W. R. Jeffery for lessons in the Milligan’s trichromestaining method.

REFERENCES

Abbate, C. and Ortolani, G. (1961). The development of Ciona eggs afterpartial removal of the cortex or ooplasm. Acta Embryol. Morphol. Exp. 4, 56-61.

Bates, W. R. and Jeffery, W. R. (1987). Localization of axial determinants inthe vegetal pole region of ascidian eggs. Dev. Biol. 124, 65-76.

Bowerman, B., Draper, B. W., Mello, C. C. and Priess, J. R. (1993). Thematernal gene skn-1 encodes a protein that is distributed unequally in early C.elegans embryos. Cell 74, 443-452.

Bowerman, B., Eaton, B. A. and Priess, J. R. (1992). skn-1, a maternallyexpressed gene required to specify the fate of ventral blastomeres in the earlyC. elegans embryo. Cell 68, 1061-1075.

Cammarata, M. (1973) Removal of the cortical regions from the ascidian egg.Acta Embryol. Exper. 115-120.

Conklin, E. G. (1905). The organization and cell lineage of the ascidian egg. J.Acad. Nat. Sci. (Philadelphia) 13, 1-119.

Deno, T. and Satoh, N. (1984). Studies on the cytoplasmic determinant formuscle cell differentiation in ascidian embryos: an attempt at transplantationof the myoplasm. Dev. Growth Diff. 26, 43-48.

Driever, W. and Nusslein-Volhard, C. (1988). The bicoid protein determinesposition in the Drosophila embryo in a concentration-dependent manner.Cell 54, 95-104.

Hainsky, A. M. and Moody, A. S. (1992). Xenopus maternal RNAs fromdorsal animal blastomere induce a secondary axis in host embryos.Development 116, 347-355.

Hirai, E. (1941). The early development of Cynthia roretzi. Sci. Rep. TohokuImp. Univ. Biol. 16, 217-232.

Jeffery, W. R. (1989). Requirement of cell division for muscle actin expressionin the primary muscle cell lineage of ascidian embryos. Development 105,75-84.

Jeffery, W. R. (1990). Ultraviolet irradiation during ooplasmic segregationprevents gastrulation, sensory cell induction, and axis formation in theascidian embryo. Dev. Biol. 140, 388-400.

3104

Jeffery, W. R. and Meier, S. (1983). A yellow crescent cytoskeletal domain inascidian eggs and its role in early development. Dev. Biol. 96, 125-143.

Jeffery, W. R. and Swalla, B. J. (1990). The myoplasm of ascidian eggs: alocalized cytoplasmic domain with multiple roles in embryonicdevelopment. Sem. Cell. Biol. 1, 373-381.

Karnovsky, M. J. and Roots, L. (1964). A ‘direct-coloring’ thiocholinemethod for cholinesterase. J. Histochem. Cytochem. 12, 219-221.

Makabe, K. W. and Satoh, N. (1989). Temporal expression of myosin heavychain gene during ascidian embryogenesis. Dev. Growth Differ. 31, 71-77.

Meedel, T. H. (1992). Development of the ascidian embryo: cell fatespecification by autonomous and inductive processes. In Morphogenesis: AnAnalysis of the Development of Biological Form (eds. E. F. Rossomand andS. Alexander), pp. 263-317. Marcel Dekker, Inc. New York.

Nakatani, Y. and Nishida, H. (1994). Induction of notochord during ascidianembryogenesis. Dev. Biol. in press.

Nishida, H. (1986). Cell division pattern during gastrulation of the ascidian,Halocynthia roretzi. Dev. Growth Diff. 28, 191-201.

Nishida, H. (1987). Cell lineage analysis in ascidian embryos by intracellularinjection of a tracer enzyme. III. Up to the tissue-restricted stage. Dev. Biol.121, 526-541.

Nishida, H. (1991). Induction of brain and sensory pigment cells in the ascidianembryo analyzed by experiments with isolated blastomeres. Development112, 389-395.

Nishida, H. (1992a). Regionality of egg cytoplasm that promotes muscledifferentiation in embryo of the ascidian, Halocynthia roretzi. Development116, 521-529.

Nishida, H. (1992b). Determination of developmental fates of blastomeres inascidian embryos. Dev. Growth Differ. 34, 253-262.

Nishida, H. (1993). Localized regions of egg cytoplasm that promoteexpression of endoderm-specific alkaline phosphatase in embryos of theascidian Halocynthia roretzi. Development 118, 1-7.

Nishida, H. (1994). Localization of egg cytoplasm that promotesdifferentiation to epidermis in embryos of the ascidian Halocynthia roretzi.Development 120, 235-243.

Nishida, H. and Satoh, N. (1985). Cell lineage analysis in ascidian embryos byintracellular injection of a tracer enzyme. II. The 16- and 32-cell stages. Dev.Biol. 110, 440-454.

Nishikata, T., Mita-Miyazawa, I., Deno, T., and Satoh, N. (1987a). Musclecell differentiation in ascidian embryos analyzed with a tissue-specificmonoclonal antibody. Development 99, 163-171.

Nishikata, T., Mita-Miyazawa, I., Deno, T., Takamura, K. and Satoh, N.(1987b). Expression of epidermis-specific antigens during embryogenesis ofthe ascidian, Halocynthia roretzi. Dev. Biol. 121, 408-416.

Nishikata, T. and Satoh, N. (1990). Specification of notochord cells in theascidian embryos analyzed with a specific monoclonal antibody. Cell Differ.Dev. 30, 43-53.

Okado, H. and Takahashi, K. (1988). A simple ‘neural induction’ model withtwo interacting cleavage-arrested ascidian blastomeres. Proc. natn. Acad.Sci. USA 85, 6197-6201.

Ortolani, G. (1958). Cleavage and development of egg fragments in ascidians.Acta Embryol. Morphol. Exp. 1, 247-272.

Reverberi, G. and Minganti, A. (1946). Fenomeni di evocatione nellesviluppo dell’uovo di ascidie. Pubbl. Staz. Zool. Napoli 20, 199-252.

Reverberi, G., Ortolani, G. and Farinella-Ferruzza, N. (1960). The causalformation of the brain in the ascidian larva. Acta Embryol. Morph. Exp. 3,296-336.

Rose, S. M. (1939). Embryonic induction in the ascidia. Biol. Bull. Mar. Biol.Lab. Woods Hole 155, 608-614

Sardet, C., Speksnidjer, J., Inoue, S. and Jaffe, L. (1989). Fertilization andooplasmic movements in the ascidian egg. Development 105, 237-249.

Satoh, N. (1979). Visualization with scanning electron microscopy of cleavagepattern of the ascidian eggs. Bull. Mar. Biol. St. Asamushi 16, 169-178.

Satoh, N. (1993). Developmental Biology of Ascidians. Cambridge: CambridgeUniv. Press.

Sawada, T. and Osanai, K. (1981). The cortical contraction related to theooplasmic segregation in Ciona intestinalis eggs. Wilhelm Roux’s Arch. Dev.Biol. 190, 208-214.

Scharf, S. R. and Gerhart, J. C. (1980). Determination of the dorsal-ventralaxis in eggs of Xenopus laevis: Complete rescue of UV-impaired eggs byoblique orientation before first cleavage. Dev. Biol. 79, 181-198.

Venuti, J. M. and Jeffery, W. R. (1986). Cell lineage and determination of cellfate in ascidian embryos. Intern. J. Dev. Biol. 33, 197-212.

Whittaker, J. R. (1980). Acetylcholinesterase development in extra cellscaused by changing the distribution of myoplasm in ascidian embryos. J.Embryol. exp. Morph. 55, 343-354.

Whittaker, J. R. (1982). Muscle lineage cytoplasm can change thedevelopmental expression in epidermal lineage cells of ascidian embryos.Dev. Biol. 93, 463-470.

Whittaker, J. R. and Meedel, T. H. (1989). Two histospecific enzymeexpressions in the same cleavage-arrested one-celled ascidian embryos. J.Exp. Zool. 250, 168-175.

Yuge, M., Kobayakawa, Y., Fujisue, M. and Yamana, K. (1990). Acytoplasmic determinant for dorsal axis formation in an early embryo ofXenopus laevis. Development 110, 1051-1056.

(Accepted 1 August 1994)

H. Nishida