Early developmental staging from the zygote stage to the gastrula is a basic step for studying embryonic development and biotechnology. We described the early embryonic development of the loach, Misgurnus anguillicaudatus, based on morphological features and gene expression. Synchronous cleavage was repeated for 9 cycles about every 27 min at 20°C after the first cleavage. After the 10th synchronous cleavage, asynchronous cleavage was observed 5.5 h post-fertilization (hpf), indicating the midblastula transition. The yolk syncytial layer (YSL) was formed at this time. Expressions of goosecoid and no tail were detected by whole-mount in situ hybridization from 6 hpf. This time corresponded to the late-blastula period. Thereafter, epiboly started and a blastoderm covered over the yolk cell at 8 hpf. At 10 hpf, the germ ring and the embryonic shield were formed, indicating the stage of early gastrula. Afterward, the epiboly advanced at the rate of 10% of the yolk cell each hour. The blastoderm covered the yolk cell completely at 15 hpf. The embryonic development of the loach resembled that of the zebrafish in terms of morphological change and gene expression. Therefore, it is possible that knowledge of the developmental stages of the zebrafish might be applicable to the loach.
The staging of early embryogenesis is a basic step for developmental biology. Developmental stages based on morphological features have been reported in many fish species (Fundulus heteroclitus; Armstrong et al., 1965, salmonid fishes, Salmo gairdneri and Salvelinus fontinalis; Ballard, 1973, medaka, Oryzias latipes; Iwamatsu, 1994, American shad, Alosa sapidissima; Shardo, 1995, ice goby (shiro-uo), Leucopsarion petersii; Arakawa et al., 1999, and others), but detailed staging based on developmental genetics has been studied only in zebrafish Danio rerio (Kimmel et al., 1995) and goldfish Carassius auratus (Yamaha et al., 1999).
According to Yamaha et al. (1999), a blastodisc cleaves synchronously in early embryonic development after fertilization, then mitotic divisions occur asynchronously. Such a turning point in cell division is referred to as the mid-blastula transition (MBT), which is characterized by cell cycle lengthening, loss of synchronism in cell divisions, activation of transcription and appearance of cell motility in Xenopus (Newport and Kirschner, 1982) and zebrafish (Kane and Kimmel, 1993). After MBT, the yolk syncytial layer (YSL) is formed and epiboly begins. The specification occurs in several regions of the blastoderm due to induction from the yolk cell (Mizuno et al., 1996; Ober and Schulte, 1999) and the expression of zygotic genome in the proliferated blastomeres (e.g., no tail at blastoderm margin, Schulte-Merker et al., 1992; goosecoid at the dorsal side of blastoderm, Stachel et al., 1993). When embryos initiate gastrulation, morphogenetic movements occur and then germ layers are formed. On the other hand, experimental embryological studies have revealed that embryonic cells before the gastrula stage have a high regulative ability (Ho and Kimmel, 1993).
Gene transfer (medaka; Ozato et al., 1992, zebrafish; Linney et al., 1999, rainbow trout, Oncorhynchus mykiss; Yoshizaki et al., 2000, mud loach, Misgurnus mizolepis; Nam et al., 2001), nuclear transplantation (European loach, Misgurnus fossilis; Gasaryan et al., 1979, medaka; Wakamatsu et al., 2001, zebrafish; Lee et al., 2002) and germ line chimera (zebrafish; Lin et al., 1992, medaka; Wakamatsu et al., 1993, goldfish; Yamaha et al., 2001, rainbow trout; Takeuchi et al., 2001, loach, Misgurnus anguillicaudatus; Nakagawa et al., 2002) have been developed and performed in several fishes from the viewpoints of basic biology as well as advanced aquaculture. These biotechnological techniques are essentially applicable to the early embryos before gastrulation because the developmental fate of the blastomeres has not yet been committed; thus various types of embryo manipulation are possible. Therefore, it is necessary to identify the embryonic stages of the target species for further biological practices.
In the present study, we describe the early embryonic stages from cleavage to the gastrula period of the loach Misgurnus anguillicaudatus, based on morphological features, cytological observation, histological results and whole-mount in situ hybridization.
MATERIALS AND METHODS
Egg and sperm
Parental fish of the loach, Misgurnus anguillicaudatus, were caught from paddy fields and water ways in Kita-village, Sorachi county, Hokkaido prefecture, Japan by the Loach Farming Cooperation of Kitamura during the spawning period from June to July 2001. We induced ovulation and spermiation to collect gametes by intraperitoneal injections of Human Chorionic Gonadotropin (20 IU/g body weight for males and females). Eggs were manually stripped on polyvinyl chloride film. Sperm was collected into hematocrit blood drawing tubes and diluted 1: 20 with loach physiological saline (7.5 g/L NaCl, 0.2 g/L KCl, 0.2 g/L MgCl2 and 0.4 g/L CaCl2, pH 7.8 by NaHCO3). Eggs were inseminated with diluted sperm and fertilized by tap water in glass petri dishes at 20°C. Fertilized eggs were dechorionated by treatment with Ringer's solution (7.5 g/L NaCl, 0.2 g/L KCl and 0.2 g/L CaCl2) containing 0.1% trypsin and 0.4% urea for about 10 min. Dechorionated eggs were transferred to Ringer's solution containing 1.6% albumen, and were incubated at 20°C after the transfer.
Visualization of nuclei
About fifty dechorionated eggs were fixed every 10 min from 60 to 360 min post-fertilization (mpf) with 2% glutaraldehyde in phosphate-buffered saline (PBS) overnight. Fixed embryos were stored in PBS at 4°C. After the removal of the blastodisc from the yolk cell, the nuclei were stained with 5 mg/ml 4′-6-diaminido-2-phenylindole (DAPI) dissolved in 10 mM Tris-HCl (pH 7.4), 5mM EDTA, and 0.15 M NaCl for 1 h or more. The blastodiscs were then washed with PBS, and observed from an animal pole under a fluorescence microscope (Olympus Model BH-2).
Embryos were fixed with Bouin's fixative for 3 h. Fixed embryos were stored in 80% ethanol and then dehydrated by butyl alcohol series and embedded in paraffin blocks. Serial sections were cut at 8 μm thickness and stained with hematoxylin-eosin.
Whole-mount in situ hybridization (WISH) analysis
From 5 to 15 h post-fertilization (hpf), dechorionated eggs were fixed every hour with 4% paraformaldehyde in PBS for 30 h. Fixed embryos were stored in 100% methanol at −20°C. WISH was performed by staining with a single color in the blastula stage, and with two colors in the gastrula stage according to Jowett and Lettica (1994) with slight modifications.
goosecoid (gsc) (Stachel et al., 1993) and no tail (ntl) (Schulte-Merker et al., 1992) of zebrafish were used as dorsal mesodermal and pan-mesodermal markers, respectively (courtesy of Dr. H. Takeda; ntl was originally cloned by Dr. S. Schulte-Merker and gsc by Drs. M. Tada and N. Ueno). For two-color in situ hybridization, RNA probes were labeled with digoxigenin for gsc and with fluorescein isothiocyanate (FITC) for ntl.
Morphological changes in early development
The first cleavage occurred about 60 min post-fertilization (mpf) at 20°C. Thereafter, cleavages occurred about every 27 min until 310 mpf (Fig. 1). Cells divided synchronously in a single blastodisc until the 10th cleavage (Fig. 1), however, the synchronism was slightly different between the marginal part and the animal pole of the blastodisc after approximately the 8th cleavage. The adjacent blastomeres showed a synchronous cell cycle. After 320 mpf, the cells divided asynchronously, even in the adjacent blastomeres of a single blastodisc (Fig. 2).
Blastomeres of the marginal part of the blastodisc connected to the cytoplasm of yolk cell before 4 hpf. At 5 hpf, nuclei in this region started to proliferate mitotically without cytoplasmic division, forming the external yolk syncytial layer (E-YSL). At the center of the bottom of the blastodisc, a multinucleate cytoplasmic layer was not observed until 5 hpf. Some blastomeres in this region were stained more deeply with hematoxylin than those of the overlaying deep blastomeres. After 6 hpf, syncytial cytoplasm was observed in the central part of the bottom of the blastoderm, and thus the internal yolk syncytial layer (I-YSL) was formed. Thereafter, the blastoderm was separated from yolk by the YSL.
After 7 hpf, the shape of the embryo changed from ellipsoidal (Fig. 3A) to spherical (Fig. 3B), afterwhich epiboly began. The I-YSL began to bulge toward the animal pole, and the marginal part moved toward the vegetal pole of the yolk cell during epiboly. At 8 hpf, the blastoderm formed a dome-like shape by bulging of the I-YSL toward the animal pole (Fig. 3C). In some embryos, the blastoderm covered approximately 30% of the yolk cell with nearly uniform thickness (Fig. 3D).
After 9 hpf, the blastoderm covered half of the animal hemisphere of the yolk cell (Fig. 4A, B and C). At 10 hpf, the marginal region of the blastoderm thickened and formed a germ ring (Fig. 4D, E and F). Almost simultaneously, the embryonic shield was formed (Fig. 4G, H and I). The exact time of embryonic shield formation was not determined. During these events, epiboly was temporarily arrested. After the shield formation, epiboly continued. At 15 hpf, the blastoderm covered the yolk cell completely.
We observed that a difference in developmental rate arose among siblings during embryonic development. It was already observed during synchronous cleavages, while clearly at 10 hpf. In this time, several embryos had already finished the formation of the embryonic shield, whereas others had not. Thus, at 15 hpf, some embryos did not achieve 100% epiboly. These differences in developmental rates among siblings tended to increase with the progress of embryonic development.
gsc and ntl expression during development
gsc and ntl transcripts were first detected at a part of the marginal blastoderm at 6 hpf by WISH (Fig. 5A and B). gsc showed graded expression on one side of the embryo (Fig. 5A). Thereafter, at 8 hpf, gsc expression was localized to part of the blastoderm margin (Fig. 5C), but ntl expression expanded over the entire blastoderm margin (Fig. 5D). During gastrulation, expressing regions of gsc and ntl were separated. However, gsc- and ntl-expressing regions overlapped at 9 hpf (data not shown). At 10 hpf, gsc-expressing regions were shifted toward the animal pole of the embryo, separating from ntl expressing regions (Fig. 5E). On the other hand, ntl-expressing regions showed no shift (Fig. 5E). Then, at 11 hpf, when gsc-expressing regions moved toward the animal poles, ntl-expressing regions thickened in the dorsal region (Fig. 5F). At 12 hpf, expression of ntl expanded to the presumptive notochord (Fig. 5G). Subsequently, gsc and ntl expressed independently without overlapping (Fig. 5G). Finally, at 15 hpf, i.e. 100% epiboly, the leading edge of the gsc-expressing region reached the animal pole, and ntl was expressed in the presumptive notochordal region (Fig. 5H).
The developmental process in the loach was already reported by Watanabe et al. (1948). They observed the development from zygote to hatch, however they did not describe particular characteristics of the embryos from cleavage to the gastrula period in detail. In teleosts, characteristic indexes of these periods have been determined only in two species, zebrafish (Kimmel et al., 1995) and goldfish (Yamaha et al., 1999). Consequentially, the period before gastrulation can be divided into two periods, the synchronous cleavage period and the late-blastula period, by the MBT. After this period, embryos enter the gastrula period when characteristic germ layers are formed by morphogenetic movements. According to the observations of the present study, developmental schedules from cleavage to gastrula in the loach M. anguillicaudatus are defined as the following periods and summarized in Table 1.
Developmental stages of loach embryos from 0 ∼ 15 h at 20°C.
The synchronous cleavage period and the MBT:
Early embryonic development is characterized by rapid and synchronous cleavage to proliferate cell number. This stage corresponded to the period from 1 cell to 1k cells in the loach. Synchronous cleavage was repeated for 9 cycles about every 27 min interval after the first cleavage, thus the first asynchronous cleavage occurred at the 11th cell cycle. Thus, we determined that the MBT occurred at 320 mpf after the 1k-cell stage, because different cell cycles were observed even in the adjacent blastomeres of a single blastoderm. The MBT is characterized by loss of synchronism in cell divisions. In zebrafish and goldfish, asynchronous cleavage begins at the 10th cell cycle (Kane and Kimmel, 1993; Yamaha et al., 1999). The number of synchronous cleavages is regulated by the nucleo-cytoplasmic ratio (Newport and Kirschner, 1982; Kane and Kimmel, 1993). The loach, zebrafish and goldfish taxonomically belong to the same order, Cypriniformes. However, the relative DNA content of the loach was about two-thirds of the other two species (loach; Zhang and Arai, 1996, zebrafish and goldfish; Ciudad et al., 2002). Therefore, it seems that the number of synchronous cleavages in loach embryos might require an increase of one cell cycle to achieve the threshold nucleocytoplasmic ratio at the MBT, as compared with zebrafish and goldfish. Furthermore, it was reported that the E-YSL was formed roughly at the time of the MBT in zebrafish (Kimmel et al., 1995; Kimmel and Law, 1985), Fundulus (Trinkaus, 1992) and goldfish (Yamaha et al., 1999). Similarly, in the loach the E-YSL was formed at around MBT.
In the loach, this period corresponded to the time from 6 to 9 hpf. In zebrafish, the expression of gsc as a dorsal mesodermal signal and ntl as a panmesodermal signal was first detected at the late-blastula stage (Stachel et al., 1993; Schulte-Merker et al., 1992; Schulte-Merker et al., 1994). In the expression of these genes, the E-YSL plays an important role in the induction of the blastoderm for mesendoderm (Mizuno et al., 1996; Chen and Kimelman, 2000; Ober and Schulte, 1999). In the loach, as the expression of mesodermal marker genes gsc and ntl was detected after 6 hpf, we determined that the onset of the late-blastula period occurred at this time. The YSL consists of two regions, the E-YSL and the I-YSL. The YSL plays an important role in epiboly (Solnica-Krezel and Driever, 1994). During epiboly, the E-YSL spreads along the yolk surface toward the vegetal pole with the enveloping layer attached to it (Trinkaus, 1984), while the I-YSL bulges toward the animal pole (Topczewski and Solnica-Krezel, 1999). In zebrafish (Kimmel et al., 1995) and goldfish (Yamaha et al., 1999), it has been reported that the I-YSL is formed after E-YSL formation. Also in the loach, the I-YSL was histologically observed at 6 hpf after E-YSL formation.
The late-blastula period in zebrafish is subdivided to ‘oblong’, ‘sphere’, ‘dome’ and ‘30% epiboly’ periods by external appearance (Warga and Kimmel, 1990). The same morphological features as zebrafish were also observed in the loach. Therefore, we applied these terms to the subdivided late-blastula period of the loach. At 6 hpf, loach embryos resembled zebrafish embryos in the ‘oblong’ stage in external appearance and the start of marker gene expression. Thus, we decided that loach embryos in 6 hpf corresponded to the ‘oblong’ stage (Fig. 3A). In the same way, most embryos at 7 hpf were staged ‘sphere’ (Fig. 3B). In zebrafish, ‘dome’ and ‘30% epiboly’ stages are characterized by bulging of the I-YSL surface and the beginning of epiboly (Warga and Kimmel, 1990) and ntl expression in the blastoderm margin (Schulte-Merker et al., 1992). Since these features were apparently observed in the loach embryos at 8 hpf, we concluded that this time was the ‘dome’ or ‘30% epiboly’ stage. However, we could not determine an accurate ‘dome’ or ‘30% epiboly’ stage because these stages simultaneously occurred at 8 hpf.
Gastrula was staged between 9 and 15 hpf. The beginning of movements to form the inner layer of cells, so called involution, defined the onset of gastrulation. Subsequently, formation of the germ ring and the embryonic shield was recognized as the external characteristic of the gastrula. With regards to zebrafish, gastrulation occurs at 50% epiboly. The formation of the germ ring and the embryonic shield of goldfish (Yamaha et al., 1999), medaka (Iwamatsu, 1994), Fundulus (Armstrong and Child, 1965) and salmonid fishes (Ballard, 1973) occurs before 50% epiboly. In the loach, gastrulation occurred at 9 hpf on 50% epiboly. The germ ring and embryonic shield were formed within 1 h after 50% epiboly. Epiboly continued after the temporal arrest of embryonic shield formation. Epiboly advanced at a rate of 10% of the yolk cell each hour. Finally, epiboly ended when the blastoderm completely covered the yolk plug. Thus, we defined 100% epiboly at 15 hpf. However, asynchrony of development was observed in all developmental stages. This asynchrony became more obvious as time passed. This phenomenon has been observed in zebrafish (Kimmel et al., 1995) and goldfish (Yamaha et al., 1999). It is thought that this problem could be explained by slight differences in incubations condition or developmental abilities of individual embryos.
The course of morphological change and gene expression in the development of the loach resembled those observed in zebrafish, though developmental speed and incubation temperature were different between the two species. Thus, the knowledge of the developmental stages of the zebrafish may applicable to the loach.
In this study, we defined developmental indexes of the early embryonic stages in the loach based on morphological features and gene expression. This information advances the biotechnology of the loach, as embryos can be used at their optimum development stage for manipulation. It was shown that the developmental fate of the blastomere has not been determined before gastrulation in zebrafish (Ho and Kimmel, 1993). In the loach, the period from 5 to 8 hpf seems to be suitable for blastoderm manipulation. For further progress of the biotechnology of the loach, it is necessary to clarify the germ cell lineage to pick up primordial germ cells selectively, as well as to clarify the pluripotency of embryonic cells in several developmental stages to use donor cells.
We wish to thank Mr. Sotozaki and Mr. Shoji, the Loach Farming Cooperation of Kitamura, for the supply of loach samples. We also thank the members of the Laboratory of Breeding Sciences, Graduate School of Fisheries Science, Hokkaido University and Nanae Fresh Water Laboratory, Field Science Center for Northern Biosphere, Hokkaido University for their help. This study was supported in part by Grants-in-Aid from the Ministry of Education, Science and Culture of Japan (No. 14360105), the Akiyama Foundation in 1999 to E.Y. and that in 2000 to K.A., and Showa Shell Sekiyu Foundation for Promotion of Environmental Research in 2001 to E.Y.