Introduction

Ascidians such as Halocynthia roretzi, Ciona intestinalis, and Phallusia mammillata have long been used as model organisms for studying embryonic development and its underlying regulatory mechanisms (McDougall et al. 2011; Satoh 1994, 2013). Due to the transparent nature of the embryo and its relatively small cell number, Conklin (1905a) used the fertilized oocytes of S. clava to track the regularity of cell division based on observations of the specific distribution and allocation of cytoplasmic pigmentum during embryogenesis, and established a cell lineage for embryonic development. Later studies found special mosaic developmental patterns in solitary ascidians such as Halocynthia roretzi and Ciona robusta (Ciona intestinalis type A) (Nishida 2005). In addition, the manipulation of embryos such as the construction of ascidian libraries and transgenes (Corbo et al. 1997), genome assembly at the chromosomal level and transcriptome profiling during development (Dehal et al. 2002; Satou et al. 2019; Wei et al. 2020), gene editing (Christiaen et al. 2009; Sasaki et al. 2014; Small et al. 2007; Stolfi et al. 2014) and other technology platforms, improved the capacity for embryo manipulation at the genetic level. Considering the special evolutionary status (Delsuc et al. 2006) and the comprehensive database (Brozovic et al. 2016; Tassy et al. 2010) of ascadians, research using these organisms can provide insights into the origin of chordate organs, and how vertebrates evolved from the common ancestor of chordates.

Styela clava is classified as a member of the class Ascidiacea, order Stolidobranchia, and family Styelidae. It was originally described by Herdman (1881) from dredged specimens in the Sea of Okhotsk (Abbott and Johnson 1972). Styela clava is a tunicate that shows global environmental adaptation (Dupont et al. 2009; Goldstien et al. 2010; Locke et al. 2007) because of its invasiveness, which may negatively affect native species and cause serious damage to commercial aquaculture such as scallop farming (Davis and Davis 2010). Adults can be collected all year round because of the high tolerance of S. clava to a wide range of temperatures (Cinar 2016).

Styela clava adults have multiple gonads, compared to a single gonad in C. robusta species. There is no outer covering of follicle cells on the eggs of S. clava whereas such a covering is present on the eggs of Ciona spp. (Bhattachan et al. 2020). The eggs of S. clava are up to 200 microns in diameter with a tough chorion. The fertilized egg needs around 10 h to develop into a hatched larva which acquires complete swimming ability instantly. After a swimming period of about 2.5 h post hatching (hph), adhesion will occur. In contrast, adhesion occurs after about 4.5 h ‘ swimming period in C. robusta (Matsunobu and Sasakura 2015). At around 3 hph, S. clava larvae have completed the swimming larval stage and the rapid tail regression process is initiated. Although there are significant structural differences between S. clava embryos and other ascidians (Dardaillon et al. 2020; Guignard et al. 2020; Hotta et al. 2007, 2020), detailed images of developmental stages are still needed to better identify the differences between species.

In this study, we constructed a standard three-dimensional (3D) developmental table and collected confocal laser scanning microscopy (CLSM) images of embryogenes in S. clava by annotating the morphology at each defined stage sequentially at 23 °C. We also provided a continuous time-lapse movie with standard developmental stages and time duration. Morphometrical information about the length of the tail-to-trunk ratio is also described for each stage. We deposited the raw data on the Resources of Ascidian Morphology Network-based (RAMNe) at https://chordate.bpni.bio.keio.ac.jp/RAMNe/latest/index_styela.html, thereby providing a comprehensive record of the spatiotemporal morphological changes that occur during the early developmental processes of S. clava. This enables easy access to 3D angles of each embryo and slices of different layers, and facilitates systematic comparisons of the morphology and development process of C. robusta and A. aspersa at the same stage, which is valuable for studying chordate development. In addition, we developed an easy procedure to remove the chorion from fertilized eggs by the chemical-washing method and applied tissue-specific fluorescent labeling in S. clava embryos.

Results

Developmental stages of S. clava

We observed and defined 22 stages of embryogenesis and larval development in S. clava, from egg fertilization to hatching of larvae, based on morphology and developmental events. For each stage, we provide corresponding differential interference contrast (DIC) images of embryos with a chorion (Supplementary Fig. S1), DIC images of embryos without a chorion (Supplementary Fig. S2), and 3D reconstruction images of embryos without a chorion using Phalloidin 488 staining (Fig. 1). Embryogenesis was partitioned into six periods: zygote, cleavage, gastrula, neurula, tailbud, and larva.

Fig. 1
figure 1

Three-dimensional reconstructed images of Styela clava embryos in the developmental time course after fertilization. Fertilized eggs were dechorionated and incubated at 23 °C. F-actin of embryos were stained with Alexa Fluor™488 Phalloidin. Stage 1 is in the zygote period. Stages 2–9 are in the cleavage period. Stages 10–13 are in the gastrula period. Stages 14–16 are in the neurula period. Stages 17–21 are in the tailbud period. Stage 22 is in the larva period. In stages 1–13, the anterior side of each embryo is upward, and in stages 14–22, the anterior end is on the left. The criteria for each stage are shown in Table 3. Scale bar: 50 μm

Full size image

The criteria for staging of the zygote to cleavage period (stages 1–10) depend on the embryo shape and the quantity of cells. After stage 10, the criteria are based primarily on embryo morphology, the tail-to-trunk ratio (recorded after stage 15), and distinct developmental event features referencing those previously described for C. robusta (Hotta et al. 2007).

The “standard development time” (in hours) was established to represent the normalized hours post-fertilization (hpf) at 23 °C by converted staging information. The optimal incubation temperatures were between 16 and 23 °C, based on our observations and previously published research (Cinar 2016). The hatching times, which were determined by time-lapse imaging of chorionic embryos co-cultured with dechorionic embryos at different temperatures, were 10 ± 0.50 h at 23 °C, 12 ± 0.38 h at 20 °C, 17 ± 0.48 h at 18 °C, and 20 ± 0.43 h at 16 °C (n = 5, 5, 5, 5 at different temperatures), respectively.

Zygote period (0–0.8 h post-fertilization (hpf) at 23 °C, stage 1)

The zygote period (0–0.8 hpf at 23 °C) consisted of only one stage, stage 1, which lasted from fertilization to the end of the first mitotic cell division (Fig. 2). The initial mitotic division is shown in Fig. 2B, B’(optical section). Although narrowed, the cleavage furrow between the two daughter cells is not yet fully formed. The egg diameter was 180 ± 5 µm (n = 10).

Fig. 2
figure 2

Styela clava embryos at the zygote stage and for most of the cleavage period. Embryos are stained with Alexa Fluor™488 Phalloidin. Zygote period stage 1 (A–B’). 3D reconstructed images of embryos (A, B) and cross-sections (A’, B’). A Showing one cell. B Showing the first cell division. Cleavage period stage 2 to stage 8 (C–K’). C–E Showing a 3D reconstructed image without dorsoventral differentiation. 3D reconstructed images in dorsal view (F–K) and ventral veiw (F’–K’), respectively. Each blastomere was identified by the number of cell lineages; two headed arrows designate a pair of daughter cells. Anterior side of each embryo is upward unless otherwise specified. A, anterior, D, dorsal, P, posterior, and V, ventral, to show the orientation of the embryo. Scale bar: 50 μm

Full size image

Cleavage period (0.8–3 hpf, stages 2–9)

The cleavage period (0.8–3.0 hpf at 23 °C) consisted of eight stages, i.e., stages 2–9 (Figs. 2C, 3A). The cleavage stage was characterized by a sequence of mitotic divisions that culminated in the formation of numerous blastomeres. The embryos underwent seven divisions during this period and were bilaterally symmetrical. At the 16-cell stage, unequal cell division (UCD) occurred resulting in the formation of two smaller cells (Hibino et al. 1998). The asynchronous vegetative half split more quickly than the animal half following this stage, forming 32-, 44-, and 64-cell stages (McDougall et al. 2019). The developmental stages during this period were defined both by cell quantity and cell morphology. After each division of the blastomere contraction process (Supplementary Movies 1, 2), the morphological changes were particularly obvious before and after the contraction of 16-cell and 32-cell stages. These two stages were further separated into two sub-stages.

Fig. 3
figure 3

Styela clava embryos at the cleavage and gastrula periods. Embryos were stained with Alexa Fluor™488 Phalloidin. The cleavage period stage 9 (A–A”) and the gastrula period stages 10–13 (B–E”). 3D reconstructed images of embryos in dorsal view (A–E), ventral view (A’–E’), and longitudinal section (A”–E”). A, B Each blastomere is indicated with the number of cell lineages; two headed arrows designate a pair of daughter cells. A Marked with obvious depressions (arrows). Anterior end of each embryo is upward in A–C’ and on the left in D–E’. A”–E” Showing the orientation of the embryo, A, anterior, D, dorsal, P, posterior, and V, ventral. The positions of midline sections are indicated by dashed lines in A”–E”. b.p., blastopore. Scale bar: 50 μm

Full size image

Stage 2

Two-cell stage (0.8 hpf, Fig. 2C). The first cleavage plane is determined by the embryo dividing into the left and right halves.

Stage 3

Four-cell stage (1.2 hpf; Fig. 2D). The second cleavage plane is determined by the embryo dividing into anterior and posterior halves. Following cell division, each blastomere was initially arranged loosely with a gap in the center (Fig. 2D). Eventually, the blastomeres compacted and the gap was filled in.

Stage 4

Eight-cell stage (1.4 hpf; Fig. 2E). The embryo underwent a third cell division to separate the vegetal plane of the cell from the animal plane. These two planes were at a slight slanted angle and were not orthogonal. We used the nomenclature established by Conklin (1905b) to label the four founder cell lineages of S. clava as follows: a, animal anterior; b, animal posterior; A, vegetal anterior; B, vegetal posterior. Each blastomere in live embryos might be recognized based on its size (Fig. 2E, B4.2 blastomere is the largest). Eventually, embryo compaction occurred so that it resembled a sphere (Supplementary Movie 1, T = 1 h 17 min).

Stage 5

16-cell stage (1.6 hpf; Fig. 2F–G). All cells in both the vegetal and animal poles underwent a fourth cleavage. The embryo's distinct anterior–posterior polarity was apparent first and the B5.2 blastomere was much smaller than the others. In C. robusta and Phallusia, the occurrence of these phenomena is caused by the effect of the centrosome-attracting body (CAB) (Dumollard et al. 2017; Negishi et al. 2007). However, in S. clava, we were unable to observe this structure clearly using phalloidin staining. This stage was divided into two parts: an early, uncompacted 16-cell stage (stage 5a, Fig. 2F–F’) and a late, compacted 16-cell stage (stage 5b, Fig. 2G–G’).

Stage 6

32-cell stage (2.0 hpf; Fig. 2H–I). All blastomeres underwent a fifth cell division, which happened earlier in the vegetal than the animal lineages. The B5.2 division was once again asymmetric, resulting in a tiny posterior B6.3 cell. In C. robusta, this was caused by the CAB’s effect, which was not observed in S. clava. This stage was divided into two parts: an early, uncompacted 32-cell stage which was almost spherical (stage 6a, Fig. 2H–H’) and a late, compacted 32-cell stage which was flatter and had a bulging B6.2 cell (stage 6b, Fig. 2I–I’).

Stage 7

44-cell stage (2.3 hpf; Fig. 2J). Only vegetal blastomeres underwent a sixth cell division during this stage. Bulging in vegetal blastomeres was observed.

Stage 8

64-cell stage (2.6 hpf; Fig. 2K). The animal blastomeres completed a sixth cell division. Due to the projection of the A7.8 and B7.4 blastomeres, the embryo appeared more angular than round when viewed from the animal pole.

Stage 9

76-cell stage (2.8 hpf; Fig. 3A). The vegetal half of the embryo had an asymmetrical pattern of cell division. Compared to animal cells, the vegetal blastomeres were taller and more columnar (Fig. 3A’’). The vegetal side of the embryo became flattened and did not begin to sink until gastrulation was initiated. This stage was the last one prior to gastrulation.

Gastrula period (3–4 hpf, stages 10–13)

The gastrula period (3.0–4.0 hpf at 23 °C) consisted of four stages: stages 10–13 (Fig. 3B–E”). Gastrulation occurred at three stages between the sixth and seventh cleavages of the ascidian embryo. The initial stage is endoderm invagination, followed by mesoderm involution, and ultimately ectoderm epiboly (Swalla 1993).

Stage 10

Initial gastrula (3.0 hpf; Fig. 3B–B’’). Gastrulation began with A7.1 blastomere apical constriction. The invagination in the middle of the vegetal pole was obvious in longitudinal sections (Fig. 3B”, arrowhead).

Stage 11

Early gastrula stage (3.2 hpf; Fig. 3C–C”). The embryo developed into a cup-shaped structure as the endoderm cells progressively invaginated. When viewed from the vegetal aspect, the embryo was horseshoe-shaped.

Stage 12

Mid-gastrula stage (3.6 hpf; Fig. 3D–D”). The embryo's blastopore was still open in the middle. Six rows of cells—three rows of a-line cells (upper) and three rows of A-line cells (lower)—were arranged in a highly regular pattern on the flat neural plate.

Stage 13

Late gastrula stage (4.0 hpf; Fig. 3E–E”). The blastopore of the embryo was almost closed at the posterior end. The embryo elongated anteriorly. Six or more rows of the neural plate formed, and the first two A-line neural rows (I and II near the blastopore) began to curve (start of neurulation).

Neurula period (4–4.8 hpf, stages 14–16)

The neurula period (4.0–4.8 hpf at 23 °C) consisted of three stages: stages 14–16 (Fig. 4A–C”). Neurulation is the transformation of a flat neuroectodermal sheet called the neural plate into a closed and extended neural tube (Nicol and Meinertzhagen 1988a, b). The neural tube then splits from the epidermis, forming the rudiments of the nervous system (Hashimoto et al. 2015).

Fig. 4
figure 4

Styela clava embryos at the neurula and tailbud periods. Embryos were stained with Alexa Fluor™488 Phalloidin. The neurula period is stages 14–16 (A–C”) and the tailbud period is stage 17 (d–D”). 3D reconstructed images of embryos in dorsal view (A–E), cross-section (A’–D’), and longitudinal section (A”–D”). Anterior end of each embryo is on the left (A–D’). A”–D” Showing the orientation of the embryo, A, anterior, D, dorsal, P, posterior, and V, ventral. The anterior edge of neural tube closure is indicated by arrowheads. The positions of midline sections are indicated by dashed lines in (A”–D”). en, endoderm; epi, epidermis; mu, muscle; noto, notochord; n.p., neuropore. Scale bar: 50 μm

Full size image

Stage 14

Early neurula stage (4.0 hpf; Fig. 4A–A”). The a-line neural plate initially bent to form a furrow bordered by raised neural folds. The blastopore was completely closed. Zippering of neural tubes started from the posterior region of the embryo (Fig. 4A).

Stage 15

Mid-neurula stage (4.3 hpf; Fig. 4B–B”). The neural tube remained open. The neural plate of the A-line cells also caused a neural fold and the embryo was oval in shape (Fig. 4B).

Stage 16

Late neurula stage (4.6 hpf; Fig. 4C–C”). The neural tube closure commenced in the posterior area (Fig. 4C, arrowhead). The embryo elongated as the notochord precursors converged and intercalated.

Tailbud period (4.8–10 hpf, stages 17–21)

The tailbud period (4.8–10 hpf at 23 °C) consisted of five stages: stages 17–21 (Figs. 4D–D”, 5A–D”’). The tailbud period was separated into four canonical stages: initial tailbud, early tailbud, mid-tailbud, and late tailbud. The late tailbud stage was divided into two sub-stages due to the prolonged development time and the significant morphological changes.

Fig. 5
figure 5

Styela clava embryos at the tailbud period. Embryos were stained with Alexa Fluor™488 Phalloidin. The tailbud period is stages 18–21 (A–C”). 3D reconstructed images of embryos in lateral view (A–E) and longitudinal section (B’–E’, A”–E”). A’ Cross-section, indicating incomplete notochord cell intercalation. B’ is a shallower lateral section than B’’. Enlarged longitudinal section of the trunk (D’) and tail (D’’, D”’): D’’ is a shallower lateral section than D’’’. Anterior end of each embryo is on the left. Embryo orientation is shown by A, anterior; D, dorsal; P, posterior; V, ventral; mu, muscle; noto, notochord. Scale bar: 50 μm

Full size image

Stage 17

Initial tailbud stage (4.8 hpf; Fig. 4D–D”). Separation between trunk and tail regions was present in embryos. The trunk and tail were equal in length. The neuropore migrated anteriorly, and the neural tube closure in the posterior area was nearly complete. Two rows of notochord cells were arranged in the left–right direction, some of which were still interdigitating and in the process of intercalation.

Stage 18

Early tailbud stage (5.2 hpf; Fig. 5A–A”). The neuropore had closed at the top of the trunk. Separation between tail and trunk regions was more distinct and the tail was significantly longer than the trunk (Tail/Trunk ratio = 1.2). Neurulation, and the intercalation of the few notochord cells that were positioned most anteriorly, were completed.

Stage 19

Mid-tailbud stage (5.9 hpf; Fig. 5B–B”). The length of the tail was 1.8 times that of the trunk (Table 3). The intercalation of notochord cells was completed. Viewed from the lateral aspect, the tail contained three rows of muscles on each side, with each medio-lateral line having seven muscles and each lateral line having four muscles. In general, a total of 36 muscle cells and 40 notochord cells were found in S. clava (see section mode of RAMNe).

Stage 20

Late tailbud stage I (7.0 hpf; Fig. 5C–C”). Pigmentation of the otoliths could be observed under an anatomical microscope (Supplementary Fig. S2). The length of the tail was 2.3 times longer that of the trunk.

Stage 21

Late tailbud stage II (8.0 hpf; Fig. 5D–D”). The length of the tail was four times that of the trunk. The epidermal cells in front of the trunk developed in a columnar fashion but did not protrude to form into palps. The notochord and muscle cells elongated along the A-P axis. Immediately prior to hatching, tail twitching was occasionally observed.

Larva period (10 hpf, stage 22)

Stage 22

Hatching larva (10.0 hpf at 23 °C, Fig. 6). The trunk grew longer and became more rectangular. The otoliths and ocellus could be observed under an anatomical microscope (Supplementary Fig. S3). Only one siphon primordium (either atrial or oral siphon primordium) was observed in the posterior region of the trunk.

Fig. 6
figure 6

Styela clava embryos at the hatching larva period. Embryos were stained with Alexa Fluor™488 Phalloidin. The hatching larva period is stage 22. A 3D reconstructed images of embryos in lateral view, showing the types of epidermal cell. B Longitudinal sections of trunk. A’’ Enlargement of trunk surface, revealing a siphon primordium. A’’’ Longitudinal sections of the tail. B Frontal views of the trunk in 3D reconstructed images, showing no protruding palps. Embryo orientation is shown by A, anterior; D, dorsal; P, posterior; V, ventral; sp, siphon primordium; epi, epidermis; mu, muscle; noto, notochord; ot, otolith; Scale bar: 50 μm

Full size image

The unique processes in early development of S. clava

We observed sereval unique organogenesis events during early development of S. clava, including the processes of notochord formation, the disappearance of palps, and tail regression.

In ascidians, notochord tissue is the most significant tissue localized at the midline region of the tail. It provides mechanochemical signals and serves as a hydrostatic skeleton during embryogenesis and the hatching larva stage. The notochord organogenesis process can be separated into three main phases in C. robusta embryogenesis (Lu et al. 2019): cell intercalation, which leads to the notochord forming a single row of cells; cell elongation, which lengthens the notochord along the A-P axis; and tubulogenesis, which provides more flexibility and rigidity to notochord tissue (Mizotani et al. 2018; Peng et al. 2020; Zhao et al. 2021). Ascidiella aspersa also seems to go through all three stages based on the available data (Funakoshi et al. 2021). In contrast, we observed only the first two phases during embryogenesis in S. clava. The intercalation of notochord cells was completed during the initial tailbud to early tailbud stage. After the mid-tailbud stage, coin-like notochord cells were elongated along the A–P axis to form drum-shaped cells (Fig. 7A). However, no lumen structure was detected in the notochord after cell elongation, even at the stage of tail regression initiation (Fig. 7A, hatching larva). These findings are not consistent with a previous study which reported luminization in S. clava (Jiang and Smith 2007). We also found that the embryos and larvae of S. clava did not form palps. In C. robusta and A. aspersa (Funakoshi et al. 2021), the anterior-most ectoderm cells thicken and protrude into three palps at the front of the nerve plate. The palps are involved in larval attachment and metamorphosis, and function in both chemo- and mechano-sensation (Wagner et al. 2014; Wakai et al. 2021). From the late tailbud stage to the hatching larva stage, we performed a 3D reconstruction of the trunk and confirmed that no protruding structures are present on the top of the trunk in S. clava embryos and larvae (Fig. 7B, yellow arrowhead). However, a thickening of the anterior ectoderm cells was detected from the late tailbud II stage (Fig. 5D’, arrowhead).

Fig. 7
figure 7

Comparison of developmental processes in Styela clava, Ciona robusta, and Ascidiella aspersa. Embryos were stained with Alexa Fluor™488 Phalloidin and Alexa Fluor™555 WGA (ThermoFisher, W32464). A Longitudinal sections and corresponding enlarged views of embryos in different stages (from left to right: early tailbud, mid-tailbud, and hatching larva stage). B 3D reconstructed images of half-frontal views of the trunk in three stages showing no palps were formed (arrows). B’ 3D reconstructed images of frontal view of the trunk at hatching larva stage (left: S. clava; right: C. robusta) showing no palps were formed in S. clava, whereas three palps are visible in C. robusta (indicated by arrows). C Longitudinal sections of the trunk in tail regression stage (left: S. clava; right: C. robusta). The notochord cells of S. clava are tubular with WGA signals on the inside (red), and the muscle cells are curled with only phalloidin signals (green). In contrast, the notochord and muscular cells of C. robusta are coiled and stacked in the trunk, forming a structure known as the internal coiling tissue (IT). en, endoderm; mu, muscle; noto, notochord

Full size image

Another obviously distinct developmental characteristic of S. clava larvae is tail regression. The tail regression process appears to differ greatly among ascidians (Cloney 1982). The time duration of tail regression and how the tail tissue is arranged in the trunk are the two key differences. Styela clava larval tail regression started at 3 hph and took only about 15 min to complete at 23 °C (from tail regression initiation to tail completely regressed into the trunk), while C. robusta larval tail regression starts at 12 hph and takes more than one hour to complete at 18 °C (Matsunobu and Sasakura 2015). Furthermore, the arrangement of the regressed-tail tissue in the trunk differs significantly between S. clava and C. robusta. The notochord cells of S. clava were concentrated in the posterior part of the trunk part and were completely seperated from other regressed-tail tissues; muscle cells rapidly buckled into a spherical shape and were scattered along the edge of the posterior part of the trunk; the trunk tissue was compressed into the anterior part of the trunk (Fig. 7C, left). In contrast, the internal tissues of C. robusta’s tail were stacked in a coiled form, called coiled internal tissues (IT) (Yamaji et al. 2020), and the muscle and notochord tissues were spirally arranged in the posterior part of the trunk with no obvious cell buckling or tissue reconstruction (Fig. 7C, right).

The present study also revealed notable differences among S. clava, C. robusta, and A. aspersa. We found only one siphon primordium (either atrial or oral) at the posterior end of the trunk in S. clava larvae (Fig. 6A, A”), whereas there are two atrial siphon primordia in C. intestinalis and A. aspersa larvae (Funakoshi et al. 2021; Hotta et al. 2007). We also compared the ratio of the tail-to-trunk length among the three species and found that C. robusta had the highest, followed by S. clava, and A. aspersa (Table 4). Among these three ascidians, a decrease in the ratio of tail-to-trunk length was observed only in S. clava. This occurred at the hatching larva stage and was caused by an increase in trunk length. Moreover, we also compared the number of lateral epidermal cells in the tail of the three species and found that C. robusta had the highest number, i.e., ~ 60 cells (Hotta et al. 2007), A. aspersa had the lowest, i.e., ~ 40 cells (Funakoshi et al. 2021), while S. clava has a middle number, i.e., ~ 50 cells. Furthermore, from stage 18 to stage 20, there was no significant ventral bending of the tail in dechorionated S. clava embryos (Fig. 5A–C), whereas such bending is present in both C. robusta and A. aspersa embryos (Funakoshi et al. 2021; Hotta et al. 2007; Kogure et al. 2022; Lu et al. 2020). It is noteworthy, however, that dechorionation might have had an effect on the developmental potential of embryos at the tailbud stages.

Resources of S. clava morphology for network-based research

We integrated the raw data of S. clava into a web-based database, “RAMNe.” The images were exported as a series of files to enable easy viewing of cross-sectional and 3D images using a web browser (Fig. 8, https://chordate.bpni.bio.keio.ac.jp/RAMNe/latest/index_styela.html). At the web interface of the developmental table, images were linked to information on developmental stages, developmental nomenclature, hour post-fertilization (hpf), % hatch, cell lineage, and time-lapse movies. Additionally, 3D reconstruction images, cross-sectional images, and DIC images of the embryos are also displayed (Fig. 8A). Users can interactively view 3D images (Fig. 8B) and z-section images (Fig. 8C) online by selecting any developmental stage from the table. Furthermore, the database also displays corresponding images from C. robusta and A. aspersa (Funakoshi et al. 2021; Hotta et al. 2007, 2020), which can easily be compared with S. clava (Fig. 8D). The detailed usage of the RAMNe database has been described previously (Funakoshi et al. 2021).

Fig. 8
figure 8

Screenshot of the database. To use the database, please refer to the website: https://chordate.bpni.bio.keio.ac.jp/RAMNe/latest/index_styela.html. A “Developmental table” refers to the two-dimensional images from a fertilized egg to the hatched larva as viewed by CLSM using an inverted microscope. B Information about section images. C Information about 3D images. D “Dual display mode” can display and compare the images of S. clava to C. robusta and A. aspersa (Funakoshi et al. 2021; Hotta et al. 2007, 2020) interactively

Full size image

The transgene method optimization and protocols

The eggs of S. clava are difficult to dechorionate due to their thick and dense chorions. These structures primarily comprise covered cylindrical box-like follicular cells on the outside and test cells dispersed irregularly on the inside, similar to Styela plicata eggs (Villa and Patricolo 2000). To establish the ideal dechorionation method, we first tested enzymes (chitosan, chitinase, and NaClO) and other components of the dechorionation solution used in C. robusta, such as altering the amounts of NaOH and sodium thioglycolate (see Table 1). Our results showed that the dechorionation solution with chitosan or chitinase cannot remove the chorion within 30 min; using NaClO caused destruction of the oocyte; and increasing the concentration of NaOH and sodium thioglycolate did not significantly reduce dechorionation duration, but the embryos were more susceptible to malformation. The optimum dechorionation solution for S. clava was found to be 1% sodium thioglycolate, 0.05% proteinase E, 55 µL 10 N NaOH, and 10 mL filtered seawater (FSW). Compared with the dechorionation method for C. robusta embryos, extra NaOH (55 μL) needed to be added to the dechorionation solution, while the concentration of protease E and sodium thioglycolate remained unchanged. Furthermore, a longer dechorionation time duration (18 min) was necessary to totally remove the chorion for S. clava embryos. Based on the above experimental results, we developed a standard protocol for S. clava embryo dechorionation. The feritilized eggs were immersed in the dechorionation solution (Fig. 9A, step 1) and, by gently pepetting for around 18 min, the chorion can be removed (Fig. 9A, step 2). Dechorionated eggs were then washed five times with FSW and transfered into an agarose-coated dish. To avoid harm and aberrant development due to the dechorionation solution, the dechorionation process must be completed within 18 min. The dechorionated eggs could either be used for the subsequent electroporation process or incubated for morphological observation.

Table 1 The effects of different dechorionation solution components on chorion removal in S. clava
Full size table
Fig. 9
figure 9

Styela clava muscle and epidermal cells were labeled by electroporating expression clones with distinct fluorescent protein tags. A, B An illustration of the process for dechorionation (A) and electroporation (B). C Two types of tissue-specific promoters fusing with fluorescent gene vectors. D Larva inducing the Sc-Ma1 > EGFP vector expressing green fluorescent protein (GFP) in muscle cells. E Larva inducing the Sc-Epi1 > tdTomato vector labeled with tdTomato in epidermal cells. D'–E’ Showing merged images with bright field and fluorescence microscopy. Anterior is to the left. Scale bars: 50 μm

Full size image

Subsequently, we explored the optimum transgene technology in S. clava by electroporation. By using the Gene Pulser Xcell System (BIO-RAD, Hercules, USA), we optimized the electroporation-related parameters, including electroporation voltage, capacitance, and plasmid concentration. The initial voltage value was calculated from the size of the egg, and the specific formula was as follows (Zeller 2018):

$${E}_{c}=\frac{{V}_{c}}{0.75\times {d}_{\mathrm{cell}}},$$

where \({E}_{c}\) is the critical field strength in volts per centimeter, \({V}_{c}\) is the membrane permeation voltage (1 at 20 °C) and \({d}_{\mathrm{cell}}\) is the cell diameter in centimeters (Multiporator manual, www.eppendorf.com). The diameter of S. clava eggs was 180 ± 5 µm (n = 10) and, together with our experimental quantification on the survival rate after electroporation, the optimal voltage value was 45 V (see Table 2). After determining the voltage, we searched for the ideal value of capacitance. A previous study showed that higher capacitance values improved transfection but reduced viability (Zeller 2018). Capacitances ranging from 1500 to 3000 µF were investigated and it was found that 2000 and 2500 µF was suitable for electroporation. We also found that a higher concentration of plasmid DNA was necessary for high transgene efficiency. Thus, we established a standard protocal for S. clava embryo electroporation, which resulted in an electroporation efficiency of 37.2 ± 9.7%. In brief, dechorionated fertilized eggs (300 μL) were mixed with 80 μg plasmid DNA in a prefabricated electrotransfer solution (0.77 mol/L mannitol in ddH2O). Then, the mixture was transfered into 4 mm cuvettes with a Gene Pulser Xcell System (BIO-RAD) and electroporated with expontential decay pulse. Electroporated eggs were recovered for around 10 min, washed once, transferred into an agarose-coated dish (Fig. 9B) and incubated at 18 °C.

Table 2 The effects of different voltage and capacitance on electroporation
Full size table

Labeling cells with a tissue-specific driver in S. clava embryos

We next attempted to introduce exogenous plasmids into fertilized eggs of S. clava. We designed a tissue-specific fluorescent fusion construct to specifically label the tissue cells. The upstream 3 kb of the C. robusta homologous genes MA1 and Epi1 (Chiba et al. 1998) were used as the driver to force a GFP or tdTomato fluorescent reporter to be expressed in specific tissue cells (Fig. 9C). After electroporation of these constructs, the results showed that the Sc-Ma1 > eGFP plasmid was expressed in tail muscle cells (Fig. 9D), whereas the Sc-Epi1 > tdTomato plasmid was expressed in the epidermal cells (Fig. 9E). Due to the mosaic expression in ascidians, the constructs were only expressed in part of the target tissue cells. The fluorescently labeled embryos could develp into larval stages in our experiments, by which the subcellular location of proteins could be visualized in a living embryo to track the dynamics of this protein.

Discussion

In this study, we developed a chemical method that can easily remove the chorionic membrane from the eggs of S. clava, allowing us to observe the early embryogenesis easily and offer the possibility to genetically manuplate the eggs. Based on this technique, we established a developmental atlas of S. clava and defined the developmental stages for this species for the first time. Combining CLSM, DIC, and time-lapse imaging, the internal and exterior structures of the embryo at each stage and time of development were meticulously documented. We accurately depicted the anatomy and developmental stages of S. clava and established a standard developmental table, which enables detailed comparison of tissue morphology and differentiation between wild-type embryos and knockout mutant embryos. The processes of embryogenesis and quantitative data of S. clava are summarized in Table 3, including the developmental duration, basic embryonic morphological characteristics, time ratio to hatch, and tail-to-trunk length ratios. We also compared the tail-to-trunk length ratios and tail length with C. robusta and A. aspersa (Table 4). Additionally, based on the mature electroporation technique in C. robusta, we applied the electroporation technique in S. clava for the first time. DNA constructs were successfully introduced into S. clava fertilized eggs. These new techniques, and the developmental atlas of S. clava embryos and larvae, provide a crucial basis for S. clava to be used as a marine model organism for reserach on developmental biology, evo-devo, ecology, and cell biology.

Table 3 Stages of early embryonic development of Styela clava at 23 °C
Full size table
Table 4 Tail/trunk ratio and Tail/trunk length of Styela clava, Ciona robusta and Ascidiella aspersa
Full size table

Styela clava is a potential model organism for studying morphological changes such as tail retraction during metamorphosis. Additionally, notochord development in S. clava embryos was straightforward during the tailbud period and did not involve vacuolation to generate the lumen, making it an ideal model for research on intercellular insertion and polarity establishment. However, it is noteworthy that dechorionation has been found to have a slight impact on morphogenesis (Oonuma et al. 2016; Kourakis et al. 2021). Nevertheless, based on the present findings, we expect chorionated specimens of S. clava to be used in future research on developmental processes.

It has been reported that the leathery sea squirt ascidian S. clava is an invasive species that shows high capacity for environmental adaptation and tolerance to temperature and salinity (Dupont et al. 2009; Goldstien et al. 2010; Locke et al. 2007). The results of studies on transcription factors and chromosome-level gene expression in S. clava have revealed a significant increase in the number of transposons and heat shock protein-related genes (Wei et al. 2020). These findings shed light on the organism's environmental adaptation mechanisms based on its genome and molecular networks. However, experimental validation and detailed molecular mechanisms are impossible without transgenic manuplation. Based on the dechornation and electroporation tenniques reported here, transgenic technologies, such as gene knockout and overexpression, could easily be applied in S. clava. This would significantly improve the operability of embryonic genes, shedding a powerful new light on the molecular mechanisms of the development of invasive ascidians and chordates.

Materials and methods

Animal collection and fertilization

Adults of S. clava (Herdman, 1881) were collected from Weihai City, China, and acclimated into seawater at 18 °C under constant light to accumulate gametes in the laboratory. The 18S rRNA gene was amplified and the PCR products were sequenced for species identification (Wei et al. 2020). The adult animals were dissected, and the mature eggs and sperm were isolated from the gonoducts from different individuals. A drop of 1 mol/L Tris pH 9.5 (Kobayashi and Satou 2018) was added to activate the sperm which was then mixed with eggs at room temperature for 15–20 min. Fertilized eggs were washed with seawater through a nylon filter to remove sperm and debris and incubated in FSW at 16 ~ 23 °C.

The dechorionation approach of S. clava

The chorion provides protection to eggs, promotes fertilization, and prevents self-fertilization and polyspermism (Villa and Patricolo 2000). Additionally, we found that eggs without the chorion cannot be fertilized. Therefore, unless needed for specific experimental purposes, we typically removed the chorion after fertilization. Fertilized eggs were washed with FSW to remove impurities and sperm, and were then transferred to a 60 mm 0.1% agarose-coated dish with dechorionation solution. The chorions can be adequately digested by the enzyme after 18 min of blowing the eggs evenly using the dropper. Eggs might be blown violently for the first 10 min, and then carefully and gently during the next 8 min. A dissecting microscope was required to examine the proportion of dechorionated eggs. When the chorions of the majority of eggs (~ 80%) were removed, the dechorionated eggs were concentrated by gently swirling the dish to gather them in the center. They were then transferred to an agarose-coated dish and washed with FSW five times to clean the residual dechorionation solution. The dechorionated eggs were used either for incubation or electrical transfer.

The electroporation process

Using a pipette, 300 μL of seawater with the required amount of dechorionated eggs was transferred into a 1.5 mL tube holding 80 μL of ddH2O, 80 μg plasmid DNA and 420 μL of electrotransfer solution (electroporation solution: 6.3 g mannitol, 5 mL FSW, and 45 mL ddH2O). The solution was then gently mixed and transferred into a cuvette, which was placed in a holder. After electroporation with suitable parameters (45 V voltage, 2000–2500 µF capacitance, and 4 mm electrode cuvette width), the sample was gently transferred into a 90 mm 0.1% agarose-coated dish with FSW and incubated for about 10 min. Gently swirlling the dish enabled the eggs to be concentrated in the center and transferred into a new dish. The embryos were then allowed to develop at 18 °C.

Construction of tissue-specific fluorescent labeling clones

Two types of tissue-specific genes of S. clava were obtained from homologous Epi1 and MA1 genes (Chiba et al. 1998) in Ciona. The upstream 3 kb promoters of these were amplified from the genome of S. clava by polymerase chain reaction (PCR) with Phanta® Max Super-Fidelity DNA Polymerase (Vazyme, P505-d1). The vectors without promoters were reverse-amplified from fluorescent protein clones pEGFP or tdTomato with an identical polymerase. After purification by electrophoresis, the promoter recombined with one fluorescent protein vector, i.e., pEGFP or ptdTomato, by homologous recombination (Vazyme ClonExpress® II One Step Cloning Kit, C112-01) to create a tissue-specific fluorescent labeling clone. The primers and templates used are presented in Supplementary Table S2. The expression clones were sequenced with the primers shown in Supplementary Table S2 and promoter transfection efficiency was verified by electroporation.

Time-lapse imaging by inverted microscope

Time-lapse imaging movies from zygote development to swimming larva after dechorionation were taken using differential interference contrast (DIC) microscopy (OLYMPUS IX73). The room temperature was maintained at 23 °C by an air conditioner and the images were acquired every 1–3 min.

Fixed embryo image collection by confocal laser scanning microscopy

Confocal images were taken with a ZEISS confocal laser scanning microscope (CLSM) equipped with 20X (Numerical aperture: 0.75) and 63X oil immersion objectives (NA: 1.4). Z-series images were taken at intervals of 0.8–1 µm, resulting in stacks of 80–100 images. Image analysis and 3D reconstruction were performed with ZEISS software packages. Some of the lengthier embryos were photographed integrally using the tile function. Adobe Photoshop and ImageJ were used to pseudocolor the images. Embryos were fixed every 0.5–1 h from the fertilization of the egg to the hatching larva stage. Distinct representative embryos in each stage (Fig. 1) were chosen based on the C. intestinalis staging criteria from CLSM data (Hotta et al. 2007). Because of the dark color of S. clava embryos, the whole embryo could not be scanned directly, and only the fluorescence signal from appromaxiately half the thickness of the embryo could be received. Therefore, the 3D reconstructed images displayed in the Results section were all half embryos, and the ventral and dorsal images of the same period were from different individuals. Cortical actin filaments were stained using Alexa Fluor™ 488 Phalloidin (ThermoFisher, A12379).