Simple and efficient transgenesis with I-SceI meganuclease in the newt, Cynops pyrrhogaster



Newts have been recognized as an ideal model for body-parts regeneration after traumatic injury since the 18th century. However, molecular mechanisms underlying regeneration remain a mystery because of technical limitations. In the current study, to break this obstacle, we established a simple and efficient transgenic protocol for the newt Cynops pyrrhogaster by adapting an I-SceI microinjection technique, as well as a two-aquarium-tank (TAT) system that allows us to constantly obtain fertilized eggs in the laboratory for transgenesis. Following our protocol, ∼20% of injected embryos would exhibit non-mosaic widespread transgene expression and survive beyond metamorphosis. This anticipated success rate is about 10 times higher than that obtained by previous protocols, reaching a practical level. Therefore, our transgenic protocol in conjunction with the TAT-system could provide a key technique to open the way to uncover the long mystery underlying body-parts regeneration of newts. Developmental Dynamics 239:3275–3284, 2010. © 2010 Wiley-Liss, Inc.


Newts are a subset of the family Salamandridae of urodele amphibians (Weisrock et al.,2006) and have been recognized as a useful model animal since the 18th century, contributing to the progress of various research fields in biology and medicine (Colucci,1891; Wolff,1895; Spemann and Mangold,1924; Taguchi et al.,1989; Kurahashi,1990; Okada,1991; Kikuyama et al.,1995; Mitashov,1996). Especially for studies on tissue repair/regeneration after a traumatic injury, newts are indispensable animals because of their unique and outstanding ability: they can regenerate, even in adulthood, missing body-parts such as eye tissues including the retina and lens, and a part of the brain and heart as well as jaws, limbs, and tail, through a mechanism of transdifferentiation during which terminally differentiated somatic cells convert into multipotent stem-like cells (Goss,1969; Brockes and Kumar,2002; Tsonis and Del Rio-Tsonis,2004; Chiba and Mitashov,2007). Unfortunately, however, the molecular mechanisms underlying such intriguing phenomena remain to be elucidated, since, in this animal, the development of tools and techniques to manipulate gene functions lag behind that for other model organisms.

So far, to analyze gene functions in vivo, researchers have tried to deliver molecular tools such as plasmid vectors (for gain of function) or antisense oligonucleotides (for loss of function) into cells in living animals by in situ transfection using chemical reagents (Madhavan et al.,2006), electroporation (Kumar et al.,2007), or particle bombardment (Pecorino et al.,1996), or by implantation of cells that have been transfected in culture (Hayashi et al.,2001; Chiba et al.,2004; Grogg et al.,2005). However, these methods have some limitations such that it is difficult to deliver molecular tools to a target cell population uniformly. Moreover, the effects of such tools are not stable, sometimes causing inconsistent consequences. On the other hand, transgenesis can be a hopeful technique to overcome such weakness and allow us to regulate gene functions at a given time and place in vivo.

To our knowledge, transgenesis of newts has been tried with a Japanese common newt (fire bellied newt; Cynops pyrrhogaster; see Fig. 1). Makita et al. (1995) reported for the first time transgenic newts produced by microinjection of DNA constructs into fertilized eggs. However, unfortunately, those larvae showed highly mosaic expression patterns of exogenous genes. To circumvent this problem, Ueda et al. (2005) applied a procedure established by Kroll and Amaya (1996) for Xenopus, i.e., transplantation of sperm nuclei carrying transgenes into unfertilized eggs, with some critical modifications, and thereby succeeded in producing transgenic newts exhibiting widespread non-mosaic expression of exogenous genes in embryos, swimming larvae, and juveniles after metamorphosis.

Figure 1.

Fire bellied newt Cynops pyrrhogaster in the mating season. A: Female. B: Male. The male characteristically has a purple tail with black spots.

However, in the meantime, efficiency of transgenesis after microinjection of DNA constructs into fertilized eggs was substantially improved in other model animals by the introduction of the I-SceI meganuclease technique established in a fish, medaka Oryzias latipes (Thermes et al.,2002). In this approach, I-SceI meganuclease, which is an intron-encoded homing endonuclease isolated from the yeast Saccharomyces cerevisiae (Jacquier and Dujon,1985), is co-injected with DNA constructs carrying a transgene cassette flanked by 18-bp I-SceI recognition sites. After microinjection, this enzyme facilitates integration of foreign genes into the host genome, allowing the generation of stable transgenic lines. Thus far, this approach has been applied successfully to several fish species [medaka (Thermes et al.,2002); stickleback Gasterosteus aculeatus (Hosemann et al.,2004); zebrafish Danio rerio (Grabher et al.,2004)], amphibians [axolotl Ambystoma mexicanum (Sobkow et al.,2006); Xenopus laevis (Pan et al.,2006); Xenopus tropicalis (Ogino et al., 2006)], and ascidians [Ciona savignyi (Deschet et al.,2003)].

If this approach were to also work in the newt, it could be an alternative way of newt transgenesis with higher efficiency and convenience. Therefore, in the current study, we attempted to apply the I-SceI meganuclease technique to the C. pyrrhogaster newt.


Two-Aquarium-Tank System Is Useful to Obtain Fertilized Eggs in the Laboratory

For successful transgenesis, it is critical to obtain fertilized eggs efficiently. However, in the C. pyrrhogaster newt, the quality and quantity of eggs/sperms are highly dependent on animals that are collected from the field (also pointed out in Ueda et al.,2005). Moreover, this species is seriously decreasing in number (press release in 2006, Ministry of Environment of Japan; php?serial=7849). Therefore, here we attempted to obtain fertilized eggs in the laboratory by setting up an aquarium system that allows the animals to grow and mate as in the fields.

We prepared the aquarium tank (60-cm width; 30-cm depth; 45-cm height) with filtration/circulation and heater/thermostatic systems, and placed them beside glass windows away from direct sunlight in an animal stock room kept at 18°C (Fig. 2A). Small rocks were placed in the tank to allow animals to hide and rest and the water was filled up to ∼15 cm in depth. The heater/thermostatic system was set at 14°C.

Figure 2.

Two-aquarium-tank system for egg collection. A: Aquarium-tank system. B: Schematic graphs showing the schedules of hormone gonadotropin injection (H) and collection of eggs using two-aquarium-tank system. The number of eggs collected in a tank declines in 2 weeks. Therefore, to obtain eggs constantly, two tanks were used alternately each 2 weeks. C: The number of eggs collected on the day of experiment (square) and the percentage of fertilized eggs in them. The averaged values calculated monthly are shown. Vertical bars indicate SEM. N, the number of rounds.

At the beginning of the mating season (November 2009), adult newts captured during the last mating season (November 2008 and April 2009) were transferred into the tank and kept at 14–18°C under natural light conditions through the windows. The number of females per tank was fixed at ∼30, and the male to female ratio was set at 1–2:3 because ratios of males lower than this range reduced males attractive for females, while higher ratios of males caused competition/disturbance between males, both leading to less successful mating. The filtration/circulation system was operated only to clean the tank (4–6 hr/day) so as to keep the water still for the majority of time. Under these conditions, animals displayed a characteristic courtship behavior followed by delivery of a spermatophore from male to female (Tsutsui,1931; Kikuyama and Toyoda,1999). From one month later (December 2009), every female was injected with a hormone gonadotropin every other day (see Experimental Procedures section). A few days after the first injection, females started to lay eggs on strips of plastic sheets. Females that laid abnormal eggs and dead animals were replaced with healthy ones.

The number of eggs collected from the same tank decreased gradually over 2 weeks. Therefore, we used two tanks, alternating them every 2 weeks (Fig. 2B); after we collected eggs from one tank (Tank-1) for 2 weeks, we allowed females there to recover without hormone treatment for the next 2 weeks during which time we collected eggs from the other tank (Tank-2). Using this two-tank system, we could obtain eggs continuously for longer than 7 months (typically, 60–80 eggs/round) (Fig. 2C). The fertilization ratio estimated by cleavage was higher than 70%. We used these fertilized eggs in the following experiments.

Optimal Microinjection Site Is Around the Animal Pole

Subsequently, as the groundwork for microinjection, we attempted to determine the best injection site by evaluating it with the survival rate until the swimming larvae (SL) stage (Fig. 3; for controls with no microinjection, see Supp. Tables S1 and S2, which are available online). Around the top (i.e., the animal pole) and side of the pigmented area of the fertilized egg (one-cell stage embryo) were tested because the intracellular space under this region contains the cytoplasm and pronucleus (Fig. 3A). A small amount (approximately 20 nl) of a tracer solution (0.01% phenol red in water) was injected through a glass micropipette whose tip was grinded to an angle of 35°s (Fig. 3B), and the injected embryos were reared at room temperature (22–24°C). When the solution was injected near the top of the pigmented area (Fig. 3C), ∼29% of embryos (n = 48) survived until the SL-stage, whereas embryos (n = 61) injected at the side of the pigmented area died before the neurula (N) stage, indicating that the microinjection site influences the survival of embryos (Fig. 3D). We also tested the non-pigmented region in the vegetal pole side, but injection was quite difficult because yolk plugged up the micropipette tip. Therefore, we decided to make microinjections near the top of the pigmented area.

Figure 3.

Optimization of the microinjection site. A: Schematic diagram showing the injection sites tested. Microinjection was tried around the top (a) and the side (b) of the pigmented hemisphere of a fertilized egg (one-cell embryo). B: The tip shape of a micropipette for injection. Inset is a magnified view of the tip. The scale on the ruler is 30 μm. C: A view of microinjection around the top of the egg. Scale bar = 1 mm. D: Effects of the site of microinjection on the survival of embryos. A tracer solution (20 nl) was injected around either the top or the side of the one-cell embryos. When the solution was injected near the top, 29.4% of embryos (n=48; N=2) survived until the swimming larvae (SL) stage, whereas embryos (n=61; N=2) injected at the side died before the neurula (N) stage. Note that the embryos after microinjection were reared at room temperature (22–24°C). C, cleavage; B, blastula; N, neurula; TB, tail-bud; SL, swimming larvae.

Incubation of I-SceI Meganuclease and DNA Constructs Before Microinjection Is Critical for Non-Mosaic Dispersed Expression of Exogenous Genes in Embryos

We tried to microinject a mixture of a reporter DNA construct pCAGGs-EGFP (Sce) and I-SceI meganuclease into one-cell-stage embryos. This construct (5.9 kb) contains the pCAGGs-EGFP cassette flanked by I-SceI recognition sites that express EGFP under the control of chicken β-actin promoter with CMV-IE, and has been applied successfully to axolotl transgenesis (Sobkow et al.,2006). We initially prepared a microinjection solution containing 0.1 μg/μl of the DNA construct, 1 unit (U)/μl of I-SceI enzyme, 0.01% phenol red, and 1 × I-SceI buffer, and injected 20 nl solution (i.e., 2 ng DNA and 0.02 U I-SceI) into the one-cell embryos, and then reared them at room temperature (22–24°C). Under this condition, most embryos started expression of EGFP from the blastula (B) stage (Fig. 4A, B). We examined expression patterns at the B-stage (Table 1), and found that 11.4% of survivors expressed EGFP uniformly (Fig. 4A); however, 54.1% exhibited mosaic patterns (Fig. 4C). For expression categories, see the Data Analysis section in the Experimental Procedures section.

Figure 4.

Transgenic newts generated by the I-SceI method with pCAGGs-EGFP(Sce) construct. A, D–G: Embryos showing an intense green fluorescence uniformly at blastula (A), neurula (D, neural plate stage; E, neural tube stage), tail-bud (F), and swimming larvae stage (G). B: Bright-light image of the same field in A. C: A blastula embryo showing a mosaic expression pattern (arrowhead). H: A magnified view of the head region of the larva in G. I–L: A juvenile newt after metamorphosis [anesthetized in 0.1% FA100 (Tanabe, Japan) for 8–10 min]. Dorsal (I, J) and ventral (K, L) views are shown. I, K: Fluorescence image; J, L: Bright-light image. For movies of the larva in G, H and the juvenile in I–L, see Supp. Movies S1 and S7, S8. Scale bars = 1 mm (A–F), 3 mm (G), 1.2 mm (H), 1 cm (I–L).

Table 1. EGFP Expression Patterns at the Blastula Stagea
   EGFP expression patterns
Injection volume (nl/egg)DNA-I-SceI incubation time (min)No. survivors at blastula stageUniformMosaic (%)ND (%)
Strong (%)Moderate Weak (%)
  • a

    Injection solution contained the DNA construct (0.1 μg/μl) and I-SceI enzyme (1 U/μl). Embryos after microinjection were reared at room temperature (22–24°C). Survival rate of embryos under these conditions is shown in Table 2. ND, no detection.


To improve this result, we incubated the injection solution at 37°C before microinjection, since insufficient enzymatic reaction of I-SceI meganuclease to DNA constructs has been suggested to be a cause of such mosaic expression patterns (Ogino et al., 2006). Expectedly, when the solution was incubated for 40 min, B-stage survivors showing uniform expression patterns increased to 38.4% (more than three-fold) [strong expression: 13.8% (8.6-fold); moderate/weak expression: 24.6% (2.5-fold)], while those showing mosaic patterns or no green fluorescence decreased (Table 1). We tested different incubation times (15, 30, 40, 60 min) on the expression pattern. Finally, a 40-min incubation yielded optimal results to obtain embryos with uniform expression patterns.

Subsequently, we examined the influence of DNA-I-SceI concentration per embryo on the expression pattern. We injected 4 nl of the same solution (i.e., 0.4 ng DNA and 0.004 U I-SceI), that was pre-incubated at 37°C for 40 min, into one-cell embryos. Under this condition, B-stage survivors showing uniform expression patterns further increased to 55.4%, while those showing mosaic patterns or no green fluorescence decreased (Table 1). In conclusion, this condition was better than before, although the % survivors showing uniform expression patterns with strong green fluorescence seemed not to be affected.

Rearing Temperature After Microinjection Significantly Influences the Survival of Embryos

For practical transgenic protocols, high levels of survival rate of embryos after microinjection are required as are uniform expression patterns. Unfortunately, under the conditions we examined so far, most embryos died before the N-stage (Table 2). As a cause of death, we suspected mechanical damage itself in the microinjection because similar phenomena were observed even after microinjection of the tracer only (Fig. 3D). Therefore, to overcome this serious problem, we subsequently explored optimal rearing conditions that would allow the injured embryos to recover. Here, to evaluate effects of parameter changes on mechanical damage, we used one-cell embryos that underwent only a poke with the tip of the micropipette (but not injection).

Table 2. Survival Rate of Embryos After Microinjectiona
Injection volume (nl/egg)DNA- I-SceI incubation time (min)Developmental stage
nC (%)B (%)N (%)TB (%)SL (%)
  • a

    Injection solution contained the DNA construct (0.1 μg/μl) and I-SceI enzyme (1 U/μl). Embryos after microinjection were reared at room temperature (22–24°C). n, total number of eggs (one-cell embryos) examined; C, cleavage; B, blastula; N, neurula; TB, tail-bud; SL, swimming larvae.

  • b

    One of them showed a moderate and uniform expression pattern of EGFP, and the other was mosaic.


We examined the influence of rearing temperature on the survival of embryos. For that, we reared the poked embryos at lower temperatures (18° or 14°C). Surprisingly, under these conditions, most embryos survived beyond the N-stage (Table 3). When the poked embryos were reared at 14°C, the survival rate at the SL-stage dramatically improved to 81.8%; the value was higher than that (55.3%) at 18°C and close to the control value (88.5%) with no poke (Supp. Table S2). Subsequently, to examine the time for recovery, we changed the time during which the poked embryos were incubated at 14°C (Fig. 5). The survival rate at the SL-stage tended to increase by extending the incubation time (Jonckheere-Terpstra test: P = 0.0004). The mean value was significantly increased with an incubation time longer than 24 hr [24 h, 61.7 ± 4.4, N = 3, n = 33; ∞ (until stage 35; Supp. Fig. S1), 82.0 ± 0.0, N = 3, n = 33] compared to that with 1-hr incubation (Kruskal-Wallis test followed by Fisher's LSD test: 24 h, P = 0.018; ∞, P = 0.001). These results suggest that the poked embryos recover from mechanical damage in temperature- and time-dependent manners. In conclusion, after microinjection rearing embryos at 14°C until stage 35 could yield optimal results. After stage 35, we could rear animals at room temperature (22–24°C).

Table 3. Influence of Rearing Temperature on the Survival of Poked Embryos
Rearing temperature (°C)Developmental stage
nC (%)B (%)N (%)TB (%)SL (%)
  • a

    Data obtained by tracer injection (circles in Fig. 2D) were arranged for comparison. For abbreviations, see Table 2.

Figure 5.

Optimization of the time allowing the embryos to recover from mechanical damage by microinjection. One-cell embryos were reared for different periods [1, 3, 9, 24 hr, and ∞ (up to stage 35)] at 14°C after a poke with the micropipette tip. Percentage of survivors at the swimming larvae (SL) stage significantly increased as the time for recovery extended. Values are the mean ± SEM (1 h, N = 3, n = 48; 3 h, N = 3, n = 28; 9 h, N = 3, n = 33; 24 h, N = 3, n = 33; ∞, N = 3, n = 33). Statistical differences from the data of 1-hr incubation, examined by the Kruskal-Wallis test followed by Fisher's LSD test, are indicated by *P<0.05 and **P<0.01.

Optimal Conditions for Transgenesis With I-SceI in the Newt

Subsequently, we tried microinjection again in conjunction with the optimal rearing temperature: we injected 4 nl of the microinjection solution (i.e., 0.4 ng DNA and 0.004 U I-SceI), which was pre-incubated at 37°C for 40 min, into one-cell embryos, and reared them at 14°C until the SL-stage. This rearing temperature obviously improved the survival of embryos after microinjection more than at room temperature, i.e., 22–24°C (Table 4); Survival rate: 47.1% (3.8-fold) at the N-stage, 39.7% (7.1-fold) at the TB-stage, 35.3% (18.6-fold) at the SL-stage. Under these conditions, we successfully obtained transgenic juveniles after metamorphosis as well as embryos at any stage; these showed intense green fluorescence in the whole-body evenly (Fig. 4D–L, and see Supp. Movie S1–9). PCR analysis with genomic DNAs isolated from single swimming larvae confirmed that the pCAGGs-EGFP cassette is inserted into the genome (Fig. 6). We explored the optimal DNA-I-SceI concentrations again and finally found that 0.4–0.8 ng DNA/egg and 0.004 U I-SceI/egg could yield optimal results (Table 5). Using 0.4 ng DNA/egg and 0.004 U I-SceI/egg, the survival rate at the SL-stage was 35.3%, and the ratio of animals exhibiting uniform EGFP expression at the SL-stage was 70.8% (19.8% of injected embryos). Using 0.8 ng DNA/egg with the same I-SceI concentration, more than 30% of the animals showing uniform expression patterns at the SL-stage (asterisk in Table 5) exhibited the most intense green florescence, although the survival rate was slightly decreased (22.7% at the SL-stage).

Figure 6.

Integration of the pCAGGs-EGFP cassette into the genome of transgenic newts. PCR analysis was carried out using genomic DNAs isolated from single swimming larvae with the Wizard Genomic DNA Purification Kit (Promega, Madison, WI). PCR primers (forward: 5′-atggt gagcaagggcgaggagctg-3′; reverse: 5′-cttgta cagctcgtccatgccgag-3′) were designed to amplify the EGFP gene (717 bp). We detected EGFP in a transgenic newt exhibiting moderate green fluorescence with a non-mosaic widespread pattern (lane T), but not in a wild type animal (lane C). The same results were obtained from two other such green larval newts. M, size marker; P, positive control with pCAGGs-EGFP(Sce) plasmid.

Table 4. Improvement of the Survival of Injected Embryos at 14°C Rearing Temperaturea
Rearing temperature (°C)Developmental stage
nC (%)B (%)N (%)TB (%)SL (%)
  • a

    Injection volume: 4 nl/egg (I-SceI: 0.004 U/egg: DNA: 0.4 ng/egg); DNA-I-SceI incubation time: 40 min.

  • b

    Data in Table 2 were listed for comparison. For abbreviations, see Table 2.

Table 5. EGFP Expression Patterns at Swimming Larvae Stage When Reared at 14°C
    EGFP expression patterns
I-SceI (U/egg)DNA (ng/egg)No. cleavaged eggsNo. survivors at swimming larvae stageUniformMosaic (%)ND (%)
Strong (%)Moderate Weak (%)
  • a

    Every larva showed very strong green fluorescence. Injection volume: 4 nl/egg; DNA- I-SceI incubation time: 40 min. ND, no detection.


In conclusion, we successfully adapted the I-SceI technique for transgenesis of the newt Cynops pyrrhogaster. The optimized procedures are summarized as a flow chart in Figure 7. The two-aquarium-tank system is effective to obtain fertilized eggs constantly in the laboratory (Step 1). Incubation of a mixture of the DNA construct [26–52 fmol (∼0.1–0.2 μg)/μl] and I-SceI enzyme (1 U/μl) at 37°C for 40 min before microinjection is necessary to obtain animals showing uniform and strong expression of transgenes (Step 3). Microinjection should be made around the animal pole of one-cell embryos (Step 5). The optimal volume of the microinjection solution per embryo is 4 nl (i.e., 0.4–0.8 ng DNA and 0.004 U I-SceI). Rearing the embryos after microinjection at 14°C significantly improves their survival (Step 5–6).

Figure 7.

A protocol for I-SceI tramsgenesis in the newt.

Following our protocol, ∼20% of injected embryos would exhibit non-mosaic widespread transgene expression and survive beyond metamorphosis. This anticipated success rate is about 10-times higher than that (0.9–3.4%) estimated in the previous transgenic protocol (Ueda et al.,2005), reaching near the values in I-SceI transgenesis in Xenopus (laevis, 20%; tropicalis, 30%; Ogino et al., 2006) and axolotl (32.7%; Sobkow et al.,2006). Biologists and surgeons have looked to newts as an ideal model for body-parts regeneration after traumatic injury since the 18th century, but the underlying molecular mechanisms had remained a mystery because of technical limitations. Our simple and efficient transgenic protocol provides a key technique to break this obstacle, opening the way to uncover this long-sought mystery.



We used sexually mature adult newts (total body length: male, ∼9 cm; female, 11–12 cm), which were collected in a restricted area (Kamogawa, Chiba, Japan) by Mr. Kazuo Ohuchi (Misato, Saitama, Japan) during the mating season (November 2008 and April 2009) and kept in plastic containers at ∼18°C as described previously (Chiba et al.,2006) until the start of this study (November 2009). It must be noted that during this period female newts underwent artificial spawning by gonadotropin injection (see below) and almost exhausted their eggs and sperm stock; later than July 2009, the total number of eggs released by one female had obviously decreased (<30) compared to that (100–200) in April, and the ratio of eggs that developed normally to swimming larvae had drastically declined (<10%). In November, as we recognized signs of a mating season, i.e., appearance of a purple color on the tail of males and a bulging abdomen of females (Fig. 1), we transferred those animals into the aquarium tanks for this study. These animals were fed with frozen fly larvae (Akamushi; Kyorin Co., Ltd., Japan) every day. From December, females were injected subcutaneously in the abdominal region with 50 μl (30 U) of gonadotropin (Gonatropin 3000; Asuka Seiyaku, Japan) every other day.

The original research reported herein was performed under the guidelines established by the University of Tsukuba animal use and care committee.


Solutions for rearing newt embryos were based on 1× Holtfreter's (H) solution, which contained (g/l) 3.5 NaCl, 0.05 KCl, 0.1 CaCl2, and 0.2 MgCl2·6H2O. For diluted H solution, pH was adjusted to ∼7.5 with NaHCO3, otherwise as mentioned. All solutions for dejellied eggs/embryos (see below) were sterilized through a syringe filter of 0.2-μm pore size (DISMIC-25cs, Cellulose Acetate; Advantec, Japan) and stored at 4°C.

Microinjection solutions contained 1× I-SceI buffer (New England Biolabs, Ipswich, MA), 0.25–1 U/μl of I-SceI meganuclease enzyme (New England Biolabs), 0.01% phenol red (stock: 0.1% phenol red dissolved with 0.3M NaOH), and 0.01–0.2 μg/μl of a plasmid DNA construct. In the current study, we used pCAGGs-EGFP (Sce) as a reporter construct (kindly provided by Dr. Elly M. Tanaka, Max Planck Institute of Molecular Cell Biology and Genetics, Leipzig, Germany; see Sobkow et al.,2006). The plasmid DNA was purified using Endofree Plasmid Maxi Kit (Qiagen, Valencia, CA), recovered in nuclease-free water, and stored at −80°C. I-SceI enzyme was aliquoted and stored at −80°C. We prepared a fresh injection solution in each experiment; solutions were incubated at 37°C for 0–60 min and placed on ice until use.

Preparation of One-Cell-Stage Embryos

The night before the day of the experiment (2–3 times/week), strips of plastic sheets that were tied into a knot were placed into the water in the aquarium tank to allow females to lay eggs on (otherwise, females hardly spawned or ate their own eggs). Fertilized eggs enclosed in a jelly capsule (“jellied” eggs) were collected early in the morning (6–8 AM) since those were mostly (∼85%) the one-cell-stage embryos. The jellied eggs were immersed in 70% ethanol for 1 min, rinsed three times with 0.5× H solution, and then treated with 2% Na-thioglycolate (194-03557; Wako, Osaka, Japan) containing 0.5× H solution (pH = 9–10) to remove the jelly capsule. Dejellied eggs were immediately transferred to 0.5× H solution with a plastic pipette, and then carefully rinsed (with no shaking) in 0.5× H solution 10 times to remove any excess Na-thioglycolate and jelly debris. These eggs were placed, animal pole side up, onto the bottom of the wells (3.5-mm diameter, 1.6-mm depth) on Terasaki plates (96-wells, sterile; Watson, Kobe, Japan) that were filled with 0.5× H solution containing 6% Ficoll PM 400 (F4375; Sigma-Aldrich, St. Louis, MO) and 1× penicillin-streptomycin (pen-strep, Cat. no. 15140; Gibco, Gaithersburg, MD). These plates were kept on ice until microinjection.

Microinjection and Embryo Rearing

Microinjection was carried out using micropipettes (outer tip ∅: 10–15 μm; inner tip ∅: 2–3.5 μm) made from glass capillaries (Model G-1, outer ∅: 1 mm; inner ∅: 0.6 mm, length: 90 mm; Narishige, Tokyo, Japan) by a micropipette vertical puller (PC-10; Narishige); micropipette tips were ground to a ∼35° angle using a micro grinder (EG-400; Narishige). The micropipette was set to a holder that was fixed to a motorized micromanipulator (MP-330; Narishige) beside a fluorescence stereomicroscope (Leica M165 FC); the holder was connected to an injector PV820 Pneumatic Picopump (World Precision Instruments, Sarasota, FL). Immediately before injection, a 2-μl drop of the injection solution was placed on a parafilm sheet (Pechiney Plastic Packaging, Chicago, IL) mounted on ice and front filled into the micropipette using an aspirator connected to the injector. After the dejellied eggs (one-cell embryos) on a Terasaki plate were placed under the stereomicroscope, the tip of the micropipette was inserted into the one-cell embryo up to ∼50 μm in depth, and a 4–30-nl solution was injected by a pressure of 4-6 PSI (pounds per square inch) for 80–300 msec. The injection volume, which was controlled by changing the duration of injection, was estimated according to the World Precision Instruments operation manual. After all embryos on Terasaki plates were injected using the same micropipette in ∼30 min, they were transferred to an incubator (CN-25C; Mitsubishi Engineering, Japan/M-200; TAITEC, Japan) and kept at 14–24°C for 24 hr. Thereafter, the embryos were gathered in a 0.8% agarose (Standard 01; Solana, EU)-coated glass Petri dish (35 mm in diameter) filled with 0.5× H solution containing 1× pen-strep, and reared at 14–24°C. Ruptured embryos with a leakage of yolk were removed. Solutions were exchanged with fresh ones by sterilized plastic pipettes immediately after removal of those embryos and every other day.

Developmental stages were determined according to the criteria of Okada and Ichikawa (1947). Here we defined embryos at stage 9–10, 17–21, 22–27, and 40–42 as blastula, neurula, tail-bud, and swimming larvae, respectively. At stage 22, tail-bud embryos were transferred to 0.1× H solution containing 1× pen-strep in agarose-coated Petri dishes. At stage 35, when forelimb buds grow and become obvious (Supp. Fig. S1), larvae were separated to wells (35 mm in diameter) filled with 0.1× H solution on 6-well plates (Falcon 3502; Becton Dickinson, Franklin Lakes, NJ). Swimming larvae grown up to ∼3 cm were then transferred to plastic cups (bottom ∅: 6–8 cm; lid ∅: 8–10 cm; hight: 4 cm; 100–200 ml). After metamorphosis, juveniles were reared in 0.1× H solution shallow enough to wet their abdomen and feet (Supp. Movie S7). Larvae and juveniles were fed with living brine shrimp or frozen fly larvae (Akamushi) a few times a week. The dishes or cups were exchanged to clean ones 12 hr after feeding.

Data Analysis

Bright-light and fluorescence images/movies were acquired using a digital video camera (C-5060; Olympus, Japan) attached onto a fluorescence stereomicroscope (Leica M165 FC, Exton, PA) with a filter set for GFP (Leica GFP-Plant; excitation filter, 470/40 nm; dichromatic beam splitter, 495 nm; barrier filter, 525/50 nml). In the current study, we divided embryos/larvae showing visible green fluorescence under the stereomicroscope into two groups, “uniform” and “mosaic.” The definition of uniform is as follows: (1) fluorescence is observed throughout the body; (2) the fluorescence intensity of cells in the body is almost even, except for pigment cells, which appear around stage 35 and absorb fluorescence. For larvae, fluorescence in the internal tissues or organs was confirmed through their transparent skin; for example, see a beating heart (Supp. Movie S5), digestive organs (Supp. Movie S6), and blood cells flowing and the lens in the eye (Supp. Movie S9). In fact, tissues or organs dissected from those animals exhibited intense green fluorescence [for example, see Supp. Fig. S2 for the stomach and the intestine (endoderm origin); see Supp. Fig. S3 for the bone and the muscle (mesoderm origin)]. The definition of mosaic is as follows: (1) fluorescence is observed only in parts or regions of the body; (2) in embryos exhibiting widespread expression patterns, the fluorescence intensity of cells in the body is obviously different. We further divided the uniform group into “very strong,” “strong,” and “moderate/weak,” on the basis of their fluorescence intensity; the fluorescence intensity was estimated as the averaged luminance by analyzing the fluorescence images of embryos/larvae that were acquired with a 40× objective lens, using a function of Photoshop Extended CS5 graphics software (Adobe, San Jose, CA). Embryos/larvae with the averaged luminance value of 100–120 were categorized as “strong,” and those with higher and lower values were as categorized as “very strong” and “moderate/weak,” respectively. Figures were prepared using the same graphics software. Image brightness, contrast, and sharpness were adjusted. Statistical data in the text were the mean ± SEM. N was the number of rounds. n was the total number of eggs (one-cell embryos) examined. Statistical significance was determined by non-parametric tests.


We express our gratitude to Dr. Elly M. Tanaka for providing us the pCAGGs-EGFP (Sce) construct. This work was supported by a Grant-in-Aid for Challenging Exploratory Research (20650060) and a Grant-in-Aid for Scientific Research (B) (21300150) from the Japan Society for the Promotion of Science (JSPS).