INTRODUCTION

Small fish species, the medaka (Oryzias latipes) and zebrafish (Danio rerio), have been receiving much attention as experimental animals in recent years. These species have many advantageous features for experiment: they are tough and easy to be maintained in small aquariums in a laboratory; they have a relatively short generation time and they frequently spawn eggs; the size of the eggs is large enough for manipulation of the embryos; and the developmental process is visible through the transparent chorion.

In addition, especially in medaka, it is possible to observe internal organs without killing fish, even in adult fish, because of a special “see-through” medaka strain lacking pigment through the lifecycle has been generated (Wakamatsu et al., 2001).

Recently, the combined use of the green fluorescent protein (GFP) and the transgenic technique is becoming popular in various organisms. Using the GFP gene as a reporter, the expression profiles of regulatory elements of interest can be visualized in live organisms without using additional chemicals and complicated treatments. It is also possible that tissues of interest can be labeled using the GFP gene linked to tissue specific promoters (Higashijima et al., 2000; Tanaka et al., 2001). The transparency of the chorion, embryos or adult fish of medaka offer a quite advantageous condition for various experiments using GFP. To introduce gene constructs containing the GFP gene into medaka embryos, microinjection has been used in most of the previous studies (Hamada et al., 1998; Kinoshita et al., 2000; Tanaka and Kinoshita, 2001; Tanaka et al., 2001). The gene transfer efficiency of this method is generally high, and it has been also indicated that introduced genes are transmitted to offspring through the germ-line (Kinoshita et al., 2000, 2001; Tanaka et al., 2001). However, the microinjection operation is performed under a microscope, in which skilled technique is required, and thus a limited number of embryos can be treated at one time. Thus, several other methods such as electroporation (Inoue et al., 1990; Murakami et al., 1994; Ono et al., 1997) and use of a retroviral vector (Lu et al., 1997) have been utilized to achieve foreign gene transfer. Recently, we proposed the particle gun method in high hopes for mass experiments (Yamauchi et al., 2000), in which the successful transfer of a foreign gene construct with high efficiency into medaka embryos and the inheritance of the introduced gene into the next generation were demonstrated. However, no information has been available on the fate of the foreign gene construct introduced by the particle gun and the stability of expression profile in later generations. Here we report on the stable inheritance of the foreign gene construct pβ-Act-GFP-N, containing a GFP gene linked to the medaka β-actin promoter, introduced by the particle gun as well as the common spacio-temporal expression profile between the F1 and F4 generations.

MATERIALS AND METHODS

Plasmid construction

Expression plasmid pβ-Act-GFP-N (Fig. 1), corresponding to pCMX-SAH/Y145F in Yamauchi et al. (2000) was prepared by the modification of pCMX-GFP-1 described by Ogawa and Umezono (1998). Namely, a 5′-upstream sequence of medaka beta actin gene reported by Takagi et al. (1994) was amplified by a polymerase chain reaction (PCR). The amplified fragment replaced the cytomegalovirus promoter sequence in pCMX-GFP-1. An additional sequence, which contains a SV40 nuclear localization signal (NLS, amino acid sequence of KKKRKV), was ligated to the 3′ end of a GFP gene.

Fig. 1

Structure of pβ-Act-GFP-N. A green fluorescence protein (GFP) gene was followed by a nuclear localization signal (NLS, amino acid sequence of KKKRKV). A PCR-amplified 5′-upstream sequence of medaka beta actin gene reported by Takagi et al. (1994) was located at the 5′-end of GFP gene.

RESULTS

Inheritance of the transgene to offspring

We generated four transgenic strains (strains A, B, C, and D) by the particle gun method in total as follows: 1691 fertilized eggs were treated with the particle gun, 75 were found to have fluorescence derived from the transgene at the hatching stage, 5 of them transmitted the fluorescence to offspring as reported previously (Yamauchi et al., 2000), and 4 stable strains were successfully established. The proportion of transgenics in the F1 generation was 3% for strain A and 5% for strain B, which suggest the mosaic distribution of the introduced transgene. In the F2 generation, the trans-gene was detected in approximately 50% of all siblings. The crossing of a transgenic male and a female of the F2 generation resulted in the Mendelian inheritance of the trans-gene to the F3 generation. By crossing transgenic F3 individuals homozygous for the transgene, all the F4 individuals became homozygous transgenic fish.

In the transgenic strain generated by microinjection, the proportion of embryos revealed that the fluorescence at the hatching stage was 40% (4 of 10 injected embryos) and one of them inherited the transgene to the F1 generation. The proportion of transgenics in F1 individuals was 23% (24 of 103 tested embryos), which also suggests the mosaic incorporation of the transgene. In this strain, the transgene was also transmitted from F1 to F2 by the back-crossing with nontransgenic fish, and F3 and F4 generations were obtained as in the four strains described above.

Fig. 2 shows the representative results of the PCR analysis for the existence of the transgene in the F4 generation. The GFP gene was clearly detected in F4 individuals derived from all four transgenic strains produced with the particle gun method, as well as those from a transgenic strain produced by the conventional microinjection method.

Fig. 2

Detection of inheritance of the GFP gene in the F4 generation. DNA was extracted from the tail fin of each transgenic strain of the F4 generation. For the detection of the GFP gene, a GFP gene-specific primer pair, which yielded an 814-bp fragment, was used. A medaka-EF-1α-A-gene-specific primer pair, which yielded a 519-bp fragment, was also used to confirm successful DNA extraction. P, introduced plasmid (pβ-Act-GFP-N); W, non-transgenic medaka; MI, transgenic strain produced by conventional microinjection; A-D, transgenic strain produced by the particle gun method.

GFP expression during embryogenesis

The pattern of expression in developing embryos and hatchlings was common among four transgenic strains generated by the particle gun method and one strain generated by microinjection. The expression pattern was also common between heterozygous F1 and homozygous F4 generations in all five strains. In transgenic embryos at a one-cell stage (stage 2b of Iwamatsu, 1994), fluorescence was not detected (Fig. 3 a-1 and a-3), although intense fluorescence of GFP was observed in some transgenic medaka strains; for example, the olvas-GFP medaka strain (Fig. 3 a-2 and a-4) (Tanaka et al., 2001). The fluorescence began to be observed ubiquitously in the embryonic body at the head forming stage (stage 17) (Fig. 3 b-1 and b-2). The ubiquitous expression of GFP was observed until the 12-somite stage (stage 23, Fig. 3 c-1 and c-2). After this stage, the fluorescence became intense only in the somite and lens. The fluorescence in other tissues, on the other hand, decreased gradually and became hardly detectable (Fig. 3 d-1 and d-2).

Fig. 3

Expression of the GFP gene in a β-Act-GFP-N transgenic fish produced by the particle gun method during embryonic development. F4 individuals of the transgenic strain A were examined by fluorescence microscopy. In a-2 and a-4, pictures of an embryo of olvas-GFP transgenic fish (Tanaka et al., 2001) were shown for comparison. At a one-cell stage, fluorescence of the GFP was not observed in an embryo of β-Act-GFP-N transgenic fish (a-3), while intense fluorescence was observed in olvas-GFP transgenic fish (a-4). The fluorescence was observed ubiquitously at the head forming stage (b-2). The ubiquitous expression of GFP was observed till the 12-somite-stage (c-2). The fluorescence became intense only in the somite and lens (d-2). Intense fluorescence was observed in the skeletal muscle in the trunk at the hatching stage (e-2). An arrowhead in e-3 shows one of the spots revealing intense fluorescence of GFP. Intense fluorescence was also observed in the lens and the muscle in jaw of hatchling (f-2). a-1, a-2, b-1, c-1, d-1, e-1, and f-1 are the bright field images; a-3, a-4, b-2, c-2, d-2, e-2, e-3, and f-2 are the fluorescent images. Yellow spots in the fluorescent images are the auto-fluorescence of pigment cells. Bars in a-1, e-1, and e-3 represent 0.25, 1, and 0.02 mm, respectively.

In hatchlings, intense fluorescence was observed in the skeletal muscle in the trunk (Fig. 3 e-1 and e-2), so that each somite was recognized clearly in the fluorescent image. The magnified image of the skeletal muscle revealed the spots (about 10 μm in length) of a more intense fluorescence of GFP (one of these spots was indicated by arrowhead in Fig. 3, e-3). Intense fluorescence was also observed in the lens (Fig. 3 e-2, f-2) and the muscle in the jaw (Fig. 3 f-1 and f-2).

GFP expression in adult tissues

The expression sites in adult fish of the F1 and F4 generations were also common among the four strains generated by the particle gun method and the one by microinjection. The external view of an adult transgenic individual of a strain generated by the particle gun was shown in Fig. 4 a-1 and a-2. Intense fluorescence was observed in the body of the trunk and lens. The intense fluorescence enables us to identify transgenic individuals more easily under the fluorescent microscope. The results of anatomical studies on adult fish were summarized in Table 1. The fluorescence was observed in the brain, gill, heart (Fig. 4)b, spleen, kidney, and ovary. The intensity of fluorescence in these tissues was considerably weaker than that of the skeletal muscle. In these tissues, however, all cells did not show green fluorescence. For example, only a small portion of the cells showed the fluorescence in the heart (Fig. 4b-1). The magnified image of the cells showing fluorescence in the heart (Fig. 4b-2) represented that the GFP was concentrated in the nucleus. We could not detect green fluorescence in the liver, gut, and testis using a stereomicroscope.

Fig. 4

Expression of the GFP gene in a β-Act-GFP-N transgenic fish produced by the particle gun method. F4 individuals (strain A) were examined by fluorescence microscopy. a, external appearance of adult individual; b-1, heart; b-2, magnified image of a GFP expressing cell indicated in b-1 by arrowhead. a-1 is a bright field image of an adult male fish. a-2, b-1, and b-2 are fluorescent images. Bars in a-1, b-1, and b-2 represent 5 mm, 300 mm, and 10 μm, respectively.

Table 1.

Expression of GFP in adult fish.

DISCUSSION

In this study, we successfully established four transgenic strains bearing the transgene pβ-Act-GFP-N containing the GFP gene controlled by a 5′-upstream sequence of the medaka β-actin gene. We demonstrated the Mendelian transmission of the transgene to the F4 generation. The expression patterns in embryos, hatchlings and adult fish were faithfully reproduced in each generation, which was the same as the patterns in the strain generated by micro-injection. Thus, we conclude that the transgene introduced by the new method, the particle gun, is stably transmitted to offspring as that introduced by the traditional microinjection technique. Considering that the procedure is much easier than microinjection, the particle gun method is one of the most promising methods for mass transfer experiments.

At present, on the other hand, there are some points needed to be improved. The frequency of germ-line-transmitting founders is one of the most important factors to select a gene delivery method for the production of a transgenic strain. In the case of the microinjection method, the frequency reported so far is 2% (3 germ-line-transmitting founders / 150 injected eggs, Kinoshita et al., 2000), 3% (13/410, Tanaka et al., 2001), and 10% (1/10, this study). On the other hand, the frequency using the particle gun method was 0.4% (2/564 at the optimum condition, Yamauchi et al., 2000), which was quite low compared with those of the microinjection method. Chimerism of germ cells in F0 founders is also a critical parameter for the selection of the gene delivery method. The chimerism observed in the particle gun method was relatively low (3% in strain A and 5% in strain B) compared with those observed in the microinjection method (23%, in this study and 5∼67%, Tanaka et al., 2001). The disadvantages of the particle gun method may be due to the following reasons: In the case of microinjection, DNA is perfectly delivered into the cytoplasm of a 1-cell-stage embryo. On the contrary, in the case of the particle gun method, some particles coated with DNA reach the cytoplasm and some stay in the chorion, perivitelline space, and/or the yolk sac. In other words, the total volume of DNA delivered into cytoplasm is not enough to elevate the frequency of germ-line-transmitting founders and chimerism up to those of the microinjection method. Another unknown feature of DNA may also reduce the frequency of the particle gun method. In the case of microinjection, the feature of DNA that is aqueous fluid makes it easy to disperse into cytoplasm. As a result, this dispersion elevates the number of cells in which the integration of the transgene may occur. In the case of the particle gun method, DNA is attached to the gold particle. Therefore, DNA must be dissolved into cytoplasmic fluid before dispersion, while cleavage is already ongoing. This delay may reduce the chance of integration of the transgene. The low frequency of germ-line-transmitting founders and chimerism requires laborious work to search a great number of individuals or eggs for the identification of germ-line-transmitting founders or transgenics. We think that the combination of a gene, which we want to introduce into medaka, with a reliable selective marker gene, for example olvas-GFP as a germ-cell-transmitter marker (Tanaka et al., 2001), helps us to identify transgenics easily and cancels the disadvantages of the particle gun method.

Since the expression pattern was common among the four strains and the one strain generated by microinjection, it is evident that the expression pattern reflects the activity of regulatory elements contained in the plasmid pβ-ActGFP-N. In the developmental process, the expression of pβ-Act-GFP-N became detectable at the head forming stage in this study. This result is in contrast with the case of the transgenic medaka bearing the GFP gene driven by a ubiquitous promoter EF-1α, in which the GFP expression was already detectable at the early gastrula stage (Kinoshita et al., 2000). It has been shown that the zygotic gene expression generally starts at the midblastula stage in developing medaka (Chong and Vielkind, 1989; Tsai et al., 1995). Taking these facts into consideration, the regulatory elements contained in pβ-Act-GFP-N becomes active with a time lag after the onset of the zygotic gene expression. It has been also reported that GFP fluorescence is already detectable before the midblastula stage in some transgenic medaka and zebrafish strains. For example, the transgenic medaka with the olvas-GFP gene (Tanaka et al., 2001), which is shown for comparison in Fig. 3 a-2 and a-4, a strong fluorescence is detected in the blastdisc of the one-cell stage. The fluorescence in a one-cell stage is, however, obviously derived from the maternal expression. In the case of pβ-ActGFP-N used in this study, the lack of expression in the cleavage stages is consistent with the very low-level of expression in the gonad of the adult fish.

Next, we added NLS of SV40 to the end of the GFP-coding region. As a result, a more intense green fluorescence was observed at the nucleus than in the cytoplasm (Fig. 3 e-3 and Fig. 4 b-2). It is evident that the NLS concentrated GFP in the nucleus and the low-level of fluorescence in the cytoplasm is supposed to be the result of the overflow from the nucleus. These results suggested that the intra-cellular localization of the transgene product could be controlled in the medaka using NLS of SV40. This element may be helpful for the transgenic studies, for example, the low level of expression may become detectable by concentrating the GFP to the nucleus.

A fish whose certain tissues are labeled, is a useful model to study the development of these tissues. For example, a transgenic zebrafish strain with Islet-1 promoter/enhancer sequence expresses GFP in the cranial motor neurons (Higashijima et al., 2000). The medaka strain with olvas-GFP expresses GFP exclusively in germ cells (Tanaka et al., 2001). In this study, transgenic strains harboring pβ-Act-GFP-N revealed a specific expression in the lens and the muscle. Since the expression pattern was uniform among the five separate strains, it is a specific characteristic of the regulatory elements contained in β-Act-GFPN. Some other transgenic medaka and zebrafish strains containing the GFP gene linked to the regulatory region of the β-actin gene have been reported previously. As for the medaka, Takagi et al. (1994), Hamada et al. (1998), and Chou et al. (2001) reported that the promoter of β-actin exhibits a ubiquitous expression in various tissues. Higashijima et al. (1997) produced two transgenic zebrafish strains; one harboring the GFP gene driven by a muscle-specific actin (alpha-actin) promoter specifically expressed the GFP throughout the muscle cells and the other harboring the GFP gene driven by the cytoskeletal β-actin gene promoter expressed the GFP throughout the body. The results in this study, in which strong expression is limited to the muscle and the lens, are different from the ubiquitous expression pattern in previous studies and are very similar to the expression pattern of the skeletal muscle actin. The reason of the difference in the expression patterns is unknown at present, however, it may be due to the difference of the length of the upstream region of the β-actin gene or differences in the construction of the downstream GFP gene. Nevertheless, the intense fluorescence in muscle cells of our transgenic strains may be helpful in the transplantation experiments to identify transplanted cells or nuclei because of its intense fluorescence.