Xenopus laevis is one of the major model systems for the analysis of vertebrate development. In the past, transgenic frogs have been used for various purposes such as (1) to dissect regulatory regions of genes (Knox et al., 1998; Casey et al., 1999), (2) to misexpress genes during development under specific spatial and temporal control (Hartley et al., 2001, 2002), and (3) to generate mutations in genes through gene trap approaches (Bronchain et al., 1999).
First attempts to create transgenic frogs had been reported in pioneering studies by Etkin et al. (1984) more than 2 decades ago. Linearized plasmid DNA was injected into fertilized Xenopus eggs and was found to be maintained in the form of long extrachromosomal concatemers during early stages of development, eventual integration into the genome occurring at later stages. Although this method is fast and easy, mosaic expression of the transgene in F0 is predominantly observed; uneven distribution of the episomal DNA before integration also results in low germ line transmission rates. Furthermore, the injected genes showed incorrect spatial and temporal regulation.
Approximately 10 years later, Kroll and Amaya (1996) developed a more efficient method for generating transgenic frogs. It involves restriction endonuclease-mediated integration (REMI) of DNA into demembranated sperm nuclei, followed by transplantation of the nuclei into unfertilized eggs. One of the major advantages of this method is that the transgene is integrated into the male genome before fertilization, resulting in high germ line transmission rates. However, the concentration and amount of the restriction enzyme used need to be well-controlled to avoid the risk of genome fragmentation. Inappropriate transplantation of the nuclei also leads to low survival rates of the embryos.
In principle, other transgenic methods reported for different model systems, such as fish, are also applicable to Xenopus embryos. These strategies include the use of DNA-NLS complexes (Liang et al., 2000) to achieve more efficient integration of transgenes, as well as the use of transposable elements, such as Sleeping Beauty (Davidson et al., 2003) and Frog Prince (Miskey et al., 2003), along with the transposase to obtain site-specific integration.
Use of the rare cutting meganuclease I-SceI to generate transgenic medaka fish has been reported recently (Thermes et al., 2002). I-SceI meganuclease is an endonuclease encoded by the mobile group I intron found in Saccharomyces cerevisiae mitochondria. The characteristic feature of this endonuclease is that it has an extended recognition site of 18 bp, which is expected to exist only once in 7 × 1010 bp of random DNA sequences. Several groups have also reported the successful use of this endonuclease for gene targeting in mammalian cells (Choulika et al., 1994; Cohen-Tannoudji et al., 1998; Richardson et al., 1999).
Here, we report the use of I-SceI meganuclease for the creation of transgenic Xenopus laevis embryos. Stable transgenic lines were established using various ubiquitous and tissues-specific promoters coupled to the green fluorescent protein (GFP) reporter. We observed high-frequency transgenesis and efficient germ line transmission.
RESULTS AND DISCUSSION
Generation of Transgenic Frog Embryos With Different Promoter Constructs Using I-SceI Meganuclease
We have applied an experimental approach to generate transgenic Xenopus laevis that was originally described for medaka fish by Thermes et al. (2002). The promoter of interest coupled to a reporter gene was subcloned into a vector containing two I-SceI recognition sites. Plasmid DNA was cleaved by and injected together with the I-SceI meganuclease into one-cell stage embryos within 1 hr after fertilization (see the Experimental Procedures section for details).
We have tested four different promoter constructs that drive the ubiquitous or tissue-specific expression of GFP: pCMV-GFP-SceI, which contains the CMV promoter (Kroll and Amaya, 1996), and pCSKA-GFP-SceI, which contains the cytoskeletal actin promoter from Xenopus borealis (Thermes et al., 2002) both driving ubiquitous expression of the GFP reporter, as well as pCry1GFP-SceI, containing the 490-bp fragment of γ-crystallin promoter (Offield et al., 2000), and pElastase-GFP-SceI, which contains the pancreatic-specific elastase promoter (Beck and Slack, 1999), driving tissue-specific expression in lens and pancreas, respectively. Transgene expression of embryos injected with pCMV-GFP-SceI and pCSKA-GFP-SceI were first detected after gastrulation and yielded highest levels at tail bud stage (stage 34). Strong GFP expression continued through tadpole stage (stage 42) and was still detectable after metamorphosis (Fig. 1A and data not shown). Transgenic embryos generated by injecting pCry1GFP-SceI started to express the transgene specifically in the lens of tail bud stage embryos (Fig. 1B), correlating with the activation of the endogenous promoter (Offield et al., 2000). Lens-specific expression was still detectable in the froglet stage. Similarly, transgenic embryos injected with pElastase-GFP-SceI showed pancreas-specific GFP expression at stage 45 (Fig. 1C). This corresponds to the endogenous expression of elastase in Xenopus laevis embryos (Beck and Slack, 1999). We were able to detect a strong and specific GFP signal during metamorphosis at stage 60 (Fig. 1C) and in the adult (data not shown).
Coinjection of DNA and I-SceI Meganuclease Increases the Frequency of Nonmosaic Transgene Expression
To optimize the experimental protocol used for the generation of transgenic Xenopus embryos, DNA and enzyme concentration were varied systematically. Plasmid pCMV-GFP-SceI was injected with or without meganuclease. In the absence of meganuclease, almost all of the GFP-positive embryos showed mosaic expression of the transgene (Table 1). When the same plasmid was injected together with meganuclease, the number of nonmosaic, GFP-expressing embryos increased significantly. Mosaic expression of GFP might be due to replication and persistence of injected plasmids as extrachromosomal episomes, which are inherited by only a subset of cells (Etkin and Pearman, 1987; Etkin et al., 1987). Mosaic expression of the transgene is gradually lost at later stages of development. Similar results were obtained when pCSKA-GFP-SceI (40 pg/embryo) and pCry1GFP-SceI were injected with or without meganuclease (Table 1).
Table 1. Transgenesis Efficiency for Different Promoter Constructs as Well as for Different DNA and Enzyme Concentrations
GFP-expressing from survival
Nonmosaic expression (%)
4 × 10−3
4 × 10−3
4 × 10−3
4 × 10−3
4 × 10−3
4 × 10−3
4 × 10−3
1 × 10−3
4 × 10−3
8 × 10−3
We have also injected different amounts of pElastase-GFP-SceI with different concentrations of I-SceI meganuclease (Table 1). Best results were obtained with the highest concentration of DNA used, further increase in plasmid concentration leads to high mortality of the embryos due to gastrulation defects. Variation of the enzyme concentration from 1 × 10−3 U/embryo to 8 × 10−3 U/embryo had no significant effect on the transgenic efficiency (Table 1). Further increase in meganuclease concentration leads to high mortality of the embryos due to gastrulation defects.
In all cases, injected embryos showed high survival rates of more than 50% and almost all of these embryos survived until after metamorphosis. This finding is an advantage over the REMI method in which injected embryos showed low survival rates upon transplantation of nuclei and most of the transgenic embryos generated using this method have difficulties surviving until the frog stage due to the generation of aneuploid embryos (Sparrow et al., 2000).
Germ-Line Transmission of Transgenes and Southern Blot Analysis of Transgenic Lines
Transgenic embryos injected with pElastase-GFP-SceI were scored for GFP expression in the pancreas. F0 transgenic embryos that were positive for GFP in the pancreas were raised to sexual maturity and mated with wild-type frogs. To check for germ-line transmission, F1 tadpoles were scored for GFP-expression in the pancreas. Six mature F0 founder frogs were raised to maturity. A significant degree of germ line transmission, which ranged from 1–76%, was observed for most of these lines (Table 2). A germline transmission rate of higher than 50% suggests that the F0 founder frog had multiple transgene-integration loci, whereas germline transmission rates that are lower than approximately 50% indicate that there is mosaicism in the F0 founder germ line. The rates close to 25% are likely to reflect that the founders were “half transgenics,” which is sometimes also seen with the REMI transgenesis method (Marsh-Armstrong et al., 1999; Hartley et al., 2002), suggesting a late integration event at the two-cell stage.
Table 2. Germline Transmission of Transgenic Carrying pElastase-GFP-SceI
F0 founder frog
Gene transmission rates
The pattern of transgene expression for the founder frogs during tadpole stages was not recorded. However, the tissue specificity and levels of transgene expression were maintained and comparable to those observed in typical F0 founder frog (Fig. 2).
The pioneering study by Etkin and Pearman (1987) had shown that the plasmid DNA-injected embryos were able to transmit the transgene only through the male germline. Here, we show that the transgene is also transmitted through the female germline using the meganuclease approach. We obtained a mature F0 female founder frog that transmitted the transgene to 30% of its F1 offspring.
To determine the nature of DNA integration, such as the number of insertion loci and the length of concatemers, we performed Southern blot analysis on genomic DNA isolated from independent F1 offspring of four different lines carrying pElastase-GFP-SceI. Isolated genomic DNA was digested with BamHI that liberates the insert from the plasmid pElastase-GFP-SceI (Fig. 3A). Analysis of four independent transgenic lines identified a 1.4-kb fragment that is common to all individuals from all lines, confirming integration of the GFP encoding portion. The restriction pattern of the offspring from line 2 suggests that the plasmid was incompletely digested (either at I-SceI cleavage site 1 or site 2) such that embryos contain insertion of the entire plasmid in tandem repeat as indicated by the presence of the corresponding characteristic fragments (3.2 kb, 1.4 kb, and 0.5 kb; Fig. 3B, lanes 1 and 2). In line 3, the absence of a correspondingly intense 3.2-kb signal indicates formation of tail to tail concatemers (Fig. 3B, lanes 3 and 4). The restriction patterns of lines 2, thus, suggests that the I-SceI cleavage sites were maintained in the genome of these transgenic lines, similar to what has been reported by Thermes et al. (2002) for medaka fish. In line 4, the absence of the 3.2-kb and 0.5-kb fragments indicates integration of the insert alone (Fig. 3B, lane 5). The presence of other fragments in all lines tested is likely to reflect different chromosomal integration sites and, thereby, generation of different junctional fragments. A GFP-negative control embryo showed no hybridization to the probe (Fig. 3B, lane 8).
The copy number of integrated concatemers was estimated using a dilution series of plasmid DNA. Assuming a DNA content of 3.1 × 109 bp per haploid genome (Tymowska, 1973), 15 pg of BamHI-digested plasmid pElastase-GFP-SceI were loaded to estimate the intensity of one copy plasmid DNA integrated into the Xenopus laevis genome. Copy numbers ranging from one or two to a maximum of approximately eight copies were observed (Fig. 3B). These numbers are significantly lower than those observed in the transgenics obtained by the REMI method, in which the linearized plasmid tends to form concatemers of more than 10 copies that also resulted in variable transgene expression levels and phenotypes (Hartley et al., 2002).
In conclusion, we have applied a novel approach to generate transgenic frogs that may have several advantages. First, the meganuclease approach is simple such that large numbers of transgenic embryos can be generated in a single experiment. Second, the transgenesis efficiency is higher than those reported for other methods. The transgenes show faithful temporal and spatial expression and high survival rates. We have also demonstrated germline transmission of the transgenes. Meganuclease coinjection, thus, provides a simple and highly efficient tool for transgenesis in Xenopus.
pCMV-GFP-SceI was generated by inserting a 2.0-kb SalI–NotI fragment (containing the CMV promoter coupled to GFP reporter sequence and a SV40 polyA tail) from pCSGFP3 into the ISceI-pBSII SK+ vector containing two I-SceI recognition sites (Thermes et al., 2002). pCSKA-GFP-SceI was as described (Thermes et al., 2002). pElastase-GFP-SceI was generated from pElastase-GFP (Beck and Slack, 1999) restricted with NotI (New England Biolabs, Ipswich, MA). The 2.1-kb NotI fragment, containing the 205-bp, pancreas-specific elastase promoter coupled to GFP and a β-globin poly A tail, was subcloned into the ISceI-pBSII SK+ vector. A 1.5-kb NheI–NotI fragment, containing 490 bp of γ-crystallin1 promoter coupled to GFP and a SV40 polyA tail, was generated from pCry1GFP3 (a gift from Robert Grainger) and subcloned into the SpeI and NotI sites of the ISceI-pBSII SK+ vector to create pCry-GFP-SceI.
Preparation of Xenopus Eggs and Embryos
Pigmented Xenopus laevis eggs were obtained by injecting the dorsal lymph sacs of females with 50 IU of human chorionic gonadotrophin (hCG) in the evening followed by 1,000 IU hCG the next morning before egg collection. Eggs were fertilized in vitro with minced testes in 0.1× MBS (1.76 mM NaCl, 48 μM NaHCO3, 20 μM KCl, 200 μM Hepes, 16 μM Mg2SO4, 8 μM CaCl2, 6 μM Ca(NO3)2, pH 7.4), dejellied with 2% cysteine hydrochloride (pH 7.8–8.0) and cultured in 0.1× MBS buffer. Embryos were staged according to Nieuwkoop and Faber (1967).
Microinjection of Plasmid DNA With Meganuclease
Plasmid DNA was digested with I-SceI meganuclease (Roche Diagnostics, Germany, or New England Biolabs) before injection [1–15 ng/μl plasmid DNA; 0.5 × commercial meganuclease buffer (Roche Diagnostics, Germany, or New England Biolabs); 1–10 U/μl I-SceI meganuclease]. The reaction mixture was incubated for 40 min at 37°C, and 4 nl of the reaction mixture was injected into one-cell stage embryos within 1 hr after fertilization, using a microinjector (Eppendorf, Germany). The reaction mixture was injected in between the sperm entry site and the center of the animal pole, the region where the pronuclei of sperm and egg will fuse, to bring the injected DNA in close proximity to the forming nucleus. Injected embryos were cultured at 12.5°C until two-cell stage and to allow integration of the transgene before first cleavage, and then transferred to 18°C for normal culturing.
Embryos were scored for GFP expression using a MZFLIII Leica dissecting microscope with a 370- to 420-nm excitation filter and a 455-nm LP emission filter. The GFP-positive embryos were sorted and bred to maturity. Mature GFP-positive frogs were mated with wild-type frogs to analyze for germline transmission of the transgene. F1 embryos were sorted according to GFP expression, and the germline transmission rates were determined.
Genomic DNA Extraction and Southern Blot Analysis
Individual 2-month-old tadpoles (F1 pElastase-GFP) were ground to a fine powder in liquid nitrogen using a prechilled mortar and pestle. Genomic DNA was extracted using the QIAGEN Blood and Cell Culture DNA kit (Qiagen, Germany). Isolated genomic DNA was digested to completion with BamHI. DNA standards were prepared by digesting pElastase-GFP-SceI used to generate the transgenic frog by BamHI. Assuming an estimated DNA content of 3.1 × 109 bp per haploid genome (Thermes et al., 2002), 15 pg of digested plasmid were loaded to estimate the signal intensity corresponding to a single gene copy. DNA samples (10 μg/lane) were separated in 0.8% agarose gels in 1× TBE and subjected to Southern blot analysis according to standard procedures (Sambrook and Russell, 2001).
DNA probes for Southern blot analysis were generated by restricting the pElastase-GFP-SceI with NotI and gel purification with Qiagen Gel Extraction and Purification Kit (Qiagen, Germany). The 2.1-kb NotI fragment was then labeled with the nonradioactive ECL Direct nucleic acid labeling and detection system (Amersham BioScience, UK).
The authors thank Dr. Jochen Wittbrodt for providing us with the ISceI-pBSII SK+ vector construct and for providing information on I-SceI–mediated transgenesis in medakafish before publication. We also thank Dr. Jonathan Slack for the pancreas-specific elastase promoter construct and Dr. Robert Grainger for the γ-crystallin1 promoter construct.