A flk-1 promoter/enhancer reporter transgenic Xenopus laevis generated using the Sleeping Beauty transposon system: An in vivo model for vascular studies

Abstract

We have used the Sleeping Beauty (SB) transposable element to generate transgenic Xenopus laevis with expression of green fluorescent protein (GFP) in vascular endothelial cells using the frog flk-1 promoter. This is the first characterization of a SB-generated transgenic Xenopus that has tissue-restricted expression. We demonstrate that the transgene integrated into single genomic loci in two independent founder lines and is transmitted through the germline at the expected Mendelian frequencies. Transgene integration occurred through a noncanonical transposition process possibly reflecting Xenopus-specific interactions with the SB system. The transgenic animals express GFP in the same spatial and temporal pattern as the endogenous flk-1 gene throughout development and into adulthood. Overexpression of xVEGF122 in the transgenic animals disrupts vascular development that is visualized by fluorescent microscopy. These studies demonstrate the convenience of the SB system for generating transgenic animals and the utility of the xflk-1:GFP transgenic line for in vivo studies of vascular development. Developmental Dynamics 236:2808–2817, 2007. © 2007 Wiley-Liss, Inc.

INTRODUCTION

Sleeping Beauty (SB) is a DNA-based transposable element that belongs to the Tc1/mariner family and has been developed for transgenesis and insertional mutagenesis in a variety of vertebrate models (Ivics and Izsvak,2004). SB has been used to deliver genes into the germline of mice, fish, and Xenopus (Dupuy et al.,2002; Davidson et al.,2003; Grabher et al.,2003; Sinzelle et al.,2006). It has also been used as a gene therapy vector in the treatment of mouse models of human genetic diseases (Hackett et al.,2005). Insertional mutagenesis screens using SB have been carried out in mice and zebrafish and have identified genes involved in cancer as well as early development (Collier et al.,2005; Dupuy et al.,2005; Sivasubbu et al.,2006).

Sleeping Beauty was reconstructed from inactive Tc1 elements in teleost fish (Ivics et al.,1997) and its mechanism of transposition has been characterized by several laboratories (Izsvak and Ivics,2004). The SB transposase recognizes terminal invert/direct repeat sequences (IR/DR) in the transposon and mobilizes it from one DNA site to another through a cut-and-paste mechanism. The open reading frame of the transposase gene was removed from the transposon to generate a nonautonomous element that is unable to mobilize until the transposase is provided in trans. The transposase is highly specific for the SB IR/DR terminal repeat elements and is unable to mobilize inactive Tc1 elements in fish or other organisms. In human cell lines, integration of the SB transposon occurs randomly at TA dinucleotides with no obvious preference for individual chromosomes or chromosomal location. These characteristics have made SB an attractive system for genetic manipulations in vertebrates.

The SB system has been shown to function in a wide range of vertebrate cell lines from fish to humans, including Xenopus laevis A6 kidney cells, suggesting that host factors essential for transposition are conserved in vertebrates (Izsvak et al.,2000). Host factors, however, are thought to modify the transposition process as the efficiency of transposition varied significantly among the cell lines. The chromatin binding protein HMGB1 was identified as a host cofactor for the SB transposase and was shown to increase transposition in mammalian cell lines (Zayed et al.,2003). In addition, the transcription factor Miz-1 was shown to interact with the SB transposase (Walisko et al.,2006). These, and potentially other host factors, may influence the efficiency of SB transposition and should be considered when using SB in different cells and organisms.

Sinzelle and colleagues described a simple method for generating transgenic X. laevis using the SB system (Sinzelle et al.,2006). A β-actin promoter:GFP transgene flanked by IR/DR sequences was co-injected together with SB transposase RNA into fertilized eggs and resulted in animals with ubiquitous expression of the reporter. They demonstrate transposase-dependent integration of the transgene into the germline of founder animals and subsequent transmission of the transposon transgene to the F1 and F2 progeny. Although the SB transposase was required for transgenesis, integration site analysis revealed that transgene insertion was by a noncanonical transposition mechanism. Their work suggests that the SB system is an attractive alternative for generating transgenics in Xenopus.

Xenopus has many characteristics that make it a useful model to study vascular development and angiogenesis. The embryos develop externally and are transparent, which enables easy visualization of vascular structures. It has also been proposed that the vascular system of Xenopus shows more similarity to higher vertebrates than do other models, such as zebrafish (Levine et al.,2003). For example, Xenopus embryos have septated atria, defined heart valves, lungs, blood islands, and a vitelline network (Kau and Turpen,1983; Kolker et al.,2000; Mohun et al.,2000). Xenopus are tetrapods and allow study of the vascular system in developmental processes limited to higher vertebrates, such as limb formation, lung development, limb or tail regeneration, and tail regression during metamorphosis.

Levine and coworkers described a detailed atlas of vascular development in Xenopus by in vivo labeling of endothelial cells with circulating fluorescently labeled low-density lipoprotein (DiI-Ac-LDL; Levine et al.,2003). Vascular endothelial cells were labeled by injecting DiI-Ac-LDL into the beating heart of tail bud stage embryos. The fluorescent label is stable for several weeks and enabled the generation of a three-dimensional atlas of vascular development from stages 33 to 46. Their technique was ideal for examining early vascular development but has limitations. The early vascular development before the heart began beating and later stages of development, during limb formation or tail regression, were inaccessible using this technique.

The receptor tyrosine kinase flk-1/vegfr2/KDR-1 is expressed in the developing embryonic vasculature as well as bipotential hemangioblasts, cells that become either blood or endothelial cells (Millauer et al.,1993; Peters et al.,1993; Yamaguchi et al.,1993; Huber et al.,2004). Targeted deletion of flk-1 results in the failure of the vascular system to develop in mice, demonstrating an essential role for this receptor tyrosine kinase in this process (Shalaby et al.,1995). In Xenopus, the expression of flk-1 occurs in the developing vascular system (Cleaver et al.,1997). Xenopus flk-1 (xflk-1) is expressed in endothelial precursor cells that will form the major blood vessels of the embryo as well as in the primary vasculature and angiogenic-derived intersomitic vessels. These data suggest that xflk-1 is a useful marker for vascular development and angiogenesis in the frog.

We isolated the X. laevis flk-1 gene to generate a vascular endothelial-specific reporter transgene. An enhanced green fluorescent protein (eGFP) reporter was cloned in frame with the first coding exon of the xflk-1 gene and a DNA fragment containing a rabbit β-globin polyadenylation signal was cloned into the second exon. The xflk-1 transgene, including 2.5 kb of the 5′ promoter region and the entire first intron, was cloned into a nonautonomous SB transposon. Transgenic founders were generated by co-injecting the SB transposon transgene and synthetic messenger RNA encoding the SB transposase into fertilized eggs as described by Sinzelle and coworkers (Sinzelle et al.,2006). We demonstrate integration of the transgene into a single genomic locus and germline transmission of the transgene. Of interest, transgene mobilization occurred through noncanonical transposition, an observation that was also reported by Sinzelle and coworkers and may reflect Xenopus host factor interactions with the SB system. Transgenic animals express GFP in the same spatial and temporal pattern as the endogenous xflk-1 gene making this an ideal model to study vascular development and angiogenesis. Disruption of vascular architecture was readily visualized in transgenic animals that were injected with Xenopus vascular endothelial growth factor (xVEGF) RNA. These studies support and extend the work of Sinzelle et al. by demonstrating that the SB system efficiently integrates transgene DNA into the genome of X. laevis and that the animals maintain tissue-restricted expression of the transgene.

RESULTS AND DISCUSSION

Germline xFlk-1:GFP Transgenic X. laevis

We have generated a vascular endothelial-specific transgenic X. laevis using the SB transposon system. The transgene was engineered from X. laevis flk-1 gene and contains 2.5 kb of upstream promoter sequence and the entire (∼3.5 kb) first intron (Fig. 1A). An eGFP was inserted into exon 1, and the rabbit β-globin polyadenylation signal (pA) was cloned downstream of the first four codons of exon 2. This arrangement of the transgene was used to promote splicing of the nascent transcript and to include the first intron of the xflk-1 gene that may contain an enhancer element. Analysis of the murine flk-1 gene indicated the presence of an enhancer element in the first intron (Kappel et al.,1999). The transgene was cloned between the Sleeping Beauty invert repeat/direct repeat sequences (IR/DR) of a nonautonomous element. Transgenic animals were generated by co-injecting the plasmid containing the transgene construct (pT2xflk-1:GFP) together with SB transposase mRNA into one-cell fertilized embryos (Fig. 1B). The injected tadpoles were raised to adulthood and outcrossed with wild-type frogs to determine germline transmission of the transgene. To date, we have outcrossed 15 of 31 animals (15 males and 16 females) that survived to adulthood. Four founders that transmit the transgene through the germline were identified and two of these founders, JD1 and JD12, were studied in detail and are presented in this manuscript. Three of the founders (JD1 ♂, JD4 ♀, and JD12 ♂) generated F1 offspring that expressed the GFP reporter throughout the vasculature (see below and data not shown). One founder, JD3 (female), produced F1 offspring that had expression of the GFP reporter restricted to the outer surface of the eye and not in the vasculature (data not shown). This finding suggests that the expression of the GFP reporter in line JD3 is influenced by the genomic locus flanking the integration site of the transposon transgene. The observed transgenesis rate was approximately 27% (4/15). This percentage may be an underestimate of the actual transgenesis rate as several of the outcrosses gave low numbers of tadpoles (<100) and, if the transmission frequency is low due to the mosaic nature of the founder germline, a large number of offspring may need to be scored to identify positive progeny. The transgenesis rate we observed is lower than that described by Sinzelle and colleagues (five founders from 12 GFP-positive adults = 40%; Sinzelle et al.,2006). The difference may be accounted for by several factors, including that we did not presort the injected embryos for GFP expression and due the large size of the xflk-1 transgene insert.

Figure 1. Generation of the Xenopus laevis flk-1 reporter transgenic frogs. A: Schematic diagram of the pT2xflk-1:GFP transposon transgene (not to scale). The transgene was generated from a BAC encoding the X. laevis flk-1 gene and contained 2.5 kb of upstream sequence, the translation start site in exon 1, the entire first intron (∼3.5 kb), and the 5′-end of exon 2. The green fluorescent protein (GFP) reporter was cloned in frame with the coding sequence of exon 1. The first four codons of the xflk-1 open reading frame were included to provide the endogenous Kozak translation start site sequence. The last three codons of exon 1 were included at the 3′-end of the GFP open reading fame to ensure that the splice donor site remained intact. A rabbit β-globin polyadenylation signal (pA) was cloned four codons downstream of the start of exon 2 to maintain the integrity of the splice acceptor. The xflk-1:GFP construct was cloned into the pT2 nonautonomous transposon, between the invert-direct repeats (IR/DR). The approximate positions of the BamHI, EcoRI and EcoRV restriction endonuclease sites within the pT2xflk-1:GFP transposon transgene are indicated. B: Schematic diagram of procedure used to generate founder, F1 and F2, transgenics. The pT2xflk-1:GFP DNA was injected together with Sleeping Beauty transposase mRNA into fertilized eggs at the one-cell stage. Tadpoles were raised to adulthood and outcrossed to wild-type animals to produce F1 and F2 generations. C: The rate of transgene transmission from two founders (JD1 P0 and JD12 P0) and two JD1 F1 progeny was determined by outcross to wild-type frogs. The transgenesis rate is calculated as the percentage of GFP-positive tadpoles in an individual clutch. D: Fluorescence in situ hybridization (FISH) on cell spreads prepared from F1 tadpoles from founders JD1 and JD12. Antisense GFP probes detect single integration sites (one hybridization signal) per nucleus for the pT2xflk-1:GFP transgene in each of the founder lines. The nuclei are counterstained with DAPI (blue). E: Southern blot analysis of DNA harvested from tadpoles generated by outcross of the two founders (JD1 and JD12) and two F1 animals (F2 tadpoles). Genomic DNA was digested with BamHI (B), EcoRI (I), or EcoRV (V), and the blot was hybridized with a radiolabeled GFP probe. Digestion with BamHI (B) releases an internal ∼3.5-kb DNA fragment that includes the xflk-1 promoter and the open reading frame of GFP (arrowhead and see A for schematic representation). The BamHI fragment is present in all GFP fluorescent-positive progeny (+) but is never detected in GFP fluorescent-negative siblings (−). Digestion of genomic DNA with restriction endonucleases that cut once within the transposon transgene (EcoRI or EcoRV) results in an ∼8.5-kb band (*) that is the predicted size of the entire transposon, suggesting that the transgene has integrated as a concatamer. L, 1-kb DNA marker lane.

Founders JD1 and JD12 were crossed to wild-type animals and transmitted the transgene to F1 offspring at a rate of 38% and 25%, respectively (Fig. 1C). F1 offspring were raised to adulthood and outcrossed to wild-type animals. Approximately 50% of the F2 offspring expressed the transgene suggesting Mendelian inheritance of a single transgene or closely linked transgenes. To confirm this observation, we performed fluorescence in situ hybridization (FISH) of F1 animals and demonstrated a single hybridization signal within each nucleus of cells harvested from heterozygous GFP-positive F2 tadpoles (Fig. 1D). These data indicated that the transposon(s) integrated into a single genomic location.

The two founder animals demonstrate noncanonical integration of the transposon. Southern blots of genomic DNA harvested from F1 and F2 tadpoles were probed with a radiolabeled GFP probe (Fig. 1E). Digestion of the transgene with BamHI releases a 3.5-kb fragment that contains the GFP sequence. The 3.5-kb band is detected in all positive animals screened and is never detected in the negative tadpoles, suggesting that transgene silencing is not occurring in the nonfluorescent animals. Digestion of the genomic DNA with either EcoRI or EcoRV, which cut once within the transgene, released a major band of ∼8.5 kb and multiple minor bands of higher and lower sizes. The ∼8.5-kb band may represent full-length transposon transgenes that had integrated as a concatamer. The weakly hybridizing bands may represent the flanking ends of the concatamer or perhaps additional integrations that occurred close to the concatamer. To identify the genomic integration site, we carried out a polymerase chain reaction (PCR) -based DNA walk on genomic DNA and identified plasmid vector sequences adjacent to transposon (data not shown). If precise SB transposition had occurred, the sequence flanking the transposon IR/DRs should be Xenopus genomic DNA and should include the target site (TA) duplication. Together with the Southern and FISH analyses, these data suggest that a concatamer of transgenes was integrated by means of a noncanonical transposition reaction. A noncanonical SB integration in X. laevis was reported by Sinzelle et al. and may reflect a Xenopus-specific characteristic of this system (Sinzelle et al.,2006). Alternatively, the ratio of transposon vector to transposase mRNA may be critical for mobilization of a single transposon from vector to genome.

GFP Transgene Expression Recapitulates Endogenous Flk-1 Expression

GFP fluorescence was first detected in F1 tadpoles at stage 26/27, predominantly in the tail region. The yolk in Xenopus eggs and tadpoles is autofluorescent and interferes with visualization of the GFP fluorescence. To overcome the problems associated with autofluorescence of the yolk, we performed in situ hybridization of transgenic embryos at early developmental stages with xflk-1 and GFP riboprobes to compare transgene expression with endogenous xflk-1 expression (Fig. 2). The expression profile of xflk-1 during early development recapitulates that described by Cleaver et al. (1997), with the exception that intense staining of the xflk-1 transcript is detected in the tail region at tail bud and tadpole stages (stage 25 and 37, respectively). In situ hybridization of sibling tadpoles with antisense GFP riboprobes revealed that the transgene and the endogenous xflk-1 expression are virtually indistinguishable at tadpole stages (stage 37; Fig. 2). At earlier stages (stage 20 and 25), however, expression of the transgene is seen at exaggerated levels in the tail region compared with the endogenous xflk-1 transcript (arrowhead; Fig. 2). At stage 20, expression of the GFP transgene and the endogenous xflk-1 are both detected in the ventral anterior region of the presumptive endocardium (arrow; Fig. 2). The expression of GFP and xflk-1 in the posterior region of the tail bud embryo resembles that of xnot2 in the tail organizer described by Gont and colleagues (Gont et al.,1993). The tail organizer is derived from the late dorsal blastopore lip and is responsible for organizing the development of the tadpole tail. Genes involved in vasculogenesis (xMsr; Devic et al.,1996) and early hematopoiesis (LMO2; Mead et al.,2001) are also expressed in this tail organizer region, suggesting that blood vessel development is integral to tail development. It is therefore not unexpected that endogenous xflk-1, and the xflk-1:GFP transgene, are expressed in the tail organizer region as well.

Figure 2. In situ hybridization for green fluorescent protein (GFP; left panels) and endogenous xflk-1 (right panels) in stage 20, 25, and 37 embryos. All embryos are shown with anterior to the left. Stage 20 embryos are shown from lateral (top) and ventral (bottom) views (top panels). Lateral view of stage 25 embryos (middle panels) and stage 37 embryos (bottom panel). At stage 20, both GFP and xflk-1 are detected in the presumptive endocardium (arrows). At swimming tadpole stages (stage 37), the expression pattern of the transgenic reporter (GFP) is similar to that of the endogenous xflk-1 gene. At earlier developmental stages, the GFP reporter is more highly expressed in the tail region of the embryo (arrowheads) than the endogenous xflk-1 gene. This finding is most obvious at stage 25 (center panels).

Xenopus tadpoles remain transparent until metamorphosis, and this finding enables visualization of the pT2xflk-1:GFP transgene throughout development. GFP fluorescence was observed in the developing vascular system at stage 35 (Fig. 3) and became easier to visualize as the autofluorescent yolk cleared at later developmental stages. At early swimming tadpoles stages, the vascular network is composed of large vessels (Fig. 3) that become progressively more branched and complex as development proceeds (Fig. 4).

Figure 3. The pT2xflk-1:GFP transgene labels vascular endothelial cells in the swimming tadpole (stage 35–36). The vascular network at early tadpole stages is comprised of large vessels with few branches. acv, anterior cardinal vein; dc, Duct of Cuvier; isv, intersomitic vessels; lh, lymph heart; pcv, posterior cardinal vein; GFP, green fluorescent protein.

Figure 4. Green fluorescent protein (GFP) fluorescent images of stage 45 transgenic tadpoles. Founders JD1 and JD12 were outcrossed with albino females and transgenic progeny were selected for microscopy. A,C: Dorsal view of a transgenic JD1 F1 tadpole (A = 60×; C = 100×). The vasculature of the head is clearly visible and several of the major vessels are labeled to orientate the reader. B,D: Ventral views of transgenic tadpoles (B = JD1 F1 at 60×; D = JD12 F1 at 100×. Note: the GFP expression patterns are indistinguishable in the F1 progeny from founders JD1 and JD12). E,F: Vasculature of the tail of stage 45 JD1 F1 tadpoles (E = 60×; F = 150×). Blood is supplied to the tail by means of the dorsal aorta and returns by means of the posterior cardinal vein. The dorsal longitudinal anastomosing vessel (dlav) drains into the posterior cardinal vein (pcv). White arrows indicate small blind-end vessels sprouting from the pcv. cl, cloaca; da, dorsal aorta; dlav, dorsal longitudinal anastomosing vessel; h, heart; ijv, internal jugular vein; mab, musculo-abdominal vein; msa, mesencephalic artery; oa, optic artery; ov optic vein; np, nasal pit; pcv, posterior cardinal vein; pcvlb, left branch of the posterior cardinal vein; pcvrb, right branch of the posterior cardinal vein.

The vascular network is well defined by stage 45 (Fig. 4). The head vasculature is clearly visible at this stage (Fig. 4A–D). The internal jugular veins (ijv) are the main dorsal head veins and are bilateral vessels that run longitudinally down the head (Fig. 4C). The ophthalmic artery (oa) and ophthalmic vein (ov) are detected running to and from the eye. The ov drains into the ijv, whereas the oa branches from the internal carotid artery (not seen). Vessels surrounding the gills are detected from the ventral surface of the tadpole (Fig. 4B,D). The musculo-abdominal veins (mab) are bilateral vessels that run from the jaw to the heart and are detected from the ventral surface of the head. At stage 45, the tail musculature is fully vascularized (Fig. 4E) and new vessels are beginning to grow into the ventral fin (Fig. 4F).

By stage 55, both the dorsal and ventral fin are vascularized. The ventral fin vessels were first seen as single, blind-end vessels growing from the posterior cardinal vein (pcv) into the fin (Fig. 5A; also see Fig. 4F). These vessels grow into loops that are continuous with the pcv (Fig. 5B), and the loops then extend and fill the fin in a lace-like network (Fig. 5C). The ventral fin vessels form before the dorsal fin and develop into a more extensive network (Fig. 5C).

Figure 5. A–C: Green fluorescent protein (GFP) fluorescent images from the tail of stage 45 (A), 50 (B), and 55 (C) F1 tadpoles. A: Blind-end vessels extend from the posterior cardinal vein (white arrows) into the ventral fin of the stage 45 tadpole. B: Vessels form closed loops in the ventral fin of stage 50 tadpoles. C: Vessels form a complex network in the ventral fin and simple loops in the dorsal fin of a stage 55 tadpole. dlav, dorsal longitudinal anastomosing vessel; pcv, posterior cardinal vein.

A small number of GFP-positive cells could be detected circulating within the vasculature in swimming tadpoles shortly after the heart begins beating (data not shown). Several reports demonstrate that flk-1 is expressed in bipotential progenitor cells that can give rise to vascular endothelial cells and blood cells (Millauer et al.,1993; Peters et al.,1993; Yamaguchi et al.,1993; Huber et al.,2004). The presence of circulating GFP-positive cells in the transgenic tadpoles may reflect the maintenance of GFP protein in cells derived from these progenitors.

Cleaver et al. noted a decline in the level of detectable xflk-1 after stage 40 (Cleaver et al.,1997). This finding is likely to reflect the limitations of whole embryo in situ hybridization for vascular staining and not a loss of xflk-1 expression in mature vessels. Flk-1 is expressed in mature vascular endothelial cells of both mouse and human. The ease of visualizing vasculature in xflk-1:GFP transgenic tadpoles highlights one of the benefits of this system for studying vasculogenesis. To overcome the problems associated with pigmentation of the skin at late tadpole stages, we outcrossed both transgenic male founders (JD1 and JD12) with albino females with the ultimate goal of incrossing the resulting progeny to generate transgenic albino animals. To our surprise, approximately 50% of the offspring from each founder lacked all body pigmentation, indicating that the founder animals were likely to be heterozygous for the albino phenotype. The transmission frequency of the GFP transgene was the same as noted earlier (Fig. 1C) and resulted in populations of albino tadpoles that contained the xflk-1:GFP transgene and thus allowed unobstructed observation of the developing vasculature at late tadpole stages (Fig. 6).

Figure 6. Founder JD1 was outcrossed with albino females and albino transgenic progeny were selected for microscopy (see text for details). Vascular morphology in the anterior region of limb bud stage embryos is revealed by the xflk-1:GFP transgene. A,B: Stage 51 albino F1 transgenic tadpoles were photographed in the ventral (A = 25×) and dorsal (B = 40×) orientations. The developing vasculature is clearly visible, and several of the major vessels and organs are labeled to orientate the reader. h, heart; hlb, hind limb bud; ijv, internal jugular vein; mab, musculo-abdominal vein; np, nasal pit; oa, optic artery; ov, optic vein; t, tentacle; th, thymus.

GFP expression is maintained in the adult transgenic animals and is clearly visible in the webbing of the hind limb (Fig. 7A). GFP expression was detected by immunohistochemistry in the internal organs of juvenile adult F1 animals. Sections of fixed tissues were stained with a monoclonal antibody to GFP to visualize expression of the transgene as the size and the optical density of the adult tissues made fluorescent microscopy difficult. We examined the expression of the transgene in highly vascularized tissues, for example the lung, to demonstrate the maintenance of GFP expression in postmetamorphic organs. Figure 7B–D shows sections of postmetamorphic lung tissue and expression of the xflk-1:GFP transgene in the lining of a blood vessel. These data indicate that the xflk-1:GFP transgenic lines generated in this study will be useful for studying blood vessel development and homeostasis throughout the lifespan of the frog in both normal and disease states.

Figure 7. Green fluorescent protein (GFP) expression in vasculature of transgenic pT2xflk-1:GFP adult F1 frogs. The webbing between the digits of the hind limb is transparent and allows visualization of the vasculature of the adult animal. A: Fluorescent vessels are clearly seen in the webbing at high magnification (A = 75×). The cartoon insert depicts the region of the foot that is shown in the figure. The clawed toe is outlined by the white dashed line. B–D: GFP expression can be visualized in adult internal organs by immunohistochemistry with GFP-specific antibodies. Lungs were dissected from postmetamorphic transgenic frogs, fixed, and embedded in paraffin. B,C: Serial 4-μm sections were prepared and immunostained with GFP antibodies and counterstained with hematoxylin. D: Alternate sections were stained with Giemsa for histological examination (D = 400×). The morphology of the postmetamorphic lung is shown in B (20×). GFP expression was readily detected in the vascular endothelial cells that line blood vessels in the lung (black arrows; C = 400×). C,D: Nucleated red blood cells (erythrocytes) are found within the blood vessel.

In Vivo Vascular Studies

To demonstrate the utility of the xflk-1 transgenic animals for in vivo vascular studies, we overexpressed VEGF, the ligand for flk-1, by injecting mRNA into early cleavage stage embryos and monitored vascular development by fluorescence microscopy. Several groups have demonstrated that overexpression of VEGF122, the secreted form of xVEGF, disrupts normal vascular development (Cleaver et al.,1997; Cleaver and Krieg,1998; Koibuchi et al.,2006). We injected xVEGF122 mRNA, together with a rhodamine–dextran fluorescent marker as a lineage tracer, into a ventral-vegetal blastomere at the four-cell stage (Fig. 8A). This targets the RNA to the posterior flank of the tadpole where the vascular network is less complex. At stage 42, we observed enlarged vessels (Fig. 8D,F) along the ventral side of the tail in what appears to be the posterior cardinal vein. The control embryos received only the fluorescent lineage tracer and displayed normal vasculature (Fig. 8B). By stage 45, 92% (24/26) of the embryos injected with xVEGF122 RNA displayed marked edema compared with 11% (3/28) of the control embryos. These results demonstrate that the xVEGF122 RNA disrupts the vasculature as reported by others, however, we were able to monitor this defect in live animals, without a lengthy in situ hybridization step. The ability to analyze expression of xflk-1 expression in vivo allows continuous observation of the experimental animals, whereas with in situ-based techniques, the tadpoles are fixed and only a “snap shot” of development can be assessed for each subject. The above experiment highlights the potential use of this transgenic line to rapidly monitor the activity of exogenous genes, or chemicals, on vascular development.

Figure 8. The pT2xflk-1:GFP transgenic lines can be used to study vascular development in an in vivo model. A: Vascular endothelial growth factor (VEGF) 122 RNA (2 ng) was injected together with a rhodamine-labeled dextran into the ventral blastomere at the four-cell stage, and embryos were examined at stage 42. B,C: Stage-matched control embryos were injected with rhodamine–dextran only. Embryos were visualized to detect the green fluorescent protein (GFP) transgene (green; B,D,F) and the rhodamine–dextran (red; C,E,G). B,C: The posterior cardinal vein (pcv) of control embryos develop normally. D–G: Ectopic expression of VEGF results in disruption of normal pcv formation. The pcv of the VEGF mRNA-injected embryos displayed extensive hyperplasia. Lateral views of the embryos are shown with the dorsal side up.

EXPERIMENTAL PROCEDURES

Plasmids

The Sleeping Beauty transposon (pT/StuI/MscI) and transposase (pSBRNAX) constructs were kind gifts from Dr. Perry Hackett (Ivics et al.,1997). Site-directed mutagenesis of the right and left arms of pT/StuI/MscI was performed using Quickchange site directed mutagenesis kit (Stratagene) to generate the pT2 transposon that was described by Cui and coworkers (2002). The modified arms were sublconed into the SacI and Acc65I sites of pBluescript (pBS SK+) to generate pT2SK+ and, thus, provide a multiple cloning site for subsequent cloning steps. The xVEGF122/pCS2 plasmid was a kind gift from Dr. Paul Krieg (Cleaver et al.,1997).

Construction of the X. laevis flk-1:GFP Transgene

Four high density filters of an arrayed X. laevis BAC library (Invitrogen catalog no. 97110M; HinDIII digest, average insert size 108 kb, pBeloBAC11 vector, 2X genomic coverage) were screened with the 5′-end of the murine flk-1 cDNA (IMAGE 4238984) to isolate the xflk-1 gene. Two overlapping BAC clones were identified in the screen (7o3 and 150e7). The 5′-end of the X. laevis flk-1 gene was isolated from clone 7o3 by subcloning and sequencing restriction enzyme fragments derived from this BAC. PCR was used to amplify a fragment containing the first four codons after the translation start site of xflk-1 and approximately 2.5 kb of upstream sequence. This fragment was ligated in frame with the eGFP cDNA (730 bp). A 3.5-kb PCR fragment was generated that extended from the last three codons of exon 1, through intron 1 and into the first four codons of exon 2. A rabbit β-globin polyadenylation sequence (561 bp) was ligated to the end of this fragment. The fragments were ligated together to generate the xflk-1 promoter:GFP reporter transgene. This transgene was then cloned into the EcoRV site between the invert/direct repeats (IR/DR) of pT2 SK+ to generate pT2xflk-1:GFP. The entire construct was sequenced to ensure that errors had not been introduced by PCR.

Generation of Transgenic X. laevis

pT2xflk-1:GFP DNA (100 pg) was co-injected with Sleeping Beauty transposase mRNA (50 pg) into fertilized X. laevis eggs at the one-cell stage. Capped RNA was synthesized from the Sleeping Beauty plasmid pSBRNAX using mMessage mMachine kits according to the manufacturers recommendations (Ambion, Austin, TX). Embryos were raised to adulthood and crossed to wild-type animals, and the F1 progeny were scored for GFP fluorescence at stage 25. A total of 31 injected tadpoles, 15 males and 16 females, survived to breeding age. The time to maturity for the males was 7 to 9 months and for the females 12 to 14 months. GFP-positive F1 animals were sorted and raised to adulthood and subsequently outcrossed to wild-type animals and scored for the rate of transmission of the transgene. The same time to maturity and gender ratios (50% male and 50% female) were observed in the F1 and F2 generations of the pT2xflk-1:GFP transgenic frogs that were observed in the founders.

Fluorescence Imaging

A Leica FLIII fluorescence dissecting microscope was used to analyze GFP expression in embryos. Images were captured with a Photometrics CoolSNAP ES camera (Roper Scientific, GMBH) and analyzed with Meta Imaging Series 6.1 software (Universal Imaging Corp.).

Southern Blot Analysis, DNA Walking, and FISH

Genomic DNA was isolated from stage 35 tadpoles by digesting single animals overnight at 60°C in 500 μl of homogenization buffer (100 μg/ml proteinase K, 1% (w/v) sodium dodecyl sulfate (SDS), 10 mM ethylenediaminetetraacetic acid (EDTA), 20 mM Tris pH 7.5, 100 mM NaCl). DNA was extracted with phenol:chloroform:isoamyalcohol (25:24:1 [v/v/v]), precipitated by adding sodium acetate pH 5.2 to 0.3 M and 1 ml of ethanol and centrifuging at 12,000 × g for 15 min. DNA was washed with 70% ethanol, air-dried, and resuspended in 50 μl of water.

Southern blots were performed by digesting 5 μg of genomic DNA with restriction enzymes (BamHI, EcoRI, or EcoRV) overnight at 37°C. DNA was separated on 0.7% (w/v) agarose horizontal gel and transferred to Hybond N+ nylon membrane (Amersham/GE Healthcare, NJ) according to the manufacturer's instructions. The membrane was prehybridized with Church and Gilbert buffer (0.25 mM sodium phosphate buffer pH 7.0, 0.25 mM NaCl, 1 mM EDTA, 7% [w/v] SDS, 5% [w/v] dextran sulfate) at 60°C for 1 hr and hybridized with probe overnight at 60°C. Radiolabeled probe was synthesized from gel purified GFP DNA fragment using the Megaprime DNA labeling system (Amersham/GE Healthcare, Princeton, NJ). Hybridization membranes were washed with 0.2× SSC, 0.1% (w/v) SDS at 55°C, and exposed to X-ray film with intensifying screens at −80°C.

PCR-mediated amplification of transposon flanking sequences was performed using DNA Walking Speedup Premix kits (Seegene, Rockville, MD), and genomic DNA was harvested from F1 transgenic animals according to the manufacturer's instructions. PCR primers were designed within the left arm of pT2 (TSP1; TGGAATTTTCCAAGCTGTT, TSP2; TTCTGACCCACTGGAATTGTGA). PCR products were cloned into the pGEM-Teasy vector (Promega, Madison, WI) and sequenced.

FISH was carried out on F1 tadpoles to detect the GFP transgene. GFP-positive and -negative tadpoles at stage 35 were digested with collagenase (1 mg/ml; Sigma) overnight at 37°C. The cell slurry was centrifuged at 700 × g, resuspended in fixative (75% methanol, 25% acetic acid), mixed at room temperature for 1 hr and a single cell suspension was dried onto glass slides. Plasmid DNA containing the GFP gene (pGem GFP) was labeled with digoxigenin dUTP by nick translation. The slides were hybridized in a solution containing 50% (v/v) formamide, 10% (w/v) dextran sulfate, and 2× SSC. Specific hybridization signals were detected using fluoresceinated anti-digoxigenin antibodies, the slides were then counterstained with 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) and analyzed.

Xenopus Embryo Manipulations, RNA injections, In Situ Hybridization, and Histology

X. laevis embryos were obtained, fertilized, and microinjected as described in Early Development of X. laevis: A Laboratory Manual (Sive,2000). Briefly, female frogs were induced to lay eggs by gonadotropin injection, fertilized in vitro with macerated testis and de-jellied with 3% (w/v) cysteine hydrochloride. Testes were harvested from an F1 transgenic male to fertilize wild-type eggs. Fertilized embryos were injected with mRNA synthesized with mMessage mMachine transcription kit (Ambion). xVEGF RNA (2 ng) was injected into a ventral-vegetal blastomere at the four-cell stage to target the RNA to the posterior flank region of the tadpole. Embryos were incubated at 18°C and monitored daily by fluorescence microscopy. Embryos were staged according to X. laevis normal tables of development (Nieuwkoop and Faber,1994). Natural matings of transgenic pT2xflk-1:GFP male and albino female frogs were used to generate transgenic albino progeny for in situ analysis. In situ hybridization was performed as described by Harland (Harland,1991) using antisense probes to Xenopus flk-1 and GFP. A full-length cDNA encoding X. laevis flk-1 was isolated from an adult spleen lambda phage library using standard procedures (Deconinck et al.,2000). For histology, lungs were dissected from postmetamorphic frogs, fixed in MEMFA (0.1 M MOPS pH 7.4, 2 mM EGTA, 1 mM MgSO₄, 3.7% formaldehyde) and embedded in paraffin. Alternate serial 4-μm sections were stained with either Giemsa or immunostained with a GFP-specific antibody. Immunostained slides were developed using diaminobenzidine and counterstained with hematoxylin.

Acknowledgements

We thank Dr. Perry Hackett for providing Sleeping Beauty plasmids and Dr. Paul Krieg for the Xenopus VEGF cDNA. We thank members of the Meadlab for helpful discussions; St. Jude Cytogenetics Shared Resource for FISH analysis; Dr. Kelli Boyd, Pamela Johnson, Brenda McGowen, and Dorothy Bush for assistance with histology and immunohistochemistry; Kevin Bergeron and Shelby Benson for expert animal husbandry.