The stable transgenesis of genes encoding functional or spatially localized proteins, fused to fluorescent proteins such as green fluorescent protein (GFP) or red fluorescent protein (RFP), is an extremely important research tool in cell and developmental biology. Transgenic organisms constructed with fluorescent labels for cell membranes, subcellular organelles, and functional proteins have been used to investigate cell cycles, lineages, shapes, and polarity, in live animals and in cells or tissues derived from these animals. Genes of interest have been integrated and maintained in generations of transgenic animals, which have become a valuable resource for the cell and developmental biology communities. Although the use of Xenopus laevis as a transgenic model organism has been hampered by its relatively long reproduction time (compared to Drosophila melanogaster and Caenorhabditis elegans), its large embryonic cells and the ease of manipulation in early embryos have made it a historically valuable preparation that continues to have tremendous research potential. Here, we report on the Xenopus laevis transgenic lines our lab has generated and discuss their potential use in biological imaging.
The first reported transgene, developed for mice, was described by Palmiter's group (Brinster et al. 1985). After that, huge numbers of transgenic studies were conducted on several other organisms by microinjecting DNA into fertilized eggs. However, transgenes introduced into Xenopus simply by injecting DNA into a fertilized egg produce only animals with mosaic gene expression. Therefore, researchers looked for a way to integrate genes into the chromosome itself. This became possible for Xenopus in 1996 by combining the restriction enzyme-mediated integration (REMI) method with sperm nuclear transplantation (Kroll & Amaya 1996). This method allowed foreign genes introduced into the Xenopus germ line to be stably expressed and controlled both temporally and spatially by various promoters and made it possible to dissect the functional role of signaling molecules in Xenopus embryogenesis. To study the developmental effects of molecules that cause severe early embryonic defects, genes have been misexpressed under promoters that are active in later development, and this technique is often used in gain-of-function experiments that focus on later developmental events such as organogenesis.
More recent advances in transgenesis include meganuclease I-SceI mediation, which is reported to be efficient for transgenesis in both Xenopus tropicalis (Ogino et al. 2006) and Xenopus laevis (Pan et al. 2006; Ishibashi et al. 2012). Methods mediated by ϕC31 integrase (Allen & Weeks 2005, 2006; Li et al. 2012) and transposons such as Tol2 (Hamlet et al. 2006) and Sleeping Beauty (SB) (Sinzelle et al. 2006) have been developed to complement REMI. These methods for transgenesis in Xenopus are compared in a review by Ogino & Ochi (2009).
Moreover, transgenesis can be used for gene-trapping in Xenopus (Bronchain et al. 1999). Transposon remobilization has great potential for enhancer-trap screens to identify novel gene functions and enhancer activities (Yergeau et al. 2010, 2011), and can be used to analyze the promoter/enhancer activity of non-coding regulatory sequences (Ochi et al. 2012).
Advances in transgenesis methods have accelerated the generation of transgenic Xenopus, and these frogs are being used in more areas of research. Therefore, it is desirable to have a system for sharing transgenic resources. The recently launched European Xenopus Resource Center (EXRC) and the National Xenopus Resource (NXR) (United States) are in charge of maintaining transgenic lines for the Xenopus research community (Pearl et al. 2012). The National BioResource Project (NBRP) for Xenopus tropicalis in Japan, supported by the Ministry of Education, Culture, Sports, Science and Technology, is expected to function as the center for Xenopus research in Asia (http://home.hiroshima-u.ac.jp/~amphibia/xenobiores_en/). These coordinated international activities of the Xenopus research community will enhance our ability to share valuable resources, including transgenic frogs, across national borders.
Here, we introduce some of the Xenopus laevis transgenic lines created in our laboratory and discuss their potential use, particularly for imaging in cell and developmental biology.
Necessity of generating transgenic lines for biological imaging
Xenopus egg- and cleavage-stage embryos are relatively large compared to those of other organisms, 0.8 mm (tropicalis) to 1.2 mm (laevis) in diameter, which is a great advantage when it comes to manipulating eggs and embryos. One manipulation, microinjection, has been particularly important in Xenopus cell and developmental biology; the overexpression of proteins by mRNA or DNA microinjection has helped to elucidate these proteins' molecular functions for several decades. However, mRNAs injected into early cleavage stage embryos at the 2- or 4-cell stage are subject to degradation, and their products gradually decrease as development proceeds. Thus, this method cannot be used to examine the effect of overexpressing or ectopically expressing these proteins in the later stages of organogenesis. DNA microinjected into cleavage embryos is generally maintained extrachromosomally, thus causing mosaicism. In addition, mRNA or DNA injected into a cleaving embryo is not necessarily distributed evenly to the daughter cells, although mRNA diffusion in the cell is thought to be fairly efficient. This uneven distribution and the transient expression of proteins, along with their fused fluorescent proteins, does not provide an ideal platform for quantitatively analyzing protein function—although it is still useful for qualitative analyses, such as identifying the spatial localization and intracellular and extracellular behavior of given proteins by biological imaging. Although transgenesis is time-consuming and laborious compared to mRNA injection, it secures stable gene expression, as well as uniform expression (to the extent that the genes are under the regulation of the appropriate promoters and are transmitted to the germ line). For these reasons, we established Xenopus laevis transgenic lines with markers for several cells and organelles, for potential use in cell and developmental biology.
Generation of transgenic Xenopus laevis
Collection of Xenopus laevis eggs and embryos and maintenance of larvae
Xenopus laevis were purchased from local suppliers and maintained at 18°C. Adult females were primed with 30 units of pregnant mare serum gonadotropin (PMSG) (ASKA Pharmaceutical, Tokyo, Japan) 3–5 days prior to egg-laying, and boosted with 400 units of human chorionic gonadotropin (hCG) (ASKA Pharmaceutical) about 15 h prior to egg-laying. Embryos were staged according to Nieuwkoop & Faber (1994). The transgenic tadpoles were fed with Sera Micron (Sera Japan, Kanagawa, Japan). The animals were handled in accordance with the guidelines of the Center for Experimental Animals of the National Institute of Natural Sciences at Okazaki.
Promoters and fluorescent proteins
The constructs used to generate our transgenic frogs are summarized in Table 1. These constructs were placed under the control of the cytomegalovirus (CMV) promoter or the CAG promoter, which is a combination of the CMV early enhancer element and chicken β-actin promoter (Niwa et al. 1991). In many cases, the CAG promoter seemed to drive stronger transgene expression than CMV promoter in Xenopus embryos (Sakamaki K., unpubl. data, 2001). For fluorescent proteins, we used enhanced GFP (EGFP), RFP, a yellow fluorescent protein (YFP) variant (Venus) (Nagai et al. 2002), or a super-enhanced cyan fluorescent protein (seCFP) (Kremers et al. 2006). We found no significant difference of toxicity among these fluorescent proteins but did observe that some transgenic embryos and larvae display visible aberrant phenotypes often associated with cell death. The extent of these effects seems to vary depending on transgenic lines rather than the difference of fluorescent proteins, suggesting that the toxicity may largely depend on the copy number and the site of integration.
Table 1. List of transgenic lines
Transgene (Fluorescent protein)
Underlines denote fluorescent proteins used. CALR, calreticulin; CMV, cytomegalovirus; COX8, cytochrome c oxidase 8; EB3, end-binding protein 3; EGFP, enhanced green fluorescent protein; LBRS, Lamin B receptor short form; RFP, red fluorescent protein; seCFP, super enhanced cyan fluorescent protein.
A simplified REMI method
Sperm nuclei were isolated as described by Kroll & Amaya (1996) with one modification: we used digitonin (Wako Biochemicals, Kyoto, Japan) instead of lysolecithin. Recently reported protocols for Xenopus transgenesis have simplified the original procedure by omitting the use of restriction enzymes, which is similar to intracytoplasmic sperm injection (ICSI) for generating transgenic mice (Perry et al. 1999; Sparrow et al. 2000; Smith et al. 2006). Although the mechanism of gene integration is unclear (Chesneau et al. 2008), in the course of several transgenesis trials, we found no significant difference in the efficiency of gene integration with or without restriction enzymes; integration efficiency ranged from 9.6% to 38.5% of animals surviving beyond the tadpole stages (Table 1). Therefore, for this series of studies, the sperm nuclei were incubated with the linear transgene and egg extract, without restriction enzymes, and the mixture was injected into unfertilized eggs dejellied in 2.5% cysteine (pH7.8) in 1× Marc's modified ringer (MMR). Sperm nuclei-transplanted embryos were selected at the 4-cell stage. At the tadpole stage, embryos expressing the transgene were screened and designated F0 founder animals. These were crossed with wild-type frogs or other transgenic frogs to obtain F1 transgenic animals.
Table 1 lists transgenic Xenopus laevis lines we have established in the past 10 years (this is also summarized in our website at http://www.nibb.ac.jp/morphgen/frog/index.html). We combined specific promoters and fluorescent proteins to express genes in various membranes and organelles of Xenopus laevis embryos and larvae.
Observation of the transgenic Xenopus embryos was performed with a laser-scanning confocal microscope (LSM510, Carl Zeiss, Oberkochen, Germany), a stereo fluorescence microscope (SZX12, Olympus, Tokyo, Japan), or by digital scanned light-sheet microscopy (DSLM) (the original was invented by Dr Enrst Steltzer et al. at the European Molecular Biology Laboratory at Heidelberg, Germany; we used a replica assembled at the National Institute for Basic Biology by Drs. Nonaka and Ichikawa) (Keller et al. 2008) as described previously (Morita et al. 2010, 2012). Fluorescence of tissues, tadpoles, and frogs in Venus and SCAT3 transgenic lines was observed as described previously (Sakamaki et al. 2005; Kominami et al. 2006).
Marking the cell membrane with membrane-targeted EGFP/RFP
The cytoplasmic membrane faithfully represents the morphology of the cell, and provides a substantial amount of information on cell type, status, and protrusive activities (Lecuit & Lenne 2007). In particular, the epithelium of the Xenopus embryo, which normally shows a two-dimensional pebbled pattern as a sheet of cells, often undergoes cell shape changes in response to various intra- and extracellular stimuli. Therefore, the analysis of cell contours over time is useful for understanding both the cytoskeletal dynamics of these cells and the mechanical influences to which they are exposed. In contrast to the surface epithelia, fluorescent imaging of the deeper tissues and structures are generally much more difficult because it is hampered by the diffraction of yolk proteins in the cells and the dark pigmentation caused by melanin.
To label the cell membrane that highlights its contours, c-Ha-Ras, a protein that bears a farnesylation signal at the carboxyl-terminus, is often fused with GFP or RFP. Ras proteins that are implicated in cancer can be localized to the plasma membrane's cytoplasmic face by a series of post-translational modifications, including the addition of the 15-carbon isoprenyl group farnesyl to cysteine in the Ras carboxyl-terminal CAAX box (where C is cysteine, A is aliphatic, and X is typically Met or Ser) by the enzyme farnesyl protein transferase (Reiss et al. 1991). Therefore, adding the CAAX box to fusion proteins can force them to localize to the plasma membrane (see example in Fig. 1 A–C).
Figure 1B shows a stage 13 embryo used in the analysis of cell shape changes during neurulation (Morita et al. 2012). Figure 1C (bottom) shows the tail region of the offspring of an EGFP-Mem transgenic frog crossed with a transgenic frog expressing H2B-RFP (Fig. 1C shows GFP labeling; see Fig. 2B for H2B-RFP). EGFP targeted to the plasma membrane clearly indicates the pebbled pattern of the epithelium. The EGFP-Mem frog and the similarly generated RFP-Mem frog were also useful for quantitative analyses of cell contour, aspect ratio, and other facets of cell morphology, which were performed by a combination of live imaging and the subsequent segmentation of the images obtained. In concept, it would be possible to visualize the boundary structure between two populations of cells if these populations were labeled with two fluorescent proteins under the regulation of their respective promoters.
We have often observed the ectopic accumulation of EGFP fluorescence in addition to the authentic localization of fusion proteins. This may represent the aggregation of EGFP proteins overproduced by the high copy numbers of DNA integrated to the genome. To avoid this artifact, transgenic lines with appropriate expression levels should be kept separate from those with abnormally high expression levels, and the better lines selected out by crossing them with wild-type frogs over several generations (Soroldoni et al. 2009).
Marking the nucleus and nuclear membrane
Since the nucleus is spatially confined to the middle of the cell, it makes a good cell position marker. The dynamic behavior of histone 2B (H2B), a conserved nuclear protein that forms the eukaryotic nucleosome core together with H2A, H3, and H4 (Wells & Brown 1991), is often used to visualize nuclei and chromosome dynamics during cell division. It is particularly useful for marking nuclei to track cells in a tissue or in an embryo undergoing drastic three-dimensional (3D) reorganization, which is rather difficult to grasp through snapshot observations of the cells themselves. In previous studies, H2B transgenic animals have been successfully used to track cell relocation (Kurotaki et al. 2007; Keller et al. 2008) and its 3D analysis (Bashar et al. 2012). Figure 2B shows the fluorescence signal of nuclei in the tail epithelial cells of an H2B-RFP transgenic tadpole. Fluorescent labeling of the nuclear membrane using Lamin B receptor (LBR) fused to EGFP (Haraguchi et al. 2000) is useful to observe the dynamics of nuclear membrane. We have generated transgenic frogs expressing LRBS, a Lamin B receptor short form containing only the first transmembrane span of human LBR (amino acids 1–238) (Ellenberg et al. 1997) fused with EGFP (Fig. 2C). We have found this probe can be used to visualize nuclear envelope breakdown and re-assembly during cell division in live embryo (Movie S1).
Marking ER, Golgi, and mitochondria
Visualizing intracellular organelles is useful, because their position and morphology are closely related to cellular functions. Endoplasmic reticulum (ER) is a membranous or cisternous network connected to the outer membrane of the nucleus. Smooth ER is involved in lipid and steroid synthesis, carbohydrate metabolism, and the release of calcium, while rough ER, to which ribosomes are attached, is essential for synthesizing secreted and membrane proteins. Calreticulin (CALR), a 60 KDa calcium-binding protein, binds to misfolded proteins to prevent their export from the ER to the Golgi apparatus. CALR is localized to the membrane through characteristic amino acid sequences, the ER-targeting (Fliegel et al. 1989) and retrieval sequence KEDL (Lys-Glu-Asp-Leu) (Munro & Pelham 1987; Pelham 1996). Therefore, fluorescent proteins harboring both sequences are used to target the ER. We constructed the transgene CMV-CALR-RFP-KEDL, composed of DNA encoding an RFP fusion protein with the ER-targeting sequence at the 5′ end and KEDL at the 3′ end (Fig. 3Aa). As shown in Figure 3B, this produced RFP signals in the ER membrane and lumen in a transgenic animal.
The Golgi apparatus, juxtaposed to the ER, is an organelle forming a stack of flat, pouch-like membranous structures, and it is essential for glycosylating secreted proteins and processing ribosomal proteins. As it mediates directed vesicle transport from the ER to extracellular space, its positioning is thought to have an active role in directed secretion, cell polarity, and wound healing (Yadav et al. 2009). Interestingly, certain extracellular stimuli are known to induce repositioning of both the Golgi apparatus and the microtubule-organizing center (MTOC) (Pu & Zhao 2005). Conventionally, Golgi apparatus was visualized by immunostaining of GM130. For capturing dynamics of Golgi apparatus by live imaging, however, fluorescence marking of Golgi markers, such as mannosidase II-GFP, galactosyltransferase (GalT)-GFP, and KDEL receptor-GFP serves as a powerful tool and is useful in cell biology (Cole et al. 1996).
In this study, we used β1,4-galactosyltransferase (β4GalT) to make a chimera with GFP (Fig. 3 Ab). Although β4GalT is also localized to the cell membrane, it has been widely used to mark the Golgi apparatus in cultured cells (Strous 1986), since visualization of the trans-Golgi network (TGN) in live cells can serve as a real-time indicator of functional cell polarity. For example, as shown in Figure 3C, crescent-shaped fluorescence representing the TGN is visible in each tailbud epithelial cell. Higher magnification of the fluorescent cells should allow the live imaging analysis of cell polarity, based on Golgi localization, as well as morphological changes in the TGN and membrane traffic pathway during embryogenesis or in response to temperature, treatment with drugs, and other environmental changes. A similar approach was previously taken to achieve the real-time analysis of membrane traffic pathways in normal rat kidney epithelial (NRK) cells stably expressing a GFP fusion protein with TGN38, a type I integral membrane protein, which cycles between the trans-Golgi network (TGN) and cell surface (Girotti & Banting 1996).
Mitochondria, organelles found in most eukaryotic cells, are essential for synthesizing adenosine triphosphate (ATP) as an energy source and play an important role in apoptosis (Green & Kroemer 2004). Mitochondria are surrounded by two lipid membrane layers, and the sorting of newly synthesized precursor proteins to the mitochondria depends on the mitochondria-targeting sequence (MTS) (reviewed in Ohmura 1998), which is a long repetitive sequence, 20–40 amino acids in length, consisting of a few hydrophobic amino acids followed by a basic amino acid. The MTS is located at the amino termini of the precursors. Therefore, fusing fluorescent proteins to the MTS, as indicated in Figure 3Ac (which shows EGFP fused to the MTS derived from subunit VIII of human cytochrome c oxidase [COX8]), directs them to the mitochondria (Manfredi et al. 2002) (see the fluorescent image in Fig. 3D). Similarly, mitochondrial labeling with EGFP fused to the MTS from a mitochondrial outer membrane protein, OMP25 (Horie et al. 2002), was recently reported; the authors successfully labeled and traced the cells to investigate the behavior of the mitochondria-rich germ plasm during Xenopus development (Taguchi et al. 2012). The number and morphology of mitochondria change with the cell cycle, apoptosis, and neurodegenerative diseases, and therefore quantitative analyses of mitochondria with these labeling tools are useful for monitoring mitochondrial functions during Xenopus development.
Microtubules: Marking plus-ends and analyzing dynamics
Like actin fibers (F-actin), microtubules are one of the most essential components of the cytoskeleton; they not only maintain cell shape, but also control cell polarity and cell motility. A microtubule is composed of two subunits of tubulin, alpha and beta, which form a cylindrical tube that provides a “track” for the intracellular transport of cell organelles and vesicles, mobilized by motor proteins such as the kinesins and dyneins. Microtubules arranged into a spindle shape are also essential for separating chromosomes during mitosis, and they are used as both a structural component of eukaryotic flagella and cilia and in the locomotion of single-celled animals. They display dynamic instability at their growing (plus) end, revealed as their assembly and disassembly.
To visualize the functional polarity generated within cells of live Xenopus embryonic tissue, we adopted end-binding (EB) proteins as a marker for microtubule growth (Fig. 4A) (Shindo et al. 2008, 2010). As the EB proteins EB1 and EB3 are known to bind preferentially to the plus end of growing microtubules, these proteins can indicate spindle formation during cell division (Fig. 4D) and indicate the orientation of microtubule growth, which represents the process of cell polarity being established (Stepanova et al. 2003). Although yolk proteins and pigmentation in early embryos make it difficult to visualize microtubule dynamics with a fluorescent probe, especially in deep-lying cells, we were able to observe such epithelial cell dynamics in the tailbud of an EB3-EGFP transgenic embryo (Fig. 4B,C, and Movie S2).
Labeling tissues and organs
Besides expressing functional proteins fused to fluorescent proteins, fluorescent proteins can be expressed alone to label and visualize cells and tissues. Using this, whole-body GFP expression has been accomplished in both Xenopus laevis and Xenopus tropicalis (Bronchain et al. 1999; Marsh-Armstrong et al. 1999; Hirsch et al. 2002); such frogs are useful for studying organogenesis during larval stages, when they are transparent. For detailed investigations at the cell and tissue levels, however, GFP observation is often hampered by endogenous fluorescence that is not filtered out. Therefore, it has been necessary to create transgenic lines with brighter exogenous fluorescence. Recently, many types of genetically engineered GFP mutants and variants have been created that emit blue, cyan, or yellow fluorescence. Among these, Venus displays the strongest fluorescence due to its rate-limited and highly efficient maturation (Nagai et al. 2002), and thus offers more sensitive and convenient detection of labeled cells and tissues.
To take advantage of this feature, we generated a transgenic line expressing Venus under the control of the CAG promoter (Fig. 5A) (Sakamaki et al. 2005). The resulting transgenic tadpoles and frogs expressed the Venus protein at significant levels throughout the body (Fig. 5B). Further analysis showed highly detectable fluorescence in most tissues throughout development and in the germ cells of both sexes (Sakamaki et al. 2005). As shown in Figure 5C–E, fluorescently labeled transgenic brain fragments derived from larval stages and transplanted into wild-type larvae were easily detected in the recipient's brain one and 3 weeks later. In addition, a transplanted Venus-positive primary spermatogonium was readily detected in a recipient's testis (Kawasaki et al. 2006). Similarly, it was possible to monitor the differentiation of animal pole cells, isolated from Venus-positive embryos at blastula stage, into eyes when they were transplanted to the eye field or flanks of stage 15 embryos and observed at larval stages (Viczian et al. 2009). Thus, we have established a more suitable transgenic frog line for monitoring marker cells in living animals.
Monitoring apoptosis during Xenopus development
Normal development requires not only cell proliferation and differentiation, but also a highly regulated process of programmed cell death, apoptosis, that removes cells by activating a cell-death pathway (Jacobson et al. 1997). Apoptosis is mainly executed by the caspase enzyme (Kumar 2007). Visualization of apoptosis in live animals has been a strong research interest for cell and developmental biologists. Recent attempts to use fluorescence resonance energy transfer (FRET) to monitor biological phenomena in living cells has led to the development of many FRET-based artificial molecules (reviewed in Chudakov et al. 2010). A FRET-based biosensor, SCAT3, was generated to monitor caspase-3 activation during apoptosis (Fig. 6A). It has been shown that the separation of seCFP and Venus occurs by proteolytic cleavage of the linker portion via activated caspase-3, resulting in the decline of the FRET efficiency in dying cells (Takemoto et al. 2003). Therefore, this biosensor allows us to visualize caspase activation in intact cells undergoing apoptosis by just monitoring the fluorescence of seCFP and Venus. Transgenic flies and mice expressing SCAT3 to identify the correlation between caspase-3 activation and cell death have been generated and they are used to detect cell death associated with caspase activation during their development (Takemoto et al. 2007; Yamaguchi et al. 2011). Based on these, we have generated transgenic frogs expressing SCAT3 (Fig. 6B) (Kominami et al. 2006) to monitor caspase-3 activation during organogeneis in Xenopus laevis. In the previous study, we have shown that a truncated Xenopus homologue of human Bid, tBid, induces apoptotic cell death in mammalian culture cells (Kominami et al. 2006). Here, we confirmed that Xenopus tBid induces cell death associated with caspase-3 activation in embryos when ectopically expressed (Fig. 6C and Fig. S1). In these embryos, we observed the decrease of the FRET ratio of SCAT3 in dying cells but not living cells (Fig. 6D,E). We also confirmed the ratiometric changes in fluorescence is caused by cleavage of SCAT3 (Fig. 6F). Using this probe, we found the in vivo apoptosis in the primary mouth of developing Xenopus (Sakamaki K., unpubl. data, 2012), which is thought to be involved in the thinning of the ectoderm and the subsequent mouth opening, can be monitored by SCAT3, confirming data obtained earlier using the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method (Dickinson & Sive 2006). Furthermore, we have used the transgenic embryos to verify the function of a recently identified Xenopus homologue of human FADD, and confirmed its strong pro-apoptotic activity by detecting the cleavage of SCAT3, namely the caspase-3 activation in developing embryos (Sakamaki et al. 2012). Thus, this established transgenic frog line is useful for analyzing apoptosis in live animals.
We would like to thank Dr Tokuko Haraguchi for providing the LBR-GFP plasmid, Dr Anna Akhmanova for the EB3-GFP plasmid, and Mr Akira Miyakoshi for constructing the plasmids for EGFP-Mem, RFP-Mem, H2B-Venus, H2B-RFP, RFP-ER, RFP-Golgi, and RFP-Mito. We are also grateful to Dr Iain Mattaj for helpful discussions regarding expressing LBR in the nuclear envelope, to Dr Kazuhito Takeshima for technical guidance on transgenesis in Xenopus laevis, to Dr Takehiko Ichikawa for observation of the transgenic embryo expressing EB3-EGFP with DSLM, and to Ms Michiyo Murakami for maintaining our frog colony. Confocal images were acquired at Spectrography and Bioimaging Facility, NIBB Core Research Facilities.