The use of live imaging approaches to examine and understand the dynamic processes that take place during mouse development has become widespread. Several groups have reported their success in generating different reporter mouse lines that express a variety of fluorescent markers for imaging. However, there is currently no established database of the reporter mouse lines available for live imaging, such as the Cre transgenic lines (Cre-X-Mice). Researchers therefore often have difficulties in determining which reporter mouse line meets their research purposes. In this review, we summarize some of the reporter mouse lines that have been generated for live imaging studies, and discuss their characteristics.
Live imaging is a powerful tool that can be used to understand embryonic development. One popular animal model used for the study of mammalian development is the mouse because of the many genetic modification techniques that have been established for this animal (Nagy 2003). The first successful reports of generating transgenic mice using pronuclear micro-injection (Gordon et al. 1980) or using gene manipulating technology with embryonic stem (ES) cells came in the 1980s (Evans & Kaufman 1981; Martin 1981; Smithies et al. 1985; Gossler et al. 1986; Robertson et al. 1986; Doetschman et al. 1987; Thomas & Capecchi 1987; Thompson et al. 1989). These technologies laid the foundation for researchers to analyze the function of genes by generating gain-of-function and loss-of-function mutations in the mouse genome. Initially, gene expression was often traced using LacZ (beta-galactosidase of Escherichia coli) as a reporter, but one drawback was that LacZ staining could only be performed in fixed samples.
The next major development was the successful cloning of the gene encoding green fluorescent protein (GFP; Prasher et al. 1992) 30 years after its discovery (Shimomura et al. 1962), which was then followed by a report demonstrating the utility of GFP as a fluorescent labeling marker in vivo in E. coli and Caenorhabditis elegans (Chalfie et al. 1994). The Green mouse, a transgenic mouse that expresses GFP, was subsequently generated in 1997 (Okabe et al. 1997). Since then, fluorescent proteins have become widely used to develop different methods for visualizing cells and their behaviors in the mouse and other living animals. And we now have reporter mice that can express fluorescent proteins to mark characteristics of cells in vivo, such as gene expression, cell cycle progression, and localization of certain subcellular structures in live cells.
Before the introduction of fluorescent proteins, gene products of interest were examined by histological observation in fixed samples, but this type of analysis is not conducive for continuously observing phenomena that change dynamically in animals. The concomitant developments in genetic engineering, fluorescent technology, culturing techniques, and microscopy have now provided us with the tools to observe biological phenomena in living animals; in particular, the development of imaging equipment with built-in, finely controlled incubation systems have greatly facilitated live imaging work. Early stages of mouse embryos are suitable for live imaging because of their transparency and their established culturing methods. Many developmental studies in mouse, such as on blastocyst axis determination, anterior visceral endoderm cell migration, notochord formation, and primitive streak formation, have been performed using live imaging techniques (Srinivas et al. 2004; Kurotaki et al. 2007; Yamanaka et al. 2007; Williams et al. 2012). In this review, we summarize the different reporter mouse lines established to observe embryonic development and physiology of adult mice for various studies.
General design of reporter mouse lines expressing exogenous genes
There are two main ways to introduce exogenous genes into the mouse genome: random transgenesis, and targeted transgenesis via homologous recombination. To introduce exogenous DNA by random transgenesis, DNA fragments with promoter-cDNA cassette are either micro-injected directly into zygotes or transfected to ES cells (Gordon et al. 1980; Hadjantonakis et al. 2002). The main limitation associated with random transgenesis is the unpredictable variation of transgene expression due to effects of the genomic integration site or transgene copy number. Furthermore, transgenic mice that exhibit high expression levels of transgenes are often infertile or not viable (Stewart et al. 2009). In the ES cell transfection method, which is called embryonic stem cell-mediated transgenesis, the expression level and copy numbers of the introduced transgenes can be verified in the ES cells first before the cells are transplanted into the recipient embryos to establish the mouse lines (Hadjantonakis et al. 2002). Recently, lentiviral-mediated transgenesis and transposon-mediated transgenesis in zygotes were also developed, which introduces transgenes with high efficiency and only inserts a single copy in the genome (Lois et al. 2002; Sumiyama et al. 2010; Marh et al. 2012).
For random transgenesis to succeed, choosing an appropriate promoter is one important factor. The most commonly used promoter to ubiquitously express a transgene is the combination of the chicken beta-actin promoter and the cytomegalovirus enhancer (CAG; Niwa et al. 1991); however, CAG promoter-driven transgenes are sometimes silenced and/or show non-ubiquitous transgene expression (Rhee et al. 2006; Griswold et al. 2011). For example, the CAG promoter drives higher levels of transgene expression in heart and neural tissues (Toyoda et al. 2003). Other promoters that are also used for transgenesis to induce widespread expression are the human ubiquitin C (UBC) promoter (Schorpp et al. 1996; Fink et al. 2010) and the ROSA26 promoter (Kisseberth et al. 1999), with the transgene expression levels under the UBC promoter being higher than that under the ROSA26 promoter. Because these two promoters are derived from an endogenous gene, the expression activities are lower than that of the CAG promoter, which includes a sequence derived from virus (Chen et al. 2011). Recent publications also demonstrate that the UBC promoter and ROSA26 promoter with genomic insulators show a more ubiquitous expression of the transgene than the CAG promoter (Griswold et al. 2011; Abe et al. 2013).
It has now become possible to introduce conditional genome alterations that are spatially and temporally restricted by combining the Cre/loxP site-specific recombination system with random transgenesis technologies (Sternberg 1981; Lakso et al. 1992). The CAG-CAT-Z (chloramphenicol acetyltransferase/lacZ) construct is a well-known basic reporter for conditional transgene expression, and uses a double-reporter system to provide a precise and accurate assay for Cre excision at the cellular level (Araki et al. 1995). The cells express CAT before Cre excision, but lose CAT expression after Cre excision, and instead express LacZ (Fig. 1). Many of the other conditional transgenes use the principles of this construct (Lobe et al. 1999).
On the other hand, we can generate knock-in reporter mouse lines with precisely designed genome modifications through homologous recombination in ES cells, using the ROSA26 locus to insert reporter genes. The ROSA26 locus was first identified during a gene trap screening in ES cells (Friedrich & Soriano 1991). This locus shows ubiquitous transcriptional activity, and the loss of gene function in this locus is not lethal (Zambrowicz et al. 1997). This locus has also been established as the preferred insertion site for ubiquitous expression of transgenes during embryonic development (Soriano 1999). Soriano (1999) successfully established a reporter line for monitoring Cre recombinase activities at the ROSA26 locus (Fig. 2). A Neo expression cassette flanked by loxP sites and followed by the lacZ gene and polyadenylation (bpA) sequence were inserted at a unique XbaI site approximately 300-bp 5′-upstream of the original gene-trap integration site. Triple repeats of polyadenylation sequences were added to the 3′ end of the Neo expression cassette to prevent transcriptional read-through. To generate ROSA26 locus knock-in mouse, the pROSA26PA and pBigT vectors were developed to facilitate the insertion of the reporter cDNA with the loxP-flanked stop sequence into the ROSA26 targeting vector (Srinivas et al. 2001). The ROSA26 targeting vector was further improved by using the Gateway system (Hohenstein et al. 2008; Abe et al. 2011), which can insert the cDNA fragments into the targeting vector by λ phage recombinase-mediated transfer (Hartley et al. 2000). The sequence encoding a reporter, such as a fluorescent protein, can easily be inserted into the targeting vector within a few days, and the reporter gene expression directly controlled by the endogenous ROSA26 promoter. The transcriptional activity of this locus is weaker than that of exogenous artificial promoters such as CAG promoter (Chen et al. 2011), with some reporter expression levels being below detectable levels (Abe et al. 2011). Furthermore, the activity of the ROSA26 locus tends to be suppressed in some tissues such as in the adult brain (Madisen et al. 2010). Therefore, to enhance expression activity in the ROSA26 locus, a CAG promoter is often used in knock-in reporter lines (Fig. 2; Madisen et al. 2010; Snippert et al. 2010; Tchorz et al. 2012). Insertion of the CAG promoter into the ROSA26 locus produces an 8- to 10-fold stronger transgene expression in ES cells than with the ROSA26 promoter alone (Chen et al. 2011). The use of the ROSA26 locus for the Cre/loxP based recombination/conditional expression system is also a powerful tool to express reporters in a spatio-temporally specific manner. Because the copy number of reporter cassettes is strictly controlled in the homologous recombination-mediated insertion, the Cre-mediated excision is carried out precisely as designed. The β-actin genomic locus, which encodes a housekeeping gene, can also be used for conditional transgene expression (Jagle et al. 2007). Although homozygous disruption of the β-actin gene results in embryonic lethality (Jagle et al. 2007), a recent study reported success in establishing transgenic mice by inserting a transgene immediately downstream of the polyadenylation site of the β-actin gene (Tanaka et al. 2012).
Fluorescent proteins in reporter mouse
Green fluorescent protein and many of its spectral variants have been developed and widely used as genetically encoded indicators in cell biological applications, such as for investigating protein expression, localization, and interactions (Tsien 1998; Chudakov et al. 2010). The Green mouse, which expresses enhanced green fluorescent protein (EGFP) driven by a CAG promoter, was first established by Okabe et al. (1997) (Table 1, Native FPs). The Green mouse showed widespread EGFP expression throughout its body, with the exception of hair and red blood cells, and had no apparent defects (Okabe et al. 1997). The successful establishment of this transgenic mouse line suggested that EGFP expression is non-toxic in mouse. Spectral variants of the green fluorescent protein are now widely used in live animals for a broad range of applications as seen in the other papers in this issue. In the mouse, enhanced yellow (EYFP) and enhanced cyan (ECFP) fluorescent proteins serve as valuable tools for imaging (Hadjantonakis & Nagy 2000; Srinivas et al. 2001).
The emission spectra of GFP variants (YFP and CFP) are relatively close, making it difficult to visually differentiate between them with readily available imaging systems (Feng et al. 2000), therefore easily identifiable spectrally distinct colors, such as red, were much sought after. The first red fluorescent protein (RFP) isolated was DsRed, which was isolated from Discocoma sp. (Baird et al. 2000). Hadjantonakis and colleagues then tried generating a transgenic mouse that expressed DsRed, but were unsuccessful in establishing this line; they suggested that DsRed was not developmentally neutral or that constitutive transgene expression could not be maintained (Hadjantonakis et al. 2002). Improvements were made to DsRed, which had slow maturation times and poor solubility (Baird et al. 2000), to generate the mutant DsRed2 (Verkhusha et al. 2001). Additional modifications through random mutagenesis produced a further improved variant, DsRed.T3 (Bevis & Glick 2002). Vintersten et al. (2004) investigated the versatility of DsRed.T3 in mouse ES cells and in living animals to create a transgenic reporter mouse line (Z/RED; Table 1, Native FPs). Z/RED comprises a CAG promoter upstream of a loxP-flanked βgeo (beta-galactosidase and nenomycyin fusion), followed by three polyadenylation sites. This transgenic mouse expresses βgeo as a default, and upon Cre-mediated recombination, βgeo is removed, which results in the DsRed.T3 being attached directly to the transcriptional control region of the promoter. These transgenic mice developed normally and were fertile. DsRed.T3 expression was also transmitted to their offspring at expected Mendelian ratios. However, DsRed.T3 was still unsuitable to create a fusion protein because it formed multimers. The first actual monomeric RFP, monomeric RFP 1 (mRFP1), was generated from the DsRed sequence by Campbell et al. (2002). Long et al. (2005) later examined the expression of native mRFP in ES cells and also its germline transmission. They found that the CAG::mRFP1 transgene (Table 1, Native FPs), a construct in which mRFP1 is driven by a CAG promoter, was transmitted to F1 offspring in Mendelian ratios, suggesting that mRFP1 expression in a wide range of tissues is compatible with normal development and fertility. Moreover, they succeeded in obtaining homozygous CAG::mRFP1 transgenic mice. This led the way for researchers to use mRFP1 in mice, both in its native form and as a part of a fusion protein. There are now many monomeric RFPs available, which have been improved from DsRed or other fluorescent proteins, and they are now widely used in biology (Chudakov et al. 2010) and transgenesis in mouse. Among them, the one that shows the most potential is monomeric (m)Cherry (Shaner et al. 2004). mCherry is brighter, matures faster, and has higher photostability than mRFP1. In addition, it is codon-optimized and better suited for making N-terminal fusion proteins. Several tissue-specific and ubiquitous mCherry transgenic mouse lines have been generated and demonstrated their functionality (Larina et al. 2009; Poche et al. 2009; Sunmonu et al. 2009; Armstrong et al. 2010; Fink et al. 2010). However, it is unclear whether or not mCherry has little or no toxicity, because some transgenic animals expressing mCherry fused with Histone H2B show infertility when they get older.
To visualize a specific cell lineage or specific cell population in mouse embryos, in addition to Cre/loxP systems, photoactivatable fluorescent proteins are also a powerful tool to use. Many different photoactivatable fluorescent proteins have been developed (Ando et al. 2002; Patterson & Lippincott-Schwartz 2002; Chudakov et al. 2004; Tsutsui et al. 2005; Gurskaya et al. 2006; Subach et al. 2009), and are now used widely in research (Fig. 3). For example, three independent groups have established transgenic mouse lines that express Kikume Green-Red (KikGR; Kurotaki et al. 2007; Nowotschin & Hadjantonakis 2009; Griswold et al. 2011; Table 1, Photoactivatable FPs). KikGR is a natively green fluorophore, which can be activated to turn red upon exposure to UV light (360–410 nm; Tsutsui et al. 2005). An iUBC-KikGR mouse was generated by Griswold et al. (2011), and this transgenic mouse showed near-ubiquitous transgene expression throughout the embryonic stages and in adult tissues. They examined cell movements of the anterior visceral endoderm around stage E5.5 by time-lapse imaging, and also examined the events in organogenesis at stage E11.5. Even after culturing the organs for 15–24 h, the photoactivated signals could still be detected in most organs. By contrast, they had difficulties tracking photoactivated signals in the lung epithelium, which was likely due to dilution of the photoactivated protein during rapid cell division. In pre-implantation stages, Kurotaki et al. (2007) exposed a single two-cell blastomere of CAG-KikGR-1 transgenic mouse with UV light, and saw that the photoactivated proteins were distributed throughout the whole blastomere. These red photoactivated proteins in the blastomere were visible for approximately 3 days. These data together suggest that the amount of photoactivated red proteins and/or the frequency of cell divisions are important factors affecting the length of time the proteins can be detected.
A Kaede transgenic mouse (Table 1, Photoactivatable FPs), which expresses photoactivatable fluorescent protein Kaede under the CAG promoter, has been also generated, but the developmental features of these mice have not been fully explored yet (Ando et al. 2002; Tomura et al. 2008).
In a study comparing the use of photoactivatable fluorescent proteins PA-GFP (Patterson & Lippincott-Schwartz 2002), PS-CFP2 (Chudakov et al. 2004), Kaede, and KikGR in ES cells, KikGR appeared to be most conducive for cell labeling and lineage tracing studies in ES cells because the protein is bright, is developmentally neutral, and undergoes rapid and complete photoactivation (Nowotschin et al. 2009).
Visualization of subcellular targets
The introduction of genetically encoded fluorescent proteins in mouse transgenesis has enabled us to observe cell behaviors in living embryos and animals. As discussed below, fluorescent proteins can also be used to visualize different subcellular structures when they are fused to specific localization signal sequences (Table 1).
Nucleus: Reporters for cell lineage tracing
Cells expressing a native fluorescent protein can be visualized using microscopy, and cell morphology can also be identified when this protein is expressed in a single cell. However, cell behavior of a single cell can be difficult to determine when the same fluorescent protein is expressed in a group of cells, because native fluorescent proteins are distributed throughout the cytoplasm, and cell–cell boundaries are not clear (Table 1, Native FPs and Fig. 4, Native FP). Past studies have shown that human histone H2B, a nuclear marker, can be tagged with a fluorescent protein and be incorporated into chromatin in the nucleus without causing any adverse effects on the viability of cells in culture, which makes it possible to identify a single cell and also track the behavior of individual cells within a group. This nuclear marker can also be used to visualize cell division and cell death (Table 1 Nucleus and Fig. 4, Nucleus marker; Kanda et al. 1998). Several groups have generated reporter mice expressing H2B-tagged fluorescent proteins. The nuclear dynamics and the various mitotic phases of live ES cells and developing embryos were then captured by using time-lapse imaging, which showed these embryos are viable and the mice are fertile with no overt morphological abnormalities even with widespread nuclear marker expression (Hadjantonakis & Papaioannou 2004; Kurotaki et al. 2007; Meilhac et al. 2009; Abe et al. 2011).
In their efforts to make H2B-tagged red fluorescent fusions, Hadjantonakis & Papaioannou (2004) tried to produce H2B-DsRed2 and H2B-DsRedExpress ES cell lines. While they did recover a few lines with H2B-DsRed2 and H2B-DsRedExpress expression, they found that long-term maintenance of these cell lines in culture resulted in a continued reduction and heterogeneity in fluorescence. They were unable to establish mouse lines with sustained homogenous H2B-DsRed2 or H2B-DsRedExpress fluorescence. mRFP1 is an improved protein variant derived from DsRed to overcome the tetramerization and sluggish maturation seen in DsRed (Campbell et al. 2002), and is amenable for use in mice (Long et al. 2005). However, the photostability of mRFP1 decreases during its genetic modification. Shaner et al. (2004) developed a few monomeric red fluorescent protein derivatives from mRFP1, namely mStrawberry, mRuby, and mCherry. mCherry was not as bright as mStrawberry and mRuby, but it showed greater photostability than the other two. Furthermore, mCherry-α-tubulin fusions were successfully incorporated into microtubules of most cells, similar to results seen with GFP-coupled tubulin (Shaner et al. 2004). Thus, H2B-mCherry is now the most commonly used H2B-tagged red fluorescent fusion, and reporter mouse lines that constitutively express H2B-mCherry are also available (Abe et al. 2011). One reporter line, R26-H2B-mCherry has been used successfully for the time-lapse recording of embryonic development at pre-implantation stage and post-implantation stage (Abe et al. 2011, 2013).
Membrane: Reporters for live cell morphology
The fluorescent fusion proteins that localize to the plasma membrane provide information on the membrane dynamics and cell morphology within a population (Table 1, Membrane and Fig. 4, Membrane marker). Rhee et al. (2006) reported generating and characterizing two transgenic mouse lines that exhibit widespread expression of lipid-modified GFP-variant fusions (GFP-GPI and myr-Venus), which target the respective fluorescent proteins to the outer and inner leaflet of the plasma membrane. These transgenic mice, generated by ES cell-mediated transgenesis, were viable and fertile. Furthermore, they showed that both CAG::GFP-GPI and CAG::myr-Venus transgenic mouse lines maintain widespread transgene expression throughout embryonic development and adulthood. Both fusion reporters were found in the plasma membrane; however, GFP-GPI localized more prominently to the membrane than myr-Venus. High levels of GFP-GPI were found in the apical plasma membrane domains of the epithelium. Both fluorescent protein fusions also localized to the Golgi complex, revealing the highly dynamic nature of this organelle. Other reporter mice similar to myr-Venus or GFP-GPI, which label the plasma membrane, have also been established; there are, however, differences in their subcellular localization signal sequences (Kondoh et al. 1999; Muzumdar et al. 2007; Abe et al. 2011; Imayoshi et al. 2012). The membrane-marker reporter mice listed in Table 1 can be used to visualize cell morphology in living cells (Morris et al. 2010, 2012; Abe et al. 2011; Williams et al. 2012).
Nucleus and membrane: Co-expression of reporters
The first breakthrough in co-expression of two reporters came from using internal ribosome entry sequences (IRES; Fig. 5). The IRES was initially identified in the encephalomyocarditis virus, a member of the genus Cardiovirus, which belong to the family of small RNA viruses called Picornaviridae (Jang et al. 1988). The IRES can be used to direct ribosomes to initiate translation at internal sites within the mRNA. They can also produce multiple proteins from a single mRNA transcript because the ribosomes bind to IRES in a 5′-cap-independent manner and initiate translation. IRES are typically large (~500 nucleotides), and more importantly, the protein translation of the second open reading frame (ORF) is much lower (~10%) than the first ORF. However, this technique is rarely applied to express multiple fluorescent proteins in mouse.
In recent years, the use of the 2A peptide in multi-cistronic constructs has become a popular alternative to the IRES (De Felipe et al. 2006). The 2A peptide was identified among picorna viruses, but in a different sub-group called the Aphthoviruses, a typical example of which is the Foot-and-mouth disease virus (Robertson et al. 1985). The 2A-like sequences have since been found in other members of Picornaviridae family such as the Equine rhinitis A virus, and in unrelated viruses such as the Porcine teschovirus-1 and the insect Thosea asigna virus (TaV; Donnelly et al. 2001). In these viruses, multiple types of proteins are derived from a large polyprotein encoded by a single ORF. Recent reports demonstrated that ribosomes skip the synthesis of the glycyl-prolyl peptide bond at the C-terminus of a 2A peptide, leading to the cleavage between the 2A peptide and its immediate downstream peptide (De Felipe et al. 2011). As a result, the downstream peptide that is cleaved off has a proline residue at its N-terminus (Fig. 5, 2A peptide). Two advantages of the 2A peptide over IRES are its short length (18–25 amino acids) and its stoichiometric expression of multiple proteins flanking the 2A peptide (close to 100% in vitro; Donnelly et al. 2001). Indeed, the 2A peptide has been used to effectively mediate the cleavage of polyproteins (Szymczak et al. 2004).
As discussed above, using a fluorescent protein fused with histone H2B allows the visualization of nuclear events, while a fluorescent protein fused with a membrane localization signal sequence highlights cellular shape. Thus, a reporter gene containing both types of fluorescent fusions would make it possible to effectively monitor cell shape, movement, mitosis, and apoptosis simultaneously. Using the 2A peptide, several reporter mouse lines that express bicistronic reporters either conditionally and/or ubiquitously have been generated successfully (Trichas et al. 2008; Chen et al. 2011; Shioi et al. 2011). The CAG-TAG transgenic mouse was generated by pronuclear injection, which expressed Myr-tdTomato-2A-H2B-EGFP (TAG) under a CAG promoter (Table 1, Nucleus and membrane). A membrane-localized tdTomato gene (Myr-tdTomato) was linked to a nuclear-localized EGFP gene (H2B-EGFP) with the 2A sequence. The tdTomato is a red fluorescent protein and contains a tandem repeat of two copies of dTomato (Shaner et al. 2004). The TAG expression and 2A function have been assessed both in developmental stages and in adult organs. In this transgenic mouse, during pre-gastrulation stages, EGFP is localized exclusively to the nucleus, while tdTomato is present predominantly in the plasma membrane. H2B-EGFP and Myr-tdTomato localized to mutually exclusive cellular compartments in all of the tissues examined: the brain, heart, lung, liver, kidney, ileum, colon, and adrenal gland. tdTomato and EGFP were never found to co-localize in the transgenic mouse. The transgene is inherited in a Mendelian manner, suggesting that the widespread expression of proteins connected with the 2A peptide sequence has no deleterious effects on the health of transgenic mice.
Shioi et al. (2011) established an R26R-RG knock-in reporter line that conditionally expresses H2B-mCherry-2A-EGFP-GPI in the ROSA26 locus, and also generated the R26-RG line that expresses a reporter ubiquitously by removing loxP-flanked stop sequences in the germline (Table 1, Nucleus & membrane). The resulting R26-RG embryos exhibited red and green fluorescence in the nuclei and cell membrane, respectively. Time-lapse images of developing embryos of these mouse lines have been successfully recorded at both pre- and post-implantation stages. Furthermore, this homozygous mouse is both viable and fertile. The fertility of homozygous R26-RG mice indicates that the expression levels of this set of reporter genes are not high enough to be toxic at this locus. Indeed, the intensity of fluorescent signals was lower than in the UBC-HS-GR transgenic mice, which were infertile in a hemizygous state (Stewart et al. 2009). Another group has also recently established a R26R-GR (GR; H2B-EGFP-2A-mCherry-GPI) knock-in reporter mouse line (Chen et al. 2012).
Fluorescent fusion proteins that localize to specific structures within cells allow us to observe the dynamic subcellular events that take place in living animals. Commonly used mouse lines that express these types of fluorescent proteins are summarized in Table 1.
The labeling of microtubules reveals features such as the mitotic machinery, by illuminating the microtubule component of the cytoskeleton. It is a powerful tool for studying microtubule dynamics in a variety of cell types and cellular processes, including axon path finding, cytokinesis, and cell migration. Hadjantonakis et al. (2008) generated a transgenic mouse line, named CAG::Tau-GFP (Table 1, Microtubules), which expresses a fusion protein of GFP and bovine microtubule-binding protein tau (Callahan & Thomas 1994; Hadjantonakis et al. 2008). The Tau-GFP transgene produces a fusion protein that marks subcellular structures, such as axons and mitotic machinery. This protein has been used to observe the relatively stationary mechanosensory cilia around the periphery of the node, as well as the rapid regular rotational movement of nodal cilia. Pratt et al. (2000) reported generating a transgenic mouse using the same DNA construct and characterized the Tau-GFP expression at E3.5 and E10.5. They investigated the thalamocortical axon pathway (TCA) in mouse to evaluate the effectiveness of Tau-GFP as a marker of fiber tracts. The TCA fiber tract was labeled with Tau-GFP and exhibited a particularly strong staining in the internal capsule and intermediate zone at E16 (Pratt et al. 2000).
Visualization of filamentous actin (F-actin) in living cells is critical for the study of cellular morphogenetic processes such as cell division, cell migration, or polarization. One marker that is widely used is the Actin-GFP fusion protein (Westphal et al. 1997), although, a strong expression of Actin-GFP can be toxic for the cells. Therefore, as an alternative to Actin-GFP, researchers have been using fusions of GFP to actin-binding domains, notably from Moesin in Drosophila melanogaster (Edwards et al. 1997). Recently, Riedl et al. (2010) found that the first 17 amino acid residues of Abp140 were sufficient to mediate actin localization in a similar manner to the full-length protein, named “Lifeact”. The Abp140-GFP is the only probe that has been used to successfully label the actin cables in budding yeast (Asakura et al. 1998; Yang & Pon 2002). Lifeact is conserved among close relatives of Saccharomyces cerevisiae, but is absent from other organisms. Two transgenic mouse lines, Lifeact-EGFP and Lifeact-mRFPruby have been generated using Lifeact (Table 1, Actin cytoskeleton), with each line expressing the probe under a CAG promoter (Fischer et al. 2006; Riedl et al. 2010). Both mouse lines were viable, fertile, and phenotypically normal, confirming previous findings in cultured cells that Lifeact expression does not interfere with cellular processes.
Finally, an mtGFP-tg mouse expressing a fusion GFP of cytochrome c oxidase subunit VIII A was established to label the mitochondria, particularly for visualizing sperm mitochondria (Rizzuto et al. 1995; Shitara et al. 2001; Table 1, Mitochondria).
Most of the reporter mouse lines described above were generated by random transgenesis and designed to express reporters constitutively. Despite this progress, the number of reporter mouse lines available for research is still inadequate. As a measure to address this issue, a series of conditional knock-in reporter mouse lines were established, in which specific organelles were marked with fluorescent fusion proteins (Table 1; Abe et al. 2011). The cDNAs were inserted into the ROSA26 locus next to the loxP-flanked stop sequence, so that reporters were expressed only after the stop sequences were removed by Cre recombinase-mediated excision (Soriano 1999; Srinivas et al. 2001). Reporter mouse lines that ubiquitously expressed reporters were also generated from each conditional reporter line by Cre recombination in the germline. Subcellular localization and intensity of the fusion protein in each reporter mouse line was investigated in E7.5 embryos. Twelve mouse lines, which marked the nucleus, plasma membrane, mitochondria, Golgi apparatus, microtubules, actin filaments, or focal adhesion sites, were found suitable for imaging.
When cells are labeled by Cre-mediated recombination, it is useful to the researcher if they can make a distinction between recombined and non-recombined cells. To facilitate this identification, double reporter transgenes that express one marker in recombined cells and another marker in non-recombined cells were generated, such as CAG-CAT-Z described in (General design of reporter mouse lines expressing exogenous genes). This double marker system is useful for mosaic analysis, especially in live tissues. Muzumdar et al. (2007) reported that a double fluorescent Cre recombinase knock-in reporter mouse, mT/mG. mT/mG expresses membrane-targeted tdTomato (“mT”) before Cre excision, and expresses membrane-targeted EGFP (“mG”) after Cre excision, thereby enabling the live visualization of and discrimination between recombined and non-recombined cells (Fig. 6; Muzumdar et al. 2007; Williams et al. 2012). The combination of a strong and ubiquitous CAG promoter (Niwa et al. 1991) and the ROSA26 targeting locus (Soriano 1999) makes it possible to label the tissues with bright fluorescence. In addition, the localization of fluorescent proteins to membrane structures (“m”) can outline the cell morphology and improves the resolution of fine cellular processes.
Multiple fluorescent proteins can be combined to visualize the locations of different cell types, a method that relies on the differential activities of Cre recombinase. There was a report on labeling neurons with multiple colors using a mixture of several fluorescent proteins obtained by random Cre recombination, which produced more than 100 colorful hues that could be distinguished by fluorescence microscopy (Fig. 7a–c; Livet et al. 2007). This technique, called Brainbow, uses Cre-mediated recombination to generate a random mix of several fluorescent proteins expressed in individual neurons and is expected to facilitate deciphering complex neuronal circuits. But, because this mouse expresses fluorescent proteins under the neuron-specific Thy1 promoter, it is not useful for analyzing other organs. Two lines, Rainbow2 and R26R-Confetti, were recently generated using Brainbow technology to perform multicolor lineage-tracing in a specific tissue. Rainbow2 mouse essentially uses the same strategy as the Brainbow-1.0 mouse (Fig. 7a), expressing the Brainbow-1.0 construct under the CAG promoter (Tabansky et al. 2013). This mouse allows researchers to carry out multicolor lineage tracing in many organs at various stage of development. This mouse can produce beautiful color images by inducing recombination, using tamoxifen inducible Cre mice. The R26R-Confetti mouse harbors a Brainbow-2.1 construct encoding nGFP, YFP, RFP, and mCFP flanked by loxPs in the ROSA26 locus (Fig. 7d; Snippert et al. 2010). Before Cre activation, the cells are prevented from expressing fluorescent proteins by loxP-flanked stop sequences, whereas after Cre recombination, the cells are induced to stochastically express nGFP (nucleus-GFP), YFP, RFP, or mCFP (membrane-CFP). This mouse has been demonstrated to be useful for tracing intestinal stem cell lineage (Snippert et al. 2010; Schepers et al. 2012). Cre recombinase can also induce flipping of flox sequences, resulting in the expression of YFP or mCFP, which is originally placed in a reversed orientation (Fig. 7d).
Cell cycle indicator
Visualizing the progress of cell cycle in cells of live embryos provides valuable information to understand developmental processes such as pattern formation, morphogenesis, cell differentiation, and growth. Sakaue-Sawano et al. (2011) developed a fluorescent ubiquitination-based probe, called Fluorescent ubiquitination-based cell cycle indicator (Fucci), to visualize the cell cycle in live cells (Fig. 8). This Fucci probe was generated by fusing monomeric Kusabira Orange 2 (mKO2; Karasawa et al. 2004) and monomeric Azami Green (mAG; Karasawa et al. 2003) to the ubiquitination domains of human Cdt1 (hCdt1(30/120)) and Geminin (hGem(1/110)), respectively. These two chimeric proteins, mKO2-hCdt1(30/120) and mAG-hGem(1/110), accumulate at different stages of the cell cycle to label the G1 phase nuclei in orange, and the S, G2 and M phase nuclei in green; thus, they function as G1 and S/G2/M probes, respectively. To visualize the cell cycle in live embryos or animals, two transgenic mouse lines, CAG-mKO2-hCdt1(30/120) and CAG-mAG-hGem(1/110), were generated by pronuclear injection. The CAG-Fucci mouse line that co-harbors CAG-mKO2-hCdt1(30/120) and CAG-mAG-hGem(1/110) was obtained by crossing the two transgenic lines generated by random transgenesis, and was successfully used to monitor the cell cycle in the head region of live E13 embryos, which revealed regional differences in cell proliferation activities in the neural tissue. However, the CAG-Fucci expression was weak or failed to be detected in some tissues, such as extraembryonic tissues in early post-implantation embryos. Recently, a new Fucci probe Fucci2 was created, which uses mCherry and mVenus to improve the color contrast (Sakaue-Sawano et al. 2011). This led to the establishment of a second generation of Fucci2 reporter mouse lines, where mCherry-hCdt1(30/120) and mVenus-hGem(1/110) are conditionally expressed in the respective lines (Abe et al. 2013). R26R mice were produced by the targeted insertion of reporters in the ROSA26 locus, and the reporter genes are placed after the loxP-flanked stop sequences. These mice can then be used to analyze the cell cycle in a specific tissue by crossing them with a tissue-specific Cre mouse. In addition, a new Fucci2 transgenic mouse, R26p-Fucci2, expressing both mCherry-hCdt1(30/120) and mVenus-hGem(1/110) under ROSA26 promoters was also generated by pronuclear injection (Abe et al. 2013). This mouse was convenient for analyzing the cell cycle in a mutant background because the two reporters are expressed simultaneously from a single locus. The ubiquitous expression of Fucci2 probes was successfully confirmed in R26p-Fucci2 embryos and adult tissues, and time-lapse imaging and quantitative analysis were carried out in pre-implantation stages.
Conclusions and perspectives
We described several conventional transgenic mouse lines expressing fluorescent probes in this review. Many transgenic mouse lines that express fluorescent probes under specific promoters have already been established (Fraser et al. 2005; Kwon et al. 2006; Kwon & Hadjantonakis 2007, 2009; Viotti et al. 2011), and the use of new technologies, such as FRET-based sensors, are becoming increasingly common in live imaging experiments (Chudakov et al. 2010). In addition, the development of novel color variants of fluorescent proteins for multi-color imaging and of photoactivatable Cre for spatio-temporal labeling will help advance and refine fluorescent labeling and live imaging technology. We expect that the use of these technologies will become more widespread in mouse research (Yamaguchi et al. 2011; Kamioka et al. 2012). Collections of transgenic mouse lines for live imaging are universal tools that can be used by many researchers, and there are a few databases of mouse lines that list available reporter mouse lines (http://www.informatics.jax.org/ and http://jaxmice.jax.org/research/fluorescent_proteins_or_lacZ.html). However, it is still difficult for researchers to search for reporter mouse lines that meet their specific purposes and thus a new database of reporter mouse lines expressing fluorescent markers that would facilitate this search should be established in the near future.
We thank members of the Laboratory for Animal Resources and Genetic Engineering, RIKEN Center for Developmental Biology (CDB) for their helpful advice.