Mechanisms of Development 76 (1998) 79–90

Generating green fluorescent mice by germline transmission of green fluorescent ES cells Anna-Katerina Hadjantonakis a, Marina Gertsenstein a, Masahito Ikawa b, Masaru Okabe b, Andras Nagy a , c ,* a

Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University Avenue, Toronto, ON M5G 1X5, Canada b Research Institute for Microbial Diseases, Osaka University, Yamadoaka 3-1, Suita, Osaka 565, Japan c Department of Molecular and Medical Genetics, University of Toronto, Toronto, ON M5S 1A8, Canada Received 24 April 1998; revised version received 5 June 1998; accepted 5 June 1998

Abstract Green fluorescent protein (GFP) and its variants currently represent the only non-invasive markers available for labeling mammalian cells in culture or in a multicellular organism through transgenesis. To date this marker gene has been widely used in the study of many organisms, but as yet has not found large-scale application in mammals due to problems encountered with weak fluorescence and instability of the wild-type protein at higher temperatures. Recently, though, several mutants have been made in the wild-type (wt) GFP so as to improve its thermostability and fluorescence. EGFP (enhanced GFP) is one such wtGFP variant. As a first step in assessing the use of EGFP in ES cell-mediated strategies, we have established a mouse embryonic stem (ES) cell lines expressing EGFP, which can be propagated in culture, reintroduced into mice, or induced to differentiate in vitro, while still maintaining ubiquitous EGFP expression. From the results presented we can suggest that: (1) possible improvements in the efficiency of transgenic regimes requiring the germline transmission of ES cells by aggregation chimeras can be made by the preselection chimeric embryos at the blastocyst stage; (2) the expression of a noninvasive marker, driven by a promoter that is active during early postimplantation development, allows access to embryos during a window of embryonic development that has previously been difficult to investigate; (3) the behavior of mutant ES cells can be followed with simple microscopic observation of chimeric embryos or adult animals comprising green fluorescent cells/tissues; and (4) intercrosses of F1 mice and subsequent generations of animals show that progeny can be genotyped by UV light, such that mice homozygous for the transgene can be distinguished from hemizygotes due to their increased fluorescence.  1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: Green fluorescent protein; Wild type; Embryonic stem; Mouse

1. Introduction Cell marking is an important feature of both transgenic approaches and cell lineage analyses, with the ultimate aim being an ability to utilize a non-invasive marker, thereby circumventing the sacrifice of the experimental sample for the analysis. To date the majority of lineage markers used in vertebrates suffer from a variety of limitations; for example, ectopically applied markers such as colloidal gold, LRD (lysinated rhodamine dextran) and DiI (1,1′-dioctadecyl3,3,3′,3′-tatramethylindocarbocyanine) are diluted during development, and the protein products of marker genes used in transgenics such as those encoding the bacterial b* Corresponding author. E-mail: [email protected]

0925-4773/98/$19.00  1998 Elsevier Science Ireland Ltd. All rights reserved PII S0925-4773 (98 )0 0093-8

galactosidase gene (lacZ), horseradish peroxidase (HRP), and human placental alkaline phosphatase (hAP) require substrate reactions to be visualized (Hogan et al., 1994; Li et al., 1997a). Green fluorescent protein (GFP) is a novel genetic reporter system derived from the bioluminescent jellyfish Aequorea victoria which, when expressed in either eukaryotic or prokaryotic cells and illuminated by blue or UV light, yields a bright green fluorescence (Chalfie et al., 1994; Prasher, 1995). Therefore GFP currently represents a superior alternative, in a broad range of applications, to other cell lineage markers in that it is non-invasive and can therefore be viewed in real time in vitro or in situ in the developing embryo or adult. The GFP chromophore consists of a cyclic tripeptide which is only fluorescent when embedded within the com-

80

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

plete GFP protein (Brejc et al., 1997; Li et al., 1997b). Until recently this marker gene had been widely used in many different organisms. Interestingly, though, problems encountered with the weak fluorescence and instability of the protein at higher temperatures have generally precluded its use in vertebrate systems. Recently though, mutagenesis of the wild-type (wt) gene has yielded several mutant forms of the original wtGFP which exhibit greater thermostability and increased fluorescence (Cormack et al., 1996; Heim and Tsien, 1996), consequently overcoming the aforementioned limitations (Ikawa et al., 1995; Zernicka-Goetz et al., 1996, 1997; Okabe et al., 1997). In this study we utilized the EGFP (Clontech Laboratories, CA, USA) variant of wtGFP which contains two amino acid substitutions in the vicinity of the chromophore (Phe-64 to Leu, Ser-65 to Thr) and 190 silent base substitutions resulting in modification of the sequence for humanized codon usage. In order to investigate the fidelity of EGFP in mammals, we sought to generate EGFP expressing (green fluorescent) ES cells which after their germline transmission would give ubiquitous EGFP expressing transgenic lines of mice. Such mice would provide an easily visualized non-invasive marker which could be crossed onto established mutant or transgenic mouse lines for chimera studies, and would also provide a source of marked cells or tissues for transplant/ explant studies. Recently, several reports have described the use of wtGFP variants in vertebrates using injection of DNA into early embryos (Ikawa et al., 1995; Zernicka-Goetz et al., 1996; Okabe et al., 1997; Zhuo et al., 1997) or ES cellmediated chimeras (Zernicka-Goetz et al., 1997), though at present there is no report of germline transmission of an ES cell-mediated GFP transgene. We believe that an ES cell-based approach offers several advantages over pronuclear injection-based strategies. In particular, ES cells can be studied in vitro and are available for selectable marker aided genomic modifications prior to their re-introduction into mice. The potential for further genome modifications of green fluorescent ES cells opens up new possibilities in gene targeting (Nagy, 1996). Strategies could be envisaged where a targeting construct incorporating a gene-specific marker, such as lacZ or one of the other wtGFP variants such as enhanced blue fluorescent protein (EBFP) or enhanced yellow fluorescent protein (EYFP), is used to target green fluorescent ES cells. Thereby EGFP would provide an independent, cell-autonomous marker for mutant cells which would then also contain a second gene-specific marker. Such cells could be used for mutant cell transplantation or chimeric analyses.

2. Results 2.1. Establishing lines of green fluorescent ES cells Ubiquitously EGFP (Promega Laboratories, CA, USA)

expressing (green fluorescent) ES cell lines were established by co-electroporation of pCX-EGFP (Okabe et al., 1997), an EGFP expressing vector, where the native EGFP gene is driven by a CMV (cytomegalovirus) immediate early enhancer coupled to the chicken b-actin promoter and first intron (Niwa et al., 1991), and pPGKPuro (Tucker et al., 1996), a selectable marker containing vector. Of the surviving puromycin-resistant colonies only green clones were selected for further investigation. These clones were visualized as green fluorescing colonies under a dissecting microscope equipped with fluorescent optics, and picked into 96well gelatinized tissue culture plates. Clones were grown in duplicate with one plate being frozen down as a stock. A variation in the intensity and homogeneity of EGFP expression was observed between clones, possibly reflecting the copy number and the effect of site of integration of the transgene. ES cells grown on gelatinized tissue culture dishes maintained a high level of green fluorescence (Fig. 1A,B). In most cases maintenance of the green fluorescence was observed in differentiated ES cells (Fig. 1C), whereas embryoid bodies plated on gelatinized plates (Fig. 1D) exhibited ubiquitous fluorescence though at varying levels, supporting previous observations that the promoter used in this study is expressed in all nucleated cells but at different levels in certain lineages (Hadjantonakis et al., 1998a). We consistently observed increased fluorescence in endodermal cells situated on the surface of embryoid bodies (Fig. 1D, arrows), and suggest that these cells may be equivalent to parietal endodermal cells which were observed to give increased fluorescence over other cell types in transgenic mouse embryos established from this ES cell line. 2.2. Selection criteria for ES cell lines: an in vivo test In order to produce germline transmitting chimeras, several different lines of green fluorescent ES cells were tested for developmental potential and ubiquitous embryonic expression by tetraploid embryo–ES cell aggregation (Nagy and Rossant, 1993; Nagy et al., 1993) (see Section 2.9). Four of these lines (B4/EGFP, B5/EGFP, C1/EGFP and D4/EGFP) were subsequently chosen for diploid embryo–ES cell aggregations (Wood et al., 1993) in order to obtain germline transmitting chimeras. 2.3. Pre-selecting chimeric embryos for transfer into recipient females Aggregating ES cells with diploid eight-cell stage embryos (Fig. 2A–D) is a simple and inexpensive means of introducing genetic alterations into mice (Wood et al., 1993; Tanaka et al., 1998). Circumventing the transfer of non-chimeric embryos after the aggregation is completed makes the technology more efficient and reduces animal space requirements. A green fluorescent ES cell line now provides a convenient tool to do so, since the incorporation of ES cells into preimplantation stage embryos can be mon-

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

Fig. 1. Expression of EGFP in undifferentiated and differentiated ES cells. (A) ES cells and (B) drug-resistant colonies, exhibiting ubiquitous EGFP expression. (C) differentiated ES cells and (D) embryoid bodies also exhibit ubiquitous transgene activity, though the latter do show differential levels of expression with the surface endoderm layer often containing a population of highly fluorescent cells (indicated by an arrow).

itored prior to their transfer (Fig. 2B–E). ES cell contributions to the embryo can vary, with ES cell contributions into morula (Fig. 2F) and blastocyst (Fig. 2G–K) stage embryos after overnight incubation of eight-cell stage embryo–B5/ EGFP ES cell aggregates depicted. While ES cells will preferentially contribute to the ICM, where they can either represent the entire ICM (Fig. 2G) or just a portion of it (Fig. 2H,I), they will also occasionally be present in the trophoblast (Fig. 2J (arrow),K), though it is unclear whether this is a real contribution to the trophoblast lineage or these cells are trapped in this region of the embryo never to proliferate. Studying postimplantation stage chimeric embryos of this type will possibly answer this question. 2.4. Germline transmission of the B5/EGFP and C1/EGFP transgenes: diploid embryo–ES cell aggregation for the production of chimeras Aggregation experiments from three of the four lines selected for germline transmission resulted in transmitting chimeras, these being B5/EGFP, C1/EGFP and D4/EGFP. Here we report the results of experiments performed with two of the three cell lines selected (B5/EGFP and C1/ EGFP). The third line (D4/EGFP) gave rise to male chimeras which exclusively transmitted their transgene to all female progeny, and is described elsewhere (Hadjantonakis et al., 1998a). This result along with further breeding indicated that the D4/EGFP line harbored a transgene insertion on its X-chromosome, and has therefore been renamed D4/ XEGFP. The D4/XEGFP line exhibited ubiquitous green fluorescence, and also allowed non-invasive sexing of embryos prior to overt sexual dimorphism (Hadjantonakis et al., 1998a), and real-time visualization of X-inactivation (unpublished observations).

81

The B5/EGFP and C1/EGFP lines were initially chosen as they exhibited extremes in expression of the transgene, as assessed by tetraploid aggregation. B5/EGFP exhibiting a high level of ubiquitous green fluorescence in midgestation tetraploid aggregation-derived embryos, with C1/EGFP also exhibiting ubiquitous green fluorescence though at a much reduced level. These extremes were chosen since no data was available on the tolerance of mice expressing EGFP at high levels. Chimeras were produced by diploid aggregation. In order to observe the fidelity of EGFP as a cell autonomous marker within a mosaic population of cells, some of the transferred embryos were recovered at later (postimplantation) stages of embryonic development. Individual EGFP expressing cells can be identified in chimeric embryos at stages E7.75 (Fig. 3A) and E8.5 (Fig. 3B). We also noted that in addition to black eyes and pigmented coat (representative of the 129 mouse strain the R1 ES cells are derived from), chimeras could be identified and assessed for the strength of their ES cell contribution by viewing new-born pups from aggregation recipients under fluorescence, either by using an appropriately equipped stereo dissecting microscope, or macroscopically with a pair of eyeglasses fitted with barrier filters having samples illuminated using a strong light source (Fig. 3C,D). Male chimeras with a high ES cell contribution were crossed with CD1 or ICR females, and F1 mice were genotyped using the appropriate optics. Both the B5/EGFP and C1/EGFP chimeric mice transmitted the EGFP transgene in a Mendelian fashion. B5/EGFP mice were observed as ubiquitously green throughout embryonic development and adult life (Figs. 4–6), whereas in the C1/EGFP line, EGFP was ubiquitously expressed (though at a lower level than in the B5/ EGFP line) during early to mid-gestational embryonic development then became restricted to the pancreas and certain regions of the CNS in newborns and adults (data not shown). Therefore the data presented here will focus on the B5/EGFP line.1 2.5. Visualizing individual green fluorescent cells in chimeras It had previously been reported that individual GFP expressing cells could be identified within mixed population tissue sections (Zernicka-Goetz et al., 1996; Zhuo et al., 1997). In order to investigate whether this was the case with B5/EGFP expressing cells in the tissues of the adult chimeric mice, we attempted several procedures for producing sections. Wax sections could not be cut, as the organic solvents used to prepare the sample would obliterate the GFP activity, we also found that cryosections gave unsatisfactory results, not yielding data to a cellular resolution. The most promising results were obtained with sections made 1 The B5/EGFP mice described here will be available from the Jackson Laboratory Induced Mutant Resource (Bar Harbor, ME, USA), and are referred to therein as TgN(GFPU)5Nagy.

82

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

Fig. 2. Diploid green ES cell–wild-type embryo aggregations allowing preselection of embryos with a good ICM contribution for transfer into recipient females. (A) Schematic diagram detailing the technique. (B–E) Aggregation of a clump of green fluorescent ES cells with a wild-type embryo results in blastocyst formation after overnight incubation, with preferential contribution of the ES cells to the inner cell mass (E). (F) ES cells colonizing part of (right) and the majority of (left) morula stage embryos. (G) ES cells colonizing most of the ICM but not the trophectoderm of a blastocyst stage embryo. (H, I) ES cells colonizing a proportion of the inner cell mass (ICM) of a blastocyst. (J) ES cells localized to the majority of the ICM with a single cell (arrowed) in the trophectoderm (tro) of a blastocyst. (K) ES cells localized to part of the ICM with two cells seen in the trophectoderm.

using a vibratome. Relatively thick (50 mm) sections could be cut from minimally fixed tissues, which could subsequently be viewed using a regular epifluorescence microscope (Fig. 3E–O); alternatively a confocal or deconvolution microscope could be used for greater resolution (Zhuo et al., 1997, and unpublished observations). If samples could not be fixed (in order to be viewed whilst the tissue was still alive), slightly thicker (100–150 mm) sections needed to be cut. Individual EGFP expressing cells could be visualized in a variety of chimeric tissue sections.

EGFP expressing cell contributions to different regions of the brain are illustrated in sections of the cerebral cortex (Fig. 3E,F) of a strong and moderate (Fig. G,H) chimera. Also, pyramidal cells can be distinguished in a weak chimera (Fig. 3H). Cortical collecting duct cells leading to the kidney (Fig. 3M) can be delineated in sections made from a strong chimeric animal, with mesenchymal and epithelial cells (Fig. 3N) and mesangial cells of a glomerulus (Fig. 3O) being distinguished in the kidney of a weak chimera.

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

83

Fig. 3. Chimeras made by diploid embryo–ES cell aggregation. Embryos exhibit mosaic expression of EGFP reflecting their chimerism between wild-type (non-fluorescing) and B5/EGFP ES cell contributions. (A) Ventral view of a late headfold/presomite stage (E8) embryo, anterior is top; (B) lateral view of an early somite stage (E8.5) embryo, anterior is left; (C) dorsal view of an E8.5 embryo, anterior is left; (D) lateral view of the head and forelimb of a new-born chimera; (E) view of the hindlimbs and tail of a new-born chimera, anterior is up. E–O: sections of tissues from adult (∼2–3 months) chimeras. Cells of the cerebral cortex in a strong (E, bright field; F, dark field of same panel) and a moderate chimera (G, H), and pyramidal cells (I, dark field) of a weak chimera. Cortical cells of the collecting ducts leading to the kidney (J, bright field; M, dark field) in a strong chimera, mesenchymal and epithelial cells in the body of the kidney (K, bright field; N, dark field), and higher magnification view of a glomerulus (L, bright field; O, dark field) from a weak chimera with mesangial cells highlighted.

84

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

protein is very stable or that the maternal transcript is very abundant. Due to the high level of maternal transgene expression, all preimplantation stage embryos derived from such a cross were observed as being green fluorescent, though only half the embryos inherited the transgene (Fig. 4E–G). We also noticed that green fluorescence resulting from maternal expression is more pronounced than zygotic expression up until the blastocyst stage. A comparison of embryos derived from a hemizygous male crossed to a wildtype female (Fig. 4H (left), note that about half the embryos are green), and a wild-type male crossed to a hemizygous female (Fig. 4H (right), note that all the embryos are green though only half will carry the EGFP transgene) illustrates this observation. The residual maternal transgene expression observed in preimplantation embryos is lost after implantation, probably as a result of the degradation of the EGFP transcripts or protein, and the rapid increase of the cellular mass that is observed concurrent with implantation at E4.5–4.75, and the onset of gastrulation at E6.5. 2.7. Expression of the B5/EGFP transgene in postimplantation embryos

Fig. 4. B5/EGFP expression in preimplantation embryos. All embryos in panels A–D are derived from a cross between a male that is hemizygous for the transgene and wild-type females, whereas embryos in panels E–H are derived from a cross between a wild-type male and a female hemizygous for the transgene. (A) one- to four-cell stage embryos exhibiting no transgene expression; (B) embryos initiate expression at the morula (eightcell) stage (∼E2.5). By the blastocyst stage, expression of the transgene is clearly visible. (C) Dark-field and (D) bright-field view of E3.5 embryos. Residual transgene expression due to maternal transcripts is observed until implantation stages, two-cell and morula stage green-fluorescent embryos bright field (E), dark field (F), dark field of blastocyst stage embryos (G). At E2.5, fluorescence due to maternal transcripts (right) is stronger than zygotic (left).

2.6. Expression of the B5/EGFP transgene in preimplantation embryos The onset of zygotic transgene expression was assayed for from the two-cell stage on. No green fluorescence was observed up to the four-cell stage, in embryos derived from crosses between transgenic males obtained after germline transmission and wild-type females (Fig. 4A). Zygotic expression from the B5/EGFP transgene first becomes visible with green fluorescence first detected at the compacting morula (eight-cell) stage (Fig. 4B). By the blastocyst stage EGFP expression is strong and clearly visible (Fig. 4C,D). On the other hand, from the analyses of non-transgenic embryos derived from matings between wild-type males and hemizygous females, we found that residual maternal expression of the transgene is observed until after implantation (Fig. 4E–G), suggesting that the EGFP transcript or

Postimplantation expression of the B5/EGFP transgene in embryos and extraembryonic tissues appeared to be ubiquitous throughout development (Fig. 5). We also found that the green fluorescence facilitated the location and subsequent isolation of embryos at early postimplantation stages (E4.75–E5.75) which is otherwise extremely difficult. This means that such a green fluorescent transgenic line allows access to embryos at all stages of development; from preimplantation through postimplantation to birth. Fig. 5A shows an E4.75 (late hatched blastocyst) green fluorescent embryo derived from a cross between a B5/EGFP male and a wild-type female; the embryo has just implanted and is surrounded by some residual non-fluorescent uterine tissue. Fig. 5B shows a uterus that has been cut open at the proximal side (opposite the mesometrial side) showing early stage decidual tissue which has been torn open to expose an E5 green fluorescent embryo. Fig. 5C shows an E5.5 embryo dissected away from the uterus but still surrounded by some residual decidual tissue. By E9 (Fig. 5D,E) it is clearly visible that both the embryo proper and extraembryonic tissues express the EGFP transgene. Starting at E9 in the extraembryonic lineage (Fig. 5E), the parietal endoderm exhibits highest levels of transgene expression, this observation being supported by the differentiated embryoid body data where endoderm cells regularly exhibit increased fluorescence (Fig. 1D arrows). At all postimplantation stages no fluorescence is observed in non-transgenic embryos, as is illustrated in Fig. 5F,G. Fig. 5F shows a bright-field normal optic view of two E9.5 littermates derived from a cross between a transgenic male and a wild-type female (only one has inherited the transgene), with Fig. 5G being a dark-field fluorescent optic view where only the transgenic embryo can be seen to fluoresce.

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

85

Fig. 5. B5/EGFP expression in postimplantation embryos. All embryos derived from a B5 hemizygous male crossed with ICR female. (A) E4.75 newly hatched and implanted blastocyst attached to decidual tissue (dec); (B) E5; (C) E5.5; (D) E9 embryo with associated ectoplacental cone (EPC) partially contained within the deciduum; (E) E9 extraembryonic membranes; (F) bright-field view of two E9.5 littermates; (G) dark-field view of the same two embryos; (H) E11.5 embryo with placenta; (I) E11.5 heart and lungs; (J) E11.5 kidney rudiment and intestine.

The residual maternal transgene expression observed in preimplantation embryos is thus lost after implantation. After midgestation transgene expression is maintained in both the embryo and placenta (Fig. 5G). All tissues appear to express the transgene including the heart and lungs (Fig. 5I) in addition to the kidney rudiment and intestine (Fig. 5J). The only cells that appear not to fluoresce green are enucleated cells such as erythrocytes. Additionally cells and tissues with increased erythrocyte content, such as blood vessels, spleen, kidney and liver, exhibit reduced fluorescence as their development proceeds since the hemoglobin masks the green fluorescence.

2.8. Expression of the B5/EGFP transgene in adult mice The analysis of EGFP expressing cells in tissue sections of the chimeras suggested that the transgene was widely expressed in the adult. In order to verify whether the transgene is truly ubiquitous, we looked at tissues from adult animals which were either hemizygous or homozygous for the B5/EGFP transgene. As with transgenic embryos, the B5/EGFP transgene was found to be ubiquitously expressed in new-born (Fig. 6A–C) and adult mice (Fig. 6D–L). Transgenic new-born pups can be recognized due to fluorescence in the skin (Fig. 6A–C), but as animals mature the

86

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

fluorescence is obscured in the fur coat-covered body parts, especially if the fur is pigmented (Fig. 6F). This contrast is clearly observed between adult mice with different coat colors, and is illustrated here by albino (Fig. 6D–F, tail on the left), chinchilla and agouti (Fig. 6F, tails center and right). The entire organ system of the B5/EGFP transgenic line fluoresces green and, even though ubiquitous, the level of expression varies between different organs, with musculature (Fig. 6G), brain (Fig. 6H) and pancreas (Fig. 6K) exhibiting the highest levels of transgene expression, while intestine (Fig. 6L) expresses at a lower level. Kidney (Fig. 6I) and liver (Fig. 6J) also do not strongly fluoresce as they have a high hemoglobin content.

2.9. Mouse embryos derived solely from ES cells: wt tetraploid embryo–GFP ES cell and GFP tetraploid embryo–wt ES cell aggregations The aggregation of ES cells with tetraploid four-cell stage embryos results in the production of a completely ES cellderived postimplantation embryo, and provides a way to study the phenotype of a given ES cell line, circumventing germline transmission (Nagy and Rossant, 1993; Nagy et al., 1993; Carmeliet et al., 1996; Tanaka et al., 1998), or to separate embryonic and extraembryonic components of a compound phenotype resulting from gene ablation (Duncan et al., 1997). The technique is illustrated in Fig. 7A, with

Fig. 6. Ubiquitous green fluorescence in B5/EGFP mice and organs. (A) head and forelimb view of a new-born B5/EGFP hemizygote pup among nontransgenic littermates, head transgenic on the right and non-transgenic littermate on the left; (B) new-born B5/EGFP hemizygote transgenic (right) and nontransgenic littermate (left); (C) 1-week-old pups, transgenic on the right, non-transgenic littermate on the left; (D) 3-week-old albino B5/EGFP mouse, anterior view; (E) 3-week-old albino B/EGFP mouse, posterior view; (F) posterior view highlighting the tails of 2-month-old albino (left), chinchilla (middle), agouti (right) mice illustrating pigment-dependent fluorescence; (G) dorsal view of an incision made in a 3-week-old B5/EGFP mouse body with skin removed to reveal fluorescence in the body. Organs from a 3-week-old transgenic mouse, right/top and non-transgenic littermate, left/bottom: (H) brain, (I) kidney, (J) liver, (K) pancreas, (L) intestine.

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

87

Fig. 7. Tetraploid aggregations to produce completely ES cell-derived embryos. (A) schematic diagram detailing the tetraploid technique. (B–E) The progression of an aggregation between green ES cells–wild-type embryo to the blastocyst stage. (E, F) Embryos made by B5 ES cell–wild-type tetraploid embryo aggregation. (G, H) Embryos made by wild-type ES cell–B5 embryo aggregation. (E) Green fluorescent E9 embryo with non-fluorescent extraembryonic membranes (outline by a hatched line). (F) Green fluorescent E11 embryo with mosaic yolk sac and non-fluorescent placenta. (G) E8.5 embryo with green fluorescent extraembryonic membranes. (H) E11 embryo with mosaic yolk sac and fluorescent placenta.

Fig. 7B–E showing the process of aggregation. The lower left two panels of Fig. 7 show embryos derived from a wildtype tetraploid embryo–B5/EGFP ES cell aggregation experiment where the embryo will be green-fluorescent (Fig. 7E,F), whereas the lower right two panels illustrate the reverse a B5/EGFP–wild-type ES cell aggregation where the ectoplacental cone/placenta will be green-fluorescent (Fig. 7G,H). With simple microscopic observation the expected separation of the tetraploid and ES cell compartments to the extra embryonic tissues and embryo proper, respectively, could be confirmed. 2.10. .Using green fluorescence for genotyping animals Unlike standard transgenic regimes where genotyping is performed by either polymerase chain reaction, Southern analysis or the use of a chromogenic substrate for reporter gene visualization, we were able to identify transgenic embryos and pups by the direct visualization of green fluorescence. When a litter of pups derived from a cross between an animal hemizygous for the transgene and a non-transgenic is viewed under the appropriate light source, only half the animals will fluoresce (Fig. 8A, asterisk denotes the position of non-transgenic littermates). Such a fluores-

cence-based reporter strategy offers the advantage not only in that transgenic animals are readily and non-invasively identified, but also that animals homozygous for the transgene can be recognized due to their increased fluorescence over hemizygotes (Fig. 8B–D), a feature that can only be determined in standard transgenics by breeding animals or by cloning the site of transgene insertion.

3. Discussion We have established mouse embryonic stem (ES) cell lines expressing EGFP, which can be propagated in culture, induced to differentiate in vivo and in vitro while still maintaining EGFP expression. Germline transmission of one of these lines, the B5/EGFP ES cells, has allowed us to establish that EGFP is ubiquitously expressed in all cellular progeny, such that it is ubiquitously expressed throughout development and adulthood. Since the EGFP transgene was introduced by co-transfection with a circular puromycin expression vector which did not integrate, the cell line can be manipulated in vitro by further transgenesis or targeted gene alteration by selectable marker containing vectors. Judged by the level of chimerism

88

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

Fig. 8. Non-invasive genotyping of embryos and adults by UV–Mendelian inheritance of the B5/EGFP transgene. Genotyping hemizygotes vs. nontransgenic animals: (A) new-born pups from a cross between a B5/EGFP hemizygous male and a wild-type ICR female; between every fluorescent (transgenic) pup is a non-fluorescent non-transgenic animal (indicated by an asterisk). Genotyping hemizygotes vs. homozygotes: (B) mice from a litter of pups derived from an intercross between a male and female, both of which where hemizygous for the transgene (homozygotes are highlighted by an asterisk); (C) a close view of pups hemizygous (left) and homozygous (right) for the transgene; (D) tail tips from E11.5 embryos homozygous (top), hemizygous (middle) and non-transgenic (bottom).

obtained, the developmental potential of the cell line appeared to be equivalent with the original R1 line (Nagy et al., 1993), therefore it may give the same, high germline compatibility. These cells could be used for mutant cell transplantation experiments or chimeric analyses, where the GFP expression provides an exceptional tool to follow the behavior of the mutant cells during development or dis-

ease processes, such as cancer. The visualization of individual EGFP expressing cells within a chimeric population suggests that real-time tracking of cells can be performed, by allowing multiple experimental time points to be obtained from a single sample. Additionally, if further wtGFP variants, such as the EBFP and EYFP versions, are shown to work in mouse transgenics, one can envisage that several different cellular populations can be non-invasively and simultaneously observed in real time in a single animal. We have demonstrated that green fluorescence can be used as a means to identify and dissect early preimplantation embryos at stages (E4.75–5.75) that have previously been elusive to all but the most highly skilled researchers. Beside the above main applications we envision other important uses for our GFP tagged ES-cell/transgenicmouse system. For example, green fluorescent ES cells should increase the efficiency of chimera production by allowing the preselection of chimeric blastocysts prior to transfer into recipient females. Furthermore, co-injection of zygotes, or co-electroporation of ES cells, with the DNA of interest along with a GFP expressing plasmid could provide a means to tag a transgenic line in random integration based transgenic approaches, thereby easing the burden of genotyping. The tetraploid embryo–ES cell aggregation provides a fast and inexpensive access to the embryonic phenotypes of mutations created in ES cells and also generates polarized chimeras in which the embryo proper, amnion and yolk sac mesoderm is completely ES cell-derived, whereas the yolk sac endoderm and trophoblast lineages are tetraploid embryo-derived (Nagy et al., 1990; Tanaka et al., 1998). We have demonstrated the feasibility of this technology in an earlier study (Carmeliet et al., 1996). In such studies it is essential to tag the tetraploid cells with a reporter gene to ensure that the embryos are completely mutant ES cellderived. Until now, the Rosa-26 transgenic line (Friedrich and Soriano, 1991), providing ubiquitous expression of a lacZ reporter, has been the obvious choice. Here we demonstrate the superiority of our GFP transgenic mouse line for tagging the tetraploid compartment of an embryo. Through demonstrating some of the applications of the ubiquitous GFP reporter system in ES cells and the corresponding transgenic mouse line, we feel that the intensively evolving mouse genetic technologies will incorporate this long-awaited tool in many aspects of its methodologies.

4. Experimental procedures 4.1. Generation of green ES cells pCX-EGFP, an EGFP-containing vector, has been described previously (Okabe et al., 1997). It contains EGFP inserted into pCAGGS, a plasmid vector containing a chicken b-actin promoter and CMV enhancer, b-actin intron and rabbit b-globin polyadenylation signal (Niwa et

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

al., 1991). pPGKPuro, a puromycin resistance gene-containing vector containing a PGK promoter (Tucker et al., 1996) was used for co-electroporation since it carries an antibiotic resistance. R1 ES cells (Nagy et al., 1993) were maintained as described previously (Hadjantonakis et al., 1998b; Pirity et al., 1998). Cells were co-electroporated under standard conditions with 40–80 mg of linearized pCX-EGFP (BamHI or ScaI were used interchangeably for linearization) and 10 mg of circular pPGKPuro. Each individual electroporation was seeded onto a single 10-mm gelatin-coated tissue culture plate and resulted in approximately 80–140 puromycinresistant colonies, 8–10 of which were green as observed under the appropriate optics. Control electroporations were carried out using one or other of the plasmids singly with subsequent selection in puromycin. Selection in puromycin was initiated 36–48 h after electroporation, and continued for 3–4 days thereafter. Colonies were viewed under a dissecting microscope (Leitz MZ8) equipped with a GFP illumination source (Leica) and scored for green by etching the underside of the tissue culture plate with a needle. Colonies assigned as being green were picked into 96-well plates for expansion and freezing (Hadjantonakis et al., 1998b). Embryoid bodies were obtained by seeding ES cells into bacteriological plates and by omitting LIF and b-mercaptoethanol from the culture medium. After approximately 1 week, embryoid bodies grown in suspension were replated onto gelatinized plates, where they attached and differentiated further. 4.2. Diploid embryo–ES cell aggregations Diploid aggregations were performed as described previously (Wood et al., 1993; Tanaka et al., 1998). A detailed protocol can be obtained on the world-wide web at http:// www.mshri.on.ca/develop/nagy/Diploid/diploid.htm. After overnight incubation with cells in aggregation plates, embryos were either transferred to recipient females (for the production of germline transmitting chimeras or postimplantation embryo dissections), seeded onto feeder cell containing plates (for the analysis of blastocyst outgrowths), or placed in M2 medium in depression glass microscope slides (for photography). Germline transmitting male chimeras were crossed with ICR females. Subsequently, the B5/EGFP mice have been maintained on an ICR background in our laboratory. 4.3. Tetraploid embryo–ES cell aggregations The technique relies on the electrofusion of two-cell stage embryos placed between the electrodes of an electrode chamber slide, resulting in rendering their genome tetraploid. The resulting one-cell stage embryos are incubated in vitro until they reach the four-cell stage of development at which time they are used to ‘sandwich’ a clump of ES cells. This arrangement when incubated overnight promotes the

89

aggregation of the two embryos and the intervening ES cells, resulting in the formation of a blastocyst. Tetraploid embryo–ES cell aggregations were performed as described previously (Nagy and Rossant, 1993; Tanaka et al., 1998). A detailed protocol can be obtained on the world-wide web at http://www.mshri.on.ca/develop/nagy/Tetraploid/Tetra.htm. After aggregation, embryos were treated exactly as for diploid aggregations. 4.4. Observation of green fluorescence Oviducts or uteruses of superovulated females were flushed with M2 for recovery of preimplantation stage embryos and cultured in M16 or KSOM media (Hogan et al., 1994; Tanaka et al., 1998). Postimplantation embryos were dissected in phosphate-buffered saline (PBS) and viewed under both dark-field and bright-field light optics. We noted that certain types of plasticware were refractory to fluorescence, and that glass vessels and microscope slides gave superior clarity. For plasticware we routinely used Costar 10 cm tissue culture, Fisher 10 cm bacteriological, and Falcon 3 cm tissue culture and organ culture dishes for viewing cells and embryos under fluorescent optics. For long-term storage and sectioning, embryos and adult tissues were fixed in 4% paraformaldehyde at 4°C for between 30 min to 4 h. Fifty-mm sections were cut in icecold PBS using a vibratome. Sections were mounted under coverslips in 1:1 glycerol/PBS and were stored at 4°C. EGFP fluorescence was examined under dissecting microscopes (Leitz MZ8 and MZ12) or microscopes (Leitz DMRB or DMRXE) equipped with GFP excitation sources and appropriate filters (Leica GFP Plus fluorescence filter set). Photographs of fluorescent subjects were taken on Kodak P1600 film shot at 800 or 1600 ASA. Mice were routinely genotyped without magnification using a Volpi fiber-optic light source fitted with a blue excitation filter and viewed through a yellow (520 nm) barrier filter (Chroma). Photographs of pups and mothers were taken using a Nikon FE10 50-mm camera fitted with a 35–75 mm lens, a set of close-up filters, and a yellow (520 nm) barrier filter (chroma).

Acknowledgements We thank Dr. Janet Rossant for valuable comments on the manuscript; Tim McGill and Michelle Shackleton at Leica Canada for help and advice; and Lajos Laszlo at BLS, Hungary, for developing the light source and eyeglasses for the macroscopic observation of green fluorescent animals. A.K.H. would also like to thank Dr. Laure Jamot for expert instruction on the use of a vibratome, and Dr Sue Quaggin for nephro-geographics. This work was funded by a grant from the Multiple Sclerosis Society of Canada and the National Cancer Institute of Canada (A.N.), and the Program for Promotion of Basic Research Activities for Inno-

90

A.-K. Hadjantonakis et al. / Mechanisms of Development 76 (1998) 79–90

vative Biosciences Japan (M.O.). A.N. is an MRC/BristolMyers Squib scientist. References Brejc, K., Sixma, T.K., Kitts, P.A., Kain, S.R., Tsien, R.Y., Ormo, M., Remington, S.J., 1997. Structural basis for dual excitation and photoisomerization of the Aequorea victoria green fluorescent protein. Proc. Natl. Acad. Sci. USA 94, 2306–2311. Carmeliet, P., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L., Gertsenstein, M., Fahrig, M., Vandenhoeck, A., Harpal, K., Eberhardt, C., Declerq, C., Pawling, J., Moons, L., Collen, D., Risau, W., Nagy, A., 1996. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 380, 435–439. Chalfie, M., Tu, Y., Euskirchen, G., Ward, W.W., Prasher, D.C., 1994. Green fluorescent protein as a marker for gene expression. Science 263, 802–805. Cormack, B.P., Valdivia, R.H., Falkow, S., 1996. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173, 33–38. Duncan, S.A., Nagy, A., Chan, W., 1997. Murine gastrulation requires HNF-4 regulated gene expression in the visceral endoderm: tetraploid rescue of Hnf-4 − / − embryos. Development 124, 279–287. Friedrich, G., Soriano, P., 1991. Promoter traps in embryonic stem cells: a genetic screen to identify and mutate developmental genes in mice. Genes Dev. 5, 1513–1523. Hadjantonakis, A.-K., Gertsenstein, M., Ikawa, M., Okabe, M., Nagy, A., 1998a. Non-invasive sexing of preimplantation mammalian embryos. Nature Genet. 19, 220–222. Hadjantonakis, A.-K., Pirity, M., Nagy, A., 1998b. In: P.T. Sharpe, I.J. Mason (Eds.), Vertebrate Embryology: Methods and Protocols. Humana Press, Totowa, NJ. Heim, R., Tsien, R.Y., 1996. Engineering green fluorescent protein for improved brightness, longer wavelengths and fluorescence resonance energy transfer. Curr. Biol. 6, 178–182. Hogan, B., Beddington, R., Constantini, F., Lacy, E., 1994. Manipulating the Mouse Embryo. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Ikawa, M., Kominami, K., Yoshimura, Y., Tanaka, K., Nishimune, Y., Okabe, M., 1995. Green fluorescent protein as a marker in transgenic mice. Dev. Growth Differ. 37, 455–459. Li, X., Wang, W., Lufkin, T., 1997. Dicistronic LacZ and alkaline phosphatase reporter constructs permit simultaneous histological analysis of expression from multiple transgenes. Biotechniques 23, 874–878. Li, X., Zhang, G., Ngo, N., Zhao, X., Kain, S.R., Huang, C.C., 1997. Deletions of the Aequorea victoria green fluorescent protein define

the minimal domain required for fluorescence. J. Biol. Chem. 272, 2855–2859. Nagy, A., 1996. The power of the new mouse genetics. J. Neurochem. 66 (Suppl. 1), 76. Nagy, A., Gocza, E., Diaz, E.M., Prideaux, V.R., Ivanyi, E., Markkula, M., Rossant, J., 1990. Embryonic stem cells alone are able to support fetal development in the mouse. Development 110, 815–821. Nagy, A., Rossant, J., 1993. In: A. Joyner (Ed.), Production of Completely ES Cell-derived Fetuses. IRL Press at Oxford University Press, Oxford, pp. 147–179. Nagy, A., Rossant, J., Nagy, R., Abramow-Newerly, W., Roder, J., 1993. Derivation of completely cell culture-derived mice from earlypassage embryonic stem cells. Proc. Natl. Acad. Sci. USA 90, 8424– 8428. Niwa, H., Yamamura, K., Miyazaki, J., 1991. Efficient selection for highexpression transfectants with a novel eukaryotic vector. Gene 108, 193– 199. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., Nishimune, Y., 1997. ‘Green mice’ as a source of ubiquitous green cells. FEBS Lett. 407, 313–319. Pirity, M., Hadjantonakis, A.-K., Nagy, A., 1998. In: J.P. Mather, D. Barnes (Eds.), Embryonal Stem Cells, Creating Transgenic Animals, Vol. 57. Academic Press, San Diego, CA, pp. 279–293. Prasher, D.C., 1995. Using GFP to see the light. Trends Genet. 11, 320– 323. Tanaka, M., Hadjantonakis, A.-K., Nagy, A., 1998. In: I. Kola, M.J. Timms (Eds.), Aggregation Chimeras; Combining ES Cells, Diploid and Tetraploid Embryos. Humana Press, Totowa, NJ. Tucker, K.L., Beard, C., Dausmann, J., Jackson-Grusby, L., Laird, P.W., Lei, H., Li, E., Jaenisch, R., 1996. Germ-line passage is required for establishment of methylation and expression patterns of imprinted but not of nonimprinted genes. Genes Dev. 10, 1008–1020. Wood, S.A., Pascoe, W.S., Schmidt, C., Kemler, R., Evans, M.J., Allen, N.D., 1993. Simple and efficient production of embryonic stem cellembryo chimeras by coculture. Proc. Natl. Acad. Sci. USA 90, 4582– 4585. Zernicka-Goetz, M., Pines, J., McLean Hunter, S., Dixon, J.P.C., Siemering, K.R., Haseloff, J., Evans, M.J., 1997. Following cell fate in the living mouse embryo. Development 124, 1133–1137. Zernicka-Goetz, M., Pines, J., Ryan, K., Siemering, K.R., Haseloff, J., Evans, M.J., Gurdon, J.B., 1996. An indelible lineage marker for Xenopus using a mutated green fluorescent protein. Development 122, 3719– 3724. Zhuo, L., Sun, B., Zhang, C.-L., Fine, A., Chui, S.-Y., Messing, A., 1997. Live astrocytes visualized by green fluorescent protein in transgenic mice. Dev. Biol. 187, 36–42.