- Email: [email protected]
A gene trap integration provides an early in situ marker for hepatic specification of the foregut endoderm
Mechanisms of Development 100 (2001) 205±215
www.elsevier.com/locate/modo
A gene trap integration provides an early in situ marker for hepatic speci®cation of the foregut endoderm Alistair J. Watt a,b,1, Elizabeth A. Jones c,d, Jan M. Ure a, Diana Peddie a, David I. Wilson c,d,2, Lesley M. Forrester a,b,* b
a Centre for Genome Research, University of Edinburgh, Kings Buildings, West Mains Road, Edinburgh EH9 3JQ, UK John Hughes Bennett Laboratory, Department of Oncology, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK c Institute of Human Genetics, School of Biochemistry and Genetics, University of Newcastle upon Tyne, Ridley Building, Claremont Road, Newcastle upon Tyne NE1 7RU, UK d Centre for Liver Research, School of Clinical Medical Sciences, Framlington Place, Newcastle upon Tyne NE2 4HH, UK
Received 6 September 2000; received in revised form 23 October 2000; accepted 27 October 2000
Abstract We report the characterization of a gene trap integration that provides an in situ marker for one of the earliest events in liver development. Expression of the reporter gene is observed at the nine-somite stage in the hepatic ®eld of the foregut endoderm. At 10.5 days post-coitus expression is observed exclusively and at high levels in the majority of cells in the developing liver bud. As development proceeds the proportion of expressing cells decreases with expression in adult liver being restricted to a few sporadic cells. This therefore provides the earliest, most speci®c in situ marker of the hepatic lineage reported to date and will be useful in the further characterization of the inductive events involved in hepatic speci®cation. Molecular characterization of the gene trap insertion suggests that the expression pattern is the result of alternative promoter use in the ankyrin repeat-containing gene, gtar. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Embryonic stem cells; Embryonic development; Alternative promoter; gtar
1. Introduction Development of the mammalian liver provides a relatively simple system to study the molecular mechanisms involved in organogenesis. Moreover, insights gleaned from these studies may be directly applicable to hepatic stem cell biology and pathology. The tissue interactions necessary for the differentiation of the de®nitive endoderm into the hepatic lineage, de®ned over 20 years ago, are only now beginning to be compounded by knowledge of the molecules mediating these interactions (Zaret, 2000). In the mouse, the ®rst morphological sign of liver development is at 8.5 days post-coitum (dpc) when the ventral * Corresponding author. John Hughes Bennett Laboratory, Department of Oncology, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK. Tel.: 144-131-537-1763; fax: 144131-537-3160. E-mail address: [email protected] (L.M. Forrester). 1 Present address: Department of Cell Biology, Neurobiology and Anatomy, Medical College of Wisconsin, Milwaukee, WI 53211, USA. 2 Present address: Human Genetics, The Duthie Building, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK.
¯oor of the foregut endoderm thickens to form the liver diverticulum. Subsequently, the liver bud is formed as the proliferating epithelial cells of the liver diverticulum invade the loose mesenchyme of the septum transversum (Douarin, 1975; Zaret, 1996, 1998). The liver becomes increasingly structured as the endodermal cells, migrating as cords, are lined by the mesenchymal cells forming the basic (adult) hepatic architecture of endoderm-derived parenchyma and mesenchyme-derived blood sinusoids (Severn, 1972). Using tissue explant cultures, two separate tissue interactions have been identi®ed which drive the initial stages of hepatic development. At the four- to six-somite stage (8.0±8.5 dpc), prior to morphological signs of liver development, hepatic speci®cation occurs as the pre-cardiac mesoderm induces the expression of the hepatic marker genes a-foetoprotein (afp) and serum albumin (alb) in the adjacent foregut endoderm (Gualdi et al., 1996). Subsequently, the closure and extension of the foregut brings the hepatic endoderm into contact with the septum transversum which induces the proliferation and migration of the hepatic endoderm (Douarin, 1975; Houssaint, 1980). Several molecules have been implicated in the early differentiation of the liver.
0925-4773/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00530-X
206
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
The transcription factors GATA-4 and HNF-3b have been shown to bind to distinct sites in the liver-speci®c alb enhancer element in undifferentiated gut endoderm. This binding is thought to potentiate the expression of hepatic-speci®c genes upon hepatic speci®cation (Gualdi et al., 1996; Bossard and Zaret, 2000). Fibroblast growth factor (FGF)1 and FGF2 can substitute for the inductive effect of cardiac mesoderm on foregut endoderm expressing FGF receptors 1 and 4. The subsequent outgrowth of the liver can be mediated, in part, by FGF8 (Jung et al., 1999). Afp and alb have been used as markers of hepatocyte differentiation in studies on the processes involved in hepatic development (Cascio and Zaret, 1991; Gualdi et al., 1996; Houssaint, 1980; Jung et al., 1999). Both afp and alb are not detected in situ in the developing liver until approximately 9.5 dpc (Cascio and Zaret, 1991; Schmid and Schulz, 1990; Shiojiri, 1981; Shiojiri et al., 1991). Consequently, their use as hepatic markers for the earliest stages of liver development has required the use of sensitive PCR-based techniques. Afp transcripts are detected throughout the de®nitive endoderm from the four-somite stage (Gualdi et al., 1996) and alb transcripts in the presumptive hepatic endoderm from the seven- to eight-somite stage. We report the characterization of a gene trap integration, I114, which provides the earliest and most speci®c in situ marker of hepatic differentiation that has been described to date. Gene trapping introduces a reporter gene fused to a consensus splice acceptor (SA) sequence into the ES cell genome (Skarnes, 1990). Integration of the gene trap vector into a transcription unit in the correct orientation will result in the splicing of the gene trap vector to an endogenous upstream exon. One consequence of this is that the reporter expression re¯ects endogenous gene expression (Skarnes et al., 1992). Reporter expression associated with the I114 integration was ®rst identi®ed in the nine-somite stage embryo exclusively in the hepatic endoderm and subsequently in the foetal liver until 15.5 dpc. Molecular characterization of the I114 integration has identi®ed a novel ankyrin-like (ANK) repeat-containing gene (gtar). Foetal liver-speci®c expression is most likely to be the consequence of alternate promoter use that is predicted to produce an embryonic liver-speci®c gtar isoform. The further characterization of the liver-speci®c control elements within this genomic region will de®ne a sequence which will mark the hepatic lineage.
embryos at this stage revealed that the expression is restricted to the ventral cells of the foregut endoderm prior to fusion (Fig. 1C). At 9.5 dpc reporter gene expression is observed in the cells migrating into the septum transversum (Fig. 1D) in an identical pattern to that observed using AFP riboprobes on comparable sections (Fig. 1E). b-Gal activity is maintained exclusively in the foetal liver with high levels observed at 10.5 dpc (27±29 somites) (Fig. 1F,G). From 10.5 dpc, b-gal activity in the developing liver is maintained throughout gestation and in neonates (data not shown). Sections of 11.5 dpc foetal liver show reporter activity restricted to hepatic parenchyme and not in the haematopoietic cells which have colonized the foetal liver by this stage (data not shown). As gestation proceeds the ®rst additional site of b-gal activity out with the liver is observed in the upper lip at 15.5 dpc and in the dorsal root ganglia of the peripheral nervous system at 17.5 dpc (data not shown). These expression data were generated from embryos heterozygous for the I114 gene trap integration after ®ve generations of backcrossing to the C57BL/6 background and were not signi®cantly affected by genetic background. An identical pattern was observed in I114 homozygous embryos derived from (129/CGR £ MF1) F1 intercrosses and (129/CGR £ 129/CGR) F3 intercrosses (data not shown). X-Gal staining was performed on cryostat sections from a range of adult tissues. Reporter expression was observed in isolated single cells of the liver (Fig. 1H,I), in the oocytes of the ovaries, the seminiferous tubules of the testes and the pelvic region of the kidney (data not shown). No reporter activity was detected in the heart, gut, lung, spleen and skeletal muscle. 2.2. Phenotypic analysis
2. Results
Heterozygous I114 mice were intercrossed to determine if the gene trap integration had any effect on development in the homozygous state. Homozygous animals were produced at the expected Mendelian ratio on the MF1 outbred background at the F1 generation, the 129/CGR background at the F3 and F4 generation and the C57BL/6 background at the F5 generation (data not shown). Intercrossing of F5 generation C57BL/6 I114 homozygous mice has produced a stable line indicating that the integration has no overt effect on viability or fertility. However, we have not generated a null allele at this locus as substantial levels of wild type gtar transcripts have been detected in homozygous tissue (see Fig. 5B). This has been reported by others and is thought to result from splicing around the gene trap vector.
2.1. Reporter gene expression
2.3. Molecular analysis
The I114 gene trap cell line was identi®ed in a gene trap screen designed to identify genes with a developmentally restricted expression pattern (Forrester et al., 1996). b-Gal activity was observed in the nine-somite stage embryo at 8.0±8.5 dpc in the foregut pocket (Fig. 1A,B). Sectioning of
Cloning and sequencing of 5 0 -RACE-PCR products from I114 ES cell and I114 homozygous foetal liver RNA identi®ed two different fusion transcript sequences, termed Group I and Group II, in a ratio of 5:1, respectively. The largest Group I (L-A8) and Group II (L-69) RACE clones
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
207
Fig. 1. Reporter gene expression embryos carrying the I114 gene trap integration. (A) Lateral view of an 8.5 dpc embryo showing b-gal activity in the foregut pocket in relation to the heart tube. (B) A ventral view of a nine-somite stage embryo showing b-gal activity in the foregut pocket showing the plane of sectioning in (C). (C) Transverse sections of an 8.5 dpc embryo showing b-gal activity in the foregut endoderm and presumptive hepatic endoderm. (D) b-Gal activity in the liver of a 9.5 dpc embryo showing migrating hepatoblasts. (E) AFP in situ analysis of a 9.5 dpc embryo showing migrating AFP-positive hepatoblasts. (F) Whole-mount b-gal staining of a 10.5 dpc embryo. (G) b-Gal activity in a transverse section of a 10.5 dpc embryo. (H,I) Cryostat sections showing b-gal expression in sporadic cells of the adult liver. Foregut (f), heart (h), neural tube (nt) and septum transversum (st) are indicated and scale bars represent 500 mm (A,C±H), 300 mm (B), and 50 mm (I).
contained 79 and 58 base pairs, respectively (Fig. 2A,B). The presence of multiple fusion transcript-producing gene trap cell lines has been identi®ed previously (Chowdhury et al., 1997; Townley et al., 1997) but has not been characterized in detail. RNase protection assays were carried out to con®rm the presence of the two different fusion transcripts and to determine their expression pro®le in I114 embryonic tissues. Expression of the Group I fusion transcript is observed in the liver, head and the rest of the body (Fig. 2C), whereas the Group II fusion transcript was expressed exclusively in the foetal liver (Fig. 2D), thus correlating with the reporter activity. The liver-speci®c expression of the Group II transcript was con®rmed by RT-PCR (Fig. 2E). The expected size product was generated using I114 ES cell and I114 heterozygous embryonic liver RNA but no product
was detected in RNA isolated from the head and only a faint band was observed in the rest of the body. It is most likely that this represents liver tissue contaminating the rest of the body tissue during dissection. Characterization of the endogenous cDNA (see below) and predicted translation of the fusion transcripts revealed that the Group I fusion places the lacZ sequence out of frame, whereas the Group II fusion sequence is untranslated and would presumably generate b-gal protein from the start codon of lacZ. This prediction is supported by Western blot analysis that showed that the b-gal protein was expressed exclusively in the foetal liver of I114 embryos (data not shown). The Group II fusion transcript is therefore most likely responsible for the liver-speci®c b-gal activity observed in embryos carrying the I114 gene trap integration.
208
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
Fig. 2. Sequence and expression of the Group I and Group II fusion transcripts. (A) Group I RACE clone sequence. The lower case sequence at the 3 0 end of each fusion transcript corresponds to the en-2 exon sequence from the PT1-ATG gene trap vector. (/) corresponds to the splice junction and the underlined poly G sequence corresponds to the arti®cial tail added during 5 0 RACE-PCR. (B) Group II RACE clone sequence. As above except that the underlined poly T sequence corresponds to the arti®cial tail added during 5 0 RACE-PCR. (C) RNase protection assay showing the Group I fusion transcript protected as a 199 bp fragment consisting of 79 bp of endogenous sequence and 120 bp of en-2 SA sequence. Expression is observed in I114 ES cells (lane 2), and in the liver (L), head (H) and the rest of the body (R) of I114 homozygous (2/2) and heterozygous (1/2) 13 dpc embryos (lanes 3±8), but not in the parental cell line R1 (lane 1) nor in tissues from wild type (1/1) 13 dpc embryos (lanes 9±11). Loading control GAPdH transcripts are protected as two fragments of 65 and 67 bp. (D) RNase protection assay showing the Group II fusion transcript protected as a 188 bp fragment consisting of 58 bp of endogenous sequence and 130 bp of en-2 SA sequence. Expression is observed in I114 ES cells (lane 1) and in the liver (L) but not the head (H) and the rest of the body (R) of I114 homozygous (2/2) 13 dpc embryos (lanes 2±4) nor in wild type 13 dpc embryonic tissues (lanes 5±7). (E) RT-PCR ampli®cation of Group II fusion transcripts using primers LST1 (complementary to the Group II sequence) and 78 (complementary to lacZ). The expected 377 bp product is observed at high levels in I114 ES cells (lane 1) and in the liver (L) of I114 homozygous (2/2) 13 dpc embryos (lane 2). No product was detected in the head (lane 3) and low levels were detected in the rest of the body (R) sample (lane 4), which likely re¯ects liver tissue contamination. I114 homozygous liver RNA subjected to the RT-PCR protocol in the absence of reverse transcriptase was used to control for genomic DNA contamination (lane 5). Ampli®cation of HPRT is used to control for the amount of total RNA used in the reactions.
2.4. cDNA cloning of gtar The Group I RACE clone was used to screen a random primed D3 ES cell cDNA library (gift from Dr Hitoshi Niwa). Seven clones were isolated and sequenced providing a contiguous sequence of 6040 nucleotides (Fig. 3). Conceptual translation of this sequence identi®es an open reading frame encoding a protein of 1599 amino acids with a predicted molecular weight of 170 kDa. The predicted protein contains 25 ANK-like repeat sequences in two separate domains of 15 and 10, three putative bipartite nuclear localization signals and a potential PEST sequence which includes a hyperacidic cluster (Rogers et al., 1986). We have named this gene gtar (Gene Trap Ankyrin Repeat). The Group I RACE clone sequence is within the coding region of the cDNA, between positions 825 and 903, and transla-
tion of the Group I fusion sequence from the gtar translational start codon con®rms that lacZ would be out of frame in this fusion transcript. We can predict that the gene trap vector is inserted into an intron between the splice donor (SD) site at position 903 and the putative downstream endogenous SA site at position 904. Two gtar splice isoforms were identi®ed that lacked the sequence between nucleotides 2843 and 3594. One of these isoforms also lacked the sequence between nucleotides 3722 and 4944 which contains the second 10 repeat ANK domain and a region of the ®rst bipartite NLS. Two colour ¯uorescent in situ hybridization FISH analysis was performed on I114 ES cell metaphase chromosome spreads using FITC-labelled PT1-ATG and pGC7, a 5.1 kb cDNA fragment of gtar (residues 1025±6139) labelled with Texas Red, as probes. The gene trap vector and gtar
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
209
Fig. 3. cDNA sequence and predicted translation of gtar. A stop codon (TAG, bold, underlined) at position 28 de®nes the translational start codon at nucleotide position 523 (bold). Bold underlined nucleotides at 520(23) and 526(14) are in agreement with the Kozak consensus sequence for translational initiation (RNNatgG; R, purine). The Group I RACE clone sequence is outlined from nucleotide position 825±903. The arrow between nucleotides 903 and 904 corresponds to the predicted vector insertion site and additional splice sites are identi®ed by arrows at nucleotides 2843 (SD), 3594 (SA), 3722 (SD) and 4944 (SA). Amino acid residues 113±124 are underlined denoting the hyperacidic cluster. The putative PEST sequence is ¯anked by positive amino acid residues at position 79 (R1 . ) and 148 (,1K). Ankyrin repeats are underlined and numbered (R1±R25) with highly conserved residues in bold. The three separate NLS signals are underlined with bold residues between amino acid residues 1468 and 1495.
210
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
resolved to a single site on chromosome 5 band E1 (data not shown) supporting that fact that we have indeed integrated into the region of the gtar locus. Interestingly, the gtar cDNA sequence shows similarity to the genomic sequence from human chromosome 4 (EMBL AC053525) in a region that is known to be syntenic to mouse chromosome 5. 2.5. Genomic characterization Three clones were isolated from the RPCI21 mouse PAC library using the Group II RACE clone L-69 as a probe. Sequence analysis (Fig. 4B) of a 1.5 kb PstI restriction fragment (439-a23.fP) (Fig. 4A) from one of these clones (439-a23) revealed that the Group I and Group II sequences were separated by only 500 base pairs and were in the same orientation with respect to the SD sequences used by the gene trap vector. Both the Group I and Group II SD sites
show a good overall match to the SD consensus sequence (Alberts et al., 1994) apart from a cytosine residue that replaces an almost invariant thymine at position 905 in the intron of the Group I SD site (Fig. 4B). The Group I and Group II sequences therefore represent alternative exons of the same gene that are differentially spliced to the gene trap vector and the two resulting fusion transcripts are differentially expressed. Interestingly, the transcriptional start site Wingender (TSSW) algorithm (UK HGMP Resource Centre) predicts a putative RNA polymerase II promoter region immediately upstream of the Group II sequence with the transcriptional start site at 1320 and a potential TATA box 28 nucleotides upstream (Fig. 4B). Position 1±903 of the 439-a23.fP genomic sequence which includes the Group I RACE sequence matches exactly the gtar cDNA sequence, indicating that this sequence represents a single exon. The promoter driving its expression is presumably upstream of the 439-a23.fP genomic fragment and distinct from the putative Group II promoter at position 1320. This identi®es a possible mode of differential expression between these two isoforms of the gtar gene. Repeated screening of cDNA libraries failed to identify a cDNA containing the Group II fusion sequence but transcripts containing the Group II fusion sequence can be predicted from its relative position in the genome (Fig. 5A). The existence of this predicted transcript and its foetal liver-speci®c expression was con®rmed by RT-PCR using two different primer pairs (Fig. 5A,B) and subsequent sequencing. Fig. 5C shows the sequence comparison of the two predicted isoforms of gtar. Translation of the Group II-containing transcript identi®es an ORF corresponding to the same reading frame as predicted for gtar. However, expression from the putative promoter of the Group II-containing transcript identi®es the ®rst start codon at nucleotide position 92 corresponding to the methionine codon at nucleotide 946 (amino acid 142) of the full-length (or Group I-containing) gtar sequence. This predicted protein will lack the ®rst 141 amino acid residues of the full-length GTAR protein which includes the putative PEST sequence and the hyperacidic cluster (Fig. 5C). 2.6. Promoter analysis
Fig. 4. Genomic analysis. (A) Schematic diagram showing the relative position of the Group I and Group II sequences in the genomic fragment, 439-a23.fp. (B) Sequence of the 1.5 kb genomic PstI fragment from PAC 439-a23. The outlined sequences from position 825±903 and 1323±1379 represent the Group I and Group II sequences, respectively. The sequence from position 1±903 is identical to 1±903 of gtar, identifying this as a single exon. The ATG highlighted at position 523 corresponds to the start codon of the endogenous gene. The putative transcriptional start site (11) of the Group II promoter is at position 1320 (bold, underlined) with the TATA box (228) at position 1292 (bold, underlined).
To test the activity of the predicted Group II promoter sequence, the 1.5 kb genomic fragment (439-a23.fP) was subcloned upstream of the IRESb-geo cassette (Fig. 6A) (construct 42). Control constructs consisted of the IRESbgeo cassette alone (construct 40) and the 1.5 kb genomic fragment subcloned in the reverse orientation (construct 43). Stably transfected ES cell lines were generated by co-electroporation with linearized PGK-puromycin and resistant colonies were screened for the presence of the test plasmid by Southern blotting using a probe to the neomycin resistance gene. All ten ES cell clones that carried the putative promoter construct in the correct orientation expressed the
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
211
Fig. 5. Expression analysis and sequence of the endogenous Group II transcript. (A) RT-PCR strategy. Primers complementary to the Group II sequence (LST1) and to gtar (RTANK-2 and RTANK-3) were used to amplify the predicted Group II-containing endogenous transcript, producing products of 338 and 651 bp, respectively. (B) RT-PCR products were analyzed after 30 PCR cycles (lanes 1±4). The expected PCR products from the LST-1/RTANK-2 and the LST-1/ RTANK-3 reactions were ampli®ed from R1 ES cells and wild type 10.5 dpc liver (L) (lanes 1 and 2), but not from wild type 10.5 dpc head (H) and the rest of the body (R) (lanes 3 and 4). Ampli®cation of HPRT was used to control for the amount of RNA and to control for genomic DNA contamination, RT-PCR reactions were performed on RNA from R1 ES cells and wild type 10.5 dpc foetal liver with (lanes 5 and 7) and without (lanes 6 and 8) reverse transcriptase. Lane 9 shows that both primer sets amplify product from I114 homozygous (2/2) 10.5 dpc RNA, indicating that the gene trap vector has failed to disrupt the expression of this transcript. (C) Sequence comparison of the two predicted isoforms of gtar. The Group II sequence splices to the endogenous downstream sequence via the SD used by the gene trap vector (49) and the SA used by the Group I sequence in gtar (50). Conceptual translation of the nucleotide sequence identi®es the same ORF as gtar with the start codon at nucleotide position 92 corresponding to nucleotide 946 and methionine 142 of the full-length gtar. The nucleotides underlined in the Group II sequence correspond to the PCR primer LST-1 used to amplify the sequence. (^) indicates the splice site.
reported gene suggesting that this sequence had some promoter activity. In contrast, no reporter activity was observed in clones carrying the control constructs (Fig. 6A). One of the clones carrying the promoter construct was injected into blastocysts and chimaeric embryos were harvested at 10 dpc. LacZ reporter expression was observed throughout these chimaeric embryos (Fig. 6B). This suggests that the 1.5 kb genomic sequence has promoter activity but must require additional regulatory elements to direct embryonic liver-speci®c expression. 3. Discussion We have identi®ed and characterized an unusual gene trap event that produces two fusion transcripts from the same integration site, one of which is an early marker for liver development. Although multiple transcript-producing gene trap lines have been described previously (Chowdhury et al., 1997; Townley et al., 1997), the mechanism of multiple transcript production has not been investigated in detail. We have shown that the gene trap vector alternately splices to two exons of the gtar gene that are 500 base pairs apart in
the genome. One of the resultant fusion transcripts is expressed throughout the embryo but because the lacZ sequence is not in the correct reading frame no b-gal activity is produced. An alternative in-frame fusion transcript results in foetal liver-speci®c reporter expression and we have identi®ed a promoter sequence associated with this transcript. Such a gene trap event has not previously been reported and highlights the sensitivity of the technique for identifying alternate promoter use within genes. The I114 gene trap integration has not generated a null allele of the gtar gene and has therefore no phenotypic consequences. This has been reported previously for a number of gene trap integrations and is predicted to be a consequence of splicing around the gene trap vector (McClive et al., 1998; Sam et al., 1996; Voss et al., 1998). We predict that this phenomenon is more likely in genes like gtar that exist as a variety of alternatively spliced forms. Although we have shown that the liver-speci®c Group II sequence has an associated promoter activity, we failed to identify the sequences necessary for liver-speci®c expression. This is not surprising considering the complex nature of promoter and enhancer elements necessary for tissuespeci®c expression reported in other systems. For example,
212
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
alb expression in the liver is dependent on both the promoter sequence and an enhancer element that lies 10 kb upstream (Pinkert et al., 1987) and tissue-speci®c afp expression requires three separate enhancers and a complex promoter sequence (Chen et al., 1997). The widespread reporter expression associated with the promoter fragment that we tested could be a consequence of removing the promoter from the in¯uence of liver-speci®c enhancer or repressor elements. Identi®cation of the necessary control elements for liver-speci®c expression will provide an excellent tool to drive expression of genes exclusively to the liver. For example, it could be used to drive Cre recombinase expression allowing embryonic liver-speci®c targeting of genes, such as HNF-3b and GATA4, that play a role in hepatic development, that have lethal phenotypes prior to hepatic speci®cation (Ang and Rossant, 1994; Kuo et al., 1997; Molkentin et al., 1997). The GTAR protein contains ankyrin repeats that are involved in protein±protein interactions and have been identi®ed in a wide range of genes such as transcription factors, cytoskeletal proteins and transmembrane proteins (Bork,
1993). The liver-speci®c GTAR isoform is predicted to lack the N-terminal 141 amino acid residues of the predicted full-length GTAR protein. This region includes a stretch of glycine residues and a hyperacidic cluster which are located within a putative PEST sequence. The presence of a PEST sequence is associated with proteins with a short half-life (Rogers et al., 1986) and so it is possible that the liverspeci®c isoform is more stable than the full-length GTAR. Alternatively, the hyperacidic cluster may function to modulate the binding speci®city of the ANK repeats in a tissue-speci®c manner. A splice isoform of the ank1 gene that is lacking an acidic portion of the regulatory domain shows enhanced ANK-mediated binding to integral membrane proteins (Lux et al., 1990). The absence of the hyperacidic cluster from the liver-speci®c gtar isoform may alter the binding speci®city of GTAR in the foetal liver in a similar manner. This gene trap integration provides the earliest exclusive marker for liver development reported to date. The ®rst site of reporter expression corresponds to the location of the presumptive hepatic endoderm that has been identi®ed in
Fig. 6. Promoter analysis. (A) Analysis of ES cell colonies generated after co-electroporation of plasmids containing puromycin under the control of the PGK promoter and test promoter plasmids as indicated. Puromycin resistant colonies were screened for the presence of the test plasmid DNA by Southern blotting and subsequently tested for b-gal activity. (B) Chimaeric embryo generated from the injection of ES cells carrying test plasmid 42 showing ubiquitous b-gal activity.
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
previous studies (Houssaint, 1980; Gualdi et al., 1996). The advantage of this marker is that it can be observed in situ at the single cell level that it is more speci®c than previously used markers. Afp, for example, is also expressed in the visceral endoderm of the yolk sac and at low levels throughout the de®nitive endoderm at the earliest stages of hepatic development (Dziadek and Adamson, 1978; Dziadek and Andrews, 1983; Gualdi et al., 1996). Embryos carrying this marker will be useful in tissue explant cultures designed to identify the factors involved in hepatic speci®cation and in the phenotypic analysis of mouse mutants with defects in liver development. I114 reporter activity is maintained in the foetal liver throughout gestation but as development proceeds expression is increasingly restricted to a subset of cells and is only seen in sporadic cells of the adult. The fact that these cells exist as a small subpopulation of phenotypically indistinct cells and express a foetal marker gene would support the idea that they correspond to the putative hepatic stem cells or oval cells (Alison, 1998; Golding et al., 1995). ES cells carrying this gene trap integration could be used to establish protocols to generate liver stem cells from ES cells in vitro and the subsequent application of this technology to the human ES cells system could lead to the development of somatic cell therapy for liver disease.
213
sectioned. Age-matched wild type embryos and tissues were stained in parallel with all samples to control for background b-gal activity. 4.3. RACE-PCR Rapid ampli®cation of cDNA ends was carried out as described previously (Townley et al., 1997) using the following primers. First strand cDNA was synthesized from total RNA using the primer 5 0 -TAATGGGATAGGTTACG-3 0 (78) complementary to lacZ. The ®rst round PCR reaction used the anchor primer 5 0 -GGTTGTGAGCTCTTCTAGATGGTTTTTTTTTTTTTTTTT-3 0 (56) paired with the nested lacZ-speci®c primer 5 0 -AGTATCGGCCTCAGGAAGATCG-3 0 (79). The second round PCR reaction used the nested anchor primer 5 0 -GGTTGTGAGCTCTTCTAGATGG-3 0 (59) paired with the primer 5 0 -TGCTCTGTCAGGTACCTGTTG-3 0 (55) complementary to the en-2 exon 2 sequence used in the gene trap vector splice site. The PCR conditions were 30 cycles of 948C for 1.5 min, 608C for 1.5 min and 728C for 3 min. The second round PCR reactions were carried out with either Taq polymerase for directional cloning into pBluescript II KS 2 using SpeI and KpnI or with Pfu polymerase (GibcoBRL) for cloning into pCR w-Blunt using the Zero Blunte PCR cloning kit (Invitrogen).
4. Experimental procedures
4.4. RNase protection analysis
4.1. Animal breeding and phenotype analysis
Group I (L-A8) and Group II (L-69) RACE clones were used to generate antisense riboprobes for hybridization to I114 and wild type embryonic tissue RNA. Riboprobes were produced by in vitro transcription using 500 ng of linearized plasmid DNA in the presence of 1 mM (250 mCi) [a- 32P]cytidine triphosphate (CTP) using either T3, T7 or SP6 RNase polymerases. In vitro transcription products were separated on 6% polyacrylamide gels and the longest transcripts were puri®ed from the gel using a volume of elution buffer (0.5 M ammonium acetate, 1 mM EDTA, 0.2% SDS). For each hybridization reaction, 10 mg target RNA, 3.5 £ 10 5 cpm of eluted riboprobe and 5 £ 10 4 cpm of glyceraldehyde 3-phosphate dehydrogenase (GAPdH) loading control riboprobe were precipitated, resuspended in hybridization buffer (80% formamide, 400 mM NaCl, 1 mM EDTA, 40 mM piperazine-N,N 0 -bis(2-etanesulfonic acid) (PIPES) (pH 6.4)) and incubated overnight at 558C. RNase A and RNAse T1 were used to digest unprotected RNA. The samples were run alongside a ddT terminated sequence of 240 primed bacteriophage M13mp18 control DNA as a size marker on a 6% polyacrylamide gel which was then dried and exposed to Xomat (Eastman Kodak Co., Rochester, NY) autoradiographic ®lm to visualize protected fragments.
The I114 gene trap integration was transmitted through the germline by injection of the ES cells into C57/Bl6 blastocysts. Chimaeric males were backcrossed onto 129/CGR and C57BL/6 inbred and MF1 outbred females. Offspring were genotyped by Southern or dot blot analysis of tail DNA using a lacZ probe. Homozygotes and heterozygotes were distinguished by quantitation of band or dot intensity using a Phosphoimager using the level of actin hybridization as an internal control. 4.2. Reporter gene expression Embryos were staged by the presence of a vaginal plug and, in some cases, by the number of somite pairs. Whole embryos (up to 11.5 dpc) were stained with X-gal as described previously (Gossler et al., 1989) and subsequently sectioned by dehydrating for 2 h each in 70, 80, 90 and 100% ethanol followed by an overnight incubation in Histoclear at room temperature. Embryos were incubated twice in molten paraf®n at 568C for 2 h, followed by 2 h in molten paraf®n under vacuum, embedded in the left lateral orientation and sectioned at 5 mm intervals. Post-13.5 dpc embryos were cryostat sectioned prior to X-gal staining as described previously (McClive et al., 1998). Adult tissues were dissected from 3- to 5-month-old animals homozygous or heterozygous for the I114 gene trap integration and cryostat
4.5. Reverse transcriptase PCR (RT-PCR) RT-PCR was carried out on RNA from I114 ES cells,
214
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
liver, head and the rest of the body from I114 homozygous 13 dpc embryos using the primers LST-1 (complementary to the Group II sequence) and LacZ-78 (complementary to lacZ), or R1 parental ES cell, liver, head and the rest the of body from wild type 10.5 dpc embryos using LST-1 in combination with two different primers, RTANK-2 and RTANK-3 (complementary to gtar). The reverse transcription reaction was carried out using 1 mg of total RNA (previously denatured at 708C for 5 min and then placed on ice) in 1£ PCR buffer (Promega) (2.5 mM MgCl2, 0.125 units of random hexamers (Boehringer Mannheim), 1 mM dNTPs and 200 units of MMLV reverse transcriptase) (GibcoBRL) in a total volume of 20 ml. Samples were incubated at 258C for 10 min and 428C for 1 h and then denatured at 958C for 10 min. Then, 4 ml of the ®rst strand cDNA reaction was added to 1.6 ml 10£ PCR buffer (1.6 ml 25 mM MgCl2, 100 ng primer x, 100 ng primer y, 0.2 ml (1 unit) Taq DNA polymerase (Promega) and water) to a ®nal volume of 20 ml. Samples were overlayed with mineral oil and PCR ampli®cation was carried out using 30 cycles of 968C for 5 s, 538C for 15 s and 728C for 1 min. RT-PCR products were run on an agarose gel, Southern blotted and probed with the 500 bp en-2 exon for LST-1/LacZ-78 and the RT-ANK-2 oligonucleotide for LST-1/RTANK-2 or RT-ANK3. The primer sequences used were 5 0 GGTAGTTTTCTGTCAGTGG-3 0 (LST-1), 5 0 -TAATGGGATAGGTTACG-3 0 (LacZ-78), 5 0 -CTGTGCTCTCAGCTCTCATC-3 0 (RTANK-2), 5 0 -GAAGACTGTGCATTGACATC-3 0 (RTANK-3), 5 0 -GCTGGTGAAAAGGACCTCT-3 0 (HPRT3 0 ) and 5 0 -CACAGGACTAGAACACCTGC-3 0 (HPRT5 0 ). HPRT was used to control for the even loading of RNA in the RT-PCR reactions (Johansson and Wiles, 1995). The endogenous Group II PCR products were subcloned using the TOPO TA cloning kit (Invitrogen). 4.6. PAC isolation and analysis The RPCI21 library (Osoegawa et al., 2000) was supplied by the UK HGMP Resource Centre (http:// www.hgmp.mrc.ac.uk) on gridded ®lters. The Group II clone, L-69, was hybridized to the ®lters in Church and Gilbert solution at 658C (Church and Gilbert, 1984) and washed in 1£ SSC, 0.1% SDS at 658C and three positive clones, 352-j9, 390-m1 and 439-a23, were identi®ed. DNA was prepared using the Qiagen Maxiprep Kit according to the manufacturer's instructions with the following modi®cations. Two hundred ®fty millilitres of overnight culture (LB 1 25 mg/ml Kanamycin) was used and 20 ml (twice the recommended amount) of each of the buffers P1, P2, P3 was added. For the elution of the plasmid from the Qiagen column, the elution buffer (QF) was heated to 658C and added in 2 ml aliquots. The 1.5 kb PstI fragment (439a23.fP) that was positive for the Group I and Group II sequences was subcloned into pZERO-2 (Invitrogen) and sequenced.
4.7. Promoter constructs The gene trap vector pGT1.8IRESbgeo (Mountford and Smith, 1995) was digested with SalI and BglII to remove the en-2 SA and re-ligated to generate construct 40. Alternatively, the genomic fragment 439-a23.fP was subcloned into the SalI/BglII site in both orientations to make constructs 42 and 43. Sca-1 linearized plasmid (40 mg) and SalI linearized pPGKpurobpA (4 mg) were electroporated into E14 ES cells (240 V, 500 mF). Cells were selected in puromycin at 2 mg/ ml for 8 days and colonies were picked into 24-well plates and expanded. Cells were X-gal stained and DNA extracted for Southern blotting as described previously (Forrester et al., 1996). Acknowledgements This work was funded from grants from the MRC (L.M.F. and A.J.W.), the BBSRC (Centre for Genome Research), the Geron Corporation, the Leukaemia Research Fund (LMF) and the Wellcome Tust (E.A.J. and D.W). References Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K., Watson, J.D., 1994. Molecular Biology of the Cell, 3rd edition, Garland, New York. Alison, M., 1998. Liver stem cells: a two compartment system. Curr. Opin. Cell Biol. 10, 710±715. Ang, S.L., Rossant, J., 1994. HNF-3 beta is essential for node and notochord formation in mouse development. Cell 78, 561±574. Bork, P., 1993. Hundreds of ankyrin-like repeats in functionally diverse proteins: mobile modules that cross phyla horizontally? Proteins 17, 363±374. Bossard, P., Zaret, K.S., 1998. GATA transciption factors as potentiators of gut endoderm differentiation. Development 125, 4909±4917. Cascio, S., Zaret, K.S., 1991. Hepatocyte differentiation initiates during endodermal-mesenchymal interactions prior to liver formation. Development 113, 217±225. Chen, H., Egan, J.O., Chiu, J.F., 1997. Regulation and activities of alphafetoprotein. Crit. Rev. Eukaryot. Gene Expr. 7, 11±41. Chowdhury, K., Bonaldo, P., Torres, M., Stoykova, A., Gruss, P., 1997. Evidence for the stochastic integration of gene trap vectors into the mouse germline. Nucleic Acids Res. 25, 1531±1536. Church, G.M., Gilbert, W., 1984. Genomic sequencing. Proc. Natl. Acad. Sci. USA 81, 1991±1995. Douarin, N.M., 1975. An experimental analysis of liver development. Med. Biol. 53, 427±455. Dziadek, M., Adamson, E., 1978. Localization and synthesis of alphafoetoprotein in post-implantation mouse embryos. J. Embryol. Exp. Morphol. 43, 289±313. Dziadek, M.A., Andrews, G.K., 1983. Tissue speci®city of alpha-fetoprotein messenger RNA expression during mouse embryogenesis. EMBO J. 2, 549±554. Forrester, L.M., Nagy, A., Sam, M., Watt, A., Stevenson, L., Bernstein, A., Joyner, A.L., Wurst, W., 1996. An induction gene trap screen in embryonic stem cells: identi®cation of genes that respond to retinoic acid in vitro. Proc. Natl. Acad. Sci. USA 93, 1677±1682. Golding, M., Sarraf, C.E., Lalani, E.N., Anilkumar, T.V., Edwards, R.J., Nagy, P., Thorgeirsson, S.S., Alison, M.R., 1995. Oval cell differentiation into hepatocytes in the acetylamino¯uorene-treated regenerating rat liver. Hepatology 22, 1243±1253.
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215 Gossler, A., Joyner, A.L., Rossant, J., Skarnes, W.C., 1989. Mouse embryonic stem cells and reporter constructs to detect developmentally regulated genes. Science 244, 463±465. Gualdi, R., Bossard, P., Zheng, M., Hamada, Y., Coleman, J.R., Zaret, K.S., 1996. Hepatic speci®cation of the gut endoderm in vitro: cell signaling and transcriptional control. Genes Dev. 10, 1670±1682. Houssaint, E., 1980. Differentiation of the mouse hepatic primordium. I. An analysis of tissue interactions in hepatocyte differentiation. Cell Diff. 9, 269±279. Johansson, B.M., Wiles, M.V., 1995. Evidence for involvement of activin A and bone morphogenetic protein 4 in mammalian mesoderm and hematopoietic development. Mol. Cell. Biol. 15, 141±151. Jung, J., Zheng, M., Goldfarb, M., Zaret, K.S., 1999. Initiation of mammalian liver development from endoderm by ®broblast growth factors. Science 284, 1998±2003. Kuo, C.T., Morrisey, E.E., Anandappa, R., Sigrist, K., Lu, M.M., Parmacek, M.S., Soudais, C., Leiden, J.M., 1997. GATA4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev. 11, 1048±1060. Lux, S.E., John, K.M., Bennett, V., 1990. Analysis of cDNA for human erythrocyte ankyrin indicates a repeated structure with homology to tissue-differentiation and cell-cycle control proteins. Nature 344, 36± 42. McClive, P., Pall, G., Newton, K., Lee, M., Mullins, J., Forrester, L., 1998. Gene trap integrations expressed in the developing heart: insertion site affects splicing of the PT1-ATG vector. Dev. Dyn. 212, 267±276. Molkentin, J.D., Lin, Q., Duncan, S.A., Olson, E.N., 1997. Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev. 11, 1061±1072. Mountford, P.S., Smith, A.G., 1995. Internal ribosome entry sites and dicistronic RNAs in mammalian transgenesis. Trends Genet. 11, 179± 184. Osoegawa, K., Tateno, M., Woon, P.Y., Frengen, E., Mammoser, A.G., Catanese, J.J., Hayashizaki, Y., de Jong, P.J., 2000. Bacterial arti®cial chromosome libraries for mouse sequencing and functional analysis. Genome Res. 10, 116±128. Pinkert, C.A., Ornitz, D.M., Brinster, R.L., Palmiter, R.D., 1987. An albumin enhancer located 10 kb upstream functions along with its promoter
215
to direct ef®cient, liver-speci®c expression in transgenic mice. Genes Dev. 1, 268±276. Rogers, S., Wells, R., Rechsteiner, M., 1986. Amino acid sequences common to rapidly degraded proteins: the PEST hypothesis. Science 234, 364±368. Sam, M., Wurst, W., Forrester, L., Vauti, F., Heng, H., Bernstein, A., 1996. A novel family of repeat sequences in the mouse genome responsive to retinoic acid. Mamm. Genome 7, 741±748. Schmid, P., Schulz, W.A., 1990. Coexpression of the c-myc protooncogene with alpha-fetoprotein and albumin in fetal mouse liver. Differentiation 45, 96±102. Severn, C.B., 1972. A morphological study of the development of the human liver. II. Establishment of liver parenchyma, extrahepatic ducts and associated venous channels. Am. J. Anat. 133, 85±107. Shiojiri, N., 1981. Enzymo- and immunocytochemical analyses of the differentiation of liver cells in the prenatal mouse. J. Embryol. Exp. Morphol. 62, 139±152. Shiojiri, N., Lemire, J.M., Fausto, N., 1991. Cell lineages and oval cell progenitors in rat liver development. Cancer Res. 51, 2611±2620. Skarnes, W.C., 1990. Entrapment vectors: a new tool for mammalian genetics. Biotechnology 8, 827±831. Skarnes, W.C., Auerbach, B.A., Joyner, A.L., 1992. A gene trap approach in mouse embryonic stem cells: the lacZ reported is activated by splicing, re¯ects endogenous gene expression, and is mutagenic in mice. Genes Dev. 6, 903±918. Townley, D.J., Avery, B.J., Rosen, B., Skarnes, W.C., 1997. Rapid sequence analysis of gene trap integrations to generate a resource of insertional mutations in mice. Genome Res. 7, 293±298. Voss, A.K., Thomas, T., Gruss, P., 1998. Compensation for a gene trap mutation in the murine microtubule-associated protein 4 locus by alternative polyadenylation and alternative splicing. Dev. Dyn. 212, 258± 266. Zaret, K.S., 1996. Molecular genetics of early liver development. Annu. Rev. Physiol. 58, 231±251. Zaret, K., 1998. Early liver differentiation: genetic potentiation and multilevel growth control. Curr. Opin. Genet. Dev. 8, 526±531. Zaret, K.S., 2000. Liver speci®cation and early morphogenesis. Mech. Dev. 92, 83±88.
www.elsevier.com/locate/modo
A gene trap integration provides an early in situ marker for hepatic speci®cation of the foregut endoderm Alistair J. Watt a,b,1, Elizabeth A. Jones c,d, Jan M. Ure a, Diana Peddie a, David I. Wilson c,d,2, Lesley M. Forrester a,b,* b
a Centre for Genome Research, University of Edinburgh, Kings Buildings, West Mains Road, Edinburgh EH9 3JQ, UK John Hughes Bennett Laboratory, Department of Oncology, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK c Institute of Human Genetics, School of Biochemistry and Genetics, University of Newcastle upon Tyne, Ridley Building, Claremont Road, Newcastle upon Tyne NE1 7RU, UK d Centre for Liver Research, School of Clinical Medical Sciences, Framlington Place, Newcastle upon Tyne NE2 4HH, UK
Received 6 September 2000; received in revised form 23 October 2000; accepted 27 October 2000
Abstract We report the characterization of a gene trap integration that provides an in situ marker for one of the earliest events in liver development. Expression of the reporter gene is observed at the nine-somite stage in the hepatic ®eld of the foregut endoderm. At 10.5 days post-coitus expression is observed exclusively and at high levels in the majority of cells in the developing liver bud. As development proceeds the proportion of expressing cells decreases with expression in adult liver being restricted to a few sporadic cells. This therefore provides the earliest, most speci®c in situ marker of the hepatic lineage reported to date and will be useful in the further characterization of the inductive events involved in hepatic speci®cation. Molecular characterization of the gene trap insertion suggests that the expression pattern is the result of alternative promoter use in the ankyrin repeat-containing gene, gtar. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Embryonic stem cells; Embryonic development; Alternative promoter; gtar
1. Introduction Development of the mammalian liver provides a relatively simple system to study the molecular mechanisms involved in organogenesis. Moreover, insights gleaned from these studies may be directly applicable to hepatic stem cell biology and pathology. The tissue interactions necessary for the differentiation of the de®nitive endoderm into the hepatic lineage, de®ned over 20 years ago, are only now beginning to be compounded by knowledge of the molecules mediating these interactions (Zaret, 2000). In the mouse, the ®rst morphological sign of liver development is at 8.5 days post-coitum (dpc) when the ventral * Corresponding author. John Hughes Bennett Laboratory, Department of Oncology, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK. Tel.: 144-131-537-1763; fax: 144131-537-3160. E-mail address: [email protected] (L.M. Forrester). 1 Present address: Department of Cell Biology, Neurobiology and Anatomy, Medical College of Wisconsin, Milwaukee, WI 53211, USA. 2 Present address: Human Genetics, The Duthie Building, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK.
¯oor of the foregut endoderm thickens to form the liver diverticulum. Subsequently, the liver bud is formed as the proliferating epithelial cells of the liver diverticulum invade the loose mesenchyme of the septum transversum (Douarin, 1975; Zaret, 1996, 1998). The liver becomes increasingly structured as the endodermal cells, migrating as cords, are lined by the mesenchymal cells forming the basic (adult) hepatic architecture of endoderm-derived parenchyma and mesenchyme-derived blood sinusoids (Severn, 1972). Using tissue explant cultures, two separate tissue interactions have been identi®ed which drive the initial stages of hepatic development. At the four- to six-somite stage (8.0±8.5 dpc), prior to morphological signs of liver development, hepatic speci®cation occurs as the pre-cardiac mesoderm induces the expression of the hepatic marker genes a-foetoprotein (afp) and serum albumin (alb) in the adjacent foregut endoderm (Gualdi et al., 1996). Subsequently, the closure and extension of the foregut brings the hepatic endoderm into contact with the septum transversum which induces the proliferation and migration of the hepatic endoderm (Douarin, 1975; Houssaint, 1980). Several molecules have been implicated in the early differentiation of the liver.
0925-4773/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00530-X
206
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
The transcription factors GATA-4 and HNF-3b have been shown to bind to distinct sites in the liver-speci®c alb enhancer element in undifferentiated gut endoderm. This binding is thought to potentiate the expression of hepatic-speci®c genes upon hepatic speci®cation (Gualdi et al., 1996; Bossard and Zaret, 2000). Fibroblast growth factor (FGF)1 and FGF2 can substitute for the inductive effect of cardiac mesoderm on foregut endoderm expressing FGF receptors 1 and 4. The subsequent outgrowth of the liver can be mediated, in part, by FGF8 (Jung et al., 1999). Afp and alb have been used as markers of hepatocyte differentiation in studies on the processes involved in hepatic development (Cascio and Zaret, 1991; Gualdi et al., 1996; Houssaint, 1980; Jung et al., 1999). Both afp and alb are not detected in situ in the developing liver until approximately 9.5 dpc (Cascio and Zaret, 1991; Schmid and Schulz, 1990; Shiojiri, 1981; Shiojiri et al., 1991). Consequently, their use as hepatic markers for the earliest stages of liver development has required the use of sensitive PCR-based techniques. Afp transcripts are detected throughout the de®nitive endoderm from the four-somite stage (Gualdi et al., 1996) and alb transcripts in the presumptive hepatic endoderm from the seven- to eight-somite stage. We report the characterization of a gene trap integration, I114, which provides the earliest and most speci®c in situ marker of hepatic differentiation that has been described to date. Gene trapping introduces a reporter gene fused to a consensus splice acceptor (SA) sequence into the ES cell genome (Skarnes, 1990). Integration of the gene trap vector into a transcription unit in the correct orientation will result in the splicing of the gene trap vector to an endogenous upstream exon. One consequence of this is that the reporter expression re¯ects endogenous gene expression (Skarnes et al., 1992). Reporter expression associated with the I114 integration was ®rst identi®ed in the nine-somite stage embryo exclusively in the hepatic endoderm and subsequently in the foetal liver until 15.5 dpc. Molecular characterization of the I114 integration has identi®ed a novel ankyrin-like (ANK) repeat-containing gene (gtar). Foetal liver-speci®c expression is most likely to be the consequence of alternate promoter use that is predicted to produce an embryonic liver-speci®c gtar isoform. The further characterization of the liver-speci®c control elements within this genomic region will de®ne a sequence which will mark the hepatic lineage.
embryos at this stage revealed that the expression is restricted to the ventral cells of the foregut endoderm prior to fusion (Fig. 1C). At 9.5 dpc reporter gene expression is observed in the cells migrating into the septum transversum (Fig. 1D) in an identical pattern to that observed using AFP riboprobes on comparable sections (Fig. 1E). b-Gal activity is maintained exclusively in the foetal liver with high levels observed at 10.5 dpc (27±29 somites) (Fig. 1F,G). From 10.5 dpc, b-gal activity in the developing liver is maintained throughout gestation and in neonates (data not shown). Sections of 11.5 dpc foetal liver show reporter activity restricted to hepatic parenchyme and not in the haematopoietic cells which have colonized the foetal liver by this stage (data not shown). As gestation proceeds the ®rst additional site of b-gal activity out with the liver is observed in the upper lip at 15.5 dpc and in the dorsal root ganglia of the peripheral nervous system at 17.5 dpc (data not shown). These expression data were generated from embryos heterozygous for the I114 gene trap integration after ®ve generations of backcrossing to the C57BL/6 background and were not signi®cantly affected by genetic background. An identical pattern was observed in I114 homozygous embryos derived from (129/CGR £ MF1) F1 intercrosses and (129/CGR £ 129/CGR) F3 intercrosses (data not shown). X-Gal staining was performed on cryostat sections from a range of adult tissues. Reporter expression was observed in isolated single cells of the liver (Fig. 1H,I), in the oocytes of the ovaries, the seminiferous tubules of the testes and the pelvic region of the kidney (data not shown). No reporter activity was detected in the heart, gut, lung, spleen and skeletal muscle. 2.2. Phenotypic analysis
2. Results
Heterozygous I114 mice were intercrossed to determine if the gene trap integration had any effect on development in the homozygous state. Homozygous animals were produced at the expected Mendelian ratio on the MF1 outbred background at the F1 generation, the 129/CGR background at the F3 and F4 generation and the C57BL/6 background at the F5 generation (data not shown). Intercrossing of F5 generation C57BL/6 I114 homozygous mice has produced a stable line indicating that the integration has no overt effect on viability or fertility. However, we have not generated a null allele at this locus as substantial levels of wild type gtar transcripts have been detected in homozygous tissue (see Fig. 5B). This has been reported by others and is thought to result from splicing around the gene trap vector.
2.1. Reporter gene expression
2.3. Molecular analysis
The I114 gene trap cell line was identi®ed in a gene trap screen designed to identify genes with a developmentally restricted expression pattern (Forrester et al., 1996). b-Gal activity was observed in the nine-somite stage embryo at 8.0±8.5 dpc in the foregut pocket (Fig. 1A,B). Sectioning of
Cloning and sequencing of 5 0 -RACE-PCR products from I114 ES cell and I114 homozygous foetal liver RNA identi®ed two different fusion transcript sequences, termed Group I and Group II, in a ratio of 5:1, respectively. The largest Group I (L-A8) and Group II (L-69) RACE clones
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
207
Fig. 1. Reporter gene expression embryos carrying the I114 gene trap integration. (A) Lateral view of an 8.5 dpc embryo showing b-gal activity in the foregut pocket in relation to the heart tube. (B) A ventral view of a nine-somite stage embryo showing b-gal activity in the foregut pocket showing the plane of sectioning in (C). (C) Transverse sections of an 8.5 dpc embryo showing b-gal activity in the foregut endoderm and presumptive hepatic endoderm. (D) b-Gal activity in the liver of a 9.5 dpc embryo showing migrating hepatoblasts. (E) AFP in situ analysis of a 9.5 dpc embryo showing migrating AFP-positive hepatoblasts. (F) Whole-mount b-gal staining of a 10.5 dpc embryo. (G) b-Gal activity in a transverse section of a 10.5 dpc embryo. (H,I) Cryostat sections showing b-gal expression in sporadic cells of the adult liver. Foregut (f), heart (h), neural tube (nt) and septum transversum (st) are indicated and scale bars represent 500 mm (A,C±H), 300 mm (B), and 50 mm (I).
contained 79 and 58 base pairs, respectively (Fig. 2A,B). The presence of multiple fusion transcript-producing gene trap cell lines has been identi®ed previously (Chowdhury et al., 1997; Townley et al., 1997) but has not been characterized in detail. RNase protection assays were carried out to con®rm the presence of the two different fusion transcripts and to determine their expression pro®le in I114 embryonic tissues. Expression of the Group I fusion transcript is observed in the liver, head and the rest of the body (Fig. 2C), whereas the Group II fusion transcript was expressed exclusively in the foetal liver (Fig. 2D), thus correlating with the reporter activity. The liver-speci®c expression of the Group II transcript was con®rmed by RT-PCR (Fig. 2E). The expected size product was generated using I114 ES cell and I114 heterozygous embryonic liver RNA but no product
was detected in RNA isolated from the head and only a faint band was observed in the rest of the body. It is most likely that this represents liver tissue contaminating the rest of the body tissue during dissection. Characterization of the endogenous cDNA (see below) and predicted translation of the fusion transcripts revealed that the Group I fusion places the lacZ sequence out of frame, whereas the Group II fusion sequence is untranslated and would presumably generate b-gal protein from the start codon of lacZ. This prediction is supported by Western blot analysis that showed that the b-gal protein was expressed exclusively in the foetal liver of I114 embryos (data not shown). The Group II fusion transcript is therefore most likely responsible for the liver-speci®c b-gal activity observed in embryos carrying the I114 gene trap integration.
208
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
Fig. 2. Sequence and expression of the Group I and Group II fusion transcripts. (A) Group I RACE clone sequence. The lower case sequence at the 3 0 end of each fusion transcript corresponds to the en-2 exon sequence from the PT1-ATG gene trap vector. (/) corresponds to the splice junction and the underlined poly G sequence corresponds to the arti®cial tail added during 5 0 RACE-PCR. (B) Group II RACE clone sequence. As above except that the underlined poly T sequence corresponds to the arti®cial tail added during 5 0 RACE-PCR. (C) RNase protection assay showing the Group I fusion transcript protected as a 199 bp fragment consisting of 79 bp of endogenous sequence and 120 bp of en-2 SA sequence. Expression is observed in I114 ES cells (lane 2), and in the liver (L), head (H) and the rest of the body (R) of I114 homozygous (2/2) and heterozygous (1/2) 13 dpc embryos (lanes 3±8), but not in the parental cell line R1 (lane 1) nor in tissues from wild type (1/1) 13 dpc embryos (lanes 9±11). Loading control GAPdH transcripts are protected as two fragments of 65 and 67 bp. (D) RNase protection assay showing the Group II fusion transcript protected as a 188 bp fragment consisting of 58 bp of endogenous sequence and 130 bp of en-2 SA sequence. Expression is observed in I114 ES cells (lane 1) and in the liver (L) but not the head (H) and the rest of the body (R) of I114 homozygous (2/2) 13 dpc embryos (lanes 2±4) nor in wild type 13 dpc embryonic tissues (lanes 5±7). (E) RT-PCR ampli®cation of Group II fusion transcripts using primers LST1 (complementary to the Group II sequence) and 78 (complementary to lacZ). The expected 377 bp product is observed at high levels in I114 ES cells (lane 1) and in the liver (L) of I114 homozygous (2/2) 13 dpc embryos (lane 2). No product was detected in the head (lane 3) and low levels were detected in the rest of the body (R) sample (lane 4), which likely re¯ects liver tissue contamination. I114 homozygous liver RNA subjected to the RT-PCR protocol in the absence of reverse transcriptase was used to control for genomic DNA contamination (lane 5). Ampli®cation of HPRT is used to control for the amount of total RNA used in the reactions.
2.4. cDNA cloning of gtar The Group I RACE clone was used to screen a random primed D3 ES cell cDNA library (gift from Dr Hitoshi Niwa). Seven clones were isolated and sequenced providing a contiguous sequence of 6040 nucleotides (Fig. 3). Conceptual translation of this sequence identi®es an open reading frame encoding a protein of 1599 amino acids with a predicted molecular weight of 170 kDa. The predicted protein contains 25 ANK-like repeat sequences in two separate domains of 15 and 10, three putative bipartite nuclear localization signals and a potential PEST sequence which includes a hyperacidic cluster (Rogers et al., 1986). We have named this gene gtar (Gene Trap Ankyrin Repeat). The Group I RACE clone sequence is within the coding region of the cDNA, between positions 825 and 903, and transla-
tion of the Group I fusion sequence from the gtar translational start codon con®rms that lacZ would be out of frame in this fusion transcript. We can predict that the gene trap vector is inserted into an intron between the splice donor (SD) site at position 903 and the putative downstream endogenous SA site at position 904. Two gtar splice isoforms were identi®ed that lacked the sequence between nucleotides 2843 and 3594. One of these isoforms also lacked the sequence between nucleotides 3722 and 4944 which contains the second 10 repeat ANK domain and a region of the ®rst bipartite NLS. Two colour ¯uorescent in situ hybridization FISH analysis was performed on I114 ES cell metaphase chromosome spreads using FITC-labelled PT1-ATG and pGC7, a 5.1 kb cDNA fragment of gtar (residues 1025±6139) labelled with Texas Red, as probes. The gene trap vector and gtar
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
209
Fig. 3. cDNA sequence and predicted translation of gtar. A stop codon (TAG, bold, underlined) at position 28 de®nes the translational start codon at nucleotide position 523 (bold). Bold underlined nucleotides at 520(23) and 526(14) are in agreement with the Kozak consensus sequence for translational initiation (RNNatgG; R, purine). The Group I RACE clone sequence is outlined from nucleotide position 825±903. The arrow between nucleotides 903 and 904 corresponds to the predicted vector insertion site and additional splice sites are identi®ed by arrows at nucleotides 2843 (SD), 3594 (SA), 3722 (SD) and 4944 (SA). Amino acid residues 113±124 are underlined denoting the hyperacidic cluster. The putative PEST sequence is ¯anked by positive amino acid residues at position 79 (R1 . ) and 148 (,1K). Ankyrin repeats are underlined and numbered (R1±R25) with highly conserved residues in bold. The three separate NLS signals are underlined with bold residues between amino acid residues 1468 and 1495.
210
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
resolved to a single site on chromosome 5 band E1 (data not shown) supporting that fact that we have indeed integrated into the region of the gtar locus. Interestingly, the gtar cDNA sequence shows similarity to the genomic sequence from human chromosome 4 (EMBL AC053525) in a region that is known to be syntenic to mouse chromosome 5. 2.5. Genomic characterization Three clones were isolated from the RPCI21 mouse PAC library using the Group II RACE clone L-69 as a probe. Sequence analysis (Fig. 4B) of a 1.5 kb PstI restriction fragment (439-a23.fP) (Fig. 4A) from one of these clones (439-a23) revealed that the Group I and Group II sequences were separated by only 500 base pairs and were in the same orientation with respect to the SD sequences used by the gene trap vector. Both the Group I and Group II SD sites
show a good overall match to the SD consensus sequence (Alberts et al., 1994) apart from a cytosine residue that replaces an almost invariant thymine at position 905 in the intron of the Group I SD site (Fig. 4B). The Group I and Group II sequences therefore represent alternative exons of the same gene that are differentially spliced to the gene trap vector and the two resulting fusion transcripts are differentially expressed. Interestingly, the transcriptional start site Wingender (TSSW) algorithm (UK HGMP Resource Centre) predicts a putative RNA polymerase II promoter region immediately upstream of the Group II sequence with the transcriptional start site at 1320 and a potential TATA box 28 nucleotides upstream (Fig. 4B). Position 1±903 of the 439-a23.fP genomic sequence which includes the Group I RACE sequence matches exactly the gtar cDNA sequence, indicating that this sequence represents a single exon. The promoter driving its expression is presumably upstream of the 439-a23.fP genomic fragment and distinct from the putative Group II promoter at position 1320. This identi®es a possible mode of differential expression between these two isoforms of the gtar gene. Repeated screening of cDNA libraries failed to identify a cDNA containing the Group II fusion sequence but transcripts containing the Group II fusion sequence can be predicted from its relative position in the genome (Fig. 5A). The existence of this predicted transcript and its foetal liver-speci®c expression was con®rmed by RT-PCR using two different primer pairs (Fig. 5A,B) and subsequent sequencing. Fig. 5C shows the sequence comparison of the two predicted isoforms of gtar. Translation of the Group II-containing transcript identi®es an ORF corresponding to the same reading frame as predicted for gtar. However, expression from the putative promoter of the Group II-containing transcript identi®es the ®rst start codon at nucleotide position 92 corresponding to the methionine codon at nucleotide 946 (amino acid 142) of the full-length (or Group I-containing) gtar sequence. This predicted protein will lack the ®rst 141 amino acid residues of the full-length GTAR protein which includes the putative PEST sequence and the hyperacidic cluster (Fig. 5C). 2.6. Promoter analysis
Fig. 4. Genomic analysis. (A) Schematic diagram showing the relative position of the Group I and Group II sequences in the genomic fragment, 439-a23.fp. (B) Sequence of the 1.5 kb genomic PstI fragment from PAC 439-a23. The outlined sequences from position 825±903 and 1323±1379 represent the Group I and Group II sequences, respectively. The sequence from position 1±903 is identical to 1±903 of gtar, identifying this as a single exon. The ATG highlighted at position 523 corresponds to the start codon of the endogenous gene. The putative transcriptional start site (11) of the Group II promoter is at position 1320 (bold, underlined) with the TATA box (228) at position 1292 (bold, underlined).
To test the activity of the predicted Group II promoter sequence, the 1.5 kb genomic fragment (439-a23.fP) was subcloned upstream of the IRESb-geo cassette (Fig. 6A) (construct 42). Control constructs consisted of the IRESbgeo cassette alone (construct 40) and the 1.5 kb genomic fragment subcloned in the reverse orientation (construct 43). Stably transfected ES cell lines were generated by co-electroporation with linearized PGK-puromycin and resistant colonies were screened for the presence of the test plasmid by Southern blotting using a probe to the neomycin resistance gene. All ten ES cell clones that carried the putative promoter construct in the correct orientation expressed the
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
211
Fig. 5. Expression analysis and sequence of the endogenous Group II transcript. (A) RT-PCR strategy. Primers complementary to the Group II sequence (LST1) and to gtar (RTANK-2 and RTANK-3) were used to amplify the predicted Group II-containing endogenous transcript, producing products of 338 and 651 bp, respectively. (B) RT-PCR products were analyzed after 30 PCR cycles (lanes 1±4). The expected PCR products from the LST-1/RTANK-2 and the LST-1/ RTANK-3 reactions were ampli®ed from R1 ES cells and wild type 10.5 dpc liver (L) (lanes 1 and 2), but not from wild type 10.5 dpc head (H) and the rest of the body (R) (lanes 3 and 4). Ampli®cation of HPRT was used to control for the amount of RNA and to control for genomic DNA contamination, RT-PCR reactions were performed on RNA from R1 ES cells and wild type 10.5 dpc foetal liver with (lanes 5 and 7) and without (lanes 6 and 8) reverse transcriptase. Lane 9 shows that both primer sets amplify product from I114 homozygous (2/2) 10.5 dpc RNA, indicating that the gene trap vector has failed to disrupt the expression of this transcript. (C) Sequence comparison of the two predicted isoforms of gtar. The Group II sequence splices to the endogenous downstream sequence via the SD used by the gene trap vector (49) and the SA used by the Group I sequence in gtar (50). Conceptual translation of the nucleotide sequence identi®es the same ORF as gtar with the start codon at nucleotide position 92 corresponding to nucleotide 946 and methionine 142 of the full-length gtar. The nucleotides underlined in the Group II sequence correspond to the PCR primer LST-1 used to amplify the sequence. (^) indicates the splice site.
reported gene suggesting that this sequence had some promoter activity. In contrast, no reporter activity was observed in clones carrying the control constructs (Fig. 6A). One of the clones carrying the promoter construct was injected into blastocysts and chimaeric embryos were harvested at 10 dpc. LacZ reporter expression was observed throughout these chimaeric embryos (Fig. 6B). This suggests that the 1.5 kb genomic sequence has promoter activity but must require additional regulatory elements to direct embryonic liver-speci®c expression. 3. Discussion We have identi®ed and characterized an unusual gene trap event that produces two fusion transcripts from the same integration site, one of which is an early marker for liver development. Although multiple transcript-producing gene trap lines have been described previously (Chowdhury et al., 1997; Townley et al., 1997), the mechanism of multiple transcript production has not been investigated in detail. We have shown that the gene trap vector alternately splices to two exons of the gtar gene that are 500 base pairs apart in
the genome. One of the resultant fusion transcripts is expressed throughout the embryo but because the lacZ sequence is not in the correct reading frame no b-gal activity is produced. An alternative in-frame fusion transcript results in foetal liver-speci®c reporter expression and we have identi®ed a promoter sequence associated with this transcript. Such a gene trap event has not previously been reported and highlights the sensitivity of the technique for identifying alternate promoter use within genes. The I114 gene trap integration has not generated a null allele of the gtar gene and has therefore no phenotypic consequences. This has been reported previously for a number of gene trap integrations and is predicted to be a consequence of splicing around the gene trap vector (McClive et al., 1998; Sam et al., 1996; Voss et al., 1998). We predict that this phenomenon is more likely in genes like gtar that exist as a variety of alternatively spliced forms. Although we have shown that the liver-speci®c Group II sequence has an associated promoter activity, we failed to identify the sequences necessary for liver-speci®c expression. This is not surprising considering the complex nature of promoter and enhancer elements necessary for tissuespeci®c expression reported in other systems. For example,
212
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
alb expression in the liver is dependent on both the promoter sequence and an enhancer element that lies 10 kb upstream (Pinkert et al., 1987) and tissue-speci®c afp expression requires three separate enhancers and a complex promoter sequence (Chen et al., 1997). The widespread reporter expression associated with the promoter fragment that we tested could be a consequence of removing the promoter from the in¯uence of liver-speci®c enhancer or repressor elements. Identi®cation of the necessary control elements for liver-speci®c expression will provide an excellent tool to drive expression of genes exclusively to the liver. For example, it could be used to drive Cre recombinase expression allowing embryonic liver-speci®c targeting of genes, such as HNF-3b and GATA4, that play a role in hepatic development, that have lethal phenotypes prior to hepatic speci®cation (Ang and Rossant, 1994; Kuo et al., 1997; Molkentin et al., 1997). The GTAR protein contains ankyrin repeats that are involved in protein±protein interactions and have been identi®ed in a wide range of genes such as transcription factors, cytoskeletal proteins and transmembrane proteins (Bork,
1993). The liver-speci®c GTAR isoform is predicted to lack the N-terminal 141 amino acid residues of the predicted full-length GTAR protein. This region includes a stretch of glycine residues and a hyperacidic cluster which are located within a putative PEST sequence. The presence of a PEST sequence is associated with proteins with a short half-life (Rogers et al., 1986) and so it is possible that the liverspeci®c isoform is more stable than the full-length GTAR. Alternatively, the hyperacidic cluster may function to modulate the binding speci®city of the ANK repeats in a tissue-speci®c manner. A splice isoform of the ank1 gene that is lacking an acidic portion of the regulatory domain shows enhanced ANK-mediated binding to integral membrane proteins (Lux et al., 1990). The absence of the hyperacidic cluster from the liver-speci®c gtar isoform may alter the binding speci®city of GTAR in the foetal liver in a similar manner. This gene trap integration provides the earliest exclusive marker for liver development reported to date. The ®rst site of reporter expression corresponds to the location of the presumptive hepatic endoderm that has been identi®ed in
Fig. 6. Promoter analysis. (A) Analysis of ES cell colonies generated after co-electroporation of plasmids containing puromycin under the control of the PGK promoter and test promoter plasmids as indicated. Puromycin resistant colonies were screened for the presence of the test plasmid DNA by Southern blotting and subsequently tested for b-gal activity. (B) Chimaeric embryo generated from the injection of ES cells carrying test plasmid 42 showing ubiquitous b-gal activity.
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
previous studies (Houssaint, 1980; Gualdi et al., 1996). The advantage of this marker is that it can be observed in situ at the single cell level that it is more speci®c than previously used markers. Afp, for example, is also expressed in the visceral endoderm of the yolk sac and at low levels throughout the de®nitive endoderm at the earliest stages of hepatic development (Dziadek and Adamson, 1978; Dziadek and Andrews, 1983; Gualdi et al., 1996). Embryos carrying this marker will be useful in tissue explant cultures designed to identify the factors involved in hepatic speci®cation and in the phenotypic analysis of mouse mutants with defects in liver development. I114 reporter activity is maintained in the foetal liver throughout gestation but as development proceeds expression is increasingly restricted to a subset of cells and is only seen in sporadic cells of the adult. The fact that these cells exist as a small subpopulation of phenotypically indistinct cells and express a foetal marker gene would support the idea that they correspond to the putative hepatic stem cells or oval cells (Alison, 1998; Golding et al., 1995). ES cells carrying this gene trap integration could be used to establish protocols to generate liver stem cells from ES cells in vitro and the subsequent application of this technology to the human ES cells system could lead to the development of somatic cell therapy for liver disease.
213
sectioned. Age-matched wild type embryos and tissues were stained in parallel with all samples to control for background b-gal activity. 4.3. RACE-PCR Rapid ampli®cation of cDNA ends was carried out as described previously (Townley et al., 1997) using the following primers. First strand cDNA was synthesized from total RNA using the primer 5 0 -TAATGGGATAGGTTACG-3 0 (78) complementary to lacZ. The ®rst round PCR reaction used the anchor primer 5 0 -GGTTGTGAGCTCTTCTAGATGGTTTTTTTTTTTTTTTTT-3 0 (56) paired with the nested lacZ-speci®c primer 5 0 -AGTATCGGCCTCAGGAAGATCG-3 0 (79). The second round PCR reaction used the nested anchor primer 5 0 -GGTTGTGAGCTCTTCTAGATGG-3 0 (59) paired with the primer 5 0 -TGCTCTGTCAGGTACCTGTTG-3 0 (55) complementary to the en-2 exon 2 sequence used in the gene trap vector splice site. The PCR conditions were 30 cycles of 948C for 1.5 min, 608C for 1.5 min and 728C for 3 min. The second round PCR reactions were carried out with either Taq polymerase for directional cloning into pBluescript II KS 2 using SpeI and KpnI or with Pfu polymerase (GibcoBRL) for cloning into pCR w-Blunt using the Zero Blunte PCR cloning kit (Invitrogen).
4. Experimental procedures
4.4. RNase protection analysis
4.1. Animal breeding and phenotype analysis
Group I (L-A8) and Group II (L-69) RACE clones were used to generate antisense riboprobes for hybridization to I114 and wild type embryonic tissue RNA. Riboprobes were produced by in vitro transcription using 500 ng of linearized plasmid DNA in the presence of 1 mM (250 mCi) [a- 32P]cytidine triphosphate (CTP) using either T3, T7 or SP6 RNase polymerases. In vitro transcription products were separated on 6% polyacrylamide gels and the longest transcripts were puri®ed from the gel using a volume of elution buffer (0.5 M ammonium acetate, 1 mM EDTA, 0.2% SDS). For each hybridization reaction, 10 mg target RNA, 3.5 £ 10 5 cpm of eluted riboprobe and 5 £ 10 4 cpm of glyceraldehyde 3-phosphate dehydrogenase (GAPdH) loading control riboprobe were precipitated, resuspended in hybridization buffer (80% formamide, 400 mM NaCl, 1 mM EDTA, 40 mM piperazine-N,N 0 -bis(2-etanesulfonic acid) (PIPES) (pH 6.4)) and incubated overnight at 558C. RNase A and RNAse T1 were used to digest unprotected RNA. The samples were run alongside a ddT terminated sequence of 240 primed bacteriophage M13mp18 control DNA as a size marker on a 6% polyacrylamide gel which was then dried and exposed to Xomat (Eastman Kodak Co., Rochester, NY) autoradiographic ®lm to visualize protected fragments.
The I114 gene trap integration was transmitted through the germline by injection of the ES cells into C57/Bl6 blastocysts. Chimaeric males were backcrossed onto 129/CGR and C57BL/6 inbred and MF1 outbred females. Offspring were genotyped by Southern or dot blot analysis of tail DNA using a lacZ probe. Homozygotes and heterozygotes were distinguished by quantitation of band or dot intensity using a Phosphoimager using the level of actin hybridization as an internal control. 4.2. Reporter gene expression Embryos were staged by the presence of a vaginal plug and, in some cases, by the number of somite pairs. Whole embryos (up to 11.5 dpc) were stained with X-gal as described previously (Gossler et al., 1989) and subsequently sectioned by dehydrating for 2 h each in 70, 80, 90 and 100% ethanol followed by an overnight incubation in Histoclear at room temperature. Embryos were incubated twice in molten paraf®n at 568C for 2 h, followed by 2 h in molten paraf®n under vacuum, embedded in the left lateral orientation and sectioned at 5 mm intervals. Post-13.5 dpc embryos were cryostat sectioned prior to X-gal staining as described previously (McClive et al., 1998). Adult tissues were dissected from 3- to 5-month-old animals homozygous or heterozygous for the I114 gene trap integration and cryostat
4.5. Reverse transcriptase PCR (RT-PCR) RT-PCR was carried out on RNA from I114 ES cells,
214
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215
liver, head and the rest of the body from I114 homozygous 13 dpc embryos using the primers LST-1 (complementary to the Group II sequence) and LacZ-78 (complementary to lacZ), or R1 parental ES cell, liver, head and the rest the of body from wild type 10.5 dpc embryos using LST-1 in combination with two different primers, RTANK-2 and RTANK-3 (complementary to gtar). The reverse transcription reaction was carried out using 1 mg of total RNA (previously denatured at 708C for 5 min and then placed on ice) in 1£ PCR buffer (Promega) (2.5 mM MgCl2, 0.125 units of random hexamers (Boehringer Mannheim), 1 mM dNTPs and 200 units of MMLV reverse transcriptase) (GibcoBRL) in a total volume of 20 ml. Samples were incubated at 258C for 10 min and 428C for 1 h and then denatured at 958C for 10 min. Then, 4 ml of the ®rst strand cDNA reaction was added to 1.6 ml 10£ PCR buffer (1.6 ml 25 mM MgCl2, 100 ng primer x, 100 ng primer y, 0.2 ml (1 unit) Taq DNA polymerase (Promega) and water) to a ®nal volume of 20 ml. Samples were overlayed with mineral oil and PCR ampli®cation was carried out using 30 cycles of 968C for 5 s, 538C for 15 s and 728C for 1 min. RT-PCR products were run on an agarose gel, Southern blotted and probed with the 500 bp en-2 exon for LST-1/LacZ-78 and the RT-ANK-2 oligonucleotide for LST-1/RTANK-2 or RT-ANK3. The primer sequences used were 5 0 GGTAGTTTTCTGTCAGTGG-3 0 (LST-1), 5 0 -TAATGGGATAGGTTACG-3 0 (LacZ-78), 5 0 -CTGTGCTCTCAGCTCTCATC-3 0 (RTANK-2), 5 0 -GAAGACTGTGCATTGACATC-3 0 (RTANK-3), 5 0 -GCTGGTGAAAAGGACCTCT-3 0 (HPRT3 0 ) and 5 0 -CACAGGACTAGAACACCTGC-3 0 (HPRT5 0 ). HPRT was used to control for the even loading of RNA in the RT-PCR reactions (Johansson and Wiles, 1995). The endogenous Group II PCR products were subcloned using the TOPO TA cloning kit (Invitrogen). 4.6. PAC isolation and analysis The RPCI21 library (Osoegawa et al., 2000) was supplied by the UK HGMP Resource Centre (http:// www.hgmp.mrc.ac.uk) on gridded ®lters. The Group II clone, L-69, was hybridized to the ®lters in Church and Gilbert solution at 658C (Church and Gilbert, 1984) and washed in 1£ SSC, 0.1% SDS at 658C and three positive clones, 352-j9, 390-m1 and 439-a23, were identi®ed. DNA was prepared using the Qiagen Maxiprep Kit according to the manufacturer's instructions with the following modi®cations. Two hundred ®fty millilitres of overnight culture (LB 1 25 mg/ml Kanamycin) was used and 20 ml (twice the recommended amount) of each of the buffers P1, P2, P3 was added. For the elution of the plasmid from the Qiagen column, the elution buffer (QF) was heated to 658C and added in 2 ml aliquots. The 1.5 kb PstI fragment (439a23.fP) that was positive for the Group I and Group II sequences was subcloned into pZERO-2 (Invitrogen) and sequenced.
4.7. Promoter constructs The gene trap vector pGT1.8IRESbgeo (Mountford and Smith, 1995) was digested with SalI and BglII to remove the en-2 SA and re-ligated to generate construct 40. Alternatively, the genomic fragment 439-a23.fP was subcloned into the SalI/BglII site in both orientations to make constructs 42 and 43. Sca-1 linearized plasmid (40 mg) and SalI linearized pPGKpurobpA (4 mg) were electroporated into E14 ES cells (240 V, 500 mF). Cells were selected in puromycin at 2 mg/ ml for 8 days and colonies were picked into 24-well plates and expanded. Cells were X-gal stained and DNA extracted for Southern blotting as described previously (Forrester et al., 1996). Acknowledgements This work was funded from grants from the MRC (L.M.F. and A.J.W.), the BBSRC (Centre for Genome Research), the Geron Corporation, the Leukaemia Research Fund (LMF) and the Wellcome Tust (E.A.J. and D.W). References Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K., Watson, J.D., 1994. Molecular Biology of the Cell, 3rd edition, Garland, New York. Alison, M., 1998. Liver stem cells: a two compartment system. Curr. Opin. Cell Biol. 10, 710±715. Ang, S.L., Rossant, J., 1994. HNF-3 beta is essential for node and notochord formation in mouse development. Cell 78, 561±574. Bork, P., 1993. Hundreds of ankyrin-like repeats in functionally diverse proteins: mobile modules that cross phyla horizontally? Proteins 17, 363±374. Bossard, P., Zaret, K.S., 1998. GATA transciption factors as potentiators of gut endoderm differentiation. Development 125, 4909±4917. Cascio, S., Zaret, K.S., 1991. Hepatocyte differentiation initiates during endodermal-mesenchymal interactions prior to liver formation. Development 113, 217±225. Chen, H., Egan, J.O., Chiu, J.F., 1997. Regulation and activities of alphafetoprotein. Crit. Rev. Eukaryot. Gene Expr. 7, 11±41. Chowdhury, K., Bonaldo, P., Torres, M., Stoykova, A., Gruss, P., 1997. Evidence for the stochastic integration of gene trap vectors into the mouse germline. Nucleic Acids Res. 25, 1531±1536. Church, G.M., Gilbert, W., 1984. Genomic sequencing. Proc. Natl. Acad. Sci. USA 81, 1991±1995. Douarin, N.M., 1975. An experimental analysis of liver development. Med. Biol. 53, 427±455. Dziadek, M., Adamson, E., 1978. Localization and synthesis of alphafoetoprotein in post-implantation mouse embryos. J. Embryol. Exp. Morphol. 43, 289±313. Dziadek, M.A., Andrews, G.K., 1983. Tissue speci®city of alpha-fetoprotein messenger RNA expression during mouse embryogenesis. EMBO J. 2, 549±554. Forrester, L.M., Nagy, A., Sam, M., Watt, A., Stevenson, L., Bernstein, A., Joyner, A.L., Wurst, W., 1996. An induction gene trap screen in embryonic stem cells: identi®cation of genes that respond to retinoic acid in vitro. Proc. Natl. Acad. Sci. USA 93, 1677±1682. Golding, M., Sarraf, C.E., Lalani, E.N., Anilkumar, T.V., Edwards, R.J., Nagy, P., Thorgeirsson, S.S., Alison, M.R., 1995. Oval cell differentiation into hepatocytes in the acetylamino¯uorene-treated regenerating rat liver. Hepatology 22, 1243±1253.
A.J. Watt et al. / Mechanisms of Development 100 (2001) 205±215 Gossler, A., Joyner, A.L., Rossant, J., Skarnes, W.C., 1989. Mouse embryonic stem cells and reporter constructs to detect developmentally regulated genes. Science 244, 463±465. Gualdi, R., Bossard, P., Zheng, M., Hamada, Y., Coleman, J.R., Zaret, K.S., 1996. Hepatic speci®cation of the gut endoderm in vitro: cell signaling and transcriptional control. Genes Dev. 10, 1670±1682. Houssaint, E., 1980. Differentiation of the mouse hepatic primordium. I. An analysis of tissue interactions in hepatocyte differentiation. Cell Diff. 9, 269±279. Johansson, B.M., Wiles, M.V., 1995. Evidence for involvement of activin A and bone morphogenetic protein 4 in mammalian mesoderm and hematopoietic development. Mol. Cell. Biol. 15, 141±151. Jung, J., Zheng, M., Goldfarb, M., Zaret, K.S., 1999. Initiation of mammalian liver development from endoderm by ®broblast growth factors. Science 284, 1998±2003. Kuo, C.T., Morrisey, E.E., Anandappa, R., Sigrist, K., Lu, M.M., Parmacek, M.S., Soudais, C., Leiden, J.M., 1997. GATA4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev. 11, 1048±1060. Lux, S.E., John, K.M., Bennett, V., 1990. Analysis of cDNA for human erythrocyte ankyrin indicates a repeated structure with homology to tissue-differentiation and cell-cycle control proteins. Nature 344, 36± 42. McClive, P., Pall, G., Newton, K., Lee, M., Mullins, J., Forrester, L., 1998. Gene trap integrations expressed in the developing heart: insertion site affects splicing of the PT1-ATG vector. Dev. Dyn. 212, 267±276. Molkentin, J.D., Lin, Q., Duncan, S.A., Olson, E.N., 1997. Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev. 11, 1061±1072. Mountford, P.S., Smith, A.G., 1995. Internal ribosome entry sites and dicistronic RNAs in mammalian transgenesis. Trends Genet. 11, 179± 184. Osoegawa, K., Tateno, M., Woon, P.Y., Frengen, E., Mammoser, A.G., Catanese, J.J., Hayashizaki, Y., de Jong, P.J., 2000. Bacterial arti®cial chromosome libraries for mouse sequencing and functional analysis. Genome Res. 10, 116±128. Pinkert, C.A., Ornitz, D.M., Brinster, R.L., Palmiter, R.D., 1987. An albumin enhancer located 10 kb upstream functions along with its promoter
215
to direct ef®cient, liver-speci®c expression in transgenic mice. Genes Dev. 1, 268±276. Rogers, S., Wells, R., Rechsteiner, M., 1986. Amino acid sequences common to rapidly degraded proteins: the PEST hypothesis. Science 234, 364±368. Sam, M., Wurst, W., Forrester, L., Vauti, F., Heng, H., Bernstein, A., 1996. A novel family of repeat sequences in the mouse genome responsive to retinoic acid. Mamm. Genome 7, 741±748. Schmid, P., Schulz, W.A., 1990. Coexpression of the c-myc protooncogene with alpha-fetoprotein and albumin in fetal mouse liver. Differentiation 45, 96±102. Severn, C.B., 1972. A morphological study of the development of the human liver. II. Establishment of liver parenchyma, extrahepatic ducts and associated venous channels. Am. J. Anat. 133, 85±107. Shiojiri, N., 1981. Enzymo- and immunocytochemical analyses of the differentiation of liver cells in the prenatal mouse. J. Embryol. Exp. Morphol. 62, 139±152. Shiojiri, N., Lemire, J.M., Fausto, N., 1991. Cell lineages and oval cell progenitors in rat liver development. Cancer Res. 51, 2611±2620. Skarnes, W.C., 1990. Entrapment vectors: a new tool for mammalian genetics. Biotechnology 8, 827±831. Skarnes, W.C., Auerbach, B.A., Joyner, A.L., 1992. A gene trap approach in mouse embryonic stem cells: the lacZ reported is activated by splicing, re¯ects endogenous gene expression, and is mutagenic in mice. Genes Dev. 6, 903±918. Townley, D.J., Avery, B.J., Rosen, B., Skarnes, W.C., 1997. Rapid sequence analysis of gene trap integrations to generate a resource of insertional mutations in mice. Genome Res. 7, 293±298. Voss, A.K., Thomas, T., Gruss, P., 1998. Compensation for a gene trap mutation in the murine microtubule-associated protein 4 locus by alternative polyadenylation and alternative splicing. Dev. Dyn. 212, 258± 266. Zaret, K.S., 1996. Molecular genetics of early liver development. Annu. Rev. Physiol. 58, 231±251. Zaret, K., 1998. Early liver differentiation: genetic potentiation and multilevel growth control. Curr. Opin. Genet. Dev. 8, 526±531. Zaret, K.S., 2000. Liver speci®cation and early morphogenesis. Mech. Dev. 92, 83±88.