BMC Developmental Biology BioMed Central Research article Identification and targeted disruption of the mouse gene encoding

Background: Embryonic stem cell-specific gene (ESG) 1, which encodes a KH-domain containing protein, is specifically expressed in early embryos, germ cells, and embryonic stem (ES) cells. Previous studies identified genomic clones containing the mouse ESG1 gene and five pseudogenes. However, their chromosomal localizations or physiological functions have not been determined.


Background
Embryonic stem (ES) cells were first derived from the blastocysts of mice in 1981 [1,2] and humans in 1998 [3]. ES cells have two important properties: theability to main-tain pluripotency, which is the ability to differentiate into a wide variety of cells, and rapid proliferation. These characteristics make mouse ES cells an essential component of gene targeting technology. These qualitiesalso make human ES cells attractive sources for cell transplantation therapy to treat various diseases, including spinal cord injuries and juvenile diabetes. The molecular mechanisms underlying the pluripotency and rapid proliferation of ES cells are currently a major focus of the field of stem cell biology [4][5][6].
To identify molecules essential in ES cells for these properties, several groups have utilized transcriptome analyses to identify genes specifically expressed in ES cells and early embryos. These analyses, including DNA microarray analyses [7] and expressed sequence tag analyses [8][9][10][11][12], identified several common transcripts, including ESG1 that was also designated dppa5 or ECAT2.
ESG1 was originally identified as a transcript Ph34 that was down-regulated by retinoic acid in embryonic carci-noma cells [13]. The expression of this gene was confined in mice to early embryos and germ cells [14]. It is also expressed in pluripotent cells, including ES cells, embryonic germ cells, and multipotent germline stem cells [15]. ESG1 encodes a polypeptide of 118 amino acids that contains a single KH domain, which is an RNA-binding domain [16]. It remains unclear, however, if ESG1 functions as an RNA-binding protein or the roles it plays in ES cells and mice. Previous genomic library screening by identified genomic clones containing the mouse ESG1 gene and seven pseudogenes [17]. Two of these pseudogenes exhibit a similar exon-intron structure as the ESG1 gene, indicating their generation by gene duplication. The five remaining pseudogenes did not contain any introns, indicating that these were generated by retrotransposition of ESG1 transcripts.
ESG1 pseudogenes identified by a Blast search of mouse genomic databases  In this study, we determined the structure of the mouse gene encoding this protein and related pseudogenes. We also performed gene targeting to determine the physiological function of ESG1.

Chromosomal localization and structures of mouse ESG1 gene and psedogenes
To determine the chromosomal localizations of the mouse ESG1 gene and pseudogenes, we performed a Blast analysis of the mouse genomic database with the ESG1 cDNA sequence as a query. We identified several putative pseudogenes without introns on chromosomes 1,5,11,12,14,16,17, and X ( Figure 1). In addition, two intronless pseudogenes were identified in DNA fragments for which the chromosomal localization remained unmapped. While these pseudodgenes have significant homology to ESG1 cDNA, they could not produce functional proteins, because of critical mutations. This result suggests that there are a larger number of intron-less pseudogenes than previously anticipated. Existence of multiple retropseudogenes is a hallmark of pluripotent cellspecific genes [18].
On chromosome 9, we identified a DNA fragment similar to the ESG1 gene that included two putative introns. These putative first and second exons, however, contained (4) multiple mutations of the ESG1 cDNA sequence. The putative third exon was identical to that of the previously reported ESG1 gene. Also on chromosome 9, we identified another DNA fragment that was similar, but not identical, to the third exon of the ESG1 gene. These findings suggest that these ESG1-like sequences on chromosome 9 have not been correctly assembled.
To obtain DNA fragments containing the ESG1 gene, we screened the bacterial artificial chromosome (BAC) DNA pool by PCR using primer pairs that would only amplify the real gene, not the pseudogenes. We obtained two independent, but overlapping BAC clones. Southern blot analyses and sequencing demonstrated that these clones contained a sequence exhibiting complete identity with ESG1 cDNA that was interrupted by two putative introns ( Figure 2A). The two intron sequences begin with GT and terminate with AG, fulfilling the GT-AG rule of exonintron junctions [19]. The 5' flanking region of this DNA fragment exhibited strong promoter/enhancer activity by luciferase reporter assays in undifferentiated ES cells, but not in somatic cells (Figure 3). The same fragment showed much weaker activity after induction of differentiation by retinoic acid. We concluded that this sequence is the ESG1 gene.
We also found that the two BAC clones contained another ESG1-like sequence ( Figure 2A). The two sequences, separated by a 68 kbp intergenic sequence, were oriented in opposite directions. The ESG1-like sequence exhibited greater than 95% identity to the exons and introns of the ESG1 gene. This sequence, however, contained critical nucleotide substitutions in all of the exons and one nucleotide insertion in exon 2 ( Figure 2B). Although 675 base pairs of the 3' flanking regions were conserved between the ESG1 gene and the pseudogene, only five base pairs of the 5' flanking region were identical. This 5' flanking region (~6 kbp) did not possess any promoter/enhancer activity in luciferase reporter assays ( Figure 3). It is thus unlikely that this sequence is transcribed or translated into a functional protein. This sequence likely represents a duplication pseudogene. Bierbaum previously reported the existence of two pseudogenes with similar exon-intron organization as the ESG1 gene [17]. We could not determine which of these two pseudogenes corresponds to the one we identified or the location of the remaining pseudogene.

Targeted disruption of the mouse ESG1 gene
To study the function of ESG1, we deleted the gene by homologous recombination in mouse ES cells. We replaced the three exons with either a fusion of the neomycin-resistance and β-galactosidase genes (β-geo) or the hygromycin resistant gene (HygR) using two targeting vectors ( Figure 4A) introduced into RF8 ES cells by electroporation. We obtained eight ES cell clones with correct homologous recombination of the β-geo targeting vector, which was confirmed by Southern blot analysis ( Figure  4B). We obtained only one clone with correct homologous recombination of the HygR targeting vector.
To obtain homozygous mutant ES cells, we introduced the β-geo vector into HygR heterozygous ES cells. Of 105 G418-resistant colonies tested, 49 were homozygous for ESG1 deletion. Northern blot and western blot analyses confirmed the absence of ESG1 in these cells ( Figure 4C). In 29 clones, the β-geo vector had replaced the HygR vector, such that the cells remained heterozygous. In the remaining 27 clones, the β-geo vector was integrated at non-homologous sites.
ESG1 -/-ES cells exhibited normal morphology ( Figure  5A). These cells also proliferated at a speed comparable to that of the control (heterozygous and wild-type) cells ( Figure 5B). ESG1 -/cells differentiated normally after the removal of leukemia inhibitory factor ( Figure 5B) or retinoic acid treatment (not shown). When transplanted into hind flanks of nude mice, these cells produced teratomas, tumors containing components of all three germ layers ( Figure 5C). These results indicate that ESG1 is dispensa-ble for the self-renewal properties and pluripotency of ES cells.
We examined the gene expression profiles of ESG1 -/-ES cells using oligonucleotide-based DNA microarrays representing ~20,000 genes. In comparison to control ES cells, ESG1 was identified as the gene reduced to the greatest extent in ESG1 -/-ES cells ( Figure 6A). The expression of ES cell marker genes, such as Nanog and Oct3/4, was normal in ESG1 -/-ES cells. We confirmed normal Oct3/4 expression at protein levels by Western blot ( Figure 6B). The overall gene expression profiles were similar between control ES cells and ESG1 -/-ES cells. Several genes exhibited a greater than two-fold reduction in ESG1 -/cells, including Krt1-8, Pem, Ctgf, Ptgs2, Igf2 and Inhba. These genes might be regulated directly or indirectly by ESG1. Since ESG1 contains a KH-type RNA-binding domain, it may stabilize mRNA of these genes. Further studied are required to clarify this possibility.
We obtained germline transmission from three clones. We obtained ESG1 -/mice at the Mendelian ratios (36 wildtype, 69 ESG1 +/-, and 45 ESG1 -/-) from intercrosses of ESG1 +/mice. These animals exhibited normal development, gross appearance, and fertility (not shown). Histological examination of testis and ovary could not identify any abnormalities (not shown). These data demonstrated that ESG1 is dispensable for both mouse development and germ cell formation.
We also generated ES cells from blastocysts obtained by intercrosses of ESG1 +/males and ESG1 -/females. Of the eight ES cell lines established, two clones were ESG1 -/-. These ESG1-null ES cells demonstrated normal morphology, proliferation, and differentiation (not shown), confirming that ESG1 is dispensable in ES cells.

Conclusion
To analyze the physiological roles of ESG1, we identified the mouse gene on chromosome 9 and deleted it by homologous recombination in ES cells. Despite specific expression in early embryos, germ cells, and pluripotent
To sequence the region containing the mouse ESG1 gene and the 3' flanking region, we subcloned a ~15 kbp XhoI-SalI fragment into the pZERO-2 vector (Invitrogen). Hin-dIII-or EcoRI-digested fragments of this vector were then cloned into pBluescript KS(-) for sequencing. To sequence the ESG1 pseudogene and the 3' flanking region, an 8 kbp NotI/XhoI fragment was cloned into pBluescript KS(-). BamHI-or PstI-fragments of this vector were also cloned into pBluescript KS(-). To identify the sequence containing the 5' flanking regions of the ESG1 gene and the related pseudogenes, we used a TOPO walker kit (Invitrogen) with the pH34-T2L (5'-ACTAGTCGCAGCAG-GGATCCAGGAATATCT-3') and pH34-L394 primers. The resulting sequence was cloned into pCR2.1 (Invitrogen). We obtained a ~6 kbp band from the NsiI-digensted library; XbaI-, SpeI-, EcoRI-, and PstI-digested fragments of this band were cloned into pBluescript KS(-) for sequencing. This fragment contained the 5' flanking region of the ESG1 gene. A ~3 kbp fragment, obtained from the SacIdigested library, was cloned into pCR2.1 for sequencing. This fragment was contained the 5' region flanking the pseudogene.
Promoter/enhancer activity of the ESG1 gene and pseudogene

Southern blot screening for homologous recombination ES cells genomic DNA was extracted with PUREGENE™
Cell Lysis Solution (Gentra systems). For 5' Southern blot analysis, genomic DNA was first digested with PstI, then separated on an 0.8% agarose gel and transferred to a nylon membrane as described [20]. A 560 bp 5' probe was amplified using the ESG1S5 (5'-GATGGTGGTGGTGACT-CAGAG -3') and ESG1AS5 as (5'-CCTCCATTGCCTC-TATATCAG -3') primers. The probe specifically labeled an 18 kbp band from the wild-type locus, a 15 kbp band from the β-geo locus, and a 12 kbp band from the HygR locus. Genomic DNA was also digested with SpeI for 3' Southern blot analysis. A 1,010 bp 3' probe was amplified with the pH34U-8000 (5'-CCAACCAGCCAGAGTT-Targeted disruption of the mouse ESG1 gene Figure 4 Targeted disruption of the mouse ESG1 gene. A) Targeting strategy. Homologous regions are indicated by thick lines. Recognition sites of PstI (P) and SpeI (S), which were used for Southern blot analyses, are shown. The gene encoding diphtheria toxin A (DTA) was inserted at the 3' end of the targeting vectors to facilitate negative selection. B) Southern blot analyses confirming homologous recombination. WT, wild-type ES cells; β, β-geo +/-ES cells; H, HygR +/-ES cells; -/-, ESG1-null ES cells. Numbers indicate clone numbers. C) Northern blot (upper) and western blot (lower) analyses of wild-type ES cells (WT), ESG1-null ES cells (-/-, three clones) and heterozygous ES cells (+/-). Northern blot was performed as previously described [20]. To confirm the loading of equal amounts of RNA, ethidium bromide staining of ribosomal RNA is also shown (middle). TCAGTTAT -3') and pH34L-9000 (5'-GATAAGCTGCT-GCCAAAAGACAAG -3') primers. The probe hybridized to an 11.5 kbp band from the wild-type locus, a 12.5 kbp band from the β-geo locus, and a 9.5 kbp band from the HygR locus.

Derivation of ESG1-deficient ES cells from blastocysts
Esg1 +/or ESG1 -/mutant female mice were injected with Tamoxifen (10 µg) and Depo-provera (1 mg) subcutaneously on the third day of pregnancy. Four days later, embryos in diapause were flushed out of the uterus and cultured on STO feeder cells in four-well plates in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 20% Fetal Bovine Serum (Hyclone), 0.1 mM Non-Essential Amino Acids (Invitrogen), 2 mM L-glutamine (Invitrogen), 50 U/ml Penicillin-Streptomysin (Invitrogen), and 0.11 mM 2-mercaptoethanol (Invitrogen). After six days, the central mass of each explant was harvested, rinsed in PBS, and placed in a drop of trypsin for a few minutes. The cell mass was collected with a finely drawnout Pasteur pipette preloaded with medium, ensuring minimal carryover of the trypsin. The cells were gently transfered into a fresh well with 20% FBS-containing medium. The resulting primary ES cell colonies were individually passaged into wells of four-well plates containing Gene expression analyses of ESG1-null ES cells