Loss of occupied gene expression in homozygous mutant cells and tissues. Pfdn1 expression (a, upper panel) in embryonic fibroblast cells from wild-type (lane 1), homozygous mutant (lane 2) and heterozygous (lane 3) fibroblasts was analyzed by northern blot analysis. The blot was stripped and probed with a GAPDH sequence as a loading control (lower panel). Cradd protein expression (b, upper panel) in primary spleen cells from homozygous mutant mice (lane 1) and wild type (lane 2) was assessed by western blot analysis. As a loading control, the blots were stripped and analyzed using an anti-β-actin antibody (lower panel). Dymeclin expression (c, upper panel) in liver tissue from wild type (lane 1) and homozygous mutant mice (lane 2) was analyzed by northern blot analysis. Hybridization to a GAPDH probe (lower panel) provides a loading control.

LOH at entrapment loci following selection in 2.0 mg/ml G418. DNAs from the parental ES cells (lane 1), heterozygous mutant entrapment clones (lane 2) and clones isolated following selection in 2.0 mg/ml G418 (lanes 3–7) were genotyped by either Southern blot hybridization (j, l) or PCR (a–i, k) using gene-specific probes and primers. The mutant clones contained entrapment vectors inserted in or near (*) the following genes: Ttc18 (a) Rbm4 (b), IL8Ra* (c), M11t10 (d), Cradd (e), 1700012G11Rik (f), Rnf7 (g), LOC240669 (h), Hesx1 (i), 1810030N24Rik (j), Dscr1l1 (k) and Xrcc5 (l). Bands corresponding to the wild-type allele migrate more slowly in all cases except (b) than the allele disrupted by gene entrapment. LOH is indicated by the absence of the wild-type allele in cells selected in high G418 media.

The frequency of presumptive LOH at sites throughout the genome increases with distance from the centromere. A total of 37 ES cell clones, each containing a single GTR1.3 gene trap vector, were placed in media containing 2.0 and 0.3 mg/ml G418. The frequency of colony formation (Presumptive LOH) in the higher concentration of G418 (normalized to the number of colonies at the lower concentration) is plotted against the distance of each mutation from the centromere. The average values for three independent experiments are plotted for all 37 entrapment loci (a) and for all 8 inserts located on chromosome 4 (b). Linear regression analyses of the two groups produced R-squared values of 0.54 and 0.78, respectively. The SD values, which were 1060% of the average values, have been omitted for clarity.

Loss of Xrcc5 expression in homozygous entrapment clones. (a) RNA was isolated from the parental AC1 ES cells (lane 1), the heterozygous Xrcc5 entrapment mutant before selection in 2.0 mg/ml G418 (lane 2) and clones isolated by selection in 2.0 G418 (lanes 3–6) and northern blot analysis was performed using Xrcc5 (downstream of exon 1 cloned by 3′RACE) and β-actin specific probes (upper and lower panels, respectively). Neo–Xrcc5 fusion transcripts in heterozygous and homozygous mutant cells are generated by splicing of the Neo sequences to exon 2 of the Xrcc5 gene. (b) Radiation sensitivity of Xrcc5 heterozygous and homozygous mutant cells. Parental ES cells (1), heterozygous Xrcc5 entrapment mutant (2) homozygous Xrcc5 entrapment mutants homozygous Xrcc5 entrapment mutants (3, 5, 6) a control Xrcc6-deficient Chinese hamster ovary (CHO) cell line (4) and a radiation-sensitive CHO cell line (Xrs5) defective in Ku86 were exposed to increasing doses of γ-irradiation, and cell survival was measured in a clonogenic assay.

Tagged sequence mutagenesis with the GTR gene trap vectors. Composition of 3′ RACE sequences (a) 838 vector–fusion transcripts cloned by 3′ RACE matched sequences in the mouse genome (MM), expressed sequence tag (EST) and RefSeq databases, as shown. Matches corresponding to cellular genes as determined by UniGene and Mouse Genome Informatics entries (MGI) are also indicated. The positions of GTR inserts in cellular genes as deduced from the sequence of 3′ RACE products. (b) The majority (217) of inserts in well-characterized genes (304) spliced to the last exon of cellular genes; of these, 168 contained multiple exons and 49 contained two exons. (c) Disruption of novel gene sequences by GTR gene trap vectors. The number of inserts into unique genes (based on UniGene designations) is plotted against the total number of ES cell clones analyzed. All matches plotted on the y-axis involved non-repetitive sequences with P-values < 10⁻²⁰. ES cell clones plotted on the x-axis included sister clones from the same dish and clones from which 3′ RACE products contained repetitive (non-informative) DNA sequences.

(a) Structures of the GTR retrovirus gene trap vectors. Expression of an intron-containing Neo gene (5′ Neo + 3′ Neo) carried by the GTR1.0 poly(A) trap vector selects for inserts in which the Neo gene, expressed from the promoter of the RNA polymerase 2 gene (Pol2), splices to downstream exons of cellular genes. Transcripts of occupied cellular genes splice to a 3′ exon [consisting of the 3′ end of a puromycin-resistance gene (3′ Puro), an internal ribosome entry site (IRES), a lacZ reporter and a polyadenylation site (PA)], disrupting their expression. A wild-type loxP site (loxP, left of 3′ Puro) and mutant loxP sites (lox 5171, on either side of the 3′ Neo exon) allow the body of the provirus to be replaced by other sequences by Cre-mediated cassette exchange. An RNA instability sequence (MI, flash symbol) increases the specificity of gene entrapment by reducing the levels of unspliced Neo transcripts. The positions of the provirus long terminal repeats (3′ and 5′ LTRs) are also indicated. Viruses lacking either the message instability sequence (pGTR1.3) or the lox 5171 in the Neo intron (GTR1.2) or both elements (GTR1.1) have also been constructed. GTR1.4–1.7 are identical to GTR1.0–1.3 except they contain an enhanced green fluorescence (EGFP) reporter instead of lacZ. GTR2.0–2.3 are identical to GTR1.0–1.3 except Neo is expressed from the PGK promoter. Some elements are not drawn to scale to enhance clarity. (b) Direct cloning of 3′ RACE products. Gene entrapment by GTR vectors generates clones in which the Neo gene (white boxes) is expressed from transcripts that splice to downstream exons of cellular genes (black boxes). The intron and NotI endonuclease cleavage site in the Neo coding sequence ensure that only recombinant plasmids that contain cDNA inserts amplified from spliced Neo-cell fusion transcripts can give rise to kanamycin-resistant E.coli.