Statistical analyses were performed using Intercooled Stata (Stata, College Station, TX, USA). The efficiency of establishing reporter clones was analyzed with Fisher's exact test. Two-way factorial analysis of variance was applied for the comparison of plasmid-based versus AAV-based targeting vectors and ATG-less TV versus SV40p TV after log transformation of the data. A paired t-test was used to compare the short and long 5′ arms.Results and discussion
We initially constructed a reporter vector, the HygR-5′ EGFP reporter, carrying a fusion gene consisting of full-length HygR (5′ side) and a 5′ portion of EGFP (500 bp in length; 3′ side) fused to each other in a single frame (Figure 1). We also constructed a targeting vector, ATG-less TV (Figure 1), containing an EGFP gene without an authentic start codon to reconstitute functional EGFP via HR with the HygR-5′ EGFP fusion gene in the reporter vector. ATG-less TV uses a promoter-trap strategy for efficient gene targeting (18, 19) and thus has no promoter to drive NeoR expression. In addition, ATG-less TV is constructed with an AAV backbone, because AAV-based targeting vectors achieve more than 1000-fold higher rates of gene targeting than conventional plasmid-based targeting vectors (5).
When cells are stably transduced with the HygR-5′ EGFP reporter vector and the ATG-less TV, gene-targeting events are detectable by a green fluorescent signal. Because the H/R ratio of targeting vectors is one of the most critical determinants of the difficulty of gene targeting, we addressed the ability of this system to assay H/R ratios. We first transfected the HygR-5′ EGFP reporter vector into the HCT116 and DLD-1 colon cancer cell lines, and isolated single cell clones through hygromycin selection (hereafter referred to as reporter clones). The resulting six reporter clones from HCT116 and six from DLD-1 were then infected with ATG-less TV and selected with G418. Fluorescence flow cytometric analyses of the resulting bulk cell populations exhibited subsets of GFP-positive cells in one reporter clone from each cell line (Figure 2, A and B, and Table 1). These results strongly suggested that homologous regions within the two vectors were recombined in subsets of cells in these clones, leading to the production of functional GFP. We also confirmed the absence of appreciable GFP-positive cell populations within parental HCT116 and DLD-1 cells as well as in cells transduced with either the HygR-5′ EGFP reporter or ATG-less TV alone. Southern blot analysis of a functional reporter clone from each of these cell lines demonstrated that these reporter clones are likely to harbor single copies of the HygR-5′ EGFP reporter vector in the genome (Figure 2C).
In the other five reporter clones for each cell line, no GFP-positive cell populations were detected by flow cytometry after infection with ATG-less TV. Although the reason for this is unclear, it may be possible that the location of the HygR-5′ EGFP reporter vector within the genome affects the efficiency of HR with ATG-less TV. Another possibility is that all or part of the 3′ arm of the HygR-5′ EGFP reporter vector may be subjected to terminal deletion upon integration into the genome, which precludes gene targeting by ATG-less TV.
We next carried out cell sorting of an HCT116-derived reporter clone transduced with ATG-less TV and isolated GFP-positive and -negative populations. A total of 15 single cell clones were established from the respective populations and analyzed by PCR to determine whether GFP-positive cells had truly undergone gene targeting by ATG-less TV. PCR over the 5′ arm of ATG-less TV (F1–R1; Figure 3, A and B) indicated that gene targeting was achieved in at least 14 of the 15 single cell clones isolated from the GFP-positive cell population, whereas none of the clones from the GFP-negative population underwent gene targeting. We also confirmed the reconstitution of a full-length HygR-EGFP gene in a few GFP-positive single cell clones by DNA sequencing (data not shown). In addition, we carried out Southern blot analyses of several GFP-positive and -negative single cell clones derived from the HCT116 and DLD-1 cell lines, and further confirmed that ATG-less TV was homologously recombined with the HygR-5′ EGFP reporter vector within the genome of GFP-positive cells (Figure 3C). These results collectively indicate that GFP positivity in the HygR-5′ EGFP reporter system represents gene targeting events.