Restriction mapping of retroviral vector episomal DNA.

The infidelity of reverse transcription enzyme activity during the replication cycle of a retrovirus can cause retroviruses to rapidly mutate. For gene therapy, this raises biosafety issues, such as the deletion of a therapeutic gene from a vector and the development of a replication-competent retrovirus. In a single retroviral replication cycle, a therapeutic herpes simplex virus thymidine kinase (HSVtk) gene was inactivated in approximately 8% of Moloney murine leukemia virus (MoMLV)based vectors. In the same vector system, the mutation rates within a single retroviral vector were calculated as high as 3% per kb (9,14). In addition to the deletion of HSVtk from the retroviral vector, deletion mutations have been observed in retroviral vectors carrying various genes, including nerve growth factor receptor (7), α-rev sequence (2), human glucocerebrosidase (16) and luciferase (13). It is often difficult to distinguish a mutant vector from correct vector sequences in the genomic DNA of vector producer cells (VPC) because both vectors share large homologous DNA sequences except at the deletion or mutation site. To define the mutated region requires multiple Southern blots and analysis by restriction endonuclease mapping. This analysis can be complicated by interference with endogenous retroviral element sequences in the mammalian genome because these sequences are highly homologous to vector sequences. High frequencies of superinfection and retrotransposition (manuscripts in preparation) of retroviral vector in cultured VPC results in detectable amounts of episomal DNA. Episomal DNA is advantageous for the Southern blot analysis of vectors because it is not subject to interference from endogenous retroviral sequences. Episomal vectors or retroviral sequences have been observed with other retroviruses, including mouse mammary tumor virus (11), avian sarcoma virus (15), avian leukosis virus (12), human immunodeficiency virus type 1 (HIV-1) (10) and in avian packaging cells (5). In this study, we successfully identified an HSVtk-deleted vector from VPC by analyzing episomal DNA directly instead of by genomic DNA restriction mapping. PCR primers were therefore designed accordingly to amplify this mutated region. The same primers were used to sequence this deletion without sequencing walking. A LTKOSN.2 VPC was previously established in our group for a phase I human gene therapy clinic trial (4). pLTKOSN plasmid DNA (Figure 1A) was first introduced into the ecotropic packaging cell line GP+E86 (6) by transient transfection. Supernates from these cells were then used to transduce the amphotropic retroviral packaging line PA317 (8), which was selected in G418 (1 mg/mL) for two weeks. Twenty different VPC clones were isolated from the original pool of cells. LTKOSN.2 VPC produces viral titers of approximately 1 × 106 cfu/mL (4). A deletion of the HSVtk gene in LTKOSN.2 VPC was first detected in viral RNA collected from pelleted viral particles in Northern blot analysis by using different probes (Figure 1B). This result indicated that the titer calculated from only G418 resistant colonies did not represent the titer of the fulllength LTKOSN vector, which implies that the evaluation of the vector titer needs to first ensure that all vectors contain an intact HSVtk suicide gene. Unintegrated, episomal copies of viral DNA were used for Southern blot analysis to analyze this mutation without the interference with endogenous retroviral elements present in the cellular genomes. Small amounts of episomal DNA derived from vector sequences have been routinely detected within VPC from PA317 and GP+E86derived VPC (unpublished data). Episomal DNA was extracted from both the cytoplasmic fraction and nuclear fraction of 1 × 107 LTKOSN.2 VPC. First, cells were trypsinized and subjected to 1% Triton X-100 detergent for 5 min at room temperature to lyse the cellular but not the nuclear membrane. Nuclei were separated from the cytoplasmic fraction by centrifugation at 9500× g for 5 min at 4°C (3). Cytoplasmic fractions were subjected to phenol/chloroform extraction and ethanol precipitation to isolate purified episomal DNA. The episomal DNA in nuclei was extracted using Hirt’s method (1) with 5 M NaCl to remove genomic DNA. The supernate containing episomal DNA was isolated from cell nuclei by centrifugation (13 000× g for 15 min) and then subjected to phenol/chloroform extraction and ethanol precipitation. Episomal DNA samples extracted from both the cytoplasmic and nuclear fractions were evaluated by Southern blot analysis. The results clearly show that episomal DNA was mainly detected in