Yeast recombinase FLP functions effectively in human cells for construction of adenovirus vectors.

We have recently developed a high-efficiency method of constructing adenovirus vectors based on Cre-mediated recombination between two plasmids co-transfected into 293 cells. The simplicity and efficiency of this method should greatly expedite the construction of most recombinant vectors. However, this system would not be suitable for constructing vectors bearing loxP sites elsewhere in the genome because of undesirable Cre-mediated vector rearrangements. To address this, we have developed a similar system using FLP-mediated site-specific recombination for the construction of adenovirus vectors.

An enhanced system for construction of adenoviral vectors by the two-plasmid rescue method.

The two-plasmid rescue method of constructing Ad vectors, which relies on either homologous or Cre-mediated recombination between two plasmids cotransfected into 293 or 293Cre4 cells, respectively, offers advantages over other approaches because of its simplicity. We have improved the efficiency of vector construction by both homologous and Cre-mediated recombination by replacing the single ITR in the shuttle plasmid with a head-to-head ITR junction. We have also expanded the versatility of this method by incorporating a Cre expression cassette into the plasmids to permit high-efficiency Cre-mediated vector rescue using 293 cells, abrogating the need for Cre-expressing cell lines. This new system retains the simplicity of the original but results in an approximately 100-fold increase in the number of recombinant viruses produced, all of which contain the foreign DNA insert, and allows high-efficiency Cre-mediated vector isolation using any E1-complementing cell line.

A high-efficiency Cre/loxP-based system for construction of adenoviral vectors.

Adenovirus (Ad) vectors provide a highly efficient means of mammalian gene transfer and are widely used for high-level protein expression in mammalian cells, as recombinant vaccines and for gene therapy. A commonly used method for constructing Ad vectors relies on in vivo homologous recombination between two Ad DNA-containing bacterial plasmids cotransfected into 293 cells. While the utility of this two-plasmid approach is well established, its efficiency is low owing to the inefficiency of homologous recombination. To address this, we have developed an improved method for Ad vector construction based on Cre-mediated site-specific recombination between two bacterial plasmids, each bearing a loxP site. Ad vectors are generated as a result of Cre-mediated site-specific recombination between the two plasmids after their cotransfection into 293 cells expressing Cre recombinase. The frequency of Ad vector rescue by Cre-mediated site-specific recombination is significantly higher (approximately 30-fold) than by in vivo homologous recombination. The efficiency and reliability of this method should greatly simplify and expedite the construction of recombinant Ad vectors for mammalian gene transfer. Ad vectors are commonly constructed by homologous recombination between two plasmids cotransfected into 293 cells. This method has numerous advantages but results in low numbers of plaques owing to inefficient recombination. We have developed an improved method based on Cre-mediated site-specific recombination, which results in vector rescue at frequencies approximately 30-fold higher than by homologous recombination. This method should greatly simplify and expedite the construction of recombinant Ad vectors for mammalian gene transfer.

Sequences in E1A proteins of human adenovirus 5 required for cell transformation, repression of a transcriptional enhancer, and induction of proliferating cell nuclear antigen.

A range of deletion and other mutants in the coding region of the E1A gene of Ad5 has been assayed for transformation of baby rat kidney (BRK) cells in cooperation with ras, repression of the SV40 enhancer, and induction of proliferating cell nuclear antigen (PCNA). Transformation efficiency was drastically reduced by deletion of residues 4-25, 36-60, or 111-138 in exon 1 of the 289 residue (289R) and 243R E1A proteins. Deletion of other residues in exon 1 had little effect. With mutants in the region unique to the 289R protein, and in exon 2, the only effect on transformation seemed to be an increased tendency of mutant transformants, compared to wt, to migrate to form secondary foci. Repression assays, performed with E1A plasmids producing only the 243R protein, showed that deletion of residues 4-25 or 36-60 inhibited repression completely. Deletion of residues 128-138 reduced repression, but deletions elsewhere in exon 1 had little effect. Deletion of residues 188-204 in exon 2 reduced repression slightly, and deletion of all of exon 2 reduced it to about one-half. It is concluded that for transformation, there are two functional domains in E1A proteins, both in exon 1, both involved in binding different cellular proteins, and both probably concerned with different transforming functions. One of these domains, involving residues 4-25 and 36-60, also functions in repression, but the role of the second in repression is much less critical. All of the deletion mutants in exon 1 induced PCNA synthesis in BRK cells. This result, together with previously published work, suggests that the active site for PCNA induction either involves residues 61-69 or 82-85 in exon 1, which have not been deleted, or it does not depend on any single limited region of the E1A proteins.

Use of deletion and point mutants spanning the coding region of the adenovirus 5 E1A gene to define a domain that is essential for transcriptional activation.

To help in identifying functional domains within Ad5 E1A proteins, we have constructed a series of mutants that create deletions throughout these products. We have also produced several mis-sense point mutations in the unique 13 S mRNA region. These mutated E1A regions have been tested in plasmid form for their ability to activate transcription of an E3-promoted CAT gene. From the results, a major domain for transactivation has been identified. This begins between residues 138 and 147, ends between residues 188 and 204, and encompasses the unique 13 S region. This domain is sensitive to mis-sense mutations. Transactivation was unaffected by small deletions in the N-terminal half of E1A proteins between residues 4 and 138, but was destroyed when this whole region was deleted. The C-terminal 71 residues may affect transactivation, but the results with the mutant in which this region was deleted were variable. The results obtained with these mutants are discussed in relation to the transactivation obtained by J. W. Lillie et al. [(1987). Cell 50, 1091-1100] with a synthetic peptide similar to the domain described here.

Peptide maps and N-terminal sequences of polypeptides from early region 1A of human adenovirus 5.

Experiments exploring the reasons for a multiplicity of products from early region 1A of adenovirus 5 are described. Labeled early region 1A products from wild-type virus were synthesized in infected cells and in a cell-free system programmed with mRNA from infected cells, immunoprecipitated specifically with an antipeptide serum, E1A-C1, directed against the C-terminal sequence of E1A products, and separated by gel electrophoresis. Two-dimensional maps of [35S]methionine-labeled peptides were consistent with antigens of 52,000 daltons (52K) and 48.5K being from the 13S mRNA and antigens of 50K, 45K, and 35K from the 12S mRNA. Partial N-terminal sequences of 52K, 50K, 48.5K, and 45K synthesized in vitro showed that each of these antigens was initiated at the predicted ATG at nucleotide 560 in the DNA sequence. These results eliminate multiple initiation sites and proteolytic cleavage at the N-terminal end as sources of antigen diversity. Peptide maps and N-terminal sequences were obtained in a similar way for E1A products from the Ad5 deletion mutant dl1504, which lacks the normal initiator codon. As predicted, these polypeptides are initiated at the next ATG, 15 codons downstream in the wild-type sequence. These results are discussed in relation to Kozak's ribosomal scanning model.