Nonintegrating foamy virus vectors

Abstract

Foamy viruses (FVs), or spumaviruses, are integrating retroviruses that have been developed as vectors. Here we generated nonintegrating foamy virus (NIFV) vectors by introducing point mutations into the highly conserved DD35E catalytic core motif of the foamy virus integrase sequence. NIFV vectors produced high-titer stocks, transduced dividing cells, and did not integrate. Cells infected with NIFV vectors contained episomal vector genomes that consisted of linear, 1-long-terminal-repeat (1-LTR), and 2-LTR circular DNAs. These episomes expressed transgenes, were stable, and became progressively diluted in the dividing cell population. 1-LTR circles but not 2-LTR circles were found in all vector stocks prior to infection. Residual integration of NIFV vectors occurred at a frequency 4 logs lower than that of integrase-proficient FV vectors. Cre recombinase expressed from a NIFV vector mediated excision of both an integrated, floxed FV vector and a gene-targeted neo expression cassette, demonstrating the utility of these episomal vectors. The broad host range and large packaging capacity of NIFV vectors should make them useful for a variety of applications requiring transient gene expression.

FIG. 1.

Schematic representation of mutations introduced into integrase, the pol helper plasmid, and the NIFV vector plasmid. (A) Domain structure of the integrase protein. The highly conserved DDE residues are shown in the catalytic core. (B) Mutations producing the indicated amino acid changes were introduced into the integrase sequence (IN) of the pol helper plasmid, pCiPS, and the pol fragment of FV vector plasmids to generate the NIFV pol helper and vector plasmids shown. Locations of the CMV promoter, simian virus 40 (SV40) intron (intron), SV40 polyadenylation site (pA), internal promoter (Pro), transgenes (Gene), LTRs, and pol open reading frame (pol) are shown.

FIG. 2.

Southern blot analysis of NIFV vector stocks. Purified DNAs from vector stocks of ΔφMscvF (Δφ), NIFV-MscvF (NIFV), D100V-MscvF (100), D157A-MscvF (157), and a pseudostock control (Ps) were cut with PstI and AgeI and analyzed by Southern blotting. DNAs were also digested with DpnI to remove transfected plasmid DNA. The positions of the linear DNA genome (linear), replication intermediates (R1 and R2), and 1-LTR circles (1-LTR) are shown. Diagrams of the predicted sizes (not to scale) of LTR-containing fragments are below the Southern blot. The position of the probe is denoted by the black bar.

FIG. 3.

GFP expression profile from cells infected with NIFV vectors. Human fibroblasts were infected with ΔφMscvF, NIFV-MscvF, D100V-MscvF, or D157A-MscvF or with a pseudostock control (stock prepared without pol expression plasmid) at an MOI of 50 genome-containing particles per cell. The percentage of GFP⁺ cells (A) or the MFI (B) was monitored over time by flow cytometry. (C) Cell cycle analysis of the DNA content in dividing and arrested cell populations by propidium iodide staining. (D) Dividing and arrested human fibroblast cultures were infected with ΔφMscvF and NIFV-MscvF at an MOI of 15 genome-containing particles per cell. The percentage of GFP⁺ cells in the arrested culture divided by that in dividing cultures (A/D ratio) is shown for both infections, with solid and open columns representing separate experiments.

FIG. 4.

Analysis of episomal vector genomes. Human fibroblasts were transduced with ΔφMscvF (Δφ) and NIFV-MscvF (NIFV) at a multiplicity of infection of 1,000 genome-containing particles per cell or a pseudostock control (Ps). (A) Low-molecular-weight DNA was isolated on days 2, 5, and 8 after infection, cut with DpnI, PstI, and AgeI, and then analyzed by Southern blotting. DNA purified from each vector stock was also included (S). See Fig. 2 for diagrams of the predicted sizes. (B) The number of vector genomes present per cell over time. (C and D) The total number of vector genomes and those of individual species of linear, 1-LTR circular, and 2-LTR circular genome forms are shown for each time point after infection with NIFV (C) or Δφ (D).

FIG. 5.

Sequence of vector-chromosome junctions from NIFV vector integrants. Two integrated proviruses were rescued as bacterial plasmids, and their junctions were sequenced. In each case, the junction-containing portions of the provirus are expanded below the vector map, and the predicted vector genome sequence is aligned above the chromosomal sequences obtained, with the junction points indicated by bold black lines. Sequences present in each provirus are in uppercase, and extensions of the vector sequences not seen in the rescued DNA are in lowercase. Insertions are denoted with italics, and duplications of the chromosomal DNA are underlined. Bold font indicates regions of microhomology between the vector and the chromosome, and the terminal CA dinucleotides are boxed. Specific nucleotide positions of the chromosomal sequences are indicated.

FIG. 6.

Cre recombinase is expressed from a NIFV vector. (A) Schematic representation of an integrated ΔφMscvTKNloxP provirus containing a thymidine kinase-neo fusion gene (TKN) driven by an internal MSCV promoter and the same genomic locus after Cre-mediated excision. The positions of loxP sites and AvrII restriction sites are indicated. (B) Illustration of the NIFV-MscvCre vector with its MSCV promoter and mammal-optimized Cre transgene (Cre). (C) Graph showing the percentage of ganciclovir-resistant (GCV^R) colonies obtained after transduction of ΔφMscvTKNloxP-containing clones with the Cre protein (protein) or NIFV-MscvCre (NIFV). (D) Southern blot analysis of DNA isolated from clones 1 and 2 after no treatment with Cre (None), treatment with Cre protein (Pro), or infection with NIFV-MscvCre (NIFV). The DNAs were digested with AvrII and probed for LTR sequences. (E) Schematic representation of the COL1A2 locus showing the untargeted allele, the targeted allele with insertion of the IRES-neo cassette, and the targeted allele after excision of IRES-neo by Cre recombinase. (F) Southern blot of NdeI/SalI-digested genomic DNA isolated from the parental, OIMSC12-targeted clone (No Vector), cells infected with NIFV-MscvCre at MOIs of 5, 50, 500, and 1,000 genome-containing particles per cell, cells infected with ΔφMscvCre at MOIs of 5, 50, and 500 genome-containing particles per cell, or the highest volume equivalent of Cre pseudostock (Ps), as indicated. Expected fragment sizes are 5.8 kb (untargeted allele), 6.6 kb (targeted allele), and 5.2 kb (cre-excised locus) after probing for COL1A2 sequences as indicated in panel E.