High transduction efficiency and predictable copy number
Katherine Felts John C. Bauer Peter Vaillancourt
The high-titer retroviral vectors pFB and pFB-Neo can be used with any Moloney Murine Leukemia Virus (MMLV)-based packaging cell line. Virus produced with these vectors can be used to infect a wide range of cell types including human cells, and because MMLV-based virions are labile and readily inactivated by ethanol, ultraviolet light, and human complement, they are relatively safe and can be used in a P2 tissue culture facility. In transient virus production experiments, using a 293 cell-based producer line, we have achieved titers in excess of 10 8 colony-forming units per ml (cfu/ml) of supernatant for pFB, and greater than 107 cfu/ml for pFB-Neo. Protein expression from the viral promoter in infected cells was comparable to or higher than that for the vector pLXSN, another commercially available MMLV vector. The vector pFB has been deleted for all viral coding sequence and contains no exogenous sequence other than that for the multiple cloning site (MCS). As has been shown for other such minimal MMLV-based vectors, pFB is capable of insertions of up to 8 kb without appreciable loss of titer.1
The development of high-titer retroviral vectors that are capable of infecting a wide variety of cell types has had a tremendous impact on fields for which highly efficient gene delivery is essential.2 When tissue culture cells are infected with virus harboring a gene of interest, such experiments commonly result in transduction efficiencies of greater than 90%. An additional advantage of retroviral vectors is that the copy number of integrated provirus can be easily controlled by varying the multiplicity of infection, in contrast to standard transfection methods for which typically a small population of transfected cells are capable of the uptake and stable integration of vector DNA and for which the copy number is unpredicatable and often prohibitively high for many applications.3 The combination of high transduction efficiency and copy number control make retroviral delivery systems particularly useful for mammalian cDNA expression library production and screening. In these types of experiments, transduction of a large number of target cells with single-copy cDNA expression cassettes is highly desirable.4 The ability to efficiently transduce a wide range of hard-to-transfect cell types has also facilitated the reconstruction of transgenic animals. Additionally, retroviral vectors have become the system of choice for the in vivo delivery of therapeutic genes in the clinic.
We introduce the retroviral vectors pFB and pFB-Neo, which are based on MMLV and contain an extended packaging sequence and splice-site configuration that increases viral titer and gene expression from the viral promoter.5 These vectors can be used with any MMLV-based packaging cell line. MMLV-based virions are labile and readily inactivated by ethanol, ultraviolet light, and human complement.6 Thus, virus produced with these vectors, using packaging cell lines that allow infection of human cells (e.g., amphotropic packaging cells), are relatively safe and can be used in a P2 tissue culture facility.
We have achieved titers in excess of 108 cfu/ml with pFB and greater than 107 cfu/ml with pFB-Neo using an amphotropic packaging cell line; we found that expression levels for these vectors were generally comparable to or higher than those for other commercially available retroviral vectors.
Nonreplicating retroviral vectors contain all of the cis elements required for transcription of mRNA molecules encoding a gene of interest and packaging of these transcripts into infectious virus particles (Figure 1). The vectors typically comprise an E. coli plasmid backbone containing a pair of 600-bp viral long terminal repeats (LTRs) between which the gene of interest is inserted. The LTR is divided into three regions: The U3 region contains the retroviral promoter/enhancer, and is flanked in the 3 direction by the R region, which contains the viral polyadenylation signal (pA). The R region is followed by the U5 region, which, along with R, contains sequences that are critical for reverse transcription. Expression of the viral RNA is initiated within the U3 region of the 5 LTR and is terminated in the R region of the 3 LTR. The viral packaging signal (Y), which is required in cis for the viral RNA to be packaged into virion particles, resides between the 5 LTR and the coding sequence for the gene of interest.
To generate infectious virus particles that carry the gene of interest, specialized packaging cell lines have been constructed that contain chromosomally integrated expression cassettes for viral gag, pol, and env proteins, all of which are required in trans to make virus.2 The gag gene encodes internal structural proteins; the pol gene encodes reverse transcriptase (RT) and integrase; and the env gene encodes the viral envelope protein, which resides on the viral surface and facilitates infection of the target cell b y direct interaction with cell type-specific receptors. Thus, the host range of the virus is dictated not by the DNA vector but by the choice of the env gene used to construct the packaging cell.
The packaging cell line is then transfected with the vector DNA and, at this point, either stable viral producer cell lines may be selected (providing the vector has an appropriate selectable marker), or virus is produced transiently as mRNAs that are transiently transcribed from the vector are encapsidated, and bud off into the cell supernatant. These supernatants are collected and used to infect target cells. Upon infection of the target cell, the viral RNA molecule is reverse transcribed by RT (present in the virion particle), and the cDNA of the gene of interest, flanked by the LTRs, is integrated into the host DNA. Because the vector itself does not express viral proteins, once a target cell is infected, the LTR expression cassette is incapable of proceeding through another round of virus production.
The vectors pFB and pFB-Neo were designed to achieve the highest titers possible. They contain the bacterial origin of replication and the ampicillin resistance gene from pBR322 (Figure 2). It has been shown for the vector pMFG that a significant increase in viral titer is achieved by extending the 3 border of the Y site further into the gag coding sequence relative to other MMLV vectors.5 This extended packaging signal (Y+) has been incorporated into the pFB and pFB-Neo vectors. In addition, the MMLV splice acceptor (SA) sequence in pMFG has been inserted immediately downstream of the Y+. The SA splices to the natural splice donor (SD) located 58 bp downstream of the 5 LTR. Thus, expression from the 5 LTR in the vector (as well as in the ultimate integrated provirus) yields unspliced and spliced transcripts. The unspliced transcript contains the Y+ site and is packaged into virus particles in packaging cells. The spliced transcript has a relatively short, efficiently translated mRNA leader and allows high-level gene expression from the provirus. This splice-site configuration has also been incorporated into the pFB vectors.
The vectors pFB and pFB-Neo differ only by the presence of the IRES/neomycin-resistance gene cassette in the vector pFB-Neo. The open reading frame (ORF) for the neomycin-resistance (neo) gene is positioned downstream of the MCS, using the encephalomyocarditis virus (EMCV) internal ribosome entry site (IRES). Expression of neo from the second ORF in a dicistronic message is advantageous in that G418-selection (i.e., selection for neo-expression) ensures that the transcript containing the gene of interest is efficiently expressed. Furthermore, the potential loss of titer due to inclusion of an exogenous promoter between the LTRs6,7 is avoided by coexpressing both proteins from a single transcript.
In light of published reports showing that a significant increase in titer is achieved by minimizing the amount of extraneous sequence between the LTRs,3,7 we also constructed the vector pFB, which contains only the sequences required for virus production, in addition to an MCS. Because of limitations in insert size due to packaging constraints, a related advantage of pFB is that elimination of extraneous sequences allows the insertion of cDNA s of up to 8 kb without a significant loss of titer.1 Because the titers of these pMFG-based minimal vectors are sufficiently high, allowing greater than or equal to 90% of target cells to be infected, the need for antibiotic selection is essentially obviated.
Coding sequence for the enhanced green fluorescent protein (GFP)9 was inserted between the EcoR I and BamH I sites in pFB and pFB-Neo. The 293 cell-based amphotropic retroviral producer cell line HW293-A (unpublished data) was used for virus production. Producer cells were transiently transfected with both of the GFP constructs using the Transfection MBS mammalian transfection kit (Stratagene) with some modifications.4 Two days following transfection, viral supernatants were collected, and serial dilutions of each supernatant were used to infect dividing NIH3T3 cells. Two days later, the infected NIH3T3 cell populations were analyzed by fluorescence-activated cell sorting (FACS) (Figure 3). Titers were determined using the highest dilution for which cells containing fluorescence above background are clearly visible in the upper left (UL) quadrant of the plot (1:105 for both viruses). Titers are expressed as the product of the dilution factor multiplied by the percentage of infected cells that fluoresce above background, then multiplied by the number of cells infected. By this definition, pFB-Neo-GFP infected cells showed a titer of 1 x 107 cfu/ml of supernatant, and pFB-GFP gave a titer of 3 x 108 cfu/ml. This experiment was repeated with essentially identical results (data not shown). In addition, for the vector pFB-Neo-GFP, parallel infec ted samples were selected with 1 mg/ml of G418 for 10 days, then the colonies were stained with methylene blue and counted. The titer, as determined by G418 resistance, was identical to that determined by FACS (1 x 107 cfu/ml) (data not shown).
To assess expression levels from the pFB vectors, the GFP coding sequence was inserted into the MMLV vector pLXSN,10 and viral titers were determined for the derivative pLXSN-GFP and pFB-Neo-GFP. NIH3T3 cells were subsequently infected with an equivalent number of virus particles for both vectors, and expression levels were qualitatively measured by fluorescence intensity in the population as a whole by FACS (Figure 4, Panel A and Panel B) and in individual cells by fluorescence microscopy (Figure 4, Panel C and Panel D). As the results in Figure 4 indicate, the peak fluorescence intensity is several-fold higher for cells infected with pFB-Neo-GFP and is also significantly higher on a cell-to-cell basis for this vector.
Using the vectors pFB and pFB-Neo for high-titer virus production will, in many cases, obviate the need for scale-up production and tedious concentration steps when large numbers of virus particles are required (e.g., production of cDNA expression libraries). In addition, the relatively high expression levels achieved with these vectors will facilitate library screening and enhance general-purpose gene expression in infected cells. Because the vectors may be used with any MMLV-based packaging line, virus can be produced that infects virtually any cell type. The relatively small amount of sequence between the LTRs allows for expression of sequence of up to 8 kb without a significant loss of titer. The small size also facilitat es modifications that allow the construction of efficient, high-titer delivery vehicles for more sophisticated expression systems that allow inducible gene expression and functional knockouts.
Riviere, I. and Sadelain, M. (1997) In Methods in Molecular Medicine: Gene Therapy Protocols (Ed. Robbins, P.D.) Humana Press, Totawa.pp. 59-78.
Miller, A.D. (1997) In Retroviruses (Eds. Coffin, J.M., Hughes, S.H., andVarmus, H.E.) Cold Spring Harbor Laboratory Press, Cold Spring Harbor. pp. 437-473.
Onishi, M., et. al. (1996) Exp. Hematol. 24: 324-329.
Pear, W.S., et. al. (1997) In Methods in Molecular Medicine: Gene Therapy Protocols (Ed. Robbins, P.D.) Humana Press, Totawa. pp. 41-57.
Riviere, I., et. al. (1995) Proc. Natl. Acad. Sci. USA 92: 6733-6737.
Miller, A.D. (1992) Current. Topics Microbiol. and Immunol. 158: 1-24
Kitamura, T., et. al. (1995) Proc. Natl. Acad. Sci. USA 92: 9146-9150.
Miller, A.D., et. al. (1986) Somat. Cell. Mol. Genet. 12: 175-183.
Haas, J., et. al. (1996) Current. Biol. 6(3): 315-324.
Miller, A.D. and Rosman, G.J. (1989). BioTechniques 7(9): 980-990.