¹Department of Molecular Biology and Biochemistry, University of California, Irvine, CA

²Department of Genetics, University of Georgia, Athens, GA, USA

^aPresent address: Department of Molecular Microbiology and Immunology, Johns Hopkins University, Baltimore, MD, USA

^bPresent address: School of Medicine, Tufts University, Boston, MA, USA

^cPresent address: Department of Zoology, Oregon State University, Corvallis, OR, USA

Using RT-PCR, we show that mouse adenovirus type I (MAV-1) is capable of infecting and expressing in various cell types, specifically human endothelial cells. The capability of MAV-1 to infect and express in human endothelial cells makes it a potentially useful alternative to the use of human adenoviruses type 2/5 (Ad2/5) in virus-based gene therapy, although presently MAV-1 can only be produced at lower titers than Ad2/5. In this report, we present methods for the purification of MAV-1 DNA and use of this DNA along with a modified bacteria-based homologous recombination protocol to generate a full-length plasmid clone of MAV-1 DNA. Using various transfection procedures, we show that this plasmid MAV-1 DNA can generate plaques of MAV-1 virus, albeit at low efficiencies (about 0.2 p.f.u./g DNA). Furthermore, the construction of an MAV-1 plasmid along with its capability to express in human cells justifies the full development of MAV-1 into a system of gene therapy.

mouse adenovirus; human; endothelial; full-length clone

Introduction

Human adenovirus type 2/5 (Ad2/5) have become very useful experimental gene therapy systems for the introduction and expression of heterologous genes into animal cells and tissues.¹ A significant advantage of these systems is their established capability to direct viral expression in various fully differentiated cell types. More recently, the adenovirus system has been shown to express the large dystrophin gene effectively in muscle cells.^2,3 The development of Ad2/5 into a system of gene therapy was significantly enhanced by the earlier development of cellular and molecular reagents that allow the convenient manipulation of recombinant adenovirus as a bacterial plasmid. These reagents include clones of the full-length adenoviral genome, such as pJM17⁴ and pTG3602,⁵ which could be used with fragments of altered viral genome to co-transfect E. coli to yield recombinant viral plasmid clones. Using human adenovirus as a gene therapy system in humans, however, poses some problems. For one, most human populations have been previously exposed to Ad2/5⁶ and it would be expected that the administration of a recombinant virus based on Ad2/5 would elicit a rapid secondary immune reaction.⁷ In addition, there remains the possibility that a defective recombinant virus might be capable of replication either because some human tissues can complement the missing viral genes (such as E1A) or because recombination might occur between the adenoviral sequence elements of the recombinant virus and silent persisting adenovirus in human tissues, resulting in a reactivated full blown infection with Ad2/5. It would therefore be beneficial if a recombinant adenoviral system were developed which had no capacity for replication in human cells. Finally, it is hypothesized that adenovirus does not infect endothelial cells efficiently due to lack of viral receptor(s). Given the interest to effect endothelial cell gene expression, especially with respect to vascular disease and cancer, manipulation of the Ad2/5 capsid has been studied. Several groups have attempted to modify the fiber protein of Ad2/5 in order to confer endothelium-specific gene expression. Modifications include targeting adenovirus to heparin-containing receptors,⁸ redirecting virus using folate conjugated antibody to fiber protein,⁹ as well as genetic modifications to the fiber protein.¹⁰ However, these systems are not highly efficient. It would therefore also be beneficial if an adenoviral system for gene therapy with a natural specificity for endothelial cells could be developed.

Mouse adenovirus type 1 (MAV-1), provides a potentially useful alternative to the human adenoviral systems. For one, in the natural biology of MAV-1, endothelial tissues in the heart,¹¹ kidney,¹² and brain^13,14 are prominent cellular targets of virus replication, as opposed to the respiratory specific Ad2/5.¹⁵ However, it had not been established that MAV-1 was in fact capable of uncoating and expressing in human endothelial cells. In addition, MAV-1 replication is specific to mouse cells and does not replicate in human cells, even those expressing human adenoviral E1A gene (JP Nery, unpublished observation). As MAV-1 shares low sequence homology with Ad2/5, it poses little theoretical risk of recombining with resident Ad2/5 that might be persisting in human tissues. In the event of an illegitimate recombination between the recombinant MAV-1 and resident Ad2/5, the resultant recombinant virus should be incapable of replication. Finally, as there is only limited immunological cross-reaction between human and mouse adenovirus,⁶ MAV-1 is unlikely to elicit a strong immune response when it is used as a system for human gene therapy with a protocol in which the number of repeat administrations is limited.

MAV-1, however, has not been developed for use as a recombinant system for gene therapy. More importantly, a full-length and biologically functional plasmid clone of MAV-1 DNA is needed for transfection experiments. Although MAV-1 DNA can be cloned as a yeast artificial chromosome (KR Spindler, unpublished data), similar to Ad2/5 DNA,¹⁶ these yeast clones yield too little MAV-1 DNA to provide biochemically useful quantities of purified DNA needed for transfection experiments. Previous attempts by us (M Del Vas and LP Villarreal, unpublished; KR Spindler, unpublished) using standard plasmid cloning methods, however, have failed to produce a full-length bacterial clone, although subclones containing a large portion of MAV-1 were successfully made. It seems that MAV-1 DNA at elevated copy number may be disfavored in bacterial cells. In addition, yields of viral MAV-1 DNA, extracted by the modified Hirt method^17,18 from infected mouse cells have been very low relative to that obtained from Ad2/5.¹⁹ Another problem appears to be that MAV-1-infected cells do not accumulate progeny virus intracellularly as do Ad2/5, but release most of the virus and viral DNA into the supernate.²⁰ The absence of an infectious plasmid clone and inefficient methods for purifying MAV-1 DNA have restricted the development of MAV-1 into a system for gene therapy. We isolated MAV-1 DNA with good yield and purity and used this DNA to construct a full-length infectious MAV-1 plasmid clone, applying a modified protocol of homologous recombination in E. coli.

Results and discussion

MAV-1 expression in human endothelial cells

Since there is widespread interest in using recombinant adenoviruses for diseases affecting human endothelial cells, it is important to show that MAV-1 is capable of infecting and expressing in these cells. RT-PCR²¹ studies were therefore performed to detect MAV-1 E1A mRNA transcription in various cells types. Total RNA was isolated 20 h after infection with MAV-1 (MOI = 100) from 3T6 (mouse embryonal cell line), 293T (human embryonal kidney cell line expressing Ad2/5 E1A and SV40 large T antigen), CV-1 (green monkey kidney cell line) and Phec 97-54 (primary human umbilical endothelial cells, provided by Dr Christopher Hughes, University of California, Irvine, CA, USA).^22,23 As shown in panel I of Figure 1, all cell types infected with MAV-1 (lanes b, e, h and k) show a major 210 bp band corresponding to spliced E1A RNA. In addition, a minor 286 bp band is also observed, which could be attributed to unspliced E1A RNA or contaminating MAV-1 DNA. In fact, the no-RT controls (lanes c, f, i and l) do show a 286 bp band, of variable intensity, which corresponds to the major product amplified from MAV-1 DNA. No signal was detected from uninfected controls. Panel II shows a 514 bp fragment corresponding to -actin RNA in samples treated with RT (lanes a, b, d, e, g, h, j and k) but no product in no-RT controls (lanes c, f, i and l). These results indicate that MAV-1 is capable of infection and expression of the E1A region with correct splicing in mouse embryonal cells, monkey kidney cells, human embryonal kidney cells and most importantly primary human endothelial cells. The 210 bp PCR product does show some variation in intensity among the cell types tested, but no definitive conclusion concerning E1A expression levels in different cell types can be drawn due to the non-linearity of the present RT-PCR assay. Based on these results, MAV-1 appears to have a natural capacity to enter and express in human endothelial cells. This justifies developing methods and molecular reagents needed to make MAV-1 into a system for gene therapy.

Purification of MAV-1 DNA from phenol interphase

An early method of Shinagawa²⁴ for the isolation of Ad2/5 DNA from infected cells used a phenol interphase to purify viral DNA. Because much of MAV-1 is released into the supernate, we evaluated the possibility of using a modified version of this procedure for the purification of MAV-1 DNA. This procedure takes advantage of the fact that adenoviruses are known to have terminal proteins covalently attached to the intracellular and extracellular DNA. When denatured in TE-saturated phenol, the terminal protein with covalently attached viral DNA is partitioned into the phenol phase and the interphase, while the bulk of cellular DNA partitions into the aqueous phase. The quality of MAV-1 DNA isolated from this interphase method (lanes b and g, Figure 2) was compared with that isolated using the modified Hirt method^17,18 (lanes a and f, Figure 2), typically used for adenoviral DNA isolation. Subsequent restriction endonuclease analysis identified the high molecular weight band observed in Figure 2 to be of MAV-1 origin (data not shown). The direct application of the Shinigawa method appeared to work well relative to the Hirt method and yielded about 3 g MAV-1 DNA per 100-mm plate. However, a substantial amount of degraded cellular DNA was visible as a heterogeneous smear in the preparation. It was also apparent that most of the MAV-1 DNA is in the extracellular supernate, as expected, and that this DNA had less degraded heterogeneous cellular DNA. The extraction of large supernate volumes with phenol, however, was most inconvenient. To decrease the aqueous volume, we first precipitated the MAV-1 virus by various methods. Ammonium sulfate precipitation gave DNA with comparable yields to unprecipitated samples and contained a lower number of heterogeneous DNA contaminants (lanes c and h, Figure 2). Polyethylene glycol precipitation and resuspension in TE gave higher yields of relatively pure MAV-1 DNA (4 g/100-mm plate) and also eliminated the heterogeneous DNA (lanes d and i, Figure 2). Thus PEG precipitation coupled with phenol interphase DNA extraction has allowed us to prepare greater quantities of MAV-1 DNA than previously.

Construction of plasmids by bacterial homologous recombination

Human adenoviral genomes could readily be cloned using a system of bacterial homologous recombination.⁵ Because this system uses a linear plasmid which must be closed by recombination with the inserted DNA to be biologically active, it positively selects for the desired recombinant. It seemed that this method might be useful for generating a full-length MAV-1 clone, circumventing the difficulties in cloning we encountered using standard methods. To perform this procedure, subclones containing target sequences (left and right ends) must first be made. As shown schematically in Figure 3, pMA was constructed using left and right end genomic clones pG1HAX and pG11SAB, which had unique ApaLI sites engineered immediately flanking the end of the MAV-1 DNA sequences (GE Carney and KR Spindler, unpublished data). Our initial attempts using the pBSX-based plasmids²⁵ yielded some clones that were unstable and tended to lose MAV-1 DNA over time. Since large plasmid clones can be toxic to bacterial host cells at high copy numbers,²⁶ we chose to use plasmid pWSK29²⁶ as vector for our construction because it is derived from the low copy number pSC101 replicon (about five copies per cell). It also contains the multiple cloning site of pBluescript (Stratagene, La Jolla, CA, USA). A drawback to manipulating a large plasmid is the lack of unique restriction sites that can be used to linearize the molecule at a desired location. Recently, using a method based on an observation that an oligonucleotide can invade a supercoiled DNA molecule at the complementary sequence, forming a D-loop structure,²⁷ we succeeded in linearizing supercoiled DNA at any desired location (Castro-Peralta and Villarreal, manuscript submitted). With these modifications, the bacterial in vivo homologous recombination system has become more generally useful for manipulating large recombinant adenoviral plasmids.

Characterization of MAV-1 clone, pKMA

Several clones that had the expected size for an insert of a full-length MAV-1 clone were selected for further analysis. A series of detailed restriction enzyme analyses were performed to establish that the cells had faithfully replicated the original MAV-1 sequence (Figure 4). As shown, all the restriction patterns were as predicted, establishing that a full-length clone had been successfully obtained. From this construct, the entire MAV-1 genome can be released by ApaLI restriction digest. A Southern analysis was performed to establish that the plasmid clone contained MAV-1 DNA (data not shown). Direct sequencing confirmed that no mutation had occurred during homologous recombination at the EcoRI and PstI junctions (Figure 3), and that the two ApaLI sites are precisely at the junctions of MAV-1 and vector plasmid DNA (data not shown). Taken together, these results indicate that pKMA contains a full-length copy of MAV-1 DNA.

Biological characterization of pKMA

If the pKMA clone is indeed a faithful copy of MAV-1 DNA, it is expected to be biologically active and capable of generating infectious MAV-1 virus following transfection. However, we have previously observed that MAV-1 DNA isolated from purified virions has a relatively low transfection specific activity; usually 1 p.f.u./g DNA or less (JP Nery, unpublished observation). As adenoviral DNA isolated from virions generally has higher transfection-specific activities than do corresponding plasmid clones of viral DNA (due to the residual terminal protein peptide), it was expected that the pKMA plasmid might be inefficient in transfection assays and that direct plaque formation might be problematic. In addition, if plasmid DNA must be released and primed by terminal protein following transfection,^28,29 this process could take enough time to delay the onset of possible virus induced cytopathology. In order to avoid these potential difficulties, we decided to do one blind passage of transfected supernate and examine the resulting supernate for its capability to induce cytopathology following an undiluted passage on to subconfluent 3T6 cells. 3T6 cells were therefore transfected with 1 g of pKMA using 8 l of lipofectamine (DNA released by ApaLI and gel purified was compared with the uncut pKMA). The DNA-lipofectamine complexes were incubated in Opti-MEM I for 45 min at room temperature and layered on the cells for 6 h at 37°C and 5% CO₂. The cells were then overlayed with 2% FBS HI DMEM plus 10^-5 m dexamethasone. Dexamethasone was added to enhance plaque visualization induced by MAV-1.³⁰ After 9 days of incubation, no CPE was apparent. This supernate was collected and used to do an undiluted infection with fresh cultures of 3T6 cells. After 11 days of incubation, these cultures showed CPE characteristic of MAV-1 infection. This supernate was collected and analyzed by plaque assay and was also used to do a high multiplicity infection of another 3T6 culture for DNA isolation. Plaque isolates were later shown to have MAV-1 by DNA gel analysis. The high multiplicity infected cultures were also confirmed to have MAV-1 DNA. Control transfections, no DNA and pKMA digested with DNase, did not yield any CPE with blind passage. These results (not shown) establish that MAV-1 virus was generated from the pKMA plasmid, albeit at what appears to be a low efficiency.

Direct plaqueing of pKMA DNA

The above result indicated that pKMA might be fully biologically active but at a lowered transfection efficiency. If this is correct, it might still be possible to obtain direct plaques from pKMA DNA if the transfection conditions allowed for a low transfection efficiency of the pKMA plasmid. Standard transfection conditions use about 1 g of DNA per tissue culture plate. If pKMA is about five-fold less efficient than virion isolated DNA (specific activity of 1 p.f.u./g), then we would not expect typically to observe a plaque resulting from a pKMA transfection. We therefore attempted transfections using very large quantities of pKMA DNA (37 g) on to 3T6 cells seeded in 60-mm plates. These plates were then overlayed with 1% bacto agar MEM (Pen/Strep, MgCl₂, glutamine, NEAA, 2% FBS HI) plus 10^-5 m dexamethasone and incubated for 10 days. Neutral red staining of these plates showed that plaques were produced (about 1 p.f.u./4 g plasmid DNA). Thus pKMA appears to be a fully active clone of MAV-1.

The generation of a biologically active clone of MAV-1 DNA should be very useful for the in vitro generation of MAV-1 mutants and for the development of MAV-1 into a new recombinant adenoviral system for gene therapy, despite the problem of producing MAV-1 at comparable titers to Ad2/5. The capability of MAV-1 to infect and express naturally in human endothelial cells allows for the possible use of this virus in treatment of these cells in humans, namely for heart disease and cancer. With the availability of a bacteria-based homologous recombination system, manipulation of the full-length mouse adenoviral genome has been significantly enhanced. The development of a new method for recovery of MAV-1 viral DNA from supernate will also significantly enhance the ease of manipulating MAV-1 DNA. Since greater amounts of viral DNA and a plasmid clone are presently available, it is now possible to consider mouse adenovirus as a potential candidate for use as a vector in virus-based gene delivery.

Materials and methods

Total RNA isolation

Total RNA was isolated following the methods of Chirgwin et al,³¹ with some modifications. Growth medium was aspirated off the cell monolayer. Cells from a T150 flask were lysed with 10 ml GIT solution (5.5 m guanidine thiocyanate, 25 mm Na-citrate, pH 5, 0.5% sarkosyl, 0.5% -mercaptoethanol, 0.15% antifoam A). The cell lysate was triturated by passing through a 18-gauge needle 10 times; clarified by centrifugation (2.5 ´ 10³ g, 10 min); layered on a CsCl solution (6 m CsCl, 25 mm Na-citrate, pH 5, 10 mm Na-EDTA); and centrifuged (90 ´ 10³ g, 24 h). The RNA pellet was dissolved in water, ethanol precipitated (2.5 m NH₄-acetate, 70% ethanol, 0°C, 1 h), dried in vacuo, re-dissolved in water, quantified spectrophotometrically and stored at -80°C.

RT-PCR methods

cDNA was synthesized using oligo(dT)₁₅ and Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega, Madison, WI, USA). Two g of total RNA were mixed with 45 pmol oligo(dT)₁₅, heated (80°C, 2 min) and quickly chilled in an ice-water bath for 1 min. The remaining RT reagents were added to a final volume of 30 l (1 ´ buffer, 0.5 mm dNTPs, 1 U RNasin, 400 U M-MLV RT) and incubated at 42°C for 1.5 h. One microliter of the RT reaction was used as template in a subsequent 25 l PCR (1 ´ buffer, 1 mm MgCl₂, 0.2 mm dNTPs, 1 m primers and 0.05 U/l Taq DNA polymerase). Primers MAV664 (5'-ACTATTGAGGTGTTCCCGCCA-3') and MAV949 (5'-TGACGGGAACATTCCGTCTTG-3') used for amplifying the cDNA hybridize to nucleotides 664-684 and 929-949 of the MAV-1 genome,³² respectively, flank one intron at nucleotides 753-828.³³ Primers specific for -actin cDNA (sense, 5'-TGTGATGGTGGGAA TGGGTCAG-3'; anti-sense, 5'-TTTGATGTCACGCACGATTTCC-3'; Stratagene, La Jolla, CA, USA) were used to amplify a 514 bp -actin fragment, as control for cDNA synthesis. Amplification of E1A products was carried out in 35 cycles as follows: First cycle, 94°C, 5 min/60°C, 5 min/72°C, 30 s; next 33 cycles, 94°C, 1 min/60°C, 30 s/72°C, 30 s; last cycle, 94°C, 1 min/60°C, 30 s/72°C, 10 min. Amplification of -actin products was carried out in 35 cycles as follows: First cycle, 94°C, 5 min/54°C, 5 min/72°C, 45 s; next 33 cycles, 94°C, 1 min/54°C, 30 s/72°C, 45 s; last cycle, 94°C, 1 min/54°C, 30 s/72°C, 10 min. Seven microliters of the PCR products were electrophoresed in a 2.5% agarose TAE (mmol per liter: 40 Tris, 20 acetate, 1 Na-EDTA; pH 8) gel.

MAV-1 DNA extraction

3T6 cells were infected at 50% confluence with MAV-1 at MOI of 10. At 3-5 days after infection when cytopathic effect (CPE) was observed, virus was harvested by collecting the scraped cells and supernate. The mixture was centrifuged (4 ´ 10³ g, 4°C, 15 min) to separate cells from supernate. Cells were resuspended in an equivalent volume of TE (mmol per liter: 10 Tris, 1 Na-EDTA; pH 8).

Hirt method: Cell suspension and supernate were processed similarly as follows.^17,18 SDS was added to a final concentration of 1% and the mixture was incubated at 37°C for 10 min. NaCl was added to a final concentration of 1 m and the mixture was incubated overnight at 4°C. Following centrifugation (16 ´ 10³ g, 15 min), the supernatant was treated with RNase A (10 g/ml, 42°C, 1 h), and extracted twice with phenol (25:24:1, phenol:chloroform:isoamyl alcohol, TE-saturated). The aqueous phase was made 0.25 m with NaCl and ethanol precipitated overnight (2.5 volumes ethanol, -20°C). DNA was pelleted (10 ´ 10³ g, 30 min, 4°C) and rinsed with 70% ethanol, dried in vacuo, resuspended in TE and used in gel analysis.

Phenol interphase: In this modified Hirt method, the phenol phase and the interphase were saved, back extracted twice with TE and precipitated with ethanol overnight (1.5 volumes ethanol, -20°C). The precipitate was pelleted (10 ´ 10³ g, 30 min, 4°C), rinsed with chloroform; and treated with proteinase K (170 ng/ml proteinase K, TE, 100 mm NaCl, and 0.5% SDS, 55°C) until all the insoluble material had disappeared. After one phenol extraction and two chloroform extractions the aqueous phase was precipitated with 2-propanol overnight (0.5 volume, -20°C). DNA was recovered by centrifugation (16 ´ 10³ g, 15 min), dried in vacuo, resuspended in TE and used for gel analysis.

Ammonium sulfate-phenol: In this modified Hirt method, the cell suspension and supernate were clarified (10 ´ 10³ g, 4°C, 30 min) and precipitated overnight with ammonium sulfate (60% by weight, 4°C). The virus pellets were resuspended in TE; the mixture was clarified (16 ´ 10³ g, 4°C, 5 min) and extracted with phenol. The phenol phase and the interphase were processed as described above.

PEG-phenol: In this modified Hirt method, the cell suspension and supernate were made 5.8% with NaCl, incubated at 0°C for 1 h, clarified (10 ´ 10³ g, 4°C, 30 min), made 7% with PEG-8000, incubated at 0°C overnight and centrifuged (10 ´ 10³ g, 4°C, 30 min). The virus pellets were resuspended in TE; the mixture was clarified (16 ´ 10³ g, 4°C, 5 min) and extracted with phenol. The phenol phase and the interphase were processed as described above.

Plasmid construction

The construction scheme is summarized in Figure 3. pMA was constructed, using the left and right ITR MAV-1 clones, pG1HAX and pG11SAB as parent plasmids. pG1HAX consists of nucleotides 1-5093 (0-16.5 MU) of the left end of MAV-1 cloned in plasmid pKJB55,³³ pG1HAX has an ApaLI site at 0 MU that was introduced by PCR. pG11SAB consists of nucleotides 26 994-30 944 (87.5-100 MU) of the right end of MAV-1 cloned in plasmid pUC18CM.³³ pG11SAB has an ApaLI site at 100 MU that was introduced by PCR. pG1HAX was digested with HindIII and EcoRI to release the 2.3 kb left end of the MAV-1 genome, which was ligated into pWSK29,²⁶ to generate pWSK/HAX (not shown). pG11SAB was cut with XbaI and PstI to release the 1.7 kb right end of the MAV-1 genome, which was ligated into pWSK/HAX, to create pMA. E. coli recBC and sbcBC strain BJ5183³⁴ was co-transfected with MAV-1 DNA, purified as described above, and pMA, linearized with EcoRI and PstI, using the CaCl₂/heat shock procedure.³⁵ Plasmid DNA was isolated from ampicillin-resistant candidates and subjected to restriction enzyme analysis. pKMA is a clone containing an intact MAV-1 genome, which can be released by ApaLI.

Acknowledgements

We thank Dr Christopher CW Hughes for providing Phec 97-54 primary umbilical endothelial cells. We thank Siddiqua Hirst and Mervin R Gutierrez for excellent technical assistance. This work was supported by a grant from the Muscular Dystrophy Association to LPV Additional facility support was from the Irvine Research Unit on Animal Viruses and the Viral Vector facility.

1 Graham FL, Prevec L. Methods for construction of adenovirus vectors. Mol Biotechnol 1995; 3: 207-220, MEDLINE

2 Kochanek S et al. A new adenoviral vector: replacement of all viral coding sequences with 28 kb of DNA independently expressing both full-length dystrophin and -galactosidase. Proc Natl Acad Sci USA 1996; 93: 5731-5736, Article MEDLINE

3 Haecker SE et al. In vivo expression of full-length human dystrophin from adenoviral vectors deleted of all viral genes. Hum Gene Ther 1996; 7: 1907-1914, MEDLINE

4 McGrory WJ, Bautista DS, Graham FL. A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5. Virology 1988; 163: 614-617, MEDLINE

5 Chartier C et al. Efficient generation of recombinant adenovirus vectors by homologous recombination in Escherichia coli. J Virol 1996; 70: 4805-4810, MEDLINE

6 Hartley JW, Rowe WP. A new mouse virus apparently related to the adenovirus group. Virology 1960; 11: 645-647,

7 Gall J, Kass-Eisler A, Leinwand L, Falck-Pedersen E. Adenovirus type 5 and 7 capsid chimera: fiber replacement alters receptor tropism without affecting primary immune neutralization epitopes. J Virol 1996; 70: 2116-2123, MEDLINE

8 Wickham TJ, Carrion ME, Kovesdi I. Targeting of adenovirus penton base to new receptors through replacement of its RGD motif with other receptor-specific peptide motifs. Neuromusc Dis 1997; 7: 284-298, MEDLINE

9 Douglas JT, Curiel DT. Strategies to accomplish targeted gene delivery to muscle cells employing tropism-modified adenoviral vectors. Neuromusc Dis 1997; 7: 284-298, MEDLINE

10 Krasnykh VN, Mikheeva GV, Douglas JT, Curiel DT. Generation of recombinant adenovirus vectors with modified fibers for altering viral tropism. J Virol 1996; 70: 6839-6846, MEDLINE

11 Blailock ZR, Rabin ER, Melnick JL. Adenovirus endocarditis in mice. Science 1967; 157: 69-70, MEDLINE

12 Heck FCJ, Sheldon WG, Gleiser CA. Pathogenesis of experimentally produced mouse adenovirus infection in mice. Am J Vet Res 1972; 33: 841-846, MEDLINE

13 Guida JD et al. Mouse adenovirus type 1 causes a fatal hemorrhagic encephalomyelitis in adult C57BL/6 but not BALB/c mice. J Virol 1995; 69: 7674-7681, MEDLINE

14 Kring SC, King CS, Spindler KR. Susceptibility and signs associated with mouse adenovirus type 1 infection of adult outbred Swiss mice. J Virol 1995; 69: 8084-8088, MEDLINE

15 Engel JP. Viral upper respiratory infections. Semin Resp Infect 1995; 10: 3-13,

16 Ketner G et al. Efficient manipulation of the human adenovirus genome as an infectious yeast artificial chromosome clone. Proc Natl Acad Sci USA 1994; 91: 6186-6190, MEDLINE

17 Hirt B. Selective extraction of polyoma DNA from infected mouse cell cultures. J Mol Biol 1967; 26: 365-369, MEDLINE

18 Graham FL, Prevec L. Manipulation of adenovirus vectors. In: Murray EJ (ed). Gene Transfer and Expression Protocols. Humana Press: Clifton, NJ, 1991, pp 109-128.

19 Graham FL, van der Eb AJ. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 1973; 52: 456-467, MEDLINE

20 Larsen SH. Evolutionary variants of mouse adenovirus containing cellular DNA sequences. Virology 1982; 116: 573-580, MEDLINE

21 Stoflet ES, Koeberl DD, Sarkar G, Sommer SS. Genomic amplification with transcript sequencing. Science 1988; 239: 491-494, MEDLINE

22 Thornton SC, Mueller SN, Levine EM. Human endothelial cells: use of heparin in cloning and long-term serial cultivation. Science 1983; 222: 623-625, MEDLINE

23 Gimbrone MAJ. Culture of vascular endothelium. Prog Hemost Thromb 1976; 3: 1-28, MEDLINE

24 Shinagawa M et al. A rapid and simple method for preparation of adenovirus DNA from infected cells. Microbiol Immunol 1983; 27: 817-822, MEDLINE

25 Cox GA et al. Overexpression of dystrophin in transgenic mdx mice eliminates dystrophic symptoms without toxicity (see comments). Nature 1993; 364: 725-729, MEDLINE

26 Wang RF, Kushner SR. Construction of versatile low copy number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 1991; 100: 195-199, MEDLINE

27 Wiegand RC, Beattie KL, Holloman WK, Radding CM. Uptake of homologous single-stranded fragments by superhelical DNA. III. The product and its enzymic conversion to a recombinant molecule. J Mol Biol 1977; 116: 805-824, MEDLINE

28 Schaack J, Ho WY, Freimuth P, Shenk T. Adenovirus terminal protein mediates both nuclear matrix association and efficient transcription of adenovirus DNA. Genes Dev 1990; 4: 1197-1208, MEDLINE

29 Miyake S et al. Efficient generation of recombinant adenoviruses using adenovirus DNA-terminal protein complex and a cosmid bearing the full-length virus genome. Proc Natl Acad Sci USA 1996; 93: 1320-1324, Article MEDLINE

30 Kring SC, Spindler KR. Lack of effect of mouse adenovirus type 1 infection on cell surface expression of major histocompatibility complex class I antigens. J Virol 1996; 70: 5495-5502, MEDLINE

31 Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ. Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 1979; 18: 5294-5299, MEDLINE

32 Meissner JD et al. Completion of the DNA sequence of mouse adenovirus type 1: sequence of E2B, L1, and L2 (18-51 map units). Virus Res 1997; 51: 53-64, MEDLINE

33 Ball AO, Williams ME, Spindler KR. Identification of mouse adenovirus type 1 early region 1: DNA sequence and a conserved transactivating function. J Virol 1988; 62: 3947-3957, MEDLINE

34 Kusano K, Nakayama K, Nakayama H. Plasmid-mediated lethality and plasmid multimer formation in an Escherichia coli recBC sbcBC mutant. Involvement of RecF recombination pathway genes. J Mol Biol 1989; 209: 623-634, MEDLINE

35 Hanahan D. Studies on transformation of Escherichia coli with plasmids. J Mol Biol 1983; 166: 557-580, MEDLINE

Figure 1 RT-PCR analysis of MAV-1 E1A transcription in mouse and primate cells. Panel I: PCR products using MAV-1 E1A specific primers. Panel II: PCR products using -actin specific primers. Lanes a, b and c: 3T6, mouse embryonal cell line; lanes d, e and f: 293 T, human embryonal kidney cell line expressing Ad2/5 E1A and SV40 large T antigen; lanes g, h and i: CV-1, green monkey kidney cell line; lanes j, k and l: Phec 97-54, primary human umbilical endothelial cells; lane m: PCR products from 10 ng MAV-1 DNA. Lanes a, d, g and j are uninfected controls; lanes b, e, h and k are 20 h post-infection samples; lanes c, f, i and l are no-RT controls. Primers and assay conditions are as described in the text.

Figure 2 Comparison of MAV-1 DNA isolated by extraction methods. Lanes a and i: Hirt extraction; lanes b and h: Phenol interphase; lanes c and g: ammonium sulfate-phenol; lanes d and f: PEG-phenol; lane e: HindIII size markers. Lanes a-d: supernates. Lanes f-i: cell pellets. MAV-1 DNA was recovered from infected 3T6 cells and their corresponding supernates using the modified Hirt, PEG precipitated-phenol, (NH₄)₂SO₄ precipitated-phenol, and phenol interphase methods as described in the text. 2 g of DNA were applied to each lane of a 0.7% agarose TAE gel.

Figure 3 Construction of plasmids pMA and pKMA. pG1HAX and pG11SAB were digested with HindIII + EcoRI, and PstI + XbaI, respectively. The 2.3 kb HindIII-EcoRI fragment from pG1HAX was inserted into the HindIII + EcoRI digested pWSK29²⁶ to yield an intermediate plasmid. The 1.7 kb PstI-XbaI fragment of pG11SAB was inserted into the PstI + XbaI digested intermediate plasmid, to yield plasmid pMA. E. coli strain BJ5183 was co-transformed with MAV-1 DNA and pMA, linearized with EcoRI. In vivo recombination yields plasmid pKMA, containing the full-length MAV-1 genome, which is releasable by ApaLI digestion. 100 map units (MU) is equivalent to 30.9 kb of MAV-1. Not to scale.

Figure 4 Restriction analysis of pKMA. pKMA was digested with EcoRV, EcoRI, ApaLI, SalI and HindIII. On the right is a linear representation of pKMA starting at 0 MU SalI and HindIII are present in the multiple cloning site of pWSK29.