Plasmid Construction. FFIL was constructed in the adenovirus yeast artificial chromosome (YAC) system (1). The Ad L5 two-step gene replacement plasmid p680FIL contains, from left to right relative to the adenovirus genome, Ad2 nucleotides 29845–32791, a polylinker, the poliovirus internal ribosome entry site (IRES) [poliovirus nucleotides 3186–3536 (2)], a polylinker, the COPV L1 gene (COPV coordinates 6837–8342), and Ad2 sequences from 32733 to 33784. The adenovirus sequences differ from wild type by A-to-G transitions at positions 32775 and 32777, which ablate the Ad L5 AATAAA sequence, and at position 32811, which eliminates an AflIII site. In the COPV L1 gene, the final amino acid (K) is replaced by E, derived from the Ad2 fiber gene. To prepare p680FIL, a PCR product (PCR1) comprising Ad2 nucleotides 32644–32791 was amplified with primers PCR1 5' and PCR1 3' (see Table 2). The fiber polyadenylation-site mutations were included in PCR1 3', and a SmaI site was added by inclusion in PCR1 5'. SpeI, SmaI-digested PCR1 was cloned between the SpeI and PmlI sites of pPBS, an IRES-containing plasmid (2). A SacII, SpeI fragment from adenovirus DNA containing Ad2 nucleotides 29845–32644 was cloned between a SacII site in pPBS and the SpeI site in PCR1 to produce pPBSFiber. PCR2, which contains the 3' end of COPV L1 (COPV nucleotides 7941–8342) fused to Ad2 sequences 32773–33784, was prepared by overlap PCR with primers PCR2l 5', PCR2l 3', PCR2r 5', and PCR2r 3'. PCR2l replaces the final codon of COPV L1 with that of Ad2 fiber; PCR2r contains the mutation that destroys the AflIII site. An XhoI site is contained at the right end of PRC2l and HpaI and HindIII sites are present at the right of PCR2r. PCR2 was cloned between XhoI and HindIII sites in pBSII SK⁺, and a fragment containing most of PCR2 was subcloned into pJS55-L1 between the natural AflIII site in COPV L1 and the HindIII site in the polylinker. The XhoI, ApaI fragment containing the L1 gene and PCR2 was then cloned in p680 (1) to produce p680L1. The SacI, SalI fragment from pPBSFiber containing the fiber gene, IRES, and PCR1 was then cloned in p680L1 to produce p680FIL. The dl2121 E3 deletion mutation was introduced into FFIL independently of the altered Ad L5 region as described in ref. 1.

The remaining recombinants were produced by recombination in transfected 293 cells between Ad5.b gal.D F (3) and appropriate shuttle plasmids. The shuttle plasmid used for construction of recombinants with the gene order fiber–COPV L1 ("standard" gene order) contained, from left to right relative to the adenovirus genome, Ad5 nucleotides 27082–27940, a unique SfiI site, which replaces 2,859 base pairs of E3 sequences, Ad5 nucleotides 30801–32788, a polylinker, and Ad5 nucleotides 32789–33788. A-to-G transitions were introduced at Ad5 nucleotide positions 32784 and 32786 to ablate the Ad L5 AATAAA site; otherwise, all adenoviral sequences were wild type. To prepare the shuttle plasmid used in the construction of FFFL, three PCR products were sequentially cloned into pCR2.1 to create PCRi+ii+iii: Ad5 nucleotides 32445–32788, containing the polyadenylation site mutations and a 3' SacII site; nucleotides 30810–31042, containing a 5' SacII site and 3' XhoI and PacI sites; and nucleotides 32789–33788, containing 5' PacI, Eco47III, FseI, and NotI sites and 3' XbaI and SgrAI sites. Independently, two PCR fragments, PCRB (Ad5 27689–27940) and PCRA (Ad5 30801–31088) also were inserted into PCR 2.1 with an intervening SfiI site added in the PCR primers to create PCRBA. An NdeI-to-BstXI fragment from viral DNA (Ad5 31088–32464) was cloned into PCRBA 3' of PCRA, and the PCRi+ii+iii fragment was added 3' of that fragment between the BstXI site in the viral sequences and an XbaI site in PCR 2.1. A SpeI-to-Tth111I fragment from viral DNA (Ad5 27082–27689) was inserted between a SpeI site in PCR 2.1 and the viral Tth111I site to produce the empty standard-order shuttle vector. An XhoI-to-NotI fragment of the COPV L1 plasmid pJS55-L1 (4) was inserted in the polylinker derived from PCRi+ii+iii to complete the shuttle plasmid. Splice acceptor regions (SAs) prepared by PCR from the Ad5 hexon gene or the SV40 large T antigen intron (see legend to Table 1 for coordinates) replaced the SacII-to-XhoI fragment in this plasmid to generate other dual-SA shuttle plasmids. A mutagenized COPV L1 gene (see below) was inserted into the empty shuttle plasmid between the XhoI and NotI sites to produce "optimized" shuttle plasmids.

To prepare recombinants with the "reverse" gene order (COPV L1–fiber), an empty reverse-order plasmid was constructed that contains Ad5 nucleotides 27082–27940, the SfiI site replacement of E3 sequences described above, Ad5 nucleotides 30801–31041, a polylinker, and Ad5 nucleotides 30855–33788. To create that plasmid, the HpaI-to-AflII fragment of the standard-order shuttle plasmid, which contains the mutant Ad L5 polyadenylation site, was replaced with a corresponding wild-type fragment to create a "3' repaired" intermediate. A new polylinker then was placed upstream of fiber by replacement of the EcoRI (Ad5 position 27331) to NdeI (31088) fragment with a modified fragment generated by using overlap PCR. To create the new fragment, primers MB1 (27306–27325), MB2 (31041–31019 plus half of the polylinker), MB3 (the remaining half of the polylinker plus Ad5 nucleotides 30855–30877), and MB4 (31093–31113) were used to amplify overlapping products from the 3' repaired plasmid. Those products then were used as templates with primers MB1 and MB4 to yield a fragment that was inserted in the 3' repaired plasmid between its EcoRI and NdeI sites. The sequence of the fiber gene and downstream adenoviral DNA in this plasmid is wild type. Upstream of the fiber gene, there is a tandem duplication of the adenoviral segment 30855–31041, which contains the fiber splice acceptor but not the E3B polyadenylation site, with the repeated segments separated by the new polylinker. To generate reverse-order shuttle plasmids, the COPV L1 gene was inserted into the polylinker between the XhoI and NotI sites of the empty reverse-order plasmid. When present, an SV40 large T antigen intron prepared by PCR was inserted in the sense orientation at the XhoI site in the polylinker.

To produce shuttle plasmids for use in creating dual polyA vectors, a 600-bp fragment containing the 3' end of the COPV L1 gene was amplified from pJS55-L1 by using a 5' primer that contains a natural Eco47III site (COPV L1 position 7783) and 3' primer that inserts a single extra A immediately 3' of the COPV L1 termination codon and also includes the 3' NotI site downstream of COPV L1 in pJS55-L1 (4). The insertion creates a new AATAAA polyadenylation signal that overlaps the termination codon of the COPV L1 gene. The mutant PCR product was inserted between the Eco47III and NotI sites either directly in a reverse-order shuttle plasmid bearing a wild-type COPV L1 gene (to create the shuttle plasmid for FLAFF) or into a plasmid bearing the optimized COPV L1 gene (pJS55-opti) . In the latter case, the optimized COPV L1 gene and the inserted A was then excised from pJS55-opti with XhoI and NotI and was used to replace the COPV L1 gene in the FLAFF shuttle plasmid.

All PCR products used in the construction of shuttle plasmids were verified by sequencing, and overall structures were confirmed by extensive restriction analysis and/or sequencing of shuttle plasmids. Portions of some recombinant viral DNAs also were sequenced to confirm the presence of specific features.

Introduction of Ad L5 into Virus. Modified Ad L5 regions were introduced into the virus by recombination between shuttle plasmid DNA and DNA–protein complex (5) prepared, as described in ref. 6, from the defective fiber deletion mutant Ad5.b gal.D F (3). Ad5.b gal.D F DNA–protein complex and shuttle plasmid were cotransfected into 293 monolayers by calcium phosphate precipitation (7). Ad5.b gal.D F lacks the fiber gene and does not form plaques on 293 cells. However, recombination between Ad5.b gal.D F and a shuttle plasmid in the regions of homology that flank Ad L5 in both molecules (approximately Ad5 nucleotides 27082–27940 and 32788–33788) will produce viral genomes in which the Ad L5 region from the shuttle plasmid, including the fiber and COPV L1 genes, has been inserted in place of the fiber deletion (Fig. 1). If fiber is expressed from the modified Ad L5 region, recombinants will form plaques on 293 cells, which complement the E1 substitution present in Ad5.b gal.D F. Therefore, plaques that appeared after cotransfection were grown into small high-titer stocks and screened for the presence of recombinants by restriction–digestion of ³²P-labeled DNA (8). Essentially, all plaques contained both recombinant virus and the Ad5.b gal.D F parent. Selected plaques were subjected to a second and, in some cases, a third round of plaque purification after treatment with 0.2% Nonidet P-40 and extraction with one-fifth vol of 1,1,2-trichlorotrifluoroethane (freon) (9). Recombinant stocks lacking Ad5.b gal.D F contamination were identified, expanded, and titrated. It should be noted that, because Ad5.b gal.D F carries a substitution in the E1 region, all the COPV L1 recombinants lack a functional E1 region and have been grown and characterized on the E1-complementing 293 cell line. Use of a recipient virus that carries a wild-type E1 region in a similar experiment will give rise to recombinants that are fully viable on any cell permissive for wild-type adenoviruses.

Mutagenesis of the COPV L1 Gene. Overlap PCR was used to modify the sequence in COPV L1 from base pairs 91–133 to create the genes used to construct FFFO, FFSO, FOFF, and FOAFF. Complementary primers MUT2 and MUT3, which contained the altered sequences, were used with outside primers MUT1 (which contains vector sequences upstream and the first seven nucleotides of COPV L1) and MUT4 (COPV positions 7264–7288), respectively, and a pJS55-L1 template to produce overlapping 136- and 371-bp products. The overlapping products served as templates for a reaction with MUT1 and MUT4 primers to yield a 471-bp product, which was digested with XhoI and HpaI and inserted between the same sites in pJS55-L1. Candidate mutant clones were screened by restriction–digestion to identify a newly engineered HaeII site within the mutagenized segment, and a representative clone was sequenced to confirm its sequence. The entire mutagenized L1 gene (XhoI–NotI ) was then ligated into appropriate shuttle vectors.

Immunoblots. Growing in 24-well dishes (5 × 10⁵ cells per well), 293 cells were infected with recombinants, Ad5, or H2dl810 at a multiplicity of infection (MOI) of 10 plaque-forming units (pfu) per cell. At 48 h postinfection, cell lysates were made by the addition of 200 m l of Laemmli sample buffer [protein loading solution (PLS)] (10) to each well. Twenty-microliter samples were fractionated on 10% SDS polyacrylamide gels (10) and transferred electrophoretically to nitrocellulose (11). COPV-LI was detected by using the AU-1 mouse monoclonal antibody, which reacts with linear epitopes (12), adenovirus fiber with 4D2 monoclonal antibody (NeoMarkers, Lab Vision, Fremont, CA), adenoviral late-gene products with an antiserum directed against SDS-disrupted adenovirus capsids (13), and actin with a commercial monoclonal antibody (MP Biomedicals, Irvine, CA). Reactions with primary antibodies were performed at 4°C overnight in 5% nonfat dry milk. Reactions with horseradish peroxidase (HRP)-conjugated secondary antibodies (Amersham Pharmacia) were performed at 25°C in 3% nonfat dry milk for 45 min. Chemiluminescent detection by using ECL reagents (Amersham Pharmacia) was as recommended by the manufacturer.

Northern Blots. Probes used for Northern blots were as follows: COPV L1, COPV nucleotides 6837–8343; fiber, Ad5 nucleotides 31076–32464; and late-region 4, Ad5 nucleotides 27082–27332. Hybridization probes were labeled with the steroid hapten digoxigenin (DIG) by random hexanucleotide priming by using the DIG High Prime kit (Roche), as recommended by the manufacturer. Filters were prehybridized for 4 h and hybridized with 25 ng of probe overnight, both at 42°C. The hybridized probe was detected by chemiluminescence with an anti-DIG HRP-conjugated antibody according to the kit protocol.

Metabolic Labeling of Infected Cell Proteins. Infection of 293 cells was performed at a MOI of 0.25. At 96 h postinfection, cells were washed in medium lacking methionine and cysteine and incubated for 6 h in medium containing 40 m Ci/ml (1 Ci = 37 GBq) Promix [³⁵S]methionine and [³⁵S]cysteine. Cells were washed with PBS, lysed with 150 m l of radioimmunoprecipitation buffer [RIPA; 10 mM Tris (pH 7.5)/0.15 M NaCl/2 mM EDTA/1% sodium deoxycholate/1% Nonidet P-40/0.1% SDS], and either immunoprecipitated or combined with 150 m l of 2× PLS. Immunoprecipitates or lysates were resolved on a 9% SDS/PAGE and autoradiographed.

Immunoprecipitation of COPV L1 with Conformation-Specific Antibody. Growing in six-well dishes, 293 cells (2 × 10⁶ cells per dish) were infected at a MOI of 10 with H2dl810 or FOAFF. At 48 h postinfection, cells were washed once in PBS and lysed in 750 m l of RIPA buffer supplemented with protease inhibitors (1 m g/ml each pepstatin A, aprotinin, and leupeptin) (11). Some samples were incubated in a boiling water bath for 10 min to denature the COPV L1 present; otherwise, all spins and manipulations were performed at 4°C. Lysates were clarified by centrifugation for 10 min at 12,000 × g in a microcentrifuge, and 700 m l of samples were preadsorbed by incubation with 50 m l of a 50% suspension of Protein A immobilized on Sepharose CL-4B beads (Sigma) in RIPA buffer for 30 min with continuous agitation. Samples were briefly centrifuged to remove beads and were then divided in half and incubated with 2 m l of either normal rabbit serum or polyclonal rabbit COPV L1 conformation-specific antibody (14) at a 1:175 dilution overnight. Protein A-Sepharose beads (50 m l) were added, and the mixture was incubated for an additional hour. The beads were washed three times with 1 ml of RIPA buffer, resuspended in 150 m l of Laemmli sample buffer, and boiled for 5 min. After removal of the beads by centrifugation, the samples were electrophoresed on 10% SDS/polyacrylamide gel and transferred to nitrocellulose. COPV L1 was detected by blotting with the AU1 antibody. Loading was shown to be equivalent by immunoblotting of COPV L1 and actin in lysates without immunoprecipitation.

Preparation of VLPs and Electron Microscopy. FFSO-infected cells (ten 9-cm plates infected at a MOI of 5 pfu per cell), were collected by centrifugation at 72 h postinfection and resuspended in 20 ml of supernatant. Nonidet P-40 was added (final concentration of 0.2%), and the mixture was extracted with a one-fifth vol of freon. Clarified extracted lysate was layered over a discontinuous CsCl gradient containing 5 ml of CsCl density 1.25 g/ml and 5 ml of CsCl density 1.70 g/ml in a 35-ml centrifuge tube. Whole adenovirus, empty adenovirus particles, and COPV L1 VLPs were sedimented onto the CsCl cushion by centrifugation for 90 min at 16,500 rpm (Sorvall SV288 vertical rotor, Kendro Laboratory Products, Asheville, NC) at 4°C. Two or three bands were apparent after centrifugation; the lower band was collected individually, and the upper bands were collected together. The pooled upper bands were centrifuged again in a continuous CsCl gradient of average density 1.315 g/ml at 35,000 rpm (Sorvall AH 650 swinging bucket rotor, Kendro Laboratory Products) for 16 h at 4°C. The band at density 1.29 g/ml in the second gradient was centrifuged in a third gradient of average density 1.29 g/ml at 35,000 rpm (Sorvall TV865 vertical rotor) for 16 h at 4°C. The central portion of the gradient was collected as 200-m l fractions. The A₂₈₀ of each fraction was determined as an estimate of protein content, and the presence of adenoviral and COPV L1 proteins was assessed by immunoblotting.

Carbon-coated gold grids (400 mesh) were negatively charged with a corona discharge apparatus. Charged grids were floated on drops of COPV L1-containing fractions from CsCl gradients for 70 s. The grids were briefly rinsed on drops of water and PBS to remove CsCl and were negatively stained by floating on drops of 2% uranyl acetate for 1 min. Stain was removed from the grids by touching the edge to a piece of Whatman filter paper, and the grids were allowed to air dry. For gold labeling, rinsed unstained grids were transferred to drops of PBS containing the COPV L1 conformation-specific rabbit polyclonal antibody (diluted 1:500) for 20 min. After three brief rinses in PBS, the grids were placed on drops containing an anti-rabbit gold-conjugated antibody (6-nm particles, 1:10 dilution, Jackson ImmunoResearch) for 6 min. The grids were washed three times in PBS and twice in water. The grids were negatively stained as above and examined with a Philips CM12 electron microscope and photographed at a magnification of 52,000 or 37,500. The negatives were scanned, and Fig. 6 was assembled in PHOTOSHOP (Adobe, San Jose, CA) after adjustment of contrast.

1. Ketner, G., Spencer, F., Tugendreich, S., Connelly, C. & Hieter, P. (1994) Proc. Natl.

Acad. Sci. USA 91, 6186-6190.

2. Morgan, R. A., Couture, L., Elroy-Stein, O., Ragheb, J., Moss, B. & Anderson, W. F.

(1992) Nucleic Acids Res. 20, 1293-1299.

3. Von Seggern, D. J., Chiu, C. Y., Fleck, S. K., Stewart, P. L. & Nemerow, G. R.

(1999) J. Virol. 73, 1601-1608.

4. Suzich, J. A., Ghim, S. J., Palmer-Hill, F. J., White, W. I., Tamura, J. K., Bell, J. A., Newsome, J. A., Jenson, A. B. & Schlegel, R. (1995) Proc. Natl. Acad. Sci. USA 92, 11553-11557.

5. Robinson, A. J., Younghusband, H. B. & Bellett, A. J. (1973) Virology 56, 54-69.

6. Ketner, G. & Boyer, J. (1998) in Adenovirus Methods and Protocols, ed. Wold, W. S. M. (Humana, Totowa, NJ).

7. Graham, F. L. & van der Eb, A. J. (1973) Virology 52, 456-467.

8. Bridge, E. & Ketner, G. (1989) J. Virol. 63, 631-638.

9. Challberg, S. S. & Ketner, G. (1981) Virology 114, 196-209.

10. Laemmli, U. K. & Quittner, S. F. (1974) Virology 62, 483-499.

11. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab. Press, Woodbury, NY).

12. Lim, P. S., Jenson, A. B., Cowsert, L., Nakai, Y., Lim, L. Y., Jin, X. W. & Sundberg, J. P. (1990) J. Infect. Dis. 162, 1263-1269.

13. Ketner, G., Bridge, E., Virtanen, A., Hemstrom, C. & Pettersson, U. (1989) Nucleic Acids Res. 17, 3037-3048.

14. Chen, Y., Ghim, S. J., Jenson, A. B. & Schlegel, R. (1998) J. Gen. Virol. 79, 2137-2146.