Complementing cell lines

A High Throughput Assay for the Detection of Neutralizing Activity in Human Serum

To enable screening of a large amount of human sera for the presence of neutralizing antibodies against all adenovirus serotypes, an automated 96-wells assay was developed.

Human Sera

A panel of 100 individuals was selected. Volunteers (50% male, 50% female) were healthy individuals between ages 20 and 60 years old with no restriction for race. All volunteers signed an informed consent form. People professionally involved in adenovirus research were excluded.

Approximately 60 ml blood was drawn in dry tubes. Within two hours after sampling, the blood was centrifuged at 2500 rpm for 10 minutes. Approximately 30 ml serum was transferred to polypropylene tubes and stored frozen at −20° C. until further

Serum was thawed and heat-inactivated at 56° C. for 10 minutes and then aliquoted to prevent repeated cycles of freeze/thawing. Part was used to make five steps of twofold dilutions in medium (DMEM, Gibco BRL) in a quantity enough to fill out approximately 70 96-well plates. Aliquots of undiluted and diluted sera were pipetted in deep well plates (96-well format) and using a programmed platemate dispensed in 100 μl aliquots into 96-well plates. This way the plates were loaded with eight different sera in (100 μl/well) according to the scheme below:

Where S1/2 to S8/2 in columns 1 and 6 represent 1× diluted sera and Sx/4, Sx/8, Sx/16 and Sx/32 the twofold serial dilutions. The last plates also contained four wells filled with 100 μl fetal calf serum as a negative control. Plates were kept at −20° C. until further use.

Preparation of Human Adenovirus Stocks

Prototypes of all known human adenoviruses were inoculated on T25 flasks seeded with PER.C6 cells (Fallaux et al., 1998) and harvested upon full CPE. After freeze/thawing 1-2 ml of the crude lysates was used to inoculate a T80 flask with PER.C6 and virus was harvested at full CPE. The timeframe between inoculation and occurrence of CPE as well as the amount of virus needed to re-infect a new culture, differed between serotypes. Adenovirus stocks were prepared by freeze/thawing and used to inoculate 3-4 T175 cm 2 three-layer flasks with PER.C6 cells. Upon occurrence of CPE, cells were harvested by tapping the flask, pelleted and virus was isolated and purified by a two-step CsCl gradient as follows. Cell pellets were dissolved in 50 ml 10 mM NaPO 4 buffer (pH 7.2) and frozen at −20° C. After thawing at 37° C., 5.6 ml sodium deoxycholate (5% w/v) was added. The solution was mixed gently and incubated for 5-15 minutes at 37° C. to completely lyse the cells. After homogenizing the solution, 1875 μl 1M MgCl 2 was added. After the addition of 375 μl DNAse (10 mg/ml) the solution was incubated for 30 minutes at 37° C. Cell debris was removed by centrifugation at 1880×g for 30 minutes at RT without brake. The supernatant was subsequently purified from proteins by extraction with FREON (3x). The cleared supernatant was loaded on a 1M Tris/HCl buffered cesium chloride block gradient (range: 1.2/1.4 g/ml) and centrifuged at 21000 rpm for 2.5 hours at 10° C. The virus band is isolated after which a second purification using a 1 M Tris/HCl buffered continues gradient of 1.33 g/ml of cesium chloride was performed. The virus was then centrifuged for 17 hours at 55000 rpm at 10° C. The virus band is isolated and sucrose (50% w/v) is added to a final concentration of 1%. Excess cesium chloride is removed by dialysis (three times 1 hr at RT) in dialysis slides (Slide-a-lizer, cut off 10000 kDa, Pierce, USA) against 1.5 liter PBS supplemented with CaCl 2 (0.9 mM), MgCl 2 (0.5 mM) and an increasing concentration of sucrose (1, 2, 5%). After dialysis, the virus is removed from the slide-a-lizer after which it is aliquoted in portions of 25 and 100 μl upon which the virus is stored at −85° C.

To determine the number of virus particles per milliliter, 50 μl of the virus batch is run on a high-pressure liquid chromatograph (HPLC) as described by Shabram et al (1997). Viruses were eluted using a NaCl gradient ranging from 0 to 600 mM. As depicted in table I, the NaCl concentration by which the viruses were eluted differed significantly among serotypes.

Most human adenoviruses replicated well on PER.C6 cells with a few exceptions. Adenovirus type 8 and 40 were grown on 911-E4 cells (He et al., 1998). Purified stocks contained between 5×10 10 and 5×10 12 virus particles/ml (VP/ml; see table I).

Titration of Purified Human Adenovirus Stocks

Adenoviruses were titrated on PER.C6 cells to determine the amount of virus necessary to obtain full CPE in five days, the length of the neutralization assay. Hereto, 100 μl medium was dispensed into each well of 96-well plates. 25 μl of adenovirus stocks pre-diluted 10 4 , 10 5 , 10 6 or 10 7 times were added to column 2 of a 96-well plate and mixed by pipetting up and down 10 times. Then 25 μl was brought from column 2 to column 3 and again mixed. This was repeated until column 1 1 after which 25 μl from column 11 was discarded. This way, serial dilutions in steps of 5 were obtained starting off from a pre-diluted stock. Then 3×10 4 PER.C6 cells (ECACC deposit number 96022940) were added in a 100 μl volume and the plates were incubated at 37° C., 5% CO 2 for five or six days. CPE was monitored microscopically. The method of Reed and Muensch was used to calculate the cell culture-inhibiting dose 50% (CCID50).

In parallel, identical plates were set up that were analyzed using the MTT assay (Promega). In this assay, living cells are quantified by calorimetric staining. Hereto, 20 μl MTT (7.5 mgr/ml in PBS) was added to the wells and incubated at 37° C., 5% CO 2 for two hours. The supernatant was removed and 100 μl of a 20:1 isopropanol/triton-X100 solution was added to the wells. The plates were put on a 96-well shaker for 3-5 minutes to solubilize the precipitated staining. Absorbance was measured at 540 nm and at 690 nm (background). By this assay, wells with proceeding CPE or full CPE can be distinguished.

Neutralization Assay

96-well plates with diluted human serum samples were thawed at 37° C., 5% CO 2 . Adenovirus stocks diluted to 200 CCID50 per 50 μl were prepared and 50 μl aliquots were added to columns 1-11 of the plates with serum. Plates were incubated for 1 hour at 37° C., 5% CO 2 . Then 50 μl PER.C6 cells at 6×10 5 /ml were dispensed in all wells and incubated for 1 day at 37° C., 5% CO2. Supernatant was removed using fresh pipette tips for each row and 200 μl fresh medium was added to all wells to avoid toxic effects of the serum. Plates were incubated for another 4 days at 37° C., 5% CO 2 . In addition, parallel control plates were set up in duplo with diluted positive control sera generated in rabbits and specific for each serotype to be tested in rows A and B and with negative control serum (FCS) in rows C and D. Also, in each of the rows E-H a titration was performed as described above with steps of five times dilutions starting with 200 CCID50 of each virus to be tested. On day 5, one of the control plates was analyzed microscopically and with the MTT assay. The experimental titer was calculated from the control titration plate observed microscopically. If CPE was found to be complete, i.e. the first dilution in the control titration experiment analyzed by MTT shows clear cell death, all assay plates were processed. If not, the assay was allowed to proceed for one or more days until full CPE was apparent after which all plates were processed. In most cases, the assay was terminated at day 5. For Ad1, 5, 33, 39, 42 and 43 the assay was left for six days and for Ad2 for eight days.

A serum sample is regarded as “non-neutralizing” when, at the highest serum concentration, a maximum protection of 40% is seen compared to controls without serum.

The results of the analysis of 44 prototype adenoviruses against serum from 100 healthy volunteers are shown in FIG. 1 . As expected, the percentage of serum samples that contained neutralizing antibodies to Ad2 and Ad5 was very high. This was also true for most of the lower numbered adenoviruses. Surprisingly, none of the serum samples contained neutralizing antibodies to Ad35. Also, the number of individuals with neutralizing antibody titers to the serotypes 26, 34 and 48 was very low. Therefore, recombinant E1-deleted adenoviruses based on Ad35 or one of the other above mentioned serotypes have an important advantage compared to recombinant vectors based on Ad5 with respect to clearance of the viruses by neutralizing antibodies.

Also, Ad5-based vectors that have (parts of) the capsid proteins involved in immunogenic response of the host replaced by the corresponding (parts of) the capsid proteins of Ad35 or one of the other serotypes will be less, or even not, neutralized by the vast majority of human sera.

As can be seen in Table I, the VP/CCID50 ratio calculated from the virus particles per ml and the CCID50 obtained for each virus in the experiments was highly variable, and ranged from 0.4 to 5 log. This is probably caused by different infection efficiencies of PER.C6 cells and by differences in replication efficiency of the viruses. Furthermore, differences in batch qualities may play a role. A high VP/CCID50 ratio means that more viruses were put in the wells to obtain CPE in 5 days. As a consequence, the outcome of the neutralization study might be biased since more (inactive) virus particles could shield the antibodies. To check whether this phenomenon had taken place, the VP/CCID50 ratio was plotted against the percentage of serum samples found positive in the assay (FIG. 2 ). The graph clearly shows that there is no negative correlation between the amount of viruses in the assay and neutralization in serum.

The Prevalence of Neutralizing Activity (NA) to Ad35 is Low in Human Sera from Different Geographic Locations

In Example 1 the analysis of neutralizing activity (“NA”) in human sera from one location in Belgium was described. Strikingly, of a panel of 44 adenovirus serotypes tested, one serotype, Ad35, was not neutralized in any of the 100 sera assayed. In addition, a few serotypes, Ad26, Ad34 and Ad48 were found to be neutralized in 8%, or less, of the sera tested. This analysis was further extended to other serotypes of adenovirus not previously tested and, using a selection of serotypes from the first screen, was also extended to sera from different geographic locations.

Hereto, adenoviruses were propagated, purified and tested for neutralization in the CPE-inhibition assay as described in Example 1. Using the sera from the same batch as in Example 1, adenovirus serotypes 7B, 11, 14, 18 and 44/1876 were tested for neutralization. These viruses were found to be neutralized in, respectively, 59, 13, 30, 98 and 54% of the sera. Thus, of this series, Ad11 is neutralized with a relatively low frequency.

Since it is known that the frequency of isolation of adenovirus serotypes from human tissue as well as the prevalence of NA to adenovirus serotypes may differ on different geographic locations, we further tested a selection of the adenovirus serotypes against sera from different places. Human sera were obtained from two additional places in Europe (Bristol, UK and Leiden, NL) and from two places in the United States (Stanford, Calif. and Great Neck, N.Y.). Adenoviruses that were found to be neutralized in 20% or less of the sera in the first screen, as well as Ad2, Ad5, Ad27, Ad30, Ad38, Ad43, were tested for neutralization in sera from the UK. The results of these experiments are presented in FIG. 3 . Adenovirus serotypes 2 and 5 were again neutralized in a high percentage of human sera. Furthermore, some of the serotypes that were neutralized in a low percentage of sera in the first screen are neutralized in a higher percentage of sera from the UK, for example, Ad26 (7% vs. 30%), Ad28 (13% vs. 50%), Ad34 (5% vs. 27%) and Ad48 (8% vs. 32%). Neutralizing activity against Ad11 and Ad49 that were found in a relatively low percentage of sera in the first screen, are found in an even lower percentage of sera in this second screen (13% vs. 5% and 20% vs. 11% respectively). Serotype Ad35 that was not neutralized in any of the sera in the first screen, was now found to be neutralized in a low percentage (8%) of sera from the UK. The prevalence of NA in human sera from the UK is the lowest to serotypes Ad11 and Ad35.

For further analysis, sera obtained from two locations in the US (Stanford, Calif. and Great Neck, N.Y.) and from The Netherlands (Leiden). FIG. 4 presents an overview of data obtained with these sera and the previous data. Not all viruses were tested in all sera, except for Ad5, Ad11 and Ad35. The overall conclusion from this comprehensive screen of human sera is that the prevalence of neutralizing activity to Ad35 is the lowest of all serotypes throughout the western countries: on average 7% of the human sera contain neutralizing activity (5 different locations). Another B-group adenovirus, Ad11 is also neutralized in a low percentage of human sera (average 11% in sera from 5 different locations). Adenovirus type 5 is neutralized in 56% of the human sera obtained from 5 different locations. Although not tested in all sera, D-group serotype 49 is also neutralized with relatively low frequency in samples from Europe and from one location of the US (average 14%).

In the herein described neutralization experiments, a serum is judged non-neutralizing when, in the well with the highest serum concentration, the maximum protection of CPE is 40% compared to the controls without serum. The protection is calculated as follows: %     protection = OD     corresponding     well - OD     virus     control OD     non  -  infected     control - OD     virus     control × 100  %

As described in Example 1, the serum is plated in five different dilutions ranging from 4× to 64× diluted. Therefore, it is possible to distinguish between low titers (i.e., neutralization only in the highest serum concentrations) and high titers of NA (i.e., also neutralization in wells with the lowest serum concentration). Of the human sera used in our screen that were found to contain neutralizing activity to Ad5, 70% turned out to have high titers whereas of the sera that contained NA to Ad35, only 15% had high titers. Of the sera that were positive for NA to Ad11 only 8% had high titers. For Ad49, this was 5%. Therefore, not only is the frequency of NA to Ad35, Ad11 and Ad49 much lower as compared to Ad5, but of the sera that do contain NA to these viruses, the vast majority has low titers. Adenoviral vectors based on Ad 11, Ad35 or Ad49 have therefore a clear advantage over Ad5 based vectors when used as gene therapy vehicles or vaccination vectors in vivo or in any application where infection efficiency is hampered by neutralizing activity.

In the following examples, the construction of a vector system for the generation of safe, RCA-free Ad35-based vectors is described.

Sequence of the Human Adenovirus Type 35

Ad35 viruses were propagated on PER.C6 cells and DNA was isolated as follows: To 100 μl of virus stock (Ad35: 3.26×10 12 VP/mil), 10μl 10× DNAse buffer (130 mM Tris-HCl pH7.5; 1,2 M CaCl 2 ; 50 mM MgCl 2 ) was added. After addition of 10 μl 10 mgr/ml DNAse I (Roche Diagnostics), the mixture was incubated for 1 hr. at 37° C. Following addition of 2.5 μl 0.5M EDTA, 3.2 μl 20% SDS and 1.5μl ProteinaseK (Roche Diagnostics; 20 mgr/ml), samples were incubated at 50° C. for 1 hr. Next, the viral DNA was isolated using the GENECLEAN spin kit (Bio101 Inc.) according to the manufacturer's instructions. DNA was eluted from the spin column with 25 μl sterile MilliQ water. The total sequence was generated by Qiagen Sequence Services (Qiagen GmbH, Germany). Total viral DNA was sheared by sonification and the ends of the DNA were made blunt by T4 DNA polymerase. Sheared blunt fragments were size fractionated on agarose gels and gel slices corresponding to DNA fragments of 1.8 to 2.2 kb were obtained. DNA was purified from the gel slices by the QIAquick gel extraction protocol and subcloned into a shotgun library of pUCl9 plasmid cloning vectors. An array of clones in 96-well plates covering the target DNA 8 (+/−2) times was used to generate the total sequence. Sequencing was performed on Perkin-Elmer 9700 thermocyclers using Big Dye Terminator chemistry and AmpliTaq FS DNA polymerase followed by purification of sequencing reactions using QIAGEN DyeEx 96 technology. Sequencing reaction products were then subjected to automated separation and detection of fragments on ABI 377 XL 96 lane sequencers. Initial sequence results were used to generate a contiguous sequence and gaps were filled in by primer walking reads on the target DNA or by direct sequencing of PCR products. The ends of the virus turned out to be absent in the shotgun library, most probably due to cloning difficulties resulting from the amino acids of pTP that remain bound to the ITR sequences after proteinase K digestion of the viral DNA. Additional sequence runs on viral DNA solved most of the sequence in those regions, however it was difficult to obtain a clear sequence of the most terminal nucleotides. At the 5′ end the sequence portion obtained was 5′-CCAATAATATACCT-3′ (SEQ ID NO: 1) while at the 3′ end, the obtained sequence portion was 5′-AGGTATATTATTGATGATGGG-3′ (SEQ ID NO: 2). Most human adenoviruses have a terminal sequence 5′-CATCATCAATAATATACC-3′ (SEQ ID NO: 3). In addition, a clone representing the 3′ end of the Ad35 DNA obtained after cloning the terminal 7 kb Ad35 EcoRI fragment into pBr322 also turned out to have the typical CATCATCAATAAT . . . sequence. Therefore, Ad35 may have the typical end sequence and the differences obtained in sequencing directly on the viral DNA are due to artefacts correlated with run-off sequence runs and the presence of residual amino acids of pTP.

The total sequence of Ad35 with corrected terminal sequences is given in FIG. 5 and identified as SEQ ID NO: 39 in the SEQUENCE LISTING. Based sequence homology with Ad5 (Genbank #M72360) and Ad7 (partial sequence Genbank #X03000) and on the location of open reading frames, the organization of the virus is identical to the general organization of most human adenoviruses, especially the subgroup B viruses. The total length of the genome is 34,794 basepairs.

Construction of a Plasmid-based Vector System to Generate Recombinant Ad35-based Viruses

A functional plasmid-based vector system to generate recombinant adenoviral vectors comprises the following components:

1. An adapter plasmid comprising a left ITR and packaging sequences derived from Ad35 and at least one restriction site for insertion of an heterologous expression cassette and lacking E1 sequences. Furthermore, the adapter plasmid contains Ad35 sequences 3′ from the E1B coding region including the pIX promoter and coding sequences enough to mediate homologous recombination of the adapter plasmid with a second nucleic acid molecule.

2. A second nucleic acid molecule, comprising sequences homologous to the adapter plasmid, and Ad35 sequences necessary for the replication and packaging of the recombinant virus, that is early, intermediate and late genes that are not present in the packaging cell.

3. A packaging cell providing at least functional E1 proteins capable of complementing the E1 function of Ad35.

Other methods for generating recombinant adenoviruses on complementing packaging cells are known in the art, and may be applied to Ad35 viruses without departing from the invention. As an example, the construction of a plasmid-based system, as outlined above, is described in detail below.

1) Construction of Ad35 Adapter Plasmids.

The adapter plasmid pAdApt (described in International Patent Application WO99/55 132) was first modified to obtain adapter plasmids that contain extended polylinkers and that have convenient unique restriction sites flanking the left ITR and the adenovirus sequence at the 3′ end to enable liberation of the adenovirus insert from plasmid vector sequences. Construction of these plasmids is described below in detail:

Adapter plasmid pAdApt was digested with SalI and treated with Shrimp Alkaline Phosphatase to reduce religation. A linker, composed of the following two phosphorylated and annealed oligos: ExSalPacF 5′-TCG ATG GCA AAC AGC TAT TAT GGG TAT TAT GGG TTC GAA TTA ATT AA-3′ (SEQ ID NO: 4) and ExSalPacR 5′-TCG ATT AAT TAA TTC GAA CCC ATA ATA CCC ATA ATA GCT GTT TGC CA-3′ (SEQ ID NO: 5) was directly ligated into the digested construct, thereby replacing the SalI restriction site by Pi-PspI, SwaI and PacI. This construct was designated pADAPT+ExSalPac linker. Furthermore, part of the left ITR of pAdApt was amplified by PCR using the following primers: PCLIPMSF: 5′-CCC CAA TTG GTC GAC CAT CAT CAA TAA TAT ACC TTA TTT TGG-3′ (SEQ ID NO: 6) and pCLIPBSRGI: 5′-GCG AAA ATT GTC ACT TCC TGT G-3′ (SEQ ID NO: 7). The amplified fragment was digested with MunI and BsrGI and cloned into pAd5/Clip (described in International Patent Application WO99/55132), which was partially digested with EcoRI and after purification digested with BsrGI, thereby re-inserting the left ITR and packaging signal. After restriction enzyme analysis, the construct was digested with ScaI and SgrAI and an 800 bp fragment was isolated from gel and ligated into ScaI/SgrAI digested pADAPT+ExSalPac linker. The resulting construct, designated pIPspSalAdapt, was digested with SalI, dephosphorylated, and ligated to the phosphorylated ExSalPacF/ExSalPacR double-stranded linker previously mentioned. A clone in which the PacI site was closest to the ITR was identified by restriction analysis and sequences were confirmed by sequence analysis. This novel pAdApt construct, termed pIPspAdapt thus harbours two ExSalPac linkers containing recognition sequences for PacI, PI-PspI and BstBI, which surround the adenoviral part of the adenoviral adapter construct, and which can be used to linearize the plasmid DNA prior to cotransfection with adenoviral helper fragments.

In order to further increase transgene cloning permutations, a number of polylinker variants were constructed based on pIPspAdapt. For this purpose, pIPspAdapt was first digested with EcoRI and dephosphorylated. A linker composed of the following two phosphorylated and annealed oligos: Ecolinker+: 5′-AAT TCG GCG CGC CGT CGA CGA TAT CGA TAG CGG CCG C-3′(SEQ ID NO: 8) and Ecolinker−: 5′-AAT TGC GGC CGC TAT CGA TAT CGT CGA CGG CGC GCC G-3′ (SEQ ID NO: 9) was ligated into this construct, thereby creating restriction sites for AscI, SalI, EcoRV, ClaI and NotI. Both orientations of this linker were obtained, and sequences were confirmed by restriction analysis and sequence analysis. The plasmid containing the polylinker in the order 5′ HindIII, KpnI, Agel, EcoRI, AscI, SalI, EcoRV, ClaI, NotI, NheI, HpaI, BamHI and XbaI was termed pIPspAdapt1 while the plasmid containing the polylinker in the order HindIII, KpnI, AgeI, NotI, ClaI, EcoRV, SalI, AscI, EcoRI, NheI, HpaI, BamHI and XbaI was termed pIPspAdapt2.

To facilitate the cloning of other sense or antisense constructs, a linker composed of the following two oligonucleotides was designed, to reverse the polylinker of pIPspAdapt: HindXba+5′-AGC TCT AGA GGA TCC GTT AAC GCT AGC GAA TTC ACC GGT ACC AAG CTT A-3′ (SEQ ID NO: 10); HindXba-5′-CTA GTA AGC TTG GTA CCG GTG AAT TCG CTA GCG TTA ACG GAT CCT CTA G-3′ (SEQ ID NO: 11). This linker was ligated into HindIII/XbaI digested pIPspAdapt and the correct construct was isolated. Confirmation was done by restriction enzyme analysis and sequencing. This new construct, pIPspAdaptA, was digested with EcoRI and the previously mentioned Ecolinker was ligated into this construct. Both orientations of this linker were obtained, resulting in pIPspAdapt3, which contains the polylinker in the order XbaI, BamHI, HpaI, NheI, EcoRI, AscI, SalI, EcoRV, ClaI, NotI, Agel, KpnI and HindIII. All sequences were confirmed by restriction enzyme analysis and sequencing.

Adapter plasmids based on Ad35 were then constructed as follows:

The left ITR and packaging sequence corresponding to Ad35 wt sequences nucleotides 1 to 464 (FIG. 5) were amplified by PCR on wtAd35 DNA using the following primers:

Primer 35F1:

5′-CGG AAT TCT TAA TTA ATC GAC ATC ATC AAT AAT ATA CCT TAT AG-3′ (SEQ ID NO: 12)

Primer 35R2:

5′-GGT GGT CCT AGG CTG ACA CCT ACG TAA AAA CAG-3′ (SEQ ID NO: 13)

Amplification introduces a PacI site at the 5′ end and an AvrII site at the 3′ end of the sequence.

For the amplification, Platinum Pfx DNA polymerase enzyme (LTI) was used according to manufacturer's instructions, but with primers at 0.6 μM and with DMSO added to a final concentration of 3%. Amplification program was as follows: 2 min. at 94° C., (30 sec. 94° C., 30 sec. at 56° C., 1 min. at 68° C.) for 30 cycles, followed by 10 min. at 68° C.

The PCR product was purified using a PCR purification kit (LTI) according to the manufacturer's instructions, and digested with PacI and AvrII. The digested fragment was then purified from gel using the GENECLEAN kit (Bio 101, Inc.). The Ad5-based adapter plasmid pIPspAdApt-3 was digested with AvrII and then partially with PacI and the 5762 bp fragment was isolated in an LMP agarose gel slice and ligated with the abovementioned PCR fragment digested with the same enzymes and transformed into electrocompetent DH10B cells (LTI). The resulting clone is designated pIPspAdApt3-Ad351ITR.

In parallel, a second piece of Ad35 DNA was amplified using the following primers:

35F3: 5′-TGG TGG AGA TCT GGT GAG TAT TGG GAA AAC-3′ (SEQ ID NO: 14)

35R4: 5′-CGG AAT TCT TAA TTA AGG GAA ATG CAA ATC TGT GAG G-3′ (SEQ ID NO: 15)

The sequence of this fragment corresponds to nucleotides 3401 to 4669 of wtAd35 (FIG. 5) and contains 1.3 kb of sequences starting directly 3′ from the E1B 55k coding sequence. Amplification and purification were done as previously described herein for the fragment containing the left ITR and packaging sequence. The PCR fragment was then digested with PacI and subcloned into pNEB 193 vector (New England Biolabs) digested with Smal and PacI. The integrity of the sequence of the resulting clone was checked by sequence analysis. pNEB/Ad35pF3R4 was then digested with BglII and PacI and the Ad35 insert was isolated from gel using the QIAExII kit (Qiagen). pIPspAdApt3-Ad351ITR was digested with BglII and then partially with PacI. The 3624 bp fragment (containing vector sequences, the Ad35 ITR and packaging sequences as well as the CMV promoter, multiple cloning region and polyA signal) was also isolated using the QIAExII kit (Qiagen). Both fragments were ligated and transformed into competent DH10B cells (LTI). The resulting clone, pAdApt35IP3, has the expression cassette from pIPspAdApt3 but contains the Ad35 left ITR and packaging sequences and a second fragment corresponding to nucleotides 3401 to 4669 from Ad35. A second version of the Ad35 adapter plasmid having the multiple cloning site in the opposite orientation was made as follows:

pIPspAdapt1 was digested with NdeI and BglII and the 0.7 kbp band containing part of the CMV promoter, the MCS and SV40 polyA was isolated and inserted in the corresponding sites of pAdApt35IP3 generating pAdApt35IP1 (FIG. 6 ).

pAdApt35.LacZ and pAdApt35.Luc adapter plasmids were then generated by inserting the transgenes from pcDNA.LacZ (digested with KpnI and BamHI) and pAdApt.Luc (digested with HindIII and BamHI) into the corresponding sites in pAdApt35IP1. The generation of pcDNA.LacZ and pAdApt.Luc is described in International Patent Application WO99/55132.

2) Construction of Cosmid pWE.Ad35.pIX-rITR

FIG. 7 presents the various steps undertaken to construct the cosmid clone containing Ad35 sequences from bp 3401 to 34794 (end of the right ITR) that are described in detail below.

A first PCR fragment (pIX-NdeI) was generated using the following primer set:

35F5: 5′-CGG AAT TCG CGG CCG CGG TGA GTA TTG GGA AAA C-3′ (SEQ ID NO: 16)

35R6: 5′-COC CAG ATC GTC TAC AGA ACA G-3′ (SEQ ID NO: 17)

DNA polymerase Pwo (Roche) was used according to manufacturer's instructions, however, with an end concentration of 0.6 pM of both primers and using 50 ngr wt Ad35 DNA as template. Amplification was done as follows: 2 min. at 94° C., 30 cycles of 30 sec. at 94° C., 30 sec. at 65° C. and 1 min. 45 sec. at 72° C., follow at 68° C. To enable cloning in the TA cloning vector PCR2.1, a last incubation with 1 unit superTaq polymerase (HT Biotechnology LTD) for 10 min. at 72° C. was performed.

The 3370 bp amplified fragment contains Ad35 sequences from bp 3401 to 6772 with a NotI site added to the 5′ end. Fragments were purified using the PCR purification kit (LTI).

A second PCR fragment (NdeI-rITR) was generated using the following primers:

35F7: 5′-GAA TGC TGG CTT CAG TTG TAA TC-3′ (SEQ ID NO: 18)

35R8: 5′-CGG AAT TCG CGG CCG CAT TTA AAT CAT CAT CAA TAA TAT ACC-3′ (SEQ ID NO: 19)

Amplification was done with pfx DNA polymerase (LTI) according to manufacturer's instructions but with 0.6 lM of both primers and 3% DMSO using 10 ngr. of wtAd35 DNA as template. The program was as follows: 3 min. at 94° C. and 5 cycles of 30 sec. at 94° C., 45 sec. at 40° C., 2 min.45 sec. at 68° C. followed by 25 cycles of 30 sec. at 94° C., 30 sec. at 60° C., 2 min.45 sec. at 68° C. To enable cloning in the TA-cloning vector PCR2.1, a last incubation with 1 unit superTaq polymerase for 10 min. at 72° C. was performed. The 1.6 kb amplified fragment ranging from nucleotides 33178 to the end of the right ITR of Ad35, was purified using the PCR purification kit (LTI).

Both purified PCR fragments were ligated into the PCR2.1 vector of the TA-cloning kit (Invitrogen) and transformed into STBL-2 competent cells (LTD). Clones containing the expected insert were sequenced to confirm correct amplification. Next, both fragments were excised from the vector by digestion with NotI and NdeI and purified from gel using the GENECLEAN kit (BIO 101, Inc.). Cosmid vector pWE15 (Clontech) was digested with NotI, dephosphorylated and also purified from gel. These three fragments were ligated and transformed into STBL2 competent cells (LTI). One of the correct clones that contained both PCR fragments was then digested with NdeI, and the linear fragment was purified from gel using the GENECLEAN kit. Ad35 wt DNA was digested with NdeI and the 26.6 kb fragment was purified from LMP gel using agarase enzyme (Roche) according to the manufacturer's instructions. These fragments were ligated together and packaged using λl phage packaging extracts (Stratagene) according to the manufacturer's protocol. After infection into STBL-2 cells, colonies were grown on plates and analysed for presence of the complete insert. One clone with the large fragment inserted in the correct orientation and having the correct restriction patterns after independent digestions with three enzymes (NcoI, PvuII and ScaI) was selected. This clone is designated pWE.Ad35.pIX-rITR. It contains the Ad35 sequences from bp 3401 to the end and is flanked by NotI sites (FIG. 8 ).

3) Generation of Ad35 Based Recombinant Viruses on PER.C6.

Wild type Ad35 virus can be grown on PER.C6 packaging cells to very high titers. However, whether the Ad5-E1 region that is present in PER.C6 is able to complement E1-deleted Ad35 recombinant viruses is unknown. To test this, PER.C6 cells were cotransfected with the above described adapter plasmid pAdApt35.LacZ and the large backbone fragment pWE.Ad35.pIX-rITR. First, pAdApt35.LacZ was digested with PacI and pWE.Ad35.pIX-rITR was digested with NotI. Without further purification, 4 μgr of each construct was mixed with DMEM (LTI) and transfected into PER.C6 cells, seeded at a density of 5×10 6 cells in a T25 flask the day before, using Lipofectamin (LTI) according to the manufacturer's instructions. As a positive control, 6μgr of PacI digested pWE.Ad35.pIX-rITR DNA was cotransfected with a 6.7 kb NheI fragment isolated from Ad35 wt DNA containing the left end of the viral genome including the E1 region. The next day, medium (DMEM with 10% FBS and 10 mM MgCl 2 ) was refreshed and cells were further incubated. At day 2 following the transfection, cells were trypsinized and transferred to T80 flasks. The positive control flask showed CPE at five days following transfection, showing that the pWE.Ad35.pIX-rITR construct is functional at least in the presence of Ad35-E1 proteins. The transfection with the Ad35 LacZ adapter plasmid and pWE.Ad35.pIX-rITR did not give rise to CPE. These cells were harvested in the medium at day 10 and freeze/thawed once to release virus from the cells. 4 ml of the harvested material was added to a T80 flask with PER.C6 cells (at 80% confluency) and incubated for another five days. This harvest/re-infection was repeated for two times but there was no evidence for virus associated CPE.

From this experiment, it seems that the Ad5-E1 proteins are not, or not well enough, capable of complementing Ad35 recombinant viruses, however, it may be that the sequence overlap of the adapter plasmid and the pWE.Ad35.pIX-rITR backbone plasmid is not large enough to efficiently recombine and give rise to a recombinant virus genome. The positive control transfection was done with a 6.7 kb left end fragment and therefore the sequence overlap was about 3.5 kb. The adapter plasmid and the pWE.Ad35.pIX-rITR fragment have a sequence overlap of 1.3 kb. To check whether the sequence overlap of 1.3 kb is too small for efficient homologous recombination, a cotransfection was done with PacI digested pWE.Ad35.pIX-rITR and a PCR fragment of Ad35 wt DNA generated with the above mentioned 35F1 and 35R4 using the same procedures as previously described herein. The PCR fragment thus contains left end sequences up to bp 4669 and, therefore, has the same overlap sequences with pWE.Ad35.pIX-rITR as the adapter plasmid pAdApt35.LacZ, but has Ad35 E1 sequences. Following PCR column purification, the DNA was digested with SalI to remove possible intact template sequences. A transfection with the digested PCR product alone served as a negative control. Four days after the transfection, CPE occurred in the cells transfected with the PCR product and the Ad35 pIX-rITR fragment, and not in the negative control. This result shows that a 1.3 kb overlapping sequence is sufficient to generate viruses in the presence of Ad35 E1 proteins. From these experiments, we conclude that the presence of at least one of the Ad3 5.E1 proteins is necessary to generate recombinant Ad35 based vectors from plasmid DNA on Ad5 complementing cell lines.

1) Construction of Ad35.E1 Expression Plasmids

Since Ad5-E1 proteins in PER.C6 are incapable of complementing Ad35 recombinant viruses efficiently, Ad35 E1 proteins have to be expressed in Ad5 complementing cells (e.g., PER.C6). Alternatively, a new packaging cell line expressing Ad35 E1 proteins has to be made, starting from either diploid primary human cells or established cell lines not expressing adenovirus E1 proteins. To address the first possibility, the Ad35 E1 region was cloned in expression plasmids as described below.

First, the Ad35 E1 region from bp 468 to bp 3400 was amplified from wtAd35 DNA using the following primer set:

35F11: 5′-GGG GTA CCG AAT TCT CGC TAG GGT ATT TAT ACC-3′ (SEQ ID NO: 20)

35F10: 5′-GCT CTA GAC CTG CAG GTT AGT CAG TTT CTT CTC CAC TG-3′ (SEQ ID NO: 21)

This PCR introduces a KpnI and EcoRI site at the 5′ end and an SbfI and XbaI site at the 3′ end.

Amplification on 5 ngr. template DNA was done with Pwo DNA polymerase (Roche) using the manufacturer's instructions, however, with both primers at a final concentration of 0.6 μM. The program was as follows: 2 min. at 94° C., 5 cycles of 30 sec. at 94° C., 30 sec. at 56° C. and 2 min. at 72° C., followed by 25 cycles of 94° C., 30 sec. at 60° C. and 2 min. at 72° C., followed by 10 min. at 72° C. PCR product was purified by a PCR purification kit (LTI) and digested with KpnI and XbaI. The digested PCR fragment was then ligated to the expression vector pRSVhbvNeo (see below) also digested with KpnI and XbaI. Ligations were transformed into competent STBL-2 cells (LTI) according to manufacturer's instructions and colonies were analysed for the correct insertion of Ad35E1 sequences into the polylinker in between the RSV promoter and HBV polyA.

The resulting clone was designated pRSV.Ad35-E1 (FIG. 9 ). The Ad35 sequences in pRSV.Ad35-E1 were checked by sequence analysis.

pRSVhbvNeo was generated as follows: pRc-RSV (Invitrogen) was digested with PvuII, dephosphorylated with TSAP enzyme (LTI), and the 3 kb vector fragment was isolated in low melting point agarose (LMP). Plasmid pPGKneopA (FIG. 10; described in International Patent Application WO96/35798) was digested with SspI completely to linearize the plasmid and facilitate partial digestion with PvuII. Following the partial digestion with PvuII, the resulting fragments were separated on a LMP agarose gel and the 2245 bp PvuII fragment, containing the PGK promoter, neomycin-resistance gene and HBVpolyA, was isolated. Both isolated fragments were ligated to give the expression vector pRSV-pNeo that now has the original SV40prom-neo-SV40polyA expression cassette replaced by a PGKprom-neo-HBVpolyA cassette (FIG. 11 ). This plasmid was further modified to replace the BGHpA with the HBVpA as follows: pRSVpNeo was linearised with ScaI and further digested with XbaI. The 1145 bp fragment, containing part of the Amp gene and the RSV promoter sequences and polylinker sequence, was isolated from gel using the GeneClean kit (Bio Inc. 101). Next, pRSVpNeo was linearised with ScaI and further digested with EcoRI partially and the 3704 bp fragment containing the PGKneo cassette and the vector sequences were isolated from gel as above. A third fragment, containing the HBV polyA sequence flanked by XbaI and EcoRI at the 5′ and 3′ end respectively, was then generated by PCR amplification on pRSVpNeo using the following primer set:

HBV-F: 5′-GGC TCT AGA GAT CCT TCG CGG GAC GTC-3′ (SEQ ID NO: 22) and

HBV-R: 5′-GGC GAA TTC ACT GCC TTC CAC CAA GC-3′ (SEQ ID NO: 23).

Amplification was done with Elongase enzyme (LTI) according to the manufacturer's instructions with the following conditions: 30 seconds at 94° C., then 5 cycles of 45 seconds at 94° C., 1 minute at 42° C. and 1 minute 68° C., followed by 30 cycles of 45 seconds at 94° C., 1 minute at 65° C. and 1 minute at 68° C., followed by 10 minutes at 68° C. The 625 bp PCR fragment was then purified using the Qiaquick PCR purification kit, digested with EcoRI and XbaI and purified from gel using the GENECLEAN kit. The three isolated fragments were ligated and transformed into DH5α competent cells (LTI) to give the construct pRSVhbvNeo (FIG. 12 ). In this construct, the transcription regulatory regions of the RSV expression cassette and the neomycin selection marker are modified to reduce overlap with adenoviral vectors that often contain CMV and SV40 transcription regulatory sequences.

2) Generation of Ad35 Recombinant Viruses on PER.C6 Cells Cotransfected with an Ad35-E1 Expression Construct.

PER.C6 cells were seeded at a density of 5×10 6 cells in a T25 flask and, the next day, transfected with a DNA mixture containing:

1 μg pAdApt35.LacZ digested with PacI

1 μg pAdApt35.LacZ digested with PacI

5 μg pRSV.Ad35E1 undigested

2 μg pWE.Ad35.pIX-aITR digested with NotI

Transfection was done using Lipofectamine according to the manufacturer's instructions. Five hours after addition of the transfection mixture to the cells, medium was removed and replaced by fresh medium. After two days, cells were transferred to T80 flasks and further cultured. One week post-transfection, 1 ml of the medium was added to A549 cells and, the following day, cells were stained for LacZ expression. Blue cells were clearly visible after two hours of staining indicating that recombinant LacZ expressing viruses were produced. The cells were further cultured, but no clear appearance of CPE was noted. However, after 12 days, clumps of cells appeared in the monolayer and 18 days following transfection, cells were detached. Cells and medium were then harvested, freeze-thawed once, and 1 ml of the crude lysate was used to infect PER.C6 cells in a 6-well plate. Two days after infection, cells were stained for LacZ activity. After two hours, 15% of the cells were stained blue. To test for the presence of wt and/or replicating competent viruses, A549 cells were infected with these virused and further cultures. No signs of CPE were stained blue. To test for the replication competent viruses. These experiments show that recombinant AdApt35.LacZ viruses were made on PER.C6 cells cotransfected with an Ad35-E1 expression construct.

Ad35 recombinant viruses escape neutralization in human serum containing neutralizing activity to Ad5 viruses.

The AdApt35.LacZ viruses were then used to investigate infection in the presence of serum that contains neutralizing activity to Ad5 viruses. Purified Ad5-based LacZ virus served as a positive control for NA. Hereto, PER.C6 cells were seeded in a wells plate at a density of 2×10 5 cells/well. The next day, a human serum sample with high neutralizing activity to Ad5 was diluted in culture medium in five steps of five times dilutions. 0.5 ml of diluted serum was then mixed with 4×10 6 virus particles AdApt5.LacZ virus in 0.5 ml medium and after 30 minutes of incubation at 37° C., 0.5 ml of the mixture was added to PER.C6 cells in duplicate. For the AdApt35.LacZ viruses, 0.5 ml of the diluted serum samples were mixed with 0.5 ml crude lysate containing AdApt35.LacZ virus and after incubation 0.5 ml of this mixture was added to PER.C6 cells in duplo. Virus samples incubated in medium without serum were used as positive controls for infection. After two hours of infection at 37° C., medium was added to reach a final volume of 1 ml and cells were further incubated. Two days after infection, cells were stained for LacZ activity. The results are shown in Table II. From these results, it is clear that whereas AdApt5.LasZ viruses are efficiently neutralized, AdApt35.LasZ viruses remain infectious irrespective of the presence of human serum. This proves that recombinant Ad3 5-based viruses escape neutralization in human sera that contain NA to Ad5-based viruses.

Generation of Cell Lines Capable of Complementing E1-deleted Ad35 Viruses Generation of pIG135 and pIG270

Construct pIG.E1A.E1B (FIG. 13) contains E1 region sequences of Ad5 corresponding to nucleotides 459 to 3510 of the wt Ad5 sequence (Genbank accession number M72360) operatively linked to the human phosphoglycerate kinase promoter (“PGK”) and the Hepatitis B Virus polyA sequences. The generation of this construct is described in International Patent Application No. WO97/00326. The E1 sequences of Ad5 were replaced by corresponding sequences of Ad35 as follows. pRSV.Ad35-E1 (described in Example 5) was digested with EcoRI and Sse8387I and the 3 kb fragment corresponding to the Ad35 E1 sequences was isolated from gel. Construct pIG.E1A.E1B was digested with Sse8387I completely and partially with EcoRI. The 4.2 kb fragment corresponding to vector sequences without the Ad5 E1 region but retaining the PGK promoter were separated from other fragments on LMP agarose gel and the correct band was excised from gel. Both obtained fragments were ligated resulting in pIG.Ad35-E1.

This vector was further modified to remove the LacZ sequences present in the pUC119 vector backbone. Hereto, the vector was digested with BsaAI and BstXI and the large fragment was isolated from gel. A double stranded oligo was prepared by annealing the following two oligos:

BB1: 5′-GTG CCT AGG CCA CGG GG-3′ (SEQ ID NO: 24) and

BB2: 5′-GTG GCC TAG GCA C-3′ (SEQ ID NO: 25).

Ligation of the oligo and the vector fragment resulted in construct pIG135 (FIG. 14 ). Correct insertion of the oligo restores the BsaAI and BstXI sites and introduces a unique AvrII site. Next, we introduced a unique site at the 3′ end of the Ad35-E1 expression cassette in pIG135. Hereto, the construct was digested with Sapi and the 3′ protruding ends were made blunt by treatment with T4 DNA polymerase. The thus treated linear plasmid was further digested with BsrGI and the large vector-containing fragment was isolated from gel. To restore the 3′ end of the HBVpolyA sequence and to introduce a unique site, a PCR fragment was generated using the following primers:

270F: 5′-CAC CTC TGC CTA ATC ATC TC-3′ (SEQ ID NO: 26) and

270R: 5′-GCT CTA GAA ATT CCA CTG CCT TCC ACC-3′ (SEQ ID NO: 27).

The PCR was performed on pIG.Ad35.E1 DNA using Pwo polymerase (Roche) according to the manufacturer's instructions. The obtained PCR product was digested with BsrGI and dephosphorylated using Tsap enzyme (LTI), the latter to prevent insert dimerization on the BsrGI site. The PCR fragment and the vector fragment were ligated to yield construct pIG270 (FIG. 15 ).

Ad35 E1 Sequences are Capable of Transforming Rat Primary Cells

New born WAG/RIJ rats were sacrificed at 1 week of gestation and kidneys were isolated. After careful removal of the capsule, kidneys were disintegrated into a single cell suspension by multiple rounds of incubation in trypsin/EDTA (LTI) at 37° C. and collection of floating cells in cold PBS containing 1% FBS. When most of the kidney was trypsinized all cells were re-suspended in DMEM supplemented with 10% FBS and filtered through a sterile cheesecloth. Baby Rat Kidney (BRK) cells obtained from one kidney were plated in 5 dishes (Greiner, 6 cm). When a confluency of 70-80% was reached, the cells were transfected with 1 or 5 μgr DNA/dish using the CaPO 4 precipitation kit (LTI) according to the manufacturer's instructions. The following constructs were used in separate transfections: pIG.E1A.E1B (expressing the Ad5-E1 region), pRSV.Ad35-E1, pIG.Ad35-E1 and pIG270 (expressing the Ad35-E1 region). Cells were incubated at 37° C., 5% CO 2 until foci of transformed cells appeared. Table III shows the number of foci that resulted from several transfection experiments using circular or linear DNA. As expected, the Ad5-E1 region efficiently transformed BRK cells. Foci also appeared in the Ad35-E1 transfected cell layer although with lower efficiency. The Ad35 transformed foci appeared at a later time point: ˜2 weeks post transfection compared with 7-10 days for Ad5-E1. These experiments clearly show that the E1 genes of the B group virus Ad35 are capable of transforming primary rodent cells. This proves the functionality of the Ad35-E1 expression constructs and confirms earlier findings of the transforming capacity of the B-group viruses Ad3 and Ad7 (Dijkema, 1979). To test whether the cells in the foci were really transformed a few foci were picked and expanded. From the 7 picked foci at least 5 turned out to grow as established cell lines.

Generation of New Packaging Cells Derivedfrom Primary Human Amniocytes

Amniotic fluid obtained after amniocentesis was centrifuged and cells were re-suspended in AmnioMax medium (LTI) and cultured in tissue culture flasks at 37° C. and 10% CO 2 . When cells were growing nicely (approximately one cell division/24 hrs.), the medium was replaced with a 1:1 mixture of AmnioMax complete medium and DMEM low glucose medium (LTI) supplemented with Glutamax I (end concentration 4 mM, LTI) and glucose (end concentration 4.5 gr/L, LTI) and 10% FBS (LTI). For transfection ˜5×10 5 cells were plated in 10 cm tissue culture dishes. The day after, cells were transfected with 20 lgr of circular pIG270/dish using the CaPO 4 transfection kit (LTI) according to manufacturer's instructions and cells were incubated overnight with the DNA precipitate. The following day, cells were washed 4 times with PBS to remove the precipitate and further incubated for over three weeks until foci of transformed cells appeared. Once a week the medium was replaced by fresh medium. Other transfection agents like, but not limited to, LipofectAmine (LTI) or PEI (Polyethylenimine, high molecular weight, water-free, Aldrich) were used. Of these three agents PEI reached the best transfection efficiency on primary human amniocytes: ˜1% blue cells 48 hrs. Following transfection of pAdApt35. LacZ.

Foci are isolated as follows. The medium is removed and replaced by PBS after which foci are isolated by gently scraping the cells using a 50-200 μl Gilson pipette with a disposable filter tip. Cells contained in ˜10 μl PBS were brought in a 96 well plate containing 15 μl trypsin/EDTA (LTI) and a single cell suspension was obtained by pipetting up and down and a short incubation at room temperature. After addition of 200 μl of the above described 1:1 mixture of AmnioMax complete medium and DMEM with supplements and 10% FBS, cells were further incubated. Clones that continued to grow were expanded and analysed their ability to complement growth of E1-deleted adenoviral vectors of different sub-groups, specifically ones derived from B-group viruses specifically from Ad35 or Ad11.

Generation of New Packaging Cell Lines from HER Cells

HER cells are isolated and cultured in DMEM medium supplemented with 10% FBS (LTI). The day before transfection, ˜5×10 5 cells are plated in 6 cm dishes and cultured overnight at 37° C. and 10% CO 2 . Transfection is done using the CaPO 4 precipitation kit (LTI) according to the manufacturer's instructions. Each dish is transfected with 8-10 μmgr pIG270DNA, either as a circular plasmid or as a purified fragment. To obtain the purified fragment, pIG270 was digested with AvrII and XbaI and the 4 kb fragment corresponding to the Ad35 E1 expression cassette was isolated from gel by agarase treatment (Roche). The following day, the precipitate is washed away carefully by four washes with sterile PBS. Then fresh medium is added and transfected cells are further cultured until foci of transformed cells appear. When large enough (>100 cells) foci are picked and brought into 96-wells as described above. Clones of transformed HER cells that continue to grow, are expanded and tested for their ability to complement growth of E1-deleted adenoviral vectors of different sub-groups specifically ones derived from B-group viruses specifically from Ad35 or Ad11.

New Packaging Cell Lines Derivedfrom PER.C6

As described in Example 5, it is possible to generate and grow Ad35 E1-deleted viruses on PER.C6 cells with cotransfection of an Ad35-E1 expression construct, e.g. pRSV.Ad35.E1. However, large-scale production of recombinant adenoviruses using this method is cumbersome because, for each amplification step, a transfection of the Ad35-E1 construct is needed. In addition, this method increases the risk of non-homologous recombination between the plasmid and the virus genome with high chances of generation of recombinant viruses that incorporate E1 sequences resulting in replication competent viruses. To avoid this, the expression of Ad35-E1 proteins in PER.C6 has to be mediated by integrated copies of the expression plasmid in the genome. Since PER.C6 cells are already transformed and express Ad5-E1 proteins, addition of extra Ad35-E1 expression may be toxic for the cells, however, it is not impossible to stably transfect transformed cells with E1 proteins since Ad5-E1 expressing A549 cells have been generated.

In an attempt to generate recombinant adenoviruses derived from subgroup B virus Ad7, Abrahamsen et al. (1997) were not able to generate E1-deleted viruses on 293 cells without contamination of wt Ad7. Viruses that were picked after plaque purification on 293-ORF6 cells (Brough et al., 1996) were shown to have incorporated Ad7 E1B sequences by non-homologous recombination. Thus, efficient propagation of Ad7 recombinant viruses proved possible only in the presence of Ad7-E1B expression and Ad5-E4-ORF6 expression. The E1B proteins are known to interact with cellular as well as viral proteins (Bridge et al., 1993; White, 1995). Possibly, the complex formed between the E1B 55K protein and E4-ORF6 which is necessary to increase mRNA export of viral proteins and to inhibit export of most cellular mRNAs, is critical and in some way serotype specific. The above experiments suggest that the E1A proteins of Ad5 are capable of complementing an Ad7-E1A deletion and that Ad7-E1B expression in adenovirus packaging cells on itself is not enough to generate a stable complementing cell line. To test whether one or both of the Ad35-E1B proteins is/are the limiting factor in efficient Ad35 vector propagation on PER.C6 cells, we have generated an Ad35 adapter plasmid that does contain the E1B promoter and E1B sequences but lacks the promoter and the coding region for E1A. Hereto, the left end of wtAd35 DNA was amplified using the primers 35F1 and 35R4 (both described in Example 4) with Pwo DNA polymerase (Roche) according to the manufacturer's instructions. The 4.6 kb PCR product was purified using the PCR purification kit (LTI) and digested with SnaBI and ApaI enzymes. The resulting 4.2 kb fragment was then purified from gel using the QIAExII kit (Qiagen). Next, pAdApt35IP1 (Example 4) was digested with SnaBI and ApaI and the 2.6 kb vector-containing fragment was isolated from gel using the GeneClean kit (BIO 101, Inc). Both isolated fragments were ligated to give pBr/Ad35.leftITR-pIX (FIG.16). Correct amplification during PCR was verified by a functionality test as follows: The DNA was digested with BstBI to liberate the Ad35 insert from vector sequences and 4 μg of this DNA was co-transfected with 4 μg of NotI digested pWE/Ad35.pIX-rITR (Example 4) into PER.C6 cells. The transfected cells were passaged to T80 flasks at day 2 and again two days later CPE had formed showing that the new pBr/Ad35.leftITR-pIX construct contains functional E1 sequences. The pBr/Ad35.leftITR-pIX construct was then further modified as follows. The DNA was digested with SnaBI and HindII and the 5′ HindII overhang was filled in using Klenow enzyme. Religation of the digested DNA and transformation into competent cells (LTI) gave construct pBr/Ad35leftITR-pIXΔDE1A (FIG. 17 ). This latter construct contains the left end 4.6 kb of Ad35 except for E1A sequences between bp 450 and 1341 (numbering according to wtAd35, FIG. 5) and thus lacks the E1A promoter and most of the E1A coding sequences. pBr/Ad35.leftITR-pIXΔDE1A was then digested with BstBI and 2 μg of this construct was co-transfected with 6 μmgr of NotI digested pWE/Ad35.pIX-rITR (Example4) into PER.C6 cells. One week following transfection full CPE had formed in the transfected flasks.

This experiment shows that the Ad35-E1A proteins are functionally complemented by Ad5-E1A expression in PER.C6 cells and that at least one of the Ad35-E1B proteins cannot be complemented by Ad5-E1 expression in PER.C6. It further shows that it is possible to make a complementing cell line for Ad35 E1-deleted viruses by expressing Ad35-E1B proteins in PER.C6. Stable expression of Ad35-E1B sequences from integrated copies in the genome of PER.C6 cells may be driven by the E1B promoter and terminated by a heterologous poly-adenylation signal like, but not limited to, the HBVpA. The heterologous pA signal is necessary to avoid overlap between the E1B insert and the recombinant vector, since the natural E1B termination is located in the pIX transcription unit that has to be present on the adenoviral vector. Alternatively, the E1B sequences may be driven by a heterologous promoter like, but not limited to the human PGK promoter or by an inducible promoter like, but not limited to the 7xtetO promoter (Gossen and Bujard, 1992). Also in these cases the transcription termination is mediated by a heterologous pA sequence, e.g. the HBV pA. The Ad35-E1B sequences at least comprise one of the coding regions of the E1B 21K and the E1B 55K proteins located between nucleotides 1611 and 3400 of the wt Ad35 sequence. The insert may also include (part of the) Ad35-E1B sequences between nucleotides 1550 and 1611 of the wt Ad35 sequence.

Ad35-based Viruses Deleted for E1A and E1B-21K Genes Efficiently Propagate on Ad5 Complementing Cell Lines

The generation of Ad35-based viruses that are deleted for E1A and retain the full E1B region is described in Example 6 of this application. Such viruses can be generated and propagated on the Ad5 complementing cell line PER.C6. The E1B region comprises partially overlapping coding sequences for the two major proteins 21K and 55K (Bos et al., 1981). Whereas during productive wt adenoviral infection both 21K and 55K are involved in counteracting the apoptose-inducing effects of E1A proteins, the E1B 55K protein has been suggested to have additional functions during the late phase of virus infection. These include the accumulation of viral mRNAs, the control of late viral gene expression and the shutoff of most host mRNAs at the level of mRNA transport (Babiss et al., 1984, 1985; Pilder et al., 1986). A complex formed between E1B-55K and the ORF6 protein encoded by the adenovirus early region 4 (Leppard and Shenk, 1989; Bridge and Ketner, 1990) exerts at least part of these functions.

To analyze which of the E1B proteins is required for propagation of Ad35-E1A deleted recombinant viruses on PER.C6 packaging cells, the E1B region in construct pBr.Ad35.leftITR-pIXΔE1A (see Example 6 and FIG. 17) was further deleted. A first construct, pBr.Ad35Δ21K, retains the full E1B-55K sequence and is deleted for E1A and E1B-21K. Hereto. pBr.Ad35.leftITR-pIXΔE1A was digested with NcoI and BspEI and the 5 KB vector fragment was isolated from agarose gel using the geneclean kit (BIO 101, Inc.) according to the manufacturer's instructions. Then a PCR fragment was generated with pBr.Ad35.leftITR-pIXΔE1A as template DNA using the following primers:

35D21: 5′-TTA GAT CCA TGG ATC CCG CAG ACT C-3′ (SEQ ID NO: 28) and

35B3: 5′-CCT CAG CCC CAT TTC CAG-3′ (SEQ ID NO: 29).

Amplifacation was done using Pwo DNA polymerase (Roche) according to manufacturer's recommendations with the addition of DMSO (fmal concentration 3%) in the reaction mixture. The PCR program was as follows: 94° C. for 2′, then 30 cycles of 94° C. for 30″, 58° C. for and 72° C. for 45″and a final step at 68° C. for 8′ to ensure blunt ends.

This PCR amplifies Ad35-E1B sequences from nucl. 1908 to 2528 (sequence Ad35, FIG. 5) and introduces an NcoI site at the start codon of the E1B-55K coding sequence (bold in primer 35D21). The 620 bp PCR fragment was purified using the PCR purification kit (Qiagen) and then digested with NcoI and BspEI, purified from agarose gel as above and ligated to the above described NcoI/BspEI digested vector fragment to give pBr.Ad35Δ21K (FIG. 18 ).

Since the coding regions of the 21K and 55K proteins overlap, it is only possible delete part of the 55K coding sequences while retaining 21K. Hereto, pBr.Ad35.leftITR-pIXΔTE1A was digested with BglII and the vector fragment was religated to give pBr.Ad35Δ55K1 (FIG. 19 ). This deletion removes E1B coding sequences from nucl. 2261 to 3330 (Ad35 sequence in FIG. 5 ). In this construct the N-terminal 115 amino acids are retained and become fused to 21 additional amino acids out of the proper reading frame before a stop codon is encountered. The 21K coding region is intact in construct pBr.Ad35Δ55K1.

A third construct that has a deletion of E1A, 21K and most of the 55K sequences was generated as follows. pBr.Ad35.leftITR-pIX (FIG. 16) was digested with SnaBI and MfeI (isoschizomer of MunI) and the 5′ overhang resulting from the MfeI digestion was filled in using Klenow enzyme. The 4.4 kb vector fragment was isolated from gel using the geneclean kit (Bio 101, Inc.) according to the manufacturer's instructions and religated to give construct pBr.Ad35AsM (FIG. 20 ). In this construct, the Ad35 sequences between nucl. 453 and 2804 are deleted thus 596 nucl. of the 3′ end of E1b-55K are retained. A further deletion of 55K sequences was made in construct pBr.Ad35AE1A. AE1B by digestion of pBr.Ad35.leftITR-pIX with SnaBI and BglII, Klenow treatment to fill in the BglII cohesive ends, and religation. FIG. 21 shows a schematic representation of the above mentioned constructs.

To test whether Ad35-based viruses can be generated with these constructs, each of the constructs was cotransfected with NotI digested pWE.Ad35pIX-rITR (see Example 4) onto PER.C6 cells. Hereto, the respective fragments were PCR amplified using primers 35F1 and 35R4 (see, Example 4). This PCR amplification was done since some of the constructs were difficult to isolate in large enough quantities. In this way, equal quality of the different adapter fragments was ensured. For the amplification Pwo DNA polymerase (Roche) was used according to the manufacturer's instructions but with DMSO (3% final concentration)added to the PCR mixture. Of each template ˜50 ng DNA was used. The conditions for the PCR were as follows: 94° C. for 2′, then 5 cycles of 94° C. for 30″, 48° C. for 45″ and 72° C. for 4′ followed by 25 cycles of 94° C. for 30″, 60° C. for 30″ and 72° C. for 4′ and a final step at 68° C. for 8′. PCR fragments were generated from pBr.Ad35leftITR-pIX, pBr.Ad35.leftITR-pIXΔE1A, pBr.Ad35Δ21K, pBr.Ad35Δ55K1, pBr.Ad35ASM and pBr.Ad35ΔE1AΔE1B. All fragments were using the PCR purification kit (Qiagen) according to manufacturer's instructions and final concentrations were estimated on EtBr stained agarose gel using the Eagle Eye II Still Video system and EagleSight software (Stratagene) with the SmartLadder molecular weight marker (Eurogentec) as reference.

PER.C6 cells were seeded at a density of 2.5×10 6 cells in a T25 culturing flask in DMEM containing 10% fetal calf serum (FCS) and 10 mM MgSO 4 and cultured in a humidified stove at 37° C., 10% CO 2 . The next day, 3 mg of each of the PCR fragments was cotransfected with 5 μgr NotI digested pWE.Ad35pIX-rITR using LipofectAmine (GIBCO, Life Technologies Inc.) according to the manufacturer's instructions. Two days after the transfection, all cells were passed to a T80 flask and further cultured. Cultures were then monitored for the appearance of CPE. In line with the outcome of previous experiments described in Examples 4 and 6, pBr.Ad35.leftITR-pIX and pBr.Ad35.leftITR-pIXΔE1A showed almost full CPE within one week following transfection. Of the fragments with different E1B deletions only pBr.Ad35Δ21K showed CPE at the same time as the above two fragments. Constructs pBr.Ad35Δ55K1, pBr.Ad35ΔSM and pEr.Ad35ΔE1AΔE1B did not give CPE at all, also not after harvesting by freeze-thawing and re-infection of the crude lysate onto fresh PER.C6 cells.

From these experiments, it can be concluded that Ad35-E1B-55K, and not E1B-21K, is necessary for generation and propagation of Ad35-based viruses on Ad5 complementing cell lines. Therefore, Ad35-based viruses having a deletion of the E1A and E1B 21K genes and having the E1B-55K gene or a functional fragment thereof, can be grown on Ad5 complementing cell lines. Alternatively, Ad35-based viruses can be grown on PER.C6 cells that stably express the full E1B region or the E1B-55K gene or a functional fragment thereof The Ad35 E1B-55K gene or functional parts thereof may be expressed from a heterologous promoter, like, but not limited to, the human PGK promoter, the human cytomegalovirus immediate early promoter (CMV), Rous sarcoma virus promoter, etc. and terminated by a heterologous poly adenylation sequence (pA), like but not limited to the hepatitis B virus poly adenylation sequence (HBVpA), the Simian Virus 40 poly adenylation sequence (SV40pA), etc. As non-limiting examples PER.C6 cells that express the Ad35-E1B region driven by the E1B promoter and HBVpA, PER.C6 cells that express the Ad35-E1B region driven by the human PGK promoter and HBVpA and PER.C6 cells that express a functional fragment of Ad35 E1B-55K driven by the human PGK promoter and HBVpA are described below.

Generation of pIG35BL and pIG35BS

We describe the generation of two expression constructs, pIG.35BS and pIG.35BL, that both carry the Ad35-E1B genes and a neomycin selection marker. The two constructs differ in the length of the fragment containing the E1B promoter. In 35BL the promoter fragment is longer and includes the 3′ end of the E1A region (103 nucl. coding sequence and pA). The E1B region is terminated by the HBVpolyA, the neor gene fragment on gel and isolated using the geneclean kit (BIO 101, Inc.). After religation of

pIG.35BL was made as follows. Construct pRSV.Ad35E1 (described in Example 5, FIG. 9) was digested with NruI and HindII and the protruding ends were filled in by Klenow treatment. The 7 kb vector fragment was separated from the smaller fragment on gel and isolated using the geneclean kit (BIO 101, Inc.). After religation of the DNA and transformation into competent STBL2 cells (Gibco, LTI) correct clones were isolated. pIG.35BL (FIG. 22) contains 273 nucl. upstream of the start site of the E1B-21K coding region.

pIG.35BS was made in the same way as pIG.35BL except that pRSV.Ad35E1 was digested with NruI and Hpal (both enzymes leave blunt ends), resulting in a shorter fragment upstream of the coding region of E1B-21K: 97 nucleotides.

To generate Ad35-E1B expressing cells, PER.C6 cells were seeded in 10 cm dishes at 1×10 6 cells/dish. Two days later cells were transfected with Scal linearised constructs. Four dishes were transfected with 1 and four with 2 μg DNA (total of 16 dishes; Lipofectamine (Gibco, LTI), no carrier DNA used) according to the manufacturer's instructions. The next day, transfected cells received G418-containing medium (0.75 mg/ml). Control transfections using LacZ expression constructs (2 μg) were stained after 48 hrs and showed a transfection efficiency of ˜25%. Four days following addition of selection medium untransfected cells started to die and again three days lateer clones were becoming visible. A week later, the first clones wew picked. Transfection with 1 μg resulted in less and also initially smaller clones (total ˜20 clones/dish against >50 clones/dish for the transfection with 2 μg DNA). The positive control transfection using 2 μg pcDNA3 (Invitrogen) resulted in ˜50 clones.

In total, 120 clones were picked and 107 were succesfully established (55 from pIG35BS and 52 from pIG35BL).

Generation of pIG35Bneo

pIG35Bneo is an Ad35-E1B expression plasmid from which the E1B with Scal and BamHI and protruding ends were filled in using Klenow enzyme. The 1070 bp fragment containing part of the Ampicilin gene and the RSV promoter was isolated from gel using the geneclean kit (BIO 101, Inc.). Next, pRSVhbvNeo was digested with To achieve this, construct pRSVhbv.Neo (described in Example 5, FIG. 12) was digested with ScaI and BamHI and protruding ends were filled in using Klenow enzyme. The 1070 were then ligated to give pRSVneo4 (FIG. 23 ). Construct pIG270 (FIG. 15, described in from gel using the geneclean kit (BIO 101, Inc.). Next, pRSVhbvNeo was digested with ScaI and EcoRI, blunted with Klenow and the 3.2 kb fragment containing the neo gene, above and religated to give pIG270delE1A. This construct was digested with AvrII and were then ligated to give pRSVneo4 (FIG. 23 ). Construct pIG270 (FIG. 15, described in Example 6) was then digested with EcoRI and NcoI and sticky ends were blunted with Next, pRSVneo4 was digested with BglII, blunted with Klenow enzyme, above and religated to give pIG270delE1A. This construct was digested with AvrII and XbaI and protruding ends were filled in using Klenow enzyme. The 2.9 kb fragment containing the hPGK promoter and Ad35.E1B sequences was isolated from gel as above. Next, pRSVneo4 was digested with BglII, blunted with Klenow enzyme, dephosphorylated and isolated from gel. The blunted AvrII/XbaI Ad35.E1B fragment was then ligated with the above prepared pRSVneo4 vector fragment and resulting clones site at nucl 2949 in FIG. 24 ). was choosen and named pIG35Bneo (FIG. 24 ). Detailed analysis of this clone revealed coding region of Ad35.E1B-21K. Hereto, both the E1A and E1B-21K sequences are first site at nucl 2949 in FIG. 24 ).

Generation of pIG35.55K

Construct pIG35.55K is similar to pIG35Bneo, however, it lacks the coding region of Ad35.E1B-21K. Hereto, both the E1A and E1B-21K sequences are first deleted from pIG270 as follows:

Construct pIG270 is digested with EcoRI, treated with Klenow enzyme and purified using the PCR purification kit (Qiagen) according to the manufacturer's instructions. The recovered DNA is then digested with Agel and the ˜5 kb vector fragment was isolated from gel as above. Next, Ad35 E1B-55K sequences sre amplified by PCR on pIG270 template DNA using the following primers:

35D21: 5′-TTA GAT CCA TGG ATC CCG CAG ACT C-3′ (SEQ ID NO: 30) and

35B3: 5′-CCT CAG CCC CAT TTC CAG-3′ (SEQ ID NO: 31).

The conditions used for the amplification are as previously descrided. The PCR fragment is purified using the PCR purification kit (Qiagen) and digested with NcoI. Following Klenow treatment to fill in the protruding ends, the DNA is forther sigested with AgeI and again column purified. The thus treated PRC fragment is then cloned into the above prepared EcoRI/AgeI digested vector fragment to give pIG270.ΔE1AΔ21K. The last steps to obtain pIG35.55K (FIG. 25) are equivalent to the last steps described above for the generation of pIG35Bneo starting with pIG270.ΔE1AΔ21K instead of pIG270.ΔE1A.

pIG35.55K is then linearized with Scal and used to transfect PER.C6 cells as described above. Clones that are resistent to G418 selection are picked and analysed for their ability to complement the propagation of E1-deleted Ad35 viruses.

New packaging cell lines for the generation and propagation of E1-deleted Ad35-based vectors derived from primary human cells.

The complete morphological transformation of primary cells by adenovirus E1 genes is the result of the combined activities of the proteins encoded by the E1A and E1B regions. The roles of the differint E1 proteins encoded by the transformation have been studied extensively (reviewed in Zantema and van der Ed, 1995; White, 1995, 1996). The adenovirus E1A proteins are essential for transformation of primary cells. The E1A proteins exert this effect through direct interaction with a number of cellular proteins that are involved in regulation of transcription. These include the pRB family of proteins, p300/CBP and TATA binding protein. In addition to this E1A increases the level of p53 protein in the cells. In the absence of adenovirus E1B activity the rise in p53 levels leads to the induction of apoptosis. Both proteins encoded by the E1B region counteract the induction of apoptosis although by different mechanisms. E1B-21K seems to counteract apoptosis in a manner similar to Bcl-2 via interaction with the effector proteins downstream in the apoptosis pathway (Han et al., 1996), whereas E1B-55K fuictions through direct interaction with p53. Importantly, the molecular mechanism by which the E1B-55K proteins of Ad2 and 5 (subgroup C) and Ad12 (subgroup A) function in the ability to neutralise p53 may differ. Whereas Ad5 E1B-55K binds p53 strongly and the complex localises to the cytoplasm, Ad12 E1B-55K binds p53 weakly and both proteins are localised in the nucleus (Zantema et al., 1985; Grand et al., 1999). Both proteins, however, inhibit the transactivation of other genes by p53 (Yew and Berk, 1992).

In rodent cells, the activity of E1A together with either E1B-21K or 55K is sufficient for full transformation although expression of both E1B proteins together is twice as efficient (Rao et al., 1992;). In human cells however, the activity of the E1B-55K protein seems to be more important given the observation that E1B-55K is indispensible for the establishment of transformed cells (Gallimore, 1986).

Example 6 hereof describes the generation of pIG270. In this construct the Ad35-E1 genes are expressed from the hPGK promoter and transcription is terminated by fragment of the Hepatitis B virus genome (Simonsen and Levinson, 1983; see also sequence described by Singer-Sam et al. (1984). The HBVpA is located in a BamHI-BglII fragment of the Hepatitis B virus genome (Simonsen and Levinson, 1983; see also Genbank HBV-AF090841). As mentioned before, the promoter and polyadenylation sequences of the E1 expression constructs described in this invention may be derived from other sources without departing from the invention. Also, other functional fragments of the hPGK and HbVpA sequences mentioned above may be used.

The functionality of pIG270 was shown by transformation of primary Baby Rat Kidney cells (BRK). Comparison with an equivalent Ad5-E1 expression construct learned that Ad35-E1 genes were less efficient in transforming these cells. The same has been found for the E1 genes of Ad12 (Bemards et al., 1982).

It is unclear which E1 protein(s) determine(s) the difference in subgroups. In the case of Ad12, transfection studies with chimeric E1A/E1B genes proteins (Bemards et al., 1982). The E1B-55K protein is shown infra to contain serotype-specific functions necessary for complementation of E1-deleted adenoviruses. If function, it is unlikely that these are involved in the transformation efficiency.

Analysis of functional domains in the Ad2 or Ad5 E1B-55K proteins using insertion mutants have revealed that functions related to viral replication, late protein synthesis and host protein shut-off are not confined to specific domains but are distributed along the protein (Yew et al., 1990). Using the same set of mutants, the domains important for interaction with p53 and E4-Orf6 were found to be more restricted. In addition to one common binding region (amino acids 262 to 326), p53 binding was affected by mutations at aa 180 and E4-Orf6 binding was affected by mutations at aa 143 (Yew and Berk, 1992; Rubenwolf et al., 1997).

Altogether these results indicate that it is difficult to separate the E1B-55K functions related to transformation (p53 binding) and late protein synthesis (Orf6

Here is described new E1 constructs that combine the high efficiency of transformation of one serotype with the serotype-specific complementation function of another serotype. These new constructs are used to transform primary human embryonic retinobladt cells and human amniocytes.

The Generation of pIG535, pIG635 and pIG735

Construct pIG535 contains the Ad5 E1A region and E1B promoter sequences linked to the Ad35 E1B sequences. Hereto, pIG270 (FIG. 15; see example 6) was digested with EcoRI and NcoI. The 5.3 kb vector fragment was then isolated from gel using the geneclean kit (BIO Inc. 101) according to the instructions of the manufacturer. Next, construct pIG.E1A.E1B (FIG. 13; see example 6) was digested with EcoRI and XbaI and the resulting 890 bp fragment was isolsted as above. A third fragment was generated by PRC amplification on pIG.E1A.E1B using the following primers:

5E1A-F: 5′-GAG ACG CCC GAC ATC ACC TG-3′ (SEQ ID NO: 33).

5E1B-R: 5′-CAA GCC TCC ATG GGG TCA GAT GTA AC-3′ (SEQ ID NO: 33).

The following PCR program was used: 94° C. for 2′ followed by 30 cycles of 94° C. for 30″, 60° C. for 30″and 72° C. for 1′, and a final step at 72° C. for 10′ to ensure blunt ends.

The resulting 400 bp PCR fragment was digested with XbaI and NcoI. After gel isolation as above, the three fragments were ligated snd transformed into STBL-2bacteris. One colony containing all three fragments in the correct order was selected and designated pIG535 (FIG. 26 ).

Construct pIG635 contains the AdS E1A and a chimeric Ad5-Ad35 E1B region such that the 21K sequence is essentially from Ad5 and linked to the Ad35 E1B-55K sequences as far as not overlapping with the 21K sequences. First, part of the AdS E1 sequences are amplified by PCR using pIG.E1A.E1B as template and the following primers:

5AK: 5′-GAG CGA AGA AAC CCA TCT GAG-3′ (SEQ ID NO: 34) and

2155R: 5′-GGT CCA GGC CGG CTC TCG G-3′ (SEQ ID NO: 35).

Amplification is accomplished with Pwo DNA polymerase (Roche) according to manufacturer's instruction. The 210 bp fragments is then purified from the primer sequences using the PCR purification kit (Qiagen).

A second PCR fragment is amplified from pIG270 DNA as described above but with the following primers:

2155F: 5′-CCG AGA GCC GGC CTG GAC-3′ (SEQ ID NO: 36) and

35F10: 5′-GCT CTA GAC CTG CAG GTT AGT CAG TTT CTT CTC CAC TG-3′ (SEQ ID NO: 37).

The 1.3 kb amplified fragment is purified as above and mixed in a 1:1 molar ratio with the first PCR fragments. The mixture is then first subjected to a PCR reaction without the solution of primers using Pwo DNA polymerase and the following program: 94° C. for 2′ and then 5 cycles of 94° C. for 30′, 60° C. for 30′, 72° C. for 90″. Subsequently, primers 5AK and 35F10 are added at 0.6 μM concentration after which a last PCR amplifies a 1.5 kb fragment. Hereto, temperature was set as follows: 94° C. for 2′, then 30 cycles of 94° C. for 30′, 60° C. for 30′ and 72° C. for 90′, followed by a final step at 72° C. for 10′ to ensure blunt ends. The resulting product is purified using the PCR purification kit (Qiagen) as above and digested with KpnI and SbfI (isoschixomer of Sse8387I). The digested DNA is then isolated from gel using the geneclean kit (BIO Inc., 101). Construct pIG.E1A.E1B is digested with Kpnl and SbfI and the vector-containing fragment is isolated from gel as above. This fragment is ligated to the above prepared final PCR product and the ligation mixture is transformed into STBL-2 cells (Gibco, LTI) according to manufacturer's instructions. This gives construct pIG635 (FIG. 27 ).

In construct pIG735, the border between Ad5 derived sequences and Ad35 derived sequences is located more 3′ than in construct pIG635. First, a BspEI site is introduced in the Ad5 sequence of construct pIg.E1A.E1B without changing the amino acid sequence. Hereto, Ad5 sequences from pIG.E1A.E1B are amplified using the following PCR primers: 5AK: (SEQ ID NO: 34) see above, and Bsp-R: 5′-GCT CTA GAC CTG CAG GGT AGC AAC AAT TCC GGA TAT TTA CAA G-3′ (SEQ ID NO: 38). Amplification is accomplished using Pwo DNA polymerase (Roche) according to the manufacturer's instruction. The following PCR program is used: 94° C. for 2′ followed by 30 cycles of 94° C. for 30′, 60° C. for 30″ and 72° C. for 30″, and a final step at 72° C. for 10′ to ensure blunt ends. The resulting 0.6 kb fragment is purified as above and digested with KpnI and SbfI and ligated to the above described KpnI/SbfI digested pIG.E1A.E1B vector fragment. Selection of colonies after transformation of STBL-2 bacteria (Life Techn. Inc.) gives construct pIG.E1Δ55K. pIG. E1Δ55K is then digested with SbfI and partially with BspEI. The 6.4 kb SbfI-partial BspEI digested vector fragment is then isolated from gel using the geneclean kit (BIO 101, Inc.). Next, pIG270 is digested with BspEI and SbfI and the resulting 915 bp fragment is isolated from gel as above. This fragment is then ligated to the above prepared SbfI/partial BspEI digested pIG.E1Δ55K vector fragment and transformed into STBL-2 competent cells. This gives construct pIG735 (FIG. 28 ). Clones are analysed by restriction enzyme digestion and sequencing to ensure correct ligation of the fragments. Constructs pIG535, pIG635 and pIG735 can be used to generate complementing cell lines from primary human cells as described in Example 6.

Legend to Table I:

All human adenoviruses used in the neutralization experiments were produced on PER.C6 cells (Fallaux et al., 1998) and purified on CsCl as described in example 1. The CCID50 is shown for the 44 viruses used in this study and reflects the HPLC column is shown. Virus particles/ml (VP/ml) were calculated from an Ad5 standard. The titer in the experiment (CCID50) was determined on PER.C6 cells as described in Example 1 by titrations performed in parallel with the neutralization experiment. The CCID50 is shown for the 44 viruses used in this study and reflects the dilution of the virus needed to obtain CPE in 50% of the wells after 5 days. The ratio of VP/CCID50 is depicted in log 10 and is a measurement of the infectively of the different batches on PER.C6 cells.

References

Abrahamsen, K., Kong, H-L., Mastrangeli, A., Brough, D., Lizonova, A., Crystal, R. G. and Falck-Pedersen, E. (1997). Construction of an adenoovirus type 7a E1A − vector. J. Virol. 71, 11, p8946-8951.

Babiss, L. E. and Ginsberg, H. S. (1984). Adenovirus type 5 early region 1b gene product is required for efficient shutoff of host protein synthesis. J. Virol. 50, p202-2122

Babiss, L. E., Ginsberg, H. S. and Darnell, J. J. (1985). Adenovirus E1B proteins are required for accumulation of late viral mRNA and for effects on cellular mRNA translation and transport. Mol. Cell. Biol. 5, p2552-2558.

Bemards,R., Houweling, A. Schrier, P. I., Bos, J. L. and van der Eb, A. j. (1982). Characterization of cells transformed by Ad5/Ad12 hybrid early region 1 plasmids. Virology 120, p422-432.

Bos, J. L., Polder, L. J., Bemards, R., Schrier, P., van den Elsen, P. J., van der Eb, A. J. and van Ormondt, H. (1981). The 2.2 kb mRNA of the E1B region of human adenovirus type 12 and 5 directs the synthesis of two major tumor antigens from different AUG triplets. Cell 12, p721-732.

Bridge, E. and Ketner, G. (1990). Interaction of adenoviral E4 and E1B products in late gene expression. Virology 174, p345-353.

Bridge, E., Medghalchi, S., Ubol, S., Leesong, M. and Ketner, G. (1993). Adenovirus early region 4 and viral DNA synthesis. Virology 193, p794-80.

Brough, D. E., Lizonova, A., Hsu, C., Kulesa, V. A. and Kovesdi, I. (1996). A gene tersfer vector-cell line system for complete functional complementation of of adenovirus early regions 1 and 4. J. Virol. 70, p6497-6501.

Fallaux, F. J., Kranenburg, O., Cramer, S. J., Houweling, A., van Ormondt, H., Hoeben, R. C. and van der Eb, A. J. (1996). Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum. Gene Ther. 7 (2), p215-222.

Fallaux, F. J., Bout, A., van der Velde, I., van den Wollenberg, D. J., Hehir, K. M., Keegan, J., Auger, C., Cramer, S. J., van Ormondt, H., van der Ed, A. J., Valerio, D. and Hoeben, R. C. (1998). New helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication competent adenoviruses. Hum. Gene Ther . 9, 1909-1917.

Gallimore, P. H., Grand, R. J. A. and Byrd, P. J. (1986). Transformation of human embryo retinoblasts with simian virus 40, adenovirus and ras oncogenes. AntiCancer Res. 6, p499-508.

Gossen, M., and H. Bujard (1992) Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89; 5547-5551.

Graham, F. O., Smiley, J., Russell, W. and Naim, R. (19770. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36, p59-72.

Grand, R. J. A., Parkhill, J., Szestak, T., Rookes, S. M., Roberts, S. and Gallimore, P. H. (1999). Definition of a major p53 binding site on Ad2E1B58K protein and a possible nuclear localization signal on the Ad12E1B54K protein. Oncogene 18, p955-965.

Han, J., Sabbatini, P., Perez, D., Rao, L., Modha, D. and White, E. (1996). The E1B 19K protein blocks apoptosis by interacting with and inhibiting the p53-inducible and death-promoting Bax protein. Genes Dev. 10 (4), p461-477.

Jochemsen, A. G., Peltenburg L. T., te Pas, M. F., de Wit, C. M., Bos, J. L. and van der Eb, A. J. (1987). Activation of adenovirus 5 E1A transcription by region E1B in transformed primary rat cells. EMBO J. 6 (11), p3399-3405.

Leppard, K. N. and Shenk, T. (1989). The adenovirus E1B 55kd protein influences mRNA tranport via an intranuclear effect on RNA metabolism. EMBO J. 8, p2329-2336.

Pilder, S., Moore, M., Logan, J. and Shenk, T. (1986). The adenovirus E1B 55K transforming polyeptide modulates transport or cytopladmic stabilization of viral and host cell mRNAs. Mol. Cell. Biol. 6, p470-476.

Rao, L., Debbas, M., Sabbatini, P., Hockenbery, D., Korsmeyer, S. and White, E. (1992). The adenovirus E1A proteins induce apoptosis, which is inhibited by the E1B 19-kDa and Bcl-2 proteins. Proc. Natl. Acad. Sci. USA 89, p7742-7746.

Rubenwolf, S., Schütt, H., Nevels, M., Wolf, H. and Dobner, T. (1997). Structural analysis of the adenovirus type 5 E1B 55-kilodalton-E4orf6 protein complex. J. Virol. 71, p1115-1123.

Singer-Sam, J., Keith, D. H., Tani, K., Simmer, R. L., Shively, L., Lindsay, S., Yoshida, A. and Riggs, A. D. (1984). Sequence of the promoter region of the gene for human X-linked 3-phosphoglycerate kinase. Gene 32 (3), p409-417.

White, E. and Cipriani, R. (1990). Role of adenovirus E1B proteins in trandformation: Altered organization of intermediate filaments if transformed cells that express the 19-kilodalton protein. Mol. Cell. Biol. 10, p120-130.

White, E. (1995). Regulation of p53-dependent apoptosis by E1A and E1B. In: The molecular repertoire of adenoviruses III. Eds. Doerfler, W. and Böhm, P. Springer-Verlag Berlin Heidelberg 1995, p33-58.

White, E. (1996). Life, death, and the pursuit of apoptosis. Genes Dev. 10 (1), p1-15.

Yew, P. R., Kao, C. C. and Berk, A. J. (1990). Dissection of functional domains in the adenovirus 2 early region 1B 55K polypeptide by suppressor-linker insertional mutagenesis. Virology 179, p795-805.

Yew, P. R. and Berk, A. J. (1992). Inhibition of p53 transactivation required for transformation by adenovirus early region 1B protein. Nature 357, p82-85.

Simonsen, C. C. and Levinson, A.D. (1983). Analysis of processing and polyadenylation signal of the hepatites B virus surface antigen gene by using simian virus 40-hepatitis B virus chimeric plasmids. Mol. and Cell. Biol. 3 (12), p2250-2258.

Zantema, A., Fransen, J. A., Davis, O. A., Ramaekers, F. C., Vooijs, G. P., DeLeys, B. and van der Eb, A. J. (1985). Localization of the E1B proteins of adenovirus 5 in transformed cells, as revealed by interaction with monoclonal antibodies. Virology 142, p44-58.

Zantema, A. and van der Eb, A. J. (1995). Modulation of gene expression by adenovirus transformation, In: The moleculat repertoire of adenoviruses III. Eds. Doerfler, W. and Böhm, P. Springer-Verlag Berlin Heidelberg 1995, p1-23.

39

1

14