Vectors with modified initiation codon for the translation of AAV-Rep78 useful for production of AAV

Abstract

The present invention relates nucleic acid constructs for the production of recombinant parvoviral (e.g. adeno-associated viral) vectors in insect cells, to insect cells comprising such constructs and to methods wherein the cells are used to produce recombinant parvoviral virions. The insect cells preferably comprise a first nucleotide sequence encoding the parvoviral rep proteins whereby the initiation codon for translation of the parvoviral Rep78 protein is a suboptimal initiation codon that effects partial exon skipping upon expression in insect cells. The insect cell further comprises a second nucleotide sequence comprising at least one parvoviral (AAV) inverted terminal repeat (ITR) nucleotide sequence and a third nucleotide sequence comprising a sequences coding for the parvoviral capsid proteins.

Claims

1. A nucleic acid construct encoding parvoviral Rep protein Rep78 and Rep52, the nucleic acid construct comprising: (a) a biscistronic nucleic acid encoding a parvovirus Rep78 protein and a parvovirus Rep52 protein, the biscistronic nucleic acid comprising a nucleic acid sequence encoding the Rep78 protein that overlaps with a nucleic acid sequence encoding the Rep52 protein, such that the nucleic acid sequence encoding the parvoviral Rep52 protein is not repeated in the biscistronic nucleic acid; (b) a first expression control sequence that is active in an insect cell operably linked to the sequence encoding Rep78 protein; (c) a second expression control sequence that is active in an insect cell operably linked to the sequence encoding Rep52 protein.

2. The nucleic acid construct of claim 1, wherein both the parvoviral Rep78 protein and the parvoviral Rep52 protein are produced when the nucleic acid construct is expressed in the insect cell.

3. The nucleic acid construct of claim 2, wherein the parvoviral Rep78 protein and the parvoviral Rep52 protein are produced in a molar ratio in the range of 1:10 to 10:1.

4. The nucleic acid construct of claim 1, wherein the first expression control sequence comprises an initiation codon.

5. The nucleic acid construct of claim 4, wherein the initiation codon is a suboptimal initiation codon.

6. The nucleic acid construct of claim 5, wherein the suboptimal initiation codon is selected from the group consisting of ACG, TTG, CTG, and GTG.

7. The nucleic acid construct of claim 1, wherein the first expression control sequence comprises SEQ ID NO: 7.

8. The nucleic acid construct of claim 1, wherein the first and second expression control sequences comprise a promoter.

9. The nucleic acid construct of claim 8, wherein the promoter is selected from the group consisting of polyhedron promoter, p10 promoter, p35 promoter, IE-1 promoter, and delta-IE1 promoter.

10. The nucleic acid construct of claim 1, wherein the second expression control sequence comprises an internal ribosome entry site (IRES).

11. The nucleic acid construct of claim 1 further comprising a sequence encoding a Cap protein operably linked to an expression control sequence for expressing the Cap protein in an insect cell.

12. The nucleic acid construct of claim 1, wherein the nucleic acid construct comprises a second nucleotide sequence comprising at least one parvoviral inverted terminal repeat (ITR) sequence and a third nucleotide sequence comprising at least one parvoviral Cap protein encoding sequence operably linked to an expression control sequence for expressing the at least one Cap protein in an insect cell.

13. The nucleic acid construct of claim 1, wherein the nucleic acid construct comprises a polyadenylation sequence at its 3-end.

14. A recombinant viral vector comprising the nucleic acid construction of claim 1.

15. A baculovirus vector comprising the nucleic acid construction of claim 1.

16. An insect cell comprising the nucleic acid construction of claim 1.

17. The insect cell of claim 16, wherein the insect cell is selected from the group consisting of Se301, Seizd2109, Seucr1, sf9, Sf900+, Sf21, Bti-Tn-5b1-4, Mg-1, Tn368, Hzam1, Ha2302, Hz2e5, High five and Express+.

18. The insect cell of claim 16, wherein the insect cell further comprises a nucleotide sequence comprising at least one parvoviral inverted terminal repeat (ITR) sequence and a nucleotide sequence comprising at least one parvoviral Cap protein encoding sequence operably linked to an expression control sequence for expressing the at least one Cap protein in the insect cell.

19. A method for producing recombinant parvovirus Rep78 and Rep52 proteins in an insect cell, comprising transfecting an insect cell with the nucleic acid construct of claim 1, wherein both Rep78 protein and Rep52 protein are produced in the insect cell.

20. The method of claim 19, wherein the insect cell is transfected with a baculovirus vector comprising the nucleic acid construct.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIG. 1: A) Organisation of Rep expression in the wild type AAV genome. The Rep78 and Rep 52 genes are expressed from respectively the P5 and P19 promoter. Expression of Rep68 and Rep40 (which are the spliced variants of resp. Rep78 and Rep52) are not shown. Both expression units contain a ATG-initiation site.

(2) B) The construct of the invention has the Rep ORF under the control of a single promoter (e.g. the polyhedron (PolH) promoter). This promoter drives the expression of both Rep78 and Rep52 because the Rep78 initiation codon ATG is converted to the alternate ACG initiation codon and partially skipped by the ribosome.

(3) C) The original construct by Urabe et al. (2002, supra) drives Rep78 and Rep52 independently from two different promoters (resp. MEI and polH).

(4) FIG. 2: Western blot analysis of Rep proteins expressed from recombinant baculovirus that was passaged 5 times on insect cells. The original baculovirus designed by Urabe et al., 2002 (original REP/Bae-to-Bae) results in a slow decrease of Rep78/52 expression over 5 passages. The expression unit for Rep78 and 52 designed by Urabe et al., 2002 inserted in baculovirus backbone PSC (original REP/PSC) also results in a decrease of Rep78/52 expression following passaging on insect cells. However, the baculovirus with the REP expression unit containing the ACG initiation codon in the PSC backbone (REP-ACG/PSC) results in stable expression of Rep78/52 over at least 5 passages. Western blot analysis was performed as described in Example 1.1.3.

(5) FIG. 3: Results of Table 1 plotted in a graph.

(6) FIG. 4: Comparison of the stabilities of various rAAV constructs in insect cells._rAAV production in SF+ cells was performed as described above in Example 1. For all productions the ITR containing baculovirus and the capsid gene containing baculovirus were identical, the passage number was the same as the Rep gene containing baculoviruses. 4 different Rep gene containing baculoviruses were used: 1) The REP-ACG/PSC, 2) SLR: the original construct by Urabe et al. (2002, supra), 3) Rep52+Rep78(B2B): Two separate Bae-to-Bae baculoviruses, one containing the Rep 78 gene and the other one containing the Rep 52 gene. 4) Rep52+Rep78(PSC): Two separate protein sciences baculoviruses one containing the Rep 78 gene and the other one containing the Rep 52 gene.

(7) FIG. 5: Stability of the REP-ACG/PSC baculovirus constructs up to passage 8.

(8) rAAV productions in SF+ cells were performed as described in Example 1.

(9) FIG. 6: Comparison of the effect of passage effect on rep protein expression of the original construct from Urabe et al. (2002, supra) with a REP-ACG/PSC construct in accordance with the invention. The baculovims passages and the western blot were done as described in Example 1. During a normal passage of the rep baculovimses, samples were taken at 40 hours after addition of the baculovimses to the SF cells and western blot was performed.

EXAMPLES

Example 1: Rep

(10) 1.1. Constructs

(11) 1.1.1 Baculovirus Plasmid Construction

(12) In order to express Rep78 and Rep52 from a sole bicistronic messenger RNA, the ATG initiation codon of Rep78 situated on the expression vector pFastBacDualSLR (Urabe et al., 2002, supra) was converted to ACG. The upstream primer used was:

(13) TABLE-US-00001 (SEQIDNO.8) BamHI 5-cgcggatcctgttaagACGGCGGGGTTTTACGAGATTGTGATTAAGG TC-3 PRIMERSEQUENCEforward

(14) The 3-primer that was used in the PCR reaction was flanking the REP78 gene and contains a XbaI site (TCTAGA):

(15) TABLE-US-00002 (SEQIDNO.9) XbaI 5-AGGCTCTAGATTCGAAAGCGGCCCG-3 PRIMERSEQUENCEreverse

(16) The sequence between the above-mentioned primer set was amplified by PCR (reaction volume 50 l; 1 Pfx Amp. Buffer, 0.3 mM dNTP's, 1 mM MgSO4, 150 mM primer forw., 150 mM primer rev., 2 enhancer solution, template 50 ng (pFastBacDualSLR), 1 U Platinum Pfx (Invitrogen, Carlsbad, Calif., USA) using the following protocol: 1 cycle of 95 C., 5 min; 35 cycles of 95 C., 15 sec; 55 C., 30 sec; 72 C., 2 min; 1 cycle of 72 C., 10 min; 4 C., for ever). The PCR product was cloned in PCR-blunt II-TOPO using the Zero Blunt TOPO PCR cloning kit (Invitrogen). The Rep78 was subcloned into pFastBacDual (Invitrogen) using the restriction sites SpeI and XbaI. The mutated Rep expression cassette was finally cloned (using restriction enzymes BstZl 7I and AvrII) into the baculovims expression construct (cut open with EcoRV and XbaI) pPSClO (Protein Sciences Corporation, Meriden, Conn., USA). The sequence analysis of the construct was verified by Baseclear, Leiden, the Netherlands.

(17) 1.1.2 Recombinant Baculovirus Production

(18) Recombinant baculovimses derived from the Autographa californica nuclear polyhydrosis vims (AcNPV) were produced using the GeneXpress BaculoKIT (Protein Sciences Corporation). Transfection was performed as follows: in a round bottom 14 ml tube 200 l GRACE medium was mixed with 6 l cellfectine (Invitrogen), and in a eppendorf tube 200 l GRACE medium was mixed with 50 l viral DNA (protein sciences) and 2 g transfer plasmid (REP). The contents from the eppendorf tube were added to the tube and mixed carefully. After an incubation period of 30 minutes at RT 1,300 l GRACE was added to the transfection mix.

(19) Insect cells in a T25 flask were washed with GRACE medium and the transfection mixture was added dropwise to the cell layer. After an incubation of 6 hours at 28 C. SF900II serum supplemented with 10% FBS was added carefully and the T25 flask was put in a 28 C. stove for 5 days after which the recombinant baculovirus was harvested.

(20) 1.1.3 Western Blot Analysis

(21) Insect cells (SF+) were infected with baculovirus-REP. At 16, 40, and 64 hours post-infection cells a sample was taken and cells were lysed by adding 0.1V 10TRIS lysis buffer (1.5M NaCl, 0.5M TRIS, 0.01M MgCl, 1% TRITON X-100, pH8.5, filter sterilised) and incubated at 28 C. for 30 minutes in a shaker (Innova 44, New Brunswick). Free DNA and RNA was degraded by incubation with benzonase at 37 C. for 30 minutes. Cell lysate was centrifuged (1,900g; 15 min; 4 C.). NuPAGE LDS sample buffer (4, Invitrogen) was added to a sample of the supernatant and was loaded onto a 4-12% Bis-Tris gel (120V). Proteins were blotted onto a PVDF membrane (BioRad) for 30 minutes, 10V (Semidry blotting). Western immunochemistry was performed by blocking the membrane with Superblock-PBS blocking buffer (PIERCE) and subsequent incubation with mouse anti-Rep (303.9, Progen, Germany; dilution 1:50) and rabbit anti-mouse-HRP (DAKO, dilution 1:500). The Rep-proteins were visualized by chemoluminescent staining with lumi-light plus Western-blotting substrate (Roche).

(22) 1.2 Results

(23) The performance of the newly designed Rep-construct of the invention (REP-ACG/PSC) was compared with the original Rep constructs in both I) PSC baculovirus backbone and in 2) Bae-to-Bae baculovirus backbone (Urabe et al., 2002). All three constructs were serially passaged until passage 5. AAVI-LPL production experiments were performed using the passage 2, 3, 4 and 5 Rep-constructs in combination with an AAV-LPL and a AAV-Cap recombinant baculovirus of respectively passage 2, 3, 4 and 5 (AAV-LPL and AAV-Cap recombinant Baculovirus used here are described below in Example 2). AAVI-LPL production yields were determined by qPCR and are shown in Table 1. The original baculovirus designed by Urabe et al., 2002 (original REP/Bae-to-Bae) results in a fast decrease of AAV production over 5 passages. The expression unit for Rep designed by Urabe et al., 2002 inserted in baculovirus backbone PSC (original REP IPSC) also results in a decrease of AAV production following passaging on insect cells. However, the baculovims with the REP expression unit containing the ACG initiation codon in the PSC backbone (REP-ACG IPSC) results in stable AAV production over at least 5 passages. Therefore, reproducible production yields of AAV-LPL over several passages (e.g. 2 to 5) were only obtained using baculovimses containing the REP-ACG construct.

(24) TABLE-US-00003 TABLE 1 Production of rAAV virions using the baculovirus constructs of several passages: original REP/ REP-ACG/PSC original REP/Bae-to- passage PSC g/ml g/ml Bae g/ml 2 5.38E+09 3.04E+09 3.62E+10 3 9.57E+09 4.77E+09 7.28E+09 4 1.66E+09 7.81E+09 7.59E+08 5 7.35E+08 9.90E+09 2.03E+08 Sf9 cells were infected with three recombinant baculoviruses encoding a LPL-vector unit of passage 2, 3, 4 or 5, a Rep-expression unit of passage 2, 3, 4 or 5 and a Cap-expression unit of passage 2, 3, 4 or 5. After three days cells were harvested and AAV yields (vector genomes per ml; vg/ml) were determined by qPCR.

(25) TABLE-US-00004 TABLE 2 Q-PCR performed on the various Bae-Rep constructs following passaging on insect cells (Passage 2-5). titer (gc's/ml) Ratio Ratio ORF Rep78 Rep52 ORF/Rep78 ORF/Rep52 original REP/Bac-to-Bac 1.4E+09 2.2E+08 2.4E+08 6.42 5.82 original REP/Bac-to-Bac 6.4E+08 5.6E+07 5.0E+07 11.43 12.93 original REP/Bac-to-Bac 2.1E+09 7.1E+07 6.5E+07 29.47 32.02 original REP/Bac-to-Bac 1.7E+09 3.2E+07 2.5E+07 53.68 69.67 REP-ACG/PSC (C4) P2 3.0E+09 2.7E+09 2.9E+09 1.11 1.04 REP-ACG/PSC (C4) P3 2.3E+09 2.0E+09 2.2E+09 1.11 1.05 REP-ACG/PSC (C4) P4 2.5E+09 2.2E+09 2.3E+09 1.13 1.08 REP-ACG/PSC (C4) P5 2.7E+09 2.1E+09 2.5E+09 1.26 1.07 REP-ACG/PSC (A3) P2 2.5E+09 2.2E+09 2.5E+09 1.18 1.00 REP-ACG/PSC (A3) P3 4.2E+09 3.9E+09 4.0E+09 1.08 1.04 REP-ACG/PSC (A3) P4 2.7E+09 2.4E+09 2.5E+09 1.10 1.05 REP-ACG/PSC (A3) P5 1.5E+09 1.5E+09 1.5E+09 1.03 0.98 original REP/Bae-to-Bae 1.0E+09 1.1E+09 1.1E+09 0.95 0.87 original REP/Bae-to-Bae 7.1E+08 6.7E+08 8.1E+08 1.07 0.88 original REP/Bae-to-Bae 1.3E+08 1.1E+08 1.3E+08 1.18 1.03 original REP/Bae-to-Bae 1.3E+08 5.3E+07 6.9E+07 2.34 1.82

(26) Table 2 shows the results of a quantitative PCR (Q-PCR) assay that was designed for the Rep-expression unit in the recombinant baculoviruses and for a flanking baculovirus ORF (gene copies per ml; gc's/ml). The ratio between the Q-PCR values determines the presence of deletions in the Rep-baculovirus. A ratio of 1 theoretically means that all baculoviruses in the batch contain a recombinant Rep78 or 52-sequence. The original baculovirus designed by Urabe et al., 2002 (original REP/Bae-to-Bae) shows significant amounts of the recombinant baculovirus at passage 5 have deletions in the Rep sequences. The expression unit for Rep78 and 52 designed by Urabe et al., 2002 inserted in baculovirus backbone PSC (original REP/PSC) shows a very early and dramatic loss of recombinant baculovirus. However, the baculovirus with the REP expression unit containing the ACG initiation codon in the PSC backbone (REP-ACG/PSC) (clone C4 and A3) show stable recombinant baculoviruses over at least 5 passages.

Example 2: Cap Constructs

(27) 2.1.1 Baculovims Plasmid Construction

(28) In order to express VP1, 2, 3 from a sole polycistronic messenger RNA, the baculovims-AAV-Cap construct was designed as described by (Urabe et al., 2002, supra). Briefly, the ATG initiation codon of VP1 was mutated to ACG. A potential ATG initiation codon at position 11 has been changed to ACG. The splice acceptor site downstream of the VP1 initiation codon was destroyed (mutation at position 21 and 24). The mutated Cap expression cassette was cloned into a baculovims expression construct; pFastBacDual (pFBDAAVIVPmI I) with BamHI/StuI restriction sites. This plasmid (pFBDAAVIVPmI I) was the starting material for introduction of alternate initiation codons for VP1. The forward primer used by Urabe et al. (2002, supra) in order to introduce the mentioned mutations was:

(29) TABLE-US-00005 (SEQIDNO.1) BamHI1112124 5-cgcggatcctgttaagACGGCTGCCGACGGTTATCTACCCGATT GGCTC-3

(30) The following forward primers were used to make the expression constructs using pFBDAAVIVPmI I (Urabe et al., 2002, supra) as starting material:

(31) TABLE-US-00006 (SEQIDNO.2) 5-cgcggatcctgttaagTTGGCTGCCGACGGTTATCTACCCGATTG CTC-3 (SEQIDNO.3) 5-cgcggatcctgttaagATTGCTGCCGACGGTTATCTACCCGATTGG GCTC-3 (SEQIDNO.4) 5-cgcggatcctgttaagGTGGCTGCCGACGGTTATCTACCCGATTGG CTC-3 (SEQIDNO.5) 5-cgcggatcctgttaagCTGGCTGCCGACGGTTATCTACCCGATTGG CTC-3

(32) The backward-primer that was used in the PCR reactions with the above forward primers was directed to position 230 bp downstream of the VP1 initiation codon and contains a unique Stu I site (AGGCCT).

(33) TABLE-US-00007 (SEQIDNO.6) 5-GTCGTAGGCCTTGTCGTGCTCGAGGGCCGC-3

(34) Fragments were amplified with the above-mentioned sets of forward and backward primer pairs by PCR. Following digestion of PCR products with BamHI and StuI the PCR products were subcloned into the BamHI/StuI sites of pFBDAAVIvpmI I resulting in the various to be tested baculovirus-AAV-Cap constructs. DNA constructs were verified by sequence analysis at Baseclear, Leiden, the Netherlands.

(35) 2.1.2 Recombinant Baculovirus Production

(36) Recombinant baculoviruses derived from the Autographa californica nuclear polyhydrosis virus (AcNPV) were produced using the Bae-to-Bae baculovirus expression system (Invitrogen). rBac-Cap was amplified by infecting 210.sup.6 Sf9 cells per ml at an moi of 0.1. Three days after infection the cells were spun down and the supernatant containing the virus recovered.

(37) 2.1.3 Recombinant AAV Production

(38) rAAV batches were produced using three recombinant baculoviruses according to Urabe et al., 2002. However, for this study one baculovirus harboured an expression construct for the LPL.sup.s447x-transgene. The therapeutically active agent expressed from the transgene is a naturally occurring variant of human lipoprotein lipase, a single chain polypeptide of 448 amino acids.

(39) The LPL.sup.s447x variant has a deletion of two amino acids at the C-terminus of the protein. The second baculovirus harboured an expression construct for the AAV replication genes, Rep 78 and Rep 52. The third baculovirus harboured the AAVI capsid sequence with either an ACG or a TTG, CTG, GTG initiation codon for VP1.

(40) Mammalian-rAAV batches produced with the plasmid-transfection system were produced according to Grimm et al., 1998 (Novel tools for production and purification of recombinant adeno-associated virus vectors. Hum Gene Ther. 1998 Dec. 10; 9(18):2745-60).

(41) 2.1.3 Western Blot Analysis

(42) Insect cells were infected with baculovirus-Cap. At three days post-infection cells were centrifuged (3,000 g; 15 min). The supernatant was filtered through a 0.22 um Millex filter. NuPAGE LDS sample buffer (Invitrogen) was added to a sample of the supernatant and was loaded onto a 4-12% Bis-Tris gel. The gel was run at 100V. Proteins were blotted onto a nitrocellulose membrane (BioRad) for 1 hr, 100V, 350 mA. Western immunochemistry was performed by blocking the membrane with 1% marvel, dried skimmed milk and subsequently incubation with mouse anti-Cap (BI from Progen, Germany; dilution 1:50) and rabbit anti-mouse-HRP (DAKO, dilution 1:100). VP1, 2 and 3 were visualized by chemoluminescent staining with lumi-light plus Western-blotting substrate (Roche).

(43) 2.1.4 Biochemical Measurements

(44) Human LPL.sup.s.sup.447.sup.x activity was assayed as previously described using a radioactive trioleoylglycerol emulsion substrate (Nilsson-Ehle and Scholtz, 1976). Human LPLs.sup.447x immunoreactive mass was assayed using a sandwich ELISA with chicken IgY and mouse 5D2 anti-hLPL antibodies (Liu et al., 2000). Plasma triglyceride levels were measured by using commercial kits following manufacturer protocols (Boehringer Mannheim, #450032).

(45) 2.2 Results

(46) 2.2.1 Construction of Recombinant Baculovirus

(47) In order to introduce different alternate initiation codons for VP1 expression in the baculovirus plasmid designed by Urabe et al. (2002, supra) a series of upstream primers were designed containing a BamHI restriction site and either a TTG, ATT, GTG or CTG codon in place of the ACG initiation codon of VP1.PCR using these primers in combination with a downstream primer containing a StuI site resulted in amplified fragments that were subcloned into the BamHVStuI site of pFBDVPmI 1 (Bae-Cap). The resulting baculovims plasmids were used for the preparation of recombinant baculovimses using the Bae-to-Bae baculovims expression system. The prepared recombinant baculovimses were infected on insect cells in order to produce AAV capsids. At three days following infection viral protein expression of the different baculovims batches were determined on Western blots. From the Western blots it became clear that the baculovims construct containing the TTG initiation codon for VP1 expressed this protein to a higher level compared to the previously used ACG initiation codon. The ratio between VP1 and VP2 using the TTG codon was found to be 1:1 which is similar to what is reported for wild type AAV (not shown).

(48) 2.2.2 Infection of rAAV Batches on Cells in Culture

(49) In order to investigate the infectivity of the AAV capsids derived from recombinant baculovimses with the TTG initiation codon rAAV was generated. Also a rAAV batch was generated by plasmid transfection on mammalian HEK293 cells. A vector genome titer of both rAAV batches was determined by qPCR. This titer was used to infect HEK 293 cells in a microtiter plate at an increasing moi. At two days following infection an quantitative assay (LPL.sup.s.sup.447.sup.x-mass assay) for the transgene product (LPL.sup.s.sup.447.sup.x) was performed on the medium of the infected cells. The assay showed that the amount of LPLs.sup.447x produced by baculovims-produced rAAV was similar to the LPL produced by the plasmid-produced rAAV (not shown).

(50) 2.2.3 Injection of rAAV Batches in Mice

(51) The rAAV batches produced with the baculovims-production system and with the conventional mammalian plasmid-production system were injected intramuscularly in mice to follow LPLs.sup.447x_protein activity and triglyceride fasting in vivo. At 3 days, 7 days and at 2 weeks following injection blood samples were taken and evaluated. Between 3 and 7 days post vims administration blood-plasma sampled from both mice injected with mammalian-rAAV and one mouse injected with baculo-rAAV was turned from milky to completely clear. Blood plasma derived from one baculo-rAAV-injected mouse remained relatively milky however fat level was clearly reduced. Triglyceride levels were lowered respectively of all treated mice (not shown). On day 14 TG levels in both mammalian-AAV and baculovims-(TTG)-AAV treated mice TG levels were reduced for 96%. Plasma samples taken at two weeks after vims administration showed that the LPLs.sup.447x-activity of the mice treated with baculovims-AAV and mammalian-AAV was similar (not shown).

Example 3: Stability of rAAV Constructs with Modified Rep 78 Initiation Codon in Insect Cells

(52) 3.1 Comparison of the Stabilities of Various rAAV Constructs in Insect Cells

(53) rAAV productions in SF+ cells were performed as described above in Example 1. For all productions the ITR containing baculovirus and the capsid gene containing baculovirus were identical, the passage number was the same as the Rep gene containing baculoviruses. 4 different Rep gene containing baculovirus were used: 1) The REP-ACG/PSC, 2) SLR: the original construct by Urabe et al. (2002, supra), 3) Rep52+Rep78(B2B): Two separate Bae-to-Bae baculoviruses, one containing the Rep 78 gene and the other one containing the Rep 52 gene. 4) Rep52+Rep78(PSC): Two separate protein sciences baculoviruses one containing the Rep 78 gene and the other one containing the Rep 52 gene.

(54) Results are shown in FIG. 4 at fifth baculovirus passage rAAV production is already improved by more than a factor 10 using a REP-ACG/PSC in accordance with invention as compared to the original Rep construct and compared to the split Rep constructs.

(55) 3.2 Stability of the Baculovims Constructs Up to Passage 8

(56) rAAV productions in SF+ cells were performed as described in Example 1. For all productions the ITR containing baculovirus and the capsid gene containing baculovirus were identical, the passage number was the same as the REP-ACG/PSC baculovirus. Results are shown in FIG. 5. The REP-ACG/PSC baculovirus is stable to at least passage 8. rAAV production titers of REP-ACG/PSC are stable up to at least 8th passage of the baculovirus.

(57) 3.3 Passage Effect on Rep Protein Expression

(58) The effect of passage number on the expression of Rep protein for the original construct from Urabe et al. (2002, supra) was compared to a REP-ACG/PSC construct in accordance with the invention. The baculovims passages and the western blot were done as described in Example 1. During a normal passage of the rep baculovimses, samples were taken at 40 hours after addition of the baculovimses to the SF cells and western blot was performed. FIG. 6 clearly shows diminished Rep expression in higher passages compared to earlier passages for the original Urabe construct (SLR), while the Rep expression in the REP-ACG/PSC construct stays the same in the higher passages compared to the lower ones.