sign in sign up
<<
Home , ColE1, Gene, Origin of replication, PIR (gene), Plasmid

Gene (1999) 6, 1482–1488  1999 Stockton Press All rights reserved 0969-7128/99 $12.00 http://www.stockton-press.co.uk/gt BRIEF COMMUNICATION pCOR: a new design of vectors for nonviral therapy

F Soubrier, B Cameron, B Manse, S Somarriba, C Dubertret, G Jaslin, G Jung, C Le Caer, D Dang, JM Mouvault, D Scherman, JF Mayaux and J Crouzet Rhoˆne-Poulenc Rorer, Centre de Recherche de Vitry Alfortville, 13 Quai J Guesde, 94403 Vitry-sur-Seine, France

A totally redesigned /vector system with improved initiator , ␲ protein, encoded by the gene limiting properties in terms of safety has been developed. The its host range to bacterial strains that produce this trans- pCOR are narrow-host range plasmid vectors for acting protein; (2) the plasmid’s selectable marker is not an nonviral . These plasmids contain a con- resistance gene but a gene encoding a bacterial ditional and must be propagated in a suppressor tRNA. Optimized E. coli hosts supporting specifically engineered E. coli host strain, greatly reducing pCOR replication and selection were constructed. High the potential for propagation in the environment or in yields of supercoiled pCOR monomers were obtained (100 treated patients. The pCOR backbone has several features mg/l) through fed-batch fermentation. pCOR vectors carry- that increase safety in terms of dissemination and selec- ing the gave high levels of lucifer- tion: (1) the origin of replication requires a plasmid-specific ase activity when injected into murine skeletal muscle.

Keywords: gene therapy; plasmid DNA; conditional replication; selection marker; multimer resolution

Two different types of DNA vehicles, based criteria.4 These high copy number plasmids carry a mini- on recombinant and bacterial DNA plasmids, are amount of bacterial sequences, a conditional origin used in gene therapy. Although -based vectors are of replication that prevents dissemination during widely used and very efficient vectors, the potential risks production or therapy and do not contain an antibiotic of viral have to be seriously considered. E. resistance gene. coli plasmid-based vectors are non-infectious and versa- The pCOR backbone consists of three bacterial tile vehicles for gene therapy. The manufacture of thera- elements: the 0.4 kb R6K ␥ conditional origin of repli- peutic plasmid DNA that carries potent cation (ori ␥), which requires the R6K ␲ initiator protein becomes a reality. Thus, new approaches that take into to be functional, a 0.2 kb selectable tRNA suppressor account safety and environmental consideration must gene (sup Phe), and a 0.4 kb cer (ColE1 resolution) be developed. fragment to resolve pCOR oligomers. Vector design is a major contributor to the safety of Such plasmids can only replicate in ␲-producing bac- production and therapy. Efforts to improve the plasmid teria considerably limiting their host range. pCOR selec- backbone have focused on the improvement of tion uses expression of a synthetic amber suppressor expression1 or plasmid copy number2 but no rational tRNA gene, specific for phenylalanine (sup Phe), which design in terms of dissemination or size, has been does not require . This suppressor corrects an reported. The ColE1-derived bacterial plasmids are cur- amber in argE gene making it possible for the rently used in gene therapy experiments and clinical recombinant host strain to grow on a minimal medium trials. As far as the safety and containment of such gen- lacking arginine (Figure 1). etic material are concerned, two elements are deleterious: The pCOR conditional origin of replication comes from the ColE1 origin of replication and the antibiotic resist- a natural plasmid, R6K, rather than the widely used ance marker. Clinical use of such plasmids could poten- ColE1-derived plasmids. R6K is a 38 kb theta-replicating tially lead to their dissemination in the environment or R (resistance) plasmid. R6K replication functions5 are in the patient. This could be of the utmost importance for clustered in a 5.5 kb DNA fragment that contains three the immunodepressed or for cystic fibrosis patients who origins of replication: ␣ and ␤ that are mainly used, and suffer from chronic bronchopulmonary .3 pCOR ␥. All three are under the control of a single replication plasmids, for plasmid with conditional origin of repli- initiation protein, ␲, the product of the pir gene. This pro- cation, were specifically designed and developed for gene tein binds to direct sequence repeats (seven therapy. These new plasmids are small efficient transgene tandemly associated 22-bp direct repeats) at the origin of carriers and fulfill several quality assurance and safety replication. It is a dual regulator, exerting both a positive effect on replication and negative that maintains R6K at its characteristic copy number (15 per ). Correspondence: J Crouzet Copy-up inhibit replication less and maintain Received 21 December 1998; accepted 12 April 1999 the ␥ origin at a high copy number. pir1166 contains a pCOR plasmids for gene therapy F Soubrier et al 1483

Figure 1 pCOR host vector system. The XAC-1pir116 strain requires arginine for growth on minimal medium whereas the arginine deficiency of the pCOR host is corrected by the tRNA sup Phe from the plasmid. The pCOR plasmid, the pir116 and argEam genes, and the XAC-1pir116 ⌬ are not drawn to scale. The amber mutation is shown by a stop and its correction by a crossed-out stop. XAC-1pir116 : ara (lac-proB)x111 ⌬ Ј + gyrA argEam rpoB thi uidA( MluI)::pir116 F (proB lacI373 lacZu118am). single base mutation (CCC→CTC) that leads to a proline Comparison of pXL2666 (ori ␥-KmR) and pXL2730 (ori to leucine substitution at position 106 in ␲. ␥-KmR-cer) topology (Figure 2) showed that cer leads to The minimum genetic information required to main- monomerization of R6K-derived plasmids. A deletion12 tain R6K derivatives is the 400-bp ␥ origin (ori ␥) and the leading to the inactivation of cer, was introduced into cis or trans-acting ␲ initiator protein.5 Ori ␥ was isolated pXL2666 to generate pXL2754 (ori ␥-KmR-⌬cer). This con- from the suicide vector pUT-T7pol7 as a 0.4 kb BamHI– struct led to the production of plasmid multimers. These EcoRI fragment and was ligated to the kanamycin resist- experiments were also carried out in XAC-1pir and XAC- ance gene (KmR) from pUC4K8 to generate pXL2666. Ori 1pir116 strains (data not shown) which differ only by the ␥ replication and were assayed in pir . Thus, these results shed light on the link wild-type and copy-up backgrounds. pXL2666 was trans- between plasmid multimerization/copy-up pir ferred into a pir (BW190949) and a pir116 strain and multimer resolution via cer-mediated intramolecu- (BW196109)ofE. coli. Agarose showed lar recombination. that there were five to 10 more plasmid DNA copies in An important original feature of pCOR is the absence the pir116 than in the pir strain, consistent with previous of an antibiotic resistance gene for selection in E. coli. E. work6 (data not shown). It also pointed out that the coli amber (TAG) suppressor tRNA genes do not contain monomer/oligomer ratio of R6K-derivatives was affected any sequences derived from phages or transposons and by the presence of the copy-up mutation which resulted are very small (less than 100 bp). They are efficient plas- in the major plasmid being multimers, mostly mid selection markers, even in high-density cultures of dimers (Figure 2). Plasmid linearization by EcoRI indi- E. coli.13 A synthetic amber suppressor tRNA gene, spe- cated that these preparations contained a mixture of cific for phenylalanine (sup Phe) was used as the pCOR monomer and the corresponding oligomers. Topoisomer- marker. sup Phe is expressed from a strong constitutive ase treatment showed that the high molecular weight E. coli , Plpp. This cassette (0.17 kb NarI–HindIII species were circular, supercoiled, plasmid multimers fragment) was isolated from pCT2-F14 and was tested in rather than catenanes. These plasmids are relaxed by E. coli XAC-1.14 The chromosomal argE gene of XAC-1 DNA I and topoisomerase IV10 and no contains an amber mutation that results in a truncated decatenation by topoisomerase IV was observed, and therefore inactive product. As argE encodes N-ace- although this was active on DNA tylornithinase which is essential for arginine biosynth- (data not shown). Plasmid multimerization is not com- esis, this strain cannot grow in the absence of arginine. patible with pharmaceutical use due to low lot to lot The suppressor tRNA sup Phe causes the correct trans- reproducibility and too few molecules per unit weight. lation of mRNA and the synthesis of an active full-length As well as reducing the number of independently segre- N-acetylornithinase. Thus, XAC-1 can grow on minimal gating molecules, it may also severely reduce plasmid medium lacking arginine as long as the plasmid is stability by affecting plasmid segregation during bacterial present, ensuring plasmid selection. However, this strain division. In our effort to resolve multimers, we tested does not synthesize the R6K ␲ initiator protein and thus the cer fragment that can promote monomerization inde- cannot replicate pCOR. Therefore, the ␲-encoding gene, pendently of the involved. This cis-acting DNA pir (or the copy-up pir116), was inserted into the determinant reduces multimerization via a site-specific XAC-1 chromosome by . This recombination event that requires only E. coli-encoded method does not introduce any mobile genetic element, . Two recombinases, XerC and XerD, are neces- such as transposons or phages, that affect genome stab- sary, along with two accessory proteins, ArgR and PepA, ility. Since there is no similarity between pir and the bac- that ensure exclusively intramolecular recombination at terial genome, pir or pir116 were inserted into the the cer site.11 The 0.38 kb HpaII cer fragment12 from the chromosomal uidA gene which encodes ␤-d-glucuronid- E. coli plasmid ColE1 was inserted into pXL2666. ase (EC 3.2.1.31), as described by Metcalf et al.9 This gene pCOR plasmids for gene therapy F Soubrier et al 1484

Figure 2 Influence of cer on the topology of R6K-derived plasmids. (a) Plasmid constructs. B: BamHI; E: EcoRI; M: MluI restriction sites. (b) Agarose gel analysis of plasmid topology. Plasmids were prepared by CsCl-ethidium bromide gradient and analyzed by agarose gel electrophoresis. 1, pXL2666; 2, pXL2754; 3, pXL2730. a, linear (EcoRI digested) plasmid; b, circular plasmid.

provides sufficient sequence similarity with the E. coli restriction analysis pattern was found to conform to genome for recombination to occur. ␤-Glucuronides are theory (data not shown). not essential in standard growth conditions, therefore it Since plasmid DNA degradation could occur during is possible to inactivate uidA by replacement with the the production/purification process, a pCOR host lacking gene of interest (pir or pir116). Replacement of the gen- endonuclease I was constructed. Endonuclease I, encoded omic uidA gene with the uidA::pir116 cassette which by endA, is a periplasmic DNase which is inhibited by results in uidA inactivation, is easy to detect with a RNA and internally cleaves native DNA into oligodeoxy- chromogenic substrate, X-gluc, as the ␤-glucuronidase- ribonucleotides in a sequence-independent fashion.17 positive colonies develop a blue color. The KmR-uidA:: This enzyme is not essential and its biological role in E. pir116 cassette9 from M13wm33 (or pir from M13wm34) coli is still unknown as mutants resemble wild-type cells was inserted into the BamHI site of the M13mp1015 rep- in terms of growth rate, phage propagation and conju- licative form. This recombinant is a suicide gation properties.18 Various studies19,20 have found that vector in non-suppressor E. coli strains. The gene replace- EndA− strains produce DNA of higher quality than ment technique developed by Blum et al16 was successful EndA+ E. coli strains. A DNA fragment containing endA21 in 18–34% of the deoxycholate-resistant and kanamycin- was used to construct an M13 KmR suicide vector. The sensitive segregants as demonstrated by the loss of ␤-glu- 0.37 kb FspI–PvuII central fragment was deleted remov- curonidase activity. These clones were tested by assessing ing 52% of endA. Gene replacement of the wild-type endA replication of an R6K derivative, pBW30.9 The integrity by the deleted endA was then obtained as previously of the in the region close to the site of described for the construction of XAC-1pir116. Genomic homologous recombination was analyzed by PCR. Each modification of the was verified by PCR and the unique amplified fragment, all of which were of the absence of endonuclease I activity was demonstrated by expected size, was identified by restriction analysis. The two biochemical tests. First, gene replacement was moni- 17 other genetic markers argEam, lacZam, gyrA (nalidixic acid tored by a methyl green staining assay. Second, the resistance), rpoB (rifampicin resistance), presence of epi- absence of endonuclease activity was assayed using the somal FЈ (susceptibility to M13 ), were unaffec- breakage of supercoiled DNA in the presence of cold- ted. The resulting pCOR host strains were named XAC- shock fluid22 from mutant or wild-type strains. Plasmid 1pir and XAC-1pir116. A pCOR plasmid (pXL2760: ori ␥- DNA was totally degraded with an extract from XAC- KmR-cer-sup Phe) was introduced into these strains and 1pir116 (EndA+ strain) whereas no degradation was subsequent selection by plating on minimal medium observed with the EndA− isogenic strain (XAC-1pir116 demonstrated that these two strains allow pCOR repli- endA−). cation and selection. The yield of pCOR from a small- XAC-1pir116 contains an FЈ episome, a circular DNA scale culture of the XAC-1pir116 recombinant strain was molecule of approximately 100 kb, that carries proB+

similar to that of standard plasmids such as pUC and lacI373 lacZu118am. Many male E. coli laboratory strains carry pCOR plasmids for gene therapy F Soubrier et al 1485 a traD36 mutation on their episome, but no mutation rable with those previously described with optimized affecting FЈ transfer ability has been described for XAC- ColE1-derived luciferase reporter plasmids.27 1. traD is at the 5Ј end of one of the (transfer) Nonviral gene therapy has been based exclusively on and encodes a that is directly involved E. coli plasmids derived from the ancestral ColE1 plas- in DNA transfer and DNA ;23 furthermore, mid. These plasmids replicate in the absence of plasmid- traD mutants do not transfer DNA.24 The presence of encoded proteins and their maintenance only requires E. wild-type traD can be considered as a potential risk of coli host-encoded factors, such as DNA I and plasmid dissemination between E. coli strains by nonspe- III and DNA-dependent RNA . To restrict cific transfer of co-resident plasmids.25 Therefore, the epi- plasmid host range to a single E. coli strain, we designed somal traD gene of XAC-1pir116 was inactivated to abol- a plasmid derived from an E. coli natural narrow-host ish FЈ transfer. A 2 kb central fragment from traD, range plasmid,28 R6K. Unlike ColE1 replication, R6K comprising 92% of the gene, was replaced with the 2 kb requires synthesis of the plasmid-coded protein ␲ that omega element from pHP45⍀26 by homologous recombi- controls R6K replication through ori ␥ core sequences. nation in XAC-1pir116 endA−. The omega element con- pCOR replication is restricted to a specific bacterial host tains the aadA antibiotic resistance gene flanked by short expressing ␲. The ␥ origin from R6K has no identified inverted repeats. aadA encodes -3 adenyl- open-reading frame, no undesirable sequences such as and confers resistance to streptomycin and sequences, and no element likely to direct its spectinomycin. The omega fragment was used because it transfer to other . Although R6K contains two prematurely terminates RNA and protein synthesis lead- transfer origins,29 they are not present in the ␥ origin of ing to the inactivation of the whole traD , not as replication. Bacterial host FЈ episome transfer was abol- a selectable marker (no antibiotic added during plasmid ished by insertional inactivation of the episomal traD to production). PCR and restriction analysis of the modified avoid low efficiency illegitimate co-transfer of pCOR. traD locus showed the expected pattern of amplified frag- The pCOR bacterial hosts, XAC-1pir116 or derivatives ments demonstrating that gene replacement had such as TEX1, are specifically engineered E. coli strains occurred. The resultant strain was named TEX1. This new containing a chromosomal copy of a copy-up variant of pCOR host strain, which is more suitable for industrial pir, pir116, inserted into the nonessential uidA gene. R6K applications, is being evaluated. derivatives have a low copy number (around 15 copies Since gene therapy applications require amounts per genome) which is not compatible with industrial- of plasmid DNA, we assessed the pCOR host strain in scale production. Introduction of pir116 into the XAC-1 fermentation experiments. Complex media containing genome generates high copy number pCOR where plas- extract were used for fed-batch fermentation with mid multimerization was resolved by inserting the cer XAC-1pir116. pCOR stability (more than 50 generations) resolution site into pCOR. This is the first report of plas- makes it possible to use such non-selective media. In mid multimers in pir116 strains and of cer-mediated these conditions, XAC-1pir116 produced more than multimer resolution. We successfully cloned inserts of 8 40 g/l dry cell weight and 100 mg/l of pCOR pXL2774 kb into the pCOR backbone without any effect on plas- (described in Figure 3) could be obtained from 2-liter - mid copy number or stability indicating that pCOR or mentors (Table 1). pCOR copy number was evaluated at ColE1 derivatives appeared to have similar 400–500 plasmids per cell and the rate of plasmid DNA capacity. R6K derivatives can accommodate inserts as synthesis was constant throughout fermentation. These large as 50–100 kb in a copy up context.30 results were extrapolated to an 800-liter fermentor suit- R6K ␥ origin-bearing plasmids are widely used as suic- able for manufacturing (Table 1). To control culture con- ide vectors in various bacteria, such as E. coli (non-patho- ditions further, fermentation was also performed in the genic9 or pathogenic31 strains), ,32 Pasteur- absence of yeast extract or any raw material from ella,33 Rhodobacter spheroides,34 Pseudomonas putida or origin. Similar results (30 g/l dry cell weight and 100 .7 This strongly supports the fact that mg/l of plasmid DNA) were obtained using a defined pCOR replication is totally dependent on ␲ initiator pro- medium in 2-liter cultures with no loss of productivity tein, and is thus biologically safe. The replication of natu- (Table 1). ral R6K, which encodes ␲, is restricted to E. coli and pCOR gene transfer in vitro and in vivo was assessed closely related bacteria.28 To our knowledge, there is no and compared with commercial plasmids. Various pCOR report of a ␲-like activity that may allow the replication plasmids were constructed with luc, the gene encoding of R6K derivatives. The potential risk of dissemination of firefly luciferase, or luc+, the engineered cytoplasmic pCOR would be restricted to a bacterium containing a luciferase gene. These plasmids were efficient in standard resident R6K. However, incompatibility35 between plas- in vitro of murine or human cell lines, giving mids that share the major incompatibility factor36 present higher levels of reporter gene activity than commercial in ori ␥ will lead to the loss of one of these plasmids and plasmids (ie pGL3-basic; Promega, Madison, WI, USA) in any case, will result in only ephemeral pCOR repli- (data not shown). The same amount of luciferase cation since there is no selective advantage for bacteria (3 × 1012) in plasmids pGL3-control, to maintain a pCOR. pCOR replication is conditional and pXL2774 and pXL3133 gave 0.07, 0.4 and 16.5 ng of offers environmental advantages over current plasmids luciferase per muscle (n = 24), respectively, when injected used in gene therapy. into the tibialis anterior of OF1 mice, as determined 72 h Plasmids are maintained as stable extrachromosomal after injection (Figure 3). Thus the pCOR plasmid DNA molecules and are excellent vectors for the dissemi- pXL3133 with an improved luciferase expression cassette nation of antibiotic resistance genes37 to a wide range of is efficient at driving expression in muscle with levels of bacteria. Contamination of the final product intended for reporter gene activity at least 200-fold higher than that use in gene therapy with an antibiotic could cause of a reference commercial vector. This result was compa- allergic reactions. The use of a suppressor tRNA as a sel- pCOR plasmids for gene therapy F Soubrier et al 1486

Figure 3 In vivo transfection with pCOR-Comparison of luciferase vectors after injection into muscle. Tibialis anterior muscles of mice were injected percutaneously with 30 ␮l of DNA in saline containing 3 × 1012 luciferase expression cassettes (approximately 15 ␮g of naked DNA). Mice were killed 72 h after injection. The muscles were collected and transferred to the lysis solution from the Luciferase assay system (Promega, Madison, WI, USA). The extraction was further performed according to the manufacturer’s instructions. The luciferase activity was measured using a LUMAT LB9501 luminometer (Berthold, Evry, France). Results are indicated as nanograms (ng) of luciferase per injected muscle as mean ± standard error of the mean. Plasmid DNA was purified by CsCl-ethidium bromide gradient ultracentrifugation and analyzed by agarose gel electrophoresis. Endotoxin level ranged between 1 and 5 EU/mg of plasmid DNA. The major structural elements of the plasmids vectors are indicated. The name of the pCOR plasmids is underlined. CMV, human cytomegalovirus immediate–early1 /promoter; SV40, simian virus 40 promoter; TK leader, leader sequence from HSV1 TK gene; luc, peroxisomal luciferase gene; luc+, modified (cytoplasmic) luciferase gene; SV40 late/early, simian virus 40 (SV40) late/early signal; bGH, bovine growth hormone polyadenylation signal.

ectable marker in bacteria is therefore an interesting alter- Consistent with these observations, natural mammalian native to the commonly used drug resistance markers. amber suppressor tRNA were isolated from calf liver.41 This tRNA, for which the anticodon was mutated to We detected no apparent changes in cell viability when match the amber termination triplet, mediates plasmid pCOR were transfected in murine or human cell lines in selection in bacteria by suppression of an amber mutation vitro or in murine muscle cells in vivo. Furthermore, when in a gene essential for growth in the conditions used. This eukaryotic amber suppressor tRNA are expressed from system developed by Normanly et al14 proved to be appropriate eukaryotic expression signals, no adverse efficient even in fermentation experiments.13 An amber effects are noticed in transfected mammalian cells42,43 or suppressor tRNA may cause readthrough at a significant in whole .44 Thus, the bacterial suppressor tRNAPhe number of natural termination codons and induce abnor- is unlikely to cause any deleterious effects in human cells mal protein synthesis. However, this requires the con- and is an efficient selectable marker for pCOR in E. coli. junction of several events in : (1) sufficient tran- Short (6 bp) immunostimulatory sequences (ISS) in scription of the suppressor tRNA in mammalian cells; (2) bacterial plasmid DNA have been shown to trigger correct folding of this tRNA which is essential to its powerful immune responses and to affect gene activity; (3) correct recognition; and (4) loading of a spe- expression in mice.45 Schwartz et al46 found that unmeth- cific prokaryotic tRNA by a eukaryotic tRNA synthetase. ylated CpG motifs caused inflammation of the lower res- Several experiments suggest that this is unlikely. Jacob- piratory tract of mice that received DNA by intratracheal son38 obtained no aminoacylation of E. coli wild-type instillation. Results obtained with oligonucleotides47 gave tRNAPhe with extracts from rat liver, mouse liver, maize indication of ISS potential to stimulate human B cells but or human spleen. It has also been shown that even if an the effects in man of plasmidic ISS have still to be investi- E. coli suppressor tRNAGln is expressed in mammalian gated. However, the results obtained in mice suggest that cells, amber suppression is dependent upon co- these sequences may also cause problems in . The expression of the E. coli GlnRS aminoacyl-tRNA synthet- pCOR backbone does not contain ISS. Moreover, there ase gene.39 Expression of sup Phe can only result from are no ‘two 5Ј purines – unmethylated CpG – two 3Ј unknown cryptic eukaryotic expression signals in the pyrimidine’ motifs in the pCOR backbone. In contrast, plasmid backbone. To avoid spurious , the a plasmid containing the widely used Tn903 kanamycin eukaryotic expression cassette was inserted into pCOR resistance gene from pUC4K (Pharmacia Biotech, such that the transgene and sup Phe gene are in Uppsala, Sweden), for example VR1255,27 and the ColE1- opposite orientations. derived origin of replication from pBlueScript Amber suppression is never complete and most known (Stratagene, La Jolla, CA, USA) contains nine such motifs. mammalian genes (79%40) terminate in UAA or UGA. Thus pCOR should have less, if any, immunostimulatory pCOR plasmids for gene therapy F Soubrier et al 1487 Table 1 Properties of the high cell density fermentation of XAC- Acknowledgments 1pir116 (pXL2774) We would like especially to thank JM Masson for provid- ␥ Complex medium Defined ing sup Phe and XAC-1 and B Wanner for ori and uidA medium ::pir cassettes. W Wackernagel is acknowledged for the 2- or 7-l 800-l 2-l generous gift of the endA fragment. We would like to fermentor fermentor fermentor thank F Blanche and collaborators for topoisomerase experiments and T Ciora for synthesis EFT (h) 40 39 48 and sequencing. ␮ (h−1) 0.130 0.132 0.124 OD (600 nm) 114 100 94 X (g/l) 44 37 30 References DNA (mg/l broth) 115 100 100 DNA (mg/gX) 2.6 2.7 3.3 1 Yew N et al. Optimization of plasmid vectors for high-level expression in lung epithelial cells. Hum Gene Ther 1997; 8: X, biomass (dry cell weight); EFT, elapsed fermentation time; 575–584. OD, optical density. 2 Lahijani R et al. High-yield production of pBR322-derived plas- Fermentation studies were carried out in 2-l fermentors (Setric, mids intended for human gene therapy by employing a tem- Toulouse, France) and culture conditions were scaled up to 800- perature-controllable . Hum Gene Ther 1996; 7: l fermentors. The basal production medium contained salts 1971–1980. (KH2PO4 (8 g/l), K2HPO4 (6.3 g/l), Na2HPO4 (1.7 g/l) and trace 3 Gilligan P. Microbiology of airway disease in patients with cys- elements), 10 mg/l thiamine and glucose as the carbon source; tic fibrosis. Clin Microbiol Rev 1991; 4: 35–51. yeast extract (3 g/l) and NH4Cl (2.5 g/l) were used as the 4 Smith K, Shepherd A, Boyd J, Lees G. Gene delivery systems nitrogen source in complex and defined media, respectively. The for use in gene therapy: an overview of quality assurance and complex medium was fed with 50% glucose, 0.7% MgSO4,7H2O, safety issues. Gene Therapy 1996; 3: 190–200. 0.02% thiamine and 10% yeast extract, whereas for the defined 5 Filutowicz M et al. Regulation of replication of an iteron-contain- medium feeding, the nitrogen was brought by the ammonia ing DNA molecule. Prog Res Mol Biol 1994; 48: necessary to adjust the pH. The pH was automatically controlled 239–273. and adjusted to a pH of 6.9–7.0 with 10% (w/v) ammonia. 6 Greener A, Filutowicz M, McEachern M, Helinski D. N-terminal Aeration was kept at 75 l/h below 0.2 bars and dissolved oxygen was maintained at 40% air saturation by the impeller truncated forms of the bifunctional pi initiation protein express speed controller and, when necessary, by enrichment of the air negative activity on plasmid R6K replication. Mol Gen Genet stream with pure oxygen. All the process variables (pH, 1990; 224: 24–32. 7 Herrero M, de LV, Timmis K. Transposon vectors containing temperature, agitation, optical density, O2 and CO2 contents of exhaust gases) were monitored on-line. The feed media were non-antibiotic resistance selection markers for cloning and added according to a pre-programmed time-based profile. stable chromosomal insertion of foreign genes in gram-negative bacteria. J Bacteriol 1990; 172: 6557–6567. 8 Vieira J, Messing J. The pUC plasmids, an M13mp7-derived sys- tem for insertion mutagenesis and sequencing with synthetic universal primers. Gene 1982; 19: 259–268. 9 Metcalf W, Jiang W, Wanner B. Use of the rep technique for properties than standard ColE1-derived plasmids. If allele replacement to construct new hosts for DNA vaccination is envisaged, immunostimulatory maintenance of R6K gamma origin plasmids at different copy sequences could be added to the plasmid backbone. numbers. Gene 1994; 138: 1–7. The pCOR backbone is significantly smaller (1 kb) than 10 Ferrero L et al. Cloning and primary structure of DNA topoisomerase IV: a primary target of fluoroquino- standard ColE1-derived plasmids (2.5 kb). It contains a lones. Mol Microbiol 1994; 13: 641–653. minimal number of prokaryotic DNA sequences, the 11 Sherratt D et al. Site-specific recombination and circular chromo- functions of which are known. The reduction of the some segregation. Philos Trans R Soc Lond B Biol Sci 1995; 347: amount of prokaryotic sequences in plasmid vectors is of 37–42. importance since some of these sequences were shown in 12 Leung D, Chen E, Cachianes G, Goeddel D. Nucleotide sequence different studies27–48 to interfere with eukaryotic gene of the partition function of Escherichia coli plasmid ColE1. DNA expression. 1985; 4: 351–355. We aimed to improve biosafety during plasmid pro- 13 Simoes D et al. A sugar-inducible excretion system for the pro- duction and nonviral gene therapy by totally redesigning duction of recombinant proteins with Escherichia coli. Ann NY Acad Sci 1991; 646: 254–258. the plasmid backbone. The pCOR plasmid and host sys- 14 Normanly J et al. Construction of two Escherichia coli amber sup- tem are suitable for clinical and industrial applications pressor genes: tRNAPheCUA and tRNACysCUA. Proc Natl Acad in terms of safety (conditional origin of replication, no Sci USA 1986; 83: 6548–6552. antibiotic marker, minimal plasmid sequences, inacti- 15 Messing J, Vieira J. A new pair of M13 vectors for selecting vation of bacterial episome transfer) and either DNA strand of double-digest restriction fragments. Gene (at least as efficient in vitro and in vivo as standard ColE1- 1982; 19: 269–276. derived plasmids, smaller size). In addition, pCOR has 16 Blum P et al. Gene replacement and retrieval with recombinant proved to be appropriate for production (high yield: 100– M13mp . J Bacteriol 1989; 171: 538–546. 150 mg/l in fermentors, possibility of using defined 17 Wright M. Mutants of Escherichia coli lacking endonuclease I, media) and purification (small size, high copy number, ribonuclease I, or ribonuclease II. J Bacteriol 1971; 107: 87–94. 18 Du¨ rwald H, Hoffmann-Berling H. Endonuclease-I-deficient and elimination of endonuclease I activity of the host). ribonuclease I-deficient Escherichia coli mutants. J Mol Biol 1968; In conclusion, pCOR is an appropriate DNA vector for 34: 331–346. gene delivery which offers environmental advantages 19 Taylor R, Walker D, McInnes R. E. coli host strains significantly over current plasmids without any loss in gene transfer affect the quality of small scale plasmid DNA preparations used efficiency. for sequencing. Nucleic Acids Res 1993; 21: 1677–1678. pCOR plasmids for gene therapy F Soubrier et al 1488 20 Schoenfeld T et al. Effects of bacterial strains carrying the endA1 34 Penfold R, Pemberton J. An improved suicide vector for con- genotype on DNA quality isolated with Wizard(TM) Plasmid struction of chromosomal insertion mutations in bacteria. Gene Purification Systems. Promega Corporation: Promega Notes 1992; 118: 145–146. 1995; 53. 35 Novick R, Hoppensteadt F. On plasmid incompatibility. Plasmid 21 Jekel M, Wackernagel W. Location of the endA gene coding for 1978; 1: 421–434. endonuclease I on the physical map of the Escherichia coli K-12 36 Filutowicz M et al. DNA and protein in the regu- chromosome. J Bacteriol 1994; 176: 1550–1551. lation of plasmid replication. J Cell Sci Suppl 1987; 7: 15–31. 22 Ananthaswamy H. The release of endonuclease I from Escher- 37 Davies J, Smith D. Plasmid-determined resistance to antimicro- ichia coli by a new cold shock procedure. Biochem Biophys Res bial agents. Annu Rev Microbiol 1978; 32: 469–518. Commun 1977; 76: 289–298. 38 Jacobson K. Reaction of aminoacyl-tRNA synthetases with het- 23 Frost L, Ippen-Ihler K, Skurray R. Analysis of the sequence and erologous tRNAs. Prog Nucleic Acid Res Mol Biol 1971; 11: 461– gene products of the transfer region of the F sex factor. Microbiol 488. Rev 1994; 58: 162–210. 39 Drabkin H, Park H, Raj Bhandary U. Amber suppression in 24 Panicker M, Minkley EJ. DNA transfer occurs during a cell sur- mammalian cells dependent upon expression of an Escherichia face contact stage of F sex factor-mediated . coli aminoacyl-tRNA synthetase gene. Mol Cell Biol 1996; 16: J Bacteriol 1985; 162: 584–590. 907–913. 25 Clark A, Warren G. Conjugal transmission of plasmids. Annu 40 Nakamura Y, Gojobori T, Ikemura T. Codon usage tabulated Rev Genet 1979; 13: 99–125. from the international DNA sequence databases. Nucleic Acids 26 Prentki P, Krisch H. In vitro insertional mutagenesis with a sel- Res 1997; 25: 244–245. ectable DNA fragment. Gene 1984; 29: 303–313. 41 Valle R, Morch M, Haenni A. Novel amber suppressor tRNAs of mammalian origin. EMBO J 1987; 6: 3049–3055. 27 Hartikka J et al. An improved plasmid DNA 42 Laski F, Belagaje R, Raj Bhandary U, Sharp P. An amber sup- for direct injection into skeletal muscle. Hum Gen Ther 1996; 7: pressor tRNA gene derived by site-specific mutagenesis: cloning 1205–1217. and function in mammalian cells. Proc Natl Acad Sci USA 1982; 28 del Solar G, Alonso J, Espinosa M, Diaz-Orejas R. Broad-host- 79: 5813–5817. range plasmid replication: an open question. Mol Microbiol 1996; 43 Hudziak R et al. Establishment of mammalian cell lines contain- 21: 661–666. ing multiple nonsense mutations and functional suppressor 29 Avila P, Nunez B, de lCF. Plasmid R6K contains two functional tRNA genes. Cell 1982; 31: 137–146. oriTs which can assemble simultaneously in relaxosomes in vivo. 44 Choisne N, Martin CA, Small I. Transactivation of a target gene J Mol Biol 1996; 261: 135–143. using a suppressor tRNA in transgenic tobacco plants. J 30 Posfai G et al. In vivo excision and amplification of large seg- 1997; 11: 597–604. ments of the Escherichia coli genome. Nucleic Acids Res 1994; 22: 45 Sato Y et al. Immunostimulatory DNA sequences necessary for 2392–2398. effective intradermal gene immunization. Science 1996; 273: 31 Schmoll T, Ott M, Oudega B, Hacker J. Use of a wild-type gene 352–354. fusion to determine the influence of environmental conditions 46 Schwartz D et al. CpG motifs in bacterial DNA cause inflam- on expression of the S fimbrial adhesin in an Escherichia coli mation in the lower respiratory tract. J Clin Invest 1997; 100: . J Bacteriol 1990; 172: 5103–5111. 68–73. 32 Skrzypek E, Haddix P, Plano G, Straley S. New suicide vector 47 Liang H et al. Activation of human B cells by phosphorothioate for gene replacement in yersiniae and other gram-negative bac- oligodeoxynucleotides. J Clin Invest 1996; 98: 1119–1129. teria. Plasmid 1993; 29: 160–163. 48 Peterson D, Beifuss K, Morley K. Context-dependent gene 33 Azad A, Coote J, Parton R. Construction of conjugative shuttle expression: cis-acting negative effects of specific procaryotic and suicide vectors for Pasteurella haemolytica and P. multocida. plasmid sequences on eucaryotic genes. Mol Cell Biol 1987; 7: Gene 1994; 145: 81–85. 1563–1567.