Human Foamy Virus Integrase Fails to Catalyse the Integration of a Circular DNA Molecule Containing an LTR Junction Sequence

Home , Rous sarcoma virus

Gene Therapy (2002) 9, 1326–1332  2002 Nature Publishing Group All rights reserved 0969-7128/02 $25.00 www.nature.com/gt BRIEF COMMUNICATION Human foamy virus integrase fails to catalyse the integration of a circular DNA molecule containing an LTR junction sequence

RA Russell1, R Critchley2, G Vassaux2 and MO McClure1 1Jefferiss Research Trust Laboratories, Wright-Fleming Institute, Imperial College School of Medicine at St Mary’s Hospital, London, UK; and 2Molecular Therapy Laboratory, ICRF Molecular Oncology Unit, ICSM at Hammersmith Hospital, London, UK

The presence of closed circular forms of the linear DNA gen- ing transfection of both 293 and 293T cells indicated that the ome of human foamy virus (HFV) has not been established. level of integration was not significantly increased by the The ability of the HFV integrase (IN) to catalyse the inte- HFV IN. Moreover, correctly integrated provirus-like forms gration of these circular forms (termed 2 long terminal repeat of the input plasmid could not be detected by PCR. Taken (LTR) circles) was investigated, with a view to producing a together, these results show that the HFV IN is not able to novel hybrid vector. To this end, a construct was made con- integrate a circular molecule containing an LTR junction and, taining, in addition to the enhanced green fluorescent protein hence, the technique is not exploitable as a tool to produce (eGFP) marker gene, the last 27 bp of the 3’ U5 LTR region hybrid vectors for gene therapy. of HFV fused to the first 28 bp of the 5’ U3 LTR, the latter Gene Therapy (2002) 9, 1326–1332. doi:10.1038/sj.gt.3301795 representing a 2LTR circle. Marker gene expression follow-

Keywords: human foamy virus; 2LTR circles; integration; gene therapy; Cre recombinase

Production of terminally redundant DNA from the RNA vitro,11,12 however, they appear to be entirely non-patho- genome and the subsequent integration of this DNA into genic in vivo.13–15 The apparent lack of associated disease, the host chromosome are key stages in the retroviral in conjunction with their integrative element,16 has led to lifecycle. Early in infection in most retroviruses virally interest in the use of foamy viruses as vectors for the encoded reverse transcriptase (RT) copies the RNA into delivery of therapeutic genes in vivo. Most work has been the DNA form, adding terminal redundancies in the pro- carried out with the prototype human foamy virus (HFV) cess, termed long terminal repeat (LTR) sequences.1 This which is now known to be of chimpanzee origin.17,18 Dur- linear molecule then enters the nucleus where it inte- ing HFV integration, the presence of closed circular forms grates randomly into the host genome by means of the of the proviral DNA has not been established. We con- retroviral integrase (IN), in a precise mechanism involv- sidered whether the HFV IN could efficiently integrate a ing the removal of the terminal two base pairs (bp) of closed circular molecule containing a 5’–3’ LTR junction each LTR.2–5 In a number of retroviruses, for example sequence: the long-term aim being to exploit this in the murine leukaemia virus (MLV), spleen necrosis virus design of a hybrid vector combining HFV with an aden- (SNV), Rous sarcoma virus (RSV) and avian sarcoma and ovirus (Ad) by means of the bacteriophage P1 cre/loxP leukosis viruses (ASV,ALV), circularised forms of the lin- system.19–23 ear DNA molecule (2LTR circles) have been ident- Ad vectors have been developed extensively due to the ified.3,4,6–9 In the case of MLV, these closed circular forms large packaging capacity and broad host-range of the par- are unnecessary for integration and the direct precursor ent virus. Furthermore, they can infect both dividing and to integration of the MLV provirus is the linear mol- non-dividing cells, are produced to high titres with com- ecule.3,4,10 However, in SNV, RSV, ASV and ALV, closed parative ease and have a high in vivo transduction circular forms were shown to act as targets for efficiency. However, following infection, the genome integration,7–9 although these reports have not been remains episomal and is lost during cell division. Retrovi- substantiated. ral vectors, on the other hand, are of value because of Foamy viruses are classified as one of the seven genera their ability to integrate a transgene into the host genome of retroviruses. They infect a wide variety of cell types, allowing long-term expression. However, unlike Ad vec- producing a characteristic ‘foamy’ cytopathology in tors, they are limited by production difficulties, restricted packaging capacity, reduced ability to infect non-dividing cells and low in vitro transduction efficiency. The respective disadvantages of either vector are best over- Correspondence: Jefferiss Research Trust Laboratories, Wright-Fleming, Institute, Imperial College School of Medicine at St Mary’s Hospital, Nor- come by producing hybrid vectors incorporating the inte- folk Place, London, W2 1PG, UK gration ability of retroviruses with the high titre, broad Received 30 January 2002; accepted 29 April 2002 host-range of Ad vectors. This has already been achieved Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1327

Figure 1 DNA constructs. (a) Plasmid pLTRI contains the CMV-IE promoter linked to the eGFP gene and the SV-40 poly A, followed by an HFV LTR junction sequence, all flanked by two loxP sites, cloned into pCRII (Invitrogen Life Technologies, UK). The first loxP site, generated by the annealing of complementary oligonucleotides, was inserted into pCRII using the TA cloning technique (Invitrogen Life Technologies, Pais- ley, UK). The LTR junction sequence and the second loxP site, also generated by annealing complementary oligonucleotides, were inserted via the NotI and XhoI sites. The CMV-IE promoter, eGFP gene and poly A cassette were amplified from pEGFP-N1 (Clontech, Cowley, UK) by PCR, and cloned between the first loxP site and the LTR junction sequence via the EcoRV site. The HFV LTR junction sequence consists of the last 27 bases of the 3’ U5 region (underlined), followed by the first 28 bases of the 5’ U3 region (bold), the sequence of which is as follows: 5’AATCAAT ATACAAAATTCCATGACAATTGTGGTGGAATGCCACTAGAAAC TAGGG3’. The sequence of each loxP site is: 5’ATAACTTCGTAT AGCATACATTATACGAAGTTAT3’. (b) Plasmid pLTRII was created Figure 2 Western blot to confirm IN expression. 293T cells were seeded × 5 from pLTRI by cloning the puroR gene linked to the encephalomyocarditis at 4 10 cells/well in a six-well plate 1 day prior to transfection, which virus IRES and a synthetic intron from pIRESpuro (Clontech), down- was carried out using the Calcium Phosphate Profection Mammalian stream of eGFP, by restriction digestion. Transfection kit (Promega, Southampton, UK). The cells were transfected in duplicate with pRR2 and pcIN1 (a total of 5 ␮g plasmid DNA per well), and incubated in the transfection mixture for 8 h. At 40 h post- with MLV and first generation Ad vectors by a number transfection, the cells were harvested into SDS-protein sample buffer (62.5 of groups that have generated Ad vectors expressing mM Tris-HCl, pH 6.8 buffer containing 1% glycerol, 5% 2-mercaptoe- thanol, 2% SDS, 1 mM EDTA, 0.025% bromophenol blue) at a concen- either the retroviral packaging components or the retrovi- tration of 1 × 106 cells/50 ␮l buffer. The lysates were incubated at 95°C ral vector genome. In vitro and iv vivo transduction with for 3–4 min, centrifuged at 16 000 g for 5 min and 20 ␮l applied to a these vectors leads to the production of transient MLV 10% polyacrylamide gel for PAGE analysis.43 A Western blot was carried vector producing cells which, in turn, result in inte- out44 using an IN-specific antibody.39 The 127 kDa Pol precursor and the gration and long-term transgene expression.24–30 40 kDa IN are shown. The smaller bands are assumed to be the results An alternative strategy developed by Murphy et al31 of protein degradation. relies on the ability of the MLV IN to integrate a 2LTR circle. Although in natural infection only the linear gen- protein (eGFP) gene driven by the cytomegalovirus ome is integrated,3,4,10 Murphy et al31 have clearly demon- immediate–early (CMV-IE) promoter, with the simian strated that if only 2LTR circles are available in an virus 40 (SV-40) late polyadenylation (polyA) site, fol- infected cell they can be integrated in a provirus-specific lowed by the LTR junction sequence (see Figure 1 manner by the MLV IN. The MLV vector was then incor- legend). This junction sequence consists of the last 27 bp porated into an Ad vector and, in conjunction with the of the 3’ U5 region fused to the first 28 bp of the 5’ U3 MLV IN and the cre/loxP system,19–21 was integrated in a region of the HFV LTR.35 The plasmid pLTRII was provirus-like manner allowing prolonged transgene derived from pLTRI by inserting the puromycin resist- expression. Zheng et al32 have also shown that an MLV ance (puroR) gene linked to the encephalomyocarditis vector genome cloned into an Ad vector can be integrated virus internal ribosome entry site (IRES) and a synthetic in a provirus-like manner. However, an IN protein was intron downstream of the eGFP gene (Figure 1b). The loxP not included and the mechanism by which integration sites, flanking the HFV cassette will allow Cre recombi- occurred has not been established.33 IN proteins are also nase directed excision of the intervening DNA.19,36–38 In encoded by bacteriophages and the bacteriophage ␭ IN, the context of a linear DNA genome, such as that of an is able to excise circular molecules containing its recog- Ad, Cre recombinase will excise a circular proviral-like nition attP and attB sites.34 Although, not directly compa- molecule. It was hoped that the HFV IN would be cap- rable to retroviral integration it is a further demonstration able of integrating this into the host genome by means of the ability of integrases to utilise circular molecules. of the LTR junction sequence. A Western blot using an Building on the observations with the MLV and bac- IN-specific antibody39 confirmed that both pRR2, express- teriophage systems,31,32,34 we attempted to show inte- ing the full repertoire of Gag and Pol proteins,40 and gration of a 2LTR circle by the HFV IN before develop- pcIN1, expressing IN alone,39 were producing IN (Figure ment in an Ad vector. To test the hypothesis, pLTRI and 2) and the data of Imrich et al,39 confirmed that IN was pLTRII were generated. The plasmid pLTRI (Figure 1a) able to enter the nucleus. contains two directly oriented loxP sites flanking the HFV To assess whether IN was capable of increasing the fre- cassette, which consists of the enhanced green fluorescent quency of integration of pLTRII by utilisation of the LTR

Gene Therapy Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1328

Figure 3 Percentage of eGFP-positive cells at each passage. 293 and 293T Figure 4 Integration assay. The number of puromycin-resistant colonies cells were seeded at 4 × 105 cells/well in a six-well plate 1 day prior to produced following co-transfection of 293 and 293T cells with pLTRII and transfection with: (i) pLTRII/pACYC184 (NEB); (ii) pLTRII/pcIN1; (iii) either pACYC184 or pcIN1 or pRR2. At 48 h post-transfection the cells × 5 pLTRII/pRR2. At 8 h post-transfection the medium was changed and 40 were harvested and 2 10 were seeded into 90-mm dishes in either ␮ h post-transfection the cells from the duplicate wells were harvested, com- medium containing 2.5 g/ml puromycin (Sigma, Poole, UK) or standard bined and an aliquot analysed by flow cytometry (FACSCalibur, Becton medium. At confluency, cells were harvested and re-seeded into selection Dickinson, Oxford, UK). A 2 × 105 aliquot was seeded into 90-mm dishes. and standard medium and the process repeated for four passages. Stable At confluency cells were again analysed by flow cytometry and re-seeded. colonies were counted in the selection dishes approximately 3 weeks after This process was repeated at each passage for four passages. The percent- seeding. The number of puromycin-resistant colonies obtained from each age of eGFP-expressing cells at each passage following transfection is transfection at each passage is shown for (a) 293 cells and (b) 293T cells. shown for (a) 293 cells and (b) 293T cells. structs were replaced with the non-expressing junction sequence, a series of transfections was carried pACYC184 (NEB, UK). In the presence of either pcIN1 out in both 293 and 293T cells and the level of eGFP or pRR2, the level of eGFP expression decreased at each expression monitored by flow cytometry over four pass- passage (represented in Figure 3 by the grey and white ages. The cells were maintained in medium without any columns, respectively), in a pattern similar to that seen selection pressure and, to ensure that the DNA concen- in the absence of IN (represented by the black columns). tration and, hence the transfection efficiency,34 in the In the event of IN-mediated integration, the number of negative controls was similar, the IN expressing con- eGFP expressing cells would stabilise over serial passage

Gene Therapy Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1329 Table 1 Primer sequence binding sites within pLTRI and pLTRII

Primer name Position within construct Orientation Sequence

1c U3 5ЈLTR sense 5ЈTGTGGTGGAATGCCACTAGAAACTAGG3Ј 1d U5 3ЈLTR antisense 5ЈCCACCACAATTGTCATGGAATTTTG3Ј 1f eGFP antisense 5ЈGTGAACAGCTCCTCGCCCTTGCTC3Ј 2d CMV promoter sense 5ЈCCCTAGTTTCTAGTGGCATTCCACC3Ј to a level above that obtained in the absence of IN. Since this was not the case, integration was either occurring at a low level or not at all. Stable transformants were subsequently selected on the assumption that the number of cell clones formed under selection pressure, following transfection of pLTRII in the presence or absence of IN, related to the level of integration occurring at the LTR junction. The hypothesis is that more colonies would form in the presence of IN than in its absence, if it were capable of using the LTR junction as a target for integration. However, the number of colonies obtained per transfection, for both 293 and 293T cells was the same in the presence of IN (grey and white columns, Figure 4), as in its absence (black columns). The presence of integrated forms of both pLTRI and pLTRII was analysed by PCR to determine whether IN- mediated integration had occurred. However, as both constructs were greater than 5 kb and too large to amplify, the cre/loxP system, already included in the plasmids, was utilised. As the HFV cassette in both constructs was flanked by loxP sites, it was possible, in the presence of Figure 5 Oligonucleotide primer pairs and their expected band sizes. Cre, to excise them from the input plasmid, resulting in Primer pairs used to assess proviral integration. (a) The binding sites a 1.6 kb circle that was readily amplified by PCR. The and expected amplification products of each primer pair for proviral-like plasmid pLTRI was co-transfected into 293T cells with integration. (b) The binding sites and expected amplification products of the Cre-expressing pMC-Cre41 and either pRR2 or pcIN1. each primer pair for plasmid integration. Thus, primer pair 1c/1f are spe- As negative controls, pLTRI was transfected in the cific for the provirus-integrated form, with the LTRs separated and facing one another and primer pair 1d/2d will recognise both the proviral and absence of pMC-Cre, pRR2 and pcIN1 and also in the plasmid forms. presence of pMC-Cre alone. The plasmid pUC18 (Roche, Lewes, UK) replaced pMC-Cre, pRR2 and pc1N1 in the negative controls. and the episomal Cre-excised fragment of pLTRI, both At 96 h after transfection, passage 1 (P1), genomic of which should be lost on passaging. Primer pair 1c/1f DNA was extracted and amplified by PCR using two oli- produces a 0.7-kb amplification product if the 3’ and 5’ gonucleotide primer pairs specific to either proviral inte- LTRs have been separated by IN-mediated proviral integration or both proviral and plasmid integration (Figure gration and face one another. The whole 5.6-kb episomal 5 and Table 1). An aliquot of the cells was simultaneously pLTRI and the 1.6-kb episomal Cre-excised fragment of analysed by flow cytometry to confirm eGFP expression pLTRI before integration are also amplified, thereby (flow cytometry data not shown). Cells from each trans- showing the success of the Cre recombination (Table 2 fection experiment were passaged on two further shows the expected bands produced by each primer pair occasions and analysed by both flow cytometry and PCR with each form of integration and episomal fragment). at each passage (Figure 6). Primer pair 1d/2d recognises As before, the episomal fragment should be lost on both the proviral- and plasmid-integrated forms, produc- passaging. ing a 1.6-kb amplicon, and demonstrated that the DNA In all transfections, plasmid DNA was detected at P1 had not been lost in the extraction procedure. The pri- by primer pair 1d/2d (Figure 6a), but lost by P3, corre- mers also amplify a 1.6-kb product from episomal pLTRI sponding with the loss in eGFP expression (Figure 3).

Table 2 Expected amplicon sizes produced by each primer pair following Cre recombination and either plasmid or proviral integration

Primer pair Expected band sizes from:

Provirus-like integration Plasmid integration Episomal input plasmid Episomal Cre-excised fragment

1d/2d 1.6 kb 1.6 kb 1.6 kb 1.6 kb 1c/1f 0.7 kb 0 kb 5.6 kb 0.7 kb

Gene Therapy Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1330 the IN was not effectively mediating integration by means of an LTR junction sequence, thereby resulting in the loss of eGFP expression over time as the episomal plasmid was degraded. To distinguish episomal plasmids from true integration, the experiment was repeated using pLTRII, which allowed puromycin selection pressure to be applied. The plasmid pLTRII was co-transfected into 293T cells with pMC-Cre and pRR2. In control experiments, pLTRII was transfected in the absence of pMC- Cre and pRR2 and in the presence of pMC-Cre alone. At 96 h post-transfection (P1), genomic DNA was extracted from each transfection and amplified by PCR using primer pair 1c/1f specific for provirus-like integration and Cre recombination. As before, an aliquot of the cells was simultaneously analysed by flow cytometry to confirm eGFP expression. Cells from each transfection experiment were passaged a further three times under puromycin selection pressure and analysed by both flow cytometry and PCR at P2, P3 and P4 (Figure 7). Under puromycin selection pressure, eGFP expression was maintained in 100% of the cells up to P4. However, only the Cre-excised episomal fragment of pLTRII was amplified (Figure 7, lanes a and c). If proviral integration were occurring, the PCR product present in lane a would still be present at P4, at a similar intensity to that seen at P1, corresponding to the 100% eGFP expression seen up to P4. In addition, genomic DNA extracted from single cell clones made following transfection of 293T cells with pLTRII, pMC-Cre and either pRR2 or pcIN1 was analysed by PCR using primer pair 1c/1f. Despite 100% eGFP expression following a number of passages, the provirus-specific primers failed to amplify a band representing proviral-like integration. Figure 6 Assessment of proviral-like integration. 293T cells were trans- Taken together, these results show that while the HFV fected with either: pLTRI/pMC-Cre/pRR2, pLTRI/pMC-Cre/pcIN1, IN is able to integrate a linear form of the genome,16 it pLTRI/pUC18 or pLTRI/pMC-Cre/pUC18. At 96 h post-transfection the cells were harvested, an aliquot analysed by flow cytometry and 2 × 105 seeded into 90-mm dishes. Genomic DNA was extracted from 2 × 106 cells.45 At confluency the cells were again analysed by flow cytometry and re-seeded. Genomic DNA was also extracted from 2 × 106 cells. This process was repeated for three passages. The genomic DNA was analysed by PCR (Figure 4 and Tables 1 and 2) under the conditions 94°C, 2 min denaturation, 1 cycle followed by 94°C, 1 min denaturation; 58°C, 1 min annealing; and 72°C, 1 min 45 s, elongation, 25 cycles ending with 72°C, 10 min final elongation, one cycle. For primer pair 1c/1f an annealing temperature of 62°C for 30 s was used. For each primer pair a 293T cell and a PCR negative control, as well as a pLTRI plasmid DNA positive control, were carried out. The results for each primer pair, are shown for (a) primer pair 1d/2d and (b) primer pair 1c/1f.

This indicates that the PCR products seen at P1 and P2, in both the presence and absence of IN, were the result of episomal DNA amplification rather than the amplification of either IN-mediated proviral integration or plas- Figure 7 Assessment of proviral-like integration under puromycin selec- mid integration. The use of primer pair 1c/1f confirmed tion. 293T cells were transfected with either pLTRII/pMC-Cre/pRR2, that Cre recombination was occurring by the amplifi- pLTRII/pUC18 or pLTRII/pMC-Cre/pUC18. At 96 h post-transfection the cells were harvested and an aliquot analysed by flow cytometry (not cation of 1.6-kb bands in the presence of Cre (Figure 6b, shown) while genomic DNA was extracted from 2 × 106 cells. Cells from lanes a, b, d, e, f and h) and the amplification of the 5.6- each transfection were passaged a further three times in medium contain- kb input pLTRI in the absence of Cre (Figure 6b, lane ing 2.5 ␮g/ml puromycin and at each stage an aliquot was analysed by c). However, no product was amplified by P3. This was flow cytometry and genomic DNA was extracted. The results of the PCR, performed under the conditions 94°C, 2 min denaturation, one cycle fol- expected based on the results obtained by primer pair ° ° ° 1d/2d and the loss of eGFP expression, and again indi- lowed by 94 C, 1 min denaturation; 62 C, 30 s annealing; and 72 C, 1 min, elongation, 25 cycles ending with 72°C, 10 min final elongation, one cated that the bands seen in P1 and P2 were the result cycle, with primer pair 1c/1f are shown. A 293T cell and a PCR negative of episomal DNA amplification rather than the amplifi- control, as well as a pLTRII plasmid DNA positive control, were carried cation of an integrated molecule. These data suggest that out. The results of each transfection, at each passage, are shown.

Gene Therapy Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1331 is unable to integrate a closed circular DNA molecule 14 Schweizer M et al. Simian foamy virus isolated from an acciden- containing an LTR junction sequence. It is possible that tally infected human individual. J Virol 1997; 71: 4821–4824. the junction sequence was of insufficient length and that 15 Heneine W et al. Identification of a human population infected the inclusion of larger sections of the LTRs may be neces- with simian foamy viruses. Nat Med 1998; 4: 403–407. sary for recognition. However, in human immunodefi- 16 Vassilopoulos G, Trobridge G, Josephson NC, Russell DW. Gene transfer into murine hematopoietic stem cells with helper-free ciency virus type 1, 2LTR circles, containing full-length 42 foamy virus vectors. Blood 2001; 98: 604–609. LTRs, are artefacts that play no role in integration and 17 Herchenroder O et al. Infectious proviral clones of chimpanzee this may also be the case for HFV. Experiments carried foamy virus (SFVcpz) generated by long PCR reveal close func- out using the IN of SNV and a complete 2 LTR circle tional relatedness to human foamy virus. Virology 1995; 214: generated similar results to those obtained with HFV (not 685–689. shown), but were in contrast with early, and since unsub- 18 Herchenroder O et al. Isolation, cloning, and sequencing of sim- stantiated, SNV data of Panganiban and Temin.8 It is ian foamy viruses from chimpanzees (SFVcpz): high homology possible that provirus-like integration occurs at a level to human foamy virus (HFV). Virology 1994; 201: 187–199. below that of our detection methods, but this would be 19 Sauer B. Manipulation of transgenes by site-specific recombi- insufficient for exploitation as a means of producing nation: use of Cre recombinase. Meth Enzymol 1993; 225: 890– hybrid vectors for gene therapy. It would appear that 900. 31 20 Lieber A, He CY, Kirillova I, Kay MA. Recombinant adeno- although the MLV IN is able to integrate a 2LTR circle, viruses with large deletions generated by Cre-mediated excision this finding cannot be extrapolated to all retroviruses and exhibit different biological properties compared with first-gen- hence, may not be a strategy that can be used to produce eration vectors in vitro and in vivo. J Virol 1996; 70: 8944–8960. a variety of retrovirus/Ad virus hybrid vectors. 21 Hardy S et al. Construction of adenovirus vectors through Cre- lox recombination. J Virol 1997; 71: 1842–1849. Acknowledgements 22 Tan BT, Wu L, Berk AJ. An adenovirus-Epstein–Barr virus hybrid vector that stably transforms cultured cells with high This work was supported by The Wellcome Trust, The efficiency. J Virol 1999; 73: 7582–7589. Jefferiss Research Trust and The Imperial Cancer 23 Vanin EF et al. Development of high-titer retroviral producer Research Fund. We would like to thank Dr Richard Vile cell lines by using Cre-mediated recombination. J Virol 1997; 71: and Dr Stephen Murphy at the MAYO Institute, Roches- 7820–7826. ter, MN, USA for helpful discussion. 24 Feng M et al. Stable in vivo gene transduction via a novel adenoviral/retroviral chimeric vector. Nat Biotechnol 1997; 15: 866–870. References 25 Caplen NJ. Gene therapy: different strategies for different applications. American Society of Gene Therapy: First Annual Meet- 1Coffin JM, Hughes SH, Varmus HE (eds). Retroviruses. Cold ing, Seattle, Washington, USA, 28–31 May 1998. Mol Med Today Spring Harbor Laboratory Press: Cold Spring Harbor, 1997. 1998; 4: 374–375. 2 Varmus H. Retroviruses. Science 1988; 240: 1427–1435. 26 Reynolds PN, Feng M, Curiel DT. Chimeric viral vectors – the 3 Fujiwara T, Mizuuchi K. Retroviral DNA integration: structure best of both worlds? Mol Med Today 1999; 5:25–31. of an integration intermediate. Cell 1988; 54: 497–504. 27 Duisit G, Salvetti A, Moullier P, Cosset FL. Functional charac- 4 Brown PO, Bowerman B, Varmus HE, Bishop JM. Retroviral terization of adenoviral/retroviral chimeric vectors and their integration: structure of the initial covalent product and its pre- use for efficient screening of retroviral producer cell lines. Hum cursor, and a role for the viral IN protein. Proc Natl Acad Sci Gene Ther 10 USA 1989; 86: 2525–2529. 1999; : 189–200. 5 Brown PO. Integration of retroviral DNA. Curr Top Microbiol 28 Yoshida Y, Emi N, Hamada H. VSV-G-pseudotyped retroviral Immunol 1990; 157:19–48. packaging through adenovirus-mediated inducible gene 6 Shoemaker C et al. Structure of a cloned circular Moloney expression. Biochem Biophys Res Commun 1997; 232: 379–382. murine leukemia virus DNA molecule containing an inverted 29 Bilbao G et al. Adenoviral/retroviral vector chimeras: a novel segment: implications for retrovirus integration. Proc Natl Acad strategy to achieve high-efficiency stable transduction in vivo. Sci USA 1980; 77: 3932–3936. FASEB J 1997; 11: 624–634. 7 Luciw PA, Oppermann H, Bishop JM, Varmus HE. Integration 30 Ramsey WJ et al. Adenovirus vectors as transcomplementing and expression of several molecular forms of Rous sarcoma templates for the production of replication defective retroviral virus DNA used for transfection of mouse cells. Mol Cell Biol vectors. Biochem Biophys Res Commun 1998; 246: 912–919. 1984; 4: 1260–1269. 31 Murphy SJ et al. Novel integrating adenoviral/retroviral hybrid 8 Panganiban AT, Temin HM. Circles with two tandem LTRs are vector for gene therapy. Hum Gene Ther 2002; 13: 745–760. precursors to integrated retrovirus DNA. Cell 1984; 36: 673–679. 32 Zheng C, Baum BJ, Iadarola MJ. O’Connell B.C. Genomic inte- 9 Duyk G et al. Circles with two tandem long terminal repeats are gration and gene expression by a modified adenoviral vector. specifically cleaved by pol gene-associated endonuclease from Nat Biotechnol 2000; 18: 176–180. avian sarcoma and leukosis viruses: nucleotide sequences 33 Link CJ. Adenoviral vectors go retro. Nat Biotechnol 2000; 18: required for site-specific cleavage. J Virol 1985; 56: 589–599. 150–151. 10 Lobel LI, Murphy JE, Goff SP. The palindromic LTR-LTR junc- 34 Darquet AM et al. A new DNA vehicle for nonviral gene deliv- tion of Moloney murine leukemia virus is not an efficient sub- ery: supercoiled minicircle. Gene Therapy 1997; 4: 1341–1349. strate for proviral integration. J Virol 1989; 63: 2629–2637. 35 Enssle J et al. An active foamy virus integrase is required for 11 Hooks JJ, Detrick-Hooks B. Spumavirinae: foamy virus group virus replication. J Gen Virol 1999; 80: 1445–1452. infections: comparative aspects and diagnosis. In: Kurstak E, 36 Scott SD, Marples B. Comment on the use of the cre/loxP recom- Kurstak C (eds). Comparative Diagnosis of Viral Diseases, vol IV. binase system for gene therapy vectors. Gene Therapy 2000; 7: Academic Press: London, 1981, pp 599–618. 1706. 12 Mergia A, Leung NJ, Blackwell J. Cell tropism of the simian 37 Sauer B, Henderson N. Site-specific DNA recombination in foamy virus type 1 (SFV-1). J Med Primatol 1996; 25:2–7. mammalian cells by the Cre recombinase of bacteriophage P1. 13 Schweizer M et al. Markers of foamy virus infections in monk- Proc Natl Acad Sci USA 1988; 85: 5166–5170. eys, apes, and accidentally infected humans: appropriate testing 38 Sternberg N, Hamilton D. Bacteriophage P1 site-specific recom- fails to confirm suspected foamy virus prevalence in humans. bination. I. Recombination between loxP sites. J Mol Biol 1981; AIDS Res Hum Retrovir 1995; 11: 161–170. 150: 467–486.

Gene Therapy Circular DNA with LTR junction is not integrated by HFV integrase RA Russell et al 1332 39 Imrich H, Heinkelein M, Herchenroder O, Rethwilm A. Primate of the European Society of Gene Therapy, 2–4 November, Tur- foamy virus pol proteins are imported into the nucleus. JGen key, 2001. Virol 2000; 81: 2941–2947. 43 Laemmli UK. Cleavage of structural proteins during the 40 Russell RA et al. The R region found in the human foamy virus assembly of the head of bacteriophage T4. Nature 1970; 277: long terminal repeat is critical for both gag and pol protein 680–685. expression. J Virol 2001; 75: 6817–6824. 44 Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of 41 Gu H, Zou YR, Rajewsky K. Independent control of immuno- proteins from polyacrylamide gels to nitrocellulose sheets: pro- globulin switch recombination at individual switch regions evi- cedure and some applications. Proc Natl Acad Sci USA 1970; 76: denced through Cre-loxP-mediated gene targeting. Cell 1993; 73: 4350–4354. 1155–1164. 45 Tosswill JHC, Taylor GP, Clewley JP, Weber JN. Quantiﬁcation 42 Neil S, Martin P, Ikeda I, Collins M. Maturation-dependent of proviral DNA load in human T-cell leukaemia virus type I transduction of macrophages by HIV-based vectors. 9th Meeting infections. J Virol Meth 1998; 75:21–26.

Gene Therapy