Department of Immunology, Windeyer Building, 46 Cleveland St, London WIP 6DB, UK

Correspondence to: M KL Collins

Many retroviral vectors based on murine leukaemia virus (MLV) contain the first 420 nucleotides of the gag gene, as this was reported to increase vector titre by increasing the efficiency of RNA packaging. In this study, deletion of this gag sequence from its original location did not decrease the titre of two retroviral vectors, pBabe puro and MFG-S^-. The two vectors could be improved by replacing the gag sequence with a CTE from Mason-Pfizer monkey virus (MPMV). This substitution improved vector titre, while eliminating a region of homology between vector and packaging constructs. Gene Therapy (2000) 7, 300-305.

retrovirus; retroviral vector; RNA packaging

MLV-based retroviral vectors are the most widely used gene delivery vehicles in clinical gene therapy protocols.¹ The viral components of these vectors comprise trans-complementing genes (gag, pol and env) expressed in packaging cells and the cis-acting sequences (LTRs, primer binding sites and packaging signals) present within the vector. Inclusion of part of the gag gene sequence in vectors was shown to increase recombinant viral titre.^2,3 This gag sequence was reported to increase the amount of vector packaged into virions and was therefore thought to be an extension of the packaging signal.^3,4 However, retention of the gag sequence in vectors is not desirable as it could aid recombination in the packaging cells, leading to generation of replication-competent viruses.

Replication of retroviruses requires export of unspliced genomic RNA from nuclei to cytoplasm. The lentiviruses, such as HIV, encode a trans-acting regulatory protein (Rev), which binds to a cis-acting Rev responsive element (RRE) to facilitate the export of the unspliced genomic RNA.^5,6 Although the simple retroviruses do not encode a Rev-like protein, a cis-acting element, termed constitutive transport element (CTE), has been identified in MPMV, SRV and RSV.^7,8,9,10 These CTEs are believed to interact with cellular proteins to enhance export of genomic RNA.^11,12,13 It has been reported that such CTEs can substitute for RRE/Rev in HIV and SIV vectors.^7,10,14,15 In this study we demonstrate that the first 420 nucleotides of gag sequence in MLV can increase both cytoplasmic RNA level and packaging efficiency of a vector genome. Its substitution by the MPMV CTE results in MLV vectors with an increased titre (up to 10-fold higher).

To analyse the function of the 420 nucleotide gag sequence in MLV vectors, we deleted this from both pBabe puro and MFG-S^- as detailed in Figure 1. The titre of pBabe gag (1-2 ´ 10⁵ IU/ml) was slightly increased in comparison with the parental vector (Figure 2a) and that of MFG gag (6-8 ´ 10⁵ IU/ml) was the same as the parental vector. These titres were determined on human TE671 cells and are therefore lower than the reported titres for the parental vectors on 3T3 cells. These results do not agree with those of Bender et al³ and Morgenstern and Land¹⁶ who showed that inclusion of the gag sequence increased vector titre by five- to 10-fold. However, Kim et al¹⁷ have previously reported that removal of the gag sequence from MFG does not decrease vector titre. To examine possible functions of the gag sequence further, we also moved it close to the 3' LTRs of the vectors, resulting in pBabe Cgag (Figure 1a) and MFG Cgag (Figure 1b). The titres of these vectors were three- to four-fold higher than the parental vectors (Figure 2). Inclusion of the gag sequence in reverse orientation (MFG Cgag R, Figure 1b) did not increase vector titre (Figure 2b).

To investigate the mechanism of action of the gag sequence, we performed Northern hybridisation analysis on both cytoplasmic polyadenylated and packaged viral RNA, using a puro fragment as a probe. The deletion of the gag sequence in pBabe puro resulted in a reduction of unspliced genomic RNA accumulated in the cytoplasm (Figure 3a). However, an increased amount of genomic RNA was packaged (Figure 3b), in agreement with the increase in titre (Figure 2a). This suggests that deletion of gag sequence from its original position significantly increases RNA packaging efficiency. Insertion of the gag sequence adjacent to the 3' LTR led to a higher level of genomic RNA accumulated in the cytoplasm in comparison with pBabe gag (Figure 3a) and a small increase in cytoplasmic genomic RNA for MFG C gag (Figure 4a). Furthermore, an increased amount of genomic RNA was detected in virus particles (Figure 3b) in agreement with the further increase in titre (Figure 2a). These results demonstrate that transfer of the gag sequence adjacent to the 3' LTR augmented vector titre, both by promoting RNA accumulation in the cytoplasm and also by increasing packaging efficiency of RNA into virus particles.

A 240 nucleotide sequence from MPMV has been shown to promote the nuclear export of unspliced RNA.^7,8,11 We therefore examined whether the 240 nucleotide MPMV CTE could enhance MLV titre (Figure 1). This sequence shares no homology with the MLV packaging constructs, or human endogenous or infectious retroviruses. It is within the non-coding region of MPMV and its longest potential open reading frame is 24 amino acids. In pBabe puro, the MPMV CTE enhanced vector titre three- to six-fold when substituted in the original gag position (pBabe NCTE, Figure 2a) or 10-fold when inserted adjacent to the 3' LTR (pBabe CCTE, Figure 2a). In MFG a two- to three-fold enhancement of titre was observed when the MPMV CTE was inserted in either position, and insertion in the reverse orientation did not enhance titre (Figure 2b, MFG NCTE, MFG CCTE and MFG CCTER).

The MPMV CTE has been identified as a cis-acting nuclear export element in MPMV or HIV.^17,18 We therefore examined whether it increased MLV cytoplasmic RNA. Northern hybridisation analysis showed that insertion of the MPMV CTE, either replacing the gag sequence or adjacent to the 3' LTR, promoted the accumulation of unspliced genomic RNA in the cytoplasm. The origin of the RNA species intermediate in size between the genomic and puro RNAs in the CTE constructs is not clear (Figure 3a); it is not included in genomic RNA quantification and was not detected in packaged RNA (Figure 3b). A six-fold increase of RNA was accumulated in the cytoplasm in comparison with pBabe puro, which corresponded to a 15-fold increase in comparison with the parental pBabe gag (Figure 3a). Again a modest increase in cytoplasmic genomic RNA was observed when the MPMV CTE was inserted in MFG (Figure 4a). Because the CTE promoted accumulation of unspliced genomic RNA in the cytoplasm, it also increased the amount of RNA packaged in virions (Figure 3b, pBabe NCTE) in agreement with the observed increase in titre (Figure 2). EGFP expression in cells infected with the MFG-based viruses is shown in Figure 4b. Improved expression compared with parental MFG was detected with the gag, NCTE and CCTE vectors. The combination of improved titre and EGFP expression makes the MFG-NCTE the most attractive for future use.

In this study, deletion of the MLV gag sequence from the commonly used pBabe and MFG-S^- vectors did not decrease viral titres. This finding appears to contradict previous studies in which viral titres or RNA packaging were compared in vectors with or without the gag sequence.^2,3,4,16 However, there is another recent report showing that deletion of the gag sequence from MFG did not affect vector titre.¹⁷ The reason for the discrepancies between results remains unclear. It is possible that use of the newer packaging cell lines^19,20,21 which express higher levels of viral proteins, eliminates effects of the gag sequence.

The gag sequence has a modest (2.5-fold) effect on accumulation of vector RNA in the cytoplasm, when included in the pBABE vector. This may partly explain its observed effect on RNA packaging in a previous study, where total cellular RNA and packaged RNA were measured.³ Significantly, this MLV gag sequence has also been reported to allow rex-independent expression from HTLV-1-based vectors, by enhancing nuclear export.²² However, deletion of the gag sequence does not eliminate RNA export or viral titre, as is the case for previously characterised retroviral CTEs.^9,18 If MLV contains a classic CTE, it must lie within sequences which cannot easily be deleted because they contain LTR, primer binding or packaging functions. In our study, the gag sequence can also enhance packaging of cytoplasmic RNA, but only when it is present at the 3' end of the vector genome. This demonstrates that it does not have to be part of a continuous packaging signal extending from and that it therefore may not be involved in genome dimerisation and nucleocapsid binding.²³

We have replaced the gag sequence with the MPMV CTE to produce improved versions of the pBabe puro and MFG-S^- vectors. The increase in titre obtained by this inclusion of the MPMV CTE suggests that nuclear export of parental vector RNA is suboptimal. Even the modest increase in titre of these new vectors, three-fold in the case of MFG-S^- with the MPMV CTE, will extend the potential applications of retroviral vectors, which are often limited by low titre. In the case of pBabe puro, removal of the gag sequence also allows alignment of the start codon of the transgene with that of gag, which should enhance transgene expression as an internal ribosome entry site facilitates gag translation.²⁴ The new vectors will also allow elimination of any sequence homology between vectors and packaging constructs, eliminating the risk of recombination in packaging cells. This is particularly valuable for packaging systems which generate retroviral vectors by transient transfection,^21,25 as multiple copies of vector and packaging constructs are present within a single cell which increases the risk of recombination. It will also be crucial for the preparation of high titre viruses for clinical use.

Acknowledgements

This study was supported by an MRC Programme Grant, number G9314271.

1 Anderson WF. Human gene therapy. Nature 1998; 392: 25-30, Article MEDLINE

2 Armentano D et al. Effect of internal viral sequences on the utility of retroviral vectors. J Virol 1987; 61: 1647-1650, MEDLINE

3 Bender MA, Palmer TD, Gelinas RE, Miller AD. Evidence that the packaging signal of Moloney murine leukaemia virus extends into the gag region. J Virol 1987; 61: 1639-1646, MEDLINE

4 Adam MA, Miller AD. Identification of a signal in a murine retrovirus that is sufficient for packaging of nonretroviral RNA into virions. J Virol 1988; 62: 3802-3806, MEDLINE

5 Fischer U et al. Evidence that HIV-1 Rev directly promotes the nuclear export of unspliced RNA. EMBO J 1994; 13: 4105-4112, MEDLINE

6 Nasioulas G et al. Elements distinct from human immunodeficiency virus type 1 splice sites are responsible for the Rev dependence of env mRNA. J Virol 1994; 68: 2986-2993, MEDLINE

7 Bray M et al. A small element from the Mason-Pfizer monkey virus genome makes human immunodeficiency virus type 1 expression and replication Rev-independent. Proc Natl Acad Sci USA 1994; 91: 1256-1260, MEDLINE

8 Ernst RK, Bray M, Rekosh D, Hammarskjold ML. A structured retroviral RNA element that mediates nucleocytoplasmic export of intron-containing RNA. Mol Cell Biol 1997; 17: 135-144, MEDLINE

9 Ogert RA, Lee LH, Karen LB. Avian retroviral RNA element promotes unspliced RNA accumulation in the cytoplasm. J Virol 1996; 70: 3834-3843, MEDLINE

10 Zolotukhin AS, Valentin A, Pavlakis GN, Felber BK. Continuous propagation of RRE(-) and Rev(-) RRE(-) human immunodeficiency virus type 1 molecular clones containing a cis-acting element of simian retrovirus type 1 in human peripheral blood lymphocytes. J Virol 1994; 68: 7944-7952, MEDLINE

11 Pasquinelli AE et al. The CTE of MPMV accesses a cellular mRNA export pathway. EMBO J 1997; 16: 7500-7510, Article MEDLINE

12 Saavedra C, Felber BK, Izaurralde E. The simian retrovirus-1 constitutive transport element, unlike the HIV-1 RRE, uses factors required for cellular mRNA export. Curr Biol 1997; 7: 619-628, MEDLINE

13 Tang H et al. A cellular cofactor for the constitutive transport element of type D retrovirus. Science 1997; 276: 1412-1415, MEDLINE

14 Tabernero C et al. Identification of an RNA sequence within an intracisternal-A particle element able to replace Rev-mediated postranscriptional regulation of human immunodeficiency virus type 1. J Virol 1997; 71: 95-101, MEDLINE

15 von Gegerfelt A, Felber BK. Replacement of post-transcriptional regulation in SIVmac239 generated a Rev-independent infectious virus able to propagate in rhesus peripheral blood mononuclear cells. Virology 1997; 232: 291-299, MEDLINE

16 Morgenstern JP, Land H. Advanced mammalian gene transfer: high titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 1992; 18: 3587-3596,

17 Kim SH et al. Construction of retroviral vectors with improved safety, gene expression, and versatility. J Virol 1998; 72: 994-1004, MEDLINE

18 Rizvi TA, Lew KA, Murphy EC Jr. Schmidt RD. Role of MPMV CTE in the propagation of MPMV vectors by genetic complementation using homologous/heterologous env genes. Virol 1996; 224: 517-532,

19 Cosset F-L et al. Higher-titre packaging cells producing recombinant retroviruses resistant to human serum. J Virol 1995; 69: 7430-7436, MEDLINE

20 Danos O, Mulligan RC. Safe and efficient generation of recombinant retroviruses with amphotropic and ecotropic host range. Proc Natl Acad Sci USA 1988; 85: 3539-3543,

21 Soneoka Y et al. A transient three-plasmid expression system for the production of high titre retroviral vectors. Nucleic Acids Res 1995; 23: 628-633, MEDLINE

22 King J et al. The extended packaging sequence of MoMLV contains a constitutive mRNA nuclear export function. FEBS Lett 1998; 434: 367-371, MEDLINE

23 Prats A et al. Cis elements and transacting factors involved in dimmer formation of murine leukaemia virus RNA. J Virol 1990; 64: 774-783, MEDLINE

24 Berlioz C, Darlix J-L. An internal ribosomal entry mechanism promotes translation of MLV gag polyprotein precursors. J Virol 1995; 69: 2214-2222, MEDLINE

25 Pear W. Nolan G, Scott M, Baltimore D. Production of high titre helper-free retroviruses by transient transfection. Proc Natl Acad Sci USA 1993; 90: 8392-8396, MEDLINE

26 Brody BA, Rhee SS, Hunter E. Postassembly cleavage of a retroviral glycoprotein cytoplasmic domain removes a necessary incorporation signal and activates fusion activity. J Virol 1994; 68: 4620-4627, MEDLINE

Figure 1 pBabe (a) and MFG (b) based vectors, showing the location of gag and MPMV CTE sequences. To generate pBabe gag, the polymerase chain reaction (PCR) was used to produce an SpeI-BamHI fragment from pBabe puro¹⁶ spanning the packaging signal and the mutated ATG at nucleotide 955 with a BamHI site introduced at the 3' end. The Spel-BamHI fragment was then used to replace the SpeI-BamHI fragment of pBabe puro, resulting in a deletion of 420 bp of gag sequence. Nucleotides 8022-8240 of Mason-Pfizer monkey virus (MPMV) with BamHI and EcoRI ends were amplified by PCR from plasmid pTMOwt, kindly provided by Eric Hunter.²⁶ This CTE fragment was cloned into pBabe gag, replacing the gag sequence to generate pBabe NCTE. The same CTE fragment, generated with ClaI ends by PCR, was inserted into the pBabe gag plasmid next to the vector 3' LTR, resulting in pBabe CCTE. In a similar way, pBabe Cgag was constructed containing a 420 bp gag fragment, amplified by PCR with ClaI ends from pBabe puro. To construct MFG based vectors, a EGFP fragment, based on pEGFP-Cl (Clontech, Basingstoke, UK), was inserted into MFG-S^- (Somatix Therapy Corporation, Alameida, CA, USA) between the NcoI-BamHI sites, resulting in MFG-EGFP. MFG-EGFP was then digested with ApaI and BglII, the DNA was repaired with S1 nuclease and Klenow polymerase and ligated to generate MFG-EGFP gag. The CTE fragments were then introduced into the MFG-EGFP gag vector, either replacing gag or adjacent to the 3' LTR, resulting in MFG-EGFP NCTE and MFG-EGFP CCTE, respectively. Insertion of the gag fragment adjacent to the 3' LTR resulted in MFG-EGFP Cgag. The construction of all the plasmids was verified by sequence analysis using an ABI system.

Figure 2 Effects of gag and MPMV CTE sequences on vector titre. Vector DNA was transfected into the amphotropic packaging cell line, TEFLYA8¹⁹ using Lipofectamine (Gibco, Paisley, UK), in the presence of 1/10 molar amount of pSV2neo (Clontech) in the case of the MFG derivatives. After 48 h, the producer cells were selected for 14 days by the addition of puromycin (1 g/ml; Sigma, Poole, UK) or G418 (1 mg/ml; Gibco) for pBabe or MFG vectors, respectively. A minimum of 200 colonies was pooled, grown to confluence then incubated overnight in OptiMEM (Gibco). The virus-containing supernatant was filtered through 0.45 m filters (Gelman Sciences, Ann Arbor, MI, USA). Human TE671 cells were infected for 5 h with serially diluted viruses in OptiMEM containing polybrene (8 g/ml; Sigma). For each vector, virus titre was determined from triplicate infection with viruses harvested from two sets of transfections. For pBabe vectors, the infected cells were selected with puromycin (1 g/ml) for 14 days and resistant colonies were counted (a). For MFG derived vectors, the infection efficiency was measured after 48 h infection on a FACScan (Becton Dickinson, Heidelberg, Germany). The percentage of cells expressing EGFP was measured and the viral titre and mean EGFP expression in infected cells was calculated (b).

Figure 3 Effects of gag and MPMV CTE sequences on cytoplasmic RNA level and packaging of pBabe puro vectors. (a) Cytoplasmic RNA was isolated by the guanidine isothiocyanate/caesium chloride method, then mRNA was isolated using a poly(A) Quick kit (Stratagene Europe, Amsterdam, The Netherlands). A 500 bp puro fragment was labelled with ³²P dCTP using the Prime-a-Gene system (Promega, Southampton, UK) and used to hybridise to the RNA samples resolved on a denaturing agarose gel, then transferred to nitrocellulose. A GAPDH fragment was used as a loading control. The relative amount of virus genomic RNA to GAPDH was determined by densitometry of autoradiographs; this ratio was set to 1.0 for pBabe puro and those of the others were normalised to this. (b) To prepare viral RNA, supernatant was harvested from packaging cells, then viral particles were pelleted by centrifugation at 1500 g for 3 h and were lysed in 4 M guanidium isothiocyanate solution. DNA and proteins were extracted with phenol/chloroform in the presence of 0.2 M sodium acetate (pH 4.0). Viral RNA was then concentrated by ethanol precipitation with yeast tRNA (50 g) as a carrier. RNA was analysed by Northern hybridisation with a puromycin probe (upper panel). 1/20 volume of the same supernatant was analysed by SDS-PAGE and Western blot with an anti-Gag antibody (lower panel).

Figure 4 Effects of gag and MPMV CTE sequences on MFG vectors. (a) poly(A)⁺ RNA was isolated from bulk populations of vector-transfected packaging cells. RNA was analysed by Northern hybridisation with an 800 bp EGFP and a GAPDH probe. The relative amount of virus genomic RNA to GAPDH was determined by excision of appropriate areas of membrane and scintillation counting; this ratio was set to 1.0 for MFG and those of the others were normalised to this. (b) Supernatants from bulk populations of vector-transfected packaging cells were used to infect human TE671 cells. Infected cells were defined by FACScan as those within a region that included only 0.02% of uninfected cells, which comprised between 20 and 70% of the infected cell populations. EGFP expression in two independently infected populations was determined, the mean fluorescence ± s.e. from the two infections is shown.