BRIEF COMMUNICATION Engineering the splice acceptor for improved gene expression and viral titer in an MLV-based retroviral vector

Transcription

1 (2004) 11, & 2004 Nature Publishing Group All rights reserved /04 $ BRIEF COMMUNICATION Engineering the splice acceptor for improved gene expression and viral titer in an MLV-based retroviral vector J-T Lee 1,SSYu 2, E Han 2, S Kim 2 and S Kim 1,3 1 Institute of Molecular Biology and Genetics, Seoul National University, Seoul, Korea; 2 ViroMed Co. Ltd., Bongcheon-dong , Kwan-Ak-Gu, Seoul, Korea; and 3 School of Biological Sciences, Seoul National University, Seoul, Korea We have recently developed a retroviral vector that contains a splice acceptor from the human EF1-a gene and drives a significantly higher level of gene expression than other well known murine leukemia virus-based vectors. However, one downside of this vector is that viral titer significantly varies depending on the packaging lines used. Results from Northern blot analysis indicated that in certain cell lines the genomic transcript containing the packaging signal sequence was too efficiently spliced to the subgenomic RNA, resulting in low levels of genomic RNA and thus leading to a low viral titer. We tested the possibility of overcoming this problem by introducing mutations around the splice acceptor sequence in such a way that a delicate balance was maintained between the splicing efficiency (which determines the level of gene expression) and the amount of genomic transcript (which influences viral titer). After mutational analysis, one such mutant was found to meet this requirement. The newly constructed vector containing the engineered splice acceptor could indeed drive higher levels of expression in many therapeutic genes than other control vectors, without significantly compromising viral titer. (2004) 11, doi: /sj.gt Keywords: retroviral vector; splicing; engineered splice acceptor; level of gene expression Murine leukemia virus (MLV)-based vectors are still the most frequently used gene delivery vehicles, employed in almost 40% of approved clinical protocols worldwide. 1,2 Despite their frequent use as gene delivery vehicles and the promising therapeutic effects in the SCID-X1 trial case, 3 there are still areas that need to be improved upon in order for MLV-based vectors to be effective in actual clinical settings. 4 9 First, many of these vectors contain sequences that are also present in the packaging lines. 6,8,10 A recombination between the packaging genome and the vector can result in the generation of replication-competent retrovirus (RCR) Second, the level of gene expression in vivo driven by MLV-based vectors, for reasons not fully understood yet, is not high enough to give clear therapeutic effects in most cases. Furthermore, the viral titers achieved by the vectors, although improving, are generally too low to deliver the gene to a therapeutically significant number of cells. We have recently reported a minimum-sized MLV vector, MT, which does not contain any of the viral coding sequences. In an effort to further improve the level of gene expression, a portion of the intron and 5 0 untranslated exon sequences from the human EF1-a gene were introduced to this vector, which was indeed found to increase significantly the level of gene expression by producing a large amount of spliced RNAs. However, Correspondence: Dr S Kim, Institute of Molecular Biology and Genetics, Seoul National University, Kwan-Ak-Gu, Seoul, Korea Received 04 October 2002; accepted 07 July 2003 viral titer of this retroviral vector widely varied depending on the packaging cell lines used. 1 RNA analysis suggested that viral titer was decreased because genomic sized RNA containing the packaging signal sequence was too efficiently spliced to subgenomic RNA, lowering the amount of packageable RNA. This may restrict the use of this type of vector to specific cell lines, narrowing the range of its application in the actual industrial setting. In this experiment, we attempted to alter the splicing efficiency by introducing mutations around the splice acceptor sequence The goal was to find a mutation(s) that could decrease the splicing efficiency to the extent that an acceptable viral titer was obtained, while the level of gene expression was still maintained at a reasonably high level. After mutational analysis, one such mutation was found. When a splice acceptor containing this mutation was introduced into the retroviral vector, viral titer was increased to a normal level while still producing high levels of gene expression. MINEI was originally derived from retroviral vector MT that does not contain any viral coding sequences unlike other vectors available at the time of its development 1 (Figure 1). The immediate parental vector of MINEI is MIN, which contains an internal ribosomal entry site (IRES) from the EMCV and the bacterial neo gene as a selectable marker. In MINEI, the splice acceptor site from the human EF1-a gene (nucleotide number, þ 772 to þ 1008), consisting of a portion of the intron and 5 0 untranslated exon sequences, 1 was placed in such a way that the gene of interest is to be expressed mainly

2 95 Figure 1 Schematic representation of retroviral vectors used in this study. MT is a retroviral vector that does not contain any viral coding sequences. 1 MIN contains an IRES from the EMCV and the bacterial neo gene as a selectable marker. MINEI includes the portion of intron and splice acceptor region from the human EF1-a gene (nucleotide number, þ 772 to þ 1008). 1 DFG, containing both IRES and the bacterial neo gene, is a derivative of MFG. 16 MFG contains viral coding sequences: 420 bp for gag, 377 for pol and 99 for env. 5,7,9 MIN1, MIN2, MIN3 and MIN5 retroviral vectors were derived from MINEI by introducing various mutations as indicated. To construct these mutants, both the intron and exon of EF1-a presented in MINEI were amplified using MINEI as a template. To construct MIN1, EI5 (ACGCGTGTCGAGCTTTTGGAGTACGTCGTCTTTAGGTT) and EI3-1s (GGATCCGTTTAAACACCTGAAATGGAAGAAAAAAACTT TGAA) primers were used. Amplified fragment was initially cloned into pgem T easy (Promega, WI, USA). The MluI BamHI fragment was cloned into the MluI BamHI site of MIN, resulting in MIN1. To construct MIN2, EI5 (ACGCGTGTCGAGCTTTTGGAGTACGTCGTCTTTAGGTT) and EI3-2s (GGATCCGTTTAAACGCCTGAAATGGAAGAAAAAAA CTTTGAA) primers were used. Amplified fragment was initially cloned into pgem T easy. The MluI BamHI fragment was cloned into the MluI BamHI site of MIN, resulting in MIN2. To construct MIN3, EI5m (ACGCGTGTCGAGCTTTTGGAGTACTGCGTCTTTAGGTT) and EI3-1s (GGATCCGTTTAAACACCTGAAATGGAAGAAAAAAACTTTGAA) primers were used. Amplified fragment was initially cloned into pgem T easy. The MluI BamHI fragment was cloned into the MluI BamHI site of MIN, resulting in MIN3. To construct MIN5, EI5 (ACGCGTGTC GAGCTTTTGGAGTACGTCGTCTTTAGGTT) and EI3-2l (TTTTTGG CTTTTAGGGGTAGTTTTCACGACGGCTGAAATGGA) primers were used. Amplified fragment was initially cloned into pgem T easy. The MluI T4 polymerase-treated SacII fragment was cloned into the MluI PmeI site of MIN, resulting in MIN5. as a spliced message. Figure 1 also shows control vector DFG derived from MFG containing the splice acceptor of MLV itself present in the pol coding region. 17 When the bacterial CAT gene was used as a reporter gene, in all three packaging cell lines used in this study, MINEI produced three- to eight-fold higher levels of CAT activity than its parental vector MIN lacking the splicing acceptor (Figure 2a). In addition, it was reproducibly found that MINEI drove a higher level of gene expression than control vector DFG. However, the viral titer of MINEI varied depending on packaging cell lines used in this study. It was comparable to other vectors in 293T-based Phoenix cells, but significantly lower in FLYA13 and PG13 cells (Figure 2b). We previously reported that low viral titer of MINEI in these cell lines resulted from highly efficient splicing of Figure 2 Comparison of various vectors using different packaging cell lines. To compare the performance of MIN and MINEI in various packaging lines, respective viruses were prepared using Phoenix Ampho, FlyA13 and PG13 packaging lines, and used to compare their viral titer using DFG as a control. It is to be noted that the methods employed for measuring the levels of gene expression and viral titer for PG13 were different from those used for Phoenix and FlyA13. This was due to the low transfection efficiency of PG13 cells. (a) Comparison of the level of gene expression. Retroviral vectors were transfected to Phoenix Ampho and the FlyA13 packaging line, respectively, using FuGene6 (Roche, Germany) according to the manufacturer s instructions. The levels of CAT activity of transfected cells were determined 2 days after transfection by a standard procedure as previously described by Yu et al. 1 To determine the level of gene expression in the PG13 packaging line, 293T cells were transfected with retroviral vectors, together with amphotropic packaging constructs, pvm-gp and pvm-ae, 25 and the cell-free viral supernatants from the transfected cells were used to transduce the PG13 cells at a low MOI of around 0.1. After drug-resistant populations were obtained, the levels of CAT activity of the respective producer lines were determined. All the assays were performed by two different investigators, and each investigator performed three independent experiments. The mean values of six experiments are shown and the error bar indicates s.d. The s.d. of six experiments were lower than 22%. (b) Comparison of viral titer. Cell-free viral supernatants from the transfected Phoenix Ampho and the FlyA13 cells were collected respectively (usually 48 h after transfection), and used to transduce NIH3T3 cells plated at on a 60 mm dish the previous day. To determine viral titer, serially diluted viral supernatants were added to NIH3T3 in the presence of 8 mg/ml polybrene. The next day, transduced cells were transferred to 100 mm plates and selected in the presence of G418 until visible colonies were formed. Viral titer was determined by counting the number of drug-resistant colonies. In the case of PG13, transduced cells were plated at cells per 60 mm dish and, 3 days later, cell-free viral supernatants were prepared and used to transduce HT1080 cells to determine viral titer as described above. The s.d. of six experiments were lower than 20%, except in the case of DFG. The s.d. of DFG in Phoenix cells were up to 30%. the genomic transcript due to the presence of the heterologous splice acceptor. 1 Based on this finding, it was hypothesized that the splice acceptor site might be engineered to make its function weakened to an extent that viral titer was increased to the level of control vectors, while still maintaining a higher level of gene expression. The universally conserved sequence elements of splicing sites are the sequence (5 0 )GU, (3 0 )AG,

3 96 polypyrimidine tract and a branchsite in vertebrate. 14,18 21 Nucleotide sequences neighboring these conserved sequences appear to have a certain pattern, found at frequencies higher than expected from a random distribution. 14,18 Since the changes in the conserved sequence elements drastically alter the RNA processing pattern in general, we mutated the neighboring sequences expecting to obtain the above characteristics Mutations were introduced around the splice acceptor region as indicated in Figure 1, and tested for their effects on viral titer, the level of gene expression and the splicing pattern. For the simplicity of the assay, the bacterial CAT sequence was initially used as a reporter gene, which is coexpressed with the neo selectable marker as a bicistronic message. 293T cells were transfected with mutant constructs (MIN1, MIN2, MIN3 and MIN5), MINEI and two control vectors (MIN and DFG), together with amphotropic packaging constructs, pvm-gp and pvm-ae. 25 Cell-free viral supernatants were used to transduce PG13 cells at an MOI of o0.1, followed by G418 selection. Drugresistant populations were obtained and the levels of CAT activity were determined. As transduced cells were thought to contain on average a single copy of the integrated vector, the level of CAT activity would reflect the relative level of gene expression from a specific vector construct in transduced cells. In this assay, the cell population, instead of subclone, was purposely analyzed because it would reflect the overall characteristics of the given retroviral vector. At 3 days after seeding transduced PG13 cells on a 60 mm dish, culture supernatants were taken and used to determine viral titer in HT1080 cells. As shown, MINEI and all mutant constructs gave approximately an eight-fold higher level of gene expression in the PG13 packaging line than MIN did, and two to three times higher than DFG did (Figure 3a). (Note that for a convenient comparison between vectors, the levels of CAT activity from various vectors were standardized by that from MIN.) Interestingly, viral titer from MIN5 was found to be comparable to that from its parental construct MIN (Figure 3b), whereas other mutant constructs (MIN1, MIN2 and MIN3) produced low viral titers in PG13 cells, similar to results obtained from MINEI (Figure 3b). These results suggested that MIN5 not only maintained a high level of gene expression but also produced an acceptable viral titer. Next, we analyzed the quality and quantity of viral RNAs produced from various vectors in PG13 cells. Total RNAs were prepared and analyzed by Northern blot hybridization using the CAT sequence as a probe (Figure 4). MIN produced the genomic transcript alone because it did not have a splice acceptor sequence (Figure 4, lane 1), while MINEI and mutant constructs generated two RNA species, a genomic transcript and a spliced subgenomic transcript (Figure 4, lanes 2 6). As expected, MINEI produced a large amount of spliced RNAs relative to the genomic RNA (Figure 4, lane 5). It was clear that most of the genomic RNAs made in this vector were spliced, thus resulting in a high level of CAT activity but low viral titer. Three mutant constructs (MIN1, MIN2 and MIN3) showed a similar pattern in the ratio of genomic to subgenomic RNAs (Figure 4, lane 2 4), while MIN5 produced a significantly higher level of genomic transcript than MINEI (Figure 4, lane 6). These results suggested that mutations introduced in MIN5 changed the splicing efficiency, increasing viral titer while still maintaining a high level of gene expression. Figure 3 Effect of various mutations introduced in MINEI on viral titer and the level of gene expression in PG13 packaging cell lines. As described in the text, MINEI-derived mutant vectors and control vectors were transfected to 293T cells together with amphotropic packaging constructs VM-GP and VM-AE. After 2 days, cell-free viral supernatants were prepared and used to transduce PG13 packaging lines at an MOI of o0.1, followed by G418 selection. After a drug-resistant population was obtained, the levels of CAT activity (a) and viral titer (b) were measured as described in Figure 2. The mean values of six experiments are presented with s.d. value of o16% (a), and o10% (b), respectively. Titer differences between MINEI and MIN5 vectors are highly significant (P¼0.0102). Figure 4 RNA blot analysis of PG13 packaging cells transduced with MIN, MINEI and the newly constructed mutant vectors. Total cellular RNAs were prepared using the guanidine thiocyanate cesium chloride method. The total RNA (20 mg) was subjected to 1% formaldehyde agarose gel electrophoresis, blotted to nitrocellulose membrane (Hybond-C; Amersham, RPN303W, USA), and hybridized with a 32 P-labeled CAT probe (the BamHI CAT fragment of MIN-CAT). Cellular b-actin probe was used as a loading control. G, genomic transcript; SG, subgenomic RNA; A, b-actin RNA.

4 To test whether MIN5 maintained its characteristics in other cell lines, we repeated a similar experiment in 293T cells. MINEI, four mutant constructs and two control vectors (MIN and DFG) were transfected to 293T cells together with amphotropic packaging constructs pvm- GP and pvm-ae. After 2 days, the levels of CAT activity in transfected cells were measured. This would represent the level of gene expression produced by a given retroviral construct in 293T cells. To determine viral titer, culture supernatants were used to transduce the NIH3T3 cells. When 293T cells were used, there was a difference in the level of gene expression between MINEI and the mutant constructs, unlike in the case of PG13 cells. The level of CAT activity from MIN5 was two- to three-fold lower than that from MINEI, but still higher than that from MIN and DFG (Figure 5a). The viral titer of MIN5 was comparable to that of MIN and DFG (Figure 5b). These results suggested that MIN5 would be useful even when 293T cells were used as a packaging cell line. To test whether our observation was restricted to the bacterial CAT gene, we used the gene encoding the human interleukin 1 receptor antagonist protein (IRAP) or the glucocerebrosidase (GC) in both PG13 and 293T cells. These two cdnas were cloned into MIN5 as well as three other control vectors, DFG, MIN and MINEI, which were then compared for their viral titer and the level of gene expression. As shown in Table 1, MINEI produced the highest level of IRAP protein but the lowest viral titer in both PG13 and 293T cells. However, when MIN5 was used, viral titer was increased by nineand four-fold in PG13 and 293T cells, respectively. MIN5 could also produce a high level of IRAP protein in both cell lines, consistent with the above results (Table 1). The performance of MIN5 was more prominent when the GC gene was used. MIN5 produced the highest level of GC activity and highest viral titer in both PG13 and 293T cells. Next, we tested MIN5 in two different cell types that were correlated with human gene therapy, human erythroleukemia line K562 and primary skin fibroblasts isolated from a patient with Gaucher disease. Cells were transduced with 0.5 ml of viral supernatant of PG13 producer lines of IRAP and GC, respectively. After 2 days, the levels of IRAP protein and GC activity were measured in respective transduced cells. The level of gene expression in this assay would represent viral titer as well as the level of gene expression driven by the respective vector. As clearly shown in Table 2, MIN5 drove the highest level of gene expression in K562 cells as well as in primary skin fibroblasts. These data Table 1 Performances of MIN5 expressing human proteins Vectors Relative level of IRAP/GC Viral titer PG13 293T PG13 293T IRAP MIN MINEI MIN DFG GC MIN MINEI MIN DFG The experimental conditions were identical to those described in Figure 3 for PG13 and Figure 5 for 293T. The level of IRAP protein was measured from cell culture supernatant using a commercially available kit from R & D Systems Inc (Minneapolis, MN, USA; DRA00). The data represent the mean values of six experiments. 7indicates s.d. The level of GC activity was measured from cell lysates by the method of Ohashi et al, 32 the mean values of three experiments are presented. The expression and viral titer from MIN was set to 1. Table 2 Performances of MIN5 in transduced target cells Vectors Relative level of IRAP Relative level of GC K562 Primary skin fibroblasts Figure 5 Comparison of mutant vectors in 293T cells. MINEI-derived mutant vectors and control vectors were transfected to 293T cells together with amphotropic packaging constructs. The level of CAT activity was determined 2 days after transfection (a). At the same time, culture supernatant was prepared and used to transduce NIH3T3 cells to determine viral titer (b). The s.d. values of three experiments were lower than 20% in the level of gene expression and lower than 34% in the case of viral titer. MIN MINEI MIN DFG Culture supernatants from PG13 producer lines of IRAP and GC were used to transduce K562 cells and primary skin fibroblasts of a patient with Gaucher disease, respectively. Target cells were transduced with 0.5 ml of viral supernatant in the presence of 8 mg/ml of polybrene. The level of gene expression was measured as described in Table 1, and the mean values of three experiments are presented.

5 98 Figure 6 RNA blot analysis of K562 cells transduced with retroviral vectors DFG, MIN, MINEI and MIN5. The experimental conditions were identical to those described in Figure 4, except that RNA (20 mg) was hybridized with a 32 P-labeled Neo probe isolated from MIN-CAT. G, genomic transcript; SG, subgenomic RNA; cs, cryptic spliced RNA; A, b- actin RNA. demonstrated that MIN5 could drive a high level of gene expression without compromising viral titer, and that our observation was not restricted to a specific reporter gene or cell type. The content of RNAs produced from a retroviral vector expressing IRAP in K562 cells was analyzed. K562 cells were transduced at an MOI of o0.1 followed by G418 selection. Total RNAs were prepared from G418- resistant populations and analyzed by Northern blot hybridization using the neo sequence as a probe (Figure 6). DFG produced two RNA species since it contained the retroviral splice acceptor sequence located in the pol gene (Figure 6, lane 1). The majority of RNA produced from the DFG vector seemed to be genomic transcript. Unexpectedly, MIN vector, which does not contain a splice acceptor site, produced a large amount of spliced RNAs (Figure 6, lane 2). It turned out that the IRAP cdna sequence contains a cryptic splice site present in the coding region that was highly efficiently used, lowering both viral titer and the level of gene expression (data not shown). Consistent with previous results, MINEI produced a large amount of subgenomic RNA of expected size, but a very low level of the genomic transcript. In MIN5, the level of packageable RNA was noticeably increased. These results suggested that MIN5 expressing IRAP behaved as expected in K562 cells, and also that the presence of the heterologous splice acceptor could suppress the use of the cryptic splice site present in the coding region. It has been known that splicing is required for rapid and efficient mrna export, and the introduction of introns can enhance the level of gene expression We and others have explored the possibility of applying this concept to the context of retroviral vectors. For example, MFG, which has proven to be a very efficient retroviral vector in terms of viral titer and the level of gene expression, contains 400 bp of the 3 0 end of the pol coding region. As this region includes the retroviral branch site and splice acceptor for the env message, a proportion of the viral RNA is spliced to a subgenomic RNA, leading to enhanced expression of the transgenes of interest. 30 However, one major concern in using MFG is the possibility of RCR generation, because it contains bits of viral coding sequences including gag and env as well as pol, thus increasing the possibility of homologous recombination between the vector and the packaging genome. 25 Hildinger et al 31 have also reported retroviral vectors containing splice acceptors. In that case, a 48 bp long splice acceptor sequence from murine retrovirus SFFVp was introduced into their SF-derived vectors, resulting in increased expression of genes. However, it must be noted that 26 bp of 48 bp was identical to the pol coding region of MLV, which can again contribute to possible homologous recombination. They suggested that this problem could be overcome by abolishing the residual homology by sequence wobbling or by designing a nonretroviral splice acceptor. In our previous experiment, we demonstrated that QJ;the introduction of a heterologous cellular or viral splice acceptor site to the retroviral vector significantly increased the level of gene expression, but lowered viral titer. 1 In this set of experiments, we introduced mutations around the splice acceptor region to improve the viral titer of the MINEI vector, and our data demonstrated that the splicing efficiency of the mutant MIN5 indeed changed. MIN5 produced a slightly lower level of gene expression than MINEI did, because its splicing efficiency was decreased, but still drove a higher level of gene expression than MIN or DFG did. For the same reason, the viral titer of MIN5 increased 5 10-fold than that of MINEI, and became comparable to that of DFG or MIN in PG13 cells. Moreover, when the GC gene was used, MIN5 drove the highest level of gene expression and viral titer than any other retroviral vectors used in this study. The splicing efficiency and the effect of mutation around the splice site seemed to be influenced by various factors in the context of retroviral vectors. First, the splicing efficiency of the same splicing acceptor could vary in different cell lines. MIN1, MIN2 and MIN3 all showed high levels of gene expression but poor titer in PG13 cells. However, in 293T cells, these mutants produced poor titers and reduced expression. This is probably because the amount of various transacting factors involved in the splicing process varies depending on cell types. It is not clear yet, among the mutants, why MIN3 gave the lowest gene expression and the poorest viral titer in 293T cells. One possibility is that the mutation introduced into the intronic sequence in MIN3 might influence the assembly and/or disassembly of spliceosome. Second, the nature of nucleotide sequence of the inserted gene also seems to affect the splicing efficiency and thus the performance of the retroviral vector. Relative levels of gene expression and relative viral titer were significantly different among CAT, IRAP and GC at a given vector and a given cell line. These data imply that one has to be very careful in choosing an optimum retroviral vector for actual gene therapy. Viral titer and the level of gene expression in transduced cells are two of the key factors influencing the success of gene therapy trials. MIN5 described in this report has features as an efficient vector for gene delivery vehicle because it drives a higher level of gene expression than other currently available vectors in various cell lines, while producing comparable viral titer. Furthermore, it contains no viral coding sequence, thus minimizing the production of replication-competent retrovirus. The MIN5 retroviral vector should be useful as a gene delivery vehicle in gene therapy.

6 Acknowledgements We are grateful to V Narry Kim for critical reading of this manuscript and helpful discussion. This work was supported in part by grants from the Korean Ministry of Commerce, Industry and Energy (Grant no. ND ), the Ministry of Science and Technology (Grant no. MI ) and IVI-Affiliated Lab Program (01-1-1). References 1 Yu SS, Kim JM, Kim S. High efficiency retroviral vectors that contain no viral coding sequences. 2000; 7: Wiley The Journal of Gene Medicine Website ( 3 Cavazzana-Calvo M et al. Gene therapy of human severe combined immunodeficiency (SCID)-X1 disease. Science 2000; 288: Boggs SS et al. Prolonged systemic expression of human IL-1 receptor antagonist (hil-1ra) in mice reconstituted with hematopoietic cells transduced with a retrovirus carrying the hil-1ra cdna. 1995; 2: Bowtell DD, Cory S, Johnson GR, Gonda TJ. Comparison of expression in hemopoietic cells by retroviral vectors carrying two genes. J Virol 1988; 62: Dranoff G et al. Vaccination with irradiated tumor cells engineered to secrete murine GM-CSF stimulates potent, specific and long lasting antitumor immunity. Proc Natl Acad Sci USA 1993; 90: Jaffee EM et al. High efficiency gene transfer into primary human tumor explants without cell selection. Cancer Res 1993; 53 (10 Suppl): Ohashi T et al. Efficient transfer and sustained high expression of the human glucocerebrosidase gene in mice and their functional macrophages following transplantation of bone marrow transduced by a retroviral vector. Proc Natl Acad Sci USA 1992; 89: Palmer TD, Thompson AR, Miller AD. Production of human factor IX in animals by genetically modified skin fibroblasts: potential therapy for hemophilia B. Blood 1989; 73: Miller AD, Roseman GJ. Improved retroviral vectors for gene transfer and expression. Biotechniques 1989; 7: Miller AD, Trauber DR, Buttimore C. Factors involved in production of helper virus free retrovirus vectors. Somat Cell Mol Genet 1986; 12: Muenchau DD et al. Analysis of retroviral packaging lines for generation of replication-competent virus. Virology 1990; 176: Otto E et al. Characterization of a replication-competent retrovirus resulting from recombination of packaging and vector sequences. Hum Gene Ther 1994; 5: Padgett RA et al. Splicing of messenger RNA precursors. Annu Rev Biochem 1986; 55: review). 15 Green MR. Biochemical mechanisms of constitutive and regulated pre-mrna splicing. Annu Rev Cell Biol 1991; 7: review). 16 Carothers AM, Urlaub G, Grunberger D, Chasin LA. Splicing mutants and their second-site suppressors at the dihydrofolate reductase locus in Chinese hamster ovary cells. Mol Cell Biol 1993; 13: Zitvogel L et al. Construction and characterization of retroviral vectors expressing biologically active human interleukin-12. Hum Gene Ther 1994; 5: Keller EB, Noon WA. Intron splicing: a conserved internal signal in introns of animal pre-mrnas. Proc Natl Acad Sci USA 1984; 81: Madhani HD, Guthrie C. Dynamic RNA RNA interactions in the spliceosome. Annu Rev Genet 1994; 28: 1 26 (review). 20 Reed R. Initial splice-site recognition and pairing during premrna splicing. Curr Opin Genet Dev 1996; 6: (review). 21 Will CL, Luhrmann R. Protein functions in pre-mrna splicing. Curr Opin Cell Biol 1997; 9: (review). 22 Hornig H, Aebi M, Weissmann C. Effect of mutations at the lariat branch acceptor site on beta-globin pre-mrna splicing in vitro. Nature 1986; 324: Macklin WB, Gardinier MV, King KD, Kampf K. An AG GG transition at a splice site in the myelin proteolipid protein gene in jimpy mice results in the removal of an exon. FEBS Lett 1987; 223: (letter). 24 Hedley ML, Forman J, Tucker PW. Mutation of 3 0 splice sites in two different class I genes results in different usage of cryptic splice sites. J Immunol 1989; 143: Yu SS et al. Construction of a retroviral vector production system with the minimum possibility of a homologous recombination. 2003; 10: Palmiter RD et al. Heterologous introns can enhance expression of transgenes in mice. Proc Natl Acad Sci USA 1991; 88: Luo MJ, Reed R. Splicing is required for rapid and efficient mrna export in metazoans. Proc Natl Acad Sci USA 1999; 96: Brinster RL et al. Introns increase transcriptional efficiency in transgenic mice. Proc Natl Acad Sci USA 1988; 85: Choi T, Huang M, Gorman C, Jaenisch R. A generic intron increases gene expression in transgenic mice. Mol Cell Biol 1991; 11: Krall WJ et al. Increased levels of spliced RNA account for augmented expression from the MFG retroviral vector in hematopoietic cells. 1996; 3: Hildinger M, Abel KL, Ostertag W, Baum C. Design of 5 0 untranslated sequences in retroviral vectors developed for medical use. J Virol 1999; 73: Ohashi T et al. Characterization of human glucocerebrosidase from different mutant alleles. J Biol Chem 1991; 266: