- Email: [email protected]
Optimization of transcriptional regulatory elements for constructing plasmid vectors
Gene 272 (2001) 149±156
www.elsevier.com/locate/gene
Optimization of transcriptional regulatory elements for constructing plasmid vectors Zhi-Li Xu a,b, Hiroyuki Mizuguchi a,*, Akiko Ishii-Watabe a, Eriko Uchida a, Tadanori Mayumi b, Takao Hayakawa a b
a Division of Biological Chemistry and Biologicals, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan Department of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Osaka University, 1-6 Yamadaoka, Suita, Osaka 565-0871, Japan
Received 26 February 2001; received in revised form 24 May 2001; accepted 29 May 2001 Received by T. Sekiya
Abstract In studies regarding both gene therapy and gene function, transgene expression by plasmid vectors bene®ts from the use of transcriptional regulatory elements which permit high-level gene expression. Therefore, with respect to transgene (luciferase) expression activity both in vitro (using HeLa, HepG2, and ECV304 cells) and in vivo (mouse liver and skeletal muscle), we investigated the effective combination of commonly-used regulatory elements, such as the promoter/enhancer, intron, and polyadenylation signal (P(A)) sequence by constructing a series of plasmids that differed only in the particular sequence element being evaluated. Of the several promoter/enhancers that were tested, hybrid CA promoter/enhancer containing human cytomegalovirus immediate-early 1 gene (CMV) enhancer and chicken b-actin promoter with the b-actin intron sequence, and the improved CMV promoter/enhancer containing the largest intron of CMV (intron A) produced the highest levels of expression both in vitro and in vivo. P(A) sequences were found to have signi®cant effects on transgene expression. The effect of a multiple enhancer was also examined. Optimized plasmids of this study were pCASL3 (composed of CMV enhancer, b-actin promoter, b-actin intron, Simian virus (SV40) P(A) sequence and SV40 enhancer) and pCMVSL3 (composed of CMV enhancer, CMV promoter, intron A, SV40 P(A) sequence and SV40 enhancer). These comparative analyses could provide a systematic reference for the development of vector construction for gene therapy, vaccine development, and gene transfer experiments. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Transgene expression; Gene therapy; Promoter; Enhancer; Intron; Polyadenylation signal sequence
1. Introduction The delivery of exogenous genes into animal cells plays an important role in not only the study of gene function but in gene therapy as well. The expression levels of the introduced gene depend mostly on the strength of transcriptional regulatory elements and the transduction ef®ciency of the gene transfer vector (Mcknight and Tjian, 1986). Currently, great efforts are being made to develop gene transfer vectors, viral and non-viral vectors. Some viral vectors produce the infectious wild type virus during ampli®cation Abbreviations: BGH, bovine growth hormone; Chimeric, hybrid intron of b-globin and immunoglobin; CMV, human cytomegalovirus immediateearly 1 gene; Intron A, the largest intron of human CMV IE gene; mRBG, minimal rabbit b-globin; P(A), polyadenylation signal; PGK, human phosphoglycerate kinase; RSV, Rous sarcoma virus genome long terminal repeats; SV40, Simian virus * Corresponding author. Tel.: 181-3-3700-9089; fax: 181-3-3700-9084. E-mail address: [email protected] (H. Mizuguchi).
and induce an immune response that renders repeated administration problematic (Gunter et al., 1993; Yang et al., 1995). Increasing transgene expression by optimizing the transcriptional regulatory elements is desirable as a way to decrease the amount of viral vector required to achieve a clinical effect. For the non-viral vectors, enhanced expression ef®ciency by using optimal transcriptional regulatory elements should be a practical way of making the vector more ef®cient. Previous reports have described the strength evaluation of promoter/enhancers or other transcriptional regulatory elements. However, almost all have focused on single cell (Pasleau et al., 1985; Martin-Gallardo et al., 1988; Oellig and Seliger, 1990; Yew et al., 1997) or tissue targets (Manthorpe et al., 1993; Hartikka et al., 1996), or on transduction methods (Cheng et al., 1993; Guo et al., 1996). No systematic study which examines the function of the vector in various conditions including cultured cells and animal tissues in vivo has been reported.
0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00550-9
150
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Although various kinds of transcriptional regulatory elements have been used to construct vectors, generally only a few are considered as ubiquitously strong elements and widely used thus far. These include CMV and Rous sarcoma virus genome long terminal repeats (RSV) promoters, SV40 and bovine growth hormone (BGH) P(A) sequences, etc. However, few studies have simultaneously compared the relative strength of these so-called strong elements in various types of cell or tissue. We report here the results from a series of experiments designed to identify the relative strength of these commonly-used transcriptional regulatory elements both in vitro and in vivo. In vitro, three cell lines, HeLa from human cervical carcinoma, HepG2 from human hepatoma, and ECV304 human vascular endothelial cell from umbilical cord vein, were used. In vivo, mouse liver and skeletal muscle were transfected. All these cells or tissues are important targets for gene therapy and gene transfer experiments. Based on this systematic study we propose a speci®c strategy for selecting effective transcriptional regulatory elements that allow the construction of a highly ef®cient expression vector. 2. Materials and methods 2.1. Plasmids A series of plasmid vectors expressing the ®re¯y luciferase gene, derived from pGL3-Control (Promega, Madison, WI), under the regulatory control of various promoter/ enhancers, introns, and P(A) sequences were used (Table 1). CMV and RSV promoter/enhancers are derived from pHMCMV6 and pHMRSV6 (Mizuguchi and Kay, 1999), respectively. The hybrid CA promoter/enhancer (CMV enhancer/b-actin promoter with b-actin intron) and human phosphoglycerate kinase (PGK) promoter were kindly provided by Dr J. Miyazaki (Osaka University) (Niwa et al., 1991) and Dr M.A. Kay (Stanford University), respecTable 1 Structures of different plasmids used in this study a Plasmid name
Promoter
Intron
P(A)
Enhancer
pGL3-Control pCMVL1 pRSVL1 pPGKL pCAL3-1 pCMVL2-1 pCMVL3-1 pCMVL3 pCMVL3-2 pCMVSL3 pCAL3 pCAL3-2 pCASL3
SV40 CMV RSV PGK b-Actin CMV CMV CMV CMV CMV b-Actin b-Actin b-Actin
± ± ± ± b-Actin Chimeric Intron A Intron A Intron A Intron A b-Actin b-Actin b-Actin
SV40 BGH BGH BGH BGH BGH BGH SV40 mRBG SV40 SV40 mRBG SV40
SV40 CMV RSV ± CMV CMV CMV CMV CMV CMV 1 SV40 CMV CMV CMV 1 SV40
a
Hybrid CA promoter/enhancer is composed of CMV enhancer and bactin promoter with b-actin intron.
tively. The sequence of CMV enhancer used in hybrid CA promoter/enhancer was the same as that in CMV promoter/ enhancer. SV40 promoter, SV40 enhancer, and SV40 P(A) (late gene) sequence are derived from pGL3-Control. SV40 enhancer was positioned following the P(A) sequence (see Fig. 4). Intron A and the chimeric intron (the 5 0 -donor splice site from human b-globin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region) are derived from pGeneGrip (Gene Therapy Systems, San Diego, CA) and pALTERMAX (Promega), respectively. Minimal rabbit b-globin (mRBG) P(A) sequence and BGH P(A) sequence are derived from pGeneGrip and pHMCMV6 (or pHMRSV6), respectively. All vectors contain Kozak consensus sequences surrounding the start codon of the luciferase gene. All plasmid DNAs were prepared using Qiagen Maxi kits (Qiagen, Germany). 2.2. Cells HeLa cells (human epitheloid carcinoma of the cervix) were cultured in Dulbecco's modi®ed Eagle's medium supplemented with 10% fetal calf serum (FCS). HepG2 cells (human hepatoma) were cultured in minimum essential medium supplemented with 10% FCS. ECV304 cells (endothelial cell line derived from human umbilical cord vein) were cultured in medium 199 supplemented with 10% FCS. 2.3. Transfection into cultured cells The cells (HeLa, HepG2, and ECV304; 1 £ 10 4 cells) were seeded into a 96-well plate. On the following day, each vector (0.5 mg per well) was transfected with SuperFect (Qiagen) according to the manufacturer's instructions. Forty-eight hours later, luciferase activity in the cells was determined using a luciferase assay system (PicaGene LT2.0, Toyo Inki Co., Tokyo, Japan). All activities were corrected by transfection of the control plasmid, pUC18. Transfection was performed in triplicate and repeated at least using an independent plasmid preparation. The standard deviations were less than 20%. We did not use a second co-transfected plasmid for standardizing transfection ef®ciency, because ubiquitous promoters (including SV40 and CMV) could be in¯uenced by co-transfected plasmids (Ponder et al., 1991; Graulich et al., 1999). Protein content was determined with the bicinchoninic acid assay (Pierce Chemical Co., Rockford, IL) using bovine serum albumin as a standard. 2.4. In vivo assays In vivo transfection into mouse liver was performed using the method described by Liu and Knapp (2001) and Liu et al. (1999). In brief, the mice (female Balb/c, 6 weeks old) were injected via the tail vein with 10 mg of each plasmid in
Z.-L. Xu et al. / Gene 272 (2001) 149±156
2.0 ml of saline in 6 s. The liver was recovered 8 h postinjection, which showed maximal luciferase expression, and luciferase activity in the liver was determined. In vivo transfection into mouse muscle was performed using the method described in The DNA Vaccine Web (http://www.dnavaccine.com/Protocols/genmeth.html). In brief, the mice (female Balb/c, 6 weeks old) were injected via the tibialis anterior (TA) muscle using a plastic insulin syringe with 50 mg of each plasmid in 50 ml of phosphate buffered saline. TA muscle was recovered 7 days post-injection, which showed maximal luciferase expression. Luciferase extraction was then performed using the method described by Davis et al. (1993), and luciferase activity was determined. 3. Results To evaluate the effects of commonly-used gene transcriptional regulatory elements on luciferase expression in vitro and in vivo, various vectors were constructed, as shown in Table 1. For comparison, we assigned the expression level of a commercially available vector, pGL3-Control, as 1.0 and expressed the activity of all other vectors relative to this value. 3.1. Commonly-used promoter/enhancers CMV, RSV, SV40, PGK, and hybrid CA (CMV enhan-
151
cer/chicken b-actin promoter/b-actin intron) promoter/ enhancers are widely used in various vectors and transgenic animals. Generally CMV promoter/enhancer is regarded as one of the strongest viral promoter/enhancers (Foecking and Hofstetter, 1986). First, we investigated the relative strength of these promoter/enhancers in regard to various cultured cells (HeLa, HepG2, and ECV304) (Fig. 1A) and in vivo (mouse liver and muscle) (Fig. 1B). All vectors except pGL3-Control contained BGH P(A) sequence. In both in vitro and in vivo conditions, pCAL3-1 containing the hybrid CA promoter/enhancer showed the maximal luciferase expression, which yielded over 1.5±6-fold higher in vitro and 2-fold higher in vivo (mouse liver and skeletal muscle) expressions than did pCMVL1 containing conventional CMV promoter/enhancer (without intron sequence). RSV promoter/enhancer (pRSVL1) showed about 1/2 to 1/6 lower activity than did CMV promoter/enhancer (pCMVL1) both in vitro and in vivo. PGK promoter (pPGKL) and SV40 promoter/enhancer (pGL3-Control) showed a moderate activity in vitro, but only a little in vivo (especially in muscle). 3.2. Improved CMV promoter/enhancer Introns have been shown to increase gene expression from plasmids and in transgenic mice (Niwa et al., 1990; Choi et al., 1991). In the promoter/enhancers used in Fig. 1, the hybrid CA promoter/enhancer has the b-actin intron,
Fig. 1. Comparison of luciferase expression from different promoter/enhancers in vitro (A) and in vivo (B). (A) Each vector was transfected into HeLa (A-1), HepG2 (A-2), or ECV304 (A-3) cells. Forty-eight hours later, the luciferase activity and total protein content were determined. The data are expressed as the mean ^ SD (n 5). (B) (B-1) Mice were injected with 2.0 ml of saline containing 10 mg of each vector. Animals were killed 8 h after injection and luciferase activity in the liver was determined. The data are expressed as the mean ^ SE (n 5). (B-2) Each vector (50 mg) was injected directly into the TA muscle of the mouse. Animals were sacri®ced 7 days after injection and luciferase activity in the muscle was determined. The data are expressed as the mean ^ SE (n 8±10). The absolute luciferase expression in HeLa, HepG2, and ECV304 cells and in vivo liver transfected with pGL3-Control was 15.0, 3.0, 21.0, and 27.9 ng luciferase/mg protein, respectively. The absolute luciferase expression in TA muscle transfected with pGL3-Control was 0.4 ng luciferase/muscle. In all following ®gures, the absolute luciferase levels in cultured cells and in vivo tissues (liver and muscle) transfected with pGL3-Control were the same as in this ®gure. The luciferase levels by each vector were normalized to that of the pGL3-Control.
152
Z.-L. Xu et al. / Gene 272 (2001) 149±156
which would partly contribute to its strong activity. Therefore, we examined the effect of two introns on the activity of the CMV promoter/enhancer which was stronger than other promoter/enhancers. One is a chimeric intron of b-globin and immunoglobin (the 5 0 -donor splice site from human bglobin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region), and the other is the largest intron of the CMV immediate-early 1 gene, intron A. Both intron sequences were inserted between the CMV promoter and luciferase gene of pCMVL1 in order to generate the plasmid vectors pCMVL2-1 and pCMVL3-1, respectively. As shown in Fig. 2, both in vitro and in vivo, pCMVL1 without the intron sequence and pCMVL2-1 with the chimeric intron sequence showed comparable levels of expression. However, pCMVL3-1 with the inclusion of the intron A sequence generated 2±6-fold in vitro and 1.5±3-fold in vivo higher levels of luciferase expression than did pCMVL1, showing a comparable level of expression as did pCAL3-1 containing the hybrid CA promoter/enhancer shown in Fig. 1. Therefore, in the following studies, only the hybrid CA promoter/ enhancer and the improved CMV promoter/enhancer with intron A sequence were used to examine the effects of P(A) sequence and a second enhancer. 3.3. Effect of polyadenylation signal sequences To investigate if the origin of the P(A) sequences affects luciferase expression, we compared three commonly-used P(A) sequences, derived from the BGH gene, the SV40 late gene, and the mRBG gene, respectively. These P(A)
sequences were inserted downstream of the luciferase gene in vectors containing the improved CMV or hybrid CA promoter/enhancer. As shown in Fig. 3, in all three cell lines the vectors containing the mRBG or SV40 P(A) sequence, with the promoter/enhancer being the improved CMV promoter/enhancer (pCMVL3 and pCMVL3-2), showed almost the same level of luciferase expression, approximately 2-fold as high as that observed in the vector with the BGH P(A) sequence (pCMVL3-1). However, in the context of the hybrid CA promoter/enhancer (pCAL3, pCAL3-1 and pCAL3-2), no signi®cant difference of luciferase expression level was seen between these three P(A) sequences, though the BGH P(A) sequence tended to be the weakest. In vivo, the promoter/enhancer did not affect the effect pattern of P(A) sequence. In mouse liver, the vectors containing BGH or mRBG P(A) sequence were comparably ef®cient, approximately 2±3-fold as ef®cient as the vector with the SV40 P(A) sequence, whereas in mouse skeletal muscle, the vectors containing mRBG or SV40 P(A) sequence were at the same expression level, approximately 2±3-fold as high as those with the BGH P(A) sequence. Generally speaking, the mRBG P(A) sequence is relatively stronger, independent of promoter/enhancer or target cell/ tissues. 3.4. Effect of adding a second heterologous enhancer sequence To determine if the vector containing two heterologous enhancers could produce higher levels of expression than
Fig. 2. Effect of intron sequence on the activity of CMV promoter/enhancer in vitro (A) and in vivo (B). (A-1) HeLa cells; (A-2) HepG2 cells; (A-3) ECV304 cells. (B-1) Mouse liver; (B-2) mouse muscle. The experiment was performed as described in Fig. 1 and in Section 2. The luciferase levels expressed by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. Chimeric intron, the 5 0 -donor splice site from human b-globin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region; intron A, from the human cytomegalovirus immediate-early 1 gene.
Z.-L. Xu et al. / Gene 272 (2001) 149±156
153
Fig. 3. Effect of polyadenylation signal sequence on luciferase expression in vitro (A) and in vivo (B). (A-1) HeLa cells; (A-2) HepG2 cells; (A-3) ECV304 cells. (B-1) Mouse liver; (B-2) mouse muscle. The experiment was performed as described in Fig. 1 and in Section 2. The luciferase levels expressed by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1.
that containing one enhancer, a SV40 enhancer sequence was inserted downstream of the P(A) sequence in plasmids containing the improved CMV or hybrid CA promoter/ enhancer (Fig. 4). Luciferase expression was examined in cultured cells (Fig. 5), mouse liver (Fig. 6), and mouse skeletal muscle (Fig. 7). In the context of SV40 P(A) sequence, the addition of a second SV40 enhancer increased expression to an approximately 2-fold level in HepG2, ECV304 cells (Fig. 5) and mouse liver (Fig. 6), regardless of the promoter/enhancer. In mouse liver, though the BGH P(A) sequence was 2±3fold as strong as the SV40 P(A) sequence (Fig. 3), the addition of the second SV40 enhancer downstream of the SV40 P(A) sequence increased the expression to a level comparable to the plasmid vectors with BGH P(A) sequence (pCAL3-1 and pCMVL3-1) (Fig. 6). However, in HeLa cell and mouse skeletal muscle (Figs. 5 and 7), a second SV40 enhancer element failed to increase luciferase expression, regardless of the promoter/enhancer or P(A) sequence. The addition of a second SV40 enhancer into the vectors containing BGH or mRBG P(A) sequence did not demonstrate signi®cantly enhanced luciferase expression (data not shown). Moreover, with the addition of a second SV40 enhancer in vitro the improved CMV promoter/enhancer (pCMVSL3) was approximately 1.5±3-fold more ef®cient than the hybrid CA promoter/enhancer containing a second SV40 enhancer (pCASL3) (Fig. 5). In mouse skeletal muscle, the improved CMV promoter/enhancer was also slightly more ef®cient than the hybrid CA promoter/enhancer (Fig. 7), while in contrast, in mouse liver, the improved CMV promoter/
enhancer was slightly less ef®cient than the hybrid CA promoter/enhancer (Fig. 6).
Fig. 4. Structure of plasmids used in Figs. 5±7.
154
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Fig. 5. Effect of a second enhancer on luciferase expression in vitro. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SD (n 5). The luciferase levels by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. Black bar, pCAL3; white bar, pCASL3. (B) Vector containing the improved CMV promoter/enhancer. Black bar, pCMVL3; white bar, pCMVSL3.
4. Discussion On one hand, though plasmid vectors can be used for transfection into the cell by physical or chemical methods, e.g. electroporation (Heller et al., 2000), particle bombardment (Cheng et al., 1993) or cationic lipid (Felgner et al., 1987), and can even be injected directly into muscle or the mouse tail veil, its ef®ciency is much lower than that of viral vectors. On the other hand, while viral vectors have a relatively much higher transfection ef®ciency, toxicity and immune response become serious with increasing dosages. One simple and practical solution to both problems is to produce the most ef®cient expression-per-vector molecule.
Fig. 6. Effect of a second enhancer on luciferase expression in mouse liver. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SE (n 5). The luciferase levels by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. (B) Vector containing the improved CMV promoter/enhancer.
The comparison of several commonly-used promoter/ enhancers showed that the hybrid CA promoter/enhancer containing CMV enhancer and chicken b-actin promoter was as ef®cient as the improved CMV promoter/enhancer which has been found to be one of the strongest viral promoter/enhancers (Foecking and Hofstetter, 1986), whereas SV40 promoter/enhancer, RSV LTR promoter/enhancer, and PGK promoter were much less ef®cient (Figs. 1 and 2). The CMV intron A has been demonstrated to improve expression of heterologous genes expressed in vitro by the CMV immediate-early 1 or other promoter/enhancers (Chapman et al., 1991). The bene®cial effect of introns on expression has been ascribed primarily to an enhanced rate of RNA polyadenylation and/or nuclear transport associated with RNA splicing (Huang and Gorman, 1990). Mikkelsen et al. (1992) reported that sequences in intron A are homologous to muscle-speci®c regulatory elements and that the
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Fig. 7. Effect of a second enhancer on luciferase expression in mouse skeletal muscle. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SE (n 8±10). The luciferase levels by each vector were normalized to that of the pGL3Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. (B) Vector containing the improved CMV promoter/enhancer.
CMV promoter/enhancer with intron A conferred high-level expression in muscles of transgenic mice. In this study, our data also showed that the addition of the CMV intron A led to a 2-fold increase in luciferase expression in the context of direct injection of plasmid DNA into mouse skeletal muscle (Fig. 2B-2), but showed only a slight increase in mouse liver (Fig. 2B-1). In cultured cells (HeLa, HepG2, and ECV304), the inclusion of intron A to the CMV promoter/enhancer generated more than 2-fold higher levels of luciferase expression (Fig. 2A). These results suggest that intron A may be a positive element for the high ef®ciency of CMV promoter/enhancer. The ef®ciency of primary RNA transcript processing and polyadenylation is known to vary between the polyadenylation signal sequences of different genes. Moreover, as the rate of transcriptional initiation is increased by the use of strong promoter/enhancers, the process of transcriptional
155
termination may become rate-limiting (Proudfoot, 1989). Though the SV40 P(A) sequence is commonly used in expression vectors, it may not be the best choice for optimal expression. We found that in mouse liver, the SV40 P(A) sequence was less ef®cient than mRBG and BGH P(A) sequences (Fig. 3B-1). The mRBG P(A) sequence has been shown to be ef®ciently recognized in insect and mammalian cells (Westwood et al., 1993). Whereas the activity of both BGH and SV40 P(A) sequences depended on cell or tissue targets, in the present experiments (Fig. 3) the mRBG P(A) sequence showed high activity in almost all cell lines and tissues transfected. This result shows that the high activity of the mRBG P(A) sequence seems to be more ubiquitous than the SV40 and BGH P(A) sequences. However, with the insertion of a second SV40 enhancer downstream of the SV40 P(A) sequence of plasmids which had contained a CMV enhancer, a further increase in expression was achieved (Figs. 5±7 and data not shown). Although this kind of enhanced effect achieved by using a second SV40 enhancer did not occur in HeLa cell and mouse skeletal muscle, the combination of the SV40 P(A) sequence and a second SV40 enhancer was comparable to or more active than only BGH or mRBG P(A) sequence. Therefore, it is anticipated that our strategy of inserting a second SV40 enhancer downstream of the SV40 P(A) sequence can be utilized in the construction of ef®cient vectors. From the overall results, pCMVSL3 (composed of CMV enhancer, CMV promoter, intron A, SV40 P(A) sequence, and SV40 enhancer) and pCASL3 (composed of CMV enhancer, chicken b-actin promoter, chicken b-actin intron, SV40 P(A) sequence, and SV40 enhancer) were the most ef®cient under in vitro and in vivo conditions. In vitro, pCMVSL3 was more ef®cient than was pCASL3 (Fig. 5). In mouse liver, pCMVSL3 was less ef®cient than was pCASL3 (Fig. 6), while in mouse skeletal muscle, pCMVSL3 was more ef®cient than was pCASL3 (Fig. 7). It is worth mentioning that the b-actin promoter shows more ubiquitous activity than does the CMV promoter. Okabe et al. (1997) produced a green mouse, a transgenic mouse line containing the cDNA of green ¯uorescent protein (GFP) under the control of a chicken b-actin promoter and CMV enhancer. All of the tissues from these transgenic lines, with the exception of erythrocytes and hair, were green under excitation light. Sawicki et al. (1998) compared the promoter/enhancer ef®ciency for ubiquitous expression in the skin of adult transgenic mice. They found that a CMV enhancer/ promoter is not active in most cell types in the skin, whereas a CMV enhancer/modi®ed b-actin promoter sequence is active in the suprabasal and basal cells of the epidermis as well as in the epithelial cells of the hair follicles, sebaceous glands, and the dermal papillae. Therefore, the b-actin promoter might be the promoter of choice for high-level and ubiquitous expression in vivo. In summary, we systematically evaluated the combination of various transcriptional regulatory elements in cultured cells and in animal tissues (mouse liver and skeletal
156
Z.-L. Xu et al. / Gene 272 (2001) 149±156
muscle) in vivo. The results of our study provide important information regarding the construction of vectors for gene therapy, vaccine development, and gene transfer experiments. Acknowledgements We would like to thank Mr Jun Murai for his technical assistance. We also thank Dr S. Nakagawa (Osaka University) for his helpful discussion. We would like to thank Dr J. Miyazaki for kindly providing the cytomegalovirus enhancer/chicken b-actin promoter. This work was supported by grants from the Ministry of Health and Welfare of Japan. References Chapman, B.S., Thayer, R.M., Vincent, K.A., Haigwood, N.L., 1991. Effect of intron A from human cytomegalovirus(Towne)immediate-early gene on heterologous expression in mammalian cells. Nucleic Acids Res. 19, 3979±3986. Cheng, L., Ziegelhoffer, P.R., Yang, N.S., 1993. In vivo promoter activity and transgene expression in mammalian somatic tissues evaluated by using particle bombardment. Proc. Natl. Acad. Sci. USA 90, 4455± 4459. Choi, T., Huang, M., Gorman, C., Jaenisch, R., 1991. A generic intron increases gene expression in transgenic mice. Mol. Cell. Biol. 11, 3070±3074. Davis, H.L., Whalen, R.G., Demeneix, B.A., 1993. Direct gene transfer into skeletal muscle in vivo: factors affecting ef®ciency of transfer and stability of expression. Hum. Gene Ther. 4, 151±159. Felgner, P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W., Wenz, M., Northrop, J.P., Ringold, G.M., Danielsen, M., 1987. Lipofection: a highly ef®cient, lipid-mediated DNA-transfection procedure. Proc. Natl. Acad. Sci. USA 84, 7413±7417. Foecking, M.K., Hofstetter, H., 1986. Powerful and versatile enhancerpromoter unit for mammalian expression vectors. Gene 45, 101±105. Graulich, W., Nettelbeck, D.M., Fischer, D., Kissel, T., Muller, R., 1999. Cell type speci®city of the human endoglin promoter. Gene 227, 55±62. Gunter, K.C., Khan, A.S., Noguchi, P.D., 1993. The safety of retroviral vectors. Hum. Gene Ther. 4, 643±645. Guo, Z.S., Wang, L.H., Eisensmith, R.C., Woo, S.L.C., 1996. Evaluation of promoter strength for hepatic gene expression in vivo following adenovirus-mediated gene transfer. Gene Ther. 3, 802±810. Hartikka, J., Sawdey, M., Cornefert-Jensen, F., Margalith, M., Barnhart, K., Nolasco, M., Vahlsing, H.L., Meek, J., Marguet, M., Hobart, P., Norman, J., Manthorpe, M., 1996. An improved plasmid DNA expression vector for direct injection into skeletal muscle. Hum. Gene Ther. 7, 1205±1217. Heller, L., Jaroszeski, M.J., Coppola, D., Pottinger, C., Gilbert, R., Heller, R., 2000. Electrically mediated plasmid DNA delivery to hepatocellular carcinomas in vivo. Gene Ther. 7, 826±829. Huang, M.T., Gorman, C.M., 1990. Intervening sequences increase ef®ciency of RNA 3 0 processing and accumulation of cytoplasmic RNA. Nucleic Acids Res. 18, 937±947.
Liu, D., Knapp, J.E., 2001. Hydrodynamics-based gene delivery. Curr. Opin. Mol. Ther. 3, 192±197. Liu, F., Song, Y.K., Liu, D., 1999. Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther. 6, 1258±1266. Manthorpe, M., Cornefert-Jensen, F., Hartikka, J., Felgner, J., Rundell, A., Margalith, M., Dwarki, V., 1993. Gene therapy by intramuscular injection of plasmid DNA: studies on Fire¯y luciferase gene expression in mice. Hum. Gene Ther. 4, 419±431. Martin-Gallardo, A., Montoya-Zavala, M., Kelder, B., Taylor, J., Chen, H., Leung, F.C., Kopchick, J.J., 1988. A comparison of bovine growthhormone gene expression in mouse L cells directed by the Moloney murine-leukemia virus long terminal repeat, simian virus-40 early promoter or cytomegalovirus immediate-early promoter. Gene 70, 51±56. Mcknight, S., Tjian, R., 1986. Transcriptional selectivity of viral genes in mammalian cells. Cell 46, 795±805. Mikkelsen, T.R., Chapman, B., Din, N., Ingerslev, J., Kristensen, P., Poulsen, K., Hjorth, J.P., 1992. Expression of a cytomegalovirus IE-1-factor VIII cDNA hybrid gene in transgenic mice. Transgenic Res. 1, 164± 169. Mizuguchi, H., Kay, M.A., 1999. A simple method for constructing E1- and E1/E4-deleted recombinant adenoviral vectors. Hum. Gene Ther. 10, 2013±2017. Niwa, H., Yamamura, K., Miyazaki, J., 1991. Ef®cient selection for highexpression transfectants with a novel eukaryotic vector. Gene 108, 193± 199. Niwa, M., Rose, S.D., Berget, S.M., 1990. In vitro polyadenylation is stimulated by the presence of an upstream intron. Genes Dev. 4, 1552±1559. Oellig, C., Seliger, B., 1990. Gene transfer into brain tumor cell lines: reporter gene expression using various cellular and viral promoters. J. Neurosci. Res. 26, 390±396. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., Nishimune, Y., 1997. ªGreen miceº as a source of ubiquitous green cells. FEBS Lett. 407, 313±319. Pasleau, F., Tocci, M.J., Leung, F., Kopchick, J.J., 1985. Growth hormone gene expression in eukaryotic cells directed by the Rous sarcoma virus long terminal repeat or cytomegalovirus immediate-early promoter. Gene 38, 227±232. Ponder, K.P., Dunbar, R.P., Wilson, D.R., Darlington, G.J., Woo, S.L.C., 1991. Evaluation of relative promoter strength in primary hepatocytes using optimized lipofection. Hum. Gene Ther. 2, 41±52. Proudfoot, N.J., 1989. How RNA polymerase II terminates transcription in higher eukaryotes. Trends Biochem. Sci. 14, 105±110. Sawicki, J.A., Morris, R.J., Monks, B., Sakai, K., Miyazaki, J., 1998. A composite CMV-IE enhancer/b-actin promoter is ubiquitously expressed in mouse cutaneous epithelium. Exp. Cell Res. 244, 367±369. Westwood, J.A., Jones, I.M., Bishop, D.H., 1993. Analyses of alternative poly(A) signals for use in baculovirus expression vectors. Virology 195, 90±99. Yang, Y., Li, Q., Ertl, H.C., Wilson, J.M., 1995. Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J. Virol. 69, 2004±2015. Yew, N.S., Wysokenski, D.M., Wang, K.X., Ziegler, R.J., Marshall, J., Mcneilly, D., Cherry, M., Osburn, W., Cheng, S.H., 1997. Optimization of plasmid vectors for high-level expression in lung epithelial cells. Hum. Gene Ther. 8, 575±584.
www.elsevier.com/locate/gene
Optimization of transcriptional regulatory elements for constructing plasmid vectors Zhi-Li Xu a,b, Hiroyuki Mizuguchi a,*, Akiko Ishii-Watabe a, Eriko Uchida a, Tadanori Mayumi b, Takao Hayakawa a b
a Division of Biological Chemistry and Biologicals, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan Department of Biopharmaceutics, Graduate School of Pharmaceutical Sciences, Osaka University, 1-6 Yamadaoka, Suita, Osaka 565-0871, Japan
Received 26 February 2001; received in revised form 24 May 2001; accepted 29 May 2001 Received by T. Sekiya
Abstract In studies regarding both gene therapy and gene function, transgene expression by plasmid vectors bene®ts from the use of transcriptional regulatory elements which permit high-level gene expression. Therefore, with respect to transgene (luciferase) expression activity both in vitro (using HeLa, HepG2, and ECV304 cells) and in vivo (mouse liver and skeletal muscle), we investigated the effective combination of commonly-used regulatory elements, such as the promoter/enhancer, intron, and polyadenylation signal (P(A)) sequence by constructing a series of plasmids that differed only in the particular sequence element being evaluated. Of the several promoter/enhancers that were tested, hybrid CA promoter/enhancer containing human cytomegalovirus immediate-early 1 gene (CMV) enhancer and chicken b-actin promoter with the b-actin intron sequence, and the improved CMV promoter/enhancer containing the largest intron of CMV (intron A) produced the highest levels of expression both in vitro and in vivo. P(A) sequences were found to have signi®cant effects on transgene expression. The effect of a multiple enhancer was also examined. Optimized plasmids of this study were pCASL3 (composed of CMV enhancer, b-actin promoter, b-actin intron, Simian virus (SV40) P(A) sequence and SV40 enhancer) and pCMVSL3 (composed of CMV enhancer, CMV promoter, intron A, SV40 P(A) sequence and SV40 enhancer). These comparative analyses could provide a systematic reference for the development of vector construction for gene therapy, vaccine development, and gene transfer experiments. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Transgene expression; Gene therapy; Promoter; Enhancer; Intron; Polyadenylation signal sequence
1. Introduction The delivery of exogenous genes into animal cells plays an important role in not only the study of gene function but in gene therapy as well. The expression levels of the introduced gene depend mostly on the strength of transcriptional regulatory elements and the transduction ef®ciency of the gene transfer vector (Mcknight and Tjian, 1986). Currently, great efforts are being made to develop gene transfer vectors, viral and non-viral vectors. Some viral vectors produce the infectious wild type virus during ampli®cation Abbreviations: BGH, bovine growth hormone; Chimeric, hybrid intron of b-globin and immunoglobin; CMV, human cytomegalovirus immediateearly 1 gene; Intron A, the largest intron of human CMV IE gene; mRBG, minimal rabbit b-globin; P(A), polyadenylation signal; PGK, human phosphoglycerate kinase; RSV, Rous sarcoma virus genome long terminal repeats; SV40, Simian virus * Corresponding author. Tel.: 181-3-3700-9089; fax: 181-3-3700-9084. E-mail address: [email protected] (H. Mizuguchi).
and induce an immune response that renders repeated administration problematic (Gunter et al., 1993; Yang et al., 1995). Increasing transgene expression by optimizing the transcriptional regulatory elements is desirable as a way to decrease the amount of viral vector required to achieve a clinical effect. For the non-viral vectors, enhanced expression ef®ciency by using optimal transcriptional regulatory elements should be a practical way of making the vector more ef®cient. Previous reports have described the strength evaluation of promoter/enhancers or other transcriptional regulatory elements. However, almost all have focused on single cell (Pasleau et al., 1985; Martin-Gallardo et al., 1988; Oellig and Seliger, 1990; Yew et al., 1997) or tissue targets (Manthorpe et al., 1993; Hartikka et al., 1996), or on transduction methods (Cheng et al., 1993; Guo et al., 1996). No systematic study which examines the function of the vector in various conditions including cultured cells and animal tissues in vivo has been reported.
0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00550-9
150
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Although various kinds of transcriptional regulatory elements have been used to construct vectors, generally only a few are considered as ubiquitously strong elements and widely used thus far. These include CMV and Rous sarcoma virus genome long terminal repeats (RSV) promoters, SV40 and bovine growth hormone (BGH) P(A) sequences, etc. However, few studies have simultaneously compared the relative strength of these so-called strong elements in various types of cell or tissue. We report here the results from a series of experiments designed to identify the relative strength of these commonly-used transcriptional regulatory elements both in vitro and in vivo. In vitro, three cell lines, HeLa from human cervical carcinoma, HepG2 from human hepatoma, and ECV304 human vascular endothelial cell from umbilical cord vein, were used. In vivo, mouse liver and skeletal muscle were transfected. All these cells or tissues are important targets for gene therapy and gene transfer experiments. Based on this systematic study we propose a speci®c strategy for selecting effective transcriptional regulatory elements that allow the construction of a highly ef®cient expression vector. 2. Materials and methods 2.1. Plasmids A series of plasmid vectors expressing the ®re¯y luciferase gene, derived from pGL3-Control (Promega, Madison, WI), under the regulatory control of various promoter/ enhancers, introns, and P(A) sequences were used (Table 1). CMV and RSV promoter/enhancers are derived from pHMCMV6 and pHMRSV6 (Mizuguchi and Kay, 1999), respectively. The hybrid CA promoter/enhancer (CMV enhancer/b-actin promoter with b-actin intron) and human phosphoglycerate kinase (PGK) promoter were kindly provided by Dr J. Miyazaki (Osaka University) (Niwa et al., 1991) and Dr M.A. Kay (Stanford University), respecTable 1 Structures of different plasmids used in this study a Plasmid name
Promoter
Intron
P(A)
Enhancer
pGL3-Control pCMVL1 pRSVL1 pPGKL pCAL3-1 pCMVL2-1 pCMVL3-1 pCMVL3 pCMVL3-2 pCMVSL3 pCAL3 pCAL3-2 pCASL3
SV40 CMV RSV PGK b-Actin CMV CMV CMV CMV CMV b-Actin b-Actin b-Actin
± ± ± ± b-Actin Chimeric Intron A Intron A Intron A Intron A b-Actin b-Actin b-Actin
SV40 BGH BGH BGH BGH BGH BGH SV40 mRBG SV40 SV40 mRBG SV40
SV40 CMV RSV ± CMV CMV CMV CMV CMV CMV 1 SV40 CMV CMV CMV 1 SV40
a
Hybrid CA promoter/enhancer is composed of CMV enhancer and bactin promoter with b-actin intron.
tively. The sequence of CMV enhancer used in hybrid CA promoter/enhancer was the same as that in CMV promoter/ enhancer. SV40 promoter, SV40 enhancer, and SV40 P(A) (late gene) sequence are derived from pGL3-Control. SV40 enhancer was positioned following the P(A) sequence (see Fig. 4). Intron A and the chimeric intron (the 5 0 -donor splice site from human b-globin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region) are derived from pGeneGrip (Gene Therapy Systems, San Diego, CA) and pALTERMAX (Promega), respectively. Minimal rabbit b-globin (mRBG) P(A) sequence and BGH P(A) sequence are derived from pGeneGrip and pHMCMV6 (or pHMRSV6), respectively. All vectors contain Kozak consensus sequences surrounding the start codon of the luciferase gene. All plasmid DNAs were prepared using Qiagen Maxi kits (Qiagen, Germany). 2.2. Cells HeLa cells (human epitheloid carcinoma of the cervix) were cultured in Dulbecco's modi®ed Eagle's medium supplemented with 10% fetal calf serum (FCS). HepG2 cells (human hepatoma) were cultured in minimum essential medium supplemented with 10% FCS. ECV304 cells (endothelial cell line derived from human umbilical cord vein) were cultured in medium 199 supplemented with 10% FCS. 2.3. Transfection into cultured cells The cells (HeLa, HepG2, and ECV304; 1 £ 10 4 cells) were seeded into a 96-well plate. On the following day, each vector (0.5 mg per well) was transfected with SuperFect (Qiagen) according to the manufacturer's instructions. Forty-eight hours later, luciferase activity in the cells was determined using a luciferase assay system (PicaGene LT2.0, Toyo Inki Co., Tokyo, Japan). All activities were corrected by transfection of the control plasmid, pUC18. Transfection was performed in triplicate and repeated at least using an independent plasmid preparation. The standard deviations were less than 20%. We did not use a second co-transfected plasmid for standardizing transfection ef®ciency, because ubiquitous promoters (including SV40 and CMV) could be in¯uenced by co-transfected plasmids (Ponder et al., 1991; Graulich et al., 1999). Protein content was determined with the bicinchoninic acid assay (Pierce Chemical Co., Rockford, IL) using bovine serum albumin as a standard. 2.4. In vivo assays In vivo transfection into mouse liver was performed using the method described by Liu and Knapp (2001) and Liu et al. (1999). In brief, the mice (female Balb/c, 6 weeks old) were injected via the tail vein with 10 mg of each plasmid in
Z.-L. Xu et al. / Gene 272 (2001) 149±156
2.0 ml of saline in 6 s. The liver was recovered 8 h postinjection, which showed maximal luciferase expression, and luciferase activity in the liver was determined. In vivo transfection into mouse muscle was performed using the method described in The DNA Vaccine Web (http://www.dnavaccine.com/Protocols/genmeth.html). In brief, the mice (female Balb/c, 6 weeks old) were injected via the tibialis anterior (TA) muscle using a plastic insulin syringe with 50 mg of each plasmid in 50 ml of phosphate buffered saline. TA muscle was recovered 7 days post-injection, which showed maximal luciferase expression. Luciferase extraction was then performed using the method described by Davis et al. (1993), and luciferase activity was determined. 3. Results To evaluate the effects of commonly-used gene transcriptional regulatory elements on luciferase expression in vitro and in vivo, various vectors were constructed, as shown in Table 1. For comparison, we assigned the expression level of a commercially available vector, pGL3-Control, as 1.0 and expressed the activity of all other vectors relative to this value. 3.1. Commonly-used promoter/enhancers CMV, RSV, SV40, PGK, and hybrid CA (CMV enhan-
151
cer/chicken b-actin promoter/b-actin intron) promoter/ enhancers are widely used in various vectors and transgenic animals. Generally CMV promoter/enhancer is regarded as one of the strongest viral promoter/enhancers (Foecking and Hofstetter, 1986). First, we investigated the relative strength of these promoter/enhancers in regard to various cultured cells (HeLa, HepG2, and ECV304) (Fig. 1A) and in vivo (mouse liver and muscle) (Fig. 1B). All vectors except pGL3-Control contained BGH P(A) sequence. In both in vitro and in vivo conditions, pCAL3-1 containing the hybrid CA promoter/enhancer showed the maximal luciferase expression, which yielded over 1.5±6-fold higher in vitro and 2-fold higher in vivo (mouse liver and skeletal muscle) expressions than did pCMVL1 containing conventional CMV promoter/enhancer (without intron sequence). RSV promoter/enhancer (pRSVL1) showed about 1/2 to 1/6 lower activity than did CMV promoter/enhancer (pCMVL1) both in vitro and in vivo. PGK promoter (pPGKL) and SV40 promoter/enhancer (pGL3-Control) showed a moderate activity in vitro, but only a little in vivo (especially in muscle). 3.2. Improved CMV promoter/enhancer Introns have been shown to increase gene expression from plasmids and in transgenic mice (Niwa et al., 1990; Choi et al., 1991). In the promoter/enhancers used in Fig. 1, the hybrid CA promoter/enhancer has the b-actin intron,
Fig. 1. Comparison of luciferase expression from different promoter/enhancers in vitro (A) and in vivo (B). (A) Each vector was transfected into HeLa (A-1), HepG2 (A-2), or ECV304 (A-3) cells. Forty-eight hours later, the luciferase activity and total protein content were determined. The data are expressed as the mean ^ SD (n 5). (B) (B-1) Mice were injected with 2.0 ml of saline containing 10 mg of each vector. Animals were killed 8 h after injection and luciferase activity in the liver was determined. The data are expressed as the mean ^ SE (n 5). (B-2) Each vector (50 mg) was injected directly into the TA muscle of the mouse. Animals were sacri®ced 7 days after injection and luciferase activity in the muscle was determined. The data are expressed as the mean ^ SE (n 8±10). The absolute luciferase expression in HeLa, HepG2, and ECV304 cells and in vivo liver transfected with pGL3-Control was 15.0, 3.0, 21.0, and 27.9 ng luciferase/mg protein, respectively. The absolute luciferase expression in TA muscle transfected with pGL3-Control was 0.4 ng luciferase/muscle. In all following ®gures, the absolute luciferase levels in cultured cells and in vivo tissues (liver and muscle) transfected with pGL3-Control were the same as in this ®gure. The luciferase levels by each vector were normalized to that of the pGL3-Control.
152
Z.-L. Xu et al. / Gene 272 (2001) 149±156
which would partly contribute to its strong activity. Therefore, we examined the effect of two introns on the activity of the CMV promoter/enhancer which was stronger than other promoter/enhancers. One is a chimeric intron of b-globin and immunoglobin (the 5 0 -donor splice site from human bglobin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region), and the other is the largest intron of the CMV immediate-early 1 gene, intron A. Both intron sequences were inserted between the CMV promoter and luciferase gene of pCMVL1 in order to generate the plasmid vectors pCMVL2-1 and pCMVL3-1, respectively. As shown in Fig. 2, both in vitro and in vivo, pCMVL1 without the intron sequence and pCMVL2-1 with the chimeric intron sequence showed comparable levels of expression. However, pCMVL3-1 with the inclusion of the intron A sequence generated 2±6-fold in vitro and 1.5±3-fold in vivo higher levels of luciferase expression than did pCMVL1, showing a comparable level of expression as did pCAL3-1 containing the hybrid CA promoter/enhancer shown in Fig. 1. Therefore, in the following studies, only the hybrid CA promoter/ enhancer and the improved CMV promoter/enhancer with intron A sequence were used to examine the effects of P(A) sequence and a second enhancer. 3.3. Effect of polyadenylation signal sequences To investigate if the origin of the P(A) sequences affects luciferase expression, we compared three commonly-used P(A) sequences, derived from the BGH gene, the SV40 late gene, and the mRBG gene, respectively. These P(A)
sequences were inserted downstream of the luciferase gene in vectors containing the improved CMV or hybrid CA promoter/enhancer. As shown in Fig. 3, in all three cell lines the vectors containing the mRBG or SV40 P(A) sequence, with the promoter/enhancer being the improved CMV promoter/enhancer (pCMVL3 and pCMVL3-2), showed almost the same level of luciferase expression, approximately 2-fold as high as that observed in the vector with the BGH P(A) sequence (pCMVL3-1). However, in the context of the hybrid CA promoter/enhancer (pCAL3, pCAL3-1 and pCAL3-2), no signi®cant difference of luciferase expression level was seen between these three P(A) sequences, though the BGH P(A) sequence tended to be the weakest. In vivo, the promoter/enhancer did not affect the effect pattern of P(A) sequence. In mouse liver, the vectors containing BGH or mRBG P(A) sequence were comparably ef®cient, approximately 2±3-fold as ef®cient as the vector with the SV40 P(A) sequence, whereas in mouse skeletal muscle, the vectors containing mRBG or SV40 P(A) sequence were at the same expression level, approximately 2±3-fold as high as those with the BGH P(A) sequence. Generally speaking, the mRBG P(A) sequence is relatively stronger, independent of promoter/enhancer or target cell/ tissues. 3.4. Effect of adding a second heterologous enhancer sequence To determine if the vector containing two heterologous enhancers could produce higher levels of expression than
Fig. 2. Effect of intron sequence on the activity of CMV promoter/enhancer in vitro (A) and in vivo (B). (A-1) HeLa cells; (A-2) HepG2 cells; (A-3) ECV304 cells. (B-1) Mouse liver; (B-2) mouse muscle. The experiment was performed as described in Fig. 1 and in Section 2. The luciferase levels expressed by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. Chimeric intron, the 5 0 -donor splice site from human b-globin intron 1 plus the 3 0 -acceptor splice site from the intron of an immunoglobin gene heavy chain variable region; intron A, from the human cytomegalovirus immediate-early 1 gene.
Z.-L. Xu et al. / Gene 272 (2001) 149±156
153
Fig. 3. Effect of polyadenylation signal sequence on luciferase expression in vitro (A) and in vivo (B). (A-1) HeLa cells; (A-2) HepG2 cells; (A-3) ECV304 cells. (B-1) Mouse liver; (B-2) mouse muscle. The experiment was performed as described in Fig. 1 and in Section 2. The luciferase levels expressed by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1.
that containing one enhancer, a SV40 enhancer sequence was inserted downstream of the P(A) sequence in plasmids containing the improved CMV or hybrid CA promoter/ enhancer (Fig. 4). Luciferase expression was examined in cultured cells (Fig. 5), mouse liver (Fig. 6), and mouse skeletal muscle (Fig. 7). In the context of SV40 P(A) sequence, the addition of a second SV40 enhancer increased expression to an approximately 2-fold level in HepG2, ECV304 cells (Fig. 5) and mouse liver (Fig. 6), regardless of the promoter/enhancer. In mouse liver, though the BGH P(A) sequence was 2±3fold as strong as the SV40 P(A) sequence (Fig. 3), the addition of the second SV40 enhancer downstream of the SV40 P(A) sequence increased the expression to a level comparable to the plasmid vectors with BGH P(A) sequence (pCAL3-1 and pCMVL3-1) (Fig. 6). However, in HeLa cell and mouse skeletal muscle (Figs. 5 and 7), a second SV40 enhancer element failed to increase luciferase expression, regardless of the promoter/enhancer or P(A) sequence. The addition of a second SV40 enhancer into the vectors containing BGH or mRBG P(A) sequence did not demonstrate signi®cantly enhanced luciferase expression (data not shown). Moreover, with the addition of a second SV40 enhancer in vitro the improved CMV promoter/enhancer (pCMVSL3) was approximately 1.5±3-fold more ef®cient than the hybrid CA promoter/enhancer containing a second SV40 enhancer (pCASL3) (Fig. 5). In mouse skeletal muscle, the improved CMV promoter/enhancer was also slightly more ef®cient than the hybrid CA promoter/enhancer (Fig. 7), while in contrast, in mouse liver, the improved CMV promoter/
enhancer was slightly less ef®cient than the hybrid CA promoter/enhancer (Fig. 6).
Fig. 4. Structure of plasmids used in Figs. 5±7.
154
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Fig. 5. Effect of a second enhancer on luciferase expression in vitro. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SD (n 5). The luciferase levels by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. Black bar, pCAL3; white bar, pCASL3. (B) Vector containing the improved CMV promoter/enhancer. Black bar, pCMVL3; white bar, pCMVSL3.
4. Discussion On one hand, though plasmid vectors can be used for transfection into the cell by physical or chemical methods, e.g. electroporation (Heller et al., 2000), particle bombardment (Cheng et al., 1993) or cationic lipid (Felgner et al., 1987), and can even be injected directly into muscle or the mouse tail veil, its ef®ciency is much lower than that of viral vectors. On the other hand, while viral vectors have a relatively much higher transfection ef®ciency, toxicity and immune response become serious with increasing dosages. One simple and practical solution to both problems is to produce the most ef®cient expression-per-vector molecule.
Fig. 6. Effect of a second enhancer on luciferase expression in mouse liver. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SE (n 5). The luciferase levels by each vector were normalized to that of the pGL3-Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. (B) Vector containing the improved CMV promoter/enhancer.
The comparison of several commonly-used promoter/ enhancers showed that the hybrid CA promoter/enhancer containing CMV enhancer and chicken b-actin promoter was as ef®cient as the improved CMV promoter/enhancer which has been found to be one of the strongest viral promoter/enhancers (Foecking and Hofstetter, 1986), whereas SV40 promoter/enhancer, RSV LTR promoter/enhancer, and PGK promoter were much less ef®cient (Figs. 1 and 2). The CMV intron A has been demonstrated to improve expression of heterologous genes expressed in vitro by the CMV immediate-early 1 or other promoter/enhancers (Chapman et al., 1991). The bene®cial effect of introns on expression has been ascribed primarily to an enhanced rate of RNA polyadenylation and/or nuclear transport associated with RNA splicing (Huang and Gorman, 1990). Mikkelsen et al. (1992) reported that sequences in intron A are homologous to muscle-speci®c regulatory elements and that the
Z.-L. Xu et al. / Gene 272 (2001) 149±156
Fig. 7. Effect of a second enhancer on luciferase expression in mouse skeletal muscle. The experiment was performed as described in Fig. 1 and in Section 2. The data are expressed as the mean ^ SE (n 8±10). The luciferase levels by each vector were normalized to that of the pGL3Control. The absolute luciferase expression in each cell and tissue transfected with pGL3-Control is described in Fig. 1. (A) Vector containing the hybrid CA promoter/enhancer. (B) Vector containing the improved CMV promoter/enhancer.
CMV promoter/enhancer with intron A conferred high-level expression in muscles of transgenic mice. In this study, our data also showed that the addition of the CMV intron A led to a 2-fold increase in luciferase expression in the context of direct injection of plasmid DNA into mouse skeletal muscle (Fig. 2B-2), but showed only a slight increase in mouse liver (Fig. 2B-1). In cultured cells (HeLa, HepG2, and ECV304), the inclusion of intron A to the CMV promoter/enhancer generated more than 2-fold higher levels of luciferase expression (Fig. 2A). These results suggest that intron A may be a positive element for the high ef®ciency of CMV promoter/enhancer. The ef®ciency of primary RNA transcript processing and polyadenylation is known to vary between the polyadenylation signal sequences of different genes. Moreover, as the rate of transcriptional initiation is increased by the use of strong promoter/enhancers, the process of transcriptional
155
termination may become rate-limiting (Proudfoot, 1989). Though the SV40 P(A) sequence is commonly used in expression vectors, it may not be the best choice for optimal expression. We found that in mouse liver, the SV40 P(A) sequence was less ef®cient than mRBG and BGH P(A) sequences (Fig. 3B-1). The mRBG P(A) sequence has been shown to be ef®ciently recognized in insect and mammalian cells (Westwood et al., 1993). Whereas the activity of both BGH and SV40 P(A) sequences depended on cell or tissue targets, in the present experiments (Fig. 3) the mRBG P(A) sequence showed high activity in almost all cell lines and tissues transfected. This result shows that the high activity of the mRBG P(A) sequence seems to be more ubiquitous than the SV40 and BGH P(A) sequences. However, with the insertion of a second SV40 enhancer downstream of the SV40 P(A) sequence of plasmids which had contained a CMV enhancer, a further increase in expression was achieved (Figs. 5±7 and data not shown). Although this kind of enhanced effect achieved by using a second SV40 enhancer did not occur in HeLa cell and mouse skeletal muscle, the combination of the SV40 P(A) sequence and a second SV40 enhancer was comparable to or more active than only BGH or mRBG P(A) sequence. Therefore, it is anticipated that our strategy of inserting a second SV40 enhancer downstream of the SV40 P(A) sequence can be utilized in the construction of ef®cient vectors. From the overall results, pCMVSL3 (composed of CMV enhancer, CMV promoter, intron A, SV40 P(A) sequence, and SV40 enhancer) and pCASL3 (composed of CMV enhancer, chicken b-actin promoter, chicken b-actin intron, SV40 P(A) sequence, and SV40 enhancer) were the most ef®cient under in vitro and in vivo conditions. In vitro, pCMVSL3 was more ef®cient than was pCASL3 (Fig. 5). In mouse liver, pCMVSL3 was less ef®cient than was pCASL3 (Fig. 6), while in mouse skeletal muscle, pCMVSL3 was more ef®cient than was pCASL3 (Fig. 7). It is worth mentioning that the b-actin promoter shows more ubiquitous activity than does the CMV promoter. Okabe et al. (1997) produced a green mouse, a transgenic mouse line containing the cDNA of green ¯uorescent protein (GFP) under the control of a chicken b-actin promoter and CMV enhancer. All of the tissues from these transgenic lines, with the exception of erythrocytes and hair, were green under excitation light. Sawicki et al. (1998) compared the promoter/enhancer ef®ciency for ubiquitous expression in the skin of adult transgenic mice. They found that a CMV enhancer/ promoter is not active in most cell types in the skin, whereas a CMV enhancer/modi®ed b-actin promoter sequence is active in the suprabasal and basal cells of the epidermis as well as in the epithelial cells of the hair follicles, sebaceous glands, and the dermal papillae. Therefore, the b-actin promoter might be the promoter of choice for high-level and ubiquitous expression in vivo. In summary, we systematically evaluated the combination of various transcriptional regulatory elements in cultured cells and in animal tissues (mouse liver and skeletal
156
Z.-L. Xu et al. / Gene 272 (2001) 149±156
muscle) in vivo. The results of our study provide important information regarding the construction of vectors for gene therapy, vaccine development, and gene transfer experiments. Acknowledgements We would like to thank Mr Jun Murai for his technical assistance. We also thank Dr S. Nakagawa (Osaka University) for his helpful discussion. We would like to thank Dr J. Miyazaki for kindly providing the cytomegalovirus enhancer/chicken b-actin promoter. This work was supported by grants from the Ministry of Health and Welfare of Japan. References Chapman, B.S., Thayer, R.M., Vincent, K.A., Haigwood, N.L., 1991. Effect of intron A from human cytomegalovirus(Towne)immediate-early gene on heterologous expression in mammalian cells. Nucleic Acids Res. 19, 3979±3986. Cheng, L., Ziegelhoffer, P.R., Yang, N.S., 1993. In vivo promoter activity and transgene expression in mammalian somatic tissues evaluated by using particle bombardment. Proc. Natl. Acad. Sci. USA 90, 4455± 4459. Choi, T., Huang, M., Gorman, C., Jaenisch, R., 1991. A generic intron increases gene expression in transgenic mice. Mol. Cell. Biol. 11, 3070±3074. Davis, H.L., Whalen, R.G., Demeneix, B.A., 1993. Direct gene transfer into skeletal muscle in vivo: factors affecting ef®ciency of transfer and stability of expression. Hum. Gene Ther. 4, 151±159. Felgner, P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W., Wenz, M., Northrop, J.P., Ringold, G.M., Danielsen, M., 1987. Lipofection: a highly ef®cient, lipid-mediated DNA-transfection procedure. Proc. Natl. Acad. Sci. USA 84, 7413±7417. Foecking, M.K., Hofstetter, H., 1986. Powerful and versatile enhancerpromoter unit for mammalian expression vectors. Gene 45, 101±105. Graulich, W., Nettelbeck, D.M., Fischer, D., Kissel, T., Muller, R., 1999. Cell type speci®city of the human endoglin promoter. Gene 227, 55±62. Gunter, K.C., Khan, A.S., Noguchi, P.D., 1993. The safety of retroviral vectors. Hum. Gene Ther. 4, 643±645. Guo, Z.S., Wang, L.H., Eisensmith, R.C., Woo, S.L.C., 1996. Evaluation of promoter strength for hepatic gene expression in vivo following adenovirus-mediated gene transfer. Gene Ther. 3, 802±810. Hartikka, J., Sawdey, M., Cornefert-Jensen, F., Margalith, M., Barnhart, K., Nolasco, M., Vahlsing, H.L., Meek, J., Marguet, M., Hobart, P., Norman, J., Manthorpe, M., 1996. An improved plasmid DNA expression vector for direct injection into skeletal muscle. Hum. Gene Ther. 7, 1205±1217. Heller, L., Jaroszeski, M.J., Coppola, D., Pottinger, C., Gilbert, R., Heller, R., 2000. Electrically mediated plasmid DNA delivery to hepatocellular carcinomas in vivo. Gene Ther. 7, 826±829. Huang, M.T., Gorman, C.M., 1990. Intervening sequences increase ef®ciency of RNA 3 0 processing and accumulation of cytoplasmic RNA. Nucleic Acids Res. 18, 937±947.
Liu, D., Knapp, J.E., 2001. Hydrodynamics-based gene delivery. Curr. Opin. Mol. Ther. 3, 192±197. Liu, F., Song, Y.K., Liu, D., 1999. Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther. 6, 1258±1266. Manthorpe, M., Cornefert-Jensen, F., Hartikka, J., Felgner, J., Rundell, A., Margalith, M., Dwarki, V., 1993. Gene therapy by intramuscular injection of plasmid DNA: studies on Fire¯y luciferase gene expression in mice. Hum. Gene Ther. 4, 419±431. Martin-Gallardo, A., Montoya-Zavala, M., Kelder, B., Taylor, J., Chen, H., Leung, F.C., Kopchick, J.J., 1988. A comparison of bovine growthhormone gene expression in mouse L cells directed by the Moloney murine-leukemia virus long terminal repeat, simian virus-40 early promoter or cytomegalovirus immediate-early promoter. Gene 70, 51±56. Mcknight, S., Tjian, R., 1986. Transcriptional selectivity of viral genes in mammalian cells. Cell 46, 795±805. Mikkelsen, T.R., Chapman, B., Din, N., Ingerslev, J., Kristensen, P., Poulsen, K., Hjorth, J.P., 1992. Expression of a cytomegalovirus IE-1-factor VIII cDNA hybrid gene in transgenic mice. Transgenic Res. 1, 164± 169. Mizuguchi, H., Kay, M.A., 1999. A simple method for constructing E1- and E1/E4-deleted recombinant adenoviral vectors. Hum. Gene Ther. 10, 2013±2017. Niwa, H., Yamamura, K., Miyazaki, J., 1991. Ef®cient selection for highexpression transfectants with a novel eukaryotic vector. Gene 108, 193± 199. Niwa, M., Rose, S.D., Berget, S.M., 1990. In vitro polyadenylation is stimulated by the presence of an upstream intron. Genes Dev. 4, 1552±1559. Oellig, C., Seliger, B., 1990. Gene transfer into brain tumor cell lines: reporter gene expression using various cellular and viral promoters. J. Neurosci. Res. 26, 390±396. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., Nishimune, Y., 1997. ªGreen miceº as a source of ubiquitous green cells. FEBS Lett. 407, 313±319. Pasleau, F., Tocci, M.J., Leung, F., Kopchick, J.J., 1985. Growth hormone gene expression in eukaryotic cells directed by the Rous sarcoma virus long terminal repeat or cytomegalovirus immediate-early promoter. Gene 38, 227±232. Ponder, K.P., Dunbar, R.P., Wilson, D.R., Darlington, G.J., Woo, S.L.C., 1991. Evaluation of relative promoter strength in primary hepatocytes using optimized lipofection. Hum. Gene Ther. 2, 41±52. Proudfoot, N.J., 1989. How RNA polymerase II terminates transcription in higher eukaryotes. Trends Biochem. Sci. 14, 105±110. Sawicki, J.A., Morris, R.J., Monks, B., Sakai, K., Miyazaki, J., 1998. A composite CMV-IE enhancer/b-actin promoter is ubiquitously expressed in mouse cutaneous epithelium. Exp. Cell Res. 244, 367±369. Westwood, J.A., Jones, I.M., Bishop, D.H., 1993. Analyses of alternative poly(A) signals for use in baculovirus expression vectors. Virology 195, 90±99. Yang, Y., Li, Q., Ertl, H.C., Wilson, J.M., 1995. Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J. Virol. 69, 2004±2015. Yew, N.S., Wysokenski, D.M., Wang, K.X., Ziegler, R.J., Marshall, J., Mcneilly, D., Cherry, M., Osburn, W., Cheng, S.H., 1997. Optimization of plasmid vectors for high-level expression in lung epithelial cells. Hum. Gene Ther. 8, 575±584.