Development of DNA delivery system using ... - CiteSeerX

Appl Microbiol Biotechnol (2000) 53: 558±567

Ó Springer-Verlag 2000

ORIGINAL PAPER

T.-Y. Chen á C.-T. Hsu á K.-H. Chang á C.-Y. Ting J. Whang-Peng á C.-F. Hui á J. Hwang

Development of DNA delivery system using Pseudomonas exotoxin A and a DNA binding region of human DNA topoisomerase I

Received: 19 July 1999 / Received revision: 10 September 1999 / Accepted: 24 September 1999

Abstract Gene therapy is de®ned as the delivery of a functional gene for expression in somatic tissues with the intent to cure a disease. Thus, highly ecient gene transfer is essential for gene therapy. Receptor-mediated gene delivery can o€er high eciency in gene transfer, but several technical diculties need to be solved. In this study, we ®rst examined the DNA binding regions of the human DNA topoisomerase I (Topo I), using agarose gel mobility shift assay, in order to identify sites of noncovalent binding of human DNA Topo I to plasmid DNA. We identi®ed four DNA binding regions in human DNA Topo I. They resided in aa 51±200, 271±375, 422±596, and 651±696 of the human DNA Topo I. We then used one of the four regions as a DNA binding protein fragment in the construction of a DNA delivery vehicle. Based on the known functional property of each Pseudomonas exotoxin A (PE) domain and human DNA Topo I, we fused the receptor binding and membrane translocation domains of PE with a highly positively charged DNA binding region of the N-terminal 198 amino acid residues of human DNA Topo I. The resulting recombinant protein was examined for DNA binding in vitro and transfer eciency in cultured cells. The results show that this DNA delivery protein is a general DNA delivery vehicle without DNA sequence, topology, and cell-type speci®city. The DNA delivery T.-Y. Chen á C.-F. Hui Institute of Genetics, School of Life Science, National Yang-Ming University, Nankang, Taipei, Taiwan, ROC T.-Y. Chen á C.-F. Hui Institute of Zoology, Academia Sinica, Nankang, Taipei, 115, Taiwan, ROC Tel.: +886-2-27899217 Fax: +886-2-27826085 T.-Y. Chen á C.-T. Hsu á K.-H. Chang á C.-Y. Ting á J. Hwang (&) Institute of Molecular Biology, Academia Sinica, Nankang, Taipei, Taiwan, ROC J. Whang-Peng Clinical Research Center, National Health Research Institutes, Nankang, Taipei, Taiwan, ROC

protein could be used to target genes of interest into cells for genetic and biochemical studies. Therefore, this technique can potentially be applied to cancer gene therapy.

Introduction Gene therapy is de®ned as the delivery of a functional gene for expression in somatic tissues with the intent to cure a disease (Friedmann 1993). Thus, highly ecient gene transfer is essential for gene therapy. Several techniques, such as viral vector, receptor-mediated, and liposome-mediated gene transfer methods have been widely applied to gene transfer in laboratory (Mitani and Caskey 1993). Currently, the preferred method for gene transfer is the use of viral vectors that are able to achieve high eciency gene transfer, and so far, retroviral vectors are the best characterized viral vectors in human gene transfer. However, one limitation of this technique is that retrovirus can only infect dividing cells. Another disadvantage is that only a maximum titer of approx. 106 infectious particles/ml can be obtained. This leads to a limited number of cells that can be infected, especially in vivo. In addition, the ®nding that three monkeys developed malignant T-cell lymphomas after transplantation of retrovirus-transduced bone marrow contaminated with a helper virus raises concern regarding the safety aspect of the viral vector system (Donahue et al. 1992). Due to these disadvantages of viral vectors, receptor-mediated and liposome-mediated gene delivery methods have been used (Wu and Wu 1987, 1988; Wu et al. 1991; Yoshimura et al. 1992). The advantage of the two nonviral vector methods is that they are relatively safe compared to the viral system (Donahue et al. 1992; Mitani and Caskey 1993). However, liposome-mediated gene delivery gives low gene transfer ecacy (Mitani and Caskey 1993; Wilson et al. 1992a, b). Although receptor-mediated gene delivery can o€er high eciency in gene transfer, several technical diculties need to be solved (Wu and Wu 1987, 1988; Wu et al. 1991). One of the problems is that most

559

foreign components, after internalization into cells, have been shown to be compartmentalized and subsequently degraded in the lysosomes. The eciency of the translocation of DNA into nuclei would be greatly impaired if ligand-polylysine-DNA conjugate follows this common intracellular pathway after internalization (Mitani and Caskey 1993; Wagner et al. 1992). To avoid this pathway upon DNA entering the cell, we attempted to introduce the membrane translocation domain of Pseudomonas exotoxin A (PE) into a DNA delivery vehicle, to increase the translocation eciency of DNA into cytosol. It has been shown that using the PE membrane translocation domain, the DNA binding chimeric protein 5EG, which contains the single-chain antibody domain derived from ErbB2 monoclonal antibody, the translocation domain of PE, and the DNA binding domain of Gal 4, is capable of increasing the transfer of DNA from endosome to cytosol four- to six-fold depending on the cell type used (Fominaya et al. 1996). However, the disadvantages of this chimeric protein are that the DNA transferred needs to carry the Gal 4 DNA binding sequence, and that its transfer eciency is celltype-dependent. PE is one of the most toxic components produced by Pseudomonas aeruginosa (Eidels et al. 1983; Middlebrook and Dorland 1984). Based on the information obtained from the three-dimensional structure of PE and the study of PE using recombinant DNA methods, receptor binding, membrane translocation, and ADPribosylation domains have been identi®ed to reside in domain Ia (residues 1±252), domain II (residues 253± 364), and domain III (residues 405±613), respectively (Allured et al. 1986; Chow et al. 1989; Hwang and Chen 1989; Hwang et al. 1987). The receptor binding domain of PE binds to the LDL/a2-macroglobulin receptor on the cell surface in similar quantities (about 1 ´ 104 to 2 ´ 104) in every human cell-type (Feldman et al. 1985; Kounna et al. 1992; Leuven et al. 1979; Moestrup et al. 1992). The membrane translocation domain of PE can increase the translocation eciency of DNA. We utilized these domains of receptor binding and membrane translocation in the development of a general DNA delivery vehicle that has no cell-type speci®city. Then, we need a DNA binding region that has no DNA sequence and topological speci®city for this DNA delivery vehicle. Depending on the nature of the reactants and reaction conditions, topoisomerases catalyze DNA relaxation/supercoiling, catenation/decatenation, and knotting/unknotting reactions (Liu 1983; Wang 1985, 1996). Human DNA Topo I relaxed both negatively and positively supercoiled DNA by catalyzing the transient breakage of a phosphodiester bond in a single DNA strand (Champoux 1990). In addition, eukaryote DNA Topo I had been shown to be associated with actively transcribing genes (Gilmour and Elgin 1987; Rowe et al. 1987; Zhang et al. 1988). Studies have also suggested that topoisomerases were involved in the control of template supercoiling during RNA transcription (Shuman 1991a, b). The crystal structure of human DNA

Topo I in covalent and noncovalent complexes with DNA has been reported (Redinbo et al. 1998; Stewart et al. 1998). DNA Topo I, thus, should bind DNA without DNA sequence and topological speci®city. In order to carry out a more detailed study of the DNA binding regions of the human DNA Topo I, we constructed and puri®ed GST fusion proteins that contained di€erent human DNA Topo I fragments. Using agarose gel mobility shift assay we examined the noncovalent binding of human DNA Topo I fragments to plasmid DNA. We identi®ed four DNA binding regions, and three of them were found for the ®rst time. This DNA binding characterization allowed us to choose a general DNA binding protein fragment for the construction of a general DNA delivery vehicle. Our aim was to develop a general DNA delivery system that has no cell-type speci®city, and no DNA sequence and topological speci®city, so that DNA can be bound regardless of its sequence or its being supercoiled, linear, or nick-cycle. We, therefore, fused the receptor binding and membrane translocation domains of PE with the highly positively charged DNA binding region of the N-terminal 198 amino acid residues of human DNA Topo I. The resulting recombinant protein was examined for DNA transfer eciency. This study shows that this DNA delivery protein has the potential to be used to target genes of interest into cells for genetic and biochemical studies.

Materials and methods Construction of GST-Topo I fusion proteins Using human DNA Topo I cDNA as template, 21 fragments were modi®ed with primers that carried desired restriction sites and ampli®ed via PCR under standard reaction conditions (Chen et al. 1999a). Products were visualized by staining with ethidium bromide (EtBr) after electrophoresis in 6% polyacrylamide gels. PCR products of correct size were electroeluted from gels and cleaved with restriction enzymes. Together with the other two restriction fragments from plasmids that contained human DNA Topo I cDNA (pGTopI-NX and pGTopI-SH), these fragments were subcloned into plasmid pGEX-KG, a glutathione S-transferase (GST) fusion protein expression vector (Bharti et al. 1996; Guan and Dixon 1991; Hwong et al. 1993). After insertion, all cDNA fragments were in frame with the GST gene and con®rmed by DNA sequencing. Expression and puri®cation of GST-Topo I fusion proteins GST fusion proteins were expressed in E. coli DH5a upon isopropyl1-thio-b-D-galactopyranoside (IPTG) induction. To express GST fusion proteins in large quantities in DH5a, colonies of DH5a carrying expression plasmid were pooled and inoculated into a 2-l culture. When cell density reached 3 ´ 107 cells/ml, IPTG was added to a ®nal concentration of 0.5 mM. Cells were harvested 90 min later by centrifugation and resuspended in 25 ml PBST bu€er (phosphatebu€ered saline, 140 mM NaCl, 2 mM KCl, 5 mM Na2HPO4, 2 mM NaH2PO4, 1% Triton X-100, pH 8.0) containing 1 mM phenylmethylsulfonyl ¯uoride (PMSF) (Chen et al. 1999b). Cells were lysed with lysozyme (1 mg/ml) at 25 °C for 5 min. The lysate was centrifuged at 10,000 g for 30 min at 4 °C and the supernatant was applied to a glutathione-agarose anity column (Pharmacia). Unbound

560 proteins were washed out by PBS bu€er. The GST fusion protein was eluted with 50 mM Tris-HCl containing 10 mM glutathione, pH 8.0. Eluted products were checked by 12.5% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Puri®ed fusion proteins were further characterized by rabbit polyclonal anti-GST serum or rabbit polyclonal anti-human DNA Topo I serum (Shiah et al. 1992). These 23 recombinant human DNA Topo I proteins were subcloned in-frame into the 3¢-end of the GST gene and can be recognized by GST polyclonal antibody as well as human DNA Topo I polyclonal antibody (data not shown). Some samples show low-molecular-weight proteins in addition to the major bands in the immunoblot. The appearance of these lower bands might be due to slight degradation of these puri®ed proteins after storage since they did not appear in SDS-PAGE. Construction of DNA carrier vehicle The pET15PMx (formerly pET-15bJIx) vector containing the DNA sequence that encodes the full lengths of PE is cleaved with Sac II and Eco RI. Using human DNA Topo I cDNA as template, the fragment was modi®ed to carry Sac II at the 5¢-end and Eco RI at the 3¢-end and ampli®ed via PCR under standard reaction conditions (Chen et al. 1999a). Products were visualized by staining with EtBr after electrophoresis in 6% polyacrylamide gels. PCR products of correct size were electroeluted from gels and cleaved with restriction enzymes. The fragments were subcloned into the Sac II and Eco RI restricted pET15PMx plasmid. The resulting plasmid was named pPETOPN, which encoded the recombinant protein PE(DIII)-TOPN that contained the receptor binding and membrane translocation domains of PE, and the N-terminal 198 amino acid residues of human DNA Topo I. Expression and puri®cation of His-tag-fusion proteins His-tag-fusion protein PE(DIII)-TOPN, PE(DIII) (expressed by plasmid pET-JJ9 which contained aa 1±425 of PE), or PE(DIII)GnRH [expressed by plasmid pPEGnRH which conjugated aa 1±413 of PE and gonadotropin releasing hormone (GnRH)] were expressed in E. coli BL21 (DE3, LysS) upon IPTG induction (Chen et al. 1999a). The cells were harvested after 90 min IPTG induction, resuspended into lysis bu€er, and sonicated. The lysate was centrifuged at 10,000 g for 30 min at 4 °C, and the pellet was resuspended in 20 ml 8 M urea for 2 h (Chen et al. 1999b). The extract lysate was centrifuged at 10,000 g for 30 min at 4 °C and the supernatant was added into the 8´ binding bu€er. The ®nal supernatant contained 5 mM imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 7.9, and 6 M urea, and was then applied to a His-Bind Ni2+-anity column (Novagen). Unbound proteins were washed out with 60 mM imidazole. The His-tagfusion protein was then eluted with 150 mM imidazole. Eluted products were checked by 12.5% SDS-PAGE. The His-tag-eluted fractions were dialyzed against to 50 mM Tris-HCl, pH 7.5. The dialyzed proteins were concentrated using a YM-10 ®lter from Amicom. The concentrated proteins were then checked by SDSPAGE. Antiserum and immunoblotting Polyclonal anti-PE raised in rabbits was raised as described (Hwang and Chen 1989). Polyclonal anti-human DNA Topo I sera were raised as described (Hwong et al. 1993). Anti-GST serum was purchased from Oncogene Research. SDS-PAGE was performed according to the method of Laemmli (Laemmli 1970). Samples containing puri®ed human DNA Topo I fusion proteins or PE(DIII)-TOPN were boiled for 5 min prior to SDS-PAGE. After electrophoresis, samples were transferred to polyvinylidene di¯uoride membrane (PVDF, Millipore) by electrotransfer in Tris-glycine bu€er (2.5 mM Tris, 9.2 mM glycine, pH 8.3) containing 20% methanol under the condition of 20 mA at 4 °C for 4 h. The

membrane was incubated with a 1:5000 dilution of the anti-human DNA Topo I serum, anti-GST serum, or anti-PE serum, followed by interaction with anti-rabbit IgG alkaline phosphatase conjugated antibody as secondary antibody and stained with 5-bromo4-chloro-3-indolylphosphate (BCIP) and nitro blue tetrazolium (NBT) (Chen et al. 1999b). Agarose gel mobility shift assay Reactions (15 ll) containing pCMV plasmid (0.5 lg) DNA and GST-Topo I fusin proteins in 50 mM Tris (pH 7.5), were incubated for 5 min at room temperature (Schneider et al. 1996). Samples were analyzed by electrophoresis through 0.8% agarose gel in TBE bu€er run at 4.1 V/cm, and were visualized by staining with EtBr. Reaction conditions for pGEM-bgal plasmid (Promega, 1 lg) DNA and PE(DIII)-TOPN fusion protein or other control proteins were similar, and followed by electrophoresis and visualization by staining with EtBr. Immunoblotting of agarose gel mobility shift assay After electrophoresis of the pCMV-p53 plasmid DNA (pCMV-p53 plasmid contained a full length p53 cDNA) and PE(DIII)-TOPN fusion protein or other control protein reactions, samples were transferred to a nitrocellulose membrane in 0.5´ TBE bu€er (2.5 mM Tris, 9.2 mM glycine, pH 8.3) at room temperature for 12 h. The membrane was incubated with a 1:5000 dilution of antiPE serum, followed by interaction with anti-rabbit IgG alkaline phosphatase-conjugated antibody as secondary antibody and stained with BCIP and NBT. Cells and cell culture Human ovarian cancer cell line A2780 cells were maintained in plastic dishes containing RPMI medium with 10% fetal calf serum at 37 °C and 5% CO2. The human cervical carcinoma cell line Hela and the human hepatoma cell line Hep G2 cells were maintained in plastic dishes containing Dulbecco's modi®ed Eagle medium (DMEM). DNA carrier assay The A2780 cells were seeded to 2 ´ 104 cells in RPMI medium on 12-well tissue culture plates, and were cultured overnight at 37 °C. The cells were washed with 1´ PBS before incubation with plasmid DNA and recombinant protein. pEGFP-N1 plasmid (Clonetech, 1 lg) that encoded the green ¯uorescence protein (GFP) together with 5 lg each of the protein PE(DIII), PE(DIII)-TOPN, or GST-TOPI(1±200) (expressed from plasmid pGTOPI-F12) were incubated in 50 mM Tris-HCl pH 7.5 for 5 min at room temperature, then mixed with 1 ml RPMI medium. The incubated plasmid and recombinant protein were added to A2780 cells, and cultured for 48 h in 5% CO2, at 37 °C. UV excitation and a 490nm ®lter were used to observe the ¯uorescence, which was photographed. The Hela and Hepa G2 cells were seeded to 2 ´ 104 cells in DMEM, and all assay procedures were similar to those for the A2780 cells.

Results Screening of the human DNA topoisomerase I DNA binding regions by gel mobility shift analysis Using the agarose gel mobility shift assay, we were able to identify DNA binding regions in the Topo I

561

fragments of the GST-Topo I fusion proteins (Fig. 1). The results of the DNA binding anity of the puri®ed GST-Topo I fusion proteins are shown in Fig. 2, and these results were summarized (Fig. 1). With the exception of fragments SH, F5, F7, F18, F20, and F21, all other Topo I fragments were found to bind DNA. We analyzed and summarized the smallest regions of amino acid sequences that can bind DNA, and the results show the existence of at least four DNA binding regions. These four regions correspond to the amino acid (aa) sequences 51±200, 271±375, 422±596, and 651±696 aa of the human DNA Topo I. The highly conserved core domain (aa 198±650) contains two DNA binding regions, while the unconserved N-terminal (aa 1±197) and linker (aa 651±696) domains also contain one DNA binding region, respectively. We chose the N-terminal region (aa 3±200) of human DNA Topo I to construct the DNA binding region of the DNA delivery vehicle.

Fig. 1 Maps of GST-Topo I fusion proteins. Plasmids for expression of fusion proteins that contained the GST and various regions of human DNA Topo I are shown. The numbers on the serial human DNA Topo I fragments represent the amino acid positions on the human DNA Topo I protein. The results shown in Fig. 2 are summarized on the right-hand column by + and ) signs, representing presence or absence of DNA binding ability, respectively

Construction and protein puri®cation of DNA carrier vehicle The strategy to engineer PE into a DNA delivery vehicle is shown in Fig. 3A. The plasmid encoding the protein that composed the receptor binding and membrane translocation domains of PE and the highly positively charged domain of human DNA Topo I was constructed as shown in Fig. 3B. The pET15PMx vector containing the DNA sequence that encoded the receptor binding and membrane translocation domains of PE was ligated with the N-terminal 198 (aa 3±200) amino acid residues of the human DNA Topo I. The resulting plasmid that encoded the recombinant protein PE(DIII)TOPN was named pPETOPN. The recombinant protein that could be induced by IPTG was expressed in E. coli BL21(DE3, LysS). The induced protein was partially extracted by 8 M urea, then puri®ed by Ni2+ anity column, and was checked by SDS-PAGE gel with Coomassie blue staining. The His-tag puri®ed recombi-

562

Fig. 2 Identi®cation of DNA binding regions by agarose gel mobility shift assay. Results of the agarose gel mobility shift assay to identify the DNA binding ability of various GST-Topo I serial fusion proteins are shown as EtBr-stained agarose gel. BSA and GST proteins were used as controls

nant protein could be seen to have been puri®ed to nearhomogeneity. The solubility of the protein after removal of urea and concentration was determined to be 0.3 lg/ll. The immunoblotting results also showed that the puri®ed protein cross-reacted with anti-PE and anti-human DNA Topo I polyclonal antibodies (data not shown).

lowest mobility when protein amount reached 5 lg [the molar ratio of protein (3.72 mM)/DNA (15 lM) was approximately 250/1] (Fig. 5A). These results show that the DNA mobility shift is dependent on the recombinant protein amount added. When the same agarose gel was immunoblotted with anti-PE polyclonal antibody, the cross-reacted protein positions coincided with the DNA positions of the EtBr-stained gel (Fig. 5B). These results show that the DNA mobility shift is speci®cally caused by DNA binding to the recombinant protein PE(DIII)TOPN in a concentration-dependent manner.

Determination of DNA-carrier complex in vitro

Determination of DNA-carrier complex in cultured cells

The recombinant protein PE(DIII)-TOPN and control proteins PE(DIII) and PE(DIII)-GnRH were then tested for DNA binding activity using the agarose gel mobility shift assay and immunoblotting with anti-PE polyclonal antibody (Chow et al. 1989; Shang et al. 1996). For 1 lg supercoiled plasmid DNA (pGEM-bgal), mobility shift of the DNA bands started to occur when the recombinant protein PE(DIII)-TOPN reached 1 lg, and most of the DNA shifted to slower mobility when protein reached 5 lg (Fig. 4A). However, when control proteins such as BSA, PE, or PE(DIII)-GnRH were used, no such shift could be observed (Fig. 4B). When the amount of the recombinant protein PE(DIII)-TOPN added was increased from 1 lg to 5 lg by 1-lg increments, the DNA mobility decreased gradually and reached its

The preincubated supercoiled plasmid DNA pEGFPN1, and the recombinant protein PE(DIII), PE(DIII)TOPN, or GST-TOPI(1±200) were transferred into cultured A2780 cells. After 48 h in 5% CO2 at 37 °C, the transfer eciency was assessed using UV to excite and a 490-nm ®lter to observe the ¯uorescence emitted by the expressed GFP protein (Fig. 6). The experiment was carried out in triplicate, and on each plate of 2 ´ 104 cells, patches of 500 cells were counted for transfer ef®ciency. The results show that only PE(DIII)-TOPN was able to deliver the plasmid pEGFP-N1 into cells. The transfer eciency for PE(DIII)-TOPN was on average 10 ‹ 2 cells/500 cells, which is similar to that for liposome (9 ‹ 2 cells/500 cells). Therefore, these results show that PE(DIII)-TOPN is capable of delivering DNA

563 Fig. 3A, B Strategy for engineering PE into a DNA delivery vehicle and constructing the DNA delivery protein PE(DIII)TOPN. A The strategy for engineering PE into a DNA delivery vehicle was to retain the receptor binding and membrane translocation domains of PE and to replace the ADP-ribosylation domain of PE with a DNA binding domain. B Strategy to construct plasmid for expression of the DNA delivery protein PE(DIII)-TOPN

564

Fig. 4A, B DNA binding ability of the DNA delivery protein PE(DIII)-TOPN as determined the by agarose gel mobility shift assay. A The ability to cause a DNA mobility shift by increasing amounts of the DNA delivery protein PE(DIII)-TOPN was examined by agarose gel mobility shift assay. The DNA used was plasmid pGEM-bgal. B The ability of PE(DIII)-TOPN to cause DNA mobility shift was compared with that of control proteins such as BSA, PE, and PE(DIII)-GnRH. The amounts of DNA and protein used in each reaction were 1 lg and 5 lg respectively

into cells and its transfer eciency is comparable to that of liposome transfection. The results obtained using Hela and Hep G2 cells were comparable to those for the A2780 cells.

Discussion In this study using the agarose gel mobility shift assay we have identi®ed four DNA binding regions in human DNA Topo I. These four regions are located at the amino acid sequences 51±200, 271±375, 422±596, and 651±696 of human DNA Topo I. The highly conserved core domain contains two DNA binding regions, while the unconserved N-terminal and linker domains contain one DNA binding region each. According to our results, these four DNA binding regions should have no DNA sequence and topological speci®city (Fig. 2). Also, the N-terminal region of human DNA Topo I contained the nuclear localization signal (NLS) that might help the delivery of DNA to the nucleus (Alsner et al. 1992). Thus, we fused the receptor binding and membrane translocation domains of PE with the highly positively charged DNA binding region of the N-terminal 198 amino acid residues of human DNA Topo I. The resulting recombinant protein was examined for DNA transfer eciency. We, therefore, have developed a general DNA delivery system that has no cell-type speci®city, and no DNA sequence and topological speci®city, so that DNA can be bound regardless of its sequence or its being supercoiled, linear, or nick-cycle.

Fig. 5A, B DNA mobility shift was caused by the DNA delivery protein PE(DIII)-TOPN. A The DNA mobility shift caused by increasing amounts of the DNA delivery protein PE(DIII)-TOPN from 1 lg, 2 lg, 3 lg, 4 lg, to 5 lg was shown in an EtBr-stained agarose gel. The DNA used was plasmid pCMV-p53. B The agarose gel was then transferred to a nitrocellulose paper and reacted with anti-PE antibody to locate the DNA delivery protein PE(DIII)-TOPN positions

Many gene transfer methods have been extensively used in recent years, but they all have inherent limitations and disadvantages. Techniques including calcium phosphate precipitation, DEAE-dextran cell fusion, and electroporation have been used in in vitro experiments, while others including virus infection, microinjection, and liposome-mediated transfer have been used in in vivo experiments (Friedmann 1993; Mitani and Caskey 1993). However, ecient gene transfer such as is necessary for genetic analysis has not yet been successfully achieved. A novel method still needs to be developed so that gene can be transferred in vitro and in vivo, then genetic analysis can be carried out in higher eukaryotes. Receptor-mediated DNA transfer may be one such method (Wu and Wu 1987, 1988; Wu et al. 1991). The method of receptor-mediated DNA transfer is based on the concept of ligand binding to receptor, then entering the cell by the receptor-mediated endocytosis pathway. A similar method has been presented that involves coupling asialoorosomucoid (AsOR) (Wu and Wu 1987, 1988; Wu et al. 1991; Yoshimura et al. 1992) or transferrin (Wagner et al. 1992) to polylysine to carry a gene of interest into cells. Therefore, choosing a highly speci®c and ecient ligand should increase DNA transfer. PE is a highly speci®c ligand to the LDL/ a2-macroglobulin receptor of mammalian cells (Eidels et al. 1983). Since all mammalian cells have this receptor, the DNA delivery vehicle using this receptor binding domain should have no cell-type speci®city. The translocation domain for toxin translocation into cytosol may increase the translocation eciency of DNA into cytosol (Fominaya and Wels 1996). Employing the receptor binding and membrane translocation domains of PE as

565 Fig. 6A±E Demonstration of DNA transfection ability of the DNA delivery protein PE(DIII)TOPN in cultured cells. Plasmid pEGFP-N1 (1 lg) was transfected to the A2780 cells. A Plasmid alone. B Plasmid+PE(DIII). C Plasmid+PE(DIII)-TOPN. D Plasmid+GST-TOPI(1±200). E. Plasmid+liposome. Arrows indicate the pEGFP-N1 transfected cells. The transfected cells were excited by UV and the ¯uorescence emitted was observed using a 490-nm ®lter

a receptor-mediated DNA transfer vehicle, the modi®ed PE can bind to receptor, transfect into all cells of human, and has better translocation eciency into cytosol; but since the cell-killing ADP-ribosylation domain has been replaced by the N-terminal region of human DNA Topo I that can bind DNA, it would not kill cells. The advantage of using the N-terminal region of human DNA Topo I to bind DNA is that it has no sequence or topological speci®city towards DNA. The complexing of this recombinant protein PE(DIII)-TOPN and DNA of the gene of interest should bind to PE

receptors of the target cells and enter cell via receptormediated endocytosis. Using this general DNA delivery system, we successfully delivered the supercoiled plasmid DNA pEGFP-N1 into A2780, Hela, and Hep G2 cells with comparable eciency to liposome transfection, and observed the green ¯uorescence emitted by the expressed GFP protein. This general DNA delivery protein can carry plasmid and oligonucleotide into its target cell, and should help to produce transient or stable changes in the genome of target cells. Finally, this delivery protein may have the potential to be used in gene therapy or

566

cancer therapy, but more experiments need to be conducted in the future in order to maximize the transfer eciency of this recombinant protein. Acknowledgments This study was supported by grants from Academia Sinica and the National Science Council (NSC86-2311-B001-077 and NSC87-2312-B-001-007).

References Allured VS, Collier RJ, Carroll SF, Mckay DB (1986) Structure of exotoxin A of Pseudomonas aeruginosa at 3.0-angstrom resolution. Proc Natl Acad Sci USA 83: 1320±1324 Alsner J, Svejstrup JQ, Kjeldsen E, Sùrensen BS, Westergaard O (1992) Identi®cation of an N-terminal domain of eukaryotic DNA topoisomerase I dispensable for catalytic activity but essential for in vivo function. J Biol Chem 267: 12408±12411 Bharti AK, Olson MO, Kufe DW, Rubin EH (1996) Identi®cation of a nucleolin binding site in human topoisomerase I. J Biol Chem 271: 1993±1997 Champoux JJ (1990) Mechanistic aspects of type-I topoisomerases. In: Cozzarelli NR, Wang JC (eds) DNA topology and its biological e€ects. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp 217±242 Chen TY, Shang HF, Chen TL, Lin CP, Hui CF, Hwang J (1999a) Recombinant protein composed of Pseudomonas exotoxin A, outer membrane proteins I and F as vaccine against Pseudomonas aeruginosa infection. Appl Microbiol Biotechnol 52: 524±533 Chen TY, Lin CP, Loa CC, Chen TL, Shang HF, Hwang J, Hui CF (1999b) A nontoxic Pseudomonas exotoxin A induces active immunity and passive protective antibody against Pseudomonas exotoxin A intoxication. J Biomed Sci 6: 357±363 Chow JT, Chen MS, Wu HCP, Hwang J (1989) Identi®cation of the carboxyl terminal amino acids important for the ADP-ribosylation activity of Pseudomonas exotoxin A. J Biol Chem 264: 18818±18823 Donahue RE, Kessler SW, Bodine D, McDonnagh K, Dunbar C, Goodman S, Agricola B, Byrne E, Raeld M, Moen R, Bacher J, Zsebo KM, Nienhuis AW (1992) Helper virus induced T cell lymphoma in nonhuman primates after retroviral mediated gene transfer. J Exp Med 176: 1125±1135 Eidels L, Proia RL, Hart DJ (1983) Membrane receptors for bacterial toxin. Microbiol Rev 47: 596±620 Feldman SR, Rosenberg MR, Ney KA, Michalopoulos G, Pizzo SV (1985) Binding of a2-macroglobulin to hepatocytes: mechanism of in vivo clearance. Biochem Biophys Res Commun 128: 795±802 Fominaya J, Wels W (1996) Target cell-speci®c DNA transfer mediated by a chimeric multidomain protein: novel non-viral gene delivery system. J Biol Chem 271: 10560±10568 Friedmann T (1993) Gene therapy ± a new kind of medicine. Trends Biotechnol 11: 156±159 Gilmour DS, Elgin SC (1987) Localization of speci®c topoisomerase I interactions within the transcribed region of active heat shock genes by using the inhibitor camptothecin. Mol Cell Biol 7: 141±148 Guan KL, Dixon JE (1991) Eukaryotic proteins expressed in Escherichia coli: an improved thrombin cleavage and puri®cation procedure of fusion proteins with glutathione S-transferase. Anal Biochem 192: 262±267 Hwang J, Chen MS (1989) Structure and function relationship of Pseudomonas exotoxin A: an immunochemical study. J Biol Chem 264: 2379±2384 Hwang J, Fitzgerald DJ, Adhya S, Pastan I (1987) Functional domains of Pseudomonas exotoxin identi®ed by deletion analysis of the gene expressed in E. coli. Cell 48: 129±136

Hwong CL, Chen CY, Shang HF, Hwang J (1993) Increased synthesis and degradation of DNA topoisomerase I during the initial phase of human T lymphocyte proliferation. J Biol Chem 268: 18982±18986 Kounnas MZ, Morris RE, Thompson MR, FitzGerald DJ, Strickland DK, Saelinger CB (1992) The a2-macroglobulin receptor/low density lipoprotein receptor-related protein binds and internalizes Pseudomonas exotoxin A. J Biol Chem 267: 12420±12423 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680±685 Leuven FV, Cassiman JJ, Berghe HVD (1979) Demonstration of an a2-macroglobulin receptor in human ®broblasts, absent in tumor-derived cell lines. J Biol Chem 254: 5155±5160 Liu LF (1983) DNA topoisomerases: enzymes that catalyse the breaking and rejoining of DNA. CRC Crit Rev Biochem 15: 1±24 Middlebrook JL, Dorland RB (1984) Bacterial toxins: cellular mechanism of action. Microbiol Rev 48: 199±221 Mitani K, Caskey CT (1993) Delivering therapeutic genes ± matching approach and application. Trends Biotechnol 11: 162±166 Moestrup SK, Gliemann J, Pallesen G (1992) Distribution of the a2-macroglobulin receptor/low density lipoprotein receptor-related protein in human tissues. Cell Tissue Res 269: 375±382 Redinbo MR, Stewart L, Kuhn P, Champoux JJ, Hol WGJ (1998) Crystal structure of human topoisomerase I in covalent and noncovalent complexes with DNA. Science 279: 1504±1513 Rowe TC, Couto E, Kroll DJ (1987) Camptothecin inhibits hsp70 heat-shock transcription and induces DNA strand breaks in hsp70 gene in Drosophila. NCI Monogr 4: 49±53 Schneider H, Huse K, Birkenmieier G, Otto A, Scholz GH (1996) Gene transfer mediated by a2-macroglobulin. Nucleic Acids Res 24: 3873±3874 Shang HF, Yeh ML, Lin CP, Hwang J (1996) Characterization of monoclonal antibody B7, which neutralizes the cytotoxicity of Pseudomonas aeruginosa exotoxin A. Clin Diagn Lab Immunol 3: 727±732 Shiah HS, Chen TY, Chang CM, Chow JT, Kung HJ, Hwang J (1992) Pseudomomas exotoxin A-epidermal growth factor (EGF) mutant chimeric protein as an indicator for identifying amino acid residues important in EGF-receptor interaction. J Biol Chem 267: 24034±24040 Shuman S (1991a) Site-speci®c DNA cleavage by vaccinia virus DNA topoisomerase I: role of nucleotide sequence and DNA secondary structure. J Biol Chem 266: 1796±1803 Shuman S (1991b) Site-speci®c interaction of vaccinia virus topoisomerase I with duplex DNA: minimal DNA substrate for strand cleavage in vitro. J Biol Chem 266: 11372±11379 Stewart L, Redinbo MR, Qiu X, Hol WGJ, Champoux JJ (1998) A model for the mechanism of human topoisomerase I. Science 279: 1534±1541 Wagner E, Zatloukal K, Cotton M, Kirlappos H, Mechtler K, Curiel DT, Birnstiel ML (1992) Coupling of adenovirus to transferrin-polylysine/DNA complexes greatly enhances receptor-mediated gene delivery and expression of transfected gene. Proc Natl Acad Sci USA 89: 6099±6103 Wang JC (1985) DNA topoisomerases. Annu Rev Biochem 54: 665±697 Wang JC (1996) DNA topoisomerases. Annu Rev Biochem 65: 635±692 Wilson JM, Grossman M, Wu CH, Chowdhury NR, Wu GY (1992a) Hepatocyte-directed gene transfer in vivo leads to transient improvement of hypercholesterolemia in low density lipoprotein receptor-de®cient rabbits. J Biol Chem 267: 963±967 Wilson JM, Grossman M, Carbrera J, Wu CH, Wu GY (1992b) A novel mechanism for achieving transgene persistence in vivo after gene transfer into hepatocytes. J Biol Chem 267: 11483± 11489

567 Wu GY, Wu CH (1987) Receptor-mediated in vitro gene transformation by a soluble DNA carrier system. J Biol Chem 262: 4429±4432 Wu GY, Wu CH (1988) Receptor-mediated gene delivery and expression in vivo. J Biol Chem 263: 14621±14624 Wu GY, Wilson JM, Shalaby F, Grossman M, Shafritz DA, Wu CH (1991) Receptor-mediated gene delivery in vivo. Partial correction of genetic analbuminemia in nagase rats. J Biol Chem 266: 14338±14342

Yoshimura K, Rosenfeld MA, Nakamura H, Scherer EM, Pavirani A, Lecocq JP, Crystal RG (1992) Expression of the human cystic ®brosis transmembrane conductance regulator gene in the mouse lung after in vivo intratracheal plasmid-mediated gene transfer. Nucleic Acids Res 20: 3233±3240 Zhang H, Wang JC, Liu LF (1988) Involvement of DNA topoisomerase I in trabscription of human ribosomal RNA gene. Proc Natl Acad Sci USA 85: 1060±1064