ARTICLE FIG. 1. Heterodimer strategy. (A) Map of the B-domain (3.3 kb, the whole exon 14) deleted human F8 cdna. The arrow and the dotted line represe

Transcription

1 Expression of Human Factor VIII by Splicing between Dimerized AAV Vectors Hengjun Chao, 1,2 Liangwu Sun, 3 Andrew Bruce, 1 Xiao Xiao, 3 and Christopher E. Walsh 1,2, * 1 UNC Gene Therapy Center, University of North Carolina at Chapel Hill, North Carolina 27599, USA 2 Department of Medicine, University of North Carolina at Chapel Hill, North Carolina 27599, USA 3 University of Pittsburgh School of Medicine, Department of Molecular Genetics and Biochemistry, Pittsburgh, Pennsylvania 15261, USA *To whom correspondence and reprint requests should be addressed. Fax: (919) cwalsh@med.unc.edu. Adeno-associated virus (AAV) is a useful vector for hemophilia gene therapy, but the limited effective packaging capacity of AAV (5 kb) appears to be incompatible with factor VIII (gene symbol F8) cdna (7 kb). Although we previously demonstrated efficient packaging and expression of B-domain deleted human F8 (BDD-F8) using a single AAV vector, the packaging limit still excludes the use of large/strong regulatory elements. Here we exploited the split AAV vector technology that expands the packaging capacity of AAV through head-to-tail dimerization. To test the feasibility of AAV heterodimerization for F8 expression, we generated a 5 vector that includes a large enhancer/promoter cassette linked with exons 1 12 of the F8 cdna and a half-intron-carrying splice donor site. A complementing 3 vector contains another halfintron-carrying splice acceptor site linked with the remaining F8 cdna and a polyadenylation signal. Following coinfection of 293 and HepG2 cells, the 5 and 3 vectors together produced functional human factor VIII protein at a level of 120 mu/ml (24 ng/ml). No factor VIII protein was detected if only one of the vectors was used. Correct head-to-tail vector dimerization as well as spliced BDD-F8 mrna was detected by DNA PCR and RT-PCR, respectively. Furthermore, intraportal injection of two raav/f8 vectors in immunodeficient mice produced 2% of the normal level of factor VIII for four months. Our results demonstrate the potential use of AAV dimerization for F8 expression. Key Words: parvovirus, hemophilia, gene transfer, liver, RNA splicing INTRODUCTION Gene therapy offers a promising approach to treat genetic diseases like hemophilia A, in which a sustained therapeutic level of factor VIII protein is required. Recombinant adeno-associated virus (raav) has attracted interest as a promising vector for protein-deficient disorders such as hemophilia [1,2]. A single-stranded DNA virus member of the parvovirus family, AAV contains 4675 nt and is dependent on a helper virus for replication [1]. Not only does AAV have a broad host cell tropism, but it also induces little cellular immune response, is nonpathogenic, and can exist in a latent state [1]. A major disadvantage of raav is its limited capacity to package large (> 5.0 kb) cdnas efficiently [3,4]. The human F8 cdna consists of 26 exons; its 7-kb cdna size was thought to be too large to be packaged into raav. The fully mature protein requires cleavage of the full-length protein into heavy and light chains that self-associate. The heavy chain comprises exons 1 12 and the light chain exons The protein domains of factor VIII constitute A1, A2, B, A3, C1, and C2. The B-domain (corresponding to exon 14) is dispensable for full function [5]. BDD-F8 cdna is 4.5 kb in size, and the corresponding protein has the equivalent pharmacological properties of full-length factor VIII [5,6]. The apparent incompatibility between limited AAV packaging size and the large size of F8 can be bypassed using several approaches. We previously reported that the B-domain deleted human factor VIII cdna (4.5 kb) could be packaged and could express long-term, therapeutic levels of human factor VIII in vivo [4]. Despite this improvement, the limited raav packaging capacity still restricts the use of larger regulatory elements that may improve F8 expression. Another approach placed the coding sequences of the individual heavy and light chains of factor VIII into two separate AAV vectors [7]. Heavy- and light-chain polypeptides self-associate to form functional factor VIII [6]. However, liver cells producing either one of the chains predominated, with secretion of unbound heavy and light chains [7] /02 $35.00

2 ARTICLE FIG. 1. Heterodimer strategy. (A) Map of the B-domain (3.3 kb, the whole exon 14) deleted human F8 cdna. The arrow and the dotted line represent the site at which exon 12 was divided. (B) Map of the dual AAV/F8 vectors. The junction sequences of splicing donor (SD) and splicing acceptor (SA) are shown. Exon 12 sequence is underlined. pdlz17a3, 5 vector carrying the CBA promoter; pdlz17c, 3 vector. ITR, AAV inverted terminal repeat. P CBA, CMV enhancer plus chicken -actin promoter. Intron, truncated human FIX intron I. P(A) n, bovine growth hormone polyadenylation sequence. A B Recombinant AAV can exist in a variety of forms in transduced cells. The virus can be found as single-genome size, circular concatemers, or large linear concatemeric forms in a preferred head-to-tail orientation [8 10]. We exploited this unique feature of AAV to form head-to-tail concatemers as an alternative approach to expand raav packaging capacity [11]; this approach was also demonstrated by other groups [12,13]. We earlier demonstrated the feasibility of this technique by splitting a -galactosidase marker gene into two vectors, each carrying an intron spacer and appropriate splicing signal [11]. Coinfection of both vectors in vitro and in vivo produced appropriately spliced full-length mrna and transgene protein [11]. This novel strategy extends the raav packaging capacity up to 10 kb, providing an opportunity to use both large regulatory elements and large transgene cdna [11,14]. In light of those results, we set out to determine if the human F8 cdna could be manipulated in a similar fashion and could express functional factor VIII protein. Here we report that through heterodimerization of two distinct raav vectors, each carrying a portion of the F8 cdna, factor VIII is expressed in vitro and in vivo. RESULTS raav Dimerization and Expression of Factor VIII Protein in Vitro The F8 cdna is encoded by 26 exons, with exons 1 12 encoding the heavy chain and the remaining exons encoding the light chain [6]. Exon 14 encodes the B- domain [5,6]. The human B-domain deleted F8 (BDD- F8) cdna was split into two vectors at exon 12 (Fig. 1A). The upstream or 5 vector carries the promoter, F8 exons 1 12, and splicing intron; exon 12 was divided at the BamHI site for the insertion of a splice donor site (termed exon 12 ). The split of exon 12 and the lack of a poly(a) signal precluded the production of heavy-chain polypeptide. The 3 vector carries intronic sequence, the remaining exon 12 (exon 12 ) and exon 13, sequence, and a polyadenylation site. Several 5 vectors (pdlz17a, pdlz17a-i, pdlz17a2, pdlz17a3, pdlz17a3-i) carrying different enhancer/promoter elements were cotransfected with the 3 pdlz17c vector plasmid (Fig. 2A). To test the functionality of the expression cassette, cotransfection was carried out in 293 cells and assayed by ELISA and Coatest, for antigen and activity of human factor VIII protein, respectively (Fig. 2B). As expected, no factor VIII antigen or activity was detected from the 293 cells transfected with either 5 or 3 vector plasmid alone (data not shown). Factor VIII antigen and activity values were in agreement, verifying the expression of fully functional factor VIII using the dual raav/f8 vectors approach. Human factor VIII antigen and activity were also detected in the medium of HepG2 cells cotransfected with the dual raav/f8 plasmids, though at lower levels (data not shown). Lower transduction efficiency of HepG2 cells than 293 cells (20% versus 50%, data not shown) may account for lower factor VIII yield from HepG2 cells. The cassette carrying the cytomegalovirus (CMV) promoter (pdlz17a2) produced more factor VIII in both 293 cells (Fig. 2B) and HepG2 cells (data not shown) than either the LSP (liver-specific) or the CBA (CMV enhancer plus chicken -actin) promoter. Although the CMV promoter produced more factor VIII expression in vitro, considering the report of transcription silencing of the CMV promoter in liver in vivo [15], pdlz17a3 (carrying CBA promoter) was chosen for in vivo investigation using the dual raav/f8 vector approach. 717

3 A FIG. 2. Dimerization AAV vector contructs. (A) Maps of dual raav/f8 vectors. The 5 vectors are termed pdlz17a, A1-I, A2, A3, and A3-I containing exon 1 12 F8, FIX intron I, and promoter LSP; LSP+intron, CMV, CBA, and CBA+intron, respectively. The 3 vector is termed as pdlz17c, containing exon of F8 cdna, intron I, and poly(a) signal sequence. All cassettes were subcloned into a ptr-uf5 backbone containing two intact AAV ITRs. ITR, AAV inverted terminal repeat. LSP (liver-specific promoter) carries the thyroid-binding globulin promoter plus the human -microglobulin/bikunin enhancer. LSP+intron, liver-specific promoter plus rabbit -globin-like intron. CMV, the immediate early promoter of cytomegalovirus. CBA, CMV enhancer plus chicken -actin promoter. CBA+intron. CBA promoter plus the chicken -actin intron I. 5 Intron refers to the first half of the human FIX intron I; 3 intron refers to the second half of the FIX intron 1 sequence. P(A) n, the bovine growth hormone (bgh) polyadenylation signal sequence. (B) Human FVIII production. After plating cells in each well of six-well plates overnight, cells were transfected with 3 g of each 5 - and 3 -vector plasmid. The medium was harvested for analysis (by ELISA and Coatest) and replaced with fresh medium every 24 hours post-transfection. The data are presented as FVIII activity and antigen levels at 48 hours post-transfection. No FVIII was detected when either vector plasmid was used (data not shown). Results are mean ± SEM of four experiments, each done in duplicate. Filled bar, FVIII activity assayed by Coatest; shaded bar, FVIII antigen assayed by ELISA. B Recombinant AAV2 viral vectors were generated from plasmid pdlz17a3 and pdlz17c. Both dot-blot and infectious-center assays were used to verify that the titers of each preparation were equivalent (~ AAV vector genomes/ml). Following coinfection of 293 cells with raav/17a3 plus raav/17c, Coatest measured fully functional BDD factor VIII. The factor VIII expression level varied depending on the ratio of input of 5 and 3 raav/f8 vector genomes (Fig. 3). The ratio of 5 -vector genomes to 3 -vector genomes of 1:5 yielded the highest level of factor VIII. This ratio differs from our observation with other AAV vector pairs [11] (L.S. and X.X., unpublished data). This may be explained by differences in virus titration, tropism, and duplex formation of each vector. Molecular Confirmation of raav Dimerization To investigate the molecular conversion of the two vectors, low-molecular-weight (Hirt) DNA was isolated from the 293 cells coinfected with both of the dual raav/f8 vectors. We designed PCR primers to amplify the junction of the raav concatemers (Fig. 4B). Different combinations of the four PCR primers were used to amplify the junction fragments of homo- and hetero-concatemers of two raav vectors in head-to-tail, head-to-head, and tail-to-tail sequences (Fig. 4C). We could amplify the junction fragments of the concatemers of two raav vectors converted in homo- or hetero- head-to-tail arrays (Fig. 4A), but failed to generate PCR products reflecting the raav molecular conversion in head-to-head or tailto-tail arrays. This may mirror the prevailing patterns of the molecular conversions of AAV vectors [8,9,16 20], although we could not completely exclude failure of PCR amplification crossing the very GC-rich domain of raav inverted terminal repeats (ITRs) between two raav vectors arrayed in the head-to-head or tail-to-tail sequence. The 960-bp fragment (lane 7) of the head-to-tail dimers in raav/17a plus raav/17c sequence was amplified (Fig. 4), indicating formation of head-to-tail heterodimers of 5 and 3 AAV/F8 vectors, which accounts for expression 718

4 ARTICLE FIG. 3. Human FVIII expression from 293 cells infected with raav/dlz17a3 and raav/dlz17c at different input ratios. After plating cells in each well of sixwell plates overnight, cells were exposed to different ratios of virus plus adenovirus type 5 (MOI = 1). Equal input is represented by raav vector genomes of raav/17a3 and raav/17c per cell. The medium was harvested for analysis (by ELISA and Coatest) and replaced with fresh medium every 24 hours post-transfection. The data are expressed as FVIII activity at 48 hours post-transduction. No FVIII was detected if either vector was used alone (data not shown). Analysis was conducted using duplicate samples, and each bar represents a mean ± SEM of four experiments. of the functional factor VIII. The multiple bands of the PCR product reflect partial deletions of raav ITRs occurring with dimerization of raav vectors. To confirm that appropriate splicing of the intronic sequence had occurred, we used a pair of primers residing in the two separate raav/f8 vectors, respectively, in RT-PCR amplification. The 5 primer was located at the 3 end of the 5 portion of F8 cdna and the 3 primer located at the 5 end of the 3 portion of F8 cdna. An expected 310-bp fragment not only confirmed head-totail heterodimerization of 5 and 3 AAV/F8 vectors, but also confirmed that appropriate splicing had occurred (Fig. 5). In Vivo Expression of BDD Factor VIII in Mice We injected raav/dlz17a3 and raav/dlz17c viral particles into the portal veins of 4- week-old male NOD/SCID mice (n = 8). Analysis of plasma samples every 4 weeks postinjection detected a gradual increase in factor VIII to a maximum of 8 ng/ml (4% of normal human factor VIII, where 100% = 200 ng/ml; Fig. 6). The mean of eight animals was ~ 4 ng/ml or 2%. The plasma concentration of factor VIII has persisted for over 4 months (Fig. 6). DISCUSSION Using a single vector carrying a small promoter element and the B- domain deleted factor VIII, we previously demonstrated the production of high levels of functional factor VIII in immunodeficient mice [4]. Though successful, the packaging-size constraint of AAV type 2 limits the use of a number of enhancer/promoter elements. Here we describe a novel approach to express human factor VIII using raav2 vectors. By expanding the packaging capacity of AAV through heterodimerization, it is possible to include larger transcription regulatory elements for factor VIII expression. Here we report the successful production of human factor VIII using this novel approach. A previous report using AAV-mediated gene transfer of F8 relied on the use of two separate vectors, each A B C FIG. 4. Molecular confirmation of raav dimerization. (A) Low-molecular-weight (Hirt) DNA was isolated from the 293 or HepG2 cells coinfected with both of the dual raav/f8 vectors. (B) PCR primers were designed to amplify the junction of heterodimers in the head-to-tail sequence. (C) The expected product extends from the 5 -vector intron to the 3 -vector F8 sequence. Head-to-tail heterodimerization of 5 and 3 AAV/F8 vector produces a 960-bp DNA fragment. Smaller DNA PCR products represent partial deletions of the AAV ITRs following head-to-tail heterodimerization. Plasmid pdlz17ac, containing the entire sequence of the expected dimer product lacking the AAV ITRs, served as a positive PCR control. Hirt DNA from 293 cells infected with raav GFP was used as negative control. 719

5 FIG. 5. Human F8 mrna splicing of dual raav/f8 vectors. Total cellular RNA was extracted from transduced cells for RT-PCR. A pair of primers was designed to amplify a 310-bp human F8 mrna fragment crossing F8 exon 12 and 13. A pair of -actin primers amplified a 250-bp fragment as an internal control for each sample. Each sample was run in duplicate with or without reverse transcription (RT). encoding the heavy- and light-chain protein [7]. Limitations of that approach include transduction of cells with only one vector, which produces one polypeptide that is unstable and prone to degradation [7]. In addition, it is unclear whether the stoichiometry of the two chains produced in that way can be regulated to prevent overproduction of one or the other protein; such a result may lead to functional interference and perhaps immune sensitization. The dual-vector strategy presented here still requires that both vectors enter the same cell, and only after dimerization occurs in the proper orientation will functional protein be produced. This strategy was designed to ensure that either vector alone is incapable of protein production that could potentially induce an immune response to the improperly processed and foreign protein. Dividing F8 at exon 12 further ensures its inability to produce heavy and light chains. The polyclonal antibody used for the ELISA assay detects both heavy and light chains. Our data would suggest that there is no detectable protein from each individual vector. The amount of functional factor VIII produced in vivo using the dual-vector approach is promising; factor VIII levels equivalent to 2 4% of the normal human factor VIII level was measured in this study. We and others have reported that 5% of liver cells are stably transduced using a single AAV vector [4,9,20]. Quantification of muscle and fibroblast transgene expression in vivo (mouse) using dual AAV vectors was measured at a rate of 4 7% of a single AAV vector [21]. We generated a mean of 40 ng/ml factor VIII in SCID mice by using a single AAV vector for factor VIII production [4]. Therefore, we would expect that the dual-vector approach would produce 2 4 ng/ml of factor VIII; this is in agreement with our results (Fig. 6). This level of factor VIII is sufficient to convert a severely affected individual with frequent bleeding episodes to a patient with mild disease. However, to be truly effective the system must be optimized. The molecular mechanism of dimerization is thought to be the result of recombination at the AAV ITR termini. Concatemer formation is preferred in a head-to-tail conformation following single raav vector infection [8,9,16 20]. These concatenates exist as long multimers [8,9,19], as circular duplex intermediates [10], or as both forms [8,9,19]. The degree of concatemerization and type of concatemer intermediates may vary depending on the cell type infected (H.C. and C.E.W, unpublished data). In addition, a variety of physical, chemical, and biological interventions could also potentially increase the AAV transgene expression [21 23]. Understanding the mechanism of concatemer formation will be important if this strategy is to be improved. In this study we describe the use of AAV type 2, the best characterized of all AAV serotypes [24]. The recent description of non-aav type 2 vectors with vastly superior gene transfer capacities in skeletal muscle [25,26] suggests that AAV serotypes will be required to improve the likelihood of dual vectors infecting the same cell so that dimerization can occur efficiently. Current testing of non-type 2 AAV serotypes in the liver suggests that FIG. 6. In vivo expression of B-domain deleted factor VIII. Four-week-old male NOD/SCID mice received injections into the portal vein of raav/dlz17a3 and raav/dlz17c raav/f8 vectors. Plasma was collected every four weeks and human factor VIII antigen measured by ELISA. The data represent the mean ± SEM (n = 8 animals). 720

6 ARTICLE these alternate serotypes are preferred (H.C. and C.E.W., unpublished data). Certainly for large transgene/cdnas the dual-vector system offers an alternative approach for AAV-mediated gene transfer and underscores the plasticity of the AAV vector system. MATERIALS AND METHODS Vector constructs. A series of dual-vector cassettes were constructed by splitting the F8 expression cassette within exon 12. The upstream or 5 vector carried enhancer/promoter elements linked to exons 1 12 of the B- domain deleted human F8 (BDD-F8) cdna and the first half of the human factor IX intron I [27]. The splice donor sequence was created using a PCRbased subcloning strategy [28] at the BamHI site in exon 12. Using the same PCR strategy, the downstream or 3 vector was composed of the second half of the human factor IX intron 1 sequence linked to exons 12, 13, of the BDD-F8 cdna and the bovine growth hormone (bgh) polyadenylation signal (pdlz17c). All PCR-generated sequence was verified by direct DNA sequencing. Vector pdlz17a carries the thyroid-binding globulin promoter and the human 1-microglobulin/bikunin enhancer, termed liver-specific-promoter (LSP; a gift from Charles Ill, Immune Response, Inc.) as described [29,30]. An intronic sequence (rabbit -globin-like intron, rbg) was inserted between LSP and F8 cdna in the pdlza to generate pdlz17a- I. The LSP promoter in vector pdlz17a2 was replaced with the immediate early promoter of cytomegalovirus (CMV), pdlz17a2, and the CBA promoter, pdlz17a3, that contains the CMV enhancer and chicken -actin promoter (a gift from Terry Flotte, University of Florida). All cassettes were subcloned into a ptr-uf5 backbone, which contains two intact AAV ITRs as described [8]. Cells and culture. Three cell types (293, HeLa, and HepG2) were cultured in Dulbecco s modified Eagle s medium (DMEM; Gibco/BRL) with 10% FBS (Gibco/BRL), with or without antibiotics (penicillin and streptomycin), at 37 C with 5% CO 2. The FBS was heated at 56 C for 30 minutes to inactivate bovine factor VIII activity determined by Coatest assay. raav production and purification. Recombinant AAV production employed a three-plasmid transfection scheme as described [31]. Briefly, subconfluent 293 cells were cotransfected with the raav transgene vector, AAV helper plasmid pxx2, and adenovirus helper plasmid pxx6 using calcium phosphate precipitation. Cells were harvested 48 hours post-transfection, lysed by three cycles of freeze thawing, and sonicated to release the raav particles. Following ammonium sulfate precipitation, the virus was purified and concentrated by affinity column chromatography [32]. Viral titer was performed by dot-blot hybridization and infectious-center assays [33]. Peak fractions were pooled, dialyzed against PBS, and stored at 20 C. In vitro expression of factor VIII. A total of or HepG2 cells were plated in each well of a six-well plate overnight. Cells were transfected with plasmid (3 g) using Effectene transfection kit (Qiagen, Germany), or infected with raav virus particles/cell plus adenovirus type 5 (MOI = 1) for 1 hour. The cell medium was harvested for analysis and replaced with fresh medium every 24 hours post infection. All the media and sera were heat-inactivated and free of endogenous factor VIII. Antigen and activity assay for human factor VIII. Human factor VIII antigen was detected by ELISA as described [4]. Monoclonal sheep antihuman factor VIII antibody (Affinity Biological, Inc.) was used as the capture antibody. Plasma or medium was applied to 96-well ELISA plates. Peroxidase-conjugated sheep anti-human factor VIII antibody (Affinity Biological, Inc.) was used as secondary antibody. The factor VIII levels were calculated according to the standard curve derived from serial dilution of the pooled normal human plasma (UCRP, Fisher). The lower limit of factor VIII detection using ELISA was 0.3 ng/ml. B-domain deleted human factor VIII activity was tested by Coatest (Chromgenix AB, Sweden), done following the manufacturer s instructions. Animal procedures. The NOD/SCID mice were maintained at the animal facilities at the University of North Carolina at Chapel Hill in accordance with the guidelines of the UNC institutional Animal Care and Use Committee. Following anesthesia, intraportal injections of the raav were carried out. Animal care and plasma collections were conducted as described [4]. DNA PCR. Using a described protocol [4], low-molecular-weight DNA (Hirt) was isolated from the 293 and HepG2 cells infected with raav/dlz17a3 and raav/dlz17c, each at 10 4 particles/cell at 48 hours postinfection: primer 1, 5 -AAGTTATGTAACGCGGAACTCCATATATGGGC-3 ; primer 2, 5 -AAAAG- CAAGCTTAAGAATTGACATAAAGAGTAGG-3 ; primer 3, 5 -TATGCCACCT- CATGCAAACAAACTGACAAC-3 ; primer 4, 5 -CCCCCGTGCCTTCCTTGAC- CCTGGAAGGTGCC-3. Four primers were designed to amplify junction fragments of the concatemers between the 5 and 3 raav/f8 vectors (Fig. 4B). The PCR conditions were: 95 C for 4 minutes followed by 30 cycles with 94 C for 1 minute, 60 C for 1 minute, 72 C for 4 minutes. RNA extraction and RT-PCR. Total cellular RNA was extracted from cultured cells for RT-PCR as described [4]. A pair of primers (sense, 5 -GAATCT- GTAGATCAAAGAGG-3 ; antisense, 5 -AAACTATCAAAAACATAGCC-3 ) were designed to amplify a 310-bp human F8 cdna fragment crossing exons 12 and 13. The cycle conditions consisted of 95 C for 2 minutes followed by 30 cycles of 95 C for 1 minute, 55 C for 1 minute, and 72 C for 1 minute. -Actin primers were used as an internal control for each sample. RECEIVED FOR PUBLICATION JANUARY 25; ACCEPTED MARCH 28, ACKNOWLEDGMENTS We thank R. Kole for critical suggestions, and T. Williams for technical assistance. H.J.C. is a recipient of the Career Development Award from National Hemophilia Foundation. REFERENCES 1. Monahan, P., and Samulski, R. (2000). Adeno-associated virus vectors for gene therapy: more pros than cons? Mol. Med. Today 6: Kaufman, R. (1999). Advances toward gene therapy for hemophilia at the millennium. Hum. Gene Ther. 10: Dong, J., Fan, P., and Frizzell, R. (1996). Quantitative analysis of the packaging capacity of recombinant adeno-associated virus. Hum. Gene Ther. 7: Chao, H., Mao, L., Bruce, A., and Walsh, C. (2000). Sustained expression of human Factor VIII in mice using a parvovirus-based vector. Blood 95: Pittman, D., et al. (1993). Biochemical, immunological, and in vivo functional characterization of B-domain-deleted Factor VIII. Blood 81: Kaufman, R., Pipe, S., Tagliavacca, L., Swaroop, M., and Moussalli, M. (1998). Biosynthesis, assembly and secretion of coagulation Factor VIII. Blood Coagul. Fibrinolysis 8: S3 S Burton, M., et al. (1999). Coexpression of Factor VIII heavy and light chain adeno-associated viral vectors produces biologically active protein. Proc. Natl. Acad. Sci. USA 96: Chao, H., et al. (1999). Persistent expression of canine Factor IX in hemophilia B canines. Gene Ther. 6: Miao, C., et al. (1998). The kinetics of raav integration in the liver. Nat. Genet. 19: Duan, D., Yan, Z., Yue, Y., and Engelhardt, J. (1999). Structural analysis of adeno-associated virus transduction circular intermediates. Virology 261: Sun, L., Li, J., and Xiao, X. (2000). Overcoming adeno-associated virus vector size limitation through viral DNA heterodimerization. Nat. Med. 6: Duan, D., Yue, Y., Yan, Z., and Engelhardt, J. (2000). A new dual-vector approach to enhance recombinant adeno-associated virus-mediated gene expression through intermolecular cis activation. Nat. Med. 6: Nakai, H., Storm, T., and Kay, M. (2000). Increasing the size of raav-mediated expression cassettes in vivo by intermolecular joining of two complementary vectors. Nat. Biotechnol. 18: Samulski, R. (2000). Expanding the AAV package. Nat. Biotechnol. 18: Snyder, R., et al. (1997). Persistent and therapeutic concentrations of human Factor IX in mice after hepatic gene transfer of recombinant AAV vectors. Nat. Genet. 16: Kotin, R., and Berns, K. (1989). Organization of adeno-associated virus DNA in latently infected Detroit 6 cells. Virology 170: Ni, T., Zhou, X., McCarty, D., Zolotukhin, I., and Muzyczka, N. (1994). In vitro replication of adeno-associated virus DNA. J. Virol. 68: Fisher, K., et al. (1996). Transduction with recombinant adeno-associated virus for gene therapy is limited by leading-strand synthesis. J. Virol. 70: Xiao, X., Li, J., and Samulski, R. (1996). Efficient long-term gene transfer into muscle tissue of immunocompetent mice by adeno-associated virus vector. J. Virol. 70: Xiao, W., et al. (1998). Adeno-associated virus as a vector for liver-directed gene ther- 721

7 apy. J. Virol. 72: Duan, D., Yue, Y., and Engelhardt, J. (2001). Expanding AAV packaging capacity with trans-splicing or overlapping vectors: a quantitative comparison. Mol. Ther. 4: Ferrari, F., Samulski, T., Shenk, T., and Samulski, R. (1996). Second-strand synthesis is a rate-limiting step for efficient transduction by recombinant adeno-associated virus vectors. J. Virol. 70: Duan, D., Yue, Y., Yan, Z., Yang, J., and Engelhardt, J. (2000). Endosomal processing limits gene transfer to polarized airway epithelia by adeno-associated virus. J. Clin. Invest. 105: Rabinowitz, J., and Samulski, R. (2000). Building a better vector: the manipulation of AAV virions. Virology 278: Chao, H., et al. (2000). Several log increase in therapeutic transgene delivery by distinct adeno-associated viral serotype vectors. Mol. Ther. 2: Chao, H., Monahan, P., Liu, Y., Samulski, R., and Walsh, C. (2001). Sustained and complete phenotype correction of hemophilia B mice following intramuscular injection of AAV1 serotype vectors. Mol. Ther. 4: Kurachi, S., Hitomi, Y., Furukawa, M., and Kurachi, K. (1995). Role of intron I in expression of the human Factor IX gene. J. Biol. Chem. 270: Jones, D., and Howard, B. (1990). A rapid method for site-specific mutagenesis and directional subcloning by using the polymerase chain reaction to generate recombinant circles. Biotechniques 8: Ill, C., et al. (1997). Optimization of the human Factor VIII complementary DNA expression plasmid for gene therapy of hemophilia A. Blood Coagul. Fibrinolysis 8: S23 S Wang, L., Takabe, K., Bidlingmaier, S., Ill, C., and Verma, I. (1999). Sustained correction of bleeding disorder in hemophilia B mice by gene therapy. Proc. Natl. Acad. Sci. USA 96: Xiao, X., Li, J., and Samulski, R. (1998). Production of high-titer recombinant adenoassociated virus vectors in the absence of helper adenovirus. J. Virol. 72: Zolotukhin, S., et al. (1997). Recombinant adeno-associated virus purification using novel methods improves infectious titer and yield. Gene Ther. 6: Snyder, R., Xiao, X., and Samulski, R. (1996). Production of recombinant adeno-associated viral vectors. In Current Protocols in Human Genetics (N. Dracopoli, et al., Eds.) pp John Wiley & Sons, Inc., New York. 722