Adeno-associated Virus of a Single-polarity DNA Genome Is Capable of Transduction In Vivo

Abstract

The adeno-associated virus (AAV) is a promising vector for gene therapy. Further improvement of the virus for clinical application depends on better understanding of the molecular structure and fate of the vector genome. AAV vectors with wild-type inverted terminal repeats package either the plus- or the minus-strand DNA genomes with equal frequency. By creating a series of deletions within the, we have developed a genetic approach that can generate an AAV vector that packages its single-stranded DNA genome predominantly in a single polarity (99.4%). This novel reagent efficiently transduced muscle, brain and liver in whole animals. The transduction efficiencies were similar to those of the control mixed-polarity vectors. Our results showed that reannealing of plus- and minus-strand DNA was not required for AAV-mediated transduction in vivo, supporting the hypothesis that second-strand DNA synthesis is a primary pathway in converting the single-stranded AAV genome into double-stranded forms. The availability of the single-polarity AAV vector would aid further studies on the mechanism of AAV transduction as well as the application of AAV vector for gene replacement therapy. The adeno-associated virus (AAV) is a promising vector for gene therapy. Further improvement of the virus for clinical application depends on better understanding of the molecular structure and fate of the vector genome. AAV vectors with wild-type inverted terminal repeats package either the plus- or the minus-strand DNA genomes with equal frequency. By creating a series of deletions within the, we have developed a genetic approach that can generate an AAV vector that packages its single-stranded DNA genome predominantly in a single polarity (99.4%). This novel reagent efficiently transduced muscle, brain and liver in whole animals. The transduction efficiencies were similar to those of the control mixed-polarity vectors. Our results showed that reannealing of plus- and minus-strand DNA was not required for AAV-mediated transduction in vivo, supporting the hypothesis that second-strand DNA synthesis is a primary pathway in converting the single-stranded AAV genome into double-stranded forms. The availability of the single-polarity AAV vector would aid further studies on the mechanism of AAV transduction as well as the application of AAV vector for gene replacement therapy. IntroductionAdeno-associated virus (AAV) is a single-stranded DNA virus encapsidating either polarity of its genome at the same frequency.1Berns KI Molecular biology of the adeno-associated viruses.Curr Top Microbiol Immunol. 1974; 65: 1-20PubMed Google Scholar It has been demonstrated that both polarities of the viral genome were equally infectious.2Samulski RJ Chang LS Shenk T A recombinant plasmid from which an infectious adeno-associated virus genome can be excised in vitro and its use to study viral replication.J Virol. 1987; 61: 3096-3101Crossref PubMed Google Scholar However, because the study was conducted using a cultured transformed cell line which supported second-strand synthesis, it remained to be determined whether one polarity of the virus was sufficient or both polarities were required for AAV-mediated transgene expression in vivo, where second-strand synthesis was rate limiting. Recently published observations seemed to suggest that the above two different mechanisms were in action. First, we showed that a novel AAV vector containing a self-complementary genome expressed in mouse tissues at a faster onset than the conventional single-stranded AAV vectors did.3McCarty DM Monahan PE Samulski RJ Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis.Gene Ther. 2001; 8: 1248-1254Crossref PubMed Scopus (525) Google Scholar,4Wang Z Ma HI Li J Sun L Zhang J Xiao X Rapid and highly efficient transduction by double-stranded adeno-associated virus vectors, in vitro and in vivo.Gene Ther. 2003; 10: 2105-2111Crossref PubMed Scopus (322) Google Scholar At later time points, the convectional vector had a gradually increased level of gene expression. The difference in the onset of gene expression indicated that the single-stranded genome of the convectional AAV vector required conversion to the double-stranded form prior to transgene expression. The gradual rise of gene expression resulting from the conventional AAV vectors implied that either second-strand synthesis or single-stranded DNA annealing, or both, were taking place though at a slow rate. Srivastava and colleagues reported that the efficiency of AAV transduction correlated with the phosphorylation level of a cellular protein regulating the conversion of the single-stranded AAV genome to its double-stranded form (refs. 5Qing K Hansen J Weigel-Kelley KA Tan M Zhou S Srivastava A Adeno-associated virus type 2-mediated gene transfer: role of cellular FKBP52 protein in transgene expression.J Virol. 2001; 75: 8968-8976Crossref PubMed Scopus (107) Google Scholar,6Qing K Li W Zhong L Tan M Hansen J Weigel-Kelley KA et al.Adeno-associated virus type 2-mediated gene transfer: role of cellular T-cell protein tyrosine phosphatase in transgene expression in established cell lines in vitro and transgenic mice in vivo.J Virol. 2003; 77: 2741-2746Crossref PubMed Scopus (52) Google Scholar and accompanying paper7Zhong L Zhou X Li Y Qing K Xiao Samulski RJ et al.Single-polarity recombinant adeno-associated Virus 2 vector-mediated transgene expression In vitro and In Vivo: mechanism of transduction. Mol Ther. ePUP ahead of print, 2007Google Scholar). These findings suggest that second-strand synthesis was the mechanism responsible for AAV transgene expression in vivo. However, Kay and colleagues used a mixed population of two AAV vectors that contained minor differences labeled with various recognizable features.8Nakai H Storm TA Kay MA Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo.J Virol. 2000; 74: 9451-9463Crossref PubMed Scopus (156) Google Scholar After analyzing the proviral structures using methods including restriction digestion, they concluded that reannealing of both polarities of vector genome was responsible for AAV transduction in the mouse liver. However, they could not rule out the possibility of single-stranded vector DNA annealing during the DNA extraction process in the test tubes. In order to further characterize the molecular fate of the AAV genome in vivo, we took a genetic approach to generate AAV vectors that contained vector DNA genome predominantly of a single polarity. This novel reagent, unable to generate any significant reannealing event when used alone, was employed in this study to analyze the mechanism of AAV transduction in animals.ResultsGeneration of single-polarity AAV vectorsThe AAV packaging signal is contained within the terminal repeat and packaging of AAV DNA is unidirectional from the 3′-end to the 5′-end into pre-form empty capsids.1Berns KI Molecular biology of the adeno-associated viruses.Curr Top Microbiol Immunol. 1974; 65: 1-20PubMed Google Scholar,9King JA Dubielzig R Grimm D Kleinschmidt JA DNA helicase-mediated packaging of adeno-associated virus type 2 genomes into preformed capsids.EMBO J. 2001; 20: 3282-3291Crossref PubMed Scopus (199) Google Scholar Thus, if the packaging signal is missing from one end of a double-stranded AAV genome, only one polarity of the DNA will be packaged, resulting in the production of an AAV preparation containing only a single polarity of the genome instead of the conventional two equally mixed polarities. We hypothesized that the AAV packaging signal is located within the D element, and therefore by creating deletion mutants of the D element we should be able to generate AAV vectors with a single polarity. To test this hypothesis, we employed a series of deletion mutants based on a novel 165–base pair (bp) AAV terminal repeat consisting of a regular 145-bp AAV terminal repeat with an additional copy of the D element on the opposite side of the existing D element (Figure 1). Previously, we demonstrated that plasmids containing the 165-bp terminal repeat were resolved at the Holiday junction to form double-stranded AAV genomes.10Xiao X Xiao W Li J Samulski RJ A novel 165-base-pair terminal repeat sequence is the sole cis requirement for the adeno-associated virus life cycle.J Virol. 1997; 71: 941-948Crossref PubMed Google Scholar The series of deletion mutants included 8-, 12-, 15-, 18-, and 29-bp deletions into the terminal repeat from one end of the D element (X.X., Ph.D. thesis).11Xiao X Characterization of adeno-associated virus DNA replication and integration. PhD thesis. Univ. Pittsburgh, 1992Google Scholar This series of plasmids, named pDDΔ8-CMV-GFP, pDDΔ12-CMV-GFP, pDDΔ15-CMV-GFP, pDDΔ18-CMV-GFP(+), and pDDΔ29-CMV-GFP respectively, as well as the parental pDDwt-CMV-GFP plasmid, were generated and separately packaged into AAV viral particles. To determine the polarities of the vector genomes, viral DNA genomes were extracted and hybridized, respectively with one of the complementary oligonucleotides derived from the cytomegalovirus (CMV) promoter. The Probe CMV(+) would only hybridize to the plus-strand of the viral DNA, while the CMV(−) would only hybridize to the minus-strand of the viral DNA. As shown in Figure 2, AAV vectors with progressive deletion of the D element contained an increasing proportion of one polarity of the viral genome and a decreasing proportion of the other polarity. With a near complete deletion of the D element as in the pDDΔ18-CMV-GFP(+), 99.4% of the vector genome was of one polarity. Our result confirmed our hypothesis that the D element contained a major AAV packaging signal and that deletion of the D element from one end of a double-stranded AAV genome allowed the generation of an AAV preparation with a single polarity of the genome.Figure 2Determination of the polarity of adeno-associated virus (AAV) vector genome. AAV vector DNA was extracted and dot-blotted to two identical sets of membrane as described in the Methods. The two sets of membrane were then hybridized each with one of the two complementary oligonucleotides derived from the cytomegalovirus (CMV) promoter, Probe CMV(+) or Probe CMV(−), which detect either the plus-strand or the minus-strand of the AAV DNA genomes. The amount of each polarity of the vector genome was determined by comparison with double-stranded, CMV-promoter plasmid standards loaded on each set of the membrane. GFP, green fluorescent protein.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Transduction of mouse muscle by single-polarity AAV vectorsTo determine whether a single-polarity AAV vector is capable of expressing its transgene in muscle, we injected directly into the hindleg muscle of C57BL/6 mice with AAV-DDΔ18-CMV-GFP(+), AAV-DDΔ18-CMV-GFP(−) (containing the CMV-GFP cassette in a flipped orientation), a mix of both the above vectors in a 1:1 ratio, AAV-DDwt-CMV-GFP, or a control AAV-CMV-LacZ. Two weeks later, expression of the GFP transgene from the AAV vectors was imaged in whole mice. Representative images are shown in Figure 3. Regardless of the genome polarity, all four preparations of the AAV expressed the GFP transgene at comparable levels. The result demonstrated that AAV vectors with a single polarity were equally capable of transduction in mouse muscle compared to mixed-polarity AAV vectors, suggesting that annealing of plus- and minus-strands did not play an appreciable role in AAV-GFP gene expression.Figure 3Transduction of mouse muscle by single-polarity adeno-associated virus (AAV) vectors. The following five preparations of AAV vector at a dose of 1011 physical particles per animal were injected into the hindlegs of C57BL/6 mice: AAV-DDΔ18-CMV-GFP, AAV-DDΔ18-CMV-GFP(−), AAV-DDΔ18-CMV-GFP mixed with AAV-DDΔ18-CMV-GFP(−) in a 1:1 ratio, AAV-DDwt-CMV-GFP, and a control AAV-CMV-LacZ. Two weeks later, expression of the green fluorescent protein (GFP) transgene contained in the AAV vectors was imaged as described in the Methods. CMV, cytomegalovirus.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Transduction of rat brain by single-polarity AAV vectorsTo evaluate the expression of single-polarity AAV vectors in the brain, we injected directly into the striatum of Sprague–Dawley rats with AAV-DDΔ18-CMV-GFP(+), AAV-DDΔ18-CMV-GFP(−), or a mix of both in a 1:1 ratio. Transgene expression was determined 2 weeks after the injection. Representative images are shown in Figure 4. Judging from the number of GFP-positive neurons per field, the three preparations of AAV vectors resulted in similar levels of transduction. This result suggested that single-polarity AAV vectors were as efficient as the conventional AAV vectors in mediating gene transfer in the brain.Figure 4Transduction of mouse brain by single-polarity adeno-associated virus (AAV) vectors. The following three preparations of AAV vector at a dose of 1011 physical particles per animal were injected into the striatum of Sprague–Dawley rats: AAV-DDΔ18-CMV-GFP, AAV-DDΔ18-CMV-GFP(−), and AAV-DDΔ18-CMV-GFP mixed with AAV-DDΔ18-CMV-GFP(−) in a 1:1 ratio. Two weeks later, expression of the green fluorescent protein (GFP) transgene contained in the AAV vectors was examined as described in the Methods. CMV, cytomegalovirus.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Transduction of mouse liver by single-polarity AAV vectorsIt has been reported that the in the of AAV vectors was in transgene expression in the W J Adeno-associated virus as a vector for gene Virol. PubMed Google Scholar Thus, we the in the single-polarity vectors with the the expression of the vectors in the liver. The is a of the gene and is by The has been shown to efficiently in a of cell M L Samulski RJ expression of protein from the PubMed Scopus Google Scholar We then injected this into mice the with 1011 particles of and a mix of both in a 1:1 ratio, and a control AAV-CMV-LacZ. This amount of AAV was used because et showed that it was the of a dose that a of transgene expression in H Storm TA S S et number of a vector dose in recombinant adeno-associated liver Virol. PubMed Scopus Google Scholar weeks after the mouse were for GFP As shown in Figure all three preparations of the AAV vectors transduced with comparable at of the in the liver. This level of expression the reported of by single-stranded AAV vectors, and was to the use of the for the transgene and the for the GFP To evaluate the possibility that the transgene expression was to reannealing of the single-polarity AAV genomes with the polarity genomes, and to that the was GFP we the dose to 1011 particles per mouse and the for both GFP and weeks after injection. As shown in Figure at a vector both the and mixed-polarity vectors resulted in similar levels of GFP expression. The GFP green with the from the GFP that the green was to expression of the GFP This set of suggested that the transgene expression was to have resulted from reannealing of the single-polarity AAV genomes with the polarity genomes in the vector In other it indicated that the GFP expression was predominantly to conversion of the single-polarity AAV genome to its double-stranded of mouse liver by single-polarity adeno-associated virus (AAV) green fluorescent protein (GFP) vectors. The following four preparations of AAV vector at a dose of 1011 vector genome particles per animal were injected into C57BL/6 mice the mixed with in a 1:1 ratio, and a control AAV-CMV-LacZ. weeks later, expression of the GFP transgene from the AAV vectors was directly imaged as described in the Methods. vector 1011 were and as in that an for GFP was on to the GFP Large Image Figure ViewerDownload Hi-res image Download of AAV vectors with a single polarity allowed to directly the whether one polarity of the genome is sufficient or both polarities are required for AAV-mediated transgene expression in vivo. Our results in this study that AAV vectors of a single polarity are efficient in transduction of muscle, brain, and liver in vivo, compared to the mixed-polarity AAV vectors. These findings suggest that reannealing of the vector genome is not required for transgene and therefore not be the primary mechanism of AAV transduction in vivo. Our is different from that of a H Storm TA Kay MA Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo.J Virol. 2000; 74: 9451-9463Crossref PubMed Scopus (156) Google Scholar that the mechanism of AAV transduction was and reannealing of the complementary and single-stranded viral DNA to form the double-stranded H Storm TA Kay MA Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo.J Virol. 2000; 74: 9451-9463Crossref PubMed Scopus (156) Google Scholar As a mechanism that single-polarity AAV vectors are of efficient transduction in vivo, and that a of vectors of both polarities transduction. However, our results with the annealing hypothesis as a major mechanism of single-stranded to DNA As we rule out the possibility of reannealing of AAV genomes in as a minor et used with of viral genomes to the two different polarity W K Xiao W viral not single-stranded DNA is for recombinant adeno-associated virus Virol. PubMed Scopus Google Scholar the vectors, results to be with our However, are two major differences our two First, studies were conducted using in to primary in vivo, have a for converting the single-stranded AAV genome into the double-stranded the of in shown in Figure was than in in the liver transduction it was for reannealing to place to as a novel reagent for mechanisms of AAV the single-polarity AAV vectors to be for gene replacement therapy. et demonstrated that an which packages only one polarity of the viral genome had efficiencies of and gene by vectors based on the virus of Virol. 2003; 77: PubMed Scopus Google Scholar In this single-polarity AAV vectors packaged in various would a of gene vectors for and The of a plasmid that a novel AAV terminal repeat with two of the D element, has been X Xiao W Li J Samulski RJ A novel 165-base-pair terminal repeat sequence is the sole cis requirement for the adeno-associated virus life cycle.J Virol. 1997; 71: 941-948Crossref PubMed Google Scholar To generate deletions from one end of the novel terminal was with and then The for the deletion mutants have been X Characterization of adeno-associated virus DNA replication and integration. PhD thesis. Univ. Pittsburgh, 1992Google Scholar 8-, 12-, 15-, 18-, and 29-bp deletions into the terminal repeat from one end of the D These mutants were named and To vectors based on plasmids, a containing a CMV-GFP expression cassette excised from the was into the to the copy of the D Samulski RJ in the adeno-associated virus terminal repeat junction Virol. 2000; 74: PubMed Scopus Google Scholar with the end near the terminal repeat were as determined by and derived were in and named pDDΔ8-CMV-GFP, pDDΔ12-CMV-GFP, pDDΔ15-CMV-GFP, and pDDwt-CMV-GFP was by the CMV-GFP into the To generate was with and the were in with and then with an After with the terminal repeat was and used to the wild-type terminal which was by two in the of the in which the terminal repeat had the opposite as that in the was as determined by and This plasmid, was with and then with the containing a CMV-GFP expression cassette excised from the at the same and as those in the were as determined by and The resulting plasmid was in and named To generate and a containing the from was used to the in the and The was excised as a vector production and AAV vectors used in this study were and as described X Li J Samulski RJ of recombinant adeno-associated virus vectors in the of Virol. PubMed Google Scholar a of three plasmids consisting of an AAV vector plasmid, an AAV plasmid and an plasmid was into AAV vectors generated were with both an and a S M K et adeno-associated virus using novel methods infectious and Ther. PubMed Scopus Google Scholar The physical of each AAV vector preparation was then determined with a To determine the polarity of the AAV vector genome, two sets of identical membrane were hybridized separately with each one of the following two complementary oligonucleotides derived from the promoter, Probe and Probe The oligonucleotides were labeled prior to the using the 3′-end following the To the of the vector genome, a set of amount of DNA from to was included in each of the dot-blotted membrane. After the were to and the of each signal was and using the and the By to the DNA were determined for each polarity of the AAV vectors. the of a polarity vector could be using the following or polarity polarity polarity transduction in muscle and GFP C57BL/6 mice from were injected with 1011 particles of the following five preparations of AAV vectors in a of into the hindleg AAV-DDΔ18-CMV-GFP, AAV-DDΔ18-CMV-GFP(−), AAV-DDΔ18-CMV-GFP mixed with AAV-DDΔ18-CMV-GFP(−) in a 1:1 ratio, AAV-DDwt-CMV-GFP, and a control AAV-CMV-LacZ. preparation of AAV vector was to a of five mice. Two weeks later, the mice were with with for the and imaged for GFP expression using an from the The for the were set at for the and animal studies have been by the transduction in brain and GFP of the were Sprague–Dawley rats from The were in a and had to and Virus vector were as Xiao X Li J Samulski RJ and expression of gene transfer by an adeno-associated virus (AAV) PubMed Scopus Google Scholar were with and into a a and a of the following three AAV preparations into the striatum to to the of and The in AAV-DDΔ18-CMV-GFP AAV-DDΔ18-CMV-GFP(−) AAV-DDΔ18-CMV-GFP mixed with AAV-DDΔ18-CMV-GFP(−) in The was in place for 2 to from the after the virus vector the were with and with by in After in the were and in The were three in on and with a fluorescent GFP was using an transduction in liver and GFP to C57BL/6 mice from were injected the with AAV vectors at the preparation of the AAV vector was to a of five mice. weeks later, the were in and at for GFP of liver were using a DM The for the were set at and GFP the were and with a The were and imaged as of during PubMed Scopus Google Scholar IntroductionAdeno-associated virus (AAV) is a single-stranded DNA virus encapsidating either polarity of its genome at the same frequency.1Berns KI Molecular biology of the adeno-associated viruses.Curr Top Microbiol Immunol. 1974; 65: 1-20PubMed Google Scholar It has been demonstrated that both polarities of the viral genome were equally infectious.2Samulski RJ Chang LS Shenk T A recombinant plasmid from which an infectious adeno-associated virus genome can be excised in vitro and its use to study viral replication.J Virol. 1987; 61: 3096-3101Crossref PubMed Google Scholar However, because the study was conducted using a cultured transformed cell line which supported second-strand synthesis, it remained to be determined whether one polarity of the virus was sufficient or both polarities were required for AAV-mediated transgene expression in vivo, where second-strand synthesis was rate limiting. Recently published observations seemed to suggest that the above two different mechanisms were in action. First, we showed that a novel AAV vector containing a self-complementary genome expressed in mouse tissues at a faster onset than the conventional single-stranded AAV vectors did.3McCarty DM Monahan PE Samulski RJ Self-complementary recombinant adeno-associated virus (scAAV) vectors promote efficient transduction independently of DNA synthesis.Gene Ther. 2001; 8: 1248-1254Crossref PubMed Scopus (525) Google Scholar,4Wang Z Ma HI Li J Sun L Zhang J Xiao X Rapid and highly efficient transduction by double-stranded adeno-associated virus vectors, in vitro and in vivo.Gene Ther. 2003; 10: 2105-2111Crossref PubMed Scopus (322) Google Scholar At later time points, the convectional vector had a gradually increased level of gene expression. The difference in the onset of gene expression indicated that the single-stranded genome of the convectional AAV vector required conversion to the double-stranded form prior to transgene expression. The gradual rise of gene expression resulting from the conventional AAV vectors implied that either second-strand synthesis or single-stranded DNA annealing, or both, were taking place though at a slow rate. Srivastava and colleagues reported that the efficiency of AAV transduction correlated with the phosphorylation level of a cellular protein regulating the conversion of the single-stranded AAV genome to its double-stranded form (refs. 5Qing K Hansen J Weigel-Kelley KA Tan M Zhou S Srivastava A Adeno-associated virus type 2-mediated gene transfer: role of cellular FKBP52 protein in transgene expression.J Virol. 2001; 75: 8968-8976Crossref PubMed Scopus (107) Google Scholar,6Qing K Li W Zhong L Tan M Hansen J Weigel-Kelley KA et al.Adeno-associated virus type 2-mediated gene transfer: role of cellular T-cell protein tyrosine phosphatase in transgene expression in established cell lines in vitro and transgenic mice in vivo.J Virol. 2003; 77: 2741-2746Crossref PubMed Scopus (52) Google Scholar and accompanying paper7Zhong L Zhou X Li Y Qing K Xiao Samulski RJ et al.Single-polarity recombinant adeno-associated Virus 2 vector-mediated transgene expression In vitro and In Vivo: mechanism of transduction. Mol Ther. ePUP ahead of print, 2007Google Scholar). These findings suggest that second-strand synthesis was the mechanism responsible for AAV transgene expression in vivo. However, Kay and colleagues used a mixed population of two AAV vectors that contained minor differences labeled with various recognizable features.8Nakai H Storm TA Kay MA Recruitment of single-stranded recombinant adeno-associated virus vector genomes and intermolecular recombination are responsible for stable transduction of liver in vivo.J Virol. 2000; 74: 9451-9463Crossref PubMed Scopus (156) Google Scholar After analyzing the proviral structures using methods including restriction digestion, they concluded that reannealing of both polarities of vector genome was responsible for AAV transduction in the mouse liver. However, they could not rule out the possibility of single-stranded vector DNA annealing during the DNA extraction process in the test tubes. In order to further characterize the molecular fate of the AAV genome in vivo, we took a genetic approach to generate AAV vectors that contained vector DNA genome predominantly of a single polarity. This novel reagent, unable to generate any significant reannealing event when used alone, was employed in this study to analyze the mechanism of AAV transduction in animals.