Novel cis-acting replication element in the adeno-associated virus type 2 genome is involved in amplification of integrated rep-cap sequences

This study identifies a region of the adeno-associated virus type 2 (AAV-2) rep gene (nucleotides 190 to 540 of wild-type AAV-2) as a cis-acting Rep-dependent element able to promote the replication of transiently transfected plasmids. This viral element is also shown to be involved in the amplification of integrated sequences in the presence of adenovirus and Rep proteins.     

Evidence for packaging of rep-cap sequences into adeno-associated virus (AAV) type 2 capsids in the absence of inverted terminal repeats: a model for generation of rep-positive AAV particles

We previously reported that a 350-bp region of the adeno-associated virus (AAV) type 2 rep gene contains a cis-acting element responsible for the Rep-dependent replication of a transiently transfected rep-cap plasmid. In this study, we further report that replicated rep-cap sequences can be packaged into AAV capsids in the absence of the inverted terminal repeats.     

Nucleotide sequence of the inverted terminal repetition in adeno-associated virus DNA.

The inverted terminal repetition in adeno-associated virus type 2 DNA has been sequenced. The terminal repetition contain 145 nucleotides of which the first 125 nucleotides can self-base pair to form a T-shaped hairpin structure. Both restriction endonuclease analysis with SmaI and BglI and direct sequence analysis of the SmaI fragments provide evidence for two sequences in the region of the terminal repetition between nucleotides 44 and 81. The two sequences represent an inversion of the first 125 nucleotides of the terminal repetition. Based on these data a model for adeno-associated virus DNA replication is presented which agrees in detail with a general model for eucaryotic DNA replication originally proposed by Cavalier-Smith (T. Cavalier-Smith, Nature [London] 18:672--684, 1976).     

Inverted terminal repeat sequences are important for intermolecular recombination and circularization of adeno-associated virus genomes

The relatively small package capacity (less than 5 kb) of adeno-associated virus (AAV) vectors has been effectively doubled with the development of dual-vector heterodimerization approaches. However, the efficiency of such dual-vector systems is limited not only by the extent to which intermolecular recombination occurs between two independent vector genomes, but also by the directional bias required for successful transgene reconstitution following concatemerization. In the present study, we sought to evaluate the mechanisms by which inverted terminal repeat (ITR) sequences mediate intermolecular recombination of AAV genomes, with the goal of engineering more efficient vectors for dual-vector trans-splicing approaches. To this end, we generated a novel AAV hybrid-ITR vector characterized by an AAV-2 and an AAV-5 ITR at opposite ends of the viral genome. This hybrid genome was efficiently packaged into either AAV-2 or AAV-5 capsids to generate infectious virions. Hybrid AV2:5 ITR viruses had a significantly lower capacity to form circular intermediates in infected cells than homologous AV2:2 and AV5:5 ITR vectors despite their similar capacity to express an encoded enhanced green fluorescent protein (EGFP) transgene. To examine whether the divergent ITR sequences contained within hybrid AV2:5 ITR vectors could direct intermolecular recombination in a tail-to-head fashion, we generated two hybrid ITR trans-splicing vectors (AV5:2LacZdonor and AV2:5LacZacceptor). Each delivered one exon of a beta-galactosidase minigene flanked by donor or acceptor splice sequences. These hybrid trans-splicing vectors were compared to homologous AV5:5 and AV2:2 trans-splicing vector sets for their ability to reconstitute beta-galactosidase gene expression. Results from this comparison demonstrated that hybrid ITR dual-vector sets had a significantly enhanced trans-splicing efficiency (6- to 10-fold, depending on the capsid serotype) compared to homologous ITR vectors. Molecular studies of viral genome structures suggest that hybrid ITR vectors provide more efficient directional recombination due to an increased abundance of linear-form genomes. These studies provide direct evidence for the importance of ITR sequences in directing intermolecular and intramolecular homologous recombination of AAV genomes. The use of hybrid ITR AAV vector genomes provides new strategies to manipulate viral genome conversion products and to direct intermolecular recombination events required for efficient dual-AAV vector reconstitution of the transgene.     

DNA-binding activity of adeno-associated virus Rep is required for inverted terminal repeat-dependent complex formation with herpes simplex virus ICP8

Herpes simplex virus (HSV) helper functions for (AAV) replication comprise HSV ICP8 and helicase-primase UL5/UL52/UL8. Here we show that N-terminal amino acids of AAV Rep78 that contact the Rep-binding site within the AAV inverted terminal repeat (ITR) are required for ternary-complex formation with infected-cell protein 8 (ICP8) on AAV single-strand DNA (ssDNA) in vitro and for colocalization in nuclear replication domains in vivo. Our data suggest that HSV-dependent AAV replication is initiated by Rep contacting the AAV ITR and by cooperative binding of ICP8 on AAV ssDNA.     

Identification of nuclear proteins that specifically interact with adeno-associated virus type 2 inverted terminal repeat hairpin DNA.

A palindromic hairpin duplex containing the inverted terminal repeat sequence of adeno-associated virus type 2 (AAV) DNA was used as a substrate in gel retardation assays to detect putative proteins that specifically interact with the AAV hairpin DNA structures. Nuclear proteins were detected in extracts prepared from human KB cells coinfected with AAV and adenovirus type 2 that interacted with the hairpin duplex but not in nuclear extracts prepared from uninfected, AAV-infected, or adenovirus type 2-infected KB cells. The binding was specific for the hairpin duplex, since no binding occurred with a double-stranded DNA duplex with the identical nucleotide sequence. Furthermore, in competition experiments, the binding could be reduced with increasing concentrations of the hairpin duplex but not with the double-stranded duplex DNA with the identical nucleotide sequence. S1 nuclease assays revealed that the binding was sensitive to digestion with the enzyme, whereas the protein-bound hairpin duplex was resistant to digestion with S1 nuclease. The nucleotide sequence involved in the protein binding was localized within the inverted terminal repeat of the AAV genome by methylation interference assays. These nuclear proteins may be likely candidates for the pivotal enzyme nickase required for replication or resolution (or both) of single-stranded palindromic hairpin termini of the AAV genome.     

Arrangement of nucleotide sequences in adeno-associated virus DNA

There are two types of adeno-associated virus virions which contain complementary single-stranded DNA genomes of about 1.4 × 10⁶ daltons. The purified complementary single polynucleotide chains anneal to form duplex linear monomers, circular monomers, and linear dimers, in addition to other less well-defined structures, as identified by sedimentation and electron microscopy. All duplex species are formed by linear single polynucleotide chains of unit length, thus duplex circles and linear dimers are assumed to be held together by relatively short overlapping hydrogen-bonded regions. The initial linear monomers present after annealing the complementary single strands do not form duplex circles or oligomers when re-exposed to annealing conditions, but DNA which sediments as linear monomers after heating linear dimers at a temperature from 7 to 25 deg. C below the T_m of duplex AAV² DNA does re-form oligomers and circles when exposed to annealing conditions. Denaturation and reannealing of any duplex species leads to the formation of all forms, indicating that the over-all single strand composition of all species is equivalent. Disruption of duplex circles by limited exonucleolytic digestion using either 3′ or 5′ exonucleases leads to the conclusion that the overlap region may have either 3′ or 5′ termini and that the overlap region represents less than 6% of the length of the genome. Exonucleolytic digestion of linear monomers to the extent of 50% leaves polynucleotide chains which cannot reanneal after denaturation, thus AAV DNA is not randomly circularly permuted. Duplex linear monomers which do not form circles when exposed to annealing conditions do form duplex circles after 1% exonuclease III digestion. A model consistent with these data is one in which the linear single polynucleotide chains present in AAV virions consist of two or more permutations. All of these chains contain terminal repetitions and their start points occur within a limited region, representing < 6% of the length of the genome. According to this interpretation AAV DNA would exist within the virion as a linear single polynucleotide chain or is cleaved at a few specific sites during extraction.     

Sequence requirements for stable binding and function of Rep68 on the adeno-associated virus type 2 inverted terminal repeats.

Replication of the palindromic inverted terminal repeats (ITRs) of adeno-associated virus type 2 requires several functions of the viral nonstructural Rep proteins. These include binding to the ITR, nicking of the double-stranded replication intermediate at the terminal resolution site (trs), and then strand displacement and synthesis from the nick. This report demonstrates the ability of both recombinant fusion maltose-binding protein (MBP)-Rep68 delta produced in Escherichia coli and wild-type (wt) Rep68 to bind to a linear truncated form of the ITR, delta 57 ITR, with similar affinity as to the wt hairpin ITR. A dissociation constant for MBP-Rep68 delta of approximately 8 x 10(-10) M was determined for the wt ITR and delta 57 ITR probes. Truncation of delta 57 ITR to generate delta 28 ITR, which retains the GCTC repeat motif but not the trs, bound at least 10 times less efficiently than delta 57 ITR. Extension of delta 28 ITR with nonspecific sequence restored the ability of MBP-Rep68 delta to bind to delta 28 ITR. Thus, high-affinity binding would appear to require stabilization by flanking sequence as well as the intact GCTC repeat motif. Cleavage of the delta 57 ITR probe with DdeI, which truncates the flanking sequence and was previously shown to inhibit binding by Rep68, also inhibited the binding of MBP-Rep68 delta. The requirements for stable binding were further defined with a series of oligonucleotide probes which spanned the region protected by MBP-Rep78 in DNase I footprinting. The binding activity of either MBP-Rep68 delta or wt Rep68 to hairpin ITR or delta 57 ITR was indistinguishable. However, the binding activity of MBP-Rep68 delta to DNA does not appear to correlate with trs endonuclease activity. The nicking and covalent linkage of MBP-Rep68 delta to the nonhairpin delta 57 ITR was approximately 100-fold less efficient than its linkage to a hairpin-containing ITR. Therefore, although the hairpin portion of the ITR does not appear to play a role in recognition and stabilization of MBP-Rep68 delta binding, its presence does affect the trs cleavage activity of the protein.     

Isolation of an integrated provirus of Moloney murine leukemia virus with long terminal repeats in inverted orientation: integration utilizing two U3 sequences.

We have previously described the construction of a mutant of Moloney murine leukemia virus, in594-2, which carries a 2-base-pair insertion in the U5 region of the genome and is partially defective in forming the integrated proviral DNA. We have now recovered a cloned copy of an unusual provirus from rat cells infected with this mutant. The viral genome is flanked by long terminal repeats in inverted orientation, with U3 sequences joined to cellular DNA at both of the outer edges. In addition, the provirus is a recombinant, containing a segment of a VL30 element in inverted orientation in place of the Moloney murine leukemia virus env region. The recovery of this provirus indicates that two U3 regions can be used for viral integration and suggests that there may be no absolute requirement in the reaction for those U5 sequences outside the 13-base-pair inverted repeats.     

Factors that bind to adeno-associated virus terminal repeats.

We have identified and characterized a DNA-protein complex that forms with the adeno-associated virus (AAV) terminal repeats. The complex formed only if the terminal palindrome was in the covalently closed or hairpin configuration; little if any binding was detected with the open duplex form of the terminal repeat. This fact suggested that both secondary structure and primary sequence are essential elements of recognition. DNase I protection studies indicated that virtually all of the A-A' palindrome and significant portions of the B-B' and C-C' palindromes are protected. The postulated terminal resolution site of AAV also is protected. Restriction mapping of the sequences necessary for binding indicated that almost all of the terminal palindrome must be present for binding to occur. Hairpins which are similar in size and shape to the AAV termini did not exhibit competition for binding, and the complex formed only if AAV-infected extracts were used. Thus, the binding reaction is specific for AAV sequences. The viral-coded nonstructural proteins Rep78 and Rep68 comigrated with the DNA-protein complex on neutral acrylamide gels, suggesting that one or both of these proteins are components of the complex. The characteristics of the complex suggested that it has a role in AAV DNA replication.     

Cloning and integration of DNA fragments in human cells via the inverted terminal repeats of the adeno-associated virus 2 genome.

In current systems for molecular cloning of eukaryotic genes, bacterial cells are routinely utilized as intermediate hosts. We investigated the possibility of using a viral system for cloning DNA fragments independent of bacterial cell usage. In this report, we provide an alternative approach for molecular cloning of DNA fragments in eukaryotic cells by utilizing the inverted terminal repeats (ITRs) of the genome of a nonpathogenic human parvovirus, the adeno-associated virus 2 (AAV). We constructed a series of chimeric linear duplex DNA molecules, ranging in length from 1.8 to 7.2 kb, containing the cruciform structures of AAV-ITRs at both ends. These 'no-end' (NE) DNA structures, when transfected into adenovirus-infected human cells in the presence of AAV replication proteins (Rep), underwent DNA replication. Furthermore, in the presence of AAV capsid proteins (Cap), all replicated DNA molecules of less than 5.0 kb were packaged into mature, biologically active AAV progeny virions. When a chimeric NE DNA (NE-neo) containing a gene (neo) encoding resistance to neomycin was transfected into human cells, neoR clones could be readily isolated in the presence of G418 (Geneticin). Southern-blot analysis of genomic DNA of several independently isolated neoR clones suggested stable integration of the NE-neo DNA into the host chromosomal DNA. AAV-ITRs, therefore, offer an alternative system for molecular cloning, as well as packaging of DNA fragments in mammalian cells independent of bacterial cell usage.     

Characterization of inverted repeated sequences in Ascaris nuclear DNA

The inverted repeated sequences of the chromatin-eliminating nematode Ascaris lumbricoides var. suum have been examined by electron microscopy and by hydroxyapatite chromatography, both in the germ-line and in the somatic DNA. 38% of the inverted repeats of the germ-line DNA analysed in the electron microscope have a single-stranded loop, in comparison to about 50% of looped structures in the somatic DNA. The loops are on average 2.3 X 10(3) base pairs (bp) long. The rest of the foldback DNA consists of simple hairpins. The average length of looped and unlooped inverted repeats is of the order of 300-400 bp in the germ-line and in the somatic DNA. The content of S1-resistant foldback duplexes isolated by hydroxyapatite chromatography amounts to 1.3% in spermatids, with an average length of 350 bp, and to 1.1% in intestinal or larval cell nuclei, with a length of about 320 bp. We estimate by two different methods that there exist approximately 12500 inverted repeats per haploid germ-line genome and approximately 8000 in the haploid somatic genome. A statistical analysis of the data indicates that the great majority of the foldback sequences are randomly distributed in the Ascaris genome, with a spacing of about (40-80) X 10(3) bp, both in the germ-line and in the somatic DNA.     

Characterization of inverted repeated sequences in wheat nuclear DNA.

The properties of inverted repeated sequences in wheat nuclear DNA have been studied by HAP(1) chromatography, nuclease S1 digestion and electron microscopy. Inverted repeated sequences comprise 1.7% of wheat genome. The HAP studies show that the amount of "foldback HAP bound DNA" depends on DNA length. Inverted repeats appear to be clustered with an average intercluster distance of 25 kb. It is estimated that there are approximately 3 x 10(6) inverted repeats per haploid wheat genome. The sequences around inverted repeats involve all families of repetition frequencies. Inverted repeats are observed as hairpins in electron microscopy. 20% of hairpins are terminated by a single-stranded spacer ranging from 0.3 to 1.5 kb in length. Duplex regions of the inverted repeats range from 0.1 to 0.45 kb with number average values of 0.24 kb and 0.18 kb for unlooped and looped hairpin respectively. Thermal denaturations and nuclease S1 digestions have revealed a length of about 100 bases for duplex regions. The methods used to study inverted repeated sequences are compared and discussed.     

Intercistronic as well as terminal sequences are required for efficient amplification of brome mosaic virus RNA3.

The genome of brome mosaic virus (BMV) is divided among messenger polarity RNA1, RNA2, and RNA3 (3.2, 2.9, and 2.1 kilobases, respectively). cis-Acting sequences required for BMV RNA amplification were investigated with RNA3. By using expressible cDNA clones, deletions were constructed throughout RNA3 and tested in barley protoplasts coinoculated with RNA1 and RNA2. In contrast to requirements for 5'- and 3'-terminal noncoding sequences, either of the two RNA3 coding regions can be deleted individually and both can be simultaneously inactivated by N-terminal frameshift mutations without significantly interfering with amplification of RNA3 or production of its subgenomic mRNA. However, simultaneous major deletions in both coding regions greatly attenuate RNA3 accumulation. RNA3 levels can be largely restored by insertion of a heterologous, nonviral sequence in such mutants, suggesting that RNA3 requires physical separation of its terminal domains or a minimum overall size for normal replication or stability. Unexpectedly, deletions in a 150-base segment of the intercistronic noncoding region drastically reduce RNA3 accumulation. This segment contains a sequence element homologous to sequences found near the 5' ends of BMV RNA1 and RNA2 and in analogous positions in the three genomic RNAs of the related cucumber mosaic virus, suggesting a possible role in plus-strand synthesis.