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.     

Adeno-associated virus type 2 DNA replication in vivo: mutation analyses of the D sequence in viral inverted terminal repeats.

The adeno-associated virus type 2 (AAV) genome contains inverted terminal repeats (ITRs) of 145 nucleotides. The terminal 125 nucleotides of each ITR form palindromic hairpin (HP) structures that serve as primers for AAV DNA replication. These HP structures also play an important role in integration as well as rescue of the proviral genome from latently infected cells or from recombinant AAV plasmids. Each ITR also contains a stretch of 20 nucleotides, designated the D sequence, that is not involved in HP structure formation. We have recently shown that the D sequence plays a crucial role in high-efficiency rescue, selective replication, and encapsidation of the AAV genome and that a host cell protein, designated the D sequence-binding protein (D-BP), specifically interacts with this sequence (X.-S. Wang, S. Ponnazhagan, and A. Srivastava, J. Virol. 70:1668-1677, 1996). We have now performed mutational analyses of the D sequences to evaluate their precise role in viral DNA rescue, replication, and packaging. We report here that 10 nucleotides proximal to the HP structure in each of the D sequences are necessary and sufficient to mediate high-efficiency rescue, replication, and encapsidation of the viral genome in vivo. In in vitro studies, the same 10 nucleotides were found to be required for specific interaction with D-BP, but viral Rep protein-mediated cleavage at the functional terminal resolution site is independent of these sequences. These data suggest that AAV replication and terminal resolution functions can be uncoupled and that the lack of efficient replication of AAV DNA may not be a consequence of impaired resolution of the viral ITRs. These studies further illustrate that the D sequence-D-BP interaction plays an important role in the AAV life cycle and indicate that it may be possible to develop the next generation of AAV vectors capable of encapsidating larger pieces of DNA.     

A novel terminal resolution-like site in the adeno-associated virus type 2 genome.

The adeno-associated virus 2 (AAV) contains a single-stranded DNA genome of which the terminal 145 nucleotides are palindromic and form T-shaped hairpin structures. These inverted terminal repeats (ITRs) play an important role in AAV DNA replication and resolution, since each of the ITRs contains a terminal resolution site (trs) that is the target site for the AAV rep gene products (Rep). However, the Rep proteins also interact with the AAV DNA sequences that lie outside the ITRs, and the ITRs also play a crucial role in excision of the proviral genome from latently infected cells or from recombinant AAV plasmids. To distinguish between Rep-mediated excision of the viral genome during rescue from recombinant AAV plasmids and the Rep-mediated resolution of the ITRs during AAV DNA replication, we constructed recombinant AAV genomes that lacked either the left or the right ITR sequence and one of the Rep-binding sites (RBSs). No rescue and replication of the AAV genome occurred from these plasmids following transfection into adenovirus type 2-infected human KB cells, as expected. However, excision and abundant replication of the vector sequences was clearly detected from the plasmid that lacked the AAV left ITR, suggesting the existence of an additional putative excision site in the left end of the AAV genome. This site was precisely mapped to one of the AAV promoters at map unit 5 (AAV p5) that also contains an RBS. Furthermore, deletion of this RBS abolished the rescue and replication of the vector sequences. These studies suggest that the Rep-mediated cleavage at the RBS during viral DNA replication may, in part, account for the generation of the AAV defective interfering particles.     

Integration of the adeno-associated virus genome into cellular DNA in latently infected human Detroit 6 cells.

A clone of human cells (Detroit 6) latently infected by adeno-associated virus (AAV) has been characterized with regard to the status of the viral DNA. In both early (9 to 10) and late (118) passages of the clone, AAV-DNA was recombined with host DNA, at least in some cases as a head-to-tail tandem repeat, via the terminal sequences of the viral genome. However, it was not possible to distinguish between integration into chromosomal DNA and very large plasmids (< 20 x 10(6) molecular weight) which contain both viral and cellular DNA sequences. Although evidence for some modifications of the viral sequence was obtained, most of the integrated sequences appeared to be intact. In some cases sequences of undetermined origin separated adjacent copies of the viral genome. Free copies of the AAV genome were detectable in late passage cells, but not in early passage cells. The orientation of nucleotide sequences present in the free AAV DNA from late passage cells was indistinguishable from that of virion DNA. With the notable exception, the organization of the integrated AAV sequences as determined by restriction enzyme digestion remained constant with continued passage. Digestion with SmaI, which cleaves within the palindromic region of the terminal repetition in AAV DNA, produced reproducibly different patterns when early and late passage DNAs were compared. Several models for rescue of free copies of the genome from the integrated DNA are possible, all of which involve the terminal repetition.     

In vitro excision of adeno-associated virus DNA from recombinant plasmids: isolation of an enzyme fraction from HeLa cells that cleaves DNA at poly(G) sequences.

When circular recombinant plasmids containing adeno-associated virus (AAV) DNA sequences are transfected into human cells, the AAV provirus is rescued. Using these circular AAV plasmids as substrates, we isolated an enzyme fraction from HeLa cell nuclear extracts that excises intact AAV DNA in vitro from vector DNA and produces linear DNA products. The recognition signal for the enzyme is a polypurine-polypyrimidine sequence which is at least 9 residues long and rich in G.C base pairs. Such sequences are present in AAV recombinant plasmids as part of the first 15 base pairs of the AAV terminal repeat and in some cases as the result of cloning the AAV genome by G.C tailing. The isolated enzyme fraction does not have significant endonucleolytic activity on single-stranded or double-stranded DNA. Plasmid DNA that is transfected into tissue culture cells is cleaved in vivo to produce a pattern of DNA fragments similar to that seen with purified enzyme in vitro. The activity has been called endo R for rescue, and its behavior suggests that it may have a role in recombination of cellular chromosomes.     

Adeno-associated virus general transduction vectors: analysis of proviral structures.

We used two kinds of adeno-associated virus (AAV) vectors to transduce the neomycin resistance gene into human cells. The first of these (dl52-91) retains the AAV rep genes; the second (dl3-94) retains only the AAV terminal repeats and the AAV polyadenylation signal (428 base pairs). Both vectors could be packaged into AAV virions and produced proviral structures that were essentially the same. Thus, the AAV sequences that are required in cis for packaging (pac), integration (int), rescue (res), and replication (ori) of viral DNA are located within a 284-base-pair sequence that includes the terminal repeat. Most of the G418r cell lines (73%) contained proviruses which could be rescued (Res+) when the cells were superinfected with the appropriate helper viruses. Some produced high yields of viral DNA; other rescued at a 50-fold lower level. Most of the lines that were Res+ (79%) contained a tandem repeat of the AAV genome (2 to 20 copies) which was integrated randomly with respect to cellular DNA. Junctions between two consecutive AAV copies in a tandem array contained either one or two copies of the AAV terminal palindrome. Junctions between AAV and cellular sequences occurred predominantly at or within the AAV terminal repeat, but in some cases at internal AAV sequences. Two lines were seen that contained free episomal copies of AAV DNA. Res+ clones contained deleted proviruses or tandem repeats of a deleted genome. Occasionally, flanking cellular DNA was also amplified. There was no superinfection inhibition of AAV DNA integration. Our results suggest that AAV sequences are amplified by DNA replication either before or after integration and that the mechanism of replication is different from the one used during AAV lytic infections. In addition, we have described a new AAV general transduction vector, dl3-94, which provides the maximum amount of room for insertion of foreign DNA and integrates at a high frequency (80%).     

Characterization of Wild-Type Adeno-Associated Virus Type 2-Like Particles Generated during Recombinant Viral Vector Production and Strategies for Their Elimination

The pSub201-pAAV/Ad plasmid cotransfection system was developed to eliminate homologous recombination which leads to generation of the wild-type (wt) adeno-associated virus type 2 (AAV) during recombinant vector production. The extent of contamination with wt AAV has been documented to range between 0.01 and 10%. However, the precise mechanism of generation of the contaminating wt AAV remains unclear. To characterize the wt AAV genomes, recombinant viral stocks were used to infect human 293 cells in the presence of adenovirus. Southern blot analyses of viral replicative DNA intermediates revealed that the contaminating AAV genomes were not authentic wt but rather wt AAV-like sequences derived from recombination between (i) AAV inverted terminal repeats (ITRs) in the recombinant plasmid and (ii) AAV sequences in the helper plasmid. Replicative AAV DNA fragments, isolated following amplification through four successive rounds of amplification in adenovirus-infected 293 cells, were molecularly cloned and subjected to nucleotide sequencing to identify the recombinant junctions. Following sequence analyses of 31 different ends of AAV-like genomes derived from two different recombinant vector stocks, we observed that all recombination events involved 10 nucleotides in the AAV D sequence distal to viral hairpin structures. We have recently documented that the first 10 nucleotides in the D sequence proximal to the AAV hairpin structures are essential for successful replication and encapsidation of the viral genome (X.-S. Wang et al., J. Virol. 71:3077–3082, 1997), and it was noteworthy that in each recombinant junction sequenced, the same 10 nucleotides were retained. We also observed that adenovirus ITRs in the helper plasmid were involved in illegitimate recombination with AAV ITRs, deletions of which significantly reduced the extent of wt AAV-like particles. Furthermore, the combined use of recombinant AAV plasmids lacking the distal 10 nucleotides in the D sequence and helper plasmids lacking the adenovirus ITRs led to complete elimination of replication-competent wt AAV-like particles in recombinant vector stocks. These strategies should be useful in producing clinical-grade AAV vectors suitable for human gene therapy.     

Adeno-associated virus (AAV) type 5 Rep protein cleaves a unique terminal resolution site compared with other AAV serotypes

Adeno-associated virus (AAV) replication depends on two viral components for replication: the AAV nonstructural proteins (Rep) in trans, and inverted terminal repeat (ITR) sequences in cis. AAV type 5 (AAV5) is a distinct virus compared to the other cloned AAV serotypes. Whereas the Rep proteins and ITRs of other serotypes are interchangeable and can be used to produce recombinant viral particles of a different serotype, AAV5 Rep proteins cannot cross-complement in the packaging of a genome with an AAV2 ITR. In vitro replication assays indicated that the block occurs at the level of replication instead of at viral assembly. AAV2 and AAV5 Rep binding activities demonstrate similar affinities for either an AAV2 or AAV5 ITR; however, comparison of terminal resolution site (TRS) endonuclease activities showed a difference in specificity for the two DNA sequences. AAV2 Rep78 cleaved only a type 2 ITR DNA sequence, and AAV5 Rep78 cleaved only a type 5 probe efficiently. Mapping of the AAV5 ITR TRS identified a distinct cleavage site (AGTG TGGC) which is absent from the ITRs of other AAV serotypes. Comparison of the TRSs in the AAV2 ITR, the AAV5 ITR, and the AAV chromosome 19 integration locus identified some conserved nucleotides downstream of the cleavage site but little homology upstream.     

Partial purification of adeno-associated virus Rep78, Rep52, and Rep40 and their biochemical characterization.

We have used differential cell extraction and conventional chromatography to separate and partially purify the four adeno-associated virus (AAV) nonstructural proteins Rep78, Rep68, Rep52, and Rep40. In the cytoplasmic extracts Rep52 and Rep40 were present in greater abundance than Rep68 and Rep78, with Rep78 being the least abundant. In nuclear extracts the four Rep proteins were approximately equal in abundance. Regardless of the subcellular fraction examined, three of the Rep proteins (Rep78, Rep68, and Rep40) consisted of two protein species with slightly different mobilities during polyacrylamide gel electrophoresis. In contrast, Rep52 consisted of only one protein species. Both Rep78 and Rep68 were capable of binding efficiently to AAV terminal hairpin DNA substrates, but we could not detect site-specific DNA binding by Rep52 and Rep40. Like Rep68, Rep78 had both an ATP-dependent trs endonuclease and a DNA helicase activity. Both Rep78 and Rep68 cut the terminal AAV sequence at the same site (nucleotide 124). The binding, trs endonuclease, and DNA helicase activities comigrated during sucrose density gradient centrifugation with a mobility expected for a monomer of the protein, suggesting that the three biochemical activities were intrinsic properties of the larger Rep proteins. The chromatographic behavior and the DNA-binding properties of the four Rep proteins identified at least two domains within the rep coding region, an exposed hydrophobic domain within the C-terminal end (amino acids 578 to 621) and a region within the N terminus (amino acids 1 to 214) which was necessary for binding to the terminal repeat sequence. No site-specific nuclease activity was seen in the presence of nucleotide analogs ATP-gamma-S or AMP-PNP, suggesting that ATP hydrolysis was required for the endonuclease reaction. Furthermore, although ATP was the only cofactor which would support the trs endonuclease activity of Rep78, Rep68 nuclease activity was seen in the presence of several other nucleotide cofactors, including CTP, GTP, and UTP.     

Tandem repeats within the inverted terminal repetition of vaccinia virus DNA

A tandemly repeated sequence within the genome of vaccinia virus is cut to fragments of approximately 70 bp by Hinf I, Taq I or Mbo II. The 70 bp repetition was localized within the much larger (10,300 bp) inverted terminal repetition by restriction analysis of cloned DNA fragments and by hybridization of the purified 70 bp repeat to vaccinia virus DNA restriction fragments. The molar abundance of the 70 bp fragment corresponds to a 30 fold repetition at each end of the genome. The repeating restriction endonuclease sites were mapped by agarose gel electrophoresis of partial Hinf I digests of the terminally labeled cloned DNA fragment. The first of 13 repetitive Hinf I sites occurred approximately 150 bp from the end of the cloned DNA. After an intervening sequence of approximately 435 bp, a second series of 17 repetitive Hinf I sites occurred. The DNA between the two blocks of repetitions has a unique sequence containing single Dde I, Alu I and Sau 3A sites. Tandem repeats within the inverted terminal repetition could serve to accelerate self-annealing of single strands of DNA to form circular structures during replication.     

Structure of the human gene encoding the invariant gamma-chain of class II histocompatibility antigens

The primary structures of a cDNA and the genomic DNA of a gene selectively expressed in chronic lymphocytic leukemia were determined. A computer search of the nucleotide sequence data bank identified this gene as the invariant gamma-chain associated with class II histocompatibility antigens. The invariant gamma-chain genomic sequence spans about 11 kilobases, with eight exons and seven introns. Three of the introns contain members of the Alu repeat family. A putative cap site and promoter sequence were identified at the 5' end of the gene. One or two copies of the gene is present in each haploid genome, and no evidence for amplification or polymorphism was obtained.     

Comparison of the DNA sequence and secondary structure of the herpes simplex virus L/S junction and the adeno-associated virus terminal repeat

The defective parvovirus Adeno-associated virus (AAV) is absolutely dependent upon coinfection with either Adenovirus or Herpes Simplex Virus (HSV) for its multiplication. We have compared the terminal repeats of HSV-1F strain DNA with the terminal 200 nucleotides of AAV DNA. Our findings demonstrate similarities between portions of the HSV inverted repeats found at the L/S junction and the termini of AAV. By computer analysis we have determined potential secondary folding patterns for both genomes. The following points can be made about the a, b, and c repeats in HSV: (1) Regions b and c are complementary over a significant portion of their length. (2) The ends of a can fold back on themselves to form large secondary structures. Moreover, when the b and c homology is used to align the ends of a, the b/a and c/a junctions are within 1 base of each other. (3) The short direct repeats within a are essentially a large loop with little secondary structure. The potential implications of this structure are discussed and a model for HSV DNA replication is presented.     

Cellular recombination pathways and viral terminal repeat hairpin structures are sufficient for adeno-associated virus integration in vivo and in vitro.

The human parvovirus adeno-associated virus (AAV) is unique in its ability to target viral integration to a specific site on chromosome 19 (ch-19). Recombinant AAV (rAAV) vectors retain the ability to integrate but have apparently lost this ability to target. In this report, we characterize the terminal-repeat-mediated integration for wild-type (wt), rAAV, and in vitro systems to gain a better understanding of these differences. Cell lines latent for either wt or rAAV were characterized by a variety of techniques, including PCR, Southern hybridization, and fluorescence in situ hybridization analysis. More than 40 AAV-rAAV integration junctions were cloned, sequenced, and then subjected to comparison and analysis. In both immortalized and normal diploid human cells, wt AAV targeted integration to ch-19. Integrated provirus structures consisted of head-to-tail tandem arrays with the majority of the junction sequences involving the AAV inverted terminal repeats (ITRs). No complete viral ITRs were directly observed. In some examples, the AAV p5 promoter sequence was found to be fused at the virus-cell junction. Data from dot blot analysis of PCR products were consistent with the occurrence of inversions of genomic and/or viral DNA sequences at the wt integration site. Unlike wt provirus junctions, rAAV provirus junctions mapped to a subset of non-ch-19 sequences. Southern analysis supported the integration of proviruses from two independent cell lines at the same locus on ch-2. In addition, provirus terminal repeat sequences existed in both the flip and flop orientations, with microhomology evident at the junctions. In all cases with the exception of the ITRs, the vector integrated intact. rAAV junction sequence data were consistent with the occurrence of genomic rearrangement by deletion and/or rearrangement-translocation at the integration locus. Finally, junctions formed in an in vitro system between several AAV substrates and the ch-19 target site were isolated and characterized. Linear AAV substrates typically utilized the end of the virus DNA substrate as the point of integration, whereas products derived from AAV terminal repeat hairpin structures in the presence or absence of Rep protein resembled AAV-ch-19 junctions generated in vivo. These results describing wt AAV, rAAV, and in vitro integration junctions suggest that the viral integration event itself is mediated by terminal repeat hairpin structures via nonviral cellular recombination pathways, with specificity for ch-19 in vivo requiring additional viral components. These studies should have an important impact on the use of rAAV vectors in human gene therapy.     

Features of the adeno-associated virus origin involved in substrate recognition by the viral Rep protein.

We previously demonstrated that the adeno-associated virus (AAV) Rep68 and Rep78 proteins are able to nick the AAV origin of DNA replication at the terminal resolution site (trs) in an ATP-dependent manner. Using four types of modified or mutant substrates, we now have investigated the substrate requirements of Rep68 in the trs endonuclease reaction. In the first kind of substrate, portions of the hairpinned AAV terminal repeat were deleted. Only deletions that retained virtually all of the small internal palindromes of the AAV terminal repeat were active in the endonuclease reaction. This result confirmed previous genetic and biochemical evidence that the secondary structure of the terminal repeat was an important feature for substrate recognition. In the second type of substrate, the trs was moved eight bases further away from the end of the genome. The mutant was nicked at a 50-fold-lower frequency relative to a wild-type origin, and the nick occurred at the correct trs sequence despite its new position. This finding indicated that the endonuclease reaction required a specific sequence at the trs in addition to the correct secondary structure. It also suggested that the minimum trs recognition sequence extended three bases from the cut site in the 3' direction. The third type of substrate harbored mismatched base pairs at the trs. The mismatch substrates contained a wild-type sequence on the strand normally cut but an incorrect sequence on the complementary strand. All of the mismatch mutants were capable of being nicked in the presence of ATP. However, there was substantial variation in the level of activity, suggesting that the sequence on the opposite strand may also be recognized during nicking. Analysis of the mismatch mutants also suggested that a single-stranded trs was a viable substrate for the enzyme. This interpretation was confirmed by analysis of the fourth type of substrate tested, which contained a single-stranded trs. This substrate was also cleaved efficiently by the enzyme provided that the correct strand was present in the substrate. In addition, the single-stranded substrate no longer required ATP as a cofactor for nicking. Finally, all of the substrates with mutant trss bound the Rep protein as efficiently as the wild-type did. This finding indicated that the sequence at the cut site was not involved in recognition of the terminal repeat for specific binding by the enzyme. We concluded that substrate recognition by the AAV Rep protein involves at least two and possibly as many as four features of the AAV terminal repeat.(ABSTRACT TRUNCATED AT 400 WORDS)     

Rep-Mediated Nicking of the Adeno-Associated Virus Origin Requires Two Biochemical Activities, DNA Helicase Activity and Transesterification

The single-stranded adeno-associated virus (AAV) genome is flanked by terminal hairpinned origins of DNA replication (terminal repeats [TRs]) that are nicked at the terminal resolution site (trs) by the AAV Rep protein in an ATP-dependent, site-specific manner. Here we determine the minimal trs sequence necessary for Rep cleavage, 3'-CCGGT/TG-5', and show that this 7-base core sequence is required only on the nicked strand. We also identify a potential stem-loop structure at the trs. Interestingly, Rep nicking on a TR substrate that fixes this trs stem-loop in the extruded form no longer requires ATP. This suggests that ATP-dependent Rep helicase activity is necessary to unwind the duplex trs and extrude the stem-loop structure, prior to the ATP-independent Rep transesterification reaction. The extrusion of origin stem-loop structures prior to nicking appears to be a general mechanism shared by plant and animal viruses and bacterial plasmids. In the case of AAV, this mechanism of TR nicking would provide a possible regulatory function.     

Incompletely base-paired flip-flop terminal loops link the two DNA strands of the vaccinia virus genome into one uninterrupted polynucleotide chain

The nature of the ends of the vaccinia virus genome was determined by nucleotide sequencing. Our finding of terminal hairpins indicated that the linear double-stranded DNA molecule consists of a single continuous polynucleotide chain. The 104 nucleotide apex of the hairpin contains predominantly A and T residues and is incompletely base-paired. These loops exist in two forms, which when inverted with respect to each other are complementary in sequence. Both forms of the 104 nucleotide loop are present in nearly equimolar amounts at each end of the genome. A set of 13 tandem 70 bp repeats begins 87 bp from the proximal segment of the terminal loop, followed by a unique sequence of 325 bp, and then by a second set of 18 tandem 70 bp repeats. The sequence of the 70 bp repeats reveals a 13 bp internal redundancy. Self-priming and de novo start replication models, which involve a site-specific nick in one DNA strand proximal to the 104 nucleotide loop, account for the observed sequence inversions and incomplete base-pairing. Similar mechanisms may be involved in replication of the ends of the eucaryotic chromosome.     

The AAV origin binding protein Rep68 is an ATP-dependent site-specific endonuclease with DNA helicase activity

Genetic studies of adeno-associated virus (AAV) indicate that two AAV genes are required for viral DNA replication: the palindromic terminal repeat, which is the origin for DNA replication, and the rep gene, which codes for a family of at least four viral nonstructural proteins. To determine the biochemical function of the Rep proteins, we have purified the AAV Rep68 protein to apparent homogeneity. We find that it contains a site-specific and strand-specific endonuclease activity that specifically cuts the AAV origin at the terminal resolution site (TRS). The TRS endonuclease requires the presence of ATP for activity and becomes covalently attached to the 5' end at the cut site. In addition to the specific endonuclease activity, Rep68 also contains a DNA helicase activity. These results demonstrate that the large AAV Rep proteins have a direct role in AAV DNA replication; namely, they provide the activities required for the resolution of covalently joined AAV termini.     

An Algebraic Representation of RNA Secondary Structures

Abstract

This paper develops mathematical methods for describing and analyzing RNA secondary structures. It was motivated by the need to develop rigorous yet efficient methods to treat transitions from one secondary structure to another, which we propose here may occur as motions of loops within RNAs having appropriate sequences. In this approach a molecular sequence is described as a vector of the appropriate length. The concept of symmetries between nucleic acid sequences is developed, and the 48 possible different types of symmetries are described. Each secondary structure possible for a particular nucleotide sequence determines a symmetric, signed permutation matrix. The collection of all possible secondary structures is comprised of all matrices of this type whose left multiplication with the sequence vector leaves that vector unchanged. A transition between two secondary structures is given by the product of the two corresponding structure matrices. This formalism provides an efficient method for describing nucleic acid sequences that allows questions relating to secondary structures and transitions to be addressed using the powerful methods of abstract algebra. In particular, it facilitates the determination of possible secondary structures, including those containing pseudoknots. Although this paper concentrates on RNA structure, this formalism also can be applied to DNA     

Circular DNA of human immunodeficiency virus: analysis of circle junction nucleotide sequences.

During infection of cells by retroviruses, some of the nonintegrated viral DNA can be found as a circular form containing two tandem, directly repeated long terminal repeats. The nucleotide sequence at the point where the long terminal repeats join (the circle junction) can be used to deduce the terminal nucleotides of the linear form of the viral DNA. Comparison of the termini of linear viral DNA with sequences at the junctions between the integrated provirus and the host chromosome has revealed that for most retroviruses 2 bp are removed from each end of the linear viral DNA during integration. For human immunodeficiency virus type 1 (HIV-1), however, sequence considerations involving primer-binding sites had suggested that only 1 bp is removed during integration. We obtained the nucleotide sequences at the ends of HIV-1 DNA by using the polymerase chain reaction to amplify fragments corresponding to the HIV-1 circle junction. Of 17 clones containing amplified sequences, 10 had identical circle junctions that contained an additional 4 bp (GTAC) relative to the integrated provirus. This indicates that, as for other retroviruses, 2 bp are removed from each end of the linear HIV-1 viral DNA during integration. The remaining seven isolates contained insertions or deletions at the circle junction.