Nucleotide sequences at the termini of phi 29 DNA.

The nucleotide sequences of the first 422 base pairs from the left-hand end and the first 274 base pairs from the right-hand end of phi 29 DNA were determined by using the chemical degradation method of Maxam and Gilbert. The data indicate that phi 29 DNA has inverted terminal repetitions that are six base pairs long 5' (-A-A-A-G-T-A-). No perfectly self-complementary sequence exists within the terminal regions of phi 29 DNA, suggesting that DNA replication via a self-priming mechanism is improbable. The putative early promoter sequences were found in both ends of the phi 29 DNA. The results of the sequence determination are discussed in relation ship to models proposed for the mechanism of replication of linear DNA molecules.     

All gene-sized DNA molecules in four species of hypotrichs have the same terminal sequence and an unusual 3'' terminus.

In hypotrichous ciliates, all of the macronuclear DNA is in the form of low molecular weight molecules with an average size of approximately 2200 base pairs. Total macronuclear DNA from four hypotrichs has been shown to have inverted terminal repeats by direct sequence analysis. In Oxytricha nova, Oxytricha sp., and Stylonychia pustulata, this terminal sequence may be written as 5'-C4A4C4A4C4 ... 3'-G4T4G4T4G4T4G4T4G4 ... In Euplotes aediculatus, the sequences is similar but differs in the lengths of the duplex region (28 base pairs) and of the putative 3' extension (14 base pairs). Also in Euplotes, a second common sequence of 5 base pairs (A-A-C-T-T-T-T-G-A-A) occurs internal to the terminal repeat and a 17-base-pair heterogeneous region: 5'-C4A4C4A4C4A4C4(X)17T-T-G-A-A ... 3'-G2T4G4T4G4T4G4T4G4T4G4(X)17A-A-C-T-T ... The length of the terminal repeat sequence for O. nova was confirmed in cloned macronuclear DNA molecules.     

Inverted terminal repeats are added to genes during macronuclear development in Oxytricha nova.

In the hypotrichous ciliate Oxytricha nova all of the macronuclear DNA is in the form of low molecular weight, gene-sized molecules with an average size of 2,200 base pairs. These molecules are produced during macronuclear development by excision from micronuclear chromosomes. All, or nearly all, of the small macronuclear DNA molecules possess an inverted terminal repeat sequence consisting of 5' C4A4 3' repeats. The hypothesis that this terminal sequence serves as a recognition signal for excision of gene-sized molecules from chromosomes has been tested. A sequence containing the C4A4 repeat has been isolated and used to screen clones of micronuclear DNA for the presence of the repeat sequence. The results show that the intact repeated C4A4 sequence is not present at the ends of macronuclear sequences as they exist in the micronuclear chromosomes. Therefore, the entire terminal repeat is not a recognition sequence for gene excision but must be added to the ends of gene-sized molecules during or after the excision process.     

Terminal region recognition factor 1, a DNA-binding protein recognizing the inverted terminal repeats of the pGKl linear DNA plasmids.

The yeast linear DNA plasmids pGKl1 and pGKl2 contain inverted terminal repeats (ITRs) and terminal proteins covalently bound to the 5' termini of each plasmid. The presence of these features suggests a protein-primed mechanism of DNA replication, similar to that exemplified by mammalian adenovirus and phi 29 phage of Bacillus subtilis. In this paper, we report the identification of an activity in cytoplasmic extracts of yeast harboring the pGKl plasmids that recognizes the termini of both pGKl1 and pGKl2. We call this activity TRF1, for terminal region recognition factor 1. Deletion analyses and DNase I protection experiments demonstrate that the activity recognizes base pairs 107-183 within the ITR of pGKl1, and base pairs 126-179 within the ITR of pGKl2. The presence of T-tracts within these two regions, but otherwise dissimilar nucleotide sequences, suggests that TRF1 recognizes a common structural feature within the ITRs of the two plasmids. TRF1 has been partially purified from yeast cytoplasmic extracts and Southwestern analysis indicates that the apparent molecular mass of the protein is 16 kDa. By expressing three open reading frames from pGKl2 in Escherichia coli, we found that open reading frame 10 (ORF10) of pGKl2 encodes TRF1. The sequence of the ORF10 gene product indicates that TRF1 is a highly basic protein of small molecular mass. Comparison of TRF1 with other DNA-binding proteins known to recognize the terminal regions of linear DNAs, such as NFI and NFIII involved in adenovirus DNA replication, and phi 29 p6, involved in phi 29 DNA replication, indicates that TRF1 has a different mode of binding.     

Visualization of an inverted terminal repetition in vaccinia virus DNA.

An inverted terminal repetition was observed in DNA molecules extracted from vaccinia virus. The repeated sequence was visualized by (i) nicking the hairpin loops present of the ends of vaccinia virus DNA, (ii) separating the strands of DNA by alkali denaturation, (iii) allowing the single strands to self-anneal, and (iv) examining the DNA with an electron microscope. Single-stranded circular molecules, each of which contained a duplex projection (3.54 +/- 0.12 micron) representing the terminal repetition, readily formed. Similar size projections were also seen in heteroduplex structures formed by crosshybridization of the separated strands of the two terminal HindIII restriction fragments. Based on contour length measurements and the electrophoretic mobility of the isolated inverted terminal repetition, a molecular weight of approximately 6.9 X 10(6), equivalent to about 10,500 nucleotide base pairs, was estimated. Evidence was obtained from DNA-RNA hybridization studies that the terminal repetition is transcribed.     

Isolation and characterization of a highly repetitious inverted terminal repeat sequence from Oxytricha macronuclear DNA.

The low-complexity "gene-sized" linear DNA duplexes of the Oxytricha macronucleus sport short inverted terminal repeats; thus, each single strand is capable of forming a circle held together by a duplex "neck" [Wesley, R. D. (1975) Porc. Natl. Acad. Sci. USA 72, 678--682]. We have isolated necks from total, circularized, single-stranded macronuclear DNA by treatment with nuclease S1. Necks represent at least 2.2% of the total DNA, are homogeneous in size (23 base pairs), melt at 55 degrees in 0.18 M Na+, and reassociate extremely rapidly at 22 degrees (Cot1/2 = 1.1 X 10(-5) mol-liter-1.sec) to form hybrid necks of the same thermal stability. From these and other results, we conclude that all necks on all the many thousands of different single-stranded circles are the same. The neck sequence is therefore highly repetitious--found in multiple copies (as inverted terminal repeats at flush duplex ends and probably also internally) on each natural "gene-sized" macronuclear DNA molecule--implying the possible participation of this sequence both in the general vegetative metabolism of macronuclear DNA and in the pre-vegetative process whereby macronuclear DNA is excised from the total Oxytricha genome.     

Nucleotide sequence of the S-2 mitochondrial DNA from the S cytoplasm of maize

Mitochondria from the S male-sterile cytoplasm (cms-S) of maize contain two plasmid-like DNAs, S-1 and S-2, that appear to be prominently involved with the cytoplasmic male sterility trait. The complete nucleotide sequence of the S-2 DNA molecule was determined by the chain termination method. The linear S-2 DNA molecule contains 5,452 base pairs and is terminated by exact 208-base-pair inverted repetitions. Two large open reading frames were identified in the S-2 DNA, suggesting the possibility of protein-encoding genes. The nucleotide sequence of the S-2 termini are discussed with regard to models proposed for the replication of linear DNA molecules.     

Molecular cloning and partial characterization of unintegrated linear DNA from gibbon ape leukemia virus.

We have cloned the complete genome of an oncogenic primate retrovirus, the San Francisco isolate of gibbon ape leukemia virus, in a lambda phage vector. DNA sequence analysis and restriction endonuclease mapping of the inserted linear provirus demonstrated 9-base pair inverted repeats at its ends, flanking direct terminal repeats 470 base pairs in length. The (-) strong stop region of this DNA showed surprisingly low sequence homology to that of another gibbon ape leukemia virus isolate from an animal with similar disease. Analysis of the clone also revealed the terminal phosphate configuration of the linear provirus. The recombinant phage is suitable for direct use as a hybridization probe to detect homologous retroviral sequences in human cell lines.     

Structure of a cloned circular Moloney murine leukemia virus DNA molecule containing an inverted segment: implications for retrovirus integration.

Closed circular Moloney murine leukemia virus (M-MuLV) DNA was prepared from recently infected cells and cloned in a lambda vector. Four classes of cloned M-MuLV inserts were found: Class I, full length 8.8-kilobase (kb) inserts with two tandem long terminal repeats (LTRs) of 600 base pairs; class 2, 8.2-kb inserts with a single copy of a LTR; class 3, M-MuLV DNA inserts with various portions deleted; and class 4, an 8.8-kb insert with an internal sequence inversion. Determination of nucleotide sequence at the junction between the two LTRs from a class 1 insert suggested that circularization occurred by blunt-end ligation of an 8.8-kb linear DNA. The class 4 molecule had an inversion that was flanked by inverted LTRs, each of which had lost two terminal base pairs at the inversion end points. Also, four base pairs that were present only once in standard M-MuLV DNA were duplicated at either end of the inversion. This molecule was interpreted as resulting from an integrative inversion in which M-MuLV DNA has integrated into itself. Its analysis thus provided explicit information concerning the mechanism by which retrovirus DNA integrates into host cell DNA. Models of retrovirus integration based on bacterial DNA transposition mechanisms are proposed.     

Initiation of adenovirus DNA replication in vitro requires a specific DNA sequence.

Initiation of adenovirus DNA replication in vitro occurs on a linearized plasmid DNA containing 3,327 base pairs of the adenovirus terminal sequence. Various deletions have been constructed in the plasmid DNA and their template activities examined. Deletions from an internal restriction enzyme cleavage site that retain only 20 base pairs or more of the adenovirus terminal sequence support initiation and limited chain elongation, whereas deletions that leave 14 base pairs or less of the terminal sequence do not. On the other hand, all deletions extending from the very terminus of the adenovirus DNA destroy the template activity. The terminal 20 base pairs of adenovirus DNA contain a sequence A-T-A-A-T-A-T-A-C-C, which is perfectly conserved in the DNAs from different serotypes of human adenovirus. Base changes within the conserved sequence greatly reduce the template activity. These results suggest that the terminal 20 base pairs constitute a functional origin for the initiation of adenovirus DNA replication in vitro.     

Inverted repeats in chloroplast DNA from higher plants

The circular chloroplast DNAs from spinach, lettuce, and corn plants have been examined by electron microscopy and shown to contain a large sequence repeated one time in reverse polarity. The inverted sequence in spinach and lettuce chloroplast DNA has been found to be 24,400 base pairs long. The inverted sequence in the corn chloroplast DNA is 22,500 base pairs long. Denaturation mapping studies have shown that the structure of the inverted sequence is highly conserved in these three plants. Pea chloroplast DNA does not contain an inverted repeat. All of the circular dimers of pea chloroplast DNA are found to be in a head-to-tail confirmation. Circular dimers of spinach and lettuce were also found to have head-to-tail conformation. However, approximately 70-80% of the circular dimers in preparations of lettuce and spinach chloroplast DNA were found to be in a head-to-head conformation. We propose that the head-to-head circular dimers are formed by a recombination event between two circular monomers in the inverted sequence.     

Elaboration of telomeres in yeast: recognition and modification of termini from Oxytricha macronuclear DNA.

The termini of macronuclear DNA molecules from the protozoan Oxytricha fallax share a common sequence and structure, both of which differ markedly from those deduced for yeast telomeres. Despite these differences, terminal restriction fragments from O. fallax macronuclear DNA can support telomere formation in yeasts. Two linear plasmids (LYX-1 and LYX-2) constructed by ligating BamHI-digested total Oxytricha macronuclear DNA to a yeast vector were analyzed. One end of LYX-1 and both ends of LYX-2 are derived from the Oxytricha DNA that encodes rRNA (rDNA) whereas the other end of LYX-1 is from an Oxytricha fragment other than rDNA. After propagation in yeast, both ends of LYX-1 and LYX-2 retain the C4A4 repeat characteristic of the O. fallax terminal sequence. In addition, both ends of both plasmids acquire 300-1000 base pairs of DNA containing the sequence (C-A)n, a sequence found near the termini of yeast chromosomes. Thus, at least two different Oxytricha termini display distinctive properties in yeast cells in that linear plasmids containing them are not degraded nor are they integrated into chromosomal DNA. These Oxytricha termini may act directly as telomeres in yeast; alternatively, the Oxytricha DNA may serve as a signal that results in the elaboration of a yeast telomere on the ciliate DNA.     

Isolation of point mutations in bacteriophage Mu attachment regions cloned in a lambda::mini-Mu phage.

Twenty-one derivatives of a lambda::mini-Mu phage containing point mutations in the Mu attachment regions were isolated after mutD mutagenesis and selection for relief from Mu-specific replicative interference of lambda growth. DNA sequence analysis revealed that the single left-end mutant had suffered a T----C transition at position 1 of the Mu sequence, while the remaining 20 right-end mutants contained single base-pair insertions or deletions within the terminal 19 base pairs. A genetic assay showed that the right-end mutations revealed by sequencing were necessary for relief of the replicative inhibition of lambda growth. The properties of these mutants suggest that the terminal 2-base-pair and subterminal 8-base-pair inverted repeats are important for Mu-specific replicative transposition.     

Adenovirus-2 DNA Contains an Inverted Terminal Repetition

Denaturation and renaturation of the adenovirus-2 chromosome (a duplex rod) generates single-stranded circles of unit length. These circles can be opened into linear DNA molecules by digestion with exonuclease III, indicating that hydrogen bonding between the two ends of an adenovirus strand is responsible for maintaining the rod in a circular state.The formation of adenovirus single-stranded circles, and their sensitivity to exonuclease III, indicate that the mature adenovirus-2 DNA molecule contains an inverted terminal repetition. That is, the base sequence at one end of the molecule is inverted and appears again at the other end of the molecule. This is the first example of such a structure, and its function is unknown.     

Template requirements for the initiation of adenovirus DNA replication.

The first step in the replication of the adenovirus genome is the covalent attachment of the 5'-terminal nucleotide, dCMP, to the virus-encoded terminal protein precursor (pTP). This reaction can be observed in vitro and has been previously shown to be dependent upon either viral DNA or linearized plasmid DNA containing viral terminal sequences. Plasmids containing deletions or point mutations within the viral terminal sequence were constructed by site-directed mutagenesis. In the case of linear double-stranded templates, pTP-dCMP formation required sequences located within the first 18 base pairs of the viral genome. This sequence contains a segment of 10 base pairs that is conserved in all human adenovirus serotypes. Point mutations within the conserved segment greatly reduced the efficiency of initiation, while a point mutation at a nonconserved position within the first 18 base pairs had little effect. Single-stranded DNAs can also support pTP-dCMP formation in vitro. In contrast to the results obtained with duplex templates, experiments with a variety of single-stranded templates, including phage M13-adenovirus recombinants, denatured plasmids, and synthetic oligodeoxynucleotides, failed to reveal any requirements for specific nucleotide sequences. With single-stranded templates containing no dG residues, the specific deoxynucleoside triphosphate requirements of the initiation reaction were altered.     

Molecuar model of the DNA interaction site for the cyclic AMP receptor protein.

A topological model of the DNA binding site for the cyclic AMP receptor protein (CRP) is presented. A consensus sequence drawn from the known CRP binding sites has several symmetrical subregions that are spatially resolved onto different faces of the DNA helix. Consideration of available biochemical and genetic data suggest one particular choice among the possible symmetrical arrangements. In this case, the sequence of its helical form presents nearly the same pattern of exposed base pairs on two faces of the helix. These two faces are separated by a helix angle of 72 degrees; the similar sequences that are exposed in the grooves occur in opposite orientations on the two faces. We propose that this inverted symmetry arrangement provides each of the identical subunits of the CRP with a similar recognition region within the overall site. In gal and ara, the site appears to accommodate a single molecule of the CRP; in lac, the site repeats the symmetrical arrangement and should accommodate two molecules of the CRP.     

Inverted repetitious sequences in the macronuclear DNA of hypotrichous ciliates.

The low-molecular-weight macronuclear DNA of the hypotrichous ciliates Oxytricha, Euplotes, and Paraurostyla contains inverted repetitious sequences. Up to 89% of the denatured macronuclear DNA molecules form single-stranded circles due to intramolecular renaturation of complementary sequences at or near the ends of the same polynucleotide chain. Other ciliated protozoans, such as Tetrahymena, with high-molecular-weight macronuclear DNA and an alternative mode of macronuclear development, appear to lack these selfcomplementary sequences. The denatured macronuclear molecules of hypotrichs are held in the circular conformation by a hydrogen-bonded duplex region, which is probably less than 50 base pairs in length, since the duplex regions are not visible by electron microscopy and since the circles in 0.12 M phosphate buffer are not retained during hydroxylapatite chromatography at 60 degrees. The existence of extremely small circles, with contour lengths shorter than the smallest pieces of native DNA, suggests that inverted repetitions containing nicks (broken phosphodiester bonds in duplex DNA) or gaps (interruptions with missing nucleotides) are present at internal positions as well as at the ends of native molecules. Estimates from length measurements of native and denatured Oxytricha macronuclear DNA indicate an average of 1.7 nicks per duplex molecule. Thus, in order to account for the high frequency of circle formation, a restriction-type enzyme(s) must exist which inserts single-strand, site-specific nicks or gaps at internal positions in the macronuclear DNA of Oxytricha.     

Arrangement of sequences in the inverted terminal repetition of adenovirus 18 DNA.

In contrast to the single-stranded circular molecules produced with denatured DNA from other adenoviruses, there was associated with nearly all circular molecules of adenovirus type 18 a visible, duplex projection. These projections had a mean contour length of 0.31 +/- 0.12 mum, equivalent to approximately 3% of genome length. Individual projections ranged in size from 0.1 to 2 mum. Alkaline sucrose gradient purification of single-stranded molecules did not affect formation of these projections, and treatment of a preparation of circular molecules with Neurospora crassa single-strand specific nucleases yielded 0.34 +/- 0.09 mum duplex fragments. Single-stranded circles did not form if a limited number of nucleotides were removed from the 3' ends of native molecules by Escherichia coli exonuclease III digestion prior to denaturation and annealing. In addition, preformed single-stranded circles could be converted to linear molecules by similar treatment. Based on the formation of specific heteroduplex structures when preparations of native DNA were denatured and reannealed and the absence of branches on linear, single-stranded molecules, we conclude that projections are generated by unusually long, inverted terminal repetitions. The repetitious sequences occur in place of rather than in addition to regular sequences. These data provide direct, visual evidence for the arrangement of bases in the inverted terminal repetition of adenovirus DNA.