The structural basis of Holliday junction resolution by T7 endonuclease I

The four-way (Holliday) DNA junction is the central intermediate in homologous recombination, a ubiquitous process that is important in DNA repair and generation of genetic diversity. The penultimate stage of recombination requires resolution of the DNA junction into nicked-duplex species by the action of a junction-resolving enzyme, examples of which have been identified in a wide variety of organisms. These enzymes are nucleases that are highly selective for the structure of branched DNA. The mechanism of this selectivity has, however, been unclear in the absence of structural data. Here we present the crystal structure of the junction-resolving enzyme phage T7 endonuclease I in complex with a synthetic four-way DNA junction. Although the enzyme is structure-selective, significant induced fit occurs in the interaction, with changes in the structure of both the protein and the junction. The dimeric enzyme presents two binding channels that contact the backbones of the junction's helical arms over seven nucleotides. These interactions effectively measure the relative orientations and positions of the arms of the junction, thereby ensuring that binding is selective for branched DNA that can achieve this geometry.     

DNA bending induced by cruciform formation

Cruciform structures in DNA are of considerable interest, both as extreme examples of sequence-dependent structural heterogeneity and as models for four-way junctions such as the Holliday junction of homologous genetic recombination. Cruciforms are of lower thermodynamic stability than regular duplex DNA, and have been observed only in negatively supercoiled molecules, where the unfavourable free energy of formation is offset by the topological relaxation of the torsionally stressed molecule. From an experimental viewpoint this can be a disadvantage, as cruciform structures can be studied only in relatively large supercoiled DNA circles, and are destabilized when a break is introduced at any point. We therefore set out to construct a pseudo-cruciform junction--by generating hereroduplex formation between two inverted repeat sequences. Stereochemically, this should closely resemble a true cruciform but remain stable in a linear DNA fragment. We have now created such a junction and find that it has the expected sensitivities to endonucleases. These DNA fragments exhibit extremely anomalous gel electrophoretic mobility, the extent of which depends on the relative position of the pseudo-cruciform along the length of the molecule. Our results are very similar to those obtained by Wu and Crothers using kinetoplast DNA, and we conclude that the pseudo-cruciform junction introduces a bend in the linear DNA molecule.     

The Bloom's syndrome helicase suppresses crossing over during homologous recombination

Mutations in BLM, which encodes a RecQ helicase, give rise to Bloom's syndrome, a disorder associated with cancer predisposition and genomic instability. A defining feature of Bloom's syndrome is an elevated frequency of sister chromatid exchanges. These arise from crossing over of chromatid arms during homologous recombination, a ubiquitous process that exists to repair DNA double-stranded breaks and damaged replication forks. Whereas crossing over is required in meiosis, in mitotic cells it can be associated with detrimental loss of heterozygosity. BLM forms an evolutionarily conserved complex with human topoisomerase IIIalpha (hTOPO IIIalpha), which can break and rejoin DNA to alter its topology. Inactivation of homologues of either protein leads to hyper-recombination in unicellular organisms. Here, we show that BLM and hTOPO IIIalpha together effect the resolution of a recombination intermediate containing a double Holliday junction. The mechanism, which we term double-junction dissolution, is distinct from classical Holliday junction resolution and prevents exchange of flanking sequences. Loss of such an activity explains many of the cellular phenotypes of Bloom's syndrome. These results have wider implications for our understanding of the process of homologous recombination and the mechanisms that exist to prevent tumorigenesis.     

Aberrant chromosome morphology in human cells defective for Holliday junction resolution

In somatic cells, Holliday junctions can be formed between sister chromatids during the recombinational repair of DNA breaks or after replication fork demise. A variety of processes act upon Holliday junctions to remove them from DNA, in events that are critical for proper chromosome segregation. In human cells, the BLM protein, inactivated in individuals with Bloom's syndrome, acts in combination with topoisomerase IIIα, RMI1 and RMI2 (BTR complex) to promote the dissolution of double Holliday junctions. Cells defective for BLM exhibit elevated levels of sister chromatid exchanges (SCEs) and patients with Bloom's syndrome develop a broad spectrum of early-onset cancers caused by chromosome instability. MUS81-EME1 (refs 4-7), SLX1-SLX4 (refs 8-11) and GEN1 (refs 12, 13) also process Holliday junctions but, in contrast to the BTR complex, do so by endonucleolytic cleavage. Here we deplete these nucleases from Bloom's syndrome cells to analyse human cells compromised for the known Holliday junction dissolution/resolution pathways. We show that depletion of MUS81 and GEN1, or SLX4 and GEN1, from Bloom's syndrome cells results in severe chromosome abnormalities, such that sister chromatids remain interlinked in a side-by-side arrangement and the chromosomes are elongated and segmented. Our results indicate that normally replicating human cells require Holliday junction processing activities to prevent sister chromatid entanglements and thereby ensure accurate chromosome condensation. This phenotype was not apparent when both MUS81 and SLX4 were depleted from Bloom's syndrome cells, suggesting that GEN1 can compensate for their absence. Additionally, we show that depletion of MUS81 or SLX4 reduces the high frequency of SCEs in Bloom's syndrome cells, indicating that MUS81 and SLX4 promote SCE formation, in events that may ultimately drive the chromosome instabilities that underpin early-onset cancers associated with Bloom's syndrome.     

Homology-dependent interactions in phage lambda site-specific recombination

General recombination shows a dependence on large regions of homology between the two participating segments of DNA. Many site-specific recombination systems also exhibit a dependence on homology, although in these systems the requirement is limited to a short region (less than 10 base pairs (bp]. We have used the in vitro phage lambda integration reaction to study the role of homology in this model site-specific recombination system. We find that certain non-homologous pairings which are strongly blocked for complete recombination, nevertheless make one pair of strand-exchanges to generate a joint molecule of the Holliday structure type. This result rules out recombination models in which the only homology-dependent step is synapsis (the juxtaposing of the two recombination sites). Our results also reveal a functional asymmetry in the recombination sites. We present models for bacteriophage lambda integrative recombination which accommodate these findings.     

Unusual helical packing in crystals of DNA bearing a mutation hot spot

The target sequence of the restriction enzyme NarI (GGCGCC) is a hot spot for the -2 frameshift mutagenesis (GGCGCC----GGCC) induced by the chemical carcinogens such as N-2-acetyl-aminofluorene. Of the guanine residues, all of which show equal reactivity towards the carcinogen, only binding to the 3'-most proximal guanine within the NarI site is able to trigger the frameshift event. We selected the non-palindromic dodecamer d(ACCGGCGCCACA), whose sequence corresponds to the most mutagenic NarI site in pBR322 DNA; for X-ray structure analysis. Its molecular structure determined at 2.8 A resolution reveals significant deviations from the structure of canonical B-form DNA, with partial opening of three G-C base pairs, high propeller twist values and sequence-dependent three-centred hydrogen bonds. This crystal structure shows a novel kind of packing in which helices are locked together by groove-backbone interactions. The partial opening of G-C base pairs is induced by interactions of phosphate anionic oxygen atoms with the amino group of cytosine bases. This provides a model for close approach of DNA molecules during biological processes, such as recombination.     

Sequence-specific artificial photo-induced endonucleases based on triple helix-forming oligonucleotides

Homopyrimidine oligonucleotides bind to homopurine-homopyrimidine sequences of duplex DNA forming a local triple helix. This binding can be demonstrated either directly by a footprinting technique, gel assays, or indirectly by inducing irreversible reactions in the target sequence, such as photocrosslinking or cleavage. Binding occurs in the major groove with the homopyrimidine oligonucleotide orientated parallel to the homopurine strand. Thymine and protonated cytosine in the oligonucleotide form Hoogsteen-type hydrogen bonds with A.T and G.C Watson-Crick base pairs, respectively. Here we report that an 11-residue homopyrimidine oligonucleotide covalently attached to an ellipticine derivative by its 3' phosphate photo-induces cleavage of the two strands of a target homopurine--homopyrimidine sequence. To our knowledge, this is the first reported case of a sequence-specific artificial photoendonuclease. In addition we show that a strong binding site for a free ellipticine derivative is induced at the junction between the triplex and duplex structures on the 5' side of the bound oligonucleotide. On irradiation, cleavage is observed on both strands of DNA. This opens new possibilities for inducing irreversible reactions on DNA at specific sites by the synergistic action of a triple helix-forming oligonucleotide and an intercalating agent.     

A 1.7-kilobase single-stranded DNA that folds into a nanoscale octahedron

Molecular self-assembly offers a means of spontaneously forming complex and well-defined structures from simple components. The specific bonding between DNA base pairs has been used in this way to create DNA-based nanostructures and to direct the assembly of material on the subnanometre to micrometre scale. In principle, large-scale clonal production of suitable DNA sequences and the directed evolution of sequence lineages towards optimized behaviour can be realized through exponential DNA amplification by polymerases. But known examples of three-dimensional geometric DNA objects are not amenable to cloning because they contain topologies that prevent copying by polymerases. Here we report the design and synthesis of a 1,669-nucleotide, single-stranded DNA molecule that is readily amplified by polymerases and that, in the presence of five 40-mer synthetic oligodeoxynucleotides, folds into an octahedron structure by a simple denaturation-renaturation procedure. We use cryo-electron microscopy to show that the DNA strands fold successfully, with 12 struts or edges joined at six four-way junctions to form hollow octahedra approximately 22 nanometres in diameter. Because the base-pair sequence of individual struts is not repeated in a given octahedron, each strut is uniquely addressable by the appropriate sequence-specific DNA binder.     

Telomere-telomere recombination provides an express pathway for telomere acquisition

DNA termini from Tetrahymena and Oxytricha, which bear C4A2 and C4A4 repeats respectively, can support telomere formation in Saccharomyces cerevisiae by serving as substrates for the addition of yeast telomeric C1-3A repeats. Previously, we showed that linear plasmids with 108 base pairs of C4A4 DNA (YLp108CA) efficiently acquired telomeres, whereas plasmids containing 28-64 base pairs of C4A4 DNA also promoted telomere formation, but with reduced efficiency. Although many of the C4A4 termini on these plasmids underwent recombination with a C4A2 terminus, the mechanism of telomere-telomere recombination was not established. We now report the sequence of the C4A4 ends from the linear plasmids. The results provide strong evidence for a novel recombination process involving a gene conversion event that requires little homology, occurs at or near the boundary of telomeric and nontelomeric DNA, and resembles the recombination process involved in bacteriophage T4 DNA replication.     

Rad54 protein promotes branch migration of Holliday junctions

Homologous recombination has a crucial function in the repair of DNA double-strand breaks and in faithful chromosome segregation. The mechanism of homologous recombination involves the search for homology and invasion of the ends of a broken DNA molecule into homologous duplex DNA to form a cross-stranded structure, a Holliday junction (HJ). A HJ is able to undergo branch migration along DNA, generating increasing or decreasing lengths of heteroduplex. In both prokaryotes and eukaryotes, the physical evidence for HJs, the key intermediate in homologous recombination, was provided by electron microscopy. In bacteria there are specialized enzymes that promote branch migration of HJs. However, in eukaryotes the identity of homologous recombination branch-migration protein(s) has remained elusive. Here we show that Rad54, a Swi2/Snf2 protein, binds HJ-like structures with high specificity and promotes their bidirectional branch migration in an ATPase-dependent manner. The activity seemed to be conserved in human and yeast Rad54 orthologues. In vitro, Rad54 has been shown to stimulate DNA pairing of Rad51, a key homologous recombination protein. However, genetic data indicate that Rad54 protein might also act at later stages of homologous recombination, after Rad51 (ref. 13). Novel DNA branch-migration activity is fully consistent with this late homologous recombination function of Rad54 protein.     

Fluorescence energy transfer shows that the four-way DNA junction is a right-handed cross of antiparallel molecules

The four-way junction between DNA helices is the central intermediate in recombination, and the manner of its interaction with resolvase enzymes can determine the genetic outcome of the process. A knowledge of its structure is a prerequisite to understanding the interaction with proteins, and there has been recent progress. Here we use fluorescence energy transfer to determine the relative distances between the ends of a small DNA junction, and hence the path of the strands. Our results are consistent with the geometry of an 'X'. The interconnected helices are juxtaposed so that the continuous strands of each helix generate an antiparallel alignment, and the two interchanged strands do not cross at the centre. The acute angle of the X structure is defined by a right-handed rotation of the helical axes about the axis perpendicular to the X plane, as viewed from the centre of the X.     

Communication between segments of DNA during site-specific recombination

Some site-specific recombination systems require the interacting DNA sequences to have a specific relative orientation. This means that DNA segments can sense each other's direction even though they may be separated by many thousands of base pairs. Here, we review the surprising results of recent experiments that lead to a new model which accounts for site orientation specificity and relates it to other recombination systems where relative orientation is not critical.     

Enzymatic capture of an extrahelical thymine in the search for uracil in DNA

The enzyme uracil DNA glycosylase (UNG) excises unwanted uracil bases in the genome using an extrahelical base recognition mechanism. Efficient removal of uracil is essential for prevention of C-to-T transition mutations arising from cytosine deamination, cytotoxic U*A pairs arising from incorporation of dUTP in DNA, and for increasing immunoglobulin gene diversity during the acquired immune response. A central event in all of these UNG-mediated processes is the singling out of rare U*A or U*G base pairs in a background of approximately 10(9) T*A or C*G base pairs in the human genome. Here we establish for the human and Escherichia coli enzymes that discrimination of thymine and uracil is initiated by thermally induced opening of T*A and U*A base pairs and not by active participation of the enzyme. Thus, base-pair dynamics has a critical role in the genome-wide search for uracil, and may be involved in initial damage recognition by other DNA repair glycosylases.     

DNA bending at adenine . thymine tracts

Intrinsic bending of DNA molecules results from local structural polymorphism in regions of homopolymeric dA . dT which are at least 4 base pairs long; the A . T tracts must be repeated in phase with the helix screw. Bending, in the direction of base-pair tilt rather than roll, occurs at the junctions between the A . T tract and adjacent B-DNA, with a larger angle at the 3' than at the 5' end of the A tract.     

The locus of sequence-directed and protein-induced DNA bending

The bending locus of trypanosome kinetoplast DNA, identified by gel electrophoresis, has tracts of a simple repeat sequence (CA5-6 T) symmetrically distributed about it, with a repeat interval of 10 base pairs. The analogous bending induced when catabolite gene activating protein binds to its recognition sequence near the promoter of the Escherichia coli lac operon is centred on a site about 5-7 base pairs away from the centre of the protein binding site.     

High-resolution structure of a DNA helix containing mismatched base pairs

The concept of complementary base pairing, integral to the double-helical structure of DNA, provides an effective and elegant mechanism for the faithful transmission of genetic information. Implicit in this model, however, is the potential for incorporating non-complementary base pairs (mismatches) during replication or subsequently, for example, during genetic recombination. As such errors are usually damaging to the organism, they are generally detected and repaired. Occasionally, however, the propagation of erroneous copies of the genome confers a selective advantage, leading to genetic variation and evolutionary change. An understanding of the nature of base-pair mismatches at a molecular level, and the effect of incorporation of such errors on the secondary structure of DNA is thus of fundamental importance. We now report the first single-crystal X-ray analysis of a DNA fragment, d(GGGGCTCC), which contains two non-complementary G X T base pairs, and discuss the implications of the results for the in vivo recognition of base-pair mismatches.     

Crystal structure of the met repressor-operator complex at 2.8 A resolution reveals DNA recognition by beta-strands.

The crystal structure of the met repressor-operator complex shows two dimeric repressor molecules bound to adjacent sites 8 base pairs apart on an 18-base-pair DNA fragment. Sequence specificity is achieved by insertion of double-stranded antiparallel protein beta-ribbons into the major groove of B-form DNA, with direct hydrogen-bonding between amino-acid side chains and the base pairs. The repressor also recognizes sequence-dependent distortion or flexibility of the operator phosphate backbone, conferring specificity even for inaccessible base pairs.     

Molecular structure of a left-handed double helical DNA fragment at atomic resolution

The DNA fragment d(CpGpCpGpCpG) crystallises as a left-handed double helical molecule with Watson-Crick base pairs and an antiparallel organisation of the sugar phosphate chains. The helix has two nucleotides in the asymmetric unit and contains twelve base pairs per turn. It differs significantly from right-handed B-DNA.     

DNA twisting and the effects of non-contacted bases on affinity of 434 operator for 434 repressor.

The bacteriophage 434 repressor regulates gene expression by binding with differing affinities to the six operator sites on the phage chromosome. The symmetrically arrayed outer eight base pairs (four in each half-site) of these 14-base-pair operators are highly conserved but the middle four bases are divergent. Although these four base pairs are not in contact with repressor, operators with A.T or T.A base pairs at these positions bind repressor more strongly than those bearing C.G or G.C, suggesting that these bases are important for the repressor's ability to discriminate between operators. There is evidence that the central base pairs influence operator function by constraining the twisting and/or bending of DNA. Here we show that there is a relationship between the intrinsic twist of an operator, as determined by the sequence of its central bases, and its affinity for repressor; an operator with a lower affinity is undertwisted relative to an operator with higher affinity. In complex with repressor, the twist of both high- and low-affinity operators is the same. These results indicate that the intrinsic twist of DNA and its twisting flexibility both affect the affinity of 434 operator for repressor.