The mechanism of cruciform formation in supercoiled DNA: initial opening of central basepairs in salt-dependent extrusion.

There are two alternative pathways by which inverted repeat sequences in supercoiled DNA molecules may extrude cruciform structures, called C-type and S-type. S-type cruciforms, which form the great majority, are characterised by absolute requirement for cations to promote extrusion, which then proceeds at higher temperatures and with lower activation parameters than for C-type cruciforms. The mechanism proposed for S-type extrusion involves an initial opening of basepairs limited to the centre of the inverted repeat, formation of intra-strand basepairing and a four-way junction, and finally branch migration to the fully extruded cruciform. The model predicts that central sequence changes will be more kinetically significant than those removed from the centre. We have studied the kinetics of cruciform extrusion by a series of inverted repeats related to that of pIRbke8 by either one or two mutations in the symmetric unit. We find that mutations in the central 8 to 10 nucleotides may profoundly affect extrusion rates--the fastest being 2000-fold faster than the slowest, whereas mutations further from the centre affect rates to a much smaller extent, typically up to ten-fold. These data support the proposed mechanism for extrusion via central opening.     

Large-scale stable opening of supercoiled DNA in response to temperature and supercoiling in (A + T)-rich regions that promote low-salt cruciform extrusion.

We have studied the properties of (A + T)-rich sequences derived from ColE1 that promote cruciform extrusion at low ionic strength in supercoiled plasmids. We compared the chemical reactivity of the sequences in negatively supercoiled DNA (using osmium tetroxide and bromoacetaldehyde) with the results of two-dimensional gel electrophoresis performed under the same conditions. Taken together, the results indicate the occurrence of cooperative helix-coil transitions in the (A + T)-rich DNA at low ionic strength, to form stable, denatured regions. The extent of the open region is a function of temperature and superhelix density, with an additional local destabilization brought about by the presence of cruciform structures. We present a simple statistical mechanical model of the helix-coil transition in the (A + T)-rich DNA, from which we have obtained estimates of the free energy for average base-pair opening of 0.31 kcal mol-1 and that for the formation of a helix-coil junction of 4.9 kcal mol-1, in 45 mM Tris-borate, pH 8.3, 0.5 mM EDTA. The results offer a model for the C-type mechanism of cruciform extrusion. Inverted repeats that are incorporated into the melted region undergo hairpin loop formation below 50 degrees C, and upon closure of the melted region, by reduction of temperature or increased ionic strength, they remain as a fully extruded cruciform structure.     

Effect of base composition at the center of inverted repeated DNA sequences on cruciform transitions in DNA

We have analyzed the effect of base composition at the center of symmetry of inverted repeated DNA sequences on cruciform transitions in supercoiled DNA. For this we have constructed two series of palindromic DNA sequences: one set with differing center and one set with differing center and arm sequences. The F series consists of two 96-base pair perfect inverted repeats which are identical except for the central 10 base pairs which consist of pure AT or GC base pairs. The S series was constructed such that the overall base composition of the inverted repeats was identical but in which the positioning of blocks of AT- and GC-rich sequences varied. The rate of cruciform formation for the inverted repeats in plasmid pUC8 was dramatically influenced by the 8-10 base pairs at the center of the inverted repeat. Inverted repeats with 8-10 AT base pairs in the center were kinetically much more active in cruciform formation than inverted repeats with 8-10 GC base pairs in the center. These experiments show a dominant influence of the center sequences of inverted repeats on the rate of cruciform formation.     

The physical chemistry of cruciform structures in supercoiled DNA molecules

Inverted repeat DNA sequences extrude cruciform structures when present in negatively supercoiled molecules, stabilised by the release of torsional stress brought about by the negative twist change. We have revealed the presence of cruciform structures by means of enzyme and chemical probing experiments and topological band shift methods. The geometry of cruciform structures has been studied from two points of view. The unpairing of bases in the loop region has been investigated using bisulphite modification, with the result that the central four nucleotides have single-stranded character, and the next pair have only partially single-stranded nature. Gel electrophoretic studies of a pseudo-cruciform structure indicate that the cruciform junction introduces a pronounced bend into the molecule. The dependence of the formation of the ColE1 cruciform upon DNA supercoiling shows that it has a free energy of formation of 18.4 +/- 0.5 kcal mole-1. The kinetics of the extrusion process are complex. Most sequences extrude slowly with considerable temperature coefficients, but the detailed properties are strongly sequence-dependent. One synthetic inverted repeat sequence which we have studied in detail has an Arrhenius activation energy of 42.4 +/- 3.2 kcal mole-1. We discuss possible mechanistic pathways for the extrusion process.     

Protein HU binds specifically to kinked DNA

We have purified the main four-way junction DNA-binding protein of Escherichia coli, and have found It to be the well-known HU protein. HU protein recognizes with high-affinity one of the angles present in the junction, a molecule with the shape of an X. Other DNA structures characterized by sharp bends or kinks, like bulged duplex DNAs containing unpaired bases, are also bound. HU protein appears to inhibit cruciform extrusion from supercoiled inverted repeat (palindromic) DNA, either by constraining supercoiling or by trapping a metastable interconversion intermediate. All these properties are analogous to the properties of the mammalian chromatin protein HMG1. We suggest that HU is a prokaryotic HMG1-like protein rather than a histone-like protein.     

Cruciform-resolvase interactions in supercoiled DNA

T4 endonuclease VII, which cleaves Holliday-like junctions in DNA, specifically cleaves short inverted repeats in supercoiled plasmids. These sequences are subject to site-specific cleavage by single-strand-specific nucleases, and cruciform formation has been suggested as an explanation for this observation. This proposal is greatly strengthened by the present data, since a formal analogy between cruciform structures and Holliday junctions exists. Resolution of a variety of unrelated cruciform sequences demonstrates that the cleavage process results in a linear molecule with hairpin ends and single ligatable nicks at positions corresponding to the stem-base of the cruciform. In two examples mapped in detail, the cleavages are exclusively introduced at two or three nucleotides from the end of the symmetric sequence at the 5' side on each strand. These studies demonstrate the potential of endonuclease VII as a probe of cruciform structure and the utility of short cruciform structures as Holliday junction models.     

A dominant influence of flanking sequences on a local structural transition in DNA

We have discovered a striking dependence of a structural transition in DNA on sequences that are distanced from those directly participating in the transformation. The dominant factor determining the selection of kinetic properties of cruciform extrusion is the sequence of the DNA that flanks the inverted repeat. The sequence of the inverted repeat itself appears to have little or no influence. The critical sequences that confer the unusual kinetics exhibited by the ColE1 cruciform are very A+T-rich. A single such sequence is sufficient, which may be as short as 100 bp, and it can control inverted repeats placed at either end. The effects operate in cis, are independent of polarity, and may be effective over relatively long distances. The influence of context has wide implications, possibly including the control of gene expression.     

Long-range structural effects in supercoiled DNA: statistical thermodynamics reveals a correlation between calculated cooperative melting and contextual influence on cruciform extrusion

We have previously described [K. M. Sullivan and D. M. J. Lilley (1986) Cell 47, 817-827] a set of sequences, called C-type inducing sequences, which cause cruciform extrusion by adjacent inverted repeats to occur by an abnormal kinetic pathway involving a large denatured region of DNA. In this paper we apply statistical thermodynamic DNA helix melting theory to these sequences. We find a marked correlation between the ability of sequences to confer C-type cruciform character experimentally and their calculated propensity to undergo cooperative melting, and no exceptions have been found. The correlations are both qualitative and quantitative. Thus the ColE1 flanking sequences behave as single melting units, while the DNA of the S-type plasmid pIRbke8 exhibits no propensity to melt in the region of the bke cruciform. The results of the calculations are also fully consistent with the following experimental observations: 1. the ability of the isolated colL and colR fragments of the ColE1 flanking sequences, as well as the short sequence col30, to confer C-type character; 2. C-type induction by an A + T rich Drosophila sequence; 3. low-temperature cruciform extrusion by an (AT)34 sequence; 4. the effect of changing sequences at a site 90 base pairs (bp) removed from the inverted repeat; 5. the effects of systematic deletion of the colL sequence; and 6. the effects of insertion of various sequences in between the colL sequence and the xke inverted repeat. These studies show that telestability effects on thermal denaturation as predicted from equilibrium helix melting theory of linear DNA molecules may explain all the features that are revealed by studying the extrusion of cruciforms in circular DNA molecules subjected to superhelical stress.     

Target site recognition by a diversity-generating retroelement

Diversity-generating retroelements (DGRs) are in vivo sequence diversification machines that are widely distributed in bacterial, phage, and plasmid genomes. They function to introduce vast amounts of targeted diversity into protein-encoding DNA sequences via mutagenic homing. Adenine residues are converted to random nucleotides in a retrotransposition process from a donor template repeat (TR) to a recipient variable repeat (VR). Using the Bordetella bacteriophage BPP-1 element as a prototype, we have characterized requirements for DGR target site function. Although sequences upstream of VR are dispensable, a 24 bp sequence immediately downstream of VR, which contains short inverted repeats, is required for efficient retrohoming. The inverted repeats form a hairpin or cruciform structure and mutational analysis demonstrated that, while the structure of the stem is important, its sequence can vary. In contrast, the loop has a sequence-dependent function. Structure-specific nuclease digestion confirmed the existence of a DNA hairpin/cruciform, and marker coconversion assays demonstrated that it influences the efficiency, but not the site of cDNA integration. Comparisons with other phage DGRs suggested that similar structures are a conserved feature of target sequences. Using a kanamycin resistance determinant as a reporter, we found that transplantation of the IMH and hairpin/cruciform-forming region was sufficient to target the DGR diversification machinery to a heterologous gene. In addition to furthering our understanding of DGR retrohoming, our results suggest that DGRs may provide unique tools for directed protein evolution via in vivo DNA diversification.     

Inverted repeat sequences can influence the melting transitions of linear DNAs

The influence of inverted repeat sequences on the melting transitions of linear DNAs has been examined. Derivative melting curves (DMC) of a 514 base pair (bp) DNA, seven subfragments of this DNA, and four other DNAs have been compared to predictions of DNA melting theory. The 514-bp DNA contains three inverted repeat sequences that can form cruciform structures in supercoiled DNA. We refer to these sequences as c-inverted repeats. Previous work showed that the DMC of this DNA, unlike a number of other DNAs, is not accurately predicted by DNA melting theory. Since the theoretical model does not include hairpin-like structures, it was suggested that hairpin or cruciform formation in these inverted repeats may be responsible for this discrepancy. Our results support this hypothesis. Predicted DMCs are in good agreement with DNAs with no inverted repeats, or inverted repeats not evident in supercoiled DNA. Differences between the theoretical and experimental Tm's are less than or equal to 0.3 degrees C. DNA molecules that contain one or more of the three c-inverted repeats are not as accurately predicted. Experimental Tm values are lower than predicted values by 0.7-3.8 degrees C. It is concluded that some inverted repeat sequences can form hairpin-like structures during the melting of linear DNAs. These structures appear to lower overall DNA stability.     

Evidence for two mechanisms of palindrome-stimulated deletion in Escherichia coli: single-strand annealing and replication slipped mispairing

Spontaneous deletion mutations often occur at short direct repeats that flank inverted repeat sequences. Inverted repeats may initiate genetic rearrangements by formation of hairpin secondary structures that block DNA polymerases or are processed by structure-specific endonucleases. We have investigated the ability of inverted repeat sequences to stimulate deletion of flanking direct repeats in Escherichia coli. Propensity for cruciform extrusion in duplex DNA correlated with stimulation of flanking deletion, which was partially sbcD dependent. We propose two mechanisms for palindrome-stimulated deletion, SbcCD dependent and SbcCD independent. The SbcCD-dependent mechanism is initiated by SbcCD cleavage of cruciforms in duplex DNA followed by RecA-independent single-strand annealing at the flanking direct repeats, generating a deletion. Analysis of deletion endpoints is consistent with this model. We propose that the SbcCD-independent pathway involves replication slipped mispairing, evoked from stalling at hairpin structures formed on the single-stranded lagging-strand template. The skew of SbcCD-independent deletion endpoints with respect to the direction of replication supports this hypothesis. Surprisingly, even in the absence of palindromes, SbcD affected the location of deletion endpoints, suggesting that SbcCD-mediated strand processing may also accompany deletion unassociated with secondary structures.     

Molecular cloning and sequence of the concatemer junction from vaccinia virus replicative DNA. Viral nuclease cleavage sites in cruciform structures

The concatemer junction from replicative forms of vaccinia virus DNA was cloned into plasmid vectors and shown to be a precise duplex copy of the viral terminal hairpin structure, with each strand corresponding to one of the alternative sequence isomers. The plasmids were relaxed circles with extruded cruciforms representing two copies of the vaccinia telomere hairpin structure. Head-to-head dimers containing two copies of the vaccinia virus concatemer junction were observed to contain only one set of stem-loop structures per molecule, suggesting that the initial formation of a small cruciform, and not branch migration, was the rate-limiting step in cruciform formation. The plasmids containing the concatemer junction were converted into nicked circular, linear and cross-linked linear molecules by a nuclease isolated from vaccinia virions. The region-specific cleavage near the border of the hairpin loop and the formation of DNA cross-links in some of the molecules is consistent with the nuclease acting as a nicking-closing enzyme that participates in the resolution of mature termini from replicative concatemer intermediates.     

Inverted repeats,stem-loops,and cruciforms: Significance for initiation of DNA replication

Inverted repeats occur nonrandomly in the DNA of most organisms. Stem-loops and cruciforms can form from inverted repeats. Such structures have been detected in pro- and eukaryotes. They may affect the supercoiling degree of the DNA, the positioning of nucleosomes, the formation of other secondary structures of DNA, or directly interact with proteins. Inverted repeats, stem-loops, and cruciforms are present at the replication origins of phage, plasmids, mitochondria, eukaryotic viruses, and mammalian cells. Experiments with anti-cruciform antibodies suggest that formation and stabilization of cruciforms at particular mammalian origins may be associated with initiation of DNA replication. Many proteins have been shown to interact with cruciforms, recognizing features like DNA crossovers, four-way junctions, and curved/bent DNA of specific angles. A human cruciform binding protein (CBP) displays a novel type of interaction with cruciforms and may be linked to initiation of DNA replication. © 1996 Wiley-Liss, Inc.     

Mus81-dependent double-strand DNA breaks at in vivo-generated cruciform structures in S. cerevisiae

Long DNA palindromes are implicated in chromosomal rearrangement, but their roles in the underlying molecular events remain a matter of conjecture. One notion is that palindromes induce DNA breaks after assuming a cruciform structure, the four-way DNA junction providing a target for cleavage by Holliday junction (HJ)-specific enzymes. Though compelling, few components of the "cruciform resolution" proposal are established. Here we address fundamental properties and genetic dependencies of palindromic DNA metabolism in eukaryotes. Plasmid-borne palindromes introduced into S. cerevisiae are site-specifically broken in vivo, and the breaks exhibit unique hallmarks of an HJ resolvase mechanism. In vivo resolution requires Mus81, for which the bacterial HJ resolvase RusA will substitute. These results provide confirmation of cruciform extrusion and resolution in the context of eukaryotic chromatin. Related observations are that, unchecked by a nuclease function provided by Mre11, episomal palindromes launch a self-perpetuating breakage-fusion-bridge-independent copy number increase termed "escape."     

A cruciform-dumbbell model for inverted dimer formation mediated by inverted repeats.

Small inverted repeats (small palindromes) on plasmids have been shown to mediate a recombinational rearrangement event in Escherichia coli leading to the formation of inverted dimers (giant palindromes). This recombinational rearrangement event is efficient and independent of RecA and RecBCD. In this report, we propose a cruciform-dumbbell model to explain the inverted dimer formation mediated by inverted repeats. In this model, the inverted repeats promote the formation of a DNA cruciform which is processed by an endonuclease into a linear DNA with two hairpin loops at its ends. Upon DNA replication, this linear dumbbell-like DNA is then converted to the inverted dimer. In support of this model, linear dumbbell DNA molecules with unidirectional origin of DNA replication (ColE1 ori ) have been constructed and shown to transform E.coli efficiently resulting in the formation of the inverted dimer. The ability of linear dumbbell DNA to transform E.coli suggests that the terminal loops may be important in bypassing the requirement of DNA supercoiling for initiation of replication of the ColE1 ori.     

Construction of a DNA four-way junction: Design and NMR spectroscopy

In 1964 Holliday postulated the formation of cruciform structures (four-way junctions) in duplex DNA as intermediate in genetic recombination. Since then, many biochemical and biophysical investigations were directed at solving questions concerning structural details of stable four-way junctions. Thus far, NMR spectroscopy played a minor part in these investigations on account of the minimum size of the molecule (expressed as the number of nucleotide residues) that was thought necessary to produce a stable cruciform structure. Indeed, the smallest four-way junction studied thus far by NMR methods was built from four separate DNA strands, each containing 16 nucleotides, a total of 64. Obviously, with such a large structure one runs into assignment problems. We considered the possibility of constructing a stable four-way junction from a single strand of DNA. The underlying idea was to make use of our detailed knowledge of the building principles of stable minihairpin loops. These loops, containing only two nucleotides to bridge the gap between antiparallel strands, are maximally stable in DNA sequences like 5-d(-C-TT-G-), where C and G form a normal Watson-Crick base pair and the two T residues cross the minor groove to form the minihairpin loop. Three of such miniloops could in principle cap three arms of the cruciform. The fourth arm would have an open end. The problem to be solved is to find the minimum length that is required to insure stability of the three closed arms and of the fourth open arm. We were successful with a structure that has three short stems (four base pairs each) and an open-end stem consisting of eight base pairs, a total of 46 residues. NMR experiments, carried out on this molecule in the presence of Mg²⁺, showed details of folding which have never been observed before.     

The structure of cruciforms in supercoiled DNA: probing the single-stranded character of nucleotide bases with bisulphite.

The single-stranded character of cytosine bases in three cruciform structures has been assessed by an examination of reactivity towards sodium bisulphite. Unpaired cytosine residues undergo deamination at C4 to give deoxyuracil, and propagation in an ung Escherichia coli host results in C-G----T-A transition mutations, detectable by restriction cleavage or sequence analysis. Very high frequencies of such mutations have been found at cruciform loops, confirming their unpaired character, with almost zero background mutation frequencies elsewhere. A low level of modification was observed at the four-way junction of a cruciform. The results indicate that the optimal cruciform loop size is four bases, with loose 'breathing' at the first base pair at the top of the cruciform stem at 37 degrees C, and little or no opening of base pairs at the four-way junction.     

Cruciform extrusion in plasmids bearing the replicative intermediate configuration of a poxvirus telomere

The transition from lineform DNA to cruciform DNA (cruciformation) within the cloned telomere sequences of the Leporipoxvirus Shope fibroma virus (SFV) has been studied. The viral telomere sequences have been cloned in recombination-deficient Escherichia coli as a 322 base-pair, imperfect palindromic insert in pUC13. The inverted repeat configuration is equivalent to the arrangement of the telomere structures observed within viral DNA replicative intermediates. A major cruciform structure in the purified recombinant plasmid has been identified and mapped using, as probes, the enzymes AflII, nuclease S1 and bacteriophage T7 endonuclease I. It was extruded from the central axis of the cloned viral inverted repeat and, by unrestricted branch migration, attained a size commensurate with the superhelical density of the plasmid molecule at native superhelical densities. This major cruciform extrusion event was the only detectable duplex DNA perturbation, induced by negative superhelical torsion, in the insert viral sequences. No significant steady-state pool of extruded cruciform was identified in E. coli. However, the identification of a major deletion variant generated even in the recombination-deficient E. coli strain DB1256 (recA recBC sbcB) suggested that the cruciform may be extruded transiently in vivo. The lineform to cruciform transition has been further characterized in vitro using two-dimensional agarose gel electrophoresis. The transition was marked by a high energy of formation (delta Gf = 44 kcal/mol), and an apparently low activation energy that enabled facile transitions at physiological temperatures provided there was sufficient torsional energy. By comparing cruciformation in a series of related bidirectional central axis deletions of the telomeric insert, it has been concluded that the presence of extrahelical bases in the terminal hairpin structures contributes substantially to the high delta Gf value. Also, viral sequences flanking the extruded cruciform were shown to influence the measured delta Gf value. Several general features of poxvirus telomere structure that would be expected to influence the facility of cruciform extrusion are discussed along with the implications of the observed cruciform transition event on the replicative process of poxviruses in vivo.     

Extrusion of an imperfect palindrome to a cruciform in superhelical DNA: complete determination of energetics using a statistical mechanical model

We present a detailed study of the extrusion of an imperfect palindrome, derived from the terminal regions of vaccinia virus DNA and contained in a superhelical plasmid, into a cruciform containing bulged bases. We monitor the course of extrusion by two-dimensional gel electrophoresis experiments as a function of temperature and linking number. We find that extrusion pauses at partially extruded states as negative superhelicity increases. To understand the course of extrusion with changes in linking number, DeltaLk, we present a rigorous semiempirical statistical mechanical analysis that includes complete coupling between DeltaLk, cruciform extrusion, formation of extrahelical bases, and temperature-dependent denaturation. The imperfections in the palindrome are sequentially incorporated into the cruciform arms as hairpin loops, single unpaired bases, and complex local regions containing several unpaired bases. We analyze the results to determine the free energies, enthalpies and entropies of formation of all local structures involved in extrusion. We find that, for each unpaired structure, the DeltaG, DeltaH and DeltaS of formation are all approximately proportional to the number of unpaired bases contained therein. This surprising result holds regardless of the arrangement or composition of unpaired bases within a particular structure. Imperfections have major effects on the overall energetics of cruciform extrusion and on the course of this transition. In particular, the extent of the DeltaLk change necessary to extrude an imperfect palindrome is considerably greater than that required for extrusion of the underlying perfect palindrome. Our analysis also suggests that, at higher temperatures, significant denaturation at the base of an imperfect cruciform can successfully compete with extension of the cruciform arms.