DNA deformations near charged surfaces: electron and atomic force microscopy views.

DNA is a very important cell structural element, which determines the level of expression of genes by virtue of its interaction with regulatory proteins. We use electron (EM) and atomic force microscopy (AFM) to characterize the flexibility of double-stranded DNA ( approximately 150-950 nm long) close to a charged surface. Automated procedures for the extraction of DNA contours ( approximately 10-120 nm for EM data and approximately 10-300 nm for AFM data) combined with new statistical chain descriptors indicate a uniquely two-dimensional equilibration of the molecules on the substrate surface regardless of the procedure of molecule mounting. However, in contrast to AFM, the EM mounting leads to a noticeable decrease in DNA persistence length together with decreased kurtosis. Analysis of local bending on short length scales (down to 6 nm in the EM study) shows that DNA flexibility behaves as predicted by the wormlike chain model. We therefore argue that adhesion of DNA to a charged surface may lead to additional static bending (kinking) of approximately 5 degrees per dinucleotide step without impairing the dynamic behavior of the DNA backbone. Implications of this finding are discussed.

Blinking fluorophores: what do they tell us about protein dynamics?

The ability to detect emission from a single fluorophore presents a powerful tool to probe the dynamic properties of protein molecules during their interactions with ligands. Here, different classes of experiments are reviewed and a distinction is drawn between experiments that monitor signals from a large number of proteins, one molecule at a time, from those that follow a single protein molecule over many individual cycles. The latter approach is potentially capable of resolving dynamic heterogeneity, such as that displayed by enzymes that fluctuate between high and low activity states. Other factors that need to be considered are the origin of the fluctuations in the emission signal and the extent to which this represents the properties of the protein under investigation, as opposed to the probe itself. Most fluorophores show fluctuations in their emission rates, termed flickering, blinking or intermittency, which may occur on a similar time-scale as the event under investigation.

Dynamic interactions of p53 with DNA in solution by time-lapse atomic force microscopy.

Dynamic interactions of the tumor suppressor protein p53 with a DNA fragment containing a p53-specific recognition sequence were directly observed by time-lapse tapping mode atomic force microscopy (AFM) in liquid. The divalent cation Mg(2+) was used to loosely attach both DNA and p53 to a mica surface so they could be imaged by the AFM while interacting with each other. Various interactions of p53 with DNA were observed, including dissociation/re-association, sliding and possibly direct binding to the specific sequence. Two modes of target recognition of p53 were detected: (a) direct binding, and (b) initial non-specific binding with subsequent translocation by one-dimensional diffusion of the protein along the DNA to the specific site.

Electron and scanning force microscopy studies of alterations in supercoiled DNA tertiary structure.

The configuration of supercoiled DNA (scDNA) was investigated by electron microscopy and scanning force microscopy. Changes in configuration were induced by varying monovalent/divalent salt concentrations and manifested by variation in the number of nodes (crossings of double helical segments). A decrease in the concentration of monovalent cations from 50 mM to approximately 1 mM resulted in a significant change of apparent configuration of negatively supercoiled DNA from a plectonemic form with virtually approximately 15 nodes (the value expected for molecules of approximately 3000 bp) to one or two nodes. This result was in good agreement with values calculated using an elastic rod model of DNA and salt concentration in the range of 5-50 mM. The effect did not depend on the identity of the monovalent cation (Na(+), K(+)) or the nature of the support used for electron microscopy imaging (glow-discharged carbon film, polylysine film). At very low salt concentrations, a single denatured region several hundred base-pairs in length was often detected. Similarly, at low concentrations of divalent cations (Mg(2+), Ca(2+), Zn(2+)), scDNA was apparently relaxed, although the effect was slightly dependent on the nature of the cation. Positively supercoiled DNA behaved in a manner different from that of its negative counterpart when the ion concentration was varied. As expected for these molecules, an increase in salt concentration resulted in an apparent relaxation; however, a decrease in salt concentration also led to an apparent relaxation manifested by a slight decrease in the number of nodes. Scanning force microscopy imaging of negatively scDNA molecules deposited onto a mica surface under various salt conditions also revealed an apparent relaxation of scDNA molecules. However, due to weak interactions with the mica surface in the presence of a mixture of mono/divalent cations, the effect occurred under conditions differing from those used for electron microscopy. We conclude that the observed changes in scDNA configuration are inherent to the DNA structure and do not reflect artifacts arising from the method(s) of sample preparation.

Scanning force microscopy of the complexes of p53 core domain with supercoiled DNA.

We used scanning force microscopy to analyse the interaction of the core domain of the tumor suppressor protein p53 (p53CD, amino acid residues 94 to 312), with supercoiled DNA (scDNA) molecules. The complexes were attached to a mica substrate by the divalent cation spreading technique. p53CD bound to supercoiled plasmid pPGM1 bearing the consensus sequence 5'-AGACATGCCTAGACATGCCT-3' (p53CON) was imaged as a globular complex. Only one such complex was observed with each scDNA molecule. In contrast, binding to supercoiled pBluescript II SK(-) DNA (lacking the consensus sequence) resulted in the appearance of multiple, variable size complexes of various sizes on single DNA molecules. Addition of p53CD to scDNA containing a cruciform-forming (AT)(34) insert resulted in the binding of the protein exclusively at the cruciform. The data presented here suggest that p53CD can form stable specific and non-specific complexes with supercoiled DNA molecules, albeit of variable multimeric organization.

HMG1 protein stimulates DNA end joining by promoting association of DNA molecules via their ends.

High mobility group (HMG) 1 protein is a highly abundant and an evolutionarily conserved chromosomal protein with two homologous DNA-binding domains (HMG boxes), A and B, attached by a short basic region to an acidic C-terminal tail. The protein has been implicated in a number of fundamental biological processes including DNA replication, transcription, recombination and repair. We demonstrate that HMG1 is able to enhance cohesive-end and blunt-end DNA ligation by T4 DNA ligase via its B domain. The C-terminal flanking sequence of the B domain (seven basic residues out of approximately 18) and a number of conserved amino-acid residues within the HMG box (mainly basic or hydrophobic) are required for efficient stimulation of ligation. Pull-down assays, electron and scanning force microscopy revealed that HMG1 can associate two DNA molecules via their ends even in the absence of complementary overhangs. We propose that HMG1 protein may be involved in the rejoining of DNA breaks by different DNA ligases due to its ability to bring DNA duplexes and their termini into a close proximity while leaving the ends accessible for ligation.

DNA bending due to specific p53 and p53 core domain-DNA interactions visualized by electron microscopy.

We have used transmission electron microscopy to analyze the specificity and the extent of DNA bending upon binding of full-length wild-type human tumor suppressor protein p53 (p53) and the p53 core domain (p53CD) encoding amino acid residues 94-312, to linear double-stranded DNA bearing the consensus sequence 5'-AGACATGCCTAGACATGCCT-3' (p53CON). Both proteins interacted with high specificity and efficiency with the recognition sequence in the presence of 50 mM KCl at low temperature ( approximately 4 degrees C) while the p53CD also exhibits a strong and specific interaction at physiological temperature. Specific complex formation did not result in an apparent reduction of the DNA contour length. The interaction of p53 and the p53CD with p53CON induced a noticeable salt-dependent bending of the DNA axis. According to quantitative analysis with folded Gaussian distributions, the bending induced by p53 varied from approximately 40 degrees to 48 degrees upon decreasing of the KCl concentration from 50 mM to approximately 1 mM in the mounting buffer used for adsorption of the complexes to the carbon film surface. The p53CD bent DNA by 35-37 degrees for all salt concentrations used in the mounting buffer. The bending angle of the p53/DNA complex under low salt conditions showed a somewhat broader distribution (sigma approximately 39 degrees ) than at high salt concentration (sigma approximately 31 degrees ) or for p53CD (sigma approximately 24-27 degrees ). Together, these results demonstrate that the p53CD has a dominant role in complex formation and that the complexes formed both by p53 and p53CD under moderate salt conditions are similar. However, the dependence of the bending parameters on ambient conditions suggest that the segments flanking the p53CD contribute to complex formation as well. The problems associated with the analysis of bending angles in electron microscopy experiments are discussed.

The Cfr10I restriction enzyme is functional as a tetramer.

It is thought that most of the type II restriction endonucleases interact with DNA as homodimers. Cfr10I is a typical type II restriction enzyme that recognises the 5'-Pu decreases CCGGPy sequence and cleaves it as indicated by the arrow. Gel-filtration and analytical ultracentrifugation data presented here indicate that Cfr10I is a homotetramer in isolation. The only SfiI restriction enzyme that recognises the long interrupted recognition sequence 5'-GGCCNNNNNGGCC has been previously reported to operate as a tetramer however, its structure is unknown. Analysis of Cfr10I crystals revealed that a single molecule in the asymmetric unit is repeated by D2 symmetry to form a tetramer. To determine whether the packing of the Cfr10I in the crystal reflects the quaternary structure of the protein in solution, the tryptophan W220 residue located at the putative dimer-dimer interface was mutated to alanine, and the structural and functional consequences of the substitution were analysed. Equilibrium sedimentation experiments revealed that, in contrast to the wild-type Cfr10I, the W220A mutant exists in solution predominantly as a dimer. In addition, the tetramer seems to be a catalytically important form of Cfr10I, since the DNA cleavage activity of the W220A mutant is < 0.1% of that of the wild-type enzyme. Further, analysis of plasmid DNA cleavage suggests that the Cfr10I tetramer is able to interact with two copies of the recognition sequence, located on the same DNA molecule. Indeed, electron microscopy studies demonstrated that two distant recognition sites are brought together through the DNA looping induced by the simultaneous binding of the Cfr10I tetramer to both sites. These data are consistent with the tetramer being a functionally important form of Cfr10I.

Static curvature and flexibility measurements of DNA with microscopy. A simple renormalization method, its assessment by experiment and simulation.

We present the derivation of equations based on statistical polymer chain analysis and a method to quantify the average angle value of intrinsic bends and the local flexibility at a given locus on DNA fragments imaged by electron microscopy. DNA fragments of n base-pairs are considered as stiff chains of n jointed unit rigid rods. If the DNA fragments are composed of two branches A0Am and A0Bn, with, respectively, m and n base-pairs, where the standard deviations of the angle formed by two consecutive base-pairs are uniform over each branch, respectively, sigmathetaA and sigmathetaB, we show that the standard deviation of the angle AmA0Bn is: [formula: see text] where sigmatheta0 is the standard deviation of the angle at locus A0. This equation is established for small angular deviations by analysis of DNA at different scales and the validity of the methodology is controlled with the computation of the reduced chi2 statistical test. The length of the DNA fragments must be of the order of, or below, the persistence length, as determined by sets of statistics from computer simulations of DNA fragments. This is verified experimentally by a detailed analysis of the digitized contours of homogeneous linear 139 base-pair DNA fragments observed by electron microscopy. The images are compared to the reconstruction of DNA fragments from the measurements. The value found, sigma0=4.6 degrees/bp, is consistent with the well-accepted value for DNA in a plane. We discuss the relationship between the standard deviation of the measured angles and the flexibility at the base-pair level. This method is useful to quantify directly from microscopy techniques, such as electron or scanning force microscopy, the true bending angle, either intrinsic or induced by a ligand, and its associated flexibility at a given locus in any small DNA fragment.

High resolution mapping DNAs by R-loop atomic force microscopy.

R-loops formed by short RNA transcripts have been imaged by atomic force microscopy (AFM) at a constant force in the height mode. The technique was applied to mapping the human endogenous retrovirus K10 family (HERV-K10) long terminal repeats (LTR) within individual plasmids and cosmids. RNA probes specific for the U3 (384 nt) and U5 (375 nt) LTR regions separated by a span of 200 bp were used for R-loop formation with LTRs located within plasmid (3.8 kb) or cosmid ( approximately 40 kb) DNAs. R-loops stabilized by glyoxal treatment and adsorbed onto the mica surface in the presence of magnesium ions looked like looped out segments of RNA:DNA hybrids. The total yield of R-loops was usually approximately 95%. The RNA:DNA hybrids were found to be 12-15% shorter than the corresponding DNA:DNA duplex. The two regions of the LTR could be easily discerned in the AFM images as clearly separated loops. R-loop positions determined on cosmids by AFM were accurate to approximately 0.5% of the cosmid length. This technique might be easily adapted for mapping various sequences such as gene exons or regulatory regions and for detecting insertions, deletions and rearrangements that cause human genetic diseases.

Analysis of various sequence-specific triplexes by electron and atomic force microscopies.

Sequence-specific interactions of 20-mer G,A-containing triple helix-forming oligonucleotides (TFOs) and bis-PNAs (peptide nucleic acids) with double-stranded DNA was visualized by electron (EM) and atomic force (AFM) microscopies. Triplexes formed by biotinylated TFOs are easily detected by both EM and AFM in which streptavidin is a marker. AFM images of the unlabeled triplex within a long plasmid DNA show a approximately 0.4-nm height increment of the double helix within the target site position. TFOs conjugated to a 74-nt-long oligonucleotide forming a 33-bp-long hairpin form extremely stable triplexes with the target site that are readily imaged by both EM and AFM as protruding DNA. The short duplex protrudes in a perpendicular direction relative to the double helix axis, either in the plane of the support or out of it. In the latter case, the apparent height of the protrusion is approximately 1.5 nm, when that of the triplex site is increased by 0.3-0.4 nm. Triplex formation by bis-PNA, in which two decamers of PNA are connected via a flexible linker, causes deformations of the double helix at the target site, which is readily detected as kinks by both EM and AFM. Moreover, AFM shows that these kinks are often accompanied by an increase in the DNA apparent height of approximately 35%. This work shows the first direct visualization of sequence-specific interaction of TFOs and PNAs, with their target sequences within long plasmid DNAs, through the measurements of the apparent height of the DNA double helix by AFM.

Recruitment of transcription factors to the target site by triplex-forming oligonucleotides.

Triplex-forming oligonucleotides (TFOs) are generally designed to inhibit transcription or DNA replication but can be used for more diverse purposes. Here we have designed a hairpin-TFO able to recruit transcription factors to a target DNA. The designed oligonucleotide contains a triplex-forming sequence, linked through a nucleotide loop to a double-stranded hairpin including the SRE enhancer of the c-fos gene promoter. We show here that this oligonucleotide can specifically recognise its DNA target at physiological salt and pH conditions. The stability of the triplex formed under these conditions is very high: >90% of the triplex remains intact after 24 h of incubation. Bound to the double-stranded target DNA, the oligonucleotide retains its ability to interact specifically with transcription factors, recruiting them to the proximity of the target DNA. Our results suggest that this type of oligonucleotide may prove useful in the design of new tools for artificial modulation of gene expression.

Recombination-dependent repair of DNA double-strand breaks with purified proteins from Escherichia coli.

We have developed an in vitro system in which repair of DNA double-strand breaks is performed by purified proteins of Escherichia coli. A segment was deleted from a circular duplex DNA molecule by restriction at two sites. 3' single-stranded overhangs were introduced at both ends of the remaining linear fragment. In a first step, RecA protein catalyzed the formation of a D-loop between one single-stranded tail and a homologous undeleted supercoiled DNA molecule. In a second step, E. coli DNA polymerase II or III used the 3' end in the D-loop as a primer to copy the missing sequences of the linear substrate on one strand of the supercoiled template. Under proper conditions, the integrity of the deleted substrate was restored, as shown by analysis of the products by electrophoresis, restriction, and transformation. In this reaction, DNA synthesis is strictly dependent on recombination, and repair is achieved without formation of a Holliday junction.

A new approach to overcome potassium-mediated inhibition of triplex formation.

G,A-containing purine oligonucleotides of various lengths form extremely stable and specific triplexes with the purine-pyrimidine stretch of the vpx gene [Svinarchuk,F., Monnot,M., Merle,A., Malvy,C. and Fermandjian,S. (1995) Nucleic Acids Res., 22, 3742--3747]. The potential application of triple-helix-forming oligonucleotides (TFO) in gene-targeted therapy has prompted us to study triplex formation mimicking potassium concentrations and temperatures in cells. Triplex formation was tested by dimethyl sulphate (DMS) footprinting, gel-retardation, UV melting studies and electron microscopy. In the presence of 10 mM MgCl2, KCl concentrations up to 150 mM significantly lowered both efficiency (triplex : initial duplex) and rate constants of triplex formation. The KCl effect was more pronounced for 11mer and 20mer TFOs than for 14mer TFO. Since the dissociation half-life for the 11mer TFO decreases from 420 min in the absence of monovalent cations to 40 min in the presence of 150 mM KCI, we suggest that the negative effect could be explained by a decrease in triplex stability. In contrast, for the 20mer TFO no dissociation of the triplex was observed during 24 h of incubation either in the absence of monovalent cations or in the presence of 150 mM KCl. We suppose that in the case of the 20mer TFO the negative effect of KCI on triplex formation is probably due to the self-association of the oligonucleotide in competitive structures such as parallel duplexes and/or tetraplexes. This negative effect may be overcome by the prior formation of a short duplex either on the 3'- or 5'-end of the 20mer TFO. We refer to these partial duplexes as 'zipper' TFOs. It was demonstrated that a 'zipper' TFO can form a triplex over the full length of the target, thus unzipping the short complementary strand. The minimal single-stranded part of the 'zipper' oligonucleotide which is sufficient to initiate triplex formation can be as short as three nucleotides at the 3'-end and six nucleotides at the 5'-end. We suggest that this type of structure may prove useful for in vivo applications.

Alternate strand DNA triple helix-mediated inhibition of HIV-1 U5 long terminal repeat integration in vitro.

Integration of the human immunodeficiency virus (HIV) DNA into the host genome is an obligatory process in the replicative life cycle of the virus. This event is mediated in vitro by integrase, a viral protein which binds to specific sequences located on both extremities of the DNA long terminal repeats (LTRs). These sites are highly conserved in all HIV genomes and thus provide potential targets for the selective inhibition of integration. The integrase-binding site located on the HIV-1 U5 LTR end contains two adjacent purine tracts on opposite strands, 5' . . . GGAAAATCTCT-3'/3'-CCTTTTAGAGA . . . 5', in parallel orientations. A single strand oligonucleotide 5'-GGTTTTTGTGT-3' was designed to associate with these tracts via its ability to form a continuous alternate strand DNA triplex. Under neutral pH and physiological temperature, the oligonucleotide, tagged with an intercalator chromophore oxazolopyridocarbazole, formed a stable triplex with the target DNA. The occurrence of this unusual triplex was demonstrated by both DNase I footprinting and electron microscopy. The triplex inhibits the two steps of the integrase-mediated reactions, namely, the endonucleolytic cleavage of the dinucleotide 5'-GT-3' from the 3' end of the integration substrate and the integration of the substrate into the heterologous target DNA. The midpoints for both inhibition reactions were observed at oligonucleotide concentrations of 50-100 nM. We believe that these results open new possibilities for the specific targeting of viral DNA LTR ends with the view of inhibiting integration under physiological conditions.

Electron microscopy mapping of oligopurine tracts in duplex DNA by peptide nucleic acid targeting.

Biotinylated homopyrimidine decamer peptide nucleic acids (PNAs) are shown to form sequence-specific and stable complexes with complementary oligopurine targets in linear double-stranded DNA. The noncovalent complexes are visualized by electron microscopy (EM) without chemical fixation using streptavidin as an EM marker. The triplex stoichiometry of the PNA-DNA complexes (two PNA molecules presumably binding by Watson-Crick and Hoogsteen pairing with one of the strands of the duplex DNA) is indicated by the appearance of two streptavidin 'beads' per target site in some micrographs, and is also supported by the formation of two retardation bands in a gel shift assay. Quantitative analysis of the positions of the streptavidin 'beads' revealed that under optimized conditions PNA-DNA complexes are preferably formed with the fully complementary target. An increase in either the PNA concentration or the incubation time leads to binding at sites containing one or two mismatches. Our results demonstrate that biotinylated PNAs can be used for EM mapping of short targets in duplex DNA.

Testing the quality of electron microscope mapping data for DNA molecules with sequence-specific ligands.

A procedure for the testing of Electron Microscope (EM) mapping data for DNA molecules with site-specific bound ligands is suggested. The difficulty of distinguishing DNA molecule ends on electron micrographs indicates that their true orientations are not known. This in turn presents problems in obtaining correct maps relating to their alignment, and complicates checking the maps' validity. For these reasons a computer simulation of the EM study of double-stranded DNA molecules with site-specific bound ligands was carried out. The knowledge of the true orientations of the simulated DNA molecules allowed us to examine their final orientations after alignment. We used the number of improper-oriented molecules as the quantitative measure of the map quality. Detailed investigation based on this parameter permitted us to invent the criterion for the map validity, and to suggest the procedure for the testing of alignment of real DNA molecules. This procedure implies multiple randomization of initial orientations of the DNA molecules and minute analysis of the final maps. Most of the molecular, statistical and experimental parameters inherent to EM investigation of site-specific binding, such as the number of specific binding sites (N), the mean number of bound ligands (A), the length of the DNA molecules (L), the specific/non-specific ratio of binding (K), together with the standard deviation of DNA molecule lengths (HL) were tested for their influence upon the quality of EM mapping data. An empirical equation for the ultimate values of these parameters has been found, allowing us to predict the success of EM mapping.

Electron microscopy visualization of oligonucleotide binding to duplex DNA via triplex formation.

Using biotinylated oligonucleotides and streptavidin as a marker, we have visualized, with the help of electron microscopy, the triplex formation. We used the natural homopurine-homopyrimidine sequence from human papillomavirus 16 cloned within a plasmid. Under conditions favouring the formation of pyrimidine-purine-pyrimidine triplex the corresponding pyrimidine oligonucleotide formed a complex with the insert and this complex was detected by electron microscopy. Similarly, under conditions favouring the formation of pyrimidine-purine-purine triplex the corresponding purine oligonucleotide formed a stable complex detected by electron microscopy. In both cases the complexes we observed exhibited remarkable sequence specificity. Near 80% of DNA molecules carried the streptavidin marker in the correct position and very few cases of non-specific binding were detected. We conclude that the triplex mode of recognition may provide very efficient sequence-specific markers for electron microscopy of DNA.

DNA unwinding upon strand-displacement binding of a thymine-substituted polyamide to double-stranded DNA.

It was recently found that polyamide nucleic acid (PNA) analogues consisting of thymines attached to an aminoethylglycine backbone bind strongly and sequence-selectively to adenine sequences of oligonucleotides and double-stranded DNA [Nielsen, P. E., Egholm, M., Berg, R. H. & Buchardt, O. (1991) Science 254, 1497-1500]. It was concluded that the binding to double-stranded DNA was accomplished via strand displacement, in which the PNA bound to the Watson-Crick complementary adenine-containing strand, whereas the thymine-containing strand was extruded in a virtually single-stranded conformation. This model may provide a general way in which to obtain sequence-specific recognition of any sequence in double-stranded DNA by Watson-Crick hydrogen-bonding base-pair recognition, and it is thus paramount to rigorously establish this binding mode for synthetic DNA-binding ligands. We now report such results from electron microscopy. Furthermore, we show that binding of PNA to closed circular DNA results in unwinding of the double helix corresponding to approximately one turn of the double helix per 10 base pairs. The DNA.PNA complex, which is formed at low salt concentration (only a small portion of DNA molecules show complex formation at NaCl concentration higher than 40 mM), is exceptionally kinetically stable and cannot be dissociated by increasing salt concentration up to 500 mM.