Tetraplex formation of a guanine-containing nonameric DNA fragment

Solution structure of the human telomeric repeat d[AG3(T2AG3)3] G-tetraplex

BACKGROUND

Repeats of Gn sequences are detected as single strand overhangs at the ends of eukaryotic chromosomes together with associated binding proteins. Such telomere sequences have been implicated in the replication and maintenance of chromosomal termini. They may also mediate chromosomal organization and association during meiosis and mitosis.

RESULTS

We have determined the three-dimensional solution structure of the human telomere sequence, d[AG3(T2AG3)3] in Na(+)-containing solution using a combined NMR, distance geometry and molecular dynamics approach (including relaxation matrix refinement). The sequence, which contains four AG3 repeats, folds intramolecularly into a G-tetraplex stabilized by three stacked G-tetrads which are connected by two lateral loops and a central diagonal loop. Of the four grooves that are formed, one is wide, two are of medium width and one is narrow. The alignment of adjacent G-G-G segments in parallel generates the two grooves of medium width whilst the antiparallel arrangement results in one wide and one narrow groove. Three of the four adenines stack on top of adjacent G-tetrads while the majority of the thymines sample multiple conformations.

CONCLUSIONS

The availability of the d[AG3(T2AG3)3] solution structure containing four AG3 human telomeric repeats should permit the rational design of ligands that recognize and bind with specificity and affinity to the individual grooves of the G-tetraplex, as well as to either end containing the diagonal and lateral loops. Such ligands could modulate the equilibrium between folded G-tetraplex structures and their unfolded extended counterparts.

Conformational polymorphism in G-tetraplex structures: strand reversal by base flipover or sugar flipover

Guanine rich sequences adopt a variety of four stranded structures, which differ in strand orientation and conformation about the glycosidic bond even though they are all stabilised by Hoogsteen hydrogen bonded guanine tetrads. Detailed model building and molecular mechanics calculations have been carried out to investigate various possible conformations of guanines along a strand and different possible orientations of guanine strands in a G-tetraplex structure. It is found that for an oligo G stretch per se, a parallel four stranded structure with all guanines in anti conformation is favoured over other possible tetraplex structures. Hence an alternating syn-anti arrangement of guanines along a strand is likely to occur only in folded back tetraplex structures with antiparallel G strands. Our study provides a theoretical rationale for the observed alternation of glycosidic conformation and the inverted stacking arrangement arising from base flipover, in antiparallel G-tetraplex structures and also highlights the various structural features arising due to different types of strand orientations. The molecular mechanics calculations help in elucidating the various interactions which stabilize different G-tetraplex structures and indicate that screening of phosphate charge by counterions could have a dramatic effect on groove width in these four stranded structures.

Thrombin-binding DNA aptamer forms a unimolecular quadruplex structure in solution

We have used two-dimensional 1H NMR spectroscopy to study the conformation of the thrombin-binding aptamer d(GGTTGGTGTGGTTGG) in solution. This is one of a series of thrombin-binding DNA aptamers with a consensus 15-base sequence that was recently isolated and shown to inhibit thrombin-catalyzed fibrin clot formation in vitro [Bock, L. C., Griffin, L. C., Latham, J. A., Vermaas, E. H. & Toole, J. J. (1992) Nature (London) 355, 564-566]. The oligonucleotide forms a unimolecular DNA quadruplex consisting of two G-quartets connected by two TT loops and one TGT loop. A potential T.T bp is formed between the two TT loops across the diagonal of the top G-quartet. Thus, all of the invariant bases in the consensus sequence are base-paired. This aptamer structure was determined by NMR and illustrates that this molecule forms a specific folded structure. Knowledge of this structure may be used in the further development of oligonucleotide-based thrombin inhibitors.

Far-infrared study of the vibrational modes of 5'-GMP gels and crystals of Na+ and K+

Five Far-Infrared (50-600 cm-1) spectra are presented: the sodium and potassium salts of 5' Guanosine Monophosphate (GMP), each salt in both the gel and crystal conformations, and poly(rG). Measurements were performed at a sample temperature of 10 K under vacuum with a liquid He-cooled bolometer. The spectra were fit with Lorentzians and assignments are suggested. There are noteworthy differences in oscillator strengths and frequencies of the bands between all spectra. We report the tentative observation of a 100 cm-1 mode which is in the neighborhood of a mode observed by Raman spectroscopy in solution (1) and dried gels (2).

Conformational isomerizations of poly(dA-dT) are dramatically influenced by a substitution of a minor amount of adenine by purine or amino2purine

We have synthesized poly(dA,dPu-dT) and poly(dA,n2dPu-dT) containing, respectively, 5.7% of purine and 7.4% of amino2purine in place of adenine to demonstrate that these apparently negligible perturbations of the primary structure have dramatic consequences for the polynucleotide conformational isomerizations. The replacement of adenine by amino2purine, preserving the number of hydrogen bonds between the complementary bases, has a stronger effect on the polynucleotide conformational isomerizations than the replacement with purine that is bound only by a single hydrogen bond to thymine. Nevertheless, poly(dA,dPu-dT) forms a more thermostable duplex than poly(dA,n2dPu-dT). Furthermore the few amino2purines in poly(dA,n2dPu-dT) inhibit its isomerization into X-DNA, stabilize but modify A-DNA and stabilize Z-DNA. Kinetics of the B-Z transition of poly(dA,n2dPu-dT) is fast to indicate that the amino groups in the double helix minor groove substantially decrease the kinetic barrier between B- and Z-DNA. On the other hand, the replacement of adenine by purine destabilizes both Z-DNA and A-DNA, and the destabilization of X-DNA is weaker than with amino2purine. A-form and B-form perhaps coexist in poly(dA,dPu-dT) at high concentrations of ethanol.

Quadruplex structure of d(G3T4G3) stabilized by K+ or Na+ is an asymmetric hairpin dimer

The ends of chromosomes contain repeats of guanine-rich sequences that can assume highly compact conformations and are presumed necessary for their biological role in chromosomal stabilization and association. We have investigated the conformational behavior of d(G3T4G3) as a function of the addition of either KCl or NaCl, in the concentration range of 50-200 mM, by using a spectrum of physical techniques and conclude that these salts induce a quadruplex species composed of two strands, each in a hairpin conformation. When salt is added, a large positive signal appears near 290 nm in the CD spectra. UV thermal denaturation curves show a single concentration-dependent transition and provide data for quantitating the thermodynamics of quadruplex formation. In electrophoresis experiments, the quadruplex structure migrates as a single species and more rapidly than the unstructured single strand. NMR spectra in the presence of KCl or NaCl indicate that the structure formed is asymmetric. Equilibrium ultracentrifugation studies confirm that these quadruplexes are composed of two strands of d(G3T4G3). Possible models for this structure are discussed.

Thermodynamics and structure of a DNA tetraplex: a spectroscopic and calorimetric study of the tetramolecular complexes of d(TG3T) and d(TG3T2G3T)

We report a combined thermodynamic and structural characterization of a DNA tetraplex. Using spectroscopic and calorimetric techniques, we demonstrate that d(TG3T) and d(TG3T2G3T), in the presence of K+, form stable tetramolecular complexes. From differential scanning calorimetry measurements, we obtain the following thermodynamic profiles for formation of each tetraplex at 25 degrees C: delta G degrees = -6.9 kcal/mol of tetraplex (or -2.3 kcal/mol of tetrad; 1 cal = 4.184 J), delta H degrees = -62.6 kcal/mol of tetraplex (or -20.9 kcal/mol of tetrad), and delta S degrees = -186.9 cal.K-1.mol-1 of tetraplex (or -62.3 cal.K-1.mol-1 of tetrad) for the d(TG3T) tetraplex; and delta G degrees = -20.2 kcal/mol of tetraplex (or -3.4 kcal/mol of tetrad), delta H degrees = -123.2 kcal/mol of tetraplex (or -20.5 kcal/mol of tetrad), and delta S degrees = -346.0 cal.K-1.mol-1 of tetraplex (or -57.7 cal.K-1.mol-1 of tetrad) for the d(TG3T2G3T) tetraplex. These data demonstrate that at 25 degrees C a G-tetrad can exhibit considerable stability, comparable to or even exceeding that of most Watson-Crick nearest-neighbor interactions, with this stability resulting from a very favorable enthalpy of formation. Temperature-dependent CD measurements reveal that the melting temperatures of both tetraplexes exhibit unusually low salt dependences. This unexpected behavior may reflect a diminished charge density due to bound K+ ions. For each complex, the Na+ and K+ forms exhibit drastically different isothermal and temperature-dependent CD profiles, with the K+ forms of each tetraplex melting more sharply and at a higher temperature than the Na+ forms. Using one- and two-dimensional NMR techniques, we show that the strands in the tetramolecular complex of d(TG3T), K+ are all parallel and that the guanine glycosidic conformations are all anti.

Guanine residues in d(T2AG3) and d(T2G4) form parallel-stranded potassium cation stabilized G-quadruplexes with anti glycosidic torsion angles in solution

We report below on proton NMR studies of the G-quadruplex structure formed by the human telomere sequence d(T2AG3) and the tetrahymena telomere sequence d(T2G4) in K cation containing solution. We observe well-resolved proton NMR spectra corresponding to a G-quadruplex monomer conformation predominant at 50 mM K cation concentration and a G-quadruplex dimer conformation predominant at 300 mM K cation concentration. By contrast, d(T2AG3T) and d(T2G4T) form only the G-quadruplex monomer structures independent of K cation concentration as reported previously [Sen, D., & Gilbert, W. (1992) Biochemistry 31, 65-70]. We detect well-resolved resonances for the exchangeable guanine imino and amino protons involved in G-tetrad formation with the hydrogen-bonded and exposed amino protons separated by up to 3.5 ppm. The observed NOEs between the amino and H8 protons on adjacent guanines within individual G-tetrads support the Hoogsteen pairing alignment around the tetrad. The imino protons of the internal G-tetrads exchange very slowly with solvent H2O in the d(T2AG3) and d(T2G4) quadruplexes. The nature and intensity of the observed NOE patterns establish formation of parallel-stranded right-handed G-quadruplexes with all anti guanine glycosidic torsion angles. A model for the parallel-stranded G-quadruplex is proposed which is consistent with the experimental NOE data on the d(T2AG3) and d(T2G4) quadruplexes in solution.(ABSTRACT TRUNCATED AT 250 WORDS)

Hairpin and parallel quartet structures for telomeric sequences

The role of thymine residues in the formation of G-quartet structures for telomeric sequences has been investigated using model oligonucleotides of the type d(G4TnG4), with n = 1-4. Sequences d(G4T3G4) and d(G4T4G4) adopt a G-quartet structure formed by hairpin dimerization in 70 mM NaCl as judged by a characteristic circular dichroism signature with a 295 nm positive and 265 nm negative bands while d(G4TG4) adopts a parallel G-quartet structure like d(G12) which exhibits a strong positive band at 260 nm and a negative band at 240 nm. The sequence d(G4T2G4) exhibits a mixture of both conformations. The stability of hairpin G-quartet structures decreases with decrease in the number of intervening thymine residues. Potassium permanganate, a single strand specific probe has been used to establish the presence of loops composed of T residues in the hairpin G quartet structures formed by the oligonucleotides d(G4TnG4) with n = 2-4 in 70 mM NaCl. The formation of hairpin G quartet structure for the above sequences is further supported by the enhanced electrophoretic mobility observed on non-denaturing polyacrylamide gels. Human telomeric sequence d(TTAGGG)4 which showed enhanced electrophoretic mobility like Tetrahymena telomeric sequence d(T2G4)4 also exhibited a characteristic CD spectrum for a folded-back G-quartet structure. A detailed model for G-quartet structure involving hairpin dimer with alternating syn-anti-syn-anti conformation for the guanine residues both along the chain as well as around the G tetrad with at least two thymine residues in the loop is proposed. Intermolecular association of short telomeric sequences reported here provides a possible model for chromosomal pairing.

Tetraplex formation of d(GGGGGTTTTT): 1H NMR study in solution

The oligonucleotide d(G5T5) can in principle form a fully matched duplex with G.T pairing and/or a tetraplex. Non-denaturing gel electrophoresis, circular dichroism and NMR experiments show that the tetraplex is exclusively formed by this oligomer in solution. In the presence of its complementary strand d(A5C5) at low temperature, d(G5T5) forms the tetraplex over the normally expected Watson-Crick duplex. However, when d(G5T5) and d(A5C5) are mixed together in equimolar amounts and heated for several minutes at 85 degrees C, and then allowed to cool, the product was essentially the Watson-Crick duplex. The lack of resolution in the 500 MHz 1H NMR spectra and the presence of extensive spin diffusion do not allow us to derive a quantitative structure for the tetraplex from the NMR data. However, we find good qualitative agreement between the NOESY and MINSY data and a theoretically derived stereochemically sound structure in which the G's and T's are part of a parallel tetraplex.

Lanthanide ion luminescence as a probe of DNA structure. 1. Guanine-containing oligomers and nucleotides

Laser-induced Eu(3+) luminescence spectroscopy is used to probe the interaction of Eu(3+) ion with guanine-containing nucleotides and single-stranded oligomers. By using time-resolved and non-time-resolved Eu(3+) luminescence techniques, two classes of Eu(3+) binding site are observed in oligo(dG)10, oligo(dG)8, oligo(dG)6, oligo(dG)4, and d-GMP. One class of site binds Eu(3+) ions more strongly than the other. Since the "tight" class of bound Eu(3+) ions have two coordinated water molecules, it is inferred that six or seven atoms from the oligomers are coordinating the Eu(3+). The "weaker" class of Eu(3+) ion sites involve the coordination of six or seven water molecules and therefore, are coordinated by one or two atoms from the oligomer. The tight class of Eu(3+) binding site is attributed to an interstrand association of Eu(3+) with the oligomers forming dimeric or polymeric structures. The dissociation constants (Kd) for the 1:1 complexes Eu(d-GMP)+ and Eu(d-GTP)- have been determined as well as the Kd for the dimerization reaction of Eu(d-GMP)+. The Tb(3+) luminescence enhancement properties of these molecules are also examined in relation to their EU(3+) binding characteristics.

Nucleic acid specificity of a vertebrate telomere-binding protein: evidence for G-G base pair recognition at the core-binding site

A factor from avian cells formed complexes with telomeric sequences and other single-stranded probes that contained tracts of guanine residues. Nucleoprotein complexes with telomere probes required two or more of the telomeric repeats that were incapable of Watson-Crick base-pairing. Methylation interference and protection experiments identified guanine N7 residues that were critical for the formation of the nucleoprotein complex and for the formation of a higher-order structure that occurred in the absence of the protein. Substitutions of deoxyinosine (dI) for deoxyguanosine (dG) demonstrated that the exocyclic N2 amino groups in the internal telomeric repeat, but not the terminal repeat, were required for the formation of the chemically protected structure and for protein binding. On the basis of these data we propose that the factor specifically recognizes a hairpin DNA structure that is stabilized by intramolecular G-G base-pairing between the telomere repeats. The positions of the critical guanine N2 and N7 groups indicate a G-G base-pairing configuration, where guanines function as hydrogen bond donors at the internal telomeric repeat and hydrogen bond acceptors at the terminal telomeric repeat.

Quadruplex structure of Oxytricha telomeric DNA oligonucleotides

The telomeres of most eukaryotes contain a repeating G-rich sequence with the consensus d(T/A)1-4G1-8, of which 12-16 bases form a 3' single-strand overhang beyond the telomeric duplex. It has been proposed that these G-rich oligonucleotides associate to form four-stranded structures from one, two or four individual strands and that these structures may be relevant in vivo. The proposed structures contain Hoogsteen base-paired G-quartets, precedent for which has been in the literature for many years. Here we use 1H NMR spectroscopy to study the conformations of the DNA oligonucleotides d(G4T4G4) (Oxy-1.5) and d(G4T4G4T4G4T4G4) (Oxy-3.5) which contain the Oxytricha telomere repeat (T4G4). We find that these molecules fold to form a symmetrical bimolecular and an intramolecular quadruplex, respectively. Both structures have four G-quartets formed from nucleotides that are alternately syn and anti along each strand. This arrangement differs from earlier models in which the strands are alternately all syn or all anti. The T4 loops in Oxy-1.5 are on opposite ends of the quadruplex and loop diagonally across the G-quartet, resulting in adjacent strands being alternately parallel and antiparallel.

Retinoblastoma susceptibility genes contain 5' sequences with a high propensity to form guanine-tetrad structures

Retinoblastoma susceptibility genes contain significant runs of oligoguanine at their 5' ends. Oligonucleotides having these sequences underwent complex formation in the presence of sodium ions, in which there was association of four strands. Formation of this structure was completely prevented if guanine was replaced by 7-deazaguanine, indicating the importance of guanine N7 in the formation of the complex. Complex formation lead to protection of guanine N7 against methylation by dimethyl sulphate, but thymine bases located between oligoguanine blocks were reactive to osmium tetroxide. There was also some sensitivity to S1 nuclease to the 5' side of the oligoguanine block. The results show that the G-rich regions of the mouse and human retinoblastoma susceptibility genes have a propensity to undergo tetraplex formation of the kind demonstrated in the immunoglobulin switch region.

Multinuclear nuclear magnetic resonance studies of Na cation-stabilized complex formed by d(G-G-T-T-T-T-C-G-G) in solution. Implications for G-tetrad structures

There has been much recent interest in the self-association of short deoxyguanosine-rich motifs within single-stranded DNAs to generate monovalent cation modulated four-stranded helical segments called G-quadruplexes stabilized by hydrogen-bonded G-tetrad alignments. We have addressed structural aspects of this novel alignment and report on multinuclear 1H, 31P and 13C nuclear magnetic resonance studies on the d(G2T4CG2) deoxynonanucleotide with Na cation as counterion in aqueous solution at low temperature. This sequence forms stable structures even though it cannot align by Watson-Crick hydrogen bond formation (see the paper on d(G2T5G2) describing optical and calorimetric measurements by Jin, R., Breslauer, K. J., Jones, R. A. & Gaffney, B. L. (1990), Science, 250, 543-546). The four narrow exchangeable protons detected between 11.5 and 12.0 parts per million (p.p.m.), which are common to the d(G2T4CG2) deoxynonanucleotide and the d(G2TCG2) deoxyhexanucleotide sequences, are assigned to deoxyguanosine imino protons hydrogen-bonded to carbonyl acceptor groups. These narrow imino protons are not detected for d(IGN5IG) and d(I2N5G2), where two deoxyguanosine residues are replaced by two deoxyinosine residues in the deoxynonanucleotide sequences. This implies that the 2-amino protons of deoxyguanosine must also participate in hydrogen bond formation and stabilize the structured conformation of d(G2T4CG2) in Na cation-containing solution. We have completely assigned the base and sugar H1', H2',2'', H3', and H4' protons of the d(G2T4CG2) oligomer following analysis of two-dimensional nuclear Overhauser enhancement spectroscopy and two-dimensional correlated spectroscopy data sets in 0.1 M-NaCl, 10 mM-sodium phosphate, 2H2O solution at 0 degree C. The relative magnitude of the nuclear Overhauser enhancements (NOEs) between the base H8 and its own sugar H1' protons of individual deoxyguanosine residues establishes that G1 and G8 adopt syn orientations while G2 and G9 adopt anti orientations about the glycosidic bond in the d(G1-G2-T3-T4-T5-T6-C7-G8-G9) sequence in both Na and K cation-containing aqueous solution. Consequently, any structure proposed for the tetramolecular complex of d(G2T4CG2) must exhibit alternating G(syn) and G(anti) glycosidic torsion angles within each strand. The directionality and magnitude of the observed NOEs are consistent with the G(syn)-G(anti) steps adopting right-handed helical conformations in solution. We also note that the H8 protons of G1 and G8 (7.35 to 7.45 p.p.m.) in a syn alignment are shifted significantly upfield from the H8 protons of G2 and G9 (8.0 to 8.3 p.p.m.) in an anti alignment.(ABSTRACT TRUNCATED AT 400 WORDS)

Characterization by 1H NMR of glycosidic conformations in the tetramolecular complex formed by d(GGTTTTTGG)

We have conducted two dimensional NOESY studies on the molecule d(G2T5G2) to characterize the structure of the tetramolecular complex previously identified by calorimetric and spectroscopic studies (1). Analysis of the NOE and exchange cross peaks observed in the NOESY spectra establishes the formation of structured conformations at low temperature (5 degrees C). Significantly, within each strand of these structured conformations, the G1 and G8 residues adopt syn glycosidic torsion angles, while the G2 and G9 residues adopt anti glycosidic torsion angles. Consequently, any structure proposed for the tetramolecular complex of d(G2T5G2) must have alternating G(syn) and G(anti) glycosidic torsion angles within each strand. The implications of this observation for potential structures of the tetramolecular complex of d(G2T5G2) are discussed.

Tetramerization of an RNA oligonucleotide containing a GGGG sequence

Poly rG can form four-stranded helices. The Hoogsteen-paired quartets of G residues on which such structures depend are so stable that they will form in 5'-GMP solutions, provided that Na+ or K+ are present (see for example, refs 2-4). Telomeric DNA sequences, which are G-rich, adopt four-stranded antiparallel G-quartet conformations in vitro, and parallel tetramerization of G-rich sequences may be involved in meiosis. Here we show that RNAs containing short runs of Gs can also tetramerize. A 19-base oligonucleotide derived from the 5S RNA of Escherichia coli (strand III), 5'GCCGAUGGUAGUGUGGGGU3', forms a K(+)-stabilized tetrameric aggregate that depends on the G residues at its 3' end. This complex is so stable that it would be surprising if similar structures do not occur in nature.

The contribution of DNA single-stranded order to the thermodynamics of duplex formation

We report a direct determination of the thermodynamic contribution that DNA single-stranded order makes to DNA duplex formation. By using differential scanning calorimetry (DSC) and temperature-dependent UV absorbance spectroscopy, we have characterized thermodynamically the thermally induced disruption of the 13-mer duplex [d(CGCATGAGTACGC)].[d(GCGTACTCATGCG)] (henceforth called S1.S2) and its component single strands, [d(CGCATGAGTACGC)] (henceforth called S1) and [d(GCGTACTCATGCG)] (henceforth called S2). These spectroscopic and calorimetric measurements yield the following thermodynamic profiles at 25 degrees C: delta G degree = 20.0 kcal/mol, delta H degree = 117.0 kcal/mol, and delta S degree = 325.4 cal.degree-1.mol-1 for duplex melting of S1.S2; delta G degree = 0.45 kcal/mol, delta H degrees = 29.1 kcal/mol, and delta S degree = 96.1 cal.degree-1.mol-1 for single-strand melting of S1; delta G degree = 1.44 kcal/mol, delta H degree = 27.2 kcal/mol, and delta S degree = 86.4 cal.degree-1.mol-1 for single-strand melting of S2 (1 cal = 4.184 J). These data reveal that the two single-stranded structures S1 and S2 are only marginally stable at 25 degrees C, despite exhibiting rather substantial transition enthalpies. This behavior results from enthalpy and entropy contributions of similar magnitudes that compensate each other, thereby giving rise to relatively small free energies of stabilization for the single strands at 25 degrees C. By contrast, the S1.S2 duplex state is very stable at 25 degrees C since the favorable transition entropy associated with duplex disruption (325.4 cal.degree-1.mol-1) is more than compensated for by the extremely large duplex transition enthalpy (117.0 kcal/mol). We also measured directly an enthalpy change (delta H degree) of -56.4 kcal/mol for duplex formation at 25 degrees C using isothermal batch-mixing calorimetry. This duplex formation enthalpy of -56.4 kcal/mol at 25 degrees C is very different in magnitude from the duplex disruption enthalpy of 117.0 kcal/mol measured at 74 degrees C by DSC. Since the DSC measurement reveals the net transition heat capacity change to be close to zero, we interpret this large disparity between the enthalpies of duplex disruption and duplex formation as reflecting differences in the single-stranded structures at 25 degrees C (the initial states in the isothermal mixing experiment) and the single-stranded structures at approximately 80 degrees C (the final states in the DSC experiment). In fact, the enthalpy for duplex formation at 25 degree C (-56.4 kcal/mol) can be combined with the sum of the integral enthalpies requires to melt each single strand from 25 to 80 degree C (23.6 kcal/mol for S1 and 27.2 kcal/mol for S2) to calculate a delta H degree of -107.2 kcal/mol for the hypothetical process of duplex formation from "random-coil" "unstacked" single strands at 25 degree C. The magnitude of this predicted delta H degree value for duplex formation is in good agreement with the corresponding parameter we measure directly by DSC for duplex disruption (117.0 kcal/mol), thereby lending credence to our interpretation and analysis of the data. Thus, our results demonstrate that despite being only marginally stable at 25 degree C, single strands can exhibit intramolecular interactions that enthalpically poise them for duplex formation. For the duplex studied herein, prior to association at 25 degree C, the two complementary single strands already possess > 40% of the total enthalpy (50.8/117) that ultimately stabilizes the final duplex state. This feature of single-stranded structure near room temperature can reduce significantly the enthalpic driving force one might predict for duplex formation from nearest-neighbor data, since such data generally are derived from measurements in which the single strands are in their random-coil states. Consequently, potential contributions from single-stranded structure must be recognized and accounted for when designing hybridization experiments and when using isothermal titration and/or batch mixing techniques to study the formation of duplexes and higher-order DNA structures (e.g., triplexes, tetraplexes, etc.) from their component single strands.

Stability and structure of three-way DNA junctions containing unpaired nucleotides

Non-paired nucleotides stabilize the formation of three-way helical DNA junctions. Two or more unpaired nucleotides located in the junction region enable oligomers ten to fifteen nucleotides long to assemble, forming conformationally homogeneous junctions, as judged by native gel electrophoresis. The unpaired bases can be present on the same strand or on two different strands. Up to five extra bases on one strand have been tested and found to produce stable junctions. The formation of stable structures is favored by the presence of a divalent cation such as magnesium and by high monovalent salt concentration. The order-disorder transition of representative three-way junctions was monitored optically in the ultraviolet and analyzed to quantify thermodynamically the stabilization provided by unpaired bases in the junction region. We report the first measurements of the thermodynamics of adding an unpaired nucleotide to a nucleic acid three-way junction. We find that delta delta G degrees (37 degrees C) = +0.5 kcal/mol for increasing the number of unpaired adenosines from two to three. Three-way junctions having reporter arms 40 base-pairs long were also prepared. Each of the three reporter arms contained a unique restriction site 15 base-pairs from the junction. Asymmetric complexes produced by selectively cleaving each arm were analyzed on native gels. Cleavage of the double helical arm opposite the strand having the two extra adenosines resulted in a complex that migrated more slowly than complexes produced by cleavage at either of the other two arms. It is likely that the strand containing the unpaired adenosines is kinked at an acute angle, forming a Y-shaped, rather than a T-shaped junction.