Search

Oligonucleotide-directed triple helix formation in the purine motif involves the binding of guanine-rich oligonucleotides to duplex DNA. Although this approach has been proposed for in vivo gene inhibition, triple helix formation by guanine-rich oligonucleotides is severely inhibited by physiological concentrations of certain monovalent cations (M+), especially K+. To clarify the mechanism of this inhibition, electrophoretic gel mobility shift titrations were performed to analyze the formation and stability of a purine motif triple helix in the presence of M+ and to monitor oligonucleotide aggregation under these conditions. M+ inhibition of triplex formation exhibited a concentration and ionic radius dependence that correlates with the ability of M+ to stabilize guanine quartet structures. In the presence of inhibitory (M+), guanine-rich oligonucleotides formed aggregates having characteristics consistent with the involvement of guanine quartets. The inhibitory effects of K+ on triplex formation could not be reversed by addition of the physiological polyamines spermidine3+ or spermine4+. M+ reduced the equilibrium concentration of the triplex primarily by decreasing the rate of triplex formation, but M+ also caused a detectable increase in the rate of triplex dissociation. Together, these results suggest that triplex inhibition under physiological ionic conditions is caused by competing equilibria wherein guanine-rich oligonucle- otides form aggregates involving guanine quartets. Approaches to destabilizing aggregates of guanine- rich oligonucleotides under physiological conditions will be required before in vivo applications can be realistically considered.

DNA is often bent when complexed with proteins. Understanding the forces responsible for DNA bending would be of fundamental value in exploring the interplay of these macromolecules. A series of experiments was devised to test the hypothesis that proteins with cationic surfaces can induce substantial DNA bending by neutralizing phosphates on one DNA face. Repulsions between phosphates in the remaining anionic helix are predicted to result in an unbalanced compression force acting to deform the DNA toward the protein. This hypothesis is supported by the results of electrophoretic experiments in which DNA spontaneously bends when one helical face is partially modified by incorporation of neutral phosphate analogs. Phasing with respect to a site of intrinsic DNA curvature (hexadeoxyadenylate tract) permits estimation of the electrostatic bend angle, and demonstrates that such modified DNAs are deformed toward the neutralized surface, as predicted. Similar model systems may be useful in exploring the extent to which phosphate neutralization can account for DNA bending by particular proteins.

Similar purine/pyrimidine mirror repeat (PMR) DNA sequences have been identified in the 5'-flanking regions of the human cystic fibrosis transmembrane conductance regulator (hCFTR) and mucin (hMUC1) genes, and supercoiled (but not linearized) plasmids containing these promoter regions were previously shown to be sensitive to digestion by S1 nuclease. The PMR element derived from the hCFTR promoter region is now sub-cloned and characterized at nucleotide resolution with respect to its reactivity toward nucleases S1 and P1, and toward the chemical probes dimethyl sulfate, chloroacetaldehyde, diethylpyrocarbonate and osmium tetroxide. These probes confirm the presence, at pH 4.5 (but not at pH 7.1), of a non-B-DNA structure. This non-B-DNA structure is distinct from H-DNA, because enzymatic and chemical probing detect single-stranded character in the absence of a stable intramolecular triple helix or extruded purine strand.

Research concerning oligonucleotide-directed triple helix formation has mainly focused on the binding of DNA oligonucleotides to duplex DNA. The participation of RNA strands in triple helices is also of interest. For the pyrimidine motif (pyrimidine.purine.pyrimidine triplets), systematic substitution of RNA for DNA in one, two, or all three triplex strands has previously been reported. For the purine motif (purine.purine.pyrimidine triplets), studies have shown only that RNA cannot bind to duplex DNA. To extend this result, we created a DNA triple helix in the purine motif and systematically replaced one, two, or all three strands with RNA. In dramatic contrast to the general accommodation of RNA strands in the pyrimidine triple helix motif, a stable triplex forms in the purine motif only when all three of the substituent strands are DNA. The lack of triplex formation among any of the other seven possible strand combinations involving RNA suggests that: (i) duplex structures containing RNA cannot be targeted by DNA oligonucleotides in the purine motif; (ii) RNA strands cannot be employed to recognize duplex DNA in the purine motif; and (iii) RNA tertiary structures are likely to contain only isolated base triplets in the purine motif.

Duplex DNA recognition by oligonucleotide-directed triple helix formation is generally limited to homopurine target domains. Various approaches have been suggested for the relief of this constraint. Artificial DNA sequences have previously been used to show that adjacent homopurine domains on opposite DNA strands can be simultaneously recognized by oligonucleotide probes that switch triple helix recognition motifs between domains. Using assays of electrophoretic mobility and chemical protection, we have explored in detail whether such strategies are of benefit in designing high-affinity probes for a natural DNA sequence in the human p53 gene. This target site contains three adjacent, purine-rich domains on opposite DNA strands. Our results show that (i) a modest but statistically significant enhancement in affinity can be achieved for this sequence by designing an oligonucleotide that simultaneously recognizes all three purine domains, (ii) correction of a pyrimidine interruption in one purine domain does not dramatically alter this result, (iii) the relative energetic and structural contributions attributable to recognition of each purine domain can be assessed using probes with combinations of specific and nonspecific nucleotide sequences, and (iv) probe affinity is not correlated with the apparent number of base triplets for certain complexes. These data suggest that unfavorable free energy changes may be associated with alternation between triple helix motifs using existing strategies. In contrast to artificial DNA sequences optimized for this purpose, a substantial affinity enhancement was not observed using alternate strand DNA recognition at this natural target sequence. We therefore conclude that such enhancement is sequence dependent. Triple helix formation is being explored as an approach to site-specific duplex DNA recognition. The high specificity and stability of some triple-helical structures suggest an approach to gene-specific therapeutic and diagnostic appli- cations (Moser & Dervan, 1987; Le Doan et al., 1987; Cooney

The Williamson laboratory studies RNA structure and RNA-protein interactions using structural biology and biophysical approaches. Two areas of emphasis include understanding the structure and specific recognition of the RNA-protein complexes that regulate gene expression, and understanding the assembly of multicomponent complexes such as the ribosome.

We are interested in creating artificial gene repressors based on duplex DNA recognition by nucleic acids rather than polypeptides. An in vitro model system involving repression of bacteriophage T7 RNA polymerase initiation has been employed to demonstrate that certain DNA oligonucleotides can repress transcription by site-specific triple-helix formation at two kinds of homopurine operator sequences [Maher, L. J., III, (1992) Biochemistry 31, 7587-7594]. Recognition in the purine motif is based on antiparallel oligonucleotide binding (G.G.C and T.A.T triplets). Recognition in the pyrimidine motif is based on parallel oligonucleotide binding (C+.G.C and T.A.T base triplets). Using this system, we report that the concentration-dependence of repression by DNA oligonucleotides provides triple-helix inhibition constant (Ki) estimates of approximately 2 x 10(-7) M for both purine motif and pyrimidine motif DNA complexes. RNA oligonucleotides are shown to repress promoters overlapping pyrimidine motif operators (Ki = 6 x 10(-7) M), but not purine motif operators. Although competent to hybridize to complementary single strands, RNA oligonucleotides fail to bind the purine motif operator. Partial substitution of deoxyribose residues tends to rescue repressor activity by RNA oligonucleotides in the purine motif. These results suggest prospects for, and constraints on, natural and artificial RNA-based repressors.

Certain sequences of double-helical DNA can be recognized and tightly bound by oligonucleotides. The effects of such triple-helical structures on DNA binding proteins have been studied. Stabilities of DNA triple-helices at or near physiological conditions are sufficient to inhibit DNA binding proteins directed to overlapping sites. Such proteins include restriction endonucleases, methylases, transcription factors, and RNA polymerases. These and other results suggest that oligonucleotide-directed triple-helix formation could provide the basis for designing artificial gene repressors. The general question of whether biological systems employ RNA molecules for recognition and regulation of double-helical DNA is discussed.

Pyrimidine oligonucleotides recognize extended purine sequences in the major groove of double-helical DNA by triple-helix formation. The resulting local triple helices are relatively stable and can block DNA recognition by sequence-specific DNA binding proteins such as restriction endonucleases. Association and dissociation kinetics for the oligodeoxyribonucleotide 5'-CTCTTTCCTCTCTTTTTCCCC (bold C's indicate 5-methylcytosine residues) are now measured with a restriction endonuclease protection assay. When oligonucleotides are present in greater than 10-fold excess over the DNA target site, the binding reaction kinetics are pseudo first order in oligonucleotide concentration. Under our standard conditions (37 degrees C, 25 mM Tris-acetate, pH 6.8, 70 mM sodium chloride, 20 mM magnesium chloride, 0.4 mM spermine tetrahydrochloride, 10 mM beta-mercaptoethanol, 0.1 mg/mL bovine serum albumin) the value of the observed pseudo-first-order association rate constant, k_(2obs), is 1.8 x 10^3 +± 1.9 x 10^2 L.(mol of oligomer^-1·s^-1. Measurement of the dissociation rate constant yields an equilibrium dissociation constant of approximately 10 nM. Increasing sodium ion concentration slightly decreased the association rate, substantially increased the dissociation rate, and thereby reduced the equilibrium binding constant. This effect was reversible by increasing multivalent cation concentration, confirming the significant role of multivalent cations in oligonucleotide-directed triple-helix formation under these conditions. Finally, a small reduction in association rate, a large increase in dissociation rate, and a resulting reduction in the equilibrium binding constant were observed upon increasing the pH between 6.8 and 7.2.

Electrophoretic assays of intrinsic DNA shape and shape changes induced by ligand binding are extremely useful because of their convenience and simplicity. The development of calibrations and empirical quantitative relationships permits highly accurate measurement of DNA shape using electrophoresis. Many conventional analyses employ the unidirectional ligation of short DNA duplexes. However, many oligonucleotides (typically more than 20) must often be synthesized for a single experiment. Additionally, the length of the DNA duplex can become limiting, preventing the analysis of certain DNA sequences. We now describe a semi-synthetic electrophoretic phasing method that offers several advantages, including a reduced number of required synthetic oligonucleotides, the ability to analyze longer DNA duplexes and a simplified approach for data analysis. We characterize semi-synthetic DNA probes in electrophoretic phasing assays by ligation of synthetic duplexes containing A(5) tracts between two longer restriction fragments. Upon electrophoresis, the gel mobility is strongly correlated with the predicted DNA curvature provided by the reference A(5) tracts. Having obtained this calibration, we show that the semi-synthetic phasing assay can be readily and economically applied to analyze DNA curvature induced by DNA charge modifications and DNA bending due to peptide binding.

Background: We have been studying the proposal that laterally asymmetric charge neutralization along the DNA double helix can induce collapse toward the neutralized surface. Results of previous experiments implied that such a phenomenon can occur, suggesting a role for local interphosphate repulsive forces in DNA shape and rigidity. Results: We now show that, whereas six ammonium ions tethered to one DNA face on flexible propyl chains can induce detectable DNA curvature, tethering of ammonium ions on rigid propynyl tethers does not induce DNA curvature. Molecular modeling indicates differing propensities for phosphate salt bridge formation between propyl- and propynyl-tethered ammonium ions. Conclusions: Ammonium ion localization is suggested as a key factor in induced bending. Rigidification of the double helix by stacking of propyne groups cannot be excluded.

Evidence is presented for the existence of teratocarcinoma transplantation antigens (Gt) encoded within the H-2 complex and present also on adult tissues. It has not been possible to separate these Gt loci from H-2 by recombination, and Gt factors map to each end of the H-2 complex. Previous reports indicating separation of all Gt loci from H-2 are reinterpreted. One class of such apparent recombinants has been shown to result from the outgrowth of tumor variants in mice of resistant genotype.

HMGB proteins are abundant non-histone proteins in eukaryotic chromatin. HMGB proteins are thought to act as architectural factors that enhance DNA bending and flexibility. The dynamic aspects of DNA bending by these proteins remain elusive. It has been proposed that these proteins associate and dissociate from DNA at fast rates. This highly dynamic behavior makes it difficult to study binding and conformational dynamics by traditional biochemical techniques. In this work we use single molecule fluorescence resonant energy transfer (smFRET) to study the binding dynamics of human HMGB2A and S. cerevisiae Nhp6A sequence non-specific single box proteins to DNA substrates. Studies were done using both total internal reflection and confocal excitation. By using a FRET pair with dyes at the ends of the DNA substrates we are able to follow the bending dynamics of the substrates at the single molecule level. We carried on experiments with duplex DNA 15 bp and 18 bp long to determine binding affinity and potential binding of multiple units to the substrate. DNA substrates containing bulges were used for comparison to study the effect of pre-bent structures on binding affinity and conformational dynamics. Bending of the DNA substrates is observed by changes in FRET efficiency allowing determining the conformational dynamics of the system in real time with temporal resolution in the order of milliseconds.

In this issue of Structure , Tesmer and coworkers report the structure of a small RNA aptamer in an inhibitory complex with a G protein receptor kinase. The aptamer adopts a charming structure that inserts an adenosine residue into the ATP binding pocket of the kinase.

DNA is a curious molecule. In vitro, DNA has been shown to be a rigid polymer resistant to bending and twisting over distances of

High-mobility group (HMG) proteins are DNA binding proteins believed to play a significant role in reorganizing the conformation of chromatin, facilitating transcription, replication and DNA repair. Both HMO1 and Nhp6A are Saccharomyces cerevisiae HMGB family sequence-nonspecific DNA binding proteins containing HMG box motifs that generate strong bends in DNA. We are studying the hypothesis that binding of HMGB proteins disrupts chromatin, possibly opening binding sites for other factors. To study this architectural function, we have observed nucleosome conformation in complementary single molecule experiments involving atomic force microscopy (AFM) and optical tweezers (OT). In AFM experiments, nucleosome arrays are reconstituted and probed in liquid on mica substrates. Images reveal tightly compacted nucleosomes, which are characterized by small inter-core particle distances. Increasing concentrations of HMGB proteins initially increase average core particle distance, then decrease it as the DNA is compacted. OT experiments unfold arrays of nucleosome core particles, probing structural stability. OT data show that forces of 10 – 20 pN are required to fully disrupt nucleosome core particles, while experiments in the presence of HMGB proteins show unfolding at lower forces (at and below 10 pN). Taken together, these results indicate HMGB proteins alter chromatin structure, possibly by disrupting, but not fully displacing, nucleosomes.