DIMACS Workshop on DNA Sequence and Topology

April 19-20, 2001
DIMACS Center, Rutgers University

Wilma K. Olson,, Rutgers University, olson@rutchem.rutgers.edu
Bernard D. Coleman, Rutgers University, bcoleman@stokes.rutgers.edu
Victor B. Zhurkin, NIH, zhurkin@nih.gov
Presented under the auspices of the Special Focus on Computational Molecular Biology.



Large-Scale Chromatin Structure and Dynamics

Andrew Belmont, University of Illinois

Over the past several years, we have developed a method for
labeling specific chromosome sites using the lac repressor/
operator interaction.  This method allows direct in vivo
visualization of chromosome dynamics.  In addition it can
be combined with gene amplification to create labeled,
amplified chromosome regions, ranging in size from chromosome
bands to entire chromosome arms, with homegeneous folding
patterns.  Using immunogold labeling, the folding motifs of
these specifically labeled chromosome regions can be analyzed
by electron microscopy reconstruction methods.  Here we review
some of the progress made using this technology.  We demonstrate
large-scale chromatic fibers, representing a level of chromatin
folding beyond the 300 nm chromatin fiber, within mammalian
interphase nuclei.  The compaction and orientation of these
fibers is generally stable over a period of hours.  Over a 
shorter time of seconds, rapid but constrained mobility is 
observed and over longer periods, a reproducible pattern of
conformational changes occurs during cell cycle progression,
including changes associated with DBNA replication initiation.
Despite the general, long-term immobility of chromosome regions,
several examples of long-range motion within the nucleus of 
specific chromosome regions have been observed which are linked
to distinct stages of the cell cycle.  Targeting of 
transcriptional activators to these labeled chromosome regions
results in dramatic unfolding of large-scale chromatin structure,
which may be related to targeting of HAT and remodeling complexes
to these chromosome regions; changes in intranuclear positioning
are also observed.  Mitotic chromosome condensation is accompanied
by highly reproducible by dynamic positioning of specific sequences
within the condensed chromosone.  EM reconstructions of labeled
chromosome regions of varying size during chromosome condensation
/decondensation are in progress to reveal the underlying folding
motifs for this large-scale chromatin structure.


Topology Driven Transitions in DNA Structure:
Their Mathematical Analysis and Roles in Biology

Craig J. Benham, Mount Sinai School of Medicine

This talk will briefly describe a formally exact mathematical
method for analyzing the topologically driven destabilization
of the DNA duplex.  Our method calcuated the statistical-
mechanical equilibrium distribution of a population of 
identical DNA molecules among all the available conformational
states.  The global character of the topological constraint
couples together the transition behaviors of all regions within
the domain.  When applied to the analysis of specific DNAs,
its results agree in a quantitatively precise manner with 
experimental measurements of the locations and extents of
local strand separations as functions of the imposed linking
deficiency.  This high degree of accuracy justifies its use
to predict the duplex destabilization properties of other
DNA base sequences, on which experiments have not been performed.

The sites of predicted duplex destabilization within natural DNA
sequences do not occur at random, but instead are closely
associated with specific types of DNA regulatory elements.
Examples include sites controlling the initiation and 
termination of gene expression, origins of replication,
and positions where DNA is attached to the nuclear scaffold.
This work has implicated topology induced DNA duplex
destabilization in the mechanisms of funtion of several
types of regulatory elements.  Examples will be presented
which illustrate how these analyses have helped elucidate
several mechanisms of DNA regulation, with emphasis on
transcriptional control.

3. DNA Sequence-Structure Relationships Christopher A. Hunter, University of Sheffield The structural properties of double helical DNA are sensitive to the base sequence, and this plays an important role in determining the way in which this molecule can be manipulated. Sequence-dependent structure is utilised by proteins for sequence specific DNA recognition and controls the positioning of DNA in nucleosomes. The molecular basis of sequence-structure relationships lies in the aromatic base stacking interactions. We have used a synthetic chemical system to develop a reliable computational model for these interactions. This model in conjuction with analyses of X-ray crystal structures of DNA oligomers has enabled us to produce a fast and accurate computational method for predicting the sequence-dependent structural properties of DNA. This presentation will outline the approach and illustrate the problems which can be tackled with these computational methods.
4. Probing the Architecture of the Crossover Site Synapsis Promoted by gd Resolvase. Andres E. Leschziner, Yale University, gd resolvase catalyzes the recombination between the two copies of the gd transposon that are found in a plasmid after transposition, generating two plasmids containing one transposon each [1]. Recombination takes place within an elaborate nucleoprotein complex containing twelve resolvase monomers plus two 120 bp DNA sites (the res sites) and has stringent topological requirements [1]. Crystal structures have been solved of resolvase both in the absence [2, 3] and presence [4] of DNA. Although a wealth of information was obtained from these structures, nothing was learned concerning the interaction of the resolvase dimers responsible for the actual strand exchange - those bound at site I. Ideally, to address this question one would like to be able to study site I in isolation, thus bypassing many of the complexities of the system. This has recently become possible thanks to a novel resolvase mutant ([5]; M. Boocock, personal communication) that is able to catalyze recombination between two isolated site I's in linear DNA. Given that the structure of a resolvase dimer bound to a site I is known, one can envision a system where the relative orientation of two site I's (and therefore that of the dimers bound to them) is fixed and can be set experimentally. In such a system, one could probe for recombination to test which specific arrangement of resolvase dimers is a favorable one. We achieved this by taking advantage of another crystal structure: that of the IHF-DNA complex [6]. IHF binds and bends DNA by about 180=B0, effectively creating a U-shaped DNA. In a linear piece of DNA containing two site I's and an IHF-binding site between them, the addition of IHF should bring the two site I's in close proximity. If the substrates are used at low enough concentration, intermolecular reactions should be eliminated. The relative orientation of the two site I-bound resolvase dimers can be adjusted by varying the number of base pairs between the site I's and the IHF-binding site. We generated a series of substrates spanning over a full B-DNA turn. The results from this series indicate that the favored orientation is one where the two dimers are aligned more or less perpendicularly to the plane of the IHF-induced bend. The data also suggest that the synapsis involves a crossing of the crossover sites (as opposed to a parallel arrangement of the two DNA=92s) but do not distinguish between the two possible crossings (a positive or a negative node) and therefore did not allow us to determine whether the two dimers interacted through their catalytic or DNA-binding domains. In order to address this question, we generated a substrate (of known topology) consisting of two catenated circles of 3 kb and 200 bp. Each circle contains a site I. Depending on the writhe of the crossover during the resolvase-catalyzed recombination, the two circles would be fused together in one of the two possible orientations and these could be easily probed by restriction digests. Reactions using this substrate showed a clear bias in favor of the product resulting from resolvase acting on a crossing over, corresponding to that having a negative writhe in the "IHF substrates" and indicating that the resolvase dimers bound at site I interact through their catalytic domains. 1. Grindley, N.D.F., Resolvase-mediated site-specific recombination, in Nucleic Acids and Molecular Biology, Eckstein, F. and Lilley, D.M.J., Editors. 1994, Springer-Verlag: Berlin. p. 236-267. 2. Rice, P.A. and T.A. Steitz, Model for a DNA-mediated synaptic complex suggested by crystal packing of gamma delta resolvase subunits. EMBO J, 1994. 13(7): p. 1514-24. 3. Sanderson, M.R., et al., The crystal structure of the catalytic domain of the site-specific recombination enzyme gamma delta resolvase at 2.7 A resolution. Cell, 1990. 63(6): p. 1323-9. 4. Yang, W. and T.A. Steitz, Crystal structure of the site-specific recombinase gamma delta resolvase complexed with a 34 bp cleavage site. Cell, 1995. 82(2): p. 193-207. 5. Arnold, P.H., et al., Mutants of Tn3 resolvase which do not require accessory binding sites for recombination activity. EMBO J, 1999. 18(5): p.1407-14. 6. Rice, P.A., et al., Crystal structure of an IHF-DNA complex: a protein-induced DNA U-turn. Cell, 1996. 87(7): p. 1295-306.
5. Ab Initio Quantum Mechanical Analysis of Nucleic Acid Components Alexander MacKerell, University of Maryland Obigonucleotide structure is dictated by a balance of intrinsic energetics and environmental contributions. Ab initio quantum mechanical (QM) calculations are an ideal tool to systematically investigate the intrinsic conformational energetic contributions to DNA and RNA structure. Approaches used and results from QM calculations on a variety of model compounds representative of nucleic acid moieties will be presented. Highlights include the relationship of the intrinsic conformational energetics of rotatable bonds in the phosphodiester backbone to the conformation of duplex DNA and RNA and the influence of cytosine on the A to B form DNA equilibrium.
6. Extracting Localized Properties of the TATA-box Via Continuum Modeling of Cyclization Experiments Robert Manning, Haverford College Continuum rod models have been used to extract long-scale properties of DNA, such as persistence length, from experimental measurements. On the other hand, many important features of DNA are quite localized; for example, the localized kinking of the TATA-box sequence is presumed to play a role in the initiation of transcription. We will present a hybrid mathematical model that inserts localized kinks and/or hinges into a standard continuum rod model, with the purpose of extracting quantitative measurements of the local shape and flexibility of the TATA-box from cyclization experiments made by Jason Kahn (University of Maryland College Park).
7. Structure of a Flp Recombinase DNA Complex Phoebe A. Rice, University of Chicago Flp is a very distant member of the lambda integrase family of tyrosine-based site-specific recombinases. Flp catalyzes the inversion of a segment of DNA in the yeast m plasmid, a process that permits the amplification of plasmid copy number. The reaction involves a covalent protein-DBA intermediate and the sequential exchange of pairs of strands. Using suicide substrates based on mechanistic studies from several laboratories, we have crystallized an intermediate of this reaction. The asymmetric unit of the crystals contains a tetrameric complex, with pseudo fourfold symmetry, and the Holliday junction intermediate is in a nearly square planar conformation. The nearly square planar geometry of the Holliday junction and the fold of the C-terminal catalytic domain are similar to that of the Cre-lox complex determined by the Van Duyne group, but other important features of the complexes are different. Among these are the protein-protein contacts that mediate communication among monomers and enforce the half-of-the sites activity of the tetramer, and the connectivity of the helix bearing the active site tyrosine, which is domain-swapped in Flp but not in Cre.
8. Dynamic Simulations of Regulatory DNA/Protein Systems Tamar Schlick, New York University One of the current challenges in macromolecular structure is to bidge the all-atom level of resolution with the macroscopic view of biological networks in the cell. Computational work can help bridge these levels of modeling, as well as provide insights into structural and functional processes studied by experiment. Recent work on using dynamic simulations and molecular modeling in our laboratory to study TATA-element/TBP association (role of DNA-sequence flexibility on transcription interaction and activity), and polymerase/DNA function (kinetics of DBA polymerase catalysis, including fidelity mechanism) will be presented. The TATA DNA studies reveal, by vigorous analysis, tailoring sequence-dependent structural, energetic, and flexibility properties for TBP interaction: overall deformability, minor groove widening (with roll, rise, and shift increases at the TATA ends), untwisting within the TATA element (with large rolling at the ends) and relatively low maximal water densities around the DNA. These properties are likely related to the association/dissociation interactions within the eukaryotic transcription assembly. The polymerase DBA studies have delineated a sequence of local conformational events involved in the large-scale opening motion of the polymerase and suggested a rate-limiting step (Arg258 rotation and release of the catalytic magnesium) that may explain the maintenance of polymerase's selectivity for the correct incoming partner to the template base.
9. Single-Molecule Analysis of DNA Uncoiling by a Type II Topoisomerase Terence Strick, Cold Spring Harbor Labs Type II DNA Topoisomerases (topo) are essential ATP- dependent enzymes capable of transporting a DNA through a transient double-strand break in a second DNA segment. This enables the enzymes to untangle replicated chromosomes and relax the interwound supercoils (plectonemes) which arise in twisted DNA. Here we present a micromanipulation experiment in which we follow in real-time the action of a single D. melanogaster topo II acting on a linear DNA which is mechanically stretched and supercoiled. By monitoring the DNA's extension, we directly observe the relaxation in the presence of ATP of two supercoils during a single catalytic turnover. By controlling the force pulling on the molecule, we find that enzyme turnover decreases as the force acting on the DNA increases. This suggests that resealing the cleaved DNA is a rate-limiting step in the enzymatic cycle. Finally, in the absence of ATP, we observe the clamping of a DNA crossover by a single topo II on at least two different time scales, i.e., configurations. These results demonstrate that single molecule experiments are a powerful new tool for the study of topoisomerases and other proteins which interact with DNA.
10. DNA Conformation and Genomic Organization Masashi Suzuki, AIST-NIBHT Structural Biology Centre, Japan The complte necleotide sequences of a number of genomes have been determined. The ultimate goal of genome science is to understand the whole system of life on the basis of these sequences, but not only to understand its individual components separately. Analysis of genes coded by genomes appears to be of the central focus of many researchers. However, understanding of a genome will not be completed without studying its other active component, the genomic DNA molecule. In fact, precise identification of protein genes becomes possible only by understanding physical characteristics of the DNA. A genomic DNA molecule interacts with its gene products in many important biological processes. Thus, gene-coding regions and non-coding regions are designed, so that these could differentiate from each other in terms of their physical characteristics, for the efficient discrimination of the two by transcription and translation machinery. Such differentiation will be most important in the human genome, in which over 90% do not code for a single gene. In fact, it seems difficult to explain this high rate without assuming some active roles for non-coding regions. In order to maintain interaction of the DBNA with proteins inside organisms living at different temperatures, nucleotide combinations change, so that an appropriate level of flexibility of genomic DNA molecules will be induced at these different temperatures. More than one nucleotide combination can be used for the coding of the same amino acid sequence, and this redundancy is used for designing genomic DNA molecules. A genomic DNA molecule is not an abstract medium designed only for installation of information on protein sequences. Importantly, a DNA molecule has physical reality. Physical characteristics of DNA are coded by the nucleotide sequence. This fact is comparable with another fact that the 3D structure and function of a protein are coded by its amino acid sequence. Physical characteristics of protein and DNA have enabled life to evolve to the forms that we see now. Also they have molded life into its basic design which we now see, and maybe have restricted life inside the limitation which we have now. References: 1. Kawashima, T. et al. (2000) Archaeal adaptation to higher temperatures revealed by genomic sequence of Thermoplasma volcanium. Proc. Natl. Acad. Sci. USA 97, 14257-14262. 2. Makino, S., Amano, N., and Suzuki, M. (1999) Visual presentation of complete genomic DNA sequences, and its application to identification of gene-coding regions. Proc. Jpn. Acad. 75B, 311-316. 3. Sucko, J. M. et al. (1998) A transcription frame-based analysis of the genomic DNA sequence of a hyper- thermophilic archaeon for the identification of genes, pseudo-genes and operon structures. FEBS Lett. 426, 86-92. 4. Suzuki, M., Amano, N., Kakinuma, J., and Tateno, M. (1997) Use of a 3D structural data base for understanding sequence- dependent conformational aspects of DNA. J Mol. Biol. 274, 421-435.5. Amano, N., Ohfuku, Y., and Suzuki, M. (1997) Genomes and DNA conformation. Biol. Chem. 378, 1397-1404.
11. A Theory of DNA Elasticity that takes into Account the Dependence of Deformability on the Base-Pair Sequence. David Swigon, Rutgers University In recent colloborative research, B.D. Coleman, W.K. Olson, and the speaker have demonstrated the feasibility of calculating equilibrium configurations in the theory of a base-pair level model for the elastic properties of DNA. In this "naturally discrete model", base pairs are treated as flat rectangular objects and the configuration of a DNA segment with N+1 base pairs is specified by giving, for each of its N base-pairs, 6 kinematical variables that measure the tilt, roll, twist, shift, slide, and rise of a base pair relative to its predecessor. The elastic energy of the segment is the sum over the base-pair steps of functions of the 6 variables. From fluctuations and average values of structural parameters in crystals of pure DNA and DNA-protein complexes, Olson and co-workers* have obtained empirical representations for 10 independent functions that quantify the sequence-dependent elastic behavior of duplex DNA. Each function is determined by the necleic acid sequence of the step and contains 6 intrinsic kinematical parameters and 21 elastic parameters, or "moduli". (Some of these 27 parameters vanish when a base-pair step is self-complementary). As one can calculate the dependence of configurations on end conditions for any specification of the elastic and kinematical parameters, we now have in hand the possibility of adjusting these parameters so as to maximize the agreement of the theory with experiments on DNA in solution. Presented will be the theory of the model and examples of its applications. We have found that, even when the theory is applied to idealized hypothetical examples, numerical solution of its governing equations can give insight into the influences of intrinisic curvature and twist on the changes in global DNA configurations caused by intercalating agents and other ligands. * W.K. Olson, A.A. Gorin, X.-J. Lu, L.M. Hock, & V.B. Zhurkin, "DNA sequence-dependent deformability deduced from protein-DNA crystal complexes," Proc. Natl. Acad. Sci., USA, 95, 11163-11168 (1998)
12. Earliest Evolutionary Signatures in DNA Sequences Edward N. Trifonov, The Weizmann Institute of Science, University of Haifa Modern nucleotide and amino-acid sequences stated their evolutionary history from very simple patterns. Reconstruction of the amino-acid and codon chronology and full genome sequence analysis indicate that the very first minigenes had the duplex structure (GCC)6*(GGC), presumably RNA, coding for Gly6 and Ala6, respectively. The 3-base periodicity of guanines in the modern protein-coding sequences and expansion of triplets GCC and GCU in neurodegenerative diseases reflect this earliest stage of sequence evolution. The size (period) of 18 bases (6 aa) is detected so far only at the protein sequence level. The next characteristic size, 75-90 bp, comes from recently discovered basic loop structure of proteins, and is observed in the actual sizes of the closed loops, and in autocorrelation patterns of protein sequences. The proteins, apparently, went through the 25-30 aa loop closure stage in their evolution. Modern proteins are largely built from consecutively connected loops of that size. With the increase in protein chain lengths respective genes soon reached their own optimal circularization size, 350-450 pb, coding for the proteins of typical fold size 120-150 aa. The well known ~ 10.5 bp periodicity in DNA sequences is likely to appear first at this stage of the sequence evolution. Both DNA curvature and anysotropic bendability associated with this period would facilitate the DNA ring closure. Thus, the consecutive stages in the evolution of the nucleotide sequences have imposed on them the periodicities of 3,~ 10.5, (18), (~80), and ~400 bases, of which the bracketed ones are detected so far at the protein sequence level.
13. Measurement of DNA Intrinsic Curvature in Solution Alexander V. Vologodskii, New York University Exceptional accuracy fo DNA persistence lengh determination can be achieved by the measurement of the cyclization efficiency of short DNA fragments, about 200 bp in lengh. Such accuracy results from the very strong dependence of the cyclization efficiency of the short DNA fragments on the persistence length. This accuracy and the possibility to work with a desired DNA sequence allowed us to study the sequence-dependent intrinsic curvature of DNA molecules. The measurable persistence length of DNA molecules with typical sequences depends both on the intrinsic curvature and on the thermal fluctuations of the angles between adjacent base pairs. Thus, to address the problem we designed an "intrinsically straight" DNA molecule that consists of 10 bp segments, in which the sequence of the first five bases is repeated by the sequence of the second five bases only. The persistence length of the first five bases is repeated by the sequence of the second five bases only. The persistence length of such DNA depends, to a good approximation, on thermal fluctuations. We measured the persistence length of this "intrinsically straight" DNA and compared it with the persistence length of regular DNA molecules. By this way we evaluated the contribution of intrinsic curvature to the measurable DNA persistence length.
14. Nucleosome Positioning DNA Sequences Jonathan Widom, Northwestern University Certain DNA sequences strongly bias the positioning of the histone octamer (i.e., a nucleosome) to a particular location along DNA. Conversely, given a strongly biased nucleosome position, particular DNA sequences can strongly affect the dynamic equilibrium accessibility of buried regions of the nucleosomal DNA. The present talk will explore the DNA sequence motifs responsible for such behavior. We will discuss the results of selection experiment for nucleosome positioning sequences, the basis of DNA mechanics of the selected sequences' properties, and analysis of signals for nucleosome positioning that are present in natural genomes.
15. Antiparallel DNA Loop in the Gal Repressosome Victor B. Zhurkin, Natinal Cancer Institute, NIH Gal repressosome assembly and repression of the gal operon in E. coli occurs when two dimeric GalR proteins and the histone- like HU protein bind to cognate sites causing DNA looping. Negatively supercoiled DNA is absolutely required for the repressosome formation, which prevents analysis of this 113 bp loop by concentional techniques. Structure-based genetic analysis defined the GalR surfaces interacting to form a stacked, V-shaped, tetrameric structure (similar to that of the LacI tetramer). In principle, this tetrameric structure is compatible with four different looping modes of DNA, two parallel and two antiparallel. We constructed stereochemical models for all the four possible DNA loops. The modeling was based on the sequence-dependent structural parameters of the interoperator DNA and conformation changes caused by GalR and HU binding. The harmonic energy functions included in this model account for the sequence-specific anisotripic bending and twisting of the duplex at the level of dimeric steps (1). They also incorporate the strong cross-correlation between the DNA bending and twisting. Evaluation of the DNA elastic energies gave unambiguous preference to a novel loop structure in which the two gal operators adopt an antiparallel orientation causintg undertwisting of DNA. The calculated DNA trajectory is consistent with the AFM images of the Gal loop. The 92 bp lac loop can also adopt a similar trajectory. We suggest that the antiparallel Gal-like DNA loops may serve as putative elementary units of prokaryotic chromatin architecture. 1. Olson, W.K., Gorin, A.A., Lu, X., Hock, L. & Zhurkin, V.B. (1998) Proc. Natl. Acad. Sci., USA 95, 11163-11168.

Previous: Participation
Next: Registration
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on April 6, 2001.