Sequence Dependence in Nucleosomal DNA Conformation
Student: David McCandlish, Swarthmore College
Mentor: Wilma Olson, Department of Chemistry, Rutgers University
The primary structural unit of eukaryotic chromosomal DNA is the nucleosome. Each nucleosome consists of 146 DNA basepairs wrapped 1.65 times around a core protein called a histone. Eukaryotic chromosomal DNA is generally packaged as a long series of nucleosomes joined by shorter DNA segments. This packing scheme helps to fit the DNA within the cell nucleus.
It turns out that some DNA sequences have a greater affinity to bond to the histone protein than others. This has particular biological significance because the DNA must be unwound from the histone prior to being transcribed. Thus a particular sequence binding strongly to a histone may result in decreased gene expression in the gene containing that sequence. Our hypothesis is that the local DNA structure implied by a particular base sequence may result in a greater or lesser affinity for the histone protein. Additionally, the requirement that neucleosomal DNA must conform to a particular winding structure may constrain the possible sequence choices.
One current goal in the study of nucleic acids is to develop methods to describe nucleic acid structures. My project focuses on a scheme that decomposes the larger structure into a sequence of relationships between adjacent basepairs.
By Euler's Rotation Theorem, it takes 6 parameters, 3 translations and 3 rotations to uniquely describe the spatial relationship between the ith and i+1th basepairs in a DNA structure. These six parameters are called the step parameters. Although the step parameters are often thought of as tools to describe an experimentally determined structure, my project looks at the step parameters in a functional sense, altering the step parameters to and observing the resulting changes.
Using 3DNA and several of my own Perl scripts, I took an existing high-resolution nucleosome structure (in atomic coordinates – pdb format), analyzed it to determine the step parameters, selectively changed a subset of the step parameters to average step parameter values for protein-DNA complexes, and rebuilt a simulated structure with the new step parameters (again in atomic coordinates). I then used Matlab and some scripts written by Xiang-Jun Lu to do a least squares fitting of the new structure to the original structure and compute the root-mean-square distance between the original atomic coordinates and the new atomic coordinates. For each subset of step parameters, this procedure compares the original sequence dependent structure to a DNA sequence with some parameters set to “generic” values. The root-mean-square distance between the two structures then serves as an estimate of the sequence specific structural information encoded in those parameters.
Repeating the procedure for all possible subsets of parameters, I created the following chart:
The bar height represents the root-mean-square distance caused by changing a particular subset of the step parameters. The critical groups of parameters are labeled.
The first thing to notice is that roll is the parameter whose sequence dependence has the greatest impact on the nucleosome structure. In fact, setting the roll to a constant value does not even produce a curved DNA structure (it is well understood that roll must vary sinusoidally to produce curved DNA). In contrast to roll, notice on the left side of the chart that the sequence dependent variation in rise and shift has very little impact on the larger structure. In fact, setting rise and shift to constant values produces an error comparable to the resolution of the original structure.
The three remaining parameters, slide, tilt and twist, have a moderate sequence dependent influence on the structure. It is interesting that setting both slide and twist to constant values produces an RMS distance less than setting either slide or twist to a constant. This is a consequence of RMS distance being a good measure of how much a structure has changed without telling you how it has changed. It turns out that a two-factor cancellation between slide and twist.
To better understand the function of each parameter, I created a series of animations showing the effect on the nucleosome structure of changing certain subsets of the parameters to constant values.
Here is one such animation. It starts out with the original structure and slowly changes the twist parameter to a constant value.
From observing the various animations, it is clear that twist, tilt and slide primarily regulate the “stretching” of the nucleosome coil along what is known as the superhelical axis. Roll controls the overall curvature of the superhelix. For more animations, click here.
I am continuing to work on several fronts. Currently, I am investigating which particular basepair steps in the nucleosome have the greatest impact on the overall structure and which step parameters are most important at each step.
Other projects include finalizing and documenting several of my scripts for DNA structural data manipulation and upgrading a particular script to simulate nucleosome structures with arbitrary DNA sequences.
I would like to thank all the people that have helped me with my project this summer, particularly Andrew Colasanti, Dr. Atsushi Matsumoto, Dr. Wilma Olson, Dr. A.R. Srinivasan and the DIMACS REU 2003.
Davey, C. et al. (2002). Solvent Mediated Interactions in the Structure of the Nucleosome Core Particle at 1.9Å Resolution. J. Mol. Biol. 329, 1097-1113.
El Hassan, M. and C. Calladine (1995). The Assessment of the geometry of Dinucleotide Steps in Double Helical DNA; a New Local Calculation Scheme. J. Mol. Biol. 251, 648-664.
Olson, W. et al. (1998). DNA Sequence Dependent Deformability Deduced from Protein-DNA Crystal Complexes. Proc Natl. Acad. Sci. USA. 95, 11163-11168.
Xiang-Jun Lu and Wilma K. Olson. (2002). 3DNA: A Software Package for the Anlysis, Rebuilding and Visualization of Three-dimensional Nucleic Acid Structures. http://rutchem.rutgers.edu/~xiangjun.