Structural Biology: Theoretical Prediction and Experimental Validation of DNA-directedNucleosome Stability

Faculty mentors: Thomas C. Bishop, Jeffrey C. Hansen, Karolin Luger, Simon Tavener

Project Objectives and Aims:

The objective is to provide a learning experience integrating theoretical and experimental studies of DNA-directed nucleosome stability. The nucleosome is a protein-DNA complex that functions as the structural subunit of eukaryotic genomes (Hansen 2002). The aims of this proposal are (1) to use mathematics and computer programming to predict which genomic DNA sequences direct strong positioning of nucleosomes; and (2) experimentally test the theoretical predictions by cloning the identified DNA sequences and measuring their nucleosome positioning strength in the laboratory in vitro.

Project Background:

The nucleosome is a protein-DNA complex of eight histones and 146 bp of DNA. In vivo, genomic DNA is wrapped 1.75 times around the histone octamer to form arrays of nucleosomes that fold into chromosomes (Hansen 2002).  Nucleosomes are routinely crystallized at Colorado State by Karolin Luger  (Luger 2003), a participant in this proposal. These crystal structures provide the atomic coordinates used in the theoretical prediction algorithm described below. The central hypothesis addressed by the proposed studies is that the propensity of any given 146 bp of genomic DNA to form stable positioned nucleosomes can be accurately predicted by theory.  

Project Description:

Nucleosome stability is largely determined by the physical properties of DNA, such as its shape and deformability. Chemical properties of DNA, such as its ability to make electrostatic, van der Waals or hydrogen bonding interactions, are much less influential (Windom 2001).  Given these assumptions, one can readily predict the stability of a nucleosome based upon the energy required to distort free DNA into the conformation it has in the nucleosome.  This energy is described by the equation

E= 1 2 145 Σ i=1  [ → X   (i)- → X    0  (i)]T K(i) [ → X   (i)- → X    0  (i)].

Here, i denotes the base pair step index and X^→₀(i) is a 6-vector representing the intrinsic shape of the i-th base pair step when it is free in solution. The components of each X^→₀(i) are the DNA helical parameters (i.e., Roll, Tilt, Twist, Shift, Slide, Rise) and there are 10 different values of X^→₀(i), one for each of the 10 unique base pair steps (Olson et al. 1998). Here K(i) is a 6x6 matrix, denoting the stiffness of the i-th base pair step, and it too is sequence dependent (Lankas et al. 2003). Finally, X^→(i) is a 6-vector representing the conformation of the i-th base pair step in a nucleosome. Under supervision of Dr. Tavener, and in consultation with Dr. Bishop, students involved in this project will use a program written by Dr. Bishop in FORTRAN (~2200 lines) that uses this equation to predict the most stable positions of nucleosomes on a segment of chromosomal DNA. The program can determine a stability profile for an entire genome (e.g., yeast) in minutes.  Students will be taught the fundamental concepts incorporated into the program, how to use it and interpret results.

To test the validity of the theoretical predictions, and to gain experience in translating theory to hypothesis-driven experimental research, UBM participants under the supervision of Dr. Hansen will clone, express, and purify  the predicted DNA sequences, and reconstitute them with histone octamers in nucleosome assembly competition assays to measure their relative stability. The experimental data will be used to refine future versions of the prediction program, providing a sustaining source of UBM research projects in the areas of structural biology and structural genomics.