The Interaction And Folding Of RNAs And Other Biomolecules

Proteins are essential constituents of all living organisms and participate in virtu-ally every process of life activity within cells. Proteins are polymers made of amino acids arranged in a linear chain and folded into a globular, stable and native struc-tures. These unique native structures are usually assumed to be responsible for their functions. Proteins can also work together to achieve a particular function, and they often associate with RNA or DNA to form functional complexes. All these activities of protein are fundamental problems in molecular biology.Due to the complexity of the physicochemical interactions in proteins, people de-veloped various versions of simplified interactions for protein studies. In our first work, we describe a simplified residue-residue interaction potential. This potential is a statis-tical one in nature. It does not necessarily correspond to the true physical interactions, but an ensemble average of the contributions of the complex protein-solvent system, the solvent entropy effect and the interaction enthalpy. In this study, a statistical method called "ENERGI" was used to constructed the residue-residue interaction matrix. It is an iterative method that extracts the pairwise additive amino acid "energy" score from a protein structural database. The successful rate of identifying native structures out of decoys in the testing set is 95.8%. The correlation coefficient between our set of parameters and the MJ and BT matrix was 0.74 and 0.80, respectively. The potential matrix can be used as the parameters of contacts in the coarse-gained model for protein dynamics. It is an advantage that the interaction matrix can be used to assign each residue a score, which correlates well with the hydrophobicity of the residue.Ribonucleic acid (RNA) is another major macromolecules that are essential for all known forms of life. Although it has been generally assumed that most genetic information is transacted by proteins, recent evidence suggests that the majority of the genomes of mammals and other complex organisms is transcribed into non-coding RNAs. It is now believed that these non-coding RNAs comprise a hidden layer of internal signals that control various levels of gene expression, determining most of our complex characteristics, and constituting an unexplored world of genetic variation both within and between species. To fully appreciate the functions of non-coding RNAs, it is necessary to study the folding/unfolding mechanisms by which they achieve three-dimensional functional structures.In our second study, we deliberately choose a pseudoknot within gene 32 mRNA of bacteriophage T2 and study its unfolding process using an ensemble of molecular dynamics simulations. This RNA pseudoknot is a good model for studying the fold-ing/unfolding process, in that it has both secondary and tertiary structures and also contains complicated interactions/structures such as noncanonical base pairs, triplex-es, coaxial stackings, and sharp turns. The study of pseudoknots may benefit predic-tions of RNA secondary structure as well, since their pseudoknotted topology and the noncanonical interactions therein present major obstacles to the development of both sampling algorithms and free-energy estimation rules. Our simulations revealed large diversities in both unfolding pathways and transition states, indicating the possible failure of the widely used two-state model in describing the folding/unfolding of this RNA pseudoknot. They also demonstrated the close interplay between the canonical and noncanonical interactions in determining the thermodynamic stabilities of struc-tural elements and altering the kinetic pathways. Moreover, the simulations provided an atomistic picture for the unzipping mechanism, the early break of the coaxial s-tacking, and the possible driving role of structural tension in the initial stage. The dynamical behaviors of water molecules were also studied and found to be coupled with unfolding through the expulsion or concurrent mechanism in the early unfolding stage. Water molecules also help mediating the interactions between nucleotides and contribute to the stability of compact unfolded structures. This study provides for the first time an atomistic and statistical analysis of the unfolding process of an RNA pseu-doknot. The unprecedentedly detailed atomistic picture helps us better understanding the underlying kinetics and the contributions of various physical and chemical factors. It may also inspire new ideas in developing new structure prediction algorithms, as we have witnessed in protein folding area.In our third work, we continue to study the folding/unfolding process of the RNA pseudoknot within gene 32 mRNA of bacteriophage T2. To characterize the free ener- gy landscape of this RNA pseudoknot, we employed an advance sampling method, i.e., the replica exchange molecular dynamics method and performed a large-scale parallel simulations to gain a sufficient sampling of the phase space. The total simulation time was approximately 16.4μs. Combined with the weighted histogram analysis method, the free energy projections on the order paramaters Q1 and Q2 at different temper-atures were calculated. These free energy landscapes revealed several intermediate states, which have their first helix almost intact while lose the second helix. The ex-istence of intermediate states suggest that for some folding pathways, the folding may follow a three-state mechanism. This study is a first computation that attempts to depict the whole free energy landscape through all-atom simulations for the folding/unfolding process of RNA pseudoknots. It presents atomistic details for the structures of inter-mediate and transition states, and the folding/unfolding pathways.The contents of my thesis are arranged below. In chapter one, I introduce the fundamental knowledge of the structure, thermodynamics and kinetics of proteins and RNAs. In Chapter two, I describe our work on the development of a statistical residue-residue interaction potential, which is based on a method called "ENERGI". In Chapter three, I present our work on the unfolding process of the RNA pseudoknot by using an ensemble of unfolding simulations. This study is followed immediately by an even larger scale simulations with replica exchange method; the obtained free energy land-scapes and the structures of intermediate states are described in Chapter four. Finally, in the last Chapter I draw the conclusions of this thesis and give suggestions for future work.