The number of nucleic acid sequences in GenBank is increasing much faster than the 3D structures in the PDB. Thus, there is a need to develop methods for predicting nucleic acid 3D structure from sequence. To address this problem, we are developing a suite of computational tools for nucleic acid sequence alignment, secondary structure prediction, de novo 3D structure prediction, 3D homology modeling (see figure), and structural analysis. Despite similarities to the protein folding problem, we have found that nucleic acids have unique characteristics that require new ideas. We are developing a new classical molecular simulation forcefield for nucleic acids (including modified residues) and new conformational search methods. An important target for structure prediction are the ribosomes from bacterial and eukaryotic pathogens as well as the human ribosome. The ribosome, however, is a huge protein-RNA complex with many moving parts and cofactors. We are interested in developing methods that can automatically model both the RNA and protein components of ribosomes given the whole organism genome sequence. In addition, we are working to develop modeling methods for interactions of RNA with small molecules, which will enable virtual screening of new classes of anti-bacterials and other drugs.

Our group uses NMR spectroscopy to solve atomic-resolution structures of biologically important RNAs. We collaborate with Dr. Philip Cunningham (Biological Sciences), who has developed a novel method called Instant Evolution, which allows for functional rRNA mutants to be genetically selected from a pool of completely randomized regions of 16S rRNA. Application of this method to highly conserved regions of rRNA and NMR structure determination of the wild-type and instant-evolution mutants has revealed key structural features that are required for ribosome function. We have solved the structures several independently folding ribosomal RNA domains. We are also studying the structure of RNA complexes with bound peptide ligands derived from library screening (collaboration with Dr. Christine Chow). The structures will be used to deduce the rules for RNA recognition and for rational design of peptidomimetics and refined drugs with improved pharmacokinetic properties.

An important goal of nanotechnology is to generate complex self-assembling molecular systems with predictable shape, dynamics, and catalytic properties. Because of the strength of Watson-Crick base pairing molecular recognition, both DNA and RNA have unique properties that make them well suited to nanotechnology applications. However, there has been a lack of understanding of 3D structure, molecular dynamics, hybridization kinetics and thermodynamics of nucleic acid self-assembly. Toward this problem, we are using our knowledge of nucleic acid 3D structure and also principles of hybridization and folding thermodynamics determined previously in the SantaLucia Lab. The goal is to allow for the fully automated design of DNA and RNA nanostructures. Beyond nucleic acid structure is the characterization of the time domain of nucleic acid structural fluctuations, which can in principle be exploited to allow for controlled movement of DNA and RNA nano-machines and catalytic properties. We also want to understand the kinetics of structural self-assembly and use this information to improve the yields and purity of the desired structures. Experimentally, we use a wide variety of techniques to characterize the structures including X-ray crystallography, electron microscopy, time-resolved fluorescence resonance energy transfer, NMR, and gel electrophoresis. Our goal is not only to improve the fundamental understanding of nucleic acid nano-structures, but also to make practical materials and devices.