Amino Acid--Base Pairs
Hydrogen bonding is an important force that helps determine the specificity of nucleic acid interactions, either within a nucleic acid structure or between proteins and nucleic acids. In efforts to better understand the contributions of hydrogen bonding to nucleic acid recognition and to assist in modeling or designing nucleic acid structures and interacting molecules, we have generated databases of all sterically possible base-base and base-amino acid interactions. "Complete" libraries have been calculated for base pairs and base triples, as well as for amino acid-base and amino acid-base pair interactions, which may be accessed from these web pages. A partial library of base quartets is also presented to give a feel for the types of interactions possible.
(see paper by BW Walberer, AC Cheng, and AD Frankel)
The calculation of base-base interactions uses simple geometric criteria to form hydrogen bonds and to avoid steric clashes between bases. Hydrogen bonds are defined by three geometric parameters, the distance between the heavy donor and acceptor atoms and the angles at the acceptor and donor sites. For every pairwise combination of nucleotide bases (A, G, C, and U, as well as the protonated bases A+ and C+), and for every combination of donor and acceptor groups on the bases, a linear hydrogen bond is initially formed. The two angles at the acceptor and donor sites are then systematically and independently varied in the plane of the base pair without breaking the preformed hydrogen bond. At each step the distances and angles for all other combinations of donor and acceptor groups on the two bases are calculated and a steric check is performed to identify additional, geometrically allowed hydrogen bonds. All different patterns of hydrogen bonds found in this manner are stored in a database file. This includes the patterns containing just the initially formed single hydrogen bonds. By extending this calculation to combinations of three and four bases, complete databases of planar base triplets and partial databases of base quartets have been constructed. The initially calculated databases were further refined to remove arrangements that cannot accommodate the sugar-phosphate backbone or that have isolated functional groups which are inaccessible to water molecules for solvation. The pages here display only the possible base-base interactions that are conformationally "fixed", with at least two hydrogen bonds in a base pair or three in a base triple.
Amino Acid-Base Interactions
(see paper by AC Cheng, WW Chen, CN Fuhrmann, and AD Frankel)Hydrogen bonds are an important component of most specific nucleic acid-protein interactions. We have systematically modeled all possible amino acid hydrogen bonding interactions to the individual bases and to the 53 possible RNA base pairs identified above (including the protonated bases). The approach is similar to that described for base interactions, except that the conformational search was performed in three dimensions, rather than restricted to one plane. Only the hydrogen-bonding moieties of the amino acid side chains were examined, but additional tests were performed to ensure that full side chains, as well as the nucleotide sugars, could be sterically accomodated. As for the bases, only "fixed" interactions involving at least two hydrogen bonds are presented on these pages. For base pairs, we show only "spanning interactions", which are arrangements in which an amino acid hydrogen bonds to both partners of a base pair. These types of interactions in principle may be used to uniquely specify unusual base pairs found in RNA structures. The interactions to single bases may be seen in bulge or loop regions of an RNA where bases are unpaired. While many conformations for a given hydrogen bonding arrangement are possible, only one is presented for each, chosen to represent as planar a conformation as possible.
Filtering the databases
The databases of interactions between bases, or bases and amino acids, may be filtered by a variety of criteria, such as the number of hydrogen bonds, arrangement of donors and acceptors, isomorphism of base arrangements, etc., as described in the papers by BW Walberer, AC Cheng, and AD Frankel (2002) and AC Cheng, WW Chen, CN Fuhrmann, and AD Frankel (2002). Interactions containing bifurcated hydrogen bonds or single hydrogen bonds that are not as conformationally constrained as the "fixed" arrangements shown on this site also may be accessed via the filters. For information about sorting through the larger databases or using the additional filters, please contact Alan Frankel (email@example.com).
You can view a QuickTime movies demonstrating how these libraries were generated. One is a search for G-C+ base pairs (601 Kb) where two hydrogen bonds are found. There is also a search for A-U base pair (666 Kb) where only one additional hydrogen bond is found.