Three-dimensional Structures of Small RNA Molecules
Principal Investigator
Peter Schuster
Co-Investigator:
Peter F. Stadler
Co-workers:
Alexander Renner, Stefan Kopp.
Support:
Fonds zur Förderung der Wissenschaftlichen Forschung
Project No. 11065-CHE
Begin 1995
Abstract
The three-dimensional structures of RNA molecules are investigated using
both theoretical methods such as Molecular Modeling, Molecular Mechanics
or Molecular Dynamics, and NMR spectroscopy. Secondary structure prediction,
either in terms of the ground state or in terms of the partition function
of the Boltzmann ensembles of all structures, serve as starting points
for the 3D structure prediction. The double helical stacks are introduced
into the computations with constant geometries as determined by the
base pairing patterns of the secondary structures.
In the case of small RNA molecules with up to 30 nucleotides that lack
extremely flexible structural elements (such as large loops, joints, or
long free ends) we are considering to calculate the minimum free energy
structures by means of direct energy minimization. Preliminary tests with
small hairpin molecules with three- or four-base loops and chain lengths
of 11 and 12 bases, respectively, showed two major outcomes:
- the structures are in good agreement with existing experimental data
for the same three- and four-base loops, and
- the geometries of all optimized structures follow a common principle of
minimizing the contact surface between the heterocyclic bases and the
solvent.
NMR spectroscopy experiments (L.R. Brown, IMB Jena; K. Wuethrich, ETH)
Zuerich) will be carried out on the same small molecules and permit a direct
comparison with the computed geometries. If there is basic agreement we plan
to iteratively refine the structure predictions by means of mutual
corrections of the structures obtained by NMR spectroscopy and Molecular Mechanics.
In the case of RNA molecules with extremely flexible structural parts we want
to supplement the structure prediction with Molecular Dynamics computations.
Since we expect major problems it will be necessary to incorporate NMR data as
structural constraints in the modeling. We plan to choose a sample of small
flexible RNAs as models and study them extensively to gain experience with
the flexible parts of molecules.
The results of the investigations of small RNA molecules together with known
structure data for the constant geometries of partial structures (A-RNA double
helix, rigid geometries of small loops, etc.) and NMR data shall be used to
predict the structures of larger RNAs with sequence lengths up to 100 bases.
The preliminary final aim of this project is the development of a software
package that is capable of predicting the 3D structure of an RNA molecule
(using NMR data as input) without user-interventions.