Department of Structure and Dynamics of Nucleic Acids

We use a wide spectrum of state-of-the-art computational techniques, including explicit solvent molecular dynamics simulations, advanced ab initio quantum-chemical calculations and modern bioinformatics.

            The main research aim is to provide unique insights into the role of molecular interactions in structure, dynamics, function and evolution of nucleic acids. In that sense, advanced computations can substantially complement experimental techniques. Our research is highly interdisciplinary and we wish to bring modern physical-chemistry insights into structural and molecular biology, biochemistry and bioinformatics. Our research has impact also in some other areas of chemistry, such as physical, supramolecular and bioinorganic chemistry.

            Ab initio quantum chemical (QM) technique is state of the art physical-chemistry methodology that provides accurate and physically complete description of small model systems. The technique reveals direct structure – energy relations that cannot be obtained by any other technique. Such calculations have indispensable role in reference evaluation of nature and magnitude of all kinds of molecular energy contributions that shape up nucleic acids, such as base stacking, base pairing, backbone conformational preferences, etc. QM calculations allow to study chemical reactions at a level of electronic structure changes.

            Classical explicit solvent molecular dynamics (MD) simulations characterize structural dynamics of nucleic acids with size of up to 100 or more nucleotides. The currently available simulations can be extended to ~1 microsecond time scale. With this technique, nucleic acids are modeled at the atomistic level of descriptions using classical potential functions known also as empirical force fields. Approximations inherent to the force fields represent the main limitation of this approach.

            Structural bioinformatics aims, among other things, to provide classification of molecular interactions in nucleic acids based on structural and sequence data. 

            Due to the unprecedented development of hardware and software in the last decade, modern computational techniques represent a useful tool to complement the experiments. This fast-expanding research field changed completely in the last 10 years. The continuing progress in developing better and better computers guaranties that the impact of computational field is continuing to grow. In many cases, when properly applied, computations can bring key insights into the fascinating world of biomolecules that cannot be obtained by experimental techniques. Computations, although limited by a number of approximations, represent an excellent analytical approach to understand the biomolecules, their structure and function. Our aim is to carry out high-quality, reliable computational studies that successfully target relevant structural biology and biochemical problems. We believe that our recent papers (dealing with key ribosomal RNA elements, guanine quadruplex DNA molecules, ribozymes, metal cation - nucleic acids interactions, basic role of base pairing and stacking in nucleic acids, prebiotic chemical reactions, testing and parametrization of force fields, etc.) prove that this ambitious task can be achieved. We cautiously try to avoid overselling the computational results, which we consider to be one of the major problems of contemporary computational field. To achieve our basic scientific goals, we intensely cooperate with many theoretical as well as experimental laboratories (see the list of collaborations). We belong to the most cited laboratories working in the Czech Republic.

Our current research includes intense MD simulation studies of structure and dynamics of functional RNA molecules. They include various ribosomal RNA segments, several catalytic RNAs (ribozymes) and some other molecules.

Figure 1. Ribosome and some segments of ribosomal RNA studied in our laboratory. Ribosome is a complex biomolecular machine that is dynamical. Molecular simulations combined with structural bioinformatics aim to analyze intrinsic flexibilities of distinct segments of ribosomal RNA and to suggest their specific roles in the protein synthesis. In this manner simulations can complement the static and averaged pictures of ribosome obtained by experimental structural techniques.

Figure 2. Small catalytic RNA molecule (Hairpin ribozyme) and detail of its catalytic pocket as seen during the atomistic explicit solvent molecular dynamics run. Computational studies aim to improve our understanding of the catalytic mechanism and the role of various base substitutions.

Figure 3. Molecular complex between RNA three-way junction and RNA-binding protein.

Figure 4. Molecular complex between RNA kink-turn (a recurrent elbow-like RNA motif) and protein. Our simulations revealed that some kink-turns can act as flexible elbows that could facilitate various motions in large ribonucleoprotein systems such as the ribosome.

Figure 5. Another view on RNA kink-turn, with tertiary A-minor interaction between the stem of the molecular elbow shown in red. Dynamical substates of this tertiary interaction significantly contribute to the unusual flexibility of the kink-turn.

 

Besides RNA, we also study B-DNA and non-canonical DNA forms such as guanine quadruplex.

Figure 6. Four-stranded quadruplex DNA (G-DNA) is the most important noncanonical DNA with significant biological, biochemical and pharmacological implications. Parallel stranded four-quartet quadruplex stem is shown.

Figure 7. Molecular simulations were used to propose plausible structures of kinetic intermediates during formation of the quadruplex stem. This investigation illustrates that simulations can be efficiently used to provide predictions going beyond the experimental insights. Our model has been subsequently verified by independent experimental studies.

 

Very important part of our work is methodology development. Specifically, we carry out extensive studies of the applicability and limitations of empirical force fields that are used for MD simulations. These studies are followed by efforts to tune and refine the force fields. This is area of research where we fully capitalize all our knowledge ranging from the highest-accuracy QM calculations to unique experience from simulation studies of all kinds of nucleic acids systems.

The quantum chemical research is aimed at basic characterization of key molecular interactions in nucleic acids, such as base stacking, base pairing and metal-cation binding.

Figure 8. Reference QM calculations provide accurate and physically complete description of nature and magnitude of key molecular interactions in nucleic acids.

Our calculations are widely used as reference data in many laboratories around the world. Current intense efforts are increasingly focused at investigations of intrinsic conformational preferences of the nucleic acids sugar–phosphate backbone. The QM branch of our projects is critically important for refinement of nucleic acids force field.

Intense QM studies are also devoted to chemical problems related to the Origin of life theory, especially to the prebiotic synthesis of nucleic acids and their components.

Figure 9.  Initial state (a) and transition state (b) complex for the phosphate-mediated ring closure reaction of the adduct formed from cyanamide and glycolaldehyde, which is one of the key steps on the road to cytidine-2’,3’-cyclic phosphate according to Sutherland et al.  Computations have shown that participation of a phosphate ion is indispensable for this reaction, since it acts simultaneously as general acid and base and the reaction proceeds in a concerted manner.

The main strength of the QM technique in this field is in its simplicity: we are able to extract information related to the behavior of single molecules, which is often impossible with experiments performed in rather complex matrices representing the prebiotic environment.  For example, our studies uncovered the physicochemical reasons that could contribute to the selection of ribose among the four aldopentoses as a building unit of the first informational polymers.  With regard to the prebiotic nucleotide synthesis our computations have shown that the classical reaction route, assuming glycosylation of the nucleobases, is prebiotically not plausible, unless it is preceded by the phosphorylation of ribose.