A Molecular Dynamics Simulation Study of Coaxial Stacking in RNA
Christoph Schneider and Jürgen Sühnel*
Biocomputing, Institut für Molekulare Biotechnologie, Postfach 100813, D-07708 Jena / Germany
Abstract
We report on unrestrained molecular dynamics simulations of an RNA tetramer binding to a tetra-nucleotide overhang at the 5'-end of an RNA hairpin (nicked structure) and of the corresponding continuous hairpin with Na+ as counterions. The simulations lead to stable structures and in this way a structural model for the coaxially stacked RNA hairpin is generated. The stacking interface in the coaxially stacked nicked hairpin structure is characterized by a reduced twist and shift and a slightly incread propeller twist as compared to the continuous system. This leads to an incread overlap between C22 and G23 in the stacking interface of the nicked structure. In the simulations the continuous RNA hairpin has an almost straight helical axis. On the other hand, the corresponding axis for the nicked structure exhibits a marked kink of 39°. The stacking interface exhibits no incread flexibility as compared to the corresponding ba pair step in the continuous structure.
皂罗袍
* Phone: +49-3641-656200; Fax: +49-3641-656210; E-mail: jsuehnel@imb-jena.de.
Introduction
2018年春节联欢晚会RNA can store genetic information and is involved in many other biological process like translation, regulation and catalysis. This functional diversity is also reflected in the structural diversity of RNA. One of the structural motifs is coaxial stacking of helical stems. It occurs in structures of tRNA, of the hammerhead ribozyme and of pudoknots, for example (1). Coaxial stacking interactions are usually not included in free-energy minimization algorithms for RNA condary structure prediction. Recently, however, model systems have been studied to estimate the thermodynamic contribution of this motif to the total free energy of RNA folding (2-4). The model systems consisted of oligomers binding to a hairpin stem with a four- or five-nucleotide overhang (Figure 1). In the studies it has been shown that the oligomers bind approximately up to 1000-fold more tightly than predicted for a free tetramer or pentamer duplex. Therefore, the authors have concluded that coaxial stacking provides large, favorable free energy contributions to RNA stability.
Thus far, no structural information on the RNA model systems investigated in the thermodynamic studies is available. Moreover, the coaxial stacking motif in known three-dimensional RNA structures is incorporated into a larger nucleic acid environment and may thus differ from the structure of the model systems. Some structural information for nicked DNA structures is available, however. A low-r
esolution X-ray structure of a nicked dodecamer DNA adopts a conformation similar to tho found in the intact structure (5). This indicates that the stacking and hydrogen bonding interactions are sufficient to overcome the effect exerted by the disruption of the backbone. It has been concluded from a gel electrophoresis study on a 139 ba pair DNA fragment with a single-stranded break that the nicked duplex is kinked (6).Nuclear magnetic resonance (NMR) spectroscopy studies and
restrained molecular dynamics (MD) simulations on nicked and gapped 14-mer DNA have also shown that the nicked structure is clo to a canonical B-DNA but exhibits enhanced local flexibility (7). On the other hand, gel electrophoresis experiments on DNA (8) and crosslinking studies on RNA (9) indicate that nicked structures retain much of the rigid character of the nucleic acids with uninterrupted strands whereas gapped structures exhibit an incread flexibility.
MD simulations are a theoretical tool for the microscopic description of structural and dynamic properties of molecules (10). Recently, the reliability of MD simulations for biopolymer structures has been substantially improved. One reason is the usage of the particle-mesh Ewald method (PME) to properly describe long-range electrostatic interactions (11). This has been especially important for the highly charged nucleic acids (12). Recent MD simulations on RNA have provided uful information complementing experimental studies (13-15) .
红酒开瓶器怎么用
脚底老茧
Here, we prent the results of two unrestrained MD simulations of a tetramer binding to a 4-nucleotide overhang at the 5'-end of a hairpin (nicked structure) and of the corresponding continuous hairpin (Figure 1) with simulation times of 3 and 2 ns. The aim of this study is to analyze structure and dynamics of the coaxial stacking motif and to compare the results to the related continuous structure.
Methods
政务专网
An initial hairpin structure was built from 2 subunits, loop and stem. The GCAA loop was taken from an experimental NMR tetraloop structure (PDB entry: 1zih; 16). The stem
调侃structure was built de novo with SYBYL (Tripos Ass., Inc.) in a canonical A-RNA conformation and merged with the loop. A nicked structure (in the following denoted as n) was generated by removing the phosphodiester linkage between C22 and G23 from the continuous backbone hairpin (in the following denoted as c). The model structures were solvated with ~2800 TIP3P water molecules. A coulombic potential on a grid was calculated and 29 (cNa+) / 28 (nNa+) Na+ and 4 Cl- counterions were placed according to this potential with xLEaP from the AMBER package to neutralize the system (17). The concentration of Na+/Cl- ions in the simulation was 0.51/0.07 M. All simulations w
ere run using the Sander module of AMBER 5 with the Cornell force field (18, 19). The nonbonded pair list was updated every 10 steps and a 10 Å cutoff was applied to the Lennard-Jones interactions. All structures were minimized for 1000 steps, afterwards water molecules and ions were relaxed, while the solute was kept fixed. Then all atoms were allowed to move. The system was gradually heated up to 300 K and equilibrated for additional 110 ps. Then production runs were performed for 2 (cNa+) and 3 (nNa+) ns at constant pressure (1 atm) and constant temperature (300 K) with a 2 fs time step. The simulation time for nNa+ was longer, becau a larger geometrical change during the simulation was expected in this ca. The PME method was ud for calculating the electrostatic interactions with a grid spacing of approximately 1Å (11). SHAKE was applied to all bonds involving hydrogen. The simulated systems had a density of 1.06 g/cm3. For both simulated systems we have generated average structures taking into account the last 500 ps of simulation time. The CURVES 5.1 program has been ud for the analysis of helical parameters (20). Only the 11 ba pairs of the stem were included in the analysis. Moreover, the ba C25 in the nNa+ average structure was ignored for the determination of the helical axis, becau an opening is found for the G2-C25 ba pair and this would significantly affect the orientation of the global helical axis. G2 was taken into account becau it has a similar orientation as the neighboring bas. Only global rotational
and translational helix parameters are prented in this work. In view of the fact that different analysis algorithms may lead to different helical parameter values (21), the data were, however, compared to the corresponding local CURVES parameters and to parameters determined with the program FREEHELIX (22).
Results and Discussion
鸡中翅The stability of simulated structures can be evaluated by comparing their geometries with the geometry of the starting structure by means of the root mean square deviation (RMSd) of heavy atoms. The time cour of the RMSd values for the simulated systems (cNa+, nNa+) is shown in Figure 2. In the final part of the simulation the RMSd values fluctuate about an almost constant average value. This indicates that the simulation leads to stable structures.
In Figure 3 the values of the root mean square fluctuations (RMSf) of the structure for the last 500 ps of the simulation time are shown. The RMSf values reveal the extent of motion of an atom around its equilibrium position. Hence, the data provide information on flexibility or rigidity of molecules or parts of them. For almost all nucleotides the peaks ari from the sugar-phosphate backbone. The incread flexibility of the backbone as compared to the bas is well-known from other MD simulati
besides什么意思ons on nucleic acids (15, 23).As expected the greatest flexibility is found for the end and loop regions of the hairpin structure. On the other hand, the stem is relatively rigid. In the context of coaxial stacking, it is important to identify possible differences in flexibility between the continuous and nicked structures. From the data shown in Figure 3 for nNa+ and cNa+ it is evident that the differences are small. Interestingly, the RMSf differences of the nucleotides forming the stacking interface (nucleotides 4,5,22,23) are especially small. In the remaining part of the molecule the flexibility of the nicked structure is
generally slightly larger as compared to the continuous hairpin. However, this difference is smaller than 0.3 Å in the stem region and only slightly incread in the loop and end regions. This means that according to the MD simulations the lack of the phosphate group in the nicked structure does not lead to a substantially incread flexibility. The results are in line with conclusions from gel electrophoresis and crosslinking studies (8, 9) but contrast with the NMR and restrained MD investigations by Roll et al. (7).
In order to detect structural differences between the two simulated systems we have performed a comprehensive analysis of geometrical ba pair, ba step and backbone parameters. The backbone parameters did not show any marked differences between the continuous and nicked hair
pin structures. Hence, only data for rotational and translational helical parameters are displayed in Figures 4 and 5. The opening found in the G2-C25 ba pair of the cNa+ average structure is reflected in many helix parameters. It is, however, ignored in the following discussion.
The great majority of helical parameters is similar for the continuous (cNa+) and nicked (nNa+) structures. A marked local effect at the stacking interface is en for the ba step parameters slide, shift and twist. All three are reduced in the nicked structure. In addition, the propeller twist for the two ba pairs at the interface exhibits a slightly incread angle in the nicked structure.Recently, it has been pointed out that different algorithms for analyzing nucleic acid conformations may yield contradictory results (21). We have thus compared the global CURVES results to local CURVES and FREEHELIX parameters (22). Except for the slide the local CURVES and the FREEHELIX results are similar to the differences in the helical parameters calculated with the global CURVES approach. From Figures 4 and 5 it can be further en that for the nicked structure the Y displacement, inclination and tip angles
