Conformational Analysis of Modified DNA and RNA by Methods of Molecular Dynamics Simulation
Summary of PhD thesis written in October 1999 under supervision of Prof. Dr. Ryszard W. Adamiak
Laboratory of Structural Chemistry of Nucleic Acids
Institute of Bioorganic Chemistry of the Polish Academy of Sciences
Noskowskiego 12/14, PL-61-704
Poznan
INTRODUCTION
Over the last decade, molecular modelling has become a method widely
accepted in structural chemistry and structural molecular biology. Various
simulation techniques exist for objects of a wide range of molecule size.
Among those, molecular dynamics (MD) simulation methods have proved their
applicability for elucidating the structure and dynamics of biopolymers,
both when used in the final stages of studies accomplished by NMR or X-Ray
methods, and as an independent strategy of conformational analysis directed
for objects not yet synthesised. For this latter approach, the problem
of the reliability of simulation protocols and representation of the molecular
system is crucial.
The main aim of my PhD thesis was conformational analysis of short
RNA and DNA fragments containing non-standard structural elements in order
to evaluate their influence on the local and global duplex conformation
using various strategies of molecular dynamics simulations as well as a
variety of procedures for structural analysis and visualization of the
results.
METHODS
All simulations were performed using the
Amber forcefield implemented
into the Discover (MSI) and Sander (UCSF) programs. Model building procedures
available in the InsightII (MSI) and AMBER 4.0 (UCSF) suites of programs
were used. For parametrisation of the modified nucleosides, the Gaussian94
and Resp (UCSF) programs were applied. The experiments were performed using
the advanced strategies of in aqua simulations: both the nonbonded cutoff
approach and the one including the PME method for treatment of nonbonded
interactions. The simulation time varied from 200 to 1000 ps, dependent
on the complexity of the system. The simulations were carried out on two
CRAY machines (YMP-EL and J-916) working in the Poznan Supercomputing and
Networking Center affiliated to our Institute. Subsequent conformational
analysis of the MD trajectories has been performed on the CRAY and SGI
machines using several programs, e.g. CURVES 5.1 (Lavery & Sklenar),
SAHN (program for cluster analysis applied for evaluating the internal
conformational similarity within the trajectories) as well as our own procedures,
mostly designed for the InsightII environment, e.g. to monitor the stacking
interactions within the nucleic strands or to evaluate the nucleic acid-solvent
interactions. Other procedures were created for visualisation of the resultant
structures as well as for presentation of structural parameters.
DNA DUPLEXES CONTAINING 1,N(6)-ETHENOADENINE
A series of 11 bp DNA duplexes containing a single 1,N(6)-ethenodeoxyadenosine
(edA) residue in the middle position of the
molecule was examined. Two cases were selected where the modified nucleo-base
was involved in stable hydrogen bond formation with the opposite base.
In the syn-edA - anti-dC case, one hydrogen
bond was identified over the considerable part of the trajectory. In the
other case, two stable hydrogen bonds were found between syn-edA
and anti-dG, which conformed previous NMR and X-Ray data. For this model
and for the analoguous unmodified duplex with the dC-dG pair, detailed
in aqua MD simulations were performed. The stability of the trajectories
obtained clearly confirmed the better reliability of the PME method over
the nonbonded cutoff approach. The resultant structures showed the ability
of the modified residue to adapt to the B-DNA duplex conformation by forming
a base pair with the opposite guanosine, which mimics a standard Watson-Crick
pair. The approach based on the cluster analysis method applied to the
trajectory conformations made it possible to identify typical conformations
which seem to predominate during the simulation.
a-ANOMERIC NUCLEOSIDE IN A DNA DUPLEX
Little is known about the influence of a single a-nucleoside
residue on the overall conformation of a DNA duplex. Our special interest
was drawn by the question why nature had “chosen” only b-nucleosides
as the structural elements of all known nucleic acids. I performed a series
of in vacuo and in aqua molecular dynamics simulations in order to test
the hypothesis of a duplex axis bending caused by such a modification,
which was suggested by certain physical analyses of molecules of this kind.
Paralelly, these experiments helped me to optimise the simulation protocols
used in my research. Preliminary simulations (especially in vacuo and those
including the nonbonded cutoff approach) caused various kinds of massive
distortions within the objects tested (both unmodified and a-anomer-containing
DNA). In most cases, however, the duplexes modified with a single a-anomeric
nucleoside (a-adenosine or a-1,N(6)-ethenodeoxyadenosine)
showed
a considerable degree of main axis bending when compared to the reference
(unmodified) molecules. Still, the most advanced experiments including
the PME method have not provided a doubtless answer as to the existence
of such an effect. On the other hand, I have found a clear ability of the
a-anomeric
residue to form base pairs with the opposite
b-nucleosides, which mimic the respective
beta-beta pairs, both in the molecules containing dA-dT and edA-dG
pairs.
DNA DUPLEXES CONTAINING THE LUMINARINE FLUOROPHORE
An experiment was performed to describe the conformational dynamics
of a DNA duplex containing a single deoxyluminarosine residue (the strongest
fluorophore applied in our group as a conformational probe for structural
research). It suggested that this three-ring system easily adapts itself
to the stacking interactions within the DNA strand and, to some extent,
it is also capable of forming hydrogen bonds with opposite guanosine and
thymidine.
BULGE RNA DUPLEXES MODIFIED WITH 2-AMINOPURINE
As part of a larger project aimed at elucidating the structure and
dynamics of short RNA duplexes containing purine bulges, I performed molecular
dynamics experiments in order to simulate the conformation of such molecules
containing adenosine or 2-aminopurine riboside (its fluorescent isomer)
residues in the bulge region. The in aqua experiments, both using the cutoff
and PME protocol, resulted in trajectories which suggested a considerable
bending of the bulged duplex, whereas the bulge residues tend to adopt
an orientation towards the solvent. No difference between the behaviour
of the
2-aminopurine system compared with the native adenosine containing
bulge was observed. This confirms earlier thermodynamical results from
our group concerning the behaviour of 2-aminopurine substituted for adenine.
HIV-1 TAR RNA AND ITS MODIFICATIONS
Experiments focused on the HIV-1 TAR RNA molecules were aimed at identifying
the conformational dynamics within the unpaired region of this molecule.
I performed three PME in aqua MD simulations of molecules differing with
the sequence of the apical loop, which started from the same initial conformation
acquired from PDB (conformation No. 17 selected from the 1ANR set provided
by the Varani group). One molecule was consistent with the native TAR RNA
sequence and two others were modified with 2-aminopurine in the G34 and
A35 positions, respectively. The resultant trajectories helped me to describe
the dynamics of the residues within the loop and suggested that 2-aminopurine
can replace adenine, but not guanine, with little effect on the local structure.
In contrast to the NMR data, the molecular dynamics experiment shows a
strucuralization of the apical loop. The final conformation of the unmodified
molecule was also used for an additional molecular mechanics analysis of
the probability of magnesium cation binding near the trinucleotide -UCU-
loop. This experiment showed two sites of possible binding of Mg2+.
RNA HYDRATION
The trajectories of the experiments described above which concerned
RNA molecules were also analysed in order to identify the specific hydration
pattern, especially for the unpaired regions of RNA. I calculated the relative
hydration of particular atoms, including hydrogens from C-H groups. This
analysis was correlated with the conformation of unpaired regions of the
molecules, especially in regard to the tendency of nucleosides to adopt
orientation towards the inside or outside of the molecule. I have also
identified several water bridges linking highly hydrated atoms on the surface
of the bulged regions of the RNA molecules.
CONCLUSIONS
The simulations presented in my PhD covered a wide variety of problems
which may be effectively researched using the molecular dynamics simulation
methods. The detailed results of the experiments performed have so far
been published in the following papers:
1. T. Kulinski, M. Olejniczak, H. Huthoff, L. Bielecki, K. Pachulska-Wieczorek, A. T. Das, B. Berkhout & R. W. Adamiak: The Apical Loop of the HIV-1 TAR RNA Hairpin Is Stabilized by a Cross-loop Base Pair. J. Biol. Chem. (Vol. 278 (40)) 2003, pp. 38892-38901.
2. M. Olejniczak, Z. Gdaniec, A. Fischer, T. Grabarkiewicz, L. Bielecki, R. W. Adamiak: The bulge region of HIV-1 TAR RNA binds metal ions in solution. Nucleic Acids Res. (Vol. 30) 2002, pp. 4241-4249.
3. T. Kulinski, L. Bielecki, R. W. Adamiak: Structure and dynamics of adenosine loops in RNA bulge duplexes as revealed by linked application of thermodynamics, spectrofluorimetry and simulation of molecular dynamics. Nucleic Acids Res. Supplement (No. 1) 2001, pp. 139-140.
4. L. Bielecki, R. W. Adamiak: Structure and dynamics of DNA duplexes containing single a-adenosine residues. Acta Biochimica Polonica (Vol. 48) 2001, pp 103-111.
5. L. Bielecki, B. Skalski, I. Zagorowska, R. E. Verrall & R. W. Adamiak: Fluorescent a-anomeric 1,N(6)-ethenodeoxyadenosine in DNA duplexes. The a-edA / dG pair. Nucleosides, Nucleotides & Nucleic Acids (Vol. 19) 2000, pp. 1735-1750.
6. L. Bielecki, T. Kulinski & R. W. Adamiak: Structure and dynamics of adenosine loops in RNA bulge duplexes. RNA hydration at the bulge site. In: RNA Biochemistry and Biotechnology. J. Barciszewski and B. F. C. Clark (eds.), NATO ASI Series, Kluwer Academic Publishers 1999, pp. 73-87.
7. T. Kulinski, L. Bielecki, M. Olejniczak, I. Zagorowska &
R. W. Adamiak: Structure and dynamics of the apical loop region of
29-mer hairpin of the TAR RNA HIV-1 sequence. Collection Symposium
Series (Vol. 2) 1999, pp 191-196.
8. A. Fischer, Z. Gdaniec., M. Olejniczak, L. Bielecki, R. W. Adamiak: Does 29-mer RNA hairpin of the HIV-1 TAR RNA sequence bind magnesium? Nucleic Acids Symposium Series (Vol. 42) 1999, pp. 117-118.
9. T. Kulinski, L. Bielecki, I. Zagorowska & R. W. Adamiak, R. Rigler: Dynamics of RNA bulge duplexes. 2-Aminopurine labelled adenosine loops. Spectroscopy of Biological Molecules: Modern Trends. Annex, 1997 UNED Madrid, Spain, pp. 39-40.
10. T. Kulinski, L. Bielecki, I. Zagorowska & R. W. Adamiak: Introductory data on dynamics of RNA bulge duplexes. 2-Aminopurine labelled adenosine loops. Collect. Czech Chem. Commun. (Vol 61) Special Issue 1996, pp. 265-267.
11. R. W. Adamiak & L. Bielecki: An optimized protocol for in vacuo molecular dynamics simulation and trajectory analysis of modified DNA duplexes. Computational Meth. Sci. Tech. (Vol. 2) 1996, pp. 7-16.
Also available:
Comments, remarks or questions