186
M. Ilchenko and I. Dubey
mechanics (MM) [45]. QM methods provide more accurate modelling results than
MM-based approaches, although they are generally much more demanding in terms
of a computation power. It is now widely recognized that both methods reinforce
one another in an attempt to understand chemical and biochemical behaviour of biomolecules at the molecular level. From a practical point of view, the complexity of
the system, time limits, available computation resources and other limiting factors
determine which method is feasible [46].
MM may be used to model biomacromolecular systems to which even semiempirical QM calculations can be applied effectively. In MM, molecular motions
are determined by the masses of atoms and the forces acting on them, whereas wave
functions or electron densities are not computed. MM is widely used in chemistry
and biochemistry to obtain molecular models since this approach is much faster and
requires less computation power than QM methods. It allows the modelling of large
molecular systems. However, MM energies have little meaning as absolute values
and can be used rather to compare relative energies obtained for several molecular
structures [46]. Moreover, MM approaches often cannot succeed with molecular
systems where electronic interactions are dominant, including π-π-stacking interactions which are the basis of most ligands binding to G-quadruplex DNA structures.
In this case, QM calculations should be used to obtain accurate results. At the same
time, despite some severe intrinsic limitations in MD approaches, base stacking can
be reasonably approximated based on well-calibrated force fields [47].
It should be noted that the molecular mechanics provide only a static view of
the flexible molecular system. The most common approach used for the simulation
of biomolecules motion on the atomic level is the molecular dynamics (MD) [48].
The forces acting on atoms are usually calculated here using MM methods. MD can
provide information on the possible conformations and dynamics of the system, as
well as its thermodynamic parameters.
MD simulations have some intrinsic limitations, e.g. force-field imperfections
and often insufficient simulation times. Nevertheless, MM-based methods, including MD, have become very popular research tools in biochemistry and drug design,
including the studies on G-quadruplexes and G4-ligand complexes. MD simulation of G-quadruplex structure and dynamics [49–52] and the interactions of quadruplexes of various topologies with cations [49, 50, 53–56] and low-molecular
organic compounds [50–52, 57–63] has been widely used to understand the basic
properties of quadruplex DNA and to improve the recognition of quadruplexes by
small molecules in the design of efficient G4 ligands, as well as to complement
available experimental data (see e.g. [25] and references therein). Such common
approaches to modern drug design as molecular docking and virtual screening have
also been successfully applied to the development of potential G4 ligands as antitumor agents [57, 62–67]. However, in the present review we will mainly concentrate
on purely quantum chemical (i. e. quantum mechanical) methods being currently
used to model G-quadruplex DNA and quadruplex-drug complexes with high accuracy. Moreover, non-QM methods of molecular modelling of G4 structures and
G4-ligand complexes, in particular MD approaches, have been recently reviewed
in a number of works [50, 52, 60, 63–66], including a detailed methodology review
by Haider and Needle [68].
M. Ilchenko and I. Dubey
mechanics (MM) [45]. QM methods provide more accurate modelling results than
MM-based approaches, although they are generally much more demanding in terms
of a computation power. It is now widely recognized that both methods reinforce
one another in an attempt to understand chemical and biochemical behaviour of biomolecules at the molecular level. From a practical point of view, the complexity of
the system, time limits, available computation resources and other limiting factors
determine which method is feasible [46].
MM may be used to model biomacromolecular systems to which even semiempirical QM calculations can be applied effectively. In MM, molecular motions
are determined by the masses of atoms and the forces acting on them, whereas wave
functions or electron densities are not computed. MM is widely used in chemistry
and biochemistry to obtain molecular models since this approach is much faster and
requires less computation power than QM methods. It allows the modelling of large
molecular systems. However, MM energies have little meaning as absolute values
and can be used rather to compare relative energies obtained for several molecular
structures [46]. Moreover, MM approaches often cannot succeed with molecular
systems where electronic interactions are dominant, including π-π-stacking interactions which are the basis of most ligands binding to G-quadruplex DNA structures.
In this case, QM calculations should be used to obtain accurate results. At the same
time, despite some severe intrinsic limitations in MD approaches, base stacking can
be reasonably approximated based on well-calibrated force fields [47].
It should be noted that the molecular mechanics provide only a static view of
the flexible molecular system. The most common approach used for the simulation
of biomolecules motion on the atomic level is the molecular dynamics (MD) [48].
The forces acting on atoms are usually calculated here using MM methods. MD can
provide information on the possible conformations and dynamics of the system, as
well as its thermodynamic parameters.
MD simulations have some intrinsic limitations, e.g. force-field imperfections
and often insufficient simulation times. Nevertheless, MM-based methods, including MD, have become very popular research tools in biochemistry and drug design,
including the studies on G-quadruplexes and G4-ligand complexes. MD simulation of G-quadruplex structure and dynamics [49–52] and the interactions of quadruplexes of various topologies with cations [49, 50, 53–56] and low-molecular
organic compounds [50–52, 57–63] has been widely used to understand the basic
properties of quadruplex DNA and to improve the recognition of quadruplexes by
small molecules in the design of efficient G4 ligands, as well as to complement
available experimental data (see e.g. [25] and references therein). Such common
approaches to modern drug design as molecular docking and virtual screening have
also been successfully applied to the development of potential G4 ligands as antitumor agents [57, 62–67]. However, in the present review we will mainly concentrate
on purely quantum chemical (i. e. quantum mechanical) methods being currently
used to model G-quadruplex DNA and quadruplex-drug complexes with high accuracy. Moreover, non-QM methods of molecular modelling of G4 structures and
G4-ligand complexes, in particular MD approaches, have been recently reviewed
in a number of works [50, 52, 60, 63–66], including a detailed methodology review
by Haider and Needle [68].