Coarse-Grained Modeling and Simulations of Thermoresponsive …
39
To find alternative computational approaches, there have been numerous efforts
made toward developing coarse-grained (CG) models. CG model development for
systems with dominant directional interactions has been limited mostly to biomacromolecules. In particular, to capture the length scales and time scales of protein aggregation within simulations, many groups have developed CG models where two–three
CG beads represent the alpha-carbon, the side chain groups and the backbone amide
groups. Many models do not represent h-bonds explicitly; the directionality induced
by h-bonds is implicitly captured through prior knowledge of known secondary structures via “pseudo”/ “ghost” bonds and/or through dihedral bonded potentials along
the chain [6]. Polysaccharides (e.g., cellulose) are another class of macromolecules
with directional h-bonds along the chains [7]. Lattice-based statistical mechanical
models have been used to study cellulose, yielding agreement with both atomistic
MD simulations of crystalline cellulose and temperature-dependent infrared spectroscopy measurements. Several CG models have been developed for nucleic acids,
[8, 9] such as the three-site-per-nucleotide (3SPN) [10–12] and OxDNA [13, 14]
models. These models reproduce several experimentally observed structural, thermal
and mechanical properties of DNA and RNA. In the context of capturing h-bonds with
the least amount of chemical details, in some CG models small attractive CG beads
embedded within larger CG beads have been used to mimic directional hydrogen
bonds between the complementary bases. These “patchy particle” type CG models
successfully capture the thermodynamics of hybridization but miss the chemically
detailed secondary structure of the duplex. Examples of studies using such models
include work by Sciortino and coworkers [15–17], Travesset and coworkers [18], and
Jayaraman and coworkers [19–22]. These minimalist CG models have successfully
captured the melting curves for DNA [19, 20] as well as RNA folding thermodynamics [23] for system sizes that are experimentally relevant. The focus of this chapter
is on such minimalist CG models that we have developed recently for simulating
experimentally relevant systems of oligomers of nucleic acids (e.g., DNA, RNA,
etc.), collagen-mimetic peptides or collagen-like peptides and polymer nanocomposites. This chapter is organized in three sections with each section independently
focused on each of these topics. Each section has a background subsection where we
describe briefly relevant past literature on the topic, a subsection describing the CG
model, a subsection describing the key features of the simulation protocol and data
analyses, a subsection where we highlight some important results from our work,
and a sub section on limitations and future directions.
2 Oligomers of Nucleic Acids
2.1 Background
Deoxyribonucleic acid, DNA, is a biological macromolecule/polymer that stores
and transfers genetic information in living organisms. Each repeat unit/monomer in
39
To find alternative computational approaches, there have been numerous efforts
made toward developing coarse-grained (CG) models. CG model development for
systems with dominant directional interactions has been limited mostly to biomacromolecules. In particular, to capture the length scales and time scales of protein aggregation within simulations, many groups have developed CG models where two–three
CG beads represent the alpha-carbon, the side chain groups and the backbone amide
groups. Many models do not represent h-bonds explicitly; the directionality induced
by h-bonds is implicitly captured through prior knowledge of known secondary structures via “pseudo”/ “ghost” bonds and/or through dihedral bonded potentials along
the chain [6]. Polysaccharides (e.g., cellulose) are another class of macromolecules
with directional h-bonds along the chains [7]. Lattice-based statistical mechanical
models have been used to study cellulose, yielding agreement with both atomistic
MD simulations of crystalline cellulose and temperature-dependent infrared spectroscopy measurements. Several CG models have been developed for nucleic acids,
[8, 9] such as the three-site-per-nucleotide (3SPN) [10–12] and OxDNA [13, 14]
models. These models reproduce several experimentally observed structural, thermal
and mechanical properties of DNA and RNA. In the context of capturing h-bonds with
the least amount of chemical details, in some CG models small attractive CG beads
embedded within larger CG beads have been used to mimic directional hydrogen
bonds between the complementary bases. These “patchy particle” type CG models
successfully capture the thermodynamics of hybridization but miss the chemically
detailed secondary structure of the duplex. Examples of studies using such models
include work by Sciortino and coworkers [15–17], Travesset and coworkers [18], and
Jayaraman and coworkers [19–22]. These minimalist CG models have successfully
captured the melting curves for DNA [19, 20] as well as RNA folding thermodynamics [23] for system sizes that are experimentally relevant. The focus of this chapter
is on such minimalist CG models that we have developed recently for simulating
experimentally relevant systems of oligomers of nucleic acids (e.g., DNA, RNA,
etc.), collagen-mimetic peptides or collagen-like peptides and polymer nanocomposites. This chapter is organized in three sections with each section independently
focused on each of these topics. Each section has a background subsection where we
describe briefly relevant past literature on the topic, a subsection describing the CG
model, a subsection describing the key features of the simulation protocol and data
analyses, a subsection where we highlight some important results from our work,
and a sub section on limitations and future directions.
2 Oligomers of Nucleic Acids
2.1 Background
Deoxyribonucleic acid, DNA, is a biological macromolecule/polymer that stores
and transfers genetic information in living organisms. Each repeat unit/monomer in