Using Data Mining Techniques to Probe the Role of Hydrophobic Residues in Protein Folding and Unfolding Simulations - Evolving Application Domains of Data Warehousing and Mining - page 261

Figure 2. Schematic secondary structure representations of the WT-TTR monomer along a typical 10 ns
molecular dynamics protein unfolding simulation. The structure at the left represents the experimental
crystal structure (PDB entry 1tta). The eight β-strands are labelled A (residues 12 to 18), B (residues
28 to 35), C (residues 41 to 49), D (residues 54 to 55), E (residues 67 to 73), F (residues 91 to 97), G
(residues 104 to 112) and H (residues 115 to 123). The simulated time along the MD trajectory is indi-
cated underneath each structure

associated with the formation of amyloid deposits
of different proteins. Human transthyretin (TTR)
is a homotetrameric protein involved in amyloid
pathologies such as familial amyloid polyneu-
ropathy (FAP), familial amyloid cardiomyopathy
(FAC) and senile systemic amyloidosis (SSA).
Each TTR monomer consists of 127 amino-acids
arranged in a well-characterized β-sandwich to-
pology comprising β-strands DAGH and CBEF
(Figure 2). Several single point mutations enhance
the amyloidogenicity of TTR and lead to disease.
Of all known mutations, Leu55→Pro (a proline
replacing a leucine in position 55) is one of the
most amyloidogenic, and Val30→Met (a methi-
onine replacing a valine in position 30) is one of
the most prevalent (Brito, 2003).
In recent years, we have been dedicating
particular attention to the characterization of
the molecular species present in the aggregation
pathway of transthyretin, using both experimen-
tal (Quintas, 2001) and computational (Correia,
2006; Rodrigues & Brito, 2004; Rodrigues, 2009)
methodologies. To explore the unfolding routes
of monomeric species of TTR, high temperature
molecular dynamics simulations were performed

in our laboratory. MD high temperature simula-
tions are able to capture the essence of the un-
folding process without changing the unfolding
behaviour of the protein (Day, 2002), and allowing
for shorter simulation times (tens of ns). As an
example, 10 representative structures along one
unfolding trajectory are shown in Figure 2, where
it is clear the loss of native secondary structure
and the loss of the β-sandwich topology of TTR,
as the simulation progresses.

Md Simulations details

We performed five independent molecular dynam-
ics unfolding simulations of wild-type transthy-
retin (WT-TTR). For all simulations, the initial
atom coordinates were taken from chain B of the
WT-TTR structure (PDB entry 1TTA; Hamilton
(1993)), and hydrogen atoms were added. The
final system also included solvent water molecules
and Na + Cl − ions. The initial atom velocities were
assigned by a random number generator using the
constraint of the Maxwell-Boltzmann distribution,
and were different for each simulation. The NAMD
(Phillips, 2005) molecular dynamics code was used