Database Reference
In-Depth Information
Ala25, Phe33, Leu58 and Ala91. On one hand,
association rules derived for Runs 1, 3 and 4,
describe associations between residues that show
a high increase of SASA values around the 4th ns
of simulation, but in these data sets Ala25, Phe33
and Ala91 start with SASA values close to 0% but
move almost immediately to positions of solvent
exposure around 30-40%. On the other hand, in
SASA variation profiles described by data sets 1, 4
and 5, the residue Pro11 moves rapidly to positions
of high exposure, and Leu58 fluctuates between
positions of low and medium exposure. Thus, it
is clear that some residues exhibit conservative
behaviour across all data sets, while others follow
alternative routes.
The results show that residues in β-strands
A, B, E and G of the WT-TTR monomer display
strongly correlated solvent exposure behaviour,
which is consistent with previous experimental
work by Liu (2000), where it is suggested that
β-strands A, B, E and G of WT-TTR undergo a
cooperative unfolding process, and that most of the
residues coupling β-strands B with E, and A with
G, play a critical role in the stabilization of TTR.
The distribution of these residues in the secondary
structure elements of WT-TTR leads us to believe
that these residues may play a central role in the
folding and unfolding processes of the protein.
Moreover, Hammarström & Carlsson (2000)
stated that detecting residual structures in unfolded
proteins can yield important clues for the identifi-
cation of initiation sites of protein folding. It has
been shown that a substantial number of studied
proteins possess residual structure in hydrophobic
regions clustered together in the protein core which
may work as seeds in the folding process. Thus,
the residues identified here using both hierarchi-
cal clustering and association rules may in high
probability be regarded as hydrophobic seeds in
the folding/unfolding process of TTR.
Another interesting result regarding the role
of residues Ala36, Asp39, and Glu42 on the sta-
bilization of the N-terminus of β-strand C was
obtained by the use of the hierarchical clustering
procedure. This issue has been discussed before
by other authors (Liu, 2000; Yang, 2006). It has
been reported that the unfolding of WT-TTR
starts with the disruption of β-strand D, followed
by the unfolding of β-strands F and H. Prior to
the complete unfolding of the protein, a residual
structure comprising the N-terminus of β-strand
C and β-strands B and E can be identified (Ro-
drigues, 2009; Yang, 2006). Particular attention
has been given to the stability of the N-terminus
of β-strand C. Liu (2000) using an experimental
approach, found that the residues in the C-terminus
of β-strand C are weakly protected from the sol-
vent, while residues in the N-terminus, Trp41 and
Glu42, are highly protected. These experimental
dIScuSSIon
The packing of amino-acid residues in proteins
is very important in determining protein stability
(Samanta, 2002). The solvent accessible surface
area (SASA) variation profile of an amino-acid
residue during unfolding simulations provides
a way to assess the changes in residue packing
upon protein unfolding. Hence, careful analysis
of SASA variation profiles and even more im-
portantly the identification of residues that vary
their SASA profiles in a synchronised manner,
may allow a better understanding of the unfolding
process in proteins, and in particular in amyloido-
genic proteins.
Here, we demonstrated the application of two
distinct data mining methods to identify clusters
and association of residues showing similar global
solvent exposure profiles across multiple MD
unfolding simulations. The two methods were
applied to five data sets originated from five inde-
pendent MD unfolding simulations of the protein
WT-TTR. The results reported demonstrate that
the methods are particularly helpful when used to
compare and contrast multiple data sets allowing
the identification of interesting correlations among
different sets of amino-acid residues.
Search WWH ::
Custom Search