7 Combinatories and Topology of theβ-Sandwich andβ-Barrel Proteins 1437.4.12 The Two Hydrophobic Tetrahedrals Present the Structural
Invariant of Barrel Proteins
The relative positions of the hydrophobic tetrahedral in the two subsheets can
be described by two parameters: angle (φ) between the hydrogen bonds that
connect the residues at positions S and T in one tetrahedral and positions
P and R in the other, and distance between the “centers of gravity” of the
two tetrahedrals (Fig. 7.4). The pair of tetrahedral makes up the invariant
substructure of the barrel proteins. Each barrel protein contains at least one
pair of tetrahedral, but some contain several such pairs with approximately
same distances between centers of gravity and angles.
7.5 Conclusion
The investigation carried out thus far allowed us to find the invariant sequence
and structural invariants in the vast diversity of beta structures. Analysis of
the supersecondary structures revealed the constraints in the arrangements of
the strands. It supports the conclusion that proteins which grouped together
on the basis of common architecture like sandwich-like or barrel-like proteins
have commonality on the level of supersecondary structure.
In both groups of proteins – sandwich and barrel structures the arrange-
ment of strands gives two invariants substructure – hydrophobic tetrahedrals.
In fact, the tetrahedral, that make up of residues of neighboring strands in
the beta sheet can be considered as the structural unit of the beta structures.
In the sandwich-like structure two tetrahedrals form interlock. The residues
of the interlock lie at the center of the interface between the beta sheets and
form the common geometrical core of sandwich proteins.
In the barrel structure two tetrahedrals form another geometrical figure –
tetrahedral. Residues of the tetralock lie at the edge of two subsheets. We can
suggest that these residues are responsible for the distortion of theβ-sheet in
the barrel-like structure, which results in the formation of two subsheets and
leads to the closed structure. The tetralock can be considered as the common
geometrical core of the barrel proteins.
Analysis of a broad groups of proteins, such as sets of superfamilies, yields
a set of sequence determinants of a group of nonhomologous proteins. These
sequence determinants form the basis of computer algorithm for classification
of novel proteins.
A direct corollary of our approach is that complexity of protein sequence
search algorithms and 3D structure predictions can be dramatically reduced:
instead of carrying out searches with whole protein sequences, we may now
carry out searches with predefined sets of several key residues. This is analo-
gous to searching for a suspect by his fingerprints, rather than by a long list
of nonunique descriptors. Our data on sandwich-like proteins shows that the