Biology Reference
In-Depth Information
To our great satisfaction, we could locate thermodynamic data for most of the
docking benchmark, although some complexes had to be replaced by homologs that
also satisfied condition (a). The K d values, which cover a wide range from 10 −5 to
10 −14 M, are derived from either a titration, mostly ITC (isothermal titration calorimetry),
or from the binding kinetics (surface plasmon resonance); a few are from enzymic
inhibition. The present version of the structure/affinity benchmark comprises 144
complexes, and includes nine pairs that have very similar structures and very differ-
ent affinities, due to differences in conformation or in sequence. For each entry, the
benchmark cites PDB codes for the complex and its components, the K d and ΔG d
values with the method and experimental conditions of their measurement, and the
relevant literature references (Kastritis et al. 2011 , and Table 5.1 ).
5.7
Conclusion
The major achievement of protein-protein docking has been its contribution to our
understanding of macromolecular interaction. Docking simulations demonstrate
that the shape and chemical complementarity of the molecular surfaces is the major
determinant in rigid-body recognition, which is a valid approximation in a number
of biological systems. Then, docking has a high predictive value, confirmed by
CAPRI and by experiments in which novel interactions are rationally designed
de novo . However, many processes of great biological importance rely on flexible
recognition, in which case the molecular surfaces become complementary only as a
result of conformation changes. The CAPRI targets that display flexible recognition
have stimulated new developments in the field of docking. Albeit still be far from
routine, methods to predict and simulate conformation changes have reached the
stage where they can produce useful models, and this has relevance to other fields.
In structural biology, much effort is made to fit the atomic resolution structure of
assembly components into lower resolution images from cryo-electron microscopy,
or an envelope derived from small-angle X-ray scattering, while allowing the structure
to change. This is a typical flexible docking problem, to which some docking algo-
rithms have already been applied. In drug design, the target proteins often make
other interactions than the one of interest. This may induce conformation changes
and allosteric effects that should be taken into account in the design procedure.
Similarly, computational biologists may want to study how protein folding is
affected by external interactions, in a homodimer for instance. Beyond the structure,
we want to understand what governs the specificity of macromolecular recognition
and the stability of protein assemblies. This implies that we should be able to model
the thermodynamics and the mechanism of the association reaction. The recent
attempt to predict affinity within the CAPRI experiment suggests that present force
fields are inadequate, and new methods must be developed. The structure/affinity
benchmark assembled on this occasion should help biophysicists to correlate func-
tion to structure, and remind them that the structure may change as new interactions
are formed.
Search WWH ::




Custom Search