Computational Approaches in Peptide and Protein Design: An Overview - Peptide and Protein Design for Biopharmaceutical Applications - page 7

development of technology for co-crystallization of ligand-receptor com-
plexes (especially enzymes and their substrates/inhibitors), as well as
advances in X-ray crystallography of proteins. Seemingly, experimental
determination of ligand-receptor complexes by X-ray crystallography
abolishes the need for computational approaches. Indeed, detailed infor-
mation on both the recognition motif (3D pharmacophore) and the bind-
ing mode of the ligand within the receptor is readily available from the
X-ray structure of the complex. In reality, however, there are several
important limitations that still require emphasis on computational

approaches in target-based design.
First, many biologically active peptides (and over 30% of drugs in clinical
use [9]) act through interactions with GPCRs, which are integral membrane
proteins that include seven transmembrane helical stretches (TM helices)
connected by loops that form the intracellular (IC) and extracellular (EC)
domains, together with the fragments containing the N- and C-termini.
Being membrane proteins, GPCRs are extremely difficult to express and to
extract from the membrane in quantity, and have resisted chemical synth-
esis. Accordingly, X-ray structures are known presently only for four
GPCRs, namely the photoreceptor rhodopsin, the b2- and b1-adrenergic
receptors and the A2A adenosine receptor [10,11, 114, 115]. Therefore, the
only way to address ligand-receptor interactions involving other GPCRs
(the largest human gene family) for target-based design is to apply computa-
tional approaches to model (either by homology or by de novo approaches)
the 3D structure of GPCR.
Second, while many ligand-receiving sites in receptors feature more or
less well-defined 3D 'pockets' inside the protein globule (as, for example,
in enzymes), other recognition sites are formed by flexible loops protrud-
ing away from the bulk of the protein (as, for example, in antibodies or
GPCRs). In the latter case, one faces the same problem as in determining
the 3D structure of a flexible peptide: namely, even if the X-ray structure
of the loops in the receptor-ligand complex is resolved, the X-ray snap-
shot may capture only a single specific conformation out of several that
may be more characteristic for the given complex and more representa-
tive of a functional complex. For instance, the conformations of the IC
loops connecting TM helices in the five X-ray structures of rhodopsin
published so far drastically differ from one another. Again, determining
the set of plausible conformations of the loops in question requires use of
computational modelling.
Third, depending on the shape and rigidity of the receiving site of the
receptor, the binding modes of the ligand also may be nonunique. That may
be especially important for small ligands with medium affinity toward the