Protein structure space

As outlined above the chemical structure space accessible to small drug-like molecules is so vast that it cannot be covered by chemical synthesis in a comprehensive and meaningful manner. During evolution nature herself has explored only a tiny fraction of chemical space in the biosynthesis of low molecular weight natural products. The same is true for the evolution of the targets bound and modulated by natural products, for example proteins. It has been estimated that during the evolution of a protein consisting of about 100 amino acids only a tiny fraction of all amino acid combinations could have been biosynthesized. However, in protein evolution, structure is even more conserved than the sequence since similar structures can be formed by very different sequences. Thus the protein structure space explored by nature is limited in size. ... 

From classifications representative lists for threading as well as assessment criteria for prediction results can be obtained. In addition, involved prediction methods can take advantage of a global view of protein structure space to better discriminate between the available alternatives than can be done on alignment scores of a sequence to an individual fold alone. We expect more comprehensive global knowledge of protein structure space to significantly improve protein structure prediction accuracy, in the future. 

The goal of structural genomics projects is to solve experimental structures of all major classes of protein folds systematically independent of some functional interest in the proteins [238, 239]. The aim is to chart the protein structure space efficiently. Functional annotations and/or assignment are made afterwards. 

Illustration of the PROTOMAP approach and an example of a PROTOMAP graph of protein structure space. Nodes represent sequence clusters edges represent proximities between clusters in sequence. Lightly and darkly shaded nodes... 

The three major protein structure classifications are SCOP, CATH, and DDD. SCOP is derived manually and is recognized as a valuable resource of detailed evolutionary information. CATH provides useful geometric information. It also introduces the concept of architecture, which reveals broad features of the protein structure space. CATH relies on partial automation and as such is subject to inaccuracies introduced by fixed thresholds. The DDD is a fully automatic classification continually updated. It is not as popular as SCOP and CATH, probably because its automatic levels are not as intuitive and require more input from the users to be interpreted. 

DA Elmds, M Levitt. Exploring conformational space with a simple lattice model for protein structure. J Mol Biol 243 668-682, 1994. 

All pictorial representations of molecules are simplified versions of our current model of real molecules, which are quantum mechanical, probabilistic collections of atoms as both particles and waves. These are difficult to illustrate. Therefore we use different types of simplified representations, including space-filling models ball-and-stick models, where atoms are spheres and bonds are sticks and models that illustrate surface properties. The most detailed representation is the ball-and-stick model. However, a model of a protein structure where all atoms are displayed is confusing because of the sheer amount of information present (Figure 2.9a). 

The folding of a single polypeptide chain in three-dimensional space is referred to as its tertiary structure. As discussed in Section 6.2, all of the information needed to fold the protein into its native tertiary structure is contained within the primary structure of the peptide chain itself. With this in mind, it was disappointing to the biochemists of the 1950s when the early protein structures did not reveal the governing principles in any particular detail. It soon became apparent that the proteins knew how they were supposed to fold into tertiary... 

Proteins derive their powerful and diverse capacity for molecular recognition and catalysis from their ability to fold into defined secondary and tertiary structures and display specific functional groups at precise locations in space. Functional protein domains are typically 50-200 residues in length and utilize a specific sequence of side chains to encode folded structures that have a compact hydrophobic core and a hydrophilic surface. Mimicry of protein structure and function by non-natural ohgomers such as peptoids wiU not only require the synthesis of >50mers with a variety of side chains, but wiU also require these non-natural sequences to adopt, in water, tertiary structures that are rich in secondary structure. 

Figure 5-1. Ramachandran plot of the main chain phi (< ) and psi (T) angles for approximately 1000 nonglycine residues in eight proteins whose structures were solved at high resolution. The dots represent allowable combinations and the spaces prohibited combinations of phi and psi angles. (Reproduced, with permission, from Richardson JS The anatomy and taxonomy of protein structures. Adv Protein Chem 1981 34 167.)...

The term biology-oriented synthesis (BIOS) [45] has been used to describe the design of compound libraries based on biologically relevant chemical space [46]. The areas in protein structures that participate in productive protein-ligand interactions have been, for the most part, already defined by natural products and drugs. Thus libraries inspired by natural products and other bioactive molecules are expected to have a higher probability of biologically activity than randomly synthesized molecules [47,48]. 

The unit cell considered here is a primitive (P) unit cell that is, each unit cell has one lattice point. Nonprimitive cells contain two or more lattice points per unit cell. If the unit cell is centered in the (010) planes, this cell becomes a B unit cell for the (100) planes, an A cell for the (001) planes a C cell. Body-centered unit cells are designated I, and face-centered cells are called F. Regular packing of molecules into a crystal lattice often leads to symmetry relationships between the molecules. Common symmetry operations are two- or three-fold screw (rotation) axes, mirror planes, inversion centers (centers of symmetry), and rotation followed by inversion. There are 230 different ways to combine allowed symmetry operations in a crystal leading to 230 space groups.12 Not all of these are allowed for protein crystals because of amino acid asymmetry (only L-amino acids are found in proteins). Only those space groups without symmetry (triclinic) or with rotation or screw axes are allowed. However, mirror lines and inversion centers may occur in protein structures along an axis. 

---