TRIAD database

Programs of Dr. Paul A. Bartlett to convert Cambridge Structural Database to one with bond vectors and to search the latter for specified vector relationships. UNIX workstations. TRIAD database of more than 400,000 energy-minimized tricyclic structures for automated design and ILIAD database of more than 100,000 energy-minimized linking structures in MacroModel, CAVEAT, MDL, SYBYL, and PDB formats. Silicon Graphics, IBM RS/6000, and Sun. 

The TRIAD database was designed to provide ideas for rigid templates, and it is clearly not optimal for problems in which large distances have to be spanned, in which a simple bridging unit is sought (e.g., the thermolysin problem above), or in which ease-of-synthesis is an important criterion from the start of the process. To address these issues, we created a different database, which we call ILIAD (to rhyme with TRIAD, of course, and rationalized as Inter-Linkers for Automated Design...). The molecules in... 

Search the Enzyme Structure Database for y-chymotrypsin active site (by the aid of the active-site-modified enzyme or active-site-specific inhibitor-enzyme complex) to identify and depict (save pdb file) the catalytic triad of y-chymotrypsin. 

In the past, considerable significance was attached to the syn-carboxylate interaction. Candour (1981) suggested that the syn electron pair may be more basic than the anti pair by a factor 10 -10. This seemed to explain the enhanced basicity of imidazole. However, studies of databases (Allen and Kirby, 1991) and model compounds (Zimmerman et ai, 1991) indicate that the difference in the relative basicities between syn- and anti-carboxylates is marginal (0.4—0.6 units). The hPL triad and the Asp-His couple in the active center of phospholipase A2 appear to support this view. Thus, the preference toward syn-type bonds observed in most serine proteinases may be due to packing effects in the active centers rather than to stereoelectronic effects. 

Especially for enzymes, common mechanisms and associated active sites are often observed. This allows searching for appropriate 3D patterns in a database of structures and for checking models to contain such a pattern (for a well known example, see SER-HIS-ASP serine protease catalytic triad [262]). 

Another possibility is to search for constellations of amino acids that can confirm a similar function even when the proteins containing them adopt completely different folds. Probably the best known example of this type of similarity is the serine protease catalytic triad (Ser, His, Asp), which occurs in at least 10 different folds. Several approaches have been developed to search for these, including PINTS [42] (http //pints.embl.de), the Catalytic Site Adas [43] (http //www.ebi.ac.uk/thomton-srv/databases/CSA), and Spasm/Rigor [44]. For some protein structures these similarities can be very illuminating as to function by either finding a convergendy evolved site on a new fold [45] or confirming a prediction of function based on a weak similarity to a known fold [46,47]. 

Figure 8.9. Output of a PASTA search, (a) Hit list from a PASTA search with human histidine triad (HIT) protein (SWISS-PROT P49789) as the query against the swissprot database. The search was performed using ktup = 1. (b) Optimal local alignment of the query to one of the database entries (marked by arrow in hit list) containing the sequence of rat galactose-1-phosphate uridylyltransferase (GalT). Although the sequence similarity is weak, these proteins have been shown to share structural similarity.

To demonstrate the improved sensitivity of the PSl-BLAST approach, the sequence of histidine triad (HIT) protein was used as a database search query. Simi-... 

In the following an overview on experimentally determined base polyad containing nucleic add structures deposited at the PDB is given, see Appendix (Table 2). In many cases the triplex or tetraplex fold formed jBrom on two, three or foiu strands is the major determinant of a structure. The compilation in Table 2 is int ded to be comprehensive for this structure type. On the other hand, especially in RNA structures only a few or even only one pol (s) may occur. Comprehensiveness for these structures would require a systematic anal is of dl currently known RNA structures. Recently, a database of non-canonical RNA base pairs has been set up. In addition to base pairs it also lists 331 base triads and 2S base tetrads (Mdch 7,2002). We have done a few random checks of the database and noticed that it represents a rather good starting point for further analysis but is not comprehensive. Table 2 includes only a few representative cases of polyads in RNA structures. 

While Bioster appears to be without predecessor, several commercial and in-house databases and software tools dealing with bioisosterism and other common replacement techniques have been developed during the past decade (for details, see relevant publications [16-21] and reviews [5, 22-26]). Of the earlier related efforts, the CAVEAT program, based on large structure collections of tricyclic rings (TRIAD) and... 

Occurrence of complexes with metaNigand multiple bonds by triad. Data from the Cambridge Structural Database in 2000. Adapted with permission from Cundari, T. R. Chem. Rev. 2000,100,807. 

The glutaminase domain consists of two subdomains initially identified by sequence comparisons of several CPSases and other amidotransferases (36-38). The carboxyl half of the domain is homologous to the trpG or triad-type amidotransferases (39) and contains several conserved residues implicated in catalysis. The amino half has a unique sequence that does not resemble any other protein in the database. The GLN subdomains were clearly identifible in the E. coli CPSase x-ray structure (26),... 

---