Promise: a new database of information on prosthetic centres and metal ions in protein active sites.

K.N.Degtyarenko1, A.C.T.North and J.B.C.Findlay combination of the ENZYME and COMPOUND databases. It contains information such as structural formulae of low molecuDepartment of Biochemistry and Molecular Biology, University of Leeds, lar mass biological ligands: enzyme prosthetic centres, coLeeds LS2 9JT, UK enzymes, substrates and products and links to the corresponding 1To whom correspondence should be addressed protein sequences. A number of electron-transfer proteins, e.g. ferredoxins and cytochromes, are also considered if they are involved in an enzyme reaction. Structural or bibliographic The PROMISE (Prosthetic centres and metal ions in protein information is not provided. active sites) database project has been launched to gather Thus, there has been no single database which is focused together comprehensive sequence, structural, functional and on protein active site structure and on the relationship between bibliographic information on proteins which possess prosthetic prosthetic centre and protein molecule, combining the relevant centres, with an emphasis on active site structure and function. sequence, structural and physico-chemical information. In an PROMISE version 1.0 comprises data on iron-containing attempt to create such a database, we have constructed proteins. PROMISE and have now made available version 1.0, which comprises data on iron-containing proteins. An important Background characteristic of PROMISE is that it has been designed from the outset for interrogation via the WorldWide Web and There are several regularly updated, publicly accessible dataincorporates hypertext links to other relevant databases. bases which deal specifically with protein domains. PROSITE (Bairoch et al., 1996) is based on heuristically defined sequence Structure of PROMISE entries motifs which are predominantly represented as short regular expressions, defining the positions of residues that are invariant Entries in the PROMISE database follow a hierarchy: at the top level are major groups (such as haem proteins and iron– or of limited structural class; several PROSITE entries are also represented by sequence profiles. PRINTS (Attwood et al., sulphur proteins); these consist of classes (such as cytochromes, haem peroxidases and Fe2S2 proteins), which in turn consist 1996) also is based on heuristically and objectively defined motifs but uses fingerprints, often of several distinct motifs of families (such as Class II cytochromes c, animal haem peroxidases and adrenodoxin-type ferredoxins). The classificaand which are more robust than PROSITE regular expressions. BLOCKS (Pietrokovski et al., 1996) is essentially based on tion is not rigid and some intermediate or alternative levels could easily be introduced. the PROSITE domains but uses an extended representation of short conserved regions (blocks) for the sequences. Both A typical group entry (e.g. iron–sulphur proteins) contains: · Classification by function PROSITE and PRINTS entries are annotated and contain bibliographic information and cross-references to sequence · Classification by prosthetic group type and/or coordination and relevant class entries in the PROMISE database (e.g. and structural databases. SCOP (acronym for Structure Classification Of Proteins) is a Fe2S2 proteins) · Classification by type and number of prosthetic groups database facilitating understanding of and access to information available for protein three-dimensional (3-D) structures · Relevant entries in motif databases: PRINTS, PROSITE and BLOCKS (with hypertext links to them) (Murzin et al., 1995). The classification of proteins in SCOP is based on evolutionary relationships and on the characteristics · Relevant family entries in the PROMISE database · Classification by type of HET (hetero) groups in the PDB of their 3-D structures. Monodomain proteins are treated as a whole, whereas the domains in multidomain proteins are · References (with hypertext links, where available) Selection of the link to the Fe2S2 protein class displays: usually classified individually. The ‘place’ of a given protein in the classification sequence family/superfamily/fold/class is · Table summarizing properties of the prosthetic group typical for the class: its chemical structure and formal oxidation states unambiguously defined. CATH (acronym for Class, Architecture, Topology, Homology) is a hierarchical classification of · Short description of protein class · Relevant family entries in the PROMISE database (e.g. protein structural relationships derived using a combination of automatic and manual methods (Michie et al., 1995). CATH and adrenodoxin) · Relevant entries in motif databases: PRINTS, PROSITE and SCOP have some differences in definition of their hierarchical levels. In addition, for each entry CATH provides a link to a BLOCKS (with hypertext links to them) · References (with hypertext links, where available) protein–ligand interaction graphical summary constructed with the LIGPLOT program (Wallace et al., 1995). Selection of the link to the adrenodoxin family gives (Figure 1): · Table summarizing properties of prosthetic group(s): chemThe Ligand Chemical Database (Suyama et al., 1993) is designed to provide a linkage between chemical and biological ical structure, coordination, amino acid ligands of metal ion(s), formal oxidation states aspects of life in the light of enzymatic reactions and is a