A structural census of metabolic networks for E. coli

A structural survey of the Escherichia coli proteins occurring in metabolic networks in the KEGG database (release 19 of LIGAND) has been carried out. A measure of structural coverage of a network is defined and calculated fo r each network. Twentyfour networks have 50% or more of the enzyme steps as signed in E. coli and of these 21 have a structural coverage of 50% or more . For those proteins that have a region matching a SCOP domain 50% fall on or below the 30% sequence identity threshold and represent non-trivial comp arative modelling targets highlighting the need for experimental structure determination studies. The survey reveals the predominance of alpha/beta an d alpha + beta folds for enzymes involved in metabolic pathways and that th is general trend is maintained at the level of each pathway. The most popul ar superfamilies are coenzyme binding domains and are involved in the suppl y of energy to reactions. Although a few superfamilies are found in many pa thways, in general there is a specificity of a particular superfamily for a particular pathway. (C) 2001 Academic Press.