Figure 3

Annotation and composition of multimodular proteins. (a) Annotation is complicated by multimodular proteins. An E. coli protein (gi1786183) contains two modules, an N-terminal aspartokinase and a C-terminal homoserine dehydrogenase. Two single module proteins from L. lactis and B. halodurans (gi12723655 and gi10174117) align to the N-terminal aspartokinase module of the E. coli protein. Based on the sequence alignments, both of these proteins should be annotated as aspartokinases. However, errors are seen in the annotation of the L. lactis and B. halodurans proteins stemming from transfer of functions between multimodular proteins and partially aligned sequences without taking into account the alignment regions. (b) Different combinations of modules are seen in multimodular proteins of different organisms. While aspartokinase is fused to homoserine dehydrogenase in E. coli it is fused to DAP decarboxylase in X. fastidiosa. In both organisms the fusions are between enzymes of metabolic pathways, threonine biosynthesis for E. coli and lysine biosynthesis in X. fastidiosa.