NRPS-PKS: a knowledge-based resource for analysis of NRPS/PKS megasynthases

Abstract

NRPS-PKS is web-based software for analysing large multi-enzymatic, multi-domain megasynthases that are involved in the biosynthesis of pharmaceutically important natural products such as cyclosporin, rifamycin and erythromycin. NRPS-PKS has been developed based on a comprehensive analysis of the sequence and structural features of several experimentally characterized biosynthetic gene clusters. The results of these analyses have been organized as four integrated searchable databases for elucidating domain organization and substrate specificity of nonribosomal peptide synthetases and three types of polyketide synthases. These databases work as the backend of NRPS-PKS and provide the knowledge base for predicting domain organization and substrate specificity of uncharacterized NRPS/PKS clusters. Benchmarking on a large set of biosynthetic gene clusters has demonstrated that, apart from correct identification of NRPS and PKS domains, NRPS-PKS can also predict specificities of adenylation and acyltransferase domains with reasonably high accuracy. These features of NRPS-PKS make it a valuable resource for identification of natural products biosynthesized by NRPS/PKS gene clusters found in newly sequenced genomes. The training and test sets of gene clusters included in NRPS-PKS correlate information on 307 open reading frames, 2223 functional protein domains, 68 starter/extender precursors and their specific recognition motifs, and also the chemical structure of 101 natural products from four different families. NRPS-PKS is a unique resource which provides a user-friendly interface for correlating chemical structures of natural products with the domains and modules in the corresponding nonribosomal peptide synthetases or polyketide synthases. It also provides guidelines for domain/module swapping as well as site-directed mutagenesis experiments to engineer biosynthesis of novel natural products. NRPS-PKS can be accessed at http://www.nii.res.in/nrps-pks.html.

Figure 1

Flowchart depicting the organization of NRPS-PKS.

Figure 2

Depiction of the bacitracin cluster in NRPSDB. Each domain has been represented using a coloured box, the inter-domain linker regions have been depicted as filled black lines connecting the domains and the N- and C-terminal linkers have been represented as shaded lines in violet. The PUBMED button leads to the abstract of the publication describing the experimental characterization of the bacitracin cluster.

Figure 3

Screen dumps showing usage of NRPSDB for retrieval of the FASTA sequence, threading alignment and active site residues of a selected adenylation domain.

Figure 4

Schematic diagram showing the multiple template method for detection of NRPS domains. For each domain type a diverse set of sequences from the training set are used as templates. Domains in the query are identified by simple BLAST using these multiple templates and picking the best match.

Figure 5

Typical use of the query interface of NRPSDB for analysis of a hybrid NRPS/PKS cluster. Upon submission of a query sequence, the program gives a pictorial depiction of domain organization similar to those for characterized clusters in NRPS-PKS. The red-coloured linker between PKS and NRPS modules indicates the possible presence of domains which could not be detected by NRPS-PKS. Clicking on the image of a domain leads to further details. On selecting an adenylation domain in the query, the user can get its sequence, active site residues, predicted specificity, list of homologous domains along with their degree of similarity and specificity, and also information on homologous structures from PDB.

Figure 6

Screen shot depicting usage of the query interface of NRPS-PKS for extracting the active site residues of a selected condensation domain and obtaining its pair alignment with homologous condensation domains in NRPSDB. For each homologous C domain, the program also lists the amino acid pair it is condensing, and whether it is involved in a cyclization or epimerization reaction.

Figure 7

Features of NRPS-PKS for analysis of PKS domains in the hybrid NRPS/PKS query. The software provides interfaces for extracting its sequence in FASTA format, its pair alignment with any other AT domain in PKSDB or ITERDB, alignment with homologous structures in PDB and a list of homologous AT domains from ITERDB and PKSDB along with their active site residues. Iterative AT matches are marked with a pink ball and modular ones with a blue ball. From the alignment with homologous domains, the program also provides link to the depiction of the PKS cluster containing the homologous domain.