STRUCTURAL INVESTIGATION AND CLASSIFICATION OF ENNIATIN SYNTHETASE FROM DIFFERENT FUSIARIUM SPECIES

Fusarium genus is able to produce several metabolites including the emerging mycotoxins beauvericin (BEA) and enniatins (ENs). Due to their ionophoric property, BEA and ENs exert many biological properties, including antimicrobial, insecticidal, and cytotoxic activity in human cell lines, so that they are recently proposed as novel anticancer drugs. BEA and ENs are cyclic hexadepsipeptides with an alternating sequence of three N-metyl-L-amino acids and three D-?-hydroxyisovaleric acids; in BEA the amino acid residues are aromatic N-metyl-pheylalanines, whereas in ENs the amino acid residues are aliphatic N-metyl-valine, -leucine or -isoleucine. Both mycotoxins are synthesized by the multifunctional enzyme enniatin synthetase (ESYN1) representing hybrid system of peptide synthetase and S-adenosyl-L-methionine-dependent N-methyltransferase. Several Fusarium species have been reported to produce ENs, BEA or both, and our hypothesis is that the different production profile depends on the esyn1 sequence. Our aim was to investigate this relation by mean of a bioinformatics approach, based on structural investigations. The esyn1 sequences of 18 Fusarium isolates belonging to 8 different species were extracted from published and unpublished genomes by BLASTN search using as query the available sequences of Fusarium scirpi and Fusarium proliferatum. The protein sequences were predicted using the Sequence Translation tools of the EMBOSS Programs (EMBL-EBI), manually curated using exon/intron boundary predictions from SpliceView (http://bioinfo4.itb.cnr.it/), and confirmed by sequencing the RT-PCR products. The selected sequences have been processed by the web server RaptorX to obtain structural predictions by homology modelling and threading methods. While the full sequences have been submitted for prediction, it has been accomplished only for part of the whole enzyme. In particular, the region containing the N-methyltransferase activity has been not completely structured. A comprehensive analysis of the predicted structures has allowed identifying common structural domains, which have been compared by considering their dihedral backbone angles and by using the tool T-PAD, developed to characterize the protein flexibility and identify hot spot residues responsible for hinge motions. The residue-by-residue flexibility profiles have been analysed by multivariate analysis, to produce a classification by Principal Component Analysis, followed by hierarchical clustering. We have thus identified the conserved structural regions and the pivotal residues responsible for the structural variability. The common domains for each of the 18 esyn1 sequences have been grouped in clusters, and the residues responsible for the classification have been singled out. Overall, we have achieved a comprehensive view of the structural features of the analysed esyn1 sequences, where the structural variability has been related to the sequence variability, and interpreted in terms of functionally relevant movements. Chemical analysis are ongoing to confirm the hypothesized correlation between structural predictions and metabolic profiles.