In order to improve plant resistance to phytopathogenic fungi, hevein-like peptides have been expressed in tobacco [33] and [52], tomato [31] and Arabidopsis plants [51] and [52]. These peptides can therefore be included in the selective class of promiscuous peptides, where a peptide or a peptide

family can have multiple activities under different environmental Serine Protease inhibitor conditions [16]. In the case of family promiscuity, the multiple functions are related to different exposed residues in the same scaffold, which in turn are stabilized by their disulfide bonds [16]. Due to the conservation of disulfide bonds, these classes are good targets for mining protein databases. This kind of approach has been applied to cyclotides [42] and defensins [65] and has revealed novel aspects about them. Identification of novel hevein-like Alectinib peptides may bring to light new possibilities for their use as well as knowledge about their functions. To this end, this work reports the identification of novel hevein-like

peptide precursors through computational methods. Sequences from plants and also from a phytopathogenic fungus were identified and their structures and possible functions were predicted. The results presented here may also suggest new prospects for hevein domain interactions that are applicable to chitin studies. The data set of hevein-like peptides was constructed by using an automatic search system. Briefly, the system here proposed runs the Blast Sucrase software [2], reads its output, gets the retrieved sequences and subsequently runs Blast once more with these retrieved sequences. This process

was repeated until no novel sequences were obtained, as described by Zhu [65] with minor modifications. Additionally, the system was set to filter fragments and sequences larger than 130 amino acid residues. The initial sequence used for searching was the Ac-AMP2′s precursor, identified from Amaranthus caudatus [9] (UniProt ID: P27275), since it has antimicrobial and antifungal activities. The search was performed in SwissProt database [56]. The final data set was manually curated, selecting only the sequences annotated as fungicidal. The software Pratt 2.1 [27] was used for pattern identification into the hevein-like data set, using the default parameters (number of consecutive wild cards, maximum number of flexible spacers and maximum number of consecutive wild cards set to five, two and two, respectively). The pattern with the highest fitness value was used for searching against NCBI’s non-redundant protein database (NR), through regular expressions and PERL scripts. The script was set to select sequences annotated as hypothetical, unnamed and/or unknown proteins, restricting the maximum size to 130 amino acid residues.