Cyclotides are a family of head-to-tail cyclized peptides containing three conserved disulfide bonds, in a structural scaffold also known as a cyclic cysteine knot. Due to the high degree of cysteine conservation, novel members from this peptide family can be identified in protein databases through a search through regular expression (REGEX). In this work, six novel cyclotide-like precursors from the Poaceae were identified from NCBI's non-redundant protein database by the use of REGEX. Two out of six sequences (named Zea mays L and M) showed an Asp residue in the C-terminal, which indicated that they could be cyclic. Gene expression in maize tissues was investigated, showing that the previously described cyclotide-like Z. mays J is expressed in the roots. According to molecular dynamics, the structure of Z. mays J seems to be stable, despite the putative absence of cyclization. As regards cyclotide evolution, it was hypothesized that this is an outcome from convergent evolution and/or horizontal gene transfer. The results showed that peptide screening from databases should be performed periodically in order to include novel sequences, which are deposited as the databases grow. Indeed, the advances in computational and experimental methods will together help to answer key questions and reach new horizons in defense-related peptide identification.

No abstract

Antimicrobial peptides (AMPs) are small molecules present in all living beings. Despite their huge sequence variability, AMPs present great structural conservation, mainly in cysteine‐stabilized families. Moreover, in non‐model plants, it is possible to detect cysteine‐stabilized AMPs (cs‐AMPs) with different sequences not covered by conventional searches. Here, we described a threading application for cs‐AMP identification in the non‐model arum lily (Zantedeschia aethiopica) plant, exploring the spathe transcriptome. By using the predicted proteins from the Z. aethiopica transcriptome as our primary source of sequences, we have filtered by using structural alignments of 12 putative cs‐AMP sequences. The two unreported sequences were submitted to PCR validation, and ZaLTP7 gene was confirmed. By using the structure alignments, we classified ZaLTP7 as an LTP type 2‐like. The successful threading application for cs‐AMP identification is an important advance in transcriptomic and proteomic data mining. Besides, the same approach could be applied to the use of NGS public data to discover molecules to combat multidrug‐resistant bacteria.