Genome mining demonstrates the widespread occurrence of gene clusters encoding bacteriocins in cyanobacteria

Abstract

Cyanobacteria are a rich source of natural products with interesting biological activities. Many of these are peptides and the end products of a non-ribosomal pathway. However, several cyanobacterial peptide classes were recently shown to be produced through the proteolytic cleavage and post-translational modification of short precursor peptides. A new class of bacteriocins produced through the proteolytic cleavage and heterocyclization of precursor proteins was recently identified from marine cyanobacteria. Here we show the widespread occurrence of bacteriocin gene clusters in cyanobacteria through comparative analysis of 58 cyanobacterial genomes. A total of 145 bacteriocin gene clusters were discovered through genome mining. These clusters encoded 290 putative bacteriocin precursors. They ranged in length from 28 to 164 amino acids with very little sequence conservation of the core peptide. The gene clusters could be classified into seven groups according to their gene organization and domain composition. This classification is supported by phylogenetic analysis, which further indicated independent evolutionary trajectories of gene clusters in different groups. Our data suggests that cyanobacteria are a prolific source of low-molecular weight post-translationally modified peptides.

Figure 1. The widespread distribution of putative bacteriocin gene clusters in cyanobacteria.

The neighbor-joining tree is based on concatenated 16S and 23S rRNA genes from 55 cyanobacterial genomes. The strains which have at least one bacteriocin gene cluster are indicated with a gray background. Phylogenetic analyses were conducted in MEGA4 by using the Maximum Composite Likelihood model and with 50000 bootstrap replications for each branch. The bootstrap values are shown next to the branches. Outgroup taxa Gammaproteobacterium HdN1, Bradyrhizobium japonicum USDA 110, and Escherichia coli UMN026 were used to root the tree, which is drawn to scale. Strains Arthrospira PCC 8005, Leptolyngbya valderiana BDU 20041, and Prochlorococcus marinus MIT 9202 are absent from this tree because they are partial genomes and have no complete rRNA genes.

Figure 2. Organization of cyanobacterial bacteriocin gene clusters.

The putatively identified gene clusters in this study were classified into seven groups (from I to VII) based on their gene organization and domain composition. ORF sizes and directions are shown in relative scale with color definition as precursor in red, ABC transporter in blue, HlyD in orange, SurA in green, LanM in pink, S8 peptidase-containing protein in yellow, other modification enzymes in purple, adjacent ORFs in black, and tRNA gene in light green. Domains involved in cyanobacterial bacteriocin production and modification are demonstrated within the ORFs with different colors, domain names are derived from the Conserved Domain Database . (I) An example structure of group I from strain Synechococcus PCC 7335, and the locus_tag of hlyD is S7335_4080. (II) from Anabaena variabilis ATCC 29413, and the locus_tag of hlyD is Ava_4382. (III) from Nostoc sp. 7120, and the locus_tag of hlyD is alr5148. (IV) from Nostoc punctiforme PCC 73102, and the locus_tag of hlyD is Npun_F5048. (V) from Nostoc punctiforme PCC 73102, and the locus_tag of hlyD is Npun_R1804. (VI) from Anabaena variabilis ATCC 29413, and the locus_tag of the S8 peptidase domain-containing protein is Ava_4226. (VII) from Trichodesmium erythraeum IMS101, and the locus_tag of hlyD is Tery_0894.

Figure 3. Sequence logos of double-glycine motif generated from cyanobacterial precursors in NHLP and N11P families.

There is a conserved region found near the peptide cleavage site with Gly-Gly motif from the precursor peptides. Relative frequency of acidic residues of the conserved sequences from families (A) NHLP and (B) N11P are demonstrated. This figure was generated by web-based software .

Figure 4. Diverse structures of cyanobacterial bacteriocin precursors from known families.

(A) Thirty-nine selected NHLP precursors are shown in a ClustalW alignment . The locus tag is given to the left of the sequence and the amino acid position is given on the right. The cleavage site of the leader peptides is indicated by an arrow. (B) Twenty-four selected N11P precursors shown in a ClustalW alignment. The coloring scheme and notation are identical to section A.

Figure 5. Evolutionary relationships of C39 peptidase domain-containing ABC transporters in cyanobacteria.

The phylogenetic analysis is based on C39 peptidase domain-containing ABC transporters in the seven gene cluster groups, except those with disrupted ORFs. The midpoint neighbor-joining tree was constructed by using MEGA4 with Poisson correction model and 50000 bootstrap replications for each branch. The tree is drawn to scale, and bootstrap values of major branches are shown. The name of each taxon is constituted by gene cluster group, accession number, and strain name of the C39 peptidase domain-containing ABC transporters. Major clades of the tree are composed of proteins distinctly from respective gene cluster groups and are named as the corresponding groups with different background colors. Independent evolutionary histories were inferred between the gene clusters in different groups. Proteins from group IV scattered among clades of III and V+VII, and are highlighted in orange. The group I protein found in clade II is shown in green.