Phylum-wide comparative genomics unravel the diversity of secondary metabolism in Cyanobacteria

Abstract

Background: Cyanobacteria are an ancient lineage of photosynthetic bacteria from which hundreds of natural products have been described, including many notorious toxins but also potent natural products of interest to the pharmaceutical and biotechnological industries. Many of these compounds are the products of non-ribosomal peptide synthetase (NRPS) or polyketide synthase (PKS) pathways. However, current understanding of the diversification of these pathways is largely based on the chemical structure of the bioactive compounds, while the evolutionary forces driving their remarkable chemical diversity are poorly understood.

Results: We carried out a phylum-wide investigation of genetic diversification of the cyanobacterial NRPS and PKS pathways for the production of bioactive compounds. 452 NRPS and PKS gene clusters were identified from 89 cyanobacterial genomes, revealing a clear burst in late-branching lineages. Our genomic analysis further grouped the clusters into 286 highly diversified cluster families (CF) of pathways. Some CFs appeared vertically inherited, while others presented a more complex evolutionary history. Only a few horizontal gene transfers were evidenced amongst strongly conserved CFs in the phylum, while several others have undergone drastic gene shuffling events, which could result in the observed diversification of the pathways.

Conclusions: Therefore, in addition to toxin production, several NRPS and PKS gene clusters are devoted to important cellular processes of these bacteria such as nitrogen fixation and iron uptake. The majority of the biosynthetic clusters identified here have unknown end products, highlighting the power of genome mining for the discovery of new natural products.

Figure 1

Distribution of shared and orphan NRPS/PKS gene clusters detected in Cyanobacteria. The species tree was generated by a concatenation of twenty-nine conserved proteins using a Maximum Likelihood method. The clades a to h of the phylogenetic tree are supported by a bootstrap of ≥70%. The species tree is connected to the distribution pattern by lines. The lines are plain for complete genome and dashed for unfinished genomes and indicate the habitat of the strains: blue for fresh water, green for marine and black for other. On the distribution pattern, the cluster families involved in the biosynthesis of known product are first indicated in the grey shadowed area from CF-1 to CF-19, followed by the shared ones and encoding potentially unknown product from CF-20 to CF-61 with a white background while the last column indicates the number of orphans. The PKS clusters are indicated in blue, hybrid in red and NRPS in green. Each cluster is indicated by a dot, except for two gene clusters present in double copy in the genome of PCC 9339 that are indicated by a larger dot. The colored clusters on the same vertical line are related to each other and define a family of clusters. Details on the species tree and on the clusters and cluster families are available on Additional file 2: Figure S1 and Additional file 1: Table S1, S2 and S3.

Figure 2

Anatoxin-a biosynthetic gene cluster (CF-9) and produced variants of in PCC 7417. Anatoxin-a pathway identified in the genome of Cylindrospermum sp. PCC 7417 compared to homologous gene clusters from anatoxin-a producing cyanobacteria [8, 26, 27]. Genes with corresponding functions and domain organization are colored the same and connected by grey areas: anaA, proline adenylation, anaB, proline deshydrogenase, anaC, Type II thioesterase; anaD, acyl carrier protein, anaE, anaF and anaG are modular type I polyketide synthase with KS, β-ketoacyl synthase; AT, acyltransferase; KR, ketoacyl reductase; ACP, acyl carrier protein; DH, dehydratase; ER, enoyl reductase and CM, C-methyltransferase. The cyclase, named here anaI, is systematically associated to the anatoxin-a pathway. In addition, the last PKS anaG2 in PCC 7417 lacks the methyltransferase, and an oxidoreductase anaJ was detected. The transposase is present in the surrounding of the cluster only in PCC 6506. The detection of dihydroanatoxin-a from PCC 7417 is presented in Additional file 2: Figure S2.

Figure 3

Proportion of NRPS/PKS genes clusters in cyanobacterial genomes and their closest homologs. A. Percent of genome dedicated to encode NRPS/PKS gene clusters highlights the burst in the late-branching clades in comparison to the early-branching ones of the phylogenetic species tree (Figure  1), B. Distribution of the BLAST hits of the 2668 proteins contained in the 452 clusters compared to the non-redundant database of NCBI (identity > =30% and e-values <10e^-20).

Figure 4

NRPS/PKS genes clusters in Cyanobacteria: mobility, gene cluster size, proportion of GC% deviated genes and dinucleotide average absolute relative abundance difference. The size of the 452 gene clusters is indicated for each strain with their corresponding percentage of genes presenting below a GC% deviation and dinucleotide average absolute relative abundance difference (δ*-differences) and above associated with presence of mobility traces in their genomic context. The genomes are ordered accordingly to the phylogenetic tree presented in Figure  1.

Figure 5

Evolution of the CF-39 gene cluster. The CF-39 gene clusters are ordered according to the species phylogeny. The gene clusters and their flanking sequences are pictured: NRPS genes are surrounded in green with their domain composition detailed as well as the substrate specificity of A domain (parentheses), the conserved hydrolase gene is indicated in red, transposase one in yellow, and other genes are pictured in white and annotated when their function is known. C and A domain phylogenies were reported as lines that link the homologous domains by the same type of lines.

Figure 6

Phylogeny of bacterial ketoacyl synthase domains and their distribution in cyanobacterial PKS PUFA gene clusters. A. Maximum likelihood phylogenetic tree of bacterial KS based on 235 conserved amino acids from 217 gene clusters encoding secondary lipids biosynthesis and enediynes. The clades of the phylogenetic tree supported by a bootstrap of ≥70% are indicated by a black dot. B. PKS gene clusters dedicated to PUFA and enediynes in Cyanobacteria. The background color of each KS refers to the corresponding clade of the adjacent phylogenetic tree. PKS genes are indicated in blue, whereas homologous genes are indicated by similar colors. KS, ketoacyl synthase; AT, acyltransferase; KR, ketoacyl reductase; ACP, acyl carrier protein; DH, dehydratase; ER, enoyl reductase; TER, thioester reductase.

Figure 7

Model for the formation of the siderophore anachelin 1. Hybrid gene cluster of Anabaena cylindrica PCC 7122 with a NRPS architecture in agreement with the assembly of the characteristic three hydrophilic amino acids (L-Thr-D-Ser-L-Ser) and the two units responsible for binding iron (Atha and Dmaq) of the anachelin 1 structure [37]. In addition, PKS tailoring and transport-related genes are present. Symbol: Cy, cyclisation; A, adenylation; T, thiolation; KACP, KSIII acyl carrier protein; SalS, salicylate synthase; CoAL, salicylyl CoA/PCP ligase; KS, β-ketoacyl synthase; AT, acyltransferase; KR, ketoacyl reductase; ACP, acyl carrier protein; E, epimerase; Red, reductase; MT, methyl transferase; AmT, aminotransferase; TyrH, putative tyrosine hydrolase; SBP, siderophore binding protein; Srec, siderophore receptor; Sal, salicylic acid; Atha, 6-amino-3, 5, 7-trihydroxyheptanoic acid; Thr, threonine; Ser, serine; Dmaq, 1,1-dimethyl-3-amino-1,2,3,4-tetrahydro-6,7-dihydroxyquinolinium. The NRPS genes are indicated in green, the PKS genes are in blue.