Cyanobacteria produce a high number of chemically diverse oligopeptides exhibiting various types of bioactivities ranging from mild enzyme inhibition to initiation of acute toxic effects in pro- or eukaryotes . More than 600 individual compounds have been described and probably many peptides have remained undiscovered. Most oligopeptides can be assigned to chemical classes, of which aeruginosins, anabaenopeptins, cyanopeptolins, microcystins, microginins, and microviridins are among the most recognized ones . Some classes have well characterized biosynthetic pathways [3–6], while for others, hardly anything is known. The biological functions of cyanobacterial oligopeptides are unknown, despite knockout studies for microcystins and cyanopeptolins [3, 4, 7]. However, gene knockouts in the relevant cyanobacterial strains have turned out to be difficult. Thus, the link between a given oligopeptide and the gene cluster has generally been difficult to establish. Since the precise functions of the peptides are unknown, the reasons for the vast structural peptide diversity (both within- and between classes of oligopeptides) remain obscure.
Several of the oligopeptides classes are produced by nonribosomal peptide synthetases (NRPSs). NRPS pathways have been verified for microcystins [3, 6, 8], cyanopeptolins [4, 9, 10] and aeruginosins . NRPSs have a modular structure with distinct activation domains (A-domains), thiolation domains (T-domains), and condensation domains (C domains) that are easily identified due to signature sequences . Conserved sequence motifs also allow identification of additional modules and domains that often are present in NRPS enzyme complexes, including metyltransferases (M-domains), epimerases (E-domains), thioesterase (TE domains), halogenases, and ABC transporters. The oligopeptide products of NRPS gene clusters may be predicted in silico based on binding pocket analyses of A-domains [12, 13], phylogeny and the co-linearity rule, i.e. the order of A-domains is co-linear with the amino acid sequence of the finished peptide . Similar predictions of secondary metabolites from NRPS gene clusters have been performed successfully previously in Streptomyces (). Some oligopeptides are, however, synthesized ribosomally [16–18], and it is at present unknown if this is the case also for other classes. Due to insufficient genomic information this question has been hard to address.
Evolutionary and phylogenetic studies of individual NRPS gene clusters have revealed frequent horizontal gene transfer (HGT) between related strains [19–21]. This is likely to generate new or recurrent variants of the enzymatic modules, leading to a change in oligopeptide profiles. In addition to HGT (intergenomic recombination), recombination between sequences within the same genome (intragenomic) may occur even between different classes of NRPS clusters due to a general high genetic similarity (see Majewski and Cohan  and Papke et al. ) among the building blocks (i.e. the modules) [9, 10, 24] of NRPS gene clusters. Recombination events and point mutations may be reinforced by positive selection for the new variant, indicating that the new oligopeptide variant may have biological significance .
Welker et al, 2006  indicated that production of oligopeptides is concentrated within certain genera of cyanobacteria. Why do most strains belonging to in the NRPS producing genera produce several classes of oligopeptides and multiple variants of the same oligopeptide class? Further, what is the biological significance of the many recombination events within the NRPS gene clusters? So far, most studies have investigated single NRPS operons (and their oligopeptides) or compared variants of the same class of operons in different strains. It is likely that all the different oligopeptides within a strain contribute to its survival and fitness, therefore the most relevant approach to take is to examine the entire genome of a strain for NRPS genes. Such an approach should establish relationship between each oligopeptide and operon, and consequently reduce the need for knock-out mutants. Furthermore, the possibility of ribosomal synthesis of some oligopeptide classes can be addressed. In addition, it will be possible to investigate whether exchange of modules between different classes of NRPS gene clusters within the same genome occurs.
Here, we have selected a Planktothrix rubescens strain, NIVA CYA 98, that produces all major classes of oligopeptides according to Welker et al  and shotgun sequenced the genome by massive parallel pyrosequencing (454) to high depth (18.5×). We have characterized the oligopeptides of this strain in detail by LC-MS-MS. This has enabled a correlation of NRPS gene clusters found within the genome with the identified oligopetides. A putative gene cluster could be identified for all structurally characterized oligopeptides. However, for two NRPS gene clusters no oligopeptide were detected. Overall, the data show that the combined LC-MS-MS and 454 genome sequencing is a very powerful approach for finding putative secondary metabolite genes.