Genome mining: Prediction of lipopeptides and polyketides from Bacillus and related Firmicutes

Abstract

Bacillus and related genera in the Bacillales within the Firmicutes harbor a variety of secondary metabolite gene clusters encoding polyketide synthases and non-ribosomal peptide synthetases responsible for remarkable diverse number of polyketides (PKs) and lipopeptides (LPs). These compounds may be utilized for medical and agricultural applications. Here, we summarize the knowledge on structural diversity and underlying gene clusters of LPs and PKs in the Bacillales. Moreover, we evaluate by using published prediction tools the potential metabolic capacity of these bacteria to produce type I PKs or LPs. The huge sequence repository of bacterial genomes and metagenomes provides the basis for such genome-mining to reveal the potential for novel structurally diverse secondary metabolites. The otherwise cumbersome task to isolate often unstable PKs and deduce their structure can be streamlined. Using web based prediction tools, we identified here several novel clusters of PKs and LPs from genomes deposited in the database. Our analysis suggests that a substantial fraction of predicted LPs and type I PKs are uncharacterized, and their functions remain to be studied. Known and predicted LPs and PKs occurred in the majority of the plant associated genera, predominantly in Bacillus and Paenibacillus. Surprisingly, many genera from other environments contain no or few of such compounds indicating the role of these secondary metabolites in plant-associated niches.

Keywords: Genome mining; Lipopeptides; Non-ribosomal protein synthetase; Paenibacillus; Polyketides; Structure prediction.

Fig. 1

Architectures of type I polyketide synthases (PKS) showing similarities and dissimilarities in known and predicted PKs. Iterative domains: ACP, acyl carrier protein; PCP, peptidyl carrier protein; A, adenylation; KS, ketosynthase; DH, dehydratase; MT, methyl transferase; KR, ketoreductase; TE, thioesterase. Further details of domains are described in Table 1. Modules and recruited amino acids indicated below, gene names indicated above each illustration. (A) Gene clusters of the three types of well-known PKS from B. amyloliquefaciens FZB42: (a) difficidin, (b) macrolactin, (c) bacillaene. Modular regions of predicted PKS: (a) bacillaene variant from P. pini 16418; number and order of the domains differ from B.amyloliquefaciens FZB42 bacillaene, (b) novel PKS from P. polymyxa E681; an adenylation domain specifies glycine, (c) paenimacrolidine like PKS from P. durus DSM 1735, (d) novel PKS form Brevibacillus brevis NBRC 100599; adenylation domains specify ala and ser, also contains the methylation domains- oMT and nMT. These predicted PKS machinery in Paenibacillus may work without thioesterase.

Fig. 2

Chemical structures of polyketides of Bacillus and Paenibacillus. (A) Polyketides from B. amyloliquefaciens FZB42 (a, b, c) and Bacillus sp. AH159-1 (c): (a) difficidins, (b) bacillaenes and (c) macrolactins. Stereochemistry not shown. (B) Polyketides from Paenibacillus: (a) Paenimacrolidin from Paenibacillus sp. F6-B70. Stereochemistry unknown. (b) Paenilamicin from P. larvae DSM25430.

Fig. 3

Organization of the non-ribosomal peptide synthetases (NRPS) encoding lipopeptides in Paenibacillus and Bacillus. Iterative domains: A, adenylation; T, thiolation; E, epimerization; MCT, malonyl-CoA transacylase; ACL, acyl-coA ligase; AMT, aminotransferase; dab, 2,4-diaminobutyric acid; orn, ornithine; KS, keto synthetase; TE, thioesterase. Further details of domains are described in Table 1. Modules and recruited amino acids indicated below, gene names indicated above each illustration. (A) Organization of the known NRPS (a) polymyxin A in P. polymyxa E681, (b) fusaricidin in P. polymyxa E681 and (c) tridecaptin A in P. terrae NRRL B-30644. (B) Organization of the predicted novel NRPS encoding (a) a heptapeptide in P. polymyxa E681; modular architecture is similar to the known Iturin but predicted amino acid composition is completely different and (b) organization of the known mycosubtilin operon , an iturin member from B. subtilis for comparison.

Fig. 4

Chemical structures of lipopeptides from Bacillus and Paenibacillus. (A) Lipopeptides from B. amyloliquefaciens FZB42 (a,b,c): (a) surfactin, (b) bacillomycin (an iturin member), (c) plipastatin (a fengycin member) and (d) kurstakin from B. thuringiensis kurstaki HD-1. (B) Lipopeptides from Paenibacillus: (a) polymyxin A from P. polymyxa E681, (b) fusaricidin C from P. polymyxa E681, (c) paenibacterin from Paenibacillus sp. OSY-SE (d) tridecaptin from P. terrae NRRL B-30644.