Genetic analysis of the biosynthesis of non-ribosomal peptide- and polyketide-like antibiotics, iron uptake and biofilm formation by Bacillus subtilis A1/3.

The Bacillus subtilis strain A1/3 shows exceptionally diverse antibiotic capacities compared to other B. subtilis strains. To analyze this phenomenon, mutants for the putative pantotheinyltransferase gene (pptS), and for several genes involved in non-ribosomal peptide synthesis and polyketide synthesis were constructed and characterized, using bioassays with blood cells, bacterial and fungal cells, and mass spectrometry. Among at least nine distinct bioactive compounds, five antibiotics and one siderophore activity were identified. The anti-fungal and hemolytic activities of strain A1/3 could be eliminated by mutation of the fen and srf genes essential for the synthesis of fengycins and surfactins. Both pptS- and dhb -type mutants were defective in iron uptake, indicating an inability to produce a 2,3-dihydroxybenzoate-type iron siderophore. Transposon mutants in the malonyl CoA transacylase gene resulted in the loss of hemolytic and anti-fungal activities due to the inhibition of bacillomycin L synthesis, and this led to the discovery of bmyLD-LA-LB* genes. In mutants bearing disruption mutations in polyketide (pksM- and/or pksR -like) genes, the biosynthesis of bacillaene and difficidins, respectively, was inactivated and was accompanied by the loss of discrete antibacterial activities. The formation of biofilms (pellicles) was shown to require the production of surfactins, but no other lipopeptides, indicating that surfactins serve specific developmental functions.

Construction and in vitro analysis of a new bi-modular polypeptide synthetase for synthesis of N-methylated acyl peptides.

BACKGROUND: Many active peptides are synthesized by nonribosomal peptide synthetases (NRPSs), large multimodular enzymes. Each module incorporates one amino acid, and is composed of two domains: an activation domain that activates the substrate amino acid and a condensation domain for peptide-bond formation. Activation domains sometimes contain additional activities (e.g. N-methylation or epimerization). Novel peptides can be generated by swapping domains. Exchange of domains containing N-methylation activity has not been reported, however. RESULTS: The actinomycin NRPS was used to investigate domain swapping. The first two amino acids of actinomycin are threonine and valine. We replaced the valine activation domain of module 2 with an N-methyl valine (MeVal) activation domain. The recombinant NRPS (AcmTmVe) catalyzes the formation of threonyl-valine. In the presence of S-adenosyl-methionine, valine was converted to MeVal but subsequent dipeptide formation was blocked. When acyl-threonine (the natural intermediate) was present at module 1, formation of acyl-threonine-MeVal occurred. The epimerization domain of AcmTmVe was impaired. CONCLUSIONS: A simple activation domain can be replaced by one with N-methylation activity. The same condensation domain can catalyze peptide-bond formation between N-methyl and nonmethylated amino acids. Modification of the upstream amino acid (i.e. acylation of threonine), however, was required for condensation with MeVal. Steric hindrance reduces chemical reactivity of N-methyl amino acids - perfect substrate positioning may only be achieved with acylated threonine. Loss of the epimerase activity of AcmTmVe suggests N-methyltransferase and epimerase domains, not found together naturally, are incompatible.

Evidence for an ergot alkaloid gene cluster in Claviceps purpurea.

A gene (cpd1) coding for the dimethylallyltryptophan synthase (DMATS) that catalyzes the first specific step in the biosynthesis of ergot alkaloids, was cloned from a strain of Claviceps purpurea that produces alkaloids in axenic culture. The derived gene product (CPD1) shows only 70% similarity to the corresponding gene previously isolated from Claviceps strain ATCC 26245, which is likely to be an isolate of C. fusiformis. Therefore, the related cpd1 most probably represents the first C. purpurea gene coding for an enzymatic step of the alkaloid biosynthetic pathway to be cloned. Analysis of the 3'-flanking region of cpd1 revealed a second, closely linked ergot alkaloid biosynthetic gene named cpps1, which codes for a 356-kDa polypeptide showing significant similarity to fungal modular peptide synthetases. The protein contains three amino acid-activating modules, and in the second module a sequence is found which matches that of an internal peptide (17 amino acids in length) obtained from a tryptic digest of lysergyl peptide synthetase 1 (LPS1) of C. purpurea, thus confirming that cpps1 encodes LPS1. LPS1 activates the three amino acids of the peptide portion of ergot peptide alkaloids during D-lysergyl peptide assembly. Chromosome walking revealed the presence of additional genes upstream of cpd1 which are probably also involved in ergot alkaloid biosynthesis: cpox1 probably codes for an FAD-dependent oxidoreductase (which could represent the chanoclavine cyclase), and a second putative oxidoreductase gene, cpox2, is closely linked to it in inverse orientation. RT-PCR experiments confirm that all four genes are expressed under conditions of peptide alkaloid biosynthesis. These results strongly suggest that at least some genes of ergot alkaloid biosynthesis in C. purpurea are clustered, opening the way for a detailed molecular genetic analysis of the pathway.

Phosphinothricin-tripeptide synthetases from Streptomyces viridochromogenes.

Phosphinothricyl-alanyl-alanine (Pt tripeptide (Ptt), bialaphos) is a metabolite produced by Streptomyces viridochromogenes and Streptomyces hygroscopicus. It contains the unique phosphinoamino acid phosphinothricin (Pt), which after cleavage from Ptt is active as an inhibitor of glutamine synthetase. We have isolated three enzymes that assemble the building block of the Ptt peptide backbone in a nonribosomal mechanism. The first enzyme, named Ptt-synthetase I (PTTS I), activates N-acetyldemethylphosphinothricin (AcDMPt) as adenylate and thioester. Pt is not activated. PTTS I can also activate N-acetylphosphinothricin (AcPt) or N-acetylglutamate as structural analogues of AcDMPT. Native PTTS I has an estimated size of 62 kDa whereas the denatured form displays a size of 76 kDa. Immunoblot analysis and determination of its N-terminal protein sequence revealed that PTTS I is identical with the gene product of phsA. The phsA gene was previously identified near the Pt-resistance gene pat in the Ptt biosynthesis gene cluster in S. viridochromogenes. Besides PTTS I, two alanine-activating enzymes (PTTS II/III) were partially purified from S. viridochromogenes with estimated native sizes of ca. 120 kDa (enzyme 1) and ca. 140 kDa (enzyme 2). Both enzymes bind alanine as a thioester via the corresponding adenylate. Level of PTTS II/III and product formation were correlated with each other in several different strains of S. viridochromogenes. These results indicate that Ptt is synthesized by three peptide synthetases, each activating one single amino acid. The data also confirm previous genetic data, which suggest that AcDMPt-Ala-Ala is the precursor of Ptt.