Sequence and analysis of the genetic locus responsible for surfactin synthesis in Bacillus subtilis

The chromosomal region of Bacillus subtilis comprising the entire srfA operon, sfp and about four kilo-bases in between have been completely sequenced and functionally characterized. The srfA gene codes for three large subunits of surfactin synthetase, 402, 401 and 144 kDa, respectively, arranged in a series of seven amino acid activating domains which, as shown in the accompanying communication, recognize and bind the seven amino acids of the surfactin peptide. The srfA amino acid activating domains share homologies with similar domains of other peptide synthetases; in particular, regions can be identified which are more homologous in domains activating the same amino acid. A fourth gene in srfA encodes a polypeptide homologous to grsT. Four genes are positioned between srfA and sfp, the disruption of which does not affect surfactin biosynthesis.     

Analysis of engineered multifunctional peptide synthetases. Enzymatic characterization of surfactin synthetase domains in hybrid bimodular systems.

The combinatorial reorganization of distinct modules of multimodular peptide synthetases is of increasing interest for the generation of new peptides with optimized bioactive properties. Each module is at least composed of enzymatic domains responsible for the adenylation, thioester formation, and condensation of an amino acid residue of the final peptide product. We analyzed various possible fusion sites for the recombination of peptide synthetases and evaluated the impact of different recombination strategies on the amino acid adenylation and acyl-thioester formation activities of peptide synthetase modules. Hybrid bimodular peptide synthetases were generated by recombination of the corresponding reading frames encoding for L-glutamic acid- and L-leucine-specific modules of surfactin synthetase SrfA-A at presumed inner- and intradomainic regions. We demonstrate that fusions at a previously postulated hinge region, dividing the amino acid adenylating domains of peptide synthetase modules into two subdomains, and at the highly conserved 4'-phosphopantetheine binding motif in acyl-thioester forming domains resulted in enzymatically active hybrid domains. By contrast, most manipulations in condensation domains like deletions, the complete exchange or the construction of chimeric domains considerably reduced or completely abolished the amino acid adenylation and thioester formation activity of the hybrid module.     

Modification of biologically active peptides: production of a novel lipohexapeptide after engineering of Bacillus subtilis surfactin synthetase

The Bacillus subtilis strain ATCC 21332 produces the lipoheptapeptide surfactin, a highly potent biosurfactant synthesized by a large multimodular peptide synthetase. We report the genetic engineering of the surfactin biosynthesis resulting in the production of a novel lipohexapeptide with altered antimicrobial activities. A combination of in vitro and in vivo recombination approaches was used to construct a modified peptide synthetase by eliminating a large internal region of the enzyme containing a complete amino acid incorporating module. The remaining modules adjacent to the deletion were recombined at different highly conserved sequence motifs characteristic of amino acid incorporating modules of peptide synthetases. The primary goal of this work was to identify permissive fusion sites suitable for the engineering of peptide synthetase genes by genetic recombination. Analysis of the rearranged enzymes after purification from B. subtilis and from the heterologous host Escherichia coli revealed that the selection of the recombination site is of crucial importance for a successful engineering. Only the recombination at a specific HHII x DGVS sequence motif resulted in an active peptide synthetase. The expected lipohexapeptide was produced in vivo and first evidence of a reduced toxicity against erythrocytes and an enhanced lysis of Bacillus licheniformis cells was shown.     

Targeted alteration of the substrate specificity of peptide synthetases by rational module swapping

Analysis of the primary structure of peptide synthetases involved in the non-ribosomal synthesis of peptide antibiotics has revealed a highly conserved and ordered modular arrangement. A module contains at least two domains, involved in ATP-dependent substrate activation and thioester formation. The occurrence and arrangement of these functional building blocks is associated with the number and order of the amino acids incorporated in the peptide product. In this study, we present data on the targeted exchange of the leucine-activating module within the three-module surfactin synthetase 1 (SrfA-A) of Bacillus subtilis. This was achieved by engineering several hybrid srfA-A genes, which were introduced into the surfactin biosynthesis operon by in vivo recombination. We examined the hybrid genes for expression and investigated the enzymatic activities of the resulting recombinant peptide synthetases. For the first time, we demonstrate directly that an individual minimal module, of bacterial or fungal origin, confers its amino acid-specific activity on a multi-modular peptide synthetase. Furthermore, it is shown that directed incorporation of ornithine at the second position of the peptide chain induces a global alteration in the conformation of surfactin and may result in premature cyclization or a branched cyclic structure. Received: 10 July 1997 / Accepted: 11 September 1997     

The tyrocidine biosynthesis operon of Bacillus brevis: complete nucleotide sequence and biochemical characterization of functional internal adenylation domains.

The cyclic decapeptide antibiotic tyrocidine is produced by Bacillus brevis ATCC 8185 on an enzyme complex comprising three peptide synthetases, TycA, TycB, and TycC (tyrocidine synthetases 1, 2, and 3), via the nonribosomal pathway. However, previous molecular characterization of the tyrocidine synthetase-encoding operon was restricted to tycA, the gene that encodes the first one-module-bearing peptide synthetase. Here, we report the cloning and sequencing of the entire tyrocidine biosynthesis operon (39.5 kb) containing the tycA, tycB, and tycC genes. As deduced from the sequence data, TycB (404,562 Da) consists of three modules, including an epimerization domain, whereas TycC (723,577 Da) is composed of six modules and harbors a putative thioesterase domain at its C-terminal end. Each module incorporates one amino acid into the peptide product and can be further subdivided into domains responsible for substrate adenylation, thiolation, condensation, and epimerization (optional). We defined, cloned, and expressed in Escherichia coli five internal adenylation domains of TycB and TycC. Soluble His6-tagged proteins, ranging from 536 to 559 amino acids, were affinity purified and found to be active by amino acid-dependent ATP-PPi exchange assay. The detected amino acid specificities of the investigated domains manifested the colinear arrangement of the peptide product with the respective module in the corresponding peptide synthetases and explain the production of the four known naturally occurring tyrocidine variants. The Km values of the investigated adenylation domains for their amino acid substrates were found to be comparable to those published for undissected wild-type enzymes. These findings strongly support the functional integrities of single domains within multifunctional peptide synthetases. Directly downstream of the 3' end of the tycC gene, and probably transcribed in the tyrocidine operon, two tandem ABC transporters, which may be involved in conferring resistance against tyrocidine, and a putative thioesterase were found.     

Cloning, sequencing, and characterization of the iturin A operon

Bacillus subtilis RB14 is a producer of the antifungal lipopeptide iturin A. Using a transposon, we identified and cloned the iturin A synthetase operon of RB14, and the sequence of this operon was also determined. The iturin A operon spans a region that is more than 38 kb long and is composed of four open reading frames, ituD, ituA, ituB, and ituC. The ituD gene encodes a putative malonyl coenzyme A transacylase, whose disruption results in a specific deficiency in iturin A production. The second gene, ituA, encodes a 449-kDa protein that has three functional modules homologous to fatty acid synthetase, amino acid transferase, and peptide synthetase. The third gene, ituB, and the fourth gene, ituC, encode 609- and 297-kDa peptide synthetases that harbor four and two amino acid modules, respectively. Mycosubtilin, which is produced by B. subtilis ATCC 6633, has almost the same structure as iturin A, but the amino acids at positions 6 and 7 in the mycosubtilin sequence are D-Ser-->L-Asn, while in iturin A these amino acids are inverted (i.e., D-Asn-->L-Ser). Comparison of the amino acid sequences encoded by the iturin A operon and the mycosubtilin operon revealed that ituD, ituA, and ituB have high levels of homology to the counterpart genes fenF (79%), mycA (79%), and mycB (79%), respectively. Although the overall level of homology of the amino acid sequences encoded by ituC and mycC, the counterpart of ituC, is relatively low (64%), which indicates that there is a difference in the amino acid sequences of the two lipopeptides, the levels of homology between the putative serine adenylation domains and between the asparagine adenylation domains in the two synthetases are high (79 and 80%, respectively), implying that there is an intragenic domain change in the synthetases. The fact that the flanking sequence of the iturin A synthetase coding region was highly homologous to the flanking sequence that of xynD of B. subtilis 168 and the fact that the promoter of the iturin A operon which we identified was also conserved in an upstream sequence of xynD imply that horizontal transfer of this operon occurred. When the promoter was replaced by the repU promoter of the plasmid pUB110 replication protein, production of iturin A increased threefold.     

Nucleotide sequence of 5' portion of srfA that contains the region required for competence establishment in Bacillus subtilus.

The nucleotide sequence of the 20,535 base pairs of the 5' end of the srfA operon, containing the region required for competence development, was determined. This included the srfA promoter region, the first open reading frame, srfAA, encoding surfactin synthetase I and part of the second open reading frame, srfAB, encoding surfactin synthetase II. Three amino acid-activating domains characteristic of those found in peptide synthetases could be discerned in both srfAA (activating Glu, Leu and D-Leu) and srfAB (activating Val, Asp, and D-Leu). The presence of a conserved spacer motif in the amino-terminal end of srfAA suggests that the srfAA product may not initiate surfactin synthesis. The portion of srfA that contains the region required for competence is composed of srfAA and the first amino acid-activating domain of srfAB.     

Genetic evidence for a role of thioesterase domains, integrated in or associated with peptide synthetases, in non-ribosomal peptide biosynthesis in Bacillus subtilis

Next to almost all prokaryotic operons encoding peptide synthetases, which are involved in the nonribosomal synthesis of peptide antibiotics, distinct genes have been detected that encode proteins with strong sequence similarity to type II fatty acid thioesterases of vertebrate origin. Furthermore, sequence analysis of bacterial and fungal peptide synthetases has revealed a region at the C-terminal end of modules that are responsible for adding the last amino acid to the peptide antibiotics; that region also exhibits significant similarities to thioesterases. In order to investigate the function of these putative thioesterases in non-ribosomal peptide synthesis of the lipopeptide antibiotic surfactin in Bacillus subtilis, srfA fragments encoding the thioesterase domain of the surfactin synthetase 3 and the thioesterase-like protein SrfA-TE were deleted. This led to a 97 and 84% reduction of the in vivo surfactin production, respectively. In the double mutant, however, no surfaction production was detectable. These findings demonstrate for the first time that the C-terminal thioesterase domains and the SrfA-TE protein are directly involved in nonribosomal peptide biosynthesis. Received: 30 September 1997 / Accepted: 4 December 1997     

Genes encoding synthetases of cyclic depsipeptides, anabaenopeptilides, in Anabaena strain 90

Anabaena strain 90 produces three hepatotoxic heptapeptides (microcystins), two seven-residue depsipeptides called anabaenopeptilide 90A and 90B, and three six-residue peptides called anabaenopeptins. The anabaenopeptilides belong to a group of cyanobacterial depsipeptides that share the structure of a six-amino-acid ring with a side-chain. Despite their similarity to known cyclic peptide toxins, no function has been assigned to the anabaenopeptilides. Degenerate oligonucleotide primers based on the conserved amino acid sequences of other peptide synthetases were used to amplify DNA from Anabaena 90, and the resulting polymerase chain reaction (PCR) products were used to identify a peptide synthetase gene cluster. Four genes encoding putative anabaenopeptilide synthetase domains were characterized. Three genes, apdA, apdB and apdD, contain two, four and one module, respectively, encoding a total of seven modules for activation and peptide bond formation of seven L-amino acids. Modules five and six also carry methyltransferase-like domains. Before the first module, there is a region similar in amino acid sequence to formyltransferases. A fourth gene (apdC), between modules six and seven, is similar in sequence to halogenase genes. Thus, the order of domains is co-linear with the positions of amino acid residues in the finished peptide. A mutant of Anabaena 90 was made by inserting a chloramphenicol resistance gene into the apdA gene. DNA amplification by PCR confirmed the insertion. Mass spectrometry analysis showed that anabaenopeptilides are not made in the mutant strain, but other peptides, such as microcystins and anabaenopeptins, are still produced by the mutant.     

Recombinant acylheptapeptide lichenysin: high level of production by Bacillus subtilis cells

Peptide synthetases are multi-domain proteins that catalyze the assembly, from amino acids and amino acid derivatives, of peptides and lipopeptides, some of which exhibit activities (pharmaceutical, surfactant, etc.) of considerable biotechnological importance. Although there is substantial interest in the generation of greater peptide diversity, in order to create new biotechnologically interesting products, attempts reported so far to exchange amino acid-activating minimal modules between enzymes have only yielded hybrid catalysts with poor activities. We report here the replacement of an entire first, L-Glu-, and fifth, L-Asp-incorporating modules of surfactin synthetase, to create a fully active hybrid enzyme that forms a novel peptide in high yields. Whole encoding regions of lichenysin A synthetase modules were introduced into surfactin biosynthesis operon between His140/His1185 of SrfAA and His1183/His2226 of SrfAB, the amino acid residues of a proposed active-site motif (HHXXXDG) of the condensation domains which is involved in the catalysis of nonribosomal peptide bond formation (Stachelhaus et al., 1998). When the lipopeptides produced by the recombinant Bacillus subtilis strains were purified and characterized, they appeared to be expressed approximately at the same level of the wild type surfactin and to be identical by their fatty acid profiles. We thereby demonstrate the utility of whole module swapping for designing novel peptides, for creating peptide diversity, and for redesigning existing peptides produced in performant production strains in high yields to correspond to desired peptides produced in low yields, or from strains unsuitable for production purposes.     

Identification and cloning of peptide synthetase genes of thermostable bacilli using the polymerase chain reaction

Fourteen thermophilic and thermostable strains of the genus Bacillus were studied. Total DNA was isolated from these strains and used as a template to identify and clone peptide synthetase genes by means of polymerase chain reaction. Amplified DNA fragments were cloned into a phasmid vector, and nucleotide sequences of cloned fragments were determined. Stringent thermophilic strains were shown to lack genetic systems, which are responsible for the synthesis of secondary metabolites and homologous to the known peptide synthetase genes. On the contrary, thermostable strains had peptide synthetases and produced antimicrobial secondary metabolites. Analysis of nucleotide sequences and deduced amino acid sequences of cloned PCR fragments from B. licheniformis strains VK2, VK21, and VK2101 showed that they are absolutely identical. The cloned DNA fragment was found to be a portion of the open reading frame, which we termed ORF1. Data from analysis of a partial nucleotide sequence of the peptide synthetase gene of strain VK21 indicated that a 9.5-kb region of chromosomal DNA contains sequences of two genes homologous to the B. subtilis peptide synthetase genes dhbB and dhbF. Strains VK2, VK21, and VK2101 were shown to synthesize siderophores. A method for screening bacteria with peptide synthetase genes has been developed.     

Four homologous domains in the primary structure of GrsB are related to domains in a superfamily of adenylate-forming enzymes

The entire nucleotide sequence of the Bacillus brevis grsB gene encoding the gramicidin S synthetase 2, which activates and condenses the four amino acids proline, valine, ornithine and leucine has been determined. The gene contains an open reading frame of 13,359 bp which encodes a protein of 4453 amino acids with a predicted Mr of 510,287. The gene is located within the gramicidin S biosynthetic operon, also containing the genes grsT and grsA, whose nucleotide sequences have been determined previously. Within the GrsB amino acid sequence four conserved and repeated domains of about 600 amino acids (45-50% identity) have been identified. The four domains are separated by non-homologous sequences of about 500 amino acids. The domains also share a high degree of similarity (20-70%) with eight peptide synthetases of bacterial and fungal origin as well as with conserved sequences of nine other adenylate-forming enzymes of diverse origin. On the basis of sequence homology and functional similarities, we infer that those enzymes share a common evolutionary origin and present a phylogenetic tree for this superfamily of domain-bearing enzymes.     

Aminoacylation of RNA Minihelices: Implications for tRNA Synthetase Structural Design and Evolution

Abstract

The genetic code is based on the aminoacylation of tRNA with amino acids catalyzed by the aminoacyl-tRNA synthetases. The synthetases are constructed from discrete domains and all synthetases possess a core catalytic domain that catalyzes amino acid activation, binds the acceptor stem of tRNA, and transfers the amino acid to tRNA. Fused to the core domain are additional domains that mediate RNA interactions distal to the acceptor stem. Several synthetases catalyze the aminoacylation of RNA oligonucleotide substrates that recreate only the tRNA acceptor stems. In one case, a relatively small catalytic domain catalyzes the aminoacylation of these substrates independent of the rest of the protein. Thus, the active site domain may represent a primordial synthetase in which polypeptide insertions that mediate RNA acceptor stem interactions are tightly integrated with determinants for aminoacyl adenylate synthesis. The relationship between nucleotide sequences in small RNA oligonucleotides and the specific amino acids that are attached to these oligonucleotides could constitute a second genetic code.     

A component of the multisynthetase complex is a multifunctional aminoacyl-tRNA synthetase.

In higher eukaryotes, nine aminoacyl-tRNA synthetases are associated within a multienzyme complex which is composed of 11 polypeptides with molecular masses ranging from 18 to 150 kDa. We have cloned and sequenced a cDNA from Drosophila encoding the largest polypeptide of this complex. We demonstrate here that the corresponding protein is a multifunctional aminoacyl-tRNA synthetase. It is composed of three major domains, two of them specifying distinct synthetase activities. The amino and carboxy-terminal domains were expressed separately in Escherichia coli, and were found to catalyse the aminoacylation of glutamic acid and proline tRNA species, respectively. The central domain is made of six 46 amino acid repeats. In prokaryotes, these two aminoacyl-tRNA synthetases are encoded by distinct genes. The emergence of a multifunctional synthetase by a gene fusion event seems to be a specific, but general attribute of all higher eukaryotic cells. This type of structural organization, in relation to the occurrence of multisynthetase complexes, could be a mechanism to integrate several catalytic domains within the same particle. The involvement of the internal repeats in mediating complex assembly is discussed.     

Bioinformatics and molecular approaches to detect NRPS genes involved in the biosynthesis of kurstakin from <Emphasis Type="Italic">Bacillus thuringiensis</Emphasis>

Degenerated primers designed for the detection by polymerase chain reaction of nonribosomal peptide synthetases (NRPS) genes involved in the biosynthesis of lipopeptides were used on genomic DNA from a new isolate of Bacillus thuringiensis CIP 110220. Primers dedicated to surfactin and bacillomycin detection amplified sequences corresponding respectively to the surfactin synthetase operon and to a gene belonging to a new NRPS operon identified in the genome of B. thuringiensis serovar pondicheriensis BSCG 4BA1. A bioinformatics analysis of this operon led to the prediction of an NRPS constituted of seven modules beginning with a condensation starter domain and which could be involved in the biosynthesis of a heptalipopeptide similar to kurstakin. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-ToF-MS) performed on whole cells of B. thuringiensis CIP 110220 confirmed the production of kurstakin by this strain. The kurstakin operon was thus used to design a new set of degenerated primers specifically to detect kurstakin genes. These primers were used to screen kurstakin producers in a collection of nine B. thuringiensis strains isolated from different areas in Algeria and two from the Pasteur Institute collection. For eight among the 11 tested strains, the amplified fragment matched with an operon similar to the kurstakin operon and found in the newly sequenced genome of Bacillus cereus or B. thuringiensis serovar pulsiensis, kurstaki, and thuringiensis. Kurstakin production was detected by MALDI-ToF-MS on whole cells for six strains. This production was compared with the spreading of the strains and their antimicrobial activity. Only the spreading can be correlated with the kurstakin production.     

A mammalian tryptophanyl-tRNA synthetase shows little homology to prokaryotic synthetases but near identity with mammalian peptide chain release factor

Determination of the amino acid sequence of beef pancreas tryptophanyl-tRNA synthetase was undertaken through both cDNA and direct peptide sequencing. A full-length cDNA clone containing a 475 amino acid open reading frame was obtained. The molecular mass of the corresponding peptide chain, 53,728 Da, was in agreement with that of beef tryptophanyl-tRNA synthetase, as determined by physicochemical methods (54 kDa). Expression of this clone in Escherichia coli led to tryptophanyl-tRNA synthetase activity in cell extracts. The open reading frame included two sequences analogous to the consensus sequences, HIGH and KMSKS, found in class I aminoacyl-tRNA synthetases. The homology with prokaryotic and yeast mitochondrial tryptophanyl-tRNA synthetases was low and was limited to the regions of the consensus sequences. However, a 90% homology was observed with the recently described rabbit peptide chain release factor (eRF) [Lee et al. (1990) Proc. Natl. Acad. Sci. 87, 3508-3512]. Such a strong homology may reveal a new group of genes deriving from a common ancestor, the products of which could be involved in tRNA aminoacylation (tryptophanyl-tRNA synthetase) or translation termination (eRF).     

Identification and Cloning of Peptide Synthetase Genes of Thermostable Bacilli Using the Polymerase Chain Reaction

Fourteen thermophilic and thermostable strains of the genus Bacillus were studied. Total DNA was isolated from these strains and used as a template to identify and clone peptide synthetase genes by means of polymerase chain reaction. Amplified DNA fragments were cloned into a phasmid vector, and nucleotide sequences of cloned fragments were determined. Stringent thermophilic strains were shown to lack genetic systems, which are responsible for the synthesis of secondary metabolites and homologous to the known peptide synthetase genes. On the contrary, thermostable strains had peptide synthetases and produced antimicrobial secondary metabolites. Analysis of nucleotide sequences and deduced amino acid sequences of cloned PCR fragments from B. licheniformis strains VK2, VK21, and VK2101 showed that they are absolutely identical. The cloned DNA fragment was found to be a portion of the open reading frame, which we termed ORF1. Data from analysis of a partial nucleotide sequence of the peptide synthetase gene of strain VK21 indicated that a 9.5-kb region of chromosomal DNA contains sequences of two genes homologous to the B. subtilis peptide synthetase genesdhbB and dhbF. Strains VK2, VK21, and VK2101 were shown to synthesize siderophores. A method for screening bacteria with peptide synthetase genes has been developed.     

Hydroxamate-based colorimetric assay to assess amide bond formation by adenylation domain of nonribosomal peptide synthetases

We demonstrated the usefulness of a hydroxamate-based colorimetric assay for predicting amide bond formation (through an aminoacyl-AMP intermediate) by the adenylation domain of nonribosomal peptide synthetases. By using a typical adenylation domain of tyrocidine synthetase (involved in tyrocidine biosynthesis), we confirmed the correlation between the absorbance at 490 nm of the l-Trp–hydroxamate–Fe³⁺ complex and the formation of l-Trp–l-Pro, where l-Pro was used instead of hydroxylamine. Furthermore, this assay was adapted to the adenylation domains of surfactin synthetase (involved in surfactin biosynthesis) and bacitracin synthetase (involved in bacitracin biosynthesis). Consequently, the formation of various aminoacyl l-Pro formations was observed.