Assembly of dimeric variants of coumermycins by tandem action of the four biosynthetic enzymes CouL, CouM, CouP, and NovN

Coumermycin A(1) is a member of the aminocoumarin family of antibiotics. Unlike its structural relatives, novobiocin and clorobiocin, coumermycin A(1) is a dimer built on a 3-methyl-2,4-dicarboxypyrrole scaffold and bears two decorated noviose sugar components which are the putative target binding motifs for DNA gyrase. Starting with this scaffold, we have utilized the ligase CouL for mono- and bisamide formation with aminocoumarins to provide substrates for the glycosyltransferase CouM. CouM was subsequently shown to catalyze mono- and bisnoviosylation of the resulting CouL products. CouP was shown to possess 4'-O-methyltransferase activity on products from tandem CouL, CouM assays. A fourth enzyme, NovN, the 3'-O-carbamoyltransferase from the novobiocin operon, was then able to carbamoylate either or both arms of the CouP product. The tandem action of CouL, CouM, CouP, and NovN thus generates a biscarbamoyl analogue of the pseudodimer coumermycin A(1). Starting from alternative dicarboxy scaffolds, these four enzymes can be utilized in tandem to create additional variants of dimeric aminocoumarin antibiotics.     

Genetic analysis of the biosynthesis of the pyrrole and carbamoyl moieties of coumermycin A1 and novobiocin

The aminocoumarin antibiotic coumermycin A(1) contains a central and two terminal pyrrole moieties. The coumermycin gene cluster in Streptomyces rishiriensis contains three genes (couN3, couN4 and couN5) that show sequence similarity to genes involved in the biosynthesis of the pyrrole moieties of pyoluteorin in Pseudomonas fluorescens and of undecylprodiginine in S. coelicolor. The gene couN3, which codes for a putative L-prolyl-S-PCP dehydrogenase, and the gene couN4, which encodes a putative L-prolyl-AMP ligase, were disrupted using in-frame deletion and insertional inactivation, respectively. HPLC analysis of culture extracts showed that formation of the two terminal pyrrole moieties was abolished in the couN3 (-) und couN4 (-) mutants. The mutants accumulated coumermycin D, which contains only the central pyrrole moiety. This result not only confirmed the involvement of couN3 and couN4 in the biosynthesis of the terminal pyrrole-2-carboxylic acid moieties of coumermycin A(1), but also indicated, for the first time, that the central 3-methylpyrrole-2,4-dicarboxylic acid unit of the coumermycins is formed by a biosynthetic pathway that differs from that used to assemble the terminal pyrrole moieties. novN, a putative carbamoyl transferase gene from the gene cluster for novobiocin biosynthesis in S. spheroides was expressed in the couN3 (-) mutant. This led to the formation of bis-carbamoylated coumermycin D, a novel compound of the coumermycin series.     

Overexpression,purification and characterization of SimL,an amide synthetase involved in simocyclinone biosynthesis

Simocyclinone D8 is a potent inhibitor of bacterial gyrase, produced by Streptomyces antibioticus Tü 6040. It contains an aminocoumarin moiety, similar to that of novobiocin, which is linked by an amide bond to a structurally complex acyl moiety, consisting of an aromatic angucycline polyketide nucleus, the deoxysugar olivose and a tetraene dicarboxylic acid. We have now investigated the enzyme SimL, responsible for the formation of the amide bond of simocyclinone. The gene was cloned, expressed in S. lividans T7, and the protein was purified to near homogeneity, and characterized. The 60 kDa protein catalyzed both the ATP-dependent activation of the acyl component as well as its transfer to the amino group of the aminocoumarin ring, with no requirement for a 4-phosphopantetheinyl cofactor. Besides its natural substrate, simocyclinone C4, SimL also accepted a range of cinnamic and benzoic acid derivatives and several other, structurally very diverse acids. These findings make SimL a possible tool for the creation of new aminocoumarin antibiotics.     

CouO and NovO: C-methyltransferases for tailoring the aminocoumarin scaffold in coumermycin and novobiocin antibiotic biosynthesis

During the biosynthesis of the streptomycete aminocoumarin antibiotics novobiocin and the dimeric coumermycin A(1), the bicyclic coumarin scaffold is C-methylated adjacent to the phenolic oxygen. The SAM-dependent C-methyltransferases NovO and CouO have been heterologously expressed and purified from Escherichia coli and shown to act after the aminocoumarin ring has been constructed by prior action of Nov/CouHIJK. Neither C-methyltransferase works on the tyrosyl-derived S-pantetheinyl intermediates tethered to NovH or on the subsequently released free aminocoumarin. NovL ligates the aminocoumarin to prenylhydroxybenzoate to yield novobiocic acid, which is the substrate for NovO before it is O-glycosylated by NovM. In coumermycin assembly, the corresponding ligase CouL makes the bis-amide by tandem ligation of two aminocoumarins to a dicarboxypyrrole. CouO works on both the mono- and bis-amides for mono- and di-C-methylation adjacent to the phenolic hydroxyl before it is glycosylated by CouM. Thus, the specific timing of C-methylation in the aminocoumarin antibiotic pathways is established.     

Characterization of the aminocoumarin ligase SimL from the simocyclinone pathway and tandem incubation with NovM,P,N from the novobiocin pathway

Simocyclinone D(8) consists of an anguicycline C-glycoside tethered by a tetraene diester linker to an aminocoumarin. Unlike the antibiotics novobiocin, clorobiocin, and coumermycin A(1), the phenolic hydroxyl group of the aminocoumarin in simocyclinone is not glycosylated with a decorated noviosyl moiety that is the pharmacophore for targeting bacterial DNA gyrase. We have expressed the Streptomyces antibioticus simocyclinone ligase SimL, purified it from Escherichia coli, and established its ATP-dependent amide bond forming activity with a variety of polyenoic acids including retinoic acid and fumagillin. We have then used the last three enzymes from the novobiocin pathway, NovM, NovP, and NovN, to convert a SimL product to a novel novobiocin analogue, in which the 3-prenyl-4-hydroxybenzoate of novobiocin is replaced with a tetraenoate moiety, to evaluate antibacterial activity.     

Cloning, overexpression, and purification of novobiocic acid synthetase from Streptomyces spheroides NCIMB 11891

Novobiocic acid synthetase, a key enzyme in the biosynthesis of the antibiotic novobiocin, was cloned from the novobiocin producer Streptomyces spheroides NCIMB 11891. The enzyme is encoded by the gene novL, which codes for a protein of 527 amino acids with a calculated mass of 56,885 Da. The protein was overexpressed as a His(6) fusion protein in Escherichia coli and purified to apparent homogeneity by affinity chromatography and gel chromatography. The purified enzyme catalyzed the formation of an amide bond between 3-dimethylallyl-4-hydroxybenzoic acid (ring A of novobiocin) and 3-amino-4,7-dihydroxy-8-methyl coumarin (ring B of novobiocin) in an ATP-dependent reaction. NovL shows homology to the superfamily of adenylate-forming enzymes, and indeed the formation of an acyl adenylate from ring A and ATP was demonstrated by an ATP-PP(i) exchange assay. The purified enzyme exhibited both activation and transferase activity, i.e. it catalyzed both the activation of ring A as acyl adenylate and the subsequent transfer of the acyl group to the amino group of ring B. It is active as a monomer as determined by gel filtration chromatography. The reaction was specific for ATP as nucleotide triphosphate and dependent on the presence of Mg(2+) or Mn(2+). Apparent K(m) values for ring A and ring B were determined as 19 and 131 micrometer respectively. Of several analogues of ring A, only 3-geranyl-4-hydroxybenzoate and to a lesser extent 3-methyl-4-aminobenzoate were accepted as substrates.     

Colorimetric inorganic pyrophosphate assay using a double cycling enzymatic method

The biosynthesis of aminocoumarin antibiotics involves the action of amide synthetases which construct amide bonds between aminocoumarins and various acyl moieties. Libraries of aminocoumarin analogues have been generated by in vivo fermentation, via feeding known amide synthetase substrates into producing microbial strains. Critically, such feeding studies rely on the inherent or engineered substrate promiscuity of each amide synthetase. We have initiated a program of directed evolution in order to create mutant amide synthetases for the synthesis of new nonnatural amino coumarin analogues. We used the clorobiocin enzyme CloL as a model amide synthetase to design and validate a fluorimetric high-throughput screen, which can be used to report the activity of mutant amide synthetases toward a broad range of coumarin and acyl donor substrates. Our assay monitors the decrease in fluorescence of aminocoumarins on acylation. The utility of the assay was illustrated by screening a library of amide synthetase mutants created by error-prone PCR. The substrate specificity of an amide synthetase was also rapidly probed using this assay, affording several newly identified substrates. It is anticipated that this high-throughput screen will accelerate the creation of amide synthetase mutants with new specificities by directed evolution.     

Chemoenzymatic formation of novel aminocoumarin antibiotics by the enzymes CouN1 and CouN7

The aminocoumarin antibiotics novobiocin, clorobiocin, and coumermycin A1 are highly potent inhibitors of the bacterial type II topoisomerase DNA gyrase. The key pharmacophore of both clorobiocin and coumermycin A1, the 5-methyl-2-pyrrolylcarbonyl moiety, targets the ATP-binding site of GyrB. The 5-methyl-2-pyrrolylcarbonyl group is transferred by the acyltransferases Clo/CouN7 from the carrier proteins Clo/CouN1 to the 3'-hydroxyl of the l-noviosyl scaffold during the late steps of clorobiocin and coumermycin A1 biosynthesis. We first examined the substrate specificity of the purified thiolation domain protein CouN1 in becoming primed by the phosphopantetheinyltransferase Sfp using a variety of synthetic CoA analogues of the 5-methyl-2-pyrrolylcarbonyl moiety. The acyl-S-CouN1 thioesters were then assayed as donors to the 3'-OH group of descarbamoylnovobiocin by the acyltransferase CouN7, resulting in 21 novel variants with heterocyclic acyl groups installed on the noviosyl moiety of the aminocoumarin scaffold. Scaleup of a 5-methylthiophene derivative yielded a compound with activity against both Gram-negative and Gram-positive bacteria. The minimal inhibitory concentration found for the Gram-positive bacteria was comparable to that of novobiocin.     

Novobiocin biosynthesis: inactivation of the putative regulatory gene<Emphasis Type="Italic"> novE</Emphasis> and heterologous expression of genes involved in aminocoumarin ring formation

The left ends of the biosynthetic gene clusters of novobiocin ( nov), clorobiocin ( clo) and coumermycin A(1) ( cou) from Streptomyces spheroides (syn. S. caeruleus) NCIMB 11891, S. roseochromogenes var. oscitans DS 12.976 and S. rishiriensis DSM 40489 were cloned and sequenced. Sequence comparison suggested that novE, cloE and couE, respectively, represent the borders of these three clusters. Inactivation of novE proved that novE does not have an essential catalytic role in novobiocin biosynthesis, but is likely to have a regulatory function. The gene products of novF and cloF show sequence similarity to prephenate dehydrogenase and may produce 4-hydroxyphenylpyruvate (4HPP) as a precursor of the substituted benzoate moiety of novobiocin and clorobiocin. Coumermycin A(1) does not contain this benzoate moiety, and correspondingly the coumermycin cluster was found not to contain a functional novF homologue. The coumermycin biosynthetic gene cluster apparently evolved from an ancestral cluster similar to those of novobiocin and clorobiocin, and parts of the ancestral novF homologue have been deleted in this process. No homologue to novC was identified in the gene clusters of clorobiocin and coumermycin, questioning the postulated involvement of novC in aminocoumarin biosynthesis. Heterologous expression of novDEFGHIJK in Streptomyces lividans resulted in the formation of 2,4-dihydroxy-alpha-oxy-phenylacetic acid, suggesting that at least one of the proteins encoded by these genes may participate in a hydroxylation reaction.     

A novel Dps-type protein from insect gut bacteria catalyses hydrolysis and synthesis of N-acyl amino acids

A novel type of a microbial N-acyl amino acid hydrolase (AAH) from insect gut bacteria was purified, cloned and functionally characterized. The enzyme was obtained from Microbacterium arborescens SE14 isolated from the foregut of larvae of the generalist herbivore Spodoptera exigua. The substrates of AAH are N-acyl-glutamines previously reported to elicit plant defence reactions after introduction into the leaf during feeding. The isolated AAH catalyses the hydrolysis of the amide bond (K(m) = 36 micromol l(-1)) and, less efficient, the formation (K(m) = 3 mmol l(-1)) of the elicitor active N-acyl amino acids. The AAH from M. arborescens SE14 shows no homology to known fatty acyl amidases (EC 3.5.1.4) but belongs to the family of Dps proteins (DNA-binding protein from starved cell). In line with other DPS proteins AAH is a homododecamer (monomer 17 181 Da) and contains iron atoms (c. 1-16 iron atoms per subunit). Unlike genuine DPS proteins the enzyme does not significantly bind DNA. Amino acid hydrolase is the first member of the DPS family that catalyses the cleavage or formation of amide bonds. The participation of a microbial enzyme in the homeostasis of N-acyl-glutamines in the insect gut adds further complexity to the interaction between plants and their herbivores.     

Cloning and analysis of the simocyclinone biosynthetic gene cluster of <Emphasis Type="Italic">Streptomyces antibioticus</Emphasis> Tü 6040

The biosynthetic gene cluster of the aminocoumarin antibiotic simocyclinone D8 was cloned by screening a cosmid library of Streptomyces antibioticusTü 6040 with a heterologous probe from a gene encoding a cytochrome P450 enzyme involved in the biosynthesis of the aminocoumarin antibiotic novobiocin. Sequence analysis of a 39.4-kb region revealed the presence of 38 ORFs. Six of the identified ORFs showed striking similarity to genes from the biosynthetic gene clusters of the aminocoumarin antibiotics novobiocin and coumermycin A(1). Simocyclinone also contains an angucyclinone moiety, and 12 of the ORFs showed high sequence similarity to biosynthetic genes of other angucyclinone antibiotics. Possible functions within the biosynthesis of simocyclinone D8 could be assigned to 23 ORFs by comparison with sequences in GenBank. Experimental proof for the function of the identified gene cluster was provided by a gene inactivation experiment, which resulted in the abolishment of the formation of the aminocoumarin moiety of simocyclinone. Feeding of the mutant with the aminocoumarin moiety of novobiocin led to a new, artificial simocyclinone derivative.     

Installation of the pyrrolyl-2-carboxyl pharmacophore by CouN1 and CouN7 in the late biosynthetic steps of the aminocoumarin antibiotics clorobiocin and coumermycin A1

The 5-methyl-2-pyrrolylcarbonyl moiety of the aminocoumarin antibiotics clorobiocin and coumermycin A1 is the key pharmacophore for targeting the ATP-binding site of GyrB for inhibition of the bacterial type-II topoisomerase DNA gyrase. During the late stage of clorobiocin and coumermycin A1 biosynthesis, the pyrrolyl-2-carboxyl group is transferred from the peptidyl carrier proteins Clo/CouN1 to the 3'-hydroxyl of the 4-methoxy-L-noviosyl scaffold by the action of the acyltransferases Clo/CouN7. CouN1 and CouN7 have now been heterologously expressed and purified from Escherichia coli. The apo form of CouN1 is converted to the acyl-holo form by loading with pyrrolyl-2-carboxyl-S-pantetheinyl moieties from synthetic pyrrolyl- and 5-methylpyrrolyl-CoAs by the action of the phosphopantetheinyl transferase Sfp. CouN7 acts as an acyltransferase, moving the pyrrole acyl moieties from CouN1 to the noviose sugar of descarbamoylnovobiocin. When the 5-methylpyrrolyl-2-carboxyl-thioester of CouN1 is the cosubstrate, the in vitro product differs from clorobiocin only in a CH3 for Cl group change on the coumarin ring. Double transfer of this acyl moiety by CouN7 to the penultimate intermediate in coumermycin A1 assembly completes that antibiotic biosynthetic pathway.     

Characterization of the formation of the pyrrole moiety during clorobiocin and coumermycin A1 biosynthesis

The aminocoumarin antibiotics clorobiocin and coumermycin A(1) target the B subunit of DNA gyrase by presentation of the 5-methyl-pyrrolyl-2-carboxy ester moiety in the ATP-binding site of the enzyme. The pyrrolyl pharmacophore is derived by a four electron oxidation of a prolyl unit while tethered in phosphopantetheinyl thioester linkage to a peptidyl carrier protein (PCP) subunit. l-Proline is selected and activated as l-prolyl-AMP by adenylation domain enzymes (CloN4 and CouN4) and then installed as the thioester on the holo form of the PCP proteins CloN5 and CouN5. Enzymatic oxidation of the prolyl-S-PCP by the flavoprotein dehydrogenase CloN3 can be followed by rapid quench and subsequent electrospray ionization-Fourier transform mass spectrometry analysis of the acyl-S-protein substrate/product mixture to establish that a two-electron oxidized pyrrolinyl-S-enzyme transiently accumulates on the way to the four-electron oxidized, heteroaromatic pyrrolyl-2-carboxy-S-PCP acyl enzyme product.     

Protein acylation in Tetrahymena

Examination of exhaustively delipidated Tetrahymena mimbres cells by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed the presence of several protein bands containing covalently linked fatty acids. Palmitic (16:0) and stearic (18:0) acids together accounted for approximately 90% of the protein-linked acyl chains, with myristic acid (14:0) comprising most of the remainder. Each of these three fatty acids was present mainly in alkali-stable linkage, indicating that unlike most other systems examined, fatty acids are attached to proteins of Tetrahymena principally by amide bonds. Smaller proportions of the acyl chains were susceptible to release by hydroxylaminolysis or by alkaline hydrolysis as would be expected from an ester linkage. The protein-bound acyl chains accounted for 0.3% of the cells' total fatty acids. They closely resembled in composition the highly saturated free fatty acid pool but not the vast pool of glycerolipid-associated fatty acids, which were mainly unsaturated. Cells subjected to thermal stress by rapid chilling from 39 to 15 degrees C responded by sharply increasing the ratio of palmitate to stearate in covalent association with proteins.     

Myristyl and palmityl acylation of the insulin receptor

The presence of covalently bound fatty acids in the insulin receptor has been explored in cultured human (IM-9) lymphocytes. Both alpha (Mr = 135,000) and beta (Mr = 95,000) subunits of the receptor incorporate [3H]myristic and [3H]palmitic acids in a covalent form. The effects of alkali and hydroxylamine on the labeled subunits indicate the existence of two different kinds of fatty acid linkage to the protein with chemical stabilities compatible with amide and ester bonds. The alpha subunit contains only amide-linked fatty acid while the beta subunit has both amide- and ester-linked fatty acids. Analysis by high performance liquid chromatography after acid hydrolysis of the [3H]myristate- and [3H]palmitate-labeled subunits demonstrates the fatty acid nature of the label. Furthermore, both [3H]myristic and [3H]palmitic acids are found attached to the receptor subunits regardless of which fatty acid was used for labeling. The incorporation of fatty acids into the insulin receptor is dependent on protein synthesis and is also detectable in the Mr = 190,000 proreceptor form. Fatty acylation is a newly identified post-translational modification of the insulin receptor which may have an important role in its interaction with the membrane and/or its biological function.     

Fatty acylation of heparan sulfate proteoglycan from human colon carcinoma cells

A number of transmembrane proteins have been recently reported to be modified by the covalent addition of saturated fatty acids which may contribute to membrane targeting and specific protein-lipid interactions. Such modifications have not been reported in cell-associated heparan sulfate proteoglycans, although these macromolecules are known to be hydrophobic. Here, we report that a cell surface heparan sulfate proteoglycan is acylated with both myristate and palmitate, two long-chain saturated fatty acids. When colon carcinoma cells were labeled with [3H]myristic acid, a significant proportion of the label was shown to be specifically incorporated into the protein core of the proteoglycan. Characterization of fatty acyl moiety in the purified proteoglycan by reverse-phase high pressure liquid chromatography revealed that approximately 60% of the covalently bound fatty acids was myristate. We further show that this relatively rare 14-carbon fatty acid was bound to the protein core via a hydroxylamine- and alkali-resistant amide bond. The remaining 40% was the more common 16-carbon palmitate, which was bound via a hydroxylamine- and alkali-sensitive thioester bond. Palmitate appeared to be added post-translationally and derived in part from intracellular elongation of myristate, a process that occurred within the first two hours and was insensitive to inhibition of protein synthesis. Acylation of heparan sulfate proteoglycan represents a novel modification of this gene product and could play a role in a number of biological functions including specific interactions with membrane receptors and ligand stabilization.     

Cloning and functional characterization of a phospholipid:diacylglycerol acyltransferase from Arabidopsis

A new pathway for triacylglycerol biosynthesis involving a phospholipid:diacylglycerol acyltransferase (PDAT) was recently described (Dahlqvist A, Stahl U, Lenman M, Banas A, Lee M, Sandager L, Ronne H, Stymne S, [2000] Proc Natl Acad Sci USA 97: 6487-6492). The LRO1 gene that encodes the PDAT was identified in yeast (Saccharomyces cerevisiae) and shown to have homology with animal lecithin:cholesterol acyltransferase. A search of the Arabidopsis genome database identified the protein encoded by the At5g13640 gene as the closest homolog to the yeast PDAT (28% amino acid identity). The cDNA of At5g13640 (AtPDAT gene) was overexpressed in Arabidopsis behind the cauliflower mosaic virus promoter. Microsomal preparations of roots and leaves from overexpressers had PDAT activities that correlated with expression levels of the gene, thus demonstrating that this gene encoded PDAT (AtPDAT). The AtPDAT utilized different phospholipids as acyl donor and accepted acyl groups ranging from C10 to C22. The rate of activity was highly dependent on acyl composition with highest activities for acyl groups containing several double bonds, epoxy, or hydroxy groups. The enzyme utilized both sn-positions of phosphatidylcholine but had a 3-fold preference for the sn-2 position. The fatty acid and lipid composition as well as the amounts of lipids per fresh weight in Arabidopsis plants overexpressing AtPDAT were not significantly different from the wild type. Microsomal preparations of roots from a T-DNA insertion mutant in the AtPDAT gene had barely detectable capacity to transfer acyl groups from phospholipids to added diacylglycerols. However, these microsomes were still able to carry out triacylglycerol synthesis by a diacylglycerol:diacylglycerol acyltransferase reaction at the same rate as microsomal preparations from wild type.     

Fatty Acid Composition of Myelin Proteolipid Protein During Vertebrate Evolution

The hydrophobic myelin proteolipid protein (PLP) contains covalently bound long-chain fatty acids which are attached to intracellular cysteine residues via thioester linkages. To gain insight into the role of acylation in the structure and function of myelin PLP, the amount and pattern of acyl groups attached to the protein during vertebrate evolution was determined. PLP isolated from brain myelin of amphibians, reptiles, birds and several mammals was subjected to alkaline methanolysis and the released methyl esters were analyzed by gas-liquid chromatography. In all species studied, PLP contained approximately the same amount of covalently bound fatty acids (3% w/w), and palmitic, palmitoleic, oleic and stearic acids were always the major acyl groups. Although the relative proportions of these fatty acids changed during evolution, the changes did not necessarily follow the variations in the acyl chain composition of the myelin free fatty acid pool, suggesting fatty acid specificity. The phylogenetic conservation of acylation suggests that this post-translational modification is critical for PLP function.     

Enzymatic peptide synthesis by the recombinant proline-specific endopeptidase from Flavobacterium meningosepticum and its mutationally altered Cys-556 variant

Proline-specific endopeptidase (PSE) (EC 3.4.21.26) was investigated for its potential as a catalyst in peptide synthesis. Using an activated peptide ester or a peptide amide as the acyl component, the enzyme catalyzed kinetically controlled aminolysis and transpeptidation respectively, with various amino acid amides as acyl acceptors. To a certain extent the nucleophile preference reflected the amino acid preference in the S₁-position of the enzyme in peptide hydrolysis: the highest fractions of aminolysis were obtained using amino acid amides with hydrophobic side-chains (e.g. Leu-NH₂, Phe-NH₂). PSE also catalyzed the thermodynamically controlled condensation of short peptides with a free carboxyterminus and various amino acid amides. This enabled us to examine the acceptance of different acyl components in the substrate-binding site of the enzyme with regard to their amino acid composition: In the S₁ position proline was clearly favored, but alanine was also accepted, whereas the S₂ subsite accepted various amino acids rather unspecifically. Since PSE was shown to be extremely sensitive against water-miscible organic solvents, an alternative approach was used to increase yields in enzymatic peptide synthesis: a derivative of PSE in which the catalytic Ser-556 is converted to a Cys was constructed by protein engineering. This mutant (PSE_cys) exhibited a dramatically increased peptide ligase activity in aqueous solution.