Exploring Peptide Ligase Orthologs in Actinobacteria-Discovery of Pseudopeptide Natural Products, Ketomemicins.

We recently identified a novel peptide ligase (PGM1), an ATP-grasp-ligase, that catalyzes amide bond formation between (S)-2-(3,5-dihydroxy-4-methoxyphenyl)-2-guanidinoacetic acid and ribosomally supplied oligopeptides in pheganomycin biosynthesis. This was the first example of an ATP-grasp-ligase utilizing peptides as nucleophiles. To explore the potential of this type of enzyme, we performed a BLAST search and identified many orthologs. The orthologs of Streptomyces mobaraensis, Salinispora tropica, and Micromonospora sp. were found in similar gene clusters consisting of six genes. To probe the functions of these genes, we heterologously expressed each of the clusters in Streptomyces lividans and detected novel and structurally similar pseudotripeptides in the broth of all transformants. Moreover, a recombinant PGM1 ortholog of Micromonospora sp. was demonstrated to be a novel dipeptide ligase catalyzing amide bond formation between amidino-arginine and dipeptides to yield tripeptides; this is the first report of a peptide ligase utilizing dipeptides as nucleophiles.

Identification and analysis of the resorcinomycin biosynthetic gene cluster.

Resorcinomycin (1) is composed of a nonproteinogenic amino acid, (S)-2-(3,5-dihydroxy-4-isopropylphenyl)-2-guanidinoacetic acid (2), and glycine. A biosynthetic gene cluster was identified in a genome database of Streptoverticillium roseoverticillatum by searching for orthologs of the genes responsible for biosynthesis of pheganomycin (3), which possesses a (2)-derivative at its N-terminus. The cluster contained a gene encoding an ATP-grasp-ligase (res5), which was suggested to catalyze the peptide bond formation between 2 and glycine. A res5-deletion mutant lost 1 productivity but accumulated 2 in the culture broth. However, recombinant RES5 did not show catalytic activity to form 1 with 2 and glycine as substrates. Moreover, heterologous expression of the cluster resulted in accumulation of only 2 and no production of 1 was observed. These results suggested that a peptide with glycine at its N-terminus may be used as a nucleophile and then maturated by a peptidase encoded by a gene outside of the cluster.

A peptide ligase and the ribosome cooperate to synthesize the peptide pheganomycin.

Peptide antibiotics are typically biosynthesized by one of two distinct machineries in a ribosome-dependent or ribosome-independent manner. Pheganomycin (PGM (1)) and related analogs consist of the nonproteinogenic amino acid (S)-2-(3,5-dihydroxy-4-hydroxymethyl)phenyl-2-guanidinoacetic acid (2) and a proteinogenic core peptide, making their origin uncertain. We report the identification of the biosynthetic gene cluster from Streptomyces cirratus responsible for PGM production. Unexpectedly, the cluster contains a gene encoding multiple precursor peptides along with several genes plausibly encoding enzymes for the synthesis of amino acid 2. We identified PGM1, which has an ATP-grasp domain, as potentially capable of linking the precursor peptides with 2, and validate this hypothesis using deletion mutants and in vitro reconstitution. We document PGM1's substrate permissivity, which could be rationalized by a large binding pocket as confirmed via structural and mutagenesis experiments. This is to our knowledge the first example of cooperative peptide synthesis achieved by ribosomes and peptide ligases using a peptide nucleophile.

A fungal prenyltransferase catalyzes the regular di-prenylation at positions 20 and 21 of paxilline.

A putative indole diterpene biosynthetic gene cluster composed of eight genes was identified in a genome database of Phomopsis amygdali, and from it, biosynthetic genes of fusicoccin A were cloned and characterized. The six genes showed significant similarities to pax genes, which are essential to paxilline biosynthesis in Penicillium paxilli. Recombinants of the three putative prenyltransferase genes in the cluster were overexpressed in Escherichia coli and characterized by means of in vitro experiments. AmyG is perhaps a GGDP synthase. AmyC and AmyD were confirmed to be prenyltransferases catalyzing the transfer of GGDP to IGP and a regular di-prenylation at positions 20 and 21 of paxilline, respectively. AmyD is the first know example of an enzyme with this function. The Km values for AmyD were calculated to be 7.6 ± 0.5 µM for paxilline and 17.9 ± 1.7 µM for DMAPP at a kcat of 0.12 ± 0.003/s.

Functional analysis of a prenyltransferase gene (paxD) in the paxilline biosynthetic gene cluster.

Paxilline is an indole-diterpene produced by Penicillium paxilli. Six genes (paxB, C, G, M, P, and Q) in paxilline biosynthetic gene cluster were previously shown to be responsible for paxilline biosynthesis. In this study, we have characterized paxD, which is located next to paxQ and has weak similarities to fungal dimethylallyl tryptophan synthase genes. PaxD was overexpressed in Escherichia coli and the purified enzyme was used for in vitro analysis. When paxilline and dimethylallyl diphosphate were used as substrates, one major and one minor product, which were identified as di-prenyl paxilline and mono-prenyl paxilline by liquid chromatography-mass spectrometry analysis, were formed. The structure of the major product was determined to be 21,22-diprenylated paxilline, showing that PaxD catalyzed the successive di-prenylation. Traces of both products were detected in culture broth of P. paxilli by liquid chromatography-mass spectrometry analysis. The enzyme is likely to be a dimer and required no divalent cations. The optimum pH and temperature were 8.0 and 37 °C, respectively. The Km values were calculated as 106.4 ± 5.4 µM for paxilline and 0.57 ± 0.02 µM for DMAPP with a kcat of 0.97 ± 0.01/s.

Regiospecificities and prenylation mode specificities of the fungal indole diterpene prenyltransferases AtmD and PaxD.

We recently reported the function of paxD, which is involved in the paxilline (compound 1) biosynthetic gene cluster in Penicillium paxilli. Recombinant PaxD catalyzed a stepwise regular-type diprenylation at the 21 and 22 positions of compound 1 with dimethylallyl diphosphate (DMAPP) as the prenyl donor. In this study, atmD, which is located in the aflatrem (compound 2) biosynthetic gene cluster in Aspergillus flavus and encodes an enzyme with 32% amino acid identity to PaxD, was characterized using recombinant enzyme. When compound 1 and DMAPP were used as substrates, two major products and a trace of minor product were formed. The structures of the two major products were determined to be reversely monoprenylated compound 1 at either the 20 or 21 position. Because compound 2 and ß-aflatrem (compound 3), both of which are compound 1-related compounds produced by A. flavus, have the same prenyl moiety at the 20 and 21 position, respectively, AtmD should catalyze the prenylation in compound 2 and 3 biosynthesis. More importantly and surprisingly, AtmD accepted paspaline (compound 4), which is an intermediate of compound 1 biosynthesis that has a structure similar to that of compound 1, and catalyzed a regular monoprenylation of compound 4 at either the 21 or 22 position, though the reverse prenylation was observed with compound 1. This suggests that fungal indole diterpene prenyltransferases have the potential to alter their position and regular/reverse specificities for prenylation and could be applicable for the synthesis of industrially useful compounds.

Reconstitution of biosynthetic machinery for indole-diterpene paxilline in Aspergillus oryzae.

Indole-diterpenes represented by paxilline share a common pentacyclic core skeleton derived from indole and geranylgeranyl diphosphate. To shed light on the detailed biosynthetic mechanism of the paspaline-type hexacyclic skeleton, we examined the reconstitution of paxilline biosynthetic machinery in Aspergillus oryzae NSAR1. Stepwise introduction of the six pax genes enabled us to isolate all biosynthetic intermediates and to synthesize paxilline. In vitro and in vivo studies on the key enzymes, prenyltransferase PaxC and cyclase PaxB, allowed us to elucidate actual substrates of these enzymes. Using the isolated and the synthesized epoxide substrates, the highly intriguing stepwide epoxidation/cyclization mechanism for the construction of core structure has been confirmed. In addition, we also demonstrated "tandem transformation" to simultaneously introduce two genes using a single vector (paxG/paxB, pAdeA; paxP/paxQ, pUNA). This may provide further option for the reconstitution strategy to synthesize more complex fungal metabolites.

Molecular breeding of a fungus producing a precursor diterpene suitable for semi-synthesis by dissection of the biosynthetic machinery.

Many clinically useful pharmaceuticals are semi-synthesized from natural products produced by actinobacteria and fungi. The synthetic protocols usually contain many complicated reaction steps and thereby result in low yields and high costs. It is therefore important to breed microorganisms that produce a compound most suitable for chemical synthesis. For a long time, desirable mutants have been obtained by random mutagenesis and mass screening. However, these mutants sometimes show unfavorable phenotypes such as low viability and low productivity of the desired compound. Fusicoccin (FC) A is a diterpene glucoside produced by the fungus Phomopsis amygdali. Both FC and the structurally-related cotylenin A (CN) have phytohormone-like activity. However, only CN exhibits anti-cancer activity. Since the CN producer lost its ability to proliferate during preservation, a study on the relationship between structure and activity was carried out, and elimination of the hydroxyl group at position 12 of FC was essential to mimic the CN-like activity. Based on detailed dissection of the biosynthetic machinery, we constructed a mutant producing a compound without a hydroxyl group at position 12 by gene-disruption. The mutant produced this compound as a sole metabolite, which can be easily and efficiently converted into an anti-cancer drug, and its productivity was equivalent to the sum of FC-related compounds produced by the parental strain. Our strategy would be applicable to development of pharmaceuticals that are semi-synthesized from fungal metabolites.

An enzyme catalyzing O-prenylation of the glucose moiety of fusicoccin A, a diterpene glucoside produced by the fungus Phomopsis amygdali.

Isoprenoids form the largest family of compounds found in nature. Isoprenoids are often attached to other moieties such as aromatic compounds, indoles/tryptophan, and flavonoids. These reactions are catalyzed by three phylogenetically distinct prenyltransferases: soluble aromatic prenyltransferases identified mainly in actinobacteria, soluble indole prenyltransferases mostly in fungi, and membrane-bound prenyltransferases in various organisms. Fusicoccin A (FC A) is a diterpene glycoside produced by the plant-pathogenic fungus Phomopsis amygdali and has a unique O-prenylated glucose moiety. In this study, we identified for the first time, from a genome database of P. amygdali, a gene (papt) encoding a prenyltransferase that reversibly transfers dimethylallyl diphosphate (DMAPP) to the 6'-hydroxy group of the glucose moiety of FC A to yield an O-prenylated sugar. An in vitro assay with a recombinant enzyme was also developed. Detailed analyses with recombinant PAPT showed that the enzyme is likely to be a monomer and requires no divalent cations. The optimum pH and temperature were 8.0 and 50 °C, respectively. K(m) values were calculated as 0.49±0.037 µM for FC P (a plausible intermediate of FC A biosynthesis) and 8.3±0.63 µM for DMAPP, with a k(cat) of 55.3±3.3×10⁻³ s. The enzyme did not act on representative substrates of the above-mentioned three types of prenyltransferase, but showed a weak transfer activity of geranyl diphosphate to FC P.

Dioxygenases, key enzymes to determine the aglycon structures of fusicoccin and brassicicene, diterpene compounds produced by fungi.

Fusicoccin A and cotylenin A are structurally related diterpene glucosides and show a phytohormone-like activity. However, only cotylenin A induces the differentiation of human myeloid leukemia cells. Since the cotylenin A producer lost its ability to proliferate during preservation, a study on the relationship between structure and activity was carried out and a modified fusicoccin A with hydroxyl group at the 3-position showed a similar biological activity with that of cotylenin A. We then searched for an enzyme source that catalyzes the introduction of a hydroxyl group into the 3-position and found that brassicicene C, which is structurally related to fusicoccin A with hydroxyl group at the 3-position, was produced by Alternaria brassicicola ATCC96836. We recently cloned a brassicicene C biosynthetic gene cluster including the genes encoding fusicocca-2,10(14)-diene synthase and two cytochrome P450s, which were responsible for the formation of fusicocca-2,10(14)-diene-8ß,16-diol. In this study, we report that a α-ketoglutarate dependent dioxygenase, the gene coding for which was located in the cluster, catalyzed a hydroxylation at the 3-position of fusicocca-2,10(14)-diene-8ß,16-diol. On the other hand, a α-ketoglutarate-dependent dioxygenase, which had been identified in a fusicoccin A biosynthetic gene cluster, catalyzed the 16-oxidation of fusicocca-2,10(14)-diene-8ß,16-diol to yield an aldehyde (8ß-hydroxyfusicocca-1,10(14)-dien-16-al), although both dioxygenases had 51% amino acid sequence identity. These findings suggested that the dioxygenases played critical roles for the formation of the fusicoccin A-type and cotylenin A-/brassicicene C-type aglycons. Moreover, we showed that short-chain dehydrogenase/reductase located in the fusicoccin A biosynthetic gene cluster catalyzed the reduction of the aldehyde to yield fusicocca-1,10(14)-diene-8ß,16-diol.

Crystal structure of heterodimeric hexaprenyl diphosphate synthase from Micrococcus luteus B-P 26 reveals that the small subunit is directly involved in the product chain length regulation.

Hexaprenyl diphosphate synthase from Micrococcus luteus B-P 26 (Ml-HexPPs) is a heterooligomeric type trans-prenyltransferase catalyzing consecutive head-to-tail condensations of three molecules of isopentenyl diphosphates (C(5)) on a farnesyl diphosphate (FPP; C(15)) to form an (all-E) hexaprenyl diphosphate (HexPP; C(30)). Ml-HexPPs is known to function as a heterodimer of two different subunits, small and large subunits called HexA and HexB, respectively. Compared with homooligomeric trans-prenyltransferases, the molecular mechanism of heterooligomeric trans-prenyltransferases is not yet clearly understood, particularly with respect to the role of the small subunits lacking the catalytic motifs conserved in most known trans-prenyltransferases. We have determined the crystal structure of Ml-HexPPs both in the substrate-free form and in complex with 7,11-dimethyl-2,6,10-dodecatrien-1-yl diphosphate ammonium salt (3-DesMe-FPP), an analog of FPP. The structure of HexB is composed of mostly antiparallel α-helices joined by connecting loops. Two aspartate-rich motifs (designated the first and second aspartate-rich motifs) and the other characteristic motifs in HexB are located around the diphosphate part of 3-DesMe-FPP. Despite the very low amino acid sequence identity and the distinct polypeptide chain lengths between HexA and HexB, the structure of HexA is quite similar to that of HexB. The aliphatic tail of 3-DesMe-FPP is accommodated in a large hydrophobic cleft starting from HexB and penetrating to the inside of HexA. These structural features suggest that HexB catalyzes the condensation reactions and that HexA is directly involved in the product chain length control in cooperation with HexB.

In vitro synthesis of high molecular weight rubber by Hevea small rubber particles.

Hevea brasiliensis is one of few higher plants producing the commercial natural rubber used in many significant applications. The biosynthesis of high molecular weight rubber molecules by the higher plants has not been clarified yet. Here, the in vitro rubber biosynthesis was performed by using enzymatically active small rubber particles (SRP) from Hevea. The mechanism of the in vitro rubber synthesis was investigated by the molecular weight distribution (MWD). The highly purified SRP prepared by gel filtration and centrifugation in the presence of Triton((R)) X-100 showed the low isopentenyl diphosphate (IPP) incorporation for the chain extension mechanism of pre-existing rubber. The MWD of in vitro rubber elongated from the pre-existing rubber chains in SRP was analyzed for the first time in the case of H. brasiliensis by incubating without the addition of any initiator. The rubber transferase activity of 70% incorporation of the added IPP (w/w) was obtained when farnesyl diphosphate was present as the allylic diphosphate initiator. The in vitro synthesized rubber showed a typical bimodal MWD of high and low molecular weight fractions in GPC analysis, which was similar to that of the in vivo rubber with peaks at around 10(6) and 10(5) Da or lower. The reaction time independence and dependence of molecular weight of high and low molecular weight fractions, respectively, indicated that the high molecular weight rubber was synthesized from the chain extension of pre-existing rubber molecules whereas the lower one was from the chain elongation of rubber molecules newly synthesized from the added allylic substrates.

Cloning and functional analysis of cis-prenyltransferase from Thermobifida fusca.

cis-Prenyltransferase catalyzes the synthesis of Z,E-mixed prenyl diphosphates by a condensation of isopentenyl diphosphate to an allylic diphosphate. A novel gene encoding a cis-prenyltransferase is cloned from Thermobifida fusca. It showed a unique substrate specificity accepting dimethylallyl diphosphate as a shortest allylic substrate, and synthesizes polyprenyl products up to C(70).

Effect of mutagenesis at the region upstream from the G(Q/E) motif of three types of geranylgeranyl diphosphate synthase on product chain-length.

(All-E) geranylgeranyl diphosphate synthases have been classified into three types based on the characteristic sequences around the first aspartate rich motif, which is highly conserved among the enzymes. In type I geranylgeranyl diphosphate synthases, which consist of archaeal enzymes, a bulky amino acid residue at the 5th position upstream from the motif plays a main role in the product determination, by blocking further elongation of prenyl chain as the bottom of the reaction cavity. On the other hand, type III geranylgeranyl diphosphate synthases, which consist of the enzymes from eukaryotes except for plants, use a bulky amino acid residue at the 2nd position upstream from the conserved G(Q/E) motif for product chain-length determination. Thus we introduced mutations into the region upstream from the G(Q/E) motif of geranylgeranyl diphosphate synthases of the three different types to confirm the importance of the region for the product chain-length determination. The results of the mutational analyses indicated that not only the 2nd but also the 3rd position upstream from the G(Q/E) motif is involved in the product chain-length determination mechanism in types I and III geranylgeranyl diphosphate synthases, while the amino acid substitution in this region did not affect the chain-length of the products of type II geranylgeranyl diphosphate synthase, which consist of the enzymes from bacteria and plants. The region upstream from the G(Q/E) motif possibly contributes to the product determination in the wide range of geranylgeranyl diphosphate synthases, as well as that around the first aspartate rich motif.

Product chain-length determination mechanism of Z,E-farnesyl diphosphate synthase.

cis-Prenyltransferases catalyze the consecutive condensation of isopentenyl diphosphate (IPP) with allylic prenyl diphosphates, producing Z,E-mixed prenyl diphosphate. The Mycobacterium tuberculosis Z,E-farnesyl diphosphate synthase Rv1086 catalyzes the condensation of one molecule of IPP with geranyl diphosphate to yield Z,E-farnesyl diphosphate and is classified as a short-chain cis-prenyltransferase. To elucidate the chain-length determination mechanism of the short-chain cis-prenyltransferase, we introduced some substitutive mutations at the characteristic amino acid residues of Rv1086. Among the mutants constructed, L84A showed a dramatic change of catalytic function to synthesize longer prenyl chain products than that of wild type, indicating that Leu84 of Rv1086 plays an important role in product chain-length determination. Mutagenesis at the corresponding residue of a medium-chain cis-prenyltransferase, Micrococcus luteus B-P 26 undecaprenyl diphosphate synthase also resulted in the production of different prenyl chain length from the intrinsic product, suggesting that this position also plays an important role in product chain-length determination for medium-chain cis-prenyltransferases.

Cloning and functional analysis of novel short-chain cis-prenyltransferases.

cis-Prenyltransferase catalyzes the synthesis of Z,E-mixed prenyl diphosphates by sequential condensation of isopentenyl diphosphate with allylic diphosphate. cis-Prenyltransferases can be classified into three subgroups: short-, medium-, and long-chain cis-prenyltransferase, according to their product chain lengths. cis-Farnesyl diphosphate synthase from Mycobacterium tuberculosis has been the only example as short-chain cis-prenyltransferase so far characterized. In this study, we cloned the novel short-chain cis-prenyltransferases from three different bacteria, and characterized their enzymatic activities to compare and elucidate a common feature of the short-chain cis-prenyltransferases. Furthermore, we identified a specific isoleucine that is conserved in short-chain cis-prenyltransferases and located in close proximity of the omega-end of the geranyl diphosphate. Several site-directed mutants with respect to the isoleucine residue synthesized longer prenyl chain products and showed broader allylic substrate specificity. These results suggested that the isoleucine plays an important role in the substrate specificity and chain length determination mechanism of cis-prenyltransferase.

The product chain length determination mechanism of type II geranylgeranyl diphosphate synthase requires subunit interaction.

The product chain length determination mechanism of type II geranylgeranyl diphosphate synthase from the bacterium, Pantoea ananatis, was studied. In most types of short-chain (all-E) prenyl diphosphate synthases, bulky amino acids at the fourth and/or fifth positions upstream from the first aspartate-rich motif play a primary role in the product determination mechanism. However, type II geranylgeranyl diphosphate synthase lacks such bulky amino acids at these positions. The second position upstream from the G(Q/E) motif has recently been shown to participate in the mechanism of chain length determination in type III geranylgeranyl diphosphate synthase. Amino acid substitutions adjacent to the residues upstream from the first aspartate-rich motif and from the G(Q/E) motif did not affect the chain length of the final product. Two amino acid insertion in the first aspartate-rich motif, which is typically found in bacterial enzymes, is thought to be involved in the product determination mechanism. However, deletion mutation of the insertion had no effect on product chain length. Thus, based on the structures of homologous enzymes, a new line of mutants was constructed in which bulky amino acids in the alpha-helix located at the expected subunit interface were replaced with alanine. Two mutants gave products with longer chain lengths, suggesting that type II geranylgeranyl diphosphate synthase utilizes an unexpected mechanism of chain length determination, which requires subunit interaction in the homooligomeric enzyme. This possibility is strongly supported by the recently determined crystal structure of plant type II geranylgeranyl diphosphate synthase.

An alternative mechanism of product chain-length determination in type III geranylgeranyl diphosphate synthase.

(All-E) prenyl diphosphate synthases catalyze the consecutive condensation of isopentenyl diphosphates with allylic prenyl diphosphates, producing products with various chain-lengths that are unique for each enzyme. Some short-chain (all-E) prenyl diphosphate synthases, i.e. farnesyl diphosphate synthases and geranylgeranyl diphosphate synthases contain characteristic amino acid sequences around the allylic substrate binding sites, which have been shown to play a role in determining the chain-length of the product. However, among these enzymes, which are classified into several types based on the possessive patterns of such characteristics, type III geranylgeranyl diphosphate synthases, which consist of enzymes from eukaryotes (excepting plants), lack these features. In this study, we report that mutagenesis at the second position before the conserved G(Q/E) motif, which is distant from the well-studied region, affects the chain-length of the product for a type III geranylgeranyl diphosphate synthase from Saccharomyces cerevisiae. This clearly suggests that a novel mechanism is operative in the product determination for this type of enzyme. We also show herein that mutagenesis at the corresponding position of an archaeal medium-chain enzyme also alters its product specificity. These results provide valuable information on the molecular evolution of (all-E) prenyl diphosphate synthases.