Application of a Newly Identified and Characterized 18-O-Acyltransferase in Chemoenzymatic Synthesis of Selected Natural and Nonnatural Bioactive Derivatives of Phoslactomycins

Phoslactomycins (PLMs) and related leustroducsins (LSNs) have been isolated from a variety of bacteria based on antifungal, anticancer, and other biological assays. Streptomyces sp. strain HK 803 produces five PLM analogs (PLM A and PLMs C to F) in which the C-18 hydroxyl substituent is esterified with a range of branched, short-alkyl-chain carboxylic acids. The proposed pathway intermediate, PLM G, in which the hydroxyl residue is not esterified has not been observed at any significant level in fermentation, and the only route to this potentially useful intermediate has been an enzymatic deacylation of other PLMs and LSNs. We report that deletion of plmS₃ from the PLM biosynthetic cluster gives rise to a mutant which accumulates the PLM G intermediate. The 921-bp plmS₃ open reading frame was cloned and expressed as an N-terminally polyhistidine-tagged protein in Escherichia coli and shown to be an 18-O acyltransferase, catalyzing conversion of PLM G to PLM A, PLM C, and PLM E using isobutyryl coenzyme A (CoA), 3-methylbutyryl-CoA, and cyclohexylcarbonyl-CoA, respectively. The efficiency of this process (k_cat of 28 ± 3 min⁻¹ and K_m of 88 ± 16 μM) represents a one-step chemoenzymatic alternative to a multistep synthetic process for selective chemical esterification of the C-18 hydroxy residue of PLM G. PlmS₃ was shown to catalyze esterification of PLM G with CoA and N-acetylcysteamine thioesters of various saturated, unsaturated, and aromatic carboxylic acids and thus also to provide an efficient chemoenzymatic route to new PLM analogs.Attachment of either short (C₂ to C₆) or medium (C₈ to C₁₂) acyl chains to both amine and alcohol moieties on polyketide and polypeptide natural products can represent a key step in generating the final biologically active molecule. This step is often, but not always, one of the later biosynthetic steps and is catalyzed by an acyltransferase. The corresponding gene is typically associated with the polyketide or polypeptide biosynthetic gene cluster. Despite the importance of this step, a relatively small number of these acyltransferases from actinomycetes have been identified, and very few have been fully characterized (2, 10, 15, 18).Studies of biosynthetic processes where there is an acylation of a polyketide chain, have indicated that the enzymes have various degrees of promiscuity for the carboxylic acid substrates. For instance, genetic evidence has shown that mdmB and acyA encode 3-O-acyltransferases which transfer either acetyl or propionyl groups to position 3 in 16-membered macrolides such as midecamycin, spiramycin, carbomycin, and tylosin (3, 10). The asm19 gene product has been identified as the 3-O-acyltransferase which catalyzes the attachment of the biologically essential acyl group in the macrocyclic ansamitocins (18). The asm19 mutant accumulates an N-demethyl-4,5-desepoxymytansinol, indicating that acylation of the macrocycle precedes N methylation and epoxidation. Escherichia coli cell extracts containing a recombinant Asm19 protein have been shown to catalyze acylation of this mytansinol intermediate using a range of short straight- and branched-chain acyl coenzyme A (CoA) thioesters (C₂ to C₅). Finally, LovD, which catalyzes the acylation of the C-8 hydroxyl group of monacolin J to yield the natural product lovastatin, a pharmaceutically important fungal polyketide product produced by Aspergillus nidulans, has been characterized (30). This enzyme is able to utilize a wide range of different acyl donors activated as CoA, N-acetylcysteamine (NAC), or methyl thioglycolate esters and thus offers an economically attractive route for generating novel lovasotatin analogs for treatment of hypercholesterolemia.One or multiple O-acyltransferases have been implied to be involved in the post-polyketide synthase tailoring steps leading to a series of natural products known as the phoslactomycins (PLMs) (Fig. ​(Fig.1)1) (6). These compounds (also known as leustroducsins [LSNs], phospholines, and phosphazomycins) have been isolated from various actinomycetes, and their structures are all identical with the exception of the acyl substituent at C-18 (4, 5, 12, 13, 20, 27, 28). Streptomyces sp. strain HK 803 produces at least five such acylated analogs (PLM A and PLMs C to F) (Fig. ​(Fig.1)1) as well as PLM B, in which the C-18 hydroxyl substituent is absent (4, 5, 27).Open in a separate windowFIG. 1.Proposed biosynthetic relationship between PLM products made by Streptomyces sp. strain HK 803. A cytochrome P450 monooxygenase (PlmS₂) catalyzes C-18 hydroxylation of PLM B to generate PLM G, which is subsequently 18-O acylated by PlmS₃.These natural products have been isolated based on their potent activity (as low as 0.008 μg/ml) against some phytopathogenic fungi (27, 28). The compounds also have relatively weak antitumor activity (50% inhibitory concentration of 2 to 3 μg/ml against L1210, P38,8, and El-4 cell lines) (19) which may arise from their activity as selective inhibitors of protein phosphatase 2A. (26). These natural products also show induction of a colony-stimulating factor (12) via NF-κB activation and thrombopoiesis (14). This array of promising biological activities has stimulated research into the field of PLMs for treatments of various diseases. Low yields and the presence of multiple acylated products from fermentations have provided a barrier to this work, and circuitous routes to obtaining individual compounds have been described. For instance, there have been no reports of any actinomycetes which produce PLM G (LSN H) (Fig. ​(Fig.1),1), in which the C-18 hydroxyl residue is present but not acylated, and this compound has been reported to be obtained only by cleavage of the acyl groups of a mixture of other PLMs and LSNs using porcine liver esterase (24). A multistep synthetic route to selectively acylate PLM G with 6-methyloctanoic acid (producing LSN B) has also been described (17).Recently the entire 75-kbp Plm biosynthetic gene cluster has been cloned, sequenced, and analyzed (21) and has provided an opportunity to study the enzymatic hydroxylation and acylation processes which give rise to the range of PLM products. Deletion of the plmS₂ open reading frame (ORF), showing high sequence similarity to bacterial cytochrome P450 monooxygenases, has resulted in an NP1 mutant producing only PLM B (Fig. ​(Fig.1)1) (21). The plmS₂ ORF has been expressed as an N-terminally polyhistidine-tagged protein in Streptomyces coelicolor, and the purified protein has been shown to catalyze conversion of PLM B to PLM G (7). This work in conjunction with other studies (1, 23) has led to a proposal that the final two biosynthetic steps involve hydroxylation of PLM B (to give PLM G) and subsequent acylation with a broad range of acyl-CoA substrates (to give PLM A and PLMs C to F). The acylation step is required for potent antifungal activity of the PLMs. Initial analysis of the PLM biosynthetic gene cluster (21) did not reveal a candidate gene or genes whose products might be responsible for this acylation.Here we identify the plmS₃ gene product as the singular acyltransferase in Streptomyces sp. strain HK 803 responsible for C-18 acylation of PLM G. Generation of a plmS₃ deletion mutant results in selective production of PLM G, supporting the proposed role of this gene product and providing the first direct fermentation method to access this intermediate. The 921-bp plmS₃ ORF was cloned and expressed as an N-terminally polyhistidine-tagged protein in E. coli, and the recombinant purified protein was shown to catalyze acylation of PLM G with isobutyryl-CoA, 3-methylbutyryl-CoA, and cyclohexylcarbonyl-CoA to give PLM A, PLM C, and PLM E, respectively. This efficient one-step enzymatic process offers an attractive alternative to the multistep synthetic process for selective acylation of PLM G. PlmS₃ was also shown to catalyze esterification of PLM G with CoA and NAC thioesters with a remarkably wide range of various saturated, unsaturated, and aromatic carboxylic acids and thus provides an efficient chemoenzymatic route to new PLM analogs.     

Non-contiguous finished genome sequence of Aminomonas paucivorans type strain (GLU-3)

Aminomonas paucivorans Baena et al. 1999 is the type species of the genus Aminomonas, which belongs to the family Synergistaceae. The species is of interest because it is an asaccharolytic chemoorganotrophic bacterium which ferments quite a number of amino acids. This is the first finished genome sequence (with one gap in a rDNA region) of a member of the genus Aminomonas and the third sequence from the family Synergistaceae. The 2,630,120 bp long genome with its 2,433 protein-coding and 61 RNA genes is a part of the GenomicEncyclopedia ofBacteria andArchaea project.     

Complete genome sequence of Thermaerobacter marianensis type strain (7p75a)

Thermaerobacter marianensis Takai et al. 1999 is the type species of the genus Thermaerobacter, which belongs to the Clostridiales family Incertae Sedis XVII. The species is of special interest because T. marianensis is an aerobic, thermophilic marine bacterium, originally isolated from the deepest part in the western Pacific Ocean (Mariana Trench) at the depth of 10.897m. Interestingly, the taxonomic status of the genus has not been clarified until now. The genus Thermaerobacter may represent a very deep group within the Firmicutes or potentially a novel phylum. The 2,844,696 bp long genome with its 2,375 protein-coding and 60 RNA genes consists of one circular chromosome and is a part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Olsenella uli type strain (VPI D76D-27C)

Olsenella uli (Olsen et al. 1991) Dewhirst et al. 2001 is the type species of the genus Olsenella, which belongs to the actinobacterial family Coriobacteriaceae. The species is of interest because it is frequently isolated from dental plaque in periodontitis patients and can cause primary endodontic infection. The species is a Gram-positive, non-motile and non-sporulating bacterium. The strain described in this study was isolated from human gingival crevices. This is the first completed sequence of the genus Olsenella and the fifth sequence from a member of the family Coriobacteriaceae. The 2,051,896 bp long genome with its 1,795 protein-coding and 55 RNA genes is a part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Acidaminococcus fermentans type strain (VR4)

Acidaminococcus fermentans (Rogosa 1969) is the type species of the genus Acidaminococcus, and is of phylogenetic interest because of its isolated placement in a genomically little characterized region of the Firmicutes. A. fermentans is known for its habitation of the gastrointestinal tract and its ability to oxidize trans-aconitate. Its anaerobic fermentation of glutamate has been intensively studied and will now be complemented by the genomic basis. The strain described in this report is a nonsporulating, nonmotile, Gram-negative coccus, originally isolated from a pig alimentary tract. Here we describe the features of this organism, together with the complete genome sequence, and annotation. This is the first complete genome sequence of a member of the family Acidaminococcaceae, and the 2,329,769 bp long genome with its 2,101 protein-coding and 81 RNA genes is part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Meiothermus ruber type strain (21)

Meiothermus ruber (Loginova et al. 1984) Nobre et al. 1996 is the type species of the genus Meiothermus. This thermophilic genus is of special interest, as its members share relatively low degrees of 16S rRNA gene sequence similarity and constitute a separate evolutionary lineage from members of the genus Thermus, from which they can generally be distinguished by their slightly lower temperature optima. The temperature related split is in accordance with the chemotaxonomic feature of the polar lipids. M. ruber is a representative of the low-temperature group. This is the first completed genome sequence of the genus Meiothermus and only the third genome sequence to be published from a member of the family Thermaceae. The 3,097,457 bp long genome with its 3,052 protein-coding and 53 RNA genes is a part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Haloterrigena turkmenica type strain (4k)

Haloterrigena turkmenica (Zvyagintseva and Tarasov 1987) Ventosa et al. 1999, comb. nov. is the type species of the genus Haloterrigena in the euryarchaeal family Halobacteriaceae. It is of phylogenetic interest because of the yet unclear position of the genera Haloterrigena and Natrinema within the Halobacteriaceae, which created some taxonomic problems historically. H. turkmenica, was isolated from sulfate saline soil in Turkmenistan, is a relatively fast growing, chemoorganotrophic, carotenoid-containing, extreme halophile, requiring at least 2 M NaCl for growth. Here we describe the features of this organism, together with the complete genome sequence, and annotation. This is the first complete genome sequence of the genus Haloterrigena, but the eighth genome sequence from a member of the family Halobacteriaceae. The 5,440,782 bp genome (including six plasmids) with its 5,287 protein-coding and 63 RNA genes is part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Methanoplanus petrolearius type strain (SEBR 4847)

Segniliparus rotundus Butler 2005 is the type species of the genus Segniliparus, which is currently the only genus in the corynebacterial family Segniliparaceae. This family is of large interest because of a novel late-emerging genus-specific mycolate pattern. The type strain has been isolated from human sputum and is probably an opportunistic pathogen. Here we describe the features of this organism, together with the complete genome sequence and annotation. This is the first completed genome sequence of the family Segniliparaceae. The 3,157,527 bp long genome with its 3,081 protein-coding and 52 RNA genes is part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Complete genome sequence of Sebaldella termitidis type strain (NCTC 11300)

Sebaldella termitidis (Sebald 1962) Collins and Shah 1986, is the only species in the genus Sebaldella within the fusobacterial family 'Leptotrichiaceae'. The sole and type strain of the species was first isolated about 50 years ago from intestinal content of Mediterranean termites. The species is of interest for its very isolated phylogenetic position within the phylum Fusobacteria in the tree of life, with no other species sharing more than 90% 16S rRNA sequence similarity. The 4,486,650 bp long genome with its 4,210 protein-coding and 54 RNA genes is part of the Genomic Encyclopedia of Bacteria and Archaea project.     

Genetic manipulation of the biosynthetic process leading to phoslactomycins, potent protein phosphatase 2A inhibitors

Phoslactomycins (PLMs) represent an unusual structural class of natural products secreted by various streptomycetes, containing an α,β-unsaturated δ-lactone, an amino group, phosphate ester, conjugated diene and a cyclohexane ring. Phosphazomycins, phospholines and leustroducsins contain the same structural moieties, varying only in the acyl substituent at the C-18 hydroxyl position. These compounds possess either antifungal or antitumor activities or both. The antitumor activity of the PLM class of compounds has been attributed to a potent and selective inhibition of protein phosphatase 2A (PP2A). The cysteine-269 residue of PP2Ac-subunit has been shown to be the site of covalent modification by PLMs. In this article, we review previous work on the isolation, structure elucidation and biological activities of PLMs and related compounds and current status of our work on both PLM stability and genetic manipulation of the biosynthetic process. Our work has shown that PLM B is surprisingly stable in solution, with a pH optimum of 6. Preliminary biosynthetic studies utilizing isotopically labeled shikimic acid and cyclohexanecarboxylic acid (CHC) suggested PLM B to be a polyketide-type antibiotic synthesized using CHC as a starter unit. Using a gene (chcA) from a set of CHC-CoA biosynthesis genes from Streptomyces collinus as a probe, a 75 kb region of 29 ORFs encoding PLM biosynthesis was located in the genome of Streptomyces sp. strain HK803. Analysis and subsequent manipulation of plmS ₂ and plmR ₂ in the gene cluster has allowed for rational engineering of a strain that produces only one PLM analog, PLM B, at ninefold higher titers than the wild type strain. A strain producing PLM G (the penultimate intermediate in PLMs biosynthesis) has also been generated. Current work is aimed at selective in vitro acylation of PLM G with various carboxylic acids and a precursor-directed biosynthesis in a chcA deletion mutant with the aim of generating novel PLM analogs.