Skip to main content
Microbiology logoLink to Microbiology
. 2015 Jul;161(Pt 7):1338–1347. doi: 10.1099/mic.0.000095

Engineered biosynthesis of enduracidin lipoglycopeptide antibiotics using the ramoplanin mannosyltransferase Ram29

Ming-Cheng Wu 1, Matthew Q Styles 1, Brian J C Law 1, Anna-Winona Struck 1, Laura Nunns 1, Jason Micklefield 1,
PMCID: PMC4635501  PMID: 25878261

Abstract

The lipopeptides ramoplanin from Actinoplanes sp. ATCC 33076 and enduracidin produced by Streptomyces fungicidicus are effective antibiotics against a number of drug-resistant Gram-positive pathogens. While these two antibiotics share a similar cyclic peptide structure, comprising 17 amino acids with an N-terminal fatty acid side chain, ramoplanin has a di-mannose moiety that enduracidin lacks. The mannosyl substituents of ramoplanin enhance aqueous solubility, which was important in the development of ramoplanin as a potential treatment for Clostridium difficile infections. In this study we have determined the function of the putative mannosyltransferase encoded by ram29 from the ramoplanin biosynthetic gene cluster. Bioinformatics revealed that Ram29 is an integral membrane protein with a putative DxD motif that is suggested to bind to, and activate, a polyprenyl phosphomannose donor and an extracytoplasmic C-terminal domain that is predicted to bind the ramoplanin aglycone acceptor. The ram29 gene was cloned into the tetracycline inducible plasmid pMS17 and integrated into the genome of the enduracidin producer S. fungicidicus. Induction of ram29 expression in S. fungicidicus resulted in the production of monomannosylated enduracidin derivatives, which are not present in the WT strain. Tandem MS analysis showed that mannosylation occurs on the Hpg11 residue of enduracidin. In addition to confirming the function of Ram29, these findings demonstrate how the less common, membrane-associated, polyprenyl phosphosugar-dependent glycosyltransferases can be used in natural product glycodiversification. Such a strategy may be valuable in future biosynthetic engineering approaches aimed at improving the physico-chemical and biological properties of bioactive secondary metabolites including antibiotics.

Introduction

There is an urgent need for new antibiotics to combat emerging drug-resistant pathogens. The majority of antibiotics that are used to treat infectious diseases are natural products or derivatives thereof (Wu et al., 2012). Natural products often require further chemical modification, to improve their biological activities or physico-chemical properties for therapeutic applications. However, many of the most promising natural products, such as the polyketides and nonribosomal peptides, are highly complex molecules that offer limited opportunity for semisynthesis and are invariably inaccessible through total synthesis on the scale required for drug development. Consequently, alternative biosynthetic engineering approaches are required that can enable the structural diversification and optimization of promising natural product scaffolds.

Biosynthetic engineering utilizes knowledge of the structure, organization and function of biosynthetic enzymes to reprogramme biosynthetic pathways at the genetic level, to produce new and potentially improved natural products (Wilkinson & Micklefield, 2007; Wu et al., 2012). Previously, we adopted biosynthetic engineering approaches to diversify the calcium-dependent lipopeptide antibiotics (CDAs), leading to a library of ‘non-natural’ CDAs (Amir-Heidari et al., 2008; Lewis et al., 2011; Mahlert et al., 2007; Milne et al., 2006; Neary et al., 2007; Powell et al., 2007a, b; Thirlway et al., 2012; Uguru et al., 2004). In this work we focus on the biosynthesis of related lipoglycopeptide antibiotics: ramoplanin from Actinoplanes sp. ATCC 33076 (McCafferty et al., 2002), and the enduracidins produced by Streptomyces fungicidicus (Yin & Zabriskie, 2006).

Ramoplanin (Fig. 1) is a highly effective antibiotic against a number of Gram-positive bacterial pathogens such as methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci (McCafferty et al., 2002; Walker et al., 2005), and has entered phase III clinical trials for the treatment of Clostridium difficile infections. Enduracidin shares a similar structure with ramoplanin (Fig. 1) and is also a potent antibiotic; both antibiotics bind to lipid II, blocking the subsequent transglycosylation reactions during peptidoglycan biosynthesis of the bacterial cell wall (Cudic et al., 2002; Fang et al., 2006; Hamburger et al., 2009). However, enduracidin has poor aqueous solubility compared with ramoplanin (Cudic et al., 2002). Consequently, whilst ramoplanin has entered clinical trials, enduracidin is instead utilized as an additive in animal feed (Castiglione et al., 2005; McCafferty et al., 2002; Yin & Zabriskie, 2006). The main structural difference between the two antibiotics is the presence of mannosyl groups on hydroxyphenylglycine11 (Hpg11) that contribute significantly to the solubility of ramoplanin but are absent in enduracidin. The mannosyl groups of ramoplanin are not required for antibacterial activity (Cudic et al., 2002) as the aglycone retains similar potency to the parent lipoglycopeptide. However the mannosyl groups of ramoplanin, in addition to enhancing solubility, contribute to the hydrolytic stability of the peptide core (Cudic et al., 2002). In addition the mannosyl groups can provide a chemical handle for semisynthesis. For example, the mannosyl groups of the mannopeptimycins have been selectively derivatized to generate semisynthetic products with increased activity (He, 2005).

Fig. 1. Structures of the nonribosomal peptide antibiotics enduracidin, ramoplanin, mannopeptimycin and the dalbavancin precursor (A40926). All, except enduracidin, possess mannosyl substituents (in green) derived from putative polyprenyl phosphomannose-dependent glycosyltransferases Ram29, MppH/I and Dbv20 respectively.

Fig. 1.

Within the ramoplanin (ram) biosynthetic gene cluster there is a gene, ram29, encoding an integral membrane protein, that is absent from the enduracidin (end) cluster. Genes encoding similar proteins are also present within the gene clusters for other important mannosylated products, including the mannopeptimycins (mppH/I) (Magarvey et al., 2006), and the commercial glycopeptide antibiotics teicoplanin (tcp15) (Sosio et al., 2004) and dalbavancin (dbv20) (Sosio et al., 2003) (Fig. 1). The putative mannosylating enzymes have not been characterized, but are suggested to transfer mannose (Man) from a polyprenyl phosphomannose (PPM) donor within the membrane (Magarvey et al., 2006). A recent report (Chen et al., 2013) described the deletion of ram29 from the ramoplanin producer, Actinoplanes sp. ATCC 33076. The resultant deletion strain was shown to produce the ramoplanin aglycone, suggesting that Ram29 is responsible for the incorporation of the mannose sugars in ramoplanin. In this paper, we have established conditions for the heterologous expression of ram29 in the enduracidin-producing strain of S. fungicidicus, confirming the mannosyltransferase function of Ram29, and also providing access to novel mannosylated enduracidin variants.

Methods

Bioinformatics analysis

The topology of Ram29 was predicted using bioinformatics programs which are applied widely to membrane protein analysis, namely tmpred (Hofmann & Stoffel, 1993), hmmtop (Tusnády & Simon, 2001), tmhmm (Möller et al., 2001), scampi (Bernsel et al., 2008), prodiv (Viklund & Elofsson, 2004), pro (Viklund & Elofsson, 2004), octopus (Viklund & Elofsson, 2008) and topcons (Viklund & Elofsson, 2004). The DNA sequence was analysed using Vector NTI software V-10.1.1 (Invitrogen), and the blast and alignment analyses were conducted using the alignment program from the same software.

Bacterial strains, microbiological methods and genetic manipulations

The bacterial strains S. fungicidicus (ATCC 31731) and Actinoplanes sp. (ATCC 33076) were purchased from the ATCC bioresource centre (LGC Standards) and cultured in ISP Medium 2 and ISP Medium 1 broth respectively (Shirling & Gottlieb, 1966). Prior to preparation of spore stocks, both strains were cultured on mannitol soya (MS) flour media for 2 weeks (Hobbs et al., 1989). Mycelia from 1 ml of culture were used for genomic DNA extraction using a Qiagen Genomic DNA extraction kit (Qiagen) following the manufacturer's protocol for bacterial genomic DNA extraction. The genomic DNA was redissolved in 100 μl of 10 mM Tris/HCl buffer pH 8.5 and used as a template for PCR. The gene-specific primers ram29-F1, ram29-R1, ram29-F2 and ram29-R2 (Table S1, available in the online Supplementary Material) were used to amplify the ram29 sequence. The resultant PCR product was digested with BamHI alone or doubly digested with EcoRV and NsiI and cloned into the similarly digested pIJ86 and pMS17 to yield pIJ86-ram29 and pMS17-ram29 (Fig. S1). pIJ86 is a non-integrative vector with a constitutive ermE* promoter which has been widely used for protein expression in Streptomyces spp. (Bibb et al., 1985). pMS17 is a conjugative and integrative plasmid containing a tetracycline (Tc)-inducible promoter (kindly provided by Professor M. Smith, University of York) (Rodríguez-García et al., 2005; Wehmeier et al., 2009). Escherichia coli TOP10 (Invitrogen) was used for cloning procedures. The integrity of pIJ86-ram29 and pMS17-ram29 was confirmed by DNA sequencing (GATC Biotech) prior to being transformed into the dam, dcm and hsdM methylation-deficient E. coli strain ET12567 carrying the non-transmissible pUZ8002 (MacNeil et al., 1992). Intergeneric conjugation was carried out between E. coli ET12567/pUZ8002 harbouring either pIJ86-ram29 or pMS17-ram29 and S. fungicidicus (Kieser et al., 2000). Following 3–6 days of incubation at 30 °C exconjugants were transferred to selective MS agar plates supplemented with apramycin (100 μg ml− 1) and nalidixic acid (25 μg ml− 1). Following 2 weeks of incubation at 30 °C, spores derived from the exconjugants were transferred to bouillon broth (1 % meat extract, 1 % peptone, 0.5 % NaCl, 0.2 % K2HPO4, pH 7.0; modified from ATCC Medium 841) and incubated at 30 °C for 2–4 days. The mycelia were pelleted by centrifugation and the genomic DNA extracted for use in sequence confirmation.

Isolation of enduracidin from S. fungicidicus fermentation media

The spores of WT or exconjugant S. fungicidicus strains (105–107) were used to inoculate 75 ml of seed medium (Bouillon broth) and cultured for 3 days at 30 °C. Ten millilitres of seed culture was then transferred to 300 ml of enduracidin fermentation medium (Nogami et al., 1984) supplemented with the appropriate antibiotics and cultured at 30 °C with shaking for 14–21 days. For exconjugants harbouring pMS17-ram29, tetracycline was added to a final concentration of 0.1 μg ml− 1 every 3 days. Following fermentation, cultures were transferred to 50 ml Falcon tubes and the mycelia pelleted by centrifugation (15 min, 10 000 g). Pellets were resuspended in 70 % acidic methanol (pH 2.0) and stirred for 3 h. Supernatants of the methanol extracts were obtained by centrifugation (10 min, 10 000 g) and proteins and lipids extracted with an equal volume of ethyl acetate. Following separation, the pH of the aqueous partition was adjusted to 8.2 and the enduracidin extracted using 300 ml (3 × 100 ml) of butanol. The butanol phase, containing enduracidin, was washed with water (3 × 100 ml), and then the enduracidin was partitioned into 300 ml (3 × 100 ml) of 5 mM HCl aqueous solution (pH 2.0). The pH of the aqueous layer was adjusted to 8.2 and the enduracidin extracted with 300 ml (3 × 100 ml) of butanol. The butanol layer containing enduracidin was separated and concentrated by evaporation under reduced pressure, leaving a crude white powder, which was redissolved in 20 % acetonitrile. Enduracidin was further purified by reverse-phase high-performance liquid chromatography (RP-HPLC) using a Gemini C18 column (10 × 250 mm, 5 μm; Phenomenex) using a gradient composed of water with 0.1 % formic acid (A) and acetonitrile with 0.1 % formic acid (B) as follows: 10 min linear gradient from A/B (95 : 5 %) to A/B (80 : 20 %), 30 min linear gradient to A/B (60 : 40 %), 5 min linear gradient to A/B (5 : 95 %), held at A/B (5 : 95 %) for 4 min, 3 min linear gradient to A/B (95 : 5 %) before re-equilibration at A/B (95 : 5 %) over 8 min. A flow rate of 3 ml min− 1 was used and elution of enduracidin was monitored at A230. Fractions were collected and further analysed by liquid chromatography (LC)-MS.

Analysis of enduracidin extracts using LC-MS

LC-MS analysis was carried out on a Micromass LCT TOF mass spectrometer (Waters), equipped with an electrospray ionization source run in positive mode (scanning from m/z 700 to 2500) combined with a Waters 2790 separation module. Samples were dissolved in acetonitrile/water/formic acid (50/50/0.1). HPLC was carried out using a Gemini C18 column (4.6 × 150 mm, 3 μm; Phenomenex) at a flow rate of 1 ml min− 1 using a gradient composed of water with 0.1 % formic acid (A) and acetonitrile with 0.1 % formic acid (B) as follows: 10 min linear gradient from A/B (80 : 20 %) to A/B (30 : 70 %), 1 min gradient to A/B (5 : 95 %) and held at A/B (5 : 95 %) for further 4 min, before re-equilibration at A/B (80 : 20 %) for 4 min.

Tandem MS

Tandem MS analysis was carried out on a Q-TOF Ultima Global instrument (Waters) by direct infusion of a purified sample dissolved in acetonitrile/water/formic acid (50/50/0.1). Borosilicate nanoemitters (Proxeon) were loaded with sample using gel loading pipette tips. The nanoemitters were mounted onto the source and then activated at 0.4 kV capillary voltage by applying pressure against the sample cone, snapping the glass fibre. Once activated, the capillary voltage was ramped steadily until a consistent ion beam was observed, typically between 1.6 and 1.8 kV. Each nanoemitter holds approximately 20 μl and provides a spray for approximately 30 min duration, giving a flow rate of approximately 0.66 μl min− 1. The instrument was calibrated using the tandem mass spectrum of [Glu]-Fibrinopeptide B.

Once an ion beam was achieved, the MS source was tuned to the product by ramping the capillary voltage and the cone voltage until the signal intensity peaked. An MS scan was carried out to determine a suitable m/z value to fragment. An MS-MS scan was then carried out by selecting the triply charged [M+3H]3+ ion (785 for the WT enduracidin and 839 for the mannosylated enduracidin) as the parent ion for analysis, and raising the collision energy within the collision cell until the parent ion was a minority peak on the resulting spectrum. The data were collected over 20 min, or until a sufficient spectrum was achieved for sequencing analysis. The subsequent spectrum was background subtracted (polynomial order = 4, 40 % below curve, tolerance = 0.010), smoothed (Savitzky Golay smoothing, smooth channels = ± 5, number of smooths = 2), and centred (minimum peak width at half height = 4, 80 % centroid top) by the MassLynx software (Version 4.0, Waters Ltd) prior to analysis.

Enduracidin sequencing

A database of theoretical m/z values was set up based on all the possible fragmentation patterns of enduracidin, assuming two cleavages yielding a linear peptide ion. Additionally, a further database was set up assuming that the first cleavage occurs at the ester bond linkage; thus additional fragmentation events yielded a, b, y and z ions at either one or both ends (as both the C- and N-termini have undergone fragmentation). Tandem MS data were then compared with these theoretical values to generate ion series that can be used to characterize the peptide, and signature losses were identified by observation.

Results and Discussion

Bioinformatics and proposed mannosylation mechanism

Sequence similarity searches predict that Ram29 is an integral membrane protein, from the GT-C superfamily of glycosyltransferases, which is likely to be similar to other predicted mannosyltransferases including MppH/I (Magarvey et al., 2006), Tcp15 (Sosio et al., 2004) and Dbv20 (Sosio et al., 2003). Ram29 was further analysed using eight bioinformatics programs designed to predict the topology of membrane proteins. Depending on the algorithm used, Ram29 is predicted to possess between 10 and 14 transmembrane segments (TMS) spanning approximately 450 amino acids within the N-terminal region of the protein. Although the bioinformatics algorithms did not converge on the precise number of TMS, all of these programmes predict that the C-terminal domain, comprising approximately 150 amino acids, is extracytoplasmic. We suggest that this putative extracytoplasmic C-terminal domain, which shows little or no similarity with any other known protein sequences, is responsible for binding the ramoplanin aglycone (Figs. 2 and S2). This is consistent with earlier findings (Borghi et al., 1991) showing that the demannosylated teicoplanin aglycone can be remannosylated by incubating with the teicoplanin producing strain, possessing the putative mannosyltransferase Tcp15, demonstrating that mannosylation occurs on the outside of the cell membrane.

Fig. 2. Proposed two-component system for mannosylation of the glycopeptide antibiotics. PPM synthase (Ppm1) is located in the membrane and catalyses the synthesis of PPMs on the cytoplasmic membrane surface, before being translocated to the extracytoplasmic membrane surface. Subsequently, the membrane protein mannosyltransferase Ram29 transfers the mannose moiety from PPM to the antibiotic aglycone, which has been transported from the cytoplasm.

Fig. 2.

Other members of GT-C superfamily possess similar TMS, with acidic (DxD or DDx) motifs in either the first or second extracytoplasmic loop (Berg et al., 2005; Liu & Mushegian, 2003). The DxD motif is suggested to bind to, and activate, polyprenyl phosphosugar donors (Berg et al., 2007), catalysing transfer of a range of different sugars, including glucose (Glu) (Song et al., 2009), galactosamine (GalN) (Skovierová et al., 2010), arabinofuranose (Araf) (Berg et al., 2005) and rhamnose (Rha) (Birch et al., 2009), during cell wall oligosaccharide biosynthesis. Sequence alignments of Ram29 led to the identification of a putative DxD motif in the second predicted extracytoplasmic loop, containing acidic residues (E162, D163 & D175) which are conserved in other putative mannosyltransferases (Tcp15 and Dbv20) (Fig. S3). Taken together, we suggest that Ram29 would use its unique C-terminal domain to recognize and bind the ramoplanin aglycone outside of the cell membrane, and the extracytoplasmic loop containing the DxD motif would recognize and transfer the mannosyl group from PPMs, which are known to be biosynthesized from GDP-mannose and polyprenyl phosphates, of varying chain length, by the PPM synthases (Ppm1) (Wehmeier et al., 2009) to form mannosylated antibiotics (Fig. 2).

Optimizing expression of ram29 in the enduracidin producer S. fungicidicus

The in vitro characterization of the putative mannosyltransferase Ram29, like other members of the GT-C superfamily enzymes, is complicated by difficulties associated with the overproduction, isolation and purification of membrane proteins as well as the lack of available PPM donor substrates with the required polyprenyl chain length. In light of this, we chose to establish the heterologous expression system for ram29 in the enduracidin producer S. fungicidicus. Given the structural similarity of enduracidin to the ramoplanin aglycone, we predicted that the endogenous enduracidin and PPM might function as substrates for Ram29 leading to mannosylated enduracidins. In addition to confirming the function of Ram29, this approach could potentially provide new mannosylated enduracidin antibiotics with improved properties, particularly those with enhanced aqueous solubility.

Accordingly, an expression cassette with the putative mannosyltransferase encoding gene ram29, including its native upstream Shine–Dalgarno (SD) sequence, was cloned into the high copy number plasmid pIJ86 under the control of the strong constitutive promoter ermE* (Bibb et al., 1985). The LC-MS analysis of cellular extracts and culture supernatants derived from S. fungicidicus pIJ86-ram29 exconjugants failed to lead to the detection of any mannosylated enduracidins. Given that Ram29 is predicted to be an integral membrane protein, the higher levels of expression afforded by pIJ86 could result in production of insoluble, non-functional Ram29. Consequently ram29 was placed under the control of a tetracycline inducible promoter by cloning the expression cassette into the integrative plasmid pMS17 (Rodríguez-García et al., 2005; Wehmeier et al., 2009). However, even with inducible control over the expression of the ram29 gene integrated onto the chromosome of S. fungicidicus, we still failed to detect any mannosylated enduracidin, which we rationalized may be due to a suboptimal sd sequence and GTG start codon of ram29. We therefore replaced the 5′-leader sequence of ram29, including the native sd sequence and start codon, with the corresponding sequence of an eGFP expression construct from the plasmid pIJ8668 (kindly provided by Professor M. Bibb, John Innes Centre), which gives robust expression of eGFP (Fig. S4). The resulting S. fungicidicus pMS17-ram29 exconjugants were shown to produce mannosylated enduracidins (vide infra), demonstrating the importance of the sd sequence and the distance between the sd sequence and the alternative ATG start codon for establishing functional expression of ram29.

Isolation of mannosylated enduracidins from S. fungicidicus pMS17-ram29

The optimized S. fungicidicus pMS17-ram29 was grown in liquid media for 2 weeks, the enduracidin isolated using a butanol extraction procedure and further purified by RP-HPLC (Fig. 3). Fractions containing enduracidin were identified by MALDI-TOF (Fig. S5) and further analysed by LC-MS (Fig. 4). It was revealed that the strain S. fungicidicus pMS17-ram29 produced two additional new minor products differing in mass by +162 Da to enduracidins A and B, which were also evident. The +162 Da mass difference is consistent with the addition of a single mannose group to enduracidins A and B. The two new putative mannosylated enduracidin A and B derivatives were absent from extracts derived from WT S. fungicidicus fermentations.

Fig. 3. RP-HPLC analysis of the S. fungicidicus crude extract. (a) Crude extract from the fermentation broth of the S. fungicidicus pMS17-ram29 transformant. (b) Enduracidin A and B standards.

Fig. 3.

Fig. 4. ESI (electrospray ionisation)-LC-MS analysis of the S. fungicidicus pMS17-ram29 fermentation products. The spectra show the triply charged [M+3H]3+ parent ions for WT enduracidins A (m/z: observed 785.5; expected 785.3) and B (m/z: observed 790.2; expected 790.0) and monomannosylated enduracidins A (m/z: observed 839.6, expected 839.3) and B (m/z: observed 844.3, expected 844.0). Doubly charged [M+2H]2+ molecular ions were also observed by ESI-LC-MS (Fig. S5), and the singly charged [M+H]+ ions were observed by MALDI-TOF (Fig. S6), consistent with the proposed structures.

Fig. 4.

The position of the putative mannose group was determined by tandem MS. The de novo sequencing of cyclic nonribosomal peptides can be non-trivial, as each linear fragment is the product of at least two fragmentation events, and the subset of possible fragments is dependent on the position of the first event. By running a series of experiments on both enduracidin A and the new glycosylated analogue in parallel, it was possible to deduce a y-ion series for both molecules under identical conditions. In the spectra, a prominent y-ion series can be observed for the WT enduracidin A peptide fragments: End15–Hpg17 (y 3, m/z 393.2), Gly14–Hpg17 (y 4, m/z 450.2), Ser12–Hpg17 (y 5, m/z 754.3), Hpg11–Hpg17 (y 7, m/z 903.3), End10–Hpg17 (y 8, m/z 1057.4), Cit9–Hpg17 (y 9, m/z 1214.4), Thr8–Hpg17 (y 10, m/z 1315.5) and Hpg7–Hpg17 (y 11, m/z 1464.5) (Fig. 5). In contrast, analysis of the monomannosylated enduracidin peptide revealed new fragment ions with m/z corresponding to: man–Hpg11–Hpg17 (y 7, m/z 1065.3), End10–Hpg17 (y 8, m/z 1219.4), Thr8–Hpg17 (y 10, m/z 1477.5) (Fig. 5). Comparison of the key observed ions y 7, y 8, and y 10 shows a mass difference of +162 Da between the WT and the mannosylated sample, corresponding to the addition of a single mannose group in the latter. Additional ions corresponding to non-hydrated fragment ions, and doubly charged fragments ions (both hydrated and non-hydrated) are described in the supplementary information (Tables S2–S5, Figs S7 and S8). The MS-MS spectra provide strong evidence as to the regioselectivity of Ram29, with the mannosylation occurring on the Hpg11 residue of enduracidin: the +162 Da mass difference was carried throughout the mannosylated peptide sample from y 10 to y 7; however once Hpg11 was lost, thus generating the y 6 ion, this and all subsequent fragment ions were identical with those of the WT enduracidin (Fig. 5). Thus, Ram29 appears to mannosylate enduracidin at the same position as on ramoplanin. Moreover, Ram29 is evidently promiscuous and can mannosylate alternative lipopeptides to ramoplanin, such as the enduracidins A and B.

Fig. 5. Tandem MS of WT and monomannosylated enduracidin A. (a) Localization of the mannosyl group on monomannosylated enduracidin A by MS-MS characterization. The y ions highlighted in red denote a mass difference of +162 Da from WT to mannosylated. The key y 7 ion is the last fragment to carry the +162 mannosyl group mass. (b) Structure of the monomannosylated enduracidin A analogue, based on MS-MS data. Additional MS-MS data (Tables S2–S5) and spectra (Figs S7 and S8) supporting the structure of monomannosylated enduracidin A can be found within the supplementary information.

Fig. 5.

The deletion of ram29 from the ramoplanin producer, Actinoplanes sp. ATCC 33076 (Chen et al., 2013), was shown to produce the ramoplanin aglycone which was consistent with the findings presented here. Given that Ram29 is the only apparent glycosyltransferase in the ramoplanin biosynthetic gene cluster, this suggests that Ram29 may act iteratively to introduce both the mannosyl groups of ramoplanin (Chen et al., 2013). Interestingly, we show that expression of Ram29 in S. fungicidicus results in only monomannosylated enduracidins. Thus whilst Ram29 clearly recognizes the phenol group of Hpg11 in both enduracidin and the ramoplanin aglycone as its acceptor substrate, it is possible that a second mannosyltransferase, which is not encoded from within the ramoplanin biosynthetic gene cluster, may be required to introduce a second distinct α1,2-dimannosyl glycosidic linkage. Alternatively, some Streptomyces species have been reported to have membrane-associated α-mannosidases which remove mannosyl groups from antibiotics. For example, the mycelia of Streptomyces GE 91081 can remove one mannose unit from ramoplanin (Di Palo et al., 2007), whilst the mycelia of Nocardia orientalis NRRL 2450 and Streptomyces candidus NRRL 3218 have been found to remove a mannose unit from teicoplanin (Borghi et al., 1991). S. fungicidicus contains a putative mannosidase encoding gene (orf13) within the enduracidin gene cluster (Yin & Zabriskie, 2006), which may function to remove a mannose group from any dimannosylated enduracidin that may be produced. Mannosidase activity could also potentially account for the low yields of mannosylated enduracidins produced by the S. fungicidicus pMS17-ram29 strain.

Conclusion

In summary, we have developed protocols for the functional expression of the mannosyltransferase Ram29 in the enduracidin producer S. fungicidicus and isolated novel monomannosylated enduracidin variants. Using tandem MS, we have shown that mannosylation occurs at Hpg11 of enduracidin, which supports the function of Ram29 in the biosynthesis of ramoplanin. Previously there have been considerable advances in the development of natural product glycodiversification strategies, which have largely focused on glycosylation pathways that involve the more common Leloir-type (sugar nucleotide-dependent) glycosyltransferases. The findings presented here provide the first example of how the less common, membrane associated PPM-dependent glycosyltransferases can also be used in glycodiversification of natural product scaffolds; this could be valuable in future efforts to develop bioactive natural products, including improved enduracidin variants with enhanced aqueous solubility.

Acknowledgements

This work was supported by BBSRC (BB/L002299/1 & BB/H015914/1). We thank Professor Maggie Smith (University of York) for providing the plasmid pMS17, and Professor Mervyn Bibb (John Innes Centre) for the plasmid pIJ8668.

Abbreviations:

Man

mannose

PPM

polyprenyl phosphomannose

Footnotes

Edited by: I. A. O'Neill

References

  1. Amir-Heidari B., Thirlway J., Micklefield J. (2008). Auxotrophic-precursor directed biosynthesis of nonribosomal lipopeptides with modified tryptophan residues Org Biomol Chem 6 975–978 10.1039/b718766c . [DOI] [PubMed] [Google Scholar]
  2. Berg S., Starbuck J., Torrelles J.B., Vissa V.D., Crick D.C., Chatterjee D., Brennan P.J. (2005). Roles of conserved proline and glycosyltransferase motifs of EmbC in biosynthesis of lipoarabinomannan J Biol Chem 280 5651–5663 10.1074/jbc.M411418200 . [DOI] [PubMed] [Google Scholar]
  3. Berg S., Kaur D., Jackson M., Brennan P.J. (2007). The glycosyltransferases of Mycobacterium tuberculosis – roles in the synthesis of arabinogalactan, lipoarabinomannan, and other glycoconjugates Glycobiology 17 35R–56R 10.1093/glycob/cwm010 . [DOI] [PubMed] [Google Scholar]
  4. Bernsel A., Viklund H., Falk J., Lindahl E., von Heijne G., Elofsson A. (2008). Prediction of membrane-protein topology from first principles Proc Natl Acad Sci U S A 105 7177–7181 10.1073/pnas.0711151105 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bibb M.J., Janssen G.R., Ward J.M. (1985). Cloning and analysis of the promoter region of the erythromycin resistance gene (ermE) of Streptomyces erythraeus Gene 38 215–226 10.1016/0378-1119(85)90220-3 . [DOI] [PubMed] [Google Scholar]
  6. Birch H.L., Alderwick L.J., Rittmann D., Krumbach K., Etterich H., Grzegorzewicz A., McNeil M.R., Eggeling L., Besra G.S. (2009). Identification of a terminal rhamnopyranosyltransferase (RptA) involved in Corynebacterium glutamicum cell wall biosynthesis J Bacteriol 191 4879–4887 10.1128/JB.00296-09 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Borghi A., Ferrari P., Gallo G.G., Zanol M., Zerilli L.F., Lancini G.C. (1991). Microbial de-mannosylation and mannosylation of teicoplanin derivatives J Antibiot (Tokyo) 44 1444–1451 10.7164/antibiotics.44.1444 . [DOI] [PubMed] [Google Scholar]
  8. Castiglione F., Marazzi A., Meli M., Colombo G. (2005). Structure elucidation and 3D solution conformation of the antibiotic enduracidin determined by NMR spectroscopy and molecular dynamics Magn Reson Chem 43 603–610 10.1002/mrc.1606 . [DOI] [PubMed] [Google Scholar]
  9. Chen J.-S., Wang Y.-X., Shao L., Pan H.-X., Li J.-A., Lin H.-M., Dong X.-J., Chen D.-J. (2013). Functional identification of the gene encoding the enzyme involved in mannosylation in ramoplanin biosynthesis in Actinoplanes sp Biotechnol Lett 35 1501–1508 10.1007/s10529-013-1233-3 . [DOI] [PubMed] [Google Scholar]
  10. Cudic P., Behenna D.C., Kranz J.K., Kruger R.G., Wand A.J., Veklich Y.I., Weisel J.W., McCafferty D.G. (2002). Functional analysis of the lipoglycodepsipeptide antibiotic ramoplanin Chem Biol 9 897–906 10.1016/S1074-5521(02)00191-6 . [DOI] [PubMed] [Google Scholar]
  11. Di Palo S., Gandolfi R., Jovetic S., Marinelli F., Romano D., Molinari F. (2007). A new bacterial mannosidase for the selective modification of ramoplanin and its derivatives Enzyme Microb Technol 41 806–811 10.1016/j.enzmictec.2007.07.013. [DOI] [Google Scholar]
  12. Fang X., Tiyanont K., Zhang Y., Wanner J., Boger D., Walker S. (2006). The mechanism of action of ramoplanin and enduracidin Mol Biosyst 2 69–76 10.1039/B515328J . [DOI] [PubMed] [Google Scholar]
  13. Hamburger J.B., Hoertz A.J., Lee A., Senturia R.J., McCafferty D.G., Loll P.J. (2009). A crystal structure of a dimer of the antibiotic ramoplanin illustrates membrane positioning and a potential Lipid II docking interface Proc Natl Acad Sci U S A 106 13759–13764 10.1073/pnas.0904686106 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. He H. (2005). Mannopeptimycins, a novel class of glycopeptide antibiotics active against Gram-positive bacteria Appl Microbiol Biotechnol 67 444–452 10.1007/s00253-004-1884-z . [DOI] [PubMed] [Google Scholar]
  15. Hobbs G., Frazer C.M., Gardner D.C.J., Cullum J.A., Oliver S.G. (1989). Dispersed growth of Streptomyces in liquid culture Appl Microbiol Biotechnol 31 272–277. [Google Scholar]
  16. Hofmann K., Stoffel W. (1993). tmbase – A database of membrane spanning protein segments Biol Chem Hoppe Seyler 374 166. [Google Scholar]
  17. Kieser T., Bibb M.J., Buttner M.J., Chater K.F., Hopwood D.A. (2000). Practical Streptomyces Genetics., Norwich: The John Innes Foundation. [Google Scholar]
  18. Lewis R.A., Nunns L., Thirlway J., Carroll K., Smith C.P., Micklefield J. (2011). Active site modification of the β-ketoacyl-ACP synthase FabF3 of Streptomyces coelicolor affects the fatty acid chain length of the CDA lipopeptides Chem Commun (Camb) 47 1860–1862 10.1039/C0CC03444D . [DOI] [PubMed] [Google Scholar]
  19. Liu J., Mushegian A. (2003). Three monophyletic superfamilies account for the majority of the known glycosyltransferases Protein Sci 12 1418–1431 10.1110/ps.0302103 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. MacNeil D.J., Gewain K.M., Ruby C.L., Dezeny G., Gibbons P.H., MacNeil T. (1992). Analysis of Streptomyces avermitilis genes required for avermectin biosynthesis utilizing a novel integration vector Gene 111 61–68 10.1016/0378-1119(92)90603-M . [DOI] [PubMed] [Google Scholar]
  21. Magarvey N.A., Haltli B., He M., Greenstein M., Hucul J.A. (2006). Biosynthetic pathway for mannopeptimycins, lipoglycopeptide antibiotics active against drug-resistant Gram-positive pathogens Antimicrob Agents Chemother 50 2167–2177 10.1128/AAC.01545-05 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Mahlert C., Kopp F., Thirlway J., Micklefield J., Marahiel M.A. (2007). Stereospecific enzymatic transformation of α-ketoglutarate to (2S,3R)-3-methyl glutamate during acidic lipopeptide biosynthesis J Am Chem Soc 129 12011–12018 10.1021/ja074427i . [DOI] [PubMed] [Google Scholar]
  23. McCafferty D.G., Cudic P., Frankel B.A., Barkallah S., Kruger R.G., Li W. (2002). Chemistry and biology of the ramoplanin family of peptide antibiotics Biopolymers 66 261–284 10.1002/bip.10296 . [DOI] [PubMed] [Google Scholar]
  24. Milne C., Powell A., Jim J., Al Nakeeb M., Smith C.P., Micklefield J. (2006). Biosynthesis of the (2S,3R)-3-methyl glutamate residue of nonribosomal lipopeptides J Am Chem Soc 128 11250–11259 10.1021/ja062960c . [DOI] [PubMed] [Google Scholar]
  25. Möller S., Croning M.D., Apweiler R. (2001). Evaluation of methods for the prediction of membrane spanning regions Bioinformatics 17 646–653 10.1093/bioinformatics/17.7.646 . [DOI] [PubMed] [Google Scholar]
  26. Neary J.M., Powell A., Gordon L., Milne C., Flett F., Wilkinson B., Smith C.P., Micklefield J. (2007). An asparagine oxygenase (AsnO) and a 3-hydroxyasparaginyl phosphotransferase (HasP) are involved in the biosynthesis of calcium-dependent lipopeptide antibiotics Microbiology 153 768–776 10.1099/mic.0.2006/002725-0 . [DOI] [PubMed] [Google Scholar]
  27. Nogami I., Shirafuji H., Matsumura S. (1984). Production of enduracidin and micro-organisms therefor US Patent 4465771.
  28. Powell A., Al Nakeeb M., Wilkinson B., Micklefield J. (2007a). Precursor-directed biosynthesis of nonribosomal lipopeptides with modified glutamate residues Chem Commun (Camb) (26), 2683–2685 10.1039/b706224a . [DOI] [PubMed] [Google Scholar]
  29. Powell A., Borg M., Amir-Heidari B., Neary J.M., Thirlway J., Wilkinson B., Smith C.P., Micklefield J. (2007b). Engineered biosynthesis of nonribosomal lipopeptides with modified fatty acid side chains J Am Chem Soc 129 15182–15191 10.1021/ja074331o . [DOI] [PubMed] [Google Scholar]
  30. Rodríguez-García A., Combes P., Pérez-Redondo R., Smith M.C.A., Smith M.C.M. (2005). Natural and synthetic tetracycline-inducible promoters for use in the antibiotic-producing bacteria Streptomyces Nucleic Acids Res 33 e87 10.1093/nar/gni086 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Shirling E.B., Gottlieb D. (1966). Methods for characterization of Streptomyces species Int J Syst Bacteriol 16 313–340 10.1099/00207713-16-3-313. [DOI] [Google Scholar]
  32. Skovierová H., Larrouy-Maumus G., Pham H., Belanová M., Barilone N., Dasgupta A., Mikusová K., Gicquel B., Gilleron M., other authors (2010). Biosynthetic origin of the galactosamine substituent of arabinogalactan in Mycobacterium tuberculosis J Biol Chem 285 41348–41355 10.1074/jbc.M110.188110 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Song F., Guan Z., Raetz C.R.H. (2009). Biosynthesis of undecaprenyl phosphate-galactosamine and undecaprenyl phosphate-glucose in Francisella novicida Biochemistry 48 1173–1182 10.1021/bi802212t . [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Sosio M., Stinchi S., Beltrametti F., Lazzarini A., Donadio S. (2003). The gene cluster for the biosynthesis of the glycopeptide antibiotic A40926 by Nonomuraea species Chem Biol 10 541–549 10.1016/S1074-5521(03)00120-0 . [DOI] [PubMed] [Google Scholar]
  35. Sosio M., Kloosterman H., Bianchi A., de Vreugd P., Dijkhuizen L., Donadio S. (2004). Organization of the teicoplanin gene cluster in Actinoplanes teichomyceticus Microbiology 150 95–102 10.1099/mic.0.26507-0 . [DOI] [PubMed] [Google Scholar]
  36. Thirlway J., Lewis R., Nunns L., Al Nakeeb M., Styles M., Struck A.W., Smith C.P., Micklefield J. (2012). Introduction of a non-natural amino acid into a nonribosomal peptide antibiotic by modification of adenylation domain specificity Angew Chem Int Ed Engl 51 7181–7184 10.1002/anie.201202043 . [DOI] [PubMed] [Google Scholar]
  37. Tusnády G.E., Simon I. (2001). The hmmtop transmembrane topology prediction server Bioinformatics 17 849–850 10.1093/bioinformatics/17.9.849 . [DOI] [PubMed] [Google Scholar]
  38. Uguru G.C., Milne C., Borg M., Flett F., Smith C.P., Micklefield J. (2004). Active-site modifications of adenylation domains lead to hydrolysis of upstream nonribosomal peptidyl thioester intermediates J Am Chem Soc 126 5032–5033 10.1021/ja048778y . [DOI] [PubMed] [Google Scholar]
  39. Viklund H., Elofsson A. (2004). Best α-helical transmembrane protein topology predictions are achieved using hidden Markov models and evolutionary information Protein Sci 13 1908–1917 10.1110/ps.04625404 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Viklund H., Elofsson A. (2008). octopus: improving topology prediction by two-track ANN-based preference scores and an extended topological grammar Bioinformatics 24 1662–1668 10.1093/bioinformatics/btn221 . [DOI] [PubMed] [Google Scholar]
  41. Walker S., Chen L., Hu Y., Rew Y., Shin D., Boger D.L. (2005). Chemistry and biology of ramoplanin: a lipoglycodepsipeptide with potent antibiotic activity Chem Rev 105 449–476 10.1021/cr030106n . [DOI] [PubMed] [Google Scholar]
  42. Wehmeier S., Varghese A.S., Gurcha S.S., Tissot B., Panico M., Hitchen P., Morris H.R., Besra G.S., Dell A., Smith M.C.M. (2009). Glycosylation of the phosphate binding protein, PstS, in Streptomyces coelicolor by a pathway that resembles protein O-mannosylation in eukaryotes Mol Microbiol 71 421–433 10.1111/j.1365-2958.2008.06536.x . [DOI] [PubMed] [Google Scholar]
  43. Wilkinson B., Micklefield J. (2007). Mining and engineering natural-product biosynthetic pathways Nat Chem Biol 3 379–386 10.1038/nchembio.2007.7 . [DOI] [PubMed] [Google Scholar]
  44. Wu M.C., Law B., Wilkinson B., Micklefield J. (2012). Bioengineering natural product biosynthetic pathways for therapeutic applications Curr Opin Biotechnol 23 931–940 10.1016/j.copbio.2012.03.008 . [DOI] [PubMed] [Google Scholar]
  45. Yin X., Zabriskie T.M. (2006). The enduracidin biosynthetic gene cluster from Streptomyces fungicidicus Microbiology 152 2969–2983 10.1099/mic.0.29043-0 . [DOI] [PubMed] [Google Scholar]

Articles from Microbiology are provided here courtesy of Microbiology Society

RESOURCES