How To Make a Glycopeptide: A Synthetic Biology ... - ACS Publications

Aug 5, 2016 - How To Make a Glycopeptide: A Synthetic Biology Approach To. Expand Antibiotic Chemical Diversity. Grace Yim, Wenliang Wang, Maulik N. T...
0 downloads 4 Views 3MB Size
Article pubs.acs.org/journal/aidcbc

How To Make a Glycopeptide: A Synthetic Biology Approach To Expand Antibiotic Chemical Diversity Grace Yim, Wenliang Wang, Maulik N. Thaker,† Stephanie Tan, and Gerard D. Wright* M. G. DeGroote Institute for Infectious Disease Research, Department of Biochemistry and Biomedical Sciences, McMaster University, Hamilton, Ontario L8S 4K1, Canada S Supporting Information *

ABSTRACT: Modification of natural product backbones is a proven strategy for the development of clinically useful antibiotics. Such modifications have traditionally been achieved through medicinal chemistry strategies or via in vitro enzymatic activities. In an orthogonal approach, engineering of biosynthetic pathways using synthetic biology techniques can generate chemical diversity. Here we report the use of a minimal teicoplanin class glycopeptide antibiotic (GPA) scaffold expressed in a production-optimized Streptomyces coelicolor strain to expand GPA chemical diversity. Thirteen scaffold-modifying enzymes from 7 GPA biosynthetic gene clusters in different combinations were introduced into S. coelicolor, enabling us to explore the criteria for in-cell GPA modification. These include identifying specific isozymes that tolerate the unnatural GPA scaffold and modifications that prevent or allow further elaboration by other enzymes. Overall, 15 molecules were detected, 9 of which have not been reported previously. Some of these compounds showed activity against GPA-resistant bacteria. This system allows us to observe the complex interplay between substrates and both non-native and native tailoring enzymes in a cell-based system and establishes rules for GPA synthetic biology and subsequent expansion of GPA chemical diversity. KEYWORDS: glycopeptide, antibiotic, heterologous expression, synthetic biosynthesis, tailoring enzymes

G

genes within biosynthetic clusters encodes tailoring enzymes (also referred to as scaffold-modifying enzymes). Tailoring enzymes are responsible for decorating the heptapeptide backbone by chlorination, glycosylation, acylation, methylation, and sulfation. Exemplar enzymes have been well-studied in vitro2−5 and in the native host,6−9 and more recently sulfotransferases have been studied in a heterologous host.10 Due to their medical importance and the dramatic decrease in nucleotide-sequencing costs, the biosynthetic clusters of over a dozen GPAs have been described.11 With the increase in antimicrobial resistance,12 the need for new antibiotics has never been so urgent. A survey of the plethora of clinically employed β-lactam and macrolide class antibiotics demonstrates that chemical derivatization of existing scaffolds to expand the arsenal of available antibiotics is a proven strategy for generating new pharmaceuticals. For example, compared to its parent molecule vancomycin, telavancin has enhanced potency against strains with reduced vancomycin susceptibility, presumably by creating increased membrane depolarization and permeability.13 However, due to their elegantly complex structures, natural product derivatization can sometimes be synthetically challenging. On the other

lycopeptide antibiotics (GPAs) are important natural products used for the treatment of infections caused by Gram-positive pathogens. Between 2000 and 2010, global GPA consumption rose 233%, reflecting the increased prevalence of antibiotic-resistant bacteria such as methicillin-resistant Staphylococcus aureus (MRSA) for which GPAs are especially useful.1 Five GPAs are currently available for therapeutic use: vancomycin, teicoplanin, dalbavancin, telavancin, and oritavancin. The core heptapeptide scaffolds of these drugs differ in their N-terminal amino acids. Vancomycin, oritavancin, and telavancin contain aliphatic amino acids at their N-termini. Teicoplanin and dalbavancin are exclusively composed of aromatic amino acids and possess an additional biaryl ether ring. Vancomycin and teicoplanin are natural products isolated from the soil bacteria Amycolatopsis orientalis and Actinoplanes teichomyceticus, respectively. Dalbavancin, oritavancin, and telavancin are semisynthetic antibiotics resulting from chemical derivatization of the natural product GPA scaffolds A40926 (teicoplanin-like scaffold), chloroeremomycin (vancomycin-like scaffold), and vancomycin, respectively. GPAs are heptapeptides synthesized by large multifunctional nonribosomal peptide synthetases (NRPS). The biosynthesis of GPAs is encoded by large gene clusters ranging from ∼70 to 100 kb that include NRPS genes and a host of other genes encoding a wide variety of functions. One pertinent group of © 2016 American Chemical Society

Received: June 13, 2016 Published: August 5, 2016 642

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

Figure 1. Chemical structures of scaffolds and successfully produced target compounds from one and two gene integrations produced in the S. coelicolor platform.

because it is a genetically tractable organism, which is similar to our native producer, and is more likely to contain biosynthetic machinery and precursors required for biosynthesis that may not be encoded in the cluster when compared to other heterologous hosts such as Escherichia coli. The biosynthetic machinery for two versions of a simplified teicoplanin type glycopeptide biosynthetic cluster A47934 (compound 1, Figure 1) and desulfo-A47934 (2, Figure 1), modified with 13 tailoring genes from seven different GPA producer organisms, served as devices and parts. The 68 kb biosynthetic cluster and tailoring genes integrate intragenically into the chromosome at the phiC31 (SCO3798) and phiBT1 (SCO4848) attB sites of S. coelicolor, respectively.17 Additional genes can be added using the SV1 attB integration site located in SCO438318 or multiple pseudo-attB sites present at various locations in the genome offering opportunity to further expand biosynthetic potential.19 Platform Construction. The biosynthetic cluster of 1 has a teicoplanin type scaffold GPA, but is relatively pared down as it lacks glycosylation and acylation. The 68 kb gene cluster was originally identified in the producer strain Streptomyces toyocaensis.20 A library of S. toyocaensis genomic DNA was constructed in the P1-derived artificial chromosome (PAC) vector pESAC13. The PAC conjugative vector, pESAC13,21 has an origin of replication for propagation in E. coli and carries machinery for chromosomal integration into the phiC31 attB site. PAC vectors can accommodate >100 kb size inserts providing a strategic advantage over the limited payload capacity of cosmids. The library was screened by PCR to locate the A47934 cluster. A clone was identified with a 140 kb

hand, natural product modification enzymes have evolved to interact with and amplify biological complexity. Although bioinformatic analysis of biosynthetic clusters has made great strides,14 it is still difficult to predict the substrate promiscuity of any given enzyme without detailed biochemical study. This knowledge gap can be addressed by using a large number of diverse enzymes in combination with a given scaffold and through systematic testing to understand the order of scaffold modification, in an attempt to increase the chances of substrate modification. GPAs present a model opportunity for such a strategy because a large number of biosynthetic clusters and their associated tailoring genes have been characterized. Here we have heterologously expressed two versions of a minimal GPA scaffold in a production-optimized Streptomyces coelicolor host and have introduced tailoring genes from GPA clusters from different producer organisms to expand the chemical diversity of this medically important class of antibiotic and to explore the rules of their biosynthetic derivatization.



RESULTS AND DISCUSSION The concepts of synthetic biology applied to natural product biosynthesis provide great opportunities for rapidly expanding the chemical diversity of natural products and for increasing the ease by which biosynthetic clusters are manipulated and expressed. We have previously described the potential application of synthetic biology to natural product biosynthesis using GPAs as an example.15 In this study, we have applied those concepts. We chose a production-optimized S. coelicolor strain, M1146,16 as our chassis for heterologous expression 643

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

fragment encoding the complete biosynthetic cluster lacking only ∼2 kb at the 3′ end that encodes the vanHAX selfresistance genes. These were subsequently added using lambda red-mediated recombination to create pA47934 encoding all necessary genes for biosynthesis, resistance, and export. The sulfotransferase, staL, was deleted from pA47934, to create a minimized GPA scaffold encoding the biosynthesis of 1 but lacking a sulfate group, compound 2 (Figure 1). PAC vectors were mobilized from E. coli into S. coelicolor M1146 by conjugation. M1146 is a strain with a minimized secondary metabolite profile because four endogenous biosynthetic clusters have been deleted.16 Chromosomal integration of the A97934 cluster at the attB site was confirmed by PCR amplification of regions at the ends and middle of the cluster. Genes encoding enzymes for scaffold modification from other GPA clusters were then integrated into the two scaffoldproducing strains using the vector pIJ10257,22 an E. coli− Streptomyces conjugative shuttle vector that integrates into a distinct region of the Streptomyces chromosome, the phiBT1 attB site. Transcription of genes cloned into the multiplecloning site of pIJ1025722 is driven by the constitutive promoter ermEp*. Heterologous Production of Scaffolds: A47934 and DS-A47934. Heterologous biosynthesis of both 1 and 2 was observed in S. coelicolor M1146, confirming that both repair of the A47934 biosynthetic cluster to include self-resistance genes and deletion of staL were successful (Figure 2). HPLC analysis

Figure 3. RP-HPLC chromatograms of fermentations after affinity chromatography of 2a−2g produced in the S. coelicolor/pA47934 ΔstaL background showing the relative conversion from 2 to the modified derivatives. After D-Ala-D-Ala affinity chromatography and before semipreparative HPLC, partially purified fermentations of 2a (a), 2d (c), 2e (d), 2f (e), and 2g (f)were separated on identical linear gradients. After affinity chromatography and before semipreparative HPLC, partially purified fermentations of 2b (black trace) and 2c (gray trace) were separated by isocratic elution (b). All peaks had UV profiles similar to that of 2 (Figure 2c).

monochlorinated A40926 but with subsequent modifications, suggesting that halogenation is not required for the subsequent action of the other A40926 tailoring enzymes (glycosyl- and acyltransferases).23 The production of many precursors in this platform likely reflects an evolved optimization of coordinated expression between regulators present within the biosynthetic cluster and global regulators present outside the biosynthetic cluster in the native producer to produce one major compound that is absent in the heterologous host. Alternatively, the increased presence of GPA precursors during heterologous expression may also be the result of a promiscuous exporter present in the heterologous producer. This lack of optimization or increased exporter promiscuity was beneficial in our system, allowing increased production and export of a larger variety of GPA molecules. Production of DS-A47934 Derivatives. To begin diversification of our two GPA scaffolds (2 and 1), 10 tailoring genes were separately cloned from six different glycopeptide producers into pIJ10257. Structures of the different GPAs made by the various native producers are shown in Figure S2. Tailoring enzymes chosen included glycosyltransferases transferring either glucose, mannose, or N-acetyl-glucosamine moieties, sulfotransferases, and C-, N-, and O-methyltransferases. Upon introduction of the 10 tailoring genes into production background 2, 7 of the 10 strains produced the predicted glycopeptide with various degrees of conversion (Figure 3; Table 1). A summary of all genes tested in the platform and whether their integration resulted in a modification of 2 is depicted in Figure 4a. Six of the seven compounds are new GPAs, not previously reported. Compound structures were confirmed by high-resolution mass spectrometry (HRMS, Table S1), 1D and 2D NMR spectra (Figure S3−29; Table S2), and MS/MS (Figure S30). In some fermentations, several prominent peaks appeared in addition to the expected product peak (Figure 3), which were identified as tailored derivatives lacking one to three chlorines (Figure S31).

Figure 2. RP-HPLC chromatograms of fermentations after affinity chromatography of heterologous and native production: production of 1 (a) from S. coelicolor/pA47934 (solid line) and S. toyocaensis (dashed line); production of 2 (b) from S. coelicolor/pA47934 ΔstaL (solid line) and S. toyocaensis ΔstaL (dashed line); production from S. coelicolor M1146 lacking a PAC vector showing no visible peaks (a, b; dark gray line). (c) Ultraviolet spectrum of compound 2.

for the purposes of determining yield and turnover was conducted on fractions from a D-Ala-D-Ala affinity column (mimics the target of GPAs) before final purification by semipreparative HPLC. Compounds eluted from the affinity column when analyzed by HPLC and LC-MS; all peaks had UV profiles similar to that of compound 2 (Figure 2c). Thus, peaks shown in Figures 2, 3 and 5 likely represent all of the GPA-like compounds produced in the given fermentation. Production of 1 and 2 has been previously observed in the native producer S. toyocaensis and a ΔstaL mutant.7 When heterologous productions of 1 and 2 were compared between the native producer and M1146, additional products with earlier retention times than those of the target compounds were observed in M1146 (Figure 2a,b). Upon examination of these fermentations by LC-MS, most of the additional masses were consistent with aglycone precursors lacking one or more chlorines (Figure S1). Previous work with mutants of Nonomuraea sp. ATCC 39727, producer of the GPA A40926 (Figure S2), showed production of dechlorinated and 644

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

identical and have 76 and 72% protein identity, respectively, but only one of each pair was effective in our system. Although modifying enzymes such as glycosyltransferases are known to have substrate promiscuity in terms of the sugar substrates they utilize,2 substrate promiscuity is clearly not a universal property of modifying enzymes. Other studies have shown that glycosyltransferase substrate promiscuity for GPAs could not be predicted by the similarity between native substrates. For example, GtfE from the vancomycin cluster and GtfB from the chloroeremomycin cluster both glucosylate their native scaffolds at amino acid 4. However, despite the similarity of their native substrates, GtfEvan can glucosylate both teicoplanin and vancomycin aglycone very efficiently, but GtfB poorly glucosylates teicoplanin aglycone.24 These studies underscore the diversity and unpredictability of substrate promiscuity in functionally homologous enzymes and point to the need for systematic sampling of biosynthetic genes in synthetic biology strategies to expand chemical diversity. Production of A47934 Derivatives. When tailoring genes were introduced into the producer of 1, a slightly larger substrate than 2 containing an additional sulfate group (Figure 2a,b), fermentation products were more complex. We had predicted the production of one dominant molecule, the tailored derivative of 1. Instead, we observed differing relative levels of four possible molecules (Table 1): compounds 2, 1, and modified derivatives of both. These fermentation results fell into one of four categories (Figure 4b; Table 1). First, no derivatives were observed; fermentations looked similar to those of 1 without any tailoring genes added (Figure 2b), where only 2 and 1 are observed. A similar profile of products was observed from fermentations of the strains possessing tgtf B, tgtfA, teg13, and stf pek. Second, good conversion to the tailored 1 was observed with very little 2 and 1 remaining, as was the case when MtfApek was expressed (Table 1). Third, there was partial conversion to both derivatives, and all four products were observed as was observed with strains encoding agtf B, gtf Evan, orf19ris, and orf 23ris. For example, when the agtf B construct was moved into the producer of 1, fermentations contained all of the products observed in the 2 producer background as well as a set of replicated peaks with longer retention times representing the tailored sulfated scaffolds (Figure 5c). Lastly, in one case, when the strain containing orf 22ris was fermented, 1 and partial conversion to modified 2 were observed, but tailored 1 was not detected (Figure 4b). Two novel compounds, 1b and 1d, were produced from the fermentations of tailored 1 (Figures 1, 4b). 1b is a product of C-methyltransferase activity at amino acid 3. 1d is a product of glycosylation with N-acetyl-glucosamine (GlcNAc) at amino acid 4. Addition of GlcNAc is especially interesting because it is the first step involved in the formation of the lipoGPAs teicoplanin and A40926.4 All derivatives of 1 were partially purified and then analyzed by HRMS (Table S1) and MS/MS (Figure S30). Two Gene Integrations with pIJ10257. To explore additional opportunities to expand chemical diversity, additional genes were cloned into pIJ10257 derivative plasmids already containing one tailoring gene to create two gene cassettes. Three new tailoring genes (teg14, orf14pek, and auk20) were cloned in various combinations with pre-existing vectors to create seven new plasmids, each containing two modifying genes. Upon introduction of the constructs into the S. coelicolor M1146 producers of 2 and 1 by conjugation, 3 of the 14 strains constructed were able to synthesize the expected product with

Table 1. Relative Conversions of 2 and 1 to Their Tailored Derivatives M1146 pA47934 ΔstaL gene(s) integrated into phiBT1 attB orf19ris orf23ris mtfApek agtf B gtf Evan orf22ris stf pek two genes integrated gtf Evan orf14pek orf 23ris teg14 mtfApek auk20

no.

tailored 2a

2a 2b 2c 2d 2e 2f 2g

++ + ++++ ++ +++ ++ +

2h

+++

2i

+

2j

+

M1146 pA47934 no.

tailored 2, 1a

2a,1a 2b, 1b 2c, 1c 2d, 1d 2e, 1e 2f, −

+, + +, + ++, +++ +, + ++, + ++, −

−, 0−5%; +, 5−25%; ++, 26−50%; +++, 51−75%; ++++, 76−100% conversion. a

Figure 4. Summary of successful and unsuccessful conversions of 2 (a) and 1 (b) resulting from single-gene integrations into the two S. coelicolor M1146 background strains. Arrows indicate the site of modification for the corresponding enzyme, and “X” indicates that the respective enzyme did not modify at that position. Green font indicates novel molecules were produced. In panel b, “∗” indicates that 2 and its modified derivative are also present in the fermentation. Numbers in the aromatic rings indicate the amino acid positions from the N to the C terminus of the heptapeptide. Numbers in parentheses indicate the compound number for the resulting derivative. The gray circle indicates the differentiating sulfate group of compound 1 compared to 2.

Surprisingly, when two functionally homologous enzymes from two different clusters that add the same moiety at the same position on their natural GPA backbones were introduced into the platform, the observed substrate specificity for the unnatural GPA substrate sometimes differed between the two enzymes (Figure 4). For example, when the glycosyltransferases from the teicoplanin (tgtf B) and A40926 (agtf B) biosynthetic clusters that add N-acetyl-glucosamine to amino acid 4 were expressed in producer 2, the glycosylated product was observed only in the strain carrying agtf B but not tgtf B (Figure 4a). Similarly, when the sulfotransferase that acts on amino acid 6 from the TEG (teg13) and pekiskomycin (stf pek) biosynthetic clusters was conjugated into producer 2, only the strain that possessed stf pek produced the sulfated derivative (Figure 4a). We could not predict a priori which enzyme teg13/stf pek oragtf B/tgtf B would be able to utilize our unnatural glycopeptide substrates. These enzyme pairs are functionally 645

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

antibiotic activity and resistance evasion.26 We have shown that secondary sequential modifications can occur at the sugar at amino acid 4, producing 2h. This suggests that more complex glycosylation patterns such as those observed in the GPAs vancomycin and ristocetin (see Figure S2 for structures) that have di- and tetrasaccharide moieties may also be engineered. When tailoring enzymes are combined to make GPAs with multiple modifications, the timing of biosynthetic steps lends insight into which conversions might be more successful. The order of events during GPA biosynthesis differs slightly between teicoplanin type, including A47934, and vancomycin type GPAs (for a recent review see ref 11). The heptapeptide backbone is first synthesized by the NRPS, followed by crosslinking by four27 or three28 mono-oxygenases for teicoplanin or vancomycin type scaffolds, respectively. Halogenation is thought to occur sometime after the first cross-linking but before glycosylation.29 In teicoplanin and chloroeremomycin, glycosylation occurs first at amino acid 4 and then at amino acid 6.30,31 Mannosylation at amino acid 7 can occur both on the aglycone and on the scaffold glycosylated at amino 4.9 Sulfation is likely to be one of the last modifications that occur on the heptapeptide scaffold, but it is unclear if sulfation occurs before or after glycosylation because most Stf studies have been conducted on the aglycone, compound 1.7 However, our data suggest that mannosylation occurs before sulfation, because both sulfation and mannosylation occur efficiently on 2 but the strain carrying both the sulfotransferase and mannosyltransferase, orf22ris in M1146 pA47934, produces only the mannosylated product lacking sulfation (Table 1; Figure 4). As more modifications were made to the GPA scaffold, the frequency of successful conversion decreased. This was observed when the derivatization of 2 was compared to that of 1 (1 has one additional sulfotransferase, StaL) with both one (Figure 4) and two (Table S3) genes integrated into the phiBT1 attB site. The number of derivatized molecules was highest with the simplest substrate, 2, with one added tailoring gene, and the lowest number of molecules was obtained with the more complex substrate, 1, with two added tailoring genes. The consequences of this phenomenon parallel the outcomes of chemical synthesis, where each step rarely occurs with 100% conversion, resulting in a reduced yield of final product as the number of intermediate steps is increased. Families of compounds, dominant products along with many lower abundance structurally similar derivatives, are often observed in fermentation broths,32 which are likely the result of 128 64 >256

0.5 >128 1 128

1 >128 1 >256

0.5 >128 0.5 256

2 >128 4 256

4 >128 4 256

2 >128 2 128

1 >128 1 64

2 >128 4 256

compared to 2 and its derivatives (Figure S32). Likewise, vancomycin MIC values were higher than those of 2 and its derivatives when tested against vancomycin-resistant enterococci and S. coelicolor M1146 (Table 2), confirming that these new antibiotics suppress resistance and offering a route to GPAs that are less affected by existing resistance in the clinic. Encoded within the genomes of many microbes is the capacity to biosynthesize diverse and complex natural products, small specialized metabolites with a plethora of biological activities. With the advent of rapid genome sequencing and improved bioinformatic analysis of biosynthetic programs, we have ready access to biosynthetic genetic diversity. Here we show how this diversity can be harnessed to expand chemical diversity. Previous efforts to expand natural product chemical diversity have largely relied on chemical synthesis25 or in vitro enzymatic2,24,37 modification of scaffolds. The platform described here is complementary to these traditional strategies for creating chemical libraries, which can be screened for any given bioactivity of choice. We have applied our synthetic biology approach to generate GPAs with improved ability to avoid resistance, offering an orthogonal strategy to address the challenge of antibiotic resistance. The success of this platform in mining biochemical diversity to generate new natural products is imminently applicable to many therapeutic areas and is not limited to antibiotics.

Unpredictable substrate promiscuity has implications for large-scale approaches to natural product combinatorial biosynthesis. Without in vitro analysis, it is difficult to predict the substrate promiscuity of a given modifying enzyme. This difficulty can be addressed by using a large number of enzymes in combination with a given chemical scaffold. Previous studies have used cDNA libraries constructed from diverse viruses and eukaryotes to incorporate a broad range of enzymes into their platforms to derivatize several natural product backbones, resulting in biologically active small scaffold sized molecules ranging from 200 to 350 Da.33 Larger modified molecules were not observed, suggesting that enzymes from cDNA libraries were unable to modify the larger, complex scaffolds. In combination with our observations, this suggests that enzyme libraries should be large and also biased toward enzymes having native substrates that are similar to the scaffold being derivatized to increase the chances of successful conversion of complex starter molecules. Induction of van Resistance Genes. GPA sulfation results in attenuated induction of vancomycin resistance and van operon expression;34 therefore, novel derivatives of 2, including 2g, which is sulfated at amino acid 6, were tested for their ability to induce van gene expression. Two assays were used to assess van gene expression and GPA resistance induction, a teicoplanin resistance induction and a van gene expression reporter assay. Both resistance induction and expression of the van genes as measured by these two assays gave comparable results, showing resistance induction only at a higher concentration in response to vancomycin when compared to compound 2 and its derivatives. In GPA resistance assays with S. coelicolor M1146, vancomycin induction of resistance at higher drug concentrations was shown by larger zones of growth when compared to compound 2 and its derivatives (Figure 6). The vanJ promoter is often used as a reporter for van gene expression and vancomycin resistance because vanSR, vanJ, vanK, and vanHAX transcripts are all positively regulated by the VanRS two-component system.35,36 When a vanJp::gusA reporter was used for vanSRJKHAX expression in S. coelicolor M1146, GusA expression in response to vancomycin occurred at higher concentrations and over a wider range of concentrations when



MATERIALS AND METHODS General Methods, Culture Methods, and Reagents. E. coli and Streptomyces were cultivated using media and methods described previously.37,38 The following antibiotics were used to supplement media for selection as needed: hygromycin B (40 μg/mL in Streptomyces; 100 μg/mL in E. coli), kanamycin (50 μg/mL), chloramphenicol (35 μg/mL), nalidixic acid (25 μg/mL), and thiostrepton (50 μg/mL). All antibiotics and media components were purchased from Sigma (St. Louis, MO, USA). Oligonucleotide synthesis and DNA sequencing were all performed at The MOBIX Lab Central Facility (McMaster University). Oligonucleotides used in this study are listed in Table S4. PCR products for cloning were generated using Phusion Polymerase (Invitrogen). Sanger sequencing was used 647

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

μm, C18 column (10 × 100 mm) and an Atlantis T3, 5 μm, C18 column (10 × 100 mm) as described previously.37 Yields of the singly modified (one gene integrated) and doubly modified (two genes integrated) compounds analyzed by NMR were 1−10 mg/L. Total GPA content of M1146 pA47934 and M1146 pA47934 ΔstaL and all derivative strains were similar. Yields of the final modified product reflect the degree of conversion as shown in Figure 3. Analytical RP-HPLC, LC-MS, MS/MS, and NMR Analyses. After elution from the D-Ala-D-Ala affinity column, all partially purified fermentations were analyzed by RP-HPLC using an XSelect CSH, 5 μm, C18 column (4.6 × 100 mm). Percent conversion to products was calculated using the area under the curve at 220 and/or 280 nm. See supplemental methods in the Supporting Information for further details. LC-ESI-MS, MS/MS, and NMR data were collected and analyzed as described previously37 except that 2d, 2e, 2f, and 2h were solubilized in NMR solvent using trifluoroacetic acid vapor and 2, 2b, and 2g were solubilized in NMR solvent with 0.3% NH4OH. Because the 6e proton and carbon signals (as defined in Figure S28) of 2d, 2e, and 2h were difficult to determine in the D2O/ACN-d3 solvent system, NMR spectra were also determined in DMSO-d6. For compounds 2d, 2e, and 2h, before spectra were determined in DMSO-d6, protons were exchanged in D2O/DMSO-d6 (3:2, v/v) overnight at room temperature one to two times. Rotating frame Overhauser effect spectroscopy (ROESY) was also performed on 2, to distinguish between 3b/3f and 4b/4f protons (as defined in Figure S28). MIC Determination. To determine the MIC, Streptomyces were grown in SIM media using the broth microdilution method as described previously.42 Enterococcal MICs were determined using the broth microdilution method in BBL brain−heart infusion broth according to NCCLS protocols. GUS Assays and GPA Resistance Assays. For GUS plate assays, starter cultures of M1146 vanJp::gusA were grown in 2XYT broth38 supplemented with apramycin for 2 days, homogenized using a glass homogenizer, diluted to an OD450 of 0.1−0.3, and swabbed onto Bennett’s agar plates containing 160 μg/mL X-Gluc (ThermoFisher Scientific, Canada). Sterile 6 mm filter disks (AMD Manufacturing, Ontario, Canada) containing 1 or 10 μg of compounds were placed on agar plates. Plates were incubated at 30 °C for 3 days and then photographed. For GPA resistance assays, starter cultures of M1146 were grown as with GUS assay plates but swabbed again onto SIM plates containing 2.5 μg of teicoplanin. Plates were incubated at 30 °C for 4 days and then photographed.

to verify that constructs, including gene additions and deletions, were error free. Construction of a PAC Containing the A47934 Biosynthetic Cluster. A library of clones containing 100− 150 kb fragments of S. toyocaensis NRRL 15009 genomic DNA were cloned into the PAC vector pESAC1321 and screened by Bio S&T (Montreal, Canada). Clones were screened by PCR using oligonucleotides hybridizing to the middle and the ends of the biosynthetic cluster. No clones were positive for all three primer sets. One clone was identified with a 140 kb fragment encoding almost all of the A47934 cluster, pA47934partial, missing the 3′ end of vanA and all of vanX. Oligonucleotides used in the study are listed in Table S4. The lamba red system39 was used to complete the biosynthetic cluster, creating pA47934, and to delete the staL gene from pA47934 to produce pA47934 ΔstaL. Cloning of Tailoring Genes. Tailoring genes were cloned into the integrating and expression vector, pIJ10257, which carries the strong constitutive promoter ermEp* and integrates into phiBT1 attB.40 Most tailoring genes were PCR amplified from genomic DNA: gtf Evan from the vancomycin producer Amycolatopsis orientalis ATCC 19575; stf pek, orf14pek, and mtfApek from the pekiskomycin producer Streptomyces malachitospinus WAC4229;11 orf19ris, orf 22ris, and orf 23ris from the ristocetin producer A. lurida NRRL 2430; tgtfA and tgtf B from the teicoplanin producer Actinoplanes teichomyceticus ATCC 31121; and agtf B from the A4096 producer Actinomadura sp. ATCC 39727. The genes teg14 and teg13 were originally identified from metagenomic samples,10 whereas auk20 is from the UK68,597 producer Acintoplanes sp. ATCC 53533.37 The genes teg14, teg13, and auk20 were amplified from synthesized DNA to remove common restriction sites and cloned into pUC57. Genes were cloned into pIJ0257 as described in the Supporting Information (supplemental methods). Error-free constructs were transformed into ET12567/pUZ8002 for conjugation into Streptomyces. Triparental Mating and E. coli−Streptomyces Conjugation. To create donor strains, PAC vectors were moved into E. coli ET12567 using triparental mating as described previously with the driver plasmid pR9406.21 Streptomyces exconjugants were selected with thiostrepton and PCR-verified using primers that hybridize the ends and middle of the biosynthetic cluster. Plasmids carrying tailoring genes cloned into pIJ10257 were transformed into E. coli ET12567/pUZ8002 cells using standard methods. Resulting transformants were conjugated into S. coelicolor M1146 carrying pA47934 or pA47934 ΔstaL using standard methods.38,40 Exconjugants were selected with hygromycin and thiostrepton and PCR-verified using the primers that amplify the multiple cloning site of pIJ10257. Fermentation and Purification. Fermentation of S. coelicolor was conducted as described previously with S. toyocaensis7 with the following modifications. Starter fermentation cultures in Streptomyces Vegetative Medium were supplemented with hygromycin and thiostrepton. Diaion HP20 (Sigma-Aldrich) resin was used to extract GPAs from conditioned media. Neutralized cell extracts and HP-20 fractions were further purified using affinity chromatography followed by semipreparative HPLC. Conditioned media and cell extract fractions were combined and fractionated by affinity chromatography as described previously using D-Ala-D-Ala coupled to Affigel 10 Gel resin (Bio-Rad).41 Semipreparative RP-HPLC separation was carried out using an XSelect CSH, 5



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsinfecdis.6b00105.



Figures S1−S33, Tables S1−S4, supplemental methods (PDF)

AUTHOR INFORMATION

Corresponding Author

*(G.D.W.) E-mail: [email protected]. 648

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

Present Address

interaction with the cell wall precursor lipid II. Antimicrob. Agents Chemother. 53, 3375−3383. (14) Weber, T., Blin, K., Duddela, S., Krug, D., Kim, H. U., Bruccoleri, R., Lee, S. Y., Fischbach, M. A., Muller, R., Wohlleben, W., Breitling, R., Takano, E., and Medema, M. H. (2015) antiSMASH 3.0 − a comprehensive resource for the genome mining of biosynthetic gene clusters. Nucleic Acids Res. 43, W237−243. (15) Thaker, M. N., and Wright, G. D. (2015) Opportunities for synthetic biology in antibiotics: expanding glycopeptide chemical diversity. ACS Synth. Biol. 4, 195−206. (16) Gomez-Escribano, J. P., and Bibb, M. J. (2011) Engineering Streptomyces coelicolor for heterologous expression of secondary metabolite gene clusters. Microb. Biotechnol. 4, 207−215. (17) Gregory, M. A., Till, R., and Smith, M. C. (2003) Integration site for Streptomyces phage phiBT1 and development of site-specific integrating vectors. J. Bacteriol. 185, 5320−5323. (18) Fayed, B., Younger, E., Taylor, G., and Smith, M. C. (2014) A novel Streptomyces spp. integration vector derived from the S. venezuelae phage, SV1. BMC Biotechnol. 14, 51. (19) Combes, P., Till, R., Bee, S., and Smith, M. C. (2002) The streptomyces genome contains multiple pseudo-attB sites for the (phi)C31-encoded site-specific recombination system. J. Bacteriol. 184, 5746−5752. (20) Pootoolal, J., Thomas, M. G., Marshall, C. G., Neu, J. M., Hubbard, B. K., Walsh, C. T., and Wright, G. D. (2002) Assembling the glycopeptide antibiotic scaffold: the biosynthesis of A47934 from Streptomyces toyocaensis NRRL15009. Proc. Natl. Acad. Sci. U. S. A. 99, 8962−8967. (21) Jones, A. C., Gust, B., Kulik, A., Heide, L., Buttner, M. J., and Bibb, M. J. (2013) Phage p1-derived artificial chromosomes facilitate heterologous expression of the FK506 gene cluster. PLoS One 8, e69319. (22) Hong, H. J., Hutchings, M. I., Hill, L. M., and Buttner, M. J. (2005) The role of the novel Fem protein VanK in vancomycin resistance in Streptomyces coelicolor. J. Biol. Chem. 280, 13055−13061. (23) Beltrametti, F., Lazzarini, A., Brunati, C., Marazzi, A., Jovetic, S., Selva, E., and Marinelli, F. (2003) Production and characterization of monochlorinated and dechlorinated A40926 derivatives. J. Antibiot. 56, 773−782. (24) Losey, H. C., Peczuh, M. W., Chen, Z., Eggert, U. S., Dong, S. D., Pelczer, I., Kahne, D., and Walsh, C. T. (2001) Tandem action of glycosyltransferases in the maturation of vancomycin and teicoplanin aglycones: novel glycopeptides. Biochemistry 40, 4745−4755. (25) Butler, M. S., Hansford, K. A., Blaskovich, M. A., Halai, R., and Cooper, M. A. (2014) Glycopeptide antibiotics: back to the future. J. Antibiot. 67, 631−644. (26) Liu, Y. C., Li, Y. S., Lyu, S. Y., Hsu, L. J., Chen, Y. H., Huang, Y. T., Chan, H. C., Huang, C. J., Chen, G. H., Chou, C. C., Tsai, M. D., and Li, T. L. (2011) Interception of teicoplanin oxidation intermediates yields new antimicrobial scaffolds. Nat. Chem. Biol. 7, 304−309. (27) Hadatsch, B., Butz, D., Schmiederer, T., Steudle, J., Wohlleben, W., Sussmuth, R., and Stegmann, E. (2007) The biosynthesis of teicoplanin-type glycopeptide antibiotics: assignment of p450 monooxygenases to side chain cyclizations of glycopeptide a47934. Chem. Biol. 14, 1078−1089. (28) Stegmann, E., Pelzer, S., Bischoff, D., Puk, O., Stockert, S., Butz, D., Zerbe, K., Robinson, J., Sussmuth, R. D., and Wohlleben, W. (2006) Genetic analysis of the balhimycin (vancomycin-type) oxygenase genes. J. Biotechnol. 124, 640−653. (29) Schmartz, P. C., Wolfel, K., Zerbe, K., Gad, E., El Tamany el, S., Ibrahim, H. K., Abou-Hadeed, K., and Robinson, J. A. (2012) Substituent effects on the phenol coupling reaction catalyzed by the vancomycin biosynthetic P450 enzyme OxyB. Angew. Chem., Int. Ed. 51, 11468−11472. (30) Lu, W., Oberthur, M., Leimkuhler, C., Tao, J., Kahne, D., and Walsh, C. T. (2004) Characterization of a regiospecific epivancosaminyl transferase GtfA and enzymatic reconstitution of the antibiotic chloroeremomycin. Proc. Natl. Acad. Sci. U. S. A. 101, 4390−4395.



(M.N.T.) Synthetic Biology Team, Global Discovery Chemistry, Novartis Institutes for Biomedical Research, 700 Main St., Cambridge, MA 02139, USA. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS



REFERENCES

We thank Flavia Alves for obtaining Orbitrap MS/MS spectra. We thank Mervyn Bibb for his kind gift of the strain S. coelicolor M1146. This research was funded by Canadian Institutes of Health Research (MT-14981) and by a Canada Research Chair (to G.D.W.). G.Y. was supported by a M.G. DeGroote Fellowship Award and a CIHR postdoctoral fellowship.

(1) Van Boeckel, T. P., Gandra, S., Ashok, A., Caudron, Q., Grenfell, B. T., Levin, S. A., and Laxminarayan, R. (2014) Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data. Lancet Infect. Dis. 14, 742−750. (2) Fu, X., Albermann, C., Jiang, J., Liao, J., Zhang, C., and Thorson, J. S. (2003) Antibiotic optimization via in vitro glycorandomization. Nat. Biotechnol. 21, 1467−1469. (3) Losey, H. C., Jiang, J., Biggins, J. B., Oberthur, M., Ye, X. Y., Dong, S. D., Kahne, D., Thorson, J. S., and Walsh, C. T. (2002) Incorporation of glucose analogs by GtfE and GtfD from the vancomycin biosynthetic pathway to generate variant glycopeptides. Chem. Biol. 9, 1305−1314. (4) Ho, J. Y., Huang, Y. T., Wu, C. J., Li, Y. S., Tsai, M. D., and Li, T. L. (2006) Glycopeptide biosynthesis: Dbv21/Orf2 from dbv/tcp gene clusters are N-Ac-Glm teicoplanin pseudoaglycone deacetylases and Orf15 from cep gene cluster is a Glc-1-P thymidyltransferase. J. Am. Chem. Soc. 128, 13694−13695. (5) Kruger, R. G., Lu, W., Oberthur, M., Tao, J., Kahne, D., and Walsh, C. T. (2005) Tailoring of glycopeptide scaffolds by the acyltransferases from the teicoplanin and A-40,926 biosynthetic operons. Chem. Biol. 12, 131−140. (6) Puk, O., Huber, P., Bischoff, D., Recktenwald, J., Jung, G., Sussmuth, R. D., van Pee, K. H., Wohlleben, W., and Pelzer, S. (2002) Glycopeptide biosynthesis in Amycolatopsis mediterranei DSM5908: function of a halogenase and a haloperoxidase/perhydrolase. Chem. Biol. 9, 225−235. (7) Lamb, S. S., Patel, T., Koteva, K. P., and Wright, G. D. (2006) Biosynthesis of sulfated glycopeptide antibiotics by using the sulfotransferase StaL. Chem. Biol. 13, 171−181. (8) Shi, R., Lamb, S. S., Zakeri, B., Proteau, A., Cui, Q., Sulea, T., Matte, A., Wright, G. D., and Cygler, M. (2009) Structure and function of the glycopeptide N-methyltransferase MtfA, a tool for the biosynthesis of modified glycopeptide antibiotics. Chem. Biol. 16, 401−410. (9) Yushchuk, O., Ostash, B., Pham, T. H., Luzhetskyy, A., Fedorenko, V., Truman, A. W., and Horbal, L. (2016) Characterization of the post-assembly line tailoring processes in teicoplanin biosynthesis. ACS Chem. Biol., DOI: 10.1021/acschembio.6b00018. (10) Banik, J. J., Craig, J. W., Calle, P. Y., and Brady, S. F. (2010) Tailoring enzyme-rich environmental DNA clones: a source of enzymes for generating libraries of unnatural natural products. J. Am. Chem. Soc. 132, 15661−15670. (11) Yim, G., Thaker, M. N., Koteva, K., and Wright, G. (2014) Glycopeptide antibiotic biosynthesis. J. Antibiot. 67, 31−41. (12) World Health Organization. (2014) Antimicrobial Resistance: Global Report on Surveillance, Geneva, Switzerland. (13) Lunde, C. S., Hartouni, S. R., Janc, J. W., Mammen, M., Humphrey, P. P., and Benton, B. M. (2009) Telavancin disrupts the functional integrity of the bacterial membrane through targeted 649

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650

ACS Infectious Diseases

Article

(31) Howard-Jones, A. R., Kruger, R. G., Lu, W., Tao, J., Leimkuhler, C., Kahne, D., and Walsh, C. T. (2007) Kinetic analysis of teicoplanin glycosyltransferases and acyltransferase reveal ordered tailoring of aglycone scaffold to reconstitute mature teicoplanin. J. Am. Chem. Soc. 129, 10082−10083. (32) Fischbach, M. A., and Clardy, J. (2007) One pathway, many products. Nat. Chem. Biol. 3, 353−355. (33) Klein, J., Heal, J. R., Hamilton, W. D., Boussemghoune, T., Tange, T. O., Delegrange, F., Jaeschke, G., Hatsch, A., and Heim, J. (2014) Yeast synthetic biology platform generates novel chemical structures as scaffolds for drug discovery. ACS Synth. Biol. 3, 314−323. (34) Kalan, L., Perry, J., Koteva, K., Thaker, M., and Wright, G. (2013) Glycopeptide sulfation evades resistance. J. Bacteriol. 195, 167− 171. (35) Kwun, M. J., Novotna, G., Hesketh, A. R., Hill, L., and Hong, H. J. (2013) In vivo studies suggest that induction of VanS-dependent vancomycin resistance requires binding of the drug to D-Ala-D-Ala termini in the peptidoglycan cell wall. Antimicrob. Agents Chemother. 57, 4470−4480. (36) Hong, H. J., Hutchings, M. I., Neu, J. M., Wright, G. D., Paget, M. S., and Buttner, M. J. (2004) Characterization of an inducible vancomycin resistance system in Streptomyces coelicolor reveals a novel gene (vanK) required for drug resistance. Mol. Microbiol. 52, 1107− 1121. (37) Yim, G., Kalan, L., Koteva, K., Thaker, M. N., Waglechner, N., Tang, I., and Wright, G. D. (2014) Harnessing the synthetic capabilities of glycopeptide antibiotic tailoring enzymes: characterization of the UK-68,597 biosynthetic cluster. ChemBioChem 15, 2613−2623. (38) Hopwood, D. A., Kieser, T., Bibb, M. J., Buttner, M. J., and Chater, K. (2000) Practical Streptomyces Genetics, John Innes Foundation, Norwich, UK. (39) Datsenko, K. A., and Wanner, B. L. (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97, 6640−6645. (40) Sherwood, E. J., and Bibb, M. J. (2013) The antibiotic planosporicin coordinates its own production in the actinomycete Planomonospora alba. Proc. Natl. Acad. Sci. U. S. A. 110, E2500−2509. (41) Sitrin, R. D., and Wasserman, G. F. (1989) Affinity and HPLC purification of glycopeptide antibiotics. J. Chromatogr. Libr. 43, 111− 152. (42) D’Costa, V. M., McGrann, K. M., Hughes, D. W., and Wright, G. D. (2006) Sampling the antibiotic resistome. Science 311, 374−377.

650

DOI: 10.1021/acsinfecdis.6b00105 ACS Infect. Dis. 2016, 2, 642−650