Identification and Analysis of the Putative Persiathiacin BGC

To identify the persiathiacin BGC, the genome of Actinokineospora sp. UTMC 2448 was sequenced using single-molecule real-time (SMRT) sequencing. A complete circular genome sequence consisting of 7,012,397 bp was obtained using this approach (GenBank accession number CP031087). Analysis of the sequence using antiSMASH identified 32 putative specialized metabolite BGCs (Table S3).²⁵ Among these, a cluster containing 33 genes (cluster 11; Table S3), several of which encode homologues of enzymes involved in the biosynthesis of other thiopeptides, was postulated to direct persiathiacin biosynthesis (Figure ​Figure33). Sequence comparisons showed that the products of perA–perP have a significant degree of similarity to the proteins encoded by nosA–nosP and nocA–nocP in the nosiheptide and nocathiacin BGCs, respectively (Table S4). Homologues of five additional genes in the nocathiacin BGC (nocR and nocT–nocV), absent from the nosiheptide cluster, are present in the putative persiathiacin cluster (perR and perT–perV, respectively; Figure ​Figure33). Moreover, the putative persiathiacin BGC contains 12 genes (perS1–perS12) hypothesized to be responsible for the biosynthesis and attachment of four 6-deoxysugars to the thiopeptide core (Table S4).

Open in a separate window

Figure 3

(A) Comparison of the nosiheptide (nos), nocathiacin (noc), and putative persiathiacin (per) biosynthetic gene clusters. Genes are colored as follows. Blue: thiazole formation; green: Ser/Thr dehydration; orange: DMIA formation and attachment; red: cytochromes P450; gray: 6-deoxysugar biosynthesis and attachment; brown: methyltransferases. (B) The biosynthesis of the thiopeptide core of the persiathiacins is proposed to commence with transcription and translation of perM to yield a precursor peptide comprised of an N-terminal leader peptide (LP) fused to a C-terminal core peptide (structure depicted). The core peptide undergoes a series of post-translational modifications catalyzed by several enzymes encoded by the persiathiacin biosynthetic gene cluster. See main text for further details.

Detailed sequence analysis of perA–perV and perS1–perS12 enabled us to propose a biosynthetic pathway for persiathiacins A and B (Figures ​Figures33 and ​and4).4). First, perM is transcribed and translated into a 49 amino acid (aa) precursor peptide, consisting of a 36 aa N-terminal LP fused to a 13 aa C-terminal core peptide with the sequence SCTTCECSCSCSS, which is fully consistent with the thiopeptide core structure of the persiathiacins deduced from the spectroscopic data.

Open in a separate window

Figure 4

The enzymes encoded by perS1–perS12 are proposed to be responsible for the biosynthesis and attachment of the 6-deoxysugars to the persiathiacin aglycone. (A) Proposed pathway for assembly of TDP-6-deoxy-α-d-glucose 8, TDP-α-d-olivose 10, and TDP-α-d-amicetose 12. (B) The glycosyltransferases encoded by perS4, perS6, persS8, and perS9 are proposed to decorate the persiathiacin core peptide with l-rhamnose, 6-deoxy-d-glucose, d-olivose, and d-amicetose (persiathiacin A 4), or l-rhamnose, 6-deoxy-d-glucose, and d-olivose (persiathiacin B 5). The methyltransferases encoded by perS3, perS5, and perS7 are hypothesized to O-methylate the l-rhamnose and 6-deoxy-d-glucose residues to produce the mature antibiotics. The timing of these transformations remains to be determined.

Cyclodehydration of the cysteine residues is proposed to be catalyzed by PerG and PerH, followed by dehydrogenation catalyzed by PerF to yield the five thiazoles in the persiathiacins.⁷ Putative dehydratases PerD and PerE are proposed to further modify the core peptide by catalyzing selective dehydration of Ser1, Ser10, Ser12, Ser13, and Thr4.

Four enzymes encoded by perI, perK, perL, and perN are proposed to be responsible for the production of 3,4-dimethylindolic acid (DMIA) from l-tryptophan and its attachment to the core peptide (Figures ​Figures33 and S15). PerL is a putative radical S-adenosylmethionine (SAM) enzyme that is homologous to NosL, which transforms l-tryptophan into 3-methyl-2-indolic acid (MIA) via unusual constitutional isomerization of a C–C bond.²⁶⁻³⁴ In nosiheptide biosynthesis, the ATP-dependent NosI enzyme adenylates MIA and loads it onto the phosphopantetheine thiol of the acyl carrier protein (ACP) NosJ. NosK then transfers the MIA residue to Cys8 of the nosiheptide core peptide (Figure S15).³⁵⁻³⁷ The persiathiacin and nocathiacin BGCs both lack nosJ homologues. Sequence comparisons of NosJ with PerI and PerK revealed similarity between NosJ and the C-terminus of PerK. Similarly, it has previously been noted that the C-terminus of NocK is homologous to NosJ.³⁵ Thus, it appears that in persiathiacin and nocathiacin biosynthesis, the C-terminal ACP domains of PerK and NocK are loaded with MIA by PerI and NocI, respectively. The N-terminal domains of PerK and NocK then catalyze attachment of the MIA residue to Ser8 of the persiathiacin and nocathiacin core peptides, respectively (Figure S15). Subsequently, the putative radical SAM methylase PerN is proposed, by analogy with the well-characterized mechanism of NosN,³⁵ to catalyze methylenation of C4 in the MIA residue. The resulting electrophilic intermediate is attacked by the Glu6 carboxylate to form an ester linkage.³⁸ Finally, PerO, which has >50% sequence identity to NosO and NocO, is hypothesized to be responsible for formation of the macrocycle and pyridine in the persiathiacins via a [4 + 2] cycloaddition.^39,40

Of the six putative cytochromes P450 (CYPs) encoded by the persiathiacin BGC, two (PerB and PerC) are homologous to NosB/NocB and NosC/NocC, which hydroxylate C3 of Glu6 and the pyridine, respectively.¹³ The CYPs encoded by perV, perU, and perT are similar in sequence to NocV, NocU, and NocT, respectively, encoded by the nocathiacin BGC. Genes encoding homologues of these enzymes are absent from the nosiheptide BGC. PerV is proposed to perform an analogous function to NocV—i.e., formation of the ether linkage between the indole and core peptide via a mechanism yet to be elucidated.⁴¹ Similarly, PerU is hypothesized to catalyze N-hydroxylation of the indole, by analogy with the proposed function of NocU.⁴² Comparison of the structures of persiathiacin A, nocathiacin I and nosiheptide (Figure ​Figure44) suggests the putative CYPs encoded by perT/nocT and methyltransferases encoded by perQ/nocQ catalyze hydroxylation and subsequent O-methylation of the dehydrobutyrine residue to form the corresponding O-methyl-Dht residue. The only CYP-encoding gene in the persiathiacin BGC that does not have a homologue in either the nosiheptide or nocathiacin BGCs is perX. Structural comparison of persiathiacin A with nocathiacin I and nosiheptide suggests that the enzyme encoded by this gene catalyzes hydroxylation of C5 in thiazole 3, to create the attachment site for the trisaccharide (Figure ​Figure55).

Open in a separate window

Figure 5

Comparative analysis of the structures of nosiheptide 1, nocathiacin I 2, and persiathiacin A 4 and the functions of the CYPs encoded by their BGCs. Blue dashed boxes highlight hydroxyl groups proposed to be installed by homologous CYPs (NosB/NocB/PerB and NosC/NocC/PerC) encoded by all three BGCs. Red dashed circles highlight hydroxyl groups proposed to be introduced by CYPs (NocT/PerT, NocV/PerV, and NocU/PerU) encoded by the nocathiacin and persiathiacin BGCs, but not the nosiheptide BGC. A purple dashed circle highlights the hydroxyl group proposed to be introduced by the CYP encoded by perX, which is only present in the persiathiacin BGC.

The final enzyme proposed to be involved in the assembly of the thiopeptide core of the persiathiacins is PerA. This enzyme is homologous to NosA, which catalyzes dealkylative cleavage of the C-terminal Dha residue in nosiheptide biosynthesis, resulting in formation of the corresponding amide.⁴³ PerA is proposed to catalyze an analogous reaction in persiathiacin biosynthesis (Figure ​Figure33).

The enzymes encoded by perS1–perS12 are proposed to assemble the glycosyl residues and catalyze their attachment to the thiopeptide core of the persiathiacins. The biosynthesis of these 6-deoxysugars is proposed to commence with the conversion of thymine diphosphate (TDP)-α-d-glucose 6 to TDP-4-keto-6-deoxy-α-d-glucose 7 catalyzed by PerS2, which shows sequence similarity to TDP-glucose-4,6-dehydratases (Figure ​Figure44). At this point, the pathway appears to bifurcate. While PerS11, which is similar to TDP-hexose-4-ketoreductases, is proposed to catalyze the formation of TDP-6-deoxy-α-d-glucose 8 from TDP-4-keto-6-deoxy-α-d-glucose 7, PerS10 and PerS1, which are homologues of TDP-4-keto-6-deoxy-d-glucose-2,3-dehydratases and TDP-4-keto-6-deoxy-d-glucose-3-ketoreductases, respectively, are hypothesized to convert 7 to TDP-4-keto-2,6-deoxy-α-d-glucose 9. A further bifurcation then occurs. PerS11 catalyzes the conversion of TDP-4-keto-2,6-deoxy-α-d-glucose 9 to TDP-d-olivose 10, whereas PerS12, which is similar to TDP-4-keto-2,6-deoxy-d-glucose-3-dehydratases, converts 9 to TDP-4-keto-2,3,6-deoxy-α-d-glucose 11. Finally, PerS11 catalyzes the reduction of TDP-4-keto-2,3,6-deoxy-α-d-glucose 11 to TDP-d-amicetose 12. This analysis is consistent with the assignment of sugars 2, 3, and 4 as d-amicetose, d-olivose, and 3,4-di-O-methyl-6-deoxy-β-d-glucose, respectively, in persiathicin A 4, and the substitution of d-amicetose by a second d-olivose residue in persiathiacin B 5.

Philipimycin 3, which bears the highest degree of structural similarity among known thiopeptide antibiotics to the persiathiacins, is proposed to be decorated with d-amecitose and two O-methylated l-rhamnose derivatives.¹⁷ Moreover, no 6-deoxysugar biosynthetic genes, beyond those hypothesized to be involved in the assembly of TDP-d-amicetose, TDP-d-olivose, and TDP-6-deoxy-β-d-glucose, are present in the persiathiacin BGC (Figure ​Figure33 and Table S4). Because l-rhamnose is ubiquitously incorporated into bacterial cell surface carbohydrates,⁴⁴ dedicated genes for TDP-l-rhamnose biosynthesis are invariably absent from rhamnosylated natural product BGCs. Taken together, these observations are consistent with the assignment of sugar 1 as an O-dimethylated l-rhamnose derivative. The deoxysugar residues contain a total of four methoxy groups (appended to C2 and C3 in sugar 1 and C3 and C4 in sugar 4), but only three putative O-methyltransferase-encoding genes (perS3, perS5, and perS7) are present in the persiathiacin BGC. It therefore appears that one of PerS3, PerS5, and PerS7 catalyzes the O-methylation of two distinct hydroxyl groups in the biosynthesis of one of these sugars.

Four genes (perS4, perS6, perS8, and perS9) encode putative glycosyltransferases, each of which is hypothesized to append one glycosyl residue to the thiopeptide core. Glycosyltransferases are known to possess broad substrate tolerance,⁴⁵ explaining why small amounts of persiathiacin B 5, in which sugar 2 is d-olivose rather than d-amicetose, are produced in addition to persiathiacin A 4.

Given that (i) the core peptide sequence encoded by perM is in complete accord with the amino acid residues found in the thiopeptide core of persiathiacins A and B; (ii) the nocathiacins and persiathiacins have identical thiopeptide core structures, and the putative persiathiacin BGC encodes homologues of the full complement of enzymes needed to assemble the nocathiacin thiopeptide core from the precursor peptide; (iii) the persiathiacin BGC encodes an additional CYP (PerX) not found in the nocathiacin BGC, all the expected enzymes for assembly of the 6-deoxysugars in the persiathiacins, and four glycosyltransferases (compared to only one in the nocathiacin BGC), fully consistent with the tetraglycosylated structure of the persiathiacins; and (iv) the complete genome sequence of Actinokineospora sp. UTMC 2448 does not contain any other gene clusters with a significant similarity to known thiopeptide BGCs, it is highly probable that the BGC we have identified directs persiathiacin biosynthesis. Notwithstanding this, we endeavored to experimentally verify this hypothesis.

PerX Catalyzes Hydroxylation of Nosiheptide

Due to a lack of genetic tools for the Actinokineospora genus, we were unable to obtain experimental evidence for the involvement of the nocathiacin-like BGC in Actinokineospora sp. UTMC 2448 in persiathiacin assembly via targeted disruption of one of the putative biosynthetic genes. Instead, we decided to investigate the ability of the putative CYP PerX to hydroxylate thiopeptides. Recombinant His₆-tagged PerX was overproduced in Escherichia coli and purified using nickel-affinity chromatography. The identity of the purified protein, including the presence of a haem prosthetic group, was confirmed by ESI-Q-TOF-MS analysis (Figure S16). The purified protein was incubated with commercially available nosiheptide 1, spinach ferredoxin, spinach ferredoxin reductase, and NADPH at room temperature for 3 h. UHPLC-ESI-Q-TOF-MS analysis of the reaction mixture revealed a species with m/z = 1238.1493, corresponding to the [M + H]⁺ ion for a compound with the molecular formula C₅₁H₄₃N₁₃O₁₃S₆ (calculated m/z = 1238.1500 for C₅₁H₄₄N₁₃O₁₃S₆⁺) that was absent from a control reaction containing heat-inactivated enzyme. The molecular formula of this species is consistent with the insertion of an oxygen atom into the nosiheptide backbone (measured m/z = 1222.1542; calculated m/z = 1222.1551 for C₅₁H₄₄N₁₃O₁₂S₆⁺) (Figure ​Figure66). These data indicate that PerX can hydroxylate substrate analogues with significant modifications to the persiathiacin/nocathiacin core thiopeptide structure, suggesting it may hold promise for development into a new tool for targeted thiopeptide structural modification.

Open in a separate window

Figure 6

(A) Reaction catalyzed by purified recombinant PerX with nosiheptide 1, in the presence of spinach ferredoxin (Fd), spinach ferredoxin reductase (Fr) and NADPH. The proposed site of oxygen atom insertion, based on the assigned function of PerX in persiathiacin biosynthesis, is highlighted in red. (B) Extracted ion chromatograms at m/z = 1222.1551 and 1238.1500, corresponding to [M + H]⁺ for nosiheptide and its hydroxylated derivative, respectively, from UHPLC–ESI-Q-TOF-MS analyses of: culture extracts of S. actuosus ATCC25421 expressing perX under the control of the strong constitutive ermE* promoter (top chromatogram); nosiheptide 1 after incubation for 3 h with purified recombinant PerX, Fd, Fr, and NADPH (middle chromatogram); and nosiheptide 1 after incubation for 3 h with heat-denatured PerX, Fd, Fr and NADPH (bottom chromatogram).

In an attempt to obtain sufficient quantities of the oxygenated nosiheptide derivative for NMR spectroscopic analysis, perX was expressed under the control of the constitutive ermE* promoter in the nosiheptide producer Streptomyces actuosus ATCC25421. Although the same nosiheptide derivative as that produced in the in vitro experiments was observed in UHPLC–ESI-Q-TOF-MS analyses of extracts from this strain (Figure ​Figure66), it was not possible to isolate sufficient quantities of the compound for full characterization by NMR spectroscopy.

PerS4 Catalyzes Glycosylation of Nosiheptide

To further validate the involvement of the identified gene cluster in persiathiacin biosynthesis, we investigated the ability of the putative glycosyltransferase PerS4 to glycosylate thiopeptides. A homologue of PerS4 from Actinobacteria fastidiosa JCM3276 has recently been reported to rhamnosylate nosiheptide.⁴⁶ Recombinant His₆-tagged PerS4 was overproduced in E. coli and purified using nickel-affinity chromatography and its identity was confirmed by ESI-Q-TOF-MS analysis (Figure S17). The purified protein was incubated with commercially available nosiheptide 1 and TDP-α-d-glucose at 30 °C for 12 h. UHPLC–ESI-Q-TOF-MS analysis of the reaction mixture revealed a species with m/z = 1384.2092, corresponding to the [M + H]⁺ ion for a compound with the molecular formula C₅₇H₅₂N₁₃O₁₇S₆ (calculated m/z = 1384.2079 for C₅₇H₅₃N₁₃O₁₇S₆⁺) that was absent from a control reaction containing heat-inactivated enzyme. The molecular formula of this species is consistent with the attachment of a d-glucose residue to one of the hydroxyl groups in nosiheptide (Figure ​Figure77). Nosiheptide 1 contains the 3-hydroxypyridine moiety that is glycosylated with the dimethylated 6-deoxy-d-glucose derivative in the persiathiacins but lacks the hydroxylated thiazole that serves as the attachment site for the trisaccharide. We therefore tentatively conclude that PerS4 catalyzes transfer of the d-glucose residue to the hydroxypyridine moiety of nosiheptide (Figure ​Figure77), but further experiments will be required to confirm this.

Open in a separate window

Figure 7

(A) Reaction proposed to be catalyzed by purified recombinant PerS4 with nosiheptide 1 and TDP-α-d-glucose. (B) Extracted ion chromatograms at m/z = 1384.2080, corresponding to [M + H]⁺ for glycosylated nosiheptide, from UHPLC–ESI-Q-TOF-MS analyses of incubations of nosiheptide 1 with TDP-α-d-glucose and PerS4 (bottom) and a control reaction containing heat-inactivated PerS4. (C) Comparison of the simulated mass spectrum for C₅₇H₅₃N₁₃O₁₇S₆⁺ (top) with the measured spectrum of the species eluting at 16.1 min, corresponding to glycosylated nosiheptide (bottom).

Strain Isolation and Identification

Actinokineospora sp. UTMC 2448 was isolated from mud sample collected from Bushehr, Iran. The sample was dried at 50 °C, ground to a powder and passed through a 2 mm sieve. Strains were isolated on solid Reasoner’s 2A medium⁵⁵ after 3 weeks of incubation at 28 °C. Solid ISP2 medium was then used to purify the strains. Purified strains were preserved in 30% glycerol at −70 °C. To identify the strains, 16S rRNA genes were amplified using a set of universal primers (27F, 1100F, 1100R, 1525R). Amplified DNA obtained from the reactions was purified using a PCR purification kit (Roti-Prep PCR Purification). The 16S rRNA gene sequence of the strains was BLASTed against the GenBank and EzTaxon databases.⁵⁶

Production, Extraction, and HPLC Purification of Persiathiacins A and B

Actinokineospora sp. UTMC 2448 was grown on solid ISP2 medium (4 g/L glucose, 4 g/L yeast extract, 10 g/L malt extract, 2 g/L CaCO₃, 15 g/L Bacto agar) for 7 days at 30 °C. The agar cultures were chopped and extracted with EtOAc. The extract was dried on a rotary evaporator and preadsorbed to C18-bonded silica, and then packed into a stainless steel HPLC guard cartridge (10 × 30 mm) attached to a semipreparative reverse-phase C18 Betasil column (21.2 × 150 mm). The column was eluted with 5% acetonitrile for 5 min, then a linear gradient from 5 to 100% acetonitrile was applied over 45 min, and the column was eluted for an additional 10 min with 100% acetonitrile. The flow rate was 9 mL/min. Sixty fractions were collected in 1 min increments over 60 min. Pure persiathiacin A was obtained in fraction 35. Persiathiacin B was purified from a mixture of persiathiacins A and B in fraction 34 using a reverse-phase C18 Betasil column (21.2 × 150 mm). Isocratic elution with 40% acetonitrile for 5 min followed by a linear gradient to 65% over 45 min was used to achieve separation of persiathiacin B (fraction 26) from persiathiacin A (fraction 28).

PacBio Library Preparation and Sequencing

Genomic DNA was extracted from Actinokineospora sp. UTMC2448. A SMRTbell template library was prepared according to the manufacturer’s instructions (Pacific Biosciences, Menlo Park, CA, USA), following the Procedure & Checklist—Greater Than 10 kb Template Preparation. Briefly, for preparation of 15 kb libraries, 8 μg genomic DNA was sheared using g-tubes (Covaris, Woburn, MA, USA) according to the manufacturer's instructions. DNA was end-repaired and ligated overnight to hairpin adapters by applying components from the DNA/Polymerase Binding Kit P6 (Pacific Biosciences, Menlo Park, CA, USA). Reactions were carried out according to the instructions of the manufacturer. BluePippin size-selection to greater than 4 kb was performed according to the manufactureŕs instructions (Sage Science, Beverly, MA, USA). Conditions for annealing of sequencing primers and binding of polymerase to purified SMRTbell template were assessed with the Calculator in RS Remote (Pacific Biosciences, Menlo Park, CA, USA). SMRT sequencing was carried out on the PacBio RSII (Pacific Biosciences, Menlo Park, CA, USA), taking one 240 min movie for one SMRT cell using the P6 Chemistry. Sequencing resulted in 76,562 postfiltered reads with a mean read length of 10,180 bp.

Genome Assembly, Error Correction, and Annotation

SMRT cell data were assembled using the “RS_HGAP_Assembly.3” protocol included in SMRT Portal version 2.3.0 using default parameters. The assembly resulted in a single circular chromosome. Error correction was performed by a mapping of 7 million paired-end Illumina reads of 2 × 100 bp onto the genome using BWA (PMID 19451168)⁵⁷ with subsequent variant and consensus calling using VarScan (PMID 22300766).⁵⁸ A consensus concordance of QV60 could be confirmed for the genome. Finally, annotation was carried out using Prokka 1.8 (PMID 24642063).⁵⁹ Prediction of specialised metabolite BGCs was made using antiSMASH v3.0 (Table S3). The putative persiathiacin biosynthetic gene cluster was subjected to detailed manual annotation via comparative sequence analysis (Table S4).

Overproduction and Purification of PerX

The gene encoding PerX was PCR-amplified from Actinokineospora sp. UTMC 2448 gDNA using Phusion DNA polymerase (NEB) and primers 5′-GTGCCGCGCGGCAGCCATATGCTTCCCGAGCCGTACACCCCCGAGTTCT-3′ and 5′-TCGACGGAGCTCGAATTCTCATCGCGTCACCCGCAGCTCGGCCA-3′ (regions complementary to the gene sequence underlined). The linear pET28a (NEB) vector backbone was PCR-amplified with primers 5′-TGAGAATTCGAGCTCCGTCGACAAGCTTG-3′ and 5′-CATATGGCTGCCGCGCGGCAC-3′. PCR products were separated on a 1% agarose gel and bands were excised and purified with a GeneJET Gel Extraction Kit (Thermo Scientific). Cloning of the pure insert into the NdeI/EcoRI restriction sites of the linear pET28a vector was accomplished by Gibson assembly following the manufacturer’s instructions (NEB). The resulting vector was used to transform E. coli TOP10 cells (Invitrogen) and plated on LB agar containing kanamycin (50 μg/mL). Colonies were picked and grown overnight in liquid LB medium. Plasmids were isolated from the culture using a GeneJET Plasmid Miniprep Kit (Thermo Scientific) and inserts were sequenced to verify their integrity. The correct pET28a plasmid containing perX was used to transform E. coli BL21(DE3) cells. A single colony was used to inoculate liquid LB medium (10 mL) containing kanamycin (50 μg/mL), which was incubated overnight at 37 °C and 180 rpm; this was then used to further inoculate liquid LB medium (1 L) containing kanamycin (50 μg/mL). The resulting culture was incubated at 37 °C and 180 rpm until OD_595 nm reached 0.6, then IPTG (0.5 mM) was added, and expression was continued overnight at 15 °C and 180 rpm. The cells were harvested by centrifugation (5000 rcf, 20 min, 4 °C) and resuspended in buffer (30 mM HEPES, 500 mM NaCl, 10% glycerol, pH 7.5) at 20 mL/L of growth medium, then lysed using sonication (Vibra-Cell Ultrasonic Liquid Processor; Sonics & Materials, Inc.). The lysate was centrifuged (30,000 rcf, 60 min, 4 °C) and the resulting supernatant was passed through a 0.45 μm filter (Sartorius). An ÄKTA pure FPLC (GE Healthcare) was used to purify PerX as follows. The supernatant was loaded onto a 1 mL HisTrap HP column (GE Healthcare), which had been equilibrated with resuspension buffer (30 mM HEPES, 500 mM NaCl, 10% glycerol, pH 7.5). Proteins were eluted in a stepwise manner using increasing concentrations of imidazole (0–150 mM) in resuspension buffer. The presence of the protein of interest in the elution fractions was confirmed by SDS–PAGE. Fractions containing the pure protein were pooled and concentrated to ~100 μM using a 50 kDa MWCO Vivaspin centrifugal concentrator (Sartorius). Aliquots of 50 μL were snap-frozen in liquid N₂ and stored at −80 °C until further use.

Hydroxylation of Nosiheptide by PerX

A 200 μL reaction mixture containing nosiheptide (100 μM), spinach ferredoxin-NADP⁺ reductase (0.1 U/mL), spinach ferredoxin (50 μg/mL), NADPH (1 mM), and PerX (10 μM) in Tris–HCl (25 mM, pH 8) was incubated at room temperature for 3 h. The reaction was terminated by adding 200 μL of methanol, and after separating the precipitate by centrifugation (16,000 rcf, 10 min) the supernatant was analyzed by UHPLC–ESI-Q-TOF-HRMS. For the negative control, PerX was inactivated by boiling at 100 °C for 15 min.

Expression of perX in the Nosiheptide-Producing Strain S. actuosus

perX was amplified from Actinokineospora sp. UTMC 2448 gDNA using Phusion DNA polymerase (NEB) and primers 5′-CAGCATATGGTGCTTCCCGAGCCGTAC-3′ and 5′-GACGAATTCTCATCGCGTCACCCGC-3′. The PCR product was digested with NdeI and EcoRI and cloned into the corresponding sites of pIB139 under the control of the ermE* constitutive promoter. The integrity of the construct was confirmed by sequencing and the resulting plasmid was used to transform E. coli ET12567/pUZ8002 cells by electroporation. A mixture of apramycin (50 μg/mL), kanamycin (50 μg/mL) and chloramphenicol (35 μg/mL) was used for selection on LB agar. The pIB139 vector containing perX was then introduced by conjugation into S. actuosus ATCC25421. The overnight culture was plated on SFM agar medium and overlaid with 1 mL of antibiotic solution mixture containing apramycin (50 μg/mL) and nalidixic acid (25 μg/mL). After 3 days, four colonies were picked and spread separately onto SFM agar medium containing apramycin (50 μg/mL) and nalidixic acid (25 μg/mL) and then further subcultured on five plates to produce spores. Spores from the resulting stocks were cultured in liquid medium containing corn steep liquor (10 g/L), soy flour (20 g/L), yeast extract (3 g/L), NaCl (4 g/L), KNO₃ (0.2 g/L), CaCO₃ (4 g/L), pH 7.0. Production of the hydroxylated nosiheptide derivative was confirmed by UHPLC–ESI-Q-TOF-MS analysis.

Overproduction, Purification and Characterization of PerS4

The gene encoding PerS4 was synthesized and cloned into pET28a (+) by GenScript. The resulting C-terminal hexa-histidine fusion protein was overproduced in E. coli BL21(DE3) as described for PerX, except 0.4 mM IPTG was used and the culture was incubated at 15 °C for 16 h. The cells were lysed and the protein was purified from the cell lysate as described for PerX, except the cell lysate was suspended in 20 mM Tris–HCl, 100 mM NaCl, pH 8.0 and the protein was eluted stepwise using increasing concentrations of imidazole buffer (20–300 mM).

200 μM of purified recombinant PerS4 was incubated with 150 μL of nosiheptide in DMSO (2.4 mg/mL), 150 μL of a solution of TDP-α-d-glucose, prepared from thymidine monophosphate (27 mM) and α-d-glucose-1-phosphate as described previously,⁵⁰ in 50 mM Tris–HCl (pH 7.5, total volume 1 mL). After incubation at 30 °C for 12 h the reaction was quenched by the addition of an equal volume of methanol. The precipitate was removed by centrifugation at 12,000 rpm for 10 min, and the supernatant was analyzed by UHPLC–ESI-Q-TOF-HRMS.

MIC Assays Against M. tuberculosis

DMSO, glycerol, isoniazid, resazurin sodium salt, and rifampicin were purchased from Sigma-Aldrich (USA). Middlebrook 7H9 was purchased from Difco (USA) and albumin dextrose catalase from Chemie Brunschwig AG (Switzerland). The M. tuberculosis reference strain H37Rv was obtained from Institut Pasteur, Paris, and clinical specimens from patients were obtained from the Lausanne University Hospital (CHUV) and Geneva University Hospital (HUG).

The resazurin reduction microplate assay was performed as described previously.⁵⁰ 2-fold serial dilutions of each test compound were prepared in 96-well plates from 10 mg/mL stocks in DMSO. Frozen aliquots of replicating tubercule bacilli (reference strains and clinical isolates) were thawed and diluted to an OD₆₀₀ of 0.0001 (3×10⁴ cells/mL) and added to the plates to obtain a total volume of 100 μL. Plates were incubated for 6 days at 37 °C before adding resazurin (0.025% w/v to 1/10 of well volume). After overnight incubation, fluorescence of the resazurin metabolite resorufin was determined by excitation at 560 nm and emission at 590 nm, as measured by a TECAN infinite M200 microplate reader. The MIC was defined visually as the lowest concentration to prevent resazurin turnover from blue to pink and was confirmed by the level of measured fluorescence. MIC values were calculated using GraphPad Prism version 7.0 (GraphPad Software, Inc., La Jolla, CA, USA). The experiment was performed twice, and all the compounds were tested in triplicate (total of six replicates).

The Bruker MaXis Impact and MaXis II UHPLC-ESI-Q-TOF-MS instruments used in this research were funded by the BBSRC (BB/K002341/1 and BB/M017982/1, respectively, to G.L.C.). G.L.C. was the recipient of a Wolfson Research Merit Award from the Royal Society (WM130033). Y.D. was supported by a grant from the MRC (MR/N501839/1 to G.L.C.). We thank Ivan Prokes and Lijiang Song for assistance with recording NMR and UHPLC–ESI-Q-TOF-MS data, respectively. Simone Schrader and Nicole Heyer are thanked for excellent technical assistance for PacBio genome sequencing. Prof Hung Wen Liu is thanked for providing plasmids encoding the enzymes responsible for converting thymidine triphosphate and α-d-glucose-1-phosphate into TDP-α-d-glucose. Y.Z. thanks Dr. Munro Passmore for assistance with preparation of Figure 7.