Rifamycin W Analogues from Amycolatopsis mediterranei S699 Δrif-orf5 Strain

Rifamycin W, the most predominant intermediate in the biosynthesis of rifamycin, needs to undergo polyketide backbone rearrangement to produce rifamycin B via an oxidative cleavage of the C-12/C-29 double bond. However, the mechanism of this putative oxidative cleavage has not been characterized yet. Rif-Orf5 (a putative cytochrome P450 monooxygenase) was proposed to be involved in the cleavage of this olefinic moiety of rifamycin W. In this study, the mutant strain Amycolatopsis mediterranei S699 Δrif-orf5 was constructed by in-frame deleting the rif-orf5 gene to afford thirteen rifamycin W congeners (1–13) including seven new ones (1–7). Their structures were elucidated by extensive analysis of 1D and 2D NMR spectroscopic data and high-resolution ESI mass spectra. Presumably, compounds 1–4 were derivatized from rifamycin W via C-5/C-11 retro-Claisen cleavage, and compounds 1–3, 9 and 10 featured a hemiacetal. Compounds 5–7 and 11 showed oxygenations at various sites of the ansa chain. In addition, compounds 1–3 exhibited antibacterial activity against Staphylococcus aureus with minimal inhibitory concentration (MIC) values of 5, 40 and 0.5 µg/mL, respectively. Compounds 1 and 3 showed modest antiproliferative activity against HeLa and Caco-2 cells with half maximal inhibitory concentration (IC50) values of about 50 µM.


Introduction
Ansamycins are a family of macrolactam antibiotics that are synthesized by type I polyketide synthase (PKS), which are structurally characterized by an aromatic moiety bridged at nonadjacent positions by an aliphatic chain (ansa chain) [1,2]. As the representative members of the ansamycin family, rifamycins were first isolated from Amycolatopsis mediterranei S699 in 1957 [3][4][5]. Semi-synthetic rifamycin derivatives, such as rifampicin, rifapentine and rifambutin, have long been the first-line antituberculosis drugs since the mid-1960s, and are effective in combating leprosy and tuberculosis involved in AIDSrelated mycobacterial infections [6][7][8][9]. However, Mycobacterium tuberculosis has developed significantly increased resistance to rifamycin antibiotics due to their extensive clinical use during recent decades [10,11].
The biosynthesis of rifamycins has been continuously studied since the 1980s, which can be divided into three stages. During the first two stages, the biosynthesis of start unit 3-amino-5-hydroxybenzoic acid (AHBA) and the polyketide skeleton were investigated, respectively. The third stage is still in progress, involving exploring hypotheses concerning post-PKS modifications [12][13][14][15][16]. As the most predominant intermediate in rifamycin biosynthesis, rifamycin W must undergo C-12/C-29 double bond oxidative cleavage to form 27-O-demethyl-25-O-deacetyl-rifamycin S (DMDARS) that is the basic rifamycin B polyketide skeleton. However, the mechanism of this putative oxidative cleavage has not been characterized yet [16,17].

Bacterial Strains, Plasmids and Culture Media
The Amycolatopsis mediterranei S699 strain, isolated in 1957 in St. Raphael, France [5], was stored in our lab. The A. mediterranei S699 Δrif-orf5 strain was constructed by deleting the rif-orf5 gene through homologous recombination. The A. mediterranei S699 Δrif-orf5::orf5 strain was constructed by transformation of the rif-orf5 gene into the Δrif-orf5 mutant through electroporation. These strains were grown on YMG (yeast extract 4 g, malt extract 10 g, glucose 4 g, 20 g agar, ddH2O 1000 mL, pH 7.2) agar media at 28 °C for the production of rifamycins.
The Escherichia coli DH5α strain was used for plasmid propagation. Suicide vector pOJ260 was used for gene knock-out. Integrating vector pSET152 was used for gene complementation [18]. E. coli strains were maintained in LB (tryptone 10 g, yeast extract 5 g, NaCl 10 g, ddH2O 1000 mL, pH 7.2) media at 37 °C. Apramycin was added into media at a final concentration of 50 µg•mL −1 . Cells were stocked with 20% glycerol and stored at −80 °C.

Bacterial Strains, Plasmids and Culture Media
The Amycolatopsis mediterranei S699 strain, isolated in 1957 in St. Raphael, France [5], was stored in our lab. The A. mediterranei S699 ∆rif -orf5 strain was constructed by deleting the rif -orf5 gene through homologous recombination. The A. mediterranei S699 ∆rif-orf5::orf5 strain was constructed by transformation of the rif -orf5 gene into the ∆rif -orf5 mutant through electroporation. These strains were grown on YMG (yeast extract 4 g, malt extract 10 g, glucose 4 g, 20 g agar, ddH 2 O 1000 mL, pH 7.2) agar media at 28 • C for the production of rifamycins.
The Escherichia coli DH5α strain was used for plasmid propagation. Suicide vector pOJ260 was used for gene knock-out. Integrating vector pSET152 was used for gene complementation [18]. E. coli strains were maintained in LB (tryptone 10 g, yeast extract 5 g, NaCl 10 g, ddH 2 O 1000 mL, pH 7.2) media at 37 • C. Apramycin was added into media at a final concentration of 50 µg·mL −1 . Cells were stocked with 20% glycerol and stored at −80 • C. First, the rif -orf5 gene knock-out vector pOJ260-orf5 was constructed. Two ca. 2 kb DNA fragments flanking upstream and downstream of the target gene were amplified from the genomic DNA of A. mediterranei S699, and named HF1 and HF2, respectively. The purified homologous fragments HF1 and HF2 were digested with HindIII/XbaI and XbaI/EcoRI, and cloned into linearized HindIII/EcoRI digested pOJ260. The ligation product was transformed into DH5α-competent cells. Positive clones were verified by restriction enzyme digestion and sequencing ( Figures S1A and S2). The gene knock-out vector pOJ260-orf5 was introduced into the rifamycin-producing strain A. mediterranei S699 by electrotransformation [19]. Apramycin-resistant (AprR) colonies were selected and confirmed to be single cross-over mutants by PCR amplification (Figures S1B and S3A). Apramycin-sensitive (AprS) colonies were counterselected from the initial AprR single cross-over colonies after several rounds of nonselective growth, and confirmed to be double cross-over gene knock-out mutant ∆rif -orf5 by PCR amplification (Figures S1B and S3B).

2.2.2.
Construction of the rif -orf5 Gene Complementation Mutant ∆rif -orf5::orf5 First, the rif -orf5 gene complementation vector pSET152-orf5 was constructed. The targeted gene rif -orf5 was amplified using the genomic DNA of A. mediterranei S699 as a template. The purified PCR fragment was digested with NdeI and XbaI, and cloned into the downstream of the rifKp promoter in pSET152 through Gibson assembly [20]. Similarly, the assembled product was transformed into DH5α-competent cells, and positive clones were verified by restriction enzyme digestion and sequencing ( Figure S4). The gene complementation vector pSET152-orf5 was transformed into the rif -orf5 gene knock-out mutant ∆rif -orf5 by electroporation. Apramycin-sensitive (AprS) colonies were selected and confirmed to be the rif -orf5 gene complementation mutant ∆rif -orf5::orf5 by PCR amplification ( Figure S5).
Primers used in this study are shown in Table S1.

HPLC Detection of the Metabolites in Mutants
A. mediterranei S699 mutants were inoculated on YMG agar media (100 mL) and cultivated for 7 days at 28 • C. The culture was diced and extracted overnight with EtOAc at room temperature. The concentrated crude extract was dissolved in 1 mL MeOH, and analyzed by high-pressure liquid chromatography (HPLC; Agilent 1200, Santa Clara, CA, USA) in a gradient system consisting of ddH 2 O + 0.5% formic acid as solvent A and acetonitrile as solvent B. The program of solvent gradient was as follows: 20-35% B in the first 5 min, 35-55% B from 5 to 19 min, 55-65% B from 19 to 23 min, 65-100% B from 23 to 27 min. Flow rate was 1 mL/min, and UV detection was monitored at 254 nm ( Figure S6).

Fermentation, Extraction and Isolation of the Metabolites from the ∆rif -orf5 Strain
The fermentation (20 L) was performed on YMG agar Petri dishes for 7 d at 28 • C. The culture was diced and extracted overnight with EtOAc/MeOH (4:1, v/v) at room temperature three times. The crude extract was partitioned between H 2 O and EtOAc (1:1, v/v) until the H 2 O layer was colorless. The EtOAc extract was partitioned between 95% aqueous MeOH and petroleum ether (PE) to afford the defatted MeOH extract. The MeOH extract was fractionated by medium-pressure liquid chromatography (MPLC) over RP C 18 silica gel (130 g) eluted with gradient aqueous CH 3 CN (30%, 50%, 70% and 100% CH 3 CN, 500 mL each) to give Fr. A-J.

Antimicrobial Assay
Compounds 1-13 were assayed for their antimicrobial activity against Staphylococcus aureus ATCC 25923, Mycobacterium smegmatis mc 2 155, Pseudomonas aeruginosa PA01 and Proteusbacillus vulgaris CPCC 160013 with the paper disk diffusion assay as previously described [21]. The tested compounds (20 µg/µL, 2 µL each) were absorbed onto individual paper disks (Ø 6 mm) and placed on the surface of the agar. The assay plates were incubated for 24 h at 37 • C and examined for the presence of inhibitory zones.
The MIC values of active compounds against the growth of Staphylococcus aureus ATCC 25923 were measured through the microbroth dilution method [22]. Microorganisms were cultured in LB media in 96-well plates at a concentration of 1 × 10 6 CFU/mL. The MIC values were obtained after incubating for 12 h at 37 • C with the tested compounds (concentration ranging from 320 to 0.039 µg/mL).

Cytotoxicity Assay
The in vitro antiproliferative activity against HeLa and Caco-2 cells was measured as previously reported [23,24]. Briefly, cells were seeded in 96-well plates at 7 × 10 3 cells/well and treated for 24 h with different concentrations of compounds 1-13. Then, 10 µL Cell Counting Kit-8 (CCK-8) was added to each well and incubated for another 4 h. The absorbance was read at 480 nm by Spark 30086376 (TECAN, Männedorf, Switzerland).

Results
Compound 1 was determined to have the molecular formula C 36  . The presence of a naphthaquinone chromophore was indicated by the HMBC correlations from H-3 (δ H 7.64) to C-1 (δ C 184.7), C-2 (δ C 143.0) and C-10 (δ C 132.6), and from H-14 (δ H 2.09) to C-6 (δ C 165.9), C-7 (δ C 120.3) and C-8 (δ C 161.7). The MeO-6 (δ H 4.00) was supported by the HMBC correlations from MeO-6 to C-6 and NOE correlations from MeO-6 to H-5 (δ H 7.18) (Tables 1 and 2, Tables S2 Figure 2). The hydroxylation of C-34a and oxidization to an aldehyde group followed by hemiacetal formation with the hydroxyl group at C-25 were determined based on the 1 H NMR of H-34a (δ H 4.54/5.08) (Table 1, Figure 2). The ansa chain was determined to undergo retro-Claisen cleavage between C-5 and C-11 on the basis of the chemical shift of C-11 downfield, the presence of the extra aromatic proton H-5 compared to that of normal rifamycins and HMBC from H-5 to C-7 (Tables S2 and S3). Hence, the planar structure of 1 was established. The stereochemistry of the hemiacetal existed as a pair of epimers (1a and 1b) at C-34a, and 1a was determined to be α-form on the basis of the coupling constants J 34a,28 = 8.4 Hz and the NOE correlations from H-34a (δ H 4.54) to H-25 (δ H 3.55) and H-27 (δ H 3.18), and between H-25 and H-27. Accordingly, 1b was determined to be β-form on the basis of the coupling constants J 34a,28 = 3.2 Hz (Figure 2). The stereochemistry of other carbons was assumed to be the same as that of rifamycin W-hemiacetal [25] based on biosynthetic logic [12]. Thus, compound 1a was named 34a-α-6-O-methyl-rifamycin W-M1-hemiacetal and 1b was named 34a-β-6-O-methyl-rifamycin W-M1-hemiacetal. The NMR spectroscopic data of 2 were similar to that of 1, except that C-34 was a hydroxymethyl (H 4.33, 4.34, C 66.0) instead of a methyl group, and the 6-hydroxyl group was free (Tables 1 and 2). The relative configuration of 2 was proposed to be identical to  (Tables 1 and 2). The relative configuration of 2 was proposed to be identical to ) and NOESY (↔) correlations of 1.
Compound 2 was confirmed to have the molecular formula C 35 H 45 NO 13 on the basis of the HRESIMS quasi molecular ion peaks at m/z 688.2956 [M + H] + and 710.2781 [M + Na] + . The NMR spectroscopic data of 2 were similar to that of 1, except that C-34 was a hydroxymethyl (δ H 4.33, 4.34, δ C 66.0) instead of a methyl group, and the 6-hydroxyl group was free (Tables 1 and 2). The relative configuration of 2 was proposed to be identical to that of 1, and the hemiacetal existed as a pair of epimers (2a and 2b) (Tables S4  and S5) as well. Thus, compound 2 was determined to be 34a-α-30-hydroxyrifamycin W-M1-hemiacetal (2a) and 34a-β-30-hydroxyrifamycin W-M1-hemiacetal (2b), respectively.
The molecular formula of 5 was confirmed to be C 37 H 47 NO 12 by the HRESIMS quasi molecular ion peaks at m/z 698.3170 [M + H] + and 720.2986 [M + Na] + . A close NMR comparison with that of rifamycin W (12) (Tables S9 and S12) [27] revealed that 5 was 34a-O-acetyl-rifamycin W, which was confirmed by the HMBC correlations between H-34a (δ H 4.01, 4.00) and the acetyl carbon (δ C 172.9).
The molecular formula of 6 was elucidated as C 35  Na] + , revealing one more oxygen atom than that of rifamycin W. NMR comparison (Tables S11 and S12) determined 7 to be 20-hydroxyrifamycin W, which was supported by the chemical shift of C-30 (δ C 77.0).
Compounds 1−13 were assayed for their antimicrobial activity against Staphylococcus aureus ATCC 25923, Mycobacterium smegmatis mc 2 155, Pseudomonas aeruginosa PA01 and Proteusbacillus vulgaris CPCC 160013. The results showed that new compounds 1-3 and known compounds 11 and 13 exhibited inhibitory activity against S. aureus ATCC 25923, while other compounds showed no antimicrobial activity ( Figure S56). Thus, new compounds 1-3 were further tested for their antibacterial activity against S. aureus ATCC 25923 using the microbroth dilution method [22], and their MIC values were determined to be 5, 40 and 0.5 µg/mL, respectively (Table S14).
In view of no evident bactericidal activity, compounds 1-13 were evaluated for their antiproliferative activity against HeLa and Caco-2 cells using Cell Counting Kit-8 (CCK-8) (Bimake, Houston, TX, USA) and etoposide (VP-16) as a positive control. Compounds 1 and 3 showed modest activity in inhibiting the proliferation of HeLa and Caco-2 cells with IC 50 values of about 50 µM (Table S15, Figures S57 and S58).

Discussion
Post-PKS modifications play an important role in increasing the structural diversity and improving the biological activity of rifamycins. As the proposed earliest macrocyclic intermediate in rifamycin post-PKS biosynthesis, proansamycin X tended to undergo dehydration to form putative protorifamycins (without C-8 hydroxyl group) or undergo dehydrogenation to form rifamycin W [24,32,33]. Rifamycin W undergoes a rearrangement of the polyketide backbone to produce rifamycin B via the oxidative cleavage of the C-12/C-29 double bond. The mechanism of this oxidative cleavage has not been characterized yet. For the rif -orf5 gene, when cloned and heterologously expressed in E. coli, the recombinant protein showed spectra typical of P450 cytochromes [34]. Thus, the rif -orf5 gene was confirmed to code for a cytochrome P450 enzyme, which is the key step for oxygen incorporation in rifamycin B biosynthesis and may be involved in the cleavage of the olefinic moiety of rifamycin W [16,17].
In this study, systematic isolation of the fermentation products of the mutant strain ∆rif -orf5 afforded thirteen rifamycin W derivatives besides the main product rifamycin W (12), indicating that the rif -orf5 gene was probably involved in the oxidative cleavage of the C-12/C-29 double bond. Compounds 1-4 all undergo C-5/C-11 retro-Claisen cleavage, as observed in the biosynthesis of proansamycin B-M1 and protorifamycin I-M1 [24,35], hygrocins I and J [36], divergolides R and S [37] and microansamycins G-I [38]. This C-5/C-11 cleavage probably occurred due to an over-accumulation of rifamycin W, which serves as a detoxification mechanism. Compounds 1-3, 9 and 10 featured a hemiacetal, in which 9 and 10 existed in β-form according to the chemical shift of C-34a and the coupling constants between C-34a and C-28, while 1-3 existed as epimer pairs (Table S13), which may be due to the feasibility of polyketide chain cleavage in C-5/C-11. Additionally, the hemiacetal containing compounds 1-3, 9 and 10, as well as the lactone-containing rifamycin Z (8), indicated that the oxidation of C-34a alcohol to the carboxyl group may occur before the C-12/C-29 olefinic bond cleavage. In addition, compared to 8-deoxy rifamycins [24], compounds 5, 6, 7 and 11 also oxygenated at C-34a, C-23, C-20 and C-30, which suggested that the rifamycin ansa chain is prone to oxidization in these specific sites during fermentation (Figure 3).  Figure 3. Proposed biosynthetic pathway of compounds from mutant Δrif-orf5.

Conclusions
In this study, the cytochrome P450 monooxygenase gene rif-orf5 was confirmed to be involved in the oxidative cleavage of the ansa chain of rifamycin W through in vivo gene

Conclusions
In this study, the cytochrome P450 monooxygenase gene rif -orf5 was confirmed to be involved in the oxidative cleavage of the ansa chain of rifamycin W through in vivo gene inactivation and isolation of the main product rifamycin W. Systematic isolation of the fermentation products of the mutant strain ∆rif -orf5 afforded seven new rifamycin W congeners, from which 1-3 featured two epimeric forms of hemiacetal at C-34a, and C-5/C-11 retro-Claisen cleavage. Compounds 1-3 exhibited antibacterial activity against Staphylococcus aureus, and 1 and 3 showed modest antiproliferative activity against HeLa and Caco-2 cells.