English

• 1.   Introduction

  Rifamycins have been used for more than 50 years in clinic and are still mainstay in the treatment of tuberculosis, leprosy and AIDS-related mycobacterial infections. Biosynthetic study revealed that the backbone of rifamycins is constructed by type Ⅰ polyketide synthase (PKS) using 3-amino-5-hydroxy benzoic acid (AHBA) as starter unit and malonyl-CoA and methylmalonyl-CoA as extending units.^{[1, 2]} The macrolactam backbone further undergos 7, 8-dehydrogenation and 34a-hydroxylation to produce the important intermediate rifamycin W.^[3] The latter is oxidized by cytochrome P450 Orf5 on Δ29, 12 olefinic bond, ^[4] followed by 25-O-acetylation catalyzed by Orf20^[5] and 27-O-methylation catalyzed by Orf14, ^[6] to produce rifamycin S, which is converted into rifamycin SV through reduction on naphtoquinone and finally into end product rifamycin B under the catalysis of transketolase Orf15 and cytochrome P450 enzyme Orf16.^{[7, 8]}

  Several steps in the post-PKS modification of rifamycins are still ambiguous after decades of study because of the complicated reaction route and redundant biosynthetic gene cluster. In the maturing of rifamycins, oxidation on C34a followed by oxidative cleavage of C12-C29 double bond is essential step to produce the intermediate harboring ketal moiety on C12, which ensures the proper molecular conformation of rifamycins to effectively bind to RNA polymerase of bacteria via the formation of hydrogen bonds.^[9] Cytochrome P450 encoded by rif-orf5 was proved to participate in the oxidative conversion of Δ29, 12 olefinic bond through in vivo experiments, ^[4] and Rif-Orf11 and Orf17 (both are flavin dependent oxidoreductases) were suspected to be responsible for the hydroxylation on C34a.^[10] However, no experimental data have been published to prove this hypothesis till now. Yoon et al.^[11] proposed that Orf0 catalyzed the hydroxylation of C34a in proansamycin X based on the accumulation of proansamycin X (Scheme 1) in orf0 deletion mutant, but the same result was not accomplished by Moor et al.^[4] or ourselves (unpublished data). Here, we demonstrate that the cytochrome P450 encoded by rif-orf13 is responsible for the hydroxylation of 34a-deoxyrifamycin W to produce rifamycin W based on intermediate feeding experiments and in vitro enzymatic assays.

  Scheme 1

  

  Scheme 1.  Biosynthetic pathway of rifamycins

  Enzymes labeled with question mark are proposed to catalyze the biosynthetic steps based on functional prediction

  下载: 全尺寸图片 幻灯片

  2.   Results

  Recently, we discovered several rifamycin analogs, including compounds 1, 2 and 3 (Scheme 1), from Micromonospora sp. TP-A0468.^[12] These compounds are structurally identical to the known rifamycins except for the absence of the methyl substitute on C16. Homologous genes of orf11 and orf17 are missing in the biosynthetic gene cluster of 16-demethylrifamycins, while the 34a-hydroxylated intermediate 2 is produced. We inferred that a cytochrome P450 instead of Orf11/Orf17 probably catalyzed the hydroxylation of C34a.

  Functionally unidentified cytochrome P450 encoding genes, rif-orf0, rif-orf4 and rif-orf13, were cloned from another rifamycin producer, Amycolatopsis (A.) tolypomycina, because the original producer of rifamycins, A. mediterranei S699, was not available in our laboratory. These genes displayed extremely high similarity to their counterparts from A. mediterranei S699 (Table S3). Rif-orf0, rif-orf4 and rif-orf13 were cloned into shuttle plasmid pSETe under the control of constitutive promoter ermE*, and the recombinant plasmids were transferred into Streptomyces (S.) lividans 1326 through conjugation between S. lividans and E. coli S17-1. Verified conjugants were cultured in ISP-2 medium and added with compound 1, 2, 3 or 4 during the cultivation. The cultures were extracted and subject to HPLC and LC-MS analysis with the standard controls of 1, 2, 3 or 4. Recombinant strain containing blank plasmid pSETe was cultured and analyzed in the same way as negative control.

  HPLC analysis showed that most of compound 1 remained in the extract of S. lividans 1326 harboring pSETe (Figure 1, Ⅲ). But in the strain containing rif-orf13, compound 1 added into the culture broth was completely disappeared, and a new compound arose in the extract which exhibited identical retention time and mass spectrum with 2 on HPLC and LC-MS analysis (Figure 1, Ⅳ). Strains containing orf0 or orf4 did not convert the additive compound 1 into any other detectable product. When compound 2, 3 or 4 was used as substrate, no product was detected in the culture of the recombinant strain containing orf13. These results demonstrate that rif-orf13 is responsible for the hydroxylation on C34a of compound 1 while orf0 and orf4 are irrelevant to this step.

  Figure 1

  

  Figure 1.  Heterologous expression and feeding experiments of rif-orf13

  The cytochrome encoding gene was heterologously expressed in Streptomyces lividans 1326 and cultured in ISP-2 medium with purified compound 1 added. Extracts of feeding experiments were analyzed at 423 nm on HPLC. Sl, Streptomyces lividans 1326; Sl/-, Sl containing plasmid pSETe, used as negative control; Sl/orf13, Sl containing recombinant plasmid pAT13he. (Ⅰ) standard compound 1; (Ⅱ) standard compound 2; (Ⅲ) Sl/-fed with 1 during cultivation; (Ⅳ) Sl/orf13 fed with 1 during cultivation. , unknown metabolite (m/z 362) produced by Streptomyces lividans 1326

  下载: 全尺寸图片 幻灯片

  To further characterize the function of Orf13, we cloned rif-orf13 into pET28a and expressed the gene in E. coli BL21(DE3). Soluble His₆-tagged protein was prepared through induced expression of rif-orf13 in E. coli followed by Nickel affinity chromatography and dialysis. A small aliquot of the characteristic dark red solution of Orf13 was analyzed with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), in which a homogenous band at about 49 kDa (expected 49.1 kDa) was detected, suggesting that Orf13 was successfully prepared. In reduced CO-bound difference spectrum, the absorption peak of Orf13 displayed obvious shift from 420 nm of untreated sample to 450 nm of CO treated sample, confirming that the cytochrome P450 Orf13 was correctly functionalized with active cofactor heme.^[13]

  Enzymatic assays were carried out using 1 as substrate and Orf13 as catalyst. Since no dedicated ferredoxin and ferredoxin reductase was found encoded by rifamycin biosynthetic gene cluster, the partner proteins from Spinacia oleracea were used in the in vitro assays of Orf13 to deliver electrons from NADPH to the cytochrome P450.^[14] After 2 h of reaction, a tiny amount of 2 was detected in HPLC analysis (Figure 2A, Ⅱ), which was further verified by LC-MS analysis with standard 2 as control (Figure 2B). Extending the reaction course to 17 h (Orf13 was thereafter unstable and denatured) led to slight increase of the production of 2 (Figure 2A, Ⅰ), while no product 2 was detected in the reactions minus Orf13 or ferredoxin/reductase or NADPH (Figure 2A, Ⅲ~Ⅴ). These assays confirmed that Orf13 catalyzed the hydroxylation of 1 to produce 2 with the assist of ferredoxin/ferredoxin reductase and NADPH, despite the extremely low efficiency in the in vitro assays.

  Figure 2

  

  Figure 2.  Functional verification of rif-orf13 through in vitro enzymatic assays

  (A) HPLC analysis of the enzymatic assays of Orf13; (Ⅰ) complete reaction mixture incubated in 30 ℃ for 17 h before analysis; (Ⅱ) complete reaction mixture incubated in 30 ℃ for 2 h before analysis; (Ⅲ) control assay that lack Orf13 in the reaction; (Ⅳ) control assay that lack ferredoxin in the reaction; (Ⅴ) control assay that lack NADPH in the reaction; (B) LC-MS analysis of the product in the enzymatic assay of Orf13

  下载: 全尺寸图片 幻灯片

  To enhance the reaction rate and reduce the formation of side products, self-sufficient ferredoxin RhFRED and RhFRED-Fdx^{[15, 16]} were employed to assist Orf13 in the enzymatic hydroxylation of 1, but no improved efficiency was observed (data not shown).

  3.   Discussion

  It's quite unexpected to find Orf13 catalyzing the hydroxylation of 1 because former research pointed that the disruption orf13 in A. mediterranei S699 had no effect on the production of rifamycin B, ^[5] and gene inactivation and complementation experiments suggested that rif-orf0 likely participated in the hydroxylation on C34a of proansamycin X.^[11] Furthermore, there is only one cytochrome P450, Sare_1259 (homologue of Orf5), is involved in the conversion of 34a-deoxy rifamycin W to desacetyl-demethylri- famycin S in Salinispora arenicola CNS-205, ^[4] thus ruled out a dedicated cytochrome P450 monooxygenase responsible for the hydroxylation of C34a in the biosynthesis of rifamycins in S. arenicola CNS-205. A reasonable proposal is that the function of Orf13 could be undertook by the homologue Orf5 in S. arenicola CNS-205, and Orf5 further catalyzes the oxidative elimination of C34a and cleavage of C12, C29 double bond (Scheme 2). Based on this hypothesis, it is logical that rifamycin W is accumulated in orf5 inactivated mutant and rifamycin B is normally produced in orf13 disrupted mutant of A. mediterranei S699.

  Scheme 2

  

  Scheme 2.  Proposed function of Orf13 in the biosynthesis of rifamycins

  下载: 全尺寸图片 幻灯片

  Phylogenetic analysis of cytochrome P450s from different rifamycin biosynthetic gene clusters showed that Orf13 displayed high homology to Orf5 (67% identity, 81% similarity), and they formed a separated clad in the phylogenetic tree. Given the fact that rifamycin W is accumulated in orf13-missing strains (S. arenicola CNS-205 and Micromonospora sp. TP-A0468)^{[4, 12]} but not in orf13-harboring strains (A. mediterranei and A. tolypomycina), ^{[5, 17]} it could be deduced that orf13 is a reduplicate copy evolved from orf5 to promote the biosynthetic rate of rifamycins. Our work therefore clarifies the function of rif-orf13 in the biosynthesis of rifamycin and explains the contradictory results of previous studies.

  It must be pointed out that the catalytic efficiencies of Rif-Orf13 in vivo and in vitro were dramatically different from each other. In the complex of cytochrome P450, ferredoxin and ferredoxin reductase, cytochrome P450 determines the substrate specificity and the reaction mode while the partner proteins account for the reaction rate.^[14] Generally, only a few partner proteins are encoded in a genome, while plenty of P450 enzymes maybe exist. We failed to identify the authentic ferredoxin/ferredoxin reductase of Orf13 but used heterologous partners from Spinacia oleracea to ensure the supplying of electrons, which probably resulted in the low efficiency of hydroxylation in the enzymatic assays. However, compound 1, the proposed substrate of Rif-Orf13, was totally converted into the product in vivo. It could be proposed that ferredoxin and ferredoxin reductase possessing better compatibility with Rif-Orf13 were harbored in Streptomyces lividans. Meanwhile, higher stability of Rif-Orf13 and smooth supply of NADPH were probably more available in the intracellular environment of Streptomyces lividans, and thus higher catalytic efficiency was acquired in vivo.

  4.   Experimental section

  Genomic sequence of A. tolypomycina DSM 44544 was downloaded from genbank (Accession No. FNSO0000-0000.1), and the homologues of rif-orf0, rif-orf4 and rif-orf13 were identified as gene SAMN04489727_0263, SAMN04489727_0292 and SAMN04489727_0305, respectively, using Basic Local Alignment Search Tool (BLAST, https://blast.ncbi.nlm.nih.gov/Blast.cgi). To heterologously express these genes, we constructed pSET152 derivate plasmid pSETe which contains constitutive promoter ermE^* in the multiple cloning site. Target genes were amplified from genomic DNA of A. tolypomycina through polymerase chain reaction (PCR) with specific primers of AT-orf0-F/AT-orf0-R, AT-orf4-F/AT-orf4-R or AT-orf13-F/AT-orf13-R, and then digested and cloned into NdeI/XbaI site of pSETe to give recombinant plasmids pAT0he, pAT4he and pAT13he. The recombinant plasmids were verified by DNA sequencing and then transferred into E. coli S17-1 through transformation, and finally into S. lividans 1326 via conjugation between E. coli S17-1 and S. lividans. Conjugants harboring pAT0he, pAT4he or pAT13he were identified by the resistance to apramycin. Parent plasmid pSETe was transferred into S. lividans 1326 to make blank control (Sl/-) in the following substrate feeding experiments.

  Recombinant strains Sl/orf0, Sl/orf4, Sl/orf13 and negative control Sl/- were cultured in 50 mL of ISP-2 broth containing 50 μg/mL apramycin in 250 mL flask at 30 ℃ and 220 round per minute. After 2 days of cultivation, the broths were added with 1 mg of compound 1, 2, 3 or 4 (in 100 μL methanol), respectively, and then continued to cultivate for 24 h under the same condition. Next, the cultures were individually extracted with 50 mL of ethyl acetate twice. The organic extracts were combined and evaporated to dry to yield yellow pastes, which were dissolved in 1 mL of methanol and subject to HPLC and LC-MS analysis.

  HPLC analysis was carried out on an Agilent Technology 1260 infinity with Dickma Diamonsil C18 column (5 μm, 4.6 mm×250 mm), using H₂O added with 0.1% formic acid (solvent A) and CH₃CN added with 0.1% formic acid (solvent B) as mobile phase. The samples were washed with a flow of 1 mL/min under the gradient elution program as follow: 0 min, 24% solvent B; 24 min, 60% solvent B, 28 min, 80% solvent B; 30 min, 95% solvent B; 33 min, 95% solvent B and 35 min, 24% solvent B. LC-MS analysis of in vitro assays was carried out on a Thermo Scientific LTQ XL Mass spectrometer with Zorbax Eclipse XDB-C18 column (5 μm, 4.6 mm×150 mm), and the same mobile phase and flow rate as above but using gradient elution program as follow: 0 min, 40% solvent B; 4.5 min, 50% solvent B; 5 min, 70% solvent B; 9 min, 95% solvent B; 10 min, 40% solvent B and 11 min, 40% solvent B.

  To characterize the in vitro function of Orf13, rif-orf13 was amplified from genomic DNA of A. tolypomycina with specific primers AT-13e-F/AT-13e-R, and then cloned into the NdeI/HindⅢ site of pET28a to give recombinant plasmid pAT13p. This plasmid was verified by DNA sequencing and finally transferred into heterologous expression host E. coli BL21 (DE3).

  E. coli BL21 (DE3)/pAT13p was cultivated in 700 mL of LB broth in 2.5 L flask with 50 μg/mL kanamycin added at 37 ℃ and 250 round per minute. The temperature was adjusted to 16 ℃ when the optical density at 600 nm (OD₆₀₀) arrived at 0.5, and 30 min later final concentration of 100 μmol/L IPTG, 50 mg/L ALA and 200 mg/L NH₄FeSO₄ were added into the culture broth to induce the production of Orf13. The induced culture was precipitated after 20 h to harvest cells. The sediment was washed and then suspended in 40 mL of phosphate buffer (50 mmol/L NaH₂PO₄, 500 mmol/L NaCl and 10 mmol/L imidazole, pH 8.0). The suspension was subject to digestion with 1 mg/mL lysozyme for 1 h and then ultrasonication in ice-bath condition to thoroughly smash the cells. After centrifugation, the lysate supernatant was mixed with nickel affinity beads to adsorb target his-tagged Orf13, then filtered to gather beads and washed with phosphate buffer containing imidazole gradient from 50 mmol/L to 500 mmol/L. The eluents containing pure target protein were combined and dialyzed against tris-HCl buffer (50 mmol/L tris-HCl, 50 mmol/L NaCl, 10% glycerol, pH 8.0) to yield Orf13 product.

  Purified Orf13 was subject to reduced CO-bound difference spectrum measurement to check whether the cofactor heme was properly loaded in the active center. The protein solution was bubbled with CO for 1 min in the presence of 1 mmol/L sodium hyposulfite and the absorption spectra of treated and untreated protein were measured with a Thermo Scientific NanoDrop 2000c spectrophotometer.

  The enzymatic activity of Orf13 was tested in 50 μL of solution containing 50 mmol/L sodium phosphate buffer (pH 6.5), 1 mmol/L 1, 3 mmol/L NADPH, 50 μmol/L Orf13 and 10 μmol/L ferredoxins and ferredoxins reductases at 30 ℃ for 2~17 h. The reaction was quenched by adding 100 μL of methanol and vortex. The product was analyzed on HPLC and LC-MS after removing the denatured protein by centrifugation.

  Supporting Information  Strains, plasmids and primers used in this study, more feeding experiments, construction of recombinant plasmids and bioinformatic analysis of genes are available free of charge via the Internet at http://sioc-journal.cn/.

• 1. [1]

     Schupp, T.; Toupet, C.; Engel, N.; Goff, S. FEMS Microbiol. Lett. 1998, 159, 201. doi: 10.1111/fml.1998.159.issue-2

  2. [2]

     August, P. R.; Tang, L.; Yoon, Y. J.; Ning, S.; Muller, R.; Yu, T. W.; Taylor, M.; Hoffmann, D.; Kim, C. G.; Zhang, X.; Hutchinson, C. R.; Floss, H. G. Chem. Biol. 1998, 5, 69. doi: 10.1016/S1074-5521(98)90141-7

  3. [3]

     Yu, T. W.; Shen, Y.; Doi-Katayama, Y.; Tang, L.; Park, C.; Moore, B. S.; Hutchinson, C. R.; Floss, H. G. Proc. Natl. Acad Sci. U. S. A. 1999, 96, 9051. doi: 10.1073/pnas.96.16.9051

  4. [4]

     Wilson, M. C.; Gulder, T. A.; Mahmud, T.; Moore, B. S. J. Am. Chem. Soc. 2010, 132, 12757. doi: 10.1021/ja105891a

  5. [5]

     Xu, J.; Wan, E.; Kim, C. J.; Floss, H. G.; Mahmud, T. Microbiology 2005, 151, 2515. doi: 10.1099/mic.0.28138-0

  6. [6]

     Xu, J.; Mahmud, T.; Floss, H. G. Arch. Biochem. Biophys. 2003, 411, 277. doi: 10.1016/S0003-9861(03)00004-3

  7. [7]

     Yuan, H.; Zhao, W.; Zhong, Y.; Wang, J.; Qin, Z.; Ding, X.; Zhao, G. P. Acta Biochim. Biophys. Sin. (Shanghai) 2011, 43, 948.

  8. [8]

     Qi, F.; Lei, C.; Li, F.; Zhang, X.; Wang, J.; Zhang, W.; Fan, Z.; Li, W.; Tang, G. L.; Xiao, Y.; Zhao, G.; Li, S. Nat. Commun. 2018, 9, 2342. doi: 10.1038/s41467-018-04772-x

  9. [9]

     Campbell, E. A.; Korzheva, N.; Mustaev, A.; Murakami, K.; Nair, S.; Goldfarb, A.; Darst, S. A. Cell 2001, 104, 901. doi: 10.1016/S0092-8674(01)00286-0

  10. [10]

      Floss, H. G.; Yu, T. W. Chem. Rev. 2005, 105, 621. doi: 10.1021/cr030112j

  11. [11]

      Lee, S. K.; Choi, C. Y.; Ahn, J. S.; Cho, J. Y.; Park, C. S.; Yoon, Y. J. J. Microbiol. Biotechnol. 2004, 14, 356.

  12. [12]

      Zhou, Q.; Luo, G.-C.; Zhang, H.; Tang, G.-L. Appl. Environ. Microbiol. 2018, DOI: 10.1128/AEM.02597-18.

  13. [13]

      Omura, T.; Sato, R. J. Biol. Chem. 1964, 239, 2370.

  14. [14]

      Cook, D. J.; Finnigan, J. D.; Cook, K.; Black, G. W.; Charnock, S. J. Adv. Protein Chem. Struct. Biol. 2016, 105, 105. doi: 10.1016/bs.apcsb.2016.07.003

  15. [15]

      Li, S.; Podust, L. M.; Sherman, D. H. J. Am. Chem. Soc. 2007, 129, 12940. doi: 10.1021/ja075842d

  16. [16]

      Zhang, W.; Liu, Y.; Yan, J.; Cao, S.; Bai, F.; Yang, Y.; Huang, S.; Yao, L.; Anzai, Y.; Kato, F.; Podust, L. M.; Sherman, D. H.; Li, S. J. Am. Chem. Soc. 2014, 136, 3640. doi: 10.1021/ja4130302

  17. [17]

      Kishi, T.; Yamana, H.; Muroi, M.; Harada, S.; Asai, M. J. Antibiot. (Tokyo) 1972, 25, 11.

• 

  Scheme 1  Biosynthetic pathway of rifamycins

  Enzymes labeled with question mark are proposed to catalyze the biosynthetic steps based on functional prediction

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Heterologous expression and feeding experiments of rif-orf13

  The cytochrome encoding gene was heterologously expressed in Streptomyces lividans 1326 and cultured in ISP-2 medium with purified compound 1 added. Extracts of feeding experiments were analyzed at 423 nm on HPLC. Sl, Streptomyces lividans 1326; Sl/-, Sl containing plasmid pSETe, used as negative control; Sl/orf13, Sl containing recombinant plasmid pAT13he. (Ⅰ) standard compound 1; (Ⅱ) standard compound 2; (Ⅲ) Sl/-fed with 1 during cultivation; (Ⅳ) Sl/orf13 fed with 1 during cultivation. , unknown metabolite (m/z 362) produced by Streptomyces lividans 1326

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Functional verification of rif-orf13 through in vitro enzymatic assays

  (A) HPLC analysis of the enzymatic assays of Orf13; (Ⅰ) complete reaction mixture incubated in 30 ℃ for 17 h before analysis; (Ⅱ) complete reaction mixture incubated in 30 ℃ for 2 h before analysis; (Ⅲ) control assay that lack Orf13 in the reaction; (Ⅳ) control assay that lack ferredoxin in the reaction; (Ⅴ) control assay that lack NADPH in the reaction; (B) LC-MS analysis of the product in the enzymatic assay of Orf13

  下载: 全尺寸图片 幻灯片
  

  Scheme 2  Proposed function of Orf13 in the biosynthesis of rifamycins

  下载: 全尺寸图片 幻灯片
•