Abstract
Pradimicins are antifungal and antiviral natural products from Actinomadura hibisca P157-2. The sugar moieties play a critical role in the biological activities of these compounds. There are two glycosyltransferase genes in the pradimicin biosynthetic gene cluster, pdmS and pdmQ, which are putatively responsible for the introduction of the sugar moieties during pradimicin biosynthesis. In this study, we disrupted these two genes using a double crossover approach. Disruption of pdmS led to the production of pradimicinone I, the aglycon of pradimicin A, which confirmed that PdmS is the O-glycosyltransferase responsible for the first glycosylation step and attaching the 4',6'-dideoxy-4'-amino-D-galactose or 4',6'-dideoxy-4'-methylamino-D-galactose moiety to the 5-OH. Disruption of pdmQ resulted in the production of pradimicin B, indicating that this enzyme is the second glycosyltransferase that introduces the D-xylose moiety to the 3'- OH of the first sugar moiety. Insertion of an integrative plasmid before pdmO might have interfered with the dedicated promoter, yielding a mutant that produces pradimicin C as the major metabolite, which suggested that PdmO is the enzyme that specifically methylates the 4'- NH2 of the 4',6'-dideoxy-4'-amino-D-galactose moiety. Functional characterization of these sugar-decorating and –incorporating enzymes thus facilitates the understanding of the pradimicin biosynthetic pathway.
Keywords: Pradimicin A, Glycosyltransferase, Methyltransferase, Gene disruption, Polyketide
Pradimicins A-C (1-3, Figure 1) are aromatic polyketide natural products from the soil bacterium Actinomadura hibisica P157-2. Since their discovery in 1988, these molecules have been intensively studied. 1 is a promising lead compound due to its combined antifungal/antiviral properties. It was found to be active against a broad-spectrum of opportunistic and pathogenic fungi. This compound also interferes with the recognition of HIV-1 to its target cells. The mechanism of action of 1 emphasizes its lectin-like property in the presence of Ca2+.1 The moieties of 1 form a primary cavity with C-14 and C-15 from the benzo[α]naphthacenequinone and several hydroxyl groups of D-mannopyranoside.2 Based on the intermolecular distance in the proposed model, it is believed that the free carboxyl group at C- 18 of two molecules of 1 interacts with one Ca2+ ion.2 Another study on the anticandidal mode of action was done with the semisynthetic pradimicin derivative, BMY-28864.3 It was concluded that the sugar moieties of pradimicins, especially thomosamine or 4',6'-dideoxy-4'-methylamino-D-galactose, were critical for sugar-recognition and involved in binding to the specific mannan.
Figure 1.
The structures of pradimicins (1-3) and their aglycon (4).
The pradimicin (pdm) biosynthetic gene cluster has previously been reported (NCBI accession number: EF151801.1). This reported gene cluster consists of 28 open reading frames.1 Through combinatorial biosynthesis and heterologous expression, we have identified the enzymes that synthesize the pentangular intermediate G-2A, including PdmA, B, C, D, G, H, K, and L.2 Two cytochrome P450 enzymes, PdmJ and PdmW, were functionally characterized as the dedicated hydroxylases that specifically introduce a hydroxyl group to C-5 and C-6, respectively.3 We have also identified PdmN as an amino acid ligase that induces a D-alanine moiety at C-16. This enzyme was found to possess broad substrate and donor specificity.3, 4 Synergistic actions of tailoring enzymes in pradimicin biosynthesis commonly exit to ensure an efficient assembly process. In this work, we show the functional characterization of three late tailoring enzymes, including PdmS, PdmQ and PdmO, using a homologous gene recombination approach. PdmS and PdmQ were characterized as the glycosyltransferases (GTs) that successively introduce the 4',6'-dideoxy-4'-methylamino-Dgalactose/ 4',6'-dideoxy-4'-amino-D-galactose and D-xylose moieties to the structure, while PdmO is responsible for the N-methylation of 4',6'-dideoxy-4'-amino-D-galactose at 4'-NH2.
As in the pradimicin family, many other important antibiotics such as erythromycin naturally contain deoxysugar moieties attached to the aglycon. Because these sugar moieties often play a critical role in the bioactivities of these natural products, it is of great interest to identify the GTs responsible for transferring these deoxysugars to the aglycons. The pdm biosynthetic gene cluster contains two putative GT genes, pdmS and pdmQ. Sequence analysis indicated that they belong to a family of GTs that catalyzes the final stage of the biosynthesis of glycosylated natural products such as vancomycin and chloroeremomycin.
BLAST analysis revealed that PdmS (GenBank accession number ABK58688) is a putative GT that contains 437 amino acids (aa). Two proteins with the highest homology include two proposed GTs from Paenibacillus dendritiformis (GenBank accession number WP_006678995, 423 aa, 61% identity) and Bacillus cereus (GenBank accession number WP_002061747, 416 aa, 59% identity). However, none of these are functionally characterized. To understand the role of PdmS in pradimicin biosynthesis, we designed two sets of primers to inactivate this gene using a double crossover approach. A 1482-bp left arm and a 1404-bp right arm were cloned from the genome of A. hibisca P157-2 (Fig. 2A). These two fragments were ligated to the thermal sensitive plasmid pKC1139 between HindIII and XbaI as well as XbaI and EcoRI, respectively, to yield a disruption plasmid pKN82. The primers used and plasmids constructed in this work are shown in Tables S1 and S2, respectively. The intergeneric conjugation between E. coli and A. hibisca has been described by Kiser et al.5 and pKN82 was introduced into A. hibisca P157-2 through a similar way.6 Correct transformants of A. hibisca/pKN82 were cultivated in YM broth (4 g L−1 Dglucose, 4 g L−1 yeast extract, and 10 g L−1 malt extract) in the presence of 50 mg L−1 apramycin at 28°C in a rotary shaker, and then cultivated on ISP4 plates with 50 mg L−1 apramycin at 37°C for a period of 10 days to allow the plasmid to integrate into the genome. Conjugants were picked to YM broth without any antibiotics and incubated for 10 days at 28°C. The cultures were then spread onto ISP4 plates without antibiotics and growth for about 10 days at 28°C. The selection process for positive double crossover mutants was performed using a replica plate technique for colonies sensitive to 50 mg L−1 apramycin. Positive colonies were cultivated in YM broth in a rotary shaker at 28°C and were subjected to PCR screening using a set of pdmS-specific primers. As shown in Figure 2B (panel i), a 1.3-kb fragment of pdmS can be amplified from the genome of the wild type, while a disrupted gene pdmS* with a length of 0.9 kb was cloned from the correct double crossover mutant. This confirmed that a 0.4-bp fragment of pdmS has been successfully deleted from the genome of A. hibisca P157-2.
Figure 1.
Disruption of pdmS, pdmQ and pdmO in A.hibisca. (A) The single and double crossover strategies used for different genes. (B) Verification of the correct mutants by PCR. (i) Amplification of pdmS from the wild type (1.3 kb) and mutant (0.9 kb). M: marker; 1: wild type; 2: A. hibisca/pKN82 double crossover mutant. (ii) Amplification of pdmQ from the wild type (1.2 kb) and mutant strain (1.0 kb). M: marker; 1: wild type; 2: A. hibisca/pKN88 double crossover mutant. (iii) Verification of the insertion of pKN99 into the genome of A. hibisca. M: marker; 1: primer 1/RV-M PCR product from the A. hibisca/pKN99 single crossover mutant; 2: primer 2/M13-47 PCR product from the A. hibisca/pKN99 single crossover mutant; 3: primer 1/primer 2 PCR product from the A. hibisca/pKN99 single crossover mutant; 4: primer 1/primer 2 PCR product from the wild type.
The ΔPdmS mutant of A. hibisca P157-2 was then grown in YM medium for product analysis. The fermentation broth was centrifuged to separate the cells and supernatant, and the latter was injected into LC-MS for analysis. As shown in Figure 3 (trace i), the wild type strain produces 1 as the major metabolite, with 2 and 3 as minor products depending on the culture time. In contrast, the ΔPdmS mutant did not produce 1-3. Instead, a new peak 4 was produced as a dominant product at 44 min (Fig. 3, trace ii). The UV spectrum 4 is similar to that of 1 (Fig. S1), indicating that they have the same chromophore. Its molecular weight was found to be 549, according to the ion peaks [M-H]- at m/z 548 and [M+H]+ at m/z 550.1 in the ESI-MS spectra (Fig. S2). This suggested that 4 is the pradimicin aglycon that has no sugar moieties. We then purified 4 7 and recorded its NMR spectra. The 1H and 13C NMR spectra indicated that there are only signals of the pradimicin aglycon in 4, confirming that it is pradimicinone I. The 1H and 13C NMR signals were assigned based on the 2D NMR spectra including HSQC, HMBC (Fig. S3) and ROESY (Fig. S3), and are shown in Table 1. These data were identical with those of reported for pradimicinone I.8 Production of 1-3 by A. hibisca revealed that pradimicins with one or two sugar moieties are naturally synthesized, suggesting that the two sugar moieties are successively introduced. Identification of the product of the ΔPdmS mutant revealed the function of PdmS as the O-GT that is responsible for the introduction of the first sugar moiety, 4',6'-dideoxy-4'-amino-D-galactose or 4',6'-dideoxy-4'-methylamino-D-galactose.
Figure 3.
HPLC analysis of the products of A. hibisca (i), A. hibisca ΔPdmS (ii), A. hibisca ΔPdmQ (iii), and A. hibisca/pKN99 single crossover mutant (iv) at 460 nm. Samples were eluted with 20–35% acetonitrile-water (containing 0.1% trifluoroacetic acid) over 40 min at 1 mL min−1.
Table 1.
1H (300 MHz) and 13C (75 MHz) NMR data for 4 (in DMSO-d6).
| No. | δH | δC |
|---|---|---|
| 1 | - | 150.9 |
| 2 | - | 127.1 |
| 3 | - | 137.4 |
| 4 | 7.04 (1H, s) | 118.0 |
| 4a | - | 140.9 |
| 5 | 4.22 (1H, brs) | 71.5 |
| 6 | 4.22 (1H, brs) | 72.4 |
| 6a | - | 149.9 |
| 7 | 8.07 (1H, s) | 115.6 |
| 7a | - | 131.2 |
| 8 | - | 187.5 |
| 8a | - | 110.1 |
| 9 | - | 164.8 |
| 10 | 6.87 (1H, brs) | 106.9 |
| 11 | - | 166.0 |
| 12 | 7.21 (1H, brs) | 107.6 |
| 12a | - | 134.3 |
| 13 | - | 185.2 |
| 13a | - | 115.4 |
| 14 | - | 156.8 |
| 14a | - | 125.9 |
| 14b | - | 113.7 |
| 15 | 2.30 (1H, s) | 19.2 |
| 16 | - | 167.2 |
| 17 | 4.37 (1H, m) | 47.7 |
| 18 | - | 174.2 |
| 19 | 1.30 (1H, d, J = 6.9 Hz) | 16.9 |
| 11-OMe | 3.89 (3H, s) | 56.5 |
| 9-OH | 12.87 (1H, s) | - |
| 16-NH | 8.50 (1H, d, J = 6.5 Hz) | - |
We next sought to identify the O-GT that transfers the D-xylose moiety. PdmQ (GenBank accession number ABK58687) is the second putative GT in the pdm gene cluster that contains 435 aa. BLAST analysis showed that it is homologous to many GTs that are proposed to catalyze the transfer of sugar moieties from activated donor molecules to specific aglycons. The closest homologs found in GenBank are two GTs from Spirillospora albida (GenBank accession number WP_030163570, 414 aa, 61% identity) and Thermomonospora curvata (GenBank accession number WP_012855060, 418 aa, 62% identity), respectively. Again, none of these homologs are identified. Using a similar double crossover approach (Fig. 2A), we constructed a pKC1139-derived disruption plasmid pKN88 (Table S2) after the amplification of a 1224-bp left arm and a 1445-bp right arm. With this plasmid, we obtained a ΔPdmQ mutant of A. hibisca P157-2. As shown in Figure 2B (panel ii), the intact pdmQ gene is 1.3 kb, while its mutant pdmQ* is 1.0 kb, confirming that a 0.3-kb fragment has been deleted from pdmQ. This mutant was grown in YM medium and the fermentation broth was analyzed by LC-MS. The double crossover mutant of A. hibisca/pKN88 did not produce 1 and 3, but a dominant product with the same retention time as 2 (trace iii of Fig. 3). Further analysis of the UV absorption spectrum (Fig. S1) and MS (Fig. S2) confirmed that this product is indeed pradimicin B (2). Thus, disruption of PdmQ led to the production of a major product with only the 4',6'-dideoxy-4'-methylamino-D-galactose moiety, which confirmed the function of this enzyme as the O-GT responsible for transferring the second sugar moiety, D-xylose. 4 was observed as a minor product in this mutant, likely due to the incomplete glycosylation of this aglycon by PdmS.
The structure of 1 contains a 4',6'-dideoxy-4'-methylamino-D-galactose moiety, while that of 3 does not have the N-CH3 group. A methyltransferase (MT) must be responsible for the methylation of this sugar moiety at the specific 4'-NH2 group. There are three putative MTs in the pdm gene cluster, including pdmF, pdmT and pdmO. BLAST analysis suggested that PdmF and PdmT are two O-MTs, while PdmO is a putative N-MT. PdmO (GenBank accession number ABM21743.1) contains 238 aa. It is homologous to a number of class I MTs that are dependent of S-adenosylmethionine such as the sugar N, N-dimethyltransferases DesVI from Streptomyces venezuelae (67% similarity and 53% identity)9 and TylM1 from Streptomyces fradiae (66% similarity and 52% identity).10 Both enzymes are involved in the biosynthesis of aminodeoxysugars. Thus, PdmO is a candidate enzyme that catalyzes the specific N-methylation of 4',6'-dideoxy-4'-amino-D-galactose.
To determine the function of PdmO, we constructed an integrative plasmid pKN99 (Table S2) from the pKC1139 vector and a 1.3-kb insert that contains the C-terminal portion of pdmO and N-terminal portion of pdmS (Fig. 2A). The plasmid pKN99 was introduced into A. hibisca through E. coli ET12567 (pUZ8002) intergeneric conjugation, as described for the disruption of pdmS. The conjugants were selected on ISP4 media with 50 μg mL−1 apramycin at 37°C. Integration of pKN99 into the genome of A. hibisca was confirmed by PCR analysis using genome- and vector-specific primers (Table S1). With primer 1 and RV-M, a 1.6-kb fragment was amplified from the genome of A. hibisca/pKN99 (Fig. 2B, panel iii, lane 1). Similarly, a 1.6-kb PCR product was obtained with primer 2 and M13-47 from the mutant (lane 2, panel iii, Fig. 2B, panel iii, lane 2). M13-47 and RV-M primers are unique sequences from the vector pKC1139. Consistently, these two 1.6-kb fragments could not be amplified from the wild type (data not shown). Primers 1 and 2 amplified a 1.4-kb band from the genomic DNA of the wild type A. hibisca, while no PCR product was obtained from the mutant A. hibisca/pKN99 (Fig. 2B, panel iii, lanes 4 and 3). This clearly shows the insertion of a long segment of DNA between primer 1 and primer 2. Thus, these PCR results confirmed that pKN99 was successfully inserted into the genome of A. hibisca.
The fermentation broth of A. hibisca/pKN99 in YM medium supplemented with 50 mg L−1 apramycin was analyzed by LCMS. As shown in Figure 3 (trace iv), this strain produced a major product 3 that is slightly polar than 1. This compound has a UV spectrum (Fig. S1) similar to that of 1, while ESI-MS revealed that its molecular weight is 826 (Fig. S2), 14 mass units smaller than 1. It has the same retention time as that of pradimicin C. Thus, it can be concluded that A. hibisca/pKN99 produced 3 as the major product, with 1 as the minor product. Given the fact that 3 lacks the N-CH3, PdmO was deduced to be the dedicated N-MT in pradimicin biosynthesis that catalyzes the specific N-methylation of the 4,6-dideoxy-4-amino-D-galactose moiety. Integration of pKN99 into the genome of A. hibisca did not disrupt pdmO (Fig. 2A). Consequently, the synthesis of 1 was still observed. However, insertion of a large fragment upstream of the intact pdmO may have influenced the function of the promoter that controls the expression of PdmO. This explains why 3 was generated as the major metabolite and is similar to observations in our previous work on HerF in herboxidiene biosynthesis.11 4 was also a minor metabolite in this mutant, which might be caused by the incomplete glycosylation of the aglycon.
In summary, using a double crossover recombination approach, we disrupted two putative GT genes in the pradimicin biosynthetic gene cluster, leading to the production of pradimicin analogs with no or one sugar moiety. Based on the corresponding product of the mutants, PdmS and PdmQ were characterized as the GTs that are respectively responsible for the introduction of the first and second sugar moieties to the aglycon 4 (Fig. 4). By inserting a large gene fragment into the genome of A. hibisca, we also found that PdmO methylates the amino group of the 4',6'-dideoxy-4'-amino-D-galactose moiety in pradimicin biosynthesis (Fig. 4). Unveiling the roles of these GTs and the N-MT further facilitates the understanding and rational engineering of the pradimicin biosynthetic pathway.
Figure 4.
The biosynthetic pathway of 1.
Supplementary Material
Acknowledgments
This work was supported by the National Institute of Allergy and Infectious Diseases (R15AI089347).
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplementary Material
Supplementary material that may be helpful in the review process should be prepared and provided as a separate electronic file. That file can then be transformed into PDF format and submitted along with the manuscript and graphic files to the appropriate editorial office.
References and notes
- 1.Nakagawa Y, Doi T, Takegoshi K, Igarashi Y, Ito Y. Bioorg. Med. Chem. Lett. 2012;22:1040. doi: 10.1016/j.bmcl.2011.11.106. [DOI] [PubMed] [Google Scholar]
- 2.Nakagawa Y, Doi T, Taketani T, Takegoshi K, Igarashi Y, Ito Y. Chem. Eur. J. 2013;19:10516. doi: 10.1002/chem.201301368. [DOI] [PubMed] [Google Scholar]
- 3.Fukagawa Y, Ueki T, Numata K-i, Oki T. Actinomycetologica. 1993;7:1. [Google Scholar]
- 4.Kim BC, Lee JM, Ahn JS, Kim BS. J. Microbiol. Biotechnol. 2007;17:830. [PubMed] [Google Scholar]
- 5.Zhan J, Watanabe K, Tang Y. ChemBioChem. 2008;9:1710. doi: 10.1002/cbic.200800178. [DOI] [PubMed] [Google Scholar]
- 6.Napan K, Zhang S, Morgan W, Anderson T, Takemoto JY, Zhan J. ChemBioChem. 2014;15:2289. doi: 10.1002/cbic.201402306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zhan J, Qiao K, Tang Y. ChemBioChem. 2009;10:1447. doi: 10.1002/cbic.200900082. [DOI] [PubMed] [Google Scholar]
- 8.Kieser T, Bibb MJ, Buttner MJ, Chater KF, Hopwood DA. Practical streptomyces genetics: John Innes Foundation. 2000 [Google Scholar]
- 9.E. coli ET12567 (pUZ8002) was transformed with pKN82 as the conjugation donor. 100 mL of the culture of E. coli ET12567(pUZ8002)/pKN82 or pKN88 was grown in LB broth to an OD600 of 0.4–0.6 in the presence of 50 mg L−1 apramycin, 50 mg L−1 kanamycin, and 25 mg L−1 chloramphenicol. The cells were then collected by centrifugation and washed twice with blank LB medium to remove the antibiotics, and then suspended in 0.1 mL of LB medium. A. hibisca spores were collected with 2× YT broth after growing for 15 days on YM medium at 30°C. The A. hibisca spores were heat shocked at 50°C for 10 min. Serial dilutions of the germinated spore suspension were made and 100 µL of each dilution was mixed with an equal volume of E. coli ET12567 (pUZ8002)/pKN82 on MS plates (20 g L−1 D-mannitol, 20 g L−1 soya flower, 20 g L−1 glucose, and 20 g L−1 agar), which were grown for 12 h at 30°C. After incubation, the plates were overlayed with 1 mL of sterile water containing 30 mg L−1 nalidixic acid and 50 mg L−1 apramycin, and incubated at 30°C until the exconjugants appeared. Exconjugants were visible after 4–5 days and started producing spores in 8–10 days.
- 10.The culture of the double crossover mutant of A. hibisca/pKN82 in 500 mL of YM medium was centrifuged and the supernatant was subjected to Diaion HP-20 resin column chromatography, eluted with 0%, 25%, 50%, 75%, and 100% aqueous methanol (v/v) to yield five fractions. HPLC analysis revealed that the 75% fraction contained the target peak. This fraction was further separated by reverse-phase HPLC (Agilent Zorbax SB-C18, 5 µm, 21.2 mm × 150 mm), eluted with 35% aqueous acetonitrile containing 0.1% formic acid, to yield 5.0 mg of 4.
- 11.Sawada Y, Tsuno T, Yamamoto H, Nishio M, Konishi M, Oki T. J. Antibiot. 1990;43:1367. doi: 10.7164/antibiotics.43.1367. [DOI] [PubMed] [Google Scholar]
- 12.Burgie ES, Holden HM. Biochemistry. 2008;47:3982. doi: 10.1021/bi800063j. [DOI] [PubMed] [Google Scholar]
- 13.Carney AE, Holden HM. Biochemistry. 2011;50:780. doi: 10.1021/bi101733y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Yu D, Xu F, Zhang S, Shao L, Wang S, Zhan J. Bioorg. Med. Chem. Lett. 2013;23:5667. doi: 10.1016/j.bmcl.2013.08.023. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.