Pentostatin (PTN, deoxycoformycin) and arabinofuranosyladenine (Ara-A, vidarabine) are purine nucleoside antibiotics used clinically to treat hematological cancers and human DNA virus infections, respectively. PTN has a 1,3-diazepine ring, and Ara-A is an adenosine analog with an intriguing epimerization at the C-2' hydroxyl group. However, the logic underlying the biosynthesis of these interesting molecules has long remained elusive. Here, we report that the biosynthesis of PTN and Ara-A employs an unusual protector-protégé strategy. To our surprise, we determined that a single gene cluster governs PTN and Ara-A biosynthesis via two independent pathways. Moreover, we verified that PenB functions as a reversible oxidoreductase for the final step of PTN. Remarkably, we provided the first direct biochemical evidence that PTN can protect Ara-A from deamination by selective inhibition of the host adenosine deaminase. These findings expand our knowledge of natural product biosynthesis and open the way for target-directed genome mining of Ara-A/PTN-related antibiotics.

BackgroundAurantimycin (ATM), produced by Streptomyces aurantiacus JA 4570, is a potent antimicrobial and antitumor antibiotic. Although the chemical structure of ATM is highly distinctive and features a cyclohexadepsipeptide scaffold attached with a C14 acyl side chain, little is known about its biosynthetic pathway and regulatory mechanism.ResultsIn this work, we report the identification and characterization of the ATM biosynthetic gene cluster from S. aurantiacus JA 4570. Targeted inactivation of artG, coding for a NRPS enzyme, completely abolished ATM production, thereof demonstrating the target gene cluster (art) is responsible for ATM biosynthesis. Moreover, four NRPS adenylation (A) domains including a freestanding enzyme ArtC have been characterized in vitro, whose substrate specificities are consistent with in silico analysis. Further genetic analysis of the two regulatory genes artB and artX unambiguously suggested both of them play positive roles in ATM biosynthesis, and ATM-A production was thus rationally enhanced to about 2.5 fold via tandem overexpression of artB and artX in S. aurantiacus JA 4570.ConclusionsThese results will provide the basis for the understanding of precise mechanisms for ATM biosynthesis, and open the way for both rational construction of high-production ATM producer and orient-directed generation of designer ATM derivatives via synthetic biology strategies.Electronic supplementary materialThe online version of this article (doi:10.1186/s12934-016-0559-7) contains supplementary material, which is available to authorized users.

No abstract

Polyoxin, produced by Streptomcyes cacaoi var. asoensis and Streptomyces aureochromogenes, contains two non-proteinogenic amino acids, carbamoylpolyoxamic acid (CPOAA) and polyoximic acid. Although the CPOAA moiety is highly unusual, its biosynthetic logic has remained enigmatic for decades. Here, we address CPOAA biosynthesis by reconstitution of its pathway. We demonstrated that its biosynthesis is initiated by a versatile N-acetyltransferase, PolN, catalyzing L-glutamate (1) to N-acetyl glutamate (2). Remarkably, we verified that PolM, a previously annotated dehydrogenase, catalyzes an unprecedented tandem reduction of acyl-phosphate to aldehyde, and subsequently to alcohol. We also unveiled a distinctive acetylation cycle catalyzed by PolN to synthesize α-amino-δ-hydroxyvaleric acid (6). Finally, we report that PolL is capable of converting a rare sequential hydroxylation of α-amino-δ-carbamoylhydroxyvaleric acid (7) to CPOAA. PolL represents an intriguing family of Fe(II)-dependent α-ketoglutarate dioxygenase with a cupin fold. These data illustrate several novel enzymatic reactions, and also set a foundation for rational pathway engineering for polyoxin production.

To investigate the possibility of removing titanium dioxide nanoparticles (TiO2 NPs) from water by coagulation, as well as to find the optimal coagulant and experimental conditions for TiO2 NP removal, four types of coagulant were adopted: polyferric sulfate (PFS), ferric chloride (FeCl3), polyaluminum chloride (PACl), and alum (Al2(SO4)3). It was found that the removal of TiO2 NPs by coagulation was affected by ionic strength, alkalinity, as well as types and dosages of coagulants. PFS and FeCl3 achieved much higher removal efficiency of TiO2 NPs than PACl and Al2(SO4)3 did. For 30 mg/L TiO2 NPs, a dosage of 0.3 mM PFS (as Fe) achieved 84% removal after coagulation followed by 30 min settlement. Optimal ionic strength (0.1 M NaCl or 0.03 M CaCl2) is of vital importance for the performance of PFS. Na2SO4 is unfavorable for the performance of PFS. Optimal alkalinity (0.01-0.03 M NaHCO3) is necessary for FeCl3 to remove TiO2 NPs. Natural organic matter, as represented by humic acid (HA) up to 11 mg/L, reduces the removal of TiO2 NPs by coagulation. These findings indicate that coagulation is a good option for the removal of TiO2 NPs from water, and more attention should be paid to the effects of water quality when using coagulation to remove TiO2 NPs from aqueous matrices. This provides a possible solution to alleviate the potential hazard caused by TiO2 NPs.