Cloning and characterization of the biosynthetic gene cluster of the bacterial RNA polymerase inhibitor tirandamycin from marine-derived Streptomyces sp. SCSIO1666

Abstract

Tirandamycins are bacterial RNA polymerase inhibitors holding great potential for antibacterial agent design. To elucidate the biosynthetic machinery and generate new derivatives, the tirandamycin biosynthetic gene cluster was cloned and sequenced from marine-derived Streptomyces sp. SCSIO1666. The biosynthetic gene cluster of tirandamycin spans a DNA region of ∼56 kb and consists of 15 open reading frames (ORFs) which encode three type I polyketide synthases (TrdAI, AII, AIII), one non-ribosomal peptide synthetase (TrdD), one phosphopantetheinyl transferase (TrdM), one Type II thioesterase (TrdB), one FAD-dependent oxidoreductase (TrdL), one cytochrome P450 monooxygenase (TrdI), three proteins related to resistance and regulations (TrdHJK), and four proteins with unknown function (TrdCEFG). To investigate the roles of the genes played in the biosynthetic machinery, seven genes (trdAI and trdBDFHIK) were inactivated via in frame replacement with an apramycin gene cassette using λ-RED recombination technology. The ΔtrdAI and ΔtrdD mutants targeting the ketosynthase and adenylation domain of TrdAI and TrdD, respectively, abolished the production of tirandamycins, confirming their involvement in the tirandamycin biosynthesis. TrdH showed high homology to LuxR family transcriptional regulatory proteins, disruption of which abolished the production of tirandamycins, indicating that TrdH is a positive regulator for tirandamycin biosynthesis. On the other hand, TrdK showed high homology to TetR-family transcriptional regulatory proteins, disruption of which significantly increased the yields of tirandamycins almost one-fold, implicating that TrdK is a negative regulator for tirandamycin biosynthesis. Disruption of the gene trdI resulted in the accumulation of the intermediate tirandamycin C (3) and a trace amount of new product tirandamycin C2 (5). A model of tirandamycin biosynthesis was proposed based on bioinformatics analyses, gene inactivation experiments and intermediates isolated from the mutants. These findings set the stage for further study of the tirandamycin biosynthetic mechanism and rationally engineer new tirandamycin analogues.

Introduction

The tirandamycins (1–4, Fig. 1) belong to a small group of naturally occurring tetramic acid (2, 4-pyrrolidinedione) antibiotics that exhibited broad biological activities, such as antiviral, antiulcerative, antitumor, and antibacterial properties [1]. The structures of tirandamycins are further featured with a conjugated dienoyltetramic acid chromophore and an intriguing bicyclic ketal scaffold. Other structurally closely related natural products include streptolydigin (6), tirandalydigin (7), Bu-2313A (8), Bu-2313B (Nocamycin I, 9) and Nocamycin II (10) (Fig. 1) [1]. This small focused group of antibiotics all displayed antimicrobial activity by targeting bacterial RNA polymerases (RNAPs) and blocking the transcriptional initiation and elongation process [1], [2].

Tirandamycin A (1) was first isolated in 1971 from Streptomyces tirandis[3] and was re-isolated from Streptomyces flaveolus in 1976 along with tirandamycin B (2) [4]. In recent years, 1 and 2 were again isolated from Streptomyces sp. 307–9 [5] and Streptomyces sp. SCSIO1666 [6], two marine-derived actinomycete bacteria isolated from samples collected in Salt Kay, US Virgin Islands and Sanya Island, South China Sea, respectively. Tirandamycins C (3) and D (4) were also purified as trace elements by using resin-capture method [5]. Our early work has revealed that 2 is the predominant product under normal fermentation conditions in Streptomyces sp. SCSIO 1666, and the yield of 1 was improved 250-fold when XAD-16 resin was added to the fermentation broth to trap biosynthetic intermediates, indicating that 2 is the final and most highly oxidized product of the tirandamycin biosynthetic pathway [6].

Recent years, the complex crystal structures of RNAPs from Escherichia coli and Thermus thermophilus with 6 have been solved, providing molecular evidence for the structural basis of the RNAP interaction mechanism [7]. To date, rifampicin is the only RNAP inhibitor currently used in clinical practice for the treatment of tuberculosis, requiring us to develop more effective RNAP inhibitors to combat the increasing drug resistant bacteria. Understanding the biosynthetic pathway of the structurally unique tetramic acid-containing antibiotics as well as the regulatory and resistant mechanisms would have great impact on guiding us to engineer and design novel analogs of this class as antibacterial agents. Due to the unusual structural features, unique mode of action and potential for drug development, we started to clone the tirandamycin gene cluster from Streptomyces sp. SCSIO1666. While the work was progressing, Salas et al. has reported the biosynthetic gene cluster for 6 from Streptomyces lydicus NRRL2433 and generated four glycosylated derivatives [8], and Sherman et al. has reported a tirandamycin gene cluster from Streptomyces sp. 307–9 [9]. In this paper, we report: (i) the cloning, sequencing, analyses and confirmation of the tirandamycin biosynthetic gene cluster from Streptomyces sp. SCSIO 1666; (ii) the identification of TrdH as a positive regulator and TrdK as a negative regulator involved in the tirandamycin biosynthesis; (iii) inactivation of TrdI resulting in accumulation of 3 and a trace amount of new product tirandamycin C2 (5); and (iv) a proposed model for tirandamycin biosynthesis that is supported by bioinformatics analyses and gene inactivation experiments.