Discovery and development of antibiotics by genome mining, protein characterization and biosynthetic pathway manipulation

Abstract

After the golden era of antibiotic discovery in the 1940s-1960s, we are currently facing an emerging antibiotic crisis as bacteria develop resistance to antibacterial agents in use. The declining antibiotic drug development pipeline aggravates this. The development of resistance after exposure to an inhibiting compound like an antibiotic is a natural consequence of the stress response of living organisms to survive. Even worse, the resistance gene(s) can be transferred between bacteria resulting in a fast spreading. Natural products produced by bacteria and fungi still significantly contribute to the discovery and development of new antibiotics. However, the antibiotics discovery pipeline is not adequately filled, due to: (1) re-discovery of known compounds, (2) a majority of bacteria is unculturable under laboratory conditions, and (3) the biosynthetic gene clusters (BGCs) encoding active compounds are not expressed. In this work, the biosynthesis of the ambigols produced by Fischerella ambigua was characterized in vitro. The ambigols show antibacterial, antifungal and molluscicidal effects, and inhibit cyclooxygenase and HIV reverse transcriptase. Enzymes involved in the biosynthesis were cloned and heterologously expressed in E. coli cells. 3-deoxy-7-phosphoheptulonate (DAHP) synthase (Ab7) is involved in chorismate biosynthesis by the shikimate pathway. Chorismate in turn is further converted by a dedicated chorismate lyase (Ab5), yielding 4-hydroxybenzoic acid (4-HBA). The stand-alone adenylation domain Ab6 is necessary to activate 4-HBA, which is subsequently tethered to the acyl carrier protein (ACP) Ab8. The Ab8-bound substrate is chlorinated by Ab10 in meta position yielding 3-Cl-4-HBA, which is then transferred by the condensation (C) domain to the peptidyl carrier protein and released by the thioesterase (TE) domain of Ab9. The released product is then expected to be the dedicated substrate of the halogenase Ab1 producing 2,4-dichlorophenol as the monomeric building block for ambigol biosynthesis, catalyzed by the previously characterized P450 enzymes Ab2 and Ab3. Furthermore, to discover new compounds possessing antibiotic properties, a genome mining approach was implemented resulting in discovery of polyoxyperuins (POPs) and monocyclic (mc-) darobactin. The pop BGC (PKS-NRPS-hybrid) was identified from our in-house genomic database originated from the marine-derived Streptomyces sp. s120. To characterize its products, homologous expression and comparative MS analysis of extracts from the native producer and a knockout mutant led to the identification of the metabolites polyoxyperuin A seco acid (1) and polyoxyoeruin A (2) corresponding to the pop BGC. Furthermore, by overexpression of a regulatory element, i.e. a LmbU-like transcriptional activator, the production yield of 1 and 2 was increased, enabling isolation and structure elucidation using high-resolution mass spectrometry and NMR spectroscopy. Compound 1 exhibited a low antibiotic effect against Micrococcus luteus, while 2 showed a strong Gram-positive antibiotic effect in micro-broth-dilution assays. A BGC showing high similarity to the darobactin BGC was mined from NCBI public database. The BGC was identified from a chromosomal genomic sequence of Sulfidibacter corallicola M133T. Since the natural producer is not in hand, a heterologous expression approach was used to express the BGC and to generate its product. Therefore, the BGC was generated synthetically, cloned in the pRSFduet expression vector and transferred to E. coli as a host. Comparative MS analysis of extracts from the heterologous producer and control led to the identification of two ionized masses corresponding to the putative product of the mcd BGC. The product is predicted to be the core peptide McdB (W-R-W-S-W-P-F) with one additional oxygen and loss of two hydrogens. The modification is suspected to correlate to the formation of a C-O-C (ether bond) like it is the case for darobactin. The product was named mono cyclic (mc-) darobactin, since the latter carries an additional ring formed between two carbon atoms. Moreover, as darobactin is a lead compound for a new antibiotic, a method to increase its production yield, a comprehensive understanding of the biosynthesis and derivatization to obtain more active derivatives and get first insights into its structure-activity-relationship were performed. Using a heterologous expression system, production yield was increased 10-fold and the production time decreased 5-fold. Furthermore, the minimum BGC was identified to consist of only darA and darE. By an in vivo study, modification of the heterologously expressed tagged DarA by DarE was investigated. The result suggested that DarE catalyzes both ring formations in DarA. Modifications occurred before the core peptide was released from the DarA-leader and both ring modifications could happen independently. For the derivatization of darobactin, its biosynthesis was manipulated by mutating codons encoding the residues in the core peptide of DarA. In that way the ribosomally synthesized and post-translationally modified peptide (RiPP) darobactin A was derivatized yielding 69 derivatives of which 37 showed inhibitory activity against E. coli. From the initial test, one of them shows better antimicrobial activity against P. aeruginosa and A. baumannii strains than Darobactin A and B.