The zeamine antibiotics: unusual polyamine-polyketide-nonribosomal peptide secondary metabolites from Serratia plymuthica RVH1

The aim of this PhD project was to identify the antimicrobial metabolites produced by Serratia plymuthica RVH1 and to study their structure, mode of action, biosynthetic gene cluster and biosynthetic pathway, as well as rates and mechanisms of resistance. Bioactivity-guided purification and structure elucidation efforts led to the identification of the zeamines, an unusual class of cationic polyamine-polyketide-nonribosomal peptide natural products with potent activity against a broad spectrum of Gram-positive and Gram-negative bacteria. The zeamine antibiotic complex from S. plymuthica RVH1 was initially believed to consist of three compounds: zeamine and zeamine II (previously identified as metabolites of the phytopathogen Dickeya zeae), as well as zeamine I (a novel analogue). Zeamine II was found to be less effective (8-fold higher MIC) at inhibiting the growth of Staphylococcus aureus ATCC27661 (MIC = 32 µg/ml) and Escherichia coli MG1655 (MIC = 64 µg/ml) than zeamine and zeamine I. With an MBC/MIC ratio of 2, the zeamines can be considered bactericidal antibiotics. The zeamine metabolites further demonstrated potent cytostatic activity against human and murine cancer cells with IC50 values ranging from 1.7 to 4.1 μM.Examination of their specific mode of action and cellular target revealed that zeamine, zeamine I and zeamine II disrupt the integrity of cell membranes. The antibiotics induce rapid and concentration-dependent leakage of carboxyfluorescein from small unilamellar vesicles of different phospholipid composition, demonstrating their ability to directly interact with lipid bilayers in the absence of a specific receptor. In this process, the zeamines were found to have a higher selectivity for bacterial over eukaryotic model membranes. DNA, RNA, fatty acid and protein biosynthesis cease rapidly and simultaneously in E. coli and S. aureus upon treatment with sub-inhibitory concentrations of the antibiotics, presumably as a direct consequence of membrane perturbation. The zeamines are also able to permeabilize the outer membrane of Gram-negative bacteria, facilitating the uptake of small molecules, such as 1-N-phenylnaphtylamine, in a dose-dependent manner. The valine-linked polyketide moiety in zeamine and zeamine I was found to increase the efficiency of this process. In contrast, bulky hydrophilic fluorescent proteins were not able to cross the outer membrane upon zeamine exposure, suggesting that the zeamines induce subtle perturbations of the outer membrane rather than drastic alterations or large pore formation. At concentrations well above the MIC, the zeamines were shown to cause membrane lysis, as indicated by time-lapse microscopy. Together, these results indicate that the bactericidal activity of zeamine, zeamine I and zeamine II derives from generalized membrane permeabilization, which is likely initiated by hydrophobic and electrostatic interactions with negatively charged membrane components.To evaluate the potential for resistance development to the zeamines, five independent lineages of S. aureus strain Newman were serially exposed to sub-inhibitory concentrations of the antibiotics. No mutants with decreased susceptibility to the zeamines could be obtained. Thus, rapid acquisition of zeamine resistance via genomic mutations seems unlikely. In a search for existing resistance determinants, the genome of Serratia proteamaculans strain LMG7884 was examined. This strain has a higher intrinsic resistance to the zeamines. One particular locus was found to encode a putative GCN5-related N-acetyltransferase that induces an eight-fold increase in the MIC of the zeamines against S. proteamaculans DSM4543.By using a combination of random transposon mutagenesis, genomic BAC-library screening and targeted gene deletions, zeamine biosynthesis was mapped to a 50 kb gene cluster (zmn5, zmn9-22) in S. plymuthica RVH1 which appears to be conserved across different enterobacterial species and has likely been acquired via horizontal gene transfer. In addition to tailoring and export-related enzymes, two separate assembly lines are encoded by genes within the zeamine gene cluster: an iterative type I fatty acid synthase (FAS)/polyketide synthase (PKS) homologous to polyunsaturated fatty acid (PUFA) synthases, and a hybrid nonribosomal peptide synthetase (NRPS)/PKS system. By using a combination of targeted gene deletions, high resolution LC-MS(/MS) analyses, in vitro biochemical assays and feeding studies, the pathway for zeamine biosynthesis was characterized. The pentaamino-hydroxyalkyl chain is synthesized by the multidomain FAS/PKS assembly line encoded by zmn10-13. The terminal, stand-alone thioester reductase Zmn14 is proposed to reductively release a tetraamino-hydroxyalkyl thioester as an aldehyde, using NADH as a co-substrate. Despite the intrinsic ability of this enzyme to fully reduce aldehyde substrates to alcohols, the aldehyde intermediate is believed to be preferentially transaminated, yielding zeamine II. In a parallel pathway, hexapeptide-monoketide and hexapeptide-diketide thioesters are produced by the hybrid NRPS/PKS formed by Zmn16-18 and subsequently conjugated to zeamine II by the freestanding condensation domain Zmn19. The resulting prezeamines, additional members of the zeamine antibiotic complex containing an N-terminal pentapeptide, were structurally elucidated and implicated as precursors to zeamine and zeamine I. The prezeamines are partially converted into zeamine and zeamine I by the action of an acylpeptide hydrolase that specifically cleaves off the N-terminal pentapeptide in a post-assembly proteolytic processing step. Thus, the zeamine antibiotics are assembled by a unique combination of NRPS, type I modular PKS and PUFA synthase-like multienzyme complexes. This unusual biosynthetic machinery may represent a valuable contribution to the synthetic biology toolbox for combinatorial biosynthesis and rational design of novel bioactive compounds. Furthermore, it offers valuable insights into atypical chain release mechanisms from type I PKS and PUFA synthase-like multienzymes and increases our current knowledge and understanding of secondary lipid biosynthetic processes.

Publication year:2015