BURNS, J. 2020. Exploitation of underused Streptomyces through a combined metabolomics-genomics workflow to enhance natural product diversity. Robert Gordon University [online], PhD thesis. Available from: https://openair.rgu.ac.uk Exploitation of underused Streptomyces through a combined metabolomics-genomics workflow to enhance natural product diversity. BURNS, J. 2020 The author of this thesis retains the right to be identified as such on any occasion in which content from this thesis is referenced or re-used. The licence under which this thesis is distributed applies to the text and any original images only – re-use of any third-party content must still be cleared with the original copyright holder. This document was downloaded from https://openair.rgu.ac.uk Exploitation of underused Streptomyces through a combined metabolomics-genomics workflow to enhance natural product diversity Joshua Burns A thesis submitted in partial fulfilment of the requirements of the Robert Gordon University for the degree of Doctor of Philosophy This research programme was carried out in collaboration with NCIMB Ltd. May 2020 ACKNOWLEDGEMENTS IV LIST OF ABBREVIATIONS V ABSTRACT XI 1 INTRODUCTION 3 1.1 An overview of modern microbial antibiotic discovery 3 1.2 Rise of antimicrobial resistance 22 1.3 Streptomyces and specialised metabolism 28 1.4 Antimicrobial discovery using Streptomyces 37 1.5 Thesis aim 45 2 METABOLOMIC SCREENING OF S. COELICOLOR A3(2) 51 2.1 Introduction 51 2.2 Materials and Methods 60 2.3 Results and Discussion 71 2.4 Conclusions 93 3 METABOLOMICS-BASED PROFILING AND SELECTION OF UNEXPLOITED STREPTOMYCES STRAINS 97 3.1 Introduction 97 3.2 Materials and Methods 101 3.3 Results and Discussion 108 3.4 Conclusions 132 4 CHARACTERISATION OF THE S. COSTARICANUS GENOME AND ITS BIOSYNTHETIC GENE CLUSTERS 135 4.1 Introduction 135 4.2 Materials and Methods 142 4.3 Results and Discussion 144 4.4 Conclusions 175 ii 5 SCALE-UP CULTURE OF S. COSTARICANUS FOR EXTRACTION AND ISOLATION OF BIOACTIVE METABOLITES 179 5.1 Introduction 179 5.2 Materials and Methods 184 5.3 Results and Discussion 189 5.4 Conclusions 228 6 GENERAL DISCUSSION 231 6.1 Thesis aim 231 6.2 Suitability of methods 231 6.3 Future work 237 6.4 Conclusions 238 7 REFERENCES 240 8 APPENDIX 270 8.1 Appendix figures 270 8.2 Appendix tables 295 iii ACKNOWLEDGEMENTS Firstly, I would like to thank my project supervisors Christine Edwards and Linda Lawton at RGU and Sam Law at NCIMB. Their guidance and support at all stages of my PhD, and the Knowledge Transfer Partnership on which it was based, have been essential for the entirety of the project and I am extremely grateful for it. The other members of CyanoSol in RGU were hugely helpful so thanks go to Len Montgomery, Aakash Welgama, Carlos Pestana, Calum McNerney, Julia Waack, Declan Maxwell, and Joe Palmer for all their help and advice at every stage of the process. I am also indebted to Joanna Reed, Philippa Hulme, and Coline Buanic for their help in data generation for Chapter 5, and to Dan Swan at NCIMB for all his help with Streptomyces genome sequencing and input on Chapter 4. Next are all the other RGU students who made my stay in Aberdeen enjoyable through a mix of their friendship, willingness to sit through complaints, and (usually) tolerating my jokes. Amongst others, I’d most like to thank Calum McNerney, Matteo Scipioni, Franzi Pohl, Hazel Ramage, Mhairi Paul, Zoi Papadatou, Teo Stoyanova, Martin Corsie, Quirin Werthner, Tesnime Jebara, Ahmed Salaheldin, PJ Barron and Qas Ali. I would also like to thank the staff at RGU, especially Andrea Macmillan and Dorothy McDonald for all their help with admin and large amounts of free tea. I also need to thank the other staff at NCIMB, in particular Carol Phillips for her contributions to the project and advice on the industry and business aspects. I am deeply grateful to Innovate UK for funding the project through KTP grant 10124, and for the guidance of the KTP advisors Mark Abbs and Ian Heywood. Finally, I want to thank my family for the constant love and support that they have given me over the last few years and at every other time. To my Mum, Dad, Caroline, Rebecca, Ofra, Maya, Eldad and Lottie, thank you, and I wouldn’t have been able to do any of it without you. iv LIST OF ABBREVIATIONS A Adenylation ACP Acyl carrier protein ACT Actinorhodin AMR Antimicrobial resistance ANI Average nucleotide identity antiSMASH Antibiotics and secondary metabolite analysis shell AT Acyl transferase autoMLST Auto multi locus sequence typing BA Bennett’s agar BEH Ethylene bridged hybrid BGC Biosynthetic gene cluster BLAST Basic local alignment search tool BLASTN Basic local alignment search tool-nucleotide BLASTP Basic local alignment search tool-protein v BUSCO Benchmarking universal single copy orthologs C Condensation CCR Carbon catabolite repression CDA Calcium dependent antibiotic CDS Coding sequences CS Corn steep liquor-soya flour DSMZ Deutsche sammlung von mikroorganismen und zellkulturen ESBL Extended spectrum beta-lactamase ENV+ Polystyrene-divinylbenzene EP Extended phenotype FASTA Fast-all GlcNAc N-acetyl glucosamine GNPS Global natural products social molecular networking GYM Glucose yeast-malt HP High polarity pool vi ISP5 International Streptomyces project medium 5 KS Ketosynthase LAP Linear azoline containing peptide LC-MS Liquid chromatography-mass spectrometry LP Lower polarity pool M19 Medium 19 M400 Medium 400 MHA Muller-Hinton agar MHB Muller-Hinton broth MiBIG Minimum information about a biosynthetic gene cluster MM Molasses-meat extract MMM Minimal mannitol medium MP Medium polarity pool MRSA Methicillin-resistant Staphylococcus aureus MS/MS Mass spectrometry/mass spectrometry vii MYM Maltose yeast-malt NaBu Sodium butyrate NCIMB National collection of industrial and marine bacteria NGS Next generation sequencing NRPS Non-ribosomal polyketide synthetase NUT Nutrient agar OM Oatmeal medium OSMAC One strain many compounds PBP Penicillin-binding protein PCA Principal component analysis PCP Peptidyl carrier protein PE Paired end PepS Peptone-sucrose medium PKS Polyketide synthetase PLS-DA Partial least squares-discriminant analysis viii ProYM Proline yeast-malt QToF Quadrupole time of flight QUAST Quality assessment tool for genome assemblies RAST Rapid annotation using subsystem technology RiPP Ribosomally synthesized and post-translationally modified peptides RED Undecylprodigiosin SCCmec Staphylococcal cassette chromosome mec SE Single end SEA Soil extract Agar SGT Soybean-glucose-tryptone SMet Specialised metabolite SM Sucrose-meat extract SPNT Supernatant TSA Tryptone soya agar UPLC Ultra-performance liquid chromatography ix WHO World health organisation WGS Whole genome sequencing XTT 2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5- [(phenylamino) carbonyl]-2H-tetrazolium hydroxide x ABSTRACT The genus Streptomyces is the source of approximately two thirds of all clinically used antibiotics. Despite being the source of so many specialised metabolites, genomic analysis indicates that most Streptomyces strains have the potential to produce around 25 bioactive metabolites, some of which may be the basis of novel therapies. This makes culture collections of Streptomyces spp. an easily accessible but underused resource to mine for genomic and metabolomic variety. Therefore, the main aim of this project was to initiate exploitation of the culture collection at NCIMB Ltd. by expanding the available chemical space from underutilised Streptomyces for the production of novel antibiotics. This primarily used a mixture of metabolomic and genomic methods. A high-throughput culture parameter screen was designed around multiple carbon sources, nitrogen sources, and extraction sample times. This was tested on the model species S. coelicolor A3(2) to compare differences in the production of known specialised metabolites, using UPLC-MS to analyse crude extracts from growth on agar. Data was analysed using MZmine and putative metabolites identified using freely available MS/MS databases, primarily GNPS. This showed clear variation in production of 9 identified metabolites including deferoxamines, germicidins, undecylprodigiosin and coelichelin as a result of different culture parameters. Therefore, the screen successfully expanded the available chemical space, so was applied to non-model Streptomyces strains. The screen was used to compare the total metabolomic variety produced by 3 Streptomyces isolated from different environments in order to select a strain for further investigation. Comparing metabolomic features using principal component analysis showed the Costa Rican soil isolate S. costaricanus to produce the most variety versus the other 2 Streptomyces strains. The metabolite family most responsible for principal component separation was identified as the actinomycins. Scale-up of both agar and broth culture was used for metabolite dereplication and bioassays against multidrug resistant Acinetobacter baumannii, one of the bacteria on the World Health Organisation’s list of pathogens most urgently requiring new therapies. Fractions were derived from broth culture supernatant and agar crude extract by flash chromatography, resulting in semi-purified fractions. The predominant metabolite families in xi fractions were actinomycins and deferoxamines, which were further split by polarity into separate fractions. This resulted in rapid purification of metabolites, with 1 fraction comprising 80% deferoxamine B by weight. Fractions were tested against A. baumannii using the 2,3-bis