UC San Diego UC San Diego Electronic Theses and Dissertations

Title The chemical ecology of marine bacteria and the sediments they inhabit

Permalink https://escholarship.org/uc/item/7594396g

Author Tuttle, Robert

Publication Date 2018

Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital Library University of California

UNIVERSITY OF CALIFORNIA SAN DIEGO

The chemical ecology of marine bacteria and the sediments they inhabit

A dissertation submitted in partial satisfaction of the requirements for the degree Doctor of Philosophy

Marine Biology

Robert Nicholas Tuttle

Committee in charge: Professor Paul Jensen, Chair Professor Gregory Rouse Professor Pieter Dorrestein Professor Chambers Hughes Professor Stuart Sandin

Robert Nicholas Tuttle, 2018

The dissertation of Robert Nicholas Tuttle is approved, and it is acceptable

in quality and form for publication on microfilm and electronically:

______Chair

University of California San Diego

2018

iii

Dedication

To my grandfather, Dr. Harry J. Woehr

You taught me that greatness often comes from humble beginnings.

EPIGRAPH

Nothing in life is to be feared, it is only to be understood. Now is the time to understand more, so that we may fear less.

-Marie Curie

TABLE OF CONTENTS SIGNATURE PAGE ...... …..iii DEDICATION ...... …..iv EPIGRAPH...... …...v TABLE OF CONTENTS ...... …..vi LIST OF FIGURES ...... …..ix LIST OF TABLES ...... …..xi ACKNOWLEDGMENTS ...... ….xii VITA ...... ….xv ABSTRACT OF THE DISSERTATION ...... xvii Chapter 1 Introduction ...... …..1 1.1 Introduction to chemical ecology ...... …..2 1.2 Marine chemical ecology – Eukaryotes ...... …..3 1.3 Marine chemical ecology – Prokaryotes ...... …..5 1.3.1 Nutrient limitation ...... …..6 1.3.2 Prokaryote- prokaryote allelopathy...... …..7 1.3.3 Prokaryote - Eukaryote allelopathy ...... …..8 1.4 Induction of bacterial specialized metabolites ...... …..9 1.4.1 Cell to cell contact ...... …10 1.4.2 Nutrient supplementation ...... …11 1.4.3 Small molecules ...... …11 1.5 Sediment bacteria – focusing on Salinispora ...... …12 1.6 Advancements in mass spectrometry ...... …14 1.7 Summary of thesis ...... …16 1.8 References ...... …17 Chapter 2 The Detection of Specialized Metabolites and Their Producers in Ocean Sediments .. …23 2.1 Abstract ...... …24 2.2 Introduction ...... …25 2.3 Material and methods ...... …26 2.3.1 Locations and samples ...... …26 2.3.2 Resin extractions and analysis ...... …27 2.3.3 Staurosporine quantification ...... …31 2.3.4 C. elegans feeding deterrence ...... …32 2.3.5 Brine shrimp lethality assay ...... …32 2.3.6 Bacterial community analysis ...... …33 2.3.7 Bacterial cultivation, identification, and metabolomics ...... …34 2.3.8 Metagenomics ...... …34 2.3.9 Phylogenetic analysis ...... …35 2.4 Results ...... …35 2.4.1 Sediment metabolomics ...... …35 2.4.2 Compound identification ...... …36

2.4.3 Linking compounds to community composition ...... …44 2.4.4 Staurosporine quantification ...... …47 2.4.5 Metagenomic analysis ...... …49 2.4.6 The ecological roles of staurosporine ...... …51 2.5 Discussion ...... …52 2.6 Acknowledgments ...... …56 2.7 References ...... …56 2.8 Appendix ...... …65

Chapter 3 Salinispora secondary metabolite mediated feeding deterrence ...... …69 3.1 Abstract ...... …70 3.2 Introduction ...... …71 3.3 Methods ...... …75 3.3.1 Salinispora strain selection and culturing ...... …75 3.3.2 C. elegans maintenance and sterilization ...... …76 3.3.3 Live Salinispora spp. assay setup ...... …77 3.3.4 Large scale cultivation and extraction ...... …79 3.3.5 Salinispora extract feeding deterrence of C. elegans ...... …79 3.3.6 Bioactivity-guided compound isolation ...... …80 3.3.7 Lomaiviticin C quantification and deterrence calculation ...... …81 3.3.8 Ophryotrocha habitat preference assay ...... …82 3.3.9 Marine nematode assays ...... …83 3.3.10 Statistical analyses ...... …83 3.4 Results ...... …84 3.4.1 Salinispora mediated C. elegans feeding preference ...... …84 3.4.2 Feeding assay using marine predators ...... …92 3.5 Discussion ...... …95 3.6 Acknowledgments ...... …99 3.7 References ...... …99 Chapter 4 Elucidating the Salinispora “inductome” through prokaryotic co-culturing ...... 104 4.1 Abstract ...... 105 4.2 Introduction ...... 106 4.3 Methods ...... 110 4.3.1 Bacterial selection and culturing ...... 110 4.3.2 Salinispora co-culturing ...... 111 4.3.3 LC MS/MS data acquisition ...... 113 4.3.4 Molecular networking ...... 114 4.3.5 Determination of upregulated mass features ...... 115 4.3.6 Siderophore induction assays ...... 116 4.4 Results ...... 118 4.4.1 Salinispora co-cultures ...... 118 4.4.2 Induced features determined by molecular networking ...... 118 4.4.3 Induced features determined by Metaboanalyst ...... 121 4.4.4 Siderophore upregulation in S. tropica strain CNB440 ...... 124 4.5 Discussion ...... 128 4.6 Acknowledgments ...... 131 4.7 References ...... 132

vii

Chapter 5 Concluding thoughts ...... 136 5.1 Concluding thoughts ...... 137 5.2 References ...... 144

viii

LIST OF FIGURES Chapter 2 Figure 2.1 Dereplication false discovery rate estimation ...... …31 Figure 2.2 Euler diagram comparing sediment metabolome composition ...... …37 Figure 2.3 Molecular network visualizing the sediment metabolomes from all 5 sites ...... …39 Figure 2.4 Compounds identified in the sediment metabolomes ...... …40 Figure 2.5 Retention time, mass, and mirror plot fragmentation spectral matching for the Standards of monactin, staurosporine, cocamidopropyl betaine, and erucamide ...... …44 Figure 2.6 Mirror plot showing the fragmentation spectrum for dynactin, trinactin, tetranactin, pheophytin, and smenospongidine ...... …45 Figure 2.7 Correlation between Salinispora and staurosporine abundance in marine sediments ...... …47 Figure 2.8 Staurosporine production by S. arenicola strain isolated ...... …47 Figure 2.9 Staurosporine detection from all 5 sediment sites ...... …49 Figure 2.10 C. elegans and brine shrimp assays ...... …53 Figure A2.1 Staurosporine standard curve including linear regression equation and R2 value ...... …65 Figure A2.2 Example biosynthetic gene clusters assembled from the site 2 metagenome ...... …67 Figure A2.3 Representative phylogenetic tree for cyanobacterial housekeeping genes found flanking cyanobacterial BGCs ...... …68

Chapter 3 Figure 3.1 Workflow for assays ...... …78 Figure 3.2 Plates containing live Salinispora spp. to test C. elegans feeding deterrence ...... …78 Figure 3.3 Salinispora / E. coli feeding preference assay ...... …85 Figure 3.4 Salinispora spp. feeding preference assay ...... …85 Figure 3.5 Salinispora spp. secreted metabolite mediated feeding assays ...... …87

Figure 3.6 Bioassay guided fractionation scheme leading to the isolation of lomaiviticins as the deterrent compounds in the S. tropica extracts ...... …89

Figure 3.7 Activity of the CNY012 fractions against C. elegans feeding assays and lomaiviticin C standard minimum deterrence calculation ...... …90

Figure 3.8 UV profile showing lomaiviticin presence in S. tropica extracts ...... …91 Figure 3.9 Quantification of lomaiviticin C in the S. tropica strains ...... …91 Figure 3.10 Results of Ophryotrocha habitat preference assay ...... …93 Figure 3.11 Survivorship curves for benthic marine nematodes on agar with Salinispora spp. extracts ...... …95

Chapter 4 Figure 4.1 Map of secondary metabolite pathways present in the 20 Salinispora strains used in the study ...... 112 Figure 4.2 Workflow for detecting specialized metabolite induction in Salinispora co-cultures ... .113 Figure 4.3 Molecular network consisting of features that were only found induced in at least two of three replicates ...... 119 Figure 4.4 Induced features per bacterial strain and Salinispora species determined by molecular networking...... 121 Figure 4.5 Heatmap depicting the average of each of the top 50 upregulated features in the CNB440 co-cultures and monocultures based on an ANOVA analysis ...... ……………125 Figure 4.6 Upregulation of desferrioxamines in S. tropica strain CNB440 in response to Streptomyces elicitors ...... 126 Figure 4.7 Mirror spectra plots comparing the fragmentation spectra from the GNPS library to the spectra from the CNB440 + supernatant extract for the amphiphilic DFOs...... 127

LIST OF TABLES Chapter 2 Table 2.1 Compound Identification from the Belizean sediment resin extracts ...... …41 Table 2.2 Staurosporine biological activity ...... …43 Table 2.3 BLAST analysis of actinomycete strain isolates from the sediments ...... …48 Table 2.4 Average staurosporine concentration per site, the percent of the sediment that was inorganic material, and the biologically relevant concentration of staurosporine ...... …49 Table A2.1 BLAST analysis of house keeping genes flanking eight biosynthetic gene clusters (BGCs) assembled from the metagenome ...... …66

Chapter 3 Table 3.1 Salinispora strains used for multi-strain plates ...... …76

Chapter 4 Table 4.1 Bacterial challenger strains and the rationale for why they were chosen ...... 113 Table 4.2 Mass features that were significantly upregulated in the co-cultures ...... 123

ACKNOWLEDGMENTS

There’s an African proverb, “It takes a village to raise a child”, that I believe should be amended to “It takes a village to raise a child… and a Ph.D. student”. Without that “village” I would never have come close to starting a Ph.D. let alone finishing one. I am eternally grateful for everyone who pushed me along and helped point me in the direction of the finish line.

First and foremost a huge thank you to Prof. Paul Jensen for taking a risk and inviting me into the lab. You were always there helping me along, throwing curveball projects my way, and believing in me even when I did not believe in myself. I would not be the scientist I am today without your constant guidance. You showed me the world of microbial chemical ecology, and natural products, something I will forever cherish.

Thank you, as well, to my committee, Prof. Pieter Dorrestein, Prof. Stuart Sandin, Prof.

Greg Rouse, and Prof. Chambers Hughes. Pieter, you opened your facilities to me, allowing me access to equipment vital for my research. Stuart, my understanding of statistics is in large part to you. Greg, thank you for all your help in isolating tiny critters and keeping them alive so I could develop my feeding deterrence assays. Also thank you for the wisdom and guidance you provided over the years, it was always refreshing to hear your insight. Chambers, you taught me so much about analytical chemistry and opened your lab and resources to me. I would be a fraction of the scientist I am today if it was not for your generosity and help.

I also could not have completed this degree without the help of every lab member that has cycled through the Jensen lab during my time at SIO. I especially need to thank Anne-Catrin

Letzel, Nadine Hughes, Kelley Gallagher, Leesa Habener, and Henrique Machado for sharing an office with me and constantly fielding my sometimes-asinine questions. Being able to freely communicate my science with knowledgeable and helpful lab mates has helped me immensely.

xii

Also, a huge thank you to Nastassia Patin for carpooling with me. These rides were some of the most insightful in my Ph.D., from figuring out life to dreaming up new projects. I also want to thank Doug Sweeney for being a great friend and lab mate. The conversations we had in the lab and over lunch were some of my favorite moments at SIO.

I could not have finished the majority of my projects without the help of the entire

Hughes lab, especially Gabriel Castro who taught me a lot of analytical chemistry techniques.

Thank you for being so patient and supportive while we spent hours trying to fix broken HPLCs so we could fractionate my samples. Also thank you to Grant and Trevor for being a constant source of amusement as well as a valuable resource when it came to anything chemistry related.

Thank you to the SIO community who welcomed me in and put up with me over the years, I don’t know how you did it. I have been lucky enough to form some amazing, indelible relationships with so many people at SIO; you have made my time in San Diego unforgettable.

Outside of SIO there were numerous people who coached me and helped me get through the emotional turmoil of a Ph.D. A huge thank you to my best friend, companion, and wife (as of a week ago!), MaiAda Carpano. You pushed me to pursue my dreams and constantly supported me in my decisions.

An enormous, gargantuan thanks to my mom, Leslie Tuttle, who from the time I dreamed of being a scientist in first grade believed in me and, sometimes stubbornly, pushed me to achieve everything I’ve done in my life. I would have never made it this far without your guidance and support. Also, a huge thanks to Gaeton, and Alex, you both listened to my practice talks, and read my papers giving me help when I needed it most.

Chapter 2 is currently being formatted into a manuscript that was submitted in September of 2018. Robert Tuttle, Alyssa M. Demko, Nastassia Patin, Cliff Kapono, Mohamed S. Donia,

xiii

Pieter Dorrestein, Paul R. Jensen. The dissertation author was the primary investigator on these studies.

Chapter 3 is currently being formatted into a manuscript that will be submitted in fall of

2018. Robert Tuttle, Katrina Forrest, Gabriel Castro, Chambers Hughes, Greg Rouse, Paul Jensen.

The dissertation author was the primary investigator on these studies.

Chapter 4 could not have been completed without the generous help of Fisher Price, an undergraduate at UCSD. Also, thanks to Leesa Habener for introducing me to the Metaboanlayst platform.

xiv

VITA

EDUCATION

2012 B.S., summa cum laude Major – Biological sciences Minor – Ecology Drexel University

2018 Ph.D. in Marine Biology Scripps Institution of Oceanography University of California San Diego RESEARCH EXPERIENCE

Ph.D. Student at University of California, San Diego 2013 - 2018 Scripps Institution of Oceanography, La Jolla, CA, USA Supervisor: Dr. Paul Jensen Thesis: The role of the marine actinomycete genus Salinispora in the chemical ecology of sediment communities

Microbiology / Natural Products Intern 2017 Sirenas Phamaceuticals Supervisor: Paige Stout Research Project: Helped develop the microbiology initiative at Sirenas, which focuses on specialized metabolite discovery from sediment bacteria

Sea Turtle Field Biologist 2012-2013 The Leatherback Trust, Playa Grande, Costa Rica Supervisor: Dr. James Spotilla Research Project: Monitored the beach for Leatherback, Olive Ridley, and Pacific Green sea turtles

Research Assistant 2012 WHOI Research Cruise NH1208 Supervisor: Dr. Aleck Wang Research Project: Analyzed carbonate chemistry of water samples collected from CTD casts in the Pacific ocean.

Research Assistant 2011-2012 Drexel University, Philadelphia, PA, USA Supervisor: Dr. Jacob Russell Research Project: Determined the presence of beneficial endosymbionts in pea aphids collected from farms in New York.

PUBLICATIONS

Amos GCA, Awakawa T, Tuttle RN, Letzel A-C, Kim MC, Kudo Y, et al. (2017). Comparative transcriptomics as a guide to natural product discovery and biosynthetic gene cluster functionality. Proc Natl Acad Sci 114: E11121–E11130.

Machado H, Tuttle RN, Jensen PR. (2017). Omics-based natural product discovery and the lexicon of genome mining. Curr Opin Microbiol 39: 136–142.

Petras D, Nothias L, Quinn RA, Alexandrov T, Tuttle RN, Bandeira N, Bouslimani A, et al. (2016). Mass Spectrometry-Based Visualization of Molecules Associated with Human Habitats. Anal Chem 88: 10775–10784.

Smith AH, Łukasik P, O'Connor MP, Lee A, Mayo G, Drott MT, Doll S, Tuttle RN, Disciullo RA, Messina A, Oliver KM, Russell JA (2015), Patterns, causes and consequences of defensive microbiome dynamics across multiple scales. Molecular Ecology, 24: 1135– 1149.

xvi

ABSTRACT OF THE DISSERTATION

The chemical ecology of marine bacteria and the sediments they inhabit

Robert Nicholas Tuttle

Doctor of Philosophy in Marine Biology

University of California San Diego, 2018

Professor Paul Jensen, Chair

Since the first studies looking into how plant specialized metabolites provide defense from herbivory, the field of chemical ecology has explored how organisms interact within their environment, and how these interactions structure the ecosystem. Over the last century, we have come to understand a lot about chemical signaling in plants and animals, but our understanding of bacterial chemical signaling is still in its infancy. This is partially because most of the focus has been placed on understanding bacteria potential for drug development as opposed to understanding their ecological roles. This thesis consists of five chapters and explores the interplay between bacterial specialized metabolites in marine sediments, their

xvii

producers, and their ecological roles, by focusing on the model marine bacterial genus

Salinispora. What follows are three research chapters preceded by an introduction to chemical ecology. The introduction focuses on our current understanding of marine bacterial chemical ecology, followed by a brief description of cutting edge mass spectrometry techniques that are now being exploited to further our understanding of the chemical landscape in the environment.

Chapter 2 summarizes a study concerning the identification of specialized metabolites in situ and the correlation between these metabolites and their potential producers.

Heterotrophic marine sediment bacteria are prolific producers of natural products, but surprisingly little is known about the compounds they produce in the environment and their effects on co-occurring microbes. Using mass spectrometry and molecular networking, characterization of the sediment metabolome was undertaken from Belizean reef habitats.

Dereplication results revealed numerous compounds could be detected directly from the sediments including synthetic, sponge, algal, and bacterial metabolites. Interestingly, one of these compounds, the cytotoxin staurosporine, was further quantified and found to occur in the sediments at abundances higher than those established to inhibit protein kinases as well as marine organisms. A 16S rRNA community analysis as well as culturing helped correlate the production of staurosporine to the obligate marine actinomycete species Salinispora arenicola.

These results indicate that microbial organisms are likely capable of producing cytotoxins in situ at appreciable quantities that likely impact the community structure of the sediment biome.

Since the cytotoxin staurosporine, which is produced by Salinispora, was found to be abundant in the sediments, chapter 3 explored the potential for Salinispora to deter predatory eukaryotic organisms. S. arenicola and S. tropica co-occur in the Caribbean and share greater than 99% 16s rRNA. A recent sequencing effort revealed that each species maintains a unique set BGCs, with S. arenicola containing the BGC to produce staurosporine and S. tropica the BGCs

xviii

to produce lomaiviticins, and salinosporamides. Numerous assays were developed to assess the ability of Salinispora spp. to deter feeding by the bacterivore C. elegans. Results indicated that

S. tropica strains can produce a suite of lomaiviticins that deter C. elegans feeding at ecologically relevant concentrations, however; S. arenicola strains do not produce deterrent allelochemicals at ecologically relevant concentrations. Follow up studies using more ecologically relevant organisms indicated that this trend was still prevalent when tested against the marine polychaete Ophryotrocha n sp. as well as marine nematodes.

In chapter 3, I saw that Salinispora spp. exhibit different chemodeterrent strategies against bacterivorous eukaryotes. However, this study relied on constitutively produced specialized metabolites, even though there are numerous examples where microbial specialized metabolites can be induced in response to biotic stressors in the environment. To address this chapter 4 focuses on a high throughput method I developed to look at induction of specialized metabolites in co-cultures containing Salinispora and marine bacterial challengers. Results indicate that induction is prevalent in cocultures using all Salinispora strains tested. The induced mass features were also unique to cocultures indicating that induction is strain and not species specific. Efforts to identify compounds that were induced in cocultures resulted in the identification of a suite of desferrioxamines that were upregulated in some Salinispora strains in response to specific Streptomyces challengers, indicating that some Salinispora strains can modulate the production of iron scavenging specialized metabolites in response to stressors.

Chapter 5 provides conclusions related to the thesis and how it fits into the broader context of marine microbial chemical ecology.

xix

CHAPTER 1

Introduction

1.1 Introduction to Chemical Ecology

Most organisms rely on chemosignaling to mediate interactions with each other and the environment (Paul and Ritson-Willliam, 2008; Hartmann, 2008). These interactions include fundamental ecological processes such as predation, communication, defense, larval recruitment, etc. (Paul and Ritson-Williams, 2008). The field of chemical ecology was born out of the desire to understand how these signals facilitate interactions between organisms in the environment.

Chemical ecology was pioneered by evolutionary and functional biologists in the mid to late 1800’s, who noticed that interactions between plants and herbivores seemed to be chemically mediated (Hartmann, 2008). However, much of their work was ignored for about a century, after which the study of chemical ecology, buoyed by developments in analytical techniques, started to flourish. Today thousands of chemically mediated interactions have been described, predominantly among macroorganisms (Hartmann, 2008). Some of the earliest work in the 20th century looked at sex pheromones, which have been shown to be integral to mating within insect species. One of the first examples of this was the pheromone driven recruitment of male conspecifics in certain moth species (Karlson and Butenandt 1959). The use of pheromones was then shown to be widespread among sexually reproducing animal phyla

(Takken and Dicke, 2006). Chemical signaling is not limited to interactions within species. Plants have been shown to possess a wealth of chemicals that aid in fitness and survival (Kappers et al., 2008; Ode, 2013). For example, geranium petals produce quisqualic acid which paralyzes predatory beetles (Popillia japonica) upon ingestion (Ranger et al., 2011). The effects of chemical signals on fitness has been realized across all domains of life from fungi to trees and from ants to elephants.

It has been well documented that microorganisms are also prolific producers of chemical signals. One of the earliest examples of this was the isolation of the antibiotic penicillin from the fungus Penicillium notatum by Alexander Flemming (Demain and Sanchez, 2009).

Flemming’s discovery ushered in a new era of chemical discovery from microorganisms known as the “Golden age of antibiotics”. Within a century, the understanding of how microbes interact within the environment had fundamentally changed (Raguso et al., 2015). Symbioses between fungus growing ants and microorganisms showed that co-opting microbial allelochemicals was not unique to humans, and that other organisms were capable of exploiting microbially derived compounds (Currie et al., 1999). Even plants were found to grow symbiotically with microbes that could confer resistance to soil borne plant pathogens. This symbiosis has been documented countless times in suppressive soils where crops are protected from pests due to bacterial allelopathy (Schlatter et al., 2017). One example is seen in the suppression of take-all disease in wheat by phenazine 1-carboxylic acid, a compound produced by Pseudomonas spp. in the rhizosphere (Thomashow et al., 1990). The interplay between organisms and their associated chemistry is well established in terrestrial ecosystems, but due to difficulties in accessing and studying marine organisms, the field of marine chemical ecology is still, comparatively, in its infancy.

1.2 Marine Chemical Ecology - Eukaryotes

Although marine chemical ecology is a relatively young field compared to its terrestrial counterpart, there is a wealth of information that has been elucidated. Chemical cues have been shown to help determine feeding, habitat, eukaryotic mating choices, as well as aid in defense across numerous taxonomic levels (Hay, 2009). The marine realm contains vast oligotrophic waters that require organisms to traverse great distances to find food sources. Organisms have evolved incredible chemosensory abilities to track down food sources in this environment

(DeBose and Nevitt, 2008). For instance numerous foraging birds that feed on zooplankton are known to sense areas of the ocean giving off high levels of dimethyl sulfide (DMS), a compound secreted in areas where zooplankton are grazing on phytoplankton (Nevitt and Bonadonna,

2005). However, some consumers, known as specialists, have specific food preferences and can be attracted to unique chemical signals given off only by their specific food source. For example, the crab Caphyra rotundifrons has been shown to feed on the alga Chlorodesmis fasrtigiata, which produces the cytotoxic diterpenoid chlorodesmin. Although this compound deters most herbivorous fish species, it serves as a chemoattractant to the crab (Hay et al., 1989).

When it comes to defense from consumers, the marine realm is teaming with examples of chemically mediated protection. Numerous organisms produce toxic or noxious compounds to deter predation (Hay, 1996). Ascidians are nonmotile, fleshy organisms that are rarely consumed by predators (Stoecker, 1980). Studies have shown that this is due to their production of noxious specialized metabolites (Paul et al., 1990; Pisut and Pawlik, 2002). Even phytoplankton have been shown to produce allelochemicals to suppress predation (Wolfe,

2000). For example, Phaeodactylum triconutum produce a suite of apo-fucoxanthinoids that deter predation from the copepod Tigriopus californicus (Shaw et al., 1995).

Proper habitat selection is imperative for organisms that undergo metamorphosis from larvae to nonmotile adults due to their need to settle in areas where they can either protect themselves, be close to potential mates, or find food sources. Some organisms, such as barnacles, rely on chemical cues released by established adults in order to find a suitable substrate (Dreanno et al., 2006). Coral and polychaete larvae are known to settle and even metamorphize in response to chemical conditions in the substrate (Pawlik, 1992). There is little known about what cues are responsible for larval recruitment; however, there is growing

evidence that microbial assemblages and their chemistry are responsible for the settlement of sponge and coral larvae (Pawlik, 1992; Whalan and Webster, 2014).

Similar to pheromones released by terrestrial organisms, chemical signals are also used to attract or find mates in the ocean (Bushmann and Atema, 2000; Hay, 2009; Snell and Morris,

1993). Copepods have been shown to increase swimming speed when they come across a female pheromone trail, swimming back and forth at rapid speeds until they locate a female, and attempt to mate (Bagøien and Kiørboe, 2005). Crustaceans also depend heavily on chemical signaling for mate attraction, where males gravitate towards sexually mature females due to pheromones sensed in the environment (Breithaupt and Thiel, 2011). These cues are so dominant that males have been shown trying to copulate with sponges soaked in minimal amounts of the stimulating compound (Hay, 2009).

1.3 Marine Chemical Ecology - Prokaryotes

As our ability to study the ocean has increased with the advent of more advanced technologies, we are beginning to understand that numerous eukaryotic interactions are actually mediated by bacterially derived specialized metabolites (Armstrong et al., 2001; Piel,

2009). Sponges are home to a consortium of bacteria that produce untold numbers of cytotoxins, protecting them from potential predators and infection (Hentschel et al., 2012; Piel,

2009). However, when comparing the number of bacterially derived specialized metabolites that have been elucidated to those ascribed to an ecological role, it is clear that there is still a lot to be learned.

Studying bacteria that are not associated with eukaryotes is much more difficult since many free living, heterotophic bacterial species occur at a low abundance in the environment, which has hampered our ability to figure out the ecological roles of their specialized metabolites.

Even though few marine bacterial secondary metabolites have been linked to ecological functions, many of those that have fall into several broad categories including: communication, nutrient sequestration, or antagonism (Hider and Kong, 2010; Straight and Kolter, 2009; Wietz et al., 2013).

Quorum sensing is a method by which bacteria use chemical signaling for intraspecies communication. It has been shown that quorum sensing mediates biofilm formation, secondary metabolite production, motility, and bioluminescence (Miller and Bassler, 2001). Bacteria also produce specialized metabolites to aid in acquiring nutrients that may be limited in the environment. Since bioavailable iron species typically occur at low concentrations in the marine realm, iron scavenging siderophores are quite common in bacteria (Hider and Kong, 2010).

Specialized metabolite mediated antagonism is the focus of this thesis and will be explored in more detail in the remaining section. Marine bacterial antagonistic natural products have been shown to work in numerous ways including: sequestering limited nutrients, inhibiting microbial organisms, or impeding eukaryote predation (Bernier and Surette, 2013; Jousset,

2012; Wietz et al., 2013). Therefore, bacterial community structure can be affected by the antagonistic compounds produced.

1.3.1 Nutrient limitation

Due to the density of microbes that occur in the ocean, competition for limiting nutrients can greatly affect community composition. Typically, oceans have low levels of bioavailable iron, creating demand for this limiting resource. In response, some microbes have evolved the ability to produce siderophores, which can act as extracellular iron sequestration compounds (Butler, 2005; Hider and Kong, 2010). Siderophore producing organisms are thus able to draw down iron to the point where other microbes are unable to grow (Hider and Kong,

2010). For instance, there is evidence that closely related Vibrio species are capable of deterring one another through nutrient competition (Pybus et al., 1994). A strain of V. anguillarum was shown to inhibit the growth of V. ordalii through siderophore production and the addition of bioavailable iron negated this effect (Pybus et al., 1994). Although numerous siderophores have been discovered from marine microbes, little is known about how these compounds specifically alter the community composition around their producers.

1.3.2 Prokaryote – prokaryote allelopathy

Bacteria also use secondary metabolites to directly inhibit other bacterial community members (Wietz et al., 2013). For instance, in biofilms, where bacterial densities are greatly increased in comparison to the surrounding environment, bacteria have developed specialized metabolites to inhibit other prokaryotes from settling. An example of this is seen in Phaeobacter sp., which produce indigoidine, a compound that inhibits biofilm formation in Vibrio fischeri

(Cude et al., 2012). Inhibition of quorum sensing through the production of “quorum- quenching” compounds also provides a route for microbes to compete with other bacterial species. For instance, tumonic acids produced by the cyanobacterial species Blennothrix cantharidosmum were shown to inhibit bioluminescence in Vibrio harveyi, a phenomena that is dependent on quorum sensing (Clark et al., 2008). Despite elucidating numerous roles that specialized metabolites can serve in prokaryote-prokaryote allelopathy, few bioactive metabolites have been linked to these roles, leaving an enormous number of compounds that serve unknown functions.

1.3.3 Prokaryote – Eukaryote allelopathy

There is little known about eukaryote predator deterrence mediated by secondary metabolites produced by marine bacteria. It has been well established that marine bacteria are

prolific producers of cytotoxic compounds, but few of these compounds have been tested for ecological function. Most of our current understanding stems from cyanobacteria, where numerous cytotoxins have been shown to deter grazing (Cruz-Rivera and Paul, 2007; Nagle and

Paul, 1998). For instance, the cyanobacterial species Lyngbya majuscule (Moorea producens) produces both malingamide and majusculamides which have been shown to deter pufferfish grazing (Pennings et al., 1996). However, heterotrophic marine bacteria can also produce compounds capable of deterring predation. Pseudomonas luteoviolacea for instance produces violacein, a purple alkaloid that inhibits protozoan feeding at sub-mM concentrations (Matz et al., 2008). Marine sediments are home to numerous secondary metabolite rich, heterotrophic bacteria that are capable of producing cytotoxic compounds. However, there has yet to be a study linking any of these compounds to potential ecological roles.

One limitation with studies elucidating the ecological roles of bacterial specialized metabolites is that they are typically conducted in laboratory settings. However, the production of these compounds in nature and how they impact the sediment community is poorly understood. In order to understand the effects of a bacterial secondary metabolites on the microbial community, a multifaceted approach needs to be considered including community analysis, metabolomics, and ecological testing of chemical compounds on members of the community.

1.4 Induction of bacterial specialized metabolites

Many bacterial chemical ecology studies are typically conducted axenically using unnatural culturing conditions which usually fail to consider environmental stressors that may impact specialized metabolism. The notion that stressors can induce compound production is something that is only starting to be understood (Abdelmohsen et al., 2015; Okada and

Seyedsayamdost, 2017). The advent of genomic sequencing has provided unparalleled amounts of information concerning the biosynthetic gene cluster content of bacteria. This has led to the understanding that numerous bacterial species possess more BGCs than compounds that have been isolated using traditional culturing techniques (Bentley et al., 2002; Janso and Carter, 2010;

Letzel et al., 2017; Nett et al., 2009). Molecular techniques have been implemented to express these orphan or cryptic BGCs in hosts that are more genetically tractable (Chiang et al., 2011); however, these techniques can be labor intensive and have limited success. Therefore, efforts are also being taken to use inducers that can turn on or upregulate specialized metabolite production (Bertrand et al., 2013; Okada and Seyedsayamdost, 2017; Seyedsayamdost, 2014;

Traxler et al., 2013). The use of chemical elicitors to stimulate specialized metabolite production has shown marked success, with high throughput screening methods being used to challenge bacterial species with hundreds to thousands of chemical elicitors at a time (Craney et al., 2012;

Seyedsayamdost, 2014). Typically these studies use small compound libraries which include compounds of non-biologic origin including synthetic compounds, heavy metals, rare earth metals, etc. (Abdelmohsen et al., 2015). For instance, screening of Streptomycetes against

30,569 chemical elicitors resulted in the identification of 19 induced compounds (Craney et al.,

2012).

Induction can also be brought about by manipulating the expression of biosynthetic gene clusters in a non-targeted manner. For example, ribosome engineering is a technique by which antibiotics that target the ribosome or RNA polymerase are added to a culture to create mutations in the cellular machinery resulting in upregulated expression of cryptic BGCs.

Successes have included the upregulation of actinorhodin from Streptomyces treated with rifampicin, due to a mutation in the gene encoding the β-subunit of the ribosome (Xu et al.,

2002).

Other techniques used to induce metabolite production look to the natural environment to find stressors. Introducing bacterial competitors and biologically derived small molecules to a traditional culturing paradigm has been shown to induce numerous specialized metabolites through various means such as cell-cell contact, nutrient supplementation, and small molecule interactions (Okada and Seyedsayamdost, 2017). The following section will explore successes from each of these examples.

1.4.1 Cell to cell contact

Induction of specialized metabolites as a result of cell to cell contact has been widely explored in the fungal realm, where numerous compounds have been found to be induced when fungi come into contact with bacteria (Cueto et al., 2001; Nützmann et al., 2011; Schroeckh et al., 2009). However, to a lesser extent, it has been shown that certain bacteria can regulate their metabolism when in contact with another organism. Recently, it was shown that when

Streptomyces come into contact with mycolic acid containing bacteria, they alter their biosynthesis of specialized metabolites (Onaka et al., 2011). However, these changes are not seen if cell free supernatant or chemical extracts are added to the Streptomyces culture. When coculturing Streptomyces lividans with T. pulmonis, phenotypic changes include the production of pigmented compounds as well as the novel antibiotic alchivemycin A (Onaka et al., 2011).

1.4.2 Nutrient supplementation

As mentioned previously, specialized metabolites can aid in nutrient acquisition.

Interestingly some of these compounds can be promiscuous, allowing for numerous bacterial species to uptake them (Okada and Seyedsayamdost, 2017). Some iron chelators or siderophores are capable of being used by multiple actinomycetes. These siderophores can provide the requisite nutrients to induce specialized metabolism within bacterial species even if

they were not the producers of the siderophore. For instance, the siderophore desferrioxamine

E produced by Streptomyces griseus was shown to enhance secondary metabolism and growth in Streptomyces tanashiensis (Yamanaka, 2005).

1.4.3 Small molecules

Since numerous bacteria are prolific producers of specialized metabolites, it is not surprising that these compounds can act as inducers. For instance, the oligopeptide goadsporin, produced by actinomycetes, has been shown to act as a general inducer of Streptomyces specialized metabolism as well as the production of the blue pigment actinorhodin (Igarashi et al., 2001). However, it is not always bacterial compounds that can act as inducers; fungal compounds have been shown to induce the production of antifungal compounds in bacteria.

This can be seen in a Streptomyces sp. that overproduces the antifungal compound irumamycin in response to heat-killed fungi addition (Fourati-Ben Fguira et al., 2008).

Overall, there appears to be a plethora of unrealized biosynthetic potential in microbes.

Given that traditional culturing rarely introduces environmental factors typically realized by bacteria in their native habitats, it is of no surprise that bringing the environment into the laboratory has led to the isolation of numerous novel metabolites. However, these studies have barely scratched the surface, and numerous prolific natural product producers have not been subjected to any sort of coculturing studies. Some sediment actinomycetes have been shown to be prolific natural product producers; however, there has been little effort to understand the ecological roles of secondary metabolites and how that can be affected by biotic stressors (Wietz et al., 2013).

1.5 Sediment bacteria – Focusing on Salinispora

Marine sediment environments are dynamic habitats that contain an extraordinary amount of biomass. The microbial abundance in a gram of sand is in the magnitude of 1011 cells

(Alongi, 1988), which helps support numerous bacterivorous eukaryotes including polychaetes, nematodes, and protozoans (Andresen and Kristensen, 2002; Moens et al., 1999; Pernthaler,

2005). Also, shallow benthic regions have drawn analogies to the rainforest biome because they promote extensive biodiversity, which has led to the evolution of many species that are stuck in a constant arms race to maintain their niche (Bouchet et al., 2002; Reaka-Kudla et al., 1996; Vos,

2009). For instance, individual sand grains are thought to harbor anywhere from 3,000-6,000 unique bacterial species, with species distributions varying greatly from grain to grain (Probandt et al., 2018). This structure is thought to be maintained partially due to the production of antagonistic secondary metabolites by microorganisms (Patin et al., 2015). However, very little is known about the interactions mediated by sediment bacteria, partially due to the difficulty in finding suitable organisms to conduct studies.The marine obligate actinomycete genus

Salinispora provides a useful model to address the ecological functions of secondary metabolites from heterotrophic sediment bacteria. So far three Salinispora species have been named,

S. tropica, S. arenicola, and S. pacifica. All three species are prolific producers of secondary metabolites (Jensen et al., 2015, 2007). Compounds have been isolated from Salinispora spp. that exhibit antimicrobial properties, such as rifamycin, as well as compounds with cytotoxic properties, such as the staurosporines, lymphostin, lomaiviticins, and salinasporamides (Jensen et al., 2015). Most of these compounds have been tested for pharmaceutically relevant activity, but there have been few efforts to address the possible ecological roles that these compounds may serve. Interestingly, S. tropica and S. arenicola seem to have diverged despite co-occurring in tropical sediments and containing >99% 16S rRNA identity, thus indicating that they must be

ecologically distinct. One reason may be that they represent different ecologically cohesive units or ecotypes, and therefore inhabit different niches within the sediment (Cohan, 2002). Previous studies have shown that when it comes to prokaryotic competition, S. arenicola participates in antagonistic allelopathy by producing antibiotics, whereas S. tropica typically grows quicker and competes through resource limitation (Patin et al., 2015). This indicates that these closely related species are differentiated due to ecological specialization defined by their secondary metabolism.

A recent large scale sequencing effort of 119 Salinispora strains from all three species was undertaken and an analysis of the Salinispora pan genome has allowed the enumeration of over 170 unique secondary metabolite pathways (Letzel et al., 2017; Ziemert et al., 2014).

However, this number is markedly higher than the number of compounds that have been isolated from Salinispora, indicating that this species has additional biosynthetic potential. Efforts to heterologously express these pathways have met with some success (Bonet et al., 2015), however ecological approaches could yield new compounds and provide insight into their functional roles in marine sediments.

There are several reasons for the discrepancy between the number of BGCs within the

Salinispora genus compared to the number of specialized metabolites that have been identified.

Very little is known about the transcription and translation of these BGCs, which means that the compounds they code for may never be produced. Also, limitations in analytical techniques may hinder our abilities to discover these specialized metabolites. However, thanks to recent developments in analytical techniques, specifically in the field of mass spectrometry, these limitations are now being overcome. The following section will cover some of these advancement in the field of mass spectrometry.

1.6 Advancements in Mass Spectrometry

The difficulty in finding quick and easy ways to dereplicate metabolites from complex chemical mixtures has plagued the natural product world for many years. The ability to simply identify what molecules are present in a rather robust and timely manner is only now in its infancy. Partially, this was due to lack of public databases containing mass information, as well as lack of sensitive machinery that could process complex samples (Bouslimani et al., 2014).

However, the last 20 years have seen numerous advances in the field of mass spectrometry that are paving the way for detection of single molecule compounds (Naik et al., 2009) at an accuracy of 1 part per billion (much lower than the accepted 5 ppm accuracy for mass detection)

(Bouslimani et al., 2014). Although high mass accuracy at a single molecule detection limit may not be the standard used in most labs, the current capabilities of mass spectrometers are allowing for detection of minute compounds in complex mixtures at below 5 ppm accuracy. This technology is now being exploited to generate pipelines that can quickly compare and dereplicate samples (Wang et al., 2016). Using these tools, one can quickly figure out what features are present within samples and how samples vary from one another.

The advent of mass spectrometry has paved way for the collection of massive amounts of spectra from relatively complex samples. One of the current bottlenecks is how to compare these enormous datasets (Bouslimani et al., 2014). For instance, in human metabolomics, the difference between a disease state and healthy state may be signified by a few metabolites.

However, analyzing samples from healthy and diseased patients may generate thousands to millions of MS features (Martinez et al., 2017). Therefore, emphasis has been placed on developing statistical programs to quickly sift through MS chromatograms to identify novel or unique MS features. One such program, entitled Metaboanalyst, was generated to help process this wealth of data, run numerous statistical analyses, and provide visualization tools to help

interpret complex mass spec-based datasets. The need for these sort of pipelines is made obvious by the use that Metaboanalyst receives; ~1.8 million jobs in the last 12 months (Chong et al., 2018).

However, most mass spectrometers can also provide unique structural information about ionized compounds by subjecting them to multiple rounds of fragmentation. Only recently has the use of MSn been incorporated into analysis pipelines to facilitate dereplication.

Where MS1 can provide information about the relative quantity or presence / absence of a unique feature in a sample, MS2 can further provide information about the structural elements and chemical relatedness of two entities. The principle exploited by these analyses is that compounds that possess similar chemical structures should fragment in a relatively analogous manner (Kind and Fiehn, 2010). Therefore, by comparing fragmentation spectra, connections can be drawn between the chemical relatedness of different features (Wang et al., 2016).

However, these sorts of analyses can be time consuming and computationally intensive.

Programs are being developed to circumvent these issues, including MS-Cluster which condenses identical ions thus deconvoluting large datasets (Frank et al., 2008). The Global

Natural Products Social Molecular Networking pipeline (GNPS) synthesizes these programs elegantly into an online package that helps in the generation of molecular networks, or networks that group compounds based on structural similarity as determined by MS/MS fragmentation

(Wang et al., 2016). Combining this network with a MS/MS fragmentation library has allowed for a relatively quick way to dereplicate a dataset. This allows a higher degree of confidence when it comes to compound identification since the accurate mass and mass fragments are used in tandem to verify compound identity. Also, these networks can be used to identify analogues of known compounds, since they produce similar fragmentation spectra and will network closely together (Wang et al., 2016).

1.7 Summary of Thesis

In my thesis, I aim to provide new insight into the chemical ecology of Salinispora secondary metabolism using these mass spectrometry-based approaches. For the first research chapter (chapter 2), I examined the tropical benthic metabolome to identify specialized metabolites in situ. This chapter is largely in the form of a manuscript that has been prepared and submitted for publication. I then linked these metabolites to potential producers using metagenomics and a 16S rRNA community analysis. Interestingly cytotoxic specialized metabolites produced by Salinispora were dereplicated from the sediment metabolome, indicating that they may be serving as a feeding deterrent. Due to these findings, I decided to look at the effects of the Salinispora metabolome on predation for chapter 3. Both S. arenicola and S. tropica were tested, revealing that S. tropica compounds deterred feeding whereas S. arenicola extracts actually acted as chemoattractants. Bioassay guided fractionation indicated that the cytotoxin lomaiviticin was the main deterrent in the S. tropica extracts.

One of the caveats with classic chemical ecology studies using bacteria is that culturing typically does not reflect interactions realized in the environment. Therefore, for chapter 4, I decided to test the effects of biotic stressors on Salinispora natural product production. Large scale genome sequencing efforts have revealed that Salinispora contain more BGCs than specialized metabolites seen under typical culturing efforts. Therefore, I cocultured Salinispora with ecologically relevant bacterial stressors to see the effects on the overall “inductome” of

Salinispora cocultures. Results of these cocultures indicated that induction is common in cocultures using Salinispora, and that each coculture condition appears to induce unique mass features, indicating strain specificity when it comes to induction. Desferrioxamines were dereplicated in cocultures, and further studies showed that these siderophores were upregulated when S. tropica strain CNB440 was treated with Streptomyces.

1.8 References Abdelmohsen UR, Grkovic T, Balasubramanian S, Kamel MS, Quinn RJ, Hentschel U. 2015. Elicitation of secondary metabolism in actinomycetes. Biotechnol Adv 33:798–811. doi:10.1016/j.biotechadv.2015.06.003 Andresen M, Kristensen E. 2002. The importance of bacteria and microalgae in the diet of the deposit-feeding polychaete Arenicola marina. Ophelia 56:179–196. doi:10.1080/00785236.2002.10409498 Armstrong E, Yan L, Boyd KG, Wright PC, Burgess JG. 2001. The symbiotic role of marine microbes on living surfacesHydrobiologia. pp. 37–40. doi:10.1023/A:1012756913566 Bagøien E, Kiørboe T. 2005. Blind dating - Mate finding in planktonic copepods. I. Tracking the pheromone trail of Centropages typicus. Mar Ecol Prog Ser 300:105–115. doi:10.3354/meps300105 Bentley S, Chater K, Cerdeño-Tárraga A-M, Challis GL, Thomson NR, James KD, Harris DE, Quail M a, Kieser H, Harper D, Bateman A, Brown S, Chandra G, Chen CW, Collins M, Cronin A, Fraser A, Goble A, Hidalgo J, Hornsby T, Howarth S, Huang C-H, Kieser T, Larke L, et. al. 2002. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417:141–147. doi:10.1038/417141a Bernier SP, Surette MG. 2013. Concentration-dependent activity of antibiotics in natural environments. Front Microbiol 4:1–14. doi:10.3389/fmicb.2013.00020 Bertrand S, Schumpp O, Bohni N, Bujard A, Azzollini A, Monod M, Gindro K, Wolfender JL. 2013. Detection of metabolite induction in fungal co-cultures on solid media by high-throughput differential ultra-high pressure liquid chromatography-time-of-flight mass spectrometry fingerprinting. J Chromatogr A 1292:219–228. doi:10.1016/j.chroma.2013.01.098 Bonet B, Teufel R, Crüsemann M, Ziemert N, Moore BS. 2015. Direct Capture and Heterologous Expression of Salinispora Natural Product Genes for the Biosynthesis of Enterocin. J Nat Prod 78:539–542. doi:10.1021/np500664q Bouchet P, Lozouet P, Maestrati P, Heros V. 2002. Assessing the magnitude of species richness in tropical marine environments: exceptionally high numbers of molluscs at a New Caledonia site. Biol J Linn Soc 75:421–436. doi:10.1046/j.1095-8312.2002.00052.x Bouslimani A, Sanchez LM, Garg N, Dorrestein PC. 2014. Mass spectrometry of natural products: current, emerging and future technologies. Nat Prod Rep 31:718. doi:10.1039/c4np00044g Breithaupt T, Thiel M. 2011. Chemical Communication in Crustaceans, Chemical Communication in Crustaceans. New York, NY: Springer New York. doi:10.1007/978-0-387-77101-4 Bushmann PJ, Atema J. 2000. Chemically mediated mate location and evaluation ln the lobster, Homarus americanus. J Chem Ecol 26(4):883–899. Butler A. 2005. Marine Siderophores and Microbial Iron Mobilization. BioMetals 18:369–374. doi:10.1007/s10534-005-3711-0 Chiang Y-M, Chang S-L, Oakley BR, Wang CC. 2011. Recent advances in awakening silent biosynthetic gene clusters and linking orphan clusters to natural products in

microorganisms. Curr Opin Chem Biol 15:137–143. doi:10.1016/j.cbpa.2010.10.011 Chong J, Soufan O, Li C, Caraus I, Li S, Bourque G, Wishart DS, Xia J. 2018. MetaboAnalyst 4.0: towards more transparent and integrative metabolomics analysis. Nucleic Acids Res gky310. doi:10.1093/nar/gky310 Clark BR, Engene N, Teasdale ME, Rowley DC, Matainaho T, Valeriote FA, Gerwick WH. 2008. Natural Products Chemistry and Taxonomy of the Marine Cyanobacterium Blennothrix cantharidosmum. J Nat Prod 71:1530–1537. doi:10.1021/np800088a Cohan FM. 2002. What are Bacterial Species? Annu Rev Microbiol 56:457–487. doi:10.1146/annurev.micro.56.012302.160634 Craney A, Ozimok C, Pimentel-Elardo SM, Capretta A, Nodwell JR. 2012. Chemical perturbation of secondary metabolism demonstrates important links to primary metabolism. Chem Biol 19:1020–1027. doi:10.1016/j.chembiol.2012.06.013 Cruz-Rivera E, Paul VJ. 2007. Chemical deterrence of a cyanobacterial metabolite against generalized and specialized grazers. J Chem Ecol 33:213–217. doi:10.1007/s10886-006- 9212-y Cude WN, Mooney J, Tavanaei AA, Hadden MK, Frank AM, Gulvik CA, May AL, Buchan A. 2012. Production of the Antimicrobial Secondary Metabolite Indigoidine Contributes to Competitive Surface Colonization by the Marine Roseobacter Phaeobacter sp. Strain Y4I. Appl Environ Microbiol 78:4771–4780. doi:10.1128/AEM.00297-12 Cueto M, Jensen PR, Kauffman C, Fenical W, Lobkovsky E, Clardy J. 2001. Pestalone, a new antibiotic produced by a marine fungus in response to bacterial challenge. J Nat Prod 64:1444–1446. doi:10.1021/np0102713 Currie CR, Scottt JA, Summerbell RC, Malloch D. 1999. Fungus-growing ants use antibiotic- producing bacteria to control garden parasites. Nature. doi:10.1038/19519 DeBose J, Nevitt G. 2008. The use of Odors at Different Spatial Scales. J Chem Ecol 34:86. Demain AL, Sanchez S. 2009. Microbial drug discovery: 80 years of progress. J Antibiot (Tokyo) 62:5–16. doi:10.1038/ja.2008.16 Dreanno C, Kirby RR, Clare AS. 2006. Locating the barnacle settlement pheromone: Spatial and ontogenetic expression of the settlement-inducing protein complex of Balanus amphitrite. Proc R Soc B Biol Sci 273:2721–2728. doi:10.1098/rspb.2006.3649 Fourati-Ben Fguira L, Smaoui S, Karray-Rebai I, Bejar S, Mellouli L. 2008. The antifungal activity of the terrestrial Streptomyces US80 strain is induced by heat-killed fungi. Biotechnol J 3:1058–1066. doi:10.1002/biot.200700155 Frank AM, Bandeira N, Shen Z, Tanner S, Briggs SP, Smith RD, Pevzner P a. 2008. Clustering Millions of Tandem Mass Spectra research articles. J proteome Res 113–122. doi:10.1021/pr070361e Hartmann T. 2008. The lost origin of chemical ecology in the late 19th century. Proc Natl Acad Sci. doi:10.1073/pnas.0709231105 Hay ME. 2009. Marine Chemical Ecology: Chemical Signals and Cues Structure Marine

Populations, Communities, and Ecosystems. Ann Rev Mar Sci 1:193–212. doi:10.1146/annurev.marine.010908.163708 Hay ME. 1996. Marine chemical ecology: What’s known and what’s next? J Exp Mar Bio Ecol. doi:10.1016/S0022-0981(96)02659-7 Hay ME, Pawlik JR, Duffy JE, Fenical W. 1989. Seaweed-herbivore-predator interactions: host- plant specialization reduces predation on small herbivores. Oecologia 81:418–427. doi:10.1007/BF00377093 Hentschel U, Piel J, Degnan SM, Taylor MW. 2012. Genomic insights into the marine sponge microbiome. Nat Rev Microbiol. doi:10.1038/nrmicro2839 Hider RC, Kong X. 2010. Chemistry and biology of siderophores. Nat Prod Rep 27:637. doi:10.1039/b906679a Igarashi Y, Kan Y, Fujii K, Fujita T, Harada K, Naoki H, Tabata H, Onaka H, Furumai T. 2001. Goadsporin, a chemical substance which promotes secondary metabolism and Morphogenesis in streptomycetes. II. Structure determination. J Antibiot (Tokyo) 54:1045– 53. doi:10.7164/antibiotics.54.1036 Janso JE, Carter GT. 2010. Biosynthetic Potential of Phylogenetically Unique Endophytic Actinomycetes from Tropical Plants. Appl Environ Microbiol 76:4377–4386. doi:10.1128/AEM.02959-09 Jensen PR, Moore BS, Fenical W. 2015. The marine actinomycete genus Salinispora: a model organism for secondary metabolite discovery. Nat Prod Rep 00:1–14. doi:10.1039/c4np00167b Jensen PR, Williams PG, Oh DC, Zeigler L, Fenical W. 2007. Species-specific secondary metabolite production in marine actinomycetes of the genus Salinispora. Appl Environ Microbiol 73:1146–1152. doi:10.1128/AEM.01891-06 Jousset A. 2012. Ecological and evolutive implications of bacterial defences against predators. Environ Microbiol 14:1830–1843. doi:10.1111/j.1462-2920.2011.02627.x Kappers IF, Dicke M, Bouwmeester HJ. 2008. Terpenoids in Plant Signaling: Chemical EcologyWiley Encyclopedia of Chemical Biology. Hoboken, NJ, USA: John Wiley & Sons, Inc. doi:10.1002/9780470048672.wecb652 Kind T, Fiehn O. 2010. Advances in structure elucidation of small molecules using mass spectrometry. Bioanal Rev 2:23–60. doi:10.1007/s12566-010-0015-9 Letzel A-C, Li J, Amos GCA, Millán-Aguiñaga N, Ginigini J, Abdelmohsen UR, Gaudêncio SP, Ziemert N, Moore BS, Jensen PR. 2017. Genomic insights into specialized metabolism in the marine actinomycete Salinispora. Environ Microbiol 19:3660–3673. doi:10.1111/1462- 2920.13867 Martinez KB, Leone V, Chang EB. 2017. Microbial metabolites in health and disease: Navigating the unknown in search of function. J Biol Chem 292:8553–8559. doi:10.1074/jbc.R116.752899 Matz C, Webb JS, Schupp PJ, Phang SY, Penesyan A, Egan S, Steinberg P, Kjelleberg S. 2008.

Marine Biofilm Bacteria Evade Eukaryotic Predation by Targeted Chemical Defense. PLoS One 3:e2744. doi:10.1371/journal.pone.0002744 Miller MB, Bassler BL. 2001. Quorum Sensing in Bacteria. Annu Rev Microbiol 55:165–199. doi:10.1146/annurev.micro.55.1.165 Moens T, Verbeeck L, De Maeyer A, Swings J, Vincx M. 1999. Selective attraction of marine bacterivorous nematodes to their bacterial food. Mar Ecol Prog Ser 176:165–178. doi:10.3354/meps176165 Nagle DG, Paul VJ. 1998. Chemical defense of a marine cyanobacterial bloom. J Exp Mar Bio Ecol 225:29–38. doi:10.1016/S0022-0981(97)00205-0 Naik AK, Hanay MS, Hiebert WK, Feng XL, Roukes ML. 2009. Towards single-molecule nanomechanical mass spectrometry. Nat Nanotechnol 4:445–450. doi:10.1038/nnano.2009.152 Nett M, Ikeda H, Moore BS. 2009. Genomic basis for natural product biosynthetic diversity in the actinomycetes. Nat Prod Rep. doi:10.1039/b817069j Nevitt GA, Bonadonna F. 2005. Sensitivity to dimethyl sulphide suggests a mechanism for olfactory navigation by seabirds. Biol Lett 1:303–305. doi:10.1098/rsbl.2005.0350 Nützmann H, Reyes-dominguez Y, Scherlach K, Schroeckh V, Horn F. 2011. Bacteria-induced natural product formation in the fungus Aspergillus nidulans requires Saga / Ada-mediated histone acetylation. Proc Natl Acad Sci 108:14282–14287. doi:10.1073/pnas.1103523108/- /DCSupplemental.www.pnas.org/cgi/doi/10.1073/pnas.1103523108 Ode PJ. 2013. Plant Defences and Parasitoid Chemical EcologyChemical Ecology of Insect Parasitoids. pp. 9–36. doi:10.1002/9781118409589.ch2 Okada BK, Seyedsayamdost MR. 2017. Antibiotic dialogues: induction of silent biosynthetic gene clusters by exogenous small molecules. FEMS Microbiol Rev 41:19–33. doi:10.1093/femsre/fuw035 Onaka H, Mori Y, Igarashi Y, Furumai T. 2011. Mycolic Acid-Containing Bacteria Induce Natural- Product Biosynthesis in Streptomyces Species. Appl Environ Microbiol 77:400–406. doi:10.1128/AEM.01337-10 Patin N V, Duncan KR, Dorrestein PC, Jensen PR. 2015. Competitive strategies differentiate closely related species of marine actinobacteria. ISME J 10:1–13. doi:10.1038/ismej.2015.128 Paul V, Lindquist N, Fenical W. 1990. Chemical defenses of the tropical ascidian Atapozoa sp. and its nudibranch predators Nembrotha spp. Mar Ecol Prog Ser 59:109–118. doi:10.3354/meps059109 Paul VJ, Ritson-Williams R. 2008. Marine chemical ecology. Nat Prod Rep. doi:10.1039/b702742g Pawlik JR. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Ocean Mar Biol Annu Rev 30:273–335. Pennings SC, Weiss AM, Paul VJ. 1996. Secondary metabolites of the cyanobacterium Microcoleus lyngbyaceus and the sea hare Stylocheilus longicauda: palatability and

toxicity. Mar Biol 126:735–743. doi:10.1007/BF00351340 Pernthaler J. 2005. Predation on prokaryotes in the water column and its ecological implications. Nat Rev Microbiol 3:537–546. doi:10.1038/nrmicro1252 Piel J. 2009. Metabolites from symbiotic bacteria. Nat Prod Rep. doi:10.1039/b703499g Pisut DP, Pawlik JR. 2002. Anti-predatory chemical defenses of ascidians: Secondary metabolites or inorganic acids? J Exp Mar Bio Ecol 270:203–214. doi:10.1016/S0022-0981(02)00023-0 Probandt D, Eickhorst T, Ellrott A, Amann R, Knittel K. 2018. Microbial life on a sand grain: From bulk sediment to single grains. ISME J 12:623–633. doi:10.1038/ismej.2017.197 Pybus V, Loutit MW, Lamont IL, Tagg JR. 1994. Growth inhibition of the salmon pathogen Vibrio ordalii by a siderophore produced by Vibrio anguillarum strain VL4355. J Fish Dis 17:311– 324. doi:10.1111/j.1365-2761.1994.tb00227.x Raguso RA, Agrawal AA, Douglas AE, Jander G, Kessler A, Poveda K, Thaler JS. 2015. The raison d’être of chemical ecology. Ecology 96:617–630. doi:10.1890/14-1474.1 Ranger CM, Winter RE, Singh AP, Reding ME, Frantz JM, Locke JC, Krause CR. 2011. Rare excitatory amino acid from flowers of zonal geranium responsible for paralyzing the Japanese beetle. Proc Natl Acad Sci 108:1217–1221. doi:10.1073/pnas.1013497108 Reaka-Kudla ML, Wilson DE, Wilson EO. 1996. Biodiversity II: understanding and protecting our biological resources. Joseph Henry Press. Schlatter D, Kinkel L, Thomashow L, Weller D, Paulitz T. 2017. Disease Suppressive Soils: New Insights from the Soil Microbiome. Phytopathology 107:1284–1297. doi:10.1094/PHYTO- 03-17-0111-RVW Schroeckh V, Scherlach K, Nutzmann H-W, Shelest E, Schmidt-Heck W, Schuemann J, Martin K, Hertweck C, Brakhage AA. 2009. Intimate bacterial-fungal interaction triggers biosynthesis of archetypal polyketides in Aspergillus nidulans. Proc Natl Acad Sci 106:14558–14563. doi:10.1073/pnas.0901870106 Seyedsayamdost MR. 2014. High-throughput platform for the discovery of elicitors of silent bacterial gene clusters. Proc Natl Acad Sci 111:7266–7271. doi:10.1073/pnas.1400019111 Shaw BA, Andersen RJ, Harrison PJ. 1995. Feeding deterrence properties of apo-fucoxanthinoids from marine diatoms. I. Chemical structures of apo-fucoxanthinoids produced by Phaeodactylum tricornutum. Mar Biol 124:467–472. doi:10.1007/BF00363921 Snell TW, Morris PD. 1993. Sexual communication in copepods and rotifers. Hydrobiologia 255– 256:109–116. doi:10.1007/BF00025828 Stoecker D. 1980. Chemical Defenses of Ascidians Against Predators. Ecology 61:1327–1334. doi:10.2307/1939041 Straight PD, Kolter R. 2009. Interspecies Chemical Communication in Bacterial Development. Annu Rev Microbiol 63:99–118. doi:10.1146/annurev.micro.091208.073248 Takken W, Dicke M. 2006. Chemical Ecology: a multidisciplinary approachChemical Ecology: From Gene to Ecosystem. pp. 1–8.

Thomashow LS, Weller DM, Bonsall RF, Pierson LS. 1990. Production of the antibiotic phenazine- 1-carboxylic acid by fluorescent Pseudomonas species in the rhizosphere of wheat. Appl Environ Microbiol 56:908–912. Traxler MF, Watrous JD, Alexandrov T, Dorrestein PC, Kolter R. 2013. Interspecies Interactions Stimulate Diversification of the Streptomyces coelicolor Secreted Metabolome. MBio 4:e00459-13. doi:10.1128/mBio.00459-13 Vos M. 2009. Why do bacteria engage in homologous recombination? Trends Microbiol 17:226– 32. doi:10.1016/j.tim.2009.03.001 Wang M, Carver JJ, Phelan V V, Sanchez LM, Garg N, Peng Y, Nguyen DD, Watrous J, Kapono CA, Luzzatto-Knaan T, Porto C, Bouslimani A, Melnik A V, Meehan MJ, Liu W-T, Crüsemann M, Boudreau PD, Esquenazi E, Sandoval-Calderón M, Kersten RD, Knight R, Jensen PR, Palsson BØ, Pogliano K, Linington RG, Gutiérrez M, Lopes NP, Gerwick WH, Moore BS, Dorrestein PC, Bandeira N, et. al. 2016. Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking. Nat Biotechnol 34:828–837. doi:10.1038/nbt.3597 Whalan S, Webster NS. 2014. Sponge larval settlement cues: The role of microbial biofilms in a warming ocean. Sci Rep 4. doi:10.1038/srep04072 Wietz M, Duncan K, Patin N V., Jensen PR. 2013. Antagonistic Interactions Mediated by Marine Bacteria: The Role of Small Molecules. J Chem Ecol 39:879–891. doi:10.1007/s10886-013- 0316-x Wolfe G V. 2000. The chemical defense ecology of marine unicellular plankton: Constraints, mechanisms, and impacts. Biol Bull 198:225–244. doi:10.2307/1542526 Xu J, Tozawa Y, Lai C, Hayashi H, Ochi K. 2002. A rifampicin resistance mutation in the rpoB gene confers ppGpp-independent antibiotic production in Streptomyces coelicolor A3(2). Mol Genet Genomics 268:179–189. doi:10.1007/s00438-002-0730-1 Yamanaka K. 2005. Desferrioxamine E produced by Streptomyces griseus stimulates growth and development of Streptomyces tanashiensis. Microbiology 151:2899–2905. doi:10.1099/mic.0.28139-0 Ziemert N, Lechner A, Wietz M, Millan-Aguinaga N, Chavarria KL, Jensen PR. 2014. Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci 111:E1130–E1139. doi:10.1073/pnas.1324161111

Chapter 2

The Detection of Specialized Metabolites and Their Producers in Ocean Sediments

2.1 Abstract

Thousands of specialized metabolites have been identified from cultured microorganisms yet evidence of their production in nature has proven elusive. Technological advances in mass spectrometry combined with public databases are providing new opportunities for the rapid identification of compounds from complex mixtures derived directly from environmental samples. Here we used adsorbent resins, tandem mass spectrometry, and next generation sequencing to provide a first glimpse into the metabolome of marine sediments and its relationship to bacterial community structure as well as its potential for natural product biosynthesis. We identified specialized metabolites previously reported from bacteria, providing evidence that these compounds are produced in situ. We also identified a compound previously reported from a marine sponge, suggesting it may be of microbial origin, and compounds of anthropogenic origin, suggesting this approach can be used as an indicator of environmental impact. The bacterial metabolite staurosporine was quantified and shown to reach physiologically relevant concentrations in marine sediments, suggesting that microbial specialized metabolites can influence community structure and predator-prey relationships.

Staurosporine concentrations correlated with the relative abundance of the staurosporine producing bacterial genus Salinispora with production confirmed in strains cultured from the same location providing links between compounds and candidate producers. These results provide evidence that coupling environmental metabolomics with community analyses can reveal relationships between specialized metabolites, the organisms that produce them, and their roles in shaping community structure.

2.2 Introduction

Bacterial natural products account for many of today’s most useful medicines (Bérdy,

2005). These “specialized metabolites” are traditionally isolated from bacteria cultured in the laboratory, leaving us with little understanding of when, why, and at what concentrations they are produced in nature (Davies, 2006). Laboratory culture conditions also bear little resemblance to those experienced by microbes in nature and thus may lack cues required to induce specialized metabolite biosynthesis. This possibility is supported by the large number of

“orphan” biosynthetic gene clusters (BGCs) detected in bacterial genomes (Letzel et al., 2017;

Milshteyn et al., 2014; Rutledge and Challis, 2015) and mounting evidence that interactions with other microbes induce specialized metabolite production (Burgess et al., 1999; Luti and

Mavituna, 2011; Patin et al., 2018; Slattery et al., 2001; Traxler et al., 2013; Trischman et al.,

2004).

Relative to the thousands of natural products discovered from cultured bacteria, few have been detected directly in nature. Exceptions include compounds reported from marine invertebrates that were subsequently shown to be of microbial origin (König et al., 2006; Wilson et al., 2014). Suppressive soils provide another example (Mazzola, 2002; Raaijmakers et al.,

2002). In this case, bacterial metabolites have been detected in soils and further linked to the inhibition of plant pathogens (Raaijmakers et al., 1999; Thrane et al., 2000), providing a rare example in which ecological effects have been assigned. The natural ability of soil microbes to suppress plant pathogens via the production of biologically active natural products is an essential component of healthy soil ecosystems, contributes to the success of commercial agriculture, and demonstrates that these compounds can affect community composition.

Microbial natural products have also been detected in the context of monitoring marine toxins

(Lane et al., 2010). Here, polyaromatic resins were used to extract water-borne toxins such as

domoic acid (Kudela, 2011; Lane et al., 2010; Takahashi et al., 2007). While this approach has been further developed to study allelopathy among soil microorganisms (Weidenhamer et al.,

2014), it has yet to be more broadly applied to the detection of bacterial specialized metabolites in marine ecosystems. Methodological advances in mass spectrometry coupled with pipelines for the visualization of complex datasets (Dunn et al., 2013; Petras et al., 2016; Wang et al.,

2016) provide new opportunities to assess environmental metabolomics and the ecological functions of specialized metabolites.

Bacteria cultured from ocean sediments represent a relatively new source of specialized metabolites (Blunt et al., 2015; Fenical and Jensen, 2006; Manivasagan et al., 2014; Mondol et al., 2013). While poorly studied relative to seawater bacteria, they occur in high abundance

(Schmidt et al., 1998) and form complex communities (Ghosh et al., 2010; Uthicke and McGuire,

2007). While there is evidence that natural products mediate interactions among marine bacteria (Angell et al., 2006; Dong et al., 2007; Hibbing et al., 2010; Kinkel et al., 2014; Simões et al., 2008; Wietz et al., 2013), their effects on the microbial community remain poorly studied.

Here we employed adsorbent resins and mass spectrometry to analyze the metabolomes associated with marine sediment communities across a range of coastal habitats off Belize. The results reveal complex chemical landscapes and the identification of numerous metabolites including those of microbial, invertebrate, and anthropogenic origin. The compounds identified included staurosporine, which occurred at concentrations that indicate its potential to function as an allelochemical. Coupling bacterial community analyses with environmental metabolomics provides new opportunities to correlate specialized metabolites with producers and a new window with which to address questions in marine chemical ecology.

2.3 Materials and Methods

2.3.1 Locations and samples

Five sites near the Smithsonian field station at Carrie Bow Cay, Belize were selected for metabolome and sediment sampling between 3-7 September 2015. The sites were characterized as: 1) shallow reef (depth: 8 m, 16° 47.901’ N, 88° 04.995’ W), 2) reef slope (depth:

20 m, 16° 48.191’ N 88° 04.691’ W), 3) mangrove channel (depth: 1 m, 16° 49.546’ N 88° 06.408’

W), 4) lagoon reef (depth: 6 m, 16° 47.726’ N 88° 07.349’ W), and 5) seagrass bed (depth: 1 m,

16° 48.174’ N 88° 04.925’ W). For sediment metabolome analysis, nine hand-made Miracloth

(EMD Millipore) bags each containing ca. 80 g of MeOH activated HP20 adsorbent resin (Diaion) were buried just below the surface of the sediment over an area ca. 1 m2 at each site. After 24 hours, the bags were retrieved and stored frozen (-20°C) prior to elution. One similarly prepared resin bag was used as a negative control. Five replicate sediment samples (ca. 80 g each) were also collected in sterile Whirl-pak® bags (Sigma-Aldrich) from each site for bacterial cultivation and additional metabolomic analyses. A subsample from each site was transferred into a 50 mL falcon tube with RNAlater (ThermoFisher) for molecular (DNA) analysis. All samples were stored frozen (-20°C) prior to processing.

2.3.2 Resin extraction and analysis

The HP20 resin samples from each site were removed from the Miracloth bags, combined, and suspended in 500 mL of 1:1 DCM:MeOH for 30 min in a 2.8 L Fernbach flask while shaking at 190 rpm on a G10 Gyrotary Shaker (New Brunswick Scientific). The solvent was collected by filtration through qualitative P8 coarse fluted filter paper (Fisher Scientific,

Hampton, NH), dried by rotary evaporation, and re-extracted with EtOAc followed by MeOH.

The combined extracts were filtered (0.2 μm) to remove particulates and sites 2-5 subjected to

C18 reversed phase flash chromatography using an acetonitrile/water gradient (100% H2O, 20%

ACN, 40% ACN, 60% ACN, 80% ACN, and 100% ACN). The fractions were dried by rotary evaporation, dissolved in MeOH at 1 mg/mL, and subjected to LC-MS/MS analysis by injecting 5

µL into an Agilent 1290 UHPLC system. The site 1 crude extract was similarly analyzed.

Chromatographic separation was achieved at 40°C using a flow rate of 0.5 mL/min and a 1.7 µm

C18 (50 x 2.1 mm) Kinetex UHPLC column (Phenomenex). A linear gradient of solvents A [water and 0.1% formic acid (v/v)] and B [acetonitrile and 0.1% formic acid (v/v)] was used: 0-0.5 min

(5% B), 0.5-8 min (5-95% B), 8-9 min (95% B). The column eluent was introduced directly into a

Bruker Daltonics microTOF focus II quadrupole-time-of-flight mass spectrometer (mtq) equipped with an Apollo II electrospray ionization source and controlled via otofControl v3.4

(build 16) and Hystar v3.2 software packages (Bruker Daltonics). The mtq instrument was first externally calibrated using ESI-L Low Concentration Tuning Mix (Agilent Technologies) prior to sample analysis and the internal calibrant (lock mass) hexakis (1H,1H,2H- difluoroethoxy)phosphazene (Synquest Laboratories), m/z 622.0295089613, was continuously introduced during the entirety of each LC/MS run. Data were collected in positive ion mode, scanning from 100-2000 m/z. Instrument source parameters were set as follows: nebulizer gas

(nitrogen) pressure, 3 Bar; capillary voltage, 4,500 V; ion source temperature, 200°C; dry gas flow, 9 L/min. MS1 spectral acquisition rate was set at 2 Hz and MS/MS acquisition rate was variable (5-10 Hz) depending on precursor intensity. Data-dependent MS/MS acquisition was programmed to the top ten most intense precursors per MS1 scan and precursors were actively excluded for 0.5 minute after being fragmented twice. Each MS/MS scan was the average of 4 collision energies, paired optimally with specific collision RF (or “ion cooler RF”) voltages and transfer times in order to maximize the qualitative structural information from each precursor.

The MS/MS data were converted to mzXML and uploaded to the MASSive server

(https://massive.ucsd.edu/) via the FTP client Filezilla (https://filezilla-project.org). Data are publicly available at http://gnps.ucsd.edu under accession number

MSV000080241 and MSV000079620.

MS data were analyzed using the Global Natural Products Social Molecular Networking

(GNPS) pipeline (Wang et al., 2016) using a modification of previously validated parameters

(Duncan et al., 2015). Only parent ions that fragmented at least twice were included. Networks were visualized using Cytoscape (http://www.cytoscape.org/cy3.html) with the built in solid layout (Smoot et al., 2011). Nodes with a cosine score >0.95 were combined into consensus spectra. The algorithm parameters included mass tolerance for fragment peaks (0.3 Da), parent mass tolerance (1.0 Da), a minimum number of matched peaks per spectral alignment (4), a maximum component size of 1, and a minimum cosine score of 0.6. Nodes in the top ten cosine scores (K parameter) in both directions were connected by an edge. MS/MS spectra from “no injection blanks” and negative controls were subtracted from the network to remove both instrument noise and parent ions associated with the resin extraction protocol.

Compounds were identified using the GNPS reference library, which contains over

220,000 curated spectra (Wang et al., 2016). To minimize false positives, a false discovery rate

(FDR) analysis was conducted using the Passatutto test on GNPS. This test generates a decoy library to assess the rate at which a user’s submitted spectra match spectra in the decoy library for a range of minimum matched peaks and cosine (similarity) scores (Scheubert et al., 2017).

Parameters were set to maintain a false discovery rate below 1% (Fig. 2.1). Library matches were assigned if they shared at least four MS/MS peaks, were within one Dalton, and had a similarity score above 0.65. Mirror spectra plots generated in GNPS were used to manually investigate all spectral matches. Positive matches to short amino acid sequences or primary metabolites were

excluded.

Four of the nine compounds identified were compared to authentic standards

[staurosporine (NETA Scientific), erucamide (VWR), cocamidopropyl betaine (Spectrum), and monactin (Santa Cruz Biotechnology)] using the Chemical Analysis Working Group’s established criteria (Sumner et al., 2007). These included retention time, fragmentation spectrum, and mass matching using a Dionex UltiMate 3000 HPLC (ThermoFisher Scientific, Waltham, MA) coupled to a Bruker impact HD qTOF mass spectrometer. Chromatographic separation was achieved at

40°C using a flow rate of 0.5 mL/min and a Kinetex C18 (1.7 μm, 100Å) column (50 mm x 2.1 mm) (Phenomenex). A linear gradient of solvents A [water and 0.1% formic acid (v/v)] and B

[acetonitrile and 0.1% formic acid (v/v)] was used: 0-1 min (5% B), 1-2 min (5-40% B), 2-8 min

(40-100% B), 8-9 min (100% B).

For the mass spec analysis, data were collected in positive ion mode, scanning from 800-

2000 m/z. Instrument source parameters were set as follows: nebulizer gas (nitrogen) pressure,

3 Bar; Capillary voltage, 4,500 V; ion source temperature, 200°C; dry gas flow, 9 L/min. MS1 spectral acquisition rate was set at 3Hz and MS/MS acquisition rate was variable depending on precursor intensity. Data-dependent MS/MS acquisition was programmed to the top ten most intense precursors per MS1 scan and precursors were actively excluded for 0.5 minute after being fragmented twice. Each MS/MS scan was subjected to five collision energies, 3eV and stepped 50%, 75%, 150%, and 200% in order to maximize the qualitative structural information from each precursor. HP-921 lock mass was introduced during the entirety of the mass spec run.

Figure 2.1 Estimated false discovery rates based on the minimum number of product ions used to match sediment spectra to library spectra for different cosine (similarity) scores. Lines are drawn at 1% and 5% FDR rates.

2.3.3 Staurosporine quantification

Three replicate sediment samples ranging from 6-11 mL from each site were extracted with 50 mL of 1:1 DCM:MeOH. The extracts were filtered, the solvent removed under a stream of nitrogen, and the remaining water dried via lyophilization. The dried extracts were solubilized in MeOH and adsorbed onto 100 mg C18, loaded onto 2 g Isolute SPE C18 columns (Biotage), and subjected to flash chromatography using a 20% acetonitrile:water step gradient. The fractions were dried by rotary evaporation, dissolved in 2 mL MeOH, and 30 µL injected into an

Agilent 1100 series HPLC with a Phenomenex Kinetex C18 reversed-phase column (2.6 mm, 100 x 4.6 mm) using a flow rate of 0.7 mL/min. A linear gradient of solvents A [water and 0.1% formic acid (v/v)] and B [acetonitrile and 0.1% formic acid (v/v)] was used with 0-2 min (5% B), 2–14 min (5–100% B), 14-15 min (100% B). The divert valve was set to waste for the first 2 min.

Staurosporine was identified based on retention time and UV-spectral matching in comparison with an authentic standard (NETA Scientific). The average area under the curve for the staurosporine peak (350 nm) was calculated for triplicate extracts from each location and converted to concentration using the best fit line equation generated in R (R Development Core

Team, 2011) for a standard curve generated from an authentic standard.

Biologically relevant sediment volumes were calculated by subtracting the inorganic volume from the total sediment volume of each sample. Inorganic volumes were determined at each site using standard protocols (Parker, 1983) by first oven drying (60°C for 48 h) triplicate 5 mL volumes of wet sediment. The dried sediments were then treated with excess 20% H2O2, heated (60°C) for 30 minutes to oxidize organics, re-dried, and the volume of remaining inorganic material calculated via volumetric displacement and subtracted from the initial wet sediment volume. Molar staurosporine concentrations were calculated by converting mg/mL staurosporine concentration calculated from the HPLC-UV data to μM using the biologically relevant sediment volumes.

2.3.4 C. elegans feeding deterrence

Staurosporine was dissolved in 50 μL DMSO to generate concentrations of 210 μM, 160

μM, and 110 μM (corresponding to 0.1 mg/mL, 0.075 mg/mL, and 0.05 mg/mL respectively) in sterile 10 mL glass culture tubes. These concentrations were picked because they bracketed the

MIC at which staurosporine deterred C. elegans feeding. E. coli strain OP-50 (Troemel Lab, UCSD) was grown in 50 mL of LB broth Miller (Fisher Scientific) for three days, autoclaved, and 950 µL added to the staurosporine dilutions and DMSO controls. 25 μL of the treated and control E. coli were then placed 1 cm apart in a 60 cm diameter petri dish containing 4 mL of nematode growth medium (Research Products International Corp., Mount Prospect, IL). Sterile C. elegans embryos

were generated as previously described (Girard et al., 2007) and ca. 75 placed equidistant from treatment and control foods, with an average of 6 plates per trial. The number of living C. elegans per food source was counted under a dissecting scope 24 h after embryo addition, converted to percentage C. elegans per food source, and differences between treatments and controls tested using a non-parametric t-test (Wilcoxon rank-sum test) in R version 3.4.1

(Wilcoxon, 1945).

2.3.5 Brine shrimp assay

The brine shrimp assay was performed as previously described (Solis et al., 1993).

Briefly, Artemia eggs (Fisher Scientific) were hatched in brine (1 L DI water, 40 g instant ocean) for 48 hours at room temperature with excess air flow. Approximately 25 Artemia were then transferred into each well of a 24 well plate containing 1.95 mL of brine, after which serially diluted staurosporine dissolved in 50 µL DMSO was added to the wells (0.16 µM-16 µM final concentrations) in triplicate along with DMSO controls. After 24 hours, the number of dead and living brine shrimp were counted in each well under a dissecting microscope.

2.3.6 Bacterial community analysis

Environmental DNA (eDNA) was extracted from 24 sediment samples (5 replicates per site for sites 1-4 and 4 replicates for site 5). Approximately 1 g of sediment per sample was added into FastPrep tubes and subjected to a physical (bead beating) and chemical (phenol- chloroform) DNA extraction protocol (Patin et al., 2013). All extractions were done in duplicate and combined during the purification process. The v4 region of the 16S rRNA gene was PCR amplified in triplicate for each sample from 1 μL of 5 ng/μL DNA template using the Phusion Hot

Start Flex 2x Master Mix, an annealing temperature of 60°C, and the primers 515F

(TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGYCAGCMGCCGCGGTAA) and 806Rb

(GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGGACTACNVGGGTWTCTAAT) (Caporaso et al.,

2012). Replicate reactions were pooled and the products cleaned using ExoSap-IT. Illumina barcodes were added using the following program: 98°C for 1 min followed by five cycles of 98°C for 10 sec, 65°C for 20 sec, and 72°C for 30 sec with a final extension at 72°C for 2 min. Libraries were constructed following Illumina protocols and PCR products cleaned with ExoSap-IT and pooled in equimolar concentrations, cleaned using ampure beads, and sequenced on a Miseq v2 PE250 instrument at the Institute for Genomic Medicine (IGM), University of California, San

Diego.

Sequences were trimmed using the paired-end command in Trimmomatic (Bolger et al.,

2014) with a quality control sliding window of 4 bases, a minimum quality of 25, and a minimum length of 50 bases. All samples were standardized to a rarefication depth of 95k reads, resulting in three replicates (one from site 3, one from site 4 and one from site 5) being removed due to inadequate sequencing depth. Remaining reads were then clustered into operational taxonomic units (OTUs) based on 99% similarity and taxonomy assigned using the Silva rRNA Database

(Quast et al., 2013) in Qiime version 1.9.1 (Caporaso et al., 2010). To assess Salinispora relative abundance, replicates were analyzed with chloroplast sequences excluded (Hanshew et al.,

2013).

2.3.7 Bacterial cultivation, identification, and metabolomics

Actinomycete cultivation was performed following standard protocols (Mincer et al.,

2002). Briefly, 1 g frozen sediment from each of the five sites was air dried for 48 h and blotted onto petri dishes containing medium A1 (10 g starch, 4 g yeast extract, 2 g peptone, 16 g agar,

750 mL 0.2 μm filtered seawater, 250 mL deionized water). The petri dishes were incubated at room temperature for two weeks after which Salinispora-like colonies were purified by dilution

streaking. Colony PCR was performed by suspending a single colony in 10 μL DMSO and using 1

μL as the PCR template. Each PCR consisted of 10x PCR Buffer (Applied Biosciences, Foster City,

CA, USA), 2.5 mM MgCl2 (Applied Biosciences), 0.7% DMSO dimethyl, 10 mM dNTPs, 1.5 U

AmpliTaq Gold DNA Polymerase (Applied Biosciences) and 10 μmol of each 16S rRNA primer

(FC27 (5’-AGAGTTTGATCCTGGCTCAG-3’) forward and RC1492 (5’TACGGCTACCTTGTTACGACTT-

3’) reverse). PCR conditions were as follows: 5 min of initial denaturation at 95°C followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min. The resulting amplicons were sequenced in the forward and reverse directions (Eton

Biosciences, San Diego), assembled, and trimmed in Geneious (Geneious v10.2.3) (Kearse et al.,

2012) yielding ca. 1300 bp products. Strains were identified based on NCBI BLAST analysis

(Johnson et al., 2008).

Metabolite analyses were performed on single actinomycete colonies removed from A1 agar plates and extracted with 5 mL of MeOH in 15 mL falcon tubes. Extracts were dried by rotary evaporation, weighed, re-dissolved at a concentration of 5 mg/mL in MeOH, and subjected to HPLC analysis on the Agilent 1100 series instrument using previously described parameters.

2.3.8 Metagenomics

Mechanically-sheared metagenomic DNA from site 2 (average size ca. 500 bps) was used to prepare an Illumina sequencing library using the Apollo 324TM NGS Library Prep System and the PrepX DNA library kit (Wafergen, CA). This library was then sequenced on an Illumina HiSeq

2500 Rapid Flow cell to yield 36,894,038 paired-end reads (2X175 bps). PRINSEQ was used to filter and trim the resulting raw Illumina reads, discarding reads that have an average quality score less than 30 or a length less than 87 (Schmieder and Edwards, 2011). SPAdes (default

parameters) was used to assemble reads that passed quality filtering (Bankevich et al., 2012).

Metagenomic BGCs were detected in the resulting scaffolds (>5000 bps) using antiSMASH and further analyzed manually (Weber et al., 2015).

2.3.9 Phylogenetic analyses

Eight putative housekeeping genes found to be flanking BGCs of interest were blasted against the NCBI database. FASTA sequences for the top 100 hits were downloaded and each set aligned with the corresponding query sequence using the MUSCLE 3.8.31 (Edgar et al., 2004) alignment option in Mesquite 3.04 (Maddison and Maddison, 2008). RaxmlGUI v1.5 was used with default parameters and no set out-group to construct phylogenetic trees with 50 bootstrap replicates (Silvestro and Michalak, 2012)

2.4 Results:

2.4.1 Sediment metabolomics

The adsorbent resin HP20 was used to extract organic compounds from marine sediments at five sites around the Smithsonian field station at Carrie Bow Cay, Belize. For sites

2-5 the extracts were fractionated then subjected to mass spectrometry and analyzed using the

GNPS pipeline (Wang et al., 2016). Site 1 crude was not fractionated however it was analyzed similarly. In total, 5,803 MS/MS spectra were collected of which 5,242 passed the filtering requirements. Identical fragmentation spectra were clustered resulting in the generation of 512 unique parent ions (nodes) after the removal of resin controls and solvent only blanks. These nodes represent parent ions with masses (m/z values) that ranged from 227-1424 Daltons, with most falling between 350–900 Daltons. The number of parent ions detected at each site ranged from 106 (site 5: seagrass bed) to 335 (site 3: mangrove channel) with an average of 204 per site. The data were visualized as a euler diagram, which revealed extraordinary levels of small

molecule diversity, with only nine parent ions shared among all sites and nearly half (47%) observed at only one site (Fig. 2.2). The decision to combine the replicates prevented a more thorough analysis of inter-site variation, however the considerable number of parent ions detected indicates that the resin technique developed for this study was effective for the extraction of a diverse set of metabolites from marine sediments.

Figure 2.2 Euler diagram comparing sediment metabolome composition at five sites. Total number of nodes (parent ions) per site is listed in parentheses after the site ID.

2.4.2 Compound identification

A molecular network generated from the MS/MS data revealed numerous molecular families with no matches to the >220,000 metabolites present in the GNPS database (Fig. 2.3).

Nine parent ions showed confident database matches including some that were detected at multiple sites (Table 2.1). A subsequent analysis performed using the GNPS standalone dereplication pipeline helped confirm these matches and led to the additional identification of cocamidopropyl betaine. The greater sensitivity of the stand-alone tool is likely due to the

clustering algorithm used to generate the molecular network, which can reduce match scores.

The nine compounds identified represent an array of diverse structures (Fig. 2.4) and origins including compounds previously reported from sponges, algae, and bacteria as well as of synthetic origin (Table 2.1). Strict criteria established by the Chemical Analysis Working Group

(CAWG) (Sumner et al., 2007) were applied to increase confidence in the matches. Standards were used to assign a “level 1” classification to four compounds based on retention time, mass, and fragmentation spectra according to CAWG recommendations (Fig. 2.5). The remaining five compounds were “putatively annotated” (level 2) according to CAWG guidelines based on spectral matching to library standards (Fig. 2.6). These results demonstrate that specialized metabolites can be identified directly in marine sediments thus providing evidence for in situ production and new opportunities to explore their ecological functions.

Figure 2.3 Sediment metabolomes visualized as a molecular network. Parent ions from five sites are colored based on distributions: green nodes only found at one site (site number indicated), red nodes found at multiple sites (total number indicated). Parent ions matching known compounds are indicated and the m/z values provided for each node.

Figure 2.4 Compounds identified in the sediment metabolomes.

Table 2.1 Compound Identification. M/z error describes how different the parent ion and library ion difference are in ppm (which contains both low resolution and high-resolution data). The similarity score is a metric from 0-1 generated by GNPS with >0.65 considered a positive match. The number of shared peaks indicates the number of MS/MS fragments that matched between the library and the parent ion fragmentation spectra. Known producers of the compound are listed with associated references.

The detection of the sponge metabolite smenospongidine was unexpected since the sediments did not contain any obvious sponge biomass. This compound was originally reported from Smenospongia sp. (Kondracki and Guyot, 1989) and subsequently from Hippospongia (Oda et al., 2007) and Dysidea spp. (Iguchi et al., 1990). While these latter genera occur in the

Caribbean (Djura et al., 1980; Pawlik et al., 1995; Wulff, 2012), there is no record of Caribbean sponges producing smenospongidine. Sesquiterpene quinones such as smenospongidine are also a well known class of bacterial metabolites (Gallagher et al., 2010), suggesting that the compound detected in marine sediments may be of microbial origin. Alternatively, it may be

excreted from sponges in proximity to the resin. Smenospongidine represents one of the few compounds detected at all five sites (Fig. 2.3) and belongs to a class of redox-active compounds whose members have been proposed to function as extracellular electron shuttles during periodic hypoxia events (Gallagher et al., 2017). Interestingly, the network analysis places smenospongidine in a small molecular family suggesting that structurally related compounds await identification (Fig. 2.3). Smenospongidine exhibits weak antimicrobial and cytotoxic activity (Rodríguez et al., 1992). Its ubiquity in marine sediments suggests it plays important, albeit yet to be defined roles in sediment microbial ecology.

Bacterial metabolites belonging to the nactin family of macrotetrolide antibiotics were also identified (monactin, dinactin, trinactin, tetranactin). These closely related macrocyclic polyketides are produced by Streptomyces species and exhibit antimicrobial, miticidal (Phillies et al., 1975), and cytotoxic activities (Graven et al., 1966). Similarities in their fragmentation spectra are visualized as a large molecular family in a network analysis (Fig. 2.1). Comparison with an authentic standard provides strong support for the identification of these compounds

(Fig. 2.5). Staurosporine was also identified using the GNPS dereplication protocols and confirmed by comparison with an authentic standard (Fig. S4). It was detected at three sites and appeared in the network as a single node (Fig. 1). Staurosporine is a protein kinase inhibitor originally isolated from Streptomyces staurosporeus (Omura et al., 1977) with activity against fungi, bacteria, and mammalian cell lines (Table 2.2). It has subsequently been reported from marine ascidians and mollusks (Cantrell et al., 1999; Schupp et al., 1999) as well as other actinomycetes (Williams et al., 1999) including the marine actinomycete Salinispora arenicola

(Jensen et al., 2007). The staurosporine isolated from marine invertebrates is proposed to be of microbial origin (Steinert et al., 2015), however this has yet to be experimentally verified. The final natural product that could be confidently identified was pheophytin, which was detected

at site 2. This flavonoid glycoside acts as an electron carrier in photosystem II (Klimov, 2003) and is also a breakdown product of chlorophyll (Lorenzen, 1967). It is found in cyanobacteria and eukaryotic phototrophs (Shoubaky et al., 2016) and may be indicative of decaying seaweeds or microalgae.

Table 2.2 Staurosporine biological activity. A list of the IC50 values previously determined for staurosporine. All values come from previously published studies except the C. elegans predation deterrence assay and the brine shrimp lethality assay.

In addition to the detection of natural products, two of the parent ions matched synthetic compounds and thus are likely of anthropogenic origin. Cocamidopropyl betaine is an additive to skin care products including sun protectants (Jacob and Sadegh, 2008) and has been found in numerous habitats associated with human activity (Petras et al., 2016). It was identified using the stand-alone GNPS dereplication tool and confirmed by comparison with an authentic standard (Fig. 2.5). The detection of this compound at three sites is alarming given that low levels of sun screen promote coral bleaching (Danovaro et al., 2008). Another potential contaminant is erucamide, a fatty acid derivative used as a slip agent in the manufacturing of plastics (Bhunia et al., 2013), waxes (Molnar, 1974), and other consumer goods. Erucamide

appeared in a large molecular family (Fig. 2.3) and was confirmed by comparison with an authentic standard (Fig. 2.5). It was detected at all five sites but not observed in methodological controls, suggesting it was present in the environment. Since erucamide is a derivative of the natural product erucic acid (Molnar, 1974), it cannot be ruled out whether or not the compound detected is of natural origin.

Figure 2.5 Retention time, mass, and mirror plot fragmentation spectral matching for the standards of monactin, staurosporine, cocamidopropyl betaine, and erucamide (green) and the parent ion detected (black).

Figure 2.6 Mirror plot showing the fragmentation spectrum for dynactin, trinactin, tetranactin, pheophytin, and smenospongidine from the GNPS library (green) against the matching parent ion from a sediment sample (black).

2.4.3 Linking compounds to community composition

In an attempt to link compounds identified in the sediment metabolomes to candidate producers, bacterial diversity was assessed by analyzing 16S rRNA gene amplicon libraries at a rarefied depth of 95,000 reads for three to five replicates from each of the five sites. The number of observed OTUs (99% sequence identity) among sites was high, ranging from 16,063 ± 344

(site 1) to 17,917 ± 222 (site 4). Given that Streptomyces spp. are well known for the production of specialized metabolites including staurosporine and nactins, we searched the sediment libraries for evidence of this genus. While no streptomycete OTUs were detected, the marine actinomycete genus Salinispora, which includes the known staurosporine producer S. arenicola

(Jensen et al., 2007), was detected at all five sites (Fig. 2.7A). The relative abundance of this genus was low, ranging from 0.001 ± 0.001% (site 5) to 0.010 ± 0.004% (site 2) of the community.

Assuming bacterial concentrations in marine sediments of 108-109/mL (Schmidt et al., 1998), the culture independent abundances detected here correspond with previous reports of Salinispora colony forming units occurring at concentrations of 103/mL sediment (Mincer et al., 2005, 2002).

Although the among site differences in Salinispora abundance were not significant (ANOVA,

F4,16=1.285; p = 0.317), sites 2 and 4 had the highest average Salinispora relative abundance.

Due to the region of the 16S rRNA gene sequenced, it was not possible to distinguish among the three named Salinispora species. We therefore used culture-dependent methods to test for the presence of S. arenicola and successfully isolated strains belonging to this species from sites 2 and 4 (Table 2.3). Metabolomic analyses revealed that these strains produced staurosporine

(Fig. 2.8), thus confirming the presence of bacteria capable of producing a compound detected in the sediment metabolome.

Figure 2.7 Correlation between Salinispora and staurosporine abundance in marine sediments. A) Salinispora relative abundance (%) and staurosporine abundance (µM) at five sites. B) Linear regression correlating Salinispora abundance with staurosporine concentration. Error bars are standard deviation

Figure 2.8 Staurosporine production by S. arenicola strains. LC traces (360 nm) of culture extracts show peaks at 5 min with UV absorbance spectra that closely match that of a staurosporine standard. 47

Table 2.3. BLAST analysis of actinomycete strains. Both isolates had a 100% match for S. arenicola strain CNH643.

2.4.4 Staurosporine quantification

Given that the resin method developed for the metabolomic analyses is not quantitative, the relationships between Salinispora relative abundance and staurosporine concentrations was determined using volumetric extraction protocols. Surprisingly, staurosporine was identified by HPLC-UV in extracts generated from all five sites (Fig. 2.9), thus allowing concentrations to be calculated based on comparisons to a standard curve (Fig. A2.1).

Given that the inorganic component of the sediment is unlikely to contain this compound, the volumetric sediment concentrations were divided by the percent organic material in the sample

(32-41%) to establish biologically relevant staurosporine concentrations, which ranged from 5-

17 µM and established the highest concentrations at sites 2 and 4 (Table 2.4). Salinispora relative abundance correlated with staurosporine concentration at the five sites (R2 = 0.56) (Fig. 2.7B), supporting the hypothesis that S. arenicola contributes to the staurosporine pool detected in marine sediments.

Figure 2.9 Staurosporine detection A) UV trace (360 nm) for sediment extracts from all five sites. Inset: absorbance spectrum corresponding to the major peak is shown with a high match score to a staurosporine library standard. B) Extracted ion chromatogram from the site 4 sediment extract for mass range 466.80-467.80 with the major peak matching staurosporine (exact mass 466.531).

Table 2.4 Average staurosporine concentration per site, the percent of the sediment that was inorganic material, and the biologically relevant concentration of staurosporine, i.e. the concentration of staurosporine if the inorganic fraction of the sediment is removed from the staurosporine quantification.

2.4.5 Metagenomic analyses

Metagenomics represents an increasingly feasible approach to link natural products to their producers (Crits-Christoph et al., 2018). This is facilitated by the growing number of experimentally characterized biosynthetic gene clusters (BGCs) and databases such as the MiBIG repository where they are deposited (Medema et al., 2015). In an effort to link compounds detected in the metabolomics data to Salinispora spp., a 37M read metagenomic library generated from site 2 was mapped to the complete genome sequences of S. tropica CNB440 and S. aeronicola CNS205. At a BLAST e-value threshold of 1x10 e-50 and a DNA sequence identity of >96%, 14 reads mapped to S. aeronicola, 22 mapped to S. tropica, and four mapped to both.

Half of the S. aeronicola reads and third of the S. tropica reads are >99% identical to the genomes, providing strong support for the species linkage. These results suggest an overall

Salinispora abundance of 0.5-1 part per million of the DNA in the sediment microbiome, which is in general agreement with the culture independent results (0.01-0.001%). Given the low relative abundance of Salinispora sequences, it is not surprising that Salinispora-derived sta genes could not be detected, despite our ability to culture staurosporine-producing Salinispora strains from the sediment samples. Given these results, it can be estimated that ca. 1000X more reads would be needed to confidently obtain 1X coverage of a Salinispora genome, thus demonstrating the challenges associated with detecting biosynthetic gene clusters from rare members of the microbial community using metagenomics approaches.

We next queried the metagenome for other potential sources of staurosporine using the staD gene, which is integral to staurosporine production (Freel et al., 2011). A tBLASTn analysis against the 1,147,515 >300 bp contig metagenomic assembly identified 14 contig hits

(e-value cutoff of 1x10 e-10) that ranged in length from 277-7834 nucleotides and coverage from

0.9-6.3X. The closest NCBI matches originated in cyanobacteria, a phylum known to produce

closely related indolocarbazoles (Sánchez et al., 2006). Moreover, the longest contig contained a FAD-dependent amino acid oxidase (L-tryptophan oxidase) known to be involved in indolocarbazole biosynthesis, supporting our partial recovery of a canonical indolocarbazole

BGC from the metagenome. Cyanobacteria were in considerably greater relative abundance

(6%) than Salinispora at site 2, suggesting they could represent a major source of the staurosporines detected. These results provide evidence that cyanobacteria represent another potential source of indolocarbazoles including staurosporine in marine sediments, although the community analysis did not reveal any known indolocarbazole producers. Similarly, BGCs associated with macrotetrolide (nactin) biosynthesis were not identified in the metagenomic analysis despite the detection of these compounds, suggesting the producers were below the detection limit or similar compounds are produced by different biosynthetic routes.

In total, 83 different BGCs were assembled from the site 2 metagenome (Fig. A2.2).

These encompass a wide range of biosynthetic paradigms including 23 RiPPs (predicted to encode bacteriocins, lantipeptides, cyanobactins, and lassopeptides), 26 PKSs, 10 NRPSs, 1 PKS-

NRPS hybrid, 14 terpenes, and 9 classified a “other” based on antiSMASH analyses. While 75% of these appear to originate from cyanobacteria based on BLAST analyses, BGCs were also potentially linked to diverse taxonomic groups including Proteobacteria (15%), Bacteroidetes

(5%), Aquificae (1%), Firmicutes (1%), and Planctomycetes (2%). To further explore the identity of the host strains, we identified housekeeping genes flanking eight of the assembled BGCs. All eight housekeeping genes were most closely related to cyanobacteria, with percent identities from 67-85% (Table S3). Phylogenetic trees were then generated containing the 100 top related genes based on an NCBI BLAST analysis. Some sequences claded with the genus Moorea (Fig.

A2.3), which is a prolific source of natural products (Leao et al., 2017). Other sequences claded with less well studied genera such as Myosarcina and Pleurocapsa. Interestingly, five of the

housekeeping genes did not cluster with any known cyanobacteria, suggesting they may belong to previously undescribed cyanobacterial lineages.

2.4.6 The ecological roles of staurosporine

Staurosporine is a low nanomolar kinase inhibitor that is highly toxic to mammalian cell lines, bacteria, and fungi (Table 2.2). To assess the potential effects of staurosporine in marine sediments, we compared the biologically relevant sediment concentrations observed in this study (Table 2.4) to previously published toxicity values (Table 2.2). The staurosporine concentrations at all five sites were above the IC50 values for HeLa cell (4 nM), protein kinase (2-

20 nM), and mycobacterium protein kinase inhibition (600 nM). Sites 2 and 4 contained staurosporine levels above those shown to inhibit fungal growth (9.6 µM). Staurosporine concentrations also exceeded brine shrimp IC50 values (812 nM) suggesting it may deter eukaryotic predators (Fig 2.10A), although these concentrations did not deter feeding by C. elegans assay (Fig. 2.10B). These results suggest that staurosporine can occur in nature at inhibitory concentrations, thus providing evidence that microbial specialized metabolites can have allelopathic effects in nature.

Figure 2.10 C. elegans and brine shrimp assays. A) Effects of staurosporine concentration on brine shrimp survival. Error bars are +/- one standard deviation of the mean. The exponential

regression line and equation are shown as well as the calculated IC50 value. B) Effects of staurosporine concentration on C. elegans feeding preference. C. elegans were given a choice between control (solvent only) and treatment (solvent + staurosporine) food sources at three different concentrations. Results presented as % C. elegans feeding on treatment

(black) vs. control (grey). The asterisk indicates a significant difference (t-test, p-value <0.05).

2.5 Discussion

Thousands of specialized metabolites have been isolated from cultured bacteria (Bérdy,

2005), yet little is known about their production and ecological functions in nature. While many of these compounds possess antibiotic and other biomedically relevant activities, it has long been questioned if they reach toxic concentrations in nature, leading to the suggestion they are more likely to function as signaling molecules than allelochemicals (Davies, 2006). Although some compounds clearly function as signals (Fajardo and Martínez, 2008), there is compelling but limited evidence that antibiotics can function as such in nature, as evidenced by the suppression of plant pathogens in soils (Raaijmakers et al., 2002). Nonetheless, challenges

associated with the detection of antibiotics in nature have hindered the ecological interpretation of their function. These challenges can now be addressed with the aid of environmental metabolomics (Viant, 2008). When applied to the human microbiome, this approach has led to the recognition that human-associated microbes produce antibiotics and other biologically active compounds (Donia and Fischbach, 2015) that play important roles in human health (Mazmanian et al., 2008; Wang et al., 2012; Zschüttig et al., 2012). Verifying production of these compounds however has only recently been addressed (Bouslimani et al.,

2015). These approaches have yet to be applied to ocean sediments, where specialized metabolites could play a major role in structuring microbial communities.

Here we demonstrate that metabolomics can be used to detect a vast array of yet to be identified metabolites in marine sediments without the need for cultivation. The ability to detect compounds directly in nature draws parallels with the first culture-independent assessments of bacterial communities, which revealed extensive diversity that had yet to be cultured or described (Hugenholtz et al., 1998). This approach has enormous potential to reveal community level chemical interactions and establish correlations between compounds and community composition. While many of the unidentified compounds could simply be missing from the dereplication database, this limitation will improve as more standards are deposited into the

GNPS library. Notably, the bacterial metabolite staurosporine was confidently identified both in the sediments and from S. arenicola strains cultured from the same sites, providing a link between compound and source organism. The low relative abundance at which the genus

Salinispora was detected supports the concept that members of the “rare biosphere” (Sogin et al., 2006) play important roles in structuring microbial communities (Ramirez et al., 2018), in this case via the production of biologically active specialized metabolites. The surprisingly high concentrations of staurosporine detected compared to the low relative abundance of

Salinispora spp. could indicate that this compound accumulates over time. The detection of cytotoxic concentrations of staurosporine at all five sites provides compelling evidence that it can function as an allelochemical in nature. While other biogenic sources remain a strong a possibility, combining sediment metabolomics with community analyses provides new opportunities to link microbial diversity with poorly characterized functional traits.

The other compounds identified provide interesting avenues for future research. The detection of smenospongidine, a sesquiterpene quinone originally reported from a marine sponge (55), suggests that sediment metabolomics may help identify compounds of microbial origin that have previously been ascribed to marine invertebrates (Wilson et al., 2014). The nactin family of macrotetrolide antibiotics were a prominent feature of site 2, yet the associated biosynthetic gene cluster (Walczak et al., 2000) was not detected in the metagenome suggesting they may originate from a source other than Streptomyces species. The detection of two synthetic compounds was not anticipated and indicates that the sediment metabolomics technique described here can be used to assess anthropogenic influence even in relatively pristine sites.

Environmental metabolomics provides a unique window with which to assess specialized metabolite diversity in nature. When coupled with community analyses, links between compounds and their producers can be established and subsequently tested. The extensive chemical diversity detected in marine sediments suggests that this approach could be used to establish baseline metabolomic fingerprints, link them to microbiome composition, and monitor baseline shifts following perturbation. It may also be possible to scale up the resin extraction protocol for discovery purposes, providing a retro-biosynthetic approach in which new compounds are first discovered and the structural information used to subsequently identify the producing organism from metagenomic assemblies. The detection of physiologically

relevant concentrations of biologically active specialized metabolites in marine sediments suggests that these compounds can have a major impact on community structure even if they are produced by rare members of the community.

2.6 Acknowledgments

E. coli strain OP-50 was kindly provided by the Troemel laboratory, UCSD. The protocol for the brine shrimp assay was provided by Christopher Leber, UCSD. This research was supported by the National Science Foundation under Grant Number (OCE-1235142). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Chapter 2 is currently being formatted into a manuscript that was submitted in September of 2018. Robert Tuttle, Alyssa M. Demko, Nastassia Patin, Cliff Kapono, Mohamed S. Donia,

Pieter Dorrestein, Paul R. Jensen. The dissertation author was the primary investigator on these studies.

2.7 References

Angell S, Bench BJ, Williams H, Watanabe CMH. 2006. Pyocyanin Isolated from a Marine Microbial Population: Synergistic Production between Two Distinct Bacterial Species and Mode of Action. Chem Biol 13:1349–1359. doi:10.1016/j.chembiol.2006.10.012 Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, Lesin VM, Nikolenko SI, Pham S, Prjibelski AD, Pyshkin A V., Sirotkin A V., Vyahhi N, Tesler G, Alekseyev MA, Pevzner PA. 2012. SPAdes: A New Genome Assembly Algorithm and Its Applications to Single-Cell Sequencing. J Comput Biol 19:455–477. doi:10.1089/cmb.2012.0021 Bérdy J. 2005. Bioactive Microbial Metabolites. J Antibiot (Tokyo) 58:1–26. doi:10.1038/ja.2005.1 Bhunia K, Sablani SS, Tang J, Rasco B. 2013. Migration of chemical compounds from packaging polymers during microwave, conventional heat treatment, and storage. Compr Rev Food Sci Food Saf 12:523–545. doi:10.1111/1541-4337.12028 Blunt JW, Copp BR, Munro MHG, Northcote PT, Prinsep MR. 2015. Marine natural products. Nat Prod Rep 28:196–268. doi:10.1039/C005001F

Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics 30:2114–2120. doi:10.1093/bioinformatics/btu170 Bouslimani A, Porto C, Rath CM, Wang M, Guo Y, Gonzalez A, Berg-Lyon D, Ackermann G, Moeller Christensen GJ, Nakatsuji T, Zhang L, Borkowski AW, Meehan MJ, Dorrestein K, Gallo RL, Bandeira N, Knight R, Alexandrov T, Dorrestein PC. 2015. Molecular cartography of the human skin surface in 3D. Proc Natl Acad Sci 112:E2120–E2129. doi:10.1073/pnas.1424409112 Burgess JG, Jordan EM, Bregu M, Mearns-Spragg A, Boyd KG. 1999. Microbial antagonism: a neglected avenue of natural products research. Prog Ind Microbiol 35:27–32. doi:10.1016/S0079-6352(99)80094-0 Cantrell CL, Groweiss A, Gustafson KR, Boyd MR. 1999. A New Staurosporine Analog from the Prosobranch Mollusk Coriocella Nigra. Nat Prod Lett 14:39–46. doi:10.1080/10575639908045433 Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, Fierer N, Peña AG, Goodrich JK, Gordon JI, Huttley G a, Kelley ST, Knights D, Koenig JE, Ley RE, Lozupone C a, Mcdonald D, Muegge BD, Pirrung M, Reeder J, Sevinsky JR, Turnbaugh PJ, Walters W a, Widmann J, Yatsunenko T, Zaneveld J, Knight R. 2010. correspondence QIIME allows analysis of high- throughput community sequencing data Intensity normalization improves color calling in SOLiD sequencing. Nat Publ Gr 7:335–336. doi:10.1038/nmeth0510-335 Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer M, Gormley N, Gilbert JA, Smith G, Knight R. 2012. Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. ISME J 6:1621– 1624. doi:10.1038/ismej.2012.8 Crits-Christoph A, Diamond S, Butterfield CN, Thomas BC, Banfield JF. 2018. Novel soil bacteria possess diverse genes for secondary metabolite biosynthesis. Nature. doi:10.1038/s41586- 018-0207-y Danovaro R, Bongiorni L, Corinaldesi C, Giovannelli D, Damiani E, Astolfi P, Greci L, Pusceddu A. 2008. Sunscreens cause coral bleaching by promoting viral infections. Environ Health Perspect 116:441–447. doi:10.1289/ehp.10966 Davies J. 2006. Are antibiotics naturally antibiotics?Journal of Industrial Microbiology and Biotechnology. pp. 496–499. doi:10.1007/s10295-006-0112-5 Djura P, Stierle DB, Sullivan B, Faulkner DJ, Arnold E V., Clardy J. 1980. Some metabolites of the marine sponges Smenospongia aurea and Smenospongia (.ident.Polyfibrospongia) echina. J Org Chem 45:1435–1441. doi:10.1021/jo01296a019 Dong Y, Wang L, Zhang L-H. 2007. Quorum-quenching microbial infections: mechanisms and implications. Philos Trans R Soc Lond B Biol Sci 362:1201–11. doi:10.1098/rstb.2007.2045 Donia MS, Fischbach MA. 2015. Small molecules from the human microbiota. Science (80- ) 349:1254766–1254766. doi:10.1126/science.1254766 Duncan KR, Crüsemann M, Lechner A, Sarkar A, Li J, Ziemert N, Wang M, Bandeira N, Moore BS, Dorrestein PC, Jensen PR. 2015. Molecular Networking and Pattern-Based Genome Mining Improves Discovery of Biosynthetic Gene Clusters and their Products from Salinispora

Species. Chem Biol 22:460–471. doi:10.1016/j.chembiol.2015.03.010 Dunn WB, Erban A, Weber RJM, Creek DJ, Brown M, Breitling R, Hankemeier T, Goodacre R, Neumann S, Kopka J, Viant MR. 2013. Mass appeal: Metabolite identification in mass spectrometry-focused untargeted metabolomics. Metabolomics 9:44–66. doi:10.1007/s11306-012-0434-4 Edgar RC, Drive RM, Valley M. 2004. Muscle: Multiple Sequence Alignment With High Accuracy and High Throughpu. Nucleic Acids Res 32:1792–7. doi:10.1093/nar/gkh340 Fajardo A, Martínez JL. 2008. Antibiotics as signals that trigger specific bacterial responses. Curr Opin Microbiol. doi:10.1016/j.mib.2008.02.006 Fenical W, Jensen PR. 2006. Developing a new resource for drug discovery: Marine actinomycete bacteria. Nat Chem Biol. doi:10.1038/nchembio841 Freel KC, Nam SJ, Fenical W, Jensen PR. 2011. Evolution of secondary metabolite genes in three closely related marine actinomycete species. Appl Environ Microbiol 77:7261–7270. doi:10.1128/AEM.05943-11 Gallagher KA, Fenical W, Jensen PR. 2010. Hybrid isoprenoid secondary metabolite production in terrestrial and marine actinomycetes. Curr Opin Biotechnol. doi:10.1016/j.copbio.2010.09.010 Gallagher KA, Wanger G, Henderson J, Llorente M, Hughes CC, Jensen PR. 2017. Ecological implications of hypoxia-triggered shifts in secondary metabolism. Environ Microbiol 19:2182–2191. doi:10.1111/1462-2920.13700 Ghosh A, Dey N, Bera A, Tiwari A, Sathyaniranjan K, Chakrabarti K, Chattopadhyay D. 2010. Culture independent molecular analysis of bacterial communities in the mangrove sediment of Sundarban, India. Saline Systems 6:1. doi:10.1186/1746-1448-6-1 Girard LR, Fiedler TJ, Harris TW, Carvalho F, Antoshechkin I, Han M, Sternberg PW, Stein LD, Chalfie M. 2007. WormBook: the online review of Caenorhabditis elegans biology. Nucleic Acids Res 35:D472–D475. doi:10.1093/nar/gkl894 Graven SN, Lardy HA, Johnson D, Rutter A. 1966. Antibiotics As Tools for Metabolic Studies. V. Effect of Nonactin, Monactin, Dinactin, and Trinactin on Oxidative Phosphorylation and Adenosine Triphosphatase Induction. Biochemistry 5:1729–1735. doi:10.1021/bi00869a040 Hanshew AS, Mason CJ, Raffa KF, Currie CR. 2013. Minimization of chloroplast contamination in 16S rRNA gene pyrosequencing of insect herbivore bacterial communities. J Microbiol Methods 95:149–155. doi:10.1016/j.mimet.2013.08.007 Hibbing ME, Fuqua C, Parsek MR, Peterson SB. 2010. Bacterial competition: surviving and thriving in the microbial jungle. Nat Rev Microbiol 8:15–25. doi:10.1038/nrmicro2259 Hugenholtz P, Goebel BM, Pace NR. 1998. Erratum: Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity (Journal of Bacteriology (1998) 180:18 (4765-4774)). J Bacteriol. doi:0021-9193/98/$04.00+0 Iguchi K, Sahashi A, Yamada Y, Kohno J. 1990. New sesquiterpenoid hydroquinone and quinones

from the Okinawan marine sponge (Dysidea sp.). Chem Pharm Bull (Tokyo) 38:1121–1123. doi:10.1248/cpb.38.1121 Jacob SE, Sadegh A. 2008. Cocamidopropyl betaine. Dermatitis. doi:10.2310/6620.2008.06043 Jensen PR, Williams PG, Oh DC, Zeigler L, Fenical W. 2007. Species-specific secondary metabolite production in marine actinomycetes of the genus Salinispora. Appl Environ Microbiol 73:1146–1152. doi:10.1128/AEM.01891-06 Johnson M, Zaretskaya I, Raytselis Y, Merezhuk Y, McGinnis S, Madden TL. 2008. NCBI BLAST: a better web interface. Nucleic Acids Res 36. doi:10.1093/nar/gkn201 Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Meintjes P, Drummond A. 2012. Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28:1647–1649. doi:10.1093/bioinformatics/bts199 Kinkel LL, Schlatter DC, Xiao K, Baines AD. 2014. Sympatric inhibition and niche differentiation suggest alternative coevolutionary trajectories among Streptomycetes. ISME J 8:249–256. doi:10.1038/ismej.2013.175 Klimov V V. 2003. Discovery of pheophytin function in the photosynthetic energy conversion as the primary electron acceptor of Photosystem II. Photosynth Res 76:247–53. doi:10.1023/A:1024990408747 Kondracki M-L, Guyot M. 1989. Biologically active quinone and hydroquinone sesquiterpenoids from the sponge smenospongia sp. Tetrahedron 45:1995–2004. doi:10.1016/S0040- 4020(01)80062-4 König GM, Kehraus S, Seibert SF, Abdel-Lateff A, Müller D. 2006. Natural products from marine organisms and their associated microbes. ChemBioChem. doi:10.1002/cbic.200500087 Kudela RM. 2011. Characterization and deployment of Solid Phase Adsorption Toxin Tracking (SPATT) resin for monitoring of microcystins in fresh and saltwater. Harmful Algae 11:117– 125. doi:10.1016/j.hal.2011.08.006 Lane JQ, Roddam CM, Langlois GW, Kudela RM. 2010. Application of Solid Phase Adsorption Toxin Tracking (SPATT) for field detection of the hydrophilic phycotoxins domoic acid and saxitoxin in coastal California. Limnol Oceanogr Methods 8:645–660. doi:10.4319/lom.2010.8.0645 Leao T, Castelão G, Korobeynikov A, Monroe EA, Podell S, Glukhov E, Allen EE, Gerwick WH, Gerwick L. 2017. Comparative genomics uncovers the prolific and distinctive metabolic potential of the cyanobacterial genus Moorea. Proc Natl Acad Sci 114:3198–3203. doi:10.1073/pnas.1618556114 Letzel A-C, Li J, Amos GCA, Millán-Aguiñaga N, Ginigini J, Abdelmohsen UR, Gaudêncio SP, Ziemert N, Moore BS, Jensen PR. 2017. Genomic insights into specialized metabolism in the marine actinomycete Salinispora. Environ Microbiol 19:3660–3673. doi:10.1111/1462- 2920.13867 Lorenzen C. 1967. Determination of chlorophyll and pheopigments:spectrophotometric equations. Limnol Oceanogr 12:343–346. doi:10.4319/lo.1967.12.2.0343

Luti KJK, Mavituna F. 2011. Streptomyces coelicolor increases the production of undecylprodigiosin when interacted with Bacillus subtilis. Biotechnol Lett 33:113–118. doi:10.1007/s10529-010-0401-y Maddison WP, Maddison DR. 2008. Mesquite: A modular system for evolutionary analysis. Evolution (N Y) 62:1103–1118. doi:http://mesquiteproject.org Manivasagan P, Venkatesan J, Sivakumar K, Kim SK. 2014. Pharmaceutically active secondary metabolites of marine actinobacteria. Microbiol Res. doi:10.1016/j.micres.2013.07.014 Mazmanian SK, Round JL, Kasper DL. 2008. A microbial symbiosis factor prevents intestinal inflammatory disease. Nature 453:620–625. doi:10.1038/nature07008 Mazzola M. 2002. No Title. Antonie Van Leeuwenhoek 81:557–564. doi:10.1023/A:1020557523557 Medema MH, Kottmann R, Yilmaz P, Cummings M, Biggins JB, Blin K, De Bruijn I, Chooi YH, Claesen J, Coates RC, Cruz-Morales P, Duddela S, Düsterhus S, Edwards DJ, Fewer DP, Garg N, Geiger C, Gomez-Escribano JP, Greule A, Hadjithomas M, Haines AS, Helfrich EJN, Hillwig ML, Ishida K, Jones AC, Jones CS, Jungmann K, Kegler C, Kim HU, Kötter P, Krug D, Masschelein J, Melnik A V., Mantovani SM, Monroe EA, Moore M, Moss N, Nützmann HW, Pan G, Pati A, Petras D, Reen FJ, Rosconi F, Rui Z, Tian Z, Tobias NJ, Tsunematsu Y, Wiemann P, Wyckoff E, Yan X, Yim G, Yu F, Xie Y, Aigle B, Apel AK, Balibar CJ, Balskus EP, Barona- Gómez F, Bechthold A, Bode HB, Borriss R, Brady SF, Brakhage AA, Caffrey P, Cheng YQ, Clardy J, Cox RJ, De Mot R, Donadio S, Donia MS, Van Der Donk WA, Dorrestein PC, Doyle S, Driessen AJM, Ehling-Schulz M, Entian KD, Fischbach MA, Gerwick L, Gerwick WH, Gross H, Gust B, Hertweck C, Höfte M, Jensen SE, Ju J, Katz L, Kaysser L, Klassen JL, Keller NP, Kormanec J, Kuipers OP, Kuzuyama T, Kyrpides NC, Kwon HJ, Lautru S, Lavigne R, Lee CY, Linquan B, Liu X, Liu W, Luzhetskyy A, Mahmud T, Mast Y, Méndez C, Metsä-Ketelä M, Micklefield J, Mitchell DA, Moore BS, Moreira LM, Müller R, Neilan BA, Nett M, Nielsen J, O’Gara F, Oikawa H, Osbourn A, Osburne MS, Ostash B, Payne SM, Pernodet JL, Petricek M, Piel J, Ploux O, Raaijmakers JM, Salas JA, Schmitt EK, Scott B, Seipke RF, Shen B, Sherman DH, Sivonen K, Smanski MJ, Sosio M, Stegmann E, Süssmuth RD, Tahlan K, Thomas CM, Tang Y, Truman AW, Viaud M, Walton JD, Walsh CT, Weber T, Van Wezel GP, Wilkinson B, Willey JM, Wohlleben W, Wright GD, Ziemert N, Zhang C, Zotchev SB, Breitling R, Takano E, Glöckner FO. 2015. Minimum Information about a Biosynthetic Gene cluster. Nat Chem Biol. doi:10.1038/nchembio.1890 Milshteyn A, Schneider JS, Brady SF. 2014. Mining the metabiome: Identifying novel natural products from microbial communities. Chem Biol. doi:10.1016/j.chembiol.2014.08.006 Mincer TJ, Fenical W, Jensen PR. 2005. Culture-dependent and culture-independent diversity within the obligate marine actinomycete genus Salinispora. Appl Environ Microbiol 71:7019–7028. doi:10.1128/AEM.71.11.7019-7028.2005 Mincer TJ, Jensen PR, Kauffman C a, Fenical W. 2002. Widespread and persistent populations of a major new marine actinomycete taxon in ocean sediments. Appl Environ Microbiol 68:5005–11. doi:10.1128/AEM.68.10.5005 Molnar N. 1974. Erucamide. J Am Oil Chem Soc 51:84–87. doi:10.1007/BF00000019 Mondol MAM, Shin HJ, Islam MT. 2013. Diversity of Secondary Metabolites from Marine Bacillus

Species: Chemistry and Biological Activity. Mar Drugs 11:2846–2872. doi:10.3390/md11082846 Oda T, Wang W, Ukai K, Nakazawa T, Mochizuki M. 2007. A sesquiterpene quinone, 5-Epi- smenospongine, promotes TNF-α production in LPS-stimulated RAW 264.7 Cells. Mar Drugs 5:151–6. doi:10.3390/md20070011 Omura S, Iwai Y, Hirano a, Nakagawa a, Awaya J, Tsuchya H, Takahashi Y, Masuma R. 1977. A new alkaloid AM-2282 OF Streptomyces origin. Taxonomy, fermentation, isolation and preliminary characterization. J Antibiot (Tokyo) 30:275–282. doi:10.7164/antibiotics.30.275 Parker JG. 1983. A Comparison of Methods used for the Measurement of Organic Matter in Marine Sediment. Chem Ecol 1:201–209. doi:10.1080/02757548308070802 Patin N V., Kunin V, Lidström U, Ashby MN. 2013. Effects of OTU Clustering and PCR Artifacts on Microbial Diversity Estimates. Microb Ecol 65:709–719. doi:10.1007/s00248-012-0145-4 Patin N V, Floros DJ, Hughes CC, Dorrestein PC, Jensen PR. 2018. The role of inter-species interactions in Salinispora specialized metabolism. Microbiology. Pawlik JR, Chanas B, Toonen RJ, Fenical W. 1995. Defenses of Caribbean sponges against predatory reef fish. I. Chemical deterrency. Mar Ecol Prog Ser 127:183–194. doi:10.3354/meps127183 Petras D, Nothias L, Quinn RA, Alexandrov T, Bandeira N, Bouslimani A, Castro-Falcón G, Chen L, Dang T, Floros DJ, Hook V, Garg N, Hoffner N, Jiang Y, Kapono CA, Koester I, Knight R, Leber CA, Ling T, Luzzatto-Knaan T, McCall L-I, McGrath AP, Meehan MJ, Merritt JK, Mills RH, Morton J, Podvin S, Protsyuk I, Purdy T, Satterfield K, Searles S, Shah S, Shires S, Steffen D, White M, Todoric J, Tuttle R, Wojnicz A, Sapp V, Vargas F, Yang J, Zhang C, Dorrestein PC. 2016. Mass Spectrometry-Based Visualization of Molecules Associated with Human Habitats. Anal Chem 88:10775–10784. doi:10.1021/acs.analchem.6b03456 Phillies GDJ, Asher IM, Eugene H. 1975. Nonactin, Monactin, Dinactin, Trinactin, and Tetranactin. 14:2311–2327. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, Peplies J, Glöckner FO. 2013. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools. Nucleic Acids Res 41. doi:10.1093/nar/gks1219 R Development Core Team R. 2011. R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing. doi:10.1007/978-3-540-74686-7 Raaijmakers JM, Bonsall RF, Weller DM. 1999. Effect of Population Density of Pseudomonas fluorescens on Production of 2,4-Diacetylphloroglucinol in the Rhizosphere of Wheat. Phytopathology 89:470–475. doi:10.1094/PHYTO.1999.89.6.470 Raaijmakers JM, Vlami M, de Souza JT. 2002. Antibiotic production by bacterial biocontrol agents. Antonie Van Leeuwenhoek 81:537–47. doi:10.1023/A:1020501420831 Ramirez KS, Knight CG, De Hollander M, Brearley FQ, Constantinides B, Cotton A, Creer S, Crowther TW, Davison J, Delgado-Baquerizo M, Dorrepaal E, Elliott DR, Fox G, Griffiths RI, Hale C, Hartman K, Houlden A, Jones DL, Krab EJ, Maestre FT, McGuire KL, Monteux S, Orr

CH, Van Der Putten WH, Roberts IS, Robinson DA, Rocca JD, Rowntree J, Schlaeppi K, Shepherd M, Singh BK, Straathof AL, Bhatnagar JM, Thion C, Van Der Heijden MGA, De Vries FT. 2018. Detecting macroecological patterns in bacterial communities across independent studies of global soils. Nat Microbiol 3:189–196. doi:10.1038/s41564-017- 0062-x Rodríguez J, Quiñoá E, Riguera R, Peters BM, Abrell LM, Crews P. 1992. The structures and stereochemistry of cytotoxic sesquiterpene quinones from dactylospongia elegans. Tetrahedron 48:6667–6680. doi:10.1016/S0040-4020(01)80012-0 Rutledge PJ, Challis GL. 2015. Discovery of microbial natural products by activation of silent biosynthetic gene clusters. Nat Rev Microbiol 13:509–523. doi:10.1038/nrmicro3496 Sánchez C, Méndez C, Salas JA. 2006. Indolocarbazole natural products: occurrence, biosynthesis, and biological activity. Nat Prod Rep 23:1007–1045. doi:10.1039/B601930G Scheubert K, Hufsky F, Petras D, Wang M, Nothias L-F, Duehrkop K, Bandeira N, Dorrestein P, Boecker S. 2017. Significance estimation for large scale untargeted metabolomics annotations. bioRxiv 109389. doi:10.1101/109389 Schmidt JL, Deming JW, Jumars PA, Keil RG. 1998. Constancy of bacterial abundance in surficial marine sediments. Limnol Oceanogr 43:976–982. doi:10.4319/lo.1998.43.5.0976 Schmieder R, Edwards R. 2011. Quality control and preprocessing of metagenomic datasets. Bioinformatics 27:863–864. doi:10.1093/bioinformatics/btr026 Schupp P, Eder C, Proksch P, Wray V, Schneider B, Herderich M, Paul V. 1999. Staurosporine Derivatives from the Ascidian Eudistoma t oealensis and Its Predatory Flatworm Pseudoceros sp. J Nat Prod 62:959–962. doi:10.1021/np980527d Shoubaky GA El, Abdel-daim MM, Mansour MH, Salem EA. 2016. Isolation and Identification of a Flavone Apigenin from Marine Red Alga Acanthophora spicifera with Antinociceptive and Anti-Inflammatory Activities 21–29. doi:10.4137/JEN.S25096.TYPE Silvestro D, Michalak I. 2012. RaxmlGUI: A graphical front-end for RAxML. Org Divers Evol 12:335–337. doi:10.1007/s13127-011-0056-0 Simões M, Simoes LC, Pereira MO, Vieira MJ. 2008. Antagonism between Bacillus cereus and Pseudomonas fluorescens in planktonic systems and in biofilms. Biofouling 24:339–349. doi:10.1080/08927010802239154 Slattery M, Rajbhandari I, Wesson K. 2001. Competition-Mediated Antibiotic Induction in the Marine Bacterium Streptomyces tenjimariensis. Microb Ecol 41:90–96. doi:10.1007/s002480000084 Smoot ME, Ono K, Ruscheinski J, Wang PL, Ideker T. 2011. Cytoscape 2.8: New features for data integration and network visualization. Bioinformatics 27:431–432. doi:10.1093/bioinformatics/btq675 Sogin ML, Morrison HG, Huber J a, Mark Welch D, Huse SM, Neal PR, Arrieta JM, Herndl GJ. 2006. Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl Acad Sci U S A 103:12115–20. doi:10.1073/pnas.0605127103

Solis PN, Wright CW, Anderson MM, Gupta MP, Phillipson JD. 1993. A microwell cytotoxicity assay using Artemia salina (brine shrimp). Planta Med. doi:10.1055/s-2006-959661 Steinert G, Taylor MW, Schupp PJ. 2015. Diversity of Actinobacteria Associated with the Marine Ascidian Eudistoma toealensis. Mar Biotechnol 17:377–385. doi:10.1007/s10126-015- 9622-3 Sumner LW, Amberg A, Barrett D, Beale MH, Beger R, Daykin CA, Fan TWM, Fiehn O, Goodacre R, Griffin JL, Hankemeier T, Hardy N, Harnly J, Higashi R, Kopka J, Lane AN, Lindon JC, Marriott P, Nicholls AW, Reily MD, Thaden JJ, Viant MR. 2007. Proposed minimum reporting standards for chemical analysis: Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics 3:211–221. doi:10.1007/s11306- 007-0082-2 Takahashi E, Yu Q, Eaglesham G, Connell DW, McBroom J, Costanzo S, Shaw GR. 2007. Occurrence and seasonal variations of algal toxins in water, phytoplankton and shellfish from North Stradbroke Island, Queensland, Australia. Mar Environ Res 64:429–442. doi:10.1016/j.marenvres.2007.03.005 Thrane, Harder Nielsen T, Neiendam Nielsen M, Sørensen, Olsson. 2000. Viscosinamide- producing Pseudomonas fluorescens DR54 exerts a biocontrol effect on Pythium ultimum in sugar beet rhizosphere. FEMS Microbiol Ecol 33:139–146. doi:10.1016/S0168- 6496(00)00054-4 Traxler MF, Watrous JD, Alexandrov T, Dorrestein PC, Kolter R. 2013. Interspecies Interactions Stimulate Diversification of the Streptomyces coelicolor Secreted Metabolome. MBio 4:e00459-13. doi:10.1128/mBio.00459-13 Trischman JA, Oeffner RE, De Luna MG, Kazaoka M. 2004. Competitive induction and enhancement of indole and a diketopiperazine in marine bacteria. Mar Biotechnol 6:215– 220. doi:10.1007/s10126-003-0010-z Uthicke S, McGuire K. 2007. Bacterial communities in Great Barrier Reef calcareous sediments: Contrasting 16S rDNA libraries from nearshore and outer shelf reefs. Estuar Coast Shelf Sci 72:188–200. doi:10.1016/j.ecss.2006.10.017 Viant MR. 2008. Recent developments in environmental metabolomics. Mol Biosyst 4:980. doi:10.1039/b805354e Walczak RJ, Woo AJ, Strohl WR, Priestley ND. 2000. Nonactin biosynthesis: The potential nonactin biosynthesis gene cluster contains type II polyketide synthase-like genes. FEMS Microbiol Lett 183:171–175. doi:10.1016/S0378-1097(99)00662-X Wang M, Carver JJ, Phelan V V, Sanchez LM, Garg N, Peng Y, Nguyen DD, Watrous J, Kapono CA, Luzzatto-Knaan T, Porto C, Bouslimani A, Melnik A V, Meehan MJ, Liu W-T, Crüsemann M, Boudreau PD, Esquenazi E, Sandoval-Calderón M, Kersten RD, Pace LA, Quinn RA, Duncan KR, Hsu C-C, Floros DJ, Gavilan RG, Kleigrewe K, Northen T, Dutton RJ, Parrot D, Carlson EE, Aigle B, Michelsen CF, Jelsbak L, Sohlenkamp C, Pevzner P, Edlund A, McLean J, Piel J, Murphy BT, Gerwick L, Liaw C-C, Yang Y-L, Humpf H-U, Maansson M, Keyzers RA, Sims AC, Johnson AR, Sidebottom AM, Sedio BE, Klitgaard A, Larson CB, Boya P CA, Torres-Mendoza D, Gonzalez DJ, Silva DB, Marques LM, Demarque DP, Pociute E, O’Neill EC, Briand E, Helfrich EJN, Granatosky EA, Glukhov E, Ryffel F, Houson H, Mohimani H, Kharbush JJ, Zeng

Y, Vorholt JA, Kurita KL, Charusanti P, McPhail KL, Nielsen KF, Vuong L, Elfeki M, Traxler MF, Engene N, Koyama N, Vining OB, Baric R, Silva RR, Mascuch SJ, Tomasi S, Jenkins S, Macherla V, Hoffman T, Agarwal V, Williams PG, Dai J, Neupane R, Gurr J, Rodríguez AMC, Lamsa A, Zhang C, Dorrestein K, Duggan BM, Almaliti J, Allard P-M, Phapale P, Nothias L-F, Alexandrov T, Litaudon M, Wolfender J-L, Kyle JE, Metz TO, Peryea T, Nguyen D-T, VanLeer D, Shinn P, Jadhav A, Müller R, Waters KM, Shi W, Liu X, Zhang L, Knight R, Jensen PR, Palsson BØ, Pogliano K, Linington RG, Gutiérrez M, Lopes NP, Gerwick WH, Moore BS, Dorrestein PC, Bandeira N. 2016. Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking. Nat Biotechnol 34:828– 837. doi:10.1038/nbt.3597 Wang X, Du L, You J, King JB, Cichewicz RH. 2012. Fungal biofilm inhibitors from a human oral microbiome-derived bacterium. Org Biomol Chem 10:2044. doi:10.1039/c2ob06856g Weber T, Blin K, Duddela S, Krug D, Kim HU, Bruccoleri R, Lee SY, Fischbach M a., Muller R, Wohlleben W, Breitling R, Takano E, Medema MH. 2015. antiSMASH 3.0--a comprehensive resource for the genome mining of biosynthetic gene clusters. Nucleic Acids Res 1–7. doi:10.1093/nar/gkv437 Weidenhamer JD, Mohney BK, Shihada N, Rupasinghe M. 2014. Spatial and Temporal Dynamics of Root Exudation: How Important is Heterogeneity in Allelopathic Interactions? J Chem Ecol 40:940–952. doi:10.1007/s10886-014-0483-4 Wietz M, Duncan K, Patin N V., Jensen PR. 2013. Antagonistic Interactions Mediated by Marine Bacteria: The Role of Small Molecules. J Chem Ecol 39:879–891. doi:10.1007/s10886-013- 0316-x Wilcoxon F. 1945. Individual Comparisons by Ranking Methods. Biometrics Bull 1:80. doi:10.2307/3001968 Williams DE, Bernan VS, Ritacco F V., Maiese WM, Greenstein M, Andersen RJ. 1999. Holyrines A and B, possible intermediates in staurosporine biosynthesis produced in culture by a marine actinomycete obtained from the North Atlantic Ocean. Tetrahedron Lett 40:7171– 7174. doi:10.1016/S0040-4039(99)01495-1 Wilson MC, Mori T, Rückert C, Uria AR, Helf MJ, Takada K, Gernert C, Steffens UAE, Heycke N, Schmitt S, Rinke C, Helfrich EJN, Brachmann AO, Gurgui C, Wakimoto T, Kracht M, Crüsemann M, Hentschel U, Abe I, Matsunaga S, Kalinowski J, Takeyama H, Piel J. 2014. An environmental bacterial taxon with a large and distinct metabolic repertoire. Nature 506:58–62. doi:10.1038/nature12959 Wulff J. 2012. Ecological Interactions and the Distribution, Abundance, and Diversity of SpongesAdvances in Marine Biology. pp. 273–344. doi:10.1016/B978-0-12-387787- 1.00003-9 Zschüttig A, Zimmermann K, Blom J, Goesmann A, Pöhlmann C, Gunzer F. 2012. Identification and characterization of microcin S, a new antibacterial peptide produced by probiotic Escherichia coli G3/10. PLoS One 7:e33351. doi:10.1371/journal.pone.0033351

2.8 Appendix

Figure A2.1 Staurosporine standard curve including linear regression equation and R2 value.

Table A2.1. BLAST analysis of house keeping genes flanking eight biosynthetic gene clusters (BGCs) assembled from the metagenome. Shown is the name of the house keeping gene, the four top related BLAST hits, and the percent identity.

(103–106)

Figure A2.2 Example biosynthetic gene clusters assembled from the site 2 metagenome. BGCs were identified from the SPAdes-assembled metagenome using antiSMASH (scaffolds > 5000 bps), and further annotated manually through BLASTp searches for each gene against NCBI. For BGC-03 and BGC-06,homologs identified in sequenced cyanobacterial genomes are shown.

Figure A2.3 Representative phylogenetic tree for cyanobacterial house keeping genes found flanking cyanobacterial BGCs. This phylogenetic tree is for the house keeping gene “UTP-- glucose-1-phosphate uridylyltransferase” containing the 100 closest BLAST matches. The sequence isolated from the metagenome is highlighted in blue

Chapter 3

Salinispora Secondary Metabolite Mediated Feeding Deterrence

3.1 Abstract

Numerous natural products have been isolated from marine bacteria; a large majority from the order Actinomycetales. However, very little effort has focused on determining the ecological roles these metabolites provide for their producers. Salinispora, an obligate marine actinomycete genus, has been studied extensively due to its production of specialized metabolites, making this a model genus for chemical ecology experiments. Interestingly, each of the three described Salinispora spp. produce unique cytotoxic compounds. The aims of this study were to determine if S. tropica and S. arenicola secondary metabolites have a deterrent effect on invertebrate predators as well as to determine if Salinispora species exhibit different anti-predatory strategies. Choice-based feeding experiments were designed using the model bacterivorous nematode Caenorhabditis elegans. Results of these choice-based feeding assays indicated that live Salinispora cultures were avoided by C. elegans. Salinispora chemical extracts were then tested in preference assays with C. elegans indicating that S. tropica extracts exhibit deterrent properties, whereas S. arenicola crude extracts have chemoattractant properties.

These results demonstrate that strategies of predator deterrence differ between Salinispora spp. Bioactivity guided isolation revealed that a suite of lomaiviticins, compounds produced by

S. tropica but not S. arenicola, were responsible for the observed deterrence. Further preference assays were performed using both a marine polychaete and marine nematodes. Results of these assays supported those observed using C. elegans, with S. tropica exhibiting chemodeterrence against both potential predators.

3.2 Introduction

In chapter 2, I detected cytotoxins in sediments off the Belize coast; specifically, the compound staurosporine. This protein kinase inhibitor was further quantified and found to be present at inhibitory concentrations. A tentative link was made between staurosporine and the marine actinomycete Salinispora arenicola, which is known to produce this compound.

Specialized metabolites from sediment bacteria have been shown to inhibit predators; however, it is not known why Salinispora spp. produce these compounds and what sort of ecological roles they may serve. The aim of this chapter was to determine if Salinispora spp. produce compounds that deter bacterivorous eukaryotes.

Marine microbes are fundamental to ecosystem processes as they play an integral part in myriad processes including nutrient cycling, primary production, and pollution remediation

(Arrigo, 2005; Azam et al., 1983; Cohen, 2002). One of the reasons these organisms can aid in such large scale processes is because of their relative abundance in the ocean environment, where microbial loads are estimated at 1011 cells per gram of sediment (Alongi, 1988). Numerous organisms have evolved to prey on these bacteria including viruses, protozoans, nematodes, and polychaetas (Pernthaler 2005; Moens et al. 1999; Tsuchiya & Kurihara 1979; Bouvier & del

Giorgio 2007). In top-down control, bacteria are under tremendous predation pressure and as such have evolved several strategies to avoid predation (Gasol et al., 2002; Jousset, 2012). These strategies can include morphological adaptations, such as increasing cell size to avoid ingestion

(Corno and Jürgens, 2006), or the formation of colonies and biofilms that are relatively impervious to protozoans and small eukaryotic predators (Queck et al., 2006; Weitere et al.,

2005). For instance, co-culturing a freshwater bacterial species of Flectobacillus with predatory protists resulted in the generation of a predation resistant bacterial morphotype (Corno and

Jürgens, 2006).

Another strategy that microbes have developed to avoid predation is chemically mediated defense (Wietz et al., 2013). There have been several studies linking compounds from terrestrial microbes to predator defense (Jousset et al., 2006; Klobutcher et al., 2006; Lainhart et al., 2009; Mazzola et al., 2009). However there have been fewer examples of predator chemodeterrence identified from marine microorganisms, and most have been from mat forming cyanobacteria (Cruz-Rivera and Paul, 2007; Nagle et al., 1996; Nagle and Paul, 1998;

Pennings et al., 1996), which form large colonies and are thus more analogous to macroalgae than bacteria. Ypaoamide for instance, produced by the cyanobacterium Lyngbya majuscule

(Moorea producens) is capable of deterring feeding from both parrotfish and sea urchins at naturally occurring concentrations (Nagle and Paul, 1998). The cyanobacterial secondary metabolites malingamide and majusculamide also exhibit sea urchin and pufferfish feeding deterrence (Pennings 1996). There are several examples of compounds from microbes in the water column that exhibit similar anti-feeding properties. Pseudomonas luteoviolacea produces violacein, a purple alkaloid that inhibits protozoan feeding at sub-millimolar concentrations.

Even ingestion of 1-2 pseudomonad cells has been shown to cause nanoflagellate death (Matz et al., 2008).

Overall, anti-predatory allelopathic interactions have not been very well characterized from sediment inhabiting marine bacteria, but are likely prevalent due to the wealth of cytotoxic compounds that have been isolated them (Blunt et al., 2015; Fenical and Jensen, 2006).

Numerous bacteria have been cultured from these environments, including several actinomycete species that produce many secondary metabolite compounds (Fiedler et al., 2005;

Jensen et al., 2015). However, to my knowledge, there has not been a single study that links any of these compounds to predation defense.

The marine obligate actinomycete genus Salinispora has been studied extensively due

to its prolific natural product production (Jensen et al., 2015). To date, three Salinispora species have been formally described, S. arenicola, S. tropica, and S. pacifica (Maldonado et al. 2005,

Ahmed et al. 2013). A recent analysis of 119 Salinispora genome sequences identified over 170 secondary metabolite biosynthetic gene clusters (BGCs) and led to estimates that the genus contains upwards of 250 BGCs (Letzel et al., 2017). Interestingly, several BGCs are unique to specific Salinispora species, indicating that Salinispora secondary metabolites may be linked to ecological differentiation (Jensen et al., 2007). For instance, the BGCs encoding the cytotoxic compounds lomaiviticin and salinisporamide A appear to be fixed in S. tropica whereas the BGC encoding the compound lymphostin is found ubiquitously in all three species. The cytokinase inhibitor staurosporine however, is found in all S. arenicola strains and is absent from S. tropica

(Ziemert et al., 2014). Recently it was shown that S. arenicola and S. tropica exhibit differential competitive strategies when dealing with microbial competitors in the benthic environment and this was linked to disparity in secreted metabolites between species (Patin et al., 2015). S. arenicola employed antimicrobial secondary metabolites to inhibit neighboring microbes whereas S. tropica used faster growth to outcompete potential competitors (Patin et al., 2015).

This provided evidence that these species are different ecotypes, allowing them to co-occur in similar habitats.

C. elegans is a bacterivorous nematode that has been widely adopted as a model organism for numerous fields of study including developmental, cell, and evolutionary biology

(Girard et al., 2007; Riddle et al., 1997). Since its isolation in 1963, C. elegans has been studied extensively leading to a wealth of information pertaining to its genetic makeup and neuronal network (Bargmann, 2006; Brenner, 1974). C. elegans has been shown to be highly chemotactic with 32 of 302 neurons dedicated to chemosensation (Bargmann, 2006). This system has been exploited in studies looking at C. elegans’ response to toxins and pathogens (Schulenburg and

Ewbank, 2007). C. elegans exhibits food preference, indicating that organisms will migrate between bacterial colonies to find an optimal food source (Shtonda and Avery, 2006). With their relatively short growth cycle, ease of culturing, and highly developed chemotactic ability

(Stiernagle, 2006), C. elegans is an ideal organism to use in feeding deterrence assays. Although

C. elegans does not occur in the marine environment, they can facilitate assay development because of how easy they are to culture, which makes them ideal for the use in bioassay-guided fractionation schemes to identify deterrent compounds.

Polychaetes represent another group of bacterivorous organisms common in marine sediments. Polychaeta is a paraphyletic class of annelid worms that include mostly marine species with varied lifestyles and morphologies (Struck et al., 2011). Numerous species occur in marine sediments, where it has been shown that bacteria are an essential part of their diets

(Andresen and Kristensen 2002). This makes polychaetes ideal predators to use in developing an assay addressing the chemical defense of Salinispora against predation. An unnamed species of Ophryotrocha (Dorvilleidae, Eunicida), which grazes on detritus and bacteria (Brown and Ellis

1971, Fauchald et al. 1979, Thornhill et al. 2009), was picked by Greg Rouse (Scripps institution of Oceanography) for use in developing chemical deterrence assays .

To assess the ecological roles of specialized metabolites, it is important to pick test organisms that are not only marine but also cooccur with the producing organisms. Tropical sediments are teaming with marine nematodes, many of which feed on bacteria, making them a perfect candidate for anti-predation based assays as well (Heip et al., 1985).

One of the main hypotheses addressed in this chapter was that Salinispora spp. produce cytotoxic specialized metabolites that deter predation by bacterivorous eukaryotes. Since there is evidence that S. tropica and S. arenicola exhibit different competitive strategies, I also

hypothesized that Salinispora spp. possess different predator deterrence strategies.

Interestingly, C. elegans showed a marked preference for S. arenicola over S. tropica, indicating that one spp. may be less palatable to bacterivores. Chemical extracts generated from S. tropica strains deterred C. elegans whereas S. arenicola extracts acted as chemoattractants indicating that Salinispora species do indeed possess different predator deterrence strategies. The primary deterrent compounds isolated from S. tropica strains were the known cytotoxic lomaiviticins.

Since C. elegans are not a marine species, and thus not ecologically relevant test organisms, I then looked at the deterrent effects of Salinispora secondary metabolites on marine polychaetes and nematodes. The results mirrored those observed in the C. elegans feeding assays when polychaetes were given a choice between habitats with and without Salinispora spp. Developing assays to test marine nematodes feeding or habitat preference proved difficult but introducing these predators to S. tropica extracts resulted in mortality over a relatively short time period. However, S. arenicola extracts were much less lethal, indicating that even with ecologically relevant bacterivores, S. tropica appears to be better chemically defended than S. arenicola.

3.3 Methods

3.3.1 Salinispora strain selection and culturing

A total of ten strains of S. arenicola and eight strains of S. tropica, all with sequenced and annotated genomes, were selected to test potential feeding deterrence against C. elegans

(Table 3.1). These strains were selected based on their shared and unique biosynthetic gene clusters with the aim of maximizing unique BGCs in the test panel. Cultures were grown in liquid

A1 marine media (10 g starch, 4 g yeast extract, 2 g peptone, 16 g agar, 750 mL 0.2 μm filtered

seawater, 250 mL deionized water) for seven days prior to use in assays.

Table 3.1 Salinispora strains used for multi-strain plates

S. tropica S. arenicola

CNT250 CNY679

CNY012 CNX482

CNS197 CNS205

CNT261 CNS820

CNS416 CNY231

CNB536 CNH963

CNR699 CNQ884

CNH898 CNY280

CNY678 CNY244

CNB440

3.3.2 C. elegans maintenance and sterilization

C. elegans N2 was acquired from the Troemel laboratory at UCSD and maintained on 25 mL agar plates of nematode growth media (NGM) (Fisher scientific) seeded with 300 μL of E. coli

OP50. C. elegans cultures were kept on the lab bench at 20-25°C and transferred to new plates every three days. To generate axenic cultures of C. elegans for use in feeding assays, an adapted protocol from Wormbook was used (Stiernagle, 2006). This procedure involves the lysis of E. coli and C. elegans, leaving only viable C. elegans embryos. Briefly, C. elegans cultures were thoroughly washed with 1 mL of DI water after which 700 μL was transferred to a 1.5 mL

Eppendorf tube. 300 μL of a 2:1 solution of 5% sodium hypochlorite and 5 M sodium hydroxide was aliquoted into the Eppendorf tubes to lyse the adult C. elegans and E. coli. The slurry was vortexed vigorously every 2 minutes for 10 minutes then spun at 1300 x g for 30 seconds in a desktop centrifuge. 800 μL of supernatant was then removed gently so as not to disturb the pellet. 800 μL of fresh DI water was added to the pellet and the sample was vortexed before repeating the centrifugation step three times to thoroughly wash away any remaining bleach/sodium hydroxide solution. After vortexing to distribute the viable embryos in DI water,

10 μL of the egg solution was examined under the dissecting scope to determine egg concentration. The solution was then diluted with sterile DI water to generate a final concentration of 8 embryos/μL.

3.3.3 Live Salinispora spp. assay setup

75 mL of A1 marine media was poured into 150 mm x 15 mm petri plates. Salinispora strains were then cultured alongside an E. coli strain OP50 to provide C. elegans with a choice between the two food sources (Fig. 3.1A). To generate the assay plates, 25 μL of each Salinispora strain (culture conditions described previously) was inoculated onto the agar plates. The cultures were spaced 4 cm apart and allowed to establish for 7 days. One day before C. elegans addition, 25 μL of E. coli strain OP50, grown in 50 mL LB broth Miller (Fisher Scientific) for 2 days, was added 2 cm away from each Salinispora colony so that there was a food source every 2 cm in a row on the agar plates (Fig. 3.2). ~75 C. elegans eggs were then seeded equidistant between each pair of food sources (E. coli vs S. tropica or E. coli vs S. arenicola) (Fig. 3.2). Since C. elegans migrate towards food sources, the number of live C. elegans on each culture was counted using a dissection scope 24 hrs after C. elegans embryo addition as a proxy for feeding. Six replicate

plates were prepared for each Salinispora strain.

Figure 3.1 Workflow for assays studying effects of Salinispora on eukaryotic predators. A) C. elegans feeding preference using live Salinispora spp, either S. tropica (St) or S. arenicola (Sa). B) C. elegans (CE) feeding deterrence by Salinispora extracts, when given a choice between food with Salinispora extract (T) or media control (C). C) effects of Salinispora on Ophryotrocha habitat preference D) marine nematode survivorship in response to Salinispora extracts.

A B

x x x x x x x x x x

Figure 3.2 Live culture C. elegans feeding preference. Orange circles are Salinispora colonies. A) Plate with 10 S. tropica strains added. Black dots indicate sites where E. coli was added. Red x’s mark the placement of C. elegans eggs. B) Assay results for one plate seeded with 9 S. arenicola strains.

To determine if one species of Salinispora was preferred over the other, multi-species plates similar to the ones previously discussed were generated; this time S. tropica was cultured adjacent to S. arenicola (Fig. 3.1A). Again, cultures were grown for 7 days in liquid A1 media, then 25 μL of each culture was spotted 2 cm apart on A1 plates. Strain placement was random for each plate except that an S. arenicola strain was always cultured next to an S. tropica strain.

C. elegans embryos were transferred to the plates 7 days after Salinispora inoculation and the numbers of C. elegans individuals on each colony counted after 24 hrs.

3.3.4 Large scale cultivation and extraction

Four strains of S. tropica and five strains of S. arenicola were selected for large-scale fermentation and extraction to generate crude organic extracts for the feeding assays and bioassay-guided fractionation. These strains were selected randomly from 12 S. tropica strains and 62 S. arenicola strains with sequenced genomes in an effort to get an unbiased representation of the Salinispora spp. Starter cultures were grown from cryovials in 50 mL of A1 marine media (75% seawater) for seven days. For all nine strains, 1 mL from each culture was then transferred and spread on six large petri plates containing 75 mL of A1 with 1.4% agar and grown for 15 days. Agar plates were then chopped into 1x1 cm blocks and shaken with EtOAc in a 2.5 L culture flask for 2 hours. After shaking, the ethyl acetate (EtOAc) was filtered (qualitative p8 fluted filter paper, Fisher) to remove any cell and agar debris, transferred into a round bottom flask, and dried on a rotary evaporator. The extracts were then re-dissolved in 5 mL of EtOAc, transferred to pre-weighed vials, re-dried, and stored at -4°C.

3.3.5 Salinispora extract feeding deterrence

One milligram of extract was dissolved in 50 μL DMSO and transferred to sterile 10 mL glass culture tubes. 1 mL of autoclaved E. coli was then added to each extract generating an extract concentration of 1 mg/mL. This concentration represented, on average, a 2x concentration of the initial extract weights per liter of culture. Media controls were similarly prepared. Since the media control resulted in relatively low yields of extract compared to

Salinispora cultures, volumetric equivalents of media were used instead in the assays. I.e., if 1 mg of Salinispora extract was derived from a 25 mL A1 agar plate, then the control used would be the total extract from 25 mL of A1 agar.

E. coli was grown in 50 mL LB media for 2 days then autoclaved to sterilize the cultures.

25 μL of autoclaved E. coli was mixed with Salinispora crude extract to generate a final concentration of 1 mg/mL crude extract and transferred to a 60 mm x 15 mm petri dish containing 4 mL of NGM. A media control was placed 1 cm apart from the treatment. ~75 C. elegans embryos were placed on each plate equidistant from both food sources, with an average of 6 plates per trial. The number of C. elegans per food source was counted 24 hrs. after C. elegans embryo addition (Fig 1B).

3.3.6 Bioactivity-guided compound isolation

Salinispora strains that produced chemodeterrent extracts against C. elegans were grown in 2 L scale (A1 without agar) to provide enough material for bioassay-guided isolation.

Salinispora starter cultures were generated as described previously. After 7 day, 25 mL of each starter culture was added to a 2.8 culture flasks containing 1 L of A1 marine media. After 12 days with shaking at 195 rpm, ~5 g sterile 1:1 mixture of XAD-7 and XAD-16 Amberlite resin (Fisher

Scientific) was added to each flask. The cultures grew for another 3 days after which the resin and cell mass was filtered out using cheese cloth and eluted using 500 mL of acetone. The

acetone extracts were then dried and tested for feeding deterrence activity at 1 mg/mL. Active crudes were then dissolved in 1:1 hexanes and EtOAc, loaded onto silica, dried down, and subjected to normal phase fractionation on a silica column using the following fractionation scheme: 100% Hexanes, 50%:50% EtoAc: Hexanes, 100% EtOAc, 5%:95 MeOH:DCM, 20%:80%

MeOH:DCM, 100% MeOH. Fractions were then dried under nitrogen and subjected to feeding deterrence assays using concentrations that C. elegans would encounter had they been exposed to the Salinispora cultures. To determine this concentration, the mass of the crude extract or fraction was divided by the volume of the culture (typically 1 L).

Active fractions were subjected to additional rounds of fractionation to isolate the active compound. This was accomplished using reversed phase preparative HPLC with the following conditions: 0-100% acetonitrile (0.1% TFA): 100-0% H2O (0.1% TFA) over 20 minutes with the first 2 minutes of the run discarded. In some cases, HPLC fractions were subjected to another round of preparative HPLC under isocratic conditions: 30:70 H2O (0.1%TFA): ACN

(O.1%TFA). Active fractions were analyzed on the Agilent 1100 series HPLC and the Agilent 6500

Q-TOF MS/MS to get an understanding of the compounds present. MS/MS parameters are described in detail in chapter 2.

3.3.7 Lomaiviticin C quantification and activity

Lomaiviticin C was provided by the Seth Herzon lab (Yale). To quantify lomaiviticin C in the S. tropica cultures, a standard curve was generated by making a dilution series of lomaiviticin

C standard at 0.5 mg/mL, 0.1 mg/mL, 0.075 mg/mL, 0.05 mg/mL, 0.01 mg/mL, and 0.005 mg/mL.

The dilution series was injected on an Agilent HPLC 1100 series in triplicate using parameters described previously, and the area under the curve for the lomaiviticin C peak at 450 nm was calculated. These data were plotted and fit with a linear regression (R2=0.99). The four S. tropica

strain were then run on the same HPLC and the area under the curve for the peak matching that of lomaiviticin C was used to calculate concentration.

The concentration at which lomaiviticin C deterred C. elegans feeding was determined by testing a dilution series (0.5 mg/mL, 0.1 mg/mL, 0.075 mg/mL, 0.05 mg/mL, and 0.01 mg/mL) in the previously described E. coli assay.

3.3.8 Ophryotrocha habitat preference assay

Ophryotrocha n. sp. were maintained in culture by the Greg Rouse lab. Briefly

Ophrytrocha were collected from the effluent line of the moon jelly tanks at the Birch Aquarium,

SIO. Individuals were transferred to containers with frozen spinach as a food source with a seawater drip. Habitat preference assays were designed to provide the test organisms with two agar-based habitats with which they could interact (Fig. 3.1C). For this assay, half of the agar plate contained live Salinispora culture while the other half contained media controls. The live cultures were prepared by as described previously.

To make the bifurcated agar plates, first A1 marine media was added to molten agar at

55°C to create a final concentration of 1.2% agar and 8 mL were poured in 60 mm plates. Once dry, half the agar was removed, and 2 mL of molten agar in seawater (2.4%) at 55°C was combined with 2 mL of live Salinispora culture and added to the plates. Once fully dried, 5

Ophryotrocha were added to the middle of the plates, so that they were touching both the control and treatment substrates and the plate was flooded with 1.5 mL of sterile seawater to facilitate Ophryotrocha movement. Time-lapse photography was used to monitor Ophryotrocha position over one hour and the number of Ophryotrocha present in each habitat counted every five minutes. Five plates were run per trial and plate orientation was randomized to negate influence from abiotic factors such as light sources or airflow. Habitat preference was measured

as a percent of Ophryotrocha per substrate per hour. To test if the activity was due to extractable chemical compounds, this assay was repeated using extracts of Salinispora at 1 mg/mL next to volumetric equivalent concentrations of A1 media control (described previously). Extracts were added to molten 75% seawater agar that had cooled to 55°C to minimize heat induced compound degradation. 8 mL of molten agar containing media extracts (control) were poured in 60 mm plates. Once dry, half the agar was removed and 4 mL of molten agar containing

Salinispora crude extract (1 mg/mL final concentration) was added to the portion of the plate without media. Ophryotrocha were introduced to the plates and monitored as described previously.

3.3.9 Marine nematode assays

Sediment was collected by bucket around the Carrie Bow Cay field station in Belize in

September 2015. The sediment was rigorously mixed by hand and the overlying seawater decanted through a 2 mm mesh to collect suspended biota. The samples were transferred to a petri dish and examined under a dissecting scope. Marine nematodes were recognized based on morphology, picked out of the samples by forceps and transferred to a container with seawater. To determine the effects of Salinispora extracts on these nematodes, marine agar plates were prepared containing 1 mg/mL concentration of either an S. tropica (CNB440) extract or an S. arenicola (CNS679) extract. Control plates containing volumetric equivalents of A1 media extract were also prepared. Ten nematodes were placed on each plate and three plates were prepared for each species. Nematode survivorship was monitored over six hours on each plate. Nematodes were considered dead if they were unable to respond to prodding (Fig 3.1D).

3.3.10 Statistical analysis

T-tests or Wilcoxon rank-sum tests were used to analyze the data. If sample sizes were

large (n>8) and the distribution was relatively normal based on a histogram analysis in R, a student’s t-test was used to evaluate significant differences between treatments or treatments and controls. For sample sizes with n<8, a non-parametric t-test was used (Wilcoxon rank-sum test) (Wilcoxon, 1945). These analyses were conducted in R (R Development Core Team, 2011), and graphs were generated in Microsoft Excel. All graphs contain standard deviation error bars.

3.4 Results and Discussion

3.4.1 Salinispora mediated C. elegans feeding preference

C. elegans feeding preference was tested against a total of 18 Salinispora strains. The first assay was designed to determine if C. elegans preferred live cultures of Salinispora species or E. coli as a food source (Fig. 3.1A). The results reveal a marked preference for E. coli cultures

(Fig. 3.3). 89.5% of the C. elegans localized to the E. coli cultures, which can be viewed as a proxy for feeding, whereas only 10.5% were observed on the Salinispora colonies (p=0.00001, student’s T-test, n=42). Preference was monitored over several time points (24, 30, and 36 hrs.) of the first trial, but no noticeable change in C. elegans preference was observed. Therefore, 24 h was used for all future trials. The results provided evidence that live Salinispora cultures are less favored by C. elegans and open the possibility that this difference is chemically mediated.

Interestingly, I noted morphological differences between the two Salinispora species, with S. arenicola cultures typically forming numerous small colonies whereas S. tropica cultures were more densely compacted with little space between colonies (Fig. 3.2).

Since Salinispora species have unique chemical and morphological phenotypes, I next evaluated if C. elegans preferred S. tropica or S. arenicola. In this case, C. elegans were given the choice between colonies of S. tropica and S. arenicola on agar plates. Significantly more C.

elegans were observed interacting with the S. arenicola colonies than with the S. tropica colonies in both trials (Fig. 3.4). For trial one, 73% of C. elegans were on S. arenicola colonies

(n=54, p<0.01), and in trial two, 84% of C. elegans were on S. arenicola colonies (n=54, p<0.0001). This indicates that C. elegans has a clear preference for S. arenicola over S. tropica.

Figure 3.3 Salinispora / E. coli feeding preference assay. A) Percent of C. elegans on A1 agar plates containing either S. tropica or E. coli food sources. B) Percent of C. elegans on A1 agar plates containing either S. tropica or E. coli food sources. Error bars represent standard error.

Figure 3.4 Salinispora spp. feeding preference assay. Bar graphs represent percent of C. elegans per food source with error bars representing standard error A) Percent of C. elegans on either S. tropica or S. arenicola food sources for trial 1 B) Percent of C. elegans on either S. tropica or S. arenicola food sources for trial 2

To determine if Salinispora feeding preference was due to extractable metabolites produced by the less preferred species, it was necessary to develop a feeding assay that could accommodate organic extracts. Since Salinispora spp. extracts are rich in bioactive metabolites, using live E. coli cultures as the food source was not an option as the extracts would likely be lethal to the E. coli. As an alternative, I was able to show that C. elegans will readily feed on autoclaved E. coli and was able to develop an alternative assay using this food source.

Preliminary studies showed that C. elegans was capable of surviving on the autoclaved food source until it was depleted (>2 days for 15 µL with ~75 individuals).

Salinispora extracts were tested at 1 mg/mL, which, on average, was a 2x concentration of the extract. Extracts were concentrated 2x to be slightly liberal with the initial feeding deterrence profiling to emphasize the effects of compounds that may exhibit weak deterrent properties. S. tropica crude extracts at 1 mg/mL deterred feeding by C. elegans relative to the E. coli control in three of four S. tropica strains tested (Fig. 3.5B). Although strain CNB440 was not deterrent (student’s t-test p= 0.397), S. tropica as a species exhibits significantly more feeding deterrence (student’s t-test p value= 2.5x10-6) than S. arenicola, suggesting that specific metabolites produced by this species may be responsible for the deterrence. The levels of deterrence detected in the live culture assays were not reproduced in the extract assays indicating that there may also be other factors that contribute to the activity or that the extraction process was not exhaustive in capturing the deterrent compounds.

For three of the five S. arenicola strains tested, crude chemical extracts exhibited chemoattractant properties relative to media controls, indicating that C. elegans has a preference for crude extracts from strains CNH820, CNY679, and CNS205 at the concentration tested (n=8, p<0.05) (Fig. 3.5A). All three of these strains possess the BGC for the highly cytotoxic kinase inhibitor staurosporine, making C. elegans’ preference for these extracts surprising. This

lack of deterrence could be a result of several things: staurosporine is not very effective against

C. elegans, the S. arenicola strains were producing this compound at low concentrations or not at all, or S. arenicola produces compounds with such strong chemoattractant properties that C. elegans feed on these extracts despite their cytotoxic properties. Since mortality rates were not monitored in this assay, it is not known if these chemoattractive extracts ultimately resulted in

C. elegans mortality.

Figure 3.5 Results of Salinispora spp. secreted metabolite mediated feeding assays. Bar graphs represent percent of C. elegans on treatment or control autoclaved E. coli food source. A1 media was used as a control extract, and the right most pairing in the bar graph represents percent C. elegans on either a food with A1 extract or solvent control. All error bars represent standard error and asterisks denote Student’s T-test p values <0.05 A) Percent of C. elegans feeding on either autoclaved E. coli containing S. arenicola crude extracts or A1 media extracts. B) Percent of C. elegans feeding on either autoclaved E. coli containing S. tropica crude extracts or A1 media extracts. Since S. tropica extracts exhibited feeding deterrence, I wanted to determine if any specific compound(s) could be linked to this activity. All three deterrent S. tropica strains were regrown in 2 L volume, extracted, and the extracts tested for activity. The crude extract from all three strains was active, so strain CNY012, which produced the most crude material, was picked for bioactivity guided fractionation. Normal phase flash fractionation resulted in one fraction

(100% MeOH) containing all observed activity (Fig. 3.6, 3.7A). Further purification using

reversed phase HPLC separation under a simple acetonitrile:water gradient resulted in the generation of three fractions. Activity testing indicated that the second fraction (8-14 minutes) contained the feeding deterrence compounds (Fig. 3.7B). Although this fraction was still rather complex, subsequent fractionation did not result in better purification, and it was determined that the active compound(s) were breaking down. Analysis of fraction 2 indicated that it contained lomaiviticins A and C, which were confirmed by UV and mass (Fig. 3.6b, 3.6C).

However, there also appeared to be other lomaiviticins present including likely breakdown products in this fraction. All of the described lomaiviticin compounds have been shown to have cytotoxic properties with lomaiviticin A inhibiting numerous cell lines at sub-nanomolar concentrations (Woo et al., 2012).

The lomaiviticins were unstable and co-eluted during fractionation, which hindered our ability to successfully purify the analogues. Therefore, a lomaiviticin C standard was acquired from the Seth Herzon lab (Yale) and tested for bioactivity. Lomaiviticin C, although the least cytotoxic lomaiviticin analogue reported to date due to the lack of one diazo group (Woo et al.,

2012), still exhibited C. elegans feeding deterrence at 0.05 mg/mL (Fig. 3.7D).

Figure 3.6 Bioassay guided fractionation leading to the isolation of lomaiviticins as the deterrent compounds in the S. tropica extracts. A) Bioassay guided fractionation scheme from the crude CNY012 extract to the active fraction containing a suite of lomaiviticins. Indicated are the conditions for each round of fractionation. A green check mark indicates activity on the C. elegans feeding deterrence assay. Also provided are the weights of each fraction. B) Base peak chromatogram of the 100% MeOH Fr 2 active fraction containing lomaiviticins. The masses are provided for two of the most abundance peaks in the chromatogram which match lomaiviticin C (670.59 m/z) and lomaiviticin D (677.59 m/z). C) Absorbance spectra for the 100% MeOH Fr2 fraction, both ELSD and 254 nm are shown. The major peak at 254 nm matches the standard for lomaiviticin C as indicated by the library spectra match highlighted. Efforts were then taken to identify and quantify lomaiviticin C in the S. tropica extracts.

The three deterrent S. tropica strains all contained lomaiviticin C, however HPLC analysis indicated that there was no lomaiviticin C in the extract from strain CNB440, the only S. tropica strain that did not deter C. elegans (Fig. 3.8). Lomaiviticin C abundance was quantified using a standard curve (Fig 3.9A). Strain CNY012 extracts contained the highest quantity of lomaiviticin

C at 8.2 µM, with CNS197 and CNH898 extracts containing 5.2 µM and 3.4 µM respectively (Fig.

3.9B), falling well above previously published IC50 values for lomaiviticin C cytotoxicity against cancer cell lines (0.2-0.9 µM) (Woo et al., 2012). However, these concentrations were lower than those that inhibited C. elegans (0.05 mg/mL or 36 µM). There were other lomaiviticins present in the sample (e.g., lomaiviticin A) that possess more potent cytotoxicity, which may work in consortium with lomaiviticin C to deter feeding by C. elegans. Purifying other lomaiviticin analogues is needed to determine if and how these compounds work together to deter C. elegans.

Figure 3.7 Activity of the CNY012 fractions against C. elegans feeding assays and lomaiviticin C standard minimum deterrence calculation. Bar graphs represent percent of C. elegans on treatment (T) or control (C) autoclaved E. coli food source. A1 media extract was used as a control. All error bars represent standard error and asterisks denote Student’s T-test p values <0.05. A) Percent of C. elegans feeding on either autoclaved E. coli containing CNY012 normal phase fractions or A1 media extracts. B) Percent of C. elegans feeding on either autoclaved E. coli containing Reverse phase fractions of the CNY012 100% MeOH fraction or A1 media extracts. C) Percent of C. elegans feeding on either autoclaved E. coli containing lomaiviticin C at several different concentrations or A1 media extracts.

Retention time (min)

Figure 3.8 UV profile at 450 nM for the four S. tropica strains tested in the C. elegans deterrence assays as well as the lomaiviticin C standard. Peaks that matched the lomaiviticin C standard are circled. Retention time is offset to better visualize the individual peaks.

Figure 3.9 Quantification of lomaiviticin C in the S. tropica strains. A) Lomaiviticin C standard curve with equation and R2 value. B) Bar graph depicting concentration of lomaiviticin C (µM) in the four S. tropica strains tested. Error bars are standard deviation

3.4.2 Feeding assays using marine predators

An assay was designed to assess whether Salinispora cultures or organic extracts affected the habitat preference of the polychaete Ophryotrocha n. sp. Since Ophryotrocha are grazers, they migrate towards areas rich in palatable bacteria and detritus. The purpose of this assay was to determine if Ophryotrocha avoided areas containing Salinispora cultures or extracts. Bifurcated agar plates containing live Salinispora or Salinispora extract on one half of the plate, and the other half containing media controls were seeded with Ophryotrocha and their position monitored over one hour. Given the choice of substrates infused with live S. tropica cultures or media controls, Ophryotrocha spent the majority of time in the control habitat (n=13, p<0.0005) (Fig. 3.10A). To determine if the preference was due to secreted compounds, S. tropica extracts were similarly tested and were also determined to have deterrent properties (n=10, p<0.001) (Fig. 3.10B). S. arenicola plates were also avoided compared to controls (live S. arenicola n=13, p<0.05, S. arenicola extract n=10, p=0.26) (Fig.

3.10C&D). However, S. tropica extracts were less preferred than those of S. arenicola. The fact that S. tropica extracts are less preferred by a polychaete and deter feeding by C. elegans suggests it produces secondary metabolites that deter diverse types of bacterivorous eukaryotes.

Figure 3.10 Results of Ophryotrocha habitat preference assays. Bar graphs represent percent of time over an hour that Ophryotrocha spent on each habitat with error bars representing standard error. Asterisks indicate statistical significance based on a student’s t-test A) Percent of time over an hour Ophryotrocha spent on either a habitat containing S. tropica live culture or A1 media (control). B) Percent of time over an hour Ophryotrocha spent a habitat containing S. tropica extract or A1 media extract (control). C) Percent of time over an hour Ophryotrocha spent on either a habitat containing S. arenicola live culture or A1 media (control). D) percent of time over an hour Ophryotrocha spent on a habitat containing S. arenicola extract or A1 media extract.

Since C. elegans is not a marine species, and the Ophryotrocha used were not derived from the tropical sediments where Salinispora are typically found, I decided to test the effects of Salinispora extracts on nematodes obtained from sediment samples collected off the coast of Belize where Salinispora has been cultured (from Chpt 2). Unfortunately, due to the way these nematodes interacted with the agar substrate, it was not possible to perform habitat preference assays similar to those used for Ophryotrocha sp. However, I noticed that any marine nematode that interacted with agar containing S. tropica extracts died within several hours. Building from

this observation, I performed nematode survivorship assays on treated and control agar plates over a time course of 6 hours. When subjected to agar containing 1 mg/mL of crude extract from

S. tropica strain CNB440, marine nematode survivorship was reduced to 0% within 6 hours during trial 1 and 15% in trial 2 (Fig 3.11). In contrast, nematode survivorship on control plates containing volumetrically equivalent concentrations of media extract was 100% and they survived on these plates for several days.

S. arenicola strain CNS205 crude extract did not exhibit similar nematode mortality levels as those generated from S. tropica, and at 1 mg/mL concentrations survivorship remained above 90% (Fig 3.11). However, if the extract concentration was increased three-fold to 6 mg/mL, the marine nematodes perished (data not shown). This indicates that S. arenicola produces cytotoxic metabolites, but they could be relatively low in concentration or not nearly as potent as those produced by S. tropica. S. tropica strain CNB440 was used in these nematode assays because they were conducted before the C. elegans crude extract trials revealed that this strain had the weakest activity of all S. tropica strains tested. If one of the other S. tropica strains from the C. elegans assay is tested in a future field collection trip, nematode survivorship will likely be affected at a much lower concentration.

100%

80%

A1 control 60% S. tropica trial 1 S. tropica trial 2 40% S. arenicola trial 1

20% Percent of nematodes alive ofnematodes Percent 0% 1 2 3 4 5 6

Hours after inoculation

Figure 3.11 Survivorship curves for benthic marine nematodes on agar with either S. tropica strain CNB440 crude extract at 1 mg/ml, S. arenicola strain CNS679 crude extract at 1 mg/ml, or agar with a volumetric equivalent of A1 media extract. Error bars represent standard deviation over the replicates, and the assay was performed twice with S. tropica extracts and once with S. arenicola extracts, with both results depicted on the same graph (trial 1 and trial 2).

3.5 Discussion

It has been well documented that sediment bacteria are prolific producers of cytotoxic specialized metabolites (Burja et al., 2001; Jensen et al., 2015), however the ecological roles of these compounds remain largely speculative. Cytotoxins produced by phototrophic cyanobacterial mats have been shown to deter grazing from macro-organisms (Cruz-Rivera and

Paul, 2007), but there has been little effort to address the ecological functions of cytotoxic compounds produced by heterotrophic sediment bacteria. This study implemented a series of novel predator preference assays to determine the effects of the sediment inhabiting marine actinomycete genus Salinispora on the feeding preferences of eukaryotic bacterivores. To our

knowledge, this is the first evidence that compounds produced by heterotrophic, sediment- inhabiting marine bacteria deter eukaryotic predators.

A limitation with microbial chemical ecology studies looking at predator deterrence is in the lack of established assays to study these intradomain interactions. For instance, C. elegans bacterial preference has been studied in detail, but there has not been an easily adoptable assay developed to determine the effects of specialized metabolites on C. elegans food preference.

Further, no assays had been developed to assess the effects of allelochemicals on Ophyrotrocha feeding or migration patterns. This study was the first to address these issues, resulting in the development of simple assays that can be conducted in field to help identify chemodeterrence.

Nonetheless, there are several limitations to these analyses in terms of ecological relevance. For one, the Salinispora strains were grown in nutrient rich media. As a result, they likely produce metabolites in a much higher abundance than that found in nature. In addition, it is unknown if the lomaiviticins are produced in situ, so they may not aid in predator deterrence. A targeted metabolomic analysis of the sediments looking for lomaiviticins, may add additional support to their proposed ecological role. Efforts are needed to better determine ecologically relevant test concentrations.

Molecular analyses have revealed that the sediment bacterial community possess an incredible amount of diversity. Interestingly, numerous species tend to co-occur despite possessing similar genetic makeup, indicating that there are forces driving speciation at a relatively small spatial scale. Bacterial lineages may fall into ecologically cohesive units or ecotypes that experience higher levels of homologous gene transfer within than between clades, leading to the co-occurrence of relatively similar bacterial species (Cohan, 2002). These ecotypes are maintained by ecological pressures. An example of this was recently shown within

the genus Salinispora. Both S. arenicola and S. tropica co-occur with >99% 16S similarity, suggesting they inhabit different niches allowing them to remain unique species. it was shown that these species exhibit different competitive strategies against prokaryotes in the environment. S. arenicola, by the production of antagonistic specialized metabolites and S. tropica, by fast growth and the production of siderophores to deplete iron in the environment

(Patin et al., 2015). This study further supports these findings, revealing that S. tropica inhibits predation, despite S. arenicola remaining susceptible. This would indicate that S. arenicola’s specialized metabolome makes it better suited for habitats rich in prokaryotes but with low predation, whereas S. tropica can better compete and grow in areas with higher eukaryotic predation.

It was not surprising that the lomaiviticins were implicated in S. tropica’s feeding deterrence. These compounds are potent cytotoxins which cause double stranded DNA breaks at nanomolar to subnanomolar concentrations (Colis et al., 2014). Also, the lomaiviticin BGC is present in all S. tropica strains with sequenced genomes, whereas the lom BGC is absent in all S. arenicola strains (Letzel et al., 2017). The fact that the lom BGC is fixed in S. tropica indicates that the products it encodes likely serve an important ecological role hence its conservation within the species. However, it cannot be ruled out that other compounds may aid in the observed chemodeterrence. Although these compounds may not exhibit deterrent properties at the concentrations they are produced when tested in isolation, similar to the lomaiviticins, they may aid in the observed deterrence of the crude extract through some unexplored synergistic property, thus adding to the complexity of metabolite mediated ecological specialization. Future studies should examine the effects of multiple specialized metabolites in consortium on the effects of predator deterrence.

In chapter 2 I discovered that staurosporine was abundant in the benthic environment and that S. arenicola was likely a producer of this cytotoxin. This led me to believe that S. arenicola would likely deter predation through the production of staurosporine. Interestingly, I found that S. arenicola extracts were not nearly as deterrent as those produced by S. tropica. In fact, the Salinispora compound that appeared to act as the main predator deterrent was lomaiviticin, not staurosporine. However, staurosporine still exhibited chemodeterrent properties at high concentrations, as seen in chapter 2, which may be realized in the sediment at the spatial scale that Salinispora and its predators occur. it would be interesting to see if lomaiviticins were being produced in the sediments, despite them not being detected in the sediment metabolomic analysis. Lomaiviticins typically do not ionize well and little is known about their binding/extraction efficiency, so control experiments should be conducted to determine the best means to isolate these compounds from the environment.

Since Salinispora spp. are capable of producing compounds that affect the habitat and/or feeding preference of eukaryotic predators, they could affect the structure of the overall benthic microbial community. Eukaryotic community dynamics are probably influenced not only by available food sources but also the environmental conditions created by other community members. Therefore, if Salinispora strains are secreting compounds that create a spatial barrier to eukaryotes, other more susceptible microbes are likely shielded from predation through associational defense (Hay, 1986). These results not only indicate the importance in understanding the effects of secreted metabolites on the community, but also indicate the need to understand the relative abundance of these compounds in nature. If cytotoxins are only being produced at sub-inhibitory concentrations, then they may have more subtle effects on predators, such as behavioral effects or habitat avoidance (Wahl et al., 1994). Further research

is needed to better understand if these compounds are present and if they are, their concentrations.

3.6 Acknowledgments

Thanks to the Troemel lab (UCSD) for donating the C. elegans and E. coli strain OP50 for use in the deterrence assays, as well as Bethany Kolody for the inspiration to use C. elegans. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Chapter 3 is currently being formatted into a manuscript that will be submitted in fall of

2018. Robert Tuttle, Katrina Forrest, Gabriel Castro, Chambers Hughes, Greg Rouse, Paul Jensen.

The dissertation author was the primary investigator on these studies.

3.7 References

Alongi DM. 1988. Bacterial productivity and microbial biomass in tropical mangrove sediments. Microb Ecol 15:59–79. doi:10.1007/BF02012952

Andresen M, Kristensen E. 2002. The importance of bacteria and microalgae in the diet of the deposit-feeding polychaete Arenicola marina. Ophelia 56:179–196. doi:10.1080/00785236.2002.10409498

Arrigo KR. 2005. Marine microorganisms and global nutrient cycles. Nature 437:343–348. doi:10.1038/nature04158

Azam F, Fenchel T, Field J, Gray J, Meyer-Reil L, Thingstad F. 1983. The Ecological Role of Water- Column Microbes in the Sea. Mar Ecol Prog Ser 10:257–263. doi:10.3354/meps010257

Bargmann CI. 2006. Chemosensation in C. elegans. WormBook 1–29. doi:10.1895/wormbook.1.123.1

Blunt JW, Copp BR, Munro MHG, Northcote PT, Prinsep MR. 2015. Marine natural products. Nat Prod Rep 28:196–268. doi:10.1039/C005001F

Bouvier T, del Giorgio PA. 2007. Key role of selective viral-induced mortality in determining marine bacterial community composition. Environ Microbiol 9:287–297. doi:10.1111/j.1462-2920.2006.01137.x

Brenner S. 1974. The genetics of Caenorhabditis elegans. Genetics 77:71–94. doi:10.1002/cbic.200300625

Burja AM, Banaigs B, Abou-Mansour E, Grant Burgess J, Wright PC. 2001. Marine cyanobacteria - A prolific source of natural products. Tetrahedron. doi:10.1016/S0040-4020(01)00931-0

Cohan FM. 2002. What are Bacterial Species? Annu Rev Microbiol 56:457–487. doi:10.1146/annurev.micro.56.012302.160634

Cohen Y. 2002. Bioremediation of oil by marine microbial mats. Int Microbiol 5:189–193. doi:10.1007/s10123-002-0089-5

Colis LC, Woo CM, Hegan DC, Li Z, Glazer PM, Herzon SB. 2014. The cytotoxicity of (-)-lomaiviticin A arises from induction of double-strand breaks in DNA. Nat Chem. doi:10.1038/nchem.1944

Corno G, Jürgens K. 2006. Direct and indirect effects of protist predation on population size structure of a bacterial strain with high phenotypic plasticity. Appl Environ Microbiol 72:78–86. doi:10.1128/AEM.72.1.78-86.2006

Cruz-Rivera E, Paul VJ. 2007. Chemical deterrence of a cyanobacterial metabolite against generalized and specialized grazers. J Chem Ecol 33:213–217. doi:10.1007/s10886-006- 9212-y

Fenical W, Jensen PR. 2006. Developing a new resource for drug discovery: Marine actinomycete bacteria. Nat Chem Biol. doi:10.1038/nchembio841

Fiedler HP, Bruntner C, Bull AT, Ward AC, Goodfellow M, Potterat O, Puder C, Mihm G. 2005. Marine actinomycetes as a source of novel secondary metabolites. Antonie van Leeuwenhoek, Int J Gen Mol Microbiol 87:37–42. doi:10.1007/s10482-004-6538-8

Gasol JM, Pedrós-Alió C, Vaqué D. 2002. Regulation of bacterial assemblages in oligotrophic plankton systems: results from experimental and empirical approaches. Antonie Van Leeuwenhoek 81:435–452. doi:10.1023/a:1020578418898

Girard LR, Fiedler TJ, Harris TW, Carvalho F, Antoshechkin I, Han M, Sternberg PW, Stein LD, Chalfie M. 2007. WormBook: the online review of Caenorhabditis elegans biology. Nucleic Acids Res 35:D472–D475. doi:10.1093/nar/gkl894

Hay ME. 1986. Associational Plant Defenses and the Maintenance of Species Diversity: Turning Competitors Into Accomplices. Am Nat 128:617. doi:10.1086/284593

Heip C, Vincx M, Vrank. 1985. The ecology of marine nematodes. Ocean Mar Biol Ann Rev.

Jensen PR, Moore BS, Fenical W. 2015. The marine actinomycete genus Salinispora: a model organism for secondary metabolite discovery. Nat Prod Rep 00:1–14. doi:10.1039/c4np00167b

Jensen PR, Williams PG, Oh DC, Zeigler L, Fenical W. 2007. Species-specific secondary metabolite

100

production in marine actinomycetes of the genus Salinispora. Appl Environ Microbiol 73:1146–1152. doi:10.1128/AEM.01891-06

Jousset A. 2012. Ecological and evolutive implications of bacterial defences against predators. Environ Microbiol 14:1830–1843. doi:10.1111/j.1462-2920.2011.02627.x

Jousset A, Lara E, Wall LG, Valverde C. 2006. Secondary metabolites help biocontrol strain Pseudomonas fluorescens CHA0 to escape protozoan grazing. Appl Environ Microbiol 72:7083–90. doi:10.1128/AEM.00557-06

Klobutcher L a, Ragkousi K, Setlow P. 2006. The Bacillus subtilis spore coat provides “eat resistance” during phagocytic predation by the protozoan Tetrahymena thermophila. Proc Natl Acad Sci U S A 103:165–70. doi:10.1073/pnas.0507121102

Lainhart W, Stolfa G, Koudelka GB. 2009. Shiga Toxin as a Bacterial Defense against a Eukaryotic Predator, Tetrahymena thermophila. J Bacteriol 191:5116–5122. doi:10.1128/JB.00508-09

Matz C, Webb JS, Schupp PJ, Phang SY, Penesyan A, Egan S, Steinberg P, Kjelleberg S. 2008. Marine Biofilm Bacteria Evade Eukaryotic Predation by Targeted Chemical Defense. PLoS One 3:e2744. doi:10.1371/journal.pone.0002744

Mazzola M, De Bruijn I, Cohen MF, Raaijmakers JM. 2009. Protozoan-induced regulation of cyclic lipopeptide biosynthesis is an effective predation defense mechanism for Pseudomonas fluorescens. Appl Environ Microbiol 75:6804–6811. doi:10.1128/AEM.01272-09

Moens T, Verbeeck L, De Maeyer A, Swings J, Vincx M. 1999. Selective attraction of marine bacterivorous nematodes to their bacterial food. Mar Ecol Prog Ser 176:165–178. doi:10.3354/meps176165

Nagle DG, Paul VJ. 1998. Chemical defense of a marine cyanobacterial bloom. J Exp Mar Bio Ecol 225:29–38. doi:10.1016/S0022-0981(97)00205-0

Nagle DG, Paul VJ, Roberts MA. 1996. Ypaoamide, a new broadly acting feeding deterrent from the marine cyanobacterium Lyngbya majuscula. Tetrahedron Lett 37:6263–6266. doi:10.1016/0040-4039(96)01391-3

Patin N V, Duncan KR, Dorrestein PC, Jensen PR. 2015. Competitive strategies differentiate closely related species of marine actinobacteria. ISME J 10:1–13. doi:10.1038/ismej.2015.128

Pennings SC, Weiss AM, Paul VJ. 1996. Secondary metabolites of the cyanobacterium Microcoleus lyngbyaceus and the sea hare Stylocheilus longicauda: palatability and toxicity. Mar Biol 126:735–743. doi:10.1007/BF00351340

101

Pernthaler J. 2005. Predation on prokaryotes in the water column and its ecological implications. Nat Rev Microbiol 3:537–546. doi:10.1038/nrmicro1252

Queck S-Y, Weitere M, Moreno AM, Rice SA, Kjelleberg S. 2006. The role of quorum sensing mediated developmental traits in the resistance of Serratia marcescens biofilms against protozoan grazing. Environ Microbiol 8:1017–25. doi:10.1111/j.1462-2920.2006.00993.x

R Development Core Team R. 2011. R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing. doi:10.1007/978-3-540-74686-7

Riddle DL, Blumenthal T, Meyer BJ, Priess JR. 1997. C. elegans IICold Spring Harbor (NY): Cold Spring Harbor Laboratory Press. doi:NBK20183 [bookaccession]

Schulenburg H, Ewbank JJ. 2007. The genetics of pathogen avoidance in Caenorhabditis elegans. Mol Microbiol 66:563–570. doi:10.1111/j.1365-2958.2007.05946.x

Shtonda BB, Avery L. 2006. Dietary choice behavior in Caenorhabditis elegans. J Exp Biol 209:89– 102. doi:10.1242/jeb.01955

Stiernagle T. 2006. Maintenance of C. elegans. WormBook 1–11. doi:10.1895/wormbook.1.101.1

Struck TH, Paul C, Hill N, Hartmann S, Hösel C, Kube M, Lieb B, Meyer A, Tiedemann R, Purschke G, Bleidorn C. 2011. Phylogenomic analyses unravel annelid evolution. Nature 471:95–98. doi:10.1038/nature09864

Tsuchiya M, Kurihara Y. 1979. The feeding habits and food sources of the deposit-feeding polychaete, Neanthes japonica (Izuka). J Exp Mar Bio Ecol 36:79–89. doi:10.1016/0022- 0981(79)90101-1

Wahl M, Jensen PR, Fenical W. 1994. Chemical control of bacterial epibiosis on ascidians. Mar Ecol Prog Ser 110:45–58. doi:10.3354/meps110045

Weitere M, Bergfeld T, Rice SA, Matz G, Kjelleberg S. 2005. Grazing resistance of Pseudomonas aeruginosa biofilms depends on type of protective mechanism, developmental stage and protozoan feeding mode. Environ Microbiol 7:1593–1601. doi:10.1111/j.1462- 2920.2005.00851.x

Wietz M, Duncan K, Patin N V., Jensen PR. 2013. Antagonistic Interactions Mediated by Marine Bacteria: The Role of Small Molecules. J Chem Ecol 39:879–891. doi:10.1007/s10886-013- 0316-x

Wilcoxon F. 1945. Individual Comparisons by Ranking Methods. Biometrics Bull 1:80. doi:10.2307/3001968

Woo CM, Beizer NE, Janso JE, Herzon SB. 2012. Isolation of Lomaiviticins C–E, Transformation of Lomaiviticin C to Lomaiviticin A, Complete Structure Elucidation of Lomaiviticin A, and Structure–Activity Analyses. J Am Chem Soc 134:15285–15288. doi:10.1021/ja3074984

102

Ziemert N, Lechner A, Wietz M, Millan-Aguinaga N, Chavarria KL, Jensen PR. 2014. Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci 111:E1130–E1139. doi:10.1073/pnas.1324161111

103

Chapter 4

Elucidating the Salinispora “inductome” through prokaryotic co-culturing

104

4.1 Abstract

The genomics era has provided unprecedented insight into the number of biosynthetic gene clusters found in bacterial species. This has led to the understanding that many bacterial species possess enormous biosynthetic potential to make specialized metabolites. However, compound isolation from classic culturing techniques typically does not yield the number of metabolites that would be suggested based on genomic content. One reason for this discrepancy is that specialized metabolite production can be induced by environmental stressors that are not integrated into classic culturing paradigms. Numerous species from the chemically important bacterial order Actinomycetales have been shown to have a highly inducible metabolome when co-cultured with other bacteria. However, few studies have been conducted in a high throughput manner to determine the overall inductome of actinomycetes in response to bacterial challengers. Using six bacterial challenge strains, I conducted a high throughput co-culturing study with 20 strains of the obligate marine actinomycete genus

Salinispora. Results indicate that induction is both strain specific with different induced mass features found in co-cultures using every Salinispora strain as well as every bacterial challenger strain. Efforts were then taken to dereplicate known compounds from the Salinispora coculture inductome, revealing that desferrioxamines were shown to be induced in a Salinispora tropica strain in response to Streptomyces challengers as well as solvent stress. This indicates that that some Salinispora strains regulate siderophore production in response to stressors including specific bacterial competitors.

105

4.2 Introduction

In chapter 3, I provided evidence that Salinispora tropica secondary metabolites deter predatory eukaryotes. For these assays, crude extracts derived from Salinispora monocultures were tested and used for bioassay guided isolation of the deterrent compounds. This study focused solely on constitutively expressed secondary metabolites and ignored the possibility that Salinispora strains regulate compound expression in response to environmental stimuli. In this chapter, I looked at the possibility of inducing secondary metabolite production in response to bacterial competitors.

With the recent advances in genetic sequencing technology coupled with the rapid decrease in cost, an unprecedented number of genomes have become available for data mining

(Benson et al., 2015). An interesting finding from these data sets is that numerous bacterial species possess more biosynthetic gene clusters (BGCs) that likely code for natural products than the compounds that have been isolated from them to date would suggest (Bentley et al., 2002;

Janso and Carter, 2010; Letzel et al., 2017; Omura et al., 2001; Ziemert et al., 2014). One reason for this discrepancy is that these “orphan” BGCs no longer produce compounds. However, the bacterial genome is typically under strong selective pressure to eliminate traits that do not provide a selective advantage and thus vestigial or noncoding genes are often removed

(Lawrence et al., 2001). Another theory is that these pathways are under strict regulatory control and are therefore only produced under specific conditions. Since the production of secondary metabolites tends to be an energetically intensive process, constitutive expression of chemical compounds would likely hinder producer fitness (Callahan et al., 2008). It would thus make evolutionary sense for metabolite producing microbes to be able to regulate chemical synthesis such that induction occurs in response to specific biotic stressors (Callahan et al., 2008). Recent studies have shown metabolite induction is quite prevalent in the microbial realm (Bertrand et

106

al., 2013; Garbeva et al., 2011; Jousset et al., 2010, 2008; Patin et al., 2018; Seyedsayamdost,

2014; Slattery et al., 2001; Trischman et al., 2004). One of the first examples of this was the enhancement of antimicrobial activity when marine microbes were co-cultured with terrestrial pathogens (Mearns-Spragg et al., 1998). Other co-culturing experiments have further demonstrated the inducibility of chemical compounds, including the production of pyocyanin by

P. aeruginosa when grown in solution with Enterobacter species (Angell et al., 2006), and the production of antimicrobial compounds by P. flourescens when co-cultured with Pedobacter and

Brevundimonas strains (Garbeva et al., 2011).

Actinomycetes are not only prolific producers of secondary metabolites, but also have been shown to have an inducible metabolome. Numerous bacterial species have been shown to induce compound production in actinomycetes including mycolic acid containing bacteria,

Gram-negative species, and specialized metabolite producers (Okada and Seyedsayamdost,

2017). S. coelicolor has been shown to recognize iron chelators from microbial competitors in the environment and produce its own suite of ferrioxamines in response. This induction can be incredibly specific with unique siderophores being produced in response to each competitor, indicating that compound induction can be species specific in actinomycetes (Traxler et al.,

2013). Mycolic acid containing bacteria have been shown to induce pigment production as well as novel antibiotics in S. lividens (Onaka et al., 2011). Even Vibrio species have been shown to induce compound production in actinomycetes (Patin et al., 2018). Induction is not limited to just bacteria-bacteria interactions but has also been observed in response to predation. For instance, Pseudomonas fluorescens lipopetide biosynthesis was induced in response to protozoan predation (Mazzola et al., 2009; Song et al., 2015).

A recent analysis of 119 Salinispora genome sequences identified 176 distinct secondary metabolite pathways and led to estimates that the genus contains upwards of 250 (Letzel et al.,

107

2017). Yet only 28 compounds have been discovered to date from this genus. However a transcriptomic analysis on several Salinispora strains revealed that a large number of secondary metabolite BGCs are not expressed when the strains are grown in mono-culture (Amos et al.,

2017). The discrepancy between compounds isolated and BGCs could be attributed to several reasons: some compounds may be produced in too low yield to be detected by the methods employed, some compounds may not be extractable using classic extraction protocols, some pathways may be nonfunctioning, and some pathways may require be regulated and require signals to induce production.

Salinispora inhabit tropical marine sediments. Microbial diversity in sediments is quite high with up to 109 individuals occurring in a gram of sediment (Alongi, 1988). Bacteria have developed numerous strategies to compete in these environments. Salinispora have been shown to produce metabolites that inhibit predatory eukaryotes, antimicrobial agents, and chelator compounds, (Patin et al., 2015; Roberts et al., 2012). However, due to the dynamic nature of these environments, there is likely temporal variability in biotic pressures on

Salinispora. It would thus be energetically favorable for Salinispora to express compounds on a more targeted basis. There is some evidence to support this, with secondary metabolite production in one strain of S. tropica increasing when co-cultured with a gram-negative Vibrio strain (Patin et al., 2018).

Induction can be assessed using MS-based metabolomics, such as molecular networking which allows for the comparison of metabolomes. Molecular networking is based on the acquisition of metabolomic data from bacterial extracts that are analyzed by high pressure liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). LC-MS/MS has proven to be an effective technique to separate and acquire mass as well as structural information about individual compounds found in crude extracts. Because the application of MS2 to crude

108

extracts can result in hundreds of fragmentation spectra, analyses can be difficult and time consuming (Frank et al., 2008). Recent applications such as MS-cluster and the online Global

Natural Products Social Molecular Networking tool (GNPS) (Wang et al., 2016) have allowed for better organization and visualization of MS2 data (Bandeira et al., 2007; Frank et al., 2008; Stein and Scott, 1994). MS-cluster cuts down on the number of spectra by grouping identical fragmentation spectra together and GNPS creates similarity indices between compounds with related fragmentation patterns. When visualized using programs such as Cytoscape (Smoot et al., 2011), mass spectra can be easily compared between extracts and analyzed for the identification of novel mass features. Using these tools, molecular profiles from various cultures can be compared so that differences can be quickly identified, and compounds targeted for further analysis.

Mass spectrometer protocols typically involve active exclusion parameters that limit the number of times specific ions are fragmented (MS2) and maximize the number of features that are subjected to fragmentation. Most ions are only fragmented 2-3 times regardless of their abundance providing no information as to their relative quantity. As a result, molecular networking, which relies on fragmentation spectra, does not provide information concerning the relative abundance of specific ions. Without relative abundance data, identifying ions that are upregulated due to co-culturing becomes rather difficult. However, MS1 events still contain information as to an ions intensity and as such can be used to determine the relative abundance of similar features between different samples. Since manually comparing MS1 features to look for upregulated ions between numerous data sets would be extraordinarily laborious, the program Metaboanlayst was generated. Metaboanalyst is an online pipeline that provides an easy to use, statistical platform to quickly pull out features that are significantly different between two or more samples (Chong et al., 2018).

109

Since Salinispora has numerous “silent” gene clusters, I hypothesized that co-culturing this genus alongside biotic stressors would result in the induction of secondary metabolites. In order to test this hypothesis, I developed a semi-high-throughput method to look for induction in response to a panel of challenge strains that possess properties that have been shown to induce secondary metabolite production. Using the GNPS pipeline for the generation of molecular networks, as well as Metaboanalyst, I identified the induced metabolome or

“inductome” that resulted when Salinispora was challenged with other bacterial species. Results indicate that co-culturing Salinispora with bacterial competitors resulted in the upregulation of compounds for every pair tested, and that this induction was unique to each coculturing pair, with S. arenicola cocultures exhibiting the highest number of induced features. Further experiments verified that desferrioxamine production was upregulated in S. tropica strain

CNB440 in response to stressors.

4.3 Methods

4.3.1 Bacterial selection and culturing

Twenty Salinispora strains with sequenced genomes were selected for induction studies based on specialized metabolite potential due to the presence of diverse biosynthetic gene clusters (Fig. 4.1). One and a half mL of each strain frozen in glycerol was added to 50 mL of A1 marine media (formula in chapter 2) in 125 mL culture flasks and shaken for 12 days prior to the addition of the challenge strain. 6 bacterial challenge strains were selected to co-culture with

Salinispora due to properties that had previously been shown to elicit induction (Table 4.1). One and a half mL of each strain that was frozen in glycerol was added to 50 mL A1 marine media in a 125 mL flask and shaken for 3 days prior to inoculation alongside the Salinispora strains.

110

Salinispora co-culturing

Ten μL aliquots of each Salinispora strain were transferred to 24 well plates (Evergreen scientific), with each well containing 1.5 mL A1 marine media with agar (0.8%). In total for each

Salinispora strain, 21 wells (3 replicates to be cultured with each of the 6 challenger strains and

3 replicates as a monoculture control) were inoculated (Fig 4.2). Several of these strains were also inoculated in 12 more wells and extracted over a time course as a control. Salinispora were left on the plate for four days, after which 10 μL of each bacterial challenger strain was inoculated adjacent the Salinispora colony in three wells. Ten μL of each bacterial challenger was also inoculated in wells in triplicate, and they were left as a monoculture control. The co- cultures were given four days to interact after which they were extracted by adding 1.5 mL of

1:1 DCM:MeOH to the wells. The solvent was left in the wells for 20 minutes after which it was transferred to 2 mL amber scintillation vials and left in a chemical hood for 3 days to dry. Once dry, the extracts were dissolved in 1 mL of MeOH and centrifuged at 10,000 x g to remove particulate. 200 µL of each supernatant was transferred to a 1.5 mL scintillation vial and analyzed by LC-MS/MS.

111

CNY244 CNY685 CNY694 CNT849 CNT859 CNS991 CNY666 CNY202 CNT403

CNH996 CNX891 CNH941 CNH905 CNX482 CNH964 CNB440 CNB536 CNY012

CNS205 CNY281 CNQ884 CNH898

iCNT150

S. tropica S. tropica S. tropica S. tropica S.

S. pacifica S. pacifica S. pacifica S.

S. pacifica S.

S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S. arenicola S.

Rif Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese lym Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS1A Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese pks1C Prese Prese Prese Prese Prese Prese Prese Prese PKS2 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS3A Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS3B Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS4 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS5 Prese Prese Prese Prese Prese Prese Prese Prese Prese Sid1 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKSNRPS Prese Prese Prese Prese Prese Prese STpks1 Prese Prese sal Prese Prese Prese Prese Prese Prese Prese STPKS2 Prese Prese Prese Prese Prese Prese spo Prese Prese slm Prese Prese Prese Prese Prese cya Prese Prese Prese Prese Prese Sid3 Prese Prese Prese Prese Prese fas1 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS7 Prese Prese Prese Prese Prese NRPS39 Prese Prese Prese Prese Prese PKS9 Prese PKS12 Prese PKS15 Prese Prese Prese Prese Prese Prese NRPS4 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese PKS17 Prese Prese PKS18 Prese Prese PKS25 Prese PKS16 Prese Prese Prese Prese Prese Prese FAS3 Prese Prese PKS28 Prese FAS5 Prese Prese Prese Prese Prese PKS30 Prese PKS31 Prese PKS38 Prese PKS41 Prese PKS42 Prese Prese PKS43 PKS44 Prese PKS60 Prese PKS61 Prese PKS63 Prese PKS67 Prese PKS68 Prese PKS69 Prese Cym Prese NRPS 1 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese NRPS2 Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese NRPS3 Prese Prese Prese Prese Prese PKS1B Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese Prese NRPS14 Prese Prese Prese Prese Prese Prese NRPS15 Prese Prese NRPS16 Prese Prese Prese NRPS17 Prese NRPS19 Prese Prese Prese Prese NRPS20 Prese Prese Prese NRPS27 Prese Prese Sid4 Prese Prese Prese Prese Prese NRPS33 Prese NRPS35 Prese Prese Prese NRPS36 Prese Prese Prese Prese Prese NRPS37 Prese NRPS38 Prese NRPS41 Prese NRPS45 Prese NRPS46 Prese NRPS 61 Prese NRPS62 Prese NRPS63 Prese NRPS67 Prese NRPS68 Prese

Figure 4.1 Map of secondary metabolite pathways present in the 20 Salinispora strains used in the study. Black box indicates presence of a BGC.

112

Table 4.1 Bacterial challenger strains and the rationale for why they were chosen.

Elicitor species Rationale Reference Vibrio sp. Strain CUA 759E Previously shown to induce specialized metabolites in Patin et al. , 2018 S. tropica Gordonia sp. strain CNJ 752 Mycolic acid induce Onaka et al. , 2011 Bacillus Sp strain CNJ 758 gram negative species Streptomyces sp. strain CNQ 525 produces numerous specialized metabolites Traxler et al. , 2013 Micromonospora sp. Strain CNZ 286 produces numerous specialized metabolites / most Mincer et al. , 2002 closely related genus to Salinispora Streptomyces sp. strain CNZ 306 produces numerous specialized metabolites Traxler et al. , 2013

Figure 4.2 Workflow for detecting specialized metabolite induction in Salinispora co-cultures.

4.3.2 LC MS/MS data acquisition

LC-MS/MS settings were based on those that have proven effective for these types of analyses (Duncan et al., 2015). Samples were analyzed with a Bruker Iontrap amaZon SL coupled to an Agilent 1260 LC system (Santa Clara, CA. USA]. A Phenomenex Kintex (Torrance, CA) C18

113

reversed-phase HPLC column (2.6 mm, 100 x 4.6 mm) was attached to the Agilent 1260 system and the samples were run under the following LC conditions: 1–2 min: 10% MeCN (0.1% FA) in

H2O (0.1% FA), 2–14 min: 10–100% MeCN (0.1 % FA), 14-15 min: 100 % MeCN (0.1 % FA) for 0.75 mL/min. The divert valve was set to waste for the first 3 min. Iontrap MS settings during the LC gradient were as follows: Acquisition: positive ion mode, mass range m/z 200–1600, MS scan rate 3/s, MS/MS scan rate 5/s, fixed collision energy 20eV; Source: gas temperature 300°C, gas flow 11 L/min; Nebulizer 45 psig; Scan source parameters: VCap 3000, Fragmentor 100,

Skimmer1 65, OctopoleRFPeak 750. MS data was analyzed with Bruker data analysis software

(Bruker).

4.3.4 Molecular networking

The MS/MS data from all replicates was converted to mzXML from Bruker files (.d) using the Trans-Proteomic pipeline (Institute for Systems Biology, Seattle, USA) (Deutsch et al.

2010). Files were then uploaded to the MASSive server (UCSD) via the FTP client Filezilla

(https://filezilla-project.org). Uploaded files were accessed and networked using the Global

Natural Products Social Molecular Networking (GNPS) pipeline, which helps in the dereplication of compounds by comparing their fragmentation spectrum against a spectral library. GNPS also creates a molecular network which clusters features together based on their chemical similarity as determined by their fragmentation patterns (Wang et al., 2016). GNPS utilizes MS-Cluster, an algorithm that compares MS2 spectra with similar parent masses using an ion tolerance of 0.5

Da and assigns a cosine similarity score to each pair. Pairs with a cosine score higher than 0.95 were combined into consensus spectra and treated as a single feature during the networking process (Frank et al., 2007). After MS-Cluster grouped consensus spectra, the spectral networks algorithm compared consensus MS/MS spectra, to network features with similar fragmentation spectra. The algorithm parameters included a mass tolerance for fragment peaks (0.5 Da),

114

parent mass tolerance 0.5 Da, the minimum number of matched peaks per spectral alignment

(4), a minimum cluster size of 2, and a minimum cosine score of 0.6. When comparing two spectra, a higher cosine score represents higher similarity of MS/MS spectra and therefore the corresponding molecules. The data were then visualized in the program Cytoscape, which created a network of nodes and edges (Smoot et al., 2011). The built in solid layout was applied to the visualized networks, and nodes containing parent ions from solvent blanks and controls were subtracted from the network. Nodes that only contained induced parent ions were represented as pie graphs with the colors indicating what sample the MS/MS spectra came from.

A standalone dereplication tool provided by GNPS was also used to identify known features in the network. Parameters were the same as those described above. Potential hits were manually verified by looking at the mirror spectra plots for each match.

4.3.5 Determination of upregulated mass features

Metaboanalyst was used to identify features that were upregulated in the co-cultures.

The program requires .csv feature tables that contain mass features present in each sample and their relative intensities provided there are at least 3 replicates per treatment. To generate these files and filter out noise from the MS1 chromatograms, the program MZmine v. 2.31 was used (Pluskal et al., 2010). The following MZmine parameters were applied in sequential order to generate the feature table: mass detection, chromatogram builder, chromatogram deconvolution, order peak lists, join aligner, peak finder, and peaks list rows filter. The parameters for mass detection were MS level 1 from 0-12 minutes with the mass detector set to centroid and the noise threshold at 2x105. For the chromatogram builder the parameters were MS level 1 from 0-12 minutes with a minimum time span of 0.02 minutes, a minimum height of 5.0x105 and an m/z tolerance of 0.5. The chromatogram deconvolution parameter was configured using the noise amplitude algorithm with a minimum peak height of 5.0x105 and an

115

amplitude of noise at 8.0x105. For the join aligner the parameters included an m/z tolerance of

0.1 Daltons with an m/z weight of 20, a retention time (RT) tolerance of 5.0% and a RT weight of 10. The peak finder was configured with an intensity tolerance of 40%, an m/z tolerance of

0.2 Daltons, and a retention time tolerance of 3.0%. Finally, the peak list rows filter was applied with default settings. The files were then assigned groups based on the origin of the samples

(I.e. monoculture, co-culture, etc.) and exported as a Metaboanalyst file.

The feature tables were uploaded to Metaboanlayst and run on the statistical analysis pipeline (Chong et al., 2018). They were subjected to log transformation and Pareto scaling to generate a relatively normal distribution of mass feature intensities. Once the files were transformed and filtered, a heat map was generated in Metaboanalyst showing the 100 features with the most significant difference in abundance between samples as determined by ANOVA and Fisher’s Least significant difference test. Results were downloaded as a table and features that showed signs of induction were manually verified in the Bruker Data Analysis platform to confirm upregulation and assess novelty.

4.3.6 Siderophore induction assays

To determine if the challenge strains or the Salinispora strains were producing compounds that were induced in the co-culture assays, Salinispora strains CNB440, CNT150,

CNZ306, and CNQ525 were each grown for 2 weeks in 50 mL A1 marine media after which 200

µL was transferred to 12 (100 mm x 16 mm) petri dishes (Fisher scientific) containing 20 mL of

A1 media with agar and left for 4 days. The Streptomyces strains were tested against CNB440 and CNT150, and the Salinispora were tested against CNZ306 and CNQ525. To determine if the induction was in response to cellular contact or secreted metabolites, cell free supernatant, autoclaved culture, and crude organic extracts were generated from each strain and Salinispora

116

treatments were added to Streptomyces cultures and Streptomyces treatments were added to

Salinispora cultures. Cell free supernatant was generated by centrifuging 15 mL of 14-day old culture in a falcon tube (Fisher scientific) on a Bechman culture centrifuge at 5,000 g for 5 minutes. The supernatant was then filtered through a 0.2 micron filter (Whattman). 100 µL of supernatant was added to each monoculture. The remaining cell pellet was then washed three times in 10 mL of 0.2 micron filter sterilized sea water, resuspended in 15 mL of seawater and autoclaved. 100 µL of autoclaved, washed cells were then added to the monocultures to determine if live cells were needed for induction. To generate crude chemical extracts, 10 mL of each culture was extracted with 20 mL of EtOAc. The extract was then dried and dissolved in 5 mL of MeOH. 100 µL of extract dissolved in MeOH was spotted adjacent to the colonies on the plates. 100 µL of MeOH was also spotted as a solvent control. After treating the cultures with autoclaved cells, supernatant, MeOH, or chemical extract, the plates were left for four days, after which they were transferred to 125 mL flasks and extracted with 30 mL of 1:1 DCM:MeOH for 20 minutes. The DCM:MeOH extract was filtered through a P8 coarse fluted filter paper

(Fisher Scientific, Hampton, NH) into 8 mL glass scintillation vials and dried under nitrogen, and the remaining water was frozen and lyophilized. Extracts were redissolved in 10 mL of MeOH, filtered through a 0.2 micron filter, and subjected to LC-MS/MS analysis as described previously.

Results were uploaded to Metaboanlayst and analyzed using the previously described parameters. Files were also uploaded to GNPS for dereplication and all matches manually compared using Metaboanalyst.

117

4.4 Results

4.4.1 Salinispora co-cultures

20 Salinispora strains were co-cultured against six different bacterial challenge strains to assess the Salinispora inductome. The Salinispora strains were selected due to their BGC diversity (Fig. 4.1) and consist of 13 S. arenicola strains, 3 S. tropica strains, and 4 S. pacifica strains. Of the 119 Salinispora strains with sequenced genomes, S. tropica appears to be the most clonal when it comes to BGC composition, so the fewest strains were chosen from this species. The challenge strains were chosen because they possess properties previously shown to induce specialized metabolism in bacteria. This includes a Vibrio sp. that induced compound production in S. tropica, a mycolic acid containing strain, specialized metabolite producers, and fast-growing Gram-negative species (Table 2.1).

4.4.2 induced features determined by molecular networking

Samples were analyzed by mass spectrometry and fragmentation spectra were uploaded to GNPS for analysis. Overall 8,638 unique parent ions were fragmented after the removal of blanks. Of those, 833 parent ions (9.64%) appeared to be induced in at least two of three replicates. These features were visualized as a molecular network, where parent ions were represented as nodes with the color corresponding to the Salinispora strain used in the co- culture that produced the parent ion (Fig. 4.3). Nodes that contained features that appeared in multiple co-cultures were represented as pie graphs, with the size of the slices equal to the relative number of times that ion was fragmented per co-culture. This can be used to get an idea of the relative abundance of a parent ion in one co-culture against the other co-cultures. Several compounds were dereplicated when compared to the GNPS library, including several amphiphilic desferrioxamines (DFO), which contained acyl side chains of various carbon lengths.

118

These desferrioxamines showed up in several strains and species but appeared to be most prominent in co-cultures containing S. tropica strain CNB440 (Fig. 4.3).

Figure 4.3 Molecular network consisting of features that were only found induced in at least two of three replicates. Parent ions are represented as nodes with edges connecting ions with similar fragmentation spectra. Nodes are represented as pie graphs with colors corresponding to the Salinispora strain used in the co-culture that produced the parent ion. Two molecular families are highlighted that matched library fragmentation spectra for desferrioxamines (DFO) containing acyl side chains with carbon lengths denoted. The molecular features were then compared to determine how induction varied between co-cultures using different Salinispora strains and species. There was huge variability in the number of induced features per strain, with some Salinispora strains showing induction of around 30 features, whereas others had upwards of 200 features (Fig. 4.4A). As for the challengers, there was less disparity between number of induced features when co-cultured

119

with the Salinispora strains. Co-cultures with the Vibrio sp. elicited the fewest number of features at 292, and co-cultures with the Gordonia sp. induced the most features at 388 (Fig.

4.4B). The mean number of induced features was 341 for the challenge strains with a standard deviation of 35.5 features.

The average number of features per Salinispora spp. was also not very different. S. arenicola produced the highest number of induced features per strain at 99, S. tropica second at 84 and S. pacifica with 76 (Fig. 4.4C). However, the induction per strain was so variable that these trends were not significant. For instance, the range of induced features between co- cultures with S. arenicola strains varied from 21 to 196 (Fig. 4.4A). Looking at the diversity of features across the samples revealed that co-cultures with S. arenicola contained the highest number of unique features at 426, although this is likely due to the larger sampling size compared to the other two species (Fig. 4.4D). The number of features shared among the strains was 43, indicating that some induced features may be common to the genus. There were also over 200 features unique to two of the three Salinispora spp. (Fig. 4.4D). However, it cannot be determined if these features were produced by the challenge strains in response to Salinispora, as this study was not able to differentiate bacterial producers in the co-culture.

120

Figure 4.4 Number of induced features from cocultures containing the different Salinispora species determined by molecular networking. A) Pie graph representing the number of induced features from the cocultures of each Salinispora strain. B) Pie graph representing the number of induced features per challenge strain cocultures. C) Bar graph depicting the average number of induced features in cocultures containing one of the three Salinispora spp. error bars are standard deviation. D) Venn diagram showing the relationship between induced features for the Salinispora cocultures

4.4.3 induced features determined by Metaboanalyst

Since over 800 induced features were detected in the co-cultures using GNPS, attempts were made to determine which induced features were most abundant. The GNPS pipeline does not have the ability to differentiate upregulated features in co-cultures compared to monocultures. Instead, feature tables were uploaded to Metaboanalyst, a program that conducts statistical analyses to distinguish features that are significantly different between

121

samples. Once the data is transformed and normalized, Metaboanalyst conducts an ANOVA on features as well as post hoc tests which can be visualized as a heatmap or series of boxplots. The features that were significantly upregulated in the co-cultures were compiled into a table consisting of 58 features across the 20 Salinispora cultures (Table 2.2). There were far fewer features that appeared to be significantly upregulated after the Metaboanalyst analysis (58) in comparison to the GNPS molecular networking analysis (866). This is likely due the fact that rather conservative filtering parameters were used in the generation of the Metaboanalyst file and because only the top 100 upregulated features were analyzed between the samples, which included numerous features in the monocultures. These settings were used to prioritize the most abundant upregulated features to make future isolation easier.

Interestingly, there were few upregulated features shared between co-cultures with different Salinispora strains, with only one feature being shared between four strains, and two features between three strains. The same was seen for the challenge strains, where most features were induced in co-cultures with a single challenger. The most induced features were seen in the co-culture containing Streptomyces strain CNQ525, however features were seen from co-cultures with all challengers. These results support the notion that induction is highly variable among strains.

122

Table 4.2 Features that were significantly upregulated in the co-cultures. Shown are the features (m/z), as well as the strains used in the co-culture(s) that produced the features

123

4.4.2 Siderophore upregulation in S. tropica strain CNB440

Several of the most significantly upregulated features from the Metaboanalyst analysis were the same features that GNPS dereplicated as desferrioxamines (Fig 4.3). These masses appeared to be upregulated to the greatest extent when CNB440 was co-cultured with either of the two Streptomyces challenge strains (CNQ525 and CNZ306) (Fig. 4.5). Efforts were taken to identify which of the strains in the co-culture were producing these compounds. Previous work from the Kolter lab showed that when certain strains of Streptomyces are grown next to other actinomycetes they start producing a suite of desferrioxamines with acyl side chains varying in length from 7-20 carbons, (Traxler et al., 2013), similar to those identified in this study. To determine which strains were responsible for the production of the siderophores, cell free supernatant, chemical extract, and autoclaved cells from the Streptomyces strains were introduced to the Salinispora cultures, and the same Salinispora treatments were added to the Streptomyces strains. Surprisingly, upregulation was only seen when Streptomyces treatments were added to Salinispora strain CNB440, which resulted in the upregulation of

C11-15 acyl-desferrioxamine, C10 amphiphilic ferrioxamine, and desferrioxamine (FIG 4.6, 4.7).

The addition of MeOH as well as cell free supernatant appeared to induce the greatest upregulation in siderophore production. Since MeOH can permeate the cell membrane and damage the cell, this may indicate that siderophore production is a general stress response in

Salinispora. However, the increase in each siderophore analogue was not uniform indicating that Salinispora may be capable of upregulating production of specific siderophores depending on the environmental stressor. Out of the four treatments, autoclaved cell pellets brought about the lowest production of siderophores, suggesting that siderophore upregulation may be in response to secreted chemical signals found in the cell free supernatant or chemical extract.

124

Figure 4.5 Clustered heatmap depicting the average of each of the top 50 upregulated features in the CNB440 co-cultures and monocultures based on an ANOVA analysis. Color indicates the fold change in intensity for each feature compared across the samples after log transformation and Pareto scaling, with red indicating a higher intensity and blue a lower intensity. Blue boxes indicate upregulated features that match desferrioxamines.

125

1.5

0.5

-0.5

-1

-1.5

Normalized fold change fold Normalized C11 C12 C13 C14 C15 C10 -2 DFO A-DFO A-DFO A-DFO A-DFO A-DFO Amphiphillic ferrioxamine

Control Cell MeOH Supernatant Extract

Figure 4.6 Upregulation of desferrioxamines in S. tropica strain CNB440 in response to Streptomyces elicitors. The bars represent the fold change between the relative intensities of the features after they were log transformed and subjected to Pareto scaling. Only features that contained statistically significant upregulation between the monoculture control and at least one treatment were selected. Blue bars represent monocultures, orange bars autoclaved cells, grey bars MeOH solvent controls, yellow bars are cell free supernatant, and light blue bars are crude chemical extracts. Error bars are standard deviation.

126

Figure 4.7 Mirror spectra plots comparing the fragmentation spectra from the GNPS library to the spectra from the CNB440 + supernatant extract for the amphiphilic DFOs. The green spectra are from the library, whereas the black spectra are from the sample.

127

4.5 Discussion

Secondary metabolite induction has been well documented in the microbial realm, but more exhaustive studies of the “inductome” remain warranted (Abdelmohsen et al., 2015;

Okada and Seyedsayamdost, 2017). Genomic insight has led to the understanding that many bacterial species have enormous untapped biosynthetic potential, yet the underlying ecological signals to bring about their production are poorly understood (Abdelmohsen et al., 2015; Letzel et al., 2017). Thanks to advances in analytical chemistry, metabolomes can easily be analyzed and annotated allowing for high-throughput screening (Seyedsayamdost, 2014). The marine actinomycete genus has enormous biosynthetic potential, with over 170 annotated BGCs, yet the number of compounds isolated from this genus pales in comparison (Letzel et al., 2017).

What has been learned to date about natural product production in the genus Salinispora is mostly based on metabolites that have been isolated from monocultures grown in the laboratory (Jensen et al., 2015). These results provide little understanding of how competition from the natural environment impact secondary metabolism. Recent efforts have shown that

Salinispora mediate specialized metabolite production in response to environmental isolates; however, this study took a relatively narrow approach studying the interaction between a

Salinispora tropica strain and various bacterial isolates (Patin et al., 2018). This study is the first to more comprehensively study the effects of bacterial challengers on the metabolome of the three described Salinispora species. Together these results indicate that the inductome of co- cultures using Salinispora is relatively large and variable among strains. The fact that so few induced metabolites matched spectral libraries indicates that there is likely abundant chemistry to be elucidated from Salinispora cocultures despite decades of concerted efforts studying the genus’s metabolome when grown axenically. One difficulty in isolating and characterizing specialized metabolites from bacterial cultures is acquiring enough material. Although

128

molecular networking helps visualize the inductome, it provides little information as to a features abundance. This can hinder efforts to isolate compounds since it is hard to determine which features would provide enough material for isolation. The integration of Metaboanalyst into the analysis pipeline allows for the determination of significantly upregulated features rather quickly, which can be helpful in prioritizing masses for further isolation.

Unfortunately, decoupling the metabolomes from the two co-cultured bacteria remains a difficult task and as such individual features cannot be assigned to a specific bacterial producer but instead only a co-culture. Production of compounds that appeared in co-cultures from numerous Salinispora strains but only a single challenger may suggest it’s produced by the challenger but without conducting follow up studies definitive assignments cannot be made.

However, co-cultures can easily be scaled so that significantly upregulated features can be isolated for drug discovery purposes. Also, chemical extracts or cell free supernatant from challenge strains can be added to Salinispora that exhibit the most interesting / highest levels of induction in the co-cultures to determine if the Salinispora is the actual producer of the induced compounds.

Since the bacterial environment is teaming with biotic and abiotic stressors, it is hard to determine which signals induce specialized metabolism. This study only tests the effects of individual challenge strains on Salinispora induction. However, in the environment, Salinispora is exposed to chemical signals from the entire microbial community, and this synergism has untold effects on Salinispora specialized metabolism. Even adding a third bacterial species to a co-culture has effects on antibiotic production, as seen in S. coelicolor (Abrudan et al., 2015). To fully understand the inductome of Salinispora many co-culture experiments will need to be conducted using a myriad of elicitors in consortium.

129

The creation of relatively easy and high throughput methods to induce secondary metabolism will likely increase the discovery of bioactive compounds. Studies that use defined chemical elicitors to induce specialized metabolism have already been met with success

(Seyedsayamdost, 2014). However, this is one of the first times that the model genus Salinispora has been used in a high throughput induction screen.

Using model organisms such as Salinispora provide numerous benefits since they have been studied in the lab for decades. For instance it has been shown that different culturing parameters influence the secreted metabolome of Salinispora, and that culturing members of this genus on agar substrates results in the production of a more diverse metabolome

(Crüsemann et al., 2017). Also, the genomes of over 100 strains have already been sequenced, which could allow the marriage of both the inductome and genome datasets to identify new compounds as well as their BGCs. Future endeavors could include bioassay guided fractionation to prioritize compounds with potential pharmaceutical promise, as well as transcriptomics to identify upregulated BGCs in co-culture conditions. Taken together, this could be a promising way to identify novel chemistry and quickly link it to its biosynthetic gene clusters.

The hydrophilicity of siderophores can be modified by the addition of acyl moieties at various carbon lengths. When the acyl side chain is relatively short with 10 or less carbons, the siderophore diffuses in the aqueous environment, however siderophores containing an acyl moiety between 12-16 carbons can form vesicles and micelles in the environment (Butler, 2005;

Martinez, 2000). It has been documented that in areas where iron is not limiting, bacterial species produce siderophores that readily diffuse in the aqueous environment, however when iron is extremely limiting such as in near shore sediments, bacteria produce amphiphilic siderophores, which is thought to keep these iron scavenging compounds closer to the producing cell (Boiteau et al., 2016). Salinispora occurs in tropical sediments where iron is

130

limited so production of these amphiphilic desferrioxamine species fits with these hypotheses.

Siderophore production has been shown to be rampant in both terrestrial and marine bacteria, with some studies showing that unique siderophore production can be induced in response to competition (Traxler et al., 2013). Although desferrioxamine production is well documented in Salinispora, especially strain CNB440, the amphiphilic siderophores identified in this study have not been documented from CNB440 in the past (Ejje et al., 2013; Roberts et al.,

2012). These results mirror those seen in Streptomyces, where the biosynthesis of desferrioxamine with acyl moieties appears to be rather flexible (Traxler et al., 2013). The production of these siderophores by certain Salinispora strains in response to bacterial competitors indicates that nutrient sequestration is a competition strategy exhibited by some

Salinispora strains. The fact that only certain Salinispora strains appear to exert this competitive strategy indicates competition can be strain specific within Salinispora. Interestingly, S. tropica strain CNB440 also upregulated numerous desferrioxamines in response to solvent stress, indicating that siderophore regulation may also be a sort of SOS response to cellular harm.

This high throughput coculturing experiment indicates the potential for using novel culturing techniques to gain a better understanding of how the environment influences the metabolome from organisms that have been studied for decades. Coculturing experiments such as this one shows that we may be missing enormous amounts of chemical potential buried in our freezers. By considering ecologically relevant stressors, a better understanding of the chemical potential of bacteria can be garnered.

4.6 Acknowledgments

131

Chapter 4 could not have been completed without the generous help of Fisher Price, an undergraduate at UCSD. Also thanks to Leesa Habener for introducing me to the Metaboanlayst platform.

4.7 References

Abdelmohsen UR, Grkovic T, Balasubramanian S, Kamel MS, Quinn RJ, Hentschel U. 2015. Elicitation of secondary metabolism in actinomycetes. Biotechnol Adv 33:798–811. doi:10.1016/j.biotechadv.2015.06.003 Abrudan MI, Smakman F, Grimbergen AJ, Westhoff S, Miller EL, van Wezel GP, Rozen DE. 2015. Socially mediated induction and suppression of antibiosis during bacterial coexistence. Proc Natl Acad Sci 112:11054–11059. doi:10.1073/pnas.1504076112 Alongi DM. 1988. Bacterial productivity and microbial biomass in tropical mangrove sediments. Microb Ecol 15:59–79. doi:10.1007/BF02012952 Amos GCA, Awakawa T, Tuttle RN, Letzel A-C, Kim MC, Kudo Y, Fenical W, S. Moore B, Jensen PR. 2017. Comparative transcriptomics as a guide to natural product discovery and biosynthetic gene cluster functionality. Proc Natl Acad Sci 114:E11121–E11130. doi:10.1073/pnas.1714381115 Angell S, Bench BJ, Williams H, Watanabe CMH. 2006. Pyocyanin Isolated from a Marine Microbial Population: Synergistic Production between Two Distinct Bacterial Species and Mode of Action. Chem Biol 13:1349–1359. doi:10.1016/j.chembiol.2006.10.012 Bandeira N, Tsur D, Frank A, Pevzner PA. 2007. Protein identification by spectral networks analysis. Proc Natl Acad Sci 104:6140–6145. doi:10.1073/pnas.0701130104 Benson DA, Clark K, Karsch-Mizrachi I, Lipman DJ, Ostell J, Sayers EW. 2015. GenBank. Nucleic Acids Res 43:D30–D35. doi:10.1093/nar/gku1216 Bentley S, Chater K, Cerdeño-Tárraga A-M, Challis GL, Thomson NR, James KD, Harris DE, Quail M a, Kieser H, Harper D, Bateman A, Brown S, Chandra G, Chen CW, Collins M, Cronin A, Fraser A, Goble A, Hidalgo J, Hornsby T, Howarth S, Huang C-H, Kieser T, Larke L, Murphy L, Oliver K, O’Neil S, Rabbinowitsch E, Rajandream M, Rutherford K, Rutter S, Seeger K, Saunders D, Sharp S, Squares R, Squares S, Taylor K, Warren T, Wietzorrek A, Woodward J, Barrell BG, Parkhill J, Hopwood D a. 2002. Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2). Nature 417:141–147. doi:10.1038/417141a Bertrand S, Schumpp O, Bohni N, Bujard A, Azzollini A, Monod M, Gindro K, Wolfender JL. 2013. Detection of metabolite induction in fungal co-cultures on solid media by high-throughput differential ultra-high pressure liquid chromatography-time-of-flight mass spectrometry fingerprinting. J Chromatogr A 1292:219–228. doi:10.1016/j.chroma.2013.01.098 Boiteau RM, Mende DR, Hawco NJ, McIlvin MR, Fitzsimmons JN, Saito MA, Sedwick PN, DeLong EF, Repeta DJ. 2016. Siderophore-based microbial adaptations to iron scarcity across the eastern Pacific Ocean. Proc Natl Acad Sci. doi:10.1073/pnas.1608594113 Butler A. 2005. Marine Siderophores and Microbial Iron Mobilization. BioMetals 18:369–374.

132

doi:10.1007/s10534-005-3711-0 Callahan HS, Maughan H, Steiner UK. 2008. Phenotypic Plasticity, Costs of Phenotypes, and Costs of Plasticity. Ann N Y Acad Sci 1133:44–66. doi:10.1196/annals.1438.008 Chong J, Soufan O, Li C, Caraus I, Li S, Bourque G, Wishart DS, Xia J. 2018. MetaboAnalyst 4.0: towards more transparent and integrative metabolomics analysis. Nucleic Acids Res gky310. doi:10.1093/nar/gky310 Crüsemann M, O’Neill EC, Larson CB, Melnik A V., Floros DJ, Da Silva RR, Jensen PR, Dorrestein PC, Moore BS. 2017. Prioritizing Natural Product Diversity in a Collection of 146 Bacterial Strains Based on Growth and Extraction Protocols. J Nat Prod. doi:10.1021/acs.jnatprod.6b00722 Duncan KR, Crüsemann M, Lechner A, Sarkar A, Li J, Ziemert N, Wang M, Bandeira N, Moore BS, Dorrestein PC, Jensen PR. 2015. Molecular Networking and Pattern-Based Genome Mining Improves Discovery of Biosynthetic Gene Clusters and their Products from Salinispora Species. Chem Biol 22:460–471. doi:10.1016/j.chembiol.2015.03.010 Ejje N, Soe CZ, Gu J, Codd R. 2013. The variable hydroxamic acid siderophore metabolome of the marine actinomycete Salinispora tropica CNB-440 Metallomics. doi:10.1039/c3mt00230f Frank AM, Bandeira N, Shen Z, Tanner S, Briggs SP, Smith RD, Pevzner P a. 2008. Clustering Millions of Tandem Mass Spectra research articles. J proteome Res 113–122. doi:10.1021/pr070361e Frank AM, Savitski MM, Nielsen ML, Zubarev R a., Pevzner P a. 2007. De novo peptide sequencing and identification with precision mass spectrometry. J Proteome Res 6:114–23. doi:10.1021/pr060271u Garbeva P, Silby MW, Raaijmakers JM, Levy SB, Boer W De. 2011. Transcriptional and antagonistic responses of Pseudomonas fluorescens Pf0-1 to phylogenetically different bacterial competitors. ISME J 5:973–985. doi:10.1038/ismej.2010.196 Janso JE, Carter GT. 2010. Biosynthetic Potential of Phylogenetically Unique Endophytic Actinomycetes from Tropical Plants. Appl Environ Microbiol 76:4377–4386. doi:10.1128/AEM.02959-09 Jensen PR, Moore BS, Fenical W. 2015. The marine actinomycete genus Salinispora: a model organism for secondary metabolite discovery. Nat Prod Rep 00:1–14. doi:10.1039/c4np00167b Jousset A, Rochat L, Scheu S, Bonkowski M, Keel C. 2010. Predator-prey chemical warfare determines the expression of biocontrol genes by rhizosphere-associated pseudomonas fluorescens. Appl Environ Microbiol 76:5263–5268. doi:10.1128/AEM.02941-09 Jousset A, Scheu S, Bonkowski M. 2008. Secondary metabolite production facilitates establishment of rhizobacteria by reducing both protozoan predation and the competitive effects of indigenous bacteria. Funct Ecol 22:714–719. doi:10.1111/j.1365- 2435.2008.01411.x Lawrence JG, Hendrix RW, Casjens S. 2001. Where are the pseudogenes in bacterial genomes? Trends Microbiol 9:535–540. doi:10.1016/S0966-842X(01)02198-9

133

Letzel A-C, Li J, Amos GCA, Millán-Aguiñaga N, Ginigini J, Abdelmohsen UR, Gaudêncio SP, Ziemert N, Moore BS, Jensen PR. 2017. Genomic insights into specialized metabolism in the marine actinomycete Salinispora. Environ Microbiol 19:3660–3673. doi:10.1111/1462- 2920.13867 Martinez JS. 2000. Self-Assembling Amphiphilic Siderophores from Marine Bacteria. Science (80- ) 287:1245–1247. doi:10.1126/science.287.5456.1245 Mazzola M, De Bruijn I, Cohen MF, Raaijmakers JM. 2009. Protozoan-induced regulation of cyclic lipopeptide biosynthesis is an effective predation defense mechanism for Pseudomonas fluorescens. Appl Environ Microbiol 75:6804–6811. doi:10.1128/AEM.01272-09 Mearns-Spragg a, Bregu M, Boyd KG, Burgess JG. 1998. Cross-species induction and enhancement of antimicrobial activity produced by epibiotic bacteria from marine algae and invertebrates, after exposure to terrestrial bacteria. Lett Appl Microbiol 27:142–146. Mincer TJ, Jensen PR, Kauffman C a, Fenical W. 2002. Widespread and persistent populations of a major new marine actinomycete taxon in ocean sediments. Appl Environ Microbiol 68:5005–11. doi:10.1128/AEM.68.10.5005 Okada BK, Seyedsayamdost MR. 2017. Antibiotic dialogues: induction of silent biosynthetic gene clusters by exogenous small molecules. FEMS Microbiol Rev 41:19–33. doi:10.1093/femsre/fuw035 Omura S, Ikeda H, Ishikawa J, Hanamoto A, Takahashi C, Shinose M, Takahashi Y, Horikawa H, Nakazawa H, Osonoe T, Kikuchi H, Shiba T, Sakaki Y, Hattori M. 2001. Genome sequence of an industrial microorganism Streptomyces avermitilis: Deducing the ability of producing secondary metabolites. Proc Natl Acad Sci 98:12215–12220. doi:10.1073/pnas.211433198 Onaka H, Mori Y, Igarashi Y, Furumai T. 2011. Mycolic Acid-Containing Bacteria Induce Natural- Product Biosynthesis in Streptomyces Species. Appl Environ Microbiol 77:400–406. doi:10.1128/AEM.01337-10 Patin N V, Duncan KR, Dorrestein PC, Jensen PR. 2015. Competitive strategies differentiate closely related species of marine actinobacteria. ISME J 10:1–13. doi:10.1038/ismej.2015.128 Patin N V, Floros DJ, Hughes CC, Dorrestein PC, Jensen PR. 2018. The role of inter-species interactions in Salinispora specialized metabolism. Microbiology. Pluskal T, Castillo S, Villar-Briones A, Orešič M. 2010. MZmine 2: Modular framework for processing, visualizing, and analyzing mass spectrometry-based molecular profile data. BMC Bioinformatics 11:395. doi:10.1186/1471-2105-11-395 Roberts AA, Schultz AW, Kersten RD, Dorrestein PC, Moore BS. 2012. Iron acquisition in the marine actinomycete genus Salinispora is controlled by the desferrioxamine family of siderophores. FEMS Microbiol Lett 335:95–103. doi:10.1111/j.1574-6968.2012.02641.x Seyedsayamdost MR. 2014. High-throughput platform for the discovery of elicitors of silent bacterial gene clusters. Proc Natl Acad Sci 111:7266–7271. doi:10.1073/pnas.1400019111 Slattery M, Rajbhandari I, Wesson K. 2001. Competition-Mediated Antibiotic Induction in the Marine Bacterium Streptomyces tenjimariensis. Microb Ecol 41:90–96.

134

doi:10.1007/s002480000084 Smoot ME, Ono K, Ruscheinski J, Wang PL, Ideker T. 2011. Cytoscape 2.8: New features for data integration and network visualization. Bioinformatics 27:431–432. doi:10.1093/bioinformatics/btq675 Song C, Mazzola M, Cheng X, Oetjen J, Alexandrov T, Dorrestein P, Watrous J, van der Voort M, Raaijmakers JM. 2015. Molecular and chemical dialogues in bacteria-protozoa interactions. Sci Rep 5:12837. doi:10.1038/srep12837 Stein SE, Scott DR. 1994. Optimization and testing of mass spectral library search algorithms for compound identification. J Am Soc Mass Spectrom 5:859–866. doi:10.1016/1044- 0305(94)87009-8 Traxler MF, Watrous JD, Alexandrov T, Dorrestein PC, Kolter R. 2013. Interspecies Interactions Stimulate Diversification of the Streptomyces coelicolor Secreted Metabolome. MBio 4:e00459-13. doi:10.1128/mBio.00459-13 Trischman JA, Oeffner RE, De Luna MG, Kazaoka M. 2004. Competitive induction and enhancement of indole and a diketopiperazine in marine bacteria. Mar Biotechnol 6:215– 220. doi:10.1007/s10126-003-0010-z Wang M, Carver JJ, Phelan V V, Sanchez LM, Garg N, Peng Y, Nguyen DD, Watrous J, Kapono CA, Luzzatto-Knaan T, Porto C, Bouslimani A, Melnik A V, Meehan MJ, Liu W-T, Crüsemann M, Boudreau PD, Esquenazi E, Sandoval-Calderón M, Kersten RD, Pace LA, Quinn RA, Duncan KR, Hsu C-C, Floros DJ, Gavilan RG, Kleigrewe K, Northen T, Dutton RJ, Parrot D, Carlson EE, Aigle B, Michelsen CF, Jelsbak L, Sohlenkamp C, Pevzner P, Edlund A, McLean J, Piel J, Murphy BT, Gerwick L, Liaw C-C, Yang Y-L, Humpf H-U, Maansson M, Keyzers RA, Sims AC, Johnson AR, Sidebottom AM, Sedio BE, Klitgaard A, Larson CB, Boya P CA, Torres-Mendoza D, Gonzalez DJ, Silva DB, Marques LM, Demarque DP, Pociute E, O’Neill EC, Briand E, Helfrich EJN, Granatosky EA, Glukhov E, Ryffel F, Houson H, Mohimani H, Kharbush JJ, Zeng Y, Vorholt JA, Kurita KL, Charusanti P, McPhail KL, Nielsen KF, Vuong L, Elfeki M, Traxler MF, Engene N, Koyama N, Vining OB, Baric R, Silva RR, Mascuch SJ, Tomasi S, Jenkins S, Macherla V, Hoffman T, Agarwal V, Williams PG, Dai J, Neupane R, Gurr J, Rodríguez AMC, Lamsa A, Zhang C, Dorrestein K, Duggan BM, Almaliti J, Allard P-M, Phapale P, Nothias L-F, Alexandrov T, Litaudon M, Wolfender J-L, Kyle JE, Metz TO, Peryea T, Nguyen D-T, VanLeer D, Shinn P, Jadhav A, Müller R, Waters KM, Shi W, Liu X, Zhang L, Knight R, Jensen PR, Palsson BØ, Pogliano K, Linington RG, Gutiérrez M, Lopes NP, Gerwick WH, Moore BS, Dorrestein PC, Bandeira N. 2016. Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking. Nat Biotechnol 34:828– 837. doi:10.1038/nbt.3597 Ziemert N, Lechner A, Wietz M, Millan-Aguinaga N, Chavarria KL, Jensen PR. 2014. Diversity and evolution of secondary metabolism in the marine actinomycete genus Salinispora. Proc Natl Acad Sci 111:E1130–E1139. doi:10.1073/pnas.1324161111

(Mincer et al., 2002)

135

Chapter 5

Concluding thoughts

136

5.1 Concluding thoughts

Since the advent of modern chemical ecology studies in the 1950’s, we have come to understand that chemically mediated interactions are important in structuring the environment

(Abelson 1994; Wietz et al. 2013; Hay 2009). Even though we are beginning to understand the types of ecological roles that these chemicals possess, there is a huge swath of the compounds that have been isolated from organisms that still have unknown ecological functions (Abelson

1994). This is troubling since they may possess important roles in the ecosystem and aid organisms in ways that we have never seen previously. For instance, extracts from the marine actinomycete Salinispora were shown to alter the bacterial community composition of bacterial mesocosms from marine sediments where this genus occurs (Patin et al. 2017). Overall, marine sediment bacteria have been shown to be prolific producers of specialized metabolites; however, few of these compounds have been linked to ecological roles, and even less is known about their production in the environment (Wietz et al. 2013).

One of the fundamental tenants of bacterial chemical ecology is to realize how bacterial metabolites impact the environment (Takken and Dicke 2006). However, due to the difficulty in conducting studies in situ, which can often require difficult analyses to determine the bacterial community and the chemical interactions they are facilitating, most studies are conducted in laboratory settings. Only recently have technological advancements progressed to the point where we can elucidate the chemical interplay of bacteria. The decreasing cost of genetic sequencing has allowed for the collection of immense amounts of genomic data, which has been used to understand the community structure as well as the metagenomic makeup of organisms across many habitats (Caporaso et al. 2010; Jansson and Baker 2016; Thompson et al. 2017).

Unfortunately, understanding the genetic potential of a habitat does not always correlate to the

137

chemical interplay realized in these habitats, since numerous BGCs may not be functional, expressed, or translated (Machado, Tuttle, and Jensen 2017). One of the ways to address this issue requires the collection of environmental metabolomic data, which for a long time was impeded by the lack of instrumentation sensitive enough to accurately detect compounds.

Thanks to advances in mass spectrometry, as well as the generation of large molecular databases, we can now deconvolute and visualize metabolomes with relative ease (Wang et al.

2016; Bouslimani et al. 2014). These technological advancements are paving the way for in situ bacterial chemical ecology studies (Jansson and Baker 2016).

The aim of my thesis was to gain a better understanding of the influence of bacterial specialized metabolites on sediments, focusing specifically on heterotrophic bacterial producers. Using Salinispora as a model organism, I took a holistic approach to studying the chemical ecology of microbial specialized metabolites; by trying to first understand what metabolites are present in the environment, why they are out there, and how bacterial competition effects their production. To accomplish these aims I wanted to look directly at the sediment metabolome to see if natural products could be identified in situ, providing insight into what can be detected in the environment as opposed to in culture. I then wanted to determine who was making these compounds and what sort of ecological roles these compounds possessed. Finally, I wanted to add ecologically relevant stressors in the form of co-culturing to see if this could induce compound production in Salinispora spp.

Shallow sediments are home to an untold number of microorganisms. The heterogeneity alone is impressive, with individual sand grains being comprised of unique microbiomes containing thousands of bacterial species (Probandt et al. 2018). It is of no surprise then that these organisms have developed a means to compete for limited resources through the use of specialized metabolites (Wietz et al. 2013). The bacterial order Actinomycetales, for

138

instance, comprise numerous genera some of which are prolific producers of allelochemicals

(Lam 2006; Doroghazi and Metcalf 2013; Jensen et al. 2005). However, our understanding of the chemical warfare going on within sediments is all from observations made in the laboratory where bacteria are typically grown axenically in nutrient rich conditions. This has called into questions whether or not these compounds are even produced in the environment, and if so, are they being produced at appreciable concentrations (Davies 2006). In chapter 2, using advanced mass spectrometry techniques combined with publicly available spectral databases, I was able to identify numerous compounds from the environment, including several known microbial specialized metabolites. By relating these compounds to a 16S rRNA community analysis, I was able to correlate the presence of the known cytotoxin staurosporine to

Salinispora arenicola. Efforts to then quantify the abundance of staurosporine in the environment led to the surprising observation that this cytotoxin was being produced at millimolar concentrations, close to the concentrations typically seen in lab cultures. These values were orders of magnitude higher than those required to inhibit brine shrimp and cancer cell lines, indicating that this compound may be impacting sediment community structure.

These analyses indicate that we are getting to the point where we can look at the environmental metabolome and detect the small molecules that mediate interactions and potentially shape community structure. This sort of analysis could have huge consequences on how microbial chemical ecology is studied since we can more accurately predict what sort of chemical interactions are at play in the environment, as opposed to assuming that compounds produced or extracted in laboratory settings possess ecological relevancy. Applying these methods to other marine environments will give us a better idea of what specialized metabolites are affecting marine communities. Unfortunately, elucidating the roles of bacterial compounds found in the environment is a difficult without culturing the producing microbe and conducting

139

follow up experiments since bacterial compounds are likely produced in low quantities.

Therefore, traditional laboratory based chemical ecology approaches should be used in tandem with environmental omics to determine the functions of the compounds found in the environment. The utility of this approach is that by first understanding what compounds are out there, we can then better prioritize what compounds should be explored for their ecological functions.

One of the early goals of microbial natural product research was the discovery of chemicals that could be harnessed for human applications. This led to the discovery of numerous compounds, notably those that exhibited antibiotic and anti-cancer properties. However, there has been a dearth of compounds discovered recently that exhibit novel modes of action, which is troubling when considering the current antibiotic crisis (Fernandes 2006). One of the exciting outcomes of the sediment metabolomics study was the number of potential compounds that are present in the sediments that were not identified using the GNPS spectral libraries. Not only does this point to how poorly we understand the chemical landscape of sediment communities, but that there is untold chemical potential that is unrealized in these habitats. Applying these adsorbent resin techniques on a larger scale may allow for the isolation of novel compounds that are not currently being found in traditional culturing studies. Essentially, we can use these methods to treat the sediment environment as our petri dish for drug discovery instead of relying on laboratory-based culturing techniques.

In chapter 2, I observed that staurosporine was produced at relatively high concentrations in the sediments, and that one potential producer was Salinispora arenicola.

Since this compound is a potent cytotoxin, I wanted to determine if Salinispora spp. specialized metabolites could deter eukaryotic predation. In general, heterotrophic marine microbes have been shown to produce cytotoxic compounds, but little known about what sort of roles these

140

compounds play in the environment (Wietz et al. 2013). The Salinispora genus contains two species that co-occur in the sediments and produce unique cytotoxic compounds (Jensen,

Moore, and Fenical 2015). To investigate why these compounds may be fixed within the

Salinispora species, I designed several assays to test the effects of Salinispora compounds on bacterivorous eukaryotes. Interestingly a C. elegans feeding deterrence assay indicated that the extractable specialized metabolome from S. tropica acted as a strong feeding deterrent, whereas the metabolome from S. arenicola exhibited more chemoattractant properties. Further investigation revealed that the primary deterrent compounds from S. tropica were the lomaiviticins, potent cytotoxins that are capable of being produced by all S. tropica strains.

Assays were also created using more ecologically relevant organisms, including a polychaete,

Ophryotrocha sp. and marine nematodes. Results from these assays mirrored those seen in the

C. elegans feeding deterrence assays, where S. tropica was more deterrent than S. arenicola.

These results indicate that closely related co-occurring species may remain distinct species due to niche differentiation brought about, at least partially, by their specialized metabolome. This study adds to the growing evidence that closely related bacterial species may remain different species due, in part, to their specialized metabolome (Patin et al. 2015).

One issue with laboratory based bacterial chemical ecology studies is that they typically avoid the introduction of stressors that bacterial spp. likely encounter in the environment.

Culturing is done axenically and therefore these studies are limited to compounds that are constitutively expressed in nutrient rich culturing conditions. Our reliance on culturing in order to generate enough material for analytical studies will likely not be overcome; however, new culturing paradigms that use more natural settings are now being incorporated, such as the iChip which allows for culturing previously “unculturable” bacteria by using the natural environment as a petri dish (Nichols et al. 2010). For bacterial species that are more readily grown in the

141

laboratory, simply adding chemical elicitors to previously established culturing conditions can induce compound production (Patin et al. 2018; Okada and Seyedsayamdost 2017;

Abdelmohsen et al. 2015). This indicates that the complex chemistry found in marine sediments likely influences specialized metabolite production from the bacterial community. Knowing that traditional culturing paradigms lack these natural chemical cues led me to develop a high throughput method to test the effects of bacterial challengers on Salinispora specialized metabolite production. Results mirrored those seen in previous co-culturing studies with actinomycetes; that induction is extremely strain specific for both the producing organism and the bacterial elicitor (Traxler et al. 2013). Efforts to determine some of the induced compounds led to the identification of the iron scavenging desferrioxamines being upregulated in

Salinispora due to Streptomyces signals. These results indicate that studying Salinispora chemistry from axenic cultures may not provide an accurate representation of their specialized metabolome in the sediments. Since many bacterial species contain amazing biosynthetic potential that is unrealized in traditional culturing techniques, high throughput induction studies using ecologically relevant stressors will greatly inform our understanding of bacterial spp. true chemical input in the environment.

Overall, the work described in this thesis emphasizes the importance of combining in situ work with laboratory studies to gain a better understanding of how bacterial species interact with their environment. By first looking at the environmental metabolome, organisms that are likely producing compounds in situ can be prioritized for further study. Then the producing organisms can be brought back to the lab, in part, to determine how their specialized metabolites affect the community they come from. However, these studies require a prior knowledge concerning the potential producing organism, such as how to isolate and culture them. Using a model genera such as Salinispora facilitates these types of studies since there is

142

already extensive knowledge concerning this genus’s biosynthetic potential and how to culture the species (Jensen, Moore, and Fenical 2015). To fully understand how bacterial specialized metabolites are impacting the environment, more information about what sort of biosynthetic potential is found in sediment bacteria, as well as how to isolate and culture the organisms needs to be determined. Only once we know how to culture and manipulate the bacteria in the sediments will we be able to understand what is happening in these densely populated communities.

Microbial chemical ecology appears to be reaching an apex, in part, due to technological advancements (Jansson and Baker 2016). Thanks to advances in analytical techniques including imaging mass spectrometry and easy to use dereplication pipelines, we can now, rather quickly, get a snapshot of entire metabolomes (Bouslimani et al. 2014). Coupling these analyses with metagenomics allows us to dereplicate compounds, link them to known BGCs, and identify the possible producing organisms (Loureiro et al. 2018; Medema 2018; Nguyen et al. 2016). These advances are exciting because they indicate that the future of the field will not be limited by technology, but by asking the right questions. This allows us to shift from a more human-centric view of bacterial specialized metabolites to understanding how these compounds help structure entire communities.

143

5.2 References

Abdelmohsen, Usama Ramadan, Tanja Grkovic, Srikkanth Balasubramanian, Mohamed Salah Kamel, Ronald J. Quinn, and Ute Hentschel. 2015. “Elicitation of Secondary Metabolism in Actinomycetes.” Biotechnology Advances 33 (6): 798–811. doi:10.1016/j.biotechadv.2015.06.003. Abelson, P H. 1994. “Chemical Ecology.” Science (New York, N.Y.) 264 (5158). John Wiley & Sons: 487. Bouslimani, Amina, Laura M. Sanchez, Neha Garg, and Pieter C. Dorrestein. 2014. “Mass Spectrometry of Natural Products: Current, Emerging and Future Technologies.” Natural Product Reports 31 (6): 718. doi:10.1039/c4np00044g. Caporaso, J Gregory, Justin Kuczynski, Jesse Stombaugh, Kyle Bittinger, Frederic D Bushman, Elizabeth K Costello, Noah Fierer, et al. 2010. “Correspondence QIIME Allows Analysis of High- Throughput Community Sequencing Data Intensity Normalization Improves Color Calling in SOLiD Sequencing.” Nature Publishing Group 7 (5): 335–36. doi:10.1038/nmeth0510-335. Davies, Julian. 2006. “Are Antibiotics Naturally Antibiotics?” In Journal of Industrial Microbiology and Biotechnology, 33:496–99. doi:10.1007/s10295-006-0112-5. Doroghazi, James R., and William W. Metcalf. 2013. “Comparative Genomics of Actinomycetes with a Focus on Natural Product Biosynthetic Genes.” BMC Genomics 14 (1): 611. doi:10.1186/1471-2164-14-611. Fernandes, Prabhavathi. 2006. “Antibacterial Discovery and Development—the Failure of Success?” Nature Biotechnology. doi:10.1038/nbt1206-1497. Hay, Mark E. 2009. “Marine Chemical Ecology: Chemical Signals and Cues Structure Marine Populations, Communities, and Ecosystems.” Annual Review of Marine Science 1 (1): 193– 212. doi:10.1146/annurev.marine.010908.163708. Jansson, Janet K., and Erin S. Baker. 2016. “A Multi-Omic Future for Microbiome Studies.” Nature Microbiology 1 (5). Macmillan Publishers Limited: 16049. doi:10.1038/nmicrobiol.2016.49. Jensen, Paul R., Tracy J. Mincer, Philip G. Williams, and William Fenical. 2005. “Marine Actinomycete Diversity and Natural Product Discovery.” Antonie van Leeuwenhoek, International Journal of General and Molecular Microbiology. doi:10.1007/s10482-004- 6540-1. Jensen, Paul R, Bradley S Moore, and William Fenical. 2015. “The Marine Actinomycete Genus Salinispora: A Model Organism for Secondary Metabolite Discovery.” Natural Product Reports 00 (October): 1–14. doi:10.1039/c4np00167b. Lam, Kin S. 2006. “Discovery of Novel Metabolites from Marine Actinomycetes.” Current Opinion in Microbiology. doi:10.1016/j.mib.2006.03.004. Loureiro, Catarina, Marnix H. Medema, John van der Oost, and Detmer Sipkema. 2018. “Exploration and Exploitation of the Environment for Novel Specialized Metabolites.” Current Opinion in Biotechnology. doi:10.1016/j.copbio.2018.01.017.

144

Machado, Henrique, Robert N. Tuttle, and Paul R. Jensen. 2017. “Omics-Based Natural Product Discovery and the Lexicon of Genome Mining.” Current Opinion in Microbiology 39 (October): 136–42. doi:10.1016/j.mib.2017.10.025. Medema, Marnix H. 2018. “Computational Genomics of Specialized Metabolism: From Natural Product Discovery to Microbiome Ecology.” MSystems 3 (2). doi:10.1128/mSystems.00182-17. Nguyen, Don D., Alexey V. Melnik, Nobuhiro Koyama, Xiaowen Lu, Michelle Schorn, Jinshu Fang, Kristen Aguinaldo, et al. 2016. “Indexing the Pseudomonas Specialized Metabolome Enabled the Discovery of Poaeamide B and the Bananamides.” Nature Microbiology. doi:10.1038/nmicrobiol.2016.197. Nichols, D., N. Cahoon, E. M. Trakhtenberg, L. Pham, A. Mehta, A. Belanger, T. Kanigan, K. Lewis, and S. S. Epstein. 2010. “Use of Ichip for High-Throughput In Situ Cultivation of ‘Uncultivable’ Microbial Species.” Applied and Environmental Microbiology 76 (8): 2445– 50. doi:10.1128/AEM.01754-09. Okada, Bethany K, and Mohammad R Seyedsayamdost. 2017. “Antibiotic Dialogues: Induction of Silent Biosynthetic Gene Clusters by Exogenous Small Molecules.” Edited by Aimee Shen. FEMS Microbiology Reviews 41 (1): 19–33. doi:10.1093/femsre/fuw035. Patin, Nastassia V., Michelle Schorn, Kristen Aguinaldo, Tommie Lincecum, Bradley S. Moore, and Paul R. Jensen. 2017. “Effects of Actinomycete Secondary Metabolites on Sediment Microbial Communities.” Applied and Environmental Microbiology 83 (4). doi:10.1128/AEM.02676-16. Patin, Nastassia V, Katherine R Duncan, Pieter C Dorrestein, and Paul R Jensen. 2015. “Competitive Strategies Differentiate Closely Related Species of Marine Actinobacteria.” The ISME Journal 10 (2). Nature Publishing Group: 1–13. doi:10.1038/ismej.2015.128. Patin, Nastassia V, Dimitrios J Floros, Chambers C Hughes, Pieter C Dorrestein, and Paul R Jensen. 2018. “The Role of Inter-Species Interactions in Salinispora Specialized Metabolism.” Microbiology. http://mic.microbiologyresearch.org/content/journal/micro/10.1099/mic.0.000679.v1. Probandt, David, Thilo Eickhorst, Andreas Ellrott, Rudolf Amann, and Katrin Knittel. 2018. “Microbial Life on a Sand Grain: From Bulk Sediment to Single Grains.” ISME Journal 12 (2): 623–33. doi:10.1038/ismej.2017.197. Takken, Willem, and Marcel Dicke. 2006. “Chemical Ecology: A Multidisciplinary Approach.” In Chemical Ecology: From Gene to Ecosystem, 1–8. https://pdfs.semanticscholar.org/8ad0/804a4149704acb29b32d07eccee5565280e7.pdf. Thompson, Luke R., Jon G. Sanders, Daniel McDonald, Amnon Amir, Joshua Ladau, Kenneth J. Locey, Robert J. Prill, et al. 2017. “A Communal Catalogue Reveals Earth’s Multiscale Microbial Diversity.” Nature. doi:10.1038/nature24621. Traxler, Matthew F., Jeramie D. Watrous, Theodore Alexandrov, Pieter C. Dorrestein, and Roberto Kolter. 2013. “Interspecies Interactions Stimulate Diversification of the Streptomyces Coelicolor Secreted Metabolome.” MBio 4 (4): e00459-13. doi:10.1128/mBio.00459-13.

145

Wang, Mingxun, Jeremy J Carver, Vanessa V Phelan, Laura M Sanchez, Neha Garg, Yao Peng, Don Duy Nguyen, et al. 2016. “Sharing and Community Curation of Mass Spectrometry Data with Global Natural Products Social Molecular Networking.” Nature Biotechnology 34 (8): 828–37. doi:10.1038/nbt.3597. Wietz, Matthias, Katherine Duncan, Nastassia V. Patin, and Paul R. Jensen. 2013. “Antagonistic Interactions Mediated by Marine Bacteria: The Role of Small Molecules.” Journal of Chemical Ecology 39 (7): 879–91. doi:10.1007/s10886-013-0316-x.

146