Microbial and Chemical Diversity of Tetrodotoxin-Producing Marine Animals

Home , Brachycephalus pernix, Colostethus inguinalis, Gliotoxin, Lophozozymus, Parasagitta elegans, Takifugu xanthopterus

Microbial and chemical diversity of tetrodotoxin-producing marine animals

Rocky Chau

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Biotechnology and Biomolecular Sciences

Faculty of Science

August 2013

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ……………………………………………......

Date ……………………………………………......

AUTHENTICITY STATEMENT

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed …………………………………………………………….

Date …………………………………………………………….

ORIGINALITY STATEMENT

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed …………………………………………………………….

Date …………………………………………………………….

Abstract

Marine molluscs are known to employ a variety of defence systems to improve their survivability, such as the production of bioactive molecules. Recently, microbes have been identified as the true producers of these compounds.

Similarly, tetrodotoxin (TTX) is hypothesised to have a bacterial origin, however, much controversy still exists. This thesis attempts to shed further light on the biosynthetic origins of TTX with a focus on Hapalochalaena sp. and

Pleurobranchaea maculata. This thesis also investigates the potential for biosynthesis of other natural products by bacteria living in association with these marine molluscs. Culture-based studies attempting to isolate TTX-producing bacterium within Hapalochalaena sp. and P. maculata were unable to isolate any such strains. Furthermore, we attempted to replicate the production of TTX by published TTX-producing bacteria, however, TTX was unable to be detected in these strains via spectrometric methods. Nevertheless, culture-independent methods were able to identify four taxa that were strongly correlated to TTX- concentration in P. maculata. Further experiments, however, are required to isolate or characterise these strains. Molecular screening for natural product biosynthesis genes, putatively involved in the biosynthesis of TTX and other natural products, revealed many candidate bacteria. These experiments also identified a bacterium with significant biosynthetic potential, Pseudoalteromonas sp. HM-SA03. Members of Pseudoalteromonas are known to produce many bioactive compounds. Mining of the HM-SA03 genome identified a total of seven novel NRPS and PKS biosynthesis gene clusters. Bioinformatic analysis of these gene clusters revealed putative novel pathways for the assembly of

i bromoalterochromide, alteramide and pseudoalterobactin-like compounds.

However, production of these compounds in laboratory-cultures of

Pseudoalteromonas sp. HM-SA03 could not be confirmed by chemical studies.

Nonetheless, chemical studies were able to identify the production of eight diketopiperazines. These bioactive molecules have been observed to function as cell-signalling molecules, and are proposed to function as such in the complex microbial community identified within the Hapalochalaena octopus. Taken together, these results indicate that marine molluscs are a rich source of biosynthetically potent microbes that deserve further attention for the elucidation of new natural products.

ii Acknowledgements

First and foremost, I would like to thank my supervisor, Professor Brett A. Neilan for his support, motivation and most of all, his patience when things did not go to plan. I am most fortunate to study in his laboratory, under his tutelage.

My co-supervisor was Dr. John A. Kalatizis, the chemistry extraordinaire, who was always willing to bear my frustrations and offer me advice when things went wrong. He has tirelessly, and without complaint, proofed many of my manuscripts.

The warmest gratitude goes to my colleagues at the Cawthron Institute,

Dr Susanna Wood and Paul McNabb. They assisted me with all the work with

Pleurobranchaea maculata, including LC-MS analysis. I am grateful for the kind hospitality you showed during my visits to New Zealand.

Thanks to Dr. Mark Brown, who first introduced me to the world of bioinformatics and microbial ecology and Dr. Russell Pickford for his assistance with LC-MS.

Thank you to all the members of my laboratory, especially Jason and Sarah.

Working alongside you all has been a pleasure; you guys keep the lab fun!

Thanks to my rock climbing and F1 buddies, Apurv included, who have kept me distracted (and sane) outside of work.

Of course, I could not have gone through this without the support of my loving family.

The final and most special thanks must go to my beloved Rebecca, who has quietly and patiently supported me wholeheartedly throughout all these years.

iii List of publications and presentations

Publications:

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2014. Tetrodotoxin. In: Dongyou, L. (ed.)

Manual of Security Sensitive Microbes and Toxins, GBR, Taylor and Francis. In press.

Chau, R., Kalaitzis, J.A., Wood, S., Neilan, B., 2013. Diversity and biosynthetic potential of culturable microbes associated with toxic marine animals. Marine

Drugs 11, 2695-2712.

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2011. On the origins and biosynthesis of tetrodotoxin. Aquatic Toxicology 104, 61-72.

Kalaitzis, J.A., Chau, R., Kohli, G.S., Murray, S.A., Neilan, B.A., 2010. Biosynthesis of toxic naturally-occurring seafood contaminants. Toxicon 56, 244-258.

Presentations:

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2012. Genome of an octopus-derived

Pseudoalteromonas reveals unprecedented natural product gene clusters.

International Congress on Natural Products Research, New York.

iv Abstract ...... i Acknowledgements ...... iii List of publications and presentations ...... iv List of abbreviations ...... ix List of figures ...... xi List of tables ...... xiii Chapter 1: Introduction to the biosynthesis of natural products ...... 15 1.1. Biosynthesis of secondary metabolites ...... 17 1.1.1. Polyketide synthases ...... 19 1.1.2. Non-ribosomal peptide synthetases ...... 20 1.1.3. Structural diversity through biosynthetic exchangeability ...... 21 1.2. Genome mining for biosynthesis gene clusters ...... 21 1.2.1. Molecular screening for biosynthesis genes ...... 23 1.2.2. Identification of biosynthesis gene clusters through genome analysis .. 24 1.3. Biosynthesis of tetrodotoxin ...... 26 1.3.1. The guanidinium moiety of tetrodotoxin...... 27 1.1.3.1. Amidinotransferase involvement in tetrodotoxin biosynthesis ...... 27 1.1.3.2. Non-ribosomal peptide synthetase involvement in tetrodotoxin biosynthesis ...... 29 1.3.2. The carbon backbone of tetrodotoxin ...... 30 1.2.3.1. A polyketide derived carbon backbone...... 30 1.2.3.2. A sugar derived carbon backbone fragment ...... 31 1.2.3.3. A terpene derived carbon backbone fragment ...... 31 1.4. Thesis aims and rationale...... 33 Chapter 2: Diversity and biosynthetic potential of culturable microbes associated with toxic marine animals ...... 35 2.1. Introduction ...... 35 2.2. Materials and methods ...... 42 2.2.1. Specimen collection ...... 42 2.2.2. Specimen dissection ...... 42 2.2.3. Bacterial culturing and genomic DNA extraction ...... 43 2.2.4. Identification of bacterial isolates ...... 44 2.2.5. PCR screening of bacterial isolates for biosynthesis genes ...... 45

v 2.2.6. Phylogenetic analysis of 16S rRNA gene sequences ...... 46 2.2.7. Toxin extraction from organ homogenates and bacterial isolates ...... 47 2.2.8. Analysis of extracts by Liquid Chromatography-Mass Spectrometry ..... 47 2.2.9. Antimicrobial bioassayof HM-SA03 crude extracts ...... 48 2.3. Results and discussion ...... 50 2.3.1. Bacterial diversity in toxic Hapalochlaena sp. and Pleurobranchaea maculata ...... 50 2.3.2. Overestimation of TTX-producing bacteria in the literature ...... 57 2.3.3. Mining host-associated bacteria for proposed TTX biosynthesis genes 61 2.4. Concluding remarks ...... 64 Chapter 3: Bacteria from the Rhodobacteriaeceae correlate with the presence of tetrodotoxin in the grey side-gilled sea slug Pleurobranchaea maculata ...... 65 3.1. Introduction ...... 65 3.2. Materials and methods ...... 66 3.2.1. Pleurobranchaea maculata specimen collection ...... 66 3.2.2. Tetrodotoxin extraction and analysis ...... 67 3.2.3. DNA extraction ...... 67 3.2.4. Sequencing and quality control ...... 68 3.2.5. Statistical analyses ...... 69 3.3. Results and discussion ...... 69 3.3.1. Amplicon sequencing and processing ...... 69 3.3.2. Correlation between total microbial population and TTX ...... 70 3.3.3. Correlation between individual OTUs and TTX ...... 73 3.4. Concluding remarks ...... 81 Chapter 4: Genome sequencing of a Pseudoalteromonas bacterium reveals an unprecedented abundance of secondary metabolite gene clusters ...... 82 4.1. Introduction ...... 82 4.2. Materials and methods ...... 84 4.2.1. Sample preparation and genome sequencing ...... 84 4.2.2. Genome assembly ...... 85 4.2.3. Gene prediction and annotation ...... 86 4.2.4. Phylogenetic tree reconstruction of condensation domains ...... 86 4.2.5. Small molecule extraction ...... 86

vi 4.2.6. Mass spectrometry ...... 87 4.3. Results and discussion ...... 88 4.3.1. Genome sequencing and assembly ...... 88 4.3.2. Gene annotation and genome mining ...... 88 4.3.3. Identification and in-silico characterisation of an alterochromide biosynthesis pathway ...... 89 4.3.3.1. Identification of a putative alterochromide biosynthesis cluster ..... 89 4.3.3.2. Formation and loading of cinnamic acid ...... 93 4.3.3.3. Polyketide extension of cinnamate and lipoinitiation ...... 97 4.3.3.4. Non-ribosomal extension and heterocyclisation...... 100 4.3.3.5. Tailoring and diversity of alterochromide structures ...... 102 4.3.4. Identification and in-silico characterisation of NRPS siderophore biosynthesis pathways ...... 104 4.4.3.1. Identification of siderophore biosynthesis gene clusters ...... 104 4.4.3.2. Putative biosynthesis gene cluster for pseudoalterobactin-like compounds ...... 105 4.4.3.3. The sid biosynthesis gene cluster ...... 111 4.3.5. Hybrid bacterial iterative type I PKS-NRPS ...... 115 4.3.6. Other biosynthesis gene clusters encoded in HM-SA03 ...... 121 4.4. Concluding remarks ...... 124 Chapter 5: Characterisation of Pseudoalteromonas sp. HM-SA03 secondary metabolism ...... 125 5.1. Introduction ...... 125 5.2. Materials and methods ...... 127 5.2.1. Culturing and extraction ...... 127 5.2.2. Siderophore assay ...... 128 5.2.3. Nuclear magnetic resonance spectroscopy ...... 129 5.3. Results and discussion ...... 132 5.3.1. Siderophores from Pseudoalteromonas sp. HM-SA03 ...... 132 5.3.2. Diketopiperazines from Pseudoalteromonas sp. HM-SA03 ...... 132 5.2.3.1. Structure elucidation ...... 135 5.2.3.2. Biosynthesis pathways ...... 142 5.2.3.3. Ecological roles ...... 145 5.4. Concluding remarks ...... 147

vii Chapter 6: General discussion ...... 148 6.1. Research motivation and objectives ...... 148 6.2. Key findings ...... 149 6.2.1. Biosynthesis of TTX by laboratory-cultured bacteria ...... 149 6.2.2. Biosynthesis of TTX by uncultured microorganisms ...... 150 6.2.3. Genome mining as a tool for natural product discovery ...... 150 6.2.4. Identification of Pseudoalteromonas-derived secondary metabolites . 151 6.3. Future directions ...... 152 6.3.1. Re-examination of TTX-production in laboratory cultures ...... 152 6.3.2. Comparison of microbial communities in all TTX-producing higher organisms ...... 153 6.3.3. Heterologous expression of HM-SA03 gene clusters ...... 154 6.4. Closing remarks ...... 155 References ...... 156

viii List of abbreviations

A adenylation AHL N-acylhomoserine lactones AMT amidinotransferase ARISA Automated Ribosomal Intergenic Spacer Analysis AT acyltransferase BLAST Basic Local Alignment Search Tool C condensation CAS chrome azurol S CDD Conserved Domains Database CDPS cyclic dipeptide synthase CLF chain length factor COSY correlation spectroscopy Cy cyclisation d doublet DH dehydratase DHB 2,3-dihydroxybenzoate DKP 2,5-diketopiperazine DMAPP dimethylallyl pyrophosphate DNA deoxyribonucleic acid E epimerase ER enoylreductase ESI electrospray ionisation HAL histidine ammonia lyase HMBC heteronuclear multiple bond correlation HMM Hidden Markov Model HPLC high performance liquid chromatography HSQC heteronuclear single quantum coherence IPP isopentenyl diphosphate KR ketoreductase

ix KS ketosynthase LC liquid chromatography m multiplet MS mass spectrometry MT methyltransferase nMDS nonmetric multidimensional scaling NMR nuclear magnetic resonance NRPS non-ribosomal peptide synthetase ORF open reading frame OTU operational taxanomic unit PAL phenylalanine ammonia lyase PCR polymerase chain reaction PEG protein encoding gene PKS polyketide synthase PNTA 9-phenyl-nonatetraenoic acid q quartet quin quintet RAST Rapid Annotation using Subsystem Technology rRNA ribosomal ribonucleic acid s singlet STX saxitoxin T thiolation t triplet TAL tyrosine ammonial lyase TE thioesterase TTX tetrodotoxin

x List of figures

Figure 1.1 The structural diversity of natural products ...... 16

Figure 1.2 Outline of NRPS and PKS biosynthesis and associated tailoring modifications ...... 18

Figure 1.3 Shimizu and co-worker's proposed biosyntheses of tetrodotoxin ...... 27

Figure 1.4 Involvement of amidinotransferases in toxin biosynthesis pathways .... 29

Figure 1.5 Polyketide and terpenoid biosynthesis pathways ...... 32

Figure 2.1 Structures of tetrodotoxin and structurally-related toxins from marine algae and cyanobacteria ...... 36

Figure 2.2 Maximum likelihood phylogenetic tree of 16S rRNA gene sequences of isolates from this study and related bacteria ...... 56

Figure 2.3 Inhibitory effect of HM-SA03 methanol-derived crude extracts ...... 63

Figure 3.1 Rarefaction analysis, plotting number of observed OTUs against total sequences obtained, for each Pleurobranchaea maculata sample ...... 70

Figure 3.2 Non-metric multi-dimensional scaling plots showing clustering of the bacterial communities from Pleurobranchaea maculata ...... 72

Figure 4.1 The NRPS biosynthesis gene cluster and structure of bromoalterochromide ...... 91

Figure 4.2 Biosynthetic pathway for the production of the phenyl-polyketide starter unit for alterochromide ...... 94

Figure 4.3 Amino acid alignment of aromatic amino acid ammonia-lyases ...... 96

Figure 4.4 Phylogenetic tree of condensation domain subtypes ...... 99

Figure 4.5 NRPS pathway for alterochromide...... 101

Figure 4.6 Positive mode ESI mass spectra of HM-SA03 crude extract ...... 103

Figure 4.7 Biosynthesis pathway for the typical siderophore starter units in salicylate and DHB ...... 108

Figure 4.8 Proposed biosynthesis pathway for a pseudoalterobactin-like compound ...... 110

xi Figure 4.9 Hybrid NRPS/PKS biosynthesis pathway for the putative siderophore, HM-SA03-S022-1 ...... 114

Figure 4.10 Proposed Biosynthesis pathway of alteramide A ...... 119

Figure 4.11 Gene map showing gene clusters homologous to the alm gene cluster of Pseudoalteromoas sp. HM-SA03 ...... 120

Figure 4.12 Chemical structures of HM-SA03 compounds as predicted by bioinformatic analysis of NRPS and PKS gene clusters...... 122

Figure 5.1 Isolation scheme for HM-SA03 grown in liquid media ...... 130

Figure 5.2 Isolation scheme for HM-SA03 grown on solid media ...... 131

Figure 5.3 Diktopiperazines isolated from Pseudoalteromonas sp. HM-SA03...... 134

Figure 5.4 Connectivity of the benzyl group ...... 136

Figure 5.5 Connectivity of the 4-hydroxyprolyl group ...... 137

Figure 5.6 Connectivity of the diketopiperazine ring to the benzyl and hydroxyprolyl groups ...... 138

Figure 5.7 The CDPS pathway for 2,5-diketopiperazine biosynthesis ...... 1444

xii List of tables

Table 2.1 Phylogenetic distribution of organisms reported to be sources of TTX (excluding bacteria) ...... 40

Table 2.2 Distribution of TTX-producing bacteria ...... 41

Table 2.3 Primers used for identification and screening of bacterial isolates ...... 46

Table 2.4 Bacterial identification and screening for polyketide synthase (PKS), nonribosomal peptide synthetase (NRPS) and amidinotransferase (AMT) genes potentially related to tetrodotoxin (TTX)-biosynthesis in bacterial isolates from Hapalochlaena sp...... 54

Table 2.5 Bacterial identification and screening for polyketide synthase (PKS), nonribosomal peptide synthetase (NRPS) and amidinotransferase (AMT) genes potentially related to tetrodotoxin (TTX)-biosynthesis in bacterial isolates from Pleurobranchaea maculata ...... 55

Table 2.6 Screening of Nassarius semiplicatus bacterial isolates for putative TTX- biosynthesis genes and a comparison of ELISA and LC-MS methods for the detection of TTX ...... 58

Table 2.7 Detection of tetrodotoxin (TTX) in whole-organ homogenates ...... 58

Table 3.1 Analysis of similarity statistics comparing bacterial community composition based on Pleurobranchaea maculata sample factors ...... 73

Table 3.3 Forty most abundant bacterial families in toxic P. maculata ...... 77

Table 3.4 Common OTUs between P. maculata egg masses and adults ...... 80

Table 4.1 Proposed encoded functions of alterochromide-like biosynthesis gene cluster ORFs ...... 92

Table 4.2 Proposed encoded functions of pseudoalterobacin biosynthesis gene cluster ORFs ...... 109

Table 4.3 Proposed functions encoded by the sid siderophore biosynthesis gene cluster ORFs ...... 113

xiii Table 4.4 Proposed functions encoded by the proposed alteramide biosynthesis gene cluster ORFs ...... 118

Table 4.5 Domain architectures and predicted products of other NRPS/PKS biosynthesis gene clusters in HM-SA03 ...... 123

Table 5.1 NMR data for cyclo(D-4-hydroxyprolyl-L-phenylalanyl), 2 ...... 133

Table 5.2 NMR data for compounds 1, 3-8 ...... 139

xiv

Chapter 1

Chapter 1: Introduction to the biosynthesis of natural products

Parts of this chapter have been published as:

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2011. On the origins and biosynthesis of tetrodotoxin. Aquatic Toxicology 104, 61-72

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2014. Tetrodotoxin. In: Dongyou, L. (ed.) Manual of Security Sensitive Microbes and Toxins, GBR, Taylor and Francis. In press

The oceans are a rich source of biologically active and structurally novel natural products derived from a myriad of organisms and biosynthetic pathways (Blunt et al., 2010). Many of these natural products serve important ecological roles in the marine environment, such as species survival, and act as chemical defenses, camouflage agents, or anti-foulants (Hay and Fenical, 1996). The bioactivities of such compounds have been adapted for use by the pharmaceutical and biotechnology industries to drive drug discovery and develop chemical tools for the benefit of humans (Glaser and Mayer, 2009). In contrast, the same complement of organisms biosynthesise some of the most toxic substances known to mankind, ranging from those with neurotoxic and paralytic effects to others causing skin irritations (Kalaitzis et al., 2010). Often, the true producers of these marine natural products are microbes living in association with these animal hosts. These microbes possess complex biosynthetic pathways, which give rise to the highly unusual chemical structures found in marine natural products (Figure 1.1).

Examples include the alkaloids tetrodotoxin (TTX) and halitulin, the diterpenoid ineleganolide, and the large cyclic peptide, theonellamide F (Figure 1.1). Two of the most highly represented class of compounds are non-ribosomal peptides and

Chapter 1

polyketides, which have a remarkable scope for chemical diversity owing to their simple, yet diverse, biosynthesis pathways.

Figure 1.1 The structural diversity of natural products

Chapter 1

1.1. Biosynthesis of secondary metabolites

The chemical structures of non-ribosomal peptides and polyketides are structurally complex, however, their biosyntheses are derived from relatively simple substrates, such as polyketides carboxyl-CoA (polyketides) and amino acids

(non-ribosomal peptides). These compounds are biosynthesised by multimodular enzymes: polyketide synthase (PKS) and non-ribosomal peptide synthetase

(NRPS), respectively ( Fischbach and Walsh, 2006). Modular type I PKS and NRPS are multifunctional enzymes consisting of catalytic domains organised into modules. A minimal module for a modular type I PKS and NRPS consists of three essential domains: a domain for activating the monomer substrate, a carrier protein to which activated substrates are covalently tethered and a domain catalysing bond formation between the enzyme-bound substrate and the nascent polyketide chain. Each module of the biosynthetic enzyme facilitates the incorporation of a single structural monomer onto the growing non-ribosomal peptide or polyketide (Figure 1.2, A, B). Together, these elegant microbial biosynthetic systems generate a staggering diversity of polyketides, non-ribosomal peptides and hybrid compounds.

Chapter 1

A1 T C A2 T A1 T C A2 T A

amido acid A amido acid B

linear peptide

B AT T KS AT T AT T KS AT T

acyl thioester malonate linear polyketide C TE

nMT

Cy Ox Hypothetical cyclised compound Cy

KR DH ER KR DH

KR cMT

AMT Hal

Figure 1.2 Outline of NRPS and PKS biosynthesis and associated tailoring modifications. Mechanism of one iteration of NRPS (A) or PKS (B) chain extension. A hypothetical compound, tethered to a thioesterase (TE), highlighting the structural diversity generated by NRPS and PKS tailoring domains (C); MT: methyltransferase, E: epimerase, Cy: cyclisation-condensation, Ox: oxidase, DH: dehydratase, KR: ketoreductase, ER: enoylreductrase, AMT: aminotransferase, Hal: halogenase.

Chapter 1

1.1.1. Polyketide synthases

Polyketides represent a large class of natural products with vast structural diversity. Their biosynthesis pathways are similarly diverse and multiple subclasses of PKS exist. Modular type I PKS resemble fatty acid synthases, and are organised in a modular fashion, with each biosynthetic module responsible for the incorporation of a single carboxylacyl monomer into the final compound

(Fischbach and Walsh, 2006; Moffitt and Neilan, 2003). Within each biosynthetic module are catalytic domains which perform specific reactions in the biosynthesis pathway. As mentioned previously, three essential domains are required in a minimal biosynthethic module. In a PKS, these are the acyltransferase (AT), thiolation (T) and ketosynthase (KS) domains. The AT domain facilitates the selection and attachment of acyl-CoA thioesters to a phosphopantetheinyl chain present on the T domain. The KS domain catalyses C-C bond formation between the T domain bound substrate and the growing polyketide chain attached on the preceding T domain. These two biosynthetic steps represent one cycle of polyketide elongation. Within each biosynthetic module, additional catalytic domains may be present, which govern modifications to the growing polyketide chain. Such tailoring domains include ketoreductase (KR), dehydratase (DH), and enoylreductase (ER) domains that govern bond reduction in the polyketide chain

(Figure 1.2, C). The fully mature linear polyketide chain is transferred to an active site serine on a thioesterase (TE), which catalyses cyclisation of the polyketide and its release from the enzymatic machinery.

PKS systems also encompass iterative biosynthesis pathways, which include iterative type I PKS and type II PKS systems. Iterative type I PKS systems contain

Chapter 1

the same core domains as their modular counterparts, however multiple rounds of chain elongation are catalysed by a single set (or module) of catalytic domains.

Similarly, type II PKS facilitate iterative chain extension, however the enzymatic architecture of these PKS are different to type I PKS. Where type I PKS catalytic domains are localised on a single multifunctional enzyme and act in cis, the biosynthetic enzymes of type II PKS systems are monofunctional, with each functional domain encoded on a discrete protein and act in trans (Fischbach and

Walsh, 2006). Type III PKS, typically found in plants, also utilise an iterative biosynthetic strategy. Type III PKS enzymes incorporate acyl-CoA monomers in the absence of T domains ( Shen, 2003; Watanabe et al., 2007). Although the substrates of PKS and NRPS differ, the principles behind the biosynthetic mechanisms of these biosynthetic systems share many similarities.

1.1.2. Non-ribosomal peptide synthetases

The NRPS pathway is an alternate system to ribosomes for the generation of peptidyl compounds in prokaryotic organisms. While the substrate selection for ribosome-dependent protein biosynthesis is restricted to the 21 proteinogenic amino acids, NRPS-dependent biosynthesis does not share this restriction and is able to utilise non-proteinogenic amino acids. This vast array of substrates available to NRPSs gives rise to the potential for a wide variety of peptide configurations, such as nostopeptolide, which incorporates the non-proteinogenic amino acids L-4-methylproline and L-leucylacetate (Hoffmann et al., 2003).

Interestingly, although the chemical structures of polyketides and non-ribosomal peptides are different, their biosyntheses share many similarities. Homologous to modular type I PKS, NRPS consist of multiple catalytic domains organised in a

Chapter 1

modular fashion ( Mootz and Marahiel, 1999; Mootz et al., 2002). The adenylation

(A), thiolation (T) and condensation (C) domains are the minimal requirements for a NRPS module. The A domain catalyses the activation of amino acids, covalently tethering these monomers onto the phosphopantetheinyl “arm” on the T domain.

Each A domain is selective for the activation of a specific amino acid. Following this, the C domain catalyses C-N bond formation between the ketone carbon and amide nitrogen of substrates tethered on T domains of adjacent modules. Finally, termination and release of the compound from the enzyme is facilitated by a TE.

Tailoring domains encoded within an NRPS module can modify the growing peptide chain in a variety of ways. These include epimerisation (E) domains that alter the stereochemistry of aminoacyl groups and cyclisation (Cy) domains that convert serine and cysteine to their oxazole and thiozole derivatives, respectively

(Figure 1.2, C).

It can be seen that both PKS and NRPS share similar biosynthetic strategies. Hence, it is unsurprising that numerous examples of hybrid natural PKS-NRPS products utilising both PKS and NRPS biosynthetic enzymes exist. The biosynthetic flexibility of individual NRPS or PKS pathways are magnified when these two complimentary biosynthetic systems are unified to form hybrid NRPS-PKS-derived natural products ( Du et al., 2001).

1.1.3. Structural diversity through biosynthetic exchangeability

Individually, both PKS and NRPS have incredible scope for the generation of diverse structural chemistries (Figure 1. 2, C). Therefore, the combination of these

Chapter 1

two biosynthetic systems is able to generate hybrid PKS-NRPS products with the potential for extensive structural diversity (Du et al., 2001).

Adding to the structural diversity of PKS/NRPS products is the utilisation of unusual substrates in their biosynthesis. There are many examples of natural products that incorporate unusual starter substrates, which have been biosynthesised through separate pathways present in the organism. Lipopeptides utilise fatty-acyl CoA starter units, which are biosynthesised through the fatty-acid biosynthesis pathways, in their biosynthesis. These substrates are recruited into

NRPS and PKS pathways to form lipomolecules, such as surfactin (Kraas et al.,

2010) and daptomycin (Wittmann et al., 2008).

Polyketides, especially, have great structural diversity due to their chain length variability, extent of bond reduction and potential for cyclisation at many sites in the structure. Numerous unusual polyketides, such as enterocin (Piel et al., 2000), hyperforin (Adam et al., 2002) and lathrophytoic acid A (de Almeida et al., 2011) contain unusual carbon skeletons which arise as a result of complex biosynthetic cyclisations. These structures contain a caged adamantane-like carbon backbone, which is relatively rare in nature. Of particular interest is the polyketide, TW93h, which is produced by an engineered strain of Streptomyces coelicolor (Shen et al.,

1999). This polyketide contains a 2,4-dioxaadamantane structure that is extremely rare in nature. The only natural product that shares this cage-like carbon skeleton is the marine toxin, tetrodotoxin (TTX). Little is known about the biosynthesis of this unusual molecule. The elucidation of the biosynthesis of such an unusual chemical structure may reveal novel enzymology that can be harnessed in the

Chapter 1

generation of new natural products. Therefore, TTX is an excellent candidate for biosynthetic studies.

1.2. Genome mining for biosynthesis gene clusters

The study of biosynthetic pathways first requires the identification of genes involved in the assembly of natural products. Searching for the genes responsible for the biosynthesis of TTX and other microbial natural products can be performed through two routes, molecular screening and genome analysis.

1.2.1. Molecular screening for biosynthesis genes

Our efforts to understand the biosynthesis of natural products and the genetics underpinning their assembly are based on observation that the genes encoding

NRPS and PKS are clustered together in the genome. Therefore, a positive identification of any one of the biosynthesis genes would allow for the identification of the entire gene cluster ( Udwary et al., 2007). PCR-based approaches facilitate the identification of biosynthesis gene clusters. KS and A domains, which are essential domains in PKS and NRPS biosynthesis pathways, respectively, are functionally and genetically conserved ( Schwarzer and Marahiel,

2001). Therefore, the genes encoding these regions are amicable to PCR-based screening strategies using degenerate oligonucleotides targeting conserved protein motifs within these regions (Burns et al., 2005; Ehrenreich et al., 2005).

Screening for the presence of NRPS and PKS genes is simple using PCR-based methods. Therefore, molecular screening is a rapid method for assessing the biosynthetic potential of microbial isolates. However, obtaining the complete gene sequence of the entire cluster is time consuming and expensive using traditional

Chapter 1

Sanger sequencing technology. With the increasing availability of whole-genome sequencing and new computational methods for genome assembly and genome mining, next-generation sequencing and analysis methods for the identification of complete biosynthesis gene clusters is proving more efficient than traditional methods.

1.2.2. Identification of biosynthesis gene clusters through genome analysis

The increasing accessibility of genome sequencing has led to a marked increase in availability of genome data. Currently, over 10,000 bacterial genomes are deposited in sequence repositories such as the NCBI. The advancement of bioinformatics tools used to analyse this trove of data is also progressing rapidly, leading to the development of tools to allow for the mining of natural products biosynthesis gene clusters in these genomes.

It is currently known that microbial genomes encode the biosynthesis of more compounds than are observed from normal laboratory fermentations ( Zerikly and

Challis, 2009). However, using in silico methods we are able to identify these gene clusters and predict their encoded products. Inferred physicochemical properties of these predicted compounds can then be used to assist in the identification of these compounds from complex fermentation extracts. The identification of these gene clusters using bioinformatics methods is facilitated by a few concepts intrinsic to NRPS and PKS biosynthesis.

In many cases, the number of modules in an NRPS or PKS gene cluster corresponds to the number of substrates incorporated into the final compound (Corre and

Challis, 2009). Additionally, the order in which these substrates are incorporated

Chapter 1

into the compound is representative of the order in which these genes are encoded

(Corre and Challis, 2009). Tailoring enzymes such as methyltransferases are generally encoded within the module incorporating the site of modification.

Therefore, site modification by tailoring domains is amicable to prediction (Zerikly and Challis, 2009). Finally, the large amount of sequence data available for NRPS and PKS biosynthesis genes allows the prediction of A and KS domains substrate specificities (Medema et al., 2011). Therefore, in many cases, we are able to accurately predict the final product produced by NRPS and PKS gene clusters.

The predicted structure of the final compound can be used to assist in the elucidation of these compounds ( Zerikly and Challis, 2009). Physicochemical properties such as polarity and molecular weight assist the correct extraction and separation strategies for these compounds. The presence of functional groups with characteristic properties can assist in the selection of proper detection methods for the predicted compound. Additionally, knowledge of the function or activity of the compound could assist in selecting biological assays suitable for detection of the predicted natural product. A similar genome-guided approach was successfully used for the isolation of salinilactam A (Udwary et al., 2007). Genome mining revealed the presence of a lysine-primed PKS gene cluster. Fermentation broths were examined for UV chromophores that were characteristic of the predicted polyketide and led to the subsequent isolation of salinilactam A. It is clear that genome mining is a powerful tool for the discovery of new natural products and their biosyntheses.

Chapter 1

1.3. Biosynthesis of tetrodotoxin

The initial focus of this thesis is the biosynthesis of tetrodotoxin. Tetrodotoxin, whose name is derived from Tetraodontidae, the family of puffer fish from which it was first isolated ( Yokoo, 1950) is found in many organisms including but not limited to, frogs ( Kim et al., 1975), blue-ringed octopus (Sheumack et al., 1984), salamanders (Brown and Mosher, 1963) and gastropods (McNabb et al., 2010). Its presence in such a wide diversity of organisms suggested that TTX was microbial in origin. TTX is a neurotoxin with a highly unusual structure, whose biosynthesis has yet to be elucidated nearly a century after its discovery. Only a few pathways to

TTX have been proposed (Figure 1.3). Furthermore, there is no convincing data in the literature to support any of these hypotheses. Its unique structure, and thus the lack of comparative biosynthesis studies, has hindered the development of molecular tools for the purpose of unravelling the pathway. There are several plausible routes to TTX biosynthesis, including the involvement of NRPS and PKS enzymes. It is likely that some biosynthesis enzymes involved in the assembly of similar structural moieties could, by analogy, be functioning in a similar manner in a TTX producer. Some possible TTX biosynthesis enzymes will be discussed, in terms of the structural fragments incorporated into the final product, in the following sections.

Chapter 1

Figure 1.3 Shimizu and co-worker's proposed biosyntheses of tetrodotoxin (Kotaki and Shimizu, 1993). The structure of TTX was proposed to derive from a condensation between arginine and either an apiose-type C5 sugar (a) or isoprene (b).

1.3.1. The guanidinium moiety of tetrodotoxin.

The guanidinium moiety of TTX (Figures 1.3 and 1.4) is important for its toxicity and serves as a good target, initially, for predicting the biosynthesis of the molecule due to its rarity in secondary metabolites. Two routes to the incorporation of a guanidinium moiety in TTX are plausible. The moiety could be either transferred from an amidino-donor via an amidinotransferase (AMT), or incorporated via an NRPS module incorporating arginine. Both of these routes have parallels in characterised natural product biosyntheses, and are equally likely to occur. Feeding studies using guanido-14C labeled arginine have been performed

(Shimizu and Kobayashi, 1983) but were unable to elucidate the biosynthesis of

TTX. Apart from this, there has been virtually no research into the origin of the guanidinium moiety of TTX.

1.1.3.1. Amidinotransferase involvement in tetrodotoxin biosynthesis

Amidinotransferases are enzymes that are responsible for the transfer of guanidino groups from donors to acceptors. In the biosynthesis of saxitoxin, the 27

Chapter 1

amidinotransferase, SxtG, transfers the amidino group from L-arginine to the saxitoxin precursor 4,7-diguanidino-3-oxoheptane (Kellmann et al., 2008) (Figure

1.4). Saxitoxin shares many similarities with TTX, they are both relatively small toxins, saxitoxin (C10H17N7O4) is 299 g/mol, and TTX (C11H17N3O8) is 319 g/mol, both possess a guanidino group, and both are sodium channel blockers which bind to the same site on voltage-gated sodium channels (Lipkind and Fozzard, 1994).

Importantly, both are believed to be biosynthesised by microorganisms.

Interestingly, many animals such as xanthid crabs, Atergatis floridus, and puffer fish, Takifugu oblongus and Fugu pardalis, which are proposed TTX producers have also been linked to saxitoxin production (Arakawa et al., 1995; Jang and Yotsu-

Yamashita, 2006; Ngy et al., 2009). Therefore, it is plausible that the biosynthesis of saxitoxin and TTX may involve similar mechanisms. In TTX biosynthesis, an amidinotransferase could facilitate the attachment of an amidino group onto β- alanine to form a guanidinopropionate (Figure 1.4). This substrate could then be utilised as the precursor for subsequent biosynthesis steps, possibly encoded by

PKS genes.

Chapter 1

Figure 1.4 Involvement of amidinotransferases in toxin biosynthesis pathways. Amidinotransfer in the biosynthesis of saxitoxin (A) and as proposed here for tetrodotoxin (B).

1.1.3.2. Non-ribosomal peptide synthetase involvement in tetrodotoxin biosynthesis

The guanidinium moiety of TTX may be derived from arginine as originally predicted by Shimizu (Kotaki and Shimizu, 1993). Arginine is proposed to condense with either a branched apiose sugar or an isoprene unit to form the general structure of TTX (Figures 1.3 and 1.5). The condensation of arginine in this manner would likely be catalysed by an NRPS. In such a pathway, the NRPS would likely be clustered with other biosynthesis enzymes to assemble the carbon backbone, which may be polyketide, sugar or terpene derived as previously proposed (Kotaki and Shimizu, 1993; Woodward and Gougoutas, 1964; Yasumoto et al., 1988).

Chapter 1

1.3.2. The carbon backbone of tetrodotoxin

TTX contains a unique, highly oxygenated, 2,4-dioxaadamantane carbon backbone.

Due to its unique structure, many hypotheses regarding the origin of the carbon backbone of TTX have been proposed, including polyketide (Woodward and

Gougoutas, 1964), C5 branched sugar (Kotaki and Shimizu, 1993), and C5 isoprene origins (Yasumoto et al., 1988). Incorporation of C5 units would require priming by arginine rather than an amidinotransferase catalysed guanidine transfer to account for all carbon atoms in TTX.

1.2.3.1. A polyketide derived carbon backbone

A polyketide pathway, as proposed by Woodward (Woodward and Gougoutas,

1964), is feasible if the guanidinium is derived from an amidinotransfer onto a three carbon unit acceptor (Figure 1.5) which is then extended by malonyl CoA derived acetates, to account for all 11 carbons in the backbone (Figure 1.5).

Conversion of the nascent polyketide into the caged structure of TTX does not appear straight-forward and there are no biosynthetic precedents in the literature.

As previously mentioned, Moore and co-workers have reported a polyketide,

TW93h, containing a 2,4-dioxaadamantane structure (Figure 1.5), similar to that of

TTX indicating that such structures are possible via a PKS pathway (Shen et al.,

1999). It is notable that TW93h, however, is assembled by a type II PKS rather than the above-mentioned type I PKS and such systems should not be ignored in examining the TTX biosynthesis pathway. Examples of hybrid NRPS/type II PKSs, however, have not been reported.

Chapter 1

1.2.3.2. A sugar derived carbon backbone fragment

Shimizu proposed that the highly oxygenated structure of TTX may be derived via the condensation of arginine with an oxygenated, branched C5 sugar, such as apiose, which is known to occur in marine environments (Kotaki and Shimizu,

1993). Apiose incorporation into natural products is rare and only plant natural products, such as conyzasaponin produced by Conyza blinii (Su et al., 2001), have been reported to possess apiose sugars. Additionally, the identification of deoxy-

TTX precursors (Kotaki and Shimizu, 1993) indicates that apiose incorporation is unlikely. The incorporation of apiose into TTX would indeed result in a highly oxygenated carbon backbone, such as that found in TTX, however, the isolation of deoxy-TTX indicates that a less oxygenated substrate, such as an isoprene unit rather than an apiose sugar is more likely (Kotaki and Shimizu, 1993).

1.2.3.3. A terpene derived carbon backbone fragment

Natural products of the terpene class are derived from C5 building blocks such as dimethylallyl pyrophosphate (DMAPP) and isopentenyl diphosphate (IPP)

(Dewick, 2002). Yasumoto proposed that a mixed pathway incorporating an isoprene unit and a single arginine to form TTX could be plausible (Yasumoto et al.,

1988). Stepwise oxygenation reactions would then follow to finally yield the highly oxygenated carbon backbone present in TTX (Figure 1.5). Regardless of whether

TTX biosynthesis incorporates a polyketide, sugar or terpene substrate, the pathway employed would be unique, and unlike any other pathway investigated to date.

Chapter 1

Figure 1.5 Polyketide and terpenoid biosynthesis pathways. Proposed biosynthesis of the caged, type II PKS product, TW93h in Streptomyces coelicolor (A). Incorporation of an IPP derived isoprene into TTX (B) as proposed by Shimizu (Kotaki and Shimizu, 1993).

Chapter 1

1.4. Thesis aims and rationale

Initially, the aim of this thesis was to elucidate the biosynthetic origins of TTX.

Accordingly, genetic screening of bacterial isolates from TTX-producing molluscs,

Hapalochlaena sp. and Pleurobranchaea maculata for the presence of genes that are putatively involved in the biosynthesis of TTX was undertaken (chapter 2).

Such a molecular approach to understanding the biosynthesis of TTX has not been described in the literature. Indeed, very little work on the biosynthesis of TTX has been ever undertaken. This approach was able to assess the diversity of culturable microbes from TTX-producing organisms and assess their biosynthetic potential.

Another area of TTX research unique to this study is the involvement of unculturable or yet-to-be-cultured microbes in TTX biosynthesis. An effort to identify the total microbial community present in the TTX producer, P. maculata, was undertaken (chapter 3). This approach revealed putative correlations between animal-associated microorganisms and TTX production.

Although a TTX-producing organism was unable to be isolated from either mollusc species, the presence of such a large proportion of organisms with high biosynthetic potential prompted further investigation. If TTX was not being produced by these host-associated microorganisms, what was, and why?

Therefore, an analysis of the genome of a Pseudoalteromonas isolate, which was one of the most biosynthetically rich and dominant microbes in Hapalochlaena sp., was performed (chapter 4). Seven novel gene clusters encoding NRPS and PKS biosynthesis pathways were identified for which products were hitherto unknown.

Chapter 1

Concurrent to this, an investigation of secondary metabolites produced by the

Pseudoalteromonas isolate was performed (chapter 5). This study aimed to identify whether any secondary metabolites could be linked to the NRPS and PKS gene clusters identified in chapter 4, or whether any non-NRPS or PKS products were being produced.

Ultimately, this thesis attempts to better understand the biogenic origins of TTX and characterise the microbial and chemical diversity present in TTX-producing organisms.

Chapter 2

Chapter 2: Diversity and biosynthetic potential of culturable microbes associated with toxic marine animals

This chapter has been published as:

Chau, R., Kalaitzis, J.A., Wood, S., Neilan, B., 2013. Diversity and Biosynthetic Potential of Culturable Microbes Associated with Toxic Marine Animals. Marine Drugs 11, 2695-2712

2.1. Introduction

Soft-bodied marine organisms, particularly sessile or slow moving invertebrates, have evolved complex chemical defence systems to enhance their competitive fitness and, ultimately, survival (Pawlik, 1993). Some of the common, but often complex, chemical defences serve to function as anti-foulants, camouflaging agents, or UV-absorbing sunscreens. The same marine organisms possess little or no means of physical defence against predation and therefore aggressively combat predators with lethal toxins. The production of toxins by marine invertebrates is one of the most intriguing and intensely studied facets of natural chemical defence systems. The focus of this study is one of the better known marine toxins, tetrodotoxin (TTX; Figure 2.1).

TTX is one of the most toxic natural substances known and symptoms of human

TTX poisoning include numbness of the face and extremities, paralysis, respiratory failure, circulatory collapse, and death (Ahasan et al., 2004). TTX is a sodium channel blocker and some TTX-bearing organisms, including Taricha newts and

Tetraodontidae fish, nullify the auto-effect of the toxin by possessing TTX-resistant sodium channels (Venkatesh et al., 2005), while other species, for example the

Chapter 2

shore crab Hemigrapsus sanguineus, have been found to produce TTX-binding compounds that are able to neutralise the effects of TTX (Shiomi et al., 1992).

Figure 2.1 Structures of tetrodotoxin and structurally-related toxins from marine algae and cyanobacteria.

TTX has been reported from taxonomically diverse organisms across eight different phyla (Table 2.1) (Chau et al., 2011), including puffer fish from the family

Tetraodontidae from which it takes its name, the blue-ringed octopus

Hapalochlaena maculosa (Sheumack et al., 1984), the grey side-gilled sea slug

Pleurobranchaea maculata (McNabb et al., 2010) and marine gastropod Nassarius semiplicatus (Wang et al., 2008). This widespread occurrence strongly suggests that microorganisms are the true source of TTX in nature. It is unlikely that the

Chapter 2

complex biosynthetic machinery, responsible for TTX production, would have co- evolved in such numerous phylogenetically and environmentally diverse higher organisms and thus it has been proposed that TTX biosynthesis is exogenous to the host. The broad distribution of TTX may be due to horizontal transfer, via acquisition through the food chain, of TTX-producing bacteria (Noguchi and

Arakawa, 2008; Noguchi et al., 2006b). This hypothesis is supported by studies in which TTX-producing bacteria have been isolated from many different host organisms (reviewed in Chau, et al., 2011).

TTX-producing bacteria belonging to the genera Vibrio, Bacillus and Pseudomonas have been isolated from numerous organisms including blue-ringed octopus

(Hapalochlaena sp.), puffer fish (Fugu spp.), and deep sea sediments (Table 2.2)

(Chau et al., 2011). Numerous studies have identified multiple species of toxic bacteria from a single host organism, suggesting that TTX -producing bacteria are abundant in these organisms. Twenty TTX-producing bacterial strains from at least two genera have been isolated from the common puffer fish Fugu rubripes (Wu et al., 2005a). Similarly, 21 TTX-producing bacterial strains belonging to five genera have been isolated from the gastropod N. semiplicatus (Wang et al., 2008). It is pertinent to note that most of these studies were performed prior to the 1990s, and thus 16S ribosomal RNA gene sequence analysis was not used as a tool for determining bacterial taxonomy and phylogeny, as it is today. Because of the inconsistencies which are often encountered when classifying microorganisms using a polyphasic approach, there is an urgent need to reassess microbial communities derived from TTX-containing host organisms.

Chapter 2

Despite evidence of bacterial production, argument over TTX’s origin still exists.

The endogenous theory of production proposes that TTX-containing organisms have the ability to produce TTX independently. While TTX has been isolated from bacteria in some animals, it has been reported that some TTX-associated organisms, such as Taricha torosa, do not harbour toxic endosymbionts (Lehman and Brodie, 2004). Furthermore, Cardall and co-workers (Cardall et al., 2004) demonstrated that Taricha granulosa release TTX when electro-stimulated and that subsequently, many newts, fed a non-TTX containing diet, are able to regenerate TTX over a period of nine months, in captivity.

Modern molecular tools have been developed that enable the identification of a microorganism’s genetic potential to biosynthesise discrete small bioactive molecules. Ultimate proof of TTX’s proposed microbial origin, or indeed that of any microbial natural product, could be achieved by identifying biosynthesis genes involved in its assembly. However, in the case of TTX, its biosynthetic pathway remains elusive, as does a stable and reliable microbial producer of the molecule. A number of TTX biosynthesis pathway proposals have been documented (Kotaki and Shimizu, 1993; Woodward and Gougoutas, 1964; Yasumoto et al., 1988), however, no supporting or experimental evidence has been published. In addition, biosynthetic feeding studies incorporating 14C-labeled acetate and guanido-14C labelled arginine have failed to reveal the mechanism or pathway for TTX biosynthesis (Shimizu and Kobayashi, 1983). A focus of this research is to understand the biosynthesis of small, naturally occurring toxins by microorganisms at the molecular level. No conclusive evidence has been published for the biosynthesis of TTX, however, based on our genetic understanding of biosynthesis of the guanidine-containing toxins, saxitoxin and cylindrospermopsin, 38

Chapter 2

and the dioxoadamantane-containing polyketide, TW93h (Kellmann et al., 2008;

Mihali et al., 2008; Shen et al., 1999) (Figure 2.1), we speculate that TTX is assembled by a hybrid polyketide synthase (PKS)/non-ribosomal peptide synthetase (NRPS) enzyme complex which possibly incorporates an amidinotransferase (AMT) (Chau et al., 2011). The genes encoding TTX biosynthesis enzymes along with those associated with toxin regulation and transport are proposed to be clustered on a single genome, in a similar manner to other toxin biosynthesis pathways in bacteria. This study describes the initial isolation of microbial populations from two TTX-associated host organisms,

Hapalochlaena sp. and P. maculata. Additionally, LC-MS methods were used to reassess the production of TTX from N. semiplicatus bacterial isolates that had been previously identified as TTX-producers using immunoassay-based methods

(Wang et al., 2008).

In order to identify biosynthesis genes putatively involved in small molecule assembly, including TTX, we mined the genomes of a library of diverse bacteria, isolated from these organisms, using a degenerate PCR-based screening approach.

Here we present the outcomes of our bacterial screening and report on the diversity of the culturable microbial communities associated with these hosts.

Chapter 2

Table 2.1 Phylogenetic distribution of organisms reported to be sources of TTX (excluding bacteria)

Phylum Class Family Binomial name Reference Chordata Amphibia Dendrobatidae Colostethus inguinalis (Daly et al., 1994) (includes frogs, Brachycephalidae Brachycephalus pernix (Pires et al., 2005) puffer fish, Brachycephalus ephippium (Pires et al., 2002) gobies and newts) Bufonidae Atelopus oxyrhynchus (Mebs and Schmidt, 1989) Salamandridae Notophthalmus viridescens (Yotsu-Yamashita and Taricha torosa Mebs, 2003) (Brown and Mosher, 1963) Taricha cynops (Hanifin, 2010) Taricha paramesotriton Taricha triturus Actinopterygii Tetraodontidae Takifugu xanthopterus (Nagashima et al., 2001) Takifugu niphobles (Yu et al., 2004) Fugu rubripesa (Wu et al., 2005a; Wu et al., 2005b) Fugu vermicularisa (Lee et al., 2000; Noguchi et al., 1987) Fugu pardalisa (Yasumoto et al., 1986) Fugu poecilonotusa (Yasumoto et al., 1986; Yotsu et al., 1987) Gobiidae Gobius criniger (Noguchi and Hashimoto, 1973) Mollusca Cephalopoda Octopodidae Hapalochlaena maculosaa (Hwang et al., 1989; (gastropods) Sheumack et al., 1984) Gastropoda Nassariidae Niotha clathrata (Jeon et al., 1984) Nassarius semiplicatusa (Wang et al., 2008) Muricidae Rapana rapiformis (Hwang et al., 1991) Rapana venosa venosa (Hwang et al., 1991) Ranellidae Charonia sauliae (Narita et al., 1981) Buccinidae Babylonia japonica (Noguchi et al., 1981) Naticidae Polinices didyma (Shiu et al., 2003) Bursidae Tutufa lissostoma (Noguchi et al., 1984) Nemertea Anopla Cephalothricidae Cephalothrix rufifronsa (Carroll et al., 2003) (nematodes) Lineidae Lineus longissimusa (Carroll et al., 2003) Echinodermata Stelleroidea Astropectinidae Astropecten latespinosus (Maruyama et al., 1984) (starfish) Astropecten polyacanthusa (Miyazawa et al., 1985)

Chaetognatha Sagittoidea Sagittidae Flaccisagitta enflata (Thuesen and Kogure, (arrow worms) 1989) Parasagitta elegans (Thuesen and Kogure, 1989) Pterosagittidae Zonosagitta nagae (Thuesen and Kogure, 1989) Eukrohniidae Eukrohnia hamata (Thuesen and Kogure, 1989) Arthropoda Merostomata Cleroidea Carcinoscorpius (Dao et al., 2009; (crabs) rotundicauda Kungsuwan et al., 1987) Malacostraca Carpiliidae Lophozozymus pictor (Tsai et al., 1995) Xanthoidea Atergatis floridus (Noguchi et al., 1983) Platyhelminthes Turbellaria Planoceridae Planocera multitentaculata (Miyazawa et al., 1986) (flatworms) Dinoflagellata Dinophyceae Goniodomataceae Alexandrium tamarense (Kodama et al., 1996) a bacterial production of TTX reported (Table 2.2).

Chapter 2

Table 2.2 Distribution of TTX-producing bacteria

TTX producing bacteria Source of bacterial isolate Reference Vibrio spp. Atergatis ﬂoridus (reef crab) (Noguchi et al., 1983) Fugu vermicularis vermicularis (common (Noguchi et al., 1987) puffer) Four species of Chaetognaths (Thuesen and Kogure, (arrowworms) 1989) Astropecten polyacanthus (comb seastar) (Narita et al., 1987) Hapalochlaena maculosa (blue-ringed (Hwang et al., 1989) octopus) Deep sea sediment (Do et al., 1990) Fugu vermicularis radialis (common puffer) (Lee et al., 2000) Seven species of nemertean worms (Carroll et al., 2003) Puffer fish species (Yu et al., 2004) Nassarius semiplicatus (sea snail) (Wang et al., 2008) Niotha clathrata (marine gastropod) (Cheng et al., 1995) Pseudomonas spp. Jania spp. (red alga) (Yasumoto et al., 1986) Fugu poecilonotus (common puffer) (Yotsu et al., 1987) Hapalochlaena maculosa (Hwang et al., 1989) Niotha clathrata (Cheng et al., 1995) Bacillus spp. Hapalochlaena maculosa (Hwang et al., 1989) Deep sea sediment (Do et al., 1990) Freshwater sediment (Do et al., 1993) Fugu rubripes (common puffer) (Wu et al., 2005a; Wu et al., 2005b) Alteromonas spp. Hapalochlaena maculosa (Hwang et al., 1989) Deep sea sediment (Do et al., 1990) Aeromonas spp. Deep sea sediment (Do et al., 1990) Niotha clathrata (Cheng et al., 1995) Micrococcus spp. Deep sea sediment (Do et al., 1990) Freshwater sediment (Do et al., 1993) Pseudoalteromonas spp. Meoma ventricosa (sea urchin) (Ritchie et al., 2000) Serratia marcescens Puffer fish species (Yan et al., 2005; Yu et al., 2004) Shewanella putrefaciens Takifugu niphobles (common puffer) (Matsui et al., 1989) Acinetobacter spp. Deep sea sediment (Do et al., 1990) Streptomyces spp. Marine sediment (Do et al., 1991) Caulobacter spp. Freshwater sediment (Do et al., 1993) Flavobacterium spp. Freshwater sediment (Do et al., 1993) Plesiomonas spp. Niotha clathrata (Cheng et al., 1995) Microbacterium Marine Puffer fish (Yu et al., 2004) arabinogalactanolyticum Nocardiopsis dassonvillei Fugu rubripes (Wu et al., 2005a; Wu et al., 2005b) Actinomycete spp. Fugu rubripes (Wu et al., 2005a; Wu et al., 2005b) Marinomonas spp. Nassarius semiplicatus (Wang et al., 2008) Tenacibaculum spp. Nassarius semiplicatus (Wang et al., 2008)

Chapter 2

2.2. Materials and methods

2.2.1. Specimen collection

Hapalochlaena sp. was purchased from Peter Fearnside (Seafish Aquarium Life), who collected the specimen from Moreton Bay, Australia. This specimen was immediately transported to the laboratory in a plastic bag containing seawater

(300 mL). The total time from specimen collection until arrival in the laboratory was approximately 24 h. It was housed in a 10 L aquarium filled with 4 L sterile seawater at 20°C for 48 h prior to dissection.

The P. maculata specimen was collected from the sediment surface in 2-3 m deep water at Narrow Neck Beach, Auckland, New Zealand. It was placed in a plastic bag containing seawater (300 mL) and transported to the laboratory in a thermo- insulated container. The specimen was maintained in a 19 L aquarium filled with

11 L of filtered seawater at 20°C and aerated using a fish tank pump. It was fed twice weekly on Greenshell mussel (Perna canaliculus). After 12 d, P. maculata and its laid egg-mass were removed from the tank prior to dissection.

2.2.2. Specimen dissection

The Hapalochlaena sp. specimen was placed at 4°C for 2 h prior to dissection. The posterior salivary gland, beak and surrounding soft tissue, digestive glands, ovaries and eggs were aseptically dissected from the specimen and sterilised (using 70% ethanol (v/v)). Organs were homogenised in 10 mL sterile Milli-Q water

(Millipore) using a Janke and Kunkel Ultra Turrax (IKA Laboritechnik) for 10 s at maximum amplitude.

Chapter 2

The P. maculata specimen was placed in ice for 2 h and then dissected using a sterile scalpel. The gonad, digestive tract, and reproductive organs were removed and homogenised. The egg mass was washed several times in 70% ethanol (v/v) and homogenised in 5 mL sterile seawater. A sub-sample of each organ and egg homogenate was frozen for subsequent TTX analysis.

2.2.3. Bacterial culturing and genomic DNA extraction

To ensure adequate coverage of unique bacterial isolates, 10-fold serial dilutions of organ homogenates were each spread on Marine (Bacto BD), Vaatanen Nine Salts

Solution (VNSS), Thiosulfate Citrate Bile Salts Sucrose (TCBS) (Acumedia), and

Tryptic Soy Agar (Bacto BD) solid media and incubated at 20°C. Morphologically unique colonies that were observed on each plate were further sub-cultured until pure cultures were obtained. Bacterial isolates were assigned an identifier using a six character code. The first two letters indicate the host organism, either

Hapalochlaena sp. (HM) or P. maculata (PM). The next two characters indicate the organ from which the bacteria was isolated and its isolate number, either posterior salivary gland (SA), beak (BE), digestive glands (LI), ovary and eggs (OE), egg sac

(EG), digestive tract (DT), reproductive tract (RT) or gonads (GO). Genomic DNA was extracted from these isolates using a xanthogenate-SDS protocol as previously described (Tillett and Neilan, 2000). Approximately 100 mg of cells were reuspended in 500 µL XS buffer (1% potassium ethyl xanthogenate, 800 mM ammonium acetate, 100 mM Tris-HCl pH 7.4, 20 mM EDTA, 1% sodium dodecylsulfate) and inverted several times to mix. The tubes were incubated for

120 min at 65°C. The tubes were vortexed for 10 s and incubated for a further

60 min at 65°C. The samples were placed on ice for 10 min, then centrifuged for

Chapter 2

10 min at 12,000 x g to remove cell debris. The resulting supernatant was washed twice with phenol:chloroform:isoamyl alcohol (25:24:1). DNA was then precipitated with 2 volumes absolute ethanol and centrifuged for 30 min at

14,000 x g. The resulting DNA pellet was washed with 70% ethanol, centrifuged for

30 min at 14,000 x g and resuspended in TE buffer (10 mM Tris-HCl, pH 7.4; 1 mM

EDTA, pH 8). Three bacterial isolates from N. semiplicatus were kindly provided by

Rencheng Yu, Institute of Oceanology, Chinese Academy of Sciences (Wang et al.,

2008).

2.2.4. Identification of bacterial isolates

PCR amplification of approximately 1400 bp of the 16S rRNA gene was performed using the bacterial-specific primers 27fl and 1494rc (Neilan et al., 1997). Each PCR reaction mixture contained 2.5 mM MgCl2, 0.15 mM each dNTP, 10 pmol each primer, 0.2 U of BioTaq DNA polymerase, the appropriate PCR buffer (Bioline) and approximately 1 ng DNA. Thermal cycling conditions were as follows: initial denaturation at 94°C for 2 min, 35 cycles of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, extension at 72°C for 1 min, followed by a final extension step at 72°C for 7 min. PCR amplification products were electrophoresed through a 1% agarose gel, visualised by staining with 0.5 μg/mL ethidium bromide, and documented with a Gel Doc XR camera using Quantity One 4.6.1 software

(BioRad). PCR amplicons were precipitated for 15 min using 2 volumes ice-cold

95% ethanol and 1/10 volumes 3M sodium acetate. Precipitated DNA was washed with 70% ethanol and resuspended in Milli-Q water.

Automated sequencing reactions were performed using approximately 50 ng of purified PCR product and 3.2 pmol of primer (26fl/1494rc) using the Prism Big

Chapter 2

Dye cycle-sequencing system and ABI 3730 DNA analyser sequencer (Applied

Biosystem). Isolates with identical nucleotide sequences were considered to be the same and a single representative was selected for downstream analysis. Closest

16S rRNA gene homologies were identified using the BLAST search program

(NCBI).

2.2.5. PCR screening of bacterial isolates for biosynthesis genes

All isolates were screened for the presence of PKS genes using the degenerate primer set DKF/DKR (Table 2.3) (Moffitt and Neilan, 2001), NRPS genes using the degenerate primer set MTF2/MTR2 (Neilan et al., 1999) and AMT genes using the degenerate primer set ATFI/ATRI (Kellmann, 2005). Expected sizes for PCR amplification products were 700 bp for PKS and AMT, and 1000 bp for NRPS products. PCR reactions were carried out as described above, except 25 pmol of each primer was used. Thermal cycling conditions were as followed: initial denaturation at 92°C for 2 min, 35 cycles of denaturation at 92°C for 10 s, annealing as outlined in Table 5 for 30 s, extension at 72°C for 1 min, followed by a final extension step at 72°C for 7 min. Purification of PCR amplification products,

DNA sequencing and gene identification using the BLAST server was performed as described above.

PCR screening of Hapalochlaena sp. isolate HM-SA03 revealed multiple NRPS and

PKS genes, hence cloning of these genes was performed. PKS and NRPS PCR amplification products were gel-purified and ligated into pCR2.1-TOPO vector using a TOPO-TA Cloning kit following protocols provided by the manufacturer

(Invitrogen) and transformed into chemically competent Escherichia coli DH5α.

Ten clones from each clone library were randomly selected and PCR amplified

Chapter 2

using vector-directed primers (MpF and MpR). Purification, sequencing and analysis of PCR products were performed as described above.

Table 2.3 Primers used for identification and screening of bacterial isolates

Primer Sequence Tma ATb Reference 27fl AGAGTTTGATCCTGGCTCAG 61 55 (Neilan et al., 1997) 1494rc TACGGCTACCTTGTTACGAC 59 55 (Neilan et al., 1997) DKF GTGCCGGTNCCRTGNGYYTC 67 55 (Moffitt and Neilan, 2001) DKR GCGATGGAYCCNCARCARMG 65 55 (Moffitt and Neilan, 2001) MTF2 GCNGGYGGYGCNTAYGTNCC 64 52 (Neilan et al., 1999) MTR2 CCNCGDATYTTNACYTG 47 52 (Neilan et al., 1999) ATfwd1 GTVTGYCCWMGSGAYGTVATG 57 55 (Kellmann, 2005) ATrev1 ATRTCCCAWRTBCRCARTG 62 55 (Kellmann, 2005) aTm is the theoretical melting temperature of the PCR primers, given in °C. bAT is the annealing temperature used in PCRs containing these primers, given in °C.

2.2.6. Phylogenetic analysis of 16S rRNA gene sequences

Partial sequences of 16S rRNA genes obtained from this study, as well as 42 reference sequences obtained from NCBI Genbank, were aligned using Muscle

(Edgar, 2004). A phylogenetic tree was constructed from a 539 bp alignment with maximum likelihood algorithm, PhyML (Guindon and Gascuel, 2003), using the general time reversible with gamma distribution (GTR+G) base substitution model determined using the web-based application, FindModel

(http://www.hiv.lanl.gov/content/sequence/findmodel/findmodel.html). The tree was rooted with Synechocystis sp. PCC 6803 (Genbank accession number

AY224195.1) as an outgroup. The aLRT SH-like approach was used as an estimate of branch reliability (Guindon et al., 2010).

Chapter 2

2.2.7. Toxin extraction from organ homogenates and bacterial isolates

Sub-samples (1 g, wet weight was used where the total mass exceeded this amount) of each organ homogenate were extracted with 3 volumes (approximately

3 mL) of HPLC-grade methanol (0.1% v/v acetic acid) and centrifuged (3000 x g,

10 min) to remove cell debris.

Cultured organisms that possessed PKS genes, as determined by PCR amplification, were grown in 1 L volumes, as were bacteria isolated from P. maculata eggs, to observe whether TTX was produced in culture. The cultures were grown aerobically at 20°C for 3 d, with shaking in peptone (0.5%; Oxoid) in filtered seawater. This was used as a culture media as it has been shown to promote the production of TTX in bacteria (Gallacher and Birkbeck, 1993). Cultures were centrifuged (3000 x g , 10 min) and the supernatant removed. Cell pellets were then extracted as described for organ homogenates. N. semiplicatus bacterial isolates were grown and chemically extracted using conditions described previously (Wang et al., 2008).

2.2.8. Analysis of extracts by Liquid Chromatography-Mass Spectrometry

To determine whether TTX was present in methanol extracts prepared from the organs and associated microbes of P. maculata, LC-MS was performed as previously described (McNabb et al., 2010). Extracts of Hapalochlaena sp. and its associated bacterial isolates were analyzed by LC-MS using a Thermo Finnigan

Surveyor HPLC and autosampler equipped with a Thermo Finnigan LCQ Deca XP

Plus ﬁtted with an electrospray source. Separation of analytes was obtained on a

Phenomenex Luna 3 μm C18 column (2.1 mm x 150 mm) at a flow rate of

Chapter 2

100 μL/min. After an initial period of 10 min using a solvent gradient of 95% water

(10 mM heptafluorobutyric acid), the mixture was ramped to 100% acetonitrile over 30 min.

To determine if the bacterial pellet matrix resulted in any suppression or enhancement of the TTX signal during LC-MS analysis a spiked recovery experiment was undertaken. Samples were spiked in duplicate with pure TTX

(Tocris Bioscience, Cat. No: 1078) to give final concentrations of 1 ng/mL,

2 ng/mL, 10 ng/mL and 100 ng/mL. No matrix effects were measured.

2.2.9. Antimicrobial bioassay of HM-SA03 crude extracts

Liquid cultures of Pseudoalteromonas sp. HM-SA03 were grown as described in section 2.2.7. Compounds in the bacterial supernatant were adsorbed with 20 g/L

XAD-7, shaking at 250 rpm for 3 h. The chromatographic resin was filtered and washed twice with 3 volumes of milliQ water to remove any media constituents.

Adsorbed compounds were eluted twice with 3 volumes of HPLC-grade methanol and the combined fractions were dried by rotary evaporation. Bacterial cell pellets were extracted twice with 2 volumes of HPLC-grade methanol and the combined fractions were dried by rotary evaporation. All crude extracts were resuspended in

DMSO for antimicrobial bioassays.

Staphylococcus aureus NCTC 6571 was used to test the antimicrobial activity of

HM-SA03 crude extracts. Prior to testing, the bacterial test strains were grown in nutrient broth (Sigma) at 30°C. When cultures reached the exponential phase of growth, as indicated by an optical density of 0.6 at 600 nm (equivalent to 1x107 cells/mL by McFarland standards), the cells were collected by centrifugation at

Chapter 2

5000 x g for 5 min and resuspended in half their original volume. Fifty microlitre aliquots of prepared cells (equivalent to 5x105 cells) were seeded into a sterile flat- bottomed 96-well microtitre plate and diluted with an equal volume of the prepared extract. The two final concentrations of the extracts in the wells were

100 µg/mL and 10 µg/mL respectively. Bacterial microtitre plates were incubated for 24 h, at the appropriate temperature. At the conclusion of the incubation period, the optical density was read at 600 nm using a SpectraMax 340 plate reader (Molecular Devices). Experiments were performed in triplicate with controls included in every plate. The controls included media only, cells treated with 0.2% DMSO and ampicillin (10 µg/mL; Sigma-Aldrich).

The inhibition of cell proliferation, as a percentage of the untreated control, was calculated using the formula:

% Inhibition = [1 - A 600(sample) / A600(control)] * 100

Where % Inhibition is the proliferation of test cells, as a percentage of the proliferation of untreated control cells, A600(sample) = absorbance of test sample,

A600(control) = absorbance of untreated control sample.

Chapter 2

2.3. Results and discussion

2.3.1. Bacterial diversity in toxic Hapalochlaena sp. and Pleurobranchaea maculata

A total of 27 and 22 unique bacterial isolates were cultured from Hapalochlaena sp. and P. maculata, respectively. In order to classify these isolates, approximately

1400 bp of the 16S rRNA gene was successfully amplified and sequenced from each isolate (GenBank accession numbers JN618116 to JN618164) and analysed using the NCBI BLASTn algorithm. This revealed that Alteromonadales were predominant, with 30 of 49 isolates accounted for by this taxonomic order (Tables

2.4 and 2.5). Pseudoalteromonas and Alteromonas were the major representatives of Alteromonadales in both host organisms.

Phylogenetic analyses revealed that 16S rRNA gene sequences grouped into seven distinct clades representing the orders Rhodobacterales, Oceanospirillales,

Alteromonadales, Vibrionales and Bacillales (Figure 2.2). In general, Hapalochlaena sp. sequences grouped separately from P. maculata sequences, indicating distinct microbial communities in these two animals. One exception to this was the P. maculata isolate PM EG11 that grouped phylogenetically with the Alteromonas sp. from Hapalochlaena sp., HM BE02 and HM OE03.

Alteromonadales isolates grouped into three separate clades and more isolates from the Alteromonadales III clade tested positive for biosynthesis genes than in all other clades of isolates. This clade includes the previously reported TTX-producer

Pseudoalteromonas tetraodonis IAM 14160 (Gallacher and Birkbeck, 1993). None

Chapter 2

of the isolates from the Vibrionales or Bacillales clades possessed complex (NRPS or PKS) biosynthesis genes, as determined by PCR.

It is well known that specific environmental conditions are critical for the biosynthesis of natural products (Clardy and Walsh, 2004; Pettit, 2009) such as

TTX (Gallacher and Birkbeck, 1993), as are the synergistic contributions by multiple organisms that lead to substantial increases in metabolite production

(Kurosawa et al., 2008; Oh et al., 2007). For these reasons, it is important that the resident microbial communities of TTX-containing organisms are classified and converted to stable laboratory cultures. Microorganisms are the proposed biogenic source of TTX, however, there have been very few studies describing the diversity of bacteria in hosts known to contain TTX. In the past, there have been numerous descriptions of culturable TTX-producing bacteria isolated from TTX-containing organisms (Chau et al., 2011), yet none of these studies reported the entire culturable microbial community. As a result the bacterial diversity in these toxic marine animals has remained largely unknown. In trying to identify a microbial producer of TTX, it is important to understand the entire microbial diversity found within host organisms as it defines the environment where TTX-biosynthesis can occur. This study represents the first study to report the culturable bacterial populations of toxic Hapalochlaena sp. and P. maculata.

The isolation of bacteria from Hapalochlaena sp. and P. maculata belonging to seven bacterial genera significantly builds upon the previously published studies investigating TTX-producing bacteria from toxic hosts, the majority of which only describe a single species (Chau et al., 2011). Interestingly, two species of mollusc,

Niotha clathrata and Nassarius semiplicatus were reported (Cheng et al., 1995;

Chapter 2

Wang et al., 2008) to possess highly diverse microbial communities, with five bacterial genera identified in each. Vibrio, Tenacibaculum, Marinomonas,

Shewanella and Aeromonas spp. were identified in N. semiplicatus, while Vibrio,

Pseudomonas, Pasteurella, Plesiomonas and Aeromonas spp. were found in N. clathrata. This is comparable to the diversity in the molluscs used in this study, and suggests that they may have greater bacterial diversity than other TTX- associated hosts.

The predominance of Pseudoalteromonas in both Hapalochlaena sp. and P. maculata suggested that this microbial genus may play a key role in the physiology of their hosts and possibly TTX biosynthesis. To our knowledge, this is the first description of any Pseudoalteromonas spp. from molluscs. Furthermore, reports of organ-associated bacteria from molluscs have only described single microorganism associations (Ruby, 2008) rather than diverse communities as detailed here.

To date, microbial TTX-producers have been reported from 23 bacterial genera

(Campbell et al., 2009; Chau et al., 2011; Simidu et al., 1990; Wang et al., 2010;

Yang et al., 2010). Of these, Vibrio spp. have been identified as the TTX-producer in over half the toxic host specimens. The predominance of Vibrio spp. is likely due to their abundance in nature and their inherent ability to be cultured from the marine environment. We identified three isolates as Vibrio spp. from the two animal hosts.

Two of these were isolated from Hapalochlaena sp. and one from and P. maculata, none of which tested positive for TTX (refer to section 2.2). Unfortunately, useful phylogenetic comparisons were unable to be performed between isolates from this study and those from previous studies of TTX-associated symbionts. Of the 25

Chapter 2

studies on TTX production by bacteria, only four of these utilised molecular taxonomic methods as part of their bacterial identification procedure. Only two of these studies (Ritchie et al., 2000; Wang et al., 2008) have submitted their data to public databases, hence comparison of 16S rRNA gene sequences from these Vibrio strains was not possible.

We have shown here that both animals host a diverse, culturable, microbial community. Studies regarding TTX production by bacteria, including the present study, have only focused on pure cultures. When TTX-producing bacteria have been identified, the strains have yielded only very low amounts of TTX (Noguchi and Arakawa, 2008; Wang et al., 2008) which could not account for the levels of toxin present in the host organisms.

Chapter 2

ID PCR TTX Strain Closest 16S rRNA gene BLAST matcha (%) screening detectionb PKS NRPS AMT HM BE02 Alteromonadales bacterium HOT3G5 (HQ537362.1) 99 - - - - HM BE03 Alteromonas sp. S2542 (FJ457277.1) 99 - - - - HM BE04 Alteromonas sp. S2542 (FJ457277.1) 99 - - - - HM BE05 Pseudoalteromonas sp. AmSamW21 (GU903211.1) 99 + - - - HM BE06 Vibrio rotiferianus isolate AP17 (HE584775.1) 99 - - - - HM LI01 Nautella italica strain R-28753 (AM944522.1) 99 - - + NT HM LI02 Alteromonas sp. S2542 (FJ457277.1) 99 - - - - HM LI03 Pseudoalteromonadaceae S3 (HQ164448.1) 100 - - - NT HM LI04 Nautella italica strain R-28753 (AM944522.1) 99 - - - NT HM LI05 Alteromonas sp. S2542 (FJ457277.1) 99 - - - NT HM LI06 Pseudoalteromonas sp. AKA07-4 (AB571944.1) 99 - - - NT HM OE02 Pseudoalteromonadaceae S3 (HQ164448.1) 100 - - + NT HM OE03 Alteromonas sp. CF14-3 (FJ170033.1) 99 - - - - HM OE04 Thalassomonas sp. PaD1.04 (GQ391976.1) 98 - - - - HM OE07 Pseudoalteromonas sp. TB51 (JF273853.1) 99 - - - - HM OE08 Colwellia sp. KMD002 (EU599214.3) 98 - - - NT HM OE09 Pseudoalteromonadaceae S3 (HQ164448.1) 99 + - + NT HM SA02 Pseudoalteromonas sp. AKA07-4 (AB571944.1) 99 + + + - HM SA03 Pseudoalteromonadaceae S3 (HQ164448.1) 99 + + + - HM SA04 Pseudoalteromonas sp. NW4327 str. NW (FR839670.1) 99 + - - NT HM SA05 Marinomonas communis LMG 2864 (DQ011528.1) 99 - - - NT HM SA06 Alteromonas sp. KB19 (HM583350.1) 99 - - - - aNCBI Genbank accession numbers are indicated in brackets bNT indicates that the sample was not tested for its ability to produce TTX as they were unable to be revived from cryogenic stocks. The detection limit of TTX using the LC-MS method was 0.1 ng/mL.

Chapter 2

ID PCR TTX Strain Closest 16S rRNA gene BLAST matcha (%) screening detectionb PKS NRPS AMT PM DT05 Photobacterium kishitanii strain S-27 (JF412253.1) 99 - - - - PM DT08 Shewanella schlegeliana (AB081761.1) 99 - - - - PM DT10 Photobacterium sp. DFC2.17 (FR873783.1) 99 - - - - PM DT11 Photobacterium kishitanii strain S-27 (JF412253.1) 99 - - - NT PM DT12 Shewanella schlegeliana (AB081761.1) 99 + - - - PM EG02 Vibrio sp. YDO5 (GU586127.1) 98 - - - - PM EG04 Staphylococcus warneri strain Na58 (HQ831387.1) 99 - - - - PM EG05 Staphylococcus warneri strain Na59 (HQ831388.1) 99 - - - - PM EG08 Pseudoalteromonas sp. 114Z-11 (GU584139.1) 99 - - - - PM EG09 Pseudoalteromonas sp. BSs20138 (EU365489.1) 99 - - - - PM EG11 Alteromonas sp. N98(2010) (HQ188650.1) 99 - + - - PM EG13 Pseudoalteromonas sp. LJ1 (FJ665500.1) 99 - - - - PM EG14 Halomonas sp. Pper-Hx-1972 (EU123940.1) 99 + + - - PM EG15 Vibrio sp. S4639 (FJ457601.1) 99 - - - - PM EG17 Pseudoalteromonas sp. 114Z-11 (GU584139.1) 98 - - - - PM EG18 Pseudoalteromonas porphyrae str. HK1 (FJ205736.1) 99 + + - - PM GO01 Shewanella schlegeliana (AB081761.1) 99 + - - - PM GO04 Vibrio splendidus isolate PB1-10rrnM (EU091337.1) 99 - - - - PM GO05 Photobacterium kishitanii calba.5.9 (AY642170.1) 99 - - - - PM GO06 Shewanella schlegeliana (AB081761.1) 99 + - - - PM GO08 Photobacterium kishitanii calba.5.9 (AY642170.1) 99 - - - - PM GO09 Vibrio fischeri strain SI1E (AY292949.1) 99 - - - NT PM GO12 Photobacterium kishitanii calba.5.9 (AY642170.1) 99 - - - NT PM GO13 Vibrio fischeri strain SI1Ecomplete (AY292949.1) 99 - - - NT PM RT05 Pseudoalteromonas sp. LJ1 (FJ665500.1) 99 - + - - PM RT07 Alteromonas sp. N98(2010) (HQ188650.1) 99 - + - - PM RT10 Vibrio parahaemolyticus strain NMGB2 (JN561593.1) 99 - - - NT aNCBI Genbank accession numbers are indicated in brackets bNT indicates that the sample was not tested for its ability to produce TTX as they were unable to be revived from cryogenic stocks. The detection limit of TTX using the LC-MS method was 0.1 ng/mL.

Chapter 2

Figure 2.2 Maximum likelihood phylogenetic tree of 16S rRNA gene sequences of isolates from this study and related bacteria.The Synechocystis sp. PCC 6803 16S rRNA gene sequence was used as an outgroup. Hapalochlaena sp. sequences (prefix, HM) are indicated with blue text, P. maculata sequences (prefix, PM) are indicated with red text. All sequences were submitted to the GenBank database (accession numbers JN618116 to JN618164). Isolates identified as possessing PKS, NRPS or AMT genes are also indicated.

Chapter 2

2.3.2. Overestimation of TTX-producing bacteria in the literature

TTX was not detected by liquid chromatography-mass spectrometry (LC-MS) in any of the Hapalochlaena sp. organ extracts. The conditions the specimen was kept in prior to its dissection may not have been conducive for toxin production. Many variables are present in the natural environment (e.g. nutrient concentrations, environmental stress) that cannot be accounted for in this study. These factors may have influenced the production of TTX by the octopus. A non-toxic species of

Hapalochlaena has not been reported in the literature. However, a microbial TTX- producer would still contain the genes required for its biosynthesis. Therefore, molecular screening of these microbes would reveal the presence of putative TTX- biosynthesis genes. TTX was present in the P. maculata samples with a uniform distribution throughout those organs tested (Table 2.7). The P. maculata egg- derived extracts were also found to contain TTX (Table 2.7). TTX or its analogues were not detected via LC-MS in any of the bacteria-derived methanol extracts tested (Tables 2.4 and 2.5). Bacteria-derived methanol extracts from N. semiplicatus isolates were not found to contain TTX or related analogues via LC-MS

(Table 2.6). This is contrary to a previous study (Wang et al., 2008), which reported these strains as TTX producers, via immunoassay approaches.

Chapter 2

Table 2.6 Screening of Nassarius semiplicatus bacterial isolates for putative TTX- biosynthesis genes and a comparison of ELISA and LC-MS methods for the detection of TTX Straina PCR screening TTX conc. TTX conc. PKS NRPS AMT ELISA (ng/g)b (LC-MS) Vibrio sp. 34HU9 (EU268265) + + - 169 BDLc Marinomonas sp. 38JIA1 (EU268259) - + - 98 BDL Tenacibaculum sp. 30ORI8 (EU268261) - + - 54 BDL a Strains are as described in Wang, X., et al., 2008 (Wang et al., 2008). b ELISA results shown are those previously reported (Wang et al., 2008). The detection limit of TTX using the ELISA method was 5 ng/mL (equivalent to 15ng/g of TTX in tissue). c BDL indicates that the concentration of TTX was below the detectable limit of the instrumentation. The detection limit of TTX using the LC-MS method was 0.1 ng/mL (equivalent to 0.3 ng/g of TTX in tissue).

Table 2.7 Detection of tetrodotoxin (TTX) in whole-organ homogenates

Host Organism Tissue tested TTX concentration (mg/kg) Hapalochlaena sp. Posterior salivary gland BDLa Beak BDL Pleurobranchaea maculata Eggs 5.19 Digestive tract 4.23 Reproductive tract 3.84 Gonads 3.54 a BDL indicates that the concentration of TTX was below the detectable limit of the instrumentation. The detection limit of TTX using the LC-MS method was 0.1 ng/mL (equivalent to 3x10-4 mg/kg of TTX in tissue).

Chapter 2

The failure to identify and culture a TTX-producing microbe in this study (Tables

2.4, 2.5) despite numerous reports of isolations from many other TTX-associated hosts was not completely unexpected. Captivity of the P. maculata specimen for 12 d prior to dissection and bacterial isolation may have influenced the microbiota present in our isolation process. Additionally, P. maculata have been shown to have high variability in TTX concentrations within individuals (Wood et al., 2012a) suggesting an exogenous source of TTX. Whether this is due to environmental factors or variations in the abundance of bacterial producers is unknown.

However, changes in environmental conditions or bacterial diversification may lead to loss of TTX-producing capability. Such a loss of toxin production capability has been reported in other toxin-producers, including the cyanobacteria

Cylindrospermopsis raciborskii, which produces cylindrospermopsin (Saker and

Griffiths, 2000). These factors may have contributed to the lack of TTX-production by bacteria in this study. Furthermore, it is possible that an unculturable bacterium may be responsible for the production of TTX in these organisms.

Many previous studies have focused exclusively on the detection of TTX in bacterial cells (Cheng et al., 1995; Wang et al., 2008; Wu et al., 2005a; Yang et al.,

2010). Similarly, this study did not investigate the presence of TTX in bacterial culture supernatants. The presence of TTX at high levels in the supernatant, however, would have been represented by the presence of TTX in bacterial cells.

In the many instances where chemical defences are conferred to progeny, the amount of toxin invested into the eggs or larvae is much lower than that of the adult (Williams et al., 2011). However, in this and previous studies (McNabb et al.,

2010; Wood et al., 2012b) the level of TTX detected in P. maculata eggs was

Chapter 2

comparable to that found in the adult. It has also been shown that P. maculata larvae possess concentrations of TTX equivalent to that of the adult (McNabb et al.,

2010). However, a lack of TTX in the P. maculata egg-derived bacterial isolates indicates that vertical transmission of TTX-producing bacteria is unlikely in P. maculata (Chaston and Goodrich-Blair, 2010).

Reported TTX-producers include species from Vibrio, Pseudomonas, Alteromonas,

Marinomonas, and Shewanella genera, and bacteria of the aforementioned genera have been previously isolated from H. maculosa (Hwang et al., 1989). Many studies describing the isolation of TTX-producing bacteria from host animals based their findings on the detection of TTX-like activity using non-specific mouse bioassays, or chemically by GC-MS (Cheng et al., 1995; Do et al., 1990; Hwang et al., 1989).

Ions characteristic of TTX observed from GC-MS analyses have also been attributed to peptone, a common growth media constituent (Matsumura, 1995).

Furthermore, the majority of reports that identify TTX in bacteria were not supported by structure characterisation using NMR or MS/MS of a purified compound. This is highlighted here, whereby N. semiplicatus isolates that had previously tested positive for TTX using ELISA (Wang et al., 2008) were not found to produce TTX using more sensitive spectrometric methods. The N. semiplicatus specimen from which the bacteria were originally isolated had tested positive for

TTX using LC-MS (Wang et al., 2008) indicating that the bacteria may not be the true producers of TTX in this organism. Taken together, these observations may have contributed to an over-estimation of the number of bacterial TTX producers in the literature.

Chapter 2

2.3.3. Mining host-associated bacteria for proposed TTX biosynthesis genes

Seven isolates from Hapalochlaena sp. were found to contain at least one of the targeted biosynthesis genes (Table 2.4). Two Pseudoalteromonas isolates, HM SA2 and HM SA3, were positive for PKS, NRPS and AMT genes and one

Pseudoalteromonas isolate, HM OE9, contained both PKS and AMT genes.

Furthermore, three Pseudoalteromonas isolates contained either a PKS or NRPS genes.

DNA sequencing of PKS and NRPS clone libraries of the Pseudoalteromonas isolate

HM-SA03 revealed four unique PKS ketosynthase domains and six unique NRPS adenylation domains. Many of these domains showed little amino acid similarity

(less than 60% identity) to characterised PKS and NRPS domains and are therefore proposed to be novel.

A total of eight P. maculata isolates contained at least one of the targeted biosynthesis gene types (Table 2.5). These genes had similarities (70-97%) to known sequences in the NCBI database, as determined by BLAST analysis. Three

Shewanella isolates contained only PKS genes, while two Alteromonas isolates and a Pseudoalteromonas isolate contained only NRPS genes. Single Halomonas and

Pseuoalteromonas strains contained both PKS and NRPS genes. AMT genes were not PCR amplified from any of the P. maculata isolates.

All three N. semiplicatus isolates contained NRPS genes, however, only the Vibrio isolate, HU9, contained PKS genes and none of the N. semiplicatus isolates screened contained AMT genes. It is unlikely that TTX is biosynthesised via an NRPS alone

(Chau et al., 2011). Hence, it is possible that these strains are not true producers of

Chapter 2

TTX, but may produce a structurally similar compound that may test positive with antibody-based assays and provide a false positive result.

Sequence analysis of AMT PCR products revealed significant similarity to AMT (e.g. cyrA) from toxic Cylindrospermopsis species (Mihali et al., 2008). This is significant because CyrA is involved in the biosynthesis of the potent toxin, cylindrospermopsin (Figure 2.1), which also incorporates an arginine as is proposed for TTX. An alternate biosynthetic route to TTX, which does not require an AMT, but rather incorporates an intact arginine has been proposed (Kotaki and

Shimizu, 1993; Shimizu and Kobayashi, 1983; Yasumoto et al., 1988). However, feeding studies with guanido-14C labeled arginine could not confirm the origin of the guanidine moiety in TTX (Shimizu and Kobayashi, 1983), and thus incorporation of intact arginine via an NRPS cannot be discounted. Gene sequences derived from NRPS and PKS PCR were indicative of adenylation and ketosynthase domains, respectively, but they did not show significant similarity to any characterised gene clusters and thus cannot be directly linked to the biosynthesis of any known toxin. Therefore, although NRPS and PKS are predicted to be involved in TTX biosynthesis, we could not confirm the identification of putative

TTX biosynthesis genes in this study. These findings suggested that AMT are indeed useful targets for PCR-directed screening of microorganisms for toxic alkaloid biosynthesis gene clusters due to their rarity in microorganisms compared to NRPS and PKS-coding genes.

Two isolates from Hapalochlaena sp., HM-SA02 and HM-SA03, derived from the octopus’ salivary glands, and identified as Pseudoalteromonas spp., tested positive for all three types of biosynthesis genes. Sequencing of HM-SA03-derived clone

Chapter 2

libraries containing amplified PKS and NRPS gene fragments, revealed unusual genetic diversity, indicative of the potential to biosynthesise small and possibly novel molecules. Additionally, crude chemical extracts derived from the HM-SA03 isolate showed moderate inhibitory effects against Staphylococcus aureus (Figure

2.3). Though we did not detect TTX or its analogues in the extracts, the isolate is of interest from a small molecule biosynthesis standpoint. Genome sequencing in the future may allow us to ascertain whether this organism possesses a candidate gene cluster for the production of TTX.

8 0 1 0 0 µ g /m l 1 0 µ g /m l 6 0

4 0

Inhibition20 (%)

) (B ic illin ) a c t (A ) ra c t

e x t 3 e x tr tro l (a m p -S A 3 M-SA C o n H HM

Figure 2.3 Inhibitory effect of HM-SA03 methanol-derived crude extracts. Both biomass (A) and supernatant-derived (B) extracts were inhibitory against Staphylococcus aureus . Assay was normalised to the DMSO control (0% inhibition).

Chapter 2

2.4. Concluding remarks

This study has provided a better understanding of the microbial diversity in two

TTX-associated molluscs and revealed the potential for the production of PKS and

NRPS-derived natural products from their isolated bacteria. We set out to isolate culturable microorganisms with the aim of positively identifying a TTX producer.

Despite our efforts we are yet to identify such a microorganism although a number of candidates have been selected for further investigation of their natural products.

It is plausible that the TTX producing organism may not be easily cultured under standard laboratory conditions, as is the case with the majority of all bacteria

(Handelsman, 2004). It is also possible that we may well have isolated a bacterium with the genetic potential to biosynthesise TTX but production was not achievable ex situ. Biosynthetic precedents of structurally similar compounds suggest an NRPS or PKS origin for TTX, however, there is no experimental evidence to support this.

In addition, the highly unusual structure of TTX may be the product of an unusual pathway utilising novel enzymes. Characterisation of the culturable microbial communities of Hapalochlaena sp. and P. maculata indicated a greater than anticipated level of diversity. It is also reasonable to presume that an unculturable, or yet-to-be cultured, TTX-producer exists. To this end, the following chapter

(chapter 3) will describe a molecular approach to study the entire microbial communities of these molluscs. This will not only extend our knowledge of their microbial communities, but also provide us with clues as to how culture members of these communities in the laboratory, aiding the search for the elusive TTX producer and its biosynthetic pathway.

Chapter 3

Chapter 3: Bacteria from Rhodobacteraeceae correlate with the presence of tetrodotoxin in the grey side-gilled sea slug

Pleurobranchaea maculata

3.1. Introduction

TTX is a highly potent neurotoxin produced by phylogenetically diverse organisms in both aquatic and terrestrial environments. Although TTX has been known and studied for over 50 years, its biogenic origin is still a topic of much controversy

(Chau et al., 2011). The endogenous theory proposes that TTX-containing organisms have the ability to produce TTX independently, while the exogenous theory proposes that bacteria associated with TTX-producing marine organisms are responsible for TTX production.

Support for an exogenous origin of TTX in Takifugu rubripes and P. maculata are based on studies that have shown a decrease in TTX concentration over time when these organisms are raised in captivity (McNabb et al., 2010; Noguchi et al., 2006a).

In P. maculata, TTX has been shown to depurate over time when cultured under controlled, laboratory conditions and fed with non-TTX containing diets (Wood et al., 2012a). This suggested an exogenous source of TTX, either dietary or through symbiotic bacteria. A bacterial origin of TTX in P. maculata has yet to be proven, because attempts at isolating and identifying TTX-producing bacteria from P. maculata have not been successful (Chau et al., 2013).

Molecular studies have been used to track changes in cyanobacterial diversity during toxic blooms and to correlate cyanobacterial abundance with changes in

Chapter 3

toxin concentrations within water systems (Te and Gin, 2011). Such approaches are able to overcome difficulties in culturing bacteria and can identify the majority of microbiological content within a sample. In an effort to elucidate the putative biogenic origin of TTX in P. maculata, we employed a ribosomal-tagged pyroseqeuncing approach to investigate the bacterial community associated with

P. maculata.

Bacterial 16S ribosomal RNA (rRNA)-based 454 pyrosequencing was performed on toxic and non-toxic whole-P. maculata homogenates. Tetrodotoxin concentrations for each sample were determined using liquid chromatography- mass spectrometry (LC-MS). The 16S rRNA gene abundance and TTX data were analysed using statistical methods to identify whether the total bacterial community as a whole, or individual microbial taxa were correlated to the presence and concentration of TTX in P. maculata.

3.2. Materials and methods

3.2.1. Pleurobranchaea maculata specimen collection

Tetrodotoxin-containing P. maculata adults (n=6) and egg masses (n=2) were collected from Auckland, New Zealand, on 16 December 2010, while non-toxic P. maculata adults (n=3) and egg masses (n=2) were collected from Nelson, New

Zealand on 5 November 2010. Specimens were placed in plastic bags containing seawater (300 mL) and transported to the laboratory at the Cawthron Institute

(Nelson, New Zealand) in a thermo-insulated container. All samples except a subset (n=3) of toxic P. maculata adults were frozen (-20°C) in RNALater®

(Amicon) upon arrival at the laboratory. The unpreserved subset of P. maculata

Chapter 3

was raised in a 19 L aquarium filled with 11 L of filtered seawater, aerated using a fish tank pump. Specimens were fed twice weekly on Greenshell™ mussel

(Perna canaliculus). These samples were raised for 141 to 166 d until death, upon which they were frozen as described above. All frozen samples were transported to the University of New South Wales (Sydney, Australia) where they were thawed and rinsed in 70% ethanol to remove surface-associated bacteria. Each sample was homogenised whole and frozen until further processing.

3.2.2. Tetrodotoxin extraction and analysis

A subsample (2 g) of each homogenised toxic P. maculata sample was extracted with 18 mL of Milli-Q water, containing 0.1% (v/v) acetic acid. Each sample was homogenised (1 min for adult samples or 30 s for egg masses; Ultra-turrax T8, KA-

Werke, China). Samples were centrifuged (3000 x g, 10 min) and an aliquot of the supernatant (1 mL) added to 9 mL of methanol, containing 0.1% (v/v) acetic acid, and frozen (-20°C) for at least 1 h. Samples were centrifuged (3000 x g, 10 min) and diluted 1:200 with methanol containing 0.1% (v/v) acetic acid. Samples were analysed for TTX using LC-MS as described by McNabb, et al. (McNabb et al., 2010).

The method was calibrated using pure TTX (Tocris Bioscience, Cat. No: 1078).

3.2.3. DNA extraction

DNA was extracted from P. maculata homogenates using a xanthogenate-SDS protocol as previously described in section 2.2.3 (Tillett and Neilan, 2000).

Approximately 100 mg of organ homogenate was resuspended in 500 µL XS buffer

(1% potassium ethyl xanthogenate, 800 mM ammonium acetate, 100 mM Tris-HCl pH 7.4, 20 mM EDTA, 1% sodium dodecylsulfate) and inverted several times to

Chapter 3

mix. The tubes were incubated for 120 min at 65°C, vortexed (10 s) and incubated

(60 min at 65°C). The samples were placed on ice (10 min), and then centrifuged

(12000 x g, 10 min), to remove cellular debris. The resulting supernatant was washed twice with phenol:chloroform:isoamyl alcohol (25:24:1). DNA was then precipitated with 2 volumes of absolute ethanol and pelleted by centrifuged

(14000 x g, 10 min). The resulting DNA pellet was washed with 70% ethanol, centrifuged (14000 x g, 10 min) and resuspended in TE buffer (10 mM Tris-HCl, pH

7.4; 1 mM EDTA, pH 8 .0). PCR amplification of partial 16S rRNA genes was performed using primers 27fl and 1049rc, as previously described (Neilan et al.,

1997), to check for the presence of polymerase-inhibitors. DNA samples that failed to PCR amplify were further purified using the Soil DNA Extraction kit (MP

Biosciences) according to manufacturer’s specifications.

3.2.4. Sequencing and quality control

Bacterial tag-encoded FLX amplicon pyrosequencing was carried out at the

Research and Testing Laboratory (Lubbock, TX). Primers 28F and 519R (Lane,

1991) were used to PCR amplify sequences spanning the V2-V3 variable regions of the bacterial 16S rRNA gene. Sequencing was carried out on a Roche 454 FLX instrument with Titanium reagents and procedures. Generated sequences were quality filtered using Mother v.1.23.1 software (Schloss et al., 2009). Sequences containing more than one mismatch in the sequencing barcode, two mismatches in the 5’ primer region, ambiguous nucleotides, or homopolymers greater than eight base pairs (bp) were discarded. Filtered sequences were aligned to the SILVA- based reference alignment of 16S rRNA genes (Schloss, 2009) and trimmed to ensure all reads and columns overlapped within the alignment space. Chimeric

Chapter 3

sequences were removed using the UCHIME (Edgar et al., 2011) wrapper in

Mothur, using a SILVA gold reference alignment. To further reduce the effect of sequencing error, pre-clustering of sequences was performed, allowing for two mismatches per sequence. Finally, a subset of 556 sequences, representing the size of the smallest dataset, was extracted from each sample for downstream analysis.

3.2.5. Statistical analyses

Non-metric multidimensional scaling (nMDS) was used to visualise the Bray-Curtis similarities of bacterial 16S rRNA gene sequences present in the P. maculata samples, based on presence and abundance of OTUs (defined at 97% similarity.

Analysis of similarity (ANOSIM) was utilised to assess the significance of observed difference between samples in the nMDS plot. PRIMER v6 software (Clarke and

Gorley, 2006) was used to perform a non-parametric RELATE test to determine the statistical significance of correlation between bacterial community composition in toxic wild and captive P. maculata and their TTX concentrations. Pairwise associations between each OTU and TTX concentration in P. maculata were represented using the Spearman correlation coefficient (ρ), calculated using the R software package (R Core Team, 2013).

3.3. Results and discussion

3.3.1. Amplicon sequencing and processing

Amplicon sequencing of 16S rRNA genes from the four P. maculata egg masses and nine adults produced 34,408 high-quality, total sequences with an average sequence length of 341 bp. Subsampling 556 sequences from each sample, representing the number of sequences from the smallest sample, resulted in 7228

Chapter 3

sequences. Clustering of sequences at 97% identity produced 1274 OTUs.

Rarefaction analysis indicated that a sampling depth of one new OTU per 100 sequences analysed was not reached for any of the samples in this study (Figure

3.1).

5 0 0 T E 1 T E 2 4 0 0 N E 1 N E 2 3 0 0 N S 1 N S 2 2 0 0 N S 3

N o . O T U s T S 1 1 0 0 T S 2 T S 3

0 T T 1 0 2 0 0 0 4 0 0 0 6 0 0 0 T T 2 T T 3 No. Sequences

Figure 3.1 Rarefaction analysis, plotting number of observed OTUs against total sequences obtained, for each Pleurobranchaea maculata sample. Abbreviations are: non-toxic egg masses (NE1-2), toxic egg masses (TE1-2), non-toxic, wild, P. maculata (NS1-3), toxic, wild, P. maculata (TS1-3), captive P. maculata (TT1-3). Solid lines represent a 5% similarity level between samples, dashed lines represent 15% similarity between samples.

3.3.2. Correlation between total microbial population and TTX

The nMDS analysis showed clustering of samples based on their life stage (egg mass versus adult) at a 15% similarity level (Figure 3.2), except two of three wild, toxic adults that did not cluster with any other sample. Clustering based on toxicity or habitat was not observed.

Chapter 3

Furthermore, analysis of similarity (ANOSIM) showed that microbial communities associated with P. maculata egg masses and adults were statistically dissimilar

(R=0.642, P<0.001) (Table 3.1). However, statistical dissimilarity was not observed between community compositions based on any other factors. These observations were maintained when eggs were removed from the analysis. Furthermore, using the RELATE test no significant correlation (ρ=-0.01, P>0.05) between the concentration of TTX and bacterial community composition was determined.

Therefore, this study demonstrated that the presence of TTX in P. maculata is not related to the overall bacterial community composition, or vice versa. These results corroborate a previous study which used Automated Ribosomal Intergenic

Spacer Analysis (ARISA) to compare bacterial community structure with TTX concentration in P. maculata maintained in aquaria for 126 d and fed a non-toxic diet (Wood et al., 2012a). Wood et al. (2012a) observed a shift in the specific bacterial community in P. maculata throughout the experiment. However, the trend and rate of change did not match with the depuration of TTX. Additionally, the large variability of TTX concentration in P. maculata at the beginning of the experiment was not reflected in large variations in the microbial community.

Therefore, it was concluded that other factors, such as diet, are likely to be the main driving force for observed differences in bacterial communities.

However, this data does not dispel the possibility an exogenous origin of TTX in P. maculata. In almost all cases of microbially produced natural products, the biogenic origin can be attributed to a single bacterial strain (Berdy, 2005). A bacterial strain with an abundance that is positively correlated to TTX concentration may indicate a candidate TTX producer in P. maculata. Therefore,

Chapter 3

we examined the correlation between individual OTU abundance and TTX concentration.

Figure 3.2 Non-metric multi-dimensional scaling plots showing clustering of the bacterial communities from Pleurobranchaea maculata. Abbreviations are: non- toxic egg masses (NE1-2), toxic egg masses (TE1-2), non-toxic, wild, P. maculata (NS1-3), toxic, wild, P. maculata (TS1-3), captive P. maculata (TT1-3). Red solid lines represent a 5% similarity level between samples, green dashed lines represent 15% similarity between samples.

Chapter 3

3.3.3. Correlation between individual OTUs and TTX

Spearman correlations comparing the concentration of TTX and abundance of individual OTUs determined that 11 OTUs had significant (P<0.05) correlations with the concentration of TTX in P. maculata (Table 3.2). Four of these were positively correlated, that is, the concentration of TTX increased as bacterial abundance increased. These groups included unclassified bacteria (n=2) and

Rhodobacteriaeceae (n=2).

Table 3.1 Analysis of similarity statistics comparing bacterial community composition based on Pleurobranchaea maculata sample factors. Analysis was performed on all samples (adults and egg masses) as well as a subset excluding egg mass samples. R approaches 1 when samples are dissimilar.

Total sample-set Excluding egg masses Group Group R P-value Group Group R P-value Egg Adult 0.642 0.001 Wild Captive -0.053 0.612 Wild Captive 0.16 0.202 Toxic Non-toxic 0.09 0.233 Toxic Non-toxic 0.173 0.19

Chapter 3

Table 3.2 Operational Taxonomic Units (OTUs) with significant (P<0.05) Spearman correlation constants in comparison to TTX concentration and their percentage composition in Pleurobranchaea maculata specimens. Non-toxic egg masses (NE1- 2), toxic egg masses (TE1-2), non-toxic, wild, P. maculata (NS1-3), toxic, wild, P. maculata (TS1-3), captive P. maculata (TT1-3).

% composition of OTU1 OTU Taxonomy ρ P-value NS1 NS2 NS3 TS1 TS2 TS3 TT1 TT2 TT3 0767 Bacteria 0.856 0.003 1.4 0.4 2.9

0766 Bacteria 0.798 0.01 29.5 2.3 1.6 23.2

0712 Rhodobacteraceae 0.737 0.023 0.2 0.2

0770 Rhodobacteraceae 0.72 0.029 0.7 0.5

0431 Bacteria -0.672 0.047 0.4 1.3 0.4 0.2

0408 Cyanobacteria_GpIIa -0.720 0.029 4.0 4.5 17.8 0.2 0.2

0405 Rhodobacteraceae -0.816 0.007 21.8 6.8 0.7

0406 Bacteria -0.821 0.007 28.6 28.6 14.6

0422 Cyanobacteria_GpIIa -0.821 0.007 0.5 0.2 0.2

0437 Rhodobacteraceae -0.836 0.005 0.2 0.2 0.2

0466 Bacteria -0.836 0.005 0.2 0.2 0.2

P. maculata TTX concentration (mg/kg) 26.0 8.1 7.7 15.4 3.8 31.0 1blank values represent an absence of the OTU

Chapter 3

No member of the Rhodobacteraceae, a family within the class

Alphaproteobacteria, has been linked to TTX production. However, one

Alphaproteobacteria, a Caulobacter isolated from the sediment of Lake Suwa,

Japan, has been reported to produce TTX (Do et al., 1993). Members of

Rhodobacteraceae are only related at a class level, to Caulobacter, which are members of the Caulobacteraceae family. Perhaps the most promising candidate for TTX production is an unclassified bacterium, present in four toxic P. maculata samples and not present in any of the non-toxic P. maculata. In the two most toxic samples, this OTU comprised between 20 and 30% of the total microbiota.

For further identification of these “unclassified” bacteria (OTUs 766 and 767), representative sequences were retrieved and compared to sequences in the NCBI nucleotide and 16S rRNA gene databases. The BLAST search against the full nucleotide database returned “unclassified bacterium” as the best alignment to our query. However, a similar search against a refined prokaryotic 16S rRNA gene database provided matches to the Mycoplasma genus. These cell-wall-lacking organisms typically live intracellularly association with eukaryotic host cells

(Razin et al., 1998) and have been linked to disease in bivalve molluscs (Azevedo,

1993). This genus is not currently known to produce TTX. The unique microbial physiology of Mycoplasma requires strict growth requirements, making them difficult to culture outside of their host cells. If this OTU is a true producer of TTX, it is not surprising that traditional growth media used for marine microorganisms were unable to cultivate these bacterium.

Ten of nineteen previously reported TTX-producing bacteria belong to the

Gammaproteobacteria. In particular, the Vibrionaceae have been closely associated

Chapter 3

with TTX and have been isolated from many TTX-producing marine animals

(Noguchi et al., 1987). Vibrionaceae were relatively rare in our study, present in only two toxic P. maculata samples and consisting of less than 1% of the microbiota of each of these (Table 3.3). Currently, members of 19 bacterial families are known to produce TTX in laboratory conditions, however, only 13 are represented in P. maculata and represent only 10.9% of the total microbial diversity in toxic P. maculata. Of particular interest are a Vibrio, Pseudoalteromonas and Bacillus, which were present in the most toxic P. maculata sample, albeit at under 1% abundance. As these bacteria are ubiquitous in the marine environment, it is difficult to attribute TTX production to their presence in this case. Additionally, bacteria from Microbacteriaceae, Caulobacteraceae, Streptomycetaceae,

Alteromonadaceae and Actinobacteria are known to produce TTX. However, their abundances in P. maculata were low, and were outside the top 40 most abundant families (Figure 3.3).

Chapter 3

Table 3.3 Forty most abundant bacterial families in toxic P. maculata P. maculata sample TS1 TS2 TS3 TT1 TT2 TT3 TTX Concentration 26.0 8.1 7.7 15.4 3.8 31.0 % Abundance1 Bacteria 44.1 8.6 13.5 17.8 24.3 42.3 Rhodobacteraceae 4.9 1.3 20.7 13.5 32.9 Staphylococcaceae 0.9 0.9 41.7 1.6 0.2 Sphingomonadaceae 32.4 Chromatiaceae 23.6 0.9 Flavobacteriaceae 0.2 0.2 0.4 22.1 1.3 Bradyrhizobiaceae 21.4 0.5 0.4 Micrococcaceae 0.2 5.8 13.3 0.5 Halomonadaceae 11.9 5.9 Nocardioidaceae 13.8 0.5 Xanthomonadaceae 1.4 7.7 1.1 0.2 3.8 Bacteroidetes 8.5 0.4 1.8 0.9 Cyanobacteria_Family_II 9.4 0.9 0.2 0.4 Comamonadaceae 0.4 5.0 2.0 1.1 0.2 Nitrospiraceae 6.3 0.9 Flavobacteriales 6.1 0.7 Acidobacteria_Gp4 2.3 4.1 Iamiaceae 6.1 Rhizobiales 2.5 3.6 Fusobacteriaceae 5.9 Methylobacteriaceae 5.9 Enterobacteriaceae 0.7 1.8 0.2 2.0 0.7 Bacillaceae 3.1 1.6 0.4 0.2 Acidobacteria_Gp16 4.5 0.7 Burkholderiales 4.7 0.4 Pseudomonadaceae 1.8 0.2 0.5 1.8 0.4 Gammaproteobacteria 2.5 0.4 0.4 1.1 Alphaproteobacteria 0.2 3.4 Proteobacteria 0.5 2.5 0.2 0.4 Acidobacteria_Gp6 3.4 Propionibacteriaceae 0.4 0.7 0.7 0.4 1.1 Planctomycetaceae 1.3 1.8 Campylobacteraceae 0.4 2.5 Actinomycetales 0.7 1.1 0.2 0.4 Vibrionaceae 0.9 0.9 Clostridiales 0.7 0.5 0.2 0.2 Deltaproteobacteria 0.9 0.7 Betaproteobacteria 0.4 1.1 Moraxellaceae 0.9 0.2 0.4 Pseudoalteromonadaceae 1.3 1blank values represent an absence of the OTU. Bolded values represent families with previously described TTX-producers.

Chapter 3

In some cases of natural products from microbial symbionts, the compound- producing microbe is enriched in the host. For example, the pederin-producing bacteria Pseudomonas sp. makes up the majority of the microbial flora in the

Paederus fuscipes beetles (Brady et al., 2009). Conversely, in the present study, two of the four OTUs that were positively correlated with TTX levels were not abundant in toxic P. maculata. However, these bacteria may still be candidates for the production of TTX. Marine invertebrates, including the squid Euprymna scolopes which harbours V. fischeri, have a compartmentalised organ where specific symbiotic bacteria reside. Therefore, these bacteria may be enriched within a specific location, but not necessarily in high overall abundance. The broad intra-organismal distribution of TTX in P. maculata (McNabb et al., 2010) and

Hapalochalaena spp. (Williams et al., 2012; Yotsu-Yamashita et al., 2007) could indicate a mechanism by which TTX producers are located within a specific tissue but TTX is transported throughout the host animal. A similar translocation process is observed in endophytes and their plant hosts (Spiering et al., 2005).

The possibility of TTX being bioaccumulated through the food chain cannot be overlooked. Recent surveys of benthic material collected at sites where P. maculata was abundant detected only trace levels of TTX in organisms that were unlikely to be dietary sources ( McNabb et al., 2010). This indicates that the production of TTX is either endogenous, produced by symbiotic microbes, or is sequestered from microbes ingested by P. maculata. Sequestration of chemical defences from the diets of marine invertebrates has been reported for marine organisms such as Sea hares of the genera Stylocheilus, Dolabella, and Aplysia. These organisms actively ingest cyanobacteria and sequester the cyanobacterial toxins that can persist for over 18 d (Harrigan and Goetz, 2002; Pennings and Paul, 1993). 78

Chapter 3

Although the methodology used in this study does not facilitate direct isolation of

TTX-producing bacteria, it serves as a guide for future bacterial isolation experiments. Future attempts to isolate TTX-producing candidates from P. maculata should focus on optimising culturing conditions selective for Mycoplasma and Rhodobacteriaceae. Localisation of these bacteria via 16S rRNA hybridisation may also assist in identifying individual tissues within P. maculata. Our previous attempts to culture microbes from P. maculata (chapter 2) yielded mainly

Vibrionaceae (12 isolates) and Pseudoalteromondaceae (6 isolates). However, these bacteria represented only 0.3% and 0.2% of the microflora, respectively, across all six toxic P. maculata. Indeed, the total diversity observed from our culture-based study represented only 11.1% of the total microbial diversity across the six toxic P. maculata samples. As such, traditional culturing methods are not representative of the total microflora present in P. maculata.

Our previous culture-based study (chapter 2) suggested that vertical transmission of TTX-producing bacteria was unlikely. This was corroborated by a comparison of the OTUs present in adult P. maculata and egg masses (Table 3.4). Only five OTUs were common between adults and egg masses of toxic, wild P. maculata and an additional 11 OTUs were common between the non-toxic adults and egg masses. Of these, Gammaproteobacteria and Enterobacteraceae contain members known to produce TTX. However, these two OTUs were less than 1% of the microbial abundance in adult P. maculata. These results are in contrast to a similar study which showed 126 ARISA fragments were common between adults and egg masses

(Wood et al., 2012a). Therefore, we believe that vertical transmission of TTX- producing bacteria is unlikely in P. maculata. This is also supported by the

Chapter 3

observation that second generation laboratory-reared P. maculata do not contain

TTX (Wood et al., 2012a).

Table 3.4 Common OTUs between P. maculata egg masses and adults

Averaged OTU Classification abundance (%) TE TS 0005 Gammaproteobacteria 2.1 0.5 0019 Sphingomonas 0.3 10.1 0027 Stenotrophomonas 0.3 2.3 0033 Ralstonia 0.2 0.1 0041 Enterobacteraceae 0.2 0.6

Averaged OTU Classification abundance (%) NE NS 0003 Rhodobacteraceae 5.5 0.8 0015 Rhodobacteraceae 10.1 0.5 0020 Rhodobacteraceae 0.3 0.2 0112 Rhodobacteraceae 0.4 0.5 0137 Alphaproteobacteria 9.3 0.2 0249 Gammaproteobacteria 0.3 0.1 0259 Flavobacteriales 0.1 0.2 0297 Alphaproteobacteria 0.3 0.2 0322 Rhodobacteraceae 0.4 0.5 0345 Alphaproteobacteria 0.9 0.1 0398 Bacteria 0.1 0.1

Chapter 3

3.4. Concluding remarks

Differences in microbial consortia of P. maculata, as a whole, do not correlate to differences in TTX concentrations. However, the identification of four OTUs that correlated to TTX concentration indicate that a bacterial origin of TTX in P. maculata is plausible. This study used a new approach in the search for the biogenic origin of TTX in marine animals. This methodology can be similarly applied in other environments to assist in the elucidation of the origin of TTX and other natural products. Although we were unable to isolate a TTX producer in chapters 2 and 3, an initial screen of the isolated microbes indicated many with potential for the biosynthesis of NRPS and PKS natural products. In particular, a

Pseudoalteromonas isolate was observed to contain many NRPS and PKS genes. In the following chapter, the genome sequencing and bioinformatic analysis of these genes and their associated biosynthetic pathways is described.

Chapter 4

Chapter 4: Genome sequencing of a Pseudoalteromonas bacterium reveals an unprecedented abundance of secondary metabolite gene clusters

4.1. Introduction

The genus Pseudoalteromonas, formerly classified as Alteromonas (Gauthier et al.,

1995), is a group of bacteria with significant ecological importance. They are commonly found in association with eukaryotic organisms and are known to produce numerous biologically active natural products (Bowman, 2007).

Pseudoalteromonas spp. are divided into two clades, non-pigmented and yellow to red pigmented. It has been observed that pigmented Pseudoalteromonas often produce a greater diversity of biologically active compounds. These compounds include the bioactive bromoalterochromides A and B from Ps. maricoloris isolated from sponges ( Speitling et al., 2007). Ps. tunicata isolated from the surface of marine algae have been observed to produce a tambjamine-like compound, YP1, which inhibits the settlement of invertebrate larvae (Franks et al., 2005).

Furthermore, Ps. tetraodonis cultivated from toxic puffer fish have been shown to be producers of TTX (Simidu et al., 1990). Other natural products isolated from

Pseudoalteromonas include isatin, brominated pyrroles, depsipeptides and diketopiperazines (Bowman, 2007).

Remarkably, the biosynthesis of only two of these natural products, YP1 and pentabromopseudilin, has been proposed. The identification of a biosynthetic gene cluster for YT1, facilitated by functional genomics, has enabled the proposal of its

Chapter 4

biosynthetic pathway ( Burke et al., 2007). The biosynthesis pathway of the pyrrole-containing pentabromopseudilin was elucidated using a feeding study using labelled acetate, amino acids and glucose (Hanefeld et al., 1994; Peschke et al., 2005). However, a gene cluster encoding for pentabromopseudilin biosynthesis has yet to be elucidated. Indeed, biosynthesis of most natural products produced by Pseudoalteromonas is unknown and biosynthetic gene clusters encoding their production have yet to be identified.

The reduction in cost of genome sequencing coupled with faster and more powerful bioinformatic methods has facilitated the expedited discovery of novel natural products and their biosynthesis in other microbial genera. The recent publications of the Salinispora tropica and S. arenicola genomes (Penn et al., 2009;

Udwary et al., 2007) aided in the discovery of 49 natural product biosynthesis clusters and assisted in the structure elucidation of salinilactam A (Udwary et al.,

2007). Such genomic approaches have also been used to great success in the actinomycete, Actinosynnema mirum, where genome-guided approaches facilitated the discovery of an unusual siderophore and the first identified biosynthetic gene cluster in the species (Giessen et al., 2012). Mining for biosynthetic gene clusters is made possible because the genes responsible for the metabolism of a natural product are usually encoded within operon-like gene clusters (Fischbach and

Walsh, 2006). The collinearity of non-ribosomal peptide synthetase (NRPS) and polyketide synthase (PKS) pathways, combined with knowledge of their biosynthetic mechanisms, enables the prediction of novel chemical structures derived from these gene clusters (Fischbach and Walsh, 2006).

Chapter 4

In this chapter, we use similar approaches to facilitate the discovery of new natural products and natural product biosynthetic gene clusters in Pseudoalteromonas.

The strain, Pseudoalteromonas sp. HM-SA03, closely related to P. piscicida, isolated from the blue-ringed octopus, Hapalochalaena sp., was identified as having high biosynthetic potential, due to the presence of multiple NRPS and PKS and AMT genes (chapter 2). Except the bromoalterochromides, Pseudoalteromonas natural products are not known to be derived from NRPS or PKS biosynthesis pathways.

Furthermore, Pseudoalteromonas are known to produce bioactive natural products. Methanol-derived crude extracts from HM-SA03 showed moderate inhibitory effects against Staphylococcus aureus (chapter 2). Therefore, analysis of its genome was performed in an attempt to reveal novel biosynthesis gene clusters. The interrogation of the HM-SA03 genome yielded numerous new biosynthesis gene clusters and potential natural products. The genome of HM-

SA03 has also enabled the identification of natural product biosynthesis pathways for previously identified Pseudoalteromonas natural products. This chapter reports this mining of the Pseudoalteromonas sp. HM-SA03 genome for novel secondary metabolism gene clusters.

4.2. Materials and methods

4.2.1. Sample preparation and genome sequencing

Pseudoalteromonas sp. HM-SA03 was grown in 0.5% peptone in filtered seawater at 23°C for 24 h. The cell culture was centrifuged at 4200 x g and a subset of the biomass was used in DNA extraction as previously described (Wilson, 2001). The media supernatant was stored at -20°C for downstream chemical analysis.

Chapter 4

Genome sequencing and comparative analyses were performed at the Clive and

Vera Ramaciotti Center for Gene Analysis at The University of New South Wales.

Genomic DNA was sequenced using the Illumina HiSeq following the standard

Illumina protocol. The sample was prepared using the Illumina paired-end sample preparation kit. Briefly, 5 µg of DNA was fragmented by nebulisation followed by end–repaired ligation of adaptors. Size selection was performed using 2% agarose gels and 350 bp fragments were recovered from the gel. Ten cycles of PCR were used to enrich the adapter-modified DNA fragments. The library was purified using a QIAquick PCR purification kit, diluted with Elution Buffer (Qiagen) to a final concentration of 10 nM, and stored at −20°C until used. The sample was run at 8 pM of paired-end 102 bp chemistry. The run was performed using the Genome analyser sequencing control software (SCS) v2.6.

4.2.2. Genome assembly

The SolexaQA package (Cox et al., 2010) was used to trim reads to the longest contiguous read segment above a 0.05 P-value. Quality-trimmed reads shorter than

50 bp were discarded. De novo genome assembly was performed with SOAPdenovo

(Luo et al., 2012) using k-mer values between 21 and 91. These k-mer values represent the minimum read overlap required to during the assembly of contigs.

Contiguous DNA sequences (contigs) shorter than 200 bp were discarded from the final assembly. For comparison, reads were also assembled using both Velvet

(Zerbino and Birney, 2008) and ABYSS (Simpson et al., 2009) genome assemblers.

Gene prediction and annotation was performed using the best assembly produced by the software.

Chapter 4

4.2.3. Gene prediction and annotation

The Rapid Annotation using Subsystem Technology (RAST) web application server was used for gene prediction and draft annotation (Aziz et al., 2008). Additionally, secondary metabolite biosynthesis clusters were identified using a combination of

2metDB (Bachmann and Ravel, 2009) and AntiSMASH (Medema et al., 2011). Both these software used profile Hidden Markov Models (pHMMs) of known biosynthesis gene domains to identify secondary metabolite genes and their domain architecture in query sequences. Substrates for PKS ketosynthase, NRPS adenylation and CoA ligase domains were also predicted using these programs. All secondary metabolite gene clusters retrieved were manually checked and further confirmation of domain architecture was performed using NCBI Conserved

Domain Database (CDD) Search (Marchler-Bauer et al., 2011).

4.2.4. Phylogenetic tree reconstruction of condensation domains

Amino acid sequences of condensation domains were obtained (Rausch et al.,

2007) and aligned with the AlcK N-terminus condensation domain amino acid sequence using MUSCLE (Edgar, 2004). Phylogenetic tree reconstruction using the

PhyML web server (http://www.atgc-montpellier.fr/phyml/) was performed using the JTT amino acid substitution model and a gamma-distribution with four rate substitution categories. Branch support measures were estimated using aLRT.

4.2.5. Small molecule extraction

HM-SA03 media supernatant was extracted by adsorption onto 20 g/L Amberlite

XAD-7HP resin for 1 h. The resin was filtered and washed with 10 ml MilliQ water to remove interfering media components. Adsorbed compounds were eluted twice

Chapter 4

with 10 mL methanol, and the combined washes were evaporated to dryness under reduced pressure. An uninnoculated culture was extracted using the same methodology and used as a control, for comparative purposes, during downstream analysis.

4.2.6. Mass spectrometry

Extracts were analysed using liquid chromatography-mass spectrometry (LC-MS) and were separated on a BEH C18 2.1 mm x 50 mm 1.9 µM UHPLC column (Waters,

MI) using a gradient of 0.1% formic acid in MilliQ water against acetonitrile at

400 µL/min using a ThermoFisher Scientific Accela pump. The gradient system was 0-5 min 0% B, linearily ramped to 100% B at 25 min, held for 1 min and returned to 0% B for 3 min. Column eluate was ionised using a positive mode electrospray source.

Chapter 4

4.3. Results and discussion

4.3.1. Genome sequencing and assembly

The genome assembly produced by SOAPdenovo, using a k-mer value of 71, resulted in a 5,248,267 bp assembly consisting of 119 scaffolds and 494 unscaffolded contigs with a N50 of 106,644 and a maximum contig length of

182,387 bp. In comparison, the next best assembly was performed by Velvet using a k-mer value of 63, producing a genome size of 5,218,927 bp consisting of 91 scaffolds and 298 unscaffolded contigs with an N50 of 101,219 and maximum contig length of 165,931 bp. The SOAPdenovo assembly was chosen for further analyses because it produced more scaffolds that were also longer than those produced by Velvet.

4.3.2. Gene annotation and genome mining

Both the Rapid Annotation using Subsystem Technology (RAST) in-house gene predictor and the Glimmer-3 module were used to identify protein-encoding genes within the HM-SA03 genome. Results from the RAST built-in gene prediction software were used for downstream analyses. Both methods produced similar results, with RAST predicting 4735 protein encoding genes and 90 RNAs. Through a combination of software-assisted annotation (AntiSMASH, 2metDB) and manual annotation, 28 genes organised within seven gene clusters were identified as being related to NRPSs or PKSs. Of these, 22 genes encoded NRPS functionality, 4 genes encoded PKSs and 2 encoded hybrid NRPS/PKSs. These NRPS and PKS genes were homologous to cloned PCR products from HM-SA03 sequenced in chapter 2. The identification of such a high number of biosynthesis genes in HM-SA03 represents

Chapter 4

one of the most potentially biosynthetically rich Pseudoalteromonas strains known.

In comparison, the recently sequenced Ps. piscicida JCM20779 genome also contains seven NRPS/PKS gene clusters, while Ps. sp. NJ631 contains three

NRPS/PKS gene clusters. Some of the NRPS/PKS genes from HM-SA03 and their associated biosynthesis gene clusters share high homology to those from published

Pseudoalteromonas genomes. However, structure prediction and biosynthetic pathway analyses have not been performed on these gene clusters, and thus are termed “orphans”.

4.3.3. Identification and in-silico characterisation of an alterochromide biosynthesis pathway

4.3.3.1. Identification of a putative alterochromide biosynthesis cluster

Mining of the HM-SA03 genome revealed the presence of truncated NRPS genes on the ends of two separate scaffolds, 52 and 119 (Figure 4.1A). Amino acid substrate specificity searches indicated that these genes were likely to incorporate thr, val

(scaffold 52, modules 1 and 2), asn and leu (scaffold 119, modules 4 and 5). This amino acid composition showed similarities to bromoalterochomides (Figure

4.1B) from Pseudoalteromonas isolated from sponges (Speitling et al., 2007), which had a cyclo-thr-val-asn-asn-leu peptide structure. From this, we believe the two scaffolds were contiguous, and that the NRPS genes located on these scaffolds were part of a single biosynthetic gene cluster. Analysis of publicly available genomic databases revealed that “contig 2” from Ps. piscicida JCM20779 (accession number:

AHCC01000002) shared significant (>95%) homology to scaffolds 52 and 119 from HM-SA03 (Figure 4.1).

Chapter 4

An attempt was made to retrieve the missing nucleotide sequence, which we predicted to be 3868 bp, separating scaffolds 52 and 119 (Figure 4.1A). A BLAST homology search between the unscaffolded HM-SA03 contigs to “Ps. piscicida

JCM20779 contig 2” revealed that a single contig from HM-SA03, C3605, had high homology to a portion of “Ps. piscicida JCM20779 contig 2” (Figure 4. 1A). C3605, spanning 2919 bp, was found to contain the missing adenylation domain encoding for the first asparagine incorporation in bromoalterochromide biosynthesis

(discussed further in section 4.3.3.4). The 71 bp on the 5’ end of C3605 were identical to those on the 5’ end of scaffold 119, which may have disrupted the assembly between scaffolds 52, C3605 and scaffold 119. Sanger sequencing was utilised to close the gaps between scaffold 52, C3605 and scaffold 119 and assemble the complete gene cluster.

In addition to the heterocyclic amino acid residues, bromoalterochromides contain a 9-(4-hydroxyphenyl)-nonatetraenoic acid (hydroxy-PNTA) moiety that is either mono- or dibrominated at the 3 and 5 positions of the benzyl ring. The genes that encode for the biosynthesis of this part of the compound were not detected by

AntiSMASH. However, BLAST and CDD searches of the genes directly upstream from the three NRPS genes revealed genes encoding for an assortment of polypeptide biosynthesis enzymes and tailoring enzymes that are proposed to be involved in the biosynthesis of the phenyl-polyketide chain of alterochromide

(Table 4.1). A detailed analysis of the biosynthesis pathway and genes involved are discussed in the next section.

Chapter 4

A Pseudoalteromonas piscicida JCM20779 contig 2 5' 3'

C - A1 - T - E C - A2 - T - E - C - A3 - T C - A4 - T - E - C - A5 - T - E - TE

C - A1 - T - E C - A2 - T - E - C - - A3 - T C - - A4 - T - E - C - A5 - T - E - TE 5' 3' scaffold 52 C3605 scaffold 119 Pseudoalteromonas sp. HM-SA03 96% 95% 95% Nucleotide sequence identity B

A2:Val

A3:Asn

A4:Asn A1:Thr

A5:Leu

Figure 4.1 The NRP S biosynthesis gene cluster and structure of bromoalterochromide. Alignment of Ps. piscicida JCM20779 contig 2 with HM-SA03 contigs containing putative alterochromide NRPS genes. Initial de-novo genome mining only identified scaffolds 52 and 119 as containing putative alterochromide NRPS genes. Subsequent mining of the HM-SA03 genome using JCM20779 as a reference identified C3605 as a contig containing the missing gene sequence of the NRPS cluster. All of the HM-SA03 contigs had high (>95%) identity to JCM20779 contig 2 (A). Structural formula of bromoalterochromide. Dotted red lines separate the amino acids on the NRPS derived portion of the compound. Adenylation domains responsible for amino acid incorporation and their substrate specificities are indicated in red. The 9-(3-bromo-4-hydroxyphenyl)-nonatetraenoic acid (bromohydroxy-PNTA) moiety common to most alterochromides is shown attached to the threonine residue (B).

Chapter 4

Table 4.1 Proposed encoded functions of alterochromide-like biosynthesis gene cluster ORFs

ID Length Closest homology (BLASTP/nr) Identity/ Accession Proposed Similarity function AlcA 540 histidine ammonia-lyase 99/99 WP_010606259.1 PAL

[Pseudoalteromonas flavipulchra] AlcB 98 hypothetical protein 97/97 WP_010369398.1 T domain [Pseudoalteromonas flavipulchra] AlcC 499 AMP-dependent synthetase and 94/96 WP_010606258.1 Loading T-A ligase [Pseudoalteromonas

flavipulchra] AlcD 728 3-oxoacyl-ACP synthase 98/99 WP_010606257.1 KS [Pseudoalteromonas flavipulchra] AlcE 256 ABC transporter ATPase 100/100 WP_010369407.1 Transporter [Pseudoalteromonas flavipulchra] AlcF 372 ABC transporter permease 100/100 WP_010369409.1 Transporter [Pseudoalteromonas flavipulchra] AlcG 133 3-hydroxyacyl-ACP dehydratase 100/100 WP_010369411.1 DH [Pseudoalteromonas flavipulchra] AlcH 244 3-oxoacyl-ACP reductase 100/100 WP_010369414.1 KR [Pseudoalteromonas piscicida] AlcI 236 hypothetical protein PflaJ_09712 97/98 WP_010606255.1 Unknown

[Pseudoalteromonas flavipulchra] function AlcJ 252 thioesterase 98/100 WP_010606254.1 TE

[Pseudoalteromonas flavipulchra] AlcK 1518 non-ribosomal peptide synthase 98/98 WP_017218128.1 C-A-T-E PNJ4335 [Pseudoalteromonas sp.

NJ631] AlcL ~2300 ornithine racemase, partial 97/98 WP_017218128.1 C-A-T-E-C

[Pseudoalteromonas flavipulchra] AlcM 2574 Non-ribosomal peptide synthase 96/97 WP_010369430.1 A-T-E-C-A-T- [Pseudoalteromonas piscicida] E-TE

Chapter 4

4.3.3.2. Formation and loading of cinnamic acid

The first step in the proposed biosynthesis of alterochromide is the deamination of phenylalanine to form cinnamic acid (Figure 4.2). The deamination of amino acids is catalysed by the putative ammonia-lyase enzyme, AlcA. These enzymes are common in plant biosynthesis pathways, and provide a key intermediate in the shikimate pathway. AlcA has closest homology to a histidine ammonia-lyase (HAL) from Ps. flavipulchra JG1 (Figure 4.3). We propose that this enzyme is a phenylalanine ammonia-lyase (PAL), rather than a histidine ammonia-lyase. It has been shown that the amino acid residue at position 89 (referenced to Rhodobacter sphaeroides TAL) is key to determining the selectivity of the lyase substrate (Louie et al., 2006; Watts et al., 2006). Generally, tyrosine ammonia-lyases (TALs) have a histidine residue at position 89, while HALs possess a serine and PALs, a phenylalanine. The exception to this is the R. toruloides PAL which contains a histidine residue at this position and is able to deaminate both phenylalanine and tyrosine (Jiang et al., 2005). AlcA has a tyrosine at position 89, which has not been observed in structurally characterised PALs. However, it is uncertain how a phenylalanine to tyrosine substitution would alter the substrate specificity of the lyase. Sequence analysis of other PALs identified leucine at position 90, whereas the HM-SA03 amino acid ammonia-lyase contains a histidine. It has been observed that a substitution at this position does not affect the substrate specificity of this enzyme. Mutation of the leucine to histidine at position 90 in the P. crispum PAL

(corresponding to the R. sphaeroides TAL position 90) did not alter PAL to HAL activity (Poppe and Rétey, 2005; Röther et al., 2002). EncP, a biochemically characterised PAL from the enterocin polyketide pathway shows no amino acid

Chapter 4

similarity in the selectivity switch to other characterised PAL (Xiang and Moore,

2005). Furthermore, a putative aromatic amino acid ammonia-lyase from

Stenotrophomonas maltophilia Ab55555 (accession number: J7VQS8), which contains Tyr89-His90, has been annotated as a PAL. From the analysis above, it is proposed that the HM-SA03 is a PAL and utilises phenylalanine as a substrate in the production of cinnamic acid.

AlcA AlcC AlcD AlcH AlcG AlcJ PAL T A KS KR DH TE

T Lipoinitiation TE-mediated release

AlcB AT? T Acyl- transferase

NRPS 3x malonyl extension

Acyl-CoA Ligase

Figure 4.2 Biosynthetic pathway for the production of the phenyl-polyketide starter unit for alterochromide. Phenylalanine is converted to cinnamic acid via PAL, which undergoes malonyl extension and is reduced to form 9-phenyl- nonatetraenoic acid (PNTA). This is primed for incorporation into the NRPS pathway in a process known as lipoinitation.

Chapter 4

The loading of carboxylic acid precursors, such as cinnamic acid, onto NRPS or PKS assembly systems have been observed to occur via two methods. The first and best documented, process is facilitated by a coenzyme A (CoA) ligase. The biosynthesis of many hybrid peptide and polyketide products utilise this system, including that of enterocin (Izumikawa et al., 2006; Piel et al., 2000). The biosynthesis of enterocin follows a similar starting strategy to that we propose for alterochromide.

The PAL facilitates the production of cinnamic acid, which is activated by EncH, a cinnamoyl-CoA ligase to form cinnamoyl-CoA. EncI and EncJ catalyse the conversion of cinnamate to benzoate that is ultimately loaded onto the ketosynthase (KS) via an aryl-CoA ligase, EncN. The alternative route is seen in rifamycin biosynthesis, where an N-terminus protein component of RifA contains a

T-A didomain that facilitates the selection and CoA-independent thiolation of 3- amino-5-hydroxybenzoate directly upstream of the first polyketide condensing module (Admiraal et al., 2001). AlcC shows high homology to an AMP-dependent synthetase and ligase from Ps. flavipulchra JG1 and encodes for a T-A didomain.

This standalone T-A didomain is proposed to prime the PKS with the cinnamic acid in a CoA-independent fashion similar to that of rifamycin. Cinnamic acid is activated as a cinnamoyl adenylate by the adenylation domain of AlcC. Cinnamoyl adenylate is then thiolated directly to form cinnamoyl-S-enzyme intermediate for downstream extension by the type II PKS. Due to the absence of a nearby CoA- ligase, this latter mechanism, facilitated by AlcC, is proposed for alterochromide

(Figure 4.2).

As mentioned above, many A-T didomains are encoded within the same gene as the downstream polyketide condensing modules. However, in this study, AlcC is a stand-alone protein. This is explained by the downstream type II PKS genes, which 95

Chapter 4

are typically encoded as single proteins encoding for a distinct domain activity

(Fischbach and Walsh, 2006). Rather than a type I PKS linear assembly line as in the previous examples, type II PKS assemble polyketides in an iterative manner, where a domain is utilised multiple times for the extension of a single polyketide chain.

89 149 R. sphaeroides TAL L QAN L VHH L A S. . . . .G T VGA S GD L T P L P.putida HAL LQRS LV LSHAA. . . . .G SVGASGDLAP L P. crispum PAL LQKE L I RF LNA. . . . .G T I TASGDLVP L R. toruloides PAL LQKA L L EHQL C. . . . .G T I SASGD L S P L A. variabilis PAL LQTN LVWF L KT. . . . .G S I GA SGD LV P L N.punctiforme PAL L QT N L IWF L K S. . . . .G S I GA S GD L V P L P. HM-SA03 PAL L QQN L I A Y H G C . . . . .G S V G A S G D L T P L Selectivity switch MIO

Figure 4.3 Amino acid alignment of aromatic amino acid ammonia-lyases. HutH from Rhodobacter sphaeroides (Q3IWB0), HutH from Pseudomonas putida (P21310), PAL1 from Petroselinum crispum (P24481), PAL from Rhodosporidium toruloides (P11544), AvPAL from Anabaena variabilis (Q3M5Z3), NpPAL from Nostoc punctiforme (B2J528) and this study (Ps. HM-SA03 PAL). Numbering corresponds to that of R. sphaeroides TAL. The His89 position (yellow) has been shown to determine substrate specificity in aromatic amino acid ammonia-lyases. Ala 149-Ser 140-Glu 151 (blue) corresponds to the conserved residues that are spontaneously cyclised and dehydrated to form 4-methylidene-imidazole-5-one (MIO), a cofactor for these enzymes.

Chapter 4

4.3.3.3. Polyketide extension of cinnamate and lipoinitiation

The biosynthesis of the PNTA moiety of alterochromide is completed via PKS chain elongation. Cinnamoyl-S-enzyme intermediate undergoes 3 rounds of malonyl extension catalysed by the type II PKS genes, AlcB, D, H and G. The dimeric ketosynthase-chain length factor (KS-CLF), AlcD catalyses C-C bond formation between the cinnamoyl thioester and a malonyl carbanionic acyl acceptor bound onto AlcB. This process is performed three times, extending the polyketide chain by six carbon atoms. Following each condensation step, reduction of the ketone is performed sequentially by AlcH and AlcG. Following polyketide extension, the resultant PNTA is released by AlcJ, a putative thoesterase. Chain termination is regulated by the chain length factor (CLF), also known as KSβ which forms a dimeric complex with the ketosynthase, KSα. This dimeric enzyme contains a pocket at the interface between KS and CLF and only products with a sufficient chain length to fill this pocket are transferred to the thioesterase (Dreier and

Khosla, 2000; Fischbach and Walsh, 2006).

PNTA is primed for incorporation in a process known as lipoinitiation. A variety of strategies for the incorporation of long chain fatty acids onto NRPS biosynthesis pathways of lipopeptides, such as surfactin and daptomycin, have been observed

(Chooi and Tang, 2010). Decanoic acid, a mycosubtilin precursor, is activated by a fatty acyl ligase and loaded directly onto a thiolation domain, by-passing the decanoyl-CoA thioester intermediate (Hansen et al., 2007). In daptomycin, a discrete fatty acyl ligase (DptE) catalyses the activation and transfer of fatty acid precursors onto a thiolation domain (DptF), to be incorporated by a N-terminus condensation domain onto a downstream amino acid (Wittmann et al., 2008). The

Chapter 4

simplest of these lipoinitiation strategies is that of surfactin (Kraas et al., 2010).

The N-terminal condensation of the surfactin synthetase is able to directly catalyse the condensation of a fatty acyl-CoA to downstream amino acid moiety. Unlike daptomycin biosynthesis, the fatty acyl moiety is not required to be tethered to a thiolation domain. The presence of a N-terminus condensation domain without genetically clustered fatty acyl ligase or thiolation domain genes suggests a lipoinitation strategy similar to daptomycin or surfactin. The fatty acyl ligases and/or acyl carrier protein may act in trans, encoded elsewhere in the genome.

These enzymes may be borrowed from fatty acid biosynthesis systems in the organism (Fischbach and Walsh, 2006). Supporting this hypothesis, HMM analysis of AlcK, using a profile constructed for four subtypes of condensation domains

(starter, LCL, DCL, dual), indicates the N-terminus C domain is more likely to be a starter condensation domain rather than a LCL or DCL. Furthermore, phylogenetic tree reconstruction (Figure 4. 4) indicated that the N-terminus C domain of AlcK groups with the clade representing starter C domains and is more closely related to other starter-unit C domains than to other C domains. Therefore, we propose a lipoinitiation strategy for alterochromide biosynthesis that is similar to that of surfacin biosynthesis. PNTA is both activated and condensed to a downstream threonine via a starter C domain encoded by AlcK.

Chapter 4

Figure 4.4 Phylogenetic tree of condensation domain subtypes. Sequences were obtained from a previous study (Rausch et al., 2007) and realigned with the sequences from HM-SA03 using MUSCLE (Edgar, 2004). The proposed starter C domain of AlcK is indicated by the bolded red line.

Chapter 4

4.3.3.4. Non-ribosomal extension and heterocyclisation

As previously discussed, the first NRPS module, encoded by AlcK catalyses the selection, activation and condensation of PNTA onto L-threonine (Figure 4.5), the first amino acid moiety of alterochromides. Epimerisation of L-threonine to D- threonine is catalysed by the C-terminus epimerisation domain of AlcK. AlcL encodes two NRPS modules, facilitating the incorporation of L-valine and L- asparagine, which are also epimerised to their D-isoforms. The final gene in the biosynthesis of alterochromide is AlcM, which encodes an NRPS incorporating a second asparagine and leucine onto the growing peptide chain. These L-amino acids are converted into D-amino acids through the action of intra-modular epimerases. This linear lipopeptide chain is cyclised and released by the C- terminus thioesterase domain of AlcM. The thioesterase domain catalyses the formation of the C-N bond between the leucine amino group and the carboxylic alcohol of threonine, cyclising the amino acid residues and releasing the final product from the enzymatic complex.

100

Chapter 4

AlcK AlcL AlcM

C* Athr T E C Aval T E C Aasn T C Aasn T E C Aleu T E TE

+ L-Thr

+ L-Val

+ L-Asn

+ L-Leu

Cyclisation

Figure 4.5 NRPS pathway for alterochromide.The N-terminus condensation domain (marked with an asterisk) selects for PNTA and catalyses its condensation to L-threonine. Subsequent condensation reactions incorporate L-valine, two L- asparagines and L-leucine. The compound is released and cyclised via the C- terminus thioesterase domain on AlcM.

101

Chapter 4

4.3.3.5. Tailoring and diversity of alterochromide structures

The reported structures of alterochromides are exclusively brominated on the 3 and/or 5 positions on the tyrosine derived aromatic ring (Speitling et al., 2007). It is believed that for bromoalterochromides, the deamination of tyrosine, rather than phenylalanine, as proposed in this study, is the first selective step in its biosynthesis. Furthermore, analysis of the mass spectra (Figure 4.6) of an organic extract of HM-SA03 revealed four parent clusters that were tentatively assigned as non-brominated alterochromides. The typical isotopic distribution of bromide was not observed in the mass spectra supporting an unbrominated structure.

Therefore, the incorporation of phenylalanine rather than tyrosine may have led to the lack of bromination. Together, this evidence supports the identification of a

PAL as the first enzyme in the biosynthesis pathway and the incorporation of phenylalanine in alterochromide biosynthesis. Furthermore, the m/z [M+H]+ of alterochromide parent ions indicate an unsaturated or monosaturated fatty acid carbon chain, instead of the fully unsaturated carbon chains observed in the literature (Speitling et al., 2007). Reduction of double bonds in the fatty acid chain is proposed to be catalysed in trans by an enoyl reductase, possibly sequestered from fatty acid biosynthesis pathways in the bacterium.

Mass spectra supporting the production of alterochromides incorporating variable fatty acid precursors have also been observed (Figure 4.6). These could highlight the promiscuity of the initiation C domain for substrates other than PNTA. The relaxed specificity of a starter C domain to fatty acyl-CoA substrates has been observed in the biosynthesis of the lipopeptide antibiotic, calcium-dependent antibiotic (CDA) ( Kraas et al., 2012). The identification of a starter C domain with

102

Chapter 4

relaxed specificity for fatty acid substrates is promising for the combinatorial biosynthesis of novel lipopeptides.

Figure 4.6 Positive mode ESI mass spectra of HM-SA03 crude extract. [M+H]+ parent ions indicate the production of alterochromide derivatives by HM-SA03. The chemical structure enclosed in red is proposed to be produced by the pathway discussed in this study. Other structures may be derived from different fatty acid chain lengths and side group.

103

Chapter 4

4.3.4. Identification and in-silico characterisation of NRPS siderophore biosynthesis pathways

4.4.3.1. Identification of siderophore biosynthesis gene clusters

Many microbes require iron for optimum growth (Neilands, 1995). Electron transport systems and many enzymes require iron for their proper function.

However, iron in its soluble form is rare in the environment, and is usually found as insoluble oxide or hydroxide forms. Siderophores are iron-chelating compounds produced by microbes to assist in the sequestration of iron from the environment.

The biosynthesis of these siderophores can occur via two biosynthesis pathways,

NRPS and NRPS-independent pathways (Barry and Challis, 2009).

Two biosynthetic gene clusters encoding for the production of siderophores via

NRPS pathways were identified in HM-SA03. Both gene clusters were assigned as siderophore biosynthesis gene clusters due to the proximity of genes encoding for iron receptors and siderophore transport proteins. Both of the siderophore biosynthesis pathways identified in HM-SA03 incorporate either 2,3- dihydroxybenzoate or salicylate in their biosynthesis. These starter units are common in many siderophores, and play an important role in ferric ion coordination. The biosynthesis of these starter units is derived from the shikimate pathway, via salicylate or 2,3-dihydroxybenzoate (DHB) (Figure 4.7). The production of DHB from isochorismate, a biosynthetic precursor derived from shikimate, is a two-step process catalysed by isochorismatase which cleaves pyruvate to form 2,3-dihydro-2,3-dihydroxylbenzoate which is then reduced to provide the fully aromatic DHB.

104

Chapter 4

4.4.3.2. Putative biosynthesis gene cluster for pseudoalterobactin-like compounds

The first siderophore biosynthesis cluster, pab (Table 4.2, Figure 4.8) is predicted to produce a pseudoalterobactin-like compound. A biosynthetic gene cluster for pseudoalterobactin (Kanoh et al., 2004), or the structurally similar alterobactin

(Deng et al., 1995), has not been previously identified, let alone characterised. The proposed biosynthetic gene cluster for this pseudoalterobactin-like compound spans 47 kb and contains seven NRPS domain-containing genes and one PKS gene.

Three genes, pabQOM, likely encoding for the biosynthesis of the DHB starter unit are present in the gene cluster. Numerous siderophore and iron receptor, regulation and transport proteins, pabDEKLR, are also encoded within the gene cluster. This cluster is also preceded by an MbtH domain-containing protein, PabA, which is found in many NRPS-dependent siderophore biosynthesis clusters

(Lautru et al., 2007; Quadri et al., 1998) and is thought to be essential for the correct biosynthesis of siderophores in vivo (Lautru et al., 2007). The presence of an MbtH domain-containing protein reinforces the classification of this genomic locus as an NRPS-dependent siderophore biosynthesis gene cluster.

The biosynthesis of the DHB starter unit, as outlined in the previous section, is facilitated by PabQOM. Activation of DHB is facilitated by a CoA-ligase domain, located at the N-terminus of PabB. The C-domain of PabB catalyses bond formation between the carbonyl and ε-amino groups of DHB and L-lysine, respectively.

Polyketide extension followed by the incorporation of L-asparagine, L-lysine, a variable amino acid, L-asparatate and L-glycine, catalysed by PabBCFGIJ. A consensus for the substrate specificity of the second adenylation domain of PabG was unable to be achieved and is proposed to encode for either L-ornithine or an

105

Chapter 4

acidic amino acid. A terminal thioesterase of PabJ catalyses the cyclisation of the peptide chain between the glycine carbonyl and the free amino group of the second lysine side chain. The function of PabH, a putative dioxygenase, is unknown, but may play a role in the β-hydroxlation of L-asparatate.

Intriguingly, the activation of the DHB starter unit seems to be encoded by a redundant set of proteins, PabP, PabO and PabN. These three genes are encoded adjacent to the DHB biosynthesis genes, downstream and in reverse orientation to the NRPS and PKS biosynthesis genes. PabP is an adenylation domain-containing protein with substrate specificity for DHB. PabO encodes a 2,3-dihydro-2,3- dihydroxybenzoate synthetase and also contains a thiolation domain. This domain may be involved in the tethering of DHB to the NRPS. PabN encodes a condensation domain with homology to starter condensation domains. As discussed for alterochromide biosynthesis, these starter condensation domains have substrate specificity for unusual starter units, including benzoates and fatty acids. The reason for the presence of two redundant pathways for the incorporation of DHB into the molecule is unknown, however, Streptomyces maritimus also has two independent biosynthetic pathways for the incorporation of benzoate into enterocin (Izumikawa et al., 2006).

Another unusual feature of this gene cluster is the presence of consecutive C- domains, one in module 5 and one in module 7. The second C-domain of module 5 is believed to be inactive, due to a mutation in the second histidine of the conserved HHxxxDG motif, that is critical to the correct function of the condensation domain (Sieber and Marahiel, 2005). Consecutive C-domains are found in NRPS pathways of other natural products, including bleomycin (Shen et

106

Chapter 4

al., 2001) and tallysomycin (Tao et al., 2006). The function of this second C-domain has not been discussed in these papers, however, one possibility is that the first C domain encodes for an epimerase, which is structurally similar to a condensation domain, and is usually located immediately prior to a condensation domain

(Rausch et al., 2007).

The predicted compound produced by the pab gene cluster in HM-SA03 is structurally similar to pseudoalterobactin A and B. However, the structure prediction from this study lacks an aspartate residue between the asparagine residue and cyclic peptide ring. Furthermore, there is no evidence to suggest sulfonation of DHB, as seen in pseudoalterobactins, however, these enzymes may act in trans and therefore may not be clustered with the NRPS/PKS biosynthesis genes. Apart from these differences, the structure of the predicted compound produced by the pab gene cluster is similar to pseudoalterobactins A and B. Both contain the DHB-Lys-malonyl side chain and a Lys-amino acid-Asp-Gly cyclic peptide. The location of the variable amino acid in pseudoalterobactins, alterobactins and the predicted product of the pab cluster is between lysine2 and aspartate1 (aspartate2 in the case of pseudoalterobactin and alterobactin). This suggests that the second adenylation domain of PabG (module 5) and its homologues in the pseudoalterobactin and alterobactin gene clusters have relaxed substrate specificities.

107

Chapter 4

isochorismate isochorismatase dehydrogenase synthase

2,3-dihydro- 2,3-dihydroxybenzoate chorismate isochorismate 2,3-dihydroxybenzoate

salicylate synthase

salicylate

Figure 4.7 Biosynthesis pathway for the typical siderophore starter units in salicylate and DHB.

108

Chapter 4

Table 4.2 Proposed encoded functions of pseudoalterobacin biosynthesis gene cluster ORFs

Protein Length Closest homology (BLASTP/nr) Identity/ Accession ID Proposed Similarity function PabA 68 MbtH domain-containing protein 99/99 ZP_10289491.1 Unknown [Pseudoalteromonas piscicida JCM 20779] PabB 1742 amino acid adenylation domain 97/99 ZP_10289492.1 AL-T-C-A-T [Pseudoalteromonas piscicida JCM 20779] PabC 1464 KR domain-containing protein 96/99 ZP_10289493.1 KS-AT-KR-T [Pseudoalteromonas piscicida JCM 20779] PabD 807 peptidase S45 penicillin amidase 97/99 ZP_10289494.1 Regulation [Pseudoalteromonas piscicida JCM 20779] PabE 722 outer membrane receptor for ferric iron uptake 98/99 ZP_10289495.1 Transport [Pseudoalteromonas piscicida JCM 20779] PabF 1539 long-chain-fatty-acid--CoA ligase 95/98 ZP_10289496.1 C-A-T-C [Pseudoalteromonas piscicida JCM 20779] PabG 3027 peptide synthase 95/97 ZP_10289497.1 C-A-T-E-C-A- [Pseudoalteromonas piscicida JCM 20779] T-C PabH 354 non ribosomal peptide synthase 98/99 ZP_10289498.1 Dioxygenase [Pseudoalteromonas piscicida JCM 20779] PabI 1037 amino acid adenylation protein 96/98 ZP_10289499.1 C-A-T [Pseudoalteromonas piscicida JCM 20779] PabJ 1319 peptide synthetase 95/97 ZP_10289500.1 C-A-T-TE [Pseudoalteromonas piscicida JCM 20779] PabK 694 TonB-dependent siderophore receptor 97/98 ZP_10289501.1 Receptor [Pseudoalteromonas piscicida JCM 20779] PabL 448 PepSY-associated TM helix family protein 70/82 ZP_10300783.1 Receptor [Pseudoalteromonas spongiae UST010723-006] PabM 255 2,3-dihydroxybenzoate-2,3-dehydrogenase 96/97 ZP_10289504.1 DHB [Pseudoalteromonas piscicida JCM 20779] biosynthesis PabN 436 hypothetical protein PpisJ2_11150 96/97 ZP_10289505.1 C-domain [Pseudoalteromonas piscicida JCM 20779] PabO 289 2,3-dihydro-2,3-dihydroxybenzoate synthetase 97/98 ZP_10289506.1 DHB [Pseudoalteromonas piscicida JCM 20779] biosynthesis PabP 539 enterobactin synthase subunit E 97/98 ZP_10289507.1 A-domain [Pseudoalteromonas piscicida JCM 20779] PabQ 405 isochorismate synthase 97/98 ZP_10289508.1 DHB [Pseudoalteromonas piscicida JCM 20779] biosynthesis PabR 247 ferric iron reductase 92/96 ZP_10289509.1 Regulation [Pseudoalteromonas piscicida JCM 20779]

109

Chapter 4

TE T Cyclisation gly A C Gly - + L + 2 T asp A C or COOH or CONH COOH or or Asp - 2 DOx + L + C R=NH T . Structural pseudoalterobactindifferences. between Structural ??? A acid C Predicted structure: structure: Predicted E + amino + T lys like compound like - A C Lys - + L + C T 2 asn A 2 C Asn - NH 2 + L + R=CH R=NHC(NH)NH

T . KR red AT KS + mal + Pseudoalterobactin A: Pseudoalterobactin B: Pseudoalterobactin T lys A C Lys - T + L +

AL Proposed biosynthesis pathway for a pseudoalterobactin pathway for Proposed biosynthesis

8 . 4

Figure Figure in shown are and structure the predicted

110

Chapter 4

4.4.3.3. The sid biosynthesis gene cluster

The second proposed siderophore biosynthesis pathway, encoded by the sid gene cluster (Table 4.3, Figure 4.9) produces a compound with no structural homology to any known siderophores. The first step in the biosynthesis of this siderophore is the generation of the salicylic acid starter unit. The formation of salicylic acid is catalysed by SidH and SidR. The proposed isochorismate synthase, Sid1H, catalyses the conversion of chorismate to isochorismate (Figure 4. 7). Hydrolysis of the pyruvate side chain of isochorismate, catalysed by SidR, results in the formation of salicylate, the starter unit for the biosynthesis of the sid siderophore. Salicylic acid is activated via the first adenylation domain of SidI, a putative NRPS. The substrate specificity of the first adenylation domain of SidI was unable to be predicted, however, its location at the start of the biosynthesis gene cluster and its proximity to salicylate biosynthesis genes suggests a role in the adenylation of salicylic acid.

Similarly, in P. aeruginosa PAO1, PhcBA encoding for salicylate biosynthesis, is clustered with the genes involved in the initial steps of pyochelin biosynthesis

(Serino et al., 1997).

SidJ-P encode for alternating PKS-NRPS genes, catalysing the incorporation of cysteine, acetate, threonine, acetate, cysteine, acetate and threonine. Numerous examples of mixed NRPS-PKS biosynthesis gene clusters are known, however, the occurrence of six alternating NRPS and PKS genes, which are each encoded on separate genes is unusual. An explanation could be that a non-collinear pathway is followed in the sid cluster.

The cysteine-incorporating C domains of SidI and SidM are proposed to have cyclisation (Cy) functionality. These Cy domains facilitate the formation of a

111

Chapter 4

thiazole ring. This ring is further oxidised to a fully aromatic ring by the action of a putative oxidase, SidM. The PKS, SidL, contains an O-methyltransferase, which gives rise to the methoxylated threonine. The mature hybrid peptide-polyketide chain is released from the peptide chain via a thioesterase domain. The peptide- polyketide chain can be released as a linear acid, or cyclised to form a heterocyclic peptide. In the case of the sid cluster compound a linear peptide is depicted (Figure

4.7), because the site of cyclisation could not be predicted.

Numerous examples of phenolate-containing siderophores with thiazole rings exist, including yersiniabactin (Pfeifer et al., 2003) and pyochelin (Serino et al.,

1997). Like mycobactin (Quadri et al., 1998) and acinetobactin (Mihara et al.,

2004), the predicted product of the sid cluster contains oxazole moieties derived from the cyclisation of serine and or threonine. A collinearly assembled compound encoded by the sid gene cluster is likely to be novel.

112

Chapter 4

Table 4.3 Proposed functions encoded by the sid siderophore biosynthesis gene cluster ORFs

Protein Length Closest homology (BLASTP/nr) Identity/ Accession ID Proposed Similarity function SidA 733 TonB-dependent siderophore receptor 48/67 ZP_05041021.1 Siderophore subfamily regulation [Alcanivorax sp. DG881] SidB 150 L-ornithine 5-monooxygenase 51/71 YP_003978635.1 Regulation [Achromobacter xylosoxidans A8] SidC 139 ExbD 49/71 AAP49339.1 Transport [uncultured bacterium] SidD 241 biopolymer transport protein 62/77 ZP_08569482.1 Transport [Rheinheimera sp. A13L] SidE 90 TonB family protein 60/80 ZP_10297230.1 Transport [Pseudoalteromonas spongiae UST010723-006] SidF 109 alpha-amylase-pullulanase 62/81 BAA05832.1 Regulation [Bacillus sp. XAL601] SidG 535 TonB-dependent receptor 31/49 ZP_10300453.1 Regulation [Pseudoalteromonas spongiae UST010723-006] SidH 238 hypothetical protein C22711_4006 58/72 EGT69974.1 Isochorismate [Escherichia coli O104:H4 str. C227-11] synthase SidI 1740 amino acid adenylation domain protein 39/57 ZP_08494728.1 A-T-Cy-A-T [Microcoleus vaginatus FGP-2] SidJ 1809 KR 42/62 YP_001869920.1 KS-AT-DH- [Nostoc punctiforme PCC 73102] KR-T SidK 1202 Nonribosomal peptide synthase family 47/64 ZP_05029356.1 C-A-T [Microcoleus chthonoplastes PCC 7420] SidL 1878 polyketide synthase module 40/59 ZP_08432360.1 KS-AT-oMT- [Moorea producta 3L] KR-T SidM 1399 nonribosomal polyketide synthase 45/62 YP_001520882.1 Cy-A-Ox-T protein [Acaryochloris marina MBIC11017] SidN 2029 NpnA 46/64 AEU11005.1 KS-AT-DH- [Nostoc sp. 152] ER-KR SidO 94 Dehydrogenase 49/73 ZP_08431750.1 T [Moorea producta 3L] SidP 1424 long-chain-fatty-acid--CoA ligase 42/58 YP_005369901.1 C-A-T [Corallococcus coralloides DSM 2259] SidQ 256 putative gramicidin S biosynthesis 47/64 ZP_08641535.1 TE protein GrsT [Brevibacillus laterosporus LMG 15441] SidR 109 isochorismate-pyruvate lyase 56/79 NP_762772.1 Salicylate [Vibrio vulnificus CMCP6] synthase

113

Chapter 4

. 1 - T thr A S022 - C SA03 - KR T ER AT DH KS T Ox cys A Cy T KR AT oMT

KS T thr A C T KR AT DH KS T cys Hybrid NRPS/PKS biosynthesis pathway for the putative siderophore, for HM Hybrid NRPS/PKSpathway biosynthesis

A 9 . 4 Cy T Figure Figure ? A

114

Chapter 4

4.3.5. Hybrid bacterial iterative type I PKS-NRPS

One of the smallest biosynthesis gene clusters in HM-SA03, alm encodes a hybrid

PKS-NRPS (Table 4.4, Figure 4.10). Genome mining identified a gene encoding a hybrid biosynthesis enzyme consisting of an iterative type I PKS module and an

NRPS module with ornithine specificity. Manual annotation of the genes flanking the hybrid PKS-NRPS identified the presence of two putative phytoene dehydrogenases enzymes and a hydroxylase. AntiSMASH results indicated that these genes, including the hybrid PKS-NRPS were syntenous in numerous genomes, mainly from the phylum Actinobacteria (Figure 4.11). Literature searches revealed that these homologous gene clusters were involved in the biosynthesis of polycyclic tetramate macrolactams (Blodgett et al., 2010). These compounds include HSAF (heat-stable antifungal factor) from Lysobacter enzymogenes strain C3 (Yu et al., 2007), fronalamide from Streptomyces sp. SPB78

(Blodgett et al., 2010) and alteramide A from an unidentified Alteromonas

(Shigemori et al., 1992).

Intriguingly, the PKS modules of the hybrid PKS-NRPS involved in the biosynthesis of these macrolactams were proposed to be a hybrid iterative type I PKS. These

PKSs differ from their modular counterparts because they assemble polyketide chains via a cyclical process, similar to type II PKS systems. Iterative type I PKS are common in fungal polyketide biosynthesis, but relatively uncommon in bacteria.

However a recent study by Clardy and co-workers ( Blodgett et al., 2010) has identified these gene clusters in multiple bacterial genomes, including the frontalamide-producing Streptomyces. The architecture of the alm gene cluster is similar to other gene clusters for frontalamide-like compounds. All of these

115

Chapter 4

clusters consist of a hybrid PKS-NRPS gene encoding for malonyl and ornithine amino acid specificity, at least one phytoene dehydrogenase located downstream from the PKS-NRPS and a hydroxylase. A key difference is the location of the hydroxylase enzyme, which is usually found upstream of the hybrid PKS-NRPS, but is downstream from the hybrid PKS-NRPS in the alm cluster. Additionally, alcohol dehydrogenases and cytochrome P450 monoxygenases, which are present in many gene clusters for frontalamide-like compounds, were absent from the alm gene cluster.

The biosynthesis pathway for frontalamide was recently proposed (Blodgett et al.,

2010) using a combination of gene cluster analysis and biosynthetic precedence from fungal iterative type I PKSs. Due to the similarity of the ftl and alm gene clusters, a similar biosynthetic approach is expected for alteramide A in HM-SA03.

The first iteration of PKS biosynthesis, catalysed by the N-terminus PKS of AltB generates a 12-carbon chain polyketide, the ketones at the 5,7 positions are proposed to be reduced to double bonds by the KR and DH domains of the PKS.

This polyketide chain is condensed to the α-amino of L-ornithine, tethered to the

NRPS thiolation domain. A second iteration of polyketide biosynthesis produces another 12-carbon polyketide chain, with the 3,5 ketone groups reduced to double bonds. This second polyketide chain is condensed to the δ-amino of the ornithine- polyketide chain produced during the first iteration of biosynthesis. This completes the carbon backbone of alteramide A, and the resulting thioester is transferred to the active site serine of the C-terminus thioesterase. A proposed

Dieckmann condensation forms the tetramic acid moiety of alteramide A and subsequently releases the mature peptide-polyketide chain from the enzyme complex. Subsequent formation of the 5-membered rings of alteramide A and 116

Chapter 4

reduction of the polyketide chains of alteramide A is catalysed by AlmCDE. A hydroxylase, AlmF, is proposed to add a hydroxyl group at the α-methylene group adjacent to the tetramic acid moiety. The location of AlmF downstream to the hybrid PKS-NRPS and dehydrogenase genes indicates co-linear biosynthesis, in contrast to the previously characterised frontalamide-type gene clusters where the hydroxylase enzyme is encoded upstream from the hybrid PKS-NRPS.

The identification of an iterative type I PKS module in HM-SA03 represents only the second iterative type I PKS identified in gammaproteobacteria. Shared gene homology between the alm genes in HM-SA03 and other Pseudoalteromonas spp. indicates that these iterative type I PKS may be present in other

Pseudoalteromonas and are not restricted to fungi.

117

Chapter 4

Table 4.4 Proposed functions encoded by the proposed alteramide biosynthesis gene cluster ORFs

Protein Length Closest homology (BLASTP/nr) Identity/ Accession ID Proposed Similarity function AlmA 435 succinyl-diaminopimelate 93/96 ZP_11228314.1 Ornithine desuccinylase metabolism [Pseudoalteromonas flavipulchra] AlmB 3114 polyketide synthase module-like 99/99 ZP_11228315.1 KS-AT-DH-KR- protein T-C-A-T-TE [Pseudoalteromonas flavipulchra] AlmC 551 phytoene dehydrogenase-like 99/99 ZP_11228316.1 Dehydrogenase protein [Pseudoalteromonas flavipulchra] AlmD 178 flavodoxin FldB 98/100 ZP_10288905.1 Unknown/FMN [Pseudoalteromonas piscicida] reductase AlmE 554 phytoene dehydrogenase -like 58/77 ZP_11228318.1 Dehydrogenase protein [Pseudoalteromonas flavipulchra] AlmF 355 hypothetical protein PflaJ_02205 59/76 ZP_11228319.1 Hydroxylase [Pseudoalteromonas flavipulchra] AlmG 444 MATE efflux family protein 55/75 ZP_11228320.1 Transport [Pseudoalteromonas flavipulchra]

118

Chapter 4

TE T orn A C

T . AlmB red KR DH AT KS polyketide chain cyclised internally and internally polyketide chain cyclised - TE T orn membered rings of alteramide A. The structure of alteramide of structure membered A. rings The - A C AlmF T AlmB oduce the 5 oduce the KR 3 Successive rounds of PKS biosynthesis produce partially reduced produce PKSpartially of rounds biosynthesis Successive

5 . DH AT KS TE AlmCDE T orn A C T AlmB KR 5 DH Proposedalteramide pathway Biosynthesis of A

7 AT 10 . 4 KS

Figure ornithine. of amino peptide groups The are condensed polyketides onto the which two cyclisationsprFurther the polyketidereleased. between chains in of frontalamide shown A are structures alteramide and between Differences the frontalamide is shown.

119

Chapter 4

Pseudoalteromonas sp. HM-SA03 almD almB almC almE almF

Streptomyces sp. SPB74

Salinispora arenicola

Saccharophagus degredans

Figure 4.11 Gene map showing gene clusters homologous to the alm gene cluster of Pseudoalteromoas sp. HM-SA03. Genes encoding for the hybrid iterative type I PKS- NRPS (red), phytoene dehydrogenase (green), hydroxylase (grey), alcohol dehydrogenase (blue) and monooxygenase (yellow) are shown. Genes in black are non-homologous.

120

Chapter 4

4.3.6. Other biosynthesis gene clusters encoded in HM-SA03

The remaining three biosynthetic gene clusters identified in the HM-SA03 genome encode for chemical structures that have no similarity to previously identified compounds (Table 4.5). Furthermore, their functional classes could not be identified via interrogation of genes adjacent to these clusters. All three of these gene clusters encode for NRPS biosynthesis, with one being a hybrid PKS-NRPS.

The presence of numerous amino acid residues with carboxylic and amide side chains, which are iron-coordinating groups, in the structures of 1 and 3 (Figure

4.12) suggests their possible roles as siderophores. However, the lack of siderophore and iron regulatory genes offers no support to these predictions.

Furthermore, many of these gene clusters contain adenylation domains with unknown substrate specificities. This ambiguity in adenylation domain substrate prediction arises due to difficulties differentiating similar amino acid side chains

(e.g. aspartate and asparagine) or if the adenylation domain utilises an unusual substrate that has no precedence in other NRPS biosynthesis pathways, such as a non-proteinogenic amino acids. These ambiguous amino acid specificities challenge chemical structure predictions in the HYBR1, NRPS1 and NRPS2 gene clusters (Table 4.5).

121

Chapter 4

Figure 4.12 Chemical structures of HM-SA03 compounds as predicted by bioinformatic analysis of NRPS and PKS gene clusters.

122

Chapter 4

Table 4.5 Domain architectures and predicted products of other NRPS/PKS biosynthesis gene clusters in HM-SA03

Gene Genomic Type Domain architecture Proposed Predicted cluster location product structure HYBR1 scaffold004 Hybrid AL-T-KS-AT-T-AMT-C-A-T Putative 1 93155-124970 PKS-NRPS C-A-T-C-A-T-C-A-T lipopeptide A-T-C-A-T-E-C-A-T-TE NRPS1 scaffold099 NRPS C*-A-T-C-A?-T Pentapeptide 3 25923-42369 C-A-T C-A-T-C-A-T-TE NRPS2 scaffold113 NRPS C*-A-T-C-A-T-C-A-T Heptapeptide 2 43598-66962 C-A-T-C-A-T-C-A-T-C-A-T-TE *Starter condensation domain

123

Chapter 4

4.4. Concluding remarks

The results of this chapter reinforce the significant biosynthetic potential of the genus Pseudoalteromonas for the production of PKS and NRPS-derived small molecules. The identification of seven biosynthesis gene clusters encoding for PKS and NRPS products in HM-SA03 represents a Pseudoalteromonas strain with one of the highest biosynthetic potentials, only equalled by Ps. piscicida JCM20779.

Bioinformatics-assisted structure prediction of the products encoded by these gene clusters has allowed the link between biosynthesis of bromoalterochromide, alteramide and pseudoalterobactin-like compounds, which had previously unknown pathways, to their biosynthetic gene clusters.

Hundreds of new genomes are sequenced every month, however, few of these genomes are investigated for natural product-encoding gene clusters. Similar bioinformatics approaches to those used in this study could be applied to the numerous publically available genomes to connect gene clusters to known compounds with unknown biosyntheses. The identification of such biosynthetic gene clusters could reveal novel enzymology and facilitate the heterologous expression of these gene clusters for the efficient production of pharmaceutically relevant compounds. Furthermore, this study has identified four gene clusters with no known homology to characterised gene clusters and their products would also therefore be novel. The next chapter describes the isolation of the natural products, possibly encoded by these biosynthetic gene clusters from

Pseudoalteromonas sp. HM-SA03.

124

Chapter 5

Chapter 5: Characterisation of Pseudoalteromonas sp. HM-SA03 secondary metabolism

5.1. Introduction

The marine environment is a rich source of natural products (Blunt et al., 2013). In particular, marine bacteria-host symbioses have been shown to be excellent niches for the discovery of new natural products (Piel, 2004). Many soft-bodied molluscs produce natural products as part of a chemical defence system, to increase their survivability (Pawlik, 1993). Dolastatin and tetrodotoxin are natural products found in the sea hare, Dolabella auricularia (Luesch et al., 2002) and

Hapalochlaena sp. (Sheumack et al., 1984), respectively. The biosynthesis of dolastatin has recently been attributed to a cyanobacteria living in association with

D. auricularia (Luesch et al., 2002), and TTX has also been hypothesised to be bacterial in origin, as discussed earlier.

Pseudoalteromonas is a genus of bacteria that is commonly found in association with marine eukaryotes, such as marine plants (Yoshikawa et al., 1997), pufferfish

(Simidu et al., 1990) and octopus (Hwang et al., 1989) and are known to produce bioactive metabolites, for example tambjamines (Lindquist and Fenical, 1991) and

TTX (Ritchie et al., 2000). Pseudoalteromonas are also known to be involved in the production of cell-signalling molecules that influence biofilm development (Davies et al., 1998). The complex microbial communities within biofilms confer advantages to its members such as greater nutrient accessibility and resistance to physicochemical stresses (O'Toole et al., 2000). Such biofilms can consist of

125

Chapter 5

multiple organisms, which are able to communicate with each other via a process known as quorum sensing. These cell density-mediated signalling pathways play an important role in the regulation of gene expression in some microorganisms

(Davey and O'toole, 2000).

A variety of biosynthesis pathways exist for the assembly of natural products, including quorum sensing molecules, the bioactive agent dolastatin, and TTX. Many natural products are derived from NRPS or PKS biosynthetic pathways. Molecular screening of HM-SA03, a Pseudoalteromonas isolated from Hapalochlaena sp., for genes associated with natural product biosynthesis identified numerous NRPS and

PKS gene clusters (chapter 2). Bioinformatics-based genome analysis (chapter 4) identified three biosynthesis gene clusters, putatively encoding the assembly of known natural products, bromoalterochromides, pseudoalterobactins, alterobactins and alteramides. These compounds have been previously isolated from the bacterial genera Pseudoalteromonas or Alteromonas (Deng et al., 1995;

Kanoh et al., 2004; Shigemori et al., 1992; Speitling et al., 2007), however, their biosynthetic pathways were previously unknown. Gene clusters encoding biosynthetic pathways for these compounds were identified in Pseudoalteromonas sp. HM-SA03 isolated from the blue ringed octopus, Hapalochlaena sp. (chapter 4).

This prompted us to investigate the natural products of HM-SA03. Additionally, the characterisation of secondary metabolites in HM-SA03 could identify cell- signalling molecules known to be produced by Pseudoalteromonas. Hence, an investigation into the chemistry of HM-SA03 would reveal clues into its possible role in the production of bioactive compounds and the ecology of Hapalochlaena sp.

126

Chapter 5

5.2. Materials and methods

5.2.1. Culturing and extraction

Liquid cultures of Pseudoalteromonas sp. HM-SA03 (Chapter 2) were grown aerobically at 20°C for 7 d, shaking at 250 rpm in peptone (0.5%; Oxoid) in 0.33

µm-filtered seawater and tryptic soy broth (Oxoid) with 0.2% casein hydrolysate

(Oxoid). Cultures were centrifuged (3000 x g, 10 min) and the supernatant removed. Compounds in the bacterial supernatant were adsorbed with 20 g/L

XAD-7 or XAD-4 resin, shaking at 250 rpm for 3 h (Figure 5. 1). The chromatographic resin was filtered and washed twice with 3 volumes of milliQ water to remove any media constituents. Adsorbed compounds were eluted twice with 3 volumes of HPLC-grade methanol and the combined fractions were dried by rotary evaporation.

For the isolation of siderophore compounds (Chapter 4), Pseudoalteromonas sp.

HM-SA03 was cultured on solid media, which has been shown to promote the production of bromoalterochromide (Speitling et al., 2007). A liquid culture of HM-

SA03 was subcultured on 400 marine agar plates, and incubated at 20°C for 7 d.

Bacterial cells were scraped into 50 mL fa lcon tubes and extracted twice with 2 volumes (w/v) of HPLC-grade methanol and the combined fractions were dried by rotary evaporation (Figure 5.2).

The crude extracts derived from bacterial cultures grown on liquid and solid media were redissolved in methanol and were separated by reverse-phase HPLC or fractionated via size-exclusion chromatography (Figures 5.1 and 5.2) on a

127

Chapter 5

Sephadex LH-20 (Pharmacia) column using methanol as the elutant with a flow rate of approximately 8 mL/min prior to HPLC purification.

Reverse-phase HPLC was performed on a Dionex Ultimate 3000 with a

Phenomenex Sphereclone 5 µm ODS 250 x 10 mm column (analytical) or

Phenomenex Luna 5 µm C18 250 x 4.6 mm column (semi-preparative). Deionised

MilliQ water (0.22 µm filtered; buffer A) and HPLC-grade acetonitrile (buffer B) were used as mobile phases, with flow rates of 1 mL/min and 4 mL/min for analytical and semi-preparative separations, respectively. Monitoring was performed at 210, 256 and 280 nm using a Dionex Ultimate 3000 Diode Array

Detector and charged species were detected using a Corona ultra RS. Fractions were collected using a Dionex Ultimate 3000 Automated Fraction Collector, and were dried by rotary evaporation.

5.2.2. Siderophore assay

The chrome azurol S (CAS) assay was employed for the detection of iron chelating compounds (Schwyn and Neilands, 1987). The assay reagent was prepared by measuring 6 mL of 10 mM hexadecyltrimethylammonium (HDTMA) solution into a

100 mL volumetric flask and diluted with water. A mixture of 1.5 ml iron (III) solution (1 mM FeCl3.6H2O, 10 mM HCl) and 7.5 mL 2 mM aqueous CAS solution was slowly added under stirring. Anhydrous piperazine (4.307 g) was dissolved in water and 6.25 mL of 12 M hydrochloric acid was carefully added. This solution was rinsed into the volumetric flask that was then filled with water to afford 100 mL of CAS assay solution.

128

Chapter 5

Methanol and acetonitrile produce a colour change when mixed with CAS assay solution, therefore, compounds to be tested (0.1 mg) were transferred to 1.5 mL tubes and dried on a Savant Speed Vac DNA110 before adding the assay solution.

To these dried compounds, 0.2 mL of CAS assay solution was added and mixed by vortex. Colour change in the assay was observed, manually, after 30 min. A colour change from blue to yellow indicated a strongly positive result, while samples with no observed colour change were recorded as inactive.

5.2.3. Nuclear magnetic resonance spectroscopy

Crude and purified fractions were redissolved in deuterated methanol, dimethylsulfoxide or chloroform (Cambridge Isotope Laboratories). All NMR spectra were recorded on either a Bruker Avance III 300 (300 MHz for 1H, 75 MHz for 13C) or Bruker Avance III 600 (600 MHz for 1H, 150 MHz for 13C). Spectra were referenced to their solvent signals based on literature data for these solvents

(Gottlieb et al., 1997).

129

Chapter 5

Tryptone Soy Broth with 0.2% Caesin Hydrolysate

XAD-7 MeOH Extract

No activity

Sephadex LH-20

Fraction 4 Fraction 5 58 min 65 min

RP-HPLCa RP-HPLCb RP-HPLCc

cyclo(Phe-Pro) cyclo(Val-Val) cyclo(Phe-Val) cyclo(Tyr-Pro) 13.6 min 15.8 min 15.8 min 9.8 min 28.0 mg 0.9 mg* 0.9 mg* 1.1 mg

0.2% Peptone in Seawater

XAD-4 MeOH XAD-7 MeOH Extract Extract

No activity No activity

SepPak C18

Fraction 3 40% MeOH 16.2 mg RP-HPLCf RP-HPLCd RP-HPLCe

cyclo(Phe-Pro-OH) cyclo(Phe-Pro-OH) cyclo(Val-Pro-OH) cyclo(Val-Pro-OH) 11.7 min 15.7 min 23.3 min 23.3 min 4.4 mg 1.8 mg 3.1mg* 1.8 mg*

Figure 5.1 Isolation scheme for HM-SA03 grown in liquid media and analysed by RP-HPLC.a,b,c5-100% B over 30 min,† d20-60% B over 30 min,† e,f5-9% B over 15 min, held for 10 min. Ramped to 15% over 15 min.† Buffers A and B are water and acetonitrile, respectively. †Gradients were followed by a wash and equilibration: Ramped up to 100% B over 5 min and held for 10 min. %B was lowered to the initial starting conditions over 5 min, and held for 10 min to equilibrate the column. *These compounds were isolated as mixtures. Mass amounts were estimated based on ratio of NMR signals. “No activity” indicates a lack of detectable colour change in the CAS siderophore assay.

130

Chapter 5

Marine Agar

MeOH Extract

Sephadex LH-20

No activity Strong siderophore No activity

Fraction 2 Fraction 1 Fractions 3-5 13-19 min

RP-HPLC

No activity Strong siderophore No activity

Fraction 2 Fraction 1 Fraction 3-5 12-24 min

RP-HPLCa

Insufficient quantities for futher structural elucidation

Figure 5.2 Isolation scheme for HM-SA03 grown on solid media and analysed by RP-HPLC. a The solvent gradient was 25-35% acetonitrile/water (buffer B/A) over 10 min, then ramped to 50% B over 5 min, then increased to 60% over 45 min. Following this, the gradient was ramped up to 100% B over 5 min and held for 10 min. %B was lowered to the initial starting conditions over 5 min, and held for 10 min to equilibrate the column. “No activity” indicates a lack of detectable colour change in the CAS siderophore assay.

131

Chapter 5

5.3. Results and discussion

5.3.1. Siderophores from Pseudoalteromonas sp. HM-SA03

The presence of seven NRPS, PKS or NRPS/PKS hybrid biosynthesis gene clusters in the genome of Pseudoalteromonas HM-SA03 (chapter 4) prompted the investigation of the chemical diversity of this strain. Some of these gene clusters were proposed to encode siderophore biosynthesis, including a putative bromoalterochromide biosynthesis cluster (chapter 4). Bromoalterochromides have been previously isolated from Pseudoalteromonas (Sobolevskaya et al., 2005;

Speitling et al., 2007). The fermentation of Pseudoalteromonas sp. HM-SA03 on marine agar plates, which has been shown to enable bromoalterochromide production (Speitling et al., 2007), led to the isolation of a dark brown extract.

Biochemical analysis using the chrome azurol S assay, which is a colourimetric assay for siderophore activity (Schwyn and Neilands, 1987), indicated the presence of iron chelating compounds. Addition of the CAS assay solution to 0.1 mg of the dried extract produced a blue to yellow colour change. Furthermore, addition of CAS assay solution to yellow-pigmented extracts, which did not have iron-chelating compounds, did not produce a colour change. Subsequent purification of this CAS-positive extract yielded fractions with insufficient quantities for further NMR and biochemical analysis.

5.3.2. Diketopiperazines from Pseudoalteromonas sp. HM-SA03

During HPLC separation, many compounds with strong absorbance at 280 nm were observed in the XAD7 and XAD4 culture supernatant crude extracts.

Purification and subsequent structure elucidation of these compounds by NMR

132

Chapter 5

revealed eight diketopiperazines (Figure 5. 3). Identification of these structures as diketopiperazines was simplified by the characteristic 13C NMR chemical shifts of the two amide carbons (160-170 ppm) and their corresponding methines (~55 ppm). The complete structure elucidation, biosynthesis and significance of these compounds are discussed in the following sections.

Table 5.1 NMR data for cyclo(D-4-hydroxyprolyl-L-phenylalanyl), 2, isolated from the extract of HM-SA03, a bacterium associated with Hapalochlaena sp.

2, cyclo(D-4-hydroxyprolyl-L-phenylalanyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings COSY HMBC [Hz] 58.2 4.24 3 1 t 5.1x(2) H10a, H10b C10, 135.4, C2, C5 39.5 3.03 10a 1 dd 14.4, 5.1 H10b, H3 C3, C13, 135.4, C2 39.5 3.21 10b 1 dd 14.4, 5.1 H10a, H3 C3, C13, 135.4, C2 55.8 2.81 6 1 t 8.7x(2) H7a, H7b C7, C5 127.1 7.33 14 1 m H13, H12 C12, C13, C11 128.3 7.32 12, 12' 2 m H13, H14 C10, C14, C12, C13 129.8 7.21 13, 13' 2 m H12, H14 C13 36.6 1.93 7a 1 ddd 14.4, 8.7, 5.8 H7b, H6, H8 C9, C6, C8, C5 36.6 2.25 7b 1 m 14.4, 8.7, 5.8 H7a, H6, H8 C9, C6, C8, C5 67.1 4.25 8 1 t 5.8x(2) H7a, H7b, H9a, H9b C9, C6, C2, C5 52.6 3.30 9a 1 dd 12.9, 6.2 H9b, H8 C7, C2* 52.6 3.62 9b 1 dd 12.9, 3.9 H9a, H8 C7, C6, C8, C2* 135.4 11 166.3 2 169.6 5 * Indicates a weak HMBC

133

Chapter 5

cyclo(L-prolyl-L-tyrosyl); 1 cyclo(D-4-hydroxyprolyl-L-phenylalanyl); 2

cyclo(L-4-hydroxyprolyl-L-leucyl); 3 cyclo(L-4-hydroxyprolyl-D-leucyl); 4

cyclo(L-leucyl-L-leucyl); 5 cyclo(L-leucyl-L-phenylalanyl); 6

cyclo(L-prolyl-L-phenylalanyl); 7 cyclo(L-4-hydroxyprolyl-L-phenylalanyl); 8

Figure 5.3 Diketopiperazines isolated from Pseudoalteromonas sp. HM-SA03.

134

Chapter 5

5.2.3.1. Structure elucidation

The diketopiperazine, compound 2 , was isolated as a yellow solid. The 1H and 13C

NMR spectra (Table 5.1) indicated the presence of 14 hydrogen atoms and 14 carbon atoms, of which two were equivalent sets of carbon atoms (C12, 13). A phase-sensitive, heteronuclear single quantum coherence experiment (HSQC), which identifies single-bond 1H-13C correlations, revealed three methylenes and eight methines.

Correlation spectroscopy (COSY), identifies 1H-1H correlations across two (e.g. same carbon) and three (e.g. adjacent carbons) bonds while Heteronuclear multiple-bond correlation spectroscopy (HMBC) correlations, identifies 1H-13C correlations across two or more bonds. These experiments allowed the assignment of five downfield methines (H12-H14) to a benzyl ring. HMBCs from H12-C10 and

H10 and H3 to C11, combined with COSYs between the H3 methine and H10 methylene allowed the assignment of the benzyl fragment (Figure 5.4).

COSY was used to determine the connectivity of the remaining two methylenes (C7 and C9) and two methines (C6 and C8; figure 5.5). The downfield carbon shift of the methine at C8 was consistent with a hydroxyl group at that position. The chemical shifts of C3, C6 and C9 (58.2, 55.8, 52.6 ppm, respectively), along with two downfield carbon shifts (166.3, 169.6 ppm) alluded to the presence of two amides. Connectivity between these fragments was achieved via HMBC analysis

(Figure 5.6). A search of chemical databases of the proposed structure, 2, confirmed that the NMR data was consistent with literature values for the compound, cyclo(D-4-hydroxylprolyl-L-phenylalanyl).

135

Chapter 5

H10a/H10b

H3/H10b

H3/H10a

H12,13,14

Figure 5.4 Connectivity of the benzyl group of the diketopiperazine, compound 2. Bolded bonds represent COSY correlations. Signals off the diagonal indicate that H10a is correlated (they are on the same or adjacent carbon atoms) to H10b, and H3. Similarly, H10b is correlated to H10a and H3. The magnified insert shows the connectivity of the phenyl ring (C12, 13, 14). Arrows represent HMBC correlations, indicating the connectivity of the phenyl ring to H10a/b and H3.

136

Chapter 5

A similar approach was also used for the structural elucidation of the other diketopiperazines (compounds 1, 3-8) isolated in this study (Table 5. 2). Literature values confirmed the isolation of antimicrobials, cyclo(L-4-hydroxyprolyl-L- leucyl), 3, (Furtado et al., 2005; Ienaga et al., 1987), and cyclo(L-4-hydroxyprolyl-

L-phenylalanyl), 8, (Cronan Jr et al., 1998; Furtado et al., 2005), cell-signalling molecules, cyclo(L-prolyl-L-tyrosyl), 1, (Jayatilake et al., 1996; Rudi et al., 1994) and cyclo(L-prolyl-L-phenylalanyl), 7, (Furtado et al., 2005), plant-growth promotor, cyclo(L-4-hydroxyprolyl-D-leucyl), 4, (Cronan Jr et al., 1998), and diketopiperazines with unknown function, cyclo(D-4-hydroxyprolyl-L- phenylalanyl), 2, (Shigemori et al., 1998), cyclo(L-leucyl-L-leucyl) , 5, (Exner and

Kostelnik, 1977) and cyclo(L-leucyl-L-phenylalanyl) , 6, (Kanzaki et al., 2000).

H7a/H7b

H7b/H6 H7a/H6

H9a/H9b

H9b/H8 H9a/H8 H7b/H8 H7a/H8

Figure 5.5 Connectivity of the 4-hydroxyprolyl group of the diketopiperazine, compound 2. Bolded bonds represent COSY correlations.

137

Chapter 5

H8/C2 H9'/C2 H9/C2

H10'/C2 H10/C2

H8/C5 H6/C5 H7'/C5 H7/C5

Figure 5.6 Connectivity of the diketopiperazine ring to the benzyl and hydroxyprolyl groups of the diketopiperazine, compound 2. Arrows represent

HMBC correlations. H8, 9, 10 are located 2 or 3 bonds away from C2, and H6, 7, 8 are located 2 or 3 bonds away from C5.

138

Chapter 5

Table 5.2 NMR data for compounds 1, 3-8, isolated from the extract of HM-SA03, a bacterium associated with Hapalochlaena sp.

1, cyclo(L-prolyl-L-tyrosyl); NMR (CDCl3) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 28.14 1.88 7a 1 m H7b , H6 C8 22.14 1.91 8a 1 m H9 C6, C5 22.14 2.02 8b 1 m H9 C7 28.14 2.33 7b 1 m H7a, H6 C8, C9, C6 35.91 2.89 10a 1 dd 14.5, 8.9 H10b, H3 C3, C9, C12, C2 35.91 3.40 10b 1 dd 14.5, 4.0 H10a, H3 C3, C9, C12, C2, C5 45.49 3.60 9 2 m H8a, H8b C8, C7, C6, C2* 58.92 4.11 6 1 br dt 8.1 H7a, H7b C7, C5 56.38 4.30 3 1 ddd 9.1, 3.9, 1.4 H10a, H10b C10, C9, C2, C5 115.79 6.80 13 2 m H12 C13, C9, C12, C14 130.63 7.07 12 2 m H13 C10, C13, C12, C14 126.2 9 155.1 12 165.2 2 170.3 5 - 6.64 4-NH 1 br s H6*, H3* C10, C3, C6, C2, C5

3, cyclo(L -4-hydroxyprolyl -L - leucyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 20.79 0.98 12 3 dd 6.4 3.6 H10a, H11 C12, C13, C11, C10 21.9 0.98 13 3 dd 6.4 3.6 H10a, H11 C12, C13, C11, C10 38 1.54 10a 1 ddd 13.5, 7.8, 5.2 H12, H13, H10b, H3 C12, C13, C11, C3, C2 24.38 1.90 11 1 m H12, H13 C12, C13, C10, C3 38 1.94 10b 1 m H10a, H3 C12, C13, C11, C3, C2, C5 36.76 2.11 7a 1 ddd 13.3, 11.2, 4.3 H7b, H8, H6 C6, C8, C5 36.76 2.30 7b 1 ddt 13.3, 6.4, 0.9x(3) H7a, H9a, H8, H6 C9, C6, C8, C2 53.76 3.47 9a 1 m H7b, H9b C7, C6, C8, C2 53.76 3.68 9b 1 m H9a, H8 C7, C6, C2 53.21 4.19 3 1 m H10a, H10b, H6 C10, C2 67.71 4.49 8 1 br t 4.1x(2) H7a, H7b, H9b C9, C6 57.3 4.54 6 1 ddd 11.1, 6.6, 1.1 H7a, H7b, H3 C7, C5 167.63 2 171.65 5

139

Chapter 5

4, cyclo(L-4-hydroxyprolyl-D-leucyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 20.53 0.98 12 3 dd 6.4, 3.6 H7b C10 21.91 1.01 13 3 m 6.6 H11 C12, C11, C10 41.98 1.60 10a 1 ddd 13.7, 8.4, 5.5 H10b, H11, H3 C12, C13, C11, C3 41.98 1.71 10b 1 ddd 13.7, 9.6, 5.6 H3, H10a C12, C13, C11, C3 24.18 1.80 11 1 m 6.8x(6), 1.7 H13, H10a C12, C13, C10 36.28 2.25 7a 1 quinq 6.4x(4), 1.1x(3) H7b, H6, H8 C9, C6, C8, C5 36.28 2.50 7b 1 ddd 13.6, 8.6, 5.5 H12, H13, H7a, H6, H8 C9, C6, C8, C5 52.85 3.47 9a 1 m H8 C7, C2 52.85 3.68 9b 1 m H8 C7, C6, C8, C2 55.53 3.89 3 1 dd 9.7, 5.5 H10a, H10b C11, C10, C5 56.15 4.37 6 1 bt t 8.0x(2) H7a, H7b C7, C5 67.45 4.44 8 1 quind 4.9x(4), 1.2 H7a, H7b, H9a, H9b C9, C6 168.04 2 169.84 5

5, cyclo(L -leucyl -L-leucyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 20.39 0.96 6, 6' 6 d 6.6 H5 C7, C5, C4* 22.06 0.98 7, 7' 6 d 6.6 H5 C6, C5, C4* 44.39 1.62 4a, 4a' 2 ddd 13.7, 9.0, 5.3 H4b, H3 C6, C5, C4*, C3* 44.39 1.71 4b, 4b' 2 m 13.7, 8.7, 4.3 H4A, H3 C6, C5, C4*, C3* 23.85 1.84 5, 5' 2 m H6, H7, H4b 53.19 3.91 3, 3' 2 dd 9.0, 4.6 H4A, H4b C5, C4, C2 169.70 2, 2'

6, cyclo(L -leucyl -L-phenylalanyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSy HMBC 43.76 0.07 7a 1 ddd 13.6, 9.6, 5.1 H7b, H6 C8, C6, C5* 21.86 0.69 10 3 d 6.7 H8 C9, C8, C7* 19.87 0.73 9 3 d 6.7 H8 C10, C8, C7* 43.76 0.87 7b 1 ddd 13.7, 9.5, 4.4 H7a, H8, H6 C8, C5 23.14 1.42 8 1 dquind 9.5, 6.6x(4), 5.1 H10, H9, H7b 38.75 2.95 11a 1 dd 13.8, 4.8 H11b, H3 C3, C13, C12*, C2* 38.75 3.28 11b 1 dd 13.9, 3.8 H11a, H3 C3, C13, C12* 52.62 3.66 6 1 ddd 9.8, 4.4, 0.7 H7a, H7b C8, C7, C2* 55.89 4.31 3 1 td 4.3x(2), 0.9 H11a, H11b C5 130.31 7.20 13 2 m H14 C11, C15, C13* 127.00 7.27 15 1 m C13 128.16 7.31 14 2 m H13 C14, C12 135.15 12 167.24 2 169.14 5

140

Chapter 5

7, cyclo(L-prolyl-L-phenylalanyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 28.00 1.26 7a 1 m H7b, H10, H6 C3, C8 21.39 1.83 8 2 m H7, H7b, H9a, H9b C7, C3, C9 28.00 2.12 7b 1 m H7, H8, H6 C9, C3 36.77 3.20 10 2 d 5.0 H3 C6, C14, C11, C2, C5 44.55 3.41 9a 1 dd 12.3, 6.4 H8, H9b C8, C7, C3 44.55 3.57 9b 1 t 8.2x(2) H8, H9A C8, C7, C3 56.29 4.09 6 1 ddd 10.8, 6.4, 1.8 H7, H7b C7, C5, C8 58.69 4.47 3 1 br t 5.0 H10 C10, C11, C2, C5 126.69 7.26 12 2 m C10, C12 128.07 7.27 13 2 m C12, C13, C14, C11 129.66 7.29 14 1 m C12, C11 135.98 11 165.52 2 169.53 5

8, cyclo(L -4-hydroxyprolyl -L - phenylalanyl); NMR (CD3OD) Shift [ppm] Pos #Hs Multiplet Structure Connectivity Correlations 13C 1H Type J-Couplings [Hz] COSY HMBC 38.0 1.54 7a 1 ddd 13.0, 11.3, 4.5 H7b, H8, H6 C6, C5, C8* 38.0 1.96 7b 1 dd 12.9, 6.4 H7a, H6 C8, C9 35.1 3.08 10 2 dd 9.2, 5.0 H3 C12, C11, C3, C2* 54.3 3.18 9a 1 d 12.6 H9b C7*, C6*, C8 54.3 3.54 9b 1 dd 12.6, 4.9 H9a, H8 C7, C6, C2 67.1 4.21 8 1 br s H7a, H9b C9, C6 57.3 4.32 6 1 ddd 10.8, 6.1, 1.0 H7a, H7b C7*, C5 56.1 4.42 3 1 br t 5.0x(2) H10 C2, C11, C5*, C10 126.8 7.24 14 1 m# C12 128.5 7.26 12 2 m# C10, C14, C12 128.2 7.30 13 2 m# C12*, C11, C13 137.7 11 165.6 2 169.9 5 5.10 8-OH 1 d 3.0 H8 8.00 4-NH 1 s C2, C6 # 1H signal could not be resolved; assignments are based on 2D correlations * Indicates a weak correlation

141

Chapter 5

5.2.3.2. Biosynthesis pathways

Two pathways exist for the biosynthesis of diketopiperazines (DKP). The first is the NRPS-dependent pathway, similar to those discussed in chapter 4. These NRPs consist of two biosynthetic modules that serve to incorporate two amino acids into the DKP structure. Unlike traditional NRPS, these dimodular enzymes lack a C- terminus TE domain, which is usually responsible for release of the final nonribosomal product (Balibar and Walsh, 2006). Instead, these DKP-dedicated NRPSs contain a C-terminus condensation domain, which is believed to catalyse the formation of the N-C bond generating the DKP ring (Belin et al., 2012).

DKPs may also be a byproduct from the biosynthesis of NRPS-derived natural products. The production of cyclo(D-phenylalanyl-L-prolyl) has been observed during the biosynthesis of tyrocidine (Stachelhaus et al., 1998). The early steps of tyrocidine biosynthesis involve the formation of a linear L-prolyl-L-phenylalanyl product, which is further elongated and cyclised to produce tyrocidine, an antibiotic. The formation of such DKPs arise via a spontaneous cyclisation of the L- prolyl-L-phenylalanyl linear peptide (Stachelhaus et al., 1998).

The biosynthesis of these DKPs can also be catalysed by a newly characterised class of enzymes, cyclic dipeptide synthases (CDPS) (Figure 5.7). This NRPS- independent pathway for the biosynthesis of DKPs was first discovered in

Streptomyces noursei in 2002 (Lautru et al., 2002). The first step in the biosynthesis of albonoursin involves the formation of the DKP, cyclo(L- phenylalanyl-L-leucyl), catalysed by the CDPS, AlbC. Similar enzymes have also been characterised in Bacillus licheniformis (Bonnefond et al., 2011) and

Myobacterium tuberculosis (Vetting et al., 2010). These enzymes catalyse the

142

Chapter 5

transfer of an aminoacyl moiety from aminoacyl-tRNA to its active site. The enzyme then interacts with a second aminoacyl-tRNA and catalyses peptide bond formation between the first and second aminoacyl moieties. This releases the tRNA from the enzyme along with the cyclised DKP product. Binding of the first tRNA substrate has been experimentally characterised, however, the interaction of the second tRNA substrate and peptide bond formation has yet to be experimentally characterised (Belin et al., 2012).

A search for either NRPS or CDPS-catalysed DKP biosynthesis pathways was conducted in HM-SA03. Genome mining for NRPS biosynthesis genes failed to reveal any dimodular NRPS clusters. Furthermore, examination of the amino acid specificities of adenylation domains in the 7 NRPS-containing clusters identified in

HM-SA03 did not identify any biosynthetic pathways where the spontaneous cyclisation of linear dipeptide intermediates would lead to the formation of DKPs isolated in this study. The search for a CDPS-catalysed pathway yielded similarly disappointing results. A HMM search for these enzymes using a custom pHMM made from three experimentally characterised DKPs also yielded no results (data not shown). The lack of these biosynthesis genes in HM-SA03 may arise through a lack of genomic coverage, since CDPS are generally very small and are encoded by only 239 amino acids (Lautru et al., 2002). Furthermore, molecular screening for these enzymes is made more difficult by the limited number of experimentally characterised CDPS and the high level of degeneracy between CDPS amino acid sequences. Additionally, the diversity of DKPs observed in this study may be biosynthesised entirely by a single CPDS. Studies of DKP production in recombinant E. coli strains expressing CDPSs showed that each enzyme was able to produce several different DKPs (Gondry et al., 2009; Seguin et al., 2011). 143

Chapter 5

Leu Phe CDPS bond formation bond CDPS-catalysed CDPS-catalysed DKP tRNA - aa O Leu Phe CDPS substrate

Interaction between secondbetween Interaction

aminoacyl-tRNA and and CDPS-bound aminoacyl-tRNA Phe CDPS complex CDPS-amino acid CDPS-amino diketopiperazine biosynthesis - tRNA - aa O Phe tRNA and and CDPS tRNA - CDPS The CDPS pathway for CDPS2,5 pathway for The

7 . 5 Interaction between Interaction aminoacyl

Figure

144

Chapter 5

5.2.3.3. Ecological roles

Many DKPs have been shown to have numerous biological activities. Compounds 3,

7, 8 have been shown to have activity against Staphylococcus aureus and

Micrococcus luteus (Furtado et al., 2005), while compounds 1 and 7 have been shown to have antimicrobial activity towards Gram-negative bacteria (Milne et al.,

1998). Interestingly, compound 8 has been shown to stimulate antibiotic production in Pseudoalteromonas luteoviolacea (Jiang et al., 2000).

Although DKPs have been shown to have antimicrobial activity, this may not be the main role of these compounds in the environment. Studies have shown that many

DKPs, mainly cyclic dipeptides, function as cell-signalling molecules (Degrassi et al., 2002; Holden et al., 1999). Quroum sensing is the use of small molecules, in bacteria, to regulate gene expression when microbial population density reaches a threshold, and typically occurs during exponential phase growth. The most studied class of signalling molecules used as quorum sensors are N-acylhomoserine lactones (AHLs). However, studies have shown that DKPs cyclo(L-prolyl-L-tyrosyl) and cyclo(L-prolyl-L-phenylalanyl) activate LuxR-based AHL biosensors (Degrassi et al., 2002; Holden et al., 1999). Furthermore, these compounds also interact with the mammalian central nervous system, suggesting a possible role in bacterial-host interaction (Prasad, 1995).

The role of DKPs as signalling molecules is supported by their biosynthetic machinery. NRPS are relatively large enzymes, while CDPS are comparatively small

(Belin et al., 2012). Therefore, the production of DKPs for cell signalling via a CDPS would more economical compared to an NRPS route. Furthermore, DKPs derived from CDPS are produced during exponential phase growth (Gondry et al., 2009),

145

Chapter 5

possibly due to the high availability of tRNAs. This growth period coincides with maximal change in microbial density and therefore, activation of quorum sensing pathways.

As discussed in chapter 2, the microbial environment from which

Pseudoalteromonas sp. HM-SA03 was isolated is quite complex. Therefore, it is possible that DKPs are produced either as an antimicrobial to increase the competitive fitness of the bacterium, or as a signalling molecule to respond to changes in the microbial environment.

The production of DKPs by HM-SA03 may also be a mechanism for interaction with the Hapalochlaena sp. host. Instances of DKP mediated interactions between bacteria and eukaryotes have been observed. The DKP, cyclo(histidyl-prolyl) is known to act as a neurotransmitter in mammalian cells (Prasad, 1995), while compounds 3, 4 and 8 have been shown to function as plant growth promoters

(Cronan Jr et al., 1998). The precise reason for DKP production in bacteria is unknown, however, it is clear that these molecules have diverse roles in nature.

146

Chapter 5

5.4. Concluding remarks

This study has characterised the most abundant small molecule compounds produced by Pseudoalteromonas sp. HM-SA03. However, it was not possible to confirm the biosynthesis of the siderophores via genomic prediction. The expression of NRPS or PKS biosynthetic pathways was not seen, possibly since many have low yields or may be entirely unexpressed in the native

Pseudoalteromonas host. Such “cryptic” or “silent” gene clusters may require expression in a heterologous host to enable the production of their encoded natural products (Ongley et al., 2013). Although no NPRS or PKS products were detected, eight DKPs from HM-SA03 were isolated. To our knowledge, the native production of such a wide diversity of DKPs is unprecedented in any microorganism. These DKPs are known antimicrobials and may also be involved in cell signalling, which suggests a mutualistic or commensal relationship between

HM-SA03 and its Hapalochlaena sp. host.

147

Chapter 6

Chapter 6: General discussion

6.1. Research motivation and objectives

The ultimate goal of natural products research is the discovery of new drug candidates for therapeutic use. An understanding of the biosynthetic processes that lead to these natural products is a key aspect of natural products research.

Knowledge of how biosynthesis pathways operate in nature can be a source of inspiration for the biomimetic synthesis of new natural products, or modification of known compounds. Additionally, the discovery of novel biosynthetic enzymes that give rise to unusual chemical structures can add to our repertoire of knowledge for the combinatorial engineering of new natural product pathways.

Therefore, we investigated the possible biosynthetic origins of a biosynthetic enigma, TTX, which contains an unusual chemical structure (described in chapter

2). We then investigated the possibility of difficult-to-culture microbes as the true producer of TTX (chapter 3). Concurrent to this, a microbe, HM-SA03, from a genus known to produce bioactive compounds and rich in natural product biosynthesis pathways was isolated from the posterior salivary gland of a Hapalochlaena octopus. This prompted further molecular and bioinformatic investigation of these pathways, facilitating the assignment of these pathways to the biosynthesis of several known compounds (chapter 4). In turn, this led us to the investigation of natural products produced by the isolate, HM-SA03 (chapter 5), and a discussion of its secondary metabolism in the context of its symbiotic habit.

148

Chapter 6

6.2. Key findings

6.2.1. Biosynthesis of TTX by laboratory-cultured bacteria

A TTX-producing bacterium was unable to be isolated from either Hapalochlaena sp. or P. maculata, both of which are reported to produce TTX (chapter 2). This conflicts with the current literature, where multiple studies have shown that TTX- producing bacteria are readily culturable from their host organisms (Table 2.2).

However, many of these studies were performed before technical advances that led to the development of modern spectrometric methods for the detection of toxins and other compounds. This prompted us to scrutinise the results of other studies of TTX biosynthesis in bacteria. Surprisingly, we were unable to replicate the biosynthesis of TTX by any of the purported “TTX-producing” strains that we sourced (chapter 2).

Biosynthetic precedent indicated that NRPS or PKS may be involved in the biosynthesis of TTX (chapter 1). Therefore, we sought to identify whether these genes were present in bacteria isolated from Hapalochlaena sp. and P. maculata.

Although the culturing conditions may have been unsuitable for the biosynthesis of

TTX ex situ, genes encoding its biosynthesis should still be identifiable. Genetic screening identified many bacteria possessing NRPS and PKS genes however their role in TTX biosynthesis was unable to be confirmed. One bacterium,

Pseudoalteromonas sp. HM-SA03 was selected for further studies based on its rich biosynthetic potential.

149

Chapter 6

6.2.2. Biosynthesis of TTX by uncultured microorganisms

Studies into the biosynthesis of TTX in bacteria have focused exclusively on microorganisms that were culturable under laboratory conditions. This may have been due to technical deficiences, as many of these studies were performed prior to the advent of high-throughput sequencing technologies for investigation of unculturable microorganisms. Nevertheless, such a large gap in scientific knowledge required further exploration. To this extent, we utilised molecular methods to identify the total microbial consortium of P. maculata (chapter 3). Prior to this work, such methodology had not been applied to TTX-producing animals in order to identify a TTX-producing microorganism. Combined with data on TTX concentration in each of our P. maculata samples, we were able to identify four taxa that were strongly positively correlated with TTX concentration. Of these, two were unable to be taxonomically classified, while two could only be classified to the family level (Rhodobacteriaceae). Although these taxa could not be directly cultured and hence, linked to TTX biosynthesis, such a molecular approach could be readily applicable to understanding the origin of TTX in other organisms or for other natural products.

6.2.3. Genome mining as a tool for natural product discovery

The identification of a Pseudoalteromonas bacterium with high biosynthetic potential was of great interest, as members of this genus are known to produce interesting bioactive compounds ( Bowman, 2007), and are candidate TTX- producers. Many of these natural products had not yet been linked to a biosynthetic gene cluster. Therefore, we sequenced the genome of this bacterium with the aim of discovering new biosynthesis pathways for novel natural products,

150

Chapter 6

or linking the biosynthesis of known natural products to a biosynthesis gene cluster. Ultimately, seven biosynthesis gene clusters were identified and the chemical structures assembled by each of these pathways were predicted. Of these, four of the predicted structures were novel, while three of the predicted compounds were known compounds without previous biosynthetic assignment.

The genomic information allowed the biosynthesis of (bromo)alterochromide, alteramide and pseudoalterobactin-like compounds to be proposed. Of particular interest was the biosynthesis of alteramide by an iterative type I PKS, which in bacteria are more rare than their non-iterative counterparts. Furthermore, many of these compounds are produced in low quantities in Pseudoalteromonas

(Bowman, 2007). This study facilitates the future heterologous expression of these gene clusters to improve yield, and therefore, their applicability as pharmaceutical drug leads, such as new antimicrobials and siderophores.

6.2.4. Identification of Pseudoalteromonas-derived secondary metabolites

Attempts to isolate natural products assembled by biosynthetic gene clusters identified in HM-SA03 did not reveal any PKS or NRPS products. However, eight diketopiperazines were identified, of which six had not been reported previously from Pseudoalteromonas species. The identification of eight diketopiperazines, produced natively, in a single microorganism is unprecedented.

Pseudoalteromonas was isolated from a blue ringed octopus containing numerous other Pseudoalteromonas species. These species are known to form complex microbial communities and are known for their cell-cell signaling mechanisms. The identification of numerous cell-signalling molecules confirms the complex microbial community present in Hapalochlaena and possibly other marine

151

Chapter 6

organisms. Furthermore, the discovery of numerous DKPs may represent an opportunity to study quorum sensing by non-AHL molecules.

6.3. Future directions

6.3.1. Re-examination of TTX-production in laboratory cultures

Organisms spanning eight phyla across two kingdoms of life are known to harbour

TTX (Chau et al., 2011). Such a staggering diversity of producers is unprecedented for any natural product. Previous studies (Kim et al., 1975) have speculated on whether there is a common food source that is widely distributed enough whereby all the different TTX-containing organisms could access it. Under this scenario the most likely producers would be simple single cell organisms, such as bacteria and algae. Since the discovery of TTX, we have seen the isolation of numerous bacteria that are purportedly able to biosynthesise TTX. Therefore, the bacterial origin of

TTX became widely accepted in the scientific community. However, despite several attempts to identify microbial producers, some tetrodotoxic taxa, including newts, have not yielded any microbial candidates for TTX production. Furthermore, modern and more specific TTX -detection methods were unable to verify the production of TTX from published research (chapter 2). In consideration of the above, the number and diversity of TTX-producing bacteria reported in the literature may have been overestimated. Therefore, a reexamination of TTX- production, using modern and highly specific TTX detection methods, such as mass spectrometry and NMR, from all published studies should be pursued.

152

Chapter 6

6.3.2. Comparison of microbial communities in all TTX-producing higher organisms

In consideration of the ambiguity surrounding TTX-production from cultured bacteria, a logical progression in TTX research is an investigation of the total microbiota, including laboratory-unculturable microbes, in all known TTX- producing animals. The results from chapter 3 have shown that such an approach is both scientifically feasible, and cost-effective. An investigation encompassing all known TTX-producers is highly ambitious, therefore a focus on organisms at lower trophic levels may be desirable. These organisms are more likely to be true TTX producers, as bioaccumulation of TTX through diet is minimised. The analysis of microbial abundance data, combined with information on TTX-concentrations in these organisms and other metadata, such as environmental variables, would allow us to identify trends that may be relevant to TTX-biosynthesis. A common microbial vector, if existent, might be readily identified in these organisms. Using modern cell-sorting techniques coupled with advances in single-cell genome amplification and sequencing ( Kalisky and Quake, 2011), these microbes can be isolated for culture-based investigations or genome sequencing. Furthermore, analysis of the environmental conditions of these habitats may give us clues into abiotic factors that affect TTX-production and promote the production of TTX under laboratory conditions. Such an investigation would greatly further our understanding of the biogenic origins of TTX.

153

Chapter 6

6.3.3. Heterologous expression of HM-SA03 gene clusters

Genomic analyses identified seven natural product gene clusters encoding for

NRPS and PKS products. However, no natural products deriving from these gene clusters were isolated, possibly due to the low, or non-expression of these gene clusters. Therefore, the heterologous expression of these gene clusters could improve the expression of these gene clusters (Ongley et al., 2013) and hence, confirm the biosynthesis of bromoalterochromide, alteramide and pseudoalterobacin-like compounds, as well as the isolation of novel natural products derived from the identified gene clusters.

Although chemical studies did not find NRPS or PKS compounds, numerous DKPs were isolated, which could not be linked to any biosynthesis genes. The construction of a metagenomic library and screening of this library for quorum sensing function may be able to elucidate the biosynthesis clusters for these DKPs.

A CDPS pathway has yet to be identified in Psueudoalteromonas. The identification and characterisation of a novel CDPS gene cluster would greatly aid in our understanding of this new class of enzymes.

154

Chapter 6

6.4. Closing remarks

This thesis describes an investigation of natural products from bacteria associated with TTX-producing marine molluscs. Our research on the origins of TTX indicated that isolation of TTX-producing microbes are more difficult than suggested by the current literature, and may require further scrutiny. Additionally, we have shown that “unculturable” bacteria may be responsible for its production and is an area that requires further investigation. Furthermore, apart from TTX -producers, marine molluscs harbor complex microbial communities that may be rich environments for the discovery of natural products. One such example was a

Pseudoalteromonas that contained seven gene clusters for the biosynthesis of nonribosomal peptides and polyketides. These gene clusters were identified using next-generation sequencing and data mining methods that are becoming invaluable in the search for new natural products. Recent advances in these computational methods combined with the rapid developments of chemical and synthetic biology represent an exciting new era in natural products research.

155

References

Adam, P., Arigoni, D., Bacher, A., Eisenreich, W., 2002. Biosynthesis of hyperforin in Hypericum perforatum. Journal of Medicinal Chemistry 45, 4786-4793.

Admiraal, S.J., Walsh, C.T., Khosla, C., 2001. The loading module of rifamycin synthetase is an adenylation-thiolation didomain with substrate tolerance for substituted benzoates. Biochemistry 40, 6116-6123.

Ahasan, H., Mamun, A.A., Karim, S.R., Bakar, M.A., Gazi, E.A., Bala, C.S., 2004. Paralytic complications of puffer fish (tetrodotoxin) poisoning. Singapore Medical Journal 45, 73-74.

Arakawa, O., Nishio, S., Noguchi, T., Shida, Y., Onoue, Y., 1995. A new saxitoxin analogue from a xanthid crab Atergatis floridus. Toxicon 33, 1577-1584.

Azevedo, C., 1993. Occurrence of an unusual branchial mycoplasma-like infection in cockle Cerastoderma edule (Mollusca, Bivalvia). Diseases of Aquatic Organisms 16, 55-59.

Aziz, R., Bartels, D., Best, A., DeJongh, M., Disz, T., Edwards, R., Formsma, K., Gerdes, S., Glass, E., Kubal, M., 2008. The RAST Server: rapid annotations using subsystems technology. BMC Genomics 9, 75.

Bachmann, B.O., Ravel, J., 2009. Chapter 8 Methods for in silico prediction of microbial polyketide and nonribosomal peptide biosynthetic pathways from DNA sequence data, in: David, A.H. (Ed.), Methods in Enzymology. Academic Press, pp. 181-217.

Balibar, C.J., Walsh, C.T., 2006. GliP, a multimodular nonribosomal peptide synthetase in Aspergillus fumigatus, makes the diketopiperazine scaffold of gliotoxin. Biochemistry 45, 15029-15038.

Barry, S.M., Challis, G.L., 2009. Recent advances in siderophore biosynthesis. Current Opinion in Chemical Biology 13, 205-215.

156

Belin, P., Moutiez, M., Lautru, S., Seguin, J., Pernodet, J.-L., Gondry, M., 2012. The nonribosomal synthesis of diketopiperazines in tRNA-dependent cyclodipeptide synthase pathways. Natural Product Reports 29, 961-979.

Berdy, J., 2005. Bioactive microbial metabolites. Journal of Antibiotics 58, 1-26.

Blodgett, J.A., Oh, D.-C., Cao, S., Currie, C.R., Kolter, R., Clardy, J., 2010. Common biosynthetic origins for polycyclic tetramate macrolactams from phylogenetically diverse bacteria. Proceedings of the National Academy of Sciences 107, 11692- 11697.

Blunt, J.W., Copp, B.R., Keyzers, R.A., Munro, M.H., Prinsep, M.R., 2013. Marine natural products. Natural Product Reports 30, 237-323.

Blunt, J.W., Copp, B.R., Munro, M.H.G., Northcote, P.T., Prinsep, M.R., 2010. Marine natural products. Natural Product Reports, 165-237.

Bonnefond, L., Arai, T., Sakaguchi, Y., Suzuki, T., Ishitani, R., Nureki, O., 2011. Structural basis for nonribosomal peptide synthesis by an aminoacyl-tRNA synthetase paralog. Proceedings of the National Academy of Sciences 108, 3912- 3917.

Bowman, J.P., 2007. Bioactive compound synthetic capacity and ecological significance of marine bacterial genus Pseudoalteromonas. Marine Drugs 5, 220- 241.

Brady, S.F., Simmons, L., Kim, J.H., Schmidt, E.W., 2009. Metagenomic approaches to natural products from free-living and symbiotic organisms. Natural Product Reports 26, 1488-1503.

Brown, M.S., Mosher, H.S., 1963. Tarichatoxin: Isolation and purification. Science 140, 295-296.

Burke, C., Thomas, T., Egan, S., Kjelleberg, S., 2007. The use of functional genomics for the identification of a gene cluster encoding for the biosynthesis of an antifungal tambjamine in the marine bacterium Pseudoalteromonas tunicata. Environmental Microbiology 9, 814-818.

157

Burns, B.P., Seifert, A., Goh, F., Pomati, F., Jungblut, A.D., Serhat, A., Neilan, B.A., 2005. Genetic potential for secondary metabolite production in stromatolite communities. FEMS Microbiology Letters 243, 293-301.

Campbell, S., Harada, R.M., DeFelice, S.V., Bienfang, P.K., Li, Q.X., 2009. Bacterial production of tetrodotoxin in the pufferfish Arothron hispidus. Natural Product Research 23, 1630-1640.

Cardall, B.L., Brodie, E.D., Hanifin, C.T., 2004. Secretion and regeneration of tetrodotoxin in the rough-skin newt (Taricha granulosa). Toxicon 44, 933-938.

Carroll, S., McEvoy, E., Gibson, R., 2003. The production of tetrodotoxin-like substances by nemertean worms in conjunction with bacteria. Journal of Experimental Marine Biology and Ecology 288, 51-63.

Chaston, J., Goodrich-Blair, H., 2010. Common trends in mutualism revealed by model associations between invertebrates and bacteria. FEMS Microbiology Reviews 34, 41-58.

Chau, R., Kalaitzis, J., Wood, S., Neilan, B., 2013. Diversity and biosynthetic potential of culturable microbes associated with toxic marine animals. Marine Drugs 11, 2695-2712.

Chau, R., Kalaitzis, J.A., Neilan, B.A., 2011. On the origins and biosynthesis of tetrodotoxin. Aquatic Toxicology 104, 61-72.

Cheng, C.A., Hwang, D.F., Tsai, Y.H., Chen, H.C., Jeng, S.S., Noguchi, T., Ohwada, K., Hasimoto, K., 1995. Microflora and tetrodotoxin-producing bacteria in a gastropod, Niotha clathrata. Food and Chemical Toxicology 33, 929-934.

Chooi, Y.-H., Tang, Y., 2010. Adding the lipo to lipopeptides: do more with less. Chemistry & Biology 17, 791-793.

Clardy, J., Walsh, C., 2004. Lessons from natural molecules. Nature 432, 829-837.

Clarke, K.R., Gorley, R.N., 2006. PRIMER V6. Primer-E Ltd. Plymouth.

158

Corre, C., Challis, G.L., 2009. New natural product biosynthetic chemistry discovered by genome mining. Natural Product Reports 26, 977-986.

Cox, M.P., Peterson, D.A., Biggs, P.J., 2010. SolexaQA: At-a-glance quality assessment of Illumina second-generation sequencing data. BMC Bioinformatics 11, 485.

Cronan Jr, J.M., Davidson, T.R., Singleton, F.L., Colwell, R.R., Cardellina, J.H., 1998. Plant growth promoters isolated from a marine bacterium associated with Palythoa sp. Natural Product Letters 11, 271-278.

Daly, J., Gusovsky, F., Myers, C., Yotsu-Yamashita, M., Yasumoto, T., 1994. First occurrence of tetrodotoxin in a dendrobatid frog (Colostethus inguinalis), with further reports for the bufonid genus Atelopus. Toxicon 32, 279.

Dao, H., Takata, Y., Sato, S., Fukuyo, Y., Kodama, M., 2009. Frequent occurrence of the tetrodotoxin-bearing horseshoe crab Carcinoscorpius rotundicauda in Vietnam. Fisheries Science 75, 435-438.

Davey, M.E., O'toole, G.A., 2000. Microbial biofilms: from ecology to molecular genetics. Microbiology and Molecular Biology Reviews 64, 847-867.

Davies, D.G., Parsek, M.R., Pearson, J.P., Iglewski, B.H., Costerton, J.W., Greenberg, E.P., 1998. The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 280, 295-298. de Almeida, M.F., Guedes, M.L., Cruz, F.G., 2011. Lathrophytoic acids A and B: two novel polyprenylated phloroglucinol derivatives from Kielmeyera lathrophyton. Tetrahedron Letters 52, 7108-7112.

Degrassi, G., Aguilar, C., Bosco, M., Zahariev, S., Pongor, S., Venturi, V., 2002. Plant growth-promoting Pseudomonas putida WCS358 produces and secretes four cyclic dipeptides: cross-talk with quorum sensing bacterial sensors. Current Microbiology 45, 250-254.

Deng, J., Hamada, Y., Shioiri, T., 1995. Total synthesis of alterobactin A, a super siderophore from an open-ocean bacterium. Journal of the American Chemical Society 117, 7824-7825.

159

Dewick, P., 2002. The mevalonate and deoxyxylulose phosphate pathways: Terpenoids and steroids, Medicinal Natural Products: A biosynthetic approach. John Wiley & Sons Inc, pp. 167-285.

Do, H., Hamasaki, K., Ohwada, K., Simidu, U., Noguchi, T., Shida, Y., Kogure, K., 1993. Presence of tetrodotoxin and tetrodotoxin-producing bacteria in freshwater sediments. Applied and Environmental Microbiology 59, 3934.

Do, H., Kogure, K., Imada, C., Noguchi, T., Ohwada, K., Simidu, U., 1991. Tetrodotoxin production of actinomycetes isolated from marine sediment. Journal of Applied Microbiology 70, 464-468.

Do, H., Kogure, K., Simidu, U., 1990. Identification of deep-sea-sediment bacteria which produce tetrodotoxin. Applied and Environmental Microbiology 56, 1162.

Dreier, J., Khosla, C., 2000. Mechanistic analysis of a type II polyketide synthase. Role of conserved residues in the β-ketoacyl synthase-chain length factor heterodimer. Biochemistry 39, 2088-2095.

Du, L., Sánchez, C., Shen, B., 2001. Hybrid peptide–polyketide natural products: Biosynthesis and prospects toward engineering novel molecules. Metabolic Engineering 3, 78-95.

Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32, 1792-1797.

Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194-2200.

Ehrenreich, I.M., Waterbury, J.B., Webb, E.A., 2005. Distribution and diversity of natural product genes in marine and freshwater cyanobacterial cultures and genomes. Applied and Environmental Microbiology 71, 7401-7413.

Exner, M.M., Kostelnik, R.J., 1977. Synthesis and carbon-13 nuclear magnetic resonance spectra of Cyclo (D-leucyl-L-leucine) and Cyclo (L-leucyl-L-leucine). Biopolymers 16, 1387-1395.

160

Fischbach, M.A., Walsh, C.T., 2006. Assembly-line enzymology for polyketide and nonribosomal peptide antibiotics: logic, machinery, and mechanisms. Chemical Reviews 106, 3468-3496.

Franks, A., Haywood, P., Holmström, C., Egan, S., Kjelleberg, S., Kumar, N., 2005. Isolation and structure elucidation of a novel yellow pigment from the marine bacterium Pseudoalteromonas tunicata. Molecules 10, 1286-1291.

Furtado, N.A., Pupo, M.T., Carvalho, I., Campo, V.L., Duarte, M.C.T., Bastos, J.K., 2005. Diketopiperazines produced by an Aspergillus fumigatus Brazilian strain. Journal of the Brazilian Chemical Society 16, 1448-1453.

Gallacher, S., Birkbeck, T., 1993. Effect of phosphate concentration on production of tetrodotoxin by Alteromonas tetraodonis. Applied and Environmental Microbiology 59, 3981.

Gauthier, G., Gauthier, M., Christen, R., 1995. Phylogenetic analysis of the genera Alteromonas, Shewanella, and Moritella using genes coding for small-subunit rRNA sequences and division of the genus Alteromonas into two genera, Alteromonas (emended) and Pseudoalteromonas gen. nov., and proposal of twelve new species combinations. International Journal of Systematic Bacteriology 45, 755-761.

Giessen, T.W., Franke, K.B., Knappe, T.A., Kraas, F.I., Bosello, M., Xie, X., Linne, U., Marahiel, M.A., 2012. Isolation, structure elucidation, and biosynthesis of an unusual hydroxamic acid ester-containing siderophore from Actinosynnema mirum. Journal of Natural Products 75, 905-914.

Glaser, K.B., Mayer, A.M.S., 2009. A renaissance in marine pharmacology: From preclinical curiosity to clinical reality. Biochemical Pharmacology 78, 440-448.

Gondry, M., Sauguet, L., Belin, P., Thai, R., Amouroux, R., Tellier, C., Tuphile, K., Jacquet, M., Braud, S., Courcon, M., Masson, C., Dubois, S., Lautru, S., Lecoq, A., Hashimoto, S.-i., Genet, R., Pernodet, J.-L., 2009. Cyclodipeptide synthases are a family of tRNA-dependent peptide bond-forming enzymes. Nature Chemical Biology 5, 414-420.

161

Gottlieb, H.E., Kotlyar, V., Nudelman, A., 1997. NMR chemical shifts of common laboratory solvents as trace impurities. The Journal of Organic Chemistry 62, 7512- 7515.

Guindon, S., Dufayard, J.-F., Lefort, V., Anisimova, M., Hordijk, W., Gascuel, O., 2010. New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Systematic Biology 59, 307-321.

Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52, 696-704.

Handelsman, J., 2004. Metagenomics: application of genomics to uncultured microorganisms. Microbiology and Molecular Biology Reviews 68, 669-685.

Hanefeld, U., Floss, H.G., Laatsch, H., 1994. Biosynthesis of the marine antibiotic pentabromopseudilin. Part 1. The benzene ring. The Journal of Organic Chemistry 59, 3604-3608.

Hanifin, C.T., 2010. The chemical and evolutionary ecology of tetrodotoxin (TTX) toxicity in terrestrial vertebrates. Marine Drugs 8, 577-593.

Hansen, D.B., Bumpus, S.B., Aron, Z.D., Kelleher, N.L., Walsh, C.T., 2007. The loading module of mycosubtilin: An adenylation domain with fatty acid selectivity. Journal of the American Chemical Society 129, 6366-6367.

Harrigan, G.G., Goetz, G., 2002. Symbiotic and dietary marine microalgae as a source of bioactive molecules-experience from natural products research. Journal of Applied Phycology 14, 103-108.

Hay, M., Fenical, W., 1996. Chemical ecology and marine biodiversity: Insights and products from the sea. Oceanography 9, 10-20.

Hoffmann, D., Hevel, J.M., Moore, R.E., Moore, B.S., 2003. Sequence analysis and biochemical characterization of the nostopeptolide A biosynthetic gene cluster from Nostoc sp. GSV224. Gene 311, 171-180.

162

Holden, M.T., Ram Chhabra, S., De Nys, R., Stead, P., Bainton, N.J., Hill, P.J., Manefield, M., Kumar, N., Labatte, M., England, D., 1999. Quorum-sensing cross talk: isolation and chemical characterization of cyclic dipeptides from Pseudomonas aeruginosa and other Gram-negative bacteria. Molecular Microbiology 33, 1254-1266.

Hwang, D., Arakawa, O., Saito, T., Noguchi, T., Simidu, U., Tsukamoto, K., Shida, Y., Hashimoto, K., 1989. Tetrodotoxin-producing bacteria from the blue-ringed octopus Octopus maculosus. Marine Biology 100, 327-332.

Hwang, D., Lu, S., Jeng, S., 1991. Occurrence of tetrodotoxin in the gastropods Rapana rapiformis and R. venosa venosa. Marine Biology 111, 65-69.

Ienaga, K., Nakamura, K., Goto, T., 1987. Bioactive compounds produced in animal tissues (I); two diketopiperadine plant growth regulators containing hydroxyproline isolated from rabbit skin tissue extract. Tetrahedron Letters 28, 1285-1286.

Izumikawa, M., Cheng, Q., Moore, B.S., 2006. Priming type II polyketide synthases via a type II nonribosomal peptide synthetase mechanism. Journal of the American Chemical Society 128, 1428-1429.

Jang, J., Yotsu-Yamashita, M., 2006. Distribution of tetrodotoxin, saxitoxin, and their analogs among tissues of the puffer fish Fugu pardalis. Toxicon 48, 980-987.

Jayatilake, G.S., Thornton, M.P., Leonard, A.C., Grimwade, J.E., Baker, B.J., 1996. Metabolites from an Antarctic sponge-associated bacterium, Pseudomonas aeruginosa. Journal of Natural Products 59, 293-296.

Jeon, J., Narita, H., Nara, M., Noguchi, T., Maruyama, J., Hashimoto, K., 1984. Occurrence of tetrodotoxin in a gastropod mollusk, “araregai” Niotha clathrata. Bulletin of the Japanese Society of Scientific Fisheries 50, 2099-2102.

Jiang, H., Wood, K.V., Morgan, J.A., 2005. Metabolic engineering of the phenylpropanoid pathway in Saccharomyces cerevisiae. Applied and Environmental Microbiology 71, 2962-2969.

163

Jiang, Z., Boyd, K.G., Mearns-Spragg, A., Adams, D.R., Wright, P.C., Burgess, J.G., 2000. Two diketopiperazines and one halogenated phenol from cultures of the marine bacterium, Pseudoalteromonas luteoviolacea. Natural Product Letters 14, 435-440.

Kalaitzis, J.A., Chau, R., Kohli, G.S., Murray, S.A., Neilan, B.A., 2010. Biosynthesis of toxic naturally-occurring seafood contaminants. Toxicon 56, 244-258.

Kalisky, T., Quake, S.R., 2011. Single-cell genomics. Nature Methods 8, 311-314.

Kanoh, K., Kamino, K., Leleo, G., Adachi, K., Shizuri, Y., 2004. Pseudoalterobactin A and B, new siderophores excreted by marine bacterium Pseudoalteromonas sp. KP20-4. ChemInform 35, no-no.

Kanzaki, H., Yanagisawa, S., Nitoda, T., 2000. Biosynthetic intermediates of the tetradehydro cyclic dipeptide albonoursin produced by Streptomyces albulus KO- 23. The Journal of Antibiotics 53, 1257.

Kellmann, R., 2005. The molecular genetics of Cylindrospermopsis and saxitoxin biosynthesis, School of Biotechnology and Biomolecular Sciences. University of New South Wales, Sydney.

Kellmann, R., Mihali, T., Jeon, Y., Pickford, R., Pomati, F., Neilan, B., 2008. Biosynthetic intermediate analysis and functional homology reveal a saxitoxin gene cluster in cyanobacteria. Applied and Environmental Microbiology 74, 4044.

Kim, Y.H., Brown, G.B., Mosher, H.S., Fuhrman, F.A., 1975. Tetrodotoxin: occurrence in atelopid frogs of Costa Rica. Science 189, 151-152.

Kodama, M., Sato, S., Sakamoto, S., Ogata, T., 1996. Occurrence of tetrodotoxin in Alexandrium tamarense, a causative dinoflagellate of paralytic shellfish poisoning. Toxicon 34, 1101-1105.

Kotaki, Y., Shimizu, Y., 1993. 1-Hydroxy-5,11-dideoxytetrodotoxin, the first N- hydroxy and ring-deoxy derivative of tetrodotoxin found in the newt Taricha granulosa. Journal of the American Chemical Society 115, 827-830.

164

Kraas, F.I., Giessen, T.W., Marahiel, M.A., 2012. Exploring the mechanism of lipid transfer during biosynthesis of the acidic lipopeptide antibiotic CDA. FEBS Letters 586, 283-288.

Kraas, F.I., Helmetag, V., Wittmann, M., Strieker, M., Marahiel, M.A., 2010. Functional dissection of surfactin synthetase initiation module reveals insights into the mechanism of lipoinitiation. Chemistry & Biology 17, 872-880.

Kungsuwan, A., Nagashima, Y., Noguchi, T., Shida, Y., Suvapeepan, S., Suwansakornkul, P., Hashimoto, K., 1987. Tetrodotoxin in the horseshoe crab Carcinoscorpius rotundicauda inhabiting Thailand. Nippon Suisan Gakkaishi 53, 261-266.

Kurosawa, K., Ghiviriga, I., Sambandan, T.G., Lessard, P.A., Barbara, J.E., Rha, C., Sinskey, A.J., 2008. Rhodostreptomycins, antibiotics biosynthesized following horizontal gene transfer from Streptomyces padanus to Rhodococcus fascians. Journal of the American Chemical Society 130, 1126-1127.

Lane, D.J., 1991. 16S/23S rRNA sequencing, in: Stackebrandt , E., Goodfellow, M. (Eds.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley and Sons, New York, pp. 115-175.

Lautru, S., Gondry, M., Genet, R., Pernodet, J.-L., 2002. The albonoursin gene cluster of S. noursei: Biosynthesis of diketopiperazine metabolites independent of nonribosomal peptide synthetases. Chemistry & Biology 9, 1355-1364.

Lautru, S., Oves-Costales, D., Pernodet, J.-L., Challis, G.L., 2007. MbtH-like protein- mediated cross-talk between non-ribosomal peptide antibiotic and siderophore biosynthetic pathways in Streptomyces coelicolor M145. Microbiology 153, 1405- 1412.

Lee, M., Jeong, D., Kim, W., Kim, H., Kim, C., Park, W., Park, Y., Kim, K., Kim, H., Kim, D., 2000. A tetrodotoxin-producing Vibrio strain, LM-1, from the puffer fish Fugu vermicularis radiatus. Applied and Environmental Microbiology 66, 1698.

165

Lehman, E.M., Brodie, E.D., 2004. No evidence for an endosymbiotic bacterial origin of tetrodotoxin in the newt Taricha granulosa. Toxicon 44, 243-249.

Lindquist, N., Fenical, W., 1991. New tamjamine class alkaloids from the marine ascidian Atapozoa sp. and its nudibranch predators. Origin of the tambjamines in Atapozoa. Experientia 47, 504-506.

Lipkind, G., Fozzard, H., 1994. A structural model of the tetrodotoxin and saxitoxin binding site of the Na+ channel. Biophysical Journal 66, 1-13.

Louie, G.V., Bowman, M.E., Moffitt, M.C., Baiga, T.J., Moore, B.S., Noel, J.P., 2006. Structural determinants and modulation of substrate specificity in phenylalanine- tyrosine ammonia-lyases. Chemistry & Biology 13, 1327-1338.

Luesch, H., Harrigan, G., Goetz, G., Horgen, F., 2002. The cyanobacterial origin of potent anticancer agents originally isolated from sea hares. Current Medicinal Chemistry 9, 1791-1806.

Luo, R., Liu, B., Xie, Y., Li, Z., Huang, W., Yuan, J., He, G., Chen, Y., Pan, Q., Liu, Y., 2012. SOAPdenovo2: an empirically improved memory-efficient short-read de novo assembler. GigaScience 1, 18.

Marchler-Bauer, A., Lu, S., Anderson, J.B., Chitsaz, F., Derbyshire, M.K., DeWeese- Scott, C., Fong, J.H., Geer, L.Y., Geer, R.C., Gonzales, N.R., 2011. CDD: a Conserved Domain Database for the functional annotation of proteins. Nucleic Acids Research 39, D225-D229.

Maruyama, J., Noguchi, T., Jeon, J., Harada, T., Hashimoto, K., 1984. Occurrence of tetrodotoxin in the starfish Astropecten latespinosus. Cellular and Molecular Life Sciences 40, 1395-1396.

Matsui, T., Taketsugu, S., Kodama, K., Ishii, A., Yamamori, K., Shimizu, C., 1989. Production of tetrodotoxin by the intestinal bacteria of a puffer fish Takifugu niphobles. Nippon Suisan Gakkaishi 55, 2199-2203.

Matsumura, K., 1995. Reexamination of tetrodotoxin production by bacteria. Applied and Environmental Microbiology 61, 3468.

166

McNabb, P., Selwood, A.I., Munday, R., Wood, S.A., Taylor, D.I., MacKenzie, L.A., van Ginkel, R., Rhodes, L.L., Cornelisen, C., Heasman, K., Holland, P.T., King, C., 2010. Detection of tetrodotoxin from the grey side-gilled sea slug - Pleurobranchaea maculata, and associated dog neurotoxicosis on beaches adjacent to the Hauraki Gulf, Auckland, New Zealand. Toxicon 56, 466-473.

Mebs, D., Schmidt, K., 1989. Occurrence of tetrodotoxin in the frog Atelopus oxyrhynchus. Toxicon 27, 819.

Medema, M.H., Blin, K., Cimermancic, P., de Jager, V., Zakrzewski, P., Fischbach, M.A., Weber, T., Takano, E., Breitling, R., 2011. antiSMASH: rapid identification, annotation and analysis of secondary metabolite biosynthesis gene clusters in bacterial and fungal genome sequences. Nucleic Acids Research 39, W339-W346.

Mihali, T., Kellmann, R., Muenchhoff, J., Barrow, K., Neilan, B., 2008. Characterization of the gene cluster responsible for cylindrospermopsin biosynthesis. Applied and Environmental Microbiology 74, 716.

Mihara, K., Tanabe, T., Yamakawa, Y., Funahashi, T., Nakao, H., Narimatsu, S., Yamamoto, S., 2004. Identification and transcriptional organization of a gene cluster involved in biosynthesis and transport of acinetobactin, a siderophore produced by Acinetobacter baumannii ATCC 19606T. Microbiology 150, 2587- 2597.

Milne, P.J., Hunt, A.L., Rostoll, K., Van Der Walt, J.J., Graz, C.J.M., 1998. Medicinal chemistry: The biological activity of selected cyclic dipeptides. Journal of Pharmacy and Pharmacology 50, 1331-1337.

Miyazawa, K., Jeon, J., Maruyama, J., Noguchi, T., Ito, K., Hashimoto, K., 1986. Occurrence of tetrodotoxin in the flatworm Planocera multitentaculata. Toxicon 24, 645-650.

Miyazawa, K., Noguchi, T., Maruyama, J., Jeon, J., Otsuka, M., Hashimoto, K., 1985. Occurrence of tetrodotoxin in the starfishes Astropecten polyacanthus and A. scoparius in the Seto Inland Sea. Marine Biology 90, 61-64.

167

Moffitt, M.C., Neilan, B.A., 2001. On the presence of peptide synthetase and polyketide synthase genes in the cyanobacterial genus Nodularia. FEMS Microbiology Letters 196, 207-214.

Moffitt, M.C., Neilan, B.A., 2003. Evolutionary affiliations within the superfamily of ketosynthases reflect complex pathway associations. Journal of Molecular Evolution 56, 446-457.

Mootz, H.D., Marahiel, M.A., 1999. Design and application of multimodular peptide synthetases. Current Opinion in Biotechnology 10, 341-348.

Mootz, H.D., Schwarzer, D., Marahiel, M.A., 2002. Ways of assembling complex natural products on modular nonribosomal peptide synthetases. ChemBioChem 3, 490-504.

Nagashima, Y., Tanaka, N., Shimakura, K., Shiomi, K., Shida, Y., 2001. Occurrence of tetrodotoxin-related substances in the nontoxic puffer Takifugu xanthopterus. Toxicon 39, 415-418.

Narita, H., Matsubara, S., Miwa, N., Akahane, S., Murakami, M., Goto, T., Nara, M., Noguchi, T., Saito, T., Shida, Y., 1987. Vibrio alginolyticus, a TTX-producing bacterium isolated from the starfish Astropecten polyacanthus. Nippon Suisan Gakkaishi 53, 617-621.

Narita, H., Noguchi, T., Maruyama, J., Ueda, Y., Hashimoto, K., Watanabe, Y., Hida, K., 1981. Occurrence of tetrodotoxin in a trumpet shell, "Boshubora" Charonia sauliae. Bulletin of the Japanese Society of Scientific Fisheries 47, 935-941.

Neilan, B.A., Dittmann, E., Rouhiainen, L., Bass, R.A., Schaub, V., Sivonen, K., Borner, T., 1999. Nonribosomal peptide synthesis and toxigenicity of cyanobacteria. Journal of Bacteriology 181, 4089-4097.

Neilan, B.A., Jacobs, D., Therese, D.D., Blackall, L.L., Hawkins, P.R., Cox, P.T., Goodman, A.E., 1997. rRNA sequences and evolutionary relationships among toxic and nontoxic cyanobacteria of the genus Microcystis. International Journal of Systematic and Evolutionary Microbiology 47, 693.

168

Neilands, J.B., 1995. Siderophores: Structure and function of microbial iron transport compounds. Journal of Biological Chemistry 270, 26723-26726.

Ngy, L., Yu, C., Taniyama, S., Takatani, T., Arakawa, O., 2009. Co-occurrence of tetrodotoxin and saxitoxin in Cambodian marine pufferfish Takifugu oblongus. African Journal of Marine Science 31.

Noguchi, T., Arakawa, O., 2008. Tetrodotoxin - distribution and accumulation in aquatic organisms, and cases of Human intoxication. Marine Drugs 6, 220-242.

Noguchi, T., Arakawa, O., Takatani, T., 2006a. Toxicity of pufferfish Takifugu rubripes cultured in netcages at sea or aquaria on land. Comparative Biochemistry and Physiology Part D: Genomics and Proteomics 1, 153-157.

Noguchi, T., Arakawa, O., Takatani, T., 2006b. TTX accumulation in pufferfish. Comparative Biochemistry and Physiology - Part D: Genomics and Proteomics 1, 145-152.

Noguchi, T., Hashimoto, Y., 1973. Isolation of tetrodotoxin from a goby Gobius criniger. Toxicon 11, 305-307.

Noguchi, T., Hwang, D., Arakawa, O., Sugita, H., Deguchi, Y., Shida, Y., Hashimoto, K., 1987. Vibrio alginolyticus, a tetrodotoxin-producing bacterium, in the intestines of the fish Fugu vermicularis vermicularis. Marine Biology 94, 625-630.

Noguchi, T., Maruyama, J., Narita, H., Hashimoto, K., 1984. Occurrence of tetrodotoxin in the gastropod mollusk Tutufa lissostoma (frog shell). Toxicon 22, 219.

Noguchi, T., Maruyama, J., Ueda, Y., Hashimoto, K., Harada, T., 1981. Occurrence of tetrodotoxin in the Japanese ivory shell Babylonia japonica. Bulletin of the Japanese Society of Scientific Fisheries 47, 901-913.

Noguchi, T., Uzu, A., Koyama, K., Hashimoto, K., 1983. Occurrence of tetrodotoxin as the major toxin in xanthid crab Atergatis floridus. Bulletin of the Japanese Society of Scientific Fisheries 49, 1887-1892.

169

O'Toole, G., Kaplan, H.B., Kolter, R., 2000. Biofilm formation as microbial development. Annual Reviews in Microbiology 54, 49-79.

Oh, D.-C., Kauffman, C.A., Jensen, P.R., Fenical, W., 2007. Induced production of emericellamides A and B from the marine-derived fungus Emericella sp. in competing co-culture. Journal of Natural Products 70, 515-520.

Ongley, S.E., Bian, X., Neilan, B.A., Muller, R., 2013. Recent advances in the heterologous expression of microbial natural product biosynthetic pathways. Natural Product Reports 30, 1121-1138.

Pawlik, J.R., 1993. Marine invertebrate chemical defenses. Chemical Reviews 93, 1911-1922.

Penn, K., Jenkins, C., Nett, M., Udwary, D.W., Gontang, E.A., McGlinchey, R.P., Foster, B., Lapidus, A., Podell, S., Allen, E.E., Moore, B.S., Jensen, P.R., 2009. Genomic islands link secondary metabolism to functional adaptation in marine Actinobacteria. The ISME Journal 3, 1193-1203.

Pennings, S.C., Paul, V.J., 1993. Sequestration of dietary secondary metabolites by three species of sea hares: location, specificity and dynamics. Marine Biology 117, 535-546.

Peschke, J.D., Hanefeld, U., Laatsch, H., 2005. Biosynthesis of the marine antibiotic pentabromopseudilin. 2. The pyrrole ring. Bioscience, Biotechnology and Biochemistry 69, 628-630.

Pettit, R., 2009. Mixed fermentation for natural product drug discovery. Applied Microbiology and Biotechnology 83, 19-25.

Pfeifer, B.A., Wang, C.C.C., Walsh, C.T., Khosla, C., 2003. Biosynthesis of yersiniabactin, a complex polyketide-nonribosomal peptide, using Escherichia coli as a heterologous host. Applied and Environmental Microbiology 69, 6698-6702.

Piel, J., 2004. Metabolites from symbiotic bacteria. Natural Product Reports 21, 519-538.

170

Piel, J., Hertweck, C., Shipley, P.R., Hunt, D.M., Newman, M.S., Moore, B.S., 2000. Cloning, sequencing and analysis of the enterocin biosynthesis gene cluster from the marine isolate Streptomyces maritimus: evidence for the derailment of an aromatic polyketide synthase. Chemistry & Biology 7, 943-955.

Pires, O., Sebben, A., Schwartz, E., Largura, S., Bloch, C., Morales, R., Schwartz, C., 2002. Occurrence of tetrodotoxin and its analogues in the Brazilian frog Brachycephalus ephippium (Anura: Brachycephalidae). Toxicon 40, 761-766.

Pires, O., Sebben, A., Schwartz, E., Morales, R., Bloch, C., Schwartz, C., 2005. Further report of the occurrence of tetrodotoxin and new analogues in the Anuran family Brachycephalidae. Toxicon 45, 73-79.

Poppe, L., Rétey, J., 2005. Friedel-Crafts-yype mechanism for the enzymatic elimination of ammonia from histidine and phenylalanine. Angewandte Chemie International Edition 44, 3668-3688.

Prasad, C., 1995. Bioactive cyclic dipeptides. Peptides 16, 151-164.

Quadri, L.E., Sello, J., Keating, T.A., Weinreb, P.H., Walsh, C.T., 1998. Identification of a Mycobacterium tuberculosis gene cluster encoding the biosynthetic enzymes for assembly of the virulence-conferring siderophore mycobactin. Chemistry & Biology 5, 631-645.

R Core Team, 2013. R: A language for data analysis and graphics. R Foundation for Statistical Computing, Vienna, Austria.

Rausch, C., Hoof, I., Weber, T., Wohlleben, W., Huson, D.H., 2007. Phylogenetic analysis of condensation domains in NRPS sheds light on their functional evolution. BMC Evolutionary Biology 7, 78.

Razin, S., Yogev, D., Naot, Y., 1998. Molecular biology and pathogenicity of mycoplasmas. Microbiology and Molecular Biology Reviews 62, 1094-1156.

Ritchie, K., Nagelkerken, I., James, S., Smith, G., 2000. Environmental microbiology: a tetrodotoxin-producing marine pathogen. Nature 404, 354.

171

Röther, D., Poppe, L., Morlock, G., Viergutz, S., Rétey, J., 2002. An active site homology model of phenylalanine ammonia-lyase from P. crispum. European Journal of Biochemistry 269, 3065-3075.

Ruby, E.G., 2008. Symbiotic conversations are revealed under genetic interrogation. Nature Reviews Microbiology 6, 752-762.

Rudi, A., Kashman, Y., Benayahu, Y., Schleyer, M., 1994. Amino acid derivatives from the marine sponge Jaspis digonoxea. Journal of Natural Products 57, 829-832.

Saker, M.L., Griffiths, D.J., 2000. The effect of temperature on growth and cylindrospermopsin content of seven isolates of Cylindrospermopsis raciborskii (Nostocales, Cyanophyceae) from water bodies in northern Australia. Phycologia 39, 349-354.

Schloss, P.D., 2009. A high-throughput DNA sequence aligner for microbial ecology studies. PLoS ONE 4, e8230.

Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., 2009. Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology 75, 7537-7541.

Schwarzer, D., Marahiel, M.A., 2001. Multimodular biocatalysts for natural product assembly. Naturwissenschaften 88, 93-101.

Schwyn, B., Neilands, J.B., 1987. Universal chemical assay for the detection and determination of siderophores. Analytical Biochemistry 160, 47-56.

Seguin, J., Moutiez, M., Li, Y., Belin, P., Lecoq, A., Fonvielle, M., Charbonnier, J.-B., Pernodet, J.-L., Gondry, M., 2011. Nonribosomal peptide synthesis in animals: The cyclodipeptide synthase of Nematostella. Chemistry & Biology 18, 1362-1368.

Serino, L., Reimmann, C., Visca, P., Beyeler, M., Chiesa, V.D., Haas, D., 1997. Biosynthesis of pyochelin and dihydroaeruginoic acid requires the iron-regulated

172

pchDCBA operon in Pseudomonas aeruginosa. Journal of Bacteriology 179, 248- 257.

Shen, B., 2003. Polyketide biosynthesis beyond the type I, II and III polyketide synthase paradigms. Current Opinion in Chemical Biology 7, 285-295.

Shen, B., Du, L., Sanchez, C., Edwards, D., Chen, M., Murrell, J., 2001. The biosynthetic gene cluster for the anticancer drug bleomycin from Streptomyces verticillus ATCC15003 as a model for hybrid peptide–polyketide natural product biosynthesis. Journal of Industrial Microbiology & Biotechnology 27, 378-385.

Shen, Y., Yoon, P., Yu, T.-W., Floss, H.G., Hopwood, D., Moore, B.S., 1999. Ectopic expression of the minimal whiE polyketide synthase generates a library of aromatic polyketides of diverse sizes and shapes. Proceedings of the National Academy of Sciences 96, 3622-3627.

Sheumack, D., Howden, M., Spence, I., 1984. Occurrence of a tetrodotoxin-like compound in the eggs of the venomous blue-ringed octopus (Hapalochlaena maculosa). Toxicon 22, 811-812.

Shigemori, H., Bae, M.A., Yazawa, K., Sasaki, T., Kobayashi, J., 1992. Alteramide A, a new tetracyclic alkaloid from a bacterium Alteromonas sp. associated with the marine sponge Halichondria okadai. The Journal of Organic Chemistry 57, 4317- 4320.

Shigemori, H., Tenma, M., Shimazaki, K., Kobayashi, J.i., 1998. Three New Metabolites from the Marine Yeast Aureobasidium pullulans. Journal of Natural Products 61, 696-698.

Shimizu, Y., Kobayashi, M., 1983. Apparent lack of tetrodotoxin biosynthesis in captured Taricha torosa and Taricha granulosa. Chemical and Pharmaceutical Bulletin 31, 3625-3631.

Shiomi, K., Yamaguchi, S., Kikuchi, T., Yamamori, K., Matsui, T., 1992. Occurrence of tetrodotoxin-binding high molecular weight substances in the body fluid of shore crab (Hemigrapsus sanguineus). Toxicon 30, 1529-1537.

173

Shiu, Y., Lu, Y., Tsai, Y., Chen, S., Hwang, D., 2003. Occurrence of tetrodotoxin in the causative gastropod Polinices didyma and another gastropod Natica lineata collected from western Taiwan. Journal of Food and Drug Analysis 11, 159-163.

Sieber, S.A., Marahiel, M.A., 2005. Molecular mechanisms underlying nonribosomal peptide synthesis: Approaches to new antibiotics. Chemical Reviews 105, 715-738.

Simidu, U., Kita-Tsukamoto, K., Yasumoto, T., Yotsu, M., 1990. Taxonomy of four marine bacterial strains that produce tetrodotoxin. International Journal of Systematic and Evolutionary Microbiology 40, 331-336.

Simpson, J.T., Wong, K., Jackman, S.D., Schein, J.E., Jones, S.J.M., Birol, İ., 2009. ABySS: A parallel assembler for short read sequence data. Genome Research 19, 1117-1123.

Sobolevskaya, M.P., Smetanina, O.F., Speitling, M., Shevchenko, L.S., Dmitrenok, P.S., Laatsch, H., Kuznetsova, T.A., Ivanova, E.P., Elyakov, G.B., 2005. Controlling production of brominated cyclic depsipeptides by Pseudoalteromonas maricaloris KMM 636T. Letters in Applied Microbiology 40, 243-248.

Speitling, M., Smetanina, O.F., Kuznetsova, T.A., Laatsch, H., 2007. Bromoalterochromides A and A′, unprecedented chromopeptides from a marine Pseudoalteromonas maricaloris strain KMM 636T. The Journal of Antibiotics 60, 36- 42.

Spiering, M.J., Lane, G.A., Christensen, M.J., Schmid, J., 2005. Distribution of the fungal endophyte Neotyphodium lolii is not a major determinant of the distribution of fungal alkaloids in Lolium perenne plants. Phytochemistry 66, 195-202.

Stachelhaus, T., Mootz, H.D., Bergendahl, V., Marahiel, M.A., 1998. Peptide bond formation in nonribosomal peptide biosynthesis: Catalytic role of the condensation domain. Journal of Biological Chemistry 273, 22773-22781.

Su, Y., Koike, K., Guo, D., Satou, T., Liu, J., Zheng, J., Nikaido, T., 2001. New apiose- containing triterpenoid saponins from Conyza blinii. Tetrahedron 57, 6721-6726.

174

Tao, M., Wang, L., Wendt-Pienkowski, E., George, N.P., Galm, U., Zhang, G., Coughlin, J.M., Shen, B., 2006. The tallysomycin biosynthetic gene cluster from Streptoalloteichus hindustanus E465-94 ATCC 31158 unveiling new insights into the biosynthesis of the bleomycin family of antitumor antibiotics. Molecular BioSystems 3, 60-74.

Te, S.H., Gin, K.Y.-H., 2011. The dynamics of cyanobacteria and microcystin production in a tropical reservoir of Singapore. Harmful Algae 10, 319-329.

Thuesen, E., Kogure, K., 1989. Bacterial production of tetrodotoxin in four species of Chaetognatha. Biological Bulletin 176, 191.

Tillett, D., Neilan, B.A., 2000. Xanthogenate nucleic acid isolation from cultured and environmental cyanobacteria. Journal of Phycology 36, 251-258.

Tsai, Y., Hwang, D., Chai, T., Jeng, S., 1995. Occurrence of tetrodotoxin and paralytic shellfish poison in the Taiwanese crab Lophozozymus pictor. Toxicon 33, 1669- 1673.

Udwary, D.W., Zeigler, L., Asolkar, R.N., Singan, V., Lapidus, A., Fenical, W., Jensen, P.R., Moore, B.S., 2007. Genome sequencing reveals complex secondary metabolome in the marine actinomycete Salinispora tropica. Proceedings of the National Academy of Sciences 104, 10376-10381.

Venkatesh, B., Lu, S.Q., Dandona, N., See, S.L., Brenner, S., Soong, T.W., 2005. Genetic basis of tetrodotoxin resistance in pufferfishes. Current Biology 15, 2069-2072.

Vetting, M.W., Hegde, S.S., Blanchard, J.S., 2010. The structure and mechanism of the Mycobacterium tuberculosis cyclodityrosine synthetase. Nature Chemical Biology 6, 797-799.

Wang, J., Fan, Y., Yao, Z., 2010. Isolation of a Lysinibacillus fusiformis strain with tetrodotoxin-producing ability from puffer fish Fugu obscurus and the characterization of this strain. Toxicon 56, 640-643.

175

Wang, X., Yu, R., Luo, X., Zhou, M., Lin, X., 2008. Toxin-screening and identification of bacteria isolated from highly toxic marine gastropod Nassarius semiplicatus. Toxicon 52, 55-61.

Watanabe, K., Praseuth, A.P., Wang, C.C., 2007. A comprehensive and engaging overview of the type III family of polyketide synthases. Current Opinion in Chemical Biology 11, 279-286.

Watts, K.T., Mijts, B.N., Lee, P.C., Manning, A.J., Schmidt-Dannert, C., 2006. Discovery of a substrate selectivity switch in tyrosine ammonia-lyase, a member of the aromatic amino acid lyase family. Chemistry & Biology 13, 1317-1326.

Williams, B.L., Hanifin, C.T., Brodie, E.D., Caldwell, R.L., 2011. Ontogeny of tetrodotoxin levels in blue-ringed octopuses: maternal Investment and apparent independent production in offspring of Hapalochlaena lunulata. Journal of Chemical Ecology, 1-8.

Williams, B.L., Stark, M.R., Caldwell, R.L., 2012. Microdistribution of tetrodotoxin in two species of blue-ringed octopuses (Hapalochlaena lunulata and Hapalochlaena fasciata) detected by fluorescent immunolabeling. Toxicon 60, 1307-1313.

Wilson, K., 2001. Preparation of genomic DNA from bacteria, Current Protocols in Molecular Biology. John Wiley & Sons, Inc.

Wittmann, M., Linne, U., Pohlmann, V., Marahiel, M.A., 2008. Role of DptE and DptF in the lipidation reaction of daptomycin. FEBS Journal 275, 5343-5354.

Wood, S., Casas, M., Taylor, D., McNabb, P., Salvitti, L., Ogilvie, S., Cary, S., 2012a. Depuration of tetrodotoxin and changes in bacterial communities in Pleurobranchea maculata adults and egg masses maintained in captivity. Journal of Chemical Ecology, 1-9.

Wood, S.A., Taylor, D.I., McNabb, P., Walker, J., Adamson, J., Cary, S.C., 2012b. Tetrodotoxin concentrations in Pleurobranchaea maculata: Temporal, spatial and individual variability from New Zealand populations. Marine Drugs 10, 163-176.

176

Woodward, R., Gougoutas, J., 1964. The structure of tetrodotoxin. Journal of the American Chemical Society 86, 5030-5030.

Wu, Z., Xie, L., Xia, G., Zhang, J., Nie, Y., Hu, J., Wang, S., Zhang, R., 2005a. A new tetrodotoxin-producing actinomycete, Nocardiopsis dassonvillei, isolated from the ovaries of puffer fish Fugu rubripes. Toxicon 45, 851-859.

Wu, Z., Yang, Y., Xie, L., Xia, G., Hu, J., Wang, S., Zhang, R., 2005b. Toxicity and distribution of tetrodotoxin-producing bacteria in puffer fish Fugu rubripes collected from the Bohai Sea of China. Toxicon 46, 471-476.

Xiang, L., Moore, B.S., 2005. Biochemical characterization of a prokaryotic phenylalanine ammonia lyase. Journal of Bacteriology 187, 4286-4289.

Yan, Q., Yu, P., Li, H., 2005. Detection of tetrodotoxin and bacterial production by Serratia marcescens. World Journal of Microbiology and Biotechnology 21, 1255- 1258.

Yang, G., Xu, J., Liang, S., Ren, D., Yan, X., Bao, B., 2010. A novel TTX-producing Aeromonas isolated from the ovary of Takifugu obscurus. Toxicon 56, 324-329.

Yasumoto, T., Yasumura, D., Yotsu, M., Michishita, T., Endo, A., Kotaki, Y., 1986. Bacterial production of tetrodotoxin and anhydrotetrodotoxin. Agricultural and Biological Chemistry 50, 793-795.

Yasumoto, T., Yotsu, M., Murata, M., Naoki, H., 1988. New tetrodotoxin analogs from the newt Cynops ensicauda. Journal of the American Chemical Society 110, 2344- 2345.

Yokoo, A., 1950. Study on chemical purification of tetrodotoxin (3) - purification of spheroidine. Journal of The Chemical Society of Japan 71, 590-592.

Yoshikawa, K., Takadera, T., Adachi, K., Nishijima, M., Sano, H., 1997. Korormicin, a novel antibiotic specifically active against marine gram-negative bacteria, produced by a marine bacterium. The Journal of Antibiotics 50, 949-953.

177

Yotsu-Yamashita, M., Mebs, D., 2003. Occurrence of 11-oxotetrodotoxin in the red- spotted newt, Notophthalmus viridescens, and further studies on the levels of tetrodotoxin and its analogues in the newt's efts. Toxicon 41, 893-897.

Yotsu-Yamashita, M., Mebs, D., Flachsenberger, W., 2007. Distribution of tetrodotoxin in the body of the blue-ringed octopus (Hapalochlaena maculosa). Toxicon 49, 410-412.

Yotsu, M., Yamazaki, T., Meguro, Y., Endo, A., Murata, M., Naoki, H., Yasumoto, T., 1987. Production of tetrodotoxin and its derivatives by Pseudomonas sp. isolated from the skin of a pufferfish. Toxicon 25, 225.

Yu, C., Yu, P., Chan, P., Yan, Q., Wong, P., 2004. Two novel species of tetrodotoxin- producing bacteria isolated from toxic marine puffer fishes. Toxicon 44, 641-647.

Yu, F., Zaleta-Rivera, K., Zhu, X., Huffman, J., Millet, J.C., Harris, S.D., Yuen, G., Li, X.-C., Du, L., 2007. Structure and biosynthesis of heat-stable antifungal factor (HSAF), a broad-spectrum antimycotic with a novel mode of action. Antimicrobial Agents and Chemotherapy 51, 64-72.

Zerbino, D.R., Birney, E., 2008. Velvet: Algorithms for de novo short read assembly using de Bruijn graphs. Genome Research 18, 821-829.

Zerikly, M., Challis, G.L., 2009. Strategies for the discovery of new natural products by genome mining. ChemBioChem 10, 625-633.

178