Leon Hartman Thesis

Manipulation of prokaryotic communities in the coral model organism, Exaiptasia diaphana

A thesis submitted for the degree of Doctor of Philosophy

June 2020

Leon Hartman

Principal supervisor: Professor Linda Blackall

Associate supervisors: Professor Madeleine van Oppen

Professor Damien Hicks

Abstract

Coral reefs have sustained damage of increasing scale and frequency due to the manifestations of climate change, most notably increased sea surface temperature, leading to the breakdown of corals’ intracellular symbiosis with dinoflagellate algae of the family Symbiodiniaceae i.e., bleaching. This has intensified the need to understand and enhance coral thermal tolerance. The sea anemone, Exaiptasia diaphana, has proven an ideal model for many coral studies due to its close phylogenetic relationship and shared traits, particularly its symbiosis with Symbiodiniaceae. However, established E. diaphana clonal lines are not available in Australia, thus limiting the ability of Australian scientists to conduct research with this model. Cultures of E. diaphana originally sourced from the Great Barrier Reef (GBR) have been established at The University of Melbourne (UoM), Australia, to address this gap.

In this thesis, I describe work undertaken to investigate the bacterial associates of E. diaphana as it has been hypothesised that these holobiont members influence cnidarian thermal tolerance. Anemones from the UoM E. diaphana collection were used to test this hypothesis and to aid their development as Australian coral models. First, the bacterial associates of four UoM E. diaphana genotypes were characterised by metabarcoding of the 16S rRNA genes to provide baseline bacterial microbiome data, and to assess the similarity of the genotypes to the existing models and each other. Second, the influence of thermal stress on the bacterial communities of one genotype was explored, again by metabarcoding. Third, the feasibility of generating gnotobiotic E. diaphana cultures by antibiotic exposure was tested, since gnotobiotes could help explain the roles of bacteria in cnidarian health and thermal tolerance. Finally, E. diaphana were inoculated with free radical scavenging bacteria before exposure to thermal stress to determine whether probiotic inoculation could mitigate bleaching.

According to my findings, bacterial communities of GBR-origin E. diaphana genotypes are not significantly different from each other and are comparable to those of other E. diaphana models. Environmental stressors drive changes in E. diaphana’s bacterial communities, but the communities are generally stable under thermal stress until a temperature threshold is exceeded. Whilst antibiotic treatment significantly reduces bacterial load, it does not generate true E. diaphana gnotobiotes. The ability of probiotic inoculation to mitigate bleaching in thermally stressed E. diaphana is unclear due to non-retention of probiotic bacteria in the treated anemones. Therefore, according to this work the hypothesis of bacterial influence on i

cnidarian thermal tolerance remains unproven. Nevertheless, these findings improve our understanding of cnidarian microbiome dynamics and advance the recently established GBR- origin E. diaphana cultures as models for coral research in Australia.

Acknowledgments

“Ma te huruhuru te manu ka rere”

It is the feathers that allow the bird to fly

– Maori proverb –

I would first like to thank my supervisor, Prof Blackall, for providing this opportunity. My PhD has its roots in her belief in me, for which I am very grateful. I would also like to thank Prof van Oppen for her invaluable input and guidance over the last three years.

A highlight of my research experience has been the collegiality I have enjoyed with so many lab members, and the support they have given me. This includes my friends from Swinburne University; Hugh, Ben, Hoang, and Alex, and from The University of Melbourne; Giada, Sam, Giulia, Louis, Roy, Sophie, Ruby, Hisham, Roshan, Patrick, Keren, Justin and Wing. I reserve a special thank you to Ashley for her friendship, assistance, and leadership of the group.

I would also like to acknowledge Hannah and Kat for helping to develop methods used by the group, and Kat again for our mutual morale-boosting exchanges.

I would like to express my gratitude to the technicians who made my life easier because they don’t just talk the talk, but also walk the walk: Rebecca, Jason, Stephen and Franca. They were all generous with their time and expert advice.

My time at the Max Planck Institute for Marine Microbiology was the highlight of my PhD. Therefore, I thank my host, Pier, for making this a special experience through his patience, generosity and friendship, and for starting me on my R journey.

Finally, I dedicate this work to my mother, father and sister who have always been there for me with love and support, and to Louise whose love and positivity kept me going through the highs and lows.

iii

Declaration

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 Swinburne or any other educational institution, except where due acknowledgement is made in the manuscript. Any contribution made to the research by others, with whom I have worked at Swinburne or elsewhere, is explicitly acknowledged in the report. I also declare that the intellectual content of this report 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: 27 January 2020

Table of Contents

Abstract ...... i Acknowledgments ...... iii Declaration ...... iv Table of Contents ...... v List of Figures ...... viii List of Tables ...... xi Chapter 1: General introduction ...... 1 1.1 Coral reefs ...... 1 1.2 The coral organism ...... 2 1.3 Coral symbioses ...... 3 1.3.1 Symbiodiniaceae ...... 3 1.3.2 Bacteria ...... 4 1.3.3 Bacterial functions: Pathogen protection ...... 7 1.3.4 Bacterial functions: Nutrient cycling ...... 8 1.3.5 Bacterial functions: Host development ...... 10 1.4 Impacts on coral reefs ...... 11 1.4.1 Natural events ...... 11 1.4.2 Human activity ...... 11 1.5 Solutions ...... 14 1.5.1 Climate change mitigation ...... 14 1.5.2 Assisted evolution ...... 15 1.6 Cnidarian model animals ...... 16 1.6.1 Hydra ...... 17 1.6.2 Nematostella vectensis ...... 17 1.6.3 Aiptasia pallida ...... 18 1.7 Exaiptasia diaphana ...... 19 1.7.1 Global distribution ...... 19 1.7.2 Lifestyle and anatomy ...... 19 1.7.3 Reproduction ...... 20 1.7.4 Microbiome: Symbiodiniaceae ...... 22 1.7.5 Microbiome: Bacteria ...... 23 1.7.6 Model development: ‘omic resources ...... 23 1.7.7 Model development: Standardisation ...... 24 1.8 Organisation of the thesis ...... 25 Chapter 2: Microbiota characterisation of Exaiptasia diaphana from the Great Barrier Reef ...... 27 2.1 Introduction ...... 28 2.2 Materials and methods ...... 31 2.2.1 E. diaphana culture collection ...... 31 2.2.2 Sampling and sample processing ...... 31 2.2.3 Sequencing data workflow ...... 33 2.2.4 Diversity analyses ...... 34 2.2.5 AIMS1–4 core bacterial community member analysis ...... 35 2.2.6 Phenotypic potential analysis ...... 35 2.3 Results ...... 36 2.3.1 Bacteria metabarcoding ...... 36 2.3.2 Diversity analyses of anemone bacterial associates ...... 37 2.3.3 Symbiodiniaceae ...... 43 2.3.4 Anemone core bacteria ...... 43 2.3.5 Phenotypic potential analysis ...... 44

2.4 Discussion ...... 45 2.4.1 Bacterial associates of anemones ...... 45 2.4.2 E. diaphana core bacteria ...... 48 2.4.3 Phenotypic potential of anemone-associated microbiota corresponds to culture environment ...... 49 2.5 Conclusion ...... 51 Chapter 3: The effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana ...... 52 3.1 Introduction ...... 53 3.2 Materials and methods ...... 55 3.2.1 Experimental conditions and sample processing ...... 55 3.2.2 Sequencing data workflow ...... 57 3.2.3 Physiological and microbiome diversity data analyses ...... 58 3.2.4 Analysis of changes in abundance of selected bacterial taxa ...... 59 3.2.5 Indicator species identification ...... 60 3.3 Results ...... 60 3.3.1 Sequencing data and bacterial community characteristics ...... 60 3.3.2 Phenotypic changes in the anemones ...... 61 3.3.3 Changes in alpha diversity of the bacterial microbiomes ...... 62 3.3.4 Changes in beta diversity of the bacterial microbiomes ...... 64 3.3.5 Changes in abundance of selected bacterial taxa ...... 65 3.3.6 Indicator Species Identification ...... 67 3.4 Discussion ...... 68 3.4.1 Factors underpinning bleaching ...... 68 3.4.2 Environmental stressors reduced alpha diversity ...... 69 3.4.3 Turnover of low-abundance ASVs drive shifts in beta diversity ...... 69 3.4.4 Changes in bacterial associates at a high taxonomic level were apparent ...... 70 3.4.5 Were Vibrio victims of competition? ...... 71 3.4.6 Specific bacteria as biomarkers for thermal stress ...... 71 3.5 Conclusion ...... 73 Chapter 4: Towards the generation of gnotobiotic Exaiptasia diaphana ...... 74 4.1 Introduction ...... 75 4.2 Materials and methods ...... 77 4.2.1 Experimental set-up and antibiotic treatment ...... 77 4.2.2 Sampling and DNA extraction ...... 79 4.2.3 Anemone–Symbiodiniaceae density measurement ...... 79 4.2.4 Bacterial load assessment ...... 79 4.2.5 Sample and data processing for bacterial community analysis ...... 81 4.2.6 Data analyses ...... 82 4.3 Results ...... 83 4.3.1 Anemone algal cell density ...... 83 4.3.2 Bacterial load (B/H ratio) ...... 84 4.3.3 Bacterial community data processing ...... 86 4.3.4 Bacterial community characterisation ...... 86 4.3.5 Bacterial genera with high or low changes in abundance ...... 89 4.4 Discussion ...... 91 4.4.1 Antibiotic exposure reduces Symbiodiniaceae density ...... 91 4.4.2 Antibiotic treatment partially eliminates bacteria ...... 92 4.4.3 Antibiotic treatment increases bacterial diversity ...... 93 4.4.4 Antibiotic treatment promotes the growth of some bacterial genera ...... 93 4.4.5 Recommendations for improved bacterial depletion in E. diaphana ...... 94 4.5 Conclusion ...... 95

Chapter 5: Assessment of a ROS-targeted bacterial probiotic designed to improve thermal tolerance in the sea anemone, Exaiptasia diaphana ...... 96 5.1 Introduction ...... 97 5.2 Materials and methods ...... 98 5.2.1 Probiotic preparation ...... 98 5.2.2 Experimental set-up ...... 100 5.2.3 Treatment schedule ...... 100 5.2.4 Symbiodiniaceae photosynthetic efficiency measurement ...... 101 5.2.5 Anemone tissue processing ...... 102 5.2.6 Symbiodiniaceae cell density measurement ...... 102 5.2.7 ROS assay ...... 102 5.2.8 Metabarcoding sample preparation ...... 103 5.2.9 Metabarcoding data processing ...... 104 5.2.10 Data analysis ...... 104 5.3 Results ...... 106 5.3.1 CFU data ...... 106 5.3.2 Photosynthetic efficiency of Symbiodiniaceae ...... 107 5.3.3 Symbiodiniaceae cell density ...... 109 5.3.4 ROS assay ...... 111 5.3.5 Metabarcoding data processing ...... 112 5.3.6 Bacterial community shifts ...... 112 5.3.7 Incorporation of the probiotic bacteria ...... 113 5.4 Discussion ...... 117 5.4.1 Probiotic inoculation did not improve thermal tolerance ...... 117 5.4.2 Probiotic uptake by E. diaphana was short-lived ...... 117 5.4.3 Probiotic uptake by E. diaphana was uneven ...... 118 5.4.4 Bleaching susceptibility differed by genotype ...... 119 5.4.5 Limitations of the study, and recommendations for future probiotic work ...... 119 5.5 Conclusion ...... 120 Chapter 6: General summary ...... 122 6.1 GBR-origin E. diaphana have bacterial microbiomes similar to their cnidarian cousins ...... 122 6.2 The bacterial microbiome of heat-stressed E. diaphana is dynamic ...... 123 6.3 Generation of gnotobiotic E. diaphana remains a work in progress ...... 124 6.4 The ability of probiotic inoculation to mitigate bleaching in E. diaphana remains unknown . 125 6.5 Future directions ...... 126 Appendix 1 ...... 129 Appendix 2 ...... 142 Appendix 3 ...... 150 Appendix 4 ...... 164 Appendix 5 (Authorship Indication) ...... 173 List of Publications ...... 175 References ...... 176

vii

List of Figures

Figure 1.1: Coral assemblage in Palmyra Atoll, Central Pacific ...... 1 Figure 1.2: Coral polyp body plan and cell types ...... 3 Figure 1.3: Coral-associated prokaryotes (>1% relative sequence abundance only) at phylum level, and class level ...... 7 Figure 1.4: Nutrient cycling performed by coral-associated microbial communities ...... 9 Figure 1.5: (a) Coral near Moorea, French Polynesia, in March 2019 with some signs of bleaching, and (b) in May 2019 after widespread bleaching due prolonged exposure to elevated SST ..... 13 Figure 1.7: Simplified depiction of inoculation with BMCs to restore coral health ...... 17 Figure 1.8: Phylogenetic relationships of selected genera within Cnidaria ...... 19 Figure 1.9: E. diaphana: (a) polyp, and (b) cut-away showing anatomical features ...... 20 Figure 1.10: E. diaphana’s primary reproductive modes ...... 21 Figure 2.1: ASVs common to the AIMS1–4 genotypes ...... 37 Figure 2.2: Alpha diversity of sample types: (a) average number of observed ASVs, (b) Simpson diversity, and (c) Shannon diversity index ...... 38 Figure 2.3: Heatmaps of the top 20 taxa by relative abundance at (a) class, and (b) genus levels...... 41 Figure 2.4: Comparison of bacterial communities in Ep1–4 anemones samples by nMDS on Bray- Curtis distances ...... 42 Figure 2.5: Comparison of bacterial communities in water and anemones by nMDS (Bray-Curtis distance) in the (a) AIMS1, (b) AIMS2, (c) AIMS3 and (d) AIMS4 genotypes ...... 43 Figure 2.6: Phenotypic potential inferred from bacterial data by METAGENassist...... 45 Figure 3.1: Sampling schedule for the (a) control anemones, and (b) heat-treated anemones ...... 56 Figure 3.2: Relative abundance of reads assigned to each class ...... 61

Figure 3.3: (a) Dark-adapted quantum yield (Fv/Fm), and (b) Symbiodinaceae cell density ...... 62 Figure 3.4: (a) Average number of observed amplicon sequence variants (ASVs), (b) Simpson index values, and (c) Shannon index values ...... 63 Figure 3.5: Unique and common bacterial ASVs at Day 0 and 14 in (a) control, or (b) heat-treated anemones ...... 64 Figure 3.6: (a) nMDS ordination of the anemone-associated bacterial communities based on Bray- Curtis distances, and (b) plots showing datapoints from the original ordination for each day ...... 65 Figure 3.7: Changes in the six most abundant bacterial classes across all samples: (a) Alphaproteobacteria, (b) Gammaproteobacteria, (c) Bacteroidia, (d) Deltaproteobacteria, (e) Spirochaetia, and (f) Pla3 Lineage ...... 66 Figure 3.8: Changes in relative abundance of Vibrio sp. ASVs ...... 67 Figure 4.1. Symbiodiniaceae density in the control and treated anemones ...... 84

viii

Figure 4.2: (a) Change in bacterial load in the antibiotic-treated anemones, and (b) bacterial load in the untreated and treated A. salina ...... 85 Figure 4.3: Change in bacterial load in the control anemones ...... 86 Figure 4.4: (a) Number of observed ASVs per anemone, (b) Simpson index values, and (c) Shannon index values ...... 87 Figure 4.5: (a) nMDS ordination of bacterial communities in control and antibiotic-treated anemones at Day 0 and Day 21, (b) Venn diagram showing numbers of common and unique ASVs in the Day 21 control anemones, and (c) Venn diagram showing numbers of common and unique ASVs in the Day 21 antibiotic-treated anemones ...... 88 Figure 4.6: Proportions of family-level taxa in (a) six Day 21 control and (b) six Day 21 antibiotic- treated anemones ...... 89 Figure 4.7: Bacterial genera in the untreated Day 0 and antibiotic-treated Day 21 anemones with significant L2FC values (α = 0.01) and changes in relative abundance >0.5% ...... 90 Figure 5.1: Inoculation and sampling schedule for each inoculation-temperature combination: (a) ambient (26 °C), and (b) elevated (26 °C–31.5 °C) ...... 101

Figure 5.2: Dark-adapted Fv/Fm at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all experimental treatments within each genotype ...... 107 Figure 5.3: Symbiodiniaceae cell density at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all treatments within each genotype ...... 109 Figure 5.4: ROS levels at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all treatments within each genotype ...... 111 Figure 5.5: PCoA ordinations for each genotype ...... 113 Figure 5.6: Relative abundance of probiotic ASVs in the AIMS2 anemones ...... 114 Figure 5.7: Relative abundance of probiotic ASVs in the AIMS3 anemones ...... 115 Figure 5.8: Relative abundance of probiotic ASVs in the AIMS4 anemones ...... 116 Figure A 1: E. diaphana acquired from the AIMS SeaSim in late 2014 ...... 129 Figure A 2: E. diaphana acquired from the AIMS SeaSim in early 2016 ...... 130 Figure A 3: E. diaphana culture collection at UoM ...... 131 Figure A 4: (a) Outflow with anemones, and (b) AIMS holding tank ...... 132 Figure A 5: Rarefaction curves for all microbiome characterisation samples ...... 133 Figure A 6: Relative proportions of family level ASVs ...... 134 Figure A 7: Legend of family-level ASVs ...... 135 Figure A 8: Heatmaps of the top 20 AIMS1–4 taxa by relative abundance at (a) order, and (b) family level ...... 139 Figure A 9: Rarefaction curves for all samples ...... 146 Figure A 10: Relative abundance of reads assigned to genus in each mock community sample ...... 147

Figure A 11: Unique and common bacterial ASVs in control and heat-treated anemones at Day 0 or 14 ...... 147 Figure A 12: ASVs identified by IndVal and considered potential indicator species: (a) Saprospiraceae, (b) Gammaproteobacteria, (c) Terasakiellaceae, (d) Terasakiellaceae, (e) Spirochaeta 2, and (f) Rhizobiaceae ...... 148 Figure A 13: ASVs identified by IndVal but discounted as potential indicator species due to erratic changes in relative abundance and/or high variance: (a) Oligoflexaceae, (b) Rubinisphaeraceae, (c) Saprospiraceae, (d) Chitinophagales, (e) Oligoflexaceae, and (f) Simkaniaceae ...... 149 Figure A 14: Example of ddPCR fluorescence output for four paired host-bacteria reactions ...... 151 Figure A 15: Rarefaction curves for all samples ...... 158 Figure A 16: nMDS (weighted unifrac) of all samples ...... 159 Figure A 17: DPPH qualitative assay results: (a) Positive indication of FRS ability for a bacterial isolate due to appearance of halo in <1 min; (b) Negative indication of FRS ability for a bacterial isolate due to absence of halo after >3 min...... 165 Figure A 18: Water temperature daily averages (2 m depth) from St Crispin Reef ...... 167

List of Tables

Table 1.1: E. diaphana genotypes commonly used in research...... 25 Table 2.1: Sampling summary for bacterial microbiota analysis...... 32 Table 2.2: Sampling summary for Symbiodiniaceae analysis...... 33 Table 2.3: Number of bacterial ASVs in each sample type...... 37 Table 2.4: Phyla identified in the collective AIMS1–4 animals, wild proxies, CC71, H21, and previously reported ...... 40 Table 2.5: Relative abundance of bacterial ASVs present in every AIMS1–4 sample, and co-occurrence in other sample types...... 44 Table 4.1: Maintenance, antibiotic treatment and sampling schedule...... 78 Table 4.2: Antibiotics used against E. diaphana- and A. salina-associated bacteria...... 78 Table 4.3: Primers used to estimate host and bacterial cell numbers...... 80 Table 4.4: ‘Stable’ bacterial genera detected in the untreated Day 0 and antibiotic-treated Day 21 anemones with relative abundances >1.0% and changes in relative abundance <2.0% .... 91 Table 5.1: Bacteria used in the positive and negative probiotics ...... 99 Table 5.2: Cell densities calculated from CFU counts of bacteria cultured for probiotic dosing on Days 0, 2, and 7 ...... 106

Table 5.3: ANOVA output for Fv/Fm in each genotype ...... 108 Table 5.4: ANOVA output for Symbiodiniaceae cell density in each genotype ...... 110 Table 5.5: ANOVA output for ROS levels in each genotype...... 112 Table 5.6: p-values from GLM analyses comparing bacterial communities of anemones receiving different probiotic treatments ...... 117 Table A 1: Contaminant ASVs removed from the characterisation dataset...... 133 Table A 2: Tukey’s HSD p-values from pair-wise Shannon value comparison ...... 136 Table A 3: Relative abundance of phyla in each AIMS1–4 genotype ...... 137

Table A 4: Pairwise binary log fold change (L2FC) for the 20 most abundant class-level taxa ...... 138

Table A 5: Pairwise binary log fold change (L2FC) for the 20 most abundant genus-level taxa ...... 138

Table A 6: Pairwise binary log fold change (L2FC) for the 20 most abundant order-level taxa ...... 140

Table A 7: Pairwise binary log fold change (L2FC) for the 20 most abundant family-level taxa ...... 140 Table A 8: Output from GLM-based analysis comparing the AIMS1–4 bacterial community compositions (‘tank’ nested within ‘genotype’) ...... 141 Table A 9: Output from GLM-based analysis comparing the AIMS1–4 bacterial community compositions (‘tank’ and ‘genotype’ specified separately) ...... 141 Table A 10: Mock community members ...... 142

Table A 11: GLM analysis of differences in bacterial community beta diversity based on treatment (control vs heat-treated) and time ...... 142 Table A 12: Putative contaminant ASVs removed from the dataset ...... 143 Table A 13: GLM analyses comparing bacterial community beta diversity at each sampling timepoint ...... 144 Table A 14: Potential indicator species identified in an IndVal analysis ...... 145 Table A 15: Overall differences in Symbiodiniaceae cell density between control and treated anemones tested by a GLS model ...... 150 Table A 16: Homogeneity of variance and normality test p-values for algal cell densities ...... 150 Table A 17: Mann-Whitney U test p-values of control vs treated anemone algal cell density at each sampling timepoint ...... 150 Table A 18: Mann-Whitney U test p-values for differences in control and treated anemone algal cell density for Day 0 versus Day 21 ...... 150 Table A 19: ddPCR data for the bacterial/16S reactions ...... 152 Table A 20: ddPCR data for the E. diaphana/Ef1-α reactions ...... 154 Table A 21: Overall differences in bacterial load ratios between control and treated anemones tested by a GLS model ...... 156 Table A 22: Homogeneity of variance and normality test p-values for bacterial load ratio data ...... 156 Table A 23: Mann-Whitney U test p-values of control vs treated bacterial load ratios at each sampling timepoint ...... 156 Table A 24: Mann-Whitney U test p-values for differences in bacterial load ratios within the control and treated anemones at selected timepoints ...... 156 Table A 25: Analysis of statistical difference between untreated and treated A. salina ...... 157 Table A 26: Contaminant ASVs removed from the dataset ...... 158 Table A 27: Analysis of overall differences in number of observed ASVs between control and treated anemones tested by a GLS model ...... 160 Table A 28: Homogeneity of variance and normality test p-values for number of observed ASVs .... 160 Table A 29: Kolmogorov-Smirnov test p-values for number of observed ASVs in control vs treated anemones at each sampling timepoint ...... 160 Table A 30: Kolmogorov-Smirnov test p-values for number of observed ASVs in control and treated anemones for day 0 versus day 21 ...... 160 Table A 31: Analysis of overall differences in Simpson index values between control and treated anemones tested by a GLS model ...... 161 Table A 32: Homogeneity of variance and normality test p-values for Simpson Index values ...... 161 Table A 33: Kolmogorov-Smirnov test p-values for Simpson Index values in control vs treated anemones at each sampling timepoint ...... 161 Table A 34: Kolmogorov-Smirnov test p-values for number of Simpson Index values in control and treated anemones for Day 0 versus Day 21 ...... 161 xii

Table A 35: Analysis of overall differences in Shannon index values between control and treated anemones tested by a GLS model ...... 162 Table A 36: Homogeneity of variance and normality test p-values for Shannon index values ...... 162 Table A 37: Kolmogorov-Smirnov test p-values for Shannon index values in control vs treated anemones at each sampling timepoint ...... 162 Table A 38: Kolmogorov-Smirnov test p-values for Shannon index values in control and treated anemones for day 0 versus day 21 ...... 162 Table A 39: Differences in beta diversity between control and antibiotic-treated E. diaphana bacterial communities tested by Generalised Linear Models ...... 163 Table A 40: R2A broth adjusted to suit marine bacteria ...... 166

Table A 41: Fv/Fm comparisons for AIMS2 anemones at Day 43 ...... 168

Table A 42: Fv/Fm comparisons for AIMS3 anemones at Day 43 ...... 168

Table A 43: Fv/Fm comparisons for AIMS4 anemones at Day 43 ...... 168 Table A 44: Symbiodiniaceae cell density comparisons for AIMS2 anemones at Day 43 ...... 169 Table A 45: Symbiodiniaceae cell density comparisons for AIMS3 anemones at Day 43 ...... 169 Table A 46: Symbiodiniaceae cell density comparisons for AIMS4 anemones at Day 43 ...... 169 Table A 47: ROS comparisons for AIMS2 anemones at Day 43 ...... 170 Table A 48: ROS comparisons for AIMS3 anemones at Day 43 ...... 170 Table A 49: ROS comparisons for AIMS4 anemones at Day 43 ...... 170 Table A 50: The 20 most abundant contaminant ASVs identified by decontam across all anemone samples ...... 171 Table A 51: GLM analysis of bacterial community differences in AIMS2 based on probiotic and temperature treatments ...... 172 Table A 52: GLM analysis of bacterial community differences in AIMS3 based on probiotic and temperature treatments ...... 172 Table A 53: GLM analysis of bacterial community differences in AIMS4 based on probiotic and temperature treatments ...... 172

xiii

Chapter 1: General introduction

1.1 Coral reefs

Coral reefs are highly productive, diverse ecosystems, and are vitally important to the marine environment. Despite occupying <0.1% of the marine environment, typically along coastal areas between latitudes 25°N and 25°S, they harbour ~25% of known marine organisms (Hoegh-Guldberg 1999). One third of all marine fish species are also estimated to spend part of their lives on coral reefs (Choat & Bellwood 1991).

In addition to providing a habitat and nursery environment for many marine species, coral reefs provide a number of other services. For example, they absorb energy from tidal surges and storms, thereby protecting shorelines and shoreline communities (Ferrario et al. 2014; van Zanten, van Beukering & Wagtendonk 2014). Reefs also provide food to an estimated 1 billion people in those communities (Carilli 2013). Moreover, their economic value is substantial, with the Great Barrier Reef (GBR) contributing more than $6 billion to Australia’s economy per annum (Deloitte Access Economics 2017).

Coral reefs may also provide benefits that are yet to be realised. For example, due to their high species diversity reefs may contain bioactive compounds with properties that could be useful in medicine or biotechnology (Bruckner 2002). Together with coral reefs’ intrinsic value as part of the natural world, these factors highlight the importance of coral reef systems, and the need to protect them and improve our understanding of them.

Figure 1.1: Coral assemblage in Palmyra Atoll, Central Pacific (Maragos 2011). 1

1.2 The coral organism

Corals are marine invertebrates of the phylum Cnidaria, and class Anthozoa. The cnidarians that build modern reefs, emerged ~250 million years ago. However, the phylum has ancient roots, with Cnidaria and Bilateria diverging from a common ancestor more than 700 million years ago (Lipps & Stanley 2016).

Anthozoans comprise those cnidarians that are almost exclusively sessile, and reproduce either sexually without a medusa stage, or asexually primarily by budding. They are further divided into three sub-classes, Hexacorallia, Octocorallia and Ceriantharia, with Hexacorallia containing the so-called hard corals. It is these hard corals of the order Scleractinia that create the reef’s physical foundation (Kitahara et al. 2016; van Iten et al. 2016). Therefore, scleractinian corals are the primary ecosystem engineers of coral reefs, and as such, are essential for their growth and persistence (Wild et al. 2011).

Cnidarians have a simple, radially symmetrical body plan, with a central mouth surrounded by tentacles containing stinging cells known as cnidocytes that are used to catch prey using harpoon-like nematocysts (Figure 1.2). Coral colonies are modular and typically comprise many genetically identical modules, the polyps. Corals build the reef’s foundation by secreting a calcium carbonate skeleton, which accumulates over successive generations. Polyps have two layers of cells, the epidermis and gastrodermis, which are separated by an acellular mesogloea and attached to a porous skeleton. The epidermis is covered by an exterior surface mucus layer (SML), which acts as a protective barrier (Brown & Bythell 2005).

Tentacles Symbiodiniaceae Surface mucus layer

Epidermis

Mouth Gastrodermis Gastrovascular Mesenterial cavity flaments Coernosarc Mesogloea

Calcium Cnidocyte carbonate skeleton

Figure 1.2: Coral polyp body plan and cell types. The living tissue of a coral colony is like a veneer atop their secreted skeletons, and those of their colonial predecessors (adapted from Elder 2015).

1.3 Coral symbioses

1.3.1 Symbiodiniaceae

At the heart of corals’ ability to live in the oligotrophic waters typical of reef environments is the mutualistic relationship they form with dinoflagellate algae of the family, Symbiodiniaceae. These algae, encapsulated within vacuoles called symbiosomes, photosynthetically fix carbon, most of which is passed to the host (Gordon & Leggat 2010). This supplements the carbon that corals obtain by capturing prey to provide a substantial portion of the host animal’s daily carbon requirements (Muscatine, McCloskey & Marian 1981), thus enabling scleractinian corals to synthesize their calcium carbonate skeletons (Lough 2011).

Due to their morphological uniformity, Symbiodiniaceae were originally thought to comprise a single species, Symbiodinium microadriaticum (Stat, Carter & Hoegh-Guldberg 2006). However, genetic analyses have revealed at least nine genera, and many species (LaJeunesse et al. 2018). As different types live symbiotically within coral hosts or independently in the marine environment (LaJeunesse 2002), it is possible that many more await identification (Parkinson, Coffroth & LaJeunesse 2015).

Symbiodiniaceae acquisition by coral can occur by either horizontal, or vertical (i.e., maternal) transfer. In horizontal transfer, aposymbiotic (Symbiodiniaceae-free) larvae or early recruits acquire Symbiodiniaceae from the environment, whereas in vertical transfer, Symbiodiniaceae are inherited from the parent (Stat, Carter & Hoegh-Guldberg 2006). There is evidence for host-Symbiodiniaceae specificity in some coral species (Howells et al. 2020; Thornhill et al. 2009), and flexibility in others (Gómez-Cabrera et al. 2008). Coral may also be capable of switching, or altering the relative abundance of their algal symbionts (i.e. ‘shuffling’) in response to environmental stress (Lewis, Neely & Rodriguez-Lanetty 2019), therefore, it has been suggested that host-Symbiodiniaceae specificity is best viewed as a gradient among coral species (Silverstein, Correa & Baker 2012). Changes in Symbiodiniaceae communities harboured by coral colonies may improve tolerance, but can also reduce host growth rate (Little, van Oppen & Willis 2004).

Symbiodiniaceae also supply oxygen, and amino acids used to synthesise proteins, fats and carbohydrates (Barott et al. 2015; Wang & Douglas 1999); therefore, the coral- Symbiodiniaceae relationship is essential to corals’ survival. However, Symbiodiniaceae is but one of many microorganisms that intimately coexist with coral. Like many multicellular organisms, corals harbour communities of prokaryotes (bacteria and archaea), microscopic eukaryotes (e.g. fungi), and viruses, which together with Symbiodiniaceae constitute the coral microbiome (Whipps, Lewis & Cooke 1988). The coral host, combined with its microbiome, is commonly called the coral holobiont (Rohwer et al. 2002); a term coined to describe an organism and the full complement of other organisms that reside upon, and within it (Margulis 1991).

1.3.2 Bacteria

Bacteria are ubiquitous throughout the earth’s biosphere, and have colonised all habitable niches, including the water, sediments, plants, and animals of the marine environment. They also play a central role in biogeochemical cycles (Whitman, Coleman & Wiebe 1998). Therefore, the part they play in making nutrients and trace elements available to higher organisms in the oligotrophic waters of reef systems is vital (Bourne & Webster 2013), and identifying the bacterial members of the microbiome, and how they contribute to coral function is an important endeavour. 4

Despite their importance, studying bacterial communities in coral holobiont research has historically been difficult because most bacteria have defied cultivation. Impediments may include cell dormancy, and medium conditions not suited to microorganisms that are fastidious or adapted to oligotrophic environments (Connon & Giovannoni 2002). Fortunately, the advent of technologies, such as high throughput sequencing that allows identification based on phylogenetically informative marker gene sequences, has enabled bacterial community characterisation (Bassis, Young & Schmidt 2013).

Subsequent investigations of coral-associated bacteria have produced conflicting results (Hester et al. 2016). For example, some studies have found evidence for host-species specificity and geographic homogeneity under normal conditions (Kemp et al. 2015; Kvennefors et al. 2010; Rohwer et al. 2002), whereas others have found that some coral species’ bacteria vary across different geographic locations (Littman et al. 2009; Rodriguez- Lanetty et al. 2013).

Variability under non-normal conditions is not surprising as coral microbiomes, like most ecological communities, are dynamic and undergo change in response to environmental factors (Kvennefors et al. 2010; McFall-Ngai et al. 2013; Zaneveld et al. 2016). These factors may be of anthropogenic origin, or naturally occurring phenomena, such as seasonal weather patterns, which alter dissolved oxygen content and temperature (Chen et al. 2011; Li et al. 2014).

Despite apparent variability, some studies have revealed the common occurrence of bacterial phylotypes among coral lineages, or across species and geographic locations, suggesting that a core coral microbiome may exist and be necessary for maintaining host health (Ainsworth et al. 2015; Fernando et al. 2014). A study exploring this concept, investigated the bacterial communities of Pachyseris speciosa in two adjacent reef systems. The findings led the authors to propose that the coral microbiome comprises three distinct fractions: a small collection of symbiotic bacterial phylotypes (i.e., a core microbiome) that is stable across broad geographic scales; a larger collection of phylotypes (<100) present in ecosystem inhabitants that are necessary for success under local conditions; and a very large collection of phylotypes (>100,000) that vary with changing environmental conditions (Hernandez-Agreda et al. 2016). This model accommodates both the core microbiome concept, and also microbiome variability. 5

Studies of other systems, most notably the human gut microbiome, have recommended that obligate, or ‘core’ community membership is best viewed in terms of functionality, rather than taxonomic identity (Boon et al. 2014; Gevers et al. 2012; Krediet et al. 2013). Indeed, coral microbiome researchers who have found groups of consistently occurring bacteria have cautioned that this may simply be evidence of selection based on function, rather than species specificity (Li et al. 2014; Li et al. 2013).

The apparent partitioning of microbial assemblages further complicates efforts to understand the coral microbiome. An early investigation found that bacterial populations differ between the environment (i.e. water column), and host (Frias-Lopez et al. 2002). Although the methodology of that work was questioned (Rohwer et al. 2002), later studies confirmed this finding, and also revealed microbiome partitioning within the coral organism between the SML, tissue and skeleton (Apprill, Weber & Santoro 2016; Li et al. 2014). This latter finding suggests that viewing a coral microbiome as a single entity may oversimplify it (Bourne, Morrow & Webster 2016).

Nevertheless, overall community data does provide insight into bacterial species that commonly associate with coral. For example, a meta-analysis using publicly available sequence data from research on coral-associated prokaryotes was conducted by Blackall, Wilson and van Oppen (2015) to explore overall diversity. Next generation sequencing (NGS) datasets, from nine studies of coral SML, tissue and skeleton microbiomes, were pooled to create an overview of coral-associated prokaryotes (Figure 1.3). The authors noted the dominance of Proteobacteria, and the high relative abundance of Gammaproteobacteria, and Alphaproteobacteria from samples across the two databases. However, there was considerable variation at lower taxonomic levels. The authors cautioned however that the data could have been skewed by differences in experimental methods, or a focus by researchers on pathogenic bacteria.

A A rchaea B Thaumarchaeota 3% Acidomicrobiia 3% Thermoplasmata 2% Actinobacteria 3% Acidimicrobiia 2% Gammaproteobacteria 38% Gammaproteobacteria 35% Actinobacteria 3% Flavobacteriia 4% Cytophagia 2% Clostridia 1% Bacteroidia 2%

Flavobacteriia 6% Bacilli 10%

Anaerolineae 1%

Synechococcophycideae 2%

Oscillatoriophycideae 2% Deltaproteobacteria 3%

Bacilli 2%

Clostridia 2% Alphaproteobacteria 8% Betaproteobacteria 2%

Deltaproteobacteria 3%

Alphaproteobacteria 22% Betaproteobacteria 20%

Proteobacteria Cyanobacteria Bacteroidetes Euryarchaeota Firmicutes Chloroflexi Actinobacteria Crenarchaeota

Figure 1.3: Coral-associated prokaryotes (>1% relative sequence abundance only) at phylum level (colour-ways), and class level (wedges). The charts compare 16S rRNA gene sequence data held in the (a) GenBank, or (b) NCBI SRA databases (Blackall, Wilson & van Oppen 2015).

A more-recent meta-analysis using data from 16 culture-dependent and independent studies was reported (Sun, Anbuchezhian & Li 2016). No assessment of relative abundance was made, however, the prevalence of Proteobacteria across the surveyed studies was again noted (Sun, Anbuchezhian & Li 2016). Yet another review highlighted the dominance of Proteobacteria, and also Actinobacteria, Bacteroidetes and Cyanobacteria among coral symbiont surveys (Bourne, Morrow & Webster 2016). The authors suggested that while informative, deeper understanding of coral host-microbe relationships, and the coral holobiont’s response to perturbation, might be more easily achieved by identifying and focussing on the few core microbiome members that persist spatially and temporally. This suggestion echoes calls for a core-focussed approach, which could overcome the complexity of analysing naturally variable coral microbiomes (Hernandez-Agreda et al. 2016; Sharp, Distel & Paul 2012).

1.3.3 Bacterial functions: Pathogen protection

Bacterial members of the coral microbiome provide services essential for maintaining host health, including protection from coral pathogens. For example, in a study using bacterial

isolates cultured from Acropora palmata SML, 8% of the isolated bacteria inhibited growth of the bacterium Serratia marcascens, which causes white pox disease (Ritchie 2006). In a later study, 12% of bacteria isolated from Montastrea annularis were able to inhibit Vibrio shiloi, which proliferates at elevated sea surface temperatures (SST) and attacks the Symbiodiniaceae symbiont (Rypien, Ward & Azam 2010). Although this attack had been suggested as the cause of Symbiodiniaceae expulsion from the host (Rosenberg & Falkovitz 2004), a later study refuted this, claiming instead that the elevated SST was the driver, and that Vibrio opportunistically proliferated after the host had been weakened (Ainsworth et al. 2008).

In another example of pathogen protection, germ free (GF) members of the cnidarian freshwater animal Hydra succumbed to infection when exposed to fungi of the genus, Fusarium, but if inoculated with native microbiomes they were largely protected. Interestingly, GF animals inoculated with only one or two native bacteria were generally not protected (Fraune et al. 2015).

Predation of pathogenic bacteria by benign bacteria has also been suggested as a protective mechanism. A study by Welsh et al. (2015) found that Halobacteriovorax, isolated from Montastraea cavernosa, was able to attack and ingest three pathogenic Vibrio bacteria. Halobacteriovorax are widely distributed in the marine environment (Pineiro et al. 2007) and have been detected in other coral species (Wegley et al. 2007), therefore they may be common, beneficial members of the coral microbiome.

Benign bacteria are also thought to protect hosts from pathogens through competition for nutrients and occupation niches (Bourne, Morrow & Webster 2016; Rohwer et al. 2002). This mechanism, referred to as colonisation resistance or competitive exclusion, is commonly observed in nature, for example in the mammalian gut epithelium (Buffie & Pamer 2013), plant leaves (Lindow 1987) and rhizospheres (Cordovez et al. 2015).

1.3.4 Bacterial functions: Nutrient cycling

As shown below (Figure 1.4), nutrient cycling between the environment and host, and within the host itself, is an important service provided by the coral microbiome (Bourne & Webster 2013; Mouchka, Hewson & Harvell 2010). For example, although corals can obtain nitrogen

by heterotrophic feeding, nitrogen fixation is largely dependent on diazotrophic bacteria and archaea (Bourne & Webster 2013). The reduction of gaseous N2 to ammonium by diazotrophs is essential for both the host and Symbiodiniaceae. Interestingly, as ammonium availability is a limiting factor in Symbiodiniaceae growth, it has been suggested that nutrient limitation may be one of the methods used by coral hosts to regulate Symbiodiniaceae populations (Falkowski et al. 1993; Rädecker et al. 2015).

Figure 1.4: Nutrient cycling performed by coral-associated microbial communities (Bourne & Webster 2013).

Despite also being essential for life, less is known about sulphur and phosphorus cycling than nitrogen cycling in the coral holobiont (Sun, Anbuchezhian & Li 2016). However, research has found evidence that bacteria are likely involved in these processes. For example, Raina et al. (2009) observed that bacteria from the genera Roseobacter, Spongiobacter, Vibrio and 9

Alteromonas, that were isolated from Montipora aequituberculata and Acropora millepora, were capable of using dimethylsulfoniopropionate (DMSP) or dimethyl sulfide (DSP) as a nutrient source. Identification of bacterial genes necessary for DMSP degradation suggested that these bacteria were involved in sulphur cycling.

Similarly, Wegley et al. (2007), in their investigation of bacteria associated with Porites astreoides, noted a high abundance of genes for transport and degradation of glutathione among the bacteria identified. As glutathione can be used as an organic source of sulphur and is abundant in eukaryotic cells, the authors suggested that these genes indicated the use of glutathione as a sulphur source by bacteria.

Phosphonoacetate constitutes up to 25% of the dissolved organic phosphorus in seawater, and is an important phosphorus source for marine organisms (Sun, Anbuchezhian & Li 2016). However, it must first be cleaved to release acetate and phosphate. This can be done by phosphonoacetate hydrolase, which is encoded by the phnA gene. DNA taken from thirteen coral species, including one hard coral of the genus Turbinaria, and amplified with primers targeting phnA, identified eight phnA+ bacteria, demonstrating that they could potentially catabolise phosphonoacetate to make phosphorus available to the host (Thomas et al. 2010). Among the phnA+ bacteria identified was one Vibrio species. Interestingly, in this study, and in the Raina et al. (2009) investigation of sulphur, Vibrio species were identified as contributors to nutrient cycling suggesting that, providing they are kept in check, Vibrio can benefit the host.

1.3.5 Bacterial functions: Host development

Bacteria play important roles in coral development by providing the protective and nutritive services described above (Ceh et al. 2013). Nutrient cycling, for example, may be particularly important to offspring of spawning corals as they are initially aposymbiotic.

Vertical or horizontal bacterial acquisition by coral offspring has generally been associated with brooding or spawning reproductive strategies respectively (Sharp, Distel & Paul 2012; Sharp et al. 2010). However, new studies have found evidence for mixed modes of acquisition in both coral types (Bernasconi et al. 2019; Epstein et al. 2019b). It is possible that correlations between the microbiomes of spawners and their offspring are due to their shared

environment, rather than parental transfer. Nevertheless, vertically transmitted bacteria could also enhance larval survival by ensuring the intergenerational maintenance of co- evolved microbial partners (Sharp, Distel & Paul 2012) whilst horizontally acquired bacteria could improve adaptation of coral offspring to their immediate environment (Epstein et al. 2019b).

The bacterial communities of corals are highly dynamic over their early life stages (Damjanovic et al. 2019a). Coral offspring initially harbour diverse, variable bacterial communities, but undergo a ‘winnowing’ as they develop, resulting in simpler bacterial communities comparable to their parents (Bernasconi et al. 2019; Epstein et al. 2019b). This suggests that microbiome flexibility is an inherent trait of the coral holobiont. Bacteria of the genera Alteromonas and Roseobacter have been consistently found in high abundance in the offspring of both brooding and spawning corals (Ceh, van Keulen & Bourne 2013; Damjanovic et al. 2019a; Freire et al. 2019). This points to an important role in development and larval settlement by these bacteria, however a mechanistic link is yet to be identified.

1.4 Impacts on coral reefs

1.4.1 Natural events

The primary natural causes of damage to coral are periodic tidal emersions, which expose coral to bright sunlight leading to thermal stress or dehydration, and storm damage, which can physically damage coral, particularly those with branching morphologies. However, coral have evolved in the presence of these disturbances, and have shown recovery rates in line with the historical occurrence of average-sized events, therefore they do not cause permanent damage to reefs systems (Barnes & Hughes 2009).

1.4.2 Human activity

Corals have not, however, evolved in the presence of the man-made threats now facing them, and as a result, reefs have dramatically declined. For example, over-harvesting has altered the ecological balance of reef environments, removing species vital for the maintenance of food webs and coral health (Zaneveld et al. 2016).

The crown of thorns starfish, Acanthaster planci, which eats hard corals, illustrates the importance of maintaining balance within ecosystems. Since it first rose to prominence in the 1960s it has become a major threat to coral species on the GBR as a result of anthropogenic influences. It is believed that the removal of the starfish predator Charonia tritonis by shell collectors, in addition to nutrient runoff, are among the causes that have led to periodic proliferation of A. planci, resulting in substantial coral loss (Reichelt, Bradbury & Moran 1990).

Nutrient runoff also allows pathogenic bacteria to proliferate. In aquatic plants, pathogenic bacteria are kept in check through competition for nutrients. However, this protective mechanism is disrupted when excess nutrients, particularly nitrogen and phosphorus, are introduced (Smith 2002). For example, a correlation has been observed between the incidence of the coral disease atramentous necrosis in Montipora aequituberculata on the GBR and seasonal runoff (Haapkylä et al. 2011). Although seasonal runoff is a natural event, its nutrient content has increased substantially due to modern agricultural practices, such as cattle grazing and fertilised cropping (Brodie & Mitchell 2005).

However, foremost among the anthropogenic threats now facing coral reefs is climate change linked to increased atmospheric CO2 (Hoegh-Guldberg 1999; Hughes et al. 2017a). As corals tend to exist near the upper limit of their temperature range, they are particularly sensitive to SST increases (Coles, Jokiel & Lewis 1976). Therefore, exposure to unusually high temperatures resulting from climate change can trigger expulsion of the algal symbiont in a process referred to as bleaching (Figure 1.5) (Baker, Glynn & Riegl 2008). Due to its reliance on its algal partner, bleaching can then lead to the host’s death unless the symbiont is replaced, and normal environmental conditions return (McClanahan et al. 2009). The point at which thermally induced bleaching is triggered varies between coral species, and within species in different locations, but it can be induced by increases of as little as 1 °C above the normal mean temperature if maintained over several weeks (Eakin, Lough & Heron 2009).

(a) (b)

Figure 1.5: (a) Coral near Moorea, French Polynesia, in March 2019 with some signs of bleaching, and (b) in May 2019 after widespread bleaching due to prolonged exposure to elevated SST (Rocha 2019a, b).

The oxidative theory of coral bleaching asserts that an accumulation of reactive oxygen species (ROS) causes a breakdown of the symbiotic relationship between Symbiodiniaceae and coral (Mydlarz, McGinty & Harvell 2010). ROS are produced under elevated temperature and irradiance by both symbiont and host as a result of physical damage to cellular components, and subsequent disturbance of metabolic pathways. In Symbiodiniaceae, thylakoid membranes are damaged, which disrupts ATP and NADPH production. This leads to a surplus of electrons in the chloroplasts, which promotes ROS formation. In the host, elevated temperature and irradiance causes mitochondrial damage, which also results in ROS formation. Although the holobiont has protective enzymes, such as catalase and superoxide dismutase to convert ROS to water and oxygen, under bleaching conditions these systems become overwhelmed and free ROS circulate causing damage to the cellular components of both partners. The expulsion of the symbiont has been likened to an innate immune response by the host to minimise exposure to symbiont-generated ROS, however it is unclear whether the symbiont is expelled, or actively leaves the host (Weis 2008).

Another by-product of increased atmospheric CO2 is ocean water pH reduction known as

“ocean acidification”. The oceans have absorbed almost one third of anthropogenic CO2 emissions since pre-industrial times leading to a reduction in seawater pH and carbonate ion concentration. Both outcomes impact coral. The skeletons of hard corals are made from the soluble form of calcium carbonate, aragonite, which renders them vulnerable to reduced pH.

Aragonite deposition by coral polyps is also impeded by reduced carbonate ion concentration (Meyer, Cardini & Wild 2015).

SST increases and ocean acidification both cause shifts in the coral microbiome, which, when combined, may have consequences for the host. For example, Vega Thurber et al. (2009) showed that high SST can cause shifts in microbiome composition. Porites compressa was subjected to temperatures 5 °C above normal, resulting in a shift from a mutualistic or commensal microbiome, to one that contained more bacteria and fungi associated with disease. However, based on their previous work the authors speculated that the changes were transitory, therefore coral disease associated with elevated SST may be the result of simultaneous exposure to other stressors, such as reduced pH. The findings of Meron et al. (2011) may support this hypothesis. These researchers found that exposing Acropora eurystoma to reduced pH also led to higher relative abundance of disease-related bacteria. This suggests that together, elevated SST and reduced pH could shift and then maintain microbiomes with pathogenic bacteria. This scenario is in line with predictions that stressors impacting coral are likely to act synergistically (Veron et al. 2009).

As a result of the multiple stressors impacting coral, coral cover across the Caribbean basin declined by 80% in the 30 years up to 2002 (Gardner et al. 2003), and by 50% on the GBR between 1985 and 2012 (De’ath et al. 2012). Mass bleaching events from 2015 to 2017 have impacted coral reefs further, dramatically reducing coral cover and damaging reef ecosystems worldwide (Hughes et al. 2018; Hughes et al. 2019).

1.5 Solutions

1.5.1 Climate change mitigation

The connection between climate change and coral reef decline is now almost universally accepted. Increased SST resulting from climate change currently poses the greatest threat to coral health, and is projected to increase until it surpasses coral’s upper thermal threshold (Hughes et al. 2003). However, governments have struggled to introduce policies that adequately address the causes of climate change, such as greenhouse gas emissions (Howlett & Kemmerling 2017), or adhere to targets already set (Maltais 2014). Therefore, scientific solutions that can improve coral’s thermal tolerance must be investigated. 14

1.5.2 Assisted evolution

Microbiomes that provide the right balance of function, including metabolism, are necessary for host development and ongoing health. This has been observed in insect and animal models such as Drosophila, zebrafish and mice, and also in humans (Gevers et al. 2012; Round 2013). However, if an imbalance that affects host health occurs (i.e., ‘dysbiosis’), transfer of the microbiota from healthy donors can restore normal function (Biedermann & Rogler 2015). Interestingly, in some non-dysbiotic situations, probiotic supplementation of normal microbiomes can also confer beneficial traits, such as drought resistance in plants (Coleman- Derr & Tringe 2014), and heat-stress tolerance in fish (Taoka, Maeda & Jo 2006).

This has led researchers to consider whether such an approach could be used to improve corals’ ability to withstand environmental challenges (Epstein et al. 2019a; Wilkins et al. 2019). The seeds of this idea were likely sown by Reshef et al. (2006) in their proposal of the Coral Probiotic Hypothesis, which posits that the dynamic nature of the coral holobiont promotes selection for prokaryotes that improve host fitness. Based on their hypothesis, they predicted that bacterial communities transferred from a stress-adapted coral donor to a non-adapted coral could accelerate the recipient’s adaptation to the stressor.

Similarly, in an evaluation of methods for protecting transplanted corals in reef restoration, Teplitski and Ritchie (2009) suggested that inoculating transplants with bacteria from corals in the destination reef (i.e., those already adapted to the destination conditions), might help protect the transplanted corals from pathogens and stressors in the new location. It is likely that without inoculation the transplanted corals’ bacterial communities would transform naturally over time in response to the new environment and exposure to the microorganisms in it. However, the rationale is that inoculation could help direct and accelerate that outcome to improve transplant survival rates. Given the genetic diversity among prokaryotes and other coral-associated microorganisms, it is plausible that microbiome transplantation or supplementation could enhance adaptive capability.

This type of manipulation was included in a range of natural, albeit human-assisted, methods for improving coral stress tolerance proposed by van Oppen et al. (2015), which also included induced acclimatisation, selective breeding, and hybridisation. Each method has the potential to produce variants better adapted to environmental conditions resulting from climate

change, and as they can occur as natural evolutionary processes, were dubbed ‘assisted evolution’ (van Oppen et al. 2015).

Peixoto et al. (2017) also proposed inoculation with what they termed ‘beneficial microorganisms for coral’ (BMCs) to improve coral health and resilience. The authors suggested that host-derived bacteria with selected traits could be assembled into probiotic consortia for use on corals under stress (Figure 1.7). They acknowledged however, that identifying, isolating and administering a consortium would be challenging, as would accounting for the complex interactions of holobiont members. Nevertheless, Rosado et al. (2018) showed that such an approach could mitigate bleaching in a thermally-stressed coral challenged with pathogenic infection. These results, and the perilous situation facing coral reefs, suggest that bleaching mitigation strategies using probiotic inoculation warrant further investigation.

Figure 1.7: Simplified depiction of inoculation with BMCs to restore coral health (Peixoto et al. 2017).

1.6 Cnidarian model animals

Growth and maintenance of coral in laboratory settings is hampered by the requirement for large tanks, pumps, lights and a constant supply of clean seawater. Completing experimental work quickly, and collecting sample tissue, is also impeded by coral’s slow growth and

calcareous skeleton. However, other cnidarians without these characteristics can be used to study cnidarian biology. Although there are several candidates, not all are ideal.

1.6.1 Hydra

Cnidarian animals of the genus Hydra (class: Hydrozoa; Order: Anthomedusae; Family: Hydridae) may be considered the first marine model organisms. They were first investigated in detail by Abraham Trembley in 1744, and helped bring marine biology into the laboratory (Ratcliff 2012). Since then, they have been studied extensively, and are now used in a broad range of scientific research, including stem cell and developmental biology research (Galliot 2012). Hydra’s prolific ability to reproduce asexually from severed parts and cell aggregates allows the production of large clonal populations for experimental work. Culturing and general experimental methods are also well developed (see Lenhoff 1983). However, only one Hydra species, H. viridis, harbours an algal symbiont, and it is a member of the taxonomic family Chlorellaceae (genus, Chlorella), not Symbiodiniaceae (Habetha et al. 2003). Moreover, H. viridissima is a freshwater species. Therefore, no member of the Hydra genera can be regarded an ideal coral model organism.

1.6.2 Nematostella vectensis

Compared to Hydra, the saltwater anemone, Nematostella vectensis (Class: Anthozoa; Order: Actiniaria; Family: Edwardsiidae), is a relative newcomer (Darling et al. 2005). A mere twenty- eight years ago Hand and Uhlinger (1992) proclaimed its potential as a cnidarian research model. This was based partly on its easy cultivability. For example, it can reproduce asexually or sexually, is capable of reaching sexual maturity in <2 months, and can be induced to spawn as frequently as every 8 days (Hand & Uhlinger 1992; Hand & Uhlinger 1994). As an anthozoan it is more closely related to coral than Hydra, however, it does not harbour an algal symbiont, and therefore cannot be used to study bleaching, nor any aspect of Symbiodiniaceae.

1.6.3 Aiptasia pallida

Sea anemones of the genus Aiptasia (Class: Anthozoa; Order Actiniaria, Family Aiptasiidae) have also been used to study aspects of cnidarian biology and development for several decades, and for other purposes, for example as pollution bio-monitors (Howe, Reichelt- Brushett & Clark 2012; Nur, Reichelt-Brushett & Howe 2019). Aiptasia’s rapid growth rate, its tolerance of different growing conditions, and asexual and sexual modes of reproduction have made it a tractable laboratory animal. These traits, as well as its relatedness to scleractinian corals (Figure 1.8), the absence of a calcareous skeleton, and critically, the symbiotic relationship it forms with Symbiodiniaceae, which breaks down under stress in a manner comparable to coral, have seen Aiptasia pallida proposed and widely adopted as a model organism for coral research (Voolstra 2013; Weis et al. 2008).

Figure 1.8: Phylogenetic relationships of selected genera within Cnidaria. Asterisks indicate genera that harbour Symbiodiniaceae symbionts (Sabourault et al. 2009)

Since Aiptasia pallida’s nomination, Family Aiptasiidae has undergone a revision. Six years ago, seven former Aiptasia species were synonymised under a newly erected genus, Exaiptasia, based on their morphological similarity, and renamed Exaiptasia pallida (Grajales & Rodríguez 2014). Although the name, E. pallida, became quickly established in the literature, it was not ratified by the International Commission on Zoological Nomenclature (ICZN 2017). Instead, the ICZN deemed E. diaphana the correct designation according to the Principle of Priority. Therefore, E. diaphana is used here. 18

1.7 Exaiptasia diaphana

1.7.1 Global distribution

E. diaphana is found worldwide in shallow tropical and sub-tropical marine environments. A study of genetic diversity among E. diaphana found evidence for two genetically distinct populations; one global, and the other isolated to the waters around Florida, USA (Thornhill et al. 2013). The authors speculated that, based on the genetic uniformity of the Symbiodiniaceae species shared by the global population, their distribution was likely due to vectored introductions, for example by ship ballast or fouling. A later study supported the description of two E. diaphana networks and added a species from Panama and Brazil to the Exaiptasia genus, E. brasilliensis (Grajales & Rodríguez 2015).

1.7.2 Lifestyle and anatomy

Unlike coral, E. diaphana is solitary, rather than colonial, and whilst primarily sessile it has the ability to move small distances. These are among the traits that set actiniarian sea anemones apart from coral. Like coral polyps however, E. diaphana has a typical anthozoan body plan and tissue structure (Figure 1.9a). It is radially symmetrical, with a central slit-like mouth surrounded by a ring of tentacles, and tissue comprising two layers of cells, the epidermis and gastrodermis, separated by a gelatinous layer, the mesogloea. It also has a SML. The two major types of the cnidocyte stinging cells that are characteristic of cnidarians are present: nematocysts that bear a toxic barb, and sticky spirocysts. Captured prey is directed into the mouth by the tentacles, and passes into the gastrovascular cavity, which is a blind sac (Figure 1.9b). Once there, mesentaries aid absorption of food by increasing surface area. At the base is the pedal disc, which allows attachment to soft or hard surfaces (Fox 2006).

Tentacle

Oral disk

Mouth

Pharynx Retractor muscle Gastrovascular cavity Mesentary

Pedal disk (a) (b)

Figure 1.9: E. diaphana: (a) polyp (Brown & Rodriguez-Lanetty 2015), and (b) cut-away showing anatomical features (Ruppert, Fox & Barnes 2004).

1.7.3 Reproduction

E. diaphana can reproduce sexually via natural female/male gamete fertilisation, or via parthenogenesis, wherein female gametes develop without fertilisation (Bocharova & Kosevich 2011). It was historically considered gonochoric (Schlesinger et al. 2010), however, research has since confirmed the occurrence of hermaphroditism (Armoza-Zvuloni et al. 2014).

E. diaphana does not have true gonads. Instead, gametes accumulate in the mesentarial tissue until breaks in the epithelium allow them to enter the gastrovascular cavity where they can be discharged through the mouth. Fertilisation then occurs externally after male/female broadcast spawning, or internally within the female gastric cavity (‘brooding’) (Bocharova 2016). As shown below (Figure 1.10), unlike coral, Anthozoans such as E. diaphana do not have a medusoid stage, but develop directly from embryo to planula larvae, then polyp.

Figure 1.10: Exaiptasia diaphana’s primary reproductive modes (Grawunder et al. 2015).

Sexual reproduction in E. diaphana is influenced by seasonal conditions. Gametogenesis is initiated by an increase in SST in spring, and spawning follows at the onset of summer in response to summertime diurnal rhythms and SST (Schlesinger et al. 2010). In the laboratory, E. diaphana can be artificially induced to spawn by exposure to elevated temperatures and blue light cues that mimic lunar cycles. This manipulation allows repeated and controlled production of larvae by sexual reproduction, which is advantageous if larvae are required in high numbers for experimental work. The aposymbiotic state of the larvae is also useful where control over the Symbiodiniaceae symbiont type is desired. However, it has been found that larvae produced this way do not proceed to metamorphosis and settlement stages (Grawunder et al. 2015). Therefore, while useful for studying cnidarian reproduction, early life stages, and the establishment of the E. diaphana-Symbiodiniaceae relationship (Wolfowicz et al. 2016), the full potential of this method is yet to be realised.

As shown above, E. diaphana is also capable of asexual reproduction by pedal laceration. Described as “sticky locomotion”, pedal laceration occurs when cells tear from the pedal disc as the polyp moves (Geller, Fitzgerald & King 2005). The ‘lacerates’ contain fragments of the parental pedal disc, mesenteries and body walls. These cells fuse to position the internal and external cells correctly. New tentacles and mesenteries then begin to form along the fused edge (Bocharova 2016). Pedal laceration is the most studied form of asexual reproduction in

E. diaphana, however longitudinal fission is also known to occur, and transverse fission has been recorded within the closely related Aiptasia genus (Bocharova & Kosevich 2011).

In nature, asexual reproduction occurs year round, but environmental conditions can increase frequency (Bocharova 2016). For example, pedal laceration increases with summertime diurnal rhythms and SST (Schlesinger et al. 2010). This mode of reproduction is useful in research as all offspring are clones, which improves experimental reproducibility. Therefore, work has been done to identify the laboratory conditions that promote it. It has been found that low animal density, darkness, feeding with A. salina nauplii, rather than metanauplii, and cultivation at 26 °C is optimal (Leal et al. 2012). The increase in pedal laceration in darkness is unusual as this reduces photosynthesis by the symbiont, and therefore deprives the host of fixed carbon, and hence energy (Hunter 1984). However, it has been hypothesised that prolonged darkness induces anemones to move in an attempt to find light, thereby increasing the occurrence of laceration (Geller, Fitzgerald & King 2005).

1.7.4 Microbiome: Symbiodiniaceae

A survey of 46 global sites showed that E. diaphana associates predominantly with the Symbiodiniaceae species Breviolum minutum, and that a high level of host-symbiont specificity exists. However, two Caribbean samples were found to harbour Symbiodinium linuchae, and one sample from Brazil harboured B. psygmophilum (Grajales, Rodríguez & Thornhill 2016). These findings corroborated those of an earlier study (Thornhill et al. 2013). However, the earlier study also revealed rare incidences of mixed symbioses among samples from Florida; a small number of samples harboured Breviolum and Symbiodinium species simultaneously, whilst one sample was found harbouring Symbiodiniaceae species from Symbiodinium and Cladocopium.

E. diaphana can also harbour the same Symbiodiniaceae genera found in many coral species, and expels them when stressed (Chen et al. 2016; Leal et al. 2015). Therefore, the Symbiodiniaceae-E. diaphana symbiosis has received considerable attention due to interest in coral bleaching mechanisms. Bleaching in E. diaphana is induced by exposure to cold or heat, with thermal limits comparable to coral (Gates, Baghdasarian & Muscatine 1992). In laboratory settings it can also be bleached by exposure to menthol (Matthews et al. 2016) or

prolonged darkness without feeding (Steen & Muscatine 1987). Once bleached, it can be maintained in an aposymbiotic state, or reinfected with different Symbiodiniaceae species (Xiang et al. 2013). This trait makes it especially useful where the influence of Symbiodiniaceae type on growth or bleaching is being investigated.

1.7.5 Microbiome: Bacteria

The bacterial component of E. diaphana’s microbiome has received less attention than Symbiodiniaceae. Two early studies described bacterial aggregates in E. diaphana tissue, but did not identify the bacteria present (McKinstry et al. 1989; Palincsar et al. 1989). Subsequent studies have focussed on characterising the bacterial associates of lab-reared (Herrera et al. 2017; Röthig et al. 2016) and wild E. diaphana (Brown et al. 2017), or tracking changes in the bacterial communities of E. diaphana exposed to thermal stress (Ahmed et al. 2019; Plovie 2010).

According to the characterisation studies, E. diaphana harbours predominantly Proteobacteria at levels comparable to corals (50–70%) (Blackall, Wilson & van Oppen 2015). However, there were differences between the studies regarding the less prevalent taxa, with Bacteroidetes, Firmicutes and Actinobacteria featuring as the second most abundant phylum, depending on sample origin. Although Herrera et al. (2017) claimed that their findings, based on a small number of lab-reared anemones, indicated host-bacteria specificity, evidence from a more comprehensive study, suggested that culture environment was more likely to determine bacterial community composition (Brown et al. 2017). This was supported by Ahmed et al. (2019) who observed increased bacterial diversity in E. diaphana exposed to elevated temperature. Bacterial community richness estimates and the existence of a core microbiome also varied between the studies, but the results appeared to be influenced by sample size and origin. The lack of clarity around E. diaphana’s bacterial microbiome suggests that further research is needed to improve estimates of diversity, specificity and stability.

1.7.6 Model development: ‘omic resources

When E. diaphana was proposed as a coral model, obtaining its full genome was deemed a research priority (Weis et al. 2008), however the transcriptome was published first as “cost 23

effective alternative” (Sunagawa et al. 2009b). This revealed genes putatively involved in bleaching and neutralisation of reactive oxygen species associated with photoinhibition. The genome was eventually published in 2015 and helped reveal suspected horizontal gene transfer between E. diaphana and Symbiodiniaceae or bacteria (Baumgarten et al. 2015), and genetic variation among global E. diaphana populations (Bellis et al. 2018; Bellis, Howe & Denver 2016). Proteomic (Oakley et al. 2016; Oakley et al. 2017) and metabolomic studies have also been performed (Hillyer et al. 2016; Rädecker et al. 2018). As well as providing researchers with a suite of multi-omic data, this information can reveal similarities and differences between E. diaphana and coral to clarify its suitability as a coral model.

1.7.7 Model development: Standardisation

E. diaphana's ability to reproduce asexually has been exploited to generate clonal lines, which have helped standardise the E. diaphana genotypes used in research. The lines most commonly used are summarised below (Table 1.1). Much of the ‘omics data described above, including the microbiome data, was obtained using CC7 and H2 anemones, making these genotypes especially useful to researchers.

Table 1.1: E. diaphana genotypes commonly used in research.

Name Origin Global Homologous Gender Australian References distribution Symbiodiniaceae availability CC7 US South US South Symbiodinium Male Unavailable (Grawunder et al. 2015; Atlantic Atlantic linucheae Sunagawa et al. 2009b) H2 Pacific Ocean, Worldwide Breviolum minutum Female Unavailable (Grawunder et al. 2015; Coconut Island, tropical – Xiang et al. 2013) Hawaii sub-tropical F003 US South US South Symbiodinium Female Unavailable (Grawunder et al. 2015) Atlantic Atlantic linucheae, Breviolum minutum RS Red Sea, Al Lith, Unknown Symbiodinium Unknown Unavailable (Cziesielski et al. 2018) Saudi Arabia microadriaticum

A notable feature of the genotypes shown in Table 1.1 is their current unavailability in Australia. Although this has not prevented Australian researchers using E. diaphana (Howe, Reichelt-Brushett & Clark 2012; Howe et al. 2017), the use of uncharacterised anemones limits experimental reproducibility and comparison with other studies. The absence of the established models in Australia is also remarkable given the importance of Australia’s coral

reefs and uncertain future they face. However, efforts have recently begun to develop an Australian model from Great Barrier Reef-sourced anemones, which will assist Australian coral researchers and expand the E. diaphana model range (Dungan et al. 2020).

1.8 Organisation of the thesis

The main objective of this thesis is to investigate the bacterial associates of E. diaphana and their hypothesised influence on host thermal tolerance. A second objective is to contribute to the development of a tractable Australian coral model by using E. diaphana anemones originally sourced from the GBR and now maintained in culture at The University of Melbourne for this work. The research questions that direct this investigation are:

1. What are the bacterial communities of GBR-origin E. diaphana?

2. How do E. diaphana’s bacterial communities respond to thermal stress?

3. Can probiotic inoculation improve E. diaphana’s thermal tolerance?

In Chapter 2 I address the first research question by characterising the bacterial communities of four GBR-origin E. diaphana genotypes using 16S rRNA gene metabarcoding. The results are compared to data from existing E. diaphana models and anemones from aquaria at the Australian Institute of Marine Science (AIMS) in Townsville, Australia, to provide an insight into bacterial diversity among E. diaphana.

In Chapter 3 I address the second research question by exposing E. diaphana to increasing temperature over two weeks. Changes in the host-associated bacterial communities are assessed using 16S rRNA gene metabarcoding to compare the bacteria of heat stressed anemones to those of anemones maintained at ambient temperature. in Chapter 4 I investigate the feasibility of generating gnotobiotic E. diaphana by exposure to antibiotics. Such cultures could help explain the roles of bacteria in the cnidarian holobiont and would be a valuable addition to the coral research toolbox, for example in work involving manipulation of the bacterial communities to alter host phenotypes. The ability of antibiotic treatment to eliminate the host-associated bacteria is assessed using 16S rRNA gene metabarcoding and digital droplet PCR.

In Chapter 5 I address the third research question by inoculating E. diaphana with bacteria with high free radical scavenging ability, then exposing them to elevated temperature. Incorporation of probiotic bacteria into the E. diaphana holobiont is assessed using 16S rRNA gene metabarcoding. The impact of thermal stress on the treated anemones is assessed by measuring the dark-adapted quantum yield (Fv/Fm) and density of Symbiodiniaceae in the host’s tissue.

In Chapter 6 I summarise the key findings of the thesis and provide suggestions for future research.

Data chapters (Chapters 2–5) have been written as standalone manuscripts and have been submitted or are intended for publication. Therefore, some repetition may occur in the introductions and descriptions of the methods.

Chapter 2: Microbiota characterisation of Exaiptasia diaphana from the Great Barrier Reef

This chapter is inserted without abstract as currently accepted for publication by the journal Animal Microbiome, pending revisions:

Hartman, LM, van Oppen, MJH & Blackall, LL: ‘Microbiota characterisation of Exaiptasia diaphana from the Great Barrier Reef’.

All authors designed the experiment. L. Hartman performed the experimental work, analysed the data, wrote and edited the manuscript. M. van Oppen and L. Blackall reviewed and edited the manuscript.

2.1 Introduction

Coral reefs are reservoirs of enormous biodiversity, are essential for the maintenance of marine and coastal ecosystems (Moberg & Folke 1999), and their economic and social values are vast (Deloitte Access Economics 2017). Alarmingly, the loss of the coral’s energy-producing Symbiodiniaceae endosymbionts, a process known as bleaching, has increased in frequency and severity due to elevated sea surface temperatures (SST) caused by rising atmospheric greenhouse gas concentrations (Oliver, Berkelmans & Eakin 2018). This has led to widespread coral mortality and damage to reef systems (Hughes et al. 2018), and has heightened the need to investigate mitigation solutions. As reefs succumb to the impacts of climate change, the need for a model organism to assist coral research has never been greater. Fortunately, this need has been met in the form of the tropical sea anemone, Exaiptasia diaphana (previously Aiptasia pallida (Grajales & Rodríguez 2014; ICZN 2017)), which was proposed as a coral model organism 12 years ago (Weis et al. 2008). Since then, the traits that made it an attractive candidate, particularly its intracellular symbiosis with members of the same algal family harboured by corals (Symbiodiniaceae) and lost under stress conditions, have seen it widely adopted by the research community. Subsequently, several clonal lines have become established; however, none are available in Australia thus hampering the ability of Australian researchers to perform laboratory-based research on this model. Therefore, a native Australian E. diaphana model is urgently needed.

E. diaphana CC7 and H2 are the clonal lines primarily employed in research elsewhere. CC7 was developed from a single propagule of Atlantic Ocean origin obtained from Carolina Biological Supply (Burlington, North Carolina) (Sunagawa et al. 2009b). H2 was developed from a founder collected at Coconut Island, Hawaii (Xiang et al. 2013). Distinguishing features of each are their gender and algal symbiont; CC7 is male and harbors Symbiodiniaceae of the genus Symbiodinium, whereas H2 is female and harbors Symbiodiniaceae of the genus Breviolum (Grawunder et al. 2015). Baseline multi-omics data describing E. diaphana has been generated to improve our understanding of cnidarian physiology (Baumgarten et al. 2015; Oakley et al. 2016; Rädecker et al. 2018; Sunagawa et al. 2009b). A key element of this work has been characterisation of E. diaphana’s associated bacteria (Brown et al. 2017; Herrera et al. 2017; Röthig et al. 2016), an important component of coral microbiota (Rohwer et al. 2002). We characterized the bacteria of GBR-sourced E. diaphana because improved understanding

of them could assist in coral bleaching mitigation research. For example, manipulation of coral-associated bacteria has been proposed as a mechanism for improving coral’s tolerance to elevated SST (van Oppen et al. 2015), and initial testing has shown some promise (Rosado et al. 2018).

Recently, Röthig et al. (2016) investigated the bacterial associates of lab-reared CC7 by metabarcoding of the V5 to V6 region of the 16S rRNA genes. They compared bacteria associated with symbiotic and aposymbiotic (lacking Symbiodiniaceae) CC7, and reported the phenotypic potential inferred from the taxonomy of all bacterial associates via the tool METAGENassist (Arndt et al. 2012), as well as a core bacterial community. The number of bacterial OTUs per symbiotic anemone (n = 5) ranged from 109 to 133, which was modest compared to some corals (Li et al. 2013; Sunagawa, Woodley & Medina 2010). Almost all OTUs belonged to the phyla Proteobacteria (67%), Actinobacteria (26%), Bacteroidetes (3%) or Firmicutes (2%). A core community i.e., OTUs present in every sample, of 37 OTUs was reported. However, the probability of an OTU being present in all samples was high because the samples originated from the same culture collection, and few animals were compared. The presence or absence of Symbiodiniaceae endosymbionts appeared to drive differences in inferred bacterial phenotype. For example, the bacteria of aposymbiotic anemones showed depletion in sulfur metabolizing capacity, which was attributed to the absence of dimethylsulfoniopropionate (DMSP), which is generally produced by algal symbionts (Raina et al. 2013).

Brown et al. (2017) described E. diaphana’s associated bacteria with samples originating from the North Pacific, Atlantic Ocean and the Caribbean Sea taken from museum collections, laboratory aquaria and pet shops. Metabarcoding of the V1–V4 of the 16S rRNA genes generated 12 585 operational taxonomic units (OTUs) from 49 samples. The samples with the highest bacterial community richness were raised in artificial environments (1 358–1 671 OTUs), whereas those with the fewest were from Hawaii’s coastal waters (409 OTUs). The relative abundance of bacterial phyla varied greatly among the samples but were consistent within environment types. For example, Proteobacteria (>50%) dominated among wild Pacific, Atlantic and Caribbean samples, whereas Firmicutes (~70%) dominated the bacteria of samples from an outdoor laboratory aquarium. A core bacterial community was not found, but six genera (Vibrio, Nautella, Ruegeria, Marinobacter, Lentisphaera and Flavobacterium)

were common. The authors concluded that E. diaphana’s bacterial communities are highly variable and shaped largely by their environment. However, the unusual sample origin and treatment of some (e.g., ethanol preserved museum specimens), disparate rearing conditions, and small sample sizes for each sample type (≤4) may have influenced these findings.

Herrera et al. (2017) analysed the bacterial microbiota of lab-reared H2 (n = 5) and CC7 (n = 5). Both genotypes were inoculated with Breviolum minutum to standardize the algal symbiont harboured by the anemones. Bacterial community richness was lower in H2 than CC7 with 96 versus 118 OTUs, respectively. Proteobacteria was the most abundant phylum in both H2 (53%) and CC7 (70%), which matched the dominance of Proteobacteria in CC7 (67%) previously reported (Röthig et al. 2016). However, the relative abundance of Bacteroidetes (37%) and Actinobacteria (10%) in H2 differed markedly from CC7 (2% versus 26% respectively). Approximately 40% of OTUs were present in all five H2 samples. Once again, the common environment and small sample number increased the probability of identifying a core contingent. The phenotypic potential of each genotype’s bacterial associates was inferred via METAGENassist (Arndt et al. 2012). Due to their different compositions, the inferred phenotypes of H2 and CC7’s bacterial associates differed substantially. The most distinctive differences were in utilisation of chitin, xylan, sugars and propionate, where H2 was depleted and CC7 enriched in each case. The different geographic origin and genders of the genotypes, rather than genotype itself, were offered as possible explanations for the distinct bacterial communities. This partly supported the previous conclusion (Brown et al. 2017) that environment shapes E. diaphana’s bacterial microbiota. However, it also assumed long-term stability of bacterial communities in culture, which has not been tested in E. diaphana.

These three reports (Brown et al. 2017; Herrera et al. 2017; Röthig et al. 2016) provide the first insights into E. diaphana’s bacterial associates, but some aspects, such as richness estimates and the influence of host genotype, remain unclear. This may be due to study limitations, such as small sample sizes and disparate sample handling. Information on E. diaphana sourced from the GBR is absent. Here, we established cultures of GBR-sourced E. diaphana, and characterized their associated bacteria using a 16S rRNA gene metabarcoding approach. We incorporated data from the earlier studies (Brown et al. 2017; Herrera et al. 2017; Röthig et al. 2016) to provide a broad E. diaphana story. We explored genotypic influence on bacterial community composition and bacterial inferred phenotypes.

Our establishment of a GBR-sourced E. diaphana model and characterisation of its bacterial associates will help clarify E. diaphana associated microbiota variability and assist Australian coral researchers who have not had access to this rising star of coral research.

2.2 Materials and methods

2.2.1 E. diaphana culture collection

Anemones used in this study were taken from the E. diaphana culture collection at the University of Melbourne (UoM), Australia. This collection was established with anemones sourced from aquaria in the National Sea Simulator (SeaSim) at the Australian Institute of Marine Science (AIMS), Townsville, Australia, which are stocked with material from the GBR (Figure A 1; Figure A 2). The UoM collection contains four E. diaphana genotypes: AIMS1 (female), AIMS2 (male), AIMS3 (female) and AIMS4 (female) (Dungan et al. 2020). All are maintained at 26 °C under 12–20 µmol photons m–2 s–1 on a 12h:12h light-dark cycle and fed twice-weekly with freshly-hatched Artemia salina (Salt Creek, Premium GSL, USA). Water is 100% changed once weekly with seawater reconstituted from Red Sea Salt™ (R11065, Red Sea, USA) at ~34 parts per thousand.

2.2.2 Sampling and sample processing

Anemones were sampled in November 2017 for bacteria and Symbiodiniaceae characterisation. Six anemones were randomly collected with 3 mL plastic sterile Pasteur pipettes from each of 12 culture tanks comprising three replicate tanks per genotype (Figure A 3). Anemones were placed into sterile 1.5 mL centrifuge tubes, snap frozen in liquid nitrogen and stored at –80 °C until processing. One litre of water was siphoned from each tank and filtered through a 100 µm cell strainer (352360, Sigma Aldrich, Australia) into a sterile filter- unit (NALDS0320-5045, Thermo Fisher, Australia), then through a Pall 47 mm 0.2 µm membrane (66234, Bio-Strategy, Australia). The membranes were individually stored in sterile, covered Petri dishes (153-533, WestLab, Australia) at –20 °C until processing. As the UoM A. salina feedstock was not presumed sterile, 3 x 75 µL of a dense A. salina suspension was sampled to identify associated bacteria.

E. diaphana is difficult to locate in the wild due to its cryptic nature, therefore five E. diaphana polyps were collected from the outflow of a 4 000 L outdoor holding tank containing live corals, snails, sea cucumbers and fish at the AIMS SeaSim (Figure A 4). They were included to estimate the bacterial composition of GBR-sourced E. diaphana maintained in a complex marine environment and thus served as wild animal proxies. A 1L water sample was also collected from the SeaSim holding tank and filtered through a Sterivex filter cartridge (SVGP01050, Sigma Aldrich, Australia), which was stored at –20 °C until processing (Table 2.1).

Table 2.1: Sampling summary for bacterial microbiota analysis. Sample type Number of samples E. diaphana (UoM cultures): 6 per tank per genotype (6 x 3 x 4) 72 Water (UoM cultures): 1 L per tank per genotype (1 x 3 x 4) 12 DNA extraction blanks: tissue x 1; water x 1 2 No-template PCR controls: x 3 3 Artemia salina: x 3 3 Wild proxy E. diaphana (from an AIMS SeaSim aquarium): x 5 5 Water (AIMS SeaSim aquarium): x 1 1 Total 98

Sample DNA was extracted according to (Wilson et al. 2002) but modified by 15 min incubation with 20 mL of 10 mg/mL lysozyme after sample homogenisation, and 20 sec bead beating at 30 Hz (Tissue-Lyser II, Qiagen, Australia) with 100 mg of sterile glass beads (G8772, Sigma Aldrich, Australia). For each water sample, filter membranes were sliced into thin strips with a sterile blade and treated as an animal tissue sample as previously described (Röthig et al. 2016). Blank extractions without sample material were used to test for reagent and plasticware contamination. Extracts were checked for DNA by agarose gel electrophoresis.

Bacterial DNA was amplified by PCR in triplicate using primers with Illumina adapters (not shown) targeting the V5-V6 regions of the 16S rRNA gene: 784F [5ʹ AGGATTAGATACCCTGGTA 3ʹ], 1061R [5ʹ CRRCACGAGCTGACGAC 3ʹ] (Andersson et al. 2008), as previously used (Herrera et al. 2017; Röthig et al. 2016). Three no-template PCR blanks were included to test for reagent and plasticware contamination. To identify the anemone’s intracellular Symbiodiniaceae, DNA from 12 UoM E. diaphana (one from each tank) sampled for the bacterial analysis, and all five

wild proxy anemone samples (Table 2.2), was amplified by PCR in triplicate using primers with Illumina overhangs (not shown) targeting the rDNA internal transcribed spacer region 2 (ITS2): ITS2-Dino-forward [5ʹ GTGAATTGCAGAACTCCGTG 3ʹ] (Pochon et al. 2001), ITS2rev2 [5ʹ CCTCCGCTTACTTATATGCTT 3ʹ] (Stat et al. 2009). Separate PCRs for bacteria and Symbiodiniaceae were performed in 20 µL volumes comprising 1 µL template DNA, 10 µL of 10 µM MyTaq HS Mix polymerase (Bioline), 0.5 µL of 10 µM forward primer, 0.5 µL of 10 µM reverse primer, and 8 µL MilliQ water. Thermal-cycler settings were: 1 cycle at 95.0 °C for 3 min, 30 cycles at 95.0 °C, 55.0 °C and 72.0 °C for 15 sec each, and 1 cycle at 72 °C for 3 min. Triplicate PCR products were pooled and checked by agarose gel electrophoresis. The SeaSim water sample was removed from the analysis as no PCR product was visible on the agarose gels.

Table 2.2: Sampling summary for Symbiodiniaceae analysis. Sample type Number of samples E. diaphana (UoM cultures): 1 per tank (1 x 12) 12 Wild proxy E. diaphana from SeaSim aquarium: x 5 5 Total 17

A volume of 25 µL of PCR product from each sample pool was sent to the Ramaciotti Centre for Genomics (RCG), Sydney, Australia for sequencing on a single Illumina MiSeq V2 2 x 250 run. RCG performed PCR product clean-up and normalisation as part of sequencing library preparation.

2.2.3 Sequencing data workflow

Raw, demultiplexed MiSeq reads were joined in QIIME2 v2018.4.0 (Bolyen et al. 2019). Sequence denoising, chimera checking and trimming was performed in DADA2 (Callahan et al. 2016) to correct sequencing errors, remove primer sequences and low quality bases. Resulting amplicon sequence variants (ASVs) with a single representative sequence were removed. Prokaryote taxonomy was assigned in QIIME2 against a SILVA database (v 132) trained with a naïve Bayes classifier (Bokulich et al. 2018; Pedregosa et al. 2011; Quast et al. 2013; Wang et al. 2007). ASVs identified as eukaryotes, mitochondria, or chloroplasts were removed. 33

Symbiodiniaceae sequences, processed in DADA2 as above, were clustered into OTUs at 99% sequence similarity by closed-reference OTU picking in vsearch (Rognes et al. 2016). A database adapted from a study of Symbiodiniaceae diversity (Arif et al. 2014) was used for taxonomic classification and to seed the OTU clusters. The raw bacterial sequencing reads from the E. diaphana clonal genotypes CC7 and H2 previously reported by Röthig et al. (2016) and Herrera et al. (2017), respectively, were downloaded from NCBI’s Sequence Read Archive and processed as above for comparison.

All analyses described hereafter were performed in R v3.6.0 (R Core Team 2018) with the packages vegan v2.5-6 (Oksanen et al. 2018), phyloseq v1.29.0 (McMurdie & Holmes 2013), microbiome v1.7.2 (Lahti & Shetty 2017), mvabund v4.0.1 (Wang et al. 2012) ggplot2 v3.2.1 (Wickham 2019), DESeq2 v1.23.10 (Love, Huber & Anders 2014) and decontam v1.5.0 (Davis et al. 2018). A significance threshold of α = 0.05 was used for all statistical tests, unless otherwise stated. Tabulated ASV counts, taxonomic assignments and metadata were imported into R, and rarefaction curves were generated to confirm that sequencing captured species diversity. Contaminant ASVs introduced during sample preparation were identified using the bacterial community data from the negative control samples using decontam’s ‘prevalence’ method and default threshold (p = 0.1). Stacked bar-charts of family-level ASVs were generated to assess sample bacterial community compositions.

2.2.4 Diversity analyses

The bacterial metabarcoding data were normalized by subsampling to 12 810 sequences per sample. Bacterial community richness was assessed according to the number of observed ASVs per sample. Simpson (Simpson 1949) and Shannon index values (Shannon & Weaver 1949) were used to describe and compare alpha diversity across the sample types. Differences in Shannon diversity between the sample types were evaluated by one-way analysis of variance (ANOVA) after checking for data normality and homogeneity of variance by Shapiro- Wilk (Shapiro & Wilk 1965) and Levene’s tests (Levene 1960) respectively. Post hoc pair-wise comparisons were performed using Tukey’s HSD (Tukey 1949). Relative proportions of bacterial phyla in all sample types were calculated and tabulated. Heatmaps of the 20 most abundant class-genus bacterial taxa in the AIMS1–4 and wild proxy anemones were generated and the magnitude (log2 fold change, L2FC) and significance of pairwise differences were 34

calculated to investigate whether differences in bacterial composition corresponded to sample type. nMDS ordinations (Hellinger transformation; Bray-Curtis dissimilarity) were generated of the AIMS1–4 samples to assess whether the genotypes’ bacterial communities were distinct. The genotype–bacteria relationship was explored using Generalized Linear Models (GLMs) with the data fitted to a negative binomial distribution and tested across 999 iterations. nMDS ordinations (Hellinger transformation; Bray-Curtis dissimilarity) of each AIMS1–4 genotype were plotted to investigate whether patterns indicative of non-random variation in the anemone-associated bacterial communities occurred within each genotype.

2.2.5 AIMS1–4 core bacterial community member analysis

The AIMS1–4 bacterial communities were surveyed for core members. An ASV was deemed ‘core’ if it was present in every AIMS1–4 sample in accordance with the ‘shared membership’ criteria (Shade et al. 2014). The core bacterial members in AIMS1–4 were also investigated in the water, wild proxy, CC7 and H2 anemones and the A. salina microbiota.

2.2.6 Phenotypic potential analysis

The phenotypic potential of the sample’s bacterial associates was determined by the online tool METAGENassist (Arndt et al. 2012), which maps taxonomy to phenotype using information from the BacMap (Cruz et al. 2012), GOLD (Reddy et al. 2015) and NCBI (Federhen 2012) databases. Data processing followed reported methods (Herrera et al. 2017; Röthig et al. 2016). Briefly, ASV count, taxonomic assignment, and sample type data were imported into METAGENassist. ASVs with identical taxonomic assignments were combined, and the data were filtered using interquartile range filtering to improve resolution and control the false discovery rate (Hackstadt & Hess 2009). The remaining ASVs were normalized by sum for sample-to-sample comparison, and by Pareto scaling for taxon-to-taxon comparison. Data describing the phenotypic capability of each sample type in 15 categories previously assessed (Herrera et al. 2017; Röthig et al. 2016) were exported from METAGENassist as a histogram.

2.3 Results

2.3.1 Bacteria metabarcoding

Sequencing produced 3 601 241 raw reads across AIMS1–4 and wild proxy E. diaphana, UoM water, A. salina feedstock and negative control samples: minimum 20 027; mean 37 126, maximum 60 798 reads per sample. After merging, denoising and chimera filtering, 2 516 454 reads remained: minimum 12 810, mean 25 943, maximum 44 033 reads per sample. A total of 4 052 ASVs were identified in all sample types. Incorporation of the CC7 and H2 data increased the number of ASVs in the dataset to 4 587.

Rarefaction curves for bacterial sequences from all anemone and water samples plateaued, suggesting that sequencing depth was sufficient to capture bacterial species diversity (Figure A 5). Decontam (Davis et al. 2018) identified seven contaminant ASVs (0.06% of bacteria in the AIMS1–4 anemone samples, 0.07% in the AIMS1–4 water samples, and 2.08% in the wild proxies) and all were removed (Table A 1). Stacked bar-charts of bacterial family-level ASVs revealed three samples with over-represented putatively-contaminant bacterial taxa from the Enterobacteriaceae and Vibrionaceae (Figure A 6). These three samples were removed from further analyses, leaving 4 401 ASVs across all sample types and 2 238 ASVs associated with the AIMS1–4 anemones.

The wild proxy anemones contained on average, two and four times as many ASVs as the long- term lab-cultured AIMS1–4, and the CC7 and H2 E. diaphana genotypes, respectively (Table 2.3). Each AIMS1–4 water sample contained ~25% of the number of ASVs identified in the E. diaphana genotype grown in that water. A. salina had comparatively fewer bacterial associates, with only 27 ASVs.

Table 2.3: Number of bacterial ASVs in each sample type. Sample Total ASVs AIMS1 (n = 18)1 967 (205)2 AIMS 2 (n = 18) 860 (208) AIMS 3 (n = 17) 810 (202) AIMS 4 (n = 17) 758 (226) Wild proxy E. diaphana from SeaSIM aquarium (n = 4) 1 507 CC7 (n = 5) 439 H2 (n = 5) 317 Artemia salina (n = 3) 27 1. Sample numbers after samples deemed to have been contaminated were removed. 2. Number of bacterial ASVs in water samples in which the AIMS genotypes were raised are in parentheses.

Although half the ASVs in each AIMS1–4 genotype were unique, they accounted for <5% relative abundance of the communities (Figure 2.1). However, ASVs common to AIMS1–4 averaged 83.54% relative abundance, which suggested high similarity between the AIMS1–4 bacterial communities.

Figure 2.1: ASVs common to the AIMS1–4 genotypes. The relative abundance of ASVs unique to each AIMS1–4 genotype are shown in brackets.

2.3.2 Diversity analyses of anemone bacterial associates

There was high variation in the number of bacterial ASVs observed within each sample type. The wild proxy anemones contained, on average, more than twice as many ASVs as the UoM 37

lab-cultured anemones (Figure 2.2a; 521 versus 202). However, the Simpson values indicated that these two anemone groups had similar bacterial community evenness (Figure 2.2b). The evenness of the H2 anemones was comparatively low, indicating dominance by a small number of ASVs. Collectively, the wild proxy anemones had considerably higher Shannon index values than the other anemones (Figure 2.2c), demonstrating higher bacterial community diversity in wild proxies compared to lab-cultured anemones. The relatively low Shannon index value for H2 reflected its low observed ASV and Simpson index values.

Observed Simpson Shannon 1.00 6

500 0.95 5

400 0.90

4 300 0.85

200 3 0.80

100 0.75 2 H2 H2 H2 WP WP WP CC7 CC7 CC7

(a) AIMS1 AIMS2 AIMS3 AIMS4 (b) AIMS1 AIMS2 AIMS3 AIMS4 (c) AIMS1 AIMS2 AIMS3 AIMS4

Figure 2.2: Alpha diversity of sample types. Alpha diversity was assessed by: (a) average number of observed ASVs — higher values indicate greater richness, (b) Simpson diversity index — higher values indicate greater evenness, and (c) Shannon diversity index — higher values indicate greater overall alpha diversity. WP = wild proxies.

Significant differences existed between sample types according to ANOVA (F(6, 77) = 65.19, p < 0.001). Pairwise testing by Tukey’s HSD, indicated that alpha diversity as described by Shannon index did not differ significantly among the AIMS1–4 genotypes (Table A 2). However, the wild proxy anemones, with their comparatively high bacterial richness and evenness, differed significantly from all other anemone types. H2 differed significantly from the other sample types due to its low bacterial richness and evenness.

Proportions of bacterial phyla among the anemone sample types were compared with the AIMS1–4 samples pooled to provide a general overview of dominant phyla in the sample types (Table 2.4). Eighteen phyla were identified in the AIMS1–4 samples; ~92% of the total bacterial community were members of Proteobacteria (~76%) or Bacteroidetes (~16%). Spirochaetes (2.54%) and Acidobacteria (1.83%) were the third and fifth most abundant bacterial phyla in AIMS1–4, but these phyla were considerably lower in the wild proxy, CC7 or H2 anemones. In contrast, very low levels of Actinobacteria were detected in AIMS1–4 (0.58%) compared to the other anemones. The wild proxy anemones were associated with 24 bacterial phyla. This was higher than in AIMS1–4 (18 phyla), CC7 (18 phyla) or H2 (10 phyla), which was not surprising given the considerably higher richness observed in the wild proxy anemones.

Table 2.4: Phyla identified in the collective AIMS1–4 animals, wild proxies, CC71, H21, and previously reported (Brown et al. 2017). See Table A 3 for individual AIMS1–4 values.

AIMS1–4 Wild proxies CC7 H2 Reported in Brown Phylum (%) (%) (%) (%) et al. (2017) Proteobacteria 76.21 64.59 67.18 52.94 P2 Bacteroidetes 15.58 25.7 2.73 37.11 P

Spirochaetes 2.53 0.17 0.02 P Planctomycetes 1.89 3.13 0.09 0.04 P

Acidobacteria 1.83 1.22 0.01 P Chlamydiae 1.18 0.63 1.18 0.19 Actinobacteria 0.58 3.31 26.27 9.33 P Firmicutes 0.14 0.16 2.28 0.26 P

Calditrichaeota 0.02 0.03

Verrucomicrobia 0.01 0.12 P

Cyanobacteria 0.01 0.08 P

Gemmatimonadetes 0.01 0.35 P

Dependentiae <0.01 0.07 0.10 Patescibacteria <0.01 0.08 0.02 <0.01

WPS-2 <0.01

Elusimicrobia <0.01 0.04

Lentisphaerae <0.01 <0.01 P Fusobacteria <0.01 0.04 0.03 0.01 P

Marinimicrobia (SAR406 clade) 0.03

Deferribacteres 0.01

Tenericutes 0.04 P

Chloroflexi 0.05 P

Kiritimatiellaeota 0.11

Nitrospirae 0.01 0.01 P

Armatimonadetes 0.03 0.05 P

Latescibacteria 0.01

PAUC34f 0.01

Omnitrophicaeota 0.03

Fibrobacteres 0.04 0.02 P

Epsilonbacteraeota 0.01 <0.01 <0.01

1. Values for CC7 and H2 may differ from those previously reported in Herrera et al. (2017) and (Röthig et al. 2016) due to differences in bioinformatic methods. 2. P = detected; empty cell = not detected.

A heatmap of bacterial ASVs at the class level (Figure 2.3) showed differences in the community compositions of the AIMS1–4 and wild proxy anemones. For example, whilst the phylum level data indicated a high relative abundance of Spirochaetes in AIMS1–4 (2.53%) (Table 2.4), the class-level heatmap showed that Spirochaetia occurred almost exclusively in 40

AIMS2 and AIMS4. Analysis by DESeq2 confirmed that the difference in Spirochaetia between AIMS2 and AIMS4 and all other sample types was significant (Table A 4). Further, members of Subgroup 22 (phylum Acidobacteria) occurred in AIMS1 and AIMS3 but were rare in other anemones, and Pla3 bacteria (phylum Planctomycetes) were absent from AIMS1 and wild proxies but present in AIMS2–4. These differences were also significant. At the genus-level (Figure 2.3), Alteromonas was highly abundant in AIMS1–4 but absent in the wild proxy anemones, and Ruegeria was highly abundant in the wild proxy anemones but present only at low levels in AIMS1–4. Differences in Spirochaeta 2 abundance followed the pattern observed for Spirochaetia. These differences were significant. See Appendix 1 for order and family-level heatmaps (Figures A 8a–b) and L2FC data (Tables A 4–5).

Alphaproteobacteria Thalassobius Gammaproteobacteria Marinobacter Bacteroidia Alteromonas Deltaproteobacteria Sedimentitalea Spirochaetia (Oligoflexaceae) uncultured Planctomycetacia Nonlabens Chlamydiae (Rhodobacteraceae) uncultured Holophagae Spirochaeta 2 Abundance Acidimicrobiia Halioglobus + Actinobacteria (Nannocystaceae) uncultured Subgroup 22 (Chitinophagales) uncultured Thermoanaerobaculia Leisingera Pla3 Erythrobacter – Class (20 most abundant) Rhodothermia Ruegeria Genus (20 most abundant) Bacilli (Cyclobacteriaceae) uncultured Phycisphaerae Aestuariibacter BD2−11 terrestrial Labrenzia Subgroup 26 OM27 clade Acidobacteriia Peredibacter Ignavibacteria Methylotenera WP WP AIMS1 AIMS2 AIMS3 AIMS4 AIMS1 AIMS2 AIMS3 AIMS4 (a) Sample type (b) Sample type

Figure 2.3: Heatmaps of the top 20 taxa by relative abundance at (a) class, and (b) genus levels. Although the heatmaps suggested the existence of a genotype-specific bacterial association in AIMS1–4, an nMDS ordination showed little separation by genotype (Figure 2.4).

NMDS2 NMDS3 NMDS3

stress = 0.18 NMDS1 NMDS1 NMDS2 AIMS1 AIMS2 AIMS3 AIMS4

Figure 2.4: Comparison of bacterial communities in Ep1–4 anemone samples by nMDS on Bray-Curtis distances. All projections of 3D nMDS ordination are shown. Ellipses indicate 95% confidence intervals.

The similarity of the AIMS1–4 bacterial communities was also demonstrated in testing with a GLM. Due to a significant interaction between genotype and tank (mvabund, LRT = 3 506, p < 0.001) (Table A 8) these variables were assessed separately. Testing revealed that whilst, overall, there was no significant difference in community composition based on anemone genotype (mvabund, LRT = 0, p = 0.997), there was a significant tank-wise difference between the samples (mvabund, LRT = 5 941, p < 0.001) (Table A 9). This suggested the presence of a tank effect. Accordingly, nMDS ordinations of the data split by genotype showed general tank- wise sample separation (Figure 2.5). The proximity of water and anemone datapoints also indicated a consistent relationship between waterborne and anemone bacteria.

AIMS1 AIMS2 AIMS3 AIMS4 NMDS2 NMDS2 NMDS2 NMDS2

stress = stress = stress = stress = 0.14 0.14 0.17 0.21 (a) NMDS1 (b) NMDS1 (c) NMDS1 (d) NMDS1 tank 1 anemones tank 1 water tank 2 anemones tank 2 water tank 3 anemones tank 3 water

Figure 2.5: Comparison of bacterial communities in water and anemones by nMDS (Bray-Curtis distance) in the (a) AIMS1, (b) AIMS2, (c) AIMS3 and (d) AIMS4 genotypes. Ellipses indicate 95% confidence intervals. Clustering by tank suggests the presence of a tank effect.

2.3.3 Symbiodiniaceae

Sequencing to identify Symbiodiniaceae produced 339 727 raw reads across the 12 representative AIMS1–4 samples and the five wild proxy anemones (minimum 12 097; mean 19 984, maximum 34 839). After merging, denoising and chimera filtering, 320 108 reads remained (minimum 11 092, mean 18 830, maximum 33 461). At a 99% sequence similarity clustering threshold, 307 402 reads formed a single OTU identified as Breviolum minutum (previously Symbiodinium Clade B, sub-clade B1) which was present in all UoM lab-cultured and wild proxy anemones. Two as-yet unnamed Breviolum OTUs (previously Symbiodinium sub-clades B1i and B1L) were also identified, with each containing 697 (2.1%) and 445 (1.3%) reads, respectively. B1i was identified in 16 samples, whereas B1L was identified in one. In a single wild proxy anemone, two OTUs of seven reads each were assigned to the Symbiodiniaceae genera Cladocopium (previously Symbiodinium Clade C) and Durusdinium (previously Symbiodinium Clade D) (LaJeunesse et al. 2018). The remaining 11 550 reads were unassigned.

2.3.4 Anemone core bacteria

Seventeen core ASVs were identified in the AIMS1–4 samples (Table 2.5). All core ASVs were also detected in the water samples, with one Lewinella bacterium being particularly prevalent (20.18%), however none were present in every sample type. None of the AIMS1–4 core ASVs

were identified in the wild proxy anemones, and only two and five AIMS1–4 core ASVs were present in the CC7 and H2 sample types, respectively. Eleven ASVs could not be identified to species level. Alteromonas macleodii was present in high abundance in all but the wild proxy samples and comprised 51.84% of the A. salina bacterial community. One other ASV, identified only to family level, Rhodobacteraceae, was common to AIMS1–4, CC7 and H2. Two Thalassobius core ASVs were particularly abundant in the AIMS1–4 samples, accounting for 7.74% relative abundance in those samples.

Table 2.5: Relative abundance of bacterial ASVs1 present in every AIMS1–4 sample, and co- occurrence in other sample types2.

Phylum – Family; Genus species AIMS AIMS Wild CC7 H2 A. GenBank 1–4 1–4 proxies salina reference water (%) (%) (%) (%) (%) (%) 1 Proteobacteria – Rhodobacteraceae; Thalassobius 5.09 0.31 MH283840

2 (FCB Group) Bacteroidetes – Lewinellaceae; Lewinella 3.98 20.18 NR_158054

3 Proteobacteria – Rhodobacteraceae 3.59 0.78 0.06 0.76 3.80 MH283808

4 Proteobacteria – Alteromonadaceae; Alteromonas macleodii 3.30 1.28 2.44 2.16 51.84 MK100428

5 Proteobacteria – Alteromonadaceae; Marinobacter salsuginis 2.74 0.80 0.62 KP645206

6 Proteobacteria – Rhodobacteraceae; Thalassobius 2.65 0.15 MH086000

7 Proteobacteria – Oligoflexaceae 2.01 0.10 0.06 0.59 FJ425635

8 Proteobacteria – Alteromonadaceae; Marinobacter adhaerens 1.93 4.59 LT935778

9 Proteobacteria – Rhodobacteraceae; Leisingera aquaemixtae 1.93 0.18 0.18 0.38 MH283840

10 Planctomycetes – Rubinisphaeraceae 1.42 0.10 KU623965

11 Proteobacteria – Hyphomonadaceae; Hyphomonas 1.22 2.23 0.66 0.60 CP017718

12 Proteobacteria – Bacteriovoracaceae; Peredibacter 1.14 0.62 0.01 FJ648610

13 Proteobacteria – Pseudohongiellaceae; Pseudohongiella 1.12 3.63 FJ666208

14 Proteobacteria – Rhodobacteraceae; Donghicola eburneus 1.04 1.71 8.73 MF580372

15 Proteobacteria – Microbulbiferaceae; Microbulbifer elongatus 0.75 0.61 KY034403

16 Proteobacteria – Alcanivoracaceae; Alcanivorax 0.45 0.30 0.01 0.01 CP032351

17 Proteobacteria – (uncultured Deltaproteobacteria) 0.25 0.38 0.01 FJ516839

1. ASVs are described to the deepest taxonomic level possible. 2. Empty cell = not detected.

2.3.5 Phenotypic potential analysis

The phenotypic potential of bacteria associated with the AIMS1–4, wild proxy, CC7 and H2 anemones was determined in METAGENassist (Arndt et al. 2012). After processing, 99 metabolism variables were retained and all anemones were compared across 15 metabolism variables previously described (Herrera et al. 2017; Röthig et al. 2016) (Figure 2.6). There was

high variability in inferred phenotypic potential among the anemones, but all scored highly in nitrogen and sulfur metabolism, and dehalogenation. AIMS1–4 bacteria were enriched in iron oxidation capability, and depleted in sugar fermentation, and propionate and atrazine metabolism, compared to other anemones. The phenotypic potential of the AIMS1–4 samples was generally depleted compared to the wild proxy anemones.

Carbon Nitrogen Sulfur Other 60

% 30

Iron oxidizer Nitrite reducer Sulfur oxidizerSulfur reducer hydrocarbons Xylan degrader Sulfate reducerSulfide oxidizer Dehalogenation Sugars fermenter Nitrogen fixation Chitin degradationDegrades aromatic Ammonia oxidizer Atrazine metabolism Propionate metabolism

AIMS1-4 Wild proxies H2 CC7

Figure 2.6: Phenotypic potential inferred from bacterial data by METAGENassist.

2.4 Discussion

2.4.1 Bacterial associates of anemones

In this study of GBR-sourced E. diaphana bacteria, UoM lab-cultured AIMS1–4 anemones were associated with considerably fewer ASVs (789-970) than AIMS wild proxy anemones (1 662) (Table 2.3). The lab-cultured anemones were maintained for several years in semi-sterile sea water (Red Sea Salt prepared with RO water) and fed commercial A. salina with few bacterial associates (27 ASVs), whereas the wild proxies were from an environment containing a large variety of marine animals likely resulting in exposure to diverse bacteria. Therefore, we conclude that the culture environment (UoM lab-maintained versus AIMS aquarium) explains the bacterial community differences. 45

Average observed ASVs (Figure 2.2a), Simpson (Figure 2.2b) and Shannon index values (Figure 2.2c) revealed high uniformity in alpha diversity between the AIMS1–4 genotype bacterial communities, but also highlighted differences between lab-cultured and wild proxy anemones. Tukey’s HSD analyses of Shannon index values found no significant differences in alpha diversity between the AIMS1–4 anemones but did find a significant difference between them and the wild proxies. The high alpha diversity of the wild proxy bacteria compared to the lab-cultured bacteria suggests a reduction in community complexity over time in culture. CC7 and H2 have been cultured for ~10 years and had the least diverse bacterial associates, which supports this hypothesis. The suggestion of a shift towards bacterial simplicity in culture is supported by an earlier study (Brown et al. 2017) where the bacteria of E. diaphana transferred from aquaria containing complex species to a laboratory environment dropped from 884±104 OTUs to 523±209 OTUs after 4 months. Such a reduction in the diversity of anemone-associated bacteria may be due to the simple, stable nature of the culturing system. It could also be that during lab-culturing the bacterial diversity reduced to a ‘minimal microbiome’, or to “the smallest set of microbes and/or microbial functions needed to develop a stable community” (de Vos 2013).

Bacterial communities in the AIMS1–4 tank water were simple compared to the resident anemones, suggesting that conditions (e.g., nutrient levels, pH) compared to those in and around the anemones (e.g., in the SML and gastrovascular cavity) supported lower bacterial diversity. It could also be due to some bacterial members being strict anemone symbionts, or dilution due to the regular full water changes, which meant bacterial seeding to the water was only from the anemones and the air. Since the SeaSim water sample generated no PCR product, we could not compare the bacteria of the wild proxy anemones and their environment.

Thirteen of the 18 bacterial phyla identified in AIMS1–4 accounted for <1% of the bacterial community (Table 2.4), highlighting the dominance of five phyla, particularly Proteobacteria (76.21%) and Bacteroidetes (15.58%). These two phyla also dominated the wild proxy anemone bacterial associates and are common in corals (Huggett & Apprill 2019). The third most prevalent AIMS1–4 associated bacterial phylum (Spirochaetes; 2.54%), was unusual in its high relative abundance compared to the other anemones. However, this was only found in AIMS2 and AIMS4 whose culture histories differed from AIMS1 and AIMS3 (refer Appendix

1). Spirochaetes has been reported in high abundance in some coral species (van de Water et al. 2016), but their role was not elucidated. nMDS ordinations and a GLM analysis indicated that, overall, the microbiota of the AIMS1–4 genotypes were not significantly different. Although patterns of genotype-related community composition seemed apparent, some, such as the predominance of Spirochaetes in AIMS2 and AIMS4 and absence of Pla3 from AIMS1, matched the genotypes’ collection and culture histories. It is possible however, that genotype-driven differences did exist but were obscured by variability caused by a tank effect, which could be addressed in future studies by increasing biological replication.

The influence of species versus environment in shaping coral-associated microbiota has been assessed in more than 60 studies, with an almost even split between findings of either species- specific or spatio-temporal-driven association (Hester et al. 2016). The data from the present study suggest that E. diaphana and its bacterial members fall into the second category. However, the aforementioned binary assignment oversimplifies the nature of cnidarian bacterial communities, which are renowned for their complexity (van Oppen & Blackall 2019). For example, coral microbiota vary depending on host compartment (e.g. gastrovascular cavity versus SML) and life-stage (Sharp, Distel & Paul 2012; Sweet, Croquer & Bythell 2011). Subsequently, studies that sample E. diaphana of different ages and from different compartments are needed to clarify the host-bacterial relationship.

In AIMS1–4, 99.6% of ITS2 reads were taxonomically assigned to the Symbiodiniaceae Breviolum minutum, which was detected in all samples. This high level of host-symbiont specificity was consistent with previous findings that Pacific Ocean E. diaphana associate exclusively with B. minutum (Thornhill et al. 2013). The detection of OTUs identified as Breviolum types B1i and B1L, and the Symbiodiniaceae Cladocopium and Durusdinium in the present work may suggest that GBR E. diaphana live in symbioses with a mix of Symbiodiniaceae types. However, the Cladocopium and Durusdinium OTUs contained only seven reads each and were found in a single wild proxy anemone. Therefore, they may have been planktonic Symbiodiniaceae that were sampling bycatch.

2.4.2 E. diaphana core bacteria

Although the AIMS1–4 genotypes shared 226 bacterial ASVs (Table 2.5), only 17 met the criteria for core status of presence in every sample. This was fewer than the shared ASVs reported for CC7 (53) or H2 (37). However, the small number of samples in the CC7 and H2 studies increased the probability of ASVs being common to all samples. The absence of the AIMS1–4 core ASVs in the wild proxy anemones illustrated again the difference between bacterial associates in lab-cultured E. diaphana versus those from a more complex environment. The presence of all core ASVs in the AIMS1–4 water samples likely indicates that the culture environment favoured their growth and hence their ubiquitous association with E. diaphana. For example, Lewinella, which was highly abundant (20.18%) in the water, is a ‘defining member of biofilms’ on plastics in the marine environment (Oberbeckmann, Osborn & Duhaime 2016) so may have found the culture tank walls an ideal growth substrate (Zettler, Mincer & Amaral-Zettler 2013).

Core bacteria may include those detectable and important to holobiont function (Busby et al. 2017). They may also include ‘conditionally rare taxa’ that are present at levels below detection that proliferate under favourable conditions (Shade et al. 2014), or taxa that are introduced and become established. The presence of Alteromonas macleodii in AIMS1–4 (3.30%), CC7 (2.43%) and H2 (2.16%), but absence from the wild proxy anemones, suggests proliferation or introduction, rather than functional importance. For example, A. macleodii is widely distributed in the marine environment and is an r-strategist, i.e., it opportunistically blooms when nutrients are in high concentration (López-Pérez et al. 2012). Therefore, its abundance and ubiquity could be due to high nutrient levels in the culturing systems that allowed it to proliferate. The A. salina feedstock may have increased nutrient levels and introduced A. macleodii as it was regularly added and was >50% of its associated bacteria. This may also apply to Marinobacter adhaerens (1.93%), M. salsuginis (2.74%) and Donghicola eburneus (1.08%) which were also A. salina bacterial associates. Although D. eburneus has been found at a similar level (0.8%) in the sea anemone Nematostella vectensis (Har et al. 2015), we cannot assume that it or the other two named taxa are important players in the E. diaphana holobiont due to their correspondence with the A. salina bacteria.

A. macleodii was one of two AIMS1–4 core ASVs that were also detected in CC7 and H2 (Table 2.5). The other ASV common to AIMS1–4, CC7 and H2 could only be identified as a 48

Rhodobacteraceae, which are common marine bacteria frequently involved in biogeochemical cycling (Simon et al. 2017). They have also been detected in high abundance in diseased coral (Kellogg et al. 2013) and are thought to opportunistically occupy and proliferate in health- compromised corals (Sunagawa et al. 2009a). The latter is not the case in our lab-cultured E. diaphana as the animals were healthy.

There is little information on Leisingera aquaemixtae (Table 2.5) and none that links it to cnidarians. It is present in all AIMS1–4 anemones at 1.93% relative abundance and thus could be a promising target in future analyses exploring its function. Two core Thalassobius ASVs comprised 7.74% of the collective AIMS1–4 bacterial communities. However, they were not detected in the other anemones and thus did not match the Thalassobius ASV reported as a core member in CC7 (1.79%) and H2 (2.26%) (Herrera et al. 2017; Röthig et al. 2016). Thalassobius have been identified in many coral studies, particularly studies on coral bleaching (Koren & Rosenberg 2008) and diseases (Sekar et al. 2006; Séré 2014), but no specific role, pathogenic or otherwise, was suggested. However, according to genomic analysis some strains degrade dimethylsulfoniopropionate (DMSP) (Pujalte et al. 2018), which is produced by Symbiodiniaceae (Raina et al. 2017) and coral (Raina et al. 2013), but not E. diaphana (van Alstyne, Dominique & Muller-Parker 2009). Therefore, the core Thalassobius ASVs may be involved in sulfur cycling. AIMS1–4 core member Microbulbifer elongatus (0.75%) is an alginate-, carrageenan- and agar-degrading marine bacterium (Khambhaty, Mody & Jha 2007). Microbulbifer species have been reported in the SML of coral (Sharon & Rosenberg 2008), therefore M. elongatus may reside in, and use polysaccharides in the SML of E. diaphana as an energy source. Studies to localize bacteria within corals’ SML, tissue and skeleton have been conducted (Apprill, Weber & Santoro 2016; Li et al. 2014; Sweet, Croquer & Bythell 2011), and a similar analysis of E. diaphana could clarify the location of M. elongatus.

2.4.3 Phenotypic potential of anemone-associated microbiota corresponds to culture environment

An analysis of the bacterial phenotypic potential with METAGENassist (Arndt et al. 2012) found that the AIMS1–4 bacterial associates were generally depleted compared to the wild proxy anemones (Figure 2.6), which may be due to their lower bacterial diversity. However, chitin degradation, which might provide carbon for metabolism, was marginally higher in 49

AIMS1–4 than their wild proxies. Carbon is acquired by coral primarily from its intracellular algal symbiont as excess photosynthate or through heterotrophy (Houlbrèque & Ferrier-Pagès 2009) but resident bacteria are also important in carbon cycling (Bourne & Webster 2013), including through chitin-degradation (Ducklow & Mitchell 1979). Therefore, bacteria may cycle carbon in the E. diaphana holobiont by metabolising chitin, such as from A. salina exoskeletons, as a food source (Beier & Bertilsson 2013). In contrast, xylan-degradation was somewhat higher in the wild proxy anemones than AIMS1–4. Xylan is a cell wall component in many green algae (Domozych 2016), including those found in coral reef systems (Haas & Wild 2010). Thus, xylan-degrading bacterial associates of AIMS1–4 may degrade green algae that are pest species in the culturing system.

Research has suggested that corals can assimilate nitrogen directly through uptake and processing of dissolved ammonium (Pernice et al. 2012). However, since most reefs exist in nitrogen-limited ecosystems, the contribution of nitrogen-fixing bacteria to coral holobiont function is critical (Lema, Willis & Bourne 2012; Rädecker et al. 2015). Therefore, it was not surprising that high nitrogen-processing potential was reported for the bacterial communities of all sample types, or that known coral-associated nitrogen-processing bacteria were identified in them, including Cyanobacteria (Lesser et al. 2004) and Rhizobiales (Hester et al. 2016) (nitrogen fixation), and Planctomycetes (Zhang et al. 2015).

Sulfur cycling is another important service provided to the host by coral-associated bacteria, and sulfate reducing potential registered highly in bacteria from all anemones. DMSP production by Symbiodiniaceae and processing by bacteria is known to be central to holobiont sulfur cycling and host acquisition (Raina et al. 2017; Raina et al. 2009), and bacteria of the genera Roseobacter, Vibrio, and Alteromonas are capable of degrading DMSP to make sulfur available (Raina et al. 2009). Taxa from each genus were present in the AIMS1–4 and wild proxy anemones, particularly Alteromonas, which was the third most abundant genus in AIMS1–4 (collectively 5.4%) (Figure 2.3b) due mainly to high levels of A. macleodii (Table 2.5).

Iron oxidation was the only category in which AIMS1–4 possessed higher phenotypic potential than the other anemones. Iron-oxidizing bacteria generally belong to the phylum Proteobacteria, and class Zetaproteobacteria (Makita 2018). Although 76.21% of taxa in AIMS1–4 were members of Proteobacteria, none were Zetaproteobacteria. However, some Gammaproteobacteria of the genus Marinobacter also oxidize iron (Edwards et al. 2003). This 50

was one of the most abundant genera in AIMS1–4 with 6.38% relative abundance compared, for example, to the wild proxy anemones with only 0.03% (data not presented in tables), which may explain the relatively high iron-oxidizing potential for the AIMS1–4 bacterial associates.

Atrazine metabolism by the AIMS1–4 and wild proxy anemones was an interesting feature reported by METAGENassist. Atrazine is an herbicide used extensively in the Queensland sugar cane industry that finds its way into the GBR via terrestrial run-off where it poses a risk to coral through its impact on Symbiodiniaceae (Davis et al. 2012; Jones et al. 2003; Lewis et al. 2009). Therefore, the ability of resident bacteria to degrade atrazine would be highly beneficial for anemones and corals on the GBR.

2.5 Conclusion

The GBR-sourced, lab-cultured E. diaphana in this study were generally consistent with previous model and wild proxy E. diaphana in terms of dominant bacterial associates, and resident Symbiodiniaceae. Bacterial richness was similar to other model E. diaphana samples but lower than wild proxy anemones, suggesting a loss of bacterial diversity in culture. The impact of this on E. diaphana health is unknown but could reduce holobiont phenotypic capability. Whilst there were differences in the bacterial associates hosted by different anemone genotypes (AIMS1–4), community-level differences were not statistically significant although these may have been obscured by tank-tank variation among the samples. Nevertheless, compositional differences provided an indication of E. diaphana bacterial community shifts due to environment, and for membership flexibility, which was further evidenced by the small core bacterial composition. By establishing GBR-sourced E. diaphana in lab-culture and producing baseline bacterial associate data we have laid the foundation for future laboratory-based research with this model organism in Australia and elsewhere, particularly if resident bacterial communities and their influence on holobiont function and resilience is of interest.

Chapter 3: The effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana

This chapter is inserted without abstract as published in the journal Microorganisms:

Hartman, LM, van Oppen, MJH & Blackall, LL 2019: 'The effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana', Microorganisms, vol. 8, no. 1, 20.

All authors designed the experiment. L. Hartman performed the experimental work, analysed the data, wrote and edited the manuscript. M. van Oppen and L. Blackall reviewed and edited the manuscript.

3.1 Introduction

One of the most notable manifestations of human-induced climate change has been an increase in sea surface temperature (SST) (Abram et al. 2016). The effect of elevated temperature on reef-building corals that live close to their thermal limit has been catastrophic, causing mass bleaching events worldwide (Coles, Jokiel & Lewis 1976; Hughes et al. 2017b). Bleaching, wherein corals lose the intracellular algae (Symbiodiniaceae) that provide the majority of their nutrition through photosynthesis, typically leads to starvation and death of the host animal (Tremblay et al. 2012a). The subsequent loss of coral cover rapidly converts previously productive reef systems into patchy remnants or marine deserts, particularly when bleaching has occurred on a mass scale (Ostrander et al. 2000).

The coral holobiont comprises the host and its intracellular Symbiodiniaceae, as well as prokaryotes, viruses and fungi (Rohwer et al. 2002), which all contribute to host health and resilience, for example through nutrient provisioning and pathogen protection (Krediet et al. 2013). Therefore, investigating how corals’ bacterial communities change in response to elevated SST can help us understand the role they play in host survival (Sharp & Ritchie 2012).

Studies of coral-associated bacteria have found that communities can be highly dynamic and may change with seasonal temperature shifts (Cai et al. 2018; Sharp et al. 2017). Changes seemingly predictive of heat stress survival have also been recorded (Ziegler et al. 2017). However, understanding the influence of bacteria on coral bleaching has been difficult as all combinations of bacterial community stability or change, and bleaching resistance or susceptibility during thermal stress have been observed (Gajigan, Diaz & Conaco 2017; Lee et al. 2016; Tracy et al. 2015). This highlights the need for controlled, laboratory-based experiments to clarify the relationship between temperature-related bacterial community shifts and bleaching in cnidarians (Bourne et al. 2009; Mouchka, Hewson & Harvell 2010; Thompson et al. 2015).

The sea anemone, Exaiptasia diaphana, is a much-used model for coral symbiosis studies (Voolstra 2013; Weis et al. 2008). Its ability to propagate asexually for rapid growth of clonal populations, basic maintenance requirements and coral-like bleaching response to environmental stressors have seen it widely adopted by the research community and several clonal lines of different geographic origin and algal symbiont type have been established.

However, studies using E. diaphana to explore cnidarian responses to heat stress have focused largely on aspects of the host-Symbiodiniaceae relationship (Bieri et al. 2016; Gates, Baghdasarian & Muscatine 1992; Gegner et al. 2017; Hillyer et al. 2016; Núñez-Pons, Bertocci & Baghdasarian 2017; Tolleter et al. 2013), whilst its bacterial microbiome has been almost wholly neglected and data from only two studies are available.

In a 2010 Master’s thesis, E. diaphana of unspecified origin were exposed to temperature ramped from 26 °C to 31 °C over ten days, then held at 31 °C for four days (Plovie 2010). No significant differences between the associated bacterial communities of control and treated anemones were detected across the study period. However, poor resolution of chosen molecular biology methods negatively impacted the findings. In addition, exposing the anemones to a maximum of 31 °C meant the anemones may not have been thermally stressed, and no measurements of algal cell density or photosynthetic performance were taken to assess their condition. Although the study revealed few insights, it is acknowledged as the first investigation of heat-related changes in the E. diaphana microbiome.

More recently, differences between the bacterial associates of three E. diaphana clonal lines maintained at 32 °C for >2 years or grown at 25 °C were reported (Ahmed et al. 2019). Four E. diaphana-Symbiodiniaceae pairings were analysed: CC7 of Atlantic Ocean origin harbouring Symbiodinium linucheae or Breviolum minutum, H2 of Pacific Ocean origin harbouring Breviolum minutum, and RS of Red Sea origin harbouring Symbiodinium microadriaticum. Significant differences in some alpha diversity indices suggested higher bacterial community richness and lower evenness in the anemones grown at 32 °C and referred to by the authors as “heat stressed”. Notably, bacterial beta diversity and variability was higher in the heat- exposed anemones, apparently providing an example of the Anna Karenina Principle wherein microbiomes from healthy host species are similar, but those from unhealthy hosts are dysbiotic in their own way (Zaneveld, McMinds & Vega Thurber 2017). However, no evidence of thermal stress was provided and long-term maintenance of the anemones at 32 °C suggests they had become acclimated, particularly as the authors described them as having “reached a final stable state”. Therefore, the bacterial associates of the anemones at 32 °C likely exhibited an adjustment, albeit an inconsistent one, of the host rather than a response to temperature increase as would occur during a natural summer heatwave. This suggests that the response of E. diaphana’s bacterial associates to environmentally relevant heat stress

remains unexplored. The present study addressed this knowledge gap by investigating bacterial changes in E. diaphana, originally sourced from Australia’s Great Barrier Reef (GBR), under thermal stress conditions comparable to those found in nature.

3.2 Materials and methods

3.2.1 Experimental conditions and sample processing

Clonal GBR origin E. diaphana (genotype = AIMS2; n = 144) harbouring their natural symbiont, B. minutum, were randomly selected from a single tank in The University of Melbourne (UoM) culture collection. Genotyping and symbiont identification for these cultures has been previously described (Dungan et al. 2020). The selected anemones were individually relocated into single wells within sterile 12-well plates (Costar 3513, Corning, USA), and placed in two Hi-Point 740 incubators, each with lighting intensity and spectra matching the culture collection. The anemones were then acclimated for 10 days at 26 °C in autoclaved seawater reconstituted from Red Sea Salt™ (hereafter, ‘RSS-water’) at a salinity of 34 parts per thousand and fed freshly hatched Artemia salina nauplii once during the acclimation period. The RSS- water was changed every two days. Lighting throughout the experiment was 31.8–33.8 µmol m−2 s−1 on a 12 h:12 h light–dark cycle. Following the 10 day acclimation period, 72 (heat- treated) anemones were exposed to temperature increasing from 26 °C to 33 °C over 14 days with a programmed increase of 0.5 °C per day, and 72 (control) anemones were maintained at 26 °C (Figure 3.1). Throughout the experiment, each incubator’s temperature was monitored by internal sensors, an independent electronic temperature probe and data logger (Saveris T3D, Testo, Germany), and glass thermometers (Initial, Brannan, England) placed in water-filled 500 mL Schott bottles. The anemones were not fed after acclimation to minimize the introduction of bacteria that could have contributed to their bacterial compositions. Starvation was considered reasonable as it is a common practice in E. diaphana studies (Bieri et al. 2016; Hillyer et al. 2016; Marty-Rivera, Yudowski & Roberson 2018; Röthig et al. 2016; Zaragoza et al. 2014).

(a) Control anemones 26°C = sampling point Day 0 2 4 6 8 10 12 14

6 × 16S : 6 6 6 6 6 6 6 6 Sym. : 3 3 3 3 3 3 3 3

Fv/Fm : all all all all all all all all

(b) Heat-treated 33°C anemones 26°C = sampling point Day 0 2 4 6 8 10 12 14 26°C 27°C 28°C 29°C 30°C 31°C 32°C 33°C

6 × 16S : 6 6 6 6 6 6 6 6 Sym. : 3 3 3 3 3 3 3 3

Fv/Fm : all all all all all all all all

Figure 3.1: Sampling schedule for the (a) control anemones, and (b) heat-treated anemones. The number of anemones sampled at each timepoint is listed for metabarcoding of the 16S rRNA genes (16S), Symbiodiniaceae cell counts (Sym), and iPAM measurements (Fv/Fm).

On sampling days, the anemones were removed from the incubators 30–60 min after the daylight cycle and dark-adapted for 10 min before maximum quantum yield (Fv/Fm) was measured with an imaging pulse amplitude modulation (iPAM) fluorometer (IMAGING-PAM M-Series, Heinz Walz, Germany). This allowed non-invasive monitoring of holobiont health by assessing damage to photosystem II (PSII) in the photosynthetic apparatus of B. minutum (Warner, Fitt & Schmidt 1999). iPAM settings for all samples were: measuring light intensity

2, frequency 1; gain 2; damping 2. After each Fv/Fm measurement, nine randomly selected control and heat-treated anemones were snap frozen in liquid nitrogen, three of each for Symbiodiniaceae density analysis to determine the extent of bleaching, and six of each for bacterial community analysis. All samples were stored at −80 °C until processing.

Anemones taken for Symbiodiniaceae density analysis were homogenized then centrifuged, and an aliquot of supernatant was removed for total protein measurement by Bradford Assay (Bradford 1976). The pellet was washed twice and resuspended in filtered RSS-water. The Symbiodiniaceae cell density in the suspension was measured in triplicate on an automated 56

cell counter (Life Technologies Countess II FL, Thermo Fisher, Australia), and values were normalized to total protein to account for anemone size differences. Sample DNA for bacterial analysis was extracted from the anemones following a protocol previously described (Wilson et al. 2002) but modified with the inclusion of 15 min incubation with 20 mL of 10 mg/mL lysozyme after sample homogenization, and 20 s bead beating at 30 Hz (Tissue-Lyser II, Qiagen, Chadstone, Australia) with 100 mg of sterile glass beads (G8772, Sigma Aldrich, Australia). In preparation for DNA sequencing, sample DNA was amplified by PCR using primers with Illumina adapters (underlined) targeting the V5–V6 regions of the 16S rRNA gene 784F [5ʹ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAGGATTAGATACCCTGGTA 3ʹ]; 1061R [5ʹ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCRRCACGAGCTGACGAC 3ʹ] (Andersson et al. 2008). Triplicate PCRs were performed in 20 µL volumes of 1 µL template DNA, 10 µL MyTaq HS Mix polymerase (Bioline, Australia), 0.5 µL of 10 µM 784F, 0.5 µL of 10 µM 1061R, and 8 µL MilliQ water. Thermal-cycler settings were: 1 cycle at 95.0 °C for 3 min, 30 cycles at 95.0 °C, 55.0 °C and 72.0 °C for 15 s each, and 1 cycle at 72 °C for 3 min. Each triplicate was pooled, then the product checked by 1% agarose gel electrophoresis. To identify contaminants introduced during sample preparation, blank DNA extractions and no-template PCRs were included as negative controls.

A volume of 25 µL of pooled PCR product from each sample, and three 25 µL aliquots of a 16 member mock community (Table A 10), which was included to assess sample-sample sequencing consistency, were sent to the Ramaciotti Centre for Genomics (RCG), Sydney, Australia for sequencing on a single Illumina MiSeq v2 (2 × 250 bp) run. RCG performed PCR product clean-up and normalization as part of library preparation prior to sequencing. The resulting Illumina MiSeq data were deposited in the NCBI Sequence Read Archive under accession number PRJNA576764.

3.2.2 Sequencing data workflow

Raw, demultiplexed MiSeq reads were joined in QIIME2 v2018.4.0 (Bolyen et al. 2019). Denoising, chimera filtering, and trimming was performed in DADA2 (Callahan et al. 2016) to correct sequencing errors, remove primer sequences, and low quality bases. Amplicon sequence variants (ASVs) with one representative sequence were removed. Taxonomy was assigned in QIIME2 against a SILVA database (v 132) trained with a naïve Bayes classifier 57

(Bokulich et al. 2018; Pedregosa et al. 2011; Quast et al. 2013; Wang et al. 2007). ASVs identified as eukaryotes, mitochondria, or chloroplasts were removed.

All subsequent data analyses were performed in R v3.6.0 (R Core Team 2018) with differences considered significant at α = 0.05. Tabulated ASV counts, taxonomic assignments and metadata were imported into R and converted into a phyloseq object for ongoing analyses (McMurdie & Holmes 2013). Rarefaction curves were generated in vegan (Oksanen et al. 2018) to assess whether the samples had been sequenced sufficiently to capture species diversity. Potential contaminants were identified using the ‘prevalence’ method in decontam (Davis et al. 2018) with the default threshold of p = 0.1. ASVs not present in every mock community sample were deemed contaminants and were removed from those samples. After removal of putative contaminants, stacked bar-charts describing the number of reads taxonomically assigned to class in the anemone samples, and species in the mock communities, were plotted with ggplot2 (Wickham 2019) to inspect bacterial community compositions.

3.2.3 Physiological and microbiome diversity data analyses

Average Fv/Fm values and algal cell densities for the control and heat-treated samples were plotted over time. Overall changes were evaluated by repeated measures ANOVA after checking assumptions of normality by Shapiro-Wilk (Shapiro & Wilk 1965), and homogeneity of variance by Levene’s tests (Levene 1960) with the R package car (Fox, Weisberg & Price 2019). If significant differences between control and heat-treated data or a group-by-time interaction was detected, paired t-tests for each timepoint were performed to determine when differences occurred. Paired, or pairwise Student’s t-tests (Student 1908) with Benjamini-Hochberg correction (Benjamini & Hochberg 1995), were also performed on the control and heat-treated data to identify when values differed significantly within each sample category. If data deviated from normality and/or homogeneity of variance, Kruskal-Wallis (Kruskal & Wallis 1952), Welch’s (Welch 1947) or Mann-Whitney U tests (Whitney 1997) were performed.

Alpha diversity metrics for the anemone-associated bacterial communities were plotted against time after normalizing the count data by sub-sampling to 12 649 reads per sample.

The number of observed ASVs was used as a measure of richness. Simpson index was used as a measure of community evenness (Simpson 1949). Shannon index was calculated for assessment of overall alpha diversity (Shannon & Weaver 1949). Differences between the control and heat-treated anemones, and across time within each category, were assessed for statistical significance as described above. The number of co-occurring ASVs within control and heat-treated anemones between Day 0 and 14 of the study period were visualized in Venn diagrams (Warnes et al. 2016) to assess bacterial community member transience.

To visualize changes in beta diversity of bacteria in the control and heat-treated anemones across time, an nMDS ordination was generated of all samples from Bray-Curtis dissimilarities of Hellinger transformed data. The ordination was separated by day to assist visual interpretation. Overall differences between the bacterial community compositions of the control and heat-treated anemones were assessed using Generalized Linear Models (GLM) of ASVs in the R package mvabund (Wang et al. 2012). Analysis was performed on ASVs collapsed to genus, and the explanatory variables ‘treatment’ (i.e., control or heat-treated) and ‘time’. A negative binomial distribution was confirmed as appropriate for the data by visualization of the model residuals. Likelihood ratio tests (LRT) were used to determine the deviance (i.e., goodness of fit) of the competing models across 999 sampling iterations. As a treatment-by- time interaction was detected (Table A 11), separate analyses were performed for each timepoint at the ASV level against ‘treatment’ to determine when significant differences between the bacterial communities of the control and heat-treated anemones occurred.

3.2.4 Analysis of changes in abundance of selected bacterial taxa

Line plots of the six most abundant bacterial classes for the control and heat-treated anemones were generated to assess changes in community composition across the study period at a high taxonomic level. Bacteria of the genus Vibrio frequently cause disease in corals and E. diaphana at temperatures above 27 °C due to the upregulation of virulence factors (Kimes et al. 2012; Zaragoza et al. 2014). Therefore, changes in the relative abundance of Vibrio ASVs were investigated to assess their prevalence, and hence disease-causing potential in GBR E. diaphana at elevated temperature. For the bacterial class-level and Vibrio analyses, significant differences between the control and heat-treated anemones, and across time within each category, were assessed as described above. 59

3.2.5 Indicator species identification

Individual taxa that differed significantly both in the heat-treated samples between Day 0 and Day 14, and between the control and heat-treated anemones at Day 14, were identified in an IndVal (Indicator Value) analysis implemented in the R package labdsv (Roberts 2016). IndVal is recommended for discovering potential bacterial biomarkers in the microbiomes of corals subjected to environmental stressors (Glasl, Webster & Bourne 2017), and has been used in previous coral research (Astudillo-García et al. 2017; Glasl, Herndl & Frade 2016; Li et al. 2014). It combines specificity, defined as the mean abundance of a species within a sample type, and fidelity, defined as the relative frequency of occurrence of that species within sample types, to calculate the probability that the species discriminates between samples (Dufrêne & Legendre 1997). Relative abundances of IndVal-nominated taxa were plotted across the sampling timepoints to assess change.

3.3 Results

3.3.1 Sequencing data and bacterial community characteristics

Sequencing produced 4 543 989 raw reads across the 96 microbiome and three mock community samples: minimum 23 447; mean 45 899, maximum 70 071 reads per sample. After merging, denoising and chimera filtering, 2 972 541 reads remained (minimum 12 649, mean 30 026, maximum 46 995 reads per sample) and 4 313 ASVs were identified. Rarefaction curves for all samples plateaued, indicating that sequencing captured bacterial diversity (Figure A 9). Decontam removed eleven ASVs deemed contaminants, which constituted ~0.02% relative abundance of the bacterial communities in the anemone samples (Table A 12). Compositions of the three replicate mock community samples were almost identical, indicating high sample–sample sequencing consistency (Figure A 10). Stacked bar-charts of taxonomic classes detected in the anemones showed moderate variation in relative abundance in the samples with dominance by Alphaproteobacteria or Gammaproteobacteria taxa in most (Figure 3.2). Half of the 40 classes detected contained <0.02% relative abundance each.

100

50 Reads assigned to Class (%)

0 tt01 tt02 tt03 tt04 tt05 tt06 tt21 tt22 tt23 tt24 tt25 tt26 tt41 tt42 tt43 tt44 tt45 tt46 tt61 tt62 tt63 tt64 tt65 tt66 tt81 tt82 tt83 tt84 tt85 tt86 tc01 tc02 tc03 tc04 tc05 tc06 tc21 tc22 tc23 tc24 tc25 tc26 tc41 tc42 tc43 tc44 tc45 tc46 tc61 tc62 tc63 tc64 tc65 tc66 tc81 tc82 tc83 tc84 tc85 tc86 tt101 tt102 tt103 tt104 tt105 tt106 tt121 tt122 tt123 tt124 tt125 tt126 tt141 tt142 tt143 tt144 tt145 tt146 tc101 tc102 tc103 tc104 tc105 tc106 tc121 tc122 tc123 tc124 tc125 tc126 tc141 tc142 tc143 tc144 tc145 tc146 Day 0 2 4 6 8 10 12 14 Day 0 2 4 6 8 10 12 14 26 °C 26 °C 27 °C 28 °C 29 °C 30 °C 31 °C 32 °C 33 °C Control samples Heat-treated samples

Class (total relative abundance)

Alphaproteobacteria (44.95%) Calditrichia (0.41%) Clostridia (0.01%) Babeliae (<0.01) Gammaproteobacteria (21.12%) Bacilli (0.31%) Oxyphotobacteria (0.01%) Subgroup 26 (<0.01) Bacteroidia (12.14%) Leptospirae (0.09%) Lentisphaeria (0.01%) 028H05−P−BN−P5 (<0.01) Deltaproteobacteria (9.98%) Phycisphaerae (0.08%) Thermoleophilia (<0.01) Fimbriimonadia (<0.01) Spirochaetia (4.11%) Rhodothermia (0.08%) Gracilibacteria (<0.01) vadinHA49 (<0.01) Pla3 lineage (3.19%) Thermoanaerobaculia (0.05%) OM190 (<0.01) Coriobacteriia (<0.01) Chlamydiae (1.49%) Subgroup 22 (0.02%) Acidobacteriia (<0.01) Ignavibacteria (<0.01) Planctomycetacia (0.77%) Acidimicrobiia (0.02%) BD7−11 (<0.01) Anaerolineae (<0.01) Holophagae (0.73%) Verrucomicrobiae (0.02%) Fusobacteriia (<0.01) Negativicutes (<0.01) Actinobacteria (0.39%) uncultured bacterium (0.02%) Gemmatimonadetes (<0.01) Rubrobacteria (<0.01)

Figure 3.2: Relative abundance of reads assigned to each class. The mean relative abundance of each class is shown in brackets. Mock community and negative control samples are omitted.

3.3.2 Phenotypic changes in the anemones

Although the difference in dark-adapted quantum yield (Fv/Fm) of B. minutum between the control and heat-treated anemones was significant at Day 8 (Student’s t-test, p = 0.010) and

Day 10 (Student’s t-test, p = 0.043), there was little difference in Fv/Fm between the sample types until the temperature exceeded 31 °C on Day 10 (Figure 3.3a). Thereafter, Fv/Fm values

in the heat-treated anemones dropped sharply, clearly indicating the onset of damage to PSII in B. minutum, whereas Fv/Fm values in the control anemones remained relatively stable.

0.5 5 * * * * 34 * * * * 34 0.4 4

32 Temp (°C) 32 Temp (°C) Control m 0.3 • 3 ) / mg Heat-treated 30 6 30 / F • v

F 0.2 2 28 28

0.1 cells (1 0 1 26 26 0.0 0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (a) Day (b) Day

Figure 3.3: (a) Dark-adapted quantum yield (Fv/Fm) (for each datapoint, n = 6), and (b) Symbiodinaceae cell density (cells (106)/mg) (for each datapoint, n = 3). Error bars ± 1 SEM. Asterisks indicate significant differences between control and heat-treated values (Student’s t-tests, α = 0.05).

Symbiodiniaceae cell densities in the heat-treated anemones declined steadily from Day 0, however values did not differ significantly from the control anemones until Day 8 (Student’s t-test, p = 0.013) (Figure 3.3b). Overall, Symbiodiniaceae cell density in the heat-treated anemones underwent a significant decline from 3.8 × 106 to 1.3 × 106 cells/mg host protein (Student’s t-test, p < 0.001). Although Symbiodinaceae cell density fluctuated in the control anemones, there was little difference between Day 0 and Day 14 values (3.2 × 106 versus 3.0

6 × 10 cells/mg host protein). The consistency of Fv/Fm and cell densities in the control anemones throughout the treatment period suggest that Day 0 values approximate normal levels.

3.3.3 Changes in alpha diversity of the bacterial microbiomes

Bacterial community richness fell significantly from Day 0 to Day 14 in both the control (Student’s t-test, p = 0.008) and heat-treated (Student’s t-test, p = 0.004) anemones (Figure 3.4a). At Day 0, the average number of bacterial ASVs in the control and heat-treated anemones was almost identical, but by Day 14 the bacterial ASVs in the control and heat-

treated anemones had dropped on average to 181 and 139, respectively. Whilst the average number of bacterial ASVs in the heat-treated anemones followed a general downward trend, those in the control anemones almost halved between Day 6 (237) to Day 8 (135), before recovering. Despite their different trajectories, there was a significant difference in average observed ASVs between the control and heat-treated anemones only at Day 14 (Student’s t- test, p = 0.036). A survey of unique bacterial ASVs present in the control and heat-treated anemones at Days 0 or 14 further illustrated the initial similarity in richness in the sample types, and higher overall reduction in richness in the heat-treated anemones (Figure A 11). Bacterial community evenness in the control and heat-treated anemones remained high and within a narrow range (Simpson index: 0.97-0.94) (Figure 3.4b). Although an overall drop in evenness for the heat-treated anemones was significant (Student’s t-test, p = 0.035), this did not result in a significant difference in Simpson evenness between the control and heat- treated anemones at Day 14. Shannon index values showed that overall alpha diversity of the bacteria in the control and heat-treated anemones fell throughout the treatment period, and the trend was generally comparable between the sample types (Figure 3.4c). Whilst the overall drop in Shannon diversity for the heat-treated anemone bacteria was significant (Student’s t-test, p = 0.042), the difference in Shannon diversity between the bacteria in the control and heat-treated anemones at Day 14 was not.

* 0.98

250 4.25 0.97

200 0.96 4.00 Simpson Index Shannon Index Observed ASVs 0.95 150 3.75

• Control 0.94 Heat-treated 100 • 3.50 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (a) Day (b) Day (c) Day

Figure 3.4: (a) Average number of observed amplicon sequence variants (ASVs), (b) Simpson index values, and (c) Shannon index values. For each datapoint, n = 6. Error bars ± 1 SEM. Asterisks indicate significant differences between control and heat-treated values (Student’s t-tests, α = 0.05).

A survey of unique bacterial ASVs present in the control or heat-treated anemones at Days 0 and 14 provided an initial insight into changes in beta diversity within each sample type (Figure 3.5). Only a subset of ASVs in each sample type persisted to the end of the experiment, however they were the dominant community members (82.2%–96.2% relative abundance). Bacterial ASVs detected at Day 14 but not Day 0 were likely low-abundance community members that increased to detectable levels, rather than bacteria introduced during the experiment, as the only input was sterile RSS-water.

Control anemones Heat-treated anemones

Day 488 250 259 Day Day 581 196 168 Day 0 14 0 14 5.0 95.0 96.2 3.8 17.8 82.2 95.5 4.5

(a) (b)

Figure 3.5: Unique and common bacterial ASVs at Day 0 and 14 in (a) control, or (b) heat- treated anemones. Inset numbers indicate relative abundance (%) on Day 0 or 14.

3.3.4 Changes in beta diversity of the bacterial microbiomes

Datapoints in an nMDS ordination of the anemone microbiomes converged during the study period, indicating that the bacterial communities of the control and heat-treated anemones became more similar over time (Figure 3.6a). Thus, time rather than treatment was the primary grouping factor. However, as seen in plots showing the original ordination separated by day (Figure 3.6b), the datapoints appeared to cluster in a group-wise manner at Day 12, suggesting that the control and heat-treated anemones were developing distinct bacterial community compositions. This trend continued through to Day 14, and a GLM analysis indicated that by Day 14 the bacterial communities of the control and heat-treated anemones had become significantly different (manyGLM, p = 0.041) (Table A 13).

4 6 Control 10 10 4 4 12 Treated 14 4 14 12 12 4 2 14 12 6 10 10 2 2 0 1410 2 4 1414 14 10 2 1212 104 8 2 4 14 4 0 Day 14 886 14 6 02 6 14 10 8 6 8 8 6 2 0 10 8 810 4 10 12 124 0 0 Day 0 (26 °C, 26 °C) Day 2 (26 °C, 27 °C) Day 4 (26 °C, 28 °C) Day 6 (26 °C, 29 °C) 0 6 2 12 2 12 2 NMDS2 12 2 0 148 10 6 6 0 4 2 8 6 12 0 0 8 0 0 6 8 8 4 10 12 6 2D stress = 0.19 14 p = 0.041 (a) NMDS1 (b) Day 8 (26 °C, 30 °C) Day 10 (26 °C, 31 °C) Day 12 (26 °C, 32 °C) Day 14 (26 °C, 33 °C)

Figure 3.6: (a) nMDS ordination of the anemone-associated bacterial communities based on Bray-Curtis distances, and (b) plots showing datapoints from the original ordination for each day. Each datapoint represents one sampling unit.

3.3.5 Changes in abundance of selected bacterial taxa

The six most abundant taxonomic classes across all samples were Alphaproteobacteria (44.95%), Gammaproteobacteria (21.12%), Bacteroidia (12.14%), Deltaproteobacteria (9.98%), Spirochaetia (4.1%) and Pla3 Lineage (3.2%), which accounted for >95% of all bacterial taxa (Figure 3.7). The relative abundance of most class-level bacterial taxa in control and heat-treated anemones was similar from Days 0 to 10. However, in some classes there was a divergence between the control and heat-treated anemones after Day 10, with significant increases in Alphaproteobacteria (Student’s t-test, p = 0.025) and Bacteroidia (Student’s t-test, p = 0.005), and decreases in Deltaproteobacteria (Welch’s t-test, p = 0.009) and Spirochaetia (Mann-Whitney U test, p = 0.031) in the heat-treated anemones compared to the control anemones. From Day 0 to Day 14, the control and heat-treated anemones underwent comparable decreases in Gammaproteobacteria taxa from 28.0% to 19.5%, although only the overall change in the control anemones was significant (Student’s t-test, p = 0.008), and comparable increases in Pla3 Lineage taxa from 0.6% to 5.0%, which was a significant increase in both the control (Mann-Whitney U test, p = 0.031) and heat-treated anemones (Mann-Whitney U test, p = 0.031).

60 Alphaproteobacteria * 60 Gammaproteobacteria * 34 34 50 50

32 Temp (°C) 32 Temp (°C) 40 40

30 30 30 30

20 28 20 28

10 10 26 26 Relative Abundance (%) Relative Abundance (%) 0 0

0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (a) Day (b) Day

60 Bacteroidia * * 60 Deltaproteobacteria * 34 34 50 50

32 Temp (°C) 32 Temp (°C) 40 40

30 30 30 30

20 28 20 28

10 10 26 26 Relative Abundance (%) Relative Abundance (%) 0 0

0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (c) Day (d) Day

60 Spirochaetia * * 60 Pla 3 Lineage 34 34 50 50

32 Temp (°C) 32 Temp (°C) 40 • Control 40 • Heat-treated 30 30 30 30

20 28 20 28

10 10 26 26 Relative Abundance (%) Relative Abundance (%) 0 0

0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (e) Day (f) Day

Figure 3.7: Changes in the six most abundant bacterial classes across all samples: (a) Alphaproteobacteria, (b) Gammaproteobacteria, (c) Bacteroidia, (d) Deltaproteobacteria, (e) Spirochaetia, and (f) Pla3 Lineage. For each datapoint, n = 6. Error bars ± 1SEM. Asterisks indicate significant differences between control and heat-treated values (see main text for test types, α = 0.05).

The bacterial communities of the control and heat-treated anemones experienced similar declines in ASVs of the genus Vibrio from Day 0, which generally continued throughout the study period. At Day 0, Vibrio averaged 4.7% of the bacteria in both control and heat-treated 66

anemones (Figure 3.8). However, by Day 14, Vibrio had dropped significantly in the control anemones to 0.42% (Mann-Whitney U test, p = 0.031), and in the heat-treated anemones to 0.04% (Mann-Whitney U test, p = 0.031).

• Control 34 Heat-treated 6 •

32 Temp (°C)

4 30

2 28

26 Relative Abundance (%) 0 0 2 4 6 8 10 12 14 Day

Figure 3.8: Changes in relative abundance of Vibrio sp. ASVs. For each datapoint, n = 6. Error bars ± 1SEM.

3.3.6 Indicator Species Identification

Twelve bacterial species were identified in an IndVal analysis (Table A 14). However, only six showed changes in relative abundance suggestive of a response to elevated temperature (Figure A 12). The others displayed high variability throughout the treatment period, making interpretation of their abundance changes difficult, and were thus discounted as potential indicator species (Figure A 13). Of the six considered valid indicator species, an ASV from the family Saprospiraceae was moderately abundant in both the control and heat-treated anemones until Day 12, but from Day 12 to Day 14 it increased substantially in the heat- treated anemones from 4.9% to 13.3% relative abundance (Figure A 12a). Two ASVs from the class Gammaproteobacteria and family Terasakiellaceae, respectively, were relatively stable until Days 8–10 (Figure A 13b,c). Thereafter, both underwent rapid increases in relative abundance in the heat-treated anemones. A second Terasakiellaceae ASV increased in abundance in the heat-treated anemones compared to control anemones until Day 12, then dropped sharply; a pattern that was replicated to a lesser extent in the control anemones (Figure A 12d). An indicator species of the genus Spirochaeta 2 (Figure A 12e) was prevalent in both the control and heat-treated anemones at Day 0, averaging 4.9% relative abundance

across all samples. Although the relative abundance of Spirochaeta 2 fluctuated in both groups, it dramatically decreased in the heat-treated anemones from Day 8 (7.1%) to Day 14 (0.1%). A sixth indicator species, from the family Rhizobiaceae (Figure A 12f), also fluctuated but generally increased in abundance in the heat-treated anemones across the study period.

3.4 Discussion

3.4.1 Factors underpinning bleaching

E. diaphana that were exposed to rising temperature (26 °C to 33 °C) maintained near-normal

Fv/Fm values until 31 °C was exceeded, thus suggesting an upper thermal limit for quantum efficiency of photosystem II in the Symbiodiniaceae harboured by these anemones. However, a notable albeit non-significant drop in Symbiodiniaceae cell density as soon as the temperature increased above ambient, showed that the anemones had low resistance to thermal bleaching.

The rapid onset of bleaching in the heat-treated anemones may have been linked to lack of food. Starvation has been practiced in previous E. diaphana experiments (Bieri et al. 2016; Hillyer et al. 2016; Marty-Rivera, Yudowski & Roberson 2018; Röthig et al. 2016; Zaragoza et al. 2014) despite observed reductions in Symbiodiniaceae cell density following food deprivation (Clayton & Lasker 1984; Cook, D'Elia & Muller-Parker 1988; Davy & Cook 2001), but the possible impact of this on experimental outcomes has rarely been acknowledged (Lehnert et al. 2014). Our data suggest that, in future work, continued feeding is advisable, but at levels that maintain normal thermal tolerance rather than enhance it as seen in some coral species under heterotrophic conditions (Borell et al. 2008; Ferrier-Pagès et al. 2010; Grottoli, Rodrigues & Palardy 2006).

Although the incubator conditions were matched to the culture collection environment, relocation of the anemones may have also increased bleaching susceptibility as 10 days may have been inadequate for full acclimation. Consequently, longer acclimation periods may be advisable in studies with E. diaphana to avoid confounding.

3.4.2 Environmental stressors reduced alpha diversity

Overall, bacterial community richness decreased in the heat-treated anemones, which was contrary to previous reports (Ahmed et al. 2019). However, such a comparison is difficult as the E. diaphana in (Ahmed et al. 2019) were held at 32 °C for >2 years and although referred to as ‘heat-stressed’, no evidence of this, such as reduced Symbiodiniaceae cell density compared to anemones at ambient temperature was provided, and no bleaching was reported. Reductions in bacterial alpha diversity have been seen in corals exposed to short- term (Grottoli et al. 2018) or long-term heat stress (Tracy et al. 2015), and in other microbiomes subjected to environmental stressors, including heat (Rocca et al. 2019). Our results concur with these findings. However, increases in alpha diversity among heat-stressed corals are more common (Bourne et al. 2008; Santos et al. 2014; Tout et al. 2015). This may indicate that the behaviour of E. diaphana’s bacterial microbiome is atypical among cnidarians, or that other factors influenced the bacterial community changes.

Shifts in bacterial alpha diversity in the control and heat-treated anemones were largely congruent, which could infer that heat was not, or was only partly responsible for changes in richness. The possible influence of starvation and incomplete acclimation on bleaching has been noted, with acclimation potentially playing a particular role in the initial changes seen in bacterial composition. For example, a study of corals transferred from a reef to aquaria found that the bacterial communities in the coral surface mucus layer (SML) took 14–28 days to stabilize (Pratte, Richardson & Mills 2015). Although the E. diaphana in the present study did not undergo such a dramatic relocation, the decline in richness among all the anemones could represent the late stages of acclimation, with the recovery in richness in the control anemones from Day 10 indicating a return to a normal state. As this return was not matched by the heat- treated anemones, it is reasonable to suggest that beyond Day 12 (i.e., above 32 °C) temperature was the main driver of change in alpha diversity for the heat-treated anemones.

3.4.3 Turnover of low-abundance ASVs drive shifts in beta diversity

Transience appears to be a common trait among coral and E. diaphana microbiome members (Brown et al. 2017; Hester et al. 2016; Sweet et al. 2017), and this was evident in the present study as the majority of ASVs detected in the control and heat-stressed samples at Day 0 were

not seen at Day 14, and vice versa (Figure 3.5). These transient bacteria, whilst comprising a small proportion of their bacterial communities (3.8%–17.8%), were high in number, suggesting the presence of many species below the limit of detection that multiplied as conditions became favourable. Such a reservoir may benefit the host during stress by allowing their bacterial communities to restructure with members better at supporting holobiont homeostasis as proposed by the coral probiotic hypothesis (Reshef et al. 2006).

The overall loss of richness in the control and heat-treated anemone bacterial associates likely led to the reduction in dissimilarity, and hence a reduction in beta diversity across the samples, by removing low-abundance bacteria that had inflated sample-sample dissimilarity. This is a common phenomenon (Shade et al. 2014). However, differences in beta diversity emerged at Day 12 and became significant at Day 14, demonstrating that heat-stressed E. diaphana have bacterial associates distinct from E. diaphana at 26 °C. Further testing is needed to determine whether shifts induced by thermal stress stabilize, and whether the new bacterial communities can support the anemones for long periods at high temperature. Previous findings (Ahmed et al. 2019) suggest that high variability rather than uniformity eventuates in bacterial communities of E. diaphana exposed to high temperature. These changes may assist long-term survival at temperatures associated with bleaching in E. diaphana (Gates, Baghdasarian & Muscatine 1992).

3.4.4 Changes in bacterial associates at a high taxonomic level were apparent

The relative abundance of Alphaproteobacteria and Gammaproteobacteria in the control and heat-treated anemones was comparable up to 31 °C, then diverged significantly. However, the differences did not remain significant above 32 °C. Nevertheless, similar changes in these bacterial classes have been observed in thermally-stressed coral, which were attributed to shifts in the sugar composition of the coral SML (Lee et al. 2016). Overall changes in relative abundance of some bacterial classes that were comparable in both control and heat-treated anemones, such as the decrease in Gammaproteobacteria and increase in Pla3 Lineage taxa, could be indicative of ongoing acclimation and normalization of the bacterial communities after relocation. However, relative increases in Bacteroidia and decreases in Deltaproteobacteria and Spirochaetia in the heat-treated anemones from Days 10–12, point

to a temperature-influenced response and thermal tipping point for bacterial stability of 31– 32 °C for GBR E. diaphana.

3.4.5 Were Vibrio victims of competition?

The near elimination of Vibrio sp. from the heat treated anemone bacterial communities was unexpected as elevated temperature has been shown to increase Vibrio abundance in coral (Tout et al. 2015; Vezzulli et al. 2010), posing a threat to coral health through temperature- induced upregulation of virulence factors (Kimes et al. 2012; Munn 2015). Unexpected also was the parallel decline of Vibrio in the control and heat-treated anemones, which suggests factors common to all the bacterial associates, such as predation or antagonism by other bacteria, determined the fate of Vibrio. Halobacteriovorax sp. prey upon Vibrio (Welsh et al. 2015) but none were detected. However, Roseobacter sp., which has members with antagonist activity against Vibrio sp. (Ruiz-Ponte et al. 1999) were present in control and heat- treated anemones at all timepoints. Stressors common to both control and heat-treated anemones, relocation and nutrient deprivation, may have created a situation in which Vibrio were displaced by more competitive bacterial associate members (Hibbing et al. 2010). Although possibly indicative of dysbiosis, the removal of potentially pathogenic Vibrio sp. could benefit thermally stressed E. diaphana.

3.4.6 Specific bacteria as biomarkers for thermal stress

Bacterial indicator species have been recommended as biomarkers for coral stress (Glasl, Webster & Bourne 2017). In the present study, the relative abundance of six bacterial species changed with rising temperature in ways that suggest they could be used to monitor the response of GBR E. diaphana exposed to thermal stress. The response of each indicator species might be due to temperature moving towards or away from a growth optimum, or other mechanisms.

An ASV from the family Saprospiraceae was the most abundant of the proposed indicator species. Saprospiraceae species have been shown to increase in abundance in heat-sensitive corals exposed to thermal stress (Ziegler et al. 2017). Some prey on algae (Furusawa et al. 2003; Shi et al. 2006) or bacteria (Lewin 1997). Therefore, the availability of released 71

Symbiodiniaceae or an increase in bacterial prey in the heat-treated anemones could explain the increase of this ASV.

Two indicator species belonged to Terasakiellaceae, a bacterial family with members potentially involved in nitrogen cycling in some nutrient-limited corals (Weiler, Verhoeven & Dufour 2018). Under ambient conditions, nitrogen availability is thought to be limited by cnidarian hosts to control Symbiodiniaceae division (Hoegh-Guldberg & Smith 1989; Muscatine et al. 1989). The increase of Terasakiellaceae ASVs may therefore signify opportunistic growth in a system where host-symbiont nitrogen regulation has been disrupted due to thermal stress.

An indicator species of the genus Spirochaeta 2 was almost eliminated in the heat-treated anemones above 30 °C, suggesting sensitivity to temperature. Bacteria from the phylum Spirochaetes have been identified in E. diaphana (Brown et al. 2017) and corals (Closek et al. 2014; Kimes et al. 2013; Lawler et al. 2016), including corals with high thermal tolerance (Ziegler et al. 2017). However, the rapid decline of this indicator species above 30 °C, shows it may serve as an early indicator of thermal stress in GBR E. diaphana.

The relative abundance of an indicator species from the family Rhizobiaceae increased as bleaching progressed. Bacteria in the family Rhizobiaceae may be intracellular associates of marine alga (Hollants et al. 2013; Schwenk, Nohynek & Rischer 2014). However, the increase of this ASV alongside a diminishing Symbiodiniaceae population may indicate that increasing temperature was a stronger promoter of growth than Symbiodiniaceae association. Despite this, it is interesting to note the possible relationship of several other indicator species with Symbiodiniaceae. Although the response of Symbiodiniaceae to temperature increase was rapid compared to the bacterial communities, thus suggesting the behaviour of each was independent, it would be naïve to believe that the former might not impact the latter, either directly or through an overall impact on the holobiont.

Although a Gammaproteobacteria indicator species appeared to respond to temperature, lack of taxonomic identification limits speculation of its behaviour. Another limitation of this and the other indicator species as universal biomarkers in future studies is a requirement for them to occur in the E. diaphana microbiomes. Due to the transience of many E. diaphana associated bacteria, this is uncertain. A further limitation, particularly for low-abundance species, is the use of relative rather than absolute abundance to describe changes in 72

prevalence as this may skew abundances (Jackson 1997). Bacterial load can be a strong indicator of stress or disease (Vandeputte et al. 2017), and when used to transform count data into absolute abundance can provide a more comprehensive picture of microbiome dynamics (Props et al. 2017). Newly proposed methods such as spike-in of synthetic DNA during sample preparation could extend and improve microbiome data interpretability in future studies (Stämmler et al. 2016; Tourlousse et al. 2017).

3.5 Conclusion

The bacterial microbiome of GBR E. diaphana is impacted by environmental stressors. In the present study, a reduction in bacterial community richness in the anemones at both ambient and elevated temperatures, and lowered bleaching resistance, may have been linked to incomplete acclimation or nutrient deprivation. However, differences between bacterial associates of control and heat-treated anemones in richness, beta diversity and taxon abundances that emerged above 31 °C and became significant above 32 °C suggest that temperature drives change above this threshold. Prolonged exposure to thermal stress may lead to further changes, such as increased beta diversity as proposed elsewhere, and this may support functions relevant to holobiont health. Some bacteria respond to thermal stress in ways that suggest they could be used to assess the impact of elevated temperature on GBR E. diaphana. These data improve our understanding of the E. diaphana bacterial microbiome, and hence this model organism’s utility in cnidarian bleaching research.

Chapter 4: Towards the generation of gnotobiotic Exaiptasia diaphana

L. Hartman, M. van Oppen and L. Blackall designed the experiment. L. Hartman performed the experimental work, analysed the data, wrote and edited the text. M. van Oppen and L. Blackall reviewed and edited the text.

4.1 Introduction

The tropical sea anemone Exaiptasia diaphana (previously Aiptasia pallida (Grajales & Rodríguez 2014; ICZN 2017)) has become an important coral model; its intracellular symbiosis with photosynthetic algae of the family Symbiodiniaceae making it useful for studying cnidarian host-symbiont relationships (Voolstra 2013; Weis et al. 2008). The loss of Symbiodiniaceae from host tissue (i.e., bleaching) has been particularly well studied (Gates, Baghdasarian & Muscatine 1992; Goulet, Cook & Goulet 2005; Núñez-Pons, Bertocci & Baghdasarian 2017; Perez, Cook & Brooks 2001; Tolleter et al. 2013), spurred by mass coral bleaching events caused by environmental conditions linked to climate change (Hughes et al. 2018).

E. diaphana’s ability to survive in a Symbiodiniaceae-free (i.e., aposymbiotic) state has clarified cnidarian metabolic processes by separating host and symbiont to reveal the roles played by each in nutrient transfer (Lehnert et al. 2014; Medrano et al. 2019; Oakley et al. 2016), and their individual responses to environmental stress (Bieri et al. 2016; Burriesci, Raab & Pringle 2012; Matthews et al. 2017). However, studies often ignore the influence of bacteria on the holobiont; a functional entity comprising the host and all its microbial partners (Margulis 1991).

The bacterial component of the cnidarian holobiont influences host health through, for example, its involvement in nutrient cycling (Ceh et al. 2013; Lesser et al. 2007; Raina et al. 2009; Thomas et al. 2010) and pathogen protection (Brown & Rodriguez-Lanetty 2015; Krediet et al. 2013; Mao-Jones et al. 2010). Therefore, removing bacteria from the holobiont represents an important next step in the elucidation of host and symbiont functional processes (Rohwer & Kelley 2004).

Although the generation of germfree Symbiodiniaceae have been reported (Xiang et al. 2013), similar coral cultures or cell lines have not (Rohwer & Kelley 2004). E. diaphana may be able fill to fill this gap. However, creating germfree E. diaphana may not be feasible as the anemones might not survive without bacteria. Instead, bacteria may be required for normal host health and development, as in many other organisms (Brune 2014; Gonzalez-Perez et al. 2016; Heijtz et al. 2011). Thus, E. diaphana that harbour reduced bacterial communities could represent a practical alternative to germfree cultures. Strictly, these anemones would be

described as gnotobiotic, that is, organisms with precisely defined microbial communities (Basic & Bleich 2019). Gnotobiotic plants (Müller et al. 2016), animals (Gordon & Pesti 1972) and insects (Luckey 1963b) are well-established and have helped progress microbiome research away from purely correlational studies (Basic & Bleich 2019; Fritz et al. 2013; Neville, Forster & Lawley 2018; Vorholt et al. 2017).

Generating gnotobiotic organisms is typically via sterilisation followed by either conventionalisation through inoculation with a characterised ‘normal’ bacterial community, or inoculation with a simple synthetic community (Mooser, Gomez de Agüero & Ganal- Vonarburg 2018), for example the ‘altered Schaedler Flora’ as used in mice (Dewhirst et al. 1999). Gnotobiotes can also be generated through decontamination; the process of reducing the number of hosted bacterial species, often with antibiotics (Luckey 1963a). In each case, the aim is simplification and uniformity of the bacteria to reduce their influence and the biological variability between hosts (Macpherson & McCoy 2015).

There has been one report of germfree E. diaphana, wherein anemones were exposed to the antibiotics ampicillin and streptomycin to render them “aseptic” (Wang & Douglas 1999). Germfree status was determined by absence of bacterial growth on nutrient-supplemented agar inoculated with treated anemone tissue and culture medium, and by examination of anemone tissue using light and election microscopy. However, as many bacteria defy culture (Stewart 2012), and the extent to which an anemone can be screened by microscopy is limited by practical constraints, the germfree status of the anemones is uncertain.

In a recent step towards development of gnotobiotic E. diaphana, a protocol was published describing a method for bacterial depletion (Costa, Cárdenas & Voolstra 2019). Depletion was achieved via exposure to four antibiotics; Rifampicin, Nalidixic acid, Carbenicillin and Chloramphenicol, each with a different mechanism of action. Detection of bacteria in the treated anemones was by culture methods and PCR. The absence of bacterial growth on marine agar inoculated with treated anemone homogenates and culture water after five days of incubation, and absence of PCR amplification from DNA extracts using primers targeting the bacterial 16S rRNA gene, was used as evidence of bacterial depletion. However, it was noted that depletion would only be maintained if treatment was ongoing. In addition, bacterial load was not quantified, and the bacterial communities were not characterised, leaving the extent of the reduction, and the uniformity and composition of the resulting bacterial communities 76

unknown. Consequently, the efficacy of antibiotic approaches for generating gnotobiotic E. diaphana cultures is unclear.

Here, I report the generation of bacteria-reduced E. diaphana, by exposure to antibiotics. My aims were to (1) evaluate the impact of antibiotic treatment on anemone health by measuring changes in host tissue Symbiodiniaceae density, (2) to quantify the reduction in bacterial load of treated anemones, and (3) to compare the bacterial community compositions of treated anemones to gauge the uniformity of the resulting cultures. In so doing, I sought to assess the efficacy of antibiotic approaches for bacterial depletion in E. diaphana. The methods and data presented here will assist future efforts to create germfree or gnotobiotic E. diaphana cultures.

4.2 Materials and methods

4.2.1 Experimental set-up and antibiotic treatment

Clonal adult GBR-origin E. diaphana (n = 108; genotype AIMS2 (Dungan et al. 2020)) were randomly selected from a single tank in the University of Melbourne (UoM) anemone culture collection and transferred into single wells within sterile 12-well plates (Costar 3513, Corning, USA). The anemones were maintained in seawater reconstituted from Red Sea Salt™ (R11065, Red Sea, USA) with reverse osmosis water, at a salinity of ~34 parts per thousand. The plates were kept in clear plastic zip-lock bags to prevent contamination, with paper towels moistened with sterile water to minimise evaporation, and randomly positioned in a Hi-Point 740 incubator (Thermo Fisher, Australia) at 26 °C with lighting at ~33 µmol m–2 s–1 on a 12 h:12 h light-dark cycle. During a one-week acclimation period the anemones were fed twice with freshly hatched Artemia salina nauplii (Salt Creek, Premium GSL, USA), and the water was changed three times (Table 4.1). After the acclimation period, antibiotic dosing and sampling commenced. All subsequent operations were performed using aseptic techniques. During the treatment period, water changes were performed with RSS-water sterilised by autoclaving and filtered (0.45 µM) to remove precipitated salt (hereafter, ‘sRSS-water’). When sampling coincided with dosing or water changes, samples were collected first.

Table 4.1: Maintenance, antibiotic treatment and sampling schedule.

Monday Tuesday Wednesday Thursday Friday Sat. Sun.

Transfer anemones Feed anemones Change water Hatch A. salina Change water Hatch A. salina Feed anemones Day 0 Day 1 Day 2 Day 3 Day 4

Sampling Sampling Change water Sampling Change water Change water Feed anemones Treatment Hatch A. salina Treatment Treatment Feed anemones Hatch A. salina Day 7 Day 8 Day 9 Day 10 Day 11

Sampling Feed anemones Change water Hatch A. salina Change water Change water Treatment Treatment Treatment Feed anemones Hatch A. salina Day 14 Day 15 Day 16 Day 17 Day 18

Sampling Feed anemones Change water Hatch A. salina Change water Change water Treatment Treatment Treatment Feed anemones Hatch A. salina Day 21

Sampling

On treatment days, half the anemones (n = 54) were exposed to antibiotics selected for their different mechanisms of action and activity against Gram-positive or Gram-negative bacteria, and their previous use on cnidarians, Symbiodiniaceae or sponges (Table 4.2). Maximum tolerable doses were determined in pre-treatment testing by exposing anemones to increasing dilutions of the antibiotics until they maintained normal appearance, growth and feeding across an 18-day test period.

Table 4.2: Antibiotics used against E. diaphana- and A. salina-associated bacteria.

Antibiotic Dose Target / Mechanism of action Gram +/– References (µg/mL) activity Carbenicillin 25 DD-transpeptidase / – (Costa, Cárdenas & Voolstra 2019; Reyes-Bermudez & Miller Inhibits cell wall synthesis 2009; Soffer, Gibbs & Baker 2008; Sweet, Croquer & Bythell 2014; Weiland-Bräuer et al. 2015)1 Chloramphenicol 25 23S rRNA / +/– (Costa, Cárdenas & Voolstra 2019; Rahat & Dimentman 1982; Inhibits protein synthesis Reyes-Bermudez & Miller 2009; Weiland-Bräuer et al. 2015) Nalidixic Acid 15 Gyrase / – (Costa, Cárdenas & Voolstra 2019; Mills et al. 2013; Inhibits DNA replication Richardson et al. 2012) Neomycin 10 30S rRNA assembly / +/– (Glasl, Herndl & Frade 2016; Polne-Fuller 1991; Weiland- Inhibits protein synthesis Bräuer et al. 2015) Polymyxin B 10 Increases gram negative cell wall – (Glasl, Herndl & Frade 2016; Polne-Fuller 1991; Weiland- permeability Bräuer et al. 2015) Rifampicin 10 RNA polymerase / + (Costa, Cárdenas & Voolstra 2019; Glasl, Herndl & Frade Inhibits transcription 2016; Rahat & Dimentman 1982) Streptomycin 25 16S rRNA / + (D’Agostino 1975; Glasl, Herndl & Frade 2016; Polne-Fuller Inhibits protein synthesis 1991; Rahat & Dimentman 1982; Reyes-Bermudez & Miller 2009; Soffer, Gibbs & Baker 2008; Wang & Douglas 1999; Xiang et al. 2013) 1. References to the use of Penicillin family antibiotics with the same mechanism of action.

Control anemones were fed with A. salina hatched in sRSS-water. Treated anemones were fed with A. salina hatched in sRSS-water containing antibiotics at concentrations matching those used for the anemones.

4.2.2 Sampling and DNA extraction

On sampling days, six control and six treated anemones were taken for bacterial community analysis to track changes in community composition, and three control and three treated anemones were taken for Symbiodiniaceae density analysis to assess whether the treatment induced a stress response (i.e., bleaching) in the anemones. Three aliquots of a dense suspension of control and treated A. salina nauplii were also collected for bacterial community analyses. All samples were snap frozen in liquid nitrogen and stored at –80 °C until processing. DNA was extracted from the anemones and A. salina samples using an existing protocol (Wilson et al. 2002) modified according to Hartman, van Oppen and Blackall (2019).

4.2.3 Anemone–Symbiodiniaceae density measurement

Anemones collected for Symbiodiniaceae density assessment were homogenised in 1 mL of sRSS-water, and 200 µL of the homogenate was removed for total protein analysis by the Bradford assay (Bradford 1976), performed in triplicate. The remaining homogenate was centrifuged, and the supernatant removed. The pellet was washed twice with 1 mL sRSS- water, then resuspended in 800 µL sRSS-water. Triplicate counts of Symbiodiniaceae in the suspension were in performed on an automated cell counter (Countess II FL, Life Technologies). Counts were normalised to total anemone protein (mg) to account for anemone size differences.

4.2.4 Bacterial load assessment

Bacterial load in the anemones and A. salina was quantified by digital droplet PCR (ddPCR). Load was described by the number of bacterial 16S rRNA gene copies normalised to host reference gene copies in each DNA extract according to the B/H ratio (Sender, Fuchs & Milo 2016) or similarly, the S (symbiont)/H ratio (Mieog et al. 2009), with ratios expressed as

decimals to aid sample-sample comparison. This approach enables normalisation of bacterial data from samples with low mass or volume and has been used in insect studies (Catterson et al. 2018; Deb, Nair & Agashe 2019; Zhukova et al. 2017). Primers targeting single or low copy reference genes in E. diaphana and A. salina were used for host cell quantification (Table 4.3). Primers targeting a conserved 98-nucleotide sequence between the V2 and V3 regions of the bacterial 16S rRNA gene were used to estimate bacterial cell numbers (Table 4.3). No correction was made for 16S rRNA gene copy number variation between different bacterial taxa; therefore, the method is semi-quantitative.

Table 4.3: Primers used to estimate host and bacterial cell numbers.

Target (gene) Primer name Primer sequence Product size ddPCR annealing References (nucleotides) temperature (°C) E. diaphana (EF1-α) Ef1-α-fwd AGCACTGAGCCACCATACAG 88 60 (Hawkins et al. 2016) Ef1-α -rev TTGGGTTATAGCCGGTCTTC 88 60 (Hawkins et al. 2016) A. salina (ß-Actin) art-actin-fwd GGTCGTGACTTGACGGACTATCT 147 60 (Jiang et al. 2007; Niu et al. 2014; Valverde et al. 2019) art-actin-rev AGCGGTTGCCATTTCTTGTT 147 60 (Jiang et al. 2007; Niu et al. 2014; Valverde et al. 2019) Universal bacteria 259-fwd GGTAAHRGCYYACCAAG 98 54 (Wang & Qian 2013) (conserved inter V2-V3 357-rev CTGCTGCCTCCCGTAGGAG 98 54 Reverse complement 16S rRNA gene region) of “primer 1” (Muyzer, de Waal & Uitterlinden 1993)

Before performing ddPCR, DNA was digested to improve droplet encapsulation of individual DNA fragments, and signal generation from low-concentration bacterial DNA (Dong et al. 2014). Sample DNA was digested for 1.5 hr at 37 °C in a volume of 20 µL comprising 7 µL sterile water, 2 µL 10x restriction enzyme buffer, 10 µL DNA, and 1 µL (~20 U) HindIII (R3104S-HF, New England BioLabs, Australia). The digested DNA was then quantified by PicoGreen (P11496, Thermo Fisher, Australia) and diluted ≥1:4 to 10-20 ng/µL to create practical working concentrations and prevent PCR inhibition by the enzyme buffer. ddPCRs for each DNA sample were prepared in an initial volume of 44 µL comprising 24 µL EvaGreen Supermix (QX200, Bio-Rad, Australia), sterile water, and ~30 ng of digested DNA to ensure that DNA concentrations for each bacteria-host reaction pair were within the dynamic range of the ddPCR system. The mixture was then split into 2 × 22 µL aliquots, one for host cell quantification (i.e., E. diaphana or A. salina), and one for bacteria. One microlitre each of the appropriate 5 nM forward and reverse primers (Table 4.3) were then added to each

reaction aliquot, giving final primer concentrations of 200 nM and volumes of 24 µL. From each 24 µL volume, 20 µL was loaded into a DG8 cartridge (1864008, Bio-Rad, Australia), followed by 70 µL of droplet generation oil for EvaGreen (1864005, Bio-Rad, Australia), and droplets were generated in a droplet-generator (QX200, Bio-Rad, Australia). A volume of 40 µL per reaction was then transferred to a 96-well plate and foil-sealed (1814040, Bio-Rad, Australia) with a thermal plate-sealer (PX1, Bio-Rad, Australia). One no-template control (NTC) reaction was included per plate. Thermal cycler settings were optimised according to (Witte et al. 2016): 1 cycle at 95.0 °C × 5 min; 50 cycles at 95 °C x 1 min + 54 °C or 60 °C (refer Table 4-3) × 2 min; 1 cycle at 4.0 °C × 5 min, 1 cycle at 90 °C × 5 min; 12 °C hold. All ramp rates were 1 °C/s. Droplets were read on a Bio-Rad QX200 droplet reader, and fluorescence data were analysed in QuantaSoft v1.7.4.0917 — output data are provided in Appendix 3 in accordance with the digital MIQE (Minimum Information for publication of Quantitative digital PCR Experiment) requirements (Huggett et al. 2013).

4.2.5 Sample and data processing for bacterial community analysis

In preparation for bacterial community analysis by metabarcoding, sample DNA was first amplified by end-point PCR using primers with Illumina adapters (not shown) targeting the V5-V6 regions of the 16S rRNA gene: 784F [5ʹ AGGATTAGATACCCTGGTA 3ʹ]; 1061R [5ʹ CRRCACGAGCTGACGAC 3ʹ] (Andersson et al. 2008). Triplicate PCRs were performed in 20 µL volumes comprising 1 µL template DNA, 10 µL MyTaq HS Mix polymerase (Bioline), 0.5 µL of 10 µM 784f, 0.5 µL of 10 µM 1061R, and 8 µL MilliQ water. Thermal-cycler settings were: 1 cycle at 95.0 °C × 3 min; 30 cycles at 95.0 °C, 55.0 °C and 72.0 °C × 15 s each; 1 cycle at 72 °C × 3 min; 12° C hold. Each triplicate was pooled, then checked by 1% agarose gel electrophoresis. No-sample DNA extractions and no-template PCRs were performed to identify contaminants introduced during sample preparation. A volume of 25 µL of pooled PCR product from each sample was sent to the Ramaciotti Centre for Genomics (RCG), Sydney, Australia for sequencing on a single Illumina MiSeq 2 x 250 bp run. RCG performed PCR product clean-up and normalisation as part of library preparation prior to sequencing. Demultiplexed MiSeq reads were joined in QIIME2 v2018.4.0 (Bolyen et al. 2019). Denoising, chimera filtering, and trimming was performed in DADA2 to correct sequencing errors, remove primer sequences, and low-quality bases (Callahan et al. 2016). Amplicon sequence variants (ASVs) with one 81

representative sequence were removed. Taxonomy was assigned in QIIME2 against a SILVA database (v 132) trained with a naïve Bayes classifier (Bokulich et al. 2018; Pedregosa et al. 2011; Quast et al. 2013; Wang et al. 2007). ASVs identified as eukaryotes, mitochondria, or chloroplasts were removed.

4.2.6 Data analyses

All data analyses were performed in R v3.6.0 (R Core Team 2018), with differences considered significant at α = 0.05 unless otherwise stated. Due to the irregular sampling timepoints, overall differences in Symbiodiniaceae cell densities and bacterial loads within and between the control and treated anemones over time were evaluated by generalised least square (GLS) models with the R package, nlme (Pinheiro et al. 2019). If significant differences by treatment or time, or treatment-by-time interactions were detected, Mann-Whitney U (Wilcoxon 1945) or Student’s t-tests (Student 1908) were performed to determine where differences existed. Tabulated ASV counts, taxonomic and meta data were imported into R and converted into a phyloseq object for bacterial community analysis (McMurdie & Holmes 2013). Rarefaction curves were generated, with the R package vegan (Oksanen et al. 2018) to assess whether the samples had been sequenced sufficiently to capture species diversity. Potential contaminating ASVs were identified and removed using the ‘prevalence’ method in the R package decontam (Davis et al. 2018) at the default threshold of p = 0.1. The community data were visualised in an nMDS ordination (weighted unifrac (Lozupone et al. 2011)) to identify samples with highly divergent bacterial compositions based on ASV abundance and phylogeny. Such samples were deemed outliers and were removed from the dataset. Alpha diversity metrics in the anemone- associated bacterial communities were calculated in vegan (Oksanen et al. 2018) and plotted over time with the R package ggplot2 (Wickham 2019) after sub-sampling the sequence data to 14 174 reads per sample. Community richness was described by number of observed ASVs per anemone. Community evenness was described using Simpson index (Simpson 1949). General alpha diversity was described using Shannon index (Shannon & Weaver 1949). Overall differences in alpha diversity values were assessed by GLS models, as above, and differences between control and treated anemones at each timepoint were assessed by paired Kolmogorov-Smirnov tests (Massey 1951) due to data non-normality and unequal sample sizes. Relationships between the untreated and treated anemones at Day 0 and day 21 were 82

visualised in an nMDS ordination based on unweighted unifrac distance. Unweighted unifrac was chosen over weighted unifrac to incorporate phylogenetic information of the bacteria without skewing the data due to bacteria with high abundance but low ubiquity. Bacterial community differences between the untreated and treated anemones at Day 0 and Day 21 were tested for significance using generalised linear models (GLM) of the count data collapsed to genus, against the explanatory variables ‘treatment’ and ‘day’ (Wang et al. 2012). After checking that a negative binomial distribution was appropriate for the data by visualisation of the residuals, likelihood ratio tests (LRT) were used to determine the deviance (i.e., goodness of fit) of the competing models across 999 sampling iterations. Common and unique ASVs in the Day 21 control and treated anemones were visualised in Venn diagrams to compare the complexity of untreated and bacterial communities. Stacked bar-charts of reads assigned to taxonomic families in the six Day 21 control anemones and six treated anemones were plotted (Wickham 2019) to compare bacterial community uniformity at the end of the treatment period. An analysis of differential pairwise abundances of bacterial genera in the untreated Day 0 anemones versus the antibiotic-treated Day 21 anemones was performed to identify taxa with significant binary log fold changes (L2FC). L2FC testing was performed using the R package DESeq2 (Love, Huber & Anders 2014) with α = 0.01 and Benjamini-Hochberg correction (Benjamini & Hochberg 1995). Genera with a change in relative abundance of <0.5% were filtered from the L2FC output due to the high number of low abundance genera. An analysis of ‘stable’ bacterial genera was also performed. Stable genera were deemed those with a relative abundance of >1.0% in the untreated anemones at Day 0 and in the antibiotic- treated anemones at Day 21, and a change in relative abundance from Day 0 to Day 21 of

<2.0%. The L2FC and stable output was compared to the A. salina bacterial communities to determine whether identified genera could have originated from the antibiotic-treated feedstock.

4.3 Results

4.3.1 Anemone algal cell density

The density of Symbiodiniaceae cells in the tissue of antibiotic-treated anemones dropped during the first two weeks of treatment, suggesting a treatment-related stress response

(Figure 4.1). Consequently, Symbiodiniaceae cell density was significantly lower in the treated anemones compared to the control anemones at Day 3 (Mann-Whitney, p = 0.012), Day 14 (Mann-Whitney, p = 0.004) and Day 21 (Mann-Whitney, p = 0.004) (Table A 17). Although there was a significant general reduction in Symbiodiniaceae in the antibiotic-treated anemones from Day 0 to Day 21 (Mann-Whitney, p = 0.004) (Table A 18), density doubled between Day 14 (2.27 cells x 106/mg) and Day 21 (4.40 cells x 106/mg) suggesting a recovery of the E. diaphana-Symbiodiniaceae symbiosis. Despite the apparent stress of antibiotic exposure, all treated anemones survived, and were similar in appearance and feeding behaviour to untreated anemones.

* * * 10

/mg total protein) 8 6

4 • Control Treated 2 •

Symbiodiniaceae (cells x 10 0 1 3 7 14 21 Day

Figure 4.1. Symbiodiniaceae density (cells × 106/mg total protein) in the control and treated anemones. For each datapoint, n = 3. Error bars ± 1 SEM. Asterisks indicate significant differences, α = 0.05.

4.3.2 Bacterial load (B/H ratio)

Reductions in bacteria in the treated anemones were evident in the ddPCR fluorescence output (Figure A 14), and bacterial load declined significantly in the treated anemones (Figure 4.2a) from Day 0 to Day 21 (Mann-Whitney, p = 0.031) (Table A 24). The most substantial drop in bacterial load in the treated anemones occurred from Day 0 to Day 1 after the first treatment, and this was also significant (Mann-Whitney, p=0.031). Bacterial load in the untreated A. salina (B/H = 0.094) was comparable to the average value for the pre-treated (i.e., Day 0) anemones (B/H = 0.095) (Figure 4.2b). Antibiotic treatment reduced bacterial load

in the A. salina approximately 4-fold to B/H = 0.025 (Student’s t-test, p = 0.029), which was comparable to the treated anemones at the end of the experiment (B/H = 0.021).

0.15 0.15 * • Treated 0.10 0.10

B/H ratio 0.05 B/H ratio 0.05

0.00 0.00 0 1 3 7 14 21 Untreated Treated (a) Day (b) A. salina A. salina

Figure 4.2: (a) Change in bacterial load in the antibiotic-treated anemones. For each datapoint, n = 6; (b) Bacterial load in the untreated and treated A. salina. For each bar, n = 3. Error bars ± 1 SEM. Asterisks indicate significant differences, α = 0.05.

There was a temporary non-significant decrease in bacterial load in the control anemones from Day 0 to Day 1 (Mann-Whitney, p = 0.563) (Figure 4.3). This may have occurred because Day 1 sampling occurred four days after feeding, compared to two-three days for all other sampling timepoints (Table 4.1). Bacterial load in mammals and insects decreases with time since feeding (Catterson et al. 2018; Sonoyama et al. 2009), and this may also occur in E. diaphana. Despite this anomaly, overall trends in the data were unambiguous, and a general increase in bacterial load in the control anemones suggested that the culture conditions favoured bacterial growth. Nevertheless, bacterial reduction was maintained in the treated anemones until the end of the experiment.

0.6 * * * 0.5 • Control 0.4 • Treated

0.3 B/H ratio 0.2

0.1

0.0 0 1 3 7 14 21 Day

Figure 4.3: Change in bacterial load in the control anemones. Data from the antibiotic-treated anemones, as shown in Figure 4.2a, is included for scale. For each datapoint, n = 6. Error bars ± 1 SEM. Asterisks indicate significant differences, α = 0.05.

4.3.3 Bacterial community data processing

Sequencing produced 3 747 373raw reads across the 72 anemone and 6 A. salina bacterial community analysis samples (minimum 3 825; mean 48 043, maximum 88 457 reads per sample). After merging, denoising and chimera filtering, 2 656 039 reads remained (minimum 2 417, mean 34 052, maximum 55 287 reads per sample) and 4 421 ASVs were identified. Rarefaction curves for all samples plateaued, indicating that sequencing captured bacterial diversity (Figure A 15). Decontam removed eight ASVs deemed contaminants, which constituted 0.076% and 0.005% relative abundance of the anemone and A. salina bacterial communities, respectively (Table A 26). An nMDS ordination (weighted unifrac) revealed two outlier samples, which were removed from the analysis (Figure A 16).

4.3.4 Bacterial community characterisation

There was no significant difference in the average number of bacterial ASVs observed in the control and antibiotic-treated anemones at Day 0 (Kolmogorov-Smirnov, p = 1.000) (Figure 4.4a). Although observed ASVs in the control anemones decreased after Day 0, this was temporary and there was no significant difference in observed ASVs for the control anemones between Day 0 and Day 21 (Kolmogorov-Smirnov, p = 0.931) despite the increase in bacterial load noted above. In contrast, there was a significant increase in observed ASVs in the 86

antibiotic-treated anemones despite their decrease in bacterial load (Kolmogorov-Smirnov, p = 0.004).

* * * * * 350 4.5 0.96 300 • Control Treated 4.0 250 • 0.92

200 Simpson Index 3.5 Observed ASVs 0.88 Sihannon Index 150

100 3.0 0 1 3 7 14 21 0 1 3 7 14 21 0 1 3 7 14 21 (a) Day (b) Day (c) Day

Figure 4.4: (a) Number of observed ASVs per anemone; (b) Simpson index values — higher values indicate higher evenness; (c) Shannon index values — higher values indicate higher overall alpha diversity. For each datapoint, n = 5–6. Error bars ±1 SEM. Asterisks indicate significant differences, α = 0.05.

A significant overall difference in evenness between the treated and control anemone bacterial communities was detected (#2 = 3.950, p = 0.047), but this was not supported by post hoc tests performed for each timepoint (Table A 33). There was a general increase in evenness in the treated anemones compared to the controls (Figure 4.4b), however this was not significant (Kolmogorov-Smirnov, p = 0.238). Despite this, at the end of the treatment period alpha diversity (Shannon index) was significantly higher in the treated anemones compared to the control anemones due to the increase in observed ASVs (Kolmogorov- Smirnov, p = 0.004) (Figure 4.4c).

Grouping of all Day 0 datapoints in an nMDS ordination of the bacterial community data suggested that the experimental anemones were highly similar at the beginning of the experiment and GLM-based analyses confirmed that they were not significantly different (LRT = 431.5, p = 0.200) (Figure 4.5a). However, by Day 21 the bacterial communities of the control and antibiotic-treated anemones had become significantly different from their Day 0 counterparts (control: LRT = 1 001, p = 0.006; treated: LRT = 1 312, p = 0.003) and each other (LRT = 1 432, p = 0.002). Tight clustering of the Day 21 control datapoints suggested that despite undergoing compositional shifts, the bacterial communities of the control anemones were still highly uniform after 21 days. In contrast, separation of the datapoints for the Day

21 treated anemones indicated decreased uniformity in their bacterial communities. This may be partially explained by the increase in observed ASVs for the treated anemones described in Figure 4.4a, above. A survey of common and unique ASVs in the Day 21 control and treated anemones explained this further by showing that, compared to the control anemones (Figure 4.5b), each treated anemone (Figure 4.5c) harboured a high number of unique ASVs. Together, these data indicated an increase in bacterial beta diversity among the antibiotic- treated anemones.

2D stress = 0.13 Day 21 control anemones Day 21 treated anemones

control 17 285 treated 27 33 166 249

NMDS2 Day 0 42 45 Day 21 69 45 272 172 60 168

(a) NMDS1 (b) Total ASVs = 439 (c) Total ASVs = 1513

Figure 4.5: (a) nMDS ordination (unweighted unifrac) of bacterial communities in control (n = 5–6) and antibiotic-treated anemones (n = 6) at Day 0 and Day 21; (b) Venn diagram showing numbers of common and unique ASVs in the Day 21 control anemones (n = 6); (c) Venn diagram showing numbers of common and unique ASVs in the Day 21 antibiotic-treated anemones (n = 6).

Increased variation (i.e., higher beta diversity) in the bacterial communities of the antibiotic- treated anemones compared to the control anemones was also evident in bar plots of family- level taxa (Figure 4.6). Bacteria of the Rhodobacteraceae and Alteromonadaceae generally predominated in the treated anemones at Day 21. However, a high fraction (24–46%) of the bacterial communities for each antibiotic-treated anemone consisted of low abundance ASVs (<1% per anemone) with low ubiquity from >150 families, creating high variability among the anemones.

100 100

75 75

50 50

25 25 Reads assigned to family (%) Reads assigned to family (%)

0 0 (a) Day 21 control anemones (b) Day 21 treated anemones

<1% Cyclobacteriaceae Pseudoalteromonadaceae Oligoflexaceae Cryomorphaceae Rhodobacteraceae Colwelliaceae Xanthomonadaceae Micrococcaceae Litoricolaceae Alteromonadaceae Bdellovibrionaceae Hyphomonadaceae Marinobacteraceae Tenderiaceae Terasakiellaceae Magnetospiraceae Nannocystaceae Enterobacteriaceae Coxiellaceae Stappiaceae Rhizobiaceae Flavobacteriaceae Spirochaetaceae unidentified Saprospiraceae Acanthopleuribacteraceae Methylophilaceae Microbacteriaceae

Figure 4.6: Proportions of family-level taxa in (a) six Day 21 control and (b) six Day 21 antibiotic-treated anemones. ASVs with <1% relative abundance are shown as a single group. Only families containing ASVs >1% relative abundance are listed in the legend.

4.3.5 Bacterial genera with high or low changes in abundance

A differential abundance analysis comparing the bacterial genera of the untreated Day 0 anemones and antibiotic-treated Day 21 anemones identified 16 genera with significant binary log fold changes (L2FC) (α = 0.01) and overall changes in relative abundance >0.5%

(Figure 4.7). This included nine genera with positive L2FC values, each with a relative abundance of 0.0–0.5% at Day 0 and up to 5.5% relative abundance at Day 21, and seven genera with negative L2FC values. The genera with negative L2FC values were collectively 38.5% of E. diaphana’s bacterial communities at Day 0 but were only 4.7% by Day 21. Ruegeria (9.1–0.1%) and an unidentified Oligoflexaceae genus (9.8–2.0%) experienced the highest reductions. Eleven genera were also associated with the antibiotic-treated A. salina feedstock.

Rel. ab. (%) 1 10 3 5 FC) 2 7 5 10

0 Binary log fold change (L

−5

PseudoalteromonasMicrococcusEscherichiaRalstoniaMicrobacteriumMethyloteneraunidentifiedThalassotalea ChitinophagalesTropicibacterCoxiellaLabrenziaNitratireductorunidentifiedSpirochaeta Oligoflunidentifiede 2 Ruegeria PB19

* * * * * * * * * xaceae Genus * *

Figure 4.7: Bacterial genera in the untreated Day 0 and antibiotic-treated Day 21 anemones with significant L2FC values (α = 0.01) and changes in relative abundance >0.5%. In the plot, genera that increased from Day 0 to Day 21 extend above 0, and genera that decreased from

Day 0 to Day 21 extend below 0. Line lengths indicate L2FC. Bubbles indicate relative abundance. Asterisks indicate genera also detected in the antibiotic-treated A. salina.

Nine bacterial genera with conserved relative abundances of ≥1.0% and changes in relative abundance of <2.0% after antibiotic treatment for three weeks were identified (Table 4.4). Due to their perseverance, these genera were considered ‘stable’ community members. All except an unidentified Terasakiellaceae genus were associated with the antibiotic-treated A. salina feedstock (Table 4.4).

Table 4.4: ‘Stable’ bacterial genera detected in the untreated Day 0 and antibiotic-treated Day 21 anemones with relative abundances >1.0% and changes in relative abundance <2.0%.

Phylum Class Order Family Genus Day 0 Day 21 Change A. salina % % % assoc.? Bacteroidetes Bacteroidia Chitinophagales Saprospiraceae unidentified 3.66 3.29 0.37 yes

Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Sedimentitalea 3.64 1.70 1.94 yes

Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Thalassobius 12.62 11.76 0.86 yes

Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Leisingera 2.31 1.68 0.63 yes

Proteobacteria Alphaproteobacteria Rhodospirillales Terasakiellaceae unidentified 3.55 2.14 1.40 no

Proteobacteria Gammaproteobacteria Alteromonadales Marinobacteraceae Marinobacter 1.87 2.27 -0.40 yes

Proteobacteria Gammaproteobacteria Pseudomonadales Pseudomonadaceae Pseudomonas 2.98 2.83 0.15 yes

Proteobacteria Gammaproteobacteria Vibrionales Vibrionaceae Vibrio 1.00 1.17 -0.17 yes

Proteobacteria Gammaproteobacteria Xanthomonadales Xanthomonadaceae Stenotrophomonas 1.78 3.19 -1.41 yes

4.4 Discussion

Antibiotic treatment has been used previously to remove (Wang & Douglas 1999) or deplete (Costa, Cárdenas & Voolstra 2019) bacteria from the coral model, E. diaphana, but with limited assessment of its efficacy and effect on the host, and no information about its impact on the bacterial associates. The present study sought to address these gaps by measuring the impact of antibiotic treatment on E. diaphana and its bacterial communities.

4.4.1 Antibiotic exposure reduces Symbiodiniaceae density

All antibiotic-treated E. diaphana survived treatment. However, their Symbiodiniaceae densities declined 48 hours after the first dose, indicating a stress response by the anemones. This may have been induced by the loss of bacteria required for critical functions, such as nutrient cycling, or by toxicity of the antibiotics to the anemones or Symbiodiniaceae. For example, chloramphenicol is lethal to Symbiodiniaceae at 50 µg/mL (Soffer, Gibbs & Baker 2008), and whilst used in the present study at only 25 µg/mL, may have been detrimental. Despite this, a doubling of Symbiodiniaceae density in the treated anemones from Day 14 to Day 21 demonstrated a recovery of the E. diaphana-algae symbiosis during treatment.

The increase in Symbiodiniaceae density may have been due to acclimation of the holobiont to antibiotic treatment via selection for resilient types or subtypes of Symbiodiniaceae. However, this assumes the presence of multiple Symbiodiniaceae types or subtypes, which

was not investigated. It would also require in hospite Symbiodiniaceae growth rates higher than have been previously recorded in either stressed (Fransolet et al. 2013) or non-stressed E. diaphana (Wilkerson, Muller & Muscatine 1983). Nevertheless, the partial recovery of Symbiodiniaceae density suggests that the antibiotics did not severely harm the anemones, and that prolonged treatment could be viable.

4.4.2 Antibiotic treatment partially eliminates bacteria

A rapid and significant reduction in bacterial load in the treated anemones and A. salina demonstrated the effectiveness of the antibiotic treatment. Similarly rapid declines in prokaryotic cell abundance have been reported in antibiotic-treated coral mucus (Glasl, Herndl & Frade 2016). However, according to the ddPCR assay, bacteria were not completely eliminated from the anemones or A. salina. It is possible that the amplified DNA originated from dead bacteria as DNA can persist long after cell death (Salo et al. 1994; Young et al. 2007), particularly if cells are intact (Brundin et al. 2010). Despite this, marine invertebrate DNA has been shown to degrade to levels below detection by ddPCR in ≤ 94 hr (Wood et al. 2020). Furthermore, frequent water changes were performed on the anemones, which would have removed free DNA and dead bacteria. Using culture techniques to test treated anemones and A. salina as done previously (Costa, Cárdenas & Voolstra 2019; Wang & Douglas 1999) could show whether the ddPCR signal originated from viable bacteria, although the absence of growth would not be conclusive due to the uncultivability of many bacteria (Stewart 2012).

Another explanation for the survival of some bacteria, is the formation of bacterial biofilms on the culture plate surfaces, which may have protected bacteria (Stewart 2002) or free DNA (Wood et al. 2020). These may have formed because the culture plates were not cleaned to minimise the risk of introducing bacteria or stressing the anemones, which may also explain the proliferation of bacteria in the control samples. In future, it may be advisable to transfer the anemones to new plates periodically to avoid this.

A further possibility is that bacteria were protected from antibiotics by their envelopment as aggregates within E. diaphana tissue. Palincsar et al. (1989) found that treatment of E. diaphana with high doses of chloramphenicol (125 mg/mL) or streptomycin (25 mg/mL) over three weeks reduced aggregates by ~90% and ~50%, respectively. This incomplete

elimination of aggregates despite aggressive antibiotic treatment suggests that generating germfree anemones by antibiotics may require prolonged treatment or may not be feasible.

The presence of Symbiodiniaceae could also confound efforts to generate germfree E. diaphana as Symbiodiniaceae are thought to contain intracellular bacteria (Ainsworth et al. 2015). In addition, the anemones used in the present study harboured Symbiodiniaceae of the genus Breviolum (Dungan et al. 2020), which have defied a previous attempt to render them germfree with antibiotics (Xiang et al. 2013). If Symbiodiniaceae provide their intracellular bacteria with protection against antibiotics, the use of aposymbiotic E. diaphana may improve bacterial depletion.

4.4.3 Antibiotic treatment increases bacterial diversity

If germfree E. diaphana are not achievable, microbiologically standardised gnotobiotic cultures with low bacterial loads and diversity would be highly valuable (Gordon & Pesti 1972). However, the bacteria in the antibiotic-treated E. diaphana underwent significant increases in alpha and beta diversity despite reductions in bacterial load. The increase in bacterial richness in the treated anemones points to a large, diverse pool of bacteria in E. diaphana that are normally below detection. The relative increase of these bacteria suggests they were held in check by different mechanisms, including competition with previously abundant bacteria, and had higher antibiotic resistance than those they superseded. The high number of unique ASVs detected in each treated anemone also suggests high variation between each pool. These traits will complicate efforts to generate E. diaphana with simplified and uniform bacterial communities, however it may be possible to limit variation by generating cultures from single, founder anemones.

4.4.4 Antibiotic treatment promotes the growth of some bacterial genera

Pseudoalteromonas underwent the highest binary log fold increase in the antibiotic-treated anemones (L2FC = 12.6). Members of Pseudoalteromonas have been noted for their resistance against three of the antibiotics used in the present study (chloramphenicol, nalidixic acid and streptomycin) (Dang et al. 2007), involvement in biofilm formation (Saravanan et al. 2006), and ability to use starch produced by dinoflagellates as a carbon source (Gobet et al. 2018). 93

Therefore, the culture conditions coupled with reduced bacterial competition in the treated anemones may have promoted Pseudoalteromonas growth. The antimicrobial activity of coral-derived Pseudoalteromonas species is also well-documented (Kvennefors et al. 2011; Nissimov, Rosenberg & Munn 2009; Shnit-Orland, Sivan & Kushmaro 2012). Although this feature did not reduce the total number of bacterial species, it may have helped them displace other community members.

Although the binary log fold change for Tropicibacter was comparatively low (L2FC = 3.7), their increase in relative abundance was substantial (0.5–5.5%). Some members of Tropicibacter also grow in association with dinoflagellates and have resistance against three of the antibiotics used in the present study (carbenicillin, chloramphenicol and streptomycin) (Wang et al. 2019). These results suggest that antibiotic treatment provided Pseudoalteromonas and Tropicibacter with an advantage, thus illustrating the limitation of antibiotic use against some bacteria.

Almost all genera with significant, positive L2FC values or stable relative abundances were also A. salina associates, thus implicating A. salina as the source, particularly as they were only partially sterilised. Correlations between the microbiomes of E. diaphana and their A. salina feedstock have been previously observed (Chapter 2, Table 2.5), which supports the hypothesis that many E. diaphana bacteria originate from A. salina. To avoid introducing bacteria that are antibiotic resistant or capable of forming biofilms, complete sterilisation of A. salina is necessary. Hatching A. salina in the antibiotic solution did not achieve this, indicating that stronger treatment of A. salina cysts is required using higher antibiotic concentrations or, for example, the chemical sterilisation method described by Sorgeloos et al. (1977).

4.4.5 Recommendations for improved bacterial depletion in E. diaphana

As noted, bacterial depletion could be increased by improving culture vessel cleanliness, proper feedstock sterilisation, prolonged antibiotic treatment, or by using aposymbiotic E. diaphana. Generating populations from antibiotic-treated E. diaphana pedal lacerates, cell aggregates (i.e., artificial lacerates), or immature rather than mature anemones, might also lower bacterial diversity as smaller treatment subjects might carry fewer bacteria and have

no, or few, bacterial aggregates. Closing the cycle of sexual reproduction in lab-reared E. diaphana has so far proven elusive (Grawunder et al. 2015). However, if this could be achieved, it would enable sterilisation of fertilised eggs or larvae, as performed on other organisms (Forberg & Milligan-Myhre 2017; Leigh, Liberti & Dishaw 2016; Ridley, Wong & Douglas 2013), and could provide a way forward for generating germfree or gnotobiotic E. diaphana.

4.5 Conclusion

Antibiotic exposure for three weeks significantly reduces the bacterial load of E. diaphana. However, treatment also increases the complexity and variability of the resulting bacterial communities. Consequently, antibiotic-treated E. diaphana lack bacterial simplicity and uniformity, so cannot be defined as gnotobiotic. Although treatment induces a bleaching response in the anemones, this is not lethal, and recovery of the E. diaphana-Symbiodiniaceae symbiosis suggests that prolonged treatment is viable. Extended antibiotic treatment could improve bacterial depletion, simplicity and uniformity, providing culture vessels and food are sterile. However, the practicality of using antibiotics could ultimately be limited by the diversity of native bacteria, some of which likely possess antibiotic resistance. Therefore, using treatment subjects with naïve bacterial communities might be necessary for generating germfree or gnotobiotic E. diaphana which, if produced, would represent a major leap forward in cnidarian symbiosis research.

Chapter 5: Assessment of a ROS-targeted bacterial probiotic designed to improve thermal tolerance in the sea anemone, Exaiptasia diaphana

This chapter describes work performed jointly by L. Hartman and A. Dungan.

L. Hartman, A. Dungan, M. van Oppen and L. Blackall designed the experiment. L. Hartman and A. Dungan performed the experimental work, analysed the data, wrote and edited the text. M. van Oppen and L. Blackall reviewed and edited the text.

5.1 Introduction

The persistence of coral reefs is threatened by environmental conditions caused by climate change, particularly elevated sea surface temperature (SST) and high irradiance (Hughes et al. 2017a). Coral’s obligate symbiosis with dinoflagellate algae of the family Symbiodiniaceae lies at the heart of this problem. The interaction of elevated temperature and light impacts Symbiodiniaceae through photoinhibition of photosynthesis and photodamage to photosystem II (PSII) (Tchernov et al. 2004; Warner, Fitt & Schmidt 1999). This leads to increased levels of reactive oxygen species (ROS) in Symbiodiniaceae, which are thought to diffuse into host cells, causing cellular damage (Downs et al. 2002; Lesser 1997). In a subsequent act of mutual self-preservation, the Symbiodiniaceae and host cells dissociate in a process known as ‘bleaching’ (Weis 2008). As corals receive most of their carbon requirements from Symbiodiniaceae (Muscatine, Falkowski & Dubinsky 1984), this separation is fatal unless environmental conditions ease and the coral-algae symbiosis is reinstated. Thus, with average SST predicted to rise further (Aral & Guan 2016), finding effective bleaching mitigation methods has become crucial.

Many innovative approaches have been proposed to mitigate bleaching and enhance coral survival during thermal stress (Peixoto et al. 2017; Rinkevich 2019; van Oppen et al. 2015). However, directly targeting ROS has rarely been investigated despite evidence that exposure to exogenous antioxidants can reduce host ROS levels and bleaching in thermally stressed corals and coral models (Lesser 1997; Marty-Rivera, Yudowski & Roberson 2018). In addition, the photosynthetic performance of thermally stressed Symbiodiniaceae has been shown to improve within hours of receiving a single antioxidant dose (Lesser 1996).

Although these results were promising, applying antioxidant compounds directly onto corals may not be practical due to the short lifespan of antioxidants when exposed to seawater (King et al. 2016) and ultraviolet radiation (Compton et al. 2019), or desirable due to their potential impact on non-target organisms. Instead, increasing natural antioxidant generation within the coral holobiont may be more effective and prudent. How might this be achieved?

Corals are colonized by a multitude of microorganisms that profoundly influence the animal’s health, for example by cycling nutrients (Lesser et al. 2007; Rädecker et al. 2015; Raina et al. 2009) and protecting hosts from pathogens (Höhener, Huap & Müller 2016; Kuek et al. 2015;

Shnit-Orland & Kushmaro 2009). Research has shown that cnidarian-associated bacterial communities can be supplemented by probiotic inoculation to protect or support hosts challenged by pathogenic (Alagely et al. 2011) or environmental stressors such as oil pollution (dos Santos et al. 2015). Inoculation of coral with native bacteria, including catalase positive isolates, has also produced improved bleaching tolerance in corals exposed to elevated temperature (Rosado et al. 2018). Although the results of this work were promising, the reason for improved bleaching tolerance was unclear due to the selected bacteria’s broad range of traits. For example, the stressed corals may have also been supported indirectly, through heterotrophic feeding on the introduced bacteria (Meunier et al. 2019; Tremblay et al. 2012b). Consequently, the influence of antioxidant-producing bacteria on cnidarian bleaching requires further investigation.

To address this, we inoculated with a probiotic consisting of free radical scavenging (FRS) bacteria to test whether this inoculum reduced ROS and mitigated bleaching in a coral model, the sea anemone Exaiptasia diaphana (previously Aiptasia pallida (Grajales & Rodríguez 2014; ICZN 2017)) when exposed to thermal stress. Holobiont response to thermal stress was assessed by measuring Symbiodiniaceae photosynthetic performance and cell densities. Holobiont ROS levels were measured to determine whether ROS was reduced by inoculation during a simulated thermal stress event. Metabarcoding of bacterial 16S rRNA genes was used to track incorporation of the probiotic bacteria into the E. diaphana microbiome and changes in bacterial community structure across time and treatments. The influence of host genotype was also explored by comparing the responses of three E. diaphana genotypes. This work extends probiotic-based bleaching mitigation research and expands knowledge of E. diaphana, revealing its potential as a coral bleaching model when coral assisted evolution research is becoming more relevant and urgent.

5.2 Materials and methods

5.2.1 Probiotic preparation

Six Exaiptasia-sourced bacterial isolates with high free radical scavenging (FRS) ability were selected for inclusion in a ‘positive’ probiotic cocktail (see Appendix 4 for probiotic development methodology). Six closely related strains with poor FRS ability were also selected

for inclusion in a ‘negative control’ probiotic cocktail. The bacteria were identified from their full length 16S rRNA gene sequences by Sanger sequencing (Table 5.1). According to these data, only the Micrococcus positive and negative FRS bacteria were distinct at the species level.

Table 5.1: Bacteria used in the positive and negative probiotics. Family Genus Species Positive Strain ID/ Negative Strain ID/ NCBI Accession NCBI Accession Number Number Alteromonadaceae Alteromonas oceani MMSF01163/ MMSF00404/ MN540711 MN540719 Alteromonadaceae Alteromonas macleodii MMSF00958/ MMSF00257/ MN540717 MN540713 Alteromonadaceae Marinobacter salsuginis MMSF01190/ MMSF00964/ MN540716 MN540714 Flavobacteriaceae Winogradskyella poriferorum MMSF00046/ MMSF00910/ MN540715 MN540718 Micrococcaceae Micrococcus luteus (+), MMSF00068/ MMSF00107/ yunnanensis (–) MN540712 MN540722 Rhodobacteraceae Labrenzia aggregata MMSF00132/ MMSF00249/ MN540721 MN540720

To prepare the probiotics, isolates were grown from preserved cells in 50 mL of supplemented Reasoner's 2A (R2A) broth (Table A 40) made with reconstituted Red Sea Salt™ seawater (R11065, Red Sea, USA), hereafter ‘RSS’. The cells were incubated at 37 °C for 48 hrs at 150 rpm in an orbital incubator (OM11, Ratek, Australia). Uninoculated broth was also incubated to confirm media sterility. Three replicate cultures were grown per isolate. After 48 hrs, the culture and media blanks were assessed for cell concentration by OD600 measurement (CLARIOstar PLUS plate-reader, BMG Labtech, Australia), and centrifuged at 3000 × g at 4 °C for 30 min to pellet the bacteria. The pelleted cells were washed three times by resuspending in 10 mL 0.2 µm-filtered RSS seawater reconstituted at ~34 parts per thousand (ppt), hereafter ‘fRSS’, and centrifuging at 3000 × g at 4 °C for 15 min. After the final centrifugation, the pellet was resuspended in 5 mL fRSS and the three replicate cultures per isolate were combined. OD600 measurements were taken of the pooled triplicates at a 1:10 dilution in fRSS, the total number of cells per isolate was calculated based on E. coli values (Volkmer & Heinemann 2011), and the pools were centrifuged and resuspended to a density of 109 cells mL–1. The bacteria at 109 cells mL–1 were pooled by probiotic treatment so that 1 mL of each pool would produce a final concentration of 106 bacterial cells mL–1 if added to 300 mL of RSS. 99

Cell densities of the pre-pooled isolates were confirmed by counting colony forming units (CFU). To assess bacterial cell viability, three replicate plates of Marine Agar (Difco™ 2216, Thermo Fisher, Australia) were spread inoculated with 50 µL of four serial dilutions (10–5 to 10–8) of each isolate. The plates were then incubated at 26 °C for seven days before CFU were counted.

5.2.2 Experimental set-up

Anemones from populations of three Great Barrier Reef (GBR)-sourced E. diaphana genotypes, AIMS2, AIMS3 and AIMS4, were randomly selected from The University of Melbourne culture collection (n = 450 per genotype) (Dungan et al. 2020). Anemones from each genotype were equally distributed among 18 × 300 mL lidded glass culture jars (total jars = 54) and the jars were evenly split between two experimental incubators (Hi-Point 740FHC, Thermo Fisher, Australia) fitted with red, white, and infrared light emitting diode (LED) lights. All anemones were maintained in RSS at ~34 ppt and fed ad libitum twice weekly with freshly hatched Artemia salina (Salt Creek, Premium GSL, USA). Jars were cleaned weekly by loosening algal debris with seawater pressure applied through sterile plastic pipettes followed by full RSS changes. All anemones were transferred to clean jars twice during the study period to combat algal growth. Seawater temperatures were monitored using submersible data loggers (Hobo UA-001-08, OneTemp, Australia). The anemones were acclimated to the experimental incubators for three weeks. Initial light levels were 12 μmol photons m–2 s–1 to match the stock culture conditions, then gradually increased to 28 µmol photons m–2 s–1 over 72 h on a 12 h:12 h light:dark cycle. Light levels were selected to correspond with previous studies evaluating bleaching via thermal stress on E. diaphana (Bieri et al. 2016; Tolleter et al. 2013), whilst ensuring that elevated temperature was the only stressor acting on the anemones as the AIMS anemones have demonstrated sensitivity to high irradiance (Dungan et al. 2020).

5.2.3 Treatment schedule

The experiment began after the three-week acclimation period at a timepoint designated Day 0 (Figure 5.1). On Day 0, probiotic-treated anemones were inoculated by addition of 1 mL

100

of the negative or positive cocktail to their rearing water, and no-inoculum control anemones were not inoculated. The probiotic-treated anemones were inoculated again on Days 2 and 7. All anemones were exposed for 43 days to ambient temperature (26 °C), or at temperature increasing from ambient to 31.5 °C between Day 8 and Day 31 (0.25 °C/day), then held at 31.5 °C. The temperature ramp rate was designed to approximate GBR summer heatwave conditions (Figure A 18 (AIMS 2019)). The experimental design included three biological replicates per genotype for each inoculation-temperature combination. The temperature designation of the incubators was swapped fortnightly and the anemones were randomly rearranged each week to remove incubator and jar position as confounding factors.

3 × 3 × 31.5°C

3 × 3 × 26°C 26°C 3 × 3 × –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43

3 × 3 × 31.5°C

3 × 3 × 26°C 26°C 3 × 3 × –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43

3 × 3 × 31.5°C

3 × 3 × 26°C 26°C 3 × 3 × –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 (a) Day (b) Day

Figure 5.1: Inoculation and sampling schedule for each inoculation-temperature combination. Temperature conditions: (a) ambient (26 °C), and (b) elevated (26 °C–31.5 °C). Positive probiotic inoculations (blue pipettors) and negative probiotic inoculations (pink pipettors) were performed on Days 0, 2 and 7. Stars indicate collection of one anemone from each replicate jar (n = 3) for three E. diaphana genotypes (AIMS2, AIMS3, AIMS4). At Day –21, nine anemones per genotype were collected from stock aquaria for baseline data collection (n = 27). Total anemone samples, n = 783.

5.2.4 Symbiodiniaceae photosynthetic efficiency measurement

Photosynthetic efficiency, according to maximum quantum yield (Fv/Fm) of PSII in

Symbiodiniaceae, is a commonly-used proxy for general holobiont health as lowered Fv/Fm indicates PSII damage resulting from thermal stress or high irradiance (Ralph et al. 2016).

101

Fv/Fm was measured on the intracellular Symbiodiniaceae by pulse-amplitude-modulated (PAM) fluorometry with an imaging-PAM system (IMAG-MAX/L, Walz Heinz, Germany). On sampling days, measurements were taken from the bodies and proximal tentacles of five anemones per jar at 10:00 a.m. (4 hr after daylight cycle start) after 30 min dark adaptation. PAM settings were: saturating pulse intensity 8, measuring light intensity 2 (frequency 1), actinic light intensity 3, damping 2, and gain 2.

5.2.5 Anemone tissue processing

On sampling days, one anemone per jar was sacrificed for Symbiodiniaceae cell density, ROS and protein quantification, and DNA extraction for bacterial community analysis. Each anemone was individually homogenized in a sterile glass homogenizer in 1 mL of fRSS and an aliquot of 250 µL was flash frozen in liquid nitrogen for DNA extraction. The remaining homogenate was centrifuged at 5000 × g for 5 min at 4 °C to pellet the Symbiodiniaceae, and 50 µL of the supernatant was stored at –20 °C for anemone (host) protein measurement, while 600 µL of supernatant was taken for ROS quantification.

5.2.6 Symbiodiniaceae cell density measurement

Pelleted Symbiodiniaceae were washed twice with 700 µL fRSS and centrifuged at 5000 × g for 5 min at 4 °C, then resuspended in 700 µL fRSS. Triplicate cell counts (cells mL–1) were performed on 10 µL of sample within 48 hrs of sample collection with an automated cell counter (Countess II FL, Life Technologies, Australia). Cell counts were normalized to host protein (mg mL–1) determined by Bradford assay (Bradford 1976). Each protein sample was assayed in triplicate and measured at 595 nm (EnSpire MLD2300 plate-reader, Perkin Elmer, Australia) against bovine serum albumin standards (500-0207, Bio-Rad, Australia).

5.2.7 ROS assay

Anemone ROS levels were quantified with CellROX® Orange (C10443, Thermo Fisher, Australia); a ROS-quantifying reagent used in previous cnidarian studies (Chakravarti, Beltran & Oppen 2017; Gegner et al. 2019; Levin et al. 2016; Marty-Rivera, Yudowski & Roberson

102

2018). For each sample, 594 µL of supernatant, collected as above, was vortexed briefly with 6 µL CellROX® Orange (final concentration 5 µM) and incubated in darkness at 37 °C for 30 min. Three 200 µL technical replicates of each sample were transferred into a 96-well black, clear-bottom culture plate (Nunc 165305, Thermo Fisher, Australia). Fluorescence readings at excitation 545 nm, emission 565 nm (EnSpire MLD2300, Perkin Elmer, Australia) were then taken and the average reading per sample was normalized to raw Symbiodiniaceae cell counts to give relative ROS values in arbitrary fluorescent units.

5.2.8 Metabarcoding sample preparation

DNA was extracted from the anemones for bacterial community analysis by metabarcoding of the 16S rRNA genes. Anemone tissue samples were collected as above. DNA was extracted according to Wilson et al. (2002) with modifications described by Hartman, van Oppen and Blackall (2019). Sample-free negative controls (n = 12) were also processed to identify contaminants introduced during DNA extractions. Extracted DNA and negative controls were amplified by PCR in triplicate using bacterial primers with Illumina Nextera adapters (underlined) targeting the V5-V6 regions of the 16S rRNA genes: 784F [5ʹ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAGGATTAGATACCCTGGTA 3ʹ]; 1061R [5ʹ

GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCRRCACGAGCTGACGAC 3ʹ] (Andersson et al. 2008). Each PCR reaction contained 0.5U MyTaq HS Mix polymerase (Bioline, Australia), 2 µL of DNA template, 0.4 µM of each primer, and nuclease-free water up to 25 µL. PCR conditions were: 1 cycle × 95 °C for 3 min; 18 cycles × 95 °C for 15 s, 55 °C for 30 s, and 72 °C for 30 s; 1 cycle × 72°C for 7 min; hold at 4 °C. The triplicate PCR products from each sample were then pooled.

To prepare DNA sequencing libraries, 20 µL of each PCR product pool was purified by size- selection using Ampure XP magnetic beads (Agencourt, Beckman Coulter, Australia). The purified DNA was resuspended in 40 µL nuclease-free water. Indexing PCR reactions were created by combining 10 µL of each DNA suspension with 10 µL 2x Taq master mix (M0270S, New England BioLabs, Australia) and 0.25 µM of forward and reverse indexing primers. PCR conditions were: 1 cycle × 95 °C for 3 min; 24 cycles × 95 °C for 15 s, 60 °C for 30 s, and 72 °C for 30 s; 1 cycle × 72°C for 7 min; hold 4 °C. Template-free negative controls (n = 6) were included to identify contaminants introduced during indexing PCR. Four representative 103

samples were checked for product size and quantity (2200 TapeStation, Agilent Technologies, Australia). Sequencing libraries were created by pooling 5 µL from each reaction and performing a final bead clean-up on 50 µL. Each library was checked for quality and quantity to guide pool normalization (2200 TapeStation), then sequenced across three Illumina Mi-Seq runs using v3 (2 × 300 bp) reagents at the Walter and Eliza Hall Institute (WEHI), Melbourne, Australia. The Illumina Mi-Seq data were deposited under NCBI BioProject ID PRJNA630329.

5.2.9 Metabarcoding data processing

Raw 16S rRNA gene sequences were imported into QIIME2 v2019.10.0 (Bolyen et al. 2019) and demultiplexed on a per-sequencing-run basis. Primers were removed with cutadapt v2.6 (Martin 2011). Filtering, denoising, and chimera checking was performed using DADA2 (Callahan et al. 2016) in the QIIME2 environment to correct sequencing errors, remove low quality bases (mean Qscore <30), and generate bacterial amplicon sequence variants (ASVs). The data from each sequencing run were then merged, and taxonomy for each ASV was assigned against a SILVA database (version 132) trained with a naïve Bayes classifier against the same V5-V6 region targeted for sequencing (Bokulich et al. 2018).

5.2.10 Data analysis

All data were analysed in R v3.6.2 (R Core Team 2018). Statistical tests were considered significant at α = 0.05, unless otherwise stated. ASV, taxonomy, metadata and phylogenetic tree files were imported into R and combined into a phyloseq object (McMurdie & Holmes 2013). Potential contaminant ASVs were identified and removed from the dataset according to their abundance in the extraction (n = 12) and PCR (n = 12) negative controls relative to the anemone samples using the prevalence method in the R package decontam with p = 0.1 (Davis et al. 2018). Dark-adapted Fv/Fm, Symbiodiniaceae cell density, and ROS measurements were plotted using the R package ggplot2 (Wickham 2019) with data separated by genotype and temperature condition. The data were analysed for overall differences using linear mixed effects models (LME) against the variables ‘probiotic’ and ‘time’, with ‘jar’ specified as a random effect, in the R package nlme (Pinheiro et al. 2019). Post hoc pairwise comparisons were performed using Tukey’s HSD (Tukey 1949) in the R package emmeans (Searle, Speed &

104 Milliken 1980) with Tukey’s adjustment for multiple comparisons. Shifts in the bacterial community compositions of each genotype were visualised in principal coordinate analysis (PCoA) ordinations based on Jaccard distance (Jaccard 1912). The influence of time (Day 0 to Day 43 only), probiotic and temperature treatments on changes in community composition within each genotype was investigated by comparison of generalised linear models (GLM) for each ASV using likelihood ratio tests across 999 iterations in the R package mvabund (Wang et al. 2012). For PCoAs and GLMs, replicate samples were merged, and the data were collapsed to genus for clarity and computational efficiency. Incorporation of the probiotic bacteria into the anemone microbiomes was visualised in bubble plots at the ASV level for the probiotic members in each genotype, temperature and probiotic condition over time. Further tests with GLMs were performed, as above, to assess whether differences between the bacterial communities for each genotype-temperature-probiotic combination were significant at selected timepoints.

105

5.3 Results

5.3.1 CFU data

Cell densities of probiotic bacteria calculated from CFU were generally within one order of magnitude of the target density (109 cells mL–1). Exceptions included the Day 7 Winogradskyella (negative probiotic) culture, which failed to produce colonies, the Day 2 Micrococcus and A. oceani, and Day 7 A. oceani (positive probiotic) cultures, which were of low CFU density (Table 5.2).

Table 5.2: Cell densities calculated from CFU counts of bacteria cultured for probiotic dosing on Days 0, 2, and 7. Bacterial cultures <108 cells mL–1 are in bold. Positive Strain Dosing Density Negative Strain Dosing Day Density Day (cells mL–1) (cells mL–1) MMSF00046 0 6.40 × 107 MMSF00910 0 1.56 × 109 Winogradskyella 2 2.34 × 108 Winogradskyella 2 3.27 × 109 7 1.61 × 109 7 No growth MMSF00068 0 3.72 × 108 MMSF00107 0 1.33 × 109 Micrococcus 2 3.80 × 107 Micrococcus 2 2.37 × 109 7 1.29 × 109 7 1.53 × 109 MMSF00132 0 2.23 × 109 MMSF00249 0 6.73 × 109 Labrenzia 2 2.60 × 108 Labrenzia 2 6.00 × 108 7 1.30 × 108 7 7.73 × 108 MMSF00958 0 2.19 × 108 MMSF00257 0 4.80 × 109 Alteromonas 2 1.43 × 108 Alteromonas 2 2.53 × 109 macleodii 7 4.37 × 108 macleodii 7 2.73 × 109 MMSF01163 0 1.01 × 109 MMSF00404 0 8.67 × 108 Alteromonas 2 2.33 × 107 Alteromonas 2 3.87 × 108 oceani 7 2.53 × 107 oceani 7 4.07 × 109 MMSF01190 0 1.87 × 109 MMSF00964 0 9.27 × 108 Marinobacter 2 2.45 × 108 Marinobacter 2 4.67 × 108 7 3.39 × 109 7 1.32 × 109

106

5.3.2 Photosynthetic efficiency of Symbiodiniaceae

Dark-adapted Fv/Fm of Symbiodiniaceae declined during the three-week acclimation period in all anemones, potentially demonstrating the impact of relocation on Symbiodiniaceae health (Figure 5.2). However, almost all cultures recovered to baseline levels by Day 12, suggesting a true acclimation period of approximately five weeks. All anemones at ambient temperature

(26 °C) maintained near-normal Fv/Fm levels after Day 12 but experienced a spike in Fv/Fm at

Day 31 compared to Day 28. As there were no changes in Fv/Fm of similar magnitude on adjacent days, the Day 31 values may represent a measurement error.

0.6 AIMS2 : Ambient Temperature 0.6 AIMS2 : Elevated Temperature Probiotic 0.5 0.5 treatment

m m Control

/ F 0.4 / F 0.4 (no inoculum) v v F F Negative 0.3 0.3 Postive 0.2 0.2 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

0.6 AIMS3 : Ambient Temperature 0.6 AIMS3 : Elevated Temperature

0.5 0.5 m m

/ F 0.4 / F 0.4 v v F F

0.3 0.3

0.2 0.2 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

0.6 AIMS4 : Ambient Temperature 0.6 AIMS4 : Elevated Temperature

0.5 0.5 m m

/ F 0.4 / F 0.4 v v F F

0.3 0.3

0.2 0.2 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

Figure 5.2: Dark-adapted Fv/Fm at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all experimental treatments within each genotype. Probiotic inoculations were performed at Days 0, 1 and 7. For each datapoint, n = 3. Error bars ± 1 SEM.

Fv/Fm for the anemones in each genotype at ambient or elevated temperature differed significantly according to time, but not with treatment (Table 5.3), suggesting that probiotic treatment did not influence changes in Fv/Fm. Pairwise comparisons showed that at Day 43

Fv/Fm was significantly lower in the anemones of each genotype at elevated temperature 107

compared to those at ambient temperature (Table A 41, Table A 42, Table A 43). Hence, declines in Fv/Fm were not mitigated by probiotic treatment. There were also no significant differences in Fv/Fm among the positive probiotic, negative probiotic and control anemones at elevated temperature, confirming that probiotic treatment had no influence on Fv/Fm values.

Table 5.3: ANOVA output for Fv/Fm in each genotype. Significant values are in bold; α = 0.05. Genotype Temperature Main effect df F-value p-value AIMS2 Ambient Probiotic 2, 6 1.271 0.3466 Time 13, 77 16.493 <0.0001 Elevated Probiotic 2, 6 0.580 0.5866 Time 13, 78 15.210 <0.0001 AIMS3 Ambient Probiotic 2, 6 6.370 0.0328 Time 13, 77 48.010 <0.0001 Elevated Probiotic 2, 6 0.087 0.9175 Time 13, 76 18.132 <0.0001 AIMS4 Ambient Probiotic 2, 6 2.700 0.1458 Time 13, 77 7.546 <0.0001 Elevated Probiotic 2, 6 0.540 0.6068 Time 13, 77 12.670 <0.0001

108

5.3.3 Symbiodiniaceae cell density

Symbiodiniaceae cell density declined among most anemones during the acclimation period (Figure 5.3). Once again, this may indicate stress induced by relocation and exposure to increased light intensity. The AIMS3 anemones underwent the greatest decline with a 52% loss of Symbiodiniaceae, on average, whilst AIMS2 and AIMS4 lost 31% and 38%, respectively. However, the higher reduction in AIMS3 may have been due to their relatively high Symbiodiniaceae density at Day –21, rather than sensitivity to relocation as all anemones had similar Symbiodiniaceae cell densities of 0.5–1.0 x 107 cells mg–1 host protein at Day 0.

2.0e+07 AIMS2 : Ambient Temperature 2.0e+07 AIMS2 : Elevated Temperature Probiotic 1.5e+07 1.5e+07 treatment Control 1.0e+07 1.0e+07 (no inoculum)

Negative 0.5e+07 0.5e+07 Postive 0.0e+07 0.0e+07 Symbiodiniaceae cells / mg protein Symbiodiniaceae cells / mg protein −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

2.0e+07 AIMS3 : Ambient Temperature 2.0e+07 AIMS3 : Elevated Temperature

1.5e+07 1.5e+07

1.0e+07 1.0e+07

0.5e+07 0.5e+07

0.0e+07 0.0e+07 Symbiodiniaceae cells / mg protein Symbiodiniaceae cells / mg protein −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

2.0e+07 AIMS4 : Ambient Temperature 2.0e+07 AIMS4 : Elevated Temperature

1.5e+07 1.5e+07

1.0e+07 1.0e+07

0.5e+07 0.5e+07

Figure 5.3: Symbiodiniaceae cell density at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all treatments within each genotype. Probiotic inoculations were performed at Days 0, 1 and 7. For each datapoint, n = 3. Error bars ± 1 SEM.

Symbiodiniaceae cell density for the anemones in each genotype at ambient or elevated temperature differed significantly according to time, but not with treatment (Table 5.4), suggesting that probiotic treatment did not influence changes in cell density. Pairwise comparisons showed that at Day 43, Symbiodiniaceae cell density was not significantly 109

different in the AIMS2 anemones at elevated temperature compared to those at ambient temperature, indicating that the AIMS2 anemones did not bleach (Table A 44, Table A 45, Table A 46). However, cell density was significantly lower in AIMS3 anemones, regardless of probiotic treatment. Although the AIMS4 comparisons showed a significant reduction in cell density only for the negative probiotic anemones, all heat stressed AIMS4 anemones appeared to undergo substantial declines (Figure 5.3). Therefore, the lack of statistical significance may have been due to high variance. Despite differences in response to elevated temperature, at Day 43 and there were no significant differences in Symbiodiniaceae cell density within each genotype based on probiotic treatment, confirming that probiotic treatment did not influence changes in cell density.

Table 5.4: ANOVA output for Symbiodiniaceae cell density in each genotype. Significant values are in bold; α = 0.05. Genotype Temperature Variable df F-value p-value AIMS2 Ambient Probiotic 2, 6 1.7177 0.2571 Time 13, 78 8.6399 <0.0001 Elevated Probiotic 2, 6 0.2345 0.7979 Time 13, 78 6.9984 <0.0001 AIMS3 Ambient Probiotic 2, 6 2.056 0.2088 Time 13, 77 13.933 <0.0001 Elevated Probiotic 2, 6 5.3142 0.0470 Time 13, 78 29.7595 <0.0001 AIMS4 Ambient Probiotic 2, 6 2.2639 0.1851 Time 13, 78 5.9935 <0.0001 Elevated Probiotic 2, 6 2.3981 0.7171 Time 13, 78 11.2786 <0.0001

110

5.3.4 ROS assay

There was no clear difference in ROS levels between the anemones based on temperature or treatment (Figure 5.4). However, trends related to treatment may have been obscured by high variation in the data. Fluctuations in ROS levels in all anemones may indicate stress induced by frequent changes in light and temperature due to handling. Fluctuations may also be due to the ephemeral nature of ROS (Diaz et al. 2016), or the low concentrations of ROS in the anemones, which may have been near the CellROX® assay’s limit of detection. Universal spikes in ROS at some timepoints, for example Day 34, may indicate measuring batch effects, rather than responses to thermal stress.

0.015 AIMS2 : Ambient Temperature 0.015 AIMS2 : Elevated Temperature Probiotic treatment 0.010 0.010 Control (no inoculum) 0.005 0.005 Negative

Postive 0.000 0.000 ROS / Symb cell Anemone ROS / Symb cell Anemone −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

0.015 AIMS3 : Ambient Temperature 0.015 AIMS3 : Elevated Temperature

0.010 0.010

0.005 0.005

0.000 0.000 ROS / Symb cell Anemone ROS / Symb cell Anemone −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day

0.015 AIMS4 : Ambient Temperature 0.015 AIMS4 : Elevated Temperature

0.010 0.010

0.005 0.005

0.000 0.000 ROS / Symb cell Anemone ROS / Symb cell Anemone −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 −21 −15 −9 01 3 8 12 16 20 24 28 31 34 37 40 43 Day Day Figure 5.4: ROS levels at ambient (26 °C) and elevated (26 °C–31.5 °C) temperature for all treatments within each genotype. Probiotic inoculations were performed at Days 0, 1 and 7. For each datapoint, n = 3. Error bars ± 1 SEM.

ROS levels for anemones in each genotype at ambient or elevated temperature differed significantly according to time, but not probiotic (Table 5.5), suggesting that probiotic treatment did not influence changes in ROS. Despite seemingly higher ROS levels among some

111

heat-stressed anemones at Day 43 (Figure 5.4), according to pairwise comparisons ROS was not significantly higher in the heat-stressed anemones compared to those at ambient temperature (Table A 47, Table A 48, Table A 49). These results may have been due to high variance as the data were highly variable throughout the experiment. There were also no significant differences in ROS within each genotype based on probiotic treatment, indicating that probiotic treatment had no influence on ROS levels.

Table 5.5: ANOVA output for ROS levels in each genotype. Significant values are in bold; α = 0.05.

Genotype Temperature Variable df F-value p-value AIMS2 Ambient Probiotic 2, 6 0.49973 0.6299 Time 13, 72 9.93504 <0.0001 Elevated Probiotic 2, 6 1.8388 0.2383 Time 13, 70 12.1565 <0.0001 AIMS3 Ambient Probiotic 2, 6 0.8167 0.4856 Time 13, 72 14.5667 <0.0001 Elevated Probiotic 2, 6 1.62681 0.2726 Time 13, 71 6.40335 <0.0001 AIMS4 Ambient Probiotic 2, 6 2.27224 0.1842 Time 13, 70 10.52050 <0.0001 Elevated Probiotic 2, 6 1.15426 0.3766 Time 13, 70 5.72760 <0.0001

5.3.5 Metabarcoding data processing

Sequencing produced 17 544 542 reads across 783 E. diaphana samples (minimum 2; mean 23 003, maximum 112 307 reads per sample). After merging, denoising and chimera filtering, 11 904 748 reads remained (minimum 0, mean 15 479, maximum 91 465 reads per sample). Four anemone samples were identified with <20 reads per sample and were removed from the dataset. Decontam identified 209 putative contaminant ASVs, which constituted 1.36% relative abundance of the bacterial communities, and were removed (Table A 50). After all filtering steps, there were 4 555 ASVs across the remaining samples.

5.3.6 Bacterial community shifts

Principal component analyses (PCoA) based on Jaccard distance showed a general temporal shift in membership of the bacterial communities within each genotype (Figure 5.5). There

112

was no apparent clustering or separation of datapoints based on probiotic treatment or temperature condition. Therefore, the primary driver of change in bacterial community composition appeared to be time. However, GLM analyses of the bacterial communities within each genotype revealed that, whilst the communities did not differ significantly based on temperature, there were differences according to probiotic treatment (manyglm, p < 0.05) (Table A 51, Table A 52, Table A 53).

AIMS2 AIMS3 AIMS4 0.4

0.2 0.2 0.2

0.0 0.0 .2 [8.6%] .2 [8.1%] 0.0 .2 [9.4%] Axi s Axi s −0.2Axi s −0.2 −0.2 −0.4

−0.25 0.00 0.25 0.50 −0.2 –0.0 0.2 0.4 −0.2 0.0 0.2 0.4 Axis.1 [20.2%] Axis.1 [13.7%] Axis.1 [13.4%]

Probiotic Control (no inoculum) Ambient (26 °C) treatment Negative Day –21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 Positive Elevated (26–31.5 °C)

Figure 5.5: PCoA ordinations (Jaccard distance) for each genotype. For each datapoint, n = 3.

5.3.7 Incorporation of the probiotic bacteria

ASV-level bubble plots show shifts in relative abundance of the probiotic bacteria within the anemones (Figure 5.6, Figure 5.7, Figure 5.8). Probiotic A. oceani ASVs showed substantial increases after the first and second inoculations for the positive probiotic treatment, regardless of genotype, and this ASV remained low in abundance in the no-inoculum controls. In contrast, the relative abundance of the probiotic A. macleodii ASVs increased substantially in all the inoculated anemones and no-inoculum controls. Probiotic Labrenzia ASVs increased in all inoculated anemones of all genotypes compared to the no-inoculum controls, suggesting incorporation from the probiotic cocktails. An increase of Marinobacter in the inoculated anemones also suggested incorporation from the cocktails despite some moderate, albeit unsustained, increases in the control anemones. There were small but notable increases in Micrococcus probiotic ASVs in most inoculated anemones, particularly in the negative probiotic trestment, but also increases in many uninoculated anemones. Increases in 113

Winogradskyella in the inoculated anemones were clear but modest. Apart from A. macleodii and Marinobacter ASVs, none of the probiotic members were maintained at higher than Day 0 levels for the duration of the experiment.

AIMS2 – control (no probiotic) AIMS2 – positive probiotic AIMS2 – negative probiotic

A. oceani

A. macleodii

Labrenzia

Marinobacter Relative abundance Marinobacter (%)

10 Micrococcus 20

Micrococcus 30

Winogradskyella 40

Winogradskyella 50

Day -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 Inoculation

Figure 5.6: Relative abundance of probiotic ASVs in the AIMS2 anemones. Blue names = data for ambient temperature anemones (26 °C). Pink names = data for elevated temperature anemones (26 °C–31 °C). Probiotic inoculations were performed at Days 0, 1 and 7 and are indicated by stars. For each bubble, n = 3.

114

AIMS3 – control (no probiotic) AIMS3 – positive probiotic AIMS3 – negative probiotic

A. oceani

A. macleodii

Labrenzia

Marinobacter Relative abundance Marinobacter (%)

10 Micrococcus 20

Micrococcus 30

Winogradskyella 40

Winogradskyella 50

Day -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 Inoculation

Figure 5.7: Relative abundance of probiotic ASVs in the AIMS3 anemones. Blue names = data for ambient temperature anemones (26 °C). Pink names = data for elevated temperature anemones (26 °C–31 °C). Probiotic inoculations were performed at Days 0, 1 and 7 and are indicated by stars. For each bubble, n = 3.

115

AIMS4 – control (no probiotic) AIMS4 – positive probiotic AIMS4 – negative probiotic

A. oceani

A. macleodii

Labrenzia

Marinobacter Relative abundance Marinobacter (%)

10 Micrococcus 20

Micrococcus 30

Winogradskyella 40

Winogradskyella 50

Day -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 -21 0 1 3 8 12 16 20 24 28 31 34 37 40 43 Inoculation

Figure 5.8: Relative abundance of probiotic ASVs in the AIMS4 anemones. Blue names = data for ambient temperature anemones (26 °C). Pink names = data for elevated temperature anemones (26 °C–31 °C). Probiotic inoculations were performed at Days 0, 1 and 7 and are indicated by stars. For each bubble, n = 3.

The bubble plots suggested there was little difference in the relative abundance of probiotic bacteria based on temperature, which supported the findings of the genotype-level GLM analyses described above. However, differences appeared to exist between the anemones based on probiotic treatment, particularly in the relative abundance of the probiotic Labrenzia ASVs immediately after Day 0. Paired GLM analyses confirmed that the bacterial communities of each genotype were significantly different based on treatment (Table 5.6). Importantly though, this was short-lived, with differences generally on Days 1 and 3 only. Therefore, when the anemones were exposed to elevated temperature, there was no significant difference between the bacterial communities of the inoculated and uninoculated anemones.

116

Table 5.6: p-values from GLM analyses comparing bacterial communities of anemones receiving different probiotic treatments (selected days only). Significant values are in bold; α = 0.05. Genotype Comparison Day 0 Day 1 Day 3 Day 8 Day 20 Day 31 Day 43 AIMS2 Uninoculated control vs Positive 0.058 0.021 0.028 0.078 0.253 0.286 0.255 Uninoculated control vs Negative 0.146 0.027 0.020 0.144 0.300 0.205 0.212 AIMS3 Uninoculated control vs Positive 0.086 0.005 0.017 0.009 0.329 0.172 0.484 Uninoculated control vs Negative 0.190 0.046 0.027 0.058 0.187 0.457 0.315 AIMS4 Uninoculated control vs Positive 0.124 0.020 0.034 0.082 0.175 0.628 0.320 Uninoculated control vs Negative 0.156 0.006 0.072 0.149 0.194 0.488 0.281

5.4 Discussion

5.4.1 Probiotic inoculation did not improve thermal tolerance

There were significant reductions in Fv/Fm in all anemones at elevated temperature regardless of probiotic treatment, indicating that probiotic treatment did not mitigate the impact of thermal stress on PSII of the Symbiodiniaceae. Although the AIMS2 anemones resisted bleaching at elevated temperature in both the control and inoculation treatments, significant reductions in Symbiodiniaceae cell density of thermally stressed AIMS3 and AIMS4 anemones showed that probiotic treatment did not mitigate bleaching. The similarity of ROS levels in all anemones regardless of treatment, suggested that inoculation with high FRS bacteria did not reduce ROS. Therefore, no evidence for improved thermal tolerance of E. diaphana due to probiotic inoculation was found. It is noteworthy however, that the anemones did not show increased bleaching susceptibility based on probiotic inoculation, indicating that the probiotic bacteria did not harm the anemones.

5.4.2 Probiotic uptake by E. diaphana was short-lived

Post-inoculation increases in genera of the probiotic bacteria and significant differences between the bacterial communities of inoculated and uninoculated anemones on Days 1 and 3 suggested successful probiotic association with the anemones. However, this was short-

117

lived. Consequently, the failure of probiotic inoculation to confer improved thermal tolerance could be explained by the absence of probiotic bacteria after Day 3.

This absence may be due to inadequate dosing. The three-dose strategy employed in the present study was similar to previous cnidarian probiotic experiments. For example, Rosado et al. (2018) inoculated coral samples with a bacterial cocktail twice, five days apart, and Damjanovic et al. (2019b) inoculated coral larvae seven times, at three to four day intervals. However, neither study sampled the target species after each inoculation, and therefore the uptake of each probiotic after dosing was unknown.

An alternative explanation for the transience of the probiotic bacteria is antagonism of the members towards each other. Bacterial antagonism can be exploited when designing probiotics to protect hosts against pathogenic bacteria and has been used to select bacteria to protect anemones and corals against Serratia marcescens and Vibrio coralliilyticus, respectively (Alagely et al. 2011; Rosado et al. 2018). If the probiotic members used in the present study were antagonistic against each other this may have impeded the ability of some members to persist.

Although previous studies have demonstrated the ability of probiotic inoculation to protect E. diaphana from pathogenic infection (Alagely et al. 2011; Zaragoza et al. 2014), incorporation of the probiotic bacteria was not assessed. Therefore, to the best of our knowledge, our study is the first time an inoculation-based modification of the E. diaphana bacterial microbiome has been described.

5.4.3 Probiotic uptake by E. diaphana was uneven

An increase in the Labrenzia ASV provided the strongest evidence for probiotic association with E. diaphana. Labrenzia are naturally abundant (~5%) in all the AIMS genotypes (Chapter 2, Table 2.3b) and are core members of the Symbiodiniaceae microbiome (Lawson et al. 2018), so may have faced low inhibition from antagonistic interactions with resident bacteria (Rypien, Ward & Azam 2010). Poor uptake of the A. oceani, Microccocus and Winogradskyella ASVs could have been caused by below-target densities of those bacterial cultures (Table 5.2), emphasising the need to dose at consistently high cell densities, and possibly higher densities for some bacteria. In contrast, A. macleodii increased rapidly, becoming dominant in all

118

anemones, including the uninoculated anemones, indicating the increase did likely not result from the probiotic inoculum. A. macleodii comprise ~50% of A. salina bacteria (Chapter 2, Table 2.5). Therefore, potentially non-probiotic A. macleodii may have originated from the A. salina feedstock and multiplied due to the culture conditions.

5.4.4 Bleaching susceptibility differed by genotype

As the probiotic bacteria were not retained by the anemones, no correlation can be drawn between differences in bleaching tolerance and inoculation. However, the maintenance of near-normal Symbiodiniaceae cell density in the AIMS2 anemones at elevated temperature versus bleaching of the AIMS3 and AIMS4 anemones is noteworthy. Initial Symbiodiniaceae cell densities in the AIMS3 and AIMS4 anemones were higher than AIMS2, which may support a suggestion that higher Symbiodiniaceae cell density is linked to increased bleaching sensitivity (Cunning & Baker 2013). Alternatively, differences in bleaching tolerance could have been driven by intraspecific diversity of the Symbiodiniaceae in each genotype (Hawkins et al. 2016; Howells et al. 2012). However, this possibility is unknown as previous characterisation of Symbiodiniaceae in the AIMS2–4 genotypes was performed using methods unable to provide taxonomic resolution below species level (Dungan et al. 2020). Nevertheless, observations from the present study indicate that genotype choice will be critical in future experiments seeking to induce or mitigate bleaching in the AIMS genotypes.

5.4.5 Limitations of the study, and recommendations for future probiotic work

Although there was evidence for increased relative abundance of some probiotic members, the key limitation of the present study was the inability of probiotic bacteria to persist after inoculation. Therefore, continual dosing to maintain high levels of probiotic bacteria in the target organisms, as practiced in aquaculture (Hai 2015), is recommended. Another method that could be borrowed from aquaculture is the delivery of probiotic bacteria via bioencapsulation in A. salina (Hai, Buller & Fotedar 2010). This can ensure the bacteria are ingested, although subsequent proliferation is not guaranteed (De et al. 2014).

Antagonism of the probiotic members against each other was not tested. The impact of this on the incorporation and maintenance of the probiotic bacteria is unknown but should be 119 assessed in future studies to ensure that probiotic members are working together, not in opposition.

The original assessments of FRS ability were performed on bacteria grown in pure culture. However, changes in environment can drive phenotypic plasticity resulting in a loss or decrease of previously observed bacterial traits (Kümmerli et al. 2009). Consequently, the persistence of FRS by the selected bacteria may have been impacted by contact with the complex holobiont community. In addition, the broth cultures grown for the probiotic cocktails were not tested for maintenance of the FRS phenotype. Given the rapid genetic evolution of bacteria, FRS ability could have been lost over subsequent generations during sub-culturing to produce each broth culture. Therefore, each culture and cocktail should be checked to ensure FRS ability has been maintained.

Resolving the probiotic bacteria to ASV level allowed us to determine whether they were incorporated by E. diaphana. However, future assessments could be enhanced by also using qPCR to quantify the probiotic bacteria in the samples. Further, qPCR primers based on the genomic region deemed responsible for the FRS phenotype of each probiotic bacterium could be used to confirm that the probiotic capability had been retained after culturing.

An element of the present study that we recommend as good practice is the use of a negative probiotic. Studies have suggested that increased heterotrophy can reduce the impact of thermal stress on corals (Aichelman et al. 2016; Grottoli, Rodrigues & Palardy 2006). Unlike comparable probiotic studies, inclusion of a negative probiotic allowed us to account for differences in response to thermal stress due to the introduction of an additional source of nutrition. Further recommendations include frequent dosing to maintain elevated levels of probiotic bacteria, testing of probiotic candidate bacteria for antagonistic activity, and the use of coral or E. diaphana genotypes that produce typical, predictable stress responses to elevated temperature. Increases in probiotic genera in the uninoculated control anemones also highlight the importance of sample segregation.

5.5 Conclusion To date, few studies have investigated the ability of probiotic bacteria to increase cnidarian thermal tolerance. In this study we were able show that probiotic inoculation using bacteria 120 with high or low FRS abilities did not harm E. diaphana. Additionally, some members of our probiotic were accepted by the host after dosing, as evidenced by increases in relative abundance of several probiotic ASVs. However, the observed increases were short-lived, and the probiotic bacteria did not persist above normal levels. Therefore, whilst probiotic inoculation did not mitigate bleaching in heat-exposed anemones, this cannot be attributed to an inability of the inoculum to confer improved thermal tolerance, and Instead remains unproven. Future studies that increase the frequency of probiotic dosing to maintain elevated levels of probiotic bacteria will provide clearer insights into the potential of the coral bleaching mitigation strategy proposed here.

121

Chapter 6: General summary

The studies performed for this thesis addressed each research question. However, not all provided conclusive answers. Despite this, each study produced data that will guide future use of the recently established AIMS E. diaphana genotypes. Key findings and suggestions for future investigations are presented below.

6.1 GBR-origin E. diaphana have bacterial microbiomes similar to their cnidarian cousins

Characterising the bacterial associates of the AIMS1–4 anemones was a logical starting point for this thesis due to interest in the influence of host-associated bacteria on cnidarian health. According to my study, described in Chapter 2, the bacteria of the AIMS1–4 anemones did not differ substantially from corals or two established E. diaphana models at a high taxonomic level, with most bacteria being members of Proteobacteria. Overall, the bacterial associates of the AIMS1–4 E. diaphana genotypes did not differ significantly from each other either, although differences may have been obscured by tank effects. A comparison of the bacterial microbiomes of lab-cultured and wild proxy anemones suggested a loss of bacterial diversity in culture, which may indicate a shift towards a minimal microbiome due to the simplicity and stability of culture environments.

Eliminating tank effects will be necessary to clarify the relationship between environmental influences and bacterial composition. This could be addressed, for example, by co-housing the genotypes. However, each individual anemone would need to be genotyped as they are visually indistinguishable. To achieve this, a new genotyping assay would be required as a previous phylogenetic analysis of the AIMS1–4 anemones (Dungan et al. 2020) that targeted marker genes used in other E. diaphana studies (Bellis & Denver 2017; Thornhill et al. 2013) was unable to separate AIMS3 and AIMS4. Alternatively, a comparison of the bacterial microbiomes could be performed using co-housed anemones from AIMS1 and AIMS2, and one of AIMS3 or AIMS4 only.

Analysing the communities that reside in the different compartments of the holobiont would also be a useful addition to future characterisation analyses. Although E. diaphana does not have a skeleton, identification of bacteria specific to the surface mucus layer, tissue and 122

gastrovascular cavity would improve our understanding of bacterial distribution in the holobiont and could shed light on the role bacteria play in host protection or nutrition. These data could also reveal differences between E. diaphana and corals in the way bacterial communities are partitioned.

The data gathered in this study, helped position the AIMS1–4 anemones alongside the main E. diaphana models, CC7 and H2, and anemones from more ecologically rich environments. It also provided valuable baseline information for the ongoing use of these anemones in microbiome-related research and expanded the global E. diaphana knowledgebase.

6.2 The bacterial microbiome of heat-stressed E. diaphana is dynamic

Bacterial community shifts in cnidarians exposed to thermal stress are of interest due to recent increases in coral bleaching linked to climate change and suggestions that host- associated bacteria influence thermal tolerance. However, the bacterial communities of some corals remain stable under heat stress (Gajigan, Diaz & Conaco 2017; Grottoli et al. 2018), whereas others undergo change (Lee et al. 2016; Tracy et al. 2015), and neither response is a predictor for host bleaching susceptibility. Therefore, the relationship between cnidarian bleaching and their bacteria is enigmatic and requires further investigation. The behaviour of E. diaphana’s bacteria to elevated temperature is largely unknown but must be characterised to improve its utility in these studies.

In my study of the microbiome dynamics of heat-treated E. diaphana, described in Chapter 3, AIMS2 anemones exposed to thermal stress began bleaching immediately upon a temperature increase above ambient conditions (26 °C). Bacterial alpha and beta diversity in the anemones exposed to elevated or ambient temperature underwent similar declines regardless of temperature treatment. However, when the temperature exceeded 32 °C significant differences between the communities in each treatment emerged. A small subset of bacteria underwent changes in relative abundance that appeared to be temperature dependent and could therefore serve as biomarkers for thermal stress in E. diaphana.

The immediate onset of bleaching in the heat-exposed anemones indicated low thermal tolerance for the AIMS2 anemones used in the experiment. Although temperature became the primary driver of bacterial change above 32 °C, it was plausible that bleaching 123

susceptibility was linked to starvation and incomplete adjustment of the anemones to the experimental environment despite 10 days acclimation. The impact of these stressors was not predicted as starvation is commonly practiced in experiments with E. diaphana (Bieri et al. 2016; Hillyer et al. 2016; Marty-Rivera, Yudowski & Roberson 2018; Röthig et al. 2016; Zaragoza et al. 2014) and the allowance for acclimation was relatively generous. This suggested that E. diaphana from the GBR are sensitive to environmental changes, and may be highly reliant on heterotrophy despite the Symbiodiniaceae translocating >95% of their photosynthate to the host according to one study (Muller-Parker & Davy 2001). The results indicated that future studies using the AIMS1–4 anemones will need to accommodate these traits to avoid confounding due to unintended stress.

Changes in abundance of some bacteria may have been linked to the loss of Symbiodiniaceae. Data describing associations between Symbiodiniaceae and bacteria may support this (Bernasconi et al. 2018; Lawson et al. 2018). However, published data from Symbiodiniaceae isolated from the AIMS1–4 anemones are needed to explain whether the changes in bacterial abundance seen in this study were related to bleaching. Nevertheless, this study provided new insights into E. diaphana’s bacterial response to thermal stress and valuable lessons for the future use of the AIMS anemones.

6.3 Generation of gnotobiotic E. diaphana remains a work in progress

Gnotobiotic organisms have proven useful for elucidating the roles of bacteria in holobiont function; for example, the relationship between insects (Kešnerová et al. 2017; Mikaelyan et al. 2015) or animals (Laycock et al. 2012; Rawls, Samuel & Gordon 2004), and the bacteria that influence their development and survival. Therefore, gnotobiotic E. diaphana cultures could reveal the importance of bacteria in cnidarian health, including nutrition, protection from pathogens and thermal stress response. However, there are few descriptions of gnotobiotic E. diaphana generation in the literature. Consequently, methods for creating gnotobiotic E. diaphana need development, with antibiotic exposure being the simplest approach.

In my study, described in Chapter 4, E. diaphana dosed with a cocktail of seven antibiotics for three weeks did not generate gnotobiotic anemones. Although the bacterial load of the anemones was significantly reduced, the remaining bacterial communities were highly diverse

124

and lacked uniformity, and hence the anemones could not be considered gnotobiotic. The anemones also bleached following treatment. However, this appeared to be temporary and no anemones died, which indicates that prolonged treatment is feasible.

The increase in bacterial diversity suggests that many low abundance bacteria with antibiotic resistance proliferated after the elimination of dominant, but non-resistant, bacteria. To overcome this, the antibiotic cocktail used in the study may need modification. Alternatively, treatment may only be effective on subjects with few resident bacteria, such as E. diaphana larvae or pedal lacerates, and these avenues should be explored. In future, bacterial seeding from the A. salina feedstock will also need to be addressed through more extensive sterilization. Fortunately, A. salina sterilization methods are well described (Costa, Cárdenas & Voolstra 2019; Sorgeloos et al. 1977).

A noteworthy element of the study was development of a digital droplet PCR (ddPCR) assay for estimating bacterial load in the anemones. The relevance of bacterial load to coral health has been cited in many studies but seldom measured (e.g. Ainsworth & Hoegh-Guldberg 2009; Sharp & Ritchie 2012; Smith et al. 2015). Techniques suggested for measuring bacterial load in coral samples include quantitative PCR (qPCR), flow cytometry, fluorescence in-situ hybridisation (FISH) and bacterial cell culturing (Cooke et al. 2019). Each has benefits. For example, flow cytometry is highly accurate, FISH enables localisation of bacteria in coral tissue, and cell culturing allows quantification of viable cells, providing they are culturable. However, compared to ddPCR, these techniques are complex or time-consuming. Therefore, the ddPCR assay described in Chapter 4 could provide a convenient way for researchers to obtain a more complete picture of bacterial influence on cnidarian hosts.

6.4 The ability of probiotic inoculation to mitigate bleaching in E. diaphana remains unknown

Enhancing cnidarian tolerance to environmental stress by microbiome manipulation has been proposed by several authors (Peixoto et al. 2017; Teplitski & Ritchie 2009; van Oppen et al. 2015). One study has since shown that the microbiomes of coral larvae can be manipulated by bacterial inoculation (Damjanovic et al. 2019b). Two further studies have demonstrated that inoculation with probiotic bacteria can protect cnidarians from disease and bleaching

125

caused by pathogenic bacteria and thermal stress (Alagely et al. 2011; Rosado et al. 2018), thus paving the way for exploration of cnidarian trait enhancement by probiotic inoculation.

Our investigation of thermal tolerance enhancement in E. diaphana by probiotic inoculation, described in Chapter 5, was inconclusive. Although the relative abundance of some of the free radical scavenging (FRS) bacteria used in the inoculum increased in the dosed anemones, this was short-lived. Therefore, we could not judge the ability of the probiotic to mitigate bleaching by reducing ROS.

Nevertheless, important outcomes of the study included the different responses of the AIMS2–4 genotypes under thermal stress. Whilst AIMS3 and AIMS4 bleached when exposed to elevated temperature (31.5 °C), the AIMS2 anemones did not. Another, noteworthy observation was the five-week acclimation period of the anemones after relocation, as judged by changes in Fv/Fm, which may confirm suggestions that incomplete acclimation influenced results obtained in the heat-stress experiment (Chapter 3). Awareness of these findings will be critical in future work with the AIMS1–4 anemones.

6.5 Future directions

The outcomes of the studies performed for this thesis may have been impacted by the use of relative rather than absolute abundance data. Describing bacterial communities in terms of relative abundance potentially obscures important patterns due to the compositional nature of the data (Gloor et al. 2017). Moreover, because compositionality violates assumptions of data independence, it may also invalidate statistical analyses (Jiang et al. 2019). Therefore, to strengthen our interpretation of bacterial community structure, progress must be made towards generation of absolute abundance data. As noted in Chapter 3, at least one method has been developed to facilitate this; synthetic DNA spike-in (Stämmler et al. 2016; Tourlousse et al. 2017). This method was recently used to reveal differences in plant rhizosphere microbiomes that would have otherwise been hidden (Guo et al. 2020), thus demonstrating its value in microbiome research.

Bacterial communities influence host phenotypes, for example, through their impact on host immunity, development and digestion (Lynch & Hsiao 2019), but linking community membership and host phenotype has proven difficult. However, systems biology, wherein 126

data from different ‘omics methods are combined, has become a powerful tool for overcoming this (Abram 2015). Among the array of ‘omics methods available, metabolomics could be particularly useful for assessing the impact of thermal stress on cnidarians as it reveals the functional status and cellular response of organisms and their symbionts to biotic and abiotic stimuli (Viant 2007). Importantly, when integrated with bacterial community data, metabolite profiles could help identify taxa associated with metabolic fluxes (Noecker et al. 2019). Consequently, if cnidarian probiotic experiments were analysed using both metabarcoding and metabolomic data, the contribution of inoculum members to thermal stress responses, and therefore the efficacy of the treatment, could be more clearly resolved.

Bacteria used in the probiotic inocula described in Chapter 5 were selected not just for FRS ability, but also their cultivability. This likely excluded many FRS bacteria present in the donor material (Stewart 2012). However, genetic modification has been proposed as a method for expanding the number of culturable bacteria with desired traits by using host-associated probiotic bacteria to guide gene selection, and gene-editing tools, such as CRISPR-Cas9, to transform so-called ‘chassis’ bacteria (Mimee, Citorik & Lu 2016; van Oppen & Blackall 2019). Previous examples of bacterial engineering for in situ delivery of therapeutic compounds have been primarily based on Escherichia coli and lactic acid bacteria (Bober, Beisel & Nair 2018) but provide proof of concept for application in native cnidarian bacteria and their hosts, and warrant future investigation.

A likely advantage of using host-derived chassis bacteria would be the stability of host colonization by transgenic bacteria, although testing would be needed to ensure this trait was retained after transformation (Inda et al. 2019). A second important advantage could be the specificity of the relationship between the bacteria and the intended target. This would reduce the risk of transmission to non-target organisms, which has been noted by bacterial engineering advocates (van Oppen & Blackall 2019), and groups scrutinising the methods proposed for coral reef conservation (National Academies of Sciences Engineering and Medicine 2019). Weighing up the risks versus the potential benefits of interventions, such as those proposed under assisted evolution (e.g., microbiome manipulation and hybridization) (van Oppen & Blackall 2019; van Oppen et al. 2015), will become more important as coral reefs succumb to climate change and the push for action increases.

127

In conclusion, this thesis produced data that increased our understanding of E. diaphana bacterial communities, including their general characteristics and response to environmental stressors. Investigating the feasibility of gnotobiotic generation by antibiotic exposure and bleaching mitigation by probiotic inoculation has also laid the groundwork for further exploration of these methods. An important feature of the thesis was the use of the AIMS1–4 genotypes. As these anemones are utilised in more studies, particularly in transcriptomic, metabolomic or proteomic work that generates descriptive data, their value as research tools will increase. However, the recent publication of two studies using the AIMS1–4 anemones, including the study described in Chapter 3, shows that they are already on their way to making an impact as coral models (Hartman, van Oppen & Blackall 2019; Tortorelli et al. 2020).

128

Appendix 1

Culture collection establishment

The Exaiptasia diaphana cultures in this study were taken from the culture collection at the University of Melbourne (UoM), Australia. Creation and maintenance of the collection, and details of species verification, genotyping and gender determination are described in (Dungan et al. 2020). However, for the present study, it should be noted that the anemones in the collection were acquired in two stages, and therefore have different culture histories. An initial group of E. diaphana were supplied to Swinburne University of Technology (SUT) in Melbourne, Victoria in late 2014 when coral skeleton fragments with attached anemones (Figure A 1) were obtained from aquaria in the National Sea Simulator (SeaSim) at the Australian Institute of Marine Science (AIMS). These anemones were found to comprise two genotypes and genders, subsequently designated AIMS2 (male) and AIMS4 (female).

Figure A 1: E. diaphana acquired from the AIMS SeaSim in late 2014. Original founder anemones are shown in two large beakers, and asexually generated offspring are shown in small beakers as at February 2016. Aeration was provided by orbital agitation at low speed.

129

In early 2016 a group of E. diaphana were supplied by AIMS to UoM, as above, and kept in a 6 L glass tank (Figure A 2). These anemones were found to comprise two further genotypes and one gender, subsequently designated AIMS1 (female) and AIMS3 (female). In March 2017, all SUT anemones were transferred to UoM to create a single culture collection.

Figure A 2: E. diaphana acquired from the AIMS SeaSim in early 2016. Anemones are shown as at October 2016. Aeration was provided by bubbling unfiltered air from an aquarium pump.

Culture collection maintenance

The culture collection is maintained in a temperature-controlled walk-in incubator in 6 L clear polycarbonate tanks, with three tanks assigned per genotype to create three independent populations per genotype (Figure A 3). The tanks are rotated between the shelves each week. Lighting is provided by white light LEDs strips at 12-20 µmol photons m–2 s–1 on a 12 h:12 h light-dark cycle. Temperature is maintained at 26 °C. The water is 100% changed once weekly with reconstituted seawater made from Red Sea Salt™ (R11065, Red Sea, USA) at ~34 parts per thousand. Feeding is performed twice weekly with freshly hatched Artemia nauplii (Salt Creek, Premium GSL, USA). At the time of sampling for the present study, aeration was provided by bubbling unfiltered air from small aquarium pumps through air-stones.

130

Figure A 3: E. diaphana culture collection at UoM. Genotypes AIMS1–4 are located left to right, with three replicate tanks located top to bottom. For the present study, six whole anemones and one 1 L water sample were taken from each tank.

Wild proxy anemone acquisition

In October 2016, five anemone polyps were collected as wild proxies of E. diaphana from the outflow (Figure A 4a) of a 4000 L outdoor holding tank containing live corals, snails, sea cucumbers and fish at the AIMS SeaSim, Townsville, Australia (Figure A 4b). The tank received seawater from the SeaSim recirculation system with inline 0.45 µm filtration and fractionation to remove particulates and protein respectively. The water temperature was stable at 27 °C, and 120 mL of A. salina nauplii feedstock was added daily.

131

5 cm (a) (b) Figure A 4: (a) Outflow with anemones, and (b) AIMS holding tank – outflow at left.

For anemone species verification, E. diaphana-specific 18S rRNA gene primers (18S_NA [5’ TAAGCACTTGT CTGTGAAACTGCGA 3’]; 18S_NB [5’ AGGAGTCCTCACTAAACCAT 3’] (Grajales & Rodríguez 2015)) were used to generate amplicons by single PCRs on DNA extracts (see main text, Methods). The PCR product was checked by 1% agar gel electrophoresis, purified (Bioline II PCR and Gel Kit BIO-52059), diluted (5 ng/µL), and sent to the Australian Genome Research Facility, Melbourne, for Sanger sequencing using the external 18S PCR primers described above, and four internal 18S rRNA gene primers: 18S_NL [5’ AACAGCCCGGTCAGTAACACG 3’]; 18S_NC [5’ AATAACAATACAGGGCTTTTCTAAGTC 3’]; 18S_NY [5’ GCCTTCCTGACTTTGGTTGAA 3’]; 18S_NO [5’ AGTGTTATTGGATGACCTCTTTGGC 3’] (Grajales & Rodríguez 2015). The sequence data was imported into Geneious (v 10.0.4) (Kearse et al. 2012). The chromatograms were visually inspected, and the sequences manually trimmed then de novo assembled. The top BLAST (Altschul et al. 1990) hits for the consensus sequences against the NCBI database (Federhen 2012) were for Aiptasia pulchella or E. pallida. However, as A. pulchella is synonymous with E. pallida (Grajales & Rodríguez 2014), and Exaiptasia diaphana is the currently approved nomenclature for these anemones (ICZN 2017), all samples were designated E. diaphana. The assembled 18S data are available under NCBI BioProject PRJNA575811.

132

wp5 wp2

wp1

500 wp4 400

300 e13a

e13be23b e11f e11ae11be41e e11c e21ee33f e43b wp3 e31ce41ae13fe11e e43ee22fe31ae33c e31e e21fe23ee13ee23ce41f e11de23de43fe13de43d Sample ASV s e12fe41ce31fe12ee21b e23a e33ee43ce42fe13c 200 e21ce32c e42b e21de33de42ee23fe41de31be32f e42ae41b h24 e21a e22e cc72 e31d e32de42de12ce22be33be32ee12ae22ae43a e22ce32a cc75 cc73 e42ce32be22d h21 w41e33a e12d cc71 cc74 cw.bw33lankw13 h25 e12bw11 w23 w43 w42w31 h23 w21 w22

100 w32 h22 w12

ce.blank cneg1 arat2arrt3t1 cneg2

0 cneg3

0 50000 100000 150000 Sample reads

Figure A 5: Rarefaction curves for all microbiome characterisation samples. Curves for all anemone and water samples plateaued, indicating that sequencing depth was sufficient to capture bacterial species diversity (sub-sampling level: 12 000; step: 600).

Table A 1: Contaminant ASVs removed from the characterisation dataset. Seven ASVs potentially introduced during sample processing were identified with decontam (Davis et al. 2018) using data from the negative control samples. All were removed from the dataset.

Phylum Class Order Family Genus AIMS1-4 AIMS1-4 Wild proxy anemones water anemones Rel. ab. (%) Rel. ab. (%) Rel. ab. (%) 1 Proteobacteria Alphaproteobacteria Caulobacterales Caulobacteraceae Brevundimonas 0.0033 0.0000 0.0478

2 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas 0.0063 0.0000 0.0000

3 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas 0.0041 0.0000 0.0000

4 Planctomycetes Planctomycetacia Pirellulales Pirellulaceae Rhodopirellula 0.0001 0.0000 0.0408

5 Actinobacteria Actinobacteria Micrococcales Micrococcaceae Unclassified 0.0058 0.0000 0.0000

6 Proteobacteria Gammaproteobacteria Pseudomonadales Moraxellaceae Enhydrobacter 0.0051 0.0000 0.0000

7 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Ralstonia 0.0350 0.0699 1.9880

0.0597 0.0699 2.0766

133

Family-level bar-charts revealed three samples with overrepresented ASVs (Figure A 7). The wild proxy “wp3” and AIMS3 “e33a” samples contained ~25% and ~57% Enterobacteriaceae respectively, compared to 0.5%, on average, for the other anemone samples. Enterobacteriaceae was absent from the A. salina feedstock, therefore it was assumed to be a contaminant. The AIMS4 sample “e43f” contained a high proportion of Vibrionaceae (~39%), which were further identified as Vibrio sp.. Although Vibrio sp. are known coral associates, all other anemone samples contained <4% Vibrionaceae, and most <1%. Consequently, wp3, e33a and e43f were removed from the analysis.

100

50 Percentage of reads

0 h21 h22 h23 h24 h25 art1 art2 art3 w11 w12 w13 w21 w22 w23 w31 w32 w33 w41 w42 w43 wp1 wp2 wp3 wp4 wp5 e11f e12f e13f e21f e22f e23f e31f e32f e33f e41f e42f e43f e11a e11c e11e e12a e12c e12e e13a e13c e13e e21a e21c e21e e22a e22c e22e e23a e23c e23e e31a e31c e31e e32a e32c e32e e33a e33c e33e e41a e41c e41e e42a e42c e42e e43a e43c e43e cc71 cc72 cc73 cc74 cc75 e11b e11d e12b e12d e13b e13d e21b e21d e22b e22d e23b e23d e31b e31d e32b e32d e33b e33d e41b e41d e42b e42d e43b e43d AIMS1 AIMS2 AIMS3 AIMS4 UoM water Samples

Figure A 6: Relative proportions of family level ASVs. ▲ Red triangles indicate samples deemed contaminated. Control samples were omitted. Bars do not reach 100% because some ASVs could not be classified to family level.

134

01D2Z36 Bogoriellaceae Desulfovibrionaceae Halomonadaceae metagenome Oligoflexaceae Rhodothermaceae Tenderiaceae 0319−6G20 Bradymonadaceae Desulfuromonadaceae Halorhodospiraceae Methylococcaceae P3OB−42 Rickettsiaceae Terasakiellaceae 67−14 Brevibacteriaceae DEV007 Halothiobacillaceae Methyloligellaceae Paenibacillaceae Rikenellaceae Thalassospiraceae A0839 Burkholderiaceae Devosiaceae Helicobacteraceae Methylomonaceae Paludibacteraceae Rubinisphaeraceae Thermoactinomycetaceae A4b Caldilineaceae Dietziaceae Holophagaceae Methylophagaceae Paracaedibacteraceae Rubritaleaceae Thermoanaerobaculaceae AB1 Calditrichaceae Diplorickettsiaceae Holosporaceae Methylophilaceae Parachlamydiaceae Rubrobacteriaceae Thioalkalispiraceae Acanthopleuribacteraceae Campylobacteraceae Dongiaceae Hydrogenophilaceae Micavibrionaceae Paraspirulinaceae Ruminococcaceae Thiohalorhabdaceae Acetobacteraceae Carnobacteriaceae Dysgonomonadaceae Hymenobacteraceae Microbacteriaceae Parvibaculaceae S25−593 Thiomicrospiraceae Acidobacteriaceae (Sub 1) Caulobacteraceae Ectothiorhodospiraceae Hyphomicrobiaceae Microbulbiferaceae Parvularculaceae Saccharospirillaceae Thiotrichaceae Actinomarinaceae Cellulomonadaceae Eel−36e1D6 Hyphomonadaceae Micrococcaceae Pasteurellaceae Salinisphaeraceae TRA3−20 Actinomycetaceae Cellvibrionaceae Eggerthellaceae Iamiaceae Micromonosporaceae Pasteuriaceae Sandaracinaceae UASB−TL25 AEGEAN−169 marine group Chitinophagaceae Elsteraceae Idiomarinaceae Micropepsaceae Pedosphaeraceae Sanguibacteraceae uncultured Aerococcaceae Chlamydiaceae Endozoicomonadaceae Ilumatobacteraceae Microscillaceae Peptostreptococcaceae Saprospiraceae uncultured Acidobacteria Aeromonadaceae Chromatiaceae Enterobacteriaceae Intrasporangiaceae Microtrichaceae Phaselicystidaceae SAR116 clade uncultured alpha proteobacterium Alcanivoracaceae Chthoniobacteraceae Enterococcaceae Isosphaeraceae MidBa8 Phycisphaeraceae SC−I−84 uncultured bacterium Algiphilaceae Clade I env.OPS 17 KD3−10 Midichloriaceae Pirellulaceae Schleiferiaceae uncultured delta proteobacterium Alicyclobacillaceae Clade II Erysipelotrichaceae KD3−93 mle1−27 Piscirickettsiaceae Sedimenticolaceae uncultured Flexibacter sp. Alteromonadaceae Clade IV Euzebyaceae Kiloniellaceae Moraxellaceae Planococcaceae Shewanellaceae uncultured gamma proteobacterium Amoebophilaceae Clostridiaceae 1 Family X Kineosporiaceae Muribaculaceae Polyangiaceae Simkaniaceae uncultured Gemmatimonadetes Aquaspirillaceae Clostridiaceae 2 Family XI Kiritimatiellaceae MWH−CFBk5 Porphyromonadaceae SM2D12 uncultured marine bacterium Arcobacteraceae Clostridiaceae 3 Family XII Kordiimonadaceae Mycobacteriaceae Porticoccaceae Sneathiellaceae uncultured organism Arenicellaceae Colwelliaceae Family XIII Koribacteraceae Mycoplasmataceae Prevotellaceae Solibacteraceae (Sub 3) uncultured Planctomyces sp. Atopobiaceae Coriobacteriaceae Fibrobacteraceae Lachnospiraceae Myxococcaceae Prolixibacteraceae Solimonadaceae uncultured planctomycete Azospirillaceae Corynebacteriaceae Fimbriimonadaceae Lactobacillaceae Nakamurellaceae Promicromonosporaceae Solirubrobacteraceae uncultured SAR324 cluster bacterium B122 Coxiellaceae Flammeovirgaceae Legionellaceae Nannocystaceae Propionibacteriaceae Sphingobacteriaceae uncultured Verrucomicrobia Babeliaceae Crocinitomicaceae Flavobacteriaceae Lentimicrobiaceae Neisseriaceae PS1 clade Sphingomonadaceae Unknown Family Bacillaceae Cryomorphaceae Fodinicurvataceae Lentisphaeraceae Nisaeaceae Pseudoalteromonadaceae Spirochaetaceae Veillonellaceae bacteriap25 Cryptosporangiaceae Francisellaceae Leptospiraceae Nitriliruptoraceae Pseudohongiellaceae Spirosomaceae Vermiphilaceae Bacteriovoracaceae Cyclobacteriaceae Frankiaceae Leptotrichiaceae Nitrincolaceae Pseudomonadaceae Spongiibacteraceae Vibrionaceae Bacteroidaceae Cytophagaceae Fusobacteriaceae Leuconostocaceae Nitrosococcaceae Pseudonocardiaceae Sporichthyaceae Weeksellaceae Balneolaceae D05−2 Gaiellaceae Limnotrichaceae Nitrosomonadaceae Psychromonadaceae Sporolactobacillaceae Woeseiaceae Bdellovibrionaceae Deferribacteraceae Geminicoccaceae Listeriaceae Nitrospiraceae Reyranellaceae Staphylococcaceae Xanthobacteraceae Beggiatoaceae Demequinaceae Gemmatimonadaceae Litoricolaceae Nocardiaceae Rhizobiaceae Stappiaceae Xanthomonadaceae Beijerinckiaceae Dermabacteraceae Geodermatophilaceae LiUU−11−161 Nocardioidaceae Rhizobiales Incertae Sedis Steroidobacteraceae Bifidobacteriaceae Dermacoccaceae Gimesiaceae Longimicrobiaceae Nocardiopsaceae Rhodanobacteraceae Streptococcaceae BIrii41 Desulfarculaceae Gracilibacteraceae Magnetospiraceae Nodosilineaceae Rhodobacteraceae Streptomycetaceae Blastocatellaceae Desulfobacteraceae Halanaerobiaceae marine metagenome NS11−12 marine group Rhodocyclaceae Sulfurospirillaceae Blattabacteriaceae Desulfobulbaceae Haliangiaceae Marinilabiliaceae NS9 marine group Rhodopirillaceae Syntrophaceae Blfdi19 Desulfohalobiaceae Halieaceae Marinobacteraceae Oleiphilaceae Rhodospirillaceae Tannerellaceae

Figure A 7: Legend of family-level ASVs.

135

According to Levene’s test, the variance of the Shannon data for the sample types did not differ significantly from homogeneity: (F(6, 77) = 0.93, p = 0.48). However, tests for normality by Shapiro-Wilk suggested that the data for three sample types, AIMS1 (w = 0.87, p = 0.02), AIMS4 (w = 0.89, p = 0.04) and the wild proxies (w = 0.70, p = 0.01), were not normally distributed. Inspection of the plotted residuals showed that AIMS1 and AIMS4 did not deviate substantially from normality but contained one and two outliers, respectively. The wild proxy result was likely exacerbated by the low number of samples (four). As ANOVA is robust to non- normality (Blanca et al. 2017), the data was analysed by ANOVA (results reported in main text) and Tukey’s HSD (Table A 2).

Table A 2: Tukey’s HSD p-values from pair-wise Shannon value comparison. Significant p-values are in bold, α = 0.05.

AIMS1 AIMS2 AIMS3 AIMS4 Wild proxies CC7 AIMS2 0.116 AIMS3 0.642 0.956 AIMS4 0.999 0.077 0.521 Wild proxies 0.000 0.000 0.000 0.000 CC7 0.000 0.060 0.010 0.000 0.000 H2 0.000 0.000 0.000 0.000 0.000 0.000

136

Table A 3: Relative abundance of phyla in each AIMS1–4 genotype. Empty cell = not detected.

AIMS1 AIMS2 AIMS3 AIMS4 Phylum (%) (%) (%) (%) Proteobacteria 75.89 75.08 77.19 76.80 Bacteroidetes 17.05 13.78 18.47 13.35 Spirochaetes 0.04 5.73 0.02 4.13 Planctomycetes 1.65 2.43 1.50 1.93 Acidobacteria 1.88 1.53 1.36 2.50 Chlamydiae 2.63 0.60 0.81 0.60 Actinobacteria 0.69 0.58 0.42 0.60 Firmicutes 0.15 0.15 0.20 0.06 Calditrichaeota <0.01 0.07 <0.01 0.01

Verrucomicrobia <0.01 0.02 0.01

Gemmatimonadetes <0.01 <0.01 0.02

Cyanobacteria 0.01 0.02 <0.01

Dependentiae 0.01 0.01 <0.01

Patescibacteria <0.01 <0.01 0.00

WPS-2 <0.01 <0.01

Elusimicrobia <0.01 <0.01

Lentisphaerae <0.01 <0.01

Fusobacteria <0.01

137

Table A 4: Pairwise binary log fold Table A 5: Pairwise binary log fold change (L2FC) for the 20 most abundant change (L2FC) for the 20 most abundant class-level taxa. Only results with genus-level taxa. Only results with p < 0.05 and L2FC > 2.00 are shown. p < 0.05 and L2FC > 2.00 are shown.

Class pair L2FC adjusted p-value Genus pair L2FC adjusted p-value Acidimicrobiia AIMS1 v WP -5.25 1.39E-05 Aestuariibacter AIMS2 v WP 4.32 1.26E-03 AIMS2 v WP -4.84 3.27E-05 AIMS4 v WP 2.67 4.62E-02 AIMS3 v WP -5.76 6.12E-07 Alteromonas AIMS1 v WP 10.66 1.56E-16 AIMS4 v WP -5.26 5.16E-06 AIMS2 v WP 10.90 2.48E-17 Acidobacteriia AIMS2 v AIMS4 6.91 1.35E-02 AIMS3 v WP 10.73 7.18E-17 AIMS1 v WP 2.24 1.85E-02 AIMS4 v WP 11.31 7.98E-19 AIMS2 v WP 2.34 1.08E-02 Chitinophagales (uncultured) AIMS3 v WP 2.64 1.09E-02 AIMS4 v WP 2.39 4.34E-03 Cyclobacteriaceae (uncultured) AIMS1 v WP -8.15 4.20E-03 BD2-11 terrestrial group AIMS1 v WP -8.01 2.88E-02 AIMS2 v WP -9.52 6.38E-04 AIMS4 v WP -9.39 4.34E-03 AIMS3 v WP -11.54 1.63E-05 Holophagae AIMS1 v AIMS4 -3.24 3.34E-04 AIMS4 v WP -11.13 2.37E-05 AIMS1 v WP 3.82 1.99E-02 Labrenzia AIMS3 v WP 2.01 1.88E-02 AIMS2 v WP 5.85 8.20E-05 Leisingera AIMS1 v WP 9.18 3.98E-13 AIMS3 v WP 4.98 1.73E-03 AIMS2 v WP 9.62 1.62E-14 AIMS4 v WP 7.06 2.71E-06 AIMS3 v WP 9.94 1.34E-15 Ignavibacteria AIMS1 v WP -9.27 4.57E-02 AIMS4 v WP 10.54 1.28E-17 Pla3 lineage AIMS1 v AIMS2 -10.39 3.21E-31 Marinobacter AIMS1 v WP 6.21 7.37E-21 AIMS1 v AIMS3 -8.89 3.86E-22 AIMS2 v WP 6.68 2.47E-24 AIMS1 v AIMS4 -7.98 4.63E-18 AIMS3 v WP 6.93 3.31E-26 AIMS2 v AIMS4 2.41 1.98E-02 AIMS4 v WP 6.42 9.72E-23 AIMS2 v WP 9.48 6.62E-08 Methylotenera AIMS1 v AIMS3 2.35 1.07E-03 AIMS3 v WP 7.98 6.10E-06 AIMS1 v WP 9.42 1.04E-09 AIMS4 v WP 7.06 4.94E-05 AIMS2 v AIMS4 -2.58 1.94E-04 Spirochaetia AIMS1 v AIMS2 -10.05 7.42E-44 AIMS2 v WP 7.99 3.53E-07 AIMS1 v AIMS3 3.05 8.20E-03 AIMS3 v AIMS4 -3.51 1.50E-07 AIMS1 v AIMS4 -9.41 1.83E-37 AIMS3 v WP 7.06 7.51E-06 AIMS1 v WP -4.19 4.28E-03 AIMS4 v WP 10.57 2.18E-12 AIMS2 v AIMS3 13.10 1.78E-53 Nonlabens AIMS1 v WP 4.53 3.10E-03 AIMS2 v WP 5.86 2.44E-06 AIMS2 v WP 4.62 2.24E-03 AIMS3 v AIMS4 -12.46 1.74E-47 AIMS3 v WP 4.59 1.94E-03 AIMS3 v WP -7.24 1.30E-07 AIMS4 v WP 5.39 1.50E-04 AIMS4 v WP 5.22 2.54E-05 Peredibacter AIMS4 v WP 2.23 3.54E-03 Subgroup 22 AIMS1 v AIMS2 7.26 9.39E-18 Rhodobacteraceae (uncultured) AIMS1 v WP -3.92 8.54E-08 AIMS1 v AIMS4 7.11 7.85E-17 AIMS2 v WP -2.61 8.11E-04 AIMS1 v WP 4.45 5.48E-03 AIMS3 v WP -4.40 1.01E-09 AIMS2 v AIMS3 -6.75 8.40E-15 AIMS4 v WP -3.92 5.12E-08 AIMS3 v AIMS4 6.60 5.52E-14 Ruegeria AIMS1 v AIMS2 5.44 6.30E-04 AIMS3 v WP 3.93 1.62E-02 AIMS1 v WP -6.37 1.87E-02 Thermoanaerobaculia AIMS2 v WP -3.86 1.32E-02 AIMS2 v WP -11.81 1.09E-06 AIMS3 v WP -4.30 7.22E-03 AIMS3 v WP -9.04 2.83E-04 AIMS4 v WP -4.14 4.34E-03 AIMS4 v WP -9.06 1.63E-04 Sedimentitalea AIMS1 v WP 4.43 2.29E-08 AIMS2 v AIMS4 2.12 6.94E-06 AIMS2 v WP 5.69 1.19E-13 AIMS3 v WP 4.28 4.99E-08 AIMS4 v WP 3.56 7.73E-06 Spirochaeta 2 AIMS1 v AIMS2 -10.28 3.55E-38 AIMS1 v AIMS3 2.75 2.06E-02 AIMS1 v AIMS4 -9.80 3.67E-34 AIMS1 v WP -5.73 4.83E-05 AIMS2 v AIMS3 13.03 1.13E-46 AIMS2 v WP 4.55 1.36E-03 AIMS3 v AIMS4 -12.55 5.32E-43 AIMS3 v WP -8.48 2.44E-09 AIMS4 v WP 4.07 3.13E-03 Thalassobius AIMS1 v WP 2.37 8.67E-11 AIMS2 v WP 2.35 7.39E-11 AIMS3 v WP 2.76 9.14E-15 AIMS4 v WP 2.21 9.57E-10

138

The relative abundance of the 20 most abundant order and family-level taxa was generally consistent between the AIMS1–4 genotypes with the exception of Spirochaetales- Spirochaetaceae, which featured strongly in AIMS2 and AIMS4 only. This difference may be related to the different culture histories of AIMS1 and AIMS3 versus AISM2 and AIMS4.

Rhodobacterales Rhodobacteraceae Alteromonadales Alteromonadaceae Chitinophagales Saprospiraceae Flavobacteriales Marinobacteraceae Rhizobiales Flavobacteriaceae Cellvibrionales Halieaceae Myxococcales Cyclobacteriaceae Cytophagales Nannocystaceae Abundance Bdellovibrionales Oligoflexaceae + Oceanospirillales (Chitinophagales) uncultured Oligoflexales Hyphomonadaceae Caulobacterales Sphingomonadaceae Betaproteobacteriales Spirochaetaceae –

Order (20 most abundant) Sphingomonadales Rhizobiaceae Family (20 most abundant) Spirochaetales Bdellovibrionaceae Rhodospirillales Methylophilaceae Planctomycetales Cryomorphaceae Chlamydiales Bacteriovoracaceae Acanthopleuribacterales Stappiaceae Ectothiorhodospirales Rubinisphaeraceae WP WP AIMS1 AIMS2 AIMS3 AIMS4 AIMS1 AIMS2 AIMS3 AIMS4 (a) Sample type (b) Sample type

Figure A 8: Heatmaps of the top 20 AIMS1–4 taxa by relative abundance at (a) order and (b) family level. WP = Wild Proxies.

139

Table A 6: Pairwise binary log fold change Table A 7: Pairwise binary log fold change

(L2FC) for the 20 most abundant (L2FC) for the 20 most abundant order-level taxa. Only results with family-level taxa. Only results with p < 0.05 and LFC > 2.00 are shown. p < 0.05 and LFC > 2.00 are shown.

Order pair L2FC adjusted p-value Family pair L2FC adjusted p-value Acanthopleuribacterales AIMS1 v AIMS4 -2.69 7.40E-03 Alteromonadaceae AIMS2 v WP 2.77 1.09E-07 AIMS1 v WP 3.99 1.60E-02 AIMS4 v WP 2.44 2.20E-06 AIMS2 v WP 5.52 4.73E-04 Cryomorphaceae AIMS1 v WP -3.05 9.19E-05 AIMS3 v WP 4.84 2.83E-03 AIMS2 v WP -2.04 2.16E-02 AIMS4 v WP 6.68 1.62E-05 AIMS3 v WP -2.44 2.04E-03 Alteromonadales AIMS1 v WP 3.26 1.02E-17 AIMS4 v WP -2.78 3.17E-04 AIMS2 v WP 3.75 1.35E-23 Cyclobacteriaceae AIMS1 v WP -2.85 6.18E-03 AIMS3 v WP 3.51 2.44E-20 AIMS2 v WP -4.76 1.19E-06 AIMS4 v WP 3.36 5.78E-19 AIMS3 v WP -3.63 3.83E-04 Betaproteobacteriales AIMS1 v WP 3.17 1.51E-03 AIMS4 v WP -3.47 4.44E-04 AIMS2 v WP 2.97 1.98E-03 Flavobacteriaceae AIMS1 v WP -2.09 2.60E-02 AIMS3 v AIMS4 -2.26 8.38E-05 AIMS2 v WP -2.11 3.35E-02 AIMS4 v WP 3.93 2.58E-05 Halieaceae AIMS1 v WP -2.49 1.49E-03 Chlamydiales AIMS1 v AIMS2 2.03 8.10E-10 AIMS3 v WP -2.51 1.16E-03 Cytophagales AIMS1 v AIMS2 2.06 3.27E-04 Marinobacteraceae AIMS1 v WP 6.48 1.29E-23 AIMS1 v WP -2.12 3.84E-02 AIMS2 v WP 6.76 1.37E-25 AIMS2 v WP -4.18 8.27E-06 AIMS3 v WP 7.16 7.39E-29 AIMS3 v WP -3.13 1.23E-03 AIMS4 v WP 6.46 1.07E-23 AIMS4 v WP -3.17 7.50E-04 Methylophilaceae AIMS1 v WP 9.97 1.68E-13 Ectothiorhodospirales AIMS2 v AIMS4 3.86 2.65E-03 AIMS2 v AIMS4 -2.57 4.04E-07 AIMS3 v AIMS4 3.26 1.43E-02 AIMS2 v WP 8.40 1.97E-09 Oceanospirillales AIMS1 v WP 3.60 5.94E-08 AIMS3 v AIMS4 -2.73 7.00E-08 AIMS2 v WP 3.10 4.94E-06 AIMS3 v WP 8.24 2.19E-09 AIMS3 v WP 3.21 2.05E-06 AIMS4 v WP 10.97 1.67E-16 AIMS4 v WP 3.41 3.94E-07 Nannocystaceae AIMS1 v WP 2.12 4.50E-03 Oligoflexales AIMS1 v WP 2.22 7.12E-03 AIMS3 v WP 2.33 1.34E-03 Rhodospirillales AIMS1 v WP 2.62 1.19E-03 Rhizobiaceae AIMS1 v WP -2.50 2.60E-02 AIMS2 v WP 2.81 2.61E-04 Spirochaetaceae AIMS1 v AIMS2 -9.96 7.42E-40 AIMS3 v WP 2.74 4.15E-04 AIMS1 v AIMS3 3.10 6.44E-03 AIMS4 v WP 3.03 5.75E-05 AIMS1 v AIMS4 -9.22 1.04E-33 Spirochaetales AIMS1 v AIMS2 -9.74 1.39E-41 AIMS1 v WP -5.24 1.15E-04 AIMS1 v AIMS3 3.14 8.90E-03 AIMS2 v AIMS3 13.06 1.47E-50 AIMS1 v AIMS4 -9.09 1.25E-35 AIMS2 v WP 4.72 6.49E-04 AIMS1 v WP -4.59 8.15E-04 AIMS3 v AIMS4 -12.32 1.51E-44 AIMS2 v AIMS3 12.88 1.87E-52 AIMS3 v WP -8.34 1.65E-09 AIMS2 v WP 5.15 4.08E-05 AIMS4 v WP 3.98 2.77E-03 AIMS3 v AIMS4 -12.23 7.08E-47 Stappiaceae AIMS1 v WP 2.23 6.18E-03 AIMS3 v WP -7.73 9.44E-09 AIMS3 v WP 2.19 6.54E-03 AIMS4 v WP 4.50 3.49E-04

140

Table A 8: Output from GLM-based analysis comparing the AIMS1–4 bacterial community compositions (‘tank’ nested within ‘genotype’).

Analysis of Deviance Table Model: manyglm(formula = data ~ data$genotype/data$tank, family = "negative_binomial") Multivariate test: Res.Df Df.diff Dev Pr(>Dev) (Intercept) 69 combo$ genotype 66 3 2434 0.001 *** combo$genotype:combo$tank 58 12 3506 0.001 *** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘. 0.1 ‘ ’ 1

Table A 9: Output from GLM-based analysis comparing the AIMS1–4 bacterial community compositions (‘tank’ and ‘genotype’ specified separately).

Analysis of Deviance Table Model: manyglm(formula = data ~ data$tank + data$genotype, family = "negative_binomial") Multivariate test: Res.Df Df.diff Dev Pr(>Dev) (Intercept) 69 combo$tank 58 11 5941 0.001 *** combo$genotype 58 3 0 0.997 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘. 0.1 ‘ ’ 1

141

Appendix 2

Table A 10: Mock community members. The community was assembled from amplified near- complete 16S rRNA genes from the DNA of pure bacterial cultures. Taxonomic identification of community members was by Sanger sequencing of the 16S rRNA genes, then BLASTing manually trimmed sequences against the NCBI blastn database. Strain codes are in brackets.

Alteromonas lipolytica (MMFS00606) Bacillus pseudofirmus (MMFS00610)

Erythrobacter flavus (MMFS00930) Exiguobacterium aestuarii (MMFS00683)

Labrenzia aggregata (MMFS00933) Leisingera aquamixtae (MMFS00623)

Microbacterium hydrocarbonoxydans (MMFS00614) Micrococcus aloeverae (MMFS00068)

Oceanicola litoreus (MMFS00034) Pelagibaca bermudensis (MMFS00670)

Phaeobacter caeruleus (MMFS00616) Pseudoalteromonas shioyasakiensis (MMFS00019)

Rhodococcus fascians (MMFS006007) Tericoccus solisilvae (MMFS00675)

Thalassomonas ganghwensis (MMFS00626) Vibrio alginolyticus (MMFS00650)

Table A 11: GLM analysis of differences in bacterial community beta diversity based on treatment (control vs heat-treated) and time. Analysis of Deviance Table Model: manyglm(formula = comboMva ~ combo$treatment * combo$time, family = "negative_binomial") Multivariate test: Res.Df Df.diff Dev Pr(>Dev) (Intercept) 95 combo$treatment 94 1 729 0.034 * combo$time 87 7 5414 0.001 *** combo$treatment:time 80 7 2610 0.001 *** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘. 0.1 ‘ ’ 1.

142

Table A 12: Putative contaminant ASVs removed from the dataset.

Relative Relative Abundance in Abundance in anemone samples mock community Phylum Class Order Family Genus samples (%) (%) 1 Proteobacteria Alphaproteobacteria Rhizobiales Xanthobacteraceae Bradyrhizobium 0.0002 0 2 Proteobacteria Alphaproteobacteria Rhizobiales Rhizobiaceae Mesorhizobium <0.0001 0 3 Proteobacteria Alphaproteobacteria Rhizobiales Xanthobacteraceae Afipia 0.0002 0 4 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas 0.0051 0 5 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Pelomonas 0.0003 0 6 Firmicutes Bacilli Lactobacillales Enterococcaceae Enterococcus 0.0025 0 7 Actinobacteria Actinobacteria Propionibacteriales Propionibacteriaceae Cutibacterium 0.0041 0 8 Bacteroidetes Bacteroidia Flavobacteriales Flavobacteriaceae NS3a marine group 0.0074 0 9 Bacteroidetes Bacteroidia Sphingobacteriales Sphingobacteriaceae Sphingobacterium 0.0001 0 10 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Burkholderia 0.0015 0 11 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Ralstonia <0.0001 0 Total (%): 0.0214 Total (%): 0

143

Table A 13: GLM analyses comparing bacterial community beta diversity at each sampling timepoint. Significant values are in bold (α = 0.05). Day Control (°C) Heat-treated (°C) res.df df dev p(>dev) signif

0 26 26 10 1 1 725 0.156 2 26 27 10 1 1 187 0.203 4 26 28 10 1 1 417 0.181 6 26 29 10 1 1 608 0.710 8 26 30 10 1 805 0.147 10 26 31 10 1 1 003 0.103 12 26 32 10 1 1 103 0.052 . 14 26 33 10 1 1 217 0.041 * Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 p-value calculated using 999 sampling iterations.

144

Table A 14: Potential indicator species identified in an IndVal analysis. Heat-treated Day 14 only: anemones: Control vs Heat-treated Day 0 vs Day 14 Phylum Class Order Family Genus IndVal stat p-val IndVal stat p-val 1 Bacteroidetes Bacteroidia Chitinophagales Saprospiraceae 0.8863684 0.003 0.7797039 0.016 2 Proteobacteria Gammaproteobacteria 0.9878049 0.002 0.8766234 0.023 3 Proteobacteria Alphaproteobacteria Rhodospirillales Terasakiellaceae 0.9589579 0.004 0.8659362 0.007 4 Proteobacteria Alphaproteobacteria Rhodospirillales Terasakiellaceae 0.7888006 0.006 0.7036694 0.026 5 Spirochaetes Spirochaetia Spirochaetales Spirochaetaceae Spirochaeta 2 0.9972709 0.003 0.9977034 0.006 6 Proteobacteria Alphaproteobacteria Rhizobiales Rhizobiaceae 0.8244275 0.004 0.6933535 0.012 7 Proteobacteria Deltaproteobacteria Oligoflexales Oligoflexaceae 0.8023567 0.021 0.9661520 0.007 8 Planctomycetes Planctomycetacia Planctomycetales Rubinisphaeraceae 0.7425532 0.034 0.8097484 0.036 9 Bacteroidetes Bacteroidia Chitinophagales Saprospiraceae 0.9940728 0.004 0.9895366 0.003 10 Bacteroidetes Bacteroidia Chitinophagales 0.7589839 0.026 0.7792621 0.019 11 Proteobacteria Deltaproteobacteria Oligoflexales Oligoflexaceae 0.8023567 0.021 0.8052620 0.018 12 Chlamydiae Chlamydiae Chlamydiales Simkaniaceae 0.9413919 0.006 0.7793103 0.020

145

350 tc64 tt03 tc45 tc02

300 tt24 tc61 tc03 tt01 tc66 tt65 tc06 tc44 tt06 tt02 tc04 tc146

250 tc42 tc23 tc24 tc46 tt66 tt125 tc106 tc22 tt41 tt45 tc142 tt21tc65 tc102 tc25 tt43 tt104 tc143 tt46 tt04tc41 tt102 tc26 200 tc124 tc62 tt23 tt44tt25tt63tc05tt105 tt61 tt106 tt82 tt122 tt64 tc103tt81 tt84 tt121tc145tc122 tt146tc126 tt22 tt62 tt103 tc121tc123tc85tt26 tc125 tt145tc63 tc101tt123 tc21tc43 tc83tt144tc82 tc141tt142 Sample ASV s 150 tt86 tt126 tc144 tt42 tc01 tt85 tt124 tc86 tc81 tt83 tt05 tc105tc104 tt101 tt143 tc84 tt141 100

50 te.blank mock3mock2mock1 tneg1 tneg3 tneg2 0

0 10000 20000 30000 40000 Sample reads Figure A 9: Rarefaction curves for all samples. Sub-sampling: 12,000; step: 600.

Guide to sample names

16S samples (e.g. tc125): first letter (t)= thermal stress experiment second letter (c or t) = control or heat-treatment sample first 1-2 digits (0-14) = sampling day 0-14 last digit = replicate no. 1-6 Negative PCR controls (e.g. tneg2): first letter (t) = thermal stress experiment + neg + replicate no. (1-3) Mock community samples (e.g. mock2): mock + replicate no. (1-3) DNA extraction negative control = te.blank

146

100 * Alteromonas (2.99%) Anaerobacillus (3.30%) Arthrobacter (0.40%) * Bacillus (6.97%) 75 * Erythrobacter (10.46%) * Exiguobacterium (5.81%) * Labrenzia (6.14%) * Leisingera (12.56%)

undance (%) 50 * Microbacterium (7.72%) e a b * Micrococcus (1.80%)

Relati v * Pseudoalteromonas (7.45%) * Rhodococcus (6.17%) 25 Thalassobius (2.76%) Thalassotalea (8.55%) * Vibrio (7.83%)

0 Unidentified (9.09%) mock1 mock2 mock3

Figure A 10: Relative abundance of reads assigned to genus in each mock community sample. Asterisks indicate genera present in the original mock community. Numbers in brackets are average relative abundances. ‘Unidentified’ contained three ASVs not identified to genus level.

Control anemones Heat-treated anemones Control anemones Heat-treated anemones

Day 459 279 498 Day Day 309 200 164 Day 0 0 14 14 4.0 96.0 96.1 3.9 5.1 94.9 94.0 6.0

(a) (b)

Figure A 11: Unique and common bacterial ASVs in control and heat-treated anemones at Day 0 or 14. Inset numbers indicate relative abundance (%) for each sample type. (a) Bacterial richness was highly similar at Day 0, with a high overlap in shared ASVs. (b) Richness in both sample types declined throughout the treatment period, with the heat-treated anemones undergoing a more substantial drop by Day 14. However, the number of shared ASVs remained high.

147

Saprospiraceae ASV 0.25 Gammaproteobacteria ASV 15 34 34

0.20

32 Temp (°C) 32 Temp (°C) 10 0.15 30 30 0.10 5 28 28 0.05 26 26 Relative Abundance (%) Relative Abundance (%) 0 0.00 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (a) Day (b) Day

Terasakiellaceae ASV 4 Terasakiellaceae ASV 2.0 34 34

Temp (°C) 3 Temp (°C) 1.5 32 32

30 30 1.0 2

28 28 0.5 1 26 26 Relative Abundance (%) Relative Abundance (%) 0.0 0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (c) Day (d) Day

Spirochaeta 2 ASV 1.5 Rhizobiaceae ASV 34 34

10 • Control Heat-treated 32 Temp (°C) 32 Temp (°C) • 1.0

30 30

5 28 0.5 28

26 26 Relative Abundance (%) Relative Abundance (%) 0 0.0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (e) Day (f) Day

Figure A 12: ASVs identified by IndVal and considered potential indicator species: (a) Saprospiraceae, (b) Gammaproteobacteria, (c) Terasakiellaceae, (d) Terasakiellaceae, (e) Spirochaeta 2, and (f) Rhizobiaceae. For each datapoint, n = 6. Error bars ± 1SEM.

148

Oligoflexaceae ASV Rubinisphaeraceae ASV 34 34 2.0 4 32 Temp (°C) 32 Temp (°C) 1.5 30 30 1.0 2 28 28 0.5 26 26 Relative Abundance (%) Relative Abundance (%) 0 0.0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (a) Day (b) Day

Saprospiraceae ASV Chitinophagales ASV 34 34

2 1.5

32 Temp (°C) 32 Temp (°C)

30 1.0 30

1 28 28 0.5

26 26 Relative Abundance (%) Relative Abundance (%) 0 0.0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (c) Day (d) Day

1.25 Oligoflexaceae ASV Control Simkaniaceae ASV • 34 34 • Heat-treated 0.6 1.00

32 Temp (°C) 32 Temp (°C)

0.75 0.4 30 30 0.50 28 0.2 28 0.25 26 26 Relative Abundance (%) Relative Abundance (%) 0.00 0.0 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 (e) Day (f) Day

Figure A 13: ASVs identified by IndVal but discounted as potential indicator species due to erratic changes in relative abundance and/or high variance. Simkaniaceae was discounted due to the large difference in relative abundance at Day 0 compared to all other timepoints: (a) Oligoflexaceae, (b) Rubinisphaeraceae, (c) Saprospiraceae, (d) Chitinophagales, (e) Oligoflexaceae, and (f) Simkaniaceae. For each datapoint, n = 6. Error bars ±1SEM.

149

Appendix 3

Table A 15: Overall differences in Symbiodiniaceae cell density between control and treated anemones tested by a GLS model. Significant values are in bold. Df Chisq Pr(>Chisq) treatment 1 144.451 < 2.2e-16 *** day 1 37.760 8.001e-10 *** treatment:day 1 8.137 0.004338 ** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Anemone cell densities were significantly different by treatment and day. However, there was also a significant interaction. The data were checked for homogeneity of variance and normality before performing paired tests to determine where differences within and between the datasets were significant.

Table A 16: Homogeneity of variance and normality test p-values for algal cell densities. Day 0 1 3 7 14 21 Levene’s (control vs treated) 0.402 0.136 0.675 0.421 0.900 0.436 Shapiro-Wilk: control 0.113 0.281 0.368 0.997 0.947 0.224 Shapiro-Wilk: treated 0.028 0.031 0.155 0.404 0.164 0.946

As some data were not normally distributed, comparisons were performed using Mann- Whitney U tests.

Table A 17: Mann-Whitney U test p-values of control vs treated anemone algal cell density at each sampling timepoint. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 p-value 0.074 0.910 0.012 0.074 0.004 0.004

Table A 18: Mann-Whitney U test p-values for differences in control and treated anemone algal cell density for Day 0 versus Day 21. Significant values are in bold (α = 0.05). Days –group 0-21 – control 0-21 – treated p-value 0.250 0.004

150

In visualisations of the ddPCR output produced in QuantaSoft, a decrease over time in the number of positive droplets for the bacterial samples of the treated anemones (left to right in Figure A 14) illustrated a decrease in bacteria with ongoing treatment. Despite a moderate amount of rain in many of the bacterial ddPCRs, QuantaSoft automatically delineated negative and positive droplets in all but one reaction, for which a manual threshold was set.

Figure A 14: Example of ddPCR fluorescence output for four paired host-bacteria reactions:

Lane Timepoint (Day) Target/Primers A04 0 E. diaphana/Ef1-α H04 0 Bacteria/16S A05 3 E. diaphana/Ef1-α H05 3 Bacteria/16S A06 7 E. diaphana/Ef1-α H06 7 Bacteria/16S A07 14 E. diaphana/Ef1-α H07 14 Bacteria/16S

151

Table A 19: ddPCR data for the bacterial/16S reactions. Copies/Well calculation based on 20 µL reactions volumes. No correction for the 24 µL reaction volumes was made as the data was used to calculate ratios only. NTC = no template control.

Sample Condition Day Sample Type Concentration Copies/Well Positives Negatives Accepted Droplets

NTC – – – 6.2 124 75 14197 14272

gc01 control 0 anemone 44.8 896 549 14146 14695

gc02 control 0 anemone 34.7 694 358 11964 12322

gc03 control 0 anemone 98.0 1960 1092 12575 13667

gc04 control 0 anemone 17.6 352 197 13074 13271

gc05 control 0 anemone 213.0 4260 2338 11783 14121

gc06 control 0 anemone 662.0 13240 5928 7845 13773

gc11 control 1 anemone 104.0 2080 1254 13520 14774

gc12 control 1 anemone 42.7 854 506 13696 14202

gc13 control 1 anemone 51.2 1024 607 13658 14265

gc14 control 1 anemone 160.0 3200 1723 11827 13550

gc15 control 1 anemone 62.0 1240 804 14864 15668

gc16 control 1 anemone 59.2 1184 723 13999 14722

gc31 control 3 anemone 238.0 4760 2609 11647 14256

gc32 control 3 anemone 499.0 9980 5057 9583 14640

gc33 control 3 anemone 80.0 1600 944 13364 14308

gc34 control 3 anemone 88.0 1760 1049 13522 14571

gc35 control 3 anemone 144.0 2880 1706 13122 14828

gc36 control 3 anemone 88.0 1760 1141 14734 15875

gc71 control 7 anemone 340.0 6800 3548 10597 14145

gc72 control 7 anemone 294.0 5880 3015 10628 13643

gc73 control 7 anemone 157.0 3140 1777 12455 14232

gc74 control 7 anemone 81.3 1626 1067 14918 15985

gc75 control 7 anemone 49.7 994 624 14458 15082

gc76 control 7 anemone 90.0 1800 1082 13536 14618

gc141 control 14 anemone 176.0 3520 1912 11856 13768

gc142 control 14 anemone 512.0 10240 4993 9152 14145

gc143 control 14 anemone 518.0 10360 4535 8198 12733

gc144 control 14 anemone 1632.0 32640 10688 3560 14248

gc145 control 14 anemone 93.0 1860 1087 13241 14328

gc146 control 14 anemone 110.0 2200 1306 13277 14583

gc211 control 21 anemone 505.0 10100 4700 8766 13466

gc212 control 21 anemone 1494.0 29880 10720 4188 14908

gc213 control 21 anemone 1008.0 20160 7987 5895 13882

gc214 control 21 anemone 287.0 5740 3153 11407 14560

gc215 control 21 anemone 771.0 15420 6969 7534 14503

gc216 control 21 anemone 175.0 3500 1923 12023 13946 152

gt01 treated 0 anemone 139.0 2780 1734 13837 15571 gt02 treated 0 anemone 33.8 676 388 13314 13702 gt03 treated 0 anemone 303.0 6060 3052 10379 13431 gt04 treated 0 anemone 28.0 560 355 14762 15117 gt05 treated 0 anemone 140.0 2800 1655 13058 14713 gt06 treated 0 anemone 26.3 526 342 15103 15445 gt11 treated 1 anemone 12.3 246 168 16003 16171 gt12 treated 1 anemone 22.2 444 286 15045 15331 gt13 treated 1 anemone 112.0 2240 1495 14999 16494 gt14 treated 1 anemone 34.8 696 448 14936 15384 gt15 treated 1 anemone 72.8 1456 954 14948 15902 gt16 treated 1 anemone 29.4 588 350 13833 14183 gt31 treated 3 anemone 59.7 1194 730 14023 14753 gt32 treated 3 anemone 67.6 1352 853 14424 15277 gt33 treated 3 anemone 87.5 1743 767 10741 10741 gt34 treated 3 anemone 34.3 686 444 14992 15436 gt35 treated 3 anemone 36.7 734 472 14898 15370 gt36 treated 3 anemone 52.0 1040 654 14479 15133 gt71 treated 7 anemone 16.8 336 195 13556 13751 gt72 treated 7 anemone 25.6 512 328 14938 15266 gt73 treated 7 anemone 9.2 184 102 13003 13105 gt74 treated 7 anemone 16.4 328 212 15070 15282 gt75 treated 7 anemone 37.5 750 444 13708 14152 gt76 treated 7 anemone 17.6 352 198 13106 13304 gt141 treated 14 anemone 16.7 334 221 15432 15653 gt142 treated 14 anemone 31.5 630 365 13442 13807 gt143 treated 14 anemone 22l5 450 286 14840 15126 gt144 treated 14 anemone 28.5 570 280 11400 11680 gt145 treated 14 anemone 20.2 404 236 13640 13876 gt146 treated 14 anemone 18.4 368 237 15023 15260 gt211 treated 21 anemone 11.5 230 141 14333 14474 gt212 treated 21 anemone 24.3 486 347 16597 16944 gt213 treated 21 anemone 52.4 1048 675 14814 15489 gt214 treated 21 anemone 11.7 234 131 13075 13206 gt215 treated 21 anemone 4.4 88 55 14582 14637 gt216 treated 21 anemone 14.3 286 190 15501 15691 gca1 control – artemia 8.3 166 133 15916 16029 gca2 control – artemia 6.7 134 85 14913 14998 gca3 control – artemia 5.6 112 64 13432 13496 gta1 treated – artemia 4.9 98 58 13909 13967 gta2 treated – artemia 4.5 90 55 14351 14406 gta3 treated – artemia 3.0 60 22 8498 8520 153

Table A 20: ddPCR data for the E. diaphana/Ef1-α reactions. Copies/Well calculation based on 20 µL reactions volumes. No correction for the actual 24 µL reaction volumes was made as the data was used to calculate ratios only. NTC = no template control.

Sample Condition Day Sample Type Concentration Copies/20uL Well Positives Negatives Accepted Droplets

NTC – – – No Call 0 0 15301 15301

gc01 control 0 anemone 1275 25500 9198 4702 13900

gc02 control 0 anemone 1246 24920 10187 5405 15592

gc03 control 0 anemone 1734 34680 10684 3174 13858

gc04 control 0 anemone 213 4260 2142 10779 12921

gc05 control 0 anemone 1802 36040 11070 3053 14123

gc06 control 0 anemone 1874 37480 11814 3015 14829

gc11 control 1 anemone 1640 32800 10874 3588 14462

gc12 control 1 anemone 1707 34140 10020 3067 13087

gc13 control 1 anemone 1158 23160 9635 5746 15381

gc14 control 1 anemone 1566 31320 10402 3735 14137

gc15 control 1 anemone 1022 20440 8331 6023 14354

gc16 control 1 anemone 1392 27840 9419 4160 13579

gc31 control 3 anemone 2205 44100 12177 2208 14385

gc32 control 3 anemone 1494 29880 10217 3990 14207

gc33 control 3 anemone 2940 58800 13088 1175 14263

gc34 control 3 anemone 2870 57400 12874 1233 14107

gc35 control 3 anemone 1675 33500 10993 3487 14480

gc36 control 3 anemone 1892 37840 11193 2804 13997

gc71 control 7 anemone 1293 25860 8618 4305 12923

gc72 control 7 anemone 2092 41840 12258 2491 14749

gc73 control 7 anemone 1549 30980 10555 3866 14421

gc74 control 7 anemone 1261 25220 8869 4616 13485

gc75 control 7 anemone 1924 38480 12306 2979 15285

gc76 control 7 anemone 2540 50800 12420 1628 14048

gc141 control 14 anemone 1640 32800 10313 3401 13714

gc142 control 14 anemone 945 18900 7437 6030 13467

gc143 control 14 anemone 1805 36100 10229 2813 13042

gc144 control 14 anemone 2730 54600 11338 1231 12569

gc145 control 14 anemone 2172 43440 12293 2304 14597

gc146 control 14 anemone 1556 31120 10149 3686 13835

gc211 control 21 anemone 1632 32640 9766 3251 13071

gc212 control 21 anemone 1686 33720 10474 3283 13757

gc213 control 21 anemone 1491 29820 9425 3696 13121

gc214 control 21 anemone 3030 60600 13406 1108 14514

gc215 control 21 anemone 1383 27660 9471 4227 13698

gc216 control 21 anemone 1872 37440 10984 2809 13793 154

gt01 treated 0 anemone 1536 30720 10638 3954 14592 gt02 treated 0 anemone 952 19040 7571 6072 13643 gt03 treated 0 anemone 1543 30860 10915 4024 14939 gt04 treated 0 anemone 1120 22400 8954 5625 14579 gt05 treated 0 anemone 868 17360 7323 6709 14032 gt06 treated 0 anemone 1192 23840 10528 6003 16531 gt11 treated 1 anemone 277 5540 3509 13200 16709 gt12 treated 1 anemone 1949 38980 11327 2669 13996 gt13 treated 1 anemone 1884 37680 12694 3205 15899 gt14 treated 1 anemone 1415 28300 10027 4306 14333 gt15 treated 1 anemone 1471 29420 11946 4793 16739 gt16 treated 1 anemone 1372 27440 9770 4423 14193 gt31 treated 3 anemone 1237 24740 9339 5019 14358 gt32 treated 3 anemone 4180 83600 14192 419 14611 gt33 treated 3 anemone 1690 33800 10835 3379 14214 gt34 treated 3 anemone 1967 39340 11679 2702 14381 gt35 treated 3 anemone 2181 43620 12464 2315 14779 gt36 treated 3 anemone 1659 33180 11222 3626 14848 gt71 treated 7 anemone 1551 31020 10465 3825 14290 gt72 treated 7 anemone 1484 29680 10988 4341 15329 gt73 treated 7 anemone 1275 25500 9010 4607 13617 gt74 treated 7 anemone 1367 27340 9945 4530 14475 gt75 treated 7 anemone 2049 40980 11449 2432 13881 gt76 treated 7 anemone 1805 36100 11161 3069 14230 gt141 treated 14 anemone 1310 26200 9333 4561 13894 gt142 treated 14 anemone 1308 26160 9622 4715 14337 gt143 treated 14 anemone 2580 51600 13600 1702 15302 gt144 treated 14 anemone 1811 36220 12124 3310 15434 gt145 treated 14 anemone 1779 35580 11440 3235 14675 gt146 treated 14 anemone 1333 26660 9536 4532 14068 gt211 treated 21 anemone 1045 20900 8216 5740 13956 gt212 treated 21 anemone 1795 35900 11126 3092 14218 gt213 treated 21 anemone 1036 20720 8776 6213 14989 gt214 treated 21 anemone 1309 26180 8027 3931 11958 gt215 treated 21 anemone 142 2840 1707 13270 14977 gt216 treated 21 anemone 1534 30680 10787 4021 14808 gca1 control – artemia 62 1244 848 15612 16460 gca2 control – artemia 84 1680 1005 13560 14565 gca3 control – artemia 83 1660 316 11201 12017 gta1 treated – artemia 167 3340 1902 12439 14341 gta2 treated – artemia 165 3300 1915 12735 14650 gta3 treated – artemia 155 3100 1588 11242 12830 155

Table A 21: Overall differences in bacterial load ratios between control and treated anemones tested by a GLS model. Df Chisq Pr(>Chisq) day 1 3.4827e+09 < 2.2e-16 *** treatment 1 3.4089e+11 < 2.2e-16 *** day:treatment 1 1.1998e+10 < 2.2e-16 *** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

There were significant differences in bacterial load across time and by treatment. However, there was also a significant interaction. The data were checked for homogeneity of variance and normality before performing paired tests to determine whether differences within and between the control and treated data were significant.

Table A 22: Homogeneity of variance and normality test p-values for bacterial load ratio data. Significant values are in bold (α = 0.05).

Day 0 1 3 7 14 21 Levene’s (control vs treated) 0.686 0.682 0.180 0.031 0.005 0.001 Shapiro-Wilk: control 0.014 0.572 0.013 0.291 0.163 0.503 Shapiro-Wilk: treated 0.448 0.665 0.067 0.216 0.346 0.035

Homogeneity of variance and normality assumptions were not met for some data. Therefore, comparisons were performed using Mann-Whitney U tests.

Table A 23: Mann-Whitney U test p-values of control vs treated bacterial load ratios at each sampling timepoint. Significant values are in bold (α = 0.05). Day 0 1 3 7 14 21 p-value 1.000 0.156 0.156 0.031 0.031 0.031

Table A 24: Mann-Whitney U test p-values for differences in bacterial load ratios within the control and treated anemones at selected timepoints. Significant values are in bold (α = 0.05). Days – group 0-1 – treated 0-21 – treated 0-1 – control 0-21 – control p-value 0.031 0.031 0.563 0.094

Somewhat surprisingly, the control Day 0 vs Day 21 comparison was not significant. This was possibly due to high variance in the data. 156

The A. salina bacterial load ratio data were checked for homogeneity of variance and normality before testing to determine whether the difference between the control and treated data was significant. As the data met all parametric assumptions, it was analysed using a two-sample Student’s t-test.

Table A 25: Analysis of statistical difference between untreated and treated A. salina. Significant values are in bold (α = 0.05). Test: Comparison p-value Levene’s: Control vs Treated 0.312 Shapiro-Wilk: Control 0.357 Shapiro-Wilk: Treated 0.363 Student’s t-test: Control vs Treated 0.029

157

gt212 gt216

400 gt213 gt144 gt142 gt214 gt211 300 gt215 gt03gt11 gt76 gt143 gta2 gt34 gc02 gt33 gc216 gc145 gt74 gt146

200 gt35 gc215 gt12 gc214gt15 gc05 gt141gt16 gt13gt32gc04gca1 gc15 gc03gc12gc74 gt01gt02gt04gc143gt36gc01 gc142 gc211gt73 gt71gt75 gc14gt72gt145gc146gc76gt14gc212gc36gt31 gc141 gta3 gc34gc33gc71gc213gc72gt05gc35 gt06 Sample ASVs gc144gc13 gc32gc16gc11gc31 gta1 gca2gc73 gc06 100 gc75 ge.blank gca3 gneg1gneg3gneg2 0

0 10000 20000 30000 40000 50000 Sample reads Figure A 15: Rarefaction curves for all samples (sub-sampling level: 12 000; step: 600).

Curves for most samples plateaued, indicating that sequencing depth was sufficient to capture bacterial diversity. One A. salina control sample, gca3, contained relatively few reads (2 417) and, consequently, a low number of ASVs (33). However, this sample was retained in the dataset as A. salina were not the focus of the study.

Table A 26: Contaminant ASVs removed from the dataset. Eight ASVs potentially introduced during sample processing were identified with the R package decontam using data from the negative control samples. All were removed from the dataset.

Phylum Class Order Family Genus Anemones A. salina Rel. ab. Rel. ab. (%) (%) 1 Proteobacteria Alphaproteobacteria Rhizobiales Xanthobacteraceae Bradyrhizobium 0.0036 0.0000

2 Proteobacteria Alphaproteobacteria Rhizobiales Rhizobiaceae Mesorhizobium 0.0006 0.0000

3 Proteobacteria Alphaproteobacteria Rhizobiales Xanthobacteraceae Afipia 0.0081 0.0051

4 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Pelomonas 0.0032 0.0000

5 Actinobacteria Actinobacteria Corynebacteriales Corynebacteriaceae Corynebacterium 0.0020 0.0000

6 Bacteroidetes Bacteroidia Sphingobacteriales Sphingobacteriaceae Sphingobacterium 0.0003 0.0000

7 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Ralstonia 0.0023 0.0000

8 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Ralstonia 0.0555 0.0000

0.0756 0.0051

158

gt11 gt03

Group Control Treated 0

Day 0 NMDS2 1 3 −1 7 14 21

−2

2D stress = 0.16

−1 0 1 2 3 NMDS1 Figure A 16: nMDS (weighted unifrac) of all samples. Two outlier samples (highlighted) were identified and removed from the dataset.

159

Table A 27: Analysis of overall differences in number of observed ASVs between control and treated anemones tested by a GLS model. Significant values are in bold. Df Chisq Pr(>Chisq) treatment 1 54.358 1.671e-13 *** day 1 36.885 1.253e-09 *** treatment:day 1 30.603 3.165e-08 *** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

The anemones were significantly different by treatment and day, but there was a significant interaction between the variables. The data were checked for homogeneity of variance and normality before performing paired tests to determine where significant differences occurred between and within the datasets.

Table A 28: Homogeneity of variance and normality test p-values for number of observed ASVs. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 Levene’s (control vs treated) 0.369 0.164 0.006 0.592 0.024 0.005 Shapiro-Wilk: control 0.641 0.246 0.565 0.807 0.224 0.202 Shapiro-Wilk: treated 0.100 0.308 0.429 0.014 0.600 0.165

The variance and distribution of some data were not homogeneous or normal. Sample sizes also differed at Days 1 and 0. Therefore, comparisons were performed using non-parametric two-sample Kolmogorov-Smirnov tests.

Table A 29: Kolmogorov-Smirnov test p-values for number of observed ASVs in control vs treated anemones at each sampling timepoint. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 p-value 0.503 0.178 0.002 0.031 0.139 0.002

Table A 30: Kolmogorov-Smirnov test p-values for number of observed ASVs in control and treated anemones for day 0 versus day 21. Significant values are in bold (α = 0.05). Days –group 0-21 – control 0-21 – treated p-value 0.931 0.004

160

Table A 31: Analysis of overall differences in Simpson index values between control and treated anemones tested by a GLS model. Df Chisq Pr(>Chisq) treatment 1 3.9497 0.04688 * day 1 3.3648 0.06660 . treatment:day 1 5.2239 0.02228 * Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

The anemones were significantly different by treatment. There was also a significant interaction between the variables. The data were checked for homogeneity of variance and normality before performing paired tests to determine where significant differences occurred between and within the datasets.

Table A 32: Homogeneity of variance and normality test p-values for Simpson Index values. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 Levene’s (control vs treated) 0.206 0.581 0.103 0.206 0.301 0.828 Shapiro-Wilk: control 0.776 0.217 0.921 0.007 0.262 0.008 Shapiro-Wilk: treated 0.174 0.076 0.142 0.661 0.084 0.372

The distribution of some data was not normal. Sample sizes also differed at Days 1 and 0. Therefore, comparisons were performed using non-parametric Kolmogorov-Smirnov tests, but despite the GLS results, no significant differences were detected in the subsequent tests.

Table A 33: Kolmogorov-Smirnov test p-values for Simpson Index values in control vs treated anemones at each sampling timepoint. Day 0 1 3 7 14 21 p-value 0.238 0.688 0.143 0.143 0.474 0.474

Table A 34: Kolmogorov-Smirnov test p-values for number of Simpson Index values in control and treated anemones for Day 0 versus Day 21. Days –group 0-21 – control 0-21 – treated p-value 0.474 0.238

161

Table A 35: Analysis of overall differences in Shannon index values between control and treated anemones tested by a GLS model. Significant values are in bold. Df Chisq Pr(>Chisq) treatment 1 0.0423 0.837012 day 1 7.2137 0.007235 ** treatment:day 1 9.6863 0.001856 ** Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

The samples did not differ significantly based on treatment, but there was a significant difference over time and a significant interaction between the variables. The data were checked for homogeneity of variance and normality before performing paired tests to determine where significant differences occurred between and within the datasets.

Table A 36: Homogeneity of variance and normality test p-values for Shannon index values. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 Levene’s (control vs treated) 0.541 0.871 0.012 0.565 0.455 0.403 Shapiro-Wilk: control 0.736 0.148 0.169 0.392 0.906 0.177 Shapiro-Wilk: treated 0.044 0.516 0.204 0.007 0.775 0.597

Table A 37: Kolmogorov-Smirnov test p-values for Shannon index values in control vs treated anemones at each sampling timepoint. Significant p-values are in bold (α = 0.05). Day 0 1 3 7 14 21 p-value 0.238 0.688 0.474 0.474 0.931 0.026

Table A 38: Kolmogorov-Smirnov test p-values for Shannon index values in control and treated anemones for day 0 versus day 21. Significant p-values are in bold (α = 0.05) Days –group 0-21 – control 0-21 – treated p-value 0.474 0.004

162

Table A 39: Differences in beta diversity between control and antibiotic-treated E. diaphana bacterial communities tested by Generalised Linear Models. Comparison res.df df dev p(>dev) signif. Day 0 control vs Day 0 treated 10 1 431.5 0.200 Day 0 control vs Day 21 control 10 1 1 001 0.006 ** Day 0 treated vs Day 21 treated 10 1 1 312 0.003 ** Day 21 control vs Day 21 treated 10 1 1 432 0.002 ** Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 p-value calculated using 999 sampling iterations.

163

Appendix 4

Isolation and identification of candidate probiotic bacteria

Great Barrier Reef-origin Exaiptasia diaphana were taken from the University of Melbourne anemone culture collection (Dungan et al. 2020) as source material for the probiotic bacteria. Sixteen replicates from each of four E. diaphana genotypes (AIMS1-4) were collected using sterile pipettes and transferred to sterile 12-well plates (CLS3513, Corning, USA) filled with 0.2 µm-filtered Red Sea Salt™ (RSS) seawater reconstituted with reverse osmosis water at ~34 parts per thousand (ppt), hereafter ‘fRSS’. The anemones were left for 30 min to remove loosely associated bacteria, then transferred into sterile glass Dounce homogenizers filled with 1 mL of fRSS and ground into a slurry. Serial dilutions were created from each homogenate at 10-1, 10-2, 10-3 and 10-4. From each dilution, 50 µL was spread onto three marine agar plates (PP2315, Thermo Fisher, Australia) and three R2A agar plates (CM0906, Thermo Fisher, Australia) made with 40 gL-1 RSS. All plates were incubated at the anemone rearing temperature, 26 °C. After one-week of incubation, cells were picked from individual colonies on plates with <100 colony forming units, and sub-cultured to purification on media consistent with their original growth. Each resulting isolate was suspended in 40% glycerol and aliquoted into 1.2 mL cryotubes for storage at –80 °C.

Each bacterial culture was taxonomically identified by Sanger sequencing of the 16S rRNA genes. DNA for sequencing was prepared as follows: cells from each pure culture were suspended in 20 µL sterile H2O and denatured at 95 °C for 10 min. Each suspension was then centrifuged at 2,000 × g at 4 °C for 2 min. PCRs were set up using 2 µL of the resulting supernatant as template in 40 µL reactions also containing 0.5 U Mango Mix™ (Bioline, Australia), and 0.25 µM of universal bacterial primers 27f (5’ AGAGTTTGATCMTGGCTCAG 3’) and 1492r (5’ TACGGYTACCTTGTTACGACTT 3’) (Lane 1991). PCR thermal cycler settings were: 1 × 95°C for 5 min; 35 × 95 °C for 1 min, 50 °C for 1 min, and 72 °C for 1 min; 1 × 72 °C for 10 min. The PCR products were then purified and sequenced on an ABI sequencing instrument at Macrogen Inc. (Seoul, South Korea) using the 1492r primer. The sequence data for each isolate was trimmed and aligned against sequences in the NCBI GenBank database with BLAST (Altschul et al. 1990) to establish taxonomic identity.

164

Candidate selection by qualitative free radical scavenging (FRS) assay

To qualitatively assess isolate free radical scavenging (FRS) ability, each isolate was grown on solid media, as described above, then incubated overnight at 26 °C with a sterile Whatman #1 filter in contact with the bacterial colonies. The filters were removed with forceps and allowed to dry for 30 min, then wet with 500 µL of a 0.2 mM DPPH-MetOH (D9132, Sigma-Aldrich, Australia) solution (Figure 1). DPPH is a stable free radical that it is purple in its oxidized state but becomes white-yellow when reduced by antioxidants and has been used to identify antioxidant marine bacteria (Takao et al. 1994; Velho-Pereira, Parvatkar & Furtado 2015). Several drops of 0.1% (w/v) L-ascorbic acid (A7631, Sigma-Aldrich, Australia) were also applied to a filter paper and tested as a positive control. The response of each isolate to the DPPH solution was recorded for 3 min after application. Appearance of a white-yellow halo around individual colonies within 1 min was judged a positive response (Figure A 17a). Appearance of a halo after 1-3 min was judged a weak positive response. Failure to form a halo was judged a negative response (Figure A 17b).

(a) (b)

Figure A 17: DPPH qualitative assay results: (a) Positive indication of FRS ability for a bacterial isolate due to appearance of halo in <1 min; (b) Negative indication of FRS ability for a bacterial isolate due to absence of halo after >3 min.

Candidate selection by quantitative free radical scavenging (FRS) assay

To quantitatively assess isolate FRS ability, each was grown on sold media, as described above. From each isolate, cells were picked from single colonies and inoculated into triplicate 250 mL conical flasks containing 50 mL sterile R2A broth (Table A 40). The flasks were incubated at 37 °C for 48 hrs at 150 rpm in an orbital incubator (OM11, Ratek, Australia). Uninoculated

165

broth was simultaneously incubated to confirm the absence of media contamination. After 48 hrs, OD600 measurements were collected from each culture and media blank (CLARIOstar PLUS plate-reader, BMG Labtech, Australia). The cultures were then centrifuged at 3000 × g at 4 °C for 30 min to pellet the bacterial cells. The cell free supernatant (CFS) was collected, frozen at –80 ˚C, and freeze-dried (Alpha 1-4 LDplus, Martin Christ, Germany). The freeze- dried CFS was kept under inert gas in darkness until analysis. Antioxidants from the CFS were extracted by suspending in 100% MetOH to a concentration of 50 mg/mL, sonicating for 5 min, and then centrifuging at 3000 x g for 5 min at 4˚C. Quantitative DPPH assays were then performed on each CFS extract mixed with 0.2 mM DPPH-MetOH in a 1:1 ratio to a volume of 1 mL. Each sample was vortexed and incubated for 30 min at room temperature, then vortexed again briefly. Three 300 µL replicates of each sample were transferred to a 96 well plate, and absorbance was measured at 517 nm. Percentage FRS was calculated according to the following formula, where control = no-sample 0.2 mM DPPH-MetOH:

% FRS activity = (control – sample) / control × 100

Reaction blanks of 100% methanol, and positive controls of 0.01%–0.001% (w/v) L-ascorbic acid were included in each plate.

Table A 40: R2A broth adjusted to suit marine bacteria. Final pH = 7.2 +/- 0.2 at 25 °C. Component grams L–1 Supplier

Casein acid hydrolysate 0.500 C0501, Sigma Aldrich, Australia

Yeast extract 0.500 LP0021, Oxoid, Thermo Fisher, Australia

Proteose peptone 0.500 211684, Thermo Fisher, Australia

Dextrose 0.500 G360, Austratec, Australia

Starch, soluble 0.500 AJA526, Univar, Australia

Dipotassium phosphate 0.300 P3786, Sigma Aldrich, Australia

Magnesium sulfate 0.024 M2643, Sigma Aldrich, Australia

Sodium pyruvate 0.300 P2256, Sigma Aldrich, Australia

Red Sea Salt™ 40.00 R11065, Red Sea, USA

166

Click and drag in the plot area to zoom in

Zoom 1m 3m 6m YTD 1y All

From Jan 11, 2016 To Feb 11, 2016 W a t 30 e r

T e m p e r

28 a t u r e

26 18-01-2016 25-01-2016 01-02-2016 08-02-2016

Figure A 18: Water temperature daily averages (2 m depth) from St Crispin Reef (AIMS 2019). The graph is illustrative for summer2016 heatwave conditions and201 8shows an increase in SST for a northern part of the GBR in summer 2016; a year that saw “unprecedented” mass coral bleaching (Hughes et al. 2017b).

167

Table A 41: Fv/Fm comparisons for AIMS2 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 12 3.450 0.0048 Negative 12 3.069 0.0097 Positive 12 2.655 0.0210 Day 43: Control-Positive Elevated 12 1.208 0.4709 Day 43: Control-Negative Elevated 12 1.336 0.4035 Day 43: Positive-Negative Elevated 12 0.127 0.9911

Table A 42: Fv/Fm comparisons for AIMS3 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 9 3.998 0.0031 Negative 9 4.664 0.0012 Positive 9 2.946 0.0163 Day 43: Control-Positive Elevated 9 0.238 0.9695 Day 43: Control-Negative Elevated 9 0.160 0.9861 Day 43: Positive-Negative Elevated 9 0.420 0.9083

Table A 43: Fv/Fm comparisons for AIMS4 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 11 2.021 0.0683 Negative 11 2.352 0.0384 Positive 11 1.982 0.0730 Day 43: Control-Positive Elevated 12 0.468 0.8873 Day 43: Control-Negative Elevated 12 0.203 0.9776 Day 43: Positive-Negative Elevated 12 0.750 0.7396

168

Table A 44: Symbiodiniaceae cell density comparisons for AIMS2 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 12 2.123 0.0552 Negative 12 0.588 0.5672 Positive 12 1.723 0.1106 Day 43: Control-Positive Elevated 12 0.686 0.7757 Day 43: Control-Negative Elevated 12 0.562 0.8422 Day 43: Positive-Negative Elevated 12 0.124 0.9916

Table A 45: Symbiodiniaceae cell density comparisons for AIMS3 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 12 7.680 <.0001 Negative 12 7.795 <.0001 Positive 12 7.202 <.0001 Day 43: Control-Positive Elevated 12 1.774 0.2194 Day 43: Control-Negative Elevated 12 0.781 0.7213 Day 43: Positive-Negative Elevated 12 0.993 0.5947

Table A 46: Symbiodiniaceae cell density comparisons for AIMS4 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 12 0.960 0.3558 Negative 12 2.287 0.0412 Positive 12 1.447 0.1735 Day 43: Control-Positive Elevated 12 1.602 0.2824 Day 43: Control-Negative Elevated 12 1.876 0.1879 Day 43: Positive-Negative Elevated 12 0.274 0.9596

169

Table A 47: ROS comparisons for AIMS2 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 11 0.923 0.3756 Negative 11 0.782 0.4507 Positive 11 0.474 0.6449 Day 43: Control-Positive Elevated 11 0.353 0.9340 Day 43: Control-Negative Elevated 11 0.305 0.9501 Day 43: Positive-Negative Elevated 11 0.053 0.9984

Table A 48: ROS comparisons for AIMS3 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 11 1.265 0.2320 Negative 11 1.664 0.1244 Positive 11 2.090 0.0606 Day 43: Control-Positive Elevated 11 0.120 0.9921 Day 43: Control-Negative Elevated 11 0.369 0.9283 Day 43: Positive-Negative Elevated 11 0.278 0.9584

Table A 49: ROS comparisons for AIMS4 anemones at Day 43. Significant values are in bold; α = 0.05. Data: Contrast Treatment df t.ratio p.value Day 43: Ambient-Elevated Control 9 0.670 0.5195 Negative 9 1.215 0.2554 Positive 9 1.685 0.1263 Day 43: Control-Positive Elevated 9 0.585 0.8314 Day 43: Control-Negative Elevated 9 1.051 0.5659 Day 43: Positive-Negative Elevated 9 0.425 0.9062

170

Table A 50: The 20 most abundant contaminant ASVs identified by decontam across all anemone samples. Some genera, for example, Cutibacterium, have members associated with human skin microbiota, suggesting the origin of those putative contaminants.

Phylum Class Order Family Genus Relative abundance (%)

1 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Pelomonas 0.44

2 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas 0.15

3 Proteobacteria Gammaproteobacteria Pseudomonadales Moraxellaceae Acinetobacter 0.13

4 Proteobacteria Gammaproteobacteria Xanthomonadales Xanthomonadaceae Stenotrophomonas 0.08

5 Proteobacteria Gammaproteobacteria Pseudomonadales Pseudomonadaceae Pseudomonas 0.06

6 Actinobacteria Actinobacteria Propionibacteriales Propionibacteriaceae Cutibacterium 0.03

7 Firmicutes Clostridia Clostridiales Ruminococcaceae Ruminococcaceae 0.03

8 Proteobacteria Gammaproteobacteria Pseudomonadales Moraxellaceae Enhydrobacter 0.03

9 Firmicutes Bacilli Lactobacillales Streptococcaceae Streptococcus 0.02

10 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Ralstonia 0.02

11 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas 0.02

12 Proteobacteria Gammaproteobacteria Pseudomonadales Moraxellaceae Acinetobacter 0.02

13 Actinobacteria Actinobacteria Propionibacteriales Propionibacteriaceae Cutibacterium 0.02

14 Firmicutes Clostridia Clostridiales Family XI Finegoldia 0.01

15 Proteobacteria Gammaproteobacteria Betaproteobacteriales Burkholderiaceae Cupriavidus 0.01

16 Bacteroidetes Bacteroidia Flavobacteriales Weeksellaceae Chryseobacterium 0.01

17 Actinobacteria Actinobacteria Corynebacteriales Corynebacteriaceae Corynebacterium 0.01

18 Actinobacteria Actinobacteria Propionibacteriales Propionibacteriaceae Cutibacterium 0.01

19 Firmicutes Bacilli Bacillales Staphylococcaceae Staphylococcus 0.01

20 Firmicutes Bacilli Bacillales Bacillaceae Bacillus 0.01

Total 1.12

21-209 Other 0.24

Total 1.36

171

Table A 51: GLM analysis of bacterial community differences in AIMS2 based on probiotic and temperature treatments. Significant values are in bold. Analysis of Deviance Table Res.Df Df.diff Dev Pr(>Dev) (Intercept) 83 temperature 82 1 860.9 0.054 . probiotic 81 1 1055.4 0.004 ** temperature:probiotic 80 1 356.2 0.309 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Table A 52: GLM analysis of bacterial community differences in AIMS3 based on probiotic and temperature treatments. Significant values are in bold. Analysis of Deviance Table Res.Df Df.diff Dev Pr(>Dev) (Intercept) 83 temperature 82 1 878.2 0.054 . probiotic 81 1 912.1 0.018 * temperature:probiotic 80 1 407.9 0.216 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Table A 53: GLM analysis of bacterial community differences in AIMS4 based on probiotic and temperature treatments. Significant values are in bold. Analysis of Deviance Table Res.Df Df.diff Dev Pr(>Dev) (Intercept) 83 temperature 82 1 764.4 0.106 probiotic 81 1 877.0 0.024 * temperature:probiotic 80 1 393.5 0.197 Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

172

Appendix 5 Swinburne Research

Authorship Indication Form

For HDR students

NOTE This Authorship Indication form is a statement detailing the percentage of the contribution of each author in each published ‘paper’. This form must be signed by each co-author and the Principal Supervisor. This form must be added to the publication of your final thesis as an appendix. Please fill out a separate form for each published paper to be included in your thesis.

DECLARATION We hereby declare our contribution to the publication of the ‘paper’ entitled:

The______effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana

First Author

Name______Signature:Leon M Hartman ______

Percentage of contribution: ____%70 Date: _2 7_ / _01 _ / _2020 _ _ _

Brief description of contribution to the ‘paper’ and your central responsibilities/role on project: This author designed the experiment; performed the experimental work; analysed and interpreted the data; wrote the first draft; reviewed, edited and approved the final draft.

Second Author

Name:______Signature:______Madeleine J H van Oppen

Percentage of contribution: ____%15 Date: _2 7_ / _01 _ / _2020 _ _ _

Brief description of your contribution to the ‘paper’: This author assisted in research planning/design, research results interpretation and manuscript writing/editing.

Third Author

Name:______Signature:Linda L Blackall ______

Percentage of contribution: ____%15 Date: _2 7_ / _01 _ / _2020 _ _ _

Brief description of your contribution to the ‘paper’:

This author assisted in research planning/design, research results interpretation and manuscript writing/editing. 173

Fourth Author

Name:______Signature:______

Percentage of contribution: ____% Date: _ _ / _ _ / _ _ _ _

Brief description of your contribution to the ‘paper’:

Principal Supervisor:

Name:______Signature:______Prof Linda Blackall

Date: _27 _ / 01_ _ / _2020 _ _ _

In the case of more than four authors please attach another sheet with the names, signatures and contribution of the authors.

Authorship Indication Form

174

List of Publications

Articles:

*Hartman, LM, van Oppen, MJH & Blackall, LL 2020, 'Microbiota characterization of Exaiptasia diaphana from the Great Barrier Reef', Animal Microbiome, vol. 2, no. 1, 10.

* Published during examination of the thesis.

Hartman, LM, van Oppen, MJH & Blackall, LL 2019, 'The effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana', Microorganisms, vol. 8, no. 1, 20.

Dungan, AM†, Hartman, LM†, Tortorelli, G, Belderock, R, Lamb, AM, Pisan, L, McFadden, G, Blackall, LL & van Oppen, MJH 2019, 'Exaiptasia diaphana from the Great Barrier Reef: a valuable resource for coral symbiosis research' Symbiosis, vol. 80, pp. 195-206.

† Equal first authors.

Posters:

‘Great Barrier Reef-sourced anemones (Exaiptasia pallida) as a model for developing probiotics to mitigate coral bleaching’, 17th International Symposium on Microbial Ecology (ISME17), Leipzig, Germany, 12–17 August 2018.

‘Development of a bacterial probiotic from Great Barrier Reef-sourced anemones (Exaiptasia pallida) to enhance bleaching tolerance’, 1st Marine ‘omics Workshop & Symposium, Delmenhorst, Germany, 16–17 July 2018.

‘The sea anemone Exaiptasia pallida as a model for coral microbiome studies’, 1st Australian Microbial Ecology conference (AusME 2017), Melbourne, Australia, 13–15 February 2017.

175

References

Abram, F 2015, 'Systems-based approaches to unravel multi-species microbial community functioning', Computational and Structural Biotechnology Journal, vol. 13, pp. 24-32.

Abram, NJ, McGregor, HV, Tierney, JE, Evans, MN, McKay, NP, Kaufman, DS, Thirumalai, K, Martrat, B, Goosse, H, Phipps, SJ, Steig, EJ, Kilbourne, KH, Saenger, CP, Zinke, J, Leduc, G, Addison, JA, Mortyn, PG, Seidenkrantz, M- S, Sicre, M-A, Selvaraj, K, Filipsson, HL, Neukom, R, Gergis, J, Curran, MAJ & von Gunten, L 2016, 'Early onset of industrial-era warming across the oceans and continents', Nature, vol. 536, pp. 411-417.

Ahmed, HI, Herrera, M, Liew, YJ & Aranda, M 2019, 'Long-term temperature stress in the coral model Aiptasia supports the “Anna Karenina Principle” for bacterial microbiomes', Frontiers in Microbiology, vol. 10, 975.

Aichelman, HE, Townsend, JE, Courtney, TA, Baumann, JH, Davies, SW & Castillo, KD 2016, 'Heterotrophy mitigates the response of the temperate coral Oculina arbuscula to temperature stress', Ecology and Evolution, vol. 6, no. 18, pp. 6758-6769.

AIMS 2019, St Crispin Reef water temperature from 11 Jan 2016 to 11 Feb 2016, Australian Institute of Marine Science, viewed 1 Dec 2018, .

Ainsworth, TD, Fine, M, Roff, G & Hoegh-Guldberg, O 2008, 'Bacteria are not the primary cause of bleaching in the Mediterranean coral Oculina patagonica', International Society for Microbial Ecology Journal, vol. 2, no. 1, pp. 67-73.

Ainsworth, TD & Hoegh-Guldberg, O 2009, 'Bacterial communities closely associated with coral tissues vary under experimental and natural reef conditions and thermal stress', Aquatic Biology, vol. 4, no. 3, pp. 289-296.

Ainsworth, TD, Krause, L, Bridge, T, Torda, G, Raina, J-B, Zakrzewski, M, Gates, RD, Padilla-Gamino, JL, Spalding, HL, Smith, C, Woolsey, ES, Bourne, DG, Bongaerts, P, Hoegh-Guldberg, O & Leggat, W 2015, 'The coral core microbiome identifies rare bacterial taxa as ubiquitous endosymbionts', International Society for Microbial Ecology Journal, vol. 9, no. 10, pp. 2261-2274.

Alagely, A, Krediet, CJ, Ritchie, KB & Teplitski, M 2011, 'Signaling-mediated cross-talk modulates swarming and biofilm formation in a coral pathogen Serratia marcescens', International Society for Microbial Ecology Journal, vol. 5, no. 10, pp. 1609-1620.

Altschul, SF, Gish, W, Miller, W, Myers, EW & Lipman, DJ 1990, 'Basic local alignment search tool', Journal of Molecular Biology, vol. 215, pp. 403-410.

Andersson, AF, Lindberg, M, Jakobsson, H, Bäckhed, F, Nyrén, P & Engstrand, L 2008, 'Comparative analysis of human gut microbiota by barcoded pyrosequencing', PLoS ONE, vol. 3, no. 7, e2836.

Apprill, A, Weber, LG & Santoro, AE 2016, 'Distinguishing between microbial habitats unravels ecological complexity in coral microbiomes', mSystems, vol. 1, no. 5, e00143-16.

Aral, MM & Guan, J 2016, 'Global sea surface temperature and sea level rise estimation with optimal historical time lag data', Water, vol. 8, 519.

Arif, C, Daniels, C, Bayer, T, Banguera-Hinestroza, E, Barbrook, A, Howe, CJ, LaJeunesse, TC & Voolstra, CR 2014, 'Assessing Symbiodinium diversity in scleractinian corals via next-generation sequencing-based genotyping of the ITS2 rDNA region', Molecular Ecology, vol. 23, no. 17, pp. 4418-4433.

176

Armoza-Zvuloni, R, Kramarsky-Winter, E, Loya, Y, Schlesinger, A & Rosenfeld, H 2014, 'Trioecy: a unique breeding strategy in the sea anemone Aiptasia diaphana and its association with sex steroids', Biology of Reproduction, vol. 90, no. 6, 122.

Arndt, D, Xia, J, Liu, Y, Zhou, Y, Guo, AC, Cruz, JA, Sinelnikov, I, Budwill, K, Nesbo, CL & Wishart, DS 2012, 'METAGENassist: a comprehensive web server for comparative metagenomics', Nucleic Acids Research, vol. 40, pp. W88-W95.

Astudillo-García, C, Bell, JJ, Webster, NS, Glasl, B, Jompa, J, Montoya, JM & Taylor, MW 2017, 'Evaluating the core microbiota in complex communities: a systematic investigation', Environmental Microbiology, vol. 19, no. 4, pp. 1450-1462.

Baker, AC, Glynn, PW & Riegl, B 2008, 'Climate change and coral reef bleaching: an ecological assessment of long-term impacts, recovery trends and future outlook', Estuarine, Coastal and Shelf Science, vol. 80, no. 4, pp. 435-471.

Barnes, RSK & Hughes, RN 2009, 'Coral reefs', in RSK Barnes (ed.) An introduction to marine ecology, Blackwell Publishing, Oxford, pp. 117-141.

Barott, KL, Venn, AA, Perez, SO, Tambutté, S & Tresguerres, M 2015, 'Coral host cells acidify symbiotic algal microenvironment to promote photosynthesis', Proceedings of the National Academy of Sciences, vol. 112, no. 2, pp. 607-612.

Basic, M & Bleich, A 2019, 'Gnotobiotics: past, present and future', Laboratory Animals, vol. 53, no. 3, pp. 232- 243.

Bassis, C, Young, V & Schmidt, T 2013, 'Methods for characterizing microbial communities associated with the human body', in DN Fredricks (ed.) The Human Microbiota, John Wiley, NJ, pp. 51-74.

Baumgarten, S, Simakov, O, Esherick, LY, Liew, YJ, Lehnert, EM, Michell, CT, Li, Y, Hambleton, EA, Guse, A, Oates, ME, Gough, J, Weis, VM, Aranda, M, Pringle, JR & Voolstra, CR 2015, 'The genome of Aiptasia, a sea anemone model for coral symbiosis', Proceedings of the National Academy of Sciences, USA, vol. 112, no. 38, pp. 11893-11898.

Beier, S & Bertilsson, S 2013, 'Bacterial chitin degradation-mechanisms and ecophysiological strategies', Frontiers in Microbiology, vol. 4, 149.

Bellis, ES & Denver, DR 2017, 'Natural variation in responses to acute heat and cold stress in a sea anemone model system for coral bleaching', Biological Bulletin, vol. 233, no. 2, pp. 168-181.

Bellis, ES, Edlund, RB, Berrios, HK, Lessios, HA & Denver, DR 2018, 'Molecular signatures of host specificity linked to habitat specialization in Exaiptasia sea anemones', Ecology and Evolution, vol. 8, no. 11, pp. 5413- 5426.

Bellis, ES, Howe, DK & Denver, DR 2016, 'Genome-wide polymorphism and signatures of selection in the symbiotic sea anemone Aiptasia', BMC Genomics, vol. 17, no. 1, pp. 1-14.

Benjamini, Y & Hochberg, Y 1995, 'Controlling the false discovery rate: a practical and powerful approach to multiple testing', Journal of the Royal Statistical Society: Series B (Statistical Methodology), vol. 57, no. 1, pp. 289-300.

Bernasconi, R, Stat, M, Koenders, A & Huggett, MJ 2018, 'Global networks of Symbiodinium-bacteria within the coral holobiont', Microbial Ecology, vol. 77, pp. 794-807.

177

Bernasconi, R, Stat, M, Koenders, A, Paparini, A, Bunce, M & Huggett, MJ 2019, 'Establishment of coral-bacteria symbioses reveal changes in the core bacterial community with host ontogeny', Frontiers in Microbiology, vol. 10, 1529.

Biedermann, L & Rogler, G 2015, 'The intestinal microbiota: its role in health and disease', European Journal of Pediatrics, vol. 174, no. 2, pp. 151-167.

Bieri, T, Onishi, M, Xiang, T, Grossman, AR & Pringle, JR 2016, 'Relative contributions of various cellular mechanisms to loss of algae during cnidarian bleaching', PLoS ONE, vol. 11, no. 4, e0152693.

Blackall, LL, Wilson, B & van Oppen, MJH 2015, 'Coral-the world's most diverse symbiotic ecosystem', Molecular ecology, vol. 24, no. 21, pp. 5330-5347.

Blanca, MJ, Alarcon, R, Arnau, J, Bono, R & Bendayan, R 2017, 'Non-normal data: is ANOVA still a valid option?', Psicothema, vol. 29, no. 4, pp. 552-557.

Bober, JR, Beisel, CL & Nair, NU 2018, 'Synthetic biology approaches to engineer probiotics and members of the human microbiota for biomedical applications', Annual Review of Biomedical Engineering, vol. 20, no. 1, pp. 277-300.

Bocharova, E 2016, 'Reproduction of sea anemones and other hexacorals', in S Goffredo and Z Dubinsky (eds), The Cnidaria, past, present and future: the world of Medusa and her sisters, Springer, Cham, pp. 239-248.

Bocharova, E & Kosevich, IA 2011, 'Modes of reproduction in sea anemones (Cnidaria, Anthozoa)', Zoologicheskiĭ zhurnal, vol. 90, no. 11, pp. 1283-1295.

Bokulich, NA, Kaehler, BD, Rideout, JR, Dillon, M, Bolyen, E, Knight, R, Huttley, GA & Caporaso, JG 2018, 'Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier plugin', Microbiome, vol. 6, 90.

Bolyen, E, Rideout, JR, Dillon, MR, Bokulich, NA, Abnet, C, Al-Ghalith, GA, Alexander, H, Alm, EJ, Arumugam, M, Asnicar, F, Bai, Y, Bisanz, JE, Bittinger, K, Brejnrod, A, Brislawn, CJ, Brown, CT, Callahan, BJ, Caraballo-Rodríguez, AM, Chase, J, Cope, E, Da Silva, R, Dorrestein, PC, Douglas, GM, Durall, DM, Duvallet, C, Edwardson, CF, Ernst, M, Estaki, M, Fouquier, J, Gauglitz, JM, Gibson, DL, Gonzalez, A, Gorlick, K, Guo, J, Hillmann, B, Holmes, S, Holste, H, Huttenhower, C, Huttley, G, Janssen, S, Jarmusch, AK, Jiang, L, Kaehler, B, Kang, KB, Keefe, CR, Keim, P, Kelley, ST, Knights, D, Koester, I, Kosciolek, T, Kreps, J, Langille, MGI, Lee, J, Ley, R, Liu, Y-X, Loftfield, E, Lozupone, C, Maher, M, Marotz, C, Martin, BD, McDonald, D, McIver, LJ, Melnik, AV, Metcalf, JL, Morgan, SC, Morton, J, Naimey, AT, Navas-Molina, JA, Nothias, LF, Orchanian, SB, Pearson, T, Peoples, SL, Petras, D, Preuss, ML, Pruesse, E, Rasmussen, LB, Rivers, A, Robeson, IIMS, Rosenthal, P, Segata, N, Shaffer, M, Shiffer, A, Sinha, R, Song, SJ, Spear, JR, Swafford, AD, Thompson, LR, Torres, PJ, Trinh, P, Tripathi, A, Turnbaugh, PJ, Ul-Hasan, S, van der Hooft, JJJ, Vargas, F, Vázquez-Baeza, Y, Vogtmann, E, von Hippel, M, Walters, W, Wan, Y, Wang, M, Warren, J, Weber, KC, Williamson, CHD, Willis, AD, Xu, ZZ, Zaneveld, JR, Zhang, Y, Zhu, Q, Knight, R & Caporaso, JG 2019, 'Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2', Nature Biotechnology, vol. 37, pp. 852–857.

Boon, E, Meehan, CJ, Whidden, C, Wong, DHJ, Langille, MGI & Beiko, RG 2014, 'Interactions in the microbiome: communities of organisms and communities of genes', FEMS Microbiology Reviews, vol. 38, no. 1, pp. 90-118.

Borell, EM, Yuliantri, AR, Bischof, K & Richter, C 2008, 'The effect of heterotrophy on photosynthesis and tissue composition of two scleractinian corals under elevated temperature', Journal of Experimental Marine Biology and Ecology, vol. 364, no. 2, pp. 116-123.

Bourne, D, Iida, Y, Uthicke, S & Smith-Keune, C 2008, 'Changes in coral-associated microbial communities during a bleaching event', International Society for Microbial Ecology Journal, vol. 2, no. 4, pp. 350-363.

178

Bourne, DG, Garren, M, Work, TM, Rosenberg, E, Smith, GW & Harvell, CD 2009, 'Microbial disease and the coral holobiont', Trends in Microbiology, vol. 17, no. 12, pp. 554-562.

Bourne, DG, Morrow, KM & Webster, NS 2016, 'Insights into the coral microbiome: underpinning the health and resilience of reef ecosystems', Annual Review of Microbiology, vol. 70, no. 1, pp. 317-340.

Bourne, DG & Webster, NS 2013, 'Coral reef bacterial communities', in E Rosenberg, EF DeLong, S Lory, E Stackebrandt and F Thompson (eds), The prokaryotes: prokaryotic communities and ecophysiology, Springer, Berlin, pp. 163-187.

Bradford, MM 1976, 'A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding', Analytical Biochemistry, vol. 72, pp. 248-254.

Brodie, JE & Mitchell, AW 2005, 'Nutrients in Australian tropical rivers: changes with agricultural development and implications for receiving environments', Marine and Freshwater Research, vol. 56, no. 3, pp. 279-302.

Brown, B & Bythell, J 2005, 'Perspectives on mucus secretion in reef corals', Marine Ecology Progress Series, vol. 296, pp. 291-309.

Brown, T, Otero, C, Grajales , A, Rodríguez, E & Rodriguez-Lanetty, M 2017, 'Worldwide exploration of the microbiome harbored by the cnidarian model, Exaiptasia pallida (Agassiz in Verrill, 1864) indicates a lack of bacterial association specificity at a lower taxonomic rank', PeerJ, vol. 5, e3235.

Brown, T & Rodriguez-Lanetty, M 2015, 'Defending against pathogens – immunological priming and its molecular basis in a sea anemone, cnidarian', Scientific Reports, vol. 5, p. 17425.

Bruckner, AW 2002, 'Life-saving products form coral reefs', Issues in Science and Technology, vol. 18, no. 3, pp. 39-44.

Brundin, M, Figdor, D, Roth, C, Davies, JK, Sundqvist, G & Sjögren, U 2010, 'Persistence of dead-cell bacterial DNA in ex vivo root canals and influence of nucleases on DNA decay in vitro', Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology, and Endodontology, vol. 110, no. 6, pp. 789-794.

Brune, A 2014, 'Symbiotic digestion of lignocellulose in termite guts', Nature Reviews: Microbiology, vol. 12, pp. 168-180.

Buffie, CG & Pamer, EG 2013, 'Microbiota-mediated colonization resistance against intestinal pathogens', Nature Reviews: Immunology, vol. 13, no. 11, pp. 790-801.

Burriesci, MS, Raab, TK & Pringle, JR 2012, 'Evidence that glucose is the major transferred metabolite in dinoflagellate–cnidarian symbiosis', Journal of Experimental Biology, vol. 215, no. 19, pp. 3467-3477.

Busby, PE, Soman, C, Wagner, MR, Friesen, ML, Kremer, J, Bennett, A, Morsy, M, Eisen, JA, Leach, JE & Dangl, JL 2017, 'Research priorities for harnessing plant microbiomes in sustainable agriculture', PLoS Biology, vol. 15, no. 3, e2001793.

Cai, L, Zhou, G, Tong, H, Tian, R-M, Zhang, W, Ding, W, Liu, S, Huang, H & Qian, P-Y 2018, 'Season structures prokaryotic partners but not algal symbionts in subtropical hard corals', Applied Microbiology and Biotechnology, vol. 102, no. 11, pp. 4963-4973.

Callahan, BJ, McMurdie, PJ, Rosen, MJ, Han, AW, Johnson, AJA & Holmes, SP 2016, 'DADA2: high-resolution sample inference from Illumina amplicon data', Nature Methods, vol. 13, no. 7, pp. 581-583.

Carilli, J 2013, Why are coral reefs important?, Nature, viewed 3 July 2018, .

179

Catterson, JH, Khericha, M, Dyson, MC, Vincent, AJ, Callard, R, Haveron, SM, Rajasingam, A, Ahmad, M & Partridge, L 2018, 'Short-term, intermittent fasting induces long-lasting gut health and TOR-independent lifespan extension', Current Biology, vol. 28, no. 11, pp. 1714-1724.

Ceh, J, Kilburn, MR, Cliff, JB, Raina, J-B, van Keulen, M & Bourne, DG 2013, 'Nutrient cycling in early coral life stages: Pocillopora damicornis larvae provide their algal symbiont (Symbiodinium) with nitrogen acquired from bacterial associates', Ecology and Evolution, vol. 3, no. 8, pp. 2393-2400.

Ceh, J, van Keulen, M & Bourne, D 2013, 'Intergenerational transfer of specific bacteria in corals and possible implications for offspring fitness', Microbial Ecology, vol. 65, no. 1, pp. 227-231.

Chakravarti, LJ, Beltran, VH & Oppen, MJH 2017, 'Rapid thermal adaptation in photosymbionts of reef-building corals', Global Change Biology, vol. 23, no. 11, pp. 4675-4688.

Chen, C-P, Tseng, C-H, Chen, CA & Tang, S-L 2011, 'The dynamics of microbial partnerships in the coral Isopora palifera', International Society for Microbial Ecology Journal, vol. 5, no. 4, pp. 728-740.

Chen, W-NU, Hsiao, Y-J, Mayfield, AB, Young, R, Hsu, L-L & Peng, S-E 2016, 'Transmission of a heterologous clade C Symbiodinium in a model anemone infection system via asexual reproduction', PeerJ, vol. 4, e2358.

Choat, JH & Bellwood, DR 1991, 'Reef fishes: their history and evolution', in PF Sale (ed.) The Ecology of Fishes on Coral Reefs, Academic Press, San Diego, pp. 39-66.

Clayton, WS & Lasker, HR 1984, 'Host feeding regime and Zooxanthellal photosynthesis in the anemone, Aiptasia pallida (Verrill)', Biological Bulletin, vol. 167, no. 3, pp. 590-600.

Closek, CJ, Sunagawa, S, DeSalvo, MK, Piceno, YM, DeSantis, TZ, Brodie, EL, Weber, MX, Voolstra, CR, Andersen, GL & Medina, M 2014, 'Coral transcriptome and bacterial community profiles reveal distinct Yellow Band Disease states in Orbicella faveolata', International Society for Microbial Ecology Journal, vol. 8, pp. 2411-2422.

Coleman-Derr, D & Tringe, SG 2014, 'Building the crops of tomorrow: advantages of symbiont-based approaches to improving abiotic stress tolerance', Frontiers in Microbiology, vol. 5, 283.

Coles, SL, Jokiel, PL & Lewis, CR 1976, 'Thermal tolerance in tropical versus subtropical Pacific reef corals', Pacific Science, vol. 30, no. 2, pp. 159-166.

Compton, DL, Evans, KO, Appell, M & Goodell, JR 2019, 'Protection of antioxidants, vitamins E and C, from ultraviolet degradation using feruloylated vegetable oil', Journal of the American Oil Chemists' Society, vol. 96, no. 9, pp. 999-1009.

Connon, SA & Giovannoni, SJ 2002, 'High-throughput methods for culturing microorganisms in very-low- nutrient media yield diverse new marine isolates', Applied and Environmental Microbiology, vol. 68, no. 8, pp. 3878-3885.

Cook, CB, D'Elia, CF & Muller-Parker, G 1988, 'Host feeding and nutrient sufficiency for zooxanthellae in the sea anemone Aiptasia pallida', Marine Biology, vol. 98, no. 2, pp. 253-262.

Cooke, I, Mead, O, Whalen, C, Boote, C, Moya, A, Ying, H, Robbins, S, Strugnell, JM, Darling, A, Miller, D, Voolstra, CR & Adamska, M 2019, 'Molecular techniques and their limitations shape our view of the holobiont', Zoology, vol. 137, 125695.

Cordovez, V, Carrion, VJ, Etalo, DW, Mumm, R, Zhu, H, van Wezel, GP & Raaijmakers, JM 2015, 'Diversity and functions of volatile organic compounds produced by Streptomyces from a disease-suppressive soil', Frontiers in Microbiology, vol. 6, 1081.

180

Costa, RM, Cárdenas, A & Voolstra, C 2019, Protocol for bacterial depletion of Aiptasia anemones - towards the generation of gnotobiotic/germ-free cnidarian host animals, protocols.io, viewed 1 Oct 2019, .

Cruz, J, Liu, Y, Liang, Y, Zhou, Y, Wilson, M, Dennis, JJ, Stothard, P, Van Domselaar, G & Wishart, DS 2012, 'BacMap: an up-to-date electronic atlas of annotated bacterial genomes', Nucleic acids research, vol. 40, pp. D599-D604.

Cunning, R & Baker, AC 2013, 'Excess algal symbionts increase the susceptibility of reef corals to bleaching', Nature Climate Change, vol. 3, no. 3, pp. 259-262.

Cziesielski, MJ, Liew, YJ, Cui, G, Schmidt-Roach, S, Campana, S, Marondedze, C & Aranda, M 2018, 'Multi-omics analysis of thermal stress response in a zooxanthellate cnidarian reveals the importance of associating with thermotolerant symbionts', Proceedings of the Royal Society B: Biological Sciences, vol. 285, 20172654.

D’Agostino, A 1975, 'Antibiotics in cultures of invertebrates,' Proceedings of the first conference on culture of marine invertebrate animals 1975, Greenport, Springer, Boston, pp. 109-133.

Damjanovic, K, Menéndez, P, Blackall, LL & van Oppen, MJH 2019a, 'Early life stages of a common broadcast spawning coral associate with specific bacterial communities despite lack of internalized bacteria', Microbial Ecology, vol. 79, pp. 706-719.

Damjanovic, K, van Oppen, MJH, Menéndez, P & Blackall, LL 2019b, 'Experimental inoculation of coral recruits with marine bacteria indicates scope for microbiome manipulation in Acropora tenuis and Platygyra daedalea', Frontiers in Microbiology, vol. 10, 1702.

Dang, H, Zhang, X, Song, L, Chang, Y & Yang, G 2007, 'Molecular determination of oxytetracycline-resistant bacteria and their resistance genes from mariculture environments of China', Journal of Applied Microbiology, vol. 103, no. 6, pp. 2580-2592.

Darling, JA, Reitzel, AR, Burton, PM, Mazza, ME, Ryan, JF, Sullivan, JC & Finnerty, JR 2005, 'Rising starlet: the starlet sea anemone, Nematostella vectensis', BioEssays, vol. 27, no. 2, pp. 211-221.

Davis, AM, Lewis, SE, Bainbridge, ZT, Glendenning, L, Turner, RDR & Brodie, JE 2012, 'Dynamics of herbicide transport and partitioning under event flow conditions in the lower Burdekin region, Australia', Marine Pollution Bulletin, vol. 65, pp. 182-193.

Davis, NM, Proctor, DM, Holmes, SP, Relman, DA & Callahan, BJ 2018, 'Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data', Microbiome, vol. 6, no. 1

Davy, S & Cook, C 2001, 'The relationship between nutritional status and carbon flux in the zooxanthellate sea anemone Aiptasia pallida', Marine Biology, vol. 139, no. 5, pp. 999-1005.

De, B, Meena, D, Behera, B, Das, P, Das Mohapatra, P & Sharma, A 2014, 'Probiotics in fish and shellfish culture: immunomodulatory and ecophysiological responses', Fish Physiology and Biochemistry, vol. 40, no. 3, pp. 921- 971. de Vos, WM 2013, 'Fame and future of faecal transplantations – developing next-generation therapies with synthetic microbiomes', Microbial Biotechnology, vol. 6, no. 4, pp. 316-325.

De’ath, G, Fabricius, KE, Sweatman, H & Puotinen, M 2012, 'The 27–year decline of coral cover on the Great Barrier Reef and its causes', Proceedings of the National Academy of Sciences, USA, vol. 109, no. 44, pp. 17995- 17999.

Deb, R, Nair, A & Agashe, D 2019, 'Host dietary specialization and neutral assembly shape gut bacterial communities of wild dragonflies', PeerJ, vol. 7, e8058. 181

Deloitte Access Economics 2017, At what price? The economic, social and iconic value of the Great Barrier Reef, viewed 1 Feb 2019, .

Dewhirst, FE, Chien, CC, Paster, BJ, Ericson, RL, Orcutt, RP, Schauer, DB & Fox, JG 1999, 'Phylogeny of the defined murine microbiota: altered Schaedler flora', Applied and environmental microbiology, vol. 65, no. 8, pp. 3287-3292.

Diaz, JM, Hansel, CM, Apprill, A, Brighi, C, Zhang, T, Weber, L, McNally, S & Xun, L 2016, 'Species-specific control of external superoxide levels by the coral holobiont during a natural bleaching event', Nature Communications, vol. 7, 13801.

Domozych, DS 2016, 'Biosynthesis of the cell walls of the algae', in MA Borowitzka, J Beardall and JA Raven (eds), The Physiology of Microalgae, Springer, London, pp. 47-65.

Dong, L, Meng, Y, Wang, J & Liu, Y 2014, 'Evaluation of droplet digital PCR for characterizing plasmid reference material used for quantifying ammonia oxidizers and denitrifiers', Analytical and Bioanalytical Chemistry, vol. 406, no. 6, pp. 1701-1712. dos Santos, HF, Duarte, GAS, Rachid, CTdC, Chaloub, RM, Calderon, EN, Marangoni, LFdB, Bianchini, A, Nudi, AH, do Carmo, FL, van Elsas, JD, Rosado, AS, Castro, CB & Peixoto, RS 2015, 'Impact of oil spills on coral reefs can be reduced by bioremediation using probiotic microbiota', Scientific Reports, vol. 5, 18268.

Downs, CA, Fauth, JE, Halas, JC, Dustan, P, Bemiss, J & Woodley, CM 2002, 'Oxidative stress and seasonal coral bleaching', Free Radical Biology and Medicine, vol. 33, no. 4, pp. 533-543.

Ducklow, HW & Mitchell, R 1979, 'Bacterial populations and adaptations in the mucus layers on living corals', Limnology and Oceanography, vol. 24, no. 4, pp. 715-725.

Dufrêne, M & Legendre, P 1997, 'Species assemblages and indicator species: the need for a flexible asymmetrical approach', Ecological Monographs, vol. 67, no. 3, pp. 345-366.

Dungan, AM, Hartman, LM, Tortorelli, G, Belderock, R, Lamb, AM, Pisan, L, McFadden, G, Blackall, LL & van Oppen, MJH 2020, 'Exaiptasia diaphana from the Great Barrier Reef: a valuable resource for coral symbiosis research', Symbiosis, vol. 80, pp. 195-206.

Eakin, CM, Lough, JM & Heron, SF 2009, 'Climate variability and change: Monitoring data and evidence for increased coral bleaching stress', in MJH van Oppen and JM Lough (eds), Coral Bleaching, vol. 205, Springer, Heidelberg, pp. 41-67.

Edwards, KJ, Rogers, DR, Wirsen, CO & McCollom, TM 2003, 'Isolation and characterization of novel psychrophilic, neutrophilic, Fe-oxidizing, chemolithoautotrophic α- and γ-Proteobacteria from the deep sea', Applied and Environmental Microbiology, vol. 69, no. 5, pp. 2906-2913.

Elder, C 2015, 'Stony coral anatomy' [image], in Design, christineelder.com, viewed 1 Oct 2019, .

Epstein, HE, Smith, HA, Torda, G & van Oppen, MJ 2019a, 'Microbiome engineering: enhancing climate resilience in corals', Frontiers in Ecology and the Environment, vol. 17, no. 2, pp. 100-108.

Epstein, HE, Torda, G, Munday, PL & van Oppen, MJH 2019b, 'Parental and early life stage environments drive establishment of bacterial and dinoflagellate communities in a common coral', The ISME Journal, vol. 13, no. 6, pp. 1635-1638.

Falkowski, PG, Dubinsky, Z, Muscatine, L & McCloskey, L 1993, 'Population-control in symbiotic corals', Bioscience, vol. 43, no. 9, pp. 606-611. 182

Federhen, S 2012, 'The NCBI taxonomy database', Nucleic acids research, vol. 40, pp. D136-D143.

Fernando, SC, Wang, J, Sparling, K, Garcia, GD, Francini-Filho, RB, Moura, RL, Paranhos, R, Thompson, FL & Thompson, JR 2014, 'Microbiota of the major south atlantic reef building coral Mussismilia', Microbial Ecology, vol. 69, no. 2, pp. 267-280.

Ferrario, F, Beck, MW, Storlazzi, CD, Micheli, F, Shepard, CC & Airoldi, L 2014, 'The effectiveness of coral reefs for coastal hazard risk reduction and adaptation', Nature Communications, vol. 5, 3794.

Ferrier-Pagès, C, Rottier, C, Beraud, E & Levy, O 2010, 'Experimental assessment of the feeding effort of three scleractinian coral species during a thermal stress: effect on the rates of photosynthesis', Journal of Experimental Marine Biology and Ecology, vol. 390, no. 2, pp. 118-124.

Forberg, T & Milligan-Myhre, K 2017, 'Gnotobiotic fish as models to study host-microbe interactions', in TR Schoeb and KA Eaton (eds), Gnotobiotics, Academic Press, London, pp. 369-383.

Fox, J, Weisberg, S & Price, B 2019, car: companion to applied regression [software] R package version 3.0-3, viewed 1 Nov 2019, .

Fox, RS 2006, Aiptasia pallida and Metridium senile anemones, Lander University, viewed 3 Aug 2019, .

Fransolet, D, Roberty, S, Herman, A-C, Tonk, L, Hoegh-Guldberg, O & Plumier, J-C 2013, 'Increased cell proliferation and mucocyte density in the sea anemone Aiptasia pallida recovering from bleaching', PLoS ONE, vol. 8, no. 5, e65015.

Fraune, S, Anton-Erxleben, F, Augustin, R, Franzenburg, S, Knop, M, Schröder, K, Willoweit-Ohl, D & Bosch, TCG 2015, 'Bacteria–bacteria interactions within the microbiota of the ancestral metazoan Hydra contribute to fungal resistance', International Society for Microbial Ecology Journal, vol. 9, no. 7, pp. 1543-1556.

Freire, I, Gutner-Hoch, E, Muras, A, Benayahu, Y & Otero, A 2019, 'The effect of bacteria on planula-larvae settlement and metamorphosis in the octocoral Rhytisma fulvum fulvum', PloS ONE, vol. 14, no. 9, e0223214.

Frias-Lopez, J, Zerkle, AL, Bonheyo, GT & Fouke, BW 2002, 'Partitioning of bacterial communities between seawater and healthy, black band diseased, and dead coral surfaces', Applied and Environmental Microbiology, vol. 68, no. 5, pp. 2214-2228.

Fritz, JV, Desai, MS, Shah, P, Schneider, JG & Wilmes, P 2013, 'From meta-omics to causality: experimental models for human microbiome research', Microbiome, vol. 1, 14.

Furusawa, G, Yoshikawa, T, Yasuda, A & Sakata, T 2003, 'Algicidal activity and gliding motility of Saprospira sp. SS98-5', Canadian Journal of Microbiology/Revue Canadienne de Microbiologie, vol. 49, no. 2, pp. 92-100.

Gajigan, AP, Diaz, LA & Conaco, C 2017, 'Resilience of the prokaryotic microbial community of Acropora digitifera to elevated temperature', MicrobiologyOpen, vol. 6, no. 4, e00478.

Galliot, B 2012, 'Hydra, a fruitful model system for 270 years', International Journal of Developmental Biology, vol. 56, no. 6-8, pp. 411-423.

Gardner, TA, Cote, IM, Gill, JA, Grant, A & Watkinson, AR 2003, 'Long-term region-wide declines in Caribbean corals', Science, vol. 310, no. 5635, pp. 958-960.

Gates, RD, Baghdasarian, G & Muscatine, L 1992, 'Temperature stress causes host-cell detachment in symbiotic cnidarians: Implications for coral bleaching', Biological Bulletin, vol. 182, no. 3, pp. 324-332.

183

Gegner, HM, Rädecker, N, Ochsenkühn, M, Barreto, MM, Ziegler, M, Reichert, J, Schubert, P, Wilke, T & Voolstra, CR 2019, 'High levels of floridoside at high salinity link osmoadaptation with bleaching susceptibility in the cnidarian-algal endosymbiosis', Biology Open, vol. 8, no. 12, bio045591.

Gegner, HM, Ziegler, M, Rädecker, N, Buitrago-López, C, Aranda, M & Voolstra, CR 2017, 'High salinity conveys thermotolerance in the coral model Aiptasia', Biology Open, vol. 6, pp. 1943-1948.

Geller, JB, Fitzgerald, LJ & King, CE 2005, 'Fission in sea anemones: integrative studies of life cycle evolution', Integrative and Comparative Biology, vol. 45, no. 4, pp. 615-622.

Gevers, D, Knight, R, Petrosino, JF, Huang, K, McGuire, AL, Birren, BW, Nelson, KE, White, O, Methé, BA & Huttenhower, C 2012, 'The Human Microbiome Project: a community resource for the healthy human microbiome', PLoS Biology, vol. 10, no. 8, e1001377.

Glasl, B, Herndl, GJ & Frade, PR 2016, 'The microbiome of coral surface mucus has a key role in mediating holobiont health and survival upon disturbance', International Society for Microbial Ecology Journal, vol. 10, no. 9, pp. 2280-2292.

Glasl, B, Webster, NS & Bourne, DG 2017, 'Microbial indicators as a diagnostic tool for assessing water quality and climate stress in coral reef ecosystems', Marine Biology, vol. 164, no. 4, p. 91.

Gloor, GB, Macklaim, JM, Pawlowsky-Glahn, V & Egozcue, JJ 2017, 'Microbiome datasets are compositional: and this is not optional', Frontiers in Microbiology, vol. 8, 2224.

Gobet, A, Barbeyron, T, Matard-Mann, M, Magdelenat, G, Vallenet, D, Duchaud, E & Michel, G 2018, 'Evolutionary evidence of algal polysaccharide degradation acquisition by Pseudoalteromonas carrageenovora 9T to adapt to macroalgal niches', Frontiers in Microbiology, vol. 9, 2740.

Gómez-Cabrera, MdC, Ortiz, JC, Loh, WKW, Ward, S & Hoegh-Guldberg, O 2008, 'Acquisition of symbiotic dinoflagellates (Symbiodinium) by juveniles of the coral Acropora longicyathus', Coral Reefs, vol. 27, pp. 219- 226.

Gonzalez-Perez, G, Hicks, AL, Tekieli, TM, Radens, CM, Williams, BL & Lamousé-Smith, ESN 2016, 'Maternal antibiotic treatment impacts development of the neonatal intestinal microbiome and antiviral immunity', Journal of Immunology, vol. 196, no. 9, pp. 3768-3779.

Gordon, BR & Leggat, W 2010, 'Symbiodinium-invertebrate symbioses and the role of metabolomics', Marine Drugs, vol. 8, no. 10, pp. 2546-2568.

Gordon, HA & Pesti, L 1972, 'The gnotobiotic animal as a tool in the study of host microbial relationships', Bacteriological Reviews, vol. 35, no. 4, pp. 390-429.

Goulet, TL, Cook, CB & Goulet, D 2005, 'Effect of short-term exposure to elevated temperatures and light levels on photosynthesis of different host–symbiont combinations in the Aiptasia pallida/Symbiodinium symbiosis', Limnology and Oceanography, vol. 50, no. 5, pp. 1490–1498.

Grajales, A & Rodríguez, E 2014, 'Morphological revision of the genus Aiptasia and the family Aiptasiidae (Cnidaria, Actinaria, Metridioidea)', Zootaxa, vol. 3826, no. 1, pp. 55-100.

Grajales, A & Rodríguez, E 2015, 'Elucidating the evolutionary relationships of the Aiptasiidae, a widespread cnidarian–dinoflagellate model system (Cnidaria: Anthozoa: Actiniaria: Metridioidea)', Molecular Phylogenetics and Evolution, vol. 94, pp. 252-263.

Grajales, A, Rodríguez, E & Thornhill, DJ 2016, 'Patterns of Symbiodinium spp. associations within the family Aiptasiidae, a monophyletic lineage of symbiotic of sea anemones (Cnidaria, Actiniaria)', Coral Reefs, vol. 35, no. 1, pp. 345-355. 184

Grawunder, D, Hambleton, EA, Bucher, M, Wolfowicz, I, Bechtoldt, N & Guse, A 2015, 'Induction of gametogenesis in the cnidarian endosymbiosis model Aiptasia sp', Scientific Reports, vol. 5, 15677.

Grottoli, AG, Dalcin Martins, P, Wilkins, MJ, Johnston, MD, Warner, ME, Cai, W-J, Melman, TF, Hoadley, KD, Pettay, DT, Levas, S & Schoepf, V 2018, 'Coral physiology and microbiome dynamics under combined warming and ocean acidification', PLoS ONE, vol. 13, no. 1, e0191156.

Grottoli, AG, Rodrigues, LJ & Palardy, JE 2006, 'Heterotrophic plasticity and resilience in bleached corals', Nature, vol. 440, no. 7088, pp. 1186-1189.

Guo, X, Zhang, X, Qin, Y, Liu, Y-X, Zhang, J, Zhang, N, Wu, K, Qu, B, He, Z, Wang, X, Zhang, X, Hacquard, S, Fu, X & Bai, Y 2020, 'Host-associated quantitative abundance profiling reveals the microbial load variation of root microbiome', Plant Communications, vol. 1, no. 1, 100003.

Haapkylä, J, Unsworth, RKF, Flavell, M, Bourne, DG, Schaffelke, B & Willis, BL 2011, 'Seasonal rainfall and runoff promote coral disease on an inshore reef', PLoS ONE, vol. 6, no. 2, e16893.

Haas, AF & Wild, C 2010, 'Composition analysis of organic matter released by cosmopolitan coral reef- associated green algae', Aquatic Biology, vol. 10, pp. 131-138.

Habetha, M, Anton-Erxleben, F, Neumann, K & Bosch, TC 2003, 'The Hydra viridis/Chlorella symbiosis: growth and sexual differentiation in polyps without symbionts', Zoology, vol. 106, pp. 101-108.

Hackstadt, AJ & Hess, AM 2009, 'Filtering for increased power for microarray data analysis', BMC Bioinformatics, vol. 10, no. 1, p. 11.

Hai, NV 2015, 'The use of probiotics in aquaculture', Journal of Applied Microbiology, vol. 119, no. 4, pp. 917- 935.

Hai, NV, Buller, N & Fotedar, R 2010, 'Encapsulation capacity of Artemia nauplii with customized probiotics for use in the cultivation of western king prawns (Penaeus latisulcatus Kishinouye, 1896)', Aquaculture Research, vol. 41, no. 6, pp. 893-903.

Hand, C & Uhlinger, KR 1992, 'The culture, sexual and asexual reproduction, and growth of the sea-anemone Nematostella vectensis', Biological Bulletin, vol. 182, no. 2, pp. 169-176.

Hand, C & Uhlinger, KR 1994, 'The unique, widely distributed, estuarine sea anemone, Nematostella vectensis Stephenson: a review, new facts, and questions', Estuaries, vol. 17, no. 2, pp. 501-508.

Har, J, Helbig, T, Lim, J, Fernando, S, Penn, K, Reitzel, A & Thompson, J 2015, 'Microbial diversity and activity in the Nematostella vectensis holobiont: insights from 16S rRNA gene sequencing, isolate genomes, and a pilot- scale survey of gene expression', Frontiers in Microbiology, vol. 6, no. 818, pp. 818.01-818.22.

Hartman, LM, van Oppen, MJH & Blackall, LL 2019, 'The effect of thermal stress on the bacterial microbiome of Exaiptasia diaphana', Microorganisms, vol. 8, no. 1, 20.

Hawkins, TD, Hagemeyer, JCG, Hoadley, KD, Marsh, AG & Warner, ME 2016, 'Partitioning of respiration in an animal-algal symbiosis: implications for different aerobic capacity between Symbiodinium spp', Frontiers in Physiology, vol. 7,

Heijtz, RD, Wang, S, Anuar, F, Qian, Y, Björkholm, B, Samuelsson, A, Hibberd, ML, Forssberg, H & Pettersson, S 2011, 'Normal gut microbiota modulates brain development and behavior', Proceedings of the National Academy of Sciences, USA, vol. 108, no. 7, pp. 3047-3052.

Hernandez-Agreda, A, Leggat, W, Bongaerts, P & Ainsworth, TD 2016, 'The microbial signature provides insight into the mechanistic basis of coral success across reef habitats', mBio, vol. 7, no. 4, e00560-16. 185

Herrera, M, Ziegler, M, Voolstra, C & Aranda, M 2017, 'Laboratory-cultured strains of the sea anemone Exaiptasia reveal distinct bacterial communities', Frontiers in Marine Science, vol. 4, 115.

Hester, ER, Barott, KL, Nulton, J, Vermeij, MJA & Rohwer, FL 2016, 'Stable and sporadic symbiotic communities of coral and algal holobionts', International Society for Microbial Ecology Journal, vol. 10, no. 5, pp. 1157-1169.

Hibbing, ME, Fuqua, C, Parsek, MR & Peterson, SB 2010, 'Bacterial competition: surviving and thriving in the microbial jungle', Nature Reviews: Microbiology, vol. 8, no. 1, pp. 15-25.

Hillyer, KE, Tumanov, S, Villas-Bôas, S & Davy, SK 2016, 'Metabolite profiling of symbiont and host during thermal stress and bleaching in a model cnidarian–dinoflagellate symbiosis', Journal of Experimental Biology, vol. 219, no. 4, pp. 516-527.

Hoegh-Guldberg, O 1999, 'Climate change, coral bleaching and the future of the world's coral reefs', Marine and Freshwater Research, vol. 50, no. 8, pp. 839-866.

Hoegh-Guldberg, O & Smith, GJ 1989, 'Influence of the population density of zooxanthellae and supply of ammonium on the biomass and metabolic characteristics of the reef corals Seriatopora hystrix and Stylophora pistillata', Marine Ecology Progress Series, vol. 57, no. 2, pp. 173-186.

Höhener, T, Huap, AC & Müller, M 2016, 'Role of bacterial communities in coral’s defence against a causative agent of coral bleaching: Vibrio coralliilyticus', Malaysian Journal of Microbiology, vol. 12, no. 4, pp. 279-283.

Hollants, J, Leliaert, F, Verbruggen, H, Willems, A & De Clerck, O 2013, 'Permanent residents or temporary lodgers: characterizing intracellular bacterial communities in the siphonous green alga Bryopsis', Proceedings of the Royal Society B: Biological Sciences, vol. 280, 20122659.

Houlbrèque, F & Ferrier-Pagès, C 2009, 'Heterotrophy in tropical scleractinian corals', Biological Reviews, vol. 84, no. 1, pp. 1-17.

Howe, PL, Reichelt-Brushett, AJ & Clark, MW 2012, 'Aiptasia pulchella: a tropical cnidarian representative for laboratory ecotoxicological research', Environmental Toxicology and Chemistry, vol. 31, no. 11, pp. 2653-2662.

Howe, PL, Reichelt-Brushett, AJ, Clark, MW & Seery, CR 2017, 'Toxicity estimates for diuron and atrazine for the tropical marine cnidarian Exaiptasia pallida and in-hospite Symbiodinium spp. using PAM chlorophyll-a fluorometry', Journal of Photochemistry and Photobiology B: Biology, vol. 171, pp. 125-132.

Howells, EJ, Bauman, AG, Vaughan, GO, Hume, BCC, Voolstra, CR & Burt, JA 2020, 'Corals in the hottest reefs in the world exhibit symbiont fidelity not flexibility', Molecular Ecology, vol. 29, no. 5, pp. 899-911.

Howells, EJ, Beltran, VH, Larsen, NW, Bay, LK, Willis, BL & van Oppen, MJH 2012, 'Coral thermal tolerance shaped by local adaptation of photosymbionts', Nature Climate Change, vol. 2, no. 2, pp. 116-120.

Howlett, M & Kemmerling, A 2017, 'Calibrating climate change policies: the causes and consequences of sustained under-reaction', Journal of Environmental Policy & Planning, vol. 19, no. 6, pp. 625-637.

Huggett, JF, Foy, CA, Benes, V, Emslie, K, Garson, JA, Haynes, R, Hellemans, J, Kubista, M, Mueller, RD, Nolan, T, Pfaffl, MW, Shipley, GL, Vandesompele, J, Wittwer, CT & Bustin, SA 2013, 'The digital MIQE guidelines: Minimum Information for publication of Quantitative digital PCR Experiments', Clinical Chemistry, vol. 59, no. 6, pp. 892-902.

Huggett, MJ & Apprill, A 2019, 'Coral microbiome database: integration of sequences reveals high diversity and relatedness of coral-associated microbes', Environmental Microbiology Reports, vol. 11, no. 3, pp. 372-385.

Hughes, T, Baird, A, Bellwood, D & Card, M 2003, 'Climate change, human impacts, and the resilience of coral reefs', Science, vol. 301, no. 5635, pp. 929-933. 186

Hughes, TP, Anderson, KD, Connolly, SR, Heron, SF, Kerry, JT, Lough, JM, Baird, AH, Baum, JK, Berumen, ML, Bridge, TC, Claar, DC, Eakin, CM, Gilmour, JP, Graham, NAJ, Harrison, H, Hobbs, J-PA, Hoey, AS, Hoogenboom, M, Lowe, RJ, McCulloch, MT, Pandolfi, JM, Pratchett, M, Schoepf, V, Torda, G & Wilson, SK 2018, 'Spatial and temporal patterns of mass bleaching of corals in the Anthropocene', Science, vol. 359, no. 6371, pp. 80-83.

Hughes, TP, Barnes, ML, Bellwood, DR, Cinner, JE, Cumming, GS, Jackson, JBC, Kleypas, J, van de Leemput, IA, Lough, JM, Morrison, TH, Palumbi, SR, van Nes, EH & Scheffer, M 2017a, 'Coral reefs in the Anthropocene', Nature, vol. 546, pp. 82-90.

Hughes, TP, Kerry, JT, Álvarez-Noriega, M, Álvarez-Romero, JG, Anderson, KD, Baird, AH, Babcock, RC, Beger, M, Bellwood, DR, Berkelmans, R, Bridge, TC, Butler, IR, Byrne, M, Cantin, NE, Comeau, S, Connolly, SR, Cumming, GS, Dalton, SJ, Diaz-Pulido, G, Eakin, CM, Figueira, WF, Gilmour, JP, Harrison, HB, Heron, SF, Hoey, AS, Hobbs, J- PA, Hoogenboom, MO, Kennedy, EV, Kuo, C-y, Lough, JM, Lowe, RJ, Liu, G, McCulloch, MT, Malcolm, HA, McWilliam, MJ, Pandolfi, JM, Pears, RJ, Pratchett, MS, Schoepf, V, Simpson, T, Skirving, WJ, Sommer, B, Torda, G, Wachenfeld, DR, Willis, BL & Wilson, SK 2017b, 'Global warming and recurrent mass bleaching of corals', Nature, vol. 543, pp. 373-377.

Hughes, TP, Kerry, JT, Baird, AH, Connolly, SR, Chase, TJ, Dietzel, A, Hill, T, Hoey, AS, Hoogenboom, MO, Jacobson, M, Kerswell, A, Madin, JS, Mieog, A, Paley, AS, Pratchett, MS, Torda, G & Woods, RM 2019, 'Global warming impairs stock-recruitment dynamics of corals', Nature, vol. 568, pp. 387-390.

Hunter, T 1984, 'The energetics of asexual reproduction: pedal laceration in the symbiotic sea anemone Aiptasia pulchella (Carlgren, 1943)', Journal of Experimental Marine Biology and Ecology, vol. 83, no. 2, pp. 127- 147.

ICZN 2017, 'Opinion 2404 (Case 3633) – Dysactis pallida Agassiz in Verrill, 1864 (currently Aiptasia pallida; Cnidaria, Anthozoa, Hexacorallia, Actiniaria): precedence over Aiptasia diaphana (Rapp, 1829), Aiptasia tagetes (Duchassaing de Fombressin & Michelotti, 1864), Aiptasia mimosa (Duchassaing de Fombressin & Michelotti, 1864) and Aiptasia inula (Duchassaing de Fombressin & Michelotti, 1864) not approved', The Bulletin of Zoological Nomenclature, vol. 74, no. 1, pp. 130-132.

Inda, ME, Broset, E, Lu, TK & de la Fuente-Nunez, C 2019, 'Emerging frontiers in microbiome engineering', Trends in Immunology, vol. 40, no. 10, pp. 952-973.

Jaccard, P 1912, 'The distribution of the flora in the alpine zone', New Phytologist, vol. 11, no. 2, pp. 37-50.

Jackson, DA 1997, 'Compositional data in community ecology: the paradigm or peril of proportions?', Ecology, vol. 78, no. 3, pp. 929-941.

Jiang, D, Armour, CR, Hu, C, Mei, M, Tian, C, Sharpton, TJ & Jiang, Y 2019, 'Microbiome multi-omics network analysis: statistical considerations, limitations, and opportunities', Frontiers in Genetics, vol. 10, 995.

Jiang, L, Hou, L, Zou, X, Zhang, R, Wang, J, Sun, W, Zhao, X & An, J 2007, 'Cloning and expression analysis of p26 gene in Artemia sinica', Acta Biochimica et Biophysica Sinica, vol. 39, no. 5, pp. 351-325.

Jones, RJ, Muller, J, Haynes, D & Schreiber, U 2003, 'Effects of herbicides diuron and atrazine on corals of the Great Barrier Reef, Australia', Marine Ecology Progress Series, vol. 251, pp. 153-167.

Kearse, M, Moir, R, Wilson, A, Stones-Havas, S, Cheung, M, Sturrock, S, Buxton, S, Cooper, A, Markowitz, S, Duran, C, Thierer, T, Ashton, B, Meintjes, P & Drummond, A 2012, 'Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data', Bioinformatics, vol. 28, no. 12, pp. 1647-1649.

Kellogg, CA, Piceno, YM, Tom, LM, DeSantis, TZ, Gray, MA, Zawada, DG & Andersen, GL 2013, 'Comparing bacterial community composition between healthy and White Plague-like disease states in Orbicella annularis using Phylochip™ G3 microarrays', PLoS ONE, vol. 8, no. 11, e79801. 187

Kemp, DW, Rivers, AR, Kemp, KM, Lipp, EK, Porter, JW & Wares, JP 2015, 'Spatial homogeneity of bacterial communities associated with the surface mucus layer of the reef-building coral Acropora palmata', PLoS ONE, vol. 10, no. 12, e0143790.

Kešnerová, L, Mars, RAT, Ellegaard, KM, Troilo, M, Sauer, U & Engel, P 2017, 'Disentangling metabolic functions of bacteria in the honey bee gut', PLoS Biology, vol. 15, no. 12, e2003467.

Khambhaty, Y, Mody, K & Jha, B 2007, 'Purification and characterization of κ-carrageenase from a novel γ- proteobacterium, Pseudomonas elongata (MTCC 5261) syn. Microbulbifer elongatus comb. Nov', Biotechnology and Bioprocess Engineering, vol. 12, no. 6, pp. 668-675.

Kimes, NE, Grim, CJ, Johnson, WR, Hasan, NA, Tall, BD, Kothary, MH, Kiss, H, Munk, AC, Tapia, R, Green, L, Detter, C, Bruce, DC, Brettin, TS, Colwell, RR & Morris, PJ 2012, 'Temperature regulation of virulence factors in the pathogen Vibrio coralliilyticus', International Society for Microbial Ecology Journal, vol. 6, no. 4, pp. 835- 846.

Kimes, NE, Johnson, WR, Torralba, M, Nelson, KE, Weil, E & Morris, PJ 2013, 'The Montastraea faveolata microbiome: ecological and temporal influences on a Caribbean reef-building coral in decline', Environmental Microbiology, vol. 15, no. 7, pp. 2082-2094.

King, DW, Berger, E, Helm, Z, Irish, E & Mopper, K 2016, 'Measurement of antioxidant activity toward superoxide in natural waters', Frontiers in Marine Science, vol. 3,

Kitahara, MV, Fukami, H, Benzoni, F & Huang, D 2016, 'The new systematics of Scleractinia: integrating molecular and morphological evidence', in S Goffredo and Z Dubinsky (eds), The Cnidaria, past, present and future: the world of Medusa and her sisters, Springer, Cham, pp. 41-59.

Koren, O & Rosenberg, E 2008, 'Bacteria associated with the bleached and cave coral Oculina patagonica', Microbial Ecology, vol. 55, pp. 523-529.

Krediet, CJ, Ritchie, KB, Paul, VJ & Teplitski, M 2013, 'Coral-associated micro-organisms and their roles in promoting coral health and thwarting diseases', Proceedings of the Royal Society B: Biological Sciences, vol. 280, 20122328.

Kruskal, WH & Wallis, WA 1952, 'Use of ranks in one-criterion variance analysis', Journal of the American Statistical Association, vol. 47, no. 260, pp. 583-621.

Kuek, FWI, Lim, L-F, Ngu, L-H, Mujahid, A, Lim, P-T, Leaw, C-P & Müller, M 2015, 'The potential roles of bacterial communities in coral defence: a case study at Talang-Talang reef', Ocean Science Journal, vol. 50, no. 2, pp. 269-282.

Kümmerli, R, Jiricny, N, Clarke, LS, West, SA & Griffin, AS 2009, 'Phenotypic plasticity of a cooperative behaviour in bacteria', Journal of Evolutionary Biology, vol. 22, no. 3, pp. 589-598.

Kvennefors, ECE, Sampayo, E, Kerr, C, Vieira, G, Roff, G & Barnes, AC 2011, 'Regulation of bacterial communities through antimicrobial activity by the coral holobiont', Microbial Ecology, vol. 63, no. 3, pp. 605-618.

Kvennefors, ECE, Sampayo, E, Ridgway, T, Barnes, AC & Hoegh-Guldberg, O 2010, 'Bacterial communities of two ubiquitous Great Barrier Reef corals reveals both site- and species-specificity of common bacterial associates', PLoS ONE, vol. 5, no. 4, e10401.

Lahti, L & Shetty, S 2017, microbiome: tools for microbiome analysis in R [software] R package version 1.7.2, viewed 1 Oct 2019, .

LaJeunesse, T 2002, 'Diversity and community structure of symbiotic dinoflagellates from Caribbean coral reefs', Marine Biology, vol. 141, no. 2, pp. 387-400. 188

LaJeunesse, TC, Parkinson, JE, Gabrielson, PW, Jeong, HJ, Reimer, JD, Voolstra, CR & Santos, SR 2018, 'Systematic revision of Symbiodiniaceae highlights the antiquity and diversity of coral endosymbionts', Current Biology, vol. 28, no. 16, pp. 2570-2580.

Lane, DJ 1991, '16S/23S rRNA sequencing', in E Stackebrandt and M Goodfellow (eds), Nucleic acid techniques in bacterial systematics, John Wiley, NY, pp. 115-175.

Lawler, SN, Kellogg, CA, France, SC, Clostio, RW, Brooke, SD & Ross, SW 2016, 'Coral-associated bacterial diversity is conserved across two deep-sea Anthothela species', Frontiers in Microbiology, vol. 7, 458.

Lawson, CA, Raina, J-B, Kahlke, T, Seymour, JR & Suggett, DJ 2018, 'Defining the core microbiome of the symbiotic dinoflagellate, Symbiodinium', Environmental Microbiology Reports, vol. 10, no. 1, pp. 7-11.

Laycock, G, Sait, L, Inman, C, Lewis, M, Smidt, H, van Diemen, P, Jorgensen, F, Stevens, M & Bailey, M 2012, 'A defined intestinal colonization microbiota for gnotobiotic pigs', Veterinary Immunology and Immunopathology, vol. 149, no. 3, pp. 216-224.

Leal, MC, Hoadley, K, Pettay, DT, Grajales, A, Calado, R & Warner, ME 2015, 'Symbiont type influences trophic plasticity of a model cnidarian–dinoflagellate symbiosis', Journal of Experimental Biology, vol. 218, no. 6, pp. 858-863.

Leal, MC, Nunes, C, Engrola, S, Dinis, MT & Calado, R 2012, 'Optimization of monoclonal production of the glass anemone Aiptasia pallida (Agassiz in Verrill, 1864)', Aquaculture, vol. 354, pp. 91-96.

Lee, STM, Davy, SK, Tang, S-L & Kench, PS 2016, 'Mucus sugar content shapes the bacterial community structure in thermally stressed Acropora muricata', Frontiers in Microbiology, vol. 7, 371.

Lehnert, EM, Mouchka, ME, Burriesci, MS, Gallo, ND, Schwarz, JA & Pringle, JR 2014, 'Extensive differences in gene expression between symbiotic and aposymbiotic cnidarians', G3: Genes|Genomes|Genetics, vol. 4, no. 2, pp. 277-295.

Leigh, BA, Liberti, A & Dishaw, LJ 2016, 'Generation of germ-free Ciona intestinalis for studies of gut-microbe interactions', Frontiers in Microbiology, vol. 7, 2092.

Lema, KA, Willis, BL & Bourne, DG 2012, 'Corals form characteristic associations with symbiotic nitrogen-fixing bacteria', Applied and Environmental Microbiology, vol. 78, no. 9, pp. 3136-3144.

Lenhoff, H (ed.) 1983, Hydra: research methods, Springer, New York.

Lesser, MP 1996, 'Elevated temperatures and ultraviolet radiation cause oxidative stress and inhibit photosynthesis in symbiotic dinoflagellates', Limnology and Oceanography, vol. 41, no. 2, pp. 271-283.

Lesser, MP 1997, 'Oxidative stress causes coral bleaching during exposure to elevated temperatures', Coral Reefs, vol. 16, no. 3, pp. 187-192.

Lesser, MP, Falcón, LI, Rodríguez-Román, A, Enriíquez, S, Hoegh-Guldberg, O & Iglesias-Prieto, R 2007, 'Nitrogen fixation by symbiotic cyanobacteria provides a source of nitrogen for the scleractinian coral Montastraea cavernosa', Marine Ecology Progress Series, vol. 346, pp. 143-152.

Lesser, MP, Mazel, CH, Gorbunov, MY & Falkowski, PG 2004, 'Discovery of symbiotic nitrogen-fixing Cyanobacteria in corals', Science, vol. 305, no. 5686, pp. 997-1000.

Levene, H 1960, 'Robust tests for equality of variances', in I Olkin, SG Ghurye, W Hoeffding, WG Madow and HB Mann (eds), Contributions to probability and statistics: essays in honor of Harold Hotelling, Stanford University Press, Menlo Park, pp. 278-292.

189

Levin, RA, Beltran, VH, Hill, R, Kjelleberg, S, McDougald, D, Steinberg, PD & van Oppen, MJH 2016, 'Sex, scavengers, and chaperones: transcriptome secrets of divergent symbiodinium thermal tolerances', Molecular Biology and Evolution, vol. 33, no. 9, pp. 2201-2215.

Lewin, RA 1997, 'Saprospira grandis: a flexibacterium that can catch bacterial prey by ``ixotrophy''', Microbial Ecology, vol. 34, no. 3, pp. 232-236.

Lewis, C, Neely, K & Rodriguez-Lanetty, M 2019, 'Recurring episodes of thermal stress shift the balance from a dominant host-specialist to a background host-generalist zooxanthella in the threatened pillar coral, Dendrogyra cylindrus', Frontiers in Marine Science, vol. 6, 5.

Lewis, SE, Brodie, JE, Bainbridge, ZT, Rohde, KW, Davis, AM, Masters, BL, Maughan, M, Devlin, MJ, Mueller, JF & Schaffelke, B 2009, 'Herbicides: a new threat to the Great Barrier Reef', Environmental Pollution, vol. 157, no. 8, pp. 2470-2484.

Li, J, Chen, Q, Long, L-J, Dong, J-D, Yang, J & Zhang, S 2014, 'Bacterial dynamics within the mucus, tissue and skeleton of the coral Porites lutea during different seasons', Scientific Reports, vol. 4, 7320.

Li, J, Chen, Q, Zhang, S, Huang, H, Yang, J, Tian, X-P & Long, L-J 2013, 'Highly heterogeneous bacterial communities associated with the South China Sea reef corals Porites lutea, Galaxea fascicularis and Acropora millepora', PLoS ONE, vol. 8, no. 8, e71301.

Lindow, SE 1987, 'Competitive exclusion of epiphytic bacteria by Ice¯ Pseudomonas syringae mutants', Applied and Environmental Microbiology, vol. 53, no. 10, pp. 2520-2527.

Lipps, JH & Stanley, GD 2016, 'Reefs through time: an evolutionary view', in KD Hubbard, SC Rogers, HJ Lipps and JDG Stanley (eds), Coral reefs at the crossroads, Springer, Dordrecht, pp. 175-196.

Little, AF, van Oppen, MJH & Willis, BL 2004, 'Flexibility in algal endosymbioses shapes growth in reef corals', Science, vol. 304, no. 5676, pp. 1492-1494.

Littman, RA, Willis, BL, Pfeffer, C & Bourne, DG 2009, 'Diversities of coral-associated bacteria differ with location, but not species, for three acroporid corals on the Great Barrier Reef', FEMS Microbiology Ecology, vol. 68, no. 2, pp. 152-163.

López-Pérez, M, Gonzaga, A, Martin-Cuadrado, A-B, Onyshchenko, O, Ghavidel, A, Ghai, R & Rodriguez-Valera, F 2012, 'Genomes of surface isolates of Alteromonas macleodii: the life of a widespread marine opportunistic copiotroph', Scientific Reports, vol. 2, 696.

Lough, JM 2011, 'Climate change and coral reefs', in D Hopley (ed.) Encyclopedia of modern coral reefs, Springer, Dordrecht, pp. 198-210.

Love, MI, Huber, W & Anders, S 2014, 'Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2', Genome Biology, vol. 15, no. 12, 550.

Lozupone, C, Lladser, ME, Knights, D, Stombaugh, J & Knight, R 2011, 'UniFrac: an effective distance metric for microbial community comparison', International Society for Microbial Ecology Journal, vol. 5, pp. 169–172.

Luckey, TD 1963a, 'Decontamination', in Germfree life and gnotobiology, Academic Press, New York, pp. 206- 209.

Luckey, TD 1963b, 'Germfree invertebrates', in Germfree life and gnotobiology, Academic Press, New York, pp. 43-60.

Lynch, JB & Hsiao, EY 2019, 'Microbiomes as sources of emergent host phenotypes', Science, vol. 365, no. 6460, pp. 1405-1409. 190

Macpherson, AJ & McCoy, KD 2015, 'Standardised animal models of host microbial mutualism', Mucosal immunology, vol. 8, no. 3, pp. 476-486.

Makita, H 2018, 'Iron-oxidizing bacteria in marine environments: recent progresses and future directions', World Journal of Microbiology & Biotechnology, vol. 34, 110.

Maltais, A 2014, 'Failing international climate politics and the fairness of going first', Political Studies, vol. 62, no. 3, pp. 618-633.

Mao-Jones, J, Ritchie, KB, Jones, LE & Ellner, SP 2010, 'How microbial community composition regulates coral disease development', PLoS Biology, vol. 8, no. 3, e1000345.

Maragos, J 2011, 'Coral reef at Palmyra Atoll National Wildlife Refuge' [image], in United States Fish and Wildlife Service – Pacific region, flickr, viewed 16 Dec 2019, .

Margulis, L 1991, 'Symbiogenesis and symbionticism', in L Margulis and R Fester (eds), Symbiosis as a source of evolutionary innovation: speciation and morphogenesis, MIT Press, Cambridge, pp. 1-14.

Martin, M 2011, 'Cutadapt removes adapter sequences from high-throughput sequencing reads', EMBnet, vol. 17, no. 1, pp. 10-12.

Marty-Rivera, M, Yudowski, G & Roberson, L 2018, 'Mitigation of coral bleaching by antioxidants' [preprint] bioRxiv, viewed 1 Mar 2019, .

Massey, FJ 1951, 'The Kolmogorov-Smirnov Test for goodness of fit', Journal of the American Statistical Association, vol. 46, no. 253, pp. 68-78.

Matthews, JL, Crowder, CM, Oakley, CA, Lutz, A, Roessner, U, Meyer, E, Grossman, AR, Weis, VM & Davy, SK 2017, 'Optimal nutrient exchange and immune responses operate in partner specificity in the cnidarian- dinoflagellate symbiosis', Proceedings of the National Academy of Sciences, USA, vol. 114, no. 50, pp. 13194- 13199.

Matthews, JL, Sproles, AE, Oakley, CA, Grossman, AR, Weis, VM & Davy, SK 2016, 'Menthol-induced bleaching rapidly and effectively provides experimental aposymbiotic sea anemones (Aiptasia sp.) for symbiosis investigations', Journal of Experimental Biology, vol. 219, no. 3, pp. 306-310.

McClanahan, TR, Weil, E, Cortés, J, Baird, AH & Ateweberhan, M 2009, 'Consequences of coral bleaching for sessile reef organisms', in MJH van Oppen and JM Lough (eds), Coral bleaching: patterns, processes, causes and consequences, Springer, Berlin, pp. 121-138.

McFall-Ngai, M, Hadfield, MG, Bosch, TCG, Carey, HV, Domazet-Lošo, T, Douglas, AE, Dubilier, N, Eberl, G, Fukami, T, Gilbert, SF, Hentschel, U, King, N, Kjelleberg, S, Knoll, AH, Kremer, N, Mazmanian, SK, Metcalf, JL, Nealson, K, Pierce, NE, Rawls, JF, Reid, A, Ruby, EG, Rumpho, M, Sanders, JG, Tautz, D & Wernegreen, JJ 2013, 'Animals in a bacterial world, a new imperative for the life sciences', Proceedings of the National Academy of Sciences, USA, vol. 110, no. 9, pp. 3229-3236.

McKinstry, MJ, Chapman, GB, Spoon, DM & Peters, EC 1989, 'The occurrence of bacterial colonies in the epidermis of the tentacles of the sea anemone Aiptasia pallida (Anthozoa: Actinaria)', Transactions of the American Microscopical Society, vol. 108, no. 3, pp. 239-244.

McMurdie, PJ & Holmes, S 2013, 'phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data', PLoS ONE, vol. 8, no. 4, e61217.

191

Medrano, E, Merselis, DG, Bellantuono, AJ & Rodriguez-Lanetty, M 2019, 'Proteomic basis of symbiosis: a heterologous partner fails to duplicate homologous colonization in a novel cnidarian-Symbiodiniaceae mutualism', Frontiers in Microbiology, vol. 10, 1153.

Meron, D, Atias, E, Iasur Kruh, L, Elifantz, H, Minz, D, Fine, M & Banin, E 2011, 'The impact of reduced pH on the microbial community of the coral Acropora eurystoma', International Society for Microbial Ecology Journal, vol. 5, no. 1, pp. 51-60.

Meunier, V, Bonnet, S, Pernice, M, Benavides, M, Lorrain, A, Grosso, O, Lambert, C & Houlbrèque, F 2019, 'Bleaching forces coral’s heterotrophy on diazotrophs and Synechococcus', International Society for Microbial Ecology Journal, vol. 13, pp. 2882–2886.

Meyer, FW, Cardini, U & Wild, C 2015, 'Ocean acidification and related indicators', in RH Armon and O Hänninen (eds), Environmental Indicators, Springer, Dordrecht, pp. 723-742.

Mieog, JC, van Oppen, MJH, Berkelmans, R, Stam, WT & Olsen, JL 2009, 'Quantification of algal endosymbionts (Symbiodinium) in coral tissue using real-time PCR', Molecular Ecology Resources, vol. 9, no. 1, pp. 74-82.

Mikaelyan, A, Thompson, CL, Hofer, MJ & Brune, A 2015, 'Deterministic assembly of complex bacterial communities in guts of germ-free cockroaches', Applied and Environmental Microbiology, vol. 82, no. 4, pp. 1256-1263.

Mills, E, Shechtman, K, Loya, Y & Rosenberg, E 2013, 'Bacteria appear to play important roles in both causing and preventing the bleaching of the coral Oculina patagonica', Marine Ecology Progress Series, vol. 489, pp. 155-162.

Mimee, M, Citorik, RJ & Lu, TK 2016, 'Microbiome therapeutics – advances and challenges', Advanced Drug Delivery Reviews, vol. 105, pp. 44-54.

Moberg, F & Folke, C 1999, 'Ecological goods and services of coral reef ecosystems', Ecological Economics, vol. 29, no. 2, pp. 215-233.

Mooser, C, Gomez de Agüero, M & Ganal-Vonarburg, SC 2018, 'Standardization in host–microbiota interaction studies: challenges, gnotobiology as a tool, and perspective', Current Opinion in Microbiology, vol. 44, pp. 50- 60.

Mouchka, ME, Hewson, I & Harvell, CD 2010, 'Coral-associated bacterial assemblages: current knowledge and the potential for climate-driven impacts', Integrative and Comparative Biology, vol. 50, no. 4, pp. 662-674.

Müller, DB, Vogel, C, Bai, Y & Vorholt, JA 2016, 'The plant microbiota: systems-level insights and perspectives', Annual Review of Genetics, vol. 50, pp. 211-234.

Muller-Parker, G & Davy, SK 2001, 'Temperate and tropical algal-sea anemone symbioses', Invertebrate Biology, vol. 120, no. 2, pp. 104-123.

Munn, CB 2015, 'The role of Vibrios in diseases of corals', Microbiology Spectrum, vol. 3, no. 4, VE-0006-2014.

Muscatine, L, Falkowski, PG & Dubinsky, Z 1984, 'Fate of photosynthetic fixed carbon in light- and shade- adapted colonies of the symbiotic coral Stylophora pistillata', Proceedings of the Royal Society B: Biological Sciences, vol. 222, no. 1227, pp. 181-202.

Muscatine, L, Falkowski, PG, Dubinsky, Z, Cook, PA, McCloskey, LR & Smith, DC 1989, 'The effect of external nutrient resources on the population dynamics of zooxanthellae in a reef coral', Proceedings of the Royal Society B: Biological Sciences, vol. 236, no. 1284, pp. 311-324.

192

Muscatine, L, McCloskey, L & Marian, RE 1981, 'Estimating the daily contribution of carbon from zooxanthellae to coral animal respirarion', Limnology and Oceanography, vol. 26, no. 4, pp. 601-611.

Muyzer, G, de Waal, EC & Uitterlinden, AG 1993, 'Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA', Applied and Environmental Microbiology, vol. 59, no. 3, pp. 695-700.

Mydlarz, LD, McGinty, ES & Harvell, CD 2010, 'What are the physiological and immunological responses of coral to climate warming and disease?', Journal of Experimental Biology, vol. 213, no. 6, pp. 934-945.

National Academies of Sciences Engineering and Medicine 2019, A research review of interventions to increase the persistence and resilience of coral reefs, The National Academies Press, Washington, DC.

Neville, BA, Forster, SC & Lawley, TD 2018, 'Commensal Koch's postulates: establishing causation in human microbiota research', Current Opinion in Microbiology, vol. 42, pp. 47-52.

Nissimov, J, Rosenberg, E & Munn, C 2009, 'Antimicrobial properties of resident coral mucus bacteria of Oculina patagonica', FEMS Microbiology Letters, vol. 292, no. 2, pp. 210-215.

Niu, Y, Defoirdt, T, Baruah, K, Van de Wiele, T, Dong, S & Bossier, P 2014, 'Bacillus sp. LT3 improves the survival of gnotobiotic brine shrimp (Artemia franciscana) larvae challenged with Vibrio campbellii by enhancing the innate immune response and by decreasing the activity of shrimp-associated vibrios', Veterinary Microbiology, vol. 173, no. 3, pp. 279-288.

Noecker, C, Chiu, H-C, McNally, CP & Borenstein, E 2019, 'Defining and evaluating microbial contributions to metabolite variation in microbiome-metabolome association studies', mSystems, vol. 4, no. 6, e00579-19.

Núñez-Pons, L, Bertocci, I & Baghdasarian, G 2017, 'Symbiont dynamics during thermal acclimation using cnidarian-dinoflagellate model holobionts', Marine Environmental Research, vol. 130, pp. 303-314.

Nur, A, Reichelt-Brushett, A & Howe, P 2019, 'Nickel uptake in Symbiodynium spp. and host cnidarian – assessment using the model cnidarian Exaiptasia pallida', IOP Conference Series: Earth and Environmental Science, vol. 339, 012025.

Oakley, CA, Ameismeier, MF, Peng, L, Weis, VM, Grossman, AR & Davy, SK 2016, 'Symbiosis induces widespread changes in the proteome of the model cnidarian Aiptasia', Cellular Microbiology, vol. 18, no. 7, pp. 1009-1023.

Oakley, CA, Durand, E, Wilkinson, SP, Peng, L, Weis, VM, Grossman, AR & Davy, SK 2017, 'Thermal shock induces host proteostasis disruption and endoplasmic reticulum stress in the model symbiotic cnidarian Aiptasia', Journal of Proteome Research, vol. 16, no. 6, pp. 2121-2134.

Oberbeckmann, S, Osborn, AM & Duhaime, MB 2016, 'Microbes on a bottle: substrate, season and geography influence community composition of microbes colonizing marine plastic debris', PLoS ONE, vol. 11, no. 8, e0159289.

Oksanen, J, Blanchet, FG, Kindt, R, Legendre, P, Minchin, PR, O'Hara, RB, Simpson, GL, Solymos, P, Stevens, MHH & Wagner, H 2018, vegan: community ecology package [software] R package version 2.5-6, viewed 1 Jul 2019, .

Oliver, JK, Berkelmans, R & Eakin, CM 2018, 'Coral bleaching in space and time', in MJH van Oppen and JM Lough (eds), Coral bleaching: patterns, processes, causes and consequences, Springer, Cham, pp. 27-49.

Ostrander, GK, Armstrong, KM, Knobbe, ET, Gerace, D & Scully, EP 2000, 'Rapid transition in the structure of a coral reef community: the effects of coral bleaching and physical disturbance', Proceedings of the National Academy of Sciences, USA, vol. 97, no. 10, pp. 5297-5302.

193

Palincsar, EE, Jones, WR, Palincsar, JS, Glogowski, MA & Mastro, JL 1989, 'Bacterial aggregates within the epidermis of the sea anemone Aiptasia pallida', Biological Bulletin, vol. 177, no. 1, pp. 130-140.

Parkinson, JE, Coffroth, MA & LaJeunesse, TC 2015, 'New species of Clade B Symbiodinium (dinophyceae) from the greater Caribbean belong to different functional guilds: S. aenigmaticum sp. Nov., S. antillogorgium sp. Nov., S. endomadracis sp. Nov., and S. pseudominutum sp. Nov', Journal of Phycology, vol. 51, no. 5, pp. 850- 858.

Pedregosa, F, Varoquaux, G, Gramfort, A, Michel, V, Thirion, B, Grisel, O, Blondel, M, Prettenhofer, P, Weiss, R, Dubourg, V, Vanderplas, J, Passos, A, Cournapeau, D, Brucher, M, Perrot, M & Duchesnay, E 2011, 'Scikit-learn: machine learning in Python', Journal of Machine Learning Research, vol. 12, no. Oct, pp. 2825-2830.

Peixoto, RS, Rosado, PM, Leite, DCA, Rosado, AS & Bourne, DG 2017, 'Beneficial microorganisms for corals (BMC): proposed mechanisms for coral health and resilience', Frontiers in Microbiology, vol. 8, 341.

Perez, SF, Cook, CB & Brooks, WR 2001, 'The role of symbiotic dinoflagellates in the temperature-induced bleaching response of the subtropical sea anemone Aiptasia pallida', Journal of Experimental Marine Biology and Ecology, vol. 256, no. 1, pp. 1-14.

Pernice, M, Meibom, A, Van Den Heuvel, A, Kopp, C, Domart-Coulon, I, Hoegh-Guldberg, O & Dove, S 2012, 'A single-cell view of ammonium assimilation in coral-dinoflagellate symbiosis', International Society for Microbial Ecology Journal, vol. 6, pp. 1314-1324.

Pineiro, SA, Stine, OC, Chauhan, A, Steyert, S, Smith, R & Williams, H 2007, 'Global survey of diversity among environmental saltwater Bacteriovoracaceae', Environmental Microbiology, vol. 9, no. 10, pp. 2441-2450.

Pinheiro, J, Bates, D, DebRoy, S, Sarkar, D & R Core Team 2019, nlme: linear and nonlinear mixed effects models [software] R Package version 3.1-141, viewed 1 Dec 2019, .

Plovie, A 2010, Comparison of bacterial communities associated with healthy and bleached Aiptasia pallida, a novel model organism for coral studies: implications and variation during bleaching, Masters thesis, Department of Biochemistry and Microbiology, Ghent University, Belgium, viewed 30 Nov 2018, .

Pochon, X, Pawlowski, J, Zaninetti, L & Rowan, R 2001, 'High genetic diversity and relative specificity among Symbiodinium-like endosymbiotic dinoflagellates in soritid foraminiferans', Marine Biology, vol. 139, no. 6, pp. 1069-1078.

Polne-Fuller, M 1991, 'A novel technique for preparation of axenic cultures of Symbiodinium (pyrrophyta) through selective digestion by amoebae', Journal of Phycology, vol. 27, no. 4, pp. 552-554.

Pratte, ZA, Richardson, LL & Mills, DK 2015, 'Microbiota shifts in the surface mucopolysaccharide layer of corals transferred from natural to aquaria settings', Journal of Invertebrate Pathology, vol. 125, pp. 42-44.

Props, R, Kerckhof, FM, Rubbens, P, De Vrieze, J, Hernandez Sanabria, E, Waegeman, W, Monsieurs, P, Hammes, F & Boon, N 2017, 'Absolute quantification of microbial taxon abundances', International Society for Microbial Ecology Journal, vol. 11, no. 2, pp. 584-587.

Pujalte, MJ, Lucena, T, Rodrigo-Torres, L & Arahal, DR 2018, 'Comparative genomics of Thalassobius Including the description of Thalassobius activus sp. nov., and Thalassobius autumnalis sp. nov', Frontiers in Microbiology, vol. 8, 2645.

Quast, C, Pruesse, E, Yilmaz, P, Gerken, J, Schweer, T, Yarza, P, Peplies, J & Glöckner, FO 2013, 'The SILVA ribosomal RNA gene database project: improved data processing and web-based tools', Nucleic Acids Research, vol. 41, pp. D590-D596.

194

R Core Team 2018, R: a language and environment for statisitical computing [software] version 3.6.0, viewed 1 Feb 2019, .

Rädecker, N, Pogoreutz, C, Voolstra, CR, Wiedenmann, J & Wild, C 2015, 'Nitrogen cycling in corals: the key to understanding holobiont functioning?', Trends in Microbiology, vol. 23, no. 8, pp. 490-497.

Rädecker, N, Raina, J-B, Pernice, M, Perna, G, Guagliardo, P, Kilburn, MR, Aranda, M & Voolstra, CR 2018, 'Using Aiptasia as a model to study metabolic interactions in cnidarian-Symbiodinium symbioses', Frontiers in Physiology, vol. 9, 214.

Rahat, M & Dimentman, C 1982, 'Cultivation of bacteria-free Hydra viridis: missing budding factor in nonsymbiotic hydra', Science, vol. 216, no. 4541, pp. 67-68.

Raina, J-B, Clode, PL, Cheong, S, Bougoure, J, Kilburn, MR, Reeder, A, Forêt, S, Stat, M, Beltran, V, Thomas-Hall, P, Tapiolas, D, Motti, CM, Gong, B, Pernice, M, Marjo, CE, Seymour, JR, Willis, BL & Bourne, DG 2017, 'Subcellular tracking reveals the location of dimethylsulfoniopropionate in microalgae and visualises its uptake by marine bacteria', eLife, vol. 6, e23008.

Raina, J-B, Tapiolas, DM, Foret, S, Lutz, A, Abrego, D, Ceh, J, Seneca, FO, Clode, PL, Bourne, DG, Willis, BL & Motti, CA 2013, 'DMSP biosynthesis by an animal and its role in coral thermal stress response', Nature, vol. 502, no. 7473, pp. 677-680.

Raina, JB, Tapiolas, D, Willis, BI & Bourne, DG 2009, 'Coral-associated bacteria and their role in the biogeochemical cycling of sulfur', Applied and Environmental Microbiology, vol. 75, no. 11, pp. 3492-3501.

Ralph, PJ, Hill, R, Doblin, MA & Davy, SK 2016, 'Theory and application of pulse amplitude modulated chlorophyll fluorometry in coral health assessment', in CM Woodley, CA Downs, AW Bruckner, JW Porter and SB Galloway (eds), Diseases of coral, Wiley, Hoboken, pp. 506-523.

Ratcliff, MJ 2012, 'The Trembley Effect or the birth of marine zoology', International Journal of Developmental Biology, vol. 56, pp. 425-436.

Rawls, JF, Samuel, BS & Gordon, JI 2004, 'Gnotobiotic zebrafish reveal evolutionarily conserved responses to the gut microbiota', Proceedings of the National Academy of Sciences, USA, vol. 101, no. 13, pp. 4596-4601.

Reddy, TBK, Thomas, AD, Stamatis, D, Bertsch, J, Isbandi, M, Jansson, J, Mallajosyula, J, Pagani, I, Lobos, EA & Kyrpides, NC 2015, 'The Genomes OnLine Database (GOLD) v.5: a metadata management system based on a four level (meta)genome project classification', Nucleic Acids Research, vol. 43, pp. D1099-D1106.

Reichelt, RE, Bradbury, RH & Moran, PJ 1990, 'The crown-of-thorns starfish, Acanthaster planci, on the Great Barrier Reef', Mathematical and Computer Modelling, vol. 13, no. 6, pp. 45-60.

Reshef, L, Koren, O, Loya, Y, Zilber-Rosenberg, I & Rosenberg, E 2006, 'The coral probiotic hypothesis', Environmental Microbiology, vol. 8, no. 12, pp. 2068-2073.

Reyes-Bermudez, A & Miller, DJ 2009, 'In vitro culture of cells derived from larvae of the staghorn coral Acropora millepora', Coral Reefs, vol. 28, no. 4, pp. 859-864.

Richardson, C, Hill, M, Marks, C, Runyen-Janecky, L & Hill, A 2012, 'Experimental manipulation of sponge/bacterial symbiont community composition with antibiotics: sponge cell aggregates as a unique tool to study animal/microorganism symbiosis', FEMS Microbiology Ecology, vol. 81, no. 2, pp. 407-418.

Ridley, EV, Wong, ACN & Douglas, AE 2013, 'Microbe-dependent and nonspecific effects of procedures to eliminate the resident microbiota from Drosophila melanogaster', Applied and Environmental Microbiology, vol. 79, no. 10, pp. 3209-3214.

195

Rinkevich, B 2019, 'The active reef restoration toolbox is a vehicle for coral resilience and adaptation in a changing world', Journal of Marine Science and Engineering, vol. 7, no. 7, 201.

Ritchie, KB 2006, 'Regulation of microbial populations by coral surface mucus and mucus-associated bacteria', Marine Ecology Progress Series, vol. 322, pp. 1-14.

Roberts, DW 2016, labdsv: ordination and multivariate analysis for ecology [software] R package version 1.8.0, viewed 31 Dec 2018, .

Rocca, JD, Simonin, M, Blaszczak, JR, Ernakovich, JG, Gibbons, SM, Midani, FS & Washburne, AD 2019, 'The Microbiome Stress Project: toward a global meta-analysis of environmental stressors and their effects on microbial communities', Frontiers in Microbiology, vol. 9, 3272.

Rocha, L 2019a, 'Moorea1' [image], in California Academy of Sciences, Twitter, viewed 18 May 2019, .

Rocha, L 2019b, 'Moorea2' [image], in California Academy of Sciences, Twitter, viewed 18 May 2019, .

Rodriguez-Lanetty, M, Granados-Cifuentes, C, Barberan, A, Bellantuono, AJ & Bastidas, C 2013, 'Ecological inferences from a deep screening of the complex bacterial consortia associated with the coral, Porites astreoides', Molecular Ecology, vol. 22, no. 16, pp. 4349-4362.

Rognes, T, Flouri, T, Nichols, B, Quince, C & Mahé, F 2016, 'VSEARCH: a versatile open source tool for metagenomics', PeerJ, vol. 4, e2584.

Rohwer, F & Kelley, S 2004, 'Culture-independent analyses of coral-associated microbes', in E Rosenberg and Y Loya (eds), Coral Health and Disease, Springer, Berlin, pp. 265-277.

Rohwer, F, Seguritan, V, Azam, F & Knowlton, N 2002, 'Diversity and distribution of coral-associated bacteria', Marine Ecology Progress Series, vol. 243, pp. 1-10.

Rosado, PM, Leite, DCA, Duarte, GAS, Chaloub, RM, Jospin, G, Nunes da Rocha, U, P. Saraiva, J, Dini-Andreote, F, Eisen, JA, Bourne, DG & Peixoto, RS 2018, 'Marine probiotics: increasing coral resistance to bleaching through microbiome manipulation', International Society for Microbial Ecology Journal, vol. 13, pp. 921–936.

Rosenberg, E & Falkovitz, L 2004, 'The Vibrio shiloi/Oculina patagonica model system of coral bleaching', Annual Review of Microbiology, vol. 58, no. 1, pp. 143-159.

Röthig, T, Costa, RM, Simona, F, Baumgarten, S, Torres, AF, Radhakrishnan, A, Aranda, M & Voolstra, CR 2016, 'Distinct bacterial communities associated with the coral model Aiptasia in aposymbiotic and symbiotic states with Symbiodinium', Frontiers in Marine Science, vol. 3, 234.

Round, JL 2013, 'From fly to human: understanding how commensal microorganisms influence host immunity and health', in DN Fredricks (ed.) The Human Microbiota, John Wiley, Hoboken, pp. 255-272.

Ruiz-Ponte, C, Samain, JF, Sánchez, JL & Nicolas, JL 1999, 'The benefit of a Roseobacter species on the survival of scallop larvae', Marine Biotechnology, vol. 1, no. 1, pp. 52-59.

Ruppert, EE, Fox, RS & Barnes, RB 2004, Invertebrate zoology: a functional evolutionary approach, 7th edn., Brooks, Belmont, MI.

Rypien, KL, Ward, JR & Azam, F 2010, 'Antagonistic interactions among coral-associated bacteria', Environmental Microbiology, vol. 12, no. 1, pp. 28-39.

196

Sabourault, C, Ganot, P, Deleury, E, Allemand, D & Furla, P 2009, 'Comprehensive EST analysis of the symbiotic sea anemone, Anemonia viridis', BMC Genomics, vol. 10, no. 1, 333.

Salo, WL, Aufderheide, AC, Buikstra, J & Holcomb, TA 1994, 'Identification of Mycobacterium tuberculosis DNA in a pre-Columbian Peruvian mummy', Proceedings of the National Academy of Sciences, USA, vol. 91, no. 6, pp. 2091-2094.

Santos, HF, Carmo, FL, Duarte, G, Dini-Andreote, F, Castro, CB, Rosado, AS, van Elsas, JD & Peixoto, RS 2014, 'Climate change affects key nitrogen-fixing bacterial populations on coral reefs', International Society for Microbial Ecology Journal, vol. 8, no. 11, pp. 2272-2279.

Saravanan, P, Nancharaiah, Y, Venugopalan, V, Subba Rao, T & Jayachandran, S 2006, 'Biofilm formation by Pseudoalteromonas ruthenica and its removal by chlorine', Biofouling, vol. 22, pp. 371-381.

Schlesinger, A, Kramarsky-Winter, E, Rosenfeld, H, Armoza-Zvoloni, R & Loya, Y 2010, 'Sexual plasticity and self- fertilization in the sea anemone Aiptasia diaphana', PLoS ONE, vol. 5, no. 7, e11874.

Schwenk, D, Nohynek, L & Rischer, H 2014, 'Algae–bacteria association inferred by 16S rDNA similarity in established microalgae cultures', MicrobiologyOpen, vol. 3, no. 3, pp. 356-368.

Searle, SR, Speed, FM & Milliken, GA 1980, 'Population marginal means in the linear model: an alternative to least squares means', The American Statistician, vol. 34, no. 4, pp. 216-221.

Sekar, R, Mills, DK, Remily, ER, Voss, JD & Richardson, LL 2006, 'Microbial communities in the surface mucopolysaccharide layer and the black band microbial mat of black band-diseased Siderastrea siderea', Applied and Environmental Microbiology, vol. 72, no. 9, pp. 5963-5973.

Sender, R, Fuchs, S & Milo, R 2016, 'Revised estimates for the number of human and bacteria cells in the body', PLoS Biology, vol. 14, no. 8, e1002533.

Séré, MG 2014, Identification of a primary pathogen involved in white patch syndrome, a newly-reported disease affecting the massive coral Porites lutea in the Western Indian Ocean, PhD thesis, Universite de la Reunion, Reunion Island, viewed 17 Oct 2019, .

Shade, A, Jones, SE, Caporaso, JG, Handelsman, J, Knight, R, Fierer, N & Gilbert, JA 2014, 'Conditionally rare taxa disproportionately contribute to temporal changes in microbial diversity', mBio, vol. 5, no. 4, e01371-14.

Shannon, CE & Weaver, W 1949, The mathematical theory of communication, University of Illinois Press, Champaign, IL.

Shapiro, SS & Wilk, MB 1965, 'An analysis of variance test for normality (complete samples)', Biometrika, vol. 52, no. 3-4, pp. 591-611.

Sharon, G & Rosenberg, E 2008, 'Bacterial growth on coral mucus', Current Microbiology, vol. 56, no. 5, pp. 481- 488.

Sharp, KH, Distel, D & Paul, VJ 2012, 'Diversity and dynamics of bacterial communities in early life stages of the Caribbean coral Porites astreoides', International Society for Microbial Ecology Journal, vol. 6, no. 4, pp. 790- 801.

Sharp, KH, Pratte, ZA, Kerwin, AH, Rotjan, RD & Stewart, FJ 2017, 'Season, but not symbiont state, drives microbiome structure in the temperate coral Astrangia poculata', Microbiome, vol. 5, 120.

Sharp, KH & Ritchie, KB 2012, 'Multi-partner interactions in corals in the face of climate change', Biological Bulletin, vol. 223, no. 1, pp. 66-77.

197

Sharp, KH, Ritchie, KB, Schupp, PJ, Ritson-Williams, R & Paul, VJ 2010, 'Bacterial acquisition in juveniles of several broadcast spawning coral species', PLoS ONE, vol. 5, no. 5, p. e10898.

Shi, M, Zou, M, Liu, X, Gao, Y, Zhang, Z, Wu, W, Wen, D, Chen, Z & An, C 2006, 'A novel bacterium Saprospira sp. strain PdY3 forms bundles and lyses cyanobacteria', Frontiers in Bioscience, vol. 11, pp. 1916-1923.

Shnit-Orland, M & Kushmaro, A 2009, 'Coral mucus-associated bacteria: a possible first line of defense', FEMS Microbiology Ecology, vol. 67, no. 3, pp. 371-380.

Shnit-Orland, M, Sivan, A & Kushmaro, A 2012, 'Antibacterial activity of Pseudoalteromonas in the coral holobiont', Microbial Ecology, vol. 64, no. 4, pp. 851-859.

Silverstein, RN, Correa, AMS & Baker, AC 2012, 'Specificity is rarely absolute in coral–algal symbiosis: implications for coral response to climate change', Proceedings of the Royal Society B: Biological Sciences, vol. 279, no. 1738, pp. 2609-2618.

Simon, M, Scheuner, C, Meier-Kolthoff, JP, Brinkhoff, T, Wagner-Döbler, I, Ulbrich, M, Klenk, H-P, Schomburg, D, Petersen, J & Göker, M 2017, 'Phylogenomics of Rhodobacteraceae reveals evolutionary adaptation to marine and non-marine habitats', International Society for Microbial Ecology Journal, vol. 11, pp. 1483-1499.

Simpson, EH 1949, 'Measurement of diversity', Nature, vol. 163, p. 688.

Smith, D, Leary, P, Craggs, J, Bythell, J & Sweet, M 2015, 'Microbial communities associated with healthy and white syndrome-affected Echinopora lamellosa in aquaria and experimental treatment with the antibiotic ampicillin', PLoS ONE, vol. 10, no. 3, e0121780.

Smith, VH 2002, 'Effects of resource supplies on the structure and function of microbial communities', Antonie Van Leeuwenhoek, vol. 81, pp. 99-106.

Soffer, N, Gibbs, PDL & Baker, AC 2008, 'Practical applications of contaminant-free Symbiodinium cultures grown on solid media,' Confrence proceedings of the International Coral Reef Symposium 2008, ICRS, Ft. Lauderdale, FL, pp. 159-163.

Sonoyama, K, Fujiwara, R, Takemura, N, Ogasawara, T, Watanabe, J, Ito, H & Morita, T 2009, 'Response of gut microbiota to fasting and hibernation in Syrian hamsters', Applied and Environmental Microbiology, vol. 75, no. 20, pp. 6451-6456.

Sorgeloos, P, Bossuyt, E, Laviña, E, Baeza-Mesa, M & Persoone, G 1977, 'Decapsulation of Artemia cysts: a simple technique for the improvement of the use of brine shrimp in aquaculture', Aquaculture, vol. 12, no. 4, pp. 311-315.

Stämmler, F, Gläsner, J, Hiergeist, A, Holler, E, Weber, D, Oefner, PJ, Gessner, A & Spang, R 2016, 'Adjusting microbiome profiles for differences in microbial load by spike-in bacteria', Microbiome, vol. 4, 28.

Stat, M, Carter, D & Hoegh-Guldberg, O 2006, 'The evolutionary history of Symbiodinium and scleractinian hosts—symbiosis, diversity, and the effect of climate change', Perspectives in Plant Ecology, Evolution and Systematics, vol. 8, no. 1, pp. 23-43.

Stat, M, Pochon, X, Cowie, ROM & Gates, RD 2009, 'Specificity in communities of Symbiodinium in corals from Johnston Atoll', Marine Ecology Progress Series, vol. 386, pp. 83-96.

Steen, RG & Muscatine, L 1987, 'Low temperature evokes rapid exocytosis of symbiotic algae by a sea anemone', Biological Bulletin, vol. 172, no. 2, pp. 246-263.

Stewart, EJ 2012, 'Growing unculturable bacteria', Journal of Bacteriology, vol. 194, no. 16, pp. 4151-4160.

198

Stewart, PS 2002, 'Mechanisms of antibiotic resistance in bacterial biofilms', International Journal of Medical Microbiology, vol. 292, no. 2, pp. 107-113.

Student 1908, 'The probable error of a mean', Biometrika, vol. 6, no. 1, pp. 1-25.

Sun, W, Anbuchezhian, R & Li, Z 2016, 'Association of coral-microbes, and the ecological roles of microbial symbionts in corals', in S Goffredo and Z Dubinsky (eds), The Cnidaria, past, present and future: the world of Medusa and her sisters, Springer, Cham, pp. 347-357.

Sunagawa, S, DeSantis, TZ, Piceno, YM, Brodie, EL, DeSalvo, MK, Voolstra, CR, Weil, E, Andersen, GL & Medina, M 2009a, 'Bacterial diversity and White Plague Disease-associated community changes in the Caribbean coral Montastraea faveolata', International Society for Microbial Ecology Journal, vol. 3, pp. 512-521.

Sunagawa, S, Wilson, EC, Thaler, M, Smith, ML, Caruso, C, Pringle, JR, Weis, VM, Medina, M & Schwarz, JA 2009b, 'Generation and analysis of transcriptomic resources for a model system on the rise: the sea anemone Aiptasia pallida and its dinoflagellate endosymbiont', BMC Genomics, vol. 10, 258.

Sunagawa, S, Woodley, CM & Medina, M 2010, 'Threatened corals provide underexplored microbial habitats', PLoS ONE, vol. 5, no. 3, e9554.

Sweet, MJ, Brown, BE, Dunne, RP, Singleton, I & Bulling, M 2017, 'Evidence for rapid, tide-related shifts in the microbiome of the coral Coelastrea aspera', Coral Reefs, vol. 36, no. 3, pp. 815-828.

Sweet, MJ, Croquer, A & Bythell, JC 2011, 'Bacterial assemblages differ between compartments within the coral holobiont', Coral Reefs, vol. 30, pp. 39-52.

Sweet, MJ, Croquer, A & Bythell, JC 2014, 'Experimental antibiotic treatment identifies potential pathogens of white band disease in the endangered Caribbean coral Acropora cervicornis', Proceedings of the Royal Society B: Biological Sciences, vol. 281, no. 1788, 20140094.

Takao, T, Kitatani, F, Watanabe, N, Yagi, A & Sakata, K 1994, 'A simple screening method for antioxidants and isolation of several antioxidants produced by marine bacteria from fish and shellfish', Bioscience, Biotechnology, and Biochemistry, vol. 58, no. 10, pp. 1780-1783.

Taoka, Y, Maeda, H & Jo, JY 2006, 'Growth, stress tolerance and non-specific immune response of Japanese flounder (Paralichthys olivaceus) to probiotics in a closed recirculating system', Fisheries Science, vol. 72, pp. 310-321.

Tchernov, D, Gorbunov, MY, de Vargas, C, Narayan Yadav, S, Milligan, AJ, Häggblom, M & Falkowski, PG 2004, 'Membrane lipids of symbiotic algae are diagnostic of sensitivity to thermal bleaching in corals', Proceedings of the National Academy of Sciences, USA, vol. 101, no. 37, pp. 13531-13535.

Teplitski, M & Ritchie, K 2009, 'How feasible is the biological control of coral diseases?', Trends in Ecology & Evolution, vol. 24, no. 7, pp. 378-385.

Thomas, S, Burdett, H, Temperton, B, Wick, R, Snelling, D, McGrath, J, Quinn, J, Munn, C & Gilbert, J 2010, 'Evidence for phosphonate usage in the coral holobiont', International Society for Microbial Ecology Journal, vol. 4, no. 3, pp. 459-461.

Thompson, JR, Rivera, HE, Closek, CJ & Medina, M 2015, 'Microbes in the coral holobiont: partners through evolution, development, and ecological interactions', Frontiers in Cellular and Infection Microbiology, vol. 4, 176.

Thornhill, DJ, Xiang, Y, Fitt, WK & Santos, SR 2009, 'Reef endemism, host specificity and temporal stability in populations of symbiotic dinoflagellates from two ecologically dominant Caribbean corals', PLoS ONE, vol. 4, no. 7, e6262. 199

Thornhill, DJ, Xiang, Y, Pettay, DT, Zhong, M & Santos, SR 2013, 'Population genetic data of a model symbiotic cnidarian system reveal remarkable symbiotic specificity and vectored introductions across ocean basins', Molecular Ecology, vol. 22, no. 17, pp. 4499-4515.

Tolleter, D, Seneca, FO, DeNofrio, JC, Krediet, CJ, Palumbi, SR, Pringle, JR & Grossman, AR 2013, 'Coral bleaching independent of photosynthetic activity', Current Biology, vol. 23, no. 18, pp. 1782-1786.

Tortorelli, G, Belderok, R, Davy, S, McFadden, G & van Oppen, M 2020, 'Host genotypic effect on algal symbiosis establishment in the coral model, the anemone Exaiptasia diaphana, from the Great Barrier Reef', Frontiers in Marine Science, vol. 6, 833.

Tourlousse, DM, Yoshiike, S, Ohashi, A, Matsukura, S, Noda, N & Sekiguchi, Y 2017, 'Synthetic spike-in standards for high-throughput 16S rRNA gene amplicon sequencing', Nucleic Acids Research, vol. 45, no. 4, e23.

Tout, J, Siboni, N, Messer, LF, Garren, M, Stocker, R, Webster, NS, Ralph, PJ & Seymour, JR 2015, 'Increased seawater temperature increases the abundance and alters the structure of natural Vibrio populations associated with the coral Pocillopora damicornis', Frontiers in Microbiology, vol. 6, 432.

Tracy, AM, Koren, O, Douglas, N, Weil, E & Harvell, CD 2015, 'Persistent shifts in Caribbean coral microbiota are linked to the 2010 warm thermal anomaly', Environmental Microbiology Reports, vol. 7, no. 3, pp. 471-479.

Tremblay, P, Grover, R, Maguer, J, Legendre, L & Ferrier-Pages, C 2012a, 'Autotrophic carbon budget in coral tissue: a new C-13-based model of photosynthate translocation', Journal of Experimental Biology, vol. 215, no. 8, pp. 1384-1393.

Tremblay, P, Naumann, MS, Sikorski, S, Grover, R & Ferrier-Pagès, C 2012b, 'Experimental assessment of organic carbon fluxes in the scleractinian coral Stylophora pistillata during a thermal and photo stress event', Marine Ecology Progress Series, vol. 453, pp. 63-77.

Tukey, J 1949, 'Comparing individual means in the analysis of variance', Biometrics, vol. 5, no. 2, pp. 99-144.

Valverde, EJ, Labella, AM, Borrego, JJ & Castro, D 2019, 'Artemia spp., a susceptible host and vector for Lymphocystis disease virus', Viruses, vol. 11, no. 6, 506. van Alstyne, KL, Dominique, VJ & Muller-Parker, G 2009, 'Is dimethylsulfoniopropionate (DMSP) produced by the symbionts or the host in an anemone–zooxanthella symbiosis?', Coral Reefs, vol. 28, no. 1, pp. 167-176. van de Water, JAJM, Melkonian, R, Junca, H, Voolstra, CR, Reynaud, S, Allemand, D & Ferrier-Pagès, C 2016, 'Spirochaetes dominate the microbial community associated with the red coral Corallium rubrum on a broad geographic scale', Scientific Reports, vol. 6, 27277. van Iten, H, Leme, JM, Pacheco, MLAF, Simões, MG, Fairchild, TR, Rodrigues, F, Galante, D, Boggiani, PC & Marques, AC 2016, 'Origin and early diversification of phylum Cnidaria: key macrofossils from the Ediacaran system of North and South America', in S Goffredo and Z Dubinsky (eds), The Cnidaria, past, present and future: the world of Medusa and her sisters, Springer, Cham, pp. 31-40. van Oppen, MJH & Blackall, LL 2019, 'Coral microbiome dynamics, functions and design in a changing world', Nature Reviews: Microbiology, vol. 17, pp. 557-567. van Oppen, MJH, Oliver, JK, Putnam, HM & Gates, RD 2015, 'Building coral reef resilience through assisted evolution', Proceedings of the National Academy of Sciences, USA, vol. 112, no. 8, pp. 2307-2313. van Zanten, BT, van Beukering, PJH & Wagtendonk, AJ 2014, 'Coastal protection by coral reefs: a framework for spatial assessment and economic valuation', Ocean & Coastal Management, vol. 96, pp. 94-103.

200

Vandeputte, D, Kathagen, G, D’hoe, K, Vieira-Silva, S, Valles-Colomer, M, Sabino, J, Wang, J, Tito, RY, De Commer, L, Darzi, Y, Vermeire, S, Falony, G & Raes, J 2017, 'Quantitative microbiome profiling links gut community variation to microbial load', Nature, vol. 551, pp. 507-511.

Vega Thurber, R, Willner-Hall, D, Rodriguez-Mueller, B, Desnues, C, Edwards, RA, Angly, F, Dinsdale, E, Kelly, L & Rohwer, F 2009, 'Metagenomic analysis of stressed coral holobionts', Environmental Microbiology, vol. 11, no. 8, pp. 2148-2163.

Velho-Pereira, S, Parvatkar, P & Furtado, IJ 2015, 'Evaluation of antioxidant producing potential of halophilic bacterial bionts from marine invertebrates', Indian Journal of Pharmaceutical Sciences, vol. 77, no. 2, pp. 183- 189.

Veron, JEN, Hoegh-Guldberg, O, Lenton, TM, Lough, JM, Obura, DO, Pearce-Kelly, P, Sheppard, CRC, Spalding, M, Stafford-Smith, MG & Rogers, AD 2009, 'The coral reef crisis: the critical importance of <350 ppm CO2', Marine Pollution Bulletin, vol. 58, no. 10, pp. 1428-1436.

Vezzulli, L, Previati, M, Pruzzo, C, Marchese, A, Bourne, DG & Cerrano, C 2010, 'Vibrio infections triggering mass mortality events in a warming Mediterranean Sea', Environmental Microbiology, vol. 12, no. 7, pp. 2007-2019.

Viant, MR 2007, 'Metabolomics of aquatic organisms: the new 'omics' on the block', Marine Ecology Progress Series, vol. 332, pp. 301-306.

Volkmer, B & Heinemann, M 2011, 'Condition-dependent cell volume and concentration of Escherichia coli to facilitate data conversion for systems biology modeling', PLoS ONE, vol. 6, no. 7

Voolstra, CR 2013, 'A journey into the wild of the cnidarian model system Aiptasia and its symbionts', Molecular Ecology, vol. 22, no. 17, pp. 4366-4368.

Vorholt, JA, Vogel, C, Carlström, CI & Müller, DB 2017, 'Establishing causality: opportunities of synthetic communities for plant microbiome research', Cell Host & Microbe, vol. 22, no. 2, pp. 142-155.

Wang, JT & Douglas, AE 1999, 'Essential amino acid synthesis and nitrogen recycling in an alga–invertebrate symbiosis', Marine Biology, vol. 135, no. 2, pp. 219-222.

Wang, Q, Garrith, GM, Tiedje, JM & Cole, JR 2007, 'Naive bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy', Applied and Environmental Microbiology, vol. 73, no. 16, pp. 5261-5267.

Wang, X, Zhu, J, Feng, J, Sardar, A, Hu, Z & Wang, H 2019, 'Tropicibacter alexandrii sp. nov., a novel marine bacterium isolated from the phycosphere of a dinoflagellate, Alexandrium minutum' [in press] Antonie van Leeuwenhoek, viewed 1 Dec 2019, .

Wang, Y, Naumann, U, Wright, ST & Warton, DI 2012, 'mvabund – an R package for model-based analysis of multivariate abundance', Methods in Ecology and Evolution, vol. 3, pp. 471-474.

Wang, Y & Qian, P-Y 2013, 'Conserved regions in 16S ribosome RNA sequences and primer design for studies of environmental microbes', in KE Nelson (ed.) Encyclopedia of Metagenomics, Springer, New York.

Warner, ME, Fitt, WK & Schmidt, GW 1999, 'Damage to photosystem II in symbiotic dinoflagellates: a determinant of coral bleaching', Proceedings of the National Academy of Sciences, USA, vol. 96, no. 14, pp. 8007-8012.

Warnes, GR, Bolker, B, Bonebakker, L, Gentleman, R, Huber, W, Liaw, A, Lumley, T, Maechler, M, Magnusson, A, Moeller, S, Schwartz, M & Venables, B 2016, gplots: various R programming tools for plotting data [software] R package version 3.0.1.1, viewed 1 Oct 2019, .

201

Wegley, L, Edwards, R, Rodriguez-Brito, B, Liu, H & Rohwer, F 2007, 'Metagenomic analysis of the microbial community associated with the coral Porites astreoides', Environmental Microbiology, vol. 9, no. 11, pp. 2707- 2719.

Weiland-Bräuer, N, Neulinger, SC, Pinnow, N, Künzel, S, Baines, JF & Schmitz, RA 2015, 'Composition of bacterial communities associated with Aurelia aurita changes with compartment, life stage, and population', Applied and Environmental Microbiology, vol. 81, no. 17, pp. 6038-6052.

Weiler, BA, Verhoeven, JTP & Dufour, SC 2018, 'Bacterial communities in tissues and surficial mucus of the cold- water coral Paragorgia arborea', Frontiers in Marine Science, vol. 5, 378.

Weis, VM 2008, 'Cellular mechanisms of Cnidarian bleaching: stress causes the collapse of symbiosis', Journal of Experimental Biology, vol. 211, no. 19, pp. 3059-3066.

Weis, VM, Davy, SK, Hoegh-Guldberg, O, Rodriguez-Lanetty, M & Pringle, JR 2008, 'Cell biology in model systems as the key to understanding corals', Trends in Ecology & Evolution, vol. 23, no. 7, pp. 369-376.

Welch, BL 1947, 'The generalization of ‘Student's’ problem when several different population variances are involved', Biometrika, vol. 34, no. 1-2, pp. 28-35.

Welsh, RM, Zaneveld, JR, Rosales, SM, Payet, JP, Burkepile, DE & Thurber, RV 2015, 'Bacterial predation in a marine host-associated microbiome', International Society for Microbial Ecology Journal, vol. 10, no. 6, pp. 1540-1544.

Whipps, JM, Lewis, K & Cooke, RC 1988, 'Mycoparasitism and plant disease control', in NM Burge (ed.) Fungi in biological control systems, Manchester University Press, Manchester, pp. 161-187.

Whitman, WB, Coleman, DC & Wiebe, WJ 1998, 'Prokaryotes: the unseen majority', Proceedings of the National Academy of Sciences, USA, vol. 95, no. 12, pp. 6578-6583.

Whitney, J 1997, 'Testing for differences with the nonparametric Mann-Whitney U test', Journal of Wound, Ostomy, and Continence Nursing, vol. 24, no. 1, p. 12.

Wickham, H 2019, ggplot2: elegant graphics for data analysis [software] R package version 3.2.1, viewed 1 Feb 2019, .

Wilcoxon, F 1945, 'Individual comparisons by ranking methods', Biometrics Bulletin, vol. 1, no. 6, pp. 80-83.

Wild, C, Hoegh-Guldberg, O, Naumann, MS, Colombo-Pallotta, MF, Ateweberhan, M, Fitt, WK, Iglesias-Prieto, R, Palmer, C, Bythell, JC, Ortiz, J-C, Loya, Y & van Woesik, R 2011, 'Climate change impedes scleractinian corals as primary reef ecosystem engineers', Marine and Freshwater Research, vol. 62, no. 2, pp. 205-215.

Wilkerson, FP, Muller, G & Muscatine, PL 1983, 'Temporal patterns of cell division in natural populations of endosymbiotic algae', Limnology and Oceanography, vol. 28, no. 5, pp. 1009-1014.

Wilkins, LGE, Leray, M, Yuen, B, Peixoto, R, Pereira, TJ, Bik, HM, Coil, DA, Duffy, JE, Herre, EA, Lessios, H, Lucey, N, Mejia, LC, O'Dea, A, Rasher, DB, Sharp, K, Sogin, EM, Thacker, RW, Vega Thurber, R, Wcislo, WT, Wilbanks, EG & Eisen, JA 2019, 'Host-associated microbiomes and their roles in marine ecosystem functions' [preprint] viewed 21 Oct 2019, .

Wilson, K, Li, Y, Whan, V, Lehnert, S, Byrne, K, Moore, S, Pongsomboon, S, Tassanakajon, A, Rosenberg, G, Ballment, E, Fayazi, Z, Swan, J, Kenway, M & Benzie, J 2002, 'Genetic mapping of the black tiger shrimp Penaeus monodon with amplified fragment length polymorphism', Aquaculture, vol. 204, no. 3, pp. 297-309.

202

Witte, AK, Mester, P, Fister, S, Witte, M, Schoder, D & Rossmanith, P 2016, 'A systematic investigation of parameters influencing droplet rain in the Listeria monocytogenes prfA assay - reduction of ambiguous results in ddPCR', PLoS ONE, vol. 11, no. 12, e0168179.

Wolfowicz, I, Baumgarten, S, Voss, PA, Hambleton, EA, Voolstra, CR, Hatta, M & Guse, A 2016, 'Aiptasia sp. larvae as a model to reveal mechanisms of symbiont selection in cnidarians', Scientific Reports, vol. 6, 32366.

Wood, SA, Biessy, L, Latchford, JL, Zaiko, A, von Ammon, U, Audrezet, F, Cristescu, ME & Pochon, X 2020, 'Release and degradation of environmental DNA and RNA in a marine system', Science of The Total Environment, vol. 704, 135314.

Xiang, T, Hambleton, EA, DeNofrio, JC, Pringle, JR & Grossman, AR 2013, 'Isolation of clonal axenic strains of the symbiotic dinoflagellate Symbiodinium and their growth and host specificity', Journal of Phycology, vol. 49, no. 3, pp. 447-458.

Young, G, Turner, S, Davies, JK, Sundqvist, G & Figdor, D 2007, 'Bacterial DNA persists for extended periods after cell death', Journal of Endodontics, vol. 33, no. 12, pp. 1417-1420.

Zaneveld, JR, Burkepile, DE, Shantz, AA, Pritchard, CE, McMinds, R, Payet, JP, Welsh, R, Correa, AMS, Lemoine, NP, Rosales, S, Fuchs, C, Maynard, JA & Thurber, RV 2016, 'Overfishing and nutrient pollution interact with temperature to disrupt coral reefs down to microbial scales', Nature Communications, vol. 7, 11833.

Zaneveld, JR, McMinds, R & Vega Thurber, R 2017, 'Stress and stability: applying the Anna Karenina principle to animal microbiomes', Nature Microbiology, vol. 2, 17121.

Zaragoza, WJ, Krediet, CJ, Meyer, JL, Canas, G, Ritchie, KB & Teplitski, M 2014, 'Outcomes of infections of sea anemone Aiptasia pallida with Vibrio spp. pathogenic to corals', Microbial Ecology, vol. 68, no. 2, pp. 388-396.

Zettler, ER, Mincer, TJ & Amaral-Zettler, LA 2013, 'Life in the “Plastisphere”: microbial communities on plastic marine debris', Environmental Science & Technology, vol. 47, no. 13, pp. 7137-7146.

Zhang, Y, Ling, J, Yang, Q, Wen, C, Yan, Q, Sun, H, Van Nostrand, JD, Shi, Z, Zhou, J & Dong, J 2015, 'The functional gene composition and metabolic potential of coral-associated microbial communities', Scientific Reports, vol. 5, 16191.

Zhukova, M, Sapountzis, P, Schiøtt, M & Boomsma, JJ 2017, 'Diversity and transmission of gut bacteria in Atta and Acromyrmex leaf-cutting ants during development', Frontiers in Microbiology, vol. 8, 1942.

Ziegler, M, Seneca, FO, Yum, LK, Palumbi, SR & Voolstra, CR 2017, 'Bacterial community dynamics are linked to patterns of coral heat tolerance', Nature Communications, vol. 8, 14213.

203