UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of morphology and neighbouring seaweeds on macroalgal microbiota Chen, Melissa 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_september_chen_melissa.pdf [ 2.11MB ]
Metadata
JSON: 24-1.0351980.json
JSON-LD: 24-1.0351980-ld.json
RDF/XML (Pretty): 24-1.0351980-rdf.xml
RDF/JSON: 24-1.0351980-rdf.json
Turtle: 24-1.0351980-turtle.txt
N-Triples: 24-1.0351980-rdf-ntriples.txt
Original Record: 24-1.0351980-source.json
Full Text
24-1.0351980-fulltext.txt
Citation
24-1.0351980.ris

Full Text

THE EFFECTS OF MORPHOLOGY AND NEIGHBOURING SEAWEEDSON MACROALGAL MICROBIOTAbyMelissa ChenB.Sc., Cell and Developmental Biology, University of British Columbia, Canada, 2015A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES(Botany)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)August 2017c© Melissa Chen, 2017AbstractAbstractMacroalgae (seaweeds) have an intimate relationship with their microbial symbionts.Microbial communities associated with macroalgal surfaces (epibiota) are generallyhost-specific and, historically, there has been great interest in the role of biologicalcompounds and chemical warfare in microbial community assembly on seaweeds. However,the interaction between seaweeds and their environment may also influence communityassembly of their microbiota. In my thesis, I conduct two experiments that ask how factorsnot related to seaweed chemistry influence microbial community assembly. First, I askwhether the interaction between flow and seaweed morphology affects the settlement andstructure of microbial biofilms. In this project, I test whether three common algalmorphologies select for differential biofilm communities using artificial macroalgae units(AM units) made out of latex. I find that morphology does affect initial microbialsettlement and community structure, but that eventual dominance of substrate specialists(in our case a latex degrader) swamps the influence of morphology in long-term biofilms.The second chapter of my thesis asks whether macroalgae affect the microbial epibiota ofeach other. To test this, I co-incubate Nereocystis leutkeana meristem fragments withdifferent species of mature macroalgae. I find that although water column communitieschange significantly when incubated with mature macroalgae, seaweed surface communitiesare far more resistant to change. Overall, these results support the idea that the seaweedsurfaces are highly selective, and demonstrate that modulations on seaweed microbiotaoperate within an overarching paradigm of species specificity. With these experiments, Ihope to contribute to the larger body of knowledge on seaweed-microbe associations andimprove understanding of how, and why, we find the observed microbiota on seaweedsurfaces.iiLay SummaryLay SummarySeaweed surfaces are home to a diverse collection of microscopic organisms (microbes) thatprovide many benefits for their seaweed host. Seaweed surfaces are selective over whichmicrobes can settle and grow, and my thesis investigates factors that explain why certainmicrobes are found on seaweeds, while others are not. First, I ask whether seaweed shapeinfluences microbial biofilm development. I find that branched seaweed shapes developbiofilms faster than un-branched shapes, and I discuss how this relates to biofilms found onwild seaweeds. Second, I test whether seaweeds grown together influence each other’sbiofilms. I find that although seaweeds transmit microbes into the water, there arerelatively few microbes that are transmitted from seaweed to seaweed. In general, myfindings demonstrate that seaweed surfaces are highly selective. My thesis contributesknowledge about how microbes grow and settle on seaweeds, and will lead to a betterunderstanding of microbe-seaweed relationship dynamics.iiiPrefacePreface• Chapter 1: A general introduction to seaweed-associated microbes.• Chapter 2: Matt Lemay, Laura Parfrey and Patrick Martone had the idea for theproject. Project was based off results from work done by Matt Lemay, Laura Parfrey,and Patrick Martone. Melissa Chen conducted the experiments and performed thesequence analysis and statistical analysis with input from Laura W. Parfrey and PatrickMartone. Melissa Chen wrote a first draft of the manuscript. Laura Parfrey supervisedthe project and committee members Mary O’Connor and Patrick Martone providedfeedback on analysis, results, and discussion.This chapter is not yet published.• Chapter 3: Melissa Chen conceived of the research question. Laura Parfrey supervisedthe project and committee members, Mary O’Connor and Patrick Martone providedfeedback. All experimentation, statistical analysis, and writing of first drafts was doneby Melissa Chen.This chapter is not yet published.Throughout this dissertation the word “we” refers to Melissa Chen and Laura Parfrey unlessotherwise stated.ivTable of ContentsTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii1 Seaweeds and their microbial symbionts . . . . . . . . . . . . . . . . . . . . 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Problem statement and objectives of thesis . . . . . . . . . . . . . . . . . . . 22 The effect of seaweed morphology on microbial community settlementand composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202.5 Supplementary Figures and Tables . . . . . . . . . . . . . . . . . . . . . . . 243 Horizontal transmission of microbes between neighbouring macroalgae . 343.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493.5 Supplementary Figures and Tables . . . . . . . . . . . . . . . . . . . . . . . 52Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59vList of TablesList of TablesTable 2.1 Number of replicates at each site for AM experiment. . . . . . . . . . 7Table 2.2 Pairwise t-tests and ANOVA of richness between morphologies acrosstime. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Table 2.3 Pairwise and overall PERMANOVAs of community composition be-tween morphologies across time. . . . . . . . . . . . . . . . . . . . . . . . . 16Table 2.4 S: PERMDISP of morphologies at each time point and overall. . . . 32Table 2.5 S: ANOVA of richness across morphology by time point. . . . . . . . 32Table 2.6 S: Welch’s t-Test comparison of richness at Reed Point and Hakai bytime point. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33Table 3.1 Comparison of community dissimilarity (PERMANOVA) and richness(Welch’s t-Test) of Nereocystis, Mastocarpus, and water samples. . . . . . . 43Table 3.2 Comparison of community dissimilarity (PERMANOVA) and richness(Welch’s t-Test) of water column and NMF surface communities in M–W–NMF treatments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46Table 3.3 S: PERMANOVA results comparing Nereocystis, Mastocarpus, andwater samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Table 3.4 S: PERMANOVA results comparing water samples from M–W–NMFexperiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Table 3.5 S: PERMANOVA results comparing NMF surface samples from M–W–NMF experiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Table 3.6 S: PERMANOVA results comparing water samples from the M–Wexperiment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57Table 3.7 S: ANOVA results comparing richness of Nereocystis, Mastocarpus, andwater. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58Table 3.8 S: ANOVA and Welch’s t-Test results comparing richness of watersamples from M–W–NMF experiment. . . . . . . . . . . . . . . . . . . . . 58Table 3.9 S: ANOVA and Welch’s t-Test results comparing NMF surface samplesfrom M–W–NMF experiment. . . . . . . . . . . . . . . . . . . . . . . . . . 58viList of FiguresList of FiguresFigure 2.1 Artificial Macroalgae. . . . . . . . . . . . . . . . . . . . . . . . . . . 6Figure 2.2 Richness of biofilms on artificial macroalgae through time. . . . . . . 13Figure 2.3 Dispersion and community composition of AM morphologies throughtime. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15Figure 2.4 Heatmap of core OTUs across morphologies and time points. . . . . 17Figure 2.5 Turnover of OTUs across time points. . . . . . . . . . . . . . . . . . 18Figure 2.6 Taxa summary of OTUs collapsed by Order. . . . . . . . . . . . . . 19Figure 2.7 S: Richness of biofilms on AM units through time (all metrics). . . . 24Figure 2.8 S: Dispersion and community composition of AMmorphologies throughtime (un-weighted Unifrac). . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Figure 2.9 S: Dispersion and community composition of AMmorphologies throughtime (weighted Unifrac). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26Figure 2.10 S: Overall dispersion of AM units over time (all metrics). . . . . . . 27Figure 2.11 S: Heatmap showing “core” OTUs. . . . . . . . . . . . . . . . . . . . 28Figure 2.12 S: Heatmap showing presence or absence of “core” OTUs. . . . . . . 28Figure 2.13 S: Comparison of water diversity and AM biofilm diversity . . . . . . 29Figure 2.14 S: Comparison of AM biofilm diversity at Reed Point and Hakai. . . 30Figure 2.15 S: Taxa summaries plot at the OTU level. . . . . . . . . . . . . . . . 31Figure 2.16 S: Dye Dip Experiment Results . . . . . . . . . . . . . . . . . . . . . 31Figure 3.1 Experimental design for M–W and M–W–NMF experiments. . . . . 37Figure 3.2 Comparison of community composition and richness across macroalgalsurfaces and water samples. . . . . . . . . . . . . . . . . . . . . . . . . . . . 42Figure 3.3 Comparison of water column communities in M–W–NMF treatments. 44Figure 3.4 Comparison of NMF surface communities M–W–NMF treatments. . 45Figure 3.5 Enrichment and reduction of NMF surface and water column commu-nities in M–W–NMF treatments. . . . . . . . . . . . . . . . . . . . . . . . . 47Figure 3.6 Taxa summary plots showing enriched or reduced genera of NMFsurface and water column communities. . . . . . . . . . . . . . . . . . . . . 48Figure 3.7 S: NMDS of Nereocystis, Mastocarpus, and water samples. . . . . . 52Figure 3.8 S: Richness of Nereocystis, Mastocarpus, and water samples. . . . . . 52Figure 3.9 S: NMDS of water column communities in M–W–NMF treatments. . 53Figure 3.10 S: Richness of water column communities in M–W–NMF treatments. 53Figure 3.11 S: Composition and Richness of samples in the M–W experiment. . . 54Figure 3.12 NMDS of NMF surface communities from M–W–NMF treatments. . 55Figure 3.13 S: Richness across NMF surface communities from M–W–NMF treat-ments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Figure 3.14 S: Taxa summary of wild and lab-incubated seaweed samples . . . . 56viiAcknowledgementsAcknowledgementsA thousand ‘thank you’s and gratitudes to Laura Parfrey for guiding me through my Masters.I know I am not always on time, but I am thankful you let me delve into all the rabbit holesthat I do.I would also like to thank my committee members, Mary O’Connor and Patrick Martone,for providing input and feedback on my projects through all their various stages. Yourenthusiasm and interest are greatly appreciated.Also thank you to all the undergrads, graduate students, and post-docs that have helped mealong the way: Matt Lemay, Stilian Loucas, Andy Loudon, Jordan Lin, Cody Foley, Cassan-dra Jensen, Maryam Osmond, Gwen Griffiths, Rhea Smith, Liam Coleman, and probablymany more. Additionally, thank you to the various technicians and statisticians who havehelped me: Evan Morien, Marcus Campbell, Courtney Van den Elzen.Finally, thank you to my partner, David, for providing continued support and encouragementduring all the times that madness felt imminent.viiiSeaweeds and their microbial symbiontsChapter 1Seaweeds and their microbial symbionts‘In a system where life is the universal good, but the destruction of life thewell nigh universal occupation, an order has spontaneously risen whichconstantly tends to maintain life at the highest limit [...] Is there not, inthis reflection, solid ground for a belief in the final beneficence of the lawsof organic nature?’Stephen A. Forbes, 18871.1 IntroductionMacroalgae (seaweeds) have an intimate relationship with their microbial symbionts. Theirsurfaces are rich with biological compounds that attract or deter a variety of microbes, manyof which provide crucial functions for their seaweed host. Microbes promote macroalgalsettlement [Joint et al., 2002,Weinberger, 2007] and are often required for normal growthand morphological development [Provasoli and Pintner, 1980,Nakanishi et al., 1996,Matsuoet al., 2005,Marshall et al., 2006]. They are known to aid in nutrient acquisition by fixingnitrogen [Rosenberg and Paerl, 1981,de Oliveira et al., 2012] or producing vitamin B12 [Croftet al., 2005] for their host, and play a role in priming immune defences against pathogenicbacteria [Küpper et al., 2002,Weinberger, 2007]. Indeed, seaweeds rely on their microbialsymbionts for a wide range of functions.Historically, there has been great interest in the role of biological compounds and chemi-cal warfare in community assembly on seaweeds. Seaweeds produce a variety of polymersincluding, but not limit to, agar, carrageenan, alginate, mannose, cellulose, and pectin; allof which have associated degraders previously isolated from seaweeds [Goecke et al., 2010].The microbes that consume these biological molecules are sometimes detrimental to theirseaweed host because they cause degradation of tissue under saprotrophic conditions. How-ever, under non-saprotrophic conditions, these microbes also provide a variety of services1Seaweeds and their microbial symbiontsthat benefit the host alga.In exchange for organic compounds, some bacteria help mineralize organic substrate andsupply algae with supplemental CO2 and minerals [Brock and Clyne, 1984, Coveney andWetzel, 1989,Croft et al., 2005,Dromgoole and J, 1978,Rosenberg and Paerl, 1981,Carpenterand Cox, 1974]. Microbes also remove heavy metals and crude oil, which are harmful formacroalgae, from the water [Riquelme et al., 1997,Semenova et al., 2009,Yurkov and Beatty,1998]. Finally, some microbes promote growth by supplying growth factors [Dimitrievaet al., 2006,Berland et al., 1972,Bolinches et al., 1988,Meusnier et al., 2001]. Thus, thereis a fine balance between the beneficial and saprotrophic effects of symbionts that consumethe polymers found on the surface of seaweeds.Seaweeds also exude metabolites and compounds that are known to have antibiotic effects.Some seaweeds store anti-fouling compounds within their tissues [Armstrong et al., 2001],and these compounds can be released into the surrounding water column. In other cases,seaweeds like kelps produce oxidative bursts that selectively reduce the number of pathogenicbacteria on their surface [Weinberger et al., 1999, Dring, 2006, Küpper et al., 2002]. Theexact mechanism by which these antifouling compounds work is highly varied, but has beendescribed in several different seaweeds [Bhakuni and Rawat, 2006,Dubber and Harder, 2008].Therefore, in addition to attracting potentially beneficial symbionts, seaweeds also activelydeter pathogenic ones.Microbial communities associated with macroalgal surfaces (epibionts) are generally host-specific [Egan et al., 2013,Lachnit et al., 2009,de Oliveira et al., 2012]. However, there is highvariation in taxonomic community composition across individuals of the same species [Burkeet al., 2011b] that may correlate to environmental conditions like seasonality [Lachnit et al.,2011] or salinity [Dittami et al., 2015]. Additionally, it has been proposed that functionaltraits in epibiotic communities on seaweeds are conserved, whereas taxonomic compositionis not [Burke et al., 2011a]. Thus, the composition of microbial communities on seaweedsurfaces is modulated by many factors in the environment, and these factors interact withthe seaweed’s biology to determine the final microbial community composition found onmacroalgae in the environment1.2 Problem statement and objectives of thesisWhile the biological and chemical interactions between seaweeds and their microbial sym-bionts are obviously important drivers in determining microbial community composition on2Seaweeds and their microbial symbiontsseaweed surfaces, there are also non-biological aspects of community assembly less oftenaddressed. For example, studies in the field of microfluidics have shown that water flowmay be important in settlement rates of microbial populations [Rusconi and Stocker, 2015].However, flow is seldom considered in studies on microbial community assembly [Rusconiand Stocker, 2015]. Additionally, studies show that the presence of some macroalgae causesfaster biofilm colonization because organic exudates from the seaweed enriches certain taxain the water column. This suggests that the richness and composition of water column com-munities may have important effects on the progression of biofilms. Again, knowledge abouthow macroalgal surface communities will change as the pool of potential colonizers shift islimited. Thus, there are many aspects of microbial settlement on seaweeds that are not yetunderstood.In my thesis, I address two specific questions related to microbial community assembly onthe surface of macroalgae. First, I ask whether the morphology of seaweed affects the settle-ment and structure of microbial biofilms. This work combines aspects of both microfluidicsand community ecology to make observations about how water flow can influence microbialgrowth and settlement on surfaces in marine environments. My second chapter investigatesthe effects of nearby macroalgae on the microbial community composition of growing sea-weeds. Specifically, I test whether the presence of macroalgae alters the microbial communitycomposition of the water column and of growing Nereocystis meristem fragments throughincubation experiments. With these experiments, I hope to contribute to the larger bodyof knowledge on seaweed-microbe associations and improve understanding of how, and why,we find the observed microbiota on seaweed surfaces.3Morphology and flowChapter 2The effect of seaweed morphology onmicrobial community settlement and com-position2.1 IntroductionSeaweeds are a crucial part of coastal ecosystems and are of great ecological, cultural, andeconomic value. For example, they provide crucial habitat for a variety of animals rangingfrom juvenile fish to grazing invertebrates [Wilson et al., 2010, Bulleri et al., 2002]. Theyare also one of the largest groups of photosynthetic marine organisms in the ocean and theyfix a significant fraction of the total carbon found in coastal marine ecosystems [Schiel andFoster, 2006, Tait and Schiel, 2011]. Finally, seaweeds are valuable to humans because oftheir role in traditional and modern aquaculture [McHugh, 2003]. Given their ecological andcultural significance, it is in our best interest to understand the various factors that impactseaweed biology and ecology.Microbes influence seaweed host fitness in a variety of ways, both positive and negative.Microbial epibionts mediate seaweed settlement and growth [Marshall et al., 2006,Singh andReddy, 2014,Matsuo et al., 2005,Joint et al., 2002], improve nutrient acquisition [Rosenbergand Paerl, 1981, Croft et al., 2005, Ilead and Carpenter, 1975, Chisholm et al., 1996], andprime the seaweed immune defence against potential pathogens [Weinberger, 2007,Küpperet al., 2002, Steinberg, 2002,Armstrong et al., 2001,Dobretsov and Qian, 2002,Maximilianet al., 1998]. However, many microbes also infect or degrade algal tissue [Küpper et al.,2002, Seyedsayamdost et al., 2011, Thomas et al., 2008]. Therefore, elucidating the pro-cesses underlying microbial community assembly on seaweed surfaces is vital to a holisticunderstanding of seaweed fitness.The epibiotic (surface-associated) microbial communities associated with seaweeds are host-4Morphology and flowspecific [Lachnit et al., 2009, de Oliveira et al., 2012]. However, there can be a tremendousamount of variation in taxonomic composition across individuals of a species [Burke et al.,2011b,Tujula et al., 2010] since the microbiota of seaweed varies both seasonally [Michelouet al., 2013,Lachnit et al., 2011] and with environmental factors like salinity [Stratil et al.,2014]. Additionally, while species specificity at coarse taxonomic levels of microbial assem-blages has been found in some studies [Lachnit et al., 2009], other work has shown that itis function, not membership, that correlates with host species [Burke et al., 2011a]. Thus,the variation in microbiota observed on different seaweed species is likely a result of manyinteracting factors; only one of which is the taxonomic identity of the seaweed itself.A survey of more than 35 species of seaweeds by Lemay et al. 2016 at West Beach near theHakai Research Institute revealed an unexpected correlation between microbial communitystructure and host morphology. That is, seaweeds with similar morphology tended to havesimilar microbial communities. For example, distantly related crustose seaweeds (each withan alternating upright foliose life history stage) shared more microbiota with other cruststhan either crust to their corresponding foliose life stage.We hypothesize that differences in epibiotic community structure depend partially on mor-phology because of the way water flows around seaweeds of various morphology. Watermoving across solid surfaces form boundary layers (velocity gradients), whose thickness canaffect nutrient transfer [Characklis and Marshall, 1990,Ollos et al., 2003,Lehtola et al., 2006]and particle deposition [Abelson and Denny, 1997]. For example, water velocity increasesas it flows around obstacles, which may promote the deposition of particles denser thanwater [Abelson and Denny, 1997]. Additionally, the behaviour of flexible objects in flowcan alter the thickness and steepness of the velocity gradients found on the surfaces of sea-weeds. Obstacles in flow (like branches) produce wakes, which can reduce overall water speedat downstream branches [Johnson et al., 2001] and reduce shearing forces experienced mymicrobes.Although flow is known to affect microbial settlement in a variety of ways [Characklis andMarshall, 1990,Ollos et al., 2003,Lehtola et al., 2006], quantifying and describing the inter-action between seaweeds and the flow of water is complex and difficult. Further, seaweedmorphology is confounded by a variety of factors that are known to influence composition ofsurface microbiota, such as chemical composition, making it difficult to determine whetherthe correlation between morphology and microbial community structure observed in Lemayet al. 2016 (unpublished) was truly due to morphology or not. Thus, in order to tease apartthe relative effects of morphology from other biological traits, we conducted a microbial5Morphology and flowsettlement experiment using artificial seaweeds.Our goal for this project is to test whether morphology, in the absence of biological factors,affects settlement rate or community composition of microbial communities.2.2 MethodsArtificial Macroalgae ExperimentFigure 2.1: Artificial Macroalgae. Finely branched (left), bladed (centre) and crustose (right)representative morphologies were cut from 0.4mm latex sheeting. AM units had surface areas ofapproximately 44cm2 (per side). Finely branched and bladed morphologies were 7.5x9cm (width xlength) and crustose morphologies were 7.5x7.5cm. Finely branched morphologies were created bycutting bladed morphologies into fine strips. Finely branched and bladed morphologies were gluedto 7.5x7.5cm laminate tiles by their “stipe”, whereas crustose morphologies were glued flat againstthe tile. Replicates at each site can be found in Table 2.1Artificial macroalgae (AM) were cut out of 0.4mm latex sheeting (Radical Rubber olive-grn40, Elastica Engineering) into three morphologies: crustose, bladed, and finely branched(Fig. 2.1). The AM were glued to 7.5 x 7.5 cm laminate tiles with silicon glue. The bladedand finely branched morphologies were glued by the base of the “stipe”, whereas the crustosemorphologies were glued flat against the tiles.The surface area for each of the three mor-phologies was approximately equal at 44cm2: the finely branched morphology was simplythe bladed morphology cut into filaments, whereas the crustose morphology was created byrounding off the pointed tip of the bladed morphology.6Morphology and flowReplicates at each time pointSite Morphology 20 m 1 h 3 h 6 h 12 h 1 d 4 dReed Point FB 3 (3) W 3 (3) 3 (3) 2 (3) 3 (3) 3 (3) W -BL 3 (3) W 3 (3) 3 (3) 3 (3) 3 (3) 3 (3) W -CR 3 (3) W 3 (3) 3 (3) 3 (3) 2 (3) 3 (3) W -Hakai FB 10 (10) 10 (10) - W 10 (10) 9 (10) - 10 (10) WBL 10 (10) 8 (10) - W 10 (10) 9 (10) - 11 (10) WCR 10 (10) 9 (10) - W 9 (10) 6 (10) - 8 (10) WTable 2.1: Number of replicates at each site for AM experiment. Replicatesfor finely branched (FB), bladed (BL), and crustose (CR) morphologies are shown in thetable below. Numbers in brackets are number of replicates originally deployed. Somereplicates were lost at the sampling stage, while others at the sequence quality filteringstage. The BL morphology on day 4 has 11 replicates because one blade from time point"1 hour" was mislabelled as "4 days". "W" indicates time points when water samples weretaken.The first experiment was conducted at Reed Point Marina in the Burrard Inlet near Vancou-ver, BC while a larger second experiment was done in the Central Coast of BC on CalvertIsland, BC at the Hakai Research Institute. The two sites differed in a number a characteris-tics including salinity, season, and geographic location. At Reed Point, salinity was between12-15ppt throughout the course of the experiment and sampling was done in March 2016.The Burrard Inlet is influenced by several rivers, most notably the Fraser River, and thesalinity is highly variable over the course of the year. The second experiment was done onCalvert Island at the Hakai Research Institute dock in mid-summer (August 2016) and thesalinity was 32ppt for the entire experiment and is largely consistent seasonally.In both experiments, tiles attached to the artificial macroalgae were suspended off the docksat approximately 1m below the water surface and positions of morphologies, replicates,and time points along the docks were randomized. Sampling was destructive. The firstexperiment was conducted in March 2016 at Reed Point Marina and three replicates ofeach morphology were sampled destructively at each of five time points: 20 minutes, 3hours, 6 hours, 12 hours, and 24 hours. Three water samples were taken concurrently at20 minutes and 24 hours, each. We chose these time points to capture initial colonizationand community dynamics. Results from experiment one at Reed Point were intriguing, butsample size was small. Thus, we conducted experiment two at Hakai with more replicatesand with time points adjusted according to initial results to capture initial colonization whilealso capturing a more mature community (20 minutes, 1 hour, 6 hours, 12 hours, 4 days,water samples at 3 hours and 4 days). The different location also allowed us to assess the7Morphology and flowrobustness and generality of these findings. A table of sampling times and replicates can befound in Table 2.1.AM units were sampled at each time point by rinsing each unit for 10 seconds with filter-sterilized seawater, and then swabbing one side of the surface for 10 seconds using sterilecotton swabs (Puritan- Item#: CA10805-154). The swab heads were then snapped off into2mL cryotubes (VWR- Item#: 10018-760) and frozen at -200C until extraction. Salinitywas measured using a refractometer.At Hakai, water flow was estimated by suspending pre-weighed Lifesaver brand “Pep-o-mints” on a cotton string 1 m below the water surface for one minute. The mints were thenoven dried for at least one hour, or until dry to the touch. The dry weight before and aftersubmersion was calculated and compared to the standard curve in Anderson and Martone(2014). According to their standard, the water flow at Hakai was less than 0.5 m/s and closeto 0m/s, although we observed movement in the water visually. Water flow at Reed PointMarina was quantified visually: flow was also minimal.Water samples were taken on day 1 and day 4 at Hakai, and at 20 minutes and 24 hoursat Port Moody. Three separate 500mL samples of water at each time point were pre-filtered through 150uM sieves to minimize sampling plant-debris and large animals, and thenfiltered through 0.22um membranes (Durapore- Item#: GVWP04700) using a peristalticpump (Cole-Parmer- Item#: RK-77913-70) at approximately 180rpm (level 30). Filterswere transferred to 2mL cryotubes and either frozen immediately at -200C (Hakai) or kepton dry ice until return to the lab (Reed Point). All samples were transported frozen back toUBC for further processing.Library prepThe 96-well MoBio PowerSoil DNA extraction kit was used to extract DNA from both thewater filters and AM swabs. Filters and swabs were transferred to the extraction kit usingtweezers sterilized with 2% HCl, then ethanol and flame. Extractions followed the MoBioPowersoil DNA extraction protocol with the following modifications based on recommen-dations in the Earth Microbiome protocol (http://www.earthmicrobiome.org/). First, a10-minute incubation at 650C in a water bath was added after the addition of C1 and beforeshaking in the shaker, next, the plates were shaken at 20 shakes per second for 20 minutestotal instead of using the recommended shaker in the manual, and lastly filters were allowedto soak in C6 at room temperature for 10 minutes prior to elution into the final elution plate.The DNA elution was stored at -200C.8Morphology and flowThe V4 region of the 16S small subunit ribosomal RNA marker gene was sequenced to pro-file bacteria and archaea. The amplicon library prep was done in-lab using the followingprimers: barcoded 515 forward primers (5’-GTGYCAGCMGCCGCGGTAA-3’) and 806 re-verse primers (5’-GGACTACHVGGGTWTCTAAT-3’). Forward primers were tagged with12bp Golay barcodes, which were unique to each sample. Primers were used at final concen-trations of 0.5uM with 4-5uL of DNA extract. DNA extracts were amplified in 20uL reactionsusing Phusion High-fidelity proofreading Mastermix (Thermofisher- Item#: F548L). Reac-tions underwent the following thermocycler settings: 980C for 10 seconds, 25 cycles of 980C(1s), 500C (5s), 720C (24s), and a final extension phase of 720C for 1 minute. Samples thatwere not amplified adequately with 4uL of DNA at 25 cycles were re-amplified with 5uLDNA at 30 cycles. Lastly, the successful PCR products were quantified using Pico-green(Thermofisher- Item#: P11496) and pooled at 45ng/sample. The pooled samples were thensent to the Centre for Comparative Genomics and Evolution Bioinformatics (CGEB) at Dal-housie University for sequencing on the Illumina MiSeq platform with 2x300bp chemistry.Processing of sequence dataRaw sequences were demultiplexed using split_libraries_fastq.py in QIIME version 1.9(Quantitative Insights into Microbial Ecology, (QIIME), [Caporaso et al., 2010]) and thencombined with the entire West Beach data set from Lemay (2016). All sequences weretrimmed (fastx_trimmer) and clipped (fastx_clipper) to 250 bp and filtered with a qual-ity threshold of Q19 (fastq_quality_filter) with the Fastx Toolkit (Hannon Lab), yielding22,648,504 raw sequences. The remaining reads were processed into operational taxonomicunits (OTUs) using Minimum Entropy Decomposition (MED, [Eren et al., 2014]) with theminimum substantive abundance (-m) parameter set to 150, yielding 5,820 unique OTUsand 15,840,802 reads. The resulting OTU matrix was transcribed into a QIIME-compatibleformat. Taxonomy was then assigned to the representative sequence for each MED node,hereafter referred to as OTUs, by matching it to the SILVA 128 database clustered at 99%with assign_taxonomy.py in QIIME using uclust [Edgar, 2010]. Chloroplast, mitochondrial,and eukaryotic DNA were filtered out. Thirty-eight OTUs suspected of being contaminants(abundance in PCR controls at least 50% of the maximum abundance found in any othersample) were also removed. Twenty-five contaminants were from the genus Pseudomonas andeight from the genus Achromobacter, which are both common lab contaminants. The remain-ing contaminants were an uncultured Alphaproteobacterium, a Lactobacillus (Firmicutes), aDokdonia (Flavobacteriia), an uncultured member of Cellvibrionaceae (Gammaproteobacte-ria), and a member of Clostridiales (Firmicutes). PCR controls samples had between 1 and9,417 reads per sample. Finally, samples with less than 1000 reads per sample were removed.9Morphology and flowThe final OTU table consisted of 5,820 unique sequences and 15,840,802 reads, with a meanof 25,508 reads per sample. For alpha and beta diversity analysis, samples were rarefiedto 4,000 reads per sample. Representative sequences were aligned with PyNAST in QIIMEand a phylogenetic tree was created using FastTree [Price et al., 2009] in QIIME with themake_phylogeny.py script.Microbial richness across time and morphologiesAlpha diversity was calculated in QIIME with the metrics Chao1, PD_whole_tree, andObserved_otus. Chao1 is a metric that incorporates estimates of unobserved diversity intotheir richness calculations, which is valuable in microbial datasets because samples generallyhave many rare taxa [Chao, 1984]; PD_whole_tree calculates diversity given phylogeneticdistance [Faith and Baker, 2006]; and observed_otus is simply the richness of the sample.These three metrics were chosen because they differed in the way diversity is quantified,and thus would provide meaningful comparisons. Alpha diversity values were imported intoR [R Core Team, 2016] for statistical analysis. Overall differences in diversity between timeand morphology (ANOVA: Morph + Time + Morph*Time, non-sequential (Type III)) wereassessed using “lm” and “Anova” in the “car” package [Fox and Weisberg, 2011]. Differ-ences between morphologies were also assessed at each time point using a one-way ANOVA(“anova”: chao1 Morph “stats” package [R Core Team, 2016]) for overall comparisons andpairwise t.tests (“pairwise.t.test” :“stats” package [R Core Team, 2016]) for pairwise compar-isons. For pairwise comparisons, p-values were corrected for multiple comparisons using theBenjamini-Hochberg procedure (also known as the “False Discovery Rate” method), whichcontrols for the proportion of expected type I errors [Benjamini et al., 1995].Community composition across time and morphologiesIn order to compare community composition across morphologies and time, distance matriceswere created in beta_diversity.py (QIIME) with a rarefied OTU table using three metrics:unweighted Unifrac, weighted Unifrac [Lozupone and Knight, 2005], and Bray-Curtis [Brayand Curtis, 1957]. The Unifrac metric assesses community distance by comparing the num-ber of shared to unshared branches in a phylogenetic tree of two communities (unweightedunifrac does not account for abundance and weighted Unifrac does), while Bray-Curtis in-corporates only membership and relative abundance into its distance calculations. Distancematrices were imported into R and the “isomds” command from the “MASS” package [Ven-ables and Ripley, 2002] was used to create 2-dimensional NMDS plots. Polygons were drawnaround treatments by using “chull” in the “grDevices” package [R Core Team, 2016]. First,10Morphology and flowwe tested for overall differences in community composition by running a permutational anova(“adonis”) from the “vegan” package [Oksanen et al., 2017] with the factors time and mor-phology plus interactions (Time + Morph + Time:Morph). Then, we ran individual pairwisecomparisons between the three morphologies at each time point (“adonis”: Morph). Pair-wise comparisons were corrected for multiple comparisons using the Benjamini-Hochbergprocedure.Taxa summary plots were generated by combining taxa summaries outputs from summa-rize_taxa.py (QIIME) at the L4 level and the OTU level in base R.To create the OTU heat map, the OTU table was filtered to only include core OTUs frommorphologies and water samples at each site using a custom python script. This was doneto reduce the overall number of OTUs in the table. A “core OTU” was defined as an OTUfound in >=90% of samples in any given treatment (a single morphology at a single timepoint), whose maximum relative abundance in the entire data set is over 0.1%. The OTUtable was imported into R and reordered using “cor” (Pearson correlation coefficient) tomaximize clustering of abundant OTUs based on treatment. Finally, OTU abundances forthe filtered table were visualized using “heatmap2” in “gplots” [Warnes et al., 2016]. OTUsare scaled by row, which means that the intensity of colour shows the abundance of thatOTU relative to the average abundance of that OTU in all samples.Dispersion of morphologies across timeDispersion of AM units through time was quantified two ways. First, the mean distance (fromthe distance matrix) between all combinations of two morphologies at each time point wascalculated. This yielded three pairwise comparisons (FB:CR, CR:BL, BL:FB). Additionally,overall dispersion of each time point (PERMDISP) was calculated using “betadisper” fromthe “vegan” package [Oksanen et al., 2017]. Differences in dispersion between time pointsand morphologies was also assessed separately using “betadisper” (“vegan” package) and“anova” (“stats” package [R Core Team, 2016]).Turnover across timeTurnover was defined as the number of “new” OTUs at each time point (relative to theprevious time point) divided by the total number of OTUs. This was calculated for all timepoints (except the first time point) and plotted using plot() in base R.Test to quantify shear force on morphologies11Morphology and flowWe conducted a dye-dipping experiment to determine whether dye on different morphologiesof AM was sheared off the surface at different rates. The experiment modified a method fromHoegh-Guldberg (1988) [Hoegh-Guldberg, 1988], which used a solution of methylene blueand Triton-X (a detergent) to estimate the surface area of coralline seaweeds. A viscous dyesolution was made by dissolving approximately 0.8g of methylene blue (Michrome) and 1mLof Triton X (BDH- Item#: R06433) into 1L of tap water. The solution was then filteredthrough fiber-glass GF/F filters (Whatman- Item#: 28497-958) to remove solid particlesand impurities.Ten replicates of finely branched and bladed AM morphologies were subjected to threetreatments. In the first treatment the AM units were dipped in dye, shaken 10 times andthen put directly in 50mL falcon tubes with 25mL of tap water to measure how much dye isinitially adhered to the surface. For the second treatment, AM units were dipped, shaken 10times, and then submerged in a 1L beaker of still tap water for 5 seconds before transferringAM units into 50mL falcon tubes with 25mL of water. In the third treatment, AM unitswere dipped, shaken 10 times, and submerged in a 1L beaker full of tap water that wasstirred by a magnetic spinner at maximum speed before transferring AM units into 50mLfalcon tubes with 25mL of water. In all three treatments, the 50mL falcon tubes were shakenvigorously for 5 seconds and allowed to soak for 1 hour, then AM units were removed fromthe falcon tube, rinsed, and dried. Subsequently, the dye concentration in the water of the50 mL falcon tubes were measured by taking absorbance measurements at a wavelength of668 nm with the Jaz Spectrometer using the Spectrasuite software. Readings were done overintegration times of 100ms each, and final measurements were averaged over 100 scans. Thesame 10 AM units were used for each of three treatments to allow for direct comparisonsbetween the three treatments on a single AM unit.2.3 ResultsSequencing yielded 1,123-106,093 sequences per sample after quality filtering; the final num-ber of replicates and water samples included in the analysis after quality filtering can befound in Table 2.1.Results were consistent between richness and community distance metrics. Since all alphadiversity metrics yielded similar conclusions, only Chao1 results are shown and describedbelow. Additionally, only results from Bray-Curtis are shown because it incorporates bothmembership and abundance into its calculations- both of which appear to be important in12Morphology and flowdefining community structure across treatment groups.Community composition across morphologies through timeA. Reed Point B. Hakail l l l200400600800100012001400chao1TimeRichness (chao1)20m 1h 6h 12h***** Finely BranchedBladedCrustosell l l l l2004006008001000chao1TimeRichness (chao1)20m 1h 3h 6h 12h24h*****Finely BranchedBladedCrustoseFigure 2.2: Richness of biofilms on artificial macroalgae through time. Richness (Chao1metric) of AM biofilms through time at (A) Reed Point and (B) Hakai. Significance is indicated bystars: * = p < 0.05; ** = p < 0.01; *** = p < 0.001. Error bars are +/-1SD. Richness increaseswith time and finely branched morphologies are more diverse than bladed or crustose morphologiesat intermediate time points (See Table 2.2 for pairwise comparisons). The interaction term betweenTime and Morphology was not significant, but there is a trend where finely branched morphologiesare more diverse at intermediate time points.Finely branched morphologies initially accumulated diversity faster than bladed and crus-tose morphologies (Fig. 2.2). Although the interaction between morphology and time isnot significant (which would imply that the rate of diversity increase differed between mor-phologies) (Hakai: ANOVA MorphxTime p = 0.0575, F2,133 = 2.918; Reed Point: ANOVAMorphxTime p = 0.202, F2,46 = 1.655), finely branched AM units are significantly morediverse than bladed or crustose AM units at early (and most) time points (Table 2.2, Hakai:ANOVA Morph p < 0.001, F1,133 = 26.653; Reed Point: ANOVA Morph p < 0.001, F1,46= 38.451). At later time points, however, similar richness of microbial taxa is found on allthree morphologies.The composition of microbial communities also became more distinct across morphologiesover the first part of the time series in both the Hakai and Reed Point data sets (Fig. 2.3). In13Morphology and flowp-Values: between morphologiesSites Test type Factors 20 minutes 1 hour 3 hours 6 hours 12 hours 1 day 4 days OverallReed Point ANOVA Morph 0.039(F2,6=5.86)0.0071(F2,6=12.59)0.048(F2,6=5.23)0.43(F2,5=0.99)0.021(F2,5=9.29)0.15(F2,6=2.6)- 0.0033(F2,46=6.48)FB:BL 0.045 0.03 0.069 0.5 0.028 0.24 - 0.012Pairwise t-tests FB:CR 0.045 0.0077 0.071 0.5 0.028 0.21 - 0.0049BL:CR 0.8 0.12 0.61 0.5 0.45 0.56 - 0.55Hakai ANOVA Morph 0.16(F2,27=1.93)0.003(F2,24=7.45)- <0.001(F2,26=10.12)0.1(F2,21=2.57)- 0.67(F2,26=0.4)<0.001(F2,133=16.86)FB:BL 0.69 0.024 - <0.001 0.6 - 0.91 0.013Pairwise t-tests FB:CR 0.22 0.0031 - <0.001 0.12 - 0.69 <0.001BL:CR 0.23 0.33 - 0.94 0.15 - 0.69 0.16Table 2.2: Pairwise t-tests and ANOVA of richness between morphologies across time.Two tests were done to quantify richness over time. First, an ANOVA was done overall (Morph,Time, MorphxTime) and at each time point (Morph). Results show p values of the ‘Morph’ factoronly. Additionally, pairwise t-tests were done at each time point to determine which morphology(if any) was driving significant results in the ANOVA. All pairwise p-values are adjusted usingthe Benjamini-Hochberg procedure (False Discovery Rate). Results show that the finely branched(FB) morphology drives the significant effect for ‘Morph’ in the ANOVA results. Bladed (BL) andcrustose (CR) morphologies do not differ in richness at any time point. At the final time pointfor Reed Point and Hakai (1 day and 4 days, respectively), there are no significant differences inrichness between morphologies.both experiments, compositional differences between morphologies are most pronounced atintermediate time points (3-12 hours, depending on the data set), but communities convergeat later time points (Fig. 2.3).Plotting the core OTUs associated with each morphology over time reveals that OTUsare differentially abundant on finely branched morphologies relative to crustose or bladedmorphologies (Fig. 2.4). The enrichment of OTUs on finely branched morphologies correlateswell with a map of core OTUs (Fig. 2.4, Fig. S2.11), but poorly with a presence/absencemap (Fig. S2.12). This suggests that finely branched morphologies accumulate biofilmsthat are structurally different from crustose or bladed biofilms, not necessarily in terms ofmembership, but rather in terms of differential abundance.The pattern of overall community turnover over time differed between the two experiments.At Hakai, microbial communities across all morphologies became less dispersed over time(Fig. 2.3b, Fig. S2.10b, PERMDISP p < 0.001, F4,131=97.403), suggesting that microbialcommunities on all AM units became more similar as settlement progressed. Additionally,core OTU turnover was low at Hakai (Fig. 2.5), which means core members of the communitywere not changing. Conversely, at Reed Point, dispersion does not change over time (Fig.2.3a, Fig. S2.10a, PERMDISP p = 0.131, F5,45=1.807), suggesting that biofilms on AM unitsare just as different from each other at the beginning of the experiment as they are at the14Morphology and flowA. Reed Point0.30.50.70.9Dispersion of morphologies across timeDistance (Bray−Curtis)20 min 1 h3 h6 h12 h24 hStress: 0.12FB:CRFB:BLCR:BLllll llllln lllllllln llllllllln**l llllllln*lllllllln**llllllllln lllCRBLFBB. Hakai0.20.40.60.8Dispersion of morphologies across timeDistance (Bray−Curtis)20 min 1 h6 h12 h 4 dStress: 0.18FB:CRFB:BLCR:BLllllllllllllllllllll lllllll**llllllllllllllllllllllll**ll lllllllll llllllllllllllll**llllll l llllllllllll ll***lll llllllllllll ll lll l Stress: 0.06lllCRBLFBFigure 2.3: Dispersion and community composition of AM morphologies through time.Pairwise dispersions were calculated by averaging distances from a Bray-Curtis distance matrixfor each pair of morphologies. Error bars show +/-1SD. Below each time point is a correspondingNMDS plot of samples from that time point, coloured by morphology. Significant differencesbetween morphologies (PERMANOVA Morph) are listed below each plot and shown with stars: *= p < 0.05; ** = p < 0.01; *** = p < 0.001. Complete statistics are shown in Table 2.3. At ReedPoint (A), dispersion of morphologies (shown above) and overall dispersion (Fig. S2.10a) doesnot differ significantly with time. At Hakai (B), dispersion of morphologies (shown above) andoverall dispersion (Fig. S2.10b) decreases significantly with time. However, in both experiments,communities initially become more distinct with time. The 4-day time point of Hakai was plottedfrom a separate NMDS calculation (and thus has a different stress value) to better capture within-time-point variation between morphologies.end. There was also higher turnover of core OTUs through time at Reed Point (Fig. 2.5).15Morphology and flowp-Values: between morphologiesSites Test type Factors 20 minutes 1 hour 3 hours 6 hours 12 hours 1 day 4 days OverallReed Point PERMANOVA Morph 0.078(R2=0.36,df=2,8)0.54(R2=0.28,df=2,7)0.007(R2=0.42,df=2,8)0.011(R2=0.42,df=2,7)0.003(R2=0.48,df=2,7)0.059(R2=0.39,df=2,8) -0.001(R2=0.11,df=2,33)FB:BL 1(R2=0.16, df=1,5)1(R2=0.22, df=1,5)0.6(R2=0.27, df=1,5)0.3(R2=0.34, df=1,4)0.3(R2=0.28, df=1,5)0.3(R2=0.4, df=1,5) -0.057(R2=0.07, df=1,34)Pairwise PERMANOVA FB:CR 0.3(R2=0.39, df=1,5)1(R2=0.23, df=1,4)0.3(R2=0.36, df=1,5)0.3(R2=0.45, df=1,4)0.3(R2=0.52, df=1,4)0.3(R2=0.41, df=1,5) -0.003(R2=0.12, df=1,32)BL:CR 0.3(R2=0.28, df=1,5)1(R2=0.21, df=1,4)0.3(R2=0.42, df=1,5)0.6(R2=0.26, df=1,5)0.3(R2=0.43, df=1,4)1(R2=0.15, df=1,5) -0.063(R2=0.07, df=1,33)Hakai PERMANOVA Morph 0.009(R2=0.14,df=2,26)0.006(R2=0.14,df=2,26) -0.002(R2=0.2,df=2,28)0.001(R2=0.17,df=2,23) -0.1(R2=0.12,df=2,28)0.001(R2=0.09,df=2,95)FB:BL 0.13(R2=0.1, df=1,17)0.033(R2=0.13, df=1,17) -0.006(R2=0.2, df=1,19)0.006(R2=0.16, df=1,17) -0.99(R2=0.05, df=1,20)0.012(R2=0.04, df=1,94)Pairwise PERMANOVA FB:CR 0.075(R2=0.13, df=1,16)0.015(R2=0.12, df=1,18) -0.003(R2=0.19, df=1,18)0.012(R2=0.17, df=1,14) -0.12(R2=0.15, df=1,17)0.003(R2=0.05, df=1,87)BL:CR 0.13(R2=0.09, df=1,18)0.59(R2=0.08, df=1,16) -0.4(R2=0.08, df=1,18)1(R2=0.07, df=1,14) -0.52(R2=0.09, df=1,18)0.46(R2=0.02, df=1,88)Table 2.3: Pairwise and overall PERMANOVAs of community composition betweenmorphologies across time. Two tests were done to quantify community dissimilarity over time.First, a PERMANOVA was done overall (Morph, Time, MorphxTime) and at each time point(Morph). Results show p-values of the ‘Morph’ factor only. Additionally, pairwise PERMANOVAswere done at each time point to determine which morphology (if any) was driving significantresults in the overall PERMANOVA. P-values for pairwise comparisons were adjusted using theBenjamini-Hochberg method (False Discovery Rate). Results show that the finely branched (FB)morphology drives the significance behind the overall PERMANOVA results. Bladed (BL) andcrustose (CR) morphologies do not differ significantly in composition at any time point in eitherHakai or Reed Point samples. At the final time point, differences between morphologies are notsignificantly different at either site.However, despite differences in community turnover trends, both experiments saw microbialcommunities on differing morphologies become more distinct (Fig. 2.3). This suggests thatdistinctiveness of communities on AM morphologies (whether FB, CR, and BL morphologiesare different or not) is decoupled from gamma diversity (heterogeneity between samples andtreatments) at these two sites.Shared core OTUs between morphologies and time pointsCommon inhabitants of biofilms found on AM units at both Hakai and Port Moody includedthose from the Order Pseudomondales, Alteromonadales, Burkholderiales, Flavobacteriales,and an unidentified representative of Alphaproteobacteria (Fig. 2.6). Flavobacteriales andthe unidentified Alphaproteobacteria are also found at significant abundances in the water.Members of Pseudomondales appear to decrease as microbial communities progress, whereasFlavobacteriales increase. AM surface communities begin less taxonomically diverse than thewater column, but are significantly more diverse than the water at later time points (Fig.S2.13). Hakai AM surface communities are more diverse than ones at Reed Point (Welch’st-Test p = <0.001, t150.6 = 4.316; Fig. S2.14, Table S2.6), but water column diversity doesnot differ between Hakai and Reed Point (PERMANOVA p = 0.334, F1,8 = 1.057). Thus,through some mechanism, diversity of microbial biofilms on AM surfaces is greater at Hakai16Morphology and flowthan Reed Point despite no apparent difference in diversity of the water column.A. Reed Point B. Hakai20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 d−4 0 2 4Row Z−ScoreColor KeyllllFinely branchedBladedCrustoseWater20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d6 h4 d−2 0 2Row Z−ScoreColor KeyllllFinely branchedBladedCrustoseWaterFigure 2.4: Heatmap of core OTUs across morphologies and time points. The heatmapshows core OTUs only. Core OTUs are defined as being present in >90% of samples in eachtreatment (morphology per time point) and observed at >0.1% relative abundance at least oncein entire dataset. Colouring is scaled by rows, meaning that each OTU is coloured based on howabundant it is compared to other samples. At both Reed Point (A) and Hakai B, there is differentialenrichment of OTUs between finely branched and bladed or crustose morphologies.At early time points, there are many shared dominant taxa between sites at the Order level(Fig. 2.6), but few shared OTUs (Fig. S2.15). However, a few shared OTUs emerge late incolonization, including Oleispira, which dominates biofilm communities on day 4 at Hakai.Selection for latex-degrading bacteria is important at later time pointsCommunities sampled on day 4 are drastically different to all other time points, and aredominated by a single OTU from the genus Oleispira (Fig. 2.6). Bacteria from the genusOleispira are known to be hydrocarbon degraders [Yakimov et al., 2003,Mason et al., 2003],and we believe this OTU is degrading the latex from our AM units. There were 87 OTUs inboth data sets assigned to Oleispira, and 44 OTUs with the exact same taxonomic descriptionas the OTU described above, but only one OTU was ever observed at relative abundancesover 2% (with most at abundances below 0.1%).Oleispira was found in all Hakai water samples and on all morphologies across all timepoints, but always at abundances less than 0.07% in the water and less than 5.1% on AM17Morphology and flow0.000.100.200.30Percent community turnoverTime1 hour3 hours6 hours12 hours1 day4 daysl l l llllllllllllll llllllFBBLCRllReed PointHakaiFigure 2.5: Turnover of OTUs across time points. Percent of OTUs at each time point thatwere not found in previous time point are plotted against time– in other words, this represents thepercent turnover of communities on each morphology through time. There are consistently higherrates of turn over on AM units at Reed Point compared to Hakai. There is almost no turnover atmost points at Hakai.units (average 0.7%, excluding 4 outliers with 30%+ abundances) in time points 20min-12h. Oleispira was also found in the Reed Point data set in both the AM samples andthe water column. Like the Hakai data set, Oleispira from Reed Point are abundant in allwater column samples at abundances less than 0.07%. Also, AM unit samples in time points20min-12h at Reed Point are composed of less than 5% (average 0.6%) Oleispira, while itsabundance in the 24hour Reed Point time point is 4-20% (average 9%). Thus, we predictthat Oleispira would have dominated the communities at Reed Point like they did at Hakaiif the experiment had continued past 24 hours.We observed Oleispira in similar abundances in the water column at both sites at all timepoints, which suggests that Oleispira is found natively in the water column. However, sincewe did not sample AM units prior to either experiment (although we did dip AM units inethanol in the Hakai experiment) we cannot be certain that Oleispira was not introduced tothe AM units in the lab.Dye experiment resultsThe results from the dye–dipping experiment to test whether shear forces differed on thesurface of AM units yielded insignificant results (Fig. S2.16). Although there were trends18Morphology and flowA. Reed Point B. Hakai20m 1h 3h 6h 12h 24hFinelyBranched20m 1h 3h 6h 12h 24hBladed20m 1h 3h 6h 12h 24hCrustose20m 1dRELATIVE ABUNDANCE20m 1h 6h 12h 4dFinelyBranched20m 1h 6h 12h 4dBladed20m 1h 6h 12h 4dCrustose6h 4dWaterRELATIVE ABUNDANCELegendUnassigned;Other;Other;OtherBacteria;__Verrucomicrobia;__Verrucomicrobiae;__VerrucomicrobialesBacteria;__Proteobacteria;__Proteobacteria;__GammaproteobacteriaBacteria;__Proteobacteria;__Proteobacteria;__AlphaproteobacteriaBacteria;__Proteobacteria;__Gammaproteobacteria;__XanthomonadalesBacteria;__Proteobacteria;__Gammaproteobacteria;__VibrionalesBacteria;__Proteobacteria;__Gammaproteobacteria;__PseudomonadalesBacteria;__Proteobacteria;__Gammaproteobacteria;__OceanospirillalesBacteria;__Proteobacteria;__Gammaproteobacteria;__CellvibrionalesBacteria;__Proteobacteria;__Gammaproteobacteria;__AlteromonadalesBacteria;__Proteobacteria;__Epsilonproteobacteria;__CampylobacteralesBacteria;__Proteobacteria;__Betaproteobacteria;__MethylophilalesBacteria;__Proteobacteria;__Betaproteobacteria;__BurkholderialesBacteria;__Proteobacteria;__Alphaproteobacteria;__SphingomonadalesBacteria;__Proteobacteria;__Alphaproteobacteria;__RhodospirillalesBacteria;__Proteobacteria;__Alphaproteobacteria;__RhodobacteralesBacteria;__Proteobacteria;__Alphaproteobacteria;__RhizobialesBacteria;__Proteobacteria;Other;OtherBacteria;__Firmicutes;__Bacilli;__LactobacillalesBacteria;__Deinococcus−Thermus;__Deinococci;__DeinococcalesBacteria;__Cyanobacteria;__Cyanobacteria;__SubsectionIBacteria;__Bacteroidetes;__Sphingobacteriia;__SphingobacterialesBacteria;__Bacteroidetes;__Flavobacteriia;__FlavobacterialesBacteria;__Bacteroidetes;__Cytophagia;__CytophagalesBacteria;__Actinobacteria;__Actinobacteria;__MicrococcalesBacteria;__Acidobacteria;__Acidobacteria;__AcidobacterialesBacteria; __Proteobacteria; __Gammaproteobacteria; __Oceanospirillales; __Oceanospirillaceae; __OleispiraFigure 2.6: Taxa summary of OTUs collapsed by Order. Taxa summaries show all replicatesfrom all time points at the Order level, except for Oleispira, which is shown at the genus level.Oleispira is observed at very low abundances (<0.7%, excluding outliers) until the 12 hour timepoint in both experiments. At Hakai, it is the most abundant OTU on all morphologies by day 4.Hakai samples are more diverse than Reed Point samples (Fig. S2.14).where the bladed morphology lost more dye, these trends were not supported statistically(Welch’s t-Test “Moving water” p = 0.39, t11,x = 0.89; “Still water” p = 0.51, t9.9,x = 0.68).There was a great amount of noise in all treatment groups, so if there were any real effects19Morphology and flowof morphology it was likely swamped by the natural variation found in our data. Thus,although the results yielded insignificant results across morphology, we do not believe itmeans shear forces on the surface of AM morphologies are necessarily the same. Rather, theprecision of our methods was likely insufficient to detect any differences.2.4 DiscussionFlow and morphology interact to influence microbial settlementThe direction and speed of water flow across solid surfaces is known to influence microbialsettlement and biofilm development. For example, biofilm formation increases when waterspeeds are high [Rusconi et al., 2014,Rusconi and Stocker, 2015] and when environments arephysically heterogenous [Abelson and Denny, 1997, Singer et al., 2010]. Additionally, flowspeed [Stewart and Carpenter, 2003] and spatial heterogeneity [Singer et al., 2010] of surfacescan affect the rate of mass transfer of nutrients to biofilms, which can further influence biofilmdevelopment. However, despite the acknowledgement that water flow is potentially crucialto understanding how and when microbes colonize surfaces [Rusconi et al., 2014,Rusconi andStocker, 2015], there has been little experimental work that tests the interaction betweenflow and seaweed morphology in situ. Therefore, we wanted to experimentally test whetherseaweed morphology effects overall biofilm community development.Our results show that macroalgal morphology affects both microbial settlement and commu-nity composition. In both replicate experiments, finely branched morphologies experiencefaster microbial settlement rates than bladed or crustose morphologies (Fig. 2.2, Table2.2). This supports our hypothesis that the narrow tendrils on finely branched morpholo-gies cause greater microbial deposition, since particles denser than water will separate fromthe main stream of flow as it moves around protrusions (a phenomenon called ‘inertial im-paction’) [Abelson and Denny, 1997]. The behaviour of finely branched tendrils in flow mayalso produce slower-flowing ‘wakes’ [Johnson et al., 2001]. Slower water velocity at down-stream tendrils may additionally facilitate settlement by reducing shear forces experiencedby microbes [Rutter and Vincent, 1988].Finely branched morphologies also hosted structurally different microbial communities com-pared to bladed and crustose morphologies. We find that it is differential enrichment of coreOTUs, and not differences in membership, that drives the divergence of finely branched mor-phologies (Fig. 2.4). These trends are consistent with observations made in other systems:biofilms found in stream beds with varying levels of physical heterogeneity also differ in rel-20Morphology and flowative abundance of core members rather than membership [Singer et al., 2010]. Streamlineand turbulent flow regimes can also produce a variety of biofilm structures (chain-like versuscolony-like growth patterns, respectively) [Singer et al., 2010], and influence the mass trans-fer rates of nutrients [Stewart and Carpenter, 2003]. Therefore, we hypothesize that shifts incommunity structure found between our AM morphologies are likely due to a combination ofdifferential mass transfer of nutrients [Singer et al., 2010,Stewart and Carpenter, 2003] andphysical biofilm structure [Singer et al., 2010] induced by differential flow regimes aroundAM unit morphologies.Dominant OTU Oleispira masks signal of morphology over timeDifferences in community structure between morphologies became insignificant in the lasttime points of both experiments. This correlates to the sudden growth (and eventual dom-inance) of a few taxa, including a latex-degrader, Oleispira. There are small differences incommunity composition between AM morphologies at the 4 day time point at Hakai (seein Fig. 2.5), but the signal of morphology is seemingly overwhelmed by the signal of a fewsubstrate specialists (Oleispira and a few others).In contrast to our AM morphologies, whose biofilms were highly uneven, real seaweeds aregenerally host to a diverse collection of coexisting microbial lineages [Egan et al., 2013]. The-ories in ecology predict that habitat heterogeneity and disturbance both increase biodiversityin local communities [Huston, 1979]. Therefore, seaweed biofilms are likely more even thanlatex biofilms because they possess greater physical surface heterogeneity (many seaweedsare textured or ruffled), and also exude a variety of antimicrobial compounds, which can beconsidered a source of ‘press’ disturbance and are known to increase evenness in microbialcommunities associated with seaweeds [Persson et al., 2011].Relevance for real seaweedsLemay et al. (2016) observed a strong correlation between seaweed morphology and mi-crobial community composition in the species Mastocarpus spp.. Our experiment showsthat morphology alone is sufficient to drive differences in microbial community structurebetween artificial seaweeds. On real seaweeds, however, other factors like polysaccharidechemistry [Dininno and McCandless, 1978, Evelegh et al., 1979, Falshaw et al., 2003] andtissue age [Bengtsson et al., 2012] correlate with morphology. These factors potentially con-tribute to the differences in microbial community structure observed on Mastocarpus spp. inLemay et al. (2016). Therefore, although morphology drives differences in community struc-ture on artificial seaweeds, it is not necessarily the single driver of differential community21Morphology and flowstructure between real seaweed morphologies.On real seaweeds, initial settlement of microbes may influence final community structurethrough downstream settlement effects. It is known that many different microbes are capableof performing similar metabolic tasks within the seaweed microbiome and it was previouslyhypothesized that neutral mechanisms like the competitive lottery model may explain thehigh taxonomic variation observed between individual macroalga microbiota [Burke et al.,2011a]. However, differences in initial biofilm members driven by morphology may also helpexplain the variation found in final (or climax) communities of macroalgal biofilms. Microbesare known to produce a range of antimicrobial compounds that prevent invasion from otherbacteria [Matz et al., 2008,Egan et al., 2008], and pre-established biofilms are known to resistinvasion from several laboratory strains of bacteria [Rao et al., 2010]. Conversely, manytypes of bacteria are described has having facilitative effects on the growth and attachmentof invertebrates, algal spores, and diatoms [Huggett et al., 2006]. Thus, initial colonizationdifferences among morphologies and seaweeds may result in drastically different downstreammicrobial and epiphytic communities.Our results also open many questions about the interaction between seaweed morphologyand its epibionts. Microbial density on seaweed surfaces can vary by as much as five orders ofmagnitude [Armstrong et al., 2001,Bengtsson and Øvreås, 2010], and one might ask whethermorphological complexity correlates with microbial density. Microbial density can influenceboth the direct and indirect effects of seaweed biofilms on their host. For example, somebacteria are known to increase settlement of macro-epiphyte spores and larvae (such as otherseaweeds and invertebrates) [Huggett et al., 2006] and epiphytes increase drag experiencedby their macroalgal hosts [Anderson and Martone, 2014]. Thus, future experiments may bedone to test, for example, whether highly branched seaweeds are at greater risk for epiphytecolonization due to higher recruitment of spores and larvae by microbial epibionts. This isbut one example of how morphology and microbial settlement may interact to affect largerecological relationships.ConclusionWe propose that morphology of seaweed may modulate, limit, or encourage the assembly ofcertain members of the seaweed microbiota due to the interaction between morphology andwater flow. Our findings suggest that initial settlement differs between finely branched andbladed or crustose morphologies, and we hypothesize that these differences are a result ofdifferential rates of settlement, sheer forces, and transfer of nutrients to the seaweed surface22Morphology and flow. At latter time points however, other factors (like the presence of a substrate specialist)tend to swamp any signal from morphology. Thus, morphology is likely one of many factorsinvolved in determining final micobiral community composition on real seaweeds.There are many organisms, microbial and macrobial, that live epiphytically on seaweeds.Most of these organisms rely on a spore dispersal stage at some point in their reproductivecycle, and the dispersal of these spores usually operate at the microbial scale. Since micro-bial and non-microbial epiphytes have the potential to be both symbiotic and detrimental toseaweeds [Egan et al., 2013], knowledge about how flow, settlement, and morphology changeinteractions with potential epiphytes will be crucial to understanding community-level dy-namics on seaweeds.23Morphology and flow2.5 Supplementary Figures and TablesA. Reed Point B. Hakail l l l1020304050PD_whole_treeTimeRichness (PD_whole_tree)20m 1h 6h 12h***** Finely BranchedBladedCrustose15202530354045PD_whole_treeTimeRichness (PD_whole_tree)20m 1h 3h 6h 12h24h*****Finely BranchedBladedCrustosel l l l2004006008001000observed_otusTimeRichness (observed_otus)20m 1h 6h 12h***** Finely BranchedBladedCrustosell l l l l200300400500600700800observed_otusTimeRichness (observed_otus)20m 1h 3h 6h 12h24h*****Finely BranchedBladedCrustoseFigure 2.7: S: Richness of biofilms on AM units through time. Richness (PD_whole_treeand observed_otus metrics) of AM biofilms through time at (A) Reed Point and (B) Hakai. Sig-nificance is indicated by stars: * = p < 0.05; ** = p < 0.01; *** = p < 0.001. Error bars are+/-1SD.24Morphology and flowA. Reed Point0.350.450.55Dispersion of morphologies across timeDistance (Unweighted Unifrac)20 min 1 h3 h6 h12 h24 hStress: 0.11FB:CRFB:BLCR:BLlll lllllln lllllllln*ll llllllln*lll llllln**lllllllln**llllllllln lllCRBLFBB. Hakai0.450.550.65Dispersion of morphologies across timeDistance (Unweighted Unifrac)20 min 1 h6 h12 h 4 dStress: 0.16FB:CRFB:BLCR:BLlllllllllllllllllllllllllll*llll llllllllllllllllllllll**l lllllllllllllllllllllllll***l ll l lllll lllllllllllllll**llllllllllllllll ll lllllllllll**Stress: 0.18lllCRBLFBFigure 2.8: S: Dispersion and community composition of AM morphologies throughtime. Pairwise dispersions were calculated by averaging distances from a un-weighted Unifracdistance matrix for each pair of morphologies. Error bars show +/-1SD. Below each time point isa corresponding NMDS plot of samples from that time point, coloured by morphology.25Morphology and flowA. Reed Point0.10.30.5Dispersion of morphologies across timeDistance (Weighted Unifrac)20 min 1 h3 h6 h12 h24 hStress: 0.09FB:CRFB:BLCR:BLllllllllln lllllllln llllllll ln**lllllllln*lllllllln*ll llllllln lllCRBLFBB. Hakai0.00.10.20.30.4Dispersion of morphologies across timeDistance (Weighted Unifrac)20 min 1 h6 h12 h 4 dStress: 0.14FB:CRFB:BLCR:BLllllllllllllll llll lllllll llll ll llllll lll lllllllll**l lllllllllllllllllllll lll***l lll llllll lllllllll*ll llllll lllllllllllllll Stress: 0.04lllCRBLFBFigure 2.9: S: Dispersion and community composition of AM morphologies throughtime. Pairwise dispersions were calculated by averaging distances from a weighted Unifrac dis-tance matrix for each pair of morphologies. Error bars show +/-1SD. Below each time point is acorresponding NMDS plot of samples from that time point, coloured by morphology.26Morphology and flowA. Reed Point B. Hakai0.20.30.40.50.6Dispersion of morphologies across timeTimeDistance (Bray−Curtis)20 min 1 h3 h6 h12 h 1 dFBBLCR0.10.20.30.40.50.6Dispersion of morphologies across timeTimeDistance (Bray−Curtis)20 min 1 h6 h12 h 4 dFBBLCR0.050.100.150.200.250.300.350.40Dispersion of morphologies across timeTimeDistance (Weighted Unifrac)20 min 1 h3 h6 h12 h 1 dFBBLCR0.000.050.100.150.200.250.30Dispersion of morphologies across timeTimeDistance (Weighted Unifrac)20 min 1 h6 h12 h 4 dFBBLCR0.250.300.350.40Dispersion of morphologies across timeTimeDistance (Unweighted Unifrac)20 min 1 h3 h6 h12 h 1 dFBBLCR0.350.400.45Dispersion of morphologies across timeTimeDistance (Unweighted Unifrac)20 min 1 h6 h12 h 4 dFBBLCRFigure 2.10: S: Overall dispersion of AM units over time. Dispersion was calculated using“betadisper” in the “vegan” package in R. Overall and time-separated PERMDISP calculationscan be found in Table S2.4. Row (A) shows Reed Point; row (B) shows Hakai. Metrics used wereBray-Curtis (top), unweighted Unifrac (middle), and weighted Unifrac (bottom).27Morphology and flowA. Reed Point B. Hakai20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 d0 0.4 0.8ValueColor KeyllllFinely branchedBladedCrustoseWater20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d6 h4 d0 0.4 0.8ValueColor KeyllllFinely branchedBladedCrustoseWaterFigure 2.11: S: Heatmap showing “core” OTUs. A core OTU must be present in >90% ofsamples in treatment and also observed at least once above 0.1% relative abundance. Core OTUscorrespond with OTUs that are enriched in each morphology and time point. (Fig. 2.4).A. Reed Point B. Hakai20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 d0 0.4 0.8ValueColor KeyllllFinely branchedBladedCrustoseWater20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d6 h4 d0 0.4 0.8ValueColor KeyllllFinely branchedBladedCrustoseWaterFigure 2.12: S: Heatmap showing presence or absence of “core” OTUs. At Hakai (B),nearly all OTUs are present across all time points. Conversely, we see that core OTUs turn overthrough time at Reed Point (A).28Morphology and flowA Bllllll2004006008001000Diversity (chao1)Time20 m 1 h 3 h 6 h 12 h 1 dp=2.4e−07 p=1.6e−05 p=0.0068 p=0.0083 p=0.76 p=0.025 (t=11.17, df=10.99)  (t=7.26, df=10.98)  (t=3.45, df=9.35)  (t=3.37, df=8.97)  (t=−0.32, df=7.57)  (t=−2.68, df=8.86)lllll200400600800100012001400Diversity (chao1)Time20 m 1 h 6 h 12 h 4 dp=0.55 p=0.78 p=0.049 p=0.003 p=0.00067 (t=0.62, df=12.45)  (t=−0.29, df=15.33)  (t=−2.12, df=17.91)  (t=−3.71, df=12)  (t=6.93, df=5.45)lll lll15202530354045Diversity (PD_whole_tree)Time20 m 1 h 3 h 6 h 12 h 1 dp=0.00098 p=0.8 p=0.23 p=0.076 p=0.0041 p=0.00027 (t=4.46, df=10.91)  (t=0.26, df=10.28)  (t=−1.29, df=9.23)  (t=−2.01, df=9.03)  (t=−4.06, df=7.59)  (t=−5.8, df=8.93)lllll1020304050Diversity (PD_whole_tree)Time20 m 1 h 6 h 12 h 4 dp=0.00088 p=7.7e−06 p=4.9e−09 p=7.1e−11 p=8.7e−08 (t=−3.65, df=33.63)  (t=−5.38, df=30.47)  (t=−7.91, df=32.2)  (t=−11.06, df=23.85)  (t=19.51, df=7.63)lll lll200300400500600700800Diversity (observed_otus)Time20 m 1 h 3 h 6 h 12 h 1 dp=3e−06 p=0.014 p=0.44 p=0.5 p=0.045 p=0.0013 (t=9.32, df=10.03)  (t=3.05, df=9.03)  (t=0.81, df=8.45)  (t=0.71, df=7.82)  (t=−2.42, df=7.15)  (t=−4.78, df=8.26)lllll2004006008001000Diversity (observed_otus)Time20 m 1 h 6 h 12 h 4 dp=0.0061 p=0.00053 p=3.1e−07 p=4.2e−10 p=2.5e−05 (t=−2.92, df=33.71)  (t=−3.87, df=30.95)  (t=−6.39, df=33)  (t=−9.5, df=27)  (t=11.27, df=6.15)Figure 2.13: S: Comparison of water diversity and AM biofilm diversity Blue line repre-sents average richness of water across both sampling time points. Black points with +/-1SD errorbars represent richness of all AM biofilms at that time point. Significance values and parameters(Welch’s t-Test) are listed below each time point.29Morphology and flow2004006008001000120014001600Diversity (chao1) of Hakai vs Reed PointTimeDiversity (chao1)20 m 1 h3 h6 h12 h 1 d4 dlllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll lllllllllllllllllllllllllllllllllllllllllllllllllHakaiReed Point1020304050Diversity (PD_whole_tree) of Hakai vs Reed PointTimeDiversity (PD_whole_tree)20 m 1 h3 h6 h12 h 1 d4 dl lllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll lllllllll lllllllllllllllllll llll lllHakaiReed Point20040060080010001200Diversity (observed_otus) of Hakai vs Reed PointTimeDiversity (observed_otus)20 m 1 h3 h6 h12 h 1 d4 dllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll lllllllllllllllllllll ll llllllHakaiReed PointFigure 2.14: S: Comparison of AM biofilm diversity at Reed Point and Hakai. Blue dotsrepresent Reed Point, whereas red dots represent Hakai. Diversity is significantly different betweenthe sites at all time points (Refer to Table 2.6).30Morphology and flowRelative Abundance20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 h3 h6 h12 h 1 d20 m 1 dGammaproteobacteria−−Vibrio_Vibrio_sp.Alphaproteobacteria−−uncultured_UnassignedGammaproteobacteria−−uncultured_uncultured_gamma_proteobacteriumUnassigned−−Unassigned_UnassignedCytophagia−−Cytophaga_uncultured_bacteriumUnassigned−−Unassigned_UnassignedGammaproteobacteria−−Klebsiella_UnassignedDeinococci−−Deinococcus_uncultured_bacteriumVerrucomicrobiae−−Roseibacillus_uncultured_bacteriumFlavobacteriia−−Formosa_UnassignedAlphaproteobacteria−−Sphingomonas_UnassignedUnassigned−−Unassigned_UnassignedFlavobacteriia−−NS4_marine_group_UnassignedCytophagia−−Ekhidna_uncultured_bacteriumFlavobacteriia−−NS4_marine_group_uncultured_Bacteroidetes_bacteriumGammaproteobacteria−−JL−ETNP−Y6_uncultured_bacteriumThermoplasmata−−Marine_Group_II_uncultured_archaeonFlavobacteriia−−uncultured_uncultured_marine_bacteriumFlavobacteriia−−uncultured_uncultured_Flavobacteriia_bacteriumGammaproteobacteria−−Oceanospirillaceae_UnassignedGammaproteobacteria−−SS1−B−06−26_UnassignedFlavobacteriia−−NS5_marine_group_uncultured_bacteriumFlavobacteriia−−Polaribacter_3_uncultured_bacteriumSphingobacteriia−−uncultured_uncultured_Sphingobacteriia_bacteriumGammaproteobacteria−−Vibrio_UnassignedCyanobacteria−−Synechococcus_uncultured_bacteriumGammaproteobacteria−−Pseudomonas_UnassignedAlphaproteobacteria−−Surface_1_uncultured_bacteriumGammaproteobacteria−−uncultured_uncultured_bacteriumFlavobacteriia−−NS9_marine_group_uncultured_bacteriumAcidimicrobiia−−Candidatus_Actinomarina_UnassignedAlphaproteobacteria−−Acidiphilium_uncultured_bacteriumFlavobacteriia−−uncultured_UnassignedCyanobacteria−−Synechococcus_UnassignedGammaproteobacteria−−Luminiphilus_uncultured_marine_bacteriumGammaproteobacteria−−Pseudomonas_uncultured_Pseudomonas_sp.Cyanobacteria−−Synechococcus_uncultured_bacteriumAlphaproteobacteria−−Ascidiaceihabitans_uncultured_alpha_proteobacteriumFlavobacteriia−−Tenacibaculum_Tenacibaculum_lutimarisBetaproteobacteria−−Diaphorobacter_uncultured_bacteriumFlavobacteriia−−NS5_marine_group_uncultured_bacteriumGammaproteobacteria−−Alteromonas_UnassignedGammaproteobacteria−−OM60(NOR5)_clade_UnassignedAlphaproteobacteria−−SAR116_clade_uncultured_bacteriumAlphaproteobacteria−−Acidiphilium_uncultured_bacteriumFlavobacteriia−−Ulvibacter_UnassignedAlphaproteobacteria−−Amylibacter_uncultured_bacteriumGammaproteobacteria−−Alteromonas_UnassignedGammaproteobacteria−−Vibrio_UnassignedGammaproteobacteria−−Pseudomonas_UnassignedGammaproteobacteria−−SAR86_clade_uncultured_marine_bacteriumAlphaproteobacteria−−Surface_1_UnassignedAlphaproteobacteria−−Planktomarina_uncultured_marine_bacteriumGammaproteobacteria−−Pseudomonas_UnassignedUnassigned−−Unassigned_UnassignedGammaproteobacteria−−Oceanospirillaceae_UnassignedGammaproteobacteria−−Pseudoalteromonas_UnassignedGammaproteobacteria−−Oleispira_UnassignedRelative Abundance20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d20 m 1 h6 h12 h 4 d6 h4 dFigure 2.15: S: Taxa summaries plot at the OTU level. OTUs observed above 3% relativeabundance in at least one sample are coloured; the rest are blocked grey.BL FB0.10.20.30.40.5Dye lost after submersion in still waterMorphologyPercent dye lostBL FB0.00.10.20.30.40.5Dye lost after submersion in moving waterMorphologyPercent dye lostFigure 2.16: S: Dye Dip Experiment Results There were no significant differences betweenthe amount of dye lost between morphologies in either still (“Still water” p = 0.51, t9.9,x = 0.68)or moving water (Welch’s t-Test “Moving water” p = 0.39, t11,x = 0.89).31Morphologyandflowp-Values for PERMDISPSite 20 minutes 1 hour 3 hours 6 hours 12 hours 1 day 4 days OverallReed Point Bray-Curtis 0.77 (f=0.27,df=2) 0.61 (f=0.56,df=2) 0.31 (f=1.45,df=2) 0.74 (f=0.32,df=2) 0.18 (f=2.46,df=2) 0.17 (f=2.37,df=2) - 0.27 (f=1.34,df=2)Weighted Unifrac 0.59 (f=0.58,df=2) 0.77 (f=0.28,df=2) 0.63 (f=0.5,df=2) 0.74 (f=0.32,df=2) 0.26 (f=1.81,df=2) 0.13 (f=2.95,df=2) - 0.21 (f=1.61,df=2)Un-weighted Unifrac 0.9 (f=0.11,df=2) 0.76 (f=0.3,df=2) 0.84 (f=0.18,df=2) 0.86 (f=0.15,df=2) 0.11 (f=3.49,df=2) 0.1 (f=3.42,df=2) - 1 (f=0,df=2)Hakai Bray-Curtis 0.092 (f=2.64,df=2) 0.0012 (f=8.97,df=2) - 0.037 (f=3.74,df=2) 0.52 (f=0.68,df=2) - 0.43 (f=0.86,df=2) 0.045 (f=3.18,df=2,133)Weighted Unifrac 0.39 (f=0.99,df=2) 0.02 (f=4.61,df=2) - 0.95 (f=0.05,df=2) 0.46 (f=0.81,df=2) - 0.34 (f=1.13,df=2) 0.44 (f=0.83,df=2)Un-weighted Unifrac 0.015 (f=5.01,df=2) 0.0029 (f=7.5,df=2) - 0.1 (f=2.48,df=2) 0.22 (f=1.64,df=2) - 0.74 (f=0.31,df=2) 0.37 (f=1,df=2)Table 2.4: S: PERMDISP of morphologies at each time point and overall. PERMDISP was calculated at each timepoint to confirm PERMANOVA results. Reed Point AM unit morphologies do not have different dispersions at any time point,but Hakai samples have different dispersions at the 1 hour and 6 hour time points.p-Values for ANOVASite 20 minutes 1 hour 3 hours 6 hours 12 hours 1 day 4 days OverallReed Point Chao1 0.039 (f=5.86,df=2,6) 0.0071 (f=12.59,df=2,6) 0.048 (f=5.23,df=2,6) 0.43 (f=0.99,df=2,5) 0.021 (f=9.29,df=2,5) 0.15 (f=2.6,df=2,6) - 0.0033 (f=6.48,df=2,46)PD_whole_tree 0.064 (f=4.5,df=2,6) 0.0035 (f=16.68,df=2,6) 0.018 (f=8.45,df=2,6) 0.25 (f=1.88,df=2,5) 0.04 (f=6.59,df=2,5) 0.15 (f=2.67,df=2,6) - 0.0016 (f=7.39,df=2,46)Observed_OTUs - - - - - - 1.6e-05 (f=11.99,df=2,133) 0.0079 (f=5.38,df=2,46)Hakai Chao1 0.16 (f=1.93,df=2,27) 0.003 (f=7.45,df=2,24) - 0.00056 (f=10.12,df=2,26) 0.1 (f=2.57,df=2,21) - 0.67 (f=0.4,df=2,26) 3e-07 (f=16.86,df=2,133)PD_whole_tree 0.45 (f=0.81,df=2,27) 0.0066 (f=6.23,df=2,24) - 8e-04 (f=9.51,df=2,26) 0.024 (f=4.5,df=2,21) - 0.85 (f=0.16,df=2,26) 1.6e-05 (f=11.99,df=2,133)Observed_OTUs - - - - - - 1.6e-05 (f=11.99,df=2,133) 1.8e-06 (f=14.66,df=2,133)Table 2.5: S: ANOVA of richness across morphology by time point.32MorphologyandflowMetric 20 min 1 hour 6 hours 12 hours Overallp-value <0.001 <0.001 <0.001 0.002 <0.001Chao1 t-statistic 6.562 5.103 5.445 3.931 4.316Df 36.997 33.196 34.521 13.516 150.616p-value <0.001 <0.001 <0.001 0.198 0.376PD_whole_tree t-statistic 6.342 4.944 5.396 1.384 0.888Df 36.065 27.887 25.663 9.386 148.219p-value <0.001 <0.001 <0.001 0.009 <0.001Observed_otus t-statistic 6.577 5.102 6.35 3.107 3.785Df 35.543 33.899 34.781 11.547 150.96Table 2.6: S: Welch’s t-Test comparison of richnessat Reed Point and Hakai by time point.33Horizontal transmission of microbesChapter 3Horizontal transmission of microbes be-tween neighbouring macroalgae3.1 IntroductionMacroalgae (seaweeds) have an intimate relationship with their microbial symbionts. Somemicrobes provide benefits for their seaweed hosts by improving nutrient acquisition [Rosen-berg and Paerl, 1981, Croft et al., 2005, Ilead and Carpenter, 1975, Chisholm et al., 1996],promoting settlement and growth [Joint et al., 2002], and priming immune responses againstpotential pathogens [Küpper et al., 2002,Maximilian et al., 1998,Steinberg et al., 1997,Wein-berger, 2007,Armstrong et al., 2001,Dobretsov and Qian, 2002]. Other microbes, however,cause tissue bleaching [Case et al., 2011,Zozaya-Valdes et al., 2015] and initiate or exasperatetissue degradation [Küpper et al., 2002, Egan et al., 2013]. Since microbes influence manyaspects of seaweed biology, it is important to understand how the assemby of the macroalgalmicrobiota occurs.Macroalgae live in a rich microbial “soup” within the ocean and constantly contact a varietyof microbes. A subset of these microbes are capable of colonizing macroalgal surfaces, andseeds the assembly of seaweed microbiota. Composition of the seaweed microbiota are gen-erally species specific [Bondoso et al., 2014,Hollants et al., 2011,Lachnit et al., 2011,Staufen-berger et al., 2008] because they are modulated and regulated, both specifically and generally,through a variety of macroalgal exudates. Polysaccharides (alginate, carageenan, cellulose,etc), which compose the bulk of macroalgae biomass, are a rich source of energy and carbon,and can promote epibiont settlement and growth [Steinberg, 2002,Lachnit et al., 2011]. Con-versely, some metabolites, such as hydrogen peroxide [Küpper et al., 2002] and antibacterialfuranones [Maximilian et al., 1998], are inhibitory toward microbial settlement and growth.The combination and proportion of exudates found on seaweed surfaces impose selection oncolonizing microbes, and result in diverse microbial assemblages across seaweed species.34Horizontal transmission of microbesMacroalgae modify the surrounding water column and nearby biofilms as a result of theirexudates. Organic exudates from macroalgae increase microbial biomass and alter bothmicrobial functional profiles and community structure in the surrounding water column[Clasen and Shurin, 2014, Krumhansl and Scheibling, 2012,Miller and Page, 2012, Newellet al., 1980,Wada and Hama, 2013] and on nearby biofilms [Fischer et al., 2014,Vega Thurberet al., 2012,Zaneveld et al., 2016]. Macroalgal exudates can also alter microbial communitycomposition by inhibiting growth of bacterial lineages [Lam and Harder, 2007, Lam et al.,2008]. Therefore, we know that macroalgae alter the microbial water column communityand nearby biofilm composition in at least two ways: by contributing to the organic carbonpool in the water column, and by inhibiting microbial growth through chemical exudates.We hypothesize that macroalga affect the composition of microbial communities on the sur-faces of other macroalgae. In dense communities of seaweeds, individuals may collectivelycontribute to a shared “plume” of exudates that both reduces potential pathogens and en-riches helpful symbionts. This may in turn alter composition of epibiotic communities. Forexample, biofilms contain higher proportions of common seaweed-associated bacteria whenin close proximity to seaweeds [Fischer et al., 2014]. Therefore, the prevalence of microbesconsistently found in association with seaweeds (the ‘core’ microbiota) may be greater indense communities of macroalgae, and reinforcement of these microbes may positively influ-ence host health as the core often consists of beneficial symbionts [Shade and Handelsman,2011]. In contrast, potential pathogens may be more effectively eliminated from the watercolumn in dense communities of seaweeds. Although population density is hypothesized tocorrelate with disease transmittance at the population level, it can simultaneously increaseimmunity at the individual level [Hawley and Altizer, 2011]. Therefore, net rate of infectiondecreases if the benefits of improved immunity outweigh the increased exposure risk. Increas-ing canopy cover is known to correlate with greater shifts in nearby microbiota [Zaneveldet al., 2016], and many seaweeds are known to release a variety of antimicrobial compoundsinto the water [Lam et al., 2008, Egan et al., 2013]. Thus, we hypothesize that individualmacroalga living in large communities of seaweeds may experience lower rates of colonizationby potentially pathogenic microbes.We test the hypothesis that the microbiota of macroalgae changes when living in close prox-imity to other macroalgae and assess whether shifts in microbial communities on macroal-gae are associated with changes in growth rate. Our study investigates the influence ofmacro-ecological communities on micro-ecological assemblages, and emphasizes the connec-tion between host and symbiont ecology. The factors that govern patterns in macro- andmicro-ecology are linked, we hope to improve our understanding of ecosystem dynamics by35Horizontal transmission of microbesapplying ecological principles at a broader scale.3.2 MethodsSampling methodsSamples of macroalgae were collected on September 6th 2016 from Brockton Point, Vancou-ver, British Columbia from the intertidal at low tide. Blades from individual Nereocystisleutkeana and Mastocarpus sp. thalli were brought back to UBC in a cooler lined with wetpaper towels (species were separated), and then transferred to overnight holding tanks withsalinities of 30ppt and temperatures maintained at 160C. The Nereocystis and Mastocarpuswere kept in separate tanks. The next day, all samples were distributed into experimentaltanks.Environmental samples were taken a few days later, on September 13th. Again, severalsamples of Nereocystis and Mastocarpus were gathered from the intertidal; all from separateplants. Five Nereocystis blades, all from separate plants, were swabbed at two locationseach: meristem (10cm from blade base) and mature blade (50cm from blade base). Theblade surface was rinsed with sterile artificial seawater (ASW, always 30ppt unless notedotherwise) for 10 seconds, and then swabbed with a sterile cotton swab (Puritan- Item#:CA10805-154) for 10 seconds. The cotton swab tip was then snapped off into 2mL cryotubes(VWR- Item#: 10018-760) and kept on ice until return to the lab. Five Mastocarpus bladeswere also swabbed using the same method. Environmental (wild) samples were comparedto swabs from experimental seaweeds to test whether lab incubation significantly affectedmicrobial community composition and diversity on seaweed surfaces.Macroalgae–Water ExperimentIn the first experiment, referred to hereafter as the “Macroalgae–Water (M–W) experiment”,we assessed the degree to which microbes are transferred from seaweed to the surroundingwater column by incubating Nereocystis and Mastocarpus alone in seawater for 6 days (seeFig. 3.1A for experimental design). Ten 10L tanks were placed in a 2-layer water table. Thetemperature of all tanks was regulated by the water table and kept at 160C. Lights werekept on for 24h a day. Additionally, a bubbler was placed in each of the tanks and set tothe maximum setting. Five tanks contained only Nereocystis and the other five tanks onlyMastocarpus (Fig. 3.1A). Seaweeds were incubated in tanks for six days. On the sixth day,we sampled one random seaweed individual and took one 500mL water sample from each36Horizontal transmission of microbesA. M–W B. M–W–NMFFigure 3.1: Experimental design for M–W and M–W–NMF experiments. (A) In theM–W experiment, either Mastocarpus (n = 5) or Nereocystis (n = 5) was incubated alone in 10Ltanks for 6 days. Water was incubated at 160C and lights were kept on 24h a day. Each tank hadone bubbler on at the maximum setting. Both water (dark blue arrows) and macroalgal surfaces(green arrows) were sampled. (B) In the M–W–NMF experiment, NMF fragments were incubatedwith either Nereocystis (n = 5, middle right), Mastocarpus (n = 5, bottom left), or both (n = 5,bottom right). A NMF alone control (n = 5, middle left) and a water only control (n = 5, top)were also included. The experiment lasted 5 days. Water temperature was maintained at 160C andlight was on 24h a day. Water was sampled from every tank (dark blue arrows), and NMF surfaceswere sampled where applicable (green arrows).tank. Individuals were rinsed with ASW for 10 seconds and then swabbed for 10 secondswith a sterile cotton swab (Puritan- Item#: CA10805-154). Swabs were immediately frozenat -200C in 2mL cryotubes (VWR- Item#: 10018-760). Water samples were collected insterile 500mL PPE bottles, pre-filtered with an acid-sterilized 150um sieve, and then pumpedthrough sterile 0.22 membranes (Durapore- Item#: GVWP04700) with a peristaltic pump(Cole-Parmer- Item#: RK-77913-70) at approximately 180rpm (level 30) to collect microbialbiomass. The tubing was rinsed with 500mL of 2% HCl, followed by a rinse with 1500mLdeionized water between replicates.Macroalgae–Water–NMF experimentWe conducted a second experiment (the Macroalgae–Water–NMF (M–W–NMF) experiment)to determine how the presence of seaweed influences the surface microbial community ofneighbouring seaweeds (see Fig. 3.1 for experimental design). Specifically, our goal was37Horizontal transmission of microbesto test whether Nereocystis meristems fragments (NMFs) co-incubated with other seaweedhad different epibiotic microbial communities than meristem fragments incubated alone.Additionally, we wanted to test whether shifts in the epibiotic community structure of NMFscorresponded to any changes in growth rate. We chose to use meristematic Nereocystisfragments for two reasons. First, Nereocystis can grow up to 14cm per day [Kain, 1987],which maximizes the potential effect size for differential growth rates between treatments.Additionally, Nereocystis growth is concentrated in the meristematic region (the first 10cmof the blade or so), and areas of new growth tend to have less microbial diversity [Bengtssonet al., 2012]. Thus, the surfaces of Nereocystis meristems are optimal areas to test formeaningful shifts in epibiotic community structure because they are highly selective.In the second experiment, twenty-five tanks with 5L of 30ppt water each were incubatedin a water table held at 160C. A bubbler was placed in each tank and lights were left on24h a day. Salinity and temperature were monitored through the experiment to ensure theywere constant and uniform. Each tank contained either: (1) water only, (2) water with oneNMF fragment, (3) water with one NMF fragment and approximately 50g (wet weight) ofNereocystis blades, (4) water with one NMF fragment and approximately 50g (wet weight)of Mastocarpus blades, or (5) water with one NMF fragment and approximately 50g (wetweight) combined of Nereocystis and Mastocarpus blades. All treatments were incubatedfor five days. Dissolved oxygen, pH, temperature, and salinity were also all measured atthe beginning and end of the experiment using the Orion STAR A329 (ThermoScientific,Item#-STARA3295) and a standard refractometer.NMFs were prepared by cutting 10-cm fragments of Nereocystis meristem from the base ofeach blade with scissors. Each fragment’s length and width were measured using a measuringtape to the nearest half millimetre, and its wet weight determined by blotting twice on apaper towel and weighing it on a scale to 2 decimal places. NMFs and other algal tissue werekept physically separated by coarse plastic mesh.At the end of the incubation period, all NMFs were sampled and 500mL of water from eachtank filtered using the same methods as described for the M–W experiment.Library prepThe 96-well MoBio PowerSoil DNA extraction kit was used to extract DNA from both thewater filters and AM swabs. Filters and swabs were transferred to the extraction kit usingtweezers sterilized with 2% HCl, then ethanol and flame. Extractions followed the MoBioPowersoil DNA extraction protocol with the following modifications based on recommen-38Horizontal transmission of microbesdations in the Earth Microbiome protocol (http://www.earthmicrobiome.org/). First, a10-minute incubation at 650C in a water bath was added after the addition of C1 and beforeshaking in the shaker; next, the plates were shaken at 20 shakes per second for 20 minutestotal instead of using the recommended shaker in the manual; and lastly filters were allowedto soak in C6 at room temperature for 10 minutes prior to elution into the final elution plate.The DNA elution was stored at -200C.The 16S small subunit ribosomal RNA marker gene was sequenced to profile bacteria and ar-chaea. The amplicon library prep was done in-lab using the following 16s primers: barcoded515 forward primers (5’-GTGYCAGCMGCCGCGGTAA-3’) and 806 reverse primers (5’-GGACTACHVGGGTWTCTAAT-3’). Primers were used at final concentrations of 0.5uMwith 4uL of DNA extract. DNA extracts were amplified in 20uL reactions using PhusionFlash High-fidelity proofreading Mastermix (Thermofisher- Item#: F548L). Reactions un-derwent the following thermocycler settings: 980C for 10 seconds; 25 cycles of 980C (1s),500C (5s), 720C (24s); and a final extension phase of 720C for 1 minute. Lastly, the suc-cessful PCR products were quantified using Pico-green (Thermofisher- Item#: P11496) andpooled at 45ng/sample. The pooled samples were then sent to the Centre for ComparativeGenomics and Evolution Bioinformatics (CGEB) at Dalhousie University for sequencing onthe Illumina MiSeq platform with 2x300bp chemistry.Sequence ProcessingRaw samples were demultiplexed with split_libraries_fastq.py in QIIME version 1.9 (Quan-titative Insights into Microbial Ecology; (QIIME) [Caporaso et al., 2010]) , yielding 3,688,981reads. Sequences were trimmed (fastx_trimmer), clipped (fastx_clipper), and filtered (fastq_quality_filter)using the Fastx Toolkit (Hannon Lab) to 250bp with a minimum quality threshold of Q19.The remaining 3,661,707 raw sequences were processed into operational taxonomic units(OTUs) using Minimum Entropy Decomposition (MED; [Eren et al., 2014]) with the min-imum substantive abundance (-m) parameter set to 100, yielding 1,363 unique OTUs and3,050,864 reads. The resulting OTU matrix was transcribed into a QIIME-compatible for-mat. Taxonomy was then assigned to the representative sequence for each MED node,hereafter referred to as OTUs, by matching it to the SILVA 128 database clustered at 99%similarity with assign_taxonomy.py in QIIME using uclust V1.2.22q [Edgar, 2010]. Chloro-plast, mitochondrial, and eukaryotic DNA were filtered out and five reads suspected of beingcontaminants were removed. The identity of those five reads were one uncultured Rubritalea(Verrumicrobia), oneMarivita (Alphaproteobacteria), a Sulfitobacter (Alphaproteobacteria),one Pseudomonas (Gammaproteobacteria), and an uncultured Plantomycete. Lastly, sam-39Horizontal transmission of microbesples with less than 1000 reads per sample were removed. The final OTU table consisted of1,314 unique sequences and 2,302,993 reads, with a mean of 26,471 reads per sample. Foralpha and beta diversity analysis, samples were rarefied to 1000 reads per sample. Represen-tative sequences were aligned with PyNAST in QIIME and a phylogenetic tree was createdusing FastTree [Price et al., 2009] in QIIME with the make_phylogeny.py script.Community dissimilarityTo compare community composition across treatments, distance matrices were created inbeta_diversity.py (QIIME) with the rarefied OTU table using three metrics: weightedUnifrac, unweighted Unifrac [Lozupone and Knight, 2005], and Bray-Curtis [Bray and Cur-tis, 1957]. Matrices were imported into R and the “isomds” command from the “MASS”package [Venables and Ripley, 2002] was used to created 2-dimensional NMDS plots. Poly-gons were drawn around treatments using “chull” in the “grDevices” package [R Core Team,2016]. Pairwise permutational anovas were calculated across treatments using “adonis” fromthe “vegan” package [Oksanen et al., 2017] and p-values were adjusted for multiple com-parisons using the Benjamini-Hochberg method (also known as the “False Discovery Rate(FDR)” method) [Benjamini et al., 1995] with “p.adjust” in the “stats” package [R CoreTeam, 2016] . We also tested for differences in dispersion between groups using “betadisper”in the “vegan” packaged [Oksanen et al., 2017]. We chose to show only Bray-Curtis resultsbecause the metric accounted for both abundance and membership in microbial communities,but results were consistent across all three metrics.Alpha diversityAlpha diversity for each treatment was calculated in QIIME using the alpha_diversitypipeline. The metrics Chao1 [Chao, 1984], PD_whole_tree [Faith and Baker, 2006], andObserved_otus were used, but since results were similar between the three, only Chao1is shown in the results. Chao1 was chosen because it estimates community richness whilecorrecting for rare taxa. In contrast, PD_whole_tree calculates diversity based on phyloge-netic distance of the sample. Pairwise comparisons between treatments was calculated using“t.test” in the “stats” package [R Core Team, 2016] with the method “Welch’s t-Test” andp-value adjustments for multiple comparisons was done using the Benjamini-Hochberg (aka“False Discovery Rate”) method [Benjamini et al., 1995] with the “p.adjust” command inthe “stats” package [R Core Team, 2016]. Tables were initially created using “xtable” in thepackage “xtable” [Dahl, 2016] and then edited manually in LaTex.OTU enrichment and Taxa summaries40Horizontal transmission of microbesFold-change enrichment and reduction of genera were calculated using “DESeq2” in the Rpackage “DESeq2” [Love et al., 2014] with the “Wald” test. First, the OTU table (unrarefied)was collapsed at level 6 (Genera) using summarize_taxa_through_plots.py (QIIME). Then,genera with less than 100 counts per sample were removed. For water samples, all treat-ments (NMF only, Nereocystis alone, Mastocarpus alone, and Nereocystis + Mastocarpus)were compared to the NMF alone control separately. Additionally, all treatments for NMF-surface samples were compared to the NMF-alone control separately. All genera that weresignificantly enriched or reduced (p < 0.05, after p-value adjustment (Benjamini-Hochbergmethod (FDR)) and were observed at abundances greater than 3% at least twice in eachexperiment were kept. Enrichment/reduction results were plotted using “heatmap.2” in the“gplots” package [Warnes et al., 2016].To plot taxa summary plots, OTU tables collapsed by genera were separated into fourexperimental groups: M–W–NMF water samples, M–W–NMF NMF surface samples, M–Wsamples, and environmental samples. Within each experimental group, genera at less than3% relative abundance and with no significant enrichment in the NMF-incubation experimentare depicted as grey bars. The remaining genera are plotted in colour in Figure 3.6.3.3 ResultsSequence processing yielded a total of 2,302,993 reads between 1,314 OTUs. There werea total of 18 samples from the M-W experiment retained (9 water samples and 9 seaweedswabs); 42 samples from the M-W-NMF experiment retained (24 water samples and 18 NMFswabs); and 15 wild seaweed swabs (5Mastocarpus and 10 Nereocystis swabs) retained. Readsper sample ranged from 1,007 to 72,849 (with an average of 26,471).Comparison of experimental and in situ seaweedsFirst, we compared surface communities from lab-incubated seaweeds with wild seaweedssampled in situ to determine whether lab incubation of samples significantly impacted theirmicrobial community composition. We found that seaweed surface communities cluster byseaweed identity (species) across all in situ and laboratory samples (Fig. 3.2a), regardless oftreatment type (PERMANOVA Nereo vs Mast vs Water p = 0.001, R2 = 0.248, df = 2,74;PERMDISP p < 0.001, F2,74= 16.30). Additionally, richness of microbial communities wassimilar between all seaweed of the same species (Fig. 3.2b). Mastocarpus surface communitieswere consistently and significantly more diverse than Nereocystis surface communities (Fig.3.2b, Table 3.1). Thus, it appears that community structure and diversity is highly conserved41Horizontal transmission of microbesA. NMDS of all samples B. Richness of all samplesllll lllllllllllllllllllllllllll llll lllllllllllllllllllllllllll ll lllllll lll−0.6 −0.4 −0.2 0.0 0.2 0.4 0.6−0.50.00.5NMDS plot of all samples (BC)Stress: 0.18NMDS 1NMDS 2lllllllllNereo Meristem (lab)Nereo Meristem (wild)Nereo blade (lab)Nereo blade (wild)Mast blade (lab)Mast blade (wild)Water alone (M−W−NMF)Water (M−W−NMF)Water (M−W)llllll200300400500Alpha Diversity (chao1)Nereo Mast WaterSample TypeFigure 3.2: Comparison of community composition and richness across macroalgalsurfaces and water samples. For statistical results, refer to Table 3.1. (A) NMDS of Nereocystis(green shades, n = 32), Mastocarpus (purple shades, n = 10), and water samples (blue shades, n =33) (created from a Bray-Curtis distance matrix) . For pairwise comparisons, see 3.1. Nereocystissamples are always more similar to each other than to Mastocarpus, regardless of treatment type.(B) Diversity (Chao1 metric) of Nereocystis (green shades), Mastocarpus (purple shades), and watersamples (blue shades). Mastocarpus surfaces are richer than Nereocystis surfaces. For pair wisecomparisons, see Table 3.1.within algal species. This provides a framework for interpreting our results in a broaderecological context, and emphasizes that the effects of treatments on microbial communitystructure are subtle modulations on a more general pattern of species specificity.Water column communities across treatmentsIncubation of seaweeds significantly altered both the composition and diversity of microbesin the water column in a species specific manner. Tanks with seaweed (Nereocystis, Mastocar-pus, and both) had water column communities that were significantly different in compositionfrom the control (NMF alone) (Fig. 3.3A, Table 3.2; PERMDISP p = 0.593, F2,12 = 0.546).Further, the richness of water column communities increased in treatments where seaweedwas added (Fig. 3.3B, Table 3.2). Water column communities from Nereocystis and Masto-carpus treatments differed significantly from each other in both the M–W (PERMANOVAp = 0.049, R2 = 0.266, df = 1,8, Fig. S3.11) and M–W–NMF experiments (Table 3.2) .Additionally, microbial richness was higher in water incubated with Mastocarpus (Masto-carpus and Nereocystis + Mastocarpus treatments) than without Mastocarpus (Nereocystis,NMF alone, and water only treatments) (Welch’s t-Test p < 0.001, t19.862 = 8.034. Refer to42Horizontal transmission of microbesPERMANOVA Welch’s t-Test(Community Dissimilarity) (Richness)Group 1 Group 2 p pNereo(n = 32)Mast(n = 10) 0.001 (R2=0.191, F.model1,41=9.466) 0.00192 (t17.58=–4.14)Nereo(n = 32)Water(n = 33) 0.001 (R2=0.168, F.model1,64=12.765) 0.0375 (t62.58=–2.24)Mast(n = 10)Water(n = 33) 0.001 (R2=0.242, F.model1,42=13.06) 0.0375 (t19.52=2.23)Table 3.1: Comparison of community dissimilarity (PERMANOVA) and rich-ness (Welch’s t-Test) of Nereocystis, Mastocarpus, and water samples. PER-MANOVA (with a Bray-Curtis distance matrix) and Welch’s t-Test (Chao1 richness)with Benjamini-Hochberg (FDR) adjusted p-values of Nereocystis surfaces, Mastocarpussurfaces, and water samples. All three groups are significantly different from each other.Mastocarpus surfaces are richer than Nereocystis surfaces. Water samples are more diversethan Nereocystis surfaces, but less diverse than Mastocarpus surfaces.Table 3.2 for all pairwise comparisons). This trend is consistent with the higher richness ob-served on Mastocarpus surfaces compared to Nereocystis surfaces (Fig. 3.2B). Interestingly,although richness of water column communities correlated with whether or not Mastocarpuswas present, overall community composition clustered by presence or absence of Nereocystis(see pairwise comparisons in Table 3.2). Thus, shifts in water column community composi-tion and richness with the addition of macroalgae differed depending on which species wasused.NMF surface communities across treatmentsMicrobial communities from NMF surfaces incubated alone were different than communitiesfrom NMFs incubated with any other macroalgae (Table 3.2), but treatments with differentcombinations of macroalgal co-incubates were not different from each other (Fig. 3.4A, seeTable 3.2 for PERMANOVA results, PERMDISP p = 0.04, F1,6 = 6.824). Since the p-values observed in the treatment comparisons (treatments with Nereocystis and treatmentswithout Nereocystis) was almost significant, we also examined the results from weightedand unweighted Unifrac PERMANOVA analyses: all other metrics also yielded insignificantresults (p < 0.1 for all pairwise tests). Therefore, it appears the while there was a shift incommunity composition on NMFs when a seaweed co-incubate is added, we did not detectdifferences between Nereocystis, Mastocarpus, and Nereocystis + Mastocarpus treatments.Growth of NMFs43Horizontal transmission of microbesA. NMDS of water samples B. Richness of water samplesllllllllllllllllllllllll−0.4 −0.2 0.0 0.2 0.4−0.4−0.20.00.20.40.6NMDS of water sa ples0.16NMDS 1NMDS 2 lllllWater onlyNMF Alonewith Nereowith Mastowith Nereo + MastllWater only (4)NMF Alone (5)With Nereo (5)With Mast (5)With Both (5)200250300350400450Alpha diversity across water samplesAlpha Diversity (chao1)TreatmentFigure 3.3: Comparison of water column communities in M–W–NMF treatments. Forstatistical results, refer to Table 3.2. (A) NMDS of water community composition (created froma Bray-Curtis distance matrix) from the M–W–NMF experiment. The water only control is notsignificantly different from the NMF alone treatment, indicating the addition of an NMF does notsignificantly alter the water column community. Water column communities are significantly dif-ferent between treatments (with Nereocystis, Mastocarpus and both) and the control (NMF alone).Furthermore, water column communities with Nereocystis added are different than water columncommunities with Mastocarpus added. (B) Richness (Chao1 metric) of water column communitiesacross treatments in the M–W–NMF experiment. Water column communities are richer in NMFalone treatments than water only treatments. Additionally, all treatments with macroalgae (Nereo-cystis, Mastocarpus, and both) are richer than the NMF alone treatment. Finally, treatments withMastocarpus are richer than treatments with Nereocystis. This follows the trend observed in Figure3.2, where Mastocarpus surfaces were richer than Nereocystis surfaces.We found no significant difference in growth rates of NMFs between treatments, but all NMFsgrew. NMFs grew proportionally to their original surface area: growth in length rangedfrom 0.7-3.8cm and growth in width ranged from 0-1.7cm. Growth in NMFs indicates thatmeristem fragments were alive and productive.In summary, the addition of seaweed to tanks in the lab significantly altered the microbialcommunity composition of the water column, and these shifts were different depending onwhich macroalgal species was added. Conversely, communities on NMF surfaces were onlysensitive to the presence or absence of a macroalgal co-incubate: there were no detectabledifferences in community structure between treatments with different species of macroalgalco-incubates.44Horizontal transmission of microbesA. NMDS of NMF surfaces B. Richness of NMF surfacesllllllllllllllllll−1.0 −0.5 0.0 0.5 1.0−1.0−0.50.00.51.0NMDS of Nereo Meristem Swabs0.19NMDS 1NMDS 2llllNMF Alonewith NereoWith Mastowith Nereo + MastllNMF alone (5)With Nereo (5)With Mast (3)With Both (5)150200250300350400Alpha diversity across meristem swabsAlpha Diversity (chao1)TreatmentFigure 3.4: Comparison of NMF surface communities M–W–NMF treatments. Forstatistical results, refer to Table 3.2. (A) NMDS of NMF surface communities (created froma Bray-Curtis distance matrix) from the M–W–NMF experiment. NMF surface communities aresignificantly different when co-incubated with macroalgae (Nereocystis, Mastocarpus and both) thanincubated alone (NMF alone). NMF surface communities are not significantly different betweentreatments with different macroalgae. (B) Richness (Chao1 metric) of NMF surface communities inthe M–W–NMF experiment. We found no statistical difference in richness between treatments, butthere is a trend where treatments withMastocarpus are richer than treatments withoutMastocarpus.This is consistent with trends observed in the water column (Fig. 3.3) and on algae surfaces (Fig.3.2).Taxonomic composition of communities and enrichment of select generaWe compared the taxonomic composition of communities from the water column and asso-ciated with NMF surfaces (Fig. 3.6). Additionally, we used “DESeq2” [Love et al., 2014]to identify genera enriched in each treatment relative to the control. In our comparison ofenriched genera, we only consider groups that are both significantly enriched (corrected p <0.05) and greater than 3% relative abundance in two or more samples per treatment group.Some genera were found consistently in all water samples. These genera included NS3aand Wenyingzhuangia (Flavobacteriia); Hyphonomas, Roseibacterium, and Sulfitobacter (Al-phaproteobacteria), and Pseudohongiella (Gammoproteobacteria) (representation shown ingrey in Fig. 3.6).We also found twenty differentially enriched taxa in water column samples that were simulta-neously greater than 3% relative abundance: five genera were enriched across all treatments45Horizontal transmission of microbesPERMANOVA Welch’s t-test(Community Dissimilarity) (Richness)Group 1 Group 2 Water samples NMF surface Water samples NMF surfaceWater Only NMF Alone 0.09(R2=0.162, F.model1,8=1.35)<0.001(t6.98=6.26)NereoNMF Alone Mast 0.001(R2=0, F.model1,19=3.062)0.003(R2=0, F.model1,17=2.278)<0.001(t26.441=–5.664)0.685(t38.636=–0.409)NereoMastNereo Mast 0.012(R2=0.293, F.model1,9=3.31)0.06(R2=0.216, F.model1,7=1.655)<0.001(t7,74=–10.96)0.854(t10.85=–0.69)Nereo NereoMast 0.018(R2=0.336, F.model1,9=4.046)0.293(R2=0.124, F.model1,9=1.133)<0.001(t31.07=–4.25)0.854(t5.33=–0.33)Mast NereoMast 0.078(R2=0.154, F.model1,9=1.458)0.06(R2=0.206, F.model1,7=1.559)0.00232(t27=–3.52)0.854(t8.51=0.19)Table 3.2: Comparison of community dissimilarity (PERMANOVA) and rich-ness (Welch’s t-Test) of water column and NMF surface communities inM–W–NMF treatments. PERMANOVA (with a Bray-Curtis distance matrix) andWelch’s t-Test (Chao1 richness) with Benjamini-Hochberg (FDR) adjusted (in pairwisecomparisons only) p-values of water sample comparisons. Water column communitiesare different between the control and treatments, as well as between different treatments.Conversely, NMF surface communities only different between control and treatments, andnot between different treatments. Richness of water column communities are different inall comparisons, whereas NMF surface communities are all similar in richness.and two genera were reduced in all treatments. Enriched genera included Winogradskyellaand Polaribacter_4 (Flavobacteriia); Glaciecola and Pseudoalteromonas (Gammaproteobac-teria) and an uncultured genera of Saprospiraceae (Sphingobacteriia), whereas reduced gen-era included Pseudophaeobacter and an uncultured Rhodospirillaceae (both from Alphapro-teobacteria). There were also taxa that were only enriched when water was co-incubatedwith each species of seaweed. For example, Algibacter, a Flavobacteriia isolated from greenalgae [Nedashkovskaya, 2004], was found enriched in water co-incubated with Nereocystis,but not Mastocarpus. Treatments with Nereocystis also saw a reduction in the genera OM43(Betaproteobacteria), Marivita (Alphaproteobacteria), and Alcanivorax (Gammaproteobac-teria). There were no genera enriched or reduced in both treatments with Mastocarpus,but treatments with Mastocarpus alone showed a decrease in an uncultured Simkaniaceae(Chlamydiae), Colwellia (Gammaproteobacteria), and Rubritalea (Verrumicrobiae).Communities on NMF surfaces, in contrast to the water samples, show less differentialenrichment across treatments (Fig. 3.6, 3.5). Although not shown, there is high variabilityin the relative abundance of dominant taxa. Thus, NMF surfaces are also more variablethan water column communities. Across all treatments, only five genera were significantlyenriched or reduced from the control (while at >3% relative abundance). One of these genera46Horizontal transmission of microbes+Nereo+Mast+Nereo+Mast+Nereo+Mast+Nereo+MastAlphaproteobacteria: Rhodospirillaceae_unculturedGammaproteobacteria: Oceanospirillaceae_OceanobacterVerrucomicrobiae: Verrucomicrobiaceae_PersicirhabdusAlphaproteobacteria: Rhodobacteraceae_PseudophaeobacterFlavobacteriia: Flavobacteriaceae;Other_NAVerrucomicrobiae: Rubritaleaceae_RubritaleaGammaproteobacteria: Alcanivoracaceae_AlcanivoraxAlphaproteobacteria: Rhodobacteraceae_MarivitaGammaproteobacteria: Colwelliaceae_ColwelliaChlamydiae: Simkaniaceae_unculturedBetaproteobacteria: Methylophilaceae_OM43_cladeAlphaproteobacteria: Rhodobacteraceae_LoktanellaGammaproteobacteria: Alcanivoracaceae_KangiellaFlavobacteriia: Flavobacteriaceae_NonlabensGammaproteobacteria: Alteromonadaceae_ParaglaciecolaAlphaproteobacteria: Rhodobacteraceae_RoseobacterFlavobacteriia: Flavobacteriaceae_FlavicellaGammaproteobacteria: Granulosicoccaceae_GranulosicoccusFlavobacteriia: Flavobacteriaceae_AlgibacterGammaproteobacteria: Pseudoalteromonadaceae_PseudoalteromonasSphingobacteriia: Saprospiraceae_unculturedFlavobacteriia: Flavobacteriaceae_Polaribacter_4Gammaproteobacteria: Alteromonadaceae_GlaciecolaFlavobacteriia: Flavobacteriaceae_Winogradskyella** *** ****** * *** ** ****** ** * * *** ***** *** * ****** ** **− −**** **− − −***** ***** *** ****−** *** * **** ***** *** ****** *** ****** *** *** * ***** *** **** *** ***−5 0 5Fold−changeColor Key −  * *****Absentp <= 0.05p <= 0.01p <= 0.001_____________WATER_____________NMF SURFACE<3% AbundantFigure 3.5: Enrichment and reduction of NMF surface and water column communitiesin M–W–NMF treatments. Fold-change enrichment or reduction of microbial genera was cal-culated using “DESeq2” in the “DESeq2” package in R. Both water sample treatments and NMFsurface treatments were compared to the NMF Alone control in order to calculated fold-change oftaxa. For a genera to be retained, fold-change p-values must be < 0.05. Additionally, each generamust occur at <3% relative abundance at least twice in the data set. Stars indicate the level ofsignificance for each taxa, whereas colours indicate fold-change. Taxa with dashes represent taxathat are not found in those samples, whereas white spaces mean they were not found in at least 3%relative abundance in any two samples. There are more differentially enriched genera in the watercolumn (twenty at >3% abundance) than on NMF surfaces (six at >3% abundance). Additionally,taxa that are enriched and abundant in the water column do not correlate well with those on NMFsurfaces.(Rubritalea) was reduced across all treatments.47Horizontal transmission of microbesRelative AbundanceWATER SAMPLES0.00.20.40.60.81.0WATER ONLY NMF ALONE WITH NEREO WITH MAST WITH N + MNMF SURFACE0.00.20.40.60.81.0Verrucomicrobiae: Verrucomicrobiaceae_PersicirhabdusVerrucomicrobiae: Rubritaleaceae_RubritaleaSphingobacteriia: Saprospiraceae_unculturedGammaproteobacteria: Pseudoalteromonadaceae_PseudoalteromonasGammaproteobacteria: Oceanospirillaceae_OceanobacterGammaproteobacteria: Granulosicoccaceae_GranulosicoccusGammaproteobacteria: Colwelliaceae_ColwelliaGammaproteobacteria: Alteromonadaceae_ParaglaciecolaGammaproteobacteria: Alteromonadaceae_GlaciecolaGammaproteobacteria: Alcanivoracaceae_KangiellaGammaproteobacteria: Alcanivoracaceae_AlcanivoraxFlavobacteriia: Flavobacteriaceae;Other_NAFlavobacteriia: Flavobacteriaceae_WinogradskyellaFlavobacteriia: Flavobacteriaceae_Polaribacter_4Flavobacteriia: Flavobacteriaceae_NonlabensFlavobacteriia: Flavobacteriaceae_FlavicellaFlavobacteriia: Flavobacteriaceae_AlgibacterChlamydiae: Simkaniaceae_unculturedBetaproteobacteria: Methylophilaceae_OM43_cladeAlphaproteobacteria: Rhodospirillaceae_unculturedAlphaproteobacteria: Rhodobacteraceae_RoseobacterAlphaproteobacteria: Rhodobacteraceae_PseudophaeobacterAlphaproteobacteria: Rhodobacteraceae_MarivitaAlphaproteobacteria: Rhodobacteraceae_LoktanellaFigure 3.6: Taxa summary plots showing enriched or reduced genera of NMF surfaceand water column communities. Taxa summary plots show all genera that are >3% abun-dant in at least 2 samples. The legend lists the class, family, and genus of each taxa. Coloredgenera are those that are also significantly enriched or reduced compared to controls (See Figure3.5 or the “Methods” section for details on how this was calculated). Genera that were >3%abundant in water samples but not enriched (shown as grey bars) include: NS3a and Wenying-shuangia (Flavobacteriia); Hyphomonas, Roseibacterium, and Sulfitobacter (Alphaproteobacteria);and Pseudohongiella (Gamaproteobacteria). Genera that were >3% abundant in NMF surfacesamples but not enriched (shown as grey bars) include: Dokdonia (Flavobacteriia); an unculturedSaprospiraceae (Sphingobacteriia); Litorimonas, an uncultured Rhodobacteraceae, and Erythrobac-ter (Alphaproteobacteria); Glaciecola, Paraglaciecola, an uncultured Alteromonadaceae, Colwellia,Thalassotalea, Arenicella, and Alcanivorax (Gammaproteobacteria). Taxa that were differentiallyenriched represent less of the overall community on NMF surfaces than in the water column.There is a striking decline in abundance of the genus Rubritalea on NMF surfaces when anymacroalgae is added (Nereocystis, Mastocarpus, and Nereocystis + Mastocarpus treatments)(Fig. 3.6, 3.5). Bacteria from this genus are highly abundant on NMF surfaces when they areincubated alone, and are also highest in abundance in the water of NMF-alone tanks. It isnot abundant in the water-only control, suggesting it is NMF-surface associated. Rubritalea48Horizontal transmission of microbesis also found in significant abundances on wild Nereocystis: although abundant in both swablocations, it is more abundant in the 10cm swabs than 50cm swabs (Fig. S3.14).Changes in Water Quality During ExperimentWe measured temperature, salinity, dissolved oxygen, and pH of the water in both the M-W and M-W-NMF experiments. Temperature and salinity were not significantly differentbetween treatments. Dissolved oxygen and pH in treatments with macroalgae co-incubateswere higher in treatments without macroalgae. In the M-W-NMF experiment, these differ-ences are significant (One-way ANOVA: Dissolved oxygen p < 0.001, F4,58=15.429; pH p <0.001, F4,58=11.556) but in the M-W experiment, they are not (One-way ANOVA: Dissolvedoxygen p = 0.316, F1,18=1.0621; pH p =0.058, F1,18=4.115).3.4 DiscussionDifferent species of macroalga induce different community shifts in the water columnThe goal of our project was to determine the extent to which neighbouring macroalgaealter the microbiota of each other. Our hypothesis was based on previous observations thatsignificant microbial community shifts occur in the water column [Lam and Harder, 2007,Lamet al., 2008,Clasen and Shurin, 2014,Miller and Page, 2012] and on nearby biofilms [VegaThurber et al., 2012, Fischer et al., 2014, Zaneveld et al., 2016] when co-incubated withmacroalgae. Additionally, experiments done by Lam et al. 2008 show that different speciesof macroalgae cause differential shifts in microbial community composition in the water.Therefore our first goal was to verify that Nereocystis and Mastocarpus induced microbialcommunity shifts in the surrounding water column, and that these shifts were differentbetween species.Our results confirm that the effect of macroalga on water column communities is species-specific. In general, treatments with Nereocystis experienced more enrichments and reduc-tions in individual microbial genera than treatments with Mastocarpus. The richness ofmicrobiota found on Nereocystis is also consistently lower than on Mastocarpus. This sug-gests that Nereocystis exudates may be more selective than Mastocarpus exudates.Interestingly, Nereocystis + Mastocarpus treatments produce water column communitiesthat are more similar in composition to Nereocystis only treatments, but more similar inrichness to Mastocarpus treatments. Additionally, treatments with both macroalgae are less49Horizontal transmission of microbesrich than treatments with Mastocarpus alone, which suggests that the effect of differentmacroalgae on the water column community is not additive. Although the Nereocystis +Mastocarpus treatment is significantly less diverse than the treatment withMastocarpus only,it is unclear whether this trend is due to potential antagonistic effects between Mastocarpusand Nereocystis exudates, or if it is because there was less total Mastocarpus tissue in thecombined treatment.Microbial communities associated with macroalgal surfaces are more resistant to change thanwater columnHorizontal transmission of microbes between neighbouring macroalgae may affect enrichmentof certain members of the macroalgal epibiotic community, but we find that macroalgalsurface communities are, in general, highly resistant to change. Our data shows that, unlikein water column communities, NMF surface communities do not differ when co-incubatedwith different species of seaweed. Instead, we only detected differences between control(NMF alone) and treatment samples, which were largely driven by the reduction of a singlegenera, Rubritalea.Identification of NMF-specific generaThe genus Rubritalea (Verrumicrobia) was found on all NMF surfaces, while other generawere highly variable across samples. Rubritalea also appears to drive the shift in communitycomposition when NMF are grown with other macroalgae (Fig. 3.6, Fig. 3.5). Represen-tatives of the genus Rubritalea produce pink-orange pigments and squalene [Scheuermayeret al., 2006, Kasai et al., 2007, Yoon et al., 2008, Yoon et al., 2007], the latter of whichis a precursor to steroids and D-vitamins [Bloch, 1983]. Interestingly, both steroids andD-vitamins are known to promote growth in some species of macroalgae [Fries, 1983]. Rubri-talea was previously isolated from sponges [Scheuermayer et al., 2006] and is a close relativeto Akkermansia, which is commensal to humans. This suggests Rubritalea is a generallyhost-associated genus. The mechanism behind why Rubritalea is reduced in co-incubationtreatments is unknown. It is possible that Rubritalea are reduced because they are propor-tionally less represented in co-incubation treatments, which generally have higher microbialrichness. Alternatively, Rubritalea may be outcompeted by other members of the microbiotawhen near mature blades of macroalgae.Rubritalea is also found (at >3% relative abundance) on wild seaweeds. It is more abundantin regions closer to the meristem (10cm versus 50cm from blade base, Fig. S3.14), whichfurther supports the hypothesis that it is a Nereocystis meristem-specific microbe.50Horizontal transmission of microbesThere is high host specificity across samplesWe found a strong signal of host specificity in our data. Comparisons between all sam-ples (including wild and lab-incubated seaweed) show that the strongest driving factor ofmicrobial community composition is macroalgal species (Nereocystis vs Mastocarpus), andnot sampling location (in situ or in-lab) (Fig. 3.2). While previous studies show highvariation in microbial community membership within a single species of macroalgae [Burkeet al., 2011b, Burke et al., 2011a], our data suggests that the variation in microbial taxabetween different species of macroalgae may be even greater. Other studies that comparewithin-species with between-species variation in microbiota structure have also found thatspecies is a stronger predictor of microbial community composition than location [Lachnitet al., 2009]. Therefore, the high variation in microbial community composition observed instudies of individual species of macroalgae may be misleading because the variation is notcompared to the variation that exists between species.ConclusionIn conclusion, we found that the influence of neighbouring macroalgae on NMF epibioticcommunities is limited. We identified only six genera that were differentially enriched acrossany treatment group, and there were no significant differences between community structuresof NMF communities incubated with different species of macroalgae. This suggests thatmacroalgal surfaces are more resistant to change than the surrounding water column. Wealso place our findings in a larger context: although there is high variation between NMF andNereocystis mature blade samples, it is clear that macroalgal species is a stronger driver inmicrobial community assembly than environment or treatment. Whether the subtle changesin microbiota observed on NMFs translate to biologically important functional differencesare unknown, but future work may be done to elucidate the effects of these microbes onoverall community function.51Horizontal transmission of microbes3.5 Supplementary Figures and Tableslllllllllll lllll ll ll llllll lllllllllllllllllllllllllllllllllllllll llll ll ll−0.4 −0.2 0.0 0.2 0.4−0.3−0.2−0.10.00.10.20.3NMDS plot of all samples (UWUF)0.16NMDS 1NMDS 2lllllllllNereo Meristem (lab)Nereo Meristem (wild)Nereo blade (lab)Nereo blade (wild)Mast blade (lab)Mast blade (wild)Water alone (NMF experiment)Water (NMF experiment)Water (Single Sp. Experiment)lllllllll lllll llllll lllllllllllllllllllllllllllllllllll lllllllllllllllllll−0.5 0.0 0.5−0.50.00.5NMDS plot of all samples (WUF)0.16NMDS 1NMDS 2lllllllllNereo Meristem (lab)Nereo Meristem (wild)Nereo blade (lab)Nereo blade (wild)Mast blade (lab)Mast blade (wild)Water alone (NMF experiment)Water (NMF experiment)Water (Single Sp. Experiment)Figure 3.7: S: NMDS of Nereocystis, Mastocarpus, and water samples. Un-weighted(left) and weighted (right) unifrac metrics. Nereocystis, Mastocarpus, and water samples clusterseparately. For statistical results, refer to Table 3.3.llllll468101214Alpha Diversity (PD_whole_tree)Nereo Mast WaterSample Typellllll100150200250Alpha Diversity (observed_otus)Nereo Mast WaterSample TypeFigure 3.8: S: Richness of Nereocystis, Mastocarpus, and water samples. PD_whole_treemetric (left) and observed_otus (right) metric. Mastocarpus surfaces are consistently richer thanNereocystis surfaces. For statistical results, refer to Table 3.752Horizontal transmission of microbesllllllllllll llllllllllll−0.2 −0.1 0.0 0.1 0.2−0.15−0.10−0.050.000.050.100.15NMDS of water samples0.16NMDS 1NMDS 2 lllllWater onlyNMF Alonewith Nereowith Mastwith Nereo + Mastlllllllll lllllllllllllll−0.4 −0.2 0.0 0.2 0.4−0.3−0.2−0.10.00.10.20.3NMDS of water samples0.13NMDS 1NMDS 2 lllllWater onlyNMF Alonewith Nereowith Mastwith Nereo + MastFigure 3.9: S: NMDS of water column communities in M–W–NMF treatments. NMDSof water community composition (created from an un-weighted (left) and weighted (right) Unifracdistance matrix) from the M–W–NMF experiment. NMF alone treatments are different fromall three treatments (with Nereocystsi, Mastocarpus, or both). Treatments are also significantlydifferent from each other. For statistical results, refer to Table S3.5.llWater only (4)NMF Alone (5)With Nereo (5)With Mast (5)With Both (5)68101214Alpha diversity across water samplesAlpha Diversity (PD_whole_tree)TreatmentllNMF alone (5)With Nereo (5)With Mast (3)With Both (5)80100120140160180200Alpha diversity across meristem swabsAlpha Diversity (observed_otus)TreatmentFigure 3.10: S: Richness of water column communities in M–W–NMF treatments.Richness (PD_whole_tree metric on the left and observed_otus on the right) of water columncommunities across treatments in the M–W–NMF experiment. Richness in treatments with Mas-tocarpus is higher than treatments with Nereocystis only. Richness in treatments with macroalgaeco-incubates is higher than treatments without. For statistical results, refer to Table 3.9.53Horizontal transmission of microbesllllllllllllllllll−0.6 −0.4 −0.2 0.0 0.2 0.4−0.4−0.20.00.20.4NMDS of Incubation Experiment0.06NMDS 1NMDS 2 llllWater−NereoNereoWater−MastMastNereo Swab (4)Mast Swab (5)Nereo Water (4)Mast Water (5)200250300350400Alpha diversityAlpha Diversity (chao1_even_1000_alpha)Treatmentll llllllllllllllll−0.2 −0.1 0.0 0.1 0.2 0.3−0.2−0.10.00.10.2NMDS of Incubation Experiment0.09NMDS 1NMDS 2 llllWater−NereoNereoWater−MastMastNereo Swab (4)Mast Swab (5)Nereo Water (4)Mast Water (5)68101214Alpha diversityAlpha Diversity (PD_whole_tree_even_1000_alpha)Treatmentlllllllllll llllll l−1.0 −0.5 0.0 0.5 1.0−0.50.00.5NMDS of Incubation Experiment0.13NMDS 1NMDS 2 llllWater−NereoNereoWater−MastMastlNereo Swab (4)Mast Swab (5)Nereo Water (4)Mast Water (5)100150200250Alpha diversityAlpha Diversity (observed_otus_even_1000_alpha)TreatmentFigure 3.11: S: Composition and Richness of samples in the M–W experiment. NMDSplots of community dissimilarity (left) and box plots of richness (right) in all metrics used. NMDSmetrics are Bray-Curtis (top left), un-weighted Unifrac (centre left), weighted Unifrac (bottom left);richness metrics used are Chao1 (top right), PD_whole_tree (middle right), and observed_otus(bottom right). For statistics, refer to Table S3.654Horizontal transmission of microbeslllllll llllllll lll−0.6 −0.4 −0.2 0.0 0.2 0.4 0.6−0.4−0.20.00.2NMDS of Nereo Meristem Swabs0.08NMDS 1NMDS 2llllNMF Alonewith Nereowith Mastwith Nereo + Mastlllllllllll lllllll−0.4 −0.2 0.0 0.2 0.4−0.4−0.20.00.20.4NMDS of Nereo Meristem Swabs0.19NMDS 1NMDS 2llllNMF Alonewith Nereowith Mastwith Nereo + MastFigure 3.12: S: Comparison of NMF surface communities from M–W–NMF treat-ments. NMDS of NMF surface communities (created from an un-weighted (left) and weighted(right) Unifrac distance matrix) from the M–W–NMF experiment. NMF communities do not showsignificant differences between treatments, but NMF Alone treatments are different from NMF with(any) macroalgae treatments. For statistical results, refer to Table 3.5.llNMF alone (5)With Nereo (5)With Mast (3)With Both (5)456789Alpha diversity across meristem swabsAlpha Diversity (PD_whole_tree)TreatmentllNMF alone (5)With Nereo (5)With Mast (3)With Both (5)80100120140160180200Alpha diversity across meristem swabsAlpha Diversity (observed_otus)TreatmentFigure 3.13: S: Richness across NMF surface communities from M–W–NMF treat-ments. Richness (PD_whole_tree metric on the left and observed_otus on the right) of NMFsurface communities in the M–W–NMF experiment. There is no statistical difference betweenrichness of any treatment. NMF Alone controls are less rich than NMF with (any) macroalgaetreatments. For statistical results, refer to Table 3.9.55Horizontal transmission of microbesRelative Abundance0.00.20.40.60.81.010cm Nereocystis50cm NereocystisMastocarpusUnidentifiedVerrucomicrobiae: Verrucomicrobiaceae_PersicirhabdusVerrucomicrobiae: Rubritaleaceae_RubritaleaGammaproteobacteria: Alcanivoracaceae_AlcanivoraxGammaproteobacteria: Granulosicoccaceae_GranulosicoccusAlphaproteobacteria: Erythrobacteraceae_ErythrobacterAlphaproteobacteria: Hyphomonadaceae_LitorimonasAlphaproteobacteria: uncultured_unculturedPlanctomycetacia: Planctomycetaceae_BlastopirellulaSphingobacteriia: Saprospiraceae_unculturedSphingobacteriia: Saprospiraceae_LewinellaFlavobacteriia: Flavobacteriaceae_unculturedFlavobacteriia: Flavobacteriaceae_MaribacterFlavobacteriia: Flavobacteriaceae_DokdoniaFlavobacteriia: Flavobacteriaceae;Other_NAAcidimicrobiia: uncultured_unculturedOTHER TAXA >3% ABUNDANCE (grey): Sphingobacteriia: Saprospiraceae_PortibacterAlphaproteobacteria: Rhodobacteraceae;Other_NAAlphaproteobacteria: Rhodobacteraceae_unculturedVerrucomicrobiae: Verrucomicrobiaceae_RoseibacillusRelative Abundance0.00.20.40.60.81.0NereoMastWater(N)Water(M)UnidentifiedGammaproteobacteria: Granulosicoccaceae_GranulosicoccusGammaproteobacteria: Colwelliaceae_ColwelliaGammaproteobacteria: Alteromonadaceae_ParaglaciecolaGammaproteobacteria: Alteromonadaceae_GlaciecolaBetaproteobacteria: Methylophilaceae_OM43_cladeAlphaproteobacteria: Erythrobacteraceae_ErythrobacterAlphaproteobacteria: Rhodobacteraceae_RoseobacterAlphaproteobacteria: Rhodobacteraceae_RoseibacteriumAlphaproteobacteria: Rhodobacteraceae_LoktanellaAlphaproteobacteria: Unknown_Family_unculturedSphingobacteriia: Saprospiraceae_unculturedSphingobacteriia: Saprospiraceae_LewinellaFlavobacteriia: Flavobacteriaceae_unculturedFlavobacteriia: Flavobacteriaceae_WinogradskyellaFlavobacteriia: Flavobacteriaceae_MaribacterFlavobacteriia: Flavobacteriaceae_DokdoniaFlavobacteriia: Flavobacteriaceae_AlgibacterFlavobacteriia: Flavobacteriaceae;Other_NAOTHER TAXA >3% ABUNDANCE (grey): Flavobacteriia: Flavobacteriaceae_NS3a_marine_groupAlphaproteobacteria: Rhodobacteraceae;Other_NAAlphaproteobacteria: Rhodobacteraceae_PontivivensAlphaproteobacteria: Rhodobacteraceae_SulfitobacterAlphaproteobacteria: Rhodobacteraceae_unculturedAlphaproteobacteria: Erythrobacteraceae_AltererythrobacterGammaproteobacteria: Oceanospirillaceae_PseudohongiellaVerrucomicrobiae: Verrucomicrobiaceae_RoseibacillusFigure 3.14: S: Taxa summary of wild and lab-incubated seaweed samples. Plots showtaxa at the genus level for (A) wild and(B) lab-incubated seaweed samples.56Horizontal transmission of microbesBray-Curtis Un-weighted Unifrac Weighted UnifracGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNereo Mast 0.001 (R2=0.191,df=1,41) 0.001 0.001 (R2=0.223,df=1,41) 0.001 0.001 (R2=0.256,df=1,41) 0.001Nereo Water 0.001 (R2=0.168,df=1,64) 0.001 0.001 (R2=0.151,df=1,64) 0.001 0.001 (R2=0.146,df=1,64) 0.001Mast Water 0.001 (R2=0.242,df=1,42) 0.001 0.001 (R2=0.224,df=1,42) 0.001 0.001 (R2=0.235,df=1,42) 0.001Table 3.3: S: PERMANOVA results comparing Nereocystis, Mastocarpus, and watersamples. Pairwise PERMANOVA calculations were done separately.Bray-Curtis Un-weighted Unifrac Weighted UnifracGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNMF Water only 0.094 (R2=0.162,df=1,8) - 0.083 (R20.222,df=1,8) - 0.005 (R20.276,df=1,8) -NereoNMF Mast 0.001 (R2=0.145,df=1,19) - 0.003 (R2=0.1662,df=1,19) - 0.001 (R2=0.203,df=1,19) -Nereo + MastNereo Mast 0.005 (R20.336,df=1,9) 0.012 0.012 (R20.289,df=1,9) 0.021 0.009 (R20.323,df=1,9) 0.009Nereo Nereo + Mast 0.087 (R20.154,df=1,9) 0.087 0.764 (R20.071,df=1,9) 0.764 0.006 (R20.219,df=1,9) 0.009Mast Nereo + Mast 0.008 (R20.250,df=1,9) 0.012 0.014 (R20.265df=1,9) 0.021 0.004 (R20.199,df=1,9) 0.009Table 3.4: S: PERMANOVA results comparing water samples from M–W–NMF ex-periment.Bray-Curtis Un-weighted Unifrac Weighted UnifracGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNereoNMF Mast 0.003 (R2=0.125,df=1,17) - 0.015 (R2=0.198,df=1,17) - 0.02 (R2=0.088,df=1,17) -Nereo + MastNereo Mast 0.03 (R2=0.216,df=1,7) 0.048 0.385 (R2=0.130,df=1,7) 0.385 0.134 (R2=0.175,df=1,7) 0.2625Nereo Nereo + Mast 0.271 (R2=0.124,df=1,9) 0.271 0.284 (R2=0.122,df=1,9) 0.385 0.175 (R2=0.137,df=1,9) 0.2625Mast Nereo + Mast 0.032 (R2=0.206,df=1,7) 0.048 0.351 (R2=0.144,df=1,7) 0.385 0.61 (R2=0.138,df=1,7) 0.61Table 3.5: S: PERMANOVA results comparing NMF surface samples fromM–W–NMFexperiment.PERMANOVA Welch’s t-TestMetric p-value Metric p-valueBray-Curtis 0.039 (R2=0.266,df=1,8) Chao1 0.188 (t=1.46,df=6.94)Un-weighted Unifrac 0.023 (R2=0.374,df=1,8) PD_whole_tree 0.943 (t=0.07,df=5.8)Weighted Unifrac 0.1 (R2=0.179,df=1,8) Observed_otus 0.17 (t=1.54,df=6.57)Table 3.6: S: PERMANOVA results comparing water samples from the M–W experi-ment.57Horizontal transmission of microbesChao1 PD_whole_tree Observed_otusGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNereo Mast 0.000641 (t=4.14,df=17.58) 0.00192 4.16e-07 (t=8.26,df=15.7) 1.25e-06 0.000104 (t=5.55,df=12.61) 0.000312Nereo Water 0.0375 (t=2.23,df=19.52) 0.0375 7.27e-06 (t=6.25,df=17.71) 1.09e-05 0.000525 (t=4.49,df=13.78) 0.000787Mast Water 0.0286 (t=-2.24,df=62.58) 0.0375 0.0224 (t=-2.34,df=62.27) 0.0224 0.122 (t=-1.57,df=62.15) 0.122Table 3.7: S: ANOVA results comparing richness of Nereocystis, Mastocarpus, andwater.Chao1 PD_whole_tree Observed_otusGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNMF Water only 2.605e-07 (t=-6.26,df=6.98) - 0.00528 (t=-3.99,df=7) - 0.00117 (t=-6.35,df=5.27) -NereoNMF Mast 0.001 (t=-7.9526,df=18.061) - 2.653e-09 (t=-9.8915,df=20.741) - 9.913e-05 (t=-5.9102,df=11.068) -Nereo + MastNereo Mast 0.0121 (t=-3.76,df=5.23) 0.0363 0.000391 (t=-5.05,df=10.8) 0.00117 0.00634 (t=-3.63,df=8.25) 0.019Nereo Nereo + Mast 0.0269 (t=-3.15,df=4.78) 0.0404 0.0592 (t=-2.48,df=4.66) 0.0889 0.0442 (t=-2.4,df=7.72) 0.0663Mast Nereo + Mast 0.168 (t=1.46,df=12.53) 0.168 0.157 (t=1.53,df=9.8) 0.157 0.242 (t=1.24,df=9.88) 0.242Table 3.8: S: ANOVA and Welch’s t-Test results comparing richness of water samplesfrom M–W–NMF experiment.Chao1 PD_whole_tree Observed_otusGroup 1 Group 2 p FDR adj. p p FDR adj. p p FDR adj. pNereoNMF Mast 0.271 (t=-1.167,df=9.857) - 0.0621 (t=-2.127,df=9.044) - 0.117 (t=-1.714,df=10.038) -Nereo + MastNereo Mast 0.286 (t=-1.13, df=10.06) 0.775 0.65 (t=0.47, df=10.94) 0.65 0.682 (t=-0.42, df=10.94) 0.951Nereo Nereo + Mast 0.517 (t=-0.69, df=5.3) 0.775 0.392 (t=0.92, df=6.59) 0.65 0.951 (t=-0.06, df=6.71) 0.951Mast Nereo + Mast 0.854 (t=0.19, df=8.51) 0.854 0.63 (t=0.5, df=8.2) 0.65 0.794 (t=0.27, df=8.41) 0.951Table 3.9: S: ANOVA and Welch’s t-Test results comparing NMF surface samples fromM–W–NMF experiment.58BibliographyBibliography[Abelson and Denny, 1997] Abelson, A. and Denny, M. (1997). Settlement of Marine Organismsin Flow. Ecology, 28:317–339.[Anderson and Martone, 2014] Anderson, L. M. and Martone, P. T. (2014). Biomechanical conse-quences of epiphytism in intertidal macroalgae. The Company of Biologists, 217:1167–1174.[Armstrong et al., 2001] Armstrong, E., Yan, L., Boyd, K. G., Wright, P. C., and Burgess, J. G.(2001). The symbiotic role of marine microbes on living surfaces. Hydrobiologia, 461:37–40.[Bengtsson and Øvreås, 2010] Bengtsson, M. M. and Øvreås, L. (2010). Planctomycetes dominatebiofilms on surfaces of the kelp Laminaria hyperborea. BMC Microbiology, 10(1):261.[Bengtsson et al., 2012] Bengtsson, M. M., Sjøtun, K., Lanzén, A., and Øvreås, L. (2012). Bacterialdiversity in relation to secondary production and succession on surfaces of the kelp Laminariahyperborea. The ISME Journal, 6:2188–2198.[Benjamini et al., 1995] Benjamini, Y., Hochberg, Y., Society, R. S., Benajmini, Y., Hochberg, Y.,Society, R. S., Benjamini, y. Y., Hochberg, Y., Benajmini, Y., Hochberg, Y., Society, R. S., Be-najmini, Y., Hochberg, Y., Society, R. S., Benjamini, y. Y., and Hochberg, Y. (1995). Controllingthe False Discovery Rate : A Practical and Powerful Approach to Multiple Testing Author ( s ):Yoav Benjamini and Yosef Hochberg Source : Journal of the Royal Statistical Society . Series B( Methodological ), Vol . 57 , No . 1 Published by :. J R Statist Soc B, 57(1):289–300.[Berland et al., 1972] Berland, B. R., Bonin, D. J., and Maestrini, S. Y. (1972). Are some bacteriatoxic for marine algae? Marine Biology, 12(3):189–193.[Bhakuni and Rawat, 2006] Bhakuni, D. S. and Rawat, D. (2006). Bioactive marine natural prod-ucts. Spring Science and Business Media.[Bloch, 1983] Bloch, K. E. (1983). Sterol, Structure and Membrane Function. Critical Reviews inBiochemistry and Molecular Biology, 14:47–92.[Bolinches et al., 1988] Bolinches, J., Lemos, M. L., and Barja, J. L. (1988). Population dynamicsof heterotrophic bacterial communities associated with Fucus vesiculosus and Ulva rigida in anestuary. Microbial Ecology, 15(3):345–357.[Bondoso et al., 2014] Bondoso, J., Balague, V., Gasol, J. M., and Lage, O. M. (2014). Communitycomposition of the Planctomycetes associated with different macroalgae. FEMS MicrobiologyEcology, 88(3):445–456.[Bray and Curtis, 1957] Bray, J. R. and Curtis, J. (1957). An Ordination of the Upland ForestCommunities of Southern Wisconsin. Ecological Monographs, 27(4):325–349.[Brock and Clyne, 1984] Brock, T. D. and Clyne, J. (1984). Significance of algal excretory productsfor growth of epilimnetic bacteria. Applied and Environmental Microbiology, 47(4):731–734.[Bulleri et al., 2002] Bulleri, F., Benedetti-Cecchi, L., Acunto, S., Cinelli, F., and Hawkins, S. J.(2002). The influence of canopy algae on vertical patterns of distribution of low-shore assemblageson rocky coasts in the northwest Mediterranean. Journal of Experimental Marine Biology andEcology, 267(1):89–106.59Bibliography[Burke et al., 2011a] Burke, C., Steinberg, P., Rusch, D., Kjelleberg, S., and Thomas, T. (2011a).Bacterial community assembly based on functional genes rather than species. Proceedings of theNational Academy of Sciences of the United States of America, 108(34):14288–14293.[Burke et al., 2011b] Burke, C., Thomas, T., Lewis, M., Steinberg, P., and Kjelleberg, S. (2011b).Composition, uniqueness and variability of the epiphytic bacterial community of the green algaUlva australis. The ISME Journal, 5(4):590–600.[Caporaso et al., 2010] Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman,F. D., Costello, E. K., Fierer, N., Peña, A. G., Goodrich, J. K., Gordon, J. I., Huttley, G. a.,Kelley, S. T., Knights, D., Koenig, J. E., Ley, R. E., Lozupone, C. a., Mcdonald, D., Muegge,B. D., Pirrung, M., Reeder, J., Sevinsky, J. R., Turnbaugh, P. J., Walters, W. a., Widmann, J.,Yatsunenko, T., Zaneveld, J., and Knight, R. (2010). QIIME allows analysis of high- throughputcommunity sequencing data. Nature Publishing Group, 7(5):335–336.[Carpenter and Cox, 1974] Carpenter, E. J. and Cox, J. L. (1974). Production of Pelagic Sargas-sum and a Blue-Green Epiphyte in the Western Sargasso Sea. Limnology and Oceanography,19(3):429–436.[Case et al., 2011] Case, R. J., Longford, S. R., Campbell, A. H., Low, A., Tujula, N., Steinberg,P. D., and Kjelleberg, S. (2011). Temperature induced bacterial virulence and bleaching diseasein a chemically defended marine macroalga. Environmental Microbiology, 13(2):529–537.[Chao, 1984] Chao, A. (1984). Nonparametric Estimation of the Number of Classes in a Population.Scandinavian Journal of Statistics, 11(4):265–270.[Characklis and Marshall, 1990] Characklis, W. G. and Marshall, K. C. (1990). Biofilms. Wiley,New York.[Chisholm et al., 1996] Chisholm, J., Douga, C., Ageron, E., Grimont, P., and Jaubert, J. (1996).’Roots’ in mixotrophic algae. Nature, 381:382.[Clasen and Shurin, 2014] Clasen, J. L. and Shurin, J. B. (2014). Kelp forest size al-ters microbial community structure and function on Vancouver Island, Canada. Ecology,96(3):140815103525005.[Coveney and Wetzel, 1989] Coveney, M. F. and Wetzel, R. G. (1989). Bacterial metabolism ofalgal extracellular carbon. Hydrobiologia, 173(2):141–149.[Croft et al., 2005] Croft, M. T., Lawrence, A. D., Raux-Deery, E., Warren, M. J., and Smith,A. G. (2005). Algae acquire vitamin B12 through a symbiotic relationship with bacteria. Nature,438:90–93.[Dahl, 2016] Dahl, D. B. (2016). xtable: Export Tables to LaTeX or HTML. R package version1.8-2.[de Oliveira et al., 2012] de Oliveira, L. S., Gregoracci, G. B., Silva, G. G., Salgado, L. T., Filho,G. a., Alves-Ferreira, M. a., Pereira, R. C., and Thompson, F. L. (2012). Transcriptomic analysisof the red seaweed Laurencia dendroidea (Florideophyceae, Rhodophyta) and its microbiome.BMC Genomics, 13:487.[Dimitrieva et al., 2006] Dimitrieva, G. Y., Crawford, R. L., and Yüksel, G. Ü. (2006). The natureof plant growth-promoting effects of a pseudoalteromonad associated with the marine algae Lam-inaria japonica and linked to catalase excretion. Journal of Applied Microbiology, 100(5):1159–1169.60Bibliography[Dininno and McCandless, 1978] Dininno, V. and McCandless, E. (1978). Immunochemistry oflambda-type carrageenans from certain red algae. Carbohydrate Research, 67:235–241.[Dittami et al., 2015] Dittami, S. M., Duboscq-Bidot, L., Perennou, M., Gobet, A., Corre, E.,Boyen, C., and Tonon, T. (2015). Host-microbe interactions as a driver of acclimation to salinitygradients in brown algal cultures. The ISME journal, 10(1):51–63.[Dobretsov and Qian, 2002] Dobretsov, S. V. and Qian, P.-Y. (2002). Effect of Bacteria Asso-ciated with the Green Alga Ulva reticulata on Marine Micro- and Macrofouling. Biofouling,18(795099257):217–228.[Dring, 2006] Dring, M. (2006). Stress resistance and disease resistance in seaweeds: the role ofreactive oxygen metabolism. In Advances in Botanical Research, pages 175–207. Academic PressLRD- Elsevier Science LTD, London.[Dromgoole and J, 1978] Dromgoole, F. I. S. W. B. and J, H. B. (1978). Nitrogenase activityassociated with Codium species from New Zealand marine habitats. N.Z.Journal of Marine andFreshwater Research, 12(1):17–22.[Dubber and Harder, 2008] Dubber, D. and Harder, T. (2008). Extracts of Ceramium rubrum, Mastocarpus stellatus and Laminaria digitata inhibit growth of marine and fish pathogenicbacteria at ecologically realistic concentrations. Aquaculture, 274:196–200.[Edgar, 2010] Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST.Bioinformatics, 26(19):2460–2461.[Egan et al., 2013] Egan, S., Harder, T., Burke, C., Steinberg, P., Kjelleberg, S. S. S. S., andThomas, T. (2013). The seaweed holobiont: Understanding seaweed-bacteria interactions. FEMSMicrobiology Reviews, 37(3):462–476.[Egan et al., 2008] Egan, S., Thomas, T., and Kjelleberg, S. (2008). Unlocking the diversity andbiotechnological potential of marine surface associated microbial communities. Current Opinionin Microbiology, 11(3):219–225.[Eren et al., 2014] Eren, A. M., Morrison, H. G., Lescault, P. J., Reveillaud, J., Vineis, J. H., andSogin, M. L. (2014). Minimum entropy decomposition: Unsupervised oligotyping for sensitivepartitioning of high-throughput marker gene sequences. The ISME Journal, 9(4):968–979.[Evelegh et al., 1979] Evelegh, M., Vollmer, C., and McCandless, E. (1979). Differentiation ofthe lambda carrageenans from Rhodophyta with immunological and spectroscopic techniques.Journal of Phycology, 14(1):89–91.[Faith and Baker, 2006] Faith, D. P. and Baker, A. M. (2006). Phylogenetic diversity (PD) andbiodiversity conservation: some bioinformatics challenges. Evolutionary bioinformatics online,2:121–128.[Falshaw et al., 2003] Falshaw, R., Bixler, H., and Johndro, K. (2003). Structure and performanceof commercial kappa-2 carrageenan extracts. Part III. Structure analysis and performance in twodairy applications of extracts from the New Zealand red seaweed, Gigartina atropurpurea. FoodHydrocolloids, 17:129–139.[Fischer et al., 2014] Fischer, M., Friedrichs, G., and Lachnit, T. (2014). Fluorescence-Based Qua-sicontinuous and In Situ Monitoring of Biofilm Formation Dynamics in Natural Marine Environ-ments. Applied and Environmental Microbiology, 80(12):3721–3728.[Fox and Weisberg, 2011] Fox, J. and Weisberg, S. (2011). An R Companion to Applied Regression.Sage, Thousand Oaks CA, second edition.61Bibliography[Fries, 1983] Fries, L. (1983). D-vitamins and Their Precursors as Growth Regulators in AxenicallyCultivated Marine Macroalgae. Journal of Phycology, 20(1):62–66.[Goecke et al., 2010] Goecke, F., Labes, A., Wiese, J., and Imhoff, J. F. (2010). Chemical interac-tions between marine macroalgae and bacteria. Marine Ecology Progress Series, 409:267–299.[Hawley and Altizer, 2011] Hawley, D. M. and Altizer, S. M. (2011). Disease ecology meets ecolog-ical immunology: understanding the links between organismal immunity and infection dynamicsin natural populations. Functional Ecology, 25(1):48–60.[Hoegh-Guldberg, 1988] Hoegh-Guldberg, O. (1988). A method for determining the surface areaof corals. Coral Reefs, 7(3):113–116.[Hollants et al., 2011] Hollants, J., Leroux, O., Leliaert, F., Decleyre, H., De Clerck, O., andWillems, A. (2011). Who Is in There? Exploration of Endophytic Bacteria within the SiphonousGreen Seaweed Bryopsis (Bryopsidales, Chlorophyta). PLoS ONE, 6(10):e26458.[Huggett et al., 2006] Huggett, M. J., Williamson, J. E., De Nys, R., Kjelleberg, S., and Steinberg,P. D. (2006). Larval settlement of the common Australian sea urchin Heliocidaris erythrogrammain response to bacteria from the surface of coralline algae. Oecologia, 149(4):604–619.[Huston, 1979] Huston, M. (1979). A General Hypothesis of Species. The American Naturalist,113(1):81–101.[Ilead and Carpenter, 1975] Ilead, W. D. and Carpenter, E. J. (1975). Nitrogen fixation associatedwith the marine macroalga Codium fragile. Limnology and Oceanography, 20(5):815–823.[Johnson et al., 2001] Johnson, S., Thompson, M., and Hourigan, K. (2001). Flow past ellipticalcylinders at low Reynolds numbers. 14th Australasian Fluid . . . , pages 343–346.[Joint et al., 2002] Joint, I., Tait, K., Callow, M. E., Callow, J. A., Milton, D., Williams, P., andCámara, M. (2002). Cell-to-Cell Communication across the Prokaryote-Eukaryote Boundary.Science, 298(5596):5–6.[Kain, 1987] Kain, J. (1987). Patterns of relative growth in Nereocystis luetkeana (Phaeophyta).[Kasai et al., 2007] Kasai, H., Katsuta, A., Sekiguchi, H., Matsuda, S., Adachi, K., Shindo, K.,Yoon, J., Yokota, A., and Shizuri, Y. (2007). Rubritalea squalenifaciens sp. nov., a squalene-producing marine bacterium belonging to subdivision 1 of the phylum ’Verrucomicrobia’. Inter-national Journal of Systematic and Evolutionary Microbiology, 57(7):1630–1634.[Krumhansl and Scheibling, 2012] Krumhansl, K. A. and Scheibling, R. E. (2012). Production andfate of kelp detritus. Marine Ecology Progress Series, 467:281–302.[Küpper et al., 2002] Küpper, F. C., Müller, D. G., Peters, A. F., Kloareg, B., and Potin, P. (2002).Oligoalginate recognition and oxidative burst play a key role in natural and induced resistanceof sporophytes of Laminariales. Journal of Chemical Ecology, 28(10):2057–2081.[Lachnit et al., 2009] Lachnit, T., Blümel, M., Imhoff, J. F., and Wahl, M. (2009). Specific epibac-terial communities on macroalgae: Phylogeny matters more than habitat. Aquatic Biology,5(2):181–186.[Lachnit et al., 2011] Lachnit, T., Meske, D., Wahl, M., Harder, T., and Schmitz, R. (2011).Epibacterial community patterns on marine macroalgae are host-specific but temporally vari-able. Environmental Microbiology, 13(3):655–665.[Lam and Harder, 2007] Lam, C. and Harder, T. (2007). Marine macroalgae affect abundance andcommunity richness of bacterioplankton in close proximity. Journal of Phycology, 43(5):874–881.62Bibliography[Lam et al., 2008] Lam, C., Stang, A., and Harder, T. (2008). Planktonic bacteria and fungi areselectively eliminated by exposure to marine macroalgae in close proximity. FEMS MicrobiologyEcology, 63(3):283–291.[Lehtola et al., 2006] Lehtola, M. J., Laxander, M., Miettinen, I. T., Hirvonen, A., Vartiainen, T.,and Martikainen, P. J. (2006). The effects of changing water flow velocity on the formation ofbiofilms and water quality in pilot distribution system consisting of copper or polyethylene pipes.Water Research, 40:2151–2160.[Love et al., 2014] Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of foldchange and dispersion for rna-seq data with deseq2. Genome Biology, 15:550.[Lozupone and Knight, 2005] Lozupone, C. and Knight, R. (2005). UniFrac : a New Phyloge-netic Method for Comparing Microbial Communities UniFrac : a New Phylogenetic Method forComparing Microbial Communities. Applied and environmental microbiology, 71(12):8228–8235.[Marshall et al., 2006] Marshall, K., Joint, I., Callow, M. E., and Callow, J. A. (2006). Effect ofmarine bacterial isolates on the growth and morphology of axenic plantlets of the green algaUlva linza. Microbial Ecology, 52(2):302–310.[Mason et al., 2003] Mason, O., Hazen, T. C., Borglin, S., Chain, P. S., Dubinsky, E. A., Fortney,J. L., Han, J., Holman, H.-Y. N., and Hultman, J. (2003). Metagenome, metatranscriptomeand single-cell sequencing reveal microbial response to Deepwater Horizon oil spill. The ISMEJournal, 6(9):1715–1727.[Matsuo et al., 2005] Matsuo, Y., Imagawa, H., Nishizawa, M., and Shizuri, Y. (2005). Isolation ofan Algal Morphogenesis Inducer from Marine Bacterium. Science, 307(5715):1596.[Matz et al., 2008] Matz, C., Webb, J. S., Schupp, P. J., Phang, S. Y., Penesyan, A., Egan, S.,Steinberg, P., and Kjelleberg, S. (2008). Marine biofilm bacteria evade eukaryotic predation bytargeted chemical defense. PLoS ONE, 3(7):1–7.[Maximilian et al., 1998] Maximilian, R., De Nys, R., Holmstrom, C., Gram, L., Givskov, M.,Crass, K., Kjelleberg, S. A., and Steinberg, P. D. (1998). Chemical mediation of bacterial surfacecolonisation by secondary metabolites from the red alga Delisea pulchra. Aquatic MicrobialEcology, 15(3):233–246.[McHugh, 2003] McHugh, D. J. (2003). A guide to the seaweed industry. FAO fisheries technicalpaper, 441.[Meusnier et al., 2001] Meusnier, I., Olsen, J. L., Stam, W. T., Destombe, C., and Valero, M.(2001). Phylogenetic analyses of Caulerpa taxifolia (Chlorophyta) and of its associated bacterialmicroflora provide clues to the origin of the Mediterranean introduction. Molecular Ecology,10(4):931–946.[Michelou et al., 2013] Michelou, V. K., Caporaso, J. G., Knight, R., and Palumbi, S. R. (2013).The Ecology of Microbial Communities Associated with Macrocystis pyrifera. PLoS ONE, 8(6):1–9.[Miller and Page, 2012] Miller, R. J. and Page, H. M. (2012). Kelp as a trophic resource for marinesuspension feeders: A review of isotope-based evidence. Marine Biology, 159(7):1391–1402.[Nakanishi et al., 1996] Nakanishi, K., Nishijima, M., Nishimura, M., Kuwano, K., and Saga, N.(1996). Bacteria that induce morphogenesis in Ulva pertusa (Chlorophyta) grown under axenicconditions. Journal of Phycology, 482(479):482–479.63Bibliography[Nedashkovskaya, 2004] Nedashkovskaya, O. (2004). Algibacter lectus gen. nov., sp. nov., a novelmember of the family Flavobacteriaceae isolated from green algae. International Journal ofSystematic and Evolutionary Microbiology, 54(4):1257–1261.[Newell et al., 1980] Newell, R. C., Lucas, M. I., Velimirov, B., and Seiderer, L. J. (1980). Quantita-tive Significance of Dissolved Organic Losses Following Fragmentation of Kelp (Ecklonia maximaand Laminaria pallida). Marine Ecology Progress Series, 2(1977):45–59.[Oksanen et al., 2017] Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P.,McGlinn, D., Minchin, P. R., O’Hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H.,Szoecs, E., and Wagner, H. (2017). vegan: Community Ecology Package. R package version2.4-3.[Ollos et al., 2003] Ollos, P. J., Huck, P. M., and Slawson, R. M. (2003). Factors Affecting BiofilmAccumulation in Model Distribution Systems. American Water Works Association, 95(1):87–97.[Persson et al., 2011] Persson, F., Svensson, R., Nylund, G. M., Fredriksson, N. J., Pavia, H.,Hermansson, M., Persson, F., Svensson, R., Nylund, G. M., and Johan, N. (2011). Ecological roleof a seaweed secondary metabolite for a colonizing bacterial community. Biofouling, 27(6):579–588.[Price et al., 2009] Price, M. N., Dehal, P. S., and Arkin, A. P. (2009). Fasttree: Computinglarge minimum evolution trees with profiles instead of a distance matrix. Molecular Biology andEvolution, 26(7):1641–1650.[Provasoli and Pintner, 1980] Provasoli, L. and Pintner, I. J. (1980). Bacteria induced polymor-phism in an axenic laboratory strain of ulva lactuca (Chlorophyceae). Journal of Phycology,16:196–201.[R Core Team, 2016] R Core Team (2016). R: A Language and Environment for Statistical Com-puting. R Foundation for Statistical Computing, Vienna, Austria.[Rao et al., 2010] Rao, D., Skovhus, T., Tujula, N., Holmström, C., Dahllöf, I., Webb, J. S.,and Kjelleberg, S. (2010). Ability for \textit{Pseudoalteromonas tunicata} to colonize natu-ral biofilms and its effect on microbial community structure. FEMS Microbiol Ecol, 73:450–457.[Riquelme et al., 1997] Riquelme, C., Rojas, A., Flores, V., and Correa, J. A. (1997). Epi-phytic bacteria in a copper-enriched environment in northern Chile. Marine Pollution Bulletin,34(10):816–820.[Rosenberg and Paerl, 1981] Rosenberg, G. and Paerl, H. W. (1981). Nitrogen fixation by blue-green algae associated with the siphonous green seaweed Codium decorticatum: effects on am-monium uptake. Marine Biology, 61(2-3):151–158.[Rusconi et al., 2014] Rusconi, R., Guasto, J. S., and Stocker, R. (2014). Bacterial transport sup-pressed by fluid shear. Nature Physics, 10(3):212–217.[Rusconi and Stocker, 2015] Rusconi, R. and Stocker, R. (2015). Microbes in flow. Current Opinionin Microbiology, 25:1–8.[Rutter and Vincent, 1988] Rutter, P. and Vincent, B. (1988). Attachment mechanisms in thesurface growth of microorganisms. Physiological models in microbiology. CRC Press, Boca Raton,Fla, pages 87–107.[Scheuermayer et al., 2006] Scheuermayer, M., Gulder, T. A. M., Bringmann, G., and Hentschel, U.(2006). Rubritalea marina gen. nov., sp. nov., a marine representative of the phylum ’Verrucomi-crobia’, isolated from a sponge (Porifera). International Journal of Systematic and EvolutionaryMicrobiology, 56(9):2119–2124.64Bibliography[Schiel and Foster, 2006] Schiel, D. R. and Foster, M. S. (2006). The population biology of largebrown seaweeds : Ecological consequences life histories of multiphase in dynamic coastal envi-ronments. Annual Review of Ecology, Evolution and Systematics, 37(2006):343–372.[Semenova et al., 2009] Semenova, E., Shlykova, D., Semenov, A., Ivanov, M., Shelyakov, O., andNetrusov, A. (2009). Bacteria-Epiphytes of Brown Macro Alga in Oil Utilization in North SeaEcosystems. Moscow University Biological Sciences Bulletin, 64(3):107–110.[Seyedsayamdost et al., 2011] Seyedsayamdost, M. R., Case, R. J., Kolter, R., and Clardy, J.(2011). The Jekyll-and-Hyde chemistry of Phaeobacter gallaeciensis. Nature Chemistry, 3(4):331–335.[Shade and Handelsman, 2011] Shade, A. and Handelsman, J. (2011). Beyond the Venn diagram:the hunt for a core microbiome. Environmental Microbiology, 14(1):4–12.[Singer et al., 2010] Singer, G., Besemer, K., Schmitt-Kopplin, P., Hödl, I., and Battin, T. J.(2010). Physical Heterogeneity Increases Biofilm Resource Use and Its Molecular Diversity inStream Mesocosms. PLoS One, 5(4):e9988.[Singh and Reddy, 2014] Singh, R. P. and Reddy, C. R. K. (2014). Seaweed-microbial interactions:Key functions of seaweed-associated bacteria. FEMS Microbiology Ecology, 88(2):213–230.[Staufenberger et al., 2008] Staufenberger, T., Thiel, V., Wiese, J., and Imhoff, J. F. (2008). Phy-logenetic analysis of bacteria associated with Laminaria saccharina. FEMS Microbiology Ecology,64:65–77.[Steinberg, 2002] Steinberg, P. D. (2002). Minireview: Chemical Mediation of Colonziation ofSeaweed Surfaces. Journal of Phycology, 629:621–629.[Steinberg et al., 1997] Steinberg, P. D., Schneider, R., and Kjelleberg, S. (1997). Chemical de-fenses of seaweeds against microbial colonization. Biodegradation, 8(3):211–220.[Stewart and Carpenter, 2003] Stewart, H. L. and Carpenter, R. C. (2003). The effects of mor-phology and water flow on photosynthesis of marine macroalgae. Ecology, 84(11):2999–3012.[Stratil et al., 2014] Stratil, S. B., Neulinger, S. C., Knecht, H., Friedrichs, A. K., and Wahl, M.(2014). Salinity affects compositional traits of epibacterial communities on the brown macroalgaFucus vesiculosus. FEMS microbiology ecology, 88(2):272–9.[Tait and Schiel, 2011] Tait, L. W. and Schiel, D. R. (2011). Dynamics of productivity in natu-rally structured macroalgal assemblages: Importance of canopy structure on light-use efficiency.Marine Ecology Progress Series, 421:97–107.[Thomas et al., 2008] Thomas, T., Evans, F. F., Schleheck, D., Mai-Prochnow, A., Burke, C.,Penesyan, A., Dalisay, D. S., Stelzer-Braid, S., Saunders, N., Johnson, J., Ferriera, S., Kjelleberg,S., and Egan, S. (2008). Analysis of the Pseudoalteromonas tunicata genome reveals propertiesof a surface-associated life style in the marine environment. PLoS ONE, 3(9):1–11.[Tujula et al., 2010] Tujula, N. a., Crocetti, G. R., Burke, C., Thomas, T., Holmström, C., andKjelleberg, S. (2010). Variability and abundance of the epiphytic bacterial community associatedwith a green marine Ulvacean alga. The ISME journal, 4(2):301–311.[Vega Thurber et al., 2012] Vega Thurber, R., Burkepile, D. E., Correa, A. M. S., Thurber, A. R.,Shantz, A. A., Welsh, R., Pritchard, C., and Rosales, S. (2012). Macroalgae Decrease Growthand Alter Microbial Community Structure of the Reef-Building Coral, Porites astreoides. PLoSONE, 7(9):1–10.65Bibliography[Venables and Ripley, 2002] Venables, W. N. and Ripley, B. D. (2002). Modern Applied Statisticswith S. Springer, New York, fourth edition. ISBN 0-387-95457-0.[Wada and Hama, 2013] Wada, S. and Hama, T. (2013). The contribution of macroalgae to thecoastal dissolved organic matter pool. Estuarine, Coastal and Shelf Science, 129:77–85.[Warnes et al., 2016] Warnes, G. R., Bolker, B., Bonebakker, L., Gentleman, R., Liaw, W. H. A.,Lumley, T., Maechler, M., Magnusson, A., Moeller, S., Schwartz, M., and Venables, B. (2016).gplots: Various R Programming Tools for Plotting Data. R package version 3.0.1.[Weinberger, 2007] Weinberger, F. (2007). Pathogen-induced defense and innate immunity inmacroalgae. Biological Bulletin, 213(3):290–302.[Weinberger et al., 1999] Weinberger, F., Friedlander, M., and Hoppe, H.-g. (1999). OligoagarsElicit a Physiological Response inGracilaria Conferta (Rhodophyta). Biological Bulletin, 35:747–755.[Wilson et al., 2010] Wilson, S. K., Depczynski, M., Fisher, R., Holmes, T. H., O’Leary, R. A.,and Tinkler, P. (2010). Habitat associations of juvenile fish at Ningaloo Reef, Western Australia:The importance of coral and algae. PLoS ONE, 5(12):1–8.[Yakimov et al., 2003] Yakimov, M. M., Giuliano, L., Gentile, G., Crisafi, E., Chernikova, T. N.,Abraham, W.-R., Lunsdorf, H., Timmis, K. N., and Golyshin, P. N. (2003). Oleispira antarcticagen. nov., sp. nov., a novel hydrocarbonoclastic marine bacterium isolated from Antarctic coastalsea water. International Journal of Systematic and Evolutionary Microbiology, 53:779–785.[Yoon et al., 2007] Yoon, J., Matsuo, Y., Matsuda, S., Adachi, K., Kasai, H., and Yokota, A.(2007). Rubritalea spongiae sp. nov. and Rubritalea tangerina sp. nov., two carotenoid- andsqualene-producing marine bacteria of the family’ Verrucomicrobiaceae within the phylum Verru-comicrobia’, isolated from marine animals. International Journal of Systematic and EvolutionaryMicrobiology, 57:2337–2343.[Yoon et al., 2008] Yoon, J., Matsuo, Y., Matsuda, S., Adachi, K., Kasai, H., and Yokota, A.(2008). Rubritalea sabuli sp. nov., a carotenoid- and squalene-producing member of the familyVerrucomicrobiaceae, isolated from marine sediment. International Journal of Systematic andEvolutionary Microbiology, 58:992–997.[Yurkov and Beatty, 1998] Yurkov, V. and Beatty, J. T. (1998). Aerobic Anoxygenic PhototrophicBacteria. Microbiology and Molecular Biology Reviews, 62(3):695–724.[Zaneveld et al., 2016] Zaneveld, J. R., Burkepile, D. E., Shantz, A. A., Pritchard, C. E., McMinds,R., Payet, J. P., Welsh, R., Correa, A. M. S., Lemoine, N. P., Rosales, S., Fuchs, C., Maynard,J. A., and Thurber, R. V. (2016). Overfishing and nutrient pollution interact with temperatureto disrupt coral reefs down to microbial scales. Nature Communications, 7(May):11833.[Zozaya-Valdes et al., 2015] Zozaya-Valdes, E., Egan, S., and Thomas, T. (2015). A comprehen-sive analysis of the microbial communities of healthy and diseased marine macroalgae and thedetection of known and potential bacterial pathogens. Frontiers in Microbiology, 6(FEB):1–9.66

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0351980/manifest

Comment

Related Items