Diversity of Picoeukaryotes at an Oligotrophic Site Off the Northern

Diversity of Picoeukaryotes at an Oligotrophic Site oﬀ the Northern Red Sea Coast

Thesis by Francisco Jos´eAcosta Espinosa

In Partial Fulﬁllment of the Requirements For the Degree of Master of Science

King Abdullah University of Science and Technology

Thuwal, Kingdom of Saudi Arabia

May, 2012 2

The thesis of Francisco Jos´eAcosta Espinosa is approved by the examination committee.

Committee Chairperson: Dr. Ulrich Stingl

Committee Member: Dr. Christian Voolstra

Committee Member: Dr. Timothy Ravasi 3

Francisco Jos´eAcosta Espinosa

This work is licensed under a Creative Commons Attribution-NonCommercial 3.0

Unreported License (http://creativecommons.org/licenses/by-nc/3.0/). 4

ABSTRACT

Diversity of Picoeukaryotes at an Oligotrophic Site oﬀ the

Northern Red Sea Coast

Francisco Jos´eAcosta Espinosa

Picoeukaryotes are protist ≤3 µm belonging to a wide diversity of taxonomic groups, and they are an important constituent of the ocean microbiota, performing essential ecological roles in marine trophic chains and in nutrient and carbon budgets. Despite this, the true extent of their diversity is currently unknown, and in the last decade molecular surveys have uncovered an substantial number of previously unknown groups from all taxonomic levels. No studies on this group have been done so far on the Red Sea, a unique marine environment characterized by oligotrophic conditions and high irradiance, salinity and water temperature. We sampled the surface waters of a site near the northern Red Sea coast, and analyzed the picoeukaryotic diversity through the construction of PCR clone libraries using the 18S ribosomal gene. The community captured by our library is dominated by three main groups, the alveolates (32%), chlorophytes (32%) and Stramenopiles (20.55%). Members of Radiolaria, Cercozoans and Haptophyta were also found, although in low abundances. Photosynthetic organisms are especially diverse and abundant in the sample, with heterotrophic organism mostly composed by the mostly parasitic novel alveolates and 5 bacterivorous stramenopiles. Novel clades were detected among the Novel Alveolates- II and the photosynthetic stramenopiles taxa, which suggests that they may be part of a number of groups unique to the basin and adapted to the high salinity and temperature conditions. This is the ﬁrst study done on the Red Sea focusing on the diversity of the complete picoeukaryotic fraction, and provides a stepping stone in the characterization of the picoeukaryotic component of the microbial diversity of the basin. 6

ACKNOWLEDGMENTS

For the present thesis, I would like to extend my most sincere thanks to my advisor Dr. Ulrich Stingl, for supporting me on this project, Dr. David Ngugi, for his help and advice during the course of the research, and to the KAUST Genomic Core Facility, for the processing and sequencing of the clone libraries. I would also like to thank all other collaborators and friends of the Red Sea Research Center, which have helped and supported me during the duration of this Master degree, particularly to my colleagues on the Laboratory for Microbial Ecology - working with them has been a pleasure, and has provided me with an amazing research environment. 7

TABLE OF CONTENTS

Examination Committee Approvals 2

Abstract 3

Acknowledgments 6

List of Illustrations 9

List of Tables 10

1 Introduction 11 1.1 Main marine picoeukaryotic groups ...... 17 1.1.1 Alveolates ...... 17 1.1.2 Stramenopiles ...... 21 1.1.3 Haptophytes ...... 24 1.1.4 Chlorophyta ...... 25 1.1.5 Rhizaria ...... 26 1.1.6 Excavata ...... 28 1.1.7 Unikonta ...... 28 1.1.8 Other taxa ...... 29 1.2 The Red Sea environment ...... 31

2 Materials and Methods 35 2.1 Sampling ...... 35 2.2 PCR ampliﬁcation and puriﬁcation ...... 35 2.3 Clone library construction ...... 37 2.4 Community and phylogenetic analysis ...... 37

3 Picoeukaryotic community on the Red Sea surface waters 39 3.1 Community sampling ...... 39 3.1.1 18S gene sequences ...... 39 3.1.2 V3 region sequences ...... 46 3.1.3 Other clone libraries ...... 51 3.1.4 Suitability of diﬀerent regions of the SSU gene ...... 52 3.2 Picoeukaryotic diversity ...... 55 3.2.1 Alveolata ...... 55 8

3.2.2 Chlorophyta ...... 67 3.2.3 Stramenopiles ...... 72 3.2.4 Rhizaria ...... 84 3.2.5 Rare taxa ...... 88 3.2.6 Metazoa ...... 90

4 Summary 91 4.1 Conclusions ...... 95 4.2 Future Research Work ...... 95

Bibliography 97

Appendices 122

A Sampling site physical and environmental characteristics 122

B Papers Submitted and Under Preparation 123 9

LIST OF ILLUSTRATIONS

1.1 Diagram of eukaryotic super-groups ...... 18

3.1 Community composition of the 1.2 - 5.0 µm fraction with the 18S gene 40 3.2 Rarefaction curves for 18S gene sequences ...... 45 3.3 Community composition of the 1.2 - 5.0 µm fraction using the 18S v3 region ...... 47 3.4 Rarefaction curves for v3 region sequences ...... 50 3.5 Eukaryotic community composition of 18S subregions for all filter sizes 51 3.6 OTU diversity for different SSU regions ...... 53 3.7 SSU rDNA phylogenetic tree for Dinoflagellates ...... 56 3.8 SSU rDNA phylogenetic tree for Novel Alveolates group I ...... 58 3.9 SSU rDNA phylogenetic tree for Novel Alveolates group II (1) . . . . 60 3.10 SSU rDNA phylogenetic tree for Novel Alveolates group II (2) . . . . 61 3.11 SSU rDNA phylogenetic tree for Ciliata ...... 65 3.12 SSU rDNA phylogenetic tree for the Chlorophyta ...... 68 3.13 SSU rDNA phylogenetic tree for phototrophic Stramenopiles . . . . . 73 3.14 SSU rDNA phylogenetic tree for heterotrophic Stramenopiles . . . . . 79 3.15 SSU rDNA phylogenetic tree for Rhizaria ...... 85 3.16 SSU rDNA phylogenetic tree for Haptophytes and unknown sequences 88

A.1 Location of sampling site ...... 122 10

LIST OF TABLES

2.1 Primers used on the study ...... 36

3.1 OTU composition for the complete 18S clone library ...... 42 3.2 OTU composition for the 18S v3 region clone library ...... 48 11

Chapter 1

Introduction

Unicellular eukaryotic organisms are a ubiquitous and important component of marine ecosystems, where they interact intimately with bacterial and archaeal components in trophic and biogeochemical systems that shape the ocean’s biosphere. Yet our current knowledge of their ecological roles and diversity in these ecosystems is surprisingly low, especially compared to the other domains of life, even after more of a decade of continuous surveys on all kinds of environments. We don’t really know the true extent of protistan diversity, or even if all major groups have been discovered already[1]; the recent discoveries of novel groups, even up to the phylum level[2], whose most basic ecological characteristics remain unknown only serves to reinforce the notion that eukaryotes are one of the lesser known components of marine ecosystems. These organisms, as well as all other parts of the marine microbiota, for research purposes are generally divided into size-based classes, that range from the microplankton (20 - 200 µm) to the femtoplankton (≤0.2 µm). This classiﬁcation obeys to both practical and ecological reasons, facilitating the sorting of cells through ﬁltration and other size-sensitive methods while at the same time grouping organisms which share similar ecological niches based on water viscosity, nutrient availability and area/vol- ume cell ratio[3]. 12

Among eukaryotes, organisms in the picoplankonic fraction - the smallest class for this domain, comprising all protist of 3 µm and smaller[4, 3] - are an important and often overlooked component of marine ecosystems, where they perform a wide range of ecological roles and functions. Their presence has been documented in practically all marine ecosystems, from surface waters[5, 6, 7, 8] to extreme benthic environments such as thermal sea vents[9, 10] and hypersaline brines[11]. Despite being the most abundant eukaryotic organisms on Earth[6], their true diversity remains unknown, with each new study revealing new groups and populations of these organisms[3], a reflection of the general undersampling of this domain in marine environments. Photosynthetic picoplankton is of extreme importance to the ocean’s energy and nutrient cycles, being the size fraction with the highest productivity[12], and picoeukaryotes have been found to constitute one of its main components, contributing up to 50% of the total Chlorophyll a in coastal waters[13]. In some environmental conditions they are the main source for primary production in surface waters, such as the North Atlantic[14], Celtic Sea coastal areas[15] and the English Channel[16]. Additionally, despite being usually outnumbered in oligotrophic environments by photosynthetic bacteria such as Prochlorococcus and Synechococcus in terms of cell number, their contribution to carbon fixation, and thus to global ocean carbon budgets, is larger due to its high cell-specific rate for carbon uptake[14]. Picosized protist are the most abundant component of the eukaryotic phytoplankton, which is estimated to have a total C biomass of 80 million tonnes for the Atlantic Ocean as a whole, approximately one third of the total estimated C biomass of this ocean[17]. Heterotrophic protists such as flagellated cells belonging to the Alveolata and Stramenopiles are also a vital link for the recycling of nutrients from the bacterial compartment to other trophic levels, integrating what is known as the microbial loop[18]. In this process, Dissolved Organic Material (DOM), which is lost in the higher levels of the trophic chain, is incorporated by bacteria and archaea organisms; 13 picoeukaryotic bacterivores close this loop by ingesting these organisms, and since in turn they can be preyed upon by bigger protist, they make this fixed carbon and nutrients available again to higher tropic levels[19]. Interaction between picoeukayotes and other microbes has further ecological implications, with bacterial abundance and community composition being strongly in- fluenced by these organisms[20]. The ”top-down” model, which postulates regulation of bacterial populations through protistan grazing, has been shown to play a role on marine systems, being particularly strong in oligotrophic ones[19]. Besides primary production and bacterivory, picoeukaryotes can also influence ecological relationships through parasitism[21] and symbiosis[3] While the presence of these organisms was detected as early as 1951[5], the lack of distinctive morphological characteristics prevented a more thorough study; our understanding of the nature and diversity of these taxa improved dramatically with the advent of molecular techniques such as Polymerase Chain Reaction (PCR), Re- striction Fragment Length Polymorphism (RFLP) and similar others, which allowed researchers to directly study the genetic sequences present in environmental samples. The analysis of environmental samples using molecular techniques, first pioneered by Giovannoni[22] for Bacterial and Archaea organisms has been applied to the study the unicellular eukaryotic communities[6, 23, 24, 1], uncovering a previously unsuspected diversity that extends to the higher taxonomic levels, including the discovery of novel phyla[2, 25]. These observations are at odds with previous theories, that assumed a protistan community composed by species with virtually limitless dispersal capability, resulting in a high local diversity but a relatively small total number of species[26]. These studies use conserved marker genes like those contained in rRNA operon such as the small subunit (18S) gene, the large subunit (23S)[27] and the internal transcribed spacers (ITS)[28], as well as mitochondrial[29] and plastid[30, 25] genes. Other technologies that have been used to explore the picoeukaryotic diversity in- 14 clude microscopy[31, 32], flow cytometry[14, 33, 17, 34] and Fluorescent In Situ Hybridization(FISH)[35, 36, 13, 11] as alternatives to or complementing molecular approaches. Although approaches based on ribosomal DNA (rDNA) genes are the most widely used for the study of microbial communities in environmental samples, these methodologies have a number of inherent problems. Amplification of these genes requires the use of primer molecules that target their conserved regions, but these regions differ among the different taxa, and very few of them, if any, span the whole breadth of eukaryotic diversity. Due to this, any primer, even those considered as universal, will only capture a fraction of the diversity present in a sample, and the use of different primer pairs can yield different dominant groups[34]. Diverse strategies have been used to address this, including the use of multiple primer pairs[34, 37, 38] and degen- erate primers[39], but the fact remains that a fraction of the diversity will probably not be captured. Additionally, since primer design is done through the analysis of already sequenced molecules, there is the possibility that any novel organisms with a high number of divergences in these conserved regions will be excluded from community analyses. Other problems are the bias towards groups that are more easily amplified, the production of hybrid sequences[40] and contamination by extraneous amplifiable extracellular DNA[41]. Even sampling and DNA extraction methodologies have been shown to introduce biases on the captured microbial community[42]. Another issue to consider is that many clone libraries, done by amplification of rDNA, isolation of different sequences and sequencing by the methodology established by Sanger[43] or variations thereof, are limited in size due to costs and the labor-intensive process required. As a consequence of this many environments remain undersampled in terms of their eukaryotic microbiota, as evidenced by the rarefaction values determined for multiple studies. Diverse technologies and methods have been implemented to circumvent these issues, such as 454 pyrosequencing[1, 44, 45], which, 15 along with several other next-generation sequence techniques, has been used to increase the average sampling depth of PCR studies by multiple orders of magnitude. There are two main drawbacks to this method, the first being that the amplified fragments are drastically size limited, being usually smaller than 300 bp, a size which in many cases prevents an accurate phylogenetic analysis of the resulting clones, and that may mask part of the total diversity. However, new advances in these technologies are starting to allow sequencing of longer amplicons, with an average length of 700 bp, which resolves this issue. The second drawback is the intrinsic error rate in pyrosequencing, which is substantially higher than in Sanger sequencing[46] and that demands stringent controls for sequence quality in order to remove all ambiguous, in- complete, erroneous or generally low-quality sequences. Other alternatives to increase sampling depth of rDNA studies include the use of flow cytometry sorting coupled with single genome amplification[34], and analysis of multiple clone libraries[47]. Besides these issues, there are various unique challenges to the Eukarya domain. Most eukaryotic organisms have multiple copies of the 18S gene, producing a dis- torted perspective in diversity and community composition; the number of copies varies considerably between different taxa with some groups, like ciliates, which have as much as 10,000 copies per genome[48], causing them to be overrepresented in clone libraries from environmental samples and mask organisms with less copies. Compar- ison with single genome amplification methods have shown that the number of 18S gene copies is one of the major factors in determining clone library composition[34], and approaches focused on reverse transcription and sequencing of RNA molecules has been deemed to provide a better estimate for the significance of the diverse marine protist groups in a given location[41]. While the general consensus is that all rRNA operons in any given organism are homogeneous, due to selective pressure to ensure the homogeneity of the products they encode[49], it is now known that in at least some species this does not hold true. 16

Heterogeneity in rRNA genes have been shown is members of the Apicomplexa, where multiple copies can differ in up to 8% of their nucleotides[50], and in certain metazoan species[51]. The mechanisms through which this heterogeneity is maintained are poorly known, and the phenomena does not seem to be widely spread among eukaryotes[52, 53, 54], but given that no comprehensive studies have been done to assess its extent, it still remains a reason for caution when interpreting 18S diversity estimations. More fundamental concerns have also been raised, such as the fact that the conservation of the 18S gene appears to be more pronounced that in other parts of the genome, a phenomena more prominent in unicellular species. The region may be too conservative for accurate taxon diversity estimations, not evolving at a high enough pace to resolve species boundaries[55]. This is reinforced by studies showing that eukaryotic organisms with almost identical 18S sequences (at 99% similitude) show divergences of almost 30% in their complete genome[56] - which would imply that they may have completely different ecological roles and metabolic functions. With this in mind, it has become clear in the last decades that the traditionally employed concept of morphospecies is not applicable for protists, and especially so in picoeukaryotes. Many traditional species are now recognized as complexes of closely related cryptic species that may differ in the functions they play on ecosystems and niche and habitat preferences, such as Micromonas pusilla[57], Amoebophrya sp.[58] and Tetrahymena tropicalis[59], among others. These organisms can vary in terms of their preferred environmental conditions, their prey or host preferences, a may even differ in their physiological characteristics[60]. Other studies have found that even cryptic species with minute differences in their ribosomal sequences may experience reproductive isolation, and present subtle ultrastructural differences which cause be- havioral and ecological differences between them[61], supporting the possibility that these species complexes rose from sympatric speciation to adapt to different, specific 17

habitats.

1.1 Main marine picoeukaryotic groups

Eukaryotic systematics have always been the source of much debate, due to numerous factors: morphologically, ultrastructural characteristics are hard to assess due to the diﬃculty to determine homology among characteristics. While the use of molecular markers obviates this problems, this method presents its own complications due to the high number of mutations present in eukaryotic genomes, especially to resolve deep evolutionary relationships, a problem exacerbated by the sparse number of sequences retrieved for certain taxonomic groups. Use of multiple-gene phylogenies has been deemed as the most accurate way to infer an accurate phylogenies[62], but so far no consensus has been reached regarding any high-level taxonomic grouping. It is currently believed that there are 6 eukaryotic super-groups[63]: Chomoalveolata, Ar- chaeoplastida, Rhizaria, Excavata, Amoebozoa and Opisthokonta; these groups and the main phyla in them can be seen on Figure 1.1.

1.1.1 Alveolates

Alveolates were among the fist unicellular organisms recognized and classified, and they constitute one of the main groups of protists. They are characterized by the presence of membrane-bound cavities under the cell membrane - the alveoli - which are though to function in mechanisms of motility, defense and formation of extracellular structures[64]. This groups includes multiple phyla with many divergent lifestyles, including the dinoflagellates, which include phototrophic, heterotrophic and numerous mixotrophic species, the mostly heterotrophic ciliates, the parasitic Apicomplexans and Perkinsozoa and a number of novel phyla, collectively know as Novel Alveolates (NA). They are probably the most common protistan group on marine surveys, con- 18

Amoebozoa Excavata Acanthamoebida

Entamoebida

Mycetozoa Ichtyosporea Heterolobosea

ParabasaliaFornicata Fungi Euglenozoa Jakobida Choanoflagellata

Metazoa

Opisthokonta

StramenopilesAlveolata Cryptophyceae Haptophyceae

Cercozoa

Foraminifera Chloroplastida

Glaucophyta ArchaeplastidaRhodophyceae

Radiolaria Chromoalveolata

Rhizaria

Figure 1.1: A diagrammatic tree showing the currently acknowledged eukaryotic super-groups. The relationship amongst most of the super-groups, and in many cases among the different phyla within them are so far unresolved. stituting up to 57% of all sequences[65]. Ciliates are among the most well-identified and characterized protists, in part because their size and morphology make classification and descriptions relatively easy, compared to other taxa. However, they fall in a size range much greater than that considered for picoeukaryotes[7], with the smallest identified ciliates measuring approximately 7 µm[66] and are thus rarely found in these analyses. The same situation occurs with the classic dinoflagellates, which despite appearing occasionally in picoplanktonic surveys are consistently larger than 3 µm[4]; their presence is most likely explained by filtration artifacts, or contamination by dissolved DNA[65]. The presence of NA in marine environments has been extensively documented 19 during the last decade in virtually every marine survey[6, 67, 68, 7, 69]. NA are exclusively marine and have been identified as encompassing the Syndiniales, a group of parasitic organisms that form a basal branch of the Dinophyceae. Originally considered one of several dinoflagellate orders, syndinides are now seen as being part of several phylogenetically distinctive and extremely diverse groups, representing more than half of all sequences of Dinophyceae recovered in marine surveys[47]. First describe by Diez et al[6], so far five main groups of novel alveolates have been described[47], each one composed by numerous clades whose number is expected to increase as more environments are sampled. It is interesting to note that the presence of many of these clades seem to be not determined by environmental conditions - in many of them sequences from tropical and polar waters, sediments and anoxic water masses seem to coexist[10, 23, 70, 37, 71, 68, 69], suggesting a wide distribution and a high capacity for adaptations to local conditions by each clade. Organisms in NA-I, although present in all marine environments[67, 10, 9, 23], are proportionally more important in the aphotic zone, and are the most abundant alveolate group in hydrothermal vent ecosystems[47]. The genera identified within this group are Ichthyodinium and Duboscquella, both parasites from marine ecosystems. Duboscquella is an intracellular parasite of many ciliate organisms, particularly tintinnids, although some species can also parasite dinoflagellates[72]. Characterized members of this genus present small, diflagellated cells of 4 - 6 µm in length. The high abundance of this group in marine samples has been attributed to the fact that during reproduction it releases hundreds of small-sized dinospores from the host, which range from 2 to 6 µm in size[73] and which may constitute the source of most the NA-I genetic material in clone libraries[72, 47]; this also explains the absence of these organisms in traditional planktonic surveys. Ichthyodinium presents a similar situation, being a parasite of fish larvae, including species of economic importance such as the atlantic sardine Sardina pilchardus[74]. 20

Among the novel alveolate groups, NA-II is by far the one with the biggest diversity present, with 44 described clades[47]. The only described genus within this group so far is Amoebophrya, with all known species being intracellular parasites of dinoflagellates, including other Amoebophrya species. While there are only a few described species within this family, the ample differences between strain infectivity and host selectivity, as well as the wide genetic diversity detected in this group probably indicate that these denominations actually represent species complexes[75, 47]. As with the parasitic species of NA-I, the presence of Amoebophrya in picoplanktonic surveys has been attributed to the presence of dinospores or cysts[58], since infective cells are bigger than 10 µm even during the initial phases of host infection[76]. Groups NA-III, -IV and -V are much less common, but like the others are found in many diverse environments. NA-III is composed exclusively of environmental sequences from both surface[23] and deep water masses, including sulfidic, anoxic environments[77]. NA-IV includes the genera Syndinium and Hematodinium, both parasites of diverse crustacean species. Dinospores from these organisms are relatively big, around 7 - 12 µm[78] - a fact that may explain their absence in many studies, where samples are usually filtered before DNA extraction. NA-V is also composed by environmental sequences, but it seems to be restricted to samples taken from the euphotic zone[47]. All the described genera within this the novel alveolates are obligate parasites, suggesting that environmental sequences probably also belong to novel parasitic organisms. Given their abundance and diversity, which most likely is reflected in a wide range of hosts, they probably play an important and poorly understood role on microbial trophic interactions and community composition[65, 21]. As for the other phyla within Alveolata, they appear to be present in low abundance in all marine ecosystems studied up to date. Perkinsids are intracellular parasites for many metazoan and dinoflagellate species[79], but they are rarely detected 21 in environmental samples - exceptions include Perkinsus-like sequences detected in hydrothermal sediments[10], abyssal plains[80] and the English Channel[68].

1.1.2 Stramenopiles

Stramenopiles are an exceedingly diverse group, with members that include some of the most common phototrophic protist, the diatoms, free living heterotrophic flagellates, mold-like parasites such as the Oomycetes and others. The only shared characteristic that unifies this phylum is the presence of tripartite hairs in their flagella and tubular cristae in their mitochondria[63, 7]. The most common groups that compose the picoplankton include those phyla within the Ocrophyta, a large radiation of mostly phototrophic organisms that seem to constitute a monophyletic group[81], and several basal heterotrophic groups, among them most of a number of novel taxa represented by environmental clones, collectively called Marine Stramenopiles (MAST). The picophotosynthetic groups of Stramenopiles include the Pelagophyceae, Boli- dophyceae, Eustigmatophyceae, Chrysophyceae, Bacillariophyceae, Dictyochophyceae and Pinguiophyceae, all of which have cultured species with picosized cells[4] in the range between 1.2 - 6 µm [4, 82, 83, 84]. Pelagophyceae are among the most common organisms in the picophytoplankton in both the Atlantic and Pacific oceans[32], at least when determined using HPLC pigment signatures; despite this, they are not abundant in 18S clone libraries[4]. Regardless, they constitute an important group in marine habitats, and some species can form important blooms in coastal areas[85, 86]. Chrysophytes are an occasional abundant group in plastid[30] and rRNA[41] clone libraries, and an important number of novel groups has been identified in the phylum[87]. The other groups also contain phototrophic planktonic organisms, but they seem to be less abundant in pelagic systems, which is reflected in their proportions on molecular surveys[4, 3]. Diatoms are found in many surveys that employ small filter sizes, a phenomena 22 that can be partly explained by the approximately 20 described species of small diatoms, such as Minidiscus and Skeletonema; the former is distinctly abundant in antarctic waters[88]. Other possible reasons for this include the possible existence of undescribed picodiatoms, the presence of diatoms that can pass through the filter pores due to their morphology, and gametes or cysts from larger organisms[4]. Of the cultured heterotrophic groups, Bicosoecids are bacterivores found mostly in pelagic environments[89, 90] , although they have been found sporadically in anoxic environments such as the sulfidic water layers of the Black Sea[91]; groups associated with them have also been found in hypersaline brine pools[11]. Labyrinthulomycetes is another group with at least two picosized species, Labyrinthuloides minuta and Thraustochytrium multirudimentale[3], which is occasionally found in picoeukaryotic surveys from diverse habitats[7, 92, 67]. Known labyrinthulids are osmotrophs[93] feeding through the absorption of dissolved organic compounds. Overall, the highest diversity of marine stramenopiles falls within a series of taxa collectively called Marine Stramenopiles (MAST). So far there are 19 MAST clades[35, 94, 38, 95] that have been identified in many different environments, where they probably fulfill very different ecological roles. These groups are highly abundant, comprising up to 20% of all sequences in pelagic 18S clone libraries[7]. Some, like MAST-1, MAST-3, MAST-4 and MAST-7 seem to constitute common pelagic organisms. MAST-1 was the first of these groups to be formally described[35] and it is distributed globally, with three internal clades with varying cell sizes that range from 4 to 8 µm[94]. MAST-4 seems to be particularly abundant, having been found in many diverse epipelagic environments such as the Mediterranean, the North Atlantic, equatorial Pacific[35], the Indian ocean and the Drake Passage in the South- ern Ocean[96]; so far, its only limits in distribution seem to be related to temperature, being absent in all samples colder than 5◦C[96]. MAST-6 is related to MAST-4, but it is less abundant and seems to play a minor role in marine ecosystems; however, mem- 23 bers of this taxa constitute an important seasonal component of the nanoflagellates in the Baltic Sea[36]. In contrast, MAST-2 seem to be found in multiple environments, but always at low abundances[97]; their sizes appear to vary with temperature, with cell size increasing with lower temperatures[94]. MAST-7 and MAST-8 are also planktonic groups that appear to be phylogenetically related to MAST-4, which would suggest that they may share some physiological and ecological characteristics; however, members of MAST-7 have been found in hypersaline, extreme environments, suggesting that it’s distribution may be wider than previously thought[11]. Others groups, such as MAST-9 and -12 are found in anoxic environments, indicating that they probably are anaerobic or anoxia-tolerant. MAST-14, -15 and -16 are also related to MAST-12 and the Bicosoecida, and they seem to share their habitat preferences, since they have only been detected in the anoxic waters of the Cariaco Basin[95]. Organisms belonging to MAST-17, -18 and -19 were found in the same samples, branching within the Thraustrochytrids and Labyrinthulids. MAST-5 was initially constituted by a single sequence obtained in the Antarctic polar front[67], which appears to form one of the earlier branchings in stramenopile evolution[97]; posteriorly other members of this groups were found in anoxic water masses[95], suggesting that this clade may be more predominant in this extreme environments. Similarly to MAST-1, MAST-3 it is an abundant and ecologically important group that has been found in both surface[35, 97, 69, 57] and deep-sea anoxic[37, 95, 41] environments. It is also the only MAST group with a cultured representative, the species Solenicola setigera, an ubiquitous 4 - 7 µm flagellate that appears to function as an important grazer of diatoms, and which has been recently shown to be a member of this clade[98]. 24

Regarding the other uncultured groups, MAST-1, -2, and -4 have been detected in mixed protistan assemblages in seawater incubations, and studies with Fluorescent Labeled Bacteria (FLB) and live bacteria show that they are also bacterivorous cells and moreover, than different strains displayed different prey preferences[35, 94, 99, 100]. In contrast, MAST-6 organisms detected in the Baltic Sea were found to be algivorous[36], showing that even phylogenetically related groups with similar morphologies can occupy very different trophic roles.

1.1.3 Haptophytes

Among the haptophytes, picoprymnesiophytes are a major component of the plankton that has been traditionally ignored, with only six described species[4, 3]. Despite the abundant detection of 19’-hexanoyloxyfucoxanthin, a pigment characteristic of this group, in the Atlantic and Paciﬁc[32], this group was considered not a major component of the picoeukaryotic communities. However, recent studies with haptophyte-speciﬁc primers have shown them to be far more widespread that initially thought; they form a complex and abundant group of organism distributed worldwide[101], being remarkably abundant in high latitudes, where they constitute more than 50% of the picoplanktonic biomass[102]. Clone libraries of chloroplastic ribosomal genes do seem to support this view, showing haptophytes to conform almost 45% of all sequences on a year long time-series on the Gulf of Naples[103]. Additionally, despite being outnumbered by prokaryotic cells by orders of magnitude, their increased size and higher growth rates result in primary production rates similar to those of Prochlorococcus[102]. Part of the ecological suc- cess of this group may be due to an hypothesized mixotrophic lifestyle[101], giving them an advantage over purely heterotrophic or autotrophic cells. 25

1.1.4 Chlorophyta

Of all eukaryotic photosynthetic organisms in marine environments, Chlorophytes, and more specifically the class Prasinophyceae, are among the most abundant and well studied[39], with numerous cultured representatives. The most important group within this class are the Mamiellales, described as Clade II of the Chlorophyta by Guillou[104], that is composed by the genera Micromonas, Bathycoccus and Ostre- ococcus. These organisms represent the most abundant eukaryotic photosynthetic component of the plankton, as determined by cultures, microscopic and genetic surveys. Micromonas pusilla seems to be distinctly important in ecological processes, being a dominant species all year round and potentially shaping the entire picoplanktonic community[13] in places as dissimilar as the English Channel[13], the Mediterranean[39], the Indian ocean[89] and the entire Arctic basin[105]. Within this species we can find several phylogenetically distinctive clades, suggesting that M. pusilla is actually a cryptospecies complex that encompasses numerous related but distinct organisms[57], which may present different ecological needs and adaptations to specific ecological conditions, similarly to the ecotypes detected in other phytoplanktonic organisms suchs as Prochlorococcus[106]. In fact, analyses show that M. pusilla may have diverged as early as the late Cretaceous, making it the oldest known cryptoespecies complex[107]. While there is no complete consensus for the evolutionary mechanisms that sustain the high levels of microdiversity present in Micromonas clades, it has been argued that they may present an evolutionary mechanism against viral infections. M. pusilla, similarly to other high abundant marine organisms, has been shown to be susceptible to lytic viruses, which can dramatically change the abundance of their populations[108]; therefore this intraspecific diversity has been suggested as a mechanism of natural populations to cope with these events, by allowing the survival and proliferation of virus-resistant genotypes[13, 109]. 26

Another member of the Mamiellales, Ostreococcus is the smallest known eukaryote, with an average width of 0.8 µm[110], and it is found in all oceans worldwide, usually at low abundances[13, 39]. It is believed that its maximum abundance may lie within the Deep Chlorophyll Maximum (DCM), which is not always covered in all studies[111]. Similarly to M. pusilla, the different phylogenetic clades of this species also present differential morphology, physiology and growth conditions[28]. Other groups within the Prasinophyceae include the Pseudoscourfieldiales, whose only described species, Pycnococcus provasolii, is particularly abundant in the DCM in oligotrophic waters[112]; the Prasinococcales, sporadically detected in surveys but which can develop important seasonal populations in coastal areas[13]; the Pyrami- monadales, a relatively diverse and widespread group that has been found in the Mediterranean[113], Blanes Bay[7], the North Sea[8] and the Arctic[114]; and Clade VII, a group composed of environmental sequences[104] that still lacks a formal description. There are two other classes within this group, the first of them being the Pedino- phyceae, a small group that has not been recorded since two decades ago where it was described in the Atlantic Ocean[115], and for which no genetic information is available; it is unknown if this is an extremely rare phyla that has gone undetected so far in all available molecular surveys or if its part of the uncultured diversity detected within the chlorophytes[113]. The other class, the Trebouxiophyceae, contains numerous cultured species, but there is no data on their actual abundance in marine waters[4].

1.1.5 Rhizaria

The rhizarians constitute an extremely heterogeneous group, comprising such diverse organisms as the chlorarachniophyta, ﬂagellated cercozoans, foraminiferans and radiolarians[63], with the only common morphological attribute being the presence of 27

fine pseudopodia. These previously unrelated groups were shown by molecular studies to constitute a single monophyletic clade, first proposed by Cavalier-Smith[116], and later confirmed by genetic and ultrastructural analysis[117]. Most rhizarians fall outside the picoplankton size range, with the exception of radiolarians and cercozoans, which are well represented in clone libraries of this size fractionMassana2011a. Rhizarians are abundant in deep sea environments, especially the sea floor, where the high diversity and abundance of foraminiferan taxa have been used as markers for geological dating of sediments for many decades[118]; however this group is rarely found in community analyses using ribosomal genes, due to their high mutation rate and unique 18S structure which enforces the use of specific primers[119]. Radiolarians are a large group of rhizarians characterized by an organic capsular wall surrounding the cell, axopodia and sometimes siliceous or sulphate skeletons[63]. The phylum contains the Polycystinea, the Acantharia and the Taxopodida taxa, although the relationship between these groups has changed due to recent phylogenetic analyses, showing them to be paraphyletic[120, 121]. While there are no described picoradiolarians, several sequences from this taxa have been isolated in surveys from diverse marine systems, conforming at least 19 novel environmental clades[69, 89, 33, 95], and while some of these may be the result of contamination by small reproductive stages or dissolved DNA[65], other are likely to represent picoplanktonic organisms. Cercozoa, while occasionally found in low abundance in the euphotic zone[68, 57, 90] have been found to be more abundant in deep sea environments, such as abyssal sea floors in the Arctic and Southern oceans[122], showing a high number of unidentified phylotypes. The most commonly found class is Chlorarachniophyta, a group of photo- synthethic cercozoans that have been found in several marine environments[123, 124], suggesting a wide distribution of the group[4]. 28

1.1.6 Excavata

The Excavata constitute one of the six super-groups within eukaryotes, characterized by mostly heterotrophic flagellates possessing a unique ventral groove structure used for feeding of small particles through a flagellum-derived current[63]; they contain also numerous amitochondriate taxa. Despite containing a number of different taxa, only one of them, the Euglenozoa, seem to be abundant in marine ecosystems[93, 65]. Euglenozoa are very rare in marine environments, but this may be due to sampling biases, which are focused in samples from either the water column or from extreme benthic environments, such as hydrothermal vents. The few published studies on protistan diversity in the abyssal sea floor, an environment traditionally considered as homogeneous and low on diversity, showed a surprising number of euglenozoan phylotypes, corresponding to the Diplonemea and Kinetoplastea[80, 122]. While theorized to be parasitic organisms[122], little is know about the ecological function of these endemic sea floor euglenozoans; cultured diplonemids are free-living phagotrophs, with some strains having larvae with parasitic lifestyles[125]. At least two novel uncultured diplonemid groups have been identified from both pelagic and deep sea samples, DSPD-I and DSPD-II[126].

1.1.7 Unikonta

The Unikonta include two of the different eukaryotic super-groups, the Amoebozoa and the Opisthokonta; their grouping is based on the shared possession of only one basal body and one flagella, as well as the presence of distinctive gene fusions[116, 127]. While theorized to constitute one of the two main branches of eukaryotic diversity, which would then be rooted on the unikont-bikont divergence, this theory has been weakened by molecular and morphological data that has become available in the last decade[128, 79]. Amoebozoans are amoeboids or ameboid flagellates, with heterotrophic or para- 29 sitic lifestyles. They are more abundant on soil and freshwater environments[129], being almost completely absent from marine surveys[65] , with a few exceptions[71, 91]. The Ophisthokonta include metazoans, fungi, and a number of unicellular groups, the main one being the choanoflagellates. A sister group to animals, choanoflagellates are unicellular organisms with a single flagellum surrounded by a collar, found in marine surveys in low abundances[6, 130, 68]. They seem to be more abundant in freshwater ecosystems, even though molecular surveys have uncovered a number of novel clades endemic to marine environments[87]. Fungi are rare in the water column, appearing in very low abundances[65] and fulfilling a chemoheterotrophic role in the ecosystems, degrading complex compounds that are not accessible to other taxa[131]. Unexpectedly, the appear to play a much more important role in deep-sea hypersaline basins, such as the Thetis basin in the Mediterranean[11] where they constitute the dominant eukaryotic group. Few studies have retrieved fungi sequences from the picoplankton[7, 6], and it is not clear whether these reflect authentic planktonic fungi or are the result of contamination from external sources. Metazoans are regularly found in picoeukaryotic surveys[4], even though the smallest members of the group, the phylum Rotifera, are much larger than the 3 µm size limit for this fraction[132]. Their presence is considered an artifact, resulting from contamination from cell fragments and extracellular genetic material[133].

1.1.8 Other taxa

Cryptophytes are represented by a single species occuring in the picoplankton, Hillea marina[3]. While this group can be abundant in eutrophic coastal waters[89, 65], almost all cryptophyte sequences recovered from clone libraries have been found in temperate and polar samples, which may suggest that his organisms are poorly adapted to low nutrient conditions[4]. 30

Picobiliphytes are a recently discovered marine eukaryotic group, with affiliations to the Cryptophytes and Katablepharids[2]. They are found in low abundance in most marine environments, usually in the picoplanktonic fraction[134, 34]. While initially thought to be an exclusively autotrophic group due to the presence of phycobilin-like autoflourescence[134, 2], later studies using single-cell genomics could not identify plastid related genes in the genomes of three different members of this group, suggesting that it may contain heterotrophic organisms[62, 34]. Another recently discovered novel taxon are the rappemonads, known only for their plastid-derived 16S sequences, and which suggest that they are related to cryptophytes and haptophytes, being a sister group of the latter. They have been found in several diverse environments, both marine and freshwater, suggesting a high flexibility in habitat adaptation[25]. However, since there is no clear association in how variations in the plastid genome correlate with variations in the nuclear genome, their functional diversity and role are still unclear. Rare taxa constitute up to 30% of the total number of sequences in clone libraries, and up to 90% of the detected OTUs[135, 45]. The role and ecological importance of many of these low abundance groups is for the most part unknown - they constitute what is called the rare biosphere[136] and are believed to constitute an immense source of genetic diversity for the ocean microbial communities. Some of these organisms may mediate small but important biogeochemical processes[137], interact and influence the populations of abundant organisms, and other may have the capacity to become dominant after environmental changes[138], a relevant fact given the current climate change phenomena and its effects on marine ecosystems[139, 140]. It as also been proposed that the rare biosphere composition may be constantly changing, with some species becoming dominant and other decreasing in abundance in response to normal fluctuating environmental conditions; this would imply a constantly changing community, but one where organisms are replaced by ecological 31 equivalents, therefore allowing the persistence of normal ecological processes and functioning, such as basic biogeochemical cycles[135].

1.2 The Red Sea environment

Situated between the African Continent and the Arabian Peninsula, the Red Sea constitutes one of the most unique marine environments in the world. Despite being populated with settlements and being extensively sailed since ancient times, up to today the basin remains underexplored in many of its features, among them its microbial diversity. Being an environment with unique physico-chemical conditions, which include high irradiance, temperatures and salinity levels, low nutrients all along the basin, as well as a wide diversity of habitats and organisms, the study of the Red Sea is of exceptional importance. Despite occupying a wide latitudinal range, from 30◦N to 12◦30’N, the Red Sea present a remarkable uniformity in its climatic conditions[141], with most important seasonal variability taking place on its border regions, such as the Strait of Bab-el- Mandeb[142, 143] and the Gulf of Aqaba[144]. Inﬂuences from the Mediterranean climate in the north and from seasonal winds in the south are sporadic and with a relatively small impact. It is located in the hottest and most arid regions of the globe, resulting in environmental conditions that greatly inﬂuence its waters. Due to extremely high solar irradiation, the sea presents high surface temperatures which vary along the year, with a cooler period corresponding roughly to the northern hemisphere winter from October to April and a hotter period for the rest of the year. There is a clear temperature gradient along the basin, with the temperature increasing from north to south; average surface temperatures range seasonally from 18◦C to 35◦C in the north of the sea and from 29◦C to 40◦C in the south. Another feature is the constant warm temperatures of around 22◦C, present from a depth of 32

250 m all the way to the bottom, a situation unlike any other ocean or sea and which it is believed to be caused by warm water masses originated in the Gulf of Suez[141]. The thermoclines lies between 90 and 200 m depending on the region and the season, being generally deeper on the north and in winter. Rainfall is low along the entire basin, never exceeding more than 18 mm per year, with the northern region being generally more arid[145]. Coupled with the lack of riverine inputs of any kind and a high evaporation rate, this causes the water form the sea to present extremely high salinities, with an average of 40 Practical Salinity Units(psu). As is the case with temperature, we can see a salinity gradient with the lowest values in the southern regions (36 psu) due to the influence of the Gulf of Aden water influx to 41 psu in the north, due to the higher evaporation rates in this area[141]. Subsurface salinity is more constant because of the common origin and nature of these waters, not deviating from 40.5 - 40.6 psu below 300 m. Due to the temperature and salinity conditions, absolute oxygen values are lower than in other seas. Surface waters are close to oxygen saturation, with values from 4.5 ml-O2/l in the north to 4.0 ml-O2/l in the south. Below a depth of 100 m oxygen levels diminish until reaching a minimum at 400 - 600 m; the levels in this minimum oxygen layer change drastically with latitude, with values from 1.5 ml-O2/l in the north that decrease southwards until a minimum around 20◦N of 0.5 ml-O2/l[145]. Thermal stratification of the upper layers is pronounced, especially in the central and northern areas where the Aden surface current does not reach. Since this flow is also the main source of nutrients[142], this means that north of 19◦N the lack of nutrient influx and the presence of the barriers formed by the halocline and thermocline, which prevent the mixing of the impoverished surface waters and the comparatively nutrient-rich deep waters, cause the formation of a euphotic zone characterized by low nutrients and productivity[146]. Nutrient levels on this layer for the Northern Red Sea are low, with values of 33

0.01 - 0.05 µmol/l for phosphate, 0.1 - 0.4 µmol/l for nitrate and 0.2 - 0.4 µmol/l for silicate[147]. The low availability of nutrients is more prominent when compared with the values of the neighboring Indian Ocean, with annual average values of 0.4 µmol/l and 6 µmol/l for phosphate and nitrate, respectively[148]. There is at least one productivity peak during the year in winter, with a second one probably occurring in July in some areas of the sea; this increase in phytoplankton productivity has been associated with eutrophication phenomena, although its effects seem to be mostly restricted to coastal zones and to the ends of the sea[146]. The basin has experienced numerous environmental disturbances, including at least two periods of intense depth, temperature and salinity fluctuations during the Miocene and the Pleistocene[145, 149]. Many of the actual multicellular flora and fauna of the sea comes from posterior colonizations from the Indian Ocean; however, despite having limited evolutionary time to differentiate from the original Indian Ocean populations, relatively high rates of endemism have been detected in certain taxa, such as fishes[150], corals[151] and other organisms. Little research has been done regarding microorganisms in this aspect, although it seems that its unique oceanographic conditions have allowed the development of a distinctive microbiota, at least in the Bacteria and Archaea domains, as evidenced by a basin-wide survey[152]. Endemism has also been reported on specific microbial species, such as Streptococcus[153], Erythrobacter[154] and the Crenarchaeota[155], all of them in the northern part of the basin. Most of what is currently known on the eukaryotic microbiota concerns the most common, bloom-forming microorganisms, as well as large well-characterized protists. Blooms are more common on coastal areas, with important offshore algal blooms taking place only in the southern region, especially during the Southwest Monsoon. Blooms of Oscillatoria erythraeum are a regular feature of the central Red Sea during the same period, due to its nitrifying capabilities[156]. In the southern regions of the 34 sea, vertical convection in the winter period fertilize the euphotic zone, allowing the development of abundant planktonic communities, such as diatoms, dinoflagellates and coccolitophores[157], the latter being the most important primary producers. In general there seems to be a tendency for less diversity as latitude increases, which has been interpreted as caused by the increased distance from the Indian Ocean and its high biodiversity[146]. Unlike the situation with prokaryotes, no molecular studies have been done to study the eukaryotic populations of the Red Sea, leaving us with a gap in the knowledge of the biodiversity present in its waters. This is more pronounced in the case of picoeukaryotes, that are routinely ignored by microscopic surveys. This study was done to cover at least partially this gap, by studying the eukaryotic diversity in surface waters near the coast of the Red Sea. Located in the northern region of the basin, the chosen region presents a combination of mild temperatures and high salinity with low levels of nutrients, which have not been covered in other picoeukaryotic surveys. 35

Chapter 2

Materials and Methods

2.1 Sampling

Genetic material was available from a previous study[152]; briefly, environmental samples of seawater taken at a coastal site in the northeast coast of the Red Sea during March 2010, on the coordinates 24.96◦ N 36.35◦. Samples were taken in Niskin bottles from a 10 m depth and filtered consecutively on three different filters: 0.1 µm, 1.2 µm and 5.0 µm. DNA was extracted form each of these using the methodologies described in the respective publication

2.2 PCR ampliﬁcation and puriﬁcation

As a preliminary analysis for a prospective pyrosequencing survey, we amplified a small, 454-amenable SSU fragment in order to test the correlation between the communities portrayed by the full 18S gene and a small region within it. For this we used broad eukaryotic primers originally designed for pyrosequencing, as described by Medinger[158]; these primers, as well as all subsequent one, can be seen on Table 2.1. They amplify an 180-200 bp fragment from the conserved V3 region in the small ribosomal subunit. The genetic material from all three filters were amplified in a 36

Table 2.1: Primers used on this study. Primer Sequence (5’ - 3’) Amplicon size Reference MedF ATTAGGGTTCGATTCCGGAGAGG 180-200 [158] MedR CTGGAATTACCGCGGSTGCTG 180-200 [158] EukA AACCTGGTTGATCCTGCCAGT ∼1800 [159] EukB TGATCCTTCTGCAGGTTCACCTAC ∼1800 [159]

PCR reaction containing 2 µl of template genetic material, 2 µl of 10 µM of each primer, 2 µl of 200 mM dNTPs, 5.0 µl of BioLabs 10x ThermoPol Reaction Buffer and 0.25 µl of BioLabs TaqDNA Polymerase. The cycling profile used consisted on a denaturation step at 95◦C for 5 minutes, 30 cycles of 45 s at 94◦C, 45 s at 65◦C, and 90 s at 72◦C, and a final elongation step of 7 minutes at 72◦C. For amplification of the complete 18S gene, we used the universal eukaryotic primers EukA and EukB[159], which are the most widely used eukaryote specific primers on marine surveys[47, 6, 160, 94, 77, 161, 90, 89]. The results from the libraries prepared with the Medinger primers showed that he 1.2 µm contained the greater diversity of eukaryotic taxa, so this fraction was chosen for a more in-depth molecular analysis. The genetic material from this filters was amplified in a PCR reaction utilizing the same PCR mix than the previous amplification. The cycling profile used consisted on a denaturation step at 95C for 5 minutes, 30 cycles of 30 s at 94◦C, 90 s at 56.5◦C, and 90 s at 72◦C, and a final elongation step of 30 minutes at 72◦C. All PCR reactions were done on Eppendorf Mastercycler Pro thermal cyclers. The length size of PCR products were confirmed by electrophoresis on a 1% agarose gel. PCR products were purified using the Qiagen MinElute PCR purification Kit, following the included protocols. Final concentration of purified amplicons ranged from 20 to 80 ng/µl. 37

2.3 Clone library construction

Puriﬁed PCR products were inserted in the pCR®2.1-TOPO vector using Invitro- gen’s TOPO TA Cloning Kit into competent Escherichia coli competent cells, and the resulting clones were plated and grown in Sigma-Aldrich S-galactose/LB medium agar for 16 hours at 37◦C. Cells with successful insertions were picked using black- white selection and resulting colonies were grown in 1 mL of Alfa Aesar LB Broth for 24 hours at 37C with horizontal shaking of 750 rpm. The plates were centrifuged at 4000 rpm for 25 minutes at room temperature, the media was discarded, and the resulting pellet was sent to KAUST Genomic Core Facility for sequencing of the inserted sequences, using the M13fwd (5’-TGTAAAACGACGGCCAGT-3’) and M13rev (5’-CAGGAAACAGCTATGACC-3’) primers.

2.4 Community and phylogenetic analysis

The obtained sequences were processed using the software Sequencher 5.0 (Gene Codes Corporation). Sequences were end-clipped using the default parameters, and the vector was trimmed using the sequence provided by the manufacturer. Contigs were then assembled using the default parameters. Potential chimeric sequences were removed using a combination of three methods: (1) analysis of the sequences by the online software Bellephoron[162], (2) use of the chimera.slayer and (3) chimera.uchime commands[163] as implemented on the Mothur v1.21.1 software[164]. Additionally, potential chimeras were further conﬁrmed using the BLAST tool[165] with the standard MegaBLAST parameters on each of the individual fragments composing each full length sequence. Alignment of the contigs, Operational Taxonomic Unit (OTU) determination, community composition and taxonomic identiﬁcation were done on the software Mothur, using the eukaryotic dataset from the SILVA rRNA Database Project (avail- 38

able at http://www.mothur.org/w/images/1/1a/Silva.eukarya.zip) as an iden- tiﬁcation reference, using a minimum bootstrap value of 70% for the complete 18S sequences and 51% for those sequences covering the V3 region, in order to com- pensate for the smaller length. Diversity values were determined using the inverted Simpson diversity index, and the coverage was determined using Good’s coverage estimator[166]. Estimations for non-parametric species richness were done using the Chao1[167] and ACE[168, 169] estimators; for parametric estimations the software CatchAll v2.0[170] was used. The initial taxonomic analysis for the sequences was performed using the ARB software v5.2[171], aligning the obtained sequences with the online automatic SILVA

Incremental Aligner SINA v1.2.7 (available at http://www.arb-silva.de/aligner/) and adding them to the SILVA SSU Ribosomal RNA Database release 108[172] through the parsimony method. Identiﬁcation of the sequences by this method was preferred to the identiﬁcation via Mothur when they diverged (which ocurred only with 1% of the sequences, and only with those encompassing only the V3 region). Additionally, selected sequences were also compared using the BLAST tool, with the standard MegaBLAST parameters. A more in-depth phylogenetic analysis was done with the software Geneious v5.5.7[173]. Phylogenetic trees were build using the Maximum Likelihood (ML) method following the GTR model using the PHYML plugin[174], and the Neighbor Joining (NJ) method[175] using Jukes-Cantor distances. Bootstrap values were calculated for both methods with 1000 replicates. 39

Chapter 3

Picoeukaryotic community on the Red Sea surface waters

3.1 Community sampling

3.1.1 18S gene sequences

Our sampling of the Red Sea surface waters using broad eukaryotic primers encompassing the nearly full-length 18S gene yielded a total of 253 sequences, belonging to 4 of the 6 recognized eukaryotic super-groups. These sequences were clustered into OTUs at multiple levels of sequence similarity. Unless specified otherwise, community analysis was done using an OTU clustering criteria of 98%; with this method we obtained 101 unique different OTUs, distributed among this same taxa. Community composition can be seen on figure 3.1. 40

Figure 3.1: Eukaryotic community composition for the 1.2 - 5.0 µm fraction, analyzed using 18S sequences. Sequences were clustered into OTUs with 98% similarity.

Community structure is similar to that found in other marine environments using similar methodologies - for example, while alveolates and stramenopiles are among the most common groups in marine surface waters, they are specially overrepresented in 18S clone libraries in comparison with other methods, such as single amplified genomes[34]. The separation of environmental sequences into groups that represent biologically and ecologically distinctive species is not trivial, and much has been discussed about the genetic classification of uncultured unicellular eukaryotes. True taxon richness is believed to lie somewhere within the range of 1% and 4%[45]. A consensus sequence difference of 3% within each OTU is the most commonly held standard on bacterial communities, and the same value has been proposed as a good compromise for identifying eukaryotic taxonomic units, at least for small, specific regions within the 18S gene[176]. Caron et al, after analysis with both cultured, well defined species and environmental samples, argued that strains of the same species share in average a similarity level of 98%, with exceptions restricted to certain taxa, such as Amebo- zoa, concluding that clustering with a 95% similarity is able to provide approximate 41 species-level distinction between sequences[177]; on the other hand, other authors have argued that a 1% difference is enough to delimitate different species, with 2% representing different genera[90]. It is worth keeping in mind, as has been mentioned by other authors[69], that even though we speak of certain similarity levels as representing specific taxonomic levels within a rank-based classification, such as species, genera or families, in reality each similarity level covers instead a range of taxonomic levels, whose broadness is most likely dependent on the taxa analyzed. This is neither surprising nor unexpected if we consider that the foundations of most of protistan classification were laid based on morphological criteria, which failed to reflect, due to technological constraints, the huge cryptodiversity in apparently identical organisms[178, 61], a problem especially preponderant with small picoplanktonic eukaryotes, which often lack any defining physical characteristics[107]. Caron’s definition as been applied, using its most stringent 98% conditions, by other authors on both general eukaryotic communities[68, 11, 179, 114] and within specific taxa[180]. Based on this, we have decided to use criteria for our OTU clustering, which for our nearly full-length 18S sequences represents a difference of 36 nucleotide differences in average. This criterion is one of the most widely used in previous community analysis studies - facilitating comparisons with them - as well as presenting a good compromise between capturing all the existing relevant diversity and grouping already existing species in defined taxonomic units. As we can see on Table 3.1, the effect of increasingly more relaxed clustering criteria affected each taxa differentially. In Alveolata, the number of OTUs diminishes gradually as we decrease the intra-OTU similarity, with no clear drop in the number of units; the effect is significant, and by the time we are applying a 95% similarity the number of taxonomic units has been reduced to one half of the number of unique sequences. Looking into more detail within the group, we can see that the different 42

Table 3.1: Operational Taxonomic Units for the 18S clone library for each taxonomic group, using different clustering criteria. Seqs=number of total sequences. Taxon seqs unique 99% 98% 97% 96% 95% 90% ALVEOLATA 81 76 56 49 45 39 39 21 Ciliata 5 5 5 5 5 4 4 3 Dinophyceae 2 2 2 2 2 2 2 2 NA-I 7 6 4 4 4 3 3 2 NA-II 67 63 45 38 34 30 30 14 CHLOROPHYTA 81 63 8 6 5 4 4 2 Mamiellales 79 61 7 5 4 3 3 1 Clade VII 2 2 1 1 1 1 1 1 STRAMENOPILES 52 47 33 31 31 30 30 22 Bicosoecida 1 1 1 1 1 1 1 1 Dictyochales 4 4 3 2 2 2 2 1 Pedinellales 3 2 1 1 1 1 1 1 Pinguiophyceae 1 1 1 1 1 1 1 1 Chrysophyceae 12 12 7 7 7 7 7 5 Bolidophyceae 5 3 2 2 2 2 2 1 Novel clade 2 2 1 1 1 1 1 1 MAST 24 22 17 16 16 15 15 11 RHIZARIA 18 14 3 3 3 3 3 3 Radiolaria 17 13 2 2 2 2 2 2 Cercozoa 1 1 1 1 1 1 1 1 HAPTOPHYTA 1 1 1 1 1 1 1 1 Prymnesiophyceae 1 1 1 1 1 1 1 1 UNKNOWN 2 2 1 1 1 1 1 1 METAZOA 18 17 10 10 8 8 7 5 43 subgroups have not been affected in the same way: ciliates and member of the NA-I present very stable clustering, with individual sequences representing distinctive taxa; it is worth noting that this groups had a few sequences each, and maybe with more sequences we would have seen a more drastic grouping. A similar situation happens in NA-I, although we see a more clear progression of the clustering process; at 90% identity, we have two OTUs, which at least for our samples correspond to the clade level determined by Guillou[47], as we’ll see below. Meanwhile, NA-II, which are by far the biggest group within the alveolates, decrease constantly into fewer and fewer taxonomic units, without reaching any discernible clustering stability; even though the different clades of the group were established by the same criteria and using the same datasets as the ones in NA-I, here we don’t see a correspondence between the number of clades and any threshold for OTU definition. In contrast to the alveolates, Stramenopiles do present a relatively more stable clustering at all identity thresholds from 99% onwards, a factor constant in all identified orders within the taxon. The different identified classes seem to correspond to the OTUs at 90% similarity, with the exception being the Chrysophyceae, a class which alone contains 5 OTUs; however, it is known that environmental sequences from this group present a natural high diversity[87]. MAST present a steadily decreasing number of OTUs, and like the case with NA-II, no clustering criteria can be correlated with the present classification of novel stramenopiles. The most dramatic example occurs within the Chlorophyta, which drops from being one of the two groups with the greater number of sequences to just 8 OTUs as soon as the initial clustering conditions are set. Multiple gene studies have detected low diversity on the 18S gene between different clades of the Chlorophytes (and more specifically, for the Mamiellales, the order to which 99% of our sequences belong to)[107], especially when compared with other functional genes; this could account for the apparently low diversity that we have found in our samples. Additionally, 44 member of these groups have been reported to be an important component in coastal ecosystems[57], where they periodically form important blooms[39, 13]. It remains to be seem whether this combination of low diversity and high abundance of Chlorophyte sequences is a characteristic of the northern Red Sea, or if it reflected some transient event in the photoplanktonic community during the sampling period. The Rhizaria, while being a low abundance group in our samples, are composed of sequences belonging to distinctive separate taxa, two belonging to the Radiolarians and one to the Cercozoa, which remain stable even with a 90% identity threshold. As for the other groups, Haptophytes and unclassified organisms presented one and two sequences respectively, so clustering was trivial. Both unclassified sequences, while distinctive, clustered into one single OTU at the most stringent threshold, implying that they belong to closely related organisms. Metazoan sequences belong to multicellular organisms, so they will not be discussed further in this nor on subsequent clone libraries. Rarefaction curves for our sampling, considering different parameters for OTU clustering, can be seen on figure 3.2. As we can note, when considering unique sequences as taxonomic units we are very far from achieving sampling saturation; all other instances of sequence clustering present very similar rarefaction curves that show progressively better sampling coverage as we diminish the stringency on clustering criteria; this large drop in sequence diversity with the initial clustering relaxation is similar to what has been found in other studies[90]. 45

Figure 3.2: Rarefaction curves for our 18S gene sequences, for six diﬀerent thresholds of OTU clustering: unique (only sequences that are identical), 0.01 (sequences with 99% or more similarity), 0.02 (sequences with 98% or more similarity), 0.03 (sequences with 97% or more similarity), 0.04 (sequences with 96% or more similarity), 0.05 (sequences with 95% or more similarity) and 0.1 (sequences with 90% or more similarity).

Good’s coverage for the chosen OTU deﬁnition was of 76.6%, while for unique sequences it was a very low 22.9% - a fact consistent with our rarefaction curve values. Estimates for total species richness in our sample were 182 and 193 using the nonparametric Chao1 and ACE estimators, respectively, implying that our library captured between 52% and 56% of the total number of OTUs. Parametric estimations suggested an even higher diversity, with a total value of 260; these estimations are know to be more accurate on low coverage datasets such as the ones in this study[178, 90]. This values are consistent with that of previous studies done using comparable sized clone libraries[68, 6, 7, 181], while falling short of estimations done in studies with more intense sequencing depth - namely, those with a total number of sequences more than one order of magnitude greater than those in the present study - which fall often within the 500 - 1200 range for open and coastal ocean waters[90, 24] going up 46 to 10,000 in extreme environments, such as anoxic water basins[95] or hydrothermal vents[136].

3.1.2 V3 region sequences

Additionally to the full 18S analysis, clone libraries were also created using primers for the V3 region of the SSU, in the same filter size and using the same original samples. This was done as a preliminary study in order to evaluate the possibility of using these small subregions for a pyrosequencing analysis in the Red Sea, which would provide us with much greater coverage of the protistan diversity on multiple sample sites covering the whole basin, like previous studies have done for the Bacteria and Archaea domains[152]. The sequences obtained from this library would allow us to see if the diversity captured by V3 sequences was equivalent to that of full length sequences, and to determine how appropriate was the selected region for taxonomic assignment of sequences. After removing sequences with insufficient length, ambiguities and possible chimeras, our clone library had a total of 399 sequences. Given that the likelihood of PCR recombination increases with amplicon length[182], it can be argued that use of small, ≤500bp fragments minimizes the incidence of chimeras, and such as been the underlying assumption of previous next-generation sequencing studies[45, 55]. Still, checking of our sequences returned a total of 4 chimeras that, while being a negligible amount (<1% of the original library size) still supports the need to do chimeric evaluation before further analysis. As with the complete 18S sequences, the V3 amplicons were clustered at different similarity levels for analysis, with a 98% similarity used for as a standard for most analyses. With this criteria, our sequences were clustered into 129 unique OTUs, distributed among 3 of the 6 recognized eukaryotic supergroups. The one missing supergroup is Rhizaria, and it is very likely to be found in our samples; however, confident classification of members of this group using the short V3 region 47 proved impossible, and therefore all tentative rhizarian sequences are grouped into the unclassified fraction. General community composition can be seen on Figure 3.3.

Figure 3.3: Eukaryotic community composition for the 1.2 - 5.0 µm fraction, analyzed using the v3 region in the SSU. Sequences were clustered into OTUs with 98% similarity.

Notably, the percentage of sequences falling within the ”unclassified” category is much higher than in the complete 18S library: 45.7% of OTUs fall here, in contrast to 1.9% using the larger amplicon. This is likely due to a combination of the small size of the amplified regions, which in many cases prevents a more confident taxonomic assignment even using more relaxed classification threshold and the relatively small coverage of current eukaryotic databases in certain subregions of the rRNA. The taxonomic ambiguity is not equally distributed among all taxa, with some groups like alveolates being susceptible of confident classification employing small regions[158] while others, such as Rhizaria, being much harder to accurately classify. Our results seem to support this, as discussed below; however, once again we cannot discount the biases introduced by the reference databases: the alveolates are one of the best studied taxa within unicellular eukaryotes, with the greater number of sequences in publicly available databases, constituting for example 13% of the Eukarya domain 48

Table 3.2: Operational Taxonomic Units for the 18S v3 region clone library for each taxonomic group, using different clustering criteria. Seqs=number of total sequences. Taxon Seqs unique 99% 98% 97% 96% 95% 90% Alveolata 84 64 55 45 40 36 35 14 Stramenopiles 33 16 13 11 11 11 8 4 Chlorophyta 142 34 11 6 2 1 1 1 Metazoa 6 6 6 6 5 4 4 2 Haptophyta 4 3 3 2 2 2 2 1 Unclassified 130 84 66 59 54 53 49 35 sequences in the Silva SSUREf 108 database, and 28.4% of all unicellular eukaryotes. In Bacteria, the V3 region allows for much greater coverage of microbial diversity, while at the same time permitting a 99% accurate taxonomic classification, compared to full-length sequences[183], which was one of the reasons this regions was chosen for the study. However, based on our results it would seem that its use in eukaryotes is currently more limited. Further studies would be required to see whether this is could be improved by larger, higher-quality reference databases, or if it reflects intrinsic characteristics of the chosen ribosomal region. Regarding the effects of sequence clustering in each taxon, as shown on Table 3.2, we can see similar patterns to those observed with nearly full-length sequences: alveolates show progressively fewer groups as the similarity thresholds are increased. Notably, the total number of OTUs detected is very similar to the one detected in the full length library, suggesting that this region may be specially useful for this taxon. The same is true for the Chlorophyta, which show similar significant decreases in the number of groups, although this is even more drastic with the V3 region sequences, probably due to their shorter size. On the other hand, the stramenopile fraction is significantly smaller in this library, a phenomenon that becomes more drastic as we use more relaxed clustering. It could be argued that this could be due to biases imposed by the primers - however, we believe that this is unlikely, given that analysis of our full length sequences indicate 49 that the primers align to their specific conserved regions with no more than 1 bp mismatch for all stramenopile sequences. It would seem that this region does not provide sufficient phylogenetic information, at least with the methodology used herein, for accurate taxonomic classification on this taxon, causing numerous stramenopile sequences to fall within the Unclassified category, as it was the case with rhizarians. More haptophyte sequences could be detected using the V3 primers, although considering their low number and the sample size, we cannot infer any differential primer specificity of this fact alone. As for the Unclassified sequences, we suggest that some of them tentatively belong to unclassified stramenopiles and rhizarians; this is partially supported by analyses between the V3 clone library and an in-silico extracted V3 region of the full length sequences, showing that they share 14 OTUs in common, belonging to the following taxa: Alveolata (28.5%), Stramenopiles (28.5%), Rhizaria (7.2%), Chlorophyta(7.2%), Metazoa (21.4%) and unclassified (7.2%). Still, two thirds of the unclassified sequences remain elusive, an we cannot reach a conclusive consensus as to which proportion of them may belong to other taxa not represented in the full length 18S sequences. The rarefaction curves (Figure 3.4 for the V3 regions show that we could not achieve sampling saturation for our sample, except when considering taxonomic units with 90% similarity. We do not see the initial improvement in sample saturation when imposing the initial clustering conditions, as we saw on the complete 18S sequences, and instead we can see a more gradual descent of the curve’s slope with each successive OTU definition. That we are far from achieving a complete sampling in this library is not surprising, considering that each clustering relaxation involves a change of a much smaller number of nucleotides,due to the small size of the considered amplicon, and the fact that V3 is a well known hypervariable region. 50

Figure 3.4: Rarefaction curves for v3 region sequences, for six diﬀerent thresholds of OTU clustering: unique (only sequences that are identical), 0.01 (sequences with 99% or more similarity), 0.02 (sequences with 98% or more similarity), 0.03 (sequences with 97% or more similarity), 0.04 (sequences with 96% or more similarity), 0.05 (sequences with 95% or more similarity) and 0.1 (sequences with 90% or more similarity).

Interestingly, coverage for our V3 library compared with its full length counterpart was roughly the same when considering operational taxonomic units (78.4%), but when calculated for unique sequences it was more than double (58.4%); while the greater coverage can be attributed to the bigger size of the library, the fact that our coverage when considering OTUs was the same is an indication that even while not capturing the same number of tentative protistan species, our full length sampling was able to capture a comparable diversity at the family/genus level. Species richness values were considerably higher: nonparametric estimations were 311 and 571 for the Chao1 and ACE calculators, and 743 using the parametric Two Mixed Exponential model. 51

3.1.3 Other clone libraries

Besides the 1.2 - 5.0 µm fraction, smaller clone libraries were done with the genetic material on the 0.1 - 1.2 µm and ≥5.0 µm size fractions, as part of a preliminary analysis of the eukaryotic diversity in the diﬀerent size compartments. After removing of fragments with insuﬃcient length, sequences with ambiguous nucleotides and chimeras, these libraries contained 64 and 46 sequences, respectively. Their composition, as compared to the 1.2 - 5.0 µm, can be seen on Figure 3.5.

Figure 3.5: Eukaryotic community composition for the 0.1 µm, 1.2 µm and 5.0 µm ﬁlters, analyzed using the v3 region in the SSU. Sequences were clustered into OTUs with 98% similarity.

While the sampling coverage using Good’s estimator of these libraries is smaller to make any confident comparisons (53.1% for the 0.1 - 1-2 µm and 63.0% for the ≥5.0 µm fraction), we can see that each community has a distinctive composition. Alveolates are common in all libraries, composing one third or more of the in each one of them. Stramenopiles also have a similar composition on the smaller filter sizes, 52 although they are interestingly absent from the ≥5.0 µm community; this is unusual, considering that diatoms, one of the most well known members of this taxon, are one of the main components of plankton on this size range. However, traditional planktonic surveys have shown that the diatom diversity in the Red Sea is scarce compared to other basins, and that most of it is restricted to the south due to the inflow of surface waters from the Indian Ocean[184]. Additionally, their characteristic silicaceous shells are very resistant to breakage during DNA extraction processes, a fact that has been used to explain their absence in other molecular surveys[158], although with the robust extraction procedures used in this study that possibility is relatively low. The absence of Chlorophyta in the 5.0 µm library is expected, since most of the members of the prasinophytes, the class to which all of our chlorophyte sequences fall into, have an average diameter of 3 µm or less[104]. Also expected is the higher abundance of metazoan sequences in this filter, due to most of the cells and eggs of multicellular organisms getting trapped on it.

3.1.4 Suitability of diﬀerent regions of the SSU gene

So far, analysis of the v3 region has been employed only on freshwater environments, and applied to specific eukaryotic taxa[158]. The primers were chosen from among other options mainly because in-silico analysis showed that they covered a high percentage of almost all important eukaryotic taxa, with 86% for the forward primer and 97% for the reverse one, without retrieving sequences from other domains. As mentioned above, while they seem to give us an accurate portrayal of the eukaryotic community, compared with the EukA - EukB primer pair, the number of sequences with an undetermined taxonomic affiliation increases greatly with v3 sequences, an undesirable effect when trying to infer the ecological relevance of the different groups captured on our surveys. 53

An in-silico analysis was done by extracting most SSU regions from our complete 18S sequences, and comparing the diversity portrayed by them at different OTU clustering criteria. The results appear on figure 3.6. While it can be argued that in-silico data can substantially differ from the actual results when applied on environmental samples, we feel confident in this data, based on the results for the v3 region. OTU curves of the actual sequences obtained from our library and of the in-silico sequences are almost identical, with the exception that the in-vitro obtained clones retrieved a higher number of unique sequences, although the number of OTUs was the same when a clustering criteria of 98% similarity or lower was applied.

Figure 3.6: Diversity of OTUs for six diﬀerent regions of the 18S gene. The sequences considered are the full length 18S gene (FF), the v2 region extracted in-silico(V2 IS), the v3 region sequences from our clone library (V3 IV), the v4 region extracted in-silico(V4 IS), the v6 region extracted in-silico(V6 IS), the v9 region extracted in-silico(V9 IS) and the v6 - v9 region(V6-V9 IS).

Our data also shows that the v3 region consistently underestimates the OTU diversity, so the similar diversity values that we retrieved from both libraries can be attributed to the higher sampling depth on the v3 clone library, which contained almost twice as many sequences. Pyrosequencing surveys usually utilize the v4[44] or v9[24, 1, 122, 45] regions, since both provide sequence targets amenable to 454 54 amplification. While our preliminary analyses showed that the primers for this regions captured a smaller percentage of the eukaryotic community, surveys done with them do capture a wide diversity of protists in almost all marine eukaryotic phyla. Making a comparison between the two, the literature reports than the v4 region is more variable than v9, and also more prone to sequencing errors, given its higher number of homopolymers[55]. This contrasts with our results, where the v9 regions appeared to be hypervariable and consistently formed more OTUs than all other sequences, even the full length ones. Given that the region includes conserved subregions that allow for the design of primers with a broad eukaryotic coverage, this greater variability would allow for a more in-depth distinction of closely related organisms. However, preferred regions for analysis may also be taxa-dependent, since for ciliates the v4 region seems to correspond more accurately to the diversity and richness presented by longer sequences, making it a better choice for comparisons on community study and composition[185]. Another factor to consider when choosing regions is up to which level they allow us to make accurate taxonomic assignment of SSU sequences, which was a major issue in our analysis of the v3 clone library. In this aspect, the v9 region shows a much smaller proportion of unclassified clones than the one employed in this study, with more than 80% of the sequences unambiguously assigned to the family level at an 85% confidence threshold[122, 45]. Similar values have been reported for the v4 region[55], suggesting that either of these would be more adequate for future molecular surveys, besides allowing for a more accurate comparison with surveys on other areas. 55

3.2 Picoeukaryotic diversity

3.2.1 Alveolata

Eighty-one sequences from our clone library (32.0%) could be assigned to the phylum Alveolata. By far, most of them belonged to Dinoflagellata (94%), with the remaining 5 sequences belonging to the Ciliphora. Our tree showed good bootstrap support for all the main alveolate groups that contained clones from the Red Sea. The relationship between the different groups is not clear, although the clustering of the NA-II with the rest of the dinoflagellates is moderately supported, as has been shown in other studies[72]. The dominance of Alveolates on molecular surveys, both those done by traditional clone libraries and next generation sequencing, is well documented[1, 33, 93, 158, 68, 45], in marked contrast to morphological surveys. This discrepancy is present even when the same samples are examined[158, 186], suggesting that it may be the result of a PCR amplification bias. This is partially explained by the fact that many groups within the alveolates are known to contain a high number of copies of their ribosomal genes[48] in comparison to other eukaryotic taxa; a similar effect has been documented on bacterial species with relatively high numbers of rRNA copies[187]. Given that there is an observed correlation on eukaryotes between organism size, genome size and number of rDNA copies[188, 189], we would expect this phenomena to be diminished in picoplanktonic surveys, where the range of copies between the different taxa is small. We must also consider the possible interference of the breakage of bigger protistan cells during filtration, a phenomena that can dramatically bias the observed community, since these bigger cells can contribute with as much 18S sequences as 10000 picoplanktonic cells[39]. This is a particularly strong phenomenon with ciliates, which are prone to breakage and can have thousands of copies of the ribosomal genes. 56

100/99

99/98

100/99 100/100

100/100

53/52

100/100 Gymnodiniales

70/61

99/97

98/98

97/95 100/100

52/57

84/80

96/92

100/100 54/--

100/100 Gonyaulacales 89/86

100/100

Figure 3.7: SSU rDNA phylogenetic tree for Dinophyceae. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. On this and all subsequent trees, two unikont sequences (not shown) were used as an outgroup; the sequences were from the organisms Acanthamoeba castellanii (Accession number U07413) and Hartmannella vermiformis (Accession number AF426157).

Within the Dinoflagellata, sequences were distributed among the traditionally well-defined groups of such as the Dinophyceae, but by far most of the sequences were found in groups mostly represented by environmental sequences, namely NA-I and NA-II. Dinophyceae appeared as a monophyletic group separated from the rest of dinoflagellates, as in similar analyses[47], although the monophyly of the Syndiniales was not preserved in our trees. Only two sequences of Dinophyceae were found in our study, which is consistent with the assumption that most of this organisms are bigger than the picoplankton and were thus retained by the filtration steps. 57

One of our clones (RS.10m.172) was tentatively identified as part of a novel group within the Gonyaulacales, clustering closely with organisms isolated from the nanoplanktonic fraction in the DCM of Sargasso Sea eddies (unpublished study, accession numbers AY665007, AY665004 and AY665005), with no close characterized organisms. The Gonyaulacales are one of the few groups within the dinoflagellates that are strongly supported by both genetic and morphological analyses[190, 191]. This group is characterized by their unique gonyaulacoid tabulation, with irregular plates covering the cell[192], and some of them the main component of toxic blooms due to the secretion of phycotoxins[191]. The other sequence, RS.10m.349, appears within the Gymnodiniales, tentatively related to Amphidinium semilunatum and to environmental clones extracted from the Ross Sea, off of Antarctica. These clones belong to a novel keptoplastic dinoflagellate abundant in this region, with vegetative cells that range from 8 to 20 µm of width[193]. Despite the fact that most Gymnodiniales are autotrophic unarmored dinoflagellates whose size ranges from 11 to 47 µm[194], sequences related to this group are common in marine picoplanktonic samples[4, 23], suggesting that this group may harbor multiple small species which have yet to be characterized. Similarly to the Gonyaulacales, some species are important toxin-producing organisms associated with harmful algal blooms[195, 196]. The large size of the genus Amphidinium and the fact that most of its species are benthic[192] make the assignation of this sequence initially doubtful, but other Amphidinium-like genes have been found in picoeukaryotic surveys[68] suggesting that the genus may contain a higher morphological diversity than previously assumed. 58

100/100

100/100 100/100 100/100

100/100 96/95 66/66 Clade 5 100/100 100/100

100/100

80/80

98/98

100/100 100/100 Clade 1 100/100

100/100 57/53 100/100 100/100 79/78 98/96 100/100

100/100 99/99 56/65 100/100

100/100 100/100 100/100 94/96 100/100 98/92 100/100

Figure 3.8: SSU rDNA phylogenetic tree for members of the NA-I group. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

Organisms in the NA-I were represented by four different OTUs, as seen in Figure 3.8. These belonged to the clades NA-I-Clade 1 and NA-I-Clade 5; this is consistent with data from other marine environments where these were found to be the most common clades, constituting more than 75% of the sequences belonging to NA-I[47]. Both OTUs in Clade 1, represented by 2 and 1 sequences, were closely related to alveolate sequences found in surface waters of the Sargasso[69] and Baltic[36] seas, as well as the equatorial zone of the Pacific Ocean[23], indicating a wide global distribution of this group. Highly similar members of this clade have been characterized through the use of FISH probes as approximately 2.5 µm long flagellated cells with large nu- clei, present in concentrations of up to 600 cells per ml-1 in eutrophic, low salinity 59 environments[36]; it still remains to be seen what their morphology and distribution patterns are in the Red Sea, although they are likely to differ considering the numerous physical differences between the two habitats. Similarly, the two OTUs in NA-I Clade 5, with three and one sequences, were also closely related to Sargasso Sea and in a lesser degree to Pacific Ocean sequences; one of them was also present in samples from the English Channel[47]. We found no sequences related to the characterized genera within this group, Duboscquella and Ichthyodinium, which is consistent with other published environmental surveys. The bulk of the Syndiniales was in the NA-II group, for which we could identify 67 clones, divided into 38 unique OTUs, and distributed among 18 clades. The distribution of these can be seen on figures 3.9 and 3.10. Almost a third of the sequences (17 of them) were in the Clade 10+11, a clade composed mostly of environmental sequences from surface waters of the Sargasso and Mediterranean seas, but which also includes some organisms from deep anoxic environments, such as the Cariaco Basin[37]. In our analysis this clade appears to be consistent and well-supported, in contrast with the results of previous studies[47], although Guillou et al used partial sequences for the determination of clades, which may have affected bootstrap values; in any case it seems likely that future analyses may divide it into further subclades. 60

100/100 Clade 14 * 100/100

100/99 Clade 26 92/92

100/100 Clade 22 100/100

60/57 100/100 100/100 100/100 Clade 16 65/65 100/100

100/100 58/58 Clade 4 73/72 100/100 63/54 100/100 Clade 2 Amoebophrya 78/88 complex 100/100 63/62 100/100 Clade 3

98/98

100/100

100/100 Clade RS2

50/-- 100/100

100/100* 100/100 54/52 Clade 28 54/55

75/74 100/100

65/60 Clade 13 100/100

75/72 100/100 Clade RS3

100/100 100/100 100/100 Clade 44

75/80

Figure 3.9: SSU rDNA phylogenetic tree for members of the NA-II group. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7. 61

100/100 100/100 Clade 21 100/100 100/100

98/97 76/78 100/100

100/100 Clade 6 91/89

100/100 100/100

99/100 100/100

100/100 100/100 98/99 100/100 Clade 20 56/60 98/99 90/92

100/100

92/92

100/100 96/97 Clade 10+11 56/62 100/100

56/60

92/90 96/95 84/80

84/77

92/93 100/100

100/100 100/100 Clade 22

94/93 100/100 Clade RS1

Figure 3.10: SSU rDNA phylogenetic tree for members of the NA-II group. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

This clade also contained most of the diversity within NA-II, with 9 OTUs; our sequences were distributed among two distinctive branches, one containing six OTUs related to clones isolated from surface waters of the Northern Atlantic, Mediterranean[47] and coastal Paciﬁc[57], with the addition of an Arctic sequence from deeper waters(70 m)[71]; the other branch contains three OTUs that cluster with sequences from the same Paciﬁc and Atlantic libraries. 62

The second most abundant clade was Clade 6, which contained 8 clones. It was not as diverse as Clade 10+11, with only 3 different OTUs. Like with Clade 10+11, they are similar to sequences from surface waters of the Pacific[57] , Mediterranean[47] and Sargasso Sea[69]. This clade is abundant throughout entire water column, a distribution that may indicate a broad host range, assuming that these are indeed parasitic organisms, as has been proposed[47]. We could also find sequences in our clone library belonging to the Amoebophrya superclade, which is a monophyletic group[70], composed by Clades 1-5, 25, 33, 39, 41 and 42[47]. There are currently 7 described species of Amoebophrya, although only one of them has been genetically described, A. ceratii[47]. Given the wide genetic diversity of the group found in environmental samples, there is considerable uncertainty in how described species correspond taxonomically to the genetic groups determined by environmental sequences. Our library contained 3 OTUs for this group, conformed by 4 sequences belonging to Clades 2, 3 and 4. All Amoebophrya organisms are parasites, with considerable variability in their host range; strains with high specifity seem to be closely related to strains with non- specific infectivity[58], so ecologic assumptions based on genetic closeness to related strains may prove unreliable. Still, is encouraging to see that the closest described strains to our Clade 4 and Clade 3 sequences are organisms infecting members of the Gymnodiniales[197, 76], a group also found in our survey, suggesting a possible ecological link between these organisms. Sequences from Clade 2 include only a characterized Scrippsiella-infecting strain, plus a number of environmental sequences from coastal sites in the Pacific[57], and more importantly, the Atlantic[47, 198]. All of these were isolated from filtered samples in the euphotic zone below the 3 µm range. The proportion of sequences of the Amoebophryidae is lower that the ones found in similar studies, mostly due to the lack of clones belonging to Clade 1 - this is 63 one of the most common clades found in samples of surface waters[47], and by far the most abundant one in the Amoebophrya complex. Analyses encompassing bigger protistan fractions would help us to indicate whether this may reflect an absence of their target species, which for the identified strains include Karlodinium veneficum and Prorocentrum minimum, or some other characteristic of the sampling site. Clade 22, unlike in its original definition[47], appears divided into two unrelated branches; this is not surprising since it was established using only partial sequences and the arrangement did not show a strong bootstrap support. One branch contains 2 sequences related to picoeukaryotes from the Sargasso Sea[69], while the other contains four sequences divided into 3 OTUs related also to Sargasso Sea and Mediter- ranean sequences[47]. Clade 16, which is also considered an abundant clade[47], contained 5 sequences, most of them also highly related to Sargasso Sea sequences. The other clades, 13, 26 and 44, presented four, two and three sequences respectively, but showed very little diversity, with only one or two OTUs per clade. These organisms were mostly related to organisms from the Sargasso Sea[69] for Clade 13, the Atlantic coast for Clade 26[47] and the Mediterranean for Clade 44[47]. Other clades contained singletons - sequences that constituted their own OTU and were not related to any other sequence in our survey. One of these singletons (clone RS.10m.0014) appeared in our tree as related to Clade 14, although they appear to be quite distant from other environmental sequences, presenting low similarity (88% or lower) to any other sequences on the database. Clone RS.10m.0019 is only closely related to sequences from the Arctic[71], falling within Clade 21. Clade 20 contained was represented by one clone that did not cluster with most of the other members of the group, except for organisms from the deep area of the Sargasso Sea[69]. Finally, clone RS.10m.00183 was established as a member of Clade 28, and was closest to organisms from the wintertime Baltic sea (unpublished study, accession number FN690343). 64

Both the low-diversity and singleton clades appear to constitute a minor fraction of the picoeukaryotic fraction in our sampling site, which is consistent wih the fact that all these clades are relatively small and not very abundant in other marine environments[70, 47]. Considering this and the fact thay they appear to be phylogenetically associated with samples from such diverse areas, we cannot make assumptions whether they may play any meaningful ecological role in the coastal Red Sea ecosystem, and which role that would be. A number of the sequences found in this study do not appear to be related to any previously described clades, or they do not associate confidently with other environmental clones. Given the singular physicochemical characteristics of the Red Sea and its relative evolutionary isolation, these may represent novel clades within NA-II. Pending a more in-depth phylogenetic analysis, we have grouped these sequences into a number of potential clades with the prefix RS. We have used a variant of the criteria used for clade definition in previous studies[70, 47], namely that a clade must (1) be composed by two or more environmental sequences and (2) have a bootstrap support higher than 60% in ML analysis. The first of these groups, Clade RS1, consists of two OTUs, that appear as a sister group to Clade 10+11, being related to sequences belonging to a group denominated ”Close to 10+11”[47] and characterized by long phylogenetic branches. While our sequences also present this characteristic, they do not group confidently with other Close to 10+11 sequences, and may represent a novel, sister clade to both of these groups. Clade RS2 is composed by seven sequences divided into two clusters. It does not appear to bear any close relationship to other clades, its closest relatives being within the Amoebophrya complex; and while it falls outside the range of the characterized Amoebophrya clades, it may tentatively be composed of organisms related to this parasite. The singleton RS.10m.350 appears close to RS2 in our trees, but the bootstrap 65

values are not high enough to consider it part of this clade. The third novel clade, RS3, is composed by a single OTU with two sequences, and its closest described clade is Clade 13. Clone RS.10m.00127, does not seem to belong to any known clade, but since it is a singleton, we cannot conﬁdently say if it constitutes a novel group within NA-II.

100/100

99/99

75/82 Oligotrichia

100/100 Spirotrichea 100/100

91/90 57/74 100/99 Choerotrichia

69/71 90/82

59/--

99/94 59/--

74/84

100/100

99/98

98/97 69/98 94/67 100/100 100/100

100/99

Figure 3.11: SSU rDNA phylogenetic tree for Ciliata. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

As for the ciliates found in our survey, as we can see in ﬁgure 3.11, all the sequences detected fall within the class Spirotrichea, with organisms identiﬁed as belonging to the Choreotrichia and Oligotrichia subclasses. The two oligotrichid sequences form a distinctive group whose closest characterized organism is the Varistrombidium genus; they show a high grade of similarity (>98%) to clonal sequences found in sediments of the Arctic Ocean, near the Kongsfjorden fjord[161]. 66

One of the sequences, RS.10m.0077, presents a high similarity with isolated strains of Lynnella semiglobulosa, probably belonging to this or some closely related genus. Lynnella is a 60 µm-long ciliate that inhabits coastal areas, and whose taxonomy within the Spirotrichea has been a subject of debate, since it presents both choerotrichid and oligotrichid characteristics; molecular studies do not show a clear support for either hypothesis[199]. In our trees this genus appears as a sister group to the oligotrichia. Choerotrichids were represented by two sequences; none of them have high similarity to any published sequences, and their closest identified organisms are Rimostrom- bidium clones found in freshwater environments. One of them also groups more closely with clones obtained from Sargasso Sea eddies (Genbank accesion number AY665093), which indicates the presence of similar organisms in oligrotrophic marine environments. While ciliates are considered to have dimensions outside the range of the picoplankton, with all the species identified so far being larger than 8 µm[7, 66], sequences from this group have been found in multiple studies from environmental samples belonging to the picoplanktonic fraction in a wide range of environments, ranging from the open ocean to Antarctic ice edges[6, 7, 33, 57]; additionally, small aloricate ciliates have been found as a common component of the nanoplankton, specially members of Spirotrichea, as those found in this study[66], being only slightly larger than the size limit for picoeukaryotic surveys. Still, contamination from bigger organisms due to breakage is the most cited explanation for these sequences[57], an hypothesis supported by the fact that studies where different sample sorting techniques have been used, such as Flourescence Activated Cell Sorting (FACS) have failed to find ciliate organisms in this fraction[34]. Contamination is indeed a likely possibility in our samples, specially for the Lynnella-like sequences. However, some of them, specially those with low similarity to cultured species, may represent novel groups 67 of ciliates present in the picoplankton, as has been theorized by other authors[68], a possibility specially relevant for our Choerotrichid sequences. Members of other alveolate groups were not found, like in similar surveys; most importantly, we found no sequences from apicomplexans, a group of exclusively parasitic organisms whose presence in marine environments seems to be more associated with benthic ecosystems[80, 10]. This is probably the result of a real scarcity of this group in pelagic habitats, since being obligate parasites, they are usually found in the coelomic and gastrointestinal cavities of metazoans[200].

3.2.2 Chlorophyta

With the alveolates, chlorophytes were the most abundant group in our study, comprising 32% of the total library with 81 clones. Among this group, all of our sequences belonged to the class Prasinophyceae, one of the oldest lineages within this division and which is characterized by small, photosynthetic cells that in some genera present scales and/or flagella. Members of this group are among the most common components of the picoplankton[3]. The phylogenetic analysis of these sequences is shown in Figure 3.12. Prasinophyceans were well represented in our clone library, as is the case with similar studies[13], showing a relative low diversity when compared to the other phyla. Most of our clones (97.5%) fall within the Mamiellales, which constitute the most abundant eukaryotic photosynthetic group in marine samples, and more specifically to the species Micromonas pusilla, Ostreococcus sp. and Bathycoccus prasinos. Most of the Mamiellales found in our study belong to the species M. pusilla, constituting 37% of this group and 11.9% of the total number of sequences, making it the most abundant eukaryotic species in our clone library. There are at least five different lineages in the Micromonas genus[104, 107, 57], with our sequences falling within four of these. Nomenclature for those clades has not been standardized, and 68 for our study we will be employing the nomenclature established by Worden[57].

Clade B

Ostreococcus

86/67

86/84 100/100 95/81

73/52

88/89 Clade A 98/97

85/50

Bathycoccus

64/-- Mamiellales

99/97 100/97

78/80 Clade A.BC.1

100/100

100/100 87/86 Clade A.A.2

100/100 80/77 Micromonas 100/100 58/57 Clade B._.4

100/100

87/86

86/87 Clade B.E.3

100/100

85/85

98/98 99/99 100/100 Lineage A Clade VII 99/97

Figure 3.12: SSU rDNA phylogenetic tree for Chlorophyta. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7. 69

The structure of lineages A.BC.1 and A.A.2 is not clear in our tree, showing a greater structural complexity that the one established by previous studies. Originally established as a single lineage by Guillou et al[104], this group was posteriorly divided into these two distinctive, although not well-supported clades[57]. While these previously established groups do appear in our tree, some of the newly added Red Sea sequences fall outside these branches, suggesting the existence of new lineages or sublineages within this group. Four of our sequences clearly fall within clade A.A.2, and they are closely related to cultured strains from the Provasoli-Guillard Center for Culture of Marine Phy- toplankton (CCMP), CCMP489 and CCMP1723, isolated from the North Atlantic and the Mediterranean, respectively. Strain CCMP489 was isolated from the deep euphotic zone, at 120 m, so its close relationship with our water surface clones is unexpected. A slightly more distantly related strain from the Roscoff Culture Collec- tion (RCC), RCC299, was isolated from surface waters of the Equatorial Pacific. It is worth noting that ours are the first environmental sequences retrieved for lineage A.A.2, which previously was composed exclusively from cultured strains[57]. The other Lineage A sequences fall within already described clades, and are not closely related with cultured strains. Given the high variability among the preexist- ing sequences of these lineages, which range from the Pacific[201, 57], Atlantic[104], Mediterranean[107] and even from fjord water samples (unpublished study, accesion number FR874391), it is hard to correlate these phylogenetic clusters with specific ecological niches or habitats. Most of the cultured strains in this clade present an optimal growth temperature between 18◦C and 26◦C, which encompasses the conditions in which our samples were taken. The biggest group of our Micromonas sequences are found within clade B. .4, a clade composed only by environmental sequences from the coastal Pacific[57] and from Blanes Bay, in the Mediterranean[7]. Our clones not only increase the coverage 70

of this clade, but also show two well supported branches within the clade - one that includes a few of our sequences as well as the ones present in available databases, and another composed exclusively by Red Sea sequences. Metagenomic analyses seem to suggest that member of this clade are also present in the Sargasso Sea during the Sorcerer II expedition[57], but pending further analysis of this, this clade seem to be composed by Mamiellales adapted for coastal environments that are present in numerous basins. They may also present specific adaptations for this environment that complicate their culturing, which would partly explain why this is the only clade without any cultured strain so far. The rest of our Micromonas sequences are part of clade B.E.3, where they cluster with multiple sequences from the coastal Pacific. These and other Pacific sequences were taken in much colder conditions that those found in the Red Sea, being isolated from water samples with temperatures of 20.7 ◦C and 17.6◦C. The same clade also contains the cultured strain CCMP2099, isolated from an Arctic polynya, and which has an optimal growth temperature of 6 - 8◦C with low light saturation[105]. This strains appears to constitute a unique psychrophilic Micromas strain with a pan- Arctic distribution[105, 202]. It would seem that members of these clades present a remarkable capability for adaptations to different environmental conditions, unlike the situation with other ecologically similar prokaryotic groups such as Prochlorococ- cus, whose phylogenetic clades seem to correspond with specific ecotypes[203]. Sim- ilar conclusions have been drawn from phylogenomic analyses of multiple cultured strains[107]. Still, more comprehensive sampling and analysis in a greater diversity of environments is required in order to get a more accurate vision of the diversity and characteristics of Micromonas sp. organisms in the world oceans. M. pusilla is the only species described in its genus, and these findings seem to support the general consensus that it actually contains a number of cryptic species[107, 57], most of them highly dispersed among marine environments, with the Red Sea not 71

being an exception - actually, given the relatively small size of our clone library, it is encouraging to find such a big diversity of Micromonas sequences encompassing most of the know groups, including a number of potentially novel clades, such as those in lineage A, and the novel branch within clade B. .4 . While it is tempting to infer that these constitute endemic strains that may be uniquely adapted to the prevalent conditions on the basin, culturing efforts should be made in order to properly characterize this organisms before reaching any definitive conclusions. Unlike most other molecular surveys of phytopicoplankton, we found almost as many sequences clustering with Ostreococcus than with Micromonas (28 and 30, respectively). The high abundance of Ostreococcus in surface waters compared to other environments is unusual, although it has been found that in other coastal areas there is an alternation of dominance between these two species[57, 111]. This may also reflect the oligotrophic conditions prevalent in the Red Sea, which favour smaller organisms with less nutrient demands. Another possible hypothesis is that the physical and chemical characteristics present at the sampling site during the survey may have favoured Ostreococcus for unknown reasons; Ostreococcus is known to produce sporadic blooms[111], and that the presence of one of these blooms during the sampling may explain the patterns observed. While four clades have been recognized so far within Ostreococcus tauri using phylogenies based in the ITS region[104, 28], all of our sequences seem to be related to strains RCC143 and RCC393, isolated from the tropical Atlantic and the Tyrrhe- nian sea, respectively.These strains belong to Clade OII, as determined by on 18S sequences[204], which was initially characterized as a low-light clade composed of sequences from the deeper parts of the euphotic zone with a higher chl a/chl b ratio as well as photoinhibition of its growth rate[28]. However, posterior analyses have shown that the distribution of these clades is dependent in multiple environmental factors other than irradiance, concluding that clade OII seems to be generally abundant and 72

associated with warm, more saline waters[204], such as our sampling site; it is also the only clade located in the Indian Ocean[89]. We found 21 Bathycoccus prasinos sequences. The Bathycoccus genus contains this single species, characterized by scaled cells with an average diameter of 0.5 - 1.0 µm. Unlike other Mamiellales, it constitutes a genetically homogeneous group[104], a fact consistent with the clustering of our clones with sequences from environmental samples and cultured strains. Only two clones diverged from the main branch, appearing as basal groups: clone RS.10m.200 which presents several small indels in its SSU sequence and clone RS.10m.200 which contains a divergent v9 region, compared with all the other Bathycoccus. As for the non-Mamiellales sequences, they consist of two clones that are highly related to environmental sequences obtained from the English Channel[68] and to existing cultures of unidentified coccoid green algae from the RCC and the CCMP. These organisms belong to lineage A of Clade VII of the Prasinococcales, as described by Guillou et al[104] which includes the species Picocystis salinarum, a green alga isolated from saline ponds[205]. Described members of Clade VII are coccoid organisms with an average diameter of ≤2 µm. Studies with flow citometry and PCR amplification on the southeast- ern Pacific Ocean show that they are abundant both on surface waters and in the DCM of mesotrophic and mildly oligotrophic waters[201]. In other basins, such as the Mediterranean, they seem to be restricted to the deeper layers of the euphotic zone[113].

3.2.3 Stramenopiles

Stramenopiles are, along with the alveolates, one of the most common groups captured in molecular surveys. In our library they were represented by 52 sequences, which amounts to 20.6% of the total number of clones, a percentage similar to to those 73

reported in analogous surveys. A good percentage of these belonged to phototrophic taxa, which can be seen in Figure 3.13

100/100 98/97 Dictyochales 89/68

99/100 Dictychophyceae 86/83 Clade A 100/100 Pedinellales

100/100 Clade B 100/100

71/71 52/-- 97/97 100/100 Pinguiophyceae

63/58 100/100 Clade RS4 100/100

100/100

98/98 100/100 Clade H 100/100 100/100

100/100 Chrysophyceae 96/97 Clade G 75/75

100/100

47/32 100/100 71/80 100/99 Clade I

--/50 88/87

64/63 100/100

57/57 100/100 Bolidophyceae 76/76

100/100

Figure 3.13: SSU rDNA phylogenetic tree for phototrophic Stramenopiles. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

The greatest diversity for the Stramenopiles was within Chrysophyceae, with 12 diﬀerent sequences divided into 7 distinctive OTUs. An initial analysis of our sequences describes them as distantly related to members of the Synurophyceae, specifically to Mallomonas sp. Synurids are photosynthetic ﬂagellated protists, abundant in freshwater protistan communities[158]; while traditionally considered a class[63], this group is usually hard to distinguish from its sister clade, the Chrysophyceae, even with full-length SSU sequences[45], and recent studies have shown them to be a 74

clade within it[206, 87]. The most described chrysophyte sequences in coastal areas are related to organisms from the genus Paraphysomonas, which fall within the 4 - 8 µm size range; this genus, while easily isolated in enrichments, does not seem to be ecologically relevant in this ecosystem[84]. However, and despite the fact that this group is not very abundant in clone libraries of marine sites, at least when compared with freshwater ones[87], a number of novel environmental groups have been found in numerous molecular surveys, which challenges the initial assumptions that chrysophytes in general are of little importance marine environments. All of our clones belong to three of these clades composed exclusively by environmental sequences - Clade H, G and I, as per the classification by del Campo[87]. Clade G contains sequences from both marine and freshwater systems. Marine sequences are mostly represented by picophytoplankton from surface oligotrophic waters from the southeast Pacific gyre[201] and the Sargasso Sea[69]; we found four sequences in this clade that presented the highest similarity with the Sargasso Sea clones, with the exception of one OTU represented by a single sequence, which showed relatively low identity scores to available sequences (less than 95%). Clade I is composed exclusively by marine sequences, most of them from the same oligotrophic waters of the Pacific. Clade I was the less abundant clade in our library, with 3 sequences that represented two OTUs, both very similar to available Pacific sequences. This close relation between samples taken from the open ocean with coastal ones is common among marine eukaryotes, and is reported among the chrysophytes as well[201]; for this clade, all environmental samples come from either open ocean samples[201] or from seawater incubations[87, 207], making ours the first reported clones from an environmental survey of a coastal setting. Clade H contains a monophyletic group of marine sequences from diverse regions, with our sequences distributed in two different branches. One of these contains a 75 single OTU with three sequences related to sequences from research on the response from picoplankton to ocean acidification (Accesion number FR874556, unpublished study), while the other contained two sequences, each representing a single OTU and that are related to unamended incubations with water samples from the Norwegian sea coast[94], further confirming the wide distribution of these clades. An interesting observation of these last two OTUs is that they show a characteristic insertion of approximately 120 bp in the SSU V5 region; this insertion is also observed in chrysophyte partial sequences retrieved from the deep euphotic zone of fjords from the Pacific coast of Canada[208], which may form part, along our sequences, of a novel subclade within Clade H. Organisms belonging to these three novel clades have been found in other clone libraries in the picoplanktonic size fraction, which indicates that their presence in our surveys is not the result of contamination from bigger organisms. They seem to be most abundant in oligotrophic environments, specifically in oceanic gyres from the Pacific[201] and the Indian[89] oceans, which is consistent with our sampling site; additionally, plastid surveys in the nearby Arabian sea were dominated by chrysophyte- related sequences[209]. The three clades have been shown to contain both pigmented cells as well as cells isolated from dark incubations, which would indicate either some metabolic flexibility on these groups or the prevalence of mixotrophic lifestyles within them[87], a common characteristic on eukaryotic picophytoplankton[210] that has been reported extensively in this group[211]. The Bolidophyceae is a class composed of two described species, Bolidomonas pacifica and B. mediterranea, both of them with an approximate diameter of 1.2 µm[212]. They are biflagellated phototrophs, which form a sister clade to the classical diatoms (Bacillariophyceae) and that form the basal branch for all the phototrophic stramenopiles. Despite the fact that only one genus has been described, a number of sequences not belonging to classified organisms have been recovered from the 76

picoplankton[8]; like the chrysophytes, they have also been detected in the Arabian sea[209]. In any case, they seem to be minor component of the phytoplankton[213]. In our samples, they were represented by two OTUs. One of them can be identified as a strain of B. pacifica, being very closely related to strains OLI46SE and OLI41SA, which were isolated from mesotrophic waters at the equatorial Pacific at a depth of 15 m[213]. Classified as B. pacifica due to its origin and flagellar characteristics, more detailed genetic analysis determined that these strains form a clade separate from both B. pacifica and B. mediterranea, which was denominated B. pacifica var. eleuthera. The other OTU does not cluster with other available sequences, being only moderately related to the type strain of B. pacifica. It appears as a basal branch of all other Bolidophyceae, and it may represent either a novel variety of B. pacifica, or even a new species within the genus. Given that the Bolidophyceae acclimate well to culture conditions, future work in coastal areas of the Red Sea may prove fruitful to obtain representatives of this clade and increase our knowledge about the diversity of this group. The Dictychophyceae are phototrophic eukaryotes, usually characterized by large protists such as the silicoflagellates. They are a low abundance group in picoeukaryotic surveys[89, 69, 201, 209], although they may be more ecologically important in temperate and polar ecosystems, based on genetic surveys in these environments[6, 83] . The class is divided into three orders, and we could find clones belonging to two of them, the Dyctiochales and the Pedinellales. The Dictyochales contained two of our OTUs; this group was originally characterized by the large species of the genus Dictyocha, but it was later found to include the picoplanktonic species Florenciella parvula, initially isolated from English Channel waters[83]. Three of our sequences, forming a single OTU, belonged to this species. Described species usually measure from 3 to 6 µm. 77

A singleton, RS.10m.0020, while clustering within the Florenciella branch of the Dictyochales with high bootstrap support, does not seem to be similar to any isolated sequences from this genus. It may represent a novel organism belonging to the Florencialles, but being a single sequence we cannot make further assumptions regarding its taxonomic position, or ecologic role. Florencialles-related sequences have been found to be one of the main groups associated with phagotrophy in incubation experiments with radio-labeled bacteria, which suggests that like other groups within the stramenopiles, some species may exhibit a mixotrophic behaviour[214]. The other OTU was part of the Pedinellales, a class which contains both phototrophic and heterotrophic species, characterized by tentacles supported by a mi- crotubular structure[83]. They are mostly rare in marine habitats, with exceptions in areas like the English Channel[16] or the Baltic sea, where they show periodic blooms[36]. Like in the previous case, this OTU appears in a sister branch to all other pedinel- lid sequences, both from described species and environmental samples, including its closest sequences as determined by BLAST, which presented a low identity value (94%). There have been no novel environmental clades detected within the Dicty- chophyceae so far, a fact that may be due to its generally low abundance in marine ecosystems. Our sequences may conform the first case of a novel branch within this group, and in our tree we have divided the Pedinellales into two distinctive clades: Clade A composed by all described genera so far, which also includes all publicly available environmental sequences, and Clade B containing our Red Sea sequences. Ideally, further sampling using Pedinellales-specific primers will help to retrieve more sequences from more different environments in order to determine the true diversity of this group. First described based on their unusual composition of polyunsaturated fatty acids concentration, the Pinguiophyceae are a monophyletic group of stramenopiles with 78

a phototrophic lifestyle[215] that includes five monotypic genera. Of these, three species - Pinguiochrysis pyriformis[216], Phaeomonas parva[82] and Pinguiococcus pyrenoidosus[217] - form part of the picoplankton. While they can be cultured from water samples with relative ease, they very rarely appear on molecular surveys[30, 124] , suggesting that they probably represent a very minor component of the eukaryotic community. We could only find a Pinguiophyceae clone in our library, belonging to the genus Phaeomonas. The type species for the genus, P. parva, is a biflagellated spherical cell with a 3.2 - 4.0 µm diameter; however, our sequence differs from all other identified Phaeomonas strains, and it may represent a novel strain for this species, or even a new species within the genus. P parva was originally isolated from the equatorial western Pacific[82], so this is also the first Phaeomonas sequence isolated from a coastal site. Another OTU, represented by two sequences, appears as a sister group to the Pinguiophyceae, although low bootstrap support prevents us from including them into this class. They are not similar to other sequences with the exception of a single partial SSU clone (accesion number GU823241) isolated from the water column at the Cariaco basin[1]. The samples were taken from a depth of over 200 m, below the photic zone, which would initially suggest that they may belong to heterotrophic organisms. We are naming this novel clade tentatively as Clade RS4, as we did with our novel alveolate clades; still, further analysis and sampling would be required to properly characterize this group and its position within the stramenopiles. Low support for the position in phylogenetic trees of the phototrophic stramenopile groups has been reported multiple times, specially with the Pinguiophyceae, Chrysophyceae, Eu- stigmatophyceae and Raphidophyceae; the position for these classes is very variable, even when using multiple gene phylogenies[81], and we can expect similar uncertain- ties with the position of Clade RS4. 79

51/51 100/100

73/68

71/70 Bicosoecida 73/60 58/63

47/32

97/94 100/100

100/100 MAST-2 *

50/52

82/79

99/99 MAST-7 98/98 68/68 92/93

100/100 100/100

100/100 57/51

100/100 63/67

92/95

79/83

98/98

94/96 MAST-4

100/100

75/--

100/100 100/100

53/-- 94/93 90/81 MAST-6 100/100

100/100 56/56 MAST-11 100/100

100/100

100/100 100/100 95/97 100/100 100/100 MAST-3 100/100

99/99 100/100 100/100 87/84

100/100 98/97

98/97 100/100

100/100 MAST-12 100/100 100/100

Figure 3.14: SSU rDNA phylogenetic tree for heterotrophic Stramenopiles. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7. 80

The rest of the Stramenopiles belonged to heterotrophic groups, which can be seen in Figure 3.14. They were constituted almost exclusively by members of the marine stramenopiles (MAST) groups, with the exception of one sequence related to the Bicosoecida. Bicosoecids constitute the heterotrophic basal branch of the stramenopiles, and they are typically biflagellated[218, 206]; the most common marine species being Cafeteria roenbergensis[87]. Like other described stramenopile clades, they are rarely found in molecular surveys, specially in the euphotic zone. They appear to be more diverse and important in freshwater sites[87] as well as some extreme environments, such as anoxic water layers[91, 37] or in oil spill-associated communities[181]. We found one sequence that groups within the Bicosoecida with good support; however, it is very distant to all other available bicosoecid sequences, its closest relatives being the especies Caecitellus parvulus and Cafeteria sp., with a similarity of 83% and 82% respectively. Phylogenetically it seems to be located in the order Anoecales, that includes the genera Cafeteria and Symbiomonas[206], although the support for this position is low. While recent phylogenetic analyses have enhanced the taxonomy of these groups, uncovering novel clades, our sequence does not seem to belong to any of these groups, which were restricted to freshwater environments[87]. Approximately half of our stramenopile sequences belonged to the novel marine stramenopiles groups (MAST), which have been reported from a multitude of marine ecosystems, and that, with the exception of MAST-3, are composed exclusive of uncharacterized environmental sequences. This percentage falls within the values of similar surveys[7], although the proportion between MAST groups and described stramenopiles is unusual for marine environments, which are usually dominated by novel groups. The highest percentage of our MAST sequences was in the clades MAST-4, -7 and -3, which are more abundant in coastal areas[96, 7] and that, together with MAST-1, 81 account for up to 74% of novel stramenopile sequences in clone libraries of marine environments. MAST-4 in particular is one of the most abundant stramenopile taxon, accounting for up to 10% of all heterotrophic flagellates in non-polar ecosystems. This group constitutes a quantitatively important component of marine ecosystems, based on FISH studies and their substantial abundance in numerous clone libraries[97]. It was the same case in our samples, where MAST-4 showed the biggest diversity, with 6 OTUs. Two of these were strongly related (99%) with sequences from the eastern coastal Pacific[57], retrieved from the <2 µm fraction, while other three OTUs appeared in a sister branch to this group, which also included clones from the deep euphotic zone of the oligotrophic equatorial Pacific[23], the Mediterranean under upwelling conditions[6] and the Sargasso Sea[69]. While these related sequences were isolated from a coastal environment and from a similar sized fraction, the environmental conditions were very different from the ones present in our sampling site, with an average temperature of 19.8◦C and a salinity of 33.7 PSU[219]. The last sequence on MAST-4 is RS.10m.321, which although can be consistently identified as part of this clade, is only moderately related to the other members of it, with a similarities of 95% or less to them. It also seems to represent a basal branch to the rest of the group. This clade is composed of small flagellated heterotrophic organisms as shown by FISH studies, with a 2 - 3 µm diameter in average[207], which are predators of bacteria[99]. While the organisms in the Red Sea may present variations in terms of morphology, they probably fulfill a similar ecological role, functioning as picoplanktonic bacterivores. Not much research has been done regarding prey selectivity in the stramenopile flagellates; initial analyses indicate that MAST-4 organisms show no specific preference between bacterial species[99], although single cell sequencing has shown a tentative predation association between MAST-4 and the bacteria Pelagibac- 82

ter ubique, a member of the SAR11 clade, the single most common clade in the Red Sea[152]. Given that MAST-4 is the most common taxon in our library associated with bacterivory, this suggest a potential ecological relationship between these two abundant groups of marine microorganisms. The following group in terms of diversity was MAST-7, with 3 different OTUs. They all presented similarities with sequences from surface coastal areas; two of them were part of a cluster of coastal sequences found in the Pacific and in fjord waters(unpublished), while the other was highly similar to sequences from Blanes Bay, in the Mediterranean[7]. Sequences in MAST-3 were distributed into two branches, one of them composed by sequences from open ocean oligotrophic waters from the Sargasso Sea[69] and the Pacific[23]. The other branch is much more longer and diverges from the rest of this group, and its conformed by one of our clones and a sequence from fjord Arctic waters[161]. While MAST-3 contains the only described species for this groups, the grazer Solenicola setigera, our sequences do not fall in the branches containing this species sequences. Other members of this clade has been shown to be 2 - 3 µm bacterivores by FLB experiments[35], so there seems to be a notable ecological and functional diversity within MAST-3. MAST-11 is one of the smallest groups of novel stramenopiles, its only full length sequences being a clone isolated from the Mediterranean, in Blanes Bay, plus some partial sequences from the coastal Pacific[92]. Our sampling increases substantially the number of sequences belonging to this clade, with one OTU grouping with previously published sequences and another forming a sister branch to it. While this last OTU is composed of sequences that do not present similarity with any other particular stramenopile group (in fact, its closest sequence as determined by BLAST is the hypotrichid Rhizidiomyces apophysatus), phylogenetic analysis assigns them with good support to MAST-11, revealing an unexpected diversity within the group. 83

The rest of MAST groups appeared in low abundance, with MAST-6 and MAST- 14 having one clone each. Clone RS.10m.0056 is part of MAST-6, a small group with sequences from the Mediterranean[6], English Channel[68] and coastal Pacific[6]. At least two morphotypes have been identified for this group: a small one with 4 - 8.8 µm round cells and a larger one that can reach up to 22 µm in length[36]. The same study determined that these cells were grazers of unicellular algae, which is unusual for picoplanktonic flagellates, traditionally considered to be mostly bacterivorous. Considering the large populations of Micromonas and other prasinophytes in our samples, it could be possible that members of this groups perform a similar role in our sampling site; the related group MAST-4, while predominantly consuming bacteria, has also been shown to be capable of ingesting Micromonas and Ostreococcus cells. Meanwhile, clone RS.10m.367 was part of MAST-12, a group characterized by sequences isolated from sediments and microbial mats in low-oxygen[160] or anoxic[220, 67] environments. The presence of sequences of this group on well-oxygenated, euphotic seawater is unexpected. Members of this taxon isolated from anoxic mats in Norwegian estuaries appear as unpigmented cells with a 5 µm diameter and two anterior flagella[160], but presumably the organisms from which our sequences were isolated could present a different morphology adapted to their pelagic environment. Finally, one of the stramenopile singletons does not seem to be part of any of the recognized MAST groups. This sequence, RS.10m.00146, appears as a sister branch to MAST-2 on both NJ and ML tree topologies, but low bootstrap values prevent us from assigning it to this clade. There is the possibility that our clone is part of MAST- 2, and does not cluster with high support due to the limited number of sequences on this group; alternatively, it may belong to a novel, low-abundance MAST clade. Further sampling from the Red Sea will likely uncover more sequences related to this one, helping to clear up the phylogenetic position of it. 84

3.2.4 Rhizaria

Not as abundant as the other groups, Rhizarians represented 7% of our library, with 18 sequences. Diversity within the phyla was low however, with only 3 distinctive OTUs, two of them belonging to the Radiolaria and one the Cercozoa. Rhizarian abundance in clone libraries from marine environments presents a wide variability; some surveys present present values similar to ours[69, 89, 92], wjhile others show much higher rhizarian abundances[41, 90]. Unlike our own results, coastal surveys are usually more abundant in cercozoan sequences, with radiolarians being more predominant in open waters[7, 92] The Radiolaria-related clones included the largest OTU of the group, with 16 individual sequences, and actually, the largest OTU in our whole clone library outside the prasinophytes. This OTU presented a 99% similarity to the spumellarian Larcopyle butschlii, belonging to the family Litheliidae. This organism is a large radiolarian with an spherical body ≥80 µm of length, which is much larger than in the size fraction employed in our sampling. This species is characterized by inhabiting deep, cold and saline water with values over 34.3 PSU[221, 222]. The presence of related organisms in surface waters of the Red Sea is surprising, given the high temperatures and irradiance associated with the sampling site. 85

100/100 100/100 Chloroarachniophyta 100/100 91/92 Cercozoa

93/90 100/100

79/77 100/100

74/57 77/78 85/84

68/74 100/100 Polycystinea: Spumellarida: Litheliidae

82/83

100/99

100/100 Radiolaria 68/77

82/74

100/100 77/85 61/50 67/71 Taxopodida 100/100 100/100

62/57 95/96 100/100 100/100

83/93 100/100

100/100 98/100 68/66 RS5

100/100 100/100 100/85 100/100 100/100

100/100 Polycystinea: Spumellarida

Figure 3.15: SSU rDNA phylogenetic tree for Rhizaria. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

While radiolarians have been shown to constitute a monophyletic group by phylogenetic analyses, the same cannot be said about its main group, the class Poly- cystinea, which diverse analyses have shown to be paraphyletic[120, 69, 121]. Within Polycystinea, order Spumellarida is particularly complex, with its different families distributed among a number of branches in a clade together with classes Acantharea 86 and Taxopodida. These analyses show the Litheliidae branching together with the Taxopodida[129, 120], a position supported in our analysis. Closely related sequences have been found in very diverse environments, such as sediments in hydrothermal vents[9], Sargasso Sea eddies(unpublished, accesion number AY665072) and the Caribbean Sea[1]. Shi and colleagues also found taxopodida- related sequences in the South Pacific[201]. It is important to note that with the exception of the Pacific and Sargasso Sea surveys, all samples were taken from general eukaryote surveys, not limited to the picoplanktonic fraction. Given the high identity of our sequences to identified spec- imens to the large species L. butschlii, the general fragility of radiolarian cells[69], as well as the high number of SSU copies in this group[129], we cannot rule out the possibility of these clones being contaminants of organisms present size fractions that got trapped into the 1.2 µm filter due to cell breakage. Picoplanktonic sequences associated with the Radiolaria have been found in several studies[69, 67, 33, 95], constituting plausible novel picoradiolarian clades. While most of our sequences are more likely to constitute contamination from larger organisms, this is not the case with the second OTU, represented by one single sequence. While not belonging to any of the already mentioned novel groups, this clone appears in our tree as a sister branch to the Taxopodida-Litheliidae group, along with sequences from an unpublished study in picoplanktonic radiolarians isolated from the Pacific Ocean, and it may represent a novel clade of radiolarians so far limited to euphotic marine environments. We have designated the group as RS5, following the nomenclature conventions employed on the previous phyla. Cercozoans were represented by one sequence in our library, which corresponded to the Chlorarachniophyta. Its closest described species is Chlorarachnion reptans, but the similitude is only of 89%. The sequences appears in a parallel branch to other chlorarachniophytes, together with a closely related environmental clone from 87 the Sargasso Sea[69]; it also presents high similarity with partial sequences from the Mediterranean[7] and the English Channel[123], all of them isolated from the picoplanktonic fraction. Chlorarachniophytes are ameboid cercozoans, and the only photosynthetic group in the phylum due to a secondary acquisition of a green algal plastid[116]. They present a wide morphological diversity, but most species fall outside the size range for picoeukaryotes[223, 224] with the exception of Bigelowiella natans, with an average diameter of 5 µm[225]. They are occasionally retrieved from marine clone libraries in low abundances[123, 124], although heterotrophic cercozoans are usually more predominant. Members of the group present several different morphologies, including amoeboid, coccoid and flagellated cells, that change between species and life stages, and that differ even between strains of the same species[224], making any phylogenetic estimation regarding the morphology of uncultured organisms, such as our clone, void. Common among all its members its an autotrophic lifestyle, although mixotrophy has also been detected[225]. As with other phyla, the advent of molecular sampling dramatically increased the know diversity of cercozoans, a group which nowadays contains several novel clades, many of them represented exclusively by environmental sequences[226]. Despite this, no novel groups among the chlorarachniophytes have been described. However, given the dissimilarity of the isolated picoeukaryotic sequences to known organisms, their wide distribution and the fact that they group into phylogenetically related groups, they may represent one or more novel clades of picoplanktonic chlorarachniophytes[65] that appear to be also present in the Red Sea, as the present study shows. 88

3.2.5 Rare taxa

100/100

100/100 73/66

100/100

65/74

97/97

100/100

78/78

100/100 Haptophyta

64/66 62/67

100/100 Unknown taxon 100/100

Figure 3.16: SSU rDNA phylogenetic tree for Haptophyta and unclassiﬁed sequences. Samples from this study are shown in red color. Bootstrap values over 50% are indicated in the appropriate branches, in the order ML/NJ. Two unikont sequences (not shown) were used as an outgroup; for details see Figure 3.7.

The other eukaryotic groups found in our clone library were comprise only by three clones, as can be seen on Figure 3.16. Only one haptophyte sequence (0.3% of the total clone library) was found in our study, which would initially suggest that they constitute a group with low importance in the coastal community of the northern Red Sea; however, this abundance value is consistent with those found in other studies[41, 6, 90, 7, 89] using universal eukaryotic primers, which are now known to present a bias towards high-GC rDNA, such as the one found in haptophytes[101]. Therefore, we cannot discount the possibility of the presence of greater haptophyte community in the sampling site. Still, the observation that haptophytes seem to be generally more prevalent in open ocean environments, in contrast with coastal areas where Mamiellales are much more common[57, 227, 89] is supported by our results. The sequence that was found belonged to the Chrysochromulina genus, and in 89 fact it forms an OTU (as per the criteria established previously in this work) with sequences belonging to cultured representatives of C. scutellum; it is also highly similar to environmental sequences obtained from the equatorial[228] and western[102] Pa- cific.C. scutellum is a nanoplanktonic phototrophic organism that has been described in several environments, and falls within the B2 Clade in the Chrysochromulina genus, characterized by their saddle-shaped morphology[229]. Regarding the ecological role played by these organisms, C. scutellum is considered an obligate autotroph, although mixotrophy has been described in other species of the same genus depending on environmental circumstances[230]. Two of our sequences (clones RS.10m.00130 and RS.10m.356) belong to the same OTU and could not be assigned to any known eukaryotic group, even after exten- sive phylogenetic analysis. BLAST analysis shows them as related to a single environmental sequence found in surface waters of an unpolluted bay in the western Pacific[92] with a 95% similitude. All other hits relate them to members of the Apu- sozoa with much lower similitude values, but phylogenetic analysis discarded this relation. Trees made in both studies place this group in a clade related to the haptophytes, which would theoretically put this taxa within the proposed superclade Hacrobia[231]. Within this novel clade, there are other environmental sequences that fall within a sister branch to ours, but were not included in our tree due to their short fragment length; these other sequences belong to such dissimilar environments such as the Norwegian Sea[207], the English Channel[68] and the Southern Ocean[6], suggesting that this is a rare but widely distributed group. At least in one of this studies[207] the sequences were identified as belonging to Fungi, but neither the phylogenetic analysis nor BLAST results seem to support this. While there have been previously claims about novel eukaryotic groups, most of these has been debunked as sequences belonging to chimeras, fast-evolving organisms or already described but not sequenced species[40]. We believe that the probabil- 90

ity of these sequences being chimeras is negligible, because (a) we employed three independent methods to curate our clone library from chimeric sequences, (b) we found two distinctive but related sequences that form a consistent OTU and (c) they cluster together with environmental sequences from similar ecosystems obtained in studies using different primer pairs. It is much more likely that they belong to already known groups that are not represented in available databases and which present unusual rDNA sequences that prevent its placement within one of the existing phyla, similarly to what initially happened with the fast-evolving ciliates of the genera Myrionecta and Mesodinium[232] and members of the Apicomplexa[10]. Many well morphologically-characterized taxa are still pending molecular analysis, and these sequences may correspond to any of these groups. In any case, further analysis utilizing a different approach such as specific primers, FISH staining or other methods would be required before making any assumptions about the nature of this group.

3.2.6 Metazoa

A number of metazoan sequences were retrieved in our survey, representing contamination from larger organisms. They constituted 7.1% of our library, with 18 sequences distributed among 10 OTUs. They all belonged to common marine organisms, with most of the sequences (55.5%) representing hydrozoan cnidarians, followed by tuni- cates (27.8%) and copepods (16.7%). Metazoans are normal contaminant almost all unicellular eukaryotic surveys, including those of the picoplankton[7, 130, 201, 68], since even numerous preﬁltration steps cannot get rid of dissolved free DNA or organic material such as gametes, pellets or tissue debris[186]. This organic contamination is particularly problematic since even minute amounts of it would contribute disproportionate amount of genetic material - a single organism can contribute up to 9,000 amplicons in a 454 library[45]. 91

Chapter 4

Summary

In the present study, we characterized the picoeukaryotic diversity in surface waters oﬀ the northern Red Sea coast, which presents a distribution of taxa similar to other coastal and marine communities. Alveolates and stramenopiles constitute the main eukaryotic groups, being predominantly composed by novel environmental taxa, and together they encompass most of the diversity of the library, constituting 80% of all our OTUs. The second most abundant taxa are chlorophytes, which were found in higher abundances than in other studies, although their diversity values were much lower. Besides that, other groups including radiolarians, cercozoans and haptophytes were found in low abundances. According to our rarefaction curves and diversity estimates, we captured an approximate 75% of the picoeukaryotic community in the sampling site at the chosen OTU level of 98%. Still, there is probably a large number of taxa remaining to be uncovered in this ecosystem, specially considering the importance of the rare microbial biosphere[233], formed by hundreds of low abundance microbial species that are usually overlooked by environmental genetic surveys like these. Pyrosequencing surveys, that provide a substantial increase on sampling eﬀort compared with traditional clone libraries, have estimated that the number of rare protistan species can account for up 92

80% of the total number of OTUs[45, 136]. The true magnitude of this rare diversity is not without debate, however, as the effect of chimera formation and sequencing errors, the latter being particularly problematic with pyrosequencing, can cause an artificial overestimation of diversity of up to one order of magnitude[234, 235]. Phytoplanktonic taxa seem to be especially diverse and abundant in the sampled site, constituting 43.5% of all our sequences, and distributed between almost all identified eukaryotic phyla: phototrophic alveolates, stramenopiles, chlorophytes, rhizarians and haptophytes were found in our clone library, even without achieving complete coverage. Of these, chlorophytes are the main group, especially the Mamiellales, with clones representing all picosized genera within the group, Ostreoccocus, Bathycoccus and Micromonas. While these organisms are associated with coastal, oligotrophic environments[44, 227, 13] such as the one surveyed, it is interesting to note their co- occurrence in a single sample in similar abundances, which is not common; usually these genera are distributed differentially among nutrient and depth gradients, with dominant populations changing with environmental parameters[89, 113]. Given that we sampled a single point in time, it can be argued that the abundance of chlorophytes could be associated with transient phenomena, such as the periodic blooms so commonly recorded for the group. However, the high incidence of phototrophic stramenopile taxa is not so easily explained by this; surveys on coastal areas usually do not capture these groups[68] or find them in low abundances[7, 57]. The breadth of these groups - which include six different classes and one tentative novel group, RS4 - makes any single, temporary explanation unlikely. Other groups that were found in low abundance consisted also on phototrophic organisms, such as the cercozoan chloroarachnophytes and the prymnesiophytes; most likely, deeper sampling would uncover further sequences and diversity within these groups. In light of all this, it seems that the saline, oligotrophic conditions of the clear surface waters of the northern Red Sea are a propitious environment for primary pro- 93 duction on the picoeukaryotic fraction, producing a higher diversity of phototrophic taxa than in other coastal areas of the world. This supports the results of other studies which show that picoplanktonic organisms are the dominant component of the phytoplankton in the Red Sea, contributing up to 77% of the primary productivity[236]. As for the heterotrophic taxa, their abundance and taxonomy is consistent to that observed in other surveys. Novel alveolates are the major contributors to these groups, with NA-II sequences forming more than a quarter of our library. While the ecological role of many clades within this group is unknown, most of them are thought to represent parasites of other organisms, including ciliates, dinoflagellates, and rhizarians[47]. The presence of dinoflagellate and radiolarian sequences related to possible host species in our library supports this hypothesis, which coupled with the high diversity of NA-II groups, hints at a high incidence and diversity of parasitic relationships between different size fractions of eukaryotes in this environment. In contrast, NA-I sequences were found in low abundances, while in other coastal analysis they actually outnumber other novel alveolate groups[7, 44]. The second most common heterotrophic group are the marine stramenopiles, which account for 10% of our clone library. These organisms perform a number of trophic roles, including grazing and herbivory, although bacterivory seems to be predominant. Bacterivores in these environments have a huge number of prospective prey species, including abundant bacterial species in the basin like members of the SAR11 clade[152, 100], and given the high diversity of the MAST clades retrieved (17 OTUs for 22 sequences), they probably prey on a wide diversity of bacterial species. Both herbivorores and bacterivores form an important link for the cycling of nutrient and carbon between different trophic compartments, and they constitute, together with viruses, the most important biotic regulators for bacterial communities[99]. Besides the general community structure, the presence of numerous prospective novel taxa in our samples is noteworthy. We found 5 of these taxa, that we have 94 designated with the prefix RS-, and they all may constitute novel eukaryotic groups endemic to the Red Sea and especially adapted to the unique conditions therein. Three of these clades were in the NA-II groups, which is not particularly remarkable, since the diversity of this group keeps increasing as more and more environments are sampled[47]; it is likely that further sampling in other regions of the Red Sea will increase the number of sequences belonging to these novel clades, and may potentially lead to the discovery of new ones. Clade RS4 is much more unusual, forming an apparent new branch within the phototrophic stramenopiles, a relatively well known group of protists; while some new classes have been recently described in this groups[215, 212], these have been isolated and described based on cultured strains, since they tend to occur in natural environments in low abundances. RS4 might then be the first case when a novel ocrophyte group has been initially isolated from environmental samples, and since many of these groups are amenable to culturing, its isolation and characterization present an interesting target for future research. A novel radiolarian clade was also described, composed by both Red Sea and Pacific ocean sequences. Radiolarians so far have had few in-depth phylogenetic studies considering environmental sequences[69, 95], with most studies focusing on cultured species[129, 117, 121]. Given this, and combined with the relative lack of rhizarian sequences from described species, it is very probable that a number of novel clades remain to be categorized within the group, with clade RS5 among them. Lastly, we found an OTU that could not be ascribed to any known eukaryotic group. Intensive sequence pre-processing and the presence of phylogenetically related sequences from another marine environment[92, 207, 68] discarded the possibility that this sequences may be chimeric, or the result of erroneous sequencing. As for the other goal of our research, which was to evaluate the use of the v3 region for a prospective pyrosequencing survey, we can conclude that while the primers 95 chosen are capable of capturing a sizable eukaryotic diversity, comparable with that of the complete 18S gene, they do not provide us with sequences amenable to accurate taxonomic identification. Future studies should concentrate in other regions to ana- lyze these communities, specially the v9 region, being the most widely used in current eukaryotic studies. This approach will facilitate us to make accurate comparisons of the Red Sea with other marine environments, while at the same time allowing for a more precise taxonomic identification of sequences.

4.1 Conclusions

As with other marine environments, the application of PCR clone libraries has uncovered a highly diverse picoeukaryotic community in the Red Sea, composed by taxa covering a substantial percentage of the eukaryotic known groups. Even though our coverage of this environment was not complete, we could identify a wide range of groups, including some novel ones within the Alveolata, Stramenopiles and Rhizaria divisions, as well as intriguing sequences belonging to an unknown phyla. Generally speaking, this community is characterized by a substantial diversity of primary producers, with parasites and bacterivores conforming most of the heterotrophic groups. This study provides an important ﬁrst step in the characterization of the eukaryotic component of the Red Sea microbiota, whose continued study will undoubtedly lead to a greater understanding of the ecological processes taking place in this unique environment, as well as to our more general knowledge about the role of eukaryotes in the world’s oceans.

4.2 Future Research Work

The results presented here are an initial estimation of the picoeukaryotic microbiota on the Red Sea, capturing its diversity in a single geographical area during an specific 96 point in time. Future work should focus on improving an extending this study both geographically and temporally, which can can be done through different approaches. First, the inclusion of more sites along the north-south axis of the basin, both coastal and from open waters. Second, establishing time series to evaluate the role of sea- sonality on the marine microbiota. And third, the surveying of other marine zones beyond surface waters, including not only the deep chlorophyll maximum and the aphotic zone, but also the benthic zone and other unique environments found in the Red Sea, such as its deep-sea brine pools. Additionally, the presence of clades of novel organisms related to cultured groups such as the Pedinellales and Pinguiophyceae opens the possibility that culturing approaches may yield new species susceptible of characterization. This would considerably improve our understanding of the ecological role performed by these taxa, as well as provide us with further information to shed light on their phylogenetic position within the different eukaryotic phyla. 97

BIBLIOGRAPHY

[1] V. Edgcomb, W. Orsi, J. Bunge, S. Jeon, R. Christen, C. Leslin, M. Holder, G. T. Taylor, P. Suarez, R. Varela, and S. Epstein, “Protistan microbial observatory in the Cariaco Basin, Caribbean. I. Pyrosequencing vs Sanger insights into species richness.” The ISME journal, vol. 5, no. 8, pp. 1344–1356, Mar. 2011.

[2] M. L. Cuvelier, A. Ortiz, E. Kim, H. Moehlig, D. E. Richardson, J. F. Hei- delberg, J. M. Archibald, and A. Z. Worden, “Widespread distribution of a unique marine protistan lineage.” Environmental microbiology, vol. 10, no. 6, pp. 1621–34, Jun. 2008.

[3] A. Z. Worden and F. Not, “Ecology and diversity of picoeukaryotes,” in Mi- crobial Ecology of the Oceans, 2nd ed., D. L. Kirchman, Ed. Hoboken, New Jersey: John Wiley and Sons, Inc., 2008, ch. 6.

[4] D. Vaulot, W. Eikrem, M. Viprey, and H. Moreau, “The diversity of small eukaryotic phytoplankton (¡ or =3 microm) in marine ecosystems.” FEMS microbiology reviews, vol. 32, no. 5, pp. 795–820, Aug. 2008.

[5] E. Knight-Jones, “Preliminary studies of nanoplankton and ultraplankton systematics and abundance by a quantitative culture method,” Journal du Conseil, vol. 17, no. 2, p. 140, Apr. 1951.

[6] B. Diez, C. Pedros-Alio, and R. Massana, “Study of genetic diversity of eukaryotic picoplankton in diﬀerent oceanic regions by small-subunit rRNA gene cloning and sequencing,” Applied and Environmental Microbiology, vol. 67, no. 7, p. 2932, 2001.

[7] R. Massana, V. Balagué,L. Guillou, and C. Pedrós-Alió,“Picoeukaryotic diversity in an oligotrophic coastal site studied by molecular and culturing approaches.” FEMS microbiology ecology, vol. 50, no. 3, pp. 231–43, Nov. 2004.

[8] L. K. Medlin, K. Metﬁes, H. Mehl, K. Wiltshire, and K. Valentin, “Picoeukary- otic plankton diversity at the Helgoland time series site as assessed by three molecular methods.” Microbial ecology, vol. 52, no. 1, pp. 53–71, Jul. 2006. 98

[9] V. P. Edgcomb, D. T. Kysela, A. Teske, A. de Vera Gomez, and M. L. Sogin, “Benthic eukaryotic diversity in the Guaymas Basin hydrothermal vent environment.” Proceedings of the National Academy of Sciences of the United States of America, vol. 99, no. 11, pp. 7658–62, May 2002.

[10] P. L´opez-Garc´ıa, H. Philippe, F. Gail, and D. Moreira, “Autochthonous eukaryotic diversity in hydrothermal sediment and experimental microcolonizers at the Mid-Atlantic Ridge.” Proceedings of the National Academy of Sciences of the United States of America, vol. 100, no. 2, pp. 697–702, Jan. 2003.

[11] A. Stock, H.-W. Breiner, M. Pachiadaki, V. Edgcomb, S. Filker, V. La Cono, M. M. Yakimov, and T. Stoeck, “Microbial eukaryote life in the new hypersaline deep-sea basin Thetis.” Extremophiles : life under extreme conditions, Oct. 2011.

[12] E. Marañón,P. Holligan, R. Barciela, N. González,B. Mouriño,M. Pazó, M. Varela, and Others, “Patterns of phytoplankton size structure and productivity in contrasting open-ocean environments,” Marine Ecology Progress Series, vol. 216, no. 216, pp. 43–56, 2001.

[13] F. Not, M. Latasa, D. Marie, T. Cariou, D. Vaulot, and N. Simon, “A single species, Micromonas pusilla (Prasinophyceae), dominates the eukaryotic picoplankton in the Western English Channel.” Applied and environmental microbiology, vol. 70, no. 7, pp. 4064–72, Jul. 2004.

[14] W. Li, “Primary production of prochlorophytes, cyanobacteria, and eucaryotic ultraphytoplankton: Measurements from ﬂow cytometric sorting,” Limnology and Oceanography, vol. 39, no. 1, pp. 169–175, 1994.

[15] I. R. Joint, N. J. P. Owens, and A. J. Pomroy, “Seasonal production of photosynthetic picoplankton and nanoplankton in the Celtic Sea,” Marine Ecology, vol. 28, pp. 251–258, 1986.

[16] E. Breton, C. Brunet, B. Sautour, and J. Brylinski, “Annual variations of phytoplankton biomass in the Eastern English Channel : comparison by pigment signatures and microscopic counts,” Journal of Plankton Research, vol. 22, no. 8, pp. 1423–1440, 2000.

[17] G. Tarran, J. Heywood, and M. Zubkov, “Latitudinal changes in the standing stocks of nano- and picoeukaryotic phytoplankton in the Atlantic Ocean,” Deep Sea Research Part II: Topical Studies in Oceanography, vol. 53, no. 14-16, pp. 1516–1529, Jul. 2006. 99

[18] F. Azam, T. Fenchel, J. Field, J. Gray, L. Meyer-Reil, and F. Thingstad, “The ecological role of water-column microbes in the sea,” Marine ecology progress series, vol. 10, no. 3, pp. 257–263, 1983.

[19] J. Pernthaler, “Predation on prokaryotes in the water column and its ecological implications.” Nature reviews. Microbiology, vol. 3, no. 7, pp. 537–46, Jul. 2005.

[20] L. Jardillier, Y. Bettarel, M. Richardot, C. Bardot, C. Amblard, T. Sime- Ngando, and D. Debroas, “Eﬀects of viruses and predators on prokaryotic community composition.” Microbial ecology, vol. 50, no. 4, pp. 557–69, Nov. 2005.

[21] A. Chambouvet, M. Laabir, M. Sengco, A. Vaquer, and L. Guillou, “Ge- netic diversity of Amoebophryidae (Syndiniales) during Alexandrium catenel- la/tamarense (Dinophyceae) blooms in the Thau lagoon (Mediterranean Sea, France).” Research in microbiology, vol. 162, no. 9, pp. 959–68, Nov. 2011.

[22] S. Giovannoni, T. Britschgi, C. Moyer, and K. Field, “Genetic diversity in Sargasso Sea bacterioplankton,” Nature, vol. 345, pp. 60–63, 1990.

[23] S. Y. Moon-van der Staay, R. De Wachter, and D. Vaulot, “Oceanic 18S rDNA sequences from picoplankton reveal unsuspected eukaryotic diversity.” Nature, vol. 409, no. 6820, pp. 607–10, Feb. 2001.

[24] “A method for studying protistan diversity using massively parallel sequencing of V9 hypervariable regions of small-subunit ribosomal RNA genes.” PloS one, vol. 4, no. 7, p. e6372, Jan. 2009.

[25] E. Kim, J. Harrison, S. Sudek, M. Jones, H. Wilcox, T. Richards, A. Worden, and J. Archibald, “Newly identiﬁed and diverse plastid-bearing branch on the eukaryotic tree of life,” Proceedings of the National Academy of Sciences, vol. 108, no. 4, p. 1496, 2011.

[26] B. J. Finlay, “Global dispersal of free-living microbial eukaryote species.” Sci- ence (New York, N.Y.), vol. 296, no. 5570, pp. 1061–1063, May 2002.

[27] W. Marande, P. L´opez-Garc´ıa,and D. Moreira, “Eukaryotic diversity and phylogeny using small- and large-subunit ribosomal RNA genes from environmental samples.” Environmental microbiology, vol. 11, no. 12, pp. 3179–3188, Dec. 2009.

[28] F. Rodr´ıguez,E. Derelle, L. Guillou, F. Le Gall, D. Vaulot, and H. Moreau, “Ecotype diversity in the marine picoeukaryote Ostreococcus (Chlorophyta, 100

Prasinophyceae).” Environmental microbiology, vol. 7, no. 6, pp. 853–9, Jun. 2005.

[29] S. Robbens, E. Derelle, C. Ferraz, J. Wuyts, H. Moreau, and Y. Van de Peer, “The complete chloroplast and mitochondrial DNA sequence of Ostreococcus tauri: organelle genomes of the smallest eukaryote are examples of compaction.” Molecular biology and evolution, vol. 24, no. 4, pp. 956–68, Apr. 2007.

[30] C. Lep`ere,D. Vaulot, and D. J. Scanlan, “Photosynthetic picoeukaryote community structure in the South East Paciﬁc Ocean encompassing the most oligotrophic waters on Earth.” Environmental microbiology, vol. 11, no. 12, pp. 3105–17, Dec. 2009.

[31] I. Probert, J. Fresnel, C. Billard, M. Geisen, and J. R. Young, “Light and Elec- tron Microscope Observations of Algirosphaera Robusta (Prymnesiophyceae),” Journal of Phycology, vol. 43, no. 2, pp. 319–332, Apr. 2007.

[32] R. Andersen, R. Bidigare, and M. Latasa, “A comparison of HPLC pigment signatures and electron microscopic observations for oligotrophic waters of the North Atlantic and Paciﬁc Oceans,” Deep Sea Research Part II:, vol. 43, no. 2-3, pp. 517–537, 1996.

[33] D. Marie, F. Zhu, V. Balagu´e,J. Ras, and D. Vaulot, “Eukaryotic picoplankton communities of the Mediterranean Sea in summer assessed by molecular approaches (DGGE, TTGE, QPCR).” FEMS microbiology ecology, vol. 55, no. 3, pp. 403–15, Mar. 2006.

[34] J. L. Heywood, M. E. Sieracki, W. Bellows, N. J. Poulton, and R. Stepanauskas, “Capturing diversity of marine heterotrophic protists: one cell at a time.” The ISME journal, vol. 5, no. 4, pp. 674–684, Oct. 2010.

[35] R. Massana, L. Guillou, B. D´ıez,and C. Pedr´os-Ali´o,“Unveiling the organisms behind novel eukaryotic ribosomal DNA sequences from the ocean,” Applied and environmental microbiology, vol. 68, no. 9, p. 4554, 2002.

[36] K. Piwosz and J. Pernthaler, “Seasonal population dynamics and trophic role of planktonic nanoﬂagellates in coastal surface waters of the Southern Baltic Sea.” Environmental microbiology, vol. 12, no. 2, pp. 364–77, Feb. 2010.

[37] T. Stoeck, B. Hayward, G. T. Taylor, R. Varela, and S. S. Epstein, “A multiple PCR-primer approach to access the microeukaryotic diversity in environmental samples.” Protist, vol. 157, no. 1, pp. 31–43, Feb. 2006. 101

[38] A. Zuendorf, J. Bunge, A. Behnke, K. J.-a. Barger, and T. Stoeck, “Diver- sity estimates of microeukaryotes below the chemocline of the anoxic Mariager Fjord, Denmark.” FEMS microbiology ecology, vol. 58, no. 3, pp. 476–91, Dec. 2006.

[39] F. Zhu, R. Massana, F. Not, D. Marie, and D. Vaulot, “Mapping of picoeu- caryotes in marine ecosystems with quantitative PCR of the 18S rRNA gene.” FEMS microbiology ecology, vol. 52, no. 1, pp. 79–92, Mar. 2005.

[40] C. Berney, J. Fahrni, and J. Pawlowski, “How many novel eukaryotic ’king- doms’? Pitfalls and limitations of environmental DNA surveys.” BMC biology, vol. 2, p. 13, Jan. 2004.

[41] F. Not, J. del Campo, V. Balagu´e,C. de Vargas, and R. Massana, “New insights into the diversity of marine picoeukaryotes.” PloS one, vol. 4, no. 9, p. e7143, Jan. 2009.

[42] F. V. Wintzingerode, U. B. G¨obel, and E. Stackebrandt, “Determination of microbial diversity in environmental samples: pitfalls of PCR-based rRNA analysis,” FEMS Microbiology Reviews, vol. 21, no. 3, pp. 213–229, Jan. 2006.

[43] F. Sanger and a. R. Coulson, “A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase.” Journal of molecular biology, vol. 94, no. 3, pp. 441–8, May 1975.

[44] M. K. Cheung, C. H. Au, K. H. Chu, H. S. Kwan, and C. K. Wong, “Compo- sition and genetic diversity of picoeukaryotes in subtropical coastal waters as revealed by 454 pyrosequencing.” The ISME journal, vol. 4, no. 8, pp. 1053–9, Aug. 2010.

[45] T. Stoeck, A. Behnke, R. Christen, L. Amaral-Zettler, M. J. Rodriguez-Mora, A. Chistoserdov, W. Orsi, and V. P. Edgcomb, “Massively parallel tag sequencing reveals the complexity of anaerobic marine protistan communities.” BMC biology, vol. 7, no. i, p. 72, Jan. 2009.

[46] J. C. Dohm, C. Lottaz, T. Borodina, and H. Himmelbauer, “Substantial biases in ultra-short read data sets from high-throughput DNA sequencing.” Nucleic acids research, vol. 36, no. 16, p. e105, Sep. 2008.

[47] L. Guillou, M. Viprey, A. Chambouvet, R. M. Welsh, a. R. Kirkham, R. Mas- sana, D. J. Scanlan, and a. Z. Worden, “Widespread occurrence and genetic diversity of marine parasitoids belonging to Syndiniales (Alveolata).” Environ- mental microbiology, vol. 10, no. 12, pp. 3349–65, Dec. 2008. 102

[48] D. Prescott, “The DNA of ciliated protozoa.” Microbiology and Molecular Bi- ology Reviews, vol. 58, no. 2, pp. 233–267, 1994.

[49] G. Dover, “Molecular drive: a cohesive mode of species evolution.” Nature, vol. 299, no. 5879, pp. 111–7, Sep. 1982.

[50] A. P. Rooney, “Mechanisms underlying the evolution and maintenance of func- tionally heterogeneous 18S rRNA genes in Apicomplexans.” Molecular biology and evolution, vol. 21, no. 9, pp. 1704–11, Sep. 2004.

[51] S. Carranza, J. Baguñà,and M. Riutort, “Origin and evolution of paralogous rRNA gene clusters within the flatworm family Dugesiidae (Platyhelminthes, Tricladida).” Journal of molecular evolution, vol. 49, no. 2, pp. 250–9, Aug. 1999.

[52] R. Massana, B. Karniol, T. Pommier, I. Bodaker, and O. B´ej`a,“Metagenomic retrieval of a ribosomal DNA repeat array from an uncultured marine alveolate.” Environmental microbiology, vol. 10, no. 5, pp. 1335–43, May 2008.

[53] R. W. Litaker, M. W. Vandersea, S. R. Kibler, K. S. Reece, N. a. Stokes, F. M. Lutzoni, B. a. Yonish, M. a. West, M. N. D. Black, and P. a. Tester, “Recognizing Dinoﬂagellate Species Using ITS rDNA Sequences,” Journal of Phycology, vol. 43, no. 2, pp. 344–355, Apr. 2007.

[54] A. R. D. Ganley and T. Kobayashi, “Highly eﬃcient concerted evolution in the ribosomal DNA repeats: total rDNA repeat variation revealed by whole-genome shotgun sequence data.” Genome research, vol. 17, no. 2, pp. 184–91, Feb. 2007.

[55] T. Stoeck, D. Bass, M. Nebel, R. Christen, M. D. M. Jones, H.-W. Breiner, and T. a. Richards, “Multiple marker parallel tag environmental DNA sequencing reveals a highly complex eukaryotic community in marine anoxic water.” Molecular ecology, vol. 19 Suppl 1, pp. 21–31, Mar. 2010.

[56] G. Piganeau, A. Eyre-Walker, N. Grimsley, and H. Moreau, “How and why DNA barcodes underestimate the diversity of microbial eukaryotes.” PloS one, vol. 6, no. 2, p. e16342, Jan. 2011.

[57] A. Worden, “Picoeukaryote diversity in coastal waters of the Paciﬁc Ocean,” Aquatic Microbial Ecology, vol. 43, pp. 165–175, Jun. 2006.

[58] S. Kim, M. G. Park, K.-Y. Kim, C.-H. Kim, W. Yih, J. S. Park, and D. W. Coats, “Genetic diversity of parasitic dinoﬂagellates in the genus amoebophrya 103

and its relationship to parasite biology and biogeography.” The Journal of eukaryotic microbiology, vol. 55, no. 1, pp. 1–8, 2008.

[59] D. Nanney, C. Park, and R. Preparata, “Comparison of sequence diﬀerences in a variable 23S rRNA domain among sets of cryptic species of ciliated protozoa,” Journal of Eukaryotic Microbiology, vol. 45, no. 1, pp. 91–100, 1998.

[60] B. J. Finlay, “Protist taxonomy: an ecological perspective.” Philosophical trans- actions of the Royal Society of London. Series B, Biological sciences, vol. 359, no. 1444, pp. 599–610, Apr. 2004.

[61] A. Amato, W. H. C. F. Kooistra, J. H. L. Ghiron, D. G. Mann, T. Pr¨oschold, and M. Montresor, “Reproductive isolation among sympatric cryptic species in marine diatoms.” Protist, vol. 158, no. 2, pp. 193–207, Apr. 2007.

[62] H. S. Yoon, D. C. Price, R. Stepanauskas, V. D. Rajah, M. E. Sieracki, W. H. Wilson, E. C. Yang, S. Duﬀy, and D. Bhattacharya, “Single-Cell Genomics Reveals Organismal Interactions in Uncultivated Marine Protists,” Science, vol. 332, no. 6030, pp. 714–717, May 2011.

[63] S. M. Adl, A. G. B. Simpson, M. a. Farmer, R. a. Andersen, O. R. Anderson, J. R. Barta, S. S. Bowser, G. Brugerolle, R. a. Fensome, S. Fredericq, T. Y. James, S. Karpov, P. Kugrens, J. Krug, C. E. Lane, L. a. Lewis, J. Lodge, D. H. Lynn, D. G. Mann, R. M. McCourt, L. Mendoza, O. Moestrup, S. E. Mozley-Standridge, T. a. Nerad, C. a. Shearer, A. V. Smirnov, F. W. Spiegel, and M. F. J. R. Taylor, “The new higher level classiﬁcation of eukaryotes with emphasis on the taxonomy of protists.” The Journal of eukaryotic microbiology, vol. 52, no. 5, pp. 399–451, 2005.

[64] S. B. Gould, W.-H. Tham, A. F. Cowman, G. I. McFadden, and R. F. Waller, “Alveolins, a new family of cortical proteins that deﬁne the protist infrakingdom Alveolata.” Molecular biology and evolution, vol. 25, no. 6, pp. 1219–30, Jun. 2008.

[65] R. Massana, “Eukaryotic picoplankton in surface oceans.” Annual review of microbiology, vol. 65, pp. 91–110, Oct. 2011.

[66] E. Sherr, B. Sherr, R. Fallon, and S. Newell, “Small, aloricate ciliates as a major component of the marine heterotrophic nanoplankton,” Limnology and oceanography, vol. 31, no. 1, pp. 177–183, 1986.

[67] P. López-Garc´ıa, F. Rodr´ıguez-Valera, C. Pedrós-Alió,and D. Moreira, “Unex- pected diversity of small eukaryotes in deep-sea Antarctic plankton.” Nature, vol. 409, no. 6820, pp. 603–7, Mar. 2001. 104

[68] K. Romari and D. Vaulot, “Composition and temporal variability of picoeukaryote communities at a coastal site of the English Channel from 18S rDNA sequences,” Limnology and Oceanography, vol. 49, no. 3, pp. 784–798, 2004.

[69] F. Not, R. Gausling, F. Azam, J. F. Heidelberg, and A. Z. Worden, “Vertical distribution of picoeukaryotic diversity in the Sargasso Sea.” Environmental microbiology, vol. 9, no. 5, pp. 1233–52, May 2007.

[70] a. Groisillier, R. Massana, K. Valentin, D. Vaulot, and L. Guillou, “Genetic diversity and habitats of two enigmatic marine alveolate lineages,” Aquatic Mi- crobial Ecology, vol. 42, pp. 277–291, 2006.

[71] C. Bachy, P. L´opez-Garc´ıa, A. Vereshchaka, and D. Moreira, “Diversity and vertical distribution of microbial eukaryotes in the snow, sea ice and seawater near the north pole at the end of the polar night.” Frontiers in microbiology, vol. 2, no. May, p. 106, Jan. 2011.

[72] A. Harada, S. Ohtsuka, and T. Horiguchi, “Species of the parasitic genus Du- boscquella are members of the enigmatic Marine Alveolate Group I.” Protist, vol. 158, no. 3, pp. 337–47, Jul. 2007.

[73] D. W. Coats, T. R. Bachvaroﬀ, and C. F. Delwiche, “Revision of the family Duboscquellidae with description of Euduboscquella crenulata n. gen., n. sp. (Dinoﬂagellata, Syndinea), an intracellular parasite of the ciliate Favella pana- mensis Kofoid & Campbell.” The Journal of eukaryotic microbiology, vol. 59, no. 1, pp. 1–11, 2012.

[74] C. Gestal, B. Novoa, D. Posada, A. Figueras, and C. Azevedo, “Perkinsoide cha- belardi n. gen., a protozoan parasite with an intermediate evolutionary position: possible cause of the decrease of sardine ﬁsheries?” Environmental microbiology, vol. 8, no. 6, pp. 1105–14, Jun. 2006.

[75] D. Coats, E. Adam, C. Gallegos, and S. Hedrick, “Parasitism of photosynthetic dinoﬂagellates in a shallow subestuary of Chesapeake Bay, USA,” Aquatic Mi- crobial Ecology, vol. 11, pp. 1–9, 1996.

[76] J. H. Gunderson, S. H. Goss, and D. W. Coats, “FISH probes for the detection of the parasitic dinoﬂagellate Amoebophrya sp. infecting the dinoﬂagellate Akashiwo sanguinea in Chesapeake Bay.” The Journal of eukaryotic microbiology, vol. 48, no. 6, pp. 670–5, 2001.

[77] T. Stoeck, G. T. Taylor, and S. S. Epstein, “Novel Eukaryotes from the Per- manently Anoxic Cariaco Basin (Caribbean Sea),” Applied and Environmental Microbiology, vol. 69, no. 9, pp. 5656–5663, Sep. 2003. 105

[78] G. D. Stentiford and J. D. Shields, “A review of the parasitic dinoﬂagel- lates Hematodinium species and Hematodinium-like infections in marine crus- taceans.” Diseases of aquatic organisms, vol. 66, no. 1, pp. 47–70, Aug. 2005.

[79] G. Walker, R. G. Dorrell, A. Schlacht, and J. B. Dacks, “Eukaryotic systematics: a user’s guide for cell biologists and parasitologists.” Parasitology, pp. 1–26, Feb. 2011.

[80] F. Scheckenbach, K. Hausmann, C. Wylezich, M. Weitere, and H. Arndt, “Large-scale patterns in biodiversity of microbial eukaryotes from the abyssal sea ﬂoor.” Proceedings of the National Academy of Sciences of the United States of America, vol. 107, no. 1, pp. 115–120, Jan. 2010.

[81] I. Riisberg, R. J. S. Orr, R. Kluge, K. Shalchian-Tabrizi, H. a. Bowers, V. Patil, B. Edvardsen, and K. S. Jakobsen, “Seven gene phylogeny of heterokonts.” Protist, vol. 160, no. 2, pp. 191–204, May 2009.

[82] D. Honda and I. Inouye, “Ultrastructure and taxonomy of a marine photosynthetic stramenopile Phaeomonas parva gen. et sp. nov.(Pinguiophyceae) with emphasis on the ﬂagellar apparatus,” Phycological Research, vol. 50, no. Novem- ber 2001, pp. 75–89, 2002.

[83] W. Eikrem, K. Romari, M. Latasa, F. Gall, J. Throndsen, and D. Vaulot, “Florenciella parvula gen. et sp. nov.(Dictyochophyceae, Heterokontophyta), a small ﬂagellate isolated from the English Channel,” Phycologia, vol. 43, no. 6, pp. 658–668, 2004.

[84] E. Lim and M. Dennett, “The ecology of Paraphysomonas imperforata based on studies employing oligonucleotide probe identiﬁcation in coastal water samples and enrichment cultures,” Limnology and oceanography, vol. 44, no. 1, pp. 37– 51, 1999.

[85] D. A. Caron, M. R. Dennett, D. M. Moran, R. A. Schaﬀner, D. J. Lonsdale, C. J. Gobler, R. Nuzzi, and T. I. McLean, “Development and Application of a Monoclonal-Antibody Technique for Counting Aureococcus anophageﬀerens, an Alga Causing Recurrent Brown Tides in the Mid-Atlantic United States,” Applied and Environmental Microbiology, vol. 69, no. 9, pp. 5492–5502, Sep. 2003.

[86] E. Buskey, H. Liu, C. Collumb, and J. Bersano, “The decline and recovery of a persistent Texas brown tide algal bloom in the Laguna Madre (Texas, USA),” Estuaries and Coasts, vol. 24, no. 3, pp. 337–346, 2001. 106

[87] J. del Campo and R. Massana, “Emerging diversity within chrysophytes, choanoﬂagellates and bicosoecids based on molecular surveys.” Protist, vol. 162, no. 3, pp. 435–48, Jul. 2011.

[88] J. Kang, S. Kang, D.-y. Kim, and D. Kim, “Planktonic centric diatom Mini- discus chilensis dominated sediment trap material in eastern Bransﬁeld Strait, Antarctica,” Marine Ecology Progress Series, vol. 255, pp. 93–99, 2003.

[89] F. Not, M. Latasa, R. Scharek, M. Viprey, P. Karleskind, V. Balague, I. Onto- riaoviedo, A. Cumino, E. Goetze, and D. Vaulot, “Protistan assemblages across the Indian Ocean, with a speciﬁc emphasis on the picoeukaryotes,” Deep Sea Research Part I: Oceanographic Research Papers, vol. 55, no. 11, pp. 1456–1473, Nov. 2008.

[90] R. Massana, M. Pernice, J. a. Bunge, and J. del Campo, “Sequence diversity and novelty of natural assemblages of picoeukaryotes from the Indian Ocean.” The ISME journal, vol. 5, no. 2, pp. 184–95, Feb. 2011.

[91] C. Wylezich and K. J¨urgens,“Protist diversity in suboxic and sulﬁdic waters of the Black Sea.” Environmental microbiology, vol. 13, pp. 2939–2956, Sep. 2011.

[92] M. Cheung, K. Chu, C. Li, H. Kwan, and C. Wong, “Genetic diversity of picoeukaryotes in a semi-enclosed harbour in the subtropical western Paciﬁc Ocean,” Aquatic Microbial Ecology, vol. 53, no. December, pp. 295–305, Dec. 2008.

[93] R. Massana and C. Pedr´os-Ali´o,“Unveiling new microbial eukaryotes in the surface ocean.” Current opinion in microbiology, vol. 11, no. 3, pp. 213–8, Jun. 2008.

[94] R. Massana, R. Terrado, I. Forn, C. Lovejoy, and C. Pedrós-Alió,“Distribution and abundance of uncultured heterotrophic flagellates in the world oceans.” Environmental microbiology, vol. 8, no. 9, pp. 1515–1522, Sep. 2006.

[95] W. Orsi, V. Edgcomb, S. Jeon, C. Leslin, J. Bunge, G. T. Taylor, R. Varela, and S. Epstein, “Protistan microbial observatory in the Cariaco Basin, Caribbean. II. Habitat specialization.” The ISME journal, vol. 5, no. 8, pp. 1357–73, Aug. 2011.

[96] R. Rodr´ıguez-Mart´ınez, M. Labrenz, J. del Campo, I. Forn, K. J¨urgens, and R. Massana, “Distribution of the uncultured protist MAST-4 in the Indian Ocean, Drake Passage and Mediterranean Sea assessed by real-time quantitative PCR.” Environmental microbiology, vol. 11, no. 2, pp. 397–408, Feb. 2009. 107

[97] R. Massana, J. Castresana, V. Balagué,L. Guillou, K. Romari, A. Groisillier, K. Valentin, and C. Pedrós-Alió,“Phylogenetic and ecological analysis of novel marine stramenopiles.” Applied and environmental microbiology, vol. 70, no. 6, pp. 3528–34, Jun. 2004.

[98] F. Gómez,D. Moreira, K. Benzerara, and P. López-Garc´ıa,“Solenicola setigera is the first characterized member of the abundant and cosmopolitan uncultured marine stramenopile group MAST-3.” Environmental microbiology, vol. 13, no. 1, pp. 193–202, Jan. 2011.

[99] R. Massana, F. Unrein, R. Rodr´ıguez-Mart´ınez,I. Forn, T. Lefort, J. Pinhassi, and F. Not, “Grazing rates and functional diversity of uncultured heterotrophic ﬂagellates.” The ISME journal, vol. 3, no. 5, pp. 588–96, May 2009.

[100] M. Martinez-Garcia, D. Brazel, N. J. Poulton, B. K. Swan, M. L. Gomez, D. Masland, M. E. Sieracki, and R. Stepanauskas, “Unveiling in situ interactions between marine protists and bacteria through single cell sequencing.” The ISME journal, vol. 6, no. 3, pp. 703–7, Mar. 2012.

[101] H. Liu, I. Probert, J. Uitz, H. Claustre, S. Aris-Brosou, M. Frada, F. Not, and C. de Vargas, “Extreme diversity in noncalcifying haptophytes explains a major pigment paradox in open oceans.” Proceedings of the National Academy of Sciences of the United States of America, vol. 106, no. 31, pp. 12 803–12 808, Aug. 2009.

[102] M. L. Cuvelier, A. E. Allen, A. Monier, J. P. McCrow, M. Messi´e,S. G. Tringe, T. Woyke, R. M. Welsh, T. Ishoey, J.-H. Lee, B. J. Binder, C. L. DuPont, M. Latasa, C. Guigand, K. R. Buck, J. Hilton, M. Thiagarajan, E. Caler, B. Read, R. S. Lasken, F. P. Chavez, and A. Z. Worden, “Targeted metage- nomics and ecology of globally important uncultured eukaryotic phytoplankton.” Proceedings of the National Academy of Sciences of the United States of America, vol. 107, no. 33, pp. 14 679–14 684, Aug. 2010.

[103] S. McDonald, D. Sarno, D. Scanlan, and a. Zingone, “Genetic diversity of eukaryotic ultraphytoplankton in the Gulf of Naples during an annual cycle,” Aquatic Microbial Ecology, vol. 50, pp. 75–89, Dec. 2007.

[104] L. Guillou, W. Eikrem, M.-J. Chrétiennot-Dinet,F. Le Gall, R. Massana, K. Ro- mari, C. Pedrós-Alió,and D. Vaulot, “Diversity of picoplanktonic prasinophytes assessed by direct nuclear SSU rDNA sequencing of environmental samples and novel isolates retrieved from oceanic and coastal marine ecosystems.” Protist, vol. 155, no. 2, pp. 193–214, Jun. 2004. 108

[105] C. Lovejoy, W. F. Vincent, S. Bonilla, S. Roy, M.-J. Martineau, R. Terrado, M. Potvin, R. Massana, and C. Pedr´os-Ali´o, “Distribution, Phylogeny, and Growth of Cold-Adapted Picoprasinophytes in Arctic Seas,” Journal of Phycol- ogy, vol. 43, no. 1, pp. 78–89, Feb. 2007.

[106] G. Rocap, D. L. Distel, J. B. Waterbury, and S. W. Chisholm, “Resolution of Prochlorococcus and Synechococcus Ecotypes by Using 16S-23S Ribosomal DNA Internal Transcribed Spacer Sequences,” Applied and Environmental Mi- crobiology, vol. 68, no. 3, pp. 1180–1191, Mar. 2002.

[107] J. Slapeta, P. L´opez-Garc´ıa, and D. Moreira, “Global dispersal and ancient cryptic species in the smallest marine eukaryotes.” Molecular biology and evolution, vol. 23, no. 1, pp. 23–9, Jan. 2006.

[108] M. Cottrell, “Dynamics of a lytic virus infecting the photosynthetic marine picoﬂagellate Micromonas pusilla,” Limnology and oceanography, vol. 40, no. 4, pp. 730–739, 1995.

[109] E. Sahlsten, “Seasonal abundance in Skagerrak-Kattegat coastal waters and host speciﬁcity of viruses infecting the marine photosynthetic ﬂagellate Mi- cromonas pusilla,” Aquatic Microbial Ecology, vol. 16, pp. 103–108, 1998.

[110] C. Courties, A. Vaquer, M. Troussellier, J. Lautier, M. Chr´etiennot-Dinet, J. Neveux, C. Machado, and H. Claustre, “Smallest eukaryotic organism,” Na- ture, vol. 370, p. 255, 1994.

[111] P. D. Countway and D. A. Caron, “Abundance and Distribution of Ostreococcus sp . in the San Pedro Channel , California , as Revealed by Quantitative PCR,” Applied and environmental microbiology, vol. 72, no. 4, pp. 2496–2506, 2006.

[112] L. Campbell, L. Shapiro, and E. Haugen, “Immunochemical characterization for eukaryotic ultraplankton from the Atlantic and Paciﬁc oceans,” Journal of Plankton Research, vol. 16, no. 1, pp. 35–51, 1994.

[113] M. Viprey, L. Guillou, M. Ferr´eol,and D. Vaulot, “Wide genetic diversity of picoplanktonic green algae (Chloroplastida) in the Mediterranean Sea uncovered by a phylum-biased PCR approach.” Environmental microbiology, vol. 10, no. 7, pp. 1804–22, Jul. 2008.

[114] B. Eddie, A. Juhl, C. Krembs, C. Baysinger, and S. Neuer, “Eﬀect of environmental variables on eukaryotic microbial community structure of land-fast Arctic sea ice.” Environmental microbiology, vol. 12, no. 3, pp. 797–809, Mar. 2010. 109

[115] K. W. Estep, P. G. Davis, P. E. Hargraves, and J. M. Sieburth, “Chloroplast containing microflagellates in natural populations of North Atlantic nanoplankton, their identification and distribution; including a description of five new species of Chrysochromulina (Prymnesiophyceae),” Protistologica, vol. 20, no. 4, pp. 613–634, 1984.

[116] “The phagotrophic origin of eukaryotes and phylogenetic classiﬁcation of Proto- zoa.” International journal of systematic and evolutionary microbiology, vol. 52, no. Pt 2, pp. 297–354, Mar. 2002.

[117] S. I. Nikolaev, C. Berney, J. F. Fahrni, I. Bolivar, S. Polet, A. P. Mylnikov, V. V. Aleshin, N. B. Petrov, and J. Pawlowski, “The twilight of Heliozoa and rise of Rhizaria, an emerging supergroup of amoeboid eukaryotes.” Proceedings of the National Academy of Sciences of the United States of America, vol. 101, no. 21, pp. 8066–71, May 2004.

[118] B. K. S. Gupta, Modern Foraminifera, B. K. S. Gupta, Ed. Cornwall: Springer, 2002.

[119] A. Habura, J. Pawlowski, S. D. Hanes, and S. S. Bowser, “Unexpected foraminiferal diversity revealed by small-subunit rDNA analysis of Antarctic sediment.” The Journal of eukaryotic microbiology, vol. 51, no. 2, pp. 173–9, 2004.

[120] Y. Kunitomo, I. Sarashina, M. Iijima, K. Endo, and K. Sashida, “Molecular phylogeny of acantharian and polycystine radiolarians based on ribosomal DNA sequences, and some comparisons with data from the fossil record.” European journal of protistology, vol. 42, no. 2, pp. 143–53, Jul. 2006.

[121] A. K. Krabberø d, J. Br˚ate, J. K. Dolven, R. F. Ose, D. Klaveness, T. Kris- tensen, K. R. Bjø rklund, and K. Shalchian-Tabrizi, “Radiolaria Divided into Polycystina and Spasmaria in Combined 18S and 28S rDNA Phylogeny.” PloS one, vol. 6, no. 8, p. e23526, Jan. 2011.

[122] J. Pawlowski, R. Christen, B. Lecroq, D. Bachar, H. R. Shahbazkia, L. Amaral- Zettler, and L. Guillou, “Eukaryotic richness in the abyss: insights from pyrotag sequencing.” PloS one, vol. 6, no. 4, p. e18169, Jan. 2011.

[123] D. Marie, X. L. Shi, F. Rigaut-Jalabert, and D. Vaulot, “Use of ﬂow cytometric sorting to better assess the diversity of small photosynthetic eukaryotes in the English Channel.” FEMS microbiology ecology, vol. 72, no. 2, pp. 165–78, May 2010. 110

[124] A. R. Kirkham, L. E. Jardillier, A. Tiganescu, J. Pearman, M. V. Zubkov, and D. J. Scanlan, “Basin-scale distribution patterns of photosynthetic picoeukaryotes along an Atlantic Meridional Transect.” Environmental microbiology, vol. 13, no. 4, pp. 975–90, Apr. 2011.

[125] S. von der Heyden, E. E. Chao, K. Vickerman, and T. Cavalier-Smith, “Riboso- mal RNA phylogeny of bodonid and diplonemid ﬂagellates and the evolution of euglenozoa.” The Journal of eukaryotic microbiology, vol. 51, no. 4, pp. 402–16, 2004.

[126] E. Lara, D. Moreira, A. Vereshchaka, and P. L´opez-Garc´ıa,“Pan-oceanic distribution of new highly diverse clades of deep-sea diplonemids.” Environmental microbiology, vol. 11, no. 1, pp. 47–55, Jan. 2009.

[127] A. Stechmann and T. Cavalier-Smith, “Rooting the eukaryote tree by using a derived gene fusion.” Science (New York, N.Y.), vol. 297, no. 5578, pp. 89–91, Jul. 2002.

[128] I. B. Rogozin, M. K. Basu, M. Cs¨ur¨os,and E. V. Koonin, “Analysis of rare genomic changes does not support the unikont-bikont phylogeny and suggests cyanobacterial symbiosis as the point of primary radiation of eukaryotes.” Genome biology and evolution, vol. 1, pp. 99–113, Jan. 2009.

[129] J. Pawlowski and F. Burki, “Untangling the phylogeny of amoeboid protists.” The Journal of eukaryotic microbiology, vol. 56, no. 1, pp. 16–25, 2009.

[130] C. Lovejoy, R. Massana, and C. Pedr´os-Ali´o,“Diversity and distribution of marine microbial eukaryotes in the Arctic Ocean and adjacent seas.” Applied and environmental microbiology, vol. 72, no. 5, pp. 3085–95, May 2006.

[131] V. P. Edgcomb, D. Beaudoin, R. Gast, J. F. Biddle, and A. Teske, “Marine subsurface eukaryotes: the fungal majority.” Environmental microbiology, vol. 13, no. 1, pp. 172–83, Jan. 2011.

[132] S. Nandini and S. Sarma, “Population growth and body size in ﬁve rotifer species in response to variable food concentration,” Journal of Freshwater Ecol- ogy, vol. 22, no. 1, pp. 1–10, 2007.

[133] P. D. Countway, R. J. Gast, P. Savai, and D. a. Caron, “Protistan diversity estimates based on 18S rDNA from seawater incubations in the Western North Atlantic.” The Journal of eukaryotic microbiology, vol. 52, no. 2, pp. 95–106, 2005. 111

[134] F. Not, K. Valentin, K. Romari, C. Lovejoy, R. Massana, K. T¨obe, D. Vaulot, and L. K. Medlin, “Picobiliphytes: a marine picoplanktonic algal group with unknown aﬃnities to other eukaryotes.” Science (New York, N.Y.), vol. 315, no. 5809, pp. 253–5, Jan. 2007.

[135] D. Caron and P. Countway, “Hypotheses on the role of the protistan rare biosphere in a changing world,” Aquatic Microbial Ecology, vol. 57, no. December, pp. 227–238, Nov. 2009.

[136] M. L. Sogin, H. G. Morrison, J. a. Huber, D. Mark Welch, S. M. Huse, P. R. Neal, J. M. Arrieta, and G. J. Herndl, “Microbial diversity in the deep sea and the underexplored ”rare biosphere”.” Proceedings of the National Academy of Sciences of the United States of America, vol. 103, no. 32, pp. 12 115–20, Aug. 2006.

[137] M. Pester, N. Bittner, P. Deevong, M. Wagner, and A. Loy, “A ’rare biosphere’ microorganism contributes to sulfate reduction in a peatland.” The ISME journal, vol. 4, no. 12, pp. 1591–602, Dec. 2010.

[138] J. Sjöstedt, P. Koch-Schmidt, M. Pontarp, B. Canbäck, A. Tunlid, P. Lund- berg, A. Hagström,and L. Riemann, “Recruitment of members from the rare biosphere of marine bacterioplankton communities after an environmental dis- turbance.” Applied and environmental microbiology, vol. 78, no. 5, pp. 1361–9, Mar. 2012.

[139] O. Hoegh-Guldberg and J. F. Bruno, “The Impact of Climate Change on the World’s Marine Ecosystems,” Science, vol. 328, no. 5985, pp. 1523–1528, Jun. 2010.

[140] N. G. Meakin and M. Wyman, “Rapid shifts in picoeukaryote community structure in response to ocean acidiﬁcation.” The ISME journal, vol. 5, no. 9, pp. 1397–1405, Mar. 2011.

[141] F. J. Edwards, “Climate and Oceanography,” in Red Sea, 1st ed., A. J. Edwards and S. Head, Eds. New York: Pergamon Press, 1987, ch. 3, pp. 45–69.

[142] A. R. Longhurst, Ecological Geography of the Sea, 2nd ed. Boston: Academic Press, 2007.

[143] S. P. Murray and W. Johns, “Direct observations of seasonal exchange through the Bab el Mandab Strait,” Geophysical Research Letters, vol. 24, no. 21, p. 2557, 1997. 112

[144] M. I. Badran, “Dissolved Oxygen, Chlorophyll a and Nutrients: Seasonal Cy- cles in Waters of the Gulf of Aquaba, Red Sea,” Aquatic Ecosystem Health & Management, vol. 4, no. 2, pp. 139–150, Jun. 2001.

[145] S. M. Head, “Introduction,” in Red Sea, 1st ed., A. J. Edwards and S. M. Head, Eds. New York: Pergamon Press, 1987, ch. 1, pp. 1–21.

[146] H. Weiker, “Plankton and the Pelagic Environment,” in Red Sea, 1st ed., A. J. Edwards and S. M. Head, Eds. New York: Pergamon Press, 1987, ch. 5, pp. 90–111.

[147] L. Karbe and J. Lange, “The chemical environment,” in Mining of Metallifer- ous Sediments from the Atlantis II Deep, Red Sea: Pre-mining Environmental Conditions and Evaluation of the Risk to the Environment., L. Karbe, H. Thiel, H. Weikert, and A. Mill, Eds., Hamburg, 1981, pp. 75–99.

[148] S. Levitus, M. E. Conkright, J. L. Reid, R. G. Najjar, and A. Mantyla, “Dis- tribution of nitrate, phosphate and silicate in the world oceans,” Progress In Oceanography, vol. 31, no. 3, pp. 245–273, Jan. 1993.

[149] C. J. Braithwaite, “Geology and Palaeogeography of the Red Sea Region,” in Red Sea, 1st ed., A. J. Edwards and S. M. Head, Eds. New York: Pergamon Press, 1987, ch. 2, pp. 22–44.

[150] A. Getahun, “The Red Sea as an Extension of the Indian Ocean,” in Large Marine Ecosystems of the Indian Ocean: Assessment, Sustainability, and Man- agement, 1st ed., E. Okemwa, M. J. Ntiba, and K. Sherman, Eds. Oxford: Blackwell Science, 1995, pp. 277–283.

[151] M. M. A. Kotb, M. H. Hanafy, H. Rirache, S. Matsumura, A. Al-Sofyani, A. g. Ahmed, F. A. Al-Horan, and G. Bawazir, “Status of Coral Reefs in the Red Sea and Gulf of Aden Region,” in Status of Coral Reefs of the World: 2008, C. Wilkinson, Ed. Townsville: Global Coral Reef Monitoring Network and Reef and Rainforest Research Center, 1998, ch. 4, pp. 67–78.

[152] D. K. Ngugi, A. Antunes, A. Brune, and U. Stingl, “Biogeography of pelagic bacterioplankton across an antagonistic temperature-salinity gradient in the Red Sea.” Molecular ecology, vol. 21, no. 2, pp. 388–405, Jan. 2012.

[153] H. Kvitt and a. Colorni, “Strain variation and geographic endemism in Strep- tococcus iniae,” Diseases of Aquatic Organisms, vol. 61, pp. 67–73, 2004. 113

[154] H. X. Wu, P. Y. Lai, O. O. Lee, X. J. Zhou, L. Miao, H. Wang, and P. Y. Qian, “Erythrobacter pelagi sp. nov., a novel bacterium in the family Erythrobac- teraceae isolated from the Red Sea,” International Journal of Systematic and Evolutionary Microbiology, no. 852, Aug. 2011.

[155] G. Trommer, M. Siccha, M. T. van der Meer, S. Schouten, J. S. Sinninghe Damst´e,H. Schulz, C. Hemleben, and M. Kucera, “Distribution of Crenar- chaeota tetraether membrane lipids in surface sediments from the Red Sea,” Organic Geochemistry, vol. 40, no. 6, pp. 724–731, Jun. 2009.

[156] E. V. Belagorskaya, “Qualitative and Quantitative Distribution of Phytoplank- ton in the Red Sea and Gulf of Aden in October-November 1963.” Biologiya morja, vol. 21, pp. 133–152, 1970.

[157] B. Kimor, “Microplankton of the Gulf of Elat: aspects of seasonal and bathy- metric distribution,” Marine Biology, vol. 42, no. 1, pp. 55–67, 1977.

[158] R. Medinger, V. Nolte, R. V. Pandey, S. Jost, B. Ottenw¨alder,C. Schl¨otterer, and J. Boenigk, “Diversity in a hidden world: potential and limitation of next- generation sequencing for surveys of molecular diversity of eukaryotic microorganisms.” Molecular ecology, vol. 19 Suppl 1, pp. 32–40, Mar. 2010.

[159] L. Medlin, H. J. Elwood, S. Stickel, and M. L. Sogin, “The characterization of enzymatically ampliﬁed eukaryotic 16S-like rRNA-coding regions.” Gene, vol. 71, no. 2, pp. 491–9, Nov. 1988.

[160] K. Kolodziej and T. Stoeck, “Cellular identiﬁcation of a novel uncultured marine stramenopile (MAST-12 Clade) small-subunit rRNA gene sequence from a norwegian estuary by use of ﬂuorescence in situ hybridization-scanning electron microscopy.” Applied and environmental microbiology, vol. 73, no. 8, pp. 2718–26, Apr. 2007.

[161] W. Luo, H. Li, M. Cai, and J. He, “Diversity of microbial eukaryotes in Kongs- fjorden, Svalbard,” Hydrobiologia, vol. 636, no. 1, pp. 233–248, Oct. 2009.

[162] T. Huber, G. Faulkner, and P. Hugenholtz, “Bellerophon: a program to detect chimeric sequences in multiple sequence alignments.” Bioinformatics (Oxford, England), vol. 20, no. 14, pp. 2317–9, Sep. 2004.

[163] R. C. Edgar, B. J. Haas, J. C. Clemente, C. Quince, and R. Knight, “UCHIME improves sensitivity and speed of chimera detection.” Bioinformatics (Oxford, England), vol. 27, no. 16, pp. 2194–2200, Jun. 2011. 114

[164] P. D. Schloss, S. L. Westcott, T. Ryabin, J. R. Hall, M. Hartmann, E. B. Hollister, R. a. Lesniewski, B. B. Oakley, D. H. Parks, C. J. Robinson, J. W. Sahl, B. Stres, G. G. Thallinger, D. J. Van Horn, and C. F. Weber, “Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities.” Applied and environmental microbiology, vol. 75, no. 23, pp. 7537–41, Dec. 2009.

[165] S. F. Altschul, T. L. Madden, a. a. Sch¨aﬀer,J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman, “Gapped BLAST and PSI-BLAST: a new generation of protein database search programs.” Nucleic acids research, vol. 25, no. 17, pp. 3389–402, Sep. 1997.

[166] I. Good, “The population frequencies of species and the estimation of population parameters,” Biometrika, vol. 40, no. 3-4, pp. 237–264, 1953.

[167] A. Chao, “Nonparametric estimation of the number of classes in a population,” Scandinavian Journal of Statistics, pp. 265–270, 1984.

[168] A. Chao and S. Lee, “Estimating the number of classes via sample coverage,” Journal of the American Statistical Association, pp. 210–217, 1992.

[169] A. Chao and M. C. K. Yang, “Stopping Rules and Estimation for Recapture Debugging with Unequal Failure Rates,” Biometrika, vol. 80, no. 1, p. 193, Mar. 1993.

[170] J. Bunge, “Estimating the number of species with catchall.” Paciﬁc Symposium on Biocomputing. Paciﬁc Symposium on Biocomputing, pp. 121–30, Jan. 2011.

[171] W. Ludwig, “ARB manual,” Nucleic Acids Research, vol. 32, no. 4, pp. 1363–71, 2002.

[172] E. Pruesse, C. Quast, K. Knittel, B. M. Fuchs, W. Ludwig, J. Peplies, and F. O. Gl¨ockner, “SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB.” Nucleic acids research, vol. 35, no. 21, pp. 7188–96, Jan. 2007.

[173] A. Drummond, B. Ashton, S. Buxton, M. Cheung, A. Cooper, C. Duran, M. Field, J. Heled, M. Kearse, S. Markowitz, R. Moir, S. Stones-Havas, S. Sturrock, T. T, and W. A, “Geneious v5.5,” 2012. [Online]. Available: www.geneious.com 115

[174] S. Guindon and O. Gascuel, “A Simple, Fast, and Accurate Algorithm to Esti- mate Large Phylogenies by Maximum Likelihood,” Systematic Biology, vol. 52, no. 5, pp. 696–704, Oct. 2003.

[175] N. Saitou, “The neighbor-joining method: a new method for reconstructing phylogenetic trees.” Molecular biology and evolution, vol. 4, no. 4, pp. 406–425, 1987.

[176] S. C. Dawson and K. D. Hagen, “Mapping the protistan ’rare biosphere’.” Journal of biology, vol. 8, no. 12, p. 105, Jan. 2009.

[177] D. a. Caron, P. D. Countway, P. Savai, R. J. Gast, A. Schnetzer, S. D. Moor- thi, M. R. Dennett, D. M. Moran, and A. C. Jones, “Deﬁning DNA-based operational taxonomic units for microbial-eukaryote ecology.” Applied and environmental microbiology, vol. 75, no. 18, pp. 5797–808, Sep. 2009.

[178] S. Epstein and P. L´opez-Garc´ıa,“Missing protists: a molecular prospective,” Biodiversity and Conservation, vol. 17, no. 2, pp. 261–276, Oct. 2007.

[179] J. Pawlowski and B. Lecroq, “Short rDNA barcodes for species identiﬁcation in foraminifera.” The Journal of eukaryotic microbiology, vol. 57, no. 2, pp. 197–205, 2010.

[180] M. Nebel, C. Pfabel, A. Stock, M. Dunthorn, and T. Stoeck, “Delimiting operational taxonomic units for assessing ciliate environmental diversity using small- subunit rRNA gene sequences,” Environmental Microbiology Reports, vol. 3, no. 2, pp. 154–158, Apr. 2011.

[181] H. F. Santos, J. C. Cury, F. L. Carmo, A. S. Rosado, and R. S. Peixoto, “18S rDNA sequences from microeukaryotes reveal oil indicators in mangrove sediment.” PloS one, vol. 5, no. 8, p. e12437, Jan. 2010.

[182] R. Cronn, M. Cedroni, T. Haselkorn, C. Grover, and J. F. Wendel, “PCR- mediated recombination in ampliﬁcation products derived from polyploid cot- ton.” TAG. Theoretical and applied genetics. Theoretische und angewandte Genetik, vol. 104, no. 2-3, pp. 482–489, Feb. 2002.

[183] S. M. Huse, L. Dethlefsen, J. a. Huber, D. Mark Welch, D. M. Welch, D. a. Relman, and M. L. Sogin, “Exploring microbial diversity and taxonomy using SSU rRNA hypervariable tag sequencing.” PLoS genetics, vol. 4, no. 11, p. e1000255, Nov. 2008. 116

[184] R. R. Guillard and P. Kilham, “The Ecology of Marine Planktonic Diatoms,” in The Biology of Diatoms, W. Dietrich, Ed. Los Angeles, California: University of California Press, 1977, ch. 12, pp. 372–470.

[185] M. Dunthorn, J. Klier, J. Bunge, and T. Stoeck, “Comparing the Hyper- Variable V4 and V9 Regions of the Small Subunit rDNA for Assessment of Cil- iate Environmental Diversity.” The Journal of eukaryotic microbiology, vol. 0, no. 0, pp. 1–3, Jan. 2012.

[186] M. C. Savin, J. L. Martin, M. LeGresley, M. Giewat, and J. Rooney-Varga, “Plankton diversity in the Bay of Fundy as measured by morphological and molecular methods.” Microbial ecology, vol. 48, no. 1, pp. 51–65, Jul. 2004.

[187] V. Farrelly, F. a. Rainey, and E. Stackebrandt, “Eﬀect of genome size and rrn gene copy number on PCR ampliﬁcation of 16S rRNA genes from a mixture of bacterial species.” Applied and environmental microbiology, vol. 61, no. 7, pp. 2798–801, Jul. 1995.

[188] T. R. Gregory, “Coincidence, coevolution, or causation? DNA content, cell size, and the C-value enigma.” Biological reviews of the Cambridge Philosophical Society, vol. 76, no. 1, pp. 65–101, Feb. 2001.

[189] C. D. Prokopowich, T. R. Gregory, and T. J. Crease, “The correlation between rDNA copy number and genome size in eukaryotes.” Genome / National Re- search Council Canada = G´enome/ Conseil national de recherches Canada, vol. 46, no. 1, pp. 48–50, Feb. 2003.

[190] S. Murray, M. r. FløJørgensen, S. Y. W. Ho, D. J. Patterson, and L. S. Jermiin, “Improving the analysis of dinoﬂagellate phylogeny based on rDNA.” Protist, vol. 156, no. 3, pp. 269–86, Nov. 2005.

[191] M. D. a. Howard, G. J. Smith, and R. M. Kudela, “Phylogenetic relationships of yessotoxin-producing dinoﬂagellates, based on the large subunit and internal transcribed spacer ribosomal DNA domains.” Applied and environmental microbiology, vol. 75, no. 1, pp. 54–63, Jan. 2009.

[192] K. A. Steidinger and K. Tangen, “Dinoﬂagellates,” in Identifying Marine Phy- toplankton, C. Tomas, Ed. Florida: Academic Press, 1997, ch. 3, pp. 387–584.

[193] R. J. Gast, D. M. Moran, D. J. Beaudoin, J. N. Blythe, M. R. Dennett, and D. a. Caron, “Abundance of a Novel Dinoﬂagellate Phylotype in the Ross Sea, Antarctica1,” Journal of Phycology, vol. 42, no. 1, pp. 233–242, Feb. 2006. 117

[194] Y. Z. Tang, T. a. Egerton, L. Kong, and H. G. Marshall, “Morphological variation and phylogenetic analysis of the dinoﬂagellate Gymnodinium aureolum from a tributary of Chesapeake Bay.” The Journal of eukaryotic microbiology, vol. 55, no. 2, pp. 91–9, 2008.

[195] D. U. Hernández-Becerril,R. Alonso-Rodr´ıguez, C. Alvarez-Góngora,S. A. Barón-Campis, G. Ceballos-Corona, J. Herrera-Silveira, M. E. Meave Del Castillo, N. Juárez-Ru´ız, F. Merino-Virgilio, A. Morales-Blake, J. L. Ochoa, E. Orellana-Cepeda, C. Ram´ırez-Camarena,and R. Rodr´ıguez-Salvador, “Toxic and harmful marine phytoplankton and microalgae (HABs) in Mexican Coasts.” Journal of environmental science and health. Part A, Toxic/hazardous sub- stances & environmental engineering, vol. 42, no. 10, pp. 1349–63, Aug. 2007.

[196] C. J. Band-Schmidt, J. J. Bustillos-Guzmán,D. J. López-Cortés,I. Gárate- Lizárraga,E. J. NúñezVázquez,and F. E. Hernández-Sandoval, “Ecological and physiological studies of Gymnodinium catenatum in the Mexican Pacific: a review.” Marine drugs, vol. 8, no. 6, pp. 1935–61, Jan. 2010.

[197] J. H. Gunderson, S. H. Goss, and D. W. Coats, “The Phylogenetic Position of Amoebophrya sp. Infecting Gymnodinium sanguineum,” The Journal of Eu- karyotic Microbiology, vol. 46, no. 2, pp. 194–197, Mar. 1999.

[198] S. Lin, H. Zhang, Y. Hou, L. Miranda, and D. Bhattacharya, “Development of a dinoﬂagellate-oriented PCR primer set leads to detection of picoplanktonic dinoﬂagellates from Long Island Sound.” Applied and environmental microbiology, vol. 72, no. 8, pp. 5626–30, Aug. 2006.

[199] W. Liu, Z. Yi, X. Lin, and K. a. S. Al-Rasheid, “Morphologic and molecular data suggest that Lynnella semiglobulosa n. g., n. sp. represents a new family within the subclass Choreotrichia (Ciliophora, Spirotrichea).” The Journal of eukaryotic microbiology, vol. 58, no. 1, pp. 43–9, 2011.

[200] B. S. Leander, S. a. J. Lloyd, W. Marshall, and S. C. Landers, “Phylogeny of marine Gregarines (Apicomplexa)–Pterospora, Lithocystis and Lankesteria– and the origin(s) of coelomic parasitism.” Protist, vol. 157, no. 1, pp. 45–60, Feb. 2006.

[201] X. L. Shi, D. Marie, L. Jardillier, D. J. Scanlan, and D. Vaulot, “Groups without cultured representatives dominate eukaryotic picophytoplankton in the oligotrophic South East Paciﬁc Ocean.” PloS one, vol. 4, no. 10, p. e7657, Jan. 2009. 118

[202] C. Lovejoy, P. E. Galand, and D. L. Kirchman, “Picoplankton diversity in the Arctic Ocean and surrounding seas,” Marine Biodiversity, vol. 41, no. 1, pp. 5–12, Oct. 2010.

[203] G. Rocap, F. W. Larimer, J. Lamerdin, S. Malfatti, P. Chain, N. a. Ahlgren, A. Arellano, M. Coleman, L. Hauser, W. R. Hess, Z. I. Johnson, M. Land, D. Lindell, A. F. Post, W. Regala, M. Shah, S. L. Shaw, C. Steglich, M. B. Sul- livan, C. S. Ting, A. Tolonen, E. a. Webb, E. R. Zinser, and S. W. Chisholm, “Genome divergence in two Prochlorococcus ecotypes reﬂects oceanic niche dif- ferentiation.” Nature, vol. 424, no. 6952, pp. 1042–7, Aug. 2003.

[204] E. Demir-Hilton, S. Sudek, M. L. Cuvelier, C. L. Gentemann, J. P. Zehr, and A. Z. Worden, “Global distribution patterns of distinct clades of the photosynthetic picoeukaryote Ostreococcus.” The ISME journal, pp. 1–13, Feb. 2011.

[205] R. A. Lewin, L. Krienitz, R. Goericke, H. Takeda, and D. Hepperle, “Picocystis salinarum gen. et sp. nov. (Chlorophyta) a new picoplanktonic green alga,” Phycologia, vol. 39, no. 6, pp. 560–565, Nov. 2000.

[206] T. Cavalier-Smith and E. E.-Y. Chao, “Phylogeny and megasystematics of phagotrophic heterokonts (kingdom Chromista).” Journal of molecular evolution, vol. 62, no. 4, pp. 388–420, Apr. 2006.

[207] R. Massana, L. Guillou, R. Terrado, I. Forn, and C. Pedrós-Alió,“Growth of uncultured heterotrophic flagellates in unamended seawater incubations,” Aquatic microbial ecology, vol. 45, no. 2, pp. 171–180, 2006.

[208] W. Orsi, Y. C. Song, S. Hallam, and V. Edgcomb, “Eﬀect of oxygen minimum zone formation on communities of marine protists.” The ISME journal, pp. 1–16, Mar. 2012.

[209] N. Fuller, C. Campbell, D. Allen, F. Pitt, K. Zwirglmaier, F. Le Gall, D. Vaulot, and D. Scanlan, “Analysis of photosynthetic picoeukaryote diversity at open ocean sites in the Arabian Sea using a PCR biased towards marine algal plas- tids,” Aquatic Microbial Ecology, vol. 43, pp. 79–93, May 2006.

[210] M. V. Zubkov and G. a. Tarran, “High bacterivory by the smallest phytoplankton in the North Atlantic Ocean.” Nature, vol. 455, no. 7210, pp. 224–6, Sep. 2008.

[211] D. A. Holen and M. E. Boraas, “Mixotrophy in chrysophytes,” in Chrysophyte algae: Ecology, phylogeny and development, 1st ed., C. D. Sandgren, J. P. Smol, and J. Kristiansen, Eds. Cambridge: Cambridge University Press, 1995, ch. 6, pp. 119–140. 119

[212] L. Guillou, M. J. Chr´etiennot-Dinet,L. K. Medlin, H. Claustre, S. L. Go¨er,and D. Vaulot, “Bolidomonas: a new genus with two species belonging to a new algal class, the Bolidophyceae (Heterokonta),” Journal of phycology, vol. 35, no. 2, pp. 368–381, 1999.

[213] L. Guillou, S. Y. Moon-Van Der Staay, H. Claustre, F. Partensky, and D. Vaulot, “Diversity and abundance of Bolidophyceae (Heterokonta) in two oceanic regions.” Applied and environmental microbiology, vol. 65, no. 10, pp. 4528–36, Oct. 1999.

[214] J. Frias-Lopez, A. Thompson, J. Waldbauer, and S. W. Chisholm, “Use of stable isotope-labelled cells to identify active grazers of picocyanobacteria in ocean surface waters.” Environmental microbiology, vol. 11, no. 2, pp. 512–25, Feb. 2009.

[215] M. Kawachi, I. Inouye, D. Honda, C. O’Kelly, J. Bailey, R. Bidigare, and R. Andersen, “The Pinguiophyceae classis nova, a new class of photosynthetic stramenopiles whose members produce large amounts of omega-3 fatty acids,” Phycological research, vol. 50, no. 1, pp. 31–47, 2002.

[216] M. Kawachi, M. Atsumi, H. Ikemoto, and S. Miyachi, “Pinguiochrysis pyriformis gen. et sp. nov.(Pinguiophyceae), a new picoplanktonic alga isolated from the Paciﬁc Ocean,” Phycological, vol. 50, no. November 2001, pp. 49–56, 2002.

[217] R. Andersen and D. Pottert, “Pinguiococcus pyrenoidosus gen. et sp. nov.(Pinguiophyceae), a new marine coccoid alga,” Phycological Research, no. November 2001, pp. 57–65, 2002.

[218] L. Guillou, M. J. Chrétiennot-Dinet,S. Boulben, S. Y. Moon-van der Staay, and D. Vaulot, “Symbiomonas scintillans gen. et sp. nov. and Picophagus flagellatus gen. et sp. nov. (Heterokonta): two new heterotrophic flagellates of picoplanktonic size.” Protist, vol. 150, no. 4, pp. 383–98, Dec. 1999.

[219] S. I. of Oceanography, “Shore Stations Program,” 2011. [Online]. Available: http://shorestation.ucsd.edu/index.html

[220] S. C. Dawson and N. R. Pace, “Novel kingdom-level eukaryotic diversity in anoxic environments.” Proceedings of the National Academy of Sciences of the United States of America, vol. 99, no. 12, pp. 8324–9, Jun. 2002.

[221] Y. Ishitani, Y. Ujii´e,C. de Vargas, F. Not, and K. Takahashi, “Two distinct lineages in the radiolarian Order Spumellaria having diﬀerent ecological prefer- 120

ences,” Deep Sea Research Part II: Topical Studies in Oceanography, vol. 61-64, pp. 172–178, Feb. 2012.

[222] H. Yamashita and K. Takahashi, “Zonal and vertical distribution of radiolarians in the western and central Equatorial Paciﬁc in January 1999,” Research Part II: Topical Studies in, vol. 49, no. January 1999, pp. 2823–2862, 2002.

[223] D. Hibberd, “Cytology and Ultraestructure of Chloraracnion reptans (Chloroarachniophyta diviso nova, Chlorarachniophyceae classis nova),” Jour- nal of Phycology, vol. 20, pp. 310–330, 1984.

[224] Y. Hirakawa, A. Howe, E. R. James, and P. J. Keeling, “Morphological diversity between culture strains of a chlorarachniophyte, Lotharella globosa.” PloS one, vol. 6, no. 8, p. e23193, Jan. 2011.

[225] O. j. Moestrup and M. Sengco, “Ultrastructural studies on Bigelowiella natans, gen. et sp. nov., A Clorarachniophyte Flagellate,” Algae, vol. 37, pp. 624–646, 2001.

[226] D. Bass and T. Cavalier-Smith, “Phylum-speciﬁc environmental DNA analysis reveals remarkably high global biodiversity of Cercozoa (Protozoa).” Interna- tional journal of systematic and evolutionary microbiology, vol. 54, no. Pt 6, pp. 2393–404, Nov. 2004.

[227] F. Not, R. Massana, M. Latasa, and D. Marie, “Late summer community composition and abundance of photosynthetic picoeukaryotes in Norwegian and Barents Seas,” Limnology and Oceanography, vol. 50, no. 5, pp. 1677–1686, 2005.

[228] S. Moon-van der Staay, G. van der Staay, L. Guillou, D. Vaulot, H. Claustre, and L. Medlin, “Abundance and diversity of prymnesiophytes in the picoplankton community from the equatorial Paciﬁc Ocean inferred from 18S rDNA sequences,” Limnology and Oceanography, vol. 45, no. 1, pp. 98–109, 2000.

[229] B. Edvardsen, W. Eikrem, J. Throndsen, A. G. S´aez,I. Probert, and L. K. Medlin, “Ribosomal DNA phylogenies and a morphological revision provide the basis for a revised taxonomy of the Prymnesiales (Haptophyta),” European Journal of Phycology, vol. 46, no. 3, pp. 202–228, Aug. 2011.

[230] H. Jones, B. Leadbeater, and J. Green, “Mixotrophy in marine species of Chrysochromulina (Prymnesiophyceae): Ingestion and digestion of a small green ﬂagellate,” Journal of the Marine Biological Association of the United Kingdom, vol. 73, no. 02, pp. 283–296, 1993. 121

[231] N. Okamoto, C. Chantangsi, A. Hor´ak,B. S. Leander, and P. J. Keeling, “Molec- ular phylogeny and description of the novel katablepharid Roombia truncata gen. et sp. nov., and establishment of the Hacrobia taxon nov.” PloS one, vol. 4, no. 9, p. e7080, Jan. 2009.

[232] M. D. Johnson, T. Tengs, D. W. Oldach, C. F. Delwiche, and D. K. Stoecker, “Highly divergent SSU rRNA genes found in the marine ciliates Myrionecta rubra and Mesodinium pulex.” Protist, vol. 155, no. 3, pp. 347–59, Sep. 2004.

[233] C. Pedr´os-Ali´o,“Ecology. Dipping into the rare biosphere.” Science (New York, N.Y.), vol. 315, no. 5809, pp. 192–3, Jan. 2007.

[234] C. Quince, A. Lanz´en,T. P. Curtis, R. J. Davenport, N. Hall, I. M. Head, L. F. Read, and W. T. Sloan, “Accurate determination of microbial diversity from 454 pyrosequencing data.” Nature methods, vol. 6, no. 9, pp. 639–41, Sep. 2009.

[235] J. Reeder and R. Knight, “The ’rare biosphere’: a reality check.” Nature methods, vol. 6, no. 9, pp. 636–7, Sep. 2009.

[236] R. Gradinger, T. Weisse, and T. Pillen, “Signiﬁcance of Picocyanobacteria in the Red Sea and the Gulf of Aden,” Botanica Marina, vol. 35, no. 3, pp. 245– 250, 1992. 122

Appendix A

Sampling site physical and environmental characteristics

Figure A.1: Geographical location of sampling site.

Coordinates: 24.96◦ N 36.35◦ E. Temperature: 25.55◦C Salinity: 39.42 PSU Dissolved oxygen: 4.24 mg/l Chlorophyll a: 0.148 mg/m3 Particular Inorganic Carbon: 0.149 mmol/m3 Particular Organic Carbon: 48.54 mg/m3 123

Appendix B

Papers Submitted and Under Preparation

• Acosta F, Ngugi DK, and Stingl U, Diversity of Picoeukaryotes at an Oligotrophic Site oﬀ the Northern Red Sea Coast, manuscript in preparation