Patrick John Keeling

Dinoflagellates are a diverse group of protists with many unique traits including large genomes packaged into permanently condensed chromosomes, photosynthetic or cryptic plastids acquired vertically or horizontally in serial endosymbioses, and in some taxa, highly complex organelles like nematocysts and the eye-like ocelloid. Because these features promise to expand our understanding of eukaryotic biology, reconstructing how they evolved has become a point of interest. To infer ancestral states, robust and well-supported phylogenies generated from high-coverage transcriptomic datasets are needed. So far, these analyses have relied on transcriptome data from cultured taxa, which are mostly photosynthetic. As half of known dinoflagellate species are non-photosynthetic, current phylogenies fail to reflect the diversity that characterizes this group. Here, I generate single cell transcriptomes from over 150 rare and under-sampled dinoflagellates collected from the environment. Using these data, I explore three major heterotrophic lineages of interest: Abedinium, the Noctilucales, and the complex organelle-bearing members of the Gymnodiniales, the warnowiids. In these investigations I reveal that Abedinium is an independent, deep-branching core dinoflagellate lineage, the unique traits of the Noctilucales are derived rather than ancestral, and the heterotrophic warnowiids retain photosynthetic genes except for photosystem II and RuBisCo, suggesting that this mechanism serves an alternate function in the ocelloid. In the final chapter I generate a comprehensive dinoflagellate phylogeny that better represents the proportion of heterotrophic, athecate, and deep- branching taxa in dinoflagellates. This analysis reveals several new insights, including the early acquisition timing of two histone-like protein (HLP) types, the diversity and punctate distribution of microbial rhodopsins, and the common retention of plastid-derived electron transport genes across heterotrophic dinoflagellates.

View record

Genome evolution of bacterial symbionts has led to different evolutionary outcomes including stable symbiosis, extinction, and rarely organelle formation (e.g., mitochondria and plastids). Symbiont genome evolution has mostly been studied in animal endosymbionts, but microbial eukaryotes (protists) contain diverse symbionts with various functional roles and evolutionary trajectories. However, symbiont genome reduction and functional relationships with protist hosts are largely unexplored. Additionally, how bacteriophage (phage) infection impacts symbiont-eukaryotic interactions and symbiont genome evolution is relatively unknown in protists. Here, I used DNA and RNA sequencing and microscopy methods to characterize endosymbionts and symbiont-infecting phages from diverse protists, including marine diplonemids, freshwater cryptomonads, and termite gut parabasalids. The diplonemid and cryptomonad endosymbionts belonged to intracellular bacterial groups Rickettsiales and Holosporaceae (Alphaproteobacteria), and they harboured an arsenal of proteins with putative eukaryotic-interacting domains (e.g., leucine-rich repeats and ankyrin repeats). The diplonemid endosymbionts had extremely small genomes with reduced metabolic potential compared to other Rickettsiales and Holosporaceae endosymbionts. Despite this, they retained a small cluster of gene transfer agent (GTA) genes also highly conserved in Rickettsiales and Holosporaceae, and evidence of GTA gene expression was found. An endosymbiont-infecting phage was also characterized in an undescribed Cryptomonas species, and the temperate phage encoded several eukaryotic-interacting proteins. Finally, phages from the hindgut of wood-feeding insects were sequenced and predicted to infect protist endosymbionts and hindgut bacteria. Overall, this work highlights the complex interactions between phages, bacteria and protists from diverse environments and provides valuable insights into the evolution of all three groups.

View record

Apicomplexans are a large phylum of obligate animal parasites that contain pathogens such as Plasmodium spp. (the causative agent of malaria) and Toxoplasma gondii. While these medically relevant apicomplexans are the subject of extensive research, the bulk of the diversity of the group, particularly the lineages that infect invertebrates, remain poorly studied and largely ignored in high-throughput sequencing surveys. In this dissertation, I show that these groups are critical to gaining insights into the origins and evolution of the Apicomplexa. I begin by examining the diversity and inferred ecology of the enigmatic apicomplexan-related lineages (ARLs), and show that ARL-V is highly abundant in environmental surveys, and is tightly associated with coral tissue and mucus, suggesting that it represents a core symbiont of coral. In the following chapters, using methods of single-cell transcriptomics, I sequenced the transcriptomes of 15 invertebrate-infecting apicomplexans. Using this dataset, I constructed a robust and taxon-rich multi-gene apicomplexan phylogeny that resolves the deep phylogenetic relationships within the group, and also form a new class of apicomplexans, the Marosporida, that is sister to the Hematozoa and Coccidia. Most unexpectedly, in Chapter 2, I show that certain taxa previously classified as apicomplexans, actually represent convergently evolved animal parasites, suggesting that apicomplexan-like parasites have evolved at least four times independently. In Chapter 3, I examine the presence and function of apicoplasts (remnant plastids) across the diversity of the group using whole genome shotgun sequencing (WGS), and find that the Marosporida contain the smallest, most AT-rich, and gene poor apicoplast genomes sequenced to date. I also present the first evidence of plastids in the gregarines, and show that archigregarines retain the canonical apicomplexan plastid metabolism, whereas only one clade of marine eugregarines retains plastids that solely carry out type II fatty acid biosynthesis. Lastly in Chapter 4, I reconstruct the mitochondrial metabolism in the gregarines and squirmids, and find that eugregarines contain highly reduced respiratory chains, suggesting that they have lost their mitochondrial genomes, and possess limited energy metabolism. Altogether, the data presented here, illustrates the significance of invertebrate-infecting apicomplexans in illuminating the early evolution of the apicomplexans and myzozoans.

View record

The nucleus is a defining characteristic of eukaryotic cells which not only houses the genome but a myriad of processes that function synergistically to regulate cellular activity. Nuclear proteins are key in facilitating core eukaryotic processes such as genome compaction, nucleocytoplasmic exchange, and DNA replication, but the interconnectedness of these processes makes them challenging to dissect mechanistically. Moreover, the antiquity of the nucleus complicates evolutionary analyses, limiting our view of nuclear evolution. Despite this, a comprehensive understanding of the function and evolution of nuclear processes is essential given their central importance in disease, basic cell biology, and eukaryotic evolution. In this dissertation, I argue that insights into nuclear biology and evolution can be obtained by examining eukaryotic diversity rather than relying solely on traditional model organisms. I begin by presenting an introduction to nuclear evolution and diversity, highlighting the existence of nuclear variation across eukaryotes from a systems perspective and underscoring the potential utility of biodiversity in studying nuclear processes (Chapter 1). In the following chapters, I test my hypothesis by examining the function and evolution of different processes in a subset of divergent nuclear systems: namely, chromatin in the dinoflagellate dinokaryon, nuclear pore complexes (NPCs) in the nucleomorphs of chlorarachniophytes and cryptophytes, and DNA replication in the ciliate macronucleus. In Chapter 2, I use an experimental evolutionary approach to investigate the drivers of histone depletion in dinoflagellates, revealing the capacity for viruses to shape cellular evolution and raising questions regarding the subfunctionalization of remnant dinoflagellate histones. In Chapter 3, I reconstruct the NPCs of endosymbiotically acquired nuclei, termed nucleomorphs, in silico, and predict a highly reduced pore structure, suggesting that a complex NPC may not be required for baseline nuclear function. Lastly, in Chapter 4, I examine the diversity of motile DNA replication systems in ciliates, highlighting new models for studying DNA replication and the capacity of cytoskeletal elements to coordinate nuclear organization and processes. Ultimately this work confirms the efficacy of examining diverse nuclear systems, provides insights into the biology and evolution of nuclear processes, and encourages a re-evaluation of how we view and select model organisms.

View record

Dinoflagellates are an abundant and diverse group of aquatic eukaryotes, with members that have photosynthetic, heterotrophic, or mixotrophic life strategies, as well as a number of unique cytological features. My thesis focuses on two groups of closely related dinoflagellates: polykrikoids and warnowiids. Both include heterotrophic as well as plastid-bearing members, though the number of times photosynthesis has been lost (or gained) in each group is unclear, and the presence and provenance of plastids in some species (e.g., Nematodinium sp. and Polykrikos lebouriae) have been debated. Polykrikoids and warnowiids also contain some of the most complex subcellular structures described--such as nematocysts and, in warnowiids, eye-like ocelloids. Yet these groups are rare in nature and uncultivated, and as such, the origins of their complex organelles are unclear. For my thesis, I modified existing techniques for use on single-cell environmental isolates, and applied these techniques to wild polykrikoid and warnowiid cells. By exploiting the common splice leader sequence found on dinoflagellate transcripts, I was able to amplify a single-cell transcriptome from Polykrikos lebouriae—a dinoflagellate with aberrant plastids. Coupled with single-cell genomics using multiple displacement amplification (MDA), I demonstrated that Polykrikos lebouriae has retained peridinin-type plastids, while photosynthesis has been lost in multiple other polykrikoid species independently. Using MDA and single-cell transmission electron microscopy, I also determined that the eye-like ocelloid of Nematodinium sp. is made in part from a peridinin plastid, and also from mitochondria. Specifically, single-cell focused ion beam scanning electron microscopy (FIB-SEM) allowed me to demonstrate that a retina-like portion of the ocelloid is a small part of a much larger peridinin-plastid that ramifies throughout the Nematodinium cell. Lastly, I investigated the evolution of nematocysts in Polykrikos spp. and Nematodinium sp. using a combination of transcriptomics, TEM, SEM, and FIB-SEM, and inferred that “nematocysts” in these groups evolved independently from those in cnidarians. Thus, nematocyst-like extrusive organelles appear to have evolved multiple times in eukaryotes. The data presented in this thesis show how extreme subcellular complexity has evolved in dinoflagellates through both endosymbiotic and autogenous origins.

View record

Endosymbiosis has played a major role in shaping eukaryotic cells, their success and diversity. At the base of the eukaryotic tree, an α-proteobacterium endosymbiont in a protoeukaryotic cell was converted into the mitochondrion through its reductive evolution, endosymbiotic gene transfer (EGT) and the development of a protein targeting system to direct the products of the transferred genes to this organelle. Similar events mark the plastid evolution from a cyanobacterium. However, the primary endosymbiosis of plastid, unlike the mitochondrion, was followed by the secondary and tertiary movement of this organelle between eukaryotes through analogous endosymbiotic reduction, EGT and evolution of a protein targeting system and many subsequent independent losses from different eukaryotic lineages.The obligate tertiary diatom endosymbiont in a small group of dinoflagellates called ‘dinotoms’ is exceptional in that it retains most of its ancestral characters including a large nucleus, its own mitochondria, plastids and many other eukaryotic organelles and structures in a large cytoplasm all enclosed in and separated from its dinoflagellate host by a single membrane. This level of conservation of ancestral features in the endosymbiont suggests an early stage of integration. In order to investigate the impacts of endosymbiosis on the organelle genomes and to determine the extent of EGT and the contribution of the host nuclear genome to the proteomes of the organelles, I conducted mass pyrosequencing of the A+T-rich portion of the DNA extracted from two dinotoms, Durinskia baltica and Kryptoperidinium foliaceum, and the SL cDNA library constructed for D. baltica.The plastid and mitochondrial genomes of these two dinotoms were sequenced, and the results indicated that, despite the permanent symbiosis between the host and its endosymbiont in dinotoms and in spite of small variations, the dinotom organelle genomes have changed very little from those of free-living diatoms and dinoflagellates. There was also no sign of EGT to the host in D. baltica, suggesting a strict compartmentalization in which the host mitochondria remain reliant on the host nucleus while the endosymbiont organelles, mitochondria and plastids, stay entirely dependent on the endosymbiont nucleus with no genetic exchange between the host and endosymbiont.

View record

As obligate parasites of animals and humans, apicomplexan parasites contain many unique characteristics that are critical to their lifestyle, but bear little resemblance to other eukaryotes. Several free-living relatives of apicomplexans represent a great potential in understanding early apicomplexan evolution. The photosynthetic Chromera velia provides a particular promise in addressing the long-contentious origin of the apicomplexan plastid. The data presented here provides evidence that the photosynthetic plastids in Chromera velia and another novel alga, CCMP3155 (later named Vitrella brassicaformis), are closely and specifically related to the apicomplexan plastid, and that they together are related to plastids in dinoflagellates. The ancestral plastid went through an unusual reduction in gene content and acquired unique features such as Rubisco II and transcript oligouridylylation. The plastid genome in C. velia is interesting on its own. Two proteins of Photosystem I and ATP synthase have been split to two fragments, which are independently expressed. The genome also appears to exist prevailingly as a linear monomer. These, and additional unprecedented features, redefine our understanding of plastid genome architecture and point to intra-chromosomal recombination as a putative driving force. Assessing environmental distribution of the newly-discovered Chromera and Vitrella leads to a discovery of six additional apicomplexan-related linages (ARLs) comprising 1,316 sequences primarily from coral reefs environments. The most abundant lineage, ARL-V, is novel and exclusively associated with coral tissue and surface samples. ARL-V is present in at least 20 species of symbiotic corals across time and space, which suggests that its relationship with corals is of potential significance to the reef ecosystem. Successful culturing of five Colponema isolates provides the first molecular data for another apicomplexan relative. The genus represents two independent lineages, one of which is the closest sister to apicomplexans and dinoflagellates. Mitochondrial genome data from both lineages reveals a gene-rich content and suggests that a linear monomeric structure with telomeres was ancestral to all alveolates. Altogether, this data illustrates the significance of Chromera, Vitrella, Colponema and several uncultured lineages in illuminating early evolution in apicomplexans and alveolates.

View record

The eukaryotic translation elongation factor EFL (for EF-Like) is a paralogue of the better-known elongation factor 1-alpha (EF-1α), which brings aminoacyl-tRNAs to the ribosome during translation. This essential protein was thought to be ubiquitous in eukaryotes until the recent discovery of EFL in a small number of diverse, mainly unicellular, eukaryotic organisms that were found to lack EF-1α. Because of the great evolutionary distances between EFL-encoding lineages and the near mutual exclusivity of the two proteins, the observed complex distribution of EFL was initially attributed entirely to multiple lateral gene transfers. In the enclosed chapters, the distribution of EFL was characterized in more detail in four distantly related eukaryotic lineages at both fine and broad taxonomic scales in order to better understand the effects that endosymbiotic gene transfer, differential loss, and lateral gene transfer have had on the molecular evolution of EFL. Endosymbiotic transfer of EFL was detected in the chlorarachniophytes, a group of algae whose secondary plastids retain a vestigial nucleus, known as a nucleomorph, in their reduced eukaryotic cytoplasm, known as the periplastid compartment (PPC). The endosymbiotically transferred EFL carries a bipartite targeting sequence similar to those of plastid-targeted proteins in this group and to plastid- and PPC-targeting sequences in cryptomonads to direct it to the PPC, suggesting similarities in the way these two lineages have solved their shared challenge of targeting to complex plastids with nucleomorphs. No clear phylogenetic evidence for lateral transfer of EFL has yet emerged; rather, differential loss of EFL and EF-1α from an ancestral state of co-occurrence was characterized in euglenozoans and detected in publicly available data from heterokonts and opisthokonts, unexpectedly revealing a significant role for this process in shaping the complex distribution of EFL and EF-1α. This finding serves as a cautionary reminder that adequate taxon sampling and a robust organismal phylogenetic hypothesis are crucial in order to correctly infer lateral gene transfer.

View record