Search

Sphingolipids (SLs) are complex lipid-derived structures that are essential components of cell membranes in eukaryotes and some bacteria. SLs and their complex derivatives ceramides are known to be involved in multiple processes such as the formation of lipid rafts, cell signalling and membrane trafficking. The first step of SL biosynthesis is universal to all sphingolipid-producing organisms from bacteria to humans and is catalysed by the enzyme serine palmitoyltransferase (SPT). SPT is a member of the alpha-oxoamine synthase (AOS) family of pyridoxal- 5'-phosphate-dependent enzymes. All AOS family enzymes retain a high degree of structural homology and catalyse the decarboxylative Claisen-like condensation of amino acids with thioester substrates. The SPT enzyme catalyses the formation of the universal SL precursor, 3-ketodihydrosphingosine (KDS), by condensation of L-serine and coenzyme A-derived palmitic acid. Being the key controller in SL biosynthesis, SPT plays a big role in regulating natural and pathological processes. A lot of research interest has been recently generated by SLs isolated from bacterial members of the human microbiome and their roles in human health. Increasing evidence suggests that some of these SLs possess immunoregulatory effects and can have a direct impact on the immunity of the host. Bacteroides fragilis is a commensal gut-dwelling bacterium that belongs to a few human microbionts known to produce unique iso-branched sphingolipids (isoSLs); these have been shown to influence the human iNKT cell count. The production of SLs in B.fragilis is completely regulated by a gene product BF2461. In this work, BF2461 was expressed and purified; using a combination of UV-vis spectrometry, enzymatic assays, mass spectrometry and protein X-ray crystallography, it has been confirmed to be an SPT. The substrate specificity of the BfSPT has been assessed with a range of different chain-length substrates, including less common 15 and 17-carbon chain length coenzyme A substrates. The enzyme can produce different types of SL precursors with a preference for the 16-carbon chain substrate palmitoyl- CoA. However, at high levels of PCoA, a substrate inhibition is observed that might point to a natural control mechanism employed by the bacterium in favour of producing iso-branched SLs (isoSLs). The structure of BfSPT has been elucidated in a complex with its amino acid substrate L-serine. Search and analysis of putative SPTs from other microbiome-associated bacteria that produce isoSLs show that they share high similarity with an average amino acid conservation of 74%, suggesting they might be adapted to a particular type of substrate. In this respect, BfSPT might be the first isoSL-producing SPT to be structurally characterised, and the first one to have a direct impact on human health. Further structural data were obtained on protein complexes with L-cycloserine and L-penicillamine, some common inhibitors of the PLP-dependent enzymes. The structure obtained in the presence of L-penicillamine provides the first direct structural evidence of the inhibitory mechanism by a thiazolidine complex formation in the active site of a PLP-dependent enzyme. These findings shed light on certain aspects of the reaction and inhibition mechanisms of BfSPT as well as opening new prospects into researching this interesting target and its impact on the human microbiome.

Microbial communities play important roles in human health and disease, are essential components of terrestrial and marine ecosystems, and are crucial for producing commercially valuable molecules in industrial processes. These communities consist of hundreds of species involved in complex interactions. Mapping the interrelationships between different species in a microbial community is vital for understanding and controlling ecosystem structure and function. Advances in sequencing and other omics technologies have led to thousands of datasets containing information about microbial composition, gene expression, and metabolism in microbial communities associated with human hosts and other environments. These provide valuable information in understanding how microbes interact with each other and how their interactions affect the health of their host (e.g., human or plant). Furthermore, understanding these interactions paves the way for the rational design and modulation of synthetic communities for producing antibiotics, biofuels, and pharmaceutical products. The first part of my thesis is focused on improving the workflow for the inference of microbial co-occurrence relationships from abundance data. Toward this goal, we developed Microbial Co-occurrence Network Explorer (MiCoNE), a pipeline that infers microbial co-occurrences from 16S ribosomal RNA (16S rRNA) amplicon data. This pipeline involves numerous complex steps that require specific tools and parameter choices, posing open questions about the robustness and uniqueness of the inferred networks. Through MiCoNE, we systematically analyzed how these choices of tools affect the final network and proposed guidelines on appropriate tool selection for a particular dataset. We envisage that this pipeline could be used to integrate multiple datasets and generate comparative analyses and consensus networks that can guide our understanding of microbial community assembly in different biomes. The second part of my thesis focuses on microbe-host interactions rather than microbe-microbe associations. In particular, we sought to predict the effects of microbial metabolites on human receptors and their associated regulatory pathways. We specifically focus on the Aryl hydrocarbon receptor (AHR), a ligand-mediated transcription factor involved in tumorigenesis. In this project, we aimed to systematically predict the binding of diverse microbial metabolites secreted from microorganisms found in the human oral microbiome to the AHR to identify links between the microbiome and cancer initiation. We further build a mathematical model of the AHR regulatory pathway and model the effects of ligand binding on downstream molecules. We envision that these methods could be used to predict the impact of microbial dysbioses on human regulatory pathways. In the final part of my thesis, we turn to the question of whether computational algorithms can help control microbial community growth to achieve specific objectives. In particular, we describe the development of a reinforcement learning algorithm to learn optimal environmental control strategies to steer a microbial community towards a certain goal, such as reaching a specific taxonomic distribution or producing desired metabolites. We train the reinforcement learning framework through community-level simulations of genome-scale metabolic models (GEMs) for different microbial species in bioreactor systems. In this project, we simulate a simple case study with two auxotrophic mutants to verify the algorithm's validity. Ultimately we aim to simulate the implementation of the algorithm in experimental bioreactor systems. Overall, the work presented in this thesis demonstrates how microbe-microbe and microbe-environment (including microbe-host) interactions represent plastic system-level properties whose understanding can help unravel the role of microbial communities in specific diseases. Correspondingly, manipulating these interactions, e.g., by appropriately modifying environmental conditions, can serve as a promising strategy for steering communities towards desired states, including producing valuable molecular products.

Parkinson’s disease (PD) is a neurodegenerative disease affecting nearly 1 million people in the United States leading to a total estimated financial burden of $52 billion. The hallmarks of PD include the amyloidogenic aggregation of α-synuclein, loss of motor function and loss of dopaminergic neurons in the substantia nigra. Symptoms occur late in the course of disease and the current treatment is focused mainly on alleviating symptoms rather that preventing the disease progression. Interestingly, nearly 80% of PD patients experience gastrointestinal dysfunction like constipation, which may precede the onset of motor symptoms by years. Moreover, many studies have shown that there is a significant difference in the gut microbiome between PD patients and healthy individuals. Coupled with the early gastrointestinal dysfunction, these studies have led to the speculations that the gut microbiome may be implicated in PD. The direct evidence of gut microbiome’s role in PD was provided by a recent study in which colonization α-synuclein over-expressing (AOE) mice gut with E. coli expressing a functional amyloid, CsgA lead to enhanced PD associated pathophysiology. Moreover, the in vitro amyloidogenic aggregation of α-synuclein was accelerated by E. coli CsgA. In addition to E. coli, the human gut is home to diverse species of microorganism with potential CsgA producing microorganism. The impact of gut bacterial CsgA on α-synuclein aggregation and the mechanism behind the accelerated α-synuclein aggregation is still poorly understood. My work has identified CsgA homologs from diverse human gut bacterial species and characterized their aggregation kinetics. Gut bacterial CsgA homologs display diverse aggregation kinetics and differentially modulate α-synuclein aggregation. Interestingly, slow aggregating CsgA homologs have enhanced ability to accelerate α-synuclein compared to fast aggregating CsgA homologs. In addition to this, slow aggregating CsgA homologs form stable dimeric species compared to fast aggregating CsgA homologs and the stability of these dimeric species is correlated their ability to accelerate α-synuclein aggregation. I also report a complex between CsgA and α-synuclein that serves as a “seed” to facilitate acceleration of α-synuclein aggregation. Since the differential aggregation kinetics of gut bacterial CsgA homologs governs their ability to accelerate α-synuclein aggregation, I also address the sequence determinants of amyloidosis of gut bacterial CsgA homologs. There are certain “gatekeeper” amino acid residues in CsgA which inhibit the intrinsic amyloid formation propensity of CsgA. In addition to the earlier described negatively charged aspartic acid (D) residues, I identified new positively charged arginine (R) and lysine (K) gatekeeper residues in CsgA and demonstrate the molecular mechanism governing CsgA amyloid formation. The electrostatic interactions between the positively (R and K) and negatively charged (D) gatekeeper residues determine the nucleation rate of CsgA amyloid formation. Salt mediated charge masking or pH induced charge neutralization/reversal severely affected the inhibition of the intrinsic amyloid formation propensity of CsgA. Taken together, my work reveals the mechanism behind the role of gut bacterial functional amyloid in PD causation. Additionally, my work also provides molecular insights into gut bacterial functional amyloidogenesis paving the way to developing future therapeutics to prevent and/or treat PD.

We have known for centuries that the microbial communities inhabiting our bodies are linked to human phenotypes, including disease. Recently, the massive scale of DNA sequencing output has made the microbiome field reliant on data-driven “discovery:” seeking biological hypotheses from associations in observational data. However, identifying reproducible and interpretable relationships between host phenotype (e.g. disease) and microbial features (e.g. species, pathways, genes), has proven to be difficult. Here, we propose an efficient path through the complexities of data-driven discovery towards identifying biologically meaningful and clinically impactful host-microbe associations. First, to understand the “data” of the human microbiome, we aimed to quantify the genetic diversity of different body sites. We compared their gene content to each other as well as to other non-human ecologies. We show that at least 50% of the genes in a given sample are ecology-specific, and we identify environmentally “conserved” genes across different human body sites, organisms, and ecologies. We next aimed to construct strategies for accurately and reproducibly modeling host-microbe interactions. We first built a software tool for scalable, low-cost cloud computing. Then, using a range of machine learning approaches, we compared the ability of microbial taxonomies, pathways, and genes to predict host phenotype, identifying that the gene “data type” far outperformed other representations of the microbiome. We next built software for massive-scale sensitivity analyses, which we used to probe the robustness of reported associations, showing that ⅔ of published human gut microbiome-phenotype associations are not immediately reproducible. Finally, we demonstrated the potential impact of data science in the microbiome through two examples. First, we compared the ability of the human microbiome versus human genetic common variants to predict disease, identifying that the former outperforms the latter. Second, we identified robust, gene-level indicators of seven different phenotypes, indicating the potential for these “gene-level microbiome architectures” to serve as cross-disease diagnostics and tools for testable hypothesis generation. Overall, by integrating across statistical and algorithmic methods, we provide a roadmap for data-driven and efficient discovery in the human microbiome. We hope that our efforts will enable faster translation of microbiome data science into both clinical and experimental settings.

Bacteria inhabit every known ecosystem, from deep-sea hydrothermal vents to plant surfaces. The human body represents one such microbially-dominated ecosystem with the identity and function of human-associated microbiota being intimately connected with human health. However, few specifics are known about the processes governing the distribution and composition of host-associated microbial communities, challenging efforts to alter or maintain certain host-associated bacterial communities. Thus, studying the ecology and evolution of human-associated bacteria is critical both for improving human health and more broadly for better understanding microbial community organization and function. In this dissertation, I use the healthy human oral microbiome as a model system to investigate the processes shaping bacterial populations from an ecological and evolutionary perspective. Altogether, these investigations reveal genomic signatures indicative of fine-scale adaptation to diverse oral niches. In my first chapter, I employ a metapangenomic approach to combine Human Microbiome Project (HMP) metagenomes with pangenomes from public genomes to study the diversity of microbial residents of three oral habitats: tongue dorsum, buccal mucosa, and supragingival plaque. For two exemplar taxa, Haemophilus parainfluenzae and the genus Rothia, metapangenomes revealed distinct genomic groups based on shared genome content. H. parainfluenzae genomes separated into three distinct subgroups with differential abundance between oral habitats. Functional enrichment analyses identified genes encoding oxaloacetate decarboxylase as diagnostic for the tongue-abundant subgroup, suggesting a metabolic adaptation to the tongue habitat. For the genus Rothia, while most R. mucilaginosa were restricted to the tongue as expected, two genomes represented a cryptic population of R. mucilaginosa in many buccal mucosa samples. For both H. parainfluenzae and the genus Rothia, I identified not only limitations in the ability of cultivated organisms to represent populations in their native environment, but also specifically which cultivar gene sequences were absent or ubiquitous. These findings provide insights into population structure and biogeography in the mouth and form specific hypotheses about habitat adaptation. These results also illustrate the power of combining metagenomes and pangenomes to investigate the ecology and evolution of bacteria across analytical scales. In my second chapter, I address the process of how bacteria adapt over time, namely whether adaptation occurs principally by gene-centric or genome-centric processes. I sequenced 81 metagenomes, shotgun sequencing of total community DNA, from the tongues of 17 healthy volunteers at four to seven timepoints obtained over intervals of days to weeks. I obtained 390 high-quality metagenome-assembled genomes (MAGs) defining population genomes for 55 bacterial genera. Decomposing MAGs by metagenomic nucleotide variants revealed MAG-defined bacterial populations were composed of up to five detectable subpopulations. Bacterial subpopulations were generally stable over time, yet some swept from low abundance to dominance over a period of days, a pattern suggestive of genome-centric adaptation. At the gene level, the vast majority of genes in each MAG were tightly coupled over time based on their frequency in the metagenomes of different mouths. Some genes were uncoupled but their frequencies did not change directionally, nor did nonsynonymous codon variants become fixed within these uncoupled genes, suggesting transient effects. Altogether, these results demonstrate that both gene- and genome-wide processes occur on daily timescales but likely with different ecological ramifications. I suggest a model where genome-wide selection of ecotypes is the dominant mode of adaptation with short-term changes in gene frequency also occurring within the population. My third chapter examines individuality in the tongue microbiome from a genome-centric perspective to reveal the genomic and functional differences of bacterial populations occurring in only a single mouth. The v4 region of the 16S rRNA gene was sequenced from all 81 samples obtained in Chapter 2 and analyzed to obtain amplicon sequence variants (ASVs) for 50 bacterial genera. Sixteen out of 287 ASVs were found to be endemic, i.e., they were abundant in a single mouth but absent elsewhere. Linking MAG and haplotype data revealed that the endemic ASVs represent genetically distinct populations, differing in tens to hundreds of genes encoding diverse functions related to mobile elements, defense, carbon metabolism, and the cell envelope. Approximately 1% of the entire set of almost 700,000 genes were individual-specific, and their predicted functions revealed phage- and defense-related genes were enriched relative to genes not restricted to a single mouth. These genomic and functional differences suggest individual-specific populations occupy similar but distinct niches across individuals.

Clostridioides difficile is a Gram-positive pathogen that is one of the most common causes of health care-associated infections in the U.S. Because of its importance to human health, researchers are trying to understand risk factors for infection and identify new therapies to treat disease. A “healthy” gut microbiome has the ability to protect the host against colonization of C. difficile or other pathogens by providing colonization resistance. One proposed mechanism for colonization resistance is through competition for nutrients with pathogens by the members of the healthy microbiome. We tested whether competition for nutrients was important for colonization resistance in human fecal minibioreactor arrays (MBRAs), an in vitro model of C. difficile colonization resistance that can be used to culture fecal communities from different healthy humans. We found that C. difficile growth was limited when amino acids were depleted and that ability to metabolize proline, an amino acid shown to be important in mouse models of infection, was required for C. difficile to colonize in some fecal communities but not others. As several previous studies point to proline as potentially important niche during human infections, these results provide further support for use of this model to understand factors important for C. difficile colonization that vary across fecal communities and can be used for further development of predictive models of important nutrient niches. Future studies that build upon this work could combine sequence data, metabolomics data, and compound and pathway enrichment analysis to begin developing models that predict important nutrient niches in specific fecal communities as a way to develop personalized clinical therapy. Advisor: Jennifer Auchtung

Human microbiomes are increasingly seen as a key to understanding human biology, whether it be in studying health/disease or in documenting human diversity. Anthropological interest in human microbiomes has primarily focused on inventorying shifts microbial abundance between lifestyles and over time. While undoubtedly valuable information, these taxonomic inventories lack application of theoretical frameworks that can provide a deeper understanding of human-microbiome interactions. In my work, I integrate microbiome data with ideas typically used in the study of macroecological systems. I present a background on ecological frameworks and their general applicability to human microbiome studies in Chapter 1. In Chapters 2-4, I present unique research on human microbiome ecology. Chapter 2 is a first-of-its-kind study on ecological dynamics in ancient human microbiomes, in which I demonstrate how network analysis can inform on general microbiome community structure in ancient coprolites and dental calculus, as well as demonstrate overlap in key ecological signatures between ancient and contemporary non-industrial populations. Chapter 3 is a demonstration of the value of interrogating potential microbiome stability and resilience in a specific niche, namely short-chain fatty acid (SCFA) production. Our research indicates that non-industrial gut microbiomes are more resilient for SCFA production and industrial SCFAs are encoded by a few, closely related species. These results come in the face of substantial database bias that inhibits study of non-industrial gut microbiomes. Finally, Chapter 4 describes vaginal and gut microbiomes in women with Ovarian Cancer (OC). We determine that women with OC have lower than expected vaginal Lactobacillii dominance, which creates ecological space for an opportunistic bacteria such as Escherichia, to thrive in women who suffer quick recurrence of cancerous growth after platinum chemotherapy.

Anthropological studies of human biology are predominately field-based, and burdened, by the need for well-preserved biological samples. In the emerging application of multi-omics, definitions of “well-preserved” and preservation strategies have had limited study, particularly those data that inform the biology of the human ecology. These human ecological data are a frequent objective of metabolomics and microbiome research. Metabolomics, through exploration of the totality of small molecules known as metabolites, offers a way to directly observe the molecular phenotype. These small molecules offer just as much valuable information to molecular anthropology as DNA and RNA, but there is a surprising lack of inquiry into how these molecules are preserved in samples and how molecular preservation impacts results and interpretations. For most metabolomic studies, the standard sample collection procedure involves snap-freezing the sample within 15 minutes of collection and storing at -80ºC. However, this is often unfeasible for field-based sample collection. Metabolome taphonomy, the study of how metabolome profiles are impacted by environmental processes as well as sample collection and preservation/preparation strategies, is still poorly understood. This thesis considers sample storage, with attention to human gut microbial samples. Consequently, this thesis presents two complementary studies, one with a focus on the metabolome and one focused on the microbiome taxonomic inventory, to determine if the application of the common DNA and RNA preservative RNAlater provides a valuable method for conserving ecological data from both approaches. Ten human fecal samples previously collected and frozen at -80ºC were homogenized, aliquoted, and subjected to treatment that simulates different levels of cold storage in the field: 22-25ºC, 4ºC, and -80ºC. To assess the impact of preservation methods on metabolite and bacterial taxonomic profiles, subsets of these aliquots were further treated with different preservation techniques, such as RNAlater. Metabolomic and bacterial taxonomic profiles were characterized using liquid chromatography-tandem mass spectrometry and 16S rRNA amplicon sequencing, respectively. These results will inform field sample collection and best storage practices for human biological approaches that apply multi-omic studies.

Enterococcus can be both a friend and a foe in the human microbiome. As an opportunistic pathogen, Enterococcus is normally benign. However, Enterococcus is responsible for many hospital-acquired infections and has shown rising rates of vancomycin resistance. Bacteriophages (phages) could be an alternative to antibiotics to target these antibiotic resistant bacteria, but the evolutionary and phenotypic outcomes of phage-bacteria interactions need to be investigated more thoroughly. We begin by investigating the gut microbiome of preterm infants that had been exposed to antibiotics to learn about the scenarios in which Enterococcus blooms occur in the gut (Chapter 1). We show that antibiotic exposed gut microbiomes are dominated by facultative anaerobes such as Enterococcus and Enterobacteriaceae. Further, we show that the bacterial composition is correlated to the overall metabolite profile. Metabolomics is a powerful tool for investigating microbial metabolism, and we contribute to the effort of developing standardized practices for metabolomics in the human microbiome by showing that freezing microbial samples required for long term storage (Chapter 2). Next, we investigated how phages could be used as therapeutics for treating Enterococcus blooms and infections (Chapters 3, 4, and 5). When Enterococcus is grown with its phages in vitro, it evolves resistance to phage infection by mutating exopolysaccharide synthesis genes. These mutations alter the exopolysaccharides on the surface of the bacterial cell to prevent binding of phage. Further, this mechanism of resistance appears to be a general mechanism leading to resistance against a diverse array of phages. This work demonstrates that experimental evolution is a powerful tool for characterizing interactions between bacteria and phages.Phage therapy is often administered as a cocktail of multiple phages, but there are no rules or best practices described for combining phages to be most effective. We show that in vitro, Enterococcus phage cocktails are more effective at preventing the growth of phage-resistant mutants, but the composition of the cocktail is important. Genetically diverse phage cocktails performed better than cocktails of related phages. My work demonstrates some of the outcomes of phage-Enterococcus interactions and will move us closer to applying phage therapy to treat Enterococcus infections.

Microorganisms are critical to immune system development and physiology, yet the factors that drive initial colonization and human microbiome assembly are largely unknown. Early molecular approaches to study the microbiology of human body sites relied upon the direct amplification and sequencing of 16S rRNA genes, but this only produces a coarse catalog of the organisms present. In contrast, genome-resolved metagenomics involves recovery of genomes directly from whole community DNA, enabling prediction of biosynthetic capacities and providing the ability to differentiate the capabilities of closely related strains. However, genome-resolved metagenomics is computationally challenging and validated methods for many types of analyses are lacking. In this thesis, custom genome-resolved metagenomic methods were developed to determine the structure of microbial communities, monitor the activities of bacteria in situ, and track their evolution. The research focused primarily on the colonization and development of microbial communities that live in, and on, premature infants. Discovered patterns of fine-scale bacterial diversity, evolution, and functional potential shed light on early microbiome assembly, and highlight factors that contribute to necrotizing enterocolitis, one of the most common diseases of premature infants. Transmission pathways and reservoirs of bacteria and microbial eukaryotes that can cause nosocomial infections were identified.Due to the sensitivity of metagenomic methods, foreign DNA sequences can be detected in metagenomic datasets. A case-study involving the identification of the source of introduced sequences in metagenomes was conducted and the specific physical source of the contaminant identified. Though very detailed genome sequence comparisons it was possible to measure the in situ evolution rate of the reagent contaminant over a three year period, enabling the near correct estimation of the introduction time of the contaminant into the reagent production facility. The research established the methods for genome-resolved metagenomics-based microbial forensics and strain tracking and showed that application is possible, even in extremely complex environments like soil.High-resolution analyses are needed to determine if microbes in different environments are the result of strain transfer events. The necessary genome-wide comparisons cannot be performed with 16S rRNA sequencing, the most common method for study of the human microbiome, leaving basic questions related to the strain-level diversity and body-site specificity unanswered. To address these questions, strain-tracking analyses were applied to metagenomes derived from samples of the mouth, skin, and gut microbiomes of premature infants. The results highlight the extreme lack of body-site diversity during very early colonization of premature infants in the neonatal intensive care unit. Surprisingly, identical bacterial genomes for organisms such as Escherichia coli, Klebsiella pneumoniae and Citrobacter koseri were found in the mouth, skin, and gut microbiomes. Differential genome coverage was used to measure their bacterial population replication rates in situ. In all cases, the replication rates for same bacterial populations in different body sites were faster in mouth and skin compared to the gut, despite the fact that these bacteria are traditionally considered gut colonists. Finally, strain-level analysis of polymorphic sites across the C. koseri genome were used to define 10 subpopulations, implying initial colonization of premature infants by multiple individual cells with distinct genotypes. Methods for rapid, accurate and reproducible “de-replication”, the process of grouping recovered genomes together based on similarity and choosing the best representative genome from each group, are needed for genome-resolved metagenomic analyses. This task requires a number of steps, including mass pairwise genome comparison, evaluation of completeness and contamination, and generation of explanatory figures to visualize and validate the dereplication process. An open-source program, “dRep”, was developed and validated. The method achieves very similar results as naive pairwise clustering algorithms, but with an order of magnitude speed increase due to use of a biphasic algorithm. Importantly, individual samples in sample sets can now be assembled independently and the genomes effectively de-replicated, reducing the incidence of chimeric sequences and improving genome recovery over results for co-assemblies.Delineation of bacteria as belonging to the same versus different species is a longstanding problem in microbiology. Fundamental questions related to the existence of species and how species should be differentiated remain, and the answers have both practical and evolutionary implications. A large public dataset comprised of >5,000 genomes acquired directly from metagenomes was analyzed. In conjunction, genome-based metrics that could be used to define bacterial species boundaries were evaluated. A distinct gap in the distribution of average nucleotide identity (ANI) values at 95% ANI exists, supporting the existence of discrete species. ANI was compared with metrics of selection for non-synonymous versus synonymous substitutions and for homologous recombination to identify processes that could lead to species clusters. The 95% ANI value corresponds approximately with the genetic distance beyond which homologous recombination drops to near zero. The findings implicate sequence divergence-based breakdown in homologous recombination as the evolutionary force responsible for bacterial speciation. 50 genes were evaluated to provide a practical means to define species content when genomes are not recovered from metagenomes for most community members. Although 16S rRNA gene sequences cannot be used for this purpose, the nucleotide sequences of several ribosomal proteins were found to be reasonable proxies for the relevant genome ANI value. Microbial eukaryotes are particularly understudied in the human microbiome, yet they are considered to be emerging health threats. Genomes from microbial eukaryotes can be reconstructed from human microbiome metagenomic datasets. The results are greatly improved through use of a recently developed machine learning algorithm, EukRep, which can identify eukaryotic DNA based on the sequences alone. EukRep was used to scan thousands of metagenomes from the premature infant gut and hospital room environments and fourteen novel eukaryotic genomes were reconstructed. Two of these, for a Diptera (fly) and a Rhabtida (worm), were novel at the class level. Importantly from the perspective of tracking nosocomial agents, genomes from the same eukaryotic species were recovered from both infants and hospital room environments. Population heterogeneity and zygosity of genomes were lower in genomes recovered from the hospital room as compared to those recovered from premature infant samples, which could reflect years of inbreeding or strong selection imposed by room conditions. Together this work indicates that the hospital room, especially the sink, may be a reservoir of infant-colonizing fungal strains.Necrotizing enterocolitis (NEC) involves extreme bowel inflammation and necrosis and has a mortality rate of around 30%. Various lines of evidence point to the human microbiome as a central factor in disease development, yet no consistent microbial signal for NEC onset has been identified. We performed large-scale genome-resolved metagenomic analyses of thousands of prospectively collected premature infant fecal samples and used a machine learning classifier to identify signals that predict imminent development of NEC. Significant associations were found related to the abundance of Klebsiella, bacteria encoding fimbriae, and specific types of secondary metabolite clusters, and the in situ growth rate of bacteria overall. NEC development could be promoted by metabolic imbalances related to the rampant growth of particular bacterial strains and/or the stimulation of human TLR4 receptors by Klebsiella and fimbriae. Together, the results identify potential biomarkers for early detection of NEC and possible targets for microbiome-based therapeutics and probiotics.

Colibactin is a genotoxin produced by human gut Escherichia coli strains. This metabolite is biosynthesized by the non-ribosomal peptide synthetase (NRPS) and polyketide synthase (PKS) enzymes encoded in a 54-kilobase biosynthetic gene cluster (the pks island) where all of the NRPS and PKS genes are required for genotoxicity. E. coli strains harboring the pks island induce DNA double strand breaks in HeLa cells and are implicated in the development of colitis-associated colorectal cancer in animal models. However, despite multiple attempts, the active genotoxin has not been isolated and its chemical structure is currently unknown. In the quest to solve the structure of colibactin, in vitro analyses of the PKS and NRPS modules from the pks island have allowed us to develop strategies for isolating pks-dependent metabolites and to formulate hypotheses for structure and biosynthesis of the active genotoxin. In addition to guiding isolation, we believe application of the in vitro approach is also indispensable for gaining mechanistic insights into individual steps of colibactin biosynthesis. In the past few years, my study of pks assembly line enzymes in vitro has led to the discovery of novel assembly line enzymology and advanced biosynthetic hypotheses to better predict colibactin’s structure. The non-canonical biosynthetic logic of the pks island will be presented in the following chapters. In Chapter 2, the study of PKS modules is reported with an emphasis on the formation, recognition, and incorporation of the rare PKS extender unit aminomalonyl-acyl carrier protein in the colibactin assembly line. We demonstrated that ClbG, a freestanding acyltransferase (AT), recognizes this building block and transfers it to multiple PKS modules from the pks gene cluster, including three atypical-AT PKSs (ClbC, the PKS module of ClbK, and ClbO) and a cis-AT PKS ClbI. These findings suggest that the pks assembly line has the potential to incorporate the non-canonical extender unit more than once into colibactin’s scaffold and will inform the identification of new aminomalonate-derived candidate precolibactins. Chapter 3 details the characterization of NRPS modules in the colibactin assembly line. We confirmed the predicted substrates of all NRPS modules and discovered an unusual NRPS module, ClbH, that uses S-adenosylmethionine (SAM) as a non-proteinogenic amino acid building block for amide bond formation and ClbI-mediated cyclopropanation. Our results revealed a distinct strategy for cyclopropane formation in natural product biosynthesis that requires direct activation of the underutilized carboxylic acid of SAM. This study also shed light on the biosynthetic origin of a functional group that is likely important for colibactin’s genotoxicity. Finally, Chapter 4 describes the characterization of an unusual PKS machinery that combines an iterative PKS module and an atypical-AT PKS module to incorporate malonate units in the early stage of colibactin biosynthesis. We established that the PKS module of the NRPS-PKS hybrid ClbB is capable of catalyzing two rounds of malonate extension. Although early models proposed that the second malonate unit was loaded onto the ACP domain of ClbC, results from our site-directed mutagenesis work indicate that this domain is not required for malonate extension. Collectively, studies presented in this thesis highlight the utility of in vitro biochemistry in revealing the unusual features of NRPS and PKS assembly lines that cannot be predicted from bioinformatics and providing insights into unprecedented biosynthetic logic of natural products.

The human microbiome has been identified as a significant biological aspect of the human body with correlations to many characteristics of the host including age, sex, diet, and complex disease states such as cancer, obesity, and IBD. In addition, there is evidence that the host genome plays an important role in regulating the microbiome. However, the relationship is not clear and automatic methods will be required to uncover the connections between host genomic and metagenomic datasets. We have developed a method using LASSO regression to screen for associations between SNPs in the host genome and microbiome data. We have applied our method to data from the Human Microbiome Project and identified 97 SNPs (46 non-synonymous) in 90 host genes that show potential association to the microbiome (found in files hmp_16S_snps_genes.xlsx and hmp_mgs_kegg_modules_snps_genes.xlsx). Our method is an early step in automating detection of host-microbiome interactions within genomic data.

Various diseases have been linked to the human microbiota, but the underlying molecular mechanisms of the human microbiota in disease pathogenesis are often poorly understood. Using acne as a disease model, my study aims to understand the functional capability changes and the molecular response of the skin microbiota to host metabolite signaling in disease pathogenesis. By metagenomic analysis, I found that the genes of a pathogenic gene island in the skin bacterium, Propionibacterium acnes, were significantly enriched in acne subjects. By metatranscriptomic analysis, I found that the transcriptional profiles of the skin microbiota separated acne patients from healthy individuals. The vitamin B12 biosynthesis pathway in P. acnes, was significantly down-regulated in acne patients. Through in-vivo and in-vitro experiments, I further demonstrated that host vitamin B12 modulates the activities of the skin microbiota and plays a signaling role in acne pathogenesis. To understand the molecular mechanism underlying the transcriptional regulation of P. acnes genes by vitamin B12, I analyzed ten putative regulatory elements in multiple P. acnes genomes and identified sequence variations among different strains. It suggests that the vitamin B12 modulation of P. acnes transcription is individual microbiome-dependent. My findings suggest a novel bacterial pathogenesis pathway in acne. My study also provides evidence that metabolite-mediated interactions between the host and the skin microbiota play essential roles in disease development. To further elucidate how the skin microbiota interacts with the host in disease pathogenesis, I have developed a protocol for enriching bacterial RNA from human skin biopsy samples, which lays a foundation for simultaneously analyzing the interactions between the skin microbiota and host cells at the transcriptional level.

The new field of ‘omics’ has spawned the development of metaproteomics, an approach that has the ability to identify and decipher the metabolic functions of a proteome derived from a microbial community that is largely uncultivable. With the development and availabilities of high throughput proteomics, high performance liquid chromatography coupled to mass spectrometry (MS) has been leading the field for metaproteomics. MS-based metaproteomics has been successful in its’ investigations of complex microbial communities from soils to the human body. Like the environment, the human body is host to a multitude of microorganisms that reside within the skin, oral cavity, vagina, and gastrointestinal tract, referred to as the human microbiome. The human microbiome is made up of trillions of bacteria that outnumber human genes by several orders of magnitude. These microbes are essential for human survival with a significant dependence on the microbes to encode and carryout metabolic functions that humans have not evolved on their own. Recently, metaproteomics has emerged as the primary technology to understand the metabolic functional signature of the human microbiome. Using a newly developed integrated approach that combines metagenomics and metaproteomics, we attempted to address the following questions: i) do humans share a core functional microbiome and ii) how do microbial communities change in response to disease. This resulted in a comprehensive identification and characterization of the metaproteome from two healthy human gut microbiomes. These analyses have resulted in an extended application to characterize how Crohn’s disease affects the functional signature of the microbiota. Contrary to measuring highly complex and representative gut metaproteomes is a less complex, controlled human-derived microbial community present in the gut of gnotobiotic mice. This human gut model system enhanced the capability to directly monitor fundamental interactions between two dominant phyla, Bacteroides and Firmicutes, in gut microbiomes colonized with two or more phylotypes. These analyses revealed membership abundance and functional differences between phylotypes when present in either a binary or 12-member consortia. This dissertation aims to characterize host microbial interactions and develop MS-based methods that can provide a better understanding of the human gut microbiota composition and function using both approaches.