Elemental ratios of particulate organic matter (POM) are key to linking biogeochemical cycles. Microbial uptake and allocation of essential biogenic elements (carbon (C), nitrogen (N), and phosphorus (P)) influence the distribution of nutrients throughout the ocean. This dissertation evaluates the role of environmental stress and underlying phytoplankton diversity in driving regional variation in the ratio of particulate C:N:P. Competing hypotheses predict C:N:P equally well due to regional co-variance in environmental conditions and biodiversity. The Indian Ocean offers a unique positive temperature and nutrient supply relationship to test these hypotheses. We collected 248 POC:N:P observations in the eastern Indian Ocean along this environmental gradient. As phytoplankton community composition was constant, biodiversity changes could not explain the elemental variation. Instead, our data supports the nutrient supply hypothesis over the influence of temperature.

Nutrients concentrations are often below detection limits in subtropical ocean regions. We develop two methods to predict nutrient stress to further evaluate its role in particulate C:P regulation. In the first method, we develop a global remote sensing estimate of surface phosphate. Using a mechanistic framework, we develop an artificial neural network to provide a robust basis for developing a remote sensing estimation of surface phosphate. However, C:P predictions using only phosphate did not match observations in either the South Indian or Pacific subtropical gyres. To address this challenge, we develop a second method by applying genomic shifts among microbial communities as ‘biosensors’ for the in situ nutritional environment. We find that our genome-based trait-model significantly improves our prediction of particulate C:P across ocean regions. Furthermore, we detect previously unrecognized ocean areas of iron, nitrogen, and phosphorus stress. Ultimately, we find a combination of nutrient stress accounts for global variation in particulate C:P.

Marine bacterial metabolism of oxygen (O), carbon (C), nitrogen (N), phosphorus (P), iron (Fe), and silica (Si) are essential for elemental cycling on our planet. Because of this, understanding the spatial distribution of marine bacteria, how these bacteria are functioning and interacting with physical processes, and what environmental conditions impact them is essential for understanding current biogeochemical dynamics and for predicting how these dynamics might change in the future. My dissertation research is divided into three projects that aimed to connect large microbial genomic datasets with geochemical and physical processes in open ocean and coastal ecosystems. These studies consisted of a biogeographic survey of the Indian Ocean, an in-depth analysis of mesoscale eddies, and a coastal study at Newport Beach, California.In the first chapter, I linked broad-scale spatial patterns of marine bacteria with geochemical and physical dynamics across the eastern and western Indian Ocean. To achieve this, I incorporated 16S rRNA gene sequences from 465 samples with in-situ CTD data, nutrient concentrations, particulate organic matter (POM) concentrations, stoichiometric ratios, and remote sensing data of sea surface height and geostrophic currents. This extensive analysis revealed 23 unique community structures within the Indian Ocean. It highlighted the southeastern gyre as the area with the largest gradient in bacterial alpha-diversity, identified that the Indian Ocean microbiome was dominated by a core set of taxa, and linked changes in community structure with transitions in physical and geochemical conditions. Importantly, the study identified distinct community structures within mesoscale studies, which served as the foundation for my next chapter. In the second chapter, I investigated the role of mesoscale eddies in shaping bacterial community composition and function across various Indian Ocean regions. This project integrated genomics (16S rRNA and short-read metagenomes), nutrient concentrations, and remote sensing data of 26 eddies of varying age, intensity, and size. Within this study, mesoscale eddies were viewed as physical disturbances that altered communities through dispersal and/or environmental selection. It was found that the origin of the eddy (i.e., coastal waters versus open ocean waters) played a pivotal role in determining how dispersal and environmental selection affected microbial community outcomes. In the third chapter, I examined the interactions between pollutants, marine bacterial diversity and function, and ecosystem recovery within the coastal waters of Newport Beach, California, following the Orange County oil spill. This investigation incorporated short-read metagenomic data, flow cytometry data, polycyclic aromatic hydrocarbon (PAH) concentrations, nutrient concentrations, and POM concentrations. The acute bacterial response to the oil spill was assessed by comparing metagenomically derived taxonomic and functional trends to a 10-year time-series. Notably, there was a rapid and anomalous decline in the abundance of the dominant picophytoplankton, Synechococcus. This decline was coupled with an increase in sulfur-oxidizing and potential hydrocarbon-degrading heterotrophic lineages. There was a lagged response in taxonomy and function to peaks in total PAHs. One week after peaks in total PAH concentrations, the largest shifts in taxonomy were observed, and one week after the taxonomy shifts were observed, unique functional changes were seen. This pattern of response was observed two times during our sampling period and corresponded with the potential resuspension of PAHs. Thus, the impact of the oil spill was temporally extended and demonstrates the need for continued monitoring long after initial exposure.

Prochlorococcus, the smallest known photosynthetic bacterium, is abundant in the ocean’s surface layer despite large environmental variation. There are several phylogenetically distinct lineages within Prochlorococcus and considerable gene gain and loss throughout its evolutionary history. However, the extent to which vertical versus horizontal inheritance shapes its genome diversity across the global oceans is unknown. We observed that Prochlorococcus field populations from a global circumnavigation had a significant relationship between phylogenetic and gene content diversity including regional differences in both phylogenetic composition and gene content that were related to environmental factors. Overall we showed that the environment determines the functional capabilities of successful Prochlorococcus.

We know Prochlorococcus has extensive genetic diversity, including the presence of multiple major clades, its sister taxa Synechococcus displays similar levels of genetic diversity. Prochlorococcus has a clear phylogeography relating to environmental selective pressures, while the biogeography and environmental drivers of Synechococcus clades are more difficult to define. To better characterize Prochlorococcus and Synechococcus genetic diversity we used high throughput sequencing of the marker gene rpoC1 from 339 samples across the Pacific Ocean and Atlantic Oceans. At multiple taxonomic scales (lineage, clade, and SNP) we observed clear parallel biogeography between these two lineages. Overall, this parallel biogeography suggests similar evolutionary selective pressures for these important marine Cyanobacteria.

Oceans are warming and will continue to increase over the next 100 years due to global climate change. Adaptation will likely play a role, but it is unclear how it will impact microbial distributions and processes. To address this unknown, we experimentally evolved a member of the prevalent marine Roseobacter clade to high temperature for 500 generations. We found that this evolved Roseobacter shifted from its usual planktonic growth mode to creating more biofilm at the surface of the culture. Furthermore, this altered lifestyle was coupled with a suite of genomic changes linked to biofilm formation and increased growth in low oxygen transfer environments.

Communities vary across space and time. In addition, they may vary differently at different spatial and temporal scales. It is well established that marine bacterial communities are temporally structured by seasons. However, there are other environmental changes that occur at different temporal scales. Here, we quantified the community variation at three temporal scales across three time series. We found that communities varied at the three time scales with the majority of community variation occurred within seasons. Additionally, we found that community variation correlated with different environmental variables at different temporal scales.

Next I wanted to identify the effect phylogenetic scale has on biodiversity patterns. Most of bacterial ecology defines a taxon as a group that shares more than 97% similarity of the 16S rRNA. However, these groups are not independent of each other. They share evolutionary history charted by divergences into different niches. In order to identify the role of phylogentic resolution on population dynamics, I undertook a 4.5-year sampling project and analysed the variation in relative abundance at different taxonomic levels (genus level, clade level, and subgroup level). I found that the frequency of variation increased as I moved to finer phylogentic resolutions. In addition, the correlation with temperature also changed by changing phylogentic resolution.

Finally, I wanted to identify antibiotic resistance genes in marine environments. Natural environments are quickly being discovered to harbor a number of antibiotic resistance genes. However, marine environments have been mostly overlooked even though they cover more than 70% of Earth's surface, and hold on the order of 1029 bacterial cells. As part of my dissertation, I wished to fill this knowledge gap. By using the culture independent method of functional metagenomics, I discovered genes that conferred antibiotic resistance. Of these genes, 28% were identified as previously known antibiotic resistance genes. The majority were unknown to confer resistance, but had activities similar to antibiotic resistance genes (e.g. transport pumps, enzymatic degradation, etc.). I also identified that the majority of these genes were found in marine bacteria, such as Pelagibacter, Roseobacter, and Prochlorococcus. Therefore, I have uncovered a potential global reservoir of antibiotic resistance genes.

Nutrient supply regulates the activity of phytoplankton, but the global biogeography of nutrient limitation and co-limitation are poorly understood. Bottle incubation experiments have revealed patterns of nutrient limitation but have limited spatial coverage, and surface nutrient concentrations are below detection limits in much of the oligotrophic ocean. In my first chapter I used genomic changes as an indicator of adaptation to nutrient stress. We collected 909 surface metagenomes from the Atlantic, Pacific, and Indian Ocean, quantified the global genome content of Prochlorococcus and inferred local nutrient stress based on shifts in nitrogen, phosphorus, and iron assimilation genes. Our ‘omics-based description of phytoplankton resource use provided a nuanced and highly resolved quantification of nutrient stress in the global ocean.We see a clear association between genome content and nutrient limitation, but the underlying population genetics of these genomic differences and the mechanisms by which they arise are unknown. In my second chapter, I described the functional diversity found within Prochlorococcus and captured the link between low nutrient adaptation and phylogeography. I analyzed 630 surface ocean metagenomes and quantified global variation in gene abundance, phylogeny, and genomic structure through consensus marker genes and metagenomically assembled genomes (MAGs). The analysis of functional diversity and phylogeography revealed widespread flexibility in genomic content, with a phylogenetically conserved haplotype driven by P limitation. While omics-based analysis can give a detailed representation of a microbial community, these methods are spatially and temporally limited to the time and location of sampling. Remote sensing-based descriptions of microbial nutrient limitation have been proposed but have never been validated in situ. For my third chapter, I combined remote sensing and metagenomically derived estimates of nutrient limitation to expand both the temporal and spatial extent of our characterization of nutrient limitation. Based on this in situ validation of the remote sensing products I analyzed the past 10 years of data and captured novel seasonal variation and a recent reduction in global nutrient limitation.

It is important to understand the evolution and prevalence of antibiotic resistance as it is considered a major threat to global public health. Predicting where new antibiotic resistance (AR) genes will rise is a challenge and especially important when new antibiotics are developed. Adaptive resistance allows sensitive bacterial cells to become transiently resistant to antibiotics through changes in gene expression. This provides an opportune time for cells to develop more efficient resistance mechanisms and permanent resistance to higher antibiotic concentrations. Intraspecific genomic diversity may be a driving force in the emergence of adaptive antibiotic resistance. Here, I use an amplification assay adapted from functional metagenomics to investigate cryptic antibiotic resistance, an adaptive resistance mechanism, across eight micro-diverse Escherichia coli from clinical and laboratory origins. Cryptic (unclassified) AR genes primarily conferred resistance within clinical strains as opposed to known AR genes as hypothesized. Most genes conferring resistance within multiple strains were classified AR genes. Cryptic AR genes are highly variable as most conferred resistance in only one strain. Hydrophilic antibiotics are more likely to induce cryptic resistance as resistance occurred to all hydrophilic antibiotics tested. These studies may help detect novel AR genes that confer resistance when upregulated. Additionally, it is important to study antibiotic resistance in the environment, including coastal water. At the beach, people may ingest bacteria harboring AR genes which can lead to infection and/or transfer of genes to commensal and opportunistic pathogens resident in the human microbiome. The ingestion of seafood potentially harboring antibiotic resistant bacteria can lead to indirect contact with antibiotic resistance. Through a ten-year time series, I used metagenomics to provide a comprehensive understanding of the AR genes present within Newport Beach, CA seawater, the temporal distribution of these genes, and the factors driving their frequencies. I found that seasonal and interannual trends of AR genes vary by gene and the taxa carrying them as opposed to a general increase of most resistance genes during specific seasons. However, I found that precipitation and Enterococcus levels may be accurate indicators for total AR gene levels in Newport beach coastal water. Mostly marine taxa carry AR genes in Newport Beach coastal water, but there are also terrestrial taxa and opportunistic pathogens harboring AR genes. Non-marine taxa may be washed in with rain, people, or sewage spills. By using metagenomics, I was able to elucidate the AR gene reservoir in Newport Beach coastal water.