Browsing by Subject "Microbiology"

An essential property of microbial communities is the ability to survive a disturbance. This is readily observed in bacteria, which have developed the ability to survive every antibiotic treatment at an alarming rate, considering the timescale at which new antibiotics are developed. Thus, there is a critical need to use antibiotics more effectively, extend the shelf life of existing antibiotics and minimize their side effects. This requires understanding the mechanisms underlying bacterial drug responses. Past studies have focused on survival in the presence of antibiotics by individual cells, such as genetic mutants. Also important, however, is the fact that a population of bacterial cells can collectively survive antibiotic treatments lethal to individual cells. This tolerance can arise by diverse mechanisms, including resistance-conferring enzyme production, titration-mediated bistable growth inhibition, swarming and interpopulation interactions. These strategies can enable rapid population recovery after antibiotic treatment and provide a time window during which otherwise susceptible bacteria can acquire inheritable genetic resistance.

To further explore bacterial antibiotic responses, I focused on bacteria producing β-lactamase, an enzyme that has drastically limited the use of our most commonly prescribed antibiotics: β-lactams. Through the characterization of clinical isolates and a computational model, my Ph.D. thesis has three implications:

First, survival can be achieved through resistance, the ability to absorb effects of a disturbance without a significant change, or resilience, the ability to recover after being perturbed by a disturbance. Current practices for determining the antibiotic sensitivity of bacteria do not characterize a population as resistant and/or resilient, they only report whether the bacteria can survive the antibiotic exposure. As resistance and resilience often depend on different attributes, distinguishing between these two modes of survival could inform treatment strategies. These concepts have long been applied to the analysis of ecological systems, though their interpretations are often subject to debate. This framework readily lends itself to the dissection of the bacterial response to antibiotic treatment, where both terms can be unambiguously defined.

Second, the ability to tolerate the antibiotic treatment in the short term corresponds to resistance, which primarily depends on traits associated with individual cells. In contrast, the ability to recover after being perturbed by an antibiotic corresponds to resilience, which primarily depends on traits associated with the population.

And finally, understanding the temporal dynamics of an antibiotic response could guide the design of a dosing protocol to optimize treatment efficiency for any antibiotic-pathogen combination. Ultimately, optimized dosing protocols could allow reintroduction of a repertoire of first-line antibiotics with improved treatment outcomes and preserve last-resort antibiotics.

A multilevel high-throughput sequencing approach for identifying microbial community interactions and informing precision microbiome engineering

Although common knowledge dictates that the lichen thallus is formed solely by a fungus (mycobiont) that develops a symbiotic relationship with an alga and/or cyanobacterium (photobiont), the non-photoautotrophic bacteria found in lichen microbiomes are increasingly regarded as integral components of lichen thalli and significant players in the ecology and physiology of lichens. Despite recent interest in this topic, the phylogeny, ecology, and function of these bacteria remain largely unknown. The experiments presented in this dissertation employ culture-free methods to examine the bacteria housed in these unique environments to ultimately inform an assessment of their status with regard to the lichen symbiosis. Microbiotic surveys of lichen thalli using new oligonucleotide-primers targeting the 16S SSU rRNA gene (developed as part of this study to target Bacteria, but exclude sequences derived from chloroplasts and Cyanobacteria) revealed the identity of diverse bacterial associates, including members of an undescribed lineage in the order Rhizobiales (Lichen-Associated Rhizobiales 1; `LAR1'). It is shown that the LAR1 bacterial lineage, uniquely associated with lichen thalli, is widespread among lichens formed by distantly related lichen-forming fungi and is found in lichens collected from the tropics to the arctic. Through extensive molecular cloning of the 16S rRNA gene and 454 16S amplicon sequencing, ecological trends were inferred based on mycobiont, photobiont, and geography. The implications for using lichens as microcosms to study larger principles of ecology and evolution are discussed. In addition to phylogenetic and ecological studies of lichen-associated bacterial communities, this dissertation provides a first assessment of the functions performed by these bacteria within the lichen microbiome in nature through 454 sequencing of two different lichen metatranscriptomes (one from a chlorolichen, Cladonia grayi, and one from a cyanolichen, Peltigera praetextata). Non-photobiont bacterial genes for nitrogen fixation were not detected in the Cladonia thallus (even though transcripts of cyanobacterial nitrogen fixation genes from two different pathways were detected in the cyanolichen thallus), implying that the role of nitrogen fixation in the maintenance of chlorolichens might have previously been overstated. Additionally, bacterial polyol dehydrogenases were found to be expressed in chlorolichen thalli (along with fungal polyol dehydrogenases and kinases from the mycobiont), suggesting the potential for bacteria to begin the process of breaking down the fixed carbon compounds secreted by the photobiont for easier metabolism by the mycobiont. This first look at the group of functional genes expressed at the level of transcription provides initial insights into the symbiotic network of interacting genes within the lichen microbiome.

Programmed death is often associated with a bacterial stress response. This behavior appears paradoxical, as it offers no benefit to the individual. This paradox can be explained if the death is `altruistic': the sacrifice of some cells can benefit the survivors through release of `public goods'. However, the conditions where bacterial programmed death becomes advantageous have not been unambiguously demonstrated experimentally. Here, I determined such conditions by engineering tunable, stress-induced altruistic death in the bacterium Escherichia coli. Using a mathematical model, we predicted the existence of an optimal programmed death rate that maximizes population growth under stress. I further predicted that altruistic death could generate the `Eagle effect', a counter-intuitive phenomenon where bacteria appear to grow better when treated with higher antibiotic concentrations. In support of these modeling insights, I experimentally demonstrated both the optimality in programmed death rate and the Eagle effect using our engineered system. These findings fill a critical conceptual gap in the analysis of the evolution of bacterial programmed death, and have implications for a design of antibiotic treatment.

Nontypeable Haemophilus influenzae (NTHi) is a common commensal in the human nasopharynx that can cause localized respiratory tract diseases such as otitis media, bronchitis, and pneumonia. NTHi adheres to respiratory epithelial cells, a critical step in the process of colonization enabled by bacterial surface adhesive structures called adhesins. One group of NTHi adhesins are autotransporters, proteins that have an N-terminal signal sequence, a C-terminal β-barrel domain, and an internal passenger domain with effector function. The goal of this work was to increase our understanding of two NTHi autotransporters, Hap and Hia.

Hap is a monomeric autotransporter that mediates adherence to epithelial cells and extracellular matrix (ECM) proteins. Hap also self-associates with protein on neighboring bacteria, resulting in bacterial aggregation and microcolony formation. The Hap passenger domain contains the regions responsible for adhesive activity. To define the molecular mechanism of Hap adhesive activity, we crystallized the Hap passenger domain. Characterization of the crystal structure revealed an N-terminal globular domain and a more ordered, prism-like C-terminal domain. Interestingly, Hap crystallized as a multimer, suggesting that Hap-Hap interactions occurred in the passenger domain. Progressive deletions of the β-loops that comprise the C-terminal region disrupted Hap-Hap interactions and led to a defect in bacterial settling. To further support that the C-terminal domain was responsible for Hap-Hap interactions,

7

we purified the wild type and truncated passenger domains and conjugated the proteins to latex beads. By light microscopy we visualized bead aggregation when the wild type passenger domain was conjugated to the beads, but not when the truncated passenger domain was conjugated. These results show that the C-terminal portion of the Hap passenger domain is responsible for Hap-Hap interactions leading to multimerization. Hap multimerization could be important in microcolony formation that leads to biofilm formation in vivo.

The ECM binding domain in located in the final 511 amino acids of the Hap passenger domain. To pin-point the region of the ECM protein fibronectin that is recognized by Hap, we spotted small fragments of fibronectin onto nitrocellulose membranes and incubated the membrane with purified Hap passenger domain. Far Western analysis using Hap antibody revealed that the smallest fibronectin region necessary for binding was comprised of the first two type III repeats, FNIII(1-2). To define the regions of Hap responsible for interaction with fibronectin, we mutated motifs in the Hap passenger domain that are important for fibronectin binding in other bacterial proteins. Based on assessment by ELISA, many of the mutations located between amino acids 525-725 caused reduced bacterial binding to fibronectin. However, no mutation totally ablated binding, suggesting that a larger Hap region is involved in fibronectin binding.

8

In an additional study, we identified a relationship between Hap levels in the outer membrane and the expression of lipopolysaccharide (LPS) biosynthesis enzymes. Through Western and qPCR analysis, we found that mutation of the rfaF, pgmB, lgtC, kfiC, orfE, rfbP, lsgB and lsgD genes involved in the synthesis of LPS oligosaccharide core in H. influenzae strain Rd/HapS243A resulted in loss of Hap in the bacterial outer membrane and a decrease in hap transcript. In contrast, the same mutations had no effect on outer membrane localization of H. influenzae P5 and IgA1 protease or levels of the p5 or iga1 transcripts, suggesting a Hap-specific effect. Elimination of the HtrA periplasmic protease resulted in a return of Hap to the outer membrane and restoration of wild type levels of hap transcript. We speculate that the lack of certain LPS biosynthesis enzymes causes Hap to mislocalize and accumulate in the periplasm, where it is degraded by HtrA. This degradation then leads to a decrease in hap transcript. lgtC is one of several phase variable LPS biosynthesis genes. Using an antibody against the epitope formed in part by the lgtC gene product, we identified lgtC phase-off bacteria by Western analysis of colony blots. Consistent with our previous observations, in lgtC phase off bacteria Hap was absent from the outer membrane and hap transcript was reduced. By analyzing a lgtC/lic2A double mutant, we found that Hap localization in the outer membrane and hap transcript levels were not related to LPS size but instead to the functions of the LPS synthesis enzymes themselves. This relationship could be beneficial to bacteria in vivo as a way to regulate Hap expression.

9

Early models suggested that autotransporters do not require accessory factors for folding and OM insertion. However, mounting recent evidence has suggested that the Bam complex is required for OM localization of most β-barrel proteins, including autotransporters. We studied the role of the Bam complex in OM localization of the trimeric autotransporter Hia. We expressed Hia in E. coli strains with mutations in the Bam complex and found that BamA and BamD were needed for Hia localization, while BamB, BamC, and BamE were not necessary. In further studies, we mutated the C-terminus of Hia and found that the final and third-to-last amino acids were the most important for outer membrane localization.

In summary, this work provides insights into the regulation and adhesive activity of Hap and the outer membrane localization of Hia. We have learned important details about these factors that shed light on aspects of H. influenzae disease and could lead to new antimicrobial therapies.

Background: Recent epidemiological studies demonstrate that the prevalence of the fifth major human malaria parasite, Plasmodium knowlesi (monkey malaria), is often underestimated and misdiagnosed with standard microscopy blood film. We sought to adapt and compare a new simple molecular diagnostic method for P. knowlesi with the gold standard nested molecular assay and microscopy blood film in P. knowlesi hotspot areas in Sarawak, Malaysia. In addition, we analyzed the statistical association between P. knowlesi positive test results and demographic and behavioral/occupational risk factors.

Methods: The study was conducted at Sibu, Kapit and Sarikei Hospitals in Sarawak, Malaysia. 115 blood samples were collected from malaria suspected patients seeking treatment at these hospitals. Samples were analyzed by microscopy, Nested polymerase chain reaction (PCR) and single-step PCR. Sensitivity, specificity, and practical value of the new single-step PCR assay was calculated. Bivariate and multivariate regression was conducted to test the possible risk factors for the detection of P. knowlesi.

Results: Single-step PCR showed low sensitivity (51.92%, 95%CI 37.63 - 65.99%) compared to nested PCR and 46.03% (95%CI 33.39 - 59.06%) compared to microscopy. When compared to nested PCR, microscopy had a false positive rate of 20.6%. However, it only missed 2 cases of P. knowlesi. The mean age in the study population was 40.35. Patients enrolled at Kapit hospital had higher odds ratio for positive P. knowlesi PCR results (adjusted OR = 4.46, 95%CI 1.16 – 11.51). Age above 21 years (adjusted OR = 6.28, 95%CI 1.53 – 25.64), male gender (adjusted OR = 2.46, 95%CI 0.91 – 6.65) and living near a vegetation (Plantation, forest, fruit trees or wet rice paddy) (adjusted OR = 5.96, 95%CI 1.11 – 31.83) were associated with increased risk for P. knowlesi infection.

Conclusions: Data from this study showed that single-step PCR has a low sensitivity and thus, it is not a suitable alternative for accurate detection of P. knowlesi. Further studies are required for assessment and development of other diagnostic assays or new primer sets. Multivariate analysis revealed that adult men over the age of 21 who live near agricultural areas have the highest risk for P. knowlesi malaria infection. Large- scale descriptive studies of both non-human hosts and vectors would greatly influence prevention and control strategies of this zoonotic disease.

The 2010 Deepwater Horizon disaster remains one of the largest oil spills in history. This event caused significant damage to coastal ecosystems, the full extent of which has yet to be fully determined. Crude oil contains both toxic substances that are detrimental to microbes and compounds that may be used as food and energy resources by some microbial species. As a result, oil spills have the potential to cause significant shifts in microbial communities. In this study, we assessed the impact of oil contamination on the function of endophytic microbial communities associated with saltmarsh cordgrass (Spartina alterniflora). Soil samples were collected from two locations in coastal Louisiana, USA: one severely affected by contamination from the Deepwater Horizon oil spill and one relatively unaffected location. Spartina alterniflora seedlings were grown in both soil samples under greenhouse conditions, and GeoChip 5.0 was used to evaluate the endophytic microbial metatranscriptome shifts in response to host oil exposure. Microbial functional shifts were detected in functional categories related to metal homeostasis, organic remediation, and phosphorus utilization. These findings show that host oil exposure elicits multiple changes in metabolic response from their endophytic microbial communities, producing effects that may have the potential to impact host plant fitness.

Pseudomonas aeruginosa, known as one of the leading causes of morbidity and mortality in cystic fibrosis (CF) patients, secretes a variety of virulence-associated proteases. These enzymes have been shown to contribute significantly to P. aeruginosa pathogenesis and biofilm formation in the chronic colonization of CF patient lungs, as well as playing a role in infections of the cornea, burn wounds and chronic wounds. Our lab has previously characterized a secreted P. aeruginosa peptidase, PaAP, that is highly expressed in chronic CF isolates. This leucine aminopeptidase is also highly expressed during infection and in biofilms, and it associates with bacterial outer membrane vesicles (OMVs), structures known for their contribution to virulence mechanisms in a variety of Gram-negative species and one of the major components of the biofilm matrix. With this in mind, we hypothesized that PaAP may play a role in P. aeruginosa biofilm formation. Using a lung epithelial cell/bacterial biofilm coculture model, we show that PaAP deletion in a clinical P. aeruginosa background alters biofilm microcolony composition to increase cellular density, while decreasing matrix polysaccharide content and resistance to the antibiotic colistin. We recreate this phenotype using a pellicle biofilm model, in which bacteria are grown statically at the culture air-liquid interface, demonstrating that these phenotypes are not dependent on the coculture host cell substrate. We additionally show that OMVs from PaAP expressing strains, but not PaAP alone or in combination with PaAP deletion strain-derived OMVs, could complement this phenotype. Finally, we found that OMVs from PaAP-expressing strains cause protease-mediated biofilm detachment, leading to changes in matrix and colony composition. OMVs mediated the detachment of biofilms formed by both non-self P. aeruginosa strains and K. pneumoniae, another respiratory pathogen, showing that this process may also be relevant in polymicrobial communities and acts on non-P. aeruginosa derived substrates. Our findings represent novel roles for OMVs and the PaAP aminopeptidase in the modulation of bacterial biofilm architecture.

Background: In recent years, emerging zoonotic microbes are gaining more attention from the public and policy makers. Explosive outbreaks such as those from avian influenza viruses, severe acute respiratory syndrome (SARS) virus, swine influenza viruses, Hendra virus, Nipah virus, and Middle East respiratory syndrome (MERS) have had tremendous international economic and social impact. In particular, livestock workers have been found to be at increased infection risk and some of the first impacted by a novel pathogen. One of the main obstacles in averting outbreaks of novel microbes is detecting it when it first begins to cross species from animals to man and may not cause severe disease. Often routine diagnostics will fail to detect a new pathogen. The purpose of this research was to evaluate diagnostics for emerging coronaviruses that would be missed with routine diagnostics.

Methods: In 2016, I learned how to run new diagnostics adapted at Duke University to detect novel coronaviruses. I took this molecular technology to Sarawak, Malaysia, where I applied the assays against a panel of human clinical specimens from patients seen at three hospitals for respiratory illnesses. Our collaborators in Sarawak had previously examined these specimens with other assays against human coronaviruses but did not tell me of their results.

Results: In my hands, the new pan-species coronavirus assay detected only one coronavirus among 88 clinical specimens. After I finished my assay work, I learned from our collaborators that 27of the 88 specimens had been positive for at least one previously recognized human coronavirus. Hence, the sensitivity of the new assay in my hands found to be 3.70% (95% confidence interval 0% - 11.91%). However, the assay accurately showed negative results with a specificity of 100%

Conclusion: While this low sensitivity may have been real, it may also been influenced by a number of confounding factors such as specimen nucleic acid degradation with numerous freeze-thaw cycles, imprecise adaptation of an assay to new equipment in a new laboratory, or my or our collaborators’ operator error. It is difficult to precisely identify the cause of the discordance. Nevertheless, I learned a great deal about global health in conducting this research in Sarawak and have chronicled some of these lessons in this report.

Pyroptosis and apoptosis are two forms of regulated cell death that can defend against intracellular infection. Although pyroptosis and apoptosis have distinct signaling pathways, when a cell fails to complete pyroptosis, backup pathways will initiate apoptosis. Here, we investigated the utility of apoptosis compared to pyroptosis in defense against an intracellular bacterial infection. We previously engineered Salmonella enterica serovar Typhimurium to persistently express flagellin, and thereby activate NLRC4 during systemic infection in mice. The resulting pyroptosis clears this flagellin-engineered strain. We now show that infection of caspase-1 or gasdermin D deficient macrophages by this flagellin-engineered S. Typhimurium induces apoptosis in vitro. Additionally, we also now engineer S. Typhimurium to translocate the pro-apoptotic BH3 domain of BID, which also triggers apoptosis in macrophages in vitro. In both engineered strains, apoptosis occurred somewhat slower than pyroptosis. During mouse infection, the apoptotic pathway successfully cleared these engineered S. Typhimurium from the intestinal niche, but failed to clear the bacteria from the myeloid niche in the spleen or lymph nodes. In contrast, the pyroptotic pathway was beneficial in defense of both niches. In order to clear an infection, distinct cell types may have specific tasks that they must complete before they die. In some cells, either apoptotic or pyroptotic signaling may initiate the same tasks, whereas in other cell types these modes of cell death may lead to different tasks that may not be identical in defense against infection. We recently suggested that such diverse tasks can be considered as different cellular “bucket lists” to be accomplished before a cell dies. As demonstrated here, engineering pathogens is a useful method for discovering new details of microbial pathogenesis and host defense. However, engineering can result in off-target effects. We engineer S. Typhimurium to overexpress the secretion signal of the type 3 secretion system effector SspH1 fused with domains of other proteins as cargo. Such engineering had no virulence cost to the bacteria for the first 48 hours post infection in mice. However, after 48 hours the engineered bacteria manifest an attenuation that correlates with the quantity of the SspH1 translocation signal expressed. In IFNg-deficient mice this attenuation was weakened. Conversely, the attenuation was accelerated in the context of a pre-existing infection. We speculate that inflammatory signals change aspects of the target cell’s physiology that make host cells less permissive to S. Typhimurium infection. This increased degree of difficulty requires the bacteria to utilize its T3SS at peak efficiency, which can be disrupted by engineered effectors.

Bacteria communicate, coordinate, and cooperate as a population and this `social' behavior is key to their proliferation. Quorum sensing (QS) is the cell-cell communication mechanism by which bacteria sense their population density and modulate their target gene expression accordingly. While QS is ubiquitous among bacteria, there is tremendous diversity in terms of the sensory elements used and the biochemical and transport properties of signaling molecules. Further, the targets of QS include a wide range of cooperative actions, such as the secretion of enzymes for nutrient foraging, virulence toxins, and biofilm-forming compounds. Here I investigate what role QS and cooperation play, as universal social characteristics, in promoting bacterial proliferation.

Engineered biological circuits offer the potential to test our understanding of natural systems under well-defined contexts, by focusing on the key characteristics and components of interest. In my doctoral work, I have taken advantage of this methodology to study bacterial social behavior. Combining mathematical modeling with quantitative experiments using gene circuits, my research has (1) elucidated the `core' components of cell-cell communication across bacteria, (2) explained how communication and cooperation advantage bacterial growth, and (3) opened up the important application of this research in generating novel antibacterial therapies.

Cells from all levels of life secrete vesicles, which are nanoscale proteoliposomes packaged with a variety of proteins, lipids, and small molecule cargo. Depending on their origin, these extracellular vesicles are termed exosomes, microvesicles, exomeres, and membrane vesicles, to list a few. Vesicles released from Gram-negative bacteria bud from the outer membrane and are, therefore, referred to as outer membrane vesicles (OMVs). In mammalian systems, OMVs facilitate bacterial survival by alleviating membrane stress, serving as a decoy for bacteriophage and antibiotics, and providing a fast membrane remodeling mechanism. OMVs also contribute to virulence by delivering toxins and other soluble and insoluble cargo to the host cell. The role OMVs play in plant systems remains unknown.

Previous studies revealed that plant pathogenic bacterial vesicles contain virulence factors, type III secretion system effectors, plant cell wall-degrading enzymes, and more, suggesting that vesicles may play similar roles to those from mammalian pathogens in host-pathogen interactions. Further, OMVs elicit several markers for pathogen-associated molecular pattern triggered immunity in plants. These responses include increased transcription of defense markers such as FRK1 and production of reactive oxygen species. Building on these findings, here we show that OMVs from the plant pathogen Pseudomonas syringae and the plant beneficial Pseudomonas fluorescens elicit plant immune responses in Arabidopsis thaliana that protect against future pathogen challenge. Intriguingly, protection is independent of salicylic acid plant defense pathways and bacterial type III secretion. OMVs also inhibit seedling growth, another indication of plant immune activation.

Our initial biochemical studies suggested that the immunogenic OMV cargo was larger than 10 kDa and differed between the pathogen and beneficial species despite similar plant immunity outcomes. Interestingly, protective OMV-mediated responses were protein-independent, while the seedling growth inhibition phenotype was entirely protein dependent. Proteomics analysis confirmed that OMV protein cargo differed between P. syringae and P. fluorescens. While media culture conditions did not dramatically impact the immunogenic activity of isolated OMVs from either species, proteomics analysis revealed a significant shift in P. syringae OMV cargo between complete and minimal media conditions. P. fluorescens OMV cargo was largely the same in the two media conditions, with no significantly enriched proteins in minimal or complete media. Further analysis of the proteins enriched in the P. syringae minimal OMV condition identified one set of proteins with the same baseline abundance in P. syringae and P. fluorescens complete OMVs and another set with a lower baseline abundance compared to P. fluorescens OMVs. These two subsets could contribute to virulence and stress tolerance, respectively. Enrichment analysis uncovered particularly interesting protein categories in the subset with the same baseline abundance. Of interest, several lipoprotein and lipid binding categories were enriched, and proteins involved in synthesis of the phytotoxin coronatine were also enriched in this same-baseline subset. These results support our hypothesis that proteins enriched in P. syringae minimal OMVs with the same baseline abundance in P. fluorescens complete OMVs may contribute to OMV-mediated bacterial virulence in plants. Our findings also suggest that our forthcoming OMV metabolomic analyses may reveal non-proteinaceous cargo that is critical for OMV-mediated plant immune activation.

The work presented here lays the groundwork for future exploration of OMV-plant interactions and adds a new layer of complexity to plant-bacteria interactions. Further, these results reveal that OMVs elicit complex plant immune responses that would be difficult for pathogens to adapt to and overcome, supporting a role for bacterial OMVs in agricultural applications to promote durable resistance and revealing a new potential avenue for disease prevention and management.

The antimicrobial properties of silver have been taken advantage of by societies for thousands of years. Its use has come back in favor in the form of silver nanoparticles, which are highly efficacious antimicrobial agents. Silver nanoparticles are incorporated into a myriad of products specifically designed for clinical use, but also for general use by consumers. Silver nanoparticles can be found in textiles such as clothing and stuffed toys, and in home appliances including washing machines and curling irons. A large number of products specifically marketed for use by children are also available to consumers, including pacifiers, sippy cups, and even breast milk storage bags. The hazards and toxicities associated with silver nanoparticles are not well understood, however modes of toxicity have been reported for ionic silver. It is assumed that the main mechanism of toxicity of silver nanoparticles relates to the release of ionic silver, however studies have indicated an additional nano-effect, likely due to nanoparticle size, differential coatings, and means of sustained dosing of ionic silver. However we are sure that these silver nanoparticles will accumulate in the waste stream, likely arriving during different stages of a product’s lifespan. A main sink of these nanoparticles travelling through both natural and engineered environments is wastewater treatment plants. As a society we rely on the biological removal of nutrients, which takes place primarily in the activated sludge of secondary treatment. Studies have already indicated possible, temporary decreases in removal efficiencies as well as changes in microbial communities, including losses of diversity, following exposure to silver nanoparticles. Therefore, it is of paramount importance to examine the effects of both silver nanoparticles and ionic silver on the community and function of wastewater bacteria.

Sequencing batch reactors were operated to mimic wastewater treatment. They were fed synthetic wastewater and after reaching acclimation, were dosed over time with varying concentrations of both ionic and nanosilver. Cell samples were collected periodically to assess the presence and identity of cultivable silver resistant bacteria and to map the microbial community changes taking place under different treatments using Next Generation Sequencing. Isolates were tested for the presence of known silver resistance (sil) genes as were activated sludge samples from a collection of domestic wastewater treatment plants, by designing TaqMan probe assays and performing quantitative PCR. The silver resistant isolates were also used to test the growth implications, as well as sil gene expression changes, following treatment with ionic silver and a variety of silver nanoparticles with various coatings, all at multiple concentrations. This was accomplished by performing multiple batch experiments and then using the TaqMan assays and reverse transcription-quantitative PCR.

Overall, microbial community changes were observed in the sequencing batch reactors, and there were differences noted based on treatment, including ionic silver versus nanosilver and between the two silver nanoparticle coatings. Most notably, the possibility of nitrification in wastewater treatment being particularly susceptible was strongly indicated. Individual wastewater bacteria isolates all contained sil genes, as did the majority of the wastewater treatment plant activated sludge, however the levels of actual sil gene expression were inconsistent. This particular finding supports a current body of work indicating that there are alternate modes of bacterial silver resistance in play that we are just becoming aware of.

Microbial population growth measurements are widespread in the study of microorganisms, providing insight into areas including genetics, physiology, and engineering. The most common models of microbial population growth data are parametric, and are derived from specific assumptions about the underlying growth process. While useful in cases where these assumptions are valid, these models are inadequate in many cases typically found in microbial growth studies, including presence of significant population death and the presence of multiple growth phases (e.g. diauxie). Here, we explore the use of the Bayesian non-parametric model Gaussian processes on microbial population growth. We first develop a general hypothesis-test using Gaussian process regression and false-discovery rate corrected Bayes factor scores. We then explore a fully Bayesian model with Gaussian process priors that can capture the latent growth processes of many population measurements under a single model. Finally, we develop hierarchical Bayesian model with GP priors in order to capture random effects in microbial population growth data.

Advances in high-throughput DNA sequencing allow for rapid and affordable surveys of thousands of bacterial taxa across thousands of samples. The exploding availability of sequencing data has poised microbiota research to advance our understanding of fields as diverse as ecology, evolution, medicine, and agriculture. Yet, while microbiota data is now ubiquitous, methods for the analysis of such data remain underdeveloped. This gap reflects the challenge of analyzing sparse high-dimensional count data that contains compositional (relative abundance) information. To address these challenges this dissertation introduces a number of tools for Bayesian inference applied to microbiome data. A central theme throughout this work is the use of multinomial logistic-normal models which are found to concisely address these challenges. In particular, the connection between the logistic-normal distribution and the Aitchison geometry of the simplex is commonly used to develop interpretable tools for the analysis of microbiome data.

The structure of this dissertation is as follows. Chapter 1 introduces key challenges in the analysis of microbiome data. Chapter 2 introduces a novel log-ratio transform between the simplex and Real space to enable the development of statistical tools for compositional data with phylogenetic structure. Chapter 3 introduces a multinomial logistic-normal generalized dynamic linear modelling framework for analysis of microbiome time-series data. Chapter 4 explores the analysis of zero values in sequence count data from a stochastic process perspective and demonstrates that zero-inflated models often produce counter-intuitive results in this this regime. Finally, Chapter 5 introduces the theory of Marginally Latent Matrix-T Processes as a means of developing efficient accurate inference for a large class of both multinomial logistic-normal models including linear regression, non-linear regression, and dynamic linear models. Notably, the inference schemes developed in Chapter 5 are found to often be orders of magnitude faster than Hamiltonian Monte Carlo without sacrificing accuracy in point estimation or uncertainty quantification.

This dissertation presents a framework to improve bioremediation of soils polluted with polycyclic aromatic hydrocarbons (PAHs). PAHs are of great concern because they are recalcitrant and toxic. PAHs enter the environment from a variety of sources such as incomplete combustion and coal tar distillation. The PAHs focused on in this dissertation have polluted soils as a result of creosote-based wood treatment operations that took place at Holcomb Creosote and Atlantic Wood Industries, Inc. (AWI) both of which are now classified as Superfund sites. There are numerous sites analogous to these two Superfund sites throughout the world which have been polluted through similar wood-treatment operations, as creosote was once industry’s foremost wood preservative.

There is room for existing PAH treatment options, which are mainly physical and chemical in nature, to be expanded to include more sustainable options. Commonly used technologies include excavation, in situ stabilization, and soil washing. Historically, bioremediation strategies relying on bacteria to transform pollutants have been challenged by the tight sorption of heavy- and middle-weight PAHs to soils, as this restricts aqueous phase transport required for bacterial degradation. Multiple studies have demonstrated fungi to be capable of degrading these inaccessible pollutants and other mixtures of hydrophobic pollutants (mycoremediation). Yet, when fungi have been introduced to polluted soils (mycoaugmentation), they have not been able to outcompete the native microbiota long enough to degrade the contaminants of concern over the long term. It is possible that a thorough characterization of the indigenous fungi at a given site may provide some insights into the development of targeted in situ mycoremediation strategies.

Although incorporating site microbes has been generally acknowledged as important for some time, the techniques enabling thorough assessment of microbial ecosystems are relatively new. Consequently, little is known about PAH-associated microbiomes in general, and even less is known about PAH-associated fungal communities. The work presented in this dissertation aims to address this knowledge gap by leveraging recent advances in high-throughput sequencing technology to design and validate targeted mycoremediation strategies. To this end, the overarching goal of this dissertation was to develop and test a framework for incorporating native fungi into a bioremediation strategy to expand such sustainable remediation options to sites where they have not been relevant in the past.

In the first aim of this dissertation research, advances in high-throughput sequencing were used to identify potential biostimulation targets in soils moderately polluted with PAHs. The next generation sequencing (NGS) platform, Illumina, was utilized to sequence the large sub-unit (LSU) gene commonly used as a marker gene in fungal community studies. Relationships were examined between concentrations of over 31 different polycyclic aromatic hydrocarbons and the pollutant-associated communities to test whether there were any fungi capable of tolerating high levels of these toxic contaminants. In this aim, fungal genera were identified that contained species closely related to known PAHs and petroleum hydrocarbon degraders. In all, this work identified 32 targets for biostimulation, based on Spearman rank correlations between prevalence and mid- and high-molecular weight PAHs. Ascomycetes were found to have higher levels of diversity than any other phylum in this subset of biostimulation targets. These data suggest that ascomycete fungi are more likely to be present in heavily polluted soils than basidiomycete fungi (which had previously been subjects of much interest). Overall, this work illustrates that polluted soils harbor fungal biostimulation targets, specifically within Ascomycota.

The second aim of this thesis research was to use the precision bioremediation assessment in highly polluted soils and then to evaluate a range of amendments with the goal of identifying strategies to stimulate the fungal communities that dominate these PAH-associated fungal communities. Here we applied the approach we fine-tuned in the first aim to the AWI soils, as these soils have some of the highest documented PAH-concentrations. Again, Ascomycota were found to be more prevalent in these soils, so an isolate obtained from AWI was used to compare alternative stimulation techniques between three substrates they are known to grow on: chitin, cellulose, and wood. We used anthracene degradation as a proxy for PAH degradation, which we monitored in sacrificial simplified bioreactors responding to the three amendments. T. harzianum is also known to have enzymes which degrade PAHs, but it is unknown which ecological role uses those enzymes, and thus which ecological role we should promote. T. harzianum was grown in the presence of chitin, cellulose, and wood as substrates in liquid culture with anthracene. Chitin was found to stimulate the highest anthracene removal, with a 0.1% (w/v) amendment resulting in ~93% degradation. While ~13% less than chitin, 1% (w/v) cellulose was also found to stimulate ~46% more anthracene degradation than wood, which had no improvement over the abiotic losses (~33% on average). This is notable because the “go to” method for stimulating fungi in the past has been wood supplements. This work provided insight into alternative stimulation strategies to target specific ecological roles that may better degrade PAHs in situ.

For the third and final aim of this dissertation research, the two most promising amendments were added with and without Trichoderma harzianum spores to test several mycoremediation treatment strategies in soil bioreactors and compare them with a (no carbon added) nutrient stimulation treatment. Pollutants were added as aged Atlantic Wood Industries soil delivering aged pollutants. Triplicate reactors from each treatment were sequenced at time zero, after two weeks, and after one month. At each sampling time, RNA was extracted, converted to cDNA, and submitted to Illumina MiSeq library preparation targeting the LSU region for fungal community analysis in addition to the V4 region of the 16S rDNA for bacterial community analysis. Statistical analyses using DESeq2 identified responders among the groups of reactors subjected to the different biostimulation treatments. Taxa from both the fungal and the bacterial communities responded differentially to the amendments. Fungi were found to comprise the majority of the significant responders. This work also found that mycoaugmented strains were not successful in establishing themselves as prominent members of the active community. This represents one of the earliest studies to directly measure mycoaugmentation failure. These data propose a hypothesis about functional redundancy inhibiting establishment of augmented fungi as already established fungi outcompete them for freshly added nutrients. Over 90% degradation was observed over the course of one month regardless of treatment-interestingly, the highest degradation was found in the nutrient amendment (no carbon added) treatment. These results show similar degradation across the soil bioreactors, yet different microbial growth, which supports the hypothesis that there is community-level functional redundancy and multiple metabolic food webs that result in the observed pollutant degradation.

Overall, this dissertation work demonstrates how significant advances in sequencing technology can be implemented in design and monitoring stages of bioremediation. This work also suggests that significant advances could be possible through the application of targeted metatranscriptomic analysis. Through incorporating such insights as described in this dissertation, this research brings the field of bioremediation one step closer to successfully engineering microbiomes to degrade contaminants of concern.

Human welfare depends heavily on ecosystem services like water purification and nutrient cycling. Many of these ecosystem services, in turn, rely on reactions performed by microbes and yet remarkably little is known about how anthropogenic impacts are affecting the structure and function of microbial communities. To help address this knowledge gap, this dissertation uses field surveys and laboratory experiments to examine how watershed urbanization affects microbial communities in receiving streams. We focus on a specific functional group and its associated function - the denitrifying bacteria and denitrification. Denitrifying bacteria use reactive nitrogen and organic carbon as substrates to perform denitrification. Denitrification is one of the few ways to permanently remove reactive nitrogen from ecosystems. Since excess reactive nitrogen in water contributes to serious water quality and human health problems like toxic algal blooms and bowel cancer, denitrification in streams can be considered a valuable ecosystem service. Watershed urbanization, however, may alter the structure of denitrifying bacteria communities in ways that constrain their capacity to remove reactive nitrogen from streams.

Watershed urbanization leads to drastic changes in receiving streams, with urban streams receiving a high frequency of scouring flows, together with increased nutrient (nitrogen and carbon), contaminant (e.g., heavy metals), and thermal pollution. These changes are known to cause significant losses of sensitive insect and fish species from urban streams. Microbes like denitrifying bacteria may be similarly affected. In the first part of this dissertation, we describe results from four repeated surveys of eight central North Carolina streams affected to varying degrees by watershed urbanization. For each stream and sampling date, we characterized both overall and denitrifying bacterial communities and measured denitrification potentials. Differences in overall and denitrifying bacteria community composition were strongly associated with the urbanization gradient. Denitrification potentials, which varied widely, were not significantly associated with substrate supply. By incorporating information on the community composition of denitrifying bacteria together with substrate supply in a linear mixed-effects model, we explained 45% of the variation in denitrification potential (p < 0.001). Results suggest that 1) watershed urbanization can lead to significant changes in the composition of bacterial communities in streams and 2) such changes may have important functional consequences.

The second part of this dissertation examines how urbanization-driven changes to the structure of denitrifying bacteria communities might affect the way they respond to stress or disturbance. Some communities can resist changes to functionality in response to disturbance, potentially as a result of previous exposure and subsequent adaptation (legacy hypothesis) or high diversity (insurance hypothesis). We compare the resistance of two structurally distinct denitrifying bacteria communities to experimental disturbances in laboratory microcosms. Communities originated from either a polluted, warm urban streams or a relatively pristine, cool forest stream. In this case, the two communities had comparable compositions, but forest communities were more diverse than their urban counterparts. Urban communities experienced significant reductions in denitrification rates in response to the most severe increased pollution and temperature treatments, while forest communities were unaffected by those same treatments. These findings support the insurance, but not the legacy hypothesis and suggest that the functioning of urban streams may be more susceptible to further environmental degradation than forest streams not heavily impacted by human activities.

In the third part of this dissertation, we discuss results from a one-time survey of denitrifying bacteria communities and denitrification potentials in 49 central North Carolina streams affected to varying degrees by watershed urbanization. We use multivariate statistics and structural equation modeling to address two key questions: 1) How do different urban impacts affect the structure of denitrifying bacteria communities and 2) How do abiotic (e.g., temperature) versus biotic (denitrifying bacteria community structure) factors affect denitrification potentials in urban streams? Denitrifying bacteria community structure was strongly affected by the urban impacts measured. Community composition responded to increased temperatures, substrate supply, and contamination, while diversity responded negatively to increased temperatures and hydrologic disturbance. Moreover, increased temperatures and substrate supply had significant positive effects, while urbanization-driven changes to denitrifying bacteria community structure had significant negative effects on denitrification potential. The structural equation model captured 63% of the variation in denitrification potential among sites and highlighted the important role that microbial community structure can play in regulating ecosystem functioning. These findings provide a novel explanation for recent observations of decreasing denitrification efficiency with increasing urbanization. Ultimately, we hope findings from this dissertation will help inform more effective stream management and restoration plans and motivate ecologists to consider including microbial community structure in ecosystem models of microbe-mediated processes.

Mycobacterial infection leads to the formation of characteristic immune cell aggregates called granulomas. In humans and animal models, tuberculous granuloma formation is accompanied by dramatic remodeling of host vasculature which ultimately benefits the infecting mycobacteria, suggesting the bacteria may actively drive this host process. First, we sought to identify bacterial factors that promote granuloma vascularization. Using Mycobacterium marinum transposon mutants in a zebrafish infection model, we revealed the enzyme Proximal Cyclopropane Synthase of alpha-Mycolates (PcaA) as an important bacterial determinant of host angiogenesis. We found that PcaA-modified trehalose dimycolate, an abundant glycolipid in the mycobacterial cell wall, drives activation of host VEGF signaling and subsequent granuloma vascularization. To facilitate our continuing investigation of granuloma dynamics, we next sought to expand and improve upon the transgenic tools for studying macrophages in the zebrafish model. I describe two such tools: i) the macrophage-specific zebrafish mfap4 promoter, which allows long-term in vivo visualization and manipulation of macrophages during mycobacterial infection, and ii) the first zebrafish transgenic line with constitutive, ubiquitous Cas9 expression, as well as a transgene design capable of generating sgRNAs using macrophage-specific promoters. These tools allow CRISPR/Cas9 gene editing in vivo in the zebrafish in a macrophage-restricted manner.

My thesis work has focused on characterizing mechanisms by which human fungal pathogens regulate their adaptive cellular responses in order to survive and cause disease in the human host. Unlike most microbial fungi found in the environment, Cryptococcus neoformans has become a successful human pathogen due to two intrinsic abilities: 1) to survive and grow at human body temperature and 2) to employ virulence factors to combat host immune defenses. Over the past two decades, the fungal pathogenesis field has made enormous progress in identifying and characterizing C. neoformans proteins responsible for these adaptive cellular responses with a particular focus on enzymes, like those involved in cell cycle progression or those responsible for synthesizing components of the fungal cell surface. Although we know a substantial amount about the functions of these enzymes and their implications on fungal pathogenesis, the mechanisms by which these enzymes are regulated are less clear. I have attempted to address this gap in knowledge by focusing my thesis work on the identification and characterization of C. neoformans non-enzymatic proteins that regulate enzymes important for adaptive cellular responses. I have identified and characterized the C. neoformans arrestin proteins as regulators of enzyme ubiquitination, and likely enzyme function, in response to specific extracellular stressors (Chapters 2 & 3). I have also characterized a Cryptococcus-specific protein, Mar1, as an important modulator of host-fungal interactions due to its regulation of cell surface remodeling through maintenance of mitochondrial metabolic activity and homeostasis in response to cellular stress (Chapters 4 & 5). Furthermore, I also performed a comprehensive comparative analysis of different RNA enrichment methods for RNA sequencing applications and long non-coding RNA identification in C. neoformans, which can help researchers select appropriate tools for studying adaptive cellular responses from the RNA level (Chapter 6). These studies collectively have demonstrated that non-enzymatic proteins are important “cellular coordinators” in human fungal pathogens; they regulate the activity of many different enzymes in response to distinct extracellular signals, and as a result are required for both fungal growth and virulence factor employment in response to host-relevant stressors.

Pathological angiogenesis is a widespread biological phenomenon that influences the progression of various diseases, including autoimmune conditions, cancers, and microbial infections. One infection in particular, tuberculosis, is associated with the induction of a potent pro-angiogenic signaling cascade that facilitates bacterial growth and accelerates disease progression. A synthesis of early studies on bacterial factors that drive host angiogenesis with modern genetic findings identified the mycobacterial glycolipid trehalose 6-6'-dimycolate (TDM) as a critical factor driving vascular endothelial growth factor (VEGFA) production and angiogenesis during mycobacterial infection. Despite these recent findings, many of the underlying host response mechanisms remain unknown. The introductory chapter will serve to introduce the reader to the major concepts addressed in this work: Mycobacterium tuberculosis and the disease it causes, the role of macrophages in health and disease, the function of pattern recognition receptors in detecting microbial ligands, the specific downstream intracellular signaling pathway of interest for this work (mediated by the transcription factor, nuclear factor of activated T cells, NFAT), the contributions of angiogenesis to diverse contexts and pathologies, and the promise of host-directed therapies to overcome challenges associated with traditional treatment approaches in infectious disease. Chapter 2 describes the new and existing methodological approaches that were required to complete this work. This work utilizes the zebrafish-Mycobacterium marinum model of tuberculosis infection to facilitate in depth in vivo observation and quantitation of these phenomena. Using this model in tandem with human macrophage cell culture, I was able to model major aspects of the host-pathogen interface, enabling me to identify a critical role for a macrophage-C-type lectin receptor-NFATC2-VEGFA signaling axis required for the angiogenic response to mycobacterial infection and TDM, findings that comprise the core of this work and are detailed at length in Chapter 3. The analysis of the large amounts of data generated in this work required creative approaches to data processing and analysis. To this end, I have developed a set of novel processing modalities in Python and R that are capable of the rapid and reproducible processing of images as well as certain aspects of automated data collection therefrom. These macros, many written for the FIJI/ImageJ programming environment, serve as the infrastructure on which the rest of this work has been built. These will be detailed in Chapter 4. Finally, this body of work leaves many questions as yet unanswered. While it is clear that NFAT signaling is required for VEGFA production, the precise mechanism by which this may work is unclear and could be mediated by either direct DNA binding or indirect activation or cooperative binding with some other transcriptional activator. There also exist a variety of other potential NFAT- and angiogenesis-related phenotypes worthy of exploring using the tools and approaches I have developed. It is my hope that the findings herein stimulate further study on the contributions of NFAT signaling to the host immune response to mycobacterial infection and evaluation of the potential of NFAT inhibition as host-directed therapy to tuberculosis.