Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 226-238). / To date, there has been limited usage of biomass and agricultural residues in rural areas as a form of renewable energy, mainly due to the expensive costs involved in collecting and transporting raw biomass. A decentralized biomass torrefaction system has the potential to upgrade the quality and transportability of distributed biomass residues in situ, thereby creating additional localized economic values and mitigating the environmental consequences associated with open burning of the excess biomass residues. Nonetheless, most existing biomass torrefaction systems so far have been designed for large-scale, centralized deployment, and are unsuitable to be scaled down in decentralized applications due to their high level of sophistication and capital cost. We propose a biomass torrefaction system based on the concept of torrefaction in a low-oxygen environment. By eliminating the stringent requirements of an inert torrefaction environment, we demonstrated that we can greatly simplify the reactor design and derive a laboratory-scale system that is also scalable. We proceeded to build and validate this torrefaction system with respect to different operating conditions and types of biomass. Using a quantitative definition for torrefaction severity, we were also able to relate the various fuel user requirements in real life back to the fundamental reactor operations. By quantifying in detail the overall energy performance, pressure requirements, and transient timescales, we also demonstrated how such a reactor system can be operated at scale, as well as the various design improvements that can further boost the performance of a scaled-up system. Therefore, this work builds the foundation towards the development of a low-cost, small-scale, and portable torrefaction system that can potentially be widely deployed in rural areas. / by Kevin S. Kung. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Osteoarthritis (OA) is the most common form of joint disorder. Individuals who have sustained an acute traumatic joint injury are at greater risk for the development of OA. The mechanisms by which injury causes cartilage degradation are not fully understood, but the elevated levels of injury-induced pro-inflammatory cytokines, such as TNFa and IL-6, have been implicated to play important roles in the pathogenesis of OA. We have used in vitro models of cartilage injury to examine the interplay between mechanical and cytokine-mediated pathways and to identify processes associated with cartilage degradation following joint injury. The overall aims of this thesis were to characterize the combined effect of TNFa and IL-6/sIL6R on matrix degradation and chondrocyte gene expression in mechanically injured cartilage, and to investigate whether cartilage degradation could be inhibited by potential therapeutic approaches. TNFa and IL-6/sIL-6R interacted to cause aggrecanase-mediated proteoglycan degradation. Importantly, the combined catabolic effects of cytokines were highly potentiated by mechanical injury. Furthermore, cartilage degradation caused by the in vitro injury model appeared to be initiated at the transcriptional level, since the gene expression of matrix proteases, cytokines and iNOS were all highly elevated in the treatment conditions. The degradative effects of TNFa in injured cartilage was due, in part, to the action of endogenous IL-6, as proteoglycan degradation was partly reduced by an IL-6 blocking Fab fragment. Interestingly, cartilage degradation induced by the combinations of proinflammatory cytokines and mechanical injury was fully abrogated by short-term treatments with dexamethasone. The results of this work are significant in that they provide evidence suggesting joint injury affects cell-mediated responses as well as the transport of cytokines and proteases in extracellular matrix, making cartilage tissue more susceptible to further degradation by biochemical mediators. / by Yihong C.S. Lu. / Ph.D.

Thesis: Sc. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2015. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 157-192). / Unlike the genomic sequence, the epigenome is dynamic, responding to and influencing the cellular state. Understanding the ways in which chromatin structure exercises control over transcriptional programs during development as well as aberrant gene expression signatures in human disease is a central challenge to biology and pathology. Despite the advance of tools to chart the chromatin landscape, much is still unknown about how epigenetic signals are integrated in the transcriptional outcome of a gene. Here, a number of systematic, genome-wide approaches result in the discovery of epigenetic patterns associated with the modulation of gene expression. We begin by examining Huntington's disease (HD), a fatal neurodegenerative disorder. The discovery of the genetic mutation that causes HD has led to the development of in-vitro and in-vivo models representative of the disease. Studies of these systems and patient samples have identified transcriptional dysregulation as a major component of the early stages of HD. Thus, HD serves as an ideal case in which to examine the modulation of gene expression in disease. First, we connect the changes in DNA methylation to the dynamics of gene expression and transcription factor binding in HD. We next identify a unique signature of the histone mark H3K4me3 at genes transcriptionally repressed in HD. Targeting this signature reverses the downregulation of genes and protects against neurodegeneration. Finally, we apply a novel, machine learning approach to 16 chromatin features in 44 human cell types. By globally examining the epigenetic states of genes in a variety of tissue contexts, we discover a small set of coordinated patterns that we term "epigenetic ensembles." Genes with particular ensembles are associated with changes in gene expression during differentiation and the control of cell-type-specific gene regulators such as super-enhancers, distal-enhancer-looping, nuclear lamina, and master transcription factors. Ensembles are also disproportionately affected in both HD and Alzheimer's disease. Together, this thesis presents a toolkit for recognizing and understanding epigenetic patterns that can further insight into the regulation of genes in development and disease. / by Christopher W. Ng. / Sc. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Human adaptation of avian influenza viruses pose an enormous public health challenge as the human population is predominantly naive to avian influenza antigens. As such, constant surveillance is needed to monitor the circulating avian strains. Of particular importance are strains belonging to H5N1, H7N7, H7N2 and H9N2 subtypes that continue to circulate in birds worldwide and have on occasions caused infections in humans. A key step in influenza human adaptation is the accumulation of substitutions/mutations in the viral coat glycoprotein, hemagglutinin (HA), that changes HA's binding specificity and affinity towards glycan receptors in the upper respiratory epithelia (referred to as human receptors). Unlike for the H1, H2, H3 and more recently H5 HA a correlation between the quantitative binding of HA to human receptors and respiratory droplet transmissibility has not been established for H9 and H7 subtypes. This thesis is a systematic investigation of determinants that mediate changes in HA-glycan receptor binding specificity, with focus on the molecular environments within and surrounding the glycan receptor binding site (RBS) of avian HAs, particularly the H9 and H7 subtypes. The glycan receptor binding properties of HA were studied using a combination of biochemical and molecular biology approaches including dose dependent glycan binding, human tissue staining and structural modeling. Using these complementary analyses, it is shown that molecular interactions between amino acids in and proximal to the RBS, including interactions between the RBS and the glycan receptor converge to provide high affinity binding of avian HA to human receptors. For the H9 HA [alpha]2-->6 glycan receptor-binding affinity of a mutant carrying Thr-189-->Ala amino acid change correlated with the respiratory droplet transmission in ferrets conferred by this change. Further, it was demonstrated for the first time that two specific mutations; Gln226-->Leu and Gly228-->Ser in glycan receptor-binding site of H7 HA substantially increase its binding affinity to human receptors. These approaches and findings contribute to a framework for monitoring the evolution of HA and the development of general rules that govern human adaption applicable to strains beyond ones currently under study. / by Karunya Srinivasan. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 144-161). / Understanding structure-function relationships involved in development of disease is a critical component for the rational design of successful therapeutics, allowing researchers to target precise molecular mechanisms and to anticipate and address future challenges early in development. However, for many diseases and contexts, these precise relationships are unknown or poorly defined. In this thesis, I develop new tools and implement integrated approaches to explore structure-function relationships of disease in two biological contexts: i) the interactions of proteins with glycans that can modify disease susceptibility, and ii) the interactions of viral coat proteins with antibodies that neutralize and prevent viral infection. Section 1: Glycosylation, which is the addition of sugar moieties to proteins and peptides, is one of the most abundant post-translational modifications, modifying a diverse set of biological processes in both health and disease. However, the glycome, or ensemble of glycan structures produced by a cell, is difficult to study, given the low-affinity, high-avidity structural interactions of glycans with host proteins and their non-template-driven biosynthesis. In this section, I investigate how changes in glycomic composition affect disease, and develop and implement tools and approaches that address the aforementioned challenges. In the first part, I leverage a novel in vitro model of lung adenocarcinoma metastasis to study the changes in cell surface glycosylation that correspond with and may influence metastatic progression. Here, I implement and integrate tools, such as glycan biosynthesis gene expression analysis, glycan mass spectrometry analysis frameworks, and lectin binding arrays, to measure the changes in the glycome, identifying several key motifs and glycomic features of metastatic cells. In the second part, I leverage protein-glycan structural insights to understand and predict the susceptibility of a novel seal influenza virus, H3N8, to infect human populations. Here, we used a glycan array, an assay which measures a diversity of glycan binding motifs with biologically relevant avidity and presentation, to demonstrate that unlike human-adapted H3N2, H3N8 lacks the affinity for long a2,6-linked sialylated glycans, which have been shown to determine host specificity and tropism for humans. This result was compared to other experiments, including tissue staining and in vitro replication, to determine that this seal H3N8 virus is unlikely to infect and spread within human populations. These insights further our understanding of how complex ensembles of glycans can influence susceptibility to disease. Section 2: The neutralization of emergent viral pathogens, including Ebola and Zika viruses, by therapeutic antibodies offers the potential to prevent viral infection and to treat patients even after they have been infected. However, given the real-time nature of viral outbreaks, strategies are needed which reduce the development cycle by allowing rational design and selection of potent neutralizing antibodies with minimal susceptibility to antigen escape. Here, I develop a structural, network-based computational and experimental framework which uses information about the viral coat protein and antibodies to identify key features of epitope-paratope interactions. In the context of Ebola virus, I use this analytical framework to identify two novel epitopes on the surface of the trimeric coat glycoprotein which are highly constrained, such that antibodies targeting these regions are likely to be resistant to antigen escape. Furthermore, I use this framework to uncover key differentiators of the overlapping therapeutic antibodies 2G4, 4G7, and KZ52, and describe how these differences may affect their relative susceptibility to epitope escape mutations. In the context of Zika virus, I further develop and expand this analytical framework using a computational Zika E glycoprotein assembly, and use it to understand neutralizing epitopes on the Zika virus surface in the context of their quaternary structure. In doing so, I identify critical interface residues for the 3E31, ZV67, C8, C10, Z23 and Z3L1 antibodies, and use these insights to generate C867, a novel asymmetric IgG bispecific antibody against Zika virus. This antibody has high bispecific purity, maintains in vitro potency, and given its non-overlapping epitopes, has a higher requirement for antigen escape. Taken together, the tools, frameworks, and developments presented here are important additions to the preclinical antibody development toolkit, and enable rational design and faster response to emergent viral pathogens. / by Devin Scott Quinlan. / Ph. D.

Thesis: Sc. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2017. / This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. / Cataloged from student-submitted PDF version of thesis. / Includes bibliographical references (pages 75-83). / Reassortment is a reticulate evolutionary process that results in genome shuffling; the most prominent virus known to reassort is the influenza A virus. Methods to identify reassortant influenza viruses do not scale well beyond hundreds of isolates at a time, because they rely on phylogenetic reconstruction, a computationally expensive method. This thus hampers our ability to test systematically whether reassortment is associated with host switching events. In this thesis, I use phylogenetic heuristics to develop a new reassortment detection algorithm capable of finding reassortant viruses in tens of thousands viral isolates. Together with colleagues, we then use the algorithm to test whether reassortment events are over-represented in host switching events and whether reassortment is an alternative 'transmission strategy' for viral persistence. / by Eric J. Ma. / Sc. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Proteins and polysaccharides are of growing importance as a source for novel therapeutic compounds and target a range of diseases, from cancer to infections from pathogens. However, owing to their large and complex structures, they face a unique set of challenges, compared to small molecules, in their discovery and development as safe, efficacious drugs. Towards addressing these challenges, we describe in this thesis the implementation of structure-function relationship approaches to characterize and engineer polysaccharides and antibodies to improve their therapeutic profiles. The plant polysaccharide pectin, when modified, has demonstrated significant anticancer activity in animal models and small-scale clinical trials. Its development has been hampered, however, due to its complex structure and lack of structure-activity correlates. Using an integrated approach, we engineer a modified pectin that exhibits significant in vivo anticancer activity, which we link to specific structural attributes and cellular functional mechanisms. These results improve our structure-function understanding of anticancer modified pectin, an important step towards the clinical use of this complex polysaccharide. Applying what we learned from pectin, we develop an integrated framework to identify a contaminant in batches of heparin, a polysaccharide anticoagulant drug, associated with an outbreak of allergic-type reactions in 2007-2008. Employing orthogonal analytical approaches to overcome challenges of characterizing structurally complex pharmaceutical heparin, we determine that the structurally related glycan, oversulfated chondroitin sulfate, is the major contaminant. We link its presence to activation of the contact pathway, thereby establishing a structure-function understanding of contaminated heparin and improving the safety profile of this polysaccharide drug. Transitioning knowledge gained from the structure-function characterization of polysaccharides, we engineer, by structure-based design, a broad spectrum neutralizing antibody to dengue virus, which yearly infects more than 200 million people, causing approximately 21,000 deaths. We incorporate complementary approaches of energetics and empirical informatics methods to rationally redesign an existing antibody for greater breadth and potency, resulting in an engineered antibody with binding to all four virus serotypes and good in vitro potency. Overall, this thesis provides important insights into structure-function approaches through the use of complementary methods to characterize and engineer therapeutic polysaccharides and antibodies. / by Luke Robinson. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 137-146). / Analysis of single cells using flow cytometry techniques has created a wealth of knowledge about cellular phenomena that could not be obtained by population average measurements. As these techniques are integrated with others to increase the number of parameters that can be measured on single cells and these measurements are made more quantitative, their ability to discriminate between sub-populations of cells increases. Microfabricated sensors offer unique advantages in this area because their internal geometries can be engineered at a size scale comparable to the cell's, making them naturally well-suited for single-cell measurements. The suspended microchannel resonator (SMR) is a versatile flow cytometry platform which is capable measuring the mass of single cells with femtogram resolution. The net frequency shift of a resonant cantilever as the cell transits the fluid-filled microchannel running through it is proportional to the buoyant mass of the cell. The resonance frequency of the SMR is also highly sensitive to a cell's position along the cantilever's length. This thesis presents a new method which makes use of this property to accurately quantify the electrophoretic mobility (EPM) of cells transiting the SMR while subjected to oscillatory electric fields. Recorded resonance frequency time courses can be analyzed to extract both the buoyant mass and EPM of individual cells. This instrument has been used to simultaneously measure the EPM and buoyant masses of discrete polystyrene microspheres and Escherichia coli bacteria. As it has been applied to microspheres of known density, the integrated measurement makes it possible to compute the absolute mass and surface charge of individual microspheres. It has been shown that integrated single-microsphere mass and surface charge measurement enables differentiation of complex aqueous suspensions which is not possible by either measurement alone. / by Philip Dextras. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2014. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Nucleic acids are subject to extensive chemical modification by all organisms. These modifications display incredible structural diversity, and some are essential for survival. Intriguingly, several of these modifications are unique to bacteria, including many human pathogens. Given the enormous global disease burden due to bacterial infections, and the rapidly increasing rates of antibiotic resistance reported across the world, the need for research to address mechanisms of bacterial survival is more pressing than ever. The goal of this thesis was to determine the function of nucleic acid modifications in pathogenic bacteria, and to evaluate their impact on the three major stages of the infectious disease process: pathogenesis, diagnosis, and therapy. We first used quantitative profiling of tRNA modifications to identify novel stress responses that help mediate host invasion in the world's most common pathogen, Helicobacter pylori. This work uncovered potentially novel targets for the development of new compounds that inhibit pathogenesis. We then developed a new animal model of mycobacterial lung infection that enables drug development and biomarker screening studies in standard laboratories without high-containment facilities. We showed that infection with Mycobacterium bovis bacille Calmette-Guérin produces a granulomatous lung disease in rats that recapitulates many of the important pathological features of human tuberculosis. This model also allowed us to test the utility of nucleic acid modifications as diagnostic biomarkers. Finally, we investigated the effect of the common, transferable bacterial DNA modification phosphorothioation on oxidative and antibiotic stress responses in several pathogens. We showed that phosphorothioation can reduce the effectiveness of antibiotic therapy, which may make it an environmental source of acquired antibiotic resistance. These studies show that nucleic acid modifications play diverse roles in pathogenic bacteria, and that their modulation may be a promising target for developing new tools that can disrupt pathogenesis, improve diagnosis, and strengthen therapy. / by Brandon S. Russell. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2013 / Cataloged from PDF version of thesis. / Includes bibliographical references. / Iron is a cofactor of a number biochemical reactions that are essential for life. In the marine environment, this micronutrient is a scarce resource that limits processes of global importance such as photosynthesis and nitrogen fixation. Given that marine microorganisms play a central role in modulating such biogeochemical cycles, understanding how their distribution, diversity and activity may be affected by changes in iron availability is key. This thesis explores how the availability of iron affects the ecology of marine microbial populations and communities. At the population level, I characterized the prevalence and diversity of iron acquisition strategies in specific populations of marine vibrios with distinct micro-habitat preferences. Using a combination of genomics and functional screens, I showed that siderophore-based iron acquisition is not conserved at the organismlevel but represents a stable trait at the population level. This population-level trait further appears to play a role in driving the diversification of specific vibrio populations, especially of those that are thought to prefer particles as a micro-habitat. At the community level, I measured whole microbial community responses to iron addition in microcosm experiments in different regions of the Pacific Ocean. Using metagenomics, I characterized the impact of iron availability on the microbial community structure of the Central Equatorial Pacific Ocean. This study showed that addition of iron to an iron-limited ecosystem triggers a phytoplankton bloom dominated by Pseudo-nitZschia-like diatoms, which in turn stimulate a Bacteroidetes population functionally distinct from the ambient free-living population. In the North Pacific Subtropical Gyre, I explored how iron availability impacts microbial community gene expression dynamics. Using a metatranscriptomic approach I showed that in that environment, the impact of iron was tightly connected to the supply of other limiting macronutrients, and seems to mostly affect photosynthetic organisms. This initial study paves the way for more in depth and longer-term studies to further investigate the effects of iron on the dynamics of the microbial community in the North Pacific Subtropical Gyre. Taken together data and analyses presented in this thesis demonstrate how iron availability can shape the ecology of marine microorganisms at population, community and functional levels. / by Laure-Anne Ventouras. / Ph.D.