Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2009. / Cataloged from PDF version of thesis. / Includes bibliographical references. / There is an ever-growing demand for enantiopure chemical compounds, particularly new pharmaceuticals. Enzymes, as natural biocatalysts, possess many appealing properties as robust asymmetric catalysts for synthetic chemistry. However, their enantioselectivity toward most synthetically useful, non-natural substrates is typically low. Therefore, improving enzymatic enantioselectivity toward a given substrate is a practically important but arduous task. Here we report a highly efficient selection method for enhanced enzymatic enantioselectivity based on yeast surface display and fluorescenceactivated cell sorting (FACS). By exploiting the aforementioned method, in just three rounds of directed evolution we both greatly increased (up to 30-fold) and also reversed (up to 70-fold) the enantioselectivity of the commercially useful enzyme, horseradish peroxidase (HRP), toward a chiral phenol. In doing so, we discovered that mutations close to the active site not only preserve HRP catalytic activity but impact its enantioselectivity far greater than distal mutations. We thus examined how a single mutation near the active site (Argl78Glu) greatly enhances (by 25-fold) the enantioselectivity of yeast surface-bound HRP. Using kinetic analysis of enzymatic oxidation of various substrate analogs and molecular modeling of enzyme-substrate complexes, this enantioselectivity enhancement was attributed to changes in the transition state energy due to electrostatic repulsion between the carboxylates of the enzyme's Glu- 178 and the substrate's D enantiomer. In addition, the effect of yeast surface immobilization and influence of a fluorescent dye on controlling the enantioselectivity of the discovered HRP variants was investigated. Soluble variants were also shown to have marked improvements in enantioselectivity, which were rationalized by computational docking studies. / by Eugene Antipov. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2011. / Each page number preceded by chapter or appendix number. Cataloged from PDF version of thesis. / Includes bibliographical references (p. F-150 - F-163). / The activation of inflammatory responses, while critical for host defense, contributes to hepatic injury in numerous acute and chronic liver disease states as well as drug-induced liver injury (DILI). The interactions that mediate susceptibility to liver injury and disease, however, are still poorly understood, underscored by the complexity of immune interactions and the diverse cellular composition and functions of the liver. Using Citrobacter rodentium, a well characterized rodent-specific enteric pathogen as a source of extrahepatic inflammatory stress; host liver responses, metabolic dysregulation, and susceptibility to injury in C57BL/6 mice were investigated. For the first time, we show altered liver pathology during the early course of C. rodentium infection, characterized by periportal necrosis indicative of thrombic ischemic injury, correlating with distinct circulating and tissue specific cytokine/chemokine profiles. Using Acetaminophen (APAP), a widely used analgesic and well-characterized hepatotoxin, we evaluated liver responses in isolation and in the context of host inflammation to gain insight into the role of live bacterial infection in altering liver metabolism and susceptibility to DILI. We combined systemic and tissue-specific cytokine/chemokine levels, clinical serum chemistries, and histopathological assessments of hepatic and enteric inflammation and necrosis to measure molecularlevel responses to treatment and their physiological effect. Using principal components analysis (PCA), clustering, partial least squares regression (PLSR), and a combination mutual-information-correlation network, enabled detection and visualization of both linear and nonlinear dependencies between molecules and physiological states across tissues and timepoints. C. rodentium-induced inflammatory stress was finally investigated for its potential in altering drug pharmacokinetics (PK) of substrates varying in their metabolic biotransformation and clearance mechanisms. Infection resulted in increased systemic oral exposure (AUC) of clinically relevant xenobiotics such as verapamil, propranolol, and digoxin. Functionally, these changes were not found dependent on CYP-mediated biotransformation of parent compounds; rather, they appear driven more by proposed gut barrier compromise. In conclusion, gastrointestinal infection with C. rodentium alters systemic and hepatocytes specific responses, not previously appreciated from this enteric pathogen, making it a useful model for studying host-pathogen interactions under acute hepatic inflammatory stress and injury. / by Arkadiusz R. Raczynski. / Ph.D.

Thesis: Sc. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 213-223). / Within synthetic biology, significant progress in creating networks using transcriptional and translational control has been made, but a network comprised solely of protein-protein interactions has not yet been built. To realize this goal, new design rules for assembling and connecting protein devices into circuits are required. In this thesis, a framework for rapid assembly and delivery of genetic networks in Saccharomyces cerevisiae is described. Utilizing this scheme, design principles of modular chimeric proteins, engineered pathway convergence, and phosphorylated activated localization are developed and subsequently applied to create phosphoin/phospho-out composable protein devices in yeast. The devices implement BUFFER, NOT and OR logical operations and collectively form a functionally complete set whereas multiple instances of the devices can be connected to implement any logical expression. Furthermore, bridge devices to interface small molecule sensors and transcriptional networks with protein devices are created. To illustrate composability, two systems are engineered in yeast, the first of which interfaces a phosphorylation load driver within flanking transcriptional regulatory modules to mitigate retroactivity, exemplifying time-scale separation as a means of realizing functional modularity in biological networks. The second system, a fast bistable toggle switch, is the first synthetic network based solely on protein interactions and can be interfaced endogenously to allow controllable abrogation of yeast budding. Work in this thesis provides a set of useful design principles for engineering protein networks in eukaryotes and may improve understanding of natural signaling motifs, allow modulation of existing networks, or foster new synthetic biology applications. / by Deepak Mishra. / Sc. D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references. / Integrins are a family of heterodimeric cell surface receptors that are functionally important for cell adhesion, migration and proliferation. Certain integrins, especially those that are known to recognize the arginine-glycine-aspartate (RGD) motif, are heavily overexpressed in many cancers relative to healthy tissue, making them attractive targets for therapeutic intervention. However, prior attempts to antagonize these integrins as a cancer therapy have all failed in the clinic. In this thesis, we instead exploit integrins as a target tumor antigen in the context of immunotherapy. The engineered cysteine knot peptide, 2.5F, is highly crossreactive and capable of recognizing multiple RGD-binding integrins. Our initial attempts to utilize this binder as a targeting moiety for delivering IL-2 as an immunocytokine failed. Mathematical modeling results indicated that immunocytokines, unless adhering to specific design criteria, are unlikely to benefit from targeting and may actually exhibit limited efficacy. Therefore, we "deconstructed" this immunocytokine into its functional parts: extended half-life IL-2 and 2.5F-Fc, the antibody-like construct directed against RGD-binding integrins. This combination immunotherapeutic approach was able to synergistically control tumor growth in three syngeneic murine models of cancer, including durable cures and development of immunological memory. Contrary to prior attempts at integrin-targeting, the mechanism of action was independent of functional integrin antagonism, including effects on angiogenesis and tumor proliferation. In fact, efficacy of this therapy depended solely upon the adaptive and innate arms of immunity, specifically CD8+ T cells, macrophages, and dendritic cells. Furthermore, checkpoint blockade, the gold standard for immunotherapy to date, can further enhance the efficacy of this therapeutic approach. This signifies that the combination of IL-2 and 2.5F-Fc exerts a distinct, yet complementary immune response that opens the door for clinical translation. / by Byron H. Kwan. / Ph. D.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2010. / Cataloged from PDF version of thesis. / Includes bibliographical references. / The validity and usefulness of clustering marine group II tetranucleotide signatures using emergent self-organizing maps was investigated. Fosmids from the HF200 library were chosen for sequencing based on end-sequence tetranucleotide clustering with group II seed sequences, as well as blastx homology. Fosmids were sequenced using a single 454- titanium sequencing run, and contigs subsequently assembled in silico. A total of 99 contigs over 20kb were retrieved, at least 72 of which belong to the marine group II archaea. The phylogenetic substructure of the marine group II archaeal clusters having more than a few representatives was investigated, by clustering tetranucleotide signatures of group II contigs over 20kb, also with an emergent self-organizing map. The distribution of these clusters in the Hawaii Ocean Time Series depth profile fosmid libraries in the DeLong lab were mapped onto depth profiles from three independent cruises. / by Rachel A. Hillmer. / S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references. / The heterogeneity of individual cellular behavior in response to physical and chemical stimuli has raised increasing attention in many biological processes. There is great incentive in developing techniques for high throughput single-cell measurements and manipulations. Particularly, cell size has been recognized as an important parameter in single cell study and pericellular protease activity plays a key role in regulating the microenvironment of individual cells. Therefore, this thesis focuses on establishing new methods to address the issues of cell size and single cell protease measurement. We first develop a size-based cell separation technique using Dean-coupled inertial microfluidic sorter. Separation of cells by size before downstream assays might be beneficial in simplifying the system and facilitating the discovery of rare subpopulations through enrichment of cells with certain size range or cell cycle phase. By investigating the particle focusing and separation mechanisms in curved microfluidic channel, we develop a novel design of inertial microfluidic sorter with higher separation resolution and then demonstrate its capacity in leukocyte isolation from blood. This novel cell sorter would be a promising alternative to many other cell separation problems. We then establish a microfluidic platform for functional measurement of single cell pericellular proteases, including both those secreted and expressed on cell surface. We apply the platform to studying the PMA-mediated protease response of HepG2 cells at single-cell level and reveal the diversity in the dynamic patterns of single-cell protease activity profile upon drug stimulation. We also present the preliminary exploration of single-cell protease activity behavior in anticancer drug resistance development. Lastly, we explore the applicability of our platform for single-cell shedding measurement. Protease-mediated molecular shedding is one of the key mechanisms through which individual cells actively regulate their own microenvironment. However, the amount of molecules being shed for individual cells is extremely low, posing significant challenges in detecting shedding quantitatively. By means of analytical analysis and numerical simulations, we investigate the intrinsic noise of low-abundance molecule detection. Experimental characterizations have also been performed to evaluate the impact of practical factors on actual readout variation. / by Lidan Wu. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Biological Engineering, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 118-125). / The adaptive immune system is an extraordinarily diverse inventory comprised of highly specialized cells, the differentiation of which requires numerous lineage specifications at various developmental stages. The precise control of immune cell differentiation and the delicate balance of their population composition are crucial for effective protection against infectious environmental agents, without triggering autoimmune responses or allergies. It is therefore important to understand at the molecular level in individual cells how lineage commitment is regulated. I explored the heterogeneous gene expression during the lineage specification of single T helper cells, by quantitatively measuring mRNA and protein levels. I have discovered a paradigm of cell lineage specification governed by the signaling interplay between extracellular cues and intracellular transcriptional factors, where the strength of extracellular signaling dominates over the intracellular signaling components. In the presence of extracellular cues, T helper cells stochastically acquire any intermediate Thl/Th2 states. The states of T helper cells can be gradually tuned by depriving availability of extracellular cytokines, which are produced stochastically by a small subpopulation of cells. When extracellular cues are removed, the weak intracellular signaling network reveals its effect, leading to classic mutual exclusion of antagonistic transcriptional factors. / by Miaoqing Fang. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2018. / 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. / The human gastrointestinal tract is home to a dense and dynamic microbial community. Recent advancements in sequencing technology have revealed numerous relationships between the composition of these communities and human and health and disease. In some cases, researchers have shown causal relationships between the presence or absence of particular microorganisms and disease. These findings offer promise for using microorganisms or microbial communities to modulate health and disease, but to date, we lack tools and mechanistic insight needed for rational engineering of these communities. Understanding how microorganisms enter, colonize, grow, and disperse to new hosts present key considerations for rational engineering of the human gastrointestinal tract. In this thesis, I use experimental studies of the human and murine gastrointestinal tract to address these considerations. In the first study, I examined endospores and other resistant cell types in the gastrointestinal communities of unrelated humans to identify the ecological role of these states in the distribution of bacterial populations in healthy people. I used this information to infer shared roles for these organisms in successional states in the human gut, and identify host- and diet-derived metabolites as environmental signals mediating the growth and colonization of these organisms. In the second study, I examined the potential for using targeted manipulations of diet to favor selective growth and colonization by an introduced bacterium in the murine gastrointestinal tract. I showed that resource exclusivity of this bacterium permits its selective expansion in this environment, and negatively impacts the growth of other commensals. Central to the goal of rational engineering of the gut microbiota, these studies reveal ecological considerations that may promote or inhibit colonization by introduced commensals in this complex ecosystem. This work invites provides a conceptual framework for integrating systems microbial ecology with engineering design to manipulate the composition of the gastrointestinal microbiota. / by Sean M. Kearney. / Ph. D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references. / The high incidence of traumatic brain injury due to adverse impact events ranging from head collisions to ballistic attacks has prompted significant interest in synthetic polymer gels capable of mimicking key mechanical properties of brain tissue. These so-called brain tissue simulants are valuable tools for developing protective strategies because they can serve as test media to evaluate new helmets or optimize robotic surgery techniques. However, the so-called "soft matter" employed to date for ballistic applications, such as ballistic gelatin and clay, are crude mechanical representations of brain tissue. Therefore, there remains a need for a class of tissue simulant materials that more accurately replicates the mechanical behavior of brain tissue under impact loading, specifically in terms of deformation resistance and impact energy dissipation. This thesis focuses on design and synthesis of hierarchically structured gels, and mechanical characterization of these compliant gels for comparison with mammalian brain tissue. In particular, we use impact indentation to explore how the impact energy dissipation response varies as a function of species for brain tissue, or as a function of molecular composition and structure for synthetic gels. We find that a bilayered polydimethylsiloxane (PDMS) composite system enables the decoupling of the material's deformation resistance and energy dissipation characteristics, and can be tuned to fully match porcine brain tissue. However, given that the top PDMS layer is highly adhesive, we investigate whether adhesion plays a significant role in modulating the energy dissipation response, which has important implications in the utility of the tissue simulant material for ballistic applications. With a separate bilayered PDMS composite system, we decouple surface adhesion from bulk viscoelasticity, and quantify their individual contributions to impact energy dissipation. Through these experimental studies, in addition to a finite element computational analysis, we establish fundamental design principles and provide new insights regarding mechanisms that govern the extent of deformation and energy dissipation in compliant polymeric materials. Finally, we extend the capabilities of our impact indentation technique by demonstrating a novel analytical approach to extract viscoelastic moduli and relaxation time constants directly from the measured impact deformation response, thus significantly broadening the utility of impact indentation. With conventional characterization techniques such as shear rheology, several challenges arise when the material of interest has stiffness on the order of 1 kPa or lower, as is the case with brain tissue, largely due to difficulties detecting initial contact with the compliant sample surface. In contrast, impact indentation does not require contact detection a priori, and thus can potentially be utilized as a more accurate tool to characterize the viscoelastic properties of a wider range of soft matter for diverse biomedical or engineering applications, not limited to brain tissue simulants. This semi-analytical approach enables future studies to extract viscoelastic properties of brain tissue and tissue simulant polymers with increased accuracy and spatial resolution, in the context of traumatic brain injury, protection, and recovery. / by Bo Qing. / Ph. D.

Thesis: Ph. D., Massachusetts Institute of Technology, Department of Biological Engineering, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 107-124). / The extensive genomic diversity of complex systems, such as the human gut microbiome and the evolution of human cancer, has been revealed with advances in DNA sequencing. But we are still at an early stage in understanding this genomic diversity to expand our knowledge in biology and for biomedical applications. Taking the diverse human gut microbiome as an example, little is known about the rapid exchange of antibiotic resistance genes and virulence factors as part of the mobile gene flow between the microbes in the gut. Understanding such heterogeneous systems often involves studying the nature and behavior of the individual cells that constitute the system and their interactions. However, it is technically challenging to probe the genomic material of cells, the smallest unit of life and amplify single genomes for sequencing. Current single-cell technologies require complex instrumentation and the data quality is often confounded by biased genome coverage and chimera artifacts. We address these challenges with a new single-cell technology paradigm to make high-quality low-input genomic research accessible to scientists. We developed hydrogel-based virtual microfluidics as a simple and robust platform for the compartmentalization of nucleic acid amplification reactions. We applied whole genome amplification (WGA) to purified DNA molecules, cultured bacterial cells, human gut microbiome samples, and human cell lines in the virtual microfluidics system. We demonstrated whole-genome sequencing of single-cell WGA products with excellent coverage uniformity and markedly reduced chimerism compared with traditional methods. Additionally, we applied single-cell sequencing to identify horizontally transferred genes between the microbes in the gut and revealed human population activities' selective pressure in shaping the mobile gene pools. Altogether, we expect virtual microfluidics will find application as a low-cost digital assay platform and as a high-throughput platform for single-cell sample preparation. This work offers a significant improvement in making high-quality low-input genomic research accessible to scientists in microbiology and oncology. / by Liyi Xu. / Ph. D.