Thesis: Ph. D., Massachusetts Institute of Technology, Computational and Systems Biology Program, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 148-160). / Biological hydrogels are fundamental to life, from microbial biofilms to mucus and the nuclear pore in humans. These hydrogels exhibit complex selective permeability behavior, allowing the passage of some particles while blocking the penetration of others. This selective permeability is critical for understanding the biological and medicinal impact of mucus, which coats all non-keratinized epithelia in the body. Mucus controls the penetration of microbes, pollutants, and nanoparticles through a combination of steric and interactive (binding-based) constraints. For small molecules, binding to mucus and in particular mucin, the main gel-forming component of mucus, affects diffusive permeability and may also affect a molecule's biological or therapeutic activity. However, the molecular characteristics leading to mucus binding are not well understood. / I therefore developed a mucus binding assay with substantially greater throughput than any existing assay, and combined it with a mucin binding screen to identify a new motif as associated with binding to mucin. I also validate the link between binding to mucin and reduced activity in mucin for the antibiotic colistin. Next, I applied my binding technique to study the binding of a wide range of antibiotics and inhaled drugs to respiratory mucus, and identified previously unknown mucus binding interactions. These binding interactions could impact the activity of the drugs within the mucus or impact their lung residence time in the case of highly muco-obstructive lung diseases. The nuclear pore, which controls the passage of material between the nucleus and the cytoplasm, is similar to mucus in that it too is a selectively permeable network of disordered proteins. / Passage through the nuclear pore requires interaction with the network that was initially thought to be purely hydrophobic in character. However, there is evidence that electrostatic interactions also partly govern nuclear pore transport. Here, we apply a peptide-based system to study the interplay of hydrophobic and electrostatic interactions to further dissect the biochemistry underlying nuclear pore function. / by Jacob Julian Seid Witten. / Ph. D. / Ph.D. Massachusetts Institute of Technology, Computational and Systems Biology Program

Tuberculosis remains a major global health challenge and the increasing strength and prevalence of drug resistance threaten to undo much of the good progress made. As one of the primary, frontline anti-tuberculous drugs, growing resistance to rifampicin in particular is concerning. Sub-lethal drug exposure and the development of adaptive phenotypic drug resistance, represent an important avenue by which genetic resistance and treatment failure or relapse may occur. Proteins and general metabolites are molecular classes that are highly dynamic, responsive and essential to understanding the state of an organism, while mass spectrometry-based proteomics and metabolomics are powerful tools by which these can be examined. For mycobacteria in particular, the lipidome and cell wall are compartments of major importance with respect to virulence, adaptation, host-pathogen interactions and persistence. As such, we sought to determine the effect of sub-lethal rifampicin exposure upon the model organism Mycobacterium smegmatis over time and determine what phenotypic adaptations might be observed and explained by alterations in the proteome and lipidome, with special focus on the cell wall sub-proteome. From these data we formed several new hypotheses with respect to virulence and mechanisms of both drug resistance and sensing, which were investigated further. Finally, we examined the effect of sub-lethal rifampicin exposure, and consequent proteomic alterations, upon the M. smegmatis lipidome and propose a model by which mycobacteria respond to sub-lethal challenge with rifampicin: Upon initial insult, drug-susceptible mycobacterial growth slows and stress response networks, including the SOS response, are temporarily activated. For both susceptible and resistant bacteria, cell wall remodelling begins early through dysregulation of cell wall and lipid synthesis enzymes — such as MtrAB, Mur proteins and PimB — resulting in ultimate accumulation of lipids with composition such as to impede passive diffusion of rifampicin into the cell. Some of this lipid accumulation, namely with PIMs, may take place rapidly and so ultimately reveal extremely large increases in abundance, which possibly necessitates downregulation of enzymes such as PimB by ~4 hours post treatment. In concert with ongoing lipid dysregulation, the cell wall proteome is altered as ABC transporter proteins are generally downregulated as an additional mechanism by which to control cell wall permeability through altered cell wall composition — through removal of cell wall penetrating transport proteins — and by limiting controlled influx of exogenous compounds. Bacterial efforts to resume normal growth and adapt to rifamipicin stress involves the dysregulation of numerous virulence factors, such as PknG, which results in impaired virulence. Transcriptional and translational machinery are also gradually upregulated so as to compensate for intracellular rifampicin's inhibition of RpoB, with transcriptional activity regulated separately to that of translational machinery. Ultimately, the combination of increased transcription, translation, and cell wall impermeability allows mycobacteria to overcome rifampicin challenge and resume normal growth. In M. smegmatis specifically, all this is accompanied by the gradual upregulation of the chromosomal resistance factor Arr which, at a later timepoint, modifies extracellular rifampicin to alleviate drug pressure.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Computational and Systems Biology Program, 2010. / 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 regulation of gene expression underlies the morphological, physiological, and functional differences between human cell types, developmental stages, and healthy and disease states. Gene regulation in eukaryotes is controlled by a complex milieu including transcription factors, microRNAs (miRNAs), cis-regulatory DNA and RNA. It is the quantitative and combinatorial interactions of these regulatory elements that defines gene expression, but these interactions are incompletely understood. In this thesis, I present two new methods for determining the quantitative specificity of gene regulatory factors. First, I present a comparative genomics approach that utilizes signatures of natural selection to detect the conserved biological relevance of miRNAs and their targets. Using this method, I quantify the abundance of different conserved miRNA target types, including different seed matches and 30-compensatory targets. I show that over 60% of mammalian mRNAs are conserved targets of miRNAs and that a surprising amount of conserved miRNA targeting is mediated by seed matches with relatively low efficacy. Extending this method from mammals to other organisms, I find that miRNA targeting rules are mostly conserved, although I show evidence for new types of miRNA targets in nematodes. Taking advantage of variations in 30 UTR lengths between species, I describe general properties of miRNA targeting that are affected by 30 UTR length. Finally, I introduce a new, high-throughput assay for the quantification of transcription factor in vitro binding affinity to millions of sequences. I apply this method to GCN4, a yeast transcription factor, and reconstruct all known properties of its binding preferences. Additionally, I discover some new subtleties in its specificity and estimate dissociation constants for hundreds of thousands of sequences. I verify the utility of the binding affinities by comparing to in vivo binding data and to the regulatory response following GCN4 induction. / by Robin Carl Friedman. / Ph.D.

Thesis: Ph. D., Massachusetts Institute of Technology, Computational and Systems Biology Program, 2015. / 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. / MicroRNAs (miRNAs) are ~21-24 nt non-coding RNAs that mediate the degradation and translational repression of target mRNAs. The genomes of vertebrate organisms encode hundreds of miRNAs, each of which may regulate hundreds of mRNA targets. Thus, miRNAs are crucial post-transcriptional regulators engaged in vast regulatory networks. To date, the characteristics of these networks remain mysterious due to the difficulty of identifying miRNA targets through either experimental or computational means. To understand the physiological roles of miRNAs in animal species, it is of fundamental importance to elucidate the structure of the targeting networks in which they participate. The recognition of a miRNA target is guided largely by perfect Watson-Crick base pairing interactions between nucleotides 2-7 from the 5' end of the miRNA (i.e., the "seed" region) and complementary motifs embedded in the 3' UTRs of the target mRNAs. The prevalence of these motifs throughout the transcriptome poses a challenge to our understanding of how specificity emerges: since the presence of a motif is not sufficient to mediate target repression, what contextual features discriminate effective target sites from ineffective ones? Further complicating this is the proposition that "noncanonical" sites lacking perfect seed pairing might mediate repression, which would expand the potential number of functional target sites by orders of magnitude. In the second chapter of this work, we define the features that predict effective miRNA target sites, incorporating their relative influence into a quantitative model which can outperform existing computational models and experimental approaches in target identification. Though the molecular roles of miRNAs in gene regulation have long been appreciated, the functions of most miRNAs in living organisms has remained elusive. In the third chapter of this work, we discuss the consequences of genetic ablation of miR-196, a deeply conserved miRNA that is predicted to simultaneously repress many HOX genes, in the mouse. We propose a role for miR-196 in the spatial patterning of the vertebrate axial skeleton. Isolating the cell populations that express the miRNA during early mammalian development, we attempt to characterize the direct in vivo targets of miR-196 and dissect the molecular underpinnings of the phenotypes observed. / by Vikram Agarwal. / Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Computational and Systems Biology Program, 2013. / Cataloged from PDF version of thesis. / Includes bibliographical references. / All the cells in the body contain the same genome yet showcase drastically different phenotypes. This is the result of different transcriptional programs, which are partly controlled by epigenetic modifications, including DNA methylation. In this thesis, I analyze genome-scale DNA methylation profiles across pre-implantation development to identify the targets and characterize the dynamics of global demethylation that lead to totipotency and the subsequent changes to embryonic specification. In Chapter 1, I validate and refine the decades old model for DNA methylation in mouse embryogenesis, identify many retrotransposons with active DNA methylation signatures at fertilization, and discover many, novel differentially methylated regions between the gametes that exist transiently during early development. Notably, the majority of epigenetic events unique to mammalian pre-implantation development are characterized in mouse. In Chapter 2, 1 describe the DNA methylation dynamics in human preimplantation development and show that the regulatory principles that operate in mouse are conserved, though some of their targets are species-specific and define regions of local divergence. Finally, in Chapter 3, I compare DNA methylation dynamics of fertilization to an artificial reprogramming process, somatic cell nuclear transfer, in mouse, and find that most dynamics are conserved but occur at a smaller magnitude after artificial reprogramming. I conclude this thesis with a summary of the chapters and a brief discussion of ongoing and future work. / by Michelle M. Chan. / Ph.D.

No description available.

Vi har identifierat flexibelt utbyte och lagring av data i databaser, tillsammans med långvarig satsning på olika existerande och framtida modeller som nyckar till förståelse av det regler nätverk som utgör bron mellan geno- och fenotyp. Denna pilot studie av modellering av stora cellulära kontroll nätverk utgår från en intressant medicinsk frågeställning inom molekylär cellbiologi: Är framtvingad expression av Cdc6, aktivering av Cdk4/6 och Cdk2 tillräcklig för förankringsfri entré av cell cykelns S fas? Vi försöker konstruera en modell för att besvara denna fråga, på så sätt att vi kan detektera problem vid modellering av stora kontroll nätverk, diskutera implikationer och möjliga lösningar. Vår modell är baserad på 1447 reaktioner och innehåller 1343 olika molekyler. Vi använde graf teori för att studera dess topologi och gjorde följande fynd: Nätverket är skalfritt och avtar enligt en potensfunktion, som var väntat baserat på tidigare arbeten. Nätverket består av ett stort väl förenat kluster. Det kan inte bli modulariserat i form av starka komponenter eller block i en användbar form. Detta eftersom vi fann en stor komponent eller ett stort block som innehöll majoriteten av alla molekyler och mer än hundra små komponenter eller block med en eller några molekyler. Vårt nätverk stämmer inte överens med en hierarkisk nätverks modell bestående av block förenade av cut-vertices. / We have identified flexible exchange and storage of data in databases, together with prolonged investment in different existing and future modelling formalisms as key issues in successful understanding of the regulatory network responsible for the connection between geno- and phenotype. This pilot study of modelling of large-scale regulatory networks starts with a medically interesting question from molecular cell biology: Is enforced expression of Cdc6, activation of Cdk4/6 and Cdk2 sufficient for anchorage-independent entry of the S phase of the cell cycle? We try to construct a model for answering this question, in such a way that we can reveal obstacles of large-scale regulatory modelling, discuss their implications and possible solutions. Our model is based on 1447 reactions and contains 1343 different molecules. We used graph theory to study its topology and made the following findings: The network is scale-free and decays as a power-law, as could be expected based on earlier works. The network consists of one huge well connected cluster. It cannot be modularised into strong components or blocks in a useful way, since we get one big component or block containing a majority of all molecules and more than a hundred tiny components or blocks with one or a few molecules. Our network does not agree with a hierarchical network model consisting of blocks linked by cut-vertices.

Breast cancer represents one of the most common forms of female malignancy of the 21st century. Among the various breast cancer subtypes, triple-negative breast cancer (TNBC) is phenotypic of breast tumors lacking expression of the estrogen receptor (ER), the progesterone receptor (PR) and the human epidermal growth factor receptor 2 (HER2). As an idiosyncratic disease, TNBC displays a conspicuously aggressive and invasive clinical course, with an unexplained partiality towards women of African ancestry. Its acute heterogeneity and complexity behave as mutually reinforcing negative factors, which further complicate prognosis, thereby increasing the burden of breast cancer-related mortality. With the absence of well-defined molecular targets in TNBC, there is a heightened reliance on tri-modality therapy (surgery, radiotherapy and chemotherapy), albeit with an increasing incidence of adverse effects and disease relapse. To this end, there is an urgent need to develop an arsenal of targeted diagnostics and therapeutics, which can be synergized to cover the vast majority of triple-negative breast tumors, paving the way towards the development of personalized regimens suitable for the particular needs and disease of each patient. As such, achieving selective cytotoxicity, with minimal or no collateral damage to healthy tissues, embodies the holy grail of targeted anti-cancer therapies. For instance, the high affinity and specificity of monoclonal antibodies (mAbs) and derivatives thereof, have cemented their application as revolutionary tools in the selective delivery of drugs to malignant cells. These therapeutic proteins, also known as antibody-drug conjugates (ADCs), might exhibit several advantages compared to their small-molecule counterparts, but their widespread clinical use is hampered by various developmental considerations. Traditional conjugation strategies employed to arm mAbs with cytotoxic warheads, usually give rise to heterogeneous mixtures of ADC species, bearing non-uniform drug-to-antibody ratios (DARs), pharmacologic characteristics and safety profiles. Fortunately, the implementation of self-labeling tags (such as SNAP-tag, CLIP-tag and Halo-tag) are providing renewed impetus to homogeneous recombinant immunotherapeutics development. More precisely, SNAP-tag is an engineered mutant of the human O(6)-alkylguanine-DNA alkyltransferase, endowed with the ability to specifically and irreversibly react with benzylguanine (BG) derivatives, forming a stable product. Based on the above premises, this research aims to use SNAP-tag technology as a cutting-edge site-specific conjugation method to: (1) develop a comprehensive antibody platform, consisting of single-chain antibody fragments (scFvs) genetically fused to SNAP-tag, to specifically screen and evaluate their predictive potential for chondroitin sulfate proteoglycan 4 (CSPG4), CD44 and aspartate (aspartyl/asparaginyl) β-hydroxylase (ASPH)-positive TNBC cells, (2) generate functional recombinant ADC formulations as robust delivery systems carrying the antimitotic drug monomethyl auristatin F (MMAF/AURIF), concurrently overcoming production constraints to yield therapeutically viable and homogeneous combination products, and (3) provide a fail-safe system that overcomes the lack of specificity of photodynamic therapy (PDT), by coupling scFv-SNAP-tag to the potent near-infrared (NIR) photosensitizer (PS) IRDye700DX (IR700) and to demonstrate its selective dose-dependent cytotoxic activities in vitro. Following in silico design of the open reading frames (ORFs) coding for each construct, standardized molecular cloning techniques were implemented to generate recombinant mammalian expression plasmids, encompassing Ig-Kappa leader as an efficient protein secretory system. After confirmation of the DNA integrity, protein expression was achieved through transient transfection into HEK293T cells. Thereafter, the resulting histidine-tagged fusion proteins (αCSPG4(scFv)-SNAP, αCD44(scFv)-SNAPf and αASPH(scFv)-SNAP) were harvested from the cell culture supernatant and subjected to immobilized metal affinity chromatography (IMAC). In order to evaluate the outcome of this protein expression and purification step, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot analysis were used to confirm the presence of full-length recombinant SNAP-tag based fusion proteins based on their molecular weights. Integration of the fluorescent dye Alexa Fluor 488 into the fusion proteins was carried out to investigate the self-labeling activity of the SNAP-tag moiety, as well as to provide qualitative and quantitative insights into the binding potential of the antibody fragments towards their cognate antigens. Subsequently, the AURIF and IR700-based immunoconjugates were generated by conjugating scFv-SNAP with their respective BG-modified substrates, in a defined 1:1 stoichiometric reaction. The specific and dose-dependent biological activities of the resulting bifunctional therapeutic proteins were then assessed on TNBC cells. In this study, pCB-αCD44(scFv)-SNAPf was successfully cloned and all 3 fusion proteins were effectively expressed, although with low yields and purity, yet adequate for downstream in vitro characterization. After showcasing the self-labeling potential of the SNAP-tag component, surface binding of the fluorescently labeled product was demonstrated on antigen-positive TNBC cell lines through confocal microscopy and flow cytometry. The cell killing ability of the novel AURIF-based recombinant ADCs and IR700 photoimmunoconjugates, was illustrated by the induction of a 50% reduction in cell viability (IC50 value) at nanomolar to micromolar concentrations on target cell lines. This observable selective cytotoxicity revealed that conjugation of BG derivatives to SNAP-tag, did not affect the binding potential of the antibody fragment, nor abrogated the cytocidal activity of the payload. As a proof of concept, this research builds on existing work that promulgates the use of SNAP-tag as a state-of-the-art conjugation strategy that can circumvent the challenges associated with the use of antibodies as effective delivery systems for therapeutic molecules. By harnessing the applicability of SNAP-tag in the unambiguous generation of homogeneous and pharmaceutically acceptable immunoconjugates, the results herein presented, also highlight the prospects of such agents in disease-specific tumor suppression. While various architectural modifications could further improve cytotoxic activities of future combination products, this research underscores the duality of SNAP-tag in the development of immunodiagnostics and therapeutics, that could potentially be instrumental in instilling a shift towards a personalized medicine stratagem. In conclusion, the combination of such immunoconjugates with a robust companion diagnostic panel provided by SNAP-tag, represents a first step towards the effective management of TNBC, with potential impact on the economic, social and clinical settings.

This chapter presents the Infobiotics Workbench (IBW), an integrated software suite developed for computational systems biology. The tool is built upon stochastic P systems, a probabilistic extension of P systems, as modelling framework. The platform utilises computer-aided modelling and analysis of biological systems through simulation, verification and optimisation. IBW allows modelling and analysing not only cell level behaviour, but also multi-compartmental population dynamics. This enables comparing be tween macroscopic and mesoscopic interpretations of molecular interaction networks and investigating temporo-spatial phenomena in multicellular systems. These capabilities make IBW a useful, coherent and comprehensive in silico tool for systems biology research.

The discovery of genetic regulatory networks was an important advancement in science. Not only do they help understand how organisms behave but the development of synthetic genetic networks has aided in other fields of science and industry. Many genetic networks have been modelled deterministically by using differential equations to provide an insight into the network's behaviour. However, within a biological environment, a certain degree of intrinsic noise should be expected and the robustness of these networks should be tested. Creating and analysing a genetic network in a biological environment can be a time consuming task so applying stochastic methods, such as the Gillespie Algorithm, to a computer model will provide an important, initial insight into the behaviour of the system. One interesting genetic network is the coupled repressilator due to its relatively simplistic design and the broad, multistable dynamics it offers. The inhomogeneous solutions that it can yield are particularly interesting as they may help explain certain biological phenomena, and may be used as a tool to assist with further research into genetic networks. In this thesis, the Gillespie Algorithm will be applied to the coupled repressilator so that its robustness can be tested. Biologically feasible modifications will be made to the system to produce much more stable and predictable dynamics so that the broad range of solutions can exist within a noisy environment. The methods developed will take into account previously made assumptions and potential errors in biological data so that they can be applied to other genetic system. One further objective in this thesis is to explore computational limitations that may occur when modelling large, stochastic networks. Issues such as rounding errors and dealing with very small and very large numbers were encountered and methods to circumvent these without sacrificing computational run-time will be developed and applied.