The assembly of colloidal-sized particles into larger structures by the manipulation of inter-particle forces has been a subject of significant research towards applications in materials science, soft matter physics, and synthetic biology. To date, much of this work has utilized manipulation of electrostatic or depletion interactions to drive the aggregation of the particles. More recently, specific (bio)-chemical interactions have been harnessed, particularly the use of deoxyribonucleic acid (DNA) linkers to program particle interactions by Watson-Crick base-pairing. In this thesis, we will demonstrate the use of an alternative set of biochemical interactions, protein-protein interactions, which have useful properties (in particular, their ability to be completely genetically-programmable).
\r\n\r\nIn Chapter 2, we discuss the development of a model system for the protein-mediated assembly of colloidal micro-particles. Associative proteins are grafted onto the surface of polystyrene micro-particles, enabling their assembly into aggregates either through reversible coiled-coil interactions or by irreversible isopeptide linkages. The sizes of the resulting aggregates are tunable and can be controlled by the concentration of the immobilized associative proteins on their surface. Further, we show that particles grafted with different protein pairs show excellent self-sorting into separate aggregates. Finally, we demonstrate that these protein-protein interactions can be used to assemble complex core-shell aggregates. The principles of protein-mediated colloidal assembly learned in this chapter will be instructive as we attempt the more complex assembly of living microbial cells.
\r\n\r\nIn Chapter 3, we discuss the implementation of a protein-driven aggregation system in living bacterial cells. Similarly to Chapter 2, we demonstrate that we can drive the aggregation of bacteria by the surface display of proteins enabling reversible coiled-coil interactions or irreversible isopeptide bonds. The sizes of these aggregates are tunable by titration of surface expression levels by standard synthetic biology techniques. Finally, we show that this programmable aggregation of bacteria may have physiological consequences for the cells, in particular, the activation of a quorum sensing circuit due to a higher local concentration of bacteria.
\r\n\r\nIn Chapter 4, we further investigate how the properties of the aggregates described in Chapter 3 can be controlled and how these relate to the underlying properties of the associative proteins and shear field. we demonstrate control of the assembly kinetics and equilibrium sizes of the resulting flocs over several orders of magnitude using different associating proteins and expression levels. Finally, we show that a single point mutation in the associative protein leads to an unexpected ultra-sensitive pH-responsive coil, demonstrating the importance of molecular-scale interactions on the macro-scale properties of the aggregates.
\r\n\r\nIn Chapter 5, we discuss the ability of the bacterial aggregates described in Chapters 3 and 4 to enable substrate channeling between bacterial strains, leading to enhancement of titers in multi-step biosynthetic pathways. When biosynthetic pathways are split into separate bacterial strains, dilution of the intermediate compound into the bulk media may decrease reaction flux. By aggregating the bacteria, the intermediate compound is able to rapidly diffuse into the downstream cell without being diluted, enabling higher reaction fluxes. we demonstrate through the model flavonoid synthesis pathway that aggregation can lead to substantially higher titers of the desired compound without pathway re-engineering, and develop a mathematical model by which this result can be understood.
" }, { "name": "Vyatskikh, Andrey", "degree": "PhD", "year": "2020", "title": "Additive Manufacturing of 3D Nano-Architected Metals and Ceramics", "advisor": "Greer, Julia R.", "url": "https://resolver.caltech.edu/CaltechTHESIS:05252020-134146453", "creators": [ { "name": { "family": "Vyatskikh", "given": "Andrey" }, "id": "Vyatskikh-Andrey", "orcid": "0000-0002-6917-6931", "display_name": "Vyatskikh, Andrey" } ], "advisors": [ { "name": { "family": "Greer", "given": "Julia R." }, "id": "Greer-J-R", "role": "advisor", "display_name": "Greer, Julia R." } ], "committee": [ { "name": { "family": "Shapiro", "given": "Mikhail G." }, "id": "Shapiro-M-G", "role": "chair", "display_name": "Shapiro, Mikhail G." }, { "name": { "family": "Faber", "given": "Katherine T." }, "id": "Faber-K-T", "role": "member", "display_name": "Faber, Katherine T." }, { "name": { "family": "Gao", "given": "Wei" }, "id": "Gao-Wei", "role": "member", "display_name": "Gao, Wei" }, { "name": { "family": "Greer", "given": "Julia R." }, "id": "Greer-J-R", "role": "member", "display_name": "Greer, Julia R." } ], "option_major": [ "medeng" ], "doi": "10.7907/pdz2-dd59", "abstract": "Additive manufacturing (AM) represents a set of manufacturing processes that create complex 3D parts out of polymers, metals, and ceramics. AM of metals and ceramics is widely used to produce parts for aerospace, automotive, and medical applications. At the micro- and nano-scales, AM is poised to become the enabling technology for efficient 3D microelectromechanical systems (MEMS), 3D micro-battery electrodes, 3D electrically small antennae, micro-optical components, and photonics. Today, the minimum feature size for most commercially available metal and ceramic AM is limited to ~20-50 \u03bcm. Currently, no established processes can reliably produce complex 3D metal and ceramic parts with sub-micron features.
\r\n\r\nIn this thesis, we first demonstrate a nanoscale metal AM process that can produce ~300 nm features out of nanocrystalline, nanoporous nickel using synthesized hybrid organic-inorganic materials, two-photon lithography, and pyrolysis. We study microstructure and mechanical properties of as-fabricated nickel architectures and compare their structural strength to established AM processes. We then show how this process can be extended to other metals and metalloids, including Mg, Ge, Si, and Ti.
\r\n\r\nThis study extends further into nanoscale AM of transparent, high refractive index materials for micro-optics and photonic crystals. We develop an AM process to 3D print fully dense nanocrystalline rutile titanium dioxide (TiO\u2082) with feature dimensions down to ~120 nm. We carefully study and model the relationship between feature dimensions and process parameters to achieve a <2% variation in critical dimensions. We then use this understanding of the process to fabricate and study 3D dielectric photonic crystals with a full photonic bandgap in the infrared.
\r\n\r\nFinally, a microscale AM process of titanium dioxide is demonstrated for photocatalytic water treatment. We show how synthesized hybrid organic-inorganic materials can be applied for stereolithography to print TiO\u2082 architectures with 100 \u03bcm features. We use the developed 3D printing process to investigate the effect of 3D architecture on the efficiency of photocatalytic water treatment.
\r\n\r\nThis work establishes a versatile and efficient pathway to create three-dimensional nano-architected metals and ceramics and to investigate their properties for applications in 3D MEMS, micro-optics, photonics, and photocatalysis.
\r\n" }, { "name": "Yang, Kevin Kaichuang", "degree": "PhD", "year": "2019", "title": "Probabilistic Protein Engineering", "advisor": "Arnold, Frances Hamilton", "url": "https://resolver.caltech.edu/CaltechTHESIS:12222018-173706714", "creators": [ { "name": { "family": "Yang", "given": "Kevin Kaichuang" }, "id": "Yang-Kevin-Kaichuang", "orcid": "0000-0001-9045-6826", "display_name": "Yang, Kevin Kaichuang" } ], "advisors": [ { "name": { "family": "Arnold", "given": "Frances Hamilton" }, "id": "Arnold-F-H", "orcid": "0000-0002-4027-364X", "role": "advisor", "display_name": "Arnold, Frances Hamilton" } ], "committee": [ { "name": { "family": "Clemons", "given": "William M." }, "id": "Clemons-W-M", "orcid": "0000-0002-0021-889X", "role": "chair", "display_name": "Clemons, William M." }, { "name": { "family": "Yue", "given": "Yisong" }, "id": "Yue-Yisong", "orcid": "0000-0001-9127-1989", "role": "member", "display_name": "Yue, Yisong" }, { "name": { "family": "Tirrell", "given": "David A." }, "id": "Tirrell-D-A", "orcid": "0000-0003-3175-4596", "role": "member", "display_name": "Tirrell, David A." }, { "name": { "family": "Arnold", "given": "Frances Hamilton" }, "id": "Arnold-F-H", "orcid": "0000-0002-4027-364X", "role": "member", "display_name": "Arnold, Frances Hamilton" } ], "option_major": [ "chemeng" ], "doi": "10.7907/ZR4M-K630", "abstract": "Machine learning-guided protein engineering is a new paradigm that enables the optimization of complex protein functions. Machine-learning methods use data to predict protein function without requiring a detailed model of the underlying physics or biological pathways. They accelerate protein engineering by learning from information contained in all measured variants and using it to select variants that are likely to be improved. We begin with a review of the basics of machine learning with a focus on applications to protein engineering and protein sequence-function datasets (Chapter 1). We used the entire machine-learning guided engineering paradigm to engineer the algal-derived light-gated channel channelrhodopsin (ChR), which can be used to modulate neuronal activity with light. We build models that discover ChRs with strong plasma membrane localization in mammalian cells (Chapter 2) and unprecedented light sensitivity and photocurrents for optogenetic applications (Chapter 3). Machine learning-guided evolution requires a machine-learning model that learns the relationship between sequence and function. For machine-learning models to learn about protein sequences, protein sequences must be represented as vectors or matrices of numbers. How each protein sequence is represented determines what can be learned. We learn continuous vector encodings of sequences from patterns in unlabeled sequences (Chapter 4). Learned encodings are low-dimensional, do not require alignments, and may improve performance by transferring information in unlabeled sequences to specific prediction tasks. Alternately, we demonstrate an interpretable Gaussian process kernel tailored to biological sequences (Chapter 6). In addition to a model to predict function from sequence, engineering requires a method to use the model to choose sequences for the next round of evolution. Most machine-learning guided engineering strategies assume that selected sequences can be queried directly. However, in directed evolution it is common to design a library of sequences and then sample stochastic batches from that library. We propose a batched stochastic Bayesian optimization algorithm for iteratively designing and screening site-saturation mutagenesis libraries (Chapter 5).
" }, { "name": "Zhang, Ruijie", "degree": "PhD", "year": "2019", "title": "Engineering Heme Proteins for Olefin and Carbon\u2212Hydrogen Bond Functionalization Reactions", "advisor": "Arnold, Frances Hamilton", "url": "https://resolver.caltech.edu/CaltechTHESIS:03102019-234035587", "creators": [ { "name": { "family": "Zhang", "given": "Ruijie" }, "id": "Zhang-Ruijie", "orcid": "0000-0002-7251-5527", "display_name": "Zhang, Ruijie" } ], "advisors": [ { "name": { "family": "Arnold", "given": "Frances Hamilton" }, "id": "Arnold-F-H", "orcid": "0000-0002-4027-364X", "role": "advisor", "display_name": "Arnold, Frances Hamilton" } ], "committee": [ { "name": { "family": "Fu", "given": "Gregory C." }, "id": "Fu-G-C", "orcid": "0000-0002-0927-680X", "role": "chair", "display_name": "Fu, Gregory C." }, { "name": { "family": "Shan", "given": "Shu-ou" }, "id": "Shan-Shu-ou", "orcid": "0000-0002-6526-1733", "role": "member", "display_name": "Shan, Shu-ou" }, { "name": { "family": "Stoltz", "given": "Brian M." }, "id": "Stoltz-B-M", "orcid": "0000-0001-9837-1528", "role": "member", "display_name": "Stoltz, Brian M." }, { "name": { "family": "Arnold", "given": "Frances Hamilton" }, "id": "Arnold-F-H", "orcid": "0000-0002-4027-364X", "role": "member", "display_name": "Arnold, Frances Hamilton" } ], "option_major": [ "chemistry" ], "doi": "10.7907/2076-CX23", "abstract": "One of the most important challenges in chemistry is the creation of new catalysts. Nature excels at this: constructed from biologically available elements, enzymes are versatile catalysts which adapt quickly to changing environments in order to sustain life. The combination of adaptable proteins with abiological reagents from synthetic chemistry affords a new direction for catalyst development. This thesis describes new enzymes, derived from a cytochrome P450 monooxygenase, which catalyze nitrogen and carbon atom transfer reactions to olefins and carbon\u2212hydrogen bonds. Chapter 1 introduces directed evolution, a strategy for the laboratory optimization of proteins, in the context of improving metalloproteins for their native catalysis or for new reactions. Chapter 2 details the development of an enzyme-catalyzed transformation of olefins to aziridines, a valuable motif which is both present in bioactive molecules and used as a versatile building block for synthesis. This study establishes that when provided the appropriate reagents (e.g. styrenes and tosyl azide), heme proteins can adopt a nitrene transfer catalytic cycle to form aziridine products and that the turnover and selectivity of the catalyst can be optimized through mutation of its amino acid sequence. The activity of heme protein catalysts is extended to the functionalization of sp3 hybridized C\u2212H bonds for carbon\u2013nitrogen and carbon\u2013carbon bond formation through nitrene and carbene insertion respectively (Chapters 3 and 4). With the exception of C\u2212H oxygenation chemistry, iron complexes are under-utilized for sp3 C\u2212H functionalization reactions, despite iron being readily available and non-toxic. Combining previously engineered heme proteins with suitable substrates led to initial reaction discovery. Directed evolution of these enzymes significantly improved their C\u2212H functionalization activity (by 140-fold in Chapter 4). Characterization of evolved enzymes, including the attainment of an X-ray crystal structure (Chapter 3) and substrate scope studies (Chapters 3 and 4), were pursued. In sum, the thesis work addresses both the biological question of expanding the catalytic capabilities of existing enzymes through mutation and expands the chemistry of iron-porphyrin catalysts." } ]