Browsing by Subject "Protein Engineering"

Now showing 1 - 6 of 6
  • Thumbnail Image
    Item
    Data Driven Approach to Engineering Protein Evolvability and Developability
    (2021-08) Golinski, Alexander
    Proteins can be engineered to perform a variety of functions ranging from diagnostics and therapeutics to industrial and commercial enzymes. The ability to computationally evaluate the performance of a protein from its amino acid sequence would increase the efficiency of discovery, expanding the impact of engineered proteins. However, the problem is plagued by the immensity, complexity, and barrenness of the amino acid sequence-function landscape. The following research is focused on predicting two nontraditional protein functions: 1) Evolvability - the ability to generate novel functionality based upon the mutation of a subset of amino acid positions, and 2) Developability - the ability to be efficiently manufactured and maintain primary functionality. Limited prior understanding of these functions was available across broad swaths of sequence space. This work advanced a hybrid experimental/computational platform to provide broad and deep experimental data on sequence-function relationship. Empowered by data analytics, the dataset enabled accurate predictions and provided mechanistic insight regarding protein evolvability and developability. The first story aimed to determine which computable biophysical properties drive evolvability. Utilizing high-throughput screens for evolving specific molecular targeting, the performance of seventeen protein scaffolds were obtained for seven molecular targets. A model predicting evolvability from biophysical properties was trained, with a focus on generalizability and interpretability. Achieving a 4/6 true positive rate, a 9/11 negative predictive value, and a 4/6 positive predictive value, the predictive model analysis suggests a large, disconnected paratope (location of sequence variation) will permit evolved binding function. The second story aimed to generate a model to predict protein developability, as determined by bacterial production, from amino acid sequence. As traditional metrics of developability are often capacity limited (10^2 - 10^3), a set of three of high-throughput (10^5) assays were created to generate a sufficient dataset. The relevance of the assays to traditional metrics was certified by a model that predicts expression from assay performance 35% closer to the experimental variance and trains 80% more efficiently than a model predicting from sequence information alone. The validated assays offer the ability to identify developable proteins at unprecedented scales, reducing a bottleneck of protein commercialization. Neural networks were trained to generate a numeric developability representation (embedding) for each sequence from the high-throughput dataset and transfer the embedding to predict recombinant expression. Mimicking protein theory, our deep-learning model convolves machine-learned amino acid properties to predict expression 42% closer to the experimental variance compared to a traditional approach. Analysis of trained numeric encodings of the amino acids highlights the unique capability of cysteine, the importance of hydrophobicity and charge, and unimportance of aromaticity when aiming to improve developability of the protein scaffold Gp2. The completion of the studies supports the hypothesis that data-driven protein engineering can both accurately predict protein evolvability and developability while also providing meaningful insight into the properties driving functionality. The success of this approach is predicted to increase significantly as the capacity to parametrize protein function continues to grow. The research presents the increased ability to engineer proteins across their diverse sequence landscape using modern experimental techniques and data analytics.
  • Thumbnail Image
    Item
    Engineered Proteins for Studying and Controlling Cellular Recognition
    (2018-08) Csizmar, Clifford
    The ability to direct cell-cell interactions has tremendous value in several therapeutic fields. While genetically-encoded artificial receptors have proven efficacious, their scope is limited by the genetic engineering that underlies the approach. To circumvent some of these limitations, our group has developed a non-genetic method to modify any cell surface with a targeted protein scaffold. First, we engineered a protein ligand based upon the human tenth type III fibronectin domain (Fn3) that binds to epithelial cell adhesion molecule (EpCAM), an overexpressed tumor antigen. Using yeast surface display, mammalian cell panning, and a novel titratable avidity-reduction selection technique, we evolved Fn3 clones exhibiting high affinity and robust selectivity for cellular EpCAM. We then incorporated these Fn3s into a multivalent chemically self-assembled nanoring (CSAN). EpCAM-targeted CSANs were anchored to cell membranes through the hydrophobic insertion of phospholipids into the lipid bilayer. The targeting elements were subsequently removed from the cell surface by disassembling the CSAN with the antibiotic, trimethoprim. Using this system, we successfully directed and reversed targeted intercellular interactions in vitro. Finally, the modular CSANs were used to study how avidity impacts the apparent affinity of a multivalent scaffold. By tuning the number of Fn3 domains on the CSAN, we quantitatively described how the apparent affinity changes as a function of ligand affinity, domain valency, and antigen expression density. These results informed the development of a CSAN capable of discriminating between cells expressing different quantities of EpCAM both in vitro and in vivo. In conclusion, we developed a diverse toolkit for directing and studying cell-cell interactions. The CSAN platform is applicable to several therapeutic arenas and, by balancing affinity and avidity, may offer advantages over current cell-directing methods.
  • Thumbnail Image
    Item
    Engineering a 45-Amino Acid Protein Scaffold for Molecular Cancer Imaging
    (2017-12) Kruziki, Max
    Cancer is the second leading cause of death in the United States. Molecularly targeted cancer treatments, including monoclonal antibodies and kinase inhibitors, exhibit strong performance on a small subset of patients but are inconsistent due to tumor heterogeneity. Biopsy-based genetic and protein tumor characterization provide value but cannot address spatial or temporal variations in heterogeneity. Non-invasive methods, such as molecular imaging, to characterize cancer cells will allow for easier patient stratification and treatment monitoring. Currently, molecular imaging is limited by the modest availability of quality probes that efficiently distribute throughout the body and quantitatively localize at the site of the cancer biomarker. Engineering effective diagnostic molecular probes would provide a substantial advance in cancer characterization and personalized medicine. Protein scaffolds, which comprise a large stabilizing framework and a randomized region onto which binding interactions can be engineered, offer an efficient platform for probe engineering. More broadly, engineered binding proteins are useful in many aspects of biotechnology and medicine. In this thesis, we mined ~100,000 known protein topologies to identify candidate small protein scaffolds. We developed the 45-amino acid Gp2scaffold and evolved multiple Gp2 variants that strongly (as strong as 0.2 nM) and specifically (greater than 50:1 target:control) bind their respective target while also retaining high thermal stability (65-80 ºC thermal desaturation midpoint) . A Gp2 variant that was evolved to bind with strong affinity to epidermal growth factor receptor (EGFR), a cell surface biomarker overexpressed in multiple cancer types, was more thoroughly investigated in pre-clinical studies. This variant exhibited strong (18 nM), selective binding, and was passive on normal EGFR signaling pathways, which is important to reduce off-target side effects. PET imaging of subcutaneously xenografted tumors in mice revealed effective probe localization to EGFR-high tumors while low signal was observed in EGFR-low tumors and from non-targeted control Gp2. Gp2 evolution was studied by comparing the efficacy of different combinatorial library amino acid diversity based on high throughput sequencing data, natural Gp2 homologs, structural data, and computed stability. Multiple library designs elucidated amino acid diversity that was beneficial or detrimental in different sections of the Gp2 protein, and will aid future evolution and developability of Gp2. From these libraries, high affinity Gp2 variants targeting an additional clinically-relevant cancer biomarker, programmed death-ligand 1 (PD-L1), were evolved, isolated, and characterized. The advancements outlined here make important contributions towards improving both protein engineering tools and methodologies as well as diagnostic imaging tools.
  • Thumbnail Image
    Item
    Engineering a CD19-Based Bispecific Molecule for CAR T Cell Therapy
    (2018-10) Schrack, Ian
    Cancer is a profoundly devastating disease both globally and within the United States. Current standards of care for treating cancer often includes surgical resection, chemotherapy, and radiation, each of which associates with its own set of adversities. An emerging class of treatment, immunotherapy, aims to utilize a patient’s own functional immune system as the therapeutic agent. Adoptive T cell therapy, but more specifically, chimeric antigen receptor (CAR) expressing T cell transfer, has had notable clinical success particularly against hematological malignancies. Chimeric antigen receptors are synthetic immunoreceptors which can redirect T cells towards varying tumor associated antigens, and, as a living cell, have the aptitude to develop sustainable memory and anti-tumor efficacy. However, conventional CAR T cells lack clinical modularity afforded by other treatments because, once transfused into a patient, the modified immune cell cannot be further altered. This nuance has resulted in several adverse side-effects which can be lethal to a subset of patients. Several resolutions have been posed to solve these reported complications, one of which is genetic encoding CAR specificity towards a secondary, bispecific molecule. This split-CAR approach has the propensity to improve antigen specificity, resolve antigen loss, afford dose-able T cell activation, and more. However, while many bispecific molecules have been developed, many lack both tunability and developability, both of which are important for the complexities and ever-changing nature of cancer. To meet this demand for engineered ligands, several high-throughput ligand selection methods have been developed for discovering ligands with a desired specificity. Furthermore, the associated CAR T cells may have poor aptitude for activation and expansion due to insufficient antigen availability. Conversely, conventional anti-CD19 CAR T cells can harness both healthy or malignant CD19-positive B cells for activation and expansion and thus have an abundance of available antigen. To these points, we utilized yeast-surface display and directed evolution as a pipeline for developing an CD19-based bispecific molecule capable of harnessing the proliferative aptitude of anti-CD19 CAR T cells to target antigens conventionally associated with solid tumors. Human CD19 was evolved for improved structural integrity through conformational selections using anti-CD19 monoclonal antibodies. Improved mutants were sequenced and provided input for designing a stably expressing, generation 2 CD19 library (termed Frame2). The second-generation diversity applied experimentally determined, beneficial mutations in multi-site fashion to drive the enhanced CD19 framework towards a higher stability and/or functionality. The Frame2 CD19 library was constructed as several fusion constructs containing either an anti-EGFR fibronectin domain or an anti-HER2 scFv in both N-terminal and C-terminal orientations and selectively evolved with anti-CD19 antibodies and the ligand-respective antigen. A set of functional bispecific CD19-ligand fusions were successfully developed. In theory, because the anti-CD19 antibodies used for fusion development have an identical binding domain to several anti-CD19 CAR constructs, these fusions should be detectable by CD19-targetted CAR T cells. Moreover, if the ligand domain also retains specificity, the CD19-ligand bispecific molecule should be capable of redirecting anti-CD19 CAR T cells to EGFR or HER2 expressing tumors.
  • Thumbnail Image
    Item
    Rational improvement of the activity of a nitrite reductase model
    (2014-05) Strange, Jacob
    Nitrite reductase (NiR) is a bacterial enzyme that catalyzes the one electron reduction of nitrite (NO2-) to nitric oxide (NO). Through site-directed PCR mutagenesis, variants of the electron transfer protein azurin (See figure) were rationally designed to mimic Nitrite Reductase (NiR) in an attempt to increase our knowledge about enzyme function and design. Several variants were created by incorporation of a Type 2 copper center on the surface of the protein. Further mutations were added to aid in electron transfer as well as to match the potential of the Type 1 copper in azurin to that of the Type 1 copper in NiR. Each variant was characterized using EPR, UV-Vis and mass spectroscopy. Finally, we characterized each variant using Michaelis-Menten kinetics. This activity was determined using the Griess assay, which allows us to spectrophotometrically quantify the amount of nitrite in our reaction. We compared the activity of our NiR models. The results show that the variant designed for faster electron transfer did not increase the activity of our enzyme, but did increase activity when added with the other mutations. The two variants designed to decrease the reduction potential of the Type 1 copper site were found to increase the activity by themselves, but decrease the activity when combined.
  • Thumbnail Image
    Item
    Synthesis of carotenoid-containing polyesters and protein engineering to improve synthetic efficiency of Pseudomonas fluorescens esterase.
    (2009-10) Jiang, Yun
    The conjugated double bonds in carotenoids central chain make these compounds potential building blocks for conductive polymers. Bixin and crocetin's capability of forming carboxylester bonds at end are used to synthesis polyesters. Among 20 esterases and lipases, Candida antarctica Lipase B (CalB) is the only enzyme that catalyzes bixin esterification reaction with linear alcohols. None of these enzyme catalyzes crocetin esterification reaction. Candida antarctica Lipase B added alcohols, such as, n-propanol, polyethylene glycol 400 and 1, 10- decanediol to bixin to make bixin diesters, which worked as initiators and were added into "-caprolactone to synthesis polyesters. The corresponding polymers had the MW 10,600, 12,900 and 11,100. Esterases/lipases and some acyltransferases share a Ser-His-Asp catalytic triad and similar mechanisms. Comparison of the x-ray structures of these structurally related esterases/lipases with acyltransferases reveals a different conformation of the oxyanion loop. In esterases/lipases this loop adopts a type II ! turn conformation with C=O of the main chain facing the active site. In acyltransferases this loop adopts a type I ! turn with the N–H of the main chain facing the active site. The x-ray crystal structure of Pseudomonas fluorescens esterase containing a sulfonate transition state analog shows the C=O facing active site activate via a bridging water molecule to the attacking water molecule. While in acyltransferases an opposing interaction with the N– H may deactivate the attacking water molecule. Oxyanion turn GWLL of Pseudomonas fluorescens esterase (PFE) was engineered in order to switch to a type I # turn and favor acyltransfer reaction. Replacing GWLL with acyltransferases oxyanion turn peptides yielded several mutants with only PFE-GLRA soluble. However PFE-GLRA doesn't show activity. Saturation mutagenesis at position W28 and L29, yielded mutants L29I, L29T, L29V, L29W, with acyltransfer/hydrolysis (A/H) ratios 2.2, 2.5, 2.5, 4 fold of that of the wild type. Some of these mutants might contain type I # turns.