Poster Abstracts

Sponsor: Agilent

Premier Laboratory Partner for a Better World Analytical scientists and clinical researchers worldwide rely on Agilent to help fulfill their most complex laboratory demands. Our instruments, software, services and consumables address the full range of scientific and laboratory management needs—so our customers can do what they do best: improve the world around us.

Whether a laboratory is engaged in environmental testing, academic research, medical diagnostics, pharmaceuticals, petrochemicals or food testing, Agilent provides laboratory solutions to meet their full spectrum of needs. We work closely with customers to help address global trends that impact human health and the environment, and to anticipate future scientific needs. Our solutions improve the efficiency of the entire laboratory, from sample prep to data interpretation and management.

Sponsor: Inscripta

Inscripta is a life science technology company enabling scientists to solve some of today’s most pressing challenges with the first benchtop system for genome editing. The company’s automated Onyx™ platform, consisting of an instrument, consumables, assays, and software, makes CRISPR-based genome engineering accessible to any research lab.

Contributed posters:

#2 A Modular DNA Cloning Toolkit for Actinobacteria

Zachary Jansen, Rice

Actinobacteria are an important phylum of environmental bacteria and include numerous industrially relevant and pathogenic species. Genetic manipulation in Actinobacteria can be difficult due to the lack of genetic tools and the size, complexity, and characteristically high GC content of their genomes. A modular cloning system for Actinobacteria will facilitate synthetic biology in these organisms and advance efforts in metabolic engineering and bioremediation. We have domesticated a large set of plasmid origins from several species within the order Mycobacteriales, which contains the genera Rhodococcus, Nocardia, and Gordonia, known for their diverse metabolic activity, as well as Mycobacterium, an important pathogen of humans and livestock. We also have developed a set of modular selectable markers, constitutive and inducible promoters, and bicistronic ribosome binding sites with a range of strengths to provide synthetic biologists with a robust toolkit to control transcription and translation in these species. These parts have been standardized for Golden Gate assembly of multigene circuits and are currently being evaluated in 20 non-model species.

#3 MINE 2.0: Enhanced biochemical coverage for peak identification in untargeted metabolomics

Jonathan Strutz, Northwestern University

Although advances in untargeted metabolomics have made it possible to gather data on thousands of cellular metabolites in parallel, identification of novel metabolites from these datasets remains challenging. To address this need, Metabolic in silico Network Expansions (MINEs) were developed. A MINE is an expansion of known biochemistry which can be used as a list of potential novel structures for unannotated metabolomics peaks. MINE 1.0 outperforms tools like PubChem for novel metabolite identification. However, MINE 1.0 is still limited by the small set of reaction rules used. Here, we present MINE 2.0, which builds on MINE 1.0 by utilizing more known biochemical transformations resulting in increased database size and significantly increased coverage of unannotated peaks in untargeted metabolomics datasets.

#4 Activating silent biosynthetic gene clusters in Streptomyces spp. with CRISPRi tools

Andrea Ameruoso, Rice University

Antibiotic-resistant bacteria are a looming threat for healthcare, and illustrate the need for the discovery of novel antimicrobial agents. The bacterial genus Streptomyces represents a promising resource, as it is the most prominent producer of antimicrobial natural products. This ability lies in the expression of enzymatic pathways encoded by genomic regions called Biosynthetic Gene Clusters (BGCs). BGCs are rarely expressed under laboratory culturing conditions (i.e. are silent), often due to the repressive action of gene regulatory networks. Therefore, many BGCs are uncharacterized and a potential source of new compounds. Here we describe a novel approach to activate silent BGCs by using CRISPR tools. First, we develop a CRISPR interference (CRISPRi) transcriptional repressor to relieve repressive regulation and activate natural product synthesis. We optimize its repression capability by tuning the expression levels of its components, and then harness it to inhibit the activity of known BGC repressors and successfully induce production of natural products. Secondly, we establish a novel CRISPR activation system (CRISPRa) to activate transcription. Together, this work provides a new approach to activate BGC expression in Streptomyces. This strategy circumvents the need for time-consuming genome engineering efforts and only requires straightforward transformation of easily designable plasmids, thus providing a simple method for BGC activation.

#5 Enhancing cell-free biosynthesis with metabolically rewired yeast extracts

Blake Rasor, Northwestern University

Biological systems present appealing opportunities for sustainable chemical synthesis, but cellular metabolic engineering faces challenges including slow design cycles, competition for cellular resources, and low volumetric productivities. Alternatively, recent developments in cell-free systems offer the potential for multifaceted reaction optimization and biosynthesis in the absence of cell growth to complement strain engineering techniques for pathway prototyping and biomanufacturing. Here we describe an integrated in vivo / in vitro metabolic engineering approach to convert glucose into value-added chemical products using crude cell extracts derived from a suite of flux-enhanced Saccharomyces cerevisiae strains. As an initial model, we explored 2,3-butanediol (BDO) synthesis. First, we assessed pathway performance in metabolically active yeast extracts with and without the heterologous BDO pathway. By generating cell extracts from wildtype and metabolically rewired strains using multiplexed CRISPR-dCas9 modulation, we observed that cell-free systems reflect the metabolic potential of the respective source strain. Second, we systematically optimized the reaction environment (e.g., cofactors, pH) and strain background to improve titers by 40% and achieve volumetric productivities of 10.4 ± 0.2 mM/h. Finally, this approach was extended to itaconic acid and glycerol biosynthesis to demonstrate enhanced cell-free biosynthesis with a low-flux, heterologous pathway and a high-flux, native pathway, respectively. Looking forward, we expect the combined cellular and cell-free metabolic engineering approach to open new opportunities for synthetic biology prototyping efforts and cell-free biomanufacturing.

#6 Production of long-chain fatty alcohols in Yarrowia lipolytica W29 Cancelled

Payman Tohidifar, University of Illinois at Urbana-Champaign

#7 Engineering a programmable split-ribozyme platform for detecting RNA in living cells

Lauren Gambill, Rice University

Intracellular RNA detection technologies represent a powerful, yet relatively underexplored, area of synthetic biology. With the detection of viral infection, disease states, and antibiotic resistance critical to our ability to combat global health crises, the development of tools for RNA detection in cells is more important than ever. While a few technologies that link RNA signals to biomolecular outputs have emerged, the field has yet to converge on a tool that is easily customized to different endogenous RNAs and that can be paired with virtually any biomolecular output of choice. To address this, we have developed a novel plug-and-play RNA detection platform. This platform uses a synthetic split ribozyme, which produces an mRNA output in response to the presence of a user-specified RNA target input. Because sensing is achieved through base-pairing interactions between the target RNA and guide RNAs that are appended onto the split ribozyme, sensors with different targets can be easily designed. To engineer this system, I first identified the optimal ribozyme split site that yields a high-dynamic range of sensing. Here, we observed no detectable output protein expression in the absence of input RNA. I then established guide RNA design rules that allow for more predictable design of new sensors, and developed a modular approach whereby any protein output can be rapidly exchanged for another without system re-engineering. Finally, I used this system to sense endogenous RNAs, thereby creating sensors for dynamic detection of cellular state. Future work will focus on improving generalizability by implementing this platform in different cell types and using it for diagnostic and therapeutic applications in biotechnology.

#8 Tools for Biorenewable Chemical Production from Pyrolysate By Lactic Acid Bacteria

Swastik Sen, Iowa State

Lignocellulosic biomass is an abundant source of carbon, especially in agricultural areas like Iowa. Fast pyrolysis is one method that can be used to create fermentable sugars from cellulosic sources, but product of this process, bio-oil, is rich in (1) levoglucosan, an anhydrosugar form of glucose, and (2) phenolic and other inhibitory compounds. Here we report the engineering of Lactococcus lactis to utilize purified levoglucosan for the production of enantiomerically pure L-lactic acid and, by expressing a heterologous three-gene pathway, 1,2-propanediol. We also report media conditions that allow the production of these chemicals from diluted bio-oil and strategies to engineer tolerance to bio-oil and other "dirty" substrates. Finally, we report the development of CRISPR-assisted genome-scale engineering and library creation tools in this important industrial organism.

#9 Siderophore-Mediated Growth Dynamics of Escherichia coli and Acinetobacter baylyi ADP1

Kevin Fitzgerald, Northwestern University

Microbial contamination, a concern traditionally addressed with unit process optimization, has long been known to complicate industrial fermentations and decrease product yields. Although techniques exist to mitigate contamination, they are generally expensive, reducing profit margins. This challenge will likely be exacerbated in biorefineries capable of utilizing strategic, cost-negative feedstocks (municipal wastewater, agricultural waste, etc.). Siderophores are small-molecule iron chelators secreted by organisms in iron-limited conditions. By binding to and increasing the bioavailability of insoluble ferric iron, these molecules have been shown to play a significant role in the community dynamics of soil, marine, and human microbiomes. Here we present our efforts to investigate the interaction between genetically tractable organisms, Escherichia coli DH5a and Acinetobacter baylyi ADP1, and their respective siderophores, enterobactin and fimsbactin. This work undergirds the development of a modular, product-independent form of coculture and the mitigation of microbial contamination in industrial fermentations.

#10 Optogenetic tools for public goods control in Saccharomyces cerevisiae

Neydis Moreno, University of Wisconsin-Madison

Microorganisms live in dense and diverse communities, with interactions between cells guiding community development and phenotype. The ability to perturb specific intercellular interactions in space and time provides a powerful route to determining the critical interactions and design rules for microbial communities. Approaches using optogenetic tools to modulate these interactions offer promise, as light can be exquisitely controlled in space and time. We report new plasmids for rapid integration of an optogenetic system into Saccharomyces cerevisiae to engineer light-control of expression of a gene of interest. In a proof-of-principle study, we demonstrate the ability to control a model cooperative interaction, namely the expression of the enzyme invertase (SUC2) which allows S. cerevisiae to hydrolyze sucrose and utilize it as a carbon source. We demonstrate that the strength of this cooperative interaction can be tuned in space and time by modulating light intensity and through spatial control of illumination. Spatial control of light allows cooperators and cheaters to be spatially segregated, and we show that the interplay between cooperative and inhibitory interactions in space can lead to pattern formation. Our strategy can be applied to achieve spatiotemporal control of expression of a gene of interest in Saccharomyces cerevisiae to perturb both intercellular and interspecies interactions.

#11 Data-Driven Protein Engineering with Latent Space Generative Models

Xinran Lian, University of Chicago

Natural proteins have evolved by natural selection to perform particular biological functions essential to life processes. As result, only a small portion of a natural protein family can perform a specific function. Can we find out constraint on the functionality, and design artificial proteins to enlarge this dataset? We answered these questions by machine learning approaches, which present a means to determine these "design rules" by analyzing databases of functional protein sequences and then exploiting these rules to engineer synthetic proteins with targeted functions. In an application to Sho1-SH3, a small domain maintaining in the context of the osmotic stress signaling system that occurs in budding yeast, we (i) employed variational autoencoders (VAEs) as a deep neural network architecture to learn the design rules within a low-dimensional latent space projection of functional protein sequences, (ii) generatively sampled from this latent space to design synthetic sequences predicted to possess high and low function, and (iii) built and tested libraries of synthetic genes using a next-generation sequencing-based selective assay. By sampling globally and locally from latent space of the VAE, we designed hundreds of functional Sho1 domains from thousands of SH3 homologs with diverse functions. Our findings reveal that functionality can be represented by locality in the VAE latent space, and demonstrate the potential for data-driven design to rationally engineer novel protein sequences.

#12 Production of Propionate, a Neuroprotectant, in Probiotic E. coli Nissle

Maple Chen, Iowa State University

Parkinson’s Disease (PD) is a neurodegenerative disease that affects millions of people globally and confers symptoms including motor deficits and dementia. Recent studies have shown that propionate, a short chain fatty acid, can help ameliorate motor deficits in mice with PD when present in the gut. Others have correlated PD to decreased levels of propionate and other short chain fatty acids in stool, suggesting that PD patients experience some physiological symptoms that decrease gut propionate levels.

Thus, a potential route for treating the motor deficits associated with PD is to introduce propionate into the gut environment by employing propionate-producing probiotics, or microbes that live in the gut. Although bacteria in the genus Propionibacteria can synthesize propionate through native pathways, they grow slowly and are not very hardy, making them less than optimal probiotic candidates. Instead, genetic engineering tools enable us to introduce propionate synthesis genes into established probiotic bacteria such as Escherichia coli Nissle 1917 (EcN), which has been used to combat ulcerative colitis and inflammatory bowel disease, among other afflictions. In this study, we construct four propionate synthesis genes into a stable EcN plasmid and detect propionate through gas chromatography mass spectroscopy, resulting in propionate titres on the order of 100 mg/L. We take advantage of pMUT2, a stable plasmid maintained by EcN, to house the biosynthesis genes, avoiding the need to manipulate the bacteria’s genomic DNA while achieving higher copy numbers of the genes, allowing higher expression. We hope to eventually transfer this study to mouse models to demonstrate the physiological effects of our engineered propionate-producing probiotic strain.

#13 Activating Integrases in vitro via Cell-Free Protein Synthesis

Lauren Clark, Northwestern University

Phage integrase proteins offer an exciting tool for DNA manipulation in both an in vitro and an in vivo context. Previously, integrase proteins such as Cre and Bxb1 have been used as molecular biology tools, but with some limits in application and efficiency. Our goal is to prototype integrases for the manipulation of novel bacterial species within mixed microbial communities. Here, we present our initial efforts to express a panel of integrase proteins from different phage/host phylogeny using E. coli extracts. In addition, we assess their efficiency in converting their target DNA sequences in an in vitro context.

#14 Engineering Medium Chain Fatty Acid Production in Synechococcus sp. PCC 7002

Joshua Abraham, University of Wisconsin, Madison

Producing target chemicals sustainably in microbes is challenging because the requirement for plant-derived sugars by heterotrophic organisms competes for arable land and freshwater that could otherwise be used for food production. Photosynthetic organisms like prokaryotic cyanobacteria can directly capture CO2 to generate desired products. In particular, the model strain Synechococcus sp. PCC 7002 (henceforth PCC7002) is a well-studied strain of cyanobacteria that can grow quickly under high light and in brackish waters, making it an ideal microbe to grow on non-arable land that does not compete for food production. PCC7002 has established genetic tools and is thus an attractive replacement host for heterotrophic production.

Medium chain fatty acids (MCFAs) are an appealing class of chemicals to produce for their high value and usability as precursors for pharmaceuticals, fuels, flavors, and fragrances. Competitive production of MCFAs in heterotrophic Escherichia coli has been enabled by protein engineering of fatty acid biosynthesis (FAB)-terminating thioesterase enzymes. In this project, we generated analogous strains of PCC7002 that express these heterologous thioesterase proteins and produce up to 860 mg/L of octanoic acid (C8). We demonstrate the functionality of an organic overlay to partition free fatty acids from aqueous media and subsequently increase titers. Our future plans are use stable isotope (13C) based nonstationary metabolic flux analysis (INST-MFA) to inform additional genetic modifications that will alleviate metabolic bottlenecks in central metabolism and FAB.

#15 High-throughput discovery of the determinants for stress-induced aggregation of de novo proteins

Cydney Martell, Northwestern

Protein aggregation from physical stresses (i.e. temperature, freeze thaw cycles, and shaking) during manufacturing, storage, and transport is a major challenge in engineering proteins for therapeutic applications. Creating effective protein scaffolds requires designs to balance biophysical criteria for stable folding alongside requirements for reducing aggregation under stress. This is a difficult challenge because both the principles behind protein folding and aggregation behavior are not well understood. To address this challenge, I am developing a mass spectrometry-based assay to quantitatively measure the aggregation propensity after stress exposure for thousands of proteins in a parallel. I will use this high-throughput approach to measure the aggregation of 10,000 de novo designed proteins when exposed to different stresses, such as, high temperature and low pH. I will then integrate the experimental aggregation data with sequence and structural properties of the proteins to develop a quantitative model to predict aggregation. Given the biophysical features identified and my model for aggregation prediction, I will rationally design new proteins and alter the de novo designs used in the first iteration of experimentation to be more resistant to aggregation. Studying these datasets of highly and extremely aggregation resistant de novo designed proteins will allow me to draw conclusions about the protein sequence and structure features that are important for distinguishing these high levels of resistance. I will repeat this cycle of experimental high-throughput aggregation assay, machine learning analysis, and protein design to robustly test and improve my aggregation prediction model. The quantitative model developed from this data will enable rational design of de novo proteins with unprecedented resistance to stress, which could greatly increase the applications of designed protein therapeutics.

#16 A TEM-1 based binding affinity assay for selection of peptide-peptide interactions

Devon Kulhanek, Rice University

Protein-protein interactions are central to many biomolecular processes such as transcriptional activation, receptor signaling, enzyme or substrate recruitment, and antibody-antigen affinity. Directed evolution techniques capable of discriminating between differing binding affinities of the interacting partners are critical to tools for engineering and customizing these interactions. Here we present a binding affinity selection assay wherein the beta-lactamase, TEM-1, was destabilized by circularly permuting the gene and removing a key stabilizing motif. The altered protein requires sufficient binding between specified terminal fusion partners to rescue function and confer antibiotic resistance. We have confirmed that this selection can be used to identify binary assay outputs, and further work will determine the dynamic range and steric or sequence limitations of the system. This technique has the potential to engineer fusion partners directly, e.g. nanobodies, or indirectly, via enzymatic activity capable of linking the termini in trans, such as orthogonal translational machinery which enable the incorporation of amino acid chemistries capable of forming covalent crosslinks.

#17 An Enzyme Self-Amplification System for Ultrasensitive Detection of Biomarkers at the Point of Care

Catherine Majors, Northwestern University

Rapid, inexpensive detection of biomarkers at the point of care is vital for many clinical purposes. However, limitations in current detection platforms have prevented the sensitive detection of many protein and small molecule biomarkers; specifically, there has been a lack of innovation for methods to amplify signal from the detection of low concentration proteins and small molecules at the point of care. Biology, on the other hand, has evolved intricate mechanisms for rapidly amplifying protein signals in vivo. Towards the goal of developing rapid, ultrasensitive diagnostics, we have developed a novel assay for detecting low concentration protein and small molecule analytes via an in vitro, protein-based signaling network incorporating self-amplifying enzymatic pathways. We incorporate two mechanisms of protein signaling networks: split enzyme reconstitution and autocatalytic positive feedback loops. We have experimentally validated analyte detection via the simultaneously binding two fusion proteins (i.e. a sandwich assay in solution), bringing two halves of a split enzyme together to produce product, which is detected via a FRET-based biosensor. Additionally, we have designed an autocatalytic feedback loop that responds to enzymatic product by producing more product; in ODE-based simulations, the assay demonstrates ultrasensitive, bistable detection of analyte that is tunable over several orders of magnitude. This system will be broadly applicable for protein and small molecule detection and could be used to detect a wide range of clinical target analytes with known binding domains.

#18 Comparative omics analysis of Saccharomyces cerevisiae strains with different isobutanol pathways

Francesca Gambacorta, University of Wisconsin- Madison

Saccharomyces cerevisiae, the workhorse for industrial fermentations, has been identified as a suitable host for producing the next-generation biofuel, isobutanol. In this experiment, we aimed to understand how two different metabolic engineering strategies, compartmentalization and cofactor balancing, could affect the performance and physiology of an isobutanol producing strain. In our work, we equipped yeast with isobutanol cassettes which had either a mitochondrial or cytosolic compartmentalized isobutanol pathway and used either a cofactor imbalanced (NADPH-dependent) or cofactor balanced (NADH-dependent) KARI enzyme; a multiomics analysis was then performed on the engineered strains to elucidate the functional differences between them.

We found that pathway compartmentalization had a greater effect on isobutanol production than cofactor balancing. The strain harboring the mitochondrial localized pathway outperformed the cytosolic version by 3.7-fold, while the strain harboring the cofactor imbalanced pathway outperformed the cofactor balanced pathway by 1.5-fold. Furthermore, the proteome of the strains harboring the cytosolic localized pathways was significantly altered; some of the most dramatically altered proteins were involved in sulfur-related pathways (sulfate assimilation and cysteine/homoserine/siroheme biosynthesis). Since one of the enzymes in the isobutanol pathway, dihydroxy-acid dehydratase, requires a 2Fe-2S cluster, we hypothesize the altered sulfur metabolism is related to this cluster requirement. Our current objective is to use the knowledge gained from this multiomics study to inform the design of the next isobutanol producing strain.

#19 Engineering a Minimal Selenocysteine Biosynthesis Pathway

Andrew Gilmour, Rice University

Selenocysteine, often considered the 21st amino acid, is a rare, naturally occurring amino acid found in proteins throughout all domains of life. Similar to thiol groups in the related amino acid cysteine, selenol groups can form covalent diselenide bonds which are an important motif for protein engineering. Replacing disulfide bonds with diselenide bonds can be used to increase protein stability, which has particular importance for recombinant protein therapeutics. However, the native biosynthesis and incorporation pathways for selenocysteine are complex and often result in low yields or competing byproducts. Previous engineering efforts have overcome some of these issues, but inefficient biosynthesis and competing incorporation of serine, the immediate biosynthetic precursor remains a challenge. Our work focuses on the two key enzymes required for production of the mature selenylated tRNA (Sec-tRNASec), selenophosphate synthase, SelD, and selenocysteine synthase, SelA. Our primary goal is to identify which enzyme is rate limiting during selenocysteine biosynthesis. This will be accomplished using in vivo fluorescent assays which rely on an engineered selenocysteine-dependent fluorescent reporter (Sec-smURFP). After identifying the rate limiting step, efforts will be focused on improving the kinetic parameters of the enzyme.

#20 Rational engineering of a modular CRISPRa platform with expanded target range

Maria Claudia Villegas Kcam, Rice University

The ability to control gene expression is a transformative capability that allows to uncover gene function or induce valuable phenotypes and cell fates important for applications in biotechnology and biomedicine. CRISPR-Cas regulators have provided an excellent toolbox for this purpose. Among these, CRISPR activator (CRISPRa) systems turn on transcription through localization of activation domains (AD). However, current bacterial CRISPRa systems still present several challenges. First, different bacterial ADs offer distinct regulatory properties, meaning different AD are often required depending on the target gene or desired application. Second, the bacterial CRISPRa systems need to be targeted to sites located a specific number of nucleotides upstream of the promoter to achieve activation, which critically limits its application in non-synthetic targets. To address these limitations, we adopt a protein engineering approach to create a highly modular and target flexible bacterial CRISPRa system. We identified a novel and strong AD which we harness to explore AD recruitment strategies. Specifically, we evaluated the use of modular protein-protein interaction domains which allows different AD to be encoded on independent and easily exchangeable plasmid elements. Additionally, we evaluated the use of circularly permuted variants of the dCas9 protein and the CRISPR Type IE system to identify variants capable of activating gene expression from distinct positions previously identified as non-activating sites and therefore expand the functional target range. Our modular CRISPRa system provides a more flexible and versatile tool that will expand our ability to activate and control bacterial genomes.

#21 Towards dynamic and complex RNA circuits: A novel approach to address retroactivity with positive feedback insulation

Baiyang Liu, Rice University

The creation of de novo designed RNA transcriptional regulatory systems with high dynamic range and orthogonality is allowing for the creation of RNA-based circuits that process signals solely at the transcriptional level. While much progress has been made, challenge remains to construct more complex and dynamic circuitry analogous to the ones achieved using protein-based regulators. One of the major challenges is high retroactivity seen in RNA regulatory systems. For example, in small RNA activating systems called STARs, a regulator RNA is irreversibly consumed for each regulatory event. To address this, we propose a novel approach to overcome retroactivity by implementing a positive-feedback insulation strategy that in essence compensates for the loss of regulatory RNAs. We present here a theoretical analysis of our strategy, called RNA compensator, and demonstrate that it can effectively mitigate the retroactivity not only in multi-state RNA circuits, e.g., self-activator and toggle switches, but also in dynamic RNA circuits like oscillators. In addition, we improve the connectivity between RNA modules by reshaping their input-output responses using tandem STAR and RNA sequestration. Finally, I will present our recent work investigating the compatibility of STAR in multiple species and creating broad-host RNA transcriptional regulators that use host-agnostic phage-derived RNA polymerases. Combining the works above, a toolset is developed to create modular, robust, and broad-host RNA circuits, providing a new platform for future synthetic RNA research.

#22 Directing the metabolism of allergy protective Lachnospiraceae with novel genetic tools

Jack Arnold, University of Chicago

The number of life-threatening food allergy cases reported in the Western world have increased dramatically in recent years, and this rise in prevalence has been linked to changes in the microbial diversity of the human gut microbiome. Commensal microbes of the gut, such as butyrate-producing Lachnospiraceae, are thought to play protective roles in the gut, including inducing regulatory T cells and promoting epithelial barrier integrity. However, our explicit understanding of how microbes affect these regulatory mechanisms remains elusive. This knowledge gap is due to insufficient molecular manipulation technologies to control the metabolic activity of the microbes involved in allergic responses. We seek to develop novel molecular tools that to investigate the relationships between microbial function and host immune responses. Establishing such tools would allow for direct genotype to phenotype analyses of important allergy-related gut symbionts. Additionally, with new insight into the basic genetic mechanisms of these microbes, we can both genetically augment living therapeutics and uncover novel microbial metabolites to treat food allergies. Here, we propose a blueprint to enable genetic engineering in currently intractable symbionts of the Lachnospiraceae family. To this end, preliminary efforts to co-opt plasmid-based systems have provided basic genetic engineering capabilities in Lachnospiraceae. These tools serve as a prototyping platform for the developing a robust genetic toolkit while enabling precise manipulation of gene expression in Lachnospiraceae. We will then employ these tools to regulate the production of potentially beneficial Lachnospiraceae metabolites. The development of these tools will aid precise genetic study of many key symbiotic microorganisms, provide a basis to genetically manipulate other intractable microbes, and allow for tuning microbial function in defined contexts to investigate hypotheses that have long remained beyond our reach.

#24 Synthetic prebiotics for control of microbial dynamics

Sanjeeva Kumar Murali, Iowa State University

The synbiotic relationship between human milk oligosaccharides (HMOs) and probiotic Bifidobacteria exemplifies prebiotic control of microbial community dynamics. Inspired by this example, we have engineered the well-known probiotic, E. coli Nissle 1917, to metabolize HMOs and used this metabolism to control population dynamics and protein expression in mixed cultures of E. coli. We accomplish this using a unique whole-cell biosensor which provides linkage-specific, quantitative detection of various HMOs (Enam and Mansell, Cell Chemical Biology, 2018). Addition of these complex substrates to synthetic microbial consortia orthogonally controls growth rate or protein expression of particular strains. In this work, we seek to expand the range of available prebiotics via a two-fold strategy: (1) chemical modification of existing privileged substrates and (2) engineering probiotic strains to utilize these modified compounds. Several strategies are discussed, including genome mining for anabolic and catabolic enzymes and from the marine pangenome.

#25 Variety is the splice of life: High-throughput topological engineering of splicing ribozymes for new switches and sensors

August Staubus, Rice University

Topological engineering, including splitting and domain insertion, has provided a robust approach for generating biomolecular switches. For example, a protein’s activity can be coupled to a ligand-binding event by splitting the protein and fusing a ligand-sensing domain to each half. Alternatively, ligand sensing domains can be directly inserted into a proteins’ structure, producing in functional changes in response to a ligand-binding event. The power of topological modifications for generating synthetic biology components is augmented when combined with high throughput approaches to generating and screening these (often large) design spaces. While these techniques have been widely successful in engineering ligand-dependent proteins, their application to other biopolymers, such as RNAs, has not been investigated. Here, we report results from the first study implementing a high-throughput approach to the generation and screening of topologically engineered RNAs, focusing on splicing ribozymes. We used in vitro transposon mutagenesis combined with fluorescence activated cell sorting (FACS) to screen large libraries of topologically engineered ribozyme variants in a single one-pot experiment. This approach identified library variants amenable to splitting and domain insertion, which we then leveraged to generate novel RNA-based sensors for small molecules and host RNAs in vivo. These results demonstrate that powerful protein engineering techniques can be generalized to RNA, and provide an effective new method for generating novel RNA-based devices for synthetic biology.

#26 Characterizing assembly and function of the Type Three Secretion System following different modes of activation

Julie Ming Liang, Northwestern University

With the advent of recombinant DNA technology, proteins can now be made in large-scale as therapeutics to treat diseases, enzymes for industrial processes, and even biomaterials like spider silk. Unfortunately, industrial-scale protein production can be inefficient as extensive purification is required in order to extract and purify the protein-of-interest (POI) from the production organism. We address this issue by engineering the Salmonella Pathogenicity Island-1 Type Three Secretion System (SPI-1 T3SS), a protein secretion system, to secrete POIs directly into the culture media.

The SPI-1 T3SS is natively used by Salmonella to invade and infect host organisms and its expression is governed by several different transcriptional regulators. Activation of T3SS can be induced by synthetically controlling these regulators. We are studying how activation of the T3SS by various induction mechanisms influence (1) SPI-1 T3SS gene expression, (2) assembly of the T3SS needle apparatuses, and (3) functional output in the form of protein secretion titer. This will expand our understanding of the steps between T3SS activation and invasion and will enable us to better engineer the T3SS for heterologous protein secretion.

#27 Engineering non-growth metabolism in Escherichia coli for improved biochemical synthesis

Rashmi Raj, Northwestern University

With the increasing flexibility of molecular biology tools, the ability to produce value-added chemicals is greater than ever. Synthetic biology is greatly applicable in this aspect as it can be used to sustainably produce many chemicals of interest. However, for these environmentally friendly chemical production methods to be commercially competitive with traditional chemical processes, substantial engineering of host native metabolism and the introduction of heterologous enzymes and pathways is required. This can often result in a heavy cellular burden and undesirable lowered product yields and variable production titers.

Instead of pathway correction through numerous rounds of genetic editing, overexpression and engineering changes, an alternative strategy would be to make strains more robust by removing pathway fragility. This can be achieved by attenuating the stress response which encompasses both sensors and effectors. However, past efforts to modulate the stress response have shown that sensor deletion results in dysfunctional RNA and protein synthesis. To circumvent this, mainly the effectors are targeted, particularly the type II toxin-antitoxin pairs and ribosome hibernation proteins. Removal of these systems could prevent disruption of translation and thus increase overall product yield without excessively changing the large and complex stringent stress-sensing network.

The deletion of these factors has resulted in the increased production of chemicals of interest, namely observed through over 19-fold increase in limonene titers as compared to wildtype production strains. A finding of interest is that engineered strains appear to outperform established commercial production strains in minimal media fermentation. Using minimal media over traditional media for industrial production could have great cost advantages and the engineered strains can be utilised in the production of a wide range of chemicals of interest.

#28 Spatial Organization in Yeast Cell via Enzyme Co-localization

Katelyn Leisman, Northwestern University

Yeast is one of the oldest domesticated organisms, today supporting nearly one trillion dollars of the US economy through its key role in baking, fermentation, and biotechnology. Moreover, yeast is a model organism that is ideal for studying how cells work, with lots of carry over to other species. Thus, there is much economic and scientific interest in making the metabolic pathways of yeast more efficient. Experiments show that spatial organization of certain enzymes involved in the reactions may impact cells' efficiency. Specifically, yeast metabolism depends on a network of reactions, each catalyzed by an enzyme; each reaction's efficiency can be described by a quantity called the Gibbs Free Energy (written as $\Delta G$). By modeling the way this $\Delta G$ changes due to different concentrations of the substances involved in the reaction (called metabolites) and the spatial organization of the enzymes, we can potentially determine configurations that increase the cells' metabolic efficiency. In this poster, I will discuss modeling the spatial organization of a few enzymes, and share some preliminary results of connecting this to the optimized efficiency of the pathway.

#29 Microbially Synthesized Polymeric Amyloid Fiber Promotes β-nanocrystal Formation and Displays Gigapascal Tensile Strength

Jingyao Li, Washington University in St. Louis

Among natural macroscopic materials, dragline spider silk stands out as one of the strongest and toughest materials that has motivated decades of research to produce recombinant silk fibers at macroscales. Despite these efforts, achieving gigapascal tensile strength with higher than 150 MJ/m3 toughness has proven to be extremely difficult for these artificial silk fibers. On the other hand, the ability of amyloid proteins to form stable β-sheet nanofibrils has made them potential candidates for material innovation in nanotechnology. However, such nano-scale feature has rarely translated into attractive macroscopic properties for mechanically-demanding applications. Here we present a strategy that combines the mechanical superiorities of spider silk and amyloid nanofibrils by fusing amyloid peptides with flexible linkers from spidroin, the resulting polymeric amyloid proteins can be biosynthesized using engineered microbes and wet-spun into macroscopic fibers. Using this strategy, fibers from three different amyloid groups were fabricated. Structural analyses unveil the presence of β-nanocrystals that resemble the cross-β structure of amyloid nanofibrils. These polymeric amyloid fibers have displayed strong and molecular-weight-dependent mechanical properties. Fibers made of a protein polymer containing 128 repeats of the FGAILSS sequence displayed an average ultimate tensile strength of 0.98 ± 0.08 GPa and an average toughness of 161 ± 26 MJ/m3, surpassing most recombinant protein fibers and even some natural spider silk fibers. The design strategy and the biosynthetic approach can be expanded to create numerous functional materials, and the macroscopic amyloid fibers will enable a wide range of mechanically-demanding applications.

#30 A low-cost, thermostable, cell-free protein synthesis platform for on-demand production of vaccines Cancelled

Katherine Warfel, Northwestern University

#31 Mobile translation systems generate genomically-engineered Escherichia coli cells with improved growth phenotypes

Samuel Gowland, Northwestern University

Cellular translation is responsible for the synthesis of proteins, a highly-diverse class of macromolecules comprised of chains of amino acid residues that form the basis of biological function. In Escherichia coli, engineering of the biomolecular components of translation, such as ribosomes, transfer RNAs (tRNAs) and aminoacyl-tRNA synthetases, have allowed efficient introduction of non-canonical amino acids into proteins. However, the engineering potential of molecular translation is hampered by the limited capabilities for rapidly sampling the large genomic space necessary to evolve well-coordinated synthetic translation networks inside cells. To address this limitation, we developed a new genome engineering method driven by mobile translation elements. This mobilization method utilizes the stochastic action of recombinase flippase (FLP) to generate up to 400 million genomic insertions, deletions, or rearrangements at targeted FRT sites per mL culture per OD in living E. coli cells. As a model, we applied our approach to evolve faster-growing E. coli strains living exclusively off genomically-expressed tethered ribosomes. In an iterative “pulse-passaging scheme,” we generated genomic libraries of cells via induction of FLP recombinase (pulse) followed by passaging the population without induction of FLP to enrich the resulting population for cells with higher fitness. We observed large structural genomic diversity across these cells, with the fastest growing strains exhibiting a 71% increase in growth rate compared to the ancestral strain. We anticipate both these chassis strains and the mobilization method will be useful tools for synthetic biology efforts to engineer molecular translation systems.

#32 Investigating the impact of cofactor availability and enzyme composition on metabolic pathway performance in bacterial microcompartments

Charlotte Abrahamson, Northwestern University

Bacterial microcompartments (MCPs) are large, inducible, proteinaceous organelles that encapsulate the enzymes, substrates, and cofactors required for specific metabolic pathways in certain species of bacteria. MCPs allow for the spatial organization of pathways containing volatile or toxic intermediates that would be detrimental to the cell if produced at the same levels in the cytosol. One example of MCPs is the 1,2-propanediol utilization (Pdu) MCP found in Salmonella enterica, which sequesters the 1,2-propanediol degradation pathway, including the toxic intermediate propionaldehyde. This sequestration is desirable to repurpose for other heterologous pathways that may suffer from similar inefficiencies, but important questions about how exactly the compartments function to enhance pathway performance must be answered before compartmentalization can be widely used in metabolic engineering strategies. One function of MCPs that is hypothesized to improve pathway performance is the creation of a private encapsulated cofactor pool that is internally recycled. By purifying and then studying the MCPs in vitro using bacterial extracts, we can evaluate how model pathways are impacted by controlling the compartment environment directly, including external presence or absence of required cofactors, and addition or removal of pathway and recycling enzymes internally. This method allows for greater study of conditions that would normally not be accessible using in vivo methods.

#33 Leveraging lipid-protein interactions to sort membrane proteins into synthetic membranes Cancelled

Justin Peruzzi, Northwestern University

#34 Community science designed ribosomes with beneficial phenotypes

Antje Kruger, Northwestern University

Functional design of ribosomes with mutant ribosomal RNA (rRNA) could expand opportunities for understanding molecular translation, building cells from the bottom-up, and engineering ribosomes with altered capabilities. However, such efforts have been challenging due to viability constraints, a gigantic rRNA sequence space and a poor understanding of the relationship of rRNA sequence, folding, and function. To address these challenges, we developed an integrated community science and experimental screening approach for the rational design of rRNAs. This approach connects Eterna, an online video game that crowdsources RNA sequence design to citizen scientists in the form of puzzles based on secondary structure energetics, with in vitro ribosome synthesis, assembly, and translation to transform the traditional design-build-test-learn approach. We discovered mutant 16S and 23S rRNA sequences that improve ribosome performance under diverse environmental conditions in vitro and in vivo. This work provides new insights into ribosome rRNA sequence-function relationships, with implications for synthetic biology.

#35 Heterologous gene expression in Paraburkholderia sacchari for enhanced biodegradable plastic production.

Dianna Long, University of Nebraska-Lincoln

Generating useful chemicals from biological platforms is an environmentally responsible alternative to non-sustainable sources. One such chemical is polyhydroxybutyrate (PHB); a component of biodegradable plastic. A bacterium of interest for production of PHB is Paraburkholderia sacchari LMG 19450 LFM101. This non-model bacterium is a relatively high efficiency producer of PHB that utilizes C5 and C6 sugars, making it an attractive candidate for PHB production from cheap and complex feedstocks. In its native form, P. sacchari cannot efficiently produce enough PHB for the process to be economically feasible, however, metabolic engineering strategies could rectify this shortcoming. Here we share our expansion on synthetic biology tools through use of the RSF origin of replication in P. sacchari and expand on metabolic engineering strategies by demonstrating heterologous gene expression from another PHB producing bacterium, Cupriavidus necator. We demonstrate that the RSF origin of replication is a candidate for plasmid maintenance within P. sacchari through successful transformation and expression of fluorescent protein, providing another origin of replication option in addition to the previously used pBBR1-oriV. We have also been able to demonstrate increased production of PHB from P. sacchari by incorporating a set of genes from C. necator (phaA, phaB, phaC, and bktb) via plasmid. These genes were chosen as they code for key enzymes in the PHB pathway and we hypothesized that overexpression could increase PHB production. We saw a 3.5X increase in PHB production with the P. sacchari strain containing bktb under the constitutive control of the promoter Plac. Current and future work incorporates synthetic biology tools into further exploration of how heterologous genes can improve PHB production within P. sacchari; a direction of research toward a sustainable, environmentally responsible, and economically beneficial method for creating biodegradable plastics.

#36 High resolution calcium recording with enzymatic DNA synthesis

Alec Castinado, Northwestern University

Genetically encoded molecular recorders enable nano-scale data acquisition and storage in biological systems by coupling changes in intracellular conditions to site-specific genomic modifications. Molecular recorders have succeeded in recording cellular process on the order of hours to days, but recording faster events remains challenging due to slow DNA modification processes. To overcome this limitation, we developed a molecular recorder based on the activity on the DNA synthesis activity of terminal deoxynucleotidyl transferase (TdT), an untemplated DNA polymerase. By measuring the change in TdT nucleotide selectivity due to changing environmental cation concentrations, we were able to record and decode fluctuations in Co2+ concentration with minutes resolution in vitro. In order to expand the applicability of this system to physiological inputs, we fused TdT to the natural calcium sensor Calmodulin, producing a TdT with reduced activity in the presence of calcium. We leveraged the altered nucleotide selectivity of the engineered sensor TdT to develop a two-polymerase recording system that encodes calcium concentration through the differential activity of the sensor TdT and a calcium-insensitive reference TdT. Using this two-polymerase system, we recorded the timing of calcium perturbations in vitro to within 1 minute over a 60-minute recording. By demonstrating a system capable of recording physiologically relevant signals, this work is a significant step toward implementing polymerase-based recording in living systems.

#37 Development of a high-throughput screen for protein-protein interactions responsible for bacterial microcompartment closure

Carolyn Mills, Northwestern University

Identification of novel antibiotic targets is crucial in combatting the growing threat of antibiotic resistance. Ideal targets for antibiotic development are those that specifically exist in pathogenic bacteria, but not healthy biota. Bacterial microcompartments (MCPs) are proteinaceous organelles that specifically aid in the proliferation of enteric pathogens on niche carbon sources, and thus represent excellent candidates for targeted antibiotic development. These MCPs are delimited by a protein shell that encases an enzymatic core, where the core enzymes are responsible for carrying out metabolic processes that aid in proliferation. An abundance of protein-protein interactions (PPIs) are responsible for holding together the MCP shell; however, development of inhibitors against these interactions has been limited by knowledge of the specific PPIs involved in MCP formation as well as the low throughput of existing MCP characterization techniques. Here, we describe our discovery that closure of the 1,2-propanediol utilization (Pdu) MCP from model pathogen Salmonella enterica serovar Typhimurium LT2 is mediated by vertex protein PduN. We find that, in the absence of PduN, extended Pdu MCP structures form, leading to an elongated cell phenotype that can be rapidly screened by both flow cytometry and high-throughput microscopy. We leverage this phenotype-assembly link to explore how different point mutations disrupt PPIs responsible for MCP closure, and examine how disruption of assembly impacts S. enterica proliferation on the niche carbon source 1,2-propanediol. Further, we show that this assembly-phenotype link can be used to screen inhibitors of PPIs responsible for MCP closure, presenting the first high-throughput platform for screening against MCP assembly.

#38 Optimizing Parameter Estimation through Experimental Design and Analysis of Model Trajectories in State Space

Sasha Shirman, Northwestern University

Many biological systems and chemical networks can be described by a set of coupled differential equations with unknown or unmeasurable parameters. Estimating these unknown parameters from experimental data often involves solving a high dimensional optimization problem in which many state variables are unmeasured or hidden. Designing experiments and selecting experimental measurements that are most informative for parameter estimation is a difficult problem, particularly when the dynamics contain multiple time scales. We develop a method, informed by the mathematical model and experimental observations, to plot the model manifold in state space and determine which characteristics of the measured states are most sensitive to unknown parameters. We transform the model manifold into a “progress space” representation where progress towards steady state, not time, is the independent variable. The discovered regions of high sensitivity in progress space are used to inform experimental design. This insight is used to reduce the dimension of the observations in the parameter estimation problem, improving computational time for parameter estimation.

#39 A volatile to non-volatile signal converter for directing information flow in soil microbial consortia

Li Chieh Lu, Rice University

Microbes can be programmed to report on chemicals in soils by coupling the production of a methyl halide transferase (MHT) to a conditional promoter. However, as the volatile methyl halide (CH3X) outputs of these sensors diffuses to the soil-air interface, they are massively diluted, creating challenges for directing information flow in terrestrial matrices. To address these challenges, we have investigated whether Methylorubrum extorquens can be programmed to sense the output of gas-reporting biosensors and convert this output into a non-volatile signal. First, we found that MHT-expressing Escherichia coli can transmit a CH3X signal through a one meter vertical column of soil. Second, we found that this E. coli strain produces sufficient CH3X gas to switch on gene expression of a fluorescent protein in M. extorquens even when these strains are separated by a column of soil. Third, we engineered M. extorquens to produce short chain acyl homoserine lactone (AHL) in response to CH3X. Finally, we found that CH3X-induced AHL production in M. extorquens can be used to activate gene expression of a visual reporter in E. coli. This work represents the first demonstration of CH3X-based cell-cell communication which provides a synthetic intercellular communication system capable of functioning at longer calling distances on a centimeter-scale and with relative bioorthogonality. Ongoing studies are investigating whether microbial consortia made up of a gas-reporting biosensor, gas to AHL signal converter, and terminal receiver strains can be used in parallel to achieve control over long-distance information flow in subterranean soils and simplify the sensing of this information at the soil-air interface.

#40 Development of a high-throughput assay for measuring secreted protein titer based on a de novo fluorescent reporter

Samuel Leach, Northwestern University

Production of heterologous proteins in a microbial host can be advantageous due to their genetic tractability, fast growth, and robustness in large cultures. However, these hosts are limited because they retain the protein product intracellularly, requiring a complicated and expensive downstream purification scheme. A promising solution is to engineer the bacteria to secrete the protein product to the extracellular space, which would greatly ease downstream purification requirements. Many research groups have focused on secreting heterologous protein products through a variety of secretion pathways native to bacteria, but the titer of secreted protein is often too low for commercial interests. While many microbial engineering techniques could be used to improve secretion titer, studies are often limited by a lack of fast, quantitative, and high throughput assays for measuring secreted protein. For this purpose, we have utilized the mini fluorescence activating protein (mFAP), a small, de novo protein consisting of a single beta barrel, which reversibly binds to the substrate DFHBI to produce a fluorescent signal. In this talk, we show that this protein is readily secreted through multiple secretory pathways when fused to the relevant secretion tag, and it produces a measurable fluorescent signal in media when its substrate is added exogenously. A high-throughput plate based fluorescence assay for measuring secreted protein using this reporter can develop a full signal in minutes, can measure secreted protein amounts across three orders of magnitude, and can be used in undefined media which typically cause low signal to noise ratios. Using the Type 3 Secretion System of Salmonella typhimurium as an example, we show that this assay can accurately measure how engineering changes affect secretion titer, and we show how this assay can enable new studies which enable further improvements to secretion titer.

#41 Devising new protein engineering strategies by extrapolation of machine learning models

Sarah Fahlberg, University of Wisconsin-Madison

Proteins can be engineered for numerous applications ranging from protein therapeutics to chemical biosynthesis. However, traditional protein engineering techniques are labor, resource, and time intensive. By learning the complex relationship between sequence and function, machine learning can be used to predict the function of unobserved sequences and find improved protein variants with reduced experimental burden. Our group’s preliminary findings have previously shown that when trained on only single and double mutants, machine learning models can be extrapolated to find significantly improved protein variants ten mutations from wild-type. However, several design considerations remain unclear, including which architectures propose the fittest variants, how far models can be extrapolated to find functional variants, and whether ensembles of models increase design success. To examine these questions further, we use the B1 domain of protein G (GB1) for improved binding affinity to Immunoglobulin G (IgG) as our engineering target. We task a variety of machine learning models to produce GB1 variants with a varying number of mutations that have improved binding affinity for IgG. We synthesize and screen top predicted designs from each condition with yeast display and FACS. Our findings suggest these engineering strategies are capable of identifying highly fit sequences within a small sequence budget.

#42 Systematic Engineering of Virus-like Particles to Identify Optimal Characteristics for Nanoparticle Delivery

Bon Ikwuagwu, Northwestern University

Self-assembling protein scaffolds known as virus-like particles (VLPs) are promising delivery vehicles for diagnostic and therapeutic applications. VLPs are easily produced in a bacterial expression system, are biocompatible, and have defined chemical handles that can attach targeting ligands. However, there are some key challenges to be solved. In particular, VLPs are highly stable and must come apart upon entry to the target cell. Moreover, it is difficult to control their size relative to other delivery platforms. Here, we engineered VLPs based on the MS2 bacteriophage by combining comprehensive codon mutagenesis, one-pot selections, and high-throughput sequencing to create protein fitness maps in a strategy we termed Systematic Mutagenesis and Particle Selection (SyMAPS). Using SyMAPS, we identified several mutations that confer a desired acid instability, as well as a single point mutation that uniformly alters the size and geometry of the VLP. We were intrigued by the mutation that alters the size of the VLP because size is a known characteristic that affects accumulation in tumors. Therefore we applied similar mutations to MS2 homologs to understand if the mutation’s effects were generalizable, and went on to identify alternative mutations that would confer the same size shift. Here, we present the detailed results of this suite of studies, and discuss the role of our engineered VLPs both for theranostic applications and as tools for investigating the role of particle size on tumor penetration.

#43 Engineering a template-independent DNA polymerase for DNA information storage and biosignal recording applications

Marija Milisavljevic, Northwestern University

Terminal deoxynucleotidyl transferase (TdT) is a template-independent DNA polymerase that synthesizes single-stranded DNA (ssDNA) and is currently being explored for use in various applications, including DNA data storage, enzymatic DNA synthesis, and temporal recording of biosignals. Several unique features make TdT a compelling tool in these areas. TdT adds nucleotides to the 3’ termini of ssDNA probabilistically, with an inherent bias in its base incorporation. It can use several divalent cations for its enzymatic activity and exhibits base preference changes in response to changes in its environment. However, TdT also has properties that can be limiting. For example, TdT activity is inhibited by the presence of rNTPs and TdT has a slower base incorporation rate compared to other DNA polymerases, both limitations for in vivo applications. Therefore, it would beneficial to modulate and tailor TdT properties. However, a high-throughput screen must be established to rapidly screen mutants, as there is no obvious approach based on current knowledge of TdT structure-function relationships. Our goal is to develop an emulsion-based method for evaluating large libraries of TdT variants, where individual cells expressing unique TdT variants will be encapsulated in aqueous droplets suspended in oil. The aqueous droplets will contain all necessary components for TdT extension and recovery, including ssDNA-conjugated microbeads. Cell lysis will release the TdT into the droplet, along with the plasmid harboring the TdT gene. Active variants will extend the ssDNA and plasmids will be captured on the bead simultaneously, linking the active phenotype to the genotype. Beads will be sorted based on a fluorescent signal generated from ssDNA extension to recover active TdTs. This screen will enable TdT engineering with easy adaptation for an array of selective pressures and improve properties to overcome challenges in emerging TdT applications.

#44 Bacterial Strain Design using Active Subspaces

Andre Archer, Northwestern University

The 1,3-propanediol pathway in bacteria provides a means of utilizing crude glycerol, a waste from biodiesel production, to produce 1,3-propanediol, a common solvent used in cosmetics, printer ink, skin care products, and antifreeze. Growing bacteria on crude glycerol can assist in lowering the carbon footprint of biodiesel and 1,3-propanediol production. Bacteria expressing the 1,3-propanediol pathway convert glycerol using the dhaB1, dhaB2 and dhaT enzyme.

In collaboration with the Tullman-Ereck Lab, I develop a model of genetically modified Salmonella growing on glycerol and producing 1,3-propanediol. The bacteria are modified to produce bacterial microcompartments, small protein-based compartments. The bacterial microcompartments are the 1,3-propanediol production hubs and provide an isolated redox environment for the 1,3-propanediol pathway enzymes.

Expressing the 1,3-propanediol pathway genes at different loci varies the concentration of pathway enzymes in the microcompartment. To rank bacterial phenotypes, I extract 1,3-propanediol yield, rate of 1,3-propanediol production and toxicity level from my mathematical model of genetically modified Salmonella growing on glycerol. However, given the number of model parameters with uncertainty, evaluating the model metrics across all uncertainty ranges would be very computationally involved. In this talk, I use a sampling method that uses Active Subspaces to incorporate parameter uncertainty into the model metrics and, thus infer bacterial strain rankings over multi-dimensional parameter space while reducing the computational load.

#45 Dynamic Kinetic Modeling Tools Using Cell-Free Experiments to Predict Metabolic Network Behavior

Jacob Martin, Northwestern University

The optimization of biosynthetic production remains a challenge in metabolic engineering. This is particularly true for products made via longer heterologous pathways, which may require manual tuning of all component reactions. While cell-free systems rapidly increase the experimental throughput of testing pathway combinations, they remain complex systems and produce large amounts of difficult-to-interpret data. Toward this goal, we are developing a dynamic kinetic model to better understand this complex system and enable rapid data analysis for pathway optimization.

We are currently using this model to study butanol production via acetyl-CoA in E. coli cell-free systems. Because these experiments have exhibited complex dynamics, wherein the transient behavior of the heterologous butanol pathway interacts with core metabolism and vice versa, our model is both mechanistic and dynamic to robustly capture these experimental phenomena and predict optimal engineering solutions. However, compared to steady-state models, dynamic models have additional degrees of freedom and would require more difficult-to-obtain non-stationary flux measurements, necessitating different parameterization methods than often used for models of living cellular metabolism. To this end, we have developed a dynamic model which utilizes a variety of literature kinetic values, thermodynamic calculations, and Monte Carlo methods for parameter sampling, followed by an ensemble approach for parameter estimation. This model also simulates several phenomena unique to cell-free systems, including gas-liquid equilibrium and transient pH measurements. Our model captures several of the qualitative behaviors of the cell-free system, including some of the effects that varying butanol production has on core metabolism. Ultimately, we aim to translate these trained mechanistic parameters into models of in vivo production strains, which will accelerate model-building and product optimization workflows.

#46 Engineering BmoR as a biosensor for butyrate in the gut

Nishit Banka, University of Wisconsin-Madison

This project aims to develop BmoR, a σ54 transcription factor found in Pseudomonas Butanovora, as a biosensor for butyrate detection and regulation in the gut. We implemented a traditional directed evolution workflow followed by next-generation sequencing (NGS) and data analysis, to enable synthesis of engineered variants with improved specificity as well as sensitivity towards butyrate. We created a mutagenized library of the ligand binding region of BmoR using error-prone PCR (epPCR) and transformed them into E. Coli. To allow for high throughput screening of the library via Fluorescence-activated cell sorting (FACS), we developed a GFP based reporter assay. We analyzed the mutational frequency data of the sorted population, obtained using NGS, to discern point mutants with enhanced butyrate induction activity. We will individually test these selected variants to assess activity and choose a template for the next round of evolution. This iterative cycle of engineering and testing will allow us to generate and characterize the optimized variant for BmoR. Our future directions will involve assessing the biosensor activity in vitro as well as in vivo microbiome environments.

#47 Model-guided Design Strategies for Bioplastic Overproduction in Rhodopseudomonas palustris

Adil Alsiyabi, University of Nebraska - Lincoln

Rhodopseudomonas Palustris is a metabolically versatile Purple Non-Sulfur Bacterium (PNSB). Depending on growth conditions, R. palustris can operate on either one of the four different forms of metabolism: photoautotrophic, photoheterotrophic, chemoautotrophic, and chemoheterotrophic. R. palustris is also a facultative anaerobe, meaning it can operate both aerobically and anaerobically. Furthermore, the organism is capable of fixing nitrogen and subsequently producing hydrogen and the bioplastic precursor polyhydroxybuyrate (PHB). Recent experimental findings revealed that PHB yields in R. palustris were highly dependent on the characteristics of the utilized carbon source. PHB production significantly increased when grown on the carbon- and electron-rich lignin breakdown product p-coumarate (C9H8O3) compared to acetate when the same amount of carbon was supplied. However, the maximum yield did not improve further when grown on coniferyl alcohol (C10H12O3). To obtain a systems-level understanding of factors driving PHB yield, a model-driven investigation was performed. A thermo-kinetic analysis of the PHB synthesis pathway identified how the relative concentration of various metabolites in the pathway influenced overall productivity. These findings were incorporated into a recently constructed genome-scale metabolic model of the bacterium to understand how characteristics of the utilized carbon substrate affected PHB productivity. This model-guided approach yielded several engineering design strategies for PHB over-production, including utilizing reduced, high molecular weight substrates that bypass the thiolase reaction. Overall, these findings uncover key thermodynamic and enzyme saturation limitations controlling PHB production and lead to design strategies that can potentially be transferrable to other PHB producing bacteria.

#48 Top-bottom and bottom-up approaches for investigating the Microcystis phycosphere by single aggregate droplet encapsulation and microdroplet combinatorial decomposition

James Tan, University of Michigan

Understanding how cell-cell interactions compound into dynamics observed in natural communities is crucial for predicting and engineering microbial communities for host and environmental health. Both meta-genomic studies of natural communities and “synthetic ecology” via defined microbial communities have provided insights into the structure, function, dynamics, and underlying mechanisms. However, for microbial systems that lack well characterized and tractable representatives to construct systems by synthetic ecology, they are heavily limited to study by culture-independent techniques. Despite tremendous improvements in culture-independent techniques, it has been difficult to achieve the high resolution required to study interactions in natural communities due to the sheer complexity. To address this challenge, this work presents two complimentary approaches, one top-bottom and the other bottom-up, to provide insights into interactions between toxic bloom-forming Microcystis aeruginosa and its aquatic heterotrophs. The top-bottom approach involves using droplets to encapsulate single Microcystis-heterotroph colonies from blooms in Lake Erie to study membership and strain-level association at the local-scale. The bottom-up approach utilizes microfluidic droplets, which are sub-nanoliter to microliter scale water-in-oil emulsions. These “microdroplets” serve as miniaturized, self-contained bioreactors for microbial cultivation, rendering ultra-high throughput investigation. In particular, the small size and stochastic nature of microdroplets enables the decomposition of a mixed culture composed of Microcystis aeruginosa and heterotrophic partners to interrogate the critical pairwise interactions. These two approaches combined will help advance our understanding of how the microbial partnerships contribute to the resiliency of Microcystis¬-dominated harmful cyanobacterial blooms.

#49 A General Platform for Accelerated Enzyme Engineering

Mark Mahnke, UW-Madison

Engineered enzymes occupy a critical role in modern biotechnology and medicine. Despite their importance, the engineering process remains a time-consuming and resource-draining process. A general platform for enzyme engineering that optimizes and accelerates the process would provide great benefit to both industry and academia. Here we propose such a platform, combining biophysical modeling, droplet microfluidic enzyme screening, and machine learning to provide an efficient search of protein sequence space. A protocol in RosettaLigand, a software for modeling substrate-binding pockets, introduces mutations and scores them based on various biophysical characteristics. We plan to use this software to design a library containing thousands of different active sites which can be screened for activity using droplet microfluidics. Through microfluidic’s ultra-high throughput capacity, multiple data points can be gathered for each variant and used to train a machine learning model to predict higher performing variants multiple mutations away. We will test this new approach using CYP76AD1, a P450 enzyme that produces L-DOPA, which is an important small-molecule for treating Parkinson’s disease as well as a precursor to many drug metabolic pathways. We hypothesize that through the combination of these techniques we will be able to accelerate and optimize the enzyme engineering process.

#50 Directed evolution of superinhibitors of quorum sensing signaling in Staphylococcus aureus

Tom Mansell, Iowa State University

Methicillin-resistant Staphylococcus aureus infections are very common and resistant to most antibiotics. Like many pathogens, S. aureus uses cell-cell communication aka quorum sensing (QS) to coordinate attack and regulate expression of toxins. Using the model Gram-positive organism Bacillus megaterium, we have developed a high-throughput fluorescence-based assay for quorum sensing inhibition and activation of S. aureus signaling. Using a designed library of ~5,000 variants of the QS autoinducer peptide, we found and characterized over 150 potential inhibitors of quorum sensing and several superactivators. Because they are ribosomally templated, these peptides can be secreted by live biotherapeutics as a potential non-antibiotic treatment for S. aureus infection.

#52 A framework for studying robustness of engineered signaling programs to variability in genetic context

Hailey Edelstein, Northwestern University

Synthetic receptor systems are powerful tools for engineering mammalian cell-based devices to sense and respond to user-defined cues. A challenge in engineering mammalian cell signaling programs is developing and characterizing them in a way that enables their translation from workhorse cell lines to the context required for downstream applications. Context features, such as gene delivery method and cell type, impact the overall expression level of engineered parts in a program, relative expression of separate parts in multi-component programs, and heterogeneity of expressed parts across the cell population. Each of these effects can substantially impact a program’s ability to satisfy performance objectives. Ideally, engineered programs would be robust to context changes such that prototyping in one context can lead to meaningful conclusions about how that program would function in a different context.

Towards this goal, we developed a framework for expression level robustness-guided implementation of synthetic signaling programs across contexts. As a model system for this study, we employed the Modular Extracellular Sensor Architecture (MESA) receptor system, a self-contained synthetic receptor system for sensing soluble ligands in mammalian cells with documented sensitivity to expression level. We harness plasmid uptake heterogeneity in transient transfections to systematically survey a wide range of receptor expression levels and evaluate robustness of performance criteria across the expression space. Based on these explorations, we observe that some features of synthetic receptor design confer increased robustness to expression level. To map various regions of the expression space generated in transient transfections to the genomic context, we establish expression level benchmarks for translationally relevant genomic integration techniques. We use these observations to guide the implementation of synthetic signaling programs in translational contexts.

#53 Engineering light-switchable bacterial two-component systems to function in stationary phase

John Lazar, Rice University

Bacterial optogenetics is a powerful tool that has allowed for precise and dynamic control of bacterial gene expression at steady state growth. However, in real-world applications of optogenetics, such as dynamic control of pathway expression over the course of a metabolic fermentation, bacteria exist in a non-growth stationary phase that largely changes the circuit dynamics. Thus, it is imperative to engineer optogenetic circuits that are functional at steady state and in stationary phase. One of the most highly utilized optogenetic systems is CcaSR, a green/red light photoreversible two-component system that utilizes the heme derivative phycocyanobilin (PCB) as a cofactor. Previously, we have engineered CcaSR strains with high dynamic ranges and fast dynamics in exponential growth; however, these strains were confirmed to be nearly nonfunctional in stationary phase. Thus, we have engineered a CcaSR circuit with stability and dynamics in both growth and non-growth phases. Kinetic modeling of two-component system stationary phase behavior, expression optimization of CcaSR, and PCB stabilization strategies reveal that decreasing response regular expression and stabilizing production of PCB increases the dynamic range of the two-component system from 4 fold to over 400 fold in stationary phase. Although this newly developed strain yields stable activation of CcaSR when the cells are grown into stationary phase, the circuit did not demonstrate the ability to be deactivated in stationary phase. To address this, we implemented mflon, an orthogonal protein degradation platform, to the CcaSR circuit in order to achieve deactivation of CcaSR in stationary phase. Finally, we apply this engineered CcaSR circuit to dynamically control expression of sts-TAL, an enzyme that converts L-tyrosine into the industrially relevant precursor, p-coumarate. This work demonstrates the first bacterial optogenetic two-component system with stability and dynamics in stationary phase.

#54 Real-time detection of response regulator phosphorylation in live bacteria

Ryan Butcher, Rice University

Bacterial two component systems (TCSs), the largest family of multi-step signal transduction pathways, facilitate environmental sensing and regulate diverse cellular processes such as pathogenicity and biofilm formation. Typically, a membrane-bound sensor kinase (SK) senses stimuli and phosphorylates a cytoplasmic response regulator (RR), which dimerizes, binds DNA, and modulates transcriptional responses. While TCSs have been studied in vitro and via transcriptional output, there is no generalized method to directly monitor TCS activity in living bacteria. To achieve this, we have developed a method for detecting RR dimerization via fluorescence polarization. As fluorescently-labelled RRs dimerize, energy transfer occurs between “on-axis” and “off-axis” fluorescent probes and the resulting emission light is depolarized. In particular, we detect dimerization of representatives from the two largest RR sub-families, TorR and NarL, fused to the fluorescent protein mNeonGreen (mNG). We demonstrate that fluorescence polarization decreases seconds after TMAO addition to cultures expressing TorR-mNG alongside the SK TorS and requisite TMAO binding protein TorT. Similarly, fluorescence polarization decreases immediately after nitrate addition to cultures expressing mNG-NarL with the nitrate-sensitive SK NarX. We verify that RR phosphorylation drives these signals by repeating the experiments with phospho-null SK and RR mutants. The phosphorylation kinetics of both TCSs depend on RR and SK expression level, in agreement with computational modeling. Furthermore, we investigate a previously unreported ‘activation surge’ dynamic in NarL phosphorylation. Our method will enable researchers to obtain previously-inaccessible measurements of RR phosphorylation dynamics in living cultures, revealing interesting temporal features and bypassing confounding factors introduced by transcriptional assays.

#55 Hydrogel-encapsulated bacteria to diagnose colon inflammation

Elena Musteata, Rice University

Inflammatory bowel disease (IBD) is a chronic, painful condition correlated with increased patient susceptibility to cancer, arthritis, and other debilitating illnesses. Over the last decade, the prevalence of IBD globally has been steadily rising. In the US, an estimated 3 million adults live with IBD.

We are developing a non-invasive tool for diagnosing IBD and monitoring recurrences of gastrointestinal inflammation. Our group previously engineered a bacterial biosensor that senses a biomarker of colonic inflammation and produces a fluorescent protein reporter in response. The diagnostic potential of this sensor was demonstrated using the dextran sodium sulfate (DSS) mouse model of colitis. Currently, we design and deploy engineered bacterial biosensing strains that respond to other biomarkers of colonic inflammation. Within our platform, these strains can be deployed simultaneously to enable multiplexed biomarker detection for increased diagnostic confidence.

For efficient delivery to the animal gut, we encapsulate the bacteria within a superparamagnetic and porous hydrogel matrix. Compared to previous methods for living therapeutics, this approach improves bacterial survival and function in the gastrointestinal tract, enables in vivo biocontainment, and facilitates rapid post-digestion retrievability and analysis. The encapsulated bacteria are administered orally and retrieved from stool for diagnosis. This allows us to measure inflammation non-invasively, unlike endoscopy and colon biopsy, the current standards of care.

By reducing or eliminating the need for invasive procedures, this approach represents a significant advance for clinical IBD diagnosis and disease monitoring. In addition, the demonstrated multiplexing capabilities of our bacterial encapsulation platform can be expanded to include biosensors for other disease biomarkers, enabling the use of similar non-invasive diagnostic tools for diverse human health applications.

#56 Directed evolution to define specificity alterations in NRPSs using siderophore systems

Erin Conley, University of Wisconsin-Madison

Nonribosomal peptide synthetases (NRPSs) are mega-enzymes that produce specialized metabolites with therapeutic and industrial significance. The adenylation (A) domain is responsible for substrate selection, and has been the target of many engineering efforts to generate designer peptides. Despite isolated successes in altering substrate specificity, a reliable method for switching A domain specificity does not exist. In this study, I exploit the essential production of siderophores to define specificity alterations in A domains. A directed evolution pipeline involving generation of non-functional NRPS fusion proteins, random mutagenesis and a genetic selection under iron-limiting conditions allows for identification of enzymes that have undergone a switch in substrate specificity. This directed evolution approach will be expanded to additional siderophore systems, thereby creating a knowledgebase of specificity alterations in NRPS enzymology.

#57 Studying the Origins of Molecular Cooperation Using Prebiotic Amino Acid Reaction Networks

Hayley Boigenzahn, University of Wisconsin Madison

Origins of life studies seek to shed light on how living cells first appeared on Earth. The Miller-Urey experiments showed how simple inorganic species could combine to form amino acids, but how such molecules further reacted to make precursors of cellular life is still highly debated. Under certain conditions, peptide bonds between amino acids can form spontaneously in the absence of cells, mRNA or ribosomes. Amino acids reacting to form increasingly complex peptides produce an enormous reaction network with the potential for many intermolecular interactions. Identifying these interactions as they appear can help us understand how molecular cooperation, which underlies all biological networks, can emerge in a non-living system. The potential complexity of these reaction networks and the interactions within them is comparable to biological systems, and like with many biological systems, gathering high-quality experimental data is challenging. Modeling techniques developed to help interpret complex biological networks can also help us understand peptide reaction networks. We aim to link detailed models of small systems to larger-scale network models to investigate details of peptide formation on the early Earth, including the impact of the environment on what peptides form and how interactions between peptides may have supported the emergence of early life.

#58 Expanding the Genetic Toolkit for Plant Growth Promoting Rhizobacteria Synthetic Biology

Caleigh Roleck, Rice University

Rhizobacteria, bacteria that live in the soil surrounding plant roots, play an important role in shaping soil properties and plant health and productivity. Certain rhizobacteria known as plant growth-promoting bacteria (PGPB) have the ability to facilitate nutrient acquisition in plants, modulate phytohormone levels, and prevent growth of plant pathogens, which has caused PGBP to gain attention as environmentally-sustainable fertilizer and biocontrol agents. Azospirillum, Burkholderia, and Psuedomonas are all able to colonize the rhizosphere of a large variety of plants and, due to their potency as PGPB, have been commercialized for agricultural use, though only a small fraction of agriculture currently uses PGPB inoculants. Engineering PGPB to perform plant growth-related functions in a more robust, controllable manner could reduce some of the inconsistencies of PGPB field applications hindering more widespread adoption. Still, well-characterized biological parts are currently lacking for many PGPB, which limits the development of PGPR synthetic biology. Here, we design and characterize a constitutive promoter library with a range of expression levels in Azospirillum brasilense, Burkholderia unamae, and Psuedomonas fluorescens—the first constitutive promoter library for A. brasilense and B. unamae. We also characterized inducible promoters to allow for tunable gene expression, which are currently non-existent in A. brasilense and B. unamae and lacking in P. fluorescens. We plan to use this library to express ligand-responsive one and two-component systems and metabolic pathways. This work paves the way for engineering communication between plants and PGPB allow for controlled delivery of useful bacterial functions to the soil in accordance with environmental conditions and plant needs.

#59 High-throughput discovery of two-component signaling systems from the human gut microbiome

Kevin Lorch, Rice University

The human gut microbiome is rich in biological diversity and provides a promising source of novel biosensors. We computationally identified over 3,000 uncharacterized two-component signal transduction systems (TCSs) in the genomes of 450 common human gut bacteria. Major challenges to elucidating the inputs of TCSs for synthetic biology applications are that TCS output promoters are unknown or are transcriptionally silent in lab-tractable hosts. Our lab recently developed a technique to rewire TCSs to well-characterized output promoters by modularly swapping the response regulator DNA-binding domain. Here, we selected 543 diverse TCSs from the human microbiome, synthesized the genes, and transformed the library into E. coli. Each TCS controls the expression of a barcoded mRNA and a GFP reporter gene. We develop a next-generation sequencing (NGS) approach to high-throughput screening of this library in response to human fecal samples and discrete chemical ligands. We use unique molecular indices to multiplex ligand screening and identify candidate sensors in high-throughput. Using this approach, we have identified dozens of TCSs that respond to human feces, a novel sensor of the inflammation biomarker nitrate, and sensors of a gut-linked neuropeptide. We are now screening this library against ligands including antimicrobial peptides produced during innate immune response, colorectal cancer biomarkers, and virulence factors of common gut pathogens. We will also screen primary fecal samples from patients with gastrointestinal diseases to identify candidate diagnostic sensors for disease state. This unique TCS discovery platform will enable the engineering of diagnostic and therapeutic gut bacteria for a wide range of diseases.

#60 Model analysis and identification of Pulse Generator designs with feedback control

Ania Baetica, University of California, San Francisco

A common network motif in synthetic and systems biology is the incoherent feedforward loop (I1-FFL), in which an activator molecule regulates both a gene and a repressor of that gene. The I1-FFL circuit is known as a pulse generator since it responds to a step increase in the input with a pulse of output gene expression. While I1-FFL circuits have been hypothesized to return to basal activity after a pulse (adaptation), adaptation does not occur in our transcriptional implementations of I1-FFL in S. cerevisiae cells. The lack of adaptation in our I1-FFL circuit has motivated us to study its response and performance of after the addition of negative and positive feedback control.

In this work, we consider an I1-FFL circuit design with two activating synthetic transcription factors, GEM and SynTF, and a previously published de novo protein called degronLOCKR that implements repression. To this I1-FFL circuit, we add either a positive feedback or a negative feedback to the activating transcription factor GEM, and then a combination of both types of feedback. Our goal is to understand and characterize the properties of these four circuits with the incoherent feedforward motif and additional positive and negative feedback loops. We characterize how well these four circuits adapt after a pulse and how large the amplitude of their transient response is.

First, we develop mathematical models of the four I1-FFL circuits and validate them against the experimental data. We use our models to prove that transcriptional nonlinearities prevent the I1-FFL’s return to basal activity after a pulse and that by adding negative feedback, we can improve steady-state error. Second, we demonstrate that the addition of positive feedback results in a higher amplitude pulse, but also higher steady-state error. This series of results are a guideline on how to tune the amplitude and steady-state of the pulse created by an I1-FFL circuit using positive and negative feedback.

#61 Engineering Streptomyces to Capture Value from Lignocellulosic Biofuel Conversion Residue

Caryn Wadler, University of Wisconsin-Madison

Current methods of switchgrass hydrolysate fermentation to bioethanol leave behind about 60% of the organic material in the hydrolysate after ethanol distillation. This material is referred to as lignocellulosic conversion residue (CR). To increase the economic viability of lignocellulosic biofuels, we are engineering Streptomyces species to maximize the metabolism of CR carbon into valuable bioproducts. From a library of 75 phylogenetically distinct Streptomyces isolates, we generated a collection of Streptomyces that produce lycopene from CR as a reporter for their potential to produce isoprenoids. The genetic element used in constructing this reporter is mobilizable between Streptomyces species and we have constructed further plasmids using a combination of traditional cloning techniques and Golden Gate assembly that allow for rapid alterations in expression levels and the generated bioproduct.

Initial screens of the engineered Streptomyces reporter strains showed a wide range of lycopene production levels. We were able to further increase lycopene production by introducing two different pathways that produce isoprenoid precursors: an optimized version of the native methylerythritol phosphate (MEP) pathway or the mevalonate (MEV) pathway with a constitutive promoter. We have also begun expanding our bioproduct library to include constructs that generate isoprene, limonene, or pinene.

#62 Machine-learning guided acyl-CoA/acyl-ACP reductase engineering for production of fatty alcohols

Jonathan Greenhalgh, University of Wisconsin-Madison

Fatty acyl reductases (FARs) are key enzymes in microbial routes to fatty alcohol production. Many existing metabolic engineering strategies utilize FARs to produce fatty alcohols from acyl-CoA pools; however, producing alcohols from acyl-ACP intermediates from fatty acid biosynthesis is more direct, has a lower energetic cost and could improve the efficiency of fatty alcohol production. We have implemented an active-learning approach to iteratively search the protein fitness landscape for FAR enzymes with improved activity on acyl-ACPs in vivo. Over the course of ten design-test-learn rounds, our machine learning models converged on engineered enzymes that produce over twofold more fatty alcohols than the initial natural sequences. We also performed in vitro kinetics assays with the top engineered sequence and determined that the improvement in alcohol production is the result of improved enzyme catalysis on the acyl-ACP substrate. Additionally, we found a strong correlation between the in vivo activity on acyl-ACP and the net charge near the putative acyl-ACP binding site.

#64 Revisiting the unique structure of autonomously replicating sequences in Yarrowia lipolytica and its role in pathway engineering

Carmen Lopez, Iowa State University

Production of industrially relevant compounds can employ either a genome or a plasmid as expression platforms. Selection of plasmids as carriers of pathways is advantageous for rapid demonstration but poses the challenge of stability, especially in nonconventional yeasts such as Yarrowia lipolytica. The autonomous replicative sequence or ARS from this oleaginous yeast showcases a unique structure; however, its characterization is scarce and scientists have paid little attention to the underlying causes for its behavior.

In this work we studied how selection of ARS sequences impacts plasmid-born gene expression and plasmid stability. We show that a noncoding sequence (referred as spacer) plays a key role in the differences observed: the use of a wildtype ARS sequence yields an increase of gene expression by 2.2-fold and 1.7-fold higher plasmid stability compared to a minimal ARS. We also tested the modularity of ARS sub-elements. We found that while these synthetic ARS sequences are more amenable to optimization, their behavior is not well predicted as plasmid stability suffers and it does it in a pathway-dependent manner.

Overall, our work sheds light to an atypical element of plasmid design in Y. lipolytica and illustrates an amenable strategy to quickly improve performance and plasmid stability as long as users are aware of potential pathway-dependency occurrence.

#66 Expanding the therapeutic capacity of engineered commensal microbes

Vince Kelly, University of Illinois at Urbana-Champaign

One of the great challenges in the development of biotherapeutics is the high cost associated with these drugs, both in terms of manufacturing cost and the patient’s cost for treatment. Dosing regimens requiring repeated administration, trips to a healthcare provider, and large amounts of formulated therapeutic, are a significant driver of these costs. The goal of our research is to leverage the inherent capabilities of the diverse microbial community of the human gut to reduce these costs by acting as therapeutic factories in situ. Recent developments in microbial genetics, synthetic biology, and the study of niche-optimized, non-pathogenic members of the human gut microbiota have paved the way for the engineering of commensal bacteria to carry out novel functions. We have developed a Golden Gate DNA assembly system to generate modular, tunable gene cassettes for driving the in situ delivery of therapeutic proteins from commensal microbes. Using species-specific genetic components, such as promoters and ribosome binding sites, therapeutic production can be achieved in both transiently colonizing and persistently colonizing members of the gut microbiota. These engineered strains can be used to treat conditions localized to the gastrointestinal tract by secreting therapeutic directly to the affected area, circumventing the need for repeated dosing and costly purification or formulation. Additionally, we are currently utilizing a protein engineering approach to add unique functionality to our therapeutics that will allow them to escape the gut lumen and enter the systemic circulation for the treatment of non-gastric conditions (e.g., HER2+ breast cancer). This work uses both in vitro validation of engineered therapeutics and commensal strains, and in vivo studies in murine and porcine model systems to demonstrate a modular, cost-effective method for in situ production and delivery of effective biotherapeutics following oral dosing of engineered commensal microbes.

#67 Introducing a pair of tools for the in silico design and dynamic simulation of eukaryotic genetic circuits

Wheaton Schroeder, University of Nebraska - Lincoln

Synthetic biology often relies on the intuitive design of genetic circuits utilizing bioparts (DNA, RNA, and proteins) to accomplish a desired task; however, this field holds the promise of creating biological systems with complex and useful responses to various stimuli. As an alternative to intuitive approaches, we created two in silico tools which broaden the range, efficacy, and predictiveness of potential synthetic biology designs by applying systems biology and optimization mathematics. This approach allows non-intuitive designs and the quick in silico screening thereof to give implemented designs the greatest chance of success. The first tool, EuGeneCiD (Eukaryotic Genetic Circuit Design tool), aids in the design and/or study of genetic circuits using a bioparts database describing the potential members of genetic circuits included promotors, genes, transcripts, terminators, and enzymes in terms of their activity, regulation, and interactions. This tool is used to design dozens of gates for the model plant Arabidopsis thaliana which respond to the divalent metals Cadmium, Zinc, and/or Copper using fluorescent proteins as reporters. The second tool, known as EuGeneCiM (Eukaryotic Genetic Circuit Simulation Modeling tool) allow for dynamic simulation of designs proposed by EuGeneCiD to test the accuracy of the design, particularly with respect to time. EuGeneCiM is applied to all EuGeneCiD-designed circuits, as well as to other time-dependent synthetic biology circuits such as repressilators to demonstrate the utility of the tool. Together, these tools can hypothesize optimal genetic circuits designs and simulate their behavior to increase the chances that a plant might have the desired behavior when transformed potentially saving time and resources. For the ease of use, various programs have been developed to make EuGeneCiD/M user-friendly, including a protocol and providing all necessary code through GitHub (https://github.com/ssbio/EuGeneCiDM).

#68 Using a Biosynthesis Modular Assembly Toolkit to Engineer UDCA production in Yeast

Iesh Gujral, University of Minnesota

Ursodeoxycholic acid (UDCA) is one microbial bile acid made with ursodiol and a taurine conjugate naturally found in bear bile. UDCA is suspected to have beneficial applications in liver function, gallstone treatment, bile flow stimulation, anti-inflammatory effects. Recently it has been reported that UDCA have neuroprotective effects against Parkinson’s and Alzheimer’s disease. Currently, UDCA production is unethical to extract from bears and too complex to productionize with chemical synthesis. In this project, we attempt to alternatively produce UDCA with microbial catalysts in yeast. Yeast was selected because it produces a precursor molecule for UDCA biosynthesis. Such biosynthesis requires fifteen genes from Human and Zebrafish and hence mandates careful tuning of gene expression strength.

Thus, “A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly” is used. This toolkit presents a versatile engineering platform for yeast, containing rapid, modular assembly and basic characterized parts. It is based on Golden Gate Assembly. It provides a framework to create designs containing promoters, terminators, assembly connectors, and more in combinatorial ways for multi-gene cluster engineering.

In the context of the UDCA project, this required tedious primer design to eliminate several restriction sites in the desired DNA sequence. In the zebrafish gene of HSD17B4 alone, this required assembling 7 fragments and 13 primers to simply produce a part plasmid. In addition, overhang sequences play a significant role in binding. Thus, researchers ensure high GC content and avoid palindromic scar sequences in overhangs. Finally, the toolkit has limitations in both count and variability. It contains 19 promoters of and 6 terminators of specific strengths. However, the UDCA project entails 15 genes. The toolkit restricts biosynthesis capabilities as we intend to maximize expression while maintaining unique combinations of promoters and terminators.

#69 Integrative Data-mining Methods to inform Re-engineering of RNA Regulators for SynBio Applications

Mia Mihailovic, University of Texas at Austin

Bacterial regulatory RNA, such as small RNA (sRNA), provide rapid post-transcriptional responses to optimize survival in response to environmental state. Because of their conditional behavior and native stability, they have been acknowledged as valuable platforms for re-engineered cellular regulation. However, to fully take advantage of the large suite of native regulatory RNA as SynBio tools, understanding their larger native regulatory networks is necessary. Although thousands of sRNA sequences have been confirmed, their transcriptional regulation and post-transcriptional roles lag far behind. Even in E. coli, 60% and 40% of sRNAs have not a single documented transcriptional regulator or target, respectively. This problem largely exists due to the targeted nature of current characterization approaches. For instance, chromatin immunoprecipitation studies involve a single protein, although many DNA-binding proteins regulate cooperatively, and RNA target ligation and sequencing studies typically rely on protein chaperones for pull-down, although many sRNAs do not rely on these factors.

To rationally delineate sRNA networks in a high-throughput manner, we have developed an integrative data-mining approach that layers a variety of publicly-available omics datasets as well as bioinformatic tools to suggest regulators and targets of biochemically-confirmed sRNAs. Specifically, we incorporate stress-specific genomic protein occupancy datasets (IPOD-HR), motif-search tools (FIMO), and condition-specific RNA expression (RNA-seq) to propose sRNA transcriptional regulators that act at the DNA-binding level. To uncover sRNA targets, we filter target predictions (IntaRNA) using RNA regional accessibility (INTERFACE) data whose patterns help predict binding sites. Here we show our generalizable platform and its reliability in capturing native E.coli sRNA networks, including that of a sRNA whose cellular role has been elusive since its discovery over 15 years ago.