Poster Abstracts

Production of medium chain length oleochemicals from C2 substrates

Medium chain length (MCL) oleochemicals have the potential to serve as renewable sources of fuel owing to their cetane number and low melting point. The sustainable production of MCL oleochemicals is crucial given the current situation of global warming and diminishing natural fuel resources. In this work we have developed a renewable chemical production strategy that uses acetate as the energy carrier to link renewable energy with the synthesis of organic compounds. We demonstrated that acetate can be efficiently converted to MCL oleochemicals in Escherichia coli by the expression of medium chain thioesterase (TE), Acyl-CoA synthase (FadD) and Acyl-CoA reductase (ACR). We integrated the fatty alcohol pathway into E. coli and optimized the copy numbers. E. coli grown on acetate as a substrate has higher energy costs and no yield of reducing equivalents. We therefore chose to cofeed ethanol, a good electron, carbon, or energy source that could provide reducing equivalents when fed with acetate. E. coli is incapable of growing on ethanol; hence we expressed the best performing operon with combination of alcohol assimilating genes, from various sources and combined with improved engineered acetate activation by using an Acetyl CoA Synthetase (ACS) mutant. The engineered strains are then used to develop a fermentation process that yields high titers of oleochemicals from acetate.

Using an expanded genetic code to select cyclic peptide inhibitors of amyloid-β 42 aggregation

Genetic code expansion technology augments the chemical diversity available to ribosomally synthesized polypeptides by enabling incorporation of noncanonical amino acids (ncAAs). We apply this technology to a phage-based, in-cell selection of cyclic peptide inhibitors of amyloid-β 42 aggregation, and ask how outcomes may vary between different expanded genetic codes. A phage-encoded cyclic peptide library, where a ncAA must be incorporated into an otherwise randomized cyclic peptide sequence, was selected with six different ncAAs in parallel. An identical peptide sequence dominated two separate selections, one utilizing 1-methyltryptophan and the other 3-iodotyrosine. This cyclic peptide, with either ncAA, displayed similar levels of inhibition to a previously identified cyclic peptide inhibitor. Although this sequence is tolerant of two different ncAAs at the same position, it is not broadly permissive. Replacing either ncAA with 5-hydroxytryptophan or p-benzoylphenylalanine dramatically reduces activity. Likewise, substituting analogous canonical amino acids, tryptophan or tyrosine, results in a ~2-fold reduction in activity. Our work demonstrates the viability of selecting peptides in parallel with various expanded genetic codes, and the feasibility of identifying active ncAA-containing sequences from cellular selections.

Enhancing Lacto-N-Tetraose Utilization: A Comprehensive Approach via Protein Engineering and Selective Directed Evolution of Lacto-N-Biosidase.

Human milk oligosaccharides (HMOs) are crucial in promoting gut colonization by providing prebiotic advantage to beneficial gut bacteria. Lacto-N-tetraose, a tetrasaccharide and 2-fucosyllactose are identified as bifidogenic due to the prebiotic properties for selective growth of Bifidobacterium spp., a key contributor to host metabolism and development and have immunomodulatory properties. However, the genetic limitations of gut microbes specifically those from cesarean-section-delivered neonates result in ineffective utilization of these HMOs. Lacto-N-tetraose hydrolysis by lacto-N-biosidase (LNBase) liberates lactose and either lacto-N-biose (LNB) or N-acetyllactosamine (GNB) from the β1-3 glycosidic linkage between the type I or II respectively. While lacto-N-biosidase, a gene from Bifidobacterium spp. can hydrolyze lacto-N-tetraose type I (LNT) but have no activity on the isomer lacto-N-neotetraose (LNnT). This proposal aims to develop an optimized enzyme (lnbB or lnbX) that has enhanced dual affinity and activity to both Lacto-N-tetraose isomers. Deterministic chemical kinetic modeling predicted microbial growth and HMO utilization over time. Molecular studies revealed distinct catalysis mechanisms for lnbB and lnbX, guiding possible mutagenesis strategies. Future works would involve the introduction of mutations by error-prone PCR or rational amino acid substitutions guided by structural analysis along with directed evolution to screen large mutant libraries. Improved variants can be tested in probiotic bacteria to assess in vivo symbiotic performance. This study lays the groundwork for engineering live biotherapeutic to maximize prebiotic and nutritional benefits from HMOs for human health precisely neonates and the aged.

Innovating Thaumatin Production using a Chlamydomonas Biofactory

When the World Health Organization released The WHO Acceleration Plan to Stop Obesity, a common intervention adopted by many countries was the reevaluation of sugar-sweetened beverages. Finding non-sugar-based sweeteners will provide new options to help consumers maintain a healthy weight. Thaumatin is a low-caloric “sweet protein” isolated from the katemfe fruit of Thaumatococcus daniellii Benth, a plant native to rainforest areas of Africa. However, large-scale cultivation of T. daniellii to meet the food industry needs would be challenging, so alternatives are needed. Synthetic biology provides a system to investigate production of this sweet protein without the complex cultivation challenges of growing the T. daniellii plant outside its native zone. The work presented here explores the production of the Thaumatin protein in the eukaryotic microalga, Chlamydomonas reinhardtii. The modular MoClo technology has allowed us to construct expression cassettes that target thaumatin produced by nuclear transformation to the chloroplast, mitochondria, or the ER pathway. Understanding the nuanced expression of heterologous thaumatin in Chlamydomonas will help to grow the use of this microalgae as a production system. Optimizing thaumatin production could provide a feasible opportunity for a competitive introduction of thaumatin into the global sweetener market.

Assaying for Realized Niche Expansion in Stressed Environments with Synthetic Microbial Communities

Microbial communities carry out complex and essential functions in nature leveraging strategies such as division of labor and functional redundancy. It has been proposed that synthetically designed microbial communities (SynComs) could offer similar advantages in biotechnological applications. To realize this potential, though, the next generation of SynCom approaches ought to incorporate a broader genetic diversity and identify environments where communities are specifically suited compared to monoculture. Borrowing from approaches in ecology, here we used paired synthetic microbial communities of genetically diverse isolates taken from a field site with acid and metal contamination to identify advantageous SynComs. By assaying under stressed and non-stressed conditions, we identify microbes with the potential to expand the realized niche (survivable environment) of partner microbes. Such findings have relevance for the construction of SynComs robust to toxic environments with possible applications in remediation and waste upgrading.

Metabolic Engineering of Issatchenkia orientalis for Cost-effective Production of Citramalate

Methyl methacrylate (MMA) is a building block of poly MMA (PMMA), a material commonly recognized as acrylic glass or plexiglass. The prevailing method to manufacture PMMA utilizes petroleum and the acetone cyanohydrin process, which is considered unsustainable and raises concerns regarding the use of toxic chemicals. An alternative route using semisynthesis, combining converting microbially produced citramalate to methacrylic acid (MA) using a catalyst and final esterification of MA to MMA, could present a viable solution. Previous studies have shown large titers of citramalate in engineered Escherichia coli. However, the increased production cost and carbon footprint from having a neutralization and reacidification step makes this route less economically attractive. A pH-tolerant strain such as Issatchenkia orientalis would be an attractive alternative. Through extensive collaborative efforts, our CABBI team has previously demonstrated that I. orientalis can withstand citramalate of up to 80 g/L at pH 3. Additionally, the team engineered a strain that produces 2 g/L citramalate by integrating the citramalate synthase gene (cimA) from Methanocaldoccus jannasch into the I. orientalis genome using piggyBac. Examining bottlenecks and increasing flux towards citramalate through metabolomic study and genome scale modeling was essential in increasing citramalate production. The metabolomics studies found that there is pyruvate overflow, accumulation of intracellular citramalate, and excess ethanol byproduct. Using the genome scale model, pathways of interest were identified to address these problems. We simultaneously decreased pyruvate overflow and increased lacking acetyl-CoA by utilizing an aldehyde dehydrogenase gene (ALD6) from S. cerevisiae and a mutated acetyl-CoA synthase gene (ACSSEL641P) from Salmonella enterica. We also incorporated a multidrug transporter (QDR3) from S. cerevisiae to increase excretion of citramalate into the growing broth. Using piggyBac we integrated QDR3-ALD6- ACSSEL641P- and a more active CimA (cimA3.7) to create a library of variants. We selected this library's top producer (Q42) for further engineering. To address excess ethanol production, we employed CRISPR to delete the pyruvate decarboxylase gene (PDC). The resulting strain, Q42 Δpdc, produced 18 g/L of citramalate in shake flasks and was scaled up to a 3L bioreactor to produce 30 g/L of using synthetic complete medium with 0.6% ammonium sulfate, 5% glucose, and a trace metal supplement without needing pH or dissolved oxygen control. In addition, the CABBI team successfully generated a xylose-utilizing strain, Q42X Δpdc, for future work involving potential feedstocks such as sorghum hydrolysate and sugarcane juice. Currently, the team is actively characterizing the metabolism and genomics of I. orientalis with developing a comprehensive knockout library and updating our genome scale model. The team is also exploring new metabolic engineering strategies such as eliminating glycerol production, exploring alternative routes for acetyl-CoA production, identifying additional target genes for up- or down-regulation, and developing a light-controlled circuit.

Unraveling Transporter Functions in E. coli: Enhancing Microbial Tolerance for Sustainable Biofuel and Chemical Production

As global warming continues to advance, there is an urgent need for new, environmentally friendly methods to produce energy and chemicals. Synthetic biology, an extension of genetic engineering, has made significant progress in developing efficient ways to produce biofuels and other products. However, a critical challenge remains in understanding how microbes transport materials across their cell membranes. Toxic compounds can damage cells in various ways, making it challenging to target specific mechanisms of toxicity. A more effective approach might be to enhance cell tolerance by actively removing toxic substances from both the cell and its environment. Transporters facilitate this removal process by regulating the movement of compounds in and out of the cell, helping to maintain safe, non-toxic levels. The unknown functions of many transmembrane proteins limit our ability to modify transporter proteins for improving cell tolerance and reducing toxicity. Our research intends to reveal these functions using high-throughput techniques like CRISPRi, knockout libraries, protein variant screenings, growth competition assays, and toxicity tests. We will evaluate the impact of transporter function loss through a knockout library and test gain of function via heterologous expression. By concentrating on E. coli transporters, our objective is to identify the substrates they transport and assess their efficiency. This research has the potential to significantly advance our understanding of transporter mechanisms and improve strategies for reducing cellular toxicity, contributing to more sustainable production methods in the face of global climate challenges.

Under-Oil Open Microfluidics Systems for studying opportunistic pathogen dynamics

Dispersal -- the process by which biofilms release planktonic cells -- is a critical virulence factor in Candida albicans infections. Our understanding of Candida albicans dispersal is rudimentary, in part due to technical challenges. Under-Oil Open Microfluidics Systems (UOMSs) rely on dual exclusive liquid repellency such that a solid surface can be patterned with extremely hydrophobic and extremely hydrophilic regions to force culture media to adhere only to hydrophilic regions while sterile silicone oil adheres to hydrophobic regions. This technology reduces biofouling, enables optical access throughout the biofilm and dispersal lifecycle, and allows retrieval of dispersed cells. This has enabled us to identify that dispersal occurs as a specific stage within biofilm development coinciding with the cessation of biofilm expansion regulated by intercellular signaling. We are currently in the process of adapting transcriptional optogenetic control from Saccharomyces cerevisiae for Candida albicans in order to study the role of regulatory genes using to temporal and spatial control to avoid confounding affects on biofilm development.

Antifreeze Protein Production via Utilization of Corn Steep Liquor Grown Lactococcus lactis

Corn steep liquor (CSL) is a by-product of the corn wet-milling industry that has few applications outside of as a low-value additive to animal feed. However, it has the potential to be a fermentation broth supplement that is rich in sugar, protein, organic acids, vitamins, and minerals. While it has not been utilized extensively for this purpose, it has been used previously to biosynthesize a wide range of products including ethanol, antibiotics, and lactic acid, and has also been used to realize a nitrogen source cost savings of 75%. Our goal is to both leverage its affordability as a growth medium and utilize it to biosynthesize value-added antifreeze proteins (AFPs) in L. lactis. AFPs inhibit or modify ide crystal growth, and have potential applications in the frozen food industry and in extending crop and fish production seasons in areas with cooler climates. Our group has previously demonstrated the ability to use CSL to express AFPs in L. lactis, and seek to now improve and expand on this capability via adaptive laboratory experiments, cellular engineering, and mutational scanning of various different AFPs.

Single-Walled Carbon Nanotube Probes to Characterize Cell-Free Expression of Hydrolases

Cell-free expression systems offer a rapid platform for prototyping new proteins, such as enzymes. We present advancements in developing sensors that facilitate real-time, dynamic monitoring of enzyme activity, specifically proteases, within cell-free protein expression systems without the need for purifying the target enzyme. Single-walled carbon nanotubes (SWCNTs) exhibit native fluorescence due to their unique band-gap structure, with their emission being highly sensitive to changes in the local environment's permittivity and charge. This inherent sensitivity has been harnessed to create numerous label-free sensors, though selectivity limitations have restricted their use to clean buffers with minimal background interference. Our work advances the development of SWCNT-based sensors that maintain stability and functionality within complex cell extracts. These sensors allow for the direct measurement of the relative enzymatic activities of various serine proteases as they are synthesized in cell-free reaction mixtures. This approach enables continuous monitoring of protease activity, offering valuable insights into protein transcription, maturation timelines, and resultant activity levels. Such technology holds the potential for integration into closed-loop, build-design-test-learn cycles, accelerating the discovery of novel enzymes.

Characterization and Optimization of Split Enzyme Biosensors

The first step to effective treatment is timely and accurate diagnosis. Numerous indications progress rapidly and are terminal if not diagnosed early. Several therapeutics are only effective in the presence of a specific biomarker. Delayed diagnosis for bacterial or viral infections increases population mortality rates. While reliable diagnostic methods for nucleic acids have been established, protein detection remains a challenge in biosensor development for the point of care. This work focuses on the engineering, characterization and optimization of split enzyme protein circuits that can enhance the sensitivity, modularity and accessibility of diagnostic biosensors. Our model system for initial optimization is a split adenylate cyclase rapamycin sensor. Using protein engineering, biophysics and biochemistry tools, we are investigating the effects of different factors on split protein stability and activity. We are studying the impacts of reaction conditions, split sites and binder combinations to inform biosensor design and other in vitro applications and assays of split proteins. The modular nature of the sensor additionally lends itself to a “plug-and-play” approach to diagnostic development. We are validating various rationally designed binders in the split enzyme format for clinically relevant protein biomarkers. By leveraging a variety of design and analytical tools, we aim to establish a workflow for biosensor deployment.

Microcompartment Signal Sequences Work Cooperatively to Mediate Enzymatic Core Formation

Biomanufacturing has the potential to sustainably produce commodity chemicals but struggles to achieve high enough yields to compete with traditional chemical manufacturing. Spatial organization of pathway enzymes could overcome some obstacles that limit yield, such as toxic intermediates, diffusion barriers between enzymes, and kinetic bottlenecks. Many bacteria natively organize certain metabolic pathways using microcompartments (MCPs), proteinaceous organelles that encapsulate pathway enzymes in a semipermeable protein shell. MCPs are hypothesized to benefit these pathways by colocalizing pathway enzymes, internally recycling cofactors, sequestering toxic intermediates, and isolating intermediates from competing pathways. We aim to target non-native pathways to MCPs to confer these benefits to heterologous biosynthetic pathways. Many MCP cargo proteins are targeted to the MCP core by short N-terminal signal sequences, which can also be fused to non-native enzymes to target them to MCPs. However, both the mechanism by which these signal sequences target enzymes to the MCP and the role of signal sequences in the proper assembly and morphology of MCPs is not known, making it difficult to encapsulate heterologous enzymes without disrupting MCP morphology. We hypothesized that signal sequences may play a role in coordinating MCP core assembly. We tested this by systematically knocking out MCP signal sequences and assessing changes to MCP formation and enzyme targeting. We found that knocking out signal sequences, particularly those which give the highest levels of encapsulation, disrupts MCP assembly, but assembly can be rescued by complementing the knocked-out signal sequences fused to a heterologous protein. Through this process, we also discovered several previously unreported signal sequences which play roles in connecting the MCP cargo to the shell. Our results provide design constraints for encapsulating heterologous pathways in MCPs while maintaining proper MCP assembly and morphology.

Predicting Entropy of Polymerization Using Machine Learning to Accelerate the Design of Recyclable Plastics

Today’s plastics are difficult to recycle, and therefore plastic waste accumulates in landfills, oceans, and other natural habitats. We are now looking to synthesize intrinsically circular polymers (iCPs) from renewable feedstocks using hybrid chemical and biological pathways with the hope of eliminating the end-of-life problem and pollution of today’s plastics. These novel iCPs can undergo a process called chemical recycling through which the quality of the material is maintained as it is broken down into its building blocks for reuse, allowing the material to be recycled infinitely. In order to expedite the design of such materials, I am developing a machine learning model that can rapidly screen the vast molecular search space created by hybrid pathways for monomer candidates that have thermodynamic properties suitable for sustainable chemical recycling. I am specifically focusing on using machine learning with engineered features to expedite the prediction of entropy of polymerization, a property that is notoriously difficult to calculate using molecular modeling methods.

A genetically encoded selection for amyloid-beta oligomer binders

Soluble amyloid beta oligomers (AβOs) are a hypothesized source of neurotoxicity in Alzheimer’s Disease. Binding proteins that recognize these species may have high utility in diagnostic and therapeutic applications. However, identifying binders that recognize AβOs directly generated from the aggregation cascade is made challenging by the short lifetime and low concentrations of oligomer populations. We report a new strategy for detecting binding to AβOs as they form during Aβ42 aggregation using a genetically encoded biosensor. We show that our method enables rapid and highly reproducible measurement of the activity of existing AβO binders and can be used to select for new binders with improved potency. We uncover hits that are >20 fold more effective than reported binders at delaying secondary nucleation, the step in Aβ aggregation thought to generate the highest amounts of toxic oligomers. Our approach may greatly accelerate the discovery and characterization of binding proteins that target AβOs.

Altering terminal locations to modify the MS2 VLP with circular permutation

Virus-like particles (VLPs) are self-assembling nanoparticles derived from viruses that exhibit potentiality in vaccine development, drug delivery, and material application. We utilize the well-studied MS2 VLP, adapted from a single-stranded RNA virus, as a model system to study the self-assembly process of VLPs. Modifications to the termini of MS2 coat proteins (CPs) can lead to the failure of the MS2 VLP assembly, limiting its application in the peptide presentation and stable cargo encapsulation. We performed the circular permutation on a fused MS2 CP dimer to alter the terminal locations and create potentially accessible loci for peptide fusion. We generated a comprehensive circular permutation library of the CP construct and screened for the functional permutants capable of assembling into VLPs with new terminal locations. Using next generation sequencing (NGS), we calculated the apparent fitness landscape (AFL) to characterize the assembly competency of all circular permutants. Our AFL heatmap identifies a few positions tolerant to introducing the termini, providing information for further engineering of the MS2 VLP. Furthermore, our study quantitatively describes the essentiality of the connectivity among the secondary structure elements in protein folding and the protein complex self-assembly process.

Engineering vfFadB to Increase Medium-Chain Fatty Alcohol Production in Reverse β-Oxidation Pathway

Medium-chain fatty alcohols (mcFaOHs) are essential oleochemicals with widespread applications in the pharmaceutical and consumer product industries. Metabolic engineering provides a sustainable approach to produce mcFaOHs via the reverse beta-oxidation pathway (RβOx). Previous studies have identified the reduction from β-keto-acyl-CoA to 3-hydroxy-acyl-CoA, catalyzed by FadB, as the rate-limiting step in RβOx. Therefore, to increase mcFAOH productions, it is critical to increase the activity of FadB. Among several homologs of FadB, a homolog from Vibrio fischeri (vfFadB) has been shown to produce the highest alcohol titer, thereby selected as our starting point. This research focuses on improving vfFadB activity through error-prone PCR, generating a random mutagenesis library. By deleting all fermentation pathways in Escherichia coli, we successfully coupled cell growth and alcohol production under anaerobic production, enabling high-throughput screening of vfFadB variants. Through this study, we hope to investigate the mechanisms behind the mutants and further increase alcohol titer.

Phenotypic Classification and Prediction of Multi-Megadalton Self-Assembling Protein Complexes using a Site Saturation Mutagenesis Library.

Multi-megadalton structural protein complexes are of emerging interest to protein engineers for applications in medicine, global health, and biomanufacturing due to their biocompatibility and material properties derived from sequence-specific nanoscale architectures. Within this thrust, protein assemblies designed De Novo have largely dominated the discourse due to a rapidly expanding computational toolkit. However, De Novo protein designs need to be engingeeded to interface with living systems, and engineering their interactions with native biological systems can be computationally intensive. Engineering native systems may be a suitable alternative to imparting physiological relevance on engineered assemblies; however, the computational tools needed to re-design native protein assemblies while preserving function need to be developed further. This work develops a rigorous machine learning toolset trained on an in-vivo screen for the assembly of a homo-hexameric microcompartment shell protein to classify and predict mutant assembly phenotypes. Towards this goal, an encoding scheme for multi-megadalton structural protein complexes was developed to capture features relevant in the primary, secondary, tertiary, and quaternary structures without directly encoding the structural coordinates of the wildtype protein and mutants. The single site saturation mutagenesis library was encoded, and data labels indicative of the presence or absence of each phenotype were accordingly attributed to each data point. Four machine learning models were trained on combinations of primary and tertiary structural encodings and upon conducting grid search hyperparameter tuning, a maximum accuracy above 90% correct guesses to total guesses was achieved. In a similar, but more complex test predicting the apparent fitness of the phenotype the model best suited for predicting apparent fitness predicted above 65% correct guesses to total guesses.

Quantifying cellular mechanisms of Transcription factor dynamics and Transcriptional bursting in light-controlled Msn2 with live cell mRNA imaging

Cells respond to environmental stimuli by transmitting signals through signaling pathways that activate TFs, which in turn bind to DNA and control gene expression. Transcriptional bursting is a phenomenon observed across eukaryotes where transcription factors (TF) regulate gene expression in stochastic bursts of mRNA production. Transcription dynamics are shaped by the complex interplay of many variables. While multiple models have been proposed, a mechanistic understanding of the biological steps that cause transcriptional bursting is still missing. Stress-responsive TFs, such as the yeast TF Msn2, have been shown to translocate to the nucleus in response to environmental signals, leading to changes in the temporal dynamics of nuclear TF concentration which are then decoded by gene promoters to enact stimulus-specific gene expression programs. Recently, our lab investigated how promoter decoding of TF localization dynamics is affected by changes in the ability of the TF to bind DNA by using light-controlled mutants of the yeast TF Msn2 as a model system. We found that yeast promoters directly decode the light-controlled localization dynamics of Msn2 and that the effects of changing Msn2 binding affinity on that decoding behavior are highly promoter dependent, but the cellular mechanisms through which this occurs and how this affects transcriptional bursting requires further study with time-resolved mRNA measurements. To this end, we simultaneously control Msn2 TF localization with light and quantify bursting parameters with live cell mRNA imaging to provide insight into which steps are regulated by specific regulatory mechanisms and how this interfaces with TF concentration dynamics in S. cerevisiae.

Strategies to improve microbial medium-chain fatty alcohol production

Medium-chain fatty alcohols (MCFaOHs) are aliphatic primary alcohols containing six to twelve carbons that are widely used in materials, pharmaceuticals, and cosmetics. Microbial biosynthesis has been touted as a route to less-abundant chain-length molecules and as a sustainable alternative to current petrochemical and oil crop extraction processes. However, poor enzyme selectivity and product toxicity are arguably the largest barriers preventing microbial MCFaOH production from industrial realization. Substrates for MCFaOH biosynthesis are sourced from metabolic pathways that natively generate (fatty acid biosynthesis) or degrade (reversed β-oxidation) fatty acids. However, cells have evolved to rely almost exclusively on the long-chain (≥C16) substrates and products of these pathways. Subsequently, pathway enzymes have evolved to favor activity towards long-chain substrates, making enzymes with sufficient medium-chain activities and/or specificities scarce in nature. Further, the minimal need for medium-chain moieties has prevented cells from developing tolerance mechanisms to cope with the toxic effects of their presence. These issues necessitate reliance on engineering methods to overcome the challenges associated with large-scale biosynthesis of medium-chain compounds.

Single Point Mutations Impact Self-Assembly of PduA Supermolecular Structure

Protein self-assembly is a ubiquitous and essential process in organisms. Bacterial microcompartments—large, protein-based organelles that natively carry out processes such as niche-carbon source metabolism and cofactor recycling—are one such model for studying protein self-assembly. These structures are popular with synthetic biologists and bioengineers for their potential applications as drug delivery vehicles, for metabolic engineering, and as a novel biomaterial. These efforts are hindered by a lack of mechanistic insight into their self-assembly processes. Here, I evaluate two proteins, PduA and PduJ, which form most of the shell of the 1,2-propanediol utilization bacterial microcompartment from Salmonella enterica serovar Typhimurium LT2—one of the most studied microcompartment systems. PduA and PduJ have nearly identical primary sequences, and have been previously demonstrated to be redundant, genetically. Using cell-free protein synthesis (CFPS), I demonstrate how PduA and PduJ spontaneously assemble into distinct structures from one another, each on a scale of tens of microns to a millimeter, as protein gets synthesized. Additionally, I have characterized three mutants of PduA, which also spontaneously self-assemble into distinct structures easily distinguishable from the wild type. These can be utilized as a basis for engineering these structures in addition to elucidating mechanistic insight into how specific residues impact assembly of larger structures, such as the microcompartment self-assembly process.

Engineering a yeast cell factory for biosynthesis of tauro-ursodeoxycholic acid (TUDCA)

Microbial bile acids (MBAs) are a group of natural compounds produced through the microbial transformation of bile acids. These compounds have diverse structures derived from multiple ring scaffolds and possess a range of unknown biological functions, making them promising candidates for medicinal applications. However, their low natural production levels have hampered their investigation. Tauro-ursodeoxycholic acid (TUDCA) is an MBA that is known for its potent neuroprotective activities against Huntington's and Parkinson's diseases. Unfortunately, current methods for obtaining TUDCA, such as bear bile extraction or chemical semi-synthesis, are not sustainable and cannot support large-scale drug development. To overcome this challenge, our primary goal is to engineer the biosynthesis of TUDCA as a model system and develop a sustainable biosynthesis platform for a variety of MBAs. The Baker’s yeast, Saccharomyces cerevisiae, was chosen as the host organism because of its ability to produce the primary sterol metabolites for cholesterol synthesis and a wide range of genetic tools available for engineering. The TUDCA biosynthetic pathway is composed of at least fifteen heterologous genes from various sources, including human, fish, and gut bacteria. Cholesterol and chenodeoxycholic acid are key intermediates in the pathway and were used to divide it into three sub-pathways for parallel characterization. While the first and last sub-pathways have been verified, we are currently working on the middle sub-pathway, which is relatively more complex to construct, consisting of eleven genes compared to the other two sub-pathways, each composed of only two genes. This research aims not only to achieve the drug development of TUDCA but also to reveal the therapeutic functions of many other MBAs that have yet to be explored.

Using Cell-Free Protein Synthesis to Harness Plant Cytochrome P450 Enzymatic Activity

Plants produce a myriad of secondary metabolites that have been shown to exhibit therapeutic properties for human health. Current challenges in the chemical synthesis of these complex metabolites have pointed to the use of biocatalysts as an alternative synthesis method. Cytochrome P450s (CYPs) are a diverse group of heme-thiolate proteins that play critical roles in the biosynthesis of plant secondary metabolites. Oftentimes, functionally expressing CYPs can involve complex and time-consuming experiments. Cell-free protein synthesis (CFPS) is an appealing alternative to traditional expression systems due to its high throughput nature and the ability to optimize reaction conditions and incorporate necessary cofactors. We aim to functionally express plant-derived CYPs and characterize their activity using the CFPS platform. We intend to do this by tuning heme bioavailability and introducing membrane mimetics to encourage proper protein folding. Upon developing a platform that enables higher throughput testing of CYPs, we hope to elucidate CYPs with unknown functions and utilize their catalytic power to synthesize plant secondary metabolites on a larger scale.

Machine learning-guided optimization of bacteriophage specificity

The escalating threat of antibiotic resistance, coupled with a slowdown in the discovery of novel antibiotics, underscores the urgent need for alternative strategies to combat bacterial infections, and phage therapy holds promise as one viable alternative. Bacteriophages, viruses capable of infecting and lysing specific bacterial hosts, present one attractive avenue for targeted bacterial removal. However, research in applying phages as antibacterials has predominantly centered on mining existing phages for desired functionalities. This reliance on isolating and characterizing natural phages presents inherent challenges, including limitations in achieving high infectivity toward desired bacterial strains, tailoring phage specificity to target only pathogenic strains, and engineering phages with broader targeting profiles. To address these challenges, we propose directly designing bacteriophage function through machine learning (ML)-guided protein engineering. We focus on modeling the impact of mutations in the T7 bacteriophage receptor binding protein (RBP) on phage virulence toward five laboratory strains of Escherichia coli presenting distinct cell surface receptors. Fitness measurements for libraries of mutant phage variants were determined by incubating with bacteria and sequencing the populations before and after infection to determine the change in relative phage abundance. Using models trained on this experimental data, we design mutant phages with high infectivity, novel specificity, and generalist targeting capabilities. We achieve a high success rate for ML-directed phage design, even at high mutational distances and for complex specificities. Through analysis of the sequence trends in the phages generated for different targeting objectives, we demonstrate the capacity of neural networks to learn and navigate multifunctional protein fitness landscapes, thereby enabling the directed design of higher order protein specificity.

Exploring allostery in sensor histidine kinases

A key class of proteins bacteria use to sense and respond to their environment are sensor histidine kinases (HKs). HKs have diverse ligands, and are a rich source of potential biosensors. These proteins are membrane-embedded, with external sensor domains and cytoplasmic signaling domains, making them a fascinating case study for understanding rules of allostery in signaling proteins. Ligand binding residues, sensor domain orientation, and transmembrane domain positioning relative to the membrane can all impact the signaling state and ligand sensitivity of a histidine kinase. We have developed a high-throughput HK activity assay to comprehensively explore the function of variants in these allosterically active regions. From this screen, we will learn ways to tune ligand responsivity and adjust ligand specificity of histidine kinases. This will further development of more robust biosensors and a deeper understanding of histidine kinase signaling mechanisms.

Towards economical on-demand biomanufacturing with cell-free systems

There is a significant global need for rapid, on-demand production of medical therapeutics to address emergent biological threats. Current manufacturing processes for protein-based biologics are often time-consuming, expensive, and inaccessible to rural and developing regions. In contrast to established in vivo production platforms, cell-free protein synthesis systems take advantage of the cellular machinery in crude cell lysates to express proteins without the need for living cells and can be freeze-dried to readily store, distribute, and activate production by simply adding water. However, high reagent costs and comparatively low protein yields (~$3,500 for common reagents per gram protein) limit the widespread use of cell-free systems. Here, we seek to optimize cell-free protein synthesis systems to achieve low-cost, high-yielding, and scalable production of medically relevant proteins. We first analyze cell-free reaction metabolomics to identify problematic reagent degradation and accumulation associated with the use of low-cost, minimal reagent formulations. We then use CRISPR genome engineering to modify key steps in central metabolism to improve cell-free reagent processing and product yields. We also screen 136 purified proteins for their impact on the cell-free production of sfGFP and FDA-approved conjugate vaccine carrier proteins CRM197 and Protein D. Eighteen proteins were found to improve production of all three tested products by up to 250%. Our adjusted minimal cell-free systems are capable of producing >1.5 g/L of protein product for less than $200 of reagents per gram protein. This work furthers the development of cost-effective cell-free biomanufacturing for rapid protein therapeutic production.

An in Vitro Assay for Bacterial Microcompartment Function Enables Shell Permeability Estimation Via Model-Guided Sampling

With the advent of metabolic engineering, researchers have been optimizing metabolic pathways to sustainably produce commodity chemicals. However, more complicated metabolic pathways face challenges of low yield. Low yields can be caused by pathway intermediates participating in other pathways, toxic pathway intermediate build-up inhibiting cellular fitness, and cofactor competition with other pathways. Nature faces similar challenges with its own metabolic pathways and has evolved many strategies to address them, one being encapsulating select metabolic pathways in protein shells called bacterial microcompartments (MCPs). Encapsulation of a pathway inside of a porous protein shell creates a diffusion barrier that can potentially sequester pathway intermediates and cofactors, making them private to the encapsulated pathway. We aim to better understand the benefits of pathway encapsulation in order to confer these benefits to heterologous metabolic pathways. We use the 1,2-propanediol utilization MCP (Pdu MCP) as our model and have developed an in vitro assay to study the metabolic flux of the pathway outside of the context of the cell and allow us more direct manipulation of the reaction environment of the Pdu MCP. With this talk, we will discuss how this in vitro assay has allowed us to probe different properties of the Pdu MCP using metabolite detection and kinetic modeling.

Developing Engineering Strategies to Enhance the Genetic Stability of Fatty Alcohol-producing Strains for Production Scale-up

Industrial production requires stringent strain stability for successful scaling up, particularly when the target product is toxic to the host cells. This toxicity not only poses a metabolic burden, limiting productivity, but also promotes the growth of low-producing or non-producing subpopulations, leading to significant genetic heterogeneity and instability during large-scale fermentation. Yarrowia lipolytica has emerged as a promising host for the biosynthesis of fatty acid-derived oleochemicals due to its high lipid content. Among these compounds, fatty alcohols have garnered attention for their applications as surfactants and lubricants. However, the hydrophobic nature of fatty acyl chains causes their insertion into cellular and subcellular membranes, negatively impacting the production host. In this study, we aimed to develop systematic strategies to improve genetic stability of engineered Y. lipolitica strains producing fatty alcohols. In our mock test simulating fermentation scale-up, it was observed that the host completely lost its ability to produce fatty alcohols after five consecutive passages. To address this issue, we fused fatty alcohol reductase (FAR) with a monomeric gene, phosphoglycerate kinase I (PGK1), which is essential to the growth of the host. Additionally, a control strain expressing the FAR-GFP fusion was constructed. To assess long-term stability, the fatty alcohol-producing strains were cultured for five days, with one percent of the culture passaged into fresh medium every 24 hours over a total span of nine passages. Our results indicated that the fusion strategy between the essential PGK1 and FAR elongated the stability of fatty alcohol production by one additional passage. Interesting, fusing FAR and GFP extended genetic stability by four additional passages. In parallel, we constructed an inactive FAR mutant fused with BFP and co-cultured it with the high-producing variant (FAR’GFP) with different ratios to study the competition between a nonproducer and a high-producer. The results showed that when the non-producing mutant appeared as a frequency of 10-5, it took approximately six passages for it to become the dominant genotype in the culture. If it appeared at a frequency of 10% , it only took two passages. To understand the mechanisms of genetic instability, genomic DNA from various strains demonstrating different levels of stability and fatty alcohol productivity were isolated for deep sequencing of the FAR expression cassette. By building an understanding of the patterns, dynamics and causes of mutagenesis in engineered yeast, we are able to develop effective strategies to mitigate undesirable mutations and optimize the evolutionary stability of the high-producing strain for large-scale fermentation.

Deciphering Sequence Determinants of Specificity in Molecular Chaperones: Towards Novel Protein Folding Therapeutics

Heat shock protein 70 (Hsp70) is a ubiquitous molecular chaperone that regulates conformational protein folding in the cell. Working in concert with co-chaperones from the Hsp40 family and nucleotide exchange factors, Hsp70s undergo ATP-driven conformational changes that allow for the cyclical binding and release of client proteins to facilitate folding. Given the diverse composition of cellular proteomes across species, the highly conserved Hsp70 chaperones are known to act as generalists, displaying a wide substrate specificity and serving many distinct clients. However, despite significant advances in understanding the function of Hsp70s, the molecular rules governing biochemical promiscuity in these chaperones are largely unknown. Here we establish the molecular determinants of specificity in Hsp70s by systematically investigating the sequence-function landscape of DnaK (E. coli Hsp70) through deep mutational scanning, setting the stage for rapid rational design and engineering of Hsp70s for numerous biotechnological and medical applications.

Discovery of high efficiency oligosaccharyltransferase mutants for conjugate vaccine synthesis using a high-throughput cell-free screening platform

Glycoconjugate vaccines, composed of bacterial capsular polysaccharides or O-antigen glycans conjugated to immunogenic carrier proteins, are an effective strategy to promote immunity and prevent bacterial infections. The conventional method to synthesize conjugate vaccines uses chemical techniques to non-specifically attach bacterial glycans to a purified carrier protein. While functional, this process is expensive, requires growing pathogenic bacteria, and modifies protective immune epitopes on the carrier. Recently, enzymatic conjugation methods have been developed in both cell-based and cell-free systems to conjugate glycan antigens to carrier proteins using an oligosaccharyltransferase, such as PglB from Campylobacter jejuni. This enzymatic method addresses the limitations of chemical conjugation and enables the synthesis of a homogeneous and more highly immunogenic vaccine product. Unfortunately, enzymatic conjugation is currently limited by low glycosylation efficiency, or the proportion of aglycosylated carrier protein that becomes glycosylated in an enzymatic reaction. Improving glycosylation efficiency is therefore essential to increase vaccine yield and produce more vaccine doses per reaction. In this work, we evaluated the effect of mutating the oligosaccharyltrasnferase PglB on the enzyme’s glycosylation efficiency. We designed a library in which 15 amino acids at chosen sites within PglB were mutated to all 19 other amino acids, creating a library of 285 single-mutants. We used cell-free protein synthesis to express each mutant and added the mutants to in vitro enzymatic reactions. AlphaLISA, an in-solution bead-based ELISA assay, was then performed to detect carrier glycosylation and evaluate each mutant’s glycosylation efficiency in high-throughput. In total, 7 mutants demonstrated higher glycosylation efficiencies than wildtype PglB, and AlphaLISA results for both high- and low-signal mutants were confirmed by western blot. Furthermore, 7 unique amino acids that were chosen for mutation generated at least one mutant with signal >5x above background. This work reveals numerous novel mutations within an oligosaccharyltransferase that improve its glycosylation efficiency. Future research will expand this mutagenesis screening towards an array of pathogen glycans to improve enzymatic conjugation technology and synthesize novel conjugate vaccines.

Engineering CIRTS to Rebalance Protein Expression from mRNA in Haploinsufficiencies

RNA-binding proteins (RBPs) play a key role in the fine control of protein expression through their regulation of mRNA half-life and translation. Leveraging this functionality by the fusion of RBP domains that regulate mRNA translation with proteins that possess programmable transcript targeting would represent an exciting new modality to address challenges in correcting deficient protein expression in disease-causing haploinsufficiencies. However, the deployment of engineered RBPs directed by programmable RNA-binding platforms such as CRISPR-Cas Inspired RNA-targeting System (CIRTS) for this purpose has been hindered by the limited availability of characterized protein domains to functionalize engineered RBPs. In order to expand the regulatory capabilities of programmable RBPs we have developed novel reporters to characterize human RBP activity when directed by CIRTS to sequence elements of disease relevant transcripts. We then identified novel CIRTS fusions capable of tuning protein expression from targeted transcripts and validated their activity boosting protein expression from endogenous disease-relevant transcripts in cellulo, in primary rat neurons, and in vivo by increasing Nav1.1 expression in a SCN1a haploinsufficiency mouse model. Collectively, the newly characterized CIRTS fusions greatly expand the programmable RBP toolbox for tuning protein expression from mRNA and demonstrate generalizable activity across several tested targets. We also determine pipelines for identifying effective guide RNA for top CIRTS fusions which should support user-friendly and streamlined design of CIRTS within AAV-packaging limits capable of addressing gene-dosage related disease for basic biological research and potential therapeutic application.

Engineered Probiotics for Biomarker-Responsive Treatment of Inflammatory Bowel Disease

Whole-cell sense-and-respond systems have emerged as a platform for targeted therapeutic delivery. Because of bacteria’s natural ability to colonize the gastrointestinal tract, they are a logical choice for targeted delivery to the gut.1 Additionally, the use of a live-cell platform allows for the design of tunable therapeutic release that is coupled to sensing of biomarkers of disease. We aim to design a probiotic strain of E. coli that constitutively produces therapeutics and releases them when high levels of calprotectin, the clinical gold-standard biomarker of gut inflammation, are sensed in the gut during inflammatory bowel disease (IBD). IBD is a group of autoimmune disorders that are characterized by chronic cycles of remission and relapse of painful symptoms caused by gut inflammation.1,2 Because of this cyclic nature, development of a therapeutic that adjusts dosage in response to disease state will allow for fine-tuning of treatment to disease course. A cornerstone of IBD treatment is the use of monoclonal antibodies (McAbs) that have been designed to block pro-inflammatory cytokines implicated in IBD.3 However, the majority of these therapeutics are delivered systemically, which results in a number of serious side effects.3 Encapsulation of these McAbs in a bacterial chassis would allow for direct delivery to the gut, greatly reducing this risk. We have previously developed a whole-cell sensor for calprotectin, and we have shown that we are able to produce a number of IBD therapeutics in bacteria.

 1. Xia JY, et al. Engineered calprotectin-sensing probiotics for IBD surveillance in humans. PNAS. 2023;120(32).

 2. Gareb B, et al. Review: Local Tumor Necrosis Factor-α Inhibition in Inflammatory Bowel Disease. Pharmaceutics. 2020;12(6):539.

 3. Griffiths OR, et al. Chapter Five - Inflammatory bowel disease and targeted oral anti-TNFα therapy. In: Donev R, ed. Advances in Protein Chemistry and Structural Biology. Vol 119. Inflammatory Disorders, Part A. Academic Press; 2020:157-198.

Characterization of the Reversed β-Oxidation Pathway for Improved Fatty Alcohol Bioproduction

The β-Oxidation pathway, normally involved in the catabolism of fatty acids, can be functionally reversed to act as a fermentative, iterative, elongation pathway when driven by the activity of a trans-enoyl-CoA reductase. The carbon-carbon bond formation mediated by thiolases, together with the terminal acyl-CoA reduction that can occur on substrates with varied chain lengths, unlock the potential for the production of a wide range of aliphatic compounds (e.g., fatty acids, alcohols, β-hydroxy-carboxylic acids). The inherent properties and commercial value of these oleochemicals are dependent on their functional groups and, chain length. Therefore, tight control of the average chain length and product profile is desirable. Lacking a termination enzyme with a narrow chain length preference, we sought alternative factors that could influence the product profile and pathway flux in the cyclic pathway. In this study, we reconstituted the reversed β-oxidation (R-βOx) pathway in vitro with a purified tri-functional complex (FadBA) responsible for the thiolase, enoyl-CoA hydratase and hydroxyacyl-CoA dehydrogenase activities, a trans-enoyl-CoA reductase (TER), and an acyl-CoA reductase (ACR). Through a combination of a design-of-experiment (DOE)-based exploration and individual enzyme assays, we were able to determine the rate limiting step of the R-βOx pathway. We also demonstrated that by controlling the enzyme ratios and the ratio of NADH and NADPH, the product profile can be tuned from primarily short-chain (C4 and C6) products, up to long-chain profiles of primarily C14 and longer alcohols. Additionally, we investigated how the chain length and distribution of intermediate metabolites correlate with alcohol product profile via LC-MS quantification. Our results emphasize the necessity of careful consideration of enzyme ratios in future studies, and highlight potential enzyme engineering targets for increasing flux in the alcohol-producing R-βOx pathway.

Probing The Type Three Secretion System (T3SS) Machinery Assembly in Salmonella enterica: A High Throughput Method Towards Enhanced Heterologous Protein Secretion

The Type Three Secretion System (T3SS) in Salmonella enterica represents a novel avenue for scalable protein production in which the protein product is secreted to the media during production resulting in a relatively pure product from fermentation. This strategy potentially overcomes limitations in industrial bacterial protein production systems such as the need to refold protein from inclusion bodies and/or remove contaminant proteins introduced during lysis. However, this new production modality will be more attractive for commercialization when optimized for higher yields and under industrial fermentation conditions. Improving the efficiency and scalability of protein production through the T3SS requires precise control over T3SS activation, assembly of the secretion apparatus, and secretion of targeted proteins. To enable engineering efforts to this end, we set out to achieve a detailed understanding of the mechanism of apparatus assembly and the impact of our engineering efforts on quantity versus activity of the secretion machinery. Recent efforts have enabled the single cell visualization of individual secretion apparatuses1. We have adapted the method for Super Resolution Structured Illumination Microscopy (SR-SIM) and correlated these low-throughput measurements to a high throughput flow cytometry-based method to probe the assembly state and semi-quantify the number of secretion apparatus per cell. This has allowed us to construct a comprehensive model linking T3SS activation, the number of apparatus formed, and subsequent protein secretion yields. Our findings offer a deeper understanding of T3SS assembly, providing a foundation for its engineering in heterologous protein production applications.

1. Zhang, Y., Lara-Tejero, M., Bewersdorf, J. & Galán, J. E. Visualization and characterization of individual type III protein secretion machines in live bacteria. Proc. Natl. Acad. Sci. U. S. A. 114, 6098–6103 (2017). 

Enhancing Xylose Metabolism to Produce 3-Hydroxypropionic Acid (3-HP) from Cellulosic Hydrolysates by Engineered Issatchenkia orientalis

Xylose is a promising renewable resource for microbial bioproduction due to its abundance in cellulosic hydrolysates in addition to glucose. To advance the conversion of cellulosic hydrolysates to target chemicals, efficient and rapid utilization of xylose and glucose in microorganisms plays a key role. Most microorganisms can natively assimilate glucose, while the lack of an efficient xylose metabolic pathway remains a significant challenge for them to efficiently convert cellulosic hydrolysates to desired products. Therefore, an efficient xylose metabolic pathway was introduced by expressing corresponding genes in the non-model yeast Issatchenkia orientalis capable of producing 3-HP. The resulting strain can efficiently and rapidly utilize both glucose and xylose from cellulosic hydrolysates for 3-HP production with a high yield. This study highlights the potential of efficient conversion of cellulosic hydrolysates for producing desired chemicals.

High-Throughput Screening for Engineered Adhesive Proteins

Nature exhibits a vast array of protein sequences with adhesive properties, inspiring the development of synthetic biology tools to engineer novel proteins for advanced materials applications. This study focuses on the development of a high-throughput screening method to efficiently evaluate the adhesive capabilities of a diverse protein library. By leveraging synthetic biology, we aim to create a platform for rapid identification of protein sequences with superior adhesive properties, ultimately enabling the design of smart and high-performance materials. Our approach combines protein engineering with advanced screening techniques to accelerate the discovery of novel adhesive proteins with tailored functionalities.

Developing Light-Controlled Synthetic Zinc Finger Transcription Factors for Budding Yeast

Synthetic zinc finger transcription factors (TFs) are powerful tools that have been used to create multigene control, logic gates, and cooperativity. However, previous studies of these TFs have not fully explored optogenetics which enables rapid and precise control of nuclear localization dynamics. Based on computational modeling results of said dynamics, we constructed optogenetically controlled synthetic zinc finger TFs with differing nuclear import/export rates and nuclear TF saturation levels by balancing nuclear localization signals, nuclear export signals, and anchor proteins that sequester the TFs to the plasma membrane. In accordance with our models, a) TFs with faster import and slower export resulted in pulsatile light input yielding higher reporter expression than continuous light, and b) TFs with slower import and faster export resulted in the opposite expression pattern, showing the power of our model guided design strategy. Furthermore, we altered the number of transcription factor binding sites (TFBSs) in the cognate synthetic promoter and observed a non-linear relationship between basal level, induced expression level, and the number of TFBSs. This study lays the groundwork for designing more advanced gene expression systems with broad implications for synthetic biology and bioengineering by incorporating synthetic zinc finger TFs and cognate promoters into optogenetics.

Towards effective screening systems to identify prion degrading enzymes

Chronic Wasting Disease (CWD) is a prion disease that causes fatal neurodegenerative disorders and is becoming a widespread issue among cervids in North America. Prions are misfolded proteins that can induce the conformational change of the major prion proteins (PrP) into an infectious form (PrPSc). Prions are highly stable molecules and are found to be accumulating in the environment/soil from contaminated feces and infected cervid carcasses. Current ways for prevention include cleaning the soil with chemicals, high levels of heat or physically removing the soil from the environment. New efficient, practical and safe methods are needed. Enzymatic degradation of prions could provide a more environmentally conscious and efficient approach in removing prions present in the soil. Proteases have an intrinsic ability to break down peptide bonds of proteins, however may need engineering to increase their activity on prions and increase their environmental stability. Such engineering efforts require screening mutant proteases, and we established an assay to evaluate the activity of these proteases on amyloid-like structures. We used Hen’s Egg White lysozyme (HEWL) chemically converted into fibrils as a surrogate for prions to allow for rapid and safe identification of improved mutants. Efforts pertaining to protease engineering are currently ongoing.

Harnessing Endogenous Type II-A CRISPR System to Achieve Genome Editing in Lactocaseibacillus rhamnosus GG to Develop New Probiotic Strains

Lacticaseibacillus rhamnosus GG (LGG) is a widely applied probiotic used by the food and pharmaceutical industries. However, the genetic bases of its beneficial properties are mostly uncertain because of the lack of effective genetic manipulation tools. CRISPR-Cas systems are now widely used for genome editing and transcriptional regulation in diverse organisms. The type II system is the most abundant CRISPR-Cas system found in lactobacilli. In this study, we characterized the type II-A CRISPR system in LGG and successfully established an endogenous CRISPR-Cas9 genome-editing system. Through computer prediction and subsequent confirmation via plasmid interference assays, we identified the CRISPR array and the protospacer adjacent motif (PAM) as 5’-NGAAA-3’. With a PAM and customized single guide RNA (sgRNA) expression cassette, the native CRISPR-Cas9 system was successfully reprogrammed to achieve gene deletion. Furthermore, by harnessing this endogenous CRISPR system, we constructed a lactose-positive LGG strain, named MJM570. In contrast to its parental strain LGG, the ability of MJM570 to metabolize lactose enabled strong and fast growth in milk. This lactose-positive strain holds potential as a probiotic starter culture for dairy fermentations and could be beneficial in treating lactose intolerance when used as an additive.