Biocatalysis offers sustainable solutions to pressing challenges in chemical synthesis by exploiting the remarkable efficiency and selectivity of enzymes. Importantly, enzymes are able to accommodate non-native substrates and mediate transformations outside of their natural repertoire. Enzymes can be engineered for diverse applications by harnessing these 'promiscuous' activities and optimizing them using directed evolution (DE). The success of a DE campaign, however, depends on the availability of a protein starting point that displays detectable levels of the desired function. To find a starting point, researchers often screen libraries of protein variants for novel activities, typically with low rates of success. Here, instead, we diversified the active site of a desirable 'parent' protein and applied machine learning to generate informed, promiscuous libraries of protein variants. Specifically, we tested 26 different carbene and nitrene transfer reactions and used active learning-assisted directed evolution (ALDE) to generate optimized protoglobin variants with high activity across multiple reactions. We observed improvements in activity and selectivity for every reaction performed by the parent enzyme in at least one member of the ALDE-predicted libraries. Moreover, variants from these libraries can catalyze 5 out of 10 reactions not catalyzed by the parent protoglobin. These results indicate that supervised machine learning can help guide the construction of high-value enzyme libraries with expanded catalytic scope.

Hydrogen sulfide (H2S) is a microbiota-derived metabolite in the gastrointestinal tract implicated in a number of diseases. Its volatility and reactivity make experimentally controlling H2S concentration in vivo difficult, limiting our ability to interrogate its dose-dependent effects on host physiology. Engineered bacteria present a compelling solution, yet most probiotic metabolic engineering approaches have focused on in vitro optimization, failing to account for the complex intestinal environment. Here, we engineered E. coli strains to produce or consume H2S in specific intestinal regions by incorporating knowledge of the local metabolic environment and resident microbial activities into the design process. Analysis of human-derived ex vivo cultures revealed that glutathione (GSH) is inefficiently converted to H2S, suggesting GSH as a relatively stable substrate for engineered sulfide production. We thus engineered a GSH-dependent H2S producer, which increased levels 21-fold ex vivo. To target the nutrient-rich, microbially sparse environment of the small intestine, we optimized a H2S producer that uses L-cysteine as a sulfur source, demonstrating a 7-fold increase in H2S levels in mice. Finally, to develop strains capable of sequestering H2S, we leveraged the availability of fumarate and nitrate as electron acceptors in the large intestine by engineering a strain expressing sulfide:quinone oxidoreductase (Sqr). This enables oxidation of H2S to intracellular polysulfides and achieves higher consumption rates than alternative sequestration strategies reliant on resource-intensive GSH production. Together, this work developed engineered microbes as precision tools to modulate H2S levels and showcases a generalizable framework for region-targeted design of engineered probiotics.

Halophilic chassis has emerged as a promising biomanufacturing platform for industrial polyhydroxyalkanoate (PHA) production. However, challenges still remain in improving the production capacity, scalability and robustness, thereby lowering cost to meet market demands. Here, a high-performing halophilic strain Halomonas LY03 was isolated with over 38% glucose-to-PHA conversion rate and broad non-grain substrate utilization capability. Multidimensional tools, including algorithm-guided high-expression neutral integration site (HENIS) screening toolkit designated 'SiteSeek', stop codon (TAA)-dependent enhancement of gene expression and recombinase-mediated large-fragment (> 9 kb) genomic integration, were then developed to enable precise, efficient and interference-free genomic integrative expression. Using these tools, various chromosomally engineered strains were rapidly constructed to achieve high-level production of poly-3-hydroxybutyrate (PHB, 151 g L-1) and poly(3-hydroxybutyrate-co-4-hydroxybutyrate) (P34HB, 139 g L-1) under high cell-density fermentation (up to 186 g L-1 cell dry weight) in a 5-L bioreactor. Scalability was demonstrated at 2-m3 and 20-m3 industry-scale fermentations, yielding up to 134 g L-1 PHB and 127 g L-1 P34HB (6.1 mol% 4HB). Building on the proven robustness, a two-stage continuous fermentation (TCF) process was developed using a twin-bioreactor system at 5-L and 20-m3 scales, where stable and sustained PHA production lasted over 260 h and 160 h, respectively. Techno-economic analysis revealed a substantial cost-reduction space of 48% compared with conventional fed-batch process. This study demonstrates a successful paradigm for engineering a newly isolated strain toward robust, high-titer and cost-competitive PHA production across lab-industry scales.

Untargeted metabolomics is a powerful approach for exploring the chemical diversity and dynamics of biological systems. However, the types of questions that can be addressed depend not only on experimental design but also on the data processing and analysis workflows used, many of which require advanced computational expertise. GNPS, now transitioning to its second major implementation (GNPS2), has evolved into an expandable platform that supports the integration of modular web applications designed to simplify and enhance downstream analysis. These apps, named MetaboApps, facilitate the postprocessing of outputs of several GNPS workflows and help make repository-scale metabolomics knowledge and other areas of metabolomics more accessible to a broader community.

Biosensors enable the in situ measurement of metabolites in living systems over time and space. Fully genetically encoded metabolite biosensors (fGEMBs) use fluorescent proteins (FPs) linked to ligand binding domains (LBDs) to transduce the ligand binding event to a measurable change in the fluorescence behavior of the FP. Because these sensors are genetically encoded, they can be expressed in cells using standard protein expression approaches, and the fluorescence changes are quantified using fluorimetry, fluorescence microscopy, and/or flow cytometry. While there are general sensor design principles to follow, an fGEMB must be engineered for each metabolite based on a particular LBD. This development process can be slow, but there are strategies emerging to increase testing throughput and improve structure-guided design. While genetically-encoded FPs remain popular, there are now numerous chemigenetic and nucleic acid-based metabolite sensors (cGEMBs) that incorporate small molecule fluorophores. De novo design of LBDs is rapidly advancing as well, and the field may soon exhibit a shift away from relying on nature’s catalog of LBDs. Despite the engineering challenges, the metabolite biosensor field has expanded significantly in recent years to meet the demand for new and better-performing sensors that visualize metabolites within their cellular environments.

Antimicrobial resistance and tolerance pose escalating global health threats, necessitating reproducible and accessible tools for antimicrobial susceptibility testing (AST). While disk diffusion assays (DDAs) and Epsilometer tests (Etests) are widely used, there are limited open-source tools to analyze them. We present J-AST, a free, open-source, web-based platform for analyzing both DDAs and Etests. It provides automated and interactive annotation of regions of interest and metadata management, and quantifies microbial resistance and tolerance. J-AST outputs correlate strongly with those of existing tools, and equivalent DDA and Etest results correlate strongly with each other. The automated MIC detection achieved >90% agreement with manual readouts. J-AST is deployable both as desktop software and cloud service, unifies automated analysis with interactive review, and advances both fundamental research and clinical AST workflows. mic