Global surveys of microbial communities across biomes have shown that environmental variables such as depth and pH are strong determinants of community composition. However, we do not understand how the traits of individual taxa, and their evolutionary conservation, conspire to give rise to these patterns. Exploiting large-scale surveys of top soil and marine microbiomes, we use canonical correlation analysis (CCA) to concurrently infer directions of environmental variation and the associated compositional changes. We find that the primary canonical direction, capturing the dominant environmental gradient, exhibits a strong phylogenetic signal: individual species' responses to environmental shifts along this direction are similar among taxa with shared evolutionary history. In contrast, secondary canonical directions show weak or no phylogenetic structure. Together, these results suggest a two-scale view of microbial community assembly. Deeply evolutionarily conserved traits govern community reorganization along the main environmental driver of community composition. Additional environmentally driven changes in community composition then reflect traits that are more evolutionarily labile.

Bilirubin, the predominant product of heme catabolism in mammals, enters the intestine via the hepatobiliary system and subsequently is metabolized by the gut microbiome. This process consumes bilirubin and generates multiple downstream derivatives, such as urobilinogen and stercobilinogen. Levels of bilirubin and its derivatives are associated with susceptibility to inflammatory and metabolic disorders, but the microbial species and enzymes that metabolize bilirubin have remained largely unknown. Here, demonstrate that metabolism of bilirubin to urobilinogen requires two separate reactions that can occur in either order and identify novel enzymes and pathway intermediates required for conversion. We find that bilirubin reductase (BilR), an enzyme that was recently discovered and proposed to convert bilirubin to urobilinogen, is specific for reducing the methine bridges of bilinoids, converting bilirubin to the novel intermediate divinylurobilinogen and mesobilirubin to urobilinogen. Using transcriptomic profiling, we identify the bilinoid vinyl reductase (BilV) responsible for reducing the vinyl groups of bilirubin and divinylurobilinogen. BilV is a flavin-dependent oxidoreductase of the Old Yellow Enzyme (OYE) superfamily with a broad distribution across human gut bacteria that overlaps with but does not completely mirror the distribution of BilR. These findings establish the complete pathway for bacterial conversion of bilirubin to urobilinogen, enabling defined studies to interrogate how this metabolism contributes to human health and disease.

Commercially available make-on-demand libraries now exceed 100 billion compounds, requiring over 50 years to screen on 2,000 CPU cores using conventional docking. We present two complementary approaches to address this challenge. CombiDOCK, a combinatorial docking framework, enables exhaustive screening at the 100-billion scale within 40 days. MINT-Dock, a generative framework, accelerates navigation of this space by integrating CombiDOCK with Monte Carlo Tree Search. Benchmarked on 46 diverse targets, CombiDOCK matched full-molecule docking accuracy, and MINT-Dock achieved a 4,800-fold enrichment over random selection. Compared with prior billion-scale brute-force campaigns against σ2, VMAT2, and VAChT, prospective CombiDOCK screens of the 100-billion-molecule library yielded higher hit rates and more potent ligands, while MINT-Dock achieved comparable outcomes across single- and multi-target objectives with >20-fold computational cost reductions. Docking-predicted poses of the best VAChT-binding compounds were confirmed by cryo-EM structures. These methods provide exhaustive and generative paths for navigating the trillion-molecule frontier of drug discovery.

Acid mine drainage (AMD) waters are a global environmental threat due to their extremely low pH (<3) and high metal loads. Acidophilic sulfate-reducing bacteria (aSRB) can mitigate AMD by reducing sulfate to sulfide, a proton-consuming process that also precipitates metals as metal sulfides. Although sulfate reduction has been observed in AMD waters, most characterized aSRB are only moderately acidophilic. Here, we examined the pH tolerance and proton stress adaptation of the complete organic acid-oxidizing aSRB Acididesulfobacillus acetoxydans. Continuous chemostat cultivations were operated across a pH gradient, reaching steady states from pH 5.0 (optimum) to pH 2.9. In subsequent batch incubations, biomass from a pH 2.9 chemostat remained metabolically active at pH 2.5. Transcriptomic profiles remained remarkably stable across conditions, except for the upregulation of the K+-transporting ATPase (kdpABC) at lower pH, suggesting an increased reliance on the chemiosmotic gradient to impede proton influx. Lipid analysis revealed increased core lipid saturation, midchain methylation, and a shift in priming precursors from leucine to valine at low pH, indicating reduced membrane permeability and more energy-efficient biosynthetic pathways. Together, these adaptations likely reduce proton entry, explaining how aSRB adapt to AMD-like acidity and unlock the pH bottleneck for AMD bioremediation and metal recovery.

Plant-derived metabolites in foods and herbal medicines are frequently transformed by microbial communities before they reach systemic circulation. Consequently, the chemical forms responsible for biological activities often differ from the compounds originally present in plant matrices. This review examines the microbial biotransformation of plant metabolites through a mechanism-focused framework that integrates enzymology, reaction chemistry, and community-level context. We organize known transformations into recurring reaction classes, including hydrolytic activation, redox remodeling, functional-group tailoring, and scaffold rearrangement. We also discuss how omics approaches enable pathway attribution and enzyme prioritization in complex microbial communities. Finally, we highlight the emerging strategies that modulate these transformations through dietary composition, fermentation, and metabolite-focused interventions.

Correlations between cellular variables, such as gene-expression levels, provide insights into regulatory mechanisms. We focus here on correlations between mRNA and protein levels and re-examine previously derived analytical predictions. We test this prediction on single-cell E. coli data and see substantial disagreement. We hypothesize that this discrepancy arises from the assumption of constant cell volume and develop a theoretical framework for mRNA–protein correlations in growing and dividing cells. Within this framework, we derive an analytical expression for mRNA– protein correlations and show that explicit incorporation of growth and division substantially alters these correlations. The resulting relation is invariant to upstream transcriptional dynamics, and we validate it using stochastic simulations across multiple gene-regulatory architectures. Finally, we show that the derived predictions are consistent with the E. coli data.

Bacterial surface structures have enabled display systems with broad impact across biotechnology, but their narrow host range limits their deployment into diverse species. Conversely, Type IV Pili (TFP, T4SS) are ubiquitous, structurally conserved appendages found across bacteria, but have been minimally explored for display. Here, we describe a computational framework for predicting viable insertion sites in major pilins for stable TFP-mediated display, which we apply to the major pilin PilA1 of the cyanobacterium Synechocystis sp. PCC 6803 to enable covalent binding to living materials. By analyzing an Alphafold3-generated PilA1 monomer alongside known multimeric TFP multimers, our pipeline identifies non-interfacial, solvent-accessible, and flexible sites for optimal PilA1 display. We probe these sites with both full-length and truncated SpyCatcher003 at two different expression levels. We show that cells expressing these PilA1-SpyCatcher 003 fusions maintain up to 8-fold higher levels of cell suspension than previous C-terminal PilA1 display platforms, suggesting improved TFP assembly despite more than a two-fold increase in cargo size. Additionally, we validate SpyCatcher003 reactivity across the engineered strains, enabling covalent attachment of SpyTag 003-containing proteins on the Synechocystis surface. Lastly, we utilize this covalent patterning to achieve a four-fold increase in Synechocystis loading into a living material without compromising its viscoelastic or mechanical properties. Taken together, this work provides a predictive framework for TFP engineering, and opens the door towards programmable surface display across the breadth of bacterial species.

Precise timing of gene expression is a common feature in natural regulatory networks but is rarely implemented in synthetic pathways, where static control often limits performance. Here, we expand the use of bacterial extracytoplasmic function (ECF) σ factors by integrating their cognate anti-σ factors to create tunable threshold-gated circuits. Anti-σ overexpression has previously limited the utility of such systems due to growth inhibition. We address this by combining targeted truncations of membrane domains with chromosomal integration, yielding a library of ECF/anti-σ pairs that maintain function while minimizing toxicity. In E. coli, these genetic circuits can enable sharper OFF/ON switching, greater dynamic range, and tunable temporal delays compared with ECF-only cascades. In the best case, single-step threshold gates achieve up to 2100-fold induction with delays spanning minutes to hours. Extending the design to two steps enables programmable cascading of gene expression, with delays ranging from 30 to 409 min, although the dynamic range is generally reduced relative to single-step circuits. In the best-performing designs, matching input-output characteristics of promoters preserves dynamic range across cascade levels. Mathematical modeling supports these findings and highlights σ/anti-σ binding affinity as a key parameter for achieving high performance in longer cascades. Together, these results provide design principles for orthogonal, lower-burden timing circuits that enable controlled, sequential gene expression with minimal intervention in synthetic pathways and highlight continuing limitations of these systems.