Metabolism-driven, high-efficiency mining of ethanol-tolerant microorganisms from pit mud microbiota using Raman flow cytometry

Abstract

The discovery of stress-tolerant microorganisms from complex microbiomes is frequently constrained by low screening throughput and the inability of culture-based approaches to access single-cell functional phenotypes, particularly for rare but highly resilient taxa. Here, we establish a metabolism-driven screen-before-culture strategy by integrating D2O-labelled single-cell Raman spectroscopy (SCRS) with Raman-activated cell sorting (RACS) to directly target ethanol-tolerant cells based on metabolic vitality. By quantifying carbon-deuterium (C-D) incorporation as a single-cell readout of de novo anabolic activity under ethanol stress, this platform enabled culture-independent enrichment of a rare, high-vitality subpopulation ([~]0.2% abundance) from pit mud microbiomes at a sorting throughput of [~]2,400 cells per hour. Using this approach, six highly ethanol-tolerant strains were successfully isolated, whereas parallel conventional culture-first screening of the same samples yielded predominantly low-tolerance isolates, with an overall screening efficiency of only 22.22%. Rapid single-cell ethanol tolerance assessment based on SCRS, completed within 7 h, showed that all sorted strains exhibited strong tolerance to 8% (v/v) ethanol, with Raman Tolerance Index (RTI) values exceeding 50%. Among them, Lactiplantibacillus plantarum F4 displayed the highest tolerance (RTI = 85.05 {+/-} 3.41%). Comparative transcriptomic analyses of representative strains revealed mechanistically coherent ethanol adaptation strategies, including ethanol-derived carbon recycling, dynamic membrane lipid remodeling, and reinforced redox homeostasis. These responses directly underpin the metabolic activity captured by the Raman screening signal, validating its physiological relevance. This integrated SCRS-RACS workflow achieved orders-of-magnitude higher screening throughput, a 4.5-fold improvement in sorting accuracy, and a 6.86-fold increase in assessment efficiency compared with conventional methods. This study establishes a versatile, metabolism-based paradigm for the targeted mining of rare, stress-tolerant microorganisms from complex microbiomes, with broad implications for industrial biotechnology and microbial ecology.