Whole-cell SERS fingerprints for rapid toxicant attribution in water

Effect-based bioassays provide rapid warning of adverse impacts but rarely identify the causative stressor, whereas targeted chemical analysis identifies selected pollutants yet offers limited toxicity information, particularly for mixtures. Here, we evaluate whole-cell surface-enhanced Raman scattering (SERS) fingerprinting as a rapid, label-free analytical approach to link biological response patterns with toxicant discrimination. Specifically, we assess whether SERS-derived bacterial spectral fingerprints can differentiate chemically distinct classes of toxicants and whether classification performance is consistent across different bacterial sensor types. Gram-negative Escherichia coli and Gram-positive Staphylococcus aureus were exposed to five model toxicants, including heavy metals (Cu²⁺, Cr³⁺, and Pb²⁺) and organic toxicants (4-chlorophenol and formaldehyde). Time-resolved SERS spectra were collected using fixed-interval sampling and analyzed using principal component analysis followed by linear discriminant analysis (PCA-LDA). The resulting fingerprints were toxicant-specific and exposure-time-dependent. For a given toxicant, broadly comparable response trends were observed between E. coli and S. aureus, indicating potential cross-species consistency of the SERS fingerprinting approach within this model system. PCA-LDA achieved high multi-class classification performance, reaching 99% accuracy for E. coli at 2 h and 100% accuracy for S. aureus at 1 h. Additionally, the method was successfully applied to spiked river and lake water samples, achieving high accuracy in distinguishing various toxicants despite the presence of background interference. These results demonstrate the applicability of whole-cell SERS fingerprinting for distinguishing representative metal and organic toxicants and support the use of multi-strain biosensor strategies for rapid toxicity screening and preliminary toxicant discrimination in water monitoring applications.