Image cytometry and kinetic modelling reveal amendment-associated differences in microbial cell state and apparent...

Organic amendments are widely applied to enhance bioremediation of PAH-contaminated soils; however, remediation performance varies with amendment dosage, and the mechanisms linking dose, microbial physiological state, and degradation kinetics remain unclear. This study integrated kinetic modelling with imaging-enabled cytometry and machine learning to evaluate how aged leaf-biomass amendment is associated with microbial cell-state profiles and apparent PAH degradation potential. PAH-contaminated soil containing phenanthrene and pyrene (249.15 mg kg⁻¹ combined) was amended with pine-poplar leaf litter at 10%, 30%, and 70% (w/w), alongside unamended (SS) and leaf-only (AL) treatments, followed by 30-day aerobic aging. After 30-day soil aging, treatment-derived inocula were enriched and evaluated in liquid minimal salt medium containing 40 mg L⁻¹ phenanthrene-pyrene to assess the functional biodegradation potential of microbial communities selected under each amendment condition. Degradation kinetics were modelled using first-order decay based on UV-visible apparent residual phenanthrene-pyrene estimates, with GC-FID used for compound-specific confirmation in selected representative treatments. Imaging flow cytometry coupled with supervised machine learning classified viable and debris-associated cells using morphology- and fluorescence-derived features. Multivariate analyses quantified associations between cellular phenotypes and degradation rate constants. Liquid enrichment cultures derived from aged leaf-litter-amended treatments achieved approximately 60-80% PAH removal within 10 days, whereas unamended soil retained approximately 70% residual PAHs. The 10% amendment exhibited the highest degradation rate (k ≈ 0.153 day⁻¹; t½ ≈ 4.5 days), compared with slower kinetics in unamended soil (k ≈ 0.052 day⁻¹). Imaging cytometry revealed higher viable-cell fractions in amendment-derived treatments and increased membrane-compromised/debris-associated populations under PAH-only conditions. Machine learning classifiers showed strong discrimination of manually annotated cytometric cell-state classes (ROC-AUC ≈ 1.00), indicating clear separability of viable/intact, membrane-compromised, and debris-associated objects within the extracted image-feature space. Morphology-derived cellular metrics showed the strongest association with apparent degradation rates (ρ ≈ 0.89), exceeding viability-based associations (ρ ≈ 0.83). Moderate aged-leaf biomass amendment was associated with higher apparent PAH degradation potential and more stable microbial cytometric profiles, including higher viable-cell fractions and more regular morphology. Imaging-derived cellular phenotypes were associated with treatment-level apparent degradation performance, suggesting that amendment dose influences biodegradation outcomes through effects on microbial cell-state stability. Integrating amendment optimization with cytometry-based physiological monitoring offers a useful framework for improving PAH bioremediation assessment and guiding future field-scale validation.