Abstract Stream-bank erosion threatens freshwater ecosystems globally, yet traditional stabilization methods using cement, synthetic polymers, and riprap cause significant environmental damage through habitat degradation and greenhouse gas emissions. Biopolymers offer a sustainable alternative, but their ecotoxicological effects on aquatic organisms remain unknown. We evaluated the acute toxicity of exopolysaccharides (EPS) produced by Rhizobium tropici ATCC® 49672 on Procambarus clarkii, a burrowing crayfish that directly interacts with stream bank soils. A 96-hour static renewal test following ASTM guidelines exposed 69 crayfish individuals to EPS concentrations ranging from 0% to 33.33% (0–3333 ppm). Dose-response relationships were analyzed using logistic regression, with model validation performed through probit regression and nonparametric interpolation. The estimated LC50 was 8.78% (878 ppm; 95% CI: 780–975 ppm), classifying EPS as "practically nontoxic" according to EPA toxicity categories. Clear concentration-dependent mortality was observed, with no mortality at ≤ 5% (500 ppm) and 100% mortality at ≥ 25% (2,500 ppm). Correlation analyses confirmed that specimen weight did not influence toxicity response (P = 0.513), validating the use of mixed-size populations. The substantial safety margin between typical field application concentrations ( 1%, 100 ppm) and the LC50 value suggests minimal risk to P. clarkii and potentially similar tolerant Procambarus species at environmentally relevant exposures. These findings nevertheless cannot be extrapolated to the broader aquatic community without additional testing of more sensitive taxa, including fishes, aquatic insects, mollusks, amphibians, and flow-adapted crayfish species. This preliminary assessment provides foundational ecotoxicological data supporting the potential integration of microbially-produced biopolymers into sustainable watershed management strategies, while highlighting the need for comprehensive multispecies toxicity testing before widespread implementation.
Morgan et al. (Wed,) studied this question.