Aminobacter sp. MSH1 immobilized in an alginate matrix in porous stones was tested in a pilot system as an alternative inoculation strategy to the use of free suspended cells for biological removal of micropollutant concentrations of 2,6-dichlorobenzamide (BAM) in drinking water treatment plants (DWTPs). BAM removal rates and MSH1 cell numbers were recorded during operation and assessed with specific BAM degradation rates obtained in lab conditions using either freshly grown cells or starved cells to explain reactor performance. Both reactors inoculated with either suspended or immobilized cells showed immediate BAM removal under the threshold of 0.1 μg/L, but the duration of sufficient BAM removal was 2-fold (44 days) longer for immobilized cells. The longer sufficient BAM removal in case of immobilized cells compared to suspended cells was mainly explained by a lower initial loss of MSH1 cells at operational start due to volume replacement and shear. Overall loss of activity in the reactors though was due to starvation, and final removal rates did not differ between reactors inoculated with immobilized and suspended cells. Management of assimilable organic carbon, in addition to cell immobilization, appears crucial for guaranteeing long-term BAM degradation activity of MSH1 in DWTP units.

Aminobacter sp. strain MSH1 grows on and mineralizes the groundwater micropollutant 2,6-dichlorobenzamide (BAM) and is of interest for BAM removal in drinking water treatment plants (DWTPs). The BAM-catabolic genes in MSH1 are located on plasmid pBAM1, carrying bbdA, which encodes the conversion of BAM to 2,6-dichlorobenzoic acid (2,6-DCBA) (BbdA ϩ phenotype), and plasmid pBAM2, carrying gene clusters encoding the conversion of 2,6-DCBA to tricarboxylic acid (TCA) cycle intermediates (Dcba ϩ phenotype). There are indications that MSH1 easily loses its BAM-catabolic phenotype. We obtained evidence that MSH1 rapidly develops a population that lacks the ability to mineralize BAM when grown on nonselective (R2B medium) and semiselective (R2B medium with BAM) media. Lack of mineralization was explained by loss of the Dcba ϩ phenotype and corresponding genes. The ecological significance of this instability for the use of MSH1 for BAM removal in the oligotrophic environment of DWTPs was explored in lab and pilot systems. A higher incidence of BbdA ϩ Dcba Ϫ MSH1 cells was also observed when MSH1 was grown as a biofilm in flow chambers under C and N starvation conditions due to growth on nonselective residual assimilable organic carbon. Similar observations were made in experiments with a pilot sand filter reactor bioaugmented with MSH1. BAM conversion to 2,6-DCBA was not affected by loss of the DCBA-catabolic genes. Our results show that MSH1 is prone to BAM-catabolic instability under the conditions occurring in a DWTP. While conversion of BAM to 2,6-DCBA remains unaffected, BAM mineralization activity is at risk, and monitoring of metabolites is warranted.IMPORTANCE Bioaugmentation of dedicated biofiltration units with bacterial strains that grow on and mineralize micropollutants was suggested as an alternative for treating micropollutant-contaminated water in drinking water treatment plants (DWTPs). Organic-pollutant-catabolic genes in bacteria are often easily lost, especially under nonselective conditions, which affects the bioaugmentation success. In this study, we provide evidence that Aminobacter sp. strain MSH1, which uses the common groundwater micropollutant 2,6-dichlorobenzamide (BAM) as a C source, shows a high frequency of loss of its BAM-mineralizing phenotype due to the loss of genes that convert 2,6-DCBA to Krebs cycle intermediates when nonselective conditions occur. Moreover, we show that catabolic-gene loss also occurs in the oligotrophic environment of DWTPs, where growth of MSH1 depends mainly on the high fluxes of low concentrations of assimilable organic carbon, and hence show the ecological relevance of catabolic instability for using strain MSH1 for BAM removal in DWTPs.