Methanotrophic contribution to biodegradation of phenoxy acids in cultures enriched from a groundwater-fed rapid sand filter.

Drinking water supply is in many parts of the world based on groundwater. Groundwater often contains methane, which can be oxidized by methanotrophs upon aeration. Sand from rapid sand filters fed with methane-rich groundwater can remove some pesticides (Hedegaard and Albrechtsen in Water Res 48:71-81, 2014). We enriched methanotrophs from filter sand and investigated whether they could drive the degradation of various pesticides. To enrich for methanotrophs, we designed and operated four laboratory-scale, continuously methane-fed column reactors, inoculated with filter sand and one control column fed with tap water. When enrichments were obtained, methane was continuously supplied to three reactors, while the fourth was starved for methane for 1 week, and the reactors were spiked with ten pesticides at groundwater-relevant concentrations (2.1-6.6 µg/L). Removal for most pesticides was not detected at the investigated contact time (1.37 min). However, the degradation of phenoxy acids was observed in the methanotrophic column reactor starved for methane, while it was not detected in the control column indicating the importance of methanotrophs. Phenoxy acid removal, using dichlorprop as a model compound, was further investigated in batch experiments with methanotrophic biomass collected from the enrichment reactors. Phenoxy acid removal (expressed per gram of matrix sand) was substantially improved in the methanotrophic enrichment compared to parent filter sand. The presence of methane did not clearly impact dichlorprop removal but did impact mineralization. We suggest that other heterotrophs are responsible for the first step in dichlorprop degradation, while the subsequent steps including ring-hydroxylation are driven by methanotrophs.

Removal of phytotoxins in filter sand used for drinking water treatment.

Phytotoxins - toxins produced by plants - are contaminants with the potential to impair drinking water quality. They encompass a large group of toxic, partially persistent compounds that have been detected in seepage waters and in shallow wells used for drinking water production. If phytotoxins enter wells used for drinking water production, it is essential to know if the drinking water treatment processes will remove them from the water phase. However, it is currently unknown whether phytotoxins remain stable during traditional groundwater treatment using sand filters as the main treatment process. The objective of this study is to investigate removal potential of phytotoxins in biological sand filters and to asses if the removal potential is similar at different waterworks. Microcosms were set up with filter sand and drinking water collected at different groundwater-based waterworks. To be able to monitor phytotoxin removal ptaquiloside, caudatoside, gramine, sparteine, jacobine N-oxide, senecionine N-oxide and caffeine were applied at initial concentrations of 300 µg L-1, which is approx. two orders of magnitude higher than currently detected in environment, but expected to cover extreme environmental conditions. Removal was monitored over a period of 14 days. Despite the high initial concentration, all filter sands removed ptaquiloside and caudatoside completely from the water phase and at waterworks where pellet softening was implemented (pH 8.4) prior to rapid sand filtration, complete removal occurred within the first 30 min. All filter sands removed gramine and sparteine, primarily by a biological process, while jacobine N-oxide, senecionine N-oxide and caffeine were recalcitrant in the filter sands. During degradation of ptaquiloside and caudatoside we observed formation and subsequent removal of degradation products pterosin B and A. Filter sands with the highest removal potential were characterised by high contents of deposited iron and manganese oxides and hence large specific surface areas. Difference between bacterial communities investigated by 16S rRNA gene analyses did not explain different removal in the filter sands. All five investigated filter sands showed similar degradation patterns regardless of water chemistry and waterworks of origin. In drinking water treatment systems biological sand filters might therefore remove phytotoxin contaminants such as ptaquiloside, caudatoside, gramine, sparteine, while for other compounds e.g. jacobine N-oxide, senecionine N-oxide further investigations involving more advanced treatment options are needed.

Evidence of co-metabolic bentazone transformation by methanotrophic enrichment from a groundwater-fed rapid sand filter.

The herbicide bentazone is recalcitrant in aquifers and is therefore frequently detected in wells used for drinking water production. However, bentazone degradation has been observed in filter sand from a rapid sand filter at a waterworks with methane-rich groundwater. Here, the association between methane oxidation and removal of bentazone was investigated with a methanotrophic enrichment culture derived from methane-fed column reactors inoculated with that filter sand. Several independent lines of evidence obtained from microcosm experiments with the methanotrophic enrichment culture, tap water and bentazone at concentrations below 2 mg/L showed methanotrophic co-metabolic bentazone transformation: The culture removed 53% of the bentazone in 21 days in presence of 5 mg/L of methane, while only 31% was removed in absence of methane. Addition of acetylene inhibited methane oxidation and stopped bentazone removal. The presence of bentazone partly inhibited methane oxidation since the methane consumption rate was significantly lower at high (1 mg/L) than at low (1 µg/L) bentazone concentrations. The transformation yield of methane relative to bentazone normalized by their concentration ratio ranged from 58 to 158, well within the range for methanotrophic co-metabolic degradation of trace contaminants calculated from the literature, with normalized substrate preferences varying from 3 to 400. High-resolution mass spectrometry revealed formation of the transformation products (TPs) 6-OH, 8-OH, isopropyl-OH and di-OH-bentazone, with higher abundances of all TPs in the presence of methane. Overall, we found a suite of evidence all showing that bentazone was co-metabolically transformed to hydroxy-bentazone by a methanotrophic culture enriched from a rapid sand filter at a waterworks.

Microbial pesticide removal in rapid sand filters for drinking water treatment--potential and kinetics.

Filter sand samples, taken from aerobic rapid sand filters used for treating groundwater at three Danish waterworks, were investigated for their pesticide removal potential and to assess the kinetics of the removal process. Microcosms were set up with filter sand, treated water, and the pesticides or metabolites mecoprop (MCPP), bentazone, glyphosate and p-nitrophenol were applied in initial concentrations of 0.03-2.4 µg/L. In all the investigated waterworks the concentration of pesticides in the water decreased - MCPP decreased to 42-85%, bentazone to 15-35%, glyphosate to 7-14% and p-nitrophenol 1-3% - from the initial concentration over a period of 6-13 days. Mineralisation of three out of four investigated pesticides was observed at Sjælsø waterworks Plant II - up to 43% of the initial glyphosate was mineralised within six days. At Sjælsø waterworks Plant II the removal kinetics of bentazone revealed that less than 30 min was needed to remove 50% of the bentazone at all the tested initial concentrations (0.1-2.4 µg/L). Increased oxygen availability led to greater and faster removal of bentazone in the microcosms. After 1 h, bentazone removal (an initial bentazone concentration of 0.1 µg/L) increased from 0.21%/g filter sand to 0.75%/g filter sand, when oxygen availability was increased from 0.28 mg O2/g filter sand to 1.09 mg O2/g filter sand. Bentazone was initially cleaved in the removal process. A metabolite, which contained the carbonyl group, was removed rapidly from the water phase and slowly mineralised after 24 h, while a metabolite which contained the benzene-ring was still present in the water phase. However, the microbial removal of this metabolite was initiated over seven days.

Mecoprop (MCPP) removal in full-scale rapid sand filters at a groundwater-based waterworks.

Contamination by the herbicide mecoprop (MCPP) was detected in groundwater abstraction wells at Kerteminde Waterworks in concentrations up to 0.08µg/L. MCPP was removed to below detection limit in a simple treatment line where anaerobic groundwater was aerated and subsequently filtered by primary and secondary rapid sand filters. Water quality parameters were measured throughout the waterworks, and they behaved as designed for. MCPP was removed in secondary rapid sand filters--removal was the greatest in the sand filters in the filter line with the highest contact time (63 min). In these secondary sand filters, MCPP concentration decreased from 0.037 µg/L to below the detection limit of 0.01 µg/L. MCPP was removed continuously at different filter depths (0.80 m). Additionally, biodegradation, mineralisation and adsorption were investigated in the laboratory in order to elucidate removal mechanisms in the full-scale system. Therefore, microcosms were set up with filter sand, water and (14)C-labelled MCPP at an initial concentration of 0.2 µg/L. After 24 h, 79-86% of the initial concentration of MCPP was removed. Sorption removed 11-15%, while the remaining part was removed by microbial processes, leading to a complete mineralisation of 13-18%. Microbial removal in the filter sand was similar at different depths of the rapid sand filter, while the amount of MCPP which adsorbed to the filter sand after 48 h decreased with depth from 21% of the initial MCPP in the top layer to 7% in the bottom layer. It was concluded that MCPP was removed in secondary rapid sand filters at Kerteminde Waterworks, to which both adsorption and microbial degradation contributed.