We investigated the density and distribution of total bacteria, canonical Ammonia Oxidizing Bacteria (AOB) (Nitrosomonas plus Nitrosospira), Ammonia Oxidizing Archaea (AOA), as well as Nitrobacter and Nitrospira in rapid sand filters used for groundwater treatment. To investigate the spatial distribution of these guilds, filter material was sampled at four drinking water treatment plants (DWTPs) in parallel filters of the pre- and after-filtration stages at different locations and depths. The target guilds were quantified by qPCR targeting 16S rRNA and amoA genes. Total bacterial densities (ignoring 16S rRNA gene copy number variation) were high and ranged from 10 to 10 per gram (10 to 10 per m) of filter material. All examined guilds, except AOA, were stratified at only one of the four DWTPs. Densities varied spatially within filter (intra-filter variation) at two of the DWTPs and in parallel filters (inter-filter variation) at one of the DWTPs. Variation analysis revealed random sampling as the most efficient strategy to yield accurate mean density estimates, with collection of at least 7 samples suggested to obtain an acceptable (below half order of magnitude) density precision. Nitrospira was consistently the most dominant guild (5-10% of total community), and was generally up to 4 orders of magnitude more abundant than Nitrobacter and up to 2 orders of magnitude more abundant than canonical AOBs. These results, supplemented with further analysis of the previously reported diversity of Nitrospira in the studied DWTPs based on 16S rRNA and nxrB gene phylogeny (Gülay et al., 2016; Palomo et al., 2016), indicate that the high Nitrospira abundance is due to their comammox (complete ammonia oxidation) physiology. AOA densities were lower than AOB densities, except in the highly stratified filters, where they were of similar abundance. In conclusion, rapid sand filters are microbially dense, with varying degrees of spatial heterogeneity, which requires replicate sampling for a sufficiently precise determination of total microbial community and specific population densities. A consistently high Nitrospira to bacterial and archaeal AOB density ratio suggests that non-canonical pathways for nitrification may dominate the examined RSFs.

A mineral coating develops on the filter grain surface when groundwater is treated via rapid sand filtration in drinking water production. The coating changes the physical and chemical properties of the filter material, but little is known about its effect on the activity, colonization, diversity, and abundance of microbiota. This study reveals that a mineral coating can positively affect the colonization and activity of microbial communities in rapid sand filters. To understand this effect, we investigated the abundance, spatial distribution, colonization, and diversity of all and of nitrifying prokaryotes in filter material with various degrees of mineral coating. We also examined the physical and chemical characteristics of the mineral coating. The amount of mineral coating correlated positively with the internal porosity, the packed bulk density, and the biologically available surface area of the filter material. The volumetric NH 4 ؉ removal rate also increased with the degree of mineral coating. Consistently, bacterial 16S rRNA and amoA abundances positively correlated with increased mineral coating levels. Microbial colonization could be visualized mainly within the outer periphery (60.6 ؎ 35.6 m) of the mineral coating, which had a thickness of up to 600 ؎ 51 m. Environmental scanning electron microscopic (E-SEM) observations suggested an extracellular polymeric substance-rich matrix and submicron-sized bacterial cells. Nitrifier diversity profiles were similar irrespective of the degree of mineral coating, as indicated by pyrosequencing analysis. Overall, our results demonstrate that mineral coating positively affects microbial colonization and activity in rapid sand filters, most likely due to increased volumetric cell abundances facilitated by the large surface area of internal mineral porosity accessible for microbial colonization.

No abstract

Allylthiourea (ATU) and chlorate (ClO) are often used to selectively inhibit nitritation and nitratation. In this work we identified challenges with use of these compounds in inhibitory assays with filter material from a biological rapid sand filter for groundwater treatment. Inhibition was investigated in continuous-flow lab-scale columns, packed with filter material from a full-scale filter and supplied with NH or NO. ATU concentrations of 0.1-0.5 mM interfered with the indophenol blue method for NH quantification leading to underestimation of the measured NH concentration. Interference was stronger at higher ATU levels and resulted in no NH detection at 0.5 mM ATU. ClO at typical concentrations for inhibition assays (1-10 mM) inhibited nitratation by less than 6%, while nitritation was instead inhibited by 91% when NH was supplied. On the other hand, nitratation was inhibited by 67-71% at 10-20 mM ClO when NO was supplied, suggesting significant nitratation inhibition at higher NO concentrations. No chlorite (ClO) was detected in the effluent, and thus we could not confirm that nitritation inhibition was caused by ClO reduction to ClO. In conclusion, ATU and ClO should be used with caution in inhibition assays, because analytical interference and poor selectivity for the targeted process may affect the experimental outcome and compromise result interpretation.

With this study we provide the first in situ evidence of ecologically relevant ammonia oxidation by comammox Nitrospira in a complex microbiome and document an unexpectedly high H13CO3− uptake and growth of proteobacterial and acidobacterial taxa under ammonia selectivity. This finding raises the question of whether comammox Nitrospira is an equally important ammonia oxidizer in other environments.