Influence of hydraulic regimes on bacterial community structure and composition in an experimental drinking water distribution system

Abstract

Microbial biofilms formed on the inner-pipe surfaces of drinking water distribution systems (DWDS) can alter drinking water quality, particularly if they are mechanically detached from the pipe wall to the bulk water, such as due to changes in hydraulic conditions. Results are presented here from applying 454 pyrosequencing of the 16S ribosomal RNA (rRNA) gene to investigate the influence of different hydrological regimes on bacterial community structure and to study the potential mobilisation of material from the pipe walls to the network using a full scale, temperature-controlled experimental pipeline facility accurately representative of live DWDS.

Analysis of pyrosequencing and water physico-chemical data showed that habitat type (water vs. biofilm) and hydraulic conditions influenced bacterial community structure and composition in our experimental DWDS. Bacterial community composition clearly differed between biofilms and bulk water samples. Gammaproteobacteria and Betaproteobacteria were the most abundant phyla in biofilms while Alphaproteobacteria was predominant in bulk water samples. This suggests that bacteria inhabiting biofilms, predominantly species belonging to genera Pseudomonas, Zooglea and Janthinobacterium, have an enhanced ability to express extracellular polymeric substances to adhere to surfaces and to favour co-aggregation between cells than those found in the bulk water. Highest species richness and diversity were detected in 28 days old biofilms with this being accentuated at highly varied flow conditions. Flushing altered the pipe-wall bacterial community structure but did not completely remove bacteria from the pipe walls, particularly under highly varied flow conditions, suggesting that under these conditions more compact biofilms were generated.

This research brings new knowledge regarding the influence of different hydraulic regimes on the composition and structure of bacterial communities within DWDS and the implication that this might have on drinking water quality.

Introduction

Drinking water distribution systems (DWDS) are extreme environments with oligotrophic conditions where a disinfectant residual is commonly maintained. Despite this, microorganisms are able to survive within DWDS, in particular by attaching to the internal surfaces of pipes forming biofilms (Simoes et al., 2007a,b). Microbial biofilms have been conceptually, and under idealised test conditions, associated with various problems in DWDS such as changes in water quality (e.g. discolouration, taste and odour), adsorption and trapping of materials from the bulk water, hosting opportunistic pathogens and promoting the deterioration of pipes (Szewzyk et al., 2000; Beech and Sunner, 2004).

Discolouration is the most common cause of water quality-related customer contacts received by water companies in the UK. Discolouration is known to be associated with the mobilisation of accumulated particles, dominated by iron and manganese but with a significant organic content, from the inner-pipe walls into the bulk water due to increases in shear stress above conditioning values (Husband et al., 2008). Given the association of discolouration with pipe surface accumulations, the occurrence of biofilm on inner-pipe surfaces and the organic content of discolouration material samples it seems logical to speculate that biofilms and biological behaviour may be playing a role in discolouration processes. However, there is limited knowledge concerning the role of microbial biofilms in the process of discolouration and the biologically mediated accumulation of particulates, such as iron and manganese, in DWDS.

There are many different factors that might influence the formation and continual growth of biofilms on pipe surfaces such as flow regime, amount and type of disinfectant, concentration of organic carbon, etc. (LeChevallier et al., 1987). It has been previously suggested that normal (daily) hydraulic conditions within distribution systems are critical in determining the accumulation and subsequent detachment of biofilms (Rickard et al., 2004a; Manuel et al., 2007; Abe et al., 2012). Other research has focused on the study of how hydraulic regimes might influence biofilm formation (Liu et al., 2002; Cloete et al., 2003; Lehtola et al., 2005, 2006). However, these and similar studies generally employed idealised conditions such as bench top reactors, scaled pipeline and biological inoculation which do not realistically reproduce conditions in real DWDS (e.g. Schwartz et al., 1998; Murga et al., 2001; Batte et al., 2003). As a consequence, it is not well understood how conditioning shear stress, and other factors, might affect formation of biofilms and its microbial composition within real DWDS and neither is there substantial information about how differences in biofilm composition might contribute to the process of material mobilization within such systems. To overcome these limitations the experimental work in this study has been carried out in a unique temperature-controlled, full-scale pipeline facility at the University of Sheffield (Fig. 1). This facility can fully recreate the hydraulic and other physical, chemical and biological conditions of real distribution systems. A particular technical advantage of the facility is the inclusion of PWG coupons (Deines et al., 2010). These can be fitted along lengths of the experimental system and enable DNA-based analysis of biofilms from the inner-pipe wall.

Molecular fingerprinting techniques such as DGGE and T-RFLPs have been previously used to evaluate microbial community structure in experimental or simulated water supply systems (Emtiazi et al., 2004; Schwartz et al., 2009; Yu et al., 2010; Sekar et al., 2012), but these techniques can only assess major changes in the composition of dominant microbial species in environmental samples (Forney et al., 2004). Pyrosequencing of the 16S ribosomal RNA (rRNA) is a recently developed molecular tool that provides a more precise characterization of bacterial communities since the diversity revealed within each sample is far larger than that detected by other molecular techniques such as fingerprinting. Recent studies have used pyrosequencing to characterize bacterial communities from impeller retrieved from customer water meters (Hong et al., 2010) and in membrane filtration systems from a drinking water treatment plant (Kwon et al., 2011). To date this technique has not been applied to the analysis of bacterial communities from internal pipe surfaces.