The drinking water exposome

Peter Vikesland

Peter is a Professor of Civil and Environmental Engineering at Virginia Tech. His research interests include nanomaterials in the environment and improved sensors for drinking water. His research on the environmental implications of nanotechnology examines the effects of solution chemistry on the aggregation and dissolution of environmentally relevant nanoparticles. Peter is the co-director of the Virginia Tech Sustainable Nanotechnology Center (VTSuN) and the director of the Virginia Tech Sustainable Nanotechnology Interdisciplinary Graduate Education Program.

Lutgarde Raskin

Professor Raskin is the Altarum/ERIM Russell O'Neal Professor of Engineering at the University of Michigan, USA. She has worked on a variety of biological water and wastewater treatment processes. Her current research focuses on anaerobic microbial processes for energy recovery from waste streams, and microbial processes in drinking water systems, including biological filtration, disinfection, and microbial ecology of distribution systems and premise plumbing. In her research, she uses cutting-edge molecular tools to characterize and optimize water quality process performance.

In recent years it has become apparent that genetics fails to explain the majority of human disease. Instead, evidence now suggests that environmental factors are responsible for up to 90% of all human illnesses. In recognition of this fact, the “human exposome” was recently defined as “the sum of an individual's lifetime environmental exposures to microbes and chemicals in water, food, air, and other environmental compartments”. In this context, the exposome can be considered as the environmental complement to the human genome. In this themed issue of Environmental Science: Water Research & Technology, we asked our contributors to use this emerging and intellectually-engaging construct to consider the chemical and microbial exposures that can occur via consumption or use of drinking water.

Ever since the advent of drinking water disinfection in the early 20th century, there has been significant effort expended to kill or inactivate the organisms present within distribution systems. While the focus of this effort has been on elimination of pathogenic organisms that cause waterborne diseases, such as typhoid and cholera, an unintended consequence of disinfectant addition has been the wholesale manipulation of the chemistry and the microbiology of the distribution system. Although the chemical implications of such manipulations in terms of disinfection by-product (DBP) formation and distribution system corrosion have been studied extensively for many years, the microbial implications have historically been much less understood. The advent of new tools for the characterization of complex microbial communities, however, has revolutionized our understanding of the intricate ecosystems present within drinking water distribution systems. Next-generation DNA sequencing (NGS) technologies continue to reveal multifaceted relationships between the physical components of the drinking water distribution system, the chemistry of the water flowing through it, and the microorganisms that reside both within the water as well as adhere to pipe walls. The composition of this drinking water microbiome is diverse, complex, and highly dynamic and it is now widely recognized that the microbes in the distribution system play both positive and negative roles in terms of water quality and distribution system maintenance. As illustrated by the articles in this compilation there is significant interplay between the chemical and microbial components of the drinking water exposome, with potential important implications for human health.

Addition of a persistent disinfectant residual, such as free chlorine or monochloramine, to treated water prior to distribution is one approach to control the microbial quality of drinking water. An alternative approach, and one that has been practiced for many years throughout Europe, is to limit the concentration of biodegradable organics (i.e., assimilable organic carbon or AOC) within the distribution system and thus minimize the potential for downstream microbial (re)growth. Such an approach is reliant upon techniques that can quantify microbial concentrations both in treated and distributed water. Van der Wielen et al. (DOI: 10.1039/c6ew00007j) evaluate the potential utility of novel microbiological parameters as alternatives to the currently required heterotrophic plate counts (HPCs) and Aeromonas plate counts for microbial growth monitoring within 28 different disinfectant-free distribution systems in the Netherlands. They show that, while adenosine triphosphate (ATP) concentration and flow cytometry enabled cell counts can be rapidly acquired, they are not good indicators for microbial (re)growth distribution system monitoring. In contrast, the established HPC and Aeromonas methods provided a much better indication of (re)growth.

Within distribution systems, microbes are found both as planktonic cells suspended in the bulk liquid or as sessile cells entrained within the biofilms that adhere to pipe walls and other solid surfaces. These biofilms are highly complex, localized ecosystems that consist of a taxonomically diverse array of bacterial and archaeal cells, viruses, and eukaryotic organisms embedded within a matrix of extracellular polymeric substances (EPS) that provides mechanical and chemical stability. Due to their capacity to accumulate organic and inorganic substances, biofilms exhibit high cell densities and their presence and composition often dictate water quality in the distribution system. To date, however, there are no guidelines regarding biofilm monitoring. In their Critical Review, Fish, Osborn, and Boxall (DOI: 10.1039/c6ew00039h) summarize the state of knowledge of drinking water distribution system biofilms and discuss different approaches for their analysis. A key point the authors make is the challenge of attempting to replicate real-world distribution systems at the laboratory scale due to differences in surface area-to-volume, pipe material, flow rate, and experiment duration. While a great deal has been learned using flow reactors and pipe rigs, there is much that remains unknown at full scale. In addition, they also emphasize that while much has been learned about the bacterial component of the drinking water microbiome (i.e., the bacteriome) considerably less is known about fungi, protozoa, archaea, and viruses. The ecological interactions among all these microbial populations within biofilms play important roles in dictating the quality of the water within a distribution system network.

Biofilms are increasingly recognized as key mediators of water quality within distribution systems; however, current water quality regulations generally revolve around HPC or other culture-based measurements obtained from planktonic cells in bulk water samples. It is for this reason, as well as the acknowledged complexities of sampling real-world distribution system biofilms, that most studies at lab-scale and in the field focus on the microbiology of bulk water. In his Emerging Investigator contribution, Dr Ameet Pinto and colleagues (DOI: 10.1039/c6ew00030d) discuss the results of a meta analysis of the 14 readily accessible and currently publicly available NGS datasets that describe the microbiome of bulk drinking water. Across their 14 datasets comprising 142 sampling locations for full-scale systems, both with and without a disinfectant residual, they show that bacteria consistently constitute a majority of the bulk water microbial community, with archaea being detected only at very low levels. Proteobacteria were the dominant bacterial phylum under all conditions with Alpha- and Betaproteobacteria dominant (>80%). Actinobacteria constituted the second most abundant phylum in systems with a residual disinfectant, while Acidobacteria was second most abundant under conditions without a disinfectant. This difference, coupled with the observation that disinfectant-free samples exhibited greater bacterial diversity than samples containing a disinfectant, may be the result of the selection pressures exerted by disinfectants on the bulk water microbial community.

A key take home message of the study by Pinto et al. (DOI: 10.1039/c6ew00030d) is the observation that cross-comparison of many NGS datasets remains challenging due to differences in sample collection, DNA extraction, PCR amplification, and many other factors. Furthermore, due to the complexities of full-scale distribution system sampling and system-to-system variability, highly controlled laboratory studies remain important. In their contribution, Gomez-Alvarez et al. (DOI: 10.1039/c6ew00053c) used a semi-closed pipe-loop system to systematically evaluate how perturbations in the chemistry of drinking water affect both the bulk water and biofilm microbiome of the distribution system. Using a metagenomic approach they show that application of a free chlorine burn in response to a nitrification episode results in temporal changes in the distribution system microbiome. These changes, however, were short-lived and the core microbiome of the bulk water was re-established within a short period of time. Similar to what Pinto et al. (DOI: 10.1039/c6ew00030d) observed, Gomez-Alvarez et al. found that the core microbiome consisted primarily of Proteobacteria and Actinobacteria. Importantly, they observe that this core microbiome is complemented by conditionally rare taxa (CRT) that are generally low in concentration, but increase in number under favorable conditions (i.e., perturbations). The CRT constitutes a reservoir of microbes with unique functionality (i.e., nitrification, antibiotic resistance) that increase in relative abundance in response to disturbances in water quality.

Nitrification, or the conversion of ammonium (NH₄⁺) into nitrite (NO₂⁻) or nitrate (NO₃⁻), is a major challenge for drinking water distribution systems that utilize monochloramine as a secondary disinfectant. Historically, this process has been attributed to the collective activity of ammonia-oxidizing bacteria (AOB) and nitrite-oxidizing bacteria (NOB) as a two-step process. However, recent research suggests that the process is considerably more complex, with both complete ammonia-oxidizing (comammox) Nitrospira bacteria and ammonia-oxidizing archaea (AOA) potentially playing important roles within distribution systems. Santillana et al. (DOI: 10.1039/c5ew00273g) focus on the latter group and utilized a simulated building plumbing system to probe archaeal ammonia oxidation. In their system, entailing copper or PVC plumbing reactors, they observe that nitrification can be attributed to the activity of an archaeal–bacterial consortium that consists of AOA and NOB. While their experimental design could not fully exclude the potential role of comammox bacteria, they note that the Nitrospira organisms that were detected by NGS were not closely related to those known to have the capacity for complete nitrification.

Exposure to the microbes and chemicals present within a drinking water distribution system generally occurs at the tap and there has been a recent surge of interest in water quality within building plumbing (often referred to as premise plumbing). Water can stagnate within building plumbing for periods that are generally much longer than in distribution mains and as such the interactions between treated water and building plumbing materials can have disproportionate effects on the quality of water at the tap. Proctor et al. (DOI: 10.1039/c6ew00016a) examined how the flexible polymeric plumbing materials increasingly being used within the last few meters of a building plumbing system affect biofilm characteristics. They showed that organics leach out of flexible pipes of variable polymeric materials at different rates and generally observed that higher leaching leads to enhanced microbial growth. One outlier, which exhibited the greatest level of leaching but the lowest level of microbial growth, was explained as a result of an inorganic silver-ion coating. This result illustrates the complex interplay between the organic–inorganic chemical composition of pipes and adhered microbial communities. This study further observed that biological growth potential does not correlate with TOC leaching. Such a result makes sense given that different organic molecules are being released and each can be expected to have a varying potential for biological assimilation.

While detection and quantification of chemicals and microbes within drinking water is an important aspect of exposome science, it is equally important that actual exposure routes be delineated and quantified. In their Critical Review, Hamilton and Haas (DOI: 10.1039/c6ew00023a) discuss the development of a quantitative microbial risk assessment (QMRA) approach to assess the potential for exposure to Legionella in engineered water systems. QMRA incorporates hazard identification, exposure assessment, dose response assessment, and risk characterization. The authors note that to date there has been no standardized approach to develop a Legionella QMRA and they identify a number of research gaps that affect the QMRAs that have been reported in the literature. Among a number of issues, of great concern for Legionella and any other biological contaminants is our historical reliance on culture-based methods for microbial quantitation. For Legionella spp., as well as many other microbes, there is abundant evidence that molecular assays (e.g., qPCR) are more sensitive than culture-based approaches and accordingly risk assessments based upon cell culture will systematically underestimate risks calculated by QMRA.

As illustrated by the lead contamination crisis in Flint, Michigan and widespread arsenic contamination of drinking water throughout the world, the contamination of drinking water by heavy metals and metalloids is a global problem of continual concern. One challenge that the water community faces, however, is the capacity to rapidly and readily detect these analytes. Acknowledging this need, Das and Sarkar (DOI: 10.1039/c5ew00276a) developed a novel tool for detection of arsenate (As^V) in drinking water. Their field-deployable dipstick turns blue when exposed to As^V at levels above 10 μg L⁻¹. This result was achieved because of the formation of a blue colored antimonyl–arseno–molybdate complex that forms when As^V binds to molybdenum under reducing conditions. By loading a polymeric hydrogel with this colorimetric reagent, Das and Sarkar are able to achieve field deployability. Impressively, this sensor exhibits similar detection capacity as atomic absorption spectroscopy (AAS) for As^V in field samples from India, but without the need for expensive instrumentation.

We previously noted that disinfection with free chlorine and monochloramine is often employed and results in disinfectant-specific microbiomes. In the contribution from Emerging Investigator Dr Christina Remucal (DOI: 10.1039/c6ew00029k), she discusses an alternative disinfection strategy that relies upon the photolysis of free chlorine (hypochlorous acid, HOCl, and hypochlorite, OCl⁻) for the production of the highly reactive oxidants hydroxyl radical (˙OH), chlorine radical (Cl˙), and ozone (O₃). As a disinfection practice, chlorination is hampered by the formation of disinfection by-products (DBPs) and the inability of free chlorine to eliminate recalcitrant chemical and microbial contaminants. For this reason, many localities are looking to either replace free chlorine as their disinfectant of choice or to augment it by application of sequential disinfection. As discussed in the article by Remucal and Manley, one method to do so is to illuminate free chlorine with UV light and thus convert the disinfection system into an advanced oxidation process (AOP). In this manner, a series of highly reactive oxidants are produced that in many cases exhibit enhanced potential to inactivate or transform recalcitrant pollutants.

Garner et al. (DOI: 10.1039/c6ew00031b), in the final Critical Review in this themed issue, propose that the exposome concept provides an ideal framework for the development of risk management strategies for water reuse. Because of the increasing demands placed on a globally finite volume of potable water, there is a rapidly advancing effort to reclaim and reuse treated wastewater. [For additional information about water reuse, the reader is referred to a previous themed issue of Environmental Science: Water Research & Technology devoted exclusively to this topic. The lead editorial summarizing the themed issue is available at DOI: 10.1039/c5ew90021b]. A major challenge with water reuse, particularly direct potable reuse in which treated wastewater is used to augment a drinking water supply, is the current lack of regulations dictating appropriate practices. Using the exposome as a scaffold, Garner et al. emphasize that water reuse raises new, as yet unanswered questions about potential exposures to wastewater-derived contaminants that historically were of lesser or no concern to the drinking water community. Microbial contaminants are of particular importance due to their potential to (re)grow within the distribution system. As such, monitoring opportunistic pathogens (OPs), antibiotic resistant bacteria (ARB) and antibiotic resistance genes (ARGs), viruses, and free-living amoebae in reused waters is important. Further, while potable water standards are commonly used as the benchmark for water reuse, other exposure routes, such as inhalation and skin colonization, are likely especially important in water reuse scenarios. Garner and colleagues further note that reused wastewater exhibits important differences in the quality and quantity of organic matter that affect its potential to enhance (re)growth as well as a utility's capacity to maintain a disinfectant residual.

We conclude by noting that the drinking water exposome is a nascent concept that explicitly attempts to connect chemical and microbial drinking water-based exposures in a holistic manner. As the papers outlined herein collectively illustrate, we must think of drinking water distribution systems as highly complex engineered ecosystems, whose characteristics are affected and defined by operational and environmental parameters in ways we currently do not fully understand. As analytical techniques continue to improve, we can expect that additional types of organisms will be found and new ‘emerging’ chemical contaminants will be detected. The exposome concept provides a context in which to compare drinking water based exposures to those via other environmental routes.

---