Legionella: A Promising Supplementary Indicator of Microbial Drinking Water Quality in Municipal Engineered Water Systems

Opportunistic pathogens (OPs) are natural inhabitants and the predominant disease causative biotic agents in municipal engineered water systems (EWSs). In EWSs, OPs occur at high frequencies and concentrations, cause drinking-water-related disease outbreaks, and are a major factor threatening public health. Therefore, the prevalence of OPs in EWSs represents microbial drinking water quality. Closely or routinely monitoring the dynamics of OPs in municipal EWSs is thus critical to ensuring drinking water quality and protecting public health. Monitoring the dynamics of conventional (fecal) indicators (e.g., total coliforms, fecal coliforms, and Escherichia coli) is the customary or even exclusive means of assessing microbial drinking water quality. However, those indicators infer only fecal contamination due to treatment (e.g., disinfection within water utilities) failure and EWS infrastructure issues (e.g., water main breaks and infiltration), whereas OPs are not contaminants in drinking water. In addition, those indicators appear in EWSs at low concentrations (often absent in well-maintained EWSs) and are uncorrelated with OPs. For instance, conventional indicators decay, while OPs regrow with increasing hydraulic residence time. As a result, conventional indicators are poor indicators of OPs (the major aspect of microbial drinking water quality) in EWSs. An additional or supplementary indicator that can well infer the prevalence of OPs in EWSs is highly needed. This systematic review argues that Legionella as a dominant OP-containing genus and natural inhabitant in EWSs is a promising candidate for such a supplementary indicator. Through comprehensively comparing the behavior (i.e., occurrence, growth and regrowth, spatiotemporal variations in concentrations, resistance to disinfectant residuals, and responses to physicochemical water quality parameters) of major OPs (e.g., Legionella especially L. pneumophila, Mycobacterium, and Pseudomonas especially P. aeruginosa), this review proves that Legionella is a promising supplementary indicator for the prevalence of OPs in EWSs while other OPs lack this indication feature. Legionella as a dominant natural inhabitant in EWSs occurs frequently, has a high concentration, and correlates with more microbial and physicochemical water quality parameters than other common OPs. Legionella and OPs in EWSs share multiple key features such as high disinfectant resistance, biofilm formation, proliferation within amoebae, and significant spatiotemporal variations in concentrations. Therefore, the presence and concentration of Legionella well indicate the presence and concentrations of OPs (especially L. pneumophila) and microbial drinking water quality in EWSs. In addition, Legionella concentration indicates the efficacies of disinfectant residuals in EWSs. Furthermore, with the development of modern Legionella quantification methods (especially quantitative polymerase chain reactions), monitoring Legionella in ESWs is becoming easier, more affordable, and less labor-intensive. Those features make Legionella a proper supplementary indicator for microbial drinking water quality (especially the prevalence of OPs) in EWSs. Water authorities may use Legionella and conventional indicators in combination to more comprehensively assess microbial drinking water quality in municipal EWSs. Future work should further explore the indication role of Legionella in EWSs and propose drinking water Legionella concentration limits that indicate serious public health effects and require enhanced treatment (e.g., booster disinfection).


INTRODUCTION
Microbial drinking water quality (i.e., the presence and concentrations of microbes) in municipal engineered water systems (EWSs, including drinking water distribution systems or DWDSs and building premise plumbing systems or PSSs) strongly affects public health (Ashbolt, 2015a;Zhang and Lu, 2021). Pathogenetic microbes in EWSs may infect end consumers and cause life-threatening diseases or disease outbreaks. Microbes enter municipal EWSs with finished water from treatment plants. Drinking water treatment facilities effectively remove and inactivate microbes, especially pathogens, with disinfection (e.g., chlorination, chloramination, and ozonation) and other treatment barriers (e.g., filtration) before discharging finished water to DWDSs (McGuire, 2006;Okpara et al., 2011;Zhang and Liu, 2019). However, residual microbes in finished water could be still active or viable [or viable-but-non-culturable (VBNC)] and enter DWDSs (Haas, 1999;Schoenen, 2002;Liu et al., 2009;Marciano-Cabral et al., 2010;Falkinham et al., 2015a;Lin et al., 2017;Chen et al., 2018;Guo et al., 2021). Microbes also invade EWSs because of deficiencies in design and operations of the systems, installation and maintenance of water mains and distal pipes, and intrusion during cross-connections (actual or potential connections between potable water supplies and non-potable sources), back-pressure (occurs when the supply pressure is lower than the local pipe pressure; a type of backflow), and back-siphonage (due to local negative pressure transients or a vacuum effect; another type of Zhang  backflow) (Herrick, 1997;Payment, 1999;LeChevallier et al., 2003;Lee and Schwab, 2005;Makris et al., 2014;Beer et al., 2015;Gerba and Pepper, 2015;Lu et al., 2016;Leslie et al., 2021).
Certain microbes after entering or invading municipal EWSs survive, (re)grow, and become natural inhabitants of drinking water. Most natural inhabitants in EWSs are harmless to humans. However, an important group of natural inhabitants in EWSs, water-based opportunistic pathogens (OPs), is the leading cause of drinking-water-based disease outbreaks, significantly threatens public health, and causes serious social-economic losses (Wang et al., 2013;Whiley et al., 2014b;Ashbolt, 2015b;Beer et al., 2015;Falkinham et al., 2015b;Benedict et al., 2017;Liu et al., 2017;El-Chakhtoura et al., 2018;Cullom et al., 2020). For instance, water-based nontuberculous mycobacteria (NTM) infections annually cause over 16,000 hospitalizations and a total hospitalization cost of over 426 million United States (US) dollars in the US (Collier et al., 2012). In addition, OPs survive and (re)grow in EWSs (Hwang et al., 2012;El-Chakhtoura et al., 2018;Douterelo et al., 2019;Zhang and Liu, 2019) even in the presence of disinfectant residuals (Haas, 1999;LeChevallier, 1999;van der Kooij, 2003;Zhang et al., 2021a) and under high-shear turbulent flow conditions (Lehtola et al., 2007). Furthermore, conventional water quality indicators (e.g., total and fecal coliforms, thermotolerant coliforms, Escherichia coli, enterococci, and fecal streptococci) and fecal contaminants (i.e., enteric and primary pathogens) that commonly exist in (contaminated) natural waterbodies occur in EWSs at low frequencies and concentrations and are often absent from well-maintained systems (Falkinham, 2020;Zhang and Lu, 2021). Therefore, the presence and concentrations of OPs are the major aspect of microbial water quality in municipal EWSs. Because OPs in drinking water frequently cause disease outbreaks, monitoring the dynamics of OPs in municipal EWSs is an important and urgent task for the water industry and governmental agencies to deeply understand microbial drinking water quality and to better protect public health.
Examining all or dominant (pathogenic) microbes to fully understand microbial drinking water quality in EWSs is technically intense, expensive, time consuming, and often impractical and unnecessary (Gerba, 2015). Therefore, water utilities and public health agencies in the US and many other countries assess microbial drinking water quality mainly or even exclusively through quantifying conventional indicator microbes (i.e., fecal indicators). Fecal indicators infer fecal contamination from warm-blooded animals and the presence of enteric/primary pathogens due to treatment system failure, water pipe breaks, cross-connections, and backflow (Cohen and Shuval, 1973;Edberg et al., 2000;Ashbolt et al., 2001;Lee and Schwab, 2005;Tallon et al., 2005;Gerba, 2015;Lu et al., 2016). Examining fecal indicators using traditional culture-based methods (e.g., the Colilert ™ Test for both total coliforms and E. coli) is straightforward and cost effective. However, as discussed above, conventional fecal indicators and fecal contaminants (i.e., enteric pathogens) have low concentrations and are often absent in EWSs. Fecal contaminants and indicators are thus not the major aspect of microbial drinking water quality, especially in (developed) countries where secondary disinfection is mandatory (e.g., the US) and where municipal EWSs are well-maintained and clean (e.g., the Netherlands) (Smeets et al., 2009;Zhang and Liu, 2019;Zhang and Lu, 2021). In addition, fecal indicators behave distinctively from OPs in EWSs (Falkinham et al., 2015b;Lu et al., 2016;Wang et al., 2017;Perrin et al., 2019a;Zhang and Lu, 2021). For instance, the concentrations of OPs generally increase, while fecal indicators decay with increasing hydraulic residence time (HRT) of distributed water (Falkinham, 2020;Zhang and Lu, 2021). OPs have high resistance to disinfectant residuals (and thus have high concentrations), while fecal indicators are susceptible to disinfectant residuals (and thus have low concentrations or are even absent) in EWSs . Therefore, conventional fecal indicators do not indicate OPs (the major aspect of microbial drinking water quality) in EWSs (Lehtola et al., 2007;LeChevallier, 2019), and a supplementary or additional indicator that specifically infers the prevalence of OPs is highly needed. To date, a microorganism indicating the presence and concentrations of OPs in municipal EWSs has not been fully developed, evaluated, and validated. To the best of our knowledge, most countries do not require a routine minoring of OPs in municipal EWSs using an indicator microorganism.
This comprehensive literature review aims to purpose an appropriate indicator microorganism that can infer the presence and concentrations of water-based OPs (the major aspect of microbial drinking water quality) in municipal EWSs. We first explain why conventional fecal indicators fail to indicate OPs or microbial drinking water quality in EWSs and why an indicator of OPs is highly needed. We then propose novel critical criteria for selecting an appropriate indicator of the presence and concentrations of OPs in drinking water. We argue that selecting such a supplementary indicator from common OPs in EWSs is appropriate. We finally comprehensively compare the behavior (i.e., occurrence, growth and regrowth, spatiotemporal variations in concentrations, resistance to disinfectant residuals, and responses to physicochemical water quality parameters) of major OPs (i.e., Legionella especially L. pneumophila, Mycobacterium, and Pseudomonas especially P. aeruginosa). The comparison proves that Legionella is a promising supplementary indicator for the presence and concentrations of OPs in EWSs while other OPs do not have this indication feature. This comprehensive literature review for the first time proposes and confirms an appropriate indicator microorganism (Legionella) that well infers water-based OPs and reflects microbial drinking water quality in municipal EWSs, adding new insights into the current literature. This review also systematically summarizes the behavior of Legionella and other common OPs in municipal drinking water, whereas such information is meaningful to developing more effective OP control strategies and promoting public health.

MUNICIPAL ENGINEERED WATER SYSTEMS
Examining conventional fecal indicators is the most common practice worldwide evaluating drinking water quality in municipal EWSs. However, fecal indicators indicate only the presence of enteric pathogens and/or fecal contaminants (mainly due to treatment failure in water utilities) and cannot infer the presence and concentrations of water-based OPs (the major aspect of microbial drinking water quality). Fecal indicators fail to indicate OPs because of three reasons. First, fecal indicators and OPs in municipal EWSs have distinct microbial community compositions. Fecal indicators mainly include E. coli, Bacteroides (more common than E. coli in the gastrointestinal tract of humans; a good indicator for and V. vermiformis frequently occurred at those sites at relatively high concentrations. In addition, in the hot water from hot water distal outlets in 28 commercial buildings in the US, the concentration of HPC bacteria was uncorrelated with Legionella positivity (p 0.788) and explained only 0.68% of the variation in Legionella positivity . In the shower water from a US hospital and cooling water from pilot-scale cooling towers, the concentrations of both HPC bacteria and adenosine triphosphate failed to accurately predict Legionella positivity or colonization .
The current routine examination of microbial drinking water quality by detecting conventional fecal indicators fails to assess the negative health effects of water-based OPs in EWSs and is thus biased. First, quantifying fecal indicators is the main or even exclusive means of examining microbial drinking water quality in municipal EWSs (Edberg et al., 2000;Gerba, 2015;Wen et al., 2020 and intestinal enterococci to be 0 per 100 mL of tap water (The Council of the EU, 2020). As discussed above, fecal indicators are uncorrelated with OPs in EWSs, and the presence, absence, and/or concentrations of fecal indicators in drinking water cannot reflect the negative public health effects of water-based OPs. Second, despite the critical role of EWSs in maintaining drinking water quality, most standards regulate only the levels of fecal indicators in the water entering DWDSs (i.e., finished water or plant effluent) and leaving PPSs (e.g., tap water). In the US, the Revised Total Coliform Rule is the only regulation that requires quantifying fecal indicators (i.e., total coliforms and E. coli) throughout DWDSs (US EPA, 2013). Third, even though Legionella frequently occurs in municipal EWSs and causes drinking-water-based disease (i.e., Legionnaires' disease) outbreaks, routine monitoring of Legionella in such systems is rare (Hamilton et al., 2019). Some researchers even inadequately argued for the removal of the requirement for routine monitoring of Legionella in EWSs from all drinking water guidelines (Whiley, 2017). At least, at distal review to proposing an appropriate indicator (i.e., Legionella) of OPs that can better reflect microbial drinking water quality in municipal EWSs. Water utilities and authorities may use Legionella as a supplementary indicator in conjunction with conventional fecal indicators to more comprehensively evaluate microbial drinking water quality and to better protect public health (Zhang et al., 2021b).

SYSTEMS
Scientists have developed multiple criteria for selecting indicator microorganisms to reflect microbial water quality (Gerba, 2015;Wen et al., 2020). Existing criteria are meaningful for selecting indicators of fecal contamination and/or enteric pathogens in drinking water, surface water, recreational water, and wastewater (Gerba, 2015). However, since fecal indicators and OPs in drinking water behave distinctly (Falkinham et al., 2015b;Wang et al., 2017), not all those existing criteria are suitable for selecting an indicator of OPs in municipal EWSs. We need to revise those criteria to select an appropriate indicator microorganism for the presence and concentrations of OPs (the major aspect of microbial drinking water quality) in municipal EWSs. Such revised criteria are missing in the literature. We propose the following novel critical criteria to select an appropriate indicator microorganism of water-based OPs in EWSs:

I.
The indicator is one important or dominant member of the natural inhabitants (e.g., OPs) of drinking water. This property of the indicator ensures that the presence and concentration of the indicator suggest the presence and concentration of OPs as a whole group.

II.
The indicator occurs in municipal EWSs at a high frequency and concentration and can be feasibly examined through multiple approaches. This property of the indicator makes its detection and quantification easy, flexible (i.e., can be assayed by different methods), and affordable.

III.
The presence and concentration of the indicator in EWSs closely correlate with the presence and concentrations of water-based OPs and other critical physicochemical and microbial drinking water quality parameters. This property of the indicator ensures that when the indicator occurs at a high (or low) concentration, the overall concentration of OPs and the public health risks of the drinking water are likely high (or low).

IV.
The occurrence and concentration of the indicator indicate the efficacies of disinfectant residuals. This property links the presence and concentration of the indicator to the efficacy of secondary disinfection (a critical treatment barrier). Therefore, a high frequency of occurrence and concentration of the indicator could suggest drinking water quality deterioration (due to disinfectant residual decay and/or consumption) and the need for enhanced or additional treatment (such as booster disinfection and distribution system cleaning/upgrade).

SYSTEMS
Conventional fecal indicators behave distinctly from and are uncorrelated with water-based OPs in municipal EWSs (Falkinham, 2020). Therefore, we need to select a novel indicator that can indicate the presence and concentrations of OPs in drinking water for more effective OP control and enhanced public health protection. This novel indicator should be selected from common OPs rather than from non-pathogenic natural inhabitants in municipal EWSs ( Table 1). The occurrence and concentrations of non-pathogenic species (which are harmless to humans) do not correlate with the public health risks posed by OPs in drinking water. By contrast, a dominant OP could well indicate the health risks posed by OPs. That dominant OP is itself opportunistically pathogenic and poses public health risks. In addition, the presence and concentration of that dominant OP could well correlate with the presence and concentrations of other OPs because OPs in EWSs share many characteristics such as high resistance to disinfectant residuals, biofilm formation, and proliferation within amebae Zhang and Lu, 2021). Therefore, this review selects a novel indicator of OPs from common OPs in municipal EWSs. On the basis of a comprehensive comparison among the behavior of dominant OPs (Legionella, Mycobacterium, and Pseudomonas) in drinking water, we conclude that Legionella is a promising indicator microorganism inferring the public health risks posed by water-based OPs in municipal EWSs while other OPs lack this indication feature (Zhang et al., 2021b).
Water-based OPs in EWSs have a diverse community with multiple bacterial and amoebal species, while the most common and dominant OPs are Legionella (especially L. pneumophila), Mycobacterium (especially NTM, M. avium, and MAC), and Pseudomonas (especially P. aeruginosa) Falkinham, 2020;Lu et al., 2017;Lu et al., 2016;Lu et al., 2015;Zhang and Lu, 2021;Zhang et al., 2021b). Other OPs (such as certain amoebae) occur in drinking water at lower frequencies and/or lower concentrations and thus do not accurately suggest the overall health risks of OPs in municipal EWSs (Table 1). For instance, low concentrations of other OPs in EWSs do not imply that the public health risks of OPs in the drinking water are low, because the three dominant OPs could occur at high concentrations and have a high potential for infection of immunocompromised and immunocompetent populations. Amoebae in drinking water could have strong correlations with bacterial OPs such as Legionella and Mycobacterium (but not always) (Wang et al., 2012a;Wang et al., 2012b;Lu et al., 2016;Valciņa et al., 2019;Zhang et al., 2021b). The association between Legionella and its amoeba hosts mainly exists in pipe biofilms in municipal EWSs (Storey et al., 2004a;Declerck et al., 2007;Lau and Ashbolt, 2009;Thomas and Ashbolt, 2011;Wang et al., 2012a), whereas the detection and monitoring of Legionella are mainly for planktonic Legionella in bulk water. Therefore, the amoeba-Legionella relationships do not warrantee that amoebae are a good indicator of OPs in municipal EWSs. Pseudomonas (especially P. aeruginosa) as a common OP in EWSs have (much) lower concentrations and frequencies of occurrence than Legionella and Mycobacterium (Wang et al., 2012a;Lu et al., 2016;Lu et al., 2017;Zhang et al., 2021b). Therefore, the occurrence and concentration of Pseudomonas (especially P. aeruginosa) also Zhang and Lu Page 9 Front Environ Sci. Author manuscript; available in PMC 2022 November 10.
do not fully reflect the health risks of OPs, and Pseudomonas (especially P. aeruginosa) is not a good indicator of OPs in municipal EWSs (Table 1).
By contrast, Legionella (especially L. pneumophila) and Mycobacterium (especially NTM, M. avium, and MAC) are the two most dominant OPs in municipal EWSs with comparably high frequencies of occurrence and concentrations (Wang et al., 2012a;Lu et al., 2016;Lu et al., 2017;Zhang et al., 2021b). Legionella (L. pneumophila) and Mycobacterium (M. avium) both have a sporadic occurrence nature in EWSs (Donohue et al., 2019). When either Legionella or Mycobacterium occurs at a high concentration, the overall concentration of OPs is high, and the corresponding drinking water poses significant public health risks, requiring enhanced treatment (e.g., booster disinfection) and distribution system cleaning or upgrade. Oppositely, when the concentration of either Legionella or Mycobacterium is low, the overall concentration of OPs would be low, and the corresponding drinking water would have a low risk of causing disease outbreaks. Therefore, either Legionella (especially L. pneumophila) or Mycobacterium (especially NTM, M. avium, and MAC) could be a good indicator of water-based OPs and their public health risks in municipal EWSs (Table 1).
Compared with Mycobacterium, Legionella could be a better indicator of water-based OPs in municipal EWSs. First, Legionella (especially L. pneumophila) is the first leading cause of drinking-water-based disease outbreaks (Beer et al., 2015;Benedict et al., 2017) and correlates with more OPs than Mycobacterium. For instance, in four full-scale chloraminated DWDSs in the US, the concentration of Legionella significantly and positively correlated with those of Mycobacterium (p 0.042), Pseudomonas (p < 0.001), and total OPs (i.e., the sum concentration of all four examined OPs; p < 0.001) (Zhang et al., 2021b). Legionella also had a nearly significant positive linear correlation with V. vermiformis (p 0.054). Therefore, the presence of Legionella at a relatively high concentration indicates a well-built growth condition for other bacterial OPs (e.g., Mycobacterium and Pseudomonas especially P. aeruginosa) and their amoeba hosts (e.g., Acanthamoeba, N. fowleri, and V. vermiformis) in municipal EWSs. By contrast, Mycobacterium had a significant positive linear correlation with only Legionella (p 0.042). Similarly, a multiple linear regression reveals that Legionella closely correlated with more physicochemical water quality parameters than Mycobacterium in the four chloraminated DWDSs (Zhang et al., 2021b). Legionella had significant linear correlations with water temperature (p 0.002), total chlorine residual concentration (p 0.012), free ammonia concentration (p 0.037), and total trihalomethane (THM) concentration (p 0.019) (negative correlations except for with free ammonia concentration). Mycobacterium had significant linear correlations with only total chlorine residual concentration (p < 0.001) and total haloacetic acid concentration (p 0.018). Since Legionella correlates with more OPs and more critical physicochemical water quality parameters than Mycobacterium, the presence and concentration of Legionella reflect better the public health risks posed by OPs and even general drinking water quality in municipal EWSs. Second, Mycobacterium (M. avium) has a much higher resistance to disinfectant residuals than Legionella (L. pneumophila) and is thus insensitive to changes in disinfectant residual levels and efficacies . Therefore, when disinfectant residuals are at levels effective enough in suppressing the (re) growth of OPs (as well as enteric pathogens and/or fecal contaminants) and ensuring a good drinking water quality, the concentration of Mycobacterium (M. avium) could Zhang  be relatively high, and using Mycobacterium (M. avium) as an indicator of water-based OPs might overestimate the public health risks of OPs in drinking water. Third, among the many OPs in EWSs, Legionella (L. pneumophila) is the most widely known and studied . Numerous culture-dependent and culture-independent assays are available for Legionella (L. pneumophila) detection and quantification (Wang et al., 2017). With the fast development of modern molecular techniques (especially quantitative polymerase chain reactions or qPCRs), assessing the concentration and diversity of Legionella (L. pneumophila) in drinking water is becoming more straightforward, easier to handle, more affordable, and faster (Girones et al., 2010;Lu et al., 2015;Lu et al., 2016;Wolf-Baca and Siedlecka, 2019;Zhang et al., 2021b). In conclusion, we argue that among the many OPs in municipal EWSs, Legionella is the most appropriate indicator of the presence and concentrations OPs (the major aspect of microbial drinking water quality) ( Table 1).
Because of its implications for microbial drinking water quality and public health risks, water authorities worldwide have made guidelines, standards, and regulations to monitor and control Legionella in drinking water ( Legionella is the only water-based OP in the National Primary Drinking Water Regulations (US EPA, 1989). In Europe, the Netherlands, France, and Germany enforce regulatory monitoring for Legionella in drinking water (Ashbolt, 2015b). The revised Drinking Water Directive of the EU even requires monitoring Legionella in DWDSs for risk assessment (The Council of the EU, 2020). In addition, Legionella is a must-measure biological parameter in almost all research studies surveying microbial drinking water quality in municipal EWSs. Since Legionella is a major OP in municipal EWSs and is listed in multiple drinking water regulations, Legionella can be developed to be an indicator of microbial drinking water quality reflecting the occurrence, concentration, and behavior of OPs as a whole group in EWSs.

Legionella Shares Many Critical Characteristics With Other Water-Based Opportunistic Pathogens
Water-based OPs in EWSs share several critical characteristics . They persist and (re)grow in EWSs, infect humans, tolerate high water temperatures, resist disinfectant residuals, and form biofilms ( Figure 1). In addition, bacterial OPs such as Legionella resist the phagocytosis of and proliferate within their amoeba hosts ( Figure 1). Because Legionella shares many critical characteristics with other OPs, the occurrence and concentration of Legionella would well indicate the occurrence and concentrations of other OPs, making Legionella an appropriate indicator of the public health risks posed by OPs in municipal EWSs.
From 2007 to 2008, 36 drinking-water-based disease outbreaks occurred in 24 states in the US and Puerto Rico, and 12 (33%) of the outbreaks were due to acute respiratory illness caused by Legionella (Brunkard et al., 2011). From 2009 to 2010, Legionella caused more outbreaks: Thirty three drinking-water-based disease outbreaks and 1,040 illnesses occurred in 17 states in the US, and Legionella caused 19 (58%) outbreaks and 72 (7%) illnesses (Hilborn et al., 2013). From 2011 to 2014, more than one dozen drinking-waterbased disease outbreaks occurred annually in the US, resulting in hundreds of illnesses, more than 50 hospitalizations, and 6 to 7 deaths per year (Beer et al., 2015;Benedict et al., 2017). Legionella caused more than 50% of those outbreaks, nearly 90% of those hospitalizations, and all those deaths. The annual hospitalizations caused by communityacquired Legionnaires' disease are from 8,000 to 18,000 in the US (Marston et al., 1997), leading to an annual hospitalization cost for water-based Legionnaires' disease in the US to be over 433 million US dollars (Collier et al., 2012;Falkinham et al., 2015b).

Legionella Survives in a Wide Water Temperature Range in Municipal
Engineered Water Systems-Legionella survives in a relatively wide water temperature range, and certain Legionella species or strains are thermotolerant or thermophilic. Legionella (mainly L. pneumophila) survives and/or (re)grows in a water temperature range from approximately 20 °C to approximately 50 °C (Kusnetsov et al., 1996;Leoni et al., 2001;Fields et al., 2002;Kim et al., 2002;Huang et al., 2011;Springston and Yocavitch, 2017) with an optimal (re)growth temperature range from approximately 35 °C to approximately 45 °C (Katz and Hammel, 1987;Rogers et al., 1994;Fields et al., 2002;Mathys et al., 2008;Buse and Ashbolt, 2011;Springston and Yocavitch, 2017). Certain thermophilic/thermotolerant Legionella species or strains, including some L. pneumophila strains, even (re)grow in hot tap water at a temperature greater than 50 °C (Mathys et al., 2008;Lesnik et al., 2016). Therefore, municipal EWSs provide a suitable water temperature range for Legionella (re)growth and are the major reservoir for this genus. Zhang and Lu Page 13 Front Environ Sci. Author manuscript; available in PMC 2022 November 10.

Legionella Has a High Resistance to Disinfectant Residuals in Municipal
Engineered Water Systems-Legionella inherently has a high resistance to disinfectant residuals. In the US and many other countries, water utilities dose disinfectant residuals such as free chlorine, combined chlorine (mainly monochloramine), and chlorine dioxide residuals to DWDSs as a treatment barrier to suppress microbial (re)growth and prevent distribution system upsets (Olivieri et al., 1986;Norton and LeChevallier, 1997;Gagnon et al., 2004;Hwang et al., 2012;Gillespie et al., 2014;Zhang et al., 2021b). The relatively high resistance of Legionella to disinfectant residuals provides it with ecological advantages in municipal EWSs. Therefore, Legionella has a relatively high frequency of occurrence and a high concentration in EWSs, especially in seasons and/or at locations where that treatment barrier is ineffective because of disinfectant residual decay and/or consumption.

Legionella Proliferates Within Amoebae in Municipal Engineered Water
Systems-Legionella proliferates within its amoeba hosts in municipal EWSs (Figure 1). The proliferation significantly enhances its resistance to disinfectant residuals, persistence in drinking water, and pathogenicity (Kim et al., 2002;Campos et al., 2003;Borella et al., 2005a;Lau and Ashbolt, 2009;Marciano-Cabral et al., 2010;Berjeaud et al., 2016). For instance, L. pneumophila within A. polyphage cysts survived free chlorine at a high dose (50 mg·L −1 ) for 18 h at 25 °C (Kilvington and Price, 1990). A pure planktonic L. pneumophila culture was significantly more sensitive (p < 0.005) to free chlorine (initial concentration 2 to 3 mg·L −1 , contact time up to 1 h, 30 °C) and chlorine dioxide (initial concentration 0.4 mg·L −1 , contact time up to 1 h, 30 °C) than L. pneumophila co-cultured with Acanthamoeba (Dupuy et al., 2011). Similarly, the C·t 99.99% value of free chlorine for pure planktonic L. pneumophila serogroup 1 was only 2.5 mg·min·L −1 (initial free chlorine concentration 0.5 mg·L −1 , room temperature) (Cervero-Aragó et al., 2015). The same strain associated with Acanthamoeba was significantly more resistant to free chlorine (C·t 99.99% value from 19 to 245 mg·min·L −1 ; initial free chlorine concentration 0.5, 1.2, or 2.5·mg·L −1 ; room temperature). In addition, L. pneumophila concentration (log 10 CFU per coupon) decreased by 1.47 to 2.07 units after exposure to free chlorine (initial concentration 0.5 mg·L −1 , contact time 15 min to 24 h, 21 °C) in mixed-culture biofilms without V. vermiformis (Donlan et al., 2005). When V. vermiformis was present in the biofilms, L. pneumophila concentration decreased by only 0.30 to 0.90 units under the same chlorination condition. Compared with free chlorine, monochloramine (initial concentration 0.5 mg·L −1 , contact time 15 min to 24 h, 21 °C) killed more L. pneumophila (concentration decreased by 1.77 to 2.29 units) in the biofilms without V. vermiformis. When V. vermiformis was present in the biofilms, monochloramine was less effective in killing L. pneumophila (concentration decreased by only 0.08 to 1.59 units under the same chloramination condition). Furthermore, compared with axenic planktonic L. gormanii, the resistance of Tetrahymena pyriformis (a ciliated protozoan) ingested L. gormanii to free chlorine (pH 7.0, 25 °C) was at least 50 folds greater (King et al., 1988). Legionella within amoebae is also more resistant to heavy metal ions such as copper and silver ions (Unger and Lück, 2012). Copper ions may appear in EWSs because of the use of copper or copper-containing pipes, whereas copper-silver ionization is effective in controlling Legionella in hospital water systems because of the strong antimicrobial activities of silver and copper ions (Liu et al., 1994;Lin et al., 1998a;Stout et al., 1998;Zhang et al., 2014;Zhang et al., 2016a;Zhang et al., 2016b). L. pneumophila in an endosymbiosis phase within A. polyphage was significantly more resistant to silver ions (0.1 mg·L −1 ), copper ions (1.0 mg·L −1 ), or the combination of silver (0.1 mg·L −1 ) and copper (1.0 mg·L −1 ) ions than a pure planktonic culture of L. pneumophila (Hwang et al., 2006).
The proliferation of Legionella within amoebae also enhances its tolerance to thermal treatment (Storey et al., 2004b). For instance, thermal treatment at 50 and 55 °C killed 99.99% of pure planktonic L. pneumophila serogroup 1 after 46 and 8 min, respectively (Cervero-Aragó et al., 2015). The contact time to kill 99.99% of L. pneumophila serogroup 1 associated with Acanthamoeba was significantly longer (745 min at 50 °C and 48 min at 55 °C). However, at 60 and 70 °C, L. pneumophila serogroup 1 associated with Acanthamoeba had the same tolerance to thermal treatment as the pure planktonic culture.

Legionella Forms Biofilms in Municipal Engineered Water Systems-In addition
to the proliferation of Legionella within amoebae, biofilm formation (Figure 1) enhances the resistance of Legionella to disinfectant residuals (Momba et al., 2000;Wright, 2000;Donlan and Costerton, 2002;Kim et al., 2002;Loret et al., 2005;Bridier et al., 2011). For instance, the C·t 99% value of iodine for well-water-grown L. pneumophila serotype 1 biofilm cells (> 960 mg·min·L −1 ) was significantly greater than that for the planktonic cells of well-water-grown L. pneumophila (200 mg·min·L −1 ) (Cargill et al., 1992). Planktonic L. pneumophila took a long time (28 d) to recover in tap water amended with 0.5 mg·L −1 of free chlorine, whereas free chlorine at 200 mg·L −1 (chlorination conducted in water, contact time 1 h) inhibited only moderately the growth of one-and two-month-old L. pneumophila biofilms on stainless steel (Cooper and Hanlon, 2010). In heterotrophic drinking water biofilms, the absolute concentration of naturally occurring L. pneumophila was stable (p > 0.05) when free chlorine residual concentration increased from 0 to 0.2 and to 1.2 mg·L −1 , and the relative concentration of L. pneumophila against total cell numbers even slightly increased when free chlorine residual concentration increased from 0 to 0.2 and to 1.2 mg·L −1 (Gião et al., 2009).

Legionella Frequently Occurs in Municipal Engineered Water Systems at a Relatively High Concentration-Because Legionella survives and (re)grows in a wide
water temperature range, resists disinfectant residuals, proliferates within amebae, and forms biofilms, it has a high frequency of occurrence (greater than approximately 40%) and concentration in municipal EWSs (Borella et al., 2005b;Leoni et al., 2005;Wang et al., 2012a;Lu et al., 2017;Qin et al., 2017;Isaac and Sherchan, 2019;Zhang et al., 2021b). Legionella concentration in the bulk water in the mains of DWDSs could reach 10 7 GCN·L −1 level or be even greater (10 9 GCN·L −1 level at dead-ends) (Whiley et al., 2014a).
In the bulk water in DWSTs, Legionella concentration could reach 10 4 GCN·L −1 level (Qin et al., 2017). In the bulk water of PPSs, Legionella concentration could be greater than 10 6 GCN·L −1 (Wang et al., 2012a). In the sediments of DWSTs, Legionella concentration could reach 10 4 GCN·g −1 or 10 4 cell equivalents·g −1 level Qin et al., 2017). At the endpoints of PPSs (i.e., the inner surfaces of taps and showerheads), the highest concentration of Legionella in biofilms could be at 10 6 GCN·swab −1 level (Wang et al., 2012a). Zhang Front Environ Sci. Author manuscript; available in PMC 2022 November 10.

Legionella Closely Correlates With Drinking Water Quality Parameters in Municipal Engineered Water Systems
The Close Correlations Between Legionella and Physicochemical Drinking Water Quality Parameters-Legionella often correlates with physicochemical drinking water quality parameters in municipal EWSs. For instance, in four full-scale chloraminated DWDSs in the US, Legionella had significant correlations with water temperature (p 0.002), total chlorine residual concentration (p 0.012), free ammonia concentration (p 0.037), and total THM concentration (p 0.019) (Zhang et al., 2021b). In a research center in Germany, Legionella concentration in the hot drinking water from a showerhead positively correlated with water temperature (49.2 to 57.9 °C, R 2 0.730) although Legionella was uncorrelated with water temperature (7.6 to 16.9 °C) in the cold drinking water from the center (R 2 0.097) (Lesnik et al., 2016). In addition, Legionella (L. pneumophila) commonly has a negative correlation with copper ion concentration in building hot water systems (HWSs) (p < 0.05) (Zacheus and Martikainen, 1994;Leoni et al., 2005) probably because copper ions have strong antimicrobial activities (Rogers et al., 1994;Borella et al., 2004;Cooper and Hanlon, 2010;Bargellini et al., 2011;Serrano-Suárez et al., 2013). Legionella also commonly has a negative correlation with disinfectant residual concentrations in EWSs (Zhang et al., 2021b). For instance, the concentration of Legionella in the tap biofilms from three laboratories in northern China was negatively associated with free and total chlorine residual concentrations in the bulk water . In 11 hospitals in San Antonio (Texas, USA) where Legionella appeared in the PPSs, average free chlorine residual concentration in patientroom tap water and the proportion of Legionella-positive sites negatively correlated (linear regression, R 2 0.52, p 0.01) (Kool et al., 1999a). Therefore, a high frequency of occurrence and a high concentration of Legionella in drinking water indicate that disinfectant residuals as a treatment barrier at the current levels are ineffective in suppressing OP (re)growth and the public health risks of the drinking water could be high, requiring enhanced treatment or EWS upgrade. In other words, Legionella could indicate the public health risks of drinking water (mainly posed by water-based OPs) in municipal EWSs.

The Close Correlations Between Legionella and Microbial Drinking Water
Quality Parameters-Legionella in EWSs closely correlates with microbial drinking water quality parameters. In general, Legionella positively correlates with major OPs such as Mycobacterium, Pseudomonas (especially P. aeruginosa), V. vermiformis, and Acanthamoeba in EWSs (Wang et al., 2012a;Serrano-Suárez et al., 2013;Lu et al., 2016). For instance, in the bulk water of four full-scale chloraminated DWDSs in the US, the concentration of Legionella significantly and positively correlated with those of Mycobacterium (p 0.042), Pseudomonas (p < 0.001), and total OPs (p < 0.001) (Zhang et al., 2021b).
In conclusion, Legionella closely correlates with multiple physicochemical and microbial drinking water quality parameters in municipal EWSs. The close correlations between Legionella and other OPs suggest that the presence and concentration of Legionella could infer the presence and concentrations of other major OPs. The close correlations between Legionella and physicochemical drinking water quality parameters (such as disinfectant Zhang and Lu Page 16 Front Environ Sci. Author manuscript; available in PMC 2022 November 10. residuals) further suggest that Legionella could indicate the overall drinking water quality and the potential public health risks posed by OPs in EWSs.

Legionella Indicates the Efficacies of Disinfectant Residuals in Engineered
Water Systems-Disinfectant residuals at relatively high levels are effective in preventing the (re)growth of OPs in EWSs (Olivieri et al., 1986;Norton and LeChevallier, 1997;Gagnon et al., 2004;Hwang et al., 2012;Gillespie et al., 2014;Zhang and Lu, 2021). For instance, Legionella (L. pneumophila) was absent from the bulk water (finished water, DWSTs, water mains entering hospitals, and taps) of an EWS (Pennsylvania, USA) because free chlorine residual of approximately 0.2 mg·L −1 appeared throughout the system (States et al., 1987). The frequencies of occurrence of Legionella in tap water (26%) and water-main biofilms (0%) in a chloraminated EWS in the US were significantly lower than those (64% for tap water and 43% for biofilms) in a no-residual EWS in Norway (Waak et al., 2018).
In the effluents of simulated household water heaters, Legionella concentration negatively correlated with the concentrations of disinfectant residuals (free chlorine or chloramines) in the influents of the heaters (Spearman's rank correlation coefficients between −0.752 and −0.019) (Wang et al., 2015). Similarly, the concentration of Legionella in the bulk water of simulated DWDSs (sampled 14 months after the start-up) negatively correlated with chloramine residual concentrations (Spearman's rank correlation coefficients between −0.908 and −0.693, p < 0.01) (Wang et al., 2012b). In the hot water from 67 buildings in Finland, chlorine residual inhibited the (re)growth of Legionella (Spearman's rank correlation coefficient between chlorine residual level and Legionella concentration −0.19, p < 0.01) (Zacheus and Martikainen, 1994). A recent study found that the distal site/outlet (hot water) positivity of Legionella for 28 buildings in the US was uncorrelated with free chlorine residual level in the incoming cold water of the buildings . However, free chlorine residual still prevented the (re)growth of Legionella in those buildings: Cold water from water mains entering the buildings (free chlorine residual 0.34 mg·L −1 , Legionella concentration 1.4 × 10 4 CFU·L −1 , Legionella positivity 3.7%) and cold water storage tanks (free chlorine residual 0.23 mg·L −1 , Legionella concentration 3.8 × 10 3 CFU·L −1 , Legionella positivity 10.3%) had much greater free chlorine residual concentration than hot water from hot water return lines (free chlorine residual 0.07 mg·L −1 , Legionella concentration 3.6 × 10 5 CFU·L −1 , Legionella positivity 26.7%) and hot water distal outlets (free chlorine 0.05 mg·L −1 , Legionella concentration 4.2 × 10 5 CFU·L −1 , Legionella positivity 30.4%) and thus much lower Legionella concentration and positivity.
Since disinfectant residuals are effective in inhibiting the (re) growth of OPs in EWSs, injecting disinfectants into building water systems reduces the frequencies of occurrence and concentrations of OPs. For instance, the positivity of Legionella at the cold and hot water distal outlets significantly decreased from approximately 60% to less than 10% after two hospitals in the US injected chlorine dioxide into the incoming water mains (Zhang et al., 2009). Likewise, after a hospital (Pennsylvania, USA) injected monochloramine into the HWS of a building, the distal site/outlet positivity of culturable Legionella was significantly lower compared with the control building (no monochloramine dosing) or the Zhang  same building before the chloramination Duda et al., 2014;. In a hospital building (Modena, Italy), injecting monochloramine into a HWS (target concentration 1.5 to 3.0 mg·L −1 , continuous application) significantly reduced the frequencies of occurrence of L. pneumophila in the hot water from heaters, return loops, and distal outlets from 97.0% (of 32 samples) to 13.3% (of 60 samples) (first-year application) Marchesi et al., 2012). Meanwhile, the geometric means of L. pneumophila concentration for the positive hot water samples dramatically dropped from 1.7 × 10 4 to 3.4 × 10 2 CFU·L −1 after the dosing of monochloramine (first-year application). Over three years after the dosing of monochloramine, Legionella appeared in only 8 hot water samples (9.5% of 84 samples) with a low concentration (3.3 × 10 2 CFU·L −1 , geometric mean for the 8 positive samples) (Marchesi et al., 2013).
As discussed above, because Legionella concentration has a general negative correlation with disinfectant residual levels, Legionella well indicates the efficacy of an important treatment barrier in municipal EWSs, secondary disinfection (i.e., dosing and maintaining disinfectant residuals) (Zhang et al., 2021b). When Legionella occurs in EWSs at a high frequency and/or a high concentration, the treatment barrier is ineffective in preventing the (re)growth of Legionella and other OPs, where the OPs might have a high potential causing disease (such as legionellosis) outbreaks. Therefore, Legionella occurring at a high frequency and/or concentration indicates drinking water quality deterioration and requires enhanced treatment (such as booster disinfection) and/or EWS cleaning/upgrade. Oppositely, drinking water from an EWS where Legionella occurs at a low frequency and/or a low concentration generally has a high quality and will less likely cause disease outbreaks due to the presence of OPs.

Monochloramine Is More Effective Than Free Chlorine in Controlling
Legionella in Engineered Water Systems-Monochloramine has a long-lasting bactericidal effect and is more effective than free chlorine in controlling Legionella in municipal EWSs (Kool, 2001;Lin et al., 2011;Buse et al., 2012;Springston and Yocavitch, 2017). Therefore, the prevalence of Legionella in chloraminated EWSs is lower than that in chlorinated EWSs, suggesting that drinking water from chloraminated EWSs is less likely to cause outbreaks of legionellosis (Kool, 2001;Kim et al., 2002;Lin et al., 2011;Buse et al., 2012). Among 32 hospitals in the US where drinking-water-associated, hospitalacquired Legionnaires' disease outbreaks occurred at least once per hospital from 1979 to 1997, free chlorine was the disinfectant residual in the drinking water for 31 hospitals, and monochloramine was the disinfectant residual in the drinking water for only one (Kool et al., 1999b;Kool et al., 2000). By contrast, for 48 hospitals in the US without a drinking-water-associated, hospital-acquired Legionnaires' disease outbreak from 1979 to 1997, free chlorine was the disinfectant residual for 36 hospitals, and monochloramine was the disinfectant residual for 12. Therefore, hospitals supplied with drinking water with monochloramine residual (adjusted odds ratio for outbreaks of Legionnaires' disease 1.0) are significantly less likely to have drinking-water-associated, hospital-acquired Legionnaires' disease outbreaks than hospitals with free chlorine residual in drinking water (adjusted odds ratio for outbreaks of Legionnaires' disease 10.2). Similarly, 16% of 38 hospitals in the US that had outbreaks/cases of hospital-acquired Legionnaires' disease used drinking water with Zhang and Lu Page 18 Front Environ Sci. Author manuscript; available in PMC 2022 November 10. monochloramine residual, while 46% of 114 hospitals in the US without an outbreak/case of hospital-acquired Legionnaires' disease used drinking water with monochloramine residual (Heffelfinger et al., 2003). As a result, the adjusted odds ratio (0.20) for definite cases or outbreaks of hospital-acquired Legionnaires' disease in a hospital using drinking water with monochloramine residual is low.
However, monochloramine residual is more effective than free chlorine residual in controlling Legionella in municipal EWSs. The greater effectiveness of monochloramine residual in preventing Legionella (re)growth in EWSs is due to three reasons. First, the majority of Legionella cells in EWSs are in biofilms/sediments and within protozoans (e.g., amoebae) (Stout et al., 1985;Lin et al., 1998b;Fields et al., 2002;Batté et al., 2003;Lau and Ashbolt, 2009;Abdel-Nour et al., 2013;Ashbolt, 2015a;. Approximately 95% of the biomass in EWSs is in biofilms and sediments on pipe walls, and only 5% of the biomass is in bulk water (Block, 1992;Flemming et al., 2002;Wingender and Flemming, 2011). For instance, in a pilot-scale HWS, more than 98.5% of Legionella cells lived in the pipe biofilms (Saby et al., 2005). Monochloramine well penetrates biofilms and controls biofilm contamination (LeChevallier et al., 1988;Norton and LeChevallier, 1997;Lee et al., 2011b;Pressman et al., 2012). By contrast, the rapid reaction between free chlorine and organic matter of biofilm origin greatly prevents free chlorine from penetrating biofilms (Chen and Stewart, 1996;Grobe et al., 2002;Bridier et al., 2011). For example, exposing a 526 μm thick P. aeruginosa biofilm to a flowing free chlorine solution (18.6 mg·L −1 ) for 3 h only raised free chlorine concentration at the substratum to 1.8 mg·L −1 (i.e., approximately 10% of the bulk free chlorine level) (Chen and Stewart, 1996). Similarly, free chlorine level in dual-species biofilms of P. aeruginosa and K. pneumoniae was only 20% or less of the free chlorine level in the bulk liquid (De Beer et al., 1994). Therefore, monochloramine residual penetrates biofilms faster and more completely than free chlorine residual in municipal EWSs (Lee et al., 2018). Indeed, the maximum initial penetration rate of monochloramine to nitrifying biofilms was approximately 170 times that of free chlorine (equivalent chlorine levels) (Lee et al., 2011b). Because monochloramine penetrates biofilms faster and more completely but free chlorine only inactivates cells on biofilm surfaces contacting bulk water (Kelsey, 2014;Lee et al., 2018), monochloramine is more effective in killing biofilm cells of Legionella. For instance, monochloramine and free chlorine at the same C·t value of 270 mg·min·L −1 (initial concentration 1.5 mg·L −1 for both Zhang and Lu Page 19 Front Environ Sci. Author manuscript; available in PMC 2022 November 10. disinfectants) reduced L. pneumophila concentration in mixed-culture bacterial biofilms by approximately 99.7% and approximately 87.4%, respectively (Donlan et al., 2001). In addition, monochloramine is more effective than free chlorine in killing biofilm cells of other bacterial species: The C·t > 99.9% value of monochloramine (3 mg·min·L −1 ) against K. pneumoniae attached to glass slides was significantly lower than the C·t 99.6% value of free chlorine (150 mg·min·L −1 ) (LeChevallier et al., 1988). The slow and incomplete biofilm penetration of free chlorine is an important reason for the relatively strong resistance of biofilm cells to free chlorine (Grobe et al., 2002).

Numerous Methods Are Available to Detect and Quantify Legionella in Engineered Water Systems
Numerous methods are available to recover, identify, and/or quantify Legionella in drinking water from EWSs. Those methods include the "gold standard" culture-based assays (e.g., with special media or amoebal co-culture), molecular assays through analyzing DNA/RNA/proteins (e.g., qPCRs, high-throughput DNA sequencing, and flow cytometry), and phenotypic assays (e.g., immunoassays) (Lee et al., 2011a;Aw and Rose, 2012;Wang et al., 2017;Zhang et al., 2021a).
Therefore, standard or traditional culture-based assays underestimate the prevalence and public health risks of Legionella in municipal drinking water. On the other hand, traditional or regular qPCRs as a very common method for Legionella detection and quantification in EWSs detect total DNA or genetic markers of Legionella, including those from both live (culturable and VBNC) and dead Legionella cells (Dusserre et al., 2008;Wang et al., 2012a;Whiley et al., 2014a;Taylor et al., 2014;Lu et al., 2015;Lu et al., 2016;Lu et al., 2017;Qin et al., 2017;Waak et al., 2018;Isaac and Sherchan, 2019;Zhang et al., 2021b). Therefore, traditional qPCRs (significantly) overestimate the health risks of Legionella in EWSs because dead Legionella cells no longer infect humans (Fleischmann et al., 2021). One promising approach to resolving this overestimation issue is propidium monoazide qPCRs (PMA-qPCRs). PMA is a DNA intercalating molecule that penetrates only membrane-damaged or membrane-compromised cells (i.e., dead cells), binds to DNA within the dead cells, and inhibits DNA amplification during subsequent qPCRs (Nocker et al., 2007;Li and Chen, 2013;Golpayegani et al., 2019;Guo et al., 2021). The DNA or genetic markers of Legionella within membrane-intact (i.e., live, including culturable and VBNC) cells are not bound to PMA and are normally amplified during qPCRs. Therefore, PMA-qPCRs detect only live (culturable and VBNC) Legionella cells and more accurately estimate the public health risks of Legionella. Researchers have successfully applied PMA-qPCRs to detect live Legionella (L. pneumophila) cells in various settings including drinking water (Yáñez et al., 2011;Slimani et al., 2012;Li et al., 2015;Bonetta et al., 2017;Bonetta et al., 2018;Kontchou and Nocker, 2019). Using PMA-qPCRs to detect Legionella (and other microbes) has certain limitations such as incomplete reduction/suppression of the qPCR signals of DNA from dead cells, especially when qPCR amplicons are short or when the abundance of dead cells is high (Pan and Breidt, 2007;Kralik et al., 2010;Luo et al., 2010;Yáñez et al., 2011;Taylor et al., 2014;Ditommaso et al., 2015;Codony et al., 2020). In addition to PMA-qPCRs, reverse transcription qPCRs (RT-qPCRs) are a useful tool for detecting viable Legionella in (drinking) water (Bej et al., 1991;Boss et al., 2018).
A comprehensive summary of the traditional and novel Legionella recovery, identification, and quantification methods is beyond the scope of this literature review. Interested readers may consult a few publications for the details of those methods Buse et al., 2012;Kirschner, 2016;Whiley and Taylor, 2016;Wang et al., 2017;Párraga-Niño et al., 2018;Baume et al., 2019;Ezenarro et al., 2020;Walker and McDermott, 2021). Because of its frequent occurrence, relatively high concentration, and various quantification methods, Legionella is a promising indicator of water-based OPs (the major aspect of microbial drinking water quality) in municipal EWSs.

DISCUSSION AND FUTURE WORK
Our early work preliminarily proposed that Legionella could be an indicator of non-fecal pathogens in municipal DWDSs (Lu et al., 2016). Our recent study examined the dynamics of dominant OPs in four full-scale DWDSs and confirmed that Legionella is a promising indicator of water-based OPs and microbial drinking water quality in municipal EWSs (Zhang et al., 2021b). The current systematic review greatly expands our previous argument and provides convincing reasons for the appropriateness of Legionella as an indicator of water-based OPs (the major aspect of microbial drinking water quality) in municipal EWSs. However, this review does not imply that Legionella should be used as a sole indicator to infer microbial drinking water quality (Huang et al., 2021). First, Mycobacterium is highly resistant to disinfectant residuals  and could be more resistant to chloramines than Legionella (Lin et al., 2011). Therefore, in a chloraminated EWS, a low frequency of occurrence and a low concentration of Legionella do not necessarily indicate a low health risk from Mycobacterium. In this case, water utilities might use Mycobacterium and Legionella in combination to better examine microbial drinking water quality Huang et al., 2021). Future work needs to comprehensively compare the appropriateness of Legionella and Mycobacterium as indicators of water-based OPs and microbial drinking water quality. Evaluating the combined use of Legionella and Mycobacterium as indicators of microbial water quality is also necessary. Second, this review does not imply that we shall abandon the routine monitoring of conventional fecal indicators in municipal EWSs. Even though conventional fecal indicators do not indicate the presence and concentrations of OPs in drinking water, they do well indicate issues in treatment utilities (e.g., failure of disinfection processes), EWS problems (i.e., infiltration, backflow, and water main breaks), and the presence of enteric pathogens and/or fecal contaminants in drinking water. Therefore, this systematic review does not argue substituting Legionella for conventional fecal indicators as the sole indicator of microbial drinking water quality. Instead, this review argues that Legionella as an additional or supplementary indicator for microbial water quality should be used in conjunction with conventional fecal indicators. The combined use of Legionella and fecal indicators as microbial drinking water quality indicators would help more comprehensively understand the health risks of drinking water in municipal EWSs. Future work needs to explore how to use Legionella and conventional fecal indicators in combination to better monitor microbial drinking water quality and better protect public health.
This review focuses on convincing the water industry that Legionella is an appropriate supplementary indicator microorganism inferring the prevalence of water-based OPs (the major aspect of microbial drinking water quality) in municipal EWSs. When an indicator microorganism is proposed, carefully selecting a threshold (i.e., a numerical permissible limit) of the indicator that indicates a health risk or requires an action to be taken is important. For instance, when the concentration of the indicator exceeds a certain threshold, a boil-water-advisory should be issued, and enhanced treatment (i.e., booster disinfection) is required. Future work should propose Legionella limits (i.e., concentration limits and positivity limits) for drinking water that indicate serious public health risks and require enhanced or additional treatment. This review grouped municipal DWDSs and building PPSs to EWSs for the ease of discussion. However, microbial growth conditions between DWDSs and PPSs are significantly different (Zhang et al., 2021b). In addition, the legal landscapes of water responsibility between DWDSs and PPSs are distinct. For instance, general microbial water quality monitoring is typically done only in DWDSs rather than in PPSs. Legionella as an indicator of microbial water quality in those two systems thus has distinct roles. Future work needs to clarify how to appropriately use Legionella as an indicator of water-based OPs and microbial water quality in these two different systems.

CONCLUSION
Legionella as an important natural inhabitant frequently occurs in municipal EWSs with a relatively high concentration. Opportunistically pathogenetic species in the genus Legionella (mainly L. pneumophila but include many other species) in EWSs frequently cause drinking-water-based disease (e.g., Legionnaires' disease) outbreaks and pose significant public health risks. Quantification of Legionella in EWSs with modern molecular-based assays (especially qPCRs) is affordable and fast. Legionella in EWSs closely correlates with major physicochemical and microbial water quality parameters, especially the occurrence and concentrations of water-based OPs. Biofilms/sediments and amoebae are the main niches of Legionella in EWSs, and the association of Legionella with biofilms/sediments and amoebae significantly enhances the already high resistance of Legionella to disinfectant residuals. Legionella indicates the efficacies and explains the modes of action of disinfectant residuals in EWSs. For instance, monochloramine is more effective than free chlorine in controlling Legionella in EWSs because monochloramine penetrates biofilms/sediments/ amoebae faster and more completely and has a long-lasting antimicrobial/bactericidal effect (i.e., more stable than free chlorine). Those features of Legionella make it a promising indicator of the prevalence of water-based OPs and microbial water quality in municipal EWSs.

FIGURE 1 |.
The decay of conventional fecal indicators and the (re)growth of opportunistic pathogens (OPs) along municipal engineered water systems (EWSs