key: cord-0029272-eanzcn7w
authors: Rahmatika, Iftita; Kurisu, Futoshi; Furumai, Hiroaki; Kasuga, Ikuro
title: Dynamics of the Microbial Community and Opportunistic Pathogens after Water Stagnation in the Premise Plumbing of a Building
date: 2022-03-24
journal: Microbes Environ
DOI: 10.1264/jsme2.me21065
sha: 24951eccf5eda9f2d496ec78bd106f409820634f
doc_id: 29272
cord_uid: eanzcn7w

In premise plumbing, microbial water quality may deteriorate under certain conditions, such as stagnation. Stagnation results in a loss of disinfectant residual, which may lead to the regrowth of microorganisms, including opportunistic pathogens. In the present study, microbial regrowth was investigated at eight faucets in a building over four seasons in one year. Water samples were obtained before and after 24 h of stagnation. In the first 100‍ ‍mL after stagnation, total cell counts measured by flow cytometry increased 14- to 220-fold with a simultaneous decrease in free chlorine from 0.17–0.36‍ ‍mg L(–1) to <0.02‍ ‍mg L(–1). After stagnation, total cell counts were not significantly different among seasons; however, the composition of the microbial community varied seasonally. The relative abundance of Pseudomonas spp. was dominant in winter, whereas Sphingomonas spp. were dominant in most faucets after stagnation in other seasons. Opportunistic pathogens, such as Legionella pneumophila, Mycobacterium avium, Pseudomonas aeruginosa, and Acanthamoeba spp., were below the quantification limit for real-time quantitative PCR in all samples. However, sequences related to other opportunistic pathogens, including L. feeleii, L. maceachernii, L. micdadei, M. paragordonae, M. gordonae, and M. haemophilum, were detected. These results indicate that health risks may increase after stagnation due to the regrowth of opportunistic pathogens.

Biological stability is a critical aspect in drinking water supply systems (Prest et al., 2016) . To achieve biological stability and prevent the growth of microorganisms, including pathogens, a disinfectant residual is generally maintained throughout the drinking water distribution system in several countries, including Japan. In Japan, free chlorine (>0.1 mg L -1 ) or combined chlorine (>0.4 mg L -1 ) must be maintained at the end of the pipe. However, stagnation, which is defined as intermittent water use at individual outlets for an extended period, may occur in premise plumbing, causing longer incubation times (Proctor and Hammes, 2015; Rhoads and Hammes, 2021) . It reduces disinfectant levels, allows microbial regrowth, and, thus, changes the microbial community composition in drinking water (Lautenschlager et al., 2010; Zhang et al., 2021b) .

Premise plumbing is characterized by a small-diameter pipe with a high surface area to volume ratio, in which disinfection decay may be faster than in the water main (Hoekstra and Tsvetanova, 2010; Ling et al., 2018) . Some pipe materials release biodegradable organic matter, which further promotes the growth of planktonic bacteria and the development of biofilms on pipe surfaces (Wen et al., 2015; Rahmatika et al., 2020) . Previous studies demonstrated that stagnation contributed to bacterial proliferation. Intact cell counts and heterotrophic plate counts (HPC) were shown to increase after overnight stagnation in premise plumbing (Lautenschlager et al., 2010) . Furthermore, the significant decay of free chlorine after stagnation caused total cell counts (TCC) to increase (Ling et al., 2018; Zhang et al., 2021b) . Increases in water temperature and biofilm detachment were also found to promote microbial growth during stagnation (Zlatanovic et al., 2017; Bedard et al., 2018; Peng et al., 2020; Zhang et al., 2021a) .

Previous studies characterized microbial community compositions in premise plumbing and evaluated their dynamics under stagnation (Lautenschlager et al., 2010; Ling et al., 2018; Zhang et al., 2021b) . Some genera, including Methylobacterium spp., Sphingomonas spp., Sphingobium spp., Mycobacterium spp., and Pseudomonas spp., increased after stagnation (Zhang et al., 2021a; 2021b) . Ley et al. (2020) reported a positive correlation between Legionella spp. and stagnation times (Ley et al., 2020) . Multiple studies have linked the occurrence of premise plumbing-associated opportunistic pathogens (i.e., L. pneumophila, M. avium, and P. aeruginosa) to water-borne infections, including Legionnaires' disease, Pontiac fever, and nontuberculous pulmonary disease (Anaissie et al., 2002; Falkinham et al., 2008; Wang et al., 2013) .

Regardless of the health risks of decreased water quality associated with microbial regrowth in premise plumbing, microbial water quality after stagnation is not routinely monitored. Moreover, limited information is currently available on the factors contributing to variations in the microbial water quality of different faucets in the same building. A more detailed understanding of the characteristics of the drinking water microbiome after stagnation is crucial for developing effective strategies that ensure the safety of water for consumption. In the present study, one-year monitoring at different faucets in a building was performed to reveal the dynamics of the microbial community before and after stagnation. In addition, the occurrence of opportunistic pathogens was analyzed to assess potential health risks.

The present study was conducted at an 11-story building at the University of Tokyo, Tokyo, Japan. Drinking water samples were collected from eight cold water faucets (F1-F8) at the handwashing basins in laboratory rooms. The target faucets were located on the 3 rd and 4 th floors. The premise plumbing in that building was constructed using steel with a polyethylene powder lining. It has been used since 1995 when the building was constructed. The building stores drinking water in rooftop tanks (8 m 3 ), which is distributed to the faucets on each floor. Water is supplied from a water treatment plant that treats surface water by coagulation followed by sedimentation, primary rapid sand filtration, ozonation, biological activated carbon filtration, and secondary rapid sand filtration. The typical free chlorine concentration in treated water at the plant is 0.5-0.7 mg L -1 to maintain a free chlorine residual at faucets of >0.1 mg L -1 , which is regulated by the Japanese drinking water quality guidelines.

To evaluate seasonal variations in microbial regrowth, pre-and post-stagnation samples were collected in four seasons. Samples were collected from eight faucets (F1-F8) in summer (June-July 2018) and from six faucets (F1, F3, F4, F5, F7, and F8) in autumn (November 2018), winter (January 2019), and spring (April 2019) (52 samples) (Table S1 ). Prior to designated stagnation, a 10-L pre-stagnation sample was collected after flushing water for 5 min at an approximate flow rate of 5 L min -1 . The pre-stagnation sample represents drinking water samples with the minimized influence of stagnation. Afterwards, the faucets were closed for 24 h. Post-stagnation samples were collected from the same faucet after approximately 24 h of stagnation at a flow rate of approximately 5 L min -1 . The first 2 L of the flushed post-stagnation sample was collected in incremental samples of 100 mL.

Daily variations in stagnation-induced microbial growth were evaluated by repeating the above sample collection procedure at F1 on six consecutive days in April 2019.

Temperature and free chlorine were measured with a digital thermometer (SK-250WP II-R; Sato Keiryoki) and HACH chlorine pocket colorimetric according to the N,N-diethyl-pphenylenediamine method (Hach), respectively. The limit of quantification (LOQ) of the free chlorine measurement was 0.02 mg L -1 .

TCC were analyzed by flow cytometry. SYBR ® Green I (Invitrogen) diluted 100-fold with dimethylsulfoxide (FUJIFILM Wako Pure Chemical) was used as a working dye solution. Then, 490 μL of the sample was stained with 5 μL of the working dye solution and 5 μL of 0.5 mol L -1 EDTA (Invitrogen). In the analysis, 0.2-μm filtered samples were used as a blank. This mixture was vortexed and incubated at 37°C for 10 min under dark conditions and analyzed using a BD Accuri C6 ® flow cytometer (BD Biosciences). Electronic gating for bacterial cell quantification was selected on density plots of green fluorescence in the FL1 channel (535±15 nm) and red fluorescence in the FL3 channel (>670 nm) (Fig. S1 ). Constant gating was applied for all samples.

HPC were enumerated using the spread plate method following Standard Method 9215 C (APHA, 1998). Water samples were serially diluted in autoclaved phosphate-buffered saline (0.01 mol L -1 , pH 7.2-7.4) (FUJIFILM Wako Pure Chemical). One hundredmicroliter aliquots of non-diluted or diluted samples were spread on R2A agar medium (Beckton, Dickinson) and incubated at 20°C for 7 days.

To collect the biomass from samples, the first 1 L of poststagnation samples was filtered through 0.2-μm polycarbonate filters (Isopore; Merck Millipore). Regarding pre-stagnation samples, 10 L of samples were concentrated using a hollow fiber ultrafiltration (HFUF) unit with a surface area of 25 m 2 (APS-25SA; Asahi Kasei Medical), following a previously described method . The recovery ratio of E. coli cells by HFUF was reported as 96.5±8.5% . Prior to the filtration process using a HFUF unit, a blocking process was conducted by circulating 5% fetal bovine serum (Thermo Fisher Scientific) in the HFUF cartridge for 2 min. The HFUF cartridge containing fetal bovine serum was incubated at room temperature overnight. The sample volume was concentrated from 10 L to approximately 150 mL by applying the dead-end filtration method using a rotary pump at a flow rate of 1.0 L min -1 . After filtration, the concentrates in the filter were back-flushed into a sterile bottle and mixed with 100× elution buffer (10% Tween 80, 1% sodium polyphosphate, and 0.1% antifoam A) (Sigma-Aldrich). The mixture of concentrates and elution buffer was circulated for 2 min to further recover cells from the HFUF cartridge. The final concentrated sample was filtered through a 0.2-μm polycarbonate membrane filter. Filters were stored at -20°C until DNA extraction.

DNA was extracted from the filter using the FastDNA SPIN Kit for Soil (MP Biomedicals) following the manufacturer's instructions. Briefly, the filter was placed in a lysing matrix tube, and dissolved in phenol-chloroform-isoamyl alcohol (25:24:1) (Nippon Gene). Bead beating was performed with Fast-Prep (MP Biomedicals) at a speed of 6.0 m s -1 for 40 s. The DNA extraction efficiency of this method evaluated using E. coli was 93.8±4.9% (n=3). A negative control experiment using milli-Q water revealed that DNA contamination was negligible.

The number of gene copies of Legionella spp. (23S-5S rRNA spacer region), L. pneumophila (mip gene), Mycobacterium spp. (16S rRNA gene), M. avium (16S rRNA gene), Pseudomonas aeruginosa (regA gene), and Acanthamoeba spp. (18S rRNA gene) were quantified using TaqMan-based qPCR (Herpers et al., 2003; Riviere et al., 2006; Feazel et al., 2009; Gensberger et al., 2013) . The primers and probes used for target genes and thermal conditions are described in Table S2 . Each reaction mixture (20 μL) consisted of 7.7 μL of distilled water, 10 μL of the LightCycler ® Probes Master solution (Roche Life Science), 0.1 μL of each primer (100 pmol μL -1 ), 0.1 μL of the probe (40 pmol μL -1 ), and 2 μL of the DNA template. Standards were prepared using a 10fold serial dilution (5.0×10 0 -5.0×10 6 copies μL -1 ) of an artificially synthesized plasmid containing the target genes (Table S3 ). LOQ was set as the lowest concentration of the standards. All reactions were performed in triplicate using a LightCycler ® LC480 system (Roche Life Science). PCR reaction efficiencies were in the range of 90.15-99.50% (Table S2 ). PCR inhibition was not observed.

The microbial community structure was analyzed using amplicon sequencing targeting the V4 region of 16S rRNA genes. The V4 regions of bacterial and archaeal 16S rRNA genes were amplified by primers with adapter sequences (underlined): 515F (5′-ACACTCTTTCCCTACACGACGCTCTTCCGATCT-GTGCCAG CMGCCGCGGTAA-3′) and 806R (5′-GTGACTGGAGTTCAG ACGTGTGCTCTTCCGATCT-GGACTACHVGGGTWTCTAAT-3′) (Caporaso et al., 2011) . The thermal condition of the first PCR Rahmatika et al. consisted of 94°C for 2 min, followed by 25 cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 30 s, with a final extension at 72°C for 5 min. TaKaRa Ex-Taq HS (TaKaRa Bio) was used for the PCR reaction. The second PCR and amplicon sequencing based on an Illumina MiSeq platform with 300-bp paired-end reads were performed at the Bioengineering Lab, Japan. FASTX-Toolkit (ver. 0.0.14) (http://hannonlab.cshl.edu/fastx_toolkit/) was applied to filter reads whose sequences had a perfect match with the primer sequences. The deletion of primer sequences, paired-end reads under 250 bases, and low qualified bases and the merging of paired-end reads were processed using Quantitative Insights into Microbial Ecology 2 (QIIME 2) (Bolyen et al., 2019) with parameters in default conditions. DADA2 in the QIIME 2 pipeline (ver. 2019.10) was used to remove chimeric and noisy sequences. Representative sequences were clustered into amplicon sequence variants (ASVs) with 100% similarity by QIIME 2. Final ASVs were taxonomically classified by referring to the EzBioCloud 16S database (May 2018) (Yoon et al., 2017) .

In addition to the Illumina MiSeq platform, DNA from 12 post-stagnation samples, which showed that the relative abundance of Legionella or Mycobacterium was >1% of the total microbial community, were analyzed using nanopore sequencing targeting full-length 16S rRNA genes. The library for the 16S rRNA gene analysis was prepared using the 16S Barcoding Kit (Oxford Nanopore Technologies). Primers with adapter sequences (underlined) were used to amplify full-length 16S rRNA genes: 16S-F (5′-ATCGCCTACCGTGAC-Barcode-AGAGTTTGATCMT GGCTCAG-3′) and 16S-R (5′-ATCGCCTACCGTGAC-Barcode-CGGTTACCTTGTTACGACTT-3′). The thermal conditions of the first PCR consisted of 94°C for 2 min, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, and 65°C for 80 s, with a final extension at 65°C for 5 min. LongAmp Taq 2× Master Mix (New England BioLabs) was used for the PCR reaction. The library was loaded onto R9.4 flow cells on the GridION platform (Oxford Nanopore Technologies) at the Bioengineering Lab, Japan. Base-calling and de-barcoding were performed using GUPPY software (ver. 3.2.6) (Oxford Nanopore Technologies). Adapter sequences were trimmed using Porechop (ver. 0.2.3). The SeqIO module in Biopython was used to extract reads with 1,400-1,600 bp. After subsampling by Filtlong software (ver 0.2.0), 221,978-678,877 sequences per sample were obtained. The taxonomic assignment was performed with RDP_16S_V16_sp using Usearch (ver. 11.0.667).

All sequence data in the present study were submitted to the DNA Data Bank Japan (DDBJ) Sequence Read Archive and are available under accession number DRA011523.

Statistical analyses of TCC, HPC, and qPCR data were performed after log transformation. The paired Student's t-test and an analysis of variance (ANOVA) followed by Tukey's honestly significant difference test (HSD) were performed using R statistical software (ver. 3.4.1), and P<0.05 was considered to be significant. The relevant statistical test results, including P values, F values, and 95% CI, are shown in Table S5 , S6, and S7.

Alpha diversity indices (observed richness and the Shannon diversity index) were calculated with the vegan package (ver. 2.5-7) in R statistical software. To evaluate the significance of differences in alpha-diversity indices between sample groups, the paired Student's t-test, an ANOVA, and HSD test were performed. To identify any significant differences (P<0.05) in the microbial community between sample groups, an analysis of similarity (ANOSIM) was performed based on the Bray-Curtis dissimilarity index at the OTU level using the vegan package in R statistical software with 999 permutations. The value generated from random permutations (R value) ranged between 0 (complete similarity) and 1 (complete dissimilarity). Additionally, hierarchical clustering at the OTU level was created with the vegan package (ver. 2.5-7) in R statistical software based on the Bray-Curtis dissimilarity index. 

stagnation, free chlorine in the first 100 mL decreased to <0.02 mg L -1 , while TCC increased to 5.5×10 4 cells mL -1 . The temperature pre-stagnation was 18.3°C and increased to 20.7°C in the first 100 mL after 24 h of stagnation (Table  S4 ). TCC gradually decreased to 1.2×10 4 cells mL -1 after flushing 500 mL water, and reached 4.0×10 3 cells mL -1 after flushing 2 L. Simultaneously, the free chlorine concentration recovered to 0.17 mg L -1 after flushing 500 mL water and reached 0.33 mg L -1 after flushing 2 L. Daily variations in microbial regrowth after 24 h of stagnation were assessed on six consecutive days at F1. TCC levels after stagnation were 5.2±0.5×10 4 cells mL -1 (average±standard deviation, n=6) (Fig. S2a) . Microbial community structures after repetitive stagnation were similar (Fig. S2b) . These results indicated that the regrowth event at the same faucet was reproducible in the short term.

Fig. 2 summarizes variations in water temperature, free chlorine, TCC, and HPC in all samples collected during different seasons. The increase in water temperature in the first 100 mL of samples collected after stagnation was observed for most samples (Fig. 2a) . Differences in water temperature before and after stagnation were significant (paired Student's t-test, P=3.059×10 -5 ), with the highest increase being 8.3°C for an F1 sample collected during winter (Table S4) . While the free chlorine of pre-stagnation samples (0.17-0.36 mg L -1 ) was not significantly different among seasons, as confirmed with ANOVA test (P=0.365), the significant depletion of free chlorine to <0.02 mg L -1 in the first 100 mL sampled after stagnation (paired Student's t-test, P=9.738×10 -23 ) was consistently observed after stagnation for all samples (Fig. 2b) . Culturable bacteria were not detected in pre-stagnation samples (Fig. 2c) . HPC increased in the first 100 mL collected after stagnation across all samples. HPC in post-stagnation samples were 1.7±2.2×10 3 CFU mL -1 , and differences among seasons were not significant (ANOVA, P=0.540). Several samples, such as F1 (summer, autumn, and winter) and F3 (summer, autumn, and winter), exceeded the target levels for drinking water quality in Japan (2,000 CFU mL -1 ) ( Table S4) . TCC in pre-stagnation samples were 2.7±2.5×10 3 cells mL -1 and did not significantly different among seasons (ANOVA, P=0.521) (Fig.  2d) . A significant increase in TCC from 2.7±2.5×10 3 to 1.2±0.9×10 5 cells mL -1 in the first 100 mL of water collected after stagnation was observed for all samples collected in all four seasons (paired Student's t-test, P=8.598×10 -21 ). Average TCC levels post-stagnation were the highest in summer; however, differences were not significant (ANOVA, P=0.736). TCC levels peaked in the first 100 mL of water collected after stagnation in all samples. The first 100 mL contributed to 10-47% of all cells discharged in the first 2 L of post-stagnation samples.

Microbial community structures in pre-and poststagnation samples were analyzed. An amplicon sequencing analysis revealed that 1,591 and 872 OTUs were recovered from pre-and post-stagnation samples, respectively, ranging between 110 and 339 OTUs per sample. Alpha diversity measured by OTU richness and the Shannon diversity index showed seasonal trends in pre-and post-stagnation samples (Fig. S3) . Winter samples exhibited significantly lower OTU richness in pre-stagnation samples than autumn samples, as confirmed by Tukey's HSD test (ANOVA, P=0.0299; Tukey's HSD, P=0.039). OTU richness in poststagnation samples in winter was significantly lower than in summer and spring (ANOVA, P=8.300×10 -4 ; Tukey's HSD, P=4.334×10 -4 -0.045). In winter, microbial communities pre-and post-stagnation were also less diverse; the Shannon index was significantly lower in winter than in other seasons (ANOVA, P=0.014 and 0.015, respectively; Tukey's HSD, P=0.001-0.053). Average OTU richness decreased after stagnation in all seasons; however, the decrease was not significant in spring (paired Student's t-test, P=0.303).

To compare the microbial community composition preand post-stagnation among different seasons, a cluster analysis based on the Bray-Curtis dissimilarity index was conducted (Fig. 3) . Four major clusters were obtained as follows: (i) pre-stagnation cluster, (ii) post-stagnation cluster 1 (the majority of post-stagnation samples from F1, F2, F4, F5, and F6), (iii) post-stagnation cluster 2 (the majority of post-stagnation samples from F3, F7, and F8), and (iv) winter cluster. A clustering analysis highlighted differences in microbial community structures pre-and poststagnation in winter. A pairwise test using ANOSIM also supported significant differences in microbial communities pre-and post-stagnation between winter and other seasons (R value=0.3981-1.0000, P<0.05) ( Table S8) .

The pre-stagnation cluster consisted of the most prestagnation samples in summer, autumn, and spring, in which Phreatobacter spp. were predominant with a relative abundance of 15-46%. The abundance of this genus was significantly lower in winter (1-3%). The shift in the microbial community after stagnation in F1, F2, F4, F5, and F6 in summer, autumn, and spring was characterized by a predominance of Sphingomonas spp. The microbial community after stagnation from these samples was clustered in poststagnation cluster 1. Phreatobacter spp. decreased in abundance from 27 to 8%, while Sphingomonas spp. increased from 5 to 22%. On the other hand, the shift in dominant bacteria in F3, F7, and F8 after stagnation in summer, autumn, and spring (post-stagnation cluster 2) was generally characterized by genera in Comamonadaceae, Dechloromonas spp., Cyanobacteria, and Porphyrobacter spp. Differences in the microbial communities of post-stagnation samples between these two groups of samples in these seasons were supported by ANOSIM (R value=0.9049, P<0.05).

Most pre-stagnation samples in winter were clustered in the winter cluster, and Pseudomonas spp. were dominant in these samples, with a relative abundance of 37-59%. After stagnation, Pseudomonas spp. were still dominant at several faucets (F1, F3, F4, and F5). F7 and F8 were grouped in post-stagnation cluster 2 after stagnation in winter, and the relative abundance of Pseudomonas spp. was higher (16-27%) than that of the other samples.

The abundance of opportunistic pathogens was quantified by real-time PCR. Clinically important opportunistic pathogens, including M. avium, L. pneumophila, P. aeruginosa, and Acanthamoeba spp., were below LOQ in all faucets before and after stagnation (LOQ of pre-and post-stagnation samples=5.0×10 -2 and 5.0×10 -1 gene copies mL -1 , respectively). On the other hand, Legionella spp. and Mycobacterium spp. significantly increased after stagnation in all seasons (Fig. 4) . qPCR revealed that Legionella spp. were detected in 62% (16/26) of samples collected from several faucets before stagnation. After stagnation, the detection frequency increased to 81% (21/26), and their abundance significantly increased from 9.6±5.0×10 -1 to 9.7±4.1×10 0 gene copies mL -1 (paired Student's t-test, P=8.281×10 -13 ). Mycobacterium spp. were positive for all samples before and after stagnation (52/52), and their abundance significantly increased from 2.5±0.6×10 0 to 8.8±3.0×10 gene copies mL -1 (paired Student's t-test; P=8.658×10 -13 ). The abundance of Legionella spp. and Mycobacterium spp. post-stagnation was lower in summer; however, these differences were not significant (ANOVA; P=0.844 and 0.720, respectively).

Post-stagnation samples with a high abundance of Legionella spp. and Mycobacterium spp. were subjected to nanopore sequencing targeting full-length 16S rRNA genes in order to obtain detailed phylogenetic information. On average, 60.7% of Legionella-related sequences were assigned to species in Legionella spp. (Fig. 5) . Sequences affiliated with pathogenic L. feeleii were the most frequently detected species among the Legionella groups, with an average relative abundance of 51.6% of Legionella-related sequences. However, the relative abundance of L. feeleii at F6 (summer and autumn) and F8 (winter) was less than that of other detected species, such as L. drozanskii. Other Legionella sequences, which were closely affiliated with pathogenic L. maceachernii Dynamics of Premise Plumbing Microbiome and L. micdadei, were also detected with an average relative abundance of 0.9 and 0.02% of all Legionella-related sequences, respectively. On the other hand, only 18.7% of Mycobacterium-related sequences were assigned to species in Mycobacterium spp. (Fig. 5 ). Sequences affiliated with pathogenic M. paragordonae were the most abundant species, with an average relative abundance of 13.4% of Mycobacterium-related sequences. Other Mycobacteriumrelated sequences affiliated with pathogenic M. gordonae and M. haemophilum were also detected, with an average relative abundance of 1.0 and 0.02% of Mycobacteriumrelated sequences, respectively.

A decrease in free chlorine and increase in TCC after stagnation were observed in all sampling events, which indicated that the regrowth phenomenon at a faucet was common in different faucets in a building. Previous studies demonstrated that the decay of free chlorine after stagnation triggered microbial regrowth (Ling et al., 2018; Dias et al., 2019; Zhang et al., 2021b) . The decay of free chlorine after overnight stagnation resulted in an increase in TCC up to 1.8×10 5 cells mL -1 (Zhang et al., 2021b) . The effect of stagnation on microbial regrowth was also reported in a non-chlorinated system in which increases in intact cell counts (1.1×10 5 cells mL -1 ) and HPC (8.7×10 2 CFU mL -1 ) were observed after overnight stagnation (Lautenschlager et al., 2010) . The higher temperatures of the samples after stagnation than those before stagnation may have accelerated microbial regrowth and the decay of free chlorine. A previous study reported that the chlorine decay rate was higher at 19.4°C than at 15.7°C (García-Ávila et al., 2020) . The increase in TCC after stagnation may be attributed to the detachment of bacteria from biofilms on the pipe wall (Manuel et al., 2010) . Small-diameter pipes at the distal ends of a building were shown to enhance the dispersal of bacteria from pipe-surface biofilms (Ling et al., 2018) . In addition, biodegradable organic matter contained in drinking water or migrating from the pipe material (e.g., polyethylene-coated stainless-steel) has been suggested to affect the regrowth of bacteria in drinking water (Rahmatika et al., 2020) . Seasonal variations in microbial regrowth were assessed in eight faucets (F1-F8) from which samples were collected in four different seasons. TCC levels after stagnation were not significantly different between seasons. Higher TCC in the first 100 mL were consistently found in all samples, which indicated that the last meter of premise plumbing was vulnerable to microbial regrowth.

Seasonal differences in microbial richness and diversity were observed in the present study, in which richness (observed OTUs) and the Shannon diversity index were significantly lower in winter. Previous studies also observed lower bacterial richness in winter and variations in the microbial community structure over time (Pinto et al., 2014; Ling et al., 2016) . These findings may be attributed to seasonal variations in the microbial community in the source water and growth conditions in premise plumbing.

Pre-stagnation samples in summer, autumn, and spring were dominated by Phreatobacter spp., which was consistent with previous findings showing the dominance of Phreatobacter spp. in a drinking water system (Perrin et al., 2019a; Van Assche et al., 2019) . The lower abundance of Phreatobacter spp. in winter than in other seasons was consistent with the finding that Phreatobacter spp. were more abundant in the warm period (water temperature >15°C) than in the cold period (water temperature <15°C) in the drinking water distribution system in Paris (Perrin et al., 2019a) .

A shift in dominant bacteria from Phreatobacter spp. to Sphingomonas spp. after stagnation was observed in F1, F2, F4, F5, and F6. A previous study reported that Sphingomonas spp. increased from 0 to 10% following the overnight stagnation of tap water (Zhang et al., 2021b) . Sphingomonas spp. are resistant to chlorine and may develop biofilms along the pipe wall (Sun et al., 2013; Douterelo et al., 2014; Liu et al., 2014; Chao et al., 2015) . Sphingomonas spp. in biofilms may grow rapidly once the water quality condition changes due to stagnation. Sphingomonas have been shown to secrete exopolysaccharides, which are the major components of biofilms (Johnsen et al., 2000; Gulati and Ghosh, 2017) . Although biofilms were not directly analyzed in the present study, they may play an important role in microbial regrowth in premise plumbing Peng et al., 2020) . F3, F7, and F8 after stagnation were dominated by Comamonadaceae, Dechloromonas spp., Cyanobacteria, and Porphyrobacter spp., which are commonly found in drinking water distribution systems (Williams et al., 2005; Revetta et al., 2011; Lautenschlager et al., 2013) . These results suggest that microbial communities after stagnation are site-specific and dependent on environmental factors, such as water usage frequency. On the other hand, pre-and post-stagnation samples in winter were dominated by Pseudomonas spp. The abundance of Pseudomonas spp. was higher in the cold period than in the warm period in a drinking water distribution system in Paris (Perrin et al., 2019a) . Therefore, Pseudomonas spp. have the potential to regrow under lowtemperature conditions.

Clinically important opportunistic pathogens, including M. avium, L. pneumophila, P. aeruginosa, and 

Acanthamoeba spp., were below LOQ in all faucets before and after stagnation. Although previous studies reported the presence of these opportunistic pathogens in premise plumbing, the present study revealed that clinically important opportunistic pathogens were below LOQ, regardless of repetitive stagnation events that occurred in the premise plumbing of the building (Wang et al., 2012; Dilger et al., 2018; Liu et al., 2019; Perrin et al., 2019b; Buse et al., 2020; Huang et al., 2021) . On the other hand, Legionella spp. and Mycobacterium spp. significantly increased after stagnation in all seasons. The abundance of Legionella spp. and Mycobacterium spp. post-stagnation was lower in summer; however, these differences were not significant. The abundance of Legionella spp. and Mycobacterium spp. was lower in the present study than in a previous study on household tap water in northern China, which reported the occurrence of Legionella spp. and Mycobacterium spp., with an average concentration of 6×10 2 and 4×10 2 gene copies mL -1 , respectively (Liu et al., 2019) . Despite the similar range of the free chlorine residual in the present study, different water characteristics and water usage practices may explain differences in the abundance of Legionella spp. and Mycobacterium spp. in premise plumbing. Previous studies reported the abundance of Legionella spp. in a hot water system (Leoni et al., 2005; Dilger et al., 2018) . Since high water temperatures up to 60°C are beneficial for their growth (Leoni et al., 2005) , future studies on hot water supply lines are warranted.

The abundance of Legionella spp. and Mycobacterium spp. was significantly higher in the first draw than in postflushing samples (Wang et al., 2012) . A positive correlation between Legionella spp. and stagnation times was also previously reported (Ley et al., 2020) . These findings suggest that Legionella spp. and Mycobacterium spp. increased after stagnation, and, thus, flushing prior to water consumption may reduce the exposure risk from these genera containing opportunistic pathogens, particularly during the COVID-19 pandemic where the occupancy of buildings has markedly decreased (Hozalski et al., 2020) . The increase observed in the abundance of Legionella spp. and Mycobacterium spp. after stagnation may be associated with biofilms due to their slow growth rates and tendency to form biofilms . Mycobacterium spp. have a hydrophobic cell wall, which protects them from disinfectants and promotes surface attachment and biofilm formation (Steed and Falkinham, 2006) . Another important factor influencing the occurrence of Mycobacterium spp. is their resistance to the disinfectant. In addition, Legionella spp. and Mycobacterium spp. both survive and live inside freeliving amoeba, such as Acanthamoeba spp., which protects them from stress and promotes their growth (Declerck et al., 2005; Lau and Ashbolt, 2009; Liu et al., 2019) . Acanthamoeba spp. were not detected in the present study, which may have been due to their presence in biofilms.

Other hosts, such as Vermamoeba spp., need to be investigated in future studies.

Species other than L. pneumophila and M. avium are also human pathogens (e.g., L. feeleii, L. micdadei, M. kansasii, and M. abscessus) (Thomas et al., 1992; Mogami et al., 2016; Sfeir et al., 2018; Wang et al., 2019) . Therefore, nanopore sequencing targeting full-length 16S rRNA genes was applied to obtain detailed phylogenetic information. In the present study, sequences affiliated with pathogenic L. feeleii were the most frequently detected among Legionellarelated sequences. The difference observed in the relative abundance of L. feeleii among faucets indicated that the composition of Legionella spp. was specific at different faucets, even in the same building. L. feeleii causes Legionnaires' disease or Pontiac fever . It has been detected in several water systems in various countries, such as a hot spring in Taiwan, hot tap water in Germany, and river water being used as a source of drinking water in Japan (Hsu et al., 2006; Dilger et al., 2018; Edagawa et al., 2019) . Other Legionella sequences affiliated with pathogenic L. maceachernii and L. micdadei were also detected. L. maceachernii, which was initially isolated from a portable water system, caused pneumonia in an immunocompromised patient, while L. micdadei caused Legionnaires' disease or Pontiac fever (Thomas et al., 1992; .

The percentage of Mycobacterium-related sequences assigned to species was low in the present study. This may be because 16S rRNA genes among different Mycobacterium species were very similar (van der Wielen et al., 2013) . Therefore, the more detailed characterization of Mycobacterium in premise plumbing is needed. Mycobacterium communities may be characterized by analyzing hsp65 and rpoB genes, which demonstrate better resolution in discriminating between species and sub-species (van der Wielen et al., 2013; Haig et al., 2018) . Several Mycobacterium sequences affiliated with pathogenic M. paragordonae, M. gordonae, and M. haemophilum were detected. M. paragordonae was reported to cause peritonitis in a peritoneal dialysis patient, while M. gordonae and M. haemophilum caused pulmonary and skin infections in elderly immunocompetent patients, respectively (September et al., 2004; Lindeboom et al., 2011; Cheung et al., 2017) . Even though clinically important L. pneumophila and M. avium were not detected in the present study, the potential risks of other sequences that are closely related to pathogenic Legionella and Mycobacterium need to be intensively monitored.

Microbial water quality was monitored for one year before and after stagnation in different faucets of the premise plumbing in a building. The main conclusions of the present study are as follows: (1) The depletion of free chlorine was observed in all faucets after stagnation, which resulted in a significant increase in TCC from 10 3 to 10 4 -10 5 cells mL -1 .

(2) Distinct changes in community structures were observed before and after stagnation because the dominant species changed from Phreatobacter spp. to Sphingomonas spp., Comamonadaceae, Dechloromonas spp., Cyanobacteria, and Porphyrobacter spp. (3) Seasonal variations appeared to play a role in shaping the composition of the microbial community in drinking water, particularly in winter when Pseudomonas spp. were dominant. (4) Although clinically important opportunistic pathogens were not detected, several sequences, which were closely related to pathogenic Legionella spp. and Mycobacterium spp., were identified. Further guidelines regarding water usage practices in premise plumbing are essential for maintaining the good microbial water quality of drinking water prior to consumption.

The hospital water supply as a source of nosocomial infections: A plea for action

Standard Methods for the Examination of Water and Wastewater

Impact of stagnation and sampling volume on water microbial quality monitoring in large buildings

Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2

Legionella diversity and spatiotemporal variation in the occurrence of opportunistic pathogens within a large building water system

Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample

Diversity and functions of bacterial community in drinking water biofilms revealed by high-throughput sequencing

Mycobacterium paragordonae: a rare cause of peritonitis in a peritoneal dialysis patient

Impact of non-Legionella bacteria on the uptake and intracellular replication of Legionella pneumophila in Acanthamoeba castellanii and Naegleria lovaniensis

Identification of factors affecting bacterial abundance and community structures in a full-scale chlorinated drinking water distribution system

Legionella contamination in warm water systems: A species-level survey

Bacterial community dynamics during the early stages of biofilm formation in a chlorinated experimental drinking water distribution system: implications for drinking water discolouration

Investigations on contamination of environmental water samples by Legionella using real-time quantitative PCR combined with amoebic co-culturing

Mycobacterium avium in a shower linked to pulmonary disease

Opportunistic pathogens enriched in showerhead biofilms

Relationship between chlorine decay and temperature in the drinking water

Propidium monoazide-quantitative polymerase chain reaction for viable Escherichia coli and Pseudomonas aeruginosa detection from abundant background microflora

Biofilm forming ability of Sphingomonas paucimobilis isolated from community drinking water systems on plumbing materials used in water distribution

A high-throughput approach for identification of nontuberculous mycobacteria in drinking water reveals relationship between water age and Mycobacterium avium

Real-time PCR assay targets the 23S-5S spacer for direct detection and differentiation of Legionella spp. and Legionella pneumophila

The effect of the surface-tovolume contact ratio on the biomass production potential of the pipe products in contact with drinking water

Flushing of stagnant premise water systems after the COVID-19 shutdown can reduce infection risk by Legionella and Mycobacterium spp

Legionella prevalence in hot spring recreation areas of Taiwan

Opportunistic pathogens and their health risk in four full-scale drinking water treatment and distribution systems

Evaluation of fluorescently labeled lectins for noninvasive localization of extracellular polymeric substances in Sphingomonas biofilms

The role of biofilms and protozoa in Legionella pathogenesis: implications for drinking water

Overnight stagnation of drinking water in household taps induces microbial growth and changes in community composition

A microbiology-based multiparametric approach towards assessing biological stability in drinking water distribution networks

Legionella waterline colonization: detection of Legionella species in domestic, hotel and hospital hot water systems

Drinking water microbiology in a water-efficient building: stagnation, seasonality, and physicochemical effects on opportunistic pathogen and total bacteria proliferation

Clinical manifestations, diagnosis, and treatment of Mycobacterium haemophilum infections

Core-satellite populations and seasonality of water meter biofilms in a metropolitan drinking water distribution system

Drinking water microbiome assembly induced by water stagnation

Pyrosequencing reveals bacterial communities in unchlorinated drinking water distribution system: an integral study of bulk water, suspended solids, loose deposits, and pipe wall biofilm

Assessing the origin of bacteria in tap water and distribution system in an unchlorinated drinking water system by SourceTracker using microbial community fingerprints

One-year survey of opportunistic premise plumbing pathogens and free-living amoebae in the tap-water of one northern city of China

Consecutive ultrafiltration and silica adsorption for recovery of extracellular antibiotic resistance genes from an urban river

Unsteady state flow and stagnation in distribution systems affect the biological stability of drinking water

Pulmonary infection caused by Mycobacterium kansasii: findings on computed tomography of the chest

Assessing the contribution of biofilm to bacterial growth during stagnation in shower hoses

Microbiome of drinking water: A full-scale spatio-temporal study to monitor water quality in the Paris distribution system

Spatiotemporal survey of opportunistic premise plumbing pathogens in the Paris drinking water distribution system

Spatial-temporal survey and occupancy-abundance modeling to predict bacterial community dynamics in the drinking water microbiome

Biological stability of drinking water: Controlling factors, methods, and challenges

Drinking water microbiologyfrom measurement to management

Impacts of organic matter migrating from pipe materials on microbial regrowth in drinking water

16S rRNA gene sequence analysis of drinking water using RNA and DNA extracts as targets for clone library development

Growth of Legionella during COVID-19 lockdown stagnation

Development of a real-time PCR assay for quantification of Acanthamoeba trophozoites and cysts

Diversity of nontuberculoid Mycobacterium species in biofilms of urban and semiurban drinking water distribution systems

Mycobacterium abscessus complex infections: A retrospective cohort study

Effect of growth in biofilms on chlorine susceptibility of Mycobacterium avium and Mycobacterium intracellulare

Characterization and identification of a chlorine-resistant bacterium, Sphingomonas TS001, from a model drinking water distribution system

Fatal Legionella maceachernii pneumonia in Canada

Characterization of the bacterial community composition in water of drinking water production and distribution systems in Flanders, Belgium. MicrobiologyOpen 8: e00726. van der Wielen

Legionella feeleii: pneumonia or Pontiac fever? Bacterial virulence traits and host immune response

Molecular survey of the occurrence of Legionella spp., Mycobacterium spp., Pseudomonas aeruginosa, and amoeba hosts in two chloraminated drinking water distribution systems

Probiotic approach to pathogen control in premise plumbing systems? A review

BioMig-A method to evaluate the potential release of compounds from and the formation of biofilms on polymeric materials in contact with drinking water

Population diversity in model potable water biofilms receiving chlorine or chloramine residual

Introducing EzBioCloud: a taxonomically united database of 16S rRNA gene sequences and whole-genome assemblies

Indoor heating triggers bacterial ecological links with tap water stagnation during winter: Novel insights into bacterial abundance, community metabolic activity and interactions

Combined effects of seasonality and stagnation on tap water quality: Changes in chemical parameters, metabolic activity and co-existence in bacterial community

An experimental study on the influence of water stagnation and temperature change on water quality in a full-scale domestic drinking water system

This work was partly supported by JSPS KAKENHI (Grant Numbers 17H04940 and 20H02282).