Occurrence of pathogenic Mycobacteria avium and Pseudomonas aeruginosa in outdoor decorative fountain water and the associated microbial community

Outdoor decorative fountains are a fantastic landscape in urban areas that attract residents to visit the jetting show (Cheng et al. 2021). Generally, decorative fountain waters are supplied by natural lakes/rivers, reclaimed, and municipal taps, which have been reported to contain opportunistic pathogens (OPs) (Whiley et al. 2015; Fang et al. 2018; Huang et al. 2021). Generally, waterborne OPs could slowly grow under oligotrophic and stagnant conditions, or can even ‘hide’ in an amoeba host; moreover, they could resist heat, disinfectants, and antibiotics (Djouadi et al. 2017; Donohue 2021). Outdoor decorative fountain water is usually recycled in the fountain basin during daily operations, resulting in stagnant conditions that benefit OPs proliferation. After the atomization of fountain water from the spraying nozzle, aerosols are generated, and the specific OPs associated with fine droplets can be easily inhaled by persons, posing potential infection risks to people visiting the decorative fountain. Outbreaks of giardiasis, cryptosporidiosis, and Legionnaires disease have been reported in the US and European countries, associated with exposure to interactive and/or decorative fountain water (O'Loughlin et al. 2007; Eisenstein et al. 2008; Haupt et al. 2012). Therefore, outdoor decorative fountain water poses higher health risks than the corresponding water source.

Legionella pneumophila, Mycobacterium avium, and Pseudomonas aeruginosa are common OPs in stagnant waters, causing thousands of waterborne disease cases worldwide (Garner et al. 2018). Thus, the monitoring of OPs in outdoor decorative fountains is vital to track potential health risks. However, the standard plate counting method is challenged by inaccurate quantification because only a small fraction (less than 0.001%) of viable OPs are cultivable (Niculita-Hirzel et al. 2022). This may underestimate viable but non-culturable (VBNC) cells. In addition, the culture of pathogenic OPs is a time- and labor-consuming test that requires long incubation times (up to 15 days), professional equipment, and strict protocols (Eble et al. 2021). Culturable fecal bacteria are usually considered pathogenic indicators of the water environment. However, no correlation was observed between OPs abundance and the enumeration of these fecal indicators (Wang et al. 2018). As a result, the surrogate pathogenic indicators, such as total coliform, could not indicate the real pathogenic state exposed by the selected OPs in the decorative fountain water. Molecular monitoring approaches, such as quantitative polymerase chain reaction (qPCR), are other frequently used methods that enable the determination of gene copy numbers of specific OPs regardless of the VBNC state (Hamilton et al. 2018). Although qPCR measurements are unable to distinguish living microorganisms from non-viable cells, tracking the gene copy number variation of the selected OPs in the water environment is helpful in examining the potential microbial risks. Previous studies have reported that biomolecular approaches of qPCR and 16S ribosomal ribonucleic acid (rRNA) sequencing are advantageous for assessing the pathogenic conditions in recreational water, drinking water, natural water, and reclaimed water (El-Sayed et al. 2019; Mapili et al. 2022). However, no study has reported the occurrence and distribution of OPs in outdoor decorative fountain waters.

In this study, three large outdoor decorative fountains with different water sources in Hangzhou, China, were selected for a 1-year survey. Molecular monitoring approaches of qPCR and 16S rRNA sequencing were combined to investigate the distribution of common OPs (L. pneumophila, M. avium, and P. aeruginosa), microbial communities, and their functional potential associated with the outdoor decorative fountain waters. Furthermore, seasonal changes in the selected OPs and their correlations with physicochemical parameters were assessed. This 1-year molecular survey of the selected OPs provided insights into the potential risks posed by OPs in decorative fountain water aerosols.

The three selected decorative fountains were located on a university campus (#1), a natural lake (#2), and a residential community square (#3). Fountains #1, #2, and #3 were supplied with municipal taps, natural lakes, and reclaimed water, respectively. For Fountain #1, the plane shape was a rectangle measuring 30 m in length and 4 m in width, holding a total water volume of 120 m³; it attracted 1,000 daily visitors, consisting mainly of faculty and students. The water for Fountain #2 was sourced from the lake with a storage capacity of 14.3 million m³; the installed spray nozzle covered a length of 126 m and a width of 2 m, attracting more than 10,000 sightseers to visit daily. As for Fountain #3, the plane shape was a circle with a diameter of 40 m, holding a total water volume of 1,000 m³; it welcomed 500 residents daily.

Sampling was conducted in August 2019, November 2019, January 2020, and April 2021, representing the summer (S), autumn (A), winter (W), and spring (P) samples, respectively. Due to the COVID-19 pandemic in early 2020, the daily jetting shows of the decorative fountains were closed. Thus, the expected sampling campaign in April 2020 was suspended and reconducted in April 2021. However, Fountain #3 was destroyed at that time because of the refurbishment of the area. Consequently, only three sampling campaigns were conducted for Fountain #3. For the definition of water samples, numbers 1, 2, and 3 represent different decorative fountains in the university campus, natural lake, and community square, respectively, and capital letters S, A, W, and P represent different seasons of summer, autumn, winter, and spring, respectively. For example, sample ‘1_S’ meant the water sample from Fountain #1 in summer, and so on for other samples.

The stagnant conditions during the idle operation stage could lead to inhomogeneities of the fountain water. To gain a better understanding of the water quality of outdoor decorative fountain water, 1 L of sample was collected before, during, and after the jetting show of each fountain using sterile sampling bags. The physicochemical parameters of dissolved oxygen (DO), oxidation–reduction potential (ORP), pH, and water temperature were measured in situ. Subsequently, each water sample from the three collections was homogeneously mixed and sent to the laboratory for testing other parameters. Sampling, storage, and transportation procedures were conducted according to the Chinese National Standard Test Method for drinking water (GB/T 5750-2006).

Physicochemical parameters of pH, water temperature, DO, and ORP were determined using portable multiple meters (F2 Standard, Mettler Toledo, Switzerland) and a pocket DO meter (SANXIN, SX751, Shanghai, China), respectively. ⁠, ⁠, ⁠, total nitrogen (TN), total phosphorus (TP), and culturable total coliforms were measured in the laboratory according to Standard Methods (APHA/AWWA/WEF 2005). Free residual chlorine was quantified by the N,N-diethyl-p-phenylenediamine colorimetric method using a portable residual chlorine tester (Shunkeda Technology Co., Ltd, Beijing, China).

To enrich the total bacteria, 1 L of the sampling water was filtrated through a 0.22-μm-pore-size sterile mixed cellulose ester (MCE) membrane with a vacuum pump in a vertical clean bench under sterile conditions. As a result, most of the waterborne bacteria (>96%) could be intercepted by the MCE membrane (Harb et al. 2021). The MCE membrane was then cut into small strips for DNA extraction using a DNA Assay Kit (Invitrogen, USA) according to the manufacturer's instructions. The PCR process was performed to amplify the 16S rRNA gene in the bacterial V3–V4 region, with a forward primer of 338F (5-ACTCCTACGGGAGGCAGCA-3) and a reverse primer of 806R (5-TCGGACTACHVGGGTWTCTAAT-3), respectively. The amplification process was conducted using a thermal cycler (Applied Biosystems 2720) with the following program: initial denaturation at 98 °C for 2 min, 30 cycles of denaturation at 98 °C for 15 s, annealing at 55 °C for 30 s, extension at 72 °C for 30 s, and a final extension at 72 °C for 5 min. The DNA was quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Technology, USA) and examined by gel electrophoresis. The concentrations of the recovered DNA for all amplicons were in the range of 4.4–17.1 ng/μL, meeting the minimal requirement (0.5 ng/μL) for the subsequent sequencing.

The purified amplicons were then sent to Shanghai Personal Biotechnology Co., Ltd (Shanghai, China) for 16S rRNA sequencing using the Illumina^® MiSeq 2 × 300 bp platform. The sequencing results were then clustered into operational taxonomic units (OTU) at a 97% similarity level using QIIME software. Indicators of Chao1, Shannon–Wiener, and Simpson indices were calculated to evaluate the bacterial richness and community diversity associated with various fountain water samples. Taxonomic classifications at the phylum and genus levels were also constructed to compare the microbial communities. The functional genes and metabolic pathways of the bacterial communities were predicted using the PICRUSts software (Langille et al. 2013), and the specific steps for the analysis were provided by the online analysis platform (https://www.genescloud.cn).

The gene copy numbers of the three selected OPs, L. pneumophila, M. avium, and P. aeruginosa, were quantified by qPCR using a TIB-8600 qPCR system (Triplex International Biosciences, China). The primers and amplification programs used for the different OPs are listed in Table 1. For the TaqMan assay, the reaction system (16 μL) contained 8 μL mixture A and 8 μL template DNA. Mixture A contained 10 μL of 2 × SYBR real-time PCR premixture (Q112-02 Vazyme), 0.4 μL of 10 μM forward primer, and 0.4 μL of 10 μM reverse primer. Each qPCR assay was performed in triplicates. For each qPCR run, a series of 10-fold diluted plasmid standard curves were established (R² ≥ 0.99), and melt curves were constructed to guarantee the specificity of the primers for amplifying the target OPs genes. The limits of quantification (LOQ) for the selected OPs were below 67 gene copies/reaction.

One-way analysis of variance was used to compare the difference in log₁₀-transformed qPCR results from different water samples, and P < 0.05 was considered statistically significant. To reveal the relationships between OPs abundance and environmental variables (pH, DO, ORP, water temperature, ⁠, ⁠, ⁠, TN, and TP), redundancy analysis (RDA) was conducted with the CANOCO 5 program. RDA, using multivariate regression, can extract variations in response variables that are explained by a set of explanatory variables. This statistical analysis provides scientists with an efficient way to identify linear combinations of environmental variables that can account for the linear combinations of the concerned microbial populations (Capblancq & Forester 2021). R software (Version 4.0.2) was used to conduct a correlative heatmap analysis between different water samples.

Physicochemical parameters and total coliforms of water samples from different outdoor decorative fountains are shown in Table 2. The results indicated that most physicochemical parameters satisfied the water quality standards for scenic environment use (GB 12941-91 or GB/T 18921-2019, China). However, the positive rates of total coliforms in the three decorative fountain water samples were 25, 100, and 67%, respectively. During the summer season, total coliforms were detected in all the samples from the three decorative fountains (1_S, 2_S, and 3_S), indicating the possible fecal contamination of the fountain water and the potential pathogenic condition in the decorative fountain waters. Thus, potential health risks would be raised for people who were exposed to the fountain water aerosols, especially in summer when the jetting shows were frequent. Furthermore, the concentrations of TN and TP in Fountain #2 (sourced by the natural lake) were 1.8–7.8, and 0.03–0.05 mg/L, respectively, which were consistent with the ranges reported in the previous investigation (Bai et al. 2020). These nutrient levels exceed the threshold concentrations for triggering the potential eutrophication (Liu et al. 2022), as indicated by the reported chlorophyll-a concentrations in this natural lake ranging from 5.9 to 16 mg/L in spring, 11.8 to 39.6 mg/L in summer, 4.3 to 9.0 mg/L in autumn, and 2.1 to 7.9 mg/L in winter, respectively (Bai et al. 2020). Under these conditions, the release of organic nitrogen from algal cells would contribute to the imbalance between TN and inorganic nitrogen species (⁠⁠, ⁠, ⁠). Thus, although the outdoor decorative fountain water was fed with different water sources, eutrophication risks were present, which in turn favored the occurrence of pathogenic microorganisms such as Vibrio spp. in the recreational water environment (Canellas et al. 2021).

Table 2

Water quality parameters of various decorative fountain water samples

Samples .	Criteria values .	1_S .	1_A .	1_W .	1_P .	2_S .	2_A .	2_W .	2_P .	3_S .	3_A .	3_W .
Total coliform (CFU/mL)	/	1	0	0	3	9	30	4	8	1	0	6
Water temperature (°C)	/	27.2	18.4	11.6	23.5	26.4	18.5	12.0	23.8	27.0	18.3	12.0
pH	6.5 ∼ 8.5	8.38	8.27	7.05	8.70	7.90	6.87	7.19	7.45	9.25	8.45	7.73
ORP (mV)	/	248	263	237	165	222	279	262	188	234	244	240
DO (mg/L)	≥3	7.13	5.45	8.67	9.70	6.11	9.24	8.57	8.10	8.27	9.38	9.55
(mg/L)	≤0.5	0.20	0.49	0.17	0.15	0.45	0.50	0.31	0.26	0.22	0.32	0.55
(mg/L)	/	0.20	0.09	0.11	0.00	0.07	0.10	0.14	0.00	0.09	0.08	0.08
(mg/L)	≤1.0	0.01	0.01	0.01	0.02	0.01	0.01	0.01	0.02	0.01	0.00	0.00
TN (mg/L)	≤10	8.10	6.70	7.01	0.47	7.10	7.00	7.80	1.80	6.90	7.00	7.40
TP (mg/L)	≤0.05	0.04	0.03	0.02	0.03	0.03	0.03	0.03	0.05	0.03	0.07	0.06

Samples .	Criteria values .	1_S .	1_A .	1_W .	1_P .	2_S .	2_A .	2_W .	2_P .	3_S .	3_A .	3_W .
Total coliform (CFU/mL)	/	1	0	0	3	9	30	4	8	1	0	6
Water temperature (°C)	/	27.2	18.4	11.6	23.5	26.4	18.5	12.0	23.8	27.0	18.3	12.0
pH	6.5 ∼ 8.5	8.38	8.27	7.05	8.70	7.90	6.87	7.19	7.45	9.25	8.45	7.73
ORP (mV)	/	248	263	237	165	222	279	262	188	234	244	240
DO (mg/L)	≥3	7.13	5.45	8.67	9.70	6.11	9.24	8.57	8.10	8.27	9.38	9.55
(mg/L)	≤0.5	0.20	0.49	0.17	0.15	0.45	0.50	0.31	0.26	0.22	0.32	0.55
(mg/L)	/	0.20	0.09	0.11	0.00	0.07	0.10	0.14	0.00	0.09	0.08	0.08
(mg/L)	≤1.0	0.01	0.01	0.01	0.02	0.01	0.01	0.01	0.02	0.01	0.00	0.00
TN (mg/L)	≤10	8.10	6.70	7.01	0.47	7.10	7.00	7.80	1.80	6.90	7.00	7.40
TP (mg/L)	≤0.05	0.04	0.03	0.02	0.03	0.03	0.03	0.03	0.05	0.03	0.07	0.06

Note: CFU is the abbreviation of colony forming units. The data were the average results of triplicate detection. For the definition of water samples, the numbers of 1, 2, and 3 represented different decorative fountains sourced by municipal tap water (Fountain #1), natural lake water (Fountain #2), and reclaimed water (Fountain #3), respectively; and capital letters S, A, W, and P represented different seasons of the summer, autumn, winter, and spring, respectively. For example, sample ‘1_S’ meant the water sample from Fountain #1 in summer. Except for TN, which is derived from the Chinese national standard of GB/T 18921-2019, the criteria values for other water quality parameters were all extracted from the Chinese national standard of GB 12941-91.

View Large

M. avium and P. aeruginosa were the common OPs in urban recreational water (Fang et al. 2018), which were tested positive in all outdoor decorative fountains and seasons in this study. The log₁₀-transformed gene copy numbers of the two OPs in various fountain waters are shown in Figure 1(b) and 1(c). Generally, the M. avium abundance was from 2.4 to 3.9 log₁₀(gene copies/mL), which fluctuated little with seasons except for spring (1_P, 2_P). However, the P. aeruginosa abundance showed significant differences between various seasons (P < 0.05) and was two to three orders of magnitude higher than that of M. avium. In summer, the gene copy numbers of P. aeruginosa in the three decorative fountain water samples ranged from 5.0 to 5.3 log₁₀(gene copies/mL), which sharply increased during the autumn and winter seasons but declined during the spring season. The maximum gene copy numbers of P. aeruginosa in Fountains #1, #2, and #3 were 6.5, 7.1, and 6.8 log₁₀(gene copies/mL), respectively. The relatively lower abundance observed in Fountain #1 could be attributed to the drinking water source, in which the residual chlorine concentrations (0.05–0.11 mg/L) were significantly higher than those in Fountain #2 and Fountain #3 (<0.01 mg/L), thereby impeding the growth of P. aeruginosa (Waak et al. 2019). These results were consistent with previous studies, indicating that the DNA marker abundance of P. aeruginosa in environmental water (rainwater) was water quality-dependent, and was higher than M. avium (Zhang et al. 2021). P. aeruginosa can cause infections of open wounds and eyelids (Liu et al. 2019), implying higher health risks to injured persons visiting outdoor decorative fountains.

In recreational water, OPs occurrence and their abundance were usually reported to be temperature-dependent, and higher water temperatures would benefit the proliferation of specific microorganisms (Inkinen et al. 2016). However, in this study, the obtuse angles between the temperature arrow line and those of both OPs arrow lines were observed, suggesting that OPs abundance in the outdoor decorative fountain water was negatively correlated with water temperature (Figure 2), which was consistent with a previous study (Bland et al. 2005). The gene copy numbers of M. avium and P. aeruginosa from cold winter samples (11.6–12.0 °C) were significantly higher than those from hot summer samples (26.4–27.2 °C) and warm spring samples (23.5 –23.8 °C) (P < 0.05). Other studies also indicated that higher or warm temperatures would limit the growth of OPs in the bacterial complex, choosing for the enrichment of common microorganisms (Norton et al. 2004).

At the genus level, the top 20 genera and their relative abundances in different decorative fountain water samples are shown in Figure 4(b). The results indicate that the dominant genera associated with Fountain #1 changed seasonally. In summer, they were Deinococcus (74.88%) and changed to Bacillus (35.91%) in autumn, Deinococcus (40.31%) and Pseudomonas (10.79%) in winter, and Sporichthyaceae (30.81%) in spring. For Fountain #2, the dominant genera were Deinococcus (23.68–56.29%) in summer and autumn, Flavobacterium (68.98%) and Pseudomonas (14.38%) in winter, and hgcI_clade (40.54%) in spring. As for Fountain #3, Exiguobacterium (9.04–33.86%) and Candidatus aquiluna (11.57–14.62%) were the most common genera in the summer, autumn, and winter seasons; other dominant genera in summer were Bacillus (12.22%), in autumn were Acinetobacter (26.01%), and in winter were Flavobacterium (9.58%). Many pathogenic genera, such as Bacillus, Pseudomonas, Flavobacterium, and Acinetobacter, were detected in the three outdoor decorative fountain waters. Notably, Pseudomonas was detected in all water samples, with a relative abundance ranging from 0.03% (summer) to 14.38% (winter). This result was consistent with the qPCR results, indicating a higher abundance of DNA markers of P. aeruginosa in cold seasons (Figure 1). Many species associated with the Pseudomonas genus cause purulent infections in people with weakened immunity or those who use antibiotics for a long time (Shevelev et al. 2020). Interestingly, the relative abundance of the genus hgcI_clade in spring samples (13.93–40.54%) was much higher than in other seasons, but the gene copy numbers of M. avium and P. aeruginosa were significantly lower (P < 0.05). Table 2 shows much lower concentrations of TN, ⁠, and in spring samples (1_P, 2_P) than in others, which would induce the propagation of hgcI_clade because the abundance of the hgcI_clade was negatively correlated with nutrient concentrations (Ruprecht et al. 2021).

Notably, the KEGG-annotated functional potentials (human diseases) included ko05110 V. cholerae infection, ko05111 V. cholerae pathogenic cycle, ko05130 pathogenic E. coli infection, and ko05146 amoebiasis, which have also been reported as pathogenic functional genes in aquatic environments (Baker-Austin et al. 2018; VanMensel et al. 2022). Although these infectious disease pathways accounted for a small percentage (<0.3%) of community functional compositions, they were predicted from all water samples regardless of the water source, possibly related to the wide occurrence of OPs, including M. avium and P. aeruginosa in this study. Moreover, the relative abundance of these infectious disease pathways in the spring samples was much lower than that in other seasons, which was consistent with the qPCR results of the two detected OPs. Thus, the pathogenic conditions associated with the decorative fountain water in spring would be lower, indicating more concerns about infection risks in other seasons.

Outdoor decorative fountains are a recreational landscape in urban areas that attracts a large number of people, especially the elderly and children. During the spraying of fountain water, aerosols are generated, which can be inhaled by visitors. This study showed that decorative fountain water, sourced from municipal taps, natural lakes, or reclaimed water, had a high abundance of OPs and other pathogenic microorganisms. Therefore, people who visit decorative fountain waters should pay critical attention to the potential health risks caused by OPs. However, it was difficult to accurately predict the OPs infection probability due to the following issues: First, most OPs were in the VBNC state, resulting in inaccurate quantitation of viable bacteria using neither the standard plate counting method nor the qPCR test. In future work, ethidium monoazide (EMA)/propidium monoazide (PMA)-qPCR could be applied to selectively quantify living OPs in decorative fountain waters, because DNA molecules from dead bacteria would be modified by intercalating the dye EMA or PMA, preventing further amplification (Hamilton et al. 2018). Moreover, the sampling of OPs in the aerosols during the jetting shows and the related computational fluid dynamics (CFD) modeling for aerosol droplets should also be investigated because the CFD model is a fast and reliable method to simulate trace aerosol droplets (Sheikhnejad et al. 2022). By exploring the above exposure parameters in different scenarios, quantitative microbial risk assessment would be highly suggested to predict the infection probability of specific OPs in decorative fountain water (Jorgensen et al. 2022).

A molecular survey of the selected OPs in three large outdoor decorative fountain waters showed that M. avium and P. aeruginosa occurred widely, and their abundance was dependent on the season and physicochemical parameters. The abundance of OPs was negatively correlated with water temperature, but positively correlated with nutrients (⁠⁠, ⁠, TN, and TP). The control of eutrophication in the aquatic environment potentially reduces OPs abundance. M. avium abundance in different fountain water samples ranged from 2.4 to 3.9 log₁₀(gene copies/mL), which was several orders of magnitude lower than P. aeruginosa (6.5–7.1 log₁₀(gene copies/mL)) in winter. Microbial community analysis and functional potential prediction also revealed that pathogenic microorganisms and infectious disease metabolic pathways were associated with decorative fountain water samples. This study provided insights into the potential health risks posed by OPs in outdoor decorative fountain waters.

This work was funded by the Foundation of Key Laboratory of Yangtze River Water Environment, Ministry of Education (Tongji University), China (No. YRWEF202203), and the Zhejiang Provincial Natural Science Foundation of China (LY23E080007), and the Key Research and Development Program of Zhejiang Province, China (2023C03134).

There are no ethical issues involved in this article and no harm will be caused to individual organisms. This entry does not apply to this article.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.