Fountain Water as a Source of Opportunistic Escherichia coli and Aeromonas hydrophila Strains in Atmospheric Air

The aim of this study was to determine the effect of fountain water on the contamination of atmospheric air with opportunistic Escherichia coli and Aeromonas hydrophila, and to evaluate the resulting health risks to humans. The counts of bacteria, their resistance to eight antibiotics, the multiple antibiotic resistance (MAR) index, and the minimum inhibitory concentrations (MICs) of ciprofloxacin (CIP) and tetracycline (TE) were determined in samples of air collected in the immediate vicinity of fountains (A 0 ) and at a distance of 20 m (A 20 ), and in samples of water (W) collected from five fountains. The counts of E. coli and A. hydrophila were highest in bioaerosols (A 0 ), and they were 10–100 times lower in A 20 samples. Most E. coli and A. hydrophila strains isolated from air samples were resistant to CIP, TE, ampicillin (AMP), and trimethoprim/sulfamethoxazole (SXT), and 6.7–46.2% of the isolates exhibited multidrug resistance. The values of the MAR index were high in all air samples and similar to those noted in fountain water isolates (0.86–0.87). In total, 17–75% of E. coli and A. hydrophila strains isolated from A 0 and A 20 samples were resistant to high doses of CIP and TE. Agglomerative clustering revealed five clusters of strains isolated from atmospheric air (A 0 , A 20 ) and fountain water. The results indicate that fountains can be a source of atmospheric air contamination with potentially pathogenic E. coli and A. hydrophila bacteria.


INTRODUCTION
Fountains are ornamental architectural features that enhance and enrich the landscape of cities and residential estates.These hydrotechnical structures are often closed-circuit systems with filtering or disinfecting devices.Due to the recreational character of fountains and the absence of physical barriers, wild birds, animals, pets and humans come into direct contact with these water spots.During continuous or simultaneous operation of water fountains, the microbiota present in water droplets are ejected by the nozzles into atmospheric air.Aerated water droplets return to the fountain basin, thus contributing to aerosolization and two-way transport of microbial contaminants such as fecal bacteria.As a result, bacteria colonizing fountain water enrich the pool of airborne microbiota and pose an indirect threat to human health in the vicinity of fountains (Flores et al., 2013).According to Mescioglu et al. (2019), airborne transmission may occur relatively easily in environments with a concentrated source of bacterial contaminants, which increases the probability of infection at a large distance from the source.
To fill in the gaps in knowledge about the influence of municipal fountains on the microbiological safety of the surrounding air, the present study was undertaken to examine the prevalence of antibiotic-resistant, potentially pathogenic bacteria, Escherichia coli and Aeromonas hydrophila, which may pose a threat to human health according to the World Health Organization (WHO, 2017).Some E. coli and A. hydrophila strains colonizing water, soil, and atmospheric air can also cause Table 1.Characteristics of the studied fountains.

No
Location Fountain characteristics and remarks I Central Park Spouting fountain; surface area -315.0 m 2 , concrete basin and bottom, basin depth -up to 0.6 m, maximum jet height -4.0 m; used as a water source by birds and wildlife; II Old Town Spouting fountain; surface area -26 m 2 , granite basin and bottom, basin depth -up to 0.5 m, maximum jet height -2.0 m; used as a water source by pigeons; III Lower Town Park Spouting fountain; surface area -286.4 m 2 , concrete basin and bottom, basin depth -up to 0.5 m, maximum jet height -3.0 m; used as a water source by birds and wildlife; IV Lower Town Park Spouting fountain; surface area -167.2 m 2 , concrete basin and bottom, basin depth -up to 0.5 m, maximum jet height -2.5 m; used as a water source by birds and wildlife; V Town Hall Cascading fountain; surface area -35.6 m 2 , granite stairs with a height of 0.25 m; used as a water source by birds.(https://www.google.com/maps/@53.7766301,20.4751515,609m/data=!3m1!1e3?hl=pl) has a diameter of 1 mm and a surface area of 0.785 mm 2 .The sampler has a flow rate of 100 L air min -1 , and air intake in the sampler can be controlled in the range of 1.0-1,000 L. The impaction speed of airborne microorganisms onto the agar surface was 11 m s -1 .The MAS-100 Eco sampler has an air flow rate of 11 m s -1 which corresponds to level 5 of the Andersen sampler (Pascual et al., 2003).The sampler supports the detection of particles larger than 1 µm which play an important role in the translocation of microorganisms.Therefore, the sampler can be used to examine the bioaerosol fraction which reaches the upper and lower respiratory tract with inhaled air and poses a potential health risk.
Air was sampled at a height of 1.30 m above ground, and it was aspirated onto a 90 mm contact Petri dish containing the appropriate agar medium.In the present study, 100 to 1,000 L of air was collected, depending on the sampling site and the analyzed group of microorganisms.The air flow sampling rate was 100 L over 1 min and 1,000 L over 10 min.According to the recommendations of the MAS-100 Eco manufacturer, sampling time should not exceed 10 min to minimize the risk of agar dehydration.
All air samples were collected downwind from municipal fountains.Samples of aerosols (A0, A20) were collected in the vicinity of municipal fountains according to Polish Standards PN-89 Z-04008/08:1989 and PN-89 Z-04111/01:1989.Wind speed and direction on the day of sampling were taken into consideration.
A0, A20 and CA samples were collected in triplicate.A total of 150 A0 samples, 150 A20 samples, and 30 CA samples were collected.Water samples were collected in three replicates from fountain basins containing stagnant water.Water samples were collected directly into sterile bottles.A total of 150 water samples were collected in 2018-2019 for microbiological analyses.

Potentially pathogenic mesophilic bacteria in atmospheric air and water samples
All air (A0, A20, CA) and water (W) samples from the studied fountains (I-V) were subjected to microbiological analyses to determine the total counts of mesophilic bacteria (TMB) capable of growing at a temperature of 36 ± 2°C for 24 h on tryptone soy agar (TSA; Oxoid, Basingstoke, UK), the total counts of hemolytic mesophilic bacteria (THMB) on TSA (TSA; Oxoid, Basingstoke, UK) with 5% addition of defibrinated sheep blood incubated at 36 ± 2°C for 24 h, Escherichia coli counts by membrane filtration on mFC Agar (Merck KgaA, Darmstadt, Germany) incubated at 44.5 ± 2°C for 24 h, and Aeromonas hydrophila counts on the Aeromonas Medium Base (Ryan; Oxoid Ltd., Basingstoke, UK), incubated at 37°C for 48 h.The microbiological analyses were conducted according to Polish Standards PN-89 Z-04111/01:1989 and PN-EN ISO 9308-1:2014.
The total counts of mesophilic E. coli and A. hydrophila in the air and water samples were determined based on the number of dark blue (mFC Agar) and opaque green colonies with a dark center, respectively, that emerged in a color reaction with the Aeromonas Medium Base (Ryan).The taxonomic identity of all E. coli and A. hydrophila colonies with a characteristic color was confirmed with the in situ hybridization (FISH) method.E. coli and A. hydrophila strains were identified with the use of specific Cy3-labeled oligonucleotide probes: ECO1167 (Thermo Fisher Scientific GmbH, Ulm, Baden-Württemberg, Germany) (Neef et al., 1995) and KO 229 (Thermo Fisher Scientific GmbH, Ulm, Baden-Württemberg, Germany), respectively (Franke-Whittle et al., 2005).Next, the hemolytic activity of all E. coli and A. hydrophila strains was determined according to the methods proposed by Chang et al. (2000).All typical colonies of E. coli and A. hydrophila that were grown on mFC agar and Aeromonas Medium Base were inoculated onto sheep blood agar (Difco, Birmingham, UK) and incubated at 35 ± 2°C for 24 h.The formation of a transparent zone around the inoculated colonies was a positive indication of hemolysis.Next, the R-factor was calculated based on the ratio between the total area of the clear zone (mm) and the colony area (mm).The isolates exhibiting high hemolytic activity at R ≥ 4 were classified as opportunistic (potentially pathogenic).A total of 110 potentially pathogenic E. coli strains and 120 potentially pathogenic A. hydrophila strains were isolated from air samples (A0, A20) and 77 E. coli and 90 A. hydrophila strains were isolated from water samples for microbiological analyses.The samples were pooled and the isolates were selected from all the sample collections.Detailed information about the number of strains isolated from water and air (A0, A20) samples is presented in Table S1.
The number of colony-forming units (cfu) of the identified bacteria (TMB, THMB, E. coli and A. hydrophila) in A0, A20, and CA samples was adjusted with the use of the Feller correction table.The results were expressed in cfu per 1 m 3 of air (cfu m -3 ).The bacterial counts determined in water samples were expressed in cfu m -3 to compare bacterial abundance in two environments (air and water).
Up to several percent of all bacterial populations in bioaerosol samples are culturable.However, there is no ideal method for collecting and recovering all microorganisms in the tested sample.The MAS-100 Eco sampler and culture methods based on Polish Standards PN-89 Z-04008/08:1989 and PN-89 Z-04111/01:1989 were used in this study.
To confirm or rule out phenotypic similarities between all potentially pathogenic strains of E. coli and A. hydrophila isolated from sampling sites (W, A0, A20) in the same periods, the phenotypic characteristics of isolates were determined with the use of API 20E and API 20NE biochemical test strips (BioMérieux®, Marcy-l'Étoile, France), respectively.

Antibiotic resistance of opportunistic E. coli and A. hydrophila strains in atmospheric air and water samples
The antibiotic resistance of potentially pathogenic E. coli and A. hydrophila isolates was analyzed by the disc diffusion method on Mueller-Hinton agar (Bauer et al., 1966).Eight popular antibiotics belonging to the antimicrobial classes of penicillins, cephalosporins, carbapenems, quinolones, aminoglycosides, tetracyclines, monobactams, and other antibiotics (chloramphenicol, fusidic acid, nitrofurantoin, and trimethoprim which are the first-choice drugs in some infections) were used in the analysis.These were: ampicillin (AMP, 10 µg), ceftazidime (CAZ, 30 µg), imipenem (IMP, 10 µg), ciprofloxacin (CIP, 5 µg), gentamicin (CN, 10 µg), tetracycline (TE, 30 µg), aztreonam (ATM, 30 µg), and trimethoprim/sulfamethoxazole (SXT, 25 µg).According to the DrugBank database (https://go.drugbank.com/drugs/DB00440),ampicillin, ceftazidime, imipenem, and aztreonam inhibit bacterial cell wall synthesis by binding to specific proteins.Ciprofloxacin targets the alpha subunits of DNA gyrase, prevents supercoiling and replication of bacterial DNA.Gentamicin leads to the displacement of divalent cations, increases membrane permeability, induces mistranslation of proteins, damages bacterial cell membranes, and disrupts the binding site of the 23S rRNA of the 50S ribosomal subunit.Tetracycline binds to the 30S ribosomal subunit and prevents tRNA from binding to the mRNA-ribosome complex, thus interfering with protein synthesis.Trimethoprim/sulfamethoxazole inhibit two consecutive steps in the biosynthesis of bacterial nucleic acids and proteins.The antimicrobial susceptibility of the isolated strains was tested according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI, 2012).The multi-drug resistance of the strains from each sampling site was assessed using the multiple antibiotic resistance (MAR) index that was calculated based on the resistance of E. coli and A. hydrophila isolates to at least two antibiotics from different classes, according to Krumperman (1983).

Statistical Analysis
Statistical analyses were conducted in the Statistica 13.3 program (StatSoft Inc. 1984-2017).The Kruskal-Wallis test, a non-parametric version of classical one-way ANOVA, was used to determine the significance of differences (p ≤ 0.05) in the abundance of TMB, THMB, E. coli, and A. hydrophila in air and water sampled from the analyzed sites and seasons (spring and summer).
Spearman's rank correlation test was used to determine statistically significant (p ≤ 0.05) correlations between: (i) counts of TMB, THMB, E. coli, and A. hydrophila bacteria in air and water samples, (ii) total bacterial counts in air and water samples vs. the physicochemical parameters of air and water, (iii) number of potentially pathogenic E. coli and A. hydrophila strains isolated from air and fountain water samples vs. number of antibiotic resistant E. coli and A. hydrophila isolates, and MAR values.
The similarity of sampling sites was determined by cluster analysis in the Past 4.03 program with the Bray-Curtis distance measure based on the percentage of E. coli and A. hydrophila bacteria exposed to the standard therapeutic doses of antibiotics (AMP, CAZ, IMP, CIP, CN, TE, ATM, SXT), MIC CIP and TE values.The results of the hierarchical cluster analysis were visualized in a dendrogram with the elastic beta method (β = -0.250)according to Sneath (1969).

Counts of Potentially Pathogenic E. coli and A. hydrophila in Air and Water Samples
The counts of the examined microorganisms differed significantly (Kruskal-Wallis test; p ≤ 0.05) by several orders of magnitude between samples of atmospheric air (A0, A20) and fountain water, depending on the microbial group, type of sample (W, A0, A20), and sampling period (Table 2).According to Polish Standards (PN-89 Z-04111/01:1989), the acceptable limits of TMB counts for low, moderate, and strong bacterial contamination in atmospheric air are set at > 1.0 × 10 3 , 1.0 × 10 3 -3.0 × 10 3 , and > 3.0 × 10 3 cfu m -3 , respectively.Based on the obtained results, abnormally high TMB counts (> 3.0 × 10 3 cfu m -3 ) were determined in most A0 and A20 samples, which indicates that microbial contamination in the air surrounding the examined fountains exceeded the safe levels established in these standards.In addition, based on the recommended values of THMB for outdoor environments, 80% and 20% of the tested air samples were characterized by strong (> 50 cfu m -3 ) and moderate (1-50 cfu m -3 ) bacterial contamination, respectively.
Samples of fountain water were characterized by a higher abundance of TMB, THMB, E. coli, and A. hydrophila (10 5 -10 9 cfu m -3 ) than air samples (10 1 -10 6 cfu m -3 ).However, an increase in E. coli and A. hydrophila counts in water and air samples (A0, A20) and the absence of these bacteria in CA samples collected in the same period suggest that microbial contaminants were transferred Table 2. Counts of total mesophilic bacteria (TMB), total hemolytic mesophilic bacteria (THMB), and potentially pathogenic Escherichia coli and Aeromonas hydrophila bacteria in samples of fountain water and air collected in the immediate vicinity of fountains (A0) and at a distance of around 20 m from the fountains (A20).
Potentially pathogenic isolates accounted for 70-75% of all E. coli and A. hydrophila strains isolated from A0, A20 and water samples (I-V).Their abundance was determined at 6.0 × 10 1 -9.9 × 10 2 and 2.0 × 10 1 -3.0 × 10 2 cfu m -3 , respectively, in 1 m 3 of A0 samples.The concentrations of these bacteria were 25-80% lower in A20 samples than in A0 samples.The counts of the examined bacteria were 10-100 times higher in A0 and A20 samples than in CA samples.The numbers of opportunistic E. coli and A. hydrophila strains differed significantly (Kruskal-Wallis test; p ≤ 0.05) between A0, A20, and CA samples (Table 2).In both air (A0, A20) and water samples, significant differences (Kruskal-Wallis test; p ≤ 0.05) in the counts of both species of potentially pathogenic bacteria were observed between spring and summer (Table 2).In water and air samples (A0, A20) collected in the spring and summer of 2018 and 2019, E. coli and A. hydrophila counts were highest in summer and lowest in spring in the first weeks of fountain operation (after the winter break).In summer, higher water temperature and increased activity of wild animals and humans (including children bathing in fountains) were chiefly responsible for high water contamination with the analyzed bacteria.High temperatures in the summer months contributed to water evaporation from fountain basins and, consequently, the transmission of potentially pathogenic bacteria (E. coli and A. hydrophila) into atmospheric air.The observed increase in bacterial concentrations in air samples collected in summer could be attributed to the fact that fountains are popular cooling and bathing sites for children, pets, and wild animals.Other causes of water contamination with E. coli and A. hydrophila, including water depth, location, contamination associated with human and animal activity, or the efficacy of water disinfection systems, cannot be ruled out (Burkowska-But et al., 2013;Flores et al., 2013;Włodyka-Bergier et al., 2019).However, no significant differences in the numbers of potentially pathogenic E. coli and A. hydrophila strains were observed between water samples collected from fountains with different basin depth (0.25-0.6 m) or surface materials (concrete vs. granite).According to Sanchez-Silva and Rosowsky ( 2008), concrete and granite are characterized by similar susceptibility to microbial corrosion.Granite is less porous and less susceptible to biofilm formation, whereas concrete has a highly alkaline pH (11-12) which acts as a natural defense against microbial corrosion.
Elevated concentrations of potentially pathogenic E. coli and A. hydrophila in most air (A0, A20) and fountain water samples point to high levels of fecal contamination and/or organic pollution (Flores et al., 2013).Barna and Kádár (2012) reported that waterborne enteric pathogens such as E. coli and A. hydrophila are highly infectious already at very low concentrations of 10-200 infective units.Nag et al. (2021) observed that human exposure to airborne E. coli is influenced by the decay constant of E. coli in the air, the rate of decay, and the bio-aerosolization efficiency factor.According to Triadó-Margarit et al. (2022), the transmission risk of antibiotic-resistant bacteria and antibiotic resistance genes increases substantially when the percentage of potentially pathogenic microorganisms in total airborne microbiota exceeds 3%.In the present study, potentially pathogenic E. coli and A. hydrophila strains accounted for 2-20% of TMB in A0 and A20 samples.These bacteria were sporadically isolated from CA samples, where they accounted for up to 1.1% of TMB.In most air samples collected in the vicinity of fountains, E. coli and A. hydrophila accounted for more than 3% of TMB counts, which indicates that the air surrounding fountains could pose a health threat for humans.The degree of this epidemiological risk is difficult to estimate because it is influenced by the infectious dose (ranging from 10 1 to more than 10 4 , depending on the pathogenic species) (Robins-Browne, 2001; Strachan et al., 2001), virulence, exposure time, as well as the age and health status of human visitors.

Correlations between Bacterial Counts in Atmospheric Air and Water
Spearman's rank correlation analysis revealed significant positive correlations (N = 480; p ≤ 0.05) between TMB, THMB, E. coli and A. hydrophila counts in all water and air samples.In turn, significant (p ≤ 0.05) negative correlations were observed only between TMB counts in A0 and CA samples (Fig. 2).The presence of correlations between bacterial counts in all W, A0 and A20 samples suggests that fountain water was the main source of microbial contamination and that potentially pathogenic bacteria were transmitted from water to atmospheric air in the vicinity of the studied fountains.Due to the absence of physical barriers, pigeons, seagulls, ducks, dogs, cats and/or humans can freely access municipal fountains.As a result, potentially pathogenic bacteria can be transmitted to and from the water by fountain users (Burkowska et al., 2012;Burkowska-But et al., 2013, Flores et al., 2013).Numerous research studies have shown that various bacteria (including Enterobacter spp., Escherichia spp., Pseudomonas spp., Bacillus spp., and Aeromonas spp.) Fig. 2. Significant correlations (N = 360; p ≤ 0.05) between the total counts of mesophilic bacteria (TMB), the total counts of hemolytic mesophilic bacteria (THMB), and the counts of potentially pathogenic Escherichia coli and Aeromonas hydrophila vs. the physicochemical parameters of fountain water (temperature-TW, dissolved oxygen-DO, oxygen saturation-OS, and pH) and atmospheric air (A0, A20, CA) (temperature-TA, relative humidity-RH, wind speed-WS).Air samples were collected in the immediate vicinity of fountains (A0), at a distance of around 20 m from the fountains (A20), and in the control site (CA).Significant (p ≤ 0.05) positive and negative correlations are marked in blue and red, respectively.The diameter of blue and red circles denotes the strength of the correlation.are readily transmitted from recirculated water to atmospheric air (Gotkowska-Płachta et al., 2013;Dueker et al., 2018;Gołaś et al., 2022).In this study, the presence of significant positive correlations between bacterial counts in W, A0, and A20 samples (collected downwind) suggests that fountain water was the main source of bioaerosol emissions.The above hypothesis is also validated by the absence of correlations between bacterial counts in A0, A20, and CA samples.The studied fountains are situated in municipal parks, which explains the local character of bioaerosols in the vicinity of fountains.Dense tree stands in parks probably inhibited the transport of microbial contaminants (including E. coli and A. hydrophila) to remote sampling sites.Windborne transport of the analyzed bacteria from the surface of soil, plants, or animal feces cannot be ruled out.However, the absence of significant correlations between microbial concentrations and WS values during the collection of A0 and A20 samples suggests that the above scenario was unlikely.In turn, significant correlations (Spearman's test; p ≤ 0.05) between E. coli and A. hydrophila counts and temperature were noted in both water and air (A0 and A20) samples.These correlations indicate that higher water temperature increases the concentrations of the examined bacteria in fountain water.In turn, higher air temperature promotes water evaporation and, consequently, the transmission of potentially pathogenic bacteria (E. coli and A. hydrophila) to the surrounding air.

Physicochemical Parameters of Air and Water Samples
During the study, the values of air temperature, relative humidity and wind speed ranged from 20.5 to 31.8°C, from 30.0% to 74.1% and from 2.0 to 5.1 m s -1 , respectively.The temperature of fountain water (TW) ranged from 9.4 to 27.3°C.Parameters DO, OS, and pH were determined in the range of 8.0-12.6 mg O2 L -1 , 101.2-122.2%O2 L -1 , and 7.16-8.8,respectively (Table S2).The statistical analysis (Spearman's test; N = 480) also revealed significant (p ≤ 0.05) positive correlations between the abundance of potentially pathogenic E. coli and A. hydrophila and the values of parameters TA, RH, and TW (Fig. 2).Similar results were reported by Burkowska-But et al. (2013) and Gotkowska-Płachta et al. (2013), who found that TA and RH significantly affect the survival and activity of microorganisms in the natural environment.The results of the correlation analysis indicate that higher TA and RH values affect the concentrations of mesophilic and potentially pathogenic bacteria not only in the direct vicinity of fountains (A0), but also at a distance of 20 m from these bioaerosol sources.Higher TA promotes water evaporation and the transmission of microorganisms (including opportunistic pathogens) to atmospheric air.Intensified evaporation can also increase RH values in air samples.As demonstrated by this study, higher RH led to a significant increase in E. coli and A. hydrophila counts in A0 and A20 samples.

Antibiotic Resistance Profiles of Potentially Pathogenic E. coli and A. hydrophila Isolated from Air and Water Samples
The antibiotic resistance profiles of the analyzed potentially pathogenic bacteria are presented in Table 3.The analysis demonstrated that 51-80% of the opportunistic E. coli and A. hydrophila isolates obtained from A0 and A20 samples were resistant to TE and CIP; 23-47% of the strains were resistant to AMP, SXT, and CAZ.In turn, 11-21% of E. coli and A. hydrophila strains isolated from A0 and A20 were resistant to the IMP, ATM, and CN.In the group of E. coli strains isolated from CA samples, 16.7% of the isolates were resistant to AMP, CIP, and TE.In the group of A. hydrophila strains isolated from CA samples, 14.3% of the isolates were resistant to AMP, TE, and SXT.Regardless of the analyzed species of potentially pathogenic bacteria in aerosol samples, the highest abundance of strains resistant to the tested antibiotics was noted in all A0 samples.The abundance of antibiotic-resistant E. coli and A. hydrophila isolates was only somewhat lower in the majority of A20 samples.Regardless of sampling site and fountain type, the resistance profiles of most bacterial isolates from A0 and A20 samples were identical to the profiles of bacterial strains isolated from water from the corresponding fountains.In addition, the MAR index values of most strains isolated from samples of fountain water (I-V) and atmospheric air (A0, A20) were indicative of multidrug resistance.Both E. coli and A. hydrophila strains isolated from water samples from all fountains were characterized by the highest nominal values of the MAR index at 0.87 and 0.86, respectively.The MAR index of E. coli strains isolated from A0 and A20 samples were determined at 0.80 and 0.59, respectively.A. hydrophila isolates were characterized by Table 3. Antibiotic resistance profiles of potentially pathogenic E. coli and A. hydrophila strains isolated from fountain water (W), and air in the immediate vicinity of the examined fountains (A0), and at a distance of around 20 m from the fountains (A20), and values of MAR index.0.17 0.14 a Air samples collected in the immediate vicinity of the studied fountains.b Air samples collected at a distance of around 20 m from the fountains.c Ampicillin (10 µg); d Ciprofloxacin (5 µg); e Tetracycline (30 µg); f Trimethoprim/sulfamethoxazole (25 µg); g Gentamicin (10 µg); h Ceftazidime (30 µg); i Aztreonam (30 µg); j Imipenem (10 µg).k Control site.l Not applicable.similar values of the MAR index, which reached 0.82 in the strains isolated from A0 and 0.68 in the strains isolated from A20.The lowest values of the MAR index were noted in E. coli and A. hydrophila strains isolated from CA samples, and they were determined at 0.17 and 0.14, respectively (Table 3).The nominal values of the MAR index for E. coli and A. hydrophila strains (> 0.2) indicate that the isolates originated from a source where antibiotics are used frequently and/or in large amounts (Mthembu et al., 2019).
In the literature, there is a general scarcity of information about the antibiotic-resistance profiles of E. coli and A. hydrophila in airborne aerosols.A. hydrophila is associated mainly with water environments, and it is often identified in aquaculture (Janda and Abbott, 2010;Gołaś et al., 2019), whereas E. coli is an indicator of fecal contamination in natural water bodies and artificial water reservoirs (Burkowska-But et al., 2013;Gotkowska-Płachta et al., 2016).In other studies (Zhang et al., 2018;Mao et al., 2019;Zhao et al., 2020;Jin et al., 2022), numerous E. coli strains resistant to various antibiotics, including macrolides, betalactams, tetracyclines, fluoroquinolones, and sulfonamides, were isolated from urban atmospheric environments, depending on the study site and the potential emission source.A study by Gołaś et al. (2022) revealed that A. hydrophila isolated from water and airborne aerosols were resistant to TE (88.2%),CTX (73.3%),SXT (44.4%),C (31.1%), and E (8.9%), and that 26-82% of these isolates were resistant to multiple drugs.In the present study, the increase in the counts of multidrug-resistant E. coli and A. hydrophila and MAR index values in A0 and A20 samples suggests that fountains can act as hot spots for the airborne transmission of antibiotic-resistant strains due to antibiotic selection pressure resulting from direct exposure to other drug-resistant strains (Piotrowska and Popowska, 2015).
In the current study, significant differences (Kruskal-Wallis test; p ≤ 0.05) in the percentage of potentially pathogenic E. coli and A. hydrophila with different drug resistance profiles were noted between air (A0, A20) and water samples collected in spring and summer (Figs. 3(a) and 3(b)).In spring, more than 20% of E. coli and A. hydrophila isolates from A0 samples were resistant to CIP and TE, whereas 11-15% of these isolates were resistant to AMP and SXT.In summer samples, the percentage of isolates resistant to CIP, TE, AMP, and SXT was nearly twice higher than in spring samples.More than 40% of E. coli and A. hydrophila strains isolated from A0 and A20 samples were resistant to standard doses of CIP and TE.More than 10% to more than 20% of the analyzed isolates were resistant to SXT and AMP.Most E. coli and A. hydrophila strains isolated from W, A0 and A20 samples were resistant to multiple drugs (MAR index).In spring, the MAR index of E. coli and A. hydrophila strains in A0 samples was determined at 0.73 and 0.76, respectively, and it was similar to the values noted in W samples (0.75 and 0.78, respectively).The MAR index of E. coli and A. hydrophila strains isolated from A0 and A20 samples was higher in summer.The nominal values of the MAR index were 0.77 and 0.44, respectively, in potentially pathogenic E. coli strains, and 0.89 and 0.71, respectively, in A. hydrophila.The MAR index of all E. coli and A. hydrophila isolates from CA samples remained below 0.2 regardless of season (Figs. 3(a) and 3(b)).
Spearman's rank correlation analysis (N = 450) confirmed the presence of correlations between the numbers of potentially pathogenic E. coli and A. hydrophila isolates, their resistance to antibiotics of eight different classes, and MAR index values in samples of fountain water and atmospheric air (Fig. 4).E. coli and A. hydrophila counts were bound by significant (p ≤ 0.05) positive correlations in all samples of fountain water and atmospheric air.In all analyzed samples (W, A0, A20), significant (p ≤ 0.05) positive correlations were observed between the abundance of E. coli and A. hydrophila, their resistance to CIP, TE, and SXT, and MAR index values.The presence of strong correlations between E. coli and A. hydrophila opportunistic strains resistant to CIP, TE, and SXT in water and air samples (A0, A20) suggests that these bacteria can be transmitted from water to atmospheric air, thus contributing to the spread of antibiotic resistance among various species of airborne bacteria.As a result, these bacteria become involved in the dissemination of antibiotic resistance among pathogenic microorganisms.

Minimum Inhibitory Concentrations (MICs) of Antibiotics
All E. coli and A. hydrophila strains isolated from CA samples were characterized by the lowest values of MIC CIP (< 0.06 µg mL -1 ) and MIC TE (≤ 2 µg mL -1 ).Significantly higher MIC CIP (1.0 to > 8.0 µg mL -1 ) and MIC TE values (32 to > 512 µg mL -1 ) were noted in the strains isolated from A0, A20, and W samples from all fountains (I-V).The MIC CIP (≥ 5.0 µg mL -1 ) and MIC TE (≥ 30 µg mL -1 ) Fig. 3. Seasonal diversity of the antibiotic resistance profiles of potentially pathogenic: a) E. coli, and b) A. hydrophila strains isolated from fountain water (W) and air in the immediate vicinity of the examined fountains (A0), and at a distance of around 20 m from the fountains (A20), and the values of the MAR index.AMP-ampicillin; CAZ-ceftazidime; IMP-imipenem; CIP-ciprofloxacin; CN-gentamicin; TE-tetracycline; ATM-aztreonam; SXT-trimethoprim/sulfamethoxazole.values exceeded standard therapeutic doses in more than 25% and 65% of E. coli strains and in 17% and 75% of A. hydrophila strains isolated from A0 and A20 samples, respectively (Figs. 5(a) and 5(b)).
The MICs of CIP and TE, and the drug resistance profiles of E. coli and A. hydrophila (Table 3) could imply that these antibiotics exert selection pressure on the studied bacteria.This observation is validated by the fact that the percentage of multidrug-resistant E. coli and A. hydrophila strains with high MICs for CIP and TE was higher in A0 and A20 samples than in CA samples.The effect of fountain water on the quality of the surrounding air has not been analyzed to date.Fountains eject streams of water, and they can act as local sources of bioaerosols.According to the authors, the extent to which fountains influence the microbiological quality and epidemiological safety of atmospheric air is determined by the degree of water contamination, water distribution, water disinfection, location (downtown areas, parks, shopping centers), density of urban and/or industrial development, climate, weather conditions, and air pollution (concentrations of particulate matter PM2.5, or heavy metals such as SO2, CO, O3).These factors not only affect the quantitative composition and species diversity of airborne microbiota in the vicinity of fountains, but they can also determine potential health risks for humans.
The research hypothesis postulating that municipal fountains are a source of potentially pathogenic E. coli and A. hydrophila bacteria in atmospheric air was confirmed by the results of agglomerative clustering and the Bray-Curtis dissimilarity index (N = 211) calculated based on the percentage of E. coli and A. hydrophila isolates resistant to all studied antibiotics, as well as MIC CIP and TE values (Fig. 5).In the analysis, the examined strains formed six distinctive clusters.Five clusters were composed of bacterial strains isolated from water and air (A0, A20) sampled from each fountain.In these clusters, the greatest similarities were observed between bacterial strains isolated from water (W) and A0 samples, whereas the strains isolated from A20 samples were somewhat less similar.The sixth cluster was composed of strains isolated from CA samples which were collected at a distance of 800-1,000 m from sampling sites No. I-V (Fig. 6).

CONCLUSIONS
The study demonstrated that potentially pathogenic E. coli and A. hydrophila bacteria in fountain water directly contributed to the microbial contamination of air samples collected in the immediate vicinity of the examined fountains (A0) and at a distance of 20 m (A20) from the studied sites.High values of the MAR index (> 0.20), MIC CIP (> 4 µg mL -1 ) and MIC TE (> 32 µg mL -1 ) of potentially pathogenic E. coli and A. hydrophila strains in A0, A20 samples point to a high risk of infection and an epidemiological threat to humans, which gives serious reason for concern.In all examined fountains, multidrug-resistant, opportunistic strains were highly likely to be carried to the surrounding air, which suggests that fountains can act as hot spots for the air-borne transmission of antibioticresistant strains.The results of the study indicate that fountain water and atmospheric aerosols in the direct vicinity of fountains should be monitored for epidemiological safety.

ACKNOWLEDGMENTS
Project financially supported by the Minister of Education and Science under the program entitled "Regional Initiative of Excellence" for the years 2019-2023, Project No. 010/RID/2018/19, amount of funding PLN 12 000 000.

DISCLAIMER
Reference to any company or commercial product does not constitute its endorsement or recommendation.

Fig. 1 .
Fig. 1.Location of the studied objects: I-fountain in the Central Park, II-fountain in the Old Town, III and IV-fountains in the Lower Town Park, V-fountain near the Town Hall.Type of samples: W-fountain water, A0-air sampled in the immediate vicinity of the fountain basin, A20-air sampled at a distance of 20 m from the fountain, CA-air sampled at the control site.(https://www.google.com/maps/@53.7766301,20.4751515,609m/data=!3m1!1e3?hl=pl)

Fig. 4 .
Fig. 4. Heat map of the Spearman's correlations (p ≤ 0.05; N = 450) between the counts of potentially pathogenic Escherichia coli and Aeromonas hydrophila isolated from fountain water (W) and air samples collected in the immediate vicinity of the studied fountains (A0) and at a distance of around 20 m from the fountains (A20), the antibiotic resistance of E. coli and A. hydrophila isolates, and the values of the MAR index.AMP-ampicillin; CAZ-ceftazidime; IMP-imipenem; CIP-ciprofloxacin; CN-gentamicin; TE-tetracycline; ATM-aztreonam; SXT-trimethoprim/sulfamethoxazole.