Recreational water activities have become an integral part of society for entertainment, health, and fitness. The water quality for these activities plays an integral role in the health and safety of the participants. This study aimed to monitor the presence of Escherichia coli in selected recreational water bodies, specifically rowing in the Gauteng Province of South Africa. Water samples were collected (upstream, midstream, and downstream) from three recreational water sites, i.e., Klip Rivier, Wemmer Pan, and Germiston Lake, monthly from May to October 2022. The presence of E. coli was determined using the Colilert® Quanti-Tray®/2,000 (IDEXX) assay and ChromoSelect agar. Pathotypes were confirmed using a multiplex PCR. The MPN for E. coli from all the samples exceeded the recommended guidelines of <2,000 microbes/100 mL. Pathogenic E. coli was detected in all three sites, with EPEC, ETEC, EAEC being most prevalent. The E. coli isolates showed 66% ESBL resistance and 94% carbapenemase resistance. Risk assessment showed recreational activities (rowing and swimming) posed a significant health risk as exceeded the annual risk benchmark limit of 1 × 104. The results obtained provide insights into the health risk associated with recreational activities within these water bodies and highlight the need for seasonal monitoring.

  • The presence of Escherichia coli and total coliforms found in recreational water bodies surpassed the recommended standards for safe recreational water use.

  • Pathogenic E. coli strains detected in this study were also found to be ESBL-resistant.

  • The Quantitative Microbial Risk Assessment shows there is a health risk associated with recreational activities such as rowing and swimming in these waters.

ATCC

American Type Culture Collection

E. coli

Escherichia coli

ESBL

extended-spectrum beta-lactamase

QMRA

quantitative microbial risk assessment

Over the past several years, recreational water activities have become an important part of society not only because they entertain people but because they have been beneficial for health and fitness reasons. Because their occurrence usually takes place within an unconfined water body, the quality of the water plays an integral role in the health and safety of the participants. Engaging in activities such as swimming, paddling, rowing, and fishing has been linked to a staggering number of gastrointestinal and respiratory illnesses (Russo et al. 2020). Faecal contamination in surface waters can occur from storm water and surface runoffs, discharge from inadequately treated or raw sewage or direct disposal of industrial effluent into the receiving waters (Edokpayi et al. 2017).

Monitoring the level of contamination in water bodies is crucial in order to assess and mitigate potential health risks. Indicator organisms, like Escherichia coli, are employed as they provide valuable insights into the presence of faecal contamination. While E. coli is typically considered a commensal bacterium, virulent strains exist that possess specific genetic traits enabling them to cause diseases (Russo et al. 2020). Strains that cause disease in the human intestinal tract are referred to as diarrheagenic E. coli (DEC) and are classified into seven subtypes; enteropathogenic E. coli (EPEC), enterohaemorrhagic E. coli (EHEC), enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), enteroinvasive E. coli (EIEC), and diffusely adherent E. coli (DAEC) (Santos et al. 2020). Infection by these strains can lead to various illnesses, depending on the presence of specific virulence factors (Ebomah et al. 2018).

However, when these bacteria become resistant to antibiotics, specifically extended-spectrum beta-lactamase (ESBL) and carbapenemase, they pose a major public health threat (Ramaite et al. 2022). Though the extensive use of antibiotics is the main cause of increased bacterial resistance, the external environment plays a greater role in the transmission and development of novel resistant phenotypes (Serwecińska 2020). Resistant bacteria enter water bodies via faecal waste, making the aquatic environments a reservoir for resistant pathogenic bacteria. Resistance spreads to other bacteria through horizontal gene transfer of the resistance genes (Kusi et al. 2022).

The presence of E. coli in waters indicates the potential presence of other waterborne bacterial pathogens such as Salmonella, Shigella, Klebsiella, and Vibrio cholerae that can adversely affect human health (Cabral 2010). All these bacteria can be transmitted through the faecal oral route or through contact with contaminated water. The aim of the study was to determine the presence of pathogenic Escherichia coli and their antimicrobial profiles in these specific water bodies that are used for recreational activities specifically rowing in Gauteng as it has not been done before. Additionally, based on the presence of virulent E. coli pathotypes, a quantitative microbial risk assessment framework was applied to estimate the health risk posed to users when using the surface water for recreational activities, such as rowing and swimming, among others.

Description of sampling sites

Water samples were collected from three recreational water bodies in the Gauteng Province of South Africa: viz. Klip River, Germiston Lake, and Wemmer Pan (Figure 1) over a period of 6 months to cover autumn, winter, and early summer conditions (May–October). These sample sites were selected as a pilot study to assess the health risk associated with these water bodies as they are the place where practice for rowing for the University of Johannesburg academic team takes place. In addition, Gauteng gets summer rainfall which could contribute to diluting the pathogen load. From each water body, samples were collected from three sites: (a) upstream, (b) midstream, and (c) downstream. A total of 54 water samples were collected for this study.
Figure 1

Sampling sites for the collection of surface water from (a) Klip Rivier (26°18′10.9″S 28°00′55.6″E), (b) Wemmer Pan (26°14′00.0″S 28°03′29.0″E), and (c) Germiston Lake (26°14′01.7″S 28°09′35.4″E).

Figure 1

Sampling sites for the collection of surface water from (a) Klip Rivier (26°18′10.9″S 28°00′55.6″E), (b) Wemmer Pan (26°14′00.0″S 28°03′29.0″E), and (c) Germiston Lake (26°14′01.7″S 28°09′35.4″E).

Close modal

Sample collection and physio-chemical analysis

At each sampling site, a 500 mL of water sample was collected using the grab-sample technique, by submerging a sterile glass bottle into the water (Delair 2021). On site physio-chemical analysis (pH, temperature, and conductivity) was conducted using a Hanna-HI98129 pH and EC meter. The samples were transported on ice to the Water and Health Research Centre (WHRC) laboratory (University of Johannesburg). Sample turbidity was measured in the WHRC laboratory on the day of sampling using the Oakton WD-35635-00 T-100 Turbidity Meter Kit according to the manufacturer's instructions.

Microbiological analysis

The collected samples were concentrated by filtering 50 mL through a (0.45 μm) nitrocellulose membrane (Whatmann) using a filtration manifold. The membranes were incubated in 50-mL LB broth at 37 °C for 24 h and subsequently used to isolate E. coli by plating onto Coliform ChromoSelect agar (Millipore). E. coli-positive colonies appeared blue-violet in colour while Klebsiella-positive colonies were salmon-red in colour.

Quantification of total coliforms and E. coli

Water samples were analyzed for the presence of coliforms and more specifically E. coli using the Colilert® Quanti- Tray®/2,000 (IDEXX) as per manufacturer's instructions. Samples were analyzed in 100 mL, 10 mL and/or 1 mL, depending on the lowest obtainable counts. The Quanti-Trays were incubated at 37 °C for 18 h. All sample lots included positive and negative controls to confirm sterility and product performance. These included a coliform-positive control (Klebsiella pneumonia ATCC 31488), coliform-negative control (Pseudomonas aeruginosa ATCC 27853), and E. coli-positive control (E. coli ATCC 25922), and autoclaved distilled water as a product control. The yellow wells were identified as positive for total coliforms and all fluorescent yellow (and white) wells under UV light were identified as positive for the presence of E. coli. Following incubation, the Colilert® Quanti- Tray®/2,000 were analyzed, and the results recorded. The IDEXX MPN table was used to calculate the most probable number (MPN).

DNA extraction and multiplex PCR

The Colilert® Quantitray®/2,000 samples were analyzed under UV light. Each Colilert® Quanti-Tray®/2,000 sample was analyzed in triplicate depending on the number of positive fluorescent wells for E. coli as indicated in Table 1 (Delair 2017).

Table 1

Determining the number of wells tested per Colilert® Quanti-Tray®/2,000

No. of wells positive for E. coliNo. of 2-mL tubes extracted per Colilert® Quanti- Tray®/2,000 sample
1–30 
31–60 
61–97 
No. of wells positive for E. coliNo. of 2-mL tubes extracted per Colilert® Quanti- Tray®/2,000 sample
1–30 
31–60 
61–97 

If the Quanti-Tray had between 1 and 30 wells that were positive for E. coli, then the broth from the tray would be removed from 10 randomly selected wells to make up to 2 mL. If the sample had between 31 and 60 positive wells, then two tubes of 2 mL from 10 randomly selected wells each would be removed. For a full positive tray, the three tubes would be filled in the same way. These tubes were then used for further analysis such as DNA extraction and multiplex PCR.

DNA isolation

DNA was extracted from the samples using the 96-well semi-automated extraction method with the QIAamp 96 DNA QIAcube HT Kit (QIAGEN®) (Delair 2017). Extractions were carried out on each tube as per the number of tubes for a given sample. The protocol was as follows: 200 μL of broth was added to a 2-mL 96-well-plate (Greiner bio-one, Cat. No. 780285). To this 100 μL of VXL buffer was added to the wells and incubated at room temperature for 5 min. Thereafter 350 μL of ACB buffer was added to the plate and mixed. The lysate was then transferred to a 96-well filter plate supplied with the kit. The filter plate was placed on a manifold and a vacuum pressure was applied to filter the lysate. The filter plate was then washed with 600 μL AW 1 buffer followed by 600 μL AW2 buffer and 600 μL ethanol (96%). A vacuum pressure was applied between each wash step. A further 3 min of vacuum pressure was applied to dry the plate. A 350 μL of elution plate (Eppendorf- 951040188) was placed within the manifold and 150 μL AVE buffer was added to the wells and incubated at room temperature for 2 min. A vacuum pressure was applied for 2 min to ensure all buffers were eluted. The plate was sealed with clear seal film and stored at −20 °C until further use. The eluted DNA was used as a template in the PCR reaction.

Multiplex PCR

The extracted DNA was used as a template, for the PCR reaction. E. coli virulence genes were amplified as described by Omar & Barnard (2014) and targeted the following virulence genes; enteroaggregative E. coli heat-stable enterotoxin (astA), Heat-stable toxin (st), heat liable toxin (lt), aggregative gene (eagg), shiga toxin 1 (stx1), shiga toxin 2 (stx2), invasion associated locus gene (ial), and intimin gene (eaeA). The malate dehydrogenase (mdh) gene is a housekeeping gene for E. coli and the glyceraldehyde-3-phosphate dehydrogenase (gapdh) was used as an internal control for the PCR reaction. The primer sequences for the targeted genes are outlined in Table 2.

Table 2

Primers used in the mPCR reaction (Omar & Barnard 2014)

PathogenGenePrimer sequence (5′- 3′)Size (bp)Reference
Internal Control mdh (F) GGT ATG GAT CGT TCC GAC CT 304 Tarr et al. (2002)  
mdh (R) GGC AGA ATG GTA ACA CCA GAG T 
EIEC ial (F) GGT ATG ATG ATG ATG AGT CCA 650 López-Saucedo et al. (2003)  
 ial (R) GGA GGC CAA CAA TTA TTT CC 
EHEC/ eaeA (F) CTG AAC GGC GAT TAC GCG AA 917 Aranda et al. (2004)  
Atypical EPEC eaeA (R) CCA GAC GAT ACG ATC CAG 
Typical EPEC bfpA (F) AAT GGT GCT TGC GCT TGC TGC 410 Aranda et al. (2004)  
bfpM (R) TAT TAA CAC CGT AGC CTT TCG CTG AAG TAC CT Omar & Barnard (2014)  
EAEC eagg (F) AGA CTC TGG CGA AAG ACT GTA TC 194 Pass et al. (2000)  
eagg (R) ATG GCT GTC TGT AAT AGA TGA GAA C 
EHEC stx1 (F) ACA CTG GAT GAT CTC AGT GG 614 Moses et al. (2006)  
stx1 (R) CTG AAT CCC CCT CCA TTA TG 
stx2 (F) CCA TGA CAA CGG ACA GCA GTT 779 Moses et al. (2006)  
stx2 (R) CCT GTC AAC TGA GCA CTT TG 
ETEC lt-1 (F) TGG ATT CAT CAT GCA CCA CAA GG 360 Pass et al. (2000)  
lt-1 (R) CCA TTT CTC TTT TGC CTG CCA TC 
st-a (F) TTT CCC CTC TTT TAG TCA GTC AAC TG 160 Pass et al. (2000)  
st-a(R) GGC AGG ATT ACA ACA AAG TTC ACA 
E. coli toxin astA (F) GCC ATC AAC ACA GTA TAT CC 106 Kimata et al. (2005)  
astA (R) GAG TGA CGG CTT TGT AGT C 
External Control gapdh (F) GAG TCA ACG GAT TTG GTC GT 238 Mbene et al. (2009)  
gapdh (R) TTG ATT TTG GAG GGA TCT CG 
PathogenGenePrimer sequence (5′- 3′)Size (bp)Reference
Internal Control mdh (F) GGT ATG GAT CGT TCC GAC CT 304 Tarr et al. (2002)  
mdh (R) GGC AGA ATG GTA ACA CCA GAG T 
EIEC ial (F) GGT ATG ATG ATG ATG AGT CCA 650 López-Saucedo et al. (2003)  
 ial (R) GGA GGC CAA CAA TTA TTT CC 
EHEC/ eaeA (F) CTG AAC GGC GAT TAC GCG AA 917 Aranda et al. (2004)  
Atypical EPEC eaeA (R) CCA GAC GAT ACG ATC CAG 
Typical EPEC bfpA (F) AAT GGT GCT TGC GCT TGC TGC 410 Aranda et al. (2004)  
bfpM (R) TAT TAA CAC CGT AGC CTT TCG CTG AAG TAC CT Omar & Barnard (2014)  
EAEC eagg (F) AGA CTC TGG CGA AAG ACT GTA TC 194 Pass et al. (2000)  
eagg (R) ATG GCT GTC TGT AAT AGA TGA GAA C 
EHEC stx1 (F) ACA CTG GAT GAT CTC AGT GG 614 Moses et al. (2006)  
stx1 (R) CTG AAT CCC CCT CCA TTA TG 
stx2 (F) CCA TGA CAA CGG ACA GCA GTT 779 Moses et al. (2006)  
stx2 (R) CCT GTC AAC TGA GCA CTT TG 
ETEC lt-1 (F) TGG ATT CAT CAT GCA CCA CAA GG 360 Pass et al. (2000)  
lt-1 (R) CCA TTT CTC TTT TGC CTG CCA TC 
st-a (F) TTT CCC CTC TTT TAG TCA GTC AAC TG 160 Pass et al. (2000)  
st-a(R) GGC AGG ATT ACA ACA AAG TTC ACA 
E. coli toxin astA (F) GCC ATC AAC ACA GTA TAT CC 106 Kimata et al. (2005)  
astA (R) GAG TGA CGG CTT TGT AGT C 
External Control gapdh (F) GAG TCA ACG GAT TTG GTC GT 238 Mbene et al. (2009)  
gapdh (R) TTG ATT TTG GAG GGA TCT CG 

All virulence genes were amplified in a 20 μL reaction mixture containing 10 μL of the 2× Qiagen® mPCR master mix (Hotstart Taq DNA polymerase, 10× buffer, 2 mM MgCl2 and dNTP mix), 1 μL 5× Q-solution, 4.5 μL of PCR grade water, 2 μL MgCl2, 2 μL of template DNA and 0.5μL of the primer mix [0.1 μM of mdh and lt-1 primers (F and R), 0.2 μM of ial, gapdh, eagg, astA, and bfp primers (F and R), 0.3 μM of eaeA and stx2 primers (F and R), 0.5 μM of stx1 and st-a primers (F and R)]. For the negative control reaction mixture, the template DNA was replaced with sterile PCR grade water and the positive control reaction contained a combination of the virulence genes from extracted plasmids from transformed E. coli strains. In the case of single genes, single primer sets were used.

The virulence genes were amplified in a BIO-RAD® T100 Thermal Mycycler under the following conditions: enzyme activation at 95 °C for 15 min, 35 cycles of DNA denaturation at 94 °C for 45 s, annealing at 55 °C for 45 s, elongation at 68 °C for 2 min with a final elongation step at 72 °C for 5 min. The amplified products were separated according to their molecular size on a horizontal agarose gel slab [2.5% (w/v)] containing ethidium bromide (0.5 μg/mL) in TAE buffer. Electrophoresis was conducted at 80–100 V for 30–50 min and viewed under UV light (Gene Genius Bio Imaging System, Vacutec®). The relevant sizes of the DNA fragments were estimated by comparing their electrophoretic mobility to that of a standard 100 bp marker (Quick-Load®, New England BioLabs®) that was run parallel with the samples on each gel.

Antimicrobial susceptibility testing of single isolates

From the selective media, presumptive colonies were selected, and colony PCR was carried out using the mPCR as described above. Isolates were grown on Muller Hinton agar and incubated overnight at 37 °C. A 0.5 MacFarland standard was made for each isolate and the ID and AST was determined using the Vitek® 2 compact system. For the isolate identification, the Gram-negative (GN) ID cards (bioMérieux Inc, USA) was used and for the antimicrobial resistance profiles, the (AST-N256 (bioMérieux Inc, USA) cards were used. The samples were processed according to manufacturers' instructions. The AST cards targeted the following antibiotics Ampicillin, Amoxicillin/Clavulanic acid, Piperacillin/Tazobactam, Cefuroxime, Cefuroxime-Axetil, Cefoxitin, Cefotaxime, Ceftazidime, Cefepime, Ertapenem, Imipenem, Meropenem, Amikacin, Gentamicin, Tobramycin, Ciprofloxacin, Tigecycline, Colistin, Trimethoprim/ Sulfamethoxazole. The AST results were interpreted using the CLSI (2018) and EUCAST (2018) guidelines. Isolates were reported as susceptible (S), intermediate (I), or resistant (R). All the relevant positive and negative controls were included for all cards.

Estimation of the health risk posed by the surface water sources

Based on the detection of virulent E. coli pathotypes, a quantitative microbial risk assessment approach was used to estimate the health risk posed by the surface water sources to community members. The data generated using the Colilert® Quanti- Tray®/2,000 (IDEXX) assay were used as input parameters for the E. coli concentration at the three sampling sites.

The major exposure routes associated with the surface water sources pertained to recreational activities and included rowing and swimming. However, additional exposure scenarios commonly associated with surface water sources in rural or informal settlement communities were also assessed, and included: drinking, accidental consumption, washing laundry by hand, and washing/bathing. The various exposure scenarios (including exposure volume, frequency of occurrence, and dose calculation) that were assessed in the current study are outlined in Table 3. Based on the detection of pathogenic E. coli in surface water sources, the fraction of E. coli assumed to be human infectious was set at 0.005–0.1 (Howard et al. 2006; Schoen et al. 2014).

Table 3

Dose–response input parameters and exposure scenarios

OrganismVariables and distributionaDose response model (DRM)DRM backgroundReference
E. coli (Enteroinvasive) BC: Site-specific concentration (cells/mL) as determined using MPN (Uniform) IF%: 0.005–0.10 Beta-Poisson
N50 = 2.11 × 106
α = 1.55 × 10−1 
Model: Human
Exposure: Ingestion
Response: Infection with positive stool isolation 
Howard et al. (2006), Schoen et al. (2014), Reyneke et al. (2020)  
ActivityVolume (distribution)aDose calculationFrequency/year (distribution)Reference
Rowing Youth: 7.7–11.6 mL (Uniform)
Adult: 11.6–15.4 mL (Uniform) 
d = BC × IF% × V 52–156 (Uniform) Ngubane et al. (2022)  
Swimming Youth: 49.3–62.7 mL (Uniform)
Adult: 21.3–32.0 mL (Uniform) 
d = BC × IF% × V 1–52 (Uniform) Ngubane et al. (2022)  
Intentional drinking λ = 872.5 mL (Poisson) d = BC × IF% × V 365 (Point) Mons et al. (2007)  
Accidental consumption 50–200 mL (Uniform) d = BC × IF% × V 0.5–2 (Uniform) Busgang et al. (2018)  
Washing laundry by hand
(Aerosol ingestion) 
AV: 1.13 × 10−13/mL to 8.18 × 10−12/mL (Uniform)
AL: 0–1.07 × 105/L (Uniform)
T: 60 min (Point)
BR: 10 L/min to 20 L/min (Uniform) 
d = BC × IF% × (AV × AL × BR × T35–55 (Uniform) Fischer et al. (2019), Reyneke et al. (2020)  
Washing/ bathing
(Hand-to-mouth contact) 
FT: μ = 1.65 μm, σ = 0.32 μm (Normal)
AHS: 0.106–0.131 m2 (Uniform)
SH: 0.1 (Point)
TEHF: 0.68 (Point)
CE: 5–20 (Uniform) 
d = (BC × IF% × FT × AHS) × SH × TEHF×CE 365 (Point) Julian & Pickering (2015), Reyneke et al. (2020), USEPA (2011)  
OrganismVariables and distributionaDose response model (DRM)DRM backgroundReference
E. coli (Enteroinvasive) BC: Site-specific concentration (cells/mL) as determined using MPN (Uniform) IF%: 0.005–0.10 Beta-Poisson
N50 = 2.11 × 106
α = 1.55 × 10−1 
Model: Human
Exposure: Ingestion
Response: Infection with positive stool isolation 
Howard et al. (2006), Schoen et al. (2014), Reyneke et al. (2020)  
ActivityVolume (distribution)aDose calculationFrequency/year (distribution)Reference
Rowing Youth: 7.7–11.6 mL (Uniform)
Adult: 11.6–15.4 mL (Uniform) 
d = BC × IF% × V 52–156 (Uniform) Ngubane et al. (2022)  
Swimming Youth: 49.3–62.7 mL (Uniform)
Adult: 21.3–32.0 mL (Uniform) 
d = BC × IF% × V 1–52 (Uniform) Ngubane et al. (2022)  
Intentional drinking λ = 872.5 mL (Poisson) d = BC × IF% × V 365 (Point) Mons et al. (2007)  
Accidental consumption 50–200 mL (Uniform) d = BC × IF% × V 0.5–2 (Uniform) Busgang et al. (2018)  
Washing laundry by hand
(Aerosol ingestion) 
AV: 1.13 × 10−13/mL to 8.18 × 10−12/mL (Uniform)
AL: 0–1.07 × 105/L (Uniform)
T: 60 min (Point)
BR: 10 L/min to 20 L/min (Uniform) 
d = BC × IF% × (AV × AL × BR × T35–55 (Uniform) Fischer et al. (2019), Reyneke et al. (2020)  
Washing/ bathing
(Hand-to-mouth contact) 
FT: μ = 1.65 μm, σ = 0.32 μm (Normal)
AHS: 0.106–0.131 m2 (Uniform)
SH: 0.1 (Point)
TEHF: 0.68 (Point)
CE: 5–20 (Uniform) 
d = (BC × IF% × FT × AHS) × SH × TEHF×CE 365 (Point) Julian & Pickering (2015), Reyneke et al. (2020), USEPA (2011)  

aIF%: human infectious fraction of target organism; DRM background: experimental model used to determine the dose response model (animal model, route of exposure and health endpoint); FT: water film thickness; AHS: average hand surface area; SH: hand surface area in contact with mouth; TEHF: transfer efficiency of bacteria; CE: hand-to-mouth contact events; AV: aerosol volume; AL: aerosol concentration; T: time; BR: human breathing rate.

The beta-Poisson dose response model (Equation (1)) was used to determine the risk of infection associated with the various exposure scenarios based on the presence of E. coli (ingestion leading to infection) as previously described by Haas et al. (1999) and Ryan et al. (2014):
(1)
in which Pinf is the probability of infection from a single exposure, d is the dosage of microorganisms (number of microorganisms ingested), N50 is the median infective dosage and α is a shape factor. All parameters associated with the dose–response models are outlined in Table 3.
The final step of the QMRA was the risk characterisation, where the probability of infection (expressed as likely numbers of infections per 10,000 persons per year) was calculated for each target organism and the respective exposure scenarios, using Equation (2):
(2)
in which P is the probability of infection from n exposure events per year based on the calculated per exposure probability of infection (Pinf).

All exposure scenarios were simulated with Monte Carlo analysis using RStudio (version 1.0.153), whereby each scenario was simulated using 500,000 iterations. During the analyses, the various dose parameters (e.g., ingestion volumes and pathogen concentrations, among others) and exposure events per year (Table 3) were randomly sampled according to the corresponding distribution for each parameter to determine the estimated infection risk associated with a single exposure event and the subsequent annual exposure risk.

Physio-chemical analysis

Water samples were collected from three selected recreational water bodies from three sites over a period of 6 months totaling 54 samples in the Gauteng region. For each water sample, the physio-chemical data was analyzed and recorded. The results are outlined in Figure 2.
Figure 2

Representation of the physio-chemical data obtained during the study period.

Figure 2

Representation of the physio-chemical data obtained during the study period.

Close modal

On average, the electrical conductivity ranged between 410 and 560 mS, although in September, Klip Rivier had an outlier value as it was the lowest downstream conductivity. The temperature varied from 10–15 °C on average, but because May and September were the hottest months, it reached 19 °C during those times. In comparison to all sites the downstream of Klip Rivier had both the lowest and highest temperature. The downstream of Klip Rivier and Wemmer Pan had the highest temperatures. The pH was consistent throughout the sampling sites and fell within the range of 6.5 and 8.5. Downstream of Klip Rivier had the lowest pH. According to the South African recreational Water Quality Guidelines Vol 2 (1996) at a pH of 8.5–9 swimming is permitted, but a pH > 9 causes ear, mucous membrane, eye, and skin irritations. Turbidity at Germiston Lake downstream was higher than the other sites, which can be caused by the presence of mud or wood ashes as there is a wooden jetty situated downstream.

Microbiological detection of E. coli

The presence of E. coli was determined by the Colilert® Quanti-Tray® method and the results are outlined in Figure 3.
Figure 3

Average MPN of E. coli and total coliform data obtained during the study period. The total number of samples collected for upstream, downstream and midstream was six samples per water body.

Figure 3

Average MPN of E. coli and total coliform data obtained during the study period. The total number of samples collected for upstream, downstream and midstream was six samples per water body.

Close modal

This study investigated the presence of E. coli from water bodies used for recreational purposes in the Gauteng region. Both coliforms and E. coli were detected in all water bodies, however, in comparison to the other locations, Wemmer Pan exhibited a higher detection of E. coli and coliforms compared with the other sites (Figure 3). The South African recreational water quality guidelines (Volume 2, 1996) have established specific standards that stipulate that the concentration of E. coli, a commonly used indicator organism for faecal contamination, should not exceed 500 counts per 100 mL of water for full contact recreational activities. These surfaces are unsuited for recreational use because the most probable number of E. coli in all of the samples was higher than the advised limit of 2,000 microbes/100 mL. Although Germiston Lake had a presence of E. coli and coliforms at all points, it exhibited the lowest levels of E. coli at <1,000 microbes/100 mL, which deems the recreational water safe for activities such as rowing and canoeing.

In all samples the total coliforms were present according to the Water Quality Guidelines Vol 2 total coliform counts >130 counts/100 mL for swimming (full contact) may lead to gastrointestinal effects. Total coliform >2,000 for recreational water activities with moderate contact (rowing) for an extensive amount of time can cause gastrointestinal illnesses. The total coliform in the upstream of Wemmer Pan was the highest and the downstream of Germiston Lake was the lowest.

The presence of pathogenic E. coli was confirmed by mPCR in all three sites (Figure 4). Each isolate was classified into its pathotypes based on the presence of the virulence gene in question (Table 2). The EPEC (eaeA, bfp gene), ETEC (lt, st genes) and EAEC (eagg gene) strains were the most prevalent throughout the study period and found consistently in both the upstream and downstream of the water bodies. High counts of all the strains were found during October in all the sites. This could be attributed to high rainfall patterns. Surface waters are constantly at risk of faecal contamination by surface runoffs and human activity (Cho et al. 2023). A study by Edokpayi et al. (2017) found high levels of E. coli in river water in Limpopo province which was used for daily activities.
Figure 4

Representation of the E. coli pathotypes found at the sampling sites for each month.

Figure 4

Representation of the E. coli pathotypes found at the sampling sites for each month.

Close modal
The antimicrobial resistance profiles were determined using the Vitek® 2 compact system as depicted in Figure 5. The results showed that 66% of the isolates tested were ESBL resistant, and 11% were intermediately resistant. Of the tested isolates, 8% were carbapenem resistant, 38% were susceptible to carbapenems and 31% were intermediately resistant. From the isolates tested, 26% were resistant to the last resort antibiotic, Colistin, but this needs to be further tested. These results correlate with other studies done on river water in South Africa (Fadare et al. 2020).
Figure 5

E. coli antimicrobial susceptibility profiles for ESBL, carbapenems and colistin using the N256 Gram-negative cards on the Vitek® 2 compact system (R-resistant, S- susceptible, I- intermediate).

Figure 5

E. coli antimicrobial susceptibility profiles for ESBL, carbapenems and colistin using the N256 Gram-negative cards on the Vitek® 2 compact system (R-resistant, S- susceptible, I- intermediate).

Close modal

When certain antibiotics are no longer able to eradicate or inhibit the growth of bacteria, the bacteria are deemed to be resistant to the antibiotic. Resisatnce in microbes may be intrinsic by nature of their physiology, or it maybe aquired through practice of antibiotics as in the case of ESBL (Reygaert 2018; Pang et al. 2019). Bacterial isolates are considered to be suseptible (S), intermediate (I), or resistant (R) based on the clinical breakpoint of each organism tested (Barnes et al. 2023). A microbe is said to be susceptible to an anitibiotic when its growth is inhibited in the presence of the antibiotic. A bacteria is considered intermediately sensitive to an antibiotic when it is inhibited by an antibiotic but has uncertain therapeutic effect and would need higher, more frequent doses of antibiotics. When the antibiotic has no effect on the microbe, the microbe is said to be resistant to that antibiotic (Prinzi 2023).

Multidrug resistance in E. coli is increasingly observed worldwide (Poirel et al. 2018). E. coli has the capacity to gain resistant genes through horizontal gene transfer, the most concerning is the acquisition of genes encoding for beta lactamases (Gauba & Rahman 2023). This phenomenon enables the transmission of virulent genes between humans and animals especially in water bodies where faecal matter and human activity are common (Tao et al. 2022). Several studies looked into the presence of antimicrobial resistant bacteria in recreational waters (Nappier et al. 2020). Human exposure to antibiotic resistant E. coli strains may pose a potential health risk by injestion of these microbes which may cause illness, making treatment options limited (Poirel et al. 2018).

Health risk posed by E. coli

The annual risk of infection linked to the exposure of community members to E. coli in the surface water sources, while performing recreational and other domestic activities, is outlined in Figure 6 and Supplementary material, Figure S1, respectively. The risk of gastrointestinal illness from rowing at Germiston Lake was marginally exceeded in May for both youths and adults, whereafter the risk decreased to below the annual risk benchmark limit of 1 × 104 in June. However, the highest risk for gastrointestinal illness for both youths and adults was subsequently recorded from July to October when the benchmark limit was exceeded (Figure 6(a)). Overall, the risk associated with swimming followed a similar seasonal pattern, with the exception of reduced risk below the benchmark limit being estimated for adults swimming in Germiston Lake in June (Figure 6(b)). In comparison, results for the surface water samples collected from the Klip Rivier (Figure 6(c) and 6(d)) and Wemmer Pan (Figure 6(e) and 6(f)) indicated that both rowing and swimming posed a significant health risk to community members, as the annual risk benchmark limit of 1 × 104 was exceeded for both youths and adults over the entire monitoring period (May to October) at both sites. Similarly, Ngubane et al. (2022) reported on the associated health risk posed by pathogenic E. coli and Cryptosporidium to community members using surface water sources for recreational activities such as swimming or canoeing in the Msunduzi catchment in the KwaZulu-Natal province, South Africa. In comparison, Denissen et al. (2023) assessed the health risk posed by various surface water sources (e.g., stream water, marsh water, stormwater) in the Western Cape province (South Africa) when used for recreational purposes such as swimming or playing near/in the water. Overall results indicated that the annual risk of infection benchmark limit (1 × 104) was exceeded based on the presence of Klebsiella pneumoniae and Pseudomonas aeruginosa.
Figure 6

Annual risk of infection linked to the use of water from Germiston Lake for (a) Rowing and (b) Swimming, from Klip Rivier for (c) Rowing and (d) Swimming, and from Wemmer Pan for (e) Rowing and (f) Swimming. The outer whiskers represent the first and third quartiles, and the inner circle represents the median. The red dashed line represents the annual infection risk benchmark limit used for drinking water (1 × 10−4; one infection per 10,000 people per year).

Figure 6

Annual risk of infection linked to the use of water from Germiston Lake for (a) Rowing and (b) Swimming, from Klip Rivier for (c) Rowing and (d) Swimming, and from Wemmer Pan for (e) Rowing and (f) Swimming. The outer whiskers represent the first and third quartiles, and the inner circle represents the median. The red dashed line represents the annual infection risk benchmark limit used for drinking water (1 × 10−4; one infection per 10,000 people per year).

Close modal

Analysis of the additional domestic activities that may be practised by community members in close proximity to the sampling sites (i.e., Germiston Lake, Klip Rivier, and Wemmer Pan) indicated that the use of the surface water sources for washing laundry by hand could safely be practised at all three sites (104–107); however, activities such as intentional drinking, accidental consumption, and using the surface water for bathing posed a significant health risk (>104) to community members (Supplementary material, Figure S1). Similarly, the use of untreated environmental water sources (e.g., stream water, marsh water, stormwater, rainwater) for domestic activities has been reported to pose a significant health risk to community members based on the presence of a multitude of pathogenic or opportunistic pathogenic microorganisms, including E. coli, P. aeruginosa, Legionella spp., and Salmonella spp. (Reyneke et al. 2020; 2023), and K. pneumoniae and P. aeruginosa (Denissen et al. 2023), that may be present in the water. Overall, results indicate that surface water sources can be contaminated by a multitude of microorganisms that may pose a health risk to community members, including more pristine water sources such as the water collected from the Klipriviersberg Nature Reserve. Community members therefore need to be educated regarding the risks posed by untreated surface water sources for domestic activities or recreational activities, while active monitoring of water bodies used for recreational activities need to be prioritised by authorities to inform public safety.

The detection and identification of E. coli and other bacterial pathogens from recreational waters in Gauteng is a crucial aspect of ensuring public health and safety. This study shows that the surface waters in question were not fit for recreational purposes in the given months of analysis. In many countries such as Canada, China, India and the United States, surface waters are used for bathing and swimming with several cases of gastrointestinal illness being reported due to recreational activities in contaminated waters. It is important to note that infection by antimicrobial resistant bacteria is dependent upon the concentration of the bacteria ingested, host immunity and residual antibiotics within the intestinal tract. While epidemiological studies may provide risk assessment based on clinical data, QMRA can predict illnesses at non-sewage impacted waters.

Currently, water quality guidelines such as USEPA, EPA and even South African water quality guidelines indicate the faecal and related bacterial limits for surface water quality but do not include monitoring and action plans for antimicrobial resistant bacteria. In addition, there are no dose response models to quantify the risk of infection due to ingestion of resistant bacteria, however there have been studies that proposed a risk assessment framework to determine the health risk based on resistant bacteria. A recommendation for future QMRA work would also be to focus on the ARB genes as surrogates for pathogens when estimating health risks. By monitoring the levels of E. coli and other indicator organisms and their antimicrobial resistance profiles in surface waters, authorities and researchers can gain valuable insights into the extent of faecal contamination and the potential health risks it poses. With constant monitoring of these water bodies, adequate information can be gathered which can then be used to implement appropriate measures to safeguard public health, but also provides valuable data for decision-makers to implement necessary measures for water treatment and management. By staying ahead of potential health risks associated with contaminated waters, authorities can ensure that recreational areas remain safe for everyone to enjoy.

The team would like to express their heartfelt gratitude to Riahaanah Paulse-Thomas, Dr Lee Heine and Xylan de Jager for their assistance and technical support.

The team would like to express their heartfelt gratitude to the NRF (Grant number PSTD2205056738) for the funding.

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

The authors declare there is no conflict.

Aranda
K. R. S.
,
Fagundes- Neto
U.
&
Scaletsky
I. C.
(
2004
)
Evaluation of multiplex PCR's for diagnosis of infection with Diarrheagenic Escherichia coli and Shigella spp
,
Journal of Clinical Microbiology
,
42
,
5849
5853
.
Barnes
V. L.
,
Heithoff
D. M.
,
Mahan
S. P.
,
House
J. K.
&
Mahan
M. J.
(
2023
)
Antimicrobial susceptibility testing to evaluate minimum inhibitory concentration values of clinically relevant antibiotics
,
STAR Protocols
,
4
(
3
),
102512
.
Cabral
J. P.
(
2010
)
Water microbiology. Bacterial pathogens and water
,
International Journal Of Environmental Research And Public Health
,
7
(
10
),
3657
3703
.
Clinical Laboratory Standards Institute
. (
2018
)
M100 performance standards for antimicrobial susceptibility testing; 28th Edition, Available at: https://www.calameo.com/read/007001748ccef4b9fdfe5.
Delair
Z.
(
2017
)
Implementation of Molecular Methods for the Detection and Characterization of Pathogenic E. Coli: Industrial and Routine Monitoring Applications
.
M-tech dissertation
,
University of Johannesburg
.
Delair
Z.
(
2021
)
Genotypic, Phenotypic and Phylogenetic Characterization of E. Coli
.
D-Tech thesis
.
University of Johannesburg
.
Ebomah
K.
,
Adefisoye
M. A.
&
Okoh
A.
(
2018
)
Pathogenic escherichia coli strains recovered from selected aquatic resources in the Eastern Cape, South Africa, and its Significance to Public Health
,
International Journal of Environmental Resources and Public Health
,
15
(
7
),
1056
.
Edokpayi
J. N.
,
Odiyo
J. O.
&
Durowoju
O. S.
(
2017
)
Impact of Wastewater on Surface Water Quality in Developing Countries: A Case Study of South Africa. Water Quality. InTech. doi: 10.5772/66561
.
European Committee on Antimicrobial Susceptibility Testing
. (
2018
)
EUCAST clinical breakpoint tables v. 8.1. Available at: https://www.eucast.org/ast_of_b acteria/previous_versions_of_documents.
Fischer
F.
,
Hamdan
Y.
,
Hansson
A.
,
Josefsson
E.
&
Sundström
L.
(
2019
)
Water Quality Modelling and Quantitative Microbial Risk Assessment of Msunduzi River
.
Thesis
,
Chalmers University of Technology
,
Gothenburg, Sweden
.
Haas
C. N.
,
Rose
J. B.
&
Gerba
C. P.
(
1999
)
Quantitative Microbial Risk Assessment
.
New York, NY
:
John Wiley & Sons, Inc
.
Kimata
K.
,
Shima
T.
,
Shimizu
M.
,
Tanaka
D.
,
Isobe
J.
,
Gyobu
Y.
,
Watahiki
M.
&
Nagai
Y.
(
2005
)
Rapid categorization of pathogenic Escherichia coli by multiplex PCR
,
Microbiology Immunology
,
49
(
6
),
485
492
.
Kusi
J.
,
Ojewole
C. O.
,
Ojewole
A. E.
&
Nwi-Mozu
I.
(
2022
)
Antimicrobial resistance development pathways in surface waters and public health implications
,
Antibiotics (Basel, Switzerland)
,
11
(
6
),
821
.
López-Saucedo
C.
,
Cerna
J. F.
,
Villegas-Sepulveda
N.
,
Thompson
R.
,
Velazquez
F. R.
,
Torres
J.
,
Tarr
P. I.
&
Estrada-Garcia
T.
(
2003
)
Single multiplex polymerase chain reaction to detect diverse loci associated with diarrheagenic Escherichia coli
,
Emerging Infectious Diseases
,
9
(
1
),
127
131
.
Mbene
A. B.
,
Houreld
N. N.
&
Abrahamse
H.
(
2009
)
DNA damage after phototherapy in wounded fibroblast cells irradiated with 16 J/cm2
,
Journal of Photochemistry and Photobiology
,
94
,
131
137
.
Mons
M. N.
,
Van Der Wielen
J. M. L.
,
Blokker
E. J. M.
,
Sinclair
M. I.
,
Hulshof
K. F. A. M.
,
Dangendorf
F.
,
Hunter
P. R.
&
Medema
G. J.
(
2007
)
Estimation of the consumption of cold tap water for microbiological risk assessment: An overview of studies and statistical analysis of data
,
Journal of Water and Health.
,
5
(
1
),
151
170
.
Moses
A. E.
,
Garbati
M. A.
,
Egwu
A. O.
&
Ameh
E. J.
(
2006
)
Detection of E. coli 0157 and 026 serogroups in human immunodeficiency virus-infected patients with clinical manifestation of diarrhoea in Maiduguri, Nigeria
,
Research Journal of Medicine and Medical Sciences
,
1
(
4
),
140
145
.
Nappier
S. P.
,
Liguori
K.
,
Ichida
A. M.
,
Stewart
J. R.
&
Jones
K. R.
(
2020
)
Antibiotic resistance in recreational waters: State of the science
,
International Journal of Environmental Research and Public Health
,
17
(
21
),
8034
.
Ngubane
Z.
,
Bergion
V.
,
Dzwairo
B.
,
Troell
K.
,
Amoah
I. D.
,
Stenstrom
T. A.
&
Sokolova
E.
(
2022
)
Water quality modelling and quantitative microbial risk assessment for uMsunduzi River in South Africa
,
Journal of Water and Health
,
20
(
4
),
641
.
Pang
Z.
,
Raudonis
R.
,
Glick
B. R.
,
Lin
T. J.
&
Cheng
Z.
(
2019
)
Antibiotic resistance in Pseudomonas aeruginosa: Mechanisms and alternative therapeutic strategies
,
Biotechnology Advances
,
37
,
177
192
.
Pass
M. A.
,
Odedra
R.
, &
Batt
R. M.
(
2000
)
Multiplex PCR for identification of Escherichia coli virulence genes
,
Journal of Clinical Microbiology
,
38
,
2001
2004
.
Poirel
L.
,
Madec
J. Y.
,
Lupo
A.
,
Schink
A. K.
,
Kieffer
N.
,
Nordmann
P.
&
Schwarz
S.
(
2018
)
Antimicrobial resistance in Escherichia coli
,
Microbiology Spectrum
,
6
(
4
),
0026
2017
.
Prinzi
A.
(
2023
)
Updating Breakpoints in Antimicrobial Susceptibility Testing, American Society for Microbiology
.
(Accessed: 10 July 2024)
.
Ramaite
K.
,
Ekwanzala
M. D.
&
Momba
M. N. B.
(
2022
)
Prevalence and Molecular Characterisation of Extended-Spectrum Beta-Lactamase-Producing Shiga Toxin-Producing Escherichia coli, from Cattle Farm to Aquatic Environments, Pathogens (Basel, Switzerland), 11(6), 674
.
Reyneke
B.
,
Hamilton
K. A.
,
Fernández-Ibáñez
P.
,
Polo-López
M. I.
,
McGuigan
K. G.
,
Khan
S.
&
Khan
W.
(
2020
)
EMA-amplicon-based sequencing informs risk assessment analysis of water treatment systems
,
Science of the Total Environment
,
743
,
140717
.
Reyneke
B.
,
Khan
S.
&
Khan
W.
(
2023
)
Social perception and risk assessment of domestic uses of solar pasteurized rainwater in an informal settlement community
,
Journal of Environmental and Chemical Engineering
,
11
,
110532
.
Ryan
M. O.
,
Haas
C. N.
,
Gurian
P. L.
,
Gerba
C. P.
,
Panzl
B. M.
&
Rose
J. B.
(
2014
)
Application of quantitative microbial risk assessment for selection of microbial reduction targets for hard surface disinfectants
,
American Journal of Infection Control
,
42
(
11
),
1165
1172
.
Santos
A. C.
,
Santos
F. F.
,
Silva
R. M.
&
Gomes
T. A.
(
2020
)
Diversity of hybrid- and hetero-pathogenic Escherichia coli and their potential implication in more severe diseases
,
Frontiers in Cellular and Infection Microbiology
,
10
,
1
11
.
Schoen
M. E.
,
Xue
X.
,
Hawkins
T. R.
&
Ashbolt
N. J.
(
2014
)
Comparative human health risk analysis of coastal community water and waste service options
,
Environmental Science and Technology
,
48
,
9728
9736
.
South African Water Quality Guidelines
. (
1996
)
(second edition), Volume 2: Recreational Use
.
Tao
S.
,
Chen
H.
,
Li
N.
,
Wang
T.
&
Liang
W.
(
2022
)
The Spread of Antibiotic Resistance Genes In Vivo Model, The Canadian journal of infectious diseases & medical microbiology, 3348695
.
Tarr
C. L.
,
Large
T. M.
,
Moeller
C. L.
,
Lacher
D. W.
,
Tarr
P. I.
,
Acheson
D. W.
&
Whittam
T. S.
(
2002
)
Molecular characterization of a serotype 0121:H19 clone, a distinct Shiga toxin-producing clone of pathogenic Escherichia coli
,
Infection and Immunity
,
70
(
12
),
6853
6859
.
United States Environmental Protection Agency (USEPA)
. (
2011
)
Exposure factors handbook: 2011 Edition. National Center for Environmental Assessment, Washington, DC; EPA/600/R-09/052F. Available from the National Technical Information Service, Springfield, VA, and online at: http://www.epa.gov/ncea/efh.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).

Supplementary data