Prevalence and characterization of Escherichia coli in the Kelantan River and its adjacent coastal waters

Escherichia coli is a coliform subgroup commonly found in the intestine of both humans and animals. It is used as a marker for fecal pollution in aquatic system (Byamukama et al. 2005) and as an indicator to assess food hygiene and food safety. E. coli are mostly non-harmful but there are some pathogenic strains. These are responsible for infections of the human digestive system (e.g. enteroinvasive E. coli (EIEC), enteropathogenic E. coli (EPEC), enterotoxigenic E. coli (ETEC), enteroaggregative E. coli (EAEC), enterohaemorrhagic E. coli (EHEC), and diffusely adherent E. coli (DAEC)) (Hamelin et al. 2006), and human extraintestinal infections, such as nosocomial bacteremia, urinary tract infection, and myositis (Galvin et al. 2010). Most of these pathotypes are of public health concern because their infectious dose is low and they are easily transmitted via food and water (Matic et al. 1997).

E. coli can be categorized into phylogroups A, B1, B2, and D based on the existence or absence of the following genes: (i) chuA (required for EHEC heme transport); (ii) jyaA (unknown function); and (iii) TSPE4.C2 (unknown DNA segment) (Clermont et al. 2000). Most commensal and obligatory pathogens (EHEC, ETEC, and EIEC) are found in phylogroup A and B1. Group A is predominantly found in humans, whereas group B1 is predominantly in animals. Phylogenetic group B2 and D are mostly virulent extraintestinal strains (Luna et al. 2010), whereas the virulent groups responsible for chronic and mild diarrhea are spread across the four phylogenetic groups. Phylogroups A and B1 are persistently found in environment and have been recognized as emerging gastrointestinal strains that are closely linked to antibiotic resistance (Escobar-Páramo et al. 2004).

In this study we focused on the Kelantan River, which is 248 km long and is formed by the confluence of the Galas and Lebir Rivera that run across the state of Kelantan in Peninsular Malaysia. It has a catchment area of 13,100 km² and is dominated by sedentary soils (hills) and alluvial soils (riverine floodplains) (Adnan & Atkinson 2011). The river passes through four main townships of Kota Bharu, Kuala Krai, Pasir Mas, and Tumpat prior to discharge into South China Sea (Pradhan & Youssef 2011). The river flow is affected by monsoon rainfall throughout the year, with maximum annual rainfall capacity of 1,750 mm during the northeast monsoon season (November to January).

For Kelantan local residents, the Kelantan River serves as an important water source for domestic use, transportation, irrigation, sand mining, agriculture, and aquaculture industries. To date, there are about 128 sand mining operations along the Kelantan River, causing the water of the Kelantan River to turn turbid since the early 1990s (Yen & Rohasliney 2013). E. coli and total coliform in the Kelantan River are frequently found at concentrations that exceed limits set by the Interim Marine Water Quality Standard (IMWQS) and the Water Quality Index (Basri et al. 2015; Bamaiyi et al. 2017). This may result in the higher incidence rates of food and water-borne diseases in Kelantan relative to other states in Malaysia. Being in the monsoon belt, the incidence of flooding in Kelantan is also frequent and the high flood waters can overwhelm sanitation systems, stir up rivers, and increase occurrence of both E. coli and total coliform between the Kelantan River Delta and its adjacent coastal waters (Ahmad et al. 1996; Vignesh et al. 2013).

As there is growing concern about the occurrence of E. coli in aquatic environment (Ghaderpour et al. 2015; Sen et al. 2019), it is good to have a better understanding of E. coli epidemiology and ecology in the Kelantan River. This study aimed to determine the distribution and genetic diversity of E. coli and their association with environmental variables in Kelantan River and its adjacent coastal waters.

Twenty samplings were carried out regularly at 15 sites along the Kelantan River and its adjacent coastal sites from November 2014 until August 2016 (total number of samples = 300) (Figure 1). Replicate water samples were collected in a sterile Schott bottle and kept cold for no more than 6 h until processed in the laboratory. In situ parameters, including surface water temperature and conductivity, were measured using a conductivity, meter (YSI-30, USA) and a pH meter (Thermo Scientific Orion 4-Star, USA) was used to record pH. For dissolved oxygen (DO) measurements, water samples were collected with DO bottles and fixed instantly with manganous chloride and alkaline iodide solutions. The mixtures were then mixed and delivered to the laboratory for analysis. The DO concentration was calculated based on the Winkler titration method (Grasshoff et al. 1999).

For total suspended solids (TSS), sample volume was determined and filtered through a glass microfiber filter/F grade (GF/F) filter paper. TSS was measured as the weight increase of filter after drying at 70 °C and particulate organic matters (POM) was calculated based on weight loss after combustion in a microwave furnace (model MF-05, USA) at 500 °C for 3 h. Dissolved inorganic nutrients (nitrate (NO₃), nitrite (NO₂), ammonium (NH₂), phosphate (PO₄) and silicate (SiO₄)) were analyzed according to Parsons et al. (1984).

The isolation and enumeration of coliform and E. coli in water samples were carried out using a membrane filtration technique. The membrane was placed on CHROMagar ECC for E. coli and coliforms (in triplicates) and incubated at 37 °C overnight (modified Method EPA 1604). Presumptive E. coli (blue colonies) on CHROMagar ECC were selected and purified on CHROMagar ECC before storing in glycerol solution. Enumeration of E. coli was carried out by counting the number of blue colonies, and colonies with mauve color were counted as total coliform. Abundance of total coliform and presumptive E. coli were expressed as colony forming units per 100 milliliters (CFU 100 mL⁻¹).

Polymerase chain reactions (PCR) for the detection of phoA gene (housekeeping gene) was carried out using primers PhoA-F (5′-GTCACAAAAGCCCGGACACCATAAATGCCT-3′) and PhoA-R (5′-TACACTGTCATTACGTTGCGGATTTGGCGT-3′). This PCR amplifies a 903 bp fragment, which was used to confirm the presumptive E. coli (Yu & Thong 2009). Twenty-five microliters of reaction mixture consisting of with 1X Green GoTaq reaction buffer (pH 8.5), 0.5 U of Taq DNA polymerase (Promega, USA), 1 mM of MgCl₂, 140 μM of each deoxynucleotide triphosphate (dNTP), 0.1 μM of both primers and 5.0 μl of DNA. The PCR was carried out in a 2,720 thermal cycler (Applied Biosystems, Singapore) with the following conditions: initial denaturation at 94.0 °C for 2 min, 40 cycles at 92 °C for 0.5 min, 59 °C for 0.5 min, 72 °C for 0.5 min; and a final extension for 5 min at 72 °C. E. coli ATCC 25922 was used as the positive control and sterile water was included as the negative control.

Triplex PCR was employed to assess E. coli phylogroups (A, B1, B2, and D) (Table 1) (Clermont et al. 2000). Primer pairs used were ChuA.F (5′-GACGAACCAACGGTCAGGAT-3′) and ChuA.R (5′-TGCCGCCAGTACCAAAGACA-3′); YjaA.F (5′-TGAAGTGTCAGGAGACGCTG-3′) and YjaA.R (5′-ATGGAGAATGCGTTCCTCAAC-3′); and TspE4C2.F (5′-GAGTAATGTCGGGGCATTCA-3′) and TspE4C2.R (5′-CGCGCCAACAAAGTATTACG-3′), which generated 279, 211, and 152 bp fragments, respectively. Characterization of E. coli was based on the existence and absence of ChuA, YjaA, and TspE4.C2 (Table 1). Twenty microliters of reaction consisting of 1X Green GoTaq buffer (pH 8.5), 0.5 U of Taq DNA polymerase (Promega, USA), 120 μM dNTP, 1.5 mM of MgCl₂, 5 μL of DNA, 0.4 μM of jyaA primers and 0.24 μM of each ChuA and TSPE4-C2 primers was used for identification of phylogenetic grouping. The conditions for the PCR were initial denaturation at 94 °C for 5 min, 30 cycles of 94 °C for 0.5 min, 55 °C for 0.5 min, and 72 °C for 0.5 min; final extension at 72 °C for 7 min.

All the validated E. coli were subjected to two multiplex PCR assays (M1 and M2) to detect for virulence genes linked to E. coli strains that cause intestinal disease (Chapman et al. 2006). Twenty-five microliters of reaction mixture consisting of 1× Green GoTaq buffer (pH 8.5), 0.5 U of Taq DNA polymerase (Promega, USA), 1.65 mM of MgCl₂, 220 μM dNTP, 0.24 μM of each primer, and 5 μl DNA was used for both PCR assays. Primer sequence and PCR amplification conditions were adopted from Gómez-Duarte et al. (2009). The primer sequences used for identification of virulence genes in this study are shown in Table 2 and PCR was carried out under the following conditions: initial denaturation at 94 °C for 2 min, 40 cycles at 92 °C for 0.5 mi, 59 °C for 0.5 min, and 72 °C for 0.5 min; and final extension at 72 °C for 5 min. Six positive controls (E. coli 2060-004, E2348/69, JM221, E9034A, C1845, and EC-12) were used and sterile water was used as the negative control.

PAST software was employed to carry out statistical analysis (Hammer et al. 2001), and p < 0.05 or at 95% confidence interval were accepted as significant. Values were compared using variance and Tukey's test analysis, and relationships between variables measured were tested using correlation and linear regression analyses. Principal component analysis (PCA) was used to determine physicochemical variables, which affect the water quality of Kelantan River, estuaries, and its adjacent waters. Canonical-correlation analysis (CCA) was used to compare the categories of variables measured and phylogenetic groups, and the relationship between the categories is represented in a two-dimensional graph.

Surface water temperature was relatively invariant among the sites, ranging from 28.8 ± 2.0 °C to 31.0 ± 2.2 °C (coefficient of variation, CV = 3%). Salinity was low (<1) in the Kelantan River and its tributaries, but increased in the estuaries (4.4 ± 4.9–11.9 ± 11.3) and its adjacent coastal sites (22.3 ± 8.9–30.5 ± 2.1). The pH ranged from 7.1 ± 0.5 to 8.2 ± 0.3 (CV = 5%), and DO concentration was at healthy levels (>270 μM, CV = 8%). TSS and POM concentrations fluctuated between 55 ± 28–300 ± 241 mg/L and 17 ± 18–52 ± 36 mg/L, respectively, and were higher in the river (>150 mg/L TSS and >30 mg/L POM) compared to the estuaries and coastal sites [significance was measured using Student's t-test with 104 degrees of freedom: t(104) = 6.61, p < 0.001; t(104) = 5.71, p < 0.001] (Table 3). Of the three nitrogen species measured, NO₃ was the dominant species and its average concentrations varied over a wide range from 0.03 to 0.61 mg/L (CV = 37%), with higher concentrations detected in the river compared to the estuaries and coastal sites. A high NO₃ level was observed in Krai (0.61 mg/L). NH₄ concentrations were low among sites, and varied from 0.01 to 0.09 mg/L with the exception of Kota Bahru (0.18 mg/L); whereas NO₂ concentrations were generally low (<0.02 mg/L). For the other inorganic nutrients measured, SiO₄ fluctuated over a wider range between 1.09 and 24.24 mg/L. High SiO₄ levels were observed upstream with the highest concentration observed at Merbau. For PO_4, concentrations were low and varied from 0.01 to 0.03 mg/L (Figure 2).

Coliform and E. coli were found at all sampling sites (Figure 3(a)). The average coliform concentrations ranged from 1.0 × 10³ to 3.3 × 10⁴ CFU 100 mL⁻¹. However, the beaches of Melawi and Bachok, and the estuary at Semarak exhibited coliform concentrations that were one order of magnitude higher (2.01 × 10⁵, 3.22 × 10⁵ and 3.15 × 10⁵, CFU 100 mL⁻¹, respectively). For E. coli, the average abundance ranged from 3.1 × 10¹ to 1.6 × 10⁵ CFU 100 mL⁻¹. The highest abundance of E. coli was detected in river water at Kota Bahru, whereas for estuaries and coastal sites, the highest E. coli abundance was observed in the Semarak River.

In this study, a total of 2,341 E. coli were isolated. In the Kelantan River and its adjacent coastal waters, all four phylogenetic groups were observed (Figure 3(b)). Phylogenetic group B1 (n = 695, 29.7%) was the most dominant, followed by groups A (n = 688, 29.4%), D (n = 597, 25.5%), and B2 (n = 361, 15.4%). The phylogenetic groups were not distributed homogeneously among the sites. Phylogenetic group A was more frequently isolated from rivers and estuaries compared to its adjacent coastal sites. Higher E. coli phylogenetic group A counts were observed upstream at Kampung Batu Udang (n = 86) followed by Manik Urai (n = 62). The counts decreased until Tanah Merah (n = 28), before increasing from the downstream site of Pasir Mas (n = 70) to the estuaries where the highest count observed was at Kuala Besar (n = 129). In contrast, phylogroup B1 was most abundant in Merbau (n = 91) and the counts decreased downstream before increasing at Kota Bahru to the estuary site KW16. In comparison, group B1 dominated in the coastal sites. For phylogenetic group B2 and D, which were the least abundant, the highest counts were observed at Kampung Batu Udang (n = 70 and n = 103, respectively) and the counts were lower at coastal waters compared to river waters. In this study, pathogenic strains (EHEC, EPEC, EAEC, ETEC, DAEC, or EIEC) were not detected. However, one E. coli isolate from Manik Urai carried the eae gene whilst another from Kampung Batu Udang possessed the LT gene.

A relatively stable temperature pattern typical of tropical water was observed for all sampling sites. Salinity varied greatly between river, estuaries and coastal water but were in the range previously reported for other riverine systems in Malaysia (You et al. 2016; Lim et al. 2018). High TSS is the pervasive water quality problem in Malaysia (Dow 1995). The high TSS concentration observed at the Kelantan River contained silt and clay, which can be attributed to the upstream deforestation activities, sand mining activities, agriculture, livestock husbandry, and dredging operations (Yen & Rohasliney 2013). This resulted in a turbid and brownish river water. The floating fine silt and detritus from the catchment area carried by rainwater during monsoon season (Prasanna & Ranjan 2010) also increased the TSS and affected the water quality in the Kelantan River Delta (Table 3). In this study we observed that TSS did not correlate with DO (r² = 0.001, p > 0.10), suggesting that the TSS in the Kelantan River had generally low organic content (<20%).

NO₃ concentration in this study was higher than NO₂ [t(216) = 15.07, p < 0.001] and NH₄ [t(216) = 8.72, p < 0.001], similar to that reported by Yen & Rohasliney (2013). The detectable concentrations were within the criterion (7 mg L⁻¹) set by the Malaysian Interim National Water Quality Standard (INWQS, Department of Environment 2019). Agriculture is the second largest sector and made up 24.6% or USD1.2 billion of the gross domestic product of Kelantan in 2016 (Department of Statistics Malaysia 2017). Over the last few decades, intensive land use together with technological development have led to large-scale planting of commercial crops (rubber, oil palm, tobacco) and other agriculture farming, which directly increased the use of fertilizers, manure, and soil in Kelantan (Samsurijan et al. 2018). Furthermore, anthropogenic activities associated with domestic sewage and waste effluents containing nitrogenous compounds, may also contribute to elevated levels of NO₃ in the river and its adjacent waters (Yen & Rohasliney 2013; Shamsuddin et al. 2016). In this study, the NH₄ concentration was lower than the NO₃ concentration indicating less impact of industrial effluents along the Kelantan River, estuaries and its adjacent coasts. For SiO₄, high concentrations were detected in the river (>13 mg L⁻¹) and the concentrations decreased in the estuaries–coastal waters sites. The main source of SiO₄ in the Kelantan River is the mining activities that begin after the convergence of the two tributaries (Galas River and Lebir River) and continues until the estuary delta of the Kelantan River, which severely impacts the transport and displacement of river sediments (Yen & Rohasliney 2013). Besides natural and chemical weathering of sedimentary soils and rocks, human activities are also sources of SiO₄ in the river (Shaari et al. 2015; Wang et al. 2017). The decrease in SiO₄ concentration may be attributed to dilution, lack of silica enrichment, utilization of silica by aquatic organisms (e.g. diatoms), and plants that grow along the river. In this study, the overall inorganic nutrients detected were in the range reported in tropical and sub-tropical waters (Lee & Bong 2006; Sakai et al. 2016).

Coliform and E. coli were highly prevalent in the Kelantan River surface waters, estuaries, and its adjacent coasts. The average abundance of E. coli detected at all sampling sites exceeded both the recommended E. coli allowable limit by NWQS of class II for rivers in Malaysia (100 CFU/100 ml) and Malaysia IMWQS (Department of Environment 2019) except KW19. This is indicative of fecal pollution, which is consistent with previous studies on the Kelantan River (Bamaiyi et al. 2015; Basri et al. 2015). Fecal pollution here is mainly due to the direct untreated sewage discharge from the houses and floating toilets built along the riverbanks. Furthermore, there is inadequate sewage treatment facilities in Kelantan where the use of individual septic tanks that connect to multi-points of sewage treatment plants (SPAN 2016) only partially treat the sewage before being discharged into the river (Sakai et al. 2016). This could subsequently cause river water quality deterioration. In this study, we found that the average abundance of E. coli in estuarine and coastal waters was one order of magnitude higher compared to the river waters. This may be caused by the untreated sewage released directly from the houses and toilets built near the beach and the coastal upwelling process, which provides a nutrient source and support for E. coli. Furthermore, E. coli can rapidly adapt to, and tolerate, different abiotic (availability of nutrients, pH, moisture, temperature, salinity) and biotic (grazing) stress factors (Van Elsas et al. 2011; Alves et al. 2014), and could further enhance their fitness in aquatic environments. Studies have shown the capability of E. coli to grow and proliferate in marine environments through alkaline pH adaptation (Hughes 2008) and or change in their genetic structure (Van Elsas et al. 2011). The prevalence of E. coli in coastal, estuaries, and river waters of Kelantan may pose health risks for the local residents who have direct or indirect contact with water through recreational activities or seafood consumption.

Our present study showed that E. coli group B1 was the most prevalent followed by A, D, and B2, of which groups B1 and A comprised 59% of the total E. coli in this study. Our findings are similar to previous studies that reported environmentally persistent E. coli are of groups B1 and A rather than virulent types B2 and D (Figueira et al. 2011; Ghaderpour et al. 2015). At upstream sites of the Kelantan River, higher abundances of D phylogroup strains were observed suggesting that fecal contamination was mainly from animal origin. Studies have shown that the E. coli population structure differs significantly between humans and animals (Carlos et al. 2010). It has been reported that livestock and poultry are the main reservoir for D phylogroup. Based on statistics from the Department of Veterinary Services Malaysia (2016), poultry farming, with a total of 1.84 million population, is the major livestock activity operating in Kelantan. Poultry could therefore be a major animal fecal pollution source in the river. The difference between phylogenetic groups among the sites in this study may be attributed to hydrological conditions, different sources of pollution, selective pressures in the waters, and land use (Lyautey et al. 2010; Van Elsas et al. 2011). Distinct survival rates, together with all these parameters, will structure the E. coli community distribution and diversity in aquatic environments (Berthe et al. 2013).

In this study, PCA implied that TSS and SiO₄ were the elements that influence the Kelantan River estuaries and its adjacent coastal water quality. PCA 1 and PCA 2 explained 72.10% and 19.75% of the total variance of the explanatory physical variables measured (Figure 4(a)). Our results showed that TSS correlated with E. coli abundance (r² = 0.165, p < 0.001). The positive correlation between E. coli and TSS suggested that E. coli was transported in the river bound to particulate matter. Research indicates that the attachment of E. coli to sediment organic matter with clay content can increase their survival in aquatic environments (Pachepsky & Shelton 2011; Liang et al. 2017). Suspended solids not only provide organic and inorganic nutrients but also provide protection against adverse factors (ultraviolet radiation, metal toxicity, grazing, attack by bacteriophage) (Medema et al. 2003). On the other hand, secretion of extracellular polymeric substances by microorganisms at the outer cell surface (Liao et al. 2015) for cell aggregation, adhesion, and protection is one of the survival strategies for cells to survive and adapt in hostile environments (Vu et al. 2009; Bruckner et al. 2011). TSS also correlated with rainfall (r² = 0.342, p < 0.001). This finding concurs with other studies that showed rainfall is the primary process affecting river volume and flow, which can directly increase the level of TSS through runoff (Shen & Julien 1993; Vaze & Chiew 2003). Our study therefore suggested that rainfall indirectly affects the distribution and abundance of E. coli in the Kelantan River, estuaries and its adjacent coasts.

In order to illuminate the factors influencing the E. coli phylogroups occurrence and distribution, a CCA was conducted (Figure 4(b)). Our CCA showed distinct differences in survival among strains belonging to different phylogenetic groups. Phylogenetic group A was greater in deteriorated water containing NH₄ and NO₂, whereas phylogenetic group D was greatest with NO₃. In contrast, E. coli phylogenetic group B2 seemed to thrive in waters with higher DO. Abundance of phylogenetic group B1 appeared to have better salinity tolerance compared to other phylogenetic groups. This explained why phylogenetic group B1 dominated at coastal sites, whereas phylogenetic group D dominated upstream of Kelantan River with higher concentration of NO₃. However, more research is needed to validate these findings.

Our study of the Kelantan River, estuaries, and its adjacent coastal waters showed that NO₃ was the dominant nitrogen species and PCA analysis demonstrated that TSS and SiO₄ were the physicochemical factors that influence the water quality. The coliform and E. coli counts detected in this study exceeded INWQS and Malaysian Marine Water Quality Standards, suggesting the prevalence of fecal pollution. TSS was significantly correlated with E. coli abundance and rainfall. All four E. coli phylogroups were detected, but most were commensal groups B1 and A. Phylogenetic group D and B2 were the least abundant. CCA analysis demonstrated that phylogenetic group B2 seemed to thrive in water with higher DO. However, phylogenetic groups A and D were greater in deteriorated water containing NH₄, NO₂ and NO₃, whereas phylogenetic group B1 appeared to have better salinity tolerance among the phylogenetic groups.

This work was supported by the Institution Centre of Excellence Phase II Fund, Ministry of Higher Education (grant number: IOES-2014D), University of Malaya Centre of Excellence (grant number: RU009D-2015 and TU001F-2018), and Bilateral Cooperation of Maritime Affairs of China (contract no: HC140502).