Survival of Pathogenic Escherichia coli Strains in Sand Subjected to Desiccation

Abstract

Shiga toxin-producing Escherichia coli (STEC) and enteropathogenic E. coli (EPEC) are E. coli pathovars of particular relevance to infant health. While the intestinal tract of humans and animals constitutes their primary habitat, these bacteria can also persist in natural environments such as sand. The aim of this study was to evaluate the persistence of STEC and EPEC strains in sand microcosms under controlled conditions of heat and desiccation in order to estimate their viability in this matrix and provide evidence regarding the potential risks associated with the use of sandboxes in public spaces. The study included STEC strains belonging to clinically important serotypes (O26:H11, O103:H2, O111:H8, O121:H19, O145:NM, O157:H7 and O174:H28), animal-derived EPEC strains, and a non-pathogenic E. coli strain (NCTC 12900). The strains were inoculated into sterile sand microcosms and maintained at 37 °C. Death curves, persistence in the matrix, presence of virulence genes, and ability to produce biofilm were evaluated. The death and persistence curves varied by serotype; some strains remained viable in the viable but non-culturable state for extended periods. All strains retained their virulence-associated genetic markers throughout the assays. None of the STEC strains was classified as a biofilm producer under the experimental conditions, whereas the two EPEC strains were identified as weak and moderate biofilm producers. However, no association was found between biofilm formation and persistence in the matrix. The findings provide an initial approach and provide relevant evidence of the capacity of STEC and EPEC strains to survive in sand, which could represent a potential risk in recreational environments.

1. Introduction

Escherichia coli is a bacterium of the Enterobacteriaceae family and a common constituent of the intestinal microbiota in both humans and animals. Certain strains have acquired mobile genetic elements that encode virulence factors capable of disrupting host cellular processes, giving rise to what are known as diarrhoeagenic Escherichia coli (DEC) [1]. Among these, Shiga toxin-producing E. coli (STEC) and enteropathogenic E. coli (EPEC) are of particular concern due to their significant impact on infant health.

Animal and human intestines are considered the primary habitat of E. coli, whereas its secondary habitat is the external environment [2]. Most strains retain, within their core genome, genetic determinants that confer a remarkable adaptive capacity, allowing them to utilise a wide range of nutrient sources and persist in external environments. It has been proposed that this species exhibits a biphasic lifestyle, alternating between host-associated and free-living phases. The bacterium is excreted via faeces, where it can proliferate rapidly in fresh matter in the presence of oxygen [3,4] and may subsequently either die or persist for variable periods in soil, water, or sediments until a new host is encountered. The survival of E. coli in natural environments can be influenced by both abiotic and biotic factors [5]. Under these conditions, E. coli may enter a “viable but non-culturable” (VBNC) state, in which cells cannot be readily recovered using standard laboratory media, despite remaining viable [2]. VBNC E. coli cells may regain their ability to grow and recover their pathogenic potential, thus posing a potential risk to public health [6].

Sand has been documented as a complex matrix that harbours microbiota, among which coliforms are particularly prominent [7,8,9,10]. In addition, numerous studies have examined the survival of E. coli and the dynamics of contamination on beaches [8,11]. Notably, high coliform loads have been detected on beaches during the summer season, persisting for several months, particularly in dry sand [12]. Studies have also investigated the presence of coliforms in park sand [13,14]. However, none of these works has addressed the survival of diarrhoeagenic E. coli pathovars under dry conditions.

The aim of the present study was to evaluate the behaviour of STEC and EPEC strains in sand microcosms subjected to heat and desiccation, to estimate their persistence in this matrix and to provide evidence regarding the potential risks associated with the use of sandboxes in public parks.

2. Material and Methods

2.1. Characteristics of the Strains Under Study

The study included STEC strains corresponding to serotypes with recognised relevance to human health (O26:H11, O103:H2, O111:H8, O121:H19, O145:NM, O157:H7, and O174:H28), as well as two EPEC strains isolated from animals: one from a synanthropic rodent of the species Rattus rattus (O88:H25) and another from a canine source (O187:H16). An E. coli strain negative for DEC molecular markers (NCTC 12900, National Collection of Type Cultures, London, UK) was used as a negative control. The virulence profiles of the strains used in this study are detailed in

Table 1. Genotypic profile of the strains under study.

ID	Serotype	Virulence Factors
FF6 1	O26:H11	stx1−/stx2+/eae+/ehxA+
KK11 1	O103:H2	stx1+/stx2−/eae+/ehxA+
GG7 1	O111:H8	stx1+/stx2−/eae+/ehxA+
CC3 1	O121:H19	stx1−/stx2+/eae+/ehxA+
GIV 2	O145:NM	stx1−/stx2+/eae+/ehxA+
D2253 1	O157:H7	stx1+/stx2+/eae+
F130 2	O174:H28	stx1−/stx2+/saa+/ehxA+
R27 2	O88:H25	stx1−/stx2−/eae+/bfp−
#633 2	O187:H16	stx1−/stx2−/eae+/bfp−
NCTC 12900	O157:H7	stx1−/stx2−/eae−

References: ¹ strain from the Statens Serum Institute ² strain from the Microbiology Laboratory (FCV-UBA).

.

2.2. Assessment of Strain Persistence

Sand was collected from a sandbox in the Autonomous City of Buenos Aires. The sand was sieved, and subsequently, a suspension was prepared by mixing 20 g of sand with 180 mL of distilled water. The mixture was homogenised using a stomacher. The pH of the suspension was measured using pH indicator strips (MColorpHast, Merck, Darmstadt, Germany), and coliform presence was assessed by inoculating 2 g into 4 mL of MacConkey broth (MCB; Oxoid, Hampshire, UK). The absence of STEC and EPEC strains was confirmed by PCR targeting the stx1, stx2, and eae genes [15,16], with primer sequences and PCR conditions provided in Table S1.

Sterile glass containers with perforated lids and filters were filled with 10 g of sand each. These were sterilised by autoclaving at 121 °C for 20 min and then dried in a stove at 37 °C. The dryness of the matrix was determined by daily weighing until a constant weight was observed over three consecutive days, indicating that the sand matrix had dried completely.

Bacterial suspensions in the logarithmic growth phase, adjusted to approximately 1 McFarland standard in sterile distilled water, were prepared. Each container was inoculated with 10 mL of suspension at a 1:1 (w/v) ratio, ensuring homogeneous dispersion of the inoculum throughout the matrix. The inoculum concentration was confirmed by viable cell counts.

Persistence of the strains was monitored daily starting immediately after inoculation. During the initial days, the sand matrix was moist, as determined by daily weighing. Once the matrix reached dryness (defined by stable weight over three consecutive days), the dry phase of the matrix began. Persistence data were collected throughout both the moist and dry phases.

Matrix dehydration was assessed daily using a Mettler P10 balance (No. 191442, capacity 10,000 g, made in METTLER TOLEDO, Giessen, Germany). The results of these measurements are presented in Table S2. The assays were conducted under controlled conditions at 37 °C and performed in triplicate.

To determine bacterial persistence, the death curve (DC) based on viable counts and microbial viability was evaluated throughout the experiment.

2.2.1. Determination of Death Curves of the Strains Under Study (Culturable Bacteria)

Following inoculation, a 2 g aliquot of the inoculated sand was collected daily and suspended in 10 mL of physiological saline. Viable counts were determined by preparing serial tenfold dilutions of the suspension and plating 0.1 mL of each dilution onto tryptic soy agar (TSA; Britania, Buenos Aires, Argentina) plates using a Drigalsky spatula. Plates were incubated at 37 °C for 24 h. Plates with 30 to 300 colony-forming units (CFU) were selected for counting, and CFU per mL in the suspension was calculated and converted to CFU per gram of sand. The daily averages were calculated from three replicates per strain.

Bacterial identity from the colonies was confirmed by classical biochemical tests specific to the Enterobacteriaceae family, including oxidase (Britania, Argentina), catalase, glucose oxidation–fermentation, and nitrate reduction to nitrite (Diatabs, Rosco Diagnostica A/S, Albertslund, Denmark), as well as species-level tests such as indole production, methyl red, Voges–Proskauer, and citrate utilisation. Unless otherwise specified, biochemical media were obtained from Britania (Argentina).

Daily monitoring of sand dehydration and viable counts allowed construction of death curves reflecting culturable bacterial persistence in both moist and dry matrix phases.

2.2.2. Assessment of Strain Viability (Detection of VBNC State)

Once no bacterial growth was observed on agar plates, a suspension was prepared by mixing 2 g of the inoculated sand with 10 mL of sterile double-distilled water. A volume of 200 µL of this suspension was inoculated, in triplicate, into tryptic soy broth (TSB, Britania, Argentina) and incubated at 37 °C for 18 h. Visual assessment of culture turbidity was employed as an indicator of bacterial growth.

Positive cultures were streaked onto TSA to verify purity, and bacterial identity was confirmed by using the same panel of biochemical tests described previously.

Growth in broth despite the absence of colonies on agar plates indicated that bacteria were in a viable but non-culturable (VBNC) state. The presence of VBNC cells was monitored daily under the same conditions.

2.2.3. Categorisation of Persistence

Strain persistence in the sand matrix was defined as the total duration during which bacteria remained either culturable (detected by viable counts on agar plates) or viable but non-culturable (VBNC, detected by growth in broth culture after loss of culturability on plates).

Persistence monitoring began immediately after inoculation, encompassing both the moist and the dry phases of the sand matrix, as determined by daily weight measurements. Based on this combined persistence (moist + dry phases), strains were categorised as follows: low persistence (total persistence ≤ 1 week), moderate persistence (1 week < total persistence ≤ 3 weeks), and high persistence (total persistence > 3 weeks).

2.3. Detection of Virulence Genes

The detection of virulence genes in the strains under study was carried out throughout all stages of the experiment. DNA was extracted from the TSA cultures used for the biochemical identification tests. Based on the genotypic profile, the genes stx1, stx2, eae, and ehxA were assessed by PCR following previously described protocols [15,16,17] with primer sequences and PCR conditions detailed in Table S2.

2.4. Biofilm Production Assay

Biofilm production was evaluated using a quantitative 96-well microplate assay following the method described by Cáceres et al. (2019) [18]. Briefly, the strains were grown in Luria–Bertani (LB) broth and incubated overnight at 37 °C with shaking. Cultures were then adjusted to 0.5 McFarland. A volume of 10 µL from each dilution was added to 190 µL of LB supplemented with NaCl, in triplicate, in lidded 96-well microplates, and incubated at 37 °C for 24 h. The medium was then replaced with 200 µL of LB containing glucose, and plates were further incubated for another 24 h under the same conditions. After two incubation periods (total 48 h), the plates were washed with distilled water (dH₂O), fixed with methanol, stained with 0.1% crystal violet aqueous solution, rinsed with water, air-dried, and eluted with ethanol under agitation. Optical density was measured at 570 nm (OD₅₇₀).

Results were interpreted by correcting the mean OD₅₇₀ of each strain (in triplicate) against the cut-off value (ODc), defined as three standard deviations above the mean OD₅₇₀ of a negative control strain (non-biofilm producer, QC148 EHEC). Based on the corrected OD values, strains were classified as follows: non-biofilm producers (OD ≤ ODc), weak biofilm producers (ODc < OD ≤ 2 × ODc), moderate biofilm producers (2 × ODc < OD ≤ 4 × ODc), and strong biofilm producers (OD > 4 × ODc). The assay was performed in triplicate.

3. Results

The analysis of the sand used as the matrix in the assays revealed a pH of 6.5, the presence of coliforms (lactose fermentation with acid and gas production in MacConkey broth), and the absence of STEC and EPEC (stx1−/stx2−/eae−). Following autoclaving, sterility was confirmed by culture, with neither viable nor VBNC forms detected.

Following inoculation, the sand microcosm reached dryness within 1–3 days, depending on the assay and strain. These data are summarised in Table S1. The results of the death curves and persistence of the strains under study in the sand microcosm varied according to serotype. Among STEC strains, culturability ranged from 4 to 18 days, while persistence in the VBNC state varied from 1 to 82 days post-inoculation, depending on the strain analysed. Four STEC strains exhibited high persistence in the matrix, one showed moderate persistence, and two demonstrated low persistence. Notably, STEC strain O174:H28 remained culturable until day 16 post-inoculation, with microbial counts falling below 100 CFU/mL from day 3 onwards and persisting in the VBNC state until day 63 post-inoculation.

In the case of the EPEC strains, both exhibited moderate persistence in the sand matrix, with culturability lasting up to day 15. Strain O88:H25 persisted an additional two days in the VBNC state.

The non-pathogenic strain NCTC 12900 also showed high persistence, with culturability lasting 4 days and viability as VBNC cells extending until day 44 post-inoculation. Death curve parameters, including culturability duration, time in the VBNC state, and overall persistence in the sand microcosm for each strain, are summarised in

Table 2. Survival and persistence of E. coli strains in sand microcosms.

Serotype	Pathovar	Duration of Culturability (Days Post-Inoculation)	Duration in VBNC State (Days Post-Inoculation)	Total Persistence in Sand Microcosm 1
O26:H11	STEC	6	1	Low
O103:H2	STEC	14	31	High
O111:H8	STEC	4	0	Low
O121:H19	STEC	8	82	High
O145:NM	STEC	18	77	High
O157:H7	STEC	10	0	Moderate
O174:H28	STEC	16	63	High
O88:H25	EPEC	15	0	Moderate
O187:H16	EPEC	15	2	Moderate
O157:H7	Non-pathogenic	4	44	High

¹ Persistence was defined as the sum of culturability and VBNC durations.

.

Molecular evaluation of virulence factors revealed the presence of genes encoding Shiga toxin 1 and 2, intimin, and enterohaemolysin throughout the entire duration of the experiments, including resuscitated VBNC cells.

In the quantitative biofilm production assay, all STEC strains tested were classified as non-biofilm producers, whereas among the EPEC strains, R27 was classified as a weak biofilm producer and #633 as a moderate biofilm producer (Table S3), and no association could be established between this trait and their persistence in the sand matrix.

4. Discussion

This study assessed the persistence of STEC and EPEC strains in a sand microcosm exposed to environmental stressors such as heat, desiccation, and nutrient limitation. A temperature of 37 °C as incubation temperature was selected to standardise experimental conditions and to simulate the maximum surface temperatures that sandbox sand can reach in public spaces in Argentina during summer under direct sunlight. Although E. coli is able to grow at this temperature, the combination of desiccation and nutrient limitation in the sterile sand matrix was expected to strongly limit active multiplication, making the conditions more representative of environmental persistence rather than bacterial growth.

Although previous studies have investigated E. coli survival in sand, most have focused on non-pathogenic strains or the O157:H7 serotype [19,20,21]. Under the conditions tested—sand maintained constantly at 37 °C—death curves and persistence varied according to serotype. The STEC O121:H19 strain exhibited the greatest persistence within the matrix, with a death curve lasting eight days and viability in a viable but non-culturable (VBNC) state persisting for a further 74 days. Conversely, the O111:H8 strain showed the shortest persistence, with a death curve of four days and no subsequent detection of VBNC cells. Both EPEC strains displayed death curves extending up to 15 days in the matrix. In agreement with the findings of Rumball et al. (2021) [20], who utilised sand microcosms, the non-pathogenic E. coli strain persisted for over six weeks. It should be noted that those authors employed beach sand, meaning the physicochemical conditions—particularly salinity—differ from those of the experimental model used in the present study.

Due to the limited availability of strains representing epidemiologically relevant STEC and EPEC serotypes at the time of the study, the experimental design included only one strain per serotype. Accordingly, this work should be regarded as an exploratory study and an initial approach to evaluating the persistence of different STEC and EPEC serotypes under desiccation in sand. The inclusion of multiple strains per serotype, ideally sourced from diverse origins, would enable a more comprehensive assessment of intra-serotype variability and strengthen the robustness of the conclusions.

The persistence of E. coli in sand has previously been linked to several environmental factors, chiefly the availability of carbon and nitrogen, as well as competition with the native microbiota, which are considered key determinants in the microorganism’s survival within this matrix [20]. In this study, the strains were inoculated in suspension using sterile bi-distilled water, indicating that any nutrients available for their sustenance originated solely from the sand. It should be noted that the experimental conditions involved a sterile sand matrix; consequently, the persistence results observed may differ from those expected in a natural environment. In such natural settings, interspecific interactions among microorganisms—whether competitive or cooperative—could positively or negatively affect the survival of STEC and EPEC strains. In line with this, studies employing non-sterile beach sand microcosms did not observe similar increases in the recovery of E. coli and faecal coliforms as those reported under sterile conditions [19,20].

Furthermore, survival in sand has been associated with various genetic factors, primarily genes encoding enzymes and transport proteins, which are present across most phylogroups, albeit less frequently in phylogroup B2, a group typically associated with extraintestinal pathogenic E. coli [21].

Prolonged persistence within the matrix was associated with the presence of VBNC forms. This mode of persistence was also observed in the negative control strain, but not in the STEC serotypes O111:H8 or O157:H7, nor in the EPEC strain O88:H25. With respect to STEC O157:H7, Zhao et al. (2013) [22] demonstrated that it can enter the VBNC state upon exposure to elevated CO₂ pressures. VBNC bacteria represent a subpopulation that arises in response to stress conditions [23]. Although it remains unclear whether this state constitutes a resistance strategy against adverse conditions or an intermediate phase preceding cell death [24], it has been documented in over 60 bacterial species [6].

Classical bacteriological identification of E. coli, based on standard biochemical tests, revealed no differences in the metabolic pathways evaluated between culturable viable cells and resuscitated VBNC cells. This observation aligns with previous studies indicating that bacteria in the VBNC state maintain metabolic activity [25,26,27,28].

Conversely, molecular analysis of virulence factors in the STEC and EPEC strains under study revealed the presence of genes encoding Shiga toxins 1 and 2, as well as intimin and enterohaemolysin. Although their expression was not assessed in this work, several authors have reported that VBNC cells may retain the capacity to express these factors. Oliver (2005) [25] observed that VBNC cells can produce their toxins, while Liu et al. (2010) [29], using RT-PCR, detected increased expression of stx1 and stx2 in the VBNC state compared to the culturable state.

In quantifying biofilm production at 37 °C, none of the STEC strains evaluated were classified as biofilm producers, while the two EPEC strains were classified as weak and moderate producers, respectively. Although a positive control strain for biofilm production was not included in the present assay, the negative control strain used is routinely employed in our laboratory, including in previous assays involving strains with varying biofilm production. In those assays, this strain consistently yielded low optical density values, comparable to those obtained in the present study, supporting that the assay functioned as expected. While biofilm formation has been suggested to enhance bacterial survival in sand by providing protection against adverse environmental conditions [19], no direct association was observed in this study between persistence in the matrix and the ability to form biofilm. It is important to consider that biofilm formation is influenced by multiple factors, including surface characteristics, nutrient availability, relative humidity, pH, and ambient temperature, as well as intrinsic properties of the bacterial strains, such as the presence of adhesion structures and cell-to-cell communication via quorum sensing. Therefore, the methodology employed in this study represents an approximation that may not necessarily reflect the behaviour of the strains under natural environmental conditions.

5. Conclusions

The results obtained in this study demonstrate that E. coli STEC and EPEC strains can persist in a sterile sand matrix exposed to adverse environmental conditions such as heat, desiccation, and nutrient deprivation. Variability in persistence was observed according to serotype, with some strains remaining viable in the VBNC state for extended periods, representing a potential public health risk.

The absence of detectable biofilm production under the experimental conditions suggests that other mechanisms, such as transition to the VBNC state, may play a significant role in the survival of these bacteria in dry environments. Furthermore, the detection of virulence genes in resuscitated VBNC cells underscores the importance of considering these bacterial forms in microbiological risk assessments.

Although this study was conducted under controlled conditions using a sterile sand matrix, the results offer valuable evidence of the persistence of STEC and EPEC strains in this type of substrate. Further research in natural environments could complement these findings, enabling a more comprehensive understanding of the factors that influence the survival of these pathogens in recreational settings, such as public parks. In addition, temperatures in natural environments may be higher or lower than those tested here, potentially resulting in different survival rates. Overall, these results suggest that sand may serve as a reservoir for STEC and EPEC, potentially posing a health risk in such environments. This work should be interpreted as an exploratory study. Further studies, including a larger number of strains per serotype from diverse sources, are needed to confirm and expand upon these findings.