Pectins and pectic-oligosaccharides, as derived by controlled enzymatic hydrolysis, were evaluated for their ability to interfere with the toxicity of Shiga-like toxins from Escherichia coli O157:H7. Both types of material resulted in some degree of protection but this was significantly higher (P>0.01) with the oligosaccharide fractions (giving 90–100% cell survival, compared to 70–80% with the polymer). An effect of methylation on the protective effect was detected with lower degrees being more active. The pectic-oligosaccharides and galabiose, the minimum toxin receptor analogue, were shown to inhibit toxicity and were both protective at 10 mg ml−1, but not at lower concentrations.
Shiga toxin-producing Escherichia coli (STEC) can cause a range of illnesses from self-limiting watery diarrhoea and haemorrhagic colitis to severe conditions such as haemolytic uraemic syndrome (HUS) and thrombotic thrombocytopoenic purpura . Apart from their ability to produce Shiga-like toxin (Stx), other virulence factors such as the adhesins intimin, E. coli secreted proteins A, B and D, the 60-MD plasmid and lipopolysaccharide are also involved [1–3]. The potential severity of diseases caused by STEC can result in a high mortality rate. For example, in 1996 in central Scotland, 21 out of >500 cases resulted in death, mostly due to HUS. The overall isolation rate of STEC in England and Wales has increased from 1468 during 1992–1994 to 3429 during 1995–1998 .
Stx produced in the intestine can be translocated to the bloodstream without cellular damage and can reach renal endothelial cells causing HUS [3,5]. However, there is a variety of data  showing that the cytotoxic effect of Stx on intestinal epithelial cells is responsible for bloody diarrhoea and it has been proposed that damage of the intestinal epithelium by Stx could aid translocation of the toxin into the bloodstream .
The presence of globotriosyl ceramide cell receptors containing the Galα1→4Galβ disaccharide mediates the attachment of Stx, as elaborated by ‘STEC’ strains of E. coli in the colon [1,6,7]. Similar types of oligosaccharide mediate the binding of P-fimbriated E. coli to uroepithelial cells [6,8]. Disruption of adhesive events either before or after attachment to host tissues will interfere with colonisation as long as the pathogen has not been internalised by the host cells. Once initial contact between bacteria and mucosal surfaces has been interrupted, subsequent events in the disease process will be aborted [9,10]. Competitive inhibition of the target interactions is therefore a rational approach to the control of infection and is currently under evaluation . Antibiotic treatment of infections is problematic since destruction of the bacteria does not reduce toxic effects. Moreover, much of the toxin remains associated to the bacterial surface and cell lysis may actually increase free Stx levels available for systemic absorption in the gut lumen .
The in vitro use of soluble receptor molecules to inhibit pathogen adhesion has been reported [12–14]. The manufacture of complete receptor saccharides with optimal conformation for inhibiting bacterial attachment is both difficult and expensive . However, anti-adhesive molecules may have potential application in the functional food market. In the case of adhesion of P-fimbriated E. coli to uroepithelial cells, the Galα1→4Galβ disaccharide did not provide optimal receptor activity, but 50% inhibition of adhesion was still achieved using a concentration of 2.13 mM. Similar results were obtained when digalacturonic acid (GalAα1→4GalA) was used as a soluble receptor molecule (50% inhibition of adhesion at 2.10 mM) .
Digalacturonic acid is the principal constituent of commercial pectins, which are principally composed of polymerised (200–1000 units), partly methylated (1→4)-linked α-d-galacturonic acid . Hydrolysis products of pectins (galacturono-oligosaccharides) have been found to block the adherence of uropathogenic P-fimbriated E. coli to uroepithelial cells in vitro [14,15], even though they are not a part of the receptor involved in recognition. In this study, we therefore investigated the use of pectin-derived oligosaccharides as an economical alternative to globoseries receptors and/or Galα1→4Galβ disaccharide for preventing gastrointestinal symptoms caused by Stx from E. coli O157:H7. The colonic cell line HT29 was used in the experiments.
2 Materials and methods
High glucose Dulbecco's modified Eagle's medium (DMEM) and fetal bovine serum (FBS) were obtained from Gibco BRL (Life Technologies, Paisley, UK). Isoton II for cell counting was obtained from Beckman Coulter (High Wycombe, UK). All other chemicals were purchased from Sigma (Poole, UK) and were cell culture grade.
High methylated citrus pectin (HMP) with 66% methylation, and low methylated apple pectin (LMP) with 8% methylation were obtained from Fluka (Sigma-Aldrich, Poole, UK). The pectic-oligosaccharides were manufactured as described previously . Two different pectic-oligosaccharides were used in the experiments: POS I, which was obtained through controlled hydrolysis from HMP, and POS II, from LMP. Galabiose (Galα1→4Galβ) was obtained from Dextra Laboratories (Reading, UK). Lactose was purchased from Sigma. Purified Stx1 and Stx2 taken from VTEC-RPLA test kits were supplied by Oxoid (Basingstoke, UK).
2.2 Cell culture
The HT29 cell line was obtained from the European Collection of Cell Cultures for Applied Microbiology and Research (ECACC, Salisbury, UK). Routinely, cell stocks were cultured at 37°C under a humidified 5% (v/v) CO2 atmosphere, in a ‘standard medium’ containing high glucose DMEM supplemented with 5% (v/v) FBS, 100 U ml−1 penicillin, 0.1 mg ml−1 streptomycin, non-essential amino acids (NEAA, ‘100×’, Sigma) and 200 mM α-glutamine. Cells were re-fed every 48 h and passaged before confluence was reached.
For all the studies carried out, cells were seeded at a low density (4×104 cells cm−2) in 24- or 96-well plates. The cells were allowed to attach overnight and then fed with standard medium the following day. Cells were rendered quiescent through removal of feeding medium and its subsequent replacement with serum-free medium, 24 h before the addition of experimental media.
2.3 Sensitivity of HT29 cells to Shiga-like toxin
Quiescent HT29 cells were treated with 1% (v/v) serum standard medium and supplemented with increasing concentrations of the Stx1 and Stx2 preparations (0.01, 0.1, 1, 10 and 100 µg protein ml−1). Cells treated with medium only were used as controls. All cells were re-fed with fresh experimental medium daily. The effect on cell growth was evaluated after 48 h treatment, by counting detached and attached cells. Detached cells were prepared for counting by removing 0.5 ml of the cell medium after gentle mixing through repeated pipetting and were added to 9.5 ml Isoton II diluent. The remaining medium was then removed by aspiration and attached cells washed twice with calcium-free phosphate-buffered saline (PBS) (pH 7, 0.1 M). Attached cells were trypsinised (0.25% trypsin/EDTA, Sigma) in a 24-well plate and 0.5 ml of the cell suspension diluted in 9.5 ml Isoton II diluent before counting. Cell numbers were determined using a Coulter counter (Beckman Coulter, High Wycombe, UK). Each toxin was tested at least three times.
2.4 Toxin neutralisation assays
Serum standard medium (1%, v/v), supplemented with different concentrations of test carbohydrates (1, 10, 100 µg ml−1, 1 and 10 mg ml−1), was pre-incubated for 1 h with Stx1 or Stx2 at 37°C. The toxin concentration used was sufficient to cause death to 50% of the cell population (ID50). Cells were then treated with experimental media and appropriate controls for 48 h. Controls comprised untreated cells, cells treated with toxins only, and those treated only with carbohydrates. Lactose addition was used as negative control since it is expected not to have any anti-adhesive properties. Spent medium was removed by aspiration and fresh experimental media were added daily.
Cytotoxicity was quantified by neutral red uptake . 200 µl of 0.1% (w/v) neutral red solution was added to all 96 wells of the culture plate and incubated for 90 min at 37°C under a 5% (v/v) CO2 atmosphere. The dye was then removed by aspiration and cells were washed twice with PBS, which had been pre-warmed to 37°C. The dye was then extracted from the cells through addition of 100 ml absolute alcohol/0.1 M citrate buffer (pH 4.2), which was mixed 1:1 to each well. The plate was then agitated for 20 min at room temperature and the absorbance read at 550 nm using an SLT Spectra plate reader (Tecan, Reading, UK).
Cytotoxic effects were recorded after 48 h incubation by comparing test sample wells with control wells that did not contain the toxin. Percentage of HT29 cell death was calculated from the equation:
2.5 Data analysis
Analysis of variance was used to test the hypothesis that means from two or more treatments were equal. When differences in the treatments were found, a two-sample Student's t-test was carried out. The differences were considered significant when P<0.05.
As expected from previously reported data , HT29 cells were sensitive to the action of Stx1 and Stx2 (Fig. 1). The number of detached cells increased in a dose–response manner, whereas the attached cell numbers decreased. The toxic effects of both Stx1 and Stx2 after 48 h incubation already gave a significant reduction (P<0.001) in attached cell numbers at 0.01 µg toxin ml−1. The ID50 was 5.25 µg protein ml−1 for Stx1 and 0.08 µg protein ml−1 for Stx2.
Susceptibility of HT29 cells to Shiga toxins. HT29 cells were incubated with varying concentrations of toxins and the adherent and detached cells counted. A: Adherent HT29 cell counts. B: Detached/adherent HT29 ratio as response to various concentrations of E. coli O157 Stx1 and Stx2 after 48 h of incubation.
The ability of pectic-oligosaccharides to neutralise Stx activity was seen at a concentration of 10 mg ml−1 (Fig. 2). In general, the sugars were more active at neutralising Stx1 than Stx2. As expected, lactose did not neutralise either Stx1 (P=0.78) or Stx2 (P=0.59) toxicity. The degree of methylation had an effect on the inhibitory activity of the sugar, with less methylated POS II having the greatest effect. POS II and LMP displayed higher inhibitory activity towards Stx1 (P=0.000008 and P=0.0002 respectively) than did POS I (P=0.001) or HMP (P=0.019). A similar effect was observed with Stx2 (P=0.03 for POS I and P=0.0006 for POS II). Neither HMP nor LMP significantly affected survival of the cells that had been treated with Stx2.
Percentage of cell survival of HT29 cells on 48 h incubation with Stx1 and Stx2 from E. coli O157. HMP, LMP, POS I, POS II and lactose were included at 10 mg ml−1. *P<0.05, significant difference compared to the toxin.
Concentrations of sugars lower than 10 mg ml−1 did not neutralise any toxic effects. Similarly, galabiose, the minimum disaccharide component of the cell glycolipid receptor responsible for toxin binding, was not inhibitory at concentrations lower than 10 mg ml−1. The percentage survival of cells treated with Stx1 and Stx2 was 29% and 44% respectively, increasing to 68% (P=0.00003) and 69% (P=0.0007) when treated with toxin and galabiose.
An investigation of the therapeutic and prophylactic potential of pectic-oligosaccharides against Stx from E. coli O157:H7 was performed using the human colonic adenocarcinoma cell line HT29. A moderate susceptibility of this cell line to Stx has been previously reported , where the ID50 after 48 h incubation for Stx1 was 1 µg ml−1. This was similar to the level seen in this study (5 µg ml−1). Differences in the purification methods and protein concentration of the toxin preparations could have explained variations in the ID50. Although Stx1 and Stx2 have similar structures and modes of action [1,2], differences in their toxicity were observed. When cells were treated with Stx1, a 50-fold higher dose was needed to reduce the cell population to 50%. These differences in toxicity have also been found in mice, where Stx2 had an LD50 that was 400 times lower than that of Stx1 . Similar results have been reported in endothelial cell studies .
The ability of pectic-oligosaccharides to neutralise Stx activity was investigated at various doses. 10 mg ml−1 was found to completely inhibit Stx1 and Stx2 whereas lower concentrations did not affect toxicity, a pattern also seen with galabiose, the minimum receptor analogue. Similar concentrations have been seen to inhibit adhesion of P-fimbriated E. coli to uroepithelial cells [14,15]. High molecular mass pectins also exhibited inhibitory activity, albeit to a lesser degree than lower molecular mass oligosaccharides. This was presumably due to increased access to receptor binding sites on the toxin.
A clear effect of degree of methylation was also seen with lower degrees producing greater inhibition. Similarly, unmethylated polygalacturonic acid was found to block adherence of P-fimbriated E. coli to uroepithelial cells .
Given the structural relationship between the digalacturonic acid and the galabiose receptor, it can be assumed that neutralisation of Stx activity was due to its binding by the pectic-oligosaccharides, thus mimicking the interaction with the galabiose receptor. This study has shown that derivatives of pectin, a fibre found commonly in plants and vegetables, have the ability to inhibit Stx of E. coli O157:H7. New food developments that exploit this interaction are likely to have important consequences for the prevention and/or alleviation of gastrointestinal symptoms caused by this toxin.
The authors would like to acknowledge Uniq Convenience Foods for their financial support of this research.
(1998) Chemical synthesis of globotriose and galabiose: relative stabilities of their complexes with Escherichia coli shiga-like toxin-1 as determined by denaturation-titration with guanidinium chloride. J. Chem. Soc. Perkin Trans. 1, 2287–2294.
(1973) The structure of plant cell walls. I. The macromolecular components of the walls of suspension-cultured sycamore cells with a detailed analysis of the pectic polysaccharides. Plant Physiol. 51, 158–173.