Shiga toxin signals via ATP and its effect is blocked by purinergic receptor antagonism

Shiga toxin (Stx) is the main virulence factor of enterohemorrhagic Escherichia coli (EHEC), that cause gastrointestinal infection leading to hemolytic uremic syndrome. The aim of this study was to investigate if Stx signals via ATP and if blockade of purinergic receptors could be protective. Stx induced ATP release from HeLa cells and in a mouse model. Toxin induced rapid calcium influx into HeLa cells, as well as platelets, and a P2X1 receptor antagonist, NF449, abolished this effect. Likewise, the P2X antagonist suramin blocked calcium influx in Hela cells. NF449 did not affect toxin intracellular retrograde transport, however, cells pre-treated with NF449 exhibited significantly higher viability after exposure to Stx for 24 hours, compared to untreated cells. NF449 protected HeLa cells from protein synthesis inhibition and from Stx-induced apoptosis, assayed by caspase 3/7 activity. The latter effect was confirmed by P2X1 receptor silencing. Stx induced the release of toxin-positive HeLa cell- and platelet-derived microvesicles, detected by flow cytometry, an effect significantly reduced by NF449 or suramin. Suramin decreased microvesicle levels in mice injected with Stx or inoculated with Stx-producing EHEC. Taken together, we describe a novel mechanism of Stx-mediated cellular injury associated with ATP signaling and inhibited by P2X receptor blockade.


Detection of ATP
Detection of ATP was carried out using a bioluminescence assay. Mouse plasma was

Phosphate determination assay
HeLa cells were incubated with TRIS-buffered saline (Medicago, Uppsala, Sweden) supplemented with 1 mM CaCl 2 , 0.4 mM MgSO 4 and 20 mM HEPES for 5 min at 37°C before addition of Stx1 (1 µg/mL), A23187 calcium ionophore as the positive control (10 µM, Sigma-Aldrich) or PBS. Supernatant was collected after 40 min and phosphate reagent (Phosphate assay kit, Abcam) was added to each sample and to the control (blank) samples, containing PBS that had not been in contact with the cells, according to the manufacturer's instructions. After 30 min incubation at room temperature the absorbance of the samples was measured at 600 nm and the blank value was subtracted from each sample.
In order to rule out a direct interaction between NF449 and Stx microtiter wells were coated with NF449 (120 µM) in 0.1 M NaHCO 3 Figure S9). Stx1 A-subunit has a molecular weight of 32 kDa and its intracellular cleavage product, A 1 , is expected to have a molecular weight of 27.5 kDa. 4 To investigate whether NF449 had an effect on cleavage of the Stx1 A-subunit, Stx1 was labeled with Iodine-125 ( 125 I) using Chloramine T (1 mg/mL) as an oxidant. HeLa cells were incubated with NF449 or with PBS for 30 minutes before addition of Stx1 (7 ng/mL) together with a trace amount of Stx1-125 I for 4 hours. The cells were lysed in RIPA buffer before separation by SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was placed in an intensifying screen (Dupont Cronex Lightning Plus LE) with an X-ray film (Amersham Hyperfilm ECL, GE Healthcare) for three days at -80°C before development. In the presence or absence of NF449 there was no difference in the ratio between uncleaved and cleaved Stx1A.

Suramin
Suramin (Sigma-Aldrich) is a non-selective P2X receptor antagonist, 5 which was used for in vitro and in vivo experiments, described below.

Calcium influx assay
HeLa cells were incubated with Fluo-4 NW (Thermo Fisher Scientific) according to the manufacturer's instructions for 30 min at room temperature. In certain experiments NF449 or suramin (200 µM), was added directly into the Fluo-4 NW solution at the same time point. HeLa cells were placed in an Axio Observer.A1 microscope (Zeiss, Oberkochen, Germany) at 37°C and the cells were allowed to settle for 15 min before the start of the experiment. Each well was monitored for a total of 300 sec and images were taken at 30-sec intervals. Stx1 (1 µg/mL, this concentration was previously shown to induce calcium influx 6 ), HBSS, as the negative control, or ATP (3 µM), as the positive control, were added to the wells 30 sec after the start of the experiment and A23187 (10 µM), was added after an additional 270 sec to induce fulminant calcium influx.
Background signal was subtracted from the mean fluorescence intensity change of all cells in the field of view using ImageJ software (Version 1.48v, NIH, Bethesda).
Human platelet-rich-plasma was incubated with Fluo-4 NW, as above, and placed in a black 96-well plate with clear bottoms (Corning Inc., Corning, NY) in a GloMax Discover System at 37°C and further simultaneous stimulated with Stx1 or Stx2 (1 µg/mL) and LPS from E. coli O157 (1 µg/mL, a gift from R. Johnson, Public Health Agency, Guelph, ON, Canada), with LPS alone or with PBS. LPS was added in order to activate the platelets and enable Stx binding, as previously shown. 7,8 In experiments in which Stx2 was used apyrase (0.32 U/mL, Sigma-Aldrich) was added to enzymatically remove previously present ATP and reduce desensitization. 9 Certain wells were preincubated with NF449. Initial fluorescence values (at 475 nm and 525/25 nm excitation and emission filters) were subtracted from 2 min post-stimulation values.

Stx1B-subunit retrograde transport to the endoplasmic reticulum
Retrograde trafficking of Stx1B to the ER was detected using a previously described method 10 in which the cell is transfected with a SNAP-tag targeting the ER that covalently binds to benzylguanine conjugated to Stx1. The SNAP-tag transfected HeLa cells were treated with NF449, PBS or the intracellular calcium chelator BAPTA-AM as a control (10 µM, Thermo Fisher Scientific) for 30 min, followed by incubation with

Viability assay
HeLa cells were incubated with Stx1 or Stx2 (7 ng/mL) or PBS control for 24 h in serumfree DMEM. Certain wells were pretreated with NF449 1 h before the toxin was added.
Cell viability was assessed with Alamar Blue (Thermo Fisher Scientific) according to the manufacturer's instructions and analyzed using a GloMax Discover System. PBS-treated cells were defined as 100% viability.

Protein synthesis assay
HeLa cells were treated with NF449 or PBS for 30 min, followed by incubation with Stx1 Values are presented as counts per minute from [ 35 S] incorporated into newly synthesized protein divided by total protein concentration.

Caspase 3/7 apoptosis assay
HeLa cells were co-incubated with Stx1 (7 ng/mL) or the PBS vehicle and CellEvent Caspase-3/7 reagent (5 µM, Thermo Fisher Scientific). Certain wells were pre-treated with NF449 1 h before the toxin was added. After 24 h the cells were washed and incubated with HBSS containing 1 µg/mL NucBlue nuclear stain (Thermo Fisher Scientific) for 15 min. Cells were imaged in an Axio Observer.A1 microscope and the fluorescence emitted with a 460 -490 nm filter was measured and divided by the number of cell nuclei using ImageJ.

Isolation and detection of microvesicles from HeLa cells and platelets
HeLa cells were incubated in DMEM (Gibco, Carlsbad, CA) complemented with 0.1 % exosome-free fetal calf serum (Gibco) with or without NF449 for 1 h. Stx1B:Alexa488 (130 ng/mL) or Stx2 (200 ng/mL, both corresponding to 3 nM, a concentration that has previously been shown to cause microvesicle release from blood cells 11

Shiga toxin 2-injection mouse model
Purified Stx2 was diluted in PBS and injected intraperitoneally at a dose of 285, 142.5 or 71.25 ng/kg body weight for ATP assay and 285 ng/kg for microvesicle analysis and control mice received PBS vehicle at the same volume, as previously described. 12 Mice were monitored two to four times a day. Weight was taken daily and mice were observed for signs of disease (ruffled fur, lethargy, hunched posture, decreased activity, paralysis, tremor, ataxia and weight loss ≥ 20%), as previously described. 12 In the Stx2-injection model mice usually show symptoms from day 3 onwards. For microvesicle counts mice were sacrificed on day 3 and for ATP assay mice were sacrificed upon showing signs of symptoms or on day 7 (the defined end of the experiment). After isoflurane anesthesia blood was collected by heart puncture into citrated syringes, and the animals were sacrificed by cervical dislocation. For analysis of microvesicles blood was treated with sterile-filtered PFA 4% (Histolab) at a final concentration of 2%.

E. coli O157:H7
The Stx2-producing E. coli O157:H7 strain 86-24 (kindly provided by A. D. O'Brien, Uniformed Services University of the Health Sciences, Bethseda, MD) was previously characterized. 13 Bacteria were grown, centrifuged and resuspended to a concentration of 10 9 colony forming units (CFU)/mL, as previously described. 12 Each mouse was inoculated with 10 8 CFU in a volume of 100 µl.

Escherichia coli O157:H7-infection mouse model
Mice were infected with E. coli O157:H7 according to a previously described infection protocol. 12 Mice were monitored two to four times a day. Weight was taken daily and mice were observed for signs of disease as previously described. 12 In this mouse model mice usually develop symptoms on day 6 and onwards. Mice were sacrificed on day 3 after inoculation, before the development of symptoms, to obtain microvesicle levels at this specific time-point. Before sacrifice mice were anesthetized with isoflurane, blood was collected for microvesicle analysis as described above. All mice were tested for bacteremia using blood culture flasks (Biomérieux inc, Durham, NC) and found to be negative.

Treatment of Stx2-injected and EHEC-infected mice with Suramin
BALB/c mice were injected intraperitoneally with suramin diluted in NaCl 0.9% at a dose of 60 mg/kg bodyweight and control mice received NaCl 0.9 % (vehicle) at the same volume 16 hours before injection of Stx2 intraperitoneally or with a suramin dose of 20 mg/kg body weight, or corresponding vehicle, one hour before inoculation with E. coli O157:H7. All mice were sacrificed on day 3 post inoculation and samples were collected for microvesicle analysis.

Isolation and labeling of murine microvesicles
Microvesicles were isolated and labeled as previously described. 11 Briefly, whole blood from mice fixed in PFA was centrifuged to isolate microvesicles and to obtain a  Results are presented as positive microvesicles after subtraction of the control antibody.

Flow cytometry for detection of cells and microvesicles
A volume of 15 µl was counted for each sample on low flow rate (0.2 µl/s).

Statistical analysis
Differences between groups were assessed by the two-tailed Mann-Whitney U test, or by the Kruskal-Wallis multiple-comparison test when comparing more than two groups, followed by comparison between specific groups using the Dunn procedure. For calcium influx repeated measurements two-way repeated measures ANOVA was used. All statistical analyses were calculated using Prism 7 version 7.0a (GraphPad, La Jolla, CA).   Figure S7: Silencing of P2X1 in HeLa cells HeLa cells treated with (A) non-targeting control siRNA (siCtrl, 6 µM) (B) siRNA targeting P2X1 mRNA (siP2X1, 6 µΜ) or (C) siP2X1 (3 µM) were lysed and protein content was separated by electrophoresis. P2X1 content was detected by immunoblotting as before. Bands were detected at approximately 60 kDa, corresponding to P2X1.